Split (1)

train · 901k rows

article_title stringlengths 1 1.44k	abstract stringlengths 10 22.8k	contributor stringlengths 29 38.3k	journal stringlengths 3 246	full_text stringlengths 22 2.53M
Molecular epidemiology of DFNB1 deafness in France	<sec><title>Background</title><p>Mutations in the <italic>GJB2 </italic>gene have been established as a major cause of inherited non syndromic deafness in different populations. A high number of sequence variations have been described in the <italic>GJB2 </italic>gene and the associated pathogenic effects are not always clearly established. The prevalence of a number of mutations is known to be population specific, and therefore population specific testing should be a prerequisite step when molecular diagnosis is offered. Moreover, population studies are needed to determine the contribution of <italic>GJB2 </italic>variants to deafness. We present our findings from the molecular diagnostic screening of the <italic>GJB2 </italic>and <italic>GJB6 </italic>genes over a three year period, together with a population-based study of <italic>GJB2 </italic>variants.</p></sec><sec sec-type="methods"><title>Methods and results</title><p>Molecular studies were performed using denaturing High Performance Liquid Chromatograghy (DHPLC) and sequencing of the <italic>GJB2 </italic>gene. Over the last 3 years we have studied 159 families presenting sensorineural hearing loss, including 84 with non syndromic, stable, bilateral deafness. Thirty families were genotyped with causative mutations. In parallel, we have performed a molecular epidemiology study on more than 3000 dried blood spots and established the frequency of the <italic>GJB2 </italic>variants in our population. Finally, we have compared the prevalence of the variants in the hearing impaired population with the general population.</p></sec><sec><title>Conclusion</title><p>Although a high heterogeneity of sequence variation was observed in patients and controls, the 35delG mutation remains the most common pathogenic mutation in our population. Genetic counseling is dependent on the knowledge of the pathogenicity of the mutations and remains difficult in a number of cases. By comparing the sequence variations observed in hearing impaired patients with those sequence variants observed in general population, from the same ethnic background, we show that the M34T, V37I and R127H variants can not be responsible for profound or severe deafness.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Roux</surname><given-names>Anne-Françoise</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Pallares-Ruiz</surname><given-names>Nathalie</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Vielle</surname><given-names>Anne</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Faugère</surname><given-names>Valérie</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Templin</surname><given-names>Carine</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Leprevost</surname><given-names>Dorothée</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Artières</surname><given-names>Françoise</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Lina</surname><given-names>Geneviève</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Molinari</surname><given-names>Nicolas</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Blanchet</surname><given-names>Patricia</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Mondain</surname><given-names>Michel</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>[email protected]</email></contrib><contrib id="A12" contrib-type="author"><name><surname>Claustres</surname><given-names>Mireille</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Medical Genetics	<sec><title>Background</title><p>The genetic origin of deafness is suspected in more than half of the congenital hearing loss cases. More than 400 syndromes can include hearing loss or deficient hearing functions as a component. However, non syndromic expression of deafness is observed in more than 70 % of cases. In the non syndromic forms of hearing loss (NSHL), familial or sporadic cases are observed and the transmission is predominantly autosomal recessive. Genetic heterogeneity has been established by linkage studies: more that 50 loci associated with NSHL (including dominant and recessive autosomal, and X-linked types of transmission) have been localized, making possible the identification of a number of causative deafness genes <ext-link ext-link-type="uri" xlink:href="http://dnalab-www.uia.ac.be/dnalab/hhh/"/>. Despite the extreme genetic heterogeneity, the recessive DFNB1 locus, mapping to chromosome 13q12, is by far the most prevalent. This locus contains the two Gap junction genes <italic>GJB2 </italic>and <italic>GJB6</italic>, encoding, respectively connexin 26 (CX26) and connexin 30 (CX30). These proteins associate in hexamers to form homo- and hetero-connexons [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Two connexons from adjacent cells dock to form a functional channel that will allow, among other small molecules, the diffusion of potassium ions critical for the normal sensory hair cell excitation [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>The contribution of the <italic>GJB2 </italic>gene in NSHL varies from 0 to 40 % in diverse populations [<xref ref-type="bibr" rid="B4">4</xref>] and this genetic heterogeneity is also emphasized by the variation in frequency of specific mutations among different populations. More than 70 mutations in the <italic>GJB2 </italic>gene have been reported [<xref ref-type="bibr" rid="B5">5</xref>], and although the majority are rare or private, the prevalence of four mutations define specific ethnic origins. The 35delG mutation accounts for approximately 70 % of <italic>GJB2 </italic>mutant alleles in Northern and Southern European, as well as American Caucasian populations, with a carrier frequency of 2.3 % to 4 % [<xref ref-type="bibr" rid="B6">6</xref>-<xref ref-type="bibr" rid="B9">9</xref>]. The three other mutations, 167delT, 235delC or R143W represent the most common pathogenic alleles in Ashkenazi Jews [<xref ref-type="bibr" rid="B10">10</xref>], Asian [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B14">14</xref>] and Ghanian populations [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>], respectively.</p><p>Recently, we and other groups have identified a large 309 kb deletion that includes the 5' region of the <italic>GJB6 </italic>gene and most of its coding region [<xref ref-type="bibr" rid="B17">17</xref>]. It is unclear whether this deletion removes regulatory elements common to <italic>GJB6 </italic>and <italic>GJB2 </italic>resulting, in addition to the deletion of <italic>GJB6</italic>, in reduced expression of the wild type <italic>GJB2 </italic>gene [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B20">20</xref>]. This deletion also appears to have an ethnic specific origin as it is absent from the Siberian (manuscript in preparation), Chinese [<xref ref-type="bibr" rid="B21">21</xref>], Austrian and Italian populations [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. In this report, when we refer to <italic>GJB6 </italic>mutations we will consider only this particular mutation Δ(<italic>GJB6</italic>-D13S1830).</p><p>Molecular diagnostic testing of non syndromic deafness was initiated in Montpellier in early 2000. This testing was carried out in parallel with a molecular epidemiology study of <italic>GJB2 </italic>variants in the Languedoc Roussillon region. Although many reports have estimated the carrier frequency in French and Mediterranean populations (mostly of the 35delG mutation), differences of frequency between samples are observed (0.0 to 2.7 % in France). This is essentially due to the size and composition of control samples [<xref ref-type="bibr" rid="B23">23</xref>]; for review see [<xref ref-type="bibr" rid="B4">4</xref>]. Assessment of <italic>GJB2 </italic>variant sequence distribution in Languedoc Roussillon was necessary, as we had observed significant differences in the distribution of <italic>CFTR </italic>mutations between several French regions [<xref ref-type="bibr" rid="B24">24</xref>].</p><p>In this study, we present our results from three years of molecular diagnostic testing of <italic>GJB2/GJB6 </italic>including the clinical and associated audiologic findings and also determine the prevalence and spectrum of DFNB1 mutations in the southern France population. In addition, we report the first screening of the most frequent <italic>GJB2 </italic>variants, on several thousand dried bloodspots (Guthrie cards) from newborns and thus re-evaluate the pathogenic status of some <italic>GJB2 </italic>variants.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Patients</title><p>A total of 159 unrelated families, comprised of 184 patients with sensorineural hearing loss, were referred from the Genetic Counseling Department and/or the Ear, Nose and Throat specialized clinics (Centre Hospitalier Universitaire of Montpellier and Lyon). All patients had permanent hearing loss not caused by infections, exposure to drugs or other prenatal or perinatal etiology of deafness. Informed consent was obtained for each individual. The hearing loss could be moderate to profound, bilateral or unilateral, symmetrical or asymmetrical, stable or progressive, pre or post-lingual, syndromic or non syndromic. From family histories, 79 patients were classified as sporadic cases, 54 defined as familial cases, including 33 with autosomal recessive, 19 with autosomal dominant, one with X-linked and one ambiguous modes of inheritance. Seven families were consanguineous. For 26 patients, family histories were unknown.</p><p>Among these 159 families, 84 were affected with a non syndromic, stable, bilateral, congenital, mild to profound deafness: 55 families had a single deaf child and the others, considered as familial cases, showed an autosomal recessive (24 cases), dominant (4 cases) or unclear (1) mode of inheritance. Hearing loss was moderate in 4 families, mild in 17, severe to profound in 60 and observed with variable expression in 3 families.</p></sec><sec><title>Audiological assessment</title><p>Pure-tone audiometry (PTA) was performed on every affected family member. PTA was performed for air conduction on each ear using an Interacoustics audiometer. Air-conduction thresholds were obtained at 0.5, 1, 2, 3, 4, 6 and 8 kHz. Severity of deafness was defined as mild (20–40 dB HL), moderate (41–70 dB HL), severe (71–90 dB HL) or profound hearing loss (above 90 dB HL). In young children, behavioral audiometry was used to determine the auditory thresholds in free field conditions. All subjects had an auditory brainstem response (ABR) assessment to determine the hearing loss level for high frequencies.</p></sec><sec><title>DNA extraction from patient WBC and mutation analysis</title><p>Blood samples were obtained from deaf patients, and their parents and sibs (when possible) and DNA was extracted with the Nucleon BACC3 DNA extraction kit (Amersham Pharmacia Biotech, Piscataway, NJ).</p><p>All samples were tested for the 35delG mutation, the non coding and coding exons of <italic>GJB2 </italic>(E1 and E2 respectively) and for the Δ(<italic>GJB6</italic>-D13S1830) mutation. The experimental protocols were as follows: 35delG mutation was screened by PCR-mediated site-directed mutagenesis (PSDM assay) as previously described [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. <italic>GJB2 </italic>was screened by DHPLC and sequencing analysis as previously described [<xref ref-type="bibr" rid="B25">25</xref>]. To obtain an optimal detection of mutations, some primers have been modified and are as follows: R1-F: AGTCTCCCTGTTCTGTCCTA, R3-F: TTCTCCATGCAGCGGCTGGT, R3-R: TGAGCACGGGTTGCCTCATC. The 309 kb deletion including most of the <italic>GJB6 </italic>gene, Δ(<italic>GJB6</italic>-D13S1830), was screened by two different PCRs encompassing the deletion breakpoints using the following primers PCR1: CCACCATGCGTAGCCTTAACC /GCAGCAGGTAGCACAACTCT; PCR2: CACTGAAGTGGTTTCTTGTGC /TCTGTGCTCTCTTTGATCTC, revealing breakpoint-junction fragments of 390 and 335 bp, respectively.</p></sec><sec><title>DNA extraction from dried bloodspots and mutation analysis of <italic>GJB2</italic></title><p>Guthrie cards were obtained from the GREPAM (Center for neonatal screening) of Montpellier. All samples were anonymized and no phenotypic data could therefore be available. Spots of 3 mm diameter, punched from the cards, were distributed in 96-well plates and DNA was extracted using methanol extraction [<xref ref-type="bibr" rid="B26">26</xref>]. All experiments were then set up using the robot BioMEK 2000 (Beckman). A first set of 2,777 spots were screened using the PSDM assay, specific for the 35delG mutation. When DHPLC technology became available in the laboratory, a further 3,516 spots were analyzed for the R1' fragment (covering from the ATG codon to position 230 of the coding sequence (CDS)). 528 of these 3 516 spots were also analyzed in the R2 fragment (position 190 to 500 of the CDS). Pools of 2 DNAs were systematically used for DHPLC screening, thus eliminating the risk of missing the detection of a homozygote sample. Any abnormal DHPLC profile was re-evaluated on individual DNA samples followed by sequencing analysis.</p></sec><sec><title>Statistical analysis</title><p>Statistical analyses were performed with the Rv.1.3.1 software (The free software Fundation, Inc). Proportion, chi-square and Fisher exact test were used to test differences between groups. All p-values were taken to be significant at <0.05. When observed or expected values were below 5, a Fisher exact test was performed.</p></sec></sec><sec><title>Results</title><sec><title>Patients</title><p>The analysis of the coding and non coding <italic>GJB2 </italic>exons plus the Δ(<italic>GJB6</italic>-D13S1830) screening allowed us to genotype 30 unrelated individuals with biallelic <italic>GJB2 </italic>and/or <italic>GJB6 </italic>(Δ(<italic>GJB6</italic>-D13S1830)) mutations (Table <xref ref-type="table" rid="T1">1</xref>). Clinical and audiological evaluation showed that 27 of these patients with biallelic DFNB1 mutations had bilateral congenital severe or profound NSHL with no evidence of progressive phenotype. No mutation was identified among families with an autosomal dominant mode of inheritance.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>DFNB1 genotypes identified in 30 patients with deafness</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Number of families</bold></td><td align="center"><bold>Genotype</bold></td><td align="center"><bold>Degree of deafness</bold></td><td align="center"><bold>Mode of inheritance*</bold></td></tr></thead><tbody><tr><td align="center">10</td><td align="center">35delG/35delG</td><td align="center">10 profound</td><td align="center">5 SC; 4 AR; 1cAR; 1un</td></tr><tr><td align="center">2</td><td align="center">35delG/E47X</td><td align="center">2 profound</td><td align="center">2AR</td></tr><tr><td align="center">2</td><td align="center">35delG/312del14</td><td align="center">1 profound, 1 moderate</td><td align="center">1AR</td></tr><tr><td align="center">1</td><td align="center">35delG/N206S</td><td align="center">1 moderate</td><td align="center">1AR</td></tr><tr><td align="center">1</td><td align="center">35delG/R184P</td><td align="center">1 profound</td><td align="center">1 SC</td></tr><tr><td align="center">1</td><td align="center">35delG/W24X</td><td align="center">1 profound</td><td align="center">1AR</td></tr><tr><td align="center">1</td><td align="center">35delG/C64X</td><td align="center">1 profound</td><td align="center">1 SC</td></tr><tr><td align="center">1</td><td align="center">35delG/delE120</td><td align="center">1 moderate</td><td align="center">1AR</td></tr><tr><td align="center">1</td><td align="center">35delG/Q57X</td><td align="center">1 profound</td><td align="center">1SC</td></tr><tr><td align="center">1</td><td align="center">35delG/R143W</td><td align="center">1 profound</td><td align="center">1SC</td></tr><tr><td align="center">1</td><td align="center">35delG/W44X</td><td align="center">1 profound</td><td align="center">1SC</td></tr><tr><td align="center">1</td><td align="center">290insA/IVS1+1G>A</td><td align="center">1 profound</td><td align="center">1AR</td></tr><tr><td colspan="4"><hr></hr></td></tr><tr><td align="center">3</td><td align="center">35delG/Δ(GJB6/D13S1830)</td><td align="center">3 profound</td><td align="center">1 SC; 1AR; 1un</td></tr><tr><td align="center">2</td><td align="center">Δ(GJB6/D13S1830)/Δ(GJB6/D13S1830)</td><td align="center">2 profound</td><td align="center">1 SC; 1 AR</td></tr><tr><td align="center">1</td><td align="center">E47X/Δ(GJB6/D13S1830)</td><td align="center">1 profound</td><td align="center">1un</td></tr><tr><td align="center">1</td><td align="center">235delC/Δ(GJB6/D13S1830)</td><td align="center">1 profound</td><td align="center">1 AR</td></tr></tbody></table><table-wrap-foot><p>top: genotypes of the 23 familieswith GJB2 biallelic mutations; bottom: 7 families with DFNB1 mutations All cases presented bilateral, congenital hearing loss with no evidence of progression * novel mutation SC: Sporadic case; AR: autosomal recessive; cAR: autosomal recessive with consanguinity; un: unknown</p></table-wrap-foot></table-wrap><p>Sixteen additional patients carried 2 <italic>GJB2 </italic>mutations/variants with a controversial pathogenic effect or a single <italic>GJB2 </italic>sequence variation (Table <xref ref-type="table" rid="T2">2</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>GJB2 genotypes with unknown consequences in 16 families</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Number of families</bold></td><td align="center"><bold>Genotype</bold></td><td align="center"><bold>Hearing loss**</bold></td><td align="center"><bold>Mode of inheritance</bold></td></tr></thead><tbody><tr><td align="center">3</td><td align="center">35delG/+</td><td align="center">1P progressive; 1M; 1U</td><td align="center">1SC,1AR, 1un</td></tr><tr><td align="center">2</td><td align="center">R127H/M34T</td><td align="center">2P</td><td align="center">1AR,1SC</td></tr><tr><td align="center">1</td><td align="center">R127H/W24X</td><td align="center">P to M </td><td align="center">1AR</td></tr><tr><td align="center">1</td><td align="center">M34T/V37I</td><td align="center">1 mild</td><td align="center">1SC</td></tr><tr><td align="center">2</td><td align="center">M34T/+</td><td align="center">1 mild, 1M</td><td align="center">2AD</td></tr><tr><td align="center">1</td><td align="center">(IVS1-12C>T)2/-34T/G</td><td align="center">P</td><td align="center">un</td></tr><tr><td align="center">2</td><td align="center">IVS1-12C>T/+</td><td align="center">1P progressive, 1P</td><td align="center">1SC, 1un</td></tr><tr><td align="center">2</td><td align="center">V37I/+</td><td align="center">1M, 1P</td><td align="center">1AR, 1un</td></tr><tr><td align="center">1</td><td align="center">V153I/+</td><td align="center">M</td><td align="center">1SC</td></tr><tr><td align="center">1</td><td align="center">G160S/+</td><td align="center">P</td><td align="center">1SC</td></tr></tbody></table><table-wrap-foot><p>* novel sequence variant + designates the wild type allele * variable phenotype within the family **** <bold>P: </bold>profound; M: moderate; U: unilateral; un: unknown</p></table-wrap-foot></table-wrap><p>A total of 21 different <italic>GJB2 </italic>sequence variations (1 in frame deletion, 5 nonsense, 4 frameshift, 8 missense, 1 splicing, 1 in the 5' untranslated region (UTR) and 1 in the intron 1(IVS1)) were found in 46 unrelated subjects from the cohort presented here. All were previously reported with the exception of the C64X (c.192 C>A) mutation, the -34T>G and IVS1-12C>T sequence variations (Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref>). Ten subjects were 35delG homozygotes, 2 were Δ(<italic>GJB6</italic>-D13S1830) homozygotes. All the other patient mutations were found in the compound heterozygous state, with the 35delG accounting for 35/60 (58.3 %) of the <italic>GJB2</italic>/<italic>GJB6 </italic>mutated alleles (Table <xref ref-type="table" rid="T1">1</xref>).</p><p>Δ(<italic>GJB6</italic>-D13S1830) is the second most frequent mutation (9/60) and accounts for 15 % of the mutated alleles: in addition to the 2 homozygotes patients, 5 patients were identified as compound heterozygotes Δ(<italic>GJB6</italic>-D13S1830) /<italic>GJB2 </italic>mutation.</p><p>We assessed the polymorphism 765C>T (referred to as SNP1 [<xref ref-type="bibr" rid="B27">27</xref>]) in 159 unrelated patients (see Table <xref ref-type="table" rid="T3">3</xref>). The frequencies of the genotypes 765C>C, C>T, T>T in control samples (with no <italic>GJB2 </italic>sequence variations, N = 113) are 57.5 %, 39.8 % and 2.7 % respectively. However, all alleles carrying the 35delG mutation (n = 38) and the 309 kb <italic>GJB6 </italic>deletion (n = 9) were associated with the 765T variant.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Distribution of the C765T polymorphism among patients</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left" colspan="5"><bold>35delG</bold></td></tr></thead><tbody><tr><td></td><td align="center"><bold>C/C</bold></td><td align="center"><bold>C/T</bold></td><td align="center"><bold>T/T</bold></td><td align="center"><bold>p-values</bold></td></tr><tr><td colspan="5"><hr></hr></td></tr><tr><td align="center"><bold>Patients 35delG/35delG</bold></td><td align="center">0</td><td align="center">0</td><td align="center">10</td><td align="center">p < 10<sup>-6</sup></td></tr><tr><td align="center"><bold>Patients 35delG/other</bold></td><td align="center">0</td><td align="center">13</td><td align="center">5</td><td></td></tr><tr><td align="center"><bold>Controls</bold></td><td align="center">65</td><td align="center">45</td><td align="center">3</td><td></td></tr><tr><td align="left" colspan="5"><bold>Δ(GJB6/D13S1830)</bold></td></tr><tr><td colspan="5"><hr></hr></td></tr><tr><td></td><td align="center"><bold>C/C</bold></td><td align="center"><bold>C/T</bold></td><td align="center"><bold>T/T</bold></td><td align="center"><bold>p-values</bold></td></tr><tr><td colspan="5"><hr></hr></td></tr><tr><td align="center"><bold>Patients Δ(GJB6/D13S1830)/Δ(GJB6/D13S1830)</bold></td><td align="center">0</td><td align="center">0</td><td align="center">2</td><td align="center">p < 10<sup>-6</sup></td></tr><tr><td align="center"><bold>Patients Δ(GJB6/D13S1830)/other</bold></td><td align="center">0</td><td align="center">3</td><td align="center">2</td><td></td></tr><tr><td align="center"><bold>controls</bold></td><td align="center">65</td><td align="center">45</td><td align="center">3</td><td></td></tr></tbody></table><table-wrap-foot><p>Controls: patients with no <italic>GJB2 </italic>mutation. Chi-square test was used to test the independence between the genotype 765T/T and the control group.</p></table-wrap-foot></table-wrap></sec><sec><title>Parents of deaf individuals</title><p>No <italic>de novo </italic>mutations were detected in patients, as every parent of patients carrying bi-allelic <italic>GJB2</italic>/<italic>GJB6 </italic>mutations was heterozygous for one of the mutations. In two unrelated families, 2 normal-hearing parents of children genotyped R127H/M34T or R127H/W24X, were found to be homozygous R127H. Interestingly, in each of these families, a normal-hearing sib also carried the genotype R127H/M34T or R127H/W24X. Thus, R127H may not be associated with deafness.</p></sec><sec><title>Analysis in the general population</title><p>Twenty-two different <italic>GJB2 </italic>sequence variations (lying in R1' and R2) were identified in the general population and are listed in table <xref ref-type="table" rid="T4">4</xref>, together with the calculated carrier frequencies. Ten of these sequence variations were detected more than once (35delG, M34T, V37I, V27I, W24X, E47X, Y68C, R127H, V153I, F83L), with relative frequencies ranging between 2,3 % to 0,09 % and 7 of these were not previously reported (Y68C, IVS1-7G/A, G4D, Q7Q, T26T, H67R, D159D).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Frequencies of the sequence variations identified in the general population from Languedoc-Roussillon.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold><italic>Name</italic></bold></td><td align="center"><bold>Nucleotide change</bold></td><td align="center"><bold>Number of alleles</bold></td><td align="center"><bold>Number of chromosomes tested</bold></td><td align="center"><bold>Allele frequencies in % (95 % CI)</bold></td><td align="center"><bold>Carrier frequencies in % (95 % CI)</bold></td></tr></thead><tbody><tr><td align="center"><bold>M34T</bold></td><td align="center">101T>C</td><td align="center">81</td><td align="center">7 032</td><td align="center">1.15 (0.92–1.44)</td><td align="center">2.3 (1.83–2.86)</td></tr><tr><td align="center"><bold>35delG</bold></td><td align="center">c.35delG</td><td align="center">96</td><td align="center">12 586</td><td align="center">0.76 (0.62–0.93)</td><td align="center">1.53 (1.24–1.85)</td></tr><tr><td align="center"><bold>R127H</bold></td><td align="center">380G>A</td><td align="center">7</td><td align="center">1 056</td><td align="center">0.66 (0.29–1.42)</td><td align="center">1.33 (0.58–2.82)</td></tr><tr><td align="center"><bold>V37I</bold></td><td align="center">109G>A</td><td align="center">30</td><td align="center">7 032</td><td align="center">0.43 (0.29–0.62)</td><td align="center">0.85 (0.58–1.24)</td></tr><tr><td align="center"><bold>V153I</bold></td><td align="center">457G>A</td><td align="center">4</td><td align="center">1 056</td><td align="center">0.38 (0.12–1.04)</td><td align="center">0.76 (0.14–1.79)</td></tr><tr><td align="center"><bold>F83L</bold></td><td align="center">249C>G</td><td align="center">3</td><td align="center">1 056</td><td align="center">0.28 (0.07–0.90)</td><td align="center">0.57 (0.14–1.79)</td></tr><tr><td align="center"><bold>V27I</bold></td><td align="center">79G>A</td><td align="center">10</td><td align="center">7 032</td><td align="center">0.14 (0.07–0.27)</td><td align="center">0.28 (0.14–0.58)</td></tr><tr><td align="center"><bold>E114G</bold></td><td align="center">341A>G</td><td align="center">1</td><td align="center">1 056</td><td align="center">0.09 (0.005–0.61)</td><td align="center">0.19 (0.009–1.22)</td></tr><tr><td align="center"><bold>DelE120</bold></td><td align="center">358-360delGAG</td><td align="center">1</td><td align="center">1 056</td><td align="center">0.09 (0.005–0.61)</td><td align="center">0.19 (0.009–1.22)</td></tr><tr><td align="center"><bold>G160S</bold></td><td align="center">478G>A</td><td align="center">1</td><td align="center">1 056</td><td align="center">0.09 (0.005–0.61)</td><td align="center">0.19 (0.009–1.22)</td></tr><tr><td align="center"><bold>D159D</bold></td><td align="center">477C>T</td><td align="center">1</td><td align="center">1 056</td><td align="center">0.09 (0.005–0.61)</td><td align="center">0.19 (0.009–1.22)</td></tr><tr><td align="center"><bold>W24X</bold></td><td align="center">71G>A</td><td align="center">5</td><td align="center">7 032</td><td align="center">0.07 (0.03–0.18)</td><td align="center">0.14 (0.06–0.36)</td></tr><tr><td align="center"><bold>E47X</bold></td><td align="center">139G>T</td><td align="center">4</td><td align="center">7 032</td><td align="center">0.06 (0.02–0.16)</td><td align="center">0.11 (0.04–0.32)</td></tr><tr><td align="center"><bold>Y68C</bold></td><td align="center">203A>G</td><td align="center">3</td><td align="center">7 032</td><td align="center">0.04 (0.01–0.14)</td><td align="center">0.09 (0.02–0.28)</td></tr><tr><td align="center"><bold>W44X</bold></td><td align="center">132G>A</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>R32H</bold></td><td align="center">95G>A</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>S19T</bold></td><td align="center">56G>C</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>IVS1-7G>A</bold></td><td></td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>G4D</bold></td><td align="center">11G>A</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>Q7Q</bold></td><td align="center">21G>A</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>T26T</bold></td><td align="center">78C>T</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr><tr><td align="center"><bold>H67R</bold></td><td align="center">200A>G</td><td align="center">1</td><td align="center">7 032</td><td align="center">0.014 (0.007–0.092)</td><td align="center">0.03 (0.01–0.18)</td></tr></tbody></table><table-wrap-foot><p>: novel sequence changes.</p></table-wrap-foot></table-wrap><p>The M34T variant is the most frequent, with a carrier frequency of 1/43 (2.3 %). In addition, among the 3 516 dried bloodspots screened, one homozygote M34T/M34T was detected.</p><p>The 35delG mutation was screened in a total of 6,293 newborns (2,777 were analyzed by PSDM and 3,516 by PCR-DHPLC). Ninety-two 35delG heterozygotes and 2 35delG homozygotes were identified, resulting in a carrier frequency of 1/66 (1.53 %) for 35delG.</p><p>In addition, 3 individuals were compound heterozygotes (M34T/35delG, V37I/G160S and V27I/E114G) for <italic>GJB2 </italic>sequence variants.</p><p>We also present the allelic frequencies of some <italic>GJB2 </italic>mutations in individuals referred for NSHL with respect to the frequency in the general population (Table <xref ref-type="table" rid="T5">5</xref>). We could not detect any significant difference between the two groups for the V37I, R127H and M34T sequence variations.</p><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>Comparison of the allelic frequencies of GJB2 sequence variations in unrelated french patients with NSHL and in the general population</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>GJB2 Sequences variations</bold></td><td align="center" colspan="4"><bold>Number of unrelated chromosomes</bold></td><td align="center">Significance (p-values)</td></tr><tr><td></td><td colspan="4"><hr></hr></td><td></td></tr><tr><td></td><td align="center" colspan="2"><bold>patients (n = 318)</bold></td><td align="center" colspan="2"><bold>general population</bold></td><td></td></tr></thead><tbody><tr><td align="center">35delG</td><td align="center">38</td><td align="center">11.9 %</td><td align="center">96/12586</td><td align="center">0.76 %</td><td align="center">S (p < 10<sup>-6</sup>)</td></tr><tr><td align="center">M34T</td><td align="center">5</td><td align="center">1.6 %</td><td align="center">81/7034</td><td align="center">1.15 %</td><td align="center">NS (p = 0.66)</td></tr><tr><td align="center">R127H</td><td align="center">3</td><td align="center">0.94 %</td><td align="center">7/1056</td><td align="center">0.66 %</td><td align="center">NS (p = 0.89)</td></tr><tr><td align="center">V37I</td><td align="center">3</td><td align="center">0.94 %</td><td align="center">30/7034</td><td align="center">0.43 %</td><td align="center">NS (p= 0.36)</td></tr></tbody></table><table-wrap-foot><p>-: not studied; S: significant difference; NS: no significant difference. P-values significant at <0.05</p></table-wrap-foot></table-wrap></sec></sec><sec><title>Discussion</title><sec><title>The 35delG mutation in patients and in general population</title><p>Thirty families were clearly genotyped with causative mutations in <italic>GJB2</italic>/<italic>GJB6</italic>. Homozygosity for 35delG was found in 33.3 % (10/30) of the genotyped unrelated deaf patients, 50 % were carrying this deletion in a compound heterozygous state and 16.7 % had other mutations. The 35delG accounts for 58.3 % (35/60) of the DFNB1 mutated alleles in these families. As in previous studies, this study shows an important implication of <italic>GJB2 </italic>in non syndromic prelingual hearing loss. However, we have observed a lower frequency of the 35delG mutant allele and a higher heterogeneity of other mutations than in previous studies [<xref ref-type="bibr" rid="B28">28</xref>-<xref ref-type="bibr" rid="B30">30</xref>].</p><p>Our study region (Languedoc Roussillon) shows a significantly higher carrier rate of the 35delG (1.53 % – 1/66), compared to the North-East part of Europe (0.9 % – 1/110 among 1,212 controls) [<xref ref-type="bibr" rid="B31">31</xref>] and a lower carrier rate compared to other south European areas such as Spain (2.31 % – 1/43) [<xref ref-type="bibr" rid="B7">7</xref>], Italy (3.45 % – 1/32) [<xref ref-type="bibr" rid="B23">23</xref>] and Greece (3.54 % – 1/28) [<xref ref-type="bibr" rid="B32">32</xref>]. The epidemiologic study presented here is based on the largest number of random samples describing an unbiased general population screen and once more supports the heterogeneous composition of the Languedoc Roussillon population. This situation has direct implications for genetic counseling as well as on the development of potential diagnostic kits (as it was in the case for the design of the CF neonatal screening kit [<xref ref-type="bibr" rid="B24">24</xref>]). The explanation for such a different carrier rate lies in the heterogeneity of the migrations "landing" in Languedoc Roussillon, historically and still today. A recent study, carried out on the prevalence of Hemochromatosis gene (HFE) mutations in the Languedoc Roussillon population has annotated the origins of the four grand-parents for each newborn. It was observed that the population originated from various regions in France and also from other European or African countries [<xref ref-type="bibr" rid="B33">33</xref>].</p></sec><sec><title>Other sequence variations in the GJB2 gene</title><p>35delG remains by far the most frequent mutation in the hearing impaired population, although its frequency is lower than was expected in comparison with other Mediterranean areas. The second most common mutation in our region is the Δ(<italic>GJB6</italic>-D13S1830) with 15 % of the mutated alleles, and the E47X mutation represents 5 %. The other mutations were identified twice or only once.</p><p>This genetic heterogeneity is emphasized by the number of sequence variations observed in both the patient and general population. Ninety variations have been reported [<xref ref-type="bibr" rid="B5">5</xref>]. In this study, besides 35delG and Δ(<italic>GJB6</italic>-D13S1830), we describe 33 sequence variations, 10 for the first time (-34T>G, IVS-12C>T, IVS1-7G>A, G4D, Q7Q, T26T, C64X, H67R, Y68C, D159D tables <xref ref-type="table" rid="T1">1</xref>, <xref ref-type="table" rid="T2">2</xref>, <xref ref-type="table" rid="T4">4</xref>). The pathogenicity of 11 of these has been previously well established (E47X, 312del14, 290insA, IVS1+1G>A, W24X, W44X, delE120, Q57X, R143W, W44X, 235delC). As well, the novel mutation C64X described in this study results in a truncated protein with pathogenic consequences. Of the 21 other sequence variations observed, four were previously reported as recessive mutations (R32H, S19T, N206S, R184P), five are known non pathogenic variants (V27I, F83L, E114G, V153I, G160S), three of them are silent (D159D, Q7Q, T26T), six have unknown consequences (-34T>G, IVS1-12C>T, IVS1-7G>A, G4D, H67R, Y68C) and three are still controversial (M34T, V37I, R127H and see below). Three of the newly identified sequence variations are in the noncoding region. According to splicing prediction programs, both IVS variations should be silent. Interestingly, the IVS1-12C>T was found in a homozygous state in a patient originating from Guadeloupe with one IVS1-12C>T variation allelic to the -34T>C variation. This T to C transition has yet unknown consequences on the regulation of the <italic>GJB2 </italic>gene.</p></sec><sec><title>Founder effects and sequence variations</title><p>A founder effect for the 35delG mutation has recently been described [<xref ref-type="bibr" rid="B27">27</xref>], which explains the variable frequency of this mutation in different populations rather than resulting from a mutational hot spot. Similarly, the 235delC mutation, frequent in the Japanese population, is derived from a common ancestor [<xref ref-type="bibr" rid="B34">34</xref>]. Founder effects may account for a number of other <italic>GJB2 </italic>sequence variations, whose frequencies depend on ethnic background (such as M34T [<xref ref-type="bibr" rid="B4">4</xref>]). The polymorphism 765C>T (referred to as SNP1 [<xref ref-type="bibr" rid="B27">27</xref>]) has been systematically included in our series. The 765T allele showed complete association with the 35delG as well as the Δ(<italic>GJB6</italic>-D13S1830). The genotype comparison between <italic>GJB2 </italic>and/or Δ(<italic>GJB6</italic>-D13S1830) patients with <italic>GJB2 </italic>negative subjects is significant with P-values < 10<sup>-6 </sup>(Table <xref ref-type="table" rid="T3">3</xref>). These results are in accordance with the fact that Δ(<italic>GJB6</italic>-D13S1830) is absent in some populations and strongly suggest the existence of a founder effect as confirmed by a recent multicenter study [<xref ref-type="bibr" rid="B17">17</xref>].</p></sec><sec><title>Controversial effects of sequence variations</title><p>Among the sequence variations that have controversial consequences, a few variants have been extensively discussed based on families with deafness and on studies using <italic>in vitro </italic>expression systems. Since we have performed a population based study, in parallel with the analysis of a number of hearing impaired patients, we show that M34T, V37I and R127H represent common variants that are not responsible for severe or profound deafness.</p><p>The M34T variation was first described as a dominant mutation [<xref ref-type="bibr" rid="B35">35</xref>] and the dominant negative effect was supported by <italic>in vitro </italic>functional studies [<xref ref-type="bibr" rid="B36">36</xref>,<xref ref-type="bibr" rid="B37">37</xref>]. However, the description of normal hearing carriers abolished this hypothesis and furthermore, normal hearing patients were found to be compound heterozygotes, M34T/167delT or M34T/35delG [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B38">38</xref>]. Since the two alleles <italic>in trans </italic>of the M34T corresponded to a null allele, Griffith et al. (2000) suggested that the M34T was functional <italic>in vivo </italic>and therefore, the phenotypic consequences of the M34T allele would depend on the opposing CX26 allele variant [<xref ref-type="bibr" rid="B39">39</xref>]. The possibility of considering the M34T as a non pathogenic variant was also raised [<xref ref-type="bibr" rid="B40">40</xref>,<xref ref-type="bibr" rid="B41">41</xref>].</p><p>M34T represents the highest carrier rate in our population. We do not deal with a carrier rate of 1/116 as estimated from a small sample in Paris [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B41">41</xref>] but with 1/43 (2.3 %). We identified one M34T homozygote in the general population, as expected from the carrier frequency (1/4,444). The M34T is more frequent than the 35delG and, in contrast to the 35delG mutation, no M34T homozygote or in compound heterozygosity with a deleterious mutation was observed in our cohort of patients. This carrier rate is similar to the one observed by Green et al. [<xref ref-type="bibr" rid="B9">9</xref>]. The allele frequency of 1.15 %, based on 3,516 individuals, shows no significant discrepancy with the M34T allele frequency in the deaf population (1.6 %). Although a study based on the general population does not rule out the possibility of hearing deficiency in few individuals, these data eliminate the possibility of considering the M34T as a dominant or recessive mutation associated with severe or profound deafness.</p><p>The R127H, first described by Estivill [<xref ref-type="bibr" rid="B7">7</xref>] is also contentious and functional studies of this variant are inconsistent [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. The frequency of carrier rate of in our region is 1/75 (1.33 %), not significantly different from that of the deaf population. Moreover, two normal-hearing parents were genotyped R127H/R127H.</p><p>Finally two normal-hearing sibs were compound heterozygotes R127H/M34T or R127H/W24X emphasizing the non pathogenic nature of this sequence variation. However, the genotype R127H/M34T was identified twice in our patient cohort (2/159 1.25 %) but never in the general population (odds to be associated randomly of 1.5 × 10<sup>-4</sup>). The observed frequency in the patients is significant (p = 1.3 %), and therefore we still cannot rule out that the combined genotype R127H/M34T can act as a variant that, under certain circumstances (associated with other modifiers such as alterations in other deafness genes), would contribute to the phenotype.</p><p>Similarly, the V37I variation does not show any significant difference in frequency between the general and deaf populations. This is consistent with the fact that it was originally identified as a non pathogenic polymorphism because of its occurrence in a heterozygous state in the general population [<xref ref-type="bibr" rid="B40">40</xref>,<xref ref-type="bibr" rid="B44">44</xref>]. However, homozygosity for V37I and compound heterozygosity of V37I were often described in patients with NSHL suggesting that V37I acts as a recessive mutation. Recent studies [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B45">45</xref>,<xref ref-type="bibr" rid="B46">46</xref>] clearly indicated its pathogenicity when associated with another mutated <italic>GJB2 </italic>allele. Functional analyses [<xref ref-type="bibr" rid="B47">47</xref>] also showed that V37I is devoid of functional activity and thus may be pathogically significant.</p><p>The likelihood of M34T to be associated with V37I is 1 × 10<sup>-4</sup>, as calculated from the allele frequency, and is significantly different from the observed frequency in patients (p = 0.63 %), and once more can not rule out that the genotype M34T/V37I is not associated with mild or moderate deafness.</p><p>These data demonstrate the challenge of interpretating the association of two sequence variations. Moreover, combined genotypes with variants such as M34T, V37I, or R127H could have a phenotypic expression modulated by environmental factors or modifier genes. Conclusions from <italic>in vitro </italic>transfection assays can not be taken as complete because one sequence variant supposedly acting as a recessive mutation should be at least co-transfected with a second recessive mutation; but above all, one wonders whether the different cell types used for these <italic>in vitro </italic>experiments really reflect the molecular context existing <italic>in vivo</italic>. Complementary assays considering co-transfection of two different mutations need to be performed. These, together with compilation of observed combined genotypes correlated with phenotypes, should help in the interpretation of the molecular tests.</p></sec><sec><title>Phenotypes associated with <italic>GJB2/GJB6 </italic>mutations</title><p>Although a recent report recommended <italic>GJB2 </italic>screening in cases of progressive and recurrent sudden HL [<xref ref-type="bibr" rid="B31">31</xref>]), no <italic>GJB2 </italic>and/or <italic>GJB6 </italic>biallelic mutations were identified in patients with progressive, postlingual, asymmetrical hearing loss in this study. The contribution of the <italic>GJB2/GJB6 </italic>genes, in our cohort is exclusively found in non syndromic, prelingual, bilateral, stable deafness and is about 33 % (28/84). This is similar to the proportion found in Spain (31.6 %) [<xref ref-type="bibr" rid="B28">28</xref>], in France (39.8 %) [<xref ref-type="bibr" rid="B48">48</xref>] and in Greece (33.3 %) [<xref ref-type="bibr" rid="B30">30</xref>]. However the degree of implication of <italic>GJB2/GJB6 </italic>genes in deafness depends on the composition in degree of severity of the patient group. In this study, the rate of DFNB1-associated deafness would be 41.6 % (25/60) if only congenital profound NSHL was considered.</p></sec></sec><sec><title>Conclusions</title><p>The data presented here demonstrate that genetic counseling has been greatly improved with the identification of the Δ(<italic>GJB6</italic>-D13S1830) mutation and a total of 30 families could benefit from a <italic>GJB2 </italic>and/or <italic>GJB6 </italic>unambiguous molecular diagnosis. The genetic counseling was more difficult for 16 families (1/3) because of the compound heterozygosity of poorly-defined variations or the presence of a single <italic>GJB2 </italic>alteration. Additional data certainly need to be collected to evaluate if two sequence variations considered as non-disease causing stay neutral when associated in trans.</p><p>The number of families (129) for which no genetic counseling could be provided, based on the mutation screening of the DFNB1 genes, remains very high and demonstrates that criteria for genetic testing must be very well defined and set according to the provided test (CX26). Genetic heterogeneity is still observed when prelingual, non syndromic, stable deafness is present, and the possibility for screening mutations in other genes should be offered.</p><p>Finally, neonatal screening programs bring a great improvement in the management of deafness; however, the option to offer molecular based screening remains very uncertain and if offered should be based on specific population studies because of the genetic heterogeneity of deafness.</p></sec><sec><title>Competing interests</title><p>There are no competing interests</p></sec><sec><title>Authors' contributions</title><p>FA, PB, GL and MD referred patients; NPR, AV, VF, CT and DL carried out the molecular studies; NM performed the statistical analyses; AFR supervised the whole study in the laboratory of MC.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2350/5/5/prepub"/></p></sec>
Correction: Peak plasma interleukin-6 and other peripheral markers of inflammation in the first week of ischaemic stroke correlate with brain infarct volume, stroke severity and long-term outcome	Could not extract abstract	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Smith</surname><given-names>Craig J</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Emsley</surname><given-names>Hedley CA</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Gavin</surname><given-names>Carole M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Georgiou</surname><given-names>Rachel F</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Vail</surname><given-names>Andy</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Barberan</surname><given-names>Elisa M</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>del Zoppo</surname><given-names>Gregory J</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Hallenbeck</surname><given-names>John M</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>[email protected]</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Rothwell</surname><given-names>Nancy J</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>[email protected]</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Hopkins</surname><given-names>Stephen J</given-names></name><xref ref-type="aff" rid="I8">8</xref><email>[email protected]</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Tyrrell</surname><given-names>Pippa J</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Neurology	<p>In table 3, the correlation coefficient between peak plasma cortisol and mRS at 3 months (column 4, row 5), should read 0.48, not 0 [<xref ref-type="bibr" rid="B1">1</xref>].</p><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2377/4/5/prepub"/></p></sec>
Absence of annexin I expression in B-cell non-Hodgkin's lymphomas and cell lines	<sec><title>Background</title><p>Annexin I, one of the 20 members of the annexin family of calcium and phospholipid-binding proteins, has been implicated in diverse biological processes including signal transduction, mediation of apoptosis and immunosuppression. Previous studies have shown increased annexin I expression in pancreatic and breast cancers, while it is absent in prostate and esophageal cancers.</p></sec><sec><title>Results</title><p>Data presented here show that annexin I mRNA and protein are undetectable in 10 out of 12 B-cell lymphoma cell lines examined. Southern blot analysis indicates that the annexin I gene is intact in B-cell lymphoma cell lines. Aberrant methylation was examined as a cause for lack of annexin I expression by treating cells 5-Aza-2-deoxycytidine. Reexpression of annexin I was observed after prolonged treatment with the demethylating agent indicating methylation may be one of the mechanisms of annexin I silencing. Treatment of Raji and OMA-BL-1 cells with lipopolysaccharide, an inflammation inducer, and with hydrogen peroxide, a promoter of oxidative stress, also failed to induce annexin I expression. Annexin I expression was examined in primary lymphoma tissues by immunohistochemistry and presence of annexin I in a subset of normal B-cells and absence of annexin I expression in the lymphoma tissues were observed. These results show that annexin I is expressed in normal B-cells, and its expression is lost in all primary B-cell lymphomas and 10 of 12 B-cell lymphoma cell lines.</p></sec><sec><title>Conclusions</title><p>Our results suggest that, similar to prostate and esophageal cancers, annexin I may be an endogenous suppressor of cancer development, and loss of annexin I may contribute to B-cell lymphoma development.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Vishwanatha</surname><given-names>Jamboor K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Salazar</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Gopalakrishnan</surname><given-names>Velliyur K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Cancer	<sec><title>Background</title><p>The Annexins comprise a family of 20 calcium- and phospholipid-binding proteins. Expressed in organisms ranging from molds and plants to mammals, this family of proteins has proven evolutionarily conserved as well as functionally diverse. Structurally, annexins consist of a 70 amino acid core domain and an N-terminal domain, which is variable in both length and sequence, and imparts upon the family its functional diversity. Annexin I has been implicated to have a biological role in inhibition of phospholipase A2 [<xref ref-type="bibr" rid="B1">1</xref>], as a substrate for epidermal growth factor receptor [<xref ref-type="bibr" rid="B2">2</xref>] and intracellular calcium release [<xref ref-type="bibr" rid="B3">3</xref>], regulation of hepatocyte growth factor receptor signaling [<xref ref-type="bibr" rid="B4">4</xref>], and membrane trafficking [<xref ref-type="bibr" rid="B5">5</xref>]. Substantial evidence suggests a role for annexin I in glucocorticoid-induced immunosuppression [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>] and MAPK/ERK pathway [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. Increased expression of intracellular annexin I is seen in bronchial epithelial cells grown in the presence of dexamethasone [<xref ref-type="bibr" rid="B9">9</xref>] and secreted annexin I appears to be proteolytically degraded by the human neutrophil elastase to an inactive form [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>].</p><p>Annexin I is a critical mediator of apoptosis [<xref ref-type="bibr" rid="B12">12</xref>-<xref ref-type="bibr" rid="B15">15</xref>]. While overexpression of annexin I has been observed in pancreatic [<xref ref-type="bibr" rid="B16">16</xref>], breast and gastric cancers [<xref ref-type="bibr" rid="B17">17</xref>], reduced or no expression of annexin I has been reported in prostate and esophageal cancers [<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B21">21</xref>]. Thus differential regulation of annexin I in a tissue specific manner may be associated with the development of cancers in these sites.</p><p>Absence of annexin II expression has been reported in two B-cell lymphoma cell lines, Raji and OMA BL-1 [<xref ref-type="bibr" rid="B22">22</xref>]. While annexin II is closely related to annexin I in amino acid identity, its cellular function is clearly different [<xref ref-type="bibr" rid="B9">9</xref>]. Both annexins I and II are upregulated in pancreatic carcinoma [<xref ref-type="bibr" rid="B16">16</xref>], and recent reports have shown absence of both annexins I and II in prostate carcinoma [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. Thus, it appears that both annexins I and II may be coordinately regulated. In view of these observations, the expression of annexin I in human B-cell lymphomas and cell lines was investigated in this study.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Cell culture, drug treatment and reagents</title><p>The human B-cell lymphoma cell lines used in this study are: progenitor B-cell lines (Nalm-6, REH, HPB-Null, PBE-1), B-lymphoblast cell lines (WI-L2, TK-6, DW-10, DHL-16), Burkitt's lymphoma cell lines (Raji, Ramos, OMA-BL-1, Namalwa). TK-6 is a lymphoblast cell line that is heterozygote for thymidine kinase. TK-6 is a derivative of the WI-L2, a lymphoblast cell line. DW-10 and WI-L2 are EBV transformed mature B-cell lines. PBE-1 and NALM-6 are both precursor B cell acute lymphoblastic leukemia cell lines. NALM-6 is an established cell line and PBE1 is a line established short term from a patient with ALL at the University of Nebraska Medical Center [Please note that a DNA fingerprint analysis [<xref ref-type="bibr" rid="B25">25</xref>] of over 500 lymphoma-leukemia cell lines indicated that PBE-1 and NALM6 may be identical]. DHL-16 is a follicular B-cell lymphoma cell line [<xref ref-type="bibr" rid="B26">26</xref>]. Human adenoids were used as a source of normal B-cells, and contained >80% B-cells as determined by cell sorting and flow cytometric analysis. In other experiments, normal B-cells were isolated from PBL of a healthy volunteer using the human B-cell isolation kit (Miltenyi Biotec Inc., Auburn, CA) as per manufacturer's guidelines. SW1116, HeLa and 293T cells were used as positive controls in the indicated experiments. Cells were grown in a growth medium consisting of Eagle's minimum essential medium (GIBCO-BRL, Grand Island, N.Y.) supplemented with 10% heat-inactivated fetal bovine serum, L-glutamine (2 mM), penicillin (100 U/ml) and streptomycin (100 μg/ml). All cell lines were determined to be mycoplasma-free by a PCR-based mycoplasma detection assay. 5-aza-2'-deoxycytidine (deoxyC, Sigma Chemical Co., St. Louis, MO) was freshly prepared in distilled water. Raji and OMA BL-1 cells, growing in T25 flasks, were incubated in one of the following: 3 μM or 6 μM deoxyC for 3 days or 6 days; 2.5 to 10 μg/mL lipopolysaccharide from E. coli (Sigma) for 24 hours; or 100 μM H<sub>2</sub>O<sub>2 </sub>for 2 to 24 hours. Genomic DNA, total cellular RNA and protein were extracted from cells using previously published procedures [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>Tissue sections</title><p>Formalin-fixed, paraffin-embedded tissues representing normal tonsil (n = 2), diffuse large cell non-cleaved B-cell lymphoma (n = 2), small lymphocytic B-cell lymphoma (n = 2) and follicular mixed B-cell lymphoma (n = 2) were generously made available for these studies by Dr. Dennis Weisenburger, Department of Pathology and Microbiology, University of Nebraska Medical Center, Omaha, NE. Sequential serial sections prepared from these specimens were used in immunohistochemical analysis.</p></sec><sec><title>Antibodies</title><p>Polyclonal rabbit anti-human annexin I antisera (α646) was a gift from Dr. Blake Pepinsky (Biogen, Boston, MA.). Polyclonal anti-human phosphoglycerate kinase (PGK) antiserum (α35) was prepared as described previously [<xref ref-type="bibr" rid="B28">28</xref>]. The mouse monoclonal anti-CD 20 antibody was from Beckman/Coulter Inc. (Westbrook, MA).</p></sec><sec><title>Gel electrophoresis and immunoblotting</title><p>Protein extracts were prepared from cells using protein lysis buffer containing; 50 mM Tris-HCl pH 7.5, 2.0 mM PMSF, 5.0 mM iodoacetamide, 5.0 mM EDTA, 150 mM NaCl, 0.5% NP-40, and 0.5% Mega-9. Protease inhibitors were added just prior to use (leupeptin at 1 μg/ml and pepstatin at 2 μg/ml). Total amounts of protein extracts was quantitated using Bio-Rad protein assay (Pierce, Rockford, IL). A total of 50 μg of protein from each extract was separated on a 12% SDS-PAGE. After electrophoresis, proteins were transferred to PVDF membrane (Millipore, Bedford, MA). The membranes were blocked in 1X TTBS with 7% powdered milk overnight at 4°C and then probed for 1 hour with rabbit polyclonal anti-human annexin I (α646) at 1:500 dilution and rabbit polyclonal anti-human PGK (α35) at 1:1000 dilution. Anti-rabbit HRP (Promega, Madison, WI) at a dilution of 1:5000 was used for 1 hour at room temperature to detect antigen-antibody complexes. Membranes were developed using ECL+ (Amersham Pharmecia Biotech, Arlington Heights, IL). For quantitation of the immunoreactive band, the blot was scanned on a laser densitometer (Molecular Dynamics, Sunnyvale, CA).</p></sec><sec><title>RNA isolation and polymerase chain reaction analysis</title><p>Total RNA was isolated from the cell monolayers or tissue samples using RNeasy mini kit (Qiagen Inc., Valencia, CA) or by TRIzol method (GIBCO/BRL). Annexin I message was characterized by PCR using the enhanced Avian RT-PCR kit (Sigma). Total RNA (5 μg) was added to each RT reaction using random nonamer primers. The cDNA products were amplified using annexin I specific primers to give a 522 bp product. The primers designed for amplification were:</p><p>Annexin I forward: 5'-GCAAGAAGGTAGAGATAAAG-3',</p><p>Annexin I reverse: 5'-ATCTCTCTTCAGTTCCTCTC-3'</p><p>For verification of the integrity of the RNA samples and as a control in all the RT-PCR analyses, we examined the expression of the α-tubulin gene. The primers A-tu1 (5'-AAG AAA TCC AAG CTG GAG TTC-3') and A-tu2 (5'-GTT GGT CTG GAA TTC TGT CAG-3'), specific for the α-tubulin gene, generated a 300 bp PCR product corresponding to the α-tubulin gene.</p></sec><sec><title>Southern blot analysis</title><p>Genomic DNA was isolated from Raji, HPB-null, Nalm-6 and HeLa cell lines, and subjected to restriction endonuclease digestion. Southern blot analysis [<xref ref-type="bibr" rid="B29">29</xref>] of genomic DNA was performed using radiolabeled 522 bp annexin I cDNA product from the PCR reaction as described above.</p></sec><sec><title>Immunohistochemistry</title><p>Standard methods were used to prepare sequential 8 μM sections of paraffin-embedded tissues, which were mounted on polylysine-coated glass slides. Tissues were cleared and rehydrated with two 10-minute rinses in a ready-to-use tissue deparaffinization solution (Biogenex Inc. San Ramon, CA). Sections were then stained by the immunoperoxidase method using the ready-to-use Vectastain Quick kit (Vector Labs, Burlingame, CA). Primary antibodies used included rabbit polyclonal anti-human annexin I antibody and mouse monoclonal anti-human CD 20 antibody.</p></sec></sec><sec><title>Results</title><sec><title>Expression of annexin I in human B-cell lymphoma cell lines</title><p>Annexin I protein levels were measured by immunoblot analysis in various B-cell lymphoma cells along with and other cell lines known to express annexin I. A typical autoradiogram is shown in Figure <xref ref-type="fig" rid="F1">1a</xref>. Extracts made from adenoid cells (containing > 80% B-cells) were used as control normal cells, and HeLa cell extracts were used as positive controls. As expected HeLa cells (lane 1) expressed high levels of annexin I, and adenoids (lane 13) expressed annexin I, but at a much reduced level when compared to HeLa cells. Annexin I protein was undetectable in any of the progenitor B-cell lines (Nalm-6, REH, HPB-Null, PBE-1) and Burkitt's lymphoma cell lines (Raji, Ramos, OMA-BL-1, Namalwa). Among the B-lymphoblast cell lines (WI-L2, TK-6, DW-10), only WI-L2 expressed annexin I, while its derivative TK-6 did not express annexin I. Using a different source of normal B-cells (obtained from PBL of a healthy volunteer), we once again confirmed annexin I expression in normal B-cells (lane 1, Figure <xref ref-type="fig" rid="F1">1b</xref>). In addition, we find significantly reduced expression of annexin I in the Burkitt's lymphoma cell line Raji and the follicular lymphoma cell line DHL-16, when compared to the normal B-cells as evidenced by the levels of PGK in the same blot. The expression of 3-phosphoglycerate kinase (PGK, Figure <xref ref-type="fig" rid="F1">1</xref>) included as an internal control on the blots was seen in all the cell lines except the adenoids. Expression of proliferating cell nuclear antigen and the 70 and 90 kDa subunits of Ku autoantigen were unaffected in these cell lines (data not shown), indicating a specific lack of expression of annexin I.</p></sec><sec><title>Annexin I is transcriptionally regulated in B-cell lymphoma cell lines</title><p>In view of the lack of annexin I protein expression in the B-cell lymphoma cell lines (Figure <xref ref-type="fig" rid="F1">1</xref>), annexin I message expression was examined by RT-PCR analysis (Figure <xref ref-type="fig" rid="F2">2</xref>). RT-PCR analysis showed that annexin I message was present in adenoids (lane 1) and absent in all of the B-cell lymphoma cell lines. As expected, HeLa (lane 11), 293T (lane 12) and SW1116 (lane 13) cells showed annexin I message.</p><p>Annexin I gene was examined in selected cell lines by Southern blot analysis (Figure <xref ref-type="fig" rid="F3">3</xref>). There was no alteration in the annexin II genomic DNA as seen by no change in the hybridization pattern after cleavage by restriction enzymes, Xba I, Bam H1 and Apa I. These enzymes generated specific fragments which were identical in the Raji, HPB-null and Nalm-6 cell lines when compared to HeLa cells. Therefore, the down-regulation of annexin I expression in B-cell lymphoma cells is at the transcriptional level.</p></sec><sec><title>Effect of gene demethylation on annexin I expression</title><p>Methylation of the C<sub>p</sub>G dinucleotide has been shown to directly inhibit transcription or stabilize structural changes in chromatin that prevent transcription. The nucleotide analogue, deoxyC was used to inhibit DNA methylation in Raji cells to measure reexpression of annexin I protein. Exposure of Raji cells to either low (1 μM) or high (10 μM) deoxyC did not result in reexpression of annexin I protein after 3 days (lanes 3 and 4). However, after 6 days of treatment with deoxyC, as a weak protein band (lanes 5 and 6) that appeared at the position of annexin I was observed. There was no difference in annexin I reexpression at 1 μM or 10 μM deoxyC concentrations.</p></sec><sec><title>Raji cell line response to LPS and H<sub>2</sub>O<sub>2</sub></title><p>To determine if inflammatory stimulus provided by exposure of cells to bacterial LPS will result in expression of annexin I, Raji cells were treated with various concentrations of LPS, and annexin I protein expression was measured by immunoblot analysis. The data in Figure <xref ref-type="fig" rid="F5">5</xref> (panel A) show that exposure of Raji cells to LPS does not result in expression of annexin I in these cells. The effect of producing reactive oxygen species (ROS) in Raji cells on annexin I expression was also investigated. The data in Figure <xref ref-type="fig" rid="F5">5</xref> (panel B) show that exposure of Raji cells to H<sub>2</sub>O<sub>2 </sub>did not cause annexin I expression.</p></sec><sec><title>Immunohistochemical analysis of annexin I expression in B-cell lymphoma</title><p>Annexin I expression was studied by immunohistochemical analysis on archival specimens representing various B-cell lymphomas. Tonsil tissue sections were included as representatives of normal tissue. Positive identification of B-cells was performed by immunodetection of CD 20 antigen. As expected, CD 20 positivity was observed in the germinal centers (Figure <xref ref-type="fig" rid="F6">6</xref>, panel A) and a few CD 20 positive cells were observed in the epithelial layer. Annexin I expression was predominantly in the epithelial layer (Figure <xref ref-type="fig" rid="F6">6</xref>, panel B). Weakly positive annexin I expressing cells were found within the germinal center. At higher magnifications (Figure <xref ref-type="fig" rid="F6">6</xref>, panels C and D), distinct cytoplasmic localization of annexin I was observed in a subset of cells within the germinal center (Figure <xref ref-type="fig" rid="F6">6</xref>, panel D). These results are consistent with the immunoblot data (Figure <xref ref-type="fig" rid="F1">1</xref>) indicating a low-level of annexin I expression in the adenoids.</p><p>Annexin I expression was absent in the B-cell lymphomas (Figure <xref ref-type="fig" rid="F7">7</xref>). Distinct B-cell clusters were observed by their CD 20 positivity in diffuse large cell non-cleaved B-cell lymphoma (Figure <xref ref-type="fig" rid="F7">7</xref>, panel A), small lymphocytic B-cell lymphoma (Figure <xref ref-type="fig" rid="F7">7</xref>, panel C) and follicular mixed B-cell lymphoma (Figure <xref ref-type="fig" rid="F7">7</xref>, panel E). However annexin I expression was undetectable in each of these B-cell lymphomas (Figure <xref ref-type="fig" rid="F7">7</xref>, panels B, D and F) even at higher magnifications (data not shown). Strong epithelial cell staining of annexin I was observed in these specimens (Figure <xref ref-type="fig" rid="F7">7</xref>, panel D). These results are consistent with the absence of annexin I expression in B-cell lymphoma cell lines (Figures <xref ref-type="fig" rid="F1">1</xref> and <xref ref-type="fig" rid="F2">2</xref>).</p></sec></sec><sec><title>Discussion</title><p>In this study, the expression of annexin I in B-cell lymphoma was examined. Annexin I is a pleotrophic, calcium and phospholipid binding protein whose proposed functions include anti-inflammatory activity, mediation of apoptosis, regulation of cell differentiation, and membrane trafficking [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B30">30</xref>]. Annexin I is expressed in the secretory bronchial epithelial cells [<xref ref-type="bibr" rid="B9">9</xref>] and its anti-inflammatory N-terminus is lost in the bronchoalveolar lavage fluids from healthy smokers [<xref ref-type="bibr" rid="B11">11</xref>] indicating the importance of this protein in human health and disease. Annexin I is frequently overexpressed in human cancers including pancreatic [<xref ref-type="bibr" rid="B16">16</xref>] and breast [<xref ref-type="bibr" rid="B17">17</xref>] cancers. However, recent reports indicate that annexin I expression is down-regulated in other human tumors, particularly esophageal and prostate tumors [<xref ref-type="bibr" rid="B19">19</xref>-<xref ref-type="bibr" rid="B21">21</xref>]. Annexin I is similar to the closely related protein annexin II, even though the two proteins are proposed to carry out distinct physiological functions. Curiously, both annexins I and II are lost in prostate cancers [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. Altered expression and loss of annexin II in B-cell lymphoma cell lines has been reported previously [<xref ref-type="bibr" rid="B22">22</xref>]. In view of this, the expression of annexin I in B-cell lymphomas and cell lines were examined in the present study.</p><p>The data presented in this manuscript indicate that annexin I is present in normal B-cells (Figure <xref ref-type="fig" rid="F1">1</xref>, panels A and B). Immunohistochemical examination of normal tonsil sections (Figure <xref ref-type="fig" rid="F6">6</xref>) indicates that a subset of cells within the germinal centers distinctly express annexin I. The germinal center normally harbors highly proliferative B-cells [<xref ref-type="bibr" rid="B31">31</xref>], however it also represents a highly dynamic environment that creates intense genomic instability among B-cells [<xref ref-type="bibr" rid="B32">32</xref>]. It is unclear if the subset of B-cells that show annexin I expression represent proliferative cells or cells that are undergoing differentiation. Association of annexin I expression and cellular differentiation has been shown previously [<xref ref-type="bibr" rid="B33">33</xref>].</p><p>In contrast, most cells lines and neoplasms of pregerminal B-cells did not show detectable annexin I expression. Our data are consistent with the loss of annexin I observed in prostate and esophageal cancers [<xref ref-type="bibr" rid="B21">21</xref>]. The etiology of reduced annexin I expression was studied by examining the possible mechanisms. Southern blot analysis indicated that annexin I gene was intact in the Raji and OMA-BL1 cells, indicating that genomic deletion of annexin I is not the cause for loss of annexin I. Gene silencing by hypermethylation of annexin I promoter was examined by culturing cells in the presence of a demethylating agent, and the data indicated reexpression of annexin I protein after prolonged exposure to 10 μM concentration of deoxyC. Thus, methylation of annexin I promoter could be one of the mechanisms for annexin I silencing in these cells. These results are similar to the reexpression of annexin II observed in a prostate cancer cell line after treatment with deoxyC [<xref ref-type="bibr" rid="B23">23</xref>], indicating methylation as a general mechanism for silencing annexins I and II in cancer tissues. Treatment of cells with a pro-inflammatory agent or an apoptosis inducing agent did not result in expression of annexin I.</p></sec><sec><title>Conclusion</title><p>The data presented in this paper show that the anti-inflammatory protein annexin I is expressed in normal B-cells, but not expressed in B-cell lymphomas and cell lines, similar to the absence of annexin I in prostate and esophageal cancers. Thus, annexin I may be an endogenous suppressor of cancer development, and loss of annexin I may contribute to B-cell lymphoma development. The additional mechanism(s) by which annexin I expression is down-regulated and the physiological consequences of annexin I loss in B-cell lymphomas need further investigation.</p></sec><sec><title>Competing Interests</title><p>None Declared</p></sec><sec><title>Authors' Contributions</title><p>JKV conceived of the project and directed the experiments, analyzed the data and prepared the manuscript. ES performed the experiments as a summer research intern. VKG isolated normal B-cells and performed immunoblot analyses.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2407/4/8/prepub"/></p></sec>
Prevalence of dementia and Alzheimer's disease in elders of nursing homes and a senior center of Durango City, Mexico	<sec><title>Background</title><p>Epidemiological reports about dementia and Alzheimer's disease (AD) in elderly people from developing countries are scarce. Therefore, we sought to determine the prevalences of dementia and AD in a population of nursing home residents and senior center attendees of Durango City, Mexico, and to determine whether any socio-demographic characteristics from the subjects associated with dementia or AD exist.</p></sec><sec sec-type="methods"><title>Methods</title><p>One hundred and fifty-five residents of two nursing homes and 125 attendees of a senior center were examined for dementia and Alzheimer's disease. All subjects were tested by the mini-mental state examination, and those who scored twenty-four or less underwent psychiatric and neurological evaluations. Diagnosis of dementia, AD and vascular dementia (VaD) was based on the DSM-IV criteria. Socio-demographic characteristics from each participant were also obtained.</p></sec><sec><title>Results</title><p>Residents of nursing homes found to suffer from dementia were 25 out of 155 (16.1%). Eighteen of them (11.6%) had AD, and seven (4.5%) had VaD. None of the attendees of the senior center suffered from dementia. Dementia (pooled AD and VaD cases) correlated with white ethnicity (OR = 3.2; 95%CI = 1.28–8.31), and a history of unemployment (OR = 6.46; 95%CI = 1.42–25.97), while AD correlated with journeymen occupations (OR = 4.55; 95%CI = 1.00–19.29).</p></sec><sec><title>Conclusion</title><p>Prevalence of dementia in residents of nursing homes found in this study is much lower than reported from more industrialized countries. AD was more frequent than VaD. Ethnicity and occupation showed effects on the prevalence figures. The prevalence of dementia found has implications for the optimum kind of health care that nursing homes should provide to their residents.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Alvarado-Esquivel</surname><given-names>Cosme</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Hernández-Alvarado</surname><given-names>Ana Berthina</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Tapia-Rodríguez</surname><given-names>Rosa Oralia</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Guerrero-Iturbe</surname><given-names>Ángel</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Rodríguez-Corral</surname><given-names>Karina</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Martínez</surname><given-names>Sergio Estrada</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Psychiatry	<sec><title>Background</title><p>Dementia in general and Alzheimer's disease (AD) in particular affect elderly people all around the world [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. Medicare costs for dementia are high and increase with severity of the disease and institutionalization [<xref ref-type="bibr" rid="B5">5</xref>]. The frequentness of dementia and AD vary substantially in different countries. In the United States, approximately 10% of the population older than sixty-five years of age suffers from dementia, with AD accounting for about two-thirds of the cases [<xref ref-type="bibr" rid="B2">2</xref>]. In a study of European population-based cohorts of persons sixty-five years and older, age-standardized prevalences of 6.4% for dementia and 4.4% for AD were found [<xref ref-type="bibr" rid="B6">6</xref>]. In contrast, elderly people of comparable age from an urban community in China, prevalences of 3.49% for dementia and 1.85% for AD have been reported [<xref ref-type="bibr" rid="B7">7</xref>]. Institutionalized elderly people have shown to have a higher frequency of dementia and AD than those found in the general population. For instance, in the United States, 26.4% of nursing home residents, identified in the 1995 Massachusetts Medicaid nursing home database, had a documented diagnosis of dementia [<xref ref-type="bibr" rid="B8">8</xref>]. Remarkably, dementia syndrome prevalence may be as high as 67.4% on admission to nursing homes [<xref ref-type="bibr" rid="B9">9</xref>] or as high as 72% later [<xref ref-type="bibr" rid="B10">10</xref>]. In Japan, in a neuropathological study, 39% of the autopsied nursing home residents had shown signs of dementia during their lives, and AD accounted for 34% of the cases [<xref ref-type="bibr" rid="B11">11</xref>]. Yet in special nursing homes, which may care for more highly impaired elders i.e. those with cognitive impairment and/or disturbed behaviors, by providing more staff time and more specialized staff assigments to residents than do traditional care units, the frequency of dementia is remarkably high. In a Japanese study, 27% and 55% of the residents in nursing homes and special nursing homes, respectively, showed dementia during life [<xref ref-type="bibr" rid="B12">12</xref>]. A number of risk factors for AD have been reported including family history of dementia, marital status, history of major depression episodes [<xref ref-type="bibr" rid="B4">4</xref>], head trauma [<xref ref-type="bibr" rid="B13">13</xref>], chronic inflammatory reactions, oxidative and nitrosylative stresses, and high cholesterol levels [<xref ref-type="bibr" rid="B14">14</xref>]. In addition, the risk of suffering from AD is greater among women [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B15">15</xref>], and a number of genetic risk factors have been implicated that may increase the risk of developing AD [<xref ref-type="bibr" rid="B16">16</xref>]. There is a lack of information concerning the epidemiology of dementia and AD in elderly people of Mexico. Therefore, we conducted a cross-sectional, prospective, descriptive survey with a 2 phase screening design carried out over 2 years in order to determine the prevalences of dementia and AD in a population of nursing home residents and senior center attendees of Durango, a northern Mexican city. In addition, we sought to determine whether any socio-demographic characteristics of the subjects correlated with dementia or AD.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Study population</title><p>We have studied 280 elderly persons including 155 residents of 2 nursing homes and 125 attendees of a senior center. All elders were inhabitants of Durango City and represented the totality of residents and attendees of the two largest nursing homes and the only one senior center of Durango City. In this study, we considered a senior center as a place which elders use to attend courses such as handicraft, dancing, primary school, high school, singing, etc., and to organize local and national travels for pleasure. One nursing home takes care of women only while the other nursing home has a mixed population of women and men. Residents of the two nursing homes have the same level of care. Characteristics of facilities and the level of care, in both nursing homes and the senior center explored, are similar to those found in the majority of nursing homes and senior centers in Mexico. All 280 participants were studied from the year 2000 until 2002.</p></sec><sec><title>Socio-demographic data, mini mental state examination, and neuropsychiatric evaluation</title><p>This work was performed in two phases. Phase I was a screening survey carried out by trained psychologists who administered the validated Spanish version of the mini mental state examination (MMSE) [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>] to all 280 elders. We used the MMSE because it has proven to be an efficient, widely accepted instrument for the screening and assessment of cognitive impairment in the elderly [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B19">19</xref>-<xref ref-type="bibr" rid="B21">21</xref>]. We used a MMSE cut off point of 24 or less, thus subjects who had MMSE scores below 25 were recruited into the phase II study. Subjects with MMSE scores of 25 or more were not studied further. Although false negative subjects might exist in the study population, the likelihood that they were missed for further evaluation is low since a high MMSE cut off point was used.</p><p>All nursing home residents and senior center attendees were allowed to be studied at any time. Although some elders died during the 2 years of the screening period, we were able to study them all before they died. Therefore, none of the elders was excluded from the study. Since all subjects were screened, response bias and length-based sampling bias were avoided.</p><p>Socio-demographic data including age, gender, birth place, ethnic group, marital status, schooling, occupation during their productive age, socioeconomic level, history of alcohol use, history of smoking, and history of drug use from all 280 subjects studied were obtained [see Additional file 1]. This data was obtained during the screening stage, and of this data such as age, history of alcohol, of drug use, etc were verified during stage II. In a few cases, information from informants was used. Socioeconomic level was evaluated by using the Bronfman criteria [<xref ref-type="bibr" rid="B22">22</xref>]. Briefly, six socioeconomic variables were evaluated: number of persons in the house, number of rooms in the house, floor material of the house, availability of potable water, presence and type of plumbing in the house, and educational level of the head of the family. In nursing home residents, data corresponded to that found in the house where the person lived prior to the nursing home.</p><p>Phase II involved the assesment for dementia by a neurologist and a psychiatrist of all subjects who scored 24 or less in the MMSE. The diagnosis of dementia and AD was performed through clinical interview, and was based on the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) criteria [<xref ref-type="bibr" rid="B23">23</xref>]. Laboratory tests (complete blood count, electrolytes, glucose, calcium and TSH), medical record examinations, and information from informants were used to support the diagnosis. The diagnosis of VaD was determined by clinical examination, and was based on the DSM-IV criteria [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title>Statistical Analysis</title><p>Results were analyzed with the aid of the software Epi Info 6. For age comparison among the groups, the <italic>t </italic>student test was used. To assess the association between the characteristics of the subjects and the disease, the crude odds ratio with a 95% exact confidence interval was used. We calculated the exact confidence interval because the cell value (number of cases) was less than 5 in some comparisons. Odds ratio is the ratio between the probability of having disease among those exposed and the probability of having the disease among those not exposed [<xref ref-type="bibr" rid="B24">24</xref>]. Odds ratio was used because it is an important instrument as a measure of association [<xref ref-type="bibr" rid="B25">25</xref>]. Since the number of cases were too low, we did not perform multivariate analysis for age and gender adjustment. In addition, comparison of the frequencies between groups was performed by the χ<sup>2 </sup>test. A level of <italic>P </italic>< 0.05 was considered significant.</p></sec></sec><sec><title>Results</title><sec><title>Socio-demographic data</title><p>Of the 155 residents of the 2 nursing homes studied, 110 were women and 45 were men. The mean age was 72.5 years (range: 61 to 99 years). A low socioeconomic status was found in eighty-six participants (55.5%), a medium status in sixty-three (40.6%), and a high status in six (3.9%). Most people studied (n = 100) were mestizos (people of mixed race i.e. Mexican Indian mixed with European and/or African people), fifty-five were white. Their occupations were: fifty-three housewives, ten retired, forty-one employees, nine professionals, twenty-two journeymen, five businessmen, and fifteen had a history of unemployment. Their marital status included twenty-four married (15.5%), sixty-two widowed (40.0%), sixty-six never married (42.6%), and three divorced (1.9%). Most of these participants (84.5%) were born in Durango State, and the rest in other Mexican states (15.5%). Fifty two participants (33.5%) had a history of alcohol use, and thirty-four had a history of smoking (21.9%). None of these participants had a history of drug use.</p><p>With respect to the 125 attendees of the senior center studied, 103 were women and 22 were men. The mean age was 69.0 years (range: 50 to 88 years). A low socioeconomic status was found in sixty-seven of these participants (53.6%), a medium status in fifty-five (44.0%), and a high status in three (2.4%). Most people studied (n = 83) were mestizos, followed by whites (n = 41), and indigenous (n = 1). Their occupations were: eighty-four housewives, seventeen retired, four employees, eight professionals, seven journeymen, four businessmen, and one had a history of unemployment. Their marital status included forty-one married (32.8%), sixty widowed (48.0%), nineteen never married (15.2%), and five divorced (4.0%). Most of the participants (66.4%) were born in Durango State, and the rest in other Mexican states (32.8%) or abroad (0.8%). Twenty seven participants (21.6%) had a history of alcohol use, and eight had a history of smoking (6.4%). None of these participants had a history of drug use.</p></sec><sec><title>Mini mental state examination and neuropsychiatry</title><p>Of the 155 residents of the 2 nursing homes studied, 102 (65.8%) showed lower MMSE grades than the cut off point. All 102 individuals with MMSE grades below the cut off point were further evaluated by the physician team. Twenty-five individuals were found to suffer from dementia (16.1%). The causes of dementia were as follows: eighteen out of the twenty-five demented subjects (11.6%) had AD, and seven (4.5%) had VaD dementia. There was consensus of diagnosis in every case by the neuropsychiatry team. Table <xref ref-type="table" rid="T1">1</xref> shows a summary of the MMSE and clinical results in nursing home residents.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Summary of the Mini-Mental State Examination (MMSE) and clinical results in residents of nursing homes.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Age group</td><td align="center">Subjects tested with MMSE</td><td align="center">Subjects with grades below the 25 MMSE cut off point</td><td align="center">Patients with dementia (pooled cases)</td><td align="center">Patients with AD<sup>a</sup></td><td align="center">Patients with VaD<sup>b</sup></td></tr></thead><tbody><tr><td align="left">51–64</td><td align="center">4</td><td align="center">2 (50.0%)</td><td align="center">1 (25.0%)</td><td align="center">1 (25.0%)</td><td align="center">0 (0.0%)</td></tr><tr><td align="left">65–74</td><td align="center">26</td><td align="center">14 (53.8%)</td><td align="center">3 (11.5%)</td><td align="center">1 (3.8%)</td><td align="center">2 (7.7%)</td></tr><tr><td align="left">75–84</td><td align="center">72</td><td align="center">49 (68.1%)</td><td align="center">10 (13.9%)</td><td align="center">8 (11.1%)</td><td align="center">2 (2.8%)</td></tr><tr><td align="left">85–94</td><td align="center">46</td><td align="center">32 (69.6%)</td><td align="center">8 (17.4%)</td><td align="center">7 (15.2%)</td><td align="center">1 (2.2%)</td></tr><tr><td align="left">95–99</td><td align="center">7</td><td align="center">5 (71.4%)</td><td align="center">3 (42.9%)</td><td align="center">1 (14.3%)</td><td align="center">2 (28.6%)</td></tr><tr><td align="left">Total (%)</td><td align="center">155</td><td align="center">102 (65.8%)</td><td align="center">25 (16.1%)</td><td align="center">18 (11.6%)</td><td align="center">7 (4.5%)</td></tr></tbody></table><table-wrap-foot><p><sup>a</sup>AD: Alzheimer's disease. <sup>b</sup>VaD: Vascular dementia.</p></table-wrap-foot></table-wrap><p>With respect to the 125 attendees of the senior center studied, 30 (24.0%) showed MMSE grades below the cut off point. All thirty subjects with low MMSE grades were also further evaluated by the physician team. None of the thirty subjects suffered from dementia. There was consensus of diagnosis in every case by the neuropsychiatry team.</p><p>As seen in Table <xref ref-type="table" rid="T2">2</xref>, two socio-demographic characteristics of the elderly people studied were correlated with dementia (pooled AD and VaD cases): white ethnicity (OR = 3.2; 95% CI = 1.28–8.31), and history of unemployment (OR = 6.46; 95% CI = 1.42–25.97). Similarly, as seen in Table <xref ref-type="table" rid="T3">3</xref>, AD correlated with journeyman occupations (OR = 4.55; 95% CI = 1.00–19.29).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Comparison of characteristics of the patients with dementia (pooled cases) and the elders without dementia.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Socio-demographic Characteristics</td><td align="center">Persons with Dementia n = 25 (%)</td><td align="center">Persons without dementia n = 255 (%)</td><td align="center">OR<sup>a</sup></td><td align="center">95% CI<sup>b</sup></td></tr></thead><tbody><tr><td align="left">Gender</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Female</td><td align="center">21 (84.0%)</td><td align="center">192 (75.3%)</td><td align="center">1.72</td><td align="center">0.55–7.15</td></tr><tr><td align="left"> Male</td><td align="center">4 (16.0%)</td><td align="center">63 (24.3%)</td><td></td><td></td></tr><tr><td align="left">Place of birth</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Durango State</td><td align="center">21 (84.0%)</td><td align="center">193 (75.7%)</td><td align="center">1.69</td><td align="center">0.54–7.00</td></tr><tr><td align="left"> Other Mexican states or abroad<sup>c</sup></td><td align="center">4 (16.0%)</td><td align="center">62 (24.3%)</td><td></td><td></td></tr><tr><td align="left">Socioeconomic level</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Low</td><td align="center">14 (56.0%)</td><td align="center">139 (54.5%)</td><td></td><td></td></tr><tr><td align="left"> Medium</td><td align="center">10 (40.0%)</td><td align="center">108 (42.4%)</td><td align="center">0.92</td><td align="center">0.35–2.33</td></tr><tr><td align="left"> High</td><td align="center">1 (4.0%)</td><td align="center">8 (3.1%)</td><td align="center">1.24</td><td align="center">0.03–10.48</td></tr><tr><td align="left">Ethnic group</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Mestizo</td><td align="center">10 (40.0%)</td><td align="center">173 (67.8%)</td><td></td><td></td></tr><tr><td align="left"> White</td><td align="center">15 (60.0%)</td><td align="center">81 (31.8%)</td><td align="center">3.2</td><td align="center">1.28–8.31</td></tr><tr><td align="left"> Indigenous</td><td align="center">0 (0.0%)</td><td align="center">1 (0.4%)</td><td></td><td></td></tr><tr><td align="left">Marital Status</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Married</td><td align="center">4 (16.0%)</td><td align="center">61 (23.9%)</td><td></td><td></td></tr><tr><td align="left"> Widowed</td><td align="center">7 (28.0%)</td><td align="center">115 (45.1%)</td><td align="center">0.93</td><td align="center">0.23–4.50</td></tr><tr><td align="left"> Never married</td><td align="center">14 (56.0%)</td><td align="center">71 (27.8%)</td><td align="center">3.01</td><td align="center">0.88–13.12</td></tr><tr><td align="left"> Divorced</td><td align="center">0 (0.0%)</td><td align="center">8 (31.4%)</td><td></td><td></td></tr><tr><td align="left">Occupation (historical)</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Housewife</td><td align="center">9 (36.0%)</td><td align="center">128 (50.2%)</td><td></td><td></td></tr><tr><td align="left"> Retired</td><td align="center">2 (8.0%)</td><td align="center">25 (9.8%)</td><td align="center">1.14</td><td align="center">0.11–5.98</td></tr><tr><td align="left"> Employee</td><td align="center">2 (8.0%)</td><td align="center">43 (16.9%)</td><td align="center">0.66</td><td align="center">0.07–3.38</td></tr><tr><td align="left"> Professional</td><td align="center">2 (8.0%)</td><td align="center">15 (5.9%)</td><td align="center">1.9</td><td align="center">0.18–10.44</td></tr><tr><td align="left"> Journeyman</td><td align="center">5 (20.0%)</td><td align="center">24 (9.4%)</td><td align="center">2.96</td><td align="center">0.71–10.82</td></tr><tr><td align="left"> Businessman</td><td align="center">0 (0.0%)</td><td align="center">9 (3.5%)</td><td></td><td></td></tr><tr><td align="left"> Unemployed</td><td align="center">5 (20.0%)</td><td align="center">11 (4.3%)</td><td align="center">6.46</td><td align="center">1.42–25.97</td></tr><tr><td align="left">History of alcohol use</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="center">6 (24.0%)</td><td align="center">73 (28.6%)</td><td align="center">0.79</td><td align="center">0.25–2.16</td></tr><tr><td align="left"> No</td><td align="center">19 (76.0%)</td><td align="center">182 (71.4%)</td><td></td><td></td></tr><tr><td align="left">History of smoking</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="center">3 (12.0%)</td><td align="center">39 (15.3%)</td><td align="center">0.76</td><td align="center">0.14–2.70</td></tr><tr><td align="left"> No</td><td align="center">22 (88.0%)</td><td align="center">216 (84.7%)</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p><sup>a</sup>OR: Odds ratio. <sup>b</sup>CI: Confidence interval. <sup>c</sup>Includes 1 subject from the United States with Mexican ancestry.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Comparison of characteristics of the patients with Alzheimer's disease (AD) and the elders without AD.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Socio-demographic characteristics</td><td align="center">Persons with AD n= 18 (%)</td><td align="center">Persons without AD n= 262 (%)</td><td align="center">OR<sup>a</sup></td><td align="center">95% CI<sup>b</sup></td></tr></thead><tbody><tr><td align="left">Gender</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Female</td><td align="center">15 (83.3%)</td><td align="center">198 (75.6%)</td><td align="center">1.62</td><td align="center">0.44–8.97</td></tr><tr><td align="left"> Male</td><td align="center">3 (16.7%)</td><td align="center">64 (24.4%)</td><td></td><td></td></tr><tr><td align="left">Place of birth</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Durango State</td><td align="center">15 (83.3%)</td><td align="center">199 (76.0%)</td><td align="center">1.58</td><td align="center">0.43–8.79</td></tr><tr><td align="left"> Other Mexican states or abroad<sup>c</sup></td><td align="center">3 (16.7%)</td><td align="center">63 (24.0%)</td><td></td><td></td></tr><tr><td align="left">Socioeconomic level</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Low</td><td align="center">11 (61.1%)</td><td align="center">142 (54.2%)</td><td></td><td></td></tr><tr><td align="left"> Medium</td><td align="center">6 (33.3%)</td><td align="center">112 (42.7%)</td><td align="center">0.69</td><td align="center">0.20–2.12</td></tr><tr><td align="left"> High</td><td align="center">1 (5.6%)</td><td align="center">8 (3.1%)</td><td align="center">1.61</td><td align="center">0.03–14.06</td></tr><tr><td align="left">Ethnic group</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Mestizo</td><td align="center">9 (50.0%)</td><td align="center">174 (66.4%)</td><td></td><td></td></tr><tr><td align="left"> White</td><td align="center">9 (50.0%)</td><td align="center">87 (33.2%)</td><td align="center">2</td><td align="center">0.67–5.90</td></tr><tr><td align="left"> Indigenous</td><td align="center">0 (0.0%)</td><td align="center">1 (0.4%)</td><td></td><td></td></tr><tr><td align="left">Marital Status</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Married</td><td align="center">2 (11.1%)</td><td align="center">63 (24.0%)</td><td></td><td></td></tr><tr><td align="left"> Widowed</td><td align="center">5 (27.8%)</td><td align="center">117 (44.7%)</td><td align="center">1.35</td><td align="center">0.21–14.50</td></tr><tr><td align="left"> Never married</td><td align="center">11 (61.1%)</td><td align="center">74 (28.2%)</td><td align="center">4.68</td><td align="center">0.96–44.66</td></tr><tr><td align="left"> Divorced</td><td align="center">0 (0.0%)</td><td align="center">8 (3.1%)</td><td></td><td></td></tr><tr><td align="left">Occupation (historical)</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Housewife</td><td align="center">6 (33.3%)</td><td align="center">131 (50.0%)</td><td></td><td></td></tr><tr><td align="left"> Retired</td><td align="center">2 (11.1%)</td><td align="center">25 (9.5%)</td><td align="center">1.75</td><td align="center">0.16–10.47</td></tr><tr><td align="left"> Employee</td><td align="center">1 (5.6%)</td><td align="center">44 (16.8%)</td><td align="center">0.5</td><td align="center">0.01–4.28</td></tr><tr><td align="left"> Professional</td><td align="center">1 (5.6%)</td><td align="center">16 (6.1%)</td><td align="center">1.36</td><td align="center">0.03–12.40</td></tr><tr><td align="left"> Journeyman</td><td align="center">5 (27.7%)</td><td align="center">24 (9.2%)</td><td align="center">4.55</td><td align="center">1.00–19.29</td></tr><tr><td align="left"> Businessman</td><td align="center">0 (0.0%)</td><td align="center">9 (3.4%)</td><td></td><td></td></tr><tr><td align="left"> Unemployed</td><td align="center">3 (16.7%)</td><td align="center">13 (5.0%)</td><td align="center">5.04</td><td align="center">0.72–26.73</td></tr><tr><td align="left">History of alcohol use</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="center">3 (16.7%)</td><td align="center">76 (29.0%)</td><td align="center">0.49</td><td align="center">0.09–1.81</td></tr><tr><td align="left"> No</td><td align="center">15 (83.3%)</td><td align="center">186 (71.0%)</td><td></td><td></td></tr><tr><td align="left">History of smoking</td><td></td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="center">1 (5.6%)</td><td align="center">41 (15.6%)</td><td align="center">0.32</td><td align="center">0.01–2.14</td></tr><tr><td align="left"> No</td><td align="center">17 (94.4%)</td><td align="center">221 (84.4%)</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p><sup>a</sup>OR: Odds ratio. <sup>b</sup>CI: Confidence interval. <sup>c</sup>Includes 1 subject from the United States with Mexican ancestry.</p></table-wrap-foot></table-wrap></sec></sec><sec><title>Discussion</title><p>Studies on the prevalence of dementia and AD are encouraged. The results of such studies are useful to estimate the psychiatric and neurological needs of the population [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. The overall prevalence of dementia and AD found in this study is comparable with those reported in elders from other countries [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. The frequencies of dementia and AD were significantly higher in residents of nursing homes (16.1% and 11.6%, respectively) than those found in attendees of the senior center (0% and 0%, p < 0.0001 and p < 0.001, respectively). The prevalences of dementia and AD in residents of the nursing homes found in this study are lower than those reported in the United States [<xref ref-type="bibr" rid="B8">8</xref>] and Japan [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. In contrast, our prevalences of dementia and AD are higher that those reported in France [<xref ref-type="bibr" rid="B3">3</xref>]. It is not clear why the residents of the nursing homes showed a higher frequency of dementia than the attendees of the senior center. There are no admission policies in the nursing homes and the senior center that reject demented subjects. Nevertheless, attendees of the senior center had more stringent behavioral rules for permanency than residents of the nursing homes. Therefore, exclusion of some subjects with psychiatric illnesses might explain the lower prevalences found in the senior center. In addition, all attendees of the senior center had social interaction and intellectual stimulation whereas residents of the nursing homes did not. These factors could influence the frequency of dementia as reported by other authors [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B30">30</xref>]. Similar age between the senior center attendees and the nursing home residents was observed (p > 0.5). Since white ethnicity and history of unemployment were associated with dementia (all causes), we compared the frequency of these characteristics among the groups. The frequency of white ethnicity was significantly higher in nursing home residents than that found in attendees of the senior center (51.6% vs 32.8%, p < 0.01). The association between white ethnicity and dementia observed in this study conflicts with those reported by Gurland [<xref ref-type="bibr" rid="B31">31</xref>] and Weintraub [<xref ref-type="bibr" rid="B32">32</xref>] that found higher frequencies of dementia in Latino and African-American groups than in whites. Similarly, the frequency of a history of unemployment was significantly higher in nursing home residents than the one observed in the senior center attendees (6.5% vs 0.8%, p < 0.05). The association between dementia and unemployment found in this study agrees with data reported by Li [<xref ref-type="bibr" rid="B33">33</xref>]. It is not clear why unemployment relates to dementia, however, associated factors such as illnesses might contribute and this observation deserves further study. Remarkably, AD was associated with a history of journeyman occupations (OR = 4.55; 95% CI = 1.00–19.29). To the best of our knowledge, this association has not been reported elsewhere. Nevertheless, this observation must be taken with care, since a low educational level was present with most journeymen. Other characteristics such as never married status [<xref ref-type="bibr" rid="B34">34</xref>,<xref ref-type="bibr" rid="B35">35</xref>], reported to be related to AD, and alcohol consumption [<xref ref-type="bibr" rid="B36">36</xref>] that may reduce the risk for AD were also analyzed. There were more never married subjects in the nursing homes than in the senior center (50.3 vs 15.2%, p < 0.0000001). In addition, we observed a higher frequency of never married subjects in the group of AD patients (61.1%) than in the group of subjects without AD (28.2%), and these results agree with data reported by Kristjansson [<xref ref-type="bibr" rid="B34">34</xref>] and Helmer [<xref ref-type="bibr" rid="B35">35</xref>]. However, the difference found in our study was not statistically significant (OR = 4.68; 95% CI = 0.96–44.66). Never married elders seem to need more support than married elders who had family, this may explain why we found a higher frequency of never married elders in the nursing homes than in the senior center. History of alcohol use was equally distributed among residents and attendees, and among AD patients and non AD subjects in our study. This latter result agrees with a previous report [<xref ref-type="bibr" rid="B36">36</xref>], but conflicts with data reported by Lindsay [<xref ref-type="bibr" rid="B38">38</xref>] and Ruitenberg [<xref ref-type="bibr" rid="B36">36</xref>], since they found that alcohol consumption was associated with a reduced risk of AD. Although the frequency of smoking was low, we observed a higher frequency of smoking in males than in females. Nevertheless, we did not find any positive or negative association between smoking and AD in our elderly population study. Other authors have shown similar results [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B39">39</xref>]. The most common cause of dementia in our study population was AD (72%). This finding is comparable with those findings reported in India [<xref ref-type="bibr" rid="B1">1</xref>] and France [<xref ref-type="bibr" rid="B40">40</xref>], but our frequency is higher than those reported in Japanese [<xref ref-type="bibr" rid="B12">12</xref>] and American-Chinese populations [<xref ref-type="bibr" rid="B41">41</xref>]. In the elderly people studied, AD was more frequently observed in women than in men (although not statistically significant). This finding agrees with previous observations that AD is more common in females [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B15">15</xref>].</p><p>Characteristics of facilities and level of care of the nursing homes and the senior center of Durango are similar to those found in nursing homes and senior centers of other Mexican states. Results of this study may reflect, although to a limited extent, the frequencies of dementia and AD in elders of nursing homes and senior centers of Mexico; however, further studies should be conducted in order to determine the national magnitude of dementia as a public health problem in the elderly people of Mexico. The prevalence of dementia among the nursing home residents found in this study has implications for health care provision in nursing home residents, and resource implications for those responsible for publicly funded care. Similarly, the study identified factors to consider that may be useful when identifying subjects at risk for dementia.</p></sec><sec><title>Conclusions</title><p>We concluded that prevalences of dementia and AD in elderly people from Durango, Mexico are comparable to those reported in developed countries. AD was the most frequent cause of dementia followed by VaD. Dementia (pooled AD and VaD cases) correlated with white ethnicity, and a history of unemployment. Similarly, AD correlated with journeyman occupations.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>CAE conceived and designed the study protocol, participated in the coordination and management of the study, and wrote the manuscript. ABHA applied the questionnaires, and performed the data analysis. ROTR performed the clinical evaluation of the elders. AGI performed the clinical evaluation of the elders. KRC applied the questionnaires, and performed the data analysis. SEM performed the statistical analysis of the data.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-244X/4/3/prepub"/></p></sec>
Medical care usage and self-rated mental health	<sec><title>Background</title><p>Population studies frequently employ a single item dependent variable for overall health. Self-rated mental health has been the focus of attention less often. The purpose of this project was to investigate the relationship between use of medical care and poor mental health in an elderly population.</p></sec><sec sec-type="methods"><title>Methods</title><p>This study involved a cross-sectional telephone survey of persons over 65 years of age in West Texas, a sparsely-populated 108-county region. Independent variables included number of medical visits, race/ethnicity, age, gender and ability to pay for care. Mental health was measured by asking subjects how often they felt downhearted or blue.</p></sec><sec><title>Results</title><p>Multiple logistic regression analysis revealed that more medical visits were made by persons who were downhearted or blue. Females, persons who had difficulty paying for care, Hispanic respondents, and older persons were more likely to report poor mental health.</p></sec><sec><title>Conclusions</title><p>Elderly persons in this region who use more medical care are at greater risk of being in poor mental health. Public health agencies that are planning population-based approaches to improving mental health should consider targeting persons who are high users of medical care as well as those of limited means, women, persons of Hispanic ethnicity, and people who are of greater age.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Rohrer</surname><given-names>James E</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Public Health	<sec><title>Background</title><p>Self-rated health may be more relevant to the goals of community health programs than mortality and morbidity rates. Self-rated health reflects the degree to which people are satisfied with their health and whether they can perform their usual activities, which is more important to most people than whether they are labeled with a particular diagnosis.</p><p>The validity of self-rated overall health has been firmly established and frequently studied [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B8">8</xref>]. Self-rated mental health is important in its own right. However, the epidemiology of self-rated mental health has not been explored as extensively as overall self-rated health. The international public health community has placed increasing emphasis on mental health. Therefore, epidemiological studies such as the one reported here are of increasing relevance and importance.</p><p>Modern societies are stressful, partly due to income inequalities, and the resulting damage to population health is consistent with the theories that drive the field of social epidemiology [<xref ref-type="bibr" rid="B8">8</xref>]. The failure to recognize and address mental health problems in the population is unfortunate for those persons who suffer from such problems. In addition, demands on the medical care system may be greater than necessary as persons seek care from a system that may not be as prepared to recognize and address mental health problems as it should be. The result might be ineffectiveness and inefficiency in a service delivery system that fails to adequately address poor mental health.</p><p>Epidemiological studies of mental health problems pose special measurement problems, because of the need to cost-effectively collect a broad set of measures that are brief yet valid. Measures of physical and mental health used in epidemiological studies have evolved in recent years. Long instruments are regarded as important for studies of patients in clinical settings, but impractical for community surveys. A single item has become the norm for measuring overall health in population studies. Measurement of mental health in population studies also has evolved from complex diagnostic instruments toward shorter scales. For example, a study of older people in Spain used a single item to measure health but 20 items to measure depression [<xref ref-type="bibr" rid="B9">9</xref>]. A recent study of medical inpatients used an eight-item symptom checklist to detect anxiety and depression and a seven-item index to mean hypochondriasis [<xref ref-type="bibr" rid="B10">10</xref>]. In contrast, the American Journal of Public Health published the results of a national survey that used a single item to measure mental health. The single item was dichotomized (positive versus negative mental health) [<xref ref-type="bibr" rid="B11">11</xref>]. The Behavioral Risk Factor Surveillance Survey (BRFSS), which is required of every US state by the Centers of Disease Control, uses a single-item to measure mental health (i.e., the number of recent days when mental health was poor). Single-item measures of mental health are valid because, rather than seeking to assign a clinical diagnosis such as depression, they simply reflect the respondent's perceptions of his or her own mental health. Perceived or self-rated mental health is inherently valid because the respondent is the best judge of his or her own perceptions.</p><p>The purpose of the study reported here was to test the hypothesis that persons who make more medical visits have worse self-rated mental health. This hypothesis is based on the theory that the medical care system is suboptimal in its ability to recognize and treat mental health problems. Failure to address the true causes of the person's symptoms could lead to more use of medical care. Therefore, high-use of medical care might be a risk factor for poor mental health in population surveys.</p></sec><sec sec-type="methods"><title>Methods</title><p>The third wave of the Texas Tech 5000 survey (N = 5,006) was used for this analysis. The Texas Tech 5000 was a random digit dialing telephone survey that involved interviews of persons over age 65 who resided in the 108 counties comprising West Texas. This study was approved by the Institutional Review Board of the Texas Tech University Health Sciences Center. Persons called were screened for cognitive impairment before granting consent to participate in the survey. Seventy-nine percent of eligible persons agreed to the baseline interview, resulting in 5006 enrollees. Telesurveys Research Associates of Houston, Texas collected the data under the auspices of a contract that specified random sampling, multiple call-backs, bi-lingual interviewers, testing for cognitive impairment, and obtaining informed consent. Additional rounds were conducted over the next three years. By the end of the project, 3155 subjects were still participating.</p><sec><title>Dependent Variable</title><p>In a study of self-rated health in Vancouver, Dunn measured mental health by asking respondents: "How much of the time in the past two weeks have you felt downhearted and blue?"[<xref ref-type="bibr" rid="B4">4</xref>]. This item was used in the Rand Mental Health Inventory with the following possible responses: all of the time, most of the time, a good bit of the time, some of the time, a little of the time, none of the time [<xref ref-type="bibr" rid="B5">5</xref>]. The item is reported to have good predictive validity [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>].</p><p>In this study, the following question was used: "During the past four weeks, have you felt downhearted and blue....1) all the time (1.1 percent), 2) most of the time (2.7 percent), 3) a good bit of the time (3.5 percent), 4) some of the time (18.1 percent), 5) a little of the time (26.1 percent), or 6) none of the time" (48.5 percent). The responses were dichotomized into poor mental health or good mental health. The poor mental health category included respondents who had felt blue "all," "most", or "a good bit" of the time. Good mental health included those who felt blue "some", "a little", or "none" of the time. Categorized in this way, only 7.3 percent of the sample was in the poor mental health category. This variable was dichotomized in order preserve consistency with the approach generally used to analyze self-rated overall health. The split was made between 'a good bit of the time' and 'some of the time' because of the logical difference in the perceived severity of mental health problems that is reflected in these different responses.</p></sec><sec><title>Independent Variables</title><p>The number of medical visits in the past year was divided into 5 dummy variables: no visits, 1 visit, 2 visits, 3–4 visits, 5 or more visits with those having no visits as the referent category. The relationship between visits and poor mental health was adjusted for the effects of other independent variables. These are described below.</p><p>Gender was included as an independent variable, with females being the reference category. Race/ethnicity was indicated as either white non-Hispanic, Hispanic, black, or other race. Age was classified as 65–69, 70–74, 75–79, or 80 and over. In order to control for differences in wealth, respondents were asked if they or anyone in their families had avoided using medical care due to its cost in the last year (yes versus no).</p><p>A multivariate logistic regression was run to determine if the number of medical visits had an independent relationship with poor self-rated mental health. EpiInfo 2003 was used to perform the analysis.</p></sec></sec><sec><title>Results</title><p>Descriptive statistics are shown in Table <xref ref-type="table" rid="T1">1</xref>. The majority of respondents were non-Hispanic White. Of that group, only 5.9 percent reported poor mental health. In contrast, 17.3 percent of Hispanic respondents had felt blue.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Descriptive Statistics</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left">Percent Blue</td><td align="left">Percent Not Blue</td><td align="right">Total Cases</td></tr></thead><tbody><tr><td align="left">Ethnicity/Race</td><td></td><td></td><td></td></tr><tr><td align="left"> Hispanic</td><td align="left">17.3</td><td align="left">82.7</td><td align="right">346</td></tr><tr><td align="left"> NH black</td><td align="left">6.8</td><td align="left">93.2</td><td align="right">73</td></tr><tr><td align="left"> NH white</td><td align="left">5.9</td><td align="left">94.1</td><td align="right">2664</td></tr><tr><td align="left"> Other</td><td align="left">11.5</td><td align="left">88.5</td><td align="right">61</td></tr><tr><td align="left">Age</td><td></td><td></td><td></td></tr><tr><td align="left"> 65–69</td><td align="left">6.2</td><td align="left">93.8</td><td align="right">737</td></tr><tr><td align="left"> 70–74</td><td align="left">6.6</td><td align="left">93.4</td><td align="right">967</td></tr><tr><td align="left"> 75–79</td><td align="left">7.5</td><td align="left">92.5</td><td align="right">682</td></tr><tr><td align="left"> 80 and over</td><td align="left">9.1</td><td align="left">90.9</td><td align="right">758</td></tr><tr><td align="left">Gender</td><td></td><td></td><td></td></tr><tr><td align="left"> Female</td><td align="left">8.7</td><td align="left">92.3</td><td align="right">2230</td></tr><tr><td align="left"> Male</td><td align="left">3.8</td><td align="left">96.2</td><td align="right">914</td></tr><tr><td align="left">Avoided Medical Care Due to Cost</td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="left">16.4</td><td align="left">83.6</td><td align="right">324</td></tr><tr><td align="left"> No</td><td align="left">6.3</td><td align="left">93.7</td><td align="right">2820</td></tr><tr><td align="left">Medical Visits</td><td></td><td></td><td></td></tr><tr><td align="left"> None</td><td align="left">7.8</td><td align="left">92.2</td><td align="right">258</td></tr><tr><td align="left"> One</td><td align="left">5.2</td><td align="left">94.8</td><td align="right">464</td></tr><tr><td align="left"> Two to five</td><td align="left">6.4</td><td align="left">93.6</td><td align="right">1479</td></tr><tr><td align="left"> Six to ten</td><td align="left">6.4</td><td align="left">93.6</td><td align="right">512</td></tr><tr><td align="left"> More than ten</td><td align="left">14.3</td><td align="left">85.7</td><td align="right">329</td></tr></tbody></table></table-wrap><p>The percentage with poor mental health increased with age. Only 6.2 percent of persons between the ages of 65 and 69 had poor self-rated mental health. The percentages increased with each age bracket, reaching a high of 9.1 percent for persons 80 and over.</p><p>Gender also was related to poor mental health, with 8.7 percent of women having been blue. In contrast, only 3.8 percent of males had been blue.</p><p>People who reported that they had avoided using medical care due to cost were much more likely to have poor mental health than those who had not avoided care (16.4 percent vs. 6.3 percent).</p><p>As the number of medical visits increased so did the percentage reporting feeling blue or downhearted. Almost eleven percent of respondents reported more than ten visits in the previous year. Over 14 percent of this group also reported poor mental health-nearly double the percentage found in any of the other categories.</p><p>The results of the multiple logistic regression analysis confirmed the apparent relationships seen in the descriptive statistics. As seen in Table <xref ref-type="table" rid="T2">2</xref>, the adjusted odds of poor self-rated mental health are significantly lower for people who made no medical visits in comparison to those who made more than ten (OR = .49, p < .0160). One visit, two to five visits, and six to ten visits also had lower odds of poor mental health in comparison to ten or more visits (ORs were .37, .44, and .49, respectively, with p-values all less than .01).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Multivariate Logistic Regression Analysis of Blueness (Final-2*Log-Likelihood: 1442.36, N = 3042)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Term</td><td align="left">Odds Ratio</td><td align="left">95% C.I.</td><td align="left">P-Value</td></tr></thead><tbody><tr><td align="left">MD Visits</td><td></td><td></td><td></td></tr><tr><td align="left"> None</td><td align="left">0.4936</td><td align="left">0.2779, 0.8768</td><td align="left">0.0160</td></tr><tr><td align="left"> One</td><td align="left">0.3667</td><td align="left">0.2157, 0.6234</td><td align="left">0.0002</td></tr><tr><td align="left"> Two to five</td><td align="left">0.4402</td><td align="left">0.2987,0.6487</td><td align="left">0.0000</td></tr><tr><td align="left"> Six to ten</td><td align="left">0.4894</td><td align="left">0.3017, 0.7939</td><td align="left">0.0038</td></tr><tr><td align="left"> More than ten</td><td align="left">1.0</td><td></td><td></td></tr><tr><td align="left">Age</td><td></td><td></td><td></td></tr><tr><td align="left"> 65–69</td><td align="left">1.0</td><td></td><td></td></tr><tr><td align="left"> 70–74</td><td align="left">1.1912</td><td align="left">0.7828, 1.8126</td><td align="left">0.4141</td></tr><tr><td align="left"> 75–79</td><td align="left">1.4475</td><td align="left">0.9287, 2.2561</td><td align="left">0.1024</td></tr><tr><td align="left"> 80 AND OVER</td><td align="left">1.9954</td><td align="left">1.3089, 3.0420</td><td align="left">0.0013</td></tr><tr><td align="left">Ethnicity/Race</td><td></td><td></td><td></td></tr><tr><td align="left"> Hispanic</td><td align="left">1.0</td><td></td><td></td></tr><tr><td align="left"> NH black</td><td align="left">0.2822</td><td align="left">0.1067, 0.7463</td><td align="left">0.0108</td></tr><tr><td align="left"> NH white</td><td align="left">0.2553</td><td align="left">0.1800, 0.3622</td><td align="left">0.0000</td></tr><tr><td align="left"> Other</td><td align="left">0.4769</td><td align="left">0.1988, 1.1440</td><td align="left">0.0972</td></tr><tr><td align="left">Gender</td><td></td><td></td><td></td></tr><tr><td align="left"> Female</td><td align="left">1.0</td><td></td><td></td></tr><tr><td align="left"> Male</td><td align="left">0.4538</td><td align="left">0.3093,0.6658</td><td align="left">0.0001</td></tr><tr><td align="left">Avoided Medical Care Due to cost</td><td></td><td></td><td></td></tr><tr><td align="left"> Yes</td><td align="left">2.3956</td><td align="left">1.6672,3.4423</td><td align="left">0.0000</td></tr><tr><td align="left"> No</td><td align="left">1.0</td><td></td><td></td></tr></tbody></table></table-wrap><p>All of the control variables were demonstrated to be significantly related to poor self-rated mental health as well. Persons 80 years of age and over had an odds ratio of 2.0 (p = .0013) in comparison to those aged 65–69. In comparison to Hispanic respondents, non-Hispanic black and non-Hispanic White respondents were less likely to report poor mental health (ORs were .28 and .26 with p-values .01 and .00). Men were less likely to report poor mental health as well (OR = .45, p = .0001). Persons who had avoided using medical care due to cost were more likely to report poor mental health (OR = 2.40, p < .0000).</p></sec><sec><title>Discussion</title><p>While no causal inferences can be made from a cross-sectional study such as this one, the sample size is large and the hypothesis is strongly supported. Furthermore, the study stands at the intersection between public health and health services research. This is an important feature of the research, since the medical care system interacts with population health in a variety of ways. Unfortunately, the field of health services research at times has drifted away from public health. Measures of medical care usage would, ideally, be incorporated into more public health studies in the future.</p><p>The medical care system has been chastised for its insensitivity to underlying mental health problems among patients who present with physical symptoms [<xref ref-type="bibr" rid="B12">12</xref>]. The problem is regarded as being of sufficient magnitude that medical costs are thought to be higher because of failure to treat mental health problems. The logical conclusion is that treatment of mental health problems would lead to a reduction or 'offset' in the cost of medical care [<xref ref-type="bibr" rid="B13">13</xref>]. However, policy analysts dispute the existence of an offset effect, calling it a myth [<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Regardless of whether medical costs would actually decline if mental health problems were to be adequately addressed, few could argue that high quality primary care would recognize and treat these issues. Nevertheless, efforts directed at improving the quality of mental health services delivered in primary care settings have had mixed results [<xref ref-type="bibr" rid="B15">15</xref>]. Therefore, it is not surprising that this study reveals poor self-rated mental health to be associated with higher utilization of medical care.</p><p>Several previous reports have demonstrated a relationship between poor mental health and the number of medical visits made [<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. In each of these studies, the number of medical visits served as the dependent variable. In all of these studies, persons with poor mental health were shown to be high users of medical visits. For example, the number of days in poor mental health was related to the number of visits in a farming community [<xref ref-type="bibr" rid="B16">16</xref>] and the number of days depressed was related to visits in another rural population survey [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Studies such as these assume that mental health problems lead to medical visits, which is no doubt a correct assumption. However, public health researchers have not previously used this relationship to identify high-use of medical care as a marker for poor mental health in population studies. The study reported here is the first to do so.</p></sec><sec><title>Conclusions</title><p>The results reported here are contingent on the validity of a single-item measure of self-rated mental health, which may not be accepted by clinicians who focus their energies on diagnoses. However, since single-item measures of physical and overall health are commonly used in epidemiology, a similar approach to mental health should be acceptable in public health.</p><p>The results of this study demonstrated that persons who had more medical visits were at risk for feeling downhearted or blue among the elderly in this southwestern region of the United States. These relationships persisted after adjusting for gender, poverty, age, and ethnicity. The fact that persons with poor self rated mental health made more visits to their doctors suggests that primary care providers might be failing to adequately address poor mental health in this group of patients.</p><p>Finally, we note the need for more research into the causes of poor mental health in the elderly as well as interventional studies to test new population-level approaches that public agencies may employ to reduce the prevalence of self-rated mental health problems among elderly persons.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Author's contributions</title><p>JER planned the study, directed the data analysis and wrote the paper.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2458/4/3/prepub"/></p></sec>
Health and health care utilisation among asylum seekers and refugees in the Netherlands: design of a study	<sec><title>Background</title><p>This article discusses the design of a study on the prevalence of health problems (both physical and mental) and the utilisation of health care services among asylum seekers and refugees in the Netherlands, including factors that may be related to their health and their utilisation of these services.</p></sec><sec><title>Methods/Design</title><p>The study will include random samples of adult asylum seekers and refugees from Afghanistan, Iran and Somali (total planned sample of 600), as these are among the largest groups within the reception centres and municipalities in the Netherlands.</p><p>The questionnaire that will be used will include questions on physical health (chronic and acute diseases and somatization), mental health (Hopkins Symptoms Checklist-25 and Harvard Trauma Questionnaire), utilisation of health care services, pre- and post-migratory traumatic experiences, life-style, acculturation, social support and socio-demographic background. The questionnaire has gone through a translation process (translation and back-translation, several checks and a pilot-study) and cross-cultural adaptation. Respondents will be interviewed by bilingual and bicultural interviewers who will be specifically trained for this purpose.</p><p>This article discusses the selection of the study population, the chosen outcome measures, the translation and cross-cultural adaptation of the measurement instrument, the training of the interviewers and the practical execution of the study. The information provided may be useful for other researchers in this relatively new field of epidemiological research among various groups of asylum seekers and refugees.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Gerritsen</surname><given-names>Annette AM</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Bramsen</surname><given-names>Inge</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Devillé</surname><given-names>Walter</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>van Willigen</surname><given-names>Loes HM</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Hovens</surname><given-names>Johannes E</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>van der Ploeg</surname><given-names>Henk M</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Public Health	<sec><title>Background</title><p>In the Netherlands, health surveys are frequently conducted to assess the health of the population and the utilisation of health care services [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Due to language and cultural problems these surveys often exclude (first generation) immigrants. However, in recent years, much research has focused on the four largest immigrant groups, i.e. people from Surinam, the Netherlands Antilles, Turkey and Morocco [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Although refugees have been coming to the Netherlands since the eighties, their numbers were not large enough and their backgrounds were too diverse for them to be the subject of large-scale epidemiological research. However, it is important that research also focuses on these groups, which differ from the four largest immigrant groups because they migrated involuntarily and may have a history of loss and traumatic experiences. A differentiation can be made between refugees who have a residence permit, and asylum seekers who are still in uncertainty of achieving such a status. These two groups may also differ with regard to living arrangements, because most asylum seekers in the Netherlands live in reception centres. Both factors may cause differences in their health status and their utilisation of health care services, and therefore both groups should be studied. In general, the term 'refugees' will be used for both groups throughout the text, and the term 'asylum seekers' will only be used if a distinction between the groups is important.</p><p>Early research on refugees focused mainly on those refugees who consulted health care services [<xref ref-type="bibr" rid="B4">4</xref>]. This gave an indication of the kind of problems refugees experience, but not of the prevalence of disorders, because not all refugees have or seek medical care for health problems. Other studies focused specifically on the victims of torture, but again such people are not representative of the refugee population in general [<xref ref-type="bibr" rid="B5">5</xref>]. Most population-based studies focusing on adult refugees living in a Western country report on the prevalence of psychiatric diseases, mainly post-traumatic stress disorder (PTSD), depression and anxiety. There is a huge range in reported prevalence rates, due to the fact that the studies are very heterogeneous with respect to the study population (e.g. selection of the study population, country of origin, duration of residence in the country of resettlement, refugee status) and measurement instruments. For example, the prevalence rates for PTSD range from 4% to 70%, and similar percentages are reported for the prevalence of depression (3% to 88%) and anxiety (2% to 80%) (Table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Prevalence of PTSD, depression and anxiety in population-based studies on refugees living in a Western country</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Reference</td><td align="left">Prevalence PTSD</td><td align="left">Prevalence depression</td><td align="left">Prevalence anxiety</td><td align="left">Measurement instrument</td><td align="left">Study population</td></tr></thead><tbody><tr><td align="left">[<xref ref-type="bibr" rid="B13">13</xref>]</td><td align="left">4%</td><td align="left">3%</td><td align="left">5%</td><td align="left">CIDI</td><td align="left">1161 Vietnamese refugees living on average 11 years in Australia</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B40">40</xref>]</td><td align="left">4% (and 9% just after arrival)</td><td align="left">18%</td><td align="left">2%</td><td align="left">criteria from the DSM- Third Edition (PTSD) and Present State Examination (depression and anxiety)</td><td align="left">145 Vietnamese quota refugees interviewed 3 years after resettlement in Norway</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B41">41</xref>,<xref ref-type="bibr" rid="B42">42</xref>]</td><td align="left">11%</td><td align="left">4%</td><td></td><td align="left">criteria from the DSM-Revised Third Edition</td><td align="left">86 Iranian and 70 Turkish asylum seekers (51%) and refugees living in reception centres (62%) in the Netherlands (70% less than 1 year)</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B43">43</xref>]</td><td align="left">12%</td><td></td><td></td><td align="left">Post-traumatic stress section of the Diagnostic Interview Schedule</td><td align="left">223 Cambodian refugees living 3 months to 10 years in New Zealand</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B24">24</xref>]</td><td align="left">15%</td><td></td><td></td><td align="left">HTQ</td><td align="left">240 refugees, predominantly from former Yugoslavia, interviewed on average 10 months and 3 years after resettlement in Norway</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B44">44</xref>]</td><td align="left">18–33%</td><td align="left">21%</td><td></td><td align="left">modified version of the Post-traumatic Symptom Scale (PTSD) and a questionnaire (depression)</td><td align="left">206 refugees from Bosnia-Hercegovina living in an asylum centre in Sweden</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B8">8</xref>]</td><td align="left">22%</td><td></td><td></td><td align="left">HTQ</td><td align="left">157 refugees from Kosovo living on average 2 years in Canada</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B27">27</xref>]</td><td align="left">32%</td><td align="left">63%</td><td align="left">36%</td><td align="left">HTQ (PTSD) en HSCL-25 (depression and anxiety)</td><td align="left">54 Somalian asylum seekers (76%) and refugees living in reception centres (65% less than 6 months) in the Netherlands</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B45">45</xref>]</td><td align="left">35%</td><td align="left">57%</td><td></td><td align="left">CIDI</td><td align="left">51 Afghan refugees living on average 4 years in the Netherlands</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B25">25</xref>]</td><td align="left">35%</td><td align="left">33%</td><td align="left">23%</td><td align="left">CIDI (PTSD) and HSCL-25 (depression and anxiety)</td><td align="left">40 asylum seekers from 21 countries living on average 3 years in Australia</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B32">32</xref>]</td><td align="left">37%</td><td></td><td></td><td align="left">CAPS</td><td align="left">86 Iraqi and Kurdish refugees recently resettled in Sweden</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B12">12</xref>]</td><td align="left">45%</td><td align="left">51%</td><td></td><td align="left">Diagnostic Interview for Children and Adolescents-revised (PTSD) and National Institute of Mental Health Diagnostic Interview Schedule (depression)</td><td align="left">124 Cambodian refugees living on average 8 years in the United States</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B46">46</xref>]</td><td align="left">50%</td><td></td><td></td><td align="left">Structured Clinical Interview for DSM-Fourth Edition</td><td align="left">40 refugees from former Yugoslavia living on average 3.5 years in a refugee camp in Italy</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B47">47</xref>]</td><td align="left">61%</td><td></td><td></td><td align="left">PDS</td><td align="left">129 Kosovar refugees studied immediately upon resettlement in the United States</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B48">48</xref>]</td><td align="left">63%</td><td></td><td></td><td align="left">CAPS</td><td align="left">126 Bosnian refugees with a permanent residency status living for over 3 years (92%) in Australia</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B49">49</xref>]</td><td align="left">65%</td><td align="left">44%</td><td align="left">34%</td><td align="left">PDS (PTSD) and Beck Depression and Anxiety Inventory</td><td align="left">842 refugees from Kosovo living in reception centres in the United Kingdom</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B26">26</xref>]</td><td align="left">86%</td><td align="left">88%</td><td align="left">80%</td><td align="left">PTSD Checklist based on criteria from the DSM-Revised Third Edition and HSCL-25 (depression and anxiety)</td><td align="left">50 Cambodian refugees living on average 5 years in the United States</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B50">50</xref>]</td><td align="left">70%</td><td></td><td></td><td align="left">PDS</td><td align="left">41 Bosnian refugees living in the United States</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B51">51</xref>]</td><td></td><td align="left">6% (just after arrival) – 2% (10 years after arrival)</td><td></td><td align="left">symptom inventory</td><td align="left">608 Southeast Asian refugees living in Canada</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B52">52</xref>]</td><td></td><td align="left">25%</td><td align="left">25%</td><td align="left">HSCL</td><td align="left">180 Somali refugees (96%) and asylum seekers living on average 8 years in the United Kingdom</td></tr><tr><td align="left">[<xref ref-type="bibr" rid="B28">28</xref>]</td><td></td><td align="left">29%</td><td align="left">15%</td><td align="left">HSCL-25</td><td align="left">129 Indochinese refugees living in New Zealand</td></tr></tbody></table><table-wrap-foot><p>CIDI = Composite International Diagnostic Interview; DSM = Diagnostic and Statistical Manual of Mental Disorders; HTQ = Harvard Trauma Questionnaire; CAPS = Clinician-Administered PTSD Scale; PDS = Posttraumatic Diagnostic Scale; HSCL = Hopkins Symptom Check List</p></table-wrap-foot></table-wrap><p>Furthermore, many factors related to the mental health of refugees are reported in the literature cited in Table <xref ref-type="table" rid="T1">1</xref>: pre- and post-migratory traumatic experiences, proficiency in the language of the country of resettlement, social network, socio-demographic background, including gender, work status, duration of residence in the country of resettlement, and marital status.</p><p>Besides studies of the physical conditions detected shortly after the arrival of the refugees (e.g. infectious diseases) [<xref ref-type="bibr" rid="B6">6</xref>] and studies focusing on the physical sequelae of torture, [<xref ref-type="bibr" rid="B7">7</xref>], few studies have investigated physical complaints such as gastrointestinal diseases, musculoskeletal complaints and cardiovascular diseases [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Moreover, the utilisation of health care services has not often been addressed in surveys [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Taking all this into account, it was decided to conduct a large-scale epidemiological study on the prevalence of health problems among both refugees and asylum seekers, including not only mental but also physical health problems, and their utilisation of health care services. In addition it is the intention to study several factors (traumatic experiences, life-style, acculturation, social support, socio-demographic background) that may be related to the health problems and the utilisation of these services. Unlike many of the other studies, the study will include refugees and asylum seekers from three different countries of origin. The aim of the study is to provide some basic epidemiological data on the health and health care utilisation among this population, and thereby improve the health care that is provided (in the Netherlands) for asylum seekers and refugees.</p><p>This article discusses the design of the study: selection of the study population, the chosen outcome measures, cross-cultural adaptation of the measurement instrument, training of the interviewers and the practical execution of the study. Papers reporting the study results can not elaborate much on these issues, although this information may be useful for other researchers in this relatively new field of research, in which problems related to differences in the language and background of the population have to be faced. Describing the design of this study may also help to enhance the comparability of future studies (e.g. regarding the choice of measurement instruments). Furthermore, it permits critical assessment of the methodological quality of the study, irrespective of the outcomes. This is important, because a study is more likely to be examined for methodological limitations if the results differ from what was expected than when the results are in line with the expectations.</p><p>The study design was approved by the Medical Ethics Committee of the VU University Medical Centre in Amsterdam.</p></sec><sec><title>Methods/Design</title><sec><title>Study population</title><p>On 1<sup>st </sup>September 2002, the top 10 nationalities of residents in the Dutch reception centres were: Iraq (8,445 people), Afghanistan (7,105), Angola (6,140), former Yugoslavia (4,806), Iran (4,509), Azerbaijan (4,398), Somalia (3,888), Sierra Leone (3,300), Sudan (3,176) and Syria (2,391) [<xref ref-type="bibr" rid="B15">15</xref>]. Because it was also the intention to study refugees the number of first generation immigrants originating from these countries and living in the Netherlands on 1<sup>st </sup>January 2002, was also recorded. The number of immigrants from five countries was large enough to consider these nationalities for inclusion in the study: former Yugoslavia (55,760 people), Iraq (35,918), Afghanistan (28,448), Iran (22,998) and Somalia (21,071) [<xref ref-type="bibr" rid="B16">16</xref>]. For practical reasons (e.g. translation of the study materials, recruitment of interviewers) former Yugoslavia was not chosen, because this group includes several smaller groups with different ethnic backgrounds and different languages. Furthermore, Iraq was not chosen, because at that time a study on the mental health of asylum seekers from Iraq was being conducted by the Drenthe Mental Health Care Services in the Netherlands. Therefore, people from Afghanistan, Iran and Somalia will be included in this study. To make it possible to compare prevalence rates within sub-groups of people, the plan is to include 100 asylum seekers and 100 refugees per country of origin, resulting in a total study population of 600 people. To achieve a representative sample of all asylum seekers and refugees from these three countries, the sampling procedures described below were applied.</p><sec><title>Sample of asylum seekers</title><p>Based on the mean number of asylum seekers from Afghanistan, Iran and Somalia per reception centre it was decided to include 15 centres in the study. For practical reasons (e.g. travel distances for the interviewers) these reception centres were randomly selected from the 46 centres located in the central region of the Netherlands. One centre was excluded from the sample because it was not considered to be representative. The Community Health Services for Asylum Seekers (MOA) had undergone radical changes due to some recent incidents concerning the health of asylum seekers living in the centre. The central administration of the Dutch Agency for the Reception of Asylum Seekers (COA) was asked to provide the contact details (names, addresses, dates of birth and gender) of all people originating from Afghanistan, Iran and Somalia who were living in the 14 centres. To be eligible for inclusion these people must be 18 years of age or older. Most of them will actually be staying in one of the reception centres concerned, but some might be living in a (neighbouring) municipality. Although most of them will not be in possession of a residence permit, some might be, but are still living in a centre because of lack of alternative accommodation. Because members of the same family may have similar values for some of the outcomes studied (e.g. traumatic experiences, social support), only one person per family was randomly selected for inclusion in the study on the basis of a registration code. This resulted in a sample of at least 157 people per country of origin from all 14 centres together.</p></sec><sec><title>Sample of refugees</title><p>Per country a list was made of municipalities in which at least 200 first generation immigrants originating from that country were living [<xref ref-type="bibr" rid="B16">16</xref>]. At that time, a study on the social status and utilisation of welfare facilities among refugees was being conducted by the Institute for Sociological and Economic Research at the Erasmus University Rotterdam – which also included refugees from Afghanistan, Iran and Somalia – so it was decided not to approach the same municipalities. Furthermore, municipalities were only taken into consideration if they agreed to provide the names and addresses of the people in the sample, so that they could be contacted directly. Some of the municipalities wanted to contact these people themselves and ask for permission to pass on their contact details, but it was thought that this would lower the response rate. Three municipalities (Leiden, Zaanstad and Arnhem) were finally approached and asked to provide a random sample from the population register of 100–150 people per country of origin. Criteria for inclusion in the sample were: 18 years of age or older; born in Afghanistan, Iran or Somalia (or if the country of birth was not recorded, at least one parent born in one of these countries); in possession of a residence permit or the Dutch nationality. Finally, a randomly selection was made of one person per address for inclusion in the study, leaving at least 62 people per country of origin in the sample from each municipality.</p></sec></sec><sec><title>Outcome measures</title><p>A recently conducted systematic review described the cross-cultural validity and reliability of instruments measuring refugee trauma and health status [<xref ref-type="bibr" rid="B17">17</xref>]. The present study included some of the instruments that had either been developed for, or adapted and tested in refugee research (e.g. Harvard Trauma Questionnaire (HTQ), Hopkins Symptom Check List-25 (HSCL-25), 90-item Symptom Check List (SCL-90)). However, the validity and reliability of these instruments has not been tested in the population included in the present study. Furthermore, no cut-off scores for symptomatic status have been established for this population. This should be taken into account when interpreting the results of this study. If no instruments that had been used in refugee research were available questions used in health surveys among the Dutch general or immigrant population were used. However, the cross-cultural validity and reliability of these questions has not been tested.</p><p>A draftversion of the questionnaire was discussed with key-informants (males and females) who were refugees from Afghanistan (3), Iran (5) and Somalia (4). They were contacted with the help of refugee organisations and via the snowballmethod, and had different professional backgrounds (e.g. anthropologist, physician, social worker). They were asked to give their opinion about the items on the questionnaire and the phrasing of the questions. This version of the questionnaire was also reviewed by more than 20 professionals working with refugees (e.g. anthropologists, epidemiologists, physicians, psychologists). The questionnaire was modified with the help of the comments on the draftversion. Questions related to sexual behaviour were omitted (e.g. sexually transmitted diseases, female genital mutilation), because this subject was considered to be taboo, and the answers were expected to be unreliable. The same was thought to apply to questions concerning the use of drugs. A qualitative study design might be more suitable to investigate these issues [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Certain background variables were also omitted (e.g. reason for requesting asylum in the Netherlands, ethnic origin), because these questions might remind the respondents too much about the Immigration and Naturalisation Service (IND) interrogations, and might therefore have a negative influence on the respondents confidence in the interviewer. Furthermore, specific response items were added for some questions, because of the cultural or refugee background of the respondents: e.g. traditional healers and medications, and specific daily activities of asylum seekers living in reception centres (because they are not allowed to work, they spent a lot of time watching television, participating in sports, etc.). The final version of the questionnaire included the outcome measures described below.</p><sec><title>Health</title><sec><title>General</title><p>The current health status of respondents was measured according to the general health question on the 36-item Short-Form [<xref ref-type="bibr" rid="B20">20</xref>]. The response options ranged from '5 = excellent' to '1 = poor'. This item has been used in health surveys among the general and immigrant population of the Netherlands [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>].</p><p>Furthermore, respondents were asked to mention their main health complaints and what they thought to be the cause of these complaints (e.g. physical or mental problem, the situation in the country of origin or in the Netherlands).</p></sec><sec><title>Physical health</title><p>The respondents were asked to indicate for 28 chronic conditions whether or not they had had this condition in the previous 12 months. If so, they were asked if they had visited or been treated by a doctor for this condition during this period. The list of chronic conditions included the items in the national health surveys (e.g. cardiovascular diseases; pulmonary diseases). Furthermore, some items from the screening list used by the MOA were added (e.g. tuberculosis; hepatitis). In addition, seven acute diseases (e.g. flu; bladder infection) from the same surveys were included. Respondents were asked to indicate whether or not they had had these diseases in the previous two months and, if so, whether they had visited a general practitioner for their complaints. Respondents could also mention a chronic or acute disease that was not included in the list.</p><p>Possible somatisation was measured according to the somatic complaints sub-scale of the SCL-90-Revised [<xref ref-type="bibr" rid="B21">21</xref>]. This scale consists of 12 items (e.g. pain in the heart or chest; pain in the lower back) which are complaints that may not be explained by the presence of physical illness, but might be caused by severe stress. Items could be scored on a 5-point scale, ranging from '1 = not at all' to '5 = extremely' bothered by the complaint in the previous week. The sum of all responses divided by the number of items answered produces a mean score. Seven groups are used to classify these mean scores (from 'very high' to 'very low'). This somatisation sub-scale has been used in a population survey of the psychosocial adjustment of Hmong refugees living in the United States [<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Mental health</title><p>The HSCL-25 was used to measure symptoms of anxiety (10 items, e.g. suddenly scared for no reason; feeling fearful) and depression (15 items, e.g. blaming yourself for things; crying easily) [<xref ref-type="bibr" rid="B23">23</xref>]. Respondents were asked to indicate the extent to which they were bothered by each symptom in the previous week, ranging from '1 = not at all' to '4 = extremely'. Individuals with a mean score for anxiety and/or depression and/or the total list of symptoms > 1.75 are considered to be symptomatic. The scale has been used in several refugee studies [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B28">28</xref>]. Because an earlier study on symptoms of depressive illness concluded that the majority of Afghan patients will express death wishes rather than suicidal thoughts, this item was added to the list [<xref ref-type="bibr" rid="B29">29</xref>]. Two items describing typical syndromes of distress related to depression and anxiety in the Iranian culture were also added to the Farsi-version of the questionnaire: nârâhati-e qalb (distress of the heart) and nârâhati-e a'sâb (distress of the nerves) [<xref ref-type="bibr" rid="B30">30</xref>].</p><p>Part IV of the HTQ was used to measure PTSD [<xref ref-type="bibr" rid="B31">31</xref>]. It includes 30 symptoms, the first 16 of which were derived from the Diagnostic and Statistical Manual of Mental Disorders-Revised Third Edition (DSM-III-R) criteria for PTSD (e.g. feeling as though the event is happening again; recurrent nightmares). The other 14 items describe symptoms related to the traumatic life events of (Indochinese) refugees (e.g. difficulty in performing work or daily tasks; blaming yourself for things that have happened). The format of the response options is comparable to that of the HSCL-25. Individuals with a mean score on the 16 PTSD symptoms and/or on the total list of 30 symptoms ≥ 2.5 are considered to be symptomatic for PTSD. This scale has also been used in many studies on refugees [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p><p>An attempt was made to identify culture-specific symptoms of anxiety, depression and trauma by giving respondents the opportunity to mention symptoms that were not included in the list.</p></sec></sec><sec><title>Utilisation of health care services</title><p>The following data were recorded: 1) frequency of contact with a general practitioner, outpatient medical specialist, dentist, physiotherapist, nurse and social-physician of the MOA, in the previous two months; 2) hospital admissions, contacts with mental health services (e.g. psychologist, psychiatrist), contacts with alternative practitioners (e.g. acupuncturist, homoeopathist), in the previous year; 3) use of health care services in the country of origin or other foreign countries in the previous year, and the reason for not using the Dutch health care services; 4) use of (un)prescribed medication in the previous 14 days and type of medication. Some of these measures are also used in the national health surveys. Furthermore, respondents were asked about: who they would turn to with their health problems, their expectations with regard to the Dutch health care providers, their opinion on the Dutch health care system (ranging from '5 = excellent' to '1 = poor') and the reason for this opinion, and suggestions for possible improvements in this system.</p></sec><sec><title>Traumatic experiences</title><p>First the respondents were asked about possible stressful experiences they had had in the Netherlands. The checklist included 18 problems often reported by refugees in research on post-migratory stressors (e.g. delays in the application for a residence permit; loneliness) [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B32">32</xref>]. They were asked to indicate the extent to which any of the items had bothered them in the previous month ('1 = not at all' to '4 = extremely'). Furthermore, they were given the opportunity to mention items that were not included in the list.</p><p>Other traumatic experiences were assessed with part I of the HTQ, which includes 17 events (e.g. lack of food and water; being close to death) [<xref ref-type="bibr" rid="B31">31</xref>]. There were four possible responses for each event (experienced, witnessed, heard about it or no) and respondents were asked to check all that were applicable. Responses are summed and divided by the number of items answered to generate two scores: total number of events (sum of all items for which the response differs from 'no') and total number of events experienced (sum of all items with a positive response to 'experienced'). However, the score for 'total number of events' often approached the maximum score of 17, due to the fact that almost everybody at least answered 'heard about it' to all items. Therefore, the response scale was replaced with a simple yes/no (experienced) option in later versions, because empirical evidence revealed the primary importance of the number of experienced events. Because being a witness can also be an important traumatic event, it was decided to use the earlier response scale, but to eliminate the response 'heard about'. However, this should still make it possible to compare the results of the study with those of other studies which used either the earlier or the later version [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B33">33</xref>]. Furthermore, the list was extended to include 15 other traumatic events, specially relevant to people from Afghanistan, Iran and Somalia, which were selected from Amnesty International Annual Reports (1975–2002) (e.g. rocket attacks, bombardments; confiscation or destruction of houses, crops, water supply). Also included was part III of the HTQ focussing on four traumatic events (yes/no experienced) that may involve head injury (e.g. drowning, suffocation).</p><p>Finally, the respondents were asked to indicate those events that they considered to be the most traumatic events that they had experienced in the Netherlands, in their country of origin or during their flight, and in their whole life.</p></sec><sec><title>Life-style</title><p>Different aspects of life-style were measured as follows: 1) Body mass index (BMI) was calculated by dividing self-reported body weight (kg) by height squared (m<sup>2</sup>). Overweight is defined as a BMI of 25–30 kg/m<sup>2 </sup>and obesity as a BMI of 30 kg/m<sup>2 </sup>or more. 2) Physical activity was assessed by asking respondents how many days a week they spent at least half an hour on physical activities in relation to work, school, household and leisure. According to Dutch standards for adults, this should be at least five days a week. 3) Smoking behaviour was assessed by asking respondents whether they smoked, and if so, how much tobacco they smoked. Heavy smoking is defined as smoking 20 or more cigarettes a day. 4) Alcohol consumption was assessed by asking respondents whether they drank alcohol, and if so, how often they had six or more alcoholic drinks on one occasion. Heavy drinking is defined as having six or more alcoholic drinks at least once a weak. All these life-style measures are used in the national health surveys.</p></sec><sec><title>Acculturation</title><p>To measure the level of self-rated acculturation of the respondents a list of questions was compliled, including items that had already been used in other studies [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B34">34</xref>]. No established acculturation measure was used, because most of the scales are developed for specific groups of immigrants, and take into account the history and conditions of their migration. As a consequence, these scales include items that are not applicable to refugees from Afhanistan, Iran and Somalia. The nine questions in the list focused on: self-reported proficiency (understanding, speaking, reading, writing) in the native language, Dutch and English ('not at all', 'a little', 'sufficient', 'good'); use of language in various situations; food preferences; feeling at home in the Netherlands; which to return home; and ethnic identity (e.g. 'mainly Dutch', 'both Dutch and Afghan').</p></sec><sec><title>Social support</title><p>With regard to social support, two issues were taken into consideration: the frequency of contacts with people who may provide social support, and the perceived amount of support received. To assess the first aspect, respondents were asked about the contact frequency ('often', 'sometimes', 'never') with other people (e.g. family or friends, both in the Netherlands or elsewhere) and with whom they had the most frequent contact. To measure the perceived amount of support received, the established social support measures were not considered to be useful, because they do not take into account the particular life situation of refugees and asylum seekers (e.g. separated from family and friends, living in a reception centre). Therefore, four items were selected from the Social Support Scale (SOS) [<xref ref-type="bibr" rid="B35">35</xref>] (e.g. If I have problems there are people I can turn to) and two items from the UCLA Loneliness scale [<xref ref-type="bibr" rid="B36">36</xref>] (e.g. There are people who really understand me). Respondents were asked to indicate whether or not these statements applied to them ('yes', 'no') in the previous month.</p></sec><sec><title>Socio-demographic background</title><p>The following socio-demographic variables were recorded: gender; age; country of origin; residence permit ('no [application rejected or application under review]', 'temporary', 'permanent' or 'Dutch nationality'); period of residence in the Netherlands and, if applicable, time since obtaining a residence permit (to calculate the duration of the asylum procedure); highest level of education completed; marital status; number and age of children; whereabouts of spouse and any children; main daily activities; and religion.</p></sec></sec><sec><title>Questionnaire translation</title><p>Taking into account the available resources (time and finance), published guidelines were adhered to as possible for the cross-cultural adaptation of the questionnaire [<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B38">38</xref>]. The entire process took approximately five months. The first step was to translate the questionnaire from the original language (Dutch) into the target languages (Dari and Pashto – Afghan languages, Farsi – Iranian language, and Somali), and this was done by experienced translators. Working from the translated version of the questionnaire and totally unaware of the original version, other experienced translators then translated the questionnaire back into the original language. All discrepancies between the original Dutch questionnaire and the back-translated version were recorded by a researcher, who was not familiar with the target languages, and the two translators. These discrepancies were then discussed item-by-item and resolved by consensus. Accordingly, corrections were made, and checked again, which resulted in a second version of the translation. It was found to be important that the researcher made it clear to both translators that discussing discrepancies was in no way meant to judge their work, but to make sure that the translated version was reflecting exactly the same content as the original version. Because also a Dutch version of the questionnaire was needed, the same translation process was also followed for the HTQ, for which no Dutch translation was available.</p><p>In general, it was found that it was possible to translate most items. However, in all languages there were some terms for which no translation was available: for example some chronic diseases (e.g. pulmonary emphysema, slipped disc, angina pectoris), different types of alternative practitioners (e.g. chiropractor, paranormal healer) and the word 'traumatic'. In such cases a description was given and/or the Dutch word was added (between brackets). Furthermore, some phrases were at first translated literally whereas they were meant to be metaphorical (e.g. 'difficulties with breathing' instead of 'shortness of breath'; 'fever or cold' instead of 'hot or cold spells'). Some expressions (e.g. 'feeling on guard', 'my social relationships are superficial') were so difficult to translate that an item with a similar meaning had to be found. In some cases, items had to be substituted by others, because they did not apply to the study population (e.g. 'gardening' was replaced by 'walking' in the question on physical activities, because gardening is not something that asylum seekers do in a reception centre). Finally, there were some difficulties in the translation of large categories of ordinal responses (e.g. the differences between the five response categories of the SCL-90-R). The experienced translators also translated all other study materials (introduction letter, confidentiality statement, etc.).</p></sec><sec><title>Interviewer recruitment and training</title><p>Although the questionnaire can be self-administered, it was decided to make use of bilingual interviewers from the three countries of origin. This was expected to result in a higher response rate, because many respondents might not be familiar with surveys, and some may have difficulties in reading and writing. Furthermore, interviewers could explain the purpose of the study in the respondent's own language and (cultural and refugee) context, thus minimising the risk of misunderstanding or miscommunication.</p><p>An advertisement to recruit interviewers was distributed twice among refugee organisations, employment agencies (for refugees), key-informants, professionals working with refugees, and translators and over a 100 written applications were received. The applicants were screened on interview experience and/or relevant education and/or working experience with refugees, and approximately 65 eligible people were invited for a personal interview. Those with good communication and social skills were selected to participate in a two-day training session. A total of 33 interviewers were trained (9 Afghan – 5 females and 4 males, 15 Iranian – 11 females and 4 males and 9 Somalian – 4 females and 5 males). Most of the interviewers were students studying (para)medical or social subjects. Before the training all participants received a manual in which the content of the interview training was described. The first day of the training included an introduction to the study (background information, purpose, design), general interviewing skills and techniques (e.g. types of questions; adherence to question sequence and wording), and an explanation of the meaning of all items and response options on the questionnaire, which took place while the interview was being practised in the group. There was some role-playing in which specific situations and problems that interviewers could encounter in actual interviews were simulated (e.g. how to keep the respondent 'on track' when he or she wanders off the subject; how to deal with a respondent who is becoming emotional). On the second day the participants practised the interview in pairs in their own language. Furthermore, they practised contact procedures, introducing and ending the interview (including answering frequently asked questions and reacting to typical objections from respondents with regard to co-operation). After completing the training the interviewers could start contacting the first set of five respondents. After each set of interviews, written feedback was given to the interviewers with regard to the quality of the interviews. A researchassistant also has regular contact by phone and (e-)mail with the interviewers to monitor their progress and discuss any problems that may arise.</p><p>During the first training sessions, considerable time was spent discussing the quality of the translation of the questionnaire. Several changes were suggested to the translators, which resulted in a third version. After that training session the questionnaire and the interview procedures were pre-tested. Interviewers were instructed to record any difficulties they encountered and, for example, the time they needed to complete the interview. These difficulties were discussed during an afternoon session. As a result of the pilot study, some minor changes were made in the translation of the questionnaire. In total, 12 respondents were contacted in a reception centre which was not included in the main study. Of these, 9 were interviewed and the other 3 were unwilling to participate for various reasons. In the three municipalities 18 respondents were contacted, only 6 of whom were interviewed. The main reasons for not interviewing a respondent were: the respondent was not living at the given address; the respondent had not come to the Netherlands as a refugee; other reasons for not wanting to participate in the study.</p></sec><sec sec-type="methods"><title>Procedures</title><p>Persons selected for inclusion in the study are sent a letter, both in Dutch and in the language(s) of their country of origin, informing them about the study and announcing that an interviewer will contact them for an interview. If possible respondents will be contacted by phone, otherwise a visit will be paid to the respondent's house or room in the reception centre. If the respondent is not at home a note is left with contact details and the date and time when the interviewer will try to contact the respondent again. The interviewers are instructed to try to contact a respondent three times and all attempts are recorded. If the respondents are contacted they are told about the type of questions they can expect and about the voluntary nature of participation. Asylum seekers are assured that participation in the study will neither help nor hinder their request for asylum, in an attempt to prevent them from participating for the wrong reasons and exaggerating their problems in order to obtain a residence permit. Reasons for not completing an interview with a respondent (e.g. the respondent was never at home; the respondent was not interested in participating) are recorded by the interviewers. Respondents who are willing to participate in the study are given a statement in which the researcher and the interviewer guarantee, among other things, the strict confidentiality of responses and the anonymous reporting of the data. The reason why the respondents are not asked to sign an informed consent statement is that this may remind them of earlier confrontations with authorities. An attempt is made to have male interviewers for male respondents, and female interviewers for female respondents, and the interviews are held either in Dutch or in the respondent's native language. The interviews are estimated to take an average of 90 minutes. At the end of each interview the respondents are given the opportunity to ask questions and informed that a psychologist or physician with experience on PTSD is available if they wish to talk about any distressing feelings evoked by the interview. This service is also available for the interviewers, because they might find it difficult to listen to stories from the respondents, especially if they have experienced similar events themselves. The respondents are also asked to give written permission to the researchers to review their medical records (of the MOA and/or their general practitioner) and those of one of their children under the age of 18 (if applicable). In the second part of the study, which is conducted by the Netherlands Institute for Health Services Research (NIVEL) in Utrecht, the information on health problems and the utilisation of health care services from these records will be compared with the self-reported data obtained from the interviews. Finally, all respondents receive a financial incentive (10 euros).</p></sec></sec><sec><title>Discussion</title><p>When conducting a population-based study among asylum seekers and refugees it is important that a representative sample is included. We plan to achieve this by using random sampling procedures for retrieving contact details of potential respondents and by taking various measures to minimise non-response: contacting respondents both by letter and in person, trying to reach a respondent several times, using bilingual and bicultural interviewers who can explain the purpose of the study and the questionnaire in the respondent's own language and cultural and refugee context, using an oral informed consent procedure, and giving a financial incentive.</p><p>When choosing the methods to obtain data, measurements instruments that have been found to be valid and reliable in the cultures included in the study, or at least in other refugee populations, should first be considered. If the cross-cultural validity and reliability of the instruments is unknown, this should be taken into account when interpreting the study results. Crioss-cultural adaptation of the chosen measurement instruments is a prerequisite if this has not already taken place; translation and back-translation only is not sufficient. In this study, certain checks were included, e.g. the interview training and the pilot-study, after which several amendments were made in the translation. Furthermore, based on the available literature on expressions of symptoms in various cultures, key-informants and professionals, some items were omitted from the questionnaire while certain specific questions and response options were added.</p></sec><sec><title>Competing interests</title><p>None declared</p></sec><sec><title>Authors' contributions</title><p>All authors participated in the design of the study. AG coordinated the study and drafted the manuscript. All authors read the drafts of the manuscript and approved the final version.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2458/4/7/prepub"/></p></sec>
Ureterolithiasis after Cohen re-implantation – case report	<sec><title>Background</title><p>In the past decades, the widespread use of cross-trigonal ureteral reimplants for the treatment of children with vesicoureteral reflux has resulted in a large population of patients with transversely lying ureters. As this population gets older they will consequently be entering an age group at higher risk for stone and urothelial cancer formation. If ureteroscopy becomes necessary, the transverse position of the ureter makes ureteric access often impossible.</p></sec><sec><title>Case Presentation</title><p>We present the case of a young man who not only suffered from urolithiasis due to hyperparathyroidism, but also further jeopardized his treatment by omitting the fact that as a child he underwent Cohen reimplantation of the right ureter.</p></sec><sec><title>Conclusions</title><p>This case illustrates the particular difficulties the endoscopist may face in this group of patients. Patients with difficult ureteric access, abnormal anatomy, or those with known cross-trigonal ureteric reimplantations should be managed in a specialised endourology unit.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Chaudhary</surname><given-names>Sonal</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Lee</surname><given-names>Miranda</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Andrews</surname><given-names>Henry O</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Buchholz</surname><given-names>Noor NP</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Urology	<sec><title>Case Presentation</title><p>A 28-year-old man with known primary hyperparathyroidism presented with right-sided colicky pain in the lumbar region radiating to the groin. A XKUB demonstrated the presence of two urinary calculi, one 8 mm right renal pelvis stone, and one 10 mm stone in the right distal ureter. IVU revealed hydronephrosis of the right kidney and a dilated ureter up to the ureteric stone. The ureter distal of the stone was not opacified on any of the films. After a failed attempt by a colleague to insert a ureteric stent to de-block the right kidney, the patient was referred to our endourology service. The colleague had failed to find the right ureteric orifice.</p><p>At no point did the patient mention that he underwent bladder surgery as a child for a large bladder diverticulum on the right and had a Cohen ureteric reimplantation on that side.</p><p>A percutaneous nephrolithotomy (PCNL) and a combined retrograde-anterograde flexible ureteroscopic approach (URS) were planned. Again, an initial attempt to localize the right ureteric orifice on cystoscopy failed. A PCNL was performed and the kidney stone removed. A guidewire was passed anterogradely down the right ureter alongside the stone into the bladder. At that point, it became clear that immediately distal of the stone the right ureter angulated 90 degrees to the left within the posterior wall of the bladder and exited in the left bladder half. The stone was firmly lodged within that bend. The stone could be visualized endoscopically anterogradely and retrogradely, but due to inflammation, bleeding and lack of vision, a safe laser lithotripsy could not be attempted in that session. Finally, a thin anterograde double-J stent was inserted.</p><p>Only now, questioned again about this most unusual anatomy, the patient remembered his childhood operation. He was scheduled for another retrograde transurethral URS, and this time the operation was successful due to an easy and marked access by the double-J stent. Four weeks later, the patient underwent hyperparathyroidectomy. So far, he is doing well without stone recurrence or complications.</p></sec><sec><title>Conclusions</title><p>To date, urologists are increasingly confronted with a group of patients that had a childhood cross-trigonal reimplantation of one or both ureters. With this technique, first described by Cohen in 1975, the ureter is tunnelled cross-trigonally within the posterior wall of the bladder to exit in the contralateral bladder half. This allows in almost all patients to achieve adequate submucosal length of the ureter. These patients are now coming into an age where they are prone to develop all sorts of urological pathologies necessitating a retrograde ureteric access [<xref ref-type="bibr" rid="B1">1</xref>]. This is nicely illustrated by our case where a young man with such a reimplantation happens to develop hyperparathyroidism and urolithiasis. Not knowing about the reimplantation, several factors played together to obscure the picture for the surgeons. Firstly, the patient did not report his complete medical history, or perhaps, since this had happened in early childhood, he had simply forgotten about it. Secondly, the stone was lodged into the angulation of the ureter, therefore still projecting over the natural course of the right ureter. And finally, the stone blocked the ureter completely, thus not revealing any information about the course of the distal part of the ureter on IVU.</p><p>Cohen reimplantation has been reported as leading to difficulties in ureteric access [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B4">4</xref>]. A variety of approaches to solve the problem has been proposed such as a combination of cystoscopy and suprapubic percutaneous ureteric catheter insertion [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>], percutaneous transvesical ureteroscopy [<xref ref-type="bibr" rid="B1">1</xref>], and transurethrally by using a curved tip vascular catheter combined with an angled tip glide wire [<xref ref-type="bibr" rid="B4">4</xref>]. Where the expertise is readily available, the ureter can also be accessed anterogradely and then later, if needed, retrogradely as in our case. We also found that once the ureter is marked, the insertion of an extra stiff guidewire will straighten the ureter and make access straightforward [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Patients with difficult ureteric access, abnormal anatomy, or those with known cross-trigonal ureteric reimplantations should be managed in a specialised endourology unit.</p></sec><sec><title>Competing Interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>SC and ML researched the literature and wrote a first draft of the manuscript. HA provided clinical background, supervision and reviewed the paper. NB supervised the work and wrote the final version of the manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2490/4/2/prepub"/></p></sec>
A new approach to 'megaprimer' polymerase chain reaction mutagenesis without an intermediate gel purification step	<sec><title>Background</title><p>Site-directed mutagenesis is an efficient method to alter the structure and function of genes. Here we report a rapid and efficient megaprimer-based polymerase chain reaction (PCR) mutagenesis strategy that by-passes any intermediate purification of DNA between two rounds of PCR.</p></sec><sec><title>Results</title><p>The strategy relies on the use of a limiting concentration of one of the flanking primers (reverse or forward) along with the normal concentration of mutagenic primer, plus a prolonged final extension cycle in the first PCR amplification step. This first round of PCR generates a megaprimer that is used subsequently in the second round of PCR, along with the second flanking primer, but without the intermediate purification of the megaprimer. The strategy has been used successfully with four different plasmids to generate various mutants.</p></sec><sec><title>Conclusion</title><p>This strategy provides a very rapid, inexpensive and efficient approach to perform site-directed mutagenesis. The strategy provides an alternative to conventional megaprimer based site-directed mutagenesis, which is based on an intermediate gel purification step. The strategy gives a high frequency of mutagenesis.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Tyagi</surname><given-names>Rajiv</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Lai</surname><given-names>Richard</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Duggleby</surname><given-names>Ronald G</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Biotechnology	<sec><title>Background</title><p>Oligonucleotide-directed mutagenesis is one of the most popular methods to introduce site-specific alterations into a DNA sequence. In past few years, various protocols have been established for site-directed mutagenesis, which facilitate the analysis of gene expression and function. Currently, various PCR-based mutagenesis strategies are available but among them 'megaprimer'-based PCR has been used widely because of its simplicity and low cost [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>Kammann <italic>et al</italic>. [<xref ref-type="bibr" rid="B2">2</xref>] originally developed the megaprimer-based mutagenesis strategy. Later, various changes have been suggested [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B7">7</xref>] to the original strategy. These megaprimer-based mutagenesis strategies are easy to use and they are very cost effective. The strategies require two rounds of PCR amplification using two flanking primers and one internal mutagenic primer. In the first round of PCR, one of the flanking primers and the internal mutagenic primer (having desired base substitutions) are used to generate a megaprimer. The megaprimer is purified by gel electrophoresis and extraction before being used along with the other flanking primer in the second round of PCR to generate the complete DNA sequence with the desired mutation.</p><p>Most of the reported megaprimer mutagenesis strategies rely on the intermediate purification of megaprimer after the first round of PCR amplification to remove the unused primers. This requires the use of cumbersome and time-consuming gel extraction steps. Three megaprimer based site-directed mutagenesis strategies are available at present that circumvent the gel purification step. The first one uses various restriction enzymes to cleave the first and second PCR templates and enzymatic removal of remaining primers from the first PCR reaction [<xref ref-type="bibr" rid="B8">8</xref>]. Since this strategy is based on the use of DNA-modifying enzymes at various steps, it is complicated and expensive to use. In the second strategy the melting temperature (<italic>Tm</italic>) of the forward and reverse flanking primers play a critical role [<xref ref-type="bibr" rid="B9">9</xref>]. A megaprimer is generated using a low <italic>Tm </italic>flanking primer and a low annealing temperature. The second PCR is carried out in the same tube and uses a high <italic>Tm </italic>flanking primer that prevents priming by the low <italic>Tm </italic>primer from the first PCR reaction, thereby obviating the need to gel-purify the megaprimer after the first PCR. However, the strategy is limited by the need to design primers in which one has a <italic>Tm </italic>that is substantially higher than the other; moreover, the average mutant frequency is approximately 80%. The third single tube strategy utilizes the three primers, including mutagenic primer, in three steps of PCR amplification to generate the mutant using different combinations of primers [<xref ref-type="bibr" rid="B10">10</xref>]. The average mutagenesis frequency obtained with this strategy is about 50%. The complex, three-step PCR protocol, along with the lower average mutagenesis frequency, makes this strategy limited in use.</p><p>In this paper we describe an efficient and inexpensive single-tube two-step megaprimer-based PCR mutagenesis strategy (Fig. <xref ref-type="fig" rid="F1">1</xref>) that by-passes the cumbersome and time-consuming gel purification step, does not require additional expensive DNA-modifying enzyme treatments, is not restrictive in primer design, and has a high mutational efficiency. This strategy has been applied successfully to generate mutants of four different proteins; namely, acetohydroxyacid synthase (AHAS) from <italic>Arabidopsis thaliana</italic>, the catalytic and regulatory subunits of <italic>Saccharomyces cerevisiae </italic>AHAS, and <italic>Escherichia coli </italic>ketol-acid reductoisomerase (KARI). To illustrate this strategy, quantitative results obtained with KARI are presented in this paper, together with semi-quantitative analyses for the remaining three constructs. The average mutagenesis frequency obtained with this new strategy approaches 100% under optimal conditions.</p></sec><sec><title>Results</title><p>We tested our new strategy by making a KARI mutant. The first PCR contained 0.05, 0.1 or 1.0 pmole of the first flanking primer (T7 terminator). The second PCR products (1837 bp) (Fig. <xref ref-type="fig" rid="F2">2</xref>, lanes 1–3) were cloned into the pPROEX™HTb vector, and used to transform <italic>E. coli </italic>CU505 cells. To calculate the mutagenesis frequency of the new strategy, several colonies from each experiment were picked and sequenced (Fig. <xref ref-type="fig" rid="F2">2</xref>). A mutant frequency of 100% was observed when 0.05 pmole of the first flanking primer was used in the first PCR, but the efficiency decreased drastically (36%) when a normal amount of primer (1.0 pmole) was used. As we had anticipated, there is a lower percentage of mutant clones as the amount of the first flanking primer increases, presumably because residual first flanking primer is competing with the megaprimer in the second PCR, thereby amplifying the wild-type sequence. A limited amount of the first flanking primer is a key step in our new strategy.</p><p>The advantage of prolonged final extension steps was demonstrated in an experiment where the amount of the first flanking primer was lowered (0.05 pmole) but the prolonged final extension step was omitted. Under these conditions, almost no second PCR product was obtained (Fig. <xref ref-type="fig" rid="F3">3</xref>, lane C).</p><p>Our new strategy was applied to three other templates, using diagnostic restriction digestion to estimate the mutation efficiency without performing sequencing. In the case of the N346H mutation in the regulatory subunit of AHAS from <italic>A. thaliana </italic>[<xref ref-type="bibr" rid="B11">11</xref>], the second PCR product obtained using 0.05, 0.1 and 1.0 pmole of the first flanking primer, when digested with <italic>Kpn</italic>I, showed different relative band intensity patterns when run on 1% agarose gel. The second PCR product obtained using 0.05 pmole of the first flanking primer showed only the bands corresponding to the anticipated mutant with no band of undigested (wild-type) product. In the 0.1 and 1.0 pmole experiments, a mixture of both wild-type and mutant was observed. The intensity of bands corresponding to mutant was very strong as compared to the wild-type for the 0.1 pmole experiment, while for the 1.0 pmole experiment an opposite banding pattern was observed. These results are consistent with our quantitative measurements on the KARI mutant. The strategy was also used to construct mutants of the regulatory subunit (D224A; Pham <italic>et al</italic>., unpublished) and catalytic subunit (M354V; [<xref ref-type="bibr" rid="B12">12</xref>]) of <italic>S. cerevisiae </italic>AHAS. The D224A mutagenic primer removes an <italic>Eco</italic>RV recognition site, while the M354V primer introduces a restriction site for <italic>Apa</italic>LI. Both mutations were tested using the same three different concentrations of first flanking primers. The results obtained from both of these mutants were in agreement with those obtained with the N346H mutant of <italic>A. thaliana </italic>AHAS.</p></sec><sec><title>Discussion</title><p>The main reason for using a gel purification step after the first PCR in conventional megaprimer mutagenesis is to remove excess flanking primer. Without this step, the second PCR would contain both of the flanking primers and substantial amounts of wild-type product would be obtained. One of the novel features of our strategy (Fig. <xref ref-type="fig" rid="F1">1</xref>) is that the amount of the first flanking primer is limited, so that it is substantially depleted during the first PCR. This, in turn, leads to the second novel feature of our strategy, a prolonged extension step at the end of the first PCR. Because the amount of the first flanking primer is limited, a relatively small number of PCR cycles are sufficient to reach an end point. However, the small number of cycles constrains one of the essential features of PCR: full-length products accumulate exponentially while incomplete products accumulate arithmetically. Therefore, after just a few PCR cycles, full-length products will not overwhelm the partial products. We overcame this problem with a final and prolonged extension step, so that all of the partial products are converted to full-length products.</p><p>The second flanking primer is added after the first PCR step; no other additions or manipulations are necessary, in contrast with other single tube PCR strategies that have been described previously [<xref ref-type="bibr" rid="B8">8</xref>-<xref ref-type="bibr" rid="B10">10</xref>]. Most of the megaprimer strategies are limited in the length of the megaprimer to about 300 bp [<xref ref-type="bibr" rid="B13">13</xref>]. The length of megaprimer does not limit our strategy, which can be used for in excess of 1 kbp long megaprimers (Table <xref ref-type="table" rid="T1">1</xref>). This strategy is also useful in terms of PCR set up; it is only necessary to program the software to pause between the two PCR stages, while the second flanking primer is added.</p></sec><sec><title>Conclusions</title><p>The results of sequencing experiments in the case of the KARI mutant, and diagnostic restriction digestion in the case of other three constructs, show the versatility of our strategy. Although there are currently three single PCR tube strategies available, ours is simpler and yields mutagenesis frequencies approaching 100%. We suggest that it could be applied to most constructs making it a simple but universal strategy for site-directed mutagenesis.</p></sec><sec sec-type="methods"><title>Methods</title><p>All PCR experiments were performed on a Perkin-Elmer 480 thermal cycler. Restriction enzymes, DNA ligase and DNA polymerase were obtained from New England Biolabs. Deoxyribonucleotide triphosphates (dNTPs) were purchased from Promega and the primers were obtained from Sigma-Aldrich Pty. Ltd. The mutagenic primers (Table <xref ref-type="table" rid="T1">1</xref>) differed from the wild-type sequence by containing substitutions to change one amino acid residue. Mini preparation of plasmid DNA was performed using the Promega Plasmid Preparation Kit.</p><p>Our new strategy has been applied to four different proteins but for illustration the strategy to construct a mutant of <italic>E. coli </italic>KARI will be described in detail. Isolating clones followed by DNA sequencing was used to assess the success of the strategy. For the other three enzyme constructs, the designed mutagenic primers contained altered restriction enzyme recognition sites to permit facile screening for the mutant (Table <xref ref-type="table" rid="T1">1</xref>). In one case; the target codon change was chosen so as to create a restriction enzyme recognition site, while in a second case an existing site was removed. A third mutagenic primer contained a second substitution that is silent with respect to the protein sequence but which introduced a restriction enzyme recognition site. The resulting mutants were screened using diagnostic restriction digestion.</p><p>The <italic>ilvC </italic>gene (1482 bp) encodes KARI and the gene product is a single polypeptide with a deduced molecular weight of 54 kDa [<xref ref-type="bibr" rid="B14">14</xref>]. We have described previously the cloning of the gene into the pET30a(+) vector [<xref ref-type="bibr" rid="B15">15</xref>]. The pET-C (6879 bp) plasmid was used as the template.</p><p>This new strategy consists of two rounds of PCR amplification. In the first round, the reaction was carried out in total volume of 50 μL and contained 2 U of Vent polymerase, 1X ThermoPol PCR buffer (10 mM KCl, 10 mM (NH<sub>4</sub>)<sub>2</sub>SO<sub>4</sub>, 20 mM Tris-HCl (pH 8.8 at 25°C), 2 mM MgSO<sub>4 </sub>and 0.1% Triton<sup>® </sup>X-100), 50–100 ng of pET-C DNA, 0.2 mM dNTP, 1.0 pmole of internal mutagenic forward primer (Table <xref ref-type="table" rid="T1">1</xref>) and 0.05 pmole of the reverse flanking primer (T7 terminator, GGTTATGCTAGTTATTGCTCAGCGGTGGC). The PCR was carried out under the following conditions: denaturation, 95°C for 1 minute; annealing, 50°C for 30 seconds; and extension, 72°C for 2 minutes. This cycle was repeated 5 times and followed by an additional extension step at 72°C for 35 minutes. In the second round of PCR, the forward flanking primer (T7 promoter, CGCGAAATTAATACGACTCACTATAGGGG) was added into the same tube. After an initial denaturation of 96°C for 1 minute, 25 PCR cycles were used as follows: denaturation, 96°C for 1 minute; annealing, 55°C for 1 minute; and extension, 72°C for 2 minutes. These cycles were followed by an additional extension step at 72°C for 10 minutes. The second PCR product (1837 bp) having desired mutation was purified and digested with <italic>Bam</italic>HI and <italic>Hin</italic>dIII to give a mutated <italic>ilv</italic>C gene fragment (1482 bp). The resulting fragment was purified and cloned into the expression vector pPROEX™HTb (4779 bp) (Gibco BRL) after digesting vector with <italic>Bam</italic>HI and <italic>Hin</italic>dIII. The resulting mutant plasmid (6192 bp) was used to transform <italic>E. coli </italic>strain CU505 (Purdue University, Indiana, USA) competent cells and 10% of the transformed culture was cultured on Luria-Bertani agar plates supplemented with 10 μg/ml ampicillin. The plasmid isolated from selected colonies was sequenced using BigDye™ Terminator chemistry at the Australian Genome Research Facility (Brisbane, Queensland, Australia).</p></sec><sec><title>List of abbreviations</title><p>PCR: polymerase chain reaction; bp: base pairs; kbp: kilo base pairs; <italic>Tm</italic>: melting temperature.</p></sec><sec><title>Authors' contributions</title><p>RT has prepared the manuscript and performed all the experiments. RL gave suggestions for the initial experiments used to establish strategy. RGD gave suggestion for the experiments, helped in preparing and revising the manuscript. All of the authors have read and approved the final manuscript.</p></sec>
A Family Day program enhances knowledge about medical school culture and necessary supports	<sec><title>Background</title><p>A Family Day program was implemented at Indiana University School of Medicine to educate the families and friends of in-coming medical students about the rigors of medical school and the factors that contribute to stress.</p></sec><sec sec-type="methods"><title>Methods</title><p>Surveys that assessed knowledge, beliefs, and attitudes about medical school were administered to participants before and after the program.</p></sec><sec><title>Results</title><p>After the program, participants showed a significant improvement in their understanding of medical school culture and the importance of support systems for medical students. Post-test scores improved by an average of 29% (P < 0.001) in each of the two years this program was administered.</p></sec><sec><title>Conclusions</title><p>The inclusion of family members and other loved ones in pre-matriculation educational programs may serve to mitigate the stress associated with medical school by enhancing the students' social support systems.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Bell</surname><given-names>Mary A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Smith</surname><given-names>Paula S</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Brokaw</surname><given-names>James J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Cushing</surname><given-names>Herbert E</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Medical Education	<sec><title>Background</title><p>The significant stress experienced by both undergraduate and graduate medical trainees is well documented. In a recent literature review, Shapiro et al. [<xref ref-type="bibr" rid="B1">1</xref>] identified over 600 articles that discussed the importance of addressing the stress of medical education. As many as one-third of medical students suffer from psychological morbidity [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Examples of stressors facing students include the need to demonstrate competence, both academic and clinical; managing financial pressures and transitional issues; coping with depression and feelings of isolation; dealing with issues of life and death; making career specialty decisions; and balancing school pressures with family and social obligations.</p><p>Regarding this later aspect, Rappaport et al. [<xref ref-type="bibr" rid="B4">4</xref>] studied the effects of a general surgery residency on the spouses and children of graduate trainees. These investigators found that, at least in their residency program, the stress of medical training created significant problems for the families, further exacerbating an already difficult period. This finding underscores the important relationship between the medical trainee and his or her loved ones. Not all medical students will be married, of course, but most will have close relationships with parents, siblings, or other intimates with whom they confide and draw support. Because family members and close friends provide core social support, they are logical targets for pre-matriculation educational programs aimed at reducing stress in medical students. In a study by Malik [<xref ref-type="bibr" rid="B5">5</xref>], nearly half of the medical students surveyed believed that personal problems were adversely affecting their academic performance, yet those who sought help were more likely to seek out a friend than a faculty member. The literature suggests that strong support systems are critical if students are to successfully emerge from medical training with positive coping skills and a healthy emotional status [<xref ref-type="bibr" rid="B1">1</xref>]. Students who develop healthy adaptive behaviors in dealing with stress are more likely to avoid maladaptive behaviors (e.g., drug abuse) later in their careers [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>With these considerations in mind, we developed a Family Day program to: (1) help orient families and significant others to the culture of medical school, (2) foster insight by providing information about the process of medical education, (3) illustrate the demands on medical students, (4) provide suggestions for supporting medical students, and (5) enhance connections between students and members of their support systems.</p></sec><sec sec-type="methods"><title>Methods</title><p>The planning committee (administrators, faculty, and students) met and conducted a needs assessment about the kinds of information that would be of value to the families and friends of medical students, with a particular emphasis on those aspects of medical training poorly understood by lay people. The committee developed a half-day program that included slides, videotapes, panel discussions, and question and answer sessions. The program was scheduled on a Saturday to optimize participation.</p><p>One month prior to the event, the Dean's Office contacted each of the 140 newly-admitted students and informed them of the upcoming Family Day activities. The students were encouraged to invite their family members and friends. Attendance was entirely discretionary and no attempt was made to determine the reasons for non-participation. Thus, those who chose to attend were self-selected and may have differed in some fundamental way from those who did not attend.</p><p>The first program, conducted in fall 2000, included an overview of the School's new competency-based curriculum, a consideration of the stressors facing medical students, a demonstration of a clinical exam using a standardized patient, a description of the National Residency Matching Program, student and physician panel discussions about medical school and career selection, and a presentation on financial aid. Each of these sessions lasted 15–45 min. Fourth-year students provided campus tours.</p><p>The second program was conducted in fall 2001. Based on feedback from the previous year, the number of topics was reduced to allow more thorough coverage of the curriculum, the competencies, and the residency match. Financial aid and career specialty information was provided in handout format. The survey instrument was modified accordingly.</p><p>Pre- and post-program surveys that assessed knowledge, beliefs, and attitudes about medical school were administered to the participants of the 2000 and 2001 programs (<xref ref-type="supplementary-material" rid="S1">Additional file: 1</xref>). Question topics included the nature and structure of medical school, the psychosocial stressors most often experienced by medical students, residency procurement and length of graduate training, indebtedness and earnings potential, and other relevant information. Pre- and post-program responses were compared with the unpaired t-test. Differences were considered statistically significant if p < 0.05.</p></sec><sec><title>Results</title><p>Sixty-two people attended the 2000 program and 65 attended the 2001 program. Pre- and post-program surveys were completed by 49 of the participants in 2000 and by 45 of the participants in 2001. To what extent these responders may have differed from the non-responders is not known. Participants in both 2000 (n = 49) and 2001 (n = 45) included medical students, their parents, spouses, fiancés/fiancées, relatives, and friends. In 2000, 52% of the participants were female and 48% male; participants in 2001 were 60% female and 40% male. In both years, 22% of the participants were in-coming medical students and 78% were family members and close friends. The occupations of these non-student participants included banking, sales, nursing, law, higher education, trucking, social work, and urban planning, among others.</p><p>After the program, participants demonstrated a significant increase in their knowledge of medical school culture and the demands placed on medical students (Table <xref ref-type="table" rid="T1">1</xref>). Although the baseline pretest score in 2001 was somewhat higher than the pretest score in 2000, presumably because of changes in the survey questions, the percent improvement in post-test scores was essentially the same in both years. Survey scores increased by an average of 29.6% in 2000 (p < 0.001) and by 29.2% in 2001 (p < 0.001).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Improvement in knowledge scores of Family Day participants at Indiana University School of Medicine, 2000–2001. Data are means ± 95% confidence limits.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Family Day</bold></td><td align="center"><bold>Pretest</bold></td><td align="center"><bold>Post-Test</bold></td><td align="center"><bold>n</bold></td><td align="center"><bold>p Value</bold></td></tr></thead><tbody><tr><td align="center">2000</td><td align="center">57.1 ± 5.2</td><td align="center">74.0 ± 4.4</td><td align="center">49</td><td align="center"><.001</td></tr><tr><td align="center">2001</td><td align="center">68.5 ± 4.6</td><td align="center">88.5 ± 3.6</td><td align="center">45</td><td align="center"><.001</td></tr></tbody></table><table-wrap-foot><p>two-tailed, unpaired t-test</p></table-wrap-foot></table-wrap></sec><sec><title>Discussion</title><p>These results suggest that we succeeded in educating participants about some of the stressors facing students and that we equipped participants with strategies for providing support. Written comments from the participants (both families and students) suggest that an unmet need had been fulfilled. Comments included: "Absolutely great! Very informative and certainly worth the time. I would recommend it to every student and his/her family to attend"; "I enjoyed this very much, especially the validation of feeling somewhat isolated from your spouse (the student) and now understanding why"; "Very enlightening!"; "Very informative!"; "Learned a lot"; and "Should continue this program."</p><p>However, it is not known whether the improvements we noted in the survey scores presage improvements in long-term outcomes. A comparison with a matched cohort of students and support system members who did not participate in the program would be helpful. Following such cohorts over time could help determine if there were lasting differences between the groups. Of course, any such analysis would be subject to the bias of participant self-selection. That is, students and support system members who elect to participate in the program may already be sensitized to the need for effective stress management. Nevertheless, outcome measures that might be followed include attrition rates, leaves of absence, academic performance, and overall satisfaction with medical school.</p></sec><sec><title>Conclusions</title><p>Stress during medical school is common. Accordingly, efforts to develop multiple stress-management techniques should continue. We have shown that a pre-matriculation educational program aimed at the families and friends of in-coming medical students can improve their understanding of medical school, which may help to mitigate the stress experienced by medical students by reinforcing the students' social support.</p></sec><sec><title>Competing interests</title><p>None declared</p></sec><sec><title>Authors' contributions</title><p>MB participated in the design and implementation of the educational program, and drafted the manuscript. PS and JB performed the statistical analysis and edited the manuscript. HC participated in the design of the program and edited the manuscript. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6920/4/3/prepub"/></p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional file 1</title></caption><media xlink:href="1472-6920-4-3-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec>
Monitoring of IVF birth outcomes in Finland: a data quality study	<sec><title>Background</title><p>The collection of information on infertility treatments is important for the surveillance of potential health consequences and to monitor service provision.</p></sec><sec><title>Study design</title><p>We compared the coverage and outcomes of IVF children reported in aggregated IVF statistics, the Medical Birth Register (subsequently: MBR) and research data based on reimbursements for IVF treatments in Finland in 1996–1998.</p></sec><sec><title>Results</title><p>The number of newborns were nearly equal in the three data sources (N = 4331–4384), but the linkage between the MBR and the research data revealed that almost 40% of the reported IVF children were not the same individuals. The perinatal outcomes in the three data sources were similar, excluding the much lower incidence of major congenital anomalies in the IVF statistics (157/10 000 newborns) compared to other sources (409–422/10 000 newborns).</p></sec><sec><title>Conclusion</title><p>The differences in perinatal outcomes in the three data sets were in general minor, which suggests that the observed non-recording in the MBR is most likely unbiased.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Gissler</surname><given-names>Mika</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Klemetti</surname><given-names>Reija</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Sevón</surname><given-names>Tiina</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Hemminki</surname><given-names>Elina</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Medical Informatics and Decision Making	<sec><title>Background</title><p>IVF services and children born as a result of IVF treatments have been monitored carefully due to the ethical, legal and economic aspects of assisted reproduction and due to the suggested health risks for treated women and IVF children. In this article we define IVF to include classical IVF, its modifications (mainly ICSI) and frozen embryo transfers (FETs).</p><p>Information on IVF treatments and their outcomes has been gathered by three methods: 1) the collection of aggregated, statistical information on the number of treatments and their results, 2) the collection of individual-level information on all IVF treatments and their results, and 3) the collection of individual-level information on all children born as a result of IVF treatments [<xref ref-type="bibr" rid="B1">1</xref>]. The European reporting systems have been varyingly administered, most often by a health authority, an independent official body, or an association, typically a national fertility society. No routine monitoring system exists in some countries, but ad hoc data collection from IVF clinics has been used to obtain data for international comparisons [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>In Finland, health authorities have used the two methods to monitor IVF. First, information on all successful IVF treatments, i.e. those leading to births and newborns, has been gathered since October 1990 in the Medical Birth Register (subsequently: MBR), which is one of the mandatory health registers. Second, aggregated IVF statistics based on initiated treatment cycles have been gathered since January 1992 [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Both existing data sources have their limitations. The MBR does not include information on women whose treatment did not result in a birth. Aggregated IVF statistics do not enable background-adjusted comparisons between clinics, studies on the use and accumulation of services at individual level, or the follow-up of women or newborns after the perinatal period.</p><p>The main requirements for the use of administrative health registers are that they be complete and that their content corresponds to reality [<xref ref-type="bibr" rid="B4">4</xref>]. Even though many Finnish health registers have been shown to have high completeness and validity [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>], the case may be different for relatively rare events such as infertility treatments. A previous study suggested that information on IVF treatments in the MBR is missing for 15% of IVF children [<xref ref-type="bibr" rid="B6">6</xref>]. Reasons may include problems in distinguishing between different infertility treatments when filling in the data collection form in the delivery hospitals; a lack of data in pregnancy records; or mothers' wishes to conceal their utilisation of infertility treatments. The validity of Finnish IVF statistics has not been evaluated.</p><p>To estimate the completeness and validity of the two routinely collected data bases – MBR and IVF statistics – we compared their information to ad hoc research data for the period 1996 to 1998. The ad hoc data set was created for research purposes by using information on reimbursements for health care services and prescriptions. The main focus of this study was the newborn outcomes of IVF children, since there was extensive information on this subject in all data sources.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>IVF statistics</title><p>The collection of Finnish IVF statistics was started in 1992 on the initiative of the Finnish Society of Obstetrics and Gynecology. Since 1994, STAKES (National Research and Development Centre for Welfare and Health) has had responsibility for data compilation. The data collection is voluntary and is based on aggregated data on initiated IVF treatments [<xref ref-type="bibr" rid="B3">3</xref>]. All clinics participate in the data collection. Since 1994 the data has been collected by using the international data collection form and definitions recommended by the International Working Group for Registers on Assisted Reproduction [<xref ref-type="bibr" rid="B7">7</xref>]. The form includes questions on the number of treatments, on the age of the treated women and their cause of infertility, on the number of transferred embryos, on the results of transfers (number of clinical pregnancies, miscarriages, ectopic pregnancies, induced abortions, stillbirths, live births, gestational age) and on newborn outcomes (birth weight, perinatal mortality and congenital anomalies) [<xref ref-type="bibr" rid="B3">3</xref>].</p></sec><sec><title>Medical Birth Register (MBR)</title><p>The MBR was started in 1987, and is run by STAKES. The register includes the mother's and child's unique, personal identification numbers, and it collects information on maternal background, on care and interventions during pregnancy and delivery and on the newborn's outcome up until the age of seven days. Data are compiled at the time of birth, using mothers' prenatal cards as one of the information sources. The information on IVF (defined as IVF, ICSI, FETs and equivalent treatments in the MBR instructions) has been collected since October 1990. Additionally, since 1996, information on both IVF and other assisted reproduction (defined as insemination, ovulation induction and equivalent treatments) has been collected. Data linkage between the MBR data and the IVF research data (see later) showed that items referring to IVF and 'other assisted reproduction' could not be separated in the MBR, and the majority of children born after 'other assisted reproduction' were in fact born as a results of IVF. Thus the two items were merged in this study.</p><p>The MBR data are collected from all delivery hospitals and in the case of home births is collected by the assisting health care personnel [<xref ref-type="bibr" rid="B2">2</xref>]. Less than 1% of all newborns are missing from the MBR; information on them can be obtained by making data linkages to the Central Population Register and the Cause-of-Death Register kept by Statistics Finland, but no medical information, including information on IVF, is available for these births. After this data linkage the MBR is considered to be complete in terms of numbers of births and newborns. According to two data quality studies, the majority of the MBR content corresponds well or satisfactorily with hospital records [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B8">8</xref>].</p><p>The MBR data are based on the year of birth, while the other data sources are based on the date of conception. Therefore, the date of conception was calculated for all IVF newborns in the MBR by using the date of birth and the information on the best estimate of the gestational age [<xref ref-type="bibr" rid="B9">9</xref>].</p></sec><sec><title>Ad hoc research data</title><p>For research purposes data on IVF treatment cycles and other infertility treatments (including ovulation inductions) performed in the 1996–1998 period were collected from national insurance reimbursement files [<xref ref-type="bibr" rid="B10">10</xref>]. The first data source was information on reimbursed costs for private health care services including physicians' consultations, laboratory and radiological examinations, and infertility treatment procedures. Private services provide some 60% of initiated IVF treatments in Finland [<xref ref-type="bibr" rid="B3">3</xref>]. The reimbursements are based on physicians' itemised bills, and they are filed at the National Social Insurance Institution in an electronic register under women's personal identification numbers.</p><p>The second source was information on drugs prescribed by physicians in outpatient care, covering both the public and private sectors. The National Social Insurance Institution reimburses all drug prescriptions for IVF, and it has kept records on reimbursed prescriptions in an electronic register since 1996. The data includes the woman's personal identification number, municipality of residence, name and class of the drug prescribed, the size and numbers of packages, the recommended dose, the dates prescribed and bought, and the code of the prescribing physician. Information on the indication is not recorded. An algorithm, based on theoretical rules on the use of drugs specific to infertility treatments and their combinations, sequence and dosages, was created to classify women into two groups: women receiving IVF treatments and women receiving other ovulation inductions [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>In order to find births resulting from infertility treatments, information on treatments undertaken was linked to the MBR. Exact dates of treatments were not available, so we used the time difference between the beginning of the last treatment cycle and the birth of the child to estimate which births resulted from IVF or ovulation induction, and which births were the results of spontaneous conception. The time limit of 44 weeks was used as a standard, but another limit of 72 weeks (subsequently: loose definition) was utilised to take into account the uncertainty caused by missing information on exact treatment dates.</p></sec><sec><title>Congenital anomalies</title><p>In the IVF statistics, a short description of each major congenital anomaly – excluding e.g. minor birthmarks of the skin, postural talipes, or clicky hips – leading to a selective induced abortion or to a birth is requested from all IVF clinics. The reported congenital anomalies are reviewed by a clinical expert, and all minor anomalies or outcomes other than congenital anomalies are removed from the statistics. Since the other two data sources did not include information on induced abortions, only congenital anomalies among stillbirths and live births were included in our comparisons. The MBR and the IVF research data were combined with the Finnish Register of Congenital Malformations (subsequently: Malformation Register) by using mothers' personal identification numbers and the dates of birth, and its definitions and classifications were used (see: Definitions). The Malformation Register collects information on all newborns with a congenital anomaly or birth defect through several data sources, including a special data collection form completed by delivery hospitals, and diagnosis data from the MBR, from the Hospital Discharge Register, from the Cause-of-Death Register and from cytogenetic laboratories until the age of one year.</p></sec><sec><title>Definitions</title><p>The definition of stillbirth was the same in the MBR and in the IVF research data (a gestational age of 22 weeks or more or a birth weight of 500 grams or more), but was found to be looser in the IVF statistics (gestational age of 20 weeks or more). The same definition of SGA (small-for-gestational age), based on national standards given by Pihkala et al. [<xref ref-type="bibr" rid="B11">11</xref>], was utilised in the MBR and in the IVF research data. The Malformation Register defines a major congenital anomaly as a significant congenital structural anomaly, chromosomal defect or congenital hypothyroidism. This does not include hereditary diseases and other diseases not associated with congenital anomalies, dysfunction of organs or tissues, developmental disabilities, congenital infections, isolated minor dysmorphic features, normal variations and common less significant congenital anomalies, which are on the exclusion list utilised by the Malformation Register. The exclusion list for minor congenital anomalies is comparable to the list which is utilised by the European Surveillance of Congenital Anomalies EUROCAT [<xref ref-type="bibr" rid="B12">12</xref>].</p></sec><sec><title>Data analysis</title><p>The comparison of information in the IVF statistics with other sources could only be performed on an aggregated level, since this data source contains no personal-level data. The MBR data and the IVF research data were compared at an individual level using women's unique personal identification numbers as the linkage key. The statistical comparisons were done using the chi-square test, the t-test, the test for relative proportions, Fischer's exact test and κ-statistics.</p></sec><sec><title>Data protection issues</title><p>According to national data protection legislation, a limited number of health registers – including the MBR – can be collected using the personal identification number. Since compilation of IVF data is not included in these statues, only aggregated data can be collected without informed consent from each patient. The ad hoc research data was received from the Social Insurance Institution after a special permission for its use in scientific research was given. The data linkage between this data and the MBR was performed after the register keeping organisations and the National Data Protection Authority had authorised it.</p></sec></sec><sec><title>Results</title><p>The number of initiated cycles was 16% higher and the number of transfers 3% higher in the IVF research data than in the IVF statistics. The number of births was almost equal in all data sets. The proportion of multiple births varied between 22% and 23%, and all these differences between the three data sets were statistically insignificant. The variation in the number of newborns followed the same pattern as observed for births. There were, however, more IVF triplets in the MBR (2.1% of all births) than in the IVF statistics (1.0%; p < 0.001) and in the IVF research data (1.1%; p < 0.001) (Table <xref ref-type="table" rid="T1">1</xref>). We also performed all analyses using a looser definition for IVF in the ad hoc research data. This gave 3876 births and 4680 newborns, which respectively were 7.8% and 6.8% higher than the number received by using the strict definition. When comparing the two definitions, the proportion of multiples was somewhat lower (20.3%, p = 0.190) than that determined using the strict definition, but no statistically significant differences were found.</p><p>The data linkage between the MBR and the IVF research data indicated a substantial general concordance of the IVF item in these two data sources: the proportion of children with a correctly reported IVF status (yes/no) was 98.7% and κ-statistics 0.75 (95% confidence interval: 0.74–0.76). If only IVF children were included in the analysis, the quality was poorer: the percentage of children with a correctly reported IVF status declined substantially to 60.2%. The individual linkage revealed, that 24.8% of IVF children (N = 1088) identified in the IVF research data lacked IVF information in the MBR, but on the other hand 24.8% of IVF children (N = 1087) in the MBR were not found in the IVF research data. Thus, even though the total numbers were very similar, 39.8% of children reported to the two data sources as IVF children were different children (Table <xref ref-type="table" rid="T2">2</xref>). These results did not vary by maternal age, parity, maternal smoking and perinatal outcome (data not shown).</p><p>Identical κ-statistics were received when using the loose definition, but the proportion of correctly reported IVF children decreased to 59.0%. The proportion of IVF children in the IVF research data for whom IVF information was lacking in the MBR increased to 28.2%, and the proportion of IVF children in the MBR who were not identified in the IVF research data decreased to 23.3%.</p><p>Age distributions of women with initiated cycles could be determined from the IVF statistics (excluding FETs) and from the IVF research data. The age distributions were similar (data not shown). Also the proportion of women aged less than 25 years (2.3% vs. 2.5%, p = 0.151) and the proportion of women aged 35 years or more (43% vs. 44%, p = 0.226) were similar.</p><p>Information on the backgrounds of parturients was available from the MBR and from the IVF research data. The mean maternal age was some three months lower in the MBR (p = 0.005), and the proportions of mothers aged 35 years or more (37% in the MBR and 40% in IVF research data, p = 0.053) and of mothers aged 40 years or more (8 and 10%, respectively, p = 0.002) were lower in the MBR than in the IVF research data. There were fewer single IVF mothers in the MBR (3%) than in the IVF research data (4%) (p = 0.023). The distributions of socioeconomic position differed between the MBR and the IVF research (p < 0.001), but this difference disappeared after those with an undefined socioeconomic position were excluded from the analysis (p = 0.357). There were small differences in parturients' residence: the MBR reported more IVF in south-east Finland, but less in southern Finland (including the capital area) and in central Finland than did the IVF research data (data not shown). The differences in the number of previous pregnancies and births and in maternal smoking were minor (Table <xref ref-type="table" rid="T3">3</xref>).</p><p>All three data sources showed more perinatal health problems for IVF children compared to all children born in the study period. With the exception of congenital anomalies, the infant outcomes were similar in all three IVF data sets (Table <xref ref-type="table" rid="T4">4</xref>). The incidence of premature births varied from 17% to 18%, the incidence of low birth weight from 19% to 21%, the incidence of SGA from 6.9% to 7.0%, and the perinatal mortality rate from 12 per 1000 to 14 per 1000. None of the differences were statistically significant. The same was true for differences among singletons, but among multiples the MBR reported more premature births than did the IVF statistics (49% vs. 43%, p = 0.019) and the MBR indicated more low birth-weight children than did the IVF research data (46% vs. 43%, p = 0.013) or the IVF statistics (46% vs. 42%, p = 0.002) as a consequence of the excess number of triplets.</p><p>The IVF statistics indicated that there were 157 major congenital anomalies per 10 000 newborns. The incidences were much higher in the MBR (422/10 000 newborns, p < 0.001) and in the IVF research data (409/10 000 newborns, p < 0.001), but the difference between the figure from the MBR and the IVF research data was not statistically significant (p = 0.373) (Table <xref ref-type="table" rid="T4">4</xref>). Similar underreporting in the IVF statistics was also observed, when studying some single major congenital anomalies, which can be observed at birth, such as trisomy 21 (9.2/10 000 newborns compared to 15.4/10 000 newborns in the IVF research data, p = 0.206), cleft palates (6.9/10 000 vs. 37.3/10 000, p = 0.001), and neural tube defects (2.3/10 000 vs. 11.0/10 000, p = 0.058).</p></sec><sec><title>Discussion</title><p>Poor perinatal outcomes of IVF children may be caused by higher multiplicity rate, adverse results of IVF technology, or infertility. Previous Finnish research on IVF children has shown that the main cause of increased perinatal health problems is multiplicity, but even IVF singletons have a higher risk for adverse perinatal outcomes than did singletons in general [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B16">16</xref>]. These studies have had uncertainties, such as unclear coverage [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B14">14</xref>] and small sample size [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. This study comparing three different nation-wide data sources suggests that poorer perinatal outcomes are unlikely to be due to methodological problems.</p><p>Our data showed that the existing two IVF data sets – aggregated data from clinics and individual-level data on newborns – gave short-term outcomes largely identical to those found in the ad hoc IVF research data, with the exception of congenital anomalies. The data linkage between the MBR and the IVF research data revealed, however, that up to two out of five IVF children were not the same individual children. This discrepancy can be explained by the problems in getting information on IVF entered into the MBR; by the problems in forming the IVF research data; and by the differences in the methods used to compile the data sets and in their inclusion criteria.</p><p>For one in four IVF children the information on IVF was missing in the MBR. We can suggest three reasons for this. First, there may be difficulties in distinguishing between IVF and other assisted reproduction methods in the maternity hospitals. A closer analysis of the MBR data showed that there were several large hospitals which did not report any IVF but only other assisted reproduction. This suggests that the hospital computer programs were not updated when the last revision was made to the MBR data collection form in 1996. Therefore, we had to include all children born as a result of assisted reproduction in the IVF group, even though some (at maximum 11%) were not IVF children. This approach accounts for 45% of those "extra" IVF children in the MBR who were not identified as IVF children in the research data. Second, information on the use of infertility treatments may not reach the maternity hospitals. Creating a seamless exchange of information between IVF clinics, antenatal care clinics and maternal hospitals would solve the problem. On the other hand, some women may also be deliberately excluded if the patient has decided to conceal their use of infertility treatments. Third, the formation of the IVF research data was based on administrative register data, which may include incorrect entries, for example, in the drug information or in the treatment codes. This would mean that the IVF research data includes women who did not use IVF, and the missing information on IVF in the MBR is actually correct.</p><p>On the other hand, one in four children reported to be IVF children in the MBR were not found in the IVF research data. Three reasons may explain this. First, there is the problem of wrongly classifying a child as an IVF child in the MBR. Second, the algorithm in the research data was based on current knowledge on drugs used in IVF treatments, but the same drugs may be used for other purposes. We were conservative in defining IVF cases to avoid false negative cases, and may therefore have excluded women who actually had received IVF. Third, it is possible that the IVF research data missed some women who had been treated in the public sector and who had used drugs which were bought and reimbursed earlier to reach the annual ceiling for free medication [<xref ref-type="bibr" rid="B17">17</xref>].</p><p>Besides these main explanatory factors, a small part of the discordance may be explained by technical factors. Despite the large number of possible sources of bias, their effect on our results and conclusions was estimated to be negligible. First, the study period was defined from the start of treatment in the IVF statistics and in the IVF research data, but was retrospectively determined from the date of conception in the MBR. Second, IVF statistics lacked information on some of the less frequent treatments, such as oocyte donation and assisted hatching. Third, the inclusion criteria differed for foreigners; they were included in the IVF statistics, but not in the MBR or in the IVF research data. Private Finnish IVF clinics provide treatments which are unavailable in some neighbouring countries, such as the use of donated oocytes, and the treatment of single women and lesbian couples. On the other hand, Finnish women who received IVF services in other countries – for example in Estonia due to more inexpensive treatments – cannot be found in IVF statistics of the IVF research data, but they were included in the MBR if their births occurred in Finland. Fourth, women who were entitled to reimbursements but who did not apply for them were not in the IVF research data, but were in the two other sources. This group is assumed to be small, because drugs and treatments are expensive, and the reimbursement is usually already given in the IVF clinic and in the pharmacy. Fifth, the IVF statistics included all pregnancies after 20 weeks of gestation, while the MBR and the IVF research data included births after 22 weeks of gestation [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>More triplets were reported to the MBR than found in the IVF statistics or in the ad hoc IVF register data. We have no clear explanation for this phenomenon. It may be caused by the fact that the MBR data is collected in the delivery hospitals, and naturally conceived triplets may incorrectly be assumed to have resulted from infertility treatments. Another explanation is that Finnish women who were treated in the neighboring countries more commonly using three occytes or more per transfer gave triplet births in Finland. The most likely explanation, however, is our decision to define all children reported to have resulted from IVF or other assisted reproduction techniques as IVF children due to quality problems in the MBR: some triplets may have resulted from ovulation induction.</p><p>With the exception of major congenital anomalies the infant outcomes were similar by data source. It seems that information on congenital anomalies does not reach the IVF clinics, or the received information may be too inaccurate to confirm whether the congenital anomaly is a major one or not. This is also confirmed by recent studies on the connection between IVF and congenital anomalies: Lower proportions of children with major congenital anomalies have been reported in studies based on data collected routinely from IVF clinics, from 2.0% to 3.2% of children [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>], than in special studies using data linkages to birth and malformation registers [<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B24">24</xref>], from 4.8% to 9.0%.</p><p>A different follow-up period (one year in the Malformation Register and undefined in the IVF statistics) can explain part of the discrepancy. The monitoring of congenital anomalies related to IVF is important, since previous studies [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B24">24</xref>] have reported a higher incidence of certain anomalies among IVF children. In this study the incidences of major congenital anomalies in the MBR and in the IVF research data were also higher than that observed for the general population (288/10 000 newborns). The cases of congenital anomalies found in this study will be investigated further with an adequate control group. For this kind of study, it is essential that the Malformation Register or other data sources which can be linked to this register collects complete background information on infertility and its treatment.</p><p>Even though the IVF children in the different data sources were not the same children, most of the data in the two routine IVF data sources were comparable to the IVF research data. This suggests that drop-outs were not selected by outcome. One likely explanation is that children born after inseminations and ovulation inductions with increased risk for adverse perinatal outcome may be reported as IVF children in the MBR. Despite this drawback, our results give confidence that IVF services (IVF statistics) and most short-term infant outcomes of IVF newborns (IVF statistics and the MBR) can be reliably monitored in Finland without ad hoc data collection.</p><p>The existing routine Finnish data sources do not, however, answer all relevant study questions, such as explaining the variation in success rates by clinics in controlling for confounding factors. For these purposes, a nation-wide IVF register would be useful. Furthermore, such a register incorporating personal identification number would enable data linkages to other health outcome sources, for example getting more accurate information on congenital anomalies or to study the long-term health outcomes of treated women and children born as a result of IVF treatments.</p><p>The Finnish health information system is based on individual-level register data with the unique personal identification number that is given to all Finnish citizens and permanent residents. In general, the possibility to identify each individual with certainty improves the quality of any data collection and the utilisation of a single register, but also enables technically easy data linkages between various registers. There have been proposals to change the national data protection legislation so that a launch of an IVF register could be made possible [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B25">25</xref>], but the idea has not been explored thoroughly. Close co-operation between IVF clinics and the register-keeping organisation is required to ensure high quality register data, to minimise the extra work load in the IVF clinics and to protect the privacy and confidentiality of treated women.</p><p>IVF is highly specialised care given by a limited number of clinics, and therefore information can be gathered relatively easily. The case is different for other infertility treatments, since they are given more widely. Information on certain surgical procedures has been collected in the Finnish Hospital Discharge Register since 1986, but its use in questions related to infertility treatments is challenged by the limitations in the national classification on operative interventions. Data on inseminations and ovulation inductions has been collected for administrative purposes on an aggregated level [<xref ref-type="bibr" rid="B26">26</xref>], and regionally as ad hoc clinical information [<xref ref-type="bibr" rid="B27">27</xref>]. Since other infertility treatment methods including ovulation induction may have similar risks as IVF, it is important to monitor also these treatments and their outcomes.</p></sec><sec><title>Conclusions</title><p>The existing two IVF data sets, which are routinely collected by the Finnish health authorities, can be used in monitoring of IVF services and their short-term outcomes in general. Information on congenital anomalies as well as on long-term outcome of treated women and IVF children, however, has to be collected separately.</p></sec><sec><title>Abbreviations</title><p>ICSI intracytoplasmic sperm injection</p><p>IVF in vitro fertilisation</p><p>FET frozen embryo transfers</p><p>STAKES National Research and Development Centre for Welfare and Health</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>MG planned the study, made the data analysis regarding IVF statistics, and drafted the article. RK and EH helped with the interpretation of results and writing process. TS made the data linkages and data analysis regarding the Medical Birth Register and the ad hoc data. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6947/4/3/prepub"/></p></sec>
Phospholipse c inhibitor, u73122, stimulates release of hsp-70 stress protein from A431 human carcinoma cells	<sec><title>Background</title><p>Accumulating evidences suggest that Hsp 70, the inducible component of Hsp70 family, might release from a living cell. Here we show that a pharmacological inhibitor of phospholipase C activity U73122 caused a 2–4 fold reduction of an intracellular level of Hsp70 in A431 human carcinoma cells.</p></sec><sec><title>Results</title><p>A depletion of Hsp70 under U73122 was a result of the protein release since it was detected in cell culture medium, as was established by immunoprecipitation and precipitation with ATP-agarose. The reduction of Hsp70 level was specifically attributed to the inhibition of PLC, since the non-active inhibitor, U73343, had no effect on Hsp70 level. The PLC-dependent decrease of Hsp70 intracellular level was accompanied by the enhanced sensitivity of A431 cells to the apoptogenic effect of hydrogen peroxide. Here for the first time we demonstrated one of the possibilities for a cell to export Hsp70 in PLC-dependent manner.</p></sec><sec><title>Conclusion</title><p>From our data we suggest that phospholipase C inhibition is one of the possible mechanisms of Hsp70 release from cells.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Evdonin</surname><given-names>Anton L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Guzhova</surname><given-names>Irina V</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Margulis</surname><given-names>Boris A</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Medvedeva</surname><given-names>Natalia D</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Cancer Cell International	<sec><title>Background</title><p>The proteins belonging to Hsp70 family possess two major properties. First, they are known to elicit chaperone activity, i.e. to recognize misfolded and newly synthesized polypeptides and to participate in their intracellular transport and degradation. Another function is mainly attributed to the inducible member of Hsp70 family and constitutes protective capacity. This was convincingly demonstrated in vitro and in vivo, and the list of cytotoxic factors from which Hsp70 protects cells includes stimuli of apoptosis. The molecular mechanisms underlying the anti-apoptotic activity of Hsp70 are thoroughly investigated now. Among the proteins that may be affected by the over-expressing Hsp70 are APAF-1/caspase-9 [<xref ref-type="bibr" rid="B1">1</xref>], NFkappaB [<xref ref-type="bibr" rid="B2">2</xref>] and stress kinases [<xref ref-type="bibr" rid="B3">3</xref>], all known to participate in apoptotic signaling. Generally, Hsp70 is thought to interfere with almost all known signaling pathways, and <italic>vice versa</italic>, some of the latter can influence the rate of the chaperone expression through the stimulation of protein kinase C activity by the phorbol ester [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>] or through the action of factors up-regulating cellular Ca2+ [<xref ref-type="bibr" rid="B6">6</xref>]. The chaperonic activity of Hsp70 can also be affected by proteins participating in cell signaling; it was shown that Bag-1 in Hsp70-overexpressing hamster fibroblasts was able to switch from Raf-1/ERK signaling cascade to that based on Hsp70 chaperonic mechanism [<xref ref-type="bibr" rid="B7">7</xref>]. Many cell signaling processes are related to the activity of phospholipase C (PLC). One of the latter, phospholipase Cγ1, belongs to the family of PLCs activated in cells exposed to a variety of extracellular ligands. PLC hydrolyzes the minor membrane phospholipid, phosphatidylinositol 4,5-bisphosphate. This reaction results in the generation of the two intracellular second messengers diacylglycerol and inositol 1,4,5-trisphosphate. The latter promotes the activation of protein kinase C and the release of Ca<sup>2+ </sup>from intracellular stores [<xref ref-type="bibr" rid="B8">8</xref>]. Several groups have demonstrated that PLC may also be activated in response to various stressful insults: exposure to heat, oxidants, UVC and mechanical stress [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Stress induced PLC activation was shown to be sufficient for preventing apoptosis induced by heat-shock and oxidants and significantly enhanced cell survival [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. However, the mechanism of PLC activation in a cell physiology under stress conditions remains unclear. Since the expression and activity of Hsp70 can be controlled by the increase of intracellular Ca<sup>2+ </sup>and by protein kinase C, whereas PLC itself is able to regulate these factors, it seemed that both proteins are associated in cell signaling kaleidoscope. Moreover, both proteins are involved in a cell response to similar stressful factors. Based on these data we have suggested that Hsp70 and PLC might be functionally linked in their cellular activities, particularly in the process of a cell reaction to stress. To test this hypothesis we used the keratinocyte-derived A431 epidermoid carcinoma cells. In human skin and epidermal cell lines, such as A431 cells Hsp70 is normally expressed at high level [<xref ref-type="bibr" rid="B13">13</xref>], and these cells may be a convenient model for investigation of consequences of variations in the content of intracellular Hsp70. The goal of this study was to compare the level of Hsp70 in A431 cells treated with various factors known to activate or suppress PLC activity. Inhibition of PLC activity resulted in a substantial decrease of the content of the intracellular Hsp70. Simultaneously, Hsp70 was detected in cell culture medium as was established with the use of immuno-precipitation or precipitation with ATP-agarose. Two forms of Hsp70 were found in culture medium: one corresponding to Hsp70 and an additional one with the molecular mass of 98 kDa, the latter was proven to be the ubiquitinated form of Hsp70. It was also shown that the release of Hsp70 from A431 cells led to the increased sensitivity to the oxidative stress.</p></sec><sec><title>Results</title><p>A431 keratinocyte-derived cells constitutively expressing inducible Hsp70 and its cognate form, Hsc70. To check whether these cells are capable of eliciting a typical stress response, we subjected them to heat shock at 42°C for 40 min. After recovery for 6 h at 37°C, we analyzed Hsp70 with the aid of immunofluorescence and immunoblotting, using 2H9 antibody, which had earlier been shown to recognize only the inducible member of the Hsp70 family. Although control cells contained rather high amounts of Hsp70, this value increased 3-5-fold after heat stress (Fig. <xref ref-type="fig" rid="F1">1A</xref>). Immunofluorescence analysis showed that Hsp70 was located in the cytoplasm in control cells, whereas after heat shock became concentrated in nuclei (Fig. <xref ref-type="fig" rid="F1">1B</xref>) showing the typical post-heat shock response [<xref ref-type="bibr" rid="B14">14</xref>]. In order to check whether changes in PLC activity affected the Hsp70 level, we studied its amount in A431 cells treated with EGF, tyrphostin AG1478 (a specific inhibitor of EGF receptor tyrosine kinase) and U73122 (an inhibitor of PLC activity). The level of Hsp70 was measured using Western blot analysis. Neither EGF, nor tyrphostin AG1478 altered significantly the intracellular level of Hsp70. In contrast, treatment of the cells with 1 uM U73122 resulted in a nearly 4-fold reduction in Hsp70 content (Fig. <xref ref-type="fig" rid="F2">2A</xref>). This decrease was observed 10 min after U73122 administration, and there was no recovery until >6 h. The non-active analog of U73122, U73343, had no effect on the Hsp70 level in A431 cells.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>A431 cells originally normally respond to heat stress. A – A431 cells were heated at 42°C for 40 min and lysates of control and heat-shocked cells were analyzed by Western blot using 2H9 anti-Hsp70 antibody. The amount of Hsp70 production shown on the Western blot was quantified by densitometry and normalized to the total protein loaded into each lane of the gel. B – control and heat-stressed cells were stained with the 2H9 anti-Hsp70 monoclonal antibody. Scale bar indicates 5 μm.</p></caption><graphic xlink:href="1475-2867-4-2-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>Effect of factors influencing PLC activity on the Hsp70 level in A431 cells. A – A431 cells were treated with EGF, Tyrphostin AG1478 and U73122, able to affect PLC activity, and lysates of cells were taken in 15 min and in 6 h after administration of the agents. The lysates were subjected to PAGE and examined by Western blot for Hsp70 and Hsc70 as indicated using 2H9 anti-Hsp70 and N69 anti-Hsc70 antibody respectively. The amount of Hsp70 shown on the Western blot was quantified by densitometry and normalized to the total protein loaded into each lane of the gel. B – Hsp70 level is reduced in the A431 cells treated with U73122; immunofluorescence with the use of 2H9 antibody. C – U73343 is not able to decrease the Hsp70 amount in A431 cells; lysates of the cells treated with U73122 and its non-active analog U73343 were analyzed by Western blot, using 2H9 antibody.</p></caption><graphic xlink:href="1475-2867-4-2-2"/></fig><p>To prove the Western blotting data, immunofluorescence analysis was performed using the same antibody (Fig. <xref ref-type="fig" rid="F2">2B</xref>). In U7312-treated cells, Hsp70 almost completely disappeared from cytoplasm and partially passed into nuclei (Fig. <xref ref-type="fig" rid="F2">2B</xref>). The Hsc70 amount measured with the aid of Western blotting using N69 antibody remained constant after the treatment with U73122 (Fig. <xref ref-type="fig" rid="F2">2D</xref>).</p><p>According to the literature, some cells are able to excrete Hsp70 [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. More recently, we have demonstrated such excretion from human T98 glioblastoma cells [<xref ref-type="bibr" rid="B14">14</xref>], and suggested that a reduction of intracellular Hsp70 content following PLC inhibition might be due to a release of the protein by A431 cells. To test this hypothesis, we analyzed cell culture medium using two kinds of affinity chromatography. The technique that was effective and quantitative for detecting Hsp70 in extremely diluted solutions proved to be chromatography on ATP-agarose [<xref ref-type="bibr" rid="B14">14</xref>]. We employed this method in the studies on A431 cells, and applied conditioned medium taken from control cells and from ones incubated with U73122 for 15 min on to column with ATP-agarose gel. The proteins bound to the gels were analysed with the aid of Western blotting using 2H9 antibody. Two bands corresponding to polypeptides 70 and 98 kDa were found on the blot and their intensity was much greater in the case of cells treated with U73122 (Fig. <xref ref-type="fig" rid="F3">3</xref>, left panel).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Inhibition of phospholipase C activity causes release of Hsp70 from A431 cells. The left panel: conditioned medium from A431 cells, control and treated with U73122 PLC inhibitor, was passed over the ATP-agarose column and the eluted protein was subjected to immunoblotting, using 2H9 antibody. The central and right panels: samples of conditioned medium were incubated with 2H9 antibody attached to Protein G-Sepharose, and the immunoprecipitated protein was studied with the help of immunoblotting, using 2H9 antibody (central panel) or anti-ubiquitin antibody (right panel). The lower panel: immunoprecipitation/immunoblotting with lysates of control and U73122-treated cells.</p></caption><graphic xlink:href="1475-2867-4-2-3"/></fig><p>Another method for isolation of Hsp70 from the conditioned medium was immunoprecipitation with 2H9 antibody. The antibody was added to the medium samples and the immune complexes were precipitated with Protein G-Sepharose. Proteins attached to the immunosorbent were analyzed with the aid of Western blotting using 2H9 antibody. Comparison of immunoprecipitates from conditioned medium of control and U73122-treated cells showed that Hsp70 was present in detectable amounts only in the medium from U73122-treated cells (Fig. <xref ref-type="fig" rid="F3">3</xref>, central upper panel). An additional band with the molecular mass of nearly 98 kDa recognized by the monoclonal antibody to Hsp70 was also detected in immunoprecipitates from conditioned medium. To prove that the culture medium was not contaminated with Hsp70, the immunoprecipitation was performed with free DMEM. There were no bands on blots with these samples (data not shown). To reveal relationship between the content of intracellular Hsp70 and its amount in cell medium, the lysates obtained from the cells incubated 15 min with U73122 and from control cells were assayed by immunoprecipitation, using 2H9 antibody (Fig. <xref ref-type="fig" rid="F3">3</xref>, lower panel). The immunoblotting data showed that the accumulation of Hsp70 in culture medium was paralleled to reduction of the endogenous protein content. The result of immunoprecipitation showed that the band of protein with the mass of 98 kDa was much more abundant than that in the material attached to ATP-agarose. We suggest that the monoclonal antibody 2H9 reacted more readily with this protein and to a lesser extent with the principal antigen, Hsp70. In the attempt to identify the 98 kDa component, we followed Jiang and coauthors [<xref ref-type="bibr" rid="B17">17</xref>], who demonstrated that a member of Hsp70 family, the Hsc70 cognate heat shock protein, could be ubiquitinated. Since the 98 kDa protein found in culture medium together with Hsp70 was recognized by monoclonal 2H9 antibody in both immunoprecipitation and Western blotting assays, this protein might be an ubiquitinated form of Hsp70. To test this, immunoprecipitates obtained after the reaction of 2H9 antibody with culture medium of non-treated and U73123-treated cells were probed by Western blotting with anti-ubiquitin antibody (Fig <xref ref-type="fig" rid="F3">3</xref>, right panel). Immunoblot analysis demonstrated that p98 did indeed represent an ubiquitinated form of Hsp70. Thus, inhibition of PLC forced A431 to liberate Hsp70 both in non-modified and ubiquitinated forms.</p><p>Since the reduction of Hsp70 amount might lower cell resistance to apoptotic stimuli, we compared the effect of hydrogen peroxide on control and U73122-treated A431 cells. H<sub>2</sub>O<sub>2 </sub>was chosen due to its apoptogenic effect that was convincingly demonstrated in numerous cell models. Treatment of A431 cells with 200 uM H2O2 did not significantly alter the original level of Hsp70 or its level in cells exposed to H2O2, but in the cells pre-treated with U73122, the Hsp70 level was lower than in the two other cell groups (Fig. <xref ref-type="fig" rid="F4">4A</xref>). Treatment with U73122 alone had no influence on the cell viability and growth dynamics (Fig. <xref ref-type="fig" rid="F4">4B</xref>). Examination of the kinetics of the response to 200 uM of H2O2 treatment revealed that, within 5 h, the U-73122-treated cells began to round up and detach from the plastic. Cells stained with DAPI displayed features typical of apoptosis including condensation and fragmentation of nuclei (Fig. <xref ref-type="fig" rid="F4">4B</xref>). The percentage of apoptotic cells was nearly twice higher than that in the population of control cells (Fig. <xref ref-type="fig" rid="F4">4C</xref>). Thus, the decrease of the Hsp70 amount correlated well with an elevation of sensitivity of A431 cells to hydrogen peroxide.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>The release of Hsp70 from A431 cells makes the latter sensitive to apoptosis-inducing action of hydrogen peroxide. A – data of immunoblotting with 2H9 antibody of control cells and cells treated with hydrogen peroxide and with U73122. B – DAPI staining of A431 cells treated with U73122 alone or in combination with hydrogen peroxide. C – the number of apoptotic cells in population of control cells and cells treated with U73122 alone and with combination of U73122 and hydrogen peroxide.</p></caption><graphic xlink:href="1475-2867-4-2-4"/></fig></sec><sec><title>Discussion</title><p>The goal of this study was to establish a possible relationship between two anti-stress proteins, Hsp70 and phospholipase C. The protective effect resulting from PLC activation was confirmed using different approaches. First, mouse embryonic fibroblasts genetically deficient in PLCγ1 cell viability following heat, or H<sub>2</sub>O<sub>2 </sub>treatment was reduced, while the reconstitution of the enzyme protects cells from stress agents [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Second, overexpression of PLCγ1 was shown to inhibit apoptosis induced by UV-irradiation and by superoxides [<xref ref-type="bibr" rid="B10">10</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Third, inhibition of PLC or activation of protein kinase C, or both, during ischemia impaired significantly the postischemic myocardial recovery [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>On the other hand, it was shown that different kinds of cell damaging factors such as, heat, oxidants, osmotic shock, mechanical stress resulted in PLC activation [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Hsp70 is known to be one of the most powerful anti-stress and anti-apoptotic proteins; its protective activity was convincingly demonstrated in a great number of experiments <italic>in vitro </italic>and <italic>in vivo </italic>[<xref ref-type="bibr" rid="B19">19</xref>]. There are many cross-points, where Hsp70 and PLC can be functionally linked to each other, and to check this A431 epidermoid cell line was chosen. These cells were found to exhibit a normal heat shock response, accumulation of Hsp70 and its transient transport to the nucleus, Fig. <xref ref-type="fig" rid="F1">1</xref>. It was suggested that activation and/or suppression of PLC activity might cause changes in the amount or intracellular distribution of Hsp70. EGF that normally triggers PLC activation failed to affect the content of Hsp70; however, a selective inhibition of PLC activity with U73122 caused a pronounced decrease of the chaperone amount. This reduction occurred in 10 min after administration of the inhibitor and caused considerable loss of the protein in cells and its partial reallocation from the cytoplasm to nuclei (fig. <xref ref-type="fig" rid="F2">2B</xref>). Since the decrease of intracellular Hsp70 level might be due to a release of the protein, we analyzed conditioned medium with the aid of two methods, immunoprecipitation and affinity chromatography on ATP-agarose, the latter known to specifically bind Hsp70 and similar proteins [<xref ref-type="bibr" rid="B20">20</xref>]. As shown by immunoblotting with anti-Hsp70 antibody, immunoprecipitates and eluates from ATP-agarose after passing extracts of U73122-treated cells contained two bands; one of them was identified as Hsp70. For the last few years the idea that Hsp70, earlier considered to be exclusively intracellular molecule, can be released and/or imported by mammalian cells has become rather popular. It dates back to the data of Tytell and coworkers who demonstrated that Hsp70 was able to migrate from glia to giant axon of the sea squid [<xref ref-type="bibr" rid="B15">15</xref>]. Recently, we have shown that human glioma cells could export Hsp70 into the culture medium regardless of whether cells were in normal conditions or subjected to heat shock [<xref ref-type="bibr" rid="B14">14</xref>]. Finally, Dybdahl with coauthors demonstrated that Hsp70 could be fluxed to blood of patients undergoing arterial surgery; moreover, this efflux of Hsp70 seemed to be relevant for the physiology of the whole organism [<xref ref-type="bibr" rid="B21">21</xref>]. Hsp70 plays a major role in protection of cells against a variety of harmful agents, therefore it seemed worth checking whether a reduction of the chaperone amount might affect cell resistance to one of the former factors, superoxide. The data of experiments with the combined treatment of A431 cells with hydrogen peroxide and U73122 showed that the loss of Hsp70 rendered the cells sensitive to H2O2 (fig. <xref ref-type="fig" rid="F4">4</xref>). Our discovery of PLC-dependent extracellular transport of Hsp70 accompanied by loss of cell resistance to risky factors and observations of Dybdahl and others have allowed us to suggest that there are certain loci in the whole organism in which cells affected by a factor similar to U73122 might start to release Hsp70. This protein accumulating in such loci and known to be a strong chaperone or chaperokine [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>] may affect markedly function of adjacent cells in an unpredictable way. For instance, Hsp70 expressed by glial cells could penetrate neurons and protect them from various stressors including heat stress [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B23">23</xref>].</p><p>Another possible target for exogenous Hsp70 is dendritic cells, monocytes and macrophages. In this case Hsp70 induced CD40, TLR2 and TLR4 receptors which are necessary steps toawrds activation of the innate immune system [see [<xref ref-type="bibr" rid="B24">24</xref>] for review]. The mechanism of Hsp70 export by cells remains unclear. Proteins could integrate into the artificial lipid bilayer and a transmembrane ion flow was detected, suggesting involvement of an ion pathway [<xref ref-type="bibr" rid="B25">25</xref>]. It was demonstrated that in rat red blood cells, an hsp70-like protein is located in the cytosol and exported via exosomes during the <italic>in vivo </italic>reticulocyte maturation [<xref ref-type="bibr" rid="B26">26</xref>]. The presence of Hsp70 has been detected in vesicles (named exosomes) leaking from mammalian and avian immature red cells (i.e. reticulocytes), as well as from differentiating avian erythroleukemia cells. It is proposed that Hsp70 takes part in an exosome formation and/or release in immature red cells [<xref ref-type="bibr" rid="B27">27</xref>]. In this regard numerous data are to be mentioned that prove the PLC also is implicated in regulation of exocytosis.</p><p>First, it was shown that exocytosis of vesicles may be blocked by drugs depleting Ca2+ stores and by inhibitors of PLC [<xref ref-type="bibr" rid="B28">28</xref>]. Second, phosphatidylinositol transfer proteins were found to be essential component for PLC-mediated hydrolysis of PIP2 and for the regulation of exocytosis [<xref ref-type="bibr" rid="B29">29</xref>]. Finally, according to our preliminary data pre-treatment of A431 cells with brefeldin, an inhibitor of vesicular transport, stopped the reduction of the Hsp70 in U73122-treated cells (Evdonin et al., unpublished observations). Based on these data we suggest that release of Hsp70 provoked by PLC inhibition occurs via vesicular transport.</p><p>Another interesting finding in this study is that a considerable part of Hsp70 released from A431 cells is in the ubiquitinated form. This fact agrees well with data indicating that the ubiquitination of Hsc70 did not trigger proteasome-dependent degradation [<xref ref-type="bibr" rid="B17">17</xref>], hence, the only way for cell to remove the protein is to excrete it. A number of groups [<xref ref-type="bibr" rid="B30">30</xref>,<xref ref-type="bibr" rid="B31">31</xref>] demonstrated that ubiquitination serves not only as a signal for proteasome-dependent degradation, but also as a trigger for different transport events.</p></sec><sec><title>Conclusions</title><p>In conclusion, the data presented here show that Hsp70 can be released from viable A431 cells in the PLC-dependent manner, and that the exodus of Hsp70 results in elevated sensitivity of the cells to oxidative stress. These observations also prove that orchestration between cell mitogen signaling and stress signaling is necessary for the cell survival in response to stress.</p></sec><sec sec-type="methods"><title>Methods</title><sec sec-type="materials"><title>Materials</title><p>Culture medium and fetal calf serum were purchased in Gibco Life Science (USA), reagents for electrophoresis, ATP-agarose, protein G-Sepharose, PLC inhibitor U73122, its non-active analog U73343 anti ubiquitin antibody and secondary antibodies conjugated with horseradish peroxidase or labeled with Cy3 were from Sigma (St Louis, USA). Tyrphostin AG1478 was purchased from Calbiochem (Lucerne, Switzerland).</p></sec><sec><title>Cells</title><p>A431 cells were grown in the Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum on 10 cm plastic dishes. The cells were treated with 100 ng/ml EGF, 1 uM U73122, 1 uM U73343 or 10 nM Tyrphostine AG1478 for indicated periods of time. To detect Hsp70 in culture medium cells were washed with DMEM, and fresh medium was added followed by the administration of a particular factor.</p></sec><sec><title>SDS-PAGE, immunoprecipitation, and immunoblot analysis</title><p>For immunoprecipitation studies, cells were washed with phosphate-buffered saline (PBS, pH 7.4), lysed on ice in RIPA buffer (250 μl), and then spun at 13 K for 10 min. To the supernatant 10 μl of anti-HSP antibody were added, and the mixture was gently rotated at 4°C for 4 h. Protein G-agarose (10 μl) was then added, and the incubation was continued for an additional 90 min. The protein G-agarose was collected by centrifugation and the beads were washed three times with 50 mM Tris-HCl, pH 8.0, containing 150 mM NaCl, 1% (v/v) Nonidet P-40. The protein G-agarose slurry was suspended in 2 × Laemmli SDS sample buffer and heated at 90°C for 5 min. Immunoprecipitated proteins were analyzed on 12 % polyacrylamide gel followed by the transfer onto the polyvinylidene difluoride membrane (0.4 μm) in 25 mM Tris, 192 mM glycine, and 20% (v/v) methanol [<xref ref-type="bibr" rid="B32">32</xref>,<xref ref-type="bibr" rid="B33">33</xref>]. Membranes were incubated successively 30 min in 5% (w/v) nonfat dry milk, 1 h in 10 mM Tris-HCl, pH 8.0, 125 mM NaCl, 0.2% (v/v), Tween 20 and in the same solution containing antiHsp70 2H9 [1:1000, [<xref ref-type="bibr" rid="B34">34</xref>]], N69 anti-Hsc70 antibodies (1:1000), or anti-ubiquitin antibody (1:100). After washing in 0.05% Tween-PBS blots were incubated 30 min in secondary antibodies with peroxidase diluted 5000-fold in above solution. Protein bands on blots were visualized using enhanced chemiluminescence as described by the manufacturer (Amersham, UK). To equalize amounts of protein loaded on acrylamide gel protein concentration was measured with the use of dye-binding method [<xref ref-type="bibr" rid="B35">35</xref>].</p></sec><sec><title>DAPI staining</title><p>DAPI staining was performed as described previously [<xref ref-type="bibr" rid="B12">12</xref>]. Briefly, prior to staining, the cells were fixed with 4% paraformaldehyde for 30 min at room temperature and washed with PBS. DAPI was added to the fixed cells for 30 min, after what they were examined by fluorescence microscopy. Apoptotic cells were identified by a typical pattern with condensed and fragmented nuclei. The percentage of apoptotic cells was calculated as related to a number of total cells multiplied by 100; minimum 1000 cells were counted for each treatment.</p></sec><sec><title>Immunofluorescence</title><p>Cells grown on uncoated glass coverslips were washed 3 times in PBS. The cells were then fixed with 3.7% formaldehyde for 15 min at room temperature, followed by a brief wash in PBS. The cells were incubated in a blocking solution containing 3% bovine serum albumin in PBS. For permeabilization 0.1% Triton X-100 was added into blocking and the solutions of antibodies. 2H9 anti-Hsp70 antibody was diluted 1:500 in PBS and cells were incubated in this solution 1 h at room temperature. After washing in PBS cells were incubated with goat anti-rabbit Cy3 conjugate, diluted 1:1000, for 30 min at room temperature. The cells were then washed in PBS, and staining was and visualized with the aid of KodakE400 microscope equipped with epifluorescence optics and digital camera.</p></sec><sec><title>Precipitation with ATP-agarose</title><p>The conditioned medium from A431 was collected and TritonX100, MgCl2 and Tris HCl pH 7,6 were added to give concentrations of 0,1%, 1 mM, 20 mM respectively (buffer A). The protein solution was passed through 100 μl ATP-agarose column, followed by washing with a buffer A without Triton X-100. A column was eluted with the buffer A, containing 3 mM ATP. Eluted proteins were resolved by SDS-Page electrophoresis followed by blotting with 2H9 antibody.</p></sec></sec><sec><title>Abbreviations</title><p>The abbreviations used are: PLC – phospholipase C; EGF – epidermal growth factor.</p></sec>
Inter-allelic recombination in the <italic>Plasmodium vivax </italic>merozoite surface protein 1 gene among Indian and Colombian isolates	<sec><title>Background</title><p>A major concern in malaria vaccine development is the polymorphism observed among different <italic>Plasmodium </italic>isolates in different geographical areas across the globe. The merozoite surface protein 1 (MSP-1) is a leading vaccine candidate antigen against asexual blood stages of malaria parasite. To date, little is known about the extent of sequence variation in the <italic>Plasmodium vivax </italic>MSP-1 gene (<italic>Pvmsp-1</italic>) among Indian isolates. Since <italic>P. vivax </italic>accounts for >50% of malaria cases in India and in Colombia, it is essential to know the <italic>Pvmsp-1 </italic>gene variability in these two countries to sustain it as a vaccine candidate. The extent of polymorphism in <italic>Pvmsp-1 </italic>gene among Indian and Colombian isolates is described.</p></sec><sec sec-type="methods"><title>Methods</title><p>The sequence variation in the region encompassing the inter-species conserved blocks (ICBs) five and six of <italic>Pvmsp-1 </italic>gene was examined. PCR was carried out to amplify the polymorphic region of <italic>Pvmsp-1 </italic>and the PCR products from twenty (nine Indian and 11 Colombian) isolates were sequenced and aligned with Belem and Salvador-1 sequences.</p></sec><sec><title>Results</title><p>Results revealed three distinct types of sequences among these isolates, namely, Salvador-like, Belem-like and a third type sequence which was generated due to interallelic recombination between Salvador-like sequences and Belem-like sequences. Existence of the third type in majority (44%) showed that allelic recombinations play an important role in PvMSP1 diversity in natural parasite population. Micro-heterogeneity was also seen in a few of these isolates due to nucleotide substitutions, insertions as well as deletions.</p></sec><sec><title>Conclusions</title><p>Intergenic recombination in the <italic>Pvmsp-1 </italic>gene was found and suggest that this is the main cause for genetic diversity of the <italic>Pvmsp-1 </italic>gene.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Maestre</surname><given-names>Amanda</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Sunil</surname><given-names>Sujatha</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Ahmad</surname><given-names>Gul</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Mohmmed</surname><given-names>Asif</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Echeverri</surname><given-names>Marcela</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Corredor</surname><given-names>Mauricio</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Blair</surname><given-names>Silvia</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Chauhan</surname><given-names>Virander S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A9" corresp="yes" contrib-type="author"><name><surname>Malhotra</surname><given-names>Pawan</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Malaria Journal	<sec><title>Background</title><p>There is an intense global effort to develop an effective vaccine in addition to the malaria control measures currently in use. Several vaccine candidate antigens have been identified against different stages of the two main human malaria parasites, <italic>Plasmodium falciparum </italic>and <italic>Plasmodium vivax</italic>, and are being developed as a part of a subunit vaccine [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. One of the major concerns in malaria vaccine development is the polymorphic nature of the candidate vaccine antigens. Several <italic>in vitro </italic>and epidemiological studies have demonstrated that natural variations can abrogate immune recognition [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. Thus, studies of sequence and antigenic diversity of malaria vaccine candidate antigens become a subject of considerable importance.</p><p>The merozoite surface protein 1 (MSP-1) of <italic>Plasmodium </italic>sp. is a large polypeptide of ~200 kDa and is one of the leading asexual blood stage vaccine candidate antigens [<xref ref-type="bibr" rid="B5">5</xref>]. A number of experimental studies with native and recombinant MSP-1, particularly its C-terminal fragments, have demonstrated the vaccine potential of MSP-1 [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. Polymorphism has been reported in the <italic>P. falciparum </italic>MSP-1 (<italic>Pfmsp-1</italic>) gene among different isolates across different geographical areas [<xref ref-type="bibr" rid="B3">3</xref>]. Based on sequence variations, the <italic>Pfmsp-1 </italic>gene has been divided into a number of blocks that are conserved, semi-conserved and variable in different species and isolates [<xref ref-type="bibr" rid="B5">5</xref>]. Broadly, <italic>Pfmsp-1 </italic>gene sequences have been classified into two allelic families, Wellcome and PNG-MAD 20 type [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. In addition, intragenic recombinations have been reported among these two parental alleles, resulting in polymorphism among different isolates around the world [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p><p>The gene encoding the merozoite surface protein 1 of <italic>P. vivax </italic>(<italic>Pvmsp-1</italic>) shows many similarities with those from other malaria species. The gene consists of 10 relatively conserved blocks alternating with regions of high diversity [<xref ref-type="bibr" rid="B5">5</xref>]. Like the <italic>Pfmsp-1 </italic>gene, the <italic>Pvmsp-1 </italic>gene has also been shown to be dimorphic; identified as Belem, type 1 [<xref ref-type="bibr" rid="B11">11</xref>] and Salvador-1, type 2 [<xref ref-type="bibr" rid="B12">12</xref>] forms. In comparison to <italic>Pfmsp-1 </italic>polymorphism studies, studies on <italic>Pvmsp-1 </italic>polymorphism are limited. Several regions of <italic>Pvmsp-1 </italic>have been amplified and sequenced from field isolates collected from Sri Lanka, Colombia, Brazil, Thailand, Korea and China [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B18">18</xref>]. A third allele type (type 3) has been reported among these isolates and it has been suggested that this allele has arisen due to intragenic recombination between the Belem and Salvador alleles [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. Recently, a detailed study of <italic>P. vivax </italic>merozoite surface protein 1 (<italic>Pvmsp-1</italic>) gene of 40 isolates from different geographical areas (Thailand, Brazil, Bangladesh, South Korea, Vanuatu and Japan) revealed mosaic organization of the <italic>Pvmsp-1 </italic>gene and heterogeneity in the frequency of allelic recombination among different isolates [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>The genetic characteristics of <italic>P. vivax </italic>circulating in distant endemic regions have not been studied so far. Therefore, the molecular basis underlying phenomena such as resistance of <italic>P. vivax </italic>to chloroquine in America and Asia (20,21) and effective immune responses (which is essential for vaccine design), have not being fully understood. Since PvMSP-1 is a potential vaccine candidate for <italic>P. vivax</italic>, it is important to further analyse the extent of polymorphism from field isolates across the globe. In the present study, the extent of polymorphism in one of the variable regions of <italic>Pvmsp-1 </italic>gene among Indian and Colombian isolates was investigated. This study is important as there is no information of <italic>Pvmsp-1 </italic>alleles and their diversity among Indian isolates in spite of <italic>P. vivax </italic>being one of the major causes of malaria in India [<xref ref-type="bibr" rid="B22">22</xref>].</p></sec><sec><title>Subjects and methods</title><sec><title>Study area and blood collection</title><p>To study the polymorphism in different <italic>P. vivax </italic>isolates, a polymorphic region of the MSP-1 between blocks 5–6 corresponding to base pairs 1984–2653 (Belem strain) was amplified by PCR (Fig. <xref ref-type="fig" rid="F1">1</xref>). To begin with, venous blood samples were collected from 50 Colombian and 25 Indian symptomatically infected patients after verbal consent. Colombian isolates were from the malaria clinics of the northwest region of Colombia (Turbo). Between 1996–2000, this region recorded highest number of malaria cases with a mean parasite annual index of about 40. <italic>P. vivax </italic>accounts for 60% of the total number of malaria cases (23). Indian isolates were from the north (Aligarh) and northeast (Nagaland) regions of India, which are more than 500 km apart. In Aligarh, <italic>P. vivax </italic>prevalence is high (60%) while in Nagaland, <italic>P. vivax </italic>accounts for 40% of the total number of malaria cases (24). Blood was either collected into EDTA-containing vials or spotted onto Whatman-3 filter sheets.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Schematic representation of the <italic>P. vivax msp-1 </italic>gene. Boxes representing interspecies conserved blocks (ICBs), conserved blocks (CBs) and variable blocks are filled, hatched and opened respectively. A scale bar for nucleotide number and relative position of primers (Pv1 and Pv2) used for the polymerase chain reaction are indicated at the top and bottom respectively. Sequence and location of the primers used for PCR amplification are indicated at the bottom.</p></caption><graphic xlink:href="1475-2875-3-4-1"/></fig></sec><sec><title>DNA preparation and PCR amplification</title><p><italic>P. vivax </italic>genomic DNA was isolated from filter discs using the Chelax extraction procedure [<xref ref-type="bibr" rid="B25">25</xref>] and from infected blood by proteinase K digestion, followed by four rounds of phenol-chloroform extraction. PCR was carried out to amplify the polymorphic region of <italic>Pvmsp-1 </italic>for 30 cycles of 94°C for one minute 55°C for one minute and 72°C for two minutes and a final primer extension at 72°C for five minutes. The forward primer used was 5'-GGGAATTCTACTTGATGGTCCTC-3' (PV1), and the reverse primer was 5'-GGAATTCTTGTGACATGCGTAAGCG-3' (PV2).</p></sec><sec><title>Sequencing, alignment and data analysis</title><p>PCR products from twenty (nine Indian and 11 Colombian) isolates were run on a 2% agarose gel and the fragments were purified from gel using a Novagen gel DNA kit. Gel purified fragments were sequenced in both directions using an automated DNA sequencer, ABI Prism 310 (Applied Biosystems). Two independent PCR products were sequenced for each isolate. The deduced amino acid sequences of the PvMSP-1 from twenty of the isolates are shown in figure <xref ref-type="fig" rid="F2">2</xref>, aligned with the corresponding Belem and/or Salvador sequences. Based on sequence comparison data, the sequences were classified. These DNA sequences are deposited under GenBank and EMBL Accession Nos. AY 229861 – AY 229867, AJ 582077-79, AJ 582111-20</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Amino acid sequence alignment of <italic>PvMSP-1 </italic>polymorphic region between blocks V-VI of twenty field isolates from India and Colombia with that of Salvador-1 (Sal-1) and Belem isolates. Dots and Dashes represent identical residues and deletions respectively.</p></caption><graphic xlink:href="1475-2875-3-4-2"/></fig></sec></sec><sec><title>Results and discussion</title><sec><title>PCR amplification</title><p>PCR-amplification revealed size variations (~450 bp and ~520 bp) among these isolates. Size variations were seen in both Colombian as well as Indian isolates. Among the Colombian isolates, twenty-nine out of fifty isolates showed amplification and sixteen out of twenty five Indian isolates showed amplification. Agarose gel electrophoresis revealed a predominance of ~520 bp species among Indian isolates.</p></sec><sec><title>Classification of isolates</title><p>Based on sequence comparison data, we classified these sequences into three basic types: Type 1, Type 2 and Type 3a. This classification was done as previously described by [<xref ref-type="bibr" rid="B15">15</xref>] for the Thai isolates. Type 1 had sequences similar to the Belem allele, but with a lesser number of Gln residues in the middle. Isolates CM 12AN, CM 13N, CM 17N and CM 19N showed 16–18 glutamine residues in comparison to Belem allele that had 23 Gln residues. These four Colombian isolates also showed few other variations as a result of nucleotide substitutions: All these isolates showed Val (GTA) at position 742 instead of Asp (GAT) seen in Belem allele. Likewise isolates CM 13N, CM 17N and CM 19N had Asp (GAT) for Gly (GGT) at position 810. Similar substitutions had been reported for Thai isolates TD439A, TD424 and TD 425A. In the present study we did not find Type 1 sequence among Indian isolates. Type 2 sequences, which are similar to Salvador alleles, were found among both Indian and Colombian isolates. These sequences showed slight variations from the Salvador allele. Such variations have been previously reported for the Thai isolates [<xref ref-type="bibr" rid="B15">15</xref>]: a 3 nt insertion encoding Gln (CAA) and Pro (CCA) was seen in isolates CM 51, CM 70, IA 14 and IA 6, respectively between Gln 744 and Pro 745 of the Salvador sequence. One such insertion Gln (CAA) has been previously reported for a Thai isolate T128. In addition, four non-synonymous nucleotide substitutions were seen in these Indian as well as Colombian isolates: CM 51 and IN 8 had a change from Val (GTA)→Ala (GCA) at position 750, that was also reported in Thai isolates T117, T128 and TD414. Isolates CM 51, CM 19, CM 70, IN 8, IA 6 and IA 14 had Thr (ACT)→Ile (ATT) at position 796 as seen among Thai isolates TD 207A and T117. There was a change from Ala (GCC)→Val (GTC) at position 813 in isolates CM 19, IA 6 and IA 14 similar to the Thai isolate T 117. Likewise, CM 19, IN 1 and IA 14 had a Glu (GAG)→Gln (CAG) substitution at position 839, which had been reported in isolates from other geographical origins [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>Type 3a was the third type of sequence seen among these isolates (Fig. 3). This type of sequence is characterized by a combination of a Salvador-like sequence at the 5' half of this region and a Belem-like sequence at the 3' end. Nine of the twenty isolates (~44%) sequenced showed this type of recombination. Similar types of recombination have been reported earlier for Thai, South Korean, Chinese, Sri Lankan as well as in Colombian isolates (16). The number of Gln repeats varied between 15 to 18 among these isolates. These sequences were conserved in regions 5' to Gln repeats. However, two out of nine isolates showed amino acid changes at two positions in the 3' end (Fig <xref ref-type="fig" rid="F2">2</xref>). The presence of a type 3a sequence in more than 40% of the isolates in the present study indicates that inter-allelic recombination is one of the main causes for <italic>Pvmsp-1 </italic>gene diversity among <italic>P. vivax </italic>Indian and Colombian populations.</p><p>The main goal of the present study was to determine and compare the extent of polymorphism in <italic>Pvmsp-1 </italic>in two different geographical regions of the world (India and Colombia). This is important, as data regarding the sequence variations among <italic>P. vivax </italic>Indian field isolates is scanty. A region was selected between the two interspecies conserved blocks 5 and 6 of <italic>Pvmsp-1 </italic>for this study. Though many recombination sites have been identified within and between the variable blocks of the <italic>Pvmsp-1 </italic>gene, this region was chosen as it has emerged as an important genetic marker for <italic>P. vivax </italic>polymorphism to detect mixed infection as well as recombination events between the two parental Salvador and Belem alleles [<xref ref-type="bibr" rid="B19">19</xref>]. Results of the present study showed that in both Indian and Colombian isolates, a similar degree of polymorphism for the <italic>Pvmsp-1 </italic>gene exists and inter-allelic recombination is quite common among these isolates. It is noteworthy that most of the amino acid changes among Indian and Colombian isolates match with the changes seen in other isolates from other geographical areas in particular, the Thai isolates. Though Belem-like sequence was not observed among any Indian isolates in the present study, it may be due to limited sample size. The presence of the Type 3a sequence among Indian isolates clearly shows that Belem-like isolates exist among Indian population. There are two schools of thought with regards to <italic>Pvmsp-1 </italic>polymorphism. Some authors have stated that intragenic recombination is a rare event in <italic>P. vivax </italic>and that is not the main cause for antigenic polymorphism [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>]. Other groups argue that recombination is the main source of antigenic diversity in PvMSP-1 [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. Another feature of the recombination events in <italic>P. vivax </italic>isolates in the present study was that the recombination took place in a polymorphic segment of the gene, the polyQ region. This is very much similar to that reported for other isolates in Asia [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Thus, the recombination events in <italic>Pvmsp-1 </italic>differ from <italic>Pfmsp-1 </italic>where recombination between <italic>MSP-1 </italic>alleles occurs in the conserved segments surrounding the polymorphic regions whereas in <italic>Pvmsp-1</italic>, the recombination occurs at the polymorphic regions [<xref ref-type="bibr" rid="B8">8</xref>].</p></sec></sec><sec><title>Conclusion</title><p>The existence of a large number of Type 3a sequences among Indian and Colombian isolates generated by the fusion of Belem and Salvador sequences suggests that recombination is the main mechanism in Pvmsp-1 allelic diversity generation among the isolates around the globe. It is believed that long lasting and recurrent parasitemia often associated with relapse, a characteristic of <italic>P. vivax </italic>is responsible for this high rate of interallelic recombination. As allelic polymorphism is one of the greatest hurdles in mounting a protective immune response against the genetically diverse <italic>P. vivax</italic>, further experiments are required to determine the impact of recombination events in the immunogenicity of PvMSP-1 antigen in natural populations.</p></sec><sec><title>Authors' contributions</title><p>MA was involved in amplification of the samples. SS was involved in amplification, sequencing of the samples and preparation of the manuscript. Both MA and SS contributed equally. GA participated in sample collection in India. AM was involved in sample collection and sequence alignment. EM, CM and BS were involved in sample collection from Colombia. VSC, the group leader participated in the design and co-ordination of the study. PM supervised the overall work.</p></sec>
Epidemiology and clinical features of vivax malaria imported to Europe: Sentinel surveillance data from TropNetEurop	<sec><title>Background</title><p><italic>Plasmodium vivax </italic>is the second most common species among malaria patients diagnosed in Europe, but epidemiological and clinical data on imported <italic>P. vivax </italic>malaria are limited. The TropNetEurop surveillance network has monitored the importation of vivax malaria into Europe since 1999.</p></sec><sec><title>Objectives</title><p>To present epidemiological and clinical data on imported <italic>P. vivax </italic>malaria collected at European level.</p></sec><sec sec-type="materials\|methods"><title>Material and methods</title><p>Data of primary cases of <italic>P. vivax </italic>malaria reported between January 1999 and September 2003 were analysed, focusing on disease frequency, patient characteristics, place of infection, course of disease, treatment and differences between network-member countries.</p></sec><sec><title>Results</title><p>Within the surveillance period 4,801 cases of imported malaria were reported. 618 (12.9%) were attributed to <italic>P. vivax</italic>. European travellers and immigrants were the largest patient groups, but their proportion varied among the reporting countries. The main regions of infection in descending order were the Indian subcontinent, Indonesia, South America and Western and Eastern Africa, as a group accounting for more than 60% of the cases. Regular use of malaria chemoprophylaxis was reported by 118 patients. With 86 (inter-quartile range 41–158) versus 31 days (inter-quartile range 4–133) the median symptom onset was significantly delayed in patients with chemoprophylaxis (p < 0.0001). Common complaints were fever, headache, fatigue, and musculo-skeletal symptoms. All patients survived and severe clinical complications were rare. Hospitalization was provided for 60% and primaquine treatment administered to 83.8% of the patients, but frequencies varied strongly among reporting countries.</p></sec><sec><title>Conclusions</title><p>TropNetEurop data can contribute to the harmonization of European treatment policies.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Mühlberger</surname><given-names>N</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Jelinek</surname><given-names>T</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Gascon</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Probst</surname><given-names>M</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Zoller</surname><given-names>T</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Schunk</surname><given-names>M</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Beran</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Gjørup</surname><given-names>I</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>[email protected]</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Behrens</surname><given-names>RH</given-names></name><xref ref-type="aff" rid="I7">7</xref><email>[email protected]</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Clerinx</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I8">8</xref><email>[email protected]</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Björkman</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I9">9</xref><email>[email protected]</email></contrib><contrib id="A12" contrib-type="author"><name><surname>McWhinney</surname><given-names>P</given-names></name><xref ref-type="aff" rid="I10">10</xref><email>[email protected]</email></contrib><contrib id="A13" contrib-type="author"><name><surname>Matteelli</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I11">11</xref><email>[email protected]</email></contrib><contrib id="A14" contrib-type="author"><name><surname>Lopez-Velez</surname><given-names>R</given-names></name><xref ref-type="aff" rid="I12">12</xref><email>[email protected]</email></contrib><contrib id="A15" contrib-type="author"><name><surname>Bisoffi</surname><given-names>Z</given-names></name><xref ref-type="aff" rid="I13">13</xref><email>[email protected]</email></contrib><contrib id="A16" contrib-type="author"><name><surname>Hellgren</surname><given-names>U</given-names></name><xref ref-type="aff" rid="I14">14</xref><email>[email protected]</email></contrib><contrib id="A17" contrib-type="author"><name><surname>Puente</surname><given-names>S</given-names></name><xref ref-type="aff" rid="I15">15</xref><email>[email protected]</email></contrib><contrib id="A18" contrib-type="author"><name><surname>Schmid</surname><given-names>ML</given-names></name><xref ref-type="aff" rid="I16">16</xref><email>[email protected]</email></contrib><contrib id="A19" contrib-type="author"><name><surname>Myrvang</surname><given-names>B</given-names></name><xref ref-type="aff" rid="I17">17</xref><email>[email protected]</email></contrib><contrib id="A20" contrib-type="author"><name><surname>Holthoff-Stich</surname><given-names>ML</given-names></name><xref ref-type="aff" rid="I18">18</xref><email>[email protected]</email></contrib><contrib id="A21" contrib-type="author"><name><surname>Laferl</surname><given-names>H</given-names></name><xref ref-type="aff" rid="I19">19</xref><email>[email protected]</email></contrib><contrib id="A22" contrib-type="author"><name><surname>Hatz</surname><given-names>C</given-names></name><xref ref-type="aff" rid="I20">20</xref><email>[email protected]</email></contrib><contrib id="A23" contrib-type="author"><name><surname>Kollaritsch</surname><given-names>H</given-names></name><xref ref-type="aff" rid="I21">21</xref><email>[email protected]</email></contrib><contrib id="A24" contrib-type="author"><name><surname>Kapaun</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I22">22</xref><email>[email protected]</email></contrib><contrib id="A25" contrib-type="author"><name><surname>Knobloch</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I23">23</xref><email>[email protected]</email></contrib><contrib id="A26" contrib-type="author"><name><surname>Iversen</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I24">24</xref><email>[email protected]</email></contrib><contrib id="A27" contrib-type="author"><name><surname>Kotlowski</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I25">25</xref><email>[email protected]</email></contrib><contrib id="A28" contrib-type="author"><name><surname>Malvy</surname><given-names>DJM</given-names></name><xref ref-type="aff" rid="I26">26</xref><email>[email protected]</email></contrib><contrib id="A29" contrib-type="author"><name><surname>Kern</surname><given-names>P</given-names></name><xref ref-type="aff" rid="I27">27</xref><email>[email protected]</email></contrib><contrib id="A30" contrib-type="author"><name><surname>Fry</surname><given-names>G</given-names></name><xref ref-type="aff" rid="I28">28</xref><email>[email protected]</email></contrib><contrib id="A31" contrib-type="author"><name><surname>Siikamaki</surname><given-names>H</given-names></name><xref ref-type="aff" rid="I29">29</xref><email>[email protected]</email></contrib><contrib id="A32" contrib-type="author"><name><surname>Schulze</surname><given-names>MH</given-names></name><xref ref-type="aff" rid="I30">30</xref><email>[email protected]</email></contrib><contrib id="A33" contrib-type="author"><name><surname>Soula</surname><given-names>G</given-names></name><xref ref-type="aff" rid="I31">31</xref><email>[email protected]</email></contrib><contrib id="A34" contrib-type="author"><name><surname>Paul</surname><given-names>M</given-names></name><xref ref-type="aff" rid="I32">32</xref><email>[email protected]</email></contrib><contrib id="A35" contrib-type="author"><name><surname>Prat</surname><given-names>J Gómez i</given-names></name><xref ref-type="aff" rid="I33">33</xref><email>[email protected]</email></contrib><contrib id="A36" contrib-type="author"><name><surname>Lehmann</surname><given-names>V</given-names></name><xref ref-type="aff" rid="I34">34</xref><email>[email protected]</email></contrib><contrib id="A37" contrib-type="author"><name><surname>Bouchaud</surname><given-names>O</given-names></name><xref ref-type="aff" rid="I35">35</xref><email>[email protected]</email></contrib><contrib id="A38" contrib-type="author"><name><surname>Cunha</surname><given-names>S da</given-names></name><xref ref-type="aff" rid="I36">36</xref><email>[email protected]</email></contrib><contrib id="A39" contrib-type="author"><name><surname>Atouguia</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I37">37</xref><email>[email protected]</email></contrib><contrib id="A40" contrib-type="author"><name><surname>Boecken</surname><given-names>G</given-names></name><xref ref-type="aff" rid="I38">38</xref><email>[email protected]</email></contrib>	Malaria Journal	<sec><title>Introduction</title><p><italic>Plasmodium vivax</italic>, has a major adverse impact on global health [<xref ref-type="bibr" rid="B1">1</xref>], accounting for 70–80 million clinical cases annually. It is responsible for over 50% of malaria outside Africa, notably in Southeast Asia and Central and South America, and particularly on the Indian subcontinent. It also accounts for around 10% of cases in Eastern and Southern Africa but has only limited prevalence in West Africa, presumably due to the presence of Duffy-negative blood group variants that limit erythrocyte invasion by the parasite [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. Similar to <italic>Plasmodium falciparum, P. vivax </italic>may cause severe anaemia, but major complications like cerebral malaria, hypoglycaemia, metabolic acidosis and respiratory distress observed in <italic>P. falciparum </italic>malaria, do not occur [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B6">6</xref>].</p><p>When imported to non-endemic areas, vivax malaria, despite its uncomplicated course, requires special attention for two reasons. First, diagnosis often is complicated by the late onset of symptoms which unlike those observed in falciparum malaria, may occur several months after arrival from endemic areas [<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B10">10</xref>]. Second, case management is complicated by the fact that parasites can remain dormant in the liver as hypnozoites. Thus, even if blood stages of the parasite are cleared, reactivation of these liver forms may cause relapses within a few months [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B11">11</xref>].</p><p>Malaria is a notifiable disease in all European countries [<xref ref-type="bibr" rid="B12">12</xref>]. As shown by national surveillance reports [<xref ref-type="bibr" rid="B13">13</xref>-<xref ref-type="bibr" rid="B19">19</xref>], <italic>P. vivax </italic>is presently the second most frequent cause of imported malaria in most of Europe, except France [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B20">20</xref>], accounting for up to 40% of the annual cases in single countries. However, since <italic>P. vivax </italic>is less virulent than <italic>P. falciparum </italic>[<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B6">6</xref>], surveillance reports primarily focus on <italic>P. falciparum </italic>infections. Epidemiological and clinical details of imported <italic>P. vivax </italic>malaria are, therefore, hardly described.</p><p>Since 1999, the importation of vivax malaria into Europe has also been systematically monitored by the TropNetEurop surveillance network. The present article summarises epidemiological and clinical data of the disease collected during the first 56 months of surveillance at European level, focusing on disease frequency, patient characteristics, place of infection, course of disease and treatment. Differences between network-member countries are highlighted.</p></sec><sec><title>Patients, materials and methods</title><p>The European network TropNetEurop was founded in 1999 to effectively detect emerging infections of potential regional, national or global impact at their point of entry into the domestic population, to link clinical and epidemiological knowledge and to serve as a platform for multi-centre research. Sentinel surveillance reporting is carried out by participating sites by use of a standardised and computerised reporting system. Electronic transmission of anonymous case information, comprising standardised demographic, epidemiological and clinical data, to the central database assures rapid detection of sentinel events. Presently 46 clinical sites from 16 European countries are organised in the network. Additional information about TropNetEurop and the reporting instruments can be received from the internet at <ext-link ext-link-type="uri" xlink:href="http://www.tropnet.net"/>.</p><p>The present work summarises data of primary cases of <italic>P. vivax </italic>malaria reported within the network between January 1999 and September 2003. Relapse notifications were excluded from the database to avoid multiple-counting of importation events. Mixed-infections with <italic>P. vivax </italic>and other Plasmodium species were included. However, while analyses focusing on disease frequency, patient characteristics and place of infection were based on all observations, analyses with clinical end point were restricted to mono-infections elicited microscopically (confirmed cases) or by antibody detection (probable cases) in order to reduce bias. Patients with mixed plasmodial or other concomitant infections, cases diagnosed by clinical reasoning (suspected cases) and cases reported without specification of the underlying diagnostic proof (unclassified cases) were excluded from those analyses.</p><p>All analyses were done with the SAS software (release 8.01 by SAS Institute Inc., Cary, NC, USA). Missing values were assumed to be non-systematic. Incomplete cases were therefore disregarded in single calculations. However, sample size or missing data information is given with each result.</p></sec><sec><title>Results</title><p>Between January 1999 and September 2003, a total of 4,801 patients with travel-related malaria were reported within the TropNetEurop network. <italic>P. vivax </italic>was involved in 618 (12.9%) cases, either as sole pathogen (n = 585) or in mixed-infections with other species (n = 33), thus accounting for the second highest number of cases after <italic>P. falciparum</italic>. The reported proportion of <italic>P. vivax </italic>was rather steady with 10.9% in 1999, 12.6% in 2000, 15.1% in 2001, 12.3% in 2002 and 12.9% in 2003. All 16 TropNetEurop countries reported <italic>P. vivax </italic>malaria. However, as shown in table <xref ref-type="table" rid="T1">1</xref>, the number of cases varied strongly between countries. Germany (24.3%), Spain (15.5%) and the UK (12.0%) reported most cases, whereas reports from Switzerland (1.8%), Poland (1.6%), Finland (1.0%), Ireland (1.0%) and Portugal (0.3%) were scarce. According to diagnostic information 557 (90.1%) of the 618 infections were confirmed, two (0.3%) were probable and eight (1.3%) were suspected cases. The remaining 51 (8.3%) could not be classified, due to missing information on the underlying diagnostic procedure. Those and the suspected cases were excluded from analyses with clinical endpoints in the latter.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Frequency of P. vivax malaria by year and reporting country</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center"><bold>1999</bold></td><td align="center"><bold>2000</bold></td><td align="center"><bold>2001</bold></td><td align="center"><bold>2002</bold></td><td align="center"><bold>2003</bold></td><td align="center"><bold>Total</bold></td></tr><tr><td align="left"><bold>Country</bold></td><td align="center"><bold>n (%)</bold></td><td align="center"><bold>n (%)</bold></td><td align="center"><bold>n (%)</bold></td><td align="center"><bold>n (%)</bold></td><td align="center"><bold>n (%)</bold></td><td align="center"><bold>n (%)</bold></td></tr></thead><tbody><tr><td align="left">Austria</td><td align="center">2 (2.1)</td><td align="center"> 8 (4.9)</td><td align="center">5 (2.7)</td><td align="center">5 (3.9)</td><td align="center">3 (6.3)</td><td align="center">23 (3.7)</td></tr><tr><td align="left">Belgium</td><td></td><td align="center">17 (10.4)</td><td align="center">8 (4.3)</td><td align="center">7 (5.5)</td><td align="center">2 (4.2)</td><td align="center">34 (5.5)</td></tr><tr><td align="left">Czech Rep.</td><td align="center">9 (9.6)</td><td align="center">13 (7.9)</td><td align="center">20 (10.8)</td><td align="center">5 (3.9)</td><td align="center">3 (6.3)</td><td align="center">50 (8.1)</td></tr><tr><td align="left">Denmark</td><td align="center">14 (14.9)</td><td align="center">16 (9.8)</td><td align="center">12 (6.5)</td><td align="center">5 (3.9)</td><td align="center">1 (2.1)</td><td align="center">48 (7.8)</td></tr><tr><td align="left">Finland</td><td></td><td align="center">1 (0.6)</td><td></td><td align="center">6 (4.7)</td><td></td><td align="center">6 (1.0)</td></tr><tr><td align="left">France</td><td align="center">1 (1.1)</td><td align="center">1 (0.6)</td><td align="center">4 (2.2)</td><td align="center">7 (5.5)</td><td align="center">2 (4.2)</td><td align="center">15 (2.4)</td></tr><tr><td align="left">Germany</td><td align="center">31 (33.0)</td><td align="center">35 (21.3)</td><td align="center">51 (27.6)</td><td align="center">22 (17.3)</td><td align="center">11 (22.9)</td><td align="center">150 (24.3)</td></tr><tr><td align="left">Ireland</td><td></td><td align="center">2 (1.2)</td><td align="center">1 (0.5)</td><td align="center">1 (0.8)</td><td align="center">2 (4.2)</td><td align="center">6 (1.0)</td></tr><tr><td align="left">Italy</td><td align="center">8 (8.5)</td><td align="center">10 (6.1)</td><td align="center">16 (8.6)</td><td align="center">3 (2.4)</td><td></td><td align="center">37 (6.0)</td></tr><tr><td align="left">Norway</td><td></td><td></td><td align="center">4 (2.2)</td><td align="center">8 (6.3)</td><td align="center">1 (2.1)</td><td align="center">13 (2.1)</td></tr><tr><td align="left">Poland</td><td align="center">3 (3.2)</td><td></td><td align="center">4 (2.2)</td><td align="center">3 (2.4)</td><td></td><td align="center">10 (1.6)</td></tr><tr><td align="left">Portugal</td><td></td><td></td><td></td><td align="center">1 (0.8)</td><td align="center">1 (2.1)</td><td align="center">2 (0.3)</td></tr><tr><td align="left">Spain</td><td align="center">15 (16.0)</td><td align="center">31 (18.9)</td><td align="center">23 (12.4)</td><td align="center">23 (18.1)</td><td align="center">4 (8.3)</td><td align="center">96 (15.5)</td></tr><tr><td align="left">Sweden</td><td align="center">4 (4.3)</td><td align="center">8 (4.9)</td><td align="center">15 (8.1)</td><td align="center">8 (6.3)</td><td align="center">8 (16.7)</td><td align="center">43 (7.0)</td></tr><tr><td align="left">Switzerland</td><td align="center">4 (4.3)</td><td align="center">1 (0.6)</td><td></td><td align="center">5 (3.9)</td><td align="center">1 (2.1)</td><td align="center">11 (1.8)</td></tr><tr><td align="left">UK</td><td align="center">3 (3.2)</td><td align="center">22 (13.4)</td><td align="center">22 (11.9)</td><td align="center">18 (14.2)</td><td align="center">9 (18.8)</td><td align="center">74 (12.0)</td></tr><tr><td colspan="7"><hr></hr></td></tr><tr><td align="left"><bold>Total</bold></td><td align="center"><bold>94 (100)</bold></td><td align="center"><bold>164 (100)</bold></td><td align="center"><bold>185 (100)</bold></td><td align="center"><bold>127 (100)</bold></td><td align="center"><bold>48 (100)</bold></td><td align="center"><bold>616 (100)</bold></td></tr></tbody></table><table-wrap-foot><p> Data reported between January and September 2003</p></table-wrap-foot></table-wrap><p>Table <xref ref-type="table" rid="T2">2</xref> describes characteristics of the 618 patients. About 2/3 were male. The median age was 32 years (inter-quartile range 26–43). About 50% reported reception of pre-travel advice, 42.2% stated use of anti-malarial chemoprophylaxis. Among those with prophylaxis, 62.1% stated compliance with the recommended regimen.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Characteristics of patients with P. vivax malaria</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">N</td><td align="center">%</td></tr></thead><tbody><tr><td align="left"><bold>Sex (n = 615)</bold></td><td></td><td></td></tr><tr><td align="left"> <bold>Male</bold></td><td align="center">422</td><td align="center">68.6</td></tr><tr><td align="left"> <bold>Female</bold></td><td align="center">193</td><td align="center">31.4</td></tr><tr><td align="left"><bold>Pre-travel advice (n = 387)</bold></td><td></td><td></td></tr><tr><td align="left"> <bold>Yes</bold></td><td align="center">208</td><td align="center">53.7</td></tr><tr><td align="left"> <bold>No</bold></td><td align="center">179</td><td align="center">46.3</td></tr><tr><td align="left"><bold>Prophylaxis (n = 554)</bold></td><td></td><td></td></tr><tr><td align="left"> <bold>None</bold></td><td align="center">320</td><td align="center">57.8</td></tr><tr><td align="left"> <bold>Chloroquine</bold></td><td align="center">39</td><td align="center">7.0</td></tr><tr><td align="left"> <bold>Proguanil</bold></td><td align="center">6</td><td align="center">1.1</td></tr><tr><td align="left"> <bold>Proguanil + Chloroquine</bold></td><td align="center">45</td><td align="center">8.1</td></tr><tr><td align="left"> <bold>Mefloquine</bold></td><td align="center">123</td><td align="center">22.2</td></tr><tr><td align="left"> <bold>Doxycycline</bold></td><td align="center">12</td><td align="center">2.2</td></tr><tr><td align="left"> <bold>Proguanil + Atovaquone</bold></td><td align="center">3</td><td align="center">0.5</td></tr><tr><td align="left"> <bold>Other</bold></td><td align="center">6</td><td align="center">1.1</td></tr><tr><td align="left"><bold>Compliance (n = 190)</bold></td><td></td><td></td></tr><tr><td align="left"> <bold>Yes</bold></td><td align="center">118</td><td align="center">62.1</td></tr><tr><td align="left"> <bold>No</bold></td><td align="center">72</td><td align="center">37.9</td></tr><tr><td align="left"><bold>Patient classification (n = 615)</bold></td><td></td><td></td></tr><tr><td align="left"> <bold>Immigrants/Refugees</bold></td><td align="center">121</td><td align="center">19.7</td></tr><tr><td align="left"> <bold>Foreign visitors</bold></td><td align="center">36</td><td align="center">5.8</td></tr><tr><td align="left"> <bold>Europeans living in EC</bold></td><td align="center">406</td><td align="center">66.0</td></tr><tr><td align="left"> <bold>European expatriates</bold></td><td align="center">52</td><td align="center">8.5</td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>IQ-Range</bold></td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td align="left"><bold>Age (n = 606)</bold></td><td align="center">32</td><td align="center">26–43</td></tr></tbody></table><table-wrap-foot><p>IQ-Range = Inter-quartile range</p></table-wrap-foot></table-wrap><p>The majority of patients who imported vivax malaria into Europe were Europeans living and working in Europe (66.0%). Immigrants and refugees, summarising both those of overseas origin who may have lived in the reporting country for many years and very recent immigrants, made up the second largest group (19.7%), followed by European expatriates (8.5%) and foreign visitors (5.8%). Analysing patient classifications by reporting country revealed that immigrants and refugees accounted for distinctly more than the overall 20% proportion in Norway (61.5%), Italy (45.9%), France (40.0%), Spain (31.3%) and Denmark (25.0%).</p><p>Reasons for travel differed for Europeans and immigrants. While Europeans living in Europe mainly travelled for tourism (71.4%), followed by business (7.8%), missionary work (7.0%), research (6.3%), visits to relatives or friends (6.0%), military missions (0.8%) or other reasons (0.8%). The main reasons for travel in the immigrant group were immigration to Europe (47.0%) or visits to relatives or friends in the former home country (44.4%).</p><p>Figure <xref ref-type="fig" rid="F1">1</xref> marks 16 geographical regions from which <italic>P. vivax </italic>malaria was imported from during the 56 month surveillance period. The main regions of infection were the Indian subcontinent (17.0%), Indonesia (12.1%), South America (11.4%) and Western Africa (11.4%), as a group accounting for 52% of the cases. However, while the Indian subcontinent was the main region of infection each year, the others switched places in the annual ranking order. Further regions of importance were Eastern Africa (10.0%), Southeast Asia (8.6%), and Oceania (8.5%), contributing another 27% of the cases. Main countries of infection were Indonesia (12.1%), India (8.7%), Papua New Guinea (8.0%), Pakistan (7.8%) and Ecuador (5.7%).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Geographical origin <italic>of P. vivax </italic>malaria imported to Europe between January 1999 and September 2003 (n = 618)</p></caption><graphic xlink:href="1475-2875-3-5-1"/></fig><p>Exclusion of patients with concomitant plasmodial or other infections and cases with suspected or unknown diagnostic status left 526 patients for the analysis of clinical endpoints. Symptom information was given for 487 of those. The most frequent complaints were fever, headache, fatigue and musculo-skeletal symptoms, affecting 95.5%, 51.3%, 32.6% and 29.6% of the patients, respectively. However, a variety of other symptoms was noted, too. Further information on the course of the disease is summarised in table <xref ref-type="table" rid="T3">3</xref>. The median time from end of journey to symptom onset was 60 days (inter-quartile range 8–149). However, with 86 versus 31 days the median onset of symptoms was significantly delayed in patients with chemoprophylaxis (p<0.0001 Wilcoxon rank test). More than half of the patients (60.1%) were hospitalised, although in-patient treatment was distinctly less common in Ireland (0%), Switzerland (9.1%), Belgium (15.6%) and Spain (26.1%). The median length of hospitalization was four days (inter-quartile range 2–5). Information whether complications occurred during the course of the disease, was given for 270 of the 526 patients. Complications were reported in 30 of them, 22 mentioning recrudescence or relapse, one G6PD-deficiency and seven indicating severe disease. Specific complications in the latter group were serious spleno- or hepatomegaly (3 patients), spleen-rupture (1 patient), pancytopenia (1 patient), macrohaematuria (1 patient) and psychosis (1 patient). All 618 patients survived.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Course of disease in 526 patients with confirmed or probable P. vivax mono-infection</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center"><bold>Median</bold></td><td align="center"><bold>IQ-Range</bold></td></tr></thead><tbody><tr><td align="left"><bold>Days from end of journey to onset (nmiss = 80)</bold></td><td align="center">60</td><td align="center">8–149</td></tr><tr><td align="left"> <bold>with chemoprophylaxis</bold></td><td align="center">86</td><td align="center">41–158</td></tr><tr><td align="left"> <bold>without chemoprophylaxis</bold></td><td align="center">31</td><td align="center">4–133</td></tr><tr><td align="left"><bold>Days in hospital (nmiss = 400)</bold></td><td align="center">4</td><td align="center">2–5</td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td></td><td align="center"><bold>N</bold></td><td align="center"><bold>% of non missing</bold></td></tr><tr><td colspan="3"><hr></hr></td></tr><tr><td align="left"><bold>In-patients (nmiss = 7)</bold></td><td align="center">312</td><td align="center">60.1</td></tr><tr><td align="left"><bold>Complications (nmiss = 256)</bold></td><td align="center">30</td><td align="center">11.1</td></tr><tr><td align="left"> <bold>indicating treatment failure</bold></td><td align="center">23</td><td align="center">8.5</td></tr><tr><td align="left"> <bold>indicating severe disease</bold></td><td align="center">7</td><td align="center">2.6</td></tr><tr><td align="left"><bold>Fatalities (nmiss = 0)</bold></td><td align="center">0</td><td align="center">0.0</td></tr></tbody></table><table-wrap-foot><p>Nmiss= Number of cases disregarded due to missing data; IQ-Range = Inter-quartile range</p></table-wrap-foot></table-wrap><p>Treatment information was given for 518 confirmed and probable mono-infections. Table <xref ref-type="table" rid="T4">4</xref> presents frequencies of drugs used in the treatment of <italic>P. vivax </italic>malaria. Although primaquine and chloroquine was the most frequently used drug combination, 84 (16.2%) patients, including 61 males older than four, were not treated with primaquine. Least frequent use of primaquine was reported from France (0.0%), Ireland (20%), Poland (40.0%) and Finland (50.0%).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Drugs used in the treatment of 518 patients with P. vivax malaria</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left"><bold>Drugs</bold></td><td align="center"><bold>No. of patients</bold></td><td align="center"><bold>%</bold></td></tr><tr><td align="left">Primaquine</td><td align="center">434</td><td align="center">83.8</td></tr><tr><td align="left">Chloroquine</td><td align="center">426</td><td align="center">82.2</td></tr><tr><td align="left">Quinine</td><td align="center">48</td><td align="center">9.3</td></tr><tr><td align="left">Mefloquine</td><td align="center">36</td><td align="center">6.9</td></tr><tr><td align="left">Atovaquone/Proguanil</td><td align="center">12</td><td align="center">2.3</td></tr><tr><td align="left">Artemether/Lumefantrine</td><td align="center">4</td><td align="center">0.8</td></tr><tr><td align="left">Proguanil</td><td align="center">3</td><td align="center">0.6</td></tr><tr><td align="left">Artemisinin-derivates</td><td align="center">2</td><td align="center">0.4</td></tr><tr><td align="left">Pyriniethamine/Sulfadoxine</td><td align="center">1</td><td align="center">0.2</td></tr></tbody></table><table-wrap-foot><p><bold>Note</bold>. Multiple entries per patient were possible</p></table-wrap-foot></table-wrap></sec><sec><title>Discussion</title><p>Since network membership is self-selected, TropNetEurop surveillance data may not be representative for the whole of Europe. However, since most major referral centres of the continent are represented in the network and the number of patients treated within the network is large (~62,000 patients/year), approximate representativeness can be assumed. One objective of the present work was to look at differences between TropNetEurop-member countries. However, since some countries contributed only small numbers of cases to our database (see table <xref ref-type="table" rid="T1">1</xref>) differences found in between country comparisons should rather be understood as hints than taken for proof.</p><p>According to TropNetEurop data, <italic>P. vivax </italic>is the second most frequent cause of malaria importation, accounting for 12.9% of the case imports into Europe since 1999. Analysis of data from 612 patients treated for <italic>P. vivax </italic>malaria in Europe between January 1999 and September 2003 yielded that 118 patients had contracted the disease despite compliant chemoprophylaxis. However, since standard chemoprophylactics do not act against hypnozoites [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B23">23</xref>], this does not indicate drug failure. Primaquine, which acts against hypnozoites [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B24">24</xref>] and was shown to be effective in the prevention of falciparum and vivax malaria [<xref ref-type="bibr" rid="B25">25</xref>-<xref ref-type="bibr" rid="B27">27</xref>] has been proposed in North America as an optional agent for regular or terminal prophylaxis in certain travellers [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B28">28</xref>-<xref ref-type="bibr" rid="B30">30</xref>]. The fact that primaquine prophylaxis was not reported in any of the TropNetEurop cases reflects the fact that it is either very effective, or uncommonly used in Europe. Most likely the latter is true, since primaquine is presently not licensed for prophylactic use. About half of the patients treated for vivax malaria in Europe were Europeans, who contracted the disease on holiday. Immigrants and refugees made up the second largest patient group, with an overall proportion of about 20%. In Norway, Italy, France, Spain and Denmark they accounted for even more of the cases. The fact that more than 40% of the immigrants had contracted the disease on a visit to their former home country reveals that even those who may have lived in Europe for many still contribute considerably to malaria importation.</p><p>It seems peculiar that Western Africa was found to be a major region of infection, accounting for more cases than Eastern Africa. <italic>P. vivax </italic>prevalence was reported to be very limited in Western Africa, due to the presence of Duffy-negative blood-group variants [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. Because our surveillance data lack a true denominator it cannot be excluded, that the larger number of patients returning from Western Africa with vivax malaria results from higher travel activity to that region. However, since international tourist arrivals counted by the world tourist organisation in 1999 and 2000 <ext-link ext-link-type="uri" xlink:href="http://www.world-tourism.org"/> were twice as high in Eastern Africa, this appears to be an unlikely explanation. Another hypothetical explanation would be that a major part of the vivax cases reported from Western Africa might be misclassified infections with <italic>Plasmodium ovale</italic>, which is more common in that region and difficult to distinguish from <italic>P. vivax </italic>microscopically.</p><p>Analysis of the course of the disease revealed that half of the patients fell ill later than 60 days after arrival from an endemic area. The most common complaints were fever, headache, fatigue, and musculo-skeletal symptoms. The finding that symptom onset was significantly delayed in patients with chemoprophylaxis is consistent with recently published findings [<xref ref-type="bibr" rid="B10">10</xref>]. This can be explained by the activity of standard prophylactics against blood stages but not against hypnozoites. No fatalities and only few severe clinical complications were reported, which emphasises <italic>P. vivax's </italic>limited virulence as compared to <italic>P. falciparum </italic>[<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. More than half of the patients received in-patient treatment, indicating differences in national treatment policies rather than severe disease, since hospitalization rates varied greatly among TropNetEurop member countries.</p><p>None of the blood schizonticides used in the treatment of malaria affects hypnozoites of <italic>P. vivax</italic>, thus radical cure without relapses can only be achieved by additional administration of primaquine [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B31">31</xref>]. However, since <italic>P. vivax </italic>strains differ in their innate sensitivity to primaquine, anti-relapse treatment may fail, especially when underdosed [<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B32">32</xref>-<xref ref-type="bibr" rid="B34">34</xref>]. On the other hand, omission of antirelapse treatment does not necessarily lead to relapses [<xref ref-type="bibr" rid="B35">35</xref>]. Both uncertainties might be used as an argument to restrict primaquine use to the treatment of recurrent episodes of the disease, even in patients without G6PD-deficiency [<xref ref-type="bibr" rid="B36">36</xref>]. Still, relapses may seriously threaten patients' health, whereas primaquine, which is highly effective in relapse prevention [<xref ref-type="bibr" rid="B31">31</xref>,<xref ref-type="bibr" rid="B32">32</xref>], is well tolerated [<xref ref-type="bibr" rid="B25">25</xref>-<xref ref-type="bibr" rid="B28">28</xref>]. This indicates that unrestricted use of primaquine in patients without G6PD-deficiency might offer additional health benefits. However, this has not been evaluated in systematic risk benefit analyses. Our data on complications support the finding that primaquine treatment is well tolerated but not perfectly preventive. Within the network, 16.2% of the patients did not receive primary anti-relapse treatment with primaquine. Contraindications like G6PD-deficiency, young age or pregnancy could be ruled out as an explanation in most of these cases [<xref ref-type="bibr" rid="B21">21</xref>]. Primaquine relapse prevention was found to be common in most TropNetEurop member countries, with France, Ireland, Poland, and Finland showing clearly lower utilization of the drug. This indicates heterogeneity of national or site-specific treatment policies. An inquiry among members of TropNetEurop confirmed that some sites actually do limit primaquine to the treatment of recurrent episodes.</p></sec>
Thermodynamic phase plane analysis of ventricular contraction and relaxation	<sec><title>Background</title><p>Ventricular function has conventionally been characterized using indexes of systolic (contractile) or diastolic (relaxation/stiffness) function. Systolic indexes include maximum elastance or equivalently the end-systolic pressure volume relation and left ventricular ejection fraction. Diastolic indexes include the time constant of isovolumic relaxation – and the end-diastolic pressure-volume relation. Conceptualization of ventricular contraction/relaxation coupling presents a challenge when mechanical events of the cardiac cycle are depicted in conventional pressure, P, or volume, V, terms. Additional conceptual difficulty arises when ventricular/vascular coupling is considered using P, V variables.</p></sec><sec sec-type="methods"><title>Methods</title><p>We introduce the concept of thermodynamic phase-plane, TPP, defined by the PdV and VdP axes.</p></sec><sec><title>Results</title><p>TPP allows all cardiac mechanical events and their coupling to the vasculature to be geometrically depicted and simultaneously analyzed.</p></sec><sec><title>Conclusion</title><p>Conventional systolic and diastolic function indexes are easily recovered and novel indexes of contraction-relaxation coupling are discernible.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Karamanoglu</surname><given-names>Mustafa</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Kovács</surname><given-names>Sándor J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BioMedical Engineering OnLine	<sec><title>Introduction</title><p>Heart failure (HF) is a common and eventually lethal disease whose incidence is increasing in the population [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. It has only recently been appreciated that in nearly half of patients carrying the diagnosis of HF, left ventricular (LV) contractile function as indicated by LV ejection fraction (EF) is relatively preserved while ventricular diastolic relaxation/stiffness function is compromised and uncoupled [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. In other words, the contraction-relaxation coupling (CRC) is abnormal.</p><p>It is established that contraction-relaxation coupling is mediated primarily by intra and extracellular calcium fluxes [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B6">6</xref>]. In the cardiomyocyte, the dominant CRC is via the ionized Ca<sup>2+ </sup>flux cycling between the contractile apparatus and the sarcoplasmic reticulum, SR [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. It has been shown that in various cardiomyopathies this cycling is adversely affected [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. Although experimental techniques exist that characterize this uncoupling at the cardiomyocyte level, there is a lack of both a conceptual and an experimental framework by which this type of uncoupling can be characterized at the whole-heart level.</p><p>In principle, the Frank-Starling Law, which is an experimentally determined relationship, should characterize contraction-relaxation coupling. Because the Frank-Starling Law is based on experimental observation, rather than being derived from basic laws, there are no explicit mathematical relationships defining CRC using different contraction-relaxation indices. In this study we use a mathematical modeling approach motivated by the historical utility of P and V as variables that define ventricular pump function; the ability of phase-plane methods to characterize the properties of oscillators and the goal to express the macroscopic consequences of contraction-relaxation coupling by introduction of a new concept suitable for derivation of physiologic CRC indexes. The intent is to facilitate quantification of the presence and severity of contractile-relaxation dysfunction by the use of these CRC indexes. We tested our preliminary concepts using conductance catheter derived pressure-volume data from a limited number of human subjects. Results indicate that derived CRC indexes may differentiate between normal and abnormal contraction-relaxation states. Further clinical studies, particularly in selected pathologic subsets, are needed to fully validate the utility of the proposed concepts.</p></sec><sec sec-type="methods"><title>Methods</title><p>In proposing our conceptual approach, we assume that analysis of ventricular contraction/relaxation in the P-V plane not only should correlate with the CRC at the cellular level but must also obey the laws of thermodynamics.</p><p>First, we note that CRC at the cellular level is measured using features such as the rise and decay of the Ca<sup>2+ </sup>transient. It is known that the upstroke of the Ca<sup>2+ </sup>transient precedes the onset of tension development and starts decaying immediately. Interestingly the Ca<sup>2+ </sup>transient is negligible at the peak of the developed tension and is not affected by inhibition of the net SR Ca<sup>2+ </sup>uptake by 3 mM caffeine [<xref ref-type="bibr" rid="B10">10</xref>].</p><p>Second, we define a differential relationship that describes the instantaneous power, (erg/s) as the rate of change of energy Ê (ergs), defined as Ê = PV. In such a thermodynamic system where the heat losses during contraction are small and are therefore negligible, instantaneous power, dÊ/dt, can be written as:</p><p>dÊ = VdP-PdV     (1)</p><p>where we denote the time derivatives of Ê, P and V as dÊ, dP and dV, respectively. In expression (1) the sign of PdV is negative because we chose the system boundary as the LV volume at a given instant. The system loses energy by doing external work during systole when pressure rises and volume falls – hence the negative sign of PdV. This differential relationship defines how the energy contained in this system changes in accordance with the laws of thermodynamics. During each phase of the cardiac cycle the change in energy can be appreciated as illustrated in Figures <xref ref-type="fig" rid="F1">1</xref> and <xref ref-type="fig" rid="F2">2</xref> (the numbers in <italic>italics </italic>refer to phase of the P-V loop depicted in Figure <xref ref-type="fig" rid="F2">2</xref>): (<italic>1) </italic>isovolumic contraction when VdP > 0 and <italic>PdV </italic>= 0; (<italic>2 </italic>to <italic>3</italic>) early systolic ejection when <italic>VdP </italic>> 0, and <italic>PdV </italic>< 0; (<italic>5 </italic>to <italic>6</italic>) late systolic ejection when <italic>VdP </italic>< 0, and <italic>PdV </italic>< 0; (<italic>7</italic>) isovolumic relaxation when <italic>VdP </italic>< 0 and <italic>PdV </italic>= 0; (<italic>8</italic>) rapid filling when <italic>VdP </italic>< 0 initially, then becomes > 0 and <italic>PdV </italic>> 0; (<italic>9</italic>) diastasis when <italic>VdP </italic>= <italic>PdV </italic>= 0; and (<italic>10</italic>) passive filling when <italic>PdV </italic>> 0 <italic>VdP </italic>> 0. Because these macroscopic contraction and relaxation processes are not adiabatic, i.e. the system looses heat and generates external work, they are causally linked to energy consumption at the cardiomyocyte level.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>The cardiac cycle in thermodynamic space. Vertical lines depict the beginning and end of systole (end systolic elastance), respectively.</p></caption><graphic xlink:href="1475-925X-3-6-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p>The cardiac cycle depicted in the thermodynamic phase plane. Arrows indicate the direction of time from the onset of systole.</p></caption><graphic xlink:href="1475-925X-3-6-2"/></fig><p>In Equation 1, the <italic>VdP </italic>term denotes the <underline>potential power</underline> that is stored as potential energy in the contractile apparatus of the cardiomyocyte and in the extra-cellular connective tissue matrix and intracellularly in titin as strain energy [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Note that VdP is uniquely different from the conventionally used indexes to describe contractility dP/dt<sub>max </sub>or relaxation, dP/dt<sub>min</sub>. This potential power is generated by the strong interaction of the actin-myosin complex and thus should be dependent on fiber length, Ca<sup>2+ </sup>concentration and the Ca<sup>2+ </sup>sensitivity. The second term in the instantaneous power expression, <italic>PdV</italic>, is the <underline>kinetic power</underline> which accounts for shortening of the contractile apparatus and is the source of external work. It is a function of the weak interaction of the actin-myosin complex. Both the potential and kinetic power production terminates as Ca<sup>2+ </sup>is sequestered back into SR. Thus, it is possible to follow the course of Ca<sup>2+ </sup>cycling by considering the magnitude and time dependence of potential and kinetic power generated by the contractile apparatus. Thereby we obtain an index of contraction/relaxation coupling (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><p>To derive physiologic indexes of systolic-diastolic coupling, we review the components of the cardiac cycle using <italic>PdV </italic>and <italic>VdP </italic>as coordinates in the thermodynamic phase plane.</p><p>The potential power becomes maximal <italic>(1) </italic>after the onset of contraction and before aortic valve opening <italic>(2) </italic>During the first part of ejection the potential power decreases and kinetic power increases until kinetic. power becomes maximal. <italic>(3) </italic>The exchange between potential and kinetic power during phase <italic>(2) </italic>and <italic>(3) </italic>occurs at a constant rate, κ. After phase <italic>(3) </italic>both kinetic power and potential power decline at constant but different rates, ρ. At peak systolic pressure (dP = 0) <italic>(4) </italic>potential power becomes zero but the net power output from the contractile apparatus is still positive until the phase <italic>(5) </italic>where the contractile apparatus is incapable of delivering any further power. At aortic valve closure, the kinetic power becomes zero and the potential power becomes negative <italic>(6)</italic>. The potential power decreases to a minimum at peak negative dP and then increases again <italic>(7)</italic>. At mitral valve opening, the Doppler E-wave is inscribed <italic>(8) </italic>when kinetic power is positive, followed by diastasis <italic>(9)</italic>. Diastasis is the only point, where both the potential power and the kinetic power (dP = dV = 0) simultaneously become zero during the cardiac cycle. This is followed by atrial contraction, the Doppler A-wave, when both potential and kinetic power is positive and the cycle repeats itself <italic>(10)</italic>.</p><p>Figure <xref ref-type="fig" rid="F2">2</xref> provides a geometric perspective of the method of CRC index derivation. Specifically the angles κ, ρ and δ define the contraction, relaxation and contraction/relaxation coupling behavior of a given ventricle, respectively. These angles in the TPP provide information about the intensity and duration of peaking and decaying of the Ca<sup>2+ </sup>transient, respectively. Therefore the mapping of Ca<sup>2+ </sup>transient to the TPP is important. Accordingly, nearly the entire Ca<sup>2+ </sup>transient maps to points <italic>(1) </italic>to <italic>(7) </italic>in Figure <xref ref-type="fig" rid="F2">2</xref>.</p><p>One of the unique aspects of thermodynamic phase plane analysis is that it depicts contraction and relaxation events simultaneously and includes the way they are coupled. In this phase plane, load dependence is easy to specify. We propose that the parameters κ, ρ and δ define the contraction, relaxation and coupling properties of a given ventricle. From this definition it follows that changes in (a) preload will alter the peak potential power, peak kinetic power but will not affect the angles κ, ρ and δ; (b) afterload will alter the potential power at the peak kinetic power without changes in κ, ρ and δ; (c) LV contractile properties will alter only κ; (d) LV relaxation properties will alter only ρ; and (e) contraction/relaxation coupling will alter the angle δ. Note also that κ, ρ and δ are inherently independent of the LV mass because these indexes describe the ratio of the instantaneous potential to kinetic power, which may be dependent on the LV mass.</p><p>Depiction of the cardiac cycle, including contraction/relaxation cycling using these expressions of power in the TPP, makes it possible to better understand the physical and thermodynamic meaning of some contractility indexes used widely in the past.</p><sec><title>External Power Normalized to V<sup>2</sup></title><p>Consider the maximal external power normalized to the volume squared, (<italic>PdV</italic>)<sub>max</sub>/ V<sup>2 </sup>[<xref ref-type="bibr" rid="B13">13</xref>]. Note that during phase <italic>(2) </italic>and <italic>(3) </italic>the kinetic power and potential power are linearly related, hence one can write <italic>PdV </italic>= κ <italic>VdP </italic>+ b<sub>κ </sub>(Figure <xref ref-type="fig" rid="F2">2</xref>). Because (<italic>dP</italic>/<italic>V</italic>)<sub>max </sub>has been shown to be an index of ventricular contractility [<xref ref-type="bibr" rid="B14">14</xref>], it follows that</p><p><bold>(PdV)<sub>max</sub>/ V<sup>2 </sup>≈ κ (dP/V)<sub>max </sub>+ b<sub>κ </sub>/ V<sup>2 </sup></bold>    (2)</p><p>This novel finding indicates that previously obtained empirical findings have a solid, and deeper thermodynamic foundation.</p></sec><sec><title>Maximal elastance</title><p>End-systolic elastance had been shown to be the maximal elastance that the ventricle can attain [<xref ref-type="bibr" rid="B15">15</xref>]. But, a deeper thermodynamic explanation for this behavior has been lacking. Noting that instantaneous elastance, E(t), is defined as <italic>P(t)/(V(t)-V</italic><sub><italic>o</italic></sub>), the maximum of this function, E<sub>max</sub>, occurs when <italic>dE(t) </italic>= (<italic>V-V</italic><sub><italic>o</italic></sub>)<italic>dP-PdV </italic>= 0, or without loss of generality i.e. by a change of variable (<italic>V-V</italic><sub><italic>o</italic></sub>) → <italic>V</italic>, or by assuming the ventricular slack volume, <italic>V</italic><sub><italic>o</italic></sub>, is small and is therefore negligible:</p><p><bold><italic>VdP </italic>= P<italic>dV </italic></bold>    (3)</p><p>This indicates that the E<sub>max </sub>point is defined not only mechanistically as the maximal stiffness but also thermodynamically as the point when kinetic and potential powers are equal such that the left ventricular total thermodynamic power output is zero. This novel consequence of our approach has not been previously appreciated.</p></sec><sec><title>Time constant of isovolumic relaxation, τ</title><p>It is interesting to note that during isovolumic relaxation [<xref ref-type="bibr" rid="B16">16</xref>] the index characterizing the relationship between dP and P, is referred to as the <italic>'time-constant of isovolumic relaxation'</italic>, [<xref ref-type="bibr" rid="B17">17</xref>]. Noting in Figure <xref ref-type="fig" rid="F2">2</xref> that after the maximal kinetic power <italic>(3) </italic>and before aortic valve closure <italic>(6) </italic>one can write</p><p><italic>PdV </italic>= ρ <italic>VdP </italic>+ b<sub>ρ </sub>and <italic>dV/ V </italic>constant. Because dP/P ≈ -1/τ  then</p><p>τ ≈ - ρ     (4)</p><p>This linear relationship between ρ and τ suggests that τ is a rate constant depicting the proportional decrease of both potential and kinetic power.</p></sec><sec><title>Subjects and study protocol</title><sec><title>Subjects</title><p>After obtaining informed consent, in accordance with Washington University Medical Center Human Studies committee criteria, we recruited 18 subjects (aged 42–73 years) from referrals scheduled for elective cardiac catheterization for clinical reasons at the request of the referring cardiologist. Subjects were enrolled if they met the following criteria: (i) scheduled for elective left or right/left heart catheterization, in a fasting, non-sedated state, (ii) judged to be clinically stable (iii) willing to participate and able to give informed consent, (iv) have no mitral valve disease, atrial fibrillation, or permanent pacemakers. Clinical status at enrollment was recorded. Subjects were divided into three groups: (i) <underline>normal group (n = 6)</underline> (τ < 50 ms; LVEF > 60% and LVEDP < 12 mmHg); (ii) <underline>impaired relaxation group (n = 6)</underline> with normal or mildly reduced left ventricular systolic function (EF > 50%) and evidence of abnormal left ventricular relaxation (τ > 50 ms) and diastolic distensibility (LVEDP > 16 mmHg) [<xref ref-type="bibr" rid="B18">18</xref>] and (iii) <underline>heart failure group (n = 6)</underline> (τ > 50 ms; LVEF < 45% and LVEDP > 16 mmHg).</p><p>Pressure volume loops were recorded using a 6F, 10-electrode pressure-volume conductance pigtail catheter (Millar Instruments, Dallas, TX). Briefly, using 62 cm 6F femoral arterial sheath (Arrow, San Leandro CA) the conductance catheter was advanced through the sheath into the central aorta, and then advanced retrogradely across the aortic valve with the pigtail tip positioned near the left ventricular apex. The conductance catheter was used with a simulator-microprocessor unit (Leycom Sigma-5, CardioDynamics, Rijnsburg, The Netherlands) in dual-field stimulation mode. Pressure-volume loops were displayed on a personal computer to allow on-line signal acquisition and real-time display. The pressure and volume data was digitized on-line at a rate of 200 Hz and was stored on a computer hard disk for subsequent off-line analysis. The baseline conductance volume data was calibrated using ventriculogram-derived end-diastolic and end-systolic volumes [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>From P-V data obtained according to the above criteria, we determined if the three groups were distinguishable using thermodynamic phase plane analysis derived indexes. Hence, we calculated τ, chamber contractility using end-systolic elastance, E<sub>es</sub>, and integrated ventricular function using the slope of preload-recruitable stroke work relationship, M<sub>sw</sub>. All analyses were performed off-line in the Cardiovascular Biophysics Laboratory. Pressure-volume loops were analyzed for derivation of contractile-relaxation indexes using a custom written software in Visual Basic. The angles κ and ρ were estimated using the best linear relationship between VdP<sub>max </sub>and PdV<sub>max </sub>and PdV<sub>max </sub>and VdP<sub>min</sub>, respectively. The angle δ was calculated geometrically by considering δ = κ + ρ. Data were analyzed using ANOVA followed by Student-Newman-Keuls multiple comparison post significance test using a commercially available software (Instat, Graphpad). Data are expressed as mean ± SD. and p < 0.05 level was considered statistically significant.</p></sec></sec></sec><sec><title>Results</title><p>As expected from the inclusion criteria, other differences in these study groups were present (Table <xref ref-type="table" rid="T1">1</xref>). Despite contractility indexes similar to the normal group, the impaired relaxation group had lower heart rates, higher blood pressure and decreased dP/dt<sub>min</sub>. The heart failure group had lower contractility and relaxation indexes.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Patient Characteristics. <sup>a </sup>= p < 0.05, <sup>b </sup>= p < 0.01 Impaired Relaxation vs Normal, <sup>d </sup>= p < 0.01 Heart Failure vs Normal, <sup>e </sup>= p < 0.05, <sup>f </sup>= p < 0.01 Heart Failure vs Impaired Relaxation.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center"><bold>Normal</bold></td><td align="center"><bold>Impaired Relaxation</bold></td><td align="center"><bold>Heart Failure</bold></td></tr></thead><tbody><tr><td align="left"><bold>Age</bold></td><td align="center">years</td><td align="right">53.5 ± 7.6</td><td align="right">61.3 ± 7.8</td><td align="right">59 ± 6.6</td></tr><tr><td align="left"><bold>Sex</bold></td><td></td><td align="right">4M 2F</td><td align="right">3M 3F</td><td align="right">4M 2F</td></tr><tr><td align="left"><bold>EF</bold></td><td align="center">percent</td><td align="right">68.8 ± 8. 6</td><td align="right">69.5 ± 7.8</td><td align="right">37.5 ± 6.4<sup>d,f</sup></td></tr><tr><td align="left"><bold>LVEDP</bold></td><td align="center">mmHg</td><td align="right">7.0 ± 3.0</td><td align="right">17.5 ± 1.4<sup>b</sup></td><td align="right">22.4 ± 8.7<sup>d</sup></td></tr><tr><td align="left"><bold>τ</bold></td><td align="center">ms</td><td align="right">41.9 ± 5.2</td><td align="right">73.4 ± 14.3<sup>b</sup></td><td align="right">86.3 ± 25.6<sup>d</sup></td></tr><tr><td align="left"><bold>ESV</bold></td><td align="center">ml</td><td align="right">41.0 ± 15.7</td><td align="right">51.5 ± 15.7</td><td align="right">139.5 ± 30.7<sup>d,f</sup></td></tr><tr><td align="left"><bold>EDV</bold></td><td align="center">ml</td><td align="right">129.9 ± 34.3</td><td align="right">173.0 ± 48.9</td><td align="right">223.7 ± 46.8<sup>d</sup></td></tr><tr><td align="left"><bold>E<sub>es</sub></bold><break/></td><td align="center">mmHg/ml</td><td align="right">3.5 ± 1.8</td><td align="right">2.93 ± 0.9</td><td align="right">0.96 ± 0.3<sup>d,e</sup></td></tr><tr><td align="left"><bold>M<sub>sw</sub></bold><break/></td><td align="center">mmHg</td><td align="right">0.87 ± 0.19</td><td align="right">0.89 ± 0.20</td><td align="right">0.39 ± 0.18<sup>d,f</sup></td></tr><tr><td align="left"><bold>dP/dt<sub>max</sub></bold><break/></td><td align="center">mmHg/s</td><td align="right">1207 ± 124</td><td align="right">1102 ± 129</td><td align="right">1070 ± 185</td></tr><tr><td align="left"><bold>dP/dt<sub>min</sub></bold><break/></td><td align="center">mmHg/s</td><td align="right">-1536 ± 193</td><td align="right">-1256 ± 78<sup>a</sup></td><td align="right">-1095 ± 204<sup>d</sup></td></tr><tr><td align="left"><bold>Hr</bold></td><td align="center">beats/min</td><td align="right">73.9 ± 7.1</td><td align="right">61.7 ± 8.6</td><td align="right">71.3 ± 12.5</td></tr><tr><td align="left"><bold>SP</bold></td><td align="center">mmHg</td><td align="right">129.5 ± 18.6</td><td align="right">151.2 ± 28.0</td><td align="right">135.3 ± 20.3</td></tr><tr><td align="left"><bold>DP</bold></td><td align="center">mmHg</td><td align="right">67.7 ± 6.7</td><td align="right">79.4 ± 10.5<sup>a</sup></td><td align="right">80.7 ± 8.8</td></tr></tbody></table></table-wrap><p>Figure <xref ref-type="fig" rid="F3">3</xref> depicts the overall relationship between the diastolic power conversion ratio ρ and the isovolumic time constant, τ. As predicted by Equation 4, there was good correlation between ρ and τ. Note that τ was calculated during the isovolumic relaxation period while ρ was calculated during the ejection period. This finding, which confirms prediction (d), also provides strong evidence that ventricular relaxation begins not at the end of ejection but immediately after peak ejection [<xref ref-type="bibr" rid="B20">20</xref>].</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>The relationship between diastolic power conversion ratio, ρ, and isovolumic time constant τ, r<sup>2 </sup>= 0.27, p < 0.05</p></caption><graphic xlink:href="1475-925X-3-6-3"/></fig><p>As predicted by Equation 2, there was a close relationship between the systolic power conversion ratio κ and contractility indexes, M<sub>sw </sub>and E<sub>es </sub>(Figure <xref ref-type="fig" rid="F4">4</xref>).</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>The relationship between systolic power conversion ratio, κ, and contractility indexes: (Left) M<sub>sw</sub>, r<sup>2 </sup>= 0.51, p < 0.001, (Right) E<sub>es</sub>, r<sup>2 </sup>= 0.28, p < 0.01</p></caption><graphic xlink:href="1475-925X-3-6-4"/></fig><p>Figure <xref ref-type="fig" rid="F5">5</xref> depicts pressure-volume loops and thermodynamic phase plane analysis for a representative subject from each group. The normal and the impaired relaxation group had similar values for κ, ρ, δ and potential power at end ejection (Table <xref ref-type="table" rid="T2">2</xref>). In contrast, the heart failure group had a higher κ and lower potential power at end ejection but similar ρ and δ. This finding confirms predictions (a), (b) and (c) and suggests that in this select group of heart failure patients, not only the power generation capacity of the ventricle is diminished but also the potential to kinetic power conversion rate, κ.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>Typical pressure-volume loops (left), thermodynamic phase-plane plots (middle) and (right) best linear fits for the TPP plots from normal, impaired relaxation and heart failure subjects.</p></caption><graphic xlink:href="1475-925X-3-6-5"/></fig><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>TPP parameters calculated. <sup>a </sup>= p < 0.05, <sup>b </sup>= p < 0.01 Heart Failure vs Normal, <sup>c </sup>= p < 0.01 Heart Failure vs Impaired Relaxation</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td align="center"><bold>Normal</bold></td><td align="center"><bold>Impaired Relaxation</bold></td><td align="center"><bold>Heart Failure</bold></td></tr></thead><tbody><tr><td align="left">κ</td><td align="right">degrees</td><td align="right">50.8 ± 11.7</td><td align="right">56.4 ± 12.0</td><td align="right">69.6 ± 7.8<sup>a</sup></td></tr><tr><td align="left">b<sub>κ </sub>(×10<sup>3</sup>)<break/></td><td align="right">mmHg.ml/s</td><td align="right">108.4 ± 51.3</td><td align="right">148.0 ± 90.9</td><td align="right">188.0 ± 35.9</td></tr><tr><td align="left">ρ</td><td align="right">degrees</td><td align="right">-25.9 ± 8.5</td><td align="right">-22.5 ± 9.9</td><td align="right">-14.6 ± 10.4</td></tr><tr><td align="left">b<sub>ρ </sub>(×10<sup>3</sup>)<break/></td><td align="right">mmHg.ml/s</td><td align="right">-13.5 ± 3.8</td><td align="right">-10.4 ± 4.2</td><td align="right">12.1 ± 6.2<sup>b,c</sup></td></tr><tr><td align="left">δ</td><td align="right">degrees</td><td align="right">76.6 ± 15.7</td><td align="right">78.8 ± 13.5</td><td align="right">89.8 ± 9.1</td></tr></tbody></table></table-wrap></sec><sec><title>Discussion</title><p>It is well known that the contractile and relaxation performance of the left ventricle is coupled in healthy subjects through the Frank-Starling mechanism. Accordingly, altered contractile performance is met by altered relaxation properties of the ventricle causing rapid adaptation (within one beat) to changes in demand. It is accepted that this behavior is not a result of how the ventricle itself is geometrically constructed but an attribute intrinsic to the Ca<sup>2+ </sup>cycling and contractile apparatus within the cardiomyocytes [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B21">21</xref>] themselves. Through increased sensitivity of actin-myosin interaction to calcium and through the SR Ca<sup>2+ </sup>fluxes, the contractile force and period of relaxation is modulated. This form of contraction-relaxation coupling at the cellular level is essential for normal function at the whole heart level. There is evidence that in dilated cardiomyopathy and ischemic cardiomyopathy contraction-relaxation coupling is altered at the cellular level [<xref ref-type="bibr" rid="B9">9</xref>]. In the former case, there are adverse changes in the Ca<sup>2+ </sup>flux from the SR leading to slower shortening velocity. In the latter case the speed of Ca<sup>2+ </sup>re-uptake into SR is reduced causing delayed relaxation. Changes in Ca<sup>2+ </sup>sensitivity of the actin-myosin interaction, on the other hand, alters the speed and extent of both the contraction and relaxation events, thus preserving the contraction-relaxation coupling [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>Whether all aspects of normal contraction-relaxation coupling are preserved at the organ level in these two and other forms of HF such as cardiomyopathy is unknown. Existing experimental methods do not allow observation of in vivo Ca<sup>2+ </sup>fluxes in humans, thus only approximate inferences using gross macroscopic mechanical behavior can be made. In the systolic failure model for example, diastolic performance is somewhat preserved through utilization of full Frank-Starling mechanism [<xref ref-type="bibr" rid="B22">22</xref>]. The converse may not be true. For example in long standing hypertension where contractile function is preserved (EF > 50%) through hypertrophic remodeling, an increased incidence of diastolic failure has been observed particularly in elderly women [<xref ref-type="bibr" rid="B23">23</xref>]. This dichotomy suggests that contraction-relaxation abnormalities associated with remodeling may be a dominant feature in the latter case. Unfortunately, at the organ level the mechanisms that may account for this uncoupling of systolic and diastolic function are not well understood. The derivation and in-vivo validation of indexes that are specific and sensitive to contraction-relaxation coupling at the organ level will help elucidate and characterize the state of coupling.</p><sec><title>Established indexes of contraction(systole) and of relaxation(diastole) in the left ventricle</title><p>Apart from the gross anatomical features such as the large-scale fiber direction and orientation [<xref ref-type="bibr" rid="B24">24</xref>], the left ventricle can be thought of as an assembly of cardiomyocytes that are arranged and work in parallel and in series. It is therefore reasonable to assume that the simplest model for left ventricle entails the main features of the cardiomyocyte itself. Similar to models (Maxwell, Voigt) of the single cardiomyocyte [<xref ref-type="bibr" rid="B25">25</xref>], various simplified approximations and parameters have also been developed to quantify cardiac mechanical events. Contraction indexes include: ejection fraction, the time-varying elastance [<xref ref-type="bibr" rid="B15">15</xref>]; maximal rate of rise of pressure [<xref ref-type="bibr" rid="B14">14</xref>], maximal external power [<xref ref-type="bibr" rid="B13">13</xref>]; and preload recruitable stroke work [<xref ref-type="bibr" rid="B14">14</xref>]. Diastolic function indexes include: the isovolumic relaxation time constant, τ, for ventricular relaxation [<xref ref-type="bibr" rid="B17">17</xref>], end-diastolic pressure-volume relationship for passive left ventricular properties [<xref ref-type="bibr" rid="B26">26</xref>], conventional E-and A-wave derived indexes [<xref ref-type="bibr" rid="B27">27</xref>] and the parameterized diastolic filling, (PDF), formalism derived indexes of ventricular filling [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]</p><p>Despite the success of these conceptual and experimental descriptions of distinct phases of the cardiac cycle, a deficiency exists in that a comprehensive description applicable to the entire cardiac cycle is absent. Specifically, these paradigms and indexes were not specifically intended for deriving indexes of contraction-relaxation coupling in the intact left ventricle.</p></sec><sec><title>Existing indexes of contraction and relaxation coupling</title><p>The two gold standard indexes that independently describe left ventricular contraction and relaxation properties respectively, are the end-systolic elastance, E<sub>es</sub>, and the isovolumic relaxation time constant, τ. Although scant animal data indicates that E<sub>es </sub>and τ may be inversely related [<xref ref-type="bibr" rid="B30">30</xref>], the data do not explain why it is possible to have preserved contractile (systolic) function with concomitant relaxation (diastolic) dysfunction. It is also known that E<sub>es </sub>is relatively load-independent but τ is sensitive to the timing and the amplitude of the imposed load. Early systolic load decreases τ but late systolic load increases τ [<xref ref-type="bibr" rid="B20">20</xref>]. Increased afterload initially decreases τ but if the systolic pressure reaches 80% of isovolumic maximum pressure, τ is increased [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B30">30</xref>]. Whether the reported load sensitivity of τ is caused by methodological details [<xref ref-type="bibr" rid="B32">32</xref>,<xref ref-type="bibr" rid="B33">33</xref>] remains an issue. Other relationships should exist. Starting from a time-varying compliance concept Palladino et al also derived energy-exchange equations between the biochemical processes, mechanical work, potential energy and heat during contraction with considerable success [<xref ref-type="bibr" rid="B34">34</xref>]. We have previously shown that SW and dP-P phase plane areas and dP<sub>min </sub>and dP<sub>max </sub>are correlated [<xref ref-type="bibr" rid="B16">16</xref>]. This former finding constitutes preliminary data to support the concept that a coupling between the dP-P limit cycle area generated by each cardiac cycle and EDV exists because SW is related to EDV through PRSWR. This state of affairs underscores the need for better indexes by which contraction-relaxation coupling can be quantified.</p></sec></sec><sec><title>Limitations</title><p>Movement and positioning uncertainties of the conductance catheter introduces noise into the data. We have found that this could be addressed in part by recording longer periods, (30 seconds), of continuous data while the hemodynamics are in steady state. Beat averaging of data eliminates most of the movement and respiratory artifacts. Although the derivatives of conductance signals amplify noise, this does not adversely affect our TPP analysis because: (a) we average data for 10–16 beats, (b) conductance signals are usually free from artifacts between phases (<italic>2</italic>) and (<italic>6</italic>) (Figure <xref ref-type="fig" rid="F5">5</xref>), and (c) because the way the VdP and PdV are depicted in TPP analysis, the impact of noise due to derivatives is reduced.</p><p>Based on the inclusion criteria, we considered a high EF, increased LVEDP and increased τ would suffice to classify the impaired relaxation group. Because passive diastolic properties of the LV chamber also contributes to the LVEDP, the discrimination between the normal and impaired relaxation group may not be powerful. Other factors which directly affect P and V, such as underlying pathology, heart rate, medications, preload, contractility and afterload could modulate the CRC indexes. We have, however, found a significant correlation between τ and ρ suggesting active relaxation processes are indeed reflected in ρ. Our analysis suggests that in this select group of subjects the magnitude of relaxation impairment was not sufficient to be detectable, although a trend suggesting that ρ was lower was present. Further studies, involving larger number of subjects with documented diastolic heart failure will be carried out to determine if this trend is significant.</p></sec><sec><title>Conclusion</title><p>The proposed thermodynamic phase plane analysis method presented forms the basis of an evolving conceptual approach for quantitative contraction-relaxation assessment. Additional work in well-defined clinical pathophysiologic subsets as well as experimental animal models will aid in defining its full potential. Utilization in the aforementioned settings will facilitate complete determination of the relationship between current ventricular systolic and diastolic function assessment methods and the thermodynamic phase plane analysis method. Although our work is preliminary, and in a modest number of subjects, we found that the thermodynamic phase plane analysis method can not only quantitatively characterize the standard clinical systolic and diastolic ventricular function indexes but is able to elucidate their deeper thermodynamic basis. Our results justify further studies in order to extend these findings to subjects and animal models with and without impaired contraction/relaxation states due to various clinical (diabetes, hypertension, obesity etc) and specific molecular-cellular (titionopathies, myosinopathies, etc) etiologies.</p></sec><sec><title>Authors' contributions</title><p>MK and SK jointly conceived the 'thermodynamic phase plane" concept in terms of PdV and VdP coordinates, MK performed the data analysis and participated in study design and coordination. SK participated in study design and coordination and carried out the cardiac catheterizations. All authors read and approved the final manuscript.</p></sec>
Evaluating the disparity of female breast cancer mortality among racial groups - a spatiotemporal analysis	<sec><title>Background</title><p>The literature suggests that the distribution of female breast cancer mortality demonstrates spatial concentration. There remains a lack of studies on how the mortality burden may impact racial groups across space and over time. The present study evaluated the geographic variations in breast cancer mortality in Texas females according to three predominant racial groups (non-Hispanic White, Black, and Hispanic females) over a twelve-year period. It sought to clarify whether the spatiotemporal trend might place an uneven burden on particular racial groups, and whether the excess trend has persisted into the current decade.</p></sec><sec sec-type="methods"><title>Methods</title><p>The Spatial Scan Statistic was employed to examine the geographic excess of breast cancer mortality by race in Texas counties between 1990 and 2001. The statistic was conducted with a scan window of a maximum of 90% of the study period and a spatial cluster size of 50% of the population at risk. The next scan was conducted with a purely spatial option to verify whether the excess mortality persisted further. Spatial queries were performed to locate the regions of excess mortality affecting multiple racial groups.</p></sec><sec><title>Results</title><p>The first scan identified 4 regions with breast cancer mortality excess in both non-Hispanic White and Hispanic female populations. The most likely excess mortality with a relative risk of 1.12 (p = 0.001) occurred between 1990 and 1996 for non-Hispanic Whites, including 42 Texas counties along Gulf Coast and Central Texas. For Hispanics, West Texas with a relative risk of 1.18 was the most probable region of excess mortality (p = 0.001). Results of the second scan were identical to the first. This suggested that the excess mortality might not persist to the present decade. Spatial queries found that 3 counties in Southeast and 9 counties in Central Texas had excess mortality involving multiple racial groups.</p></sec><sec><title>Conclusion</title><p>Spatiotemporal variations in breast cancer mortality affected racial groups at varying levels. There was neither evidence of hot-spot clusters nor persistent spatiotemporal trends of excess mortality into the present decade. Non-Hispanic Whites in the Gulf Coast and Hispanics in West Texas carried the highest burden of mortality, as evidenced by spatial concentration and temporal persistence.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Ed Hsu</surname><given-names>Chiehwen</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Jacobson</surname><given-names>Holly</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Soto Mas</surname><given-names>Francisco</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib>	International Journal of Health Geographics	<sec><title>Background</title><p>According to a recently released U.S. Cancer Statistics report, breast cancer was the leading cause of cancer deaths among American women in all racial/ethnic groups in the year 2000. [<xref ref-type="bibr" rid="B1">1</xref>] National and state-specific studies indicated that the distribution of breast cancer mortality varied by race in Texas over an extended period of time. Spatiotemporal variations observed by the Cancer Atlas of the National Cancer Institute [<xref ref-type="bibr" rid="B2">2</xref>] suggested that the Houston-Galveston and Dallas-Fort Worth State Economic Areas (SEAs) had the highest breast cancer mortality among White females. Among the Black female population, Abilene and Alice SEAs had the highest breast cancer mortality between 1970 and 1994. Texas-specific cancer research also indicated that Texas counties near the Gulf Coast, Bexar and El Paso counties, among others, had an excess mortality from cancers between 1980 and 1997. [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B5">5</xref>] On the other hand, the report of U.S. Cancer Statistics suggested that disparities exist in cancer mortality among different racial groups. [<xref ref-type="bibr" rid="B1">1</xref>] The report indicated that the occurrence of breast cancer among non-Hispanic White women was almost 1.2 times higher than that of Black women, and 1.7 times higher than that among Asians/Pacific Islanders. In Texas, 26,338 females died of cancer in the last decade. Among them, 72% (18,966) were non-Hispanic White females.[<xref ref-type="bibr" rid="B6">6</xref>] Conversely, other research argued that excessive breast cancer mortality presented an uneven burden on African-Americans, as this particular racial group experienced worse breast cancer outcomes, [<xref ref-type="bibr" rid="B7">7</xref>] and that African-American and Hispanic women had poorer overall survival rates from breast cancer.[<xref ref-type="bibr" rid="B8">8</xref>] Although the literature provides inconclusive results in terms of which race/ethnicity may suffer the most from the burden of breast cancer mortality, it nevertheless underlines the importance of clarifying the spatiotemporal disparity in racial groups.</p><p>To quantify the breast cancer mortality burden by race across space and time, this study adopted a statistical approach to characterize the spatiotemporal clusters of breast cancer mortality. A "cluster", in this context, is detected within a defined geographical area during a specific timeframe when the area has a disproportionate excess in mortality, when compared to the neighbouring areas under study.[<xref ref-type="bibr" rid="B9">9</xref>] By meeting the statistical assumptions of a set of statistical models, the unusual rise or reduction of mortality in a specific spatial and temporal window (with adjustments for demographic factors such as age and gender, or other substantiated risk factors) can be characterized by statistical significance. In this context, this study used the terms "clusters" and "excess mortality" interchangeably, with both terms referring to the statistical context of both spatial and temporal dimensions of excess.</p><p>For the time period between 1990 and 2001, the present study evaluated the county-level excess of breast cancer mortality in three predominant racial groups of Texas female populations. The excess mortality burden was characterized by spatiotemporal variations. The study tested the potential continuation of excess deaths for 10 years or more to the present decade. Based on the results of analysis, the study identified each racial group and multiple groups in Texas regions that may have been most affected by the persistent mortality burden over time. These results point to priority geographic areas for policy deliberation.</p></sec><sec><title>Data collection and treatments</title><p>To identify potential breast cancer excess mortality in Texas counties between 1990 and 2001, breast cancer race-specific deaths, the female population at risk, and location data were collected and saved in three separate files. The first, "Deaths Files By Race", included female breast cancer deaths (ICD-9 Code 174 and ICD-10 Code C50), which reported the place of residence in 254 Texas counties of 4 racial groups, that were coded as categorical data (e.g., non-Hispanic White = 1, Black = 2, Hispanic = 3 and Other = 4) in each of the 12 study years. Each file contained 16 age-group categorical variables, with values ranging from 1 to 16, representing the ages of "0 to 4" to "75 and above", grouped at 5-year intervals. The data were so arranged for age adjustment and race stratification. One file for per racial group was created, including 48,768 records reflecting the number of deaths for each race among the 16 age groups in 254 counties over the 12-year study period. The second file, the "Population File", contained data on the populations at risk in the study period, (i.e., the female population in each Texas county, during the 12 study years, with respect to race and age group information that corresponded to the "Deaths File By Race"). Race and age data from the Year 2000/1990 Census were obtained from the "Summary File 1" of Census 2000/1990, originated from the American FactFinder Website of the U.S. Census Bureau. [<xref ref-type="bibr" rid="B10">10</xref>] The population data for the remaining years were obtained from population estimates made available through the Texas State Data Center and the Center of Vital Statistics of the Texas Department of Health. [<xref ref-type="bibr" rid="B11">11</xref>] The Population file contained a total of 195,072 records, representing the four races and 16 age groups in 254 counties for the 12-year study period. Age and race variables were also coded as categorical data to enable subsequent adjustment and stratification. The third file, the "Geographic File", was also obtained from the US Census Bureau. [<xref ref-type="bibr" rid="B12">12</xref>] This file contained the latitude and longitude information of Texas county centroids as a proxy that indicated the locality of each county. The Texas county shapefiles were obtained from the CDC Website (URL <ext-link ext-link-type="uri" xlink:href="http://www.cdc.gov/epiinfo/usa/tx.exe"/>) for further mapping analysis.</p></sec><sec><title>Methods of analysis</title><p>This study employed the Spatial Scan Statistic developed by Kulldorff and colleagues [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>] to detect potential excess breast cancer mortality. The test statistic was adopted previously for detecting excesses of breast and brain cancers. [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B16">16</xref>] When compared with other statistical methods for cluster detection, this statistic was found to have good power for detecting localized hot-spots type of excess events, particular those in rural areas.[<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>] This test statistic seems appropriate for detecting potential excess breast cancer mortality in the unique rural-urban combination of the state of Texas.</p><p>The Spatial Scan Statistic factors in uneven geographical population densities and conditions, and then analyzes the total number of observed breast cancer deaths. It searches for clusters of cases without specif ying their size or location ahead of time, which tests for their statistical significance while adjusting for the multiple testing inherent in such a procedure. For an analysis of rare cases/deaths, such as cancer, the Poisson model can be used for estimating the probability distribution when the number of cases/deaths is substantially smaller than that of the population at risk. When there are no covariates, the null hypothesis of the Poisson model provides that the expected death counts in each county are proportional to the population size (or person-years) in that area. The alternative hypothesis states that deaths are not randomly distributed. In the present study, for each location and size of the scanned space and time, the alternative hypothesis refers to elevated adjusted mortality rates within space and time as compared to outside areas under study. Calculations can be performed using the SatScan Program (version 4.0, freeware available from URL:<ext-link ext-link-type="uri" xlink:href="http://www.satscan.org"/>). SatScan first aggregates data with the scanning window of spatial (referring to the population at risk) – as a cylinder base, and temporal (years in this study) – corresponding to the height of the cylinder as selected by the users. For each cylinder, the scan adjusts for covariates and calculates the Log Likelihood Ratio (LLR, formula described below) by scanning through and plotting circles around geographic identifiers (prepared in the Geographic File) in a population size specified by the user across the entire study area. The base is the same as defined in spatial statistics, while the height reflects the time period of potential clusters. The cylindrical window moves in space and time and scans through each possible geographic location, defined by county centroids in the present study. The overall relative risk for each cluster, along with a set of simulated values based on the same procedure within a specific space and time, are then calculated. The latter are used as a baseline against the LLR values of the observed values. SatScan employs the Monte-Carlo simulation to estimate the LLR. When the LLR values of observed windows are higher than LLR based on simulation, SatScan determines the deaths in a particular region that are significantly different from the rest of the study area for the particular time window by rejecting the null hypothesis. Under the Poisson assumption, the Likelihood function for a specific space-time is then proportional to:</p><p><italic>LLR = (c/n)</italic><sup><italic>c</italic></sup><italic>([C-c]/[C-n])</italic><sup><italic>(C-c)</italic></sup><italic>I() </italic>[<xref ref-type="bibr" rid="B9">9</xref>]</p><p>Where C is the total number of breast cancer deaths, c is the number of cases within the space-time window, and n is the covariate(s)-adjusted expected number of deaths within the space-time analysis under the null-hypothesis. I() is an indicator function, whereby I() is equal to 1 when the timeframe has more deaths than is expected under the null-hypothesis, and is 0, otherwise. Based on a test statistic value of the LLR, a p-value is then calculated which suggests how well all the variables fit into the model at the same time. SatScan performs adjustments by indirect standardization.</p><p>For the present study, the Poisson model was used to calculate the number of expected deaths in each county. The space-time retrospective analysis was conducted without prior assumptions as to the size or location of such areas or duration of excessive mortality. The scan setting was set at a maximum spatial cluster size of 90% of the study period (i.e., 10 years) and 50% of the population at risk. The "50% of population at risk" parameter was recommended by Kulldorff [<xref ref-type="bibr" rid="B9">9</xref>] as an optimal value setting that maximizes the effect of potential cluster detection. This means that a cluster would comprise, at most, 50% of the population at risk. The study further tested the potential persistence of temporal clusters across the entire study period (i.e., 12 years) by holding constant the 50% maximum spatial cluster and scanning with the "purely spatial" option.</p><p>For data processing, we developed a Visual Basic application to automate data collection and manipulation, and output the results to geographic information systems (GIS) for performing mapping and spatial queries. The SatScan program saved the output files, including cluster locations, relative risk for each location, simulated LLRs, and the test statistic, in database (.dbf format) files. Data were stored on a Microsoft SQL Server version 7.0, and The SatScan program calculated LLR by performing 999 instances of Monte Carlo replications. The automated process, including data input, scanning, and output, took an average of 25 minutes of computer time.</p></sec><sec><title>Results</title><p>The study included 28,813 breast cancer deaths among an average female population of 9,585,195 in Texas counties across the 12-year study period. The age-and-race-adjusted annual mortality rate of all races was 25/100,000 women/year. Annual age-adjusted mortality rates for non-Hispanic Whites, Blacks, and Hispanics were 31, 29.4, and 12.3 per 100,000 females respectively. Figures <xref ref-type="fig" rid="F2">2</xref> to <xref ref-type="fig" rid="F4">4</xref> are summary choropleth maps of the age-adjusted (to 2000 US population) mortality rates by quartiles at the Texas county level. With the adjustment of age and stratification by race, a scan window of a maximum of 90% of the study period (i.e., 10 years) and 50% of the population at risk revealed 10 regions of likely excess mortality in three racial groups within Texas female population. Among these, four regions were statistically significant in terms of both spatial and temporal excess. Figures <xref ref-type="fig" rid="F1">1</xref> to <xref ref-type="fig" rid="F3">3</xref> present these likely areas of excess mortality in this set of analysis. To describe the extensive geographic regions of the state of Texas in a consistent manner, we adopt the term used in the "counties and regions cross-reference" for each public health regions of Texas, as defined by the Texas Health and Human Services Commission. (URL <ext-link ext-link-type="uri" xlink:href="http://www.hhsc.state.tx.us/about_hhsc/HHS_Regions.html"/>). For the non-Hispanic White population, three potential excess mortality regions were detected. The most likely area of excess mortality with a Relative Risk (RR) of 1.12 (p = 0.001) occurred between 1990 and 1996 in southeast Texas along Gulf Coast, Central and Upper South Texas. These included 42 counties ranging from Harris to Kleberg counties. A secondary excess mortality region (RR = 1.12, p = 0.001) was identified between 1990 and 1993 in the 42 counties of Northwest and Metroplex Texas. One potential excess area in the non-Hispanic White population was detected in 5 counties in Upper East Texas that was not statistically significant (RR = 1.36, p = 0.99). For the Black population, the most likely area of excess mortality (RR = 1.15, p = 0.12) occurred between 1991 and 1996 in southeast Texas along the Gulf Coast. These included 15 counties ranging from Gasper to Brazoria. This cluster was not statistically significant. Additionally, 5 regions identified with potential excess mortality were not statistically significant (p > 0.20). For the Hispanic population, the most likely area of excess mortality (RR = 1.18, p = 0.001) occurred between 1990 and 1998 in West Texas, ranging from the border of the state of New Mexico to Lower South Texas. This area included 127 counties. An area of secondary excess mortality was found in the Gulf Coast and Upper South Texas (RR = 3.53, p = 0.848) between 1990 and 1995. For the "Other" population, six clusters were detected, and none of these were statistically significant (p > 0.28). It was not clear whether the temporal component of the clusters were there because it was truly a cluster restricted in time, or whether it is a purely spatial cluster that showed up as a space-time cluster due to the maximum temporal cluster size restriction. To further detect whether this cluster (or others) might have persisted into the current decade, a second scan analysis was performed with a "spatial only" option to enable the detection of 12-year clusters. The results of the second scan were almost identical to those of the first set of analysis, with the p value slightly increased in the most likely cluster for Blacks (p = 0.131). In terms of spatial persistence, the number of counties originally in the secondary possible excess mortality area remained the same. The secondary cluster remained the same in terms of relative risk and p-value. The non-statistically-significant cluster remained not statistically significant. Tables <xref ref-type="table" rid="T1">1</xref> summarizes the results of analysis by potential excess mortality, duration of occurrence, observed and expected deaths, counties, and relative risk information in each excess mortality region of a statistical significance.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Potential excess of breast cancer mortality detected in the <underline>non-Hispanic White f</underline>emale population of Texas counties, 1990–2001, </bold>conducted with a scan window of a maximum of 90% of the study period and a spatial cluster size of 50% of the population at risk. Age-adjusted breast cancer mortality rates of <underline>non-Hispanic White</underline> Females in Texas Counties, 1990–1998.</p></caption><graphic xlink:href="1476-072X-3-4-1"/></fig><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Potential excess of breast cancer mortality detected in the <underline>Black</underline> female population of Texas counties, 1990–2001, </bold>conducted with a scan window of a maximum of 90% of the study period and a spatial cluster size of 50% of the population at risk. Age-adjusted breast cancer mortality rates of <underline>Black</underline> females in Texas Counties, 1990–1998.</p></caption><graphic xlink:href="1476-072X-3-4-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>Potential Excess of Breast Cancer Mortality detected in the <underline>Hispanic</underline> female population of Texas Counties, 1990–2001, </bold>conducted with a scan window of a maximum of 90% of the study period and a spatial cluster size of 50% of the population at risk. Age-adjusted breast cancer mortality rates of <underline>Hispanics</underline> females in Texas Counties, 1990–1998.</p></caption><graphic xlink:href="1476-072X-3-4-3"/></fig><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>Potential Excess of Breast Cancer Mortality detected in the <underline>Other</underline> female population of Texas Counties, 1990–2001, </bold>conducted with a scan window of a maximum of 90% of the study period and a spatial cluster size of 50% of the population at risk. Age-adjusted breast cancer mortality rates of <underline>Other</underline> females in Texas Counties, 1990–1998.</p></caption><graphic xlink:href="1476-072X-3-4-4"/></fig><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Texas counties detected with excess breast cancer mortality with statistical significance 1990–2001.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center"><bold>Potential Excess mortality</bold></td><td align="center"><bold>Annual Age-adjusted rates (per 100,000)</bold></td><td align="center"><bold>Year</bold></td><td align="center"><bold>Observed deaths</bold></td><td align="center"><bold>Expected deaths</bold></td><td align="center"><bold>Relative Risk (RR) of excess mortality</bold></td><td align="center"><bold>P Values</bold></td><td align="center"><bold>Total counties</bold></td></tr></thead><tbody><tr><td align="center">Most Likely Cluster for non-Hispanic Whites (90% study period and 50%population at risk)</td><td align="center">35.0</td><td align="center">1990 – 1996</td><td align="center">4088</td><td align="center">3623</td><td align="center">1.12</td><td align="center">0.001</td><td align="center">42</td></tr><tr><td align="center">Secondary Cluster for non-Hispanic Whites</td><td align="center">34.8</td><td align="center">1990 – 1993</td><td align="center">2240</td><td align="center">1995</td><td align="center">1.12</td><td align="center">0.001</td><td align="center">41</td></tr><tr><td align="center">Primary Cluster for Hispanic White</td><td align="center">29.0</td><td align="center">1990 – 1998</td><td align="center">1436</td><td align="center">1212</td><td align="center">1.18</td><td align="center">0.001</td><td align="center">127</td></tr><tr><td align="center">Primary Cluster for Blacks (note: not statistically significant)</td><td align="center">34.1</td><td align="center">1991 – 1996</td><td align="center">791</td><td align="center">682</td><td align="center">1.15</td><td align="center">0.129</td><td align="center">15</td></tr></tbody></table></table-wrap><p>To investigate those counties of excess mortality involving more than one racial group, we further conducted a spatial query. For both non-Hispanic Whites and Blacks, there were 3 counties in Gulf Coast Texas (Harris, Galveston and Brazoria counties) presented excess mortality, and 9 counties in Central Texas (Hays, Comal, Guadalupe, Bexar, Medina, Wilson, Atascosa, Frio and McMullen) had excess mortality among non-Hispanic Whites and Hispanics.</p></sec><sec><title>Discussion</title><p>The results indicate that between 1990 and 1998, four geographic regions were identified with excess mortality rates in Texas that were statistically significant. With respect to suspected excess mortality, the regions detected with excess breast cancer mortality were consistent with those presented in the analyses of the 1970–1994 data by the US National Cancer Institute [<xref ref-type="bibr" rid="B2">2</xref>] and the 1990–1997 data reported by Zhan. [<xref ref-type="bibr" rid="B4">4</xref>] The results rendered supporting evidence that most counties that were previously suspected of having elevated breast cancer mortality do indeed have excessive cancer mortality. The present study additionally identified West Texas counties as having excess mortality from breast cancer in the Hispanic female population that persisted for 9 years, which was not previously reported. The relative risk of this cluster was at the modest level of 1.18. Nevertheless, this region had the highest in relative risk, with the longest temporal persistence among detected potential clusters of all racial groups in this study. Based on this finding and on the comparisons of LLRs for the primary suspected clusters from both scan trials (non-Hispanic Whites LLR = 35.00 vs. Blacks LLR = 10.01 vs. Hispanic LLR = 29.01 vs. Others LLR = 8.30), it was determined that the Hispanic and non-Hispanic White female populations in the regions detected with clusters had the highest burden of breast cancer mortality, as evidenced by both temporal persistence and spatial concentration.</p><p>The verification of breast cancer excess mortality over time may prove beneficial for health policy and planning. For instance, the state of Texas has yet to reach the Healthy People 2010 Objective of 16.3 deaths per 100,000 females in the population, as provided by the U.S. Centers for Disease Control and Prevention (i.e. Objective 3-3, "reduce the breast cancer death rate", <ext-link ext-link-type="uri" xlink:href="http://www.healthypeople.gov/Document/html/tracking/od03.htm"/>). Spatiotemporal analysis such as that described in this study will be instrumental in planning and reaching the projected objective. For instance, the present analysis underscored the two regions with multiple racial groups that bear the persistent burden of breast cancer mortality, and detected a potential 9-year persistence of excess breast cancer mortality with the highest relative risk in the Hispanic population. Both regions carry a disproportional burden of excess mortality and warrant further investigation and policy intervention. The results of spatiotemporal analysis quantified disease burden over time by both spatial concentration (as determined by p values, LLRs and relative risks) and temporal persistence (as determined by the duration of detected clusters), which presented another perspective of measuring health disparity. It contributed to an understanding of the persistent burden of the disease across space and time, as well as aiding in determining whether the mortality burden that may have persisted into the current decade.</p><p>Several research notes arose from this study and warrant elaboration. First, as identified in the present study, the very modest relative risks that occurred over a large region of contiguous counties in Texas did not necessarily meet the strict definition of "clusters" of epidemic intensity. Compared with previous studies using SatScan for cluster detection, [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B19">19</xref>] the relative risks revealed in the present study were apparently lower, and no localized, hot-spot clusters (with constant, high risks in the clusters) that persisted over time were detected. On the other hand, breast cancer may have a substantial developmental period and may have potential risk/vulnerability factors, such as the stage at diagnosis, access to treatment, and the exposure to environmental toxic wastes that are not fully understood. These potential contributors were not accounted for in the present study. Given that many of these causes and risk factors may have operated over various time scales, the mortality examined here is only an endpoint in that process. While early detection of cancer is generally beneficial to survival, there is controversy over the effectiveness of breast cancer screening in reducing mortality.[<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>] Ideal interventions may also target modifiable risk factors that exist above and beyond the windows of space or time considered here. Nevertheless, this study offered baseline descriptions of persistently elevated breast cancer deaths in Texas, which may serve as a point of departure for policy deliberation and health resource allocation. Second, although this study focused primarily on statistically significant excess mortality, it by no means suggested that those non-statistically-significant regions of excess mortality were less important. To be statistically significant at the 0.05 or 0.01 level, outcome measures had to satisfy the Poisson distribution model and all independent variables of this study, including space, time and age, had to fit into the model simultaneously, and produce a large LLR as a result of spatial-temporal analysis. For example, the potential cluster detected among Blacks between 1991–1996 in Gulf Coast Texas (RR = 1.15, p = 0.12) was for all age groups. However, the results may become statistically significant if analysis was conducted with the stratification of certain age groups, such as among Black females aged 25 to 40. Therefore, the p-value derived is construed as an indicator, suggesting the level of excess mortality that calls for further investigation. Third, the choice of county level analysis entailed the strengths and weaknesses intrinsic to this level of aggregation. Although sub-county level (such as census tracts or block groups) of analysis may be preferred in cancer analysis, [<xref ref-type="bibr" rid="B22">22</xref>] we chose the county-level data because this level of aggregation was used in other studies on detecting breast cancer clusters,[<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B15">15</xref>] and also because the disproportional demographic distribution of Texas population made sub-county level analysis less feasible. For example, there were seven border counties that averaged fewer than ten breast cancer deaths, and had a population of less than 900 residents during the study period. The rates based on these small numbers of events and small population sizes tend to be unpredictable and often inflated. In particular, several of the above sparsely populated counties were in the summary choropleth maps of the age-adjusted mortality rates by race presented in Figures <xref ref-type="fig" rid="F1">1</xref> to <xref ref-type="fig" rid="F3">3</xref>. These inflated mortality rates tend to produce visual bias, as these are the counties that attract more visual attention due to the intensive colour shading. Readers are advised to use caution when trying to interpret health outcomes, including excess mortality in these sparsely-populated counties. Fourth, with reference to data management, we found that most spreadsheet programs are limited in accommodating data in a worksheet (for example, 65,536 rows or records). Conventional spreadsheet programs are insufficient for storing the aggregated data of multiple years required in the present study (N > 195 k). Instead of using spreadsheet programs, we recommend the use of a relational database for data management for spatial-temporal analysis using SatScan.</p><p>As observed in Jacquez and Greiling [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>] of the Journal, several methodological issues involving spatial analysis warrant consideration when using SatScan for cluster analysis. Among other potential limitations are the assumption that the clusters are cylindrically-shaped and the constraints that are attributable to centroids and the edge effects of the scan method. As the authors advised, the scan statistic is but one tool that one may bring to bear on the study of geographic variation in cancer. Nevertheless, th e shape of cluster detection in SatScan may be enhanced over time, and improved methods of utilizing SatScan for cluster analysis are emerging [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. As users of this program, we found that it affords a great opportunity of analyzing the unique geo-demographic composition of Texas data. Particularly appealing is the fact that the program is in the public domain. It provides an opportunity for the integration (as a calculation engine) with other mapping programs, such as EpiMap, freely available from the CDC. Currently, the authors are working to develop an integrated solution using these two programs as a health surveillance system for Texas counties.</p></sec><sec><title>Conclusion</title><p>Between 1990 and 2001, 4 regions of potential excess breast cancer mortality of statistical significance were detected in Texas counties among non-Hispanic White and Hispanic groups, for the time period between 1990 and 1998. Among all racial groups, both non-Hispanic Whites in Southeast Gulf Coast Texas and Hispanics in West Texas had the highest mortality burden of this disease, as evidenced by spatial concentration and temporal persistence. There was no evidence that the excess mortality may have persisted through the year 2000 or later in Texas. The excessive occurrence of breast cancer in 3 counties in the Gulf Coast region (for non-Hispanic Whites and Blacks) and in West Texas (for Hispanics) warrants further investigation.</p></sec><sec><title>Authors' Contributions</title><p>All authors collaborated intensively in this study. CEH contributed to the conceptualization of this study, including visual basic programming and data analysis. HJ and FSM contributed to data interpretation, health disparity analysis and the realignment of the manuscript. All read and approved the final manuscript.</p></sec>
Cholesterol-lowering properties of <italic>Ganoderma lucidum in vitro</italic>, <italic>ex vivo</italic>, and in hamsters and minipigs	<sec><title>Introduction</title><p>There has been renewed interest in mushroom medicinal properties. We studied cholesterol lowering properties of <italic>Ganoderma lucidum </italic>(<italic>Gl</italic>), a renowned medicinal species.</p></sec><sec><title>Results</title><p>Organic fractions containing oxygenated lanosterol derivatives inhibited cholesterol synthesis in T9A4 hepatocytes. In hamsters, 5% <italic>Gl </italic>did not effect LDL; but decreased total cholesterol (TC) 9.8%, and HDL 11.2%. <italic>Gl </italic>(2.5 and 5%) had effects on several fecal neutral sterols and bile acids. Both <italic>Gl </italic>doses reduced hepatic microsomal <italic>ex-vivo </italic>HMG-CoA reductase activity. In minipigs, 2.5 <italic>Gl </italic>decreased TC, LDL- and HDL cholesterol 20, 27, and 18%, respectively (P < 0.05); increased fecal cholestanol and coprostanol; and decreased cholate.</p></sec><sec><title>Conclusions</title><p>Overall, <italic>Gl </italic>has potential to reduce LDL cholesterol <italic>in vivo </italic>through various mechanisms. Next steps are to: fully characterize bioactive components in lipid soluble/insoluble fractions; evaluate bioactivity of isolated fractions; and examine human cholesterol lowering properties. Innovative new cholesterol-lowering foods and medicines containing <italic>Gl </italic>are envisioned.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Berger</surname><given-names>A</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Rein</surname><given-names>D</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Kratky</surname><given-names>E</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Monnard</surname><given-names>I</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Hajjaj</surname><given-names>H</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Meirim</surname><given-names>I</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Piguet-Welsch</surname><given-names>C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Hauser</surname><given-names>J</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Mace</surname><given-names>K</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Niederberger</surname><given-names>P</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Lipids in Health and Disease	<sec><title>Background</title><p>In Kampo Chinese folk medicine, mushrooms have been known to have medicinal properties since AD1200 [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>In recent years, there has been interest in the cholesterol lowering properties of mushrooms, including <italic>Ganoderma lucidum </italic>(Reishi-, Longevity-, or Phantom mushrooms, Biladi Top, Young-zhi, The King Of Herbs, Ling Zhi in Chinese, Saru-no-koshikake and Mannendake in Japanese) [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>], <italic>Pleurotus ostreatus </italic>(Oyster mushroom) [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B8">8</xref>], <italic>Volvariella volvacea </italic>(Straw mushroom) [<xref ref-type="bibr" rid="B9">9</xref>], <italic>Agaricus bisporus </italic>(champignon) [<xref ref-type="bibr" rid="B10">10</xref>], <italic>Agaricus campestris </italic>[<xref ref-type="bibr" rid="B11">11</xref>], <italic>Auricularia auricula </italic>(Tree-ear), <italic>Tremella fuciformis </italic>(White-jelly leaf) [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>], <italic>Grifola frondosa </italic>(Maitake mushroom) [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>], <italic>Lentinus erodes </italic>(Shiitake) and isolated fractions [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B16">16</xref>], and <italic>Polyporus confluens </italic>(Ningyotake) [<xref ref-type="bibr" rid="B17">17</xref>]. In an earlier work, Kaneda and Tokuda [<xref ref-type="bibr" rid="B18">18</xref>] studied cholesterol lowering properties of ether-, water- and ethanol extracts from caps and stems from <italic>Lentinus edodes</italic>, <italic>Auricularia polytricha </italic>(Jews-ear), <italic>Flammulina velutipes</italic>, and <italic>Agaricus bisporus</italic>. The majority of these studies were performed in rats. The cholesterol lowering properties of <italic>Cordyceps sinensis </italic>were studied in humans [<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Our focus is <italic>Gl</italic>, an important medicinal fungus belonging to the Ganodermataceae family that has been studied for its many interesting health promoting properties, including anti-tumor, anti-inflammatory, and anti-platelet aggregation [<xref ref-type="bibr" rid="B20">20</xref>-<xref ref-type="bibr" rid="B27">27</xref>]<ext-link ext-link-type="uri" xlink:href="http://kyotan.com/lectures/lectures"/>. Indeed, entire books, symposiums, organizations (<italic>e.g.</italic>, the <italic>Ganoderma </italic>International Research Institute, New York) and therapies have been devoted to <italic>Gl</italic>. As further testament to its importance, in ancient Chinese times, a Reishi Goddess (Reishi senshi) was even worshipped to bestow health, life and eternal youth.</p><p>As described, <italic>Gl </italic>has been occasionally studied for its cholesterol lowering- and hypotensive properties in the rat [<xref ref-type="bibr" rid="B2">2</xref>] and rabbit [<xref ref-type="bibr" rid="B28">28</xref>], but not in more physiological cholesterol models [<xref ref-type="bibr" rid="B29">29</xref>] such as minipigs. <italic>Gl </italic>can supposedly lower cholesterol in humans, but the work was not peer-reviewed nor adequately described [<xref ref-type="bibr" rid="B24">24</xref>]<ext-link ext-link-type="uri" xlink:href="http://www.ganotherapyusa.com/DXN/docs/whatis.htm"/>.</p><p>Like humans, minipigs are omnivours, and their lipid and steroid metabolism, and digestive and cardiovascular physiology closely resembles that of humans [<xref ref-type="bibr" rid="B30">30</xref>-<xref ref-type="bibr" rid="B32">32</xref>]; whereas in contrast to humans, rodents carry most of the cholesterol in HDL fractions unless they are fed high saturated fat and cholesterol rich diets, which has the effect of shutting down LDL receptors [<xref ref-type="bibr" rid="B29">29</xref>].</p><p>The components in <italic>Gl </italic>that may lower cholesterol are not known, but may include ganoderan-type glucans [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>], hetero-β-glucans, glucan-protein complexes (xyloglucans, uronic acid-β-glucans), other fibers, lectins [<xref ref-type="bibr" rid="B25">25</xref>], terpenoid triterpenes [<xref ref-type="bibr" rid="B35">35</xref>-<xref ref-type="bibr" rid="B38">38</xref>], ergostane sterols [<xref ref-type="bibr" rid="B39">39</xref>], and highly oxygenated ganoderic acid-type, lanostanoid triterpenes [<xref ref-type="bibr" rid="B38">38</xref>-<xref ref-type="bibr" rid="B42">42</xref>]. <italic>Gl </italic>fibrous components could affect cholesterol absorption and bile acid recycling, whereas lipophilic components could affect cholesterol synthesis.</p><p><italic>Gl </italic>may affect cholesterol synthesis at the committed 3-hydroxy-3-methylglutaryl coenzyme A reductase (HMG-CoA reductase) rate-limiting step; or at the latter lanosterol 14α-methyl demethylase:cytochrome P-450 demethylase (P-45014DM) step [<xref ref-type="bibr" rid="B43">43</xref>,<xref ref-type="bibr" rid="B44">44</xref>], catalyzing the rate limiting step in lanosterol-cholesterol conversion. In non-<italic>Gl </italic>mushroom species, inhibition of squalene synthetase by zaragozic acid fungal metabolites has also been reported in primates [<xref ref-type="bibr" rid="B45">45</xref>].</p><p>Herein, we tested the effects of <italic>Gl </italic>on cholesterol metabolism in hepatic T9A4 human cells, a hamster small animal model, and a minipig larger animal model having different lipoprotein cholesterol distribution than the hamster model. Animal models were fed cholesterol-containing diets described in Tables <xref ref-type="table" rid="T1">1</xref>, <xref ref-type="table" rid="T2">2</xref>.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Proximate analysis of Nafag 924 test diets for hamsters<sup>1</sup></p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="3">Weight %</td></tr><tr><td align="left">Component</td><td align="left">Control</td><td align="left">2.5% <italic>Gl</italic></td><td align="left">5.0% <italic>Gl</italic></td></tr></thead><tbody><tr><td align="left">Carbohydrate (by difference)</td><td align="left">31.8</td><td align="left">31.0</td><td align="left">30.3</td></tr><tr><td align="left">Starch</td><td align="left">23.6</td><td align="left">23.0</td><td align="left">22.5</td></tr><tr><td align="left">Crude Protein</td><td align="left">15.6</td><td align="left">15.2</td><td align="left">14.8</td></tr><tr><td align="left">Water</td><td align="left">8.0</td><td align="left">7.8</td><td align="left">7.6</td></tr><tr><td align="left">Crude fat</td><td align="left">4.4</td><td align="left">4.3</td><td align="left">4.2</td></tr><tr><td align="left">Ash</td><td align="left">4.7</td><td align="left">4.6</td><td align="left">4.5</td></tr><tr><td align="left">Crude fiber</td><td align="left">3.5</td><td align="left">3.4</td><td align="left">3.3</td></tr><tr><td align="left">Essential amino acids</td><td align="left">3.4</td><td align="left">3.3</td><td align="left">3.2</td></tr><tr><td align="left"><italic>Gl </italic>extract</td><td align="left">0.0</td><td align="left">2.5</td><td align="left">5.0</td></tr><tr><td align="left">Vitamin mix (includes choline)</td><td align="left">2.4</td><td align="left">2.3</td><td align="left">2.3</td></tr><tr><td align="left">Minerals (Ca, P, Mg, K, Na)</td><td align="left">2.3</td><td align="left">2.2</td><td align="left">2.2</td></tr><tr><td align="left">Trace elements</td><td align="left">0.3</td><td align="left">0.3</td><td align="left">0.3</td></tr></tbody></table><table-wrap-foot><p><sup>1</sup>Nafag 924 hamster complete diet was from Eberle Nafag AG, Gossau, Switzerland. The detailed ingredients in the diet are not known. Metabolizable energy was estimated to be 3111 kcal/kg diet. The lovastatin diet was identical to the control diet, but contained 2 mg/100 g diet lovastatin. Gl, <italic>G. lucidum.</italic></p></table-wrap-foot></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Ingredients in test diets for minipigs<sup>1</sup></p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Component (g/100 g diet, as fed basis)</td><td align="left">Weight % Control or 2.5% <italic>Gl</italic></td></tr></thead><tbody><tr><td align="left">Corn (to 100%)</td><td align="left">26.6</td></tr><tr><td align="left">Wheat shorts</td><td align="left">29.8</td></tr><tr><td align="left">Pork fat<sup>2</sup></td><td align="left">9.0</td></tr><tr><td align="left">Soy meal (44%; contains soy protein)</td><td align="left">8.5</td></tr><tr><td align="left">Bakery by products</td><td align="left">8.5</td></tr><tr><td align="left">Unsalted, melted butter<sup>3</sup></td><td align="left">4.5</td></tr><tr><td align="left">Amino acid mix<sup>4</sup></td><td align="left">3.4</td></tr><tr><td align="left">Mineral mix<sup>5</sup></td><td align="left">2.8</td></tr><tr><td align="left">Cellulose<sup>6 </sup>(control) or <italic>Gl</italic><sup>7</sup></td><td align="left">2.5</td></tr><tr><td align="left">Canola (rapeseed) meal</td><td align="left">2.5</td></tr><tr><td align="left">Poultry meal</td><td align="left">1.7</td></tr><tr><td align="left">Cholesterol<sup>8</sup></td><td align="left">0.1</td></tr><tr><td align="left">Vitamin mix (includes choline)<sup>9</sup></td><td align="left">0.1</td></tr><tr><td align="left">Trace element mix<sup>10</sup></td><td align="left">0.02</td></tr></tbody></table><table-wrap-foot><p><sup>1</sup>Diets were custom prepared by Kliba (Kaiseraugst, Switzerland) as a fat and cholesterol enriched diet 2604. Unless indicated otherwise, stocks were from Kliba. Proximate analysis of minipig test diets (wt%) was: carbohydrate as nitrogen free extract 46.6, crude protein 15.9, water 11.1, crude fat 16.8, ash (including vitamins) 5.1, and crude fiber 4.6. Total- and digestible energy of the diet were estimated to be 3985 and 3776 kcal/kg diet, respectively. <sup>2</sup>Centravo Schweinefett B 90, Centravo AG, Zurich, Switzerland. Contained 98% fat, 5% free fatty acids, 9% polyunsaturated fatty acids, protected with antioxidants. <sup>3</sup>Migros Genossenschafts-Bund SA, Zurich, Switzerland. 82% fat, 0.5% protein, 0.5% carbohydrate. <sup>4</sup>Per 100 g diet, contained 1.06% arginine, 0.75% lysine, 0.26% methionine, 0.56% methionine + cystine, 0.20% tryptophan, and 0.57% threonine [<xref ref-type="bibr" rid="B78">78</xref>]. <sup>5</sup>Per 100 g diet, contained 0.80% calcium, 0.70% phosphorus, 0.19% sodium, 0.63% potassium, 0.20% magnesium, and 0.26% chloride. <sup>6</sup>Vitacel LC 200 Cellulose, J. Rettenmaier & Söhne (JRS), GMBH + Co, Rosenberg, Holzmühle 1, Germany. 0.3% sulfate ash, pH 5.0–7.5, 300 μM fiber length. <sup>7</sup>Champitec, Payerne, Switzerland. Prepared as described in the text. Pigs that did not receive the <italic>Gl </italic>extract, received the control diet plus 80 mg lovastatin/pig/d in half an apple. <sup>8</sup>Fluka 26740, Fluka Holding AG, Buchs, Switzerland. 97% pure. <sup>9</sup>Roche Vitamins Ltd, Basel, Switzerland. In mg/100 g diet, contained 0.4 vitamin A (800 IU), 2.0 vitamin D<sub>3 </sub>(80 IU), 10.5 vitamin E, 0.3 vitamin K<sub>3</sub>, 1.2 vitamin B<sub>1</sub>, 0.8 vitamin B<sub>2</sub>, 3.0 nicotinic acid, 2.0 pantothenic acid, 0.1 folic acid, 0.7 vitamin B<sub>6</sub>, 0.0034 vitamin B<sub>12</sub>, 0.02 biotin, 65.1 choline, and 2.0 vitamin C. <sup>10</sup>In mg/100 g diet, contained 1.1 copper (mg/kg), 7.6 zinc, 11.0 iron, 0.05 iodine, 5.0 manganese, and 0.03 selenium. <italic>Gl</italic>, <italic>G. lucidum.</italic></p></table-wrap-foot></table-wrap></sec><sec><title>Results</title><sec><title>Active components in <italic>Gl </italic>and <italic>in vitro </italic>activity</title><p>Organic and aqueous <italic>Gl </italic>phases did not contain HPLC-detectable lovastatin. The organic extracted phase strongly inhibited cholesterol biosynthesis (ID<sub>50 </sub>= 1.3 μg/mL, relative to 0.4 for lovastatin), while the aqueous phase was ineffective (ID<sub>50 </sub>> 330). Various highly oxygenated lanostanoid triterpenes, and 32-methyl- and 26-oxo sterols were found in the organic phase, and likely contributed to inhibition of cholesterol synthesis. A 20% EtOAc/hexane fraction contained ganoderal A; and a 50% EtOAc/hexane contained ganoderols-A and B, and Y ganoderic acid.</p></sec><sec><title>Body and organ weights, and food intake of hamsters</title><p>Body weights ranged from 68.7–70.8 and 83.2–86.4 g for the experimental groups on D1 and D18, respectively, without significant differences relative to control, on D1, D18, or D18 minus D1. D18 liver and cecum relative weights (g organ/100 g body wt) were 2.77–2.84 and 0.52–0.56 for the various groups, respectively, without significant differences relative to control. Daily food intake was 7.1–7.8 g food/d averaged over D1-16; there were no significant differences relative to control.</p></sec><sec><title>Cholesterol and triacylglycerol in hamsters</title><p>Starting D1 TC levels did not differ among the groups, whereas there were differences in D1 TAG (Table <xref ref-type="table" rid="T3">3</xref>). <italic>Gl </italic>at 2.5 and 5.0% reduced D18 TAG (likely due to D1 TAG differential starting values). <italic>Gl </italic>at 2.5% did not reduce D18 TC, LDL or HDL. With 5.0% <italic>Gl</italic>, there was a statistical trend (P < 0.10) to reduce TC and HDL; LDL was not affected. Similarly to the higher dose of <italic>Gl</italic>, lovastatin decreased D18 TC and HDL, but not LDL. LDL/HDL ratio was not statistically significantly different for any dietary treatments relative to control.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Plasma cholesterol and triacylglycerol in hamsters treated with <italic>G. lucidum </italic>and lovastatin (mmol/L)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Group</td><td align="left">TC</td><td align="left">TC</td><td align="left">TAG</td><td align="left">TAG</td><td align="left">VLDL</td><td align="left">LDL</td><td align="left">HDL</td><td align="left">LDL/HDL</td></tr><tr><td></td><td align="left">D1</td><td align="left">D18</td><td align="left">D1</td><td align="left">D18</td><td align="left">D18</td><td align="left">D18</td><td align="left">D18</td><td align="left">D18</td></tr></thead><tbody><tr><td align="left">Control</td><td align="left">3.57</td><td align="left">3.48</td><td align="left">1.10</td><td align="left">1.08</td><td align="left">0.22</td><td align="left">0.51</td><td align="left">2.75</td><td align="left">0.19</td></tr><tr><td align="left">Lovastatin</td><td align="left">3.39</td><td align="left">3.16<sup>a</sup></td><td align="left">1.07</td><td align="left">0.98</td><td align="left">0.18<sup>a</sup></td><td align="left">0.48</td><td align="left">2.50<sup>a</sup></td><td align="left">0.19</td></tr><tr><td align="left"><italic>Gl </italic>(2.5%)</td><td align="left">3.46</td><td align="left">3.40</td><td align="left">0.64<sup>a</sup></td><td align="left">0.90<sup>a</sup></td><td align="left">0.18</td><td align="left">0.53</td><td align="left">2.69</td><td align="left">0.20</td></tr><tr><td align="left"><italic>Gl </italic>(5%)</td><td align="left">3.34</td><td align="left">3.14<sup>a</sup></td><td align="left">0.89<sup>a</sup></td><td align="left">0.92<sup>a</sup></td><td align="left">0.20</td><td align="left">0.49</td><td align="left">2.44<sup>a</sup></td><td align="left">0.20</td></tr></tbody></table><table-wrap-foot><p>TC, total cholesterol; TAG, triacylglycerol; VLDL, LDL, HDL, very low-, low-, and high-density lipoproteins; D, day; LOVA, 20 mg/kg diet. Values represent mean of 6 animals. <sup>a</sup>Significantly different from the control (P < 0.05 or P < 0.10), students, unpaired, 1-tailed,<italic>t</italic>-test, equal variances.</p></table-wrap-foot></table-wrap></sec><sec><title>Fecal bile acids and neutral sterols in hamsters</title><p><italic>Gl </italic>(2.5%) increased fecal total bile acids and chenodeoxycholate (Table <xref ref-type="table" rid="T4">4</xref>). Both <italic>Gl </italic>doses increased coprostanol 3-one, whereas, 5% <italic>Gl </italic>decreased cholestanol. Lovastatin had no significant effects on bile acids or neutral sterols examined.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Fecal bile acids and neutral sterols in hamsters treated with <italic>G. lucidum </italic>and lovastatin (nmol/g dry feces/d)</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="6">Bile acids</td><td align="center" colspan="5">Neutral Sterols</td></tr></thead><tbody><tr><td align="left">Group</td><td align="left">C</td><td align="left">LC</td><td align="left">DC</td><td align="left">CDC</td><td align="left">UDC</td><td align="left">Total BA</td><td align="left">COP-ol</td><td align="left">COP-3-one</td><td align="left">CHOL erol</td><td align="left">CHOL anol</td><td align="left">Total NS</td></tr><tr><td colspan="12"><hr></hr></td></tr><tr><td align="left">Control</td><td align="left">15.0</td><td align="left">14.3</td><td align="left">10.7</td><td align="left">5.1</td><td align="left">1.7</td><td align="left">46.7</td><td align="left">11.5</td><td align="left">ND</td><td align="left">7.2</td><td align="left">18.4</td><td align="left">37.1</td></tr><tr><td align="left">Lovastatin</td><td align="left">15.0</td><td align="left">12.8</td><td align="left">11.6</td><td align="left">6.2</td><td align="left">1.6</td><td align="left">47.1</td><td align="left">12.0</td><td align="left">ND</td><td align="left">6.7</td><td align="left">18.9</td><td align="left">37.6</td></tr><tr><td align="left"><italic>Gl </italic>(2.5%)</td><td align="left">16.9</td><td align="left">16.2</td><td align="left">14.3</td><td align="left">8.5<sup>a</sup></td><td align="left">1.9</td><td align="left">57.8<sup>a</sup></td><td align="left">11.7</td><td align="left">4.2<sup>a</sup></td><td align="left">7.7</td><td align="left">16.9</td><td align="left">40.6</td></tr><tr><td align="left"><italic>Gl </italic>(5%)</td><td align="left">12.2</td><td align="left">13.9</td><td align="left">12.0</td><td align="left">6.4</td><td align="left">1.4</td><td align="left">45.8</td><td align="left">12.0</td><td align="left">7.8<sup>a</sup></td><td align="left">7.0</td><td align="left">14.1<sup>a</sup></td><td align="left">41.0</td></tr></tbody></table><table-wrap-foot><p>Feces were collected quantitatively at the conclusion of the experiment (D18). C, cholate; LC, lithocholate; DC, deoxycholate; CDC, chenodeoxycholate, UDC, Ursodeoxycholate; COP, coprostan; CHOL, cholest; BA, bile acids; NS, neutral sterols. ND, not detected. Values represent mean of 6 determinations. <sup>a</sup>Significantly different from the control (<italic>P </italic>< 0.05), students unpaired, 2-tailed <italic>t</italic>-test, equal variances. No statistical trends (P < 0.10) existed in the data set.</p></table-wrap-foot></table-wrap></sec><sec><title><italic>Ex vivo </italic>hepatic HMG-CoA reductase activity in hamsters</title><p>Lovastatin did not affect de-phosphorylated activity, and phosphorylated activity not examined (Table <xref ref-type="table" rid="T5">5</xref>). In absence of NaF (inhibitor of phosphatase) and in presence of 2.5 and 5 % <italic>Gl</italic>, 3-hydroxy-3-methylglutaryl-CoA reductase activity in hamster hepatic microsomes (pmol/min/g liver) was reduced 2.1- and and 1.5 fold, respectively, relative to the control. In presence of NaF, 2.5% and 5% <italic>Gl </italic>reduced HMG-CoA reductase 3.5- and 1.9-fold, respectively, relative to control.</p><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>HMG-CoA reductase activity in hamster hepatic microsomes (pmol [<sup>14</sup>C]mevalonolactone/min/mg microsomal protein)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Group</td><td align="left">Activity-NaF</td><td align="left">Total activity+NaF</td></tr></thead><tbody><tr><td align="left">Control</td><td align="left">12.36</td><td align="left">9.30</td></tr><tr><td align="left">Lovastatin</td><td align="left">13.34</td><td align="left">ND</td></tr><tr><td align="left"><italic>Gl </italic>(2.5%)</td><td align="left">5.86<sup>a</sup></td><td align="left">2.66<sup>a</sup></td></tr><tr><td align="left"><italic>Gl </italic>(5%)</td><td align="left">8.08<sup>a</sup></td><td align="left">4.86<sup>a</sup></td></tr></tbody></table><table-wrap-foot><p>Refer to the text for methodologic details. [<sup>14</sup>C]mevalonate was converted to [<sup>14</sup>C]mevalonolactone with 10 M HCl before the TLC step. Values above were corrected for recovery (15–63%) with [<sup>3</sup>H]mevalonate. <italic>G. lucidum</italic>, but not lovastatin, was found to inhibit <italic>ex vivo </italic>synthesis whether: recovery was accounted for; results were expressed per gram liver weight; [<sup>14</sup>C]mevalonolactone was first extracted into organic solvent and then applied to TLC plates [<xref ref-type="bibr" rid="B79">79</xref>] rather than applying aqueous extracts directly to TLC plates [<xref ref-type="bibr" rid="B55">55</xref>] as reported above (data not shown); and whether or not endogenous phosphatase activity was inhibited with 50 mM NaF (shown above). Values represent mean of 6 determinations. <sup>a</sup>Significantly different from control, students unpaired, 2-tailed <italic>t</italic>-test, equal variances (P < 0.05). ND, not determined.</p></table-wrap-foot></table-wrap></sec><sec><title>Fractional cholesterol synthesis rate in hamsters</title><p>In hamsters, 24 h FSR values (Atom% enrichment D17-18) were 1.68 ± 0.20, 1.91 ± 0.16, 1.75 ± 0.36, and 2.29 ± 0.05 (mean of n = 6, ± 1 SEM) for control, lovastatin, 2.5%-, and 5% <italic>Gl</italic>, respectively. Values were not statistically significantly different from control.</p></sec><sec><title>Body weights of minipigs</title><p>Minipig body weights increased equivalently with <italic>Gl </italic>and lovastatin from 19.0–26.9 kg over D1-28. Similar weights per age were previously reported for experimentally-fed Göttingen minipigs [<xref ref-type="bibr" rid="B46">46</xref>].</p></sec><sec><title>Cholesterol and triacylglycerol in minipigs</title><p>The experimental diet increased TC 27–30% from D1-14 (Table <xref ref-type="table" rid="T6">6</xref>). In the <italic>Gl</italic>-fed group, TC significantly decreased 12.5% from D14-21, but not further from D21-29; the decrease in TC from D14-29 was 20% (P < 0.01). Lovastatin did not significantly decrease TC during D14-21 (P > 0.13), D21-29, nor D14-29; but TC did decrease >10% in two pigs from D14-21.</p><table-wrap position="float" id="T6"><label>Table 6</label><caption><p>Plasma cholesterol and triacylglycerol in minipigs treated with <italic>G. lucidum </italic>and lovastatin (mmol/L)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Group</td><td align="left">TC</td><td align="left">TC</td><td align="left">TC</td><td align="left">TC</td><td align="left">TAG</td><td align="left">TAG</td><td align="left">TAG</td><td align="left">TAG</td><td align="left">VLDL</td><td align="left">VLDL</td><td align="left">LDL</td><td align="left">LDL</td><td align="left">HDL</td><td align="left">HDL</td><td align="left">LDL/HDL</td><td align="left">LDL/HDL</td></tr><tr><td></td><td align="left">D1</td><td align="left">D14</td><td align="left">D21</td><td align="left">D29</td><td align="left">D1</td><td align="left">D14</td><td align="left">D21</td><td align="left">D29</td><td align="left">D14</td><td align="left">D29</td><td align="left">D14</td><td align="left">D29</td><td align="left">D14</td><td align="left">D29</td><td align="left">D14</td><td align="left">D29</td></tr></thead><tbody><tr><td align="left"><italic>Gl </italic>(2.5%<italic>)</italic></td><td align="left">2.47<sup>a</sup></td><td align="left">3.21<sup>bc</sup></td><td align="left">2.81</td><td align="left">2.58</td><td align="left">0.53</td><td align="left">0.57</td><td align="left">0.80</td><td align="left">0.69</td><td align="left">0.07</td><td align="left">0.09</td><td align="left">1.45<sup>c</sup></td><td align="left">1.08</td><td align="left">1.69<sup>c</sup></td><td align="left">1.42</td><td align="left">0.88</td><td align="left">0.79</td></tr><tr><td align="left">Lovastatin</td><td align="left">2.36<sup>a</sup></td><td align="left">3.00</td><td align="left">2.44</td><td align="left">2.81</td><td align="left">0.50</td><td align="left">0.60</td><td align="left">0.59</td><td align="left">0.71</td><td align="left">0.10</td><td align="left">0.09</td><td align="left">1.40</td><td align="left">1.29</td><td align="left">1.50</td><td align="left">1.43</td><td align="left">0.95</td><td align="left">0.91</td></tr></tbody></table><table-wrap-foot><p>Lovastatin was administered at 80 mg/d. Between D1-14, all pigs received a high cholesterol and fat control diet; from D15-29, pigs received either <italic>G. lucidum </italic>(<italic>Gl</italic>) extract or lovastatin. The same statistical conclusions were reached if all 10 pigs were compared between D1-14. Pigs were randomly selected to receive either <italic>GI </italic>or lovastatin before study commencement. Student's, paired, 1-tailed, <italic>t</italic>-test, was utilized for statistical comparisons. Statistically significant changes (P < 0.05, 1-tailed testing) in cholesterol parameters are indicated as follows: <sup>a</sup>D1 vs. 14; <sup>b</sup>D14 vs. 21; <sup>c</sup>D14 vs. 29. Abbreviations: refer to Table <xref ref-type="table" rid="T3">3</xref>. There was a slight trend for 2.5% <italic>Gl </italic>to reduce LDL/HDL ratio between D14-29 (P < 0.11, 1-tailed testing).</p></table-wrap-foot></table-wrap><p>There were no significant differences in TAG and VLDL with <italic>Gl </italic>or lovastatin (Table <xref ref-type="table" rid="T6">6</xref>). VLDL was however a minor lipoprotein pool. Lovastatin had not significant effects on LDL nor HDL; <italic>Gl </italic>decreased LDL 26% and HDL 16% (P < 0.01; D14 vs 29). <italic>Gl </italic>did not affect statistically significantly affect LDL/HDL since both individual parameters decreased from D14-29.</p></sec><sec><title>Fecal bile acids and neutral sterols in minipigs</title><p>The high cholesterol-fat diet decreased chenodeoxycholate; and increased coprostanol, coprostan 3-one, and cholesterol from D1-14 (P < 0.05 or < 0.10; Table <xref ref-type="table" rid="T7">7</xref>). <italic>Gl </italic>trended to increase cholestanol (D14 vs 29; P < 0.10).</p><table-wrap position="float" id="T7"><label>Table 7</label><caption><p>Fecal bile acids and neutral sterols in minipigs treated with <italic>G. lucidum </italic>and lovastatin (nmol/g dry feces)</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="3">Bile acids</td><td align="center" colspan="4">Neutral Sterols</td></tr></thead><tbody><tr><td align="left">Group</td><td align="left">NCT</td><td align="left">C</td><td align="left">CDC</td><td align="left">COP-ol</td><td align="left">COP-3-one</td><td align="left">CHOL erol</td><td align="left">CHOL anol</td></tr><tr><td colspan="8"><hr></hr></td></tr><tr><td align="left">All pigs (D1)</td><td align="left">1.98</td><td align="left">0.86</td><td align="left">0.66<sup>a</sup></td><td align="left">2.41<sup>a</sup></td><td align="left">0.10<sup>a</sup></td><td align="left">1.73<sup>a</sup></td><td align="left">0.96</td></tr><tr><td align="left">All pigs (D14)</td><td align="left">1.98</td><td align="left">1.33</td><td align="left">0.16</td><td align="left">3.75</td><td align="left">0.17</td><td align="left">2.96</td><td align="left">0.91</td></tr><tr><td align="left"><italic>Gl </italic>(D14)</td><td align="left">1.98</td><td align="left">1.61</td><td align="left">0.14</td><td align="left">3.61</td><td align="left">0.18</td><td align="left">3.37</td><td align="left">0.87<sup>b</sup></td></tr><tr><td align="left"><italic>Gl </italic>(D29)</td><td align="left">1.98</td><td align="left">0.81</td><td align="left">0.16</td><td align="left">4.44</td><td align="left">0.15</td><td align="left">3.15</td><td align="left">1.22</td></tr><tr><td align="left">Lovastatin (D14)</td><td align="left">1.98</td><td align="left">0.99</td><td align="left">0.19</td><td align="left">3.92</td><td align="left">0.15</td><td align="left">2.44</td><td align="left">0.96</td></tr><tr><td align="left">Lovastatin (D29)</td><td align="left">1.98</td><td align="left">1.23</td><td align="left">0.18</td><td align="left">3.28</td><td align="left">0.16</td><td align="left">2.34</td><td align="left">1.14</td></tr></tbody></table><table-wrap-foot><p>Feces were collected on D1, 14 and 29. A quantitative fecal collection was not possible, hence results are expressed per gram of feces. Abbreviations: refer to Table <xref ref-type="table" rid="T4">4</xref>, except, NCT, 23- Nor β cholanate 5α,7α,12α triol. A paired, 2-tailed students <italic>t</italic>-test, equal variances, evaluated effects of the high cholesterol and fat diet, between D1-14 (10 minipigs); and the effects of <italic>Gl </italic>or lovastatin between D15-29 (5 minipigs/group), indicated as follows: <sup>a</sup>D1 vs 14, for all minipigs combined (<italic>P </italic>< 0.001).<sup>b</sup>D14 vs 29 (P < 0.05). 0.05 < P < 0.1, to indicate statistical trends.</p></table-wrap-foot></table-wrap></sec></sec><sec><title>Discussion</title><sec><title>Active components in <italic>Gl </italic>and <italic>in vitro </italic>activity</title><p>As described, lovastatin was not detected in our <italic>Gl </italic>mushroom preparations. By contrast, statin-like compounds have been found in oyster mushrooms [<xref ref-type="bibr" rid="B47">47</xref>] and <italic>Chrysosporium pannorum </italic>[<xref ref-type="bibr" rid="B48">48</xref>].</p><p>We did however detect oxygenated lanosterol molecules such as 32-methyl- and 26-oxo sterols, ganoderols-A and B, Y ganoderic acid, and ganoderals-A and B in the organic layer. The organic layer strongly inhibited cholesterol biosynthesis from acetate. Similar or identical oxygenated lanosteroids had been previously reported in <italic>Gl </italic>[<xref ref-type="bibr" rid="B38">38</xref>-<xref ref-type="bibr" rid="B42">42</xref>], and found to inhibit conversion of 24,25-dihydrolanosterol to cholesterol at the lanosterol 14 α-demethylase step [<xref ref-type="bibr" rid="B49">49</xref>-<xref ref-type="bibr" rid="B51">51</xref>], and also indirectly to inhibit HMG-CoA reductase activity [<xref ref-type="bibr" rid="B51">51</xref>]. The fact that the aqueous phase from <italic>Gl </italic>was ineffective at inhibiting cholesterol synthesis (ID<sub>50 </sub>> 330) suggests that hydrophilic molecules such as glucans and fibers in <italic>Gl </italic>do not affect conversion of acetate to cholesterol. Such molecules may however affect cholesterol absorption and bile acid recycling.</p></sec><sec><title><italic>Ex vivo </italic>hepatic HMG-CoA reductase and fractional cholesterol synthesis rate in hamsters</title><p>The observed inhibition of <italic>ex-vivo </italic>HMG-CoA reductase activity in hamsters treated with <italic>Gl </italic>has similarly been observed with <italic>Gl </italic>in rats [<xref ref-type="bibr" rid="B51">51</xref>], and with pure lanosterol analogs [<xref ref-type="bibr" rid="B44">44</xref>,<xref ref-type="bibr" rid="B52">52</xref>]. Our lack of effect with lovastatin (4.3 μmol/kg body wt) contrasts results with the related statin, simvastatin, where 10, 30, and 60 μmol/kg body wt/d increased <italic>ex-vivo </italic>hepatic HMG-CoA reductase activity 2-, 17-, and 50-fold, respectively [<xref ref-type="bibr" rid="B53">53</xref>]. Lovastatin could have different effects on HMG-CoA reductase and other enzymes than simvastatin, and was not however examined in the above study.</p><p>Lanosterol analogs such as those found in <italic>Gl </italic>are known to inhibit translation of HMG-CoA reductase mRNA, and may also accelerate protein degradation [<xref ref-type="bibr" rid="B44">44</xref>,<xref ref-type="bibr" rid="B52">52</xref>]. <italic>Gl </italic>may also affect cholesterol biosynthesis at latter biosynthetic steps such as the conversion of lanosterol [<xref ref-type="bibr" rid="B51">51</xref>], which could in turn, indirectly inhibit HMG-CoA reductase activity, as reported for statins in minipigs [<xref ref-type="bibr" rid="B53">53</xref>]. Indeed, it was reported that repression of the lanosterol 14 α-demethylase step can result in accumulation of 3 β-hydroxy-lanost-8-en-32-al, a known translational downregulator of HMG-CoA reductase [<xref ref-type="bibr" rid="B54">54</xref>].</p><p>If <italic>Gl </italic>had direct physical effects on HMG-CoA reductase activity, this implies that even after the 16 h fast employed in hamsters, <italic>Gl </italic>components were still bound to the enzyme during the assay procedure [<xref ref-type="bibr" rid="B55">55</xref>]. After the 16 h fast, lovastatin could have been removed from the enzyme accounting for the lack of observed effects of lovastatin on <italic>ex-vivo </italic>HMG-CoA reductase activity. Due to removal of the drug, other statins have even been found to increase <italic>ex-vivo </italic>HMG-CoA reductase activity [<xref ref-type="bibr" rid="B56">56</xref>]. Hepatic <italic>ex-vivo </italic>HMG-CoA reductase activity and whole body cholesterol FSR are entirely different types of measurements. It is not clear why <italic>Gl </italic>and lovastatin did not influence cholesterol FSR in hamsters. In principle, the low saturated fat-cholesterol condition employed via use of a chow diet, should have led to a high endogenous rate of cholesterol synthesis, one that could be inhibited by <italic>Gl </italic>and lovastatin. It is conceivable that the <italic>Gl </italic>and lovastatin became decomposed in the dietary mixture. To test this hypothesis, we re-extracted <italic>Gl </italic>and lovastatin from stored diets after culmination of the experiments, and found no differences in bioactive components analyzed, compared to the original starting materials (before addition to the diets; data not shown).</p></sec><sec><title>Cholesterol and triacylglycerol in hamsters and minipigs</title><p>Hamsters were fed a low-cholesterol chow-based diet with no added exogenous cholesterol or saturated fat. Under these conditions, there was not sufficient cholesterol to redistribute cholesterol from the HDL to LDL pool [<xref ref-type="bibr" rid="B29">29</xref>]. This is why in hamsters, 5% <italic>Gl </italic>and lovastatin reduced D18 TC and HDL, but not LDL [<xref ref-type="bibr" rid="B57">57</xref>,<xref ref-type="bibr" rid="B58">58</xref>].</p><p>Using the same types of diet, lovastatin was similarly found to preferentially reduce HDL in hamsters; and only when dietary saturated fat was added, were both LDL and HDL reduced [<xref ref-type="bibr" rid="B57">57</xref>].</p><p>Another factor contributing to the lack of strong effects in hamsters, and the total lack of effect in minipigs may be that the dose of lovastatin was insufficient. In hamsters, the employed dose of 2 mg lovastatin/100 g diet is ca. 4.3 μmol lovastatin/kg body wt. Himber <italic>et al. </italic>[<xref ref-type="bibr" rid="B57">57</xref>] treated hamsters with 25 μmol lovastatin/kg body wt, which lowered HDL; or 50 μmol, which lowered LDL and HDL [<xref ref-type="bibr" rid="B57">57</xref>]. Morand <italic>et al</italic>. [<xref ref-type="bibr" rid="B53">53</xref>] found that 20–200 μmol simvastatin/kg body wt was sufficient to reduce LDL. Ma <italic>et al</italic>. [<xref ref-type="bibr" rid="B59">59</xref>] reduced lipoproteins in hamsters with 100 mg lovastatin/100 g diet. In minipigs, we utilized a dose of 80 mg lovastatin/minipig/d, which may also have been on the low side. A dose of 24–42 mg was sufficient to lower lipoproteins in Hyde Park minipigs [<xref ref-type="bibr" rid="B60">60</xref>]. Nevertheless, our particular species, strain, and location of minipigs may have responded less aggressively to lovastatin (M. Huff, Personal Communication, December 2000). In Göttingen minipigs, a dose of 80 mg simvastatin lowered LDL, whereas 240 mg lowered LDL and HDL [<xref ref-type="bibr" rid="B53">53</xref>]; simvastatin is likely more effective in minipigs than lovastatin at a similar dietary weight percent [<xref ref-type="bibr" rid="B61">61</xref>,<xref ref-type="bibr" rid="B62">62</xref>].</p><p>The reduction in TAG with <italic>Gl </italic>was likely due to lower D1 TAG values in the <italic>Gl </italic>groups relative to control. TAG reductions in hamster models typically occur under conditions of higher saturated fat intake [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B63">63</xref>]. In the only other peer-reviewed study examining cholesterol lowering properties of <italic>Gl </italic>in a small animal model, 5 dietary wt% dried Reishi mushroom powder was found to decrease TC in SHR rats; effects on VLDL, LDL and HDL were not studied [<xref ref-type="bibr" rid="B2">2</xref>]. In minipigs, with the high fat-cholesterol feeding conditions employed, a <italic>Gl</italic>-induced inhibition of cholesterol synthesis should result in less availability of hepatic cholesterol for lipoprotein synthesis. In turn, this has the potential effect of reducing plasma VLDL cholesterol secretion, reducing LDL direct secretion; and possibly reducing VLDL-LDL conversion [<xref ref-type="bibr" rid="B64">64</xref>,<xref ref-type="bibr" rid="B65">65</xref>]. In the present work, we did not observe differences in TAG or VLDL in pigs fed either <italic>Gl </italic>or lovastatin, however this effect could have been missed since the VLDL pool represented only a small lipoprotein pool and/or there was efficient VLDL-LDL conversion. The reductions in both LDL and HDL with <italic>Gl </italic>is consistent with that seen with higher statin doses [<xref ref-type="bibr" rid="B53">53</xref>].</p></sec><sec><title>Fecal bile acids and neutral sterols in hamsters and minipigs</title><p>In hamsters, <italic>Gl </italic>increased fecal total bile acids and chenodeoxycholate, whereas both doses, increased coprostanol 3-one; the 5% dose decreased cholestanol for unclear reasons. An increase in fecal chenodeoxycholate likely indicates production or recycling of chenodeoxycholate was enhanced.</p><p>Plasma levels of cholestanol are positively associated with cholesterol absorption [<xref ref-type="bibr" rid="B66">66</xref>]; whereas decreased fecal cholestanol may indicate plasma cholestanol was increased and cholesterol absorption was enhanced. In minipigs, <italic>Gl </italic>tended to increase fecal cholestanol, the opposite pattern to that of hamsters fed 5% <italic>Gl. </italic>Coprostanol and coprostanol 3-one are the bacterial products of cholesterol, which are increased when fecal cholesterol is increased, or when gut flora are altered [<xref ref-type="bibr" rid="B67">67</xref>]. Since fecal cholesterol and coprostanol levels were not changed by either dose of <italic>Gl</italic>, it is not obvious why coprostanol 3-one accumulated.</p><p>Bile salts are now known to possess many different functions acting as detergents, activators of protein kinase C and phosphatidylinositol-3 kinase; and being important gene regulators [<xref ref-type="bibr" rid="B68">68</xref>,<xref ref-type="bibr" rid="B69">69</xref>]. Chenodeoxycholate, deoxycholate, and their glycine and taurine conjugates can lead to farnesoid X receptor/retinoid X receptor (FXR/RXR)-induced activation of intestinal bile acid binding protein transcription (I-BABP), and suppression of CYP7α RNA and protein levels (FXR prevents liver X receptor (LXRα)-induced transactivation of CYP7α). CYP7α regulates the committed step in classical bile acid synthesis. Overall, an increased fecal level of chenodeoxycholate would mean less chenodeoxycholate is available to activate FXR. Less activation of FXR would lead to less bile acid recycling and less inhibition of bile acid synthesis, more hepatic cholesterol converted to bile acids, and a lowering of plasma cholesterol.</p><p>Overall, it is likely that fibrous and/or lipophilic sterol-like molecules in <italic>Gl </italic>altered the absorption and recycling of bile acids and neutral sterols, leading to altered fecal accumulation. Monitoring plasma levels of neutral sterols and bile acids, and quantifying conjugated and de-conjugated bile acids, should help to clarify the potential importance of the observed trends.</p></sec><sec><title>Comparing <italic>in vitro</italic>, <italic>ex vivo</italic>, and <italic>in vivo </italic>results</title><p>In the present work, the <italic>in vitro </italic>experiments were performed with fractionated <italic>Gl </italic>extracts, whereas the <italic>ex-vivo </italic>and <italic>in vivo </italic>work utilized intact <italic>Gl</italic>. Intact <italic>Gl </italic>contains fibrous components, which may have affected bile acid and neutral sterol absorption and recycling. Fibrous components could also impair the uptake of lipophilic components, such as those inhibiting <italic>in vitro </italic>cholesterol synthesis. An additional complexity is that lipophilic components such as ergostane sterols [<xref ref-type="bibr" rid="B39">39</xref>] could also affect bile acid and neutral sterol levels. Thus, it is difficult to directly compare our <italic>in vitro </italic>and <italic>in vivo </italic>results. Feeding fractionated and intact mushrooms should help to unravel the <italic>in vivo </italic>bioactive components, as has been accomplished for oyster mushrooms [<xref ref-type="bibr" rid="B70">70</xref>].</p></sec></sec><sec><title>Conclusions and key findings</title><p>In summary, <italic>GI </italic>was found to have cholesterol lowering potential <italic>in vitro</italic>, <italic>ex-vitro</italic>, and in two animal models, with some differences between the two animal models. It is possible that oxygenated lanosterol derivatives in <italic>Gl </italic>(partly characterized in the present work) contributed to this cholesterol lowering by decreasing cholesterol synthesis (changes in <italic>in vitro </italic>and <italic>ex-vivo</italic>, but not whole body, cholesterol synthesis were apparent in the present work). Fibrous components and glucans in <italic>Gl </italic>were likely responsible for the observed alterations in fecal neutral sterols and bile acids in both animal species, ultimately affecting cholesterol absorption and bile acid recycling and contributing to cholesterol lowering. Next steps are to examine the cholesterol lowering properties of various doses of intact and fractionated, chemically characterized, <italic>Gl </italic>components in a placebo-controlled clinical trial. Animal experimentation should also utilize fractionated materials, and ideally, elucidate mechanisms of action of each bioactive component. Positive cholesterol-lowering results in such studies will pave the way for adding <italic>Gl </italic>to new cholesterol-lowering foods and medicines, alone, and in combination with other established cholesterol-lowering ingredients and drugs.</p></sec><sec sec-type="materials\|methods"><title>Materials and methods</title><sec sec-type="materials"><title>Materials</title><p><italic>Gl </italic>was from Fermenta SA, Payerne, Switzerland. Mushrooms were cultivated on a defined formula of sawdust, wheat straw and millet grain. Substrate was sterilized at 90°C for 48 h, then incubated with <italic>Gl </italic>seed material from Mycotec Sàrl (Cernier, Neuchâtel). Cultivation was with controlled temperature, light, humidity and carbon dioxide concentration. Human hepatic T9A4 cells [<xref ref-type="bibr" rid="B71">71</xref>] were grown in LCM serum-free media under 3.5% CO<sub>2 </sub>at 37°C. Lovastatin was purchased as 20 mg Mevacor tablets (MSD Chibropharm GmbH, Haar, Germany). HMG-CoA reductase, <italic>DL</italic>-3-Glutaryl-3- [<sup>14</sup>C]-HMG-CoA (2216 MBq/mmol), R-[5-<sup>3</sup>H] mevalonic acid ammonium salt (1443 MBq/mmol), and [1-<sup>14</sup>C] acetic acid sodium salt (2070 MBq/mmol) were from Amersham (Upsala, Sweden). α-3-HMG-CoA (cold) and liquid scintillation cocktail were from Sigma (Buchs, Switzerland). LCM cell medium was from Biofluids (Rockville, MD). 5β-cholesteane-3α-ol, 5-α-cholestane and 2,3-nor-5β-cholanicacid-3α,7α,12α-triol were from Steraloids, Inc. (Newport, Rhode Island); other steroid standards were from Sigma, and Calbiochem (La Jolla, California). Methanolic HCl and Sylon HTP were from Supelco (Buchs, Switzerland). The Cobas Bio autosampler was from Hoffmann-La Roche (Basel, Switzerland) and reagents were from Roche Diagnostics (Rotkreuz, Switzerland). Total Cholesterol Kit 352 and Triacylglycerol Kit 336 were from Sigma. Deuterium was from Cambridge Isotope Laboratories (Andover, MA). Zn catalyst was from Biochemical laboratories (University Bloomington, IN). Silica gel thin layer chromatography (TLC) plates were from Merck Eurolab (Dietikon, Switzerland). Coomassie Plus-200 protein assay reagents and bovine serum albumin fraction V were from Pierce (Rockford, Illinois). All other chemicals were from Sigma.</p></sec><sec><title>Preparation of Gl for <italic>in vitro </italic>testing</title><p>Fruiting bodies from <italic>Gl </italic>(20 g) were dried, milled and macerated in 0.4 L MeOH/H<sub>2</sub>O (4:1, v/v) at room temperature for 3d. The mixture was then filtered, evaporated, re-dissolved in H<sub>2</sub>O, acidified to pH 3 with 3 M HCl, extracted 3 × with 150 mL ethyl acetate, and the organic phase evaporated under vacuum at 30°C, re-dissolved in 10 mL MeOH, and dried with Na<sub>2</sub>SO<sub>4</sub>, for HPLC analyses and <italic>in vitro </italic>testing.</p></sec><sec><title>Chemical analysis of Gl</title><p>The presence of lovastatin in <italic>Gl </italic>was determined by HPLC with a Nucleosil 100-5 C18 column (250 × 4 mm; Macherey-Nagel, Oensingen, Switzerland) and a Lichrospher 100 RP-18 post column (Merck, Glattbrugg, Switzerland). Solvent A was H<sub>3</sub>PO<sub>4</sub>/H<sub>2</sub>O (1:2000, by vol); solvent B was acetonitrile. Separation was initiated with a linear gradient of 95% A, 5% B, reaching 50% A, 50% B in 45 min, 30% A, 70% B in 46 min, 10% A, 90% B in 48 min, and 0% A, 100% B in 50 min; the run was continued isocratically 4 min. Initial conditions were maintained 6 min for re-equilibration; the flow rate was 1 mL/min. The detector was a G1315 A, series 1100 detector (Hewlett Packard, Meyrin, Switzerland); absorbance was measured at 254 nm. After selective extraction and purification with different adsorbents and solvents, ganoderols and ganoderic acids were detected by mass spectroscopy and NMR (details to be published separately).</p></sec><sec><title><italic>In vitro </italic>activity of Gl extracts</title><p>Human hepatic T9A4 cells were grown in LCM serum-free media under 3.5% CO<sub>2 </sub>at 37°C. Cells were seeded in 24-well plates and at confluence, incubated with 1 mM <sup>14</sup>C-acetate (1 mCi/mmol) for 20 h ± mushroom extracts. Lipids were extracted from cells by incubating 2 × with 1.5 mL hexane/isopropanol (3:2, by vol) for 30 min at room temperature. Combined organic extracts were dried under N<sub>2</sub>, re-dissolved in hexane, and separated by TLC with hexane/diethyl ether/acetic acid (75:25:1, by vol). Cholesterol synthesis was determined by measuring incorporation of <sup>14</sup>C from acetate to cholesterol. Radioactivity was assessed with an instant imager and expressed as percent of control.</p></sec><sec><title>Administration of Gl and lovastatin to hamsters</title><p>Male Golden Syrian hamsters (Harlan, UK), 3–4 wks, 40–60 g, were housed individually in Macrolon Type 3 cages with 12 h alternating periods of light and darkness. During 3 wks preceding treatment, hamsters were fed Nafag 924 hamster complete diet (# 3132/20, Eberle Nafag AG, Gossau, Switzerland; Table <xref ref-type="table" rid="T1">1</xref>). Following body weight randomization, groups consisted of 6 hamsters/group receiving either: Nafag diet (control), Nafag mixed with 2 mg lovastatin /100 g diet (powdered in liquid N<sub>2</sub>); or Nafag mixed with 2.5 or 5.0% dried <italic>Gl</italic>. Hamsters were fed experimental diets for 17 d. Lovastatin is an inhibitor of HMG-CoA reductase [<xref ref-type="bibr" rid="B72">72</xref>], and was used as a positive control. Dietary intake was recorded daily, body weights weekly. Feces were collected on D15-18. Hamsters were injected subcutaneously with 250 μL D<sub>2</sub>O on D17 and killed under anesthesia with isoflurane on D18. Following a 16 h fast, D1 (0.5 mL) and D18 blood (>3 mL) were obtained from the retro-orbital cavity and cardiac vein, respectively, and transferred to EDTA tubes. Plasma was prepared by centrifugation at 1500 g, 15 min, at 4°C. Plasma, and hepatic and cecum tissues were stored at -80°C. Animal procedures were authorized by Service Vétérinaire du Canton de Vaud, Switzerland, protocol 1247.</p></sec><sec><title>Administration of Gl and lovastatin to minipigs</title><p>Nine female and one male Göttingen minipig(s) (Jörg Farm in Bern Switzerland; Minipig-Primärzucht, Auswill, Switzerland) aged 6–12 mo (18–20 kg), with white (7) and black (3 minipigs) colorations, were housed in a 30 m<sup>2 </sup>box with normal light/dark cycle, and kept at room temperature. Females were chosen because they have fewer age-related lipid modifications and higher lipid concentrations than males [<xref ref-type="bibr" rid="B73">73</xref>]. One male was accidentally provided in the delivery, however its total cholesterol (TC), lipoproteins, bile acids and neutral sterols were similar to that of other minipigs. Minipigs were randomly distributed by weight into two separately housed groups, marked with a plastic label in the ear, and fed twice daily for 11 d with powdered commercial pig chow (Diet 574, Minipig-Primärzucht). During a subsequent 4 d adaptation period, minipigs were fed an acclimatization mixture of chow and increasing amounts of powdered hypercholesterolemic control diet (custom diet 2604, Kliba, Kaiseraugst, Switzerland; Table <xref ref-type="table" rid="T2">2</xref>) from 0% to 100%, in steps of 25%, designed after Burnett <italic>et al. </italic>[<xref ref-type="bibr" rid="B64">64</xref>,<xref ref-type="bibr" rid="B74">74</xref>], that was consistent with Göttingen minipig nutritional needs [<xref ref-type="bibr" rid="B75">75</xref>]. During the following 2 wks (D15-29), groups were fed control hypercholesterolemic diet pre-mixed with 2.5% <italic>Gl </italic>extract; or hypercholesterolemic diet plus 80 mg lovastatin/pig/d (in four 20 mg tablets) [<xref ref-type="bibr" rid="B53">53</xref>], hand fed to each minipig, mornings, in half an apple. For acclimatization, on D12-14, minipigs received a half apple without lovastatin. The study was blinded in that the diets were coded, and the mushroom extract was referred to as "Nestlé Special Fiber." Food intake was 3.5% of body wt/d (based on group average wt), readjusted weekly, to provide sufficient, but not excessive, calories [<xref ref-type="bibr" rid="B64">64</xref>,<xref ref-type="bibr" rid="B65">65</xref>,<xref ref-type="bibr" rid="B75">75</xref>]. Diets were distributed at 0700 and 15h00, and spread linearly on a clean cement surface to facilitate individual consummation. Distilled water was provided <italic>ad libitum</italic>. Toys and human contact were provided to avoid boredom. Fasting 16 h blood samples (10 mL; 20 mL on D29) were collected in EDTA tubes on D1, 15, 22, 28, and 29 from anterior vena cava. Plasma was prepared by centrifugation as described for hamsters, and stored at -80°C. Blood collection began at 0800 following injection of the intra-muscular relaxant Dormicum<sup>® </sup>(Hoffmann La Roche, Basel, Switzerland), then the tranqulizer Stresnil<sup>® </sup>(Janssen Pharmaceuticals, Beerse, Belgium). After blood sampling on D1, 15 and 29, minipigs were isolated for 2 h maximum for individual fecal collections. Some minipigs did not defecate during this period, whereas others defecated again following return to their groups. Hence, the morning fecal collection was qualitative. Feces were stored at -40°C under N<sub>2</sub>. Body weight was recorded weekly, and food intake recorded each morning. Minipigs were donated to the University of Geneva at the study's conclusion. Animal procedures were authorized by Service Vétérinaire du Canton de Geneve, Switzerland, protocol 1315, authorization 31.1.1014/1719/1.</p></sec><sec><title>Cholesterol and triacylglycerol measurements in hamsters and minipigs</title><p>Plasma total cholesterol and triacylglycerol were measured using commercial kits and a Roche Cobas Bio autosampler. Plasma lipoproteins were separated by size-exclusion HPLC as previously described [<xref ref-type="bibr" rid="B63">63</xref>].</p></sec><sec><title>Fractional cholesterol synthesis rate measurements in hamsters and minipigs</title><p>Measurements of water- and cholesterol deuterium enrichment were performed with a Finnigan Thermoquest Delta XL plus Isotopic Ratio Mass Spectrometer (Bremen, Germany) as previously described [<xref ref-type="bibr" rid="B76">76</xref>,<xref ref-type="bibr" rid="B77">77</xref>]. Fractional synthesis rate (FSR) of free cholesterol was calculated from a plasma sample collected 24 h after deuterium oxide subcutaneous injection as follows: FSR (in % pool/d) = 100 × (cholesterol enrichment/(water enrichment × 0.478)). Due to technical reasons, there were insufficient values in the minipig experiments to reach interpretable conclusions.</p></sec><sec><title>Fecal bile acids and neutral sterol measurements in hamsters and minipigs</title><p>Fecal neutral sterols and bile acids were extracted from lyopholized feces, deconjugated, derivatized with Sylon HTP and analyzed by gas chromatography as previously described with internal standards: 5-α-cholestane for neutral sterols; 2,3-nor-5β-cholanic acid-3α,7α,12α-triol for bile acids [<xref ref-type="bibr" rid="B63">63</xref>].</p></sec><sec><title>Hepatic <italic>ex-vivo </italic>HMG-CoA reductase measurements in hamsters</title><p>Freshly excised liver (300 mg) was collected after 16 h fast of hamsters, minced with scissors, and homogenized with 0.4 mL buffer (50 mM KH<sub>2</sub>PO4, 0.1 M sucrose, 50 mM KCl, 50 mM NaCl, 30 mM EDTA, and 2 mM dithiothreitol, ± 50 mM NaF) with a Potter-Elvehjem S homogenizer with 400 rpm/5 strokes, on ice, after Conde <italic>et al. </italic>[<xref ref-type="bibr" rid="B55">55</xref>]. NaF inhibits dephosphorylation of HMG-CoA reductase by inactivating phosphoprotein phosphatases, yielding total phosphorylated HMG-CoA reductase activity. After washing homogenizer with 0.2 mL buffer, homogenate was centrifuged at 10000 g, 15 min, at 4°C. Supernatant was decanted, 0.4 mL cold buffer added, and the tube vortexed and re-centrifuged. Pooled post-mitochondrial supernatants were spun in 1.5 mL ultracentrifuge tubes at 150000 rpm, 10 min, at 4°C in a Sorvall Discovery M 150 micro ultracentrifuge (Kendro Laboratory Products SA, Carouge-Geneva, Switzerland), and microsomal fractions stored at -80°C. Microsomal protein (200 μg, 10–18 μL) was pre-incubated 10 min at 37°C in an agitating bath, then incubated 15 min with 50 μL substrate solution (buffer plus 90 mM glucose-6-phosphate, 72 mM EDTA, 9 mM NADP, 6.2 nmol cold HMG-CoA (0.12 mM), 1.3 nmol [<sup>14</sup>C]HMG-CoA (0.0025 MBq), 0.3 IU glucose-6-phosphate dehydrogenase, and 0.024 MBq [<sup>3</sup>H]mevalonic acid as recovery standard). After 15 min, reaction was terminated with 25 μL 10 M HCL, then incubated 30 min at 37°C for mevalonate-mevalonolactone conversion. Following centrifugation at 1000 g, 1 min, at 4°C to remove denatured protein, supernatant was applied to activated (1 h, 105°C) TLC plates, developed in fresh benzene-acetone (1:1, by vol), the mevalonolactone region scraped (based on migration of cold standards and X-ray film visualization; R<sub>f </sub>0.42–0.5), and radioactivity measured in 10 mL scintillation cocktail.</p></sec><sec><title>Statistics</title><p>Differences between groups were tested by unpaired/paired, one-tailed/two-tailed, student <italic>t</italic>-tests, equal variances, as appropriate for different measurements. Statistical significance was evaluated at P < 0.05 unless stated otherwise.</p></sec></sec><sec><title>List of abbreviations</title><p>D, day; FPLC, fast protein liquid chromatography; GC, gas chromatography; <italic>Gl or G. lucidum, Ganoderma lucidum; </italic>HMG-CoA, 3-hydroxy-3-methylglutaryl coenzyme A; HPLC, high pressure liquid chromatography; TLC, thin layer chromatography; TAG, triacylglycerol; TC, total cholesterol.</p></sec><sec><title>Authors' contributions</title><p>AB wrote and compiled the majority of the manuscript, was responsible for minipig studies, and served as project leader for animal cholesterol research. DR developed methods for cholesterol lipoprotein measurements, and was responsible for hamster studies. EK, I. Monnard, and JH assisted in both animal studies, and developed methods for neutral sterols, bile acids and <italic>ex vivo </italic>measurements. HH developed methods to chemically analyze <italic>Gl</italic>.I. Meirim and CPW developed methods for cholesterol synthetic rates. KM was responsible for <italic>in vitro </italic>biological testing of <italic>Gl </italic>extracts. P.Niederberger served as overall project leader.</p></sec>
Distribution of serum lipids and lipoproteins in patients with beta thalassaemia major; an epidemiological study in young adults from Greece	<sec><title>Background</title><p>Beta-thalassaemia major (b-TM) has been defined as a combination of chronic hemolytic anemia, iron storage disease and myocarditis, and it has been associated with premature death especially due to heart failure. To the best of our knowledge the status of blood lipids in these patients has rarely been investigated. Thus, we assessed the levels of lipids and lipoproteins in a sample of cardiovascular disease free adult men and women with b-TM.</p></sec><sec sec-type="methods"><title>Methods</title><p>During 2003 we enrolled 192 consecutive patients with b-TM that visited our Institution for routine examinations. The Institution is considered the major reference center for b-TM in Greece. Of the 192 patients, 88 were men (25 ± 6 years old) and 104 women (26 ± 6 years old). Fasting blood lipid levels were measured in all participants.</p></sec><sec><title>Results</title><p>Data analysis revealed that 4% of men and 2% of women had total serum cholesterol levels > 200 mg/dl, and 11% of men and 17% of women had triglyceride levels > 150 mg/dl. In addition, mean HDL cholesterol levels were 32 ± 11 mg/dl in men and 38 ± 10 mg/dl in women, lipoprotein-a levels were 8.3 ± 9 mg/dl in men and 8.8 ± 9 mg/dl in women, apolipoprotein-A1 levels were 111 ± 17 mg/dl in men and 123 ± 29 mg/dl in women, and apolipoprotein-B levels were 60 ± 20 mg/dl in men and 59 ± 14 mg/dl in women. Total-to-HDL cholesterol ratios were 3.7 ± 1.2 and 3.8 ± 1.5 in men and women, respectively.</p></sec><sec><title>Conclusions</title><p>The majority of the patients had blood lipid levels (by the exception of HDL-cholesterol) within the normal range, and consequently the prevalence of lipid and lipoprotein abnormalities was much lower as compared to the general population of the same age. Interestingly, is that the total – to HDL cholesterol ratio was high in our patients, and may underline the importance of this index for the prognosis of future cardiac events in these patients.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Chrysohoou</surname><given-names>Christina</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Panagiotakos</surname><given-names>Demosthenes B</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Pitsavos</surname><given-names>Christos</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Kosma</surname><given-names>Konstantina</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Barbetseas</surname><given-names>John</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Karagiorga</surname><given-names>Markisia</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Ladis</surname><given-names>Ioannis</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Stefanadis</surname><given-names>Christodoulos</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Lipids in Health and Disease	<sec><title>Introduction</title><p>Beta thalassaemia major (b-TM) is a very serious blood condition since individuals with it are unable to make enough healthy red blood cells and depend on blood transfusions all their life. However, quality and duration of life of transfusion-dependent thalassaemic patients has been transformed over the last few years, with their life expectancy increasing well into the third decade and beyond, with a good quality of life. Nevertheless, cardiac symptoms and premature death from cardiac causes are still major problems since in the absence of effective iron chelation therapy, many patients develop evidence of iron-induced myocardial damage with cardiac failure, cardiac arrhythmia, sudden death, or a distressing lingering death from progressive congestive cardiac failure [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B4">4</xref>].</p><p>During the past years many scientific evidences have raised the adverse effect of abnormal blood lipid levels, like total cholesterol and other lipids and lipoproteins, on atherosclerotic disease [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. At this point it should be mentioned that the relationships between blood lipids and atherosclerosis might be influenced by several other lifestyle-related factors, like glucose intolerance; blood pressure levels, dietary habits and smoking habits [<xref ref-type="bibr" rid="B8">8</xref>].</p><p>To the best of our knowledge data regarding the distribution of blood lipids levels among patients with b-TM are lacking. Therefore, we investigated the distribution of total-, HDL-, LDL-, cholesterol, triglycerides, apolipoprotein-A1 and B and lipoprotein – (a) levels, in a sample of patients with b-TM, in Greece.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Population of the study</title><p>From January to December 2003 we enrolled all patients (i.e. 192) with b-TM who visited the "Saint Sophia" Pediatric Hospital for routine blood transfusion. The aforementioned Institute is the major reference center for b-TM in Greece. Of the 192 patients, 88 were males, aged 25 ± 6 years old (range 18–42) and 104 were females, aged 26 ± 6 years old (range 18–46). The number of enrolled participants is adequate to evaluate differences between the investigated parameters greater than 20%, achieving statistical power > 0.80 at < 0.05 probability level (P-value). Moreover, the selected sample can be considered as representative since there were only minor, insignificant, differences in sex and age distribution between the study population and the target population regarding the sex-age distribution. All patients were Caucasians, they were living in various regions of Greece and interviewed by trained personnel who used a standard questionnaire. Participants had hematologic evidence of b-TM, i.e. profound hypo chromic anemia, mean erythrocyte volume less than 75 fl, electrophoretic hemoglobin A<sub>2 </sub>higher than 3.5% of total hemoglobin and both parents had beta thalassaemia. Moreover, presence of the disease was also evaluated by genetic analysis that performed in our Institution, which confirmed the absence or the reduced levels of alpha- or beta-chain synthesis in hemoglobin. Patients received red blood cell transfusions regularly, every 2 to 3 weeks, to maintain hemoglobin level 10 to 13 g/dl. All patients were under iron chelation therapy with deferoxamine, in a dose of 30 to 50 mg/kg given five to six rimes weekly subcutaneous. Chelation treatment was monitored by frequent estimation of feritine levels and urine excretion. All patients were without any evidence of heart failure at entry, as assessed according to the New York Heart Association (NYHA) classification classes I to IV [<xref ref-type="bibr" rid="B4">4</xref>], as well as by the recent guidelines of the European Society of Cardiology [<xref ref-type="bibr" rid="B9">9</xref>]. In addition, we did not include patients with other cardiovascular or systemic diseases, including rheumatic valve disease, chronic bronchitis and cirrhosis.</p></sec><sec><title>Blood lipids and lipoproteins evaluation</title><p>The blood samples were collected from the antecubital vein between 8 to 10 a.m., in a sitting position after 12 hours of fasting and avoiding of alcohol. The biochemical evaluation was carried out in the same laboratory that followed the criteria of the World Health Organization Lipid Reference Laboratories. All biochemical examinations (serum total cholesterol, HDL-cholesterol and triglycerides) were measured using chromatographic enzymic method in a Technicon automatic analyser RA-1000 (Dade Behring, Marburg, Germany). HDL cholesterol was determined after precipitation of the Apolipoprotein B containing lipoproteins with dextran-magnesium-chloride. Lipoprotein (a) was measured by a latex enhanced turbidimetric immuno-assay. Serum for the measurement of these lipids was harvested immediately after admission. LDL cholesterol calculated using the Friedwald formula: {total cholesterol} - {HDL cholesterol} - 1/5 (triglycerides). Non-HDL cholesterol is the total cholesterol minus HDL cholesterol.</p><p>Hypercholesterolemia was defined as total serum cholesterol levels greater than 200 mg/dl or the use of lipid-lowering agents and diabetes mellitus as a blood glucose > 125 mg/dl or the use of antidiabetic medication.</p><p>An internal quality control was in place for assessing the validity of cholesterol, triglyceride and HDL methods. The intra and inter-assay coefficients of variation of cholesterol levels did not exceed 3%, triglycerides 2% and HDL 4%.</p></sec><sec><title>Demographic, clinical and lifestyle characteristics</title><p>The study's questionnaire also included demographic characteristics like age, gender, and residence of the participants. Information about smoking habits was collected using a standardized questionnaire developed for the Study. Current smokers were defined as those who smoked at least one cigarette per day. Never smokers those who have never tried a cigarette in their life and former smokers were defined as those who had stopped smoking more than one year previously. For the multivariate statistical analyses cigarette smoking was quantified in pack-years (cigarette packs per day × years of smoking), adjusted for a nicotine content of 0.8 mg/cigarette.</p><p>Body mass index was calculated as weight (in kilograms) divided by standing height (in meters squared). Obesity was defined as body mass index > 29.9 Kg/m<sup>2</sup>.</p><p>Arterial blood pressure was measured three times at the right arm (ELKA aneroid manometric sphygmometer, Von Schlieben Co, West Germany), at the end of the physical examination with subject in sitting position at least for 30 minutes. The systolic blood pressure level was determined by the first perception of sound (of tapping quality). The diastolic blood pressure level was determined by phase V when the repetitive sounds become fully muffed (disappear). Changes in loudness were not considered. Patients whose average blood pressure levels were greater or equal to 140/90 mm Hg or were under antihypertensive medication were classified as hypertensives.</p></sec><sec><title>Statistical analysis</title><p>Continuous variables are presented as mean values ± one standard deviation, while qualitative variables are presented as absolute and relative frequencies. The use of contingency tables and the calculation of chi-squared test tested associations between categorical variables. Comparisons between normally distributed continuous variables and categorical were performed by the calculation of Student's t-test as well as Analysis of Co-Variance, after testing for equality of variances (homoscedacity). In the case of asymmetric continuous variables the tested hypotheses were based on the calculations of the non-parametric test suggested by Kruskal and Wallis. Correlations between lipids levels and age, smoking habits (in pack-years) and body mass index were evaluated by the calculation of Pearson's correlation coefficient for the normally distributed variables and by the Spearman correlation coefficient for the skewed variables.</p><p>All reported <italic>P</italic>-values are based on two-sided tests and compared to a significance level of 5%. SPSS 11.0 software (SPSS Inc. 2002, Illinois, USA) was used for all the statistical calculations.</p></sec></sec><sec><title>Results</title><p>Demographic and clinical characteristics of the patients are presented in Table <xref ref-type="table" rid="T1">1</xref>. It is of particular interest the very low prevalence of hypertension, result of the low systolic and diastolic blood pressure levels observed in these patients. Additionally, mean body mass index was within normal range (i.e. < 25 kg/m<sup>2</sup>); as result, obesity prevailed in less than 2% of the patients. On the other hand, a high proportion of men and women reported current smoking habits.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Characteristics of the patients (% by gender)</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">Men (n = 88)</td><td align="center">Women (n = 104)</td><td align="center">p</td></tr></thead><tbody><tr><td align="left">Age (years)</td><td align="center">25 ± 6</td><td align="center">26 ± 6</td><td align="center">0.33</td></tr><tr><td align="left">Current smoking, n (%)</td><td align="center">34 (45%)</td><td align="center">36 (34%)</td><td align="center">0.08</td></tr><tr><td align="left">Systolic blood pressure (mmHg)</td><td align="center">120 ± 43</td><td align="center">114 ± 15</td><td align="center">0.03</td></tr><tr><td align="left">Diastolic blood pressure (mmHg)</td><td align="center">71 ± 10</td><td align="center">74 ± 10</td><td align="center">0.11</td></tr><tr><td align="left">Hypertension, n (%)</td><td align="center">0 (0%)</td><td align="center">1 (1%)</td><td align="center">-</td></tr><tr><td align="left">Glucose (mg/dl)</td><td align="center">104 ± 28</td><td align="center">88 ± 16</td><td align="center">0.06</td></tr><tr><td align="left">Diabetes mellitus, n (%)</td><td align="center">12 (15%)</td><td align="center">17 (15%)</td><td align="center">0.88</td></tr><tr><td align="left">Body mass index (kg/m<sup>2</sup>)</td><td align="center">21.8 ± 2</td><td align="center">22.2 ± 3</td><td align="center">0.56</td></tr><tr><td align="left">Obesity, n (%)</td><td align="center">2 (2%)</td><td align="center">0 (0%)</td><td align="center">0.13</td></tr><tr><td align="left">Family history of CHD, n (%)</td><td align="center">8 (10%)</td><td align="center">17 (17%)</td><td align="center">0.16</td></tr><tr><td align="left">Hematocrit (%)</td><td></td><td></td><td></td></tr><tr><td align="left">Feritine (ng/dl)</td><td align="center">2411 ± 1784</td><td align="center">2076 ± 1350</td><td align="center">0.17</td></tr><tr><td align="left">Hemoglobin (g/dl)</td><td align="center">10.4 ± 1</td><td align="center">10.6 ± 1</td><td align="center">0.77</td></tr><tr><td align="left">White blood cell (counts)</td><td align="center">9959 ± 4999</td><td align="center">9401 ± 4074</td><td align="center">0.78</td></tr></tbody></table></table-wrap><sec><title>Blood lipids distribution</title><p>The mean values of the investigated blood lipids and lipoproteins both in men and women are presented in Table <xref ref-type="table" rid="T2">2</xref>. Furthermore, Figure <xref ref-type="fig" rid="F1">1</xref> illustrates the distribution of total, HDL cholesterol and triglycerides levels. Mean total cholesterol varied within normal values (< 200 mg/dl). In addition, none of the participants had cholesterol levels above 240 mg/dl, while only 4% of men and 2% of women had total serum cholesterol levels greater than 200 mg/dl. However, oppose results were observed regarding HDL cholesterol since 42% of men and 29% of women patients had very low HDL cholesterol levels (< 30 mg/dl). A group of people with particular interest is those who have normal total cholesterol, but low HDL cholesterol levels. In the our patients, 39% of men and 30% of women who had normal total cholesterol levels (i.e. < 200 mg/dl) had HDL cholesterol levels lower than 35 and 45 mg/dl, respectively. Mean triglycerides were also low, and 11% of men and 17% of women had triglyceride levels greater than 150 mg/dl. Finally, mean LDL cholesterol levels were also low and none of men and women had LDL cholesterol levels greater than 130 mg/dl.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Blood lipids distribution in men and women with b-TM</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">Men (n = 88)</td><td align="center">Women (n = 104)</td><td align="center">P*</td></tr></thead><tbody><tr><td align="left">Total cholesterol (mg/dl)</td><td align="center">114 ± 33</td><td align="center">129 ± 30</td><td align="center">0.018</td></tr><tr><td align="left">Triglycerides (mg/dl)</td><td align="center">113 ± 43</td><td align="center">112 ± 44</td><td align="center">0.98</td></tr><tr><td align="left">HDL-cholesterol (mg/dl)</td><td align="center">32 ± 11</td><td align="center">37 ± 13</td><td align="center">0.065</td></tr><tr><td align="left">Total – HDL cholesterol ratio</td><td align="center">3.7 ± 1.2</td><td align="center">3.8 ± 1.5</td><td align="center">0.75</td></tr><tr><td align="left">LDL-cholesterol (mg/dl)</td><td align="center">128 ± 38</td><td align="center">122 ± 36</td><td align="center">0.01</td></tr><tr><td align="left">Apolipoprotein-A1 (mg/dl)</td><td align="center">123 ± 28</td><td align="center">111 ± 17</td><td align="center">0.01</td></tr><tr><td align="left">Apolipoprotein-B (mg/dl)</td><td align="center">60 ± 20</td><td align="center">59 ± 15</td><td align="center">0.10</td></tr><tr><td align="left">Lipoprotein (a) (mg/dl)</td><td align="center">6.3 ± 8.8</td><td align="center">8.3 ± 9.1</td><td align="center">0.01</td></tr></tbody></table><table-wrap-foot><p><italic>P </italic>value derived from the comparison between men vs. women after taking into account the effect of age, body mass index, and smoking habits.</p></table-wrap-foot></table-wrap><fig position="float" id="F1"><label>Figure 1</label><caption><p>Distribution of selected blood lipids in men (left column) and women (right column) with beta thalassaemia major</p></caption><graphic xlink:href="1476-511X-3-3-1"/></fig><p>Mean lipoprotein – (a) levels were considerably low both in men and women patients. Moreover, the distribution of lipoprotein – (a) was skewed to the right indicating that a very small proportion of the patients had high (> 15 mg/dl) lipoprotein – (a) values (Figure <xref ref-type="fig" rid="F2">2</xref>).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Distribution of lipoprotein – (a) in men (left column) and women (right column) with beta thalassaemia major</p></caption><graphic xlink:href="1476-511X-3-3-2"/></fig><p>It is known that age is a factor that correlates well with blood lipid levels. In our study, age was positively and significantly associated with all blood lipids measurements in both men and women, by the exception of HDL-cholesterol levels. We expanded the previous findings by evaluating the association of the investigated blood lipids with age, after controlling for other potential confounders, like sex and smoking habits. We observed that a decade difference in age was associated with 7 mg/dl higher total cholesterol levels (95% confidence interval (CI) from 5 to 9 mg/dl, p < 0.001), 12 mg/dl higher triglycerides levels (95% CI from 10 to 14 mg/dl, p < 0.001), 7 mg/dl higher LDL cholesterol levels (95% CI from 5 to 9 mg/dl, p < 0.001), 0.9 mg/dl higher lipoprotein-(a) levels (95% CI from 0.4 to 1.4 mg/dl, p < 0.001), but only 2.5 mg/dl lower HDL cholesterol levels (95% CI from -1.5 to 6.5 mg/dl, p = 0.31).</p><p>Finally, none of the investigated blood lipids or lipoproteins was associated with feritine, hemoglobin levels or white blood cell counts (data not shown in text).</p></sec></sec><sec><title>Discussion</title><p>In this work we evaluated the distribution of several blood lipids and lipoproteins in a sample of Greek adults with beta thalassaemia major. To our knowledge the distribution of blood lipids and lipoproteins among patients with this disorder are presented for the first time in the literature. We found that the majority of the participants had normal total cholesterol levels, on the contrary a considerable proportion of the patients had very low HDL cholesterol levels. In addition, LDL cholesterol, triglycerides, as well lipoprotein – (a) and apolipoproteins A1 and B were substantially low.</p><sec><title>Epidemiology of blood lipids in patients with beta thalassaemia major</title><p>Only 4% of men and 2% of women had total cholesterol levels greater than 200 mg/dl. A recent report from the ATTICA study [<xref ref-type="bibr" rid="B10">10</xref>], which enrolled a representative and adequate sample of the general healthy population from Greece, suggest that roughly 25% of men and women with the same age of our patients (i.e. < 45 years old) had high total cholesterol. Based on the previous report it could be speculated that patients with beta thalassaemia major have lower total cholesterol levels as compared to healthy individuals of the same age. In accordance with the previous findings we also observed very low mean LDL cholesterol levels in thalassaemic patients. It is of interest that none of men and women patients had LDL cholesterol levels greater than 130 mg/dl. On the contrary, the ATTICA study [<xref ref-type="bibr" rid="B10">10</xref>] reported that 17% of men and 15% of women of the same age with our patients had LDL cholesterol levels above 130 mg/dl.</p><p>Mean triglycerides were also low in beta thalassaemic patients. Roughly one out of ten men and two out of ten women had triglyceride levels greater than 150 mg/dl. If we compare these figures with the ones reported by the ATTICA study investigators, i.e. that approximately one third of men and only 13% of women had triglycerides levels higher than 150 mg/dl, we may assume that thalassaemic men had substantially lower triglyceride levels, while thalassaemic women have similar prevalence of high triglyceride as compared to normal individuals. The ATP III suggests a cut off point of 150 mg/dl for defining elevated triglycerides levels [<xref ref-type="bibr" rid="B11">11</xref>]. It seems that patients with beta thalassaemia major are at low coronary risk as regards their triglycerides levels.</p><p>On the contrary, when we focused our interest on HDL cholesterol we observed that thalassaemic patients had very low values. Particularly, 42% of men and 29% of women had HDL cholesterol levels below 30 mg/dl. Comparing these findings with young adults from the ATTICA study [<xref ref-type="bibr" rid="B10">10</xref>] we observed that the rates of low HDL cholesterol levels among our patients were substantially higher. Studies suggest that even for those with normal levels of total cholesterol, risk for myocardial infarction is high when HDL cholesterol is low [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. The later may highlight the importance of total-to-HDL cholesterol ratio for the evaluation of blood lipids and the prevention of atherosclerotic disease. It has also been reported that the total cholesterol-to-HDL cholesterol ratio predicts coronary heart disease risk regardless of the absolute LDL- and HDL-cholesterol [<xref ref-type="bibr" rid="B12">12</xref>]. We observed that 39% of men and 30% of women patients who had total cholesterol levels below 200 mg/dl had HDL-cholesterol lower than 35 and 45 mg/dl, respectively. Moreover, the average total-to-HDL cholesterol ratios in men and women patients were above the threshold indicated by the ATP III guidelines (i.e. 3.5) for high-risk people [<xref ref-type="bibr" rid="B11">11</xref>]. If we compare the previous rates with the ones presented by the ATTICA study (i.e. only 19% of men and 12% of women who had desirable total cholesterol levels had low HDL cholesterol levels), we could suggest that thalassaemic patients are at much higher coronary risk than their matched controls, because of the low HDL cholesterol production, even if they are within normal values of total cholesterol. Bersot et al [<xref ref-type="bibr" rid="B12">12</xref>] suggested that in populations at risk for coronary heart disease caused by low HDL cholesterol, qualification of subjects for treatment based on the total – to – HDL cholesterol ratio thresholds (i.e. 3.5) identifies more high-risk subjects for treatment than other cholesterol threshold values alone.</p><p>It has been reported that the distribution of lipoprotein – (a) levels in the general population are different from the bell-shaped curve of serum cholesterol [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Figure <xref ref-type="fig" rid="F1">1</xref> expands the previous suggestion among patients with beta thalassaemia major. Moreover, the mean lipoprotein – (a) values in our sample were approximately 6 mg/dl among men and 8 mg/dl among women. These figures were much lower than those reported by the ATTICA study investigators in healthy people of the same age [<xref ref-type="bibr" rid="B10">10</xref>]. However, none of our patients had lipoprotein – (a) levels higher than 30 mg/dl, while on the contrary, approximately, 10% of men and women from the ATTICA study had high lipoprotein – (a) levels. The gender difference that observed in the general Greek population seems to hold in our patients, too.</p><p>Papanastasiou et al [<xref ref-type="bibr" rid="B15">15</xref>] studied a total of 104 patients with major and minor beta thalassaemia and compared them with 112 healthy controls. The investigators reported that total cholesterol; HDL and LDL-cholesterol was significantly decreased, while triglycerides were significantly increased in the thalassaemic patients compared to the control subjects. They also found a positive correlation between age and triglycerides levels. By the exception of triglycerides we also observed similar results regarding blood lipids levels among our patients and the healthy controls from the general population of Greece (based on the ATTICA study). In accordance to these findings Maioli et al [<xref ref-type="bibr" rid="B16">16</xref>] studying 70 individuals with beta thalassaemia major from Italy found that these patients disclosed significantly lower total-cholesterol, LDL-cholesterol, HDL-cholesterol, apolipoprotein A1, and B plasma levels and higher triglyceride concentration controls people. It appears, therefore, that many factors such as iron overload, liver injury, and hormonal disturbances affects lipids pattern among patients with major form of beta-thalassaemia. Moreover, Maioli et al [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>] in previous reports suggested that accelerated erythropoiesis and increased uptake of LDL by macrophages and histiocytes of the reticuloendothelial system are the main determinants of low plasma cholesterol levels in beta thalassaemia major. In addition, Giardini et al [<xref ref-type="bibr" rid="B19">19</xref>] observed that total serum phospholipids, their fractions and cholesterol were significantly lower among patients with thalassaemia major. These changes were referred to hepatic damage and to severe anaemia, respectively. Furthermore, some serum lipid polyunsaturated fatty acids were significantly decreased among patients with beta thalassaemia major as compared to normal controls. Since these alterations are a sign of lipid oxidation, the causes of this phenomenon are discussed. These differences on blood lipids and lipoprotein levels could also attribute to the adherence of a healthier lifestyle by people with beta thalassaemia, which could include consumption of healthy foods since childhood. However, it should be mentioned that oppose results from the previous reports have been reported in a recent study [<xref ref-type="bibr" rid="B20">20</xref>] which suggested that adolescents with beta-thalassaemia minor have significantly lower cholesterol levels than patients with beta thalassaemia major. The investigators suggested that this has been related to their disorder and not influenced by age, sex, hemoglobin, or feritine levels. In these patients, needless investigations for hypolipidemia should be avoided.</p><p>At this point it should be noted that the extrapolation of our findings into other populations with beta thalassaemia major may be under scrutiny, since thalassaemia is genetically oriented and various expressions of the related polymorphisms may be involved in the distribution of blood lipids and lipoprotein levels.</p></sec></sec><sec><title>Conclusion</title><p>The present study revealed that men and women with beta thalassaemia major have their blood lipid and lipoprotein levels within the normal range, and lower than the healthy individuals of the same age and population. An exception is the observed very low HDL cholesterol levels, which may underline the importance of total-to-HDL cholesterol ratio as a prognostic factor for future cardiac events in this high-risk population.</p></sec><sec><title>Authors' contributions</title><p>CC: design of the study, drafted the manuscript, DP: design of the study, data analysis and interpretation of the findings, CP, JB: drafted the manuscript, CK: design of the study, MK, JL: field investigator, drafted the manuscript</p><p>All authors read and approved the final manuscript.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec>
Bronchial hyperreactivity and spirometric impairment in polysensitized patients with allergic rhinitis	<sec><title>Background</title><p>We previously demonstrated in a group of patients with perennial allergic rhinitis alone impairment of spirometric parameters and high percentage of subjects with bronchial hyperreactivity (BHR). The present study aimed at evaluating a group of polysensitized subjects suffering from allergic rhinitis alone to investigate the presence of spirometric impairment and BHR during the pollen season.</p></sec><sec sec-type="methods"><title>Methods</title><p>One hundred rhinitics sensitized both to pollen and perennial allergens were evaluated during the pollen season. Spirometry and methacholine bronchial challenge were performed.</p></sec><sec><title>Results</title><p>Six rhinitics showed impaired values of FEV1 without referred symptoms of asthma. FEF 25–75 values were impaired in 28 rhinitics. Sixty-six patients showed positive methacholine bronchial challenge. FEF 25–75 values were impaired only in BHR positive patients (p < 0.001). A significant difference was observed both for FEV1 (p < 0.05) and FEF 25–75 (p < 0.001) considering BHR severity.</p></sec><sec><title>Conclusions</title><p>This study evidences that an impairment of spirometric parameters may be observed in polysensitized patients with allergic rhinitis alone during the pollen season. A high percentage of these patients had BHR. A close relationship between upper and lower airways is confirmed.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Ciprandi</surname><given-names>Giorgio</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Cirillo</surname><given-names>Ignazio</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Tosca</surname><given-names>Maria A</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Vizzaccaro</surname><given-names>Andrea</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	Clinical and molecular allergy : CMA	<sec><title>Background</title><p>Close association between allergic rhinitis and asthma has been demonstrated by several studies [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. Moreover, allergic rhinitis has been demonstrated to be a strong risk factor for the onset of asthma in adults [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Asthma is characterized by a reversible airflow obstruction and forced expiratory volume/1 second (FEV1) is considered the main parameter to evaluate bronchial obstruction [<xref ref-type="bibr" rid="B5">5</xref>]. Nevertheless, there is increasing interest to consider the involvement of small airways in the pathogenesis of asthma [<xref ref-type="bibr" rid="B6">6</xref>]. Even though there is no direct parameter cap able of assessing small airways, it has been assumed that the forced expiratory flow at the 25 and 75% of the pulmonary volume (FEF 25–75) might be considered as a measure of the caliber concerning distal airways [<xref ref-type="bibr" rid="B7">7</xref>]. Particularly, subjects with mild asthma and normal FEV1 may show impaired FEF 25–75 only [<xref ref-type="bibr" rid="B8">8</xref>]. On the other hand, bronchial hyperreactivity (BHR) is a paramount feature of asthma. Moreover, BHR may be observed in a proportion of rhinitics [<xref ref-type="bibr" rid="B9">9</xref>]. In this regard, it has been hypothesized that a positive bronchial challenge to methacholine could be considered as predictive for those rhinitics would progress to develop asthma [<xref ref-type="bibr" rid="B10">10</xref>]. In addition, a seasonal variability in BHR was described in subjects sensitized to pollens [<xref ref-type="bibr" rid="B11">11</xref>]. Very recently, we demonstrated that patients with perennial allergic rhinitis alone frequently showed impaired spirometric parameters and positive methacholine challenge [<xref ref-type="bibr" rid="B12">12</xref>].</p><p>On the basis of these considerations, we aimed at evaluating a group of polysensitized patients with allergic rhinitis alone to investigate the presence of spirometric abnormalities and BHR during the pollen season.</p></sec><sec sec-type="materials\|methods"><title>Materials and methods</title><sec><title>Study design</title><p>The study was performed during the pollen season (when patients were symptomatic), from April to May. To evaluate spirometric abnormalities and the presence of BHR in patients with pure rhinitis, we included subjects with allergic rhinitis due both to pollen and perennial allergens. We excluded all the subjects who met the following exclusion criteria: asthma symptoms, including cough, wheezing, dyspnoea, chest tightness, and shortness of breathing, acute upper respiratory infections and use of nasal or oral corticosteroids, and antihistamines within the previous 4 weeks.</p><p>The study was approved by the Institutional Review Board of Navy Hospital, an informed consent was obtained from patients, and was in compliance with the Helsinki Declaration.</p></sec><sec><title>Subjects</title><p>One hundred rhinitic patients were prospectively and consecutively evaluated, all males, age 23.4 ± 3.8 years. All of them were Navy soldiers who referred to Navy Hospital for periodic fitness visit. All of them were evaluated performing both spirometry and methacholine bronchial challenge during the pollen season, i.e. in the spring, season with pollens in our geographic area [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>A detailed clinical history and a complete physical examination, including allergy evaluation, were performed. The patients were included in the study on the basis of a clinical history of allergic rhinitis. All patients were sensitized both to pollens (i.e. <italic>Parietaria officinalis</italic>, grasses, olive tree, birch, or hazel) and perennial allergens (i.e. house dust mites, cat, or dog). The diagnosis of allergic rhinitis was made on the basis of a history of nasal symptoms and positive skin prick test as described elsewhere [<xref ref-type="bibr" rid="B3">3</xref>]. None of the patients was a previous or a current smoker.</p></sec><sec><title>Skin prick test</title><p>it was performed as stated by the Italian Society of Allergy and Clinical Immunology [<xref ref-type="bibr" rid="B13">13</xref>]. The panel consisted of: house dust mites (<italic>Dermatophagoides farinae </italic>and <italic>pteronyssinus</italic>), cat, dog, grasses mix, <italic>Compositae </italic>mix, <italic>Parietaria officinalis</italic>, birch, hazel, olive tree, <italic>Alternaria Tenuis</italic>, <italic>Cladosporium</italic>, <italic>Aspergilli </italic>mix (Stallergenes, Milan, Italy).</p></sec><sec><title>Spirometry</title><p>It was performed by using a computer-assisted spirometer (Pulmolab 435-Spiro 235, Morgan, England), with optoelectronic whirl flow meter. Spirometry is performed as stated by European respiratory Society [<xref ref-type="bibr" rid="B14">14</xref>], using the European Community for Steel and Coal reference equations.</p><p>If an airway obstruction was present as detected by FEV1 values less than 80% of the predicted, a test of bronchodilatation was performed using a salbutamol metered dose of 200 mcg. Reversibility was considered if an increase of at least 12% of FEV1 from baseline was achieved, according to international guidelines [<xref ref-type="bibr" rid="B15">15</xref>].</p></sec><sec><title>Methacholine bronchial challenge</title><p>It was performed to evaluate BHR only if basal FEV1 was equal or more than 80% of predicted. Aerosol is delivered using a dosimetric computerized supply (MEFAR MB3, Marcos, Italy). Subjects inhaled increasing doses of methacholine, starting from 34 μg/mL. The scheduled doses consisted of the following: 34, 68, 68, 68, 170, 170, 340, 170, 340, 170 μg/mL as previously reported [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B12">12</xref>].</p><p>The test was interrupted when FEV1 value was reduced by more or equal than 20% of control or a maximal cumulative dose of 1,598 μg/ml was achieved. The threshold dose causing a 20% fall of FEV1 (PD20) was calculated.</p></sec><sec><title>Degree of BHR</title><p>Four arbitrary classes of BHR were considered: very mild = PD20 > 400 μg/mL, mild = PD20 ranging from 201 to 400 μg/mL, moderate = PD20 ranging from 200 to 101 μg/mL, and severe = PD20 < 100 μg/mL as previously reported [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B16">16</xref>].</p></sec><sec><title>Statistical analysis</title><p>Statistical analysis was performed using X square test, calculating confidential limits of the relative risk at 95%. Differences were considered significant if p values were <0.05. Data are presented as means.</p></sec></sec><sec><title>Results</title><p>All rhinitics were consecutive subjects meeting the inclusion and exclusion criteria and agreeing to join the study.</p><p>No adverse event was reported during the study.</p><sec><title>Sensitizations</title><p>all subjects were sensitized both to perennial allergens and pollen allergens. Twenty subjects had 2 sensitizations, 34 had 3 sensitizations, and 46 had more than 3 sensitizations. There was no relationship between number of sensitizations and spirometric data.</p></sec><sec><title>Spirometry</title><p>six patients showed a FEV1 value less than 80% of the predicted. It has to be mentioned that all of them were completely asymptomatic for complaints concerning lower airways. A bronchial reversibility was achieved in all subjects.</p><p>In addition, 7 patients showed impaired FVC values and 28 patients showed abnormal FEF 25–75 values.</p></sec><sec><title>Methacholine bronchial challenge</title><p>it was performed in 94 rhinitics. Sixty-six rhinitics showed a positive methacholine challenge. On the basis of BHR degree, we subdivided the methacholine positive patients in 4 groups: very mild, mild, moderate, and severe. Seventeen patients had a very mild degree of BHR, 16 had a mild degree, 10 had a moderate degree, and 23 a severe degree.</p><p>Then, we analyzed subjects subdividing them in two groups: patients with BHR (BHR positive group) and patients without BHR (BHR negative group). Thus, we evaluated the distribution of the patients considering FEV1, FVC, and FEF 25–75 values (Figure <xref ref-type="fig" rid="F1">1</xref>). FEV1 values were normal in both groups. Five subjects in the BHR positive group and 2 in the BHR negative group had reduced values of FVC only. FEF 25–75 values were reduced in 28 subjects of BHR positive group only (p < 0.001).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Percentage distribution of FEF 25–75 values (as % of predicted) in BHR positive and BHR negative patients.</p></caption><graphic xlink:href="1476-7961-2-3-1"/></fig><p>We considered the three spirometric parameters related with BHR degree (Figure <xref ref-type="fig" rid="F2">2</xref>). A significant difference was observed for both FEV1 and FEF 25–75 considering BHR severity in subjects with moderate BHR (p < 0.001 for FEF 25–75 only) and with severe BHR (p < 0.05 for FEV1 and p < 0.001 for FEF 25–75).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Percentage distribution of mean values of FVC, FEV1, and FEF 25–75 in comparison with BHR grade.</p></caption><graphic xlink:href="1476-7961-2-3-2"/></fig></sec><sec><title>Discussion</title><p>Allergic rhinitis and asthma should be considered as a single syndrome involving two parts of the respiratory tract, even though it is evident that these two disorders affect each other [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>Allergic rhinitics frequently present a non-specific BHR even in absence of asthmatic symptoms. In these subjects with normal FEV1 values, BHR may be envisaged as a marker of susceptibility to develop asthma. On the other hand, in mild asthmatics during intercritical periods lung function may be normal concerning FEV1 values [<xref ref-type="bibr" rid="B17">17</xref>]. Moreover, asthma is a chronic inflammatory disease of airways and using other parameters it has been demonstrated a persistence of inflammation, also in absence of symptoms, mainly involving smaller airways [<xref ref-type="bibr" rid="B18">18</xref>]. In these cases, abnormal FEF 25–75 values may be observed: it has been reported that FEF 25–75 may be reduced in asthmatics with normal FEV1 and FVC values [<xref ref-type="bibr" rid="B8">8</xref>]. It has been suggested that FEF 25–75 might be considered a marker of small airways impairment in mild asthmatics with normal FVC values [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>Very recently, we demonstrated some interesting findings in a group of 100 patients with perennial allergic rhinitis alone [<xref ref-type="bibr" rid="B12">12</xref>]. Five patients showed impaired FEV1 values (<80% of predicted), without any perceived lower respiratory symptoms [<xref ref-type="bibr" rid="B12">12</xref>]. Moreover, 72 patients showed positive methacholine challenge, and there was a significant relationship between BHR degree and FEV1 and FEF 25–75 values [<xref ref-type="bibr" rid="B12">12</xref>]. Thus, we aimed at investigating a large group of polysensitized patients with allergic rhinitis during the pollen season to evaluate spirometry and BHR.</p><p>The present findings suggest some considerations concerning the link between upper and lower airways.</p><p>Firstly, evaluating a large cohort of polysensitized subjects with allergic rhinitis alone, it is possible to single out some subjects (six) with overt bronchial obstruction, as documented by impaired FEV1 values. These patients may be considered as "poor perceiver" of their lower respiratory symptoms. In fact, all of them had a normal life playing different sports without trouble. In addition, they never felt lower respiratory symptoms nor diagnosis of asthma has been made. It is noteworthy that this finding confirms that demonstrated in perennial rhinitics (5 patients with overt bronchial obstruction).</p><p>Secondly, most of our rhinitics (66 subjects) showed BHR. This finding is not surprising if compared with literature analysis and confirm our previous findings in patients with perennial allergic rhinitis. The exposure to allergens is characterized by nasal inflammation as previously described by ourselves [<xref ref-type="bibr" rid="B19">19</xref>]. This concept may be consistent with a consequent bronchial inflammation. It is noteworthy that BHR was asymptomatic in all our rhinitics.</p><p>Thirdly, considering the evaluation of FEF 25–75 parameter we demonstrated that some rhinitics (28 subjects) shows an initial level of bronchial obstruction during the pollen season. It has to be highlighted that BHR positive patients only showed this impairment. This finding may underline the relevance of considering this parameter as it was impaired only in BHR subjects. Thus, FEF 25–75 could be envisaged as marker of bronchial involvement in pure rhinitics with BHR.</p><p>Fourthly, there is a relationship between degree of BHR and FEV1 and FEF 25–75 impairment. These last findings underline the relationship between BHR and airway caliber in patients with airway inflammation. Moreover, these data, taken together, partially confirm previous results observed in patients with perennial allergic rhinitis alone [<xref ref-type="bibr" rid="B12">12</xref>]. Polysensitized patients with allergic rhinitis, compared with patients with perennial allergic rhinitis, even more show an association with asthma, the impairment of FEF 25–75, the BHR, and the relationship between BHR grade and spirometric abnormalities. Actually, it is clear that allergic inflammation is chronic in these subjects and it is exacerbated by pollen exposure.</p></sec></sec><sec><title>Conclusions</title><p>The present study highlights the frequent coexistence of bronchial impairment in polysensitized patients with allergic rhinitis alone during the pollen season and supports the strong link between upper and lower airways. Thus, a careful evaluation of lower airways should be performed also in those patients with allergic rhinitis alone.</p></sec><sec><title>List of abbreviations</title><p>BHR: bronchial hyperreactivity</p><p>FEV1: forced expiratory volume in 1 second</p><p>FEF: forced expiratory flow</p><p>FVC: forced volume capacity</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>GC conceived of the study, and participated in its design and coordination, IC participated in the design of the study and performed the statistical analysis, MAT revised the manuscript, and AV participated in the clinical study. All authors read and approved the final manuscript.</p></sec>
To eat or not to eat: The NICE way	Could not extract abstract	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Kovvali</surname><given-names>Gopala</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Journal of Carcinogenesis	<p>Seventh international symposium on preventive oncology and intervention strategies brought several scientists and physicians together in a French city, Nice from 7–10<sup>th </sup>February 2004.</p><p>The conference covered several contemporary topics like apoptosis, angiogenesis and several interesting and insightful talks were presented by the delegates. The keynote addresses ' Cancer predisposition', on the opening day by Dr. HT Lynch, the father of lynch syndrome, a hereditary non polyposis colorectal cancer (HNPCC) is a highlight of the symposium. There were several talks on dietary influences, screening and detection, prevention and intervention, genetic and environmental interactions and susceptibility genes among many other topics.</p><p>As much as it shed spikes of useful information in the field of preventive oncology, the meeting brought the delegates to the 'crossroads' of cancer epidemiology. A thought provoking and very interesting closing address 'Cancer Epidemiology at a Crossroads' by Dr. Eduardo Franco seems to have summarized and reflected some of the private feelings and thoughts of the practicing scientists and physicians as to what is the right answer for questions like 'what is the probability of getting cancer for an individual? Does aspirin cure and prevent colds and cancers alike? Should we eat more vegetables or beware of trace amounts of pesticides reported to be present in some of them? Are tobacco and alcohol the sole responsible factors for lung and upper aero-digestive tract cancers? It all seems to be in the 'P' value we all dearly believe in, to represent 'probability' and 'prevention'. Dr. Eduardo Franco rightly suggested, clinical epidemiology is still an evolving discipline and the best of it is yet come.</p><p>A rather interesting set of presentations on the polymorphisms of DNA repair and tumor suppressor genes seems to make his point more eloquent. Polymorphisms in codon 72 of the p53 gene seem to be a novel 'hot-spot' for the 'ever hot' tumor suppressor. While correlation or lack of it with likelihood of cancer predisposition in individuals harboring the polymorphic codon is certainly receiving attention of researchers, culpability of that lone codon in human cancers is yet to be demonstrated. Is it really relevant to cancer predisposition in humans? It may be a question of a million dollar grant proposal. May be a couple.</p><p>An interesting introduction by Dr. Horton in his talk on 'HER-2/neu expression and prognosis in breast cancer' is worth noting. While he was making a case for efficacy of certain chemotherapeutic drugs in breast cancer patients, he enumerated the limitations of commonly used Immuno Histochemistry technique. He pointed out several important aspects of the technique that could lead to subjective evaluation of the expression levels of the protein leading to ambiguous conclusions. Interestingly, during my visit to posters, a researcher sadly noted that in his hands, only 28% of patients show mutations or/and over expression of p53 in cancers he is investigating while his competitors reported 83% of such cases. HPV and cervical cancers seems to be another topic of less concordant and more diffuse results.</p><p>What is the source of these variations? Demographic differences among samples, sample handling, variations in the technique used and the technician who performed the technique, experimental bias and inappropriate design of experiment or a combination of any of these? It is interesting to note that there is no uniformity in the experimental procedures and standardized and certified laboratory practices in biomedical research laboratories. There are no regulatory bodies to oversee compliance to best laboratory practices. We often hear the expression like 'in our hands, we observe'. I wonder if it is prudent to design, evolve and follow standardized laboratory techniques and practices. It might save a lot of time and resources and accelerate the process of biomedical research. It may result in conclusions that are meaningful and useful. More importantly, it may enhance the credibility of biomedical research among the general public.</p><p>On a lighter note, the welcome reception made me believe that a limited serving of fruits, peanuts, chips, cheese and wine is the right recipe for a healthy diet. But the reception on the concluding day seems to suggest that a handful of peanuts and a glass of red wine may be a sufficient source of antioxidants and polyunsaturated fatty acids in healthy diet.</p><p>There were several presentations that showed promising data on the use of soy and tomato among other health promoting foods in preventing carcinogenesis and cancer. Yet, there seems to be no NICE way to eat healthy.</p>
A comparison of the Nottingham Health Profile and Short Form 36 Health Survey in patients with chronic lower limb ischaemia in a longitudinal perspective	<sec><title>Background</title><p>Different generic quality of life instruments such as the Nottingham Health Profile (NHP) and the Short Form 36 Health Survey (SF-36) have revealed conflicting results in patients with chronic lower limb ischaemia in psychometric attributes in short-term evaluations. The aim of this study was to compare the NHP and the SF-36 regarding internal consistency reliability, validity, responsiveness and suitability as outcome measures in patients with lower limb ischaemia in a longitudinal perspective.</p></sec><sec sec-type="methods"><title>Methods</title><p>48 patients with intermittent claudication and 42 with critical ischaemia were included. Assessment was made before and one year after revascularization using comparable domains of the NHP and the SF-36 questionnaires.</p></sec><sec><title>Results</title><p>The SF-36 was less skewed and more homogeneous than the NHP. There was an average convergent validity in three of the five comparable domains one year postoperatively. The SF-36 showed a higher internal consistency except for social functioning one-year postoperatively and was more responsive in detecting changes over time in patients with intermittent claudication. The NHP was more sensitive in discriminating among levels of ischaemia regarding pain and more able to detect changes in the critical ischaemia group.</p></sec><sec><title>Conclusion</title><p>Both SF-36 and NHP have acceptable degrees of reliability for group-level comparisons, convergent and construct validity one year postoperatively. Nevertheless, the SF-36 has superior psychometric properties and was more suitable in patients with intermittent claudication. The NHP however, discriminated better among severity of ischaemia and was more responsive in patients with critical ischaemia.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Wann-Hansson</surname><given-names>Christine</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Hallberg</surname><given-names>Ingalill Rahm</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Risberg</surname><given-names>Bo</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Klevsgård</surname><given-names>Rosemarie</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Health and Quality of Life Outcomes	<sec><title>Background</title><p>During the past few decades quality of life assessment has become one central outcome in treatment of patients with chronic lower limb ischaemia. Different generic quality of life instruments such as the Nottingham Health Profile (NHP) and the Short Form 36 Health Survey (SF-36) [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>] have previously been tested, revealing conflicting results in these patients according to psychometric attributes in short-term evaluations. The strengths and weakness of the NHP and the SF-36 scales are not extensively examined and further research is needed to establish which is the more appropriate and responsive quality of life instrument for patients with chronic lower limb ischaemia in the long term. The main goal of vascular surgical treatment is the relief of symptoms and improvement in patients quality of life. A majority of the patients are elderly and have generally widespread arterial disease with numbers of symptoms due to the chronic lower limb ischaemia, which may affect the patients' quality of life [<xref ref-type="bibr" rid="B3">3</xref>-<xref ref-type="bibr" rid="B5">5</xref>]. Intermittent claudication (IC) means leg pain constantly produced by walking or muscular activity and is relieved by rest, while critical leg ischaemia (CLI) means pain even at rest and problems with non-healing ulcers or gangrene [<xref ref-type="bibr" rid="B6">6</xref>]. It is important to identify dimensions which are influenced by the severity and nature of the disease when selecting a suitable quality of life instrument [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>The World Health Organization QOL group [<xref ref-type="bibr" rid="B8">8</xref>] has identified and recommended five broad dimensions – physical and psychological health, social relationship perceptions, function and well-being – which should be included in a generic quality of life instrument. Generic instruments cover a broad range of dimensions and allow comparisons between different groups of patients. Disease-specific instruments, on the other hand, are specially designed for a particular disease, patient group or areas of function [<xref ref-type="bibr" rid="B9">9</xref>]. The functional scale, Walking Impairment Questionnaire (WIQ) [<xref ref-type="bibr" rid="B10">10</xref>] and quality of life instruments such as Intermittent Claudication Questionnaire (ICQ) [<xref ref-type="bibr" rid="B11">11</xref>] and Claudication Scale (CLAU-S) [<xref ref-type="bibr" rid="B12">12</xref>] are examples of disease-specific instruments which have been developed in recent years for patients with IC. However, at present there is no accepted disease-specific questionnaire for quality of life assessment in patients with CLI. Nevertheless, the TransAtlantic Inter-Society Consensus (TASC) [<xref ref-type="bibr" rid="B6">6</xref>] recommended that quality of life instruments should be used in all clinical trials and preferably include both generic and disease-specific quality of life measures.</p><p>Outcome measures need to satisfy different criteria to be useful as a suitable health outcome instrument in clinical practice. Construct validity is one of the most important characteristics and is a lengthy and ongoing process [<xref ref-type="bibr" rid="B13">13</xref>]. An essential consideration is the instrument's ability to discriminate between different levels of the disease; another consideration is its reliability, which means the degree to which the instrument is free from random error and all items measure the same underlying attribute [<xref ref-type="bibr" rid="B14">14</xref>]. Further, the requirement for a useful outcome measure is the responsiveness in detecting small but important clinical changes of quality of life in patients following vascular interventions [<xref ref-type="bibr" rid="B13">13</xref>]. Finally the ideal quality of life instrument must also be acceptable to patients, simple and easy to use and preferably short. Comparisons among quality of life instruments and their psychometric characteristics and performance are needed to provide recommendations about their usefulness as outcome measures for these particular groups of patients.</p><p>The aim of this study was to compare two generic quality of life questionnaires, the Nottingham Health Profile (NHP) and the Short Form 36 Health Survey (SF-36) regarding the internal consistency reliability, validity, responsiveness and suitability as outcome measures in patients with lower limb ischaemia in a longitudinal perspective.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Patients</title><p>Ninety consecutive patients from a Swedish vascular unit in southern Sweden were invited to participate in this study. The assessment took place before and 12 months after revascularization. Out of 90 patients, 24 (27%) dropped out during the follow-up period, of whom 14 suffered from CLI. Six patients (7%) died, 15 (17%) did not wish to participate and 3 (3%) had other concurrent diseases. The inclusion criteria were patients admitted for active treatment of documented lower limb ischaemia, having no communication problems and having no other disease restricting their walking capacity [<xref ref-type="bibr" rid="B1">1</xref>]. The severity of ischaemia was graded according to suggested standards for grading lower limb ischaemia [<xref ref-type="bibr" rid="B15">15</xref>]. Sixty-two (68.8%) patients were treated with a surgical bypass, 24 (26.6%) had a percutaneous angioplasty (PTA) and 4 (4.6%) had a surgical thromboendatherectomy (TEA) (Table <xref ref-type="table" rid="T1">1</xref>). Routine medical history, risk factors and clinical examinations, which included ankle blood pressure (ABP) and ankle-brachial pressure index (ABPI), were obtained before and one year after revascularization in accordance with the Swedish Vascular Registry (Swedvasc) [<xref ref-type="bibr" rid="B16">16</xref>]. The questionnaire also contains questions about sex, age, housing and civil status. Demographic characteristics and clinical data were obtained from the patients' medical records.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Demographic characteristics of the patient groups before revascularization (n = 90).</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">Claudicants n = 48</td><td align="center">Critical ischaemia n = 42</td><td align="center"><italic>P</italic>-<italic>value</italic></td></tr></thead><tbody><tr><td align="left"><bold>Age m (SD)<sup>1</sup></bold></td><td align="right">67 (10.2)</td><td align="right">71 (10.1)</td><td align="right">.05</td></tr><tr><td align="left"><bold>Sex %<sup>2</sup></bold></td><td></td><td></td><td align="right">.52</td></tr><tr><td align="left">Male/female</td><td align="right">54/46</td><td align="right">52/48</td><td></td></tr><tr><td align="left"><bold>Cohabitation n (%)<sup>2</sup></bold></td><td></td><td></td><td align="right">.33</td></tr><tr><td align="left">Living alone</td><td align="right">19 (39.6)</td><td align="right">11 (26.2)</td><td></td></tr><tr><td align="left">Living with family/relatives</td><td align="right">29 (60.4)</td><td align="right">31 (73.8)</td><td></td></tr><tr><td align="left"><bold>Severity of disease n (%)</bold></td><td></td><td></td><td></td></tr><tr><td align="left">Intermittent claudication</td><td align="right">48 (100)</td><td></td><td></td></tr><tr><td align="left">Ischaemia rest pain</td><td></td><td align="right">22 (52.2)</td><td></td></tr><tr><td align="left">Ischaemia ulcers</td><td></td><td align="right">17 (40.4)</td><td></td></tr><tr><td align="left">Ischaemia gangrene</td><td></td><td align="right">3 (7.4)</td><td></td></tr><tr><td align="left"><bold>Type of intervention n (%)<sup>2</sup></bold></td><td></td><td></td><td align="right">.02</td></tr><tr><td align="left">Angioplasty/STENT</td><td align="right">18 (37.5)</td><td align="right">6 (14.3)</td><td></td></tr><tr><td align="left">Bypass</td><td align="right">28 (58.3)</td><td align="right">34 (81.0)</td><td></td></tr><tr><td align="left">Thromboendatherectomy</td><td align="right">2 (4.2)</td><td align="right">2 (4.8)</td><td></td></tr><tr><td align="left"><bold>Level of disease n (%)</bold></td><td></td><td></td><td></td></tr><tr><td align="left">Iliac</td><td align="right">22 (45.8)</td><td align="right">13 (31.0)</td><td></td></tr><tr><td align="left">Femoral (above knee)</td><td align="right">20 (41.7)</td><td align="right">10 (23.8)</td><td></td></tr><tr><td align="left">Distal (below knee)</td><td align="right">6 (12.5)</td><td align="right">19 (45.2)</td><td></td></tr><tr><td align="left"><bold>Leg side of disease n (%)</bold></td><td></td><td></td><td></td></tr><tr><td align="left">Unilateral</td><td align="right">45 (93.7)</td><td align="right">34 (81.0)</td><td></td></tr><tr><td align="left">Bilateral</td><td align="right">3 (6.3)</td><td align="right">8 (19.0)</td><td></td></tr><tr><td align="left"><bold>Risk factors n (%)<sup>2</sup></bold></td><td></td><td></td><td></td></tr><tr><td align="left">Smoking</td><td align="right">11 (22.9)</td><td align="right">11 (26.2)</td><td align="right">.45</td></tr><tr><td align="left">Hypertension</td><td align="right">14 (29.2)</td><td align="right">14 (33.3)</td><td align="right">.42</td></tr><tr><td align="left">Heart disease</td><td align="right">9 (18.8)</td><td align="right">16 (38.1)</td><td align="right">.03</td></tr><tr><td align="left">Diabetics</td><td align="right">4 (10.4)</td><td align="right">11 (26.2)</td><td align="right">.05</td></tr><tr><td align="left">Hyperlipaemia</td><td align="right">3 (6.3)</td><td align="right">1 (2.4)</td><td align="right">.36</td></tr><tr><td align="left">Stroke/TIA</td><td align="right">4 (8.3)</td><td align="right">2 (4.8)</td><td align="right">.40</td></tr><tr><td align="left">Chronic lung disease</td><td align="right">1 (2.1)</td><td align="right">3 (7.1)</td><td align="right">.38</td></tr><tr><td align="left">Kidney disease, kreat >150</td><td align="right">2 (4.2)</td><td align="right">2 (4.8)</td><td align="right">.64</td></tr><tr><td align="left"><bold>Reoperations during follow-up n (%)</bold></td><td align="right">4 (6.1)</td><td align="right">3 (4.5)</td><td align="right">.64</td></tr></tbody></table><table-wrap-foot><p><sup>1</sup>Mann-Whitney <italic>U</italic>-test <sup>2</sup>Chi-square test Include the patients (n = 66) who completed the study one year postoperatively. <italic>P-value = <0.05</italic></p></table-wrap-foot></table-wrap></sec><sec><title>Nottingham Health Profile</title><p>The Nottingham Health Profile (NHP) was developed to be used in epidemiological studies of health and disease [<xref ref-type="bibr" rid="B17">17</xref>]. It consists of two parts. Part I contains 38 yes/no items in 6 dimensions: pain, physical mobility, emotional reactions, energy, social isolation and sleep. Part II contains 7 general yes/no questions concerning daily living problems. The two parts may be used independently and part II is not analysed in this study. Part I is scored using weighted values which give a range of possible scores from zero (no problems at all) to 100 (presence of all problems within a dimension). Swedish weights have been developed and used in this study [<xref ref-type="bibr" rid="B18">18</xref>]. The Swedish version has proved to be valid and reliable, for example, in patients with arthrosis of the hip joint [<xref ref-type="bibr" rid="B19">19</xref>] and in patients suffering from grave ventricular arrhythmias [<xref ref-type="bibr" rid="B20">20</xref>]. The NHP scale has also proved capable of measuring changes in perceived health following different treatments such as radical surgery for colorectal cancer [<xref ref-type="bibr" rid="B21">21</xref>] and after vascular interventions in lower limb ischaemia patients [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title>Short Form 36 Health Survey</title><p>The Short Form 36 Health Survey (SF-36) was developed by Ware et al [<xref ref-type="bibr" rid="B24">24</xref>] and designed to provide assessments involving generic health concepts that are not specific to age, disease or treatment group. It includes 36 items covering eight health concepts: bodily pain, physical functioning, role limitations due to physical problems, mental health, vitality, social functioning, role limitations due to emotional problems and general health. The response format is yes or no or in a three-to-six response scale. For each health concept questions scores are coded, summed and transformed on a scale from zero (worst health) to 100 (best health). In this study, the standard Swedish version was used [<xref ref-type="bibr" rid="B25">25</xref>]. The SF-36 has shown acceptable validity and reliability in population studies [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B27">27</xref>] and in various groups of patients, for example after stroke [<xref ref-type="bibr" rid="B28">28</xref>] and in patients with rheumatoid arthritis [<xref ref-type="bibr" rid="B29">29</xref>]. The SF-36 scale has also shown responsiveness to changes in health status over time in patients with critical ischaemia [<xref ref-type="bibr" rid="B30">30</xref>-<xref ref-type="bibr" rid="B32">32</xref>] and in patients with intermittent claudication following a revascularization [<xref ref-type="bibr" rid="B33">33</xref>-<xref ref-type="bibr" rid="B35">35</xref>].</p></sec><sec><title>Procedure</title><p>The patients were asked by the head nurse to fill out the NHP and the SF-36 questionnaire during their admission before treatment. At the one-year follow-up, the questionnaire was sent home to the patients with a covering letter and a prepaid envelope. The Ethics Committee of Lund University approved the study (LU 470-98, Gbg M 098-98).</p></sec><sec><title>Statistical analysis</title><p>Differences in characteristics between patients with IC and with CLI before revascularization were analysed using Chi-squared test and Mann-Whitney <italic>U</italic>-test. The prevalence of the lowest ("floor" effect) and highest ("ceiling" effect) possible quality of life score in NHP and SF-36 was also calculated.</p><p>Construct validity was evaluated for aspects of convergent and discriminant validity by the Multitrait-Multimethod matrix (MTMM) [<xref ref-type="bibr" rid="B13">13</xref>] based on five comparable domains, including pain, physical mobility, emotional reactions, energy and social isolation for the NHP and bodily pain, physical functioning, mental health, vitality and social functioning for the SF-36 (Table <xref ref-type="table" rid="T2">2</xref>). Further, the Mann-Whitney <italic>U</italic>-test was used to examine the relative ability of the two instruments to discriminate among the degrees of severity of the peripheral vascular disease. Spearman's rank correlation coefficient was used to express the correlation between quality of life scores, ABPI, type of intervention and age. The internal consistency based on correlations between items for each scale was assessed with Cronbach's alpha [<xref ref-type="bibr" rid="B36">36</xref>]. The recommended reliability standard for group-level comparisons is a reliability coefficient of 0.70, while comparisons between individuals demands a reliability coefficient of 0.90 [<xref ref-type="bibr" rid="B25">25</xref>].</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Comparable domains between the Nottingham Health Profile (NHP) and the Short Form 36 (SF-36) and number of items.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Domains</bold></td><td align="left"><bold>Nottingham Health Profile</bold></td><td align="left"><bold>Short Form 36</bold></td></tr></thead><tbody><tr><td align="left">Pain</td><td align="left">Pain (8 items)</td><td align="left">Bodily pain (2 items)</td></tr><tr><td align="left">Physical activity</td><td align="left">Physical mobility (8 items)</td><td align="left">Physical functioning (10 items)</td></tr><tr><td align="left">Psychological status</td><td align="left">Emotional reactions (9 items) Energy (3 items)</td><td align="left">Mental health (5 items) Vitality (4 items)</td></tr><tr><td align="left">Social activity</td><td align="left">Social isolation (5 items)</td><td align="left">Social functioning (2 items)</td></tr><tr><td align="left">Other</td><td align="left">Sleep</td><td align="left">General Health Physical role Emotional role</td></tr></tbody></table></table-wrap><p>The Wilcoxon Signed Ranks test was used to detect the responsiveness of within-subjects changes over time, before and one year after revascularization, in patients with IC and CLI. Data analysis was performed for overall comparisons using the statistical package SPSS 11.0 and a <italic>P </italic>value of <.05 was taken as statistically significant.</p></sec></sec><sec><title>Results</title><p>Forty-eight (53.3%) patients had IC of whom 26 (54%) were men. The remaining 42 (46.7%) suffered from CLI and 22 (52%) of them were men. There was a significant difference in age between the two groups with a mean age of 67 and 71 respectively (Table <xref ref-type="table" rid="T1">1</xref>). One year postoperatively, sixty-six (73%) patients (38 with IC and 28 with CLI) remained in the study and secondary reconstructions were made on 7 (10%) patients during the follow-up. There were no significant differences at baseline in sex, age, ABPI and quality of life scores between the drop-out patients and the patients who completed the study. Further, there were no significant differences between the drop-outs and the remaining patients regarding the method of treatment or severity of ischaemia.</p><p>Analyses between the comparable domains showed that the NHP scores were more skewed than the SF-36 scores, especially in emotional reactions, energy and social isolation (Figure <xref ref-type="fig" rid="F1">1</xref>). The "floor effect", the proportion of individuals having the lowest possible scores (SF-36 = 0, NHP = 100), was larger for the NHP scale in energy one year (19.7%) after revascularization than for the SF-36. The "ceiling effect", the proportion of individuals having the best possible scores (SF-36 = 100, NHP = 0), was also larger for the NHP scale in emotional reactions (50.0%), energy (42.4%) and social isolation (71.2%) one year after revascularization (Table <xref ref-type="table" rid="T3">3</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Frequency distribution of scores on the NHP (left side) and comparable dimensions on the SF-36 (right side) one year after revascularization. NHP scores had 100 subtracted for consistency with SF-36</p></caption><graphic xlink:href="1477-7525-2-9-1"/></fig><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Reliability and "floor" and "ceiling" effects in comparable NHP and SF-36 scales before and one year after revascularization.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>NHP<sup>a </sup>Before N = 90</bold></td><td align="left">% floor</td><td align="left">% ceiling</td><td align="left">Alpha<sup>b </sup>values</td><td align="left"><bold>SF-36 Before N = 90</bold></td><td align="left">% floor</td><td align="left">% ceiling</td><td align="left">Alpha<sup>b </sup>values</td></tr></thead><tbody><tr><td align="left">Pain</td><td align="left">5.6</td><td align="left">3.3</td><td align="left">0.71</td><td align="left">Bodily pain</td><td align="left">5.6</td><td align="left">0</td><td align="left">0.57</td></tr><tr><td align="left">Physical mobility</td><td align="left">0</td><td align="left">1.1</td><td align="left">0.69</td><td align="left">Physical functioning</td><td align="left">2.2</td><td align="left">0</td><td align="left">0.82</td></tr><tr><td align="left">Emotional reactions</td><td align="left">0</td><td align="left">36.7</td><td align="left">0.76</td><td align="left">Mental health</td><td align="left">1.1</td><td align="left">3.3</td><td align="left">0.76</td></tr><tr><td align="left">Energy</td><td align="left">27.8</td><td align="left">29.9</td><td align="left">0.71</td><td align="left">Vitality</td><td align="left">4.4</td><td align="left">0</td><td align="left">0.70</td></tr><tr><td align="left">Social isolation</td><td align="left">0</td><td align="left">65.6</td><td align="left">0.34</td><td align="left">Social functioning</td><td align="left">0</td><td align="left">17.8</td><td align="left">0.56</td></tr><tr><td colspan="8"><hr></hr></td></tr><tr><td align="left"><bold>NHP<sup>a </sup>One year n = 66</bold></td><td align="left">% floor</td><td align="left">% ceiling</td><td align="left">Alpha<sup>b </sup>values</td><td align="left"><bold>SF-36 One year n = 66</bold></td><td align="left">% floor</td><td align="left">% ceiling</td><td align="left">Alpha<sup>b </sup>values</td></tr><tr><td colspan="8"><hr></hr></td></tr><tr><td align="left">Pain</td><td align="left">1.5</td><td align="left">13.6</td><td align="left">0.76</td><td align="left">Bodily pain</td><td align="left">1.5</td><td align="left">0</td><td align="left">0.86</td></tr><tr><td align="left">Physical mobility</td><td align="left">1.5</td><td align="left">13.6</td><td align="left">0.76</td><td align="left">Physical functioning</td><td align="left">1.5</td><td align="left">0</td><td align="left">0.87</td></tr><tr><td align="left">Emotional reactions</td><td align="left">1.5</td><td align="left">50.0</td><td align="left">0.84</td><td align="left">Mental health</td><td align="left">0</td><td align="left">7.6</td><td align="left">0.87</td></tr><tr><td align="left">Energy</td><td align="left">19.7</td><td align="left">42.4</td><td align="left">0.73</td><td align="left">Vitality</td><td align="left">4.5</td><td align="left">1.5</td><td align="left">0.83</td></tr><tr><td align="left">Social isolation</td><td align="left">3.0</td><td align="left">71.2</td><td align="left">0.73</td><td align="left">Social functioning</td><td align="left">1.5</td><td align="left">34.8</td><td align="left">0.64</td></tr></tbody></table><table-wrap-foot><p><sup>a </sup>The NHP scores are reversed for consistency with the SF-36. <sup>b </sup>Cronbach's α</p></table-wrap-foot></table-wrap><sec><title>Validity</title><p>The average convergent validity coefficients exceeded 0.5 one year postoperatively except for physical mobility and physical functioning (<italic>r </italic>= -0.46) and for social isolation and social functioning (<italic>r </italic>= -0.32), indicating a considerable convergence of the SF-36 and NHP (Table <xref ref-type="table" rid="T4">4</xref>). One year postoperatively significant correlations between ABPI and physical functioning (<italic>r </italic>= 0.29) (SF-36), physical mobility (<italic>r </italic>= 0.42) and pain (<italic>r </italic>= 0.42) (NHP) were found. The severity of the ischaemia had a significant influence in the NHP-measured domain of pain (<italic>P </italic>< .003) and physical mobility (<italic>P </italic>< .03), indicating lower quality of life scores in patients with critical ischaemia. In the ability to discriminate between levels of ischaemia in the other comparable quality of life domains, no significant differences were found (Table <xref ref-type="table" rid="T5">5</xref>).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Multitrait-Multimethod matrix of correlation coefficient for the comparable scales of the SF-36 and NHP in patients with varying degrees of lower limb ischaemia one year after revascularization.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="left" colspan="5">SF-36</td><td align="left" colspan="5">NHP</td></tr></thead><tbody><tr><td></td><td align="left">Bodily pain</td><td align="left">Physical functioning</td><td align="left">Mental health</td><td align="left">Vitality</td><td align="left">Social functioning</td><td align="left">Pain</td><td align="left">Physical mobility</td><td align="left">Emotional reactions</td><td align="left">Energy</td><td align="left">Social isolation</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="left">SF-36</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Bodily pain</td><td align="left">-</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Physical functioning</td><td align="left">0.59</td><td align="left">-</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Mental health</td><td align="left">0.33</td><td align="left">0.43</td><td align="left">-</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Vitality</td><td align="left">0.41</td><td align="left">0.48</td><td align="left">0.81</td><td align="left">-</td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Social functioning</td><td align="left">0.35</td><td align="left">0.51</td><td align="left">0.47</td><td align="left">0.45</td><td align="left">-</td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">NHP</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Pain</td><td align="left"><bold>-0.53</bold></td><td align="left">-0.46</td><td align="left">-0.12</td><td align="left">-0.38</td><td align="left">-0.14</td><td align="left">-</td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Physical mobility</td><td align="left">-0.53</td><td align="left"><bold>-0.46</bold></td><td align="left">-0.12</td><td align="left">-0.38</td><td align="left">-0.14</td><td align="left">1.00</td><td align="left">-</td><td></td><td></td><td></td></tr><tr><td align="left">Emotional reactions</td><td align="left">-0.32</td><td align="left">-0.39</td><td align="left"><bold>-0.66</bold></td><td align="left">-0.62</td><td align="left">-0.38</td><td align="left">0.31</td><td align="left">0.31</td><td align="left">-</td><td></td><td></td></tr><tr><td align="left">Energy</td><td align="left">-0.45</td><td align="left">-0.60</td><td align="left">-0.59</td><td align="left"><bold>-0.65</bold></td><td align="left">-0.58</td><td align="left">0.43</td><td align="left">0.43</td><td align="left">0.68</td><td align="left">-</td><td></td></tr><tr><td align="left">Social isolation</td><td align="left">-0.21</td><td align="left">-0.40</td><td align="left">-047</td><td align="left">-0.40</td><td align="left"><bold>-0.32</bold></td><td align="left">0.03</td><td align="left">0.03</td><td align="left">0.47</td><td align="left">0.47</td><td align="left">-</td></tr></tbody></table><table-wrap-foot><p>Correlation coefficients for comparable domains of the two questionnaires are shown in bold type Correlation coefficients are negative because the two scales run in opposite directions.</p></table-wrap-foot></table-wrap><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>Differences in comparable domains in the NHP and the SF-36 between claudicants and patients with critical ischaemia before revascularization.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Dimension</td><td align="center">Claudicants n = 48</td><td align="center">Critical ischaemia n = 42</td><td align="center"><italic>P-value</italic></td></tr></thead><tbody><tr><td align="left" colspan="4"><bold>NHP* <italic>Md (q1,q3)</italic></bold></td></tr><tr><td align="left">Pain</td><td align="center">29.9 (27.2–48.5)</td><td align="center">54.2 (28.8–84.2)</td><td align="center">.003</td></tr><tr><td align="left">Physical mobility</td><td align="center">30.8 (10.3–50.2)</td><td align="center">47.3 (20.6–66.7)</td><td align="center">.03</td></tr><tr><td align="left">Emotional reactions</td><td align="center">11.4 (0.0–39.3)</td><td align="center">10.0 (0.0–25.2)</td><td align="center">.42</td></tr><tr><td align="left">Energy</td><td align="center">60.5 (0.0–100)</td><td align="center">47.0 (0.0–100)</td><td align="center">.65</td></tr><tr><td align="left">Social isolation</td><td align="center">0.0 (0.0–24.2)</td><td align="center">0.0 (0.0–25.1)</td><td align="center">.75</td></tr><tr><td align="left" colspan="4"><bold>SF-36*<italic>Md (q1,q3)</italic></bold></td></tr><tr><td align="left">Bodily pain</td><td align="center">36.5 (22.0–42.0)</td><td align="center">31.0 (22.0–41.0)</td><td align="center">.27</td></tr><tr><td align="left">Physical functioning</td><td align="center">25.0 (15.0–40.0)</td><td align="center">25.0 (15.0–45.0)</td><td align="center">.93</td></tr><tr><td align="left">Mental health</td><td align="center">68.0 (56.0–88.0)</td><td align="center">68.0 (51.0–77.0)</td><td align="center">.54</td></tr><tr><td align="left">Vitality</td><td align="center">45.0 (31.2–60.0)</td><td align="center">45.0 (20.0–60.0)</td><td align="center">.47</td></tr><tr><td align="left">Social functioning</td><td align="center">68.7 (50.0–87.5)</td><td align="center">62.5 (50.0–78.1)</td><td align="center">.41</td></tr></tbody></table><table-wrap-foot><p>A higher score (100) indicates more perceived problems. ** A higher score (100) indicates fewer perceived problems. <sup>a </sup>P-value, claudicants versus critical ischaemia patients before revascularization. Tested by Mann-Whitney <italic>U-</italic>test. <italic>p-value = <0.05</italic></p></table-wrap-foot></table-wrap></sec><sec><title>Internal consistency</title><p>Physical functioning (α = 0.82), mental health (α = 0.76) and vitality (α = 0.70) for the SF-36 and pain (α = 0.71), emotional reactions (α = 0.76) and energy (α = 0.71) for the NHP scale were reliable, with coefficients >0.70 before revascularization. For the SF-36, all of the comparable domains except for social functioning (α = 0.64) exceeded the Cronbach's alpha value of 0.8 at the one-year follow-up. For the NHP the internal consistency coefficient was less than 0.8 but still exceeded 0.70 (Table <xref ref-type="table" rid="T3">3</xref>).</p></sec><sec><title>Responsiveness</title><p>The NHP scale and SF-36 were not equally good at detecting within-patient changes over time. In patients with IC the SF-36 scale showed significant improvements in bodily pain (<italic>P </italic>< .01) and in physical functioning (<italic>P </italic>< .001) and for the patients with CLI there were significant improvements in bodily pain (<italic>P </italic>< .004) at the one-year follow-up (Figure <xref ref-type="fig" rid="F2">2</xref>). The NHP scale showed no significant improvements in patients with IC, while in patients with CLI, significant improvements in pain (<italic>P </italic>< .001) and physical mobility (<italic>P </italic>< .03) were found (Figure <xref ref-type="fig" rid="F3">3</xref>).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Changes in median score for the SF-36 in claudicants and critical ischaemia patients before and one year after revascularization. Tested by Wilcoxon Signed Ranks test. A higher score indicates fewer perceived problems. BP = Bodily pain, PF = Physical functioning, MH = Mental health, VT = Vitality and SF = Social functioning.</p></caption><graphic xlink:href="1477-7525-2-9-2"/></fig><fig position="float" id="F3"><label>Figure 3</label><caption><p>Changes in median score for the NHP in claudicants and critical ischaemia patients before and one year after revascularization. Tested by Wilcoxon Signed Ranks test. A higher score indicates more perceived problems. P = Pain, PM = Physical mobility, EM = Emotional reactions, E = Energy and SO = Social isolation.</p></caption><graphic xlink:href="1477-7525-2-9-3"/></fig></sec></sec><sec><title>Discussion</title><p>The result showed that the SF-36 was less skewed and more homogeneous with lower "floor" and "ceiling" effects than the NHP. A considerable convergence in three of the five comparable domains one year postoperatively indicates an average convergent validity. The SF-36 showed a higher internal consistency except for social functioning one-year postoperatively and was more responsive in detecting changes over time in the IC group. The NHP was more sensitive in discriminating among levels of ischaemia regarding pain and more able to detect changes in the CLI group.</p><p>The attrition or loss of subjects (27%) in this study could have affected the outcome. Further analysis showed that there were no significant differences in quality of life, sex, age, method of treatment and severity of disease between the attrition group and those who completed the study. The fact that the NHP and SF-36 differ in their nature and content may limit the study design. Therefore the analyses in this study focused only on the comparable domains of the two instruments, including the basic domains of physical, mental and social health identified by the WHOQOL group [<xref ref-type="bibr" rid="B7">7</xref>]. A suitable quality of life instrument for patients with chronic lower limb ischaemia should not only be valid, reliable and responsive but also simple for elderly people to understand and complete. In this study there was no difference in response rate between the two instruments and both seemed to be user-friendly and took about 5–10 minutes to complete. The findings strengthen earlier results suggesting that both scales are practical and acceptable to use in elderly patients [<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B38">38</xref>].</p><p>A generic quality of life instrument, designed for a variety of populations and measuring a comprehensive set of health concepts, is likely to have problems with "ceiling" and "floor" effects [<xref ref-type="bibr" rid="B24">24</xref>]. In this study the NHP showed higher "ceiling" effects in all dimensions than the SF-36. There were minor "floor" effects in both the NHP and SF-36, indicating the lowest possible quality of life. This is in accordance to Klevsgård et al, [<xref ref-type="bibr" rid="B1">1</xref>] who also showed higher "ceiling" effects in social isolation, emotional reactions and energy for the NHP than the SF-36. Other studies have also reported fewer "ceiling" and "floor" effects in the SF-36 than in the NHP in patients with chronic obstructive pulmonary disease [<xref ref-type="bibr" rid="B39">39</xref>] and after a myocardial infarction [<xref ref-type="bibr" rid="B37">37</xref>]. The advantage of the SF-36 may be due to its use of a Likert-type response format with a number of possible different scores and its ability to detect positive as well negative states of health, whereas the NHP items are dichotomous and state more extreme ends of ill health [<xref ref-type="bibr" rid="B39">39</xref>]. This could mean that a patient with acceptable initial scores fails to improve even if the improvement is obvious. Furthermore, a false negative response will be more likely when a patient perceives to having perfect functioning on a measure that only assesses severe dysfunction. The result confirms the importance of finding a quality of life instrument that does have a spectrum of dimensions which match the patients with chronic lower limb ischaemia related to the presence of numerous and often severe comorbid conditions.</p><p>In this study the internal consistency was not as high as desirable for any of the instruments before revascularization, but both instruments exceeded the minimum internal consistency value of 0.7, except for social functioning in the SF-36 one year postoperatively. The SF-36, however, had considerably higher α-values in all other dimensions. Several studies have previously reported that the SF-36 has higher Cronbach's α values than the NHP, but the domains in which the highest and lowest values were estimated differ [<xref ref-type="bibr" rid="B37">37</xref>-<xref ref-type="bibr" rid="B40">40</xref>]. The findings suggest that it is not only the magnitude of the correlation among items, but also the number of items in the scale that affects the internal consistency. For example, the domains of pain and social functioning in the NHP contain 8 and 5 items respectively, while bodily pain and social functioning in the SF-36 contain only 2 items. This is further strengthened by the fact that both the scales were not sensitive enough to identify significant within-patients changes in social isolation and social functioning. Another explanation could be that the patients were a more homogeneous group before treatment, with similar problems which affected the quality of life, but one year postoperatively they have become more heterogeneous and represent different states of recovery [<xref ref-type="bibr" rid="B13">13</xref>]. Both instruments meet the reliability standards for group-level application in most respects, although none of them achieved the degree of reliability that be would be desirable in individual-based assessment.</p><p>The result in this study showed significant convergent correlation coefficients between scores of the comparable dimensions except for physical activity and social activity, indicating a considerable convergence of the NHP and SF-36 scale. Prieto et al [<xref ref-type="bibr" rid="B39">39</xref>] and Meyer-Rosberg et al [<xref ref-type="bibr" rid="B40">40</xref>] demonstrated similar results with an average convergent validity. Thus the NHP and SF-36 are relatively equal in the validity and corroborate that the subscales probably reflect similar impacts of chronic lower limb ischaemia. However, social isolation in the NHP showed a higher correlation with mental health in the SF-36 and might measure more psychological aspects of social life, whilst social functioning in the SF-36 tends to assess social activities according to the higher correlation with energy in NHP. Similarly the physical functioning in the SF-36 showed a higher correlation with energy and may reflect physical activities of daily living rather than physical mobility. This suggests that the SF-36 and NHP measure different aspects of physical and social activities.</p><p>Validity in terms of the instruments' relative ability to discriminate among different levels of the ischaemia could only demonstrate that patients with CLI had significantly more problems with pain and physical mobility before treatment than patients with IC measured by the NHP. Klevsgård et al [<xref ref-type="bibr" rid="B1">1</xref>] showed similar results, that the NHP was more sensitive in discriminating deterioration in pain and physical mobility than the SF-36. In contrast, Brown et al [<xref ref-type="bibr" rid="B37">37</xref>] demonstrated that the SF-36 was more sensitive than the NHP for identifying people still troubled with angina or breathlessness after a myocardial infarction. Despite the lack of significant differences between patients with IC and patients with CLI, the NHP scale tends to be more sensitive in explaining the quality of life in this group of patients with regard to the dimension of pain and physical mobility. The important issue thus is to consider how well the measurement method explains health-related phenomena which are significant for the particular targeted disease or group of patients.</p><p>The SF-36 was the more responsive instrument in detecting changes in quality of life over time in patients with IC, including bodily pain and physical functioning one year postoperatively. However, in patients with CLI, the NHP was the more responsive instrument, with significant changes in pain and physical mobility, while the SF-36 showed a significant change only in bodily pain. Falcoz et al [<xref ref-type="bibr" rid="B38">38</xref>] also demonstrated that the SF-36 was more responsive than the NHP in detecting changes five weeks after cardiac surgery. In contrast, Klevsgård et al [<xref ref-type="bibr" rid="B1">1</xref>] showed that the NHP was more responsive in patients with chronic lower limb ischaemia one month after revascularization. The result of the present study supports the TransAtlantic Inter-Society Consensus (TASC) [<xref ref-type="bibr" rid="B12">12</xref>] recommendation that SF-36 should be used as a generic health outcome measure in patients with chronic lower limb ischaemia. It seems to be more sensitive for detecting changes in quality of life than the NHP in patients with IC. In the group of CLI patients who have more problems with mobility and pain, however, it is harder to evaluate whether the one questionnaire is superior to the other, the NHP could be a preferable instrument in this group of patients.</p></sec><sec><title>Conclusion</title><p>The findings indicate that both the SF-36 and the NHP have acceptable degrees of reliability for group-level comparisons, convergent and construct validity one year after revascularization. Nevertheless, the SF-36 seems generally to have more superior psychometric properties and was more suitable than the NHP for evaluating quality of life in patients with intermittent claudication. The NHP, however, discriminated better among severity of ischaemia and was more responsive in detecting changes over time in patients with critical leg ischaemia.</p></sec>
Cancer characterization and feature set extraction by discriminative margin clustering	<sec><title>Background</title><p>A central challenge in the molecular diagnosis and treatment of cancer is to define a set of molecular features that, taken together, distinguish a given cancer, or type of cancer, from all normal cells and tissues.</p></sec><sec><title>Results</title><p>Discriminative margin clustering is a new technique for analyzing high dimensional quantitative datasets, specially applicable to gene expression data from microarray experiments related to cancer. The goal of the analysis is find highly specialized sub-types of a tumor type which are similar in having a small combination of genes which together provide a unique molecular portrait for distinguishing the sub-type from any normal cell or tissue. Detection of the products of these genes can then, in principle, provide a basis for detection and diagnosis of a cancer, and a therapy directed specifically at the distinguishing constellation of molecular features can, in principle, provide a way to eliminate the cancer cells, while minimizing toxicity to any normal cell.</p></sec><sec><title>Conclusions</title><p>The new methodology yields highly specialized tumor subtypes which are similar in terms of potential diagnostic markers.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Munagala</surname><given-names>Kamesh</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Tibshirani</surname><given-names>Robert</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Brown</surname><given-names>Patrick O</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib>	BMC Bioinformatics	<sec><title>Background</title><p>A unique molecular portrait that distinguishes a cancer from any normal cell or tissue could be exploited in many different ways for diagnosis or treatment. For example, an experienced biologist may be able to "read" a particular set of molecular features as representing the activity of a metabolic or regulatory system that can be exploited for treatment. We wondered, however, whether in some cases it might be possible to use a more general approach, which would not necessarily rely upon a detailed understanding of the physiological implications of each molecular portrait. Suppose, for example, that, for any given gene product, we have a way to deliver a toxin to cells at a dose proportional to the level at which the gene is expressed in each cell. Indeed, for cell surface molecules, monoclonal antibodies can approximate such a delivery system. If, for any cancer, we can identify a set of molecular targets whose cumulative level of expression in each cancer cell exceeds their expression level in any normal cell by a sufficient therapeutic margin, then we could, in principle, use a combination of the corresponding molecularly targeted toxins to kill each cancer cell, while sparing the normal cells. This scenario, while highly speculative, serves to highlight the potential value of methods that can identify moderate-sized sets of discriminating features, and simultaneously classify or cluster samples (eg, cancers) based on the set of molecular features that discriminate then from, eg., normal cells.</p><p>In this paper we identify natural cancer sub-classes based on similarity of the sets of genes that discriminate them from the class of all normal tissues, from a large set of microarray data comprising quantitative measurements of the expression of thousands of genes in a diverse set of cancers in normal human tissues. This is done by running a hierarchical clustering procedure on top of a linear kernel classifier.</p><p>We first describe the linear classifier [<xref ref-type="bibr" rid="B1">1</xref>]. Assume that we have expression pro-files for samples in two groups: a normal class, and an abnormal class. The kernel of the method is the <italic>positive maximum margin classifier</italic>, illustrated in Figure <xref ref-type="fig" rid="F1">1</xref>. We find the linear combination of genes, with non-negative weights, that produces the largest margin (gap) between the normal and abnormal classes. This linear combination is depicted by the middle solid line in the figure. This line can be efficiently computed by a linear programming technique even when the number of genes is around 10,000 and the number of samples is around 500 (details of the formulation can be found in Appendix A). This class of problems are called <italic>packing </italic>linear programs, and have efficient solutions. A discussion of the methods for solving such problems can be found in [<xref ref-type="bibr" rid="B1">1</xref>], and are therefore omitted. We focus on positive margin classifiers as we are interested in genes showing larger expression value in the tumor samples. The protein products of such genes might be detectable in the blood stream, and can possibly be targeted for diagnosis and therapy. Though genes showing lower expression value in the tumor samples are potentially biologically interesting, we do not consider them in this study; our methodology, however, extends naturally to linear classifiers which can detect such genes as well.</p><p>For a given tumor sample, the output of this classifier is a weighted vector of genes whose combined expression is larger in this sample compared to <italic>all </italic>normal samples. This would serve as a discriminatory feature set for this tumor sample. Our main goal is to cluster tumor samples of a certain histological type (like CNS or renal) based on the similarity of their feature sets, to identify sub-groups for which the same feature set is discriminatory. This would produce feature sets of moderate size, which find, and effiectively characterize cancer classes with respect to the genes whose detected expression in the tumors distinguishes them from normal cells/tissues. We propose a method that combines margin classification with hierarchical clustering and convex polytope exploration techniques to find cancer classes, and moderately large feature sets of genes for each class, spanning the various gene classes that distinguish the cancer type. We call this procedure <italic>Discriminative Margin Clustering</italic>.</p><p>The input to our procedure is the set of tumor samples, (either labelled with their histological type, like CNS, renal, etc, or unlabeled samples) and the set of normal samples. We run the discriminative margin clustering procedure on the set of tumor samples versus all the normal samples (this is a binary problem, and not a multi-class problem), to obtain sub-types of tumors with similar feature sets. We run the procedure separately on each class of tumor samples (like CNS, renal, etc) to obtain clusters within that class which are similar in terms of their feature sets. The data set and results are present at <ext-link ext-link-type="uri" xlink:href="http://microarray-pubs.stanford.edu/margin clus/"/>. Our method can therefore be run either in a semi-supervised fashion with apriori class labels, or in an unsupervised fashion. We do an empirical validation of the quality of these feature sets in terms of uniqueness to a tumor class by evaluating their goodness on a test set of tumor samples, and show that the accuracy of predicting the correct tumor type is pretty high. The prediction accuracy, though high, is not as good as the accuracy of traditional hierarchical clustering, mainly because we are working with feature sets of small size. Nevertheless, we show (somewhat surprisingly) that small feature sets (which are just based on properties of a certain tumor class versus the normal class) are sufficient to obtain reasonably high prediction accuracies against other tumor classes as well.</p><sec><title>Classification</title><p>Several researchers have observed that margin classifiers work well in finding signature gene sets for a cancer class. Most previous work [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B6">6</xref>] based on Support Vector Machine classifiers has focused on feature sets (or clusters of feature sets [<xref ref-type="bibr" rid="B7">7</xref>]) that either separate a cancer class from all other cancer classes, or that separate a cancer class from the corresponding normal class, by minimizing the Euclidean norm of the feature vector. Several tumors have natural sub-types [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>] which clustering techniques can identify, but which these classification schemes do not identify (refer [<xref ref-type="bibr" rid="B10">10</xref>] for a detailed discussion of this issue, along with additional references to literature). The main difference between our work and previous work on clustering is that our clustering effort is focused on finding sub-types which are amenable to similar diagnostic/therapeutic methods, rather than finding sub-types which are histologically similar. The linear margin classifier we present has been previously studied in [<xref ref-type="bibr" rid="B11">11</xref>]. A feature of this classifier is that the discriminatory gene sets are of very small size (this is a property of linear classifiers, where the size of the feature set is provably no more than the number of tumor and normal samples.) Our work differs from previous work in focusing on <italic>positive </italic>margin classifiers, since the feature sets which result are potentially useful for the purposes of diagnosis and therapy.</p><p>Methods like Partial Least Squares [<xref ref-type="bibr" rid="B12">12</xref>] can be used to reduce the dimension of the feature space before applying our clustering procedure. It would be interesting to observe how such an approach would affect the quality of clustering and the relevance of the feature sets.</p></sec><sec><title>Clustering</title><p>A related objective is to group tumor samples with related feature sets. For this purpose, we need to <italic>cluster </italic>the tumor samples into groups so that tissues within a group have similar feature sets. Traditional methods of clustering tumors [<xref ref-type="bibr" rid="B13">13</xref>] use the notion of <italic>dot-product </italic>or <italic>Euclidean distance </italic>similarity between the gene expression vectors as the notion of similarity. Clustering based on overall gene expression patterns works very well in grouping together histologically similar tumor samples. However, a problem with this approach is that a disproportionate amount of weight is given to a set of genes which form part of the same broad biological function, like respiration or the cell cycle. The individual genes in these functionally themed groups need not be not <italic>discriminatory </italic>in the sense we want them to be. It is therefore not geared towards recognizing samples with similar feature sets.</p></sec><sec><title>Statistical methods</title><p>An alternative method to finding feature sets is to list the genes that are relatively highly expressed in the tumor class as compared to <italic>most </italic>normal tissue samples, using some standard statistical significance test [<xref ref-type="bibr" rid="B14">14</xref>]. The problem with this approach is that for a large set of genes, the normal tissues in which that gene is relatively highly expressed could be the same. Therefore, the set of genes might work well to distinguish a tumor from most normal tissues, but consistently fail to distinguish the tumor class from a particular set of normal tissue samples. For example, for breast cancer samples, most of the genes (more than a thousand) with large expression values would also have large expression in normal breast samples. Thus a set of genes that distinguish breast cancers from most normal tissues will generally do a poor job of discriminating breast cancers from normal breast tissue. The validity of these methods oftentimes depends on the relative abundance of samples of the various normal tissue types; our method does not suffer from this drawback.</p></sec><sec><title>Related work</title><p>The work of [<xref ref-type="bibr" rid="B10">10</xref>] is most similar to our effort. The authors use a completely unsupervised self organizing map technique to cluster gene expression vectors, and identify tumor and normal classes in the SOM. The SOM also provides an expression vector unique to each tumor class. The authors report poor accuracy in classifying breast, ovarian, coloretical and lung cancers, and the feature sets produced for these classes are small in size. Our work differs from their work in being semi-supervised (though it is possible for our method to be run in a completely unsupervised fashion); we start with class labels for each broad tumor class (like CNS, renal, etc) and the set of normal samples. Our method uses an entirely different idea and is geared towards finding specialized subtypes within each histological type which are similar in terms of discrimination. Our clustering procedure finds biologically meaningful sub-types for Breast and Ovarian cancer, where we identify genes like ERBB2, ESR1, NAT1, GATA3 and MSLN as being prominent in the feature sets. These genes are well known markers for these cancer types. We identify three natural sub-types of breast cancer which traditional hierarchical clustering also identifies. In contrast to the traditional hierarchical clustering method, our method relies on very small feature sets to identify the exact same sub-types. In other cases, our method groups tumors differently from the traditional methods. Our classification results have around 75% accuracy for most cancer types except pancreas, prostate and gastric cancer, which is comparable to the 80% accuracy obtained by [<xref ref-type="bibr" rid="B10">10</xref>] (a caveat is that we are working with different data-sets, and this may effect the results of the predictions). However, we have the advantage of finding different clusters in some cases (for instance, breast cancer and ovarian cancer), and also finding a different set of markers, some of which have been verified in literature. In addition, we expect our method to produce results even when the class of normal tissues is extremely heterogeneous, with very few samples of any histological type. This suggests our approach as an alternative procedure which may work better in some situations.</p><p>The work of Dettling and Buhlmann [<xref ref-type="bibr" rid="B15">15</xref>] attempts to cluster genes based on multiclass discrimination of cancer types. We differ in attempting to cluster samples in each tumor type based on similarity of the discriminatory genes. We therefore get multiple clusters of genes for the same broad cancer type, which would reveal fine grained variations within that cancer type.</p><p>We emphasize that traditional classification and clustering procedures would be expected to have higher prediction accuracy for unknown samples because they rely more on overall gene expression patterns (this is confirmed in [<xref ref-type="bibr" rid="B10">10</xref>]); our clustering method is oriented towards grouping tumors based on similarity of feature sets, so as to identify highly specialized tumor sub-classes. We show that these feature sets have sufficient classifying power to make them statistically meaningful. In some cases, for instance, breast cancer samples, our clustering result mirrors that of traditional hierarchical clustering. In some other cases, for instance, lung cancer, our procedure groups tumor samples of diverse histo-logical origin together, suggesting that a similar marker set or treatment may be amenable to them. In addition, it would also suggest the lack of a common treatment for a set of samples which are histologically similar. There are alternative techniques to achieve the same goal, for instance [<xref ref-type="bibr" rid="B10">10</xref>]; the high dimensionality of the gene space often leads to different results from different methods, and a combination of these methods may lead to biologically meaningful insights.</p><p>Although this work focusses on the application to global gene expression data and cancer, the discriminative margin clustering method should be generally applicable to large high dimensional data-sets in which similar classification questions arise. For more details on the relevance of our objective, and the manner in which traditional clustering and classification procedures fail to address it, along with additional references to literature, we refer the reader to [<xref ref-type="bibr" rid="B10">10</xref>].</p></sec></sec><sec><title>Results</title><sec><title>Discriminative Margin Clustering</title><p>We once again refer to our example from the previous section. The "abnormal" samples in Figure <xref ref-type="fig" rid="F1">1</xref> lie near each other, and hence are well separated from the normals by a single line. This will not always be the case: and it hence it is of interest to group members of the abnormal class with respect to their joint separability. This is the idea of discriminative margin clustering.</p><p>The process follows the same general scheme as agglomerative hierarchical clustering, but uses <italic>margin from the normal class</italic>, rather than similarity of expression profiles, as the clustering metric. The idea is schematized in Figure <xref ref-type="fig" rid="F2">2</xref>.</p><p>We start with each abnormal sample forming its own group. Then for every pair of samples, we compute the maximum positive margin classifier for that pair versus the normal samples. We find the pair whose resulting margin is largest, and agglomerate them together. This process is repeated, and at each stage agglomerating the pair of single samples or groups that produces a combined group with largest margin.</p><p>Figures <xref ref-type="fig" rid="F3">3</xref> and <xref ref-type="fig" rid="F4">4</xref> show a toy example. There are 100 genes and 20 samples, 10 each of the normal and abnormal classes. This example is to illustrate the power of our clustering method in finding specialized tumor sub-types which other clustering methods would miss out. This example is not a stochastic model of gene expression, and is presented mainly to illustrate the potential difference in hierarchical clustering and discriminative margin clustering in terms of the sub-types they are capable of detecting. Discriminative margin clustering will detect highly specialized sub-types with similar feature sets, while hierarchical clustering will detect tumor types with similar gene expression patterns. Both methods therefore have their relative merits.</p><p>The first 50 genes have high expression for the normal samples and abnormal samples 11–15, and lower expression for samples 16–20, while the last 10 genes have high expression for samples 11,13,15,17 and 19 and low expression for the others. The left panel of Figure <xref ref-type="fig" rid="F4">4</xref> shows average linkage hierarchical clustering, using the Euclidean metric. The first 50 genes dominate the clustering, and hence samples 11–15 and 16–20 represent the major groups found. The right panel of the figure shows the result of discriminative margin clustering. It has separated samples 11,13,15,17,19, as these are the ones most easily discriminated from the normal class. Each join of the dendrogram is drawn at height equal to the (negative) margin achieved for the combined groups.</p><p>This discriminative margin clustering procedure delivers another useful piece of information. At each merge, it finds a set of non-negative gene weights that best separate the current group from the set of normals. In practical terms, this might mean that each successive cluster defines a group of cancers for which a specific combination therapy (directed at the products of the genes that best distinguish this group of tumors from all normal tissues) would be useful.</p><p>For example the join at height = -3.6 has weight vector</p><p>(0.14, 0.00, 0.27, 0.07, 0.00, 0.00, 0.14, 0.17, 0.00, 0.21, 0.00)</p><p>for the last 10 genes, and zero for the rest. Hence in this example we would learn that the separation of samples 11, 13, 15, 17, and 19 is best achieved by a (weighted subset of) genes 90–100.</p><p>The dendrogram from this process can also be used for classification of new samples. Given a new expression profile <italic>z </italic>= (<italic>z</italic><sub>1</sub>, <italic>z</italic><sub>2</sub>, ... <italic>z</italic><sub><italic>p</italic></sub>), at each join in the tree we compute the margin <inline-graphic xlink:href="1471-2105-5-21-i1.gif"/> where the <inline-graphic xlink:href="1471-2105-5-21-i2.gif"/> are the estimated gene weights. The margin for <italic>z </italic>is defined to be the largest value <italic>m </italic>such that <italic>z </italic>achieves a margin of <italic>m </italic>at a join at or below height -<italic>m </italic>of the tree. If the margin <italic>m </italic>is negative, it is set to -∞.</p><p>The predicted class is "abnormal" if <italic>m </italic>is greater than some cut-point <italic>h</italic>, and "normal" otherwise. Hence samples with margin -∞ are always predicted as "normal". The optimal cut-point <italic>h </italic>is estimated from a test set or cross-validation.</p><p>Figure <xref ref-type="fig" rid="F5">5</xref> shows the training and test error curves for the toy example. A test sample of size 500 was used. An "error" would be the classification of a normal sample as "abnormal", or vice-versa. The test error is minimized at a margin about 2.0, which is reasonable in view of the dendrogram of Figure <xref ref-type="fig" rid="F4">4</xref>. Therefore, we can obtain clusters of the "abnormal" samples by chopping the tree at a margin of 2.0.</p><p>The clustering procedure therefore has the potential of finding highly specialized tumor sub-types. These groups would be expected to have useful information in terms of diagnosis and therapy.</p><sec><title>Discussion</title><p>Although it would appear that our clustering procedure simply groups together samples based on the similarity of the maximum margin feature sets, this is not strictly true. The similarity between two tumor samples could be very high even if the maximum margin feature sets have low overlap. All we require is that there exist a common feature set which gives large margin for both the samples. Therefore, our clustering procedure is different from the more simple-minded procedure that computes similarity between the maximum margin feature sets. Therefore, if a set of tumor samples has a common discriminatory feature set (this need not be the maximum margin feature set for any of the samples individually), our clustering procedure would be expected to group them together. This explains why it performs well in classifying unknown samples in the cross-validation test we describe in the next section.</p><p>One advantage of clustering is removal of sensitivity to noise. Though expression values of a particular tumor sample may be prone to error, the combined margin classifier for a class of related tumor samples would be a reliable indicator of the genes which characterize the class. For a gene to be erroneously included with large weight in the feature set, its expression value has to be abnormally high in most of the samples in the class. This is a possibility if the class has only one sample, but this event has low probability if the class has many samples. This point also illustrates the importance of finding a <italic>weighted </italic>combination as opposed to an unweighted combination. Genes with large weight are more reliable markers than genes with small weight. For instance, ERBB2 appears as a gene with a large weight for a large cluster of breast samples, and therefore, would be a good candidate to include in the feature set. We also note that (contrary to intuition) finding the best weighted combination is a computationally simpler problem than finding the best un-weighted combination (which is provably computationally intractable). We therefore focus our attention on computing weighted combinations; we discuss how our results change for unweighted combinations in the next section.</p><p>One natural question to ask is what if the margin is negative. Our experiments with the tumor and normal data set which we present in the next section shows that even if we consider the entire class of tumors versus the entire class of normals, the maximum margin is positive, showing that there exists a set of discriminatory genes for this case as well. A similar result is obtained by [<xref ref-type="bibr" rid="B10">10</xref>]. We will therefore assume that the margin is always positive.</p><p>We note that though the set of genes yielding the largest possible margin is unique, there will be many sets of genes whose weighted combinations yields close to the largest margin. In Appendix B, we discuss techniques to find larger feature sets which yield close to the maximum margin. This has potential applications to finding diagnostic/therapeutic markers. We note that in our experiments with the tumor and normal dataset, the maximum margin feature set is contained in the expanded feature set, albeit with smaller weights assigned to the corresponding genes.</p></sec></sec><sec><title>Properties of the feature sets and expanding them</title><p>For a realistic look at its potential application to real biological problems, we apply discriminative margin clustering to a large dataset of gene expression data from systematic analysis of transcript levels in normal and cancer tissues, using DNA microarrays. We use data from the Stanford Microarray Database [<xref ref-type="bibr" rid="B16">16</xref>]. The log scale, mean centered and normalized dataset can be found at <ext-link ext-link-type="uri" xlink:href="http://microarray-pubs.stanford.edu/margin clus/"/>. We impute missing values using the <italic>k</italic>-nearest neighbors method with <italic>k </italic>= 10. For our method, accurate imputation of missing values is important. We set all remaining missing values to zero.</p><p>Since we wish only to cluster samples which have the same broad histological type (for instance, lung samples or ovarian samples), we run the clustering procedure separately on each broad tumor class versus all the normal samples. There are 104 normal samples and 268 tumor samples which fall in 14 broad tumor classes – Bladder, Breast, CNS, Kidney, Liver, Lung, Lymph, Ovary, Pancreas, Prostate, Skin, Soft tissue, Stomach and Testis. There are around 7500 genes, and the data is on the log scale.</p><p>Figure <xref ref-type="fig" rid="F6">6</xref> illustrates the discriminative margin clustering for a collection of lung samples. Note that the clustering groups the squamous, adeno and small cell sub-types separately, but the large cell sub-type does not cluster as a discrete group suggesting that these tumors are heterogenous with respect to the molecular features that distinguish them from normal tissues. Combining the large cell subtype together results in a feature set with very small margin, showing the absence of a common set of molecular markers. We note that traditional hierarchical clustering using a dot-product similarity measure would group the large-cell sub-type together, which shows these samples have similar overall gene expression patterns and histological type. But, for the purpose of finding discriminating feature sets, these samples are very heterogeneous.</p><p>Cancers of the same histological type can be heterogenous in their gene expression patterns, their genetic origin and their behavior. Therefore, we do not expect the discriminating gene set to be consistent among cancers of the same histological diagnosis. Nevertheless, we tested the ability of discriminative margin clustering to define gene expression features useful for classifying cancers according to broad histological types. We do a predictive analysis for the quality of the clustering produced using the identification of known cancer classes as a test. We describe the procedure below.</p><p>To obtain clusters from the dendogram, we need to find the margin at which to chop the tree. We do a cross-validation test by randomly partitioning the samples into two-third "training" and one-third "test" groups. We run the discriminative margin clustering procedure on each tumor class in the training set versus the normal samples in the training set. This yields one dendogram for each tumor class. For each sample in the test set, we find the node with the best margin among <italic>all </italic>dendograms generated from the training data. This would assign a <italic>predicted </italic>class for the test sample as CNS, renal, normal, etc, along with the margin. Figure <xref ref-type="fig" rid="F7">7</xref> shows the predicted class of each tumor and normal sample; the color assigned to a point is the color of the predicted class. A margin of -∞ implies classification as normal. From this figure, it we can deduce that a margin of 2.0 is the threshold beyond which the classification accuracy is large. To confirm this, we fix a margin, and chop the trees off at that margin to obtain a larger set of trees. Now, we re-classify the test samples using the nodes in the new trees (all nodes have margin at least the cut-off now). Figure <xref ref-type="fig" rid="F8">8</xref> shows the error in accurately classifying the normal and tumor samples in the test set. The error is minimized when the margin is 2.0. We therefore pick this as the margin.</p><p>For the margin of 2.0, Table <xref ref-type="table" rid="T1">1</xref> shows the prediction of the test samples. Note that the accuracy of prediction is around 75% for most classes, but bad for prostate, pancreas and stomach cancer. The normal samples cannot be predicted accurately as they are extremely heterogeneous in nature (there are very few samples for most histological types), and if the training set misses out the normal sample of a particular histological type, the feature sets may not classify the test samples corresponding to that type correctly. Despite this heterogeneity, the accuracy for predicting the tumor samples is comparable to that in [<xref ref-type="bibr" rid="B10">10</xref>] (they use a data set with comparable number of normal and tumor samples), and as we show below, the feature sets have genes which are biologically relevant for that cancer class. We then consider the prediction of the test sample using just the dendograms of the corresponding histological type. This would predict whether the test sample belongs to that histological type or whether it is a normal sample. This improves the prediction results (Table <xref ref-type="table" rid="T2">2</xref>). Note that the power of the clustering in predicting broad histological type is high, even though this analysis did not take into account the considerable molecular heterogeneity within the cancer classes.</p><p>We obtain similar classification accuracy even if we make the weights of the genes in the feature set equal to 1. This shows that the <italic>combination </italic>of genes is important in addition to the weights, and for a given set of expression values, the discriminatory feature sets are not very sensitive to the weights. However, the weighted combination is more resilient to noise in the data than the unweighted combination, especially since we are considering sets of small size.</p><p>We note that traditional classifiers, for instance, the nearest neighbor classifier (which we have implemented and compared against), have higher prediction accuracy (refer [<xref ref-type="bibr" rid="B10">10</xref>]), as they are based on broader gene expression profiles. Our method, in contrast, is geared towards finding tumor sub-types which are similar in terms of having the same small discriminative set of genes, and we use just these for classification to test the significance of these genes. These discriminative sets also make no explicit effort to discriminate one cancer class from another – we simply discriminate the cancer class from the normal samples. Our main point is to find clusters of specialized tumors with a common marker set. We show, somewhat surprisingly, that these feature sets have the additional power to classify unknown tumor samples well, implying that they have statistical validity.</p><p>In some cases, the clustering helps us identify natural tumor sub-types as well. For instance, breast cancer samples cluster into three natural sub-types, which are also identified by traditional hierarchical clustering.</p><sec><title>ERBB2 sub-type</title><p>The feature set for this class gives 65% weight to ERBB2.</p></sec><sec><title>Luminal A sub-type</title><p>ESR1, NAT1 and GATA3 together account for 55% of the weight in the feature set.</p></sec><sec><title>Proliferative sub-type</title><p>Characterized by large expression values for cell-cycle related genes, suggesting a rapidly proliferating sub-type.</p><p>The feature sets of genes produced at the nodes of the cluster tree are small (about 10 genes) even at the top nodes of the tree. Though these feature sets include genes which are well studied in the context of their respective cancers, these sets may be too small to provide a sufficient set of candidates for diagnosis or treatment. For example, for a breast cancer class with 15 samples, the feature set (along with the associated weights) is listed in Table <xref ref-type="table" rid="T3">3</xref>. Although the maximum margin feature set is small, there will be feature sets with close to the optimal margin, which are equally good candidates for finding markers. Note that the presence of these feature sets does not affect the clustering, but simply yields more marker sets.</p><p>We can expand the set of genes by convex programming techniques. At a particular node in the cluster tree, we consider the polytope of all feature sets whose margin is close to the optimal margin (the "closeness" being a parameter which we can control). We then formulate a quadratic program to search for a combination in this polytope that includes as many different genes as possible. Details can be found in Appendix B. These feature sets provide a much larger set of potential markers. In our experiments, the maximum margin feature set is contained in the expanded feature set, albeit with smaller weight.</p><p>For the breast cancer cluster in represented in Table <xref ref-type="table" rid="T3">3</xref>, the expression patterns for the expanded list of genes are shown in Figure <xref ref-type="fig" rid="F9">9</xref>, (and in more detail in the web supplement) in heatmap format. Adjacent to each gene is the weight assigned to it in the expanded feature set. For each tissue, we also indicate the weighted expression value of the genes in the feature set. The first 15 samples are breast cancer samples while the rest are normal tissue samples. Please refer to the web supplement for an enlarged version of this figure. We have clubbed two sub-classes found by the clustering in this heatmap. Note that the ERBB2 sub-type (last 4 tumor samples) has high expression of ERBB2, GRB7, MLN64 and LIV-1, while the Luminal A sub-type (remaining samples) has high expression of ESR1, NAT1 and LIV-1. The expression patterns of these genes is sufficient to discriminate these two sub-types from each other.</p><p>Many genes related to proliferation and the cell cycle (C20ORF1, TOPK, L2DTL, KNSL1, NUF2R, CENPF, ...) are present in the expanded feature set. These genes do not have very high differential expression values, and are relatively highly expressed in many cancer types. They are therefore not present in the maximum margin feature sets that drive the clustering procedure, though they are present in the expanded feature sets. On another note, GRB7, which is significantly co-expressed with ERBB2, is absent in the maximum margin feature set, but is present in the expanded feature set.</p><p>The heatmap clearly illustrates the spread of expression values in the normal tissue sample. For any one normal sample, the number of highly expressed genes is relatively small, but for any gene, the probability that at least one normal tissue expresses the gene at a level comparable to that in the given tumor is high. Nevertheless, this example shows that we can find a set of genes that as a group, discriminates the tumor class from the normal classes quite well.</p><p>Although our analysis used no information about the known diagnostic or therapeutic value of the genes, the feature sets identified were strikingly enriched in genes corresponding to the established therapeutic and diagnostic targets. We illustrate these in Table <xref ref-type="table" rid="T4">4</xref>. For example, ERBB2 and ESR1 are targeted by some of the most effective treatments for breast cancer. The genes MSLN and AMACR have been identified as useful markers of Ovarian and Prostate cancers respectively, and our feature sets are heavily weighted in favor of these genes. Note that despite some similarity, many of our markers are quite different from those in [<xref ref-type="bibr" rid="B10">10</xref>]. This may be because of the fact that we use different data-sets.</p><p>A detailed collection of dendograms and feature sets are present at <ext-link ext-link-type="uri" xlink:href="http://microarray-pubs.stanford.edu/margin clus/"/>.</p></sec></sec></sec><sec><title>Discussion</title><p>We have illustrated a procedure for identifying interesting feature sets of genes to distinguish a tumor class from a set of normal tissues from large scale systematic gene expression data sets, obtained by DNA microarrays. This method has wider applicability in finding feature sets to discriminate one set of data samples from another. The method has the advantage of producing feature sets of moderate size (neither too small nor too large), while at the same time discriminating the tumor class from all normal classes. The method involves a novel hierarchical clustering procedure combined with polytope exploration techniques. The clustering diminishes sensitivity of the results to noise, while the margin classifier technique gives equal weightage to discriminating the tumor class from all normal samples. The method may be especially valuable in identifying clinically useful sets of diagnostic or therapeutic markers for defined groups of cancers.</p><p>An interesting research direction which we plan to pursue is to use the results of the margin classifier to find combinations of genes whose protein products can be targeted by antibodies. This strategy may be useful in developing screening tests and drug treatments for cancers.</p></sec><sec><title>Paper web-site</title><p><ext-link ext-link-type="uri" xlink:href="http://microarray-pubs.stanford.edu/margin clus/"/></p></sec><sec sec-type="methods"><title>Methods</title><sec><title>A Details of the Margin Classifier</title><p>For clarity of exposition, let us denote the set of tumor tissues by <italic>T</italic>, and a set of normal tissues by <italic>N</italic>. For every tissue <italic>t </italic>∈ <italic>T </italic>and <italic>n </italic>∈ <italic>N</italic>, we are given the expression values of the set <italic>G </italic>of genes. For every gene, <italic>g </italic>∈ <italic>G </italic>and tissue, <italic>x </italic>∈ <italic>T </italic>∪ <italic>N</italic>, let <italic>e</italic><sub><italic>xg </italic></sub>denote the expression of gene <italic>g </italic>in tissue <italic>t</italic>. These expression levels are on the log scale.</p><p>Our goal is to find feature sets for the tumor tissue samples using the following general framework. Suppose for a subset <italic>T' </italic>⊆ <italic>T</italic>, there exists a set of fractional weights <italic>w </italic>= {<italic>w</italic><sub><italic>g</italic></sub>\|<italic>g </italic>∈ <italic>G</italic>, Σ<sub><italic>g</italic>∈<italic>G</italic></sub><italic>w</italic><sub><italic>g </italic></sub>= 1} such that the minimum weighted expression of the genes in every <italic>t </italic>∈ <italic>T'</italic>, given by <italic>E</italic><sub><italic>wT' </italic></sub>= min<sub><italic>t</italic>∈<italic>T' </italic></sub>Σ<sub><italic>g</italic></sub><italic>w</italic><sub><italic>g</italic></sub>·<italic>e</italic><sub><italic>tg </italic></sub>is much larger than the maximum expression of the weighted expression of the genes in any normal tissue, given by <italic>E</italic><sub><italic>wN </italic></sub>= max<sub><italic>n</italic>∈<italic>N </italic></sub>Σ<sub><italic>g</italic></sub><italic>w</italic><sub><italic>g</italic></sub>·<italic>e</italic><sub><italic>ng</italic></sub>. In other words, the difference (note that we are working in the log scale) <italic>r</italic><sub><italic>wT' </italic></sub>= <italic>E</italic><sub><italic>wT' </italic></sub>-<italic>E</italic><sub><italic>wN </italic></sub>is large. This means, for example, that a drug combination whose activity is directed at the products of the genes by the weighted combination <italic>w </italic>could target the tumor tissues more effectively than any normal tissue, and therefore would, in principle, be effective in chemotherapy.</p><p>The goal is to make the difference of <italic>r</italic><sub><italic>T'w </italic></sub>as large as possible by choosing an appropriate <italic>w</italic>. We call <italic>r</italic><sub><italic>T' </italic></sub>= max<sub><italic>w</italic></sub><italic>r</italic><sub><italic>wT'</italic></sub>, the effectiveness value for tissues <italic>T'</italic>. The <italic>w </italic>which maximizes <italic>r</italic><sub><italic>T'w </italic></sub>is the feature set of the tumors <italic>T'</italic>. Given <italic>t </italic>∈ <italic>T'</italic>, the goal of finding <italic>w </italic>can be reduced to a linear program as follows:</p><p><inline-graphic xlink:href="1471-2105-5-21-i3.gif"/></p><p>The feature set is the set <italic>G' </italic>= {<italic>g\|w</italic><sub><italic>g </italic></sub>> 0}. This linear program is precisely the one used for finding separating hyperplanes minimizing the <italic>l</italic><sub>1 </sub>norm in linear Support Vector Machines [<xref ref-type="bibr" rid="B1">1</xref>]. This contrasts to the non-linear approach is used in [<xref ref-type="bibr" rid="B5">5</xref>] for cancer classification. A similar linear programming approach is used by [<xref ref-type="bibr" rid="B11">11</xref>] to find small feature sets in gene expression data. We show later how to construct feature sets of variable sizes and confidence using our methods.</p><p>It is possible to construct examples where the best margin is negative, i.e., the classes overlap, but this never happened in all of our examples with real gene expression data. A feature of a linear classifier is that the number of nonzero weighted genes is at most the number of samples in the data-set. In our experiments, the number of genes with non-zero weight is typically much smaller than this upper bound.</p><p>We use a simplex method (available in the CPLEX commercial package) to solve this problem for each candidate merge. The idea behind the simplex algorithm is to iteratively modify the weights using an approach similar to gradient descent, as long as the objective function improves. It can be shown that the algorithm converges in a small number of steps, especially for problems of the form we discuss, which are called <italic>packing linear problems</italic>. The intuition behind the small number of steps is that the algorithm can easily find a direction in the gene space in which a large alteration in weights improves the objective function significantly while maintaining feasibility of the problem.</p><p>Let <italic>M</italic>(<italic>t</italic>) be the achieved margin for sample <italic>t</italic>, <italic>M</italic>({<italic>t</italic><sub>1</sub>, <italic>t</italic><sub>2</sub>}) is the margin for the pair of samples <italic>t</italic><sub>1</sub>, <italic>t</italic><sub>2 </sub>etc. A key fact here is that merging samples cannot increase the margin</p><p><italic>M</italic>({<italic>t</italic><sub>1</sub>, <italic>t</italic><sub>2</sub>}) ≤ min [<italic>M </italic>(<italic>t</italic><sub>1</sub>), <italic>M </italic>(<italic>t</italic><sub>2</sub>)]    (1)</p><p>This allows us to draw the tree with join heights equal to the negative margins.</p><p>A computationally simpler approximation we use, which produces the same clustering results is the following:</p><p><italic>M</italic>({<italic>T</italic><sub>1</sub>, <italic>T</italic><sub>2</sub>}) ≈ min [<italic>M </italic>(<italic>T</italic><sub>1</sub>), <italic>M </italic>(<italic>T</italic><sub>2</sub>), min{<italic>M </italic>(<italic>t</italic><sub>1</sub>, <italic>t</italic><sub>2</sub>), <italic>t</italic><sub>1</sub>∈ <italic>T</italic><sub>1</sub>, <italic>t</italic><sub>2</sub>∈ <italic>T</italic><sub>2</sub>}]    (2)</p><p>This says that the margin achieved in joining groups <italic>T</italic><sub>1</sub>, <italic>T</italic><sub>2 </sub>is the minimum margin for each of the two groups, and all pairs of samples, one in group <italic>T</italic><sub>1 </sub>and the other in group <italic>T</italic><sub>2</sub>. The left hand side of (2) must be less than or equal the right hand side, and in general this seems to be a reasonable approximation. It allows us to find the best probable pair using only quantities that have already been computed. Note that as long as the clustering results are the same, the feature sets we find using the simpler approximation would be the same as those found using the original scheme, and this is what we observe in practice.</p></sec><sec><title>Finding competitors at each join in the tree</title><p>The linear program mentioned above has the problem of reporting a relatively small set of genes. It would miss out genes which are heavily weighted in some distinct feature set whose margin is close to the optimal one.</p><p>It is therefore desirable to have routines that "search" around the optimal solution to find other good feature sets. We present several approaches which find feature sets of varying sizes, with a concrete confidence measure on the importance of each set. It is also easy to incorporate additional constraints, like insisting certain genes are present to a minimum fraction, in our methods.</p><p>The first step in this process is to relax the margin slightly and define a new polytope where any weight vector is a "good" weight vector.</p><p>Suppose ε is the relaxation in the margin. Choosing a larger value of ε would yield a larger set of possible feature sets to choose from, but the margin obtained from these would be lower, implying lesser confidence in the set. For our data-set, ε = 0.4 produces feature sets of moderate size (50 – 100 genes).</p><p>For a set of samples <italic>T'</italic>, let <italic>M</italic>(<italic>T'</italic>) = <italic>R</italic>. The polytope <italic>P</italic><sub>ε </sub>(<italic>T'</italic>) is defined as:</p><p><inline-graphic xlink:href="1471-2105-5-21-i4.gif"/></p><p>Our goal now is to find weight vectors in <italic>P</italic><sub>ε </sub>(<italic>T'</italic>) with large number of non-zero dimensions (genes). We outline two methods below.</p><sec><title>Non-Linear programming approach</title><p>Our first approach is to solve the following program:</p><p><inline-graphic xlink:href="1471-2105-5-21-i5.gif"/></p><p>Though the objective function is non-linear, it is convex, and therefore can be optimized using interior point methods. The CPLEX barrier optimizer can optimize for this function.</p><p>The advantage of this function is that it "spreads" the weight over many genes and typically gives a much larger feature set than the maximum margin feature set.</p></sec><sec><title>Minimum overlap method</title><p>In this method, we start with a set of genes <italic>G' </italic>and iteratively expand this set by finding extreme points with minimum overlap with this set. Given any starting extreme point with gene set <italic>G'</italic>, consider the following linear program:</p><p><inline-graphic xlink:href="1471-2105-5-21-i6.gif"/></p><p>This will find a weight vector <inline-graphic xlink:href="1471-2105-5-21-i7.gif"/> whose total weight along the dimensions <italic>G' </italic>is as small as possible. Let <italic>G" </italic>= {<italic>g </italic>∈ <italic>G</italic>\|<italic>w</italic><sub><italic>g </italic></sub>> 0}. We set <italic>G' </italic>← <italic>G' </italic>∪ <italic>G"</italic>, and iterate.</p><p>We can stop when the objective value stops changing substantially. This technique quickly finds a large set of important genes.</p><p>This method has an added advantage. Given any feature set, if we set <italic>G' </italic>to be this set, the objective value of the linear program tells us the minimum fraction to which genes in this set must be present in <italic>any </italic>weight vector in the polytope <italic>P</italic><sub>ε</sub>(<italic>T'</italic>). This can be used as a confidence measure of the feature set.</p><p>We tested the algorithms on a group of 3 Breast tumors, with ε set to 0.4. We find the confidence of a feature set using the program described above. The maximum margin feature set with 8 genes has confidence (as defined above) 37%. The non-linear technique produces a larger feature set of around 30 genes with confidence of 70%, which is a much larger confidence than the solution produced by the maximum margin classifier.</p></sec></sec></sec>
Cultural Issues in Using the SF-36 Health Survey in Asia: Results from Taiwan	<sec><title>Background</title><p>The feasibility of using the SF-36 in non-Western cultures is important for researchers seeking to understand cultural influences upon health status perceptions. This paper reports on the performance of the Taiwan version of the SF-36, including the implications of cultural influences.</p></sec><sec sec-type="methods"><title>Methods</title><p>A total of 1191 volunteered subjects from the general population answered the translated SF-36 Taiwan version, which was developed following IQOLA project protocols.</p></sec><sec><title>Results</title><p>Results from tests of scaling assumptions and reliability generally were satisfactory. Convergent validity, as assessed by comparing the SF-36 to a mental health oriented inventory, was acceptable. Results of principal components analysis were similar to US results for many scales. However, differences were seen for the Vitality scale which was a stronger measure of mental health than physical health in Taiwan. Results are compared to those from other Asian studies and the U.S.</p></sec><sec><title>Conclusion</title><p>The results raise important questions regarding cultural influences in international studies of health status assessment. Further research into the conceptualization and components of mental health in Asian countries is warranted.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Tseng</surname><given-names>Hsu-Min</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Lu</surname><given-names>Jui-fen Rachel</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Gandek</surname><given-names>Barbara</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	Health and Quality of Life Outcomes	<sec><title>Background</title><p>The health outcomes of a population can be measured in terms of etiology and pathogenesis. Nevertheless, well-developed health outcome measurement systems have expanded the measurement of health beyond the classical endpoints of mortality and morbidity in clinical practice [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Health-related quality of life has emerged as the new reflection of modern medicine as viewed from biopsychosocial perspectives. With the fast growth in health care expenditures, this concept has been increasingly used as an important attribute in patient care and clinical studies as well as health economic evaluations [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>Over the past 20 years, health status measures have been widely applied in different medical fields such as oncology, cardiology, arthritis, and psychiatry [<xref ref-type="bibr" rid="B4">4</xref>]. Health researchers have begun to evaluate whether common standardized health status measures are technically and conceptually equivalent for various socio-cultural groups. Hence, there is an increasing need for international standards to measure health status in a manner that allows comparisons across countries, but which also are relevant within individual cultures. In particular, the well-recognized differences between Western and Eastern cultures may well be reflected in health status measurement results.</p><p>Encountering diversified cultural backgrounds, researchers hence have to take even more cautious steps in translating well-established standard instruments in Asia [<xref ref-type="bibr" rid="B5">5</xref>]. Extensive psychometric testing is also required for the translated instruments. Amongst those instruments which have become standard in the health status field, the SF-36 is one of the most widely accepted, extensively translated and tested instruments around the world. The International Quality of Life Assessment (IQOLA) Project was formed in 1991 and has developed a standard protocol for translating and psychometrically testing the SF-36 in different language versions[<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>].</p><p>As one aspect of health status assessment is to measure an individual's physical and mental state, respondents are often sensitive to wording which reflects differences in ethnicity and culture, even if the language used is the same in a broad sense. English and Chinese provide good examples of this. Although most of the words are similar, there are US English and U.K. English versions of the SF-36 Health Survey, reflecting linguistic differences in the two cultures. Similarly, the IQOLA Project has collaborated on the development of several Chinese versions of the SF-36 Health Survey. The published Chinese versions are for the US Chinese population[<xref ref-type="bibr" rid="B8">8</xref>] and Hong Kong (the HK translation is being used in some other Southeast Asian countries) [<xref ref-type="bibr" rid="B9">9</xref>]. Results of psychometric testing of these Chinese versions suggested that scaling assumptions were generally met and conceptual equivalence was achieved in comparison with forms using Western languages. However, both Chinese-American studies reported less satisfactory psychometric results for the mental health and vitality scales, as evidenced by high correlations between items in both scales. These results suggest that health-related quality of life (HRQOL) measures may need to be interpreted within a cultural framework.</p><p>Although these studies have begun to address the need for different versions of Chinese health surveys, one issue is that the samples used in these studies (either Chinese Americans or Hong-Kong Chinese) have been somewhat adapted to Western cultures from an ecological point of view. Thus, results of those studies may be influenced by a mixture of Chinese and Western cultures. Within a cultural framework, people in Taiwan are more influenced by Chinese culture than Western ones as a legacy of historical and political developments. Thus, the Taiwan population provides a distinct culture with which to address influences of Chinese cultural factors upon QOL measures. In this paper, we document the adaptation process of the SF-36 for use in Taiwan, by presenting the translation process for the SF-36 Taiwan version and the results of psychometric tests performed on <italic>three </italic>different groups. The implications of these results are then discussed within a cultural framework.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Design and Sample</title><p>The standard Taiwan version of SF-36 was administered to a total of 1,191 volunteered subjects who participated in three studies of health status surveys. The mean age of all the respondents was 28.2 (S.D. = 14.2) years; 64% were female. The first sample contained data from 614 freshmen students in one private university (mean age = 18.3, S.D. = 0.66; 46.9% female), who responded to the SF-36 Taiwan version (among other scales) in a screening survey of psychological well-being. The second was based on a study that examined the impact of organizational factors upon health. A total of 501 employees of a medical institute were approached and among them 491 employees (mean age = 34.9, S.D. = 10.06; 85.9% female) returned valid SF-36 data for analysis. The third was a group of 76 elderly people taking part in a study on the effects of Tai-Chi practice upon physical strength and balance (mean age = 65.9, S.D. = 5.43; 60.5% female). After excluding questionnaires with invalid data, a total of 1181 SF-36 profiles were available for analysis.</p><p>In the student sample, an additional questionnaire, the "Stress Coping Inventory", was administered to test the criterion validity of the <italic>mental health dimension </italic>in the translated SF-36 version. A total of 569 students returned valid data for this analysis.</p></sec><sec><title>Measures</title><sec><title>SF-36 Health Survey</title><p>The SF-36 measures eight health concepts: physical functioning (PF), role limitations due to physical problems (RP), bodily pain (BP), general health (GH), vitality (VT), social functioning (SF), role limitations due to emotional problems (RE), and mental health (MH), which were selected from the Medical Outcomes Study (MOS) inventory [<xref ref-type="bibr" rid="B10">10</xref>]. During its development, extensive and thorough psychometric testing was performed not only on the general population but also on diverse disease groups, and comparisons with other established instruments were also made[<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. Before the development of the SF-36 Taiwan standard version in 1996, there were a few researchers who had independently translated the SF-36 into Mandarin, the official language used in Taiwan. The translators of three versions were native Mandarin speakers with a fair amount educational training in the US; most possessed doctoral degrees from major US universities. The first stage of the development of the SF-36 Taiwan version was to reconcile the three existing versions into one single version. All translators and some experienced users of the SF-36 (US version) were invited to discuss the three versions in a coordination meeting. All participants in the meeting were asked to review the three translated versions and agree on the final translation item by item. The criteria of agreement were based on clarity (<italic>is it clear?</italic>), common language use (<italic>is it easy to understand?</italic>) and conceptual equivalence (<italic>is the concept measured in original US-English version captured?</italic>). In terms of common language use, the convenience of reading the translation into the local dialect, Taiwanese, was especially taken into consideration because Taiwanese is more commonly used in older populations. As an example of conceptual equivalence, the list for moderate activities in item 3 b was slightly modified to substitute "playing Tai-Chi" for "playing golf", since golf is not a common sport among the general population of Taiwan. Other activities were changed with equivalent concepts included "pushing a vacuum cleaner" in item 3 b ("mopping the floor" is used instead) and units for distance (e.g. kilometer is used to substitute for miles) in item 3 g.</p><p>After the consolidated version was developed, it was evaluated by a focus group, which was composed of experts in the fields of public health, psychology, psychometrics, nursing, social work and family medicine. In the focus group meeting, members reviewed the questionnaire item by item and evaluated the consolidated translation, based on their training and expertise. Modifications were made when it was deemed necessary, and the members agreed upon a second consolidated version which was then backward translated into English. The Principal Investigator for the Taiwan team and the IQOLA project director had an extensive discussion on problematic items and response choices. The Taiwanese SF-36 standard version then was produced and ready to be field tested.</p></sec></sec><sec><title>Stress Coping Inventory</title><p>The Stress Coping Inventory was developed by Chen and Wu (1987) [<xref ref-type="bibr" rid="B13">13</xref>], based on the bio-psychosocial model. It aims to investigate how personal resources influence the mental health outcome of stress coping among university students. The questionnaire includes 52 items and has two parts: evaluation of individual resources for stress coping and investigation of various psychological symptoms. Higher scores in the resources dimension represent persons with more resources (either personal or social) to cope with stress. On the contrary, higher psychological symptoms mean worse mental health. Personal resources are composed of three subscales, namely "self-esteem," "friendship support," and "family support". Two subscales labeled as "anxiety reaction" and "depression reaction," were related to psychological symptoms. The questionnaire has acceptable internal consistency (Cronbach's α ranged from 0.74 to 0.95).</p></sec><sec><title>Data Analysis</title><p>Psychometric analyses included tests of assumptions underlying the construction of SF-36 scales, analysis of principal components to test the hypothesized structure of the SF-36, and tests of construct validity using criterion-based and construct approaches. Following the IQOLA procedure, the Multitrait Analysis Program-Revised (MAP-R) for Windows was used to test whether the scores satisfied summated-rating scaling assumptions[<xref ref-type="bibr" rid="B14">14</xref>]. Assumptions underlying the scoring of SF-36 data from each of these studies were evaluated to determine if it was appropriate to use the method of summated ratings to score the SF-36 scales, following standard SF-36 scoring algorithms[<xref ref-type="bibr" rid="B15">15</xref>]. Internal consistency reliability for each scale score was estimated using Cronbach's alpha coefficient. Because higher levels of reliability increase statistical power, a minimum reliability of 0.70 for measures used in-group comparisons has been recommended[<xref ref-type="bibr" rid="B16">16</xref>]. In addition, the percentages of subjects achieving either the highest score (ceiling) or lowest score (floor) were calculated, because a large ceiling or floor effect will limit the ability of SF-36 to detect change over time.</p><p>Validity was tested using construct and criterion-based approaches. The SF-36 was constructed to represent two major dimensions of health – physical and mental. A second-order factor analysis (using principal component analysis) of the 8 SF-36 scale scores was carried out to test the assumption that there were two underlying factors in the SF-36. Two factors with eigenvalues greater than 1 were extracted and rotated to orthogonal simple structure using the varimax method. To interpret the two components, we then examined the strength of their correlations with the eight scales. In addition, the results were compared to published results from HK, US, and Japan[<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. An additional test of construct validity was conducted by comparing SF-36 scale scores to scores from the Stress Coping Inventory, which was constructed to investigate the concept of mental health. It is therefore anticipated that scales of the Stress Coping Inventory within the resources part will positively correlate with SF-36 scales related to the mental health dimension, while a reverse relationship will be found for those within the psychosomatic symptoms part. In addition, a stronger association is anticipated with the mental health dimension than with the physical health dimension in the SF-36 because the construction of the Stress Coping Inventory is more mentally oriented. Criterion-based validity was tested by comparing the elderly group with the university student one. It is anticipated that subscales of the physical health dimension (e.g. PF, RP, BP) will decline with increasing age, while those in the mental health dimension (i.e. MH, RE) will be less influenced by age.</p></sec></sec><sec><title>Results</title><p>The psychometric properties of the translated version are presented in terms of data quality, reliability, and validity.</p><sec><title>Data Quality and Descriptive Statistics</title><p>A total of 1,181 records were available for psychometric testing of the SF-36 Taiwan version. Table <xref ref-type="table" rid="T1">1</xref> presents scale means and standard deviations, and the percentage scoring at the ceiling and floor for the SF-36 scales. The results of the item descriptive statistics indicate that SF-36 Taiwan version have a high rate of data completeness (Table <xref ref-type="table" rid="T2">2</xref>). The rates of missing values on the item level were consistently low, ranging from 0.0% (GH1) to a high of 2.7% (GH2). As would be expected for a sample that is primarily composed of healthy respondents, response distributions tended to be skewed in the direction of positive health. This is especially evidenced by the results that substantial ceiling effects were more frequently encountered in scales measuring functional limitations (e.g. PF, RP, and RE). Conversely, the percentage of respondents scoring at the lowest scale level (i.e., floor effect) was minimal in that floor effects were observed in less than 1% of the sample for all but the two role functioning scales (RP and RE).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Descriptive Statistics for SF-36 Scales</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center">Mean</td><td align="center">SD</td><td align="center">Floor (%)</td><td align="center">Ceiling (%)</td></tr></thead><tbody><tr><td align="right">PF – Physical Functioning</td><td align="center">92.60</td><td align="center">11.47</td><td align="right">0.0%</td><td align="right">45.3%</td></tr><tr><td align="right">RP – Role–Physical</td><td align="center">83.56</td><td align="center">28.88</td><td align="right">5.9%</td><td align="right">67.9%</td></tr><tr><td align="right">BP – Bodily Pain</td><td align="center">82.38</td><td align="center">16.75</td><td align="right">0.2%</td><td align="right">31.1%</td></tr><tr><td align="right">GH – General Health</td><td align="center">67.49</td><td align="center">18.21</td><td align="right">0.1%</td><td align="right">2.3%</td></tr><tr><td align="right">VT – Vitality</td><td align="center">65.32</td><td align="center">15.15</td><td align="right">0.0%</td><td align="right">0.9%</td></tr><tr><td align="right">SF – Social Functioning</td><td align="center">79.35</td><td align="center">16.00</td><td align="right">0.1%</td><td align="right">19.5%</td></tr><tr><td align="right">RE – Role–Emotional</td><td align="center">71.32</td><td align="center">36.98</td><td align="right">14.1%</td><td align="right">55.4%</td></tr><tr><td align="right">MH – Mental Health</td><td align="center">68.43</td><td align="center">14.67</td><td align="right">0.0%</td><td align="right">1.6%</td></tr></tbody></table></table-wrap><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Item Descriptive Statistics and Pearson Item-Scale Correlations Corrected for Overlap</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Scale (Choice Range)</td><td align="center">Item</td><td align="center">Missing</td><td align="center">Mean</td><td align="center">SD</td><td align="left">PF</td><td align="left">RP</td><td align="left">BP</td><td align="left">GH</td><td align="left">VT</td><td align="left">SF</td><td align="left">RE</td><td align="left">MH</td></tr></thead><tbody><tr><td align="left">Physical Functioning (1–3)</td><td align="center">PF1</td><td align="center">1.1%</td><td align="center">2.44</td><td align="center">0.62</td><td align="left"><bold>0.51</bold><sup>a</sup></td><td align="left">0.40</td><td align="left">0.31</td><td align="left">0.34</td><td align="left">0.22</td><td align="left">0.2</td><td align="left">0.14</td><td align="left">0.13</td></tr><tr><td></td><td align="center">PF2</td><td align="center">1.1%</td><td align="center">2.87</td><td align="center">0.38</td><td align="left"><bold>0.65</bold><sup>a</sup></td><td align="left">0.33</td><td align="left">0.29</td><td align="left">0.25</td><td align="left">0.17</td><td align="left">0.18</td><td align="left">0.05</td><td align="left">0.11</td></tr><tr><td></td><td align="center">PF3</td><td align="center">1.1%</td><td align="center">2.91</td><td align="center">0.32</td><td align="left"><bold>0.59</bold><sup>a</sup></td><td align="left">0.29</td><td align="left">0.28</td><td align="left">0.18</td><td align="left">0.11</td><td align="left">0.14</td><td align="left">0.07</td><td align="left">0.06</td></tr><tr><td></td><td align="center">PF4</td><td align="center">1.0%</td><td align="center">2.75</td><td align="center">0.48</td><td align="left"><bold>0.58</bold><sup>a</sup></td><td align="left">0.32</td><td align="left">0.28</td><td align="left">0.23</td><td align="left">0.17</td><td align="left">0.15</td><td align="left">0.11</td><td align="left">0.12</td></tr><tr><td></td><td align="center">PF5</td><td align="center">1.2%</td><td align="center">2.96</td><td align="center">0.22</td><td align="left"><bold>0.60</bold><sup>a</sup></td><td align="left">0.25</td><td align="left">0.25</td><td align="left">0.18</td><td align="left">0.16</td><td align="left">0.18</td><td align="left">0.13</td><td align="left">0.10</td></tr><tr><td></td><td align="center">PF6</td><td align="center">1.0%</td><td align="center">2.88</td><td align="center">0.36</td><td align="left"><bold>0.58</bold><sup>a</sup></td><td align="left">0.29</td><td align="left">0.32</td><td align="left">0.22</td><td align="left">0.15</td><td align="left">0.16</td><td align="left">0.11</td><td align="left">0.11</td></tr><tr><td></td><td align="center">PF7</td><td align="center">0.8%</td><td align="center">2.85</td><td align="center">0.39</td><td align="left"><bold>0.53</bold><sup>a</sup></td><td align="left">0.33</td><td align="left">0.2</td><td align="left">0.23</td><td align="left">0.23</td><td align="left">0.21</td><td align="left">0.19</td><td align="left">0.18</td></tr><tr><td></td><td align="center">PF8</td><td align="center">0.8%</td><td align="center">2.92</td><td align="center">0.30</td><td align="left"><bold>0.61</bold><sup>a</sup></td><td align="left">0.36</td><td align="left">0.22</td><td align="left">0.17</td><td align="left">0.17</td><td align="left">0.22</td><td align="left">0.17</td><td align="left">0.10</td></tr><tr><td></td><td align="center">PF9</td><td align="center">0.9%</td><td align="center">2.98</td><td align="center">0.17</td><td align="left"><bold>0.51</bold><sup>a</sup></td><td align="left">0.19</td><td align="left">0.16</td><td align="left">0.12</td><td align="left">0.07</td><td align="left">0.10</td><td align="left">0.09</td><td align="left">0.05</td></tr><tr><td></td><td align="center">PF10</td><td align="center">0.8%</td><td align="center">2.98</td><td align="center">0.16</td><td align="left"><bold>0.42</bold><sup>a</sup></td><td align="left">0.18</td><td align="left">0.12</td><td align="left">0.12</td><td align="left">0.10</td><td align="left">0.13</td><td align="left">0.13</td><td align="left">0.10</td></tr><tr><td align="left">Role Physical (1–2)</td><td align="center">RP1</td><td align="center">0.9%</td><td align="center">1.83</td><td align="center">0.37</td><td align="left">0.33</td><td align="left"><bold>0.63</bold><sup>a</sup></td><td align="left">0.29</td><td align="left">0.28</td><td align="left">0.17</td><td align="left">0.29</td><td align="left">0.26</td><td align="left">0.08</td></tr><tr><td></td><td align="center">RP2</td><td align="center">1.0%</td><td align="center">1.80</td><td align="center">0.40</td><td align="left">0.31</td><td align="left"><bold>0.58</bold><sup>a</sup></td><td align="left">0.27</td><td align="left">0.33</td><td align="left">0.21</td><td align="left">0.29</td><td align="left">0.35</td><td align="left">0.14</td></tr><tr><td></td><td align="center">RP3</td><td align="center">1.0%</td><td align="center">1.84</td><td align="center">0.37</td><td align="left">0.43</td><td align="left"><bold>0.63</bold><sup>a</sup></td><td align="left">0.35</td><td align="left">0.32</td><td align="left">0.18</td><td align="left">0.28</td><td align="left">0.22</td><td align="left">0.12</td></tr><tr><td></td><td align="center">RP4</td><td align="center">1.0%</td><td align="center">1.87</td><td align="center">0.34</td><td align="left">0.42</td><td align="left"><bold>0.54</bold><sup>a</sup></td><td align="left">0.32</td><td align="left">0.31</td><td align="left">0.26</td><td align="left">0.30</td><td align="left">0.23</td><td align="left">0.19</td></tr><tr><td align="left">Bodily Pain (1–6)</td><td align="center">BP1</td><td align="center">0.2%</td><td align="center">4.88</td><td align="center">1.02</td><td align="left">0.33</td><td align="left">0.31</td><td align="left"><bold>0.60</bold><sup>a</sup></td><td align="left">0.40</td><td align="left">0.28</td><td align="left">0.29</td><td align="left">0.17</td><td align="left">0.25</td></tr><tr><td></td><td align="center">BP2</td><td align="center">1.1%</td><td align="center">4.53</td><td align="center">0.65</td><td align="left">0.37</td><td align="left">0.41</td><td align="left"><bold>0.60</bold><sup>a</sup></td><td align="left">0.37</td><td align="left">0.28</td><td align="left">0.42</td><td align="left">0.26</td><td align="left">0.23</td></tr><tr><td align="left">General Health (1–5)</td><td align="center">GH1</td><td align="center">0.0%</td><td align="center">3.05</td><td align="center">0.97</td><td align="left">0.28</td><td align="left">0.33</td><td align="left">0.33</td><td align="left"><bold>0.56</bold><sup>a</sup></td><td align="left">0.41</td><td align="left">0.24</td><td align="left">0.17</td><td align="left">0.30</td></tr><tr><td></td><td align="center">GH2</td><td align="center">2.7%</td><td align="center">3.85</td><td align="center">1.07</td><td align="left">0.19</td><td align="left">0.30</td><td align="left">0.34</td><td align="left"><bold>0.56</bold><sup>a</sup></td><td align="left">0.34</td><td align="left">0.26</td><td align="left">0.21</td><td align="left">0.27</td></tr><tr><td></td><td align="center">GH3</td><td align="center">2.5%</td><td align="center">3.94</td><td align="center">0.96</td><td align="left">0.20</td><td align="left">0.23</td><td align="left">0.25</td><td align="left"><bold>0.54</bold><sup>a</sup></td><td align="left">0.37</td><td align="left">0.21</td><td align="left">0.16</td><td align="left">0.32</td></tr><tr><td></td><td align="center">GH4</td><td align="center">2.6%</td><td align="center">3.86</td><td align="center">1.08</td><td align="left">0.29</td><td align="left">0.28</td><td align="left">0.28</td><td align="left"><bold>0.45</bold><sup>a</sup></td><td align="left">0.32</td><td align="left">0.27</td><td align="left">0.16</td><td align="left">0.28</td></tr><tr><td></td><td align="center">GH5</td><td align="center">2.5%</td><td align="center">3.80</td><td align="center">0.92</td><td align="left">0.27</td><td align="left">0.31</td><td align="left">0.37</td><td align="left"><bold>0.68</bold><sup>a</sup></td><td align="left">0.46</td><td align="left">0.30</td><td align="left">0.21</td><td align="left">0.40</td></tr><tr><td align="left">Vitality (1–6)</td><td align="center">VT1</td><td align="center">1.2%</td><td align="center">4.34</td><td align="center">1.09</td><td align="left">0.18</td><td align="left">0.23</td><td align="left">0.21</td><td align="left">0.42</td><td align="left"><bold>0.58</bold><sup>a</sup></td><td align="left">0.34</td><td align="left">0.26</td><td align="left">0.52</td></tr><tr><td></td><td align="center">VT2</td><td align="center">1.2%</td><td align="center">4.14</td><td align="center">1.09</td><td align="left">0.20</td><td align="left">0.15</td><td align="left">0.23</td><td align="left">0.44</td><td align="left"><bold>0.63</bold><sup>a</sup></td><td align="left">0.33</td><td align="left">0.25</td><td align="left">0.59</td></tr><tr><td></td><td align="center">VT3</td><td align="center">1.2%</td><td align="center">4.47</td><td align="center">0.90</td><td align="left">0.23</td><td align="left">0.24</td><td align="left">0.24</td><td align="left">0.33</td><td align="left"><bold>0.46</bold><sup>a</sup></td><td align="left">0.32</td><td align="left">0.26</td><td align="left">0.44</td></tr><tr><td></td><td align="center">VT4</td><td align="center">1.0%</td><td align="center">4.11</td><td align="center">0.90</td><td align="left">0.16</td><td align="left">0.17</td><td align="left">0.27</td><td align="left">0.35</td><td align="left"><bold>0.53</bold><sup>a</sup></td><td align="left">0.35</td><td align="left">0.23</td><td align="left">0.53</td></tr><tr><td align="left">Social Functioning (1–5)</td><td align="center">SF1</td><td align="center">0.1%</td><td align="center">4.44</td><td align="center">0.69</td><td align="left">0.18</td><td align="left">0.31</td><td align="left">0.33</td><td align="left">0.28</td><td align="left">0.29</td><td align="left"><bold>0.41</bold><sup>a</sup></td><td align="left">0.39</td><td align="left">0.41</td></tr><tr><td></td><td align="center">SF2</td><td align="center">2.5%</td><td align="center">3.91</td><td align="center">0.84</td><td align="left">0.25</td><td align="left">0.32</td><td align="left">0.32</td><td align="left">0.31</td><td align="left">0.43</td><td align="left"><bold>0.41</bold><sup>a</sup></td><td align="left">0.30</td><td align="left">0.42</td></tr><tr><td align="left">Role Emotional (1–2)</td><td align="center">RE1</td><td align="center">0.3%</td><td align="center">1.74</td><td align="center">0.44</td><td align="left">0.17</td><td align="left">0.31</td><td align="left">0.18</td><td align="left">0.20</td><td align="left">0.27</td><td align="left">0.37</td><td align="left"><bold>0.61</bold><sup>a</sup></td><td align="left">0.32</td></tr><tr><td></td><td align="center">RE2</td><td align="center">0.2%</td><td align="center">1.71</td><td align="center">0.45</td><td align="left">0.17</td><td align="left">0.32</td><td align="left">0.19</td><td align="left">0.21</td><td align="left">0.26</td><td align="left">0.33</td><td align="left"><bold>0.64</bold><sup>a</sup></td><td align="left">0.29</td></tr><tr><td></td><td align="center">RE3</td><td align="center">0.3%</td><td align="center">1.69</td><td align="center">0.46</td><td align="left">0.11</td><td align="left">0.22</td><td align="left">0.19</td><td align="left">0.20</td><td align="left">0.29</td><td align="left">0.30</td><td align="left"><bold>0.50</bold><sup>a</sup></td><td align="left">0.32</td></tr><tr><td align="left">Mental Health (1–6)</td><td align="center">MH1</td><td align="center">1.2%</td><td align="center">4.07</td><td align="center">1.10</td><td align="left">0.11</td><td align="left">0.12</td><td align="left">0.21</td><td align="left">0.28</td><td align="left">0.34</td><td align="left">0.26</td><td align="left">0.22</td><td align="left"><bold>0.40</bold><sup>a</sup></td></tr><tr><td></td><td align="center">MH2</td><td align="center">1.5%</td><td align="center">4.69</td><td align="center">0.92</td><td align="left">0.15</td><td align="left">0.12</td><td align="left">0.17</td><td align="left">0.29</td><td align="left">0.52</td><td align="left">0.38</td><td align="left">0.29</td><td align="left"><bold>0.61</bold><sup>a</sup></td></tr><tr><td></td><td align="center">MH3</td><td align="center">1.2%</td><td align="center">4.31</td><td align="center">1.02</td><td align="left">0.14</td><td align="left">0.12</td><td align="left">0.24</td><td align="left">0.34</td><td align="left">0.54</td><td align="left">0.37</td><td align="left">0.32</td><td align="left"><bold>0.56</bold><sup>a</sup></td></tr><tr><td></td><td align="center">MH4</td><td align="center">1.2%</td><td align="center">4.53</td><td align="center">0.91</td><td align="left">0.11</td><td align="left">0.12</td><td align="left">0.19</td><td align="left">0.31</td><td align="left">0.54</td><td align="left">0.41</td><td align="left">0.28</td><td align="left"><bold>0.63</bold><sup>a</sup></td></tr><tr><td></td><td align="center">MH5</td><td align="center">1.0%</td><td align="center">4.50</td><td align="center">1.07</td><td align="left">0.11</td><td align="left">0.13</td><td align="left">0.17</td><td align="left">0.34</td><td align="left">0.60</td><td align="left">0.37</td><td align="left">0.27</td><td align="left"><bold>0.61</bold><sup>a</sup></td></tr></tbody></table><table-wrap-foot><p><sup>a </sup>denotes correlation between an item and its hypothesized scale.</p></table-wrap-foot></table-wrap></sec><sec><title>Test of Scaling Assumptions</title><p>To evaluate scaling assumptions underlying scoring of the SF-36 scales, item variability, item-internal consistency and item discriminant validity were assessed. Table <xref ref-type="table" rid="T2">2</xref> summarizes the results of the item descriptive statistics and the Pearson item-scale correlations between each item and scale. All of the correlations between each item and its hypothesized scale (i.e. item-scale internal consistency) corrected for overlap exceeded 0.40, ranging from 0.40 (MH1) to 0.68 (GH5). However, these are low for the SF scale. Item-scale correlations were roughly equivalent within each scale, although correlations were slightly lower for PF10 (bathing and dressing), GH4 (expect health to get worse), RE3 (didn't do work or activities as carefully as usual), and MH1 (nervous person). Item means and standard deviations generally were roughly equivalent within a scale, with some exceptions previously noted in other studies[<xref ref-type="bibr" rid="B19">19</xref>]. The mean values of VT4 (tired) and MH1 (nervous person) were lower than expected, however.</p><p>Test of item discriminant validity focus on the integrity of hypothesized item groupings relative to the health concepts hypothesized. According to the IQOLA protocol[<xref ref-type="bibr" rid="B7">7</xref>], an item was considered to have "succeeded" in the test of item discriminant validity if the correlation between an item and its hypothesized scale is statistically and significantly higher (i.e. > 2 standard errors) than the correlations between that item and all scales other than its hypothesized scale. All items passed the test for discriminant validity except the VT and SF subscales which had rates below 100%. Almost all the items in the VT and SF subscales overlapped with those in the MH subscale. In the vitality scale, VT4 which assessed the lack of energy had the same high correlations with the mental health scale (0.53) as its hypothesized scale. In the social functioning scale, SF2 was more highly correlated with the mental health and vitality scale (0.42 and 0.43, respectively) than its hypothesized scale.</p></sec><sec><title>Reliability</title><p>Internal consistency reliability statistics for the eight SF-36 scales are presented on the diagonal of Table <xref ref-type="table" rid="T3">3</xref>. All scales met or exceeded the 0.70 level recommended for group comparisons, with the exception of the SF scale (Cronbach's alpha = 0.57). Inter-scale correlation analysis also revealed that the scale constructs for the translated Taiwan SF-36 version were generally distinct. Most of the inter-scale correlation coefficients were medium to low, and higher coefficients were found between scales which represented similar constructs (e.g. vitality and mental health) than those with competing constructs (e.g. role emotional vs. physical functioning). However, the correlation between the MH and VT scales (0.69) was nearly as high as the reliability of the two scales.</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Reliabilty Statistics and Correlations between SF-36 Scales</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Scale</td><td align="center">PF</td><td align="center">RP</td><td align="center">BP</td><td align="center">GH</td><td align="center">VT</td><td align="center">SF</td><td align="center">RE</td><td align="center">MH</td></tr></thead><tbody><tr><td align="left">PF – Physical Functioning</td><td align="center">(0.83)</td><td></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">RP – Role–Physical</td><td align="center">0.47</td><td align="center">(0.79)</td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">BP – Bodily Pain</td><td align="center">0.39</td><td align="center">0.39</td><td align="center">(0.70)</td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">GH – General Health</td><td align="center">0.34</td><td align="center">0.40</td><td align="center">0.43</td><td align="center">(0.78)</td><td></td><td></td><td></td><td></td></tr><tr><td align="left">VT – Vitality</td><td align="center">0.25</td><td align="center">0.26</td><td align="center">0.31</td><td align="center">0.51</td><td align="center">(0.75)</td><td></td><td></td><td></td></tr><tr><td align="left">SF – Social Functioning</td><td align="center">0.26</td><td align="center">0.37</td><td align="center">0.38</td><td align="center">0.35</td><td align="center">0.44</td><td align="center">(<underline>0.57</underline>)</td><td></td><td></td></tr><tr><td align="left">RE – Role–Emotional</td><td align="center">0.18</td><td align="center">0.34</td><td align="center">0.23</td><td align="center">0.25</td><td align="center">0.33</td><td align="center">0.41</td><td align="center">(0.75)</td><td></td></tr><tr><td align="left">MH – Mental Health</td><td align="center">0.17</td><td align="center">0.17</td><td align="center">0.27</td><td align="center">0.43</td><td align="center">0.69</td><td align="center">0.49</td><td align="center">0.38</td><td align="center">(0.78)</td></tr></tbody></table><table-wrap-foot><p>Note. Scale reliability was represented on the diagonal. Internal consistency reliability lower than recommended level of 0.70 were shown with underlined entries.</p></table-wrap-foot></table-wrap></sec><sec><title>Principal Component Analysis (PCA)</title><p>The SF-36 was constructed to represent two major dimensions of health – physical and mental. To test this in the Taiwan data, two factors with eigenvalues greater than 1 were extracted, which accounted for 60% of the total variance in SF-36 scale scores. As shown in Table <xref ref-type="table" rid="T4">4</xref>, correlations between SF-36 scales and the two components generally were consistent with hypotheses in Taiwan. The PF, RP, and BP scales had the highest correlations with the physical component (r = 0.80, 0.80, and 0.64 respectively) and the lowest correlations with the mental component. In addition, the MH scale and SF scale had moderate associations with the mental component (r = 0.90 and 0.61, respectively) as anticipated.</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Factor Loadings of the 8 SF-36 scales using rotated principal components in Taiwan, Japan, Hong Kong and the US</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td align="center" colspan="10">Rotated Principal Components <sup>a</sup></td></tr></thead><tbody><tr><td></td><td align="center" colspan="5">Physical</td><td align="center" colspan="5">Mental</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="left">Scale</td><td align="center">Expected</td><td align="center">Taiwan</td><td align="center">Japan</td><td align="center">HK</td><td align="center">US</td><td align="center">Expected</td><td align="center">Taiwan</td><td align="center">Japan</td><td align="center">HK</td><td align="center">US</td></tr><tr><td colspan="11"><hr></hr></td></tr><tr><td align="left">PF</td><td align="center">●</td><td align="center">0.80</td><td align="center">0.75</td><td align="center">0.78</td><td align="center">0.85</td><td align="center">○</td><td align="center">0.09</td><td align="center">0.17</td><td align="center">-0.08</td><td align="center">0.12</td></tr><tr><td align="left">RP</td><td align="center">●</td><td align="center">0.80</td><td align="center">0.86</td><td align="center">0.63</td><td align="center">0.81</td><td align="center">○</td><td align="center">0.19</td><td align="center">0.19</td><td align="center">0.34</td><td align="center">0.27</td></tr><tr><td align="left">BP</td><td align="center">●</td><td align="center">0.64</td><td align="center">0.51</td><td align="center">0.61</td><td align="center">0.76</td><td align="center">○</td><td align="center">0.28</td><td align="center">0.52</td><td align="center">0.23</td><td align="center">0.28</td></tr><tr><td align="left">GH</td><td align="center">◐</td><td align="center">0.46</td><td align="center">0.37</td><td align="center">0.70</td><td align="center">0.69</td><td align="center">◐</td><td align="center">0.56</td><td align="center">0.66</td><td align="center">0.10</td><td align="center">0.37</td></tr><tr><td align="left">VT</td><td align="center">◐</td><td align="center">0.16</td><td align="center">0.21</td><td align="center">0.48</td><td align="center">0.47</td><td align="center">◐</td><td align="center">0.84</td><td align="center">0.88</td><td align="center">0.51</td><td align="center">0.64</td></tr><tr><td align="left">SF</td><td align="center">◐</td><td align="center">0.38</td><td align="center">0.45</td><td align="center">0.20</td><td align="center">0.42</td><td align="center">●</td><td align="center">0.61</td><td align="center">0.60</td><td align="center">0.74</td><td align="center">0.67</td></tr><tr><td align="left">RE</td><td align="center">○</td><td align="center">0.30</td><td align="center">0.69</td><td align="center">0.05</td><td align="center">0.17</td><td align="center">●</td><td align="center">0.54</td><td align="center">0.34</td><td align="center">0.74</td><td align="center">0.78</td></tr><tr><td align="left">MH</td><td align="center">○</td><td align="center">0.02</td><td align="center">0.13</td><td align="center">0.10</td><td align="center">0.17</td><td align="center">●</td><td align="center">0.90</td><td align="center">0.89</td><td align="center">0.77</td><td align="center">0.87</td></tr></tbody></table><table-wrap-foot><p>● Strong association (r ≥ .70) ◐ Moderate to substantial association (.30 < r < .70) ○ Weak association (r ≤ .30) <sup>a </sup>Correlations between each scale and rotated principal component. Source: Data for US is from Ware, J., Snow, K., Kosinski, M., & Gandek, B. (1993). <italic>SF-36 Health Survey: Manual & Interpretation Guide</italic>. MA: Boston: Nimrod Press. Data for Japan is from Fukuhara S, Ware JE, Kosinski M, Wada S, Gandek B (1998). Psychometric and Clinical Tests of Validity of the Japanese SF-36 Health Survey. <italic>J Clin Epidemiol, 51</italic>, 1998; 51: 1045–1053 Data for HK is from Lam, C. L., Gandek, B., Ren, X. S., & Chan, M. S. (1998). Tests of scaling assumptions and construct validity of the Chinese (HK) version of the SF-36 Health Survey. <italic>J Clin Epidemiol, 51</italic>, 1139–1147.</p></table-wrap-foot></table-wrap><p>However, correlations between the two components and the three other SF-36 scales were less consistent with hypotheses. The RE scale did not show as strong an association with the mental component (r = 0.54) in Taiwan as observed in Western countries. The GH and VT scales, originally hypothesized to measure both physical and mental health components, appeared to represent more a mental health concept (r = 0.56 and 0.84 respectively) than a physical one (r = 0.46 and 0.16), particularly for the Vitality scale.</p><p>Cross-cultural comparisons of the principal component analysis for Taiwan, Japan, Hong Kong, and the United States are also presented in Table <xref ref-type="table" rid="T4">4</xref>. Consistent with results in previous studies, the pattern of correlations was similar across countries for the PF, RP, SF, and MH scales. For the BP scale, Taiwanese results appear to be similar to those from the US, but the BP scale is a little less physical. As evidenced by the correlation patterns, the GH and VT scales in Taiwan and Japan represent more of a mental concept than a physical one, which is different from US results. The RE scale represents more a mental concept than a physical one in Taiwan, as opposed to the Japanese study, but the association is not as strong as that observed in US.</p></sec><sec><title>Validity</title><p>Construct validity was examined by comparing SF-36 scores to the Stress Coping Inventory, which was aimed to assess a mental health construct. As expected, the results (Table <xref ref-type="table" rid="T5">5</xref>) suggest that SF-36 score profiles correlate with those of the Stress Coping Inventory in expected ways. All subscales of the SF-36 mental health dimension except role-emotional, correlated more highly with related constructs measured by the Stress Coping Inventory than did those of the SF-36 physical health dimension. The results of the correlation analysis suggested an acceptable level of convergent validity for the SF-36 Taiwan version.</p><table-wrap position="float" id="T5"><label>Table 5</label><caption><p>Correlations between the SF-36 Health Survey and the Stress Coping Inventory (N = 569)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Stress Coping Inventory SF-36</td><td align="center" colspan="3">Resources</td><td align="center" colspan="2">Psychological Symptoms</td></tr><tr><td></td><td colspan="5"><hr></hr></td></tr><tr><td></td><td align="center">Self-Esteem</td><td align="center">Family Support</td><td align="center">Friendship Support</td><td align="center">Anxiety</td><td align="center">Depression</td></tr></thead><tbody><tr><td align="left">PF – Physical Functioning</td><td align="center">0.29</td><td align="center">0.11</td><td align="center">0.22</td><td align="center"><underline>-0.31</underline></td><td align="center">-0.30</td></tr><tr><td align="left">RP – Role–Physical</td><td align="center">0.18</td><td align="center">0.13</td><td align="center">0.20</td><td align="center">-0.29</td><td align="center">-0.21</td></tr><tr><td align="left">BP – Bodily Pain</td><td align="center">0.15</td><td align="center">0.14</td><td align="center">0.17</td><td align="center">-0.28</td><td align="center">-0.26</td></tr><tr><td align="left">GH – General Health</td><td align="center"><underline>0.38</underline></td><td align="center">0.23</td><td align="center">0.29</td><td align="center"><underline>-0.39</underline></td><td align="center"><underline>-0.38</underline></td></tr><tr><td align="left">VT – Vitality</td><td align="center"><underline>0.53</underline></td><td align="center">0.26</td><td align="center"><underline>0.39</underline></td><td align="center"><underline>-0.49</underline></td><td align="center"><underline>-0.53</underline></td></tr><tr><td align="left">SF – Social Functioning</td><td align="center"><underline>0.32</underline></td><td align="center">0.20</td><td align="center">0.27</td><td align="center"><underline>-0.45</underline></td><td align="center"><underline>-0.36</underline></td></tr><tr><td align="left">RE – Role–Emotional</td><td align="center">0.18</td><td align="center">0.08</td><td align="center">0.15</td><td align="center"><underline>-0.41</underline></td><td align="center"><underline>-0.34</underline></td></tr><tr><td align="left">MH – Mental Health</td><td align="center"><underline>0.53</underline></td><td align="center">0.22</td><td align="center"><underline>0.42</underline></td><td align="center"><underline>-0.6</underline></td><td align="center"><underline>-0.57</underline></td></tr></tbody></table><table-wrap-foot><p>Note: All correlation coefficients in this table are statistically significant (p < 0.05). <underline>Underlined</underline> entries represent Moderate to Substantial associations (.30 < r < .70)</p></table-wrap-foot></table-wrap><p>Results of criterion-based validity conformed to the original hypothesis, i.e. subscales of the physical health dimension declined with increasing age, while those in the mental health dimension were less influenced by age. As shown in Figure <xref ref-type="fig" rid="F1">1</xref>, means of subscales in the physical health dimension (e.g. physical function and role physical) decreased with increasing age, while those in the mental health dimension (e.g. mental health or vitality) fluctuated less with age.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Comparison of SF-36 Scale Profiles between the Student and Elderly Sample.</p></caption><graphic xlink:href="1477-7525-1-72-1"/></fig></sec></sec><sec><title>Discussion</title><p>One important issue in the current study is to determine whether the SF-36 measurement model can be applied in Taiwan. In general, the findings of the current study provide evidence that the concepts embodied in the SF-36 can be conveyed to the Taiwanese people and are feasible to be applied in Taiwan. Most tests of the psychometric properties of the SF-36 Taiwan version were satisfactory according to criteria set by the IQOLA project protocol, suggesting the feasibility of the translated version of the SF-36 for use in Taiwan. Data quality was high across the three study samples. The percentage of missing data ranged from less than 0% to 2.7% at an item level. These rates compare favorably with those reported in the original Medical Outcomes Study in the US [<xref ref-type="bibr" rid="B18">18</xref>] and other Western countries [<xref ref-type="bibr" rid="B20">20</xref>]. Results of the multitrait scaling analysis basically supported the hypothesized scale structure of the SF-36 in Taiwan and indicated that standard scoring algorithms could be used to score the eight SF-36 scales. The ordering of item means within scales generally were clustered within scales as hypothesized, with two exceptions involving the "felt tired" (VT4) and "felt nervous" (MH1) item. Similar results of lower mean score than expected for the VT4 item were also found in several other countries[<xref ref-type="bibr" rid="B19">19</xref>]. Due to the uneven distribution of subjects in terms of age range in the present study, the results of lower means in the VT4 and MH1 item should be interpreted with caution. However, psychometric testing results of the present study also indicate specific areas of the Taiwanese SF-36 in which further refinement and work will be required. Of the SF-36 eight scales, internal consistency reliability was generally acceptable for group-level comparisons except for the Social Functioning scale. In the Taiwan version, the two SF items were correlated more highly with the MH scale than with their hypothesized scale. In addition, the SF scale had the lowest scaling success rate (87.5%). The suboptimal scaling performance of the SF scale has been observed and reported in a cross-cultural content comparison of SF-36 translations [<xref ref-type="bibr" rid="B21">21</xref>]. Due to cultural differences in the concept of social functioning, these items have been reported to be difficult to translate in some other countries [<xref ref-type="bibr" rid="B19">19</xref>]. The finding seems to suggest cultural differences in item interpretation. The concept of social functioning may be more westernized and less clear for Taiwanese people. Deeply ingrained in the Confuscian ideology of collectivism, it is culturally unacceptable for people in Taiwan to use health problems as an excuse to avoid family or social gatherings[<xref ref-type="bibr" rid="B8">8</xref>]. That is, the denial of disturbance of physical and emotional health on social activities is more salient among Asians than Americans. Therefore, a specific family functioning scale may need to be added to generic health status questionnaires used in Taiwan, to acknowledge the impact of health on family life in cultures in which family life may play a more central role in people's lives, one that is distinct from the roles friends and other contacts hold[<xref ref-type="bibr" rid="B21">21</xref>].</p><p>Differences in the SF-36 profile between young and old adults shows evidence of discriminant validity for the SF-36 Taiwan version. In the student sample, acceptable convergent validity was obtained by the results of the comparison of the SF-36 and the Stress Coping Inventory, in which higher correlation coefficients were obtained for scales measuring similar psychological constructs.</p><p>In many countries, the SF-36 has been shown to yield reliable scale scores measuring eight dimensions of health status, which have two underlying measures of physical and mental health [<xref ref-type="bibr" rid="B22">22</xref>]. In Taiwan, the results of a principal component analysis lend support to the two component models as hypothesized. However, some disparity in the pattern of correlations relative to United States was found. While primarily a physical scale, the Bodily Pain scale does not have as strong an association with the physical dimension as was found in the US. Although primarily a mental scale, the Role Emotional scale was not as purely associated with the mental component as in the US and Western Europe. However, the Role Emotional scale did have a higher loading on the mental factor than the physical factor in Taiwan. In contrast, the Role Emotional scale had a high loading on the physical factor and a low loading on the mental factor in Japan (although this finding did not hold for highly educated Japanese women) [<xref ref-type="bibr" rid="B17">17</xref>]. Differences in the factor structure of the Role Emotional scale between Asian and Western countries may reflect a reluctance to attribute limitations in role functioning to emotional states, particularly for the elderly who are less influenced by Western culture than the younger generation[<xref ref-type="bibr" rid="B23">23</xref>].</p><p>The VT and MH scales were highly intercorrelated, however, VT and MH items did have higher correlations with their hypothesized scales than all other scales. The Vitality scale had a high correlation with the mental component and a low correlation with the physical component in both Taiwan and Japan; these results contrasted with those from US, in which Vitality had a moderate to substantial association with both the mental and physical components. In addition, the Vitality scale in Taiwan was strongly correlated with scales that are related to the construct of mental health in the Stress Coping Inventory. Therefore, the Vitality scale appears to be less valid for measuring physical health in Taiwan than in the West. Such a relationship between Vitality and the Mental Health dimension has also been seen in other studies of Chinese-Americans[<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B24">24</xref>] Subjects in the abovementioned samples are all familiar with the culture of traditional Chinese medicine; hence the results may reflect how people perceive their vitality status within a cultural framework. In the traditional theory of Chinese medicine, the phrase ("JingShen") associated with the presence of vitality is used to describe "mental well-being" [<xref ref-type="bibr" rid="B25">25</xref>]. It is therefore not surprising for Ren et al[<xref ref-type="bibr" rid="B8">8</xref>] to conclude that "vitality is central to the concept of a healthy mental state for Chinese."</p><p>Some limitations of the current study should be kept in mind. The reliability of the tool is not fully established in the current study, future assessment of reproducibility and responsiveness would be necessary for this. In addition, the construct validity of the results reported in this article is mainly derived from the student sample. Further research is necessary to replicate the results in different age groups to validate the SF-36 Taiwan version.</p></sec><sec><title>Conclusions</title><p>In conclusion, we have provided empirical data to illustrate the feasibility of translating and validating the SF-36 in an Asian country. The Taiwan version of the SF-36 Health Survey appears to be a practical and reliable instrument in the general population. The finding that the Vitality scale is strongly associated with the mental health component is interpreted in a cultural framework. Previous studies have also found this pattern and raise important questions regarding cultural influences upon illness attribution and perception. Further research into conceptualization of the Vitality and/or Role Emotional scale among Asian countries within a cultural framework is warranted.</p></sec><sec><title>Authors' contributions</title><p>HMT carried out and participated in the design of the studies and performed the statistical analyses, as well as drafted the manuscript. JRL carried out and finalized the translated Taiwan version of SF-36, and drafted the manuscript. BG participated in the translation process and in the manuscript preparation. All authors read and approved the final manuscript.</p></sec>
Cloning of a novel signaling molecule, AMSH-2, that potentiates transforming growth factor β signaling	<sec><title>Background</title><p>Transforming growth factor-βs (TGF-βs), bone morphogenetic proteins (BMPs) and activins are important regulators of developmental cell growth and differentiation. Signaling by these factors is mediated chiefly by the Smad family of latent transcription factors.</p></sec><sec><title>Results</title><p>There are a large number of uncharacterized cDNA clones that code for novel proteins with homology to known signaling molecules. We have identified a novel molecule from the HUGE database that is related to a previously known molecule, AMSH (<underline>a</underline>ssociated <underline>m</underline>olecule with the <underline>SH</underline>3 domain of STAM), an adapter shown to be involved in BMP signaling. Both of these molecules contain a coiled-coil domain located within the amino-terminus region and a JAB (Domain in <underline>J</underline>un kinase <underline>a</underline>ctivation domain <underline>b</underline>inding protein and proteasomal subunits) domain at the carboxy-terminus. We show that this novel molecule, which we have designated AMSH-2, is widely expressed and its overexpression potentiates activation of TGF-β-dependent promoters. Coimmunoprecipitation studies indicated that Smad7 and Smad2, but not Smad3 or 4, interact with AMSH-2. We show that overexpression of AMSH-2 decreases the inhibitory effect of Smad7 on TGF-β signaling. Finally, we demonstrate that knocking down AMSH-2 expression by RNA interference decreases the activation of 3TP-lux reporter in response to TGF-β.</p></sec><sec><title>Conclusions</title><p>This report implicates AMSH and AMSH-2 as a novel family of molecules that positively regulate the TGF-β signaling pathway. Our results suggest that this effect could be partially explained by AMSH-2 mediated decrease of the action of Smad7 on TGF-β signaling pathway.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Ibarrola</surname><given-names>Nieves</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Kratchmarova</surname><given-names>Irina</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Nakajima</surname><given-names>Daisuke</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected] </email></contrib><contrib id="A4" contrib-type="author"><name><surname>Schiemann</surname><given-names>William P</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Moustakas</surname><given-names>Aristidis</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A6" corresp="yes" contrib-type="author"><name><surname>Pandey</surname><given-names>Akhilesh</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A7" corresp="yes" contrib-type="author"><name><surname>Mann</surname><given-names>Matthias</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Cell Biology	<sec><title>Background</title><p>TGF-β ligands belong to a large and expanding family of multifunctional cytokines which include the activins, inhibins, bone morphogenetic proteins (BMPs), growth/differentiation factors, Mullerian inhibiting substance and Nodal. Members of this family regulate a broad range of physiological processes, including organogenesis, proliferation, differentiation, adhesion, motility and apoptosis [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Mutations in downstream components of the TGF-β signaling pathways have been implicated in several human diseases [<xref ref-type="bibr" rid="B3">3</xref>].</p><p>TGF-β superfamily members signal through a receptor complex formed by the association of two transmembrane proteins designated I and II, each possessing serine/threonine kinase activity (for reviews see [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]). TGF-β binding induces the assembly of the type I and type II receptors [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>], thereby enabling the type II receptor to phosphorylate and activate the type I receptor [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. Activated type I receptors then phosphorylate Smad2 and Smad3 receptor-regulated Smad proteins (R-Smads) at the carboxy terminal motif SSxS [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. Phosphorylated R-Smads interact with Smad 4, the unique Co-Smad [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>]. This complex accumulates in the nucleus where it regulates transcriptional activity of several genes in association with transcriptional activators and corepressors (for reviews see [<xref ref-type="bibr" rid="B14">14</xref>-<xref ref-type="bibr" rid="B17">17</xref>]).</p><p>The specific transcriptional outcome in response to TGF-β is achieved by the interplay of intracellular regulators at different levels. Within the Smad family of proteins, Smad6 and 7 (I-Smads) negatively regulate TGF-β signaling [<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B20">20</xref>]. Smad6 also inhibits BMP signaling, while activin signaling is negatively regulated by Smad7 [<xref ref-type="bibr" rid="B21">21</xref>]. In addition to the Smad proteins, other regulatory molecules work as adapter and/or scaffolding proteins, such as SARA, that mediates the access of Smads to activated receptors [<xref ref-type="bibr" rid="B22">22</xref>]. Finally, the specific outcome of TGF-β signaling is modulated by several other signaling pathways (for review see [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B23">23</xref>]).</p><p>The human unidentified gene-encoded large proteins database (HUGE) contains long cDNAs and is a valuable tool to discover new proteins, and to predict structural features with a possible functional relevance [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. Through the analysis of the cDNA clones contained in this database, we have identified a novel protein that is highly similar to AMSH (<underline>a</underline>ssociated <underline>m</underline>olecule with the <underline>SH</underline>3 domain of STAM) [<xref ref-type="bibr" rid="B26">26</xref>]. Systematic analysis using specific reporters for several pathways showed that this novel molecule, designated AMSH-2, positively regulates TGF-β signaling system but not the TNF signaling pathway.</p></sec><sec><title>Results and discussion</title><sec><title>Identification of a novel signaling molecule, AMSH-2</title><p>We have adopted a systematic strategy to identify and characterize cDNAs encoding novel proteins containing domains that have a role in signal transduction pathways [<xref ref-type="bibr" rid="B27">27</xref>-<xref ref-type="bibr" rid="B29">29</xref>]. We have used the HUGE database as a primary public resource as it contains mostly uncharacterized full-length cDNAs. By analyzing the domain composition of the predicted proteins, we identified a cDNA [Genbank <ext-link ext-link-type="gen" xlink:href="AB037794">AB037794</ext-link>] with an open reading frame of 461 amino acids (Figure <xref ref-type="fig" rid="F1">1A</xref>). The domain composition of the predicted protein was obtained by searching the SMART database [<xref ref-type="bibr" rid="B30">30</xref>] and revealed a coiled-coil domain spanning amino acids 149 to 176 in the N-terminus of the molecule and a JAB domain [<xref ref-type="bibr" rid="B31">31</xref>-<xref ref-type="bibr" rid="B33">33</xref>] (amino acids 268 to 394) located in the carboxy terminal region of the molecule (Figure <xref ref-type="fig" rid="F1">1B</xref>). A search of GenBank database revealed a similar molecule, AMSH, which had been identified on the basis of its ability to interact with STAM and showed a similar domain composition. Alignment of the novel protein against human AMSH revealed a 75% identity within the JAB domain, with an overall identity of 56% between the two molecules (Figure <xref ref-type="fig" rid="F1">1A</xref>). We have designated this novel protein AMSH-2 after its paralog AMSH. Upon searching the human genomic database, we found that AMSH-2 gene was located on 10q23.33. A detailed analysis of the genomic structure of the gene encoding for AMSH-2 revealed that the coding region spans 10 exons (Figure <xref ref-type="fig" rid="F1">1C</xref>). All the intron-exon boundaries followed the GT/AG rule for splice junctions [<xref ref-type="bibr" rid="B34">34</xref>].</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Sequence, domain architecture and genomic structure analysis of a novel protein AMSH-2. </bold>A) Alignment of AMSH family of proteins. Identical residues are shaded in yellow. B) The domain architecture of AMSH-2 was analyzed by SMART program. The coiled-coil region and the JAB domains are represented by green and red figures respectively. The numbers indicate the position of the domain within the protein sequence. C) Analysis of the genomic structure of the human gene encoding for AMSH-2. Only the coding exons are represented. Each exon is represented by a box with a length scaled to the size of the exon. The number of the amino acids encoded by each exon are indicated.</p></caption><graphic xlink:href="1471-2121-5-2-1"/></fig></sec><sec><title>Expression of AMSH-2</title><p>In order to determine the tissue expression of AMSH-2, we probed a Northern blot containing poly A+ mRNA from several human tissues with a radiolabeled AMSH-2 probe. We found a single band of approximately 2.3 kb which was detected in almost all the tissues examined (Figure <xref ref-type="fig" rid="F2">2A</xref>, left panel). Probing a cancer cell line northern blot revealed a band of similar size in most of the cell lines, including HL60, Hela S3, Molt-4, Raji, SW480 and A549 (Figure <xref ref-type="fig" rid="F2">2A</xref>, right panel).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>A) Northern blot analysis of AMSH-2 in different human tissues and cell lines. Each lane contains 2 μg of poly A+ mRNA. The top panels show the result of hybridization of the membranes with a specific probe for AMSH-2 mRNA. The lower panels show β-actin reprobing of the membranes. The sizes of the transcripts are indicated in kb on the left. B) Expression of AMSH-2 protein. Cell lysates from metabolically labeled 293T and HepG2 cells were immunoprecipitated with preimmune (PI) and AMSH-2 antiserum (I). A specific band for AMSH-2 can be observed in the lanes immunoprecipitated with AMSH-2 antiserum (indicated by an arrow). C) Expression of AMSH-2 Myc construct. 293T cells were transfected with the indicated constructs and metabolically labeled with S<sup>35</sup>. Cell lysates were immunoprecipitated with 9E10 monoclonal antibody and after SDS-polyacrylamide gel separation and blotting, the membrane was exposed. The arrow points a specific band corresponding to the Myc tagged AMSH-2 protein.</p></caption><graphic xlink:href="1471-2121-5-2-2"/></fig><p>We corroborated the mRNA distribution of AMSH-2 by performing a BLAST search against the EST database. We found ESTs corresponding to AMSH-2 mRNA in brain, testes, bone, pancreas, fetal liver, kidney, colon, stomach, bone marrow, placenta, breast, pectoral muscle, hypothalamus, ovary as well as in a number of cell lines of different origins, including T cell leukemia, NT2 neuronal and B-cell CLL [<xref ref-type="bibr" rid="B35">35</xref>]. These data indicate that, like AMSH [<xref ref-type="bibr" rid="B26">26</xref>], AMSH-2 mRNA is widely expressed.</p><p>We raised an antibody against the carboxy terminus of AMSH-2 and measured the expression of AMSH-2 protein in different cell lines. For this purpose, HepG2 and 293T cells lines were metabolically labeled with <sup>35</sup>S and cell lysates were immunoprecipitated with preimmune or anti-AMSH-2 antiserum. We observed a band with a molecular weight of approximately 56 kDa, which is in agreement with the predicted molecular weight of AMSH-2. This band was specific for AMSH-2 as it was absent in the samples immunoprecipitated with pre-immune serum (Figure <xref ref-type="fig" rid="F2">2B</xref>).</p></sec><sec><title>AMSH-2 enhances gene expression of TGF-β-dependent promoters</title><p>In order to study the functional role of AMSH-2, we constructed several epitope-tagged versions of AMSH-2. The entire open reading frame of AMSH-2 was subcloned by PCR into pEF mammalian expression vectors that provide a Myc or V5 epitopes at the C-terminus. Ectopic expression of AMSH-2/Myc was undertaken and confirmed by metabolic labeling of 293T cells, followed by immunoprecipitation with anti-Myc antibodies (Figure <xref ref-type="fig" rid="F2">2C</xref>).</p><p>We next examined the effect of overexpression of AMSH-2 on a number of signaling pathways. Since AMSH-2 is expressed in HepG2 and 293T cells, we chose these cell lines for our assays. Given the similarity between AMSH-2 and AMSH, and the ability to AMSH to regulate BMP signaling, we hypothesized that AMSH-2 may affect signaling by TGF-β. We first tested the effect of AMSH-2 overexpression on TGF-β signaling pathway using p3TP-lux reporter plasmid which is a standard reporter to evaluate TGF-β signaling [<xref ref-type="bibr" rid="B6">6</xref>]. This reporter construct contains the luciferase gene under the control three consecutive TPA response elements and a 96 bp fragment of the PAI-1 promoter that is regulated by TGF-β and is also highly inducible in response to TGF-β. Overexpression of AMSH-2 in HepG2 cells increased the level of luciferase activity driven by the 3TP promoter upon TGF-β stimulation as compared to cells transfected with vector alone (Figure <xref ref-type="fig" rid="F3">3A</xref>). As this promoter is derived from the Plasminogen activator inhibitor-1 (PAI-1) promoter, we also tested the effect on PAI-1 luciferase activity [<xref ref-type="bibr" rid="B36">36</xref>]. When we cotransfected AMSH-2 into HepG2 cells together with this reporter, a similar increase in the luciferase levels was observed indicating that AMSH-2 is likely a positive regulator of TGF-β signaling (Figure <xref ref-type="fig" rid="F3">3B</xref>).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>A) Effect of AMSH-2 on TGF-β induced transcriptional activity of 3TP driven luciferase. HepG2 cells were cotransfected with 3TP-lux reporter together with the empty vector or increasing amounts of AMSH-2 Myc construct. Twenty-four hours posttransfection, cells were serum-starved and left untreated (white bars) or stimulated (black bars) with 5 ng/ml TGF-β1 for 16–20 h before the luciferase and β-galactosidase activities were measured. B) Effect of AMSH-2 on TGF-β induced transcriptional activity of PAI-1 luciferase. HepG2 cells were cotransfected with PAI-1 reporter together with the empty vector or 0.9 μg/ml of AMSH-2 Myc construct. A similar procedure as described in A was followed. C) AMSH-2 does not stimulate NF-kB signaling pathway. 293 cells were cotransfected with ELAM luciferase reporter construct together with an empty vector, 0.9 μg/ml of AMSH-2 Myc construct or 0.9 μg/ml of TRAF-2, used as a positive control. Forty-eight hours posttransfection, luciferase and β-galactosidase activities were measured.</p></caption><graphic xlink:href="1471-2121-5-2-3"/></fig><p>To account for possible non-specific effects, we tested the effect of AMSH-2 overexpression on the TNF signaling pathway. We used an ELAM-luciferase reporter construct as a reporter for activation of this pathway. This reporter contains the ELAM-1 promoter (-730 to +52) and is well known to be induced by TNF stimulation [<xref ref-type="bibr" rid="B37">37</xref>]. We therefore cotransfected 293 cells with ELAM-luciferase and either AMSH-2, empty vector or TRAF2, a prototypical member of the TRAF family of adapter proteins known to potentiate TNF signaling pathway [<xref ref-type="bibr" rid="B38">38</xref>-<xref ref-type="bibr" rid="B40">40</xref>]. We did not observe any increase in the luciferase levels in AMSH-2 transfected cells as compared with the empty vector, although overexpression of TRAF2 led to a significant upregulation of the promoter (Figure <xref ref-type="fig" rid="F3">3C</xref>). Taken together, these results imply that the AMSH-2 regulates TGF-β signaling pathway.</p></sec><sec><title>AMSH-2 negates the inhibitory effect of Smad7 on TGF-β signaling</title><p>The ability of AMSH-2 to regulate TGF-β signaling could be explained by several mechanisms, including interaction with Smad2, Smad3, Smad4, or Smad7 [<xref ref-type="bibr" rid="B41">41</xref>]. To determine whether AMSH-2 interacts with these Smads, we coexpressed AMSH-2 with epitope tagged Smad constructs in 293T cells. As shown in Figure <xref ref-type="fig" rid="F4">4A</xref>, we observed an association between AMSH-2 and Smad7 but not with Smads 3 or 4 (Figure <xref ref-type="fig" rid="F4">4A</xref>). We also observed association between Smad2 and AMSH-2. These findings are similar to that observed for AMSH, since this molecule interacts strongly with Smad6 and 7; but differ in that AMSH does not with any R-Smads [<xref ref-type="bibr" rid="B41">41</xref>].</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>A) AMSH-2 interacts with Smad2 and Smad7, but not with Smad3 or Smad4. Vector or AMSH-2 V5 construct were coexpressed in 293T cells together with the Smad proteins as indicated on the top of the figure. Lysates were immunoprecipitated with anti-V5 antibody and western blotted with anti-Flag or anti-HA antibodies. To control for Smad and AMSH-2 protein expression part of the lysates were immunoprecipitated with anti-Flag, anti-HA or anti-V5 antibodies and western blotted with the same antibodies. B) Smad7 inhibitory effect on TGF-β signaling pathway is rescued by AMSH-2. HepG2 cells were cotransfected with 3TP-lux reporter together with the empty vector or 10 ng/ml of Smad7 with or without 0.3 μg/ml of AMSH-2 Myc construct. Twenty-four hours posttransfection, cells were serum-starved and left untreated (white bars) or stimulated (black bars) with 5 ng/ml TGF-β1 for 16–20 h before luciferase and β-galactosidase activities were measured.</p></caption><graphic xlink:href="1471-2121-5-2-4"/></fig><p>Because AMSH-2 interacts with Smad7, we also analyzed the effect of AMSH-2 on the ability of Smad7 to inhibit TGF-β signaling in HepG2 cells. As expected, overexpression of Smad7 attenuated the induction of p3TP-luciferase activity by TGF-β (Figure <xref ref-type="fig" rid="F4">4B</xref>). Interestingly, overexpression of AMSH-2 negated the inhibitory effect of Smad7 on p3TP expression stimulated by TGF-β (Figure <xref ref-type="fig" rid="F4">4B</xref>). Therefore AMSH-2 can rescue the inhibitory effects of Smad7 on TGF-β signaling pathway. Given the sequence similarity between AMSH and AMSH-2, it is indeed possible that AMSH-2 also interacts with Smad6. Smad6 has been shown to inhibit the TGF-β pathway [<xref ref-type="bibr" rid="B42">42</xref>] and AMSH is able to rescue the inhibitory effect of Smad6 on BMP signaling pathway [<xref ref-type="bibr" rid="B41">41</xref>]. If AMSH-2 is able to interact with Smad6, the effect of AMSH-2 on TGF-β signaling pathway could be explained by its interaction with both Smad7 and Smad6.</p></sec><sec><title>AMSH-2 is a positive regulator of TGF-β-dependent transcriptional activation</title><p>The use of small interfering RNAs (siRNA) has been proven a powerful tool to suppress the expression of specific proteins in mammalian cells [<xref ref-type="bibr" rid="B43">43</xref>,<xref ref-type="bibr" rid="B44">44</xref>]. We have shown that overexpression of AMSH-2 positively regulates TGF-β-dependent transcriptional activation of 3TP-lux reporter. To assess the effect of depletion of endogenous AMSH-2, we designed a specific siRNA to silence AMSH-2. This siRNA was cotransfected in HepG2 cells and the effect on the transcriptional activity of the 3TP driven luciferase reporter was analyzed. As a positive control we used an siRNA targeted against Smad7. Cotransfection of siRNA specific for AMSH-2 cause a decrease in the luciferase activity upon stimulation with TGF-β when compared with cells cotransfected with a commercially available randomly ordered RNA scrambled RNA. As expected, we observed an increase in luciferase activity in the cells transfected with a siRNA specific for Smad7 (Figure <xref ref-type="fig" rid="F5">5</xref>). These knock down experiments confirm that AMSH-2 positively regulates TGF-β signaling pathway.</p><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>AMSH-2 RNA interference has a negative effect on TGF-β signaling pathway. </bold>HepG2 cells were transfected with 3TP-lux reporter and scramble, laminA/C, AMSH-2 or Smad7 siRNA duplexes as indicated in the bottom of the graph. Forty-eight hours posttransfection cells were serum-starved and left treated (black bars) or not (white bars) with 5 ng/ml TGF-β1.</p></caption><graphic xlink:href="1471-2121-5-2-5"/></fig></sec></sec><sec><title>Conclusions</title><p>Our results indicate that AMSH-2 is a positive regulator of TGF-β signaling. This effect might be mediated through the interaction with the inhibitory Smad7 as AMSH-2 is capable of negating the inhibitory effects of Smad7 in TGF-β signaling pathway. Further, use of siRNA to knock down the expression of AMSH-2 reduces signaling by TGF-β. Similar to our studies, it has been suggested that AMSH might be involved in TGF-β signaling. AMSH can interact with both I-Smads (<italic>i. e</italic>. Smads 6 and 7) and is a positive regulator of BMP signaling pathway [<xref ref-type="bibr" rid="B41">41</xref>]. The high degree of homology between these paralogs suggests that AMSH-2 might also interact with Smad6, and therefore might regulate BMP signaling. Although further research needs to be conducted to elucidate the exact mechanism of action of AMSH-2, it is clear that AMSH and AMSH-2 belong to a new family of molecules that positively regulate TGF-β signaling pathway.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>cDNAs and constructs</title><p>A human cDNA clone [Genbank <ext-link ext-link-type="gen" xlink:href="AB037794">AB037794</ext-link>] was used as described [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. To generate wild type AMSH-2 expression vector with a carboxy-terminal Myc epitope, we subcloned AMSH-2 open reading frame into the Nco I and Xho I sites of pEF/myc/cyto (Invitrogen, Carsbad, CA) by standard PCR procedures using the primers: AAAACCATGGATCAGCCTTTTACTGTGAATTC (5' primer) and AAACTCGAGCTGTTCAGATGGTGATGATGAC (3' primer). To generate AMSH-2 V5 carboxy-terminal-tagged expression vector, the open reading frame of AMSH-2 was subcloned into Nco I and Xho I sites of pEntr4 (Invitrogen) using the same 5' primer as described above and AAACTCGAGAGCTGTTCAGATGGTGATGATGACCTAG (3' primer), and transferred to pEF1/V5-HisB (Invitrogen), which has been previously made gateway compatible by inserting a cassette in the EcoRV site.</p><p>3TP-lux, PAI-1-luc, pELAMP-luc+, TRAF2, Flag-tagged Smad7, Smad3, have been previously described [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B36">36</xref>-<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B45">45</xref>] while Flag-tagged Smad2 and HA-tagged Smad4 were generously provided by Joan Massague.</p></sec><sec><title>Northern blot analysis</title><p>Human tissue northern blot II and cancer cell line blot were obtained from CLONTECH (Palo Alto, CA). We isolated and radiolabeled an AMSH-2 cDNA (nucleotides 1088–2010) produced by digestion of AMSH-2 cDNA clone with HindIII and SphI. Hybridization signals were detected on a BAS-2000 bioimaging analyzer (Fuji film, Tokyo, Japan). After exposure, the blots were stripped and reprobed with beta-actin to control for equal amounts of PolyA+ RNA loading.</p></sec><sec><title>Cell culture and antibodies</title><p>293 and 293T cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM L-glutamine and penicillin/streptomycin. HepG2 cells were grown in MEM supplemented with 10% fetal bovine serum, non essential amino-acids, sodium pyruvate and penicillin/streptomycin.</p><p>The peptide corresponding to the C-terminus of human AMSH-2 was synthesized by Boston Biomolecules, Inc (Woburn, MA). The specific AMSH-2 rabbit polyclonal antibody (anti-AMSH-2) was raised at Covance Research Products Inc. (Denver, PA). The following antibodies were obtained: (i) c-Myc mAb 9E10 (Covance Research Products Inc);(ii) anti-V5 and anti-V5 HRP antibodies (Invitrogen); (iii) anti-Flag HRP (UPSTATE, Lake Placid, NY);and (v) anti-HA 12CA5, anti-HA HRP (Roche Diagnostics Corp, Indianapolis, IN).</p></sec><sec><title>Metabolic labeling and immunoprecipitation</title><p>To test endogenous expression of AMSH-2, 293T and HepG2 cells were metabolically labeled with <sup>35</sup>S methionine and <sup>35</sup>S cysteine and lysed in modified RIPA buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 1 mM EDTA, 1% NP-40, 0.25% sodium deoxycholate) with 1 mM sodium orthovanadate and protease inhibitors. Samples were incubated either with pre-immune serum or AMSH-2 antiserum. Immunoprecipitated proteins were resolved by SDS-PAGE. To test expression of the Myc-tagged AMSH-2 construct, cells were transfected with 15 μg of DNA using the standard calcium phosphate method. Twenty-four h postransfection, the cells were metabolically labeled for 16 h. and subsequently lysed and immunoprecipitated with anti-Myc antibody as above.</p></sec><sec><title>Luciferase assays</title><p>HepG2 cells were transfected with 0.5 μg of the corresponding luciferase reporter, 50 ng of β-galactosidase reporter and 4.5 μg of DNA. Twenty-four h postransfection, the cells were depleted from serum and stimulated for 16–20 h with 5 ng/ml of purified recombinant human TGF-β1 (R&Dsystems. Inc. Minneapolis, MN). Luciferase and β-galatosidase activities were measured according to manufacturer's instructions (Tropix, Bedford, MA). Measurements were corrected by β-galactosidase activity. The experiments were repeated at least three times.</p></sec><sec><title>Coimmunoprecipitation experiments</title><p>293T cells were cotransfected with 13 μg of AMSH-2 V5 construct and 2 μg of the corresponding Smad constructs. After 48 h, the cells were lysed in lysis buffer (150 mM NaCl, 50 mM Tris, pH 7.5, 1 mM EDTA, 1% NP-40) with 1 mM sodium orthovanadate and protease inhibitors. Lysates were incubated with anti-V5 antibody, and the resulting immunocomplexes were separated in SDS-PAGE. Blots were probed with the indicated HRP conjugated antibody.</p></sec><sec><title>siRNA experiments</title><p>To target AMSH-2 and Smad 7, we designed 21 nt siRNA duplexes: AACAATTCCTTGCTGAATGTA and AACCGCAGCAGTTACCCCATC, according to [<xref ref-type="bibr" rid="B44">44</xref>]. All siRNA duplexes including lamin A/C siRNA duplexes and scrambled siRNA were obtained from Dharmacon Research. Inc. (Lafayette, Colorado).</p><p>HepG2 cells grown in 6 well plates were transfected with 250 ng of 3TP-lux reporter, 50 ng of β-galactosidase reporter and 0.12 nmol of the corresponding siRNA duplex. After 48 h of transfection cells were depleted from serum and treated with 5 ng/ml TGF-β1. Luciferase and β-galactosidase activities were measured after 17 h stimulation.</p></sec></sec>
Segmentally Variable Genes:A New Perspective on Adaptation	<p>Genomic sequence variation is the hallmark of life and is key to understanding diversity and adaptation among the numerous microorganisms on earth. Analysis of the sequenced microbial genomes suggests that genes are evolving at many different rates. We have attempted to derive a new classification of genes into three broad categories: lineage-specific genes that evolve rapidly and appear unique to individual species or strains; highly conserved genes that frequently perform housekeeping functions; and partially variable genes that contain highly variable regions, at least 70 amino acids long, interspersed among well-conserved regions. The latter we term segmentally variable genes (SVGs), and we suggest that they are especially interesting targets for biochemical studies. Among these genes are ones necessary to deal with the environment, including genes involved in host–pathogen interactions, defense mechanisms, and intracellular responses to internal and environmental changes. For the most part, the detailed function of these variable regions remains unknown. We propose that they are likely to perform important binding functions responsible for protein–protein, protein–nucleic acid, or protein–small molecule interactions. Discerning their function and identifying their binding partners may offer biologists new insights into the basic mechanisms of adaptation, context-dependent evolution, and the interaction between microbes and their environment.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Zheng</surname><given-names>Yu</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Roberts</surname><given-names>Richard J</given-names></name><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Kasif</surname><given-names>Simon</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff3"> <sup>3</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>Microbes occupy almost every habitable niche in the biosphere, highlighting their enormous capability for adaptation and survival. This adaptive ability has been refined during millennia of evolution and has resulted in genes that evolve at very different rates. Some, such as housekeeping genes that code for the essential biochemical functions of the organism, are now evolving rather slowly. Others that have to defend against potentially lethal attack by viruses or toxins and adapt to varying environmental conditions, often evolve more rapidly (<xref rid="pbio-0020081-Murphy1" ref-type="bibr">Murphy 1993</xref>; <xref rid="pbio-0020081-Moxon1" ref-type="bibr">Moxon and Thaler 1997</xref>; <xref rid="pbio-0020081-Jordan1" ref-type="bibr">Jordan et al. 2002</xref>). Pathogenic microbes, for example, face stringent tests of their adaptive potential because of the escalating efficiency of the host-defense mechanisms (<xref rid="pbio-0020081-Moxon1" ref-type="bibr">Moxon and Thaler 1997</xref>). In the arms race between pathogens and their hosts, both sides try to improve their overall fitness by deploying sophisticated strategies to generate genetic variability (<xref rid="pbio-0020081-Elena1" ref-type="bibr">Elena and Lenski 2003</xref>). Sequence divergence during rapid evolution can take many forms. Some genes change throughout their entire sequences, resulting in apparently lineage-specific genes that lack clear similar sequences in current versions of GenBank. Others show a mosaic pattern of one or more variable regions interspersed within conserved regions. This latter group is the subject of this paper and we refer to them as segmentally variable genes (SVGs). For the purpose of the current analysis, we define such variable regions as having a minimum length of 70 amino acids, which would permit them to fold into independent domains. This distinguishes them from most nonfunctional interdomain segments, which are usually shorter and whose principal function depends on length rather than specific sequence content.</p><p>An example of an SVG family is provided by the cytosine-5 DNA methyltransferases (<xref rid="pbio-0020081-Posfai2" ref-type="bibr">Posfai et al. 1989</xref>). These enzymes typically form parts of restriction-modification systems, which are key components of an important bacterial defense mechanism to protect against phage attack and other unwanted infiltration of foreign DNA (<xref rid="pbio-0020081-Cheng2" ref-type="bibr">Cheng 1995</xref>). These methyltransferases catalyze the addition of a methyl group from S-adenosylmethionine to the 5-position of cytosine and contain a highly variable region of more than 90 amino acids that is responsible for specific DNA sequence recognition (<xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>A; <xref rid="pbio-0020081-Posfai2" ref-type="bibr">Posfai et al. 1989</xref>; <xref rid="pbio-0020081-Cheng2" ref-type="bibr">Cheng 1995</xref>; <xref rid="pbio-0020081-Lange1" ref-type="bibr">Lange et al. 1996</xref>). A detailed examination of the three-dimensional (3D) structure of the variable region suggests that it folds into an independent domain, which has been shown to bind to DNA (<xref rid="pbio-0020081-Cheng3" ref-type="bibr">Cheng et al. 1993</xref>). The flanking sequences are highly conserved because they are responsible for the chemistry of methylation, which is common to all members of the family. Variability in this family has arisen because there is a need for great variation in the DNA sequences being recognized so that the specific pattern of methylation becomes a marker to distinguish innate DNA from foreign DNA.</p><fig id="pbio-0020081-g001" position="float"><label>Figure 1</label><caption><title>Variability Profile for Typical SVGs</title><p>Blocks in the lines are conserved subsequences identified using the Pfam, BLOCKS, and PRINTS databases. In the variability profile, the x-axis is the amino acid position and the y-axis is the variability index (see Materials and Methods). Variable domains are marked by the black lines over the graph.</p><p>(A) Cytosine-specific DNA methyltransferase of 355 amino acid long in <named-content content-type="genus-species">H. pylori</named-content>. Notice the variable domain in the middle and the variable segment in its N-terminal region, which is shorter than 70 amino acids and has no known function.</p><p>(B) Virulence-associated protein homolog (VacB) of 644 amino acid long in <named-content content-type="genus-species">H. pylori</named-content>. It has two variable domains at the N- and C-termini.</p></caption><graphic xlink:href="pbio.0020081.g001"/></fig><p>To the best of our knowledge, there has been no systematic attempt to identify, catalog, and classify similar SVGs in the sequenced microbial genomes. Nor has any attempt been made to find potentially common functions among genes displaying this property. Since it is known that many genes involved in defense mechanisms, such as the DNA methyltransferases and the antigens exposed on the surface of bacteria, show such variability (<xref rid="pbio-0020081-Roche1" ref-type="bibr">Roche et al. 2001</xref>), it is tempting to speculate that one might identify host-defense genes based on this property. Thus, the regional variability might reflect the influence of diversifying selection pressure that could come from constant interaction with other fast-evolving molecules in the environment. Could such genes be the predominant members of the SVG families? Or do other genes, such as those involved in basic energy metabolism and synthesis, show similar variability? In this paper we provide an initial systematic analysis. We describe our findings about the distribution of SVGs and the potential function achieved by segmental variability.</p></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>Classification of Genes into Three Broad Groups</title><p>We carried out a classification of the genes in 43 fully sequenced microbial genomes (see <xref ref-type="supplementary-material" rid="st001">Table S1</xref> for a full name list). A Web site (<ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu)">http://geneva.bu.edu)</ext-link> is also provided with results for several selected genomes, including <italic>Escherichia coli, Helicobacter pylori, Neisseria meningitidis,</italic> and several others. Each gene is accompanied with schematic diagrams from Pfam (<xref rid="pbio-0020081-Bateman1" ref-type="bibr">Bateman et al. 2002</xref>), BLOCKS (<xref rid="pbio-0020081-Henikoff1" ref-type="bibr">Henikoff et al. 1999</xref>), PRINTS (<xref rid="pbio-0020081-Attwood1" ref-type="bibr">Attwood et al. 2003</xref>), and the nongapped BLAST (<xref rid="pbio-0020081-Altschul1" ref-type="bibr">Altschul et al. 1990</xref>) analyses.</p><p>For each genome, the full proteome is compared with the nonredundant GenBank sequence set using nongapped BLAST (<xref ref-type="sec" rid="s4">see Materials and Methods</xref> for the parameters used). Based on the degree of conservation or divergence among similar genes in different species, we classify them into three broad groups. Lineage-specific genes are defined as genes with no significantly similar hits from other species in the current GenBank (<italic>E</italic>-value cutoff, 1<italic>E</italic>-5). SVGs are defined as genes containing at least one highly variable region, containing more than 70 amino acids, interspersed among well-conserved regions. In any single SVG family, the length of the variable region can differ only within a certain range (<xref ref-type="sec" rid="s4">see Materials and Methods</xref> for more details). In this paper, regions are considered to be variable if no sequence similarity can be detected against possible homologous genes, where the overall homology is determined by the conserved portions. The rest of the genes in the genome are considered as fully conserved genes. Naturally, this initial soft classification is somewhat dependent on specific thresholds and will be biased by the current state of GenBank and the quality of the annotation.</p><p>In <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref> we show a scatter plot of the three classes of genes in the <named-content content-type="genus-species">H. pylori</named-content> genome in two-dimensional (2D) space, where the x-axis shows the length of the variable region and the y-axis shows the number of possible homologs of each gene. Lineage-specific genes (filled square in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>) by definition naturally cluster on the x-axis. Most of the genes in this group are still annotated as unknown. A few genes with annotated functions in this group, such as the outer-membrane protein family in <named-content content-type="genus-species">H. pylori</named-content> (<xref rid="pbio-0020081-Tomb1" ref-type="bibr">Tomb et al. 1997</xref>), only appear in this organism and contribute to its unique biology. A second group contains fully conserved genes (filled triangle in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>) with only short variable regions. It is in this class that most “housekeeping” genes fall. Examples include the subunits of ATP synthetase F1 (atpD, atpA, atpG) and ribosomal proteins such as rps4 (<xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>), etc. The third group contains the SVGs (filled diamond in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>). A few examples in this group are labeled with their names in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref> and will be discussed later. In <xref ref-type="table" rid="pbio-0020081-t001">Table 1</xref> we list the number of genes in each category for a representative set of microbial genomes (see <xref ref-type="supplementary-material" rid="st001">Table S1</xref> for a full list).</p><fig id="pbio-0020081-g002" position="float"><label>Figure 2</label><caption><title>Classification of Three Groups of Genes from a Single Genome, <named-content content-type="genus-species">H. pylori</named-content>, in 2D Space</title><p>The x-axis is the length of the variable region and the y-axis is the number of possible homologs a gene has from a BLAST search. The variable region length for a lineage-specific gene is defined as the length of the gene so that they naturally cluster onto the x-axis. Multiple variable regions in one gene are represented separately.</p></caption><graphic xlink:href="pbio.0020081.g002"/></fig><table-wrap id="pbio-0020081-t001" position="float"><label>Table 1</label><caption><title>Classification of Genes into Three Broad Categories for a Representative Set of Microbial Genomes</title></caption><graphic xlink:href="pbio.0020081.t001"/><table-wrap-foot><fn id="nt101"><p>See Table S1 for the entire table</p></fn></table-wrap-foot></table-wrap><p>SVGs are subdivided into different types depending on whether they have one, two, or more variable regions. The number of genes with a single variable region is much larger than the number of genes with multiple ones. In <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>A we show the variation profile of an SVG containing one variable region. The variation profile is displayed together with conserved subsequences identified using the Pfam (<xref rid="pbio-0020081-Bateman1" ref-type="bibr">Bateman et al. 2002</xref>), BLOCKS (<xref rid="pbio-0020081-Henikoff1" ref-type="bibr">Henikoff et al. 1999</xref>), and PRINTS (<xref rid="pbio-0020081-Attwood1" ref-type="bibr">Attwood et al. 2003</xref>) databases. This gene is the cytosine-specific DNA methyltransferase, M.HpyAVIB, from <named-content content-type="genus-species">H. pylori</named-content>. The variability lies in its DNA recognition domain (approximately 140 amino acids), which in this case recognizes the DNA sequence CCTC (<xref rid="pbio-0020081-Lin1" ref-type="bibr">Lin et al. 2001</xref>). In <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>B we give an example with two variable regions. It is the virulence-associated protein homolog VacB from <named-content content-type="genus-species">H. pylori</named-content>, which has variable regions at both its N-terminus (approximately 200 amino acids) and C-terminus (approximately 100 amino acids). <italic>VacB</italic> has been shown to encode a 3′–5′ exoribonuclease and is necessary for expression of virulence (<xref rid="pbio-0020081-Cheng4" ref-type="bibr">Cheng and Deutscher 2002</xref>). The conserved central region (approximately 400 amino acids (Pfam domain: RNB) defines a group of homologs distributed in a number of microbial genomes (<xref rid="pbio-0020081-Zuo1" ref-type="bibr">Zuo and Deutscher 2001</xref>). Note that the C-terminal region is variable, and its <named-content content-type="genus-species">E. coli</named-content> homolog contains RNA-binding motifs (<xref rid="pbio-0020081-Zuo1" ref-type="bibr">Zuo and Deutscher 2001</xref>). Although the detailed physiological roles of VacB remain unknown (<xref rid="pbio-0020081-Cheng4" ref-type="bibr">Cheng and Deutscher 2002</xref>), the variable regions may contribute to the determination of substrate specificity of VacB in the RNA quality-control process that eliminates defective ribosomal RNA (rRNA) molecules in different species.</p><p>The number of SVGs increases as genome sizes vary, from 0.5 MB<italic>(Mycoplasma genitalium</italic>) to 8.6 MB<italic>(Streptomyces coelicolor</italic>) (<xref ref-type="table" rid="pbio-0020081-t001">Table 1</xref>). For most microorganisms included, the proportion of SVGs varies in the range of 10%–20%. The number of lineage-specific genes, on the other hand, does not appear to correlate with the genome size. Instead, it is influenced by the content of the database. For instance, a “minimal” genome, <named-content content-type="genus-species">M. genitalium</named-content>, has a relatively high content of SVGs (20%) and a low percentage of lineage-specific genes (0.2%). However, when a closely related species, <named-content content-type="genus-species">M. pneumoniae</named-content>, is excluded from the database, its proportion of lineage-specific genes rises to 14%, while the proportion of SVGs remains unchanged. In general, the genomic proportion of SVGs is less affected by the database content.</p></sec><sec id="s2b"><title>Case Studies of SVGs and Functional Implication of Variability</title><p>In the following sections, we have selected several SVG families to demonstrate the functional implication of segmental variability.</p><sec id="s2b1"><title>Outer-membrane signal transduction genes/sensor histidine kinases</title><p>In prokaryotes, two-component signal-transducing systems are common and consist of a histidine kinase (HK) and a response regulator. Most HKs are membrane-bound, homodimeric proteins with an N-terminal periplasmic sensing domain and a C-terminal cytoplasmic kinase domain. HKs usually possess a highly variable sensing domain (usually over 150 amino acids), while the cytoplasmic kinase domain is quite conserved. By diversifying the sensing domain, microorganisms can develop different two-component modules to respond to different signals and interact with small molecules from the exterior. <xref ref-type="fig" rid="pbio-0020081-g003">Figure 3</xref> displays the distance matrix calculated from the sensing domains and the kinase domains from a group of highly similar HK genes. As shown in <xref ref-type="fig" rid="pbio-0020081-g003">Figure 3</xref>, sensing domains are much more diverse than the kinase domains. Moreover, the two regions show distinct clustering patterns, of which only the one for the conserved kinase domains is close to the phylogenetic relationship inferred from 16S rRNA sequences (data not shown). Significant homologies in the sensing regions can only be found in closely related species (e.g., <named-content content-type="genus-species">Ralstonia solanacearum</named-content> [Rs] and <named-content content-type="genus-species">Ralstonia metallidurans</named-content> [Rm] in <xref ref-type="fig" rid="pbio-0020081-g003">Figure 3</xref>), suggesting rapid divergence after speciation. Other sensor genes involved in cell motility<italic>,</italic> e.g., genes encoding methyl-accepting chemotactic protein (MCP) (see <italic>tlpA, tlpC</italic> in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>), are also highly variable in their N-terminal domains. In several bacteria<italic>,</italic> e.g., <named-content content-type="genus-species">Vibrio cholerae</named-content>, there is a greater number of segmentally variable MCP genes (approximately 40) than in other genomes (see the gene list of <named-content content-type="genus-species">V. cholerae</named-content> at <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu)">http://geneva.bu.edu)</ext-link>, which must correspond to its expanded ability to detect different chemical signals and find favorable environments. Although a few conserved motifs have been detected in the sensing region (<xref rid="pbio-0020081-Galperin2" ref-type="bibr">Galperin et al. 2001</xref>), the exact sensing signals for most prokaryotic HKs are unknown.</p><fig id="pbio-0020081-g003" position="float"><label>Figure 3</label><caption><title>2D Representation of the Distance Matrix Computed from the Variable and Conserved Domains in a Group of Similar HKs</title><p>The upper triangle shows the variable domains, the lower one the conserved domains. Amino acid sequence distances are calculated by the PROTDIST program using the Dayhoff PAM matrix. The sequence from each species is the best match (<italic>E</italic>-value < 1<italic>E</italic>-10) in that genome to the query <named-content content-type="genus-species">E. coli</named-content> gene. Abbreviations for organisms: Ec, <named-content content-type="genus-species">Escherichia coli</named-content> K12; Ps, <named-content content-type="genus-species">Pseudomonas syringae</named-content> pv. syringae B728a; Rm, <named-content content-type="genus-species">Ralstonia metallidurans</named-content>; Rs, <named-content content-type="genus-species">Ralstonia solanacearum</named-content>; Li, <named-content content-type="genus-species">Listeria innocua</named-content>; Tm, <named-content content-type="genus-species">Thermotoga maritime</named-content>; Ml, <named-content content-type="genus-species">Mycobacterium leprae</named-content>; Mt, <named-content content-type="genus-species">Mycobacterium tuberculosis</named-content> CDC1551; No, <italic>Nostoc</italic> sp. PCC 7120; Ef, <named-content content-type="genus-species">Enterococcus faecalis</named-content>; Bs, <named-content content-type="genus-species">Bacillus subtilis</named-content>; Ne, <named-content content-type="genus-species">Nitrosomonas europaea</named-content>; Sy, <italic>Synechococcus</italic> sp. PCC 7942; At, <named-content content-type="genus-species">Agrobacterium tumefaciens</named-content>. The PROTDIST program is included in the PHYLIP software package version 3.5 (<xref rid="pbio-0020081-Felsenstein1" ref-type="bibr">Felsenstein 1989</xref>).</p></caption><graphic xlink:href="pbio.0020081.g003"/></fig></sec><sec id="s2b2"><title>Transporter genes and outer-membrane proteins</title><p>The biggest family of SVGs is cell envelope-related, including the ATP-binding cassette transporters (ABC transporters), outer-membrane proteins, and virulence-related gene products. For membrane proteins, since part of their sequences are exposed to the outside of the cell and interact directly with the environment, one might hypothesize that the variable portions have evolved rapidly to deal with the changing environmental conditions.</p><p>ABC transporters are essential for microorganisms because they import nutrients into the cell and export noxious substances and toxins out of the cell. A typical ABC transporter gene in a prokaryote genome has a conserved ATPase domain (approximately 150 amino acids) and a large (over 300 amino acids) variable integral membrane domain. Two examples from this group are the multidrug-resistance genes <italic>hetA</italic> and <italic>spaB</italic> shown in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 2</xref>. It is known that substrates interact with the specific binding sites inside the membrane domain (<xref rid="pbio-0020081-Holland1" ref-type="bibr">Holland and Blight 1999</xref>), which suggests that the variability in the membrane domain may have to do with substrate selectivity or with different transport kinetics. Moreover, outer-membrane transporters are binding targets for bacteriophages and bacterial toxins. For example, the vitamin B12 transporter BtuB (614 amino acids) is the receptor for bacteriophage BF23 and E-colicin (<xref rid="pbio-0020081-Bradbeer1" ref-type="bibr">Bradbeer et al. 1976</xref>; <xref rid="pbio-0020081-Mohanty1" ref-type="bibr">Mohanty et al. 2003</xref>). The crystal structure of BtuB in <named-content content-type="genus-species">E. coli</named-content> has been solved (<xref rid="pbio-0020081-Chimento1" ref-type="bibr">Chimento et al. 2003</xref>). The variable region in <named-content content-type="genus-species">E. coli</named-content> BtuB overlaps with the 22-strand β-barrel (position 150–360), while the N-terminal hatch domain (position 6–132) and the extreme C-terminal TonB-box domain (position 550–614) are conserved among many homologs (<xref ref-type="supplementary-material" rid="sg001">Figure S1</xref>). The extracellular loops between contiguous strands in the β-barrel are displayed outside the cell (<xref rid="pbio-0020081-Chimento1" ref-type="bibr">Chimento et al. 2003</xref>) and possibly serve as receptor sites for bacteriophages and toxins. The variability in these loops may be driven by attempts to defend against bacteriophages and interaction with different bacterial toxins.</p></sec><sec id="s2b3"><title>DNA/RNA-processing enzymes</title><p>DNA/RNA processing enzymes form another large family of SVGs. Characteristic examples are the restriction and modification enzymes, where the DNA methylases have a variable region designed for DNA sequence recognition (<xref rid="pbio-0020081-Cheng2" ref-type="bibr">Cheng 1995</xref>) and the restriction enzymes are almost completely variable. Here we discuss two other genes: DNA gyrase B (<italic>gyrB</italic>) and DNA topoisomerase A (<italic>topA</italic>), whose competing actions control the degree of DNA supercoiling (<xref rid="pbio-0020081-Tse-Dinh1" ref-type="bibr">Tse-Dinh et al. 1997</xref>). Schematic alignments anchored by the conserved motifs from the BLOCKS database (<xref rid="pbio-0020081-Henikoff1" ref-type="bibr">Henikoff et al. 1999</xref>) for both enzymes are shown in <xref ref-type="fig" rid="pbio-0020081-g004">Figure 4</xref>. The variable region in GyrB is an additional approximately 160 amino acids long segment that is only present in the gram-negative eubacteria (<xref ref-type="fig" rid="pbio-0020081-g004">Figure 4</xref>B). Experiments probing the role of this region in <named-content content-type="genus-species">E. coli</named-content> GyrB have demonstrated its involvement in DNA binding, although the detailed function is unknown (<xref rid="pbio-0020081-Chatterji1" ref-type="bibr">Chatterji et al. 2000</xref>). We suspect that variability in this inserted domain may determine the specificity of the interaction between GyrB and DNA or suggest interaction with other molecules. It is intriguing to see that other gyrases lacking this region are also functional.</p><fig id="pbio-0020081-g004" position="float"><label>Figure 4</label><caption><title>Schematic Alignment of TopA and GyrB</title><p>(A) TopA. (B) GyrB. Each line represents a sequence. Black boxes indicate the conserved blocks from the BLOCKS database and are aligned correspondingly. Red boxes in (A) are the zinc-finger motifs reported by Pfam. Notice that the number of occurrences of this motif varies and that there are several sequences without this motif in the C-terminal. The lines between the boxes are the variable sequences that cannot be aligned. Variable domains are labeled in the figure.</p></caption><graphic xlink:href="pbio.0020081.g004"/></fig><p>For TopA, the N-terminal region of approximately 600 amino acids shows extensive sequence similarity while the C-terminal region (over 100 amino acids) is variable both in sequence content and in length (<xref ref-type="fig" rid="pbio-0020081-g004">Figure 4</xref>A). The conserved N-terminal region of TopA has the catalytic function of relaxing negatively supercoiled DNA (<xref rid="pbio-0020081-Feinberg1" ref-type="bibr">Feinberg et al. 1999</xref>). The variable C-terminus of TopA sometimes contains multiple copies of zinc-binding motifs, although there are a few exceptions, e.g., TopA in <named-content content-type="genus-species">Mycobacterium tuberculosis</named-content> (<xref ref-type="fig" rid="pbio-0020081-g004">Figure 4</xref>A). Interestingly, there are two copies of TopA in <named-content content-type="genus-species">H. pylori</named-content> 26695; one has three zinc-binding motifs in C-terminal region and the other does not. The zinc-binding motifs in <named-content content-type="genus-species">E. coli</named-content> TopA are shown to be involved in the interaction with the β′ subunit of RNA polymerase (<xref rid="pbio-0020081-Cheng1" ref-type="bibr">Cheng et al. 2003</xref>) and in DNA binding (<xref rid="pbio-0020081-Ahumada1" ref-type="bibr">Ahumada and Tse-Dinh 1998</xref>). Since RNA polymerase β′ subunit is a fully conserved gene, the overall sequence variation in the C-terminal region of TopA seems more likely to relate to DNA binding. TopA plays an important role in adaptation to environmental challenges, such as heat shock conditions (<xref rid="pbio-0020081-Tse-Dinh1" ref-type="bibr">Tse-Dinh et al. 1997</xref>). Deletion experiments show that in <named-content content-type="genus-species">E. coli</named-content> the C-terminal region is important for the in vivo function of TopA during the osmotic stress response (<xref rid="pbio-0020081-Cheng1" ref-type="bibr">Cheng et al. 2003</xref>). All together, these facts suggest a versatile role that the C-terminal region of TopA might play in those processes.</p><p>Variable regions are sometimes found in DNA processing enzymes with essential and conserved functions. One example is DNA polymerase I, which has a variable region between the conserved C-terminal 5′–3′ polymerase domain and the N-terminal 5′–3′ exonuclease domain. In some polymerases, this region encodes a 3′–5′ exonuclease activity for proofreading replication errors, and conserved motifs can be observed (<xref rid="pbio-0020081-Derbyshire1" ref-type="bibr">Derbyshire et al. 1995</xref>). However, other polymerases in the same family that lack such proofreading activity show much sequence divergence in this region (<xref rid="pbio-0020081-Derbyshire1" ref-type="bibr">Derbyshire et al. 1995</xref>). The exact reason why sequence variability is observed in these polymerases is unknown.</p><p>Another interesting family is the aminoacyl-tRNA synthetases (AARS) (<xref rid="pbio-0020081-Ibba1" ref-type="bibr">Ibba and Söll 2000</xref>). This family of genes is well known for its precision in substrate selection. The molecules known to interact with AARS include tRNA, amino acids, and ATP. Since the same amino acids and ATP molecules are found in all organisms, variability inside the AARS sequences must relate to the recognition and interaction with the tRNAs. Correspondingly, each AARS usually contains a conserved domain for catalysis and acceptor helix interaction and a nonconserved domain that interacts with the variable distal parts of its substrate tRNA (<xref rid="pbio-0020081-Schimmel1" ref-type="bibr">Schimmel et al. 1993</xref>). For instance, in bacterial-type prolyl-tRNA synthetase (ProRS), the N-terminal catalytic domain (approximately 200 amino acids) and the C-terminal anticodon-binding domain (approximately 150 amino acids) are highly conserved, while a less conserved region of about 180 amino acids is inserted between them (<xref ref-type="supplementary-material" rid="sg002">Figure S2</xref>). This variable domain shows similarity to the YbaK domain, which is thought to be involved in oligonucleotide binding (<xref rid="pbio-0020081-Zhang1" ref-type="bibr">Zhang et al. 2000</xref>). Sporadic conserved residues in this region of <named-content content-type="genus-species">E. coli</named-content> ProRS are known to be involved in the posttransfer editing for mischarged Ala-tRNA<sup>Pro</sup> (<xref rid="pbio-0020081-Wong1" ref-type="bibr">Wong et al. 2002</xref>). ProRS is also known to possess an inherent ability to mischarge cysteine (<xref rid="pbio-0020081-Ahel1" ref-type="bibr">Ahel et al. 2002</xref>). Partial deletion of this variable region of <named-content content-type="genus-species">E. coli</named-content> ProRS results in a lower rate of proline acylation to cysteine acylation (<xref rid="pbio-0020081-Ahel1" ref-type="bibr">Ahel et al. 2002</xref>), suggesting a possible role of substrate discrimination in this region. Thus, the variability in this inserted domain of ProRS appears to contribute to substrate recognition and the editing function of the enzyme. Intriguingly, ProRS in <italic>Methanococcus jannaschii,</italic> which does not have this inserted region, also possesses editing abilities (<xref rid="pbio-0020081-Beuning1" ref-type="bibr">Beuning and Musier-Forsyth 2001</xref>). As a result, there is a possibility that this region may have another unknown function, e.g., interaction with other undetected molecules.</p></sec><sec id="s2b4"><title>Carbohydrate active enzymes</title><p>Variable regions exist in carbohydrate metabolizing enzymes, such as glycosyltransferases (GTs) and glycoside hydrolases (GHs), which respectively catalyze the biosynthesis of diverse glycoconjugates and their selective cleavage (<xref rid="pbio-0020081-Bourne1" ref-type="bibr">Bourne and Henrissat 2001</xref>). Many pathogens express outer-membrane glycosylated oligosaccharides, which closely interact with the host environment (<xref rid="pbio-0020081-Saxon1" ref-type="bibr">Saxon and Bertozzi 2001</xref>). For example, they even mimic host cell surface glycoconjugates to evade immune recognition (<xref rid="pbio-0020081-Persson1" ref-type="bibr">Persson et al. 2001</xref>). Both GTs and GHs have been classified into subfamilies based on sequence similarity (<xref rid="pbio-0020081-Bourne1" ref-type="bibr">Bourne and Henrissat 2001</xref>). Structural studies on bacterial GTs from different subfamilies always reveal two-domain molecules, such as LgtC (<xref rid="pbio-0020081-Persson1" ref-type="bibr">Persson et al. 2001</xref>), GtfB (<xref rid="pbio-0020081-Mulichak1" ref-type="bibr">Mulichak et al. 2001</xref>), MurG (<xref rid="pbio-0020081-Hu1" ref-type="bibr">Hu et al. 2003</xref>), and SpsA (<xref rid="pbio-0020081-Charnock1" ref-type="bibr">Charnock and Davies 1999</xref>), with one domain responsible for donor molecule (usually nucleotide-diphospho-sugar) binding and the other domain involved in acceptor sugar molecule binding. These genes exhibit great variability in the acceptor-binding domains and conservation in the donor-binding domains (see <xref ref-type="supplementary-material" rid="sg003">Figure S3</xref> for the example of GtfB), which agrees with the relatively limited types of donor species (usually UDP/TDP-sugar) and their conserved binding modes, but a diversity of acceptor molecules (LgtC: lactose; GtfB: vancomycin aglycone; MurG: <italic>N</italic>-acetyl muramyl pentapeptide; SpsA: unknown). Owing to the lack of homology in the acceptor binding domains, the substrate specificities encoded by these regions for most GTs are still unknown.</p></sec><sec id="s2b5"><title>Transcriptional regulators</title><p>Prokaryotic transcriptional regulators form another large group of SVGs. Transcription regulators are usually two-domain proteins with one binding to DNA and one binding to ligand. The DNA-binding domains, which usually interact with DNA via helix–turn–helix, zinc-finger, or other modes, are more conserved than ligand-binding domains. Based on the characteristic conserved DNA-binding domains, transcriptional regulators can be classified into many different families (<xref rid="pbio-0020081-Nguyen1" ref-type="bibr">Nguyen and Saier 1995</xref>; <xref rid="pbio-0020081-Rigali1" ref-type="bibr">Rigali et al. 2002</xref>). Even within each family, the ligand-binding domains are variable. For instance, the C-terminal regions involved in effector molecule binding and oligomerization (E-b/O) inside the GntR transcriptional regulator family are highly variable both in sequence content and in size (<xref rid="pbio-0020081-Rigali1" ref-type="bibr">Rigali et al. 2002</xref>). The variability in the effector molecule-binding domains enables the transcriptional regulators to sense the presence of diverse ligands and signal the regulation of the downstream genes or operons accordingly. As in most previous cases, these variable regions remain functionally uncharacterized.</p></sec><sec id="s2b6"><title>Hypothetical genes</title><p>In addition to genes with functional annotations, our method identifies a number of SVGs with unknown or hypothetical annotations in each genome (<named-content content-type="genus-species">H. pylori</named-content>: 17 genes; <named-content content-type="genus-species">N. meningitidis</named-content>: 32 genes; <named-content content-type="genus-species">V. cholerae</named-content>: 69 genes, etc.; see <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu">http://geneva.bu.edu</ext-link> for the full list). In contrast to lineage-specific hypothetical genes, these hypothetical genes contain conserved domains, which suggest their functional importance. Although most of the conserved domains in these hypothetical genes have currently unknown function, there are a few exceptions. Among them are the prokaryotic mechanosensitive channel proteins, which respond to external osmotic pressure (<xref rid="pbio-0020081-Pivetti1" ref-type="bibr">Pivetti et al. 2003</xref>). Examples include the 343 amino acid long <named-content content-type="genus-species">E. coli</named-content> B1330 and 371 amino acid long <named-content content-type="genus-species">Bacillus subtilis</named-content> YhdY, both of which are currently annotated as “hypothetical.” However, they both have the characteristic domain of mechanosensitive proteins (Pfam domain: MS_channel). The central regions (approximately 150 amino acids) of these genes are conserved while both the N-terminal region (approximately 100 amino acids) and the C-terminal region (approximately 100 amino acids) are variable (see alignment in <xref ref-type="supplementary-material" rid="sg004">Figure S4</xref>). The conserved central region encodes three transmembrane segments, and the molecules are predicted to have their N-terminus outside and C-terminus inside the cell (<xref rid="pbio-0020081-Miller1" ref-type="bibr">Miller et al. 2003</xref>). Although the C-terminus is variable, the deletion experiments show that it is indispensable for stability and activity of this protein (<xref rid="pbio-0020081-Miller1" ref-type="bibr">Miller et al. 2003</xref>). It is tempting to hypothesize that the interacting partners for both N- and C-termini might vary in different organisms.</p></sec></sec><sec id="s2c"><title>Functional Classification of SVGs</title><p>We are interested in probing the functional distribution of SVGs within a single genome. Are certain functional categories overrepresented? In <xref ref-type="fig" rid="pbio-0020081-g005">Figure 5</xref>, we show a functional classification of SVGs in three microorganisms using 18 broad functional categories of the clusters of orthologous group (COG) database (<xref rid="pbio-0020081-Tatusov1" ref-type="bibr">Tatusov et al. 1997</xref>). We calculated the percenta<italic>g</italic>e (<italic>r</italic> in <xref ref-type="fig" rid="pbio-0020081-g005">Figure 5</xref>) of SVGs within each functional class and the <italic>p</italic>-value of overrepresentation (<xref ref-type="fig" rid="pbio-0020081-g005">Figure 5</xref>). Several functional categories are overrepresented (<italic>p</italic>-value < 0.01; see <xref ref-type="fig" rid="pbio-0020081-g005">Figure 5</xref> for details): (i) cell envelope biogenesis, outer membrane; (ii) DNA replication, recombination and repair; (iii) secondary metabolite biosynthesis, transport and catabolism; (iv) cell motility and secretion; (v) cell division and chromosome partitioning. Among them, only categories (i) and (ii) are overrepresented in all three genomes. Most functional categories involved in the basic metabolic processes are not significantly overrepresented or even underrepresented. The number of overrepresented categories and the order of significance differ from one genome to another, reflecting differences in genome content and presumably the relative importance of the different specific adaptations.</p><fig id="pbio-0020081-g005" position="float"><label>Figure 5</label><caption><title>Functional Classification of SVGs in Three Microorganisms</title><p> <italic>M</italic> is the total number of genes in a COG broad functional category, and <italic>m</italic> is the number of SVGs within that category. <italic>r</italic> ( = <italic>m/M</italic>) is the proportion of SVGs in that category. The <italic>p</italic>-value is calculated using a hypergeometric distribution: let <italic>N</italic> = number of genes in the genome; <italic>n</italic> = number of SVGs identified; <italic>M</italic> = number of genes belonging to a particular category; <italic>m</italic> = number of SVGs belonging to a particular category:</p><p> <disp-formula id="pbio-0020081-e003"><graphic xlink:href="pbio.0020081.e003"/></disp-formula> </p><p>The set of lineage-specific genes has been excluded in each genome to avoid the possible skew it brings to the estimation of significance. The significance level is set at 0.01. Cells with <italic>p</italic>-value less than 0.01 are shaded.</p></caption><graphic xlink:href="pbio.0020081.g005"/></fig><p>In <xref ref-type="fig" rid="pbio-0020081-g006">Figure 6</xref> we show the relative abundance of a set of SVG families in several microorganisms based on shared keywords in the annotations. The relative enrichments in several gene families for some microbes seem to correlate with the peculiarities of niche adaptation. In particular, <named-content content-type="genus-species">H. pylori</named-content> has more SVGs involved in cell motility and chemotaxis than two other genomes with a similar genome size <italic>(N. meningitidis, Streptococcus pneumoniae). H. pylori</italic> is one of the few microbes that can colonize the highly acidic gastric environment (<xref rid="pbio-0020081-Tomb1" ref-type="bibr">Tomb et al. 1997</xref>). The motility of <named-content content-type="genus-species">H. pylori</named-content> is crucial for its infectious capability and there is evidence that poorly motile strains are less able to colonize or survive in the host (<xref rid="pbio-0020081-OaToole1" ref-type="bibr">O'Toole et al. 2000</xref>). <named-content content-type="genus-species">S. pneumoniae</named-content> has more carbohydrate-metabolizing enzymes, especially glycosyltransferases (GTs), which appear to be segmentally variable. The unique pattern of cell surface glycosylation in <named-content content-type="genus-species">S. pneumoniae</named-content> has been under extensive investigation and plays an important role in pathogenesis (Tette<xref rid="pbio-0020081-Lin1" ref-type="bibr">lin et al. 2001</xref>). The GTs are responsible for making <italic>O</italic>-linked glycosylations on surface proteins, which coat the surface of the bacterium and interact with the host (Tette<xref rid="pbio-0020081-Lin1" ref-type="bibr">lin et al. 2001</xref>).</p><fig id="pbio-0020081-g006" position="float"><label>Figure 6</label><caption><title>Abundance of SVGs in Different Functional Categories in Five Microorganisms</title><p>The approximate total gene number for each organism is as follows: <named-content content-type="genus-species">H. pylori</named-content>, 1,566 genes; <named-content content-type="genus-species">S. pneumoniae</named-content>, 2,094 genes; <named-content content-type="genus-species">N. meningitidis</named-content>, 2,065 genes; <named-content content-type="genus-species">E. coli</named-content>, 4,289 genes; <named-content content-type="genus-species">B. subtilis</named-content>, 4,100 genes.</p></caption><graphic xlink:href="pbio.0020081.g006"/></fig></sec><sec id="s2d"><title>Gene Duplication and SVGs</title><p>Duplication followed by diversification is an efficient way of generating functional innovations (<xref rid="pbio-0020081-Prince1" ref-type="bibr">Prince and Pickett 2002</xref>). Regional sequence divergence has been observed between duplicated gene copies (<xref rid="pbio-0020081-Gu1" ref-type="bibr">Gu 1999</xref>; <xref rid="pbio-0020081-Dermitzakis1" ref-type="bibr">Dermitzakis and Clark 2001</xref>; <xref rid="pbio-0020081-Marin1" ref-type="bibr">Marin et al. 2001</xref>). We thus asked the following questions: (1) What is the distribution of paralogous genes in the set of SVGs in a single genome? (2) Is there a significant association between gene duplication and SVGs?</p><p>In <xref ref-type="fig" rid="pbio-0020081-g007">Figure 7</xref>A, we show the distribution of paralogous genes among SVGs in several genomes. We consider paralogous genes to be similar genes in the same genome with a BLAST <italic>E</italic>-value less than 1<italic>E</italic>-5. As shown in <xref ref-type="fig" rid="pbio-0020081-g007">Figure 7</xref>A, in <named-content content-type="genus-species">H. pylori</named-content>, <italic>N. meningitidis,</italic> and <named-content content-type="genus-species">S. pneumoniae</named-content>, the largest group of SVGs is the one with no paralogs. However, in <named-content content-type="genus-species">E. coli</named-content>, the largest group is the one with a single paralog. <named-content content-type="genus-species">E. coli</named-content> obviously has more paralogous genes in the SVG set, probably owing to a larger genome size by duplication. In <xref ref-type="fig" rid="pbio-0020081-g007">Figure 7</xref>A (inset), we show the percentage of genes with different numbers of paralogs in each class for both segmentally variable and fully conserved genes in <named-content content-type="genus-species">E. coli</named-content>. Interestingly, over half of the fully conserved genes in <named-content content-type="genus-species">E. coli</named-content> do not have paralogs. There is a significant difference between the two distributions (χ<sup>2</sup> test, <italic>p</italic>-value < 1<italic>E</italic>-5). In <xref ref-type="fig" rid="pbio-0020081-g007">Figure 7</xref>B, we list the number of genes in a contingency table and test the significance using a χ<sup>2</sup> test. For all genomes examined, there is a strong association between gene duplication and SVGs, suggesting an SVG is more likely to have originated from a duplicated gene.</p><fig id="pbio-0020081-g007" position="float"><label>Figure 7</label><caption><title>Paralogous Genes in SVGs</title><p>(A) Paralog families in SVGs for four microorganisms. The x-axis shows the number of paralogs for each SVG. The y-axis shows the number of SVGs. The inset figure shows the percentage of genes with different numbers of paralogs for SVGs and fully conserved genes in <named-content content-type="genus-species">E. coli</named-content> genome. The x-axis is the number of paralogs, and the y-axis is the percentage.</p><p>(B) Contingency tables to examine the dependence between SVG and paralogous gene. χ<sup>2</sup> statistics are computed using standard formula.</p></caption><graphic xlink:href="pbio.0020081.g007"/></fig><p>Here we give an interesting example where one paralogous copy of a gene is segmentally variable and the other copy is fully conserved. In <named-content content-type="genus-species">H. pylori</named-content> strain 26695, gene products of <italic>HP1299</italic> (253 amino acids) and <italic>HP1037</italic> (357 amino acids) both have a conserved domain (approximately 250 amino acids; Pfam: Peptidase_M24) that is characteristic of the methionyl aminopeptidase (<italic>map</italic>) family (metalloprotease family M24) (<xref rid="pbio-0020081-Rawlings1" ref-type="bibr">Rawlings and Barrett 1995</xref>). <italic>HP1299</italic> is fully conserved in a number of microbes and is homologous to the <italic>E. coli map</italic> gene (<xref ref-type="supplementary-material" rid="sg005">Figure S5</xref>), while the product of <italic>HP1037</italic> has an extra N-terminal region (approximately 100 amino acids) that is variable among its similar genes (<xref ref-type="supplementary-material" rid="sg006">Figure S6</xref>). Additionally, <italic>HP1037</italic> is annotated as a conserved hypothetical gene. The five residues found in the <italic>E. coli map</italic> that are involved in cobalt (Co<sup>2+</sup>) binding (Asp-97, Asp-108, His-177, Glu-204, Glu-235; <xref rid="pbio-0020081-Rawlings1" ref-type="bibr">Rawlings and Barrett 1995</xref>), are conserved in both genes by examining the multiple alignment. These findings suggest that <italic>HP1037</italic> might also encode a <italic>map</italic> activity and that its variable N-terminal region might be involved in additional functional roles, e.g., interactions with other molecules. In <named-content content-type="genus-species">Saccharomyces cerevisiae</named-content>, there are two <italic>map</italic> genes and both have an extra N-terminal region compared to the <italic>E. coli map</italic> gene. One copy of the yeast <italic>map</italic> gene contains zinc-finger motifs in the N-terminal region that are indispensable for in vivo function (<xref rid="pbio-0020081-Li1" ref-type="bibr">Li and Chang 1995</xref>). A functional role involving interaction with the ribosome has also been suggested for this N-terminal domain (<xref rid="pbio-0020081-Vetro1" ref-type="bibr">Vetro and Chang 2002</xref>). In most prokaryotes, it has been assumed that there is only one copy of the <italic>map</italic> gene. The SVG family exemplified by <italic>HP1037</italic> may represent another family of <italic>map</italic> genes in prokaryotes.</p></sec></sec><sec id="s3"><title>Discussion</title><p>A major fraction of bioinformatics research on sequence analysis has focused on the conserved regions in proteins, trying to hypothesize the role of the protein by identifying sequence motifs that have been shown experimentally to correlate with a specific function. Some work has gone into cataloging the groups of lineage-specific proteins that show no similarity to other proteins in GenBank (<xref rid="pbio-0020081-Galperin1" ref-type="bibr">Galperin and Koonin 1999</xref>), but there the route to assigning function usually needs experimental approaches requiring biochemistry or genetics or more rarely by determining the crystal structure of the gene product (<xref rid="pbio-0020081-Zhang1" ref-type="bibr">Zhang et al. 2000</xref>). Unfortunately, current bioinformatics methods are only occasionally helpful in suggesting where to begin such studies.</p><p>In this paper we have initiated an effort to identify SVGs, which contain both well-conserved regions and highly variable regions. By looking carefully at a few specific examples where functional information is available from experimental data, we find that the variable region often seems to play a key role in mediating interactions with other molecules, both large and small. Sometimes the variable portions are involved in biological processes with a component of interaction between the cell and agents from the external environment. For instance, the DNA methyltransferases are part of a defense system that recognizes and clears invading foreign DNA; membrane-bound sensory HKs and mechanosensitive ion channels, etc., monitor changes of living conditions. Sometimes the variable portions are involved in intracellular processes that appear to have lineage-specific features. Thus, the variable regions inside DNA GyrB and several types of AARSs probably determine the specificity of substrate recognition. The detailed factors that introduce the molecular variability may go well beyond our explanations here and likely vary from case to case. Some variable regions may have diverged a long time ago and are now kept constant, while others may keep changing. In all of these cases, SVGs are exceptionally worthy targets of further experimental investigation, and such investigations can be greatly aided by the presence of the conserved regions that may suggest a preliminary function to be tested.</p><p>Why might certain genes contain these variable regions? Could they be simply relics left over during evolution and now serve no purpose? Are they just “pseudo-segments” with no function? There are several lines of evidence that support the hypothesis that when variable regions have been retained, they indeed serve a function. First, several studies have shown that deletions are, on average, more frequent than insertions (<xref rid="pbio-0020081-Halliday1" ref-type="bibr">Halliday and Glickman 1991</xref>). As a result, if a region is evolving under weak functional constraints, it tends to get smaller over time (<xref rid="pbio-0020081-Lipman1" ref-type="bibr">Lipman et al. 2002</xref>). Second, in a special case, one can imagine that when a variable region occurs at the C-terminus of a protein and is not being selected, it is likely to suffer random mutations including nonsense mutations or insertions/deletions that cause a shift in reading frame. Thus, we searched GenBank release 136.0 for examples of genes that matched the conserved region of an SVG, but in which the C-terminus was missing or much shorter. The DNA sequences downstream of such hits were examined for similarity to the variable region in the query gene. Of the 83 SVGs with a C-terminal variable region in <named-content content-type="genus-species">H. pylori</named-content>, none of them had hits with a disrupting stop codon in the variable region; 20 of them have hits with genes showing insertions/deletions that cause frame shifts in the variable region. However, the real number is likely to be much fewer, since, based on previous work, many of them may be the results of sequencing errors (<xref rid="pbio-0020081-Posfai1" ref-type="bibr">Posfai and Roberts 1992</xref>).</p><p>In other cases, we find that some proteins have lost the variable segment in a subset of genomes. For instance, in ProRSs, the variable segment is absent in archaea; in GyrB, the variable segment is absent in the Gram-positive bacteria. Clearly in those cases the organisms can get by without the variable domain, although they may have a compensating function in a different gene. But this again does not imply that the variable region has no function in those genes that have retained it.</p><p>SVGs are distinct from sequences with shuffled domains (<xref rid="pbio-0020081-Doolittle1" ref-type="bibr">Doolittle 1995</xref>) in that the variable region is bounded by the same sets of conserved portions, while domain shuffling usually manifests itself in a different sequential order of conserved domains. We also hypothesize that the variable regions in SVGs are not the result of multiple domain fusion events, each resulting in an insertion of a different sequence into the protein. This hypothesis is supported by the fact that the fused domains are often conserved across multiple organisms (<xref rid="pbio-0020081-Marcotte1" ref-type="bibr">Marcotte et al. 1999</xref>). Additionally, our procedure requires that the variable regions are of similar length within a family of proteins, which are also restricted to conserved length distributions. This filter suggests a mutational mechanism that originated from an ancient protein. Indeed, it is possible that originally the variable region was a result of a single or possibly relatively few ancient fusion events, but this paper does not focus on the evolutionary origin of SVGs.</p><p>Another prediction from our observations is that the variable regions are excellent candidates to bind substrates or partner macromolecules. They may be extremely helpful in discovering the networks of protein–protein or protein–nucleic acid interactions within a cell. Bioinformatics may even be able to help in this endeavor by finding genes that seem to have coevolving variable regions as a result of such interactions. Experimental data from techniques such as the yeast two-hybrid system or microarrays may provide evidence for interactions that can involve two variable regions.</p><p>Much additional bioinformatics work will be needed to explore fully the potential of this method in hypothesizing function. For instance, the size limits we have arbitrarily imposed on the variable region should be tested systematically. In our relatively simple formulation presented here, the length of the variable region and the number of proteins in the same family that do not have an alignment to the variable region are the primary factors in determining its statistical significance. Methods using other sequence analysis tools, such as multiple alignment and sequence profiles, may provide alternative ways to identify segmental pattern of variability. A fundamental problem is to differentiate random evolutionary drift from positive selection correlated to functional requirements. Although one might expect that the N- and C-termini may be more variable than the regions in the middle, our data suggest that variable regions in SVGs are not preferentially located in either end (data not shown). We have also examined the amino acid composition, codon usage, and GC content in the variable regions and the conserved regions of the same SVG. While there is no significant deviation of amino acid composition and GC content between the two regions in general, codon usage appears to be biased in the variable regions (data not shown).</p><p>SVGs usually account for 10%–20% of the total genes in a microbial genome. Currently, we think of the class of lineage-specific genes as being the key factor that distinguishes one strain or species from another. The class of SVGs that we have defined in this paper must now be added to this collection of lineage-specific genes by virtue of the unique segments that constitute their variable regions. They also appear to provide functional elements that help to differentiate among strains and species. This point is well illustrated by considering the restriction-modification systems. Here, the DNA methyltransferases, which have a variable region responsible for DNA recognition, are members of the SVG class. With the help of their companion restriction endonucleases, which typically appear as lineage-specific genes, they serve to keep foreign, unmodified DNA sequences from entering the genome. In this case, the synergy of function provided by members of the two classes highlights the key role that both sets of genes must play in defining the individuality of a strain or species.</p><p>Our analysis to date is limited to prokaryotes and archaea where SVGs are transcribed and translated as contiguous genomic segments. In eukaryotes, alternative RNA splicing introduces substantial additional complexity into the interpretation of gene structure and protein product, thereby rendering impossible the simple analysis we have applied here. It is tempting to consider alternative splicing as a highly evolved control mechanism to introduce the variability we find in the SVGs and thereby achieve the functional diversity necessary for cell survival under different conditions. In eukaryotes, alternatively spliced exons can be introduced in response to the functional demands of different cell types by merely juggling protein coding regions in the genome, thereby creating an SVG structure. If this view is correct, then it reinforces and highlights the importance of these SVGs to the workings of the cell.</p><p>In this paper we have provided an initial glimpse of SVGs, which appear to provide an important genetic layer in the adaptation of cells to novel environments and hazardous pathogens. We have focused attention on the biological significance of these genes, especially those that have highly diverged segments. We are currently trying to develop a more refined classification of these genes so as to explore the functional significance of the variability. We would like to know whether extreme variability is required for diverse function or whether more modest variation is sufficient. Such questions require that we can first distinguish positive selection acting on these variable regions from neutral evolution leading to gene decay and eventual loss. Since the variable regions we report are often not amenable to current tools available for alignment, we are exploring new methods that will help us to assess whether positive selection is driving the evolution of these genes.</p><p>In summary, we have identified an extremely useful way of classifying genes that leads to the identification of those with a high priority for both experimental and computational research.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><p>Our method for detecting SVGs includes several steps: (1) identification of similar genes followed by query-anchored multiple alignment using nongapped BLAST (<xref rid="pbio-0020081-Altschul1" ref-type="bibr">Altschul et al. 1990</xref>); (2) taxonomy clustering of similar genes to avoid bias; (3) detection of segmental variability.</p><sec id="s4a1"><title>Identification of similar genes</title><p>Given a gene, we start by searching for all its similar genes in the nonredundant database (GenBank release 136.0, 15 June 2003) using nongapped BLAST (<xref rid="pbio-0020081-Altschul1" ref-type="bibr">Altschul et al. 1990</xref>). We use the nongapped BLAST because the gapless high scoring pairs (HSPs) reported are rather conservative. The gapped BLAST, however, tends to extend HSPs over variable regions, which has been observed in several examples (e.g., DNA-recognition domain in cytosine-specific methyltransferase; data not shown). Two criteria are used to define close similarity. First, the <italic>E</italic>-value is less than 1<italic>E</italic>-10. Here we use a strict <italic>E</italic>-value threshold to avoid possible functional divergence among the homologs. Accordingly, we use the BLOSUM80 scoring matrix in the BLASTP search, although the result does not change dramatically if BLOSUM62 is used. Second, the overall length of the hit sequence does not differ significantly from the query sequence. We define the gap content (GapC) between two sequences:</p><p> <disp-formula id="pbio-0020081-e001"><graphic xlink:href="pbio.0020081.e001.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where <italic>L,l</italic> are the lengths of the protein sequences of two genes. It is a measure of the smallest percentage of gaps needed to be introduced into the pairwise alignment. Sequences with a high GapC value indicate significantly different domain structures, possibly owing to domain insertions or losses, and thus are excluded from the set of similar genes. In our current implementation, we require that GapC must be less than 0.2.</p></sec><sec id="s4a2"><title>Taxonomy clustering of the similar genes</title><p>Similar genes reported by BLASTP are not evenly distributed among different species. In many cases, highly similar genes from different strains of the same species or highly similar paralogous genes from a particular strain tend to introduce bias into the dataset. We adopted a simple taxonomy clustering by using the NCBI Taxonomy Database (<xref rid="pbio-0020081-Wheeler1" ref-type="bibr">Wheeler et al. 2003</xref>) to reduce this bias.</p><p>We collapse all the similar genes from the same species into a single group. Then we choose the gene with the best similarity score to the query sequence as the representative of that species for later calculations. The definition of species follows the hierarchical taxonomy used in the NCBI Taxonomy database (superkingdom → phylum → class → subclass → order → family → genus → species → no rank [strain]). By doing taxonomy clustering, we are able to collect a less biased sample of similar genes from different species.</p></sec><sec id="s4a3"><title>Detection of segmental variability</title><p>Query-anchored multiple alignment after taxonomy clustering is performed by aligning the HSPs reported by nongapped BLAST (see <xref ref-type="supplementary-material" rid="sg002">Figure S2</xref> and <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu)">http://geneva.bu.edu)</ext-link>. Two unaligned regions in two sequences are considered as the variable regions if they are bounded by similar HSPs at both ends (or one end, if the unaligned region is at the terminus of the gene). To avoid the possibility of a large segment containing insertions or deletions, we again require that GapC be less than 0.2 between these two unaligned regions.</p><p>For each amino acid position in the query gene, we can count the number of times <italic>(m)</italic> it is inside an HSP region and the number of times <italic>(n)</italic> it is inside a variable region. A high ratio of <italic>n</italic> over <italic>m</italic> + <italic>n</italic> suggests that this position is inside the variable region most of the time. We estimate the statistical significance (<italic>p</italic>-value) of the variability for each position by a binomial distribution:</p><p> <disp-formula id="pbio-0020081-e002"><graphic xlink:href="pbio.0020081.e002.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where <italic>q</italic> is the probability of an amino acid position being inside a HSP region. We estimate <italic>q</italic> by averaging the proportion of HSP in each hit sequence among all hits. If the <italic>p</italic>-value calculated using the above formula is less than the significance level, which we set at 0.05, we then consider this position as a variable position; otherwise, it is a conserved position. A consecutive run of variable positions forms a variable region. The next question is how long the variable region should be to be considered meaningful, as opposed to functionally unimportant regions such as linker regions, which are usually short. From our experience, there is no clear decision boundary between the length of the region and its functional importance. Any choice of cutoffs would have to balance between false positives and false negatives. However, previous studies on the length distribution of protein domains has shown that the most likely length of a protein domain is around 70 amino acids, and regions shorter than this are less likely to form a functional domain (<xref rid="pbio-0020081-Wheelan1" ref-type="bibr">Wheelan et al. 2000</xref>). Based on this, we chose 70 amino acids as the length threshold for a variable region to be considered functionally important. In <xref ref-type="supplementary-material" rid="sg007">Figure S7</xref>, we show the length distribution of the variable regions in all genes of <italic>H. pylori.</italic> </p><p>A direct way of visualizing the variability of a protein sequence is by calculating the ratio of <italic>n</italic> over (<italic>m</italic> + <italic>n</italic>) for each position and plotting it. We call such plots variability profiles. Sample variability profiles are shown in <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>. In <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>A, two obvious peaks are present: one from position 20 to 70, the other from position 160 to 300. The latter (approximately 140 amino acids) forms a separate DNA recognition domain, while the former (approximately 50 amino acids) has no known function. In <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref> we also show conserved subsequences from the Pfam (<xref rid="pbio-0020081-Bateman1" ref-type="bibr">Bateman et al. 2002</xref>), BLOCKS (<xref rid="pbio-0020081-Henikoff1" ref-type="bibr">Henikoff et al. 1999</xref>), and PRINTS (<xref rid="pbio-0020081-Attwood1" ref-type="bibr">Attwood et al. 2003</xref>) databases. The BLOCKS and PRINTS databases are relatively conservative in defining motifs. However, the Pfam domain seems to include the variable region within the conserved region, as shown in <xref ref-type="fig" rid="pbio-0020081-g001">Figure 1</xref>A.</p></sec></sec></sec><sec sec-type="supplementary-material" id="s5"><title>Supporting Information</title><sec id="s5a"><title>Data Deposit</title><p>We provide a static collection of segmentally variable genes at our Web site, <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu">http://geneva.bu.edu</ext-link>. SVGs for several representative genomes are listed there. For SVG lists in other genomes, please request more information from Y. Zheng at E-mail: <email>[email protected]</email>. All the case examples mentioned throughout the paper and Supporting Information have been compiled into one Web page, <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu/paper03.html">http://geneva.bu.edu/paper03.html</ext-link>, with hyperlinks. Readers can follow each hyperlink to access additional information from Pfam, BLOCKS, PRINTS, COG, and nongapped BLAST for each gene.</p><supplementary-material content-type="local-data" id="sg001"><label>Figure S1</label><caption><title>Multiple Alignment of BtuB and Homologs</title><p>Conservation score is plotted under the alignment (ClustalX). The conserved portions are as follows: N-terminal domain, extreme C-terminal domain, and a segment between N-terminal and C-terminal domain. The variable domain (between N-terminal and C-terminal) overlaps with the transmembrane 22-strand β-barrel regions.</p><p>(2.69 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg001.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg002"><label>Figure S2</label><caption><title>Query-Anchored Alignment of ProRS</title><p>The query protein is <named-content content-type="genus-species">H. pylori</named-content> ProRS. The blue segments are HSPs reported by nongapped BLAST. The yellow segments are the variable region. The gray region is the gap-rich region (GapC > 0.2, deletion in this alignment). See <ext-link ext-link-type="uri" xlink:href="http://geneva.bu.edu/paper03.html">http://geneva.bu.edu/paper03.html</ext-link> for a high-resolution Web figure.</p><p>(4.71 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg002.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg003"><label>Figure S3</label><caption><title>Multiple Alignment of GtfB and Its Homologs</title><p>(3.12 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg003.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg004"><label>Figure S4</label><caption><title>Multiple Alignment of <named-content content-type="genus-species">B. subtilis</named-content> Gene <italic>yhdY</italic> and Its Homologs</title><p>YhdY is currently annotated as a hypothetical protein and contains a conserved domain for mechanosensitive proteins (the middle region of the alignment) and two variable domains (N- and C-termini).</p><p>(2.86 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg004.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg005"><label>Figure S5</label><caption><title>Multiple Alignment for <named-content content-type="genus-species">H. pylori</named-content> Gene <italic>HP1299</italic> </title><p>It is the methionine aminopeptidase (type Ia <italic>map</italic>). This is an example of a fully conserved gene.</p><p>(1.87 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg005.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg006"><label>Figure S6</label><caption><title>Multiple Alignment for <named-content content-type="genus-species">H. pylori</named-content> Gene <italic>HP1037</italic> </title><p>It is currently annotated as “conserved hypothetical protein.” The N-terminal region is variable. The conserved C-terminal domain is characteristic of methionine aminopeptidase.</p><p>(2.22 MB EPS).</p></caption><media xlink:href="pbio.0020081.sg006.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg007"><label>Figure S7</label><caption><title>Length Distribution of Variable Regions in the Genome of <named-content content-type="genus-species">H. pylori</named-content> </title><p>Shown as a histogram. Only variable regions inside fully conserved genes and SVGs are included. Pink line shows the domain size distribution in 3D-structure database (data from <xref rid="pbio-0020081-Wheelan1" ref-type="bibr">Wheelan et al. 2000</xref>).</p><p>(643 KB EPS).</p></caption><media xlink:href="pbio.0020081.sg007.ps"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="st001"><label>Table S1</label><caption><title>Classification of Genes into Three Broad Categories</title><p>(62 KB DOC).</p></caption><media xlink:href="pbio.0020081.st001.doc"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec><sec id="s5b"><title>Accession Numbers</title><p>The GenBank (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/GenBank/">http://www.ncbi.nlm.nih.gov/GenBank/</ext-link>) accession numbers for the genes discussed in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 3</xref> are as follows: <italic>atpA</italic> (2314285), <italic>atpD</italic> (2314283), <italic>atpG</italic> (2314284), <italic>dnaX</italic> (2313841), <italic>flgK</italic> (2314271), <italic>ftsK</italic> (2314237), <italic>gyrB</italic> (2313611), <italic>hetA</italic> (2314367), <italic>HP1450</italic> (2314626), <italic>infB</italic> (2314195), <italic>M.hpyAVIB</italic> (2313124; REBASE [<ext-link ext-link-type="uri" xlink:href="http://rebase.neb.com">http://rebase.neb.com</ext-link>] ID M2.hpyAVI), <italic>mutS</italic> (2313742), <italic>NQO3</italic> (2314431), <italic>NQO8</italic> (2314432), <italic>polA</italic> (2314647), <italic>rps4</italic> (2314460), <italic>spaB</italic> (2313717), <italic>spoT</italic> (2313901), <italic>tlpA</italic> (2313179), and <italic>tlpC</italic> (2313162).</p><p>The GenBank accession numbers for the genes discussed in <xref ref-type="fig" rid="pbio-0020081-g002">Figure 3</xref> are as follows: <named-content content-type="genus-species">Agrobacterium tumefaciens</named-content> (15890351), <named-content content-type="genus-species">B. subtilis</named-content> (16079962), <named-content content-type="genus-species">Enterococcus faecalis</named-content> (8100675), <named-content content-type="genus-species">E. coli</named-content> K12 (16128553), <named-content content-type="genus-species">L. innocua</named-content> (16801788), <named-content content-type="genus-species">Mycobacterium leprae</named-content> (15826988), <named-content content-type="genus-species">M. tuberculosis</named-content> CDC1551 (15840173), <named-content content-type="genus-species">Nitrosomonas europaea</named-content> (22955201), <named-content content-type="genus-species">Nostoc</named-content> sp. PCC 7120 (17228666), <named-content content-type="genus-species">P. syringae</named-content> pv. syringae B728a (23470301), <named-content content-type="genus-species">Ralstonia metallidurans</named-content> (22980570), <named-content content-type="genus-species">R. solanacearum</named-content> (17548875), <named-content content-type="genus-species">Synechococcus</named-content> sp. PCC 7942 (21954778), and <named-content content-type="genus-species">Thermotoga maritime</named-content> (15644402); in case studies, <named-content content-type="genus-species">B. subtilis yhdY</named-content> (2633299), <named-content content-type="genus-species">E. coli</named-content> b1330 (1787591), <named-content content-type="genus-species">H. pylori</named-content> cytosine-specific DNA methyltransferase (2313124), <named-content content-type="genus-species">H. pylori</named-content> HP1299 (2314463), <named-content content-type="genus-species">H. pylori</named-content> HP1037 (2314181), <named-content content-type="genus-species">H. pylori</named-content> prolyl-tRNA synthetase (2313329), and <named-content content-type="genus-species">H. pylori</named-content> VacB (2314413).</p></sec></sec>
Mimotopes for Alloreactive and Conventional T Cells in a Peptide–MHC Display Library	<p>The use of peptide libraries for the identification and characterization of T cell antigen peptide epitopes and mimotopes has been hampered by the need to form complexes between the peptides and an appropriate MHC molecule in order to construct a complete T cell ligand. We have developed a baculovirus-based peptide library method in which the sequence encoding the peptide is embedded within the genes for the MHC molecule in the viral DNA, such that insect cells infected with virus encoding a library of different peptides each displays a unique peptide–MHC complex on its surface. We have fished in such a library with two different fluorescent soluble T cell receptors (TCRs), one highly peptide specific and the other broadly allo-MHC specific and hypothesized to be much less focused on the peptide portion of the ligand. A single peptide sequence was selected by the former αβTCR that, not unexpectedly, was highly related to the immunizing peptide. As hypothesized, the other αβTCR selected a large family of peptides, related only by a similarity to the immunizing peptide at the p5 position. These findings have implications for the relative importance of peptide and MHC in TCR ligand recognition. This display method has broad applications in T cell epitope identification and manipulation and should be useful in general in studying interactions between complex proteins.</p>	<contrib contrib-type="author"><name><surname>Crawford</surname><given-names>Frances</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Huseby</surname><given-names>Eric</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib><contrib contrib-type="author"><name><surname>White</surname><given-names>Janice</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Marrack</surname><given-names>Philippa</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref><xref ref-type="aff" rid="aff3"> <sup>3</sup> </xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kappler</surname><given-names>John W</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref><xref ref-type="aff" rid="aff4"> <sup>4</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>The identification of peptide epitopes associated with particular αβ T cell receptors (αβTCRs) is often still a bottleneck in studying T cells and their antigenic targets in, for example, autoimmunity, hypersensitivity, and cancer. A direct genetic or biochemical attack on this problem can be successful, especially with class I major histocompatibility complex (MHCI)-presented peptides. For example, tumor (<xref rid="pbio-0020090-Van1" ref-type="bibr">Van Der Bruggen et al. 2002</xref>) and transplantation (<xref rid="pbio-0020090-Scott1" ref-type="bibr">Scott et al. 2000</xref>; <xref rid="pbio-0020090-Simpson1" ref-type="bibr">Simpson et al. 2001</xref>; <xref rid="pbio-0020090-Shastri1" ref-type="bibr">Shastri et al. 2002</xref>; <xref rid="pbio-0020090-Sahara1" ref-type="bibr">Sahara and Shastri 2003</xref>) peptide epitopes have been found this way. Identification of the antigenic peptide in a mix of peptides stripped from MHC molecules isolated from antigen-presenting cells (APCs) has sometimes been possible using a combination of a biological assay, peptide fractionation, and peptide sequencing (<xref rid="pbio-0020090-Guimezanes1" ref-type="bibr">Guimezanes et al. 2001</xref>). However, this method is extremely labor intensive and depends on relatively high peptide frequency in the mix and a very sensitive bioassay. These conditions are not always achievable, especially with peptides presented by MHCII, in which peptide loading of surface MHC may require peptide concentrations orders of magnitude higher than those required for MHCI loading.</p><p>The reward for the labor involved in identifying peptide epitopes directly can often be the identification of the protein source of the peptide, especially as the sequencing of the genomes of many organisms approaches completion. However, in many situations, rather than identifying this precise peptide epitope, it is sufficient to identify a peptide “mimotope.” Mimotopes can be defined as peptides that are different in sequence from the actual peptide recognized in vivo, but that are nevertheless capable of binding to the appropriate MHC molecule to form a ligand that can be recognized by the αβTCR in question. These peptides can be very useful for studying the T cell in vitro, for altering the immunological state of the T cell in vivo (<xref rid="pbio-0020090-Hogquist1" ref-type="bibr">Hogquist et al. 1994</xref>), for vaccine development (<xref rid="pbio-0020090-Partidos1" ref-type="bibr">Partidos 2000</xref>), and potentially in preparing multimeric fluorescent peptide–MHC complexes for tracking T cells in vivo (<xref rid="pbio-0020090-You1" ref-type="bibr">You et al. 2003</xref>).</p><p>Mimotopes can sometimes be identified in randomized peptide libraries that can be screened for presentation by a particular MHC molecule to the relevant T cell (<xref rid="pbio-0020090-Gavin1" ref-type="bibr">Gavin et al. 1994</xref>; <xref rid="pbio-0020090-Linnemann1" ref-type="bibr">Linnemann et al. 2001</xref>; <xref rid="pbio-0020090-Sung1" ref-type="bibr">Sung et al. 2002</xref>; reviewed in <xref rid="pbio-0020090-Hiemstra1" ref-type="bibr">Hiemstra et al. 2000</xref>; <xref rid="pbio-0020090-Liu2" ref-type="bibr">Liu et al. 2003</xref>). Thus far, strategies for screening these types of libraries have involved testing individual pools of peptides from the library and then either deduction of the mimotope sequence from the pattern of responses or sequential reduction in the size of the pool until a single peptide emerges. There are several limitations to this type of approach. Again, a very sensitive T cell bioassay is needed in which the activity of the correct stimulating peptide is not masked by competition with the other peptides in the pool. Also, an APC that expresses the relevant MHC molecule, but not the relevant peptide, must be found or constructed. Finally, because the screen relies on T cell stimulation, only agonist mimotope peptides are identified.</p><p>In other applications, another powerful library method has been sequential enrichment/expansion of a displayed library of protein–peptide variants by direct ligand–receptor binding, e.g., using bacterial phage or yeast (also reviewed in <xref rid="pbio-0020090-Liu2" ref-type="bibr">Liu et al. 2003</xref>). These methods have not yet been developed for the routine identification of T cell antigen mimotopes, because of the lack of a suitable system for the display of peptide–MHCs or for screening via αβTCR binding using these organisms. In this paper, we describe such a method using modifications of previously described systems for producing soluble peptide–MHC complexes (<xref rid="pbio-0020090-Kozono1" ref-type="bibr">Kozono et al. 1994</xref>; <xref rid="pbio-0020090-Crawford1" ref-type="bibr">Crawford et al. 1998</xref>; <xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>) and αβTCRs (<xref rid="pbio-0020090-Kappler2" ref-type="bibr">Kappler et al. 1994</xref>) from baculovirus-infected insect cells. We constructed a library of peptides displayed on the surface of baculovirus-infected cells bound to the mouse MHCII molecule, IA<sup>b</sup>. The peptides in the library varied in five peptide amino acids known to be surface exposed and predicted to lie within the footprint of αβTCR interaction.</p><p>Using fluorescent αβTCRs as probes, we have identified in the library mimotopes for two types of T cells, both originally produced by immunization of mice with the same IA<sup>b</sup>–peptide combination. One of these T cells was predicted from previous data (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>) to be very dependent on all of the peptide surface exposed amino acids. Consistent with these predictions, a single peptide mimotope was identified in the library for this T cell. The sequence of this peptide was highly related to the immunizing peptide. In contrast, the other T cell was hypothesized to be very peptide promiscuous (<xref rid="pbio-0020090-Marrack1" ref-type="bibr">Marrack et al. 2001</xref>) based on its broad allo-MHC reactivity. Consistent with this hypothesis, its αβTCR selected a large family of peptide mimotopes from the library. Comparison of these peptides indicated that attention of this αβTCR was focused primarily on a single position in the peptide.</p></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>Characteristics of a Broadly Alloreactive and Conventional T Cell</title><p>For this study we selected two T cell hybridomas, both prepared from IA<sup>b</sup> mice immunized with the peptide p3K. This peptide binds well to IA<sup>b</sup> (<xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>), and its crystal structure bound to IA<sup>b</sup> has been determined (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>) (<xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>A). The hybridoma B3K-06 was produced from wild-type C57BL/6 immunized conventionally with the peptide (<xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>). Like most T cells resulting from immunization with a foreign peptide, it responds to IA<sup>b</sup>-expressing APCs in the presence, but not the absence, of p3K (<xref ref-type="fig" rid="pbio-0020090-g001">Figure1</xref>B). It does not respond to APCs expressing other alleles of the IA MHCII molecule (data not shown). Also, as is commonly seen with conventional T cells, the interaction of the αβTCR of B3K-06 with IA<sup>b</sup>-p3K is very sensitive to changes in any of the peptide amino acids exposed on the surface of the IA<sup>b</sup>-p3K complex. Mutation of Q2, K3, K5, N7, or K8 to alanine virtually eliminates recognition of p3K by B3K-06 (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>; see <xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>B).</p><fig id="pbio-0020090-g001" position="float"><label>Figure 1</label><caption><title>Structure of IA<sup>b</sup>-p3K and Properties of T Cell Hybridomas Reactive to It</title><p>(A) Ribbon structure of the α1 and β1 domains of IA<sup>b</sup> with a wire-frame representation of the bound p3K peptide (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>). Amino acids labeled in red are the five central peptide amino acids available for αβTCR interaction.</p><p>(B) The figure shows the response of 10<sup>5</sup> B3K-06 hybridoma cells to various peptides presented by 10<sup>5</sup> IA<sup>b</sup>-bearing APCs, LB-15.13.</p><p>(C) The figure shows the response of the T cell hybridoma YAe-62 to various MHCII molecules. In each case, 10<sup>5</sup> hybridoma cells were incubated overnight with MHCII presented in various ways. For IA<sup>b</sup>-p3K, soluble IA<sup>b</sup>-p3K was immobilized in the culture well before the addition of the hybridoma cells. In other cases, 10<sup>6</sup> spleen cells were used directly as APCs without additional peptide antigen. For pEα, the spleen cells came from IA<sup>b</sup>-pEα/ΔIAβ/ΔIi mice (<xref rid="pbio-0020090-Ignatowicz1" ref-type="bibr">Ignatowicz et al. 1996</xref>). For wild-type IA<sup>b</sup> and allo-MHCII, the spleen cells came from H-2 congenic mice on the C57BL/10 background. Finally, spleen cells from ΔIAβ/ΔIi C57BL/6 mice were used.</p></caption><graphic xlink:href="pbio.0020090.g001"/></fig><p>The hybridoma YAe-62 was chosen as a representative of broadly allo-reactive T cells present in mice carrying transgenes and gene knockouts that lead to expression of MHCII almost completely occupied by a single peptide (<xref rid="pbio-0020090-Ignatowicz1" ref-type="bibr">Ignatowicz et al. 1996</xref>). It was produced from IA<sup>b</sup>-p3K-immunized mice that express the IA<sup>b</sup> molecule covalently linked to pEα, a dominant IA<sup>b</sup>-binding peptide derived from the MHCII IEα chain. Its properties are shown in <xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>C. YAe-62 responds to APCs bearing IA<sup>b</sup>-p3K, but not to APCs lacking MHCII nor to IA<sup>b</sup>-pEα-bearing APCs from the mouse from which the hybridoma was derived. However, YAe-62 has additional reactivities common to many T cells isolated from these mice (<xref rid="pbio-0020090-Ignatowicz1" ref-type="bibr">Ignatowicz et al. 1996</xref>). In the absence of any added peptide, it also responds to all APCs expressing wild-type IA<sup>b</sup>, including those from mice with a much reduced MHCII peptide repertoire due to lack of the invariant chain. YAe-62 also responds well to APCs from a variety of mice carrying other alleles of IA. We have postulated that these T cells are focused on structural features of the MHCII molecule and are minimally dependent on direct peptide interaction (<xref rid="pbio-0020090-Marrack1" ref-type="bibr">Marrack et al. 2001</xref>).</p></sec><sec id="s2b"><title>Display of Functional Peptide–MHC on Baculovirus-Infected Insect Cells</title><p>We previously established methods that used baculovirus-infected insect cells to produce soluble MHC molecules with covalently bound antigenic peptides (<xref rid="pbio-0020090-Kozono1" ref-type="bibr">Kozono et al. 1994</xref>; <xref rid="pbio-0020090-Crawford1" ref-type="bibr">Crawford et al. 1998</xref>; <xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>). These constructions were the starting point for developing insect cells displaying functional peptide–MHCIIs. Several modifications were made to constructs that encoded the mouse MHCII molecule, IA<sup>b</sup>, with various bound peptides. First, to increase the stability of the molecule, an acid–base leucine zipper (<xref rid="pbio-0020090-OaShea1" ref-type="bibr">O'Shea et al. 1993</xref>) was attached to the C-termini of the extracellular portions of the MHC α and β chains, replacing what would normally be the transmembrane regions of these proteins. The basic half of the zipper was attached to the α chain (<xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>A) and the acidic half to the β chain (<xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>B). In addition, sequence encoding the transmembrane and cytoplasmic tail of the baculovirus major coat glycoprotein, gp64, was attached to the end of the acid zipper (<xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>B). Sf9 insect cells infected with virus encoding this construction produced the MHCII molecule at a high level anchored on the cell surface via the gp64 transmembrane (<xref ref-type="fig" rid="pbio-0020090-g003">Figure 3</xref>A). Also, to make Sf9 cells better APCs (<xref rid="pbio-0020090-Cai1" ref-type="bibr">Cai et al. 1996</xref>), we established a version transfected with the genes for mouse ICAM and B7.1 (<xref ref-type="fig" rid="pbio-0020090-g003">Figure 3</xref>B). When we tested the ability of Sf9 cells displaying the IA<sup>b</sup>-p3K complex to present the antigen to B3K-06 or YAe-62, the presence of ICAM/B7.1 greatly improved IL-2 production (<xref ref-type="fig" rid="pbio-0020090-g003">Figure 3</xref>C). These results showed that IA<sup>b</sup>-p3K could be displayed on the surface of insect cells in a form easily recognized by T cells. In all of the experiments described below, infected conventional Sf9 cells were used for flow cytometry and infected ICAM/B7.1-expressing Sf9 cells were used in IL-2 stimulation assays.</p><fig id="pbio-0020090-g002" position="float"><label>Figure 2</label><caption><title>Constructions Used in These Experiments</title><p>(A and B) Previously described constructions (<xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>) for the coexpression in a single baculovirus of soluble version of the α (A) and β (B) chains of IA<sup>b</sup> were modified as described in the <xref ref-type="sec" rid="s4">Materials and Methods</xref> to anchor the molecule on the surface of infected insect cells.</p><p>(C) The construction was further modified as described in the <xref ref-type="sec" rid="s4">Materials and Methods</xref> to disrupt the IA<sup>b</sup> β chain with sequence encoding enhanced GFP flanked by sites for the enzymes SbfI and CeuI.</p><p>(D and E) A degenerate DNA fragment was produced by PCR (D) and cloned into the construct replacing the GFP-encoding sequence (E) as described in the <xref ref-type="sec" rid="s4">Materials and Methods</xref>.</p></caption><graphic xlink:href="pbio.0020090.g002"/></fig><fig id="pbio-0020090-g003" position="float"><label>Figure 3</label><caption><title>Functional Display of IA<sup>b</sup>-p3K on the Surface of Insect Cells</title><p>(A) Sf9 insect cells were infected with baculovirus encoding a membrane-bound form of IA<sup>b</sup>-p3K. After 3 d, the surface expression of IA<sup>b</sup>-p3K was detected with an anti-IA<sup>b</sup> mAb using flow cytometry.</p><p>(B) The genes for mouse ICAM (CD54) and B7.1 (CD80) were cloned into an insect cell expression plasmid as described in the <xref ref-type="sec" rid="s4">Materials and Methods</xref>. The plasmids were used to cotransfect Sf9 cells, and a stable transfectant (Sf9-ICAM/B7.1) was cloned expressing both proteins detected with mAbs using flow cytometry.</p><p>(C) Either Sf9 (open bars) or Sf9-ICAM/B7.1 (closed bars) cells were infected with baculovirus expressing IA<sup>b</sup>-p3K. After 3 d, the infected insect cells were used as APCs to stimulate IL-2 production from B3K-06 and YAe-62. Uninfected cells were used as negative controls.</p></caption><graphic xlink:href="pbio.0020090.g003"/></fig></sec><sec id="s2c"><title>Detection of Displayed Peptide–MHC with Multimeric αβTCR</title><p>Next we prepared fluorescent, soluble αβTCR reagents for use in flow cytometry to detect insect cells displaying the appropriate peptide–MHCII combination. Fluorescent multivalent versions of the soluble αβTCRs of B3K-06 and YAe-62 bound to insect cells displaying the IA<sup>b</sup>-p3K, but not a control peptide–MHCII combination (<xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>A).</p><fig id="pbio-0020090-g004" position="float"><label>Figure 4</label><caption><title>Detection of IA<sup>b</sup>-p3K-Expressing Insect Cells with Polyvalent, Fluorescent αβTCRs</title><p>(A) Sf9 insect cells were infected with baculovirus encoding IA<sup>b</sup> bound either to p3K (filled histogram) or a control peptide (FEAPVAAALHAV) (unfilled histogram). After 3 d, the infected insect cells were incubated with polyvalent, fluorescent soluble αβTCRs from B3K-06 or YAe-62. The binding of each αβTCR was assessed by flow cytometry.</p><p>(B) Cells, prepared as in (A), were simultaneously analyzed with fluorescent αβTCRs and a mAb specific for IA<sup>b</sup> (17–227) that does not interfere with αβTCR–IA<sup>b</sup> interaction.</p><p>(C) The binding of the αβTCRs is shown only for those infected insect cells that bear a high level of surface IA<sup>b</sup> (dotted region in [B]).</p></caption><graphic xlink:href="pbio.0020090.g004"/></fig><p>Insect cells displaying IA<sup>b</sup>-p3K bound the αβTCR reagents very heterogeneously (<xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>A), probably owing to heterogeneous expression of IA<sup>b</sup>-p3K due to variations in the multiplicity of infection (MOI) and the lack of synchrony in viral infection and expression. To focus on cells bearing a particular level of IA<sup>b</sup>, we stained the cells simultaneously with the fluorescent αβTCR reagents and with an anti-IA<sup>b</sup> monoclonal antibody (mAb) that did not interfere with αβTCR binding. In this case, there was a direct correlation between the amount of surface IA<sup>b</sup>-p3K expressed by an individual insect cell and the amount of αβTCR bound (<xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>B) with cells bearing a particular level of IA<sup>b</sup>-p3K, binding the αβTCRs uniformly (<xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>C). Therefore, comparing the two types of staining gave us a useful tool to evaluate the relation between peptide sequence and the strength of αβTCR binding (see below).</p></sec><sec id="s2d"><title>Recovering Baculovirus Carrying a Particular Peptide–MHC Combination</title><p>Our experiments showed that fluorescent αβTCRs could be used with flow cytometry to identify insect cells infected with a baculovirus encoding a specific peptide–MHC combination. We next tested whether this system could be used to enrich baculoviruses encoding a particular peptide–MHC. Insect cells were infected at an MOI of about 1 with a mixture of baculoviruses. Of these viruses, 1% encoded the IA<sup>b</sup>-p3K molecule and 99% encoded a control molecule (an αβTCR β chain). The infected cells were stained with fluorescent YAe-62 αβTCR and analyzed by flow cytometry. Although a distinct population of brightly fluorescent cells was not seen, the 1% of the cells with the brightest fluorescence were sorted, as were an equal number of cells that were very dully fluorescent (<xref ref-type="fig" rid="pbio-0020090-g005">Figure 5</xref>A). The recovered infected cells were cultured with fresh insect cells to produce new viral stocks. These stocks were used to infect insect cells that were tested again with the fluorescent αβTCR reagent (<xref ref-type="fig" rid="pbio-0020090-g005">Figure 5</xref>B). The cells infected with virus from the few fluorescent positive cells in the original population were now nearly all brightly fluorescent, and those infected with the virus from the fluorescently dull cells were nearly all negative for binding of the αβTCR. These results showed that flow cytometry could be used with a fluorescent multimerized αβTCR to find and greatly enrich insect cells infected with a virus encoding a specific peptide–MHC combination.</p><fig id="pbio-0020090-g005" position="float"><label>Figure 5</label><caption><title>Recovery of IA<sup>b</sup>-p3K Virus-Infected Cells with Fluorescent αβTCR</title><p>(A) Sf9 cells were infected with a mixture of virus, 99% of which encoded a control protein (a TCR β chain linked to the gp64 transmembrane/cytoplasmic tail) and 1% of which encoded IA<sup>b</sup>-p3K. After 3 d, the infected cells were analyzed as in <xref ref-type="fig" rid="pbio-0020090-g003">Figure 3</xref>A for binding fluorescent αβTCR from YAe-62. The 1% of the infected cells with the brightest fluorescence was sorted (high sort, 15,700 cells). As a control, a similar number of cells that fluoresced as dully as the background fluorescence were also sorted (low sort).</p><p>(B) The sorted cells were incubated with fresh Sf9 insect cells to allow propagation of the viruses and production of new stocks. The stocks were used to infect new Sf9 cells, and after 3 d the analysis of αβTCR binding was repeated.</p></caption><graphic xlink:href="pbio.0020090.g005"/></fig></sec><sec id="s2e"><title>Construction of a Peptide Library Attached to IA<sup>b</sup> in Baculovirus</title><p>The most widely used method for introducing gene constructions into baculovirus involves assembling the construct first in an <named-content content-type="genus-species">Escherichia coli</named-content> transfer plasmid, where it is flanked by sections of baculovirus DNA. The complete construct is then introduced into baculovirus by homologous recombination using any of the commercially available modified baculovirus DNAs that require homologous recombination with the plasmid in order to generate functional circular viral DNA (<xref rid="pbio-0020090-Kitts1" ref-type="bibr">Kitts and Possee 1993</xref>). Based on this procedure, we constructed an IA<sup>b</sup>–peptide library in two steps. In the original transfer plasmid that encoded the displayed IA<sup>b</sup>-p3K, we flanked the site encoding the peptide with unique restriction sites, one in the section encoding the β chain leader and the other in the section encoding the linker from the peptide to the N-terminus of the β chain. The DNA between these sites was replaced with DNA encoding enhanced green fluorescent protein (GFP) (Clontech, Palo Alto, California, United States) in-frame with the IA<sup>b</sup> signal peptide and with a 3′ termination codon (see <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>C). Thus, cells infected with baculovirus carrying this construct produced GFP, but not an IA<sup>b</sup> molecule, because of disruption of the IA<sup>b</sup> β chain gene.</p><p>We then designed a peptide library based on the structure of p3K bound to IA<sup>b</sup> (see <xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>A) We used oligonucleotides with random nucleotides in codons encoding five peptide amino acids (p2, p3, p5, p7, and p8) corresponding to the central surface-exposed amino acids of p3K bound to IA<sup>b</sup>. Other positions were kept identical to p3K, including alanines at the four standard anchor residues at p1, p4, p6, and p9. These oligonucleotides were used in a PCR to create a DNA fragment randomized in these five codons and with 5′- and 3′-end restriction enzyme sites compatible with those in the signal peptide and linker (see <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>D). This fragment was ligated into the restricted plasmid, replacing the GFP sequence and restoring a functional IA<sup>b</sup> β chain gene (see <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>E). The mixture of plasmids was then used to transform <named-content content-type="genus-species">E. coli</named-content> and a bulk plasmid preparation was made. The plasmids were cotransfected with BaculoGold baculovirus DNA into Sf9 insect cells to produce a mixed viral stock in which each virus carried the genes for IA<sup>b</sup> with a different peptide bound. Although it is difficult to calculate the efficiency with which recombination yields infectious baculovirus, we estimate the size of this library was between 3 × 10<sup>4</sup> and 1 × 10<sup>5</sup> independent viruses.</p></sec><sec id="s2f"><title>Successive Enrichment of Baculovirus Carrying Peptide–MHC Combinations That Bind a Particular αβTCR</title><p>A large number of Sf9 insect cells were infected at an MOI of about 1, with baculovirus carrying the IA<sup>b</sup>–peptide library. After 3–4 d, the cells were analyzed with fluorescent B3K-06- or YAe-62-soluble αβTCR, as described above. Fluorescent cells were sorted and cultured with fresh uninfected Sf9 cells to create new infected cells for analysis and an enriched viral stock. This process was repeated three to four times. In each case, when no clear fluorescent population was apparent, the brightest 1% of the infected cells was sorted. In later rounds the majority of the cells in a clearly distinguishable fluorescent population were sorted. <xref ref-type="fig" rid="pbio-0020090-g006">Figure 6</xref> summarizes the successive enrichment of viruses that produced IA<sup>b</sup>–peptide combinations that could be detected with each of the fluorescent αβTCRs. Infected cells binding the B3K-06 αβTCR were apparent only after two rounds of enrichment, but eventually yielded a population with uniform binding (<xref ref-type="fig" rid="pbio-0020090-g006">Figure 6</xref>A). Infected cells that bound the YAe-62 αβTCR were detectable even with the initial library of viruses and enriched rapidly to yield a population with more heterogeneous levels of binding to the receptor (<xref ref-type="fig" rid="pbio-0020090-g006">Figure 6</xref>B).</p><fig id="pbio-0020090-g006" position="float"><label>Figure 6</label><caption><title>Summary of Successive Screening of the IA<sup>b</sup>–Peptide Libraries with Fluorescent αβTCRs</title><p>Sf9 insect cells (1 × 10<sup>7</sup> to 1.5 × 10<sup>7)</sup> were infected at a MOI of approximately 1 with an aliquot of baculovirus encoding the IA<sup>b</sup>–peptide library. After 3 d, the infected cells were analyzed for binding the αβTCR of either B3K-06 or YAe-62. Either obviously fluorescent cells or the brightest 1% of the cells were sorted (2 × 10<sup>4</sup> to 8 × 10<sup>4</sup> cells) and added to 3 × 10<sup>6</sup> fresh Sf9 cells to propagate and reexpress the viruses contained in the sorted cells. These infected cells were then reanalyzed and sorted using the fluorescent αβTCRs. This process was repeated until no further enrichment of αβTCR binding was seen. In most cases, the reanalysis was done directly from the cells that were cocultured with the sorted cells. In a few cases, an intermediate viral stock was made and then used to infect additional Sf9 cells. The turn around time per cycle was 4–7 d. The figure shows the reanalysis in a single experiment of the initial viral stocks and all of the various intermediate enriched viral stocks. Sf9 cells were infected at an MOI of less than 1 with the viral stocks and analyzed as in <xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref> for either B3K-06 (A) or YAe-62 (B) αβTCR binding.</p></caption><graphic xlink:href="pbio.0020090.g006"/></fig></sec><sec id="s2g"><title>Comparison of αβTCR-Selected Peptides from the Library</title><p>At the time of the final enrichment, single infected cells binding each of αβTCRs were sorted into individual wells of 96-well culture plates containing fresh Sf9 cells in order to prepare clonal viral stocks. These stocks were used to infect fresh Sf9 cells, which were reanalyzed for binding to the appropriate αβTCR as in <xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>. Viral DNA from the clones that showed homogeneous TCR binding at a particular level of IA<sup>b</sup> were used as template in a PCR using oligonucleotides that flanked the peptide site in the construct, and a third internal oligonucleotide was used to sequence the PCR fragment. The majority of PCR fragments yielded a single unambiguous peptide sequence. These viruses were used to infect Sf9 cells that expressed mouse ICAM and B7.1. The infected cells were used as APCs for either the B3K-06 or YAe-62 hybridoma, with IL-2 production being a measure of IA<sup>b</sup>–peptide recognition. Viruses expressing IA<sup>b</sup>–peptide combinations that produced high levels of surface IA<sup>b</sup>, but that neither bound to the αβTCR nor stimulated the T cell hybridomas, were used as negative controls, and virus producing IA<sup>b</sup>-p3K was used as the positive control. Results with a few representative virus clones are shown in <xref ref-type="fig" rid="pbio-0020090-g007">Figure 7</xref>A and <xref ref-type="fig" rid="pbio-0020090-g007">7</xref>B, and a summary of all of the results is shown in <xref ref-type="table" rid="pbio-0020090-t001">Table 1</xref>.</p><fig id="pbio-0020090-g007" position="float"><label>Figure 7</label><caption><title>Analysis of Baculovirus Clones from the αβTCR-Enriched IA<sup>b</sup>–Peptide Library</title><p>(A) Sf9 cells were infected with stock from four baculovirus clones (B9, B13, B17, and B23) isolated from the virus pool enriched with the αβTCR of B3K-06. After 3 d, an aliquot of cells from each infection was analyzed as in <xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref> to assure uniform binding of the fluorescent B3K-06 αβTCR (top). Viral DNAs prepared from other aliquots of the cells were used as templates in a PCR with oligonucleotides that flanked the DNA encoding the IA<sup>b</sup>-bound peptide. The fragment was sequenced directly with a third internal oligonucleotide (middle). The clone stock was then used to infect Sf9-ICAM/B7.1 cells. After 3 d, the infected cells were used as APCs for B3K-06 production of IL-2 (bottom). Virus encoding IA<sup>b</sup>-p3K was used as a positive control. Virus encoding pEα was used as the negative control.</p><p>(B) Same as (A), but using YAe-62 and clones (Y2, Y14, Y28, Y52) derived from the IA<sup>b</sup>–peptide library using the YAe-62 αβTCR.</p></caption><graphic xlink:href="pbio.0020090.g007"/></fig><table-wrap id="pbio-0020090-t001" position="float"><label>Table 1</label><caption><title>Summary of Peptides Selected by p3K-Reactive αβTCRs</title></caption><graphic xlink:href="pbio.0020090.t001"/><table-wrap-foot><fn id="nt101"><p> <sup>a</sup>Amino acids homologous to those in p3K are shown in red</p></fn><fn id="nt102"><p> <sup>b</sup>Determined from mean fluorescence as in <xref ref-type="fig" rid="pbio-0020090-g004">Figure 4</xref>B and <xref ref-type="fig" rid="pbio-0020090-g004">4</xref>C</p></fn><fn id="nt103"><p> <sup>c</sup>Sorted by frequency and then by level of TCR binding</p></fn></table-wrap-foot></table-wrap><p>Given our previous data indicating that the B3K-06 αβTCR interacted with all five of the p3K amino acids varied in this library (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>; see also <xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>B), we expected that mimotopes satisfying this receptor would be infrequent or perhaps even absent in a library of this size. Indeed, only one peptide was recovered from the library with the B3K-06 αβTCR, FEAQRARAARVD. It was found in all 42 clones analyzed with unambiguous αβTCR binding and peptide sequences. The sequence of this peptide was strikingly similar to that of p3K. Like p3K, it had a glutamine at p2. It had arginines at positions p3, p5, and p8, corresponding to the lysines found in these positions in p3K, most likely reflecting the importance of the positive charges at these positions. We do not know the relative importance of lysine versus arginine at these three positions, but given that there are two codons for lysine and six for arginine, there was of course a much higher probability of finding arginine than lysine. The most significant difference between this peptide and p3K was an alanine instead of asparagine found at p7.</p><p>When bound to IA<sup>b</sup> on ICAM/ B7.1-expressing Sf9 APCs, FEAQRARAARVD was able to stimulate B3K-06 to produce IL-2, but not nearly as well as did p3K. This loss of stimulating activity was caused by one or more of the lysine-to-arginine substitutions and/or the asparagine-to-alanine substitution at p7. Interestingly, the substitution of alanine for asparagine in p3K eliminated the response of B3K-06 to soluble peptide presented by an IA<sup>b</sup>-bearing mouse APC (see <xref ref-type="fig" rid="pbio-0020090-g001">Figure 1</xref>B). Perhaps the very high density of IA<sup>b</sup>–peptide on the surface of the insect cells allows for responses to peptides that would normally not be stimulatory with peptides presented by conventional APCs, although another possibility is that somehow the arginine (particularly at p8) compensated for the absence of the asparagine sidechain.</p><p>Consistent with the hypothesis that the αβTCR of YAe-62 would be more peptide promiscuous than that of B3K-06, we found 20 different peptide sequences among the analyzed clones that produced an IA<sup>b</sup>–peptide combination that bound the YAe-62 αβTCR. It is likely that many more would be identified if more clones were analyzed. Five sequences were found multiple times. Not unexpectedly, these were among those that bound the YAe-62 αβTCR most strongly. There was a 100-fold range in the intensity of αβTCR binding to the different IA<sup>b</sup>–peptide combinations, ranging from about 4-fold to 400-fold binding above that seen with a negative control peptide. One obvious property of these peptides stands out. There appeared to be a very strong selection for a basic amino acid at position 5. In 16 of 20 of the peptides, a lysine, arginine, or histidine was found at position 5, matching the lysine found in p3K. As a control, we sequenced random clones picked either from the original <named-content content-type="genus-species">E. coli</named-content> construction of the library (17 clones) or from the baculovirus library that expressed IA<sup>b</sup>–peptide well, but did not bind either αβTCR (11 clones). The frequencies of basic amino acids at p5 in these sequences were only 12% and 9%, respectively (data not shown).</p><p>There was no strong selection for amino acids homologous to those of p3K at positions p2, p3, p7, or p8. The amino acids at positions p2 and p3 appear nearly random, suggesting little or no essential contact between this part of the peptide–MHC ligand and the receptor, although these positions may contribute to the wide range of apparent αβTCR affinities seen. While not homologous to the asparagine in p3K, leucine was found at p7 in six of 20 (30.0%) of the YAe-62 αβTCR-selected peptides and three of 11 (27.2%) of the IA<sup>b</sup>-binding peptides that were not bound by the YAe-62 αβTCR, but only two of 17 (11.8%) of the random <named-content content-type="genus-species">E. coli</named-content> plasmids. The amino acid in this position is only partially exposed on the surface and can contribute significantly to peptide–MHC interaction (<xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>). After asparagine, leucine is the most common amino acid found at this position in peptides found naturally bound to IA<sup>b</sup> (<xref rid="pbio-0020090-Dongre1" ref-type="bibr">Dongre et al. 2001</xref>; <xref rid="pbio-0020090-Liu3" ref-type="bibr">Liu et al. 2002</xref>). Therefore, although more data would be required to test its significance, there may have been some slight enrichment of leucine at p7 in the expressed library prior to αβTCR selection, reflecting the role of p7 in stable peptide binding to IA<sup>b</sup>.</p><p>The amino acid at position p8 is predicted to be fully surface exposed. In the selected peptides, rather than an amino acid homologous to the lysine of p3K, there may be an overrepresentation of amino acids with small neutral sidechains (threonine, serine, alanine, glycine) at this position. Perhaps this indicates that, in general, larger sidechains can be inhibitory at this position, but again more data would be required to test this idea.</p><p>The 12 IA<sup>b</sup>–peptide combinations that bound the YAe-62 αβTCR most strongly were also the ones that were able to induce IL-2 production from YAe-62. Among these, a number with the very highest apparent affinities stimulated YAe-62 better than did p3K. However, there was not a direct correlation between apparent affinity and the level of IL-2 production; i.e., several peptides that yielded complexes with IA<sup>b</sup> with about the same apparent affinity for the αβTCR nevertheless stimulated very different levels of IL-2 production from YAe-62. This may be related to the phenomenon of altered peptide ligands (<xref ref-type="sec" rid="s3">see Discussion</xref>).</p><p>Overall, our results supported our original prediction that for conventional T cells, such as B3K-06, most of the surface-exposed residues of the peptide would be important in peptide–MHC recognition, while for broadly allo-MHC-reactive T cells, such as YAe-62, peptide recognition would be much more promiscuous.</p></sec></sec><sec id="s3"><title>Discussion</title><p>The peptide degeneracy allowed for a given αβTCR–MHC combination has been a subject of study over many years. While minor changes in the exposed amino acids sidechains of the peptide can often destroy αβTCR recognition, usually at least some variation is tolerated within the predicted footprint of the αβTCR on the peptide–MHC ligand (<xref rid="pbio-0020090-Evavold1" ref-type="bibr">Evavold and Allen 1992</xref>; <xref rid="pbio-0020090-Reay1" ref-type="bibr">Reay et al. 1994</xref>). We can understand this flexibility to some extent from the X-ray structures of αβTCR–MHC–peptide complexes that show poor or even absent interactions between some peptide amino acid sidechains and the complementarity region (CDR) loops of the receptor (reviewed in <xref rid="pbio-0020090-Garcia1" ref-type="bibr">Garcia et al. 1999</xref>).</p><p>We have reported the properties of mice that have been genetically manipulated to express their MHCII molecules virtually completely occupied by a single peptide (<xref rid="pbio-0020090-Ignatowicz1" ref-type="bibr">Ignatowicz et al. 1996</xref>; <xref rid="pbio-0020090-Marrack1" ref-type="bibr">Marrack et al. 2001</xref>). One of the most unusual features of the repertoire of T cells that develop in these animals is that they show an unusually high frequency of broadly allo-MHC–self-MHC-reactive T cells. These T cells are lost when these animals are repopulated with MHCII wild-type bone marrow cells. We have concluded that cells of this type are commonly positively selected in normal animals, but to a large extent negatively selected by self-MHC occupied by a variety of self-peptides. Their survival in single peptide–MHC mice may reflect the need for many different peptides to expose all MHC amino acids and their various conformers during T cell-negative selection. We have proposed that the αβTCRs of these cells are focused on the common conserved features of peptide–MHC complexes rather than on the specific sidechains of the exposed amino acids of the peptide (<xref rid="pbio-0020090-Marrack1" ref-type="bibr">Marrack et al. 2001</xref>). A consequence of this hypothesis is the prediction that T cells of this sort should be much more peptide promiscuous than conventional T cells.</p><p>The experiments reported here were designed to test this prediction by comparing the peptide promiscuity of one of these broadly allo-reactive T cells, YAe-62, typical of T cells from these mice, to that of a T cell with the same nominal specificity produced by immunization of conventional mice. The results support the conclusion that the broadly allo-reactive T cell has a much greater peptide promiscuity than does the conventional T cell. This question of T cell promiscuity is an important one in that it addresses the existence of a very large set of TCRs that apparently make it through positive selection, but never see the light of day in normal animals, because they are negatively selected on self-MHC with little input from the MHC-bound peptide. Thus, studying the peripheral fully negatively selected T cell repertoire gives a false impression of the interaction requirements necessary or sufficient for positive selection. These promiscuous T cells may also give us insight into possible evolutionary conserved αβTCR–MHC interactions that have been hard to sort out with conventional T cells.</p><p>While perhaps much less frequent than in single peptide–MHC mice, peptide-promiscuous T cells have been described in normal individuals (<xref rid="pbio-0020090-Brock1" ref-type="bibr">Brock et al. 1996</xref>). Consistent with the idea that this property may be linked to allo-MHC reactivity, in a parallel study we have shown that peptide-promiscuous T cells are enriched in normal mice in the population of T cells reactive to foreign MHC alleles and isotypes (<xref rid="pbio-0020090-Huseby1" ref-type="bibr">Huseby et al. 2003</xref>).</p><p>In order to study the relationship between peptide sequence and αβTCR recognition, we developed a baculovirus-based display method for rapid identification of peptides that form complexes with MHC that bind a particular αβTCR. Display is one of the most powerful library techniques available. Its underlying principle is that the protein or peptide members of the library are expressed on the surface of organisms that harbor the DNA encoding them. A binding assay that isolates all members of the library with the appropriate properties copurifies the organism and the encoding DNA. The DNA is then amplified and reexpressed and the process repeated as many time as necessary to enrich fully the relevant molecules, whose sequence can be deduced from the copurified DNA. The great advantage of display libraries is that all members of the library that satisfy the screening conditions are enriched simultaneously without the need to identify them one by one.</p><p>In order for peptides to be tested for αβTCR binding, they must be complexed with the relevant MHC molecule on a platform suitable for interaction with the T cell and/or its αβTCR. For display libraries, one aspect of this problem has been solved by the ability to express MHC molecules with sequence for a covalently attached antigenic peptide imbedded in the MHC genes (<xref rid="pbio-0020090-Kozono1" ref-type="bibr">Kozono et al. 1994</xref>; <xref rid="pbio-0020090-Mottez1" ref-type="bibr">Mottez et al. 1995</xref>; <xref rid="pbio-0020090-Uger1" ref-type="bibr">Uger and Barber 1998</xref>; <xref rid="pbio-0020090-White1" ref-type="bibr">White et al. 1999</xref>). However, the most commonly used bacterial display systems do not yet allow for the assembly and display of complex multichain MHC molecules. There is a recent report of the successful display of a single-chain peptide–MHCI on yeast cells (<xref rid="pbio-0020090-Brophy1" ref-type="bibr">Brophy et al. 2003</xref>), but our own previous attempts with yeast had failed to yield displayed peptide–MHCII in a form capable of recognition by T cell hybridomas (data not shown). Our previous success with producing soluble MHC and αβTCR molecules using a baculovirus expression system and a report of peptide libraries displayed in baculovirus (<xref rid="pbio-0020090-Ernst2" ref-type="bibr">Ernst et al. 1998</xref>) led us to adapt these methods for surface display of peptide–MHCII on insect cells. We constructed a library of peptides attached to the displayed mouse class II molecule, IA<sup>b</sup>. Using fluorescently labeled multimeric soluble αβTCRs as bait and insect cells infected with the IA<sup>b</sup>–peptide library as fish, we were able to identify rapidly the members of the library that encoded peptide mimotopes for two αβTCRs.</p><p>In these studies, the immunizing peptide (epitope) for the αβTCR was already known. However, this method should be useful as well in identifying mimotopes for αβTCRs whose peptide epitope is not known, provided that suitable peptide anchor residues for MHC binding are known. One limitation of this display method as presented here is the size of the peptide library. The bottlenecks caused by the preparation of the library in an <named-content content-type="genus-species">E. coli</named-content> plasmid and then its introduction into baculovirus by homologous recombination resulted in a library with only 3 × 10<sup>4</sup> to 1 × 10<sup>5</sup> members. This is far below the size required to have all 3.2 × 10<sup>6</sup> versions of the peptide present when randomizing five amino acids. Large libraries of this size require more efficient baculovirus-cloning methods, such as incorporation of DNA fragments directly into baculovirus DNA by ligation (<xref rid="pbio-0020090-Ernst1" ref-type="bibr">Ernst et al. 1994</xref>) or in vitro recombinase-mediated recombination (<xref rid="pbio-0020090-Peakman1" ref-type="bibr">Peakman et al. 1992</xref>). In preliminary experiments, we have constructed an IA<sup>b</sup>–peptide library with over 10<sup>7</sup> clones by directly ligating (<xref rid="pbio-0020090-Ernst1" ref-type="bibr">Ernst et al. 1994</xref>) a randomized PCR DNA fragment encoding the peptide into linearized baculovirus DNA using unique homing restriction enzyme (SceI–CeuI) sites introduced flanking the peptide-encoding region of the construct (data not shown). Since recircularized baculovirus DNA is directly infectious when introduced into insect cells by transfection, there is no theoretical reason why this method cannot be used to create libraries as large as those reported for yeast and phage.</p><p>We have developed this method using IA<sup>b</sup> as the displayed MHCII molecule carrying the peptide library. However, using the same strategy, we have successfully displayed numerous other MHCII molecules, such as murine IE<sup>k</sup> and human DR4, DR52c, and DP2 (data not shown). While the leucine zippers that we included in this construct are not strictly required for expression of IA<sup>b</sup>, they have helped considerably in expression of some of these other MHCII molecules. Moreover, we (<xref rid="pbio-0020090-White1" ref-type="bibr">White et al. 1999</xref>) and others (<xref rid="pbio-0020090-Mottez1" ref-type="bibr">Mottez et al. 1995</xref>; <xref rid="pbio-0020090-Uger1" ref-type="bibr">Uger and Barber 1998</xref>) have shown that peptides can be tethered to MHCI molecules via the N-terminus of either β2m or the heavy chain, making this approach feasible for searching for MHCI-bound peptide mimotopes as well. In preliminary experiments we have successfully displayed on the surface of Sf9 cells the mouse MHCI molecule, D<sup>d</sup>, with a β2m-tethered peptide from HIV gp120 (data not shown). Given that baculovirus has been such a successful expression system for many different types of complex eukaryotic proteins that express or assemble poorly in <named-content content-type="genus-species">E. coli</named-content>, this method may have broad applications to other receptor–ligand systems.</p><p>As opposed to methods that use T cell activation as the peptide-screening method, an advantage of display methods that use flow cytometry for screening and enrichment is that the strength of binding of receptor and ligand can be estimated and manipulated. In the results reported here, by limiting the analysis to insect cells bearing a particular level of peptide–MHC, a uniform level of αβTCR binding was seen for an individual peptide sequence, but the strength of binding varied over two orders of magnitude for different peptides, presumably reflecting the relative affinity of the receptor for different IA<sup>b</sup>–peptide combinations. Thus, depending on whether one was interested in high- or low-affinity ligands for the αβTCR, one could enrich for peptides with these properties directly during the screening of the library. Such an approach has been used with antibody (<xref rid="pbio-0020090-Boder1" ref-type="bibr">Boder and Wittrup 2000</xref>) and αβTCR (<xref rid="pbio-0020090-Shusta1" ref-type="bibr">Shusta et al. 2000</xref>) variants displayed on yeast to select directly for receptors with increased affinity.</p><p>It is worth noting that there was not a direct correlation between the strength of αβTCR binding to a particular peptide–MHC combination and the subsequent level of IL-2 secretion seen from the T cell responding to this combination. While in general the best IL-2 secretion was obtained with complexes with the highest apparent affinities, some IA<sup>b</sup>–peptide combinations with apparent high affinity stimulated IL-2 production poorly. One interesting possibility is that this observation is related to the phenomenon of altered peptide ligands in which amino acid variants of fully immunogenic peptides only partially activate or even anergize the T cell (<xref rid="pbio-0020090-Evavold2" ref-type="bibr">Evavold et al. 1993</xref>), despite minor differences in affinity. In some cases, this phenomenon has been related to αβTCR binding kinetics, rather than just overall affinity (<xref rid="pbio-0020090-Lyons1" ref-type="bibr">Lyons et al. 1996</xref>). Future experiments using surface plasmon resonance or fluorescence peptide–MHC multimers might help to test this idea.</p><p>In summary, the very properties that have made baculovirus a very successful expression system for complex eukaryotic proteins also make it suitable for library display methods, with potential application not only in T cell epitope/mimotope discovery, characterization, and manipulation, but also in studying a wide variety of other protein–protein interactions.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><sec id="s4a1"><title>Synthetic peptides, oligonucleotides, and DNA sequencing</title><p>The peptides pEα (FEAQGALANIAVD), p3K (FEAQKAKANKAVD), and various alanine-substituted variants of p3K were synthesized in the Molecular Resource Center of the National Jewish Medical and Research Center (Denver, Colorado, United States), as were all oligonucleotides used in PCR and DNA sequencing. Automated DNA sequencing was also performed in this facility.</p></sec><sec id="s4a2"><title>Cell lines and T cell hybridomas</title><p>The insect cell lines Sf9 and High Five were obtained from Invitrogen (Carlsbad, California, United States). The IA<sup>b</sup>-p3K-reactive T cell hybridoma B3K-06 was produced from C57BL/6 mice as previously described (<xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>). The IA<sup>b</sup>-expressing B cell hybridoma LB-15.13 (<xref rid="pbio-0020090-Kappler1" ref-type="bibr">Kappler et al. 1982</xref>) was used to present soluble peptides to B3K-06.</p><p>The T cell hybridoma YAe-62 (<xref rid="pbio-0020090-Marrack1" ref-type="bibr">Marrack et al. 2001</xref>) was produced from previously described (<xref rid="pbio-0020090-Ignatowicz1" ref-type="bibr">Ignatowicz et al. 1996</xref>) C57BL/6 mice that lacked expression of the endogenous IA<sup>b</sup> β gene (ΔIAβ) and the invariant chain (ΔIi) and that carried a transgene for the IA<sup>b</sup> β gene that was modified to insert sequence encoding pEα and a flexible linker between the signal peptide and the N-terminus of the β chain. These mice were immunized intravenously with 3 × 10<sup>6</sup> dendritic cells from ΔIAβ/ΔIi C57BL/6 mice. These cells had been retrovirally transduced (<xref rid="pbio-0020090-Mitchell1" ref-type="bibr">Mitchell et al. 2001</xref>; <xref rid="pbio-0020090-Schaefer1" ref-type="bibr">Schaefer et al. 2001</xref>) with the IA<sup>b</sup> β gene that was modified as above to express with a tethered p3K. T cells from these immunized mice were propagated in vitro and converted to T cell hybridomas, by standard techniques (<xref rid="pbio-0020090-White2" ref-type="bibr">White et al. 2000</xref>). The hybridomas were initially screened for binding of multivalent, fluorescent IA<sup>b</sup>-p3K (<xref rid="pbio-0020090-Crawford1" ref-type="bibr">Crawford et al. 1998</xref>; <xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>) and subsequently for IL-2 production in response to immobilized, soluble IA<sup>b</sup>-p3K, but not to spleen cells from the host ΔIAβ/ΔIi IA<sup>b</sup>-pEα transgenic mice. Further characterization of YAe-62 is described in the Results.</p></sec><sec id="s4a3"><title>Soluble αβTCRs</title><p>cDNA, prepared from B3K-06 and YAe-62, was used as template in a PCR using oligonucleotides that flanked the Vα and Vβ regions and introduced restriction enzyme sites that allowed cloning of the PCR fragments into a previously described baculovirus expression vector for soluble αβTCRs (<xref rid="pbio-0020090-Kappler2" ref-type="bibr">Kappler et al. 1994</xref>). The cloned fragments were sequenced and incorporated into baculovirus and αβTCRs were purified from the supernatants of infected High Five cells. For B3K-06, the α chain was AV0401/AJ27 and the CDR3 sequence was CALVISNTNKVVFGTG. The β chain was BV0801/BJ0103 and the CD3 sequence was CASIDSSGNTLYFGEG. For YAe-62, the α chain was AV0412/AJ11 and the CD3 sequence was CAANSGTYQRFGTG. The β chain was BV0802/JD0204 and the CD3 sequence was CASGDFWGDTLYFGAG.</p></sec><sec id="s4a4"><title>Expression of ICAM and B7.1 in Sf9 cells</title><p>DNA fragments encoding the baculovirus hr5 enhancer element, IE1 gene promoter, and IEI poly(A) addition region were synthesized by PCR using baculovirus DNA as template. The fragments were used to construct an insect cell expression vector (pTIE1) on a pTZ18R (Pharmacia, Uppsala, Sweden) backbone with the hr5 enhancer at the 5′-end, followed by the IE1 promoter, a large multiple cloning site (Esp3I, MunI, SalI, XhoI, BsrGI, HpaI, SpeI, BstXI, BamHI, BspEI, NotI, SacII, XbaI), and the IE1 poly(A) addition region. The complete sequence of the pTIE1 vector has been deposited in GenBank (see Supporting Information). DNA fragments encoding mouse ICAM and B7.1 were cloned between the XhoI and NotI sites of the multiple cloning site. Sf9 cells were transfected with a combination of the plasmids by the standard calcium phosphate method and cells expressing both molecules on their surfaces were cloned without selection at limiting dilution to establish the line Sf9-ICAM/B7.1.</p></sec><sec id="s4a5"><title>IL-2 assays</title><p>T cell hybridoma cells (10<sup>5</sup>) were added to microtiter wells containing either (1) saturating immobilized peptide–MHC, (2) 10 μg/ml peptide plus 10<sup>5</sup> LB-15.13 cells, (3) 5 × 10<sup>4</sup> Sf9-ICAM/B7.1 insect cells infected 3 d previously with baculovirus encoding a displayed peptide–MHC, (4) 10<sup>6</sup> spleen cells from IA<sup>b</sup>-pEα single peptide mice, or (5) 10<sup>6</sup> spleen cells from various knockout or MHC congenic mice. After overnight incubation the culture supernatants were assayed for IL-2 as previously described (<xref rid="pbio-0020090-White2" ref-type="bibr">White et al. 2000</xref>).</p></sec><sec id="s4a6"><title>mAbs and flow cytometry</title><p>The following mAbs were used in these studies: 17/227, a mouse IgG2a antibody, specific for IA<sup>b</sup> (<xref rid="pbio-0020090-Lemke1" ref-type="bibr">Lemke et al. 1979</xref>); ADO-304, an Armenian hamster antibody specific for an epitope on the αβTCR Cα region not accessible on the surface of T cells, but exposed on recombinant αβTCR and on CD3-dissociated, NP-40-solublized natural αβTCR (<xref rid="pbio-0020090-Liu1" ref-type="bibr">Liu et al. 1998</xref>); 3E2 (PharMingen, San Diego, California, United States), specific for mouse ICAM (CD54); and 16–10A1 (PharMingen), specific for mouse B7.1 (CD80). For flow cytometry, an unlabeled version of 17/227 was used with phycoerythrin-coupled goat anti-mouse IgG2a (Fisher Biotech, Foster City, California, United States).</p><p>To assemble multivalent fluorescent versions of the soluble αβTCRs, first a biotinylated version of ADO-304 was prepared. In brief, purified ADO-304 at 1–3 mg/ml in 0.1 M NaHCO<sub>3</sub> was labeled with Sulfo-NHS-LC-Biotin (Pierce Chemical Company, Rockford, Illinois, United States) at a molar ratio of 2.5:1 (biotin:antibody) for 4 h at room temperature. The reaction was quenched with 0.1 M lysine and the product dialyzed extensively against PBS. The resulting derivative contained about one biotin per molecule of mAb. The biotinylated mAb was complexed in excess with AlexaFlour647–streptavidin (Molecular Probes, Eugene, Oregon, United States). The complex was separated from the free biotin–antibody using Superdex-200 size exclusion chromatography (Pharmacia). In preliminary experiments, the amount of soluble αβTCR required to saturate an aliquot of a large single batch of this reagent was determined. To prepare the multivalent αβTCR, the appropriate amount of soluble αβTCR was mixed with an aliquot of the fluorescent anti-Cα reagent overnight. For staining for flow cytometry, this mix was used without further purification. Each 100 μl sample contained approximately 2 μg of the fluorescent reagent plus 10<sup>5</sup> Sf9 insect cells. This mixture was incubated at 27°C for 1–2 h. The cells were then washed for analysis. The advantages of this method for preparing fluorescent multimers over using direct enzymatic biotinylation (<xref rid="pbio-0020090-Schatz1" ref-type="bibr">Schatz 1993</xref>) of the αβTCR were that only one fluorescent reagent needed to be prepared for all αβTCRs, the mAb–streptavidin complex was very stable over a long period of time, and no special peptide-tagged version of the soluble αβTCR was required.</p><p>Analytical flow cytometry was performed with a FacsCaliber flow cytometer (Becton-Dickinson, Palo Alto, California, United States). For sorting, a MoFlo instrument was used (Dako/Cytomation, Glostrup, Denmark).</p></sec><sec id="s4a7"><title>IA<sup>b</sup> and peptide library constructions</title><p>For displaying IA<sup>b</sup> on the surface of baculovirus-infected insect cells, modifications were made as described in <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>A and <xref ref-type="fig" rid="pbio-0020090-g002">2</xref>B to a previously reported baculovirus construct for producing soluble IA<sup>b</sup>-p3K (<xref rid="pbio-0020090-Rees1" ref-type="bibr">Rees et al. 1999</xref>). Other versions of this construction were prepared encoding other IA<sup>b</sup>-binding peptides. The constructions were incorporated into baculovirus by homologous recombination using the BaculoGold system (PharMingen).</p><p>As described in <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>C, this construction was altered in the <named-content content-type="genus-species">E. coli</named-content> transfer plasmid to replace the portion encoding p3K with sequence encoding enhanced GFP, flanked by sites for the restriction enzymes SbfI and CeuI. A PCR fragment was produced as described in <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>D that encoded an IA<sup>b</sup>-binding peptide randomized at positions p2, p3, p5, p7, and p8, but identical to p3K at other positions. This sequence was flanked by sites for the restriction enzymes PstI and BstXI, such that the cohesive ends generated by these enzymes were compatible with those generated by SbfI and CeuI in the GFP-containing plasmid. Cloning the restricted fragment into this site regenerated a covalent peptide in-frame with the signal peptide and flexible linker of the IA<sup>b</sup> β chain (see <xref ref-type="fig" rid="pbio-0020090-g002">Figure 2</xref>E). After ligation of the fragment into this plasmid, a bulk transformation was performed using XL1-Blue <named-content content-type="genus-species">E. coli</named-content> (Stratagene, La Jolla, California, United States). An estimated 3 × 10<sup>4</sup> to 10 × 10<sup>4</sup> independent transformants were obtained that were used to make a mixed plasmid preparation. This mixture was incorporated into baculovirus by homologous recombination as above. In order to assure a high efficiency of conversion of plasmids to virus, 1.5 × 10<sup>7</sup> Sf9 cells were cotransfected with 6 μg of the plasmid mixture and 1.5 μg of BaculoGold DNA.</p></sec></sec></sec><sec sec-type="supplementary-material" id="s5"><title>Supporting Information</title><sec id="s5a"><title>Accession Numbers</title><p>The GenBank (<ext-link ext-link-type="uri" xlink:href="http://www.ncbi.nlm.nih.gov/Genbank/)">http://www.ncbi.nlm.nih.gov/Genbank/)</ext-link> accession numbers for the sequences described in this paper are B7.1 (AJ278965), baculovirus DNA (L22858), ICAM (X52264), and pTIE1 vector (AY522575).</p></sec></sec>
Activation of Arp2/3 Complex: Addition of the First Subunit of the New Filament by a WASP Protein Triggers Rapid ATP Hydrolysis on Arp2	<p>In response to activation by WASP-family proteins, the Arp2/3 complex nucleates new actin filaments from the sides of preexisting filaments. The Arp2/3-activating (VCA) region of WASP-family proteins binds both the Arp2/3 complex and an actin monomer and the Arp2 and Arp3 subunits of the Arp2/3 complex bind ATP. We show that Arp2 hydrolyzes ATP rapidly—with no detectable lag—upon nucleation of a new actin filament. Filamentous actin and VCA together do not stimulate ATP hydrolysis on the Arp2/3 complex, nor do monomeric and filamentous actin in the absence of VCA. Actin monomers bound to the marine macrolide Latrunculin B do not polymerize, but in the presence of phalloidin-stabilized actin filaments and VCA, they stimulate rapid ATP hydrolysis on Arp2. These data suggest that ATP hydrolysis on the Arp2/3 complex is stimulated by interaction with a single actin monomer and that the interaction is coordinated by VCA. We show that capping of filament pointed ends by the Arp2/3 complex (which occurs even in the absence of VCA) also stimulates rapid ATP hydrolysis on Arp2, identifying the actin monomer that stimulates ATP hydrolysis as the first monomer at the pointed end of the daughter filament. We conclude that WASP-family VCA domains activate the Arp2/3 complex by driving its interaction with a single conventional actin monomer to form an Arp2–Arp3–actin nucleus. This actin monomer becomes the first monomer of the new daughter filament.</p>	<contrib contrib-type="author"><name><surname>Dayel</surname><given-names>Mark J</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Mullins</surname><given-names>R. Dyche</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>The actin cytoskeleton determines the shape, mechanical properties, and motility of most eukaryotic cells. To change shape and to move, cells precisely control the location and timing of actin filament assembly by regulating the number of fast-growing (barbed) filament ends (<xref rid="pbio-0020091-Pollard1" ref-type="bibr">Pollard et al. 2000</xref>). The actin-related protein (Arp) 2/3 complex, a seven-subunit protein complex that contains two actin-related proteins, generates these new barbed ends in response to cellular signals (<xref rid="pbio-0020091-Welch1" ref-type="bibr">Welch et al. 1998</xref>; <xref rid="pbio-0020091-Machesky1" ref-type="bibr">Machesky et al. 1999</xref>; <xref rid="pbio-0020091-Rohatgi1" ref-type="bibr">Rohatgi et al. 1999</xref>). In a process called “dendritic nucleation,” the Arp2/3 complex nucleates new actin filaments from the sides of preexisting filaments to produce a rigid and highly crosslinked filament array (<xref rid="pbio-0020091-Mullins1" ref-type="bibr">Mullins et al. 1998</xref>; <xref rid="pbio-0020091-Machesky1" ref-type="bibr">Machesky et al. 1999</xref>; <xref rid="pbio-0020091-Blanchoin3" ref-type="bibr">Blanchoin et al. 2000a</xref>). Such crosslinked arrays form the core of many motile cellular structures, including the leading edges of amoeboid cells and the actin comet tails that propel endosomes and bacterial pathogens through eukaryotic cytoplasm. To understand the construction, function, and regulation of these structures, it is important to understand the molecular mechanism of Arp2/3 activation.</p><p>The Arp2/3 complex must be activated by both a member of the Wiskott–Aldrich syndrome protein (WASP) family and a preexisting actin filament before it will nucleate a new actin filament (<xref rid="pbio-0020091-Machesky1" ref-type="bibr">Machesky et al. 1999</xref>; <xref rid="pbio-0020091-Blanchoin5" ref-type="bibr">Blanchoin et al. 2001</xref>; <xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>). The structure and the orientation of the Arp2 and Arp3 subunits within the crystal structure of the complex suggest that these subunits may nucleate a new filament by forming an actin-like heterodimer that mimics the barbed end of an actin filament (<xref rid="pbio-0020091-Robinson1" ref-type="bibr">Robinson et al. 2001</xref>). In the crystal structure of the unactivated complex, however, Arp2 and Arp3 are separated by 40 Å so that formation of an actin-like dimer would require a conformational change (<xref rid="pbio-0020091-Robinson1" ref-type="bibr">Robinson et al. 2001</xref>). Binding of the Arp2/3 complex to both a preformed filament and a WASP-family protein is thought to drive at least part of this conformational change (<xref rid="pbio-0020091-Blanchoin5" ref-type="bibr">Blanchoin et al. 2001</xref>; <xref rid="pbio-0020091-Marchand1" ref-type="bibr">Marchand et al. 2001</xref>; <xref rid="pbio-0020091-Panchal1" ref-type="bibr">Panchal et al. 2003</xref>). The Arp2/3-activating region of WASP-family proteins, also known as the VCA domain, is composed of three sequences arranged in tandem: (1) an actin-binding verprolin-homology (or V) domain (also known as a WASP-homology 2 [WH2] domain), (2) a conserved “connecting” (or C) region that interacts with both the Arp2/3 complex and monomeric actin (<xref rid="pbio-0020091-Marchand1" ref-type="bibr">Marchand et al. 2001</xref>), and (3) an acidic (or A) region that binds the Arp2/3 complex. This VCA domain is both necessary and sufficient for efficient Arp2/3 activation. We and others have previously suggested that an actin monomer provided by the VCA domain to the Arp2/3 complex may drive the formation of an Arp2–Arp3–actin heterotrimer and form a nucleus for actin polymerization (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>; <xref rid="pbio-0020091-Marchand1" ref-type="bibr">Marchand et al. 2001</xref>).</p><p>Both the Arp2 and Arp3 subunits of the complex bind ATP (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>). Hydrolysis of this ATP could be used to perform work, to provide a signal, or, like the guanine triphosphate (GTP) bound to the α subunit of tubulin heterodimers, may simply stabilize a protein fold. On conventional actin, ATP hydrolysis is a timing mechanism that promotes construction of dynamic and polarized filament networks. Actin rapidly hydrolyzes ATP upon polymerization (<xref rid="pbio-0020091-Blanchoin2" ref-type="bibr">Blanchoin and Pollard 2002</xref>) and releases bound phosphate several hundred seconds later (<xref rid="pbio-0020091-Melki1" ref-type="bibr">Melki et al. 1996</xref>). ATP hydrolysis and phosphate dissociation do not cause immediate filament disassembly, but enable interaction with depolymerizing factors such as cofilin (<xref rid="pbio-0020091-Blanchoin1" ref-type="bibr">Blanchoin and Pollard 1999</xref>). ATP hydrolysis by actin thereby determines the overall rate of filament turnover.</p><p>We show here that the Arp2/3 complex rapidly hydrolyzes ATP on the Arp2 subunit upon filament nucleation. There are several events in the Arp2/3 nucleation reaction that might trigger ATP hydrolysis on Arp2: (1) binding of VCA to the Arp2/3 complex, (2) binding of VCA-Arp2/3 to the side of a preformed filament, (3) binding of a VCA-tethered actin monomer to the Arp2/3 complex, or (4) binding of a second or third actin monomer to form a stable daughter filament. We find that ATPase activity requires the combination of a preformed actin filament, a VCA domain, and an actin monomer, but does not require actin polymerization. This indicates that hydrolysis is triggered relatively early in the nucleation reaction—before completion of a stable daughter filament. Capping the pointed ends of actin filaments also stimulates Arp2 to rapidly hydrolyze ATP in the absence of monomeric actin and VCA and without branch formation. Thus, ATP hydrolysis on Arp2 is stimulated directly by interaction with conventional actin, presented to the complex either as a monomer attached to the VC domain of the WASP-family protein or as one of the subunits making up the pointed end of a preformed filament. To our knowledge this is the first direct evidence that the monomer supplied by the VCA domain is the first monomer of the new daughter filament. From these observations we propose a model for the mechanism of Arp2/3 complex activation by WASP-family proteins.</p></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>γ-<sup>32</sup>P-AzidoATP Can Be Covalently Crosslinked to Arp2 and Arp3 with Approximately Equal Efficiency</title><p>Previously we used sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) to show that UV irradiation covalently crosslinks α-<sup>32</sup>P-8-AzidoATP to the Arp2 and Arp3 subunits of the Arp2/3 complex (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>). Here we crosslink γ-<sup>32</sup>P-AzidoATP instead of α-<sup>32</sup>P to Arp2 to measure ATPase activity. Using SDS-PAGE, we can separate the subunits and simultaneously monitor cleavage of the labeled γ-phosphate from ATP bound to both Arp2 and Arp3. This technique allows us to measure ATP hydrolysis specifically on the Arp2/3 complex in spite of a 100-fold molar excess of actin, which also binds and hydrolyzes ATP. We crosslinked γ-<sup>32</sup>P-AzidoATP to the Arp2/3 complex by brief (9 s) exposure to UV light. In the presence of γ-<sup>32</sup>P-AzidoATP at concentrations above the K<sub>D</sub> for ATP (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>), γ-<sup>32</sup>P-AzidoATP crosslinks to both Arp2 and Arp3 with approximately equal efficiency (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>A). Addition of large amounts of monomeric actin to the labeled Arp2/3 distorts the shape of the Arp2 band, but the <sup>32</sup>P signal from Arp2 remains separately quantifiable, and the magnitude is unaffected (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>A). The efficiency of crosslinking for both Arp2 and Arp3 is approximately 10% (unpublished data); therefore, only 1% of the Arp2/3 complex has γ-<sup>32</sup>P-AzidoATP crosslinked to both Arp2 and Arp3. For simplicity, we refer to this partially crosslinked Arp2/3 complex as γ-<sup>32</sup>P-AzidoATP-Arp2/3. Reactions using γ-<sup>32</sup>P-AzidoATP-Arp2/3 are performed in the presence of 100 μM ATP, to occupy the noncrosslinked sites and ensure 100% of the Arp2/3 complex is active.</p><fig id="pbio-0020091-g001" position="float"><label>Figure 1</label><caption><title>Arp2 Hydrolyzes ATP Rapidly upon Filament Nucleation</title><p>(A) Arp2/3 (2 μM) was covalently crosslinked to γ-<sup>32</sup>P-AzidoATP by exposure to UV light. Both Arp2 and Arp3 crosslink with approximately equal efficiency (lane 1). Addition of 100-fold excess monomeric actin (lane 2) distorts the shape of the Arp2 band, but the Arp2 signal remains separate and quantifiable.</p><p>(B–E) γ-<sup>32</sup>P-AzidoATP-Arp2/3 (20 nM) was mixed with 2 μM monomeric actin in polymerization buffer. Samples were taken before and at indicated times after the addition of 750 nM VCA, which initiates rapid actin-filament nucleation by the Arp2/3 complex.</p><p>(B) Subunits were separated by SDS-PAGE and stained with Coomassie.</p><p>(C) <sup>32</sup>P signal shows remaining uncleaved γ-<sup>32</sup>P on Arp2 and Arp3 subunits. Arp2 rapidly loses γ-<sup>32</sup>P after addition of VCA.</p><p>(D) Cleaved γ-<sup>32</sup>P was separated from free <sup>32</sup>P-ATP and protein-<sup>32</sup>P-ATP by TLC.</p><p>(E) Quantitation of (B) to (D): Protein-ATP (closed circle), Cleaved Pi (closed square), Free ATP (closed diamonds), and Arp2-ATP from SDS-PAGE (open circle, normalized separately).</p><p>(F) Arp2 releases phosphate soon after ATP hydrolysis. Reaction conditions were the same as (B)–(E), but with the addition of 2 mM maltose and 2 U/ml maltose phosphorylase. Timepoints were quenched into formic acid and assayed by TLC. Hydrolyzed <sup>32</sup>P-ATP was quantified from the decrease in protein conjugated <sup>32</sup>P, and released <sup>32</sup>P was quantified from the <sup>32</sup>P-glucose phosphate produced.</p></caption><graphic xlink:href="pbio.0020091.g001"/></fig></sec><sec id="s2b"><title>Arp2 Hydrolyzes ATP Rapidly upon Actin Filament Nucleation</title><p>We mixed 20 nM γ-<sup>32</sup>P-AzidoATP-Arp2/3 with 2 μM monomeric actin in polymerization buffer and initiated polymerization by adding 750 nM VCA, which activates rapid actin filament nucleation by the Arp2/3 complex (t<sub>½ actin polymerization</sub> ≈ 20 s; unpublished observations). (Unless otherwise stated, VCA refers to 6-histidine [6His]-N-WASP-VCA [398-502]. Cleavage of the 6His tag did not affect the kinetics of Arp2/3-mediated actin polymerization [unpublished data].) We assayed timepoints both by SDS-PAGE and thin-layer chromatagraphy (TLC) during the same reaction to monitor remaining and cleaved <sup>32</sup>P, respectively (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>B–<xref ref-type="fig" rid="pbio-0020091-g001">1</xref>D; quantified in <xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>E). ATP is hydrolyzed by the Arp2/3 complex at the earliest timepoints after the addition of VCA (monitored by <sup>32</sup>P cleavage) and cleavage has ceased by 90 s (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>D). SDS-PAGE analysis separates the subunits and shows that the γ-<sup>32</sup>P is cleaved rapidly from Arp2 upon addition of VCA, but not significantly from Arp3 (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>C). The kinetics of ATP hydrolysis assayed by SDS-PAGE match the kinetics of phosphate cleavage by TLC (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>E). Since the nucleation reaction is autocatalytic, the rate increases over time, and therefore it is not possible to derive an exact ATPase rate constant from our data. However, we can define a conservative lower bound: k<sub>hyd</sub> > 0.1 s<sup>–1</sup>, noting that the true rate constant is probably much higher. Isolated Arp2/3 complex in polymerization buffer shows very slow spontaneous cleavage of γ-<sup>32</sup>P from both Arp2 and Arp3 (<1 × 10<sup>–4</sup> s<sup>–1</sup>) (unpublished data). As a control, <sup>32</sup>P-ATP hydrolysis is only seen when the Azido-ATP is covalently crosslinked to the Arp2/3 complex (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>D, compare open and closed circles) indicating that the signal is due only to hydrolysis of ATP covalently bound to the Arp2/3 complex and not due to ATP hydrolysis by polymerizing actin. This is further confirmed by observations of ATP hydrolysis on the Arp2/3 complex under conditions where no actin polymerization takes place (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>E and <xref ref-type="fig" rid="pbio-0020091-g003">3</xref>F; <xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>).</p><fig id="pbio-0020091-g003" position="float"><label>Figure 3</label><caption><title>A Single Actin Monomer, in the Presence of Actin Filaments and VCA, Stimulates ATP Hydrolysis on Arp2, without Requiring Actin Polymerization</title><p>(A–C) Remaining unhydrolyzed γ-<sup>32</sup>P-AzidoATP on Arp2 (closed circle) and Arp3 (open circle) was quantified to assay ATP hydrolysis (same conditions as <xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>B–<xref ref-type="fig" rid="pbio-0020091-g001">1</xref>D). γ-<sup>32</sup>P-AzidoATP-labeled Arp2/3 (20 nM) was mixed at indicated times with either 750 nM VCA then 2 μM G-actin (A), 2 μM G-actin then 750 nM VCA (B), or 2 μM F-actin then 750 nM VCA (C).</p><p>(D) Latrunculin B (open square) inhibits the ability of VCA plus monomeric actin (open circle) to stimulate ATP hydrolysis on the Arp2/3 complex in the absence of actin filaments. Also, <sup>32</sup>P ATP hydrolysis signal requires covalent crosslinking to Arp2/3. Arp2/3 was mixed with 6 μM γ-<sup>32</sup>P-AzidoATP and exposed to UV either before (closed circle) or after (open circle) the addition of excess (2 mM) unlabeled ATP. Excess ATP added before the UV exposure prevents crosslinking and abolishes the ATP hydrolysis signal, indicating that all the <sup>32</sup>P ATP hydrolysis signals measured are due to ATP hydrolysis on Arp2/3 and not from ATP hydrolysis on actin.</p><p>(E and F) In the presence of phalloidin-stabilized actin filaments, actin monomers are prevented from polymerizing by Latrunculin B, but still stimulate ATP hydrolysis on the Arp2/3 complex. 20 nM γ-<sup>32</sup>P-AzidoATP–labeled Arp2/3 was premixed with 1 μM phalloidin-stabilized actin filaments. The reaction was initiated by mixing with 750 nM N-WASP VCA, 1 μM G-actin and 4 μM Latrunculin B as indicated, cleaved γ-<sup>32</sup>P was assayed by phosphomolybdate extraction (E), and separately, actin polymerization was monitored by pyrene–actin fluorescence (F).</p></caption><graphic xlink:href="pbio.0020091.g003"/></fig><fig id="pbio-0020091-g004" position="float"><label>Figure 4</label><caption><title>Pointed-End Filament Capping Is Sufficient to Stimulate ATP Hydrolysis on Arp2 in the Absence of VCA</title><p>(A) The Arp2/3 complex prevents actin filament reannealing by capping the pointed ends. The length distribution of 2 μM Alexa-488 phalloidin-stabilized actin filaments is unaffected in the absence (i) or presence (iii) of 20 nM Arp2/3 complex. (ii) 5 min after shearing the filaments, filaments have begun to reanneal in the absence of the Arp2/3 complex, but 20 nM Arp2/3 complex (iv) maintains short filaments, preventing reannealing by capping filament pointed ends.</p><p>(B) ATP hydrolysis on Arp2 is stimulated by pointed-end capping. Crosslinked γ-<sup>32</sup>P-AzidoATP-Arp2/3 (20 nM) was mixed with 2 μM phalloidin-stabilized actin filaments. The mixture was split in two and one sample was sheared. Timepoints were taken as shown.</p><p>(C) Uncleaved <sup>32</sup>P on Arp2 (unsheared [closed circle] and sheared [closed square]) and Arp3 (unsheared [open circle] and sheared [open square]) were quantified from (B). Arp2 rapidly hydrolyzes bound ATP upon filament pointed-end capping.</p></caption><graphic xlink:href="pbio.0020091.g004"/></fig></sec><sec id="s2c"><title>Phosphate Release by Arp2 Lags Hydrolysis by Approximately 40 s</title><p>To investigate the kinetics of phosphate release fromArp2/3 during the polymerization reaction, we added maltose and maltose phosphorylase to the reaction. In the presence of <sup>32</sup>P-labeled Arp2/3 complex, maltose phosphorylase conjugates the <sup>32</sup>P-orthophosphate released from Arp2 to a hydrolyzed maltose molecule to make <sup>32</sup>P-glucose phosphate. The phosphate from adenosine diphosphate-inorganic phosphate (ADP-Pi)-bound Arp2 is inaccessible to the enzyme and remains unconjugated orthophosphate. We quantified hydrolyzed <sup>32</sup>P-ATP and released phosphate by TLC (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>F). Phosphate release from Arp2 lags behind ATP hydrolysis by approximately 40 s.</p></sec><sec id="s2d"><title>The Rate of Filament Nucleation Matches the Rate of ATP Hydrolysis by Arp2</title><p>To determine whether ATP hydrolysis on Arp2 is coupled to filament nucleation, we varied the rate of nucleation and looked to see whether the rate of ATP hydrolysis by Arp2 varied accordingly. We varied the nucleation rate by using N-WASP and Scar1 VCA domains, which stimulate different rates of Arp2/3 complex-dependent actin nucleation (<xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>). To slow the nucleation reaction and allow more accurate kinetic measurements, we used only 1 μM monomeric actin. We used pyrene–actin polymerization data (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>A) to calculate the concentration of barbed ends produced over time (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>B, open symbols) (see <xref ref-type="sec" rid="s4">Methods and Materials</xref>; <xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>). Note that this calculation is model-independent and simply uses the established kinetic parameters for actin polymerization and the change in the amount of monomeric and filamentous actin over time measured from the pyrene–actin curves. The same reagents were used to monitor ATP hydrolysis by Arp2 under the same conditions. We used loss of γ-<sup>32</sup>P labeling as a probe for ATP hydrolysis and scaled the initial labeling intensity to the Arp2/3 concentration used in the reaction (20 nM) to calibrate the stoichiometry of ATP hydrolyzed by Arp2 (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>B). Using Scar1 VCA instead of N-WASP VCA halves both the rate of nucleation of actin filaments and the rate of ATP hydrolysis on Arp2.</p><fig id="pbio-0020091-g002" position="float"><label>Figure 2</label><caption><title>ATP Hydrolysis by Arp2 Coincides with Nucleation of New Actin Filaments, and Not Filament Debranching</title><p>(A, B) The kinetics of nucleation were slowed by using only 1 μM monomeric actin (compared to 2 μM for <xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>). γ-<sup>32</sup>P-AzidoATP-Arp2/3 (20 nM) was mixed with either 750 nM N-WASP WWA (closed circle) or Scar1 WA (closed square) and 1 μM 7% pyrene-labeled monomeric actin.</p><p>(A) Actin polymerization measured by pyrene fluorescence.</p><p>(B) The concentration of new filament ends (open symbols) was calculated from the polymerization data in a model-independent way (see <xref ref-type="sec" rid="s4">Methods and Materials</xref>), and Arp2-ATP hydrolysis (closed symbols) was measured under the same reaction conditions for both N-WASP WWA (open and closed circles) and Scar1 WA (open and closed squares).</p><p>(C) ATP hydrolysis on Arp2 does not accompany filament debranching. Using a large excess (100 nM) of γ-<sup>32</sup>P-AzidoATP-Arp2/3 creates a slow hydrolysis phase that follows the rapid nucleation phase. The slow phase of ATP hydrolysis can be inhibited by excess (1.5 μM) uncrosslinked Arp2/3 added at t = 200 s, showing that the slow phase of ATP hydrolysis is from Arp2/3 being recruited from solution and not from that already incorporated in branches.</p></caption><graphic xlink:href="pbio.0020091.g002"/></fig><p>We note that the total amount of Arp2 that hydrolyzes ATP in the polymerization reaction is 30% less for Scar1 VCA than for N-WASP VCA, which we interpret as 30% fewer filaments produced. Although it is possible to calculate the <italic>rate</italic> of end production from the pyrene–actin polymerization curve in a model-independent way, it is not possible to calculate the <italic>total</italic> number of barbed ends produced, since once polymerization reaches equilibrium, the pyrene–actin curve will not change even if new barbed ends continue to be produced. From the ATP hydrolysis data, therefore, the Arp2/3 complex produces filament ends more slowly when activated by Scar1, and under our conditions, the reaction ends when monomeric actin is depleted by incorporation into the new filaments. Fewer total filaments are therefore produced by the less active VCA domain.</p></sec><sec id="s2e"><title>ATP Hydrolysis on Arp2 Does Not Accompany Filament Debranching</title><p>A previous study claimed that ATP hydrolysis on Arp2 occurs very slowly (t<sub>1/2</sub> ≈ 800 s), coincident with filament debranching (<xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. 2003</xref>). <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>) used a much higher concentration of Arp2/3 complex (100 nM) in their assays than the 5 nM Arp2/3 complex that they estimate was used up during their polymerization reaction. Using these conditions, we find that Arp2/3 complex hydrolyses ATP in two discrete phases: a fast (nucleation) phase, followed by a slow, approximately linear phase (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>C, open symbols). This slow phase does not plateau within 6000 s and is similar to the data presented in <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>). To demonstrate that this slow ATP hydrolysis is not due to the Arp2/3 complex hydrolyzing ATP upon debranching, we added an excess of unlabeled Arp2/3 complex into solution at t = 200 s, after the polymerization phase is complete. This unlabeled Arp2/3 complex competes for nucleating factors with γ-<sup>32</sup>P-AzidoATP-Arp2/3 in solution, but it does not compete with γ-<sup>32</sup>P-AzidoATP-Arp2/3 already incorporated in branches. Addition of excess unlabeled Arp2/3 complex at t = 200 s inhibits the slow phase of ATP hydrolysis (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>C, closed symbols), indicating that the slow phase is due to ATP hydrolysis on Arp2/3 complex being recruited from solution and not due to ATP hydrolysis on Arp2/3 complex already in branches. This slow ATP hydrolysis probably represents a low rate of filament nucleation by the excess unused Arp2/3 complex, the rate of nucleation being limited by the low monomeric actin concentration that remains after most of the actin has polymerized.</p></sec><sec id="s2f"><title>Both VCA and Monomeric Actin Are Required to Stimulate ATP Hydrolysis by Arp2 during the Polymerization Reaction</title><p>Although the kinetics of ATP hydrolysis on Arp2 match the kinetics of actin polymerization, these data do not rule out the possibilities that VCA alone or the filamentous actin created during the polymerization reaction stimulates the ATPase activity independent of nucleation. To more specifically determine what stimulates ATP hydrolysis on Arp2, we varied the order of addition of the components that initiate the polymerization reaction. Incubation of the Arp2/3 complex with VCA does not induce ATP hydrolysis by the complex until monomeric actin is added to the reaction (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>A), showing that VCA alone does not stimulate the ATPase activity. Similarly, monomeric actin alone does not stimulate the Arp2/3 complex to hydrolyze ATP until the addition of VCA (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>B). To test whether actin filaments themselves stimulate Arp2/3 ATP hydrolysis, we used phalloidin-stabilized actin filaments to ensure that no monomeric actin would be present and took care not to shear the filaments in order to reduce the number of free pointed ends. ATP hydrolysis is not stimulated on the Arp2/3 complex by filamentous actin, even in presence of VCA (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>C). As controls, we found that neither 5 μM phalloidin nor 20 mM phosphate inhibit the kinetics of ATP hydrolysis by Arp2 during the polymerization reaction (unpublished data).</p><p>When Arp2/3 concentration is low (20 nM), and nucleation is rapid (using N-WASP VCA), initiation of the polymerization reaction causes striking and near-complete ATP hydrolysis on Arp2 (approximately 80%, i.e., approximately 16 nM; <xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>B and <xref ref-type="fig" rid="pbio-0020091-g003">3</xref>C). We detect a small amount of ATP hydrolysis on Arp3 with similar kinetics but much lower stoichiometry (10%–20%). The decrease is not caused by the dilution effect of adding the second component (approximately 4%), which is already compensated for in the data presented.</p></sec><sec id="s2g"><title>In the Presence of Both VCA and Actin Filaments, a Nonpolymerizable Actin Monomer Is Sufficient to Trigger Rapid ATP Hydrolysis on Arp2</title><p>The timing and stoichiometry of ATP hydrolysis and the combination of factors required to stimulate it suggest that Arp2 hydrolyzes ATP during the filament nucleation reaction. Kinetic and light-microscopy data indicate that most or all Arp2/3-dependent filament nucleation occurs from Arp2/3 complex bound to the sides of filaments produced earlier in the polymerization reaction (<xref rid="pbio-0020091-Blanchoin3" ref-type="bibr">Blanchoin et al. 2000a</xref>, <xref rid="pbio-0020091-Blanchoin5" ref-type="bibr">2001</xref>; <xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>). To test whether filament side-binding is necessary for ATP hydrolysis on Arp2, we blocked filament formation with the actin-monomer binding toxin, Latrunculin B. Latrunculin B binds to monomeric actin and prevents it polymerizing, but does not affect its binding to VCA (R. D. Mullins and A. E. Kelly, unpublished data). The combination of VCA and Latrunculin B–actin monomers does not stimulate ATP hydrolysis on Arp2/3 complex (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>D, open squares), nor do preformed, phalloidin-stabilized actin filaments and Latrunculin B–actin monomers without VCA (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>E, filled squares). In the presence of preformed actin filaments and VCA, however, Latrunculin B–actin monomers stimulate rapid ATP hydrolysis on Arp2/3 (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>E, filled circles) without actin polymerization (<xref ref-type="fig" rid="pbio-0020091-g003">Figure 3</xref>F). <xref ref-type="table" rid="pbio-0020091-t001">Table 1</xref> summarizes the requirements for stimulation of ATP hydrolysis on Arp2. These data indicate that during the nucleation reaction, actin filament side-binding by Arp2/3 complex is a prerequisite for VCA and monomeric actin to stimulate ATP hydrolysis on Arp2. The observation that polymerization of the daughter filament is unnecessary implies that the VCA-mediated interaction of a single actin monomer with the Arp2/3 complex is the trigger for ATP hydrolysis on Arp2.</p><table-wrap id="pbio-0020091-t001" position="float"><label>Table 1</label><caption><title>Requirements to Stimulate ATP Hydrolysis on the Arp2 Subunit of Arp2/3 Complex</title></caption><graphic xlink:href="pbio.0020091.t001"/><table-wrap-foot><fn id="nt101"><p>Abbreviations: G-actin, monomeric actin; F-actin, actin filaments; LatB, Latrunculin B</p></fn></table-wrap-foot></table-wrap></sec><sec id="s2h"><title>Pointed-End Capping by the Arp2/3 Complex Stimulates Rapid ATP Hydrolysis by Arp2 in the Absence of Either Branch Formation or a WASP-Family VCA Domain</title><p>The Arp2/3 complex is known to cap the pointed ends of preformed actin filaments in vitro, inhibiting both polymerization and depolymerization from the pointed ends of gelsolin-capped filaments (<xref rid="pbio-0020091-Mullins1" ref-type="bibr">Mullins et al. 1998</xref>). The Arp2/3 complex does not cap the barbed ends of actin filaments and does not affect the rate of addition of monomers from the barbed ends of spectrin-capped filaments (unpublished data). We speculated that the way the Arp2/3 complex caps a free-filament pointed end in solution might mimic the way the Arp2/3 complex anchors the pointed end of the new daughter filament in a branch. If the actin monomer that triggers ATP hydrolysis during nucleation is the first monomer of the daughter filament, pointed-end capping, like nucleation, should drive interaction with this monomer and trigger ATP hydrolysis on Arp2. To test this, we sheared preformed, phalloidin-stabilized actin filaments in the presence of the Arp2/3 complex. Mechanical shearing fragments long actin filaments into many short filaments, creating many new filament ends that rapidly reanneal to produce long filaments again (<xref rid="pbio-0020091-Murphy1" ref-type="bibr">Murphy et al. 1988</xref>). This reannealing process is blocked by proteins that cap filament ends (<xref rid="pbio-0020091-Andrianantoandro1" ref-type="bibr">Andrianantoandro et al. 2001</xref>). Without shearing, the addition of 20 nM Arp2/3 complex does not alter the length distribution of phalloidin-stabilized actin filaments (<xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>A, compare [i] and [iii]). After shearing in the presence of 20 nM Arp2/3 complex, pointed-end capping by the Arp2/3 complex blocks reannealing and results in significantly shorter filaments (<xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>A, compare [ii] and [iv]). No branches form within this time—it takes several hours for even a few branches to assemble under these conditions (unpublished data). To assay for ATP hydrolysis by the complex, we incubated γ-<sup>32</sup>P-AzidoATP-Arp2/3 complex with actin filaments under the same conditions as the microscopy experiment. We split the mixture into two parts, sheared one half, and took timepoints to assay for ATP hydrolysis from both samples (<xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>B; quantified in <xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>C). No ATP hydrolysis occurs in the unsheared condition, confirming that binding to the sides of actin filaments is not sufficient to stimulate ATP hydrolysis. ATP hydrolysis occurs rapidly in the sheared condition and occurs only on Arp2 (<xref ref-type="fig" rid="pbio-0020091-g004">Figure 4</xref>C). Since this occurs well before any branches form, pointed-end capping by the Arp2/3 complex is sufficient to stimulate ATP hydrolysis on Arp2 not only in the absence VCA, but also in the absence of filament side-binding.</p></sec></sec><sec id="s3"><title>Discussion</title><p>Conventional actin and all actin-related proteins share a conserved nucleotide binding pocket. Actin monomers bind ATP but do not hydrolyze it until they are induced to polymerize. Actin polymerization triggers rapid ATP hydrolysis, followed by a slow release of cleaved phosphate from the filament (<xref rid="pbio-0020091-Blanchoin2" ref-type="bibr">Blanchoin and Pollard 2002</xref>). Arp2 also hydrolyzes its bound ATP, and we find that the conditions that promote ATP hydrolysis and the kinetics of the reaction are remarkably similar to those of conventional actin. In the presence of VCA and actin filaments, monomeric actin stimulates ATP hydrolysis on Arp2 (<xref ref-type="table" rid="pbio-0020091-t001">Table 1</xref>). We also find that binding of the Arp2/3 complex to the pointed end of a preformed actin filament is sufficient to trigger Arp2 ATP hydrolysis, even in the absence of VCA. The stimulation of Arp2 ATPase activity by both filament pointed ends and by actin monomers under nucleating conditions suggests that the geometry of the Arp2/3–actin interaction is the same in both cases.</p><p>Interaction between the Arp2/3 complex and conventional actin can occur in three distinct ways: (1) the Arp2/3 complex binds the sides of preformed actin filaments; (2) the Arp2/3 complex binds to the pointed ends of filaments, either by remaining associated with the daughter filament following nucleation or by capping preformed pointed ends; and (3) the Arp2/3 complex may interact with an actin monomer bound to the VCA domain of a WASP-family protein. There is abundant experimental evidence for filament side- and pointed-end binding by the complex (<xref rid="pbio-0020091-Mullins1" ref-type="bibr">Mullins et al. 1998</xref>; <xref rid="pbio-0020091-Blanchoin3" ref-type="bibr">Blanchoin et al. 2000a</xref>, <xref rid="pbio-0020091-Blanchoin5" ref-type="bibr">2001</xref>; <xref rid="pbio-0020091-Amann1" ref-type="bibr">Amann and Pollard 2001a</xref>, <xref rid="pbio-0020091-Amann1" ref-type="bibr">2001b</xref>). Evidence that a VCA-bound actin monomer interacts with the Arp2/3 complex is more circumstantial and is supported by four observations: (1) VCA domains can simultaneously bind both the Arp2/3 complex and monomeric actin (<xref rid="pbio-0020091-Marchand1" ref-type="bibr">Marchand et al. 2001</xref>; <xref rid="pbio-0020091-Panchal1" ref-type="bibr">Panchal et al. 2003</xref>); (2) removal of the actin monomer-binding WH2 (V) domain from a WASP-family protein severely decreases the efficiency of Arp2/3 activation (<xref rid="pbio-0020091-Marchand1" ref-type="bibr">Marchand et al. 2001</xref>); (3) kinetic modeling suggests that the Arp2/3 complex requires monomeric actin to form a filament nucleus (<xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>); and (4) Arp2/3-dependent nucleation is not limited to the end of the mother filament (<xref rid="pbio-0020091-Amann1" ref-type="bibr">Amann and Pollard 2001a</xref>), indicating that the VCA-bound actin monomer does not incorporate into the mother filament. Two of the three interactions between the Arp2/3 complex and conventional actin—nucleation and pointed-end capping—are thought to be mediated by the actin-related subunits, analogous to actin–actin interactions in a filament. Both interactions stimulate rapid ATP hydrolysis by Arp2.</p><p>Based on sequence conservation and biochemical similarities, ATP hydrolysis on Arp2 is probably driven by a mechanism similar to that which stimulates ATP hydrolysis on actin. The molecular details of how polymerization activates ATP hydrolysis on conventional actin, however, are not well understood. A leading hypothesis is that a “hydrophobic plug”—a loop between subdomains 3 and 4 of actin (residues 262–274 in yeast; <xref rid="pbio-0020091-Kuang1" ref-type="bibr">Kuang and Rubenstein 1997</xref>)—undocks from the monomer surface and binds to a hydrophobic cleft formed by adjacent monomers in the opposite strand of the two-start filament helix (<xref rid="pbio-0020091-Lorenz1" ref-type="bibr">Lorenz et al. 1993</xref>; <xref rid="pbio-0020091-Kuang1" ref-type="bibr">Kuang and Rubenstein 1997</xref>). Our data are consistent with stimulation of ATP hydrolysis by docking of a hydrophobic plug sequence on Arp2 into a hydrophobic cleft created by Arp3 and the first actin monomer of the daughter filament (<xref ref-type="fig" rid="pbio-0020091-g005">Figure 5</xref>). In the crystal structure of the inactive Arp2/3 complex, Arp2 and Arp3 are oriented like a pair of actin monomers in opposite strands of the two-start filament helix (<xref rid="pbio-0020091-Robinson1" ref-type="bibr">Robinson et al. 2001</xref>), but they are separated by a 40 Å cleft. Our data support a model in which activation of the complex involves closure of the cleft, allowing actin to polymerize from an Arp2–Arp3 heterodimer (<xref rid="pbio-0020091-Kelleher1" ref-type="bibr">Kelleher et al. 1995</xref>; <xref rid="pbio-0020091-Robinson1" ref-type="bibr">Robinson et al. 2001</xref>), which then remains attached to the pointed end of the new daughter filament, anchoring it to the branch (<xref ref-type="fig" rid="pbio-0020091-g005">Figure 5</xref>B [iv]). Based on the geometry of the subunits in the crystal structure and the hydrophobic plug model, we expect that the Arp3–actin contact creates a pocket to bind the hydrophobic plug of Arp2 (residues 265–277 in yeast Arp2). The geometry of the interaction would stimulate the ATPase activity of Arp2, but not Arp3 (<xref ref-type="fig" rid="pbio-0020091-g005">Figure 5</xref>A).</p><fig id="pbio-0020091-g005" position="float"><label>Figure 5</label><caption><title>Model for Activation of ATP Hydrolysis on the Arp2/3 Complex and Mechanism by which WASP-Family Proteins Activate the Arp2/3 Complex to Nucleate New Actin Filaments</title><p>(A) Filament pointed-end capping stimulates ATP hydrolysis on Arp2 without branch formation. (i) Arp2 and Arp3 are separated when the Arp2/3 complex is free in solution. (ii) Upon pointed-end capping, the binding energy of the actin-Arp2/3 interface drives Arp2 and Arp3 together and (iii) a conformational change on Arp2 (shown by the red the subdomain 3/4 loop flipping out) triggers ATP hydrolysis by Arp2 (filament pointed-end capping is probably not a significant function of the Arp2/3 complex in vivo).</p><p>(b) A VCA-bound actin monomer drives the activation of the Arp2/3 complex and stimulates ATP hydrolysis on Arp2. (i) The Arp2/3 complex must first be bound to the side of an actin filament, and an actin monomer is bound to the VC domain of the WASP-family protein. (ii) The VC domain of the WASP-family protein docks the first monomer of the daughter filament onto the Arp2/3 complex, stabilizing the Arp2–Arp3–actin interaction and promoting the active conformation of the complex. (cf. Aii). (iii) The active conformation of the Arp2–Arp3–actin monomer triggers a conformational change on Arp2 and ATP hydrolysis by the subunit. (iv) Actin polymerizes from the activated Arp2/3 complex. ATP hydrolysis by Arp2 may promote dissociation of the CA domain of the WASP-family protein from the Arp2/3 complex, aided by actin polymerization, which competes its WH2 domain from the first actin monomer.</p></caption><graphic xlink:href="pbio.0020091.g005"/></fig><p>Monomeric actin does not interact directly with the Arp2/3 complex in the absence of VCA, but under conditions that promote nucleation, a single actin monomer triggers VCA-dependent ATP hydrolysis on Arp2. By analogy with capping-induced ATP hydrolysis, the monomer that triggers ATPase activity is therefore the first monomer of the new daughter filament (<xref ref-type="fig" rid="pbio-0020091-g005">Figure 5</xref>B [i]–[iii]). The hydrophobic pocket formed between Arp2, Arp3, and the actin monomer would therefore promote a similar conformational change in Arp2 and stimulate ATP hydrolysis (<xref ref-type="fig" rid="pbio-0020091-g005">Figure 5</xref>B [iv]).</p><p>Interaction of the Arp2/3 complex with the sides of filaments is not sufficient to trigger Arp2 ATPase activity, even in the presence of VCA. Binding of Arp2/3 to the sides of filaments is, however, required for ATP hydrolysis on Arp2 stimulated by VCA and monomeric actin. These data suggest that binding the side of an actin filament induces a conformational change in the Arp2/3 complex that enables it to interact with the actin monomer bound to VCA. The filament side-binding activity of Arp2/3 does not require the presence of the Arp2 or Arp3 subunits and can be reconstituted by a combination of the Arc2 (p34) and Arc4 (p20) subunits (<xref rid="pbio-0020091-Gournier1" ref-type="bibr">Gournier et al. 2001</xref>). The Arc2 and Arc4 subunits contact both Arp2 and Arp3, and therefore filament side-binding might favor association of Arp2 and Arp3. The fact that Arp2-ATP hydrolysis induced by VCA and an actin monomer requires filament side-binding strongly suggests that all Arp2/3-generated actin filaments are born on the side of preformed filaments.</p><p>Our results disagree with a recent paper that claims that ATP hydrolysis on Arp2 is slow and accompanies filament debranching (<xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. 2003</xref>). Using experimental conditions similar to the previous study, we observe similar slow ATP hydrolysis kinetics (<xref ref-type="fig" rid="pbio-0020091-g002">Figure 2</xref>C) and show that this ATP hydrolysis occurs on Arp2/3 complex recruited slowly from solution. The slow hydrolysis does not reflect delayed ATP hydrolysis on Arp2/3 complex that had been rapidly incorporated into branches early in the experiment. ATP hydrolysis on Arp2, therefore, cannot be associated with debranching. <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>) claim that ATP hydrolysis does not occur during nucleation and present data with a lag of several hundred seconds between computer-simulated nucleation kinetics and measured ATP hydrolysis kinetics (<xref ref-type="fig" rid="pbio-0020091-g001">Figure 1</xref>B in <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. 2003</xref>). In this experiment, <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>) initiate polymerization in the absence of free ATP. These conditions would deactivate up to 97% of the Arp2/3 complex (the fraction that is not crosslinked to ATP on both subunits). In our experience, removal of free ATP introduces an artificial lag in polymerization that lasts until tightly bound ATP is released from monomeric actin (1/k<sub>ATP release</sub> = 330 s; <xref rid="pbio-0020091-Selden1" ref-type="bibr">Selden et al. 1999</xref>) and is free to interact with the Arp2/3 complex (unpublished data). The claim by <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>) that the absence of free ATP does not affect ATP hydrolysis kinetics is contradicted by their observation that the <sup>32</sup>P signal is unchanged by the addition of free ATP. The <sup>32</sup>P signal is only equivalent to hydrolyzed ATP in the absence of free ATP. The addition of free ATP should cause the excess of uncrosslinked Arp2/3 complex to compete with the small fraction of crosslinked <sup>32</sup>P-ATP-Arp2/3 complex and thereby significantly reduce the <sup>32</sup>P signal. The observation that the <sup>32</sup>P signal is not reduced, rather than confirming that removal of free ATP has no effect, instead confirms that contaminating ATP is present for the latter part of the “ATP-free” condition, presumably released slowly from monomeric actin. The lag in the polymerization created by the initial absence of ATP would be present in the experimental ATP hydrolysis measurement, but may not have been present in the nucleation data presented because this was generated by a model-dependent computer simulation (<xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. 2003</xref>).</p><p>We find that ATP hydrolysis and phosphate release from Arp2 (approximately 40 s) are more than an order of magnitude faster than debranching of Arp2/3-generated dendritic networks (approximately 1000 s) (<xref rid="pbio-0020091-Blanchoin4" ref-type="bibr">Blanchoin et al. 2000b</xref>). The kinetics of phosphate release from Arp2 are also about an order of magnitude faster than phosphate release from actin (1/k<sub>Pi release</sub> = 384 s for skeletal muscle actin; <xref rid="pbio-0020091-Melki1" ref-type="bibr">Melki et al. 1996</xref>), suggesting that, if phosphate release controls debranching, it is the phosphate release from the daughter actin filament that is important, not the phosphate release from Arp2. This is supported by the observation that phalloidin, which slows phosphate release from actin, slows filament debranching, and cofilin, which accelerates phosphate release from actin, accelerates filament debranching (<xref rid="pbio-0020091-Blanchoin4" ref-type="bibr">Blanchoin et al. 2000b</xref>). <xref rid="pbio-0020091-Le1" ref-type="bibr">Le Clainche et al. (2003</xref>) show that chromium-ATP Arp2/3 debranches more slowly than magnesium-ATP Arp2/3 and claim (but do not demonstrate) that chromium-ATP Arp2/3 releases phosphate more slowly. If chromium does slow the phosphate release from Arp2/3, in light of our data, this suggests that phosphate release from Arp2 may be a prerequisite for filament debranching—but is not a direct cause, since it occurs much too rapidly.</p><p>We previously showed that the Arp2/3 complex requires hydrolyzable ATP for nucleation activity (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>), and the current study adds weight to the hypothesis that ATP hydrolysis has a direct role in nucleation by showing that ATP is hydrolyzed by Arp2 upon nucleation. The separation of the Arps in the crystal structure and the very low nucleation rate of the unactivated complex probably reflect the tendency of Arp2 and Arp3 to remain separated in the absence of all the required nucleation promoting factors. This suggests that there is a large free energy barrier to the formation of an Arp2–Arp3 heterodimer. Our data indicate that there are two ways to overcome this energy barrier, both using the binding energy of actin: one using the combined binding energy of the two actin monomers at the pointed end of an actin filament during pointed-end capping, and the other the combined binding energy of the side of the mother filament, the VCA domain, and a single actin monomer. The surface area of the filament pointed end that would be buried by interaction with an Arp2–Arp3 dimer would be large (approximately 6800 Å<sup>2</sup>). This is consistent with the fact that in vitro the binding energy of this interface is sufficient to drive the interaction and promote the active conformation of the complex directly, even in the absence of VCA or a mother filament (<xref rid="pbio-0020091-Mullins1" ref-type="bibr">Mullins et al. 1998</xref>). The binding of monomeric actin alone is insufficient to overcome the free-energy barrier, which ensures that the inactive conformation of the Arp2/3 complex is robust despite high cellular concentrations of actin. Because of the free energy of all the binding partners involved in nucleation, however, the energy of ATP hydrolysis may not be needed to stabilize the nucleus. Regardless, it is very likely that ATP hydrolysis on Arp2, like actin, provides a timing signal to the system. ATP hydrolysis on Arp2/3 would promote release of VCA from the complex and allow a new actin branch to move away from the site of its creation (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>). ATP hydrolysis may also regulate the timing of the interaction of the Arp2/3 complex with other binding partners such as cortactin and cofilin. Temporal regulation of these interactions is likely to be essential to construction of functional motile structures.</p><p>The Arp2/3 ATP hydrolysis assay presented here provides a novel assay for activation of the Arp2/3 complex that does not rely, as all previous assays have done, solely on actin polymerization. Pyrene–actin polymerization is only useful over a limited range of actin concentrations because at high concentrations, spontaneous assembly obscures Arp2/3-mediated nucleation. The pyrene–actin assay also has temporal limits since it rapidly uses up one of the factors necessary for Arp2/3 activation–monomeric actin. Our observation that ATP is hydrolyzed by Arp2 rapidly during, or soon after, the nucleation reaction means that we can use ATP hydrolysis on Arp2 as an assay to study the factors required to promote activation of the Arp2/3 complex. The fact that nonpolymerizable actin monomers are competent to stimulate hydrolysis enables us to investigate the conditions for Arp2/3 complex activation under a wider range of conditions. This system will be useful for further studies of the biophysics of Arp2/3-mediated actin assembly.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><sec id="s4a1"><title>Purification of proteins</title><p>We purified Arp2/3 from <named-content content-type="genus-species">Acanthamoeba castellini</named-content> by a combination of conventional and affinity chromatography (<xref rid="pbio-0020091-Dayel1" ref-type="bibr">Dayel et al. 2001</xref>). We flash-froze Arp2/3 complex in aliquots of approximately 40 μM in 10% glycerol, 0.5 μM TCEP, and 2 mM Tris (pH 8.0), and stored them at –80°C for later use. We purified actin from <italic>Acanthamoeba</italic> by the method of <xref rid="pbio-0020091-MacLean-Fletcher1" ref-type="bibr">MacLean-Fletcher and Pollard (1980</xref>). Actin was stored in fresh G-buffer (0.5 μM TCEP, 0.1 μM CaCl<sub>2</sub>, 0.2 μM ATP, 2 mM Tris [pH 8.0]) and gel-filtered before use. Rat N-WASP VCA (398–502) and Human Scar1-VCA (489–559) with N-terminal 6His tags and TEV cleavage sites were bacterially expressed and purified by nickel affinity chromatography.</p><p>We prepared phalloidin-stabilized actin filaments by adding 1/10 volume of 10× KMEI to monomeric actin at room temperature for 20 min to initiate polymerization, then added twice the concentration of phalloidin and incubated for a further hour at room temperature (1× KMEI buffer: 50 mM KCl, 1 mM MgCl<sub>2</sub>, 1 mM EGTA, 10 mM Imidazole [pH 7.0]). We took care not to unintentionally shear the phalloidin-stabilized actin filaments by using wide-bore pipette tips.</p></sec><sec id="s4a2"><title>Arp2/3 ATPase assay</title><p>We diluted freshly thawed aliquots of Arp2/3 to 2.0 μM in 1 mM MgCl<sub>2</sub>, 50 mM KCl, 10 mM Imidazole (pH 7.0) and added 6 μM γ-<sup>32</sup>P-labeled 8-AzidoATP (Affinity Labeling Technologies, Lexington, Kentucky, United States). After a 2-min incubation to allow nucleotide exchange, we crosslinked for 9 s using a UV hand lamp (312 nm; Fisher Scientific, Hampton, New Hampshire, United States), added 1 mM ATP and 1 mM DTT to quench the reaction and buffer exchanged into 1× KMEI plus 100 μM ATP, 1 mM DTT using a NAP5 column (Amersham Pharmacia Biotech, Little Chalfont, United Kingdom). We used the Arp2/3 for assays within 10 min of crosslinking. The same actin (including 7% pyrene–actin) was used for both ATP hydrolysis assays and correlative pyrene–fluorescence polymerization assays. We took ATPase time points by mixing 400 μl of the reaction mixture with premixed 400 μl of methanol and 100 μl of chloroform. We ran the precipitated protein on SDS-PAGE gel to separate the subunits and quantified <sup>32</sup>P-labeling using a phosphoimager (Storm 840; Molecular Dynamics, Sunnyvale, California, United States). For phosphate cleavage assays, we quenched timepoints into 1/10 volume 26 M formic acid, spotted on cellulose TLC plates, and separated the components in 0.4 M KH<sub>2</sub>PO<sub>4</sub> (pH 3.4). We separately ran <sup>32</sup>P-ATP and <sup>32</sup>P-ATP treated with apyrase as standards to confirm the separation of <sup>32</sup>P-ATP and cleaved <sup>32</sup>P, respectively (unpublished data). As an alternative method of quantifying cleaved <sup>32</sup>P, phosphomolybdate was extracted as in <xref rid="pbio-0020091-Shacter1" ref-type="bibr">Shacter (1984</xref>) and quantified using a scintillation counter. To distinguish the ADP-Pi state of Arp2 from the ADP state, the kinetics of phosphate release were measured by performing the reaction in the presence of 2 mM maltose and 2 U/ml maltose phosphorylase (Sigma-Aldrich, St. Louis, Missouri, United States), which uses only the released Pi to form glucose phosphate. Glucose phosphate was separated from free ATP, protein-ATP, and Pi using TLC.</p></sec><sec id="s4a3"><title>Actin polymerization assays</title><p>We doped <italic>Acanthamoeba</italic> actin with 7% pyrene–actin to monitor actin polymerization by fluorescence (λ<sub>ex</sub> = 365 nm, λ<sub>em</sub> = 407 nm, 25°C) (<xref rid="pbio-0020091-Mullins2" ref-type="bibr">Mullins and Machesky 2000</xref>). We calculated the number of ends produced over time from [ENDS] = (d[F-actin]/dt)/([free G-actin]*10 μM s<sup>–1</sup>) (cf. <xref rid="pbio-0020091-Zalevsky1" ref-type="bibr">Zalevsky et al. 2001</xref>). Polymerization reactions were performed in G-buffer plus 1/10 volume 10× KMEI. The Ca<sup>2+</sup> cation on monomeric actin was preexchanged with Mg<sup>2+</sup> 30 s before use.</p></sec><sec id="s4a4"><title>Microscopy</title><p>We prepared filamentous actin as above and stabilized filaments with stoichiometric Alexa-488 phalloidin (Molecular Probes, Eugene, Oregon, United States). We mixed 2 μM Alexa-488 phalloidin–F-actin with 20 nM Arp2/3, passed twice through a 30-gauge needle to shear the filaments, and incubated at room temperature. Timepoints were taken by diluting 500-fold and rapidly applying to poly-L-lysine–coated coverslips for visualization. Filament images were quantified for length distribution and branch frequency by a custom MATLAB (MathWorks Inc., Natick, Massachusetts, United States) routine.</p></sec></sec></sec>
“Mosaic” Genes Highlight Forces of Genome Diversity and Adaptation	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>Microbes are arguably the most adaptable organisms on Earth, inhabiting nearly every crevice of nearly every corner of the globe. Some invade the cavities of a wide variety of insects and other invertebrates while others colonize the skin, blood, eyes, and internal organs of animals. Still others thrive in such inhospitable places as the hydrothermal vents of the ocean floor and the Dry Valleys of Antarctica. These “simple” single-celled organisms have evolved unique molecules and strategies over some 3.5 billion years that suit life on the edge. With the sequenced genomes of nearly 140 microbial species in hand, scientists are gaining valuable insights into the nature of this adaptive diversity.<xref ref-type="fig" rid="pbio-0020094-g001"/> </p><fig id="pbio-0020094-g001" position="float"><caption><title>Segmentally variable genes</title></caption><graphic xlink:href="pbio.0020094.g001"/></fig><p>Adapting to such radically different niches, it appears, has produced genes with diverse functions that evolve at very different rates. Genes that code for molecules essential for fundamental cellular functions like maintaining cell metabolism and structure tend to evolve rather slowly, while genes that make proteins charged with mediating cellular responses to internal or external changes often evolve relatively quickly. Pathogenic microbes in particular rely on a flexible genome to keep a step ahead of their hosts' similarly evolving defenses in the never-ending struggle to gain adaptive advantage. This adaptability underlies the increasing antibiotic resistance of diseases like tuberculosis, as selective pressures favor the expansion of resistant bacterial populations.</p><p>Combating such problems requires a molecular understanding of bacterial infections, yet function has been ascribed to only a fraction of the genes found in microbial genomes. One approach to improve functional analyses of genome sequences combines bioinformatics with experimental methods. With such collaborations in mind, Yu Zheng, Richard Roberts, and Simon Kasif have developed a computational approach to help filter out the genetic noise and home in on genomic regions likely to contain clues to gene function. Their method relies on a novel way of classifying genes that flags sequences likely to reward biochemical and genetic efforts to analyze gene function.</p><p>Many comparative genomic studies have focused on looking for sequence “motifs” that correlate with well-characterized protein sequences (that is, the amino acid sequence) and predicting function based on their similarity to the known protein sequences. Zheng, Roberts, and Kasif took a different approach, classifying genes based on their sequence variation. The researchers analyzed 43 fully sequenced microbial genomes and, after determining the degree of conservation or divergence among similar genes in different species, divided the genes into three broad categories: rapidly evolving genes unique to a particular species; highly conserved genes; and “segmentally variable,” or mosaic, genes. Stipulating that the boundaries between the categories are somewhat blurred, Zheng et al. define segmentally variable genes as regions that show a mosaic pattern of one or more rapidly evolving, variable regions interspersed with conserved regions. Based on evidence suggesting that retained variable regions tend to serve a function, the researchers predicted that these mosaic genes, with their highly variable, fast-evolving regions, would shed light on the forces that shape genome diversity and adaptation.</p><p>For most of the microbes analyzed, mosaic genes accounted for about 8–20% of their genomes. Selecting several large families of mosaic genes, Zheng et al. explored the relationship between genes with known function and the structure of their variable regions. Noting an overabundance of particular functional categories in different species—such as signaling proteins that come into either direct or indirect contact with the cell's environment—the researchers speculate that the variable regions may constitute an adaptive layer for the microbe, as they not only “play a key role in mediating interactions with other molecules” but also support a microbe's ability to adapt to its particular niche. Several bacteria species, for example, contain roughly 40% more mosaic sensor genes involved in cell motility, which the authors attribute to the microbes' “expanded ability to detect different chemical signals and find favorable environments.”</p><p>This regional variability appears to reflect the influence of selective pressures that fuel diversity through ongoing interactions with other rapidly evolving molecules in the environment, adding another source of genetic adaptability as cells adjust to new environments and outmaneuver pathogenic threats. While many of the mosaic genes identified encode proteins involved in host-pathogen interactions, defense mechanisms, and intracellular responses to external changes, their function is only broadly understood. While Zheng et al. cannot say to what extent variability affects function—Is extreme variability required for diversity or can modest variation suffice?—they are refining their classification of segmentally variable genes to address such questions. Until then, the authors' “mosaic” approach to understanding gene function promises to improve efforts to annotate the volumes of sequenced genomes on hand, offering biologists a much-needed tool to sift through the mountains of genomic datasets and identify promising targets for further study.</p>
Calcium Dynamics of Cortical Astrocytic Networks In Vivo	<p>Large and long-lasting cytosolic calcium surges in astrocytes have been described in cultured cells and acute slice preparations. The mechanisms that give rise to these calcium events have been extensively studied in vitro. However, their existence and functions in the intact brain are unknown. We have topically applied Fluo-4 AM on the cerebral cortex of anesthetized rats, and imaged cytosolic calcium fluctuation in astrocyte populations of superficial cortical layers in vivo, using two-photon laser scanning microscopy. Spontaneous [Ca<sup>2+</sup>]<sub>i</sub> events in individual astrocytes were similar to those observed in vitro. Coordination of [Ca<sup>2+</sup>]<sub>i</sub> events among astrocytes was indicated by the broad cross-correlograms. Increased neuronal discharge was associated with increased astrocytic [Ca<sup>2+</sup>]<sub>i</sub> activity in individual cells and a robust coordination of [Ca<sup>2+</sup>]<sub>i</sub> signals in neighboring astrocytes. These findings indicate potential neuron–glia communication in the intact brain.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Hirase</surname><given-names>Hajime</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Qian</surname><given-names>Lifen</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Barthó</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Buzsáki</surname><given-names>György</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>Astrocytes are nonneuronal cells of the brain with some known and hypothesized functions (Kettenmann and Ransom 1995; <xref rid="pbio-0020096-Fields1" ref-type="bibr">Fields and Stevens-Graham 2002</xref>). Traditionally, astrocytes have been considered to mediate supportive and protective functions in the central nervous system because of their strategic placement relative to the vasculature, and because they lack fast sodium action potentials. It is only recently that this family of glial cells has been implicated in controlling the dynamics of the neuronal networks in the central nervous system (<xref rid="pbio-0020096-Nedergaard1" ref-type="bibr">Nedergaard 1994</xref>; <xref rid="pbio-0020096-Parpura1" ref-type="bibr">Parpura et al. 1994</xref>; <xref rid="pbio-0020096-Kang1" ref-type="bibr">Kang et al. 1998</xref>; <xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>). Although the membrane potential of unidentified glial cells shows correlated changes with neuronal activity in vivo (<xref rid="pbio-0020096-Amzica2" ref-type="bibr">Amzica and Steriade 2000</xref>; <xref rid="pbio-0020096-Amzica1" ref-type="bibr">Amzica and Massimini 2002</xref>), most of our knowledge on neuron–glia and glia–glia communication comes from studies in vitro.</p><p>In cultured and acutely prepared astrocytes, free calcium concentration ([Ca<sup>2+</sup>]<sub>i</sub>) in the cytosol undergoes large changes spontaneously or in response to various physiological and pharmacological manipulations, such as mechanical stimulation, membrane potential depolarization, and activation of metabotropic glutamate receptors (<xref rid="pbio-0020096-Cornell-Bell1" ref-type="bibr">Cornell-Bell et al. 1990a</xref>; <xref rid="pbio-0020096-Pasti1" ref-type="bibr">Pasti et al. 1997</xref>). These slow events are mediated by release of Ca<sup>2+</sup> from intracellular stores (<xref rid="pbio-0020096-Charles2" ref-type="bibr">Charles et al. 1993</xref>; <xref rid="pbio-0020096-Venance1" ref-type="bibr">Venance et al. 1997</xref>). The [Ca<sup>2+</sup>]<sub>i</sub> surges can be evoked by strong neuronal activity (<xref rid="pbio-0020096-Dani1" ref-type="bibr">Dani et al. 1992</xref>; <xref rid="pbio-0020096-Porter1" ref-type="bibr">Porter and McCarthy 1996</xref>), suggesting a potential homeostatic role of astrocytes in the regulation of extracellularly accumulating neurotransmitters (<xref rid="pbio-0020096-Verkhratsky1" ref-type="bibr">Verkhratsky et al. 1998</xref>). Conversely, spontaneous [Ca<sup>2+</sup>]<sub>i</sub> changes in astrocytes have been shown to influence neuronal excitability (<xref rid="pbio-0020096-Parpura1" ref-type="bibr">Parpura et al. 1994</xref>; <xref rid="pbio-0020096-Kang1" ref-type="bibr">Kang et al. 1998</xref>; <xref rid="pbio-0020096-Pasti2" ref-type="bibr">Pasti et al. 2001</xref>). The mechanism of activity propagation among astrocytes is controversial. In tissue cultures, [Ca<sup>2+</sup>]<sub>i</sub> events can propagate among a network of astrocytes via gap junction or by elevation of adenosine triphosphate level (<xref rid="pbio-0020096-Cornell-Bell2" ref-type="bibr">Cornell-Bell et al. 1990b</xref>; <xref rid="pbio-0020096-Charles3" ref-type="bibr">Charles et al. 1991</xref>; <xref rid="pbio-0020096-Nedergaard1" ref-type="bibr">Nedergaard 1994</xref>; <xref rid="pbio-0020096-Reetz1" ref-type="bibr">Reetz et al. 1997</xref>; <xref rid="pbio-0020096-Newman1" ref-type="bibr">Newman 2001</xref>). In the in vitro slice preparation, coordination of [Ca<sup>2+</sup>]<sub>i</sub> activity appears independent of gap junctions but may require transmitter activation of N-methyl-D-aspartic acid (NMDA) and/or metabotropic glutamate receptors (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>; <xref rid="pbio-0020096-Nett1" ref-type="bibr">Nett et al.</xref>2002; <xref rid="pbio-0020096-Tashiro1" ref-type="bibr">Tashiro et al. 2002</xref>). Moreover, the extent and magnitude of these network effects vary as a function of the preparation used, and can involve correlated [Ca<sup>2+</sup>]<sub>i</sub> changes in no, or only a few, neighboring astrocytes, or the whole population (<xref rid="pbio-0020096-Porter1" ref-type="bibr">Porter and McCarthy 1996</xref>; <xref rid="pbio-0020096-Verkhratsky1" ref-type="bibr">Verkhratsky et al. 1998</xref>). Whether and how the observations in the various in vitro situations apply to the intact brain have yet to be determined.</p><p>We have used two-photon laser scanning microscopy (2-PLSM) to monitor cytosolic Ca<sup>2+</sup> concentration in astrocytes labeled with Fluo-4 acetoxymethyl (AM) ester in juvenile rats in vivo. We find that [Ca<sup>2+</sup>]<sub>i</sub> dynamics in astrocytes is rather quiescent during baseline anesthesia. However, increased population bursting, brought about by attenuating γ-aminobutyric acid (GABA<sub>A</sub>) receptor-mediated neurotransmission, leads to increased magnitude [Ca<sup>2+</sup>]<sub>i</sub> surges, and the [Ca<sup>2+</sup>]<sub>i</sub> changes become more strongly coordinated in neighboring astrocytes.</p></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>Loading of Calcium-Sensitive Dye</title><p>To examine the depth of penetration of the Fluo-4 AM, coronal brain slices (300 μm thick) were acutely prepared after the residual dye was washed off from the craniotomy. A large number of cells below the craniotomy showed fluorescence labeling (<xref ref-type="fig" rid="pbio-0020096-g001">Figure 1</xref>). On the basis of morphological appearance (see also <xref ref-type="supplementary-material" rid="sv001">Videos S1-S4</xref>), most brightly labeled cells were astrocytes, in accordance with recent observations using a pressure application of the indicator (<xref rid="pbio-0020096-Stosiek1" ref-type="bibr">Stosiek et al. 2003</xref>). The large overlap between Fluo-4 AM-loaded cells and astrocytes identified by S100B immunoreactivity provided confidence that most of the loaded cells were astrocytes (<xref ref-type="supplementary-material" rid="sv005">Video S5</xref>). In addition to astrocytes, capillary endothelial cells and pericytes, outlining microvessels, were also observed, albeit less regularly. Some processes of astrocytes contacted local vessels. To quantify the dye penetration, mean bulk fluorescence intensity was plotted for different depths from the pial surface. Most intensive labeling occurred between 50–150 μm below the surface (i.e., layers I/II), but labeled cells could be visualized at greater than 300 μm as well (<xref ref-type="fig" rid="pbio-0020096-g001">Figure 1</xref>C). The decreased fluorescence on the surface is likely due to the diluting effect of the washout procedure in the superficial tissue. Like the histological appearance, in vivo imaging revealed numerous astrocytes (<xref ref-type="fig" rid="pbio-0020096-g001">Figure 1</xref>E). Although the labeling was dense, the somata and several associated processes, including vessel-contacting end feet, of single astrocytes could be clearly revealed (<xref ref-type="fig" rid="pbio-0020096-g002">Figure 2</xref>).</p><fig id="pbio-0020096-g001" position="float"><label>Figure 1</label><caption><title>In Vivo Loading and Imaging of Astrocytes Using Fluo-4 AM</title><p>(A) Acute slice prepared 1 h after dye loading. Scale bar, 200 μm.</p><p>(B) Higher magnification reveals cells with typical astrocyte morphology. Scale bar, 20 μm.</p><p>(C) Average bulk fluorescence as a function of the depth from the pial surface.</p><p>(D) Schematic drawing of the experimental arrangement. Abbreviations: EKG, electrocardiogram. PMT, photomultiplier. LFP, glass micropipe for local field potential and multiple unit recording. The same pipette was used to deliver bicuculline.</p><p>(E) Image taken 50–150 μm below pial surface in vivo. Flattened xyz stack.</p><p>(F) Fluo-4 AM loaded cells (left) were stained for S100B immunoreactivity (right), and the images were merged (center). See <xref ref-type="supplementary-material" rid="sv003">Video S3</xref> for large-scale staining. Scale bar, 20 μm.</p></caption><graphic xlink:href="pbio.0020096.g001"/></fig><fig id="pbio-0020096-g002" position="float"><label>Figure 2</label><caption><title>Time-Lapse Imaging of Astrocytes In Vivo</title><p>Four astrocytes, from which fluorometric Ca<sup>2+</sup> imaging (0.5 Hz) was made, are outlined. A blood vessel, outlined by the astrocyte end feet, runs diagonally across the viewed area. White arrows show the end foot connected to the imaged astrocyte.</p></caption><graphic xlink:href="pbio.0020096.g002"/></fig></sec><sec id="s2b"><title>Spontaneous Calcium Events in Astrocytes</title><p>In our initial experiments, we made a large number of line scans (sampling rate ∼200 Hz) of dye-loaded cells to examine whether some of them were neurons. We never observed short-lasting [Ca<sup>2+</sup>]<sub>i</sub> transients (less than 200 ms; <xref rid="pbio-0020096-Svoboda1" ref-type="bibr">Svoboda et al. 1997</xref>; <xref rid="pbio-0020096-Garaschuk1" ref-type="bibr">Garaschuk et al. 2000</xref>), suggesting that the brightly loaded cells were likely to be non-neuronal (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Stosiek1" ref-type="bibr">Stosiek et al. 2003</xref>). In subsequent experiments (<italic>n</italic> = 8 rats), cells with astrocytic appearance (<italic>n</italic> = 185) were selected for long-term (10–20 min) monitoring. For quantitative studies, three states of [Ca<sup>2+</sup>]<sub>i</sub> activity were distinguished: (a) quiescent state with very slow (less than 0.025 Hz) oscillations of baseline fluorescence level, (b) [Ca<sup>2+</sup>]<sub>i</sub> spikes (greater than or equal to 20% increase in <italic>ΔF/F<sub>0</sub></italic> between 5–50 s), and (c) [Ca<sup>2+</sup>]<sub>i</sub> plateau potentials (greater than or equal to 20% increase in <italic>ΔF/F<sub>0</sub></italic> for greater than 50 s). [Ca<sup>2+</sup>]<sub>i</sub> spikes and [Ca<sup>2+</sup>]<sub>i</sub> plateau potentials were automatically detected. In the control (baseline) condition, 11% of astrocytes had at least one spike event, and 52% had at least one plateau event in 10 min. The mean frequency of [Ca<sup>2+</sup>]<sub>i</sub> spikes among the cells that had at least one [Ca<sup>2+</sup>]<sub>i</sub> spike was 0.121 ± 0.098 per minute (mean width at greater than or equal to 20% <italic>ΔF/F<sub>0</sub></italic>: 25.1 ± 10.31 s) and the mean frequency of [Ca<sup>2+</sup>]<sub>i</sub> plateau was 0.118 ± 0.058 per minute (mean duration: 160.4 ± 114.9 s).</p><p>To investigate whether the baseline values of [Ca<sup>2+</sup>]<sub>i</sub> dynamics were affected by increasing neuronal activity, we induced regularly occurring population bursts by local application of bicuculline (<xref rid="pbio-0020096-Schwartz1" ref-type="bibr">Schwartz and Bonhoeffer 2001</xref>; <italic>n</italic> = 7 rats). Large amplitude (0.69 ± 0.26 mV) synchronous field events (approximately 100 ms) occurred at relatively regular frequency (0.15 ± 0.06 Hz), associated with multiple unit discharges. No significant difference was observed in average heartbeat frequency between the control sessions and bicuculline sessions (4.51 ± 0.54 Hz and 4.36 ± 0.74 Hz, respectively; paired t-test, <italic>p</italic> = 0.13).</p><p>We used two methods to evaluate the effect of neuronal activity on [Ca<sup>2+</sup>]<sub>i</sub> in astrocytes (<italic>n</italic> = 214 cells). First, the incidence of [Ca<sup>2+</sup>]<sub>i</sub> spikes and plateau events was counted in the absence and presence of bicuculline-induced population bursts. Under bicuculline condition significantly more astrocytes had [Ca<sup>2+</sup>]<sub>i</sub> spikes (11% versus 24%; <italic>p</italic> < 0.001; Fisher's exact test), whereas the probability (52% versus 54%) of plateau potentials did not differ significantly. The mean duration of plateau potentials, however, was significantly longer (160.4 ± 114.9 s versus 211.12 ± 152.175 s; t-test, <italic>p</italic> < 0.001) after bicuculline treatment. Among the cells that exhibited at least one spike or plateau event, there was not a significant difference in frequency of the event occurrences (spike: 0.121 ± 0.098/min versus 0.098 ± 0.068/min; t-test, <italic>p</italic> = 0.24; plateau 0.118 ± 0.058/min versus 0.112 ± 0.049/min; t-test, <italic>p</italic> = 0.46). Thus, the major difference between control and bicuculline conditions was the higher proportion of active astrocytes under bicuculline.</p><p>The second method examined [Ca<sup>2+</sup>]<sub>i</sub> changes in the frequency domain. The <italic>ΔF/F<sub>0</sub></italic> trace was considered as a continuous process, and the power spectrum estimate was calculated with a multi-taper method for each astrocyte and averaged across cells. There was a general increase of power at all frequencies in bicuculline-treated animal. The most consistent significant increase (<italic>p</italic> < 0.05) of power appeared in the frequency range of 0.10–0.24Hz, reflecting the increased incidence of [Ca<sup>2+</sup>]<sub>i</sub> spikes. Short-term cross-correlation of neuronal field bursts and [Ca<sup>2+</sup>]<sub>i</sub> signals (± 10 s) did not show a significant time-locked relationship (<xref ref-type="fig" rid="pbio-0020096-g003">Figure 3</xref>).</p><fig id="pbio-0020096-g003" position="float"><label>Figure 3</label><caption><title>Frequency Domain Analysis of Population Dynamics of Fluorescence in Astrocytes in Control State and during Bicuculline-Induced Neuronal Hyperactivity</title><p>Insets show local field potentials in a control animal and regular spiking in a bicuculline treated mouse (scale bar: 2.0 s, 500 μV). Asterisks show significant differences (<italic>p</italic> < 0.05) between groups at various frequencies.</p></caption><graphic xlink:href="pbio.0020096.g003"/></fig></sec><sec id="s2c"><title>Spatio-Temporal Dynamics of [Ca<sup>2+</sup>]<sub>i</sub> Events</title><p>In individual experiments, propagation of synchronous activity could be observed visually (<xref ref-type="fig" rid="pbio-0020096-g004">Figure 4</xref>A; <xref ref-type="supplementary-material" rid="sv006">Video S6</xref>) but the spatio-temporal relationship of [Ca<sup>2+</sup>]<sub>i</sub> dynamics among astrocytes varied across experiments. To quantify the magnitude and spatial extent of this population effect, pair-wise cross-correlograms of <italic>ΔF/F<sub>0</sub></italic> intensity were calculated separately for nearby cell pairs (local: less than or equal to 50 μm) and distant cell pairs (greater than 50 μm). In control conditions, the temporal correlation of [Ca<sup>2+</sup>]<sub>i</sub> signals in neighboring pairs was somewhat larger than in distant pairs, but this difference was not significant (<italic>n</italic> = 374 neighbor pairs and <italic>n</italic> = 1,138 distant pairs). Nevertheless, [Ca<sup>2+</sup>]<sub>i</sub> signals in astrocytes were not completely random, since the cross-correlograms had wide central peaks at the 10–100 s scale (<xref ref-type="fig" rid="pbio-0020096-g004">Figure 4</xref>B). In contrast to the baseline condition, the temporal correlation of [Ca<sup>2+</sup>]<sub>i</sub> changes in local and distant pairs were significantly different after large population bursts were brought about by bicuculline (<xref ref-type="fig" rid="pbio-0020096-g004">Figure 4</xref>C). Correlation of distant pairs under bicuculline (<italic>n</italic> = 433 pairs) was similar to those in the control condition. However, synchrony between local pairs (<italic>n</italic> = 1,282) increased several-fold relative to both distant pairs under the same condition (<italic>t</italic>-test, <italic>p</italic> < 0.0001) and to local pairs in the baseline condition (<italic>t</italic>-test, <italic>p</italic> < 0.0001).</p><fig id="pbio-0020096-g004" position="float"><label>Figure 4</label><caption><title>Spatio-Temporal Dynamics of Astrocyte Ca<sup>2+</sup> Activity</title><p>(A) Definition of nearby (less than 50 μm) and distant (greater than 50 μm) cell pairs.</p><p>(B) Fluorescence changes in two nearby astrocytes.</p><p>(C) Cross-correlogram of fluorescent intensity.</p><p>(D) Mean cross-correlation of <italic>ΔF/F<sub>0</sub></italic> in all nearby (thick line) and distant (thin line) cell pairs in control condition (left) and in the presence of bicuculline (right). Note large increase of <italic>ΔF/F<sub>0</sub></italic> correlation in nearby cell pairs in the bicuculline condition (error bar: standard error of the mean).</p><p>(E) Relationship between distance of the two cells and the magnitude of correlation at zero timelag. Note lack of a reliable relationship in the control condition (left). Note also the significant negative correlation between the distance and correlated <italic>ΔF/F<sub>0</sub></italic> changes in cell pairs in the bicuculline-treated cortex (right).</p></caption><graphic xlink:href="pbio.0020096.g004"/></fig><p>Using a different approach, the magnitude of the zero-timelag correlation coefficient for each cell pair was plotted against distance between the cell pairs. Under control condition, no notable relationship was observed between these variables (<xref ref-type="fig" rid="pbio-0020096-g004">Figure 5</xref>; <italic>n</italic> = 1,512 cell pairs, <italic>r</italic> = 0.019, <italic>p</italic> = 0.46). In contrast, a significant negative correlation was found between the synchrony of [Ca<sup>2+</sup>]<sub>i</sub> signals in the bicuculline condition (<italic>n</italic> = 1,715; <italic>r</italic> = −0.281; <italic>p</italic> < 0.0001).</p></sec></sec><sec id="s3"><title>Discussion</title><p>Astrocytes in superficial cortical layers were successfully loaded using Fluo-4 AM by surface application up to 350 μm from the pial surface in juvenile rats. In agreement with previous literature (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Dallwig1" ref-type="bibr">Dallwig and Deitmer 2002</xref>; <xref rid="pbio-0020096-Simard1" ref-type="bibr">Simard et al. 2003</xref>), the majority of the Fluo-4-loaded cells exhibited astrocytic morphology with multipolar branching and bushy microprocesses impinging on local vasculature. 2-PLSM imaging revealed spontaneous [Ca<sup>2+</sup>]<sub>i</sub> events in individual astrocytes in vivo. Some coordination of these events was indicated by the broad cross-correlograms in the baseline condition. Increased neuronal discharge was associated with increased astrocytic activity and a robust coordination of [Ca<sup>2+</sup>]<sub>i</sub> signals in neighboring astrocytes, providing evidence for neuron–glia communication in the intact brain.</p><p>The magnitude, frequency and pattern of [Ca<sup>2+</sup>]<sub>i</sub> events observed here are qualitatively similar to those described in tissue cultures (<xref rid="pbio-0020096-Dani1" ref-type="bibr">Dani et al. 1992</xref>; <xref rid="pbio-0020096-Charles1" ref-type="bibr">Charles 1998</xref>) and acute hippocampal, neocortical, and thalamic slice preparations (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>; <xref rid="pbio-0020096-Nett1" ref-type="bibr">Nett et al.</xref>2002; <xref rid="pbio-0020096-Tashiro1" ref-type="bibr">Tashiro et al. 2002</xref>). It has been reported that the percentage of active astrocytes in brain slices showed a 2- to 3-fold decrease from early postnatal days to juvenile age (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>). In our experiments, a large portion of the imaged astrocytes were active, showing either [Ca<sup>2+</sup>]<sub>i</sub> or plateau potentials. It is unlikely that the elevated activity in vivo is due to anesthesia because urethane is known to suppress transmitter release from presynaptic vesicles and attenuate both α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) and NMDA receptors (<xref rid="pbio-0020096-Hara1" ref-type="bibr">Hara and Harris 2002</xref>). Since blockade of these receptors decreases astrocytic [Ca<sup>2+</sup>]<sub>i</sub> activity in vitro (<xref rid="pbio-0020096-Parri1" ref-type="bibr">Parri et al. 2001</xref>; <xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>), it is expected that in the drug-free animal the percentage of active cells will be even higher. A different explanation for the lower percentage of active astrocytes in the slice, relative to the in vivo situation and tissue culture preparation, is that the trauma of brain slicing attenuates spontaneous [Ca<sup>2+</sup>]<sub>i</sub> activity. Reactive astrocytes in a stab wound area show very limited [Ca<sup>2+</sup>]<sub>i</sub> activity (<xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>). In addition, the temperature at which the cells are kept may be playing an important role.</p><p>In the absence of provoking conditions, spontaneous [Ca<sup>2+</sup>]<sub>i</sub> activity in individual astrocytes does not spread among astrocytes as an intercellular Ca<sup>2+</sup> wave (<xref rid="pbio-0020096-Nett1" ref-type="bibr">Nett et al.</xref>2002). In baseline condition, the magnitude of correlated activity in nearby and distant astrocytes was quite similar. Nevertheless, the presence of zero-timelag correlation suggests that activity in the astrocytic syncytium in vivo is not random, but is under some coordinated control. Widespread but limited coordination of glial cells can be brought about by common synchronizing inputs in the intact brain, such as vascular and vegetative nervous system control or large-scale slow changes of neuronal excitability. The latter possibility is supported by the observation that ionotropic glutamate receptor antagonists and tetrodotoxin effectively decorrelated the astrocytic network without altering the number of active astrocytes (<xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>). Furthermore, the intact corticothalamic system displays substantial excitability fluctuation at the time scale of the astrocytic [Ca<sup>2+</sup>]<sub>i</sub> events (<xref rid="pbio-0020096-Jando1" ref-type="bibr">Jando et al. 1995</xref>).</p><p>Although neuronal activity is not needed to generate [Ca<sup>2+</sup>]<sub>i</sub> surges in astrocytes (<xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>; <xref rid="pbio-0020096-Nett1" ref-type="bibr">Nett et al.</xref>2002), neurotransmitters can enhance the frequency of such events. The impact of neuronal activity on the glial network is illustrated by the increased activity and enhanced local correlation of [Ca<sup>2+</sup>]<sub>i</sub> signal in astrocytes after regular population bursting of neurons was brought about by the GABA<sub>A</sub>-receptor blocker bicuculline. These changes shared similarities to those observed in hippocampal and neocortical slices (<xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>; <xref rid="pbio-0020096-Tashiro1" ref-type="bibr">Tashiro et al. 2002</xref>). In contrast to the slice situation, we did not find a time-locked triggering of astrocytic events to the neuronal bursts (see also <xref rid="pbio-0020096-Nett1" ref-type="bibr">Nett et al.</xref>2002). This discrepancy may be explained by the magnitude of the evoked neuronal bursts. Bicuculline in vitro evoked rare (greater than 30 s intervals), but very large bursts or afterdischarges (<xref rid="pbio-0020096-Tashiro1" ref-type="bibr">Tashiro et al. 2002</xref>; <xref rid="pbio-0020096-Aguado1" ref-type="bibr">Aguado et al. 2002</xref>). In vivo, synchronous events of moderate size occurred frequently (approximately 0.3 Hz). The enhanced bursts, associated with large field potentials, can be regarded as interictal epileptic spikes (<xref rid="pbio-0020096-Schwartz1" ref-type="bibr">Schwartz and Bonhoeffer 2001</xref>), but seizures were never observed. Although the exact mechanisms of neuron–astrocyte signaling remain to be disclosed, our findings indicate that neuronal and glial networks are coupled in the intact brain.</p><p>Many of the imaged astrocytes had processes (end feet) in close contact with small brain vessels (<xref rid="pbio-0020096-Peters1" ref-type="bibr">Peters et al. 1970</xref>). It has been shown that surges of [Ca<sup>2+</sup>]<sub>i</sub> in astrocytes trigger the release of vasoactive compounds (<xref rid="pbio-0020096-Bezzi1" ref-type="bibr">Bezzi et al. 1998</xref>). Furthermore, stimulation of single astrocytes in cortical slices led to delayed (greater than 30 s) and protracted dilation of the contacted arteriole (<xref rid="pbio-0020096-Zonta1" ref-type="bibr">Zonta et al. 2003</xref>). These findings support the view that a cardinal function of astrocytes in the intact brain is to regulate local circulation according to the metabolic needs of neurons. Overall, the approach introduced in this paper will be a potent tool to investigate these issues in vivo.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><sec id="s4a1"><title>Subjects and surgery</title><p>Male and female rats, 12–16 d postnatal (P12– P16), of the Sprague–Dawley strain were used in these experiments. Animals were deeply anesthetized with 1.7 g/kg urethane. An outline of the craniotomy above the primary somatosensory (barrel) cortex was marked with a dental drill. A metal frame, similar to what has been described in <xref rid="pbio-0020096-Kleinfeld1" ref-type="bibr">Kleinfeld and Denk (2000</xref>), was attached to the skull with cyanoacrylic. A craniotomy (3–4 mm diameter), centered at 1.5 mm posterior to bregma and 2.5 mm from midline, was performed and the dura mater was surgically removed. Care was taken to avoid any damage to pial vessels or the cortex.</p></sec><sec id="s4a2"><title>Dye loading</title><p>Fluo-4 AM (F-14201, 50 μg; Molecular Probes, Eugene, Oregon, United States) was mixed with 2 μl of Pluronic (P-3000, Molecular Probes) and 5 μl of dymethyl sulfoxide (D-8779; Sigma, St. Louis, Missouri, United States) for 15 min. The solution was then diluted in 18 μl of artificial cerebrospinal fluid (ACSF) (125 mM NaCl, 3 mM KCl, 10 mM glucose, 26 mM NaHCO<sub>3</sub>, 1.1 mM NaH<sub>2</sub>PO<sub>4</sub>, 2 mM CaCl<sub>2</sub>, 1 mM MgSO<sub>4</sub>; pH adjusted to 7.4) and mixed for a further 15 min. A small volume (up to 12 μl) of the dye-containing solution was applied to the cortical surface by a micropipette. The solution was retained in place by a small piece gelfoam. The unbound dye was removed 45–60 min after the surface application of Fluo-4 AM by irrigating the exposed surface with ACSF for at least 10 min. The craniotomy was then covered with 1% agar dissolved in phosphate-buffered saline (pH 7.4), and a glass coverslip was placed on a metal frame. This arrangement allowed access for a glass recording electrode from the side. Juvenile rats (P13–P15) were used because we found in preliminary experiments that in adult animals, mostly vascular cells were loaded with the current protocol.</p></sec><sec id="s4a3"><title>Electrophysiological recording</title><p>During the recording session, a heating blanket was placed under the rat to maintain body temperature at approximately 37°C. The electrocardiogram (EKG) was monitored continuously. The R wave of EKG was used to monitor brain pulsation-derived movement of artifacts during imaging. Population bursts of cortical neurons (“interictal” spikes; <xref rid="pbio-0020096-Schwartz1" ref-type="bibr">Schwartz and Bonhoeffer 2001</xref>) were induced by inserting a large-tip (20–50 μm tip diameter) glass pipette, containing 2 mM bicuculline in 0.9% (w/v) NaCl, into the deep layers of the somatosensory cortex. This electrode also served to record local field potential and multiple unit activity. Large population bursts were reliably induced 10–30 min after the insertion of the pipette.</p></sec><sec id="s4a4"><title>Imaging</title><p>A custom-made 2-PLSM was constructed as described earlier (<xref rid="pbio-0020096-Majewska1" ref-type="bibr">Majewska et al. 2000</xref>). In brief, a Ti:S laser (Mira 800F; Coherent, Santa Clara, California, United States) was pumped by a solid state CW laser (Verdi 8; Coherent) to produce a mode-locked beam (840 nm; approximately 100 fs pulse width at 76 MHz repetition rate). The beam was directed to a modified confocal scanhead (Fluoview 300; Olympus, Tokyo, Japan). The fluorescent signal was first filtered with an emission filter (HQ525, passband 525 ± 25 nm; Chroma, Rockingham, Vermont, United States) and detected by an external photo-multiplier tube (R-3896, Hamamatsu Photonics, Hamamatsu City, Japan) with a built-in preamplifier board (F-5 PSU-B; Olympus).</p></sec><sec id="s4a5"><title>Data analysis</title><p>Fluorescence signal was quantified by measuring the mean pixel value of a manually selected somatic area for each frame of the image stack using ImageJ software. The values were exported to MatLab and the fluorescence change <italic>ΔF/F<sub>0</sub></italic> was computed, where <italic>F<sub>0</sub></italic> is the mean of the lowest 20% of the somatic fluorescence signals. Sessions that had visible drifts when image sequences were replayed as animation (the majority of the cells showed correlated activity [<italic>\|r\|</italic> > 0.6], or greater than 10% fluorescence change due to the heartbeat when the cell was imaged in line scan [approximately 200 Hz]) were excluded from the analysis. For display purposes, the signal was convolved with a Hanning window of order three to smooth the signal trace. Power spectra of fluorescent signals were computed using the multi-taper method <italic>(NW = 4)</italic>. For the calcium event detection, <italic>ΔF/F<sub>0</sub></italic> signal was convolved with a Hanning window of order 15. “Spike” events were defined as transient increase of <italic>ΔF/F<sub>0</sub></italic> signal exceeding 20%, lasting 5–50 s. “Plateau” events were defined as sustained increase of <italic>ΔF/F<sub>0</sub></italic> (greater than 20%) signal longer than 50 s. Peak amplitudes of both spike and plateau events required an increase of at least 50% <italic>ΔF/F<sub>0</sub></italic> from the onset of events. Calcium events were automatically detected with the above detection. Cross-correlation between cell pairs was computed by normalizing the <italic>ΔF/F<sub>0</sub></italic> signals to unity (zero mean, unity standard deviation) so that the computed values represent the correlation coefficient between the two signals at a given timelag. All numbers are indicated as mean ± standard deviation, unless otherwise noted.</p></sec><sec id="s4a6"><title>Immunocytochemisty</title><p>Since Fluo-4 AM loading was best visible in the somatic region of the putative astrocytes, we chose S100B antibody (A5110; DakoCytomation, Glostrup, Denmark) because this antibody stains the somatic region of astrocytes as well as its processes (<xref rid="pbio-0020096-Ren1" ref-type="bibr">Ren et al. 1992</xref>). Following Fluo-4 AM loading, acute brain slices (300 μm thickness) were cut coronally around the dye-loaded area using standard procedures. Fluo-4 in cells of the acute brain slices were fixed by incubating the acute brain slices in freshly made saline containing 40 mg/ml 1-ethyl-3-(3-dimethylaminopropyl)-carbodiimide hydrochloride (EDAC, E7750; Sigma) for 30 min. Next, the slices were incubated in formalin-based fixative (4% formaldehyde, 0.1 M phosphate buffer, [pH ∼7.1]) for 30 min. Once the fixation procedures were completed, the sections were mounted on a glass slide and imaged with 2-PLSM (z-stack; wavelength, 840 nm). After imaging of calcium-loaded cells, and three subsequent washes in phosphate-buffered saline (PBS) (1.06 mM KH<sub>2</sub>PO<sub>4</sub>, 155.17 mM NaCl, 2.96 mM NaHPO<sub>4</sub>, pH approximately 7.4), the slices were treated with S100B antibody (made in rabbit, 1:50 dilution) in Triton X-PBS (0.5% Triton X in PBS) overnight. The sections were subsequently washed three times in PBS, followed by incubation with the secondary fluorescent antibody (1:1000 dilution, 711-166-152, CY3 Anti-Rabbit IgG [H + L]; Jackson ImmunoResearch Laboratories, West Grove, Pennsylvania, United States) in Triton X-PBS solution for 2 h. Simultaneous viewing of the two image stacks allowed a systematic comparison of the extent of overlap between Fluo-4 loading and S100B immunoreactivity (<xref ref-type="supplementary-material" rid="sv005">Video S5</xref>).</p></sec></sec></sec><sec sec-type="supplementary-material" id="s5"><title>Supporting Information</title><supplementary-material content-type="local-data" id="sv001"><label>Video S1</label><caption><title>Visualization of Loaded Astrocytes (Low Magnification)</title><p>The primary somatosensory cortex (P15) was stained with Fluo-4 AM in vivo and subsequently imaged in vitro. Acute slices (approximately 300 μm thickness) were cut in cold ACSF after the cells were loaded in vivo. (Z step = 1 μm; scale bar = 50 μm).</p><p>(49 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv001.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sv002"><label>Video S2</label><caption><title>Visualization of Loaded Astrocytes (High Magnification, Layer I)</title><p>Same slice as shown in <xref ref-type="supplementary-material" rid="sv001">Videos S1</xref>, but with higher magnification. Z step = 1 μm; scale bar = 20 μm.</p><p>(48 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv002.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sv003"><label>Video S3</label><caption><title>Visualization of Loaded Astrocytes (High Magnification, Layers II/III)</title><p>Detailed imaging of in vivo-loaded acute slice preparation of the primary somatosensory cortex (P15; approximately 270 μm below the pial surface). Z step = 1 μm; scale bar = 20 μm.</p><p>(39 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv003.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sv004"><label>Video S4</label><caption><title>High-Contrast Image Upper Layers (I to II/III) of the Fluo-4 AM-Loaded Somatosensory Cortex (P15)</title><p>Empty circles in layers II/III, presumed unloaded neurons (note their absence in layer I). The loaded cells have typical glial morphological appearance. Z step = 1 μm; scale bar = 50 μm.</p><p>(50 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv004.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sv005"><label>Video S5</label><caption><title>Double-Labeling of Fluo-4 AM-Loaded Astrocytes with S100B Antibody</title><p>Acute slices (300 μm thickness) were prepared from the in vivo Fluo-4 AM-loaded somatosensory cortex. The slices were subsequently incubated in EDAC containing saline followed by formalin fixation. The loaded astrocytes were identified by examination at various depths and numbered (left). Next, the slices were processed for immunocytochemistry with astrocyte marker S100B. Depth scans (1 μm between the frames) were taken again to determine immunoreactivity of cells with S100B (right movie). An overlapping set of the cells was identified to be S100B-immunoreactive, indicating that nearly all Fluo-4 AM-loaded cells were astrocytes.</p><p>(5 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv005.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sv006"><label>Video S6</label><caption><title>Imaging of Fluo-4 AM Fluorescence Activity in Astrocytes In Vivo</title><p>Movie taken from a P14 rat. Image was taken with 2 Hz sampling rate for 10 min and compressed to 36 s for display purposes. Note spatial- and light-emission-stability of the recorded cells. Note also that at frames approximately 9 s and 15 s, two of the astrocytes in the middle display transient increased fluorescence. Scale bar 50 micro μ.</p><p>(55 MB AVI).</p></caption><media xlink:href="pbio.0020096.sv006.avi"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec>
Neural Activity When People Solve Verbal Problems with Insight	<p>People sometimes solve problems with a unique process called insight, accompanied by an “Aha!” experience. It has long been unclear whether different cognitive and neural processes lead to insight versus noninsight solutions, or if solutions differ only in subsequent subjective feeling. Recent behavioral studies indicate distinct patterns of performance and suggest differential hemispheric involvement for insight and noninsight solutions. Subjects solved verbal problems, and after each correct solution indicated whether they solved with or without insight. We observed two objective neural correlates of insight. Functional magnetic resonance imaging (<xref ref-type="sec" rid="s2">Experiment 1</xref>) revealed increased activity in the right hemisphere anterior superior temporal gyrus for insight relative to noninsight solutions. The same region was active during initial solving efforts. Scalp electroencephalogram recordings (<xref ref-type="sec" rid="s2">Experiment 2</xref>) revealed a sudden burst of high-frequency (gamma-band) neural activity in the same area beginning 0.3 s prior to insight solutions. This right anterior temporal area is associated with making connections across distantly related information during comprehension. Although all problem solving relies on a largely shared cortical network, the sudden flash of insight occurs when solvers engage distinct neural and cognitive processes that allow them to see connections that previously eluded them.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Jung-Beeman</surname><given-names>Mark</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Bowden</surname><given-names>Edward M</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Haberman</surname><given-names>Jason</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Frymiare</surname><given-names>Jennifer L</given-names></name><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Arambel-Liu</surname><given-names>Stella</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Greenblatt</surname><given-names>Richard</given-names></name><xref ref-type="aff" rid="aff3"> <sup>3</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Reber</surname><given-names>Paul J</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Kounios</surname><given-names>John</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>According to legend, Archimedes shouted “Eureka!” (“I have found it!”) when he suddenly discovered that water displacement could be used to calculate density. Since then, “Eureka!,” or “Aha!,” has often been used to express the feeling one gets when solving a problem with <italic>insight</italic>. Insight is pervasive in human (and possibly animal [<xref rid="pbio-0020097-Epstein1" ref-type="bibr">Epstein et al. 1984</xref>]) cognition, occurring in perception, memory retrieval, language comprehension, problem solving, and various forms of practical, artistic, and scientific creativity (<xref rid="pbio-0020097-Davidson1" ref-type="bibr">Sternberg and Davidson 1995</xref>). The Archimedes legend has persisted over two millennia in part because it illustrates some of the key ways in which insight solutions differ from solutions achieved through more straightforward problem solving. We examine the neural bases of these different problem-solving methods.</p><p>Although many processes are shared by most types of problem solving, insight solutions appear to differ from noninsight solutions in several important ways. The clearest defining characteristic of insight problem solving is the subjective “Aha!” or “Eureka!” experience that follows insight solutions (<xref rid="pbio-0020097-Schooler2" ref-type="bibr">Schooler et al. 1993</xref>). This subjective experience can lead to a strong emotional response—according to legend, Archimedes ran home from the baths shouting “Eureka!” without donning his clothes first. In addition, problem solving with insight is characterized by the following features. (1) Solvers first come to an impasse, no longer progressing toward a solution (<xref rid="pbio-0020097-Duncker1" ref-type="bibr">Duncker 1945</xref>). Archimedes, for example, was stymied by King Hiero's challenge to determine whether his new crown was pure gold without damaging the crown. (2) Solvers usually cannot report the processing that enables them to reinterpret the problem and overcome the impasse (<xref rid="pbio-0020097-Maier1" ref-type="bibr">Maier 1931</xref>). Insight often occurs when people are not even aware they are thinking of the problem, as reportedly happened to Archimedes while in the baths. (3) Solvers experience their solutions as arising suddenly (<xref rid="pbio-0020097-Metcalfe2" ref-type="bibr">Metcalfe and Wiebe 1987</xref>; <xref rid="pbio-0020097-Smith1" ref-type="bibr">Smith and Kounios 1996</xref>) and immediately recognize the correctness of the solution (or solution path). (4) Performance on insight problems is associated with creative thinking and other cognitive abilities different from those associated with performance on noninsight problems (<xref rid="pbio-0020097-Schooler1" ref-type="bibr">Schooler and Melcher 1997</xref>). Some researchers have argued that all these characteristics of insight solutions are essentially epiphenomenal, that insight and noninsight solutions vary only in emotional intensity, and that they are attained with precisely the same cognitive (hence neural) mechanisms (<xref rid="pbio-0020097-Weisberg2" ref-type="bibr">Weisberg and Alba 1981</xref>; <xref rid="pbio-0020097-Weisberg1" ref-type="bibr">Weisberg 1986</xref>; <xref rid="pbio-0020097-Perkins1" ref-type="bibr">Perkins 2000</xref>).</p><p>Persistent questions about insight concern whether unconscious processing precedes reinterpretation and solution, whether distinct cognitive and neural mechanisms beyond a common problem-solving network are involved in insight, and whether the apparent suddenness of insight solutions reflects truly sudden changes in cognitive processing and neural activity.</p><p>Recent work suggests that people are thinking—at an unconscious level—about the solution prior to solving problems with insight. Specifically, while working on a verbal problem they have yet to solve, people presented with a potential solution word read the actual solution word faster than they read an unrelated word (<xref rid="pbio-0020097-Beeman2" ref-type="bibr">Bowden and Beeman 1998</xref>). This “solution priming” effect is greater—and in fact people make solution decisions about presented words more quickly—when words are presented to the left visual hemifield, which projects directly to the right hemisphere (RH), than when words are presented to the right visual hemifield, which projects to the left hemisphere (LH). This suggests that RH semantic processing is more likely than LH semantic processing to produce lexical or semantic information that leads to the solution. These RH advantages occur only when solvers experience insight—the “Aha!” or “Eureka!” feeling that comes with insight solutions (<xref rid="pbio-0020097-Bowden2" ref-type="bibr">Bowden and Jung-Beeman 2003a</xref>). Moreover, when subjects try to solve classic insight problems, they benefit more from hints presented to the left visual field (i.e., the RH) than from hints presented to the right visual field (i.e., the LH) (<xref rid="pbio-0020097-Fiore1" ref-type="bibr">Fiore and Schooler 1998</xref>).</p><p>Problem solving is a complex behavior that requires a network of cortical areas for all types of solving strategies and solutions, so solving problems with and without insight likely invokes many shared cognitive processes and neural mechanisms. One critical cognitive process distinguishing insight solutions from noninsight solutions is that solving with insight requires solvers to recognize distant or novel semantic (or associative) relations; hence, insight-specific neural activity should reflect that process. The most likely area to contribute to this component of insight problem solving is the anterior superior temporal gyrus (aSTG) of the RH. Language comprehension studies demonstrate that the RH is particularly important for recognizing distant semantic relations (<xref rid="pbio-0020097-Chiarello1" ref-type="bibr">Chiarello et al. 1990</xref>; <xref rid="pbio-0020097-Beeman2" ref-type="bibr">Beeman 1998</xref>), and bilateral aSTG is involved in semantic integration. For example, sentences and complex discourse increase neural activity in aSTG bilaterally (<xref rid="pbio-0020097-Mazoyer1" ref-type="bibr">Mazoyer et al. 1993</xref>; <xref rid="pbio-0020097-Stowe1" ref-type="bibr">Stowe et al. 1999</xref>), and discourse that places particular demands on recognizing or computing distant semantic relations specifically increases neural activity in RH temporal areas (<xref rid="pbio-0020097-St1" ref-type="bibr">St. George et al. 1999</xref>; <xref rid="pbio-0020097-Mason1" ref-type="bibr">Mason and Just 2004</xref>), especially aSTG (<xref rid="pbio-0020097-Meyer1" ref-type="bibr">Meyer et al. 2000</xref>; <xref rid="pbio-0020097-Kircher1" ref-type="bibr">Kircher et al. 2001</xref>). If this prediction of RH aSTG involvement is confirmed, it will help constrain neurocognitive theories of insight. Other cortical areas, such as prefrontal cortex and the anterior cingulate (AC) may also be differentially involved in producing insight and noninsight solutions.</p><p>We used functional magnetic resonance imaging (FMRI) in <xref ref-type="sec" rid="s2">Experiment 1</xref> and electroencephalogram (EEG) measurement in <xref ref-type="sec" rid="s2">Experiment 2</xref> to test the empirically and theoretically derived hypothesis that solving problems with insight requires engagement of (or increased emphasis on) distinct neural mechanisms, particularly in the RH anterior temporal lobe. Event-related experimental designs compared neural activity when people solved verbal problems with insight to neural activity when they solved problems (from the same problem set) without insight.</p><p>As in earlier behavioral work, we used a set of compound remote associate problems (<xref rid="pbio-0020097-Bowden3" ref-type="bibr">Bowden and Jung-Beeman 2003b</xref>) adapted from a test of creative cognition (<xref rid="pbio-0020097-Mednick1" ref-type="bibr">Mednick 1962</xref>). <xref ref-type="fig" rid="pbio-0020097-g001">Figure 1</xref> illustrates the sequence for each trial. Subjects saw three problem words <italic>(pine, crab, sauce)</italic> and attempted to produce a single solution word <italic>(apple)</italic> that can form a familiar compound word or phrase with each of the three problem words <italic>(pineapple, crab apple, applesauce)</italic>. We relied on solvers' reports to sort solutions into insight solutions and noninsight solutions, avoiding the complication that presumed insight problems can sometimes be solved without insight (<xref rid="pbio-0020097-Davidson1" ref-type="bibr">Davidson 1995</xref>) and circumventing the use of different types of problems requiring different cognitive operations. Thus, we made use of the most important defining characteristic of insight problems: the subjective conscious experience—the “Aha!” A similar technique revealed distinct behavioral characteristics when people recognized solutions with insight (<xref rid="pbio-0020097-Bowden2" ref-type="bibr">Bowden and Jung-Beeman 2003a</xref>). Note that this is a very “tight” comparison. In both conditions problems are solved using a network of processes common to both insight and noninsight solutions. If insight ratings reflect some distinct cognitive processes, this contrast will reveal the distinct underlying brain activity. In other words, within the cortical network for problem solving, different components will be engaged or emphasized for insight versus noninsight solutions. FMRI (<xref ref-type="sec" rid="s2">Experiment 1</xref>) should reveal neuroanatomical locations of processes that are unique to insight solutions, and EEG (<xref ref-type="sec" rid="s2">Experiment 2</xref>) should reveal the time course (e.g., whether insight really is sudden) and frequency characteristics of neurophysiological differences.</p><fig id="pbio-0020097-g001" position="float"><label>Figure 1</label><caption><title>Sequence of Events for Each Trial</title><p>(A) The “Compound” prompt was presented for 0.5 s, then persisted for a variable amount of additional time (0–2 s) until a cue from the scanner indicated the beginning of a new whole brain acquisition. (B) A three-word problem appeared in the center of the screen and persisted until subjects indicated with a bimanual button press that they had solved the problem, or until the 30-s time limit elapsed. Thus, event timing and condition were completely dependent on subjects' responses. (C) Following the button press or time limit, subjects were prompted to verbalize the solution (or press the buttons and say “Don't know” if the time limit expired prior to solution) then (D) prompted to indicate (with a bimanual button press) whether they felt insight, as described prior to the experiment. (E) Next, subjects performed 9 s of an unrelated filler task (three line-matching trials, 3 s each), allowing BOLD signal to return to baseline (in areas not involved in line matching).</p></caption><graphic xlink:href="pbio.0020097.g001"/></fig></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>Experiment 1</title><p>Subjects solved 59% of the problems presented, and pressed buttons indicating “insight” for 56% (s.d. = 18.2) of their solutions, “no insight” for 41% (s.d. = 18.9) of their solutions, and “other” for 2% of their solutions. We marked a point about 2 s (rounded to the nearest whole second) prior to each solution button press as the solution event, and examined a time window 4–9 s after this event (i.e., 2–7 s after the button press) to isolate the corresponding hemodynamic response. Solving problems and responding to them required a strict sequence of events (reading of words, solving effort, solving, button press, verbalizing the solution, insight decision), but this sequence was identical whether subjects indicated solving with or without insight, so differences in FMRI signal resulted from the degree to which distinct cognitive processes and neural systems led to insight or noninsight solutions.</p><p> <xref ref-type="fig" rid="pbio-0020097-g002">Figure 2</xref> illustrates the most robust insight effect: as predicted, insight solutions were associated with greater neural activity in the RH aSTG than noninsight solutions. The active area was slightly anterior to primary auditory cortex, posterior to temporal pole, and along the medial aspect of the aSTG, extending down the lateral edge of the descending ramus of the Sylvian fissure to midway through the middle temporal gyrus (MTG). (This site is also close to the superior temporal sulcus, which has been implicated in language). Across all 13 subjects, the peak signal difference at a single voxel within the RH aSTG was 0.25% across the 6-s window, and 0.30% at a single time to repetition (TR), i.e., the time needed to repeat the image of the whole brain. Overall signal in this region was robust, reaching 96.8% of the brainwide average (after removing voxels in other brain areas with signal below a standard criterion). Within the cluster of voxels identified across the group, 12 subjects showed from 0.03% to 0.35% greater signal for insight than for noninsight solutions; one subject showed 0.02% greater signal for the noninsight solutions. It is not likely that RH aSTG is involved only in output or in emotional response following insight solutions, because neural activity in this area also increased when subjects first encountered each problem (<xref ref-type="fig" rid="pbio-0020097-g003">Figure 3</xref>). Thus, RH aSTG is involved in processing the problem words both initially and at solution. (Of course, event-related FMRI signal occurred in many other cortical regions at problem onset, especially visual cortex). There was no insight effect in response windows immediately preceding or following the defined response window. All indications point to a striking transient event in the RH aSTG near the time when subjects solve problems with insight.</p><fig id="pbio-0020097-g002" position="float"><label>Figure 2</label><caption><title>FMRI Insight Effect in RH aSTG</title><p>(A) Voxels showing greater FMRI signal for insight than noninsight solutions, overlaid on the averaged normalized structural image of all subjects. The active area has a volume of 531 mm<sup>3</sup> (peak <italic>t</italic> = 4.89 at 44, −9, −9 in Talairach space).</p><p>(B) and (C) Group average signal change following the solution event, for insight (red line) and noninsight (blue line) solutions (yellow arrow indicates button press): (B) over entire LH aSTG region; (C) over entire RH aSTG region.</p><p>(D) Insight solution signal change minus noninsight solution signal change, in RH aSTG (error bars show the standard error of the mean of the difference at each timepoint).</p></caption><graphic xlink:href="pbio.0020097.g002"/></fig><fig id="pbio-0020097-g003" position="float"><label>Figure 3</label><caption><title>FMRI Signal in RH aSTG during Initial Solving Efforts</title><p>(A) Voxels in right temporal lobe showing baseline-to-peak event-related FMRI signal when subjects first encounter problems, overlaid on the averaged normalized structural image of all subjects. The cluster is in RH aSTG, with a volume of 469 mm<sup>3</sup>, with peak <italic>t</italic> value of 4.37 at 41, −6, −12 in Talairach space, clearly overlapping with the cluster showing an insight effect at solution.</p><p>(B) Group average signal change following problem onset (time = 0), for the cluster defined by signal at the problem onset (green line) and the cluster (illustrated in <xref ref-type="fig" rid="pbio-0020097-g002">Figure 2</xref>A) showing the insight effect at solution (white line). Error bars show the standard error of the mean of the difference at each time point.</p></caption><graphic xlink:href="pbio.0020097.g003"/></fig><p>The involvement of the RH rather than the LH for this verbal task is not due to greater difficulty in producing insight solutions: subjects produced insight solutions at least as quickly (mean solution time = 10.25 s, s.d. = 3.58 s) as they produced noninsight solutions (mean = 11.28 s, s.d. = 4.13 s) (<italic>t</italic> < 1.0, <italic>p</italic> > 0.3). More importantly, the hemodynamic responses to both insight and noninsight solutions in the homologous area of the LH are about equivalent to the response to noninsight solutions in the RH aSTG—it is the strong response to insight solutions in the RH aSTG that stands out. There is no insight effect anywhere within temporal cortex of the LH. At statistical thresholds below significant levels (<italic>p</italic> < 0.1 uncorrected), there are as many voxels in LH temporal cortex showing a noninsight effect as showing an insight effect.</p><p>Several other cortical areas showing insight effects that did not meet significance criteria are listed in <xref ref-type="table" rid="pbio-0020097-t001">Table 1</xref> (see also <xref ref-type="supplementary-material" rid="sg001">Figure S1</xref>). Some of these effects were in frontal cortex, which is notable because various frontal areas have been implicated in problem solving and reasoning. Patients with prefrontal damage have particular difficulty integrating relations in reasoning tasks (<xref rid="pbio-0020097-Waltz1" ref-type="bibr">Waltz et al. 1999</xref>), and when healthy subjects perform the same task, neural activity increases in rostrolateral prefrontal cortext (<xref rid="pbio-0020097-Christoff1" ref-type="bibr">Christoff et al. 2001</xref>). Some problem solving increases activity in dorsolateral prefrontal cortex (<xref rid="pbio-0020097-Prabhakaran1" ref-type="bibr">Prabhakaran et al. 1997</xref>), perhaps because of working memory demands. Solving of poorly structured problems seems particularly impaired following damage to the prefrontal cortex of the RH (<xref rid="pbio-0020097-Goel2" ref-type="bibr">Goel and Grafman 2000</xref>). Moreover, the inferior frontal gyrus (IFG) is highly active when people engage in directed semantic retrieval (<xref rid="pbio-0020097-Wagner1" ref-type="bibr">Wagner et al. 2001</xref>) or when they select particular semantic concepts over competing ones (<xref rid="pbio-0020097-Thompson-Schill1" ref-type="bibr">Thompson-Schill et al. 1997</xref>), e.g., to generate a response (<xref rid="pbio-0020097-Frith1" ref-type="bibr">Frith et al. 1991</xref>). Usually in these circumstances the IFG activity is stronger in the LH, even when people are reasoning about spatial problems (<xref rid="pbio-0020097-Goel3" ref-type="bibr">Goel et al. 1998</xref>), but the IFG responds particularly strongly in the RH when subjects select more distant semantic relations because of task demands (<xref rid="pbio-0020097-Seger1" ref-type="bibr">Seger et al. 2000</xref>) or comprehension goals (<xref rid="pbio-0020097-Robertson1" ref-type="bibr">Robertson et al. 2000</xref>). Because of its putative importance for problem solving, semantic retrieval, and semantic selection, IFG was an a priori region of interest. One question we had hoped to answer was whether the semantic selection of insight solutions would preferentially evoke activity in RH or LH IFG, but the insight effects in both areas were too small (in area and in reliability) to test this question. When a more lenient statistical threshold was adopted, small clusters of signal were observed in both RH and LH IFG (<xref ref-type="table" rid="pbio-0020097-t001">Table 1</xref>; <xref ref-type="supplementary-material" rid="sg001">Figure S1</xref>A). Indeed, within the small region surpassing this weak statistical threshold, signal change in the RH IFG region was moderately strong (peak = 0.21% across the whole window). However, as is often the case, FMRI signal in this region was low (about 72% of the brainwide average) and variability was high, decreasing our confidence in the effect.</p><table-wrap id="pbio-0020097-t001" position="float"><label>Table 1</label><caption><title>Full FMRI Results of Insight Effect</title></caption><graphic xlink:href="pbio.0020097.t001"/><table-wrap-foot><fn id="nt101"><p>All areas showing an “insight effect”—stronger signal for insight solutions than noninsight solutions (ordered by mean percent signal change). All cluster sizes represent active voxels at <italic>t</italic>(12) = 3.43, <italic>p</italic> < 0.005, except bilateral inferior frontal gyrus areas (*), shown at 2.83, <italic>p</italic> < 0.015, because it was an a priori region of interest. Location of cluster centers and peak <bold><italic>t</italic></bold> values are shown in Talairach coordinates</p></fn></table-wrap-foot></table-wrap><p>After RH aSTG, the second largest area showing an insight effect in FMRI signal was the medial frontal gyrus in the LH (<xref ref-type="table" rid="pbio-0020097-t001">Table 1</xref>; <xref ref-type="supplementary-material" rid="sg001">Figure S1</xref>B). Although this area was 85% as large (453 mm<sup>3</sup> at <italic>p</italic> < 0.005 threshold) as RH aSTG, the event-related signal within it was weak and the insight–noninsight difference (peak difference = 0.15%) was relatively small. (The insight effect may be attributable as much to a negative response for noninsight solutions as to a positive response for insight solutions.)</p><p>There also was an insight effect in small clusters in or near bilateral amygdala or parahippocampal gyrus. Again, regional signal was low (83% of the brainwide average), and the signal difference was small (peak = 0.16%). However, an amygdalar response may be expected, given the emotional sensation of the insight experience (Parsons and Osherson, 2001). Hippocampal or parahippocampal involvement is also plausible, if memory interacts with insight solutions differently from how it interacts with noninsight solutions. For instance, insight problems may encourage distinct memory encoding (<xref rid="pbio-0020097-Wills1" ref-type="bibr">Wills et al. 2000</xref>) or may require distinct retrieval. Finally, a small cluster in the LH posterior cingulate (PC) also showed an insight effect. There was strong, sustained FMRI signal for both solution types in this region; on the fringe of this responding region, FMRI signal began earlier following insight than noninsight solutions. The lateness of the FMRI signal across LH PC suggests that this effect began later in the response sequence, rather than during solution generation. Finally, as in most FMRI studies, signal was relatively weak in temporal pole and orbitofrontal areas due to magnetic susceptibility artifact, so we cannot rule out undetected effects in those areas.</p><p>Several cortical areas showed strong solution-related FMRI signal, but approximately equally for insight and noninsight solutions. Some of these areas (e.g., motor cortex) relate to the response sequence rather than solution processes; other areas probably reflect component processes of a problem-solving network common to both insight and noninsight solving, such as retrieving potential solutions. Two areas that may be of interest for future studies are AC and posterior middle/superior temporal gyrus. Both these areas, in the RH only, showed strong, negative solution-related signal, approximately equal in the two solution types. AC is an area that might be predicted to be involved in reorienting attention as solvers overcome impasses, given its role in performance monitoring and cognitive control (<xref rid="pbio-0020097-MacDonald1" ref-type="bibr">MacDonald et al. 2000</xref>). RH posterior MTG is active when subjects “get” jokes (<xref rid="pbio-0020097-Goel1" ref-type="bibr">Goel and Dolan 2001</xref>) and when they attempt to solve problems with deductive reasoning (<xref rid="pbio-0020097-Parsons1" ref-type="bibr">Parsons and Osherson 2001</xref>). However, in our experiment, only the RH aSTG showed a robust insight effect.</p></sec><sec id="s2b"><title>Experiment 2</title><p>A separate group of subjects participated in fundamentally the same paradigm while we continuously recorded EEGs from the scalp. We then compared time-frequency analyses of the EEGs associated with insight solutions versus noninsight solutions. EEG provides temporal resolution greatly superior to that of FMRI and thus can better elucidate the time course and suddenness of the insight effect. Furthermore, complex EEG oscillations can be parsed into constituent frequency components, some of which have been linked to particular types of neural and cognitive processes (<xref rid="pbio-0020097-Ward1" ref-type="bibr">Ward 2003</xref>).</p><p>The high temporal resolution of EEG allows us to address one of the fundamental questions raised earlier: does insight really occur suddenly, as subjective experience suggests? For problems typically solved without insight, solvers report gradually increasing closeness to solution. In contrast, for problems typically solved with insight, solvers report little or no progress until shortly before they actually solve the problem (<xref rid="pbio-0020097-Metcalfe1" ref-type="bibr">Metcalfe 1986</xref>; <xref rid="pbio-0020097-Metcalfe2" ref-type="bibr">Metcalfe and Wiebe 1987</xref>). Similarly, quantitative analyses of the distributions of response times and accuracies during anagram solving (a task frequently eliciting the experience of insight) reveal that a solution becomes available in a discrete transition from a state of little or no information about the correct response directly to the final state of high accuracy. This contrasts with various language and memory tasks not associated with insight, which yield partial outputs before processing has been completed (<xref rid="pbio-0020097-Kounios1" ref-type="bibr">Kounios and Smith 1995</xref>; <xref rid="pbio-0020097-Smith1" ref-type="bibr">Smith and Kounios 1996</xref>).</p><p>We predicted that a sudden change in neural activity associated with insight solutions would produce an EEG correlate. Specifically, we predicted that high-frequency EEG oscillations in the gamma band (i.e., greater than 30 Hz) would reflect this sudden activity, because prior research has associated gamma-band activity with the activation of perceptual, lexical, and semantic representations (<xref rid="pbio-0020097-Tallon-Baudry1" ref-type="bibr">Tallon-Baudry and Bertrand 1999</xref>; <xref rid="pbio-0020097-Pulvermuller1" ref-type="bibr">Pulvermüller 2001</xref>). Gamma-band electrical activity correlates with the blood oxygenation level–dependent (BOLD) response apparent in FMRI signal; lower-frequency EEG components do not seem to have direct correlates in FMRI signal (<xref rid="pbio-0020097-Foucher1" ref-type="bibr">Foucher et al. 2003</xref>; <xref rid="pbio-0020097-Laufs1" ref-type="bibr">Laufs et al. 2003</xref>). Consequently, based on the language literature discussed earlier and on our FMRI results, we predicted a discrete insight-related increase in gamma-band activity at electrodes over the anterior temporal lobe of the RH.</p><p>Participants solved 46% (s.d. = 8.2) of the problems correctly within the time limit. Of correctly solved problems, subjects reported more insight solutions (56%, s.d. = 8.4) than noninsight solutions (42%, s.d. = 9.0), (<italic>t</italic>[18] = 3.47, <italic>p</italic>=0.003); there was no difference in mean response times (insight solutions = 9.94 s, s.d. = 2.60; noninsight solutions=9.25 s, s.d. = 3.06; t < 1.0).</p><p>There was a burst of gamma-band activity associated with correct insight solutions (but not noninsight solutions) beginning approximately 0.3 s before the button-press solution response at anterior right temporal electrodes (<xref ref-type="fig" rid="pbio-0020097-g004">Figure 4</xref>), with no significant difference between insight and noninsight solutions over homologous LH sites. A repeated-measures analysis of variance (ANOVA) performed on log-transformed gamma-band (39 Hz) EEG power at left and right temporal electrode sites (T7 and T8, respectively) for insight and noninsight trials using two time windows (−1.52 to −0.36 s and −0.30 to −0.02 s, measured with respect to the solution response) yielded significant insight × time window (<italic>F</italic>[1,18] = 6.68, <italic>p</italic> = 0.019) and insight × time window × Hemisphere (<italic>F</italic>[1,18] = 8.11, <italic>p</italic> = 0.011) interactions. The overall interaction occurred because there was an insight × hemisphere interaction from −0.30 to −0.02 s (<italic>F</italic>[1,18] = 4.61, <italic>p</italic> = 0.046) but no effect in the −1.52 to −0.36 s time window. Within the −0.30 to −0.02 s interval for these two electrodes, there was a significant insight effect at the right temporal (T8) site (<italic>t</italic>[18] = 3.48, <italic>p</italic> = 0.003), but not at the homologous left temporal (T7) site or any other LH temporal electrode. Laplacian mapping of this effect (<xref ref-type="fig" rid="pbio-0020097-g004">Figure 4</xref>B) is remarkably consistent with the FMRI signal in RH aSTG observed in <xref ref-type="sec" rid="s2">Experiment 1</xref>. (EEG does not have the spatial resolution of FMRI. However, we used the Laplacian transform [i.e., second spatial derivative] to localize observed activity. The Laplacian derivation acts as a high-pass spatial filter that reduces the contribution from activity in distant areas of the brain to the signal at a given electrode, and therefore reflects relatively focal and proximal brain activity. Given our FMRI results and the demonstrated correspondence between high-frequency EEG activity and FMRI signal [<xref rid="pbio-0020097-Foucher1" ref-type="bibr">Foucher et al. 2003</xref>; <xref rid="pbio-0020097-Laufs1" ref-type="bibr">Laufs et al. 2003</xref>], we are confident in the localization of this effect.)</p><fig id="pbio-0020097-g004" position="float"><label>Figure 4</label><caption><title>Gamma-Band Power for Insight and Noninsight Solutions</title><p>(A) Grand average time course of EEG power (in v<sup>2</sup>) at 39 Hz estimated with the Morlet wavelet transform at right superior temporal electrode T8. The <italic>x</italic>-axis represents time (in seconds) with the yellow arrow and <italic>R</italic> marking the point in time of the solution button-press response (i.e., 0.0 s). The green horizontal bars above the <italic>x</italic>-axis represent the time intervals used in the statistical analyses and topographic maps. Note that gamma-band power for insight trials (red line) starts to increase above power on noninsight trials (blue line) by approximately 0.3 s before the button press.</p><p>(B) Time-frequency plots of the insight minus noninsight difference shown in (A). The <italic>y</italic>-axis represents frequency (in Hz); the <italic>x</italic>-axis represents time (in seconds, with respect to the button press, exactly as shown in [A]). Red areas of the plot reflect times and frequencies at which insight EEG power is greater than noninsight EEG power; blue areas reflect times and frequencies at which noninsight EEG power is greater than insight EEG power. Note the sudden emergence of increased gamma power for insight solutions approximately 0.3 s before the button press.</p><p>(C) Insight minus noninsight gamma-band differences plotted as topographic maps (LH and RH) of scalp current density (in v<sup>2</sup>/m<sup>2</sup>) estimated by a spline-based Laplacian transform computed with a realistic FMRI-derived head model. The Laplacian transform acts as a high-pass spatial filter that minimizes the contribution of activity distant from each electrode, thereby manifesting discrete, relatively superficial sources. The maps are thresholded to show foci of current density at the upper and lower 20% of the scale. Note the prominent effect of insight (effect for insight greater than effect for noninsight, in red) at the right superior temporal electrode (T8) and surrounding electrodes present from −0.30 to −0.02 s (measured with respect to the solution response) that is not present in the earlier epoch (−1.52 to −0.36 s). The blue area over left inferior parietal cortex (electrode P7) indicates that noninsight gamma power is nonsignificantly greater than insight power (<italic>F</italic>[1,19] < 1) over this region.</p></caption><graphic xlink:href="pbio.0020097.g004"/></fig><p>The gamma burst in the right temporal area cannot be attributed to motor processes involved in making the response because (A) motor activity associated with the bimanual button press would have caused a bilateral gamma burst, not a unilateral one; (B) the location of the gamma burst as determined by Laplacian mapping (<xref ref-type="fig" rid="pbio-0020097-g004">Figure 4</xref>B) is not consistent with hand-related motor cortex activity; and (C) both insight and noninsight solutions required button presses.</p><p>Other planned statistical tests (ANOVAs) examined possible insight-related frontal theta (5–8 Hz), posterior alpha (8–13 Hz), and fronto-central beta (13–20 Hz) activity. There were no statistically significant theta or beta effects. (Visual inspection and post hoc statistical tests suggested insight-related frontal 4-Hz activity, but this effect cannot be reliably distinguished from possible artifacts due to small vertical eye movements.) There was a significant posterior alpha effect, which is discussed below.</p></sec></sec><sec id="s3"><title>Discussion</title><p>Complex problem solving requires a complex cortical network to encode the problem information, search memory for relevant information, evaluate this information, apply operators, and so forth. The FMRI and EEG results reported here conclusively demonstrate that solving verbal problems with insight requires at least one additional component to this cortical network, involving RH aSTG, that is less important to solving without insight. The insight effect in RH aSTG accords with the literature on integrating distant or novel semantic relations during language comprehension. When people comprehend (read or listen to) sentences or stories, neural activity increases in aSTG or temporal pole bilaterally more than when comprehending single words (<xref rid="pbio-0020097-Mazoyer1" ref-type="bibr">Mazoyer et al. 1993</xref>; <xref rid="pbio-0020097-Bottini1" ref-type="bibr">Bottini et al. 1994</xref>; <xref rid="pbio-0020097-Stowe1" ref-type="bibr">Stowe et al. 1999</xref>; <xref rid="pbio-0020097-Humphries1" ref-type="bibr">Humphries et al. 2001</xref>; <xref rid="pbio-0020097-Meyer1" ref-type="bibr">Meyer et al. 2000</xref>). Neural activity increases in predominantly RH aSTG during tasks that emphasize integration across sentences to extract themes (<xref rid="pbio-0020097-St1" ref-type="bibr">St. George et al. 1999</xref>) or to form more coherent memories for stories (<xref rid="pbio-0020097-Mason1" ref-type="bibr">Mason and Just 2004</xref>). RH aSTG is also selectively active when subjects must generate the best ending to a sentence (<xref rid="pbio-0020097-Kircher1" ref-type="bibr">Kircher et al. 2001</xref>) or mentally repair grammatically incorrect sentences (<xref rid="pbio-0020097-Meyer1" ref-type="bibr">Meyer et al. 2000</xref>), both of which likely require intense semantic integration.</p><p>Like the results in language processing, the current results are predicted by the theory that the RH performs relatively coarse semantic coding (<xref rid="pbio-0020097-Beeman2" ref-type="bibr">Beeman 1998</xref>; similarly, <xref rid="pbio-0020097-Chiarello1" ref-type="bibr">Chiarello et al. 1990</xref>). This theory contends that when people encounter words, semantic processing in several LH areas engages in relatively fine semantic coding which produces small semantic fields—i.e., this processing strongly focuses on a few concepts closely related to the input word in the given context. This is very effective for most straightforward language processing. In contrast, the homologous RH areas engage in relatively coarse semantic coding, which produces large and weak semantic fields—i.e., this processing includes many concepts, even concepts distantly related to the input words and context. This process is ineffective for rapid interpretation or selection but increases semantic overlap among multiple semantic fields (<xref rid="pbio-0020097-Beeman3" ref-type="bibr">Beeman et al.</xref>1994), which is useful when drawing together parts of a story or conversation that are only distantly related (<xref rid="pbio-0020097-Beeman1" ref-type="bibr">Beeman 1993</xref>; <xref rid="pbio-0020097-Beeman3" ref-type="bibr">Beeman et al. 2000</xref>). In this view, the coarseness of semantic coding is largely influenced by slight asymmetries in neural microcircuitry that produce more discrete, less redundant input fields in pyramidal neurons of the LH language cortex, and more overlapping input fields in corresponding neurons in the RH (for reviews see <xref rid="pbio-0020097-Beeman2" ref-type="bibr">Beeman 1998</xref>; <xref rid="pbio-0020097-Hutsler1" ref-type="bibr">Hutsler and Galuske 2003</xref>).</p><p>We suggest that semantic integration, generally, is important for connecting various problem elements together and connecting the problem to the solution, and that coarsely coded semantic integration, computed in RH aSTG, is especially critical to insight solutions, at least for verbal problems (or problems that can be solved with verbal or semantic information). People come to an impasse on insight problems because their retrieval efforts are misdirected by ambiguous information in the problem or by their usual method for solving similar problems. Large semantic fields allowing for more overlap among distantly related concepts (or distantly associated lexical items) may help overcome this impasse. Because this semantic processing is weak, it may remain unconscious, perhaps overshadowed by stronger processing of the misdirected information (<xref rid="pbio-0020097-Schooler2" ref-type="bibr">Schooler et al. 1993</xref>; <xref rid="pbio-0020097-Smith2" ref-type="bibr">Smith 1995</xref>), and solvers remain stuck at impasse. Eventually, solution-related information bursts into awareness “in a sudden flash.” This can happen after misdirected processing decays or is suppressed, after solution-related processing grows, or after environmental cues occur—such as the water overflowing the bathtub when Archimedes got in. Archimedes had semantic and verbal knowledge about how to compute density from weight and volume, but struggled with measuring the volume of an irregularly shaped crown without harming the crown (e.g., melting it). His observation of water displacement allowed him to connect known concepts in new ways. This is the nature of many insights, the recognition of new connections across existing knowledge.</p><p>A persistent question has been whether the cognitive and neural events that lead to insight are as sudden as the subjective experience. The timing and frequency characteristics of the EEG results shed light on this question. We propose that the gamma-band insight effect in <xref ref-type="sec" rid="s2">Experiment 2</xref> reflects the sudden transition of solution-related cognitive processing from an unconscious to a conscious state. Recent research associates gamma-band oscillations with the ignition of neural cell assemblies supporting the transient feature binding necessary to activate a representation (<xref rid="pbio-0020097-Tallon-Baudry1" ref-type="bibr">Tallon-Baudry and Bertrand 1999</xref>; <xref rid="pbio-0020097-Pulvermuller1" ref-type="bibr">Pulvermüller 2001</xref>)—in this case, a phonological, lexical, or semantic representation corresponding to the solution word and its associations to the problem words. According to this hypothesis, greater synchronous gamma-band activity for insight than for noninsight solutions could reflect a more integrated or unitized solution representation. Furthermore, synchronous gamma-band activity has been hypothesized to play a critical role in the accessibility to consciousness of such representations (<xref rid="pbio-0020097-Engel1" ref-type="bibr">Engel and Singer 2001</xref>). The timing (with respect to the solution button press) of the insight gamma-band effect closely approximates estimates derived from cognitive behavioral studies of the amount of time required to access an available solution and generate a two-alternative, forced-choice button-press response (e.g., <xref rid="pbio-0020097-Kounios2" ref-type="bibr">Kounios et al. 1987</xref>; <xref rid="pbio-0020097-Meyer2" ref-type="bibr">Meyer et al. 1988</xref>; <xref rid="pbio-0020097-Smith1" ref-type="bibr">Smith and Kounios 1996</xref>). The present experiments had no response choice (i.e., always the same bimanual button press for solutions), so subjects could easily have responded 0.3 s after solving the problems. Thus, we infer that the observed gamma burst reflects the sudden conscious availability of a solution word resulting from an insight.</p><p>Suddenly recognizing new connections between problem elements is a hallmark of insight, but it is only one component of a large cortical network necessary for solving problems with insight, and recognizing new connections likely contributes to other tasks, such as understanding metaphors (<xref rid="pbio-0020097-Bottini1" ref-type="bibr">Bottini et al. 1994</xref>) and deriving a story theme (<xref rid="pbio-0020097-St1" ref-type="bibr">St. George et al. 1999</xref>). Similar tasks may depend on related cortical networks. For example, appreciating semantic jokes (<xref rid="pbio-0020097-Goel1" ref-type="bibr">Goel and Dolan 2001</xref>) and engaging in deductive reasoning that sometimes involves insight (<xref rid="pbio-0020097-Parsons1" ref-type="bibr">Parsons and Osherson 2001</xref>) both increase activity in RH posterior MTG. It is striking that the insight effect observed in the RH in our experiments occurred when people solved verbal problems, which traditional views suggest should involve mostly LH processing with little or no contribution from the RH. It is possible that insight solutions to nonverbal problems would require different cortical networks. However, the observed effect cannot be due simply to verbal retrieval, which must occur for both insight and noninsight solutions; it could be due to a type of verbal retrieval specific to insight solutions, but not involved in noninsight solutions.</p><p>We turn now to another result from the EEG time-frequency analysis, which was not predicted but nevertheless suggests a provocative interpretation. The gamma burst thought to reflect the transition of the insight solution from an unconscious to a conscious state was preceded by insight-specific activity in the alpha band (8–13 Hz). Specifically, there was a burst of alpha power (estimated at 9.8 Hz) associated with insight solutions detected over right posterior parietal cortex from approximately 1.4 s until approximately 0.4 s before the solution response, at which point insight alpha power decreased to the level of noninsight alpha power, or below (<xref ref-type="fig" rid="pbio-0020097-g005">Figure 5</xref>). An ANOVA was performed on log-transformed alpha-band (9.8 Hz) EEG power at left and right parietal-occipital electrode sites (PO7 and PO8, respectively) for insight and noninsight trials using three time windows: −2.06 to −1.56 s, −1.31 to −0.56 s, and −0.31 to 0.06 s (measured from the solution button press). This analysis yielded a significant insight × time window interaction (<italic>F</italic>[2,36] = 4.13, <italic>p</italic> = 0.027, with the Huynh-Feldt correction). Follow-up <italic>t</italic>-tests in each time window yielded significant effects of insight in the first time window at both electrode sites (PO7: <italic>t</italic>[18] = 2.32, <italic>p</italic> = 0.033; PO8: <italic>t</italic>[18] = 2.42, <italic>p</italic> = 0.026) and in the second time window only at the RH site (PO8: <italic>t</italic>[18] = 2.17, <italic>p</italic> = 0.043), with a reversal of the direction of the effect. The third time window yielded no significant effects.</p><fig id="pbio-0020097-g005" position="float"><label>Figure 5</label><caption><title>Alpha-Band Power for Insight and Noninsight Solutions</title><p>(Same conventions as in <xref ref-type="fig" rid="pbio-0020097-g004">Figure 4</xref>). (A) Time course of EEG power at 9.8 Hz (in v<sup>2</sup>) at right parietal-occipital electrode (PO8). The <italic>x</italic>-axis represents time (in seconds), with the green horizontal bars above the <italic>x</italic>-axis representing the time intervals used in the statistical analyses and topographic maps. The yellow arrow and <italic>R</italic> (at 0.0 s) signify the time of the button-press response.</p><p>(B) Time-frequency plots of the insight minus noninsight difference shown in (A).</p><p>(C) Insight minus noninsight alpha-band differences plotted as topographic maps of scalp current density (in v<sup>2</sup>/m<sup>2</sup>). Note that alpha-band power is significantly greater for insight solutions than noninsight solutions during the −1.31 to −0.56 s interval, but not during the preceding (−2.06 to −1.56 s) or subsequent (−0.31 to +0.06 s) intervals. This alpha burst was embedded in a slow decrease in alpha (see [A]), probably reflecting a general increase in cortical activity as effort increases during the course of problem solving.</p></caption><graphic xlink:href="pbio.0020097.g005"/></fig><p>Alpha rhythms are understood to reflect idling or inhibition of cortical areas (<xref rid="pbio-0020097-Pfurtscheller1" ref-type="bibr">Pfurtscheller et al. 1996</xref>). Increased alpha power measured over parietal-occipital cortex indicates idling or inhibition of visual cortex. This has been attributed to gating of visual information flowing into the perceptual system in order to protect fragile or resource-intensive processes from interference from bottom-up stimulation (<xref rid="pbio-0020097-Ray1" ref-type="bibr">Ray and Cole 1985</xref>; <xref rid="pbio-0020097-Worden1" ref-type="bibr">Worden et al. 2001</xref>; <xref rid="pbio-0020097-Jensen1" ref-type="bibr">Jensen et al. 2002</xref>; <xref rid="pbio-0020097-Cooper1" ref-type="bibr">Cooper et al. 2003</xref>; <xref rid="pbio-0020097-Ward1" ref-type="bibr">Ward 2003</xref>). This interpretation assumes that brain areas are normally highly interactive, and that allowing one process to proceed relatively independently requires active attenuation of this interaction. For instance, when subjects attend to visual space in the hemifield projecting to one hemisphere, posterior alpha increases over the other hemisphere, which receives inputs from the unattended hemifield (<xref rid="pbio-0020097-Worden1" ref-type="bibr">Worden et al. 2001</xref>). Analogously, the present results suggest selective gating of visual inputs to the RH during the interval preceding the insight-related right temporal gamma burst (<xref ref-type="fig" rid="pbio-0020097-g006">Figure 6</xref>). Hypothetically, this allows weaker processing about more distant associations between the problem words and potential solutions to gain strength, by attenuating bottom-up activation or other neural activity not related to solution that would decrease the signal-to-noise ratio for the actual solution.</p><fig id="pbio-0020097-g006" position="float"><label>Figure 6</label><caption><title>The Time Course of the Insight Effect</title><p>Alpha power (9.8 Hz at right parietal-occipital electrode PO8) and gamma power (39 Hz at right temporal electrode T8) for the insight effect (i.e., correct insight solutions minus correct noninsight solutions, in v<sup>2</sup>). The left <italic>y</italic>-axis shows the magnitude of the alpha insight effect (purple line); the right <italic>y</italic>-axis applies to the gamma insight effect (green line). The <italic>x</italic>-axis represents time (in seconds). The yellow arrow and <italic>R</italic> (at 0.0 s) signify the time of the button-press response. Note the transient enhancement of alpha on insight trials (relative to noninsight trials) prior to the gamma burst.</p></caption><graphic xlink:href="pbio.0020097.g006"/></fig><p>This interpretation of the early insight-specific alpha effect is consistent with previous behavioral research suggesting that, prior to an insight, the solution to a verbal problem can be weakly activated (<xref rid="pbio-0020097-Bowers1" ref-type="bibr">Bowers et al. 1990</xref>), especially in the RH (Bowden and <xref rid="pbio-0020097-Beeman2" ref-type="bibr">Beeman 1998</xref>; <xref rid="pbio-0020097-Bowden2" ref-type="bibr">Bowden and Jung-Beeman 2003a</xref>). Thus insight solutions may be associated with early unconscious solution-related processing, followed by a sudden transition to full awareness of the solution. We suggest that, in <xref ref-type="sec" rid="s2">Experiment 2</xref>, the early posterior alpha insight effect is an indirect correlate of the former, and the right temporal gamma effect is a direct correlate of the latter.</p><p>In sum, when people solve problems with insight, leading to an “Aha!” experience, their solutions are accompanied by a striking increase in neural activity in RH aSTG. Thus, within the network of cortical areas required for problem solving, different components are engaged or emphasized when solving with versus without insight. We propose that the RH aSTG facilitates integration of information across distant lexical or semantic relations, allowing solvers to see connections that had previously eluded them. In the two millennia since Archimedes shouted “Eureka!,” it has seemed common knowledge that people sometimes solve problems—whether great scientific questions or trivial puzzles—by a seemingly distinct mechanism called insight. This mechanism involves suddenly seeing a problem in a new light, often without awareness of how that new light was switched on. We have demonstrated that insight solutions are indeed associated with a discrete, distinct pattern of neural activity, supporting unique cognitive processes.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><sec id="s4a1"><title>Subjects</title><p>Ten men and eight women were paid to participate in <xref ref-type="sec" rid="s2">Experiment 1</xref>; 19 new subjects (nine men, ten women) were paid to participate in <xref ref-type="sec" rid="s2">Experiment 2</xref>. All were young (18–29) neurologically intact, right-handed, native English speakers; <xref ref-type="sec" rid="s2">Experiment 1</xref> participants met safety criteria for FMRI scanning. After hearing about all methods and risks and performing practice trials, they consented to participate. In <xref ref-type="sec" rid="s2">Experiment 1</xref>, data from four men and one woman were excluded due to poor FMRI signal or because subjects provided fewer than ten insight or noninsight responses. This research was approved by the University of Pennsylvania Institutional Review Board.</p></sec><sec id="s4a2"><title>Behavioral paradigm</title><p>Following practice, subjects attempted 124 compound remote associate problems during FMRI scanning. These problems (<xref rid="pbio-0020097-Bowden3" ref-type="bibr">Bowden and Jung-Beeman 2003</xref>b) can be solved quickly and evoke an “Aha!” experience, producing a distinct behavioral signature (<xref rid="pbio-0020097-Bowden2" ref-type="bibr">Bowden and Jung-Beeman 2003</xref>a), roughly half the time they are solved. <xref ref-type="fig" rid="pbio-0020097-g001">Figure 1</xref> illustrates the sequence of events for each trial. Each trial began with the task label “Compound” presented on liquid crystal diode goggles for 0.5 to 2.5 s. A gating signal from the scanner triggered the central presentation of three problem words, which persisted until subjects solved the problem or 30 s elapsed. If subjects solved the problem, they made a bimanual button press, after which the word “Solution?” prompted them to verbalize their solution. After 2 s the word “Insight?” prompted subjects to press buttons indicating whether they solved the problem with insight.</p><p>Prior to the experiment subjects were told the following: “A feeling of insight is a kind of ‘Aha!' characterized by suddenness and obviousness. You may not be sure how you came up with the answer, but are relatively confident that it is correct without having to mentally check it. It is as though the answer came into mind all at once—when you first thought of the word, you simply knew it was the answer. This feeling does not have to be overwhelming, but should resemble what was just described.” The experimenter interacted with subjects until this description was clear. This subjective rating could be used differently across subjects (or even across trials), blurring condition boundaries; yet the distinct neural correlates of insight observed across the group demonstrate that there was some consistency.</p><p>If subjects failed to solve problems within 30 s, the “Solution?” prompt appeared, and subjects pressed the “no” buttons and verbalized “Don't Know.” Then the “Insight?” prompt appeared, and subjects pressed the “no” buttons again. After the insight rating, subjects performed three line-matching trials (3 s each) to distract them from thinking about the problems, allowing the critical BOLD signal to return to baseline (<xref rid="pbio-0020097-Binder1" ref-type="bibr">Binder et al. 1999</xref>). The total time from the end of one problem to the onset of the next was 14.5–16.5 s. The condition (e.g., insight or noninsight solution) and time of events was determined by subjects' responses.</p></sec><sec id="s4a3"><title>Image acquisition</title><p>Imaging was performed at the Hospital of the University of Pennsylvania, on a 1.5 Tesla GE SIGNA scanner with a fast gradient system for echo-planar imaging and a standard head coil. Head motion was restricted with plastic braces and foam padding. Anatomical high-resolution T1-weighted axial and sagittal images were acquired while subjects performed practice trials. Functional images (21 slices, 5 mm thick; 3.75-mm × 3.75-mm in-plane resolution; TR = 2000 ms for 21 slices; time to echo = 40 ms) were acquired in the same axial plane as the anatomical images using gradient-echo echo-planar sequences sensitive to BOLD signal (<xref rid="pbio-0020097-Kwong1" ref-type="bibr">Kwong et al. 1992</xref>; <xref rid="pbio-0020097-Ogawa1" ref-type="bibr">Ogawa et al. 1992</xref>). Each functional run was preceded by a 20-s saturation period. Subjects participated in four 15-min runs and a fifth run of varying length, depending on the number of remaining problems.</p></sec><sec id="s4a4"><title>Image analysis</title><p>Images were coregistered through time with a three-dimensional registration algorithm (<xref rid="pbio-0020097-Cox1" ref-type="bibr">Cox 1996</xref>). Echo planar imaging volumes were spatially smoothed using a 7.5-mm full-width half-maximum Gaussian kernel. Within each run, voxels were eliminated if the signal magnitude changed more than 10% across successive TRs, or if the mean signal level was below a noise threshold. Functional data were transformed (<xref rid="pbio-0020097-Collins1" ref-type="bibr">Collins et al. 1994</xref>) to a standard stereotaxic atlas (<xref rid="pbio-0020097-Talairach1" ref-type="bibr">Talairach and Tournoux 1988</xref>) with a voxel size of 2.5 mm<sup>3</sup>.</p><p>Data were analyzed using general linear model analysis that extracted average responses to each trial type, correcting for linear drift and removing signal changes correlated with head motion. Each TR was divided into two 1-s images to improve time locking of the solving event and the functional image data (time-course data were temporally smoothed in Figures <xref ref-type="fig" rid="pbio-0020097-g002">2</xref> and <xref ref-type="fig" rid="pbio-0020097-g003">3</xref>). Solution-related responses were calculated using the average signal change within the window 4–9 s (to account for hemodynamic delay) after the solving event (beginning about 2 s prior to the button press). Differences between insight and noninsight solution events were estimated for each participant, then combined in a second-stage random effects analysis to identify differences consistent across all subjects. A cluster threshold was set at regions at least 500 mm<sup>3</sup> in volume (32 normalized voxels, or 7.1 original-sized voxels) in which each voxel was reliably different across subjects, (<italic>t</italic>[12] > 3.43, p < 0.005 uncorrected). Monte Carlo simulations with similar datasets reveal low false positive rates with these criteria. RH aSTG was the only cluster to exceed these criteria, and converging evidence and the a priori prediction about RH aSTG strengthen confidence in this result.</p></sec><sec id="s4a5"><title>Experiment 2</title><p>Behavioral procedures were similar to those of <xref ref-type="sec" rid="s2">Experiment 1</xref>, except that (A) problem words were presented at smaller visual angles to discourage eye movements, (B) there were 2-s delays between each event in the response sequence, and (C) subjects triggered a new problem directly after responding to the previous problem (i.e., no line task occurred between problems).</p></sec><sec id="s4a6"><title>EEG methods</title><p>Continuous high-density EEGs were recorded at 250 Hz (bandpass: 0.2–100 Hz) from 128 tin electrodes embedded in an elastic cap (linked mastoid reference with forehead ground) placed according to the extended International 10–20 System. Prior to data analysis, EEG channels with excessive noise were replaced with interpolated data from neighboring channels. Eyeblink artifacts were removed from the EEG with an adaptive filter separately constructed for each subject using EMSE 5.0 (Source Signal Imaging Inc., San Diego, California, United States). Induced oscillations were analyzed by segmenting each subject's continuous EEG into 4-s segments beginning 3 s before each solution response. (An analysis epoch beginning at an earlier point in time would have resulted in the loss of trials associated with response times of less than 3 s.)</p><p>Time-frequency transforms (performed with EMSE 5.0) were obtained by the application of complex-valued Grossmann-Morlet wavelets, which are Gaussian in both time and frequency. Following <xref rid="pbio-0020097-Torrence1" ref-type="bibr">Torrence and Campo (1998</xref>), the mother wavelet, ω<sub>0</sub>, in the time domain has the form</p><p> <disp-formula id="pbio-0020097-e001"><graphic xlink:href="pbio.0020097.e001.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where ω<sub>0</sub> is a nondimensional frequency. In this case, ω<sub>0</sub> is chosen to be 5.336, so that ∫ϕ<sub>0</sub>(<italic>t</italic>) ≅ 0. The constant π−¼ is a normalization factor such that ∫(ϕ<sub>0</sub>(<italic>t</italic>))<sup>2</sup> = 1. For the discrete time case, a family of wavelets may be obtained as</p><p> <disp-formula id="pbio-0020097-e002"><graphic xlink:href="pbio.0020097.e002.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where δ<italic>t</italic> is the sample period (in seconds), <italic>s</italic> is the scale (in seconds), and <italic>n</italic> is an integer that counts the number of samples from the starting time. The Fourier wavelength λ is given by</p><p> <disp-formula id="pbio-0020097-e003"><graphic xlink:href="pbio.0020097.e003.jpg" mimetype="image" position="float"/></disp-formula> </p><p>In the frequency domain, the (continuous) Fourier transform of <xref ref-type="disp-formula" rid="pbio-0020097-e002">Equation 2</xref> is</p><p> <disp-formula id="pbio-0020097-e401"><graphic xlink:href="pbio.0020097.e401.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where</p><p> <disp-formula id="pbio-0020097-e402"><graphic xlink:href="pbio.0020097.e402.jpg" mimetype="image" position="float"/></disp-formula> </p><p>One reasonable way to measure the “resolution” of the wavelet transform is to consider the dispersion of the wavelets in both time and frequency. Since the wavelets are Gaussian in both domains, the <italic>e</italic>-folding time and frequency may serve as quantitative measures of dispersion. Note that these dispersions are a function of the scale, <italic>s</italic>. For a selected frequency, 𝒻<sub><italic>c</italic></sub> = 1/λ, or from <xref ref-type="disp-formula" rid="pbio-0020097-e003">Equation 3</xref> </p><p> <disp-formula id="pbio-0020097-e005"><graphic xlink:href="pbio.0020097.e005.jpg" mimetype="image" position="float"/></disp-formula> </p><p>Then substituting into <xref ref-type="disp-formula" rid="pbio-0020097-e002">Equation 2</xref>, we find that the <italic>e</italic>-folding time is <inline-formula id="pbio-0020097-e006"><inline-graphic xlink:href="pbio.0020097.e006.jpg" mimetype="image"/></inline-formula> for frequency 𝒻<sub><italic>c</italic></sub>. From <xref ref-type="disp-formula" rid="pbio-0020097-e002">Equation 2</xref>, the <italic>e</italic>-folding frequency is <inline-formula id="pbio-0020097-e007"><inline-graphic xlink:href="pbio.0020097.e007.jpg" mimetype="image"/></inline-formula>. To make this concrete, we find that for a 10-Hz (alpha-band) center frequency, the <italic>e</italic>-folding time is 0.12 s and the <italic>e</italic>-folding frequency is 2.6 Hz. For a 40-Hz ( gamma-band) center frequency, the <italic>e</italic>-folding time is 0.03 s and the <italic>e</italic>-folding frequency is 10.5 Hz. Note that these <italic>e</italic>-folding parameters imply that wavelet scaling preserves the joint time-frequency resolution (equal areas in time-frequency space), with higher temporal resolution but broader frequency resolution as the wavelet scale decreases. </p><p>Segments corresponding to trials for which individual subjects produced the correct response were isolated and averaged separately according to whether or not the subject reported the experience of insight. Planned statistical tests (repeated-measure ANOVAs) were performed in order to detect insight-related effects on frontal midline theta (5–8 Hz), posterior alpha (8–13 Hz), fronto-central beta (13–20 Hz), and left and right temporal gamma (20–50 Hz). Response-locked event-related potentials (ERPs) were also computed using the same analysis epoch. Standard ERP analyses yielded no evidence of statistically significant effects, likely because ERPs reflect phase-locked activity rather than the induced (i.e., nonphase-locked) activity examined in the wavelet analyses; due to the long response times evident in this experiment, phase locking resulting from problem presentation would not be expected.</p><p>EEG effects were topographically mapped by employing spline-based Laplacian mapping with an FMRI-derived realistic head model and digitized electrode positions. Localization of EEG/ERP signals is a form of probabilistic modelling rather than direct neuroimaging. In contrast to other techniques, source estimation by Laplacian mapping indicates the presence of superficial foci of neuroelectric activity with minimal assumptions.</p></sec></sec></sec><sec sec-type="supplementary-material" id="s5"><title>Supporting Information</title><supplementary-material content-type="local-data" id="sg001"><label>Figure S1</label><caption><title>Cortical Regions Showing “Insight Effects” Below Cluster Size Threshold</title><p>The far left lane shows for each region a single slice best depicting the cluster activated above threshold; middle lane shows time course of signal following insight (red line) and noninsight (blue line) solutions, across the entire active cluster; right panel shows the “insight effect” (insight signal minus noninsight signal, error bars show the standard error of the mean of the difference at each timepoint).</p><p>(A) depicts bilateral IFG with lowered threshold (<italic>t</italic>[12] = 2.83, <italic>p</italic> < 0.015); (B–D) depict clusters of FMRI signal at the same <italic>t</italic>-threshold used in the main paper (<italic>t</italic>[12] = 3.43, <italic>p</italic> < 0.005), but the clusters are too small to surpass cluster criterion.</p><p>(B) LH medial frontal gyrus;</p><p>(C) LH PC gyrus;</p><p>(D) LH amygdala (there was also a small cluster near RH amygdala). Spatial coordinates and other are details listed in <xref ref-type="table" rid="pbio-0020097-t001">Table 1</xref>.</p><p>(914 KB PDF).</p></caption><media xlink:href="pbio.0020097.sg001.pdf"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec>
Avoiding URL Reference Degradation in Scientific Publications	Could not extract abstract	<contrib contrib-type="author"><name><surname>Kelly</surname><given-names>Desiree P</given-names></name></contrib><contrib contrib-type="author"><name><surname>Hester</surname><given-names>Eric J</given-names></name></contrib><contrib contrib-type="author"><name><surname>Johnson</surname><given-names>Kathryn R</given-names></name></contrib><contrib contrib-type="author"><name><surname>Heilig</surname><given-names>Lauren F</given-names></name></contrib><contrib contrib-type="author"><name><surname>Drake</surname><given-names>Amanda L</given-names></name></contrib><contrib contrib-type="author"><name><surname>Schilling</surname><given-names>Lisa M</given-names></name></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Dellavalle</surname><given-names>Robert P</given-names></name><email>[email protected]</email></contrib>	PLoS Biology	<p>While we applaud PloS' use of Digital Object Identifiers (“The What and Whys of DOIs,” PLoS Biol 1: e57 doi: <ext-link ext-link-type="doi" xlink:href="10.1371/journal.pbio.0000057">10.1371/journal.pbio.0000057</ext-link>), we also note the lack of provisions in your instructions for authors for preserving access to electronic information residing at a cited Internet addresses via Uniform Resource Locators (URLs). Medical and scientific literature increasingly cites information only found on the Internet. However, URLs may become inaccessible shortly after article publication.</p><p>Please consider requiring PLoS' authors to (1) submit all cited URLs to the Internet Archive (<ext-link ext-link-type="uri" xlink:href="www.archive.org">www.archive.org</ext-link>), a nonprofit organization that has been preserving electronic content since 1996, and (2) maintain a printed copy of the electronic information for future communication until the URL becomes available at the Internet Archive (about a six-month lag time).</p><p>The Internet Archive, the largest digital library of Internet sites and other digital data, stores cited Internet information at no cost to the author, reader, or publisher. By requiring PLoS' authors to submit all cited Internet-based information to the Internet Archive, PLoS will better preserve the integrity of its content for the future.</p><sec id="s2"><title>PLoS' Response</title><p>Ms. Kelly and colleagues raise an important issue about the ephemeral nature of many information sources on the Internet. In the case of online scholarly literature, information is more likely to be archived and able to be found—indeed, an open-access article is one in which, according to the Bethesda Definition, “A complete version of the work and all supplemental materials, including a copy of the permission…, in a suitable standard electronic format is deposited immediately upon initial publication in at least one online repository that is supported by an academic institution, scholarly society, government agency, or other well-established organization that seeks to enable open access, unrestricted distribution, interoperability, and long-term archiving (for the biomedical sciences, PubMed Central is such a repository).”</p><p>Other types of Internet-based information are more likely to change, move, or be removed. We agree that wherever possible we must find a way to preserve the relevant information from the sources cited in our articles.</p><p>PLoS has always encouraged authors to submit supporting information for their research articles, including raw datasets, spreadsheets, multimedia, and snapshots of Web-based interactive tools. PLoS makes this supporting information available to everyone for download and use. PLoS also requires authors to deposit all appropriate datasets, images, and information in public databases and to list the relevant accession and version numbers in the article.</p><p>The question, then, is how PLoS and its authors can preserve access to other Internet-based information, including organizational Web sites, articles in the popular media, or interactive databases.</p><p>Although submitting cited URLs to the Internet Archive is worthwhile, it is still (unfortunately) far from ideal. The Internet Archive is best at archiving simple HTML and may be the most appropriate place to archive a Web site an author has cited for its static information content. The Internet Archive does not, however, archive content with password restrictions or “crawling” restrictions, and it allows the removal of already archived content at the request of Web administrators; it would therefore not be an effective archive, for example, for popular press articles that have restricted access. In addition, the Internet Archive cannot preserve functions that interact with the originating server, so it is not an appropriate way to archive a Web site an author has cited, for example, for its useful interactive tools. Finally, there is currently no automated way for publishers to redirect links from the original address to the address on the Internet Archive.</p><p>For the time being, PLoS plans to review all electronic citations on a case-by-case basis and, when appropriate, request that authors submit the cited Web site URL to the Internet Archive and additionally submit a digital copy of the information to PLoS for internal archiving.</p><p>We would also like to encourage further input on this issue from the scientific and medical community and urge them to support the Internet Archive and other organizations working to preserve the digital record for future generations.</p></sec>
The Cytoskeleton In Vivo	Could not extract abstract	<contrib contrib-type="author"><name><surname>Fernández</surname><given-names>Beatriz García</given-names></name></contrib>	PLoS Biology	<p>As a student I always marvelled at the sight of single cells in culture moving over artificial surfaces and exhibiting membrane ruffles and protrusions. However, while I found cultured cells fascinating I always wondered how cells are able to move and regulate their shape in the context of a whole organism where so many space constraints exist and where all cellular processes have to be tightly regulated. Some answers to my questions began to emerge in a paper written by <xref rid="pbio-0020100-Baum1" ref-type="bibr">Baum and Perrimon (2001)</xref>, in which the authors showed the expression and regulation of the actin cytoskeleton and of actin binding proteins in a real epithelium.</p><p>The cytoskeleton is a meshwork of protein polymers extending throughout the cytoplasm. It not only provides structural support for the cell but also plays a central role in a range of dynamic processes from signalling to endocytosis and intracellular trafficking. A particularly clear example of this is the use of actin cytoskeleton as a “wool” for knitting multiple dynamic structures such as lamellae, filopodia, and stress fibres. These structures determine cell shape and also produce the driving force accompanying many types of cellular movements including muscle contraction and cell division. We know many details about some of the proteins that modulate the dynamics of actin in these structures. However, most of them have been found biochemically and their function has been elucidated primarily using in vitro and cell culture assays of actin assembly. What about these proteins in the context of a developing organism? How do cells generate a spatially and temporally ordered network of actin filaments represented at the tissue level? To answer these questions, we need to move to experimentally accessible multicellular organisms, such as <named-content content-type="genus-species">Drosophila</named-content>, which offers virtually unlimited possibilities as a model system for the genetic and molecular analysis of biological processes.</p><p> <xref rid="pbio-0020100-Baum1" ref-type="bibr">Baum and Perrimon (2001)</xref> analyzed the function of a number of proteins involved in actin dynamics within the context of a developing epithelium—the follicle cells that surround the germ line cyst during <named-content content-type="genus-species">Drosophila</named-content> oogenesis. These cells have a simple polarised arrangement of actin filaments, which provides a useful system to study the spatial organisation of the actin cytoskeleton.</p><p>Taking advantage of the ability to generate clones of cells lacking specific proteins, the authors identified new functional roles for actin regulators such as CAP (a <named-content content-type="genus-species">Drosophila</named-content> homologue of adenylyl cyclase-associated proteins), Enabled (Ena) and Abelson (Abl). These proteins had been well characterized in cell culture and in vitro studies, but little was known about their function in a developing organism. Clones of cells lacking CAP (<xref ref-type="fig" rid="pbio-0020100-g001">Figure 1</xref>), a protein known to inhibit actin polymerisation, maintained their epithelial polarity but had higher levels of actin and defects in the apical actin organisation. This result indicates that the inhibitory activity of CAP is restricted to one side of the cells, thus demonstrating that actin dynamics can be independently modified at opposite poles of an epithelium. Ena, a member of the Ena/VASP family proteins that catalyse filament formation, and Abl, a protein kinase that binds CAP in mammalian cells, were found to work with CAP in this process. The authors proposed that CAP, Ena, and Abl regulate the level and spatial organization of actin in the follicle cells.</p><fig id="pbio-0020100-g001" position="float"><label>Figure 1</label><caption><title>CAP Mutant Clones</title><p>Follicle cells lacking CAP accumulate actin (red) in their apical region. Ena (blue in the bottom panel), also accumulates apically in the mutant cells (looks pink in the clone of cells due to overlap with F-actin in red). The mutant cell clones are identified by the absence of GFP (green in the top panel). Using this technique the cytoskeleton of mutant cells can be analysed in the context of a wild type epithelium. (Image kindly supplied by Buzz Baum.)</p></caption><graphic xlink:href="pbio.0020100.g001"/></fig><p>In contrast to the spatially restricted functions of CAP, Ena, and Abl, profilin and cofilin were shown to regulate actin filament formation throughout the cell cortex, a more global function that matches the results obtained in cell culture experiments. In summary, this study showed how proteins can organise actin in space and began to highlight some of the differences and similarities between cells in culture and in vivo. The functions revealed in the follicular epithelium were consistent with the roles previously shown in mammalian systems, but the experiments on intact tissue began to reveal a spatial and temporal functional dimension that could not have been observed in cell culture.</p><p>These experiments could be expanded to large-scale screens (<xref rid="pbio-0020100-StJohnston1" ref-type="bibr">St Johnston 2002</xref>), but this would be time consuming and could encounter the problem that some genes will be cell lethal, preventing the analysis of their function in actin dynamics. However, two more recent reports (<xref rid="pbio-0020100-Kiger1" ref-type="bibr">Kiger et al. 2003</xref>; <xref rid="pbio-0020100-Rogers1" ref-type="bibr">Rogers et al. 2003</xref>) describe a complementary and exhaustive search for regulators of cytoskeletal dynamics by taking advantage of genomic resources and the powerful RNA interference (RNAi) technique (<xref rid="pbio-0020100-Hutvagner1" ref-type="bibr">Hutvágner and Zamore 2002</xref>). RNAi allows individual genes to be knocked out in a simple and controlled fashion.</p><p> <xref rid="pbio-0020100-Kiger1" ref-type="bibr">Kiger et al. (2003)</xref> used RNAi in two different cell lines of <named-content content-type="genus-species">Drosophila</named-content> to screen a number of genes involved in signalling and cytoskeletal dynamics. They targeted 994 genes, of which 160 produced phenotypes in the experiment. The range of phenotypes varied from specific defects in the actin and tubulin cytoskeleton to others affecting cell cycle progression, cytokinesis, and cell shape. They also showed that only about 40% of the genes had similar loss-of-function phenotypes in both cell lines. This alone indicates an important limitation of many tissue culture experiments, since the same protein can have different effects depending on the cell type. Another valuable element of this work is that clustering of genes with similar phenotypes leads to the identification of pathways and networks of genes that are involved in cytoskeletal function.</p><p> <xref rid="pbio-0020100-Rogers1" ref-type="bibr">Rogers et al. (2003)</xref>, using only one <named-content content-type="genus-species">Drosophila</named-content> cell line, studied the effects of proteins involved in the formation of lamellae. The authors looked at the effects of loss of function in 90 genes known to be involved in actin dynamics and the formation and activity of the lamella. As well as confirming the function of many proteins already known to play a role in this process, this analysis allowed them to find interactions between genes and to build genetic pathways.</p><p>Together these two studies reveal that RNAi screens in tissue culture can be a powerful tool for finding new functions of known and uncharacterized genes, and new relationships between genes. However, this is only the beginning, and the genes identified in this manner will have to be tested in vivo, in systems like that of Baum and Perrimon, where specific functions can be assessed in time and space within the confines of real organisms. The focus must be to understand how all these molecular events and regulation cascades operate in individual cells to contribute to the generation of changes in a whole individual. Increasingly, the attention of developmental biologists is being drawn from genes and their products towards cells (<xref rid="pbio-0020100-Kaltschmidt1" ref-type="bibr">Kaltschmidt and Martinez Arias 2002</xref>). The future, it seems to me, lies in the combination of in vitro systems, cell culture, and in vivo studies. I hope to apply this view in my analysis of the process of dorsal closure in <named-content content-type="genus-species">Drosophila</named-content> embryos, as an example of how signalling pathways coordinate and regulate the activity of the cytoskeleton in the generation of shape and morphogenetic movements (<xref rid="pbio-0020100-Jacinto1" ref-type="bibr">Jacinto et al. 2002</xref>).</p>
Peace Lessons from an Unlikely Source	Could not extract abstract	<contrib contrib-type="author"><name><surname>de Waal</surname><given-names>Frans B. M</given-names></name></contrib>	PLoS Biology	<p>Upon arrival from Europe, now more than two decades ago, I was taken aback by the level of violence in the American media. I do not just mean the daily news, even though it is hard getting used to multiple murders per day in any large city. No, I mean sitcoms, comedies, drama series, and movies. Staying away from Schwarzenegger and Stallone does not do it; almost any American movie features violence. Inevitably, desensitization sets in. If you say, for example, that <italic>Dances with Wolves</italic> (the 1990 movie with Kevin Costner) is violent, people look at you as if you are crazy. They see an idyllic, sentimental movie, with beautiful landscapes, showing a rare white man who respects American Indians. The bloody scenes barely register.</p><p>Comedy is no different. I love, for example, <italic>Saturday Night Live</italic> for its inside commentary on peculiarly American phenomena, such as cheerleaders, televangelists, and celebrity lawyers. But <italic>SNL</italic> is incomplete without at least one sketch in which someone's car explodes or head gets blown off. Characters such as Hans and Franz (“We're going to pump you up!”) appeal to me for their names alone (and yes, I do have a brother named Hans), but when their free weights are so heavy that their arms get torn off, I am baffled. The spouting blood gets a big laugh from the audience, but I fail to see the humor.</p><p>Did I grow up in a land of sissies? Perhaps, but I am not mentioning this to decide whether violence in the media and our ability to grow immune to it—as I also have over the years—is desirable, or not. I simply wish to draw attention to the cultural fissures in how violence is portrayed, how we teach conflict resolution, and whether harmony is valued over competitiveness. This is the problem with the human species. Somewhere in all of this resides a human nature, but it is molded and stretched into so many different directions that it is hard to say if we are naturally competitive or naturally community-builders. In fact, we are both, but each society reaches its own balance between the two. In America, the squeaky wheel gets the grease. In Japan, the nail that stands out gets pounded into the ground.</p><p>Does this variability mean, as some have argued, that animal studies cannot possibly shed light on human aggression? “Nature, red in tooth and claw” remains the dominant image of the animal world. Animals just fight, and that is it? It is not that simple. First, each species has its own way of handling conflict, with for example the chimpanzee (<named-content content-type="genus-species">Pan troglodytes</named-content>) being far more violent than that equally close relative of ours, the bonobo (<named-content content-type="genus-species">P. paniscus</named-content>) (<xref rid="pbio-0020101-deWaal1" ref-type="bibr">de Waal 1997</xref>). But also within each species we find, just as in humans, variation from group to group. There are “cultures” of violence and “cultures” of peace. The latter are made possible by the universal primate ability to settle disputes and iron out differences.</p><p>There was a time when no review of human nature would be complete without assertions about our inborn aggressiveness. The first scientist to bring up this issue, not coincidentally after World War II, was <xref rid="pbio-0020101-Lorenz1" ref-type="bibr">Konrad Lorenz (1966)</xref>. Lorenz's thesis was greeted with accusations about attempts to whitewash human atrocities, all the more so given the Nobel Prize winner's native tongue, which was German. But Lorenz was hardly alone. In the USA, science journalist <xref rid="pbio-0020101-Ardrey1" ref-type="bibr">Robert Ardrey (1961)</xref> presented us as “killer apes” unlikely to ever get our nasty side under control. Recent world events have done little to counter this pessimistic outlook.</p><p>The opposition argued, of course, that aggression, like all human behavior, is subject to powerful cultural influences. They even signed petitions to this effect, such as the controversial <italic>Seville Statement on Violence</italic> (<xref rid="pbio-0020101-Adams1" ref-type="bibr">Adams et al. 1990</xref>). In the polarized mind-set of the time, the issue was presented in either-or fashion, as if behavior cannot be both learned and built upon a biological foundation. This rather fruitless nature/nurture debate becomes considerably more complex if we include what is usually left out, which is the ability to keep aggression under control and foster peace. For this ability, too, there exist animal parallels, such as the habit of chimpanzees to reconcile after fights by means of a kiss and embrace. Such reunions are well-documented in a multitude of animals, including nonprimates, such as hyenas and dolphins. They serve to restore social relationships disturbed by aggression, and any animal that depends on cooperation needs such mechanisms of social repair (<xref rid="pbio-0020101-Aureli1" ref-type="bibr">Aureli and de Waal 2000</xref>; <xref rid="pbio-0020101-deWaal2" ref-type="bibr">de Waal 2000</xref>). There are even indications that in animals, too, cultural influences matter in this regard. This may disturb those who write culture with a capital <italic>C</italic>, and hence view it as uniquely human, but it is a serious possibility nonetheless.</p><p>Nonhuman culture is currently one of the hottest areas in the study of animal behavior. The idea goes back to the pioneering work of Kinji Imanishi, who in 1952 proposed that if individuals learn from one another, their behavior may over time grow different from that of individuals in other groups of the same species, thus creating a characteristic culture (reviewed by <xref rid="pbio-0020101-deWaal3" ref-type="bibr">de Waal 2001</xref>). Imanishi thus brought the culture concept down to its most basic feature, that is, the social rather than genetic transmission of behavior. Since then, many examples have been documented, mostly concerning subsistence techniques, such as the sweet potato washing of Japanese macaques (<named-content content-type="genus-species">Macaca fuscata</named-content>) and the rich array of tool use by wild chimpanzees, orangutans (<named-content content-type="genus-species">Pongo pymaeus</named-content>), and capuchin monkeys (<named-content content-type="genus-species">Cebus</named-content> spp.) (<xref rid="pbio-0020101-Whiten1" ref-type="bibr">Whiten et al. 1999</xref>; <xref rid="pbio-0020101-deWaal3" ref-type="bibr">de Waal 2001</xref>; <xref rid="pbio-0020101-Hirata1" ref-type="bibr">Hirata et al. 2001</xref>; <xref rid="pbio-0020101-Perry1" ref-type="bibr">Perry et al. 2003</xref>; <xref rid="pbio-0020101-vanSchaik1" ref-type="bibr">van Schaik et al. 2003</xref>). However, much less attention has been paid to <italic>social culture</italic>, which we might define as the transmission of social positions, preferences, habits, and attitudes.</p><p>Social culture is obviously harder to document than tool use. In human culture, for instance, it is easy to tell if people eat with knife and fork or with chopsticks, but to notice if a culture is egalitarian or hierarchical, warm or distant, collectivistic or individualistic takes time and is difficult to capture in behavioral measures. A well-documented monkey example of social culture is the inheritance of rank positions in macaque and baboon societies. The future position in the hierarchy of a newborn female can be predicted with almost one hundred percent certainty on the basis of her mother's rank. Females with relatives in high places are born with a silver spoon in their mouth, so to speak, whereas those of lowly origin will spend their life at the bottom. Despite its stability, the system depends on learning. Early in life, the young monkey finds out against which opponents it can expect help from her mother and sisters. When sparring with peer A she may utter screams that recruit massive support to defeat A. But against peer B she can scream her lungs out and nothing happens. Consequently, she will come to dominate A but not B. Experiments manipulating the presence of family members have found that when support dwindles dominant females are unable to maintain their positions (<xref rid="pbio-0020101-Chapais1" ref-type="bibr">Chapais 1988</xref>). In other words, the kin-based hierarchy is maintained for generation after generation through social rather than genetic transmission.</p><p>Returning to the issue of aggressive behavior, here the effects of social culture can be felt as well. Without any drugs or brain lesions, one experiment managed to turn monkeys into pacifists. Juveniles of two different macaque species were placed together, day and night, for five months. Rhesus monkeys (<named-content content-type="genus-species">Macaca mulatta</named-content>), known as quarrelsome and violent, were housed with the more tolerant and easy-going stumptail monkeys (<named-content content-type="genus-species">M. arctoides</named-content>) (<xref ref-type="fig" rid="pbio-0020101-g001">Figure 1</xref>). Stumptail monkeys easily reconcile with their opponents after fights by holding each others' hips (the so-called “hold-bottom” ritual), whereas reconciliations are rare in rhesus monkeys. Because the mixed-species groups were dominated by the stumptails, physical aggression was rare. The atmosphere was relaxed, and after a while all of the monkeys became friends. Juveniles of the two species played together, groomed together, and slept in large, mixed huddles. Most importantly, the rhesus monkeys developed peacemaking skills on a par with those of their more tolerant group mates. Even when, at the end of the experiment, both species were separated, the rhesus monkeys still showed three times more reconciliation and grooming behaviors after fights than typical of their kind (<xref rid="pbio-0020101-deWaal4" ref-type="bibr">de Waal and Johanowicz 1993</xref>). Primates thus can adopt social behavior under the influence of others, which opens the door to social culture.</p><fig id="pbio-0020101-g001" position="float"><label>Figure 1</label><caption><title>Stumptail Monkeys</title><p>Stumptail monkeys (<named-content content-type="genus-species">Macaca arctoides</named-content>) are among the most conciliatory members of the genus Macaca. They are heavily built, yet remarkably friendly and tolerant, such as here: the alpha male is eating attractive food unperturbed by an entire audience around him. When stumptail monkeys were housed with a less tolerant macaque, they modified the latter species' behavior into a more pacific direction. (Photograph by Frans de Waal, used with permission.)</p></caption><graphic xlink:href="pbio.0020101.g001"/></fig><p>Not unlike rhesus monkeys, baboons have a reputation for fierce competition and nasty fights. With the study by Robert Sapolsky and Lisa Share published in this issue of <italic>PLoS Biology</italic>, we now have the first field evidence that primates can go the flower power route (<xref rid="pbio-0020101-Sapolsky2" ref-type="bibr">Sapolsky and Share 2004</xref>). Wild baboons developed an exceptionally pacific social tradition that outlasted the individuals who established it. For years, Sapolsky has documented how olive baboons (<named-content content-type="genus-species">Papio anubis</named-content>) on the plains of the Masai Mara, in Kenya, wage wars of nerves, compromising their rivals' immune systems and pushing up the level of their blood cortisol (<xref rid="pbio-0020101-Sapolsky1" ref-type="bibr">Sapolsky 1994</xref>). An accident of history, however, selectively wiped out all the male bullies of his main study troop. As a result, the number of aggressive incidents dropped dramatically. This by itself was not so surprising. It became more interesting when it was discovered that the behavioral change was maintained for a decade. Baboon males migrate after puberty, hence fresh young males enter troops all the time, resulting in a complete turn-over of males during the intervening decade. Nevertheless, compared with troops around it, the affected troop upheld its reduced aggression, increased friendly behavior, and exceptionally low stress levels. The conclusion from this natural experiment is that, like human societies, each animal society has its own ecological and behavioral history, which determines its prevalent social style.</p><p>It is somewhat ironic that at a time when researchers on human aggression are increasingly attracted, albeit with a far more sophisticated approach, to the Lorenzian idea of a biological basis of aggression (<xref rid="pbio-0020101-Enserink1" ref-type="bibr">Enserink 2000</xref>), students of animal behavior are beginning to look at its possible cultural basis. There is no reason for animals with a development as slow as a baboon (with adulthood achieved in five or six years) not to be influenced in every way by the environment in which they grow up, including the social environment. How this influence takes place is a point of much debate, and remains unclear in the case of the peaceful male baboons in the Masai Mara. Given their mobility, the males themselves are unlikely transmitters of social traditions within their natal troop. Therefore, Sapolsky and Share look at the females for an answer—female baboons stay all their lives in the same troop. By reacting positively to certain kinds of behavior, for example, females may be able to steer male attitudes in a new direction. This complex problem is hard to unravel with a single study, especially in the absence of experimentation. Yet, the main two points of this discovery are loud and clear: social behavior observed in nature may be a product of culture, and even the fiercest primates do not forever need to stay this way.</p><p>Let us hope this applies to humanity as well.</p>
A Mechanism for Adding the First Link in a Nascent Actin Filament Chain	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>The capacity for self-generated movement is a defining characteristic of animal life. With the molecular components of cellular locomotion conserved in organisms from protozoa to vertebrates, directed cell motility appears to be an ancient cell process, likely dating back a billion years. Most directed motion relies on the assembly, or polymerization, of actin proteins into filaments. Actin is one of the most abundant proteins in cells; about half of the cellular concentration of actin is bound together in filaments at any given time while the other half floats freely as “monomers” in the cytoplasm. The erection and demolition of actin filaments directs the cell motility that lays down the remarkable million miles of nerve cells that form the nervous system and drives a variety of fundamental biological processes, from effective immune response to embryonic development. Mutations in proteins that regulate actin assembly can lead to the abnormal cell migration associated with metastatic cancer. The actin cytoskeleton also provides the structural support for animal cells that the cell wall provides for plants.<xref ref-type="fig" rid="pbio-0020103-g001"/> </p><fig id="pbio-0020103-g001" position="float"><caption><title>Actin addition</title></caption><graphic xlink:href="pbio.0020103.g001"/></fig><p>The molecular mechanisms underlying actin assembly and cell motility remained obscure until 1994, when Thomas Pollard and his colleagues discovered the protein complex that initiates actin polymerization. Called actin-related protein 2/3 (Arp2/3) complex, this molecular machine consists of seven subunits, including the two actin-related proteins. Free actin monomers are primed for rapid polymerization, but polymerization must be initiated by the Arp2/3 complex in a process referred to as nucleation. To nucleate a new filament, the Arp2/3 complex must be activated, a job accomplished by a family of proteins called WASP (after Wiskott Aldrich Syndrome, a genetic disease characterized by defects in platelet development and lymphocyte function). WASP proteins bind to both the Arp2/3 complex and an actin monomer. The Arp2/3 complex also binds two molecules of adenosine triphosphate (ATP) on the Arp2 and Arp3 subunits. ATP releases energy in a process called hydrolysis, which drives most energy-dependent processes, from actin polymerization to muscle contraction. The precise mechanisms governing Arp2/3 activation and nucleation are not known. Now Mark Dayel and Dyche Mullins show where hydrolysis occurs during this crucial first step in polymerization and use this finding to investigate the mechanisms that drive nucleation.</p><p>In previous experiments, Dayel and Mullins found that Arp2/3 appears to require hydrolysable ATP to effect nucleation. To determine when and if ATP hydrolysis occurs on the Arp2/3 complex, Dayel and Mullins developed a technique that allowed them to analyze the Arp2 and Arp3 subunits separately. Dayel and Mullins discovered that hydrolysis occurs only on the Arp2 subunit of the complex and that it happens during the step when WASP initiates the nucleation of a new filament. The researchers then used ATP hydrolysis on Arp2 to dissect the mechanism by which WASP activates the Arp2/3 complex and develop a model of nucleation. (All previous techniques required actin polymerization to monitor the activity of the Arp2/3 complex, but this technique offers a way to decouple activation from polymerization.) They find that WASP proteins activate the Arp2/3 complex by coordinating its interaction with an actin monomer—the first monomer of the new filament.</p><p>By developing a novel technique to monitor activation of the Arp2/3 complex, the authors contribute a new tool for further investigations of this central part of the cellular motility machinery. And by showing how Arp2/3 is activated, they offer important insights into the workings of a multiprotein cellular machine and the mechanisms that cells enlist to control their shape and motility—which could suggest potential drug targets to inhibit the abnormal cell movement characteristic of cancer and other diseases.</p>
Who Pays for Open Access?	Could not extract abstract	<contrib contrib-type="author"><name><surname>Doyle</surname><given-names>Helen</given-names></name></contrib><contrib contrib-type="author"><name><surname>Gass</surname><given-names>Andy</given-names></name></contrib><contrib contrib-type="author"><name><surname>Kennison</surname><given-names>Rebecca</given-names></name></contrib>	PLoS Biology	<p>In the wake of declarations supporting open access to research literature from international bodies including the Organization for Economic Cooperation and Development (OECD) and the United Nations' World Summit on the Information Society (WSIS), advocates and critics of the movement appear to have agreed that the issue warrants a robust, ongoing dialogue—a development undoubtedly in the interest of the scientific community, regardless of its ultimate outcome.</p><p>To the extent that listserv messages, editorials, and conference presentations are representative of more widespread reactions to the debate, there appear to be a number of common misconceptions about what open access is and what problems it can or cannot solve. Over the next few months in <italic>PLoS Biology</italic>, we plan to explore the more pervasive of these misunderstandings, in an effort to expose the real challenges that need to be overcome and to identify some possible solutions. Here we address the first of these—the perception that the publication-charge model puts an unfair burden on authors. Subsequently, we will address concerns about the long-term economic viability of the open-access model, the integrity and quality of work published in open-access journals, and the effect that open access will have on scholarly societies.</p><sec id="s2"><title>Publication Charges—Nothing New</title><p>By charging authors a fee to have their work published in lieu of charging readers to access articles, open-access publishers such as the Public Library of Science (PLoS) and BioMed Central (BMC) have transformed the traditional publishing system. This reliance on a seemingly untested revenue stream has generated skepticism that authors will be both willing and able to pay publication charges.</p><p>Publication fees are not a phenomenon born of the open-access movement. Many authors regularly pay several thousands of dollars in page charges, color charges, correction costs, reprint costs, and other fees to their publisher, even when such costs are entirely voluntary. In the <italic>EMBO Journal</italic>, for example, authors are allowed six pages of text free, but are then charged $200 per page beyond that. A review of recent issues shows that almost all authors exceed six pages, voluntarily paying on average over $800 to publish their articles.</p><p>Furthermore, in addition to paying other publication charges, authors may be willing to pay extra for their articles to be made open access, as several publishers have recently recognized. A recent survey of authors in the <italic>Proceedings of National Academy of Science</italic> (<italic>PNAS</italic>) found that although <italic>PNAS</italic> already makes its content freely available after six months, nearly 50% of <italic>PNAS</italic> authors expressed a willingness to pay an “open-access surcharge” of $500 or more to make their papers available for free online immediately upon publication—this above and beyond the $1,700 in page charges that the average <italic>PNAS</italic> author already pays (<xref rid="pbio-0020105-Cozzarelli1" ref-type="bibr">Cozzarelli et al. 2004</xref>).</p><p>Although we recognize that authors who submit to <italic>PLoS Biology</italic> may well be a self-selected group of enthusiastic open-access supporters, we have found that nearly 90% of those who submit manuscripts do not request a fee waiver, and the few who do still offer to pay some portion of the fee.</p><p>The concern about authors' ability to pay publication charges will become less pressing as governments, funding organizations, and institutions increasingly support open-access publication on their researchers' behalf. More funding agencies are joining the Howard Hughes Medical Institute, the Wellcome Trust, and others who have already designated funds for open-access publication. (For more information about these funders' announcements and other international policy statements relevant to open access, see <ext-link ext-link-type="uri" xlink:href="http://www.plos.org/openaccess">http://www.plos.org/openaccess</ext-link>.)</p><p>Universities, too, are supporting open access directly by setting aside funds for open-access publication through institutional memberships with BMC and PLoS or through discretionary funds that faculty can tap into to pay publication charges. Such approaches reduce authors' reliance on individual grants to support charges directly and ensure equal access to publishing options that require such payments.</p></sec><sec id="s3"><title>The Disenfranchised</title><p>Even with the steady increase in sources to pay publication fees, detractors claim that open-access publishing may lead to a situation in which some authors are simply unable to publish their work due to lack of funds. The response to this concern is that the ability of authors to pay publication charges must never be a consideration in the decision to publish their papers. To ensure that this happens, PLoS has a firewall in place such that neither the editors nor the reviewers know which authors have indicated whether or not they can pay. Because all work judged worthy of publication by peer review should be published, any open-access business model should be designed to account for fee waivers, just as publishers have always absorbed some authors' inability to pay page and color charges. PLoS grants full or partial publication-charge waivers to any author who requests them, no questions asked.</p><p>In part, the savings to institutions, hospitals, nongovernmental organizations, and universities provided by open-access publications could help to establish funds for researchers who are less well supported. In the developing world, as free online access to scientific literature is increasingly seen as a political imperative, organizations such as the World Health Organization, the Oxford-based International Network for the Availability of Scientific Publications, and Brazil's SciELO are likely to become more willing to pay open-access publication charges for authors who cannot afford them. The Open Society Institute (OSI) already pays such costs for universities and other organizations in a number of countries in which the foundation is active by way of a PLoS Institutional Membership that grants waived publication charges to authors while providing compensatory revenue for PLoS.</p><p>Perhaps the real misconception about the unfair burden that open access places on authors resides in the terminology—the term “author charge” is itself misleading. Publication fees are not borne purely by authors, but are shared by the many organizations whose missions depend on the broadest possible dissemination and communication of scientific discoveries. Some of those may provide funding for open-access publication as intermediaries between authors and journals, as OSI does. Others—including many government-financed funding agencies—do so directly through their research grants to scientists. In both cases, funding open access is an effective way to fulfill mandates for public access to and accountability over scientific research and to ensure that all worthy research is published.</p></sec>
A Pacific Culture among Wild Baboons: Its Emergence and Transmission	<p>Reports exist of transmission of culture in nonhuman primates. We examine this in a troop of savanna baboons studied since 1978. During the mid-1980s, half of the males died from tuberculosis; because of circumstances of the outbreak, it was more aggressive males who died, leaving a cohort of atypically unaggressive survivors. A decade later, these behavioral patterns persisted. Males leave their natal troops at adolescence; by the mid-1990s, no males remained who had resided in the troop a decade before. Thus, critically, the troop's unique culture was being adopted by new males joining the troop. We describe (a) features of this culture in the behavior of males, including high rates of grooming and affiliation with females and a “relaxed” dominance hierarchy; (b) physiological measures suggesting less stress among low-ranking males; (c) models explaining transmission of this culture; and (d) data testing these models, centered around treatment of transfer males by resident females.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Sapolsky</surname><given-names>Robert M</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Share</surname><given-names>Lisa J</given-names></name><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>A goal of primatology is to understand the enormous variability in primate social behavior. Early investigators examined interspecies differences, e.g., that pair-bonding is more common among arboreal than terrestrial primates (<xref rid="pbio-0020106-Crook1" ref-type="bibr">Crook and Gartlan 1966</xref>). Attention has also focused on geographical differences in behavior within species (<xref rid="pbio-0020106-Whiten2" ref-type="bibr">Whiten et al. 1999</xref>). Often, such differences reflect environmental factors (e.g., a correlation between quantities of rainfall and foraging time) or, in theory, could reflect genetic drift. However, increasing evidence suggests that group-specific traits can also represent “traditions” or “cultures” (the latter term will be used, commensurate with the near consensus among primatologists that the term can be appropriately applied to nonhuman primates).</p><p>As traditionally applied to humans, such “culture” can be defined as behaviors shared by a population, but not necessarily other species members, that are independent of genetics or ecological factors and that persist past their originators (<xref rid="pbio-0020106-Kroeber1" ref-type="bibr">Kroeber and Kluckhohn 1966</xref>; <xref rid="pbio-0020106-Cavalli-Sforza1" ref-type="bibr">Cavalli-Sforza 2000</xref>; <xref rid="pbio-0020106-de2" ref-type="bibr">de Waal 2000</xref>; <xref rid="pbio-0020106-de3" ref-type="bibr">de Waal 2001</xref>). Thus defined, transmission of culture occurs in apes (<xref rid="pbio-0020106-McGrew1" ref-type="bibr">McGrew 1998</xref>; <xref rid="pbio-0020106-Whiten2" ref-type="bibr">Whiten et al. 1999</xref>; <xref rid="pbio-0020106-van2" ref-type="bibr">van Schaik et al. 2003</xref>), monkeys (<xref rid="pbio-0020106-Kawai1" ref-type="bibr">Kawai 1965</xref>; <xref rid="pbio-0020106-Cambefort1" ref-type="bibr">Cambefort 1981</xref>; <xref rid="pbio-0020106-Perry1" ref-type="bibr">Perry et al. 2003</xref>), cetaceans (<xref rid="pbio-0020106-Noad1" ref-type="bibr">Noad et al. 2000</xref>; <xref rid="pbio-0020106-Rendell1" ref-type="bibr">Rendell and Whitehead 2001</xref>), and fish and birds (<xref rid="pbio-0020106-Laland2" ref-type="bibr">Laland and Reader 1999</xref>; <xref rid="pbio-0020106-Laland1" ref-type="bibr">Laland and Hoppitt 2003</xref>). As particularly striking examples, chimpanzees <italic>(Pan troglodytes)</italic> across Africa demonstrate variability in 39 behaviors related to tool use, grooming, and courtship (<xref rid="pbio-0020106-Whiten2" ref-type="bibr">Whiten et al. 1999</xref>), and the excavation of near-millenium-old chimpanzee tools has been reported (<xref rid="pbio-0020106-Mercader1" ref-type="bibr">Mercader et al. 2002</xref>).</p><p>Nearly all such cases of nonhuman culture involve either technology (for example, the use of hammers for nut cracking by chimpanzees), food acquisition, or communication. In this paper, we document the emergence of a unique culture in a troop of olive baboons <italic>(Papio anubis)</italic> related to the overall structure and social atmosphere of the troop. We also document physiological correlates of this troop atmosphere, the transmission of relevant behaviors past their originators, and possible mechanisms of transmission.</p></sec><sec id="s2"><title>Results/Discussion</title><sec id="s2a"><title>Circumstances Leading to the Emergence of a Unique Culture</title><p>In the early 1980s, Forest Troop slept in trees 1 km from a tourist lodge. During that period, an open garbage pit was greatly expanded at the lodge. This attracted an adjacent baboon troop, Garbage Dump Troop, which slept near the pit and foraged almost exclusively there.</p><p>By 1982, many Forest Troop males went to the garbage pit at dawn for food. While such refuse eaters did not differ in age distribution (data not shown) or average dominance rank from non–refuse eaters, they were more aggressive (<xref ref-type="table" rid="pbio-0020106-t001">Table 1</xref>); such aggressiveness could be viewed as a prerequisite in order to compete with Garbage Dump males for access to refuse. Refuse eaters were also involved in more dominance interactions within Forest Troop than were non–refuse eaters (note that frequency of dominance interactions is independent of outcome, and thus of rank).</p><table-wrap id="pbio-0020106-t001" position="float"><label>Table 1</label><caption><title>Characteristics of Forest Troop Males As a Function of Whether They Competed for Refuse with the Garbage Dump Troop</title></caption><graphic xlink:href="pbio.0020106.t001"/><table-wrap-foot><fn id="nt101"><p>Statistical comparisons by unpaired t-test, <italic>n</italic> = 7 and <italic>n</italic> = 8 for refuse eaters and remaining males, respectively. Dominance rank based on approach–avoidance criteria (<xref rid="pbio-0020106-Altmann1" ref-type="bibr">Altmann 1974</xref>). Data concerning refuse eaters were derived solely from their time in the troop, rather than including time spent with the Garbage Dump Troop. Rate of male–male aggression consisted of aggression with any other adult or subadult male in the troop. Rate of aggression directed at females included all adult and subadult females. Rates of behaviors are per 100 h of focal observation, except for grooming, which is per 10 h. Data are mean ± standard error of the mean (SEM)</p></fn></table-wrap-foot></table-wrap><p>In 1983, an outbreak of bovine tuberculosis occurred, originating from infected meat in the dump. From 1983 to 1986, most Garbage Dump animals died, as did all refuse-eating Forest Troop males (46% of adult males); no other Forest Troop animals died (<xref rid="pbio-0020106-Tarara1" ref-type="bibr">Tarara et al. 1985</xref>; <xref rid="pbio-0020106-Sapolsky4" ref-type="bibr">Sapolsky and Else 1987</xref>).</p><p>These deaths greatly altered Forest Troop composition, such that there were fewer adult males and more adult females; this more than doubled the female:male ratio (<xref ref-type="table" rid="pbio-0020106-t002">Table 2</xref>). By 1986, troop behavior had changed markedly, because only less aggressive males had survived.</p><table-wrap id="pbio-0020106-t002" position="float"><label>Table 2</label><caption><title>Troop Composition Before and After the Tuberculosis Outbreak</title></caption><graphic xlink:href="pbio.0020106.t002"/><table-wrap-foot><fn id="nt201"><p>Data from annual troop censuses. Census numbers include both “subadult” animals (undergoing the emergence of secondary sexual characteristics) and “fully adult” (fully emerged secondary sexual characteristics)</p></fn></table-wrap-foot></table-wrap><p>Because of these events, observations of the troop were stopped, and only censusing was done until 1993. Research was begun on Talek Troop, approximately 50 km away.</p><p>In 1993, informal observation of Forest Troop indicated that the behavioral features seen by 1986 had persisted. Critically, by 1993, no adult males remained from 1983–1986; all current adult males had joined the troop following 1986. Thus, the distinctive behaviors that emerged during the mid-1980s because of the selective deaths were being carried out by the next cohort of adult males that had transferred into the troop. Focal sampling on Forest Troop recommenced in 1993, in order to document this phenomenon. Data from Forest Troop 1993–1996 (henceforth, F93–96) were compared with two other data sets that served as controls: observations from 1993–1998 on the Talek Troop (henceforth T93–98), and observations of Forest Troop itself prior to the deaths (1979, 1980, 1982; henceforth F79–82). These two control data sets did not differ significantly from each other and were combined, henceforth T93–98/F79–82.</p></sec><sec id="s2b"><title>Atypical Features of the Behavior of Forest Troop Males</title><sec id="s2b1"><title>Male–male dominance interactions</title><p>Males of F93–96 and T93–98/F79–82 had similar rates of approach–avoidance dominance interactions (data not shown). Moreover, dominance stability did not differ, as measured by the percentage of approach–avoidance interactions which represented a reversal of the direction of dominance within a dyad of males of adjacent rank (16% ± 5% and 20% ± 5% for F93–96 and T93–98/F79–82, respectively, n.s.). There was also no difference in the average tenure length of the highest-ranking male (approximately a year).</p><p>Despite those similarities, dominance behavior in F93–96 differed from the two control cases in ways that, arguably, made for less stress for low-ranking males. A first example concerns approach–avoidance dominance interactions between males more than two ranks apart in the hierarchy. The overwhelming majority of such interactions were won by the higher-ranking individual. Because a male is rarely seriously threatened by an individual more than two ranks lower in the hierarchy, interactions between individuals that far apart typically represent harassment of or displacement of the subordinate by the higher-ranking male, rather than true competition. In T93–98 and F79–82, approximately 80% of approach–avoidance interactions were between males more than two ranks apart in the hierarchy. In contrast, a significantly smaller percentage of approach–avoidance interactions were soin F93–96 (<xref ref-type="fig" rid="pbio-0020106-g001">Figure 1</xref>A). Instead, a disproportionate percentage of F93–96 dominance interactions occurred among males of adjacent ranks (with, as noted, no difference in dominance stability)(<xref ref-type="fig" rid="pbio-0020106-g001">Figure 1</xref>B). Moreover, high-ranking males in F93–96 were more “tolerant” of very low-ranking males, as there was a disproportionately high number of reversals with males more than two steps lower in the hierarchy (<xref ref-type="fig" rid="pbio-0020106-g001">Figure 1</xref>C). Thus, in F93–96, with a typical level of dominance stability, approach–avoidance dominance interactions were concentrated among closely ranking animals, with low-ranking males being more tolerated and less subject to harassment and/or displacement by high-ranking males.</p><fig id="pbio-0020106-g001" position="float"><label>Figure 1</label><caption><title>Quality of Male–Male Dominance Interactions</title><p>(A) Percentage of male approach–avoidance dominance interactions occurring between males more than two ranks apart.</p><p>(B) Percentage of male approach–avoidance interactions occurring between males of adjacent ranks.</p><p>(C) Percentage of approach–avoidance interactions representing a reversal of the direction of dominance within a dyad by a male more than two steps lower ranking. Mean ± SEM, and * indicate <italic>p</italic> < 0.02 and <italic>p</italic> < 0.01, respectively, by t-test, treating each male/year as a data point. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 19 different males in F79–82. Potentially, the result in (B) could have arisen from different numbers of males in F93–96 versus the other two troops (a smaller group size does not change the number of adjacent animals available to any given subject, but decreases the number of nonadjacent animals available). However, the same results were found if the numbers of males in the three troops were artificially made equal by excluding excess males from either the top or the bottom of the hierarchy (data not shown).</p></caption><graphic xlink:href="pbio.0020106.g001"/></fig></sec><sec id="s2b2"><title>Aggression</title><p>Patterns of aggression also differed between F93–96 and T93–98/F79–82 in a way that suggested a less stressful environment for subordinates in F93–96. The troops had similar overall rates of aggressive interactions (<xref ref-type="table" rid="pbio-0020106-t003">Table 3</xref>). However, aggression in F93–96 was more likely than in the control troops to occur between closely ranked animals (i.e., within two rank steps), rather than to reflect high-ranking males directing aggression at extremely low-ranking ones; the latter type of interaction is particularly stressful for a subordinate, because of its typical unpredictability. Moreover, F93–96 males were less likely than T93–98/F79–82 males to direct aggression at females.</p><table-wrap id="pbio-0020106-t003" position="float"><label>Table 3</label><caption><title>Patterns of Aggression in Forest Troop 1993–1996 versus Talek Troop 1993–1998 and Forest Troop 1979–1982</title></caption><graphic xlink:href="pbio.0020106.t003"/><table-wrap-foot><fn id="nt301"><p> and * indicate <italic>p</italic> < 0.025 and <italic>p <</italic> 0.01 by unpaired t-test, respectively. Observed/expected ratios were derived by comparing observed frequencies of behavior with the frequencies expected with even distribution of aggressive interactions across all dyads; a ratio of 1.0 indicates the behavior occurring at the expected rate. Data from T93–98 and F79–82 did not differ significantly, and thus were pooled. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 19 different males in F79–82</p></fn></table-wrap-foot></table-wrap><p>We examined the data for reconciliative behavior (i.e., affiliative behaviors between pairs following aggressive interactions [<xref rid="pbio-0020106-de5" ref-type="bibr">de Waal and van Roosmalen 1979</xref>]) in F93–96 and T93–98/F79–82. However, we saw no male–male reconciliation in any troop, in agreement with prior reports (<xref rid="pbio-0020106-Cheney1" ref-type="bibr">Cheney et al. 1995</xref>).</p></sec><sec id="s2b3"><title>Affiliative behaviors</title><p>Quantitative data on affiliative behaviors were not available for F79–82. However, F93–96 males socially groomed more often than did control T93–98 males (<xref ref-type="fig" rid="pbio-0020106-g002">Figure 2</xref>A) this difference was due to more grooming between males and females. F93–96 males were also in close proximity to other animals more often than were T93–98 males (<xref ref-type="fig" rid="pbio-0020106-g002">Figure 2</xref>B). While males did not differ between troops in the average number of adult male neighbors (i.e., within 3 m), F93–96 males were more likely than T93–98 males to have adult females, infants, adolescents, and juveniles as neighbors.</p><fig id="pbio-0020106-g002" position="float"><label>Figure 2</label><caption><title>Quality of Affiliative Behaviors</title><p>(A) Amount of grooming involving adult males in Forest Troop 1993–1996 and Talek Troop 1993–1996. The first pair of columns represents mean time adult males spent grooming adult females; the second pair, mean time adult males were groomed by adult females.</p><p>(B) Comparison of average number of neighbors (i.e., within 3 m) of adult males in the two troops. Mean ± SEM. , , and ** indicate <italic>p</italic> < 0.05, 0.02, and 0.01, respectively, by unpaired t-test. Data were derived from a total of ten different males and 17 different females in F93–96, 31 different males and 21 different females in T93–98, and 19 different males and 23 different females in F79–82.</p></caption><graphic xlink:href="pbio.0020106.g002"/></fig></sec><sec id="s2b4"><title>Sexual behavior</title><p>Sexual behavior did not differ between F93–96 and T93–98/F79–82. The percentages of nonpregnant, nonlactating females in estrus per day did not differ (27% ± 7% and 30% ± 4%, respectively, n.s.). Moreover, the relationship between male rank and reproductive success did not differ (<italic>R<sup>2</sup></italic> of correlation between rank and reproductive success: 0.25 ± 0.25 and 0.54 ± 0.10, respectively, n.s.).</p></sec></sec><sec id="s2c"><title>Physiological Correlates of Behavioral Features of Forest Troop</title><p>Thus, F93–96 males had high rates of affiliative behaviors, and low-ranking males were subject to low rates of aggressive attack and subordination by high-ranking males. In a stable hierarchy, low-ranking baboon males show physiological indications of being stressed, including elevated basal levels of glucocorticoids (the adrenal hormones secreted in response to stress), hypertension, and decreased levels of high density lipoprotein cholesterol, growth factors, and circulating lymphocytes (<xref rid="pbio-0020106-Sapolsky1" ref-type="bibr">Sapolsky 1993</xref>; <xref rid="pbio-0020106-Sapolsky5" ref-type="bibr">Sapolsky and Share 1994</xref>; <xref rid="pbio-0020106-Sapolsky7" ref-type="bibr">Sapolsky and Spencer 1997</xref>). We tested whether subordinate males in F93–96 were spared the stress-related physiology of subordination seen in other troops.</p><p>This was the case (<xref ref-type="fig" rid="pbio-0020106-g003">Figure 3</xref>A). In F79–82, i.e., prior to the tuberculosis outbreak, subordination was associated with elevated basal levels of glucocorticoids, as in other species in which subordination entails extensive stressors and low rates of coping outlets (<xref rid="pbio-0020106-Sapolsky3" ref-type="bibr">Sapolsky 2001</xref>). While glucocorticoids aid in surviving an acute physical stressor, chronic overexposure increases the risk of glucose intolerance, hypertension, ulcers, and reproductive and immune suppression (<xref rid="pbio-0020106-Sapolsky8" ref-type="bibr">Sapolsky et al. 2000</xref>). In contrast to this picture in F79–82, in which subordination was associated with a physiology suggesting a chronic state of stress, subordinate F93–96 males did not have elevated basal glucocorticoid levels (levels were unavailable for T93–98).</p><fig id="pbio-0020106-g003" position="float"><label>Figure 3</label><caption><title>Stress-Related Physiological Profiles</title><p>(A) Basal glucocorticoid levels (μg/100 ml). Males were split into higher- and lower-ranking 50%, by approach–avoidance criteria. The primate glucocorticoid, cortisol, was measured by radioimmunoassay.</p><p>(B) Number of anxiety-related behaviors observed 10–20 min after β-carboline-3-carboxylic acid administration (M-156, Research Biochemicals International, Natick, Massachusetts, United States), after subtracting the number observed 10–20 min after vehicle administration (dextrin in 1 ml saline); 0.5 g of the drug in 1ml saline was delivered intramuscularly by dart syringe (Pneu-Dart, Inc., Williamsport, Pennsylvania, United States) fired from a blowgun at 5 m. Mean ± SEM. * and *** indicate <italic>p</italic> < 0.05 and <italic>p</italic> < 0.01, respectively, by unpaired t-test. Data were derived from a total of ten different males in F93–96, 31 different males in T93–98, and 18 different males in F79–82.</p></caption><graphic xlink:href="pbio.0020106.g003"/></fig><p>Subordinate F93–96 males were spared another stress-related physiological marker. Experimental anxiety was induced by darting males, intramuscularly, with β-carboline-3-carboxylic acid, a benzodiazepine receptor antagonist which induces behavioral and physiological indices of anxiety in primates (benzodiazepine receptors bind tranquilizers such as valium and librium and mediate their anxiolytic effects)(<xref rid="pbio-0020106-Ninan1" ref-type="bibr">Ninan et al. 1982</xref>). Males were darted on days when they had not had a fight, injury or mating. As a control, they were darted on separate days with vehicle alone (order of dartings randomized). Males were then monitored by an observer unaware of treatment.</p><p>β-carboline-3-carboxylic acid had no effect on behavior in high-ranking males in T93–98 or F93–96 (<xref ref-type="fig" rid="pbio-0020106-g003">Figure 3</xref>B).The drug increased anxiety-related behaviors in low-ranking males in T93–98 but not in F93–96 (the recorded anxiety-related behaviors were self-scratching, rhythmic head shaking, assuming a vigilant stance, repeated wiping of nose, and jaw grinding in a solitary male [<xref rid="pbio-0020106-Ninan1" ref-type="bibr">Ninan et al. 1982</xref>; <xref rid="pbio-0020106-Aureli1" ref-type="bibr">Aureli and van Schaik 1991</xref>; <xref rid="pbio-0020106-Castles1" ref-type="bibr">Castles et al. 1999</xref>]).</p><p>Thus, in the more typical F79–82 and T93–98 troops, subordination had distinctive stress-related physiological correlates. In contrast, F93–96 males lacked these rank-related differences.</p></sec><sec id="s2d"><title>Potential Mechanisms Underlying Transmission of This Culture</title><p>A decade after the deaths of the more aggressive males in the troop, Forest Troop preserved a distinct social milieu accompanied by distinct physiological correlates. Critically, as noted, no adult males in F93–96 had been troop members at the end of the tuberculosis outbreak. Instead, these males had subsequently transferred in as adolescents, adopting the local social style. A number of investigators have emphasized how a tolerant and gregarious social setting facilitates social transmission (e.g., <xref rid="pbio-0020106-van1" ref-type="bibr">van Schaik et al. 1999</xref>), exactly the conditions in F93–96.</p><p>The present case of social transmission is reminiscent of some prior cases. For example, juvenile rhesus monkeys <italic>(Macaca mulatta)</italic> housed with stumptail macaques <italic>(M. artoides)</italic> assume the latter's more conciliatory style (<xref rid="pbio-0020106-de4" ref-type="bibr">de Waal and Johanowicz 1993</xref>). Moreover, anubis baboons <italic>(Papio anubis)</italic> and hamadryas baboons <italic>(P. hamadryas)</italic> differ in social structure, and females of either species experimentally transferred into a group of the other adopt the novel social structure within hours (<xref rid="pbio-0020106-Kummer1" ref-type="bibr">Kummer 1971</xref>).</p><p>Several models have been hypothesized to explain transmission of cultures (<xref rid="pbio-0020106-Whiten2" ref-type="bibr">Whiten et al. 1999</xref>; <xref rid="pbio-0020106-de3" ref-type="bibr">de Waal 2001</xref>; <xref rid="pbio-0020106-Galef1" ref-type="bibr">Galef 1990</xref>). For clarity, it is useful to first consider their application to an established example of transmission of a “technology” before then applying them to the transmission of the social milieu of F93–96. An example of the former is the nut cracking with stone hammers by West African chimpanzees (<xref rid="pbio-0020106-Boesch3" ref-type="bibr">Boesch and Boesch 1983</xref>; <xref rid="pbio-0020106-Boesch2" ref-type="bibr">Boesch 2003</xref>), a trait transmitted transgenerationally.</p><p>In “instructional models” of chimpanzee tool use, young are actively taught hammer use. In the case of F93–96, instructional models would involve new transfer males being subject to socially rewarding interactions (e.g., grooming) or aversive ones (e.g., supplantation or attack) contingent upon their assimilating the troop tradition. In such models, a key question is who “instructs.” Much as with the term “culture” being used with respect to animal behavior, the use of the term “instruction” has also generated some controversy, with some preferring the concept of “active behavioral modification” by others bringing about the change. As a striking example of that, when young male cowbirds learn to produce their local song, they initially produce an undifferentiated repertoire of songs, and females react to the production of appropriate dialect with copulation solicitation displays, thus providing positive reinforcement and shaping those behaviors (<xref rid="pbio-0020106-Smith1" ref-type="bibr">Smith et al. 2000</xref>).</p><p>In “observational models” applied to chimpanzee tool use, young learn nut cracking by observing and copying adults. As applied to F93–96, transfer males would model behavior upon that of resident males.</p><p>In “facilitation models” of the chimpanzee example, proximity to adults and their hammers increases the likelihood of the young experimenting with hammers and deriving the skill themselves. As applied to the baboons, male F93–96 behaviors would be an implicit default state where, in the absence of the more typical rates of male aggression (either male–male or male–female), females broadly tend to become more affiliative, and in the context of more affiliative female behavior, transfer males broadly tend to become less aggressive. As perhaps a way of stating the same, the default state may emerge because of the atmosphere of a troop with a high female:male ratio (with less need for male competition for access to estrus females).</p><p>Finally, a “self-selection model” may apply to the baboons, in which particular kinds of males were more prone to transfer into such a troop (note that the fact that males transferred in from an array of surrounding troops rules out the possibility of an additional model, in which the culture was continued by genetic means).</p><p>We assessed these models by analyzing cases where adolescent males transferred on known dates and were observed for at least 2–6 mo afterward. Thus, we searched for behavioral patterns involving new transfer males that might differ between F93–96 (five such transfers) and T93–98/F79–82 (12 transfers).</p><p>Many interactions involving new transfer males did not differ (<xref ref-type="table" rid="pbio-0020106-t004">Table 4</xref>). Transfer males in F93–96, T93–98, and F79–82 all attacked and supplanted females from feeding or resting sites at equal rates. Moreover, despite the different dominance structure among resident F93–96 males, resident males in F93–96, T93–98, and F79–82 all treated new transfer males similarly. There were similar latencies until transfer males were first lunged at by residents, and transfer males were involved in dominance and aggressive interactions at similar rates in all three troops (note that because there were half as many resident males in F93–96 as in T93–98 or F79–82, the rate of such interactions within any given resident/transfer male dyad would differ). We examined instances where resident males acted aggressively towards transfer males, determining whether such behaviors were more prevalent during the 20 min after aggressive behavior by the transfer male than at other, randomly selected times (<xref rid="pbio-0020106-de6" ref-type="bibr">de Waal and Yoshihara 1983</xref>; <xref rid="pbio-0020106-de4" ref-type="bibr">de Waal and Johanowicz 1993</xref>). We found no evidence for such contingent behavior (data not shown).</p><table-wrap id="pbio-0020106-t004" position="float"><label>Table 4</label><caption><title>Behaviors of Newly Transferred Males</title></caption><graphic xlink:href="pbio.0020106.t004"/><table-wrap-foot><fn id="nt401"><p>Subjects consisted of the five transfer males in F93–96 and the 12 in T93–98/F79–82</p></fn></table-wrap-foot></table-wrap><p>We then examined affiliative interactions between females and new transfer males, and found striking differences between F93–96 and T93–98/F79–82, in that F93–96 females treated new transfer males in the same affiliative manner that they treated resident males. F93–96 transfer males had a shorter latency until first being groomed by or presented to by a female than did T93–98/F79–82 transfer males (<xref ref-type="fig" rid="pbio-0020106-g004">Figure 4</xref>A). (The differences between F93–96 and T93–98 did not arise from a single F93–96 female accounting for the much shorter latencies until presentation and grooming: three different females accounted for the first interactions with the five F93–96 transfer males). Moreover, F93–96 transfers sat in closer proximity to and had more grooming bouts with females than did T93–98/F79–82 transfers (<xref ref-type="fig" rid="pbio-0020106-g004">Figure 4</xref>B). While estrous females are more likely than nonestrous females to interact with transfer males (<xref rid="pbio-0020106-Smuts1" ref-type="bibr">Smuts 1999</xref>), the percentage of females in estrus did not differ among the troops (see above). In addition, F93–96 females did not seem to treat transfer males in a contingent manner (<xref rid="pbio-0020106-de6" ref-type="bibr">de Waal and Yoshihara 1983</xref>; <xref rid="pbio-0020106-de4" ref-type="bibr">de Waal and Johanowicz 1993</xref>). To test for this, we first examined instances where resident females were affiliative towards transfer males, determining whether this was more likely during the 20 min following an affiliative behavior on the part of the transfer male than at other, randomly selected times. Second, we determined whether females were less likely to be affiliative during the 20 min following an aggressive behavior on the part of a transfer male. We found no evidence for either pattern (data not shown).</p><fig id="pbio-0020106-g004" position="float"><label>Figure 4</label><caption><title>Quality of Interactions between Resident Females and Transfer Males</title><p>(A) Latency, in days, until a newly transferred male is first groomed by a female (left) or presented to by a female (right).</p><p>(B) Average number of adult female neighbors per scan (i.e., within 3 m; left) and average number of grooming bouts with females per 100 h of observation (right) for transfer males. Mean ± SEM. * and *** indicate <italic>p</italic> < 0.05 and <italic>p</italic> < 0.01, respectively, by unpaired t-test. Latency until first presented to by a female approached significance (<italic>p</italic> < 0.08). Data were derived from a total of ten different males and 17 different females in F93–96, and 31 different males and 21 different females in T93–98.</p></caption><graphic xlink:href="pbio.0020106.g004"/></fig><p>These data allow some insight as to the mechanisms of social transmission in F93–96 (without remotely allowing an analysis fine-grained enough to see whether these mechanisms were equally relevant to the transmission of all the components of the F93–96 culture, namely the low rates of male aggression, the high rates of female affilitation, and the relaxed dominance structure). The lack of contingency in the treatment of transfer males by residents argues against instruction; commensurate with this, there is relatively little evidence for “instruction” in nonhuman primate cultural transmission (<xref rid="pbio-0020106-de3" ref-type="bibr">de Waal 2001</xref>; for an exception, see <xref rid="pbio-0020106-Boesch1" ref-type="bibr">Boesch 1991</xref>). The similar rates of displacement behaviors by transfer males onto females in all three troops argue against self-selection (i.e., the possibility that F93–96 transfer males already behaved differently than transfer males elsewhere). This is not surprising. While adolescent male baboons may transfer repeatedly before choosing a troop (<xref rid="pbio-0020106-Pusey1" ref-type="bibr">Pusey and Packer 1986</xref>), as well as later in life (<xref rid="pbio-0020106-Sapolsky2" ref-type="bibr">Sapolsky 1996</xref>), we have seen little evidence among these animals of the systematic sampling of different troops required by a self-selection model.</p><p>The data instead support either observational or facilitative/default models. Insofar as resident males in all troops interacted with transfer males similarly, transmission in F93–96 could have involved observation only if such observations were of how resident males interacted with females or each other. Some, but not all, studies support observational models of social transmission in other primates (<xref rid="pbio-0020106-Visalberghi1" ref-type="bibr">Visalberghi and Fragaszy 1990</xref>; <xref rid="pbio-0020106-Whiten1" ref-type="bibr">Whiten 1998</xref>; <xref rid="pbio-0020106-Boesch2" ref-type="bibr">Boesch 2003</xref>; <xref rid="pbio-0020106-Whiten3" ref-type="bibr">Whiten et al. 2003</xref>); there are few data at present from baboons concerning this issue. As shown, F93–96 transfer males were had high rates of affilitative interactions with females. The preponderance of females in F93–96 is a plausible explanation for their unconditional (or, at least, less conditional) increase in tolerance of and affiliation with males (including transfer males), insofar as males in the troop had less numeric means to be aggressive to females. (Note that this skewed sex ratio continues in this troop to the present, for unknown reasons.) Thus, affilative data support a facilitative/default model only if it involves preferential sensitivity to the quality of interactions with females.</p><p>This analysis raises the possibility that there is no social transmission, but that the F93–96 pattern is merely the emergent outcome of the 2:1 female:male ratio. To test this, we analyzed the five available studies of baboon troops with adult female:male ratios of 2 or more which contained quantitative data comparable to the present data (<xref rid="pbio-0020106-Seyfarth1" ref-type="bibr">Seyfarth 1976</xref>, <xref rid="pbio-0020106-Seyfarth2" ref-type="bibr">1978</xref>; <xref rid="pbio-0020106-Strum1" ref-type="bibr">Strum 1982</xref>; <xref rid="pbio-0020106-Bercovitch1" ref-type="bibr">Bercovitch 1985</xref>; <xref rid="pbio-0020106-Noe1" ref-type="bibr">Noe 1994</xref>). The key question was whether those prior data more closely resembled those of F93–96 or the control troops. Previous data more closely resembled, and did not differ significantly from, data from the control troops for the percentage of time males groomed females (based on <xref rid="pbio-0020106-Seyfarth2" ref-type="bibr">Seyfarth 1978</xref>), the percentage of time females groomed males (<xref rid="pbio-0020106-Seyfarth2" ref-type="bibr">Seyfarth 1978</xref>), the rate of intersexual aggression (<xref rid="pbio-0020106-Seyfarth1" ref-type="bibr">Seyfarth 1976</xref>, <xref rid="pbio-0020106-Seyfarth2" ref-type="bibr">1978</xref>), the structure of male–male dominance (<xref rid="pbio-0020106-Noe1" ref-type="bibr">Noe 1995</xref>), or the structure of of male–male aggression (<xref rid="pbio-0020106-Strum1" ref-type="bibr">Strum 1982</xref>; <xref rid="pbio-0020106-Bercovitch1" ref-type="bibr">Bercovitch 1985</xref>). In contrast, no quantitative measures more closely resembled F93–96. This strongly suggests that the F93–96 pattern is unique and is being uniquely maintained, rather than being the social structure that automatically emerges whenever a female-skewed female:male ratio occurs. Thus, insofar as a facilitative/default model is operating in this troop, it cannot be a relative paucity of males which “activates” a default state; instead, it would likely be the paucity of aggressive males.</p><p>The unconditional (or less conditional) nature of the default model is puzzling, in that it requires that females be relatively affiliative to recent transfer males who, nonetheless, are initially aggressive to them. This seems counter to the long-standing emphasis in primatology on individual relations (i.e., females are unlikely to be unable to distinguish between relatively unaggressive resident males and relatively aggressive newly transferred males). Precedent for this unexpected implication comes from the social epidemiology literature concerning “social capital,” in which health and life expectancy increase in a community as a function of communitywide attributes that transcend the level of the individual or individual social networks (<xref rid="pbio-0020106-Kawachi1" ref-type="bibr">Kawachi et al 1997</xref>).</p><p>In summary, we have observed circumstances that produced a distinctive set of behaviors and physiological correlates in a troop of wild baboons. Moreover, these behaviors were taken on by new troop members; while obviously not conclusive, the data suggest that this most likely occurs through observational or facilitative/default models. Finally, somewhat uniquely in nonhuman primate studies, these findings concern the intergenerational transfer of social, rather than material culture.</p><p>These findings raise some issues. There appear to be adverse health consequences of the stress-related physiological profile of subordination in typical baboon troops (<xref rid="pbio-0020106-Sapolsky1" ref-type="bibr">Sapolsky 1993</xref>; <xref rid="pbio-0020106-Sapolsky5" ref-type="bibr">Sapolsky and Share 1994</xref>; <xref rid="pbio-0020106-Sapolsky7" ref-type="bibr">Sapolsky and Spencer 1997</xref>). The distinctive rank-related patterns of physiology in F93–96 suggest that subordinate males in that troop may be spared those pathologies. Another issue concerns the consequences of the culture of F93–96 remaining stable over some time. A hallmark of human culture is that it is cumulative (i.e., innovations are built upon each other), and there is only scant evidence, at best, for the same in nonhuman primates (<xref rid="pbio-0020106-Boesch2" ref-type="bibr">Boesch 2003</xref>). It would thus be interesting to see if additional features of the F93–96 social tradition emerge with time.</p><p>A converse issue concerns circumstances that might destroy the F93–96 culture. The culture might be destroyed if numerous males transfer into the troop simultaneously, or if a male transfers in who, rather than assuming the F93–96 culture, instead takes advantage of it. Game theory suggests that F93–96 would be vulnerable to such “cheating.” Another issue concerns the fate of natal males from F93–96 when they transfer elsewhere. Reciprocal altruism models (<xref rid="pbio-0020106-Axelrod1" ref-type="bibr">Axelrod and Hamilton, 1981</xref>) suggest that if one F93–96 male transfers elsewhere and continues his natal behavioral style, he will be at a competitive disadvantage. However, should two F93–96 males simultaneously join another troop and maintain F93–96–typical interactions between them, they might be at a competitive advantage. This might represent a means to transmit this social style between troops.</p><p>Finally, these findings raise the issue of their applicability to understanding human social behavior and its transmission. Human history is filled with examples of the selective loss of demographic subsets of societies (e.g., the relative paucity of adult men following the American Civil War or the relative paucity of girls in contemporary China due to male-biased reproductive technology practices and female-biased infanticide). The present data suggest that demographic skews may have long-term, even multigenerational consequences, including significant changes in the quality of life in a social group.</p></sec></sec><sec id="s3"><title>Materials and Methods</title><p>Subjects were a troop, Forest Troop, of olive baboons <italic>(Papio anubis)</italic> living in the Masai Mara Reserve of Kenya. Olive baboons live in multimale troops of 30–150 animals, with polygamy and considerable male–male aggression. Males change troops at puberty and, as adults, achieve ranks in somewhat fluid dominance hierarchies. In contrast, females remain in their natal troop, inheriting a rank one below that of their mother.</p><p>Subjects were observed each summer from 1978–1986, and continuously since 1993. An additional troop, Talek Troop, was observed continuously since 1984. Behavioral data were collected as 20-min focal samples (<xref rid="pbio-0020106-Altmann1" ref-type="bibr">Altmann 1974</xref>). During years of only summer observation (Forest Troop, 1978–1986), 45 samples were collected per subject per season; otherwise, an average of three samples per subject per week were collected throughout the year. Sampling was distributed throughout the day in the same fashion for each individual. During samples, social behavior, feeding, and grooming were recorded. Rankings were derived from approach–avoidance interactions, which included avoidances, supplants, and presentations, in the absence of aggression. Escalated aggression included open-mouthed lunges, chases, and bites. Nearest neighbor scans were done before and after each sample.</p><p>Reproductive success was indirectly estimated from frequencies of matings and consortships (maintenance of exclusive mating with and proximity to an estrous female for at least one sample). The value of any given consortship or mating was adjusted by the probability of a fertile mating occurring that day (<xref rid="pbio-0020106-Hendrickx1" ref-type="bibr">Hendrickx and Kraemer 1969</xref>).</p><p>Endocrine data were collected under circumstances allowing for measures of basal steroid hormone levels (<xref rid="pbio-0020106-Sapolsky6" ref-type="bibr">Sapolsky and Share 1997</xref>). Subjects were darted unaware with anesthetic from a blowgun syringe between 7 A.M. and 10 A.M., and only on days on which they were not sick, injured, in a consortship, or had not recently had a fight. Blood samples were collected within 3 min of anesthetization.</p></sec>
Semi-Supervised Methods to Predict Patient Survival from Gene Expression Data	<p>An important goal of DNA microarray research is to develop tools to diagnose cancer more accurately based on the genetic profile of a tumor. There are several existing techniques in the literature for performing this type of diagnosis. Unfortunately, most of these techniques assume that different subtypes of cancer are already known to exist. Their utility is limited when such subtypes have not been previously identified. Although methods for identifying such subtypes exist, these methods do not work well for all datasets. It would be desirable to develop a procedure to find such subtypes that is applicable in a wide variety of circumstances. Even if no information is known about possible subtypes of a certain form of cancer, clinical information about the patients, such as their survival time, is often available. In this study, we develop some procedures that utilize both the gene expression data and the clinical data to identify subtypes of cancer and use this knowledge to diagnose future patients. These procedures were successfully applied to several publicly available datasets. We present diagnostic procedures that accurately predict the survival of future patients based on the gene expression profile and survival times of previous patients. This has the potential to be a powerful tool for diagnosing and treating cancer.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Bair</surname><given-names>Eric</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Tibshirani</surname><given-names>Robert</given-names></name><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><sec id="s1a"><title>Predicting Patient Survival</title><p>When a patient is diagnosed with cancer, various clinical parameters are used to assess the patient's risk profile. However, patients with a similar prognosis frequently respond very differently to the same treatment. This may occur because two apparently similar tumors are actually completely different diseases at the molecular level (<xref rid="pbio-0020108-Alizadeh1" ref-type="bibr">Alizadeh et al. 2000</xref>; <xref rid="pbio-0020108-Sorlie1" ref-type="bibr">Sorlie et al. 2001</xref>; <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. 2002</xref>; <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>; <xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. 2004</xref>; <xref rid="pbio-0020108-Lapointe1" ref-type="bibr">Lapointe et al. 2004</xref>).</p><p>The main example discussed in this paper concerns diffuse large B-cell lymphoma (DLBCL). This is the most common type of lymphoma in adults, and it can be treated by chemotherapy in only approximately 40% of patients (<xref rid="pbio-0020108-NHLCP1" ref-type="bibr">NHLCP 1997</xref>; <xref rid="pbio-0020108-Vose1" ref-type="bibr">Vose 1998</xref>; <xref rid="pbio-0020108-Coiffier1" ref-type="bibr">Coiffier 2001</xref>). Several recent studies used DNA microarrays to study the gene expression profiles of patients with DLBCL. They found that it is possible to identify subgroups of patients with different survival rates based on gene expression data (<xref rid="pbio-0020108-Alizadeh1" ref-type="bibr">Alizadeh et al. 2000</xref>; <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. 2002</xref>; <xref rid="pbio-0020108-Shipp1" ref-type="bibr">Shipp et al. 2002</xref>).</p><p>If different subtypes of cancer are known to exist, there are a variety of existing techniques that can be used to identify which subtype is present in a given patient (<xref rid="pbio-0020108-Golub1" ref-type="bibr">Golub et al. 1999</xref>; <xref rid="pbio-0020108-Hastie1" ref-type="bibr">Hastie et al. 2001a</xref>; <xref rid="pbio-0020108-Hedenfalk1" ref-type="bibr">Hedenfalk et al. 2001</xref>; <xref rid="pbio-0020108-Khan1" ref-type="bibr">Khan et al. 2001</xref>; <xref rid="pbio-0020108-Ramaswamy1" ref-type="bibr">Ramaswamy et al. 2001</xref>; <xref rid="pbio-0020108-Nguyen1" ref-type="bibr">Nguyen and Rocke 2002a</xref>, <xref rid="pbio-0020108-Nguyen2" ref-type="bibr">2002b</xref>; <xref rid="pbio-0020108-Shipp1" ref-type="bibr">Shipp et al. 2002</xref>; <xref rid="pbio-0020108-Tibshirani1" ref-type="bibr">Tibshirani et al. 2002</xref>; <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. 2002</xref>; <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>; <xref rid="pbio-0020108-Nutt1" ref-type="bibr">Nutt et al. 2003</xref>). However, most of these techniques are only applicable when the tumor subtypes are known in advance. The question of how to identify such subtypes, however, is still largely unanswered.</p><p>There are two main approaches in the literature to identify such subtypes. One approach uses unsupervised learning techniques, such as hierarchical clustering, to identify patient subgroups. This type of procedure is called “unsupervised” since it does not use any of the clinical information about the patient. The subgroups are identified using only the gene expression data. (In contrast, “supervised learning” would use the clinical data to build the model.) For an overview of unsupervised learning techniques, see <xref rid="pbio-0020108-Gordon1" ref-type="bibr">Gordon (1999</xref>) or <xref rid="pbio-0020108-Hastie2" ref-type="bibr">Hastie et al. (2001b</xref>).</p><p>Hierarchical clustering (<xref rid="pbio-0020108-Eisen1" ref-type="bibr">Eisen et al. 1998</xref>) has successfully identified clinically relevant cancer subtypes in several different studies (<xref rid="pbio-0020108-Alizadeh1" ref-type="bibr">Alizadeh et al. 2000</xref>; <xref rid="pbio-0020108-Bhattacharjee1" ref-type="bibr">Bhattacharjee et al. 2001</xref>; <xref rid="pbio-0020108-Sorlie1" ref-type="bibr">Sorlie et al. 2001</xref>; <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. 2002</xref>; <xref rid="pbio-0020108-Lapointe1" ref-type="bibr">Lapointe et al. 2004</xref>). However, one drawback to unsupervised learning procedures is that they may identify cancer subtypes that are unrelated to patient survival. Although several different subtypes of a given cancer may exist, if the prognosis for all patients is the same regardless of which subtype they have, then the utility of this information is limited. Since unsupervised learning procedures by definition do not use the clinical data to identify subtypes, there is no guarantee that the subtypes they identify will be correlated with the clinical outcome.</p><p>The second approach to identifying subtypes of cancer is based exclusively on the clinical data. For example, patients can be assigned to a “low-risk” or a “high-risk” subgroup based on whether they were still alive or whether their tumor had metastasized after a certain amount of time. This approach has also been used successfully to develop procedures to diagnose patients (<xref rid="pbio-0020108-Shipp1" ref-type="bibr">Shipp et al. 2002</xref>; <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. 2002</xref>; <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>).</p><p>However, by dividing the patients into subgroups based on their survival times, the resulting subgroups may not be biologically meaningful. Suppose, for example, that there are two tumor cell types. Suppose further that patients with cell type 2 live slightly longer than patients with cell type 1 but that there is considerable overlap between the two groups (<xref ref-type="fig" rid="pbio-0020108-g001">Figure 1</xref>). Assume also that the underlying cell types of each patient are unknown. If we were to assign patients to “low-risk” and “high-risk” groups based on their survival times, many patients would be assigned to the wrong group, and any future predictions based on this model would be suspect. We can obtain more accurate predictions by identifying these underlying subtypes and building a model that can determine which subtype is present in future patients.</p><fig id="pbio-0020108-g001" position="float"><label>Figure 1</label><caption><title>Two Patient Subgroups with Overlapping Survival Times</title></caption><graphic xlink:href="pbio.0020108.g001"/></fig></sec><sec id="s1b"><title>Proposed Semi-Supervised Methods</title><p>To overcome these difficulties, we propose a novel procedure that combines both the gene expression data and the clinical data to identify cancer subtypes. The crux of the idea is to use the clinical data to identify a list of genes that correlate with the clinical variable of interest and then apply unsupervised clustering techniques to this subset of the genes.</p><p>For instance, in many studies, the survival times of the patients are known even though no tumor subtypes have been identified (<xref rid="pbio-0020108-Alizadeh1" ref-type="bibr">Alizadeh et al. 2000</xref>; <xref rid="pbio-0020108-Bhattacharjee1" ref-type="bibr">Bhattacharjee et al. 2001</xref>; <xref rid="pbio-0020108-Sorlie1" ref-type="bibr">Sorlie et al. 2001</xref>; <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. 2002</xref>; <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. 2002</xref>; <xref rid="pbio-0020108-Shipp1" ref-type="bibr">Shipp et al. 2002</xref>; <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. 2002</xref>; <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>; <xref rid="pbio-0020108-Nutt1" ref-type="bibr">Nutt et al. 2003</xref>; <xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. 2004</xref>). We can calculate the Cox score for each gene in the expression data—the Cox score measures the correlation between the gene's expression level and patient survival—and consider only the genes with a Cox score that exceeds a certain threshold.</p><p>Once such a list of significant genes is compiled, there are several methods we can use to identify clinical subgroups. We can apply clustering techniques to identify subgroups of patients with similar expression profiles. Once such subgroups are identified, we can apply existing supervised learning techniques to classify future patients into the appropriate subgroup. In this study, we will use the “nearest shrunken centroids” procedure of <xref rid="pbio-0020108-Tibshirani1" ref-type="bibr">Tibshirani et al. (2002</xref>), which is implemented in the package PAM (<xref rid="pbio-0020108-Tibshirani2" ref-type="bibr">Tibshirani et al. 2003</xref>). For a brief description of the procedure, see “<xref ref-type="sec" rid="s4">Materials and Methods</xref>.”</p><p>Sometimes, however, a continuous predictor of survival is desired. We also describe a supervised version of principal components analysis that can be used to calculate a continuous risk score for a given patient and identify subtypes of cancer. The resulting predictor performs very well when applied to several published datasets.</p><p>These two methods will produce satisfactory results in most datasets. However, we will describe some variations of these methods that can sometimes improve their performance. When we cluster a dataset using only a subset of the genes, it is important that we choose the correct subset of genes. Choosing the genes with the largest Cox scores is generally a good strategy, but this procedure sometimes selects some spurious genes. We will show that one can use partial least squares (PLS) to compute a “corrected” Cox score. Selecting the genes with the largest “corrected” Cox scores can produce better clusters than selecting genes with largest raw Cox scores. Additionally, we will describe two other continuous predictors of survival that we will call β˜ and γ^. For some problems, they are better predictors than the continuous predictor based on supervised principal components (<xref ref-type="supplementary-material" rid="sg001">Figures S1–S3</xref>). These methods are described in <xref ref-type="supplementary-material" rid="sd018">Protocol S1</xref>.</p></sec><sec id="s1c"><title>Related Methods in the Literature</title><p> <xref rid="pbio-0020108-Ben-Dor1" ref-type="bibr">Ben-Dor et al. (2001</xref>) and <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) attempt to identify biologically meaningful tumor subtypes from gene expression data by clustering on a subset of the genes. The important distinction between these methods and our semi-supervised clustering method is that our method uses the available clinical data to choose the subset of the genes that is used to perform the clustering. The methods of <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) and <xref rid="pbio-0020108-Ben-Dor1" ref-type="bibr">Ben-Dor et al. (2001</xref>) do not use this clinical information. We will show that utilizing the available clinical data can improve the quality of the clustering.</p><p>There are also related methods for predicting the survival of cancer patients using gene expression data. <xref rid="pbio-0020108-Nguyen1" ref-type="bibr">Nguyen and Rocke (2002a</xref>) use a form of PLS to predict survival. <xref rid="pbio-0020108-Li1" ref-type="bibr">Li and Luan (2003</xref>) use support vector machines (SVMs). However, a drawback of these methods is the fact that they use a combination of all of the genes to predict survival. Since the vast majority of the genes in a given dataset are unrelated to survival, the result is that many of the inputs to the model are superfluous, which reduces the predictive accuracy of the model. We will show that our semi-supervised methods, which use only a subset of the genes, generally perform better than these methods.</p><p>Moreover, in many applications, we would like to identify which genes are the best predictors of survival. These genes could be analyzed in the laboratory to attempt to discover how they influence survival. They could also be used to develop a diagnostic test based on immunostaining or reverse transcriptase PCR. For these applications, it is important to have a predictor of survival that is based on a small subset of the genes. This is another important advantage of our methods over existing methods.</p><p> <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. (2002</xref>) utilized an ad hoc method that fit a series of univariate Cox proportional hazards models and took a linear combination of the resulting coefficients. A brief description of their method is given in <xref ref-type="supplementary-material" rid="sd018">Protocol S1</xref>. This method is similar to our methods in that it selects a relevant subset of genes by choosing the genes with the largest Cox scores. However, this method is a purely supervised procedure. It does not apply any unsupervised methods (such as clustering or principal components analysis) to this subset of genes to identify additional patterns in the data. We will show that our semi-supervised procedures generally perform better than this method.</p></sec><sec id="s1d"><title>Summary</title><p>Our goal is to identify subtypes of cancer that are both clinically relevant and biologically meaningful. Suppose that we have 𝓃 patients, and we measure the expression level of <italic>p</italic> genes for each patient. (Note that 𝓃 ≫ <italic>p</italic>.) We assume that there are several different types (classes) of cancer, each of which responds differently to treatment, and each of which is distinct at the molecular level. Therefore, given a set of 𝓃 patients with different classes of cancer, we wish to train a classifier that can diagnose which type of cancer a future patient has, given the expression levels of the patient's <italic>p</italic> genes. We will show that it is possible to identify such subgroups using the semi-supervised learning techniques described in the previous paragraph, and that identification of such subgroups can enable us to predict the clinical outcome of cancer more accurately.</p></sec></sec><sec id="s2"><title>Results</title><sec id="s2a"><title>Fully Unsupervised Clustering</title><p>As noted in the Introduction, we needed to assign each patient to a subgroup before we could apply nearest shrunken centroids. First, we applied an unsupervised 2-means clustering procedure to the DLBCL data of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>). This dataset consisted of measurements on 7399 genes from 240 patients. Of these 240 patients, 160 were used for training the model and 80 were reserved for validating the model. A survival time was given for each patient, which ranged between 0 and 21.8 y.</p><p>We compared the survival times of the two subgroups using a log-rank test. The log-rank test statistic was 0.7, with a corresponding <italic>p-</italic>value of 0.416. Thus, conventional clustering techniques failed to identify subgroups that differed with respect to their survival times. Subgroups identified using hierarchical clustering also did not differ with respect to survival (data not shown).</p></sec><sec id="s2b"><title>Using Clinical Data Alone to Generate Class Labels</title><p>We assigned each patient in the training data to either a “low-risk” or “high-risk” subgroup based on their survival time (see “<xref ref-type="sec" rid="s4">Materials and Methods</xref>” for details.) After applying nearest shrunken centroids with crossvalidation, we selected a model that used 249 genes. We then used this model to assign each patient in the independent test data to the “low-risk” or “high-risk” group. The plots of the two survival curves associated with the two classes generated by the model are shown in <xref ref-type="fig" rid="pbio-0020108-g002">Figure 2</xref>. The <italic>p</italic>-value of the log-rank test was 0.03.</p><fig id="pbio-0020108-g002" position="float"><label>Figure 2</label><caption><title>Comparison of the Survival Curves of the “Low-Risk” and “High-Risk” Groups</title><p>These were obtained by applying nearest shrunken centroids to the DLBCL test data. Patients in the training data were assigned to either the “low-risk” or “high-risk” group depending on whether or not their survival time was greater than the median survival time of all the patients.</p></caption><graphic xlink:href="pbio.0020108.g002"/></fig></sec><sec id="s2c"><title>Supervised Clustering</title><p>In order to identify tumor subclasses that were both biologically meaningful and clinically relevant, we applied a novel, supervised clustering procedure to the DLBCL data. We ranked all of the genes based on their univariate Cox proportional hazards scores, and performed clustering using only the “most significant” genes.</p><p>Recall that when we performed 2-means clustering on the patients in the test data using all 7,399 genes and used a log-rank test to compare the survival times of the patients in the two resulting clusters, the result was not significant. To test our new clustering method, we calculated the Cox scores of all 7,399 genes based on the 160 training observations and ranked the genes from largest to smallest based on their absolute Cox scores. We then clustered the 80 test observations using only the 25 top-scoring genes. This time, the log-rank statistic comparing the survival times of the two clusters was highly significant ( <italic>p</italic> = 0.0001). A plot of the two resulting survival curves is shown in <xref ref-type="fig" rid="pbio-0020108-g003">Figure 3</xref>. A plot of the survival curves that we obtained by applying 2-means clustering to all of the genes is also shown for comparison.</p><fig id="pbio-0020108-g003" position="float"><label>Figure 3</label><caption><title>Comparison of the Survival Curves Resulting from Applying Two Different Clustering Methods to the DLBCL Data</title></caption><graphic xlink:href="pbio.0020108.g003"/></fig></sec><sec id="s2d"><title>Other Clustering Methods</title><p>Both <xref rid="pbio-0020108-Ben-Dor1" ref-type="bibr">Ben-Dor et al. (2001</xref>) and <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) used a subset of the genes to try to cluster a microarray dataset in a biologically meaningful manner. They observed that clustering using a subset of the genes can produce better results than using all of the genes. However, these methods were still fully unsupervised since they used only the gene expression data to perform the clustering. They did not use the clinical data to identify subgroups.</p><p>Although these methods do a better job of identifying biologically meaningful clusters than clustering based on all of the genes, there is no guarantee that the clusters thus identified are associated with the clinical outcome of interest. Indeed, both <xref rid="pbio-0020108-Ben-Dor1" ref-type="bibr">Ben-Dor et al. (2001</xref>) and <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) applied their procedures to a small DLBCL dataset of 40 patients (<xref rid="pbio-0020108-Alizadeh1" ref-type="bibr">Alizadeh et al. 2000</xref>). The clusters they identified did not (with a few exceptions) differ significantly from one another with respect to survival.</p><p>We applied the clustering procedure of <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) to the larger DLBCL dataset of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>). <xref ref-type="fig" rid="pbio-0020108-g004">Figure 4</xref> shows the survival curves of the two clusters generated using this method. The survival curves generated by clustering on the genes with the largest Cox scores are included for comparison. Note that the two subgroups identified using the clustering procedure of <xref rid="pbio-0020108-von1" ref-type="bibr">von Heydebreck et al. (2001</xref>) do not differ significantly with respect to survival.</p><fig id="pbio-0020108-g004" position="float"><label>Figure 4</label><caption><title>Comparison of the Survival Curves Resulting from Applying Two Different Clustering Methods to the DLBCL Data</title></caption><graphic xlink:href="pbio.0020108.g004"/></fig></sec><sec id="s2e"><title>Survival Diagnosis</title><p>We showed that the cancer subgroups identified using this supervised clustering method can be used to predict survival in future patients. The idea is straightforward. First, we identified subgroups of patients using supervised clustering. Then we trained a nearest shrunken centroid classifier to predict the subgroup to which each patient belonged. Details are given in “<xref ref-type="sec" rid="s4">Material and Methods</xref>.”</p><p>We tested this procedure on the DLBCL data. A clustering based on 343 genes produced the smallest crossvalidation error rate, so we used a classifier based on this clustering to assign each of the 80 test patients to one of the two subgroups. The survival curves of the two predicted subgroups are shown in <xref ref-type="fig" rid="pbio-0020108-g005">Figure 5</xref>; the <italic>p</italic>-value of the log-rank test comparing the two survival curves is 0.008.</p><fig id="pbio-0020108-g005" position="float"><label>Figure 5</label><caption><title>Survival Curves for Clusters Derived from the DLBCL Data</title></caption><graphic xlink:href="pbio.0020108.g005"/></fig></sec><sec id="s2f"><title>Supervised Principal Components</title><p>We used a form of the principal components of the expression matrix to predict survival. Principal components analysis is an unsupervised learning technique that is used to reduce the dimensionality of a dataset by calculating a series of “principal components.” The hope is that the first few principal components will summarize a large percentage of the variability in the entire dataset. See <xref rid="pbio-0020108-Hastie2" ref-type="bibr">Hastie et al. (2001b</xref>) for a description of principal components analysis.</p><p>Unfortunately, principal components analysis suffers from the same limitations as purely unsupervised clustering. If we perform principal components analysis using all of the genes in a dataset, there is no guarantee that the resulting principal components will be associated with survival. Thus, we propose a semi-supervised form of principal components analysis that we call “supervised principal components.” Rather than using all of the genes when we perform principal components analysis, we use only a subset of the genes that are correlated with survival.</p><p>Using the 160 training observations, we computed the Cox scores for each gene. We kept the 17 genes with Cox scores of 2.39 or greater. We calculated the principal components of the training data using only these 17 genes. Then we approximated the principal components of the test data using <xref ref-type="disp-formula" rid="pbio-0020108-e011">equation (11)</xref> (see “<xref ref-type="sec" rid="s4">Materials and Methods</xref>” for details.)</p><p> <xref ref-type="fig" rid="pbio-0020108-g006">Figure 6</xref> shows that there does appear to be a correlation between the value of the first principal component, υ^I<sub>1</sub>, and patient survival. To confirm this observation, we fit a Cox proportional hazards model to a linear combination of υ^I<sub>1</sub> and υ^I<sub>2</sub>, the estimated first and second principal components of the test data, respectively. (See “<xref ref-type="sec" rid="s4">Materials and Methods</xref>” for a description of how this linear combination was obtained.) The resulting sum was a significant predictor of survival (<italic>R</italic> <sup>2</sup> = 0.113, likelihood ratio test statistic = 9.58, 1 d.f., <italic>p</italic> = 0.00197). This predictor is stronger than the discrete predictor shown in <xref ref-type="fig" rid="pbio-0020108-g005">Figure 5</xref> (<italic>R</italic> <sup>2</sup> = 0.08, likelihood ratio test statistic = 6.7, 1 d.f., <italic>p</italic> = 0.00966).</p><fig id="pbio-0020108-g006" position="float"><label>Figure 6</label><caption><title>Plot of Survival Versus the Predictor υ^I for the DLBCL Data</title></caption><graphic xlink:href="pbio.0020108.g006"/></fig></sec><sec id="s2g"><title>A Breast Cancer Example</title><p>Thus far, all of our examples have been based on the DLBCL data of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>). We now apply our methodology to a set of breast cancer microarray data. In a recent study, <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> built a model to predict the time to metastasis of breast cancer in patients based on microarray data from 78 patients. They showed that this model could be used to predict the times to metastasis of 20 independent test patients. Later, in a separate study, this same model was applied to a much larger set of 292 patients (<xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. 2002</xref>).</p><p>Unfortunately, the expression levels of only 70 genes were available for the 292 patient dataset, making it difficult to test our methodology. However, we were able to apply our supervised principal components method. The expression levels of approximately 25,000 genes were available for the earlier study (consisting of 78 patients). After applying crossvalidation, we selected a model consisting of eight genes, five of which were included among the 70 genes in the larger dataset. Thus, we fit a supervised principal components model using these five genes and applied it to the dataset of 292 patients.</p><p>The results are shown in <xref ref-type="table" rid="pbio-0020108-t001">Table 1</xref>. (To compare the predictive power of the various models, we fit a Cox proportional hazards model to each predictor and computed the <italic>R</italic> <sup>2</sup> statistic for each model. <italic>R</italic> <sup>2</sup> measures the percentage of the variation in survival time that is explained by the model. Thus, when comparing models, one would prefer the model with the larger <italic>R</italic> <sup>2</sup> statistic.) We see that our supervised principal components method produced a stronger predictor of metastasis than the procedure described in <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref>. Furthermore, our method used only five genes, whereas the predictor of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> used 70 genes. These results held even though we did not have the expression data for three genes that we would like to have included in our model.</p><table-wrap id="pbio-0020108-t001" position="float"><label>Table 1</label><caption><title>Supervised Principal Components Applied to Breast Cancer Data</title></caption><graphic xlink:href="pbio.0020108.t001"/><table-wrap-foot><fn id="nt101"><p>Comparison of the values of the <italic>R</italic> <sup>2</sup> statistic of the Cox proportional hazards model (and the <italic>p-</italic>value of the associated log-rank statistic) obtained by fitting the times to metastasis to our supervised principal components method and the discrete predictor described in <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> </p></fn></table-wrap-foot></table-wrap><p>(Of the 78 patients used to build the model in the original study, 61 were included in the larger dataset of 292 patients. Thus, the values of <italic>R</italic> <sup>2</sup> calculated using all 292 patients are inflated, since part of the dataset used to validate the model was also used to train the model. We include these results merely to demonstrate the greater predictive power of our methodology. Moreover, we repeated these calculations using only the 234 patients that were not included in the earlier study to ensure that our results were still valid.)</p></sec><sec id="s2h"><title>Comparison With Related Methods in the Literature</title><p>We compared each of our proposed methods to several previously published methods for predicting survival based on microarray data. In particular, we examined three previously published procedures: a method based on SVMs (<xref rid="pbio-0020108-Li1" ref-type="bibr">Li and Luan 2003</xref>), a method based on PLS (<xref rid="pbio-0020108-Nguyen1" ref-type="bibr">Nguyen and Rocke 2002a</xref>), and an ad hoc procedure that calculated a “risk index” for each patient by taking an appropriate linear combination of a subset of the genes (<xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. 2002</xref>). Finally, we considered a naive procedure that split the training data into two groups by finding the bipartition that minimized the <italic>p</italic>-value of the resulting log-rank statistic. A brief description of each of these procedures is given in <xref ref-type="supplementary-material" rid="sd018">Protocol S1</xref>; for a full description of these procedures, see the original papers.</p><p>We compared these methods on four different datasets (See <xref ref-type="supplementary-material" rid="sd001">Datasets S1-S13</xref>). First, we examined the DLBCL dataset (<xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. 2002</xref>) that we used in the earlier examples. Recall that there were 7,399 genes, 160 training patients, and 80 test patients. Second, we considered a breast cancer dataset (<xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>). There were 4,751 genes and 97 patients in this dataset. We partitioned this dataset into a training set of 44 patients and a test set of 53 patients. Third, we examined a lung cancer dataset (<xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. 2002</xref>). There were 7,129 genes and 86 patients, which we partitioned into a training set of 43 patients and a test set of 43 patients. Finally, we considered a dataset of acute myeloid leukemia patients (<xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. 2004</xref>). It consisted of 6,283 genes and 116 patients. This dataset was partitioned into a training set of 59 patients and a test set of 53 patients. The results are shown in <xref ref-type="table" rid="pbio-0020108-t002">Table 2</xref>.</p><table-wrap id="pbio-0020108-t002" position="float"><label>Table 2</label><caption><title>Comparison of the Different Methods on Four Datasets</title></caption><graphic xlink:href="pbio.0020108.t002"/><table-wrap-foot><fn id="nt201"><p>Comparison of the different methods applied to the DLBCL data of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>), the breast cancer data of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref>, the lung cancer data of Beer et al (2002), and the acute myeloid leukemia (AML) data of <xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. (2004</xref>). The methods are (1) assigning samples to a “low-risk” or “high-risk” group based on their median survival time; (2) using 2-means clustering based on the genes with the largest Cox scores; (3) using the supervised principal components method; (4) using 2-means clustering based on the genes with the largest PLS-corrected Cox scores; (5) using the continuous predictor ; (6) using 2-means clustering to identify two subgroups; (7) partitioning the training data into “low-risk” and “high-risk” subgroups by choosing the split that minimizes the <italic>p-</italic>value of the log-rank test when applied to the two resulting groups; (8) using SVMs, similar to the method of <xref rid="pbio-0020108-Li1" ref-type="bibr">Li and Luan (2003</xref>); (9) using a discretized version of (8); (10) Using partial least squares regression, similar to the method of <xref rid="pbio-0020108-Nguyen1" ref-type="bibr">Nguyen and Rocke (2002a</xref>); (11) using a discretized version of (11); (12) using the method of <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. (2002</xref>) </p></fn></table-wrap-foot></table-wrap></sec><sec id="s2i"><title>A Simulation Study</title><p>We compared each of the methods we proposed above on two simulated datasets. (See <xref ref-type="supplementary-material" rid="sd001">Data S1-S4</xref>.) The first simulated dataset <italic>X</italic> had 5,000 genes and 100 samples. All expression values were generated as standard normal random numbers with a few exceptions. Genes 1–50 in samples 1–50 had a mean of 1.0. We randomly selected 40% of the samples to have a mean of 2.0 in genes 51–100, 50% of the samples to have a mean of 1.0 in genes 101–200, and 70% of the samples to have a mean of 0.5 in genes 201–300.</p><p>We then generated survival times. The survival times of samples 1–50 were generated as normal random numbers with a mean of 10.0 and a standard deviation of 2.0, and the survival times of samples 51–100 were generated as normal random numbers with a mean of 8.0 and a standard deviation of 3.0. For each sample, a censoring time was generated as a normal random number with a mean of 10.0 and a standard deviation of 3.0. If the censoring time turned out to be less than the survival time, the observation was considered to be censored. Finally, we generated another 5000 × 100 matrix of test data X˜I, which was generated the same way <italic>X</italic> was generated. Survival times for X˜I were also generated in an identical manner.</p><p>We defined samples 1–50 as belonging to “tumor type 1” and samples 51–100 as belonging to “tumor type 2.” Thus, a successful subgroup discovery procedure should assign samples 1–50 to one subgroup, and samples 51–100 to the other subgroup.</p><p>We applied the methods discussed above to identify these subgroups (and predict survival) for the simulated dataset. This simulation was repeated ten times. The results are shown in <xref ref-type="table" rid="pbio-0020108-t003">Table 3</xref>. The first column of the table shows how many samples were misclassified when the dataset was originally divided into two subgroups. The second column shows the number of crossvalidation errors that occurred when the nearest shrunken centroids model was applied to these putative class labels. The third column shows the number of incorrectly labeled samples when the optimal nearest shrunken centroids model was used to assign labels to the samples in the test data X˜I. The final column is the value of <italic>R</italic> <sup>2</sup> obtained by fitting a Cox proportional hazards model to the predicted class labels for the test data (or by fitting a Cox model to γ^ in the case of methods 4 and 6).</p><table-wrap id="pbio-0020108-t003" position="float"><label>Table 3</label><caption><title>Comparison of the Different Methods on Our Simulated Data</title></caption><graphic xlink:href="pbio.0020108.t003"/><table-wrap-foot><fn id="nt203"><p>The methods are (1) assigning samples to a “low-risk” or “high-risk” group based on their median survival time; (2) using 2-means clustering to identify two subgroups; (3) using 2-means clustering based on the genes with the largest Cox scores; (4) using the supervised principal components method; (5) using 2-means clustering based on the genes with the largest PLS-corrected Cox scores; (6) using the continuous predictor . Each entry in the table represents the mean over 10 simulations; the standard error is given in parentheses </p></fn></table-wrap-foot></table-wrap><p>In the first simulation, we found that the fully supervised and the fully unsupervised methods produced much worse results than the semi-supervised methods. (For each iteration of the “median cut” method, the crossvalidation error was minimized when all of the observations were assigned to the same class. Hence, each such model had no predictive power, and the value of <italic>R</italic> <sup>2</sup> was zero for each iteration. If we had chosen a smaller value of the tuning parameter Δ, the procedure would have performed better, although not significantly better.) The continuous predictor based on supervised principal components performed nearly as well as the methods based on semi-supervised clustering.</p><p>Next, we performed a second simulation. The second simulated dataset <italic>X</italic> had 1000 genes and 100 samples. All expression values were generated as Gaussian random variables with a mean of zero and a variance of 1.5, although again there were a few exceptions. Genes 1–50 had a mean of 0.5 in samples 1–20, a mean of 0.6 in samples 21–40, a mean of 0.7 in samples 41–60, a mean of 0.8 in samples 61–80, and a mean of 0.9 in samples 81–100. And again, we randomly selected 40% of the samples to have a mean of 2.0 in genes 51–100, 50% of the samples to have a mean of 1.0 in genes 101–200, and 70% of the samples to have a mean of 0.5 in genes 201–300. To generate the survival time of each “patient,” we took the sum of the expression levels of the first 50 genes and added a Gaussian noise term with variance 0.01. There was no censoring mechanism for the second simulation. We also generated another 1000 × 100 matrix of test data using an analogous procedure.</p><p>Under this model, there are actually five “tumor subgroups.” However, we still used 2-means clustering on this simulated dataset in order to evaluate the performance of our methods when the number of clusters is chosen incorrectly. Thus, in this simulation, it does not make sense to talk about the number of “misclassification errors;” we can only compare the methods on the basis of their predictive ability.</p><p>We applied the six different methods to this new simulated dataset and repeated this simulation ten times; the results are shown in <xref ref-type="table" rid="pbio-0020108-t004">Table 4</xref>. The supervised principal component method is the clear winner in the second simulation study. The semi-supervised methods performed poorly because there were a large number of subgroups and there was a considerable overlap between subgroups. This example demonstrates that the supervised principal component method performs well regardless of the number of tumor subclasses and that it seems to perform especially well when survival is an additive function of the expression level of certain genes.</p><table-wrap id="pbio-0020108-t004" position="float"><label>Table 4</label><caption><title>Comparison of the Different Methods on Our Simulated Data</title></caption><graphic xlink:href="pbio.0020108.t004"/><table-wrap-foot><fn id="nt301"><p>The methods are (1) assigning samples to a “low-risk” or “high-risk” group based on their median survival time; (2) using 2-means clustering to identify two subgroups; (3) using 2-means clustering based on the genes with the largest Cox scores; (4) using the supervised principal components method; (5) using 2-means clustering based on the genes with the largest PLS-corrected Cox scores; (6) using the continuous predictor . Each entry in the table represents the mean over 10 simulations; the standard error is given in parentheses </p></fn></table-wrap-foot></table-wrap></sec></sec><sec id="s3"><title>Discussion</title><p>One important goal of microarray research is to develop more powerful diagnostic tools for cancer and other diseases. Consider a hypothetical cancer that has two subtypes. One subtype is known to spread much more rapidly than the other subtype, and hence must be treated much more aggressively. We would like to be able to diagnose which type of cancer patients have and give them the appropriate treatment.</p><p>If it is known that two such subtypes of a certain cancer exist, and if we have a training set where it is known which patients have which subtype, then we can use nearest shrunken centroids or other classification methods to build a model to diagnose this cancer in future patients. However, in many cases, we do not know how many subtypes are present, nor do we know which patients belong to which subgroup. Thus, it is important to develop methods to identify such subgroups.</p><p>Unsupervised methods, such as hierarchical clustering, are popular techniques for identifying such subgroups. However, there is no guarantee that subgroups discovered using unsupervised methods will have clinical significance.</p><p>An alternative is to generate class labels using clinical data. The simplicity of the approach of dividing the patients into two subclasses based on their survival time is attractive, and there is evidence that this procedure can successfully predict survival. Indeed, this procedure produced a significant predictor of survival in four different datasets, suggesting that this approach has some utility. However, as noted in the Introduction, subgroups identified in this manner may not be biologically meaningful. When we applied this model to the DLBCL data described earlier, the misclassification error rate for the shrunken centroids model was very high (around 40%), so a diagnosis based on this procedure is likely to be inaccurate.</p><p>Supervised clustering methods can overcome these problems. We have seen that if we selected significant genes prior to clustering the data, we could identify clusters that were clinically relevant. We have also seen how knowledge of these clusters could be used to diagnose future patients.</p><p>This supervised clustering methodology is a useful prognostic tool. It is also easy to interpret. However, it has certain shortcomings as well. Recall our conceptual model shown in <xref ref-type="fig" rid="pbio-0020108-g001">Figure 1</xref>. Patients with tumor type 2 live longer than patients with tumor type 1 on average, but there is still significant variability within each tumor type. Even if we can diagnose a patient with the correct tumor type 100% of the time, the prognosis of the patient may be inaccurate if the variability in survival time within each tumor type is large. Thus, it would be desirable to find a continuous predictor of survival that accounts for this within-group variability.</p><p>One possible such predictor is our supervised principal components procedure. This procedure used the principal components of a subset of the expression matrix <italic>X</italic> as a predictor of patient survival. The chosen subset contained the genes with the largest Cox scores. This method could also be used to detect cancer subtypes, since the principal components will presumably capture the variation that exists between subtypes. It is also capable of identifying variation within these subtypes, which, as discussed above, cannot be identified using supervised clustering. We showed that this procedure could produce a stronger predictor of survival than the discrete predictor based on supervised clustering.</p><p>We compared our methods to several previously published methods for predicting survival based on microarray data. In general, our methods performed significantly better than these existing methods. In particular, our supervised principal components method gave the best results on three of the four datasets. (It performed slightly worse than our γ^ method on the DLBCL data, but it still outperformed almost all of the other methods.) Furthermore, each of our proposed methods was a significant predictor of survival (at <italic>p</italic> = 0.05) for all four datasets, which was not true for any of the other methods. Finally, if we consider only discrete predictors of survival, our semi-supervised clustering methods performed better than the other models on at least three of the four datasets.</p><p>Another important advantage of our methods is that they select a subset of the genes to use as predictors. The methods of <xref rid="pbio-0020108-Nguyen1" ref-type="bibr">Nguyen and Rocke (2002a</xref>) and <xref rid="pbio-0020108-Li1" ref-type="bibr">Li and Luan (2003</xref>), by contrast, require the use of all (or a large number) of the genes. If we can identify a small subset of genes that predict the survival of cancer patients, it may be possible to develop a diagnostic test using immunostaining or reverse transcriptase PCR. However, such a test would not be feasible if hundreds or thousands of genes were necessary to make the diagnosis.</p><p>Throughout this study, we have used survival data to help us identify genes of interest. However, other clinical variables could also be used, such as the stage of the tumor, or whether or not it has metastasized. Rather than ranking genes based on their Cox scores, one would use a different metric to measure the association between a given gene and the clinical variable of interest. For example, suppose we wished to identify a subgroup of cancer that was associated with a high risk of metastasis. For each gene, we could compute a t-statistic comparing the expression levels in the patients whose cancer metastasized to those in the patients with no metastasis. <xref rid="pbio-0020108-Tusher1" ref-type="bibr">Tusher et al. (2001</xref>) described methods for generating such “significant gene lists” for a variety of possible clinical variables. Many of these methods are implemented in the significance analysis of microarrays software package (<xref rid="pbio-0020108-Chu1" ref-type="bibr">Chu et al. 2002</xref>).</p><p>Information about the risk of metastasis (and death) for a given patient is essential to treat cancer successfully. If the risk of metastasis is high, the cancer must be treated aggressively; if the risk is low, milder forms of treatment can be used. Using DNA microarrays, researchers have successfully identified subtypes of cancer that can be used to assess a patient's risk profile. Our results show that semi-supervised learning methods can identify these subtypes of cancer and predict patient survival better than existing methods. Thus, we believe they can be a powerful tool for diagnosing and treating cancer and other genetic diseases.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><sec id="s4a1"><title>Overview of nearest shrunken centroids</title><p>The nearest shrunken centroids procedure calculates the mean expression of each gene within each class. Then it shrinks these centroids toward the overall mean for that gene by a fixed quantity, Δ. Diagonal linear discriminant analysis (LDA) is then applied to the genes that survive the thresholding. Details are given in <xref rid="pbio-0020108-Tibshirani1" ref-type="bibr">Tibshirani et al. (2002</xref>). It has successfully classified tumors based on gene expression data in previous studies. In one experiment, there were a total of 88 patients, each of which had one of four different types of small round blue cell tumors . Nearest shrunken centroids classified 63 training samples and 25 test samples without a single misclassification error (<xref rid="pbio-0020108-Tibshirani1" ref-type="bibr">Tibshirani et al. 2002</xref>).</p></sec><sec id="s4a2"><title>Generation of “median cut” class labels<italic/> </title><p>We created two classes by cutting the survival times at the median survival time (2.8 y). Any patient who lived longer than 2.8 y was considered to be a “low-risk” patient, and any patient that lived less than 2.8 y was considered to be a “high-risk” patient. In this manner, we assigned a class label to each observation in the training data.</p><p>Unfortunately, many of the patients' survival times were censored, meaning that the individual left the study before the study was completed. When this occurs, we do not know how long the patient survived; we only know how long the patient remained in the study prior to being lost to follow-up.</p><p>If an observation is censored, we may not know to which class it belongs. For example, suppose that the median survival time is 2.8 y, but that a patient left the study after 1.7 y. If the patient died in the interval between 1.7 y and 2.8 y, then the patient should be assigned to the “high-risk” group. Otherwise, the patient should be assigned to the “low-risk” group. However, there is no way to determine which possibility is correct.</p><p>Based on the Kaplan-Meier survival curve for all the patients, we can estimate the probability that a censored case survives a specified length of time (<xref rid="pbio-0020108-Cox1" ref-type="bibr">Cox and Oakes 1984</xref>; <xref rid="pbio-0020108-Therneau1" ref-type="bibr">Therneau and Grambsch 2000</xref>). For example, suppose that the median survival time is 50 months and a patient left the study after 20 months. Let <italic>T</italic> denote the survival time of this patient. Then, using the Kaplan-Meier curve, we can estimate <italic>p</italic>(<italic>T</italic>>50) and <italic>p</italic>(<italic>T</italic>>20). Then we can estimate <italic>p</italic>(<italic>T</italic>>50\|<italic>T</italic>>20) as follows:</p><p> <disp-formula id="pbio-0020108-e001"><graphic xlink:href="pbio.0020108.e001.jpg" mimetype="image" position="float"/></disp-formula> </p><p>and, of course,</p><p> <disp-formula id="pbio-0020108-e002"><graphic xlink:href="pbio.0020108.e002.jpg" mimetype="image" position="float"/></disp-formula> </p><p>In this manner, we can estimate the probability that each censored observation belongs to the “low-risk” and “high-risk” classes, respectively.</p><p>However, it is still unclear how we would train our classifier based on this information. Nearest shrunken centroids is a modified version of LDA. It is described in detail in <xref rid="pbio-0020108-Hastie2" ref-type="bibr">Hastie et al. (2001b</xref>). Like most classification techniques, LDA assumes that the class labels of the training observations are known with complete certainty. The version of LDA described in <xref rid="pbio-0020108-Hastie2" ref-type="bibr">Hastie et al. (2001b</xref>) and most other books cannot handle probabilistic class labels, where there is a certain probability that a training observation belongs to one class, and a certain probability that it belongs to a different class. We will now describe a simple modification of LDA that can be trained based on this type of data. It is similar to a technique described in <xref rid="pbio-0020108-McLachlan1" ref-type="bibr">McLachlan (1992</xref>) for training an LDA classifier when some of the training observations are missing.</p><p>Let {<bold>x</bold> <sub><italic>i</italic></sub>}<inline-formula id="pbio-0020108-e013"><inline-graphic xlink:href="pbio.0020108.e013.jpg" mimetype="image"/></inline-formula> denote the set of input variables, and let {<bold>y</bold> <sub><italic>i</italic></sub>}<inline-formula id="pbio-0020108-e014"><inline-graphic xlink:href="pbio.0020108.e014.jpg" mimetype="image"/></inline-formula> represent the corresponding response variables. Also, let <italic>g</italic> represent the number of discrete classes to which the <italic>y</italic> <sub><italic>i</italic></sub>s may belong. (If we are dividing the training data into “low-risk” and “high-risk” patients, then <italic>g</italic> = 2.) When we perform LDA when all of the <italic>y</italic> <sub><italic>i</italic></sub>s are known, the problem is to fit the mixture model </p><p> <disp-formula id="pbio-0020108-e003"><graphic xlink:href="pbio.0020108.e003.jpg" mimetype="image" position="float"/></disp-formula> </p><p>(Generally, each 𝒻<sub><italic>i</italic></sub> is a Gaussian density function, and the θ<sub><italic>i</italic></sub>s correspond to the mean of the observations in each class. The π<sub><italic>i</italic></sub>s correspond to “prior” probabilities that an observation belongs to class 𝒾.) In this case, we must fit this model on the basis of the classified (uncensored) training data, which we denote by <bold>t</bold>, and the unclassified (censored) feature vectors <bold>x</bold> <sub>𝒿</sub> ( 𝒿 = 𝓃+1, …,𝓃+𝓂), which we denote by <bold>t</bold> <sub>𝓊</sub>. (Also, note that Φ = (π′,θ′)′ denotes the vector of all unknown parameters.)</p><p>We define the latent variables 𝓏<sub><italic>ij</italic></sub> to be equal to one if the 𝒿th observation belongs to the 𝒾th class, and zero otherwise. Then the complete-data log likelihood is</p><p> <disp-formula id="pbio-0020108-e004"><graphic xlink:href="pbio.0020108.e004.jpg" mimetype="image" position="float"/></disp-formula> </p><p>The EM algorithm is applied to this model by treating <bold>z</bold> <sub><italic>j</italic></sub>(𝒿 = 𝓃+1,…, 𝓃+𝓂) as missing data. It turns out to be very simple in the case of LDA. The E-step is effected here simply by replacing each unobserved indicator variable 𝓏<sub><italic>ij</italic></sub> by its expectation conditional on <bold>x</bold> <sub><italic>j</italic></sub>. That is, 𝓏<sub><italic>ij</italic></sub> is replaced by the estimate of the posterior probability that the 𝒿th entity with feature vector <bold>x</bold> <sub><italic>j</italic></sub> belongs to <italic>G</italic> <sub><italic>i</italic></sub>(𝒾 = 1, …,<italic>G</italic>, 𝒿 = <italic>n</italic> + 1, …,<italic>n</italic>+<italic>m</italic>) (<xref rid="pbio-0020108-McLachlan1" ref-type="bibr">McLachlan 1992</xref>). We take the initial estimates of 𝓏<sub><italic>ij</italic></sub> to be the earlier estimate that the 𝒾th censored observation belongs to class 𝒿 based on the Kaplan-Meier curve.</p><p>The estimates of π<sub><italic>i</italic></sub> and μ<sub><italic>i</italic></sub> in the M-step are equally simple:</p><p> <disp-formula id="pbio-0020108-e005"><graphic xlink:href="pbio.0020108.e005.jpg" mimetype="image" position="float"/></disp-formula> </p><p>and</p><p> <disp-formula id="pbio-0020108-e006"><graphic xlink:href="pbio.0020108.e006.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where</p><p> <disp-formula id="pbio-0020108-e007"><graphic xlink:href="pbio.0020108.e007.jpg" mimetype="image" position="float"/></disp-formula> </p><p>In these expressions, τ<sub><italic>i</italic></sub>(<bold>x</bold>;Φ) is the posterior probability that the 𝒿th entity with feature vector <bold>x</bold> <sub><italic>j</italic></sub> belongs to <italic>G</italic> <sub><italic>i</italic></sub>, or, in other words,</p><p> <disp-formula id="pbio-0020108-e008"><graphic xlink:href="pbio.0020108.e008.jpg" mimetype="image" position="float"/></disp-formula> </p><p>We continue these imputations until the algorithm converges. In practice, one imputation seems to be sufficient for most problems, since each imputation is computationally intensive, and additional imputations did not seem to change the results significantly.</p></sec><sec id="s4a3"><title>Diagnosing patient survival via supervised clustering<italic/> </title><p>We calculated the Cox scores of each gene based on the 160 training observations, and obtained a list of the most significant genes. Then we performed 2-means clustering on these 160 observations using the genes with the largest absolute Cox scores and obtained two subgroups. We repeated this procedure multiple times with different numbers of genes. For each such clustering, we trained a nearest shrunken centroid classifier to assign future patients to one subgroup or the other and examined the crossvalidation error rate.</p><p>The problem of choosing the number of genes on which to perform the clustering is more complicated than it appears. The obvious way to choose the optimal number of genes on which to cluster is to simply minimize the crossvalidation error rate of the nearest shrunken centroids model based on the clustering. This works up to a certain point. It is possible that the clustering procedure will identify a cluster that is unrelated to survival. (Since we are clustering on the genes with the highest Cox scores, this is unlikely to occur. However, it is still possible, especially if the number of genes on which we are clustering is large.) Thus, we needed to build a safeguard against this possibility into our procedure. After performing clustering based on a given set of high-scoring genes, we performed a log-rank test to determine if the resulting clusters differed with respect to survival. If they did not, the clustering was discarded without further analysis. An outline of the procedure follows: (1) Choose a set <italic>G</italic> of possible values of Γ. (2) Let <italic>p</italic> <sub>min</sub> = 1 and <italic>e</italic> <sub>min</sub> = 1. (3) For each Γ in <italic>G</italic>, do the following: (4) Perform <italic>k</italic>-means clustering using only those genes with absolute Cox scores greater than Γ. (5) Perform a log-rank test to test the hypothesis that the <italic>k</italic> clusters have different survival rates. Call the <italic>p</italic>-value of this test <italic>p</italic>. (6) If <italic>p</italic>≥<italic>p</italic> <sub>min</sub>, then return to step 3. (7) Fit a nearest shrunken centroids model based on the clusters obtained in step 3. Calculate the minimum crossvalidation error rate across all values of the shrinkage parameter, and call it <italic>e</italic>. (8) If <italic>e</italic><<italic>e</italic> <sub>min</sub>, then let Γ<sub>best</sub> = Γ, and return to step 3. Otherwise return to step 3 without changing the value of Γ<sub>best</sub>. The optimal value of Γ is taken to be the value of Γ<sub>best</sub> when this procedure terminates.</p><p>Several comments about this procedure are in order. First, note that we did not recalculate the Cox scores at each fold of the crossvalidation procedure. We calculated them only once, using all of the patients in the dataset. There are several reasons for doing this. Recalculating the Cox scores at each fold would be extremely expensive computationally. Moreover, we found that the Cox score of a given gene varied depending on the number of patients (and which patients) we included in the model. Thus, if a given value of Γ produced a low crossvalidation error rate, there was no guarantee that a model based on the full dataset using this value of Γ would produce good results, since the model based on the full dataset may use a different list of genes. Other studies have found that using the entire dataset to produce a “significant gene list” prior to performing crossvalidation can produce more accurate predictions (<xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. 2002</xref>).</p><p>Also, the set <italic>G</italic> was left unspecified in the procedure. The choice of which (and how many) possible values of Γ to include in <italic>G</italic> depends on the problem at hand, as well as the computational power available. As a default, we recommend trying 100 evenly spaced values of Γ between the 90th percentile of the Cox scores and the maximum of the Cox scores. However, the optimal Γ<sub>best</sub> varies greatly from dataset to dataset, so we recommend trying several different forms of <italic>G</italic> if adequate computing power exists.</p><p>Furthermore, note that when we calculated the <italic>p</italic>-value of the log-rank test after performing the original clustering, we insisted not only that the <italic>p</italic>-value be significant, but also that it be lower than the best <italic>p</italic>-value obtained thus far. The reasons for this are twofold. First, experience suggests that if a given set of genes produces a good clustering on the training data (“good” defined as having a low <italic>p</italic>-value from a log-rank test), then it is likely to produce a good clustering on the test data. (We offer no theoretical or biological justification for this statement; it simply represents our experience. However, we have observed this result a sufficient number of times to convince us that it is not coincidental.) Moreover, this speeds up the algorithm substantially. Calculating the nearest shrunken centroids crossvalidation error rate for a given clustering is the slowest part of the procedure; the time required to perform the clustering and calculate the log-rank statistic is insignificant in comparison. Thus, by only considering clusterings which produce a log-rank statistic with a small <italic>p</italic>-value, we allow the set <italic>G</italic> to be much larger than would be feasible otherwise.</p><p>Finally, the number of clusters <italic>k</italic> was unspecified in the procedure. We have experimented with some algorithms to choose the value of <italic>k</italic> automatically, but without success. If possible, we recommend that the value of <italic>k</italic> be chosen based on prior biological knowledge. (Perhaps one could first perform hierarchical clustering, examine a dendogram of the data, and visually search for major subgroups.) If this is not possible, we recommend trying several different small values of <italic>k</italic> and choosing the one that gives the best results. (Our experience suggests that choosing <italic>k</italic> = 2 will give good results for almost all datasets.)</p></sec><sec id="s4a4"><title>Supervised principal components<italic/> </title><p>As above, let <italic>X</italic> be the <italic>p</italic>×<italic>n</italic> matrix of expression values, for <italic>p</italic> genes and <italic>n</italic> patients. Let <italic>x</italic> <sub><italic>ij</italic></sub> denote the expression level of the 𝒾th gene in the 𝒿th patient. Assume that each patient has one of two possible underlying tumor types. Without loss of generality, assume that patients 1, …,<italic>m</italic> have tumor type 1, and that patients <italic>m</italic> + 1,…,<italic>n</italic> have tumor type 2. Then assume that the genetic profiles of the two tumor types are distinct from one another, which is equivalent to assuming that the joint distribution of (<italic>x</italic> <sub>1<italic>j</italic></sub>, …,<italic>x</italic> <sub><italic>pj</italic></sub>) is different for 1 ≤ 𝒿 ≤ 𝓂 than it is for 𝓂 + 1 ≤ 𝒿 ≤ 𝓃. Thus, if we choose constants {<italic>a</italic> <sub><italic>i</italic></sub>}<inline-formula id="form3"><sup><italic>p</italic></sup><sub><italic>i</italic>=1</sub></inline-formula>, the distribution of ∑<inline-formula id="form4"><sup><italic>p</italic></sup><sub><italic>j</italic>=1</sub></inline-formula>  <italic>a<sub>j</sub>x<sub>ij</sub></italic> will be different for 1 ≤ 𝒿 ≤ 𝓂 than it is for 𝓂 + 1 ≤ 𝒿 ≤ 𝓃. (Obviously, this is not true for all values of {<italic>a<sub>i</sub></italic>}<inline-formula id="form5"><sup><italic>p</italic></sup><sub><italic>i</italic></sub>=1</inline-formula>. For example, if we let <italic>a<sub>i</sub></italic> = 0 for all 𝒾, then this statement will not hold. However, it will generally be true unless we deliberately choose a pathological set {<italic>a<sub>i</sub></italic>}<inline-formula id="form6"><sup><italic>p</italic></sup><sub><italic>i</italic></sub>=1</inline-formula>.)</p><p>In particular, consider the singular value decomposition of <italic>X</italic>:</p><p> <disp-formula id="pbio-0020108-e009"><graphic xlink:href="pbio.0020108.e009.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where <italic>U</italic> is a <italic>p</italic> × 𝓃 orthogonal matrix, <italic>D</italic> is an 𝓃 × 𝓃 diagonal matrix, and <italic>V</italic> is an 𝓃 × 𝓃 orthogonal matrix (<xref rid="pbio-0020108-Horn1" ref-type="bibr">Horn and Johnson 1985</xref>). Then the matrix <italic>V</italic> can be written as</p><p> <disp-formula id="pbio-0020108-e010"><graphic xlink:href="pbio.0020108.e010.jpg" mimetype="image" position="float"/></disp-formula> </p><p>In other words, for a given column of <italic>V</italic>, each row of <italic>V</italic> is a linear combination of the expression values in the corresponding column of <italic>X</italic>. Thus, by the argument in the preceding paragraph, rows 1 through 𝓂 should have a different distribution than rows 𝓂 + 1 through 𝓃. Hence, we propose that the first few columns of <italic>V</italic> be used as continuous predictors of survival for each patient. Formally,</p><p> <disp-formula id="pbio-0020108-e011"><graphic xlink:href="pbio.0020108.e011.jpg" mimetype="image" position="float"/></disp-formula> </p><p>Moreover, suppose that we have an independent test set X˜I. Then let</p><p> <disp-formula id="pbio-0020108-e012"><graphic xlink:href="pbio.0020108.e012.jpg" mimetype="image" position="float"/></disp-formula> </p><p>where <italic>U</italic> and <italic>D</italic> are the same as in <xref ref-type="disp-formula" rid="pbio-0020108-e011">equation (11)</xref> (i.e., derived from the singular value decomposition of the training data). In this case, the first few columns of V^I can be used to estimate survival for the patients in the independent test set.</p><p>The reason for choosing the first few columns of <italic>V</italic> is because the matrix <italic>U</italic> was chosen so that <italic>X<sup>T</sup>u</italic> <sub>1</sub> has the largest sample variance amongst all normalized linear combinations of the rows of <italic>X</italic> (<xref rid="pbio-0020108-Hastie2" ref-type="bibr">Hastie et al. 2001b</xref>). (Here, <italic>u</italic> <sub>1</sub> represents the first column of <italic>U</italic>.) Hence, assuming that variation in gene expression accounts for variation in survival, we would expect that <italic>X<sup>T</sup></italic> <sub><italic>u</italic>1</sub> captures a large percentage of the variation in survival. (Indeed, in some simple models, it can be proven that <xref ref-type="disp-formula" rid="pbio-0020108-e011">equation [11]</xref> is the best possible predictor of survival; see <xref ref-type="supplementary-material" rid="sd018">Protocol S1</xref>.)</p><p>In theory, we could calculate <italic>V</italic> using the entire dataset <italic>X</italic>, and the rows of <italic>V</italic> would have different distributions depending on the tumor type of the corresponding patient. In practice, however, many of the genes in <italic>X</italic> are unrelated to survival, and if we use the entire dataset <italic>X</italic> to compute <italic>V</italic>, the quality of the resulting predictor is poor. We can overcome this difficulty by using only the genes with the largest Cox scores. Formally, we construct a matrix <italic>X</italic>′ consisting of only those genes whose Cox scores are greater than some threshold Γ, and take the singular value decomposition of <italic>X</italic>′.</p><p>To choose the optimal value of Γ, we employ the following procedure: (1) Choose a set <italic>G</italic> of possible values of Γ. (2) For each Γ in <italic>G</italic>, split the training data into <italic>k</italic> random partitions (i.e., perform <italic>k</italic>-fold crossvalidations). For most problems (and for the rest of this discussion), we can let <italic>k</italic> = 10. (3) For each crossvalidation fold, take a singular value decomposition of <italic>X</italic>, leaving out one of the 10 partitions for validation purposes. Use only those genes with absolute Cox scores greater than Γ. (4) Calculate υ^ for the 10% of the data that was withheld, as described above. (5) Fit a Cox proportional hazards model to υ^, and calculate the chi-square statistic for the log-rank test associated with this model. Denote the chi-square statistic for the 𝒾th crossvalidation fold by 𝓌<sub><italic>i</italic></sub>. (6) Average the 𝓌<sub><italic>i</italic></sub>s over the 10 crossvalidation folds. Call this average 𝓌<sub>Γ</sub>. (7) If 𝓌<sub>Γ</sub> is greater than the value of 𝓌<sub>Γ∗</sub>, then let Γ∗ = Γ and 𝓌<sub>Γ∗</sub> = 𝓌<sub>Γ</sub>. (8) Return to step 2. The set <italic>G</italic> is left unspecified in the procedure. As a default, we recommend trying 30 evenly spaced values of Γ between the 90th percentile of the Cox scores and the maximum of the Cox scores, although this recommendation is somewhat arbitrary.</p><p>In some cases, we can improve the predictive power of our model by taking a linear combination of several columns of <italic>V</italic> (rather than simply taking the first column of <italic>V</italic>). Suppose we wish to find a predictor based on the first <italic>k</italic> columns of <italic>V</italic>. We can perform the following procedure: (1) Let <italic>X</italic> denote the training data. Take the singular value decomposition of <italic>X</italic> = <italic>UDV<sup>T</sup></italic> as described above (after selecting an appropriate subset of the genes). (2) Fit a Cox proportional hazards model using the first <italic>k</italic> columns of <italic>V</italic> as predictors. (3) Calculate the matrix V^I for the test data using <xref ref-type="disp-formula" rid="pbio-0020108-e012">equation (12)</xref> above. (4) Take a linear combination of the first <italic>k</italic> columns of V^I using the Cox regression coefficients obtained in step 2. Use the resulting sum as a continuous predictor of survival.</p></sec><sec id="s4a5"><title>Software and computational details<italic/> </title><p>All computations in this study were performed using the R statistical package, which is available on the Internet at <ext-link ext-link-type="uri" xlink:href="http://cran.r-project.org/">http://cran.r-project.org/</ext-link>. R source code for the procedures described in this paper are available from the authors upon request (see also <xref ref-type="supplementary-material" rid="sd001">Data S1–Data S4</xref>). These methods will also be implemented in a future version of the PAM microarray analysis package (<xref rid="pbio-0020108-Tibshirani2" ref-type="bibr">Tibshirani et al. 2003</xref>). (The “median cut” method has been implemented in version 1.20, which is now available.)</p></sec></sec></sec><sec sec-type="supplementary-material" id="s5"><title>Supporting Information</title><supplementary-material content-type="local-data" id="sd001"><label>Data S1</label><caption><title>Documentation of Our R Functions</title><p>This file contains a brief description of the functions contained in the semi-super.R file.</p><p>(1 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd001.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd002"><label>Data S2</label><caption><title>R Source Code</title><p>This file contains R functions for implementing the procedures we have described in our study.</p><p>(6 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd002.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd003"><label>Data S3</label><caption><title>Source Code for Simulation Study 1</title><p>This file contains the R source code that we used to perform the first simulation study in our paper.</p><p>(31 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd003.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd004"><label>Data S4</label><caption><title>Source Code for Simulation Study 2</title><p>This file contains the R source code that we used to perform the second simulation study in our paper.</p><p>(39 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd004.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd005"><label>Dataset S1</label><caption><title>Breast Cancer Expression Data: <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> Study</title><p>The gene expression data for the breast cancer study of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref>. We include only the expression levels of 4,751 genes identified in the study whose expression varied</p><p>(2.9 MB CSV).</p></caption><media xlink:href="pbio.0020108.sd005.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd006"><label>Dataset S2</label><caption><title>Breast Cancer Gene Names: <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> Study</title><p>The names of each of the 4,751 genes in the study of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref>.</p><p>(74 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd006.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd007"><label>Dataset S3</label><caption><title>Breast Cancer Survival Data: <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref> Study</title><p>The clinical data for the study of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. (2002)</xref>. The first column represents the time until metastasis (or the time until the patient left the study); the second column is 1 if the tumor metastasized and 0 if it did not.</p><p>(1 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd007.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd008"><label>Dataset S4</label><caption><title>Breast Cancer Expression Data: <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref> Study</title><p>The gene expression data for the 70 genes in the breast cancer study of <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref>.</p><p>(141 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd008.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd009"><label>Dataset S5</label><caption><title>Breast Cancer Gene Names: <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref> Study</title><p>The names of the 70 genes in the study of <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref>.</p><p>(1 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd009.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd010"><label>Dataset S6</label><caption><title>Repeated Breast Cancer Samples</title><p>A single column that is 1 if the patient was included in the earlier study (that of <xref rid="pbio-0020108-vant1" ref-type="bibr">van't Veer et al. [2002</xref>]), and 0 if the patient was not included in the earlier study.</p><p>(1 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd010.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd011"><label>Dataset S7</label><caption><title>Breast Cancer Survival Data: <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref> Study</title><p>The clinical data for the study of <xref rid="pbio-0020108-vannde1" ref-type="bibr">van de Vijver et al. (2002)</xref>. The format is the same as the format of the earlier file of clinical data.</p><p>(5 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd011.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd012"><label>Dataset S8</label><caption><title>DLBCL Expression Data</title><p>The gene expression data for the DLBCL study of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>).</p><p>(24.38 MB CSV).</p></caption><media xlink:href="pbio.0020108.sd012.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd013"><label>Dataset S9</label><caption><title>DLBCL Survival Data</title><p>The clinical data for the study of <xref rid="pbio-0020108-Rosenwald1" ref-type="bibr">Rosenwald et al. (2002</xref>). The format is the same as the format of the clinical data above.</p><p>(2 KB CSV).</p></caption><media xlink:href="pbio.0020108.sd013.csv"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd014"><label>Dataset S10</label><caption><title>Lung Cancer Gene Expression Data</title><p>This is the gene expression data for the lung cancer dataset of <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. (2002</xref>).</p><p>(5.39 MB TXT).</p></caption><media xlink:href="pbio.0020108.sd014.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd015"><label>Dataset S11</label><caption><title>Lung Cancer Survival Data</title><p>This is the clinical data for the lung cancer dataset of <xref rid="pbio-0020108-Beer1" ref-type="bibr">Beer et al. (2002</xref>). The format is the same as in the clinical data above.</p><p>(1 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd015.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd016"><label>Dataset S12</label><caption><title>AML Gene Expression Data</title><p>This is the gene expression data for the AML dataset of <xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. (2004</xref>).</p><p>(9.96 MB TXT).</p></caption><media xlink:href="pbio.0020108.sd016.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd017"><label>Dataset S13</label><caption><title>AML Survival Data</title><p>This is the clinical data for the AML dataset of <xref rid="pbio-0020108-Bullinger1" ref-type="bibr">Bullinger et al. (2004</xref>). It has the same format as the clinical data above.</p><p>(1 KB TXT).</p></caption><media xlink:href="pbio.0020108.sd017.txt"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg001"><label>Figure S1</label><caption><title>Results of Using PLS-Derived Cox Scores in the Supervised Clustering Procedure</title><p>(8.26 MB TIFF).</p></caption><media xlink:href="pbio.0020108.sg001.tif"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg002"><label>Figure S2</label><caption><title>Plot of Survival Versus the Least Squares Estimate of β˜ for the DLBCL Data</title><p>(8.33 MB TIFF).</p></caption><media xlink:href="pbio.0020108.sg002.tif"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sg003"><label>Figure S3</label><caption><title>Plot of Survival Versus the Least Squares Estimate of γ˜ for the DLBCL Data</title><p>(8.42 MB TIFF).</p></caption><media xlink:href="pbio.0020108.sg003.tif"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material><supplementary-material content-type="local-data" id="sd018"><label>Protocol S1</label><caption><title>Additional Models and Methods</title><p>(28 KB TEX).</p></caption><media xlink:href="pbio.0020108.sd018.tex"><caption><p>Click here for additional data file.</p></caption></media></supplementary-material></sec>
Neural Basis of Solving Problems with Insight	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>If you're one of those insufferable people who can finish the Saturday <italic>New York Times</italic> crossword puzzle, you probably have a gift for insight. The puzzles always have an underlying hint to solving them, but on Saturdays that clue is insanely obtuse. If you had all day, you could try a zillion different combinations and eventually figure it out. But with insight, you'd experience the usual clueless confusion, until—voilà—the fog clears and you get the clue, which suddenly seems obvious. The sudden flash of insight that precedes such “Aha!” moments is characteristic of many types of cognitive processes besides problem-solving, including memory retrieval, language comprehension, and various forms of creativity. Although different problem-solving strategies share many common attributes, insight-derived solutions appear to be unique in several ways. In this issue, researchers from Northwestern and Drexel Universities report on studies revealing a unique neural signature of such insight solutions.<xref ref-type="fig" rid="pbio-0020111-g001"/> </p><fig id="pbio-0020111-g001" position="float"><caption><title>Insight lights up the brain</title></caption><graphic xlink:href="pbio.0020111.g001"/></fig><p>Mark Jung-Beeman, John Kounios, Ed Bowden, and their colleagues recount the storied origin of the term <italic>Eureka!</italic>, which Archimedes reportedly shouted upon realizing that water displacement could be used to compute density. Illustrating the strong emotional response elicited by such a sudden insight, Archimedes is said to have run home from the baths in euphoric glee—without his clothes.</p><p>Among other characteristics that typically distinguish insight from “noninsight” solutions, people feel stuck before insight strikes; they can't explain how they solved the problem and might say they were not even thinking about it; the solution appears suddenly and is immediately seen as correct. But are the neural processes involved in arriving at a solution through insight actually distinct from those related to more mundane problem-solving?</p><p>Recent findings suggest that people think about solutions, at an unconscious level, prior to solving insight problems, and that the right cerebral hemisphere (RH) appears to be preferentially involved. Jung-Beeman et al. predicted that a particular region of the RH, called the anterior superior temporal gyrus (aSTG), is likely involved in insight because it seems critical for tasks that require recognizing broad associative semantic relationships—exactly the type of process that could facilitate reinterpretation of problems and lead to insight.</p><p>To test this hypothesis, Jung-Beeman et al. mapped both the location and electrical signature of neural activity using functional magnetic resonance imaging (fMRI) and the electroencephalogram (EEG). In the first experiment, thirteen people were given three words (<italic>pine</italic>, <italic>crab</italic>, <italic>sauce</italic>) and asked to think of one word that would form a compound word or phrase for each of the words (can you figure it out?). Neural activity was mapped with fMRI while the participants were given 124 similar word problems—which can be solved quickly with or without insight, and evoke a distinct Aha! moment about half the time they're solved. Subjects pressed a button to indicate whether they had solved the problem using insight, which they had been told leads to an Aha! experience characterized by suddenness and obviousness.</p><p>While several cortical regions showed about the same heightened activity for both insight and noninsight-derived solutions, only the aSTG in the RH showed a robust insight effect. Given that neural activity in this area also increased when subjects first encountered the problem (perhaps reflecting unconscious processing), the authors conclude that the increase does not simply reflect the emotional jolt associated with insight.</p><p>In a second experiment, 19 new participants engaged in the same type of problem-solving tasks as the first group while their brain waves were measured with an EEG. The researchers then analyzed the EEG recordings to look for differences between insight and noninsight solutions in brain wave activity. The researchers found that 0.3 seconds before the subjects indicated solutions achieved through insight, there was a burst of neural activity of one particular type: high-frequency (gamma band) activity that is often thought to reflect complex cognitive processing. This activity was also mapped to the aSTG of the RH, providing compelling convergence across experiments and methods.</p><p>Problem-solving involves a complex cortical network to encode, retrieve, and evaluate information, but these results show that solving verbal problems with insight requires <italic>at least</italic> one additional component. Further, the fact that the effect occurred in RH aSTG suggests what that process may be: integration of distantly related information. Distinct neural processes, the authors conclude, underlie the sudden flash of insight that allows people to “see connections that previously eluded them.”</p>
When Food Kills	Could not extract abstract	<contrib contrib-type="author"><name><surname>Krebs</surname><given-names>John</given-names></name></contrib>	PLoS Biology	<p>For the estimated 800 million people, living largely in developing countries, without enough food to eat, the main food risk is starvation. But if you ask, ‘When does food actually kill?’ in a country such as the United Kingdom, ‘Not that often’ is the short reply you would give after reading Hugh Pennington's book <italic>When Food Kills:</italic> BSE, E. coli, <italic>and Disaster Science</italic>. The two food-borne diseases that occupy much of the book, <named-content content-type="genus-species">Escherichia coli</named-content> 0157 and bovine spongiform encephalopathy (BSE), kill humans very rarely, although the ramifications and implications of these few deaths for science, regulators, and government are large.<xref ref-type="fig" rid="pbio-0020112-g001"/> </p><fig id="pbio-0020112-g001" position="float"><graphic xlink:href="pbio.0020112.g001"/></fig><p>As Pennington clearly explains, there is still much uncertainty in the science of BSE, and the eventual UK death toll from the human form may be as low as a few hundred, with even the most pessimistic expert assessments putting the upper bound as fewer than 5,000. Food-borne <named-content content-type="genus-species">E. coli</named-content> 0157 kills fewer than a dozen people a year in the UK.</p><p>Whilst each death is a terrible tragedy and an indescribably harrowing experience for those close to the victim, these figures are small when compared with other ways in which food kills. Epidemiologists estimate that the dietary contributions to cardiovascular disease and cancer between them kill more than 100,000 people a year in Britain. Yet we hear much more about BSE and <named-content content-type="genus-species">E. coli</named-content> as food risks. For instance, a recent study by the King's Fund (<ext-link ext-link-type="uri" xlink:href="http://www.kingsfund.org.uk/pdf/healthinthenewssummary.pdf">http://www.kingsfund.org.uk/pdf/healthinthenewssummary.pdf</ext-link>) reports that the rate of news coverage in the UK of a death from variant Creutzfeldt-Jakob disease, the human form of BSE, is nearly 23,000 times that for a death from obesity.</p><p>In his characteristically diverting and obscurely erudite way, Pennington describes this discrepancy between public perception and magnitude of risk by referring to an article on railway accidents published in 1859 by one Dionysius Lardner. The systematic and much more revealing analyses of risk perception by psychologists such as Paul Slovic over the past 25 years do not get a mention.</p><p>In fact, one of the hallmarks of Pennington's style is his enthusiasm for taking his reader down little-known historical byways. Whether it be the drowning (possibly suicide) of King Ludwig II of Bavaria in the Starnberger See or the treatment of James Norris in Bethlehem Lunatic Asylum in 1814, Pennington has an almost endless supply of anecdotes to provide peripheral colour to his main narrative. Indeed, on some occasions his delight in the detail makes it hard to see where the main narrative is leading, although his aim is to show that similar conclusions can be drawn about risk management in food, transport, oil rigs, and other fields.</p><p>Anyone who has heard Hugh Pennington speak will know that he has a remarkably direct and engaging style, which he translates into the written word with verve. Already on page 2, he gets us into the mood by referring to a sample from a five-year-old girl sent for analysis at the start of the Lanarkshire <named-content content-type="genus-species">E. coli</named-content> outbreak of 1996: ‘It was a stool. The word carries the impression of firmness, even of deliberate effort in its production. Hers was not’. His laconic sense of humour is also reflected in many of the wittily irrelevant or tangential photographs. My personal favourites are ‘Her Majesty in Gloves’ on page 44 and ‘Turds on Campsite Track’ on page 101.</p><p>The Lanarkshire <named-content content-type="genus-species">E. coli</named-content> 0157 outbreak, which in late 1996 affected 202 people and killed eight, was very much Pennington's show. He chaired the public enquiry that led eventually to a change in the law, requiring all butchers in the UK handling cooked and raw meat to be licensed. The license itself is less important than the training in food safety management principles that precedes it. The butcher John Barr (and his staff), whose shop was the primary source of the outbreak, apparently did not know that you have to keep raw meat and ready-to-eat products separate to avoid cross-contamination with dangerous pathogens, such as <named-content content-type="genus-species">E. coli</named-content> 0157, that can occur in raw meat. Pennington's authoritative and blow-by-blow account shows failings not only in the butcher (who was, incidentally, Scottish Master Butcher of the Year in 1996), but also in the inspectors who had visited his shop eight times in the previous two years. They had not, apparently, picked up that Barr and his staff employed the same knives for cutting up raw and cooked meat, nor that they used a ‘biodegradable’ cleaning fluid, not realising that this is not the same as ‘biocidal’.</p><p>The second theme, BSE, is given somewhat shorter treatment. Nevertheless, Pennington goes into some detail in assessing the prion theory of transmissible spongiform encephalopathies (he argues that a nucleic acid is not also involved). He also reviews the sequence of events that led the UK government in the early 1990s to conclude that there was not likely to be a risk to human health and to be slow to change its view. This and the concluding part of the book (see below) draw heavily on the Phillips Enquiry into BSE. Although this enquiry focussed on the response of the UK government, its lessons are relevant to other countries where BSE has emerged in recent years, including many European countries, Japan, Canada, and the United States.</p><p>In his book <italic>Mountains of the Mind</italic>, Robert Macfarlane writes: ‘[F]or the hunter risk wasn't optional—it came with the job. I sought risk out, however. I courted, in fact paid for it. This is the great shift which has taken place in the history of risk…. [I]t became a commodity’. Pennington reflects a similar shift in attitude to food risk over the past half century or so. Back in 1938, although it was known that over 2,500 people a year in Britain died from drinking raw milk, the risk was not seen as large enough to warrant legislation to make pasteurisation compulsory. We are now used to much higher standards of food safety, and we can, as a society, enjoy the luxury of fear of relatively minor risks.</p><p>Nevertheless, there are important lessons from past failures for all involved in food safety (and in other areas of risk management), and Pennington discusses some of these in his concluding chapters. He emphasises the need to continually review the evidence underpinning risk assessments, to communicate effectively with the media, to ensure that actions to manage risks are effectively implemented and audited. Notably, he refers to the importance of inclusiveness and openness about risk and uncertainty in decision-making: ‘[I]f [this] becomes the norm, it will be possible to say that good has come out of tragedy’.</p>
Troubled Waters: The Future of Global Fisheries	Could not extract abstract	<contrib contrib-type="author"><name><surname>Gewin</surname><given-names>Virginia</given-names></name></contrib>	PLoS Biology	<p>It is becoming increasingly apparent that the vast blue expanse of ocean—the last frontier—is not as inexhaustible as it once seemed. While we have yet to fully explore the reaches of the sea, technology has granted humans the ability to harvest its wealth. We can now fish anywhere, at any depth, for any species. Like the American frontier range's bison and wolf populations brought to the brink of extinction swordfish and sharks are the ocean's most pursued prizes. The disadvantages associated with the depth and dimensions of this open range, however, have long obscured the real consequences of fishing. Indeed, scientists have the formidable challenge of assessing the status of species whose home covers over 75% of the earth.</p><p>Three recent highly publicized papers—a trifecta detailing troubled waters—call attention to overfishing's contributions to the dramatic declines in global fisheries. Delving into the past, <xref rid="pbio-0020113-Jackson1" ref-type="bibr">Jeremy Jackson and colleagues (2001)</xref> combined local historic records with current estimates to detail the ecological impacts of overfishing, <xref rid="pbio-0020113-Watson1" ref-type="bibr">Reg Watson and Daniel Pauly (2001)</xref> drew attention to distortions of global catches, and <xref rid="pbio-0020113-Myers1" ref-type="bibr">Ransom Myers and Boris Worm (2003)</xref> highlighted the depletion of the majority of the largest ocean predators. While some have valid criticisms of the assumptions and aggregation of historic data used to assess the global situation, few disagree with the overriding conclusion that humans have drastically altered not only fish biodiversity, but, increasingly, the ocean itself.</p><p>Recent reports by the United Nation's Food and Agriculture Organization (FAO) which maintains the world's most complete global fisheries database, appear to validate the conclusions of these studies. The most recent FAO report states that 28% of global stocks are significantly depleted or overexploited, and 47% are either fully exploited or meet the target maximum sustainable yield. Only 24% of global stocks are either under- or moderately exploited. As the sea is increasingly harvested, many ecologists wonder how the ecosystem will continue to function (<xref rid="pbio-0020113-Jackson1" ref-type="bibr">Jackson et al. 2001</xref>). Although economic and social considerations often supercede scientific assessments, science will continuously be called upon to deliver management options that will straddle the needs for conservation and production, even in areas where there is only subsistence fishing <xref ref-type="boxed-text" rid="box1">(Box 1)</xref>. As scientists debate the details of global fisheries assessment, they are also including studies of the long-term ecosystem effects and options for recovery efforts. Like was done on the open range, shall we conserve or farm the sea—or both?</p><sec id="s2"><title>Catches, Collapses, and Controversies</title><p>The FAO began keeping fisheries records in 1950. Unfortunately, an enormous amount of data comes directly from each country's fishing industry, which is often biased as a result of unreported discarding, illegal fishing, and the misreporting of harvests. For example, mid-level Chinese government officials seeking promotions systematically enhanced China's fisheries numbers in recent years—which inflated and skewed international catch rates.</p><p>The FAO data show that catches, excluding a recent surge in anchoveta and China's suspect numbers, reached a peak of 80 million metric tons in the late 1980s and have since begun to decline. Regional studies validate these trends. “Most of the line fish around the coast of South Africa are depleted to 5%–15% of pristine levels,” says George Branch, a marine biologist from the University of Cape Town (Cape Town, South Africa). Meryl Williams, Director General of WorldFish in Penang, Malaysia, notes that the Asia-specific database called TrawlBase (<ext-link ext-link-type="uri" xlink:href="www.worldfishcenter.org/trawl/">www.worldfishcenter.org/trawl/</ext-link>) confirms that the region's commercial species have been depleted to 10%–30% of what they were 30–40 years ago.</p><p>Obtaining accurate information on highly migratory species is challenging, to say the least. It is not hard to imagine that data quality is the biggest disadvantage to any scientific assessment. Of the 50 managed stocks in the northeast Atlantic Ocean—including invertebrates, sport fishes, and major commercial finfish—data are kept on only one-fifth of the species. There are 250 fish species in the region, but only 55 species are of commercial interest and merit inquiry. “We know next to nothing about noncommercially fished species,” notes Jeff Hutchings, a conservation biologist at Dalhousie University (Halifax, Nova Scotia, Canada). And that is where fisheries have adequate access to current monitoring programs. “With the recent expansion of the Taiwanese and Chinese fleets, we don't have the kind of sampling programs needed for those kinds of fisheries,” says Rick Deriso, a fisheries scientist with the Inter-American Tropical Tuna Commission (IATTC) (La Jolla, California, United States).</p><p>Couple these inadequacies with previously unknown bycatch rates (i.e., the fish caught in addition to the target catch) and illegal catches, and it is easy to see that the task is formidable. The FAO estimates that roughly one-quarter of the marine commercial catch destined for human consumption—some 18–40 million metric tons of fish—is thrown back in the sea, a harvested catch that is never utilized or counted. It is estimated that the illegal, unreported, and unregulated (IUU) fisheries surpass allowed fishing quotas by 300%. IUU fishers operate in areas where fishing is not permitted, use banned technologies or outlawed net types, or underreport catches. “The IUU fishery for Patagonian toothfish expanded rapidly in the mid-1990s, likely on the order of 20–30 vessels,” says Andrew Constable, an ecological modeler at the Australian Antarctic Division (Kingston, Australia), who also works with the Scientific Committee of the Commission for the Conservation of Antarctic Marine Living Resources (Hobart, Australia). “These rates of IUU fishing could reduce stocks to threshold levels in some areas in two to five years,” he adds.</p><p>Often overlooked is the inescapable fact that even sustainable harvest rates reduce fish populations quickly. “If the goal is a productive fishery, we're automatically talking about up to a 70% decline in population across the board,” says Deriso. The FAO's Chief of Marine Resource Services, Jorge Csirke, states that “from a stock point of view, there is no way to preserve integrity of wild stocks and exploit them at the same time.” Indeed, the United States' National Marine Fisheries Service (NMFS) considers optimal harvest rates to be between 40%–60% of virgin levels. But once fish populations dip below the 10%–20% mark, declines are of serious concern.</p><p>Atlantic cod in Canadian waters suffered a total population collapse and are now on Canada's endangered species list (<xref ref-type="fig" rid="pbio-0020113-g001">Figure 1</xref>). From 2 billion breeding individuals in the 1960s, Atlantic cod populations have declined by almost 90%, according to Hutchings. While advisors called attention to declining cod stocks, Constable notes that by the time a significant declining trend has been detected by traditional catch assessments, stocks are likely to be in poor shape, if not already depleted.</p><fig id="pbio-0020113-g001" position="float"><label>Figure 1</label><caption><title>Cod in a High Arctic Lake in Canada</title><p>These cod resemble those of past Atlantic catches. Measuring 47–53 inches (120–135 cm) long and weighing between 44 and 57 pounds (20 and 26 kg), it is easy to see that today's 16–20 inches (40–50 cm) commercially caught cod are less than half this size. (Photo, with permission, by David Hardie, Dalhousie University.)</p></caption><graphic xlink:href="pbio.0020113.g001"/></fig><p>Given the task of compiling data on only the economically important species, fisheries biologists developed a single-species management approach in the 1960s, which assumed that fisheries affect each species in isolation. This approach, although now rife with problems, served the community and the politicians well during the decades of abundant resources. “They brought the approach of single-species management to near-perfection,” says Boris Worm, a marine ecologist at the Institute for Marine Science in Kiel, Germany. A growing discontent with the model, in addition to greater awareness of ecological interactions, however, prompted Worm and his Dalhousie University colleague Ransom Myers to question the sustainability of the single-species approach. Attempting a comprehensive assessment, their widely cited recent paper (<xref rid="pbio-0020113-Myers1" ref-type="bibr">Myers and Worm 2003</xref>) indicated that the global ocean has lost more than 90% of large predatory fishes, such as marlin, sharks, and rays.</p><p>However, this new approach to assess fish stocks is not without its critics. Fisheries biologists point out that the nuances of management contained in fisheries data—such as altered fisher behavior, the variable “catchability” of individual species, and altered gear use—were discounted in the <xref rid="pbio-0020113-Myers1" ref-type="bibr">Myers and Worm (2003)</xref> assessment and led to misinterpretations for some species, notably tropical tunas (<xref ref-type="fig" rid="pbio-0020113-g002">Figure 2</xref>). A number of tuna biologists have expressed concern that these omissions have left the mistaken impression that all tuna species are among the list of declining predators (<xref rid="pbio-0020113-Hampton1" ref-type="bibr">Hampton et al 2003</xref>). Worm acknowledges that his approach can be improved, but says, “The whole point of our paper was to aggregate species to communities to see what the overall ecosystem is doing.”</p><fig id="pbio-0020113-g002" position="float"><label>Figure 2</label><caption><title>Pole Fishing for Medium-Sized (40–50 lb or 18.1–22.7 kg) Big-Eye Tuna aboard the Live-Bait, Pole-and-Line Vessel <italic>Her Grace</italic> </title><p>(Photo, with permission, by Kurt Schaefer and Dan Fuller, IATTC.)</p></caption><graphic xlink:href="pbio.0020113.g002"/></fig></sec><sec id="s3"><title>Ecosystem Sustainability</title><p>Despite the controversy, most agree that the large predators, particularly sharks, skates, rays, and marlin, are in the most dire straits. Unlike other lower-trophic order species, the wholesale removal of top predators has enormous effects on the rest of the ecosystem. One consequence is that overall reproduction rates can potentially suffer. Fish size, gender, and age at maturity have a substantial impact on individual species' reproduction rates. Since larger fish are the most susceptible to fishing, the population's age structure can shift as individuals, particularly females, are fished out. For example, a 23-inch (59-cm) female vermilion rockfish can produce 17 times the young of a 14-inch (36-cm) fish. Given uncertainties with population dynamics, the fact that basic biological data are missing makes the job even harder. While knowledge of these components is still quite spotty, tuna inventories, for example, have started collecting gender data on catches.</p><p>Daniel Pauly, a fisheries biologist at the University of Vancouver (Vancouver, British Columbia, Canada), has shown that increased fishing has caused the industry to “fish down the food web,” or systematically move to lower trophic levels over time as higher ones were depleted (<xref rid="pbio-0020113-Pauly1" ref-type="bibr">Pauly et al. 1998</xref>). The impact to ecosystems is only beginning to be uncovered. “If you fish out an abundant predator, the species that it was eating or competing with will increase,” says Worm. “The problem is that the ecosystem may change in such a way that recovery is inhibited because a species niche space is taken or altered.”</p><p>Fisheries science has taken steps to increase the quality of data in recent years. “Traditional fishery models assumed that a fishery was a homogenous thing—like bacteria in a bottle—rather than a spatially diverse system,” says Pierre Kleiber, a fisheries biologist with the Pacific Islands Fisheries Science Center of the NMFS (Honolulu, Hawaii, United States). He adds that recent work accounts for spatial diversity. In addition, fisheries are now dealing with the inherent uncertainty of their work and are factoring that into models and decision-making. “Uncertainty didn't used to be dealt with at all in formulating fishery management advice,” confirms Keith Sainsbury, a marine ecologist with the Commonwealth Scientific and Industrial Research Organisation (CSIRO) (Clayton, South Victoria, Australia), adding that its absence gave rise to an awful lot of troubles. “Traditional models tended to assume perfect data with no holes in it,” says Kleiber. “Now we've tried to craft a model to fit the realities of missing data.”</p><p>As well as incorporating spatial diversity and uncertainty, researchers are beginning to comprehend the ecological damage caused by different types of fishing gear. Indeed, trawling the bottom of the seafloor for groundfish can destroy a half-acre footprint of habitat (<xref ref-type="fig" rid="pbio-0020113-g003">Figure 3</xref>). Detailed reports document that, depending on the habitat's stability, bottom trawling can not only remove fish from seafloor habitats, but alter bottom relief such that it compromises the ability of other fish to survive (<xref rid="pbio-0020113-NRC1" ref-type="bibr">NRC, 1002</xref>). In Australia, for example, lingcod rely on undisturbed bottom relief to lay their eggs, while other groundfish species depend on complex seafloor habitats for the majority of their food.</p><fig id="pbio-0020113-g003" position="float"><label>Figure 3</label><caption><title>The Effect of Trawling the Seafloor for Groundfish</title><p>(A) The coral community and seabed on an untrawled seamount. (B) The exposed bedrock of a trawled seamount. Both are 1,000–2,000 meters (1094–2188 yards) below the surface. (Photo, with permission, by CSIRO Marine Research.)</p></caption><graphic xlink:href="pbio.0020113.g003"/></fig><p>“Science is getting more realistic, but it is getting more difficult,” says Branch. Ecological models are far more complex than traditional fisheries models, says Csirke, adding that more model variables make it more difficult to apply to fisheries, an industry whose focus is, understandably, not conservation. Despite its incorporation into national fisheries policies, ecosystem-based management remains a loosely defined term. It is not a well-defined concept because it is not possible to optimize every species, says Deriso.</p><p>An additional concern to scientists is that of biomass resilience in the face of environmental changes. Francisco Chavez, a biologist with the Monterey Bay Aquarium Research Institute (Moss Landing, California, United States), recently demonstrated that over a 25-year period, warmer and cooler Pacific waters tilt the distribution of anchoveta versus sardines, both open-ocean dwellers (<xref rid="pbio-0020113-Chavez1" ref-type="bibr">Chavez et al. 2003</xref>). Indeed, El Niño influenced the crash of the heavily fished Peruvian anchoveta industry in the late 1970s. These examples illustrate how susceptible fisheries are to environmental fluctuations. When the biomass of a population is reduced, it is much more sensitive to environmental change. We do not know how environmental fluctuations like these will affect the natural production of young fish, says Kleiber, expressing the concern that without a better understanding of climate, fisheries scientists end up trying to estimate moving targets.</p><p>In the end, many scientists have their doubts about the influence of science on decision-making. “My personal view is that it's naïve to think that modifying and improving models will necessarily lead to improved natural resource management,” says Simon Jennings, a fisheries biologist with the United Kingdom's Centre for Environment, Fisheries and Aquaculture Science in Lowestoft. Indeed, the International Council for the Exploration of the Seas (Copenhagen, Denmark) recently recommended a total ban on North Sea and Irish Sea cod stocks, based on single-species assessment. Although the more intensive ecosystem-based models could not have produced a more stringent recommendation, politicians allowed harvests at roughly half of last year's catch.</p></sec><sec id="s4"><title>To Conserve or to Farm?</title><p>While lowering fisheries' effort seems the most logical approach to the recovery of depleted fisheries, social and economic concerns often stymie political action. Yet demand for seafood continues. Therefore, scientists also are investigating both conservation and alternative production options.</p><p>Given the social, economic, and political problems associated with that, managers have often used closures to help a hard-hit species recover. In many cases, however, the recovery time for exploited species is longer than once thought (<xref rid="pbio-0020113-Hutchings1" ref-type="bibr">Hutchings 2000</xref>). “Based on the available information, it is not unusual for fish populations to show no or little recovery even after 15 years,” says Hutchings. “All else being equal, we predict the earlier the age of maturity, the faster the rate of recovery,” he adds. And that depends on environmental conditions as well. “In the case of Antarctic species, some overexploited populations remain at less than 5% pre-exploitation abundance after 30 years,” says Constable.</p><p>One management strategy to recover species is to create marine protected areas (MPAs), zones that restrict all removal of marine life <xref ref-type="boxed-text" rid="box2">(Box 2)</xref>. A number of marine ecologists are staunch supporters of MPAs for both conservation and fishery's recovery. What looked like sustainability in the past were fisheries out of our reach—naturally protected areas—says Pauly, adding that our increasing ability to harvest fisheries necessitates the creation of MPAs now. In theory, these areas are refugia for fishes to reproduce, spilling over not only healthy adults but also potentially transporting thousands of viable young—seeding surrounding waters. To date, less than 1% of the ocean's area is protected, which hinders the ability to conclusively determine if spillover rates have the predicted impact on fishery's recovery.</p><p>A review of 89 studies of MPAs by Ben Halpern, a student at the University of California, Santa Barbara (Santa Barbara, California, United States), demonstrated that the average number of fish inside a reserve increases between 60%– and 150% (<xref rid="pbio-0020113-Halpern1" ref-type="bibr">Halpern 2003</xref>). In addition, 59% of the sites had increased diversity. While the numbers inside the reserves look good, the crucial condition of larval spillover has yet to be proven. Most scientists involved in the debate agree that MPAs should be one component in an overall management scheme, but worry that until the crucial element of fishing effort is resolved, MPAs may just displace the vast industrial fleets.</p><p>In terms of simply producing fish for global food needs, aquaculture (also known as fish farming) is another, increasingly popular, option. In 2001, the European Union produced 17% of total fishery's production via aquaculture. These numbers are projected to steadily increase, but some question whether aquaculture would be sufficient to supply what has been lost by overexploited fisheries.</p><p>Concentrated in coastal areas, aquaculture has aroused numerous concerns. Indeed, in developed countries, most operations grow carnivorous fish, which necessitates growing fish to feed fish. While the process has become more efficient in recent years, due in part to a growing reliance on vegetarian diets, it still takes about 3 pounds (1.36 kg) of fish to create 2.2 pounds (1 kg) of desirable meat (<xref rid="pbio-0020113-Aldhous1" ref-type="bibr">Aldhous 2004</xref>). Yet, the total catch of food fish continues to grow, as do concerns about nutrient runoff and estuary pollution resulting from aquaculture. Increasingly, coastal residents often complain about the aesthetics of such activities, and there is also new research that indicates that farm-raised fish harbor more cancer-causing pollutants than wild species (<xref rid="pbio-0020113-Hites1" ref-type="bibr">Hites et al. 2004</xref>).</p><p>To alleviate many of these concerns, open-ocean aquaculture is now being considered. Indeed, the NMFS is set to propose a Code of Conduct for Offshore Aquaculture, which would open up the 200-mile (322-km) United States Exclusive Economic Zone to net pens seaward of coastal state boundaries and authorities. The Sea Grant program in conjunction with interested business, is also currently assessing the carrying capacity of open-water pens as well as their potential environmental impact. Given increased industrial interest and unchanging demand for seafood, many think farming the sea may be around the corner.</p><p>Undoubtedly, scientific effort will continue to inform both conservationists and industry about fisheries' capacity and potential recovery options. As attitudes towards fisheries continue to change, increased understanding of the ecological underpinnings should help strike a more informed balance between fisheries' conservation and production. “The big mistake is suggesting that you can manage fish stocks,” says Niels Daan, a biologist with the Netherlands Institute for Fisheries Research (IJmuiden, The Netherlands). “In my opinion, we can only manage human activity.”</p></sec><sec><title/><boxed-text id="box1" position="float"><caption><title>Box 1. Fisheries Management in Developing Countries</title></caption><p>While industrial-scale fishing is a growing concern to fisheries biologists, the management of subsistence fishing in developing countries is equally complex. Indonesia alone has 1.3 million fishers. Given the lack of alternative economic options for subsistence fishers, it is much more difficult to reduce fishing because it meets immediate food and resource needs. Local scientists, often lacking in resources, have a much more difficult time assessing the effects and offering advice to governmental fisheries regulators, who have limited political influence. Kenyan researcher Tim McClanahan notes that a main problem is a lack of coordination and respect between traditional and national programs of management. Therefore, he focuses on the fishing gear used. By reconciling the impact of certain fishing gear with traditional knowledge, McClanahan has developed a basis for suggested restrictions deemed acceptable to the local community.</p></boxed-text><boxed-text id="box2" position="float"><caption><title>Box 2. The Establishment of High Profile MPAs</title></caption><p>While MPAs are heavily touted as one of the best management tools to address both conservation and fisheries management, few have been enacted. In 2001, following a strong mandate by the Australian Minister to the Environment and overwhelming political will, the Great Barrier Reef Marine Park Authority (GBRMPA) in Australia established a network of marine protected, or no-take, areas as an ecosystem-based management approach.</p><p>In setting up the reserve networks, scientists determined the most effective areas to protect biodiversity with little impact to productivity. “We tried to avoid peak use areas, while protecting at least one-third of each bioregion and minimizing the impact to users of the Great Barrier Reef Park,” says Phil Cadwallader, Director of Fisheries at the GBRMPA.</p><p>Off the coast of California, the Channel Islands network of marine reserves, established in April 2003, consists of 13 areas designed to protect biodiversity and critical habitat for breeding fish and to maintain biodiversity. The area has suffered serious declines of red snapper, angel sharks, and abalone, once plentiful off the California coast, over the past decade. Scientists designed the network to protect those productive habitats that would help ensure that larval dispersal was maintained between the individual reserves. Totaling 132 nautical square miles (342 nautical square kilometers), 11 of the areas are no-take reserves—allowing no fishing or harvest of any kind.</p></boxed-text></sec>
A Window into the Brain Demonstrates the Importance of Astrocytes	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>Did you ever wish you could peek inside someone's brain and see what was going on in there? In research reported in this issue of <italic>PLoS Biology</italic>, Hajime Hirase and his colleagues at Rutgers University have done just that by focusing their microscope on the brains of living rats in order to examine how certain cells called astrocytes function in vivo.<xref ref-type="fig" rid="pbio-0020115-g001"/> </p><fig id="pbio-0020115-g001" position="float"><caption><title>Astrocyte in the cerebral cortex</title></caption><graphic xlink:href="pbio.0020115.g001"/></fig><p>In the longstanding quest to understand how the brain works, scientists have focused on neurons. Neurons conduct action potentials, electrical signals that transmit information in the nervous system. But the brain also contains several other types of cells called glia. (<italic>Glia</italic> is derived from the Latin for “glue”; these cells were thought to “hold it all together.”) One type of glial cell, the astrocyte (named for its starlike shape), is the most populous cell in the brain and forms an intimate association with neurons and their synapses. It was thought that these cells played a supporting role in the brain, ensuring the proper chemical environment for synapses.</p><p>Recent research, however, has suggested that astrocytes and other glial cells may play a more significant role. When examining astrocytes cultured in the lab, scientists have observed behavior suggesting that astrocytes can communicate with neurons. Though astrocytes cannot propagate electrical signals like neurons do, they can sense the transmission of such signals at the synapse between two neurons. Furthermore, astrocytes are able to propagate a different kind of signal, a chemical signal based on the release of calcium ions. Calcium signaling is a mechanism of chemical signaling that has been observed in many other cell types. The exact properties of neuron–astrocyte communication, however, are not clear because different preparations of these tissues have yielded different results. It has also not been established that this type of communication occurs in the living brain.</p><p>To explore such questions, Hirase and colleagues have taken the next step by investigating the calcium signaling properties of astrocytes in the brains of living rats. To accomplish this feat, the researchers used a combination of two technologies. They monitored calcium signaling using a fluorescent dye called Fluo-4, which fluoresces in response to calcium ions. Then they used a special type of microscope called a two-photon laser scanning microscope to visualize the dye. Since this type of microscope uses a lower energy laser, it can image the dye in living tissue without causing harm.</p><p>The researchers applied the dye to the brains of anesthetized rats, washed out the excess dye that had not penetrated into cells, and then imaged the tissue under the microscope. They first confirmed that they indeed were examining astrocytes and noticed that cells displayed a moderate level of baseline calcium signaling activity. They then used a drug called bicuculline to stimulate neurons and observed a significant increase in the calcium signaling activity of the astrocytes. Because bicuculline only affects neurons, this implies that the astrocytes are responding to the activity of the neurons. The researchers also found that neighboring astrocytes often also displayed coordinated calcium signaling activity, suggesting that the communication among astrocytes is facilitated by increased neuronal activity.</p><p>This research confirms the complexity of astrocyte signaling functions in the living brain and demonstrates that astrocytes play far more than a supporting role in brain function. It also establishes an important experimental system for scientists seeking to understand how these distinct elements of the brain—neurons and astrocytes—work together. Though this research makes it clear that signaling exists both among astrocytes and between neurons and astrocytes, scientists have yet to understand the effect of this signaling. Some possibilities include regulation of synapse formation, modification of synaptic strength, or more complicated roles in information processing resulting from the coordination of neuronal activity. Future research using this and other systems will help reveal these functions.</p>
Phage Display Libraries Identify T Cells	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>Doctors and researchers often look for the rapid proliferation of T cell populations, key defensive players in the immune system, as a telltale sign that the body is working hard to fend off a foreign threat. Every one of these circulating white blood cells carries a T cell receptor (TCR) that binds to a specific protein, or antigen, when displayed on the surface of a cell. A match between TCR and displayed antigen results in the cell's death and the subsequent expansion of T cell clones, all programmed to recognize the original offending protein. Some TCRs bind and expand in response to pathogenic antigens, such as viral or bacterial proteins. But T cells can also react and proliferate inappropriately in response to the body's own proteins, leading to destructive autoimmune diseases such as multiple sclerosis, which is characterized by immune system attacks on nervous tissue. Self-recognizing TCRs, however, can also target and destroy tumors—though full activation of these T cells is inconsistent and poorly understood.<xref ref-type="fig" rid="pbio-0020117-g001"/> </p><fig id="pbio-0020117-g001" position="float"><caption><title>Peptide display</title></caption><graphic xlink:href="pbio.0020117.g001"/></fig><p>Identifying the particular antigen behind an exploding population of T cells is invaluable for finding the source of autoimmune diseases and studying immune responses to cancer. But it's a laborious and time-consuming process, as researchers are faced with the prospect of sifting through millions upon millions of possible matches between TCRs and their prospective antigen epitopes—the part of the antigenic molecule to which the receptor binds. Now, as they report in this issue of <italic>PLoS Biology</italic>, Frances Crawford and colleagues have developed a novel method for rapidly identifying TCR mimotopes—peptide sequences similar or identical to epitopes that also elicit the immune response—which can be used to determine the antigen of a given T cell population.</p><p>Working backwards, the team started off with two different T cell clones that had been previously selected for with a known antigen—a peptide called p3K. One clone was derived from mice genetically engineered to have broadly reactive T cells; the other, a conventional clone, was much more sensitive to the precise molecular structure of p3K.</p><p>Crawford and colleagues then created a “peptide library” comprising more than 30,000 baculoviruses (viruses that selectively target insect cells), each one carrying a slightly different version of the p3K gene, varied in regions of the peptide known to be important for TCR binding. These p3K genes were embedded within a major histocompatibility complex (MHC) gene—a type of cell surface protein that holds displayed antigens and is also important for proper TCR recognition. The team then unleashed their virus library onto insect cells that, once infected, began to produce the specific peptide–MHC complexes encoded on the viral DNA. The insect cells then shuttled these proteins to their surfaces, resulting in a vast array of cells that each displayed a unique variant of the p3K–MHC complex. This “display library” was then incubated with fluorescently labeled TCRs from the two different clones. By observing and isolating the insect cells that lit up, the researchers could see which of the thousands of cells displaying peptide–MHC possessed a mimotope capable of binding a TCR. Because the genetic information about the displayed complex was still stored within the virus-infected cell, the researchers could determine the full peptide sequence responsible for the identified mimotopes.</p><p>Confirming the effectiveness of their method, the results of the fluorescence experiments echoed the authors' original characterizations about the two populations of T cells. The broadly reactive TCR bound to several different uniquely displayed complexes; it had 20 mimotopes. The conventional TCR, however, bound only to one peptide–MHC complex, an almost perfect match to the original p3K peptide. Though this study was based on a known antigen and epitope (which allowed verification of the method), the baculovirus display library technique described here could easily be used on T cell populations with unknown antigens. With such a tool, researchers could, for example, identify the antigens connected with tumor-fighting T cells and, through inoculation, possibly induce the production of similar T cells in cancer patients who lack them.</p>
Evaluating Disease Trends in Marine Ecosystems	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>After the recent mad cow scare in the United States, 61% of Americans said they would start eating more fish, according to a <italic>Wall Street Journal Online</italic> poll. The respondents may not know that populations of large predatory fish, such as tuna, swordfish, and marlin, have declined 90% over the past 50 years or that less-prized species are increasingly overfished. Or that ever more fish and seafood species show rising levels of mercury contamination, rendering them unfit for human consumption—and contaminating other organisms in the ocean food chain. Humans are also affecting marine life in unexpected ways, as when large numbers of seals in Antarctica in 1955 and in Siberia in 1987 succumbed to canine distemper virus, presumably contracted from domestic dogs. In 2000, more than 10,000 Caspian seals—which also had contact with domestic dogs—died of the same virus. Such human incursions cause even more damage by exacerbating the effects of naturally occurring parasitic and pathogenic diseases that already wreak havoc as they ripple through the food chain.<xref ref-type="fig" rid="pbio-0020119-g001"/> </p><fig id="pbio-0020119-g001" position="float"><caption><title>A dead gorgonian sea fan on a wall in Palau (Photograph, with permission, by Drew Harvell)</title></caption><graphic xlink:href="pbio.0020119.g001"/></fig><p>With recent studies suggesting that disease rates have increased over the past 30 years—and are expected to increase even more, thanks to global climate change—prospects for protecting marine ecosystems depend on understanding the causes and nature of these disease outbreaks. While all indicators point to a real increase in disease rates, scientists have no baseline data to measure these increases against and so cannot directly test the hypothesis that marine diseases are increasing. Now Jessica Ward and Kevin Lafferty report a method that uses the recorded incidence of disease as a proxy for baseline data to identify disease trends in major groups of marine organisms.</p><p>Ward and Lafferty conducted an online search of 5,900 journals published from 1970 to 2001 for reports of disease in nine taxonomic groups: turtles, corals, mammals, urchins, mollusks, seagrasses, decapods (crustaceans), sharks/rays, and fishes. Their approach takes into account three potentially confounding factors in determining trends in this type of search. Fluctuations in publication numbers could skew results, since an increase in the number of scientific reports published in a particular taxonomy might not reflect a true increase in the incidence of disease; a particularly prolific author could bias the search results by turning up more cases of disease in a population than actually occurred; or a single disease event reported multiple times in different papers could create the impression that disease had suddenly increased. To normalize publication rates over time, Ward and Lafferty used a proportion of disease reports from a given population relative to the total number of reports in that group. To determine whether there was an “author effect,” they removed the most prolific author in each taxonomic group and found that an author's abundant contributions did not skew the results. Finally, they confirmed that a single disease didn't bias their results by removing multiple reports of the same disease from the literature before analyzing the trends.</p><p>When they analyzed the searches without adjusting for the total number of reports published, Ward and Lafferty found that reports of disease increased for all groups. But when they analyzed the normalized results, they found that trends varied. While there was a clear increase in disease among turtles, corals, mammals, urchins, and mollusks, they found no significant trends for seagrasses, decapods, and sharks/rays. And they found that disease reports actually decreased for fishes. (One explanation for this decrease could be that drastic reductions in population density present fewer opportunities for transmitting infection.) Ward and Lafferty tested the soundness of this approach by using a disease (raccoon rabies) for which baseline data exist and showing that normalized reports of raccoon rabies increased since 1970, just as the disease increased from one case reported in Virginia in 1977 to an “epizootic” outbreak, affecting eight mid-Atlantic states and Washington, D.C., by 1992.</p><p>The pattern of increased reports, the authors propose, confirms scientists' perceptions about the rising distress of threatened populations and thus reflects a real underlying pattern in nature. The fact that disease did not increase in all taxonomic groups suggests that increases in disease are not simply the result of increased study and that certain stressors, such as global climate change, likely impact disease in complex ways. By demonstrating that an actual change in disease over time is accompanied by a corresponding change in published reports by scientists, Ward and Lafferty have created a powerful tool to help evaluate trends in disease in the absence of baseline data. It is only by understanding the dynamics of disease outbreaks that scientists can help develop sound methods to contain them.</p>
The Elusive Baseline of Marine Disease: Are Diseases in Ocean Ecosystems Increasing?	<p>Disease outbreaks alter the structure and function of marine ecosystems, directly affecting vertebrates (mammals, turtles, fish), invertebrates (corals, crustaceans, echinoderms), and plants (seagrasses). Previous studies suggest a recent increase in marine disease. However, lack of baseline data in most communities prevents a direct test of this hypothesis. We developed a proxy to evaluate a prediction of the increasing disease hypothesis: the proportion of scientific publications reporting disease increased in recent decades. This represents, to our knowledge, the first quantitative use of normalized trends in the literature to investigate an ecological hypothesis. We searched a literature database for reports of parasites and disease (hereafter “disease”) in nine marine taxonomic groups from 1970 to 2001. Reports, normalized for research effort, increased in turtles, corals, mammals, urchins, and molluscs. No significant trends were detected for seagrasses, decapods, or sharks/rays (though disease occurred in these groups). Counter to the prediction, disease reports decreased in fishes. Formulating effective resource management policy requires understanding the basis and timing of marine disease events. Why disease outbreaks increased in some groups but not in others should be a priority for future investigation. The increase in several groups lends urgency to understanding disease dynamics, particularly since few viable options currently exist to mitigate disease in the oceans.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Ward</surname><given-names>Jessica R</given-names></name><email>[email protected]</email><xref ref-type="aff" rid="aff1"> <sup>1</sup> </xref></contrib><contrib contrib-type="author"><name><surname>Lafferty</surname><given-names>Kevin D</given-names></name><xref ref-type="aff" rid="aff2"> <sup>2</sup> </xref></contrib>	PLoS Biology	<sec id="s1"><title>Introduction</title><p>Marine organisms serve as hosts for a diversity of parasites and pathogens. Mortalities affect not only the host population, but can cascade through ecosystems. Loss of biologically engineered habitats such as seagrass beds (<xref rid="pbio-0020120-Lewis1" ref-type="bibr">Lewis 1933</xref>; <xref rid="pbio-0020120-Taylor1" ref-type="bibr">Taylor 1933</xref>) and cascading trophic effects due to removal of consumers (<xref rid="pbio-0020120-Lessios1" ref-type="bibr">Lessios 1988</xref>) can alter community structure.</p><p>Understanding marine disease and the timing of outbreaks is increasingly important given escalating anthropogenic stressors affecting marine ecosystems. Humans directly affect community structure (e.g., overfishing [<xref rid="pbio-0020120-Jackson1" ref-type="bibr">Jackson et al. 2001</xref>; <xref rid="pbio-0020120-Myers1" ref-type="bibr">Myers and Worm 2003</xref>]) and facilitate introduction of terrestrial pathogens to marine organisms (e.g., canine distemper virus in Antarctic seals [<xref rid="pbio-0020120-Bengtson1" ref-type="bibr">Bengtson and Boveng 1991</xref>]). Human-mediated climate change may also affect disease prevalence. A recent review predicts disease in both terrestrial and marine ecosystems could increase with future climate warming (<xref rid="pbio-0020120-Harvell3" ref-type="bibr">Harvell et al. 2002</xref>).</p><p>Previous literature reviews suggesting a higher rate of disease outbreaks in the last three decades (<xref rid="pbio-0020120-Epstein1" ref-type="bibr">Epstein et al. 1998</xref>; <xref rid="pbio-0020120-Harvell1" ref-type="bibr">Harvell et al. 1999</xref>), coupled with predictions of future increases due to climate change (<xref rid="pbio-0020120-Harvell3" ref-type="bibr">Harvell et al. 2002</xref>), lend new urgency to understanding causes of marine disease outbreaks. Evidence suggests the increase is real (<xref rid="pbio-0020120-Harvell1" ref-type="bibr">Harvell et al. 1999</xref>), yet lack of baseline data for most marine communities precludes a direct test of the hypothesis.</p><p>We developed a proxy method to test a prediction of the increasing disease hypothesis: that reports of disease in the scientific literature, normalized to overall publication rates, increased since 1970. We searched an online literature database (ISI Web of Science) and quantified reports of disease in natural populations of marine organisms from 1970 to 2001. Nine marine taxonomic groups were searched: turtles, corals, mammals, urchins, molluscs, seagrasses, deca-pods, sharks/rays, and fishes.</p><p>Previous analyses of ecological literature specifically assessed trends among scientists such as taxonomic bias (<xref rid="pbio-0020120-Clark1" ref-type="bibr">Clark and May 2002</xref>) and taxonomic chauvinism (<xref rid="pbio-0020120-Bonnet1" ref-type="bibr">Bonnet et al. 2002</xref>) in research. Our proxy method is to our knowledge the first quantitative use of normalized trends in the literature to investigate an ecological hypothesis. In the absence of baseline data, the literature proxy method detects important trends of disease in major groups of marine plants, invertebrates, and vertebrates.</p></sec><sec id="s2"><title>Results</title><p>The largest confounding factor when using literature searches to correlate disease events with time is temporal change in the total number of publications (related to disease or not) on the taxonomic group. To control for changes in total publication, data were normalized using a yearly proportion of disease reports from natural populations relative to total literature inputs for each taxonomic group.</p><p>Total disease reports, not normalized to literature inputs, increased in all groups (<xref ref-type="table" rid="pbio-0020120-t001">Table 1</xref>). However, normalized results varied with taxonomic group. Normalized disease reports increased in turtles, corals, mammals, urchins, and molluscs. No significant trends were detected for seagrasses, decapods, and sharks/rays (though disease occurred in these groups). Counter to the hypothesis, disease reports decreased in fishes (<xref ref-type="fig" rid="pbio-0020120-g001">Figure 1</xref>).</p><fig id="pbio-0020120-g001" position="float"><label>Figure 1</label><caption><title>Percent of Literature Reporting Disease over Time in Each Taxonomic Group</title><p> <italic>r<sub>s</sub></italic> is Spearman's ρ. α is controlled for multiple comparisons with Holm's sequential Bonferroni adjustments. (A) Turtle. (B) Coral bleaching and disease (closed square); coral disease including infectious bleaching (open circle); coral bleaching (asterisk). (C) Mammal. (D) Urchin. (E) Mollusc. (F) Seagrass. (G) Decapod. (H) Shark/ray. (I) Fish.</p></caption><graphic xlink:href="pbio.0020120.g001"/></fig><table-wrap id="pbio-0020120-t001" position="float"><label>Table 1</label><caption><title>Spearman's Rank Correlation Analysis</title></caption><graphic xlink:href="pbio.0020120.t001"/><table-wrap-foot><fn id="nt101"><p>The table shows total reports (not corrected for research effort), normalized reports, and normalized reports with most frequent author removed. <italic>r<sub>s</sub></italic> is Spearman's ρ. α is controlled for multiple comparisons with Holm's sequential Bonferroni adjustments. Bold indicates significance</p></fn></table-wrap-foot></table-wrap><p>The relevance of our approach hinges on the assumption that an actual change in disease over time is accompanied by a corresponding change in publication frequency by scientists. We evaluated this assumption by testing the protocols with a case in which the baseline was known (raccoon rabies [<xref rid="pbio-0020120-Rupprecht1" ref-type="bibr">Rupprecht and Smith 1994</xref>]). Normalized reports of raccoon rabies increased since 1970 (see <xref ref-type="table" rid="pbio-0020120-t001">Table 1</xref>) just as the disease increased from an index case in Virginia in 1977 to an epizootic affecting eight mid-Atlantic states and the District of Columbia by 1992 (<xref rid="pbio-0020120-Rupprecht1" ref-type="bibr">Rupprecht and Smith 1994</xref>). Despite improvements in search protocols, use of a literature proxy is limited by the inability to distinguish between an event that did not occur and an event that was not reported.</p><p>We tested whether particular authors contributed disproportionate primary literature inputs that could bias results. Papers by the most prolific author in each taxonomic group were removed to determine whether there was an “author effect,” and none was observed in any taxonomic group (see <xref ref-type="table" rid="pbio-0020120-t001">Table 1</xref>). Multiple reports of a single disease event could also bias the data. Multiple reports were removed from the turtle, coral, urchin, mammal, shark/ray, and seagrass literature. Removal of the reports did not alter the significance of the results; thus, multiple reports in the mollusc, decapod, and fish literature were not removed, owing to the large volume of literature in these groups.</p></sec><sec id="s3"><title>Discussion</title><p>We address an ecological hypothesis, that disease of marine organisms increased since 1970, using a quantitative literature proxy method. Although total reports of marine disease increased over time (<xref rid="pbio-0020120-Epstein1" ref-type="bibr">Epstein et al. 1998</xref>; see <xref ref-type="table" rid="pbio-0020120-t001">Table 1</xref>), a parallel increase in publication rates confounds interpretation of this pattern. Our approach normalizes data to overall publication within each group to circumvent this problem.</p><p>While an increase in disease reports was detected in many taxa, our finding that disease did not increase in all taxa has two important implications. First, the increases were not exclusively the result of increased study of disease by marine biologists. Second, factors such as global change may have complex effects on disease. Although some aspects of global change, such as warming and pollution, are predicted to make hosts more susceptible to infection (<xref rid="pbio-0020120-Scott1" ref-type="bibr">Scott 1988</xref>; <xref rid="pbio-0020120-Holmes1" ref-type="bibr">Holmes 1996</xref>), some stressors may impact parasites more than their hosts (<xref rid="pbio-0020120-Lafferty1" ref-type="bibr">Lafferty 1997</xref>). Signs of infection with coldwater disease in salmonids, for example, occur between 4°C and 10°C and disappear as water temperature increases (<xref rid="pbio-0020120-Holt1" ref-type="bibr">Holt et al. 1989</xref>). In addition, stressors that depress host population density may reduce density-dependent transmission of host-specific infectious disease by reducing contact rates between infected and uninfected individuals (<xref rid="pbio-0020120-Lafferty3" ref-type="bibr">Lafferty and Holt 2003</xref>).</p><p>New or increasing stressors, such as global warming, could increase disease if stressed hosts are more susceptible to infection. Elevated sea surface temperature due to El Niño events is a common explanation for coral bleaching (<xref rid="pbio-0020120-Williams1" ref-type="bibr">Williams and Bunkley-Williams 1990</xref>; <xref rid="pbio-0020120-Hoegh-Guldberg1" ref-type="bibr">Hoegh-Guldberg 1999</xref>) and may increase coral susceptibility to disease (<xref rid="pbio-0020120-Harvell2" ref-type="bibr">Harvell et al. 2001</xref>). Increases in turtle and mollusc disease also appear temperature-related. Green turtle fibropapilloma tumors are hypothesized to grow rapidly in summer and reach a debilitating size by winter, when cold water temperatures further stress turtles, resulting in winter strandings (<xref rid="pbio-0020120-Herbst1" ref-type="bibr">Herbst 1994</xref>). The geographic range of the oyster parasite <named-content content-type="genus-species">Perkinsus marinus</named-content> extended 500 km north owing to an increase in average winter low temperatures (<xref rid="pbio-0020120-Ford1" ref-type="bibr">Ford 1996</xref>). Pollution is another ubiquitous and increasing stressor. Bioaccumulation of lipophillic toxins in marine mammals affects the immune system and increases susceptibility to disease (<xref rid="pbio-0020120-Lafferty2" ref-type="bibr">Lafferty and Gerber 2002</xref>).</p><p>Disease could also increase if transmission increases with host density. Some sea urchins experienced increased populations due to overfishing of their predators, and these high-density populations are associated with bacterial disease (<xref rid="pbio-0020120-Lafferty4" ref-type="bibr">Lafferty and Kushner 2000</xref>). Regulations such as the United States Marine Mammal Protection Act of 1972 fully protect pinniped populations, and several species have increased in abundance to levels where transmission efficiency would be expected to increase.</p><p>The decline in infectious diseases of wild fishes over time corresponds to documented reductions in fish populations through intense fishing (<xref rid="pbio-0020120-Jackson1" ref-type="bibr">Jackson et al. 2001</xref>; <xref rid="pbio-0020120-Myers1" ref-type="bibr">Myers and Worm 2003</xref>). Fisheries that reduce the abundance of a fished species should also reduce infectious disease transmission (<xref rid="pbio-0020120-Dobson1" ref-type="bibr">Dobson and May 1987</xref>). This has been documented in experiments (<xref rid="pbio-0020120-Amundsen1" ref-type="bibr">Amundsen and Kristoffersen 1990</xref>) and in observations of parasite declines associated with overfishing (<xref rid="pbio-0020120-Sanders1" ref-type="bibr">Sanders and Lester 1981</xref>).</p><p>Grouping diseases within taxa could obscure important patterns. For example, the trend for increasing coral disease was driven by coral bleaching (<italic>r<sub>s</sub></italic> = 0.87, <italic>p</italic> < 0.0001), while infectious coral diseases, including infectious bleaching, did not increase over time (<italic>r<sub>s</sub></italic> = 0.13, <italic>p</italic> = 0.4934; see <xref ref-type="fig" rid="pbio-0020120-g001">Figure 1</xref>B). The infectious bleaching literature includes several papers since 1996. To ensure the lack of a significant coral disease trend was not due to multiple papers published on this topic at the end of the time range surveyed, an additional analysis was conducted with all infectious bleaching papers excluded; <italic>r<sub>s</sub></italic> and <italic>p</italic> values did not change (<xref ref-type="table" rid="pbio-0020120-t002">Table 2</xref>).</p><table-wrap id="pbio-0020120-t002" position="float"><label>Table 2</label><caption><title>Normalized Coral Disease Reports</title></caption><graphic xlink:href="pbio.0020120.t002"/><table-wrap-foot><fn id="nt201"><p>Original data include papers on infectious bleaching. <italic>r<sub>s</sub></italic> and <italic>p</italic> values are the same for both analyses. Italics indicate changes in proportions after removal of infectious bleaching literature</p></fn></table-wrap-foot></table-wrap><p>While we did not detect an increase in normalized coral disease reports over time, impacts of disease can be high. The recent shift of dominant corals (<italic>Acropora</italic> to <italic>Agaricia</italic>) on reefs due to white band disease was unprecedented in the last 3,000 y (<xref rid="pbio-0020120-Aronson1" ref-type="bibr">Aronson et al. 2002</xref>). Future research should take a finer-scale look at disease, particularly disease impacts, within each taxonomic group. Further investigation is also warranted to determine why some groups showed no temporal pattern in disease reports.</p><p>We examined temporal trends in disease reports since 1970 to identify groups experiencing increased outbreaks. The strong pattern of increased reports in groups such as turtles, mammals, and urchins reflects perceived changes noted by scientists (<xref rid="pbio-0020120-Harvell1" ref-type="bibr">Harvell et al. 1999</xref>). Trends in other groups, such as seagrasses and fishes, suggest that an increase in disease did not occur across all taxa. Although this proxy approach does not directly test hypotheses of temporal changes in disease, a strong signal likely reflects an underlying pattern in nature. In the absence of baseline data, this is a useful approach for detecting quantitative trends in disease occurrence. Understanding disease dynamics, including trends in disease occurrence, is fundamental to conserve ecosystems faced with rising anthropogenic stresses.</p></sec><sec id="s4"><title>Materials and Methods</title><sec id="s4a"><title/><p>We searched the Science Citation Index Expanded (5,900 journals, ISI Web of Science versions 1.1 and 1.2) for papers published from 1970 to 2001 with titles containing specific host taxonomic strings alone and in combination with a disease string (<xref ref-type="table" rid="pbio-0020120-t003">Table 3</xref>). We excluded articles clearly about disease in nonnatural settings, such as hatcheries, aquaculture, and mariculture, or about experimental or laboratory infections. Searches for corals were performed twice to quantify reports of bleaching separately from infectious bleaching (e.g., <named-content content-type="genus-species">Vibrio shiloi</named-content> [<xref rid="pbio-0020120-Israely1" ref-type="bibr">Israely et al. 2001</xref>]) and disease. Only titles were searched, as online abstracts are not available for many articles prior to 1990. Searching the complete citation would bias results after 1990 because more text of each publication would be searched.</p><table-wrap id="pbio-0020120-t003" position="float"><label>Table 3</label><caption><title>Taxonomic Groups and Search Strings</title></caption><graphic xlink:href="pbio.0020120.t003"/></table-wrap><p>Abstracts (or entire manuscripts, when necessary) were obtained for articles within the turtle, coral, urchin, mammal, shark/ray, and seagrass literature that appeared to report the same disease event (e.g., multiple reports of the Caribbean <italic>Diadema</italic> urchin mortality). If more than one paper reported an event, only the earliest published report was included in the analysis. Because significance of results was not altered, multiple reports of disease were not removed from mollusc, decapod, and fish literature owing to the large number of publications returned for each group.</p><p>Often, returned titles contained part of the search string, but were not relevant (e.g. “crab nebula” when searching “crab”). Modifications to search strings excluded most irrelevant articles, and titles were read to determine relevance. If more than 50 titles were returned, titles were randomly sorted and the greater of 20% (maximum of 200) or 50 returned titles were read. Total relevant articles were calculated as the proportion of relevant articles read times the total number returned.</p><p>Protocols were tested using raccoon rabies, a disease for which baseline data are available (<xref rid="pbio-0020120-Rupprecht1" ref-type="bibr">Rupprecht and Smith 1994</xref>). Potential biases were considered and tested. Extensive descriptive or taxonomic work early in the study of a group could bias results against a large number of disease reports. If such a bias existed, one would expect both a large number of disease reports and a large number of nondisease publications in the beginning of the literature survey period. Neither prediction is true—the number of both disease reports and nondisease publications either remains relatively constant or increases through time in all groups.</p><p>Frequent publishing by one author could bias results. Papers by the most published author in a taxonomic group were removed from the analysis to determine their effect. Papers on a particular “hot” topic could also bias results, particularly if that topic is disease and inflates normalized disease reports late in the survey period. For example, a recent mortality event could increase scientists' awareness of disease, resulting in increased publishing without a concomitant increase in the phenomenon. This likely does not affect our results because (a) disease is not the only “hot” topic experiencing increased publication rates and (b) while multiple papers on disease may be published, not all are reports of disease in natural populations.</p><p>A 3-y running mean was used to reveal trends obscured by clustered reporting (e.g., a symposium volume on a topic) and time lags between observation and publication (approximately 3 y, determined by comparing event and publication dates). Data were analyzed with Spearman's rank correlation (JMP version 5.0) with α controlled for multiple comparisons using Holm's sequential Bonferroni adjustments.</p></sec></sec>
Chronic Wasting Disease—Prion Disease in the Wild	Could not extract abstract	<contrib contrib-type="author"><name><surname>Bunk</surname><given-names>Steve</given-names></name></contrib>	PLoS Biology	<p>In 1967, mule deer in a research facility near Fort Collins, Colorado, in the United States apparently began to react badly to their captivity. At least, that was the guess of researchers working on the natural history and nutrition of the deer, which became listless and showed signs of depressed mood, hanging their heads and lowering their ears. They lost appetite and weight. Then they died—of emaciation, pneumonia, and other complications—or were euthanized. The scientists dubbed it chronic wasting disease (CWD), and for years they thought it might be caused by stress, nutritional deficiencies, or poisoning. A decade later, CWD was identified as one of the neurodegenerative diseases called spongiform encephalopathies, the most notorious example of which is bovine spongiform encephalopathy (BSE), more commonly known as mad cow disease. Nowadays, CWD is epidemic in the United States. Although no proof has yet emerged that it's transmissible to humans, scientific authorities haven't ruled out the possibility of a public health threat. The media have concentrated on this concern, and politicians have responded with escalated funding over the past two years for fundamental research into the many questions surrounding this mysterious disease.</p><p>Quite apart from how little is yet known about CWD, media interest is reason enough to step up investigation of it, says Mo Salman, a veterinary epidemiologist at Colorado State University in Fort Collins. He's been scientifically involved with BSE, since it was first discovered among cattle in the United Kingdom in 1986. He recalls predicting that lay interest in BSE would wane after five years. Instead, the disease was found in the mid-1990s to be capable of killing humans who ate tainted beef. “I was wrong, and it really changed my way of thinking, to differentiate between scientific evidence and the public perception,” Salman admits. “Because CWD is similar to BSE, the public perception is that we need to address this disease, to see if it has any link to human health.”</p><disp-quote><p> <italic>CWD is the only spongiform encephalopathy known to naturally infect both free-ranging and captive animals, a situation that greatly complicates efforts to monitor, control, or eradicate it.</italic> </p></disp-quote><sec id="s2"><title>Increasing Attention</title><p>In 2001, the United States' Department of Agriculture (USDA) declared an emergency after CWD was first diagnosed in deer east of the Mississippi River, indicating a potential nationwide problem. This year, the USDA is developing a herd certification program to help prevent the movement of infected animals in the game farming industry. This will bolster monitoring already underway in virtually every state, including postmortem examinations of game killed by hunters and by sharpshooters in mass culling operations.</p><p>By June 2003, brain tissue from more than 111,000 animals had been sampled in North America, and 629 were found to have tested positive for CWD. That's a small epidemic compared to the thousands of BSE cases detected in cattle in the United Kingdom, but CWD is thought to be slow-spreading and perhaps lurking undiscovered elsewhere. So far, the United States and Canada are the only countries in which it has been identified, apart from a few imported cases in the Republic of Korea, but surveillance has not been thorough in North America and is virtually nonexistent in the rest of the world.</p><p>Considered 100% fatal once clinical signs develop, CWD has struck three species of the cervid family—mule deer, white-tailed deer, and Rocky Mountain elk—which roam wild and are raised on farms for meat and hunting. It's the only spongiform encephalopathy known to naturally infect both free-ranging and captive animals, a situation that greatly complicates efforts to monitor, control, or eradicate it.</p><p>The economic costs are hard to quantify, but a 2001 survey by the United States Department of Commerce's Bureau of Census shows that big-game hunters nationwide spend more than US$10 billion annually for trips and equipment. By far, their main target is deer. Wildlife watching of large land mammals, principally deer, drew 12.2 million participants in 2001. The North American Deer Farmers Association represents owners of 75,000 cervid livestock raised for their meat and for velvet antler, a health-food supplement made from antlers. These animals are valued at more than US$111 million.</p><p>Over the past two years, the federal government's emphasis on CWD has been “quite high” compared to other wildlife diseases, says USDA staff veterinarian Dan Goeldner. “In no small part, that's because the disease has cropped up in new places, and those are states that have political clout.” It has now been found in ten more states beyond what became known as the endemic region of Colorado and neighboring Wyoming (<xref ref-type="fig" rid="pbio-0020121-g001">Figure 1</xref>). Last year, the USDA received US$14.8 million to monitor and manage the disease, and Goeldner says the department expects to get about US$16 million this year.</p><fig id="pbio-0020121-g001" position="float"><label>Figure 1</label><caption><title>CWD in North America</title><p>CWD has been detected in both free-ranging and captive animals in Wyoming, Colorado, South Dakota, Wisconsin, and Saskatchewan; only in captive herds in Montana, Nebraska, Kansas, Oklahoma, Minnesota, and Alberta; and only in wild animals in New Mexico, Utah, and Illinois. (Figure courtesy of Gary Wolfe and the CWD Alliance.)</p></caption><graphic xlink:href="pbio.0020121.g001"/></fig></sec><sec id="s3"><title>The Prion Diseases</title><p>Those figures don't include scientific research funded by other organizations, such as the US$42.5 million received by the United States' Department of Defense in 2002 to start up a National Prion Research Program. The prion is the protein-like agent that causes transmissible spongiform encephalopathies (TSEs). Its normal function is uncertain, but when it misfolds into an abnormal or “infectious” form, it causes the microscopic holes and globs of toxic, misshapen protein found in the brains of TSE victims. Unlike viruses, prions don't contain nucleic acids—only protein. Without DNA or RNA to issue biochemical commands, abnormal prions shouldn't be able to convert normal prions to the infectious state, but that's exactly what they do (<xref ref-type="boxed-text" rid="box1">Box 1</xref>).</p><p>Prion diseases occur in many species. In domestic sheep and goats, prion diseases occur as scrapie, which has a virtually worldwide distribution. North America and Europe have also reported rare cases of TSE in ranched mink. Humans get kuru, Creutzfeldt-Jacob disease (CJD), and Gerstmann-Sträussler-Scheinker syndrome—all rare—and BSE itself manifests in people as a variant form of CJD. Since the United Kingdom outbreak, BSE has been discovered in more than 20 countries, most recently in North America. As public fear rose of possible CWD transmission to humans who eat infected venison, the United States' Centers for Disease Control and Prevention (CDC) released a report last year of its investigation into several deaths among venison eaters who might have had a TSE. The report concluded that none of the deaths could be attributed to venison, but it nevertheless cautioned that animals showing evidence of CWD should be excluded from the human and animal feed chains (<xref ref-type="boxed-text" rid="box2">Box 2</xref>).</p><p>CWD is the least understood of all the prion diseases. Its origins are unknown and may well never be discovered. The question is largely academic, unless one hypothesis is proven true, that it derives from scrapie. In that case, the knowledge might help in efforts to control the two diseases through herd and flock management.</p><p>Researchers are working to determine the minimum incubation time of CWD before clinical signs appear, now roughly estimated at 15 months in deer and 12—34 months in elk. They're trying to discover whether CWD strains exist that can affect the length of the disease process and different regions of the brain or that can infect different species, including humans. They are also investigating the period during which the prion is passed on, as well as its modes of transmission. They want to know whether disease reservoirs exist in the bodily fluids of hosts, in the environment, or both. They're racing to develop a diagnostic test that can be performed on live animals, enabling identification of the disease before clinical signs appear, which would eliminate the need to kill thousands of apparently healthy animals in areas where CWD is detected. But among the first things they need to clarify are CWD's distribution across North America and its prevalence.</p><disp-quote><p> <italic>“…the disease has cropped up in new places, and those are states that have political clout.”</italic> </p></disp-quote></sec><sec id="s4"><title>An Initial Step: Improved Surveillance</title><p>“Before you can start to control CWD, you need to understand where it is and how much of it you have,” says veterinary pathologist Beth Williams of the University of Wyoming in Laramie. “So I think you really need surveillance.” Research on its pathogenesis and transmission will help to develop better diagnostic tools, which will improve surveillance, adds Williams, who first identified the disease as a TSE more than a quarter-century ago.</p><p>Colorado State's Salman argues that current surveillance is primarily a series of reactions to reported cases, rather than a systematic strategy designed to determine where and at what prevalence the disease exists and where it's absent. The estimated prevalence is about 1% in elk and 2.5% in deer. But Salman says, “We don't have a good idea of areas in which we are saying we haven't found the disease because these areas are not yet, in my estimation, negative for the disease. Scrapie is a wonderful example of systematic surveillance but, to be fair to the decision-makers and technical people involved with CWD, surveillance on wildlife species is very difficult.”</p><p>The USDA's Goeldner declares, “We have the goal and the hope to eradicate the disease from the farm population.” But Colorado Department of Wildlife veterinarian and CWD expert Mike Miller warns, “Given existing tools, it seems unlikely that CWD can be eradicated from free-ranging populations once established.”</p><p>The gold standard of diagnosis is based on examination of the brain for spongiform lesions and abnormal prion aggregation. Suspect animals are decapitated and their bodies incinerated. “This is an approach that nobody wants, including the people who have to implement it,” says wildlife ecologist Michael Samuel, principal investigator in the United States Geological Survey–Wisconsin Cooperative Wildlife Research Unit at the University of Wisconsin in Madison.</p><p>Nevertheless, when three white-tailed deer shot by hunters in the south-central part of that state during the fall of 2001 were diagnosed with CWD, the state government took swift action. By the spring of 2003, almost 40,000 deer had been sacrificed and sampled for the disease, both within and without a 411 square-mile (1065 square-kilometer) region dubbed the eradication zone. There the goal was to remove as many deer as possible, whereas the plan in contiguous outlying areas was to reduce density to about ten deer per square mile. CWD is thought to spread more efficiently in high-density populations, and normal densities in Wisconsin are 50–100 deer per square mile, about five times that of Colorado and Wyoming. The main objectives of the Wisconsin culling were to discover where the disease existed and its prevalence in affected areas. In the eradication zone, it was 6%–7%, although in the outlying region it was only 1%–2%. Samples elsewhere in the state tested negative.</p></sec><sec id="s5"><title>In Search of a Live Assay</title><p>A key to combating the spread of CWD is to put into widespread use a preclinical diagnostic test on live animals. Miller and colleagues recently developed and validated the first such assay, based on a biopsy of lymphoid tissue, where the infectious agent is known to incubate. They showed that tonsillar biopsies taken from live animals can confirm disease at least 20 months prior to death and up to 14 months before the onset of clinical signs. Although the method is a useful screening tool, it requires much time and training. Each deer must be anaesthetized and blindfolded, placed in a restraint, its mouth held open with a gag, the tonsil visualized with a laryngoscope, and the biopsy taken with endoscopic forceps. Lymphoid tissue sampling was first used as a preclinical test in sheep scrapie. “Many attempts have been made to develop and evaluate tests for live animals, but it is fraught with difficulties,” declares TSE specialist Danny Matthews of the United Kingdom Government's Veterinary Laboratories Agency in Weybridge. He says that a live test for BSE in cattle is likely to be evaluated shortly by the European Food Safety Authority, but warns of a major problem: test samples are collected early in the incubation, whereas brain pathology only arises two to three years later. This creates long delays in determining whether a positive preclinical test result is, in fact, accurate: “How can one do an appropriate evaluation?”</p><p>Matthews notes that blood appears to be a useful medium for testing scrapie in sheep, but current technology cannot deliver a tool applicable across a range of different scrapie genotypes. “Like sheep, elk and mule deer do have a peripheral pathogenesis, which suggests that the blood test route may have some potential, especially if the genotype variability is more restricted than in sheep.”</p></sec><sec id="s6"><title>Transmission Mysteries</title><p>Scrapie can be vertically transmitted from mother to offspring, either in the womb or from the transfer of infected germ plasm. It also can be transmitted horizontally, from any one animal to another. CWD, the only other known contagious TSE, is thought to be transmitted solely by as-yet-undetermined direct or indirect horizontal contact. It probably is not transmitted through infected feed, as is the case for BSE.</p><p>A number of scientists are currently on the trail of suspected CWD disease reservoirs. Saliva is a leading candidate, because clinical signs of CWD include excessive thirst, drinking, and drooling. Work with lab animals suggests that the infectious agent might be produced in salivary glands and, if so, it could be transmitted through social interactions. Feces is also a possible reservoir because animals nose in the ground for feed, and urine is yet another candidate, because it is involved in the scenting activities of cervids.</p><p>Soil could be an environmental reservoir, because cervids ingest dirt to supplement their diets with minerals. Bucks also lick soil on which does have urinated to ascertain their mating status. University of Wisconsin soil science professor Joel Pedersen has discovered that abnormally folded prions stick to the surface of some soil types, such as clay, resisting environmental and chemical damage. “Captive elk contracted CWD when introduced into paddocks occupied by infected elk more than 12 months earlier, despite fairly extensive efforts to disinfect the enclosures,” Pedersen notes. He has begun a five-year project to characterize interactions between infectious prions and soil particles and determine the extent to which infectivity is retained.</p><p>No matter how CWD is transmitted between cervids, the likelihood of human susceptibility seems low. Laboratory evidence has demonstrated a molecular barrier against such cross-species infection, based on the failure of abnormal cervid prions to efficiently convert normal human prions to the infectious state. Likewise, abnormal cervid prions don't easily convert normal cattle prions, suggesting that cattle won't get CWD and pass it on to humans who eat tainted beef. While cattle can contract CWD if inoculated with the infectious agent, long-term studies placing cattle in close proximity to diseased cervids have resulted in no cases of natural transmission. Williams summarizes what all this suggests: “Never say never, but based on the [molecular] work, the CDC's findings, and the epidemiology, we certainly don't have evidence that humans have gotten CWD.”</p></sec><sec><title/><fig id="pbio-0020121-g002" position="float"><label>Figure 2</label><caption><title>Identifying Animals at Risk from CWD</title><p>A raccoon family feeds on a deer carcass staked out by researchers at the University of Wisconsin, in a study aimed at determining which species could be at risk of contracting CWD. (Photo courtesy of the Wisconsin Cooperative Wildlife Research Unit, University of Wisconsin-Madison.)</p></caption><graphic xlink:href="pbio.0020121.g002"/></fig><boxed-text id="box1" position="float"><caption><title>Box 1. How Prions Confound Research</title></caption><p>The relative newness to science of CWD and mad cow disease is one reason they aren't well understood, but sheep scrapie was first identified in Great Britain in 1732—and it still isn't well-characterized. The main problem is the numerous roadblocks to researchers posed by prions, the disease agents of such TSEs.</p><p>Because normal and abnormal prions have identical amino acid sequences, the immune system neither recognizes an infection nor mounts a prion-specific response. Accordingly, an antibody specific to prions has not yet been identified. Without nucleic acid, prions can't be detected or analyzed using conventional techniques such as polymerase chain reaction. They also are extraordinarily resistant to a range of treatments that typically kill or inactivate infectious agents, such as ultraviolet and ionizing radiation, heating, and most chemical disinfectants. The infectious form is largely resistant to degradation by protease enzymes, and in laboratory animals it can incubate for months to years before clinical disease signs appear. Finally, prion diseases must compete for space in expensive, biohazard-safe labs. It's therefore unsurprising that knowledge of these diseases has not sped forward. Still, scientists hope that the recent upsurge of research into BSE, CWD, and scrapie in the United States and Europe will produce synergistic results for preventing and controlling all TSEs.</p></boxed-text><boxed-text id="box2" position="float"><caption><title>Box 2. Who Else Might Get CWD?</title></caption><p>When mad cow disease broke out in the United Kingdom in the 1980s, cattle and humans were far from the only species found to be affected. Among other bovids in zoo and research colonies that contracted spongiform encephalopathy from tainted beef were nyala, gemsbok, eland, Arabian and scimitar-horned oryx, greater kudu, and North American bison. A feline version of the disease was found in domestic cats, cougars, cheetahs, ocelots, and tigers. Among primates, rhesus macaques and lemurs were also infected.</p><p>Unlike BSE, CWD is not thought to be transmitted through feed. But three species of cervids are naturally susceptible, and the question arises of how many other species might be in danger. To help answer that question, Michael Samuel and colleagues at the University of Wisconsin are staking out deer carcasses to see which scavengers come to feed. With flashlit photography, they've discovered “an amazing cast of characters,” including hawks, owls, crows, dogs, cats, coyotes, raccoons, skunks, mink, foxes, and opossums (<xref ref-type="fig" rid="pbio-0020121-g002">Figure 2</xref>). Mammalian scavengers in the state's CWD-affected region will later be examined for disease.</p></boxed-text></sec>
A Wee Lesson in Science Communication	Could not extract abstract	<contrib contrib-type="author"><name><surname>King</surname><given-names>Emma</given-names></name></contrib>	PLoS Biology	<p>The need for effective communication of research and the promotion of science is more important then ever. Public scepticism of research is high, and the number of students studying science continues to dwindle. In an attempt to combat this, the University of Edinburgh encourages its Ph.D. students to participate in a broad programme of science communication activities, designed to develop transferable skills they can use outside pure research and to enhance public engagement in science (PES).</p><p>Included as part of the transferable skills programme, the University hosts a science communication module. The course provides an introduction to the different media, educational, ethical, and political issues surrounding the communication of science to a nonspecialist audience. Initially, the course addresses the presentation of science in written and oral forms. However, as the presentation of science to the public is not simply a practical skill, part of the course is dedicated to the tactical communication of science in society and the relationship between scientists and the media. Finally, students are introduced to ongoing PES opportunities. Any involvement requires the approval of supervisors, to ensure no adverse effects on academic performance. The activities available are diverse and flexible, and they enable students to undertake projects that reflect personal interest and availability.<xref ref-type="fig" rid="pbio-0020122-g002"/> <xref ref-type="fig" rid="pbio-0020122-g002"/> </p><fig id="pbio-0020122-g002" position="float"><graphic xlink:href="pbio.0020122.g002"/></fig><p>What opportunities are available? For enthusiastic students, membership of the Nikon–University of Edinburgh Post-Graduate Science Communication Team (PGSCT) is possible. The PGSCT requires a commitment of 15 days per academic year to paid science communication projects and events, including compulsory support of SCI-FUN, the University's established science and technology roadshow. The team is recruited annually from students of the Science and Engineering, Medicine, and Veterinary Medicine Colleges and takes a leading role in PES activities. Team members are supported by other graduate students who participate on a more casual basis.</p><p>To ensure that a broad spectrum of activities is available to audiences, the PGSCT members are actively encouraged to design and develop their own ideas. An opportunity to do this is provided through the University's Science Zone, at the Royal Museum of Scotland, for the duration of the Edinburgh International Science Festival. From successful workshops, originally piloted at the science festival, several ex-PGSCT members have gone on to establish their own projects. One example of this is the Natural Environment Science Education scheme, which later received recognition through the Royal Society of Edinburgh/Scottish Executive ‘Science in the Community’ Award for 2003. A primary aim of this scheme was to deliver in isolated and remote communities innovative hands-on earth science- and natural science-based activities beyond city venues. For students, developing and presenting workshops are incredibly rewarding and allow them to experience the enthusiasm of participants and fellow presenters.</p><p>Some activities are transferred from the Science Zone to the classroom, where students can communicate with children at all stages of their schooling. At primary school age (5–11 years) after-school science clubs, led by graduate students, provide an opportunity for presenters to share with pupils their enthusiasm for science. Alternative activities are directed at secondary school children (12–18 years), with a variety of workshops available across the different scientific disciplines. In particular, The Scottish Institute for Biotechnology Education (SIBE) has been set up within the University to facilitate PES activities such as the popular ‘Green Fingerprinting’ workshop where the principles of DNA fingerprinting are applied to an ecological scenario in the form of a practical activity (<xref ref-type="fig" rid="pbio-0020122-g001">Figure 1</xref>). The majority of workshops coordinated by SIBE have been funded by the Biotechnology and Biological Sciences Research Council (BBSRC).</p><fig id="pbio-0020122-g001" position="float"><label>Figure 1</label><caption><title>Students Studying a Stained Agarose Gel at a Green Fingerprinting Workshop (Photograph, with permission, by Douglas Robertson)</title></caption><graphic xlink:href="pbio.0020122.g001"/></fig><p>The presentations may be in person or utilise new technologies such as video conferencing. A BBSRC Dialogue Award has recently been granted to SIBE to design, develop and deliver video-conferences addressing bio-ethical issues surrounding recent advances in biotechnology. Together, the two approaches of in-person and virtual presenting, enhance the schools' curriculum and facilitate dialogue between scientists and the public at a stage where promotion of science can influence the choice of further study and career options.</p><p>The promotion of biotechnology within schools does not stop with the encouragement of school children's enthusiasm; SIBE also works closely with organisations such as ‘Science and Plants for Schools’ (SAPS) to assist with continuing professional development of biology teachers through the organisation of training days aimed at curriculum enhancement and the practical application of biotechnology in the classroom. Interested graduate students can co-present the workshops, though they are led by permanent staff. The students introduce their own research to highlight the applications of biotechnology and provide a useful technical resource.</p><p>For students who prefer to communicate through the written word, though no guarantee of publication can be given, the University is in a position to highlight science-writing opportunities, usually as a contribution to publications specifically concerned with science communication or within the scientific press. At the Institute of Cell and Molecular Biology at the University a ‘press-gang’ meets on a monthly basis to discuss research carried out in the Institute and generate press releases for publication in the university press and national newspapers where appropriate.</p><p>Complementing the University's efforts, many other organisations, such as UK Research Councils and the British Association for the Advancement of Science, endorse graduate students spending time on PES activities. Several schemes have been put in place to facilitate this; Researchers in Residence, the Science and Engineering Ambassadors Scheme and science communication courses. Hosted at the University's science campus—Kings Buildings—‘pgscicom’ provides up-to-date information on opportunities for PES at the University and beyond through regular email communications.</p><p>As a PhD student and PGSCT member, I believe the approach of the University of Edinburgh to PES is of benefit to all involved. The combination of training and practical experience provides graduate students with new and valuable skills and opportunities to develop them further. The events and activities enthuse and engage audiences with the presentation of science in an informal but informative manner.</p><fig id="pbio-0020122-g003" position="float"><graphic xlink:href="pbio.0020122.g003"/></fig>
Yeast Prions: Protein Aggregation Is Not Enough	Could not extract abstract	<contrib contrib-type="author"><name><surname>Sherman</surname><given-names>Michael Y</given-names></name></contrib>	PLoS Biology	<p>Many damaged and mutant polypeptides, as well as some normal proteins, have a tendency to aggregate in cells. Some protein aggregates are capable of “dividing” and propagating in cells, leading to formation of similar aggregates in daughter cells or even in neighboring cells due to “infection.” These self-propagating protein aggregates are called prions and constitute the basis of prion diseases. The infectious agent in these diseases is an abnormal conformation of the PrP protein (PrP<sup>Sc</sup>), which makes it protease-resistant and initiates its aggregation (<xref rid="pbio-0020125-Prusiner1" ref-type="bibr">Prusiner 1998</xref>). The abnormal aggregated species can recruit normal soluble PrP molecules into aggregates, thus inactivating them. The aggregates of PrP<sup>Sc</sup> can proliferate within cells and be transmitted to other cells and tissues, leading to the spread of neurotoxicity.</p><sec id="s2"><title>Prion Domains</title><p>While so far only one prion protein is known in mammals, several prion-like proteins capable of forming self-propagating aggregates have been found in various yeast species. The common structural feature of yeast prion proteins is the so-called prion domain, characterized by the high content of glutamines (Q) and asparagines (N) (<xref rid="pbio-0020125-DePace1" ref-type="bibr">DePace et al. 1998</xref>; <xref rid="pbio-0020125-Michelitsch1" ref-type="bibr">Michelitsch and Weissman 2000</xref>), also known as the Q/N-rich domain. The prion domains are the major structural determinants that are solely responsible for the polypeptide aggregation and propagation of the aggregates. Interestingly, the mammalian PrP<sup>Sc</sup> is fundamentally different from yeast prions, since it lacks a Q/N-rich domain, indicating that distinct structural features are responsible for its ability to form self-propagating aggregates. The Q/N-rich domains in yeast prions are transferable in that, when fused to a heterologous polypeptide, they confer prion properties to this polypeptide. With a low probability, soluble proteins with prion domains can change conformation to form self-propagating aggregates, which can be transmitted to daughter cells (<xref rid="pbio-0020125-Lindquist1" ref-type="bibr">Lindquist 1997</xref>) (<xref ref-type="fig" rid="pbio-0020125-g001">Figure 1</xref>). As with PrP<sup>Sc</sup>, yeast prions efficiently recruit soluble molecules of the same species, thus inactivating them (<xref rid="pbio-0020125-Lindquist1" ref-type="bibr">Lindquist 1997</xref>; <xref rid="pbio-0020125-Chernoff1" ref-type="bibr">Chernoff 2001</xref>; <xref rid="pbio-0020125-Wickner1" ref-type="bibr">Wickner et al. 2001</xref>). Also with low probability, the aggregation-prone conformation of yeast prion proteins can reverse to a soluble functional conformation. Certain yeast prion proteins, when in soluble conformation, function in important pathways; e.g., Sup35 (forming [PSI<sup>+</sup>] prion) controls termination of translation, and Ure2 (forming [URE3<sup>+</sup>] prion) controls some membrane transporter systems. Aggregation of these proteins leads to phenotypes (e.g., suppression of nonsense mutations or transport defects) inherited in a non-Mendelian fashion owing to the nonchromosomal basis of the inheritance.</p><fig id="pbio-0020125-g001" position="float"><label>Figure 1</label><caption><title>Aggregation, Division, and Transfer of Prions in Yeast</title></caption><graphic xlink:href="pbio.0020125.g001"/></fig></sec><sec id="s3"><title>Inheriting Variations</title><p>A remarkable feature of yeast prion proteins is their ability to produce distinct inherited “variants” of the prion. For example, [PSI<sup>+</sup>] prion could exist in several distinct forms that suppress termination of translation to different degrees. These “variants” of yeast prions are analogous to different prion “strains” of PrP<sup>Sc</sup>, which cause versions of the disease with different incubation periods and different patterns of brain pathology. The molecular nature of distinct PrP<sup>Sc</sup> strains is determined by specific stable conformations of PrP. Similarly, “variants” of yeast prions are explained by different stable conformation states of the corresponding prion proteins (<xref rid="pbio-0020125-Chien1" ref-type="bibr">Chien et al. 2003</xref>). Strict conformation requirements for aggregate formation can also explain interspecies transmission barriers, where prion domains of Sup35 derived from other yeast species cannot cause formation of [PSI<sup>+</sup>] prion in Saccharomyces cerevisiae, in spite of a high degree of homology. This observation is very intriguing, especially in light of a recent finding that prion conformation of some proteins is required for formation of prions by the other proteins. For example, for de novo formation of [PSI<sup>+</sup>] prion, a distinct prion [RNQ<sup>+</sup>] should be present in a cell (<xref rid="pbio-0020125-Derkatch1" ref-type="bibr">Derkatch et al. 2001</xref>; <xref rid="pbio-0020125-Osherovich1" ref-type="bibr">Osherovich and Weissman 2001</xref>), probably in order to cross-seed Sup35 aggregates. This is in spite of relatively limited homology between the prion domains of these proteins. The apparent contradiction between the interspecies transmission barriers of very homologous prion proteins and possible cross-seeding of aggregates by prion proteins with more limited homology represents an interesting biological problem. On the other hand, this apparent contradiction may indicate that prion formation is a more complicated process than we currently think and that it may involve many cellular factors.</p></sec><sec id="s4"><title>What Do Prions Do?</title><p>Although yeast prions have been studied for almost ten years, very little is known about their biological significance. We do not know the functions of the majority of proteins that can exist as prions. Even if a function of prion proteins, such as with Sup35 or Ure2, is known, we do not understand the biological significance of their “prionization,” i.e., that they aggregate and propagate in the aggregated form. A very intriguing and unexpected finding was that formation of [PSI<sup>+</sup>] prion causes a wide variety of phenotypic alterations, which depend on the strain background (<xref rid="pbio-0020125-True1" ref-type="bibr">True and Lindquist 2000</xref>). In fact, comparison of yeast strains of different origin, each with and without [PSI<sup>+</sup>] prion, showed that certain strains with [PSI<sup>+</sup>] prion have different sensitivity to stresses and antibiotics than their non-prion derivatives, despite their genetic identity. In some strains, cells with [PSI<sup>+</sup>] prion demonstrated better survival than their non-prion counterparts in the presence of inhibitors of translation or microtubules, heavy metals, low pH, and other deleterious conditions, which of course gives a strong advantage to the [PSI<sup>+</sup>] cells. It is likely that some genomic mutations could be suppressed and therefore become silent when termination of translation by Sup35 is partially inactivated in [PSI<sup>+</sup>] prion cells (<xref rid="pbio-0020125-Lindquist2" ref-type="bibr">Lindquist 2000</xref>; <xref rid="pbio-0020125-True1" ref-type="bibr">True and Lindquist 2000</xref>). [PSI<sup>+</sup>] could also reveal previously silent mutations or their combinations. It was hypothesized that switches between prion and non-prion forms of Sup35 enhance survival in fluctuating environments and provide a novel instrument for evolution of new traits.</p></sec><sec id="s5"><title>Q/N Does Not Necessarily a Prion Make</title><p>Searching genomes of various species demonstrated that a relatively large fraction of proteins (between 0.1% and 2%) contain Q/N-rich domains (<xref rid="pbio-0020125-Michelitsch1" ref-type="bibr">Michelitsch and Weissman 2000</xref>) or polyQ or polyN sequences. These domains are often found in transcription factors, protein kinases, and components of vesicular transport. The Q/N-rich domains usually are not evolutionary conserved and their functional role is largely unknown. Some of the Q/N-rich or polyQ domains facilitate aggregation of polypeptides, especially if expanded owing to mutations. Such expansion of the polyQ domains in certain neuronal proteins could cause neurodegenerative disorders, e.g., Huntington's disease or several forms of ataxia. Importantly, aggregates formed by polypeptides with the Q/N-rich or polyQ domains are not necessarily self-propagating aggregates, i.e., prions. In fact, there are additional structural properties of the polypeptides that provide the self-propagation (see below). Even if a protein with a polyQ domain does not form a prion, its aggregation may depend on certain prions. For example, recent experiments demonstrated that [RNQ<sup>+</sup>] prion dramatically stimulated aggregation of fragments of recombinant human huntingtin or ataxin-3 with an expanded polyQ domain cloned in yeast (<xref rid="pbio-0020125-Osherovich1" ref-type="bibr">Osherovich and Weissman 2001</xref>; <xref rid="pbio-0020125-Meriin1" ref-type="bibr">Meriin et al. 2002</xref>). [RNQ<sup>+</sup>] facilitated the nucleation phase of the huntingtin fragment aggregation, suggesting that this prion can be directly involved in seeding of the aggregates. The major question now is whether there are analogous prion-like proteins in mammalian cells that are involved in aggregation of huntingtin or ataxin-3 and subsequent neurodegenerative disease.</p><p>The first indication that mammalian proteins with Q/N-rich domains can form self-propagating prions came from recent work with a regulator of translation cytoplasmic polyadenylation element-binding protein (CPEB) from <italic>Aplysia</italic> neurons (<xref rid="pbio-0020125-Si1" ref-type="bibr">Si et al. 2003</xref>). The neuronal form of this protein has a Q/N-rich domain similar to the prion domains of yeast prions. The Q/N-rich domain from CPEB (CPEBQ), when fused to green fluorescent protein (GFP), conferred upon it prion-like properties. The CPEBQ–GFP fusion polypeptide existed in yeast cells in one of the three distinct states, i.e., soluble, many small aggregates, or few large aggregates. Mother cells almost always gave rise to daughter cells in which the CPEBQ–GFP polypeptide was in the same state, indicating the ability of these aggregates to be inherited, i.e., to self-propagate. Furthermore, full-length <italic>Aplysia</italic> CPEB protein, when cloned in yeast, can also exist in two distinct states, soluble and aggregated, which is an inherited feature. Very unexpectedly, unlike other prions, the aggregated state of CPEB was more functionally active than the soluble form (<xref rid="pbio-0020125-Si1" ref-type="bibr">Si et al. 2003</xref>). These data strongly suggest that metazoan proteins with Q/N-rich domains are potentially capable of forming prions. The challenge now will be to establish whether CPEB can exist as a self-propagating aggregate in <italic>Aplysia</italic> or mammalian neurons.</p></sec><sec id="s6"><title>Mystery of Propagation</title><p>What makes protein aggregates in yeast propagate? The key cellular element that is critical for this process is molecular chaperone Hsp104 (<xref rid="pbio-0020125-Chernoff2" ref-type="bibr">Chernoff et al. 1995</xref>). This factor is specifically required for maintenance of all known prions within generations and probably is not involved in prion formation (i.e., initial protein aggregation). [PSI<sup>+</sup>] yeast cells have about 60 seeds of this prion (although this number differed in different [PSI<sup>+</sup>] isolates), and maintenance of about this number of seeds after cell divisions requires functional Hsp104 (<xref rid="pbio-0020125-Eaglestone1" ref-type="bibr">Eaglestone et al. 2000</xref>). In fact, in the absence of Hsp104, prion aggregates continue to grow without increase in number and are rapidly lost in generations (<xref rid="pbio-0020125-Wegrzyn1" ref-type="bibr">Wegrzyn et al. 2001</xref>). Since this chaperone can directly bind to protein aggregates and promote there disassembly (<xref rid="pbio-0020125-Glover1" ref-type="bibr">Glover and Lindquist 1998</xref>), it was suggested that the main function of Hsp104 in prion inheritance is to disaggregate large prion aggregates to smaller elements, thus leading to formation of new seeds (<xref rid="pbio-0020125-Kushnirov1" ref-type="bibr">Kushnirov and Ter-Avanesyan 1998</xref>). Interestingly, although Hsp104 is conserved among bacteria, fungi, and plants, animal cells do not have this chaperone or its close homologs. Therefore, if yeast-type prions with Q/N-rich domains exist in animal cells, there should be alternative factors that disaggregate large prion aggregates into smaller species in order to keep the number of seeds relatively constant and thus maintain the prions.</p><p>The fact that some proteins with Q/N-rich domains form self-propagating aggregates, while others can aggregate but cannot form prions, suggests that there should be some structural elements either within the Q/N-rich sequence or close to it that confer the ability to propagate. In an article in this issue of <italic>PLoS Biology</italic> by <xref rid="pbio-0020125-Osherovich2" ref-type="bibr">Osherovich et al. (2004)</xref>, the authors examined sequence requirements for prion formation and maintenance of two prion proteins, Sup35 and New1. They noted that both prion proteins contain an oligopeptide repeat QGGYQ in close proximity to Q/N-rich sequences and examined the functional significance of the repeats for aggregation and maintenance of the prions. In New1, in contrast to a deletion of the N-rich domain, deletion of the repeat did not affect aggregation of the protein or formation of the prion, but abrogated inheritance of the prion. With Sup35, the situation was somewhat more complicated, since repeats adjacent to Q/N-rich domain affected both protein aggregation and prion maintenance while more distant repeats affected only the prion inheritance. The authors suggested that the oligopeptide repeats facilitate the division of aggregates, either by serving as binding sites for Hsp104 or by altering the conformation of the polypeptides in aggregates to promote access for Hsp104 (<xref ref-type="fig" rid="pbio-0020125-g002">Figure 2</xref>).</p><fig id="pbio-0020125-g002" position="float"><label>Figure 2</label><caption><title>Distinct Domains of Sup35 Are Responsible for Aggregation and Division of Aggregates</title></caption><graphic xlink:href="pbio.0020125.g002"/></fig><p>The likely possibility was that the oligopeptide repeats could be interchangeable between different prions, leading to creation of novel chimeric prions. In fact, the authors constructed an F chimera, a fusion protein having the N-rich domain of New1 and the oligopeptide repeat of Sup35. This fusion polypeptide efficiently formed prion [F<sup>+</sup>]. Furthermore, when the oligopeptide repeat sequence was added to a polyQ sequence, this fusion polypeptide also acquired the ability to form self-propagating aggregates. This work, therefore, clarifies the architecture of prions by defining two structural motifs in prion proteins that have distinct functions in aggregation and propagation. Interestingly, not all yeast prions have similar oligopeptide repeat motifs, indicating that distinct structures could confer prion properties to polypeptides that can aggregate. It would be important to identify these structures in order to understand the mechanisms of aggregate propagation. The work of <xref rid="pbio-0020125-Osherovich2" ref-type="bibr">Osherovich et al. (2004)</xref> may help to identify proteins from mammalian cells, plants, and bacteria that can potentially form prions. Finding these novel prions could be of very high significance since they may provide insight into a wide range of currently unexplained epigenetic phenomena.</p></sec>
Galvanising mental health research in low- and middle-income countries: Role of scientific journals	Could not extract abstract	<contrib id="A1" corresp="yes" contrib-type="author"><collab>Editors and WHO November 2003 Group</collab><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Annals of General Hospital Psychiatry	<sec><title/><p>The Department of Mental Health and Substance Abuse, WHO organized a meeting on <bold><italic>Mental Health Research in Developing Countries: Role of Scientific Journals </italic></bold>in Geneva on 20 and 21 November 2003 that was attended by twenty-five editors representing journals publishing mental health research. A number of other editors reviewed and contributed to the background and follow-up material. This statement is issued by all participants jointly (see Appendix 1 for the list of journals/organizations and their representatives).</p><p>Research is needed to address the enormous unmet mental health needs of low- and middle-income (LAMI) countries. Scientific journals play an important role in production and dissemination of research. However, at present, only a minute proportion of research published in widely accessible mental health and psychiatric journals is from or about these countries. Yet over 85% of the world's population lives in the 153 countries categorized as low- and middle-income, according to World Bank criteria. Even more worrying is the observation that the gap between these and high-income countries may be widening in terms of their number of publications. The meeting was aimed at finding ways of resolving this unsatisfactory situation.</p></sec><sec><title>Responsibility of scientific journals towards international mental health</title><p>Science, in its quest to accomplish valid generalisations about nature, is inherently global. Researchers from all parts of the world should, desirably, contribute to new knowledge about mental health and mental illness, and publish their reports in widely accessible journals. This process is facilitated by a shared understanding of aims and scientific methods, formats of presentation and reference to previous published work. Mental health research from LAMI countries is needed for advocacy, policy development, establishment and expansion of clinical services and to educate investigators in research skills. A steady stream of information about mental health issues in these countries would also contribute to a greater international and multicultural understanding of mental health and ill-health.</p><p>Unfortunately, substantial barriers impede publication of mental health research from LAMI countries in widely accessible journals. Researchers from LAMI countries are often unable to meet the requirements of these journals because of limited access to information, lack of advice on research design and statistics, difficulty in writing in a foreign language, and overall material, financial, policy and infrastructural constraints. Limited appreciation of the research needs of, and realities in LAMI countries and the comparative anonymity of their researchers and research centres in editorial offices of journals may constitute additional barriers. Many researchers from LAMI countries are daunted by the seemingly insurmountable chasm between their research effort and its publication in international journals.</p></sec><sec><title>Supporting mental health researchers from low- and middle-income countries</title><p>We need to face the challenge of reducing the barriers to publication of mental health research by investigators working in LAMI countries. Time, skills, resources and commitment are needed to publish relevant studies from these countries. Editors' and reviewers' experience with and interest in LAMI countries could be an asset in facilitating publication. Meeting researchers from these countries on 'their home ground' could assist this process. International journals could also help researchers improve their submissions by diligent assessment, detailed recommendations for revision and sympathetic consideration of revised versions, even if it means requesting reviewers to 'take an extra round' to make papers suitable for publication. This is not to say that journals need to lower their standards in publishing papers from LAMI countries; rather, they should devise strategies to help authors attain those standards. Other approaches to support contributions from LAMI countries could be to launch 'starter' sections such as information pages and special columns or even dedicated issues of the journal.</p><p>Capacity building is the paramount factor in the long term. Training in research methodology and scientific writing is needed. This could be done through mentoring, personal encouragement, training courses and research collaboration. Increased access to mental health research publications would, by itself, help in capacity building.</p></sec><sec><title>Supporting mental health journals from low- and middle-income countries</title><p>A major impediment in accessing mental health research from LAMI countries is the lack of visibility of journals published in these countries. Most of them are not indexed in international databases and are often not available beyond their country or region of origin. These journals are published under strained circumstances, in that they often lack sound financial support and have a hard time becoming self-sufficient. They also have difficulty in obtaining suitable articles for publication because their author pool is limited; moreover, influential authors from this pool prefer to publish their best research in indexed journals. Some authors who submit their articles to LAMI country based journals may have limited skills in conducting research and/or in writing up their reports. However, it must be stressed that some excellent work does find publication in these journals.</p><p>The task of strengthening journals in LAMI countries begins from the recognition of their role as contributors to the enhancement of the mental health knowledge base and as partners in the international research community. Editors of LAMI country based journals require support to elevate standards in editorial procedures, peer review and overall journal management since sufficient expertise and experience may be lacking. This could be achieved through their participation in the publication process of established journals, mentorship, twinning arrangements and training workshops.</p></sec><sec><title>Enhancing dissemination of mental health research publications</title><p>Many high quality mental health journals have a wide distribution, but most of their subscribers are from high-income countries. Special attention to dissemination of research findings is needed urgently in order to maximize their impact on mental health policy and practice and advance relevant research in LAMI countries. Increasing online availability is cost-effective since little additional expenditure is required to provide access to new users apart from the initial costs of posting material on a website. Free access to many categories of electronic resources is provided by many journals. Initiatives such as the WHO-led Health InterNetwork Access to Research Initiative (HINARI) offer institutions in LAMI countries electronic access to thousands of journals at no or very low cost. The Open Access model provides free online access along with the possibility of unrestricted dissemination of research materials, but charges for publication may be prohibitive for authors from LAMI countries unless support comes from funding agencies and governments, e.g. the Scientific Electronic Library Online (SciELO) project in Latin America. Governments in other LAMI countries need to be made aware of the opportunities provided by information technology for dissemination and application of research knowledge.</p></sec><sec><title>The role of various stakeholders</title><p>Editors of journals, editors' associations and international organizations, including WHO could help achieve the aforementioned objectives. A catalogue of ideas is presented in Appendix 2 to act as a starting point for specific action. Although these ideas have been developed for the field of mental health, many of them may apply to other areas of health.</p></sec><sec><title>Appendix 1: Participants</title><p><bold><italic>Acta Psychiatrica Scandinavica </italic></bold>(Povl Munk-Jorgensen), <bold><italic>American Journal of Orthopsychiatry </italic></bold>(Carlos Sluzki), <bold><italic>Annals of General Hospital Psychiatry </italic></bold>(George St. Kaprinis, Konstantinos N. Fountoulakis), <bold><italic>Anthropology and Medicine </italic></bold>(Sushrut Jadhav), <bold><italic>Australian and New Zealand Journal of Psychiatry </italic></bold>(Sidney Bloch), <bold><italic>BioMed Central Psychiatry </italic></bold>(Pritpal S. Tamber), <bold><italic>British Journal of Psychiatry </italic></bold>(Peter Tyrer), <bold><italic>BMJ </italic></bold>(Kamran Abbasi), <bold><italic>Bulletin of World Health Organization </italic></bold>(Hooman Momen), <bold><italic>Child Abuse and Neglect, The International Journal </italic></bold>(John M. Leventhal), <bold><italic>Chinese Journal of Nervous and Mental Disease </italic></bold>(Li Yingxi, Guan Jinli), <bold><italic>Comprehensive Psychiatry </italic></bold>(David L. Dunner), <bold><italic>Culture, Medicine and Psychiatry </italic></bold>(Mary-Jo Delvecchio Good), <bold><italic>Epidemiologia e Psichiatria Sociale </italic></bold>(Michele Tansella), <bold><italic>L'Evolution Psychiatrique </italic></bold>(Yves Thoret), <bold><italic>Indian Journal of Psychiatry </italic></bold>(Utpal Goswami), <bold><italic>L'Information Psychiatrique </italic></bold>(Thierry Tremine), <bold><italic>International Journal of Social Psychiatry </italic></bold>(Dinesh Bhugra), <bold><italic>International Psychiatry </italic></bold>(Hamid Ghodse), <bold><italic>Journal of Child and Adolescent Mental Health </italic></bold>(Alan Flisher), <bold><italic>Journal of Nervous and Mental Disease </italic></bold>(Eugene B. Brody, Kathy McKnight), <bold><italic>Lancet </italic></bold>(Laragh Gollogly), <bold><italic>Primary Care Psychiatry </italic></bold>(Sean Lynch), <bold><italic>Psychiatry: Interpersonal and Biological Processes </italic></bold>(Robert Ursano), <bold><italic>Psychiatry Research </italic></bold>(Monte Buchsbaum), <bold><italic>Psychological Medicine </italic></bold>(Eugene Paykel), <bold><italic>Psychology and Psychotherapy: Theory, Research and Practice </italic></bold>(Phil Richardson), <bold><italic>Psychopathologie Africaine </italic></bold>(Momar Gueye), <bold><italic>Quarterly Journal of Pakistan Psychiatric Society </italic></bold>(Amin A. Gadit), <bold><italic>Revista Brasileria de Psiquiatria </italic></bold>(Jair Mari), <bold><italic>Salud Mental </italic></bold>(Hector Perez-Rincon), <bold><italic>Social Psychiatry and Psychiatric Epidemiology </italic></bold>(Paul Bebbington), <bold><italic>South African Journal of Psychiatry </italic></bold>(Robin Emsley, Susan Hawkridge), <bold><italic>Transcultural Psychiatry </italic></bold>(Laurence J. Kirmayer), <bold><italic>World Psychiatry </italic></bold>(Mario Maj), <bold><italic>Forum for African Medical Editors </italic></bold>(James K. Tumwine), <bold><italic>Global Forum for Health Research </italic></bold>(Andres de Francisco), <bold><italic>World Association of Medical Editors </italic></bold>(Ana Marusic, Peush Sahni), <bold><italic>World Health Organization </italic></bold>(Shekhar Saxena, Pratap Sharan, Benedetto Saraceno, Barbara Aronson, Vladimir Poznyak, Izthak Levav, Edith Certain, R Srinivasa Murthy, Tikki Pang).</p><p>Shekhar Saxena, Pratap Sharan, Hooman Momen and Benedetto Saraceno organized the WHO Meeting leading to this joint statement.</p></sec><sec><title>Appendix 2: Catalogue of ideas</title><sec><title>Individual journals</title><sec><title>Giving priority to relevant mental health research from low- and middle-income countries</title><p>• Educate editors and reviewers on research needs of and research infrastructure in LAMI countries;</p><p>• Use surveys of various stakeholders such as readers (including those from other regions) for shaping journals' priorities;</p><p>• Sensitize readers and other stakeholders to international mental health issues (e.g. through special sections and dedicated issues, guest editorship and the commissioning of relevant research from LAMI countries);</p><p>• Critically re-examine the use and limitation of measures such as citation rates and impact factors;</p><p>• Adopt a multilingual approach, such as translation of relevant articles and abstracts into other languages;</p><p>• Include reviewers and correspondents with a special interest and expertise in LAMI countries on editorial boards;</p><p>• Accept a higher proportion of submissions from LAMI countries for review; and</p><p>• Encourage general medical journals to publish mental health research especially in countries/regions where no mental health journal exists at present.</p></sec><sec><title>Supporting authors/researchers from low- and middle-income countries</title><p>• Familiarize researchers from LAMI countries with the peer review process;</p><p>• Provide constructive critical feedback/detailed recommendations for revision;</p><p>• Make provision for extra rounds of editing, assistance with language and use of technical editors;</p><p>• Pay attention to the educational goals of the review process (e.g. availability of reviewer's comments to readers or recruiting young researchers in LAMI countries to referee papers);</p><p>• Provide mentorship and support prior to submission;</p><p>• Organise training workshops for LAMI country researchers and students on scientific writing and research methodology;</p><p>• Facilitate the involvement of researchers in multi-centre projects and research groups;</p><p>• Accept and process submissions online; and</p><p>• Devise strategies to prevent economic exclusion of researchers from LAMI countries in author/input paying publishing models.</p></sec><sec><title>Supporting journals from low- and middle-income countries</title><p>• Support "twinning" or "pairing" arrangements, such as invited editorials, exchange of journals, cross-publication of contents/abstracts/summaries/articles and joint publications;</p><p>• Agree to serve on editorial boards or as reviewers;</p><p>• Agree to mentor reviewers and editors;</p><p>• Provide training workshops for editors and reviewers; and</p><p>• Support national/regional journals in developing their own websites and/or seeking inclusion in specialized websites on mental health</p></sec><sec><title>Enhancing Dissemination</title><p>• Participate in electronic dissemination initiatives or provision of free/open access through the journal's website;</p><p>• Participate in "buddy system"/peer sponsoring initiatives;</p><p>• Employ user-friendly technology for easier downloads;</p><p>• Subsidize journal subscriptions for LAMI countries; and</p><p>• Explore mechanisms for publication of selected papers in more than one journal for wider dissemination.</p></sec></sec><sec><title>Editors' associations</title><p>• Develop guidelines for good editorial practice concerning publishing and research ethics and conflicts of interest;</p><p>• Facilitate access to literature and bibliographic services (e.g. through a directory of databases);</p><p>• Support authors to access appropriate specialized journals and specific audience (e.g. through a database of journals and instructions to authors);</p><p>• Facilitate mentoring for editors, reviewers and researchers;</p><p>• Organise training of editors, reviewers and researchers from LAMI countries; and</p><p>• Facilitate the multidirectional flow of articles, resources and expertise (e.g. translation of relevant articles and support with information technology).</p></sec><sec><title>International organizations</title><sec><title>Supporting mental health research, research infrastructure and publications</title><p>• Influence other international institutions to give priority to mental health research in their agendas for LAMI countries;</p><p>• Support national institutions in LAMI countries to urge their governments to give higher priority to mental health research;</p><p>• Support inclusion of researchers/editors from LAMI countries in relevant decision-making forums; and</p><p>• Facilitate capacity building for researchers and journals from LAMI countries.</p></sec><sec><title>Enhancing Dissemination</title><p>• Assess information needs in LAMI countries and raise awareness of these;</p><p>• Provide access to journals publishing mental health research (e.g. expansion of HINARI or enabling journals to be open access); and</p><p>• Encourage and facilitate the application of information technology.</p></sec><sec><title>Enhancing Collaboration</title><p>• Develop networks between editors, editorial organizations, professional bodies, publishers, funding agencies, national and international organizations and the media); and</p><p>• Adopt a systematic approach for follow up: statement of changes hoped for, development of outcome criteria, assessment of progress.</p></sec></sec></sec>
Nucleic acid amplification tests in the diagnosis of tuberculous pleuritis: a systematic review and meta-analysis	<sec><title>Background</title><p>Conventional tests for tuberculous pleuritis have several limitations. A variety of new, rapid tests such as nucleic acid amplification tests – including polymerase chain reaction – have been evaluated in recent times. We conducted a systematic review to determine the accuracy of nucleic acid amplification (NAA) tests in the diagnosis of tuberculous pleuritis.</p></sec><sec sec-type="methods"><title>Methods</title><p>A systematic review and meta-analysis of 38 English and Spanish articles (with 40 studies), identified via searches of six electronic databases, hand searching of selected journals, and contact with authors, experts, and test manufacturers. Sensitivity, specificity, and other measures of accuracy were pooled using random effects models. Summary receiver operating characteristic curves were used to summarize overall test performance. Heterogeneity in study results was formally explored using subgroup analyses.</p></sec><sec><title>Results</title><p>Of the 40 studies included, 26 used in-house ("home-brew") tests, and 14 used commercial tests. Commercial tests had a low overall sensitivity (0.62; 95% confidence interval [CI] 0.43, 0.77), and high specificity (0.98; 95% CI 0.96, 0.98). The positive and negative likelihood ratios for commercial tests were 25.4 (95% CI 16.2, 40.0) and 0.40 (95% CI 0.24, 0.67), respectively. All commercial tests had consistently high specificity estimates; the sensitivity estimates, however, were heterogeneous across studies. With the in-house tests, both sensitivity and specificity estimates were significantly heterogeneous. Clinically meaningful summary estimates could not be determined for in-house tests.</p></sec><sec><title>Conclusions</title><p>Our results suggest that commercial NAA tests may have a potential role in confirming (ruling in) tuberculous pleuritis. However, these tests have low and variable sensitivity and, therefore, may not be useful in excluding (ruling out) the disease. NAA test results, therefore, cannot replace conventional tests; they need to be interpreted in parallel with clinical findings and results of conventional tests. The accuracy of in-house nucleic acid amplification tests is poorly defined because of heterogeneity in study results. The clinical applicability of in-house NAA tests remains unclear.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Pai</surname><given-names>Madhukar</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Flores</surname><given-names>Laura L</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Hubbard</surname><given-names>Alan</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Riley</surname><given-names>Lee W</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A5" corresp="yes" contrib-type="author"><name><surname>Colford</surname><given-names>John M</given-names><suffix>Jr</suffix></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Infectious Diseases	<sec><title>Background</title><p>In the context of the human immunodeficiency virus (HIV) epidemic, clinicians frequently encounter extrapulmonary and disseminated forms of tuberculosis (TB) [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. In the USA, nearly 20% of the TB cases are extra-pulmonary [<xref ref-type="bibr" rid="B2">2</xref>]. In England and Wales, 38% of the TB cases are extrapulmonary [<xref ref-type="bibr" rid="B3">3</xref>]. Tuberculous pleuritis is a common manifestation of extrapulmonary TB [<xref ref-type="bibr" rid="B4">4</xref>]. TB is the most common cause of pleural effusion in many countries [<xref ref-type="bibr" rid="B4">4</xref>]. For example, studies from Spain [<xref ref-type="bibr" rid="B5">5</xref>], Malaysia [<xref ref-type="bibr" rid="B6">6</xref>], and Saudi Arabia [<xref ref-type="bibr" rid="B7">7</xref>] showed that TB accounted for 25%, 44%, and 37% of all effusions respectively. In the USA, the annual incidence of tuberculous pleuritis has been estimated to be about 1000 cases, and approximately one in 300 patients with TB will have tuberculous pleuritis [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B8">8</xref>]. The incidence of tuberculous effusions may be higher in patients with HIV infection [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Conventional diagnostic tests include microscopy of the pleural fluid, culture of pleural fluid, sputum pleural tissue, and pleural biopsy [<xref ref-type="bibr" rid="B4">4</xref>]. These tests have limitations. Microscopy of the pleural fluid is rarely positive (<5%) [<xref ref-type="bibr" rid="B10">10</xref>-<xref ref-type="bibr" rid="B12">12</xref>]. Culture of pleural fluid has low sensitivity (24% – 58%), and results are not available for weeks [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>]. Biopsy of pleural tissue, and culture of biopsy material are widely held to be the best methods of confirming the diagnosis [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B12">12</xref>]. This combination may lead to the diagnosis 86% of the time [<xref ref-type="bibr" rid="B10">10</xref>]. Although not perfect, culture and/or biopsy, therefore, are widely considered the standard of diagnosis [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B10">10</xref>]. However, pleural biopsy is invasive, operator-dependent, and technically difficult (particularly in children) [<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Because of the limitations of conventional tests, newer and rapid tests such as nucleic acid amplification tests – including polymerase chain reaction (PCR) – have been evaluated. Because of their high sensitivity and specificity in smear-positive respiratory specimens, these tests are now used – mainly in developed countries – for the direct detection of <italic>M. tuberculosis </italic>complex in respiratory specimens [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>]. NAA tests are categorized as commercial or in-house ("home-brew") tests. Commercial tests include the Amplified <italic>Mycobacterium tuberculosis </italic>Direct Test<sup>® </sup>(MTD) (Gen-Probe Inc, San Diego, CA), the Amplicor<sup>® </sup>MTB tests (Roche Molecular Systems, Branchburg, NJ), and the recently discontinued LCx<sup>® </sup>test (Abbott Laboratories, Abbott Park, IL). In the USA, the Amplicor test is licensed for use in smear-positive respiratory specimens; the MTD test is approved for smear-positive as well as smear-negative respiratory specimens [<xref ref-type="bibr" rid="B16">16</xref>]. No commercial test is licensed for use in non-respiratory specimens. We conducted this systematic review and meta-analysis to determine the overall accuracy of NAA tests in the diagnosis of tuberculous pleuritis, and to identify factors associated with heterogeneity of results between studies.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Search strategy</title><p>We searched the following databases: PubMed (1985 – January 2003), EMBASE (1988–2002), Web of Science (1990–2002), BIOSIS (1993–2002), Cochrane Library (2002; Issue 2), and LILACS (1990–2002). The <italic>Journal of Clinical Microbiology</italic>, a high-yield journal for TB diagnostic studies, was also hand-searched separately. All searches were up to date as of August 2002. The PubMed search was updated in January 2003. The search terms used were: tuberculosis, <italic>Mycobacterium tuberculosis</italic>, nucleic acid amplification techniques, polymerase chain reaction, sensitivity and specificity, and accuracy. Experts in the field were contacted. Bibliographies from the included studies and relevant review articles were screened. We obtained lists of citations from companies that manufacture commercial tests. Although no language restrictions were imposed initially, for the full-text review and final analysis our resources only permitted review of English and Spanish articles. Conference abstracts were excluded because of the limited data presented in them.</p></sec><sec><title>Study selection</title><p>Our search strategy was designed to include all published studies on NAA tests for the direct detection of <italic>M. tuberculosis </italic>in pleural fluid specimens. For inclusion, a study had to:</p><p>1. report a comparison of an NAA test against a reference standard, and provide data necessary for the computation of both sensitivity and specificity;</p><p>2. include at least 10 pleural fluid specimens (since studies with very few specimens are vulnerable to selection bias [<xref ref-type="bibr" rid="B17">17</xref>])</p><p>Studies on use of NAA tests on pleural biopsy and/or cytology specimens were excluded.</p><p>Two reviewers (MP and LLF) independently judged study eligibility while screening the citations. Disagreements were resolved by consensus. A list of excluded studies and a log of reasons for exclusion are available from the authors upon request.</p></sec><sec><title>Data extraction and quality assessment</title><p>Data extraction was performed by two reviewers. One reviewer (MP) extracted the data from all English studies. Another reviewer (LLF) extracted data from all Spanish articles. The second reviewer (LLF) also independently extracted data from a subset (36%) of the English articles, in order to determine the inter-rater agreement. The abstracted data included methodological quality, participant characteristics, test methods, and outcome data.</p><p>Quality assessment was performed using methods adapted from two guidelines on systematic reviews of diagnostic studies [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. For each study, the following quality criteria were scored as fulfilled or not: 1) Independent comparison of NAA test against reference standard; 2) Cross-sectional design (versus case-control design) ; 3) Blinded (single or double) interpretation of test and reference standard results; 4) Consecutive or random sampling of patients; 5) Prospective data collection; 6) Inclusion of at least 10 specimens/patients with confirmed tuberculous pleuritis. If no data on the above criteria were reported in the primary studies, we requested the information from the authors. For the purposes of analysis, responses coded as "not reported" were grouped together with "not met." A high quality study was arbitrarily defined as that which met at least 5/6 criteria; a medium quality met 3 or 4 of the 6 criteria; and a low quality study met less than 3/6 criteria. Since discrepant analysis (where discordant results between NAA test and reference test results are resolved, <italic>post-hoc</italic>, using clinical data) may be a potential source of bias, we preferentially included unresolved data.</p></sec><sec><title>Statistical analysis and data synthesis</title><p>We used standard methods recommended for meta-analyses of diagnostic test evaluations [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B19">19</xref>]. Analyses were performed using Meta-Test [<xref ref-type="bibr" rid="B20">20</xref>], and Stata version 8 (Stata Corporation, Texas). We computed the following measures of test accuracy for each study: sensitivity [true positive rate (TPR)], specificity [1-false positive rate (FPR)], positive likelihood ratio (LR+), negative likelihood ratio (LR-), and diagnostic odds ratio (DOR). These measures were pooled using the random effects model [<xref ref-type="bibr" rid="B17">17</xref>-<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Each study in the meta-analysis contributed a pair of numbers: TPR and FPR. Since TPR and FPR are not independent, we summarized their joint distribution by constructing a summary receiver operating characteristic (SROC) curve [<xref ref-type="bibr" rid="B21">21</xref>]. Unlike a traditional ROC plot used to explore the effect of varying thresholds (cut-points) on TPR and FPR in a single study, each data point in the SROC plot represents a separate study the meta-analysis. The SROC curve (and area under the curve) represents the overall performance of the test, and depicts the trade off between sensitivity and specificity. A symmetric curve suggests that the variability in accuracy between studies is explained, in part, by differences in thresholds employed by the studies.</p><p>Heterogeneity in meta-analyses refers to the degree of variability in results across studies. We used the Chi-square and Fisher's exact tests to detect statistically significant heterogeneity. Stratified (subgroup) analyses were used to identify study design and test-related factors responsible for heterogeneity in test accuracy. Studies using commercial tests were analyzed separately from those using in-house tests. Studies with commercial tests were further stratified by type of test (brand). Finally, since publication bias is of concern for meta-analyses of diagnostic studies [<xref ref-type="bibr" rid="B22">22</xref>], we tested for the potential presence of this bias using funnel plots and the Egger test [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec></sec><sec><title>Results</title><sec><title>Description of included studies</title><p>Figure <xref ref-type="fig" rid="F1">1</xref> outlines our study selection process. Thirty-eight articles [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B61">61</xref>] were included in the meta-analysis. Four articles were in Spanish [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B44">44</xref>]. Two articles reported evaluations of more than one NAA test against a common reference standard [<xref ref-type="bibr" rid="B30">30</xref>,<xref ref-type="bibr" rid="B57">57</xref>]. Each such test comparison was counted as a separate study. Thus, the total number of test comparisons (hereafter referred to as studies) was 40. Of these, 14 (35%) were studies of commercial tests, and 26 (65%) were of in-house tests. The average (median) sample size of each study in the meta-analysis was 60 pleural specimens or subjects, with a range of 15 to 375.</p></sec><sec><title>Study characteristics and quality</title><p>The mean inter-rater agreement between the two reviewers for items in the quality checklist was 0.86. Our initial data were affected by incomplete reporting in the primary studies. We contacted the authors via email, and obtained additional data for 25/40 (63%) included studies. Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref> present the descriptive data from studies with commercial and in-house tests, respectively. The tables present data on study quality, along with sample size, sensitivity and specificity estimates. In the commercial tests subgroup (Table <xref ref-type="table" rid="T1">1</xref>), 36%, 50%, and 14% of the studies were of high, medium and low quality, respectively. In the in-house tests subgroup (Table <xref ref-type="table" rid="T2">2</xref>), 39%, 42%, and 19% of the studies were of high, medium and low quality, respectively. Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref> show the variability in study quality and variability in NAA protocols employed. Although a variety of reference tests (or combinations of reference tests) were used, culture alone, or a combination of culture plus biopsy/microscopy/clinical data were used in 39 of the 40 studies.</p></sec><sec><title>Diagnostic accuracy of commercial tests</title><p>Figure <xref ref-type="fig" rid="F2">2(A)</xref> displays the sensitivity and specificity estimates from each of the 14 studies using commercial tests, stratified by type (brand) of test. Almost all studies had specificity estimates close to 1.0. In contrast, sensitivity estimates were lower and heterogeneous (range 0.20–1.0). Figure <xref ref-type="fig" rid="F3">3(A)</xref> shows the SROC curve for the commercial tests. The regression line does not trace out a typical ROC curve – the curve shows no trade-off between sensitivity and specificity. Table <xref ref-type="table" rid="T3">3</xref> presents the results of the meta-analysis. The summary measures of specificity and LR+ were very high and homogeneous. All other measures were highly heterogeneous.</p><p>Commercial tests use different amplification methods and target nucleic acid sequences: the Amplicor test employs PCR technology to amplify the 16s rRNA target; the LCx test employs ligase chain reaction to amplify the 38 kDa target; and the MTD test utilizes transcription-mediated amplification to amplify rRNA. To account for these differences, we further stratified commercial tests by type (brand) of test. Four of the studies evaluated the Amplicor test [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B54">54</xref>,<xref ref-type="bibr" rid="B55">55</xref>], four the LCx test [<xref ref-type="bibr" rid="B37">37</xref>,<xref ref-type="bibr" rid="B44">44</xref>,<xref ref-type="bibr" rid="B46">46</xref>,<xref ref-type="bibr" rid="B52">52</xref>], and six evaluated the MTD test [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B33">33</xref>-<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B61">61</xref>]. Stratification reduced the overall heterogeneity to some extent (Table <xref ref-type="table" rid="T3">3</xref> and Figure <xref ref-type="fig" rid="F2">2A</xref>). Although specificity did not vary by type of kit, sensitivity did – the Amplicor test had a lower sensitivity (0.37) than the LCx (0.72) and MTD (0.77) tests. Since 5/6 studies on the MTD test evaluated the first generation MTD-1 kit, there were insufficient numbers of MTD-2 studies to determine if the enhanced MTD-2 test (licensed for use in smear-negative respiratory specimens) performed better than the MTD-1 test.</p></sec><sec><title>Diagnostic accuracy of in-house tests</title><p>Figure <xref ref-type="fig" rid="F2">2(B)</xref> displays the sensitivity and specificity estimates from each of the 26 studies using in-house tests. Both sensitivity (range 0.20–1.0) and specificity (range 0.53–1.0) estimates were highly variable. All summary measures were grossly heterogeneous (Table <xref ref-type="table" rid="T3">3</xref>) and therefore would not be appropriately summarized. The SROC curve [Figure <xref ref-type="fig" rid="F3">3(B)</xref>] displays a ROC-type trade-off between sensitivity and specificity.</p><p>We performed stratified analyses to identify sources of heterogeneity among in-house tests. Table <xref ref-type="table" rid="T4">4</xref> presents two factors that appeared most strongly associated with the observed heterogeneity. Studies that employed a case-control design produced diagnostic odds ratio estimates nearly 2.4 times higher than studies that employed a cross-sectional design. Studies with PCR tests that used the IS6110 target sequence produced diagnostic odds ratio estimates 2.5 times higher than studies that used PCR tests with other targets. However, even after stratifying on study design and target sequence, considerable unexplained heterogeneity persisted in all the summary measures. The shape of the SROC curve suggested that variability in diagnostic thresholds (cut-points) across studies could partly explain the heterogeneity.</p></sec><sec><title>Publication bias</title><p>In the subgroup with commercial tests, the Egger test was not statistically significant (p = 0.55). However, in the in-house tests subgroup the Egger test was significant (p = 0.002), with an asymmetric funnel plot (figure <xref ref-type="fig" rid="F4">4</xref>) – evidence in favour of potential publication bias.</p></sec></sec><sec><title>Discussion</title><p>Since conventional tests are not always helpful in establishing a diagnosis of tuberculous pleuritis, several rapid tests and biomarkers have been evaluated: Adenosine Deaminase (ADA) [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B45">45</xref>,<xref ref-type="bibr" rid="B51">51</xref>,<xref ref-type="bibr" rid="B59">59</xref>,<xref ref-type="bibr" rid="B62">62</xref>], Interferon-γ (IFN-γ) [<xref ref-type="bibr" rid="B59">59</xref>,<xref ref-type="bibr" rid="B60">60</xref>,<xref ref-type="bibr" rid="B62">62</xref>,<xref ref-type="bibr" rid="B63">63</xref>], lysozyme [<xref ref-type="bibr" rid="B62">62</xref>], soluble interleukin 2 receptors [<xref ref-type="bibr" rid="B63">63</xref>], and NAA tests [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B61">61</xref>]. There has been an explosion of studies evaluating these rapid tests, and systematic reviews and meta-analyses are necessary to synthesize this growing body of literature. A recent meta-analysis summarized the evidence on ADA and IFN-γ for the diagnosis of tuberculous pleuritis [<xref ref-type="bibr" rid="B64">64</xref>]. Both ADA and IFN-γ tests were found to be reasonably accurate at detecting tuberculous pleuritis. Our meta-analysis summarizes the evidence on accuracy of NAA tests in the diagnosis of tuberculous pleuritis.</p><sec><title>Principal findings</title><p>The role of NAA tests has been reasonably well defined in pulmonary tuberculosis [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B65">65</xref>], and guidelines exist for testing of respiratory specimens [<xref ref-type="bibr" rid="B16">16</xref>]. In contrast, their role in the evaluation of specimens such as pleural fluid is not clear. Our results indicate that commercial NAA tests have high specificity and positive likelihood ratios. These test properties suggest a potential role for commercial tests in confirming (ruling in) the diagnosis of tuberculous pleuritis. These tests, however, have low and widely varying sensitivities – test properties that make them unhelpful in ruling out TB. Potential explanations for the low sensitivity include a low bacillary load in pleural fluid, or the presence of substances in the pleural fluid that inhibit amplification [<xref ref-type="bibr" rid="B65">65</xref>]. Some authors have suggested that pleural fluids should be tested with NAA methods after the specimens are adequately pre-treated to remove inhibitors [<xref ref-type="bibr" rid="B65">65</xref>]. All commercial kits appear to be designed to maximize only specificity. The MTD and LCx kits appear to have higher sensitivity than the Amplicor test. This comparison should be interpreted cautiously because it is based on few studies. Studies that directly compare these commercial tests (head-to-head) within the same study population are required to confirm these observations. The most important finding regarding in-house PCR is the significant heterogeneity across studies.</p></sec><sec><title>Clinical implications</title><p>To interpret the summary measures in a clinical context, consider a patient from a high incidence setting (e.g. countries such as Spain or Malaysia) who is estimated to have a 50% probability of pleural TB after clinical evaluation, and is evaluated with either the MTD test (LR+ of 17.4 and LR- of 0.31) or the Amplicor test (LR+ of 52.8 and LR- of 0.59). A LR+ of 17.4 for the MTD test suggests that patients with tuberculous pleuritis have a 17-fold higher chance of being MTD test positive as compared to patients without TB. If the MTD test were positive, the likelihood that this patient has TB increases from 50% to 95%, a probability that is sufficiently high to justify initiation of anti-tuberculosis treatment. A positive Amplicor test will raise the probability of TB from 50% to 97%. In contrast, if the MTD test result were negative, there is still a 24% chance that this patient has TB, probably not sufficiently low to rule out TB with confidence. In case of the Amplicor test, a negative test will reduce the probability from 50% to 40%, again not low enough to rule out TB.</p><p>Consider another patient from a low incidence setting (e.g. countries such as the USA), where the baseline probability of TB is low (e.g. 5%). If MTD test were positive, the likelihood that this patient has TB increases from 5% to 48%, a probability that justifies further investigation. A positive Amplicor test will raise the probability of TB from 5% to 75%. If the MTD<sup>® </sup>test result were negative, the baseline probability changes from 5% to 2%, a negligible shift that is unlikely to be helpful in clinical decision-making. In case of the Amplicor test, the probability changes from 5% to 4%. These examples illustrate the impact of the baseline prevalence (pre-test probability) on predictive values of the tests.</p><p>The accuracy of in-house PCR was heterogeneous across studies, and thus meaningful summary measures of accuracy could not be determined. The clinical implications, therefore, will depend on the setting. Institutions that use in-house PCR will have to rely on local data to decide on its accuracy and clinical applicability. In general, PCR for tuberculosis is known to have poor inter-laboratory reproducibility [<xref ref-type="bibr" rid="B66">66</xref>].</p><p>In addition to the effect of diagnostic thresholds seen in the SROC plot, we identified two factors that were associated with heterogeneity among in-house tests: use of a case-control design, and use of the IS6110 target sequence. Case-control studies sample patients from the extreme ends of the clinical spectrum (an ideal, "extreme contrast" setting). If the sensitivity of a test is evaluated in seriously diseased subjects, and specificity in healthy individuals, both measures will overestimate the true diagnostic accuracy [<xref ref-type="bibr" rid="B67">67</xref>]. Empiric research suggests that case-control studies overestimate the diagnostic odds ratio by a factor of 3, when compared to cross-sectional studies [<xref ref-type="bibr" rid="B68">68</xref>]. Future studies of NAA tests could reduce this bias by avoiding the case-control design and recruiting consecutive series of patients in whom the test is clinically indicated (a realistic, "clinical practice" setting). The IS6110 target sequence is widely used in <italic>M. tuberculosis </italic>fingerprinting [<xref ref-type="bibr" rid="B69">69</xref>]. Because this target is specific to the <italic>M. tuberculosis </italic>complex, and because it is usually present as multiple copies in the genome, PCR tests using this target might be more sensitive. Further research is underway to confirm this finding, in a larger meta-analysis of in-house PCR in the diagnosis of pulmonary tuberculosis.</p></sec><sec><title>Previous meta-analyses of NAA test accuracy</title><p>Our data are consistent with the results of two previous meta-analyses on the accuracy of NAA tests. Sarmiento and colleagues summarized the accuracy of PCR in the diagnosis of smear-negative pulmonary TB [<xref ref-type="bibr" rid="B70">70</xref>]. Their meta-analysis of 50 studies showed that both sensitivity and specificity estimates were heterogeneous. They concluded that PCR is not consistently accurate enough to be routinely recommended for the diagnosis of smear-negative TB. Our previous meta-analysis of 49 studies summarized the accuracy of NAA tests in the diagnosis of tuberculous meningitis [<xref ref-type="bibr" rid="B71">71</xref>]. Commercial tests were found to have high overall specificity (0.98) and low sensitivity (0.56). The accuracy of in-house PCR was not determined because of heterogeneity in study results.</p></sec><sec><title>Limitations of the review</title><p>Our review has limitations. Our analysis lacks data on the incremental gain of NAA tests over and above the diagnostic performance achieved by using only conventional methods or other rapid tests like ADA and IFN-γ. The primary studies in our review did not report such data. Also, few studies in our review directly compared the NAA test against tests such as ADA and IFN-γ[<xref ref-type="bibr" rid="B45">45</xref>,<xref ref-type="bibr" rid="B51">51</xref>,<xref ref-type="bibr" rid="B59">59</xref>]. Only one study [<xref ref-type="bibr" rid="B59">59</xref>] directly compared the three tests in the same population, and showed that ADA, IFN-γand PCR were 88%, 86%, and 74% sensitive respectively, and 86%, 97%, and 90% specific respectively, for culture or biopsy-confirmed pleural TB. Since we did not include tests such as ADA and IFN-γ in our literature searches, our review cannot identify the most accurate test. Also, publication bias was a concern with the in-house tests. Exclusion of studies published in languages other than English and Spanish could have contributed to this potential bias.</p></sec></sec><sec><title>Conclusions</title><p>In summary, our data suggest a potentially useful role for commercial NAA tests in confirming a diagnosis of tuberculous pleuritis. However, commercial kits have low and varying sensitivities, and therefore should not be used for excluding a diagnosis of tuberculous pleuritis. NAA test results, therefore, cannot replace conventional tests; they need to be interpreted in parallel with clinical findings and results of conventional tests. The accuracy of in-house PCR tests is poorly defined because of heterogeneity in study results. Clinically useful summary measures cannot be estimated for in-house PCR tests; their clinical applicability remains unclear.</p></sec><sec><title>List of abbreviations</title><p>ADA: Adenosine Deaminase</p><p>DOR: diagnostic odds ratio</p><p>FPR: false positive rate</p><p>HIV: human immunodeficiency virus</p><p>IFN-γ: Interferon-gamma</p><p>LR+: positive likelihood ratio</p><p>LR-: negative likelihood ratio</p><p>NAA: nucleic acid amplification</p><p>PCR: polymerase chain reaction</p><p>ROC: receiver operating characteristic curve</p><p>SROC: summary receiver operating characteristic curve</p><p>TB: tuberculosis</p><p>TPR: true positive rate</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>Study concept and design: MP, LLF, LWR, JMC.</p><p>Acquisition of data: MP, LLF.</p><p>Analysis and interpretation of data: MP, AH, LWR, JMC.</p><p>Drafting of manuscript: MP, JMC</p><p>Critical revision of the manuscript for important intellectual content: all authors</p><p>All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2334/4/6/prepub"/></p></sec>
Bronchospasm and its biophysical basis in airway smooth muscle	<p>Airways hyperresponsiveness is a cardinal feature of asthma but remains unexplained. In asthma, the airway smooth muscle cell is the key end-effector of bronchospasm and acute airway narrowing, but in just the past five years our understanding of the relationship of responsiveness to muscle biophysics has dramatically changed. It has become well established, for example, that muscle length is equilibrated dynamically rather than statically, and that non-classical features of muscle biophysics come to the forefront, including unanticipated interactions between the muscle and its time-varying load, as well as the ability of the muscle cell to adapt rapidly to changes in its dynamic microenvironment. These newly discovered phenomena have been described empirically, but a mechanistic basis to explain them is only beginning to emerge.</p>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Fredberg</surname><given-names>Jeffrey J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	Respiratory Research	<sec><title>Introduction</title><p>It is self-evident that acute narrowing of the asthmatic airway and shortening of the airway smooth muscle are inextricably linked. Nonetheless, it was many years ago that research on the asthmatic airway and research on the biophysics of airway smooth muscle (ASM) had a parting of the ways [<xref ref-type="bibr" rid="B1">1</xref>]. The study of smooth muscle biophysics took on a life of its own and pursued a deeply reductionist agenda, one that became focused to a large extent on myosin II and regulation of the acto-myosin cycling rate. The study of airway biology pursued a reductionist agenda as well, but one that became focused less and less on contractile functions of muscle and instead emphasized immune responses, inflammatory cells and mediators, and, to the extent that smooth muscle remained of interest, that interest centered mainly on synthetic, proliferative and migratory functions [<xref ref-type="bibr" rid="B2">2</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. Inflammatory remodeling of the airway wall was also recognized as being a key event in the asthmatic diathesis [<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B17">17</xref>]. Computational models of ever increasing sophistication were formulated in order to better understand the impact of inflammatory remodeling processes upon ASM shortening and acute airway narrowing but, remarkably, the muscle compartment of these models remained at a relatively primitive level, being represented by nothing more than the classical relationship of active isometric force <italic>vs. </italic>muscle length [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. As discussed below, this description is now considered to be problematic because the very existence of a well-defined static force-length relationship has of late been called into question, as has the classical notion that the muscle possesses a well-defined optimal length. Rather, other factors intrinsic to the muscle, especially muscle dynamics and mechanical plasticity, as well as unanticipated interactions between the muscle and its load, are now understood to be major factors affecting the ability of the muscle to narrow the airway [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B23">23</xref>-<xref ref-type="bibr" rid="B27">27</xref>].</p><p>The topics addressed in this review are intended to highlight recent discoveries that bring airway biology and smooth muscle biophysics into the same arena once again. The emphasis is biophysical properties of airway smooth muscle as they relate to excessive airway narrowing. This is appropriate because, in the end, if airway inflammation didn't cause airway narrowing, then asthma might be a tolerable disease (Julian Solway, personal communication). But asthma is not a tolerable disease. In order to understand the multifaceted problem of bronchospasm in asthma, therefore, an integrative understanding that brings together a diversity of factors will be essential.</p></sec><sec><title>Airway hyperresponsiveness</title><p>It was recognized quite early that the lung is an irritable organ and that stimulation of its contractile machinery in an animal with an open chest can cause an increase in lung recoil, air to be expelled, a rise in intratracheal pressure, and an increase in airways resistance [<xref ref-type="bibr" rid="B28">28</xref>-<xref ref-type="bibr" rid="B31">31</xref>]. However, until the second half of the last century airway smooth muscle was not regarded as being a tissue of any particular significance in respiration mechanics [<xref ref-type="bibr" rid="B28">28</xref>]. A notable exception in that regard was Salter [<xref ref-type="bibr" rid="B32">32</xref>], who, in 1859, was well aware of the existence of airway smooth muscle and its potential role in asthma. Airway smooth muscle was first described in 1804 by Reisseisen (as related by Otis [<xref ref-type="bibr" rid="B28">28</xref>]) and its functional properties first considered by Einthoven [<xref ref-type="bibr" rid="B33">33</xref>] and Dixon and Brodie [<xref ref-type="bibr" rid="B31">31</xref>]. More recent studies have shown that the fraction of the tissue volume that is attributable to contractile machinery is comparable for airways, alveolated ducts and blood vessels in the lung parenchyma [<xref ref-type="bibr" rid="B34">34</xref>]; the lung parenchyma, like the airway, is a contractile tissue [<xref ref-type="bibr" rid="B35">35</xref>-<xref ref-type="bibr" rid="B39">39</xref>]. Airway smooth muscle is now recognized as being the major end-effector of acute airway narrowing in asthma [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. There is widespread agreement that shortening of airway smooth muscle is the proximal cause of excessive airway narrowing during an asthmatic attack [<xref ref-type="bibr" rid="B17">17</xref>], with swelling of airway wall compartments and plugging by airway liquid or mucous being important amplifying factors [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B40">40</xref>]. It remains unclear, however, why in asthma the muscle can shorten excessively.</p><p>Airway hyperresponsiveness is the term used to describe airways that narrow too easily and too much in response to challenge with nonspecific contractile agonists [<xref ref-type="bibr" rid="B41">41</xref>]. Typically, a graph of airways resistance vs. dose is sigmoid in shape (Fig. <xref ref-type="fig" rid="F1">1</xref>); the response shows a plateau at high levels of contractile stimulus. Generally, the existence of the plateau is interpreted to mean that the airway smooth muscle is activated maximally and, therefore, has shortened as much as it can against a given elastic load. Once on the plateau, therefore, any further increase in stimulus can produce no additional active force, muscle shortening, or airway resistance.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>A computational result showing airway length (top) and airway resistance (bottom) as a function of agonist concentration for a tenth generation airway [<xref ref-type="bibr" rid="B151">151</xref>]. The cases shown depict airways from a normal, an asthmatic and a COPD lung. In this computation, the effects of tidal breathing and deep inspirations (6/minute) upon myosin binding dynamics are taken into account explicitly [<xref ref-type="bibr" rid="B151">151</xref>]. As explained in the text, such an airway exhibits both hyperreactivity and hypersensitivity.</p></caption><graphic xlink:href="1465-9921-5-2-1"/></fig><p>To say that airways narrow too easily, then, means that the graph of airways resistance vs. dose of a non-specific contractile stimulus is shifted to the left along the dose axis, and that airways respond appreciably to levels of stimulus at which the healthy individual would be unresponsive; this phenomenon is called hypersensitivity. By contrast, to say that the airways narrow too much means that the level of the plateau response is elevated, or that the plateau is abolished altogether, regardless of the position of the curve along the dose-axis; this phenomenon is called hyperreactivity. As distinct from hypersensitivity, it is this ability of the airways to narrow excessively, with an elevated or abolished plateau, that accounts for the morbidity and mortality associated with asthma [<xref ref-type="bibr" rid="B42">42</xref>].</p><p>It has long been thought that the factors that cause hypersensitivity <italic>vs. </italic>hyperreactivity are distinct, with the former being associated with receptor complement and downstream signaling events but the latter being associated with purely mechanical factors, including the contractile apparatus, the cytoskeleton, and the mechanical load against which the muscle shortens [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. Macklem has pointed out that once the muscle has become maximally activated it is the active force and the load that become all important, and the plateau response becomes essentially uncoupled from underlying biochemistry, signaling and cell biology [<xref ref-type="bibr" rid="B20">20</xref>-<xref ref-type="bibr" rid="B22">22</xref>]. As described below, there is reason to think that these distinctions may not be as clear as once believed, however.</p><p>Although asthma is usually defined as being an inflammatory disease, the link between the immunological phenotype and the resulting mechanical phenotype associated with disease presentation, including airways hyperresponsiveness, remains unclear; indeed, it is now established that airway hyperresponsiveness can be uncoupled from airway inflammation [<xref ref-type="bibr" rid="B44">44</xref>-<xref ref-type="bibr" rid="B47">47</xref>]. It remains equally unclear if airway hyperresponsiveness is due to fundamental changes within the smooth muscle itself, as might be caused by inflammatory mediators, chemokines and cytokines [<xref ref-type="bibr" rid="B48">48</xref>], or due to changes external to the muscle such as a reduced mechanical load against which the smooth muscle contracts. Still another possibility supported by recent evidence is that there is an interaction of the two wherein the contractile machinery within the smooth muscle cell adapts in response to a change in its mechanical microenvironment [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B50">50</xref>]. Moreover, Tschumperlin and colleagues [<xref ref-type="bibr" rid="B51">51</xref>,<xref ref-type="bibr" rid="B52">52</xref>] have provided evidence to suggest that bronchospasm can lead to mechanically-transduced pro-inflammatory signaling events in the airway epithelium, in which case inflammation may cause bronchospasm, but bronchospasm in turn may amplify or even cause inflammation.</p><p>In the balance of this review I address the classical picture of smooth muscle behavior and then go on to describe what we know about non-classical behavior in a dynamic mechanical environment driven by the tidal action of breathing. In addition, in recent years we have come to learn that the mechanical environment leads to interesting airway instabilities and adaptation in the muscle itself. Finally, I conclude by providing an emerging integrative context that seems to account for many of these properties that are not accounted for in classical perspectives of smooth muscle biophysics. I do not address the increasing evidence that now suggests that cytokines such as IL-1β and TNFα augment responses to bronchoconstrictor agonists while attenuating the bronchodilation that can be effected by hormones and paracrine agents like epinephrine and PGE<sub>2 </sub>[<xref ref-type="bibr" rid="B53">53</xref>]. Such cytokines, along with growth factors and other inflammatory mediators also result in smooth muscle hyperplasia, at least in culture systems [<xref ref-type="bibr" rid="B5">5</xref>]. In culture, extracellular matrix proteins also influence the contractile phenotype of airway smooth muscle cells [<xref ref-type="bibr" rid="B54">54</xref>,<xref ref-type="bibr" rid="B55">55</xref>]. Whether asthmatic inflammation can result in a hypercontractile phenotype remains to be established.</p></sec><sec><title>Classical behavior of airway smooth muscle and the balance of static forces</title><p>The microstructure of striated muscle is highly ordered whereas there is abundant evidence in the literature demonstrating that the cytoskeletal matrix of smooth muscle is quite disordered [<xref ref-type="bibr" rid="B56">56</xref>,<xref ref-type="bibr" rid="B57">57</xref>]; it is, after all, its amorphous structure that gives 'smooth' muscle its name. Moreover, the cytoskeletal matrix of airway smooth muscle is in a continuous state of remodeling, a point to which we return below. Despite these differences, it has been widely presumed that to a first approximation Huxley's sliding filament model of muscle contraction [<xref ref-type="bibr" rid="B58">58</xref>] describes the function of both smooth and striated muscle [<xref ref-type="bibr" rid="B59">59</xref>-<xref ref-type="bibr" rid="B61">61</xref>]. For many of the biophysical phenomena observed in airway smooth muscle, such as active force generation and shortening velocity, Huxley's model represents a useful tool for thought [<xref ref-type="bibr" rid="B61">61</xref>], while for others, like mechanical plasticity, it does not.</p><p>As in the case of striated muscle contraction, the principal biophysical parameters that characterize the case of smooth muscle contraction include the maximum active isometric force (or stress, which is simply the force carried per unit area), the length at which the muscle can attain that maximal force (i.e., the optimum length, Lo), and the shortening capacity of the muscle. The sliding filament model of Huxley is the starting point for understanding each of these phenomena. As described by Huxley [<xref ref-type="bibr" rid="B58">58</xref>], isometric force, as well as muscle stiffness, are proportional to the number of acto-myosin cross links per unit volume. This is true because, assuming rigid filaments, all bridges within a given contractile unit must act mechanically in parallel, with their displacements being identical and their forces being additive. The maximum active stress supported by smooth <italic>vs. </italic>striated muscle is approximately the same and is of the order 10<sup>5 </sup>Pascal. In striated muscle, the optimum length is attributed to the extent of overlap between the myosin filament and the actin filament, with optimum length corresponding to a maximum number of myosin heads finding themselves within striking distance of an available actin binding site, and the maximum capacity of the muscle to shorten being limited by the collision of the myosin filament end with the z-disc. Smooth muscle possesses no structure comparable to the z-disc, however, although actin filaments terminate in dense bodies, which might come into play in limiting muscle shortening. Whereas unloaded striated muscle can shorten perhaps 20% from its optimum length, unloaded smooth muscle can shorten as much as 70% [<xref ref-type="bibr" rid="B62">62</xref>-<xref ref-type="bibr" rid="B65">65</xref>]. Several physical factors may come into play to limit the capacity for unloaded shortening of smooth muscle. Small [<xref ref-type="bibr" rid="B56">56</xref>,<xref ref-type="bibr" rid="B57">57</xref>] has shown that the actin filaments of the contractile apparatus connect to the cytoskeleton at cytoplasmic dense bodies and with the longitudinal rib-like arrays of dense plaques of the membrane skeleton that couple to the extracellular matrix. Moreover, the side-polar configuration of the myosin filament [<xref ref-type="bibr" rid="B66">66</xref>,<xref ref-type="bibr" rid="B67">67</xref>] is likely to be involved. Still other factors coming into play include length-dependent activation [<xref ref-type="bibr" rid="B68">68</xref>,<xref ref-type="bibr" rid="B69">69</xref>], length-dependent rearrangements of the cytoskeleton and contractile machinery [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B70">70</xref>], and length-dependent internal loads [<xref ref-type="bibr" rid="B65">65</xref>,<xref ref-type="bibr" rid="B71">71</xref>,<xref ref-type="bibr" rid="B72">72</xref>].</p><p>What are the extramuscular factors that act to limit airway smooth muscle shortening? The basic notion, of course, is that muscle shortening stops when the total force generated by the muscle comes into a static balance with the load against which the muscle has shortened, both of which vary with muscle length. The factors setting the load include the elasticity of the airway wall, elastic tethering forces conferred by the surrounding lung parenchyma, active tethering forces conferred by contractile cells in the lung parenchyma [<xref ref-type="bibr" rid="B73">73</xref>,<xref ref-type="bibr" rid="B74">74</xref>], mechanical coupling of the airway to the parenchyma by the peribronchial adventitia, and buckling of the airway epithelium and submucosa [<xref ref-type="bibr" rid="B75">75</xref>-<xref ref-type="bibr" rid="B77">77</xref>]. In addition, the airway smooth muscle itself is a syncytium comprised mostly of smooth muscle cells, aligned roughly along the axis of muscle shortening, and held together by an intercellular connective tissue network. In order to conserve volume, as the muscle shortens it must also thicken. And as the muscle shortens and thickens, the intercellular connective tissue network must distort accordingly. Meiss has shown evidence to suggest that at the extremes of muscle shortening it may be the loads associated with radial expansion (relative to the axis of muscle shortening) of the intercellular connective tissue network that limits the ability of the muscle to shorten further [<xref ref-type="bibr" rid="B78">78</xref>].</p><p>In the healthy intact dog, airway smooth muscle possesses sufficient force generating capacity to close all airways [<xref ref-type="bibr" rid="B79">79</xref>,<xref ref-type="bibr" rid="B80">80</xref>]. This fact may at first seem to be unremarkable, but it is not easily reconciled with the observation that when healthy animals or humans are challenged with inhaled contractile agonists in concentrations thought to be sufficient to activate the muscle maximally, resulting airway narrowing is limited in extent, and that limit falls far short of airway closure [<xref ref-type="bibr" rid="B81">81</xref>,<xref ref-type="bibr" rid="B82">82</xref>]. Breathing remains unaccountably easy. Indeed, it is this lightness of breathing in the healthy challenged lung, rather than the labored breathing that is characteristic of the asthmatic lung, that in many ways presents the greater challenge to our understanding of the determinants of acute airway narrowing [<xref ref-type="bibr" rid="B83">83</xref>]. Brown and Mitzner [<xref ref-type="bibr" rid="B79">79</xref>] have suggested that the plateau of the dose-response curve reflects uneven or limited aerosol delivery to the airways. Another possibility, however, is that mechanisms exist that act to limit the extent of muscle shortening in the healthy breathing lung, whereas these mechanisms become compromised in the asthmatic lung. It has been suspected that the impairment of that salutary mechanism, if it could only be understood, might help to unlock some of the secrets surrounding excessive airway narrowing in asthma, as well as the morbidity and mortality associated with that disease [<xref ref-type="bibr" rid="B84">84</xref>-<xref ref-type="bibr" rid="B87">87</xref>]. This brings us to muscle dynamics, deep inspirations, and the potent effects of a time-varying muscle load associated with the act of breathing.</p></sec><sec><title>Shortening velocity and other manifestations of muscle dynamics</title><p>The oldest and certainly the simplest explanation of airway hyperreactivity would be that muscle from the asthmatic airway is stronger than muscle from the healthy airway, but evidence in support of that hypothesis remains equivocal [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B88">88</xref>]. Indeed, studies from the laboratory of Stephens and colleagues [<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B89">89</xref>-<xref ref-type="bibr" rid="B91">91</xref>] have emphasized that the force generation capacity of allergen-sensitized airway smooth muscle of the dog, or asthmatic muscle from the human, is no different from that of control muscle. As a result, the search for an explanation turned to other factors, and several alternative hypotheses have been advanced. These fall into three broad classes, each of which is consistent with remodeling events induced by the inflammatory microenvironment, and include an increase of muscle mass [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B77">77</xref>,<xref ref-type="bibr" rid="B92">92</xref>], a decrease of the static load against which the muscle shortens [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B75">75</xref>,<xref ref-type="bibr" rid="B77">77</xref>,<xref ref-type="bibr" rid="B92">92</xref>], and a decrease of the fluctuating load that perturbs myosin binding during breathing, described in greater detail below [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B93">93</xref>,<xref ref-type="bibr" rid="B94">94</xref>]. Aside from their effects on acto-myosin binding, changes in the static load and/or the dynamic load also lead to dramatic cytoskeletal remodeling events as the smooth muscle cell adapts to its microenvironment. Together, these hypotheses are attractive because they suggest a variety of mechanisms by which airway smooth muscle can shorten excessively even while the isometric force generating capacity of the muscle remains essentially unchanged.</p><sec><title>Setting the muscle length: a process equilibrated statically or dynamically?</title><p>Even if the force generating capacity is unchanged, a consistent association has been noted between airway hyperresponsiveness and unloaded shortening velocity of the muscle [<xref ref-type="bibr" rid="B89">89</xref>,<xref ref-type="bibr" rid="B90">90</xref>,<xref ref-type="bibr" rid="B95">95</xref>-<xref ref-type="bibr" rid="B97">97</xref>]. This association suggests the possibility that the problem with airway smooth muscle in asthma may be that it is too fast rather than too strong. But how shortening velocity – a dynamic property of the muscle – might cause excessive airway narrowing – a parameter that was thought to be determined by a balance of static forces – remains unclear. To account for increased shortening capacity of unloaded cells, Stephens and colleagues have reasoned that upon activation virtually all muscle shortening is completed within the first few seconds [<xref ref-type="bibr" rid="B97">97</xref>]. As such, the faster the muscle can shorten within this limited time window, the more it will shorten. However, in isotonic loading conditions at physiological levels of load, muscle shortening is indeed most rapid at the very beginning of the contraction, but appreciable shortening continues for at least 10 min after the onset of the contractile stimulus [<xref ref-type="bibr" rid="B24">24</xref>]. An alternative hypothesis to explain why intrinsically faster muscle might shorten more comes from consideration of the temporal fluctuations of the muscle load that are attributable to the action of spontaneous breathing [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B88">88</xref>,<xref ref-type="bibr" rid="B98">98</xref>]}. Load fluctuations that are attendant to spontaneous breathing are the most potent of all known bronchodilating agencies [<xref ref-type="bibr" rid="B99">99</xref>,<xref ref-type="bibr" rid="B100">100</xref>]. Among many possible effects, these load fluctuations perturb the binding of myosin to actin, causing the myosin head to detach from actin much sooner than it would have during an isometric contraction. But the faster the myosin cycling (i.e., the faster the muscle), the more difficult it is for imposed load fluctuations to perturb the actomyosin reaction. This is because the faster the intrinsic rate of cycling, the faster will a bridge, once becoming detached, reattach and contribute once again to active force and stiffness.</p><p>Why is muscle from the allergen sensitized animal or asthmatic subject faster? For technical reasons, in their study on biopsy specimens from asthmatic subjects Ma et al [<xref ref-type="bibr" rid="B97">97</xref>] did not measure expression of myosin light chain kinase, but their finding of increased content of message strongly implicates MLCK. Although regulation of myosin phosphorylation is a complex process with multiple kinase and phosphatase pathways, this finding substantially narrows the search for the culprit that may account for the mechanical changes observed in these cells. Also, these studies seem to rule out changes in the distribution of myosin heavy chain isoforms; content and isoform distributions of message from asthmatic cells showed the presence of smooth muscle myosin heavy chain A (SM-A) but not SM-B, the latter of which contains a seven-amino acid insert, is typical of phasic rather than tonic smooth muscle, and is by far the faster of the two isoforms [<xref ref-type="bibr" rid="B101">101</xref>,<xref ref-type="bibr" rid="B102">102</xref>]. Together, these findings confirm in muscle biopsy specimens from the asthmatic airway a number of findings from the allergen-sensitized dog model.</p></sec><sec><title>Cycling rate regulation</title><p>In smooth muscle, shortening velocity and its determinants are of particular interest [<xref ref-type="bibr" rid="B60">60</xref>,<xref ref-type="bibr" rid="B103">103</xref>]. Compared to striated muscle, the maximum unloaded shortening velocity of smooth muscle is smaller by more than an order of magnitude. This difference seems to be a adaptation to smooth muscle functions; whereas striated muscles typically produce motion or work in an efficient manner (i.e., converting chemical energy into mechanical energy with a small amount of energy lost to heat), smooth muscles are found in hollow organs where they serve to maintain tone or shape in an economical manner (i.e., doing so at a rate of chemical energy utilization that is smaller than that consumed striated muscle by about 300 fold [<xref ref-type="bibr" rid="B59">59</xref>,<xref ref-type="bibr" rid="B60">60</xref>]. Since the rate of myosin ATPase activity is tied directly to the myosin cyclic rate [<xref ref-type="bibr" rid="B104">104</xref>,<xref ref-type="bibr" rid="B105">105</xref>], a slower cycling rate implies economical maintenance of tone.</p><p>If smooth muscle is activated but held isometrically as the waiting time between the stimulus onset and a subsequent quick release is increased, the isometric force grows but the rate of ATP utilization and unloaded shortening velocity immediately after the release progressively decrease; this curious behavior represents a major distinction between smooth and striated muscle. Huxley's view of muscle contraction implies that the unloaded shortening velocity is set by the rate at which the myosin head can advance along the actin filament. Accordingly, for a given myosin step size, this means that shortening velocity is a direct measure of cross bridge cycling rate. Since the shortening velocity is found to decrease appreciably with increasing waiting time, the logical explanation following these ideas is that the cycling rate is a regulated variable and decreases as a function of time since the onset of the activation. Taken together, force maintenance, down-regulation of unloaded shortening velocity, and reduced rates of ATP utilization, comprise what is known as the latch state, or, equivalently, the latch phenomenon.</p></sec><sec><title>Latch</title><p>Force generated by any smooth muscle is sustained by cyclic interactions of myosin with actin. With onset of the contractile event, myosin-actin cycling begins and the number of interactions (bridges) increases and eventually approaches a steady state. It is widely agreed that during this process the rate of bridge cycling initially increases but then becomes substantially diminished. The mechanisms of cycling rate regulation remain very much an open question in the literature [<xref ref-type="bibr" rid="B103">103</xref>,<xref ref-type="bibr" rid="B106">106</xref>-<xref ref-type="bibr" rid="B113">113</xref>].</p><p>Among mechanisms of cycling rate regulation that have been proposed, the foremost is the latch hypothesis of Hai and Murphy, which has the attributes of being the simplest and capturing the central importance of phosphorylation of the 20 kDa myosin regulatory light chain [<xref ref-type="bibr" rid="B103">103</xref>,<xref ref-type="bibr" rid="B106">106</xref>,<xref ref-type="bibr" rid="B114">114</xref>-<xref ref-type="bibr" rid="B116">116</xref>]. Murphy and his colleagues suggested that the latch phenomenon arises as rapidly cycling cross bridges are replaced progressively by slowly cycling latch bridges if given enough time at a fixed muscle length, where the latch bridge is nothing more than a myosin head whose 20 kDa regulatory light chain becomes dephosphorylated while remaining attached to actin and maintaining both force and stiffness. The central notion is that the latch bridge has a very small rate of detachment from actin and, as a result, the latch bridge comprises an internal load against which rapidly cycling bridges must shorten [<xref ref-type="bibr" rid="B61">61</xref>].</p><p>Within the latch schema, the attainment of the isometric steady state implies that the population distribution of myosin molecules among their four possible states (attached vs. unattached to actin, phosphorylated vs. unphosphorylated regulatory light chains) have come to a binding equilibrium set by a balance of kinetic rate processes, many of which are ATP dependent. Once enough time has passed that this balance is attained and myosin has come to a binding equilibrium appropriate to isometric steady-state conditions, the muscle is then said to be in the latch state. Thus, the latch state corresponds to what I will refer to as a static equilibrium of myosin binding at the molecular level, and a balance of static forces at the mechanical level.</p><p>Whether the latch bridge might account for the latch phenomenon remains a point of some contention, however. The accessory proteins calponin and caldesmon are known to modulate the rate of muscle shortening, and have been suggested as being molecules responsible for or contributing to the latch phenomenon, but the mechanisms of action of these molecules are not well understood [<xref ref-type="bibr" rid="B108">108</xref>,<xref ref-type="bibr" rid="B117">117</xref>-<xref ref-type="bibr" rid="B119">119</xref>]. Finally, if contractile units were evanescent and the number of such units in series were to decrease progressively during a contractile event, as discussed below, it has been suggested that these adaptations might account for the latch phenomenon [<xref ref-type="bibr" rid="B25">25</xref>].</p></sec></sec><sec><title>Non-classical behavior: load fluctuations and dynamic equilibration of the muscle length</title><p>Load fluctuations are imposed continuously on airway smooth muscle and pulmonary vascular smooth muscle by the tidal action of breathing, and on muscular systemic arteries and arterioles by the pulsatile action of blood ejected from the heart. Smooth muscles in the urethra, urinary bladder, and gut are also subjected to periodic stretch. Accordingly, imposed fluctuations in muscle load seem to be a universal part of smooth muscle physiology.</p><p>It is well established that imposition of load fluctuations on smooth muscle inhibits development of active force and stiffness [<xref ref-type="bibr" rid="B80">80</xref>,<xref ref-type="bibr" rid="B98">98</xref>,<xref ref-type="bibr" rid="B120">120</xref>]. Although imposed load fluctuations induce important plastic changes in the cytoskeleton, as described later in this review [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B70">70</xref>], a major part of the force and stiffness inhibitions that are observed are attributable to direct effects of tidal stretch upon bridge dynamics [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B61">61</xref>,<xref ref-type="bibr" rid="B98">98</xref>].</p><sec><title>Perturbed equilibria of myosin binding</title><p>In the case of airway smooth muscle, the effects of oscillatory loading were first addressed by Sasaki and Hoppin and later by Gunst and colleagues, who demonstrated that imposition of tidal changes in muscle length depresses active force [<xref ref-type="bibr" rid="B98">98</xref>,<xref ref-type="bibr" rid="B120">120</xref>-<xref ref-type="bibr" rid="B123">123</xref>]. Subsequent studies [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B98">98</xref>] showed that imposed fluctuations of muscle length about a fixed mean length cause depression of muscle force and muscle stiffness (averaged over the stretch cycle); imposed length fluctuations also cause augmentation of the specific rate of ATP utilization and the hysteresivity (related to the muscle viscosity and an index of bridge cycling rate, as described below [<xref ref-type="bibr" rid="B105">105</xref>]. Thus, from both a mechanical and a metabolic point of view, the perturbed state is a hot or a melted state of the muscle [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B61">61</xref>,<xref ref-type="bibr" rid="B94">94</xref>]. Imposed force fluctuations about a fixed mean distending force systematically bias the airway smooth muscle toward lengthening; this phenomenon is called fluctuation-driven muscle lengthening [<xref ref-type="bibr" rid="B24">24</xref>]. Although tidal stretches smaller than 1% of muscle length produce only trivial mechanical effects, tidal stretches in the range of amplitudes expected during quiet tidal breathing (about 3% of muscle length) produce force inhibition that is equipotent with concentrations of isoproterenol in the range 10<sup>-7 </sup>to 10<sup>-5 </sup>M [<xref ref-type="bibr" rid="B99">99</xref>]. Molfino et al. showed in humans that brief cessation of tidal breathing causes the cross-sectional area of central airways to decrease by about half under the influence of baseline smooth muscle tone, and when tidal breathing is resumed the airway promptly dilates [<xref ref-type="bibr" rid="B124">124</xref>]. Taken together, these findings suggest that quiet tidal breathing is as effective in relaxing airway smooth muscle as is a potent relaxing agonist.</p><p>Importantly, there is increasing evidence that the potent bronchodilating response to periodic stretch and deep inspirations is impaired in asthma, and that this impairment may be the proximal cause of the loss of the plateau of the dose-response curve and the resulting morbidity of the disease [<xref ref-type="bibr" rid="B81">81</xref>,<xref ref-type="bibr" rid="B84">84</xref>,<xref ref-type="bibr" rid="B86">86</xref>,<xref ref-type="bibr" rid="B87">87</xref>,<xref ref-type="bibr" rid="B125">125</xref>,<xref ref-type="bibr" rid="B126">126</xref>]. There exists also a bronchoprotective effect of deep inspirations; deep inspirations prior to agonist challenge have been shown to blunt the subsequent contractile response [<xref ref-type="bibr" rid="B127">127</xref>-<xref ref-type="bibr" rid="B131">131</xref>]. It has been suggested that the bronchoprotective effect of deep inspirations are even more important than the bronchodilating response, and it too is profoundly impaired in asthma.</p><p>Lung inflations strain airway smooth muscle with each breath, and these periodic mechanical strains are transmitted to the myosin head and cause it to detach from the actin filament much sooner than it would have in isometric circumstances. This premature detachment profoundly reduces the duty cycle of myosin, typically by as much as 50–80% of its unperturbed isometric steady-state value, and depresses total numbers of bridges attached and active force to a similar extent. Of the full isometric force generating capacity of the muscle, therefore, only a small fraction ever comes to bear on the airway during tidal breathing, even when the muscle is activated maximally. Since bridge cycling is strongly perturbed by the imposed tidal stretches over time the number of myosin attachment and detachment events come into a dynamic balance. To distinguish it from the state that prevails in static loading conditions, this state has been called a perturbed equilibrium of myosin binding [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B61">61</xref>].</p><p>In pathological circumstances the tidal strains acting on myosin can become compromised, however. For example, in the chronically inflamed airway the peribronchial adventitia thickens [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B132">132</xref>]; this thickening decreases tidal muscle strains and thereby permits myosin binding to approach an unperturbed binding equilibrium. In doing so, the muscle would then generate the full complement of isometric force appropriate to the stimulus. Any factor that lessens peribronchial stress will decrease the force fluctuations impinging upon the muscle, including inflammatory thickening of the lamina reticulosa, thickening of the peribronchial adventitia, loss of lung elastic recoil, breathing at low lung volumes, and failure to take deep breaths [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B21">21</xref>,<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B75">75</xref>,<xref ref-type="bibr" rid="B133">133</xref>-<xref ref-type="bibr" rid="B135">135</xref>]; in addition, Colebatch and colleagues found evidence of increased rigidity of airways in asthmatic subjects [<xref ref-type="bibr" rid="B136">136</xref>]. If any of these factors pertain, then the perturbed equilibrium of myosin binding can collapse toward a static binding equilibrium. For example, if for any reason the muscle should stretch a bit less, then fewer bridges would be perturbed. Because more attached bridges working in parallel are harder to break than fewer, the muscle would then become stiffer still and therefore stretch even less, and so on. Ultimately this process reaches the limit in which the muscle may become so stiff that the physiological forces acting on it are insufficient to stretch the muscle appreciably, leaving the muscle stuck at its static equilibrium length [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B61">61</xref>]. Compared with the perturbed state, this statically equilibrated contractile state is also characterized by slow bridge cycling and a small rate of ATP utilization per bridge attached or per unit force developed [<xref ref-type="bibr" rid="B24">24</xref>]. So from both the mechanical and the metabolic point of view, it would appear to be a cold or 'frozen' contractile state. The muscle could be said to be statically equilibrated and frozen in the latch state.</p><p>This point of view leads to the hypothesis that airway hyperresponsiveness is associated with the failure of the underlying perturbed binding equilibrium to sustain itself, and an ensuing collapse of myosin binding kinetics to the binding equilibrium that pertains in static conditions and latch [<xref ref-type="bibr" rid="B23">23</xref>]. Clearly, this constellation of factors points towards dynamic instability, as described below.</p></sec><sec><title>Dynamic muscle instabilities</title><p>It is a well-established empirical fact that within any given lung or lung segment the response of airways to contractile agonists is always accompanied by 1) extreme heterogeneity of the response and 2) sensitivity of the response to the amplitude of the tidal volume and the magnitude of the load fluctuation [<xref ref-type="bibr" rid="B36">36</xref>,<xref ref-type="bibr" rid="B100">100</xref>,<xref ref-type="bibr" rid="B137">137</xref>-<xref ref-type="bibr" rid="B142">142</xref>]. It has been speculated that the observation of a smoothly graded decline of tests of lung functions (such as lung resistance) with progressively increasing doses of contractile agonist might be better explained by progressive changes in number of open airways accommodating gas flow rather than by some smoothly changing reduction in the airway caliber of each [<xref ref-type="bibr" rid="B138">138</xref>]. Indeed, the heterogeneity of the response is so extensive that peripheral airways have sometimes been thought of as being partitioned into only two quantum-like states, as it were, either open wide or almost closed entirely, with virtually no intermediate state [<xref ref-type="bibr" rid="B138">138</xref>,<xref ref-type="bibr" rid="B143">143</xref>,<xref ref-type="bibr" rid="B144">144</xref>]. Several plausible factors have been invoked to try to account for this profound heterogeneity of the contractile response, including intrinsic inhomogeneities in airway structure, muscle amounts, muscle sensitivity to agonist, and nonuniformity of agonist delivery [<xref ref-type="bibr" rid="B36">36</xref>,<xref ref-type="bibr" rid="B79">79</xref>,<xref ref-type="bibr" rid="B145">145</xref>].</p><p>Anafi and Wilson considered the narrowing of an airway containing activated airway smooth muscle subject to load fluctuations as would occur during breathing [<xref ref-type="bibr" rid="B146">146</xref>,<xref ref-type="bibr" rid="B147">147</xref>]. Their mathematical analysis shows that in some circumstances such an airway must become unstable. Of course, the generic idea that airways can be unstable is not at all new [<xref ref-type="bibr" rid="B148">148</xref>], but the analysis of Anafi and Wilson identified a new class of airway instability that is intrinsically dynamic and, in particular, is rooted in the response of airway smooth muscle to imposed load fluctuations.</p><p>To appreciate the potential importance of this muscle instability, we imagine a lung in which there are a large number of airways that are all operating in parallel and that are in every regard identical. The airway smooth muscle is activated uniformly and subjected to identical load fluctuations caused by the tidal action of breathing. Even though inter-regional differences are infinitesimally small at the outset, any small perturbation would be amplified by the processes described above. And as the differences grow the amplification grows. When the process eventually becomes dynamically equilibrated, the airways will be seen to have partitioned themselves between only two states in order to accommodate the total flow-one state effectively wide open and the other nearly closed. The analysis of Anafi and Wilson and that of Fredberg and colleagues [<xref ref-type="bibr" rid="B24">24</xref>] address overlapping but different aspects of airway narrowing; although the relationship between these perspectives remains somewhat unclear, airway smooth muscle in the closed airways might be expected to correspond to the static or 'frozen' state, and muscle in the open airways might be expected to correspond to a 'melted' state with an underlying perturbed equilibrium of myosin binding. Changes in tidal volume would affect the number of airways in each state.</p><p>In the healthy lung during spontaneous breathing, tidal volume and associated force fluctuations acting on airway smooth muscle may be large enough to keep almost all units in the open/melted state and out of jeopardy of closure. But if tidal volumes were compromised, or deep inspirations were prohibited, or transmission of force fluctuations to the airway smooth muscle were to become compromised, then a larger fraction of units would be expected to be found in the closed/frozen state. Therefore, this picture would seem to be able to explain in one stroke both a profound heterogeneity of the airway response and its sensitivity to tidal volume.</p><p>This insight on instability of smooth muscle dynamics does not preclude important contributions from intrinsic heterogeneities that exist amongst airways, of course. The Anafi-Wilson instability is attractive, though, because it shows that it is not necessary to postulate intrinsic airway heterogeneity of any kind in order to account for both a profound heterogeneity of the airway response and its sensitivity to tidal volume.</p></sec><sec><title>Deep Inspirations</title><p>Jensen and colleagues [<xref ref-type="bibr" rid="B149">149</xref>] followed the earlier studies of Fish [<xref ref-type="bibr" rid="B84">84</xref>], Lim et al. [<xref ref-type="bibr" rid="B86">86</xref>], and then Skloot [<xref ref-type="bibr" rid="B87">87</xref>] to show in normal individuals challenged with nonspecific contractile agonists that a single deep inspiration (DI) causes a prompt and dramatic decrease in airway resistance (Raw) followed by a slow return of Raw to the level observed prior to the DI. By contrast, in individuals with severe asthma a single DI causes a prompt but only modest decrease in Raw followed by rapid return of Raw to the level observed prior to the DI. The findings in challenged normal individuals are consistent with disruption of cross bridges as described above. In the asthmatic, however, it is plausible that bridges were stretched, but not enough to disrupt an appreciable fraction of the acto-myosin bonds. Immediately following the deep inspiration, the rapid recovery of Raw to pre-DI levels could reflect either a rapid elastic recoil of the stretched muscle, or an altered cytoskeletal remodeling response to the DI (described below), or reattachment of bridges that were disrupted, but at a rate that is faster than that observed in healthy individuals.</p><p>The crucial role of tidal loading of the ASM was further demonstrated by Mitzner and Brown in an ingenious use of high resolution CT in a dog model in which one lung was ventilated while the other was held isovolumic [<xref ref-type="bibr" rid="B150">150</xref>]. Compared to the lung undergoing cyclic volume changes, the isovolumic hung was found to be both hyperresponsive and refractory to the bronchodilating effects of a subsequent deep inspiration.</p><p>In order to better understand these processes and, especially, the impact of inflammatory remodeling processes upon ASM shortening and acute airway narrowing, Mijailovich and colleagues [<xref ref-type="bibr" rid="B151">151</xref>] developed a computational model which, unlike previous attempts [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>], took into account explicitly molecular dynamics of myosin and the ability of load fluctuations associated with tidal breathing and periodic deep inspirations to perturb myosin binding. Results for a generation 10 airway reveal hyperreactive airways in asthma and COPD, as was found by others using static muscle models (Fig. <xref ref-type="fig" rid="F1">1</xref>). However, this model of molecular dynamics expresses two additional phenomena that the static model does not. The first is induction of hypersensitivity – a shift of the dose-response curve to the left – which arises from myosin binding dynamics and their interaction with a fluctuating load. As noted above, mechanisms responsible for hypersensitivity <italic>vs. </italic>hyperreactivity had been though to be distinct whereas, by contrast, this result shows that myosin binding dynamics in the setting of a fluctuating mechanical load can elicit both. Compared to the case of the asthmatic airway, in the case of the normal airway the action of tidal loading upon myosin binding dynamics is sufficient to inhibit airway narrowing until the level of muscle activation is relatively large. The second phenomenon is what would appear to be a discontinuous or unstable response: with an increasing degree of muscle activation (agonist dose) there arises a value beyond which, when the airway does begin to narrow, it does so precipitously.</p><p>Taken together, these non-classical behaviors support the notion that activated airway smooth muscle is typically melted by the tidal action of breathing and, therefore, muscle length is dynamically equilibrated. Nonetheless, in pathological circumstances, as described below, muscle can become frozen in the latch state.</p></sec><sec><title>Clinical manifestations</title><p>In practical terms, specific instances in which elevated rates of bridge cycling might come into play include the differences in airway responsiveness that have been observed between normal versus allergen sensitized muscle, between certain animal strains and, in some species, between mature versus immature animals [<xref ref-type="bibr" rid="B68">68</xref>,<xref ref-type="bibr" rid="B142">142</xref>,<xref ref-type="bibr" rid="B152">152</xref>-<xref ref-type="bibr" rid="B156">156</xref>]. This rationale leads us to a plausible mechanism by which the rate of bridge cycling and its regulation may be reasonably thought to bear upon the prevalence of childhood asthma and its changes with lung maturation and allergic status.</p><p>Although far from explaining these phenomena in their entirety, the perturbed equilibrium hypothesis may help to tie together still other loose ends (Fig. <xref ref-type="fig" rid="F2">2</xref>). According to that hypothesis, greater contractile responses are favored whenever the force fluctuations acting on the airway become compromised, not only when the peribronchial adventitia undergoes cytokine-driven inflammatory thickening, but also when the lung loses elastic recoil, or tidal lung expansion becomes diminished. These instances bring immediately to mind not only asthma, but also emphysema, normative aging, restrictive disorders of the chest wall, obesity, and cervical spinal cord injury, each of which is known to be associated with a predisposition for airway hyperresponsiveness or asthma [<xref ref-type="bibr" rid="B157">157</xref>-<xref ref-type="bibr" rid="B163">163</xref>]. Reduced amplitude of force fluctuations may come into play similarly in asthma exacerbations that occur during sleep (nocturnal asthma) [<xref ref-type="bibr" rid="B164">164</xref>,<xref ref-type="bibr" rid="B165">165</xref>] and during late pregnancy [<xref ref-type="bibr" rid="B166">166</xref>,<xref ref-type="bibr" rid="B167">167</xref>] where, in both circumstances, functional residual capacities are diminished [<xref ref-type="bibr" rid="B168">168</xref>,<xref ref-type="bibr" rid="B169">169</xref>], suggesting in turn diminished lung recoil and smaller load fluctuations acting upon the airway smooth muscle [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B94">94</xref>]. Moreover, it is clear that when inflammatory remodeling of the airway does occur, the perturbed equilibrium hypothesis predicts that the resulting predisposition for airway hyperresponsiveness might persist long after the inflammation itself is resolved [<xref ref-type="bibr" rid="B46">46</xref>,<xref ref-type="bibr" rid="B170">170</xref>]. And finally, perturbed equilibria might also help to explain why the obstructive response in exercise-induced asthma typically begins only after cessation of the exercise, when tidal volumes have declined to resting levels.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>The perturbed equilibrium hypothesis connects phenotypes that were largely unexplained and had been thought to be essentially unrelated. Reduced force fluctuations and/or increased bridge cycling rates allow airway smooth muscle to more nearly approach a static equilibrium of myosin binding (latch) and the frozen state. Airway hyperresponsiveness phenotypes shown in blue correspond to circumstances in which the airway and airway smooth muscle might be normal but there is a problem with the respiratory pump, i.e., the muscles of the chest wall. Those shown purples correspond to phenotypes in which the airway smooth muscle may be normal, but there is a problem in the mechanical coupling between the respiratory pump and the myosin motor. Finally, those shown in green correspond to phenotypes in which the problem may be at the level of the myosin motor itself. At that level, the rate of bridge cycling is thought to be influenced by its isoform, the amount of myosin light chain kinase, caldesmon, calponin, Rho-kinase and other factors.</p></caption><graphic xlink:href="1465-9921-5-2-2"/></fig></sec></sec><sec><title>Mechanical plasticity: another non-classical feature of airway smooth muscle</title><p>When activated muscle in the muscle bath is subjected to progressively increasing load fluctuations approaching the magnitude and frequency expected during normal breathing, the muscle lengthens appreciably in response [<xref ref-type="bibr" rid="B24">24</xref>]. But when load fluctuations are progressively reduced, the muscle reshortens somewhat but fails to return to its original length. Incomplete reshortening after exposure to tidal loading is not accounted for by muscle injury; the original operating length can be recovered simply by removing the contractile agonist and allowing the muscle a short interval before recontracting. Neither can incomplete reshortening be accounted for by myosin dynamics; myosin dynamics by themselves predict complete reshortening when the load fluctuations are removed [<xref ref-type="bibr" rid="B24">24</xref>]. Thus, the failure of activated muscle to reshorten completely is evidence of a plasticity of the contractile response. During a sustained contraction, the operational length of the muscle for a given loading, or the force at a given length, can be reset by loading and the history of that loading [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B49">49</xref>,<xref ref-type="bibr" rid="B70">70</xref>,<xref ref-type="bibr" rid="B171">171</xref>-<xref ref-type="bibr" rid="B174">174</xref>]. In healthy individuals this plasticity seems to work in a favorable direction, allowing activated muscle to be reset to a longer length. The asthmatic, it has been argued, never manages to melt the contractile domain in the airway smooth muscle and, as such, the benefits of this plastic response are not attained.</p><p>It is now firmly established that airway smooth muscle can somehow adapt its contractile machinery, as well as the cytoskeletal scaffolding on which that machinery operates, in such a way that the muscle can maintain the same high force over an extraordinary range of muscle length [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B70">70</xref>,<xref ref-type="bibr" rid="B173">173</xref>-<xref ref-type="bibr" rid="B178">178</xref>]; ASM is characterized by its ability to disassemble its contractile apparatus when an appropriate stimulus is given, and its ability to reassemble that apparatus when accommodated at a fixed length. When exposed to contractile agonists, airway smooth muscle cells in culture reorganize cytoskeletal polymers, especially actin [<xref ref-type="bibr" rid="B179">179</xref>], and become stiffer [<xref ref-type="bibr" rid="B180">180</xref>]. Although cell stiffening is attributable largely to activation of the contractile machinery, an intact actin lattice has been shown to be necessary but not sufficient to account for the stiffening response [<xref ref-type="bibr" rid="B180">180</xref>].</p><p>Malleability of the cell and its mechanical consequences have been called by various authors mechanical plasticity, remodeling, accommodation or adaptation. Even though the force generating capacity varies little with length in the fully adapted muscle, the unloaded shortening velocity and the muscle compliance vary with muscle length in such a way as to suggest that the muscle cell adapts by adding or subtracting contractile units that are mechanically in series. The mechanisms by which these changes come about and the factors that control the rate of plastic adaptation are unknown, however.</p><p>Several hypotheses have been advanced to explain smooth muscle plasticity. Ford and colleagues have suggested that the architecture of the myosin fibers themselves may change [<xref ref-type="bibr" rid="B27">27</xref>], while Gunst and colleagues have argued that it is the connection of the actin filament to the focal adhesion plaque at the cell boundary that is influenced by loading history [<xref ref-type="bibr" rid="B70">70</xref>,<xref ref-type="bibr" rid="B171">171</xref>,<xref ref-type="bibr" rid="B172">172</xref>]. Alternatively, another notion is that secondary but important molecules stabilize the cytoskeleton, and as the contractile domain melts under the influence of imposed load fluctuations, those loads must be borne increasingly by the scaffolding itself, and thus reflects malleability of the cytoskeletal domain [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B54">54</xref>,<xref ref-type="bibr" rid="B70">70</xref>,<xref ref-type="bibr" rid="B181">181</xref>]. In that connection a role for the Rho-A pathway has been suggested [<xref ref-type="bibr" rid="B54">54</xref>,<xref ref-type="bibr" rid="B182">182</xref>] and some evidence now suggests that the p38 MAP kinase pathway may be involved [<xref ref-type="bibr" rid="B50">50</xref>]. Airway smooth muscle incubated with an inhibitor of the p38 MAP kinase pathway demonstrates a greater degree of fluctuation-driven muscle lengthening than does control muscle, and upon removal of the force fluctuations it remains at a greater length. Moreover, force fluctuations themselves activate the p38 MAP kinase pathway. It is noteworthy in that connection that heat shock protein 27 has been implicated as an essential element in the motility of airway smooth muscle cells and is a downstream target of rho and p38 [<xref ref-type="bibr" rid="B118">118</xref>,<xref ref-type="bibr" rid="B183">183</xref>-<xref ref-type="bibr" rid="B187">187</xref>]. These findings are consistent with the hypothesis that stress response pathways may somehow stabilize the airway smooth muscle cytoskeleton and limit the bronchodilating effects of deep inspirations.</p><p>Regardless of the specific molecules and mechanisms invoked to explain the plasticity of the contractile response, the melting of the contractile domain would appear to be a necessary (or permissive) event, but one that by itself is not sufficient to explain the effects of the history of tidal loading. This brings us, finally, to the notion of the cytoskeleton acting as a glassy material.</p><sec><title>An emerging question: are we built of glass?</title><p>The abilities of the cytoskeleton (CSK) to deform, to flow, and to remodel (i.e., mechanical plasticity) come into play in a wide variety of situations, from cell division, crawling and extravasation to invasion, contraction and wound healing.</p><p>Deformation, flow and remodeling of the airway smooth muscle cell can be described at the molecular level by specific modes of molecular motion and interactions between specific molecular species, at least in principle. The straight-forward approach would be to use direct numerical simulation to describe these molecular interactions within an integrated cytoskeletal lattice and then go on to compute the macroscale integrative properties that result. Direct numerical simulation faces three daunting problems, however. The first is shear complexity; the number of cytoskeletal molecular species is counted in the scores. Moreover, integrated multi-molecular assemblies in airway smooth muscle comprise a messy microstructural geometry, one characterized by a degree of long range order that is far less than that observed in ordinary solids but far greater than that found in fluids. The second is that the list of species remains incomplete and the nature of most protein-protein interactions has yet to be characterized biophysically, with the acto-myosin interaction being the exception that proves the rule. And third, there is good evidence to suggest that in several regards these systems exist far away from local thermodynamic equilibrium (LTE); elemental components do not enjoy a fixed spatial address and, instead, are closely packed and continuously jostling one another in a never-ending search for a minimum energy configuration but never managing to find one. Accordingly, these configurations are adaptable, being in a continuous state of remodeling and, in the process, consuming energy on an ongoing basis via the hydrolysis of ATP. If the first two problems are not enough to saturate the most powerful present-day computers, the non-LTE nature of the problem is particularly thorny because it adds a crucial dimension to the biophysics at the same time that it invalidates a major class of computational approaches, namely, those based on principles of energy minimization.</p><p>All three problems are addressed by the approach of Fabry et al. [[<xref ref-type="bibr" rid="B187">187</xref>,<xref ref-type="bibr" rid="B189">189</xref>]. Using airway smooth muscle cells in a culture system, Fabry discovered that integrative statistical properties of the cytoskeleton, such as measures of its ability to deform, to flow and to remodel, conform at the macroscale to a universal empirical framework, namely, that of glassy systems. Findings to date suggest that these processes depend mainly on a putative energy level in the cytoskeletal lattice, where that energy is representative of the amount of molecular agitation, or jostling, present in the lattice relative to the depth of energy wells that constrain molecular motions. This energy level can be expressed as an effective lattice temperature – as distinct from the familiar thermodynamic temperature. Even while the thermodynamic temperature is held fixed this effective temperature can change, can be manipulated, and can be measured. The higher the effective temperature, the more frequently do elemental structures trapped in one energy well manage to hop out of that well only to fall into another. The hop, therefore, can be thought of as the fundamental molecular remodeling event.</p><p>Among the many consequences of these findings, one can easily show that the rate of cytoskeletal remodeling (plastic adaptation) must scale as this effective temperature; when the matrix is "hot", as it is early an a contractile event for example, the rate of remodeling can be fast. But later in the contractile event, as the effective temperature falls and the matrix "cools", the rate of remodeling slows [<xref ref-type="bibr" rid="B188">188</xref>-<xref ref-type="bibr" rid="B190">190</xref>]. If the effective temperature falls enough that remodeling virtually comes to a standstill, the matrix behaves as if it were frozen. This latter state is consistent with and subsumes the latch state, as described above, while load fluctuations driven by the action of breathing impinge on the cell and represent another source of agitation whose action is consistent with an elevated effective temperature [[<xref ref-type="bibr" rid="B187">187</xref>,<xref ref-type="bibr" rid="B189">189</xref>]: Gunst, 2003 #1051]. Taken together, these features describe a soft glass, and the effective temperature at which all remodeling ceases is called the glass transition temperature. Data available to date conform in many ways to predictions based upon the idea of traps and hops, and conform as well to closely similar behavior that is shared by other soft materials found in nature including foams, pastes, slurries, colloids and some clays. Taken together, these are referred to in the literature as the class of soft glassy materials.</p></sec></sec><sec><title>Future directions</title><p>It has become well established that muscle length is equilibrated dynamically rather than statically, and that in a dynamic micro-environment non-classical features of muscle biophysics come to the forefront, including unanticipated interactions between the muscle and its time-varying load, as well as the ability of the muscle cell to adapt rapidly to changes in its mechanical microenvironment. Evidence supporting the notion of a highly malleable cell is accumulating rapidly, but a mechanistic basis to explain this malleability is only beginning to emerge. The notion that airway smooth muscle malleability and remodeling might be explained as a reflection of glassy behavior in the neighborhood of a glass transition is attractive because it is rather simple yet seems to tie together diverse behaviors within a unified empirical framework [<xref ref-type="bibr" rid="B188">188</xref>-<xref ref-type="bibr" rid="B190">190</xref>]. But whether this line of thinking will bear fruit remains to be seen.</p></sec>
Predicting Cancer Patient Survival with Gene Expression Data	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>Cancer specialists often talk about cancer as an umbrella term for over 200 different diseases, each having unique characteristics. But even these categories are too broad, as the same type of cancer can take very different paths in different people. It's not uncommon, for example, for a tumor to grow aggressively in one patient and stabilize or regress in another, even though their tumors are indistinguishable and are treated in the same way. Researchers have traditionally diagnosed and treated cancer based on microscopic analysis of cell size and shape, a method that's especially difficult for very closely related cancers, such as non-Hodgkin's lymphoma, which has 20 subtypes. As scientists learn more about the molecular alterations in cancer, they're beginning to establish cancer subtypes based on the underlying molecular footprint of a tumor. Four years ago, DNA microarray analysis revealed that the most common subtype of non-Hodgkin's lymphoma is in fact two separate diseases. Though the tumor cells of both cancers appear large and diffusely dispersed in a tissue sample under a microscope, each has a distinct genetic profile, possibly explaining why only 40% of patients with this subtype respond to the standard chemotherapy treatment.<xref ref-type="fig" rid="pbio-0020118-g001"/> </p><fig id="pbio-0020118-g001" position="float"><caption><title>Selecting expression profiles that can predict cancer outcome</title></caption><graphic xlink:href="pbio.0020118.g001"/></fig><p>Such molecular pathology has led to the discovery of subtypes of several different tumor types and has successfully identified patients with different survival times. But such correlations work best when cancer subtypes based on genetic profiles are already known. If you know that different subtypes exist and which patients belong to which subtype, then you can build a statistical model to diagnose such cancers in future patients. But in most situations, clinicians don't know either of these variables—or even whether such a subtype exists—information that is crucial to developing effective diagnostic and treatment protocols. Statistical methods to identify such subtypes exist, but they can generate classifications that lack clinical relevance. Now Eric Bair and Robert Tibshirani describe a procedure that combines both gene expression data and the patients' clinical history to identify biologically significant cancer subtypes and show that this method is a powerful predictor of patient survival.</p><p>Their approach uses clinical data to identify a list of genes that correspond to a particular clinical factor—such as survival time, tumor stage, or metastasis—in tandem with statistical analysis to look for additional patterns in the data to identify clinically relevant subsets of genes. In many retrospective studies, patient survival time is known, even though tumor subtypes are not; Bair and Tibshirani used that survival data to guide their analysis of the microarray data. They calculated the correlation of each gene in the microarray data with patient survival to generate a list of “significant” genes and then used these genes to identify tumor subtypes. Creating a list of candidate genes based on clinical data, the authors explain, reduces the chances of including genes unrelated to survival, increasing the probability of identifying gene clusters with clinical and thus predictive significance. Such “indicator gene lists” could identify subgroups of patients with similar gene expression profiles. The lists of subgroups, based on gene expression profiles and clinical outcomes of previous patients, could be used to assign future patients to the appropriate subgroup.</p><p>An important goal of microarray research is to identify genetic profiles that can predict the risk of tumor metastasis. Being able to distinguish the subtle differences in cancer subtype will help doctors assess a patient's risk profile and to prescribe a course of treatment tailored to that profile. A patient with a particularly aggressive tumor, for example, would be a candidate for aggressive treatment, while a patient whose cancer seems unlikely to metastasize could be spared the debilitating side effects of aggressive anticancer therapies. By providing a method to cull the thousands of genes generated by a microarray to those most likely to have clinical relevance, Bair and Tibshirani have created a powerful tool to identify new cancer subtypes, predict expected patient survival, and, in some cases, help suggest the most appropriate course of treatment.</p>
Emergence of a Peaceful Culture in Wild Baboons	Could not extract abstract	Could not extract contributor	PLoS Biology	<p>For most animal species, behavioral attributes are largely the product of interactions between genes and environment, with behavioral patterns preserved by natural selection. Birds, for example, know instinctively what type of nest to build for their offspring; salamanders don't need lessons to swim. But when it comes to primates—including humans—a good deal of behavior is learned. Primates exhibit a wide range of behaviors, not just among species but also among populations and even individuals. Yet the nature versus nurture debate still rages, particularly when it comes to understanding the roots of aggression. While bonobos are famous for using sex to resolve disputes, aggression is far more common in most primate species—again humans included. Our closest relative, the chimpanzee, has a reputation for being among the most belligerent, with rhesus monkeys and baboons not far behind. For many of these species, bouts of violence are often followed by gestures of reconciliation, such as grooming or, in the case of chimps, kissing. Since most primates live in social groups, it may be that such conciliatory measures serve to maintain some semblance of social structure, offsetting the disruptive effects of aggression. (To learn more about primate behavior and aggression, see the primer by Frans de Waal in this issue [DOI: <ext-link ext-link-type="doi" xlink:href="10.1371/journal.pbio.0020101">10.1371/journal.pbio.0020101</ext-link>].)<xref ref-type="fig" rid="pbio-0020124-g001"/> </p><fig id="pbio-0020124-g001" position="float"><caption><title>In baboons, “grooming” is a socially rewarding behavior. (Photograph, with permission, by Robert Sapolsky)</title></caption><graphic xlink:href="pbio.0020124.g001"/></fig><p>Primatologists characterize these behavioral differences as “cultural” traits, since they arise independent of genetic or environmental factors and are not only shared by a population (though not necessarily a species) but are also passed on to succeeding generations. Such cultural traditions have been documented in African chimp populations, which display over 39 behaviors related to “technology” (such as using stones to crack nuts), grooming, and courtship. While most of these cases involve either tools, foraging, or communication, Robert Sapolsky and Lisa Share report evidence of a higher order cultural tradition in wild baboons in Kenya. Rooted in field observations of a group of olive baboons (called the Forest Troop) since 1978, Sapolsky and Share document the emergence of a unique culture affecting the “overall structure and social atmosphere” of the troop.</p><p>In his book <italic>A Primate's Memoir</italic>, Sapolsky studied the activities and lifestyle of the Forest Troop to explore the relationship between stress and disease. In typical baboon fashion, the males behaved badly, angling either to assume or maintain dominance with higher ranking males or engaging in bloody battles with lower ranking males, which often tried to overthrow the top baboon by striking tentative alliances with fellow underlings. Females were often harassed and attacked. Internecine feuds were routine. Through a heartbreaking twist of fate, the most aggressive males in the Forest Troop were wiped out. The males, which had taken to foraging in an open garbage pit adjacent to a tourist lodge, had contracted bovine tuberculosis, and most died between 1983 and 1986. Their deaths drastically changed the gender composition of the troop, more than doubling the ratio of females to males, and by 1986 troop behavior had changed considerably as well; males were significantly less aggressive.</p><p>After the deaths, Sapolsky stopped observing the Forest Troop until 1993. Surprisingly, even though no adult males from the 1983–1986 period remained in the Forest Troop in 1993 (males migrate after puberty), the new males exhibited the less aggressive behavior of their predecessors. Around this time, Sapolsky and Share also began observing another troop, called the Talek Troop. The Talek Troop, along with the pre-TB Forest Troop, served as controls for comparing the behavior of the post-1993 Forest Troop. The authors found that while in some respects male to male dominance behaviors and patterns of aggression were similar in both the Forest and control troops, there were differences that significantly reduced stress for low ranking males, which were far better tolerated by dominant males than were their counterparts in the control troops. The males in the Forest Troop also displayed more grooming behavior, an activity that's decidedly less stressful than fighting. Analyzing blood samples from the different troops, Sapolsky and Share found that the Forest Troop males lacked the distinctive physiological markers of stress, such as elevated levels of stress-induced hormones, seen in the control troops.</p><p>In light of these observations, the authors investigated various models that might explain how the Forest Troop preserved this (relatively) peaceful lifestyle, complete with underlying physiological changes. One model suggests that nonhuman primates acquire cultural traits through observation. Young chimps may learn how to crack nuts with stones by watching their elders, for example. In this case, the young baboon transplants might learn that it pays to be nice by watching the interactions of older males in their new troop. Or it could be that proximity to such behavior increases the likelihood that the new males will adopt the behavior. Yet another explanation could be that males in troops with such a high proportion of females become less aggressive because they don't need to fight as much for female attention and are perhaps rewarded for good behavior. But it could be that the females had a more direct impact: new male transfers in the Forest Troop were far better received by resident females than new males in the other troops.</p><p>Sapolsky and Share conclude that the method of transmission is likely either one or a combination of these models, though teasing out the mechanisms for such complex behaviors will require future study. But if aggressive behavior in baboons does have a cultural rather than a biological foundation, perhaps there's hope for us as well.</p>
A <italic>plant natriuretic peptide-like </italic>gene in the bacterial pathogen <italic>Xanthomonas axonopodis </italic>may induce hyper-hydration in the plant host: a hypothesis of molecular mimicry	<sec><title>Background</title><p>Plant natriuretic peptides (PNPs) are systemically mobile molecules that regulate homeostasis at nanomolar concentrations. PNPs are up-regulated under conditions of osmotic stress and PNP-dependent processes include changes in ion transport and increases of H<sub>2</sub>O uptake into protoplasts and whole tissue.</p></sec><sec><title>Presentation of the hypothesis</title><p>The bacterial citrus pathogen <italic>Xanthomonas axonopodis </italic>pv. Citri str. 306 contains a gene encoding a PNP-like protein. We hypothesise that this bacterial protein can alter plant cell homeostasis and thus is likely to represent an example of molecular mimicry that enables the pathogen to manipulate plant responses in order to bring about conditions favourable to the pathogen such as the induced plant tissue hyper-hydration seen in the wet edged lesions associated with <italic>Xanthomonas axonopodis </italic>infection.</p></sec><sec><title>Testing the hypothesis</title><p>We found a <italic>Xanthomonas axonopodis </italic>PNP-like protein that shares significant sequence similarity and identical domain organisation with PNPs. We also observed a significant excess of conserved residues between the two proteins within the domain previously identified as being sufficient to induce biological activity. Structural modelling predicts identical six stranded double-psi β barrel folds for both proteins thus supporting the hypothesis of similar modes of action. No significant similarity between the <italic>Xanthomonas axonopodis </italic>protein and other bacterial proteins from GenBank was found. Sequence similarity of the <italic>Xanthomonas axonopodis </italic>PNP-like protein with the <italic>Arabidopsis thaliana </italic>PNP (AtPNP-A), shared domain organisation and incongruent phylogeny suggest that the <italic>PNP-gene </italic>may have been acquired by the bacteria in an ancient lateral gene transfer event. Finally, activity of a recombinant <italic>Xanthomonas axonopodis </italic>protein in plant tissue and changes in symptoms induced by a <italic>Xanthomonas axonopodis </italic>mutant with a knocked-out <italic>PNP-like </italic>gene will be experimental proof of molecular mimicry.</p></sec><sec><title>Implication of the hypothesis</title><p>If the hypothesis is true, it could at least in part explain why the citrus pathogen <italic>Xanthomonas campestris </italic>that does not contain a <italic>PNP</italic>-<italic>like </italic>gene produces dry corky lesions while the closely related <italic>Xanthomonas axonopodis </italic>forms lesions with wet edges. It also suggests that genes typically found in the host, horizontally transferred or heterologous, can help to explain aspects of the physiology of the host-pathogen interactions.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Nembaware</surname><given-names>Victoria</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Seoighe</surname><given-names>Cathal</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Sayed</surname><given-names>Muhammed</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" corresp="yes" contrib-type="author"><name><surname>Gehring</surname><given-names>Chris</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib>	BMC Evolutionary Biology	<sec><title>Background</title><p>Plant natriuretic peptides (PNPs) are a novel class of plant molecules with biological activity at nanomolar concentrations. The PNP-dependent responses include concentration-dependent promotion of stomatal opening [<xref ref-type="bibr" rid="B1">1</xref>], rapid and transient increases in cellular cGMP levels [<xref ref-type="bibr" rid="B2">2</xref>] and modulation of K<sup>+</sup>, Na<sup>+ </sup>and H<sup>+ </sup>net fluxes [<xref ref-type="bibr" rid="B3">3</xref>] in <italic>Zea mays </italic>root tissue. PNPs also induce rapid increases in osmoticum-dependant H<sub>2</sub>O uptake into <italic>Solanum tuberosum </italic>and <italic>Arabidopsis thaliana </italic>protoplasts [<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. We have also observed PNP-dependant increases in lateral H<sub>2</sub>O movement out of the conductive tissue (xylem) into the neighbouring parenchyma [<xref ref-type="bibr" rid="B6">6</xref>] and such a 'drawing' of water into cells and tissues together with an up-regulation under conditions of drought and NaCl stress are compatible with a role for these molecules in plant homeostasis. Incidentally, a PNP-like protein from <italic>Citrus jambhiri </italic>(CjBAp12) is expressed in root and stem tissue in response to a challenge from citrus blight [<xref ref-type="bibr" rid="B7">7</xref>] which proliferates in the conductive tissue of the host and severely affects host homeostasis eventually resulting in xylem plugging and consequent shoot wilting and host death. It is conceivable that the expression CjBAp12 is an early host response to counteract the pathogen induced limitation of water and nutrient availability.</p><p>Several lines of evidence suggest that PNPs can act systemically. Firstly, PNPs are associated with conductive tissues as demonstrated by <italic>in situ </italic>hybridisation and tissue printing [<xref ref-type="bibr" rid="B8">8</xref>]. Secondly, biologically active PNP was isolated from xylem exudates [<xref ref-type="bibr" rid="B8">8</xref>], a tissue that is associated with transport and not protein synthesis. Amino acid sequence comparisons and structural modelling predict that PNPs do not contain the putative polysaccharide-binding C-terminal domain typical for the related expansins that act on the cell wall [<xref ref-type="bibr" rid="B9">9</xref>-<xref ref-type="bibr" rid="B11">11</xref>]. The absence of such a domain presumably results in increased extracellular mobility which in turn is a precondition for a systemic mode of action.</p><p>Here we report the discovery of a gene in the completely sequenced genome of the plant pathogenic bacterium <italic>Xanthomonas axonopodis </italic>[<xref ref-type="bibr" rid="B12">12</xref>] that encodes a protein with significant sequence similarity to an <italic>Arabidopsis thaliana </italic>PNP (AtPNP-A) A. We propose that the presence of a PNP-like protein in <italic>Xanthomonas axonopodis </italic>has enabled the pathogen to affect plant homeostasis. Furthermore, we have investigated the origin of this PNP-like protein encoding gene in <italic>Xanthomonas axonopodis </italic>and have found evidence consistent with the possibility that it has been acquired by the bacterium through horizontal gene transfer. This has led us to search for other genes that show evidence of horizontal transfer with a view to establishing how many genes may have been acquired by <italic>Xanthomonas axonopodis </italic>through horizontal gene transfer from plants.</p></sec><sec><title>Presentation of the hypothesis</title><p>We found a <italic>PNP-like </italic>gene in the bacterial citrus pathogen <italic>Xanthomonas axonopodis </italic>pv. Citri str. 306 that has significant sequence similarity to the PNP encoding genes and hypothesise that the encoded protein can alter homeostasis of the host plant. Since PNP-like molecules are exported into the extracellular space, act systemically and promote significant ion and H<sub>2</sub>O uptake into cells it is very possible that the pathogen uses its PNP-like molecule to induce cell and tissue hyper-hydration in the host. Such hyper-hydration is typically seen in the wet rim of the lesions caused by <italic>Xanthomonas axonopodis </italic>and may suggest that PNP-like molecule benefits the pathogen by facilitating access to water and nutrients while severely disturbing the homeostasis of its host.</p></sec><sec><title>Testing the hypothesis</title><p>The closest homologue of the <italic>Xanthomonas axonopodis </italic>protein NP_642965.1 that motivated this study was the <italic>Arabidopsis thaliana </italic>protein AtPNP-A that we have previously shown to have an important role in plant homeostasis [<xref ref-type="bibr" rid="B13">13</xref>]. The alignment of the two protein sequences (Figure <xref ref-type="fig" rid="F1">1</xref>) shows that they are similar in length (AtPNP-A: 126 amino acids; <italic>Xanthomonas axonopodis </italic>PNP-like protein: 144 amino acids) and that both contain N-terminal transmembrane signal peptides to direct the molecules into the extracellular space, a precondition for a systemic role. Importantly, the molecules show a significantly greater amount (p < 0.05 using a Fishers' Exact Test) of conservation at sites between amino acids 33 and 66 of AtPNP-A (Figure <xref ref-type="fig" rid="F1">1</xref>) that we have previously identified as critical and sufficient for homeostatic function [<xref ref-type="bibr" rid="B5">5</xref>]. Within the entire length the of the domain (Figure <xref ref-type="fig" rid="F1">1</xref>) the identity is 36.4%, the similarity is 43.2% and the gaps are 22.7%.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Alignment of a plant natriuretic peptide from <italic>Arabidopsis thaliana </italic>(AtPNP-A; Accession No. AAD08935) and the plant natriuretic peptide-like protein from <italic>Xanthomonas axonopodis </italic>(Accession No. NP_642965). Solid triangles delineate the domain in AtPNP-A that has been shown to be sufficient to induce increased water uptake into plant protoplasts [<xref ref-type="bibr" rid="B5">5</xref>]. The gray sequence represents the signal peptide and the underlined sequence is the domain spanning the first psi loop. The α helices are marked in red, the dotted red line spans an α helix with a 3–10 helix component (between QNG). The β sheets are marked in blue. Asterisks (*) identify identical amino acid, colons (:) are conservative amino acid replacements and dots (.) are semi-conservative amino acid replacements. Arrows (↑) mark conserved cysteines, the open arrow (↑) marks a position where other PNP-like molecules have a tyrosine or a phenylalanine.</p></caption><graphic xlink:href="1471-2148-4-10-1"/></fig><p>The observed similarity between the two proteins could be due to an ancient horizontal gene transfer event [<xref ref-type="bibr" rid="B14">14</xref>] from the plants to bacteria or to convergent evolution. However, we believe that lateral transfer is more likely because the bacterial and the plant genes also show some similarity outside of the region that we have shown to be essential and sufficient for the function of the protein (Figure <xref ref-type="fig" rid="F1">1</xref>). This similarity in domains not essential for osmotic function suggests that the overall similarity between the two molecules is not just a result of shared function but reflects common ancestry.</p><p>A bootstrapped phylogenetic tree constructed using the <italic>Xanthomonas axonopodis </italic>protein NP_642965.1 and its closest homologues (Figure <xref ref-type="fig" rid="F2">2</xref>) reveals that, if the bacterial gene is indeed a product of horizontal gene transfer, the transfer event is likely to have occurred after the divergence of AtPNP-A from the rest of the expansin protein family. If indeed a plant is the source of this gene through horizontal transfer, it is likely to be the result of a relatively ancient event because the bacterial protein and its plant homologue are significantly diverged and saturated at silent sites.</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>CLUSTALW [26] was used for the multiple sequence alignment of the full length <italic>Xanthomonas axonopodis </italic>PNP-like protein and its homologs obtained from BLASTp searches against the NCBI database. An alignment of 287 amino acids with a score of 17253 was obtained and after discarding all columns with gaps the alignment length was reduced to 89 amino acids with a score of 9705. The Neighbor-Joining tree of the 89 amino acid alignment was constructed using MEGA2 [<xref ref-type="bibr" rid="B33">33</xref>]. Bootstrap values are shown on the branches and the root was placed at the mid-point of the tree. Sequences are named by abbreviations of the species name followed by the NCBI accession number. Abbreviations: At – <italic>Arabidopsis thaliana</italic>, Ca – <italic>Cicer arietium</italic>, Cj – <italic>Citrus jambhri</italic>, Es – <italic>Erucastrum strigosum</italic>, Gh – <italic>Gossypium hirsutum</italic>, Hh – <italic>Hedera helix</italic>, Os – <italic>Oryza sativa</italic>, Xa – <italic>Xanthomonas axonopodis </italic>pv. Citri str. 306 (in red).</p></caption><graphic xlink:href="1471-2148-4-10-2"/></fig><p>Since common structural features in particular within biologically active and/or catalytic domains can support a case for common functionality [<xref ref-type="bibr" rid="B15">15</xref>], we have undertaken a structure prediction approach to compare AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like protein. We used fold recognition methods in a structure prediction metaserver [<xref ref-type="bibr" rid="B16">16</xref>]. The obtained result from FUGUE [<xref ref-type="bibr" rid="B17">17</xref>] which uses structural environment-specific substitution tables and structure-dependent gap penalties reveals with certainty (Z score: >5 for AtPNP-A and >10 for the <italic>Xanthomonas axonopodis </italic>PNP-like protein) that both proteins share the same fold as the N-terminal domain of a Phl P 1 Timothy Grass Pollen Allergen. All the methods in the server gave consistent top hits for Phl P 1. A homology model illustrating the overall fold of AtPNP and the <italic>Xanthomonas axonopodis </italic>PNP-like protein was generated using MODELLER [<xref ref-type="bibr" rid="B18">18</xref>] and was based on the crystal structure of the N-terminal domain of Phl P 1 (Accession No.: P43213) determined to 2.9 Å (pdb code = 1n10) (Figure <xref ref-type="fig" rid="F3">3</xref>). MODELLER implements comparative protein structure modelling by satisfaction of spatial restraints. Furthermore, the structural alignment in FUGE [<xref ref-type="bibr" rid="B17">17</xref>] also gave significant hits (Z score: >4) for both AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like proteins with the barley wound-induced plant defence protein (Barwin). This protein is an endoglucanase-like molecule and endoglucanses have previously been shown to be related to both expansins [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>] and PNP-like molecules [<xref ref-type="bibr" rid="B11">11</xref>]. The basic common fold for these molecules (Figure <xref ref-type="fig" rid="F3">3</xref>) is a double-psi β barrel structure where a six-stranded β barrel assumes a pseudo-twofold axes in which the parallel strands form two psi structures [<xref ref-type="bibr" rid="B21">21</xref>]. The first psi loop connects strands β1 and β2, whereas the second psi loop connects strands β4 and β5 (Figure <xref ref-type="fig" rid="F3">3</xref>). In the currently known structures, the active sites of the protein cluster around the psi loops indicating that its protrusion and free main chain functional groups may be well suited to providing a framework for catalysis [<xref ref-type="bibr" rid="B21">21</xref>].</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Modelled fold of a PNP-like molecule showing the six stranded double-psi β barrel structure. Fold recognition methods predict with certainty (Z score: >5) that AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like molecule both adopt this fold. The N- and C-terminus of the protein are indicated, the α-helices are in red, the 6 β-strands are in blue and the two protruding psi loops are marked with a solid arrow (↑). The open arrows (↑) delineate the 33 amino acid long domain critical and sufficient for biological activity [<xref ref-type="bibr" rid="B5">5</xref>]. The N-terminal signal peptide that is not required for biological function outside the cell [<xref ref-type="bibr" rid="B5">5</xref>] was not included in the model. The model was generated using the software MOLSCRIPT [<xref ref-type="bibr" rid="B34">34</xref>].</p></caption><graphic xlink:href="1471-2148-4-10-3"/></fig><p>In AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like protein the first psi loop connects strands β1 and β2, whereas the second psi loop connects strands β4 and β5 (Figure <xref ref-type="fig" rid="F3">3</xref>). The sequence conservation between AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like molecule is greatest in the domain spanning the β2 and β3 strands which both flank the α helix (Figure <xref ref-type="fig" rid="F1">1</xref> and <xref ref-type="fig" rid="F3">3</xref>). In AtPNP-A this structure (β2 – α helix – β3) has been demonstrated to be within the 33 amino acid long domain that is critical and sufficient for conferring biological activity [<xref ref-type="bibr" rid="B5">5</xref>]. This domain also contains the first psi loop which is likely to be a part of the functional framework of AtPNP-A and the <italic>Xanthomonas axonopodis </italic>PNP-like molecule.</p><p>We also carried out a screen of all proteins from <italic>Xanthomonas axonopodis </italic>in order to discover whether other genes from this bacterial pathogen showed evidence of unexpected tree topology thus indicating horizontal gene transfer [<xref ref-type="bibr" rid="B14">14</xref>] from plants. All known proteins from the completely sequenced genomes of <italic>Xanthomonas axonopodis </italic>[<xref ref-type="bibr" rid="B12">12</xref>], <italic>Xanthomonas campestris </italic>[<xref ref-type="bibr" rid="B12">12</xref>], <italic>Pseudomonas putida </italic>[<xref ref-type="bibr" rid="B22">22</xref>], <italic>Escherichia coli </italic>[<xref ref-type="bibr" rid="B23">23</xref>] and <italic>Arabidopsis thaliana </italic>[<xref ref-type="bibr" rid="B24">24</xref>] were downloaded from GenBank (13/08/2003). The <italic>Xanthomonas axonopodis </italic>proteins were searched against the proteins from the remaining four organisms using BLASTp [<xref ref-type="bibr" rid="B25">25</xref>]. CLUSTALW [<xref ref-type="bibr" rid="B26">26</xref>] was used to generate multiple sequence alignments and Neighbour-Joining phylogenetic trees from 4307 sets of five proteins consisting of one protein from each of the five organisms. <italic>Xanthomonas axonopodis </italic>proteins with greater than 25% identity to their <italic>Arabidopsis thaliana </italic>homologues that clustered with the <italic>Arabidopsis thaliana </italic>homologue on a phylogenetic tree were retained for further analysis. Each of these proteins was searched against the GenBank non-redundant protein database. Homologous sequences were downloaded and phylogenetic trees were constructed using the Neighbor-Joining method [<xref ref-type="bibr" rid="B26">26</xref>].</p><p>This approach was used to determine the number of <italic>Xanthomonas axonopodis </italic>genes that showed evidence of horizontal acquisition from plants. The initial screen with the five completely sequenced organisms returned seven cases of putative horizontal transfers (Table <xref ref-type="table" rid="T1">1</xref>). However, only two of the proteins (Table <xref ref-type="table" rid="T1">1</xref>), NP_642965.1, the subject of this study, and NP_643621.1 had no significant bacterial homologs in the NCBI database using the default E-value cut-off 10. The remaining proteins had bacterial homologues that suggested that they were not likely to have been acquired through horizontal transfer. The search indicated that horizontal transfer of genes between plants and the plant pathogen <italic>Xanthomonas axonopodis</italic>, if it has indeed occurred, has been rare.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Analysis of putative laterally transferred genes obtained from the whole genome analyses.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="center" colspan="2"><bold>Proteins:</bold></td><td align="center"><bold>Function:</bold></td><td align="center"><bold>Bacterial homologs:</bold></td></tr><tr><td align="left"><bold><italic>X. axonopodis</italic></bold></td><td align="left"><bold><italic>A. thaliana</italic></bold></td><td></td><td></td></tr></thead><tbody><tr><td align="left">NP_640439</td><td align="left">NP_564216</td><td align="left">short-chain dehydrogenase</td><td align="center">yes</td></tr><tr><td align="left">NP_641062</td><td align="left">NP_196873</td><td align="left">N-acetylglucosaminidase</td><td align="center">yes</td></tr><tr><td align="left">NP_642289</td><td align="left">NP_196225</td><td align="left">3-oxoacyl-[ACP] reductase</td><td align="center">yes</td></tr><tr><td align="left">NP_642965</td><td align="left">NP_194767</td><td align="left">hypothetical protein</td><td align="center">no</td></tr><tr><td align="left">NP_643053</td><td align="left">NP_568712</td><td align="left">amine oxidase related</td><td align="center">yes</td></tr><tr><td align="left">NP_643621</td><td align="left">NP_173748</td><td align="left">expressed protein</td><td align="center">no</td></tr><tr><td align="left">NP_644089</td><td align="left">NP_182049</td><td align="left">methionine aminopeptidase</td><td align="center">yes</td></tr></tbody></table></table-wrap><p>Recently, another example of a pathogen mimicking an extracellular plant molecule has been reported [<xref ref-type="bibr" rid="B27">27</xref>]. This protein (GrEXP1), a molecule with cell wall loosening (expansin) activity previously seen in plants [<xref ref-type="bibr" rid="B28">28</xref>] and other organisms with cell walls only [<xref ref-type="bibr" rid="B29">29</xref>], was found in the plant-parasitic roundworm <italic>Globodera rostochiensis</italic>. The infective juvenile nematodes express and secrete GrEXP1 in the subventral oesophageal glands [<xref ref-type="bibr" rid="B27">27</xref>] using this 'typical plant' protein to their advantage when invading the host root system.</p><p>Finally, if the <italic>PNP-like </italic>gene was indeed horizontally transferred from a plant to <italic>Xanthomonas axonopodis </italic>it is also consistent with the complexity theory of gene transfer [<xref ref-type="bibr" rid="B30">30</xref>] which postulates that a major factor in the more frequent horizontal transfer of operational genes such as <italic>expansins </italic>and <italic>PNPs </italic>as compared to informational genes is that they are structurally and functionally less complex. This bias is explained by the increased chance of transfer of a functional unit advantageous to the recipient. It would also appear that particularly in the case of a pathogen, extracellular signals, transporters or surface components perceived by the host can cause systemic host responses that give the pathogen a significant advantage. A point in case are eukaryotic genes found in <italic>Mycobacterium tuberculosis </italic>many of which directly modulate host responses and have a role in the specific pathogenesis induced by the bacterium [<xref ref-type="bibr" rid="B31">31</xref>].</p><p>The experimental test of the hypothesis of molecular mimicry of the <italic>Xanthomonas axonopodis </italic>PNP-like molecule will require two types of investigations. In the first, a recombinant <italic>Xanthomonas axonopodis </italic>protein must be obtained and tested for effects on (host) plant tissue. Molecular mimicry would require that net H<sub>2</sub>O uptake is increased and ion transport is affected in the host tissue in response to the recombinant peptide. In a second experiment, a <italic>Xanthomonas axonopodis </italic>mutant with a knocked-out <italic>PNP-like </italic>gene must be obtained. If the mutant induces altered host symptoms and in particular an absence of watery edges of the lesions, then the hypothesis can be considered proven.</p></sec><sec><title>Implications</title><p>If the hypothesis is true, then the bacterial PNP-like protein plays a role in manipulating the homeostatic balance of the host. Such mimicry could at least in part explain why the citrus pathogen <italic>Xanthomonas campestris </italic>that does not contain a <italic>PNP-like </italic>gene produces dry corky lesions while the closely related <italic>Xanthomonas axonopodis </italic>forms wet lesions [<xref ref-type="bibr" rid="B32">32</xref>]. Furthermore, the hypothesis suggests that the presence of "typical" and functional host genes in pathogens can explain key aspects of host-pathogen interactions in general and can help elucidate the specific molecular and cellular interactions between hosts and pathogens.</p></sec><sec><title>Authors' contributions</title><p>VN and CS carried out the bioinformatics and phylogenetic analyses, MS performed the structural analysis and CG advised on the biological function of PNPs and drafted the manuscript. All authors contributed to the editing of the manuscript and approved of the final version.</p></sec>
Reversibility of stress-echo induced ST-segment depression by long-term oral n-3 PUFA supplementation in subjects with chest pain syndrome, normal wall motion at stress-echo and normal coronary angiogram	<sec><title>Background</title><p>Normal coronary arteries may coexist with abnormal coronary and systemic endothelial function in patients with chest pain. Recent work by the renowned Pisa echo-group elegantly suggests that isolated ST-segment depression during stress-echo (SE) can be used as a marker of coronary endothelial dysfunction, in the absence of stress-inducible wall motion abnormalities and in the absence of angiographically-significant coronary artery disease (CAD). The long chain n-3 polyunsaturated fatty acids (PUFAs) have been reported to possess several properties that may positively influence vascular function. The present study's hypothesis is that a 4 month-course of oral supplementation with n-3 PUFAs can reverse endothelial dysfunction.</p></sec><sec sec-type="methods"><title>Methods</title><p>Subjects were selected on the basis of the following criteria: 1) reported chest pain syndrome, 2) significant ST-segment depression during an otherwise normal SE, 3) absence of angiographically-significant CAD. Subjects underwent a 4-month course of oral supplementation with commercially available n-3 PUFA, 1 g once a day. Normalization of endothelial dysfunction was defined, at the end of the supplementation period, by the absence of significant ST-segment depression during repeat SE. We tested the aforementioned hypothesis in a very small series of consecutive subjects, with the intent to produce a hypothesis-generating study.</p></sec><sec><title>Results</title><p>Seven out of the total nine subjects enrolled (77.8%) had normal ST-segment during repeat SE performed after the 4 month course of therapy.</p></sec><sec><title>Conclusions</title><p>A striking rate of reversion of SE-induced ST-segment depression after oral n-3 PUFAs suggests reversion of coronary endothelial dysfunction; nonetheless these data need to be validated in larger, placebo-controlled studies.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Gaibazzi</surname><given-names>Nicola</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Ziacchi</surname><given-names>Vigilio</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Cardiovascular Disorders	<sec><title>Background</title><p>Normal coronary arteries may coexist with abnormal coronary and systemic endothelial function in patients with chest pain [<xref ref-type="bibr" rid="B1">1</xref>]. The functional status of systemic endothelium somewhat mirrors the coronary endothelial function [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>] and can be evaluated non-invasively with brachial ultrasound [<xref ref-type="bibr" rid="B4">4</xref>]. A novel, provocative paper by the Pisa echo-lab group [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>], suggests that endothelial dysfunction strictly correlates with ST-segment depression (STSD) during pharmacologic (as well as exercise) stress, while epicardial coronary artery stenosis exclusively affects wall motion abnormalities (WMA), although both components may coexist; they demonstrated that abnormal brachial artery flow-mediated dilation (measured with brachial artery ultrasound) was predicted by STSD during SE but not by either presence of WMA during SE or angiographically significant CAD. There was no correlation between brachial artery flow-mediated dilation and extent of angiographical CAD.</p><p>Therefore isolated STSD during SE could be used as a surrogate for coronary endothelial dysfunction, in the absence of stress-inducible WMA and in the absence of angiographically-significant CAD.</p><p>The majority of cardiovascular benefits of PUFAs are likely to be mediated in the vascular wall and at the vascular endothelium level. The long chain PUFAs have been reported to possess several properties that may positively influence vascular function. These include favorable mediator profiles (nitric oxide, eicosanoids) that influence vascular reactivity, change in vascular tone via actions on selective ion channels [<xref ref-type="bibr" rid="B7">7</xref>], and maintenance of vascular integrity. In addition to direct effects on contractility, PUFAs may affect vascular function by modifying expression of inflammatory cytokines and adhesion molecules. These properties may explain the beneficial cardiovascular protection of this family of fatty acids that have been clearly evident through epidemiological data as well as from few clinical trials [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Although many epidemiological surveys and basic research studies have suggested that "marine fish oil" cardioprotective effect is consistent and at least partly mediated through an action on the endothelium, no large clinical intervention trial has yet addressed the effect of commercially-available PUFAs preparations on a clinically-measurable direct marker of coronary endothelial dysfunction in normocholesterolemic subjects [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. A small, randomized study performed in hypercholesterolemic subjects suggests a positive effect of 4-month marine oil supplementation on brachial artery endothelium-mediated vasodilation.[<xref ref-type="bibr" rid="B12">12</xref>]</p><p>The present study's hypothesis is that a 4 month-course of oral supplementation with PUFAs can reverse coronary endothelial dysfunction. Eventual "normalization" of endothelial dysfunction was measured, at the end of the supplementation period, by the absence of both significant STSD and WMA during repeat SE.</p></sec><sec sec-type="methods"><title>Methods</title><p>Patients were enrolled in the study if they met the following inclusion criteria: (1) chest pain syndrome; (2) reference for SE; (3) presence of STSD (strict ECG criteria were applied: a) STSD ± 1 mm 0.08 s after the J point in the peripheral leads and/or b) STSD ≥ 2 mm 0.08 s after the J point in the precordial leads) during a SE with no changes in wall motion score, (4) coronary angiogram performed within 20 days after SE, (5) no nitrate therapy at the time of SE.</p><p>When meeting the aforementioned entry criteria, patients had their total cholesterol measured by means of a standard whole-blood sample.</p><p>Subjects were referred for SE because of one of the following: (a) patients with positive exercise-ECG in concomitance with conditions lowering the positive predictive value of the ECG marker of ischaemia (for instance, female gender or arterial hypertension); (b) patients with recurring chest pain in the absence of significant electrocardiographic changes during routine exercise-ECG; (c) patients unable to exercise. All of the patients underwent dipyridamole SE.</p><p>The following were exclusion criteria: (1) uninterpretable resting ECG (left bundle branch block, preexcitation, baseline STSD or elevation > 1 mm, or typical digitalis repolarization pattern); (2) technically poor acoustic window; (3) cardiomyopathy or severe valvular disease; (4) previously detected CAD; (5) statin therapy; (6) high total cholesterol levels (defined as total cholesterol ≥ 240 mg/dl as suggested by current NCEP guidelines).</p><p>The study was approved by the institutional review board. All patients gave their written informed consent before entering the study.</p><p>Stress echo was performed with dipyridamole up to 0.84 mg/kg over 10 min with atropine up to 2 mg when needed; stress echo was repeated using the same protocol after a 4-month course of oral supplementation with commercially available (Seacor<sup>®</sup>, Esapent<sup>®</sup>, Eskim<sup>®</sup>) n-3 PUFA at a dose of 1 g once a day. An imaging system with digital acquisition was used (GE-Vivid 7). All standard echocardiographic views were obtained when possible. The left ventricle was divided into 16 segments, as suggested by the American Society of Echocardiography. Segmental wall motion was graded as follows: normal = 1; hypokinetic = 2; akinetic = 3; and dyskinetic = 4. Inadequately visualized segments were not scored. Stress echo was considered positive when one left ventricular segment was increased by one grade or more at peak stress. Normalization of coronary endothelial dysfunction was defined as isoelectric ST segment 0.08 s after the J point in the peripheral leads and STSD < 0.5 mm 0.08 s after the J point in the precordial leads during repeat SE with no WMA.</p><p>Coronary angiography in multiple views was performed according to the standard Judkins or Sones technique.</p><p>The percent diameter stenosis was determined by quantitative coronary angiography. A vessel was considered to have significant obstruction if its diameter was narrowed by 50% or more, with respect to the pre-stenotic tract.</p></sec><sec><title>Results</title><p>Nine consecutive patients were prospectively enrolled according to inclusion/exclusion criteria during the time period between January 2003 and June 2003. Three patients (out of the 12 who fulfilled entry criteria) were excluded because of the presence of at least one exclusion criteria (2 subjects had high total cholesterol levels as per protocol defined and one was already on statin therapy).</p><p>Five patients were male (55%); mean age was 60(± 5).</p><p>Four patients (44%) had hypertension (treated with ACE-inhibitors, ARBs, diuretics or combinations of them), 3(33%) had diabetes (treated with a combination of metformin and a sulfonylurea), 2(22%) had total cholesterol levels in the range between 200 mg/dl and 240 mg/dl (defined as borderline high by current NCEP guidelines), 3(33%) were active smokers. None was under antiischaemic therapy (beta-blocker and/or non-dihydropiridinic calcium antagonist) at the time of SE testing. All of the patients maintained their usual therapy througout the study time period, with no exceptions.</p><p>All but 2 subjects had normalization of STSD during repeat SE performed after the 4 month course of therapy. Of the two subjects still presenting STSD at repeat SE, one was hypertensive and non-diabetic while the other was diabetic but non-hypertensive).</p></sec><sec><title>Discussion</title><p>We tested the study hypothesis in a small series of consecutive subjects, with the intent to produce a small hypothesis-generating study; testing the hypothesis in a large, randomized, placebo-controlled trial was felt premature, both because we used an innovative, but not definitely validated marker of coronary endothelial dysfunction and because lack of strong clinical data about the effect of PUFAs supplementation on human coronary endothelium.</p><p>Our study, as any small "proof of concept" study, has many limitations:</p><p>1) It is a prospective, non-randomized, uncontrolled interventional study, 2) it is performed in a very small group of subjects, 3) it is suggestive for a positive-reversal effect of PUFAs on isolated, angiographically-negative STSD during SE, but even if recent data strongly suggest this clinical pattern to be linked with coronary endothelial dysfunction (5,6), this is not a direct measure of endothelial function and, also, no data exist demonstrating that reversal of STSD indicates normalization of coronary endothelial dysfunction.</p><p>Of course no subgroup analysis can be performed in such a small group of subjects, even if it is interesting that non-responders (2 out of 9 subjects) consisted of one hypertensive, non-diabetic subject and one diabetic non-hypertensive subject, suggesting that inefficacy of therapy may not be linked to the presence/absence of these clinical conditions.</p><p>Even if evidence is strong for a negative prognostic effect of coronary endothelial dysfunction on mortality, atherosclerosis progression and plaque instability [<xref ref-type="bibr" rid="B13">13</xref>], coronary endothelial function is to date difficult to measure on clinical grounds: hence it is obvious cardiologists prefer to dilate what they can clearly see (epicardial stenosis) rather than pharmacologically engage what they can not measure (endothelium).</p><p>An easy to obtain, clinically useful marker of coronary endothelial dysfunction such as isolated STSD during SE in the absence of visible CAD would be very welcome in the cardiology arena, while the striking effect exerted by PUFAs in this small study would be even more valuable.</p></sec><sec><title>Conclusions</title><p>Striking rate of reversion of stress-induced ST-segment depression after oral n-3 PUFAs suggests reversion of coronary endothelial dysfunction; nonetheless these data need to be validated in larger, placebo-controlled studies utilizing direct measurements of coronary endothelial function.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>NG conceived the study and materially performed echocardiographic examinations, VG participated in the coordination of the study and revision of the paper. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2261/4/1/prepub"/></p></sec>
The inter-observer agreement of examining pre-school children with acute cough: a nested study	<sec><title>Background</title><p>The presence of clinical signs have implications for diagnosis, prognosis and treatment. Therefore, the aim of this study was to examine the inter-observer agreement of clinical signs in pre-school children presenting to primary care.</p></sec><sec sec-type="methods"><title>Methods</title><p>A nested study comparing two clinical assessments within a prospective cohort of 256 pre-school children with acute cough recruited from eight general practices in Leicestershire, UK. We examined agreement (using kappa statistics) between unstandardised and standardised clinical assessments of tachypnoea, chest signs and fever.</p></sec><sec><title>Results</title><p>Kappa values were poor or fair for all clinical signs (range 0.12 to 0.39) with chest signs the most reliable.</p></sec><sec><title>Conclusions</title><p>Primary care clinicians should be aware that clinical signs may be unreliable when making diagnosis, prognosis and treatment decisions in pre-school children with cough. Future research should aim to further our understanding of how best to identify abnormal clinical signs.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Hay</surname><given-names>Alastair D</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Wilson</surname><given-names>Andrew</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Fahey</surname><given-names>Tom</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Peters</surname><given-names>Tim J</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Family Practice	<sec><title>Background</title><p>Cough is the most frequently managed problem in primary care and becomes increasingly common at the extremes of age [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Cough in pre-school children is usually due to simple, self limiting respiratory tract infection, but more severe causes need to be ruled out including pneumonia, bronchiolitis, pertussis, croup and asthma[<xref ref-type="bibr" rid="B2">2</xref>]. The presence of clinical signs may have diagnostic, prognostic, and treatment implications. The absence of tachypnoea has been shown to be most useful for ruling out pneumonia[<xref ref-type="bibr" rid="B3">3</xref>], and fever is associated with poor outcome in children with cough[<xref ref-type="bibr" rid="B4">4</xref>] and otitis media[<xref ref-type="bibr" rid="B5">5</xref>]. In a study of cough in adults, antibiotics were eight times more likely to be prescribed in patients with abnormal chest signs[<xref ref-type="bibr" rid="B6">6</xref>], and in another study 93% of adults presenting with the combination of cough and chest signs received antibiotics[<xref ref-type="bibr" rid="B7">7</xref>].</p><p>The reliability and accuracy of respiratory symptoms and signs have been assessed almost exclusively in secondary care[<xref ref-type="bibr" rid="B8">8</xref>], where relatively serious illness is more prevalent[<xref ref-type="bibr" rid="B9">9</xref>]. Given the diagnostic, prognostic and treatment implications of these clinical signs, we decided to examine the inter-observer agreement between a standardised and non-standardised clinical assessment in pre-school children presenting with acute cough in primary care. These were children already recruited to a cohort study investigating duration and complications of cough[<xref ref-type="bibr" rid="B4">4</xref>,<xref ref-type="bibr" rid="B10">10</xref>].</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Practices and participants</title><p>Practice and participant recruitment have been described in detail elsewhere[<xref ref-type="bibr" rid="B10">10</xref>]. The Leicestershire Research Ethics Committee approved the study. To maximise the efficiency of child recruitment, practices with list sizes greater than 8000 were invited by letter to participate. Recruitment took place from November to April over two years between 1999 and 2001, at morning and evening surgeries rotated between practices. A researcher was located in the surgery during recruitment sessions to ensure all eligible children were invited to participate. These were children aged 0–4 years with a cough ≤ 28 days duration presenting to a General Practitioner (GP) or Nurse Practitioner (NP), without asthma (defined as recommended to be receiving preventive or regular reliever treatment) or any other chronic disease. Two observers examined each child.</p></sec><sec><title>Observer one</title><p>This was the GP or NP to whom the child presented. Our aim was not to alter the clinical assessment of observer one, but to ask the clinician to perform a routine, non-standardised, examination of the child. A standardised data collection sheet [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>] included questions about respiratory rate, the presence of fever and chest signs, but only examined items were recorded. For respiratory rate and temperature, clinicians were asked to give a global opinion of abnormality. They were not required to count breaths per minute or use a thermometer, though they could record these data if they wished. Similarly, if the clinician auscultated the chest, they were able to record if abnormal signs (wheezes or crepitations) were present.</p></sec><sec><title>Observer two</title><p>This was one general practitioner (ADH), who performed a standardised clinical assessment within 30 minutes, before or after, observer one and was blind to the results of the other assessment. Data collected differed between children presenting in the first and second winters. In the first winter, we included a global assessment of the child's respiratory rate and auscultation of all respiratory zones of the chest. However, by the second winter, it became apparent that, in addition to the global assessment, we wanted a more accurate measure of temperature and respiratory rate [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>]. We used a mercury thermometer placed in the axilla for five minutes and counted breaths over a 30 to 60 second period of settled behaviour[<xref ref-type="bibr" rid="B11">11</xref>].</p></sec><sec><title>Sample size</title><p>The sample size was determined by the primary research question, which was to quantify cough duration[<xref ref-type="bibr" rid="B10">10</xref>]. For this study, sample size is best considered through the precision attained in the agreement analyses as shown by the 95% confidence limits in Table <xref ref-type="table" rid="T2">2</xref>.</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Inter-observer agreement values</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Clinical sign</bold></td><td align="left"><bold>Number with complete data (%)</bold></td><td align="left"><bold>Observer one positive sign (%)</bold></td><td align="left"><bold>Observer two positive sign (%)</bold></td><td align="left"><bold>Kappa (chance corrected agreement with 95% CI)<sup>b</sup></bold></td><td align="left"><bold>Phi (chance independent agreement)<sup>c</sup></bold></td></tr></thead><tbody><tr><td align="left">Raised respiratory rate (observer one opinion vs. observer two opinion)</td><td align="left">214 (84%)</td><td align="left">8.8%</td><td align="left">5.5%</td><td align="left">0.29 (0.16, 0.43)</td><td align="left">0.54</td></tr><tr><td align="left">Raised respiratory rate (observer one opinion vs. observer two counted rate)</td><td align="left">93 (80%)a</td><td align="left">8.8%</td><td align="left">51.6%</td><td align="left">0.12 (0.009, 0.23)</td><td align="left">0.47</td></tr><tr><td align="left">Fever (observer one opinion vs. observer two measured)</td><td align="left">103 (89%)a</td><td align="left">10.8%</td><td align="left">3.9%</td><td align="left">0.18 (0.005, 0.35)</td><td align="left">0.42</td></tr><tr><td align="left">Abnormal chest signs (observer one opinion vs. observer two opinion)</td><td align="left">209 (82%)</td><td align="left">21.5%</td><td align="left">14.2%</td><td align="left">0.39 (0.26, 0.53)</td><td align="left">0.51</td></tr></tbody></table><table-wrap-foot><p><sup>a </sup>Second winter data only, 116 children recruited. <sup>b </sup>Strength of agreement; < 0.2 poor, 0.2 – 0.4 fair, 0.41 – 0.6 moderate, 0.61 – 0.8 good, 0.81 – 1.0 very good.<sup>15 </sup><sup>c </sup>-1 perfect disagreement, 0 agreement no better than chance, +1 perfect agreement<sup>13</sup></p></table-wrap-foot></table-wrap></sec><sec><title>Data entry and analysis</title><p>Data were single entered onto an Access database. No errors were found in 14 randomly selected cases. We used Stata version 7 to describe the clinical assessment data and generate chance adjusted (kappa) inter-observer agreement statistics[<xref ref-type="bibr" rid="B12">12</xref>]. Because kappa values decrease as the proportion of positive ratings become extreme, even when observers interpret signs consistently, we also calculated chance independent agreement values, or phi[<xref ref-type="bibr" rid="B13">13</xref>]. For the second winter data from observer two, the counted respiratory rates were converted into a binary variable using 40 breaths per minute as the upper limit of normal for children aged up to one year and 30 breaths per minute for children aged up to five years of age [<xref ref-type="bibr" rid="B14">14</xref>]. Similarly, measured temperatures were converted using an upper limit of normal of 37.5°C [<xref ref-type="bibr" rid="B11">11</xref>]. We did not compare the thermometer derived continuous measurements because of the small number of children in whom these data were available from both observers (23) and because we felt it was clinically more useful to dichotomise children into febrile or afebrile.</p></sec></sec><sec><title>Results</title><sec><title>Descriptive statistics</title><p>The cohort has been described in detail elsewhere[<xref ref-type="bibr" rid="B10">10</xref>]. We recruited 89% of eligible children presenting to 124 morning or evening surgeries at eight practices: two hundred and fifty six in total, 116 from the second winter. The two main reasons for not recruiting the 11% of eligible children were parental refusal and inability to read/write English. Sixty-one GPs and three NPs performed the role of observer one, and 96% of children were seen by a GP. Global assessment data from observer one were available in 98% of children for temperature and respiratory rate and 96% of children for chest signs. For observer two (ADH), data were available in 81% of children for respiratory rate, 85% for chest signs and 89% of children for temperature. Table <xref ref-type="table" rid="T1">1</xref> summarises the clinical data. For the first observer, one or more abnormal clinical findings were found in 80/241 (33%) of children with data complete for all three signs. Abnormal chest signs were found in 22%, fever in 11% and tachypnoea in 9%.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Clinician and researcher clinical assessments</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Variables</bold></td><td align="left"><bold>Observer one (un-standardised assessment)<sup>a,c</sup></bold></td><td align="left"><bold>Observer two (standardised assessment)<sup>c</sup></bold></td></tr></thead><tbody><tr><td align="left">Breaths per minute counted</td><td align="left">61/250 (24.4%)</td><td align="left">95/116 (81.9%)<sup>b</sup></td></tr><tr><td align="left">Counted respiratory rate raised</td><td align="left">15/61 (24.6%)</td><td align="left">49/95 (51.6%)<sup>b</sup></td></tr><tr><td align="left">Raised respiratory rate (global opinion)</td><td align="left">22/250 (8.8%)</td><td align="left">12/218 (5.5%)<sup>a</sup></td></tr><tr><td align="left">Temperature recorded using thermometer</td><td align="left">61/250 (24.4%)</td><td align="left">103/116 (88.8%)<sup>b</sup></td></tr><tr><td align="left">Temperature recorded and raised (> 37.5°C)</td><td align="left">6/61 (9.8%)</td><td align="left">4/103 (3.9%)<sup>b</sup></td></tr><tr><td align="left">Fever (global opinion)</td><td align="left">27/250 (10.8%)</td><td align="left">Not examined.</td></tr><tr><td align="left">Abnormal chest signs</td><td align="left">53/246 (21.5%)</td><td align="left">31/218<sup>a </sup>(14.2%)</td></tr></tbody></table><table-wrap-foot><p><sup>a </sup>Data collected from both winters <sup>b </sup>Data collected on consecutive children for second winter only <sup>c </sup>Denominators vary due to missing data</p></table-wrap-foot></table-wrap></sec><sec><title>Inter-observer agreement</title><p>The number of children in whom inter-observer agreement was assessed is shown in Table <xref ref-type="table" rid="T2">2</xref>. Kappa values were poor to fair for all clinical signs (range 0.12 to 0.39) with chest signs the most reliable[<xref ref-type="bibr" rid="B15">15</xref>]. Phi values showed less variation (range 0.42 to 0.51), with raised respiratory rate the most reliable.</p></sec></sec><sec><title>Discussion</title><sec><title>Summary of main results</title><p>This study shows that in usual practice, primary care clinicians found one or more abnormal sign in a third of pre-school children with cough in primary care, and used a thermometer or formally counted the respiratory rate in a quarter. The inter-observer agreement between un-standardised and standardised assessments of these signs was at best fair.</p></sec><sec><title>Interpretation of results</title><p>Children presenting to primary care are seen earlier in the natural history of their condition than those presenting to secondary care, when signs are likely to be less subtle. Although we found similar levels of inter-observer agreement to studies in secondary care, it is disappointing that the kappa values were not higher. This may in part be explained by the low proportion with abnormal signs (as judged by either observer). This leads to paradoxically low kappa values[<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. We therefore also calculated phi values and, as would be expected, these showed less sensitivity to the proportion with positive signs. In general though, the level of agreement achieved calls into question the usefulness of signs in everyday clinical practice to assist diagnosis, prognosis and antibiotic treatment. For example, kappa values of ≥ 0.6 are recommended if symptoms or signs are to be used in clinical prediction rules[<xref ref-type="bibr" rid="B18">18</xref>]. In part, it may explain the wide variation seen in diagnostic labels used for respiratory tract infection in primary care[<xref ref-type="bibr" rid="B19">19</xref>]. However, it is possible that agreement might be improved if clinicians adopt a more standardised approach to assessment.</p><p>The second observer found a higher proportion of children with tachypnoea using counted respiratory rate compared with the global assessment. Previous research suggests that this may be because, in their global assessment of respiratory rate, clinicians adjust for other factors such as the child's general condition, presence of cyanosis, respiratory effort and accessory muscle use[<xref ref-type="bibr" rid="B3">3</xref>].</p></sec><sec><title>Where this fits in with other research</title><p>Notwithstanding the levels observed, our study has demonstrated similar inter-rater agreement to previous studies using higher levels of standardisation of examination in children and adults in secondary care. Studies of infants summarised in a review found inter-rater kappa values of 0.49 for respiratory retractions, 0.59 for accessory muscle use, 0.3 for crepitations and 0.29 for wheezing[<xref ref-type="bibr" rid="B3">3</xref>]. A study of adults found inter-rater kappas of 0.25 for tachypnoea, 0.51 for wheezes, 0.41 for crackles and 0.32 for bronchial breath sounds[<xref ref-type="bibr" rid="B20">20</xref>].</p></sec><sec><title>Limitations</title><p>While we have no reason to believe that the children recruited in the second winter differ systematically from those from the first winter, the lower number of children with measured temperature and counted respiratory rate from the second winter limits the precision of these estimates in our study. Respiratory rate can fluctuate quickly and it is possible that the 30 minutes maximum between clinical assessments explains some of the poor agreement. Our desire to compare usual clinical practice with a standardised assessment means we have not been able to assess the agreement of counted respiratory rate or thermometer measured temperature or further our understanding of how the clinicians identify abnormal clinical signs. We do not know from this study whether the standardised or non-standardised assessment is more accurate at predicting diagnosis or prognosis, nor have we assessed the intra-observer agreement of clinical signs. It is possible that the data collection form altered the clinical behaviour of observer one. This may have changed the number of children identified with abnormal signs, counted respiratory rate or thermometer-measured temperature. While we used mercury thermometry for the standardised assessment, we acknowledge its use in day-to-day practice is limited by the inconvenience of prolonged measurement time.</p></sec></sec><sec><title>Conclusions</title><p>Primary care clinicians should be aware that clinical signs may be unreliable when making diagnosis, prognosis and treatment decisions in pre-school children with cough. Future research should aim to further our understanding of how best to identify abnormal clinical signs and examine the inter- and intra-observer agreement of standardised clinical assessments.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>AH and AW conceived the idea for the study and AH analysed the data. AH drafted the paper with subsequent contributions from all the authors. AH is the guarantor.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2296/5/4/prepub"/></p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional File 1</title><p>This was the data collection form used for the clinical data recorded by observer one (sections G and H, pages 5 and 6) and observer two (section P, page 10).</p></caption><media xlink:href="1471-2296-5-4-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec>
Genetic analysis of haemophilia A in Bulgaria	<sec><title>Background</title><p>Haemophilias are the most common hereditary severe disorders of blood clotting. In families afflicted with heamophilia, genetic analysis provides opportunities to prevent recurrence of the disease. This study establishes a diagnostical strategy for carriership determination and prenatal diagnostics of haemophilia A in Bulgarian haemophilic population.</p></sec><sec sec-type="methods"><title>Methods</title><p>A diagnostical strategy consisting of screening for most common mutations in the factor VIII gene and analysis of a panel of eight linked to the factor VIII gene locus polymorphisms was established.</p></sec><sec><title>Results</title><p>Polymorphic analysis for carrier status determination of haemophilia A was successful in 30 families out of 32 (94%). Carrier status was determined in 25 of a total of 28 women at risk (89%). Fourteen prenatal diagnoses in women at high risk of having a haemophilia A – affected child were performed, resulting in 6 healthy boys and 5 girls.</p></sec><sec><title>Conclusion</title><p>The compound approach proves to be a highly informative and cost-effective strategy for prevention of recurrence of haemophilia A in Bulgaria. DNA analysis facilitates carriership determination and subsequent prenatal diagnosis in the majority of Bulgarian families affected by haemophilia A.</p></sec>	<contrib id="A1" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Petkova</surname><given-names>Rumena</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Chakarov</surname><given-names>Stoian</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Kremensky</surname><given-names>Ivo</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Blood Disorders	<sec><title>Background</title><p>Haemophilia A is a common inherited disorder of blood clotting, inherited in a recessive X-linked pattern. The incidence of the disease is estimated at app. 1:8000 males.</p><p>Affected individuals develop a variable phenotype of hemorrhage into joints and muscles, easy bruising, and prolonged bleeding from wounds. The severity and frequency of bleeding in heamophilia A is inversely related to the amount of residual factor.</p><p>Haemophilia A is caused by deficiency of factor VIII, a cofactor in the activation of factor X at the middle stages of the coagulation cascade. The gene coding for factor VIII is located in the subtelomeric Xq28 region, comprises 26 exons and spans 186 kb genomic DNA [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>].</p><p>By the present moment over 400 mutations leading to haemophilia A have been identified [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. De novo mutations in the factor VIII gene constitute a significant proportion of haemophilia A cases (app. 30% of all cases) [<xref ref-type="bibr" rid="B5">5</xref>]. Half of such mutations do not derive from a single germ cell but are attributed to a germline or somatic mosaic originating from a mutation during early embryogenesis [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B7">7</xref>].</p><p>In approximately 40–50% of the severe cases (25% of all cases) the underlying molecular defect causing haemophilia A is a gross rearrangement of the gene owing to intrachromosomal homologous recombination between inverted repeats located in intron 22. and outside the gene. Recombination leads to separation of exons 1. – 22. from exons 23. – 26. and positioning these in reverse orientation to the remaining part of the gene [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. Several types of inversion have been recognized. Most often encountered is inversion type I, which involves the distal copy of the repeated unit, but types II (involves the proximal copy) and III (when more than two extragenic copies of the repeated unit are present) are also common in haemophilia A patients [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>].</p><p>Transition mutations C→T (respectively, G→A) in CpG dinucleotides of factor VIII gene occur frequently [<xref ref-type="bibr" rid="B12">12</xref>]. Thus, screening for mutations abolishing or creating restriction sites in patients with severe haemophilia A can provide direct carrier detection. Screening for unknown mutations is also accomplished by various scanning methods (such as SSCA) and subsequent sequencing of fragments which show abnormal electrophoretic pattern [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>].</p><p>Large inversions can be detected by either Southern hybridisation or nested RT-PCR spanning the breakpoint in intron 22. analysis and thus carrier and/or antenatal diagnosis is accomplished [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. Because of the enormous variety of mutations producing the disease in the remaining 75% of cases (except for the large inversions), carrier identification and prenatal diagnosis can be accomplished by indirect detection using linked to the disease locus DNA polymorphisms. A number of RFLPs and tandem repeats have been identified within the genes of factor VIII and in its extragenic regions [<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B23">23</xref>].</p><p>Rapid carriership determination and prenatal diagnosis are usually facilitated by a compound diagnostical strategy consisting of screening for the large inversions combined with DNA analysis of polymorphisms within and outside the factor VIII gene. DNA analysis using polymorphic markers linked to disease locus proves to be highly effective, as the proportion of cases in which linkage analysis fails is less than 20% of the affected families [<xref ref-type="bibr" rid="B24">24</xref>-<xref ref-type="bibr" rid="B26">26</xref>].</p><p>In Bulgaria, the incidence of haemophilia A is estimated to be about 1:18 000 live births (1:9000 males) [<xref ref-type="bibr" rid="B27">27</xref>]. The mainstay of treatment for bleeding episodes in Bulgarian haemophiliacs is "treatment on demand", e. g. application of the deficient factor at the first symptom of bleeding. Considering the short half-life of factor VIII in plasma, most of the time the haemophiliacs in Bulgaria live and function far below therapeutic levels of deficient factor. This poses a constant risk of sudden hemorrhage that might be difficult to handle. Thus, despite recent marked improvement in treatment of bleeding disorders, for Bulgarian patients, haemophilia remains a severe, crippling, sometimes life-threatening disease.</p><p>Carriership of haemophilia invariably has influence on women's reproductive plans. It is documented elsewhere that approximately 50% of female relatives of haemophilia patients may decide consciously not to have children and relate their decision to fear of passing the disease to their children [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. DNA analysis greatly improves the effectiveness of early determination of carrier status.</p><p>The main goal of this study was to investigate the genetic heterogeneity of haemophilia A in Bulgaria in order to construct effective diagnostical strategy for prevention of recurrent disease in families already affected.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>DNA samples</title><p>Thirty-two haemophilia A families consisting of 162 members with 36 haemophilia A-affected males were included in the present study. Twenty-two of the families (69%) did not have previous history of haemophilia A. The distribution of clinical severity of haemophilia A was as follows: 23 patients (64%), severe (FVIII:C <1%); 8 patients (22%), moderate (FVIII:C 1 – 5%) and 5 patients (14%), mild (FVIII:C >5%).</p><p>Written informed consent was obtained from all haemophilia patients and their relatives that participated in the study. In the case of minors, informed consent for participation in genetic studies was obtained from their parents or legal guardians.</p><p>DNA was extracted from 5 ml peripheral blood or fetal material. For PCR purposes, DNA was prepared by proteinase K digestion/phenol extraction/ethanol precipitation as described in [<xref ref-type="bibr" rid="B30">30</xref>]. For genomic hybridisation, high molecular weight DNA was obtained by embedding leucocytes in low-melting agarose blocks and subsequent digestion by proteinase [<xref ref-type="bibr" rid="B30">30</xref>].</p><p>Fetal DNA was extracted from 3–10 mg chorionic villi obtained by chorionic villus sampling (CVS) at 10.–12. week of gestation (g. w.) or from 10–15 ml amniotic liquid obtained by amniocentesis at 16.–21. g. w., after informed consent.</p></sec><sec><title>Southern hybridisation for detection of factor VIII inversions</title><p>High molecular weight genomic DNA (3–6 μg) was digested with Bcl I, separated by agarose electrophoresis and blotted onto nylon membrane. DNA probe was obtained either by digestion of p542.16 with Eco RI/Sac I [<xref ref-type="bibr" rid="B9">9</xref>]; or by amplification of the desired region of the intron 22. using genomic DNA from healthy male volunteers [<xref ref-type="bibr" rid="B31">31</xref>]. PCR fragment was purified by low-melting point agarose elution. Filling-end reaction was performed with Klenow fragment, though the probe can be used without the ends – blunting step. The hybridisation probes were labeled to high specific activity by random priming. Blots were hybridized at 64° C in modified Church's buffer: (5 X SSC, 50 mM Na<sub>2</sub>HPO<sub>4</sub>; pH = 7,2; 10% dextran sulphate; 6% SDS and washed twice with 2 X SSC, 0,1% SDS at room temperature, and twice with 0,1 X SSC, 0,1% SDS at the hybridisation temperature [<xref ref-type="bibr" rid="B9">9</xref>]. Autoradiography was carried out at -70° C.</p></sec><sec><title>Analysis of DNA polymorphisms linked to Factor VIII gene locus (gene tracking)</title><sec><title>PCR analysis of tandem repeats</title><p>Tandem repeats in extragenic region of factor VIII gene (St14) and inside the gene (intron 13. (IVS13(CA)nSTR, IVS13STR), intron 22. (IVS22(CA)nSTR, IVS22STR), intron 25. (IVS25(CA)nSTR, IVS25STR) (Universitat fur Humangenetik-Muenster, Germany-personal communication) and intron 6. (IVS6(CA), IVS6STR) were used in the present study [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B23">23</xref>]. PCR products were separated in 6–10% nondenaturing (or sequencing) polyacrylamide gels and visualised by silver staining [<xref ref-type="bibr" rid="B32">32</xref>] or by autoradiography.</p></sec><sec><title>PCR-RFLP polymorphisms</title><p>Polymorphic sites in intron 18. (Bcl I RFLP); intron 19. (Hind III RFLP), intron 22. (Xba I RFLP) of the factor VIII gene were used in the present study [<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. PCR products were subjected to restriction with the relevant restriction endonuclease, separated in nondenaturing polyacrylamide gels and visualised by silver staining.</p></sec><sec><title>G/A polymorphism in intron 7. of the factor VIII gene</title><p>The region containing the polymorphic site was amplified according to [<xref ref-type="bibr" rid="B18">18</xref>]. Samples were denatured and electrophoresed on a 14% polyacrylamide gel (37,5:1 acrylamide:bisacrylamide) at 5° C for 20 h and visualised by silver staining.</p></sec></sec><sec><title>Screening for mutations in the essential regions of the factor VIII gene</title><sec><title>SSCA analysis</title><p>Exons 11., 14. and 24. of the gene of the Factor VIII gene were screened for mutations as described in [<xref ref-type="bibr" rid="B14">14</xref>]. PCR yielded 343 bp fragment for exon 24. and 445 bp fragment for exon 11., respectively. The large exon 14. was amplified as four overlapping fragments, app. 1000 bp long. The resulting amplification fragments were hydrolyzed with the Alu I. Digested samples were denatured and electrophoresed on a 14% polyacrylamide gel (37,5:1 acrylamide:bisacrylamide) at 8°C for 20 h and visualized by silver staining. Sequencing of the fragments was performed by the dideoxy chain-termination method, on AbiGene automatic sequencer (Applied Biosystems).</p></sec><sec><title>Screening for restriction site abolishing mutations in exons 1. and 18</title><p>Screening for Taq I site abolishing mutations in exon 1. and Taq I/Rsa I site abolishing mutations in exon 18. of the Factor VIII gene was performed according to [<xref ref-type="bibr" rid="B12">12</xref>]. Briefly, the PCR amplification yielded 186 bp fragment for exon 1. and 291 bp fragment for exon 18., respectively.</p><p>Normally, Taq I digestion of the PCR fragment of exon 1. results in fragments 165 and 21 bp. C→T transition in exon 1. abolishes the existing Taq I restriction site.</p><p>In exon 18. normally exists 1 restriction site for Taq I (digestion pattern of 163 and 128 bp) and 2 sites for Rsa I (digestion pattern of 207, 57 and 32 bp). Transition C→T results in abolishment of the Taq I site and one of the Rsa I sites I (digestion pattern of 234 and 57 bp).</p></sec></sec></sec><sec><title>Results and discussion</title><sec><title>Screening for most common mutations causing severe haemophilia A</title><p>Inversion mutation was found in 13 out of 29 severe and severe-to-moderate haemophilia A patients (45%, 0.95% CI: 28,4–62,5%). Ten of these carried the common type of inversion (type I), involving the distal extragenic copy of the repeated unit (77%), two patients (15%) – inversion, involving the proximal copy of the 9,5 kb repeat (type II) and 1 patient carried the rare type IIIB inversion.</p></sec><sec><title>Analysis of DNA polymorphisms linked to factor VIII gene locus</title><p>By genealogical data, 28 women (not including mothers of haemophilia A boys) were identified to be at risk to be carriers of haemophilia A.</p><p>Polymorphic analysis for carrier status determination of haemophilia A was successful in 30 families out of 32 (94%).</p><p>These families were informative by at least one marker locus. In 5 families (16%) only the extragenic DNA marker St14 was informative. Only 2 families (6%) were uninformative by all the markers used. Rates of heterozygocity for polymorphic markers used are shown in table <xref ref-type="table" rid="T1">1</xref>.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>The panel of factor VIII gene linked polymorphisms, used as markers for carriership determination and prenatal diagnostics of haemophilia A in Bulgaria and their local heterozygocity rates.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Marker</bold></td><td align="left"><bold>ST14 VNTR</bold></td><td align="left"><bold>IVS22 STR</bold></td><td align="left"><bold>IVS13 STR</bold></td><td align="left"><bold>BclI RFLP</bold></td><td align="left"><bold>HindIII RFLP</bold></td><td align="left"><bold>XbaI RFLP</bold></td><td align="left"><bold>IVS25 STR</bold></td><td align="left"><bold>Int7 G/A</bold></td><td align="left"><bold>IVS6 STR</bold></td></tr></thead><tbody><tr><td align="left"><bold>H</bold></td><td align="left">0.78</td><td align="left">0.63</td><td align="left">0.46</td><td align="left">0.35</td><td align="left">0.35</td><td align="left">n.a.</td><td align="left">0.25</td><td align="left">0.19</td><td align="left">0.02</td></tr></tbody></table><table-wrap-foot><p><bold>H</bold>: heterozygocity rate. (H = 1 - 931;p<sub>i</sub><sup>2</sup>); p<sub>i </sub>– frequency of allele i. **The sequence containing the Xba I polymorphic site is located in the 9,5 kb repeated region of intron 22. Thus, flanking primers amplify together with the Xba I polymorphic site a fragment in the extragenic copies that usually lack the Xba I restriction site and, practically, genotypes +/+ cannot be distinguished of genotype +/-. Thus, genotype of a female must be inferred of the genotype of her children [<xref ref-type="bibr" rid="B19">19</xref>].</p></table-wrap-foot></table-wrap><p>Among markers internal to factor VIII, intron 22 dinucleotide repeat showed highest heterozygocity rate (H) of 0,63, followed by intron 13 repeat (0,46), that renders these two very useful and reliable for diagnostic purposes in Bulgarian population. Intron 22 STR was informative in 19 of 32 families (60%, 0.95% CI: 42,2–74,5%). Intron 13 STR was informative in 12 of 32 families (38%, 0.95% CI: 22,9–54,8%).</p><p>Biallelic polymorphisms Hind III and Bcl I RFLP follow as markers of choice in the Bulgarian population as their heterozygocity is 0,35. This marker system was informative in 14 of 32 families (45%, 0.95 CI: 28,1–60,7%).</p><p>Intron 25 dinucleotide repeat was informative in 7 of 32 families (21%, 0.95 CI: 11,0–38,7%). This marker exhibited a rare allele, (150 bp long, a CA<sub>22 </sub>repeat) in the cohort of Bulgarian patients. This allele was observed only in one of the 65 independent X-chromosomes studied and is presumed to be quite rare (1,5%, 0.95 CI 0,27 – 8,22%). IVS25STR has limited applicability in most cases, because of its low heterozygocity rate (0,25). Nevertheless, as the gene for factor VIII is quite large and recombination within the limits of the gene itself is possible, when informative, IVS25STR may serve as an useful genetic marker in the distal part of the gene.</p><p>Intron 7 G/A SSCA polymorphism was informative in 6 of 32 families (19%, 0.95% CI: 8,9–35,3%). This marker exhibited in Bulgarian patients' heterozygocity rate even lower than the intron 25 dinucleotide repeat system (0,19). After typing with a panel of IVS22STR – HindIII/BclI system – Xba I RFLP – IVS25STR, intron 7 SSCA haplotype data does not add to the proportion of cases in which at least one marker is informative (Fig. <xref ref-type="fig" rid="F1">1</xref>).</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>A plot of cumulative informativeness of panel of markers linked to factor VIII locus used for polymorphic DNA analysis of haemophilia A in Bulgaria. By polymorphic DNA analysis of the factor VIII gene locus carrier status of females in 30 Bulgarian families out of 32 (94%) could be determined. Intron 22. and intron 13. STR dinucleotide repeats used together are informative in 64% of haemophilia A families. Adding biallelic Hind III/Bcl I RFLP polymorphic system and Xba I PFLP results in further increase in percentage of informative families up to 70%. IVS25STR marker is informative in a limited proportion of cases (increases informativeness up to 72% of the families). Intron 7. SSCA haplotype data does not add to the proportion of cases in which at least one marker is informative. St14 tandem repeat exhibits the highest heterozygocity rate of all factor VIII gene linked markers and thus proves to be the most informative marker in the Bulgarian population. St14 is a marker of choice in cases when none of the intragenic markers is informative, nevertheless, its location apart of factor VIII gene poses risk of misdiagnosis due to marker-to-gene recombination.</p></caption><graphic xlink:href="1471-2326-4-2-1"/></fig><p>IVS6STR dinucleotide repeat exhibits very low polymorphism in the Bulgarian haemophilic population.(heterozygocity rate of 0,02). We observed three alleles with 12–14 repetitions (CA)<sub>(12–14)</sub>. Single allele ((CA)<sub>13</sub>) accounts for 99% of all alleles. This renders the marker less useful for the diagnosis of haemophilia A carriers.</p><p>St14 VNTR, with its large number of alleles (10 observed in this study out of 15 described in [<xref ref-type="bibr" rid="B21">21</xref>]) proves to be the most informative marker in the Bulgarian population. It was informative in 29 of 32 families (91%, 0.95% CI: 75,8–96,8%) and has the highest heterozygocity rate (0,78) of all the polymorphic markers included in this study.</p><p>St14 may solve the diagnostic problem in cases when none of the intragenic markers is informative. As it is located approximately 2 map units apart from the factor VIII locus, recombination rate is estimated at 2% per meiosis. Rarely, recombination between marker locus and disease gene locus may occur. We have observed one such case (data not shown).</p></sec><sec><title>Screening for mutations in the essential regions of the factor VIII gene</title><sec><title>SSCA analysis</title><p>SSCA can detect over 95% of causative mutations in moderate and mild haemophilia A and about 50% of the defects that cause severe haemophilia A (as large rearrangements account for app. 50% of the severe cases) [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B33">33</xref>].</p><p>Exons 11., 14. and 24. of the gene of the Factor VIII gene in 34 males with haemophilia A were screened by SSCA and subsequent sequencing. No pattern different from the normal was found.</p></sec><sec><title>Restriction analysis</title><p>Screening for Taq I abolishing mutations in exon 1 and Taq I/Rsa I site abolishing mutations in exon 18. of the factor VIII gene was performed. Two related patients were found to carry a Taq I site abolishing mutation in exon I, resulting in CGA-to-TGA (Arg to Stop) transition in the vicinity of transcription startpoint of the factor VIII gene. This molecular defect resulted in these patients in severe phenotype with practically undetectable plasma levels of factor VIII.</p></sec></sec></sec><sec><title>Conclusions</title><p>Severe form of haemophilia A accounts for over 60% of Bulgarian patients in this study. As no founder effect pattern could be established, this biased pattern may be due to undereferral of moderate and mild haemophiliacs. This may reflect the fact that some milder cases may escape clinical attention. In fact, some researchers estimate the real incidence of haemophilia A [<xref ref-type="bibr" rid="B1">1</xref>] to be 1:5000 males, that is, almost twice as high as the current value.</p><p>Gross inversions of the factor VIII were discovered in 45% of Bulgarian cases of severe and severe-to moderate haemophilia A (36% of all). In the remaining severe cases as well as in the moderate and mild haemophilia A, the analysis of polymorphisms linked to the factor VIII gene proved as an invaluable tool in carrier status determination and prenatal diagnosis. In Bulgarian patients, combined use of analysis of DNA polymorphisms linked to factor VIII locus and screening for inversion mutation failed only in 2 cases (6%).</p><p>Fig. <xref ref-type="fig" rid="F1">1</xref>. presents a plot of cumulative informativeness of the markers included in this study vs. increasing number of markers used for haplotyping. Additional number of intragenic markers does not increase the percentage of informative families beyond 72%. Including the external marker St14 will raise the overall informativeness up to 94%, nevertheless, the risk of recombination between extragenic marker and disease locus may become significant depending on the number of meioses in question.</p><p>Carrier status was determined in 25 of a total of 28 women at risk (89%). One woman dropped out of study when she became pregnant. In 2 women (7%) DNA analysis gave no information about the type of mutation and tracking of the defective gene failed because of lack of heterozygocity of the polymorphic markers. In thirteen women carriership of haemophilia A was ruled out. In the remaining 12 women, carriership for haemophilia A was confirmed.</p><p>Fourteen prenatal diagnoses in women at high risk of having a haemophilia A – affected child were performed. These pregnancies produced 6 boys and 5 girls (two carriers, two noncarriers, and one where no further DNA analysis after fetal sex determination was carried out). Five of the boys were healthy. In one case with pregnancy with male fetus, a key recombination between extragenic and intragenic for Factor VIII gene polymorphic markers was discovered. As the haemophiliac male was the first and only case in this family and no DNA was available from him, mutation screening could not be accomplished and no definite conclusion could be drawn from the polymorphic DNA analysis. The parents decided not to terminate the pregnancy, as the mother was in 22. g. w. by the end of analysis.</p><p>In two cases, the fetus was found to be at high risk to be affected by haemophilia A. In both cases, parents invariably chose termination of the pregnancy.</p><p>In one case CVS produced insufficient amount of material for DNA analysis and biopsy was repeated, which led to obstetrical complications, and, finally, to loss of fetus.</p><p>Screening for gross inversions disrupting factor VIII gene combined with analysis of a panel of DNA markers linked to the disease locus proves to be highly informative and cost-effective strategy for prevention of recurrence of haemophilia A in Bulgaria. This compound approach facilitates carriership determination and subsequent prenatal diagnosis in the majority of Bulgarian haemophilia A families.</p></sec><sec><title>List of abbreviations used in the text</title><p><bold>H </bold>Heterozygocity rate</p><p><bold>PCR </bold>Polymerase chain reaction</p><p><bold>RFLP </bold>Restriction fragment length polymorphism</p><p><bold>SSCA </bold>Analysis of conformation single-stranded DNA fragments</p><p><bold>STR </bold>Short tandem repeat</p><p><bold>VNTR </bold>Variable number of tandem repeats</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>RP and SC carried out the DNA analysis studies. IK provided samples for analysis. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2326/4/2/prepub"/></p></sec>
High grade lymphoma in the nasopharynx presented as sudden onset of bilateral blindness	<sec><title>Background</title><p>Sudden onset of bilateral blindness is rare; hysteria, cortical infarction or bilateral central retinal arterial occlusion can cause this.</p></sec><sec><title>Case presentation</title><p>The authors describe a single case of sudden onset bilateral blindness in a patient with nasopharyngeal carcinoma, which is unusual. Biopsy revealed a high-grade lymphoma. After treatment the patient made a complete visual recovery, with no evidence of visual sequelae and no clear reasons for this complete recovery.</p></sec><sec><title>Conclusion</title><p>CT and MR imaging did not demonstrate any lesions invading any part of the visual pathway or even indeed the occipital cortex. High dose steroids may have reduced the mass effect of the tumour or the blindness may have been hysterical but is unlikely.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Shambhu</surname><given-names>S</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Vose</surname><given-names>M</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Ophthalmology	<sec><title>Background</title><p>Nasopharyngeal carcinoma is one of the most common cancers in eastern Asia. Usually presenting with a neck mass and blood-tinged sputum [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. Headache and cranial nerve abnormalities have also been reported [<xref ref-type="bibr" rid="B3">3</xref>], presumably from brainstem involvement. With advanced nasopharyngeal carcinoma the occurrence of distant metastases are high [<xref ref-type="bibr" rid="B4">4</xref>]. Typical sites include bone, liver and lung [<xref ref-type="bibr" rid="B5">5</xref>]. Brain metastasis is rare [<xref ref-type="bibr" rid="B6">6</xref>].</p></sec><sec><title>Case presentation</title><p>A 49-year-old female presented with an 18-hour history of complete visual loss. The patient denied perception of light. Six weeks prior to presentation the patient had complained of nasal congestion. This was treated with nasal steroid with no improvement. There were no other symptoms prior to her presentation.</p><p>On examination, both pupils were fixed and dilated. Direct and consensual reflexes were absent (Amaurotic pupils). Fundoscopy revealed a healthy retina with no papilloedama or vessel engorgement. Further ocular examination revealed roaming eye movements but no evidence of cranial nerve involvement. Medically, the patient was very distressed and could not cooperate to full examination. Urgent CT scan revealed a large nasopharyngeal mass (60 mm diameter) extending from the level of the base of the skull to approximately C2. The mass also extended into the right sphenoidal sinus, left maxillary sinus and bilateral nasal cavities.</p><p>A neoplastic lesion was suspected but with such widespread parasinus involvement, infection could not be excluded. The patient was given intravenous dexamethasone (16 mg) with 1.2 g Augmentin. Regular intravenous hydrocortisone (200 mg – 4 hourly) was commenced. Preparations were made for theatre and urgent drainage of the sinuses.</p><p>In theatre, a large, firm nasopharyngeal mass was noted in the left and right nasal cavities extending from the post nasal space to the posterior choanae. The mass was seen to extend down to the level of the uvula in the oral cavity. Biopsies were taken. There was no evidence of pus on sphenoidotomy.</p><p>On day one post-operatively, perception of light was denied. Both pupils were fixed and dilated. Day three post operatively the patients' visual acuity suddenly improved to 6/12 and the pupil reactions returned to normal. By day four, vision had recovered to 6/5 and she was able to read N5 print unaided. Goldmann visual field perimetry was completely unremarkable and the patient scored fully on Ishihara test plates.</p><p>Formal CT and MRI reports were obtained. Neither demonstrated optic nerve/chiasm compression by mass or associated fluid. The MRI report stated that several "high density areas in the occipital cortex would account for the patient's visual symptoms". Visual evoked potentials (VEP) were conducted and found to be unremarkable. However, there was no demonstrable field defect on full Goldmann visual field screening. Histology of the mass confirmed high-grade Non-Hodgkins lymphoma. Chemotherapy was commenced by the physicians and a cure was anticipated.</p><p>Choroid metastasis from nasopharyngeal carcinoma has previously been reported as a cause of visual disturbance [<xref ref-type="bibr" rid="B7">7</xref>]. There have been two case reports of visual loss being the initial presentation of nasopharyngeal carcinoma [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B9">9</xref>]. These cases were reported as unilateral visual loss. In either case, the CT scan demonstrated a space-occupying lesion compromising the orbital apex area and there was no improvement in vision, even after treatment [<xref ref-type="bibr" rid="B8">8</xref>]. A case of cortical blindness associated with nasopharyngeal carcinoma has been previously reported [<xref ref-type="bibr" rid="B10">10</xref>]. In this case, the patient had bilateral occipital lobe metastases causing permanent loss of vision. Blindness has also been reported as a result of chemotherapy in treating nasopharyngeal carcinoma [<xref ref-type="bibr" rid="B11">11</xref>].</p></sec><sec><title>Conclusion</title><p>The patient made a full visual recovery with no lasting visual field defect. This is testimony to the absence of optic pathway or cortical involvement and worthy of note. The mass effect of the tumour on the brain stem may explain the patient's fixed dilated pupils, but not her blindness. Invasion of the right sphenoidal sinus by the tumour may have caused a localized visual field defect, but this proved not to be so. The CT and MRI scans specifically reported no optic chiasm involvement from tumour or associated oedema. There was no evidence of occipital cortex infarct or oedema on either CT/MRI scans. The intensive high-dose steroids may have reduced the mass effect of the tumour/oedema. The "blindness" may well have been "hysterical" but in the context of the pupil reactions this seems unlikely.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>SS was involved in the initial drafting of the manuscript. MV reviewed and modified the manuscript prior to admission. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2415/4/2/prepub"/></p></sec>
Monoamine related functional gene variants and relationships to monoamine metabolite concentrations in CSF of healthy volunteers	<sec><title>Background</title><p>Concentrations of monoamine metabolites in human cerebrospinal fluid (CSF) have been used extensively as indirect estimates of monoamine turnover in the brain. CSF monoamine metabolite concentrations are partly determined by genetic influences.</p></sec><sec sec-type="methods"><title>Methods</title><p>We investigated possible relationships between DNA polymorphisms in the serotonin 2C receptor (<italic>HTR2C</italic>), the serotonin 3A receptor (<italic>HTR3A</italic>), the dopamine D<sub>4 </sub>receptor (<italic>DRD4</italic>), and the dopamine β-hydroxylase (<italic>DBH</italic>) genes and CSF concentrations of 5-hydroxyindolacetic acid (5-HIAA), homovanillic acid (HVA), and 3-methoxy-4-hydroxyphenylglycol (MHPG) in healthy volunteers (n = 90).</p></sec><sec><title>Results</title><p>The <italic>HTR3A </italic>178 C/T variant was associated with 5-HIAA levels (p = 0.02). The <italic>DBH</italic>-1021 heterozygote genotype was associated with 5-HIAA (p = 0.0005) and HVA (p = 0.009) concentrations. Neither the <italic>HTR2C </italic>Cys23Ser variant, nor the <italic>DRD4 </italic>-521 C/T variant were significantly associated with any of the monoamine metabolites.</p></sec><sec><title>Conclusions</title><p>The present results suggest that the <italic>HTR3A </italic>and <italic>DBH </italic>genes may participate in the regulation of dopamine and serotonin turnover rates in the central nervous system.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Jönsson</surname><given-names>Erik G</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Bah</surname><given-names>Jessica</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Melke</surname><given-names>Jonas</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Abou Jamra</surname><given-names>Rami</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Schumacher</surname><given-names>Johannes</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Westberg</surname><given-names>Lars</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Ivo</surname><given-names>Roland</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Cichon</surname><given-names>Sven</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A9" contrib-type="author"><name><surname>Propping</surname><given-names>Peter</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A10" contrib-type="author"><name><surname>Nöthen</surname><given-names>Markus M</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A11" contrib-type="author"><name><surname>Eriksson</surname><given-names>Elias</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A12" contrib-type="author"><name><surname>Sedvall</surname><given-names>Göran C</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Psychiatry	<sec><title>Background</title><p>Concentrations of the major serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA), the major dopamine metabolite homovanillic acid (HVA), and the major norepinephrine metabolite 3-methoxy-4-hydroxyphenylglycol (MHPG) in lumbar cerebrospinal fluid (CSF) have been used extensively as indirect measures of monoamine turnover in the brain of humans. Studies of human twins indicate that CSF 5-HIAA and HVA levels are under familial influence of both genetic and environmental origin, whereas MHPG is under major genetic influence [<xref ref-type="bibr" rid="B1">1</xref>]. In rhesus monkeys significant portions of CSF 5-HIAA, HVA, and MHPG in the central nervous system, have been shown to be determined by genetic mechanisms [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>A number of serotonin receptors mediate the effects of serotonin. Among several functions, the serotonin receptor 5-HT<sub>2C</sub>, which is densely expressed throughout the brain [<xref ref-type="bibr" rid="B3">3</xref>], seems to be directly involved in the regulation of serotonin and norepinephrine activities in the brain [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B6">6</xref>]. The 5-HT<sub>2C </sub>gene (<italic>HTR2C</italic>) is localised to chromosome Xq24 [<xref ref-type="bibr" rid="B7">7</xref>]. An <italic>HTR2C </italic>variant giving rise to a Cystein to Serine substitution at position 23 of the protein has been identified [<xref ref-type="bibr" rid="B8">8</xref>]. This variant was shown to influence the CSF MHPG concentration in a Finnish sample of predominantly alcoholic offenders [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>In contrast to all other serotonin receptors, which are G protein-coupled, the 5-HT<sub>3 </sub>receptor is a ligand-gated ion channel [<xref ref-type="bibr" rid="B10">10</xref>]. In the brain 5-HT<sub>3 </sub>receptors are localised in areas including the amygdala, hippocampus, and caudate nucleus. In addition to their effects on serotonin-regulated physiological processes, there are data suggesting that 5-HT<sub>3 </sub>receptors influence the activity of several other neurotransmitters, including norepinephrine and dopamine [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B13">13</xref>]. The 5-HT<sub>3A </sub>gene (<italic>HTR3A</italic>) is localised to chromosome 11q23.1-q23.2 [<xref ref-type="bibr" rid="B10">10</xref>]. An <italic>HTR3A </italic>single nucleotide polymorphism (178 C/T) in the upstream regulatory region was recently discovered to be of putative functional importance, because luciferase reporter assays in human embryonal kidney cells showed a two to three times higher activity of the rare allele compared to the wildtype [<xref ref-type="bibr" rid="B14">14</xref>]. This <italic>HTR3A </italic>variant was reported to be associated with bipolar disorder [<xref ref-type="bibr" rid="B14">14</xref>] and the personality trait harm avoidance in women [<xref ref-type="bibr" rid="B15">15</xref>].</p><p>The dopamine D<sub>4 </sub>receptor has a predominantly cortical localisation in the human brain [<xref ref-type="bibr" rid="B16">16</xref>]. Among several effects, the dopamine D<sub>4 </sub>receptor seems to modulate dopamine synthesis and turnover [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. The dopamine D<sub>4 </sub>receptor gene (<italic>DRD4</italic>) is located to chromosome 11p15.5 [<xref ref-type="bibr" rid="B19">19</xref>]. Recently, a putative functional <italic>DRD4 </italic>upstream region variant (-521C/T) was discovered, where the -521C allele was reported to be 40% less active than -521T allele in a chloramphenicol acetyltransferase assay using human retinoblastoma cells [<xref ref-type="bibr" rid="B20">20</xref>]. This <italic>DRD4 </italic>variant was associated with schizophrenia [<xref ref-type="bibr" rid="B20">20</xref>] as well as the personality trait novelty seeking in some [<xref ref-type="bibr" rid="B21">21</xref>-<xref ref-type="bibr" rid="B24">24</xref>] but not all studies [<xref ref-type="bibr" rid="B25">25</xref>-<xref ref-type="bibr" rid="B31">31</xref>]. However, meta-analyses suggested association with both conditions [<xref ref-type="bibr" rid="B32">32</xref>-<xref ref-type="bibr" rid="B34">34</xref>].</p><p>The enzyme dopamine β-hydroxylase (DβH) catalyses the conversion of dopamine to norepinephrine. DβH is localised in catecholamine-containing vesicles of noradrenergic and adrenergic cells [<xref ref-type="bibr" rid="B35">35</xref>,<xref ref-type="bibr" rid="B36">36</xref>]. DβH enzyme activity has been shown to be heritable to a great extent [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B37">37</xref>]. The DβH gene (<italic>DBH</italic>) is located on chromosome 9q34 [<xref ref-type="bibr" rid="B38">38</xref>]. Recently, a <italic>DBH </italic>promoter variant (-1021 C/T) was shown to strongly influence plasma DβH-activity [<xref ref-type="bibr" rid="B39">39</xref>-<xref ref-type="bibr" rid="B41">41</xref>], indicating a functional effect. In the present study we have examined the <italic>HTR2C </italic>Cys23Ser, <italic>HTR3A </italic>178 C/T, <italic>DRD4 </italic>-521 C/T and <italic>DBH </italic>-1021 C/T variants for possible relationships to concentrations of 5-HIAA, HVA, and MHPG in lumbar CSF from healthy Swedish volunteers.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Healthy human volunteers</title><p>The study was approved by the Ethics Committee of the Karolinska Hospital, Stockholm. Informed consent of the subjects was obtained after the nature of the procedures had been fully explained.</p><p>The characteristics and assessment of the subjects participating in the present study have been described previously [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. Caucasian individuals (n = 90) were recruited predominantly among students or hospital staff. Lumbar puncture (LP) was performed in all subjects. Height was also recorded. Back length, defined as the distance between the external occipital protuberance and the insertion point of the lumbar needle with the subject in the lying position, was measured in 63 subjects. Eight to 19 years later a structured interview was performed by a psychiatrist (EJ) to assess psychiatric morbidity (DSM-III-R; [<xref ref-type="bibr" rid="B44">44</xref>]), somatic illness and presence of mental and nervous system disorders among relatives. Subjects completed a questionnaire regarding smoking habits. Hospital records were obtained and examined for diagnosis. Genealogical data for antecedents up to the third degree were obtained from parish registers to assess the origin of the individuals. Subjects who reported any lifetime psychiatric disorder were excluded.</p><p>Of the 90 subjects 52 were men and 38 women. The age range at the time of the structured interview was 29 to 56, with a mean ± standard deviation of 40.5 ± 6.4 years. The mean age ± standard deviation at LP was 27.4 ± 5.9 years, age range 18 – 43. Thirty-six were university graduates. Twenty-one subjects had a family history of major mental illness defined as at least one first or second degree relative with schizophrenia, schizoaffective disorder, bipolar disorder, recurrent unipolar disorder, other non-organic psychosis, or who had committed suicide. Of the subjects 54 were or had been regular tobacco users, 25 were non-smokers or had only used tobacco once or a few times in their life, while data were missing for 11 individuals. Of the women, 15 used oral contraceptives at LP, 21 did not, while data were missing for two individuals. Except for oral contraceptives all participants were drug free at LP. Genealogical data implicated that 89.6% and 4.7% of the genes originated in ancestors born in Sweden and Finland, respectively, and the remaining 5.7% were distributed on 8 European countries.</p></sec><sec><title>CSF monoamine metabolite concentrations</title><p>All subjects had at least 8 h of bed-rest in the hospital, abstaining from food and smoking. CSF samples were obtained by LP between 8 and 9 a.m. with the subjects in the sitting (n = 41) or recumbent (n = 47) position. Samples of 12.5 ml CSF were drawn according to a standardised sampling procedure [<xref ref-type="bibr" rid="B45">45</xref>]. Samples were stored at below -20°C and analysed within two months. 5-HIAA, HVA, and MHPG concentrations were measured by mass fragmentography with deuterium labelled internal standards [<xref ref-type="bibr" rid="B46">46</xref>].</p></sec><sec><title>Genotype analyses</title><p>Venous blood was taken from all individuals into EDTA-containing tubes. DNA was isolated as previously described [<xref ref-type="bibr" rid="B47">47</xref>]. The <italic>HTR2C </italic>Cys23Ser variant was genotyped in accordance with Lappalainen et al [<xref ref-type="bibr" rid="B8">8</xref>]. The <italic>HTR3A </italic>178 C/T, <italic>DRD4 </italic>-521 C/T, and <italic>DBH </italic>-1021 C/T variants were genotyped as previously described [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B48">48</xref>].</p></sec><sec><title>Data analyses</title><p>One way analysis of variance (ANOVA) was used for comparisons between genotypes and 5-HIAA, HVA, and MHPG concentrations, respectively. To correct monoamine metabolite levels for back length and use of oral contraceptives (among women), suggestive but discussed confounding variables for monoamine metabolite concentrations in lumbar CSF [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>,<xref ref-type="bibr" rid="B49">49</xref>], analysis of covariance (ANCOVA) was used. For those subjects where back length was not available, estimated back length values, based on the relationship between back length and height, was used as previously described [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B43">43</xref>]. Significance level was defined as a p-value lower than 0.05. Power was estimated in accordance with published methods [<xref ref-type="bibr" rid="B50">50</xref>,<xref ref-type="bibr" rid="B51">51</xref>].</p></sec></sec><sec><title>Results</title><sec><title>Relationships between HTR2C genotypes and CSF monoamine metabolite concentrations</title><p>The <italic>HTR2C </italic>genotyping was successful in 86 individuals. Among men the allele frequencies were 0.88 (Cys23) and 0.12 (Ser23). In women the allele frequencies were 0.89 (Cys23) and 0.11 (Ser23), distributed on the following genotypes: Cys23Cys (81%), Cys23Ser (16%), and Ser23Ser (3%). As the <italic>HTR2C </italic>gene is localised on the X chromosome, each gender was analysed separately. Among women, the Ser23Ser and Cys23Ser genotypes were pooled and analysed versus the Cys23Cys genotype, because of the small number of Ser23Ser subjects. There were no significant relationship between genotypes and any of the CSF monoamine metabolite concentrations neither among men or women (table <xref ref-type="table" rid="T1">1</xref>).</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Serotonin receptor 5-HT<sub>2C </sub>(<italic>HTR2C</italic>) genotypes and relationships to monoamine metabolite concentrations in human lumbar cerebrospinal fluid.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="center" colspan="3">5-HIAA</td><td align="center" colspan="3">HVA</td><td align="center" colspan="3">MHPG</td></tr><tr><td></td><td></td><td></td><td colspan="9"><hr></hr></td></tr><tr><td align="center"><italic>HTR2C </italic>Allele/genotype</td><td align="center">Sex</td><td align="center">n</td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td></tr></thead><tbody><tr><td align="center">Cys23</td><td align="center">Men</td><td align="center">43</td><td align="center">92 ± 40</td><td align="center">F = 0.27</td><td align="center">F = 0.13</td><td align="center">166 ± 75</td><td align="center">F = 0.73</td><td align="center">F = 0.50</td><td align="center">43 ± 8</td><td align="center">F = 1.75</td><td align="center">F = 1.48</td></tr><tr><td align="center">Ser23</td><td></td><td align="center">6</td><td align="center">86 ± 16</td><td align="center">p = 0.61</td><td align="center">p = 0.72</td><td align="center">144 ± 39</td><td align="center">p = 0.40</td><td align="center">p = 0.48</td><td align="center">39 ± 4</td><td align="center">p = 0.19</td><td align="center">p = 0.23</td></tr><tr><td align="center">Cys23Cys</td><td align="center">Women</td><td align="center">30</td><td align="center">105 ± 39</td><td align="center">F = 0.66</td><td align="center">F = 0.57</td><td align="center">194 ± 77</td><td align="center">F = 2.14</td><td align="center">F = 1.34</td><td align="center">40 ± 6</td><td align="center">F = 1.03</td><td align="center">F = 0.91</td></tr><tr><td align="center">Cys23Ser<sup>c</sup></td><td></td><td align="center">6</td><td align="center">118 ± 29</td><td align="center">p = 0.42</td><td align="center">p = 0.46<sup>d</sup></td><td align="center">231 ± 69</td><td align="center">p = 0.15</td><td align="center">p = 0.25<sup>e</sup></td><td align="center">44 ± 8</td><td align="center">p = 0.32</td><td align="center">p = 0.35<sup>f</sup></td></tr><tr><td align="center">Ser23Ser<sup>c</sup></td><td></td><td align="center">1</td><td align="center">107</td><td></td><td></td><td align="center">229</td><td></td><td></td><td align="center">38</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p>5-HIAA = 5-hydroxyindoleacetic acid; HVA = homovanillic acid; MHPG = 3-methoxy-4-hydroxyphenylglycol. Statistical comparisons done on monoamine metabolite residuals corrected<sup>a </sup>and uncorrected<sup>b </sup>for back length. <sup>c </sup>Cys23Ser and Ser23Ser genotypes were combined in the analyses. <sup>d </sup>Correction for use of oral contraceptives, F = 0.21, p = 0.65. <sup>e </sup>Correction for use of oral contraceptives, F = 0.84, p = 0.37. <sup>f </sup>Correction for use of oral contraceptives, F = 0.22, p = 0.64.</p></table-wrap-foot></table-wrap></sec><sec><title>Relationships between HTR3A genotypes and CSF monoamine metabolite concentrations</title><p>The <italic>HTR3A </italic>178 C/C genotype was the most frequent (67%), followed by the 178 C/T (30%) and the 178 T/T (3%) genotypes. The allele frequencies were 0.82 (178C) and 0.18 (178T). The 178 T/T and 178 C/T genotypes were pooled in the calculations, because of the small number of subjects carrying the 178 T/T genotype. In the total sample there were associations between the <italic>HTR3A </italic>variant and 5-HIAA (p = 0.002) and HVA (p = 0.006) concentrations, with higher concentrations of these monoamine metabolites in carriers of the T-containing genotypes (table <xref ref-type="table" rid="T2">2</xref>). However, when corrected for back-length the association between the <italic>HTR3A </italic>variant and lumbar HVA concentrations was reduced to a trend (p = 0.08). In the male sub-sample, no significant relationships between <italic>HTR3A </italic>variation and 5-HIAA or HVA concentrations emerged. Among women the relationship between <italic>HTR3A </italic>variation and 5-HIAA concentrations, indicating higher 5-HIAA levels in subjects carrying the 178T allele, was significant both uncorrected (p = 0.004), corrected for back-length (p = 0.006), and corrected for use of oral contraceptives (p = 0.03; table <xref ref-type="table" rid="T2">2</xref>). However, in the female sub-sample the association between <italic>HTR3A </italic>and HVA concentrations was of borderline significance (p = 0.05 uncorrected, p = 0.03 corrected for use of oral contraceptives), but was non-significant after correction for back-length (p = 0.12). Inspection of the CSF levels of the different genotype groups indicated a possible heterosis effect with regard to the CSF 5-HIAA and HVA concentrations [<xref ref-type="bibr" rid="B52">52</xref>]. We therefore also performed calculations pooling the homozygotic genotypes. However, the probability levels of significance did not exceed those obtained pooling the T/T and C/T genotypes (table <xref ref-type="table" rid="T2">2</xref>). There were no significant relationships between the <italic>HTR3A </italic>genotype and MHPG concentrations (table <xref ref-type="table" rid="T2">2</xref>).</p><table-wrap position="float" id="T2"><label>Table 2</label><caption><p>Serotonin receptor 5-HT<sub>3A </sub>(<italic>HTR3A</italic>) genotypes and relationships to monoamine metabolite concentrations in human lumbar cerebrospinal fluid.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="center" colspan="3">5-HIAA</td><td align="center" colspan="3">HVA</td><td align="center" colspan="3">MHPG</td></tr><tr><td></td><td></td><td></td><td colspan="9"><hr></hr></td></tr><tr><td align="center"><italic>HTR3A </italic>Genotype</td><td align="center">Sex</td><td align="center">n</td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td></tr></thead><tbody><tr><td align="center">T/T<sup>c</sup></td><td align="center">All</td><td align="center">3</td><td align="center">105 ± 52</td><td align="center">F = 5.75</td><td align="center">F = 10.22</td><td align="center">203 ± 124</td><td align="center">F = 3.25</td><td align="center">F = 7.95</td><td align="center">44 ± 5</td><td align="center">F = 0.45</td><td align="center">F = 0.29</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">27</td><td align="center">117 ± 36</td><td align="center">p = 0.02</td><td align="center">p = 0.002</td><td align="center">211 ± 67</td><td align="center">p = 0.07</td><td align="center">p = 0.006</td><td align="center">43 ± 8</td><td align="center">p = 0.50</td><td align="center">p = 0.75</td></tr><tr><td align="center">C/C</td><td></td><td align="center">60</td><td align="center">90 ± 35</td><td></td><td></td><td align="center">166 ± 71</td><td></td><td></td><td align="center">42 ± 7</td><td></td><td></td></tr><tr><td align="center">T/T<sup>c</sup></td><td align="center">Men</td><td align="center">1</td><td align="center">52</td><td align="center">F = 0.31</td><td align="center">F = 0.96</td><td align="center">113</td><td align="center">F = 0.75</td><td align="center">F = 1.49</td><td align="center">49</td><td align="center">F = 0.20</td><td align="center">F = 0.49</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">11</td><td align="center">106 ± 33</td><td align="center">p = 0.58</td><td align="center">p = 0.33</td><td align="center">193 ± 50</td><td align="center">p = 0.39</td><td align="center">p = 0.23</td><td align="center">44 ± 9</td><td align="center">p = 0.66</td><td align="center">p = 0.49</td></tr><tr><td align="center">C/C</td><td></td><td align="center">40</td><td align="center">90 ± 37</td><td></td><td></td><td align="center">159 ± 11</td><td></td><td></td><td align="center">43 ± 7</td><td></td><td></td></tr><tr><td align="center">T/T<sup>c</sup></td><td align="center">Women</td><td align="center">2</td><td align="center">131 ± 34</td><td align="center">F = 8.55</td><td align="center">F = 9.75</td><td align="center">248 ± 137</td><td align="center">F = 2.49</td><td align="center">F = 4.05</td><td align="center">41 ± 3</td><td align="center">F = 0.68</td><td align="center">F = 1.00</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">16</td><td align="center">124 ± 37</td><td align="center">p = 0.006</td><td align="center">p = 0.004<sup>d</sup></td><td align="center">224 ± 76</td><td align="center">p = 0.12</td><td align="center">p = 0.05<sup>e</sup></td><td align="center">42 ± 7</td><td align="center">p = 0.42</td><td align="center">p = 0.32<sup>f</sup></td></tr><tr><td align="center">C/C</td><td></td><td align="center">20</td><td align="center">92 ± 30</td><td></td><td></td><td align="center">179 ± 64</td><td></td><td></td><td align="center">40 ± 6</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p>5-HIAA = 5-hydroxyindoleacetic acid; HVA = homovanillic acid; MHPG = 3-methoxy-4-hydroxyphenylglycol. Statistical comparisons done on monoamine metabolite residuals corrected<sup>a </sup>and uncorrected<sup>b </sup>for back length. <sup>c </sup>T/T and C/T genotypes were combined in the analyses. <sup>d </sup>Correction for use of oral contraceptives, F = 5.56, p = 0.02. <sup>e </sup>Correction for use of oral contraceptives, F = 1.76, p = 0.19. <sup>f </sup>Correction for use of oral contraceptives, F = 0.17, p = 0.68. Analysing heterosis, i.e. comparing homo- vs heterozygotes: All subjects 5-HIAA: F = 10.03, p = 0.002 (F = 5.71, p = 0.02 after correction for back length). All subjects HVA: F = 7.13, p = 0.009 (F = 2.28, p = 0.10). Men 5-HIAA: F = 1.96, p = 0.17 (F = 1.00, p = 0.32). Men HVA: 2.33, p = 0.13 (F = 1.40, p = 0.24). Women 5-HIAA: F = 6.73, p = 0.01 (F = 5.65, p = 0.02 and F = 3.76, p = 0.06 corrected for back length and use of oral contraceptives, respectively). Women HVA: F = 2.49, p = 0.12 (F = 1.03, p = 0.32 and F = 0.98, p = 0.33 corrected for back length and use of oral contraceptives, respectively).</p></table-wrap-foot></table-wrap></sec><sec><title>Relationships between DRD4 genotypes and CSF monoamine metabolite concentrations</title><p>The <italic>DRD4 </italic>-521 C/T genotype was the most frequent (58%), followed by the -521 T/T (28%) and the -521 C/C genotypes (14%). The allele frequencies were 0.57 (-521T) and 0.43 (-521C). When women were analysed separately, the -521 C/C and -521 C/T genotypes were pooled, because of the small number of -521 C/C subjects. There was no significant relationship between <italic>DRD4 </italic>genotypes and any of the CSF monoamine metabolite concentrations neither in the total sample, nor among men or women (table <xref ref-type="table" rid="T3">3</xref>).</p><table-wrap position="float" id="T3"><label>Table 3</label><caption><p>Dopamine D<sub>4 </sub>receptor (<italic>DRD4</italic>) genotypes and relationships to monoamine metabolite concentrations in human lumbar cerebrospinal fluid.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="center" colspan="3">5-HIAA</td><td align="center" colspan="3">HVA</td><td align="center" colspan="3">MHPG</td></tr><tr><td></td><td></td><td></td><td colspan="9"><hr></hr></td></tr><tr><td align="center"><italic>DRD4 </italic>Genotype</td><td align="center">Sex</td><td align="center">n</td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td></tr></thead><tbody><tr><td align="center">C/C</td><td align="center">All</td><td align="center">13</td><td align="center">104 ± 43</td><td align="center">F = 0.76</td><td align="center">F = 0.22</td><td align="center">199 ± 94</td><td align="center">F = 1.39</td><td align="center">F = 0.52</td><td align="center">42 ± 7</td><td align="center">F = 2.62</td><td align="center">F = 2.68</td></tr><tr><td align="center">C/T</td><td></td><td align="center">52</td><td align="center">99 ± 34</td><td align="center">p = 0.47</td><td align="center">p = 0.80</td><td align="center">176 ± 68</td><td align="center">p = 0.25</td><td align="center">p = 0.59</td><td align="center">41 ± 7</td><td align="center">p = 0.08</td><td align="center">p = 0.07</td></tr><tr><td align="center">T/T</td><td></td><td align="center">25</td><td align="center">96 ± 41</td><td></td><td></td><td align="center">181 ± 77</td><td></td><td></td><td align="center">45 ± 8</td><td></td><td></td></tr><tr><td align="center">C/C</td><td align="center">Men</td><td align="center">12</td><td align="center">104 ± 44</td><td align="center">F = 0.99</td><td align="center">F = 0.98</td><td align="center">197 ± 98</td><td align="center">F = 1.49</td><td align="center">F = 1.60</td><td align="center">43 ± 7</td><td align="center">F = 1.93</td><td align="center">F = 2.06</td></tr><tr><td align="center">C/T</td><td></td><td align="center">26</td><td align="center">91 ± 35</td><td align="center">p = 0.38</td><td align="center">p = 0.38</td><td align="center">156 ± 63</td><td align="center">p = 0.24</td><td align="center">p = 0.21</td><td align="center">41 ± 8</td><td align="center">p = 0.16</td><td align="center">p = 0.14</td></tr><tr><td align="center">T/T</td><td></td><td align="center">14</td><td align="center">84 ± 32</td><td></td><td></td><td align="center">155 ± 50</td><td></td><td></td><td align="center">46 ± 8</td><td></td><td></td></tr><tr><td align="center">C/C<sup>c</sup></td><td align="center">Women</td><td align="center">1</td><td align="center">107</td><td align="center">F = 0.004</td><td align="center">F = 0.11</td><td align="center">229</td><td align="center">F = 0.003</td><td align="center">F = 0.48</td><td align="center">38</td><td align="center">F = 0.78</td><td align="center">F = 1.23</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">26</td><td align="center">106 ± 33</td><td align="center">p = 0.95</td><td align="center">p = 0.74<sup>d</sup></td><td align="center">195 ± 68</td><td align="center">p = 0.96</td><td align="center">p = 0.49<sup>e</sup></td><td align="center">40 ± 6</td><td align="center">p = 0.38</td><td align="center">p = 0.27<sup>f</sup></td></tr><tr><td align="center">T/T</td><td></td><td align="center">11</td><td align="center">111 ± 47</td><td></td><td></td><td align="center">214 ± 93</td><td></td><td></td><td align="center">43 ± 7</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p>5-HIAA = 5-hydroxyindoleacetic acid; HVA = homovanillic acid; MHPG = 3-methoxy-4-hydroxyphenylglycol. Statistical comparisons done on monoamine metabolite residuals corrected<sup>a </sup>and uncorrected<sup>b </sup>for back length. <sup>c </sup>C/C and C/T genotypes were combined in the analyses. <sup>d </sup>Correction for use of oral contraceptives, F = 0.38, p = 0.54. <sup>e </sup>Correction for use of oral contraceptives, F = 0.04, p = 0.84. <sup>f </sup>Correction for use of oral contraceptives, F = 1.09, p = 0.30.</p></table-wrap-foot></table-wrap></sec><sec><title>Relationships between DBH genotypes and CSF monoamine metabolite concentrations</title><p>The <italic>DBH </italic>-1021 C/C genotype was the most frequent (71%), followed by the -1021 C/T (26%) and the -1021 T/T genotypes (3%). The allele frequencies were 0.84 (-1021C) and 0.16 (-1021T). The -1021 T/T and -1021 C/T genotypes were pooled in the calculations, because of the small number of subjects with the -1021 T/T genotype. In the total sample there was an association between <italic>DBH </italic>genotype and 5-HIAA concentrations (p = 0.01 uncorrected, p = 0.003 when corrected for back-length), with higher 5-HIAA levels in subjects with the -1021T containing genotypes (table <xref ref-type="table" rid="T4">4</xref>). This relationship reached significance also among men (p = 0.06 uncorrected, p = 0.04 corrected for back length), but not among women (table <xref ref-type="table" rid="T4">4</xref>). In the total sample there was a tendency for association (p = 0.09) between the <italic>DBH </italic>variant and HVA concentrations, with higher HVA levels in -1021T carriers (table <xref ref-type="table" rid="T4">4</xref>). When corrected for back-length the strength of this difference was significant (p = 0.03). Neither in the smaller male or female sub-samples this difference reached significance (table <xref ref-type="table" rid="T4">4</xref>).</p><table-wrap position="float" id="T4"><label>Table 4</label><caption><p>Dopamine β-hydroxylase (<italic>DBH</italic>) genotypes and relationships to monoamine metabolite concentrations in human lumbar cerebrospinal fluid.</p></caption><table frame="hsides" rules="groups"><thead><tr><td></td><td></td><td></td><td align="center" colspan="3">5-HIAA</td><td align="center" colspan="3">HVA</td><td align="center" colspan="3">MHPG</td></tr><tr><td></td><td></td><td></td><td colspan="9"><hr></hr></td></tr><tr><td align="center"><italic>DBH </italic>Genotype</td><td align="center">Sex</td><td align="center">n</td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td><td align="center">Mean ± SD (nmol/L)</td><td align="center">F<sup>a </sup>p<sup>a</sup></td><td align="center">F<sup>b </sup>p<sup>b</sup></td></tr></thead><tbody><tr><td align="center">C/C</td><td align="center">All</td><td align="center">64</td><td align="center">92 ± 35</td><td align="center">F = 9.07</td><td align="center">F = 6.42</td><td align="center">172 ± 66</td><td align="center">F = 5.21</td><td align="center">F = 3.00</td><td align="center">42 ± 7</td><td align="center">F = 0.01</td><td align="center">F = 0.01</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">23</td><td align="center">118 ± 39</td><td align="center">p = 0.003</td><td align="center">p = 0.01</td><td align="center">208 ± 93</td><td align="center">p = 0.02</td><td align="center">p = 0.09</td><td align="center">42 ± 9</td><td align="center">p = 0.90</td><td align="center">p = 0.93</td></tr><tr><td align="center">T/T<sup>c</sup></td><td></td><td align="center">3</td><td align="center">78 ± 25</td><td></td><td></td><td align="center">151 ± 18</td><td></td><td></td><td align="center">40 ± 5</td><td></td><td></td></tr><tr><td align="center">C/C</td><td align="center">Men</td><td align="center">36</td><td align="center">86 ± 36</td><td align="center">F = 4.68</td><td align="center">F = 3.80</td><td align="center">156 ± 56</td><td align="center">F = 2.22</td><td align="center">F = 1.81</td><td align="center">43 ± 7</td><td align="center">F = 0.08</td><td align="center">F = 0.04</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">15</td><td align="center">107 ± 35</td><td align="center">p = 0.04</td><td align="center">p = 0.06</td><td align="center">186 ± 98</td><td align="center">p = 0.14</td><td align="center">p = 0.18</td><td align="center">43 ± 10</td><td align="center">p = 0.78</td><td align="center">p = 0.84</td></tr><tr><td align="center">T/T<sup>c</sup></td><td></td><td align="center">1</td><td align="center">95</td><td></td><td></td><td align="center">160</td><td></td><td></td><td align="center">45</td><td></td><td></td></tr><tr><td align="center">C/C</td><td align="center">Women</td><td align="center">28</td><td align="center">101 ± 31</td><td align="center">F = 4.08</td><td align="center">F = 3.37</td><td align="center">192 ± 73</td><td align="center">F = 3.72</td><td align="center">F = 1.79</td><td align="center">41 ± 6</td><td align="center">F = 0.03</td><td align="center">F = 0.08</td></tr><tr><td align="center">C/T<sup>c</sup></td><td></td><td align="center">8</td><td align="center">139 ± 38</td><td align="center">p = 0.05</td><td align="center">p = 0.07<sup>e</sup></td><td align="center">249 ± 72</td><td align="center">p = 0.06</td><td align="center">p = 0.19<sup>d</sup></td><td align="center">41 ± 7</td><td align="center">p = 0.87</td><td align="center">p = 0.78<sup>f</sup></td></tr><tr><td align="center">T/T<sup>c</sup></td><td></td><td align="center">2</td><td align="center">70 ± 30</td><td></td><td></td><td align="center">147 ± 23</td><td></td><td></td><td align="center">38 ± 4</td><td></td><td></td></tr></tbody></table><table-wrap-foot><p>5-HIAA = 5-hydroxyindoleacetic acid; HVA = homovanillic acid; MHPG = 3-methoxy-4-hydroxyphenylglycol. Statistical comparisons done on monoamine metabolite residuals corrected<sup>a </sup>and uncorrected<sup>b </sup>for back length. <sup>c </sup>T/T and C/T genotypes were combined in the analyses. <sup>d </sup>Correction for use of oral contraceptives, F = 0.88, p = 0.36. <sup>e </sup>Correction for use of oral contraceptives, F = 0.46, p = 0.50. <sup>f </sup>Correction for use of oral contraceptives, F = 0.003, p = 0.96. Analysing heterosis, i.e. comparing homo- vs heterozygotes: All subjects 5-HIAA: F = 9.56, p = 0.003 (F = 13.07, p = 0.0005 after correction for back length). All subjects HVA: F = 4.43, p = 0.04 (F = 7.24, p = 0.009). Men 5-HIAA: F = 3.85, p = 0.06 (F = 4.6187, p = 0.04). Men HVA: 1.95, p = 0.17 (F = 2.31, p = 0.14). Women 5-HIAA: F = 9.14, p = 0.005 (F = 9.74, p = 0.004 and F = 4.11, p = 0.05 corrected for back length and use of oral contraceptives, respectively). Women HVA: F = 4.43, p = 0.04 (F = 6.42, p = 0.02 and F = 1.81, p = 0.19 corrected for back length and use of oral contraceptives, respectively).</p></table-wrap-foot></table-wrap><p>We also performed calculations pooling the homozygotic genotypes, i.e. analysing possible heterosis [<xref ref-type="bibr" rid="B52">52</xref>]. In the total sample there was an association between <italic>DBH </italic>heterozygosity and 5-HIAA concentrations (p = 0.003 uncorrected, p = 0.0005 when corrected for back-length) (table <xref ref-type="table" rid="T4">4</xref>). This relationship reached significance also among men (p = 0.06 uncorrected, p = 0.04 corrected for back length) and women (p = 0.005 uncorrected, p = 0.02 corrected for back length, p = 0.05 corrected for use of oral contraceptives; table <xref ref-type="table" rid="T4">4</xref>). In the total sample there was an association (p = 0.04 uncorrected, p = 0.009 corrected for back-length) between <italic>DBH </italic>heterozygosity and HVA concentrations (table <xref ref-type="table" rid="T4">4</xref>). This association failed to obtain significance in the smaller male sub-sample. Among women there was an association (p = 0.04 uncorrected, p = 0.02 corrected for back-length), which however did not survived correction for use of oral contraceptives (p = 0.18; table <xref ref-type="table" rid="T4">4</xref>).</p><p>There were no significant relationships between the <italic>DBH </italic>genotype and MHPG concentrations (table <xref ref-type="table" rid="T4">4</xref>).</p><p>Given α = 0.05 the present study had a power of 0.93 – 0.96 (total sample), 0.67 – 0.81 (men), or 0.53 – 0.67 (women) to detect differences of a large effect size (f = 0.40). For differences of a medium effect size (f = 0.25) the power was 0.54 – 0.65 (total sample) or less.</p></sec></sec><sec><title>Discussion</title><p>The present study is, to our knowledge, the first to investigate three potentially functional candidate gene variants (<italic>HTR3A </italic>178 C/T, <italic>DRD4 </italic>-521 C/T, <italic>DBH </italic>-1021 C/T) in the context of monoamine metabolite concentrations in cerebrospinal fluid from human healthy volunteers. Association was detected between two of these gene variants (<italic>HTR3A </italic>178 C/T, <italic>DBH </italic>-1021 C/T) and the indirect measures of monoamine activity in the brain.</p><p>However, we were not able to replicate the previously reported finding of higher CSF MHPG concentrations in <italic>HTR2C </italic>Ser23 compared to Cys23 carriers among men [<xref ref-type="bibr" rid="B9">9</xref>]. On the contrary, our male sub-sample showed a non-significant relationship in the opposite direction, i.e. higher MHPG concentrations in Cys23 subjects (table <xref ref-type="table" rid="T1">1</xref>). When we added another 23 men, excluded from the main analysis because of reported life-time psychiatric disorder, mostly alcohol abuse, depressive or anxiety disorders, this difference reached nominal significance (F = 4.44, d.f. = 71, p = 0.04), a result robust for correction for back-length (p = 0.04) and presence of life-time psychiatric disorder (p = 0.04), respectively (data not shown). Reasons for the different results in the two studies may include the different selection of subjects. The previous study included 73% alcoholic violent offenders and 27% healthy controls of Finnish ethnicity, and the <italic>HTR2C </italic>genotype effect was most prominent among the offenders [<xref ref-type="bibr" rid="B9">9</xref>], whereas the present study only included healthy subjects. The present study may lack power to detect the previous relationship. Alternatively, assuming different relationships between alcoholic violent offenders and healthy controls, the Finnish study may be under-powered concerning control subjects. It is also possible that both results are valid, but the results reflect an association to a linked variant, and that the degree of linkage between the <italic>HTR2C </italic>Cys23Ser variant and the 'real' functional polymorphism differs between the two populations investigated. There may also be a difference between the two populations with regard to other genes interacting with the present to influence MHPG concentrations. It is also possible that the Finnish, the present or both results have emerged by chance.</p><p>Subjects carrying the rarer <italic>HTR3A </italic>178T allele, which has been associated with higher protein expression than the wild-type variant [<xref ref-type="bibr" rid="B14">14</xref>], displayed higher lumbar CSF 5-HIAA concentrations. This suggests that a more efficient variant of the 5-HT<sub>3 </sub>receptor, involved in the regulation of serotonin activities, enhances brain serotonin turnover, giving rise to higher levels of the serotonin degradation product in CSF.</p><p>We were not able to find any significant relationships between <italic>DRD4 </italic>-521 C/T variation and CSF monoamine metabolite concentrations. This is in accordance with previous studies, analysing a <italic>DRD4 </italic>exon 3 variable number of tandem repeat variant [<xref ref-type="bibr" rid="B42">42</xref>,<xref ref-type="bibr" rid="B53">53</xref>]. In the present report there was a trend for an association between the <italic>DRD4 </italic>-521 C/T genotypes and CSF MHPG concentrations. This may mean that the present study does not have sufficient power to detect such a relationship, or that this trend reflects a tendency to a false positive finding. The results so far obtained suggest that the <italic>DRD4 </italic>gene does not have a large impact on the monoamine turnover in the brain as reflected by the major degradation products of these compounds in healthy human subjects.</p><p>One would expect that functional variants of the gene encoding the dopamine β-hydroxylase would primarily affect the catecholamines, in particular norepinephrine. However, in the present study the strongest relationship emerged between the <italic>DBH </italic>-1021 C/T variant and CSF 5-HIAA levels. Complex interactions between the noradrenergic and serotonergic systems have been reported [<xref ref-type="bibr" rid="B54">54</xref>,<xref ref-type="bibr" rid="B55">55</xref>]. Altered noradrenergic activity may alter the firing activity of serotonergic neurons, leaving a possibility for a decreased or increased availability of norepinehprine to be involved in these interactions. One might speculate that a more effective <italic>DBH </italic>variant, giving rise to an enhanced norepinephrine formation, facilitates noradrenergic activity, which in its turn facilitates serotonergic activity, giving rise to larger amounts of the serotonin degradation product 5-HIAA. There was also an association between <italic>DBH </italic>-1021 C/T variation and the major dopamine degradation product, indicating higher HVA levels in subjects with a less effective enzyme variant. This is in accordance with the theory that a less effective conversion of dopamine to norepinephrine would lead to higher amounts of dopamine, and in turn to its degradation product HVA. However, the stronger associations between heterozygotic genotypes and 5-HIAA and HVA concentrations, examples of positive heterosis [<xref ref-type="bibr" rid="B52">52</xref>], rather indicate a more complex physiology including interactions based on hidden stratification of unknown factors or heterozygotic advantage [<xref ref-type="bibr" rid="B52">52</xref>]. In this context, an interaction with e.g. the monoamine oxidase A gene, where more effective variants have been reported to increase 5-HIAA concentrations [<xref ref-type="bibr" rid="B43">43</xref>,<xref ref-type="bibr" rid="B56">56</xref>], may be a possibility. However, these latter results further complicate the task, as they call into question the use of monoamine metabolites as straightforward, although indirect, measures of brain turnover. This may mean that high levels of 5-HIAA may reflect a more effective degradation process rather than an enhanced overall turnover giving a possibility for reduced serotonin transmission to be associated with high levels of 5-HIAA.</p><p>It is possible that the present associations may have emerged by chance. Applying Bonferroni's correction would give a p-value of < 0.0011 (0.05/45) to be considered significant. Only one of the reported relationships, i.e. the association between <italic>DBH </italic>heterozygosity and CSF 5-HIAA levels, would survive such a correction procedure. On the other hand, although relatively large to constitute a sample of healthy subjects investigated by a demanding procedure, i.e. lumbar puncture, the present sample is small from a statistical point of view. The power of the present study was adequate to detect differences of large, but not medium to small effect sizes. Thus, it cannot be excluded that relationships of smaller magnitudes may have escaped our analysis attempts. Applying strict corrections for multiple testing would make investigations like the present impossible to perform, because a sample big enough to withstand such a correction procedure would probably never be possible to obtain. This is especially true taking the detection of small effects into account.</p></sec><sec><title>Conclusions</title><p>If replicated, the present results suggests that the <italic>HTR3A </italic>and <italic>DBH </italic>variants participate differentially in the regulation of serotonin turnover in the central nervous system of human subjects. It is also suggested that the <italic>DBH </italic>variant differentially influence dopamine turnover in the brain. The results give some support for an influence of the <italic>HTR2C </italic>variant on norepinephrine turnover in men, but do not favour a major differential influence of <italic>DRD4 </italic>gene activities on monoamine metabolite concentrations in lumbar CSF.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>EGJ performed the second clinical investigation, the statistical analyses, and drafted the manuscript. JB, JM, RAJ, JS, LW, and RI performed the genotyping. SC, PP, MMN and EE participated in the design and co-ordination of the study. GCS conceived of the study and participated in its design and co-ordination. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-244X/4/4/prepub"/></p></sec>
S-Adenosyl methionine (SAMe) versus celecoxib for the treatment of osteoarthritis symptoms: A double-blind cross-over trial. [ISRCTN36233495]	<sec><title>Background</title><p>S-Adenosylmethionine (SAMe) is a dietary supplement used in the management of osteoarthritis (OA) symptoms. Studies evaluating SAMe in the management of OA have been limited to Non Steroidal Anti-inflammatory Drugs (NSAIDs) for comparison. The present study compares the effectiveness of SAMe to a cyclooxygenase-2 (COX-2) inhibitor (celecoxib) for pain control, functional improvement and to decrease side effects in people with osteoarthritis of the knee.</p></sec><sec sec-type="methods"><title>Methods</title><p>A randomized double-blind cross-over study, comparing SAMe (1200 mg) with celecoxib (Celebrex 200 mg) for 16 weeks to reduce pain associated with OA of the knee. Sixty-one adults diagnosed with OA of the knee were enrolled and 56 completed the study. Subjects were tested for pain, functional health, mood status, isometric joint function tests, and side effects.</p></sec><sec><title>Results</title><p>On the first month of Phase 1, celecoxib showed significantly more reduction in pain than SAMe (p = 0.024). By the second month of Phase 1, there was no significant difference between both groups (p < 0.01). The duration of treatment and the interaction of duration with type of treatment were statistically significant (ps ≤ 0.029). On most functional health measures both groups showed a notable improvement from baseline, however no significant difference between SAMe and celecoxib was observed. Isometric joint function tests appeared to be steadily improving over the entire study period regardless of treatment.</p></sec><sec><title>Conclusion</title><p>SAMe has a slower onset of action but is as effective as celecoxib in the management of symptoms of knee osteoarthritis. Longer studies are needed to evaluate the long-term effectiveness of SAMe and the optimal dose to be used.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Najm</surname><given-names>Wadie I</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Reinsch</surname><given-names>Sibylle</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Hoehler</surname><given-names>Fred</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Tobis</surname><given-names>Jerome S</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Harvey</surname><given-names>Phillip W</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Musculoskeletal Disorders	<sec><title>Background</title><p>Arthritis is one of the most common chronic medical conditions encountered with older age, and osteoarthritis (OA) is the most common type of arthritis, and a major cause of disability. OA of the knee is more common than any other joint. Symptoms include pain, stiffness, and decreased range of motion. Subjects often become limited in their activities and quality of life, due to pain.</p><p>To manage the symptoms of OA, subjects and healthcare providers often resort to multiple approaches, which include lifestyle modifications, medications, exercise or surgery. Non-steroidal anti-inflammatory drugs (NSAIDs) are the mainstay medications used to manage osteoarthritis pain. NSAIDs prevent inflammation and control pain by blocking cyclooxygenase (COX-1 and COX-2 enzymes) but their side effect profile is not always acceptable, particularly among the elderly. Each year more than 100,000 subjects are hospitalized for gastrointestinal complications of NSAIDs, and of those approximately 15% die from these complications [<xref ref-type="bibr" rid="B1">1</xref>].</p><p>Newer NSAIDs, known as COX-2-inhibitors such as valdecoxib (Bextra), celecoxib (Celebrex), and rofecoxib (Vioxx), were introduced recently as providing similar anti-inflammatory activity and pain relief but fewer gastrointestinal side effects, including a reduced risk of GI erosion and bleeding. However, recent evidence does not always support these claims, and a few studies have suggested a possible increased risk of heart disease [<xref ref-type="bibr" rid="B2">2</xref>]. In addition, subjects often lose the benefit from these medications once they are discontinued; however, OA is a progressive disease and as such requires continuous management.</p><p>Over the last decade many dietary supplements were introduced to the public for the management of a variety of conditions including arthritis. Advertisement and information about supplements targeting people with arthritis can be found everywhere. Acupuncture, massage, magnet therapy, and dietary supplements have been promoted for the management of arthritis.</p><p>Given the limitations of the established osteoarthritis medications/treatments, and the recent explosion of information and interest in complementary and alternative medicine (CAM), the public has turned their attention to CAM and is exploring safer alternatives to manage their symptoms. The recent American College of Rheumatology (ACR) guidelines for treating OA included dietary supplements, such as glucosamine sulfate, chondroitin sulfate, and antioxidants, as well as acupuncture and magnets as therapies under investigation [<xref ref-type="bibr" rid="B3">3</xref>]. While their final judgment of their effectiveness is yet to be scientifically proven, these therapies do seem to have the advantage of showing, possible benefit with fewer side effects. S-Adenosylmethionine (SAMe) is one of the dietary supplements that gained popularity, and was recently reported to be effective in the management of depression, liver disease and arthritis.</p><p>SAMe was first discovered in Italy in 1952 [<xref ref-type="bibr" rid="B4">4</xref>]. Soon after, it became popular in Europe and more recently in the U.S. SAMe is an important physiologic compound that is distributed throughout the body tissues and fluids. It is produced endogenously from methionine and adenosine triphosphate (ATP). It is an important methyl group donor playing an essential role in many biochemical reactions involving enzymatic transmethylation. Those play an important role in the biosynthesis of phospholipids that are important for the integrity of cell membranes. Despite our basic understanding of its role, the exact mechanism of action in different disease conditions is not well established. Oral SAMe achieves peak plasma concentrations (0.5 – 1 mg/L) 3 to 5 hours after ingestion of an enteric-coated tablet (400 – 1000 mg). The half-life is about 100 minutes, and it is excreted in urine and feces [<xref ref-type="bibr" rid="B5">5</xref>].</p><p>Biochemically, SAMe is involved in three main metabolic pathways: 1) methylation, as the principal source of methyl groups in the body; 2) transsulfuration, SAMe forms S-Adenosylhomocysteine (SAH) and then converted to homocysteine (Hcy) which can be converted to cystathionine then to cysteine and the sulfate (SO<sub>4</sub>) donated to other metabolic intermediates; and 3) aminopropylation, as SAMe plays an important role in the synthesis of polyamines which can eventually form and recycle methionine [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>The exact mechanism of SAMe in reducing pain of OA is not known, but evidence suggests that it may play a role in reducing inflammation, increasing proteoglycan synthesis or having an analgesic effect. Whether SAMe is a COX-2 inhibitor is also not known. In vitro studies using human articular chondrocytes have shown SAMe-induced increases in proteoglycan synthesis [<xref ref-type="bibr" rid="B7">7</xref>] and proliferation rates in rabbits [<xref ref-type="bibr" rid="B8">8</xref>]. SAMe may reduce inflammatory mediators thus reducing pain. This was noted in other studies with the reduction of Tumor Necrosis Factor (TNF)-alpha and fibronectin RNA expression using cultured rabbit synovial cells [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Initial studies with SAMe used the parenteral route exclusively due to the instability of the oral form. As additional work allowed the development of a stable oral form of SAMe, further studies tested the effectiveness of the oral form in the management of several medical conditions including osteoarthritis. The length of treatment differed in each study ranging from 7 to 84 days for testing oral preparations, 5 days for intravenous and 7 days for intramuscular preparations. Study designs usually implemented a parallel group design and two studies used a cross-over design [<xref ref-type="bibr" rid="B10">10</xref>][<xref ref-type="bibr" rid="B11">11</xref>]. Only one cross-over study used an oral preparation with a 5-day washout period [<xref ref-type="bibr" rid="B10">10</xref>]. The majority of studies compared SAMe to NSAIDs. Most studies used either 600 mg or 1200 mg of SAMe per day. No specific explanations were provided to justify the dose used. The overall conclusions of these studies pointed toward an effect size in favor of SAMe compared to placebo and equivalent to NSAIDs in reducing pain and improving functioning.</p><p>Three reviews analyzed available studies to date [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B13">13</xref>][<xref ref-type="bibr" rid="B14">14</xref>]. One is a meta-analysis [<xref ref-type="bibr" rid="B13">13</xref>], the other two provide a general overview of the articles. All three reviews reach similar conclusions that the results of available studies are heterogeneous and do not allow for a firm conclusion about the effectiveness of SAMe in the management of OA. However, despite this limitation the consensus is that SAMe appears to be of equivalent effectiveness to NSAIDs in reducing pain and improving functional limitations, with fewer side effects. Di Padova (1987) [<xref ref-type="bibr" rid="B15">15</xref>]concluded that SAMe had a slower onset of action than NSAIDs, with equivalent results at 4 weeks.</p><p>Studies were frequently limited though, in quality of design and implementation. Most studies were done during the 1980's, with one study [<xref ref-type="bibr" rid="B16">16</xref>] in 1994. Since at that time the majority of the studies compared SAMe to placebo or NSAIDs, there are presently no studies available that evaluate the effectiveness or safety of SAMe versus the new class of drugs used for OA, COX-2 inhibitors. The objective of the present study is to assess the therapeutic effectiveness of oral SAMe (S-adenosylmethionine), in comparison to celecoxib (Celebrex<sup>®</sup>), to relieve pain and improve function in adults with OA of the knee.</p></sec><sec sec-type="methods"><title>Methods</title><p>The present study is a prospective randomized double blind cross-over design, of subjects with knee osteoarthritis. We compared the effectiveness of 1200 mg SAMe (600 mg twice daily) with 200 mg celecoxib (Celebrex 100 mg twice daily), to reduce pain and improve functional symptoms associated with OA of the knee. Each phase of the study was for a total of 8 weeks with one-week washout between the two phases.</p><sec><title>Setting</title><p>The study was conducted at the General Clinical Research Center (GCRC) in the University of California, Irvine, Medical Center (UCIMC) from September 2000 to September 2002.</p><sec><title>Selection criteria</title><p>Inclusion criteria consisted of a) adults age 40 or older, diagnosed with OA of the knee, based on the American College of Rheumatology criteria for the classification of OA of the knee, i.e. knee pain morning stiffness up to 30 minutes, and crepitus on motion [<xref ref-type="bibr" rid="B17">17</xref>], and b) agreement to participate in the study protocol as described in the informed consent form. Two-view radiographs of the knee that was most affected by OA were obtained. Exclusion criteria consisted of any history of adverse reaction to the study drugs (SAMe, celecoxib or sulfa drugs), current pregnancy status, active infection, blood coagulopathy, use of narcotic analgesics, acute or serious illness, uncontrolled hypertension, moderate to severe CHF, neurological defects involving the lower extremity, bipolar disorder, history of adverse reaction to anti-depressants, or current treatment for depression.</p></sec></sec><sec><title>Participants</title><p>Adults with diagnosis of OA of the knee were recruited for this study. Sample size was calculated based on the assumptions of a) a minimum clinically significant difference equal to 10% of the maximum VAS pain score, b) a population standard deviation of the difference between SAMe and celecoxib equal to 27.5% of the maximum VAS pain score, c) a two-sided alpha level of 0.05 and d) a beta level of 0.2 (80% power). Using the method given by Lachin [<xref ref-type="bibr" rid="B18">18</xref>], 60 patients would be required.</p><p>During the open enrollment, four electronic recruitment announcements to all employees at the University of California, Irvine, and its medical center were sent. This resulted in 242 initial inquiries about the study with 190 telephone-screening interviews. Seventy subjects met the inclusion /exclusion criteria and were invited to be examined by the physician. Sixty-one participants agreed and were finally enrolled in this study.</p></sec><sec><title>Study design</title><p>The study was a randomized, double-blind, cross-over trial over a 4-month period. Each phase of the study lasted 8 weeks, with one week of washout in between. The Specialty Pharmacy conducted the randomization process after each participant met the inclusion criteria and was enrolled in the study. Allocation to treatment groups was concealed to participants, investigators, and research staff until all the participants had completed the study.</p></sec><sec><title>Study procedure</title><p>Participants were assessed in-person during 5 visits conducted at monthly intervals. Visit #1 was the baseline assessment, visit #2 was the phase 1 midpoint assessment, visit # 3 was the end of Phase 1 and served as baseline assessment of Phase 2, visit #4 was at the midpoint of phase 2, and visit #5 was at the end of phase 2, upon completion of the study. At visits #1, 3, and 5, assessments included a review of concurrent medications, a medical examination by a physician and a nurse; assessment of pain, activity impairment, functional health, mood, and knee strength and flexibility measurements. These encounters lasted about 1.5 hours each. At visits #2 and 4, assessments included vital signs, weight, edema measurements, a review of adverse experiences and pain. Visits #2 and 4 lasted about 20 minutes each.</p><p>During each visit, participants were provided with a calendar page to serve as a health diary for the upcoming month. Participants were asked to note any changes in their health and any additional medication they used for their knee pain. Participants were advised, when possible, to use only acetaminophen (Tylenol) for breakthrough knee pain. When stronger pain medicine was needed, they were asked to note this on their health calendar.</p><p>Each participant received a $100 honorarium at the end of the study to cover travel expenses for five visits.</p></sec><sec><title>Intervention</title><p>Participants were randomly assigned to one of two sequences: Sequence A received SAMe 600 mg bid for 8 weeks in Phase 1, followed by celecoxib (Celebrex) 100 mg bid for 8 weeks in Phase 2. Sequence B received celecoxib for 8 weeks in Phase 1, followed by SAMe for 8 weeks in Phase 2. One week of washout was allowed prior to dispensing each of the study medicines.</p><p>SAMe is biologically synthesized from L-methionine and adenosine triphosphate (ATP) mediated by the enzyme adenosylmethionine synthetase (formally known as methionine adenosyltransferase or MAT) [<xref ref-type="bibr" rid="B4">4</xref>]. The SAMe (S-adenosyl-L-methionine disulfate monotosylate salt) used in our study contained 45% S/S and 55% R/S isomer or a ratio of 45%/55%. The S/S or R/S indicates the chirality of the amino acid molecule where the first letter refers to the confirmation at the sulfur atom and the second letter refers to the confirmation at the alpha-carbon. A stereospecific colorimetric assay has been shown to identify these isomers [<xref ref-type="bibr" rid="B19">19</xref>]. The S-Adenosyl-L-Methionine (SAMe) used in this study was enteric coated 200 mg tablets from 400 mg S-Adenosyl-L-Methionine Disulfate Monotosylate salt, the "elemental" amount of SAMe in the salt mixture is 540 mg. The SAMe was verified for quality assurance and potency using a high pressure liquid chromatography (HPLC) analysis with a reference standard of S-adenosyl-methionine p-tolunenesulfonate salt (Sigma Lot 77H7053). Results of analysis yielded 99% of relative label claim with <0.005% relative % of degradation product.</p><p>Celecoxib was of pharmaceutical grade quality, available to pharmacies all over the United States. The Research Pharmacy produced "Dummy" placebo capsules and tablets, as celecoxib is marketed in capsules and SAMe in tablets. The SAMe and celecoxib (Celebrex) medications were stored under refrigeration in the research pharmacy, protected from heat and light. Medications were labeled to meet state dispensing requirements and were dispensed in a double-blind, double-dummy randomized design.</p></sec><sec><title>Main Outcome Measures</title><sec><title>Pain</title><p>A visual analog scale [<xref ref-type="bibr" rid="B20">20</xref>] was administered at visits # 1,2,3,4, 5, with four-week intervals between visits. One scale ascertained pain "today" and a second scale measured pain "in the past month".</p></sec><sec><title>Activity impairment</title><p>A 24-item questionnaire, adapted from the Rowland-Morris activity scale [<xref ref-type="bibr" rid="B21">21</xref>], was administered at visits # 1,3, and 5.</p></sec><sec><title>Functional health</title><p>Perceived functional activities were measured with the COOP questionnaire [<xref ref-type="bibr" rid="B22">22</xref>], addressing level of physical fitness, emotional distress, daily work, social activity, pain, change in health, overall health, social support, and quality of life. All nine items used pictorial representations and brief descriptions of the five levels of intensity. The COOP questionnaire has been shown to be reliable and valid [<xref ref-type="bibr" rid="B23">23</xref>]. The questionnaire was administered at visits # 1, 3, and 5.</p><p>Functional health was measured with a widely used health survey, the MOS SF-36 [<xref ref-type="bibr" rid="B24">24</xref>]. The SF-36 scales include physical functioning, role limitation due to physical health, role limitation due to emotional problems, vitality, mental health, social function, bodily pain, and general health. The SF-36 has been shown to be reliable and valid [<xref ref-type="bibr" rid="B23">23</xref>]. The questionnaire was administered at visits # 1, 3, and 5. Both instruments have been shown to have good concurrent validity.</p></sec><sec><title>Mood status</title><p>Depression was measured with the short form of the Geriatric Depression Scale [<xref ref-type="bibr" rid="B25">25</xref>], comprising 15 yes/no questions. The questionnaire was administered at visits # 1, 3, and 5.</p></sec><sec><title>Clinical assessment of OA of the knee</title><p>A physician, blinded to the participant's randomzation followed the Western Ontario Mac Master (WOMAC) protocol [<xref ref-type="bibr" rid="B26">26</xref>] at visits # 1,3, and 5. Three residents from the Department of Physical Medicine & Rehabilitation, trained to conduct the protocol, performed clinical assessments to measure tenderness, swelling, and fluid in the knee on a 4-point scale. They measured the knee circumference (in cm) at 2 cm above the superior and 2 cm below the inferior border of the patella. They asked the participants to indicate the frequency of mild, moderate and severe pain and to rate pain during walking measured on a 10-point scale. They asked participants to indicate which activities of daily living they felt impaired, from a choice of 8 activities.</p></sec><sec><title>Knee strength and mobility</title><p>An isokinetic multi-joint system, the Biodex System III (Biodex Medical Systems, Shirley, NY), was used to measure isometric strength at 60 degrees (peak flexion and peak extension); isokinetic strength at 60 degrees (peak flexion, peak extension, average flexion, average extension); and isokinetic strength at 180 degrees (peak flexion, peak extension, average flexion, average extension). Walking speed was timed electronically over a 5-meter distance.</p></sec><sec><title>Side effects</title><p>Subjects were made aware of possible side effects during the first visit. Adverse reactions were monitored with health diaries, nursing assessments and clinical interviews in-person at visits # 2, 3, 4, and 5. Subjects were also asked to phone or e-mail about adverse symptoms.</p></sec></sec><sec><title>Statistical methods</title><p>Comparisons between SAMe and celecoxib were based on Analysis of Variance (ANOVA) with factors for treatment (SAMe versus celecoxib), Time (Phase 1 versus Phase 2), Sequence (SAMe-celecoxib versus celecoxib-SAMe) and subjects within sequence. Treatment and time were within-subject effects and Sequence was a between-subjects effect. The primary analysis included only subjects with data from both the SAMe and celecoxib phase. Additional analyses were performed on all available data using the Generalized Estimating Equation (GEE) method to handle missing data.</p><p>For pain scores (measured at the middle as well as the end of each treatment period) effects of time on each treatment were analyzed using a within subject factorial ANOVA with effects of treatment (SAMe versus Celecoxib), time on the treatment (Month 1 versus Month 2) and the interaction of treatment with time on each treatment.</p><p>Comparisons between the two sequence groups at different time periods were performed using Student's t test. Comparisons of post-treatment results with baseline were performed using the paired t test. Comparisons of the incidence of adverse events used the exact binomial test for within-subject comparisons and the Fisher exact test for between-subject comparisons.</p><p>Analyses of possible effects of SAMe deterioration on pain scores were analyzed assuming a negative exponential decay function and performing an ANOVA with factors for time on the treatment (Month 1 versus Month 2) and the calculated dose of SAMe.</p><p>Statistical comparisons were based on all matched pairs data or on all available data as appropriate. Sample sizes may be lower than the total number of subjects if data were missing. Two-sided p values are reported.</p><p>Equivalence tests were performed on pain measures using the Two One-Sided Tests (TOST) procedure [<xref ref-type="bibr" rid="B27">27</xref>]. For assessment of equivalence, the minimum clinically significant difference was defined as 10 percent of the maximum VAS pain score. The 90% confidence intervals were determined using the standard error of the difference scores for all patients with data in both phases of the study.</p></sec></sec><sec><title>Results</title><sec><title>Randomization and Compliance</title><p>A total of 61 subjects met the inclusion criteria and were entered in this study (Table <xref ref-type="table" rid="T1">1</xref>). After completing the initial assessment, subjects were randomly assigned to sequence A or B. Thirty subjects were randomized to sequence A (SAMe followed by celecoxib) and 31 subjects into sequence B (celecoxib followed by SAMe).</p></sec><sec><title>Demographics and Baseline Data</title><p>There were no significant demographic differences between subjects assigned to sequence A and those assigned to Sequence B (ps > 0.10) (Table <xref ref-type="table" rid="T2">2</xref>). Subjects in this study tended to be obese, as indicated by an average body mass index greater than 30.</p><p>Sequence A (SAMe-celecoxib) comprised 30 participants assigned at random, 28 completed Phase 1 of the protocol and 27 completed Phase 2. Twenty-seven participants were analyzed for the primary outcome. Sequence B (celecoxib-SAMe) comprised 31 participants assigned at random, 29 completed Phase 1 of the protocol and 29 completed Phase 2. Twenty-nine participants were analyzed for the primary outcome. Three subjects discontinued participation in Sequence A and two in Sequence B. In each sequence, one subject withdrew during the washout period and one during the celecoxib treatment. During the washout period, the reasons for withdrawal were: insufficient pain to warrant treatment and stroke. During the celecoxib treatment, reasons for withdrawal were: one subject did not reply to follow-up scheduling, and one was diagnosed with rheumatoid arthritis 4 weeks into the treatment. During the SAMe treatment one patient withdrew after 3 days due to headache.</p></sec><sec><title>Pain Based on the VAS Scale</title><p>Pain was assessed on two separate visual analog scales. For the first scale, subjects were asked to rate pain today. For the second scale, subjects were asked to rate pain over the past month. Analyses of results from the two measures are similar. In Phase 1, at the end of the first month, the celecoxib group showed significant reductions from baseline (ps < 0.01), while the reductions of pain in the SAMe group were only marginal (ps < 0.10). Celecoxib showed significantly more reduction in pain than SAMe during the first month of treatment (p = 0.024). By the second month of Phase 1, both groups were significantly improved from baseline (ps < 0.01) and there was no significant difference between them.</p><p>In Phase 2, during the first month, the group that switched from celecoxib to SAMe was noticeably but not significantly worse than the group that shifted from SAMe to celecoxib. However, by the end of Phase 2, there was no apparent difference between the two groups.</p><p>Despite these minor differences, the overall results from the two measures on pain in both groups are similar (See Figure <xref ref-type="fig" rid="F1">1</xref>).</p><p>Because the effects of SAMe and celecoxib seem unaffected by the order of treatment, it is reasonable to combine results from Phase 1 and Phase 2. Cross-over analyses indicated that celecoxib was significantly better at Month 1 (ps ≤ 0.023) but there was no significant difference between the two groups at Month 2. There were no significant effects of time (Phase 1 versus Phase 2) or sequence (A versus B). The analysis combining Month 1 and Month 2 data showed significant effects of treatment for pain over the past month (p = 0.043) but not for pain today (p = 0.247). For both pain measures, the effects of duration of treatment and the interaction of duration with type of treatment were statistically significant (ps ≤ 0.029) indicating that the patterns shown in Figures 1-4 are reliable.</p><p>The mean difference between the two groups (SAMe minus Celecoxib) on the "pain today" measure was 9.3 at the end of Month 1 and -0.4 at the end of month 2. The difference score on the "pain over the past month" measure was 11.6 at the end of Month 1 and 1.6 at the end of Month 2. The 90% confidence intervals for the "pain today" measure were 2.7 to 15.8 at the end of Month 1 and -7.4 to 6.6 at the end of Month 2. The 90% confidence intervals for the "pain over the past month" measure were 5.2 to 18.1 at the end of Month 1 and -6.1 to 9.3 at the end of Month 2. Thus, regardless of the pain measure used, equivalence (based on a minimum clinically significant difference of 10) was demonstrated at the end of Month 2 but not at the end of Month 1 (Figure <xref ref-type="fig" rid="F2">2</xref>).</p></sec><sec><title>COOP Scores</title><p>COOP scores are shown in Table <xref ref-type="table" rid="T3">3</xref>. ANOVA showed no significant difference between SAMe and celecoxib on any measure. Effects of time (Phase 1 versus Phase 2) and sequence were also non significant. Both groups showed a significant improvement from baseline in the total COOP score as well as [work], [pain], [change in health] and [overall health]. Neither group showed a significant change from baseline in [physical fitness], [emotional] and [daily activity]. On the [social support] and [quality of life] measures, the SAMe group showed a significant improvement from baseline (ps ≤ 0.018) while the celecoxib group did not.</p></sec><sec><title>SF-36 Scores</title><p>SF-36 scores are shown in Table <xref ref-type="table" rid="T4">4</xref>. ANOVA showed no significant difference between SAMe and celecoxib on any measure. There was a significant effect of time (Phase 1 versus Phase 2) on the vitality measure (p = 0.024) but, otherwise, there were no significant time or sequence effects. Both groups showed a significant improvement from baseline on the physical function, role physical, bodily pain, and general physical measures (ps ≤ 0.045). Neither group showed a significant change from baseline on the role emotional, mental health, social function, general health or general mental measures. The vitality measure showed a significant improvement for the celecoxib group (p = 0.012) but the magnitude of this effect was quite small.</p></sec><sec><title>WOMAC, Activity and Mood</title><p>WOMAC scores, Roland-Morris activity scores and Mood scores are shown in Table <xref ref-type="table" rid="T5">5</xref>. The only measure that showed any difference between SAMe and celecoxib was circumference of the knee measured in centimeters (cm). The subjects receiving Sequence A showed a decrease from baseline of 0.3 cm in the first (SAMe) phase and an increase from baseline of 0.2 cm during the second (celecoxib) phase. The subjects receiving Sequence B showed an increase from baseline of 0.3 cm in the first (celecoxib) phase and no change from baseline in the second (SAMe) phase. This difference was statistically significant (p = 0.001). Curiously, the measures of swelling in the knee and fluid in the knee showed no such trend. The Roland-Morris activity score and the frequency of at least moderate pain showed significant time (Phase 1 versus Phase 2) effects (ps ≤ 0.035) but, otherwise, there were no significant time or sequence effects. Both groups showed significant improvement in tenderness in the knee, swelling in the knee, pain during walking, pain frequency, ADL, the Roland-Morris activity score and the mood score (ps ≤ 0.037). Neither group showed significant changes from baseline in circumference of the knee or swelling of the knee.</p></sec><sec><title>Biodex</title><p>Biodex data are shown in Tables <xref ref-type="table" rid="T6">6</xref> &<xref ref-type="table" rid="T8">8</xref>. All of the measures showed significant improvement from baseline (ps ≤ 0.001) and most of them showed significant time effects (Phase 1 versus Phase 2) in the overall ANOVA. These measures appeared to be steadily improving over the entire study period regardless of treatment. However, there was no significant difference between SAMe and celecoxib.</p></sec><sec><title>Adverse Events</title><p>Thirty-six subjects reported adverse events during the SAMe period and 46 subjects reported adverse events during the celecoxib period. The most common adverse events were gastrointestinal disorder (SAMe 4: celecoxib: 6, NS), anxiety (SAMe: 5, celecoxib 4, NS) and dyspepsia (SAMe: 1, celecoxib: 3, NS). No other adverse event was reported by more than two subjects. One patient terminated the study due to headache three days into the SAMe treatment; however this patient had a well known history of such headaches prior to enrolling in the study.</p></sec><sec><title>Loss of SAMe potency</title><p>The increased incidence of pain in the SAMe group was concentrated in the first month of Phase 2 where 12 of 29 subjects who had recently switched from celecoxib to SAMe complained of pain, while only 2 of 24 subjects recently switched from SAMe to celecoxib complained of pain (p = 0.011). The difference between groups was not statistically significant at any other time period.</p><p>Approximately 75% of the way through the study, a routine quality check of a random sample of the study medicine indicated that the SAMe had lost approximately 51% of its potency. At this time, the study was delayed until a new batch of SAMe could be obtained. In order to examine this effect, potency was crudely estimated based on the assumption of an exponential decay curve from 100% potency at the start of the study to 49% potency at day 595. This yields the equation y = exp (-0.0011989x) where y is potency and x is days from the start of the study or days from the starting time for the new batch of SAMe. An ANOVA on pain-today scores in SAMe treated subjects with a factor for duration of treatment (1 month versus 2 months) and a continuous potency factor produced a significant effect of month showing improvement from Month 1 to Month 2 (p < 0.001). The effect of SAMe potency was in the expected direction (lower estimated potency was correlated with higher pain) but the effect was not statistically significant (p = 0.188).</p></sec></sec><sec><title>Discussion</title><p>To our knowledge this is the first study comparing the effectiveness of SAMe to a COX-2 inhibitor in the management of osteoarthritis symptoms. Recent studies have demonstrated that COX-2 inhibitors have equivalent effectiveness but a better safety profile than other NSAIDs in the management of osteoarthritis. Their advantage is that by avoiding the COX-1 inhibitors they would have less tendency to produce gastrointestinal side effects. Published studies compared SAMe most commonly to 1200 mg of ibuprofen. Despite its efficacy at 1200 mg, ibuprofen is more frequently dispensed at 1600 mg/day, or higher, based on the underlying conditions. To avoid any doubt or controversy, we opted to use the most frequently used dose of SAMe and the optimal dose of celecoxib. Future studies should explore different doses of SAMe, to identify the lowest most effective and safest dose to be used.</p><p>Our results indicate that SAMe is equivalent in almost all measures to COX-2 inhibitors (celecoxib) in relieving pain and improving function in subjects with osteoarthritis of the knee. It is clear from our results that SAMe has a slower onset of action, requiring almost one month of treatment prior to achieving similar effect to celecoxib. COX-2 inhibitors and NSAIDs have a definite advantage during the first month of treatment. This is consistent with prior conclusions by di Padova [<xref ref-type="bibr" rid="B15">15</xref>]. However, it does not clearly explain findings of several previously published small studies [<xref ref-type="bibr" rid="B28">28</xref>][<xref ref-type="bibr" rid="B29">29</xref>][<xref ref-type="bibr" rid="B30">30</xref>][<xref ref-type="bibr" rid="B31">31</xref>]. These studies although limited to 1 month of duration, reported equivalent effectiveness to NSAIDs in relieving symptoms of OA.</p><p>During the second month of each phase of the study, the pain relieving effect is equivalent for both drugs. Of interest is that, while the pain relief of celecoxib was constant throughout the study, the effect of SAMe continued to increase with time (Figure <xref ref-type="fig" rid="F1">1</xref>). This raises the question whether the effect of SAMe would have continued to improve had the study been for a longer period of time? Hence a more extended study is needed to evaluate the long-term effect of SAMe and to assess the presence of a possible ceiling effect of this supplement. Another possible interpretation could be that the effect of SAMe slows the progression of the disease, hence the persistent effect after discontinuation. The same effect was not seen with celecoxib.</p><p>Although this study was not set up to evaluate the long-term effect of SAMe, the data hints that the pain relief effect obtained with SAMe persisted even after the medication was discontinued (Figure <xref ref-type="fig" rid="F1">1</xref>). In this study we noted that subjects who changed from SAMe to celecoxib had an initial decline in their pain level, however the level was back to baseline at 2 months consistent with the level of those that were started on celecoxib. Although the difference was not statistically different in both groups, this initial dip beyond the baseline pain relief obtained by celecoxib alone hint for a possible persistent effect of SAMe for 1 month beyond discontinuation of the medication. This is reminiscent of reports on the pain relieving effect observed with glucosamine sulfate. If this result could be reproduced in future large studies, the question will be raised of whether SAMe could be used as a pulsed therapy, for the management of pain in OA, after an initial period to establish a steady level?</p><sec><title>Functional health</title><p>Several health measures were assessed during this study, including medical and joint function assessments, measures of daily function and activity impairment, assessments of mood and quality of life. On most of these measures, both treatment groups showed significant improvements from baseline and isometric joint function tests appeared to be steadily improving over the entire study period as well, regardless of treatment. No statistically significant difference was seen between SAMe and celecoxib. Thus, the evidence of comparable efficacy between SAMe and celecoxib seems rather solid.</p><p>Of interest but of unclear clinical significance is the measurement of the knee circumference. The circumference differed significantly between the two treatments, with a decrease from baseline of 0.3 cm during the SAMe phase and an increase from baseline of 0.2 cm during the celecoxib phase. While this statistical difference does not seem to have much clinical meaning, future studies on the mechanism of action of SAMe might shed light on this decrease in knee circumference.</p><p>Perceived physical fitness, emotional well-being and daily activities did not improve from baseline with either treatment. Fitness and well-being as well as daily activities might represent more stable attributes than physical complaints and thus might not change in a 2-months treatment.</p></sec><sec><title>Depression</title><p>Of interest is that no participant in this study was identified with depression, using the abbreviated geriatric depression scale, during this study. Hence the results and discussion of effects of SAMe on depression in this study should be interpreted with caution and could not be generalized to a depressed population.</p><p>Despite a slight drop in the mood scale while subjects were on SAMe (see Table <xref ref-type="table" rid="T7">7</xref>), no statistical difference in the depression scale was noted while subjects were on SAMe during this study. However, one conclusion remains clear, given the absence of positive antidepressive effects seen in this study, one can conclude that the beneficial results (pain, function) noted in this study could not be explained by the antidepressive effect of SAMe. Hence other mechanisms of action are in effect, and further studies are warranted to explore these mechanism.</p></sec><sec><title>Adverse events</title><p>The type of side effects seen in this study is consistent with those reported in other trials. Gastrointestinal, psychiatric, insomnia, allergy, and rash were all reported in previous studies. Again consistent with prior data, gastrointestinal and psychiatric side effects were the most common. This is consistent with findings by Soeken et al. [<xref ref-type="bibr" rid="B13">13</xref>] that subjects on SAMe were less likely to experience side effects than those treated with NSAIDs. In her analysis, Soeken reports that this is independent of SAMe dose or length of intervention.</p><p>Of note is that the most common reported complaint in this study, although not included under adverse event, is pain. This could be attributed to either failure of the medication to provide adequate pain relief (SAMe during the first weeks of therapy); or for this study during the period where the potency of SAMe tablets was initially noted to have declined.</p></sec><sec><title>Potency and quality assurance</title><p>SAMe is produced biologically in the S/S form. Both R/S and S/S isomers are biologically active [<xref ref-type="bibr" rid="B32">32</xref>]. Under normal physiological conditions or normal storage conditions, SAMe spontaneously racemizes to form a mixture of S and R isomers, thus over time, the S/S form converts into the R/S form. This is an equilibration phenomenon and occurs in the commercial form. The racemisation at the sulfur side of the SAMe molecule has been established in the literature [<xref ref-type="bibr" rid="B33">33</xref>]. All commercial SAMe products are manufactured by combining fermentation and synthetic steps. Commercially available SAMe has been tested to have different ratios of S/S & R/S isomers ranging from 70%/30% to 45%/55% (unpublished data). Several studies have indicated that the S/S-SAMe configuration is the "active" form involved in methyltransferase catalyzed reactions [<xref ref-type="bibr" rid="B34">34</xref>]'[<xref ref-type="bibr" rid="B35">35</xref>]'[<xref ref-type="bibr" rid="B36">36</xref>] and that the R/S-SAMe configuration is the biologically "inactive" form as it may inhibit methyltransferases reactions [<xref ref-type="bibr" rid="B37">37</xref>]. In more of a functional outcome, other authors have shown that both S/S and R/S isomers have significant activity respectively in increasing blood flow and bile production in isolated perfused rat livers [<xref ref-type="bibr" rid="B38">38</xref>]. Therefore, there may not appear to be any difference in the biological effects of the different ratios of SAMe.</p><p>In this study, we did not perform analysis of the S/S & S/R isomers during quality assurance. Hence we are unable to verify whether formulations with higher or lower S/S or S/R isomers are more potent in the management of OA. Our experience highlights the need for future studies, involving dietary supplements, to develop a quality assurance safeguard.</p><p>Given the absence of clear and well established explanation for the mechanism of action for SAMe in the management of OA, future studies should explore the possible role of isomers.</p></sec><sec><title>Limitations of the study</title><p>Although this study had fewer patients than originally proposed, it retained sufficient power to conclude that a) celecoxib is significantly better than SAMe in the first month of treatment and b) by the second month of treatment, SAMe and celecoxib were clinically equivalent based on the proposed equivalence criterion.</p><p>Findings should be interpreted with caution due to the small number of participants and the short duration of the study. Toward the end of the study, SAMe lost about half of its potency and the study was delayed until a new supply of SAMe was obtained. While the lower potency was associated with higher pain, this effect was non-significant and did not affect the overall study outcome. Future long-term studies evaluating dietary supplements should implement quality checks at appropriate time intervals for adequate quality assurance.</p></sec></sec><sec><title>Conclusion</title><p>SAMe is helpful for the management of pain in osteoarthritis, and demonstrates similar effectiveness as a currently accepted COX-2 inhibitor celecoxib. Results from this study add and confirm results from prior studies indicating a possible role for SAMe in the management of osteoarthritis. It is clear however, that many questions remain unanswered. Prime among these questions is: What is the optimal (effective and safe) dose SAMe in the management of osteoarthritis? Other question that should be explored are: What is the long-term effectiveness and safety of SAMe? What is the mechanism of action of SAMe? Does SAMe effect the disease progression or could it reverse the disease process? and finally Is the combination of SAMe and a COX-2 inhibitor more effective than either alone in the management of osteoarthritis? Most of the currently available studies offer a promise, however well done randomized controlled trials are needed to answer any and all of these questions.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>JST, SR, and WIN conceived the initial research design and protocol. SR and JST carried out data collection and day-to-day coordination of this study. FH provided all the statistical and methodological support. PWH provided advice and coordinated the procurement of SAMe. All the authors contributed to the manuscript and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1471-2474/5/6/prepub"/></p></sec>
Crystal structure of the YffB protein from <italic>Pseudomonas aeruginosa </italic>suggests a glutathione-dependent thiol reductase function	<sec><title>Background</title><p>The <italic>yffB </italic>(PA3664) gene of <italic>Pseudomonas aeruginosa </italic>encodes an uncharacterized protein of 13 kDa molecular weight with a marginal sequence similarity to arsenate reductase from <italic>Escherichia coli</italic>. The crystal structure determination of YffB was undertaken as part of a structural genomics effort in order to assist with the functional assignment of the protein.</p></sec><sec><title>Results</title><p>The structure was determined at 1.0 Å resolution by single-wavelength anomalous diffraction. The fold is very similar to that of arsenate reductase, which is an extension of the thioredoxin fold.</p></sec><sec><title>Conclusion</title><p>Given the conservation of the functionally important residues and the ability to bind glutathione, YffB is likely to function as a GSH-dependent thiol reductase.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Teplyakov</surname><given-names>Alexey</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Pullalarevu</surname><given-names>Sadhana</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Obmolova</surname><given-names>Galina</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Doseeva</surname><given-names>Victoria</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Galkin</surname><given-names>Andrey</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Herzberg</surname><given-names>Osnat</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Dauter</surname><given-names>Miroslawa</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A8" contrib-type="author"><name><surname>Dauter</surname><given-names>Zbigniew</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A9" corresp="yes" contrib-type="author"><name><surname>Gilliland</surname><given-names>Gary L</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Structural Biology	<sec><title>Background</title><p>The <italic>yffB </italic>(PA3664) gene of <italic>Pseudomonas aeruginosa </italic>encodes an uncharacterized protein of 13 kDa molecular weight. Based on the amino acid sequence analysis, YffB and its homologs have been assigned to the family of arsenate reductase (AR) and related proteins (Pfam entry PF03960) [<xref ref-type="bibr" rid="B1">1</xref>]. AR participates in arsenic detoxification by catalyzing the reduction of arsenate [the oxyanion of As(V)] to arsenite [the oxyanion of As(III)], which is then exported through a specific transport system [<xref ref-type="bibr" rid="B2">2</xref>]. There are two different types of bacterial ARs. AR of Gram-negative bacteria has a distinct HX<sub>3</sub>CX<sub>3</sub>R catalytic sequence motif, belongs to the thioredoxin (Trx) structural superfamily, and is coupled to the glutathione (GSH) and glutaredoxin (Grx) system for its enzyme activity [<xref ref-type="bibr" rid="B3">3</xref>,<xref ref-type="bibr" rid="B4">4</xref>]. AR of Gram-positive bacteria also has a redox active cysteine residue but within a CX<sub>5</sub>R sequence motif embedded in a fold typical for low molecular weight protein tyrosine phosphatases, and requires Trx and Trx reductase for enzyme activity [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>].</p><p>Besides the Trx-like AR, the PF03960 family includes a number of uncharacterized proteins whose function seems unlikely to be arsenate reductase. Members of the family are widely represented in bacteria, but not in archaea or eukaryotes. The pattern of amino acid distribution along the polypeptide chain suggests that these proteins may have a common fold.</p><p>The crystal structure determination of YffB was undertaken as part of a structural genomics effort [<xref ref-type="bibr" rid="B8">8</xref>] in order to assist with the functional assignment of the protein. The project was focused on the so-called hypothetical proteins from <italic>Haemophilus influenzae</italic>. The YffB protein from <italic>P. aeruginosa </italic>has emerged as an ortholog of HI0103 to increase chances for successful crystallization. YffB was cloned, expressed, and the crystal structure determined at 1.0 Å resolution. The protein fold appeared to be similar to that of <italic>Escherichia coli </italic>AR. Analysis of the structure suggests that YffB may function as a thiol reductase.</p></sec><sec><title>Results and discussion</title><p>The atomic model of YffB contains all residues but the N-terminal methionine, which has probably been posttranslationally cleaved. The molecular weight of the SeMet protein measured by matrix-assisted laser-desorption ionization (MALDI) mass spectrometry was close to the calculated value of 13,121 Da for residues 2–115. An addition of one selenomethionine would have increased the molecular weight by 178 Da.</p><p>The structure of YffB consists of two domains (Fig. <xref ref-type="fig" rid="F1">1</xref>). One is formed by a four-stranded mixed β-sheet flanked by two α-helices on one side. The other domain is an α-helical bundle comprising residues 38–88. The overall fold is very similar to AR from <italic>E. coli </italic>encoded by the <italic>arsC </italic>gene [<xref ref-type="bibr" rid="B4">4</xref>]. The rms deviation between the structures is 2.3 Å for all 109 common Cα atoms, the Z-score calculated by DALI [<xref ref-type="bibr" rid="B9">9</xref>] is 12.6. Relative to YffB, ArsC has an additional C-terminal domain, a 3-strand meander, that covers helices α1 and α7.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>(Left) Ribbon presentation of the polypeptide fold of YffB. Active site residues are shown as ball-and-stick models (produced with MOLSCRIPT [<xref ref-type="bibr" rid="B19">19</xref>]). (Right) Electrostatic surface potential calculated with GRASP [<xref ref-type="bibr" rid="B20">20</xref>]. Positive charges are blue and negative are red.</p></caption><graphic xlink:href="1472-6807-4-5-1"/></fig><p>The α/β-domain is characteristic of a superfamily of the Trx-like proteins [<xref ref-type="bibr" rid="B10">10</xref>]. Common to all of them, a <italic>cis</italic>-proline residue is located at the N-terminus of strand β3. This residue is important for the integrity of the active site, as the <italic>cis</italic>-peptide bond promotes a turn of the polypeptide chain. The corresponding Pro93 in YffB is also in the <italic>cis</italic>-conformation.</p><p>The overall structural similarity to the members of the Trx superfamily allows the identification of the putative active site of YffB in the loop between the first β-strand and the following α-helix. A cysteine residue at the N-terminus of the α-helix acts as a catalytic nucleophile [<xref ref-type="bibr" rid="B10">10</xref>]. Its pK is lowered by the basic environment and the dipole of the helix [<xref ref-type="bibr" rid="B11">11</xref>]. In ArsC the thiol group of the cysteine attacks the arsenic of the substrate to form a covalent intermediate [<xref ref-type="bibr" rid="B4">4</xref>]. The arsenite ion is released upon binding of GSH, which is then reduced by Grx [<xref ref-type="bibr" rid="B3">3</xref>]. Unlike many Trx-like proteins, the catalytic cysteine in ArsC does not form an internal disulfide in the oxidized state. YffB also has only one cysteine residue at the active site (Cys11), which further emphasizes its similarity to ArsC and suggests that it functions in a GSH-dependent manner.</p><p>Despite the structural similarity to ArsC, the two proteins share only 16% identical residues. However, most of the residues involved in substrate binding and catalysis are conserved in these proteins. Three invariant arginine residues, Arg60, Arg94, and Arg107, bind the substrate and stabilize the reaction intermediate in ArsC [<xref ref-type="bibr" rid="B12">12</xref>]. They may also enhance the nucleophilicity of the active cysteine. Arg94 together with the ensuing cis-Pro95 is conserved in all AR-related proteins. Lys91 of YffB, despite its different location in the sequence, is spatially equivalent to Arg60 of ArsC and therefore can take part in substrate binding. The most important difference with respect to ArsC is the substitution of Gly for Arg107 that leaves the binding site without the positively charged anchor and also makes Cys11 more accessible for bulky compounds. This substitution probably reflects a difference in substrate specificity between ArsC and YffB.</p><p>One particular consequence of this Arg to Gly replacement might be the ability of YffB to bind GSH, whereas ArsC cannot bind GSH in the absence of arsenate [<xref ref-type="bibr" rid="B13">13</xref>]. The binding of GSH to YffB was detected by MALDI mass spectrometry using an oxidized form of glutathione (GSSG). The mass increase of about 300 Da indicated one GSH molecule bound per protomer. This result supports the contention that GSH is probably involved in the functional cycle of YffB.</p><p>YffB has a very polarized distribution of charges over the surface of the molecule. The active site area is thronged with basic residues, as it is in ArsC, whereas the opposite side of the molecule is predominantly negatively charged (Fig. <xref ref-type="fig" rid="F1">1B</xref>). The positive electrostatic potential would certainly favor binding of anions.</p></sec><sec><title>Conclusion</title><p>Given the structural similarity to the Trx-like proteins and particularly to ArsC, and the conservation of the functionally important residues, YffB is likely to function as a thiol reductase. The nature of the substrate remains to be established in further biochemical and biophysical studies. These studies will be facilitated by the three-dimensional structure of the protein.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Cloning, expression and purification</title><p>The <italic>yffB </italic>(PA3664) gene from <italic>Pseudomonas aeruginosa </italic>PAO1 was amplified using <italic>PfuTurbo </italic>DNA polymerase (Stratagene), genomic DNA (ATCC 47085D), and the following 5'- and 3'-end primers.</p><p>Forward: 5'-CACC<underline>CTGGTGCCGCGCGGCAGC</underline><italic>CATATG</italic>ACCTACGTTCTCTACGGCATCA-3'.</p><p>Reverse: 5'-TCAGGCCAGGGCGGC-3'.</p><p>The sequence encoding the thrombin cleavage site is underlined, and the <italic>Nde</italic>I restriction site is shown in italic. The PCR product was introduced into a pET100/D-TOPO expression vector by the TOPO directional cloning procedure (Invitrogen). Recombinant plasmids were isolated from the <italic>E. coli </italic>TOP10 strain. The expression construct for production of the native protein without a His-tag was prepared by digestion with <italic>Nde</italic>I and self ligation. For production of the selenomethionine (SeMet) protein, <italic>E. coli </italic>strain B834 (DE3) was transformed with the recombinant plasmid, and cells were grown in a minimal medium supplemented with 100 μg/mL ampicillin and 40 μg/mL SeMet until the A<sub>600 </sub>reached 0.8. At this point the cells were induced with 1 mM isopropyl β-D-thiogalactoside and harvested after 3 h.</p><p>The SeMet protein was purified by column chromatography in three steps. The cell extract was applied to a Q Sepharose HP (Pharmacia) column equilibrated with 20 mM HEPES (pH 6.7), 50 mM NaCl, and 0.5 mM EDTA. About half of the protein bound to the column and was eluted in a 50–500 mM NaCl gradient. After dialysis in 20 mM HEPES (pH 6.7) and 0.5 mM EDTA, the fractions containing the protein were applied to a Source 15S (Pharmacia) column, and eluted with a 0–450 mM NaCl gradient. The protein was concentrated to 4 mg/ml, applied to a Sephacryl S100 (Pharmacia) gel filtration column, and eluted in 20 mM HEPES (pH 7.5), 100 mM NaCl, and 0.25 mM EDTA. According to the SDS gel, the protein was at least 95% pure. For crystallization, the protein was concentrated to 15 mg/ml.</p></sec><sec><title>Crystallization and structure determination</title><p>YffB crystals were grown by the vapor diffusion hanging drop method at room temperature from 0.1 M CHES, pH 10, 26% polyethylene glycol 3350, and 5% isopropanol. They belong to the space group C2 with unit cell parameters: <italic>a </italic>= 87.45 Å, <italic>b </italic>= 43.25 Å, <italic>c </italic>= 29.06 Å, β = 93.5°. There is one protein molecule in the asymmetric unit with the solvent content of 40%. For X-ray data collection, the crystals were soaked in the mother liquor supplemented with 15% polyethylene glycol 400 and flash-frozen in liquid propane.</p><p>The structure was solved by using single-wavelength (0.9794 Å) anomalous X-ray diffraction data collected on the NCI-NIH beamline at the National Synchrotron Light Source (Upton, NY). The data (Table <xref ref-type="table" rid="T1">1</xref>) were processed with HKL2000 [<xref ref-type="bibr" rid="B14">14</xref>]. Two selenium sites were located by SHELXD and were used for phasing with SHELXE [<xref ref-type="bibr" rid="B15">15</xref>]. The polypeptide chain was automatically traced with RESOLVE [<xref ref-type="bibr" rid="B16">16</xref>]. The atomic model was completed using O [<xref ref-type="bibr" rid="B17">17</xref>] and refined with REFMAC [<xref ref-type="bibr" rid="B18">18</xref>] using anisotropic B-factors. The model includes residues 2–115 of the protein, a molecule of isopropanol, and 220 water molecules. 93% residues have main-chain torsion angles in the most favored conformation.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>X-ray data and refinement statistics.</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left">Resolution (Å)</td><td align="left">30-1.02 (1.04-1.02)<sup>1</sup></td></tr><tr><td align="left">Number of unique reflections<sup>2</sup></td><td align="left">98,866 (5,709)</td></tr><tr><td align="left">Completeness (%)</td><td align="left">91.2 (92.3)</td></tr><tr><td align="left">Redundancy</td><td align="left">3.9 (3.7)</td></tr><tr><td align="left">R<sub>sym </sub>(Σ\|I-<I>\|)/ΣI)</td><td align="left">0.051 (0.419)</td></tr><tr><td align="left"><I/σ></td><td align="left">22.0 (3.6)</td></tr><tr><td align="left">Fraction of refls with I>3σ (%)</td><td align="left">81.6 (50.2)</td></tr><tr><td align="left">R<sub>cryst </sub>(Σ\|\|F<sub>o</sub>\|-\|F<sub>c</sub>\|\|)/Σ\|F<sub>o</sub>\|)</td><td align="left">0.129 (0.155)</td></tr><tr><td align="left">R<sub>free </sub>(2% data)</td><td align="left">0.139 (0.202)</td></tr><tr><td align="left">Number of protein non-hydrogen atoms</td><td align="left">935</td></tr><tr><td align="left">Number of water molecules</td><td align="left">220</td></tr><tr><td align="left">Mean B-factor from the model (Å<sup>2</sup>)</td><td align="left">10.4</td></tr><tr><td align="left">Mean B-factor from Wilson plot (Å<sup>2</sup>)</td><td align="left">8.1</td></tr><tr><td align="left">RMSD in bonds (Å)</td><td align="left">0.011</td></tr><tr><td align="left">RMSD in angles (°)</td><td align="left">1.3</td></tr><tr><td align="left">RMSD in main-chain B factors (Å<sup>2</sup>)</td><td align="left">2.1</td></tr></tbody></table><table-wrap-foot><p><sup>1</sup>Data for the highest resolution shell are given in parentheses <sup>2</sup>Anomalous pairs not merged</p></table-wrap-foot></table-wrap><p>The atomic coordinates of YffB and structure factors were deposited in the Protein Data Bank under the accession code 1RW1.</p></sec><sec><title>Glutathione binding</title><p>Binding of glutathione was detected by MALDI mass spectrometry using a Voyager spectrometer (Applied Biosystems, Foster City, CA). The SeMet protein (10 μM) was incubated for 1 h at room temperature with 1 mM GSSG in 20 mM HEPES buffer, pH 7.<bold>T</bold>he sample was mixed 1:1 with matrix solution (10 mg/mL 3,5-dimethoxy-4-hydroxycinnamic acid, 50% aqueous acetonitrile, and 0.2% trifluoroacetic acid), deposited onto a golden plate, and allowed to dry at room temperature. Bovine myoglobin was used for molecular mass calibration.</p></sec></sec><sec><title>Authors' contributions</title><p>AT modeled, refined and analyzed the structure, performed MALDI experiments, and drafted the manuscript. SP and GO purified and crystallized the protein. VD and AG cloned and expressed the protein. MD and ZD collected and processed the diffraction data and calculated electron density maps. OH and GLG coordinated the study and provided financial support.</p></sec>
In vitro test of external Qigong	<sec><title>Background</title><p>Practitioners of the alternative medical practice 'external Qigong' generally claim the ability to emit or direct "healing energy" to treat patients. We investigated the ability of experienced Qigong practitioners to enhance the healthy growth of cultured human cells in a series of studies, each following a rigorously designed protocol with randomization, blinding and controls for variability.</p></sec><sec sec-type="methods"><title>Methods</title><p>Qigong practitioners directed healing intentionality toward normal brain cell cultures in a basic science laboratory. Qigong treatments were delivered for 20 minutes from a minimum distance of 10 centimeters. Cell proliferation was measured by a standard colony-forming efficiency (CFE) assay and a CFE ratio (CFE for treated samples/CFE for sham samples) was the dependent measure for each experiment.</p></sec><sec><title>Results</title><p>During a pilot study (8 experiments), a trend of increased cell proliferation in Qigong-treated samples (CFE Qigong/sham ratios > 1.0) was observed (P = 0.162). In a formal study (28 experiments), a similar trend was observed, with Qigong-treated samples showing on average more colony formation than sham samples (P = 0.036). In a replication study (60 experiments), no significant difference between Qigong-treated samples and sham samples was observed (P = 0.465).</p></sec><sec><title>Conclusion</title><p>We observed an apparent increase in the proliferation of cultured cells following external Qigong treatment by practitioners under strictly controlled conditions, but we did not observe this effect in a replication study. These results suggest the need for more controlled and thorough investigation of external Qigong before scientific validation is claimed.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Yount</surname><given-names>Garret</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Solfvin</surname><given-names>Jerry</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Moore</surname><given-names>Dan</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Schlitz</surname><given-names>Marilyn</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Reading</surname><given-names>Melissa</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Aldape</surname><given-names>Ken</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Qian</surname><given-names>Yifang</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Complementary and Alternative Medicine	<sec><title>Background</title><p>Healing practices that purportedly involve the mental or spiritual manipulation of some form of life energy associated with the body are popular forms of complementary and alternative medicine worldwide. The National Center for Complementary and Alternative Medicine currently supports research projects investigating such 'energy medicine' modalities [<xref ref-type="bibr" rid="B1">1</xref>]. Energy medicine practitioners generally claim the ability to emit or direct "healing energy" to treat patients. Whether the apparent efficacy of such energy medicine practices is due to processes internal or external to the patients is not clear, however. The human body has an interwoven network of central nervous system and endocrine system processes involved in modulation of the immune system in response to psychological states [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>]. Self-healing processes may thus be induced by psychological states triggered by the drama of a healing ritual or by the charisma of a practitioner. Do energy medicine modalities exert their effect (if any) through the power of suggestion, then, or does it have a direct effect on biological systems? This question may be answered using samples in the laboratory as targets of a practitioner's treatment. Such <italic>in vitro </italic>models effectively eliminate from experiments the factor of psychological cueing.</p><p>A vast majority of the energy medicine studies published over the past 20 years have been evaluations of Qigong. Qigong is a health-promoting meditative practice originating from traditional Chinese medicine. Like herbal remedies and acupuncture, its theoretical basis involves manipulation of the purported life energy, 'Qi'. According to traditional Chinese medicine principles, Qi can be regulated mentally as well as chemically (as with herbal remedies) or physically (as with acupuncture). A person's mind is believed to be capable of influencing Qi, which travels along the meridians in the body and promotes health. Most Qigong practice is self-meditation. Through daily meditation practices, sometimes involving simple movements or postures and breathing exercises, a person focuses his or her own intentionality on improving or maintaining health. Some individuals believe that, after many years of practicing such 'internal Qigong', they can develop the ability to manipulate Qi outside the body. This is the basis of 'external Qigong'. Supporting the validity of such external applications of Qigong, a database of reports from Asia translated into English contains approximately 90 abstracts [<xref ref-type="bibr" rid="B4">4</xref>] presenting evidence that external Qigong can influence biological targets <italic>in vitro</italic>.</p><p>Many in the Chinese scientific community question the credibility of these studies, however, and argue that therapeutic effects following Qigong treatment, if any, must result from the power of suggestion. Some in China have reported on experimental evidence supporting this view and have proclaimed the "end of fairy-tale external Qigong" [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>]. Chinese research on both sides of the controversy over the efficacy of external Qigong does not appear to follow accepted scientific standards however, creating doubt as to its validity. As one example, "[Publication bias] appears to be so much ingrained in Chinese publications on ... Qigong that the literature cannot be trusted" [<xref ref-type="bibr" rid="B7">7</xref>]. Poor design and lack of adequate controls also plague even the benchmark studies in the Qigong research field. Based on only four trials that lacked blinding, controls for experimental conditions, and statistical analysis, for example, a frequently referenced report of using bacteria as an <italic>in vitro </italic>target [<xref ref-type="bibr" rid="B8">8</xref>] concludes that external Qigong can stimulate or inhibit the growth of <italic>Escherichia coli</italic>. Another general weakness in the field is the lack of replication of experiments. A recent critical review that assessed a sample of 58 <italic>in vitro </italic>studies of Qigong in the Asian literature found no independent replications on any single model [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Our report presents the results of a collaborative effort between basic scientists and clinicians working with external Qigong practitioners in the United States and China to develop a rigorous <italic>in vitro </italic>protocol for evaluating potential direct effects of external Qigong. We chose a cell culture model known to be sensitive to conventional energy treatments (e.g., ionizing radiation) and sought an experimental design that would be easily reproducible in independent laboratories. Our experimental hypothesis was that experienced external Qigong practitioners could enhance the healthy growth of human cells under stressful culture conditions.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Overall study design</title><p>We conducted three independent studies of Qigong efficacy in two different geographic locations. All three studies incorporated rigorous experimental design, including (1) blinded observation and analysis of results, (2) randomized assignment to experimental groups, and (3) a nontreated control group (sham) to compare with each treatment group. In addition, we included an equal number of experiments comparing sham treatment to sham treatment (systematic negative controls) throughout the duration of the study to continually assess potential variability within the model system. All three studies used a well-established assay to measure cell proliferation in culture conditions in which a majority of the cells do not survive. The protocol was approved by the Institutional Review Board of the California Pacific Medical Center.</p><p>The first study was a pilot to evaluate the work of three Qigong practitioners who were in the San Francisco Bay Area at the time of the study. The setting was a basic science laboratory, but the room used for the experimental treatments was a nearby office. A scientist who was blinded as to the eventual status of the experiment (i.e., treatment by Qigong practitioner or sham treatment) brought randomly selected samples of cells to the treatment room. Ten minutes later, another scientist escorted a Qigong practitioner to the treatment room. The practitioners delivered Qigong treatment to the samples for 20 minutes. For each experiment, there was also a sham control session in which the cell cultures were handled identically except that no Qigong practitioner entered the treatment room. In this pilot study, a random number generator was used to decide which treatment, Qigong or sham, would be applied first. One of the practitioners performed two experiments; each of the other two performed a single experiment. Pilot experiments evaluating tumor cell killing or growth inhibition after Qigong treatment did not produce results that justified further investigation but experiments with normal cells did.</p><p>We designed the formal study to have 80% power to detect an apparent increased proliferation effect observed in cultures of normal human brain cells following Qigong treatment suggested in the pilot study. Power analysis of the results of the pilot study suggested that at least 14 Qigong/sham experiments would be needed for the formal study. This study was conducted at a basic science laboratory in Beijing, China, where more practitioners were readily available that we could verify were professionals of Traditional Chinese Medicine working within a Qigong department or division in a hospital. Three researchers traveled there to conduct the experiments and worked with nine practitioners from four traditional Chinese medicine institutions. Again, the room used for experimental treatment was a nearby office. The protocol was identical to that of the pilot study except that the sham always preceded the Qigong treatment. This variation in the protocol was at the request of the practitioners. Precedent in the Chinese literature indicates the possibility that residual Qi generated during a Qigong treatment session might linger in a treatment room and contaminate a subsequent sham treatment in the same room. To avoid such a "linger effect," the practitioners requested at the outset of the formal study that the sham treatments always precede the Qigong treatments. Nine practitioners in this study conducted a total of 14 experiments; five practitioners conducted two experiments each, and the remaining four conducted single experiments.</p><p>A third study was conducted the following year in Beijing to assess the reproducibility of the findings of the formal study. The protocol was identical to that of the formal study, except that the order of treatments (Qigong or sham) was randomized to eliminate any potential confounding of treatment and time. This time there were 30 experiments conducted at two laboratories and involving eight practitioners; three conducted three experiments, four conducted four experiments, and one conducted five experiments. Two of the eight practitioners had participated in the previous Beijing study. A unique feature of this study was the addition of a temperature probe inside each treatment box to measure temperature (to 1/100°C in 30-second intervals). This was added to address the question of whether the proximity of the Qigong practitioner to the cells could be influencing subsequent proliferation through the mechanism of increasing temperature.</p></sec><sec><title>Outcome measure and statistics</title><p>We assessed cell proliferation by quantifying the colony-forming efficiency (CFE) of normal brain cells growing in culture plates. The assessment of CFE measures a cell's ability to duplicate itself again and again, forming a colony, under various treatment conditions. The methodology, first developed in the 1950s [<xref ref-type="bibr" rid="B10">10</xref>], was considered by the 1970s to be the gold-standard assay in studies of <italic>in vitro </italic>sensitivity to therapeutic agents [<xref ref-type="bibr" rid="B11">11</xref>]. It remains a mainstay in the measurement of cell response <italic>in vitro </italic>[<xref ref-type="bibr" rid="B12">12</xref>]. In our studies, we counted colonies of 50 or more cells manually using a low-power stereoscope, and we documented images of each plate using a digital imaging system.</p><p>For each experiment, 100 cells were placed into each of 12 culture plates. The dependent variable was the ratio of the CFEs between cell cultures in two treatment boxes (six plates in box A and six plates in box B) for a given experiment. Variation within each experiment was measured and differences between experiments tested using Bartlett's test.</p><p>To find the mean, the CFEs for the six plates in each box were averaged following standard and previously published methods [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>]. This method reduces the variability of the CFE outcome measure, thus increasing the power to detect potential effects of treatment. The ratio of the averages (A/B) where then computed, giving a single data point for each experiment. Because averaging ratios can result in a skewed distribution, the statistics are presented as Log values. Unlike the mean Qigong/sham ratios, the Log values reflect whether a value is above or below chance expectation (CE), with negative values below CE and positive values above CE.</p></sec><sec><title>Cell culture</title><p>We used human astrocytes as the target of healing intentionality in these studies after evaluating these normal brain cells in parallel pilot experiments with glioma tumor cells, a common type of brain cancer. The glioma cells appeared to be less responsive to Qigong treatment. Based on a comparison of four experiments comparing Qigong treatment to sham treatment and four experiments comparing sham treatment to sham treatment (systematic negative controls), we found that the mean Qigong/sham CFE ratio for glioma cells was 1.047 (sd = 0.153) and the mean sham/sham ratio was 0.986 (sd = 0.107). Using a t-test to compare conditions yields t(6) = 0.651, which is an effect size only in the 0.02 – 0.25 range. Given the small effect size, we determined not to pursue this model. Parallel experiments with astrocytes showed a greater effect size (0.30 – 0.35) with the mean CFE ratio being 1.042 (sd = 0.046) for Qigong/sham samples and 0.972 (sd = 0.129). Comparing treatment conditions for astrocytes yields t(6) = 1.016, <italic>P </italic>= 0.349.</p><p>The astrocyte cell culture model is used widely and is responsive to conventional therapies utilizing emitted energy (i.e., ionizing radiation). The model can demonstrate both positive and negative responses (cell proliferation and cell death), depending on the treatment. Astrocytes were isolated by established methods [<xref ref-type="bibr" rid="B15">15</xref>] and were confirmed to be astrocytic by uniform staining with an anti-GFAP antibody (Beohringer Mannhiem). Cells were grown in a DME H-21-based medium with 10 ng/ml EGF, 10 nM hydrocortisone, 10 ng/ml biotin, ITS supplement, transferrin 50 μg/ml, biotin 10 ng/ml (all from Sigma), and 0.2% bovine pituitary extract (Clonetics). This was mixed with DME/10% fetal calf serum at a 10:2 ratio. Cells were passaged at confluence using trypsinization and were expanded to a large population size, aliquoted, and frozen viably for long-term storage. A fresh aliquot was thawed at the start of each experimental trial to ensure uniformity in the genetic profile of the target cells throughout the project.</p><p>Conditioning of the plastic surface of the culture plates with organic material is necessary to promote proliferation of experimental cells. We provided minimal conditioning (a small number of lethally irradiated "feeder" cells) so that as few as 1% of the cells survived and were able to form colonies under control conditions. Thus, the majority of the normal brain cells were not expected to survive the stress of the culture conditions.</p></sec><sec><title>Randomization</title><p>After one day of incubation, each cell culture plate was randomly assigned to (1) one of the two treatment boxes labeled A or B (an opaque plastic box with lid closed tightly) and (2) a specific position within its assigned box. A computer program written specifically for this task (using ZBasic for Macintosh) produced equally likely assignment of plates to boxes and positions. Assignment was constrained to assure that each box contained six cell culture plates. This program produced a printed diagram of box A and box B, with plate positions labeled by identification codes. Just before experimental treatments were to be delivered, a researcher followed the printed diagram to prepare the two boxes and placed them in the incubator to allow the CO<sub>2 </sub>concentration to equilibrate before the lid was closed. The boxes were then carried one after the other to a separate room to receive experimental treatments. We also assigned each cell culture plate randomly to a position in the incubator. A computer program written for the experiment (Microsoft Visual Basic 5.0) used a pseudo-random number generator to assign plates to positions on a grid marked on the incubator shelf. Each plate had equal likelihood of assignment to any incubator position. This allowed testing of whether placement in the incubator had any effect.</p></sec><sec><title>Qigong practitioners</title><p>One challenge in Qigong research is that there are numerous styles of Qigong practice and few practitioners have identical training. In our experience, practitioners often emphasize how different and unique their own style and ability is compared with that of others. They also emphasize how their ability may vary according to environmental factors, such as weather and stress. Because of such vast variance and the fact that there are no objective measures for the "amount" and "strength" of the Qi each practitioner emits each time (if that is indeed emitted), it is impossible to verifiably identify a group of practitioners who can give identical external Qigong treatment consistently. Since the goal of our study is to assess external Qigong treatment for patients, we looked for common denominators among the practitioners from a clinical standpoint. These are: 1) they claim that they can emit Qi externally for the purpose of treating physical illnesses, albeit with variations and through different styles, 2) they are professionals of Traditional Chinese Medicine working within a Qigong department or division in a hospital, 3) they have practiced such treatment regularly for at least five years, and 4) they have experience treating cancer patients. A total of 18 practitioners (with practice experience ranging from 5 to 17 years) participated in the three independent studies. We worked with practitioners at their convenience and were unable to balance the number of experimental sessions in which each participated. To avoid implicit endorsement, practitioners were compensated for their participation through honoraria and remain anonymous.</p></sec><sec><title>Qigong and sham treatments</title><p>External Qigong treatments by a Qigong practitioner typically involve "emission" of Qi with the intent to heal. In these studies, Qigong practitioners were instructed to try to direct emitted Qi toward cell cultures in an attempt to stimulate the growth of cell cultures inside an opaque plastic box. We described the cell culture system and the experimental hypothesis to the practitioners. For each experiment, one practitioner was escorted into a treatment room and observed to ensure that a minimum distance of 10 centimeters was maintained from the treatment box that had been placed on the laboratory bench. During a Qigong treatment, the practitioner performed a meditative practice similar to what would be done during a session with a patient. Some of the practitioners made gentle motions with their hands as if collecting or channeling invisible streams of "healing energy" toward the cells. Others stood motionless, holding a specific posture. The duration of one treatment session was 20 minutes, a typical period of time for treating a patient. A sham treatment involved all the same physical manipulations of the cell culture plates as during an experimental treatment, except that nobody entered the treatment room (the same room used for both types of treatments) during the 20-minute period.</p></sec><sec><title>Systematic negative controls</title><p>To incorporate systematic negative controls, we changed randomly between experiments in which we compared Qigong treatment to sham treatment and experiments in which we compared sham treatment to another sham treatment. By allowing quantitative assessment of potential systematic errors associated with the methodology used, systematic negative controls tested the method's accuracy and reliability throughout the experimental series [<xref ref-type="bibr" rid="B16">16</xref>]. Questions about the potential influence of variations in physical parameters between treatment sessions could therefore be addressed.</p></sec><sec><title>Blinding procedures</title><p>Experiments were conducted with blinding applied to each of the four scientists and the biostatistician involved, following previously reported methods [<xref ref-type="bibr" rid="B17">17</xref>]. Briefly, Scientist #1 handled the cell culture plates, labeled them with random identifying codes, and always placed them according to randomly assigned positions provided by Scientist #2. Scientist #1 would signal Scientist #3 via an electronic signaling system (a nonverbal signal) when samples were ready for treatment so that Scientist #3 could escort the practitioners in and out of the treatment room at the appropriate times. Thus, the scientist handling the cells knew which cells were inside treatment boxes A and B but did not know what treatments (i.e., Qigong or sham) were delivered. Likewise, the scientist aware of the treatment schedule did not know which cells were being treated at any given time (i.e., which cells were in the boxes). A fourth scientist was responsible for counting colonies in each sample using only the random identifying codes. For the studies conducted in Beijing, the cell culture plates were carried back to San Francisco after the colonies were fixed for counting. Each scientist sent blinding codes and data to an outside institute to be kept by an independent peer (a code keeper). A biostatistician, who was blinded to what treatment each group had received, conducted statistical analysis using only the code number of the culture plates.</p></sec></sec><sec><title>Results</title><p>The results show a trend toward increased cell proliferation in the treated samples in the pilot study (Qigong/sham CFE ratio > 1.0), a statistically significant trend of increased proliferation following Qigong treatment in the formal study, and a nonsignificant, slight increase in proliferation following Qigong treatment in the replication study. We pooled the results from all three studies to form summary statistics, including an overall <italic>t</italic>-test for significance. The mean for the pooled Qigong/sham data was >1.0 but was not statistically significant. The pooled sham/sham data was also not significantly different than 1.0. Table <xref ref-type="table" rid="T1">1</xref> summarizes the results of the three independent studies and the original data are provided as supplementary material [see <xref ref-type="supplementary-material" rid="S1">Additional file 1</xref>]. Figure <xref ref-type="fig" rid="F1">1</xref> plots the CFE ratios for each experiment.</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Summary of study results</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Experiment</td><td align="center">No. of Qigong Practitioners</td><td align="right">No. of Exps.</td><td align="right">Mean* CFE Ratio</td><td align="right">St Dev</td><td align="right"><italic>t</italic>-statistic</td><td align="right">2-sided p</td></tr></thead><tbody><tr><td align="left"><italic>Qigong/Sham</italic></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Pilot study</td><td align="center">3</td><td align="right">4</td><td align="right">1.041</td><td align="right">0.046</td><td align="right">1.845</td><td align="right">0.162</td></tr><tr><td align="left">Formal study</td><td align="center">9</td><td align="right">14</td><td align="right">1.099</td><td align="right">0.167</td><td align="right">2.333</td><td align="right">0.036</td></tr><tr><td align="left">Replication</td><td align="center">8</td><td align="right">30</td><td align="right">1.070</td><td align="right">0.534</td><td align="right">0.740</td><td align="right">0.465</td></tr><tr><td align="left"><bold>Pooled data</bold></td><td align="center"><bold>18</bold></td><td align="right"><bold>48</bold></td><td align="right"><bold>1.076</bold></td><td align="right"><bold>0.440</bold></td><td align="right"><bold>1.238</bold></td><td align="right"><bold>0.222</bold></td></tr><tr><td align="left"><italic>Sham/Sham</italic></td><td></td><td></td><td></td><td></td><td></td><td></td></tr><tr><td align="left">Pilot study</td><td align="center">0</td><td align="right">4</td><td align="right">0.966</td><td align="right">0.126</td><td align="right">-0.526</td><td align="right">0.635</td></tr><tr><td align="left">Formal study</td><td align="center">0</td><td align="right">14</td><td align="right">1.037</td><td align="right">0.155</td><td align="right">0.905</td><td align="right">0.382</td></tr><tr><td align="left">Replication</td><td align="center">0</td><td align="right">30</td><td align="right">0.955</td><td align="right">0.489</td><td align="right">-0.491</td><td align="right">0.627</td></tr><tr><td align="left"><bold>Pooled data</bold></td><td align="center"><bold>0</bold></td><td align="right"><bold>48</bold></td><td align="right"><bold>0.979</bold></td><td align="right"><bold>0.412</bold></td><td align="right"><bold>-0.347</bold></td><td align="right"><bold>0.730</bold></td></tr></tbody></table><table-wrap-foot><p>Mean and standard deviation based on logarithmic values.</p></table-wrap-foot></table-wrap><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Colony forming efficiency ratios for three studies. </bold>Natural Log Qigong/sham ratios are graphed for all experiments in the pilot study (left), the first formal study (middle) and the replication study (right). Qigong/sham ratios are shown as <italic>solid circles</italic>; sham/sham ratios are shown as <italic>open circles</italic>.</p></caption><graphic xlink:href="1472-6882-4-5-1"/></fig><p>A striking feature of these three studies is the increase in variability as we went from the small pilot study in San Francisco to the formal study in Beijing and then to the larger replication study in Beijing. This is likely due to the sub-optimal cell culture conditions at the Beijing laboratories (two sites for the replication study). Each increase in variability is highly statistically significant in the Qigong/sham experiments and also in the sham/sham experiments (<italic>P </italic>< 0.001 for each pair-wise variance comparison).</p><p>Measuring the temperature inside the treatment boxes (replication study only) revealed that the temperature-to-CFE relationship in the replication study is rather complex. Overall, temperature and CFE seem to correlate strongly (<italic>r</italic>[122] = 0.72, <italic>P </italic>< 0.0001). Upon closer inspection, however, this appears to be an artifact of large changes in both CFE and temperature across the consecutive days of the experiment. The within-day correlations between temperature and CFE are modest, although relatively consistent (mean <italic>r </italic>= 0.22). More importantly, neither the CFE nor the temperature is significantly related to the CFE ratio (Qigong/sham), either overall or on any given day. No significant temperature differences were found between A and B sessions. The temperature is thus unlikely to have influenced the results of the main hypothesis of this study. While the astrocytes are sensitive to ambient temperature, there is no indication that treated or untreated samples fared differently. Similarly, analysis of the randomly assigned position of cell culture plates in the incubator for the formal study showed no relation between incubator position and colony counts (<italic>P </italic>= 0.99).</p><p>We were curious whether the failure to find significant results in the replication study was because the new practitioners were less efficacious than the practitioners from the first formal study. After we analyzed the replication study data, comparing results from experiments involving practitioners who had also participated in the earlier formal study in Beijing (n = 9, mean CFE = 1.10, SD = 0.68) with those from experiments involving the new practitioners (n = 21, mean CFE = 1.08, SD = 0.49), we determined that this is unlikely. We tested for a difference between these two groups of practitioners based on a two-sided two-sample <italic>t</italic>-test and found no significant difference (<italic>P </italic>= 0.83). Another difference between the formal study and the replication study is that sham treatments always preceded Qigong treatments in the formal study, while the order of treatment was randomized in the replication study. This difference in protocols does not appear to account for the difference in outcomes of the two studies because there were no differences in the sham results, regardless of whether they preceded or followed Qigong treatment. Sham median counts were 6.6 (95% CI 1.6 to 11.5) in the formal study and 5.5 (95% CI 2.6 to 8.4) in the replication study when sham was first, and 6.8 in the replication study (95%CI 3.9 to 9.8) when sham was second. None of the pairwise differences are statistically significant based on a Wilcoxon rank sum test, which was used since there were outlier values.</p></sec><sec><title>Discussion</title><p>The systematic negative controls included in the protocol were helpful when considering the significance of an outlier in the data. One of the Qigong/sham experiments in the replication study yielded an exceptionally high CFE ratio (see Figure <xref ref-type="fig" rid="F1">1</xref>). If it were not for the information gained from the sham/sham experiments, this data point could easily be interpreted as an exceptional performance (i.e., stimulating cell proliferation) by a particular practitioner in that particular experiment. Evidence of intrinsic variability of the system of similar magnitude does not support this speculation, however. Two data points from sham/sham experiments were also outliers to a similar degree. Thus, the outlier in the Qigong/sham experiment falls within the range of variability associated with the experimental model. The variability of the model must yield to the laws of probability as long as there are sufficient numbers of experiments, careful randomization, and strict blinding to eliminate artifact and bias. In this series, the systematic negative controls provided an additional level of confidence in the validity of the results. Indeed, even three outliers in the data - all in the direction suggesting an influence of the Qigong treatment (one high Qi/sham CFE and two low sham/sham CFEs), did not sway the results into spurious significance.</p><p>Our report presents a protocol that can easily be replicated in independent laboratories to assess potential direct influences of energy medicine modalities. Following this protocol, we did not observe reproducible effects of external Qigong treatment on the colony-forming efficiency of normal astrocytes. Application of protocols of comparable rigor to other outcome measures (preferably with less variability) is necessary to address the question of the mechanisms of external Qigong. Likewise, exploration of other outcome measures that more closely approximate physiological conditions may also be warranted. If reproducible positive results are observed in other model systems, it will be important to include specific experimental conditions to control for the possibility that physical parameters associated with the proximity of a human body (practitioner) are sufficient to influence samples. Pheromones or other chemical signals, for example, might stimulate biological samples. The demonstration that single cells in mice respond to pheromones at concentrations below 10<sup>-11 </sup>molar [<xref ref-type="bibr" rid="B18">18</xref>] increases speculation along these lines. Examples of appropriate control conditions could include having a practitioner present without delivering treatment and having a nonpractitioner "confederate" present with or without mimicking the practitioner's behavior.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>GY and YQ conceived of the study and participated in its design, coordination and implementation. JS and MS participated in the design of the study. DM performed the statistical analysis. KA prepared samples and MR performed colony counts. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6882/4/5/prepub"/></p></sec><sec sec-type="supplementary-material"><title>Supplementary Material</title><supplementary-material content-type="local-data" id="S1"><caption><title>Additional File 1</title><p>Includes original data for all experiments: The supplementary table presents original data for all experiments in the three studies, including: 1) whether each experiment was comparing Qigong/sham or sham/sham, 2) the mean colony counts for each group of samples, 3) the ratio of these mean values and 4) the natural logarithm of these ratios.</p></caption><media xlink:href="1472-6882-4-5-S1.doc" mimetype="application" mime-subtype="msword"><caption><p>Click here for file</p></caption></media></supplementary-material></sec>
Assessment of single-dose benzodiazepines on insulin secretion, insulin sensitivity and glucose effectiveness in healthy volunteers: a double-blind, placebo-controlled, randomized cross-over trial [ISRCTN08745124]	<sec><title>Background</title><p>The present study aimed at investigating in healthy volunteers the effects of diazepam and clonazepam on beta-cell function, insulin sensitivity and glucose effectiveness based on the frequently sampled intravenous (0.5 gkg<sup>-1</sup>) glucose tolerance test with minimal-model analysis.</p></sec><sec sec-type="methods"><title>Methods</title><p>The study was designed as a double-blind, placebo-controlled, cross-over clinical trial. Diazepam (10 mg) and clonazepam (1 mg) were infused during 30 min to 15 male subjects with a mean age of 22 years (range: 20–29), after informed consent was given. Benzodiazepines were assayed by capillary gas chromatography with electron capture, insulin by radioimmunoassay and glucose by the enzymatic glucose oxidase method.</p></sec><sec><title>Results</title><p>Both benzodiazepines induced significant psychotropic effects. The acute insulin responses (AIR) were significantly and negatively correlated with the clonazepam plasma concentrations (r = -0.609, P < 0.05, n = 14). However, the mean AIR was not significantly different between the benzodiazepine-treated subjects and the controls. In addition, the parameters of glucose assimilation were significantly decreased as compared with placebo in the subgroup of 7 subjects with plasma clonazepam concentrations higher than 6.0 ng ml<sup>-1 </sup>(median and lower limit of effective therapeutic concentrations): 1.37 ± 0.3 <italic>versus </italic>2.84 ± 0.60 × 10<sup>-2</sup>min<sup>-1 </sup>(P = 0.028) for the coefficient of glucose tolerance (K<sub>g</sub>), 2.18 ± 0.29 <italic>versus </italic>3.71 ± 0.89 × 10<sup>-4</sup>μUml<sup>-1</sup>min<sup>-1 </sup>(P = 0.018) for insulin sensitivity (S<sub>i</sub>) and 1.80 ± 0.39 <italic>versus </italic>3.59 ± 0.71 × 10<sup>-2</sup>min<sup>-1 </sup>(P = 0.028) for glucose effectiveness at basal insulin (S<sub>g</sub>). These parameters were not significantly modified when diazepam was administered; plasma levels of this drug however, were below the effective therapeutic concentrations (300 ng ml<sup>-1</sup>) from min 15 after the end of the perfusion.</p></sec><sec><title>Conclusion</title><p>The present results suggest that a benzodiazepine, in particular clonazepam, may alter insulin secretion and insulin sensitivity after a single administration in healthy volunteers.</p></sec>	<contrib id="A1" equal-contrib="yes" corresp="yes" contrib-type="author"><name><surname>Chevassus</surname><given-names>Hugues</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Mourand</surname><given-names>Isabelle</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Molinier</surname><given-names>Nathalie</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Lacarelle</surname><given-names>Bruno</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Brun</surname><given-names>Jean-Frédéric</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Petit</surname><given-names>Pierre</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib>	BMC Clinical Pharmacology	<sec><title>Background</title><p>Benzodiazepine drugs are widely prescribed, mainly for their anxiolytic and sedative properties. In addition to these psychotropic actions, different metabolic effects have also been reported but few clinical trials on that topic are available.</p><p>A trend to increased glycemia without significant modification of insulinemia was shown after a single dose administration of diazepam in healthy volunteers [<xref ref-type="bibr" rid="B1">1</xref>]. Aggravation of hyperglycemia has been reported during benzodiazepine treatment in diabetic patients [<xref ref-type="bibr" rid="B2">2</xref>]. In patients treated with the imidazopyridine compound alpidem, which binds benzodiazepine receptors [<xref ref-type="bibr" rid="B3">3</xref>], we reported an alteration of glucose tolerance after one week administration [<xref ref-type="bibr" rid="B4">4</xref>]. Studies <italic>in vitro </italic>have shown that some benzodiazepines could affect insulin secretion differently according to the experimental model [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B6">6</xref>].</p><p>Weight gain, which is often related to insulin resistance, has also been claimed to occur in some patients treated with benzodiazepines. Different benzodiazepines have different effects on body weight gain in rats, suggesting that different subtypes of benzodiazepine receptors are involved [<xref ref-type="bibr" rid="B7">7</xref>].</p><p>In order to further explore the potential effects of benzodiazepines on beta-cell function and insulin sensitivity, we performed in healthy volunteers the frequently sampled intravenous glucose tolerance test with minimal-model analysis according to Bergman <italic>et al</italic>. [<xref ref-type="bibr" rid="B8">8</xref>]. The single-dose effects of two benzodiazepines, diazepam and clonazepam, with different affinity and selectivity for central and peripheral receptors were investigated.</p></sec><sec sec-type="methods"><title>Methods</title><p>Fifteen healthy male volunteers, aged 20–29 years (mean 22.5 years), were included in this study. They underwent a full medical examination with electrocardiogram and standard biological screening before entry into the study. All subjects were in good physical condition and of normal weight (mean: 68.6 kg; range: 55–81) and body mass index (mean: 22.3 kg/m<sup>2</sup>; range: 20.1–25.0). They had no previous history of significant disease and satisfied the inclusion/exclusion criteria. All volunteers gave written consent to participate after being informed of the aim of the research, its experimental protocol and potential risks. The study was approved by the institutional Ethics Committee of Montpellier and has been conducted in accordance with the Helsinki declaration and the European guideline for « Good Clinical Practice ».</p><p>The study was designed as a double-blind, three-way placebo-controlled cross-over investigation of diazepam, a central and peripheral benzodiazepine receptor ligand [<xref ref-type="bibr" rid="B9">9</xref>], clonazepam, a selective central receptor ligand [<xref ref-type="bibr" rid="B9">9</xref>], and placebo. Subjects were randomly assigned to one of the three treatment groups to control for potential order effect. The three treatments were single intravenous doses of diazepam (10 mg), clonazepam (1 mg) or placebo. The three treatment administrations were separated by a 7-day washout period.</p><p>In the morning of each test-day, subjects were admitted to the Clinical Investigation Center, in fasting conditions for at least 12 hours, namely from the evening before. The subjects remained in the supine position and a venous catheter was inserted in each forearm, one for the administration of the drugs and glucose, the other for blood sampling. The drugs in 125 mL saline solution or the saline solution alone (placebo) were administered during 30 min (between minutes -45 and -15). The treatments were prepared by the research nurse according to the randomization form. Fifteen minutes after the end of the perfusion, the frequently sampled intravenous glucose tolerance test (FSIVGTT) was started, with a glucose load of 0.5 g/kg in 3 min and a bolus of 0.03 U/kg insulin just before minute 20. Blood samples were taken at times -45, -15 and 0, then at 1, 3, 4, 6, 8, 10, 15, 19, 20, 22, 24, 30, 41, 70, 90 and 180 minutes after glucose injection for the measurement of plasma glucose and serum insulin concentrations. Additional samples were taken at times -45, -15, 0, then at 15, 30, 60, 120 and 180 minutes after glucose injection for the measurement of plasma diazepam, N-desmethyldiazepam, clonazepam and cortisol.</p><p>Furthermore, the subjects' mood and feelings were self-rated before (time -45 min) and after the infusion (time -15 min), and then at times 70 and 180 minutes, using the 49-item questionnaire from the Addiction Research Center Inventory (ARCI) and the Norris series of 16 bipolar visual analogue scales (VAS). At the end of the test day, subjects were discharged after a last clinical evaluation.</p><p>After immediate centrifugation of the blood samples (1500 g, 10 min), plasma or serum was separated and frozen at -20°C. Plasma glucose levels were assayed by the glucose oxidase method derived from Trinder [<xref ref-type="bibr" rid="B10">10</xref>]. Serum insulin and plasma cortisol were measured by radioimmunoassay using the marketed immunoradiometric kit INSIK-5 purchased from DiaSorin (Saluggia, Italy). Diazepam, N-desmethyldiazepam and clonazepam were assayed by a capillary gas chromatography technique with electron capture, adapted from a previously published method [<xref ref-type="bibr" rid="B11">11</xref>]. Briefly, benzodiazepines were extracted from plasma at pH = 9.0 by ethyl ether. GLC analysis was performed on a HP-5 (crosslinked 5 % PH ME siloxane) 30 m × 0.32 mm × 0.25 μm capillary column (Agilent) at an oven temperature from 170 to 250°C (2°C/min). The intra and inter-assay variability ranged from 4.7 % to 9.9 % for clonazepam, 2.7 % to 7.4 % for diazepam and 5.4 % to 9.7 % for N-desmethyldiazepam for three quality control concentrations.</p><p>Acute insulin response (AIR) was used as an index of the first-phase insulin secretion and was calculated as the mean of the serum insulin concentrations at times 1, 3, 4, 6, 8 and 10 minutes.</p><p>The least-square slope of the log of the absolute glucose concentration between 4 and 19 minutes after the glucose bolus was used as an index of glucose tolerance (K<sub>g</sub>).</p><p>Minimal-model analysis of the FSIVGTT was performed according to the method previously reported by Bergman <italic>et al</italic>. [<xref ref-type="bibr" rid="B8">8</xref>] and Yang <italic>et al</italic>. [<xref ref-type="bibr" rid="B12">12</xref>], using the software TISPAG from the Department of Clinical Physiology, Lapeyronie University Hospital, (Montpellier, France), which uses a nonlinear least-square estimation [<xref ref-type="bibr" rid="B13">13</xref>]. This program produced the values for insulin sensitivity (S<sub>i</sub>) and glucose effectiveness (S<sub>g</sub>). S<sub>i </sub>and S<sub>g </sub>are calculated from the equations dG(t)/dt = -[p1 + X(t)]G(t) + p1Gb, G(0) = Go, dX(t)/dt = -p2X(t) + p3 [I(t) - Ib], and X(0) = 0, where G(t) and I(t) are plasma glucose and insulin concentrations, X(t) is the insulin in a compartment remote from plasma (insulin action), and p1 to p3 are model parameters. Go is the glucose concentration that would be obtained immediately after injection if there was instantaneous mixing in the extracellular fluid compartment. Gb and Ib are basal values of glucose and insulin. Parameter p1 represents S<sub>g</sub>, i.e. the fractional disappearance rate of glucose independent of any insulin response, and p3 and p2 determine the kinetics of insulin transport, respectively, into and out the remote insulin compartment where insulin action is expressed. S<sub>i </sub>is an index of the influence of plasma insulin to change the glucose effect <italic>per se </italic>on glucose concentration. Thus, S<sub>i </sub>is equal to -p3/p2.</p><p>S<sub>g </sub>was divided into its two components: the contribution of hyperglycemia <italic>per se </italic>to tissue glucose utilization and the effect of basal insulin on glucose uptake. The basal insulin component of S<sub>g </sub>is BIE and can be calculated as the product of basal insulin Ib and S<sub>i</sub>: BIE = Ib × S<sub>i</sub>. Thus, the contribution of non-insulin-dependent glucose uptake (GEZI) to glucose uptake is the difference between total S<sub>g </sub>and the BIE: GEZI = S<sub>g </sub>- (Ib × S<sub>i</sub>).</p><p>The sample size was calculated from clinical data on AIR obtained in a control sample after FSIVGTT, with the 2-sided hypothesis of a difference between diazepam and placebo or clonazepam and placebo of at least 35%; the study power was set at 90% and the significance threshold at 5%. All the results are given as means ± standard error (s.e. mean). Where necessary, 95% confidence intervals on differences (95% CI) are given. The time courses of mean cortisol concentrations were compared using repeated measures ANOVA. Other kinetics data were integrated in areas under the curves (AUC), calculated including baseline by the trapezoidal rule using the Systat<sup>® </sup>6.0 software for Windows<sup>®</sup>. AUC means were compared using the nonparametric Friedman test for related data. Spearman correlations were performed between AIR and diazepam or clonazepam concentrations. <italic>Post hoc </italic>and subgroups analysis were performed where necessary using the paired <italic>t</italic>-test for cortisol and the Wilcoxon rank-sum test for other paired data. The Systat<sup>® </sup>6.0 software for Windows<sup>® </sup>was also used for the statistical analysis.</p></sec><sec><title>Results</title><p>Intravenous administration of 10 mg diazepam during 30 min yielded to plasma concentrations peaking at 370.1 ± 22.6 ngml<sup>-1 </sup>at time -15 min, just at the end of the infusion. The concentrations rapidly decreased to 229.7 ± 12.6 ngml<sup>-1 </sup>at time 0, before glucose administration (Figure <xref ref-type="fig" rid="F1">1</xref>, A). The mean concentrations of N-desmethyldiazepam ranged between 3.2 ± 0.9 ngml<sup>-1 </sup>at min -15 and 9.1 ± 1.6 ngml<sup>-1 </sup>at min 180. Intravenous infusion of 1 mg clonazepam over 30 min led to a mean maximal concentration of 15.1 ± 2.0 ngml<sup>-1 </sup>at min -15 (Figure <xref ref-type="fig" rid="F1">1</xref>, B); the mean concentrations at times 0 and 15 minutes were 6.6 ± 0.5 and 4.9 ± 0.4 ngml<sup>-1</sup>, respectively.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p>Plasma concentrations after diazepam and clonazepam infusion. A Time course of plasma diazepam and N-desmethyldiazepam concentrations after 10 mg diazepam infusion (during 30 min from time -45 min to time -15 min) during FSIVGTT (0.5 g/kg glucose load at time 0 min). ● diazepam; ○ N-desmethyldiazepam. B Time course of plasma clonazepam concentration after 1 mg clonazepam infusion (during 30 min from time -45 min to time -15 min) during FSIVGTT (0.5 g/kg glucose load at time 0 min). ◆ clonazepam.</p></caption><graphic xlink:href="1472-6904-4-3-1"/></fig><p>Concerning the psychotropic effects of the drugs, the pentobarbital-chlorpromazine-alcohol group (PCAG) scores of the ARCI, which reflect sedation, were significantly different with treatment (P < 0.001, Figure <xref ref-type="fig" rid="F2">2</xref>); a significant increase was observed for diazepam (+363% <italic>vs </italic>placebo, 95% CI: 214.0, 820.7, n = 15, P = 0.004) and clonazepam (+593% <italic>vs </italic>placebo, 95% CI: 511.2, 1146.3, n = 14, P = 0.001). The benzedrine group (BG) scores, which explore stimulation, were also significantly different with treatment (P < 0.017, Figure <xref ref-type="fig" rid="F2">2</xref>); a significant decrease was observed for clonazepam (-153% <italic>vs </italic>placebo, 95% CI: -598.2, -170.3, n = 14, P = 0.004). Concerning the other factors of the ARCI, the scores of the morphine-benzedrine group (MBG), assessing euphoria, of the LSD group, assessing dysphoric and psychotomimetic changes, and of the amphetamine group, assessing amphetamine-like effects were not significantly modified with treatment (not shown). On the VAS, the self-rating of asthenia-fatigue (factor 1) scores were also significantly modified with treatment (P = 0.002), with a similar time-course as the PCAG score of the ARCI (not shown); a significant increase was observed for diazepam (+42.3% <italic>vs </italic>placebo, 95% CI: 899.1, 4501.6, n = 15, P = 0.015) and clonazepam (+70.3% <italic>vs </italic>placebo, 95% CI: 2402.2, 6516.4, n = 14, P = 0.002).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p>Time course of PCAG- and BG-scores of the ARCI after 10 mg diazepam, 1 mg clonazepam or placebo infusion (during 30 min from time -45 min to time -15 min) during FSIVGTT (0.5 g/kg glucose load at time 0 min). ● PCAG scores after diazepam infusion; ◆ PCAG scores after clonazepam infusion; ▲ PCAG scores after placebo infusion; ○ BG scores after diazepam infusion; ◇ BG scores after clonazepam infusion; △ BG scores after placebo infusion.</p></caption><graphic xlink:href="1472-6904-4-3-2"/></fig><p>In addition, there was a significant difference (P = 0.028) in the time course of plasma cortisol concentrations between the different groups (Figure <xref ref-type="fig" rid="F3">3</xref>); a significant reduction was observed at time 60 min for diazepam (-33.3% <italic>vs </italic>placebo, 95% CI: -130.4, -38.3, n = 15, P = 0.002) and clonazepam (-24.0% <italic>vs </italic>placebo, 95% CI: -119.0, -2.6, n = 15, P = 0.042).</p><fig position="float" id="F3"><label>Figure 3</label><caption><p>Time course of plasma cortisol after 10 mg diazepam, 1 mg clonazepam or placebo infusion (during 30 min from time -45 min to time -15 min) during FSIVGTT (0.5 g/kg glucose load at time 0 min). ● diazepam; ◆ clonazepam; ▲ placebo; * P < 0.05 for diazepam <italic>vs </italic>placebo and clonazepam <italic>vs </italic>placebo.</p></caption><graphic xlink:href="1472-6904-4-3-3"/></fig><p>As shown in figure <xref ref-type="fig" rid="F4">4</xref>, the acute insulin responses (AIR) were significantly and negatively correlated with the plasma concentrations of clonazepam (r = -0.609, P < 0.05, n = 14) at time 0 minute, just before glucose injection. However, the mean AIR were not significantly different between the benzodiazepine-treated subjects and the controls (Table <xref ref-type="table" rid="T1">1</xref>).</p><fig position="float" id="F4"><label>Figure 4</label><caption><p>Correlation of acute insulin response (AIR), calculated from serum insulin concentrations obtained during the 10 first minutes after glucose injection, with the plasma concentrations of clonazepam obtained at time 0 min, just before glucose injection (r' = -0.609, P < 0.05, n = 14).</p></caption><graphic xlink:href="1472-6904-4-3-4"/></fig><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Values of the parameters of insulin secretion, glucose tolerance, insulin sensitivity and glucose effectiveness according to the different treatment-groups.</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left">Treatment</td><td align="left">AIR (μUml<sup>-1</sup>)</td><td align="left">K<sub>g </sub>(× 10<sup>-2</sup>min<sup>-1</sup>)<break/></td><td align="left">S<sub>i </sub>(× 10<sup>-4</sup>μUml<sup>-1</sup>min<sup>-1</sup>)</td><td align="left">S<sub>g </sub>(× 10<sup>-2</sup>min<sup>-1</sup>)</td></tr></thead><tbody><tr><td align="left">Diazepam</td><td align="left">52.9 ± 13.7</td><td align="left">2.27 ± 0.26</td><td align="left">3.52 ± 0.52</td><td align="left">2.98 ± 0.30</td></tr><tr><td align="left">Clonazepam</td><td align="left">51.2 ± 8.6</td><td align="left">1.78 ± 0.37</td><td align="left">3.70 ± 0.79</td><td align="left">2.39 ± 0.48</td></tr><tr><td align="left">Placebo</td><td align="left">53.9 ± 10.7</td><td align="left">2.65 ± 0.31</td><td align="left">3.50 ± 0.47</td><td align="left">3.42 ± 0.37</td></tr></tbody></table></table-wrap><p>Overall, glucose tolerance, insulin sensitivity and glucose effectiveness were not significantly different between the benzodiazepine-treated subjects and the controls (Table <xref ref-type="table" rid="T1">1</xref>). However, as shown in figure <xref ref-type="fig" rid="F5">5</xref>, a concentration-dependent effect could be observed with clonazepam. Indeed, in the subgroup of 7 subjects with plasma clonazepam concentrations at time 0 minute higher than 6.0 ngml<sup>-1 </sup>(median and lower limit of effective therapeutic concentrations), each of these parameters was significantly decreased as compared with placebo: -51.8% (95% CI: -2.58, -0.37 × 10<sup>-2</sup>min<sup>-1</sup>, P = 0.028) for the coefficient of glucose tolerance (K<sub>g</sub>), -41.2% (95% CI: -3.04, -0.02 × 10<sup>-4</sup>μUml<sup>-1</sup>min<sup>-1</sup>, P = 0.018) for insulin sensitivity (S<sub>i</sub>) and -49.9% (95% CI: -3.07, -0.51 × 10<sup>-2</sup>min<sup>-1</sup>, P = 0.028) for glucose effectiveness at basal insulin (S<sub>g</sub>). Furthermore, the two components of S<sub>g </sub>were significantly decreased with clonazepam <italic>vs </italic>placebo in the subgroup of subjects with the highest plasma clonazepam concentrations at time 0 minute: -38.1% (95% CI: -0.18, -0.03 × 10<sup>-2</sup>min<sup>-1</sup>, P = 0.018) for basal insulin effctiveness (BIE) and -50.9% (95% CI: -2.97, -0.40 × 10<sup>-2</sup>min<sup>-1</sup>, P = 0.028) for glucose effectiveness at zero insulin (GEZI).</p><fig position="float" id="F5"><label>Figure 5</label><caption><p>Effects of clonazepam on glucose tolerance, insulin sensitivity and glucose effectiveness. A Effect of clonazepam on glucose tolerance (K<sub>g</sub>) according to plasma clonazepam concentration. <italic>White bars</italic>, controls; <italic>Striped bars</italic>, treated; * P < 0.05 <italic>vs </italic>placebo. B Effect of clonazepam on insulin sensitivity (S<sub>i</sub>) according to plasma clonazepam concentration. <italic>White bars</italic>, controls; <italic>Striped bars</italic>, treated; * P < 0.05 <italic>vs </italic>placebo. C Effect of clonazepam on glucose effectiveness at basal insulin (S<sub>g</sub>) according to plasma clonazepam concentration.<italic>White bars</italic>, controls; <italic>Striped bars</italic>, treated; * P < 0.05 <italic>vs </italic>placebo.</p></caption><graphic xlink:href="1472-6904-4-3-5"/></fig></sec><sec><title>Discussion</title><p>The present study investigated the effects of a single-dose administration of different benzodiazepines on β-cell function and insulin sensitivity in healthy volunteers.</p><p>Intravenous administration of 10 mg diazepam or 1 mg clonazepam led to peak concentrations at the end of the infusion within the standard effective plasma concentrations of these drugs, namely 300–400 ngml<sup>-1 </sup>for diazepam and 5–70 ngml<sup>-1 </sup>for clonazepam [<xref ref-type="bibr" rid="B14">14</xref>]. However, whereas the plasma levels of diazepam rapidely decreased below the standard effective concentrations, the mean concentrations of clonazepam further remained around the lower limit of effective concentrations.</p><p>Both diazepam and clonazepam induced significant psychological effects, as judged by the PCAG and BG scores of the ARCI and the asthenia-fatigue factor of the VAS, which is in accordance with previously reported data on the subjective effects induced by benzodiazepines [<xref ref-type="bibr" rid="B15">15</xref>]. These effects were still significant at time 70 minutes, ie approximately one and a half hour after the end of the infusion, although at that time, the PCAG and VAS factor 1 scores were also increased in the placebo group. This latter observation may be the result of the hypoglycemia measured at that time (2.4 ± 0.1 mmoll<sup>-1</sup>) and resulting from the insulin bolus, since acute hypoglycemia has been reported to affect brain function and activation [<xref ref-type="bibr" rid="B16">16</xref>]. On the other hand, the treatment by either benzodiazepine also induced a significant biological effect on plasma cortisol concentrations, which were decreased in the benzodiazepine-treated groups as compared with controls. This result is in accordance with previous studies in the literature [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>], although not all studies have reported that a single dose of benzodiazepine affects cortisol secretion in healthy volunteers [<xref ref-type="bibr" rid="B19">19</xref>]. This effect of benzodiazepines on cortisol is in agreement with the inhibiting action of gabaergic agonists on the corticotropin releasing hormone secretion, as observed <italic>in vitro </italic>in rats [<xref ref-type="bibr" rid="B20">20</xref>]. It can also be observed that there is a relative increase in plasma cortisol concentrations following the lowest glucose concentrations (at time 70 minute), and that this cortisol response to insulin-induced hypoglycemia is not affected by either benzodiazepine, in accordance with some [<xref ref-type="bibr" rid="B21">21</xref>] but not all previously reported data [<xref ref-type="bibr" rid="B22">22</xref>].</p><p>Considering the metabolic effects, the present study shows that clonazepam is able to reduce the acute insulin response to a glucose challenge in healthy volunteers, in a concentration-dependent manner. This result is in accordance with previously reported data showing that Diazepam Binding Inhibitor (DBI), which is an endogenous ligand for both central and peripheral benzodiazepine receptors, dose-dependently inhibited glucose-induced insulin secretion <italic>in vitro </italic>in rat isolated islets [<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B24">24</xref>]. In the same way, it supports previous observations that benzodiazepines or related compounds alter glucose tolerance <italic>in vivo </italic>in the rat [<xref ref-type="bibr" rid="B25">25</xref>] or in patients [<xref ref-type="bibr" rid="B4">4</xref>]. Other experiments however have shown <italic>in vitro </italic>that drugs acting specifically at peripheral benzodiazepine receptors inhibit glucose-induced insulin secretion [<xref ref-type="bibr" rid="B5">5</xref>] by reducing oxidative metabolism [<xref ref-type="bibr" rid="B26">26</xref>], whereas clonazepam, which binds very weakly to peripheral sites [<xref ref-type="bibr" rid="B27">27</xref>], is ineffective [<xref ref-type="bibr" rid="B5">5</xref>]. The peripheral-type receptor agonist, 4'-chlordiazepam, was also reported to increase glycaemia after intraperitoneal administration in the mice [<xref ref-type="bibr" rid="B28">28</xref>]. These discrepant data concerning clonazepam may be due to species variations or the result of an indirect central effect of the drug in our study. At the concentrations obtained with the single dose administered, there was no overall effect of clonazepam on the acute insulin response according to the different groups. Diazepam was ineffective as well, which may be the result of too low plasma concentrations over time. It is noteworthy to observe that both drugs however have significant psychotropic effects as well as effects on cortisol, even after 75 min after the end of the infusion, suggesting differential potencies relative to these effects.</p><p>On the other hand, this study shows for the first time that a benzodiazepine can reduce both insulin sensitivity (Si) and non-insulin-mediated glucose disposal (Sg and its components), provided that its concentration is above the admitted therapeutic threshold. Apart from the peak concentrations, this was especially the case for clonazepam but not for diazepam. There is no clear explanation for that observation. However, since insulin sensitivity and non-insulin mediated glucose disposal have been shown to increase even shortly after muscular activity [<xref ref-type="bibr" rid="B13">13</xref>], it can be tentatively speculated that the well-known muscle relaxant action of the benzodiazepine may alter insulin sensitivity and glucose disposal.</p></sec><sec><title>Conclusions</title><p>Even if clinical metabolic events related to benzodiazepines are uncommon, the results of the present study on both insulin secretion and insulin action prompt to suggest that, on the long term, these drugs may be a potential risk factor for glycemic dysregulation.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Author's contributions</title><p>HC participated in the design of the study, managed the research, writed the standard operating procedures, handled data, conducted statistical analyses and participated in drafting the manuscript. IM collected informed consents, supervised selection procedures and clinical investigations, reported verified data on the CRF. NM participated in clinical tasks described above. BL carried out benzodiazepines assays. JFB computed glucose tolerance, insulin secretion and insulin sentivity indexes. PP conceived of the study and coordinated the data analysis and the draft writing. All authors read and approved the final manuscript.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6904/4/3/prepub"/></p></sec>
Observed-predicted length of stay for an acute psychiatric department, as an indicator of inpatient care inefficiencies. Retrospective case-series study.	<sec><title>Background</title><p>Length of stay (LOS) is an important indicator of efficiency for inpatient care but it does not achieve an adequate performance if it is not adjusted for the case mix of the patients hospitalized during the period considered. After two similar studies for Internal Medicine and Surgery respectively, the aims of the present study were to search for Length of Stay (LOS) predictors in an acute psychiatric department and to assess the performance of the difference: observed-predicted length of stay, as an indicator of inpatient care inefficiencies.</p></sec><sec sec-type="methods"><title>Methods</title><p>Retrospective case-series of patients discharged during 1999 from the Psychiatric Department from General Hospital "Hermanos Ameijeiras" in Havana, Cuba. The 374 eligible medical records were randomly split into two groups of 187 each. We derived the function for estimating the predicted LOS within the first group. Possible predictors were: age; sex; place of residence; diagnosis, use of electroconvulsive therapy; co morbidities; symptoms at admission, medications, marital status, and response to treatment. LOS was the dependent variable. A thorough exam of the patients' records was the basis to assess the capacity of the function for detecting inefficiency problems, within the second group.</p></sec><sec><title>Results</title><p>The function explained 37% of LOS variation. The strongest influence on LOS came from: age (p = 0.002), response to treatment (p < 0.0001), the dummy for personality disorders (p = 0.01), ECT therapy (p = 0.003), factor for sexual and/or eating symptoms (p = 0.003) and factor for psychotic symptoms (p = 0.025). Mean observed LOS is 2 days higher than predicted for the group of records with inefficient care, whereas for the group with acceptable efficiency, observed mean LOS was 4 days lower than predicted. The area under the ROC curve for detecting inefficiencies was 69%</p></sec><sec><title>Conclusions</title><p>This study demonstrates the importance of possible predictors of LOS, in an acute care Psychiatric department. The proposed indicator can be readily used to detect inefficiencies.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Jiménez</surname><given-names>Rosa E</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" contrib-type="author"><name><surname>Lam</surname><given-names>Rosa M</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Marot</surname><given-names>Milagros</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Delgado</surname><given-names>Ariel</given-names></name><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib>	BMC Health Services Research	<sec><title>Background</title><p>Today there is a growing interest in improving quality and efficiency of health care to the maximum, a fact which highlights the necessity of good indicators of quality and efficiency of health care.</p><p>Length of stay (LOS) has been repeatedly used as an indicator of efficiency for inpatient care, probably due to its clear meaning as one of the main sources of hospital costs and because LOS can be also deemed an indicator of quality [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. However, LOS for a certain period and facility, is not a useful basis for meaningful comparisons unless it is adjusted for the case mix of patients hospitalized during the period considered. This process is called "risk adjustment" and is thoroughly described and discussed in the book edited by Iezzoni [<xref ref-type="bibr" rid="B4">4</xref>].</p><p>Allegedly the ideal way of adjustment should be based on the difference between LOS a patient should require–provided the attention received was efficient–, predicted LOS, and the actual one, observed LOS. As a continuous variable, LOS (or a proper transformation) may be modeled by means of a linear regression approach [<xref ref-type="bibr" rid="B5">5</xref>]. If an adequate model is found, the difference between observed LOS (OLOS) and predicted LOS (PLOS) could be a proper efficiency indicator. We already evaluated such an indicator for Internal Medicine and Surgery departments with fair results [<xref ref-type="bibr" rid="B6">6</xref>].</p><p>As a matter of fact, several studies, some of them recently performed, have looked for LOS predictors in psychiatric departments with similar aims [<xref ref-type="bibr" rid="B7">7</xref>-<xref ref-type="bibr" rid="B19">19</xref>].</p><p>Diagnosis, severity of illness, age, sex, physical co morbidities, treatment issues and psychosocial characteristics have been already confirmed as LOS determinants with more or less strength across the above referred studies.</p><p>The present work entails two related and successive aims. Firstly, the search for an appropriate function to predict the optimal LOS for an inpatient in an acute psychiatric ward, according to his or her characteristics, and, secondly, explore the ability of the difference <italic>Observed-Predicted LOS </italic>to detect inefficient care.</p></sec><sec sec-type="methods"><title>Methods</title><p>The study is basically a retrospective case-series study. It is mainly descriptive although some hypothesis testing has been performed during the derivation of the function.</p><sec><title>Setting</title><p>Information was collected from the clinical records of patients discharged from the Psychiatric Department in "Hermanos Ameijeiras" General Hospital in Havana, during 1999. The hospital is a government funded and public facility, it provides secondary and tertiary medical attention within all clinical and surgical specialities for adults except Obstetrics. The Psychiatry department comprises 46 beds, seven of which are reserved for a one week anti alcoholic addiction treatment. The remaining 39 beds are used for regular hospitalized patients admitted from three sources: 1) outpatient attention in the hospital, 2) the outpatient facilities within the hospital's catchment area or 3) the emergency department in the hospital.</p></sec><sec><title>Data retrieval</title><p>Included records belonged to new patients or known psychiatric patients in an acute phase of their illness. Excluded records were from: 3 self requested discharges, 2 patients included in research protocols affecting LOS, 2 patients transferred from other hospitals, 4 patients admitted for alcoholism treatment with pre-established LOS, 4 not concluded for unknown reasons (possibly self requests not stated), 2 patients escaped from the ward, and 2 patients in which the final main diagnosis was not psychiatric. For 20 patients who had more than one admission within the period only the last one was considered.</p><p>The 374 clinical records left available for our study were split randomly into two groups of 187 each. The first group was employed to derive the optimal function to estimate LOS. In the second group, we evaluated the capacity of the function to detect inefficiency problems during their stay. Thus, in both groups we obtained information from each patient record about the following variables allegedly affecting LOS: age, sex; place of residence, marital status, main diagnosis, administered medications, use of electroconvulsive therapy, co morbidities, response to treatment and symptoms at admission. Symptoms were included to account for the patient's severity of illness at admission since there is no regular Severity Index recorded for patients in this Department. Categories of all the variables, except symptoms, are displayed and detailed in Table <xref ref-type="table" rid="T1">1</xref>. The list of all symptoms and their categories are displayed in Table <xref ref-type="table" rid="T2">2</xref>. LOS was expressed in days from admission to discharge.</p></sec><sec><title>Data for validation</title><p>In the second group each record was thoroughly examined looking for sources of inefficient care that could be retrieved from the record, namely delays due to: a) more than 2 days between the indication and the realization of laboratory tests, b) more than 4 days between the realization of laboratory tests and results return from the corresponding laboratory, c) more than 2 days between admission and diagnosis discussion (a feature of all clinical records in the hospital that should be done within 48 hours after admission), d) more than 2 days for interconsultations with another specialist within the hospital, e) more than 3 weekend leaves and e) more than 4 days for prescribed leaves. Records were then classified as: reflecting <italic>acceptable efficient care </italic>if none of the mentioned situations were found in the record, or otherwise as <italic>care with efficiency problems</italic>. This assessment and classification was made by one of the authors (RML) blindly regarding the difference <italic>Observed-Predicted LOS </italic>(OLOS-PLOS). Doubts were discussed with another author (REJ) until agreement.</p><p>Predicted LOS was obtained for each patient via the function derived with the first study group and the differences OLOS–PLOS were obtained at the end of the study, when all the information was ready for statistical processing.</p></sec><sec><title>Statistical Analysis</title><p>The whole group of 374 records was firstly described (mean, standard deviation and median of LOS) within the categories of the different variables. With the first group of 187 records, a Multiple Linear Regression model was applied for assessing the independent influence of each variable on LOS and appraising the possibility of obtaining the predicted LOS. An initial exploration of LOS distribution in this group showed a right asymmetry suggesting the natural logarithm of LOS as the dependent variable. Principal Component Analysis was applied to reduce 19 symptoms to 8 factors that explained 61% of symptom variation (Table <xref ref-type="table" rid="T3">3</xref>). The regression function was derived with the variables and the 8 factors–after a Varimax rotation–in place of symptoms [<xref ref-type="bibr" rid="B20">20</xref>]. Thus, the following explanatory (independent) variables were included in the function: age as quantitative; sex, marital status, response to treatment and electroconvulsive therapy, as binary; co morbidities, administered drugs, diagnosis and place of residence as dummy variables; and the 8 factors (principal components) accounting for symptoms at admission. The final function was thus adjusted with 185 patients (after eliminating two outliers with standardized residuals higher than 3) and 24 variables. A determination coefficient (R<sup>2</sup>) of 0.374 was obtained and considered acceptable for the next step.</p><p>The estimated function was then used to obtain the predicted LOS for each patient in the second group (187 patients). We calculated for each patient in this group its score for each of the principal components with the <italic>Factor Score Coefficient Matrix </italic>obtained with the first group of records. The difference OLOS-PLOS was also obtained for each patient in this group and the association between these differences and the classification group, according to type of attention, evaluated with one way ANOVA. Finally, an ROC curve was obtained to evaluate the capacity of the new indicator (OLOS-PLOS) to detect records with inefficiency problems. The area under the curve was the global measure of the indicator performance. Statistical analysis was performed using SPSS Version 10.0.</p><p>The Ethics for Research Committee of Hospital "Hermanos Ameijeiras" approved the research protocol provided the authors maintain the confidentiality of data retrieved from clinical records. Only two of the authors (RML and MM, both medical doctors) worked directly with the records. The identity of the patients could not be identified in the database for statistical analysis.</p></sec></sec><sec><title>Results</title><sec><title>Sample description</title><p>Table <xref ref-type="table" rid="T1">1</xref> shows the main description of all variables in the whole group of medical records. The number of patients is fairly high for all categories. Higher mean LOS was found for patients receiving ECT during their stay and those with a delayed response to treatment.</p></sec><sec><title>Symptoms principal components</title><p>Table <xref ref-type="table" rid="T3">3</xref> displays the rotated component matrix for the symptoms. Each number in the table represents the correlation between the particular symptom and the rotated factor. Though it is not the aim of the study to deepen into the internal structure of the group of symptoms, it can be considered a fine factor solution since each symptom is only highly correlated with one of the factors. Factors are also easy to interpret since each factor correlates highly to one, two or three symptoms. Eight factors account for 61% of the variation of 19 original symptoms, a fact considered satisfactory.</p></sec><sec><title>Multiple Linear Regression results</title><p>Table <xref ref-type="table" rid="T4">4</xref> displays the results of the definitive multiple linear regression model which explains 37.4% of LOS variation in the sample. Residual analysis showed a Normal distribution and no need for quadratic terms. According to standardized regression coefficients (SRC) and statistical significance, the strongest influence on LOS came from six variables: age (SRC = 0.254), response to treatment (SRC = 0.246), the dummy for personality disorders (SRC = -0.236), ECT therapy (SRC = 0.215), factor 3, <italic>sexual and eating disorders </italic>(SRC = 0.203) and factor 1. <italic>psychotic symptoms </italic>(SRC = 0.174). The coefficients for the dummy variables standing for diagnoses indicate adjusted mean LOS for patients with a diagnosis of a personality disorder is the lowest and adjusted mean LOS for those included in the category of <italic>other disorders </italic>(reference category) is the highest. Thus, it is not possible to identify the specific diagnosis with the largest adjusted LOS in our series.</p></sec><sec><title>Validation</title><p>Table <xref ref-type="table" rid="T5">5</xref> shows means and standard deviations of the indicator (OLOS–PLOS) for both groups of medical attention according to efficiency. In the group with efficient care observed LOS is in average 4 days lower than predicted meanwhile in the group with efficiency problems observed LOS is in average 2 days higher than predicted. Figure <xref ref-type="fig" rid="F1">1</xref> provides a general view of the expected positive relation between observed and predicted LOS; however, it is apparent from the table and the scatter diagram that some patients have an observed LOS rather high or low according to their predicted LOS.</p><p>Figure <xref ref-type="fig" rid="F2">2</xref> shows the ROC Curve for detecting inefficient care with the indicator. The area under the curve is 0.695. (95% CI = 0.603 – 0.786).</p><p>Table <xref ref-type="table" rid="T6">6</xref> shows sensitivity and specificity, as well as predictive values for different cut-off points of the indicator. A reasonably high specificity (87%) will be obtained with a cut-off point of 5 days, and a high sensitivity with a cut-off point of -6 days (80%). There is no optimal cut-off point with high sensitivity and specificity. With a prevalence of 24% for records with inefficient care, positive predictive values are low but negative predictive values are very high for almost any point.</p></sec></sec><sec><title>Discussion</title><p>Our results focus on the plausibility of obtaining a function that fairly estimates the LOS a given patient, admitted in a Psychiatric Department for acute patients, should have had according to his or her characteristics.</p><p>Age and gender relationship to psychiatric LOS have been reported in several studies [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. Oiesvold et al [<xref ref-type="bibr" rid="B14">14</xref>] report longer LOS for patients in the older ages and for females in psychiatric patients in hospitals of Sweden and Finland. Huntley et al [<xref ref-type="bibr" rid="B12">12</xref>] classified age as one of the five variables significantly predicting LOS steadily over time. Barnow et al [<xref ref-type="bibr" rid="B9">9</xref>] found a correlation coefficient as high as 0.73 for describing the univariate association between age and LOS for depressed patients. Richter [<xref ref-type="bibr" rid="B21">21</xref>] found diagnosis and age were responsible of 10.5% of the LOS variations. Our results agree with these authors with regard to age but not to gender. Age is a natural determinant of LOS since it is closely related with all vital events; some authors (vg. Kiesler et al [<xref ref-type="bibr" rid="B7">7</xref>]), mix it up with another demographic variables while others like Tucker and Brems [<xref ref-type="bibr" rid="B8">8</xref>] just include it as a covariate.</p><p>Diagnosis is also a variable related to LOS in Psychiatric patients but how to include it with the aim of predicting LOS, is a challenge. Diagnosis Related Groups (DRG's) have been deeply explored [<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]; they have the advantage of being just a few for Psychiatry though some authors have alleged they are not relevant for predicting LOS [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. The Diagnostic and Statistical Manual of Mental Disorders (DSM) classifications (III and IV, lately) [<xref ref-type="bibr" rid="B8">8</xref>,<xref ref-type="bibr" rid="B13">13</xref>] or International Classification of Diseases (ICD 9<sup>th </sup>or 10<sup>th</sup>) [<xref ref-type="bibr" rid="B11">11</xref>] are also used in this context. Perhaps a broad classification system would achieve the best predictions but it would imply a huge number of patients for deriving the prediction function. We used an ad hoc classification based on DSM IV that yielded differences in LOS when analyzed univariately and when it was adjusted for other variables as well.</p><p>Most of the authors find an association between LOS and diagnosis [<xref ref-type="bibr" rid="B11">11</xref>-<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B26">26</xref>] but there are some discrepancies, most authors report psychoses as responsible for the highest LOS [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B16">16</xref>] but others find major depression [<xref ref-type="bibr" rid="B13">13</xref>] as more important predictor of LOS.</p><p>It is recognized that the patient's severity of illness influence LOS independently from diagnosis [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B27">27</xref>] but finding a valid, reliable and useful way to measure it, with the aim of adjusting quality indicators, has always been and continues to be a challenge [<xref ref-type="bibr" rid="B28">28</xref>]). Various scales for measuring severity of illness in Psychiatric inpatients have arisen in the last two decades. The Brief Psychiatric Rating Scale (BPRS) in its expanded version [<xref ref-type="bibr" rid="B29">29</xref>], the Psychiatric Severity of Illness Index (PSII) [<xref ref-type="bibr" rid="B27">27</xref>], the Computerized Psychiatric Severity Index (CPSI) [<xref ref-type="bibr" rid="B30">30</xref>] and the Health of the Nation Outcome Scale (HoNOS) [<xref ref-type="bibr" rid="B31">31</xref>,<xref ref-type="bibr" rid="B32">32</xref>] are probably the most mentioned ones. However, a low reliability is to expect during their use since appraisers must categorize symptoms in various levels of severity according to their opinion. Thus, implementation of any of these scales implies a period of special training and/or detailed instructions, a fact that prevents their use in daily practice. For instance, Durbin et al [<xref ref-type="bibr" rid="B33">33</xref>], in an attempt to introduce CPSI for predicting LOS from clinical records, had evaluators participate in a 3 day training program.</p><p>Our principal component solution is a real possibility since we collected information from all 19 symptoms described in the psychiatric record routinely used in our wards and afterward converted them in 8 factors. However, we still had to implement 3 categories for 8 of the symptoms (see Appendix), a feature that should be changed in the near future for the sake of gaining reliability.</p><p>We have not found any other study that employs PCA to reduce dimension with the aim of predicting LOS but the method has already been used to reduce symptoms' dimension in the field of Psychiatry and Psychology [<xref ref-type="bibr" rid="B34">34</xref>-<xref ref-type="bibr" rid="B36">36</xref>].</p><p>ECT was also an important LOS predictor in our series. It has been included by other authors in LOS prediction models [<xref ref-type="bibr" rid="B11">11</xref>] or mentioned as a cause for longer stays [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B37">37</xref>]. The use of ECT during hospitalization is a severity indicator, but is also a cause of complications.</p><p>Response to treatment turned out to be a variable with strong influence on LOS. Among reviewed literature only Draper and Luscombe [<xref ref-type="bibr" rid="B13">13</xref>] recognize the role of this issue in LOS prolongation. We understand it is a difficult aspect to assess and include in information systems unless the physician in charge of the patient completes the discharge form, a claim that should be evaluated in future research.</p><p>Other variables included in our study have been also explored by other authors. Marital status (or as <italic>living alone</italic>) has been acknowledged as an important LOS determinant in different studies [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B16">16</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B38">38</xref>]; physical co morbidities have also been analyzed [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B39">39</xref>] and found fairly relevant. Place of residence, as distance from home to hospital was also included in a Brazilian study [<xref ref-type="bibr" rid="B40">40</xref>]; medications were also examined by Parks [<xref ref-type="bibr" rid="B10">10</xref>] and found "polypharmacy" as a LOS predictor in geropsychiatric patients.</p><p>About the goodness of fit of the regression function, 0.37 is not an encouraging determination coefficient but the difficulty to find functions that explain more than 50% of LOS variation, a variable of complex nature, is also true. Among the psychiatric domain, Creed et al were able to explain up to 49% of LOS variation including demographic data, clinical features, social measures and behavioural issues [<xref ref-type="bibr" rid="B11">11</xref>]. Richter found an R2 of only 0.11 including in its function: age, diagnosis and other clinical and sociodemographic variables [<xref ref-type="bibr" rid="B21">21</xref>]. Stoskopf and Horn found coefficients in the range of 0.10–0.14 including only diagnosis and severity [<xref ref-type="bibr" rid="B30">30</xref>]. Huntley et al achieved to explain 17% of total variance in LOS including five variables in a stepwise regression analysis [<xref ref-type="bibr" rid="B12">12</xref>]. Regarding the fitted model, we chose the logarithmic transformation of LOS since its original distribution was right tailed. LOS distribution has been explored by various authors. Priest et al analyzed LOS distribution in an acute Psychiatric department in London; he found the exponential model yielded the best fit [<xref ref-type="bibr" rid="B41">41</xref>]. Stevens et al fit the exponential model and explore the influence of several factors by means of a Cox Regression model, an approach that would not allow LOS prediction [<xref ref-type="bibr" rid="B16">16</xref>]. However, several authors [<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B18">18</xref>] choose the logarithmic transformation for the search of predictors via a regression function, and perhaps most authors fit the regression model with the original LOS observations [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B23">23</xref>,<xref ref-type="bibr" rid="B33">33</xref>]. Marazzi in a huge European study found Lognormal, Weibull, or Gamma models were fine for describing the distribution of length of stay [<xref ref-type="bibr" rid="B42">42</xref>].</p><p>Regarding the proposed indicator (OLOS-PLOS), we confirmed the tendency of observed LOS to be higher than predicted LOS when there are inefficiency problems. However, we did not achieve a highly sensitive and specific cut-off point for detecting inefficiencies, a fact that emphasizes the necessity of refining the method with more variables and larger samples. A control method for efficient care similar to the proposed here is reported in some studies but not for psychiatric areas [<xref ref-type="bibr" rid="B43">43</xref>,<xref ref-type="bibr" rid="B44">44</xref>].</p><p>It is fair to recognize that the process of detecting inefficiencies in the records was somewhat arbitrary; first of all we almost identify inefficiencies as delays though it could be argued inefficient care can be provided without any delay. However, it would be out of the scope of the study to search for another kind of inefficiencies, as, for instance the ones arising from a wrong management of the patient. In second place, some cut offs for deeming a record "inefficient" are also arbitrarily chosen; we chose the time intervals considered normal for the hospital usual performance in all departments. Perhaps the main limitation of the present approach is that the prediction function must be estimated in the same setting where allegedly inefficiencies exist. This handicap is partly solved with the elimination of outliers during the function development. Finally, a practical limitation of our method ensues from the many variables that must be reported by assistant physicians or retrieved from the records. We believe this issue can be solved with the introduction of computers and friendly computer programs at the wards.</p></sec><sec><title>Conclusions</title><p>Our work supports the importance of a series of variables as LOS predictors in a Psychiatry department. The observed-predicted length of stay can be implemented as an indicator of inefficiencies provided the appropriate cut-off point is chosen. The approach showed its validity and is adaptable to other settings although there is an obvious need to continue the effort in the search of more explanatory functions.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Authors' contributions</title><p>REJ conceived the study and participated in its design and coordination. She personally wrote the manuscript. MM as the psychiatric specialist directed, organized and revised the data collected from the psychiatric records. RML directly reviewed the clinical records and obtained the information by means of adequate templates and under the supervision of MM. AD performed the statistical analysis. All authors read and approved the final manuscript</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-6963/4/4/prepub"/></p></sec>
Following in the footsteps of smallpox: can we achieve the global eradication of measles?	<sec><title>Background</title><p>Although an effective measles vaccine has been available for almost 40 years, in 2000 there were about 30 million measles infections worldwide and 777,000 measles-related deaths. The history of smallpox suggests that achieving measles eradication depends on several factors; the biological characteristics of the organism; vaccine technology; surveillance and laboratory identification; effective delivery of vaccination programmes and international commitment to eradication.</p></sec><sec><title>Discussion</title><p>Like smallpox, measles virus has several biological characteristics that favour eradication. Humans are the only reservoir for the virus, which causes a visible illness and infection leading to life-long immunity. As the measles virus has only one genetic serotype which is relatively stable over time, the same basic vaccine can be used world-wide. Vaccination provides protection against measles infection for at least 15 years, although efficacy may be reduced due to host factors such as nutritional status. Measles vaccination may also confer other non-specific health benefits leading to reduced mortality. Accurate laboratory identification of measles cases enables enhanced surveillance to support elimination programmes. The "catch-up, keep-up, follow-up" vaccination programme implemented in the Americas has shown that measles elimination is possible using existing technologies. On 17<sup>th </sup>October 2003 the "Cape Town Measles Declaration" by the World Health Organisation and the United Nations Childrens Fund called on governments to intensify efforts to reduce measles mortality by supporting universal vaccination coverage and the development of more effective vaccination.</p></sec><sec><title>Summary</title><p>Although more difficult than for smallpox, recent experience in the Americas suggests that measles eradication is technically feasible. Growing international support to deliver these programmes means that measles, like smallpox, may very well become a curiosity of history.</p></sec>	<contrib id="A1" corresp="yes" contrib-type="author"><name><surname>Morgan</surname><given-names>Oliver WC</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC International Health and Human Rights	<sec><title>Background</title><p>On the 26th October 1977, the last known case of naturally acquired smallpox occurred in Somalia [<xref ref-type="bibr" rid="B1">1</xref>]. This promised a new era in which many infectious diseases would be eradicated. However, nearly a quarter of a century later, 1.7 million children die each year from vaccine preventable diseases, nearly half due to measles (Figure. <xref ref-type="fig" rid="F1">1</xref>). Although an effective measles vaccine has been available for almost 40 years [<xref ref-type="bibr" rid="B2">2</xref>], in 2000 there were about 30 million measles infectious worldwide and 777,000 measles-related deaths [<xref ref-type="bibr" rid="B3">3</xref>]; measles is the 5<sup>th </sup>leading cause of death in children under the age of five [<xref ref-type="bibr" rid="B4">4</xref>]. One might therefore wonder why have we not eradicated measles?</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Proportional mortality (1.7 million worldwide) due to vaccine preventable diseases among children (2000) </bold>Source: WHO (2002) [<xref ref-type="bibr" rid="B31">31</xref>]</p></caption><graphic xlink:href="1472-698X-4-1-1"/></fig><p>The history of smallpox has shown us that several factors must be considered in order to eradicate an infectious disease [<xref ref-type="bibr" rid="B5">5</xref>-<xref ref-type="bibr" rid="B7">7</xref>]; the biological characteristics of the organism; vaccine technology; surveillance and laboratory identification; effective delivery of vaccination programmes and international commitment to eradication. I will consider each of these elements, their implications for measles eradication and draw on lessons from smallpox eradication.</p></sec><sec><title>Discussion</title><sec><title>Biological characteristics of the measles virus</title><p>Measles virus is a spherical single-stranded RNA virus belonging to the paramyxoviridae family [<xref ref-type="bibr" rid="B2">2</xref>]. It is spread by airborne droplets causing rash, cough and fever which lasts for several days [<xref ref-type="bibr" rid="B8">8</xref>]. Although there is no treatment, most infected individuals recover by themselves. However, complications due to pneumonia occur in 2–27% of cases causing 56–86% of all measles deaths [<xref ref-type="bibr" rid="B2">2</xref>]. Less commonly, measles infection can cause serious neurological complications [<xref ref-type="bibr" rid="B2">2</xref>]. Compared to smallpox, the measles virus is considerably more contagious, capable of causing large outbreaks even in populations with high vaccine coverage [<xref ref-type="bibr" rid="B9">9</xref>]. Nevertheless, measles shares several biological characteristics with smallpox which favour eradication [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B7">7</xref>,<xref ref-type="bibr" rid="B10">10</xref>]: humans are the only reservoir for the virus (i.e. animals are not infected); measles causes a visible illness; infection leads to life-long immunity; cases often occur at regular intervals enabling the targeting of interventions; measles virus has only one genetic serotype which is relatively stable over time [<xref ref-type="bibr" rid="B2">2</xref>]; an effective vaccine is available and accurate laboratory identification is possible.</p></sec><sec><title>Vaccine technology</title><p>The limited variability of the measles genome means that the same basic vaccine can be used worldwide [<xref ref-type="bibr" rid="B2">2</xref>]. A single dose of the vaccine provides about 90% protection, which is increased to about 99% with a second dose [<xref ref-type="bibr" rid="B11">11</xref>]. However vaccination programmes in developing countries may not achieve such high levels of protection, either because of cold chain failures [<xref ref-type="bibr" rid="B12">12</xref>] or host factors such as poor nutritional status [<xref ref-type="bibr" rid="B13">13</xref>,<xref ref-type="bibr" rid="B14">14</xref>]. Antibody titres following vaccination are lower than following infection with wild measles virus [<xref ref-type="bibr" rid="B2">2</xref>] and decrease over time. Nevertheless, vaccination can provide immunological memory against wild measles virus for at least 15 years [<xref ref-type="bibr" rid="B15">15</xref>]. However, in contrast to immunity following infection by wild measles virus, protection may not be life-long [<xref ref-type="bibr" rid="B15">15</xref>]. Infection due to wild virus amongst individuals with reduced vaccine-induced protection may be subclinical or cause milder illness than for individuals who are unvaccinated or did not seroconvert following vaccination [<xref ref-type="bibr" rid="B2">2</xref>].</p><p>Because of the highly infectious nature of the virus, between 90–95% vaccination coverage is needed to halt measles transmission [<xref ref-type="bibr" rid="B10">10</xref>]. An additional difficulty is that maternal antibodies interfere with the vaccine, reducing seroconversion in infants between 6–9 months old [<xref ref-type="bibr" rid="B16">16</xref>]. Despite concerns in some industrialised countries about the safety of the measles vaccine, adverse vaccine-associated events are rare [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B17">17</xref>]. A study of 20 million children and adolescents receiving measles vaccination in the UK, Canada, Australia, South Korea, Costa Rica, Romania and New Zealand identified no vaccine-related deaths [<xref ref-type="bibr" rid="B9">9</xref>]. The risk of encephalitis was estimated to be 1 per 1 million doses and acute anaphylaxis less than 1 per 1 million doses [<xref ref-type="bibr" rid="B9">9</xref>].</p><p>Epidemiological research further suggests that measles vaccination (and mild measles illness) may actually confer non-specific health benefits leading to reduced childhood mortality rates [<xref ref-type="bibr" rid="B18">18</xref>,<xref ref-type="bibr" rid="B19">19</xref>]. The reduction in non-measles mortality following vaccination is greater in girls than boys [<xref ref-type="bibr" rid="B20">20</xref>]. Although there is currently no biological explanation for this, greater protection against other infections may be caused by immunological stimulation due to the measles virus or vaccine [<xref ref-type="bibr" rid="B18">18</xref>]. If this is indeed the case, the implications would be that continued measles vaccination may be beneficial even after wild measles virus is eradicated [<xref ref-type="bibr" rid="B18">18</xref>].</p><p>Measles vaccination is currently given by injection, which requires skilled personnel to administer, thereby making widespread coverage more difficult. Injections are also associated with an increased risk of blood borne disease (e.g. HIV, hepatitis), risk of infection at the injection site and they lead to large amounts of medical waste [<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. As efforts to eradicate smallpox were transformed by the development of better vaccine delivery techniques [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>], a new aerosolised measles vaccine promises to transform the battle against the measles vaccine [<xref ref-type="bibr" rid="B21">21</xref>]. The new approach, which is non-invasive and requires only minimal training to administer, has been successfully trailed in Mexico and is hoped to be in widespread use by 2009 [<xref ref-type="bibr" rid="B21">21</xref>]. Further research using a wide range of vaccine technologies such as DNA vaccines, bacterial vectors and viral vectors is also being undertaken to develop effective vaccines for infants younger than 6 months old [<xref ref-type="bibr" rid="B21">21</xref>].</p></sec><sec><title>Surveillance and laboratory identification</title><p>A strong measles surveillance network is vital to ensure that vaccine efficacy and coverage are maintained and that outbreaks or reservoirs of disease are identified [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. Identification of measles infection can be made on clinical grounds by experienced clinicians. However, experience with smallpox has shown that as the incidence of disease decreases it becomes increasingly important to investigate every suspected case (case-based surveillance) and provide laboratory confirmation. In order to achieve this, the World Health Organisation is establishing a Global Measles Laboratory Network, which will develop local reference laboratories and train staff [<xref ref-type="bibr" rid="B24">24</xref>]. Laboratory identification of measles can be performed using relatively simple methods to test for the presence of measles-specific antibodies in the blood [<xref ref-type="bibr" rid="B25">25</xref>]. Where more complex genetic techniques are available, identification of different virus strains can show whether infections are due to local transmission or importation [<xref ref-type="bibr" rid="B24">24</xref>,<xref ref-type="bibr" rid="B25">25</xref>]. However, many poorer countries currently still lack sufficient resources to develop enhanced measles surveillance and laboratory diagnosis.</p></sec><sec><title>Effective delivery of vaccination programmes</title><p>In 2000, vaccination programmes worldwide achieved about 80% coverage with one dose of measles vaccine (Table <xref ref-type="table" rid="T1">1</xref>). This was lowest in the African region (65%) and highest in the American (91%) and European (92%) regions. Only 74 countries (35%) achieved greater than 90% coverage while sixteen countries achieved coverage below 50% [<xref ref-type="bibr" rid="B4">4</xref>]. Since 2001 81% of countries have been offering a second opportunity for measles vaccination [<xref ref-type="bibr" rid="B4">4</xref>]. These efforts have had a significant impact on measles morbidity, preventing an estimated 1 million measles-related deaths per year compared to pre-vaccination levels (Table <xref ref-type="table" rid="T1">1</xref>) [<xref ref-type="bibr" rid="B4">4</xref>].</p><table-wrap position="float" id="T1"><label>Table 1</label><caption><p>Number of deaths, vaccination coverage and deaths prevented by WHO region (2000)</p></caption><table frame="hsides" rules="groups"><thead><tr><td align="left"><bold>Region</bold></td><td align="center"><bold>No of deaths</bold></td><td align="center"><bold>Vaccination coverage (%)</bold></td><td align="center"><bold>Deaths prevented</bold></td></tr></thead><tbody><tr><td align="left">African</td><td align="center">453,000</td><td align="center">65</td><td align="center">384,960</td></tr><tr><td align="left">American</td><td align="center">0</td><td align="center">91</td><td align="center">148,852</td></tr><tr><td align="left">Eastern Mediterranean</td><td align="center">81,000</td><td align="center">79</td><td align="center">129,769</td></tr><tr><td align="left">European</td><td align="center">7,000</td><td align="center">92</td><td align="center">28,692</td></tr><tr><td align="left">Southeast Asia</td><td align="center">202,000</td><td align="center">83</td><td align="center">256,633</td></tr><tr><td align="left">Western Pacific</td><td align="center">34,000</td><td align="center">86</td><td align="center">126,132</td></tr><tr><td colspan="4"><hr></hr></td></tr><tr><td align="left"><bold>Total</bold></td><td align="center"><bold>777,000</bold></td><td align="center"><bold>80</bold></td><td align="center"><bold>1,071,938</bold></td></tr></tbody></table><table-wrap-foot><p>Source: Adapted from Henao-Restrepo (2003) [<xref ref-type="bibr" rid="B4">4</xref>]</p></table-wrap-foot></table-wrap><p>However, the Americas is the only region that has made significant progress towards measles eradication; no known indigenous measles transmission has occurred since November 2002 [<xref ref-type="bibr" rid="B22">22</xref>]. This has been achieved by implementing a "catch-up, keep-up, follow-up" programme [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. The programme begins with a one-time "catch-up" mass-vaccination campaign aimed at all children aged 1–14 years old (individuals over 15 years are considered to have acquired natural immunity and infants are not included). Routine vaccination is given to all individuals at 12 months of age to "keep-up" coverage and periodic mass-vaccination campaigns provide a second "follow-up" opportunity. The universal second opportunity is important because children who miss their first vaccination and those individuals not protected by the first vaccination gradually build-up a population of susceptible individuals able to sustain transmission if the virus is re-introduced. Epidemiological studies suggest that these 'cycles of abundance' occur about every four years [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B22">22</xref>]. However, the need to achieve almost universal coverage of large geographical areas to halt measles transmission has meant that follow-up vaccinations programmes targeted at urban centres or limited geographical areas have sometimes been unsuccessful [<xref ref-type="bibr" rid="B26">26</xref>].</p><p>Implementing "catch-up, keep-up, follow-up" programmes present several challenges; the high cost can not be met by many countries while donor dependency may create unsustainable programmes in the longer term; poor healthcare infrastructure, limited access to rural areas and vaccine management are likely to reduce vaccination coverage; competing national and international interests may divert necessary financial and technical resources; campaigns may interrupt or divert resources away from routine vaccination programmes thereby reducing their efficiency; adopting a second opportunity for measles vaccination for all children will almost double the demand for measles vaccine and without adequate demand forecasting, sufficient quantities of vaccine may not be available; political and social upheaval such as armed conflict or natural disasters may interrupt on-going programmes [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B12">12</xref>,<xref ref-type="bibr" rid="B27">27</xref>].</p></sec><sec><title>International commitment to eradication</title><p>In 2001 the World Health Organisation (WHO) and United Nations Children's Fund (UNICEF) published a joint strategic plan for 2001–2005 [<xref ref-type="bibr" rid="B23">23</xref>]. The strategy aims to reduce measles mortality in 2005 by 50% compared to 1999 levels and to maintain interruption of indigenous measles transmission in large geographic areas. The strategy was endorsed in 2003 by the World Health Assembly, which includes 192 member states. On 17<sup>th </sup>October 2003, a meeting of WHO and UNICEF produced the "Cape Town Measles Declaration" calling on governments to intensify efforts to meet the Strategy's goal of mortality reduction [<xref ref-type="bibr" rid="B28">28</xref>]. Other initiatives, such as the Global Alliance for Vaccines and Immunisations (GAVI) and the Measles Initiative, are helping to ensure adequate vaccine provision [<xref ref-type="bibr" rid="B29">29</xref>,<xref ref-type="bibr" rid="B30">30</xref>]. At the end of the strategic plan in 2005, WHO and UNICEF will review progress and assess the feasibility of global measles eradication.</p></sec></sec><sec><title>Conclusions</title><p>Global eradication of measles is more difficult than for smallpox, mostly due to the greater virulence of the virus, needing almost universal vaccine coverage. However, success in the Americas has shown that measles eradication is technically feasible using existing vaccines and vaccination programmes. Growing international support, to deliver these programmes means that measles, like smallpox, can very well become a curiosity of history.</p></sec><sec><title>Summary</title><p>▪ Although an effective vaccine has been available for almost 40 years, in 2000 there were about 30 million measles infections and 777,000 measles-related deaths worldwide.</p><p>▪ The history of smallpox highlights several important factors for disease eradication; the biological characteristics of the organism; vaccine technology; surveillance and laboratory identification; effective delivery of vaccination programmes and international commitment to eradication.</p><p>▪ Although more difficult than for smallpox, consideration of these factors and recent experience in the Americas suggests that measles eradication is technically feasible.</p><p>▪ There is growing international support to deliver effective measles vaccination programmes leading to the eradication of measles in our lifetime.</p></sec><sec><title>Competing interests</title><p>None declared.</p></sec><sec><title>Pre-publication history</title><p>The pre-publication history for this paper can be accessed here:</p><p><ext-link ext-link-type="uri" xlink:href="http://www.biomedcentral.com/1472-698X/4/1/prepub"/></p></sec>
Basal jawed vertebrate phylogeny inferred from multiple nuclear DNA-coded genes	<sec><title>Background</title><p>Phylogenetic analyses of jawed vertebrates based on mitochondrial sequences often result in confusing inferences which are obviously inconsistent with generally accepted trees. In particular, in a hypothesis by Rasmussen and Arnason based on mitochondrial trees, cartilaginous fishes have a terminal position in a paraphyletic cluster of bony fishes. No previous analysis based on nuclear DNA-coded genes could significantly reject the mitochondrial trees of jawed vertebrates.</p></sec><sec><title>Results</title><p>We have cloned and sequenced seven nuclear DNA-coded genes from 13 vertebrate species. These sequences, together with sequences available from databases including 13 jawed vertebrates from eight major groups (cartilaginous fishes, bichir, chondrosteans, gar, bowfin, teleost fishes, lungfishes and tetrapods) and an outgroup (a cyclostome and a lancelet), have been subjected to phylogenetic analyses based on the maximum likelihood method.</p></sec><sec><title>Conclusion</title><p>Cartilaginous fishes have been inferred to be basal to other jawed vertebrates, which is consistent with the generally accepted view. The minimum log-likelihood difference between the maximum likelihood tree and trees not supporting the basal position of cartilaginous fishes is 18.3 ± 13.1. The hypothesis by Rasmussen and Arnason has been significantly rejected with the minimum log-likelihood difference of 123 ± 23.3. Our tree has also shown that living holosteans, comprising bowfin and gar, form a monophyletic group which is the sister group to teleost fishes. This is consistent with a formerly prevalent view of vertebrate classification, although inconsistent with both of the current morphology-based and mitochondrial sequence-based trees. Furthermore, the bichir has been shown to be the basal ray-finned fish. Tetrapods and lungfish have formed a monophyletic cluster in the tree inferred from the concatenated alignment, being consistent with the currently prevalent view. It also remains possible that tetrapods are more closely related to ray-finned fishes than to lungfishes.</p></sec>	<contrib id="A1" equal-contrib="yes" contrib-type="author"><name><surname>Kikugawa</surname><given-names>Kanae</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" equal-contrib="yes" contrib-type="author"><name><surname>Katoh</surname><given-names>Kazutaka</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Kuraku</surname><given-names>Shigehiro</given-names></name><xref ref-type="aff" rid="I1">1</xref><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Sakurai</surname><given-names>Hiroshi</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Ishida</surname><given-names>Osamu</given-names></name><xref ref-type="aff" rid="I3">3</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Iwabe</surname><given-names>Naoyuki</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A7" corresp="yes" contrib-type="author"><name><surname>Miyata</surname><given-names>Takashi</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib>	BMC Biology	<sec><title>Background</title><p>The evolutionary relationship among jawed vertebrates is currently a controversial issue. Cartilaginous fishes are traditionally considered to be ancestral to other jawed vertebrates (Figure <xref ref-type="fig" rid="F1">1A</xref>). Arnason and colleagues challenged the traditional view, based on phylogenetic analyses of complete mitochondrial sequences from several vertebrates [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. According to their mitochondrial tree (Figure <xref ref-type="fig" rid="F1">1B</xref>), cartilaginous fishes have a terminal position in the phylogeny of bony fishes (coelacanth, lungfishes, bichirs, teleost fishes and other ray-finned fishes), implying that bony fishes are ancestral to cartilaginous fishes. Furthermore, the mitochondrial tree shows a basal split between tetrapods and other jawed vertebrates.</p><p>Phylogenetic analyses based on mitochondrial sequences, however, often result in misleading trees when distantly related vertebrates are compared [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B7">7</xref>]. Some efforts have been made by several groups to obtain the robust phylogenetic trees of jawed vertebrates based on nuclear DNA-coded genes. In the LSU rRNA tree by Zardoya and Meyer [<xref ref-type="bibr" rid="B6">6</xref>], the basal position of cartilaginous fishes is not significantly supported; the bootstrap probabilities are 72%, 68% and 74%, for the maximum parsimony (MP) method, the neighbor joining (NJ) method and the maximum likelihood (ML) method, respectively. On the basis of presence or absence of insertions or deletions within conserved sequences [<xref ref-type="bibr" rid="B8">8</xref>], Venkatesh <italic>et al.</italic>[<xref ref-type="bibr" rid="B9">9</xref>] claimed to have found robust molecular evidence (molecular synapomorphy) against the mitochondrial tree [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. However, their tree is basically an unrooted tree of major groups of jawed vertebrates as pointed out by Dimmick [<xref ref-type="bibr" rid="B10">10</xref>], because none of the molecular synapomorphies they found included an outgroup (cyclostomes or lancelets). Apart from the position of bichir, the tree by Venkatesh <italic>et al. </italic>is equivalent to that by Rasmussen <italic>et al.</italic>, when compared as unrooted trees.</p><p>Martin [<xref ref-type="bibr" rid="B11">11</xref>] analyzed multiple nuclear DNA-coded genes and the hypothesis by Rasmussen <italic>et al. </italic>[<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>] could not be refuted. Hedges[<xref ref-type="bibr" rid="B12">12</xref>] analyzed 10 nuclear DNA-coded genes from two cyclostomes and three jawed vertebrates, and concluded the basal position of cartilaginous fishes in the jawed vertebrate tree. Takezaki <italic>et al. </italic>[<xref ref-type="bibr" rid="B13">13</xref>] confirmed this finding based on a comparison of 31 nuclear DNA-coded genes. Because only a single bony fish lineage represented by teleost fishes is included in these analyses, it remains possible that other bony fishes (lungfishes or bichir) are more deeply branching than cartilaginous fishes are. If it is the case, one cannot refute the hypothesis by Rasmussen and Arnason [<xref ref-type="bibr" rid="B3">3</xref>] that bony fishes are ancestral to cartilaginous fishes. The phylogenetic position of bichir is particularly important; bichir is often inferred to be the outgroup to all other jawed vertebrates in mitochondrial trees, when amphibian data is included in comparison (data not shown).</p><p>The phylogenetic relationship amongst teleost fishes and two holosteans is also controversial. Living holosteans, comprising bowfin and gar, are possible sister groups of teleost fishes [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B15">15</xref>]. All of three possible topologies (Figure <xref ref-type="fig" rid="F2">2A,2B,2C</xref>) were proposed by morphologists to date (see references cited in [<xref ref-type="bibr" rid="B15">15</xref>] and [<xref ref-type="bibr" rid="B16">16</xref>]). Partial mitochondrial and LSU rRNA data, on the other hand, do not support any of these morphology-based trees at a statistically significant level [<xref ref-type="bibr" rid="B17">17</xref>,<xref ref-type="bibr" rid="B18">18</xref>]. Venkatesh <italic>et al. </italic>[<xref ref-type="bibr" rid="B9">9</xref>] noted the possibility of an alternative tree (Figure <xref ref-type="fig" rid="F2">2D</xref>) based on a molecular synapomorphy. Inoue <italic>et al. </italic>[<xref ref-type="bibr" rid="B16">16</xref>] recently reported that this tree was supported by complete mitochondrial sequences. Mitochondrial sequences, however, may not be suitable for inferring phylogenetic relationships among such distantly related groups [<xref ref-type="bibr" rid="B6">6</xref>,<xref ref-type="bibr" rid="B19">19</xref>].</p><p>To test the mitochondrial trees at a statistically significant level, it is therefore necessary to perform phylogenetic analyses based on nuclear DNA-coded genes. There is, however, a possible error from paralogous comparisons when a nuclear DNA-coded gene tree is used for inferring the phylogenetic relationship of organisms. To avoid this, we selected basically single copy genes, such as enzymes in glycolysis, which are evolving at roughly constant rates over a wide range of animal taxa [<xref ref-type="bibr" rid="B20">20</xref>,<xref ref-type="bibr" rid="B21">21</xref>]. Since their evolutionary rates are generally low, no single gene amongst them has detailed phylogenetic information. Thus a large amount of sequence is needed for a statistically solid inference.</p><p>We have cloned and sequenced seven nuclear DNA-coded genes comprising ~3,000 amino acid residues in total, from eleven jawed vertebrate species, two cyclostomes (lamprey and/or hagfish) and a lancelet. These amino acid sequences, together with those available from the DDBJ/EMBL/GenBank databases, were subjected to phylogenetic analyses and statistical tests based on the ML method. We report here that the nuclear DNA-coded gene tree differs sharply from the mitochondrial tree on the two phylogenetic problems of jawed vertebrates; our tree supports the deepest position of cartilaginous fishes in jawed-vertebrate phylogeny, and the monophyly of holosteans.</p></sec><sec><title>Results and discussion</title><sec><title>Phylogenetic tree inference</title><p>Teleost fishes have two TPI genes (TPI-A and TPI-B) [<xref ref-type="bibr" rid="B22">22</xref>]; <italic>A. baerii </italic>has two ALDa genes (AB111402 and AB111403); mammals have two PGK genes (M11968 and X05246 for human, M15668 and M17299 for mouse); and the mouse has two G6PD genes (Z84471 and AF326207). Each of these four gene pairs is shown to have multiplied within the respective taxonomic group by preliminary phylogenetic analyses. To avoid 'long branch attraction' (LBA) artifact [<xref ref-type="bibr" rid="B23">23</xref>], the slowly evolving counterpart for each of these gene pairs was selected for phylogenetic inference: <italic>O. latipes </italic>ortholog of TPI-B, AB111402 for <italic>A. baerii </italic>ALDa, M11968 for human PGK, M15668 for mouse PGK and AF326207 for mouse G6PD. Cyclostomes have muscle and non-muscle types of aldolase (ALD) genes [<xref ref-type="bibr" rid="B24">24</xref>]. Although the relationship between these two cyclostome ALD genes and three ALD genes (a, b and c) from jawed vertebrates is not clearly resolved, each of the jawed-vertebrate ALD genes was inferred to be orthologous [<xref ref-type="bibr" rid="B25">25</xref>]. The muscle-type ALD gene of hagfish, the non-muscle-type ALD gene of hagfish and the non-muscle-type ALD gene of lamprey were used as outgroups for ALDa, ALDb and ALDc of jawed vertebrates, respectively.</p><p>For each of the seven proteins, the amino acid sequences from 15 vertebrate species listed in Materials and methods have been aligned, and phylogenetic tree analyses have been carried out for regions comprising 317 amino acid residues (aa) in ALDa, 316aa in ALDb, 317aa in ALDc, 463aa in G6PD, 940aa in GAG, 383aa in PGK and 206aa in TPI, for each of which unambiguous alignment is possible. The total data set of 2,942aa was subjected to phylogenetic analyses based on the <sc>GAMT</sc> program [<xref ref-type="bibr" rid="B26">26</xref>] as described in materials and methods.</p><p>We have selected the candidate topologies – a set of topologies with log-likelihood values close to that of the ML tree – from seven protein data sets as described in Materials and Methods. The numbers of candidate topologies selected are 379 from the ALDa data set, 91 from the ALDb data set, 1,121 from the ALDc data set, 665 from the G6PD data set, 11 from the GAG data set, 652 from the PGK data set and 2,860 from the TPI data set, and 103 from the concatenated alignment. Excluding identical topologies, a total of 5,801 topologies were subjected to further analyses as the candidate topologies. For each candidate topology, the likelihood value of totalml and that of concatenated alignment were computed.</p><p>Figure <xref ref-type="fig" rid="F3">3</xref> shows the ML tree inferred from the concatenated alignment. This tree strongly supports the basal position of cartilaginous fishes and the monophyly of holosteans, although individual trees inferred from each of seven proteins did not give statistically significant results, probably because of limited phylogenetic information held in a single gene (data not shown). Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref> show the ML topology and some topologies with large likelihood values inferred from concatenated alignment analysis and totalml analysis, respectively. Each table includes only topologies with <italic>P</italic>-values larger than 5% calculated by the Kishino-Hasegawa (KH) test. Note that the ML tree in totalml analysis (topology <italic>b </italic>in Table <xref ref-type="table" rid="T2">2</xref>) differs from that in concatenated alignment analysis (Figure <xref ref-type="fig" rid="F3">3</xref> and topology <italic>a </italic>in Table <xref ref-type="table" rid="T1">1</xref>).</p></sec><sec><title>Statistical tests</title><p>Topologies <italic>a </italic>and <italic>b </italic>in Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref> differ considerably from other topologies in their bootstrap probabilities and <italic>P</italic>-values. These two topologies have approximately equal likelihood values in each of the totalml and concatenated alignment analyses, although the ML tree in concatenated alignment analysis is the second best tree in totalml analysis, and <italic>vice versa</italic>.</p><p>In addition to the bootstrap probability and the KH test, a test based on Bayesian posterior probability (BPP) has been carried out. The resulting BPP values are self-contradictory; topology <italic>a</italic>, which was the best topology in concatenated alignment analysis (Table <xref ref-type="table" rid="T1">1</xref>), is significantly rejected in totalml analysis (Table <xref ref-type="table" rid="T2">2</xref>; the BPP value was 0.005). Thus the BPP test might be too liberal, as already pointed out [<xref ref-type="bibr" rid="B27">27</xref>,<xref ref-type="bibr" rid="B28">28</xref>]. The approximately unbiased (AU) test has also been carried out for reference.</p><p>Focusing on some phylogenetic problems, the support values for each competing hypothesis were computed based on the intact bootstrap probability (BP) analysis (see Materials and Methods), the <sc>TREE-PUZZLE</sc> (TP) program [<xref ref-type="bibr" rid="B29">29</xref>] and the <sc>MRBAYES</sc>[<xref ref-type="bibr" rid="B30">30</xref>] program (Table <xref ref-type="table" rid="T3">3</xref>). In addition, the RELL BP value, the BPP value and the <italic>P</italic>-values by the KH test and the AU test, which are based on concatenated alignment analysis described above, are also shown. The intact BP value is largely accordant to the RELL BP value, whereas low support values are observed in the TP method. This may be an artifact derived from the limited topology searches in the TP method, because the same result as shown in Table <xref ref-type="table" rid="T1">1</xref> was obtained, when the candidate topologies described above were subjected to the <sc>TREE-PUZZLE</sc> program with the 'user defined trees' option.</p></sec><sec><title>Cartilaginous fishes have a basal position among jawed vertebrates</title><p>Cartilaginous fishes are thought to be ancestral to other jawed vertebrates in the traditional view (Figure <xref ref-type="fig" rid="F1">1A</xref>). In contrast, Rasmussen and Arnason [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>] and Arnason <italic>et al. </italic>[<xref ref-type="bibr" rid="B3">3</xref>] pointed out another possibility that bony fishes are ancestral to cartilaginous fishes (Figure <xref ref-type="fig" rid="F1">1B</xref>). The present results strongly support the traditional view as shown in Figure <xref ref-type="fig" rid="F3">3</xref> and Table <xref ref-type="table" rid="T3">3</xref>. The bootstrap probabilities of the topologies having a basal position of cartilaginous fishes totaled 88.2% and 87.8% in concatenated alignment analysis and totalml analysis, respectively. The minimum log-likelihood difference between the ML tree and trees not supporting a basal split between cartilaginous fishes and remaining jawed vertebrates was 18.3 ± 13.1 (<italic>P</italic>-value = 0.09) and 15.3 ± 12.7 (<italic>P</italic>-value = 0.12), in concatenated alignment analysis and totalml analysis, respectively. The minimum log-likelihood difference between the ML tree and that supporting the bony fish origin of cartilaginous fishes was 123 ± 23.3 (<italic>P</italic>-value < 0.01) and 137 ± 29.6 (<italic>P</italic>-value < 0.01) in concatenated alignment analysis and totalml analysis, respectively, providing strong evidence against the hypothesis by Arnason's group [<xref ref-type="bibr" rid="B1">1</xref>-<xref ref-type="bibr" rid="B3">3</xref>]. When the lancelet (a distant outgroup) sequences are excluded from the analysis, the minimum log-likelihood difference between the ML tree and trees that support their hypothesis was 122 ± 25.9, still being statistically significant.</p><p>According to the phylogenetic analysis based on mitochondrial sequences, however, all topologies consistent with the present analysis are significantly rejected (<italic>P</italic>-value < 0.01). This controversial result may be due to the incompleteness of phylogenetic information retained in the mitochondrial sequences; the amino acid composition of mitochondrial DNA-coded proteins is highly biased to hydrophobic residues and thus multiple and reverse substitutions may occurs very frequently [<xref ref-type="bibr" rid="B4">4</xref>]. In addition, the evolutionary rates of mitochondrial sequences often differ greatly for different lineages; the mitochondrial sequences of most tetrapods evolve more rapidly than those of fishes [<xref ref-type="bibr" rid="B31">31</xref>,<xref ref-type="bibr" rid="B32">32</xref>]. These evolutionary features characteristic of mitochondrial sequences might result in the LBA artifact [<xref ref-type="bibr" rid="B23">23</xref>].</p></sec><sec><title>Did tetrapods originate from lobe-finned fishes?</title><p>Several molecular phylogenetic analyses were carried out to clarify the phylogenetic relationship among tetrapods, coelacanth and lungfishes, using ray-finned fishes [<xref ref-type="bibr" rid="B33">33</xref>-<xref ref-type="bibr" rid="B36">36</xref>] and/or cartilaginous fishes [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B37">37</xref>] as an outgroup. The validity of these two rootings needs to be confirmed with molecular evidence [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B5">5</xref>]. Although no coelacanth sequence is included, the present analysis provides a confirmation for the cartilaginous fish rooting. Ray-finned fishes, however, cannot be used as an outgroup, because it remains possible that tetrapods are more closely related to ray-finned fishes than to lobe-finned fishes (topology <italic>b </italic>in Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref>).</p><p>Since bichir, the basal ray-finned fishes (see below), have a pair of lungs and fleshy pectoral fins [<xref ref-type="bibr" rid="B38">38</xref>], the common ancestor of bony fishes are likely to be somewhat like lobe-finned fishes. Thus it remains possible that tetrapods originated from such ancestral ray-finned fishes or from the common ancestor of ray-finned fishes and lobe-finned fishes. Recently reported fossil records suggest that the divergence of lungfish and tetrapods occurred at least as early as 417-412 Mya [<xref ref-type="bibr" rid="B39">39</xref>,<xref ref-type="bibr" rid="B40">40</xref>]. According to Kumar and Hedges [<xref ref-type="bibr" rid="B41">41</xref>] and Hedges [<xref ref-type="bibr" rid="B12">12</xref>], the divergence of ray-finned fishes and lobe-finned fishes was estimated to have occured at 450-400 Mya, which is simultaneous with or immediately before the divergence of lungfish and tetrapods. This is consistent with the present study suggesting the almost simultaneous divergence of tetrapods, lungfishes and ray-finned fishes.</p></sec><sec><title>Living holosteans form a natural group</title><p>The phylogenetic relationship among teleost fishes and holosteans comprising bowfin and gar is controversial [<xref ref-type="bibr" rid="B15">15</xref>]. Four different tree topologies (Figure <xref ref-type="fig" rid="F2">2A,2B,2C,2D</xref>) have been proposed to date from morphological and molecular data. According to a formerly accepted view, living ray-finned fishes are divided into three major groups (Figure <xref ref-type="fig" rid="F2">2A</xref>): Chondrostei (chondrosteans including sturgeons and paddlefishes), 'Holostei' (holosteans comprising bowfin and gar), and Teleostei (teleost fishes consisting all other living ray-finned fishes). 'Holostei' is, however, a term that has fallen into disuse in formal classifications. Instead, in the currently accepted view, holosteans are considered to be paraphyletic; bowfin is thought to be more closely related to teleost fishes than gar is [<xref ref-type="bibr" rid="B14">14</xref>,<xref ref-type="bibr" rid="B38">38</xref>], as shown in Figure <xref ref-type="fig" rid="F2">2B</xref>, and therefore ray-finned fishes are classified into two monophyletic groups: Chondrostei and Neopterygii (holosteans and teleost fishes). Another possibility that gar is closely related to teleost fishes (Figure <xref ref-type="fig" rid="F2">2C</xref>) was also proposed [<xref ref-type="bibr" rid="B42">42</xref>]. Furthermore, mitochondrial sequences suggest a distinct tree topology (Figure <xref ref-type="fig" rid="F2">2D</xref>), in which holosteans and chondrosteans form a monophyletic group [<xref ref-type="bibr" rid="B16">16</xref>].</p><p>In the present analysis, holosteans are inferred to form a monophyletic group that is the sister group to teleost fishes, as shown in Figure <xref ref-type="fig" rid="F3">3</xref> and Table <xref ref-type="table" rid="T3">3</xref>. The bootstrap probabilities for the holostean clade are 92.2% and 83.8% in concatenated alignment analysis and totalml analysis, respectively. The topologies not supporting the holostean clade are relatively small in <italic>P</italic>-values (≤ 0.12), as shown in Tables <xref ref-type="table" rid="T1">1</xref> and <xref ref-type="table" rid="T2">2</xref>. This result is rather consistent with a formerly accepted view of vertebrate classification, but is inconsistent with the currently accepted view. The mitochondrial tree shown in Figure <xref ref-type="fig" rid="F2">2D</xref> was significantly rejected by the KH test, if its likelihood value was calculated using nuclear DNA-coded genes; its log-likelihood difference from the ML tree was 34.9 ± 18.0 (<italic>P</italic>-value = 0.03) and 42.3 ± 20.8 (<italic>P</italic>-value = 0.02) in concatenated alignment analysis and totalml analysis, respectively.</p><p>We also analyzed a mitochondrial data set and confirmed the monophyly of holosteans and chondrosteans. In contrast to the high support value (100%) by the <sc>MRBAYES</sc> program for this relationship, however, the RELL BP value was only 71%. The likelihood difference between topologies <italic>A </italic>and <italic>D </italic>of Figure <xref ref-type="fig" rid="F2">2</xref> was 11.9 ± 13.3 (<italic>P</italic>-value = 0.18), which is not significant, as Inoue <italic>et al. </italic>[<xref ref-type="bibr" rid="B16">16</xref>] noted. Considering the Bayesian inference often results in erroneously high support values [<xref ref-type="bibr" rid="B28">28</xref>,<xref ref-type="bibr" rid="B43">43</xref>], the inconsistency between the present inference and that based on mitochondrial sequences might be caused by the artifact of the Bayesian inference.</p></sec><sec><title>Bichir is the basal ray-finned fish</title><p>The phylogenetic position of bichir has long been controversial as well, as it shares many characters with both lobe-finned fishes and ray-finned fishes [<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B44">44</xref>]. Most morphologists currently place bichir to a basal position in ray-finned fishes [<xref ref-type="bibr" rid="B38">38</xref>,<xref ref-type="bibr" rid="B45">45</xref>], although it remains possible that bichir and chondrosteans form a monophyletic group [<xref ref-type="bibr" rid="B14">14</xref>]. Recently, Venkatesh <italic>et al. </italic>[<xref ref-type="bibr" rid="B9">9</xref>,<xref ref-type="bibr" rid="B46">46</xref>] found one molecular synapomorphy indicating that bichir is the basal ray-finned fish, under the assumption that cartilaginous fishes are basal in the jawed-vertebrate tree. Our result strongly confirms the result from molecular synapomorphies: bichir is placed at the deepest position in ray-finned fishes, and the bootstrap probabilities are 98.1% and 95.1% in concatenated alignment analysis and in totalml analysis, respectively, as shown in Figure <xref ref-type="fig" rid="F3">3</xref> and Table <xref ref-type="table" rid="T3">3</xref>. The alternative hypothesis that bichir and chondrosteans form a monophyletic group was not supported; its log-likelihood difference from the ML tree is 37.4 ± 18.5 (<italic>P</italic>-value = 0.02) and 33.3 ± 19.5 (<italic>P</italic>-value = 0.04) in concatenated alignment analysis and totalml analysis, respectively.</p></sec><sec><title>Chimaeras and other cartilaginous fishes form a monophyletic group</title><p>Some paleontologists have pointed out the possibility that chimaeras were derived from placoderms independently from other cartilaginous fishes (eg, [<xref ref-type="bibr" rid="B38">38</xref>]). To test this possibility, we have isolated the genes listed in Materials and methods, except for ALDb, from a plownose chimaera, <italic>Callorhinchus callorhynchus</italic>, and have inferred its phylogenetic position based on the concatenated alignment of 2,431 amino acid residues. The resulting tree significantly supported the monophyly of cartilaginous fishes including chimaeras (as shown by dash-dotted line in Figure <xref ref-type="fig" rid="F3">3</xref>) with the RELL bootstrap probability of 100%. Mitochondrial data also support this relationship [<xref ref-type="bibr" rid="B3">3</xref>].</p></sec></sec><sec><title>Conclusions</title><p>Molecular phylogenetic analyses of jawed vertebrates based on mitochondrial sequences often result in confusing inferences which are obviously inconsistent with generally accepted trees. To obtain a robust tree of jawed vertebrates, we have cloned and sequenced seven nuclear DNA-coded genes from thirteen vertebrate species and have carried out phylogenetic analyses including thirteen jawed vertebrates from eight major groups and an outgroup (a cyclostome and a lancelet) based on the maximum likelihood method. We have shown that (<italic>i</italic>) cartilaginous fishes are basal to other jawed vertebrates. This is consistent with generally accepted view, but is inconsistent with mitochondrial trees. (<italic>ii</italic>) Living holosteans, comprising bowfin and gar, form a monophyletic group which is the sister group to teleost fishes. This is consistent with a formerly prevalent view of vertebrate classification, but inconsistent with both of the current morphology-based and mitochondrial sequence-based trees. (<italic>iii</italic>) The bichir is the basal ray-finned fish. (<italic>iv</italic>) Tetrapods and lungfish form a monophyletic cluster in the tree inferred from the concatenated alignment, being consistent with currently accepted view. It remains also possible that tetrapods are more closely related to ray-finned fishes than to lungfishes.</p><p>The present results are statistically solid and highly consistent with traditional views based on morphological and paleontological evidence. Comparing with trees inferred from mitochondrial sequences, which often provide obviously bizarre phylogeny, these nuclear DNA-coded genes probably have more accurate phylogenetic information. More intensive taxonomic sampling, particularly inclusion of coelacanth, would provide more solid inference for the origin of tetrapods and other phylogenetic problems currently discussed mainly based on mitochondrial sequences. An extended analysis including coelacanth sequences is in progress.</p></sec><sec sec-type="materials\|methods"><title>Materials and methods</title><sec><title>Isolation and sequencing of cDNAs</title><p>We have carried out a phylogenetic analysis of jawed vertebrates based on seven nuclear DNA-coded genes from six ray-finned fishes, three tetrapods, two lobe-finned fishes, three cartilaginous fishes and an outgroup (a cyclostome and a lancelet). For plownose chimaera, only six gene sequences excluding ALDb sequence were available for analysis. The names and abbreviations of proteins used in the present analysis are as follows: ALDa, ALDb and ALDc, fructose-bisphosphate aldolase A, B and C, respectively; G6PD, glucose-6-phosphate 1-dehydrogenase; GAG, a trifunctional protein with glycinamide ribonucleotide synthetase (GARS)-aminoimidazole ribonucleotide synthetase (AIRS)-glycinamide ribonucleotide formyltransferase (GART); PGK, phosphoglycerate kinase; TPI, triosephosphate isomerase.</p><p>Species and tissues used for RNA extraction are as follows: <italic>Ambystoma mexicanum</italic>, axolotl (gill and tail); <italic>Lepidosiren paradoxa</italic>, South American lungfish (brain, liver, heart and muscle); <italic>Protopterus annectens</italic>, African lungfish (pectoral fin); <italic>Oryzias latipes</italic>, Japanese medaka (liver); <italic>Lepisosteus osseus</italic>, longnose gar (brain, liver and muscle); <italic>Amia calva</italic>, bowfin (caudal fin); <italic>Acipenser baerii</italic>, Siberian sturgeon (brain and liver); <italic>Polypterus ornatipinnis</italic>, bichir (brain, liver and muscle); <italic>Cephaloscyllium umbratile</italic>, swell shark (brain, liver and muscle); <italic>Potamotrygon motoro</italic>, freshwater stingray (brain and liver); <italic>Callorhinchus callorhynchus</italic>, plownose chimaera (embryo); <italic>Lethenteron reissneri</italic>, lamprey (larva); <italic>Eptatretus burgeri</italic>, inshore hagfish (liver); <italic>Branchiostoma belcheri</italic>, lancelet (whole body). Total RNAs were extracted using TRIZOL Reagent (GIBCO BRL). These total RNAs were reverse-transcribed to cDNAs using oligo(dT) primer with reverse transcriptase (SuperScript II, GIBCO BRL) and were used as templates for PCR amplification with Expand High-Fidelity PCR System (Roche Diagnostics). The sense and antisense degenerate primers of the seven proteins were designed from conserved amino acid residues of each gene as shown in Table <xref ref-type="table" rid="T4">4</xref>. PCR amplification was carried out under annealing condition of 43–50°C with the sense and antisense primers. Nested PCR with a proper set of sense and antisense primers was carried out with primary PCR product, when needed.</p><p>The PCR products were separated in 1.5% agarose gel containing ethidium bromide. Products of expected size were isolated as gel slices, purified using DNA purification kit (TOYOBO), and cloned into pT7Blue vector (Novagen). Then, <italic>Escherichia coli </italic>strain DH5α (TOYOBO) was transformed with a ligated vector. More than three independent clones were isolated for each gene and sequenced by dideoxy chain termination method using BigDye Terminator Cycle Sequencing Ready Kit (Applied Biosystems) and ABI PRISM 377 and 3100 DNA sequencers (Applied Biosystems).</p><p>The 3' ends of cDNAs were amplified using 3'RACE System for Rapid Amplification of cDNA Ends (GIBCO BRL). The amplified fragments were purified, subcloned and sequenced in the same way as above.</p></sec><sec><title>Sequence data</title><p>The following sequence data was taken from the DDBJ/EMBL/GenBank database: the seven gene sequences from human, mouse and <italic>Takifugu rubripes </italic>(fugu); ALD gene sequences from <italic>Eptatretus burgeri </italic>(inshore hagfish), <italic>Lethenteron japonicum </italic>(Japanese lamprey) and <italic>B. belcheri</italic>; TPI gene sequences from <italic>L. reissneri </italic>and <italic>B. belcheri</italic>. The DDBJ/EMBL/GenBank accession number of each sequence data is shown in Table <xref ref-type="table" rid="T5">5</xref>.</p></sec><sec><title>Phylogenetic tree inference</title><p>Multiple alignments of amino acid sequences were carried out by <sc>MAFFT</sc>[<xref ref-type="bibr" rid="B47">47</xref>], a multiple sequence alignment program recently developed by us, and manually inspected on the <sc>XCED</sc> sequence alignment editor.</p><p>Using the cyclostome and lancelet sequences as an outgroup, phylogenetic analyses have been carried out by <sc>GAMT</sc>[<xref ref-type="bibr" rid="B26">26</xref>], a genetic algorithm-based ML method, with the JTT-F model [<xref ref-type="bibr" rid="B48">48</xref>,<xref ref-type="bibr" rid="B49">49</xref>]. Heterogeneity of evolutionary rates among sites was modeled by a discrete Γ distribution [<xref ref-type="bibr" rid="B50">50</xref>] with the optimized shape parameter α for each protein. A limited number of candidate tree topologies were generated by the following procedure and subjected to the comparison of likelihood values and statistical tests.</p><p>For alignment constructed from each of the seven protein sequences and that of the concatenated sequences, the <sc>GAMT</sc> program [<xref ref-type="bibr" rid="B26">26</xref>] was applied. <sc>GAMT</sc> is a method for searching for the ML tree based on the genetic algorithm (GA), outputting the best tree as well as multiple alternative trees each of which has a likelihood value close to the best one. The outline of the procedure is as follows: (<italic>i</italic>) The initial population of tree topologies is generated by applying the NJ method [<xref ref-type="bibr" rid="B51">51</xref>] to alignments generated by the bootstrap resampling. (<italic>ii</italic>) The fitness value for the best tree in the population is set to a constant value <italic>N</italic><sub><italic>w</italic></sub>, that for the second best one is set to <italic>N</italic><sub><italic>w</italic></sub>-1, and so on. (<italic>iii</italic>) A new tree <italic>i </italic>is generated by applying either of two operators (mutation or crossover) to trees picked up from the current population according to their fitness values. (<italic>iv</italic>) If there is a tree <italic>j </italic>with Δ ln<italic>L</italic><sub><italic>j</italic></sub>/σ<sub><italic>j </italic></sub>> ΔlnL<sub><italic>i</italic></sub>/σ<sub><italic>i </italic></sub>in the current population, tree <italic>i </italic>replaces tree <italic>j</italic>, where Δ ln<italic>L</italic><sub><italic>i</italic></sub>(= ln<italic>L</italic><sub><italic>best</italic></sub>-ln<italic>L</italic><sub><italic>i</italic></sub>) is the log-likelihood difference between tree <italic>i </italic>and the best tree in the current population, and σ<sub><italic>i </italic></sub>is the standard deviation of Δ ln<italic>L</italic><sub><italic>i</italic></sub>. (<italic>v</italic>) The procedure from steps (<italic>ii</italic>-<italic>iv</italic>) is repeated <italic>N</italic><sub><italic>e </italic></sub>times. Parameters used in the present analysis are as follows: <italic>N</italic><sub><italic>w </italic></sub>= 100 and <italic>N</italic><sub><italic>e </italic></sub>= 10,000.</p><p>This procedure was repeated for the seven different proteins and the concatenated alignment. The resulting topologies with Δ ln<italic>L</italic><sub><italic>i </italic></sub>< σ<sub><italic>i </italic></sub>from each of the seven proteins were selected as candidate ones. Additional topologies with σ<sub><italic>i </italic></sub>≤ Δ ln<italic>L</italic><sub><italic>i </italic></sub>< 3 σ<sub><italic>i </italic></sub>were also selected from the concatenated alignment. For the all candidate topologies selected, the log-likelihood values were computed from the seven protein sequence data sets, and totaled following the procedure of Kishino <italic>et al. </italic>[<xref ref-type="bibr" rid="B52">52</xref>] using the <sc>TOTALML</sc> program from the <sc>MOLPHY</sc> package [<xref ref-type="bibr" rid="B48">48</xref>]. This procedure is hereafter referred to as 'totalml' analysis. Another type of log-likelihood value was computed from the concatenated alignment for each candidate topology. This procedure is hereafter referred to as 'concatenated alignment' analysis.</p></sec><sec><title>Statistical tests</title><p>The following known statistical tests (<italic>i-vii</italic>) were carried out in the present analysis. (<italic>i</italic>) Kishino-Hasegawa (KH) test and (<italic>ii</italic>) approximately unbiased (AU) test were carried out using the <sc>CONSEL</sc> package [<xref ref-type="bibr" rid="B53">53</xref>]. (<italic>iii</italic>) Bayesian posterior probability (BPP) value of each of the candidate topologies was also computed using the <sc>CONSEL</sc> package [<xref ref-type="bibr" rid="B53">53</xref>]. (<italic>iv</italic>) Bootstrap probability value for a hypothesis was computed by applying the RELL (resampling of estimated log-likelihoods) approximation [<xref ref-type="bibr" rid="B52">52</xref>] to the candidate topologies and then totaling the bootstrap probabilities of the candidates supporting the hypothesis. This value is referred to as RELL BP value or simply BP value. (<italic>v</italic>) The <sc>MRBAYES</sc> program [<xref ref-type="bibr" rid="B30">30</xref>] was applied to the concatenate alignment. Default settings were used except for aamodel = jones, rates = gamma, ngen = 200,000 and burnin = 100. (<italic>vi</italic>) The <sc>TREE-PUZZLE</sc> program [<xref ref-type="bibr" rid="B29">29</xref>] was applied to the concatenate alignment. (<italic>vii</italic>) Intact bootstrap probability was also computed for the concatenated alignment. The calculation procedure is simple, but time-consuming; the ML tree was inferred by the <sc>GAMT</sc> program independently for each of the 100 alignments generated by bootstrap resampling, and the number of the ML trees supporting the hypothesis was counted. This value is referred to as intact BP value. The first four methods (<italic>i-iv</italic>), which are for testing given candidate topologies, were applied to both totalml and concatenated alignment analyses, whereas the last three methods (<italic>v-vii</italic>), each of which is for inferring a phylogenetic tree, were applied to the concatenated alignment without setting any candidate topologies.</p></sec></sec><sec><title>Authors' contributions</title><p>K Kikugawa carried out the labwork and phylogenetic analysis. K Katoh carried out phylogenetic analysis and wrote the manuscript. S Kuraku and N Iwabe developed laboratory protocols, provided a part of sequence data, and supervised the labwork and phylogenetic analysis. H Sakurai and O Ishida carried out collection and preparation of plownose chimaera embryos. T Miyata designed the study and edited the manuscript.</p></sec><sec><title>List of abbreviations</title><p>ALDa, fructose-bisphosphate aldolase A; ALDb, fructose-bisphosphate aldolase B; ALDc, fructose-bisphosphate aldolase C; G6PD, glucose-6-phosphate 1-dehydrogenase; GAG, a trifunctional protein with glycinamide ribonucleotide synthetase (GARS)-aminoimidazole ribonucleotide synthetase (AIRS)-glycinamide ribonucleotide formyltransferase (GART); PGK, phosphoglycerate kinase; TPI, triosephosphate isomerase; LSU rRNA, large-subunit ribosomal RNA.</p></sec>
TLR4 signaling is essential for survival in acute lung injury induced by virulent <italic>Pseudomonas aeruginosa </italic>secreting type III secretory toxins	<sec><title>Background</title><p>The relative contributions of the cytotoxic phenotype of <italic>P. aeruginosa </italic>expressing type III secretory toxins and an immunocompromised condition lacking normal Toll-like receptor 4 (TLR4) signaling in the pathogenesis of acute lung injury and sepsis were evaluated in a mouse model for <italic>Pseudomonas aeruginosa </italic>pneumonia. By using lipopolysaccharide-resistant C3H/HeJ mice missing normal TLR4 signaling due to a mutation on the <italic>tlr4 </italic>gene, we evaluated how TLR4 signaling modulates the pneumonia caused by cytotoxic <italic>P. aeruginosa </italic>expressing type III secretory toxins.</p></sec><sec sec-type="methods"><title>Methods</title><p>We infected C3H/HeJ or C3H/FeJ mice with three different doses of either a cytotoxic <italic>P. aeruginosa </italic>strain (wild type PA103) or its non-cytotoxic isogenic mutant missing the type III secretory toxins (PA103Δ<italic>UT</italic>). Survival of the infected mice was evaluated, and the severity of acute lung injury quantified by measuring alveolar epithelial permeability as an index of acute epithelial injury and the water to dry weight ratios of lung homogenates as an index of lung edema. Bacteriological analysis and cytokine assays were performed in the infected mice.</p></sec><sec><title>Results</title><p>Development of acute lung injury and sepsis was observed in all mouse strains when the cytotoxic <italic>P. aeruginosa </italic>strain but not the non-cytotoxic strain was instilled in the airspaces of the mice. Only C3H/HeJ mice had severe bacteremia and high mortality when a low dose of the cytotoxic <italic>P. aeruginosa </italic>strain was instilled in their lungs.</p></sec><sec><title>Conclusion</title><p>The cytotoxic phenotype of <italic>P. aeruginosa </italic>is the critical factor causing acute lung injury and sepsis in infected hosts. When the <italic>P. aeruginosa </italic>is a cytotoxic strain, the TLR4 signaling system is essential to clear the batcteria to prevent lethal lung injury and bacteremia.</p></sec>	<contrib id="A1" contrib-type="author"><name><surname>Faure</surname><given-names>Karine</given-names></name><xref ref-type="aff" rid="I1">1</xref><email>[email protected]</email></contrib><contrib id="A2" corresp="yes" contrib-type="author"><name><surname>Sawa</surname><given-names>Teiji</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A3" contrib-type="author"><name><surname>Ajayi</surname><given-names>Temitayo</given-names></name><xref ref-type="aff" rid="I3">3</xref><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib><contrib id="A4" contrib-type="author"><name><surname>Fujimoto</surname><given-names>Junichi</given-names></name><xref ref-type="aff" rid="I5">5</xref><email>[email protected]</email></contrib><contrib id="A5" contrib-type="author"><name><surname>Moriyama</surname><given-names>Kiyoshi</given-names></name><xref ref-type="aff" rid="I2">2</xref><email>[email protected]</email></contrib><contrib id="A6" contrib-type="author"><name><surname>Shime</surname><given-names>Nobuaki</given-names></name><xref ref-type="aff" rid="I6">6</xref><email>[email protected]</email></contrib><contrib id="A7" contrib-type="author"><name><surname>Wiener-Kronish</surname><given-names>Jeanine P</given-names></name><xref ref-type="aff" rid="I2">2</xref><xref ref-type="aff" rid="I3">3</xref><xref ref-type="aff" rid="I4">4</xref><email>[email protected]</email></contrib>	Respiratory Research	<sec><title>Background</title><p><italic>Pseudomonas aeruginosa </italic>(<italic>P. aeruginosa</italic>) frequently causes acute pneumonia in immunocompromised individuals. Patients who are mechanically ventilated are especially at high risk for developing <italic>P. aeruginosa </italic>pneumonia [<xref ref-type="bibr" rid="B1">1</xref>,<xref ref-type="bibr" rid="B2">2</xref>]. In addition, patients with <italic>P. aeruginosa </italic>pneumonia are more likely to develop bacteremia, septic shock, and multiple organ failure, and have a higher mortality than patients with pneumonia due to other infectious agents [<xref ref-type="bibr" rid="B2">2</xref>,<xref ref-type="bibr" rid="B3">3</xref>].</p><p>A poorer prognosis of patients with <italic>P. aeruginosa </italic>pneumonia is associated with specific virulence factors of <italic>P. aeruginosa</italic>. We have reported that the ability of these bacteria to cause acute epithelial injury involves the expression of virulent toxins that are translocated directly into eukaryotic cells via the type III secretion (TTS) system [<xref ref-type="bibr" rid="B4">4</xref>-<xref ref-type="bibr" rid="B9">9</xref>]. The TTS system is utilized by most virulent pathogenic gram-negative bacteria, including <italic>Yersinia, Salmonella, Shigella</italic>, and <italic>E. coli </italic>[<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>]. The TTS system functions as a molecular syringe to deliver toxins directly into the cytosol of cells and inhibit host immune responses [<xref ref-type="bibr" rid="B12">12</xref>-<xref ref-type="bibr" rid="B14">14</xref>]. In animal models, cytotoxic <italic>P. aeruginosa </italic>strains expressing the TTS toxins were able to consistently induce septic shock by causing the dissemination of inflammatory mediators from injured lungs into the circulation [<xref ref-type="bibr" rid="B15">15</xref>]. The administration of anti-PcrV IgG blocked the TTS toxin translocation by <italic>P. aeruginosa </italic>and protected animals from lethal pneumonia [<xref ref-type="bibr" rid="B16">16</xref>-<xref ref-type="bibr" rid="B18">18</xref>]. Recently, several studies also demonstrated that patients infected with cytotoxic strains of <italic>P. aeruginosa </italic>expressing the TTS toxins have a high risk of mortality [<xref ref-type="bibr" rid="B19">19</xref>,<xref ref-type="bibr" rid="B20">20</xref>].</p><p>In infections with gram-negative bacteria including <italic>P. aeruginosa</italic>, endotoxin (lipopolysaccharide, LPS), a component of bacterial outer membranes, has been reported to be responsible for the pathogenesis of acute organ injury and sepsis by provoking host inflammatory responses and inducing systemic inflammatory syndromes [<xref ref-type="bibr" rid="B21">21</xref>]. The acute immune response to LPS occurs mainly through Toll-like receptor 4 (TLR4) [<xref ref-type="bibr" rid="B22">22</xref>,<xref ref-type="bibr" rid="B23">23</xref>]. In animal models of <italic>P. aeruginosa </italic>pneumonia, release and accumulation of inflammatory cytokines primarily occurs in the airspaces of the infected lung. The inflammatory cytokines produced in the lung are able to disseminate into the circulation across the lung epithelium damaged by the toxins secreted from cytotoxic <italic>P. aeruginosa, </italic>leading to septic shock [<xref ref-type="bibr" rid="B15">15</xref>]. The administration of interleukin-10, an anti-inflammatory cytokine, suppresses the systemic inflammatory mediators leaking into the circulation from the lungs and improves the survival of infected animals [<xref ref-type="bibr" rid="B15">15</xref>]. Therefore, two virulent components of <italic>P. aeruginosa</italic>, the TTS system and LPS, appeared to be important in causing sepsis associated with <italic>P. aeruginosa </italic>pneumonia.</p><p>In contrast, the inflammatory response to the airspace instillation of LPS alone does not cause acute lung epithelial injury [<xref ref-type="bibr" rid="B24">24</xref>]. In addition, neither lung injury nor septic shock can be induced by the airspace instillation of non-cytotoxic <italic>P. aeruginosa </italic>strains that are defective in the production of TTS toxins [<xref ref-type="bibr" rid="B15">15</xref>]. Therefore, whether the response to LPS through TLR4 becomes deleterious or beneficial for infected hosts appears to be influenced by the virulence of <italic>P. aeruginosa. </italic>Clinically, LPS-TLR4 signaling has been a therapeutic target for preventing acute lung injury and sepsis. However, the blockade of LPS-TLR4 signaling may be harmful because recognition of bacterial LPS through TLR4 and subsequent proinflammatory responses in infected hosts must be important for innate immunity against pathogens. In addition, the presence of endogenous ligands for TLRs has been reported [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. This implies that there are regulatory roles of TLR signaling in inflammatory and immune responses against not only infectious microbes but also endogenous harmful stimuli.</p><p>In this investigation, the contributions of the cytotoxic phenotype of <italic>P. aeruginosa </italic>and LPS-TLR4 signaling were evaluated in a mouse model of <italic>P. aeruginosa </italic>pneumonia. Using LPS-resistant C3H/HeJ mice missing normal TLR4 signaling due to a mutation on the <italic>tlr4 </italic>gene [<xref ref-type="bibr" rid="B27">27</xref>], we evaluated how TLR4 signaling modulated the development of <italic>P. aeruginosa </italic>pneumonia induced by the airspace instillation of either a cytotoxic <italic>P. aeruginosa </italic>strain or its non-cytotoxic isogenic mutant missing all TTS toxins. The outcomes suggested that the role of TLR4 signaling in infected hosts was important in clearance of the <italic>P. aeruginosa </italic>when the bacteria causing pneumonia is a cytotoxic strain expressing the type III secretory toxins.</p></sec><sec sec-type="methods"><title>Methods</title><sec><title>Animals</title><p>Male C3H/FeJ and C3H/HeJ mice (8- to 12-week-old) were purchased from Jackson Laboratories (Bar Harbor, ME). Animals were housed in cages with filter tops in specific pathogen-free conditions. Sterile food and water were provided <italic>ad lib</italic>. All experiments were done in compliance with the Animal Care Committee Rules of the University of California at San Francisco, and all protocols were approved.</p></sec><sec><title><italic>P. aeruginosa </italic>strains</title><p><italic>P. aeruginosa </italic>parental PA103 strain and its isogenic mutant PA103Δ<italic>UT </italic>missing both ExoU and ExoT toxins were generously provided by Dara W. Frank (Medical College of Wisconsin, Milwaukee, WI) [<xref ref-type="bibr" rid="B15">15</xref>]. Bacteria from frozen stocks were streaked onto trypticase soy agar plates and grown in a deferrated dialysate of tryptic soy broth supplemented with 10 mM nitrilotriacetic acid (Sigma Chemical), 1% glycerol, and 100 mM monosodium glutamate and grown at 33°C for 13 h in a shaking incubator. Cultures were centrifuged at 8,500 × g for 5 minutes, and the bacterial pellet was washed twice and diluted into the appropriate number of CFU/ml in lactated Ringer's solution, determined by spectrophotometry. The bacterial concentration was confirmed by plating out the known dilutions of the samples on sheep blood agar plates.</p></sec><sec><title>Intratracheal instillation</title><p>Mice were briefly anesthetized with inhaled sevoflurane (Ultane, Abbot Laboratories, Abbot Park, Illinois, USA) and placed supine, at an angle of approximately 30°. Fifty microliters of bacterial solution were instilled slowly in the left lobe of the lungs through a gavage needle (Modified animal feeding needle, 24 G, Popper & Sons, Inc., New Hyde Park, NY) inserted into the trachea via the oropharynx. The proper insertion of the needle in the trachea was confirmed by palpation of the tip of the blunt needle through the skin and the observation of movement of the solution inside the syringe with respiratory efforts [<xref ref-type="bibr" rid="B28">28</xref>].</p></sec><sec><title>Quantification of acute lung injury</title><p>The severity of epithelial lung injury was quantified by measuring the efflux of the <sup>131</sup>I-labeled albumin (the alveolar protein tracer) across the alveolar epithelium into the circulation, as previously described [<xref ref-type="bibr" rid="B28">28</xref>]. The alveolar instillate was made up of 0.5 μCi of <sup>131</sup>I-labeled human serum albumin (Merck-Frost, Quebec, Canada) and 0.05 mg of anhydrous Evan's blue in 2 ml of lactated Ringer's solution with 5% BSA. An appropriate quantity of the specified <italic>P. aeruginosa </italic>was added to the 50 μl volume, if the experiment was to include bacteria. The total radioactivity (in counts/minute/g) in the instillate was measured in a γ-ray counter (Auto-Gamma, model 5550, Packard, Downers Grove, IL). Four hours after the instillation, mice were deeply anesthetized with intraperitoneal sodium pentobarbital. Blood was collected by a carotid arterial puncture and median sternotomies were done to harvest and measure the radioactivity of pleural fluids, lungs, tracheas, oropharynxes and stomachs. The percentage of radioactive albumin that left the lung and entered the circulation was calculated by multiplying the counts per gram measured in the terminal blood sample times the blood volume of the animal (body weight × 0.07). The lungs were homogenized in sterile containers with sterile 0.9% saline solution. Lung homogenates were placed in pre-weighed aluminium pans and dried to constant weight in an oven at 80°C for 3 days. The water to dry weight ratios of lung homogenates were used as an index of lung edema.</p></sec><sec><title>Bacteriological assay</title><p>Blood was plated onto sheep blood agar plates for quantitative assessment of CFU/ml. Lung homogenates were sequentially diluted then plated onto sheep blood agar plates. Colony forming units were calculated per gram of tissue.</p></sec><sec><title>Bronchoalveolar lavage (BAL)</title><p>Mouse lungs were lavaged by instilling and withdrawing 1 ml of phosphate buffered saline (PBS). Supernatants of centrifuged BAL fluids (BALF) were collected and processed for cytokine measurement.</p></sec><sec><title>Leukocyte counts in BALF</title><p>BALF samples were mixed with the solution including 0.01% crystal violet dye and 2.7% acetic acid for leukocyte staining and erythrocyte hemolysis. The number of leukocytes in BLAF was counted by a hemacytometer under a microscope.</p></sec><sec><title>Determination of cytokine levels</title><p>Biological activity of TNFα in plasma and BALF was measured by analyzing the cytotoxicity of TNFα on the fibrosarcoma cell line WEHI-13VAR, as reported previously [<xref ref-type="bibr" rid="B15">15</xref>,<xref ref-type="bibr" rid="B29">29</xref>]. Cytokine ELISA was performed for IL-10, as previously described [<xref ref-type="bibr" rid="B28">28</xref>]. Sample concentrations were calculated by comparison with standard curves using recombinant cytokine.</p></sec><sec><title>Statistical analysis</title><p>Mantel-Cox log rank test was used for survival analysis. One-way analysis of variance (ANOVA) followed by the Bonferroni multiple comparison test or the Dunn's Multiple Comparisons test were used for comparisons among 3 groups. Student's t-test or Mann-Whitney test were used for comparisons between 2 groups. The Kruskal-Wallis non-parametric test with Dunn's test was used for the comparisons of leukocyte counts in BALF. Significance was accepted at p < 0.05.</p></sec></sec><sec><title>Results</title><sec><title>Survival study</title><p>To examine whether a lack of normal TLR4 signaling affects the mortality induced by <italic>P. aeruginosa </italic>pneumonia, we performed survival studies using LPS-resistant C3H/HeJ mice. LPS-sensitive C3H/FeJ mice (the background strain of C3H/HeJ) were used as control groups. Mice were intratracheally infected with wild type <italic>P. aeruginosa </italic>PA103 at doses of either 5 × 10<sup>3 </sup>CFU (low dose), 6 × 10<sup>4 </sup>CFU (medium dose), or 5 × 10<sup>5 </sup>CFU (high dose), or with an isogenic mutant PA103Δ<italic>UT</italic>, which is missing the expression of TTS toxins (ExoU and ExoT), at a dose of 5 × 10<sup>5 </sup>CFU (high dose).</p><p>After the instillation of a low dose (5 × 10<sup>3 </sup>CFU) wild type PA103, 30% of LPS-resistant C3H/HeJ mice died in 3 days, while all C3H/FeJ survived for a week (Fig. <xref ref-type="fig" rid="F1">1a</xref>). However, statistically significant difference was not detected between the two strains of mice. After the instillation of a medium dose (6 × 10<sup>4 </sup>CFU) of wild type PA103, 65% of LPS-resistant C3H/HeJ mice died in 4 days, while all C3H/FeJ survived for a week (Fig. <xref ref-type="fig" rid="F1">1b</xref>). There was statistically significant difference between C3H/HeJ and C3H/FeJ mice. After the instillation of high dose (5 × 10<sup>5 </sup>CFU) wild type PA103, only 30% of C3H/FeJ and only 15% C3H/HeJ mice, respectively, survived for more than 3 days (Fig. <xref ref-type="fig" rid="F1">1c</xref>). The results suggest that, in the comparison between C3H/HeJ and C3H/FeJ mice, LPS-resistant C3H/HeJ mice were more susceptible to wild type PA103 at three different doses. The result implies that defective TLR4 signaling was not beneficial for the survival of infected mice with cytotoxic PA103.</p><fig position="float" id="F1"><label>Figure 1</label><caption><p><bold>Survival study. </bold>An inoculum of either PA103 (5 × 10<sup>5 </sup>CFU or 6 × 10<sup>4 </sup>CFU) or PA103Δ<italic>UT </italic>(5 × 10<sup>5 </sup>CFU) was instilled into the lung of mice, then the survival was monitored for 7 days. n = 10 / group. <italic>a. </italic>PA103 (5 × 10<sup>5 </sup>CFU). <italic>b</italic>. PA103 (6 × 10<sup>4 </sup>CFU). <italic>c</italic>. PA103Δ<italic>UT </italic>(5 × 10<sup>5 </sup>CFU). , <italic>p </italic>< 0.05 versus C3H/FeJ PA103 6 × 10<sup>4 </sup>CFU group.</p></caption><graphic xlink:href="1465-9921-5-1-1"/></fig><p>Next, we examined whether the mortality of infected mice depended on the phenotypes of the TTS toxins of <italic>P. aeruginosa</italic>. An isogenic mutant PA103Δ<italic>UT </italic>missing ExoU and ExoT was intratracheally administered at a high dose (5 × 10<sup>5 </sup>CFU) to the airspaces of C3H/HeJ and C3H/FeJ mice. Although instilled at a high dose, none of the mice died for a week (Fig. <xref ref-type="fig" rid="F1">1d</xref>). Therefore, the mortality we observed in the mice infected with a high dose of <italic>P. aeruginosa </italic>PA103 was totally dependent upon the TTS virulence of <italic>P. aeruginosa </italic>rather than the genotype of TLR4 in mice.</p></sec><sec><title>Acute lung injury</title><p>We quantified acute lung injury in the mice infected with <italic>P. aeruginosa</italic>. Acute lung epithelial damage was quantified by measuring the efflux of alveolar protein tracers from the lungs into the bloodstream (Fig. <xref ref-type="fig" rid="F2">2a</xref>). Lung edema was quantified by measuring the water to dry weight ratios of the lung (Fig. <xref ref-type="fig" rid="F2">2b</xref>). The instillation of a high dose (5 × 10<sup>5 </sup>CFU) of PA103 led to a significant increase of lung epithelial injury (Fig. <xref ref-type="fig" rid="F2">2a</xref>) and severe increases of lung edema (Fig. <xref ref-type="fig" rid="F2">2b</xref>) in both groups of mice. When the mice were infected with a medium dose (6 × 10<sup>4 </sup>CFU) of PA103, both severe lung epithelial injury and lung edema occurred in the both mouse strains (Fig. <xref ref-type="fig" rid="F2">2a</xref>). Although the injury levels in C3H/HeJ were slightly higher, these were not statistically significant. In contrast, the intratracheal instillation of PA103Δ<italic>UT </italic>even at a high dose (5 × 10<sup>5 </sup>CFU) did not cause lung injury in either of the mice (Fig. <xref ref-type="fig" rid="F2">2a</xref> and <xref ref-type="fig" rid="F2">2b</xref>). This implies that, in our mouse model, the bacterial dose and virulence of the bacteria, but not the TLR4 phenotype of mice, determines the initial severity of acute lung injury. Because more C3H/HeJ mice infected with a medium dose (6 × 10<sup>4 </sup>CFU) of PA103 died 3 days after infection in the survival studies (Fig. <xref ref-type="fig" rid="F1">1a</xref>), we analyzed lung edema 3 days after a medium dose instillation (6 × 10<sup>4 </sup>CFU) of PA103 in the strains of mice. In contrast to C3H/FeJ mice, C3H/HeJ mice showed a significantly higher level of lung edema (Fig. <xref ref-type="fig" rid="F2">2c</xref>).</p><fig position="float" id="F2"><label>Figure 2</label><caption><p><bold>Acute lung injury. </bold>Two parameters of lung damage were considered 4 h after the instillation made of the protein tracer <sup>131</sup>I-labeled albumin and either PA103 (5 × 10<sup>5 </sup>CFU or 6 × 10<sup>4 </sup>CFU), or PA103Δ<italic>UT </italic>(5 × 10<sup>5 </sup>CFU), or no bacteria. <italic>a</italic>. Efflux of protein tracer into the bloodstream was calculated as a percentage of the total instilled amount to quantify lung epithelial injury. <italic>b</italic>. Water to dry weight ratios of lungs as index of lung edema. Data are means ± SE. n = 3 in control groups (no bacteria), and n = 5 in other groups. , p < 0.05 versus C3H/FeJ, †, p < 0.05 versus no bacteria.</p></caption><graphic xlink:href="1465-9921-5-1-2"/></fig></sec><sec><title>Lung bacterial clearance</title><p>We quantified the ability of the different strains of mice to clear the bacteria from their infected lungs and the severity of bacteremia during infection. Four hours after a high dose instillation (5 × 10<sup>5 </sup>CFU) of wild type PA103, both strains of mice had similar quantities of bacteria in their lungs (Fig. <xref ref-type="fig" rid="F3">3a</xref>). However, 4 h after the instillation of a medium dose (6 × 10<sup>4 </sup>CFU) of the wild type PA103, there was a significantly larger quantity of bacteria in the lungs of the C3H/HeJ mice. The quantity of bacteria in the lungs of the C3H/HeJ mice was significantly increased after 3 days of infection (6 × 10<sup>4 </sup>CFU) compared to the quantity of bacteria in the C3H/FeJ mice (Fig. <xref ref-type="fig" rid="F3">3b</xref>). The lungs of C3H/HeJ infected with high dose (5 × 10<sup>5 </sup>CFU) PA103Δ<italic>UT </italic>also showed a significantly higher level of bacterial burden in their lungs in comparison to the lungs of the C3H/FeJ mice (Fig. <xref ref-type="fig" rid="F3">3a</xref>) although all the mice survived for a week after instillation of the same dose of PA103Δ<italic>UT </italic>in the survival study (Fig. <xref ref-type="fig" rid="F1">1c</xref>). This result implies that C3H/HeJ mice could not clear the bacteria efficiently from the infected lungs even 3 days after the infection (Fig. <xref ref-type="fig" rid="F3">3b</xref>). Persistent lung infection observed in C3H/HeJ mice explained higher edema (Fig. <xref ref-type="fig" rid="F2">2c</xref>) and mortality (Fig. <xref ref-type="fig" rid="F1">1a</xref>) of the mice infected with a medium dose of PA103.</p><fig position="float" id="F3"><label>Figure 3</label><caption><p><bold>Bacteriology. </bold>Mice were infected with PA103 (5 × 10<sup>5 </sup>CFU or 6 × 10<sup>4 </sup>CFU) or PA103Δ<italic>UT </italic>(5 × 10<sup>5 </sup>CFU). The number of viable bacteria remaining in the infected lungs and blood was counted 4 h and 2 days after the instillation. Lung homogenates were sequentially diluted with sterile water and placed on sheep blood agar plates for 24 h at 37°C. 100 μl of blood was plated on agar plates for 24 h at 37°C. <italic>a</italic>. Number of bacteria remaining in the lungs 4 h after infection. <italic>b</italic>. Number of bacteria remaining in the lungs 2 days after PA103 6 × 10<sup>4 </sup>CFU infection. <italic>c</italic>. Bacteremia (CFU/ml of blood) 4 h after infection. <italic>d</italic>. Bacteremia (CFU/ml of blood) 2 days after PA103 6 × 10<sup>4 </sup>CFU infection. n = 5 in groups. Data are means ± SE. , p < 0.05 versus C3H/FeJ.</p></caption><graphic xlink:href="1465-9921-5-1-3"/></fig></sec><sec><title>Bacteremia</title><p>We evaluated the severity of bacteremia in C3H/HeJ and C3H/FeJ mice. Only a high dose instillation of PA103 caused detectable bacteremia 4 h after the infection in either strain of mice (Fig. <xref ref-type="fig" rid="F3">3c</xref>). Among the two strains, C3H/HeJ mice showed a significantly higher level of bacteremia 4 h after infection (Fig. <xref ref-type="fig" rid="F3">3c</xref>), although more than 70% of mice in the both strains eventually died within a week after the instillation of bacteria (Fig. <xref ref-type="fig" rid="F1">1b</xref>). Neither the instillation of a high dose of PA103Δ<italic>UT </italic>nor the instillation of a medium dose of PA103 resulted in bacteremia 4 h after the infection (Fig. <xref ref-type="fig" rid="F3">3c</xref>). However, 3 days after the instillation of the medium dose PA103, C3H/HeJ mice showed a significantly greater level of bacteremia while C3H/FeJ mice had either minimal bacteremia or none, respectively (Fig. <xref ref-type="fig" rid="F3">3d</xref>). Therefore, the results imply that decreased bacterial clearance probably lead to a delay in the onset of bacteremia in C3H/HeJ mice.</p></sec><sec><title>Leukocytes in bronchoalveolar lavage fluids</title><p>We measured the number of leukocytes in BALFs which were collected 16 h after infection from the mice infected with either PA103 or PA103Δ<italic>UT </italic>at a medium dose (Fig. <xref ref-type="fig" rid="F4">4</xref>). More leukocytes were found in the BALF from the mice infected with PA103Δ<italic>UT </italic>than in the BALF from the mice infected with PA103. More leukocytes were found in the BALF of C3H/FeJ mice than in the C3H/FeJ mice. The number of leukocytes in the BALF of C3H/FeJ mice infected with PA103 was the lowest among the four groups, and significantly lower than that of C3H/HeJ mice infected with PA103Δ<italic>UT</italic>. The results implied that C3H/HeJ mice failed to recruit inflammatory cells efficiently into the site of infection. In addition, when infection was by virulent <italic>P. aeruginosa </italic>secreting TTS toxins, few leukocytes were detected in BALF, implying that virulent <italic>P. aeruginosa </italic>suppressed the proper inflammatory cell recruitment or eliminated the recruited leukocytes by its cytotoxicity.</p><fig position="float" id="F4"><label>Figure 4</label><caption><p><bold>The number of leukocytes in bronchoalveolar lavage fluids. </bold>The numbers of leukocytes were measured in the bronchoalveolar lavage fluids collected from the mice infected with either PA103 5 × 10<sup>4 </sup>CFU or PA103Δ<italic>UT </italic>5 × 10<sup>4 </sup>CFU. Closed and open circles mean each data. n = 5 in groups. Boxes are the first quartile with medians as central bars. , p < 0.05 versus C3H/HeJ infected with PA103Δ<italic>UT </italic>by the Kruskal-Wallis non-parametric test followed by the Dunn's test.</p></caption><graphic xlink:href="1465-9921-5-1-4"/></fig></sec><sec><title>TNFα and IL-10 levels</title><p>We measured concentrations of TNFα in BALF and in the serum of infected animals. TNFα in BALF was almost undetectable in C3H/HeJ mice 4 h after the instillation of a high dose (5 × 10<sup>5 </sup>CFU) of either wild type PA103 or of PA103Δ<italic>UT </italic>(Fig. <xref ref-type="fig" rid="F5">5a</xref>). Only in the C3H/FeJ mice there was a significant difference in the TNFα in BALFs between PA103 and PA103Δ<italic>UT </italic>infections (Fig. <xref ref-type="fig" rid="F5">5a</xref>). Only C3H/FeJ mice infected with a high dose of PA103 showed a significantly increased serum concentration of TNFα compared to those in the other strains of mice infected with either PA103 or PA103Δ<italic>UT </italic>(Fig. <xref ref-type="fig" rid="F5">5b</xref>).</p><fig position="float" id="F5"><label>Figure 5</label><caption><p><bold>Cytokines. </bold>TNFα and IL-10 levels were measured 10 h after <italic>P. aeruginosa </italic>infection (PA103 or PA103Δ<italic>UT </italic>5 × 10<sup>5 </sup>CFU). <italic>a</italic>. TNFα levels in bronchoalveolar lavage fluids. <italic>b</italic>. TNFα levels in serum. c. IL-10 levels in serum 10 h after PA103 infection. Data are means ± SE. n = 8 – 10/group. *, p < 0.05 versus C3H/HeJ.</p></caption><graphic xlink:href="1465-9921-5-1-5"/></fig><p>We also measured the serum concentration of the anti-inflammatory cytokine IL-10 in the mice infected with a high dose of PA103. The concentration of IL-10 4 h after infection was higher in the C3H/HeJ mice, but the difference not statistically significant (Fig. <xref ref-type="fig" rid="F5">5c</xref>).</p></sec></sec><sec><title>Discussion</title><p>This study was designed to determine the contributions of the TTS system and LPS of <italic>P. aeruginosa </italic>in the pathogenesis of acute lung injury and bacteria-induced death. To understand the role of TLR4 signaling in acute lung injury and sepsis, we compared LPS-resistant C3H/HeJ mice, which are refractory to the biological effects of LPS due to the mutation of <italic>tlr4, </italic>to LPS-sensitive C3H/FeJ mice [<xref ref-type="bibr" rid="B27">27</xref>]. The two strains of mice received tracheal instillation of either cytotoxic wild type <italic>P. aeruginosa </italic>PA103 or an isogenic strain PA103Δ<italic>UT </italic>which is missing two critical TTS toxins, ExoU and ExoT [<xref ref-type="bibr" rid="B15">15</xref>]. Three different doses of bacteria were tested in these mice. ExoU is a cytotoxin inducing necrotic cell death by a mechanism associated with its phospholipase activity [<xref ref-type="bibr" rid="B30">30</xref>]. ExoT is a 53-kDa exoenzyme S possessing ADP-ribosyltransferase and small GTPase-activating activities. ExoT blocks epithelial wound healing [<xref ref-type="bibr" rid="B31">31</xref>]. The administration of large doses of PA103Δ<italic>UT </italic>did not cause any lung injury nor affect the mortality of any strain of mice. Large doses of wild type PA103 caused acute injury, bacteremia and death in the both strains of mice. Although there appeared to be minor differences in survival, there were no statistically significant differences in the mortality rates between the strains of mice infected with large doses of PA103. These results suggest that the administration of large doses of bacteria overwhelms host defense, even in mice with an intact LPS-signaling pathway. Furthermore, although there was abnormal clearance of the non-cytotoxic PA103Δ<italic>UT </italic>in the C3H/HeJ mice in comparison to the C3H/FeJ mice, PA103Δ<italic>UT </italic>did not cause lung injury. As no lung injury occurred, PA103Δ<italic>UT </italic>did not disseminate, and all the mice survived. These results demonstrate that organ injury and mortality were not directly affected by <italic>P. aeruginosa </italic>LPS or the TLR4 signaling of the infected mice.</p><p>In contrast, an intact TLR4 signaling pathway was clearly essential for host protection from virulent PA103. More C3H/HeJ mice died from a medium dose of PA103, and some of C3H/HeJ mice still died even from the lowest dose of PA103. When lung injury was measured after a short interval, there was no statistical difference between the acute lung injury produced by similar doses in either the C3H/HeJ or C3H/FeJ mice. Therefore, a given dose of cytotoxic bacteria produced a reproducible quantity of lung injury, in a short interval, in both mice. However, when the infections were allowed to persist for a longer interval, the number of cytotoxic bacteria in the C3H/HeJ multiplied to a greater extent and caused lung edema later. The lack of bacterial clearance in the C3H/HeJ mice ultimately led to relatively larger quantities of lung injury, even when smaller doses of cytotoxic bacteria had been administered. The bacteria were able to multiply to injurious levels, implying that, in the absence of normal TLR4 signaling, cytotoxic PA103 can avoid the host immune clearance, disseminate into the bloodstream, and cause the death of the infected animals. The low leukocyte number and a lack of TNFα production in the infected airspace in the C3H/HeJ mice could indicate the lack of a normal inflammatory response that clears the bacteria from the infected site in these mice missing normal TLR4 signaling. Therefore, immune responses modulated by TLR4 signaling are important in the protecting hosts from cytotoxic <italic>P. aeruginosa </italic>expressing TTS toxins.</p><p>LPS is a static component of gram-negative bacteria, while the virulence associated with the TTS system relies on active energy production by live bacteria; the TTS apparatus can be upregulated and TTS toxins need to be translocated using bacterial energy [<xref ref-type="bibr" rid="B10">10</xref>,<xref ref-type="bibr" rid="B11">11</xref>,<xref ref-type="bibr" rid="B13">13</xref>]. Gram-negative bacteria such as <italic>E. coli </italic>and <italic>P. aeruginosa </italic>possess LPS as a component of their cell membranes, although there are many variations in its pattern [<xref ref-type="bibr" rid="B32">32</xref>]. However, only pathogenic strains of gram-negative bacteria possess the TTS system [<xref ref-type="bibr" rid="B33">33</xref>,<xref ref-type="bibr" rid="B34">34</xref>]. In <italic>P. aeruginosa</italic>, only strains expressing the TTS system can cause significant acute necrotic cell death and tissue damage, while non-cytotoxic <italic>P. aeruginosa </italic>missing TTS but still possessing LPS do not lead to tissue damage [<xref ref-type="bibr" rid="B5">5</xref>,<xref ref-type="bibr" rid="B7">7</xref>]. Recently, there have been reports of endogenous stimulators of TLRs [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>]. For TLR4, heparan sulfate, hyaluronan, HSP60 and HSP70, surfactant protein A, and β-defensin 2 have been all implicated as possible endogenous ligands or stimulators [<xref ref-type="bibr" rid="B25">25</xref>,<xref ref-type="bibr" rid="B26">26</xref>,<xref ref-type="bibr" rid="B35">35</xref>]. These facts suggest that TLRs play a role as surveillance receptors for tissue injury and tissue remodelling as well as for infection [<xref ref-type="bibr" rid="B35">35</xref>]. If TLR4 recognizes not only LPS but also degradation products of endogenous macromolecules from necrotic cells and microorganisms, and if this recognition provokes responses to limit tissue injury and induce remodelling, a lack of normal TLR4 signaling must be disadvantageous during infections involving severe tissue damage. The activation of this surveillance system must be more important when cytotoxic <italic>P. aeruginosa </italic>infects and causes tissue injury. Without having this surveillance system, hosts may not be able to survive infection involving severe tissue damage. Members of the TLR and interleukin-1 receptor (IL-1R) superfamily share an intracytoplasmic Toll-IL-1 receptor (TIR) domain, which mediates recruitment of the interleukin-1 receptor-associated kinase (IRAK) complex via TIR-containing adapter molecules. Recently, 3 unrelated children with an inherited IRAK-4 deficiency were reported. They were prone to infections caused by pyogenic bacteria [<xref ref-type="bibr" rid="B36">36</xref>]. Their blood and fibroblast cells did not activate NF-κB or mitogen-activated protein kinase (MAPK) and failed to induce downstream cytokines in response to any of the known ligands of TIR-bearing receptors. These findings suggest that in humans the normal TLR signaling pathway is crucial for protective immunity against specific virulent bacteria and is redundant against most other microorganisms. Our results suggest that, in <italic>P. aeruginosa </italic>infections, the identification of the phenotype or genotype of <italic>P. aeruginosa </italic>could be useful clinically.</p></sec><sec><title>Conclusions</title><p>In <italic>P. aeruginosa </italic>infections, the cytotoxic phenotype of <italic>P. aeruginosa </italic>is the most critical factor in causing acute lung injury and sepsis in infected hosts. TLR4 signaling is another important factor for the efficient clearance of bacteria in hosts infected with cytotoxic <italic>P. aeruginosa </italic>strains.</p></sec><sec><title>Authors' contributions</title><p>Karine Faure and Teiji Sawa carried out animal studies. Temitayo Ajayi drafted and edited the manuscript. Junichi Fujimoto, Nobuaki Shime and Kiyoshi Moriyama participated in the animal studies. Teiji Sawa and Jeanine P. Wiener-Kronish conceived of the study, and participated in its design and coordination. All authors read and approved the final manuscript.</p></sec><sec><title>Abbreviations</title><p>BAL: Bronchoalveolar lavage, BALF: Bronchoalveolar lavage fluid, <italic>P. aeruginosa</italic>:<italic>Pseudomonas aeruginosa, </italic>TTS: type III secretion, LPS: lipopolysaccharide, TLR: Toll-like receptor.</p></sec>
Experimental modulation of capsule size in <italic>Cryptococcus neoformans</italic>	<p>Experimental modulation of capsule size is an important technique for the study of the virulence of the encapsulated pathogen <italic>Cryptococcus neoformans</italic>. In this paper, we summarize the techniques available for experimental modulation of capsule size in this yeast and describe improved methods to induce capsule size changes. The response of the yeast to the various stimuli is highly dependent on the cryptococcal strain. A high CO<sub>2</sub> atmosphere and a low iron concentration have been used classically to increase capsule size. Unfortunately, these stimuli are not reliable for inducing capsular enlargement in all strains. Recently we have identified new and simpler conditions for inducing capsule enlargement that consistently elicited this effect. Specifically, we noted that mammalian serum or diluted Sabouraud broth in MOPS buffer pH 7.3 efficiently induced capsule growth. Media that slowed the growth rate of the yeast correlated with an increase in capsule size. Finally, we summarize the most commonly used media that induce capsule growth in <italic>C. neoformans.</italic> </p>	<contrib contrib-type="author" corresp="yes"><name><surname>Zaragoza</surname><given-names>Oscar</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Casadevall</surname><given-names>Arturo</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>Microbial capsules are structures found surrounding the cell body that confer particular characteristics to encapsulated organisms. In the case of the pathogenic fungus <italic>Cryptococcus neoformans</italic>, the capsule is an important virulence factor (see review in (<xref rid="B1" ref-type="bibr">1</xref>)). The <italic>C. neoformans</italic> capsule is composed of at least two polysaccharides glucuronoxylomannan and galactoxylomannan, and a much smaller proportion of mannoprotein (see review in (<xref rid="B2" ref-type="bibr">2</xref>)). This polysaccharide capsule has a large number of effects on the host, such as complement activation and depletion, Ab unresponsiveness, inhibition of leukocyte migration and inhibition of phagocytosis (<xref rid="B3" ref-type="bibr">3</xref>-<xref rid="B9" ref-type="bibr">9</xref>). </p><p>One of the characteristics of cryptococcal cells is great variation in capsule size depending on the environmental conditions. During <italic>in vitro</italic> culture conditions (<xref rid="B10" ref-type="bibr">10</xref>) the size of the capsule is usually small, although capsular enlargement can be induced by several factors, including high CO<sub>2</sub> and low iron (<xref rid="B11" ref-type="bibr">11</xref>, <xref rid="B12" ref-type="bibr">12</xref>). During mammalian infection, the size of the capsule increases dramatically (<xref rid="B13" ref-type="bibr">13</xref>-<xref rid="B16" ref-type="bibr">16</xref>), and it is thought that this phenomenon is necessary for cryptococcal pathogenesis (<xref rid="B11" ref-type="bibr">11</xref>). Hence, modulation of capsule size is an important subject of investigation for the cryptococcal field. To date, the mechanism of capsule growth remains poorly characterized. </p><p>The purpose of this paper is to review the different ways of manipulating capsule size and to highlight the options and limitations available to cryptococcal investigators who plan to manipulate the size of a cryptococcal strain. In particularly, we will review the great variability in the behavior of strains with different genetic backgrounds and the heterogeneity in the response in the same strain. We also suggest an experimental roadmap to identify the proper conditions to induce capsule size of a particular <italic>C. neoformans</italic> strain.</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><sec><title> <bold>Yeast strain and growth conditions</bold> </title><p>The serotype A <italic>C. neoformans </italic>strain H99 was kindly provided by Dr. John Perfect (Durham, NC) and used for capsule growth experiments. The yeast cells were grown overnight in 10 ml of acid Sabouraud medium (Difco<sup>TM</sup>, Catalogue number 238130, Sparks, MD) at 30<sup>o</sup>C with moderate shaking (150 r.p.m), collected in the logarithmic phase of growth by centrifugation (5 minutes at 1200 r.p.m. at room temperature), washed twice with PBS (137 mM NaCl, 2.7 mM KCl, 1.5 mM KH<sub>2</sub>PO<sub>4</sub>, 8.5 mM Na<sub>2</sub>HPO<sub>4</sub>) and incubated in the capsule growth inducing media overnight at 37<sup>o</sup>C. The cells were placed in six-wells polystyrene plates at a cell density of 5 x 10<sup>6</sup>-10<sup>7</sup> cells/ml in the designated media (listed in Table <xref rid="T1" ref-type="table">1</xref>) and capsule induction was performed for 18-24 hours.</p></sec><sec><title> <bold>India Ink staining and light microscopy</bold> </title><p>To observe and measure the size of the capsule, 10 mL of a cell suspension were mixed with an India Ink drop (Becton Dickinson, NJ, Cat. Number 261194) and observed in an Olympus AX70 microscope. Pictures were taken with a QImaging Retiga 1300 digital camera using the QCapture Suite V2.46 software (QImaging, Burnaby BC, Canada), and processed with Adobe Photoshop 7.0 for windows (San Jose, CA). At least five different fields were randomly chosen and photographed, and 40 to 60 cells were analyzed. </p></sec><sec><title> <bold>Measurement of capsule volume</bold> </title><p>To calculate the capsule volume, the diameter of the whole cell and the cell body were each measured with Adobe Photoshop 7.0 and capsule volume was defined as the difference between the volume of the whole cell (yeast cell + capsule) and the volume of the cell body (as limited by the cell wall). Volumes were calculated using the equation for volume of a sphere as (4π/3)(D/2)<sup>3</sup>. Between 15 and 40 cells were measured for each determination. For statistical analysis, t-test was used with Unistat 5.5 software for Excel. </p></sec></sec><sec><title>Results and Discussion</title><sec><title> <bold>Capsule size in the ecological niches of <italic>C. neoformans</italic> </bold> </title><p> <italic>C. neoformans </italic>is a pathogenic yeast that is acquired from the environment during mammalian infection. This organism can be isolated from several environmental niches and is commonly found in pigeon excreta, soils, and some trees (see review in (<xref rid="B1" ref-type="bibr">1</xref>)). For environmental isolates, the size of the capsule is uniformly small. During <italic>in vitro</italic> growth conditions in the media commonly used in research laboratories (e.g. Sabouraud medium), the size of the capsule is usually small (<xref rid="B10" ref-type="bibr">10</xref>). This is in contrast with the situation found during infection whereby cells with very large capsules are commonly found in tissue (<xref rid="B13" ref-type="bibr">13</xref>-<xref rid="B16" ref-type="bibr">16</xref>). When cells infect the host, there is a rapid increase in capsule size (<xref rid="B17" ref-type="bibr">17</xref>). This observation suggests the importance of the environmental conditions in determining the size of the cryptococcal capsule. In the following sections, we summarize and discuss some factors that influence capsule size and their interplay in laboratory conditions (see Fig. <xref rid="F1" ref-type="fig">1</xref> for a summary).</p><fig id="F1"><label>Fig. 1</label><caption><title>Fig. 1: Scheme of the factors that influence capsule size in <italic>C. neoformans</italic>.</title><p> The diagram summarizes the main factors involved in the modulation of capsule size in <italic>C. neoformans</italic>. In the upper part we represent factors that increase capsule size, and in the lower part, factors that inhibit capsule growth. Bar on left panel denotes 10 microns and applies also for panel on the right.</p></caption><graphic xlink:href="bpo_v6_p10_m68f1lg"/></fig></sec><sec><title> <bold>Classical stimuli used to increase capsule size</bold> </title><p>The size of the capsule of <italic>C. neoformans</italic> in most laboratory media is small, and ranges between 2-4 microns in transversal length, without considering the distance between the cell wall. Similarly, organisms recovered from the environment usually have small capsules. In the late fifties, Littman formulated a mineral medium in which the growth of the organisms was accompanied by capsule enlargement (<xref rid="B10" ref-type="bibr">10</xref>). The composition of this medium was based on the nutrients that the yeast was likely to have available in physiological fluids, such as the cerebrospinal fluid. That study highlighted the importance of certain nutrients in determining the size of the capsule, and identified several amino acids, vitamins, and carbon sources that contributed to this process. However, this method has not been extensively used to induce capsule size, possibly because it is not successful with all strains. Several decades later, Granger <italic>et al</italic>. identified CO<sub>2</sub> as an inducing factor for capsule growth (<xref rid="B11" ref-type="bibr">11</xref>). That study used strain H99. It showed, when incubated in DMEM medium in an atmosphere rich in CO<sub>2</sub>, induced capsule growth. This observation had potentially important physiological implications because it identified a product of mammalian respiration as a stimulus for capsule growth. Consequently, CO<sub>2</sub> is one of the most commonly used stimuli to manipulate capsule size. However, not all strains responded to CO<sub>2</sub> with increased capsule growth (<xref rid="B18" ref-type="bibr">18</xref>). In 1993, Vartivarian <italic>et al</italic>. observed that a low iron concentration also induced capsule growth (<xref rid="B12" ref-type="bibr">12</xref>). This medium (known as LIM for limited iron medium) has been extensively used in the literature to induce capsule size. More recently, we described mammalian serum as a potent activator of capsule growth (<xref rid="B18" ref-type="bibr">18</xref>). </p></sec><sec><title> <bold>Other factors that affect capsule size</bold> </title><p>Capsule growth is affected by a variety of factors, such as pH, osmolarity, carbon source, nutrient concentration and temperature. Alkaline conditions facilitate capsule growth (<xref rid="B11" ref-type="bibr">11</xref>, <xref rid="B18" ref-type="bibr">18</xref>-<xref rid="B21" ref-type="bibr">21</xref>), although a basic pH is not sufficient to mediate this effect. Increasing the pH of Sabouraud medium does not enhance capsule growth (<xref rid="B18" ref-type="bibr">18</xref>). High osmolarity may block capsule growth since high glucose concentrations inhibit capsule growth (<xref rid="B21" ref-type="bibr">21</xref>), and although this effect does not apply to any solute (<xref rid="B22" ref-type="bibr">22</xref>), sodium chloride produces the same effect (<xref rid="B21" ref-type="bibr">21</xref>, <xref rid="B22" ref-type="bibr">22</xref>). Finally, temperature seems to have an effect on the size of the cell. This can affect the size of the capsule (the diameter of the cell is used in deriving the size of the capsule). However, temperature alone does not affect the relative size of the capsule (<xref rid="B18" ref-type="bibr">18</xref>, <xref rid="B23" ref-type="bibr">23</xref>).</p></sec><sec><title> <bold>Problems with the reproducibility</bold> </title><p>One of the problems of the cryptococcal field is that strain capsule size and the enlargement phenomenon is poorly reproducible from laboratory to laboratory. Consequently, this effect has not been studied extensively despite its association with virulence. Heterogeneity in the response presumably reflects the genetic background, and also the complexity in the interplay between the different inducing factors. For example, both CO<sub>2</sub> and serum are very effective as inducing factors, but only in specific media (<xref rid="B18" ref-type="bibr">18</xref>). Neither serum nor CO<sub>2</sub> are able to induce capsule growth in media that is rich in nutrients, like Sabouraud media, and they require that the yeast is placed in either DME or PBS, respectively, to induce capsule growth (<xref rid="B18" ref-type="bibr">18</xref>). This implies that capsule growth is a highly complex phenomenon that is not determined by a unique factor.</p><p>There is great variation between different strains (inter-strain heterogeneity) in the response to the different stimuli used for inducing capsule growth (<xref rid="B18" ref-type="bibr">18</xref>). Stimuli that produce capsule enlargement in some strains do not produce it in others. This variability adds more complexity to this phenomenon. For example, the capsular growth induction by CO<sub>2</sub> or iron may differ from strain-to-strain (<xref rid="B12" ref-type="bibr">12</xref>, <xref rid="B18" ref-type="bibr">18</xref>). The reason for this inter-strain heterogeneity is not known, but most probably it is due to genetic differences between the strains. Consequently, for a given strain one must evaluate several media empirically to ascertain the best conditions for achieving capsule growth. Furthermore, certain strains demonstrate intra-strain heterogeneity in the response such that all cells may not respond in the same manner (<xref rid="B11" ref-type="bibr">11</xref>, <xref rid="B17" ref-type="bibr">17</xref>-<xref rid="B18" ref-type="bibr">18</xref>). For instance, it has been reported that when strain H99 is in the lung, there is a high heterogeneity in the size of the cells and capsules of the yeast (<xref rid="B17" ref-type="bibr">17</xref>). Similar heterogeneity was also reported for strain 24067 <italic>in vitro </italic>(<xref rid="B18" ref-type="bibr">18</xref>). In the case of strain 24067, the heterogeneity is probably due to a different behavior of the buds that emerge during the incubation in the inducing medium (<xref rid="B18" ref-type="bibr">18</xref>). The mechanisms responsible for intra- and inter-strain heterogeneity are not known, but it is likely that an explanation includes genetic differences combined with growth differences for individual cells in the medium used (<xref rid="B18" ref-type="bibr">18</xref>). Alternatively, intra-strain heterogeneity could reflect a phenotypic-switching phenomenon that is manifested at the cellular level.</p></sec><sec><title> <bold>New media able to induce capsule size</bold> </title><p>During our studies of <italic>C. neoformans</italic> biology, we serendipitously discovered that yeast cells of strain H99 placed in a Sabouraud medium diluted 10 times with H<sub>2</sub>O developed large capsules (Fig. <xref rid="F2" ref-type="fig">2</xref>). That result suggested that capsule enlargement was related to the cell growth rate. In addition, we observed that when Sabouraud medium was diluted, not in water, but in buffer with basic pH, such as MOPS, HEPES or PBS, the increase in capsule size was more noticeable (Fig. <xref rid="F2" ref-type="fig">2</xref>). Moreover, increasing the pH of diluted Sabouraud in water by addition of some drops of NaOH enhanced the efficiency of the process (Fig. <xref rid="F2" ref-type="fig">2</xref>).</p><fig id="F2"><label>Fig. 2</label><caption><title>Fig. 2: Capsule growth in different media containing diluted Sabouraud.</title><p> H99 cells were grown in Sabouraud medium, collected in the logarithmic phase of growth, washed and transferred to 2 ml of media indicated in each case. In some cases, a few drops of NaOH were used to manipulate the pH, while in other experiments MOPS or HEPES buffer were used. The cells were incubated overnight at 37ºC as described in material and methods. Bar on first panel denotes 10 microns and applies also for the rest of the panels.</p></caption><graphic xlink:href="bpo_v6_p10_m68f2lg"/></fig><p>The effect was observed by six hours and appeared complete by 24 h. Longer incubations did not significantly change the size of the capsule. Among the media evaluated, the biggest capsules were observed when pH was buffered with 50 mM MOPS pH 7.3, and was associated with noticeably slower growth (data not shown). These results provide new experimental conditions to increase capsule size in the absence of serum or CO<sub>2</sub> incubators. Comparison of the efficiency of diluted Sabouraud in MOPS buffer, with CO<sub>2</sub> and serum to previously established methods revealed that our methods were more consistent in promoting capsule growth (Fig. <xref rid="F3" ref-type="fig">3</xref>).</p><fig id="F3"><label>Fig. 3</label><caption><title>Fig. 3: Comparison of capsule growth on diluted Sabouraud, serum and CO<sub>2</sub>.</title><p> H99 cells were grown in Sabouraud overnight and then transferred to the following media: Diluted Sabouraud (1/10) in 50 mM MOPS buffer (pH 7.3), or 10% FBS (in PBS) or in DME in the presence of an atmosphere of 10% CO<sub>2</sub>. The cells were incubated in these media overnight at 37ºC. A) Picture of a representative cell in an India Ink suspension. The bar in the first panel denotes 10 microns, and this scale applies to the rest of pictures. B) Measurement of the capsule size for the samples described above. The size of the capsule was expressed as the total capsule volume. P value for all comparisons was below 0.001.</p></caption><graphic xlink:href="bpo_v6_p10_m68f3lg"/></fig></sec><sec><title> <bold>Capsule size during infection</bold> </title><p>The increase of capsule size during infection is believed to be biologically significant because of the virulence-enhancing properties of capsular polysaccharide. It is thought that this increase in capsule size is necessary for the virulence of the yeast since mutants that are unable to do this are not as virulent as wild type strains (<xref rid="B11" ref-type="bibr">11</xref>, <xref rid="B24" ref-type="bibr">24</xref>-<xref rid="B25" ref-type="bibr">25</xref>). This increase in capsule size is noticeable after several hours of infection (<xref rid="B17" ref-type="bibr">17</xref>). It is interesting that after several days of infection, the increase in capsule size is accompanied by an increase in the size of both the cell and the capsule, which results in giant forms of the yeast, whose role in virulence is unknown (<xref rid="B17" ref-type="bibr">17</xref>). Finally, the increase in capsule size is dependent on the organ of infection, with capsules being larger in the lung than in the brain (<xref rid="B16" ref-type="bibr">16</xref>). This reflects the importance of the environment on the size of the capsule. With regard to organic substances, both coagulase plasma (<xref rid="B26" ref-type="bibr">26</xref>) and mammalian sera (<xref rid="B18" ref-type="bibr">18</xref>) have been reported to be effective in inducing capsule growth.</p></sec><sec><title> <bold>Physiological meaning of the factors that induce capsule growth</bold> </title><p>We have described several new stimuli for capsule growth. However, the fact that this phenomenon occurs only in certain conditions raises the question of what factors are specifically responsible for the effect. In the case of CO<sub>2</sub> induction, capsule growth appears to resemble the <italic>in vivo</italic> situation that the yeast cells encounter in the lung, which is the first organ infected after inhalation of the infectious particles. Limited iron medium is another stimulus that can be used to induce capsule growth. The mechanism by which iron limitation induces capsule growth remains unknown. Iron regulates the production of some virulence factors in bacteria (<xref rid="B27" ref-type="bibr">27</xref>, <xref rid="B28" ref-type="bibr">28</xref>), and regulation of Fe concentration in physiological fluids by the host could affect the virulence of these organisms. It has been suggested that if iron uptake is mediated through the capsule, an increase in capsule size would be a response to transport iron more efficiently, but the same authors argued that this was unlikely since acapsular mutants grow normally during iron limitation conditions (<xref rid="B12" ref-type="bibr">12</xref>). Serum is known to contain iron-binding chelators and the effect in serum could reflect lack of iron, which in turn would affect growth. However, serum-induced capsule growth was not inhibited by addition of iron to the medium (<xref rid="B18" ref-type="bibr">18</xref>), indicating that the inducing signal is different. Since <italic>C. neoformans</italic> in a mammalian host would encounter serum components in the course of infection, the phenomenon of serum-induced capsular growth may reflect the stimulus that induces capsule size during pathogenesis. Concerning the other media described in this paper, such as diluted Sabouraud medium in buffers with basic pH, we believe that there is a decrease in the growth rate of the yeast, which could trigger capsule growth, possibly as a consequence of a stress response. This result, if validated by subsequent experimental work, would link capsule growth to cell growth. Consistent with this hypothesis is the fact that capsule growth is enhanced by an alkaline pH, a condition in which fungal cells grow more slowly. Finally, and related to the <italic>in vivo</italic> situation, it has been shown that in the lung, the size of the capsule is bigger than in the brain (<xref rid="B16" ref-type="bibr">16</xref>). Although the cause of this phenomenon is not known, the inflammatory response in much higher in the lung than in the brain (<xref rid="B1" ref-type="bibr">1</xref>), which may slow the growth of the yeast in the lung and is also consistent the view that capsule growth is correlated with slower replication. In addition, brain tissue may contain more iron (<xref rid="B16" ref-type="bibr">16</xref>), which could promote faster growth. However, it is important to stress that the concept that capsule growth is inversely proportional to growth rate is currently only a hypothesis.</p></sec></sec>
Phosphopeptide mapping of proteins ectopically expressed in tissue culture cell lines	<p>Post-translational modifications such as phosphorylation play a vital role in the regulation of protein function. In our study of the basic Helix-loop-Helix (bHLH) transcription factor HAND1, it was suspected that HAND1 was being phosphorylated during trophoblast giant cell differentiation and that coexpression of a constitutively active kinase with HAND1 resulted in changes in the proteins dimerization profile. In order to accurately document HAND1 phosphorylation and identify the resides being modified, we employed metabolic cell labeling with <sup>32</sup>P of tissue culture cells coexpressing a Flag-epitope tagged HAND1 along with a number of active kinases and phosphatase subunits. We generated phosphopeptide maps of the phosphorylated HAND1 using the methods described below and linked these modifications to changes in HAND1 biological function.</p>	<contrib contrib-type="author"><name><surname>Firulli</surname><given-names>Beth A.</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Virshup</surname><given-names>David M.</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Firulli</surname><given-names>Anthony B.</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>The basic Helix-loop-Helix (bHLH) transcription factors HAND1 and HAND2 are members of the twist family of bHLH proteins (for review (<xref rid="B1" ref-type="bibr">1</xref>). HAND factors are expressed in a variety of tissues during murine development including the heart, cardiac neural crest, lateral mesoderm, extraembryonic mesoderm, maternally derived decidua and sympathetic nervous system. bHLH factors activate/repress transcription by forming dimers via their HLH motif and bind DNA through the juxtaposition of the 2 basic domains which recognize a canonical <italic>cis</italic>-element CANNTG termed an E-box (for review (<xref rid="B2" ref-type="bibr">2</xref>)). bHLH proteins can generally be separated into 2 classes. The class A bHLH factors (E-proteins) which are ubiquitously expressed and the tissue-restricted or class B bHLH factors. The established paradigm of function was that a heterodimer composed of a class A and a class B protein was required for transcriptional regulation and in fact many class B factors do not form homodimers or heterodimers with other class B factors efficiently. Recently, we and others recognized that members of the twist family of proteins exhibited a more promiscuous dimerization profile forming homodimers and heterodimers with a wide range of class B factors (<xref rid="B3" ref-type="bibr">3</xref>-<xref rid="B6" ref-type="bibr">6</xref>). With this in mind, we set out a hypothesis that tissue-specific transcriptional regulation by HAND factors was in fact driven by the specific bHLH dimer complex that could form. This hypothesis begs the question of how is HAND dimerization controlled. Most recently, we determined that the answer to this question is in part addressed by post-translational modification (phosphorylation) of specific residues in the bHLH domain and that these modifications affect the dimerization affinities of HAND proteins that allow for changes in biological function (<xref rid="B3" ref-type="bibr">3</xref>). To determine if in fact HAND factors were phosphorylated and to define the specific location of the modified residues, we coexpressed a flag epitope tagged HAND1 (FlagHAND1) with constitutively active kinases and or phosphatases in tissue culture cells (HEK293 and RCHOI), which were subsequently metabolically labeled with <sup>32</sup>P-orthophosphate, HAND1 protein was immuno-precipitated using flag antibody, subjected to trypsin digestion, peptides were separated via 2-dimensional phosphopeptide mapping and phosphopeptides were visualized via exposure of the TLC plates to phosphoimager screens. By employing HAND1 deletion and point mutants, this technique allowed for the identification of the specific residues that are phosphorylated.</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><sec><title> <bold>Constructs</bold> </title><p>PKC-7 is a constitutively active form of PKCα, which was previously described (<xref rid="B7" ref-type="bibr">7</xref>). pFC-PKA is a constitutively active PKA (Strategene). pIRES FLAG-HAND1 and HAND1 point mutants are amino FLAG-tag (Sigma) fusions cloned into pIRES NEO (Clonetech). B56α and δ cDNAs were cloned into the expression plasmid pCEP4-l. HAND1 point mutants HAND1 S98A, S109A-T107A &D were generated using the Quick-Change Mutagenesis kit (Strategene) following the manufacturers protocols.</p></sec><sec><title> <bold>Tissue culture</bold> </title><p>HEK293 cells were grown in 10% FBS containing DMEM containing antibiotics at 37<sup>o</sup>C and 5% CO<sub>2</sub>. A total of 10 μg of the indicated plasmid construct was transfected into cells using a CaPO<sub>4</sub> based transfection (<xref rid="B8" ref-type="bibr">8</xref>). Briefly DNA was resuspended in 50 μl of H<sub>2</sub>O. To this 500 μl of 2X HBS was added. While bubbling air constantly with an autopipette, 450 μl of a 2:12 dilution of 2 M CaCl<sub>2</sub> was added drop-wise and allowed to sit for 20 minutes at room temperature. Precipitates were then added to cells and allowed to incubate for 4 hours at 37<sup>o</sup>C and 5% CO<sub>2</sub>. Media was removed and replaced with 3 ml of 15% glycerol in supplemented DMEM for 1 minute. Media was immediately aspirated, cells were washed in 5 mL of 1X PBS and 10 ml of supplemented DMEM was added and cells were grown for 48 hours.</p></sec><sec><title> <bold> <italic>In vivo</italic> </bold> <bold>labeling-Immunoprecipitations</bold> </title><p>HEK293 cells were grown and transfected as described above with the indicated constructs. 48 hours post transfection cells were incubated with 1 mCi of <sup>32</sup>P orthophosphoric acid (NEN)/ml of phosphate–free DMEM supplemented with dialyzed FBS for 4 hours. Cells were washed in 20 mM HEPES pH 7.4 and 150 mM NaCl followed by lysis in 20 mM NaPO<sub>4</sub>, 150 mM NaCl, 2 mM MgCl<sub>2</sub>, 0.1% NP40, 10% glycerol, 3 mg/μl leupeptin, 3 mg/μl pepstatin, 1 mM PMSF, 50 μM NaVO<sub>4</sub>, 5 mM NaF, 100 nM Okadaic acid, and 5 mM Beta glycerol phosphate. Equal amounts of protein were immunoprecipitated with agarose-conjugated FLAG M2-beads (Sigma) for 2 hours. Samples were washed 3 times in 1X PBS with a tube change on the last wash. Samples were boiled in loading dye, run through a 12% SDS PAGE, dried and exposed to a phosphoimager screen. </p></sec><sec><title> <bold>2-dimensional phosphopeptides mapping</bold> </title><p>HEK293 cells were grown, transfected, labeled with <sup>32</sup>P, and immunoprecipitated as described above. Dried IP gels were exposed to phosphoimager. Bands were identified and cut by aligning the image with radioactive marker spots. The acrylamide bands corresponding to labeled FLAG-HAND1 were cut away from the dried gel, the filter paper was scraped away from the acrylamide, and rehydrated in 400 μl freshly made 50 mM ammonium bicarbonate. The rehydrated acrylamide was crushed into small pieces and allowed to digest overnight with 30 μg of TPCK-treated trypsin (Worthington) at 37ºC. Digests were spiked the following day. Digested peptides were removed from the crushed acrylamide, washed twice with 50 mM ammonium bicarbonate, and the peptides were concentrated in a Speed Vac (Forma). Digested peptides were washed 4 times with 1 mL ddH<sub>2</sub>O and then twice in pH 1.9 buffer (2.8% formic acid, 7.8% glacial acetic acid). Samples were then resuspended in 4 μl pH 1.9 buffer, spotted onto cellulose TLC plates and run in the 1<sup>st</sup> dimension in pH 1.9 buffer on a Hunter Thin Layer Electrophoresis apparatus (HTLE 7000, CBS Scientific, Inc.) for 35 minutes at 1300 V. Plates were dried, rotated 90 degrees, and then run in the 2<sup>nd</sup> dimension using an isobutyric acid buffer (62.5% isobutyric acid, 1.9% n-Butanol, 4.8% pyridine, 2.9% glacial acetic acid) in a TLC tank. When liquid phase migration was 1 cm from the top of the plates, they were removed, dried, and exposed to a phosphoimager screen for visualization and analysis.</p></sec></sec><sec><title>Results and Discussion</title><p>We recently demonstrated that HAND1 is phosphorylated during the differentiation of RCHOI trophoblast stem cells and that protein kinase A (PKA), protein kinase C (PKC) and the delta isoform of the B56-regulatory subunit (B56δ) of the protein phosphatase 2A (PP2A) regulate HAND1 phosphorylation status on residues T107 and S109 (<xref rid="B3" ref-type="bibr">3</xref>). Key in this study was the ability to determine the exact residues that were being modified in HAND1 so that functional analysis of HAND1 mutants in these residues could be deduced. We employed phosphopeptide analysis to address these questions as it provided a greater sensitivity to changes in HAND1 phosphorylation state than simple immunoprecipitations (IPs) of HAND1 expressed in metabolically labeled cells. We chose to use phosphopeptide mapping over mass spectrophometry as we could perform these experiments ourselves without needing to rely on instrumentation analysis by others. Results show that expression of constitutively active PKC along with HAND1 results in increased phosphorylation of 5 peptides (Fig. <xref rid="F1" ref-type="fig">1</xref>; 3). Moreover, coexpression of B56α results in no significant change to HAND1 phosphorylation while coexpression of B56δ reduces the phosphorylation on spots 8 and 9 correlating well with B56δ interaction analysis with HAND1 (Fig. <xref rid="F1" ref-type="fig">1</xref>; 3).</p><fig id="F1"><label>Fig. 1</label><caption><title>Identification of HAND1 residues phosphorylated by PKC using phosphopeptide analysis in HEK293 cells.</title><p> Panels (A) and (B) show the variation of HAND1 phosphorylation when expressed with or without constitutively active PKC. The increased signal intensity of 5-phosphopepties (5-9) indicates increased phosphorylation. Panels (C) and (D) show that coexpression of the non-interacting B56α (C) has no effect on HAND1 phosphorylation by PKC whereas expression of B56δ (D) reduces HAND1 phosphorylation of peptides 8 and 9 (marked *). Panel (E) shows that point mutagenesis of both T107 and S109 to alanine eliminates the phosphorylation of peptides 8 and 9 (marked X) confirming these sites as PKC targets and targets for dephosphorylation by B56δ-containing PP2A.</p></caption><graphic xlink:href="bpo_v6_p16_m69f1lg"/></fig><p>We next employed serine and threonine mutations based on our understanding of the trypsin restriction map of HAND1. Our designed HAND1 mutants correspond to S→A changes in consensus kinase sites for PKA and PKC. Additionally when designing the mutants, we considered if the mutations were contained within a single trypsin fragment and if other S and or T residues were also present in these peptides (<xref rid="B3" ref-type="bibr">3</xref>). In the case of T107 and S109, they were contained in a single fragment, so we decided to mutate these in combination. Phosphopeptide maps of HAND1 and HAND1T107;S109A coexpressed with or without constitutively active PKC show that 2-specific phosphopeptides are absent from the map of HAND1T107;S109A when compared to wildtype HAND1 (Fig. <xref rid="F1" ref-type="fig">1</xref>). These results correspond with similar previously published experiments in which we coexpressed constitutively active PKA (<xref rid="B3" ref-type="bibr">3</xref>). The observation that a double mutation within a single trypsin fragment resulted in the reduction of two phosphopeptides on the map can be interpreted in two ways. One is that the two spots represent mono- and diphosphorylated forms of the protein. The second is that the two peptides are in fact an artifact of incomplete trypsin digestion. Trypsin cuts at basic residues and the HAND1 basic domain contains a high concentration of basic residues. As the HAND1 basic domain is located just amino to T107 and S109 partial trypsin digestion could explain the loss of 2 phosphopeptides in the HAND1T107;S109A mutant shown in Figure <xref rid="F1" ref-type="fig">1</xref>. This phenomenon was clearly encountered in the mutation of S98 (the only S within its trypsin peptide) where 3-5 partial fragments were reduced in the maps of this HAND1 mutant (<xref rid="B3" ref-type="bibr">3</xref>). To try to address this issue directly, we made the single HAND1 mutants T107A and S109A, coexpressed these mutants in cells with constitutive kinase and compared maps to wild type HAND1 (Fig. <xref rid="F2" ref-type="fig">2</xref>). Results of these experiments show that mutation of T107 to alanine does not significantly alter phosphopeptide pattern of HAND1 when coexpressed with PKA; however, mutation of S109 eliminates phosphorylation of both peptides (Fig. <xref rid="F2" ref-type="fig">2</xref>). These results suggest that indeed like S98 partial digestion may come into play, but in this case it is also possible that phosphorylation of T107 is required for phosphorylation of S109. To address this question would require mass spectrophometery.</p><fig id="F2"><label>Fig. 2</label><caption><title>Single mutations of T107 and S109 suggest that S109 is the main target of PKA and PKC.</title><p> HAND1, HAND1T107A or HAND1S109A were coexpressed with constitutively active PKA in HEK293 cells and subjected to phosphopeptide analysis. Mutation of T107 shows no significant difference in the HAND1 phosphopeptide map specifically peptides 8 and 9 (see Fig. <xref rid="F1" ref-type="fig">1</xref>). In contrast, mutation of S109 to alanine eliminates phosphopeptides 8 and 9 and recapitulates the data observed in the double mutant. This suggests that peptides 8 and 9 are partial tryptic digests and not mono- and diphosphorylated forms of the protein.</p></caption><graphic xlink:href="bpo_v6_p16_m69f2lg"/></fig><p>Taken together, the results obtained show that the upregulation of HAND1 phosphorylation that is observed during RCHOI differentiation (<xref rid="B3" ref-type="bibr">3</xref>) can be attributed to the modifications of three residues within HAND1. It should be considered that it is possible that other residues within HAND1 may in fact be phosphorylated, but our conditions of mapping are either insufficient to resolve 2 labeled peptides or the peptides run off of the TLC plate in the solvent phase. Employing different solvents for the chromatography used in the 2nd dimension can test this latter possibility. A useful and expedient way to test different solvent conditions is to use <italic>in silico</italic> mapping simulation programs such as <italic>Mobility</italic>, which can be found on the <italic>Gene</italic> <italic>Stream</italic> web site (http://www.genestream.org/). Based on your trypsin digestion pattern the program will show you which peptides will likely be present on the peptide map using a particular solvent.</p></sec>
Negative Staining and Image Classification – Powerful Tools in Modern Electron Microscopy	<p>Vitrification is the state-of-the-art specimen preparation technique for molecular electron microscopy (EM) and therefore negative staining may appear to be an outdated approach. In this paper we illustrate the specific advantages of negative staining, ensuring that this technique will remain an important tool for the study of biological macromolecules. Due to the higher image contrast, much smaller molecules can be visualized by negative staining. Also, while molecules prepared by vitrification usually adopt random orientations in the amorphous ice layer, negative staining tends to induce preferred orientations of the molecules on the carbon support film. Combining negative staining with image classification techniques makes it possible to work with very heterogeneous molecule populations, which are difficult or even impossible to analyze using vitrified specimens.</p>	<contrib contrib-type="author"><name><surname>Ohi</surname><given-names>Melanie</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Li</surname><given-names>Ying</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><name><surname>Cheng</surname><given-names>Yifan</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Walz</surname><given-names>Thomas</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>EM has proved to be an exceptionally versatile tool to study the structure of proteins and macromolecular complexes. Biological molecules are problematic specimens for EM, because of their susceptibility to radiation damage, their poor capacity to scatter electrons, and their proneness to dehydration in the vacuum of the electron microscope. A main requirement for EM specimen preparation is to prevent the structural collapse upon sample dehydration, and ideally also to increase specimen contrast.</p><p>Negative staining, the embedding of a specimen in a layer of dried heavy metal solution, was introduced early on as a quick and easy specimen preparation technique that significantly increases the specimen contrast. Images of negatively stained molecules are amenable to image averaging techniques that increase the signal-to-noise ratio (SNR) and thus allow finer details of the molecule to be visualized. Using computational tools to combine different views, images of negatively stained specimens can also be used to produce threedimensional (3D) reconstructions of the molecule under investigation (e.g. (<xref rid="B1" ref-type="bibr">1</xref>)). Embedding of the sample in a layer of dried staining solution provides some protection against the collapse of the specimen due to dehydration, but 3D reconstructions from specimens prepared in this manner are usually still significantly flattened. Moreover, microcrystals formed by the heavy metals upon drying of the negative stain solution limit the resolution of the 3D map that can be achieved to about 20 Å.</p><p>In 1984, Dubochet and colleagues revolutionized single particle EM by introducing specimen vitrification, in which the sample is applied to a grid covered with holey carbon film and quickly frozen by plunging the grid into liquid ethane (<xref rid="B2" ref-type="bibr">2</xref>). The rapid freezing prevents the water from forming ice crystals and embeds the molecules in a layer of vitrified (or amorphous) ice. Vitrification preserves the specimen in a near-native environment, eliminating not only specimen distortions due to dehydration and adsorption but also the limitation in achievable resolution associated with negative staining. A drawback of vitrification is the poor SNR in images of vitrified specimens, which poses severe difficulties especially for the study of small molecules. Nevertheless, under favorable circumstances, high-resolution structures can be obtained even with rather small molecules in vitrified ice (e.g. (<xref rid="B3" ref-type="bibr">3</xref>-<xref rid="B5" ref-type="bibr">5</xref>)).</p><p>Molecules usually adopt more or less random orientations in an amorphous ice layer providing many different views of the molecule. These views can be exploited to calculate a 3D reconstruction with the angular reconstitution approach (<xref rid="B6" ref-type="bibr">6</xref>), making it unnecessary to record images of tilted specimens. While random orientations of the molecules are advantageous for homogeneous samples, they create a severe problem for heterogeneous samples. To obtain a reliable 3D reconstruction of a molecule it is essential that only images of identical molecules are combined. However, from projection views alone, it is usually not possible to distinguish between molecules in different orientations and molecules in different conformations.</p><p>Vitrification has proved to be such a powerful technique to determine the structure of macromolecules that many scientists have begun to consider negative staining to be old-fashioned and not worth pursuing. If applicable, vitrification is undoubtedly the best technique to obtain an undistorted 3D map of a molecule by EM. However, negative staining has unique advantages and can provide important information on biological molecules that is not easily obtained or, in the case of very heterogeneous samples, not even possible to obtain with vitrified specimens. In this paper we describe different negative staining protocols. We also demonstrate the unique information that can be gathered by applying classification methods to images of heterogeneous samples prepared by the conventional negative staining protocol. To illustrate these points, we will use mainly recent examples from our own work.</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><sec><title> <bold>Proteins</bold> </title><p>Recombinant human transferrin receptor (TfR) and transferrin (Tf) were prepared as described in (<xref rid="B5" ref-type="bibr">5</xref>). <italic>S. cerevisiae </italic>Sec23p/Sec24p was prepared as described in (<xref rid="B7" ref-type="bibr">7</xref>). Recombinant bacteriophage T7 primase-helicase was prepared as described in (<xref rid="B8" ref-type="bibr">8</xref>). Recombinant integrin aVb3 was prepared as described in (<xref rid="B9" ref-type="bibr">9</xref>). Recombinant integrin a5b1 headpiece and fibronectin fragments 7-10 (Fn7-10) and 9-10 (Fn9-10) were prepared as described in (<xref rid="B10" ref-type="bibr">10</xref>). <italic>S. cerevisiae </italic>proteasome was prepared as described in (<xref rid="B11" ref-type="bibr">11</xref>). Rabbit spleen 20S and 26S immunoproteasome and recombinant mouse PA26a were prepared as described in (<xref rid="B12" ref-type="bibr">12</xref>). <italic>Rhodobacter sphaeroides </italic>reaction center (RC) – light-harvesting complex 1 (LH1) photounits were prepared as described in (<xref rid="B13" ref-type="bibr">13</xref>).</p></sec><sec><title> <bold>Electron microscopy</bold> </title><p>Unless stated otherwise, specimens were prepared for EM using the conventional negative staining procedure. Briefly, a 2.5 ml drop of sample solution was adsorbed to a glow-discharged carbon-coated copper grid, washed with two drops of deionized water, and stained with two drops of freshly prepared 0.75% uranyl formate.</p><p>Unless stated otherwise, samples were imaged at room temperature using a Philips Tecnai T12 electron microscope equipped with an LaB6 filament and operated at an acceleration voltage of 120 kV. Images were taken at a magnification of 52,000x and a defocus value of 1.5 mm on Kodak SO-163 film using low-dose procedures. Films were developed for 12 minutes with fullstrength Kodak D-19 developer at 20°C. All micrographs were visually inspected with a laser diffractometer, and only drift-free images were selected for digitization with a Zeiss SCAI scanner using a step size of 7 mm. Micrographs were binned over 3 ¥ 3 pixels to yield a pixel size of 4.04 Å on the specimen level.</p></sec><sec><title> <bold>Image processing</bold> </title><p>Particles were selected interactively from images using the display program WEB associated with SPIDER (14), the program used for all subsequent image processing steps. Selected particles were windowed into individual images with a size depending on the molecule under investigation. The side length of the images was typically chosen to be approximately twice the length of the longest particle dimension. Particle images were first subjected to 10 rounds of alignment and classification, specifying a number of output classes depending on the heterogeneity of the particular sample. Unique averages were selected from the resulting class averages and used as references for 8 cycles of multi-reference alignment.</p><p>To compare projection averages of the <italic>S. cerevisiae </italic>proteasome with the corresponding crystal structure, the crystal structure was converted into a density map and resolution-filtered to 25 Å. Projections from the density map were calculated at angular intervals of 2º and cross-correlated with the respective projection averages. The projections with the highest cross-correlation coefficients are shown in Figure <xref rid="F4" ref-type="fig">4</xref>a.</p></sec></sec><sec><title>Results and Discussion</title><sec><title>Negative staining protocols</title><p>The conventional negative staining protocol involves the adsorption of the specimen to a glow-discharged carbon-coated EM grid, which is washed with two drops of deionized water and subsequently stained with two drops of heavy metal solution. To obtain thinner stain embedding, excess stain solution can be removed from the grid by vacuum aspiration. This basic protocol can easily be adapted if required. Buffer solution can be used instead of water to wash the grid if this is necessary for sample stability, although this generally results in a somewhat higher background. If a membrane protein is to be visualized, at least five drops of water should be used to remove the detergent from the grid, since detergents can interfere with staining.</p><p>The conventional negative staining protocol normally induces specimens to adsorb to the carbon support film in one or a limited number of preferred orientations. For example, a complex between a construct of the transferrin receptor containing only the extracellular domains but lacking the transmembrane and cytoplasmic domains (TfR) with transferrin (Tf) adsorbs to the carbon film predominantly in two orientations (Fig. <xref rid="F1" ref-type="fig">1</xref>b, insets 1 and 2). It is characteristic for this method to create stain clouds that surround the molecules, producing a strong contrast between the background and the particle (Fig. <xref rid="F1" ref-type="fig">1</xref>a). When images are tilted to provide the different views of the specimen needed to calculate a 3D map by the random conical tilt approach (<xref rid="B1" ref-type="bibr">1</xref>), these stain clouds are particularly evident (Fig. <xref rid="F1" ref-type="fig">1</xref>b).</p><fig id="F1"><label>Fig. 1</label><caption><title>Images of the TfR-Tf complex obtained with different negative staining protocols.</title><p> a and b: Images of an untilted (a) and a 60° tilted sample (b) prepared by the conventional negative staining protocol using uranyl formate. The particles are surrounded by a dark stain cloud, which is particularly evident in the image of the tilted specimen. Class averages (insets) show the two predominant orientations, in which the complex adsorbs to the carbon film, revealing a side view (inset 1) and a top view (inset 2) of the complex. c and d: Images of an untilted (c) and a 60° tilted sample (d) prepared by the carbon sandwich technique using uranyl formate. The negative stain forms a continuous layer and no stain cloud is apparent in images of untilted or tilted specimens. Class averages show that the complexes are seen in the same orientations as in the conventional negative staining protocol (insets 1 and 2). Due to the additional carbon layer, a significant number of complexes are being squashed upon drying and therefore can not be used for structure determination (insets 3 and 4). e: Image obtained with an untilted sample embedded in a mixture of glucose and ammonium molybdate, showing the image contrast to be much weaker than in the case of uranyl formate staining. Classification of particles selected from such images reveals that the molecules adsorb to the grid in random orientations. Some of the resulting class averages are shown in insets 1 to 10. The lines in panels a to d indicate the tilt axis. The scale bar corresponds to 50 nm and the inset panels have a side length of 26 nm.</p></caption><graphic xlink:href="bpo_v6_p23_m70f1lg"/></fig><p>The conventional negative staining protocol is quick and easy, but 3D maps calculated from specimens prepared in this way show severe deformations due to flattening and incomplete stain embedding. To overcome incomplete stain embedding, Frank and colleagues have developed a carbon sandwich technique, in which the specimen is embedded in a continuous layer of stain in between two carbon films. While specimens prepared using this technique usually still adsorb in preferred orientations (Fig. <xref rid="F1" ref-type="fig">1</xref>d, insets 1 and 2), images do not have stain clouds surrounding the molecules (Fig. <xref rid="F1" ref-type="fig">1</xref>c and d). While this preparation technique prevents artifacts due to incomplete stain embedding, a fraction of the molecules can become significantly squashed, leading to a spread-out appearance of the molecules (Fig. <xref rid="F1" ref-type="fig">1</xref>d, insets 3 and 4). Such particles should be excluded from structure determination. To reduce dehydration-induced flattening of the specimen or the squashing of the molecules in the carbon sandwich technique, glycerol or glucose can be added to the specimen or staining solution. If sugar or glycerol is added directly to the specimen solution, this often causes the specimen to adsorb to the carbon film in random orientations (Fig. <xref rid="F1" ref-type="fig">1</xref>e). Glycerol and sugars are also very sensitive to radiation damage. Therefore, specimens prepared with such additives need to be imaged at liquid nitrogen temperature.</p><p>In all the negative stain protocols described above the specimen is dried. The best specimen preservation is however achieved with cryo-negative staining techniques, which avoid drying of the specimen. In a protocol pioneered by Adrian and co-workers, the specimen is vitrified in a saturated ammonium molybdate solution (<xref rid="B15" ref-type="bibr">15</xref>). This technique, however, exposes the specimen to high ionic strength, which can cause the dissociation of many macromolecular complexes. In an alternative, gentler approach, developed in the Stark laboratory, the sample is mixed with glycerol, stained in a carbon layer sandwich and then frozen (<xref rid="B16" ref-type="bibr">16</xref>). Both techniques provide high contrast due to the heavy metal stain while avoiding dehydration of the specimen. Molecules prepared in either way usually adopt random orientations, so that the angular reconstitution approach must be used for 3D reconstruction (<xref rid="B6" ref-type="bibr">6</xref>).</p><p>The carbon sandwich technique, the addition of glycerol or glucose, and the cryo-negative staining approaches all improve the quality of 3D structures by reducing artifacts due to dehydration, adsorption and incomplete stain embedding. Conventional negative staining is however perfectly adequate when visualizing the 3D structure of small molecules (< 250 kDa), where the above problems are less severe, and when only projection structures are being determined. The thin layer of stain produced by conventional staining is actually an advantage when visualizing very small molecules. Moreover, the adsorption of the molecules to the carbon film in one or only a limited number of preferred orientations, which is usually observed with samples prepared by this technique, is beneficial in the analysis of heterogeneous samples. The use of projection images of negatively stained samples to analyze heterogeneous samples is the focus of this paper.</p></sec><sec><title>Choice of negative stains</title><p>A variety of heavy metal compounds are available for conventional negative staining. Among the most commonly used stains are uranyl and tungstate stains, ammonium molybdate and aurothioglucose (for a more complete list of stains, see (<xref rid="B17" ref-type="bibr">17</xref>)). It is important to note that stains are usually not inert, but have different characteristics that can lead to different staining of the specimen (e.g., (<xref rid="B18" ref-type="bibr">18</xref>-<xref rid="B20" ref-type="bibr">20</xref>)).</p><p>Tungstate stains and ammonium molybdate are negatively charged metal ions and have the advantage that the pH of the stain solutions can be neutralized. Aurothioglucose carries no charge and preserves the specimen particularly well due to its sugar component. However, aurothioglucose produces only poor image contrast and is very sensitive to radiation damage, requiring images to be taken at liquid nitrogen temperature. The positively charged uranyl stains often generate the highest contrast and have a fixative effect, but they require the stain solution to be acidic (pH ~ 4.5) in order to prevent precipitation of the stain. This is problematic when studying proteins that undergo pH-dependent conformational changes, such as viral fusion proteins. Recent studies have demonstrated, however, that a uranyl acetate solution fixes protein structure on the millisecond timescale (<xref rid="B21" ref-type="bibr">21</xref>), alleviating this problem.</p><fig id="F2"><label>Fig. 2</label><caption><title>Visualizing small molecules (< 100 kDa) prepared by the conventional negative staining protocol using uranyl formate.</title><p> a: Image of a mixture of integrin α5β1 headpieces and a fibronectin (Fn) fragment containing Fn domains 7 to 10 (Fn7-10, MW ~40 kDa). The image not only visualizes the Fn7-10 fragment bound to the α5β1 headpiece (asterisks), but also unbound Fn7-10 fragment (arrows). Class averages of the unbound Fn7-10 fragment obtained from such images resolve the four individual 10-kDa domains in the flexible Fn7-10 fragment (insets 1 and 2). b: Individual molecules can clearly be seen in images of negatively stained Tf molecules (MW~ 70 kDa). Class averages show a top view (inset 1) and a side view of the molecule (inset 2). The top view resolves the two lobes of Tf (MW ~35 kDa) as well as the two domains of each lobe (MW ~17 kDa). The scale bar corresponds to 50 nm and the inset panels have a side length of 26 nm. Insets in Figure <xref rid="F4" ref-type="fig">4</xref>a modified and reprinted from (<xref rid="B10" ref-type="bibr">10</xref>). Copyright 2003 with permission from EMBO Journal.</p></caption><graphic xlink:href="bpo_v6_p23_m70f2lg"/></fig><p>For most applications uranyl stains are the best choice. A solution of the more commonly used uranyl acetate is stable over many months. A uranyl formate solution is only stable over a few days, but it yields better staining of the specimen due to a finer grain size. This can be important when visualizing very small molecules (< 100 kDa). In our hands, visualizing the 40 kDa fibronectin Fn7-10 fragment (Fig. <xref rid="F2" ref-type="fig">2</xref>a) and the 75 kDa Tf molecule (Fig. <xref rid="F2" ref-type="fig">2</xref>b) was only possible by conventional negative staining with uranyl formate. We could not see these molecules using any other negative stain or negative staining procedure.</p></sec><sec><title>The need for image classification</title><p>Only in the most favorable cases is the sample to be studied truly homogeneous. Even if a protein appears pure and shows only a single band on an SDS PAGE, gel, the particles are likely to appear heterogeneous when viewed by negative stain EM. A minor degree of heterogeneity can be introduced by the negative staining procedure itself due to distortions upon adsorption and/or a variable degree of stain embedding. These are usually only significant problems with larger molecules. More significant heterogeneity in the molecule population may however occur because of (<italic>i</italic>) particles adsorbing to the grid in different orientations (resulting in identical molecules having a different appearance), (<italic>ii</italic>) particles assembling into different oligomeric states, and (<italic>iii</italic>) particles adopting different conformational states. While heterogeneity is usually already present in pure protein preparations, it is even more pronounced when macromolecular complexes are studied, which in many cases can fall apart. Image classification is the computational method that allows one to obtain meaningful structural information from heterogeneous specimens.</p><p>As a general rule, biochemical data such as SDS PAGE and gel filtration chromatography do not suffice to rule out sample heterogeneity. Due to the more or less random orientation of the molecules and the poor SNR of the images, it is virtually impossible to assess sample heterogeneity by cryo-EM of vitrified samples. Structure determination using vitrified samples therefore always carries the risk of producing a distorted image of the molecule under investigation due to averaging images of non-identical molecules. Negative staining yields a much better SNR in the images and usually induces molecules to adsorb to the support film in only one or a few preferred orientations. This makes it possible to assess sample heterogeneity by image classification. When working with a structurally uncharacterized molecule, it is therefore good practice to first perform a negative stain analysis prior to any attempt of working with vitrified specimens.</p><p>We illustrate this point with the example of the complex formed by the headpiece of integrin α<sub>5</sub>β<sub>1</sub> with the fibronectin Fn9-10 fragment (<xref rid="B10" ref-type="bibr">10</xref>). Gel filtration of the complex showed only two peaks corresponding to the complex and excess Fn9-10 used for complex formation, suggesting no unliganded integrin α<sub>5</sub>β<sub>1 </sub>headpieces to be present (Fig. <xref rid="F3" ref-type="fig">3</xref>a). Images of the peak fraction containing the complex prepared by negative staining immediately revealed, however, heterogeneity in the sample (Fig. <xref rid="F3" ref-type="fig">3</xref>b) due to dissociation of the ligand from the integrin headpiece, which is accompanied by a conformational change in the headpiece (Fig. <xref rid="F3" ref-type="fig">3</xref>c and <xref rid="F3" ref-type="fig">3</xref>d). This heterogeneity would not have been seen in vitrified preparations and without doubt cryo-EM of vitrified samples would have produced a flawed 3D reconstruction. By contrast, using negative stain EM in combination with image classification techniques it was straightforward to separate images of the two different molecules. Moreover, it was possible to obtain meaningful 3D reconstructions from this heterogeneous sample for both the unliganded integrin headpiece (Fig. <xref rid="F3" ref-type="fig">3</xref>e) and its complex with the Fn9-10 fragment (Fig. <xref rid="F3" ref-type="fig">3</xref>f).</p><fig id="F3"><label>Fig. 3</label><caption><title>Negative stain electron microscopy of the integrin α5β1 headpiece with and without a bound fibronectin (Fn) fragment containing Fn domains 7 to 10 (Fn9-10).</title><p> a: Elution profile from a gel filtration column used to purify the complex of α5β1 headpiece with an Fn9-10 fragment. The elution profile shows two peaks that correspond to the α5β1-Fn9-10 complex (~200 kDa) and unbound Fn9-10 fragment (~30 kDa). b: Negative stain electron microscopy reveals that the α5β1 headpiece adopts two conformations, namely a closed (black circles) and an open conformation (white circles). c and d: Class averages representing the closed (c) and the open conformation (d). Binding of Fn9-10 fragment (arrow in d) induces the open conformation of the headpiece, while the unliganded is in the closed conformation (c). e and f: 3D reconstructions of an unliganded (e) and an Fn9-10-liganded α5β1 headpiece (f) with the fit atomic structures of the αV and β3 subunit (<xref rid="B33" ref-type="bibr">33</xref>) in red and blue, respectively, and of the Fn9-10 fragment (<xref rid="B34" ref-type="bibr">34</xref>) in white. The scale bar corresponds to 50 nm and panels c to f have a side length of 22 nm. Figure panels modified and reprinted from (<xref rid="B10" ref-type="bibr">10</xref>). Copyright 2003 with permission from EMBO Journal.</p></caption><graphic xlink:href="bpo_v6_p23_m70f3lg"/></fig></sec><sec><title>Principle of image classification</title><p>A mathematical description of classification algorithms is beyond the scope of this paper and the interested reader is referred to (<xref rid="B22" ref-type="bibr">22</xref>) as an excellent introduction to the subject. Briefly, classification algorithms provide mathematical tools to quantify the similarity among different images. Variants of the K-means and hierarchical classification method are the most commonly used algorithms in single particle EM, but other algorithms, e.g. the Baysean method (<xref rid="B23" ref-type="bibr">23</xref>), are still being explored for the application to sets of images. All classification methods are in essence based on a comparison of the intensities of all pixels in one image with those of all the corresponding pixels in another image. For the best result of classification it is therefore of crucial importance that all particle images be aligned to each other as precisely as possible.</p><p>The details of how a set of images is aligned and classified strongly depend on both the software package and the classification algorithm used, and we therefore only provide a very general outline. The first step has to be a translational and rotational alignment of the particle images, for which reference-free or multi-reference alignment procedures can be used (<xref rid="B22" ref-type="bibr">22</xref>). So not to interfere with the alignment and classification algorithms, each image ideally contains only one particle. The alignment procedure is done iteratively over many cycles and is considered completed when overall image shifts and rotations no longer decrease upon further alignment cycles.</p><p>The aligned images are then subjected to classification. After specifying into how many classes the images should be sorted, the images in each class are averaged to create class averages with improved SNR. The number of required classes depends on the heterogeneity in the sample and cannot be predicted. It is therefore good practice to run multiple classifications specifying different numbers of output classes, e.g. 20, 50, and 100 classes. One good indication that sufficient output classes were selected is that a number of projection structures are represented by more than one class average. The more heterogeneous the population, the more classes are required, and accordingly the larger the image set has to be to obtain a sufficient number of images in each class to calculate meaningful averages.</p><p>To assess whether the classification procedure was successful, the images in each class can be visually compared to the corresponding class average. This is usually easy when working with images of negatively stained samples, where the particles are clearly seen due to the high image contrast. The same assessment is substantially harder for images of vitrified specimens, where the poor SNR of the images can make it difficult to even see the particle. At the same time the poor SNR of images taken from vitrified specimens reduces the probability that the images are being assigned to the correct class. This is of particular concern with heterogeneous samples and emphasizes the notion that a structurally uncharacterized protein should always first be analyzed in negative stain prior to analysis of vitrified specimens.</p><p>If desired, the unique classes can be selected from the best classification result and used as references for multi-reference alignment. Here, all the images of a data set are cross-correlated with all the references and assigned to the reference that yielded the highest correlation coefficient. All the images assigned to the various references are again averaged. Multi-reference alignment is also performed iteratively with the averages of each cycle being used as the references for the next cycle. The procedure is considered completed when the averages show no significant changes upon further cycles. No further improvement in the resolution of the averages upon further cycles, as assessed by the Fourier ring correlation or the spectral SNR criterion, is also an indication that the multi-reference alignment procedure is completed.</p><p>Various software packages are available that contain modules to perform image alignment, classification, and multi-reference alignment. IMAGIC (<xref rid="B24" ref-type="bibr">24</xref>) is commercially available, whereas SPIDER (<xref rid="B14" ref-type="bibr">14</xref>) and EMAN (<xref rid="B25" ref-type="bibr">25</xref>) are academic packages. Of these three programs, we consider SPIDER the most versatile one, allowing various classification algorithms to be applied to a data set.</p></sec><sec><title>Examples for the use of image classification</title><p>At this point we emphasize again that if the goal is to determine the 3D structure of a protein or a macromolecular complex, cryo-EM of vitrified specimens is the method of choice. If the sample is however too heterogeneous to use vitrified specimens, much care should be taken to find the negative stain and the preparation protocol that preserve the specimen with as little preparation artifacts as possible in order to obtain a reliable 3D map. The appropriate combination depends on the molecule that is being studied. This paper, however, does not focus on the determination of 3D structures. It is our intention to show that many biological questions can be addressed by simply taking projection images of negatively stained specimens. For this purpose it is in most cases sufficient to use the conventional negative stain approach with uranyl formate as the stain. All specimens used as examples for classification in the next paragraphs were prepared in this way.</p><p> <italic>Apparent heterogeneity due to different orientations</italic>. A fundamental problem in EM is the difficultly in distinguishing between molecules viewed from different orientations and molecules in different conformational states from projection views alone. There are two ways to determine whether differences in projection averages arise from different views of the molecule or from conformational variability.</p><fig id="F4"><label>Fig. 4</label><caption><title>Apparent sample heterogeneity due to different particle orientations.</title><p> a: Image of a negatively stained yeast proteasomes. Classification of the particle images yielded two class averages (insets 1 and 2). Comparison of the two class averages with projections from a resolution-limited model generated from the crystal structure (<xref rid="B26" ref-type="bibr">26</xref>) identified the two averages to correspond to a top (inset 3) and a side view (inset 4) of the proteasome. b: Image of negatively stained yeast Sec23p/Sec24p complexes. Classification of the particle images yielded a variety of slightly different class averages. 3D reconstructions of the classes shown in insets 1 to 5, calculated using images of tilted specimens, looked identical, demonstrating that the variations in the class averages are due to slightly different orientations, in which the complexes adsorbed to the carbon film. The scale bars correspond to 50 nm and the inset panels have a side length of 34 nm in a and 32 nm in b. Figure 4b modified and reprinted from (<xref rid="B7" ref-type="bibr">7</xref>). Copyright 2001 with permission from the National Academy of Sciences, USA.</p></caption><graphic xlink:href="bpo_v6_p23_m70f4lg"/></fig><p>If the crystal structure of the molecule is available, it can be used to create a resolution-limited model, typically using a resolution cut-off of about 25 Å. Projection views can then be generated from this model at regular angular intervals, e.g. 2°. If comparison by cross-correlation identifies a highly similar projection view of the molecule for all the experimental class averages, the structural heterogeneity is most likely due to different orientations of the molecules on the carbon film. This situation is illustrated by a negative stain preparation of yeast proteasomes (Fig. <xref rid="F4" ref-type="fig">4</xref>a), where classification produced two projection averages (Fig. <xref rid="F4" ref-type="fig">4</xref>a, insets 1 and 2). Comparison with projections from a 25 Å resolution-limited model generated from the crystal structure (<xref rid="B26" ref-type="bibr">26</xref>) identified the two projection averages to correspond to a top (Fig. <xref rid="F4" ref-type="fig">4</xref>a, inset 3) and a side view (Fig. <xref rid="F4" ref-type="fig">4</xref>a, inset 4) of the proteasome.</p><p>The situation is more difficult if no structural information is available on the molecule under investigation. In this case 60°/0° image pairs have to be collected to calculate 3D maps by the random conical tilt approach. The particle images from the untilted specimen are subjected to classification and individual 3D reconstructions are calculated for all or at least some of the classes using the corresponding particle images from the tilted specimen. If the 3D maps look the same, the classes represent different views rather than different conformations. This is illustrated by the Sec23p/24p complex. Visualized by negative stain EM, this complex has a bone-like appearance (Fig. <xref rid="F4" ref-type="fig">4</xref>b), but classification of the particle images produced averages with notable variations in the shape of the molecules (Fig. <xref rid="F4" ref-type="fig">4</xref>b, insets 1 to 5). Since there was no crystal structure available at the time of our EM analysis, individual 3D reconstructions were calculated for several classes, which looked essentially the same. Therefore, all images were combined to calculate a single 3D reconstruction (<xref rid="B7" ref-type="bibr">7</xref>) and the accuracy of our structure was later confirmed by a crystal structure of the Sec23p/24p complex (<xref rid="B27" ref-type="bibr">27</xref>). Because the sample was homogenous, this complex would also have been amenable to structure determination by cryo-EM of vitrified specimens.</p><p>If 3D maps obtained from different classes are different, experience is required to make the decision whether the differences arise from true conformational changes in the molecule or from preparation artifacts such as specimen flattening or incomplete stain embedding. A certain degree of variation in the 3D structures is indeed expected, because the staining pattern and the deformations due to adsorption interactions and sample dehydration all depend on the specific orientation of the particle on the carbon film. As discussed above, addition of glucose or glycerol to the sample and the use of the carbon sandwich technique minimize such variations induced by the negative stain preparation.</p><p> <italic>Heterogeneity due to different oligomeric states</italic>. Sample heterogeneity arises when a protein can exist in different oligomeric states as exemplified by the bifunctional primase-helicase of bacteriophage T7, the crystal structure of which has recently been solved (<xref rid="B8" ref-type="bibr">8</xref>). This protein oligomerizes into a ring-shaped structure and appears rather homogeneous when visualized by negative stain EM (Fig. <xref rid="F5" ref-type="fig">5</xref>). Upon image classification it becomes evident, however, that mixed populations of six- (Fig. <xref rid="F5" ref-type="fig">5</xref>, inset 1) as well as seven-membered rings (Fig. <xref rid="F5" ref-type="fig">5</xref>, inset 2) are present in the sample. Currently, it is not known what drives the equilibrium between the two oligomeric states, and an efficient way to separate the two forms has yet to be found. Such heterogeneity may not have been discernable in vitrified specimens and may have produced unreliable features in a 3D reconstruction determined by cryo-EM of vitrified specimens.</p><fig id="F5"><label>Fig. 5</label><caption><title>Sample heterogeneity due to different oligomeric states.</title><p> Image of negatively stained T7 helicase/primase in the presence of dTDP. While the particles appear rather homogeneous in the micrograph, image classification revealed the protein formed six- (inset 1) as well as seven-membered rings. The scale bar corresponds to 50 nm and the inset panels have a side length of 30 nm.</p></caption><graphic xlink:href="bpo_v6_p23_m70f5lg"/></fig><p> <italic>Heterogeneity due to different conformations</italic>. Like many proteins, integrin α<sub>V</sub>β<sub>3</sub> undergoes an extensive conformational change upon activation (<xref rid="B9" ref-type="bibr">9</xref>). For the presented analysis a construct of integrin α<sub>V</sub>β<sub>3</sub> containing only the extracellular domains but lacking the transmembrane and cytoplasmic domains was used. In the presence of inactivating Ca<sup>2+</sup> ions α<sub>V</sub>β<sub>3</sub> adopts a compact conformation (Fig. <xref rid="F6" ref-type="fig">6</xref>a), while activating Mn<sup>2+</sup> ions induce an extended conformation (Fig. <xref rid="F6" ref-type="fig">6</xref>b). It is rare to find experimental conditions that shift a conformational equilibrium completely to one side or the other, leaving at least some residual conformational heterogeneity in the particle population. In the case of the α<sub>V</sub>β<sub>3</sub> construct, about 14% of the molecules were in the extended conformation even in the presence of Ca<sup>2+</sup>, whereas about 20% of the molecules remained in the compact conformation in the presence of Mn<sup>2+</sup>. Addition of the strong ligand mimetic cyclic peptide cyclo-RGDfV induced more than 98% of the molecules to adopt the extended conformation irrespective of the cation present (<xref rid="B9" ref-type="bibr">9</xref>). These results demonstrate that by simple quantification of the subpopulations, classification can provide quantitative information for example on the potency of activating agents.</p><fig id="F6"><label>Fig. 6</label><caption><title>Conformational equilibrium of integrin αVβ3.</title><p> a: Image of αVβ3 in the presence of inhibiting Ca2+ ions, where most of the molecules adopt a compact, closed conformation (insets 1 and 2 show representative class averages). Some of the molecules however can be seen in an extended, open conformation (arrows). b: In the presence of activating Mn2+ ions, the situation is reversed and most molecules are in the extended conformation (insets 1 to 4 show representative class averages), while only few molecules adopt the compact conformation (arrows). The scale bar corresponds to 50 nm and the inset panels have a side length of 40 nm. Figure <xref rid="F4" ref-type="fig">4</xref> modified and reprinted from (<xref rid="B9" ref-type="bibr">9</xref>). Copyright 2002 with permission from Elsevier.</p></caption><graphic xlink:href="bpo_v6_p23_m70f6lg"/></fig><p>Another example for the use of quantitative classification is the determination of binding constants.</p><fig id="F7"><label>Fig. 7</label><caption><title>Quantitative classification of TfR-Tf complexes.</title><p> a: Image of a 1:1 mixture of Tf and TfR, revealing five different particle types. b: Classification of the particle images into 30 classes yielded five unique projection averages corresponding to TfR with two Tf molecules bound (label 1: side view; label 2: top view), TfR with one Tf molecule bound (label 3), TfR by itself (label 4), and Tf by itself (label 5). One projection average for each unique class was selected for multi-reference alignment (black frames). c: Final projection averages of the five classes with the number of particle images in each class noted below. The scale bar corresponds to 50 nm and the panels in b and c have a side length of 30 nm.</p></caption><graphic xlink:href="bpo_v6_p23_m70f7lg"/></fig><p>Figure <xref rid="F7" ref-type="fig">7</xref>a shows an image, where TfR at a concentration of 25 nM was mixed with Tf at a 1:1 molar ratio. The mixture was directly applied to a grid and negatively stained. Figure <xref rid="F7" ref-type="fig">7</xref>b shows the classes resulting from classification into 30 classes. Five unique classes representing unliganded TfR and Tf and TfR with one or two Tf molecules bound were selected for multi-reference alignment (Fig. <xref rid="F7" ref-type="fig">7</xref>c). By determining the percentage of the TfR-containing complexes formed at varying mixing ratios, it is possible to obtain a binding curve and thus the binding constant. The high contrast and the few preferred orientations obtained by negative staining are crucial for this approach to work, which was indeed used to determine the association constant for the binding of the 19S regulatory particle to the 20S proteasome (<xref rid="B28" ref-type="bibr">28</xref>).</p><p> <italic>Analysis of mixed complexes</italic>. Complexes can be unstable, making it impossible to purify them to homogeneity. In these cases it is possible to simply mix the proteins and adsorb the mixture to an EM grid. Although such preparations are intrinsically very heterogeneous, this is sometimes the only practical approach for obtaining structural data for an unstable complex.</p><fig id="F8"><label>Fig. 8</label><caption><title>Complexes formed in a mixture of 20S proteasome, 19S regulatory particle and the α subunit of proteasome activator PA26.</title><p> Image of a mixture of 26S proteasome with an excess of PA26α. Due to its low binding affinity for the proteasome, many unbound PA26α rings are present in this preparation (circles). Classification of images of proteasome-containing particles yielded projection averages of all the expected complexes, namely 20S proteasome with two (inset 1) or one 19S regulatory particles bound (inset 2), unliganded 20S proteasome (inset 3), 20S proteasome with one (inset 4) or two PA26α rings bound (inset 5) as well as the ternary complex of a 20S proteasome with a 19S regulatory particle and a PA26α ring (inset 6). The scale bar corresponds to 50 nm and the inset panels have a side length of 48 nm. Figure modified and reprinted from (<xref rid="B10" ref-type="bibr">10</xref>). Copyright 2003 with permission from EMBO Journal.</p></caption><graphic xlink:href="bpo_v6_p23_m70f8lg"/></fig><p>This mixing approach was successfully used to determine the structure of the ternary complex formed by the 20S proteasome with the 19S regulatory particle and the proteasome activator PA26 (<xref rid="B12" ref-type="bibr">12</xref>). Since only the α subunit of the PA26 complex was available, which has a low binding affinity for the proteasome, a large excess of PA26α had to be mixed with 26S proteasome to obtain a ternary complex, and the mixture had to be adsorbed to an EM grid without any further purification. Images of this preparation were therefore dominated by small ring-shaped molecules in the background formed by unbound PA26α, but various proteasome-containing complexes could also be seen (Fig. <xref rid="F8" ref-type="fig">8</xref>). Classification revealed all the expected complexes, namely unliganded 20S proteasome, 20S proteasome with one or two 19S regulatory particles bound, 20S proteasome with one or two PA26 bound, and 20S proteasome with one 19S and one PA26 bound (Fig. <xref rid="F8" ref-type="fig">8</xref>, insets 1 to 6). No other approach would have allowed structural analysis of such a heterogeneous sample. If required even 3D reconstructions could have been produced for all six complexes by recording images of tilted specimens.</p><p> <italic>Use of classification with images of 2D crystals</italic>. Image classification is typically associated with single particle EM, but it can also be helpful in the analysis of 2D crystals formed by structurally distinct unit cells. In this approach the unit cells are extracted from the 2D array and classified as individual particles.</p><p>An early application of classification to 2D crystals was the analysis of the binding of maltose-binding protein (MBP) to maltoporin 2D crystals (<xref rid="B29" ref-type="bibr">29</xref>). MBP bound to only one out of three symmetry-related binding sites per maltoporin trimer. Classification was therefore used to select similar unit cells of the decorated maltoporin array and a projection map could be generated showing the outline of an MBP molecule interacting with a maltoporin trimer.</p><fig id="F9"><label>Fig. 9</label><caption><title>Single-particle processing of 2D crystals formed by RC-LH1 photounits from Rhodobacter sphaeroides.</title><p> a and b: Image of a negatively stained RC-LH1 2D crystal (a) and the corresponding calculated power spectrum (b). c: The same crystal area as in panel a after Fourier-peak filtration, revealing the individual RC-LH1 complexes. d: Projection structure of the RC-LH1 complex obtained by crystallographic averaging of the image shown in a. e: Single particle average of the unit cells marked in panel c without rotational alignment. f: Single particle average of the same unit cells used to generate the average in panel e after rotational alignment. While the RC in the center of the LH1 ring has no features in averages d and e, it has a distinct shape in average f. The scale bars in a and c correspond to 100 nm, the scale bar in b to (6 nm)-1, and panels d to f have a side length of 18 nm. Figure modified and reprinted from (<xref rid="B13" ref-type="bibr">13</xref>). Copyright 1998 with permission from Elsevier.</p></caption><graphic xlink:href="bpo_v6_p23_m70f9lg"/></fig><p>A different approach was used for the analysis of 2D crystals formed by reaction center (RC) – light-harvesting complex 1 (LH1) photounits from <italic>Rhodobacter sphaeroides</italic> (Fig. <xref rid="F9" ref-type="fig">9</xref>a and b) (<xref rid="B13" ref-type="bibr">13</xref>). The crystal contacts in these arrays were mediated by the 16-membered ring formed by the LH1 molecules (Fig. <xref rid="F9" ref-type="fig">9</xref>d). The RC in the center of the LH1 ring adopted a random orientation and was smeared out upon crystallographic (Fig. <xref rid="F9" ref-type="fig">9</xref>d) or single-particle averaging without prior alignment of the unit cells (Fig. <xref rid="F9" ref-type="fig">9</xref>e). Only by rotational alignment of the individual unit cells was it possible to resolve the projection structure of the RC in the center of the LH1 ring (Fig. <xref rid="F9" ref-type="fig">9</xref>f). A similar strategy was also used for 2D crystals formed by RC-LH1 photounits from <italic>Rhodospirillum rubrum</italic> (<xref rid="B30" ref-type="bibr">30</xref>).</p><p>The most vigorous single particle processing of a 2D crystal was applied to vitrified ordered arrays of the Na<sup>+</sup>/K<sup>+</sup>-ATPase from dog kidney. The resolution of the projection map improved from about 20 Å obtained with conventional electron crystallographic image processing (<xref rid="B31" ref-type="bibr">31</xref>) to 11 Å by the single particle/classification approach (<xref rid="B32" ref-type="bibr">32</xref>). As the borders between single particle and electron crystallographic processing are starting to blur, software packages are being developed that include classification modules that will be equally easy to apply to images of single particles and 2D crystals. One example is the Zephyr package that is currently being developed in the groups of David DeRosier and Nikolaus Grigorieff at Brandeis University.</p></sec></sec>
Ribosome formation from subunits studied by stopped-flow and Rayleigh light scattering	<p>Light scattering and standard stopped-flow techniques were used to monitor rapid association of ribosomal subunits during initiation of eubacterial protein synthesis. The effects of the initiation factors IF1, IF2, IF3 and buffer conditions on subunit association were studied along with the role of GTP in this process. The part of light scattering theory that is essential for kinetic measurements is high-lighted in the main text and a more general treatment of Rayleigh scattering from macromolecules is given in an appendix. </p>	<contrib contrib-type="author"><name><surname>Antoun</surname><given-names>Ayman</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Pavlov</surname><given-names>Michael Y.</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Tenson</surname><given-names>Tanel</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Ehrenberg</surname><given-names>Måns</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>In eubacteria, association of ribosomal subunits and initiation of protein synthesis require the three initiation factors IF1, IF2 and IF3 (<xref rid="B1" ref-type="bibr">1</xref>-<xref rid="B3" ref-type="bibr">3</xref>). In eukaryotes, subunit association and initiation of translation are more complex and require at least twelve initiation factors (<xref rid="B2" ref-type="bibr">2</xref>). All three prokaryotic initiation factors have their corresponding functional homologues in eukaryotes. Initiation factors IF1 and IF2 are close sequence and functional homologues of the eukaryotic initiation factors eIF1A (<xref rid="B4" ref-type="bibr">4</xref>) and eIF5B (<xref rid="B3" ref-type="bibr">3</xref>), respectively. Initiation factor IF3 has no sequence homology with any of the eukaryotic initiation factors (<xref rid="B2" ref-type="bibr">2</xref>). It has, however, several functions in common with eukaryotic eIF3 as well as with eukaryotic eIF1. The latter factor associates with eIF3 in mammals and is one of the subunits of eIF3 in yeast (<xref rid="B5" ref-type="bibr">5</xref>).</p><p>Termination of protein synthesis in eubacteria is carried out by either one of the class-1 peptide release factors RF1 or RF2 in a stop codon dependent way (<xref rid="B6" ref-type="bibr">6</xref>). After peptide release, rapid dissociation of the class-1 release factor is accomplished by the GTP-dependent action of the class-2 release factor RF3 (<xref rid="B7" ref-type="bibr">7</xref>-<xref rid="B9" ref-type="bibr">9</xref>). Subsequently, the ribosome is split by the combined activities of RRF, EF-G and IF3 (<xref rid="B10" ref-type="bibr">10</xref>), making the ribosomal 30S and 50S subunits ready for a new round of initiation of protein synthesis. Here, the 30S subunit, in complex with IF3, binds a messenger RNA, IF1, IF2:GTP and initiator tRNA (fMet-tRNA<sup>fMet</sup>) in a 30S pre-initiation complex (<xref rid="B1" ref-type="bibr">1</xref>), which rapidly recruits the 50S subunit in the formation of a 70S initiation complex. After GTP hydrolysis, IF2 rapidly dissociates from the 70S initiation complex, thereby making the ribosome ready to form the first peptide bond in a nascent protein (<xref rid="B11" ref-type="bibr">11</xref>). Subunit joining is an essential step in initiation of protein synthesis, but has in the past received comparatively little attention. </p><p>Subunit association or dissociation can be directly monitored by light scattering (<xref rid="B12" ref-type="bibr">12</xref>, <xref rid="B13" ref-type="bibr">13</xref>) or ultracentrifugation (<xref rid="B14" ref-type="bibr">14</xref>) methods. More recently, Rayleigh light scattering, in combination with stopped-flow techniques, was used to study rapid formation of translation competent ribosomes from different pre-initiation 30S complexes and the 50S subunit (<xref rid="B11" ref-type="bibr">11</xref>).</p><p>In contrast to the rapid and non-invasive light scattering techniques, ultra centrifugation methods provide little kinetic information on subunit association or dissociation, but have been used to monitor the extent of eukaryotic 80S assembly (<xref rid="B15" ref-type="bibr">15</xref>). Ultracentrifugation methods are of non-equilibrium type and subunits, originally in ribosome complexes, may become separated during a centrifugation run. This potential problem is aggravated by the high pressure that develops in the rotor during a centrifuge run which further promotes subunit dissociation (<xref rid="B16" ref-type="bibr">16</xref>).</p><p>In this work, we describe the principles of Rayleigh light scattering and explain how this method can be combined with stopped-flow techniques to monitor the kinetics of formation or disruption of macromolecular complexes. </p><p>We apply the method, using a standard stopped-flow instrument (<xref rid="B11" ref-type="bibr">11</xref>), to initiation of protein synthesis in eubacteria, and present novel experiments that high-light the roles of IF3 and GTP on IF2 for selective and rapid 70S initiation complex formation.</p><p>The method of light scattering is general and can also be used to study the formation or disruption of macromolecular complexes other than the ribosome.</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><sec><title> <bold>Chemicals and buffers</bold> </title><p>Nucleoside triphosphates (ATP, UTP, and GTP), radioactive amino acids and unlabelled nucleotides were from Amersham (USA). Non-hydrolysable GTP analogue GDPNP (GMPPNP), CTP, phosphoenolpyruvate (PEP), myokinase (MK), pyruvate kinase (PK), putrescine, spermidine, puromycin dihydrochloride, and non-radioactive amino acids were from Sigma (USA). All other chemicals were of analytical grade from Merck (Germany). Before use in binding and exchange assays, the guanine nucleotides GTP and GDP were further purified as described (<xref rid="B8" ref-type="bibr">8</xref>). All experiments were carried out in polymix buffer (<xref rid="B17" ref-type="bibr">17</xref>) which has the following final composition: [95 mM KCl, 5 mM NH<sub>4</sub>Cl, 5 mM Mg(OAc)<sub>2</sub>, 0.5 mM CaCl<sub>2</sub>, 8 mM putrescine, 1 mM spermidine, 5 mM potassium phosphate (KP) (pH 7.5) and 1 mM DTE]. One ml of this buffer is prepared by adding 0.1 ml of 10 times polymix, 0.05 ml of 20 times KP and 0.02 ml of 50 mM DTE to 0.83 ml of water. Preparation of 10 times polymix buffer is described in the protocol section. It contains the components of the polymix buffer at 10 times concentration but does not contain KP and DTE to avoid precipitation of calcium phosphate. </p></sec><sec><title> <bold>Components of the translation system</bold> </title><p>Synthetic mMFTI mRNA, encoding the tetra-peptide Met-Phe-Thr-Ile, was prepared according to (<xref rid="B18" ref-type="bibr">18</xref>). 70S ribosomes, 50S and 30S subunits were prepared from the <italic>E. coli </italic>strain MRE 600, using sucrose gradient zonal ultracentrifugation according to (<xref rid="B19" ref-type="bibr">19</xref>). Initiation factors were purified from overproducing strains according to (<xref rid="B20" ref-type="bibr">20</xref>). [<sup>3</sup>H]fMet-tRNA<sup>fMet</sup> and Phe-tRNA synthetase (PheRS) were prepared according to (<xref rid="B7" ref-type="bibr">7</xref>). Elongation factors EF-Tu, EF-Ts and tRNA<sup>Phe</sup> were purified according to (<xref rid="B21" ref-type="bibr">21</xref>).</p></sec><sec><title> <bold>Kinetics of macromolecular complex formation analyzed by stopped-flow and light scattering </bold> </title><p>In a typical light scattering experiment to monitor a binary complex formation between particles of type <italic>A</italic> and <italic>B</italic>, a solution containing particles <italic>A</italic> is rapidly mixed with a solution containing particles <italic>B</italic> and the intensity of light scattered perpendicular to the beam of illuminating light is recorded as a function of time. Initially, the mixture contains particles <italic>A</italic> and <italic>B</italic> at concentrations <italic>a(0)</italic> and <italic>b(0)</italic>, respectively, while the concentration, <italic>c(0)</italic>, of complex <italic>C</italic> is zero. The scattering intensity, <italic>I(t),</italic> at a time <italic>t</italic> after the mixing is the sum of the scattering intensities from free <italic>A-</italic>particles, free <italic>B-</italic>particles and <italic>C-</italic>complexes:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn01ehren.jpg"/> </inline-formula> </p><p> <italic>a(t)</italic> and <italic>b(t)</italic> are the concentrations of free particles <italic>A</italic> and <italic>B</italic>, <italic>c(t)</italic> is the concentration of complexes, C. <italic>I<sub>A</sub> </italic>, <italic>I<sub>B</sub> </italic> and <italic>I<sub>C</sub> </italic> are the scattering intensities per unit concentration for the corresponding particles and complexes. Since, for every <italic>C</italic> complex that is formed, one particle <italic>A</italic> and one particle <italic>B</italic> are consumed, <italic>a(t)</italic> and <italic>b(t)</italic> are related to the initial concentrations <italic>a(0)</italic> and <italic>b(0)</italic> as: a(t) = a(0) – c(t) and b(t) = b(0) – c(t). Introducing these mass relations in expression [1] gives</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn02ehren.jpg"/> </inline-formula> </p><p>where <italic>ΔI<sub>C</sub> </italic> is the increase in light scattering intensity when the two particles <italic>A</italic> and <italic>B</italic> form a complex <italic>C</italic>. For particles with dimensions much smaller than the wave length of the illuminating light, the scattering intensity of a particle is proportional to the square of its molecular mass and does not depend on particle shape (<xref rid="B12" ref-type="bibr">12</xref>) (see also Appendix 1). Since the complex <italic>C</italic> between particles <italic>A</italic> and <italic>B</italic> is just a bigger particle, <italic>ΔI<sub>C</sub> </italic> can be estimated as:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn03ehren.jpg"/> </inline-formula> </p><p> <italic>M<sub>A</sub> </italic> and <italic>M<sub>B</sub> </italic>are the molecular masses of the <italic>A</italic> and <italic>B</italic> particles, respectively, and <italic>Z</italic> is a proportionality coefficient for the particular experimental set up. Ribosomes and their subunits do not extend more than 30 nm, which is less than one tenth of the wave length (λ= 430 nm) of illuminating light used in most light scattering experiments on these particles (<xref rid="B11" ref-type="bibr">11</xref>-<xref rid="B13" ref-type="bibr">13</xref>). Therefore, relation [3] holds very well and it follows from this expression that, with <italic>x=M<sub>A</sub>/M<sub>B</sub> </italic>, the scattering intensity increases by a factor of <italic>1+2x/(1+x<sup>2</sup>)</italic> when molecules A and B form a complex. The largest relative increase is by a factor of two, when <italic>M<sub>A</sub>=M<sub>B</sub> </italic> so that <italic>x=1</italic>.</p><p>Relation [2] shows that the increase, <italic>I(t)-I(0)</italic>, in scattering intensity with time is directly proportional to the concentration <italic>c(t)</italic> of formed complexes. When complex formation has reached equilibrium, the plateau value, <italic>I<sub>eq</sub> </italic>, of the scattered intensity is given by</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn04ehren.jpg"/> </inline-formula> </p><p>Combining the experimentally measured parameters <italic>I(0), I(t)</italic> and <italic>I<sub>eq</sub> </italic>, and using the relations [2] and [4] one gets:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn05ehren.jpg"/> </inline-formula> </p><p>The time evolution of the function <italic>f<sub>c</sub>(t)</italic>, which is one at time zero and zero at infinite time, contains all kinetic information about the complex formation. The ratio <italic>f<sub>c</sub>(t)</italic> is the difference between the current and the equilibrium concentration of the complex C, normalized to the value of this difference at time zero. Notice that <italic>f<sub>c</sub>(t)</italic> can be obtained from the experimentally measured intensities <italic>I(0), I(t)</italic> and <italic>I<sub>eq</sub> </italic> without knowledge of the absolute value of <italic>ΔI<sub>C</sub> </italic>. Therefore, kinetic experiments can be interpreted without knowledge of the coefficient <italic>Z</italic> in relation [3], which depends on the experimental set up. </p><p>To exemplify the kind of kinetic information one can get from light scattering experiments, we consider the irreversible formation of a complex <italic>C</italic> from <italic>A-</italic> and <italic>B-</italic>particles that have the same initial concentration [<italic>A</italic>]<sub>0</sub> </p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn06ehren.jpg"/> </inline-formula> </p><p>The corresponding rate equation is:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn07ehren.jpg"/> </inline-formula> </p><p>Its analytical solution is </p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn08ehren.jpg"/> </inline-formula> </p><p>as can be verified by substituting <italic>c(t)</italic> in [7] with the expression for <italic>c(t)</italic> in [8]. After a long time all particles <italic>A</italic> and <italic>B</italic> will eventually end up in complexes <italic>C</italic> and an equilibrium concentration of complexes <italic>c<sub>eq</sub>=[A]<sub>0</sub> </italic> will be reached. </p><p>Substituting the above expressions for <italic>c(t)</italic> and <italic>c<sub>eq</sub>=[A]<sub>0</sub> </italic> into [5] one obtains a very simple expression for <italic>f<sub>c</sub>(t)</italic>:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn09ehren.jpg"/> </inline-formula> </p><p>Accordingly, a plot of <italic>1/f<sub>c</sub>(t)</italic> versus time gives a straight line with slope <italic>k<sub>a</sub>[A]<sub>0</sub> </italic>, from which the association rate constant <italic>k<sub>a</sub> </italic> can be obtained from linear regression and knowledge of the initial concentrations <italic>[A]<sub>0</sub> </italic>. In general, however, it is better to use non-linear regression methods to obtain <italic>k<sub>a</sub> </italic> along with normalization parameters. For this, the scattered intensity <italic>I(t)</italic> can be written as:</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p35_eqn10ehren.jpg"/> </inline-formula> </p><p>This relation follows from Eq. [5], with <italic>a<sub>0</sub> </italic> equal to <italic>I<sub>eq</sub> </italic> and <italic>a<sub>1</sub> </italic> equal to <italic>I<sub>eq</sub> </italic> <italic>- I(0)</italic>. A best fit of this theoretical expression for <italic>I(t)</italic> to its experimental counterpart by variation of the parameters <italic>k<sub>a</sub> </italic>, <italic>a<sub>0</sub> </italic> and <italic>a <sub>1</sub> </italic>, e.g. with the Marquardt algorithm (<xref rid="B22" ref-type="bibr">22</xref>), gives an estimate of <italic>k<sub>a</sub> </italic> along with the expected error (standard deviation) of this estimate. The content of this section is all that is required to apply Rayleigh light scattering to the kinetic analysis of macromolecular complex formation. An extended and more detailed description of light scattering theory and its experimental applications can be found in Appendix 1.</p></sec><sec><title> <bold>Light scattering experiments</bold> </title><p>Association of ribosomal subunits was monitored with light scattering after their rapid mixing in an SX-18MV stopped-flow instrument (Bio-sequential SX-18MV, Applied Photophysics, Leatherhead, UK) equipped with Xenon arc light source. To study the role of IF3 in subunit association, two mixtures, A and B, were prepared. Mixture A contained 4µM mMFTI mRNA, 0.5 mM ATP, 0.5 mM of GTP, 2 µM 30S and either (i) no additional components, (ii) 4 µM IF3, (iii) 2 µM IF1, 4µM IF2 or (iv) 4 µM IF3, 2 µM IF1, 4µM IF2 and 4 µM [<sup>3</sup>H]fMet-tRNA<sup>fMet</sup> as indicated. Mixture B contained 0.5 mM ATP, 0.5 mM GTP and 2 µM 50S. To remove dust particles, the mixtures were spun for 3 min at 14000 rpm in an Eppendorf centrifuge before they were loaded into the syringes of the stopped-flow instrument and pre-incubated at 37°C for at least 5 minutes. After rapid mixing, light scattering at 436 nm at right angle to the illuminating light was recorded as a function of time. The instrument was in emission detection mode, the photomultiplier voltage set to 520 V, and the time constant of the noise reduction filter to 5 ms. The volume of mixes (around 1 ml each) loaded into the syringes of the stopped-flow instrument was sufficient for at least ten independent time traces. Kinetic parameters were obtained for each individual trace by non-linear regression (see the section “Curve fitting” below) and used to obtain average estimates of rate constants and the standard deviations of these estimates.</p><p>To study the effects of GTP and GDP on the kinetics of subunit association in the presence of IF2, two mixtures, A and B, were prepared. Mixture A contained 0.5 mM ATP, 1 mM PEP, 0.5 mM of either GTP, GDPNP or GDP, 4 µM [<sup>3</sup>H]fMet-tRNA<sup>fMet</sup>, 2 µM 30S, 4µM mMFTI mRNA, 2 µM IF1, 4 µM IF2 and 4 µM IF3. Mixture B contained 0.5 mM ATP, 1 mM PEP, 2 µM 50S. Both mixtures were centrifuged for 3 min at 14000 rpm, loaded into the syringes of the stopped-flow instrument and pre-incubated at 37°C for at least 5 min before fast mixing. </p><p>Effects of buffer composition on the kinetics of 70S initiation complex formation were studied as follows. Two mixtures, A and B, were prepared and loaded into the syringes of the stopped-flow instrument. Mixture A contained 2 µM 70S ribosomes, 4µM mMFTI mRNA, 4 µM IF1, 4 µM IF3, 0.5 mM ATP, 0.5 mM GTP and indicated concentrations of PEP and Mg (OAc)<sub>2</sub> in polymix buffer. Mixture B contained 0.5 mM ATP, 0.5 mM GTP, 4 µM [<sup>3</sup>H]fMet-tRNA<sup>fMet</sup>, 4 µM IF2 and the same concentrations of PEP and Mg(OAc)<sub>2</sub> in polymix buffer as in mixture A. Mixture A was pre-incubated for at least 10 min at 37°C to ensure ribosome dissociation into 30S and 50S subunits. The formation of 70S initiation complexes was then initiated by mixing the mixtures A and B in a stopped-flow instrument as described above. </p><p>The kinetics of association of naked ribosomal subunits was studied in the following way. Two mixtures, A and B, were prepared and loaded into the syringes of the stopped-flow instrument. Mixture A contained 2 µM 30S ribosomal subunits, 0.5 mM ATP, 0.5 mM GTP, 1.5 mM PEP and indicated concentrations of Mg(OAc)<sub>2</sub> in polymix buffer. Mixture B contained 2 µM 50S ribosomal subunits, 0.5 mM ATP, 0.5 mM GTP and the same concentrations of PEP and Mg(OAc)<sub>2</sub> in polymix buffer as in mixture A. The formation of 70S initiation complexes was then initiated by mixing mixtures A and B in a stopped-flow instrument as described above.</p></sec><sec><title> <bold>Dipeptide-Formation assay</bold> </title><p>The effect of IF3 on the formation of translation-competent 70S initiation complexes was also studied with a dipeptide formation assay. To this end, two mixtures, A and B, were first prepared. Mixture A contained 0.5 mM ATP, 2 mM PEP, 0.5 mM GTP, 1.5 µM 30S, 2.5 µM mMFTI mRNA, 2.5 µM IF1, 2.5 µM IF2. Mixture B contained 0.5 mM ATP, 2 mM PEP, 2 µM 50S, 3 µM EF-Tu, 5 µM tRNA<sup>Phe</sup>, 30 µM phenylalanine, 1 µg/ml PK, 0.1 µg/ml MK and 10 U/ml PheRS (1 U of PheRS aminoacylates one pmol of tRNA per second). Then, 2.5 µM [<sup>3</sup>H]fMet-tRNA<sup>fMet</sup> was added either to mixture A or to mixture B. After pre-incubation for 10 min at 37°C the mixtures A (0.025 ml) and B (0.025 ml) were loaded into a quench flow instrument (KinTech, USA), mixed and quenched after the indicated times by 50% formic acid. The samples were centrifuged and the amount of formed fMet-Phe-tRNA<sup>Phe</sup> in the pellet was determined by HPLC as described previously (<xref rid="B18" ref-type="bibr">18</xref>).</p></sec><sec><title> <bold>Curve fitting</bold> </title><p>The association rate constant (k<sub>a</sub>) for subunit association in the stopped-flow light scattering experiments was estimated by non-linear regression (<xref rid="B22" ref-type="bibr">22</xref>) or by the <italic>Origin Program</italic>, using the three-parameter relation Eq. [10].</p></sec></sec><sec><title>Results</title><sec><title> <bold>IF3 as anti-association factor</bold> </title><p>In the absence of initiation factors, the 30S:mRNA complex associated with the 50S subunit with an association rate constant k<sub>a</sub>=1.2 μM<sup>-1</sup> s<sup>-1</sup> (Fig. <xref rid="F1" ref-type="fig">1</xref>A). Fig. <xref rid="F1" ref-type="fig">1</xref>C shows that the addition of IF1, IF2 and GTP to the 30S:mRNA complex resulted in a faster association of 30S with 50S (k<sub>a</sub>=4.1 μM<sup>-1</sup> s<sup>-1</sup>). In the presence of only IF3, there was no complex formation between 50S and 30S:mRNA alone (Fig. <xref rid="F1" ref-type="fig">1</xref>B), or together with IF1, IF2 and GTP (not shown).</p><fig id="F1"><label>Fig. 1</label><caption><title>The anti-association activity of IF3.</title><p> The extent of 70S initiation complex formation was monitored as a function of time by light scattering after rapid mixing in the stopped-flow instrument of a volume containing 30S subunits, mRNA, GTP and initiation factors as indicated with a volume containing 50S subunits. Time traces obtained with no initiation factors added to the 30S subunits (Panel A), with only IF3 added (Panel B), with IF1 and IF2 added (Panel C) and with IF1, IF2, IF3 and [<sup>3</sup>H]fMet-tRNAfMet added (Panel D).</p></caption><graphic xlink:href="bpo_v6_p35_m71f1lg"/></fig><p>The results of these experiments, summarized in Table <xref rid="T1" ref-type="table">1</xref>, demonstrate the ability of IF3 to block subunit association when the 30S pre-initiation complex lacks initiator tRNA. When, however, fMet-tRNA<sup>fMet </sup>was present in the pre-initiation 30S:mRNA complex together with IF1, IF2 and GTP, the block was removed and the ribosomal subunits joined with an association rate constant <italic>k<sub>a</sub> </italic> <italic> = </italic>12 μM<sup>-1</sup>s<sup>-1</sup> (Fig. <xref rid="F1" ref-type="fig">1</xref>D).</p><table-wrap id="T1"><label>Table 1</label><caption><p>Association rate constants of 50S subunits with 30S:mRNA in the presence of different combinations of Initiation Factors and fMet-tRNA.</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left" colspan="1"> <bold>Factors added</bold> </td><td align="left" colspan="1"> <bold>k<sub>ass</sub>(μM<sup>-1</sup>s<sup>-1</sup>)</bold> </td></tr><tr><td align="left" colspan="1">none</td><td align="left" colspan="1">1.2 ± 0.13 </td></tr><tr><td align="left" colspan="1">+IF3</td><td align="left" colspan="1"><0.0</td></tr><tr><td align="left" colspan="1">+IF1+IF2:GTP</td><td align="left" colspan="1">4.1 ± 0.3</td></tr><tr><td align="left" colspan="1">+IF1+IF2:GTP+IF3</td><td align="left" colspan="1">< 0.01</td></tr><tr><td align="left" colspan="1">+IF1+IF2:GTP+IF3+fMet-tRNA</td><td align="left" colspan="1">12.2 ± 1.6</td></tr><tr><td align="left" colspan="1">+IF1+IF2:GTP+IF3+fMet-tRNA ()</td><td align="left" colspan="1">8.6 ± 1.2</td></tr><tr><td align="left" colspan="1">+IF1+IF2:GDP+IF3+fMet-tRNA ()</td><td align="left" colspan="1">0.13 ± 0.03</td></tr><tr><td align="left" colspan="2">The association rate constant <italic>k<sub>ass</sub> </italic> was calculated as an average (<italic>k<sub>av</sub> </italic>) over rate constants (<italic>k<sub>i</sub> </italic>) obtained from <italic>n</italic> (between six and eight) individual time traces of light scattering in a stopped-flow experiment. The error is the standard deviation σ obtained from the variance, estimated as</td></tr><tr><td align="left" colspan="2"> <disp-formula> <graphic xlink:href="bpo_v6_p35_eqn11ehren"/> </disp-formula> </td></tr><tr><td align="left" colspan="2">(*) The reaction buffer contained 1 mM phospho<italic>enol</italic>pyruvate (PEP).</td></tr></tbody></table></table-wrap><p>The vital importance of IF3 for proper initiation of protein synthesis in eubacteria is further illustrated by quench-flow experiments that monitored the rate of dipeptide formation in the absence of IF3 (Fig. <xref rid="F2" ref-type="fig">2</xref>). In one experiment, pre-initiation 30S:mRNA complex together with fMet-tRNA<sup>fMet</sup>, IF1 and IF2 was mixed with 50S complex and all factors needed for peptide-bond formation (Fig. <xref rid="F2" ref-type="fig">2</xref>A). In an otherwise identical parallel experiment, fMet-tRNA<sup>fMet</sup> was present in the 50S, rather than in the 30S, mixture when the rate of peptide bond formation was followed (Fig. <xref rid="F2" ref-type="fig">2</xref>B). The rate of peptidyl-transfer was fast in the former (Fig. <xref rid="F2" ref-type="fig">2</xref>A), but virtually zero in the latter experiment (Fig. <xref rid="F2" ref-type="fig">2</xref>B). The absence of dipeptide formation in the second experiment reflects the rapid formation of a translationally inactive 70S complex lacking initiator tRNA. This complex is of the type seen with light scattering in Fig. <xref rid="F1" ref-type="fig">1</xref>C. This inactive ribosome complex was, in other words, formed before initiator tRNA had time to bind to the 30S:mRNA pre-initiation complex. </p><fig id="F2"><label>Fig. 2</label><caption><title>Rate of initiation in the absence of IF3 monitored by di-peptide formation.</title><p> The extent of dipeptide formation was monitored as a function of time after rapid mixing in a quench flow instrument of a volume containing 30S subunits, mRNA, GTP, IF1 and IF2 with an equal volume containing 50S subunits. Initiator tRNA was present either in the 30S or in the 50S mix. Time curves obtained with [<sup>3</sup>H]fMet-tRNAfMet added with 30S (Panel A), [<sup>3</sup>H]fMet-tRNAfMet added with 50S (panel B).</p></caption><graphic xlink:href="bpo_v6_p35_m71f2lg"/></fig></sec><sec><title> <bold>The role of GTP in subunit association</bold> </title><p>It was recently shown that GTP on IF2 is important for fast association of a 30S pre-initiation complex with the 50S subunit (<xref rid="B11" ref-type="bibr">11</xref>). Those experiments were carried out with the functionally active β-form of IF2, lacking part of the N-terminal domain of the α-form of the factor (<xref rid="B23" ref-type="bibr">23</xref>, <xref rid="B24" ref-type="bibr">24</xref>). Since differences in the GTP dependency of these factors cannot be excluded, we present here a similar study, but with a his-tagged version of the full-length α-form of IF2. Pre-initiation 30S complexes were formed with IF1, IF2, IF3, mRNA and fMet-tRNA in the presence of GTP, GDP or the GTP analogue GDPNP. Subsequently, these were rapidly mixed with 50S subunits in the stopped-flow instrument and the intensity of the scattered light was recorded. The rate constant for subunit association was around 8.5 μM<sup>-1</sup> s<sup>-1</sup> with GTP (Fig. <xref rid="F3" ref-type="fig">3</xref>A) or GDPNP (not shown) and 0.13 μM<sup>-1</sup> s<sup>-1</sup> with GDP (Fig. <xref rid="F3" ref-type="fig">3</xref>B). This means that GTP accelerated subunit formation sixty-fold compared to the rate obtained with GDP, in line with our previous results with the β-form of IF2 (<xref rid="B11" ref-type="bibr">11</xref>), but in contrast to results obtained by others (<xref rid="B25" ref-type="bibr">25</xref>).</p><fig id="F3"><label>Fig. 3</label><caption><title>The effects of G-nucleotides on the association of 30S pre-initiation complex with 50S subunits.</title><p> The extent of 70S initiation complex formation was monitored as a function of time by light scattering after rapid mixing of pre-initiation 30S complexes with 50S subunits in a stopped-flow instrument. Traces obtained with GTP (Panel A) and GDP (Panel B).</p></caption><graphic xlink:href="bpo_v6_p35_m71f3lg"/></fig></sec><sec><title>Dependence of the subunit association rate constant on buffer composition</title><p>It is well known that the concentration of magnesium ions, ionic strength and composition of the buffer have a profound effect on the rate and accuracy of protein synthesis (<xref rid="B21" ref-type="bibr">21</xref>). It has also been demonstrated that the association rate constant of ‘empty’ ribosomal subunits increases by almost an order of magnitude when the Mg<sup>2+</sup> concentration in the buffer increases from 4 to 8 mM (<xref rid="B13" ref-type="bibr">13</xref>). It was therefore of considerable interest to study the effect of Mg<sup>2+</sup> and other components, like phospho-<italic>enol</italic>pyruvate (PEP), usually included in buffers for <italic>in vitro</italic> translation on the rate of formation of ‘real’ 70S initiation complexes. In the light scattering experiments described below ribosomes were first dissociated into their subunits in the presence of mRNA and initiation factors IF1 and IF3 in a buffer of indicated composition and then initiation factor IF2:GTP was added together with fMet-tRNA to dissociated 70S ribosomes in the stopped-flow instrument.</p><fig id="F4"><label>Fig. 4</label><caption><title>The effects of buffer composition on the rate of 70S initiation complex formation.</title><p> The extent of 70S complex formation was monitored as a function of time by light scattering after rapid mixing of mixture A containing dissociated 70S ribosomes together with IF1, IF3 and mRNA with mixture B containing IF2:GTP together with fMet-tRNA. Complex formation in polymix buffer (PM) with 3 mM of free Mg<sub>2+</sub> (Panel A), in PM buffer with 7 mM of free Mg<sub>2+</sub> (Panel B) and in PM buffer with 3 mM free Mg<sub>2+</sub> plus 10 mM of PEP (Panel C).</p></caption><graphic xlink:href="bpo_v6_p35_m71f4lg"/></fig><p>Fig. <xref rid="F4" ref-type="fig">4</xref> shows that the rate constant of subunit association increased by only 50% from 7 µM<sup>-1</sup>s<sup>-1</sup> to about 10 µM<sup>-1</sup>s<sup>-1</sup> when the free Mg<sup>2+</sup> concentration increased from 3 to 7 mM in our standard polymix buffer (compare Fig. <xref rid="F4" ref-type="fig">4</xref>A and <xref rid="F4" ref-type="fig">4</xref>B). At the same time, the addition of 10 mM PEP to polymix buffer (Fig. <xref rid="F4" ref-type="fig">4</xref>C) resulted in a three-fold decrease in the rate constant for formation of the 70S initiation complex to 2.3 µM<sup>-1</sup>s<sup>-1</sup>.</p><p>An unexpectedly modest effect of Mg<sup>2+</sup> on the association rate of 50S subunits with pre-initiated 30S complexes (see Table <xref rid="T2" ref-type="table">2</xref>) compared to the large effects seen for association of “naked” 30S and 50S subunits observed by Wishnia <italic>et al</italic>. (<xref rid="B13" ref-type="bibr">13</xref>) may be due to the presence of mRNA, fMet-tRNA or initiation factors in the 30S pre-initiation complex. Alternatively, the difference could be due to the different ionic milieu in the polymix buffer compared to that in the buffer used in (<xref rid="B13" ref-type="bibr">13</xref>). The latter work employed a simple TMN buffer containing 10 mM Tris pH 7.5, 50 mM NH<sub>4</sub>Cl, 7 mM β-mercaptoethanol and different concentrations of MgCl<sub>2</sub>.</p><table-wrap id="T2"><label>Table 2</label><caption><p>Dependence of association rate constants of 70S initiation complex (IC) or naked 70S ribosomes on buffer conditions. 70S initiation complexes were formed from 30S and 50S subunits in the presence of mRNA, IF1, IF2:GTP, IF3 and fMet-tRNA.</p></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left" colspan="1"> <bold>Free Mg<sup>++</sup> </bold> </td><td align="left" colspan="1"> <bold>PEP</bold> </td><td align="left" colspan="1"> <bold>k<sub>ass</sub> (μM<sup>-1</sup> s<sup>-1</sup>)</bold> </td><td align="left" colspan="1"> <bold>70S type </bold> </td></tr><tr><td align="left" colspan="1">3</td><td align="left" colspan="1">0</td><td align="left" colspan="1">6.9± 0.7</td><td align="left" colspan="1">IC</td></tr><tr><td align="left" colspan="1">7</td><td align="left" colspan="1">0</td><td align="left" colspan="1">9.6± 1.5</td><td align="left" colspan="1">IC</td></tr><tr><td align="left" colspan="1">3</td><td align="left" colspan="1">10</td><td align="left" colspan="1">2.3± 0.4</td><td align="left" colspan="1">IC</td></tr><tr><td align="left" colspan="1">3</td><td align="left" colspan="1">1.5</td><td align="left" colspan="1">10.1± 0.9</td><td align="left" colspan="1">“naked”</td></tr><tr><td align="left" colspan="1">7</td><td align="left" colspan="1">1.5</td><td align="left" colspan="1">18.3± 1.8</td><td align="left" colspan="1">“naked”</td></tr><tr><td align="left" colspan="4">The association rate constant <italic>k<sub>ass</sub> </italic> and errors were calculated as for Table <xref rid="T1" ref-type="table">1</xref>.</td></tr></tbody></table></table-wrap><fig id="F5"><label>Fig. 5</label><caption><title>The effects of the level of magnesium on the rate of naked 70S formation formation from its subunits.</title><p> The extent of 70S complex formation was monitored as a function of time by light scattering after rapid mixing of mixture A containing 30S ribosomes with mixture B containing 50S subunits. 70S formation in polymix buffer (PM) with 3 mM of free Mg<sup>2+</sup> (Panel A), in PM buffer with 7 mM of free Mg<sup>2+</sup> (Panel B).</p></caption><graphic xlink:href="bpo_v6_p35_m71f5lg"/></fig><p>To discriminate between these possibilities we have measured the association rate of naked 30S and 50S subunits in polymix buffer containing either 3 or 7 mM of free Mg<sup>2+</sup>. The results shown in Fig. <xref rid="F5" ref-type="fig">5</xref> clearly demonstrates that the increase in Mg<sup>2+</sup> concentration in polymix buffer from 3 to 7 mM results in about 80% increase in the subunit association rate from approximately 10 µM<sup>-1</sup>s<sup>-1</sup> to 18 µM<sup>-1</sup>s<sup>-1</sup> which is comparable to the 50% increase observed for pre-initiated 30S complexes and 50 subunits (see Table <xref rid="T2" ref-type="table">2</xref>). Thus, the different effect of Mg<sup>2+</sup> on subunit association in the two buffer systems is probably due to the difference in composition of the buffers and not to the presence or absence of initiation factors, mRNA or fMet-tRNA.</p></sec></sec><sec><title>Discussion</title><p>This work demonstrates the power of combining stopped-flow and light scattering techniques for experimental studies of how ribosomal subunits join during initiation of protein synthesis. Light scattering techniques were used in early experiments to determine how the equilibrium constant for subunit association depends on translation initiation factors (<xref rid="B12" ref-type="bibr">12</xref>). The kinetics of association of naked 30S and 50S subunits and its dependence on buffer conditions have previously been studied with stopped-flow techniques (<xref rid="B13" ref-type="bibr">13</xref>) and the effect of IF3 on the rate of ribosome splitting has been addressed with light scattering and manual mixing (<xref rid="B12" ref-type="bibr">12</xref>). However, under near-physiological conditions used in our <italic>in vitro</italic> experiments (<xref rid="B21" ref-type="bibr">21</xref>), subunit association catalyzed by initiation factors (Figs. <xref rid="F1" ref-type="fig">1</xref> and <xref rid="F2" ref-type="fig">2</xref>; Antoun <italic>et al</italic>. (<xref rid="B11" ref-type="bibr">11</xref>)) and ribosome splitting, catalyzed by EF-G, RRF and IF3 (<xref rid="B10" ref-type="bibr">10</xref>), are rapid processes and their study therefore requires the combination of stopped-flow techniques and light-scattering. Here, we used stopped-flow with light scattering techniques to demonstrate the anti-association property of IF3 in the absence of initiator tRNA (Fig. <xref rid="F1" ref-type="fig">1</xref>), and complemented these measurements with quench-flow experiments, performed under similar conditions, to follow the rate of formation of the first peptide bond after initiation of protein synthesis (Fig. <xref rid="F2" ref-type="fig">2</xref>). The IF3 dependent block in the association of ribosomal subunits can be removed by the presence of initiator tRNA, IF1 and IF2. In cases when the subunits associate in the absence of IF3 and initiator tRNA the formed 70S ribosomes are unable to participate in protein synthesis. This suggests that IF3 plays a fundamental role in preventing premature ribosome formation in the absence of initiator tRNA. </p><p>We also demonstrated the fundamental role of GTP for fast subunit association catalyzed by the α-form of IF2 during initiation of eubacterial protein synthesis, in line with previous results obtained with the β-form of IF2 (<xref rid="B11" ref-type="bibr">11</xref>). During exponential growth of bacteria, ribosomes load on to the 5’ end of an mRNA each four seconds (<xref rid="B26" ref-type="bibr">26</xref>). The distance between ribosomes in a polysome is around 230 nucleotides (<xref rid="B26" ref-type="bibr">26</xref>). With a rate of 20 codons/s for protein elongation (<xref rid="B27" ref-type="bibr">27</xref>) and the need to clear the occluded ribosome binding site (<xref rid="B1" ref-type="bibr">1</xref>) to allow for the binding of the next ribosome, the lower limit for the initiation rate is about 0.3 s<sup>-1</sup>. This rate includes 30S docking to mRNA, fMet-tRNA and IF2 binding and subunit joining. Taking into account the results in Table <xref rid="T1" ref-type="table">1</xref> and that the concentrations of free ribosomal subunits in the cell are around 1 μM (<xref rid="B1" ref-type="bibr">1</xref>) one can conclude the rate of subunit joining catalyzed by IF2:GTP observed here is compatible with the rate of initiation <italic>in vivo</italic>. </p><p>The presented light scattering experiments show also the importance of a proper choice of buffer conditions to study the kinetics of ribosomal reactions. All experiments presented in this paper were performed in polymix buffer that mimics the ionic milieu of the bacterial cell (<xref rid="B21" ref-type="bibr">21</xref>). We have found, however, that some additional, supposedly “neutral,” components, often included in standard <italic>in vitro</italic> translation systems, like phosphoenolpyruvate (PEP) may result in considerable alteration in the rate of 70S complex formation (Table <xref rid="T2" ref-type="table">2</xref>). Even an addition of small amounts of PEP (1 mM) to the reaction mixture results in a noticeable decrease of the rate of association of pre-initiated 30S complexes with 50S subunits (Table <xref rid="T1" ref-type="table">1</xref>). The reason for this PEP effect is not clear at present. The effect of an increase of the free Mg<sup>2+</sup> concentration from 3 to 7 mM in polymix buffer on the rate of 70S initiation complex was however much more modest, i.e. only 50%. This 50% effect is, nevertheless, quite comparable with an 80% increase in the association rate for “naked” ribosomal subunits upon the same increase in free Mg<sup>2+</sup> concentration (see Table <xref rid="T2" ref-type="table">2</xref>). Comparison with published data (<xref rid="B13" ref-type="bibr">13</xref>) on association of “naked” subunits shows, however, that the increase in free Mg<sup>2+</sup> concentration from 3 to 7 mM results in a drastic increase in the association rate of ‘naked’ subunits from 0.63 μM<sup>-1</sup>s<sup>-1</sup> to about 20 μM<sup>-1</sup>s<sup>-1</sup>. This discrepancy is likely to be due to the absence of organic polyamines such as putrescine and spermidine in the TNM buffer system [10 mM Tris pH 7.5, 50 mM NH<sub>4</sub>Cl, 2 to 8 mM MgCl<sub>2</sub> and 7 mM β-mercaptoethanol] used in the previous work (<xref rid="B13" ref-type="bibr">13</xref>). Addition of polyamines corresponds, to a first approximation, to an effective increase of Mg<sup>2+ </sup>concentration in the buffer since polyamines mimic the most important, electrostatic, contribution of Mg<sup>2+</sup> in shielding phosphates of rRNA and reducing the electrostatic repulsion between the subunits (<xref rid="B13" ref-type="bibr">13</xref>). The association rate constant of 10 μM<sup>-1</sup>s<sup>-1</sup> at 3 mM free Mg<sup>2+</sup> in polymix buffer is similar to the rate constant of 9.2 µM<sup>-1</sup>s<sup>-1</sup> observed by Wishnia <italic>et al</italic>. (<xref rid="B13" ref-type="bibr">13</xref>) at 5.5 mM of free Mg<sup>2+</sup>. It seems therefore more appropriate to compare our results with those in TNM buffer upon the increase of Mg<sup>2+ </sup>from 5.5 to 9.5 mM. Published data (<xref rid="B13" ref-type="bibr">13</xref>) show that the association rate of naked subunits plateaus around 7.5 mM Mg<sup>2+</sup> reaching 22 μM<sup>-1</sup>s<sup>-1</sup>. If this value of 22 μM<sup>-1</sup>s<sup>-1</sup> is really a plateau, we will get a very good agreement for the effect of Mg<sup>2+</sup> on subunit association in two different buffer systems<bold>. </bold> </p><p>The experiments described here were performed with a standard stopped-flow instrument (SX-18MV, Applied Photophysics, Leatherhead, UK) in fluorescence mode. The required ribosome concentration was in the μM range, and all solutions were centrifuged for 3 min at 14000 rpm to remove dust particles and aggregates before they were loaded into the syringes of the stopped-flow instrument. The stopped-flow measurements successfully covered a broad range of subunit association times, from 10 ms to 30 min.</p><p>As described, the scattering intensity is proportional to the molar concentration of particles and to the square of their molecular weight. Therefore, the scattering intensity from a 1 μM solution of 30S subunits (Mw 900 kD) equals the scattering intensity from a 100 μM solution of 90 kD proteins. Accordingly, light scattering methods can be used to monitor the association kinetics also of proteins with considerably smaller molecular weights than the ribosome and its subunits, albeit at higher protein concentrations and with a larger investment in the total amount of protein.</p></sec>
Organotypic cocultures as skin equivalents: A complex and sophisticated in vitro system	<p>To assess the role of genes required for skin organogenesis, tissue regeneration and homeostasis, we have established in vitro skin equivalents composed of primary cells or cell lines, respectively. In these organotypic cocultures keratinocytes generate a normal epidermis irrespective of the species and tissue origin of fibroblasts. The combination of cells derived from mouse and human tissues facilitates the identification of the origin of compounds involved in epidermal tissue reconstitution and thus the precise analysis of growth regulatory mechanisms.</p>	<contrib contrib-type="author"><name><surname>Stark</surname><given-names>Hans-Jürgen</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Szabowski</surname><given-names>Axel</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib><contrib contrib-type="author"><name><surname>Fusenig</surname><given-names>Norbert E.</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author" corresp="yes"><name><surname>Maas-Szabowski</surname><given-names>Nicole</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>In skin, epithelial cells develop an orderly structured and well-organized epithelium consisting of basal, spinous, granular, and cornified strata. Their development during embryogenesis, regeneration in wound healing, as well as the maintenance of the homoeostasis of the epidermis depends on epithelial interactions with the underlying connective tissue, the dermis. The continuous process involving keratinocyte proliferation and terminal differentiation is mainly regulated by mesenchymal influences (<xref rid="B1" ref-type="bibr">1</xref>). To examine cellular growth, signalling processes and cell-cell interactions, culture systems have been developed combining the two major cell types of skin, keratinocytes and fibroblasts. In simple two-dimensional feeder-layer cocultures combining postmitotic dermal fibroblasts (feeder cells) and epidermal keratinocytes, properly stratified epithelia are not formed. Only in advanced three-dimensional in vitro systems keratinocytes develop well-ordered epithelia, thus offering an opportunity to analyse the cellular mechanisms of tissue formation, such as cell-cell interactions, the regulation of proliferation and differentiation as well as the reepithelialisation process after wounding (<xref rid="B2" ref-type="bibr">2</xref>). To reconstruct skin <italic>in vitro</italic> several organ-like culture systems have been established (<xref rid="B3" ref-type="bibr">3</xref>-<xref rid="B6" ref-type="bibr">6</xref>). In so called organotypic cocultures, epidermal keratinocytes grow air-exposed on a matrix of native collagen type I, extracted from rat tail tendon or calf skin. To become functional dermal equivalents, the collagen gels contain viable fibroblasts which reorganize this matrix by producing extracellular matrix components comparable to the wound situation (<xref rid="B7" ref-type="bibr">7</xref>, <xref rid="B8" ref-type="bibr">8</xref>). The cultures are nourished by diffusion from below, optionally in serum containing medium or in defined medium (<xref rid="B9" ref-type="bibr">9</xref>). </p><p>The characterisation of paracrine acting factors produced by either fibroblasts or keratinocytes is difficult in a homologous skin model containing human cells in both compartments. On the other hand, in a heterologous skin equivalent harbouring cell types of different species, the source of the respective gene products can be easily and unequivocally distinguished and assigned to the specific cell types (<xref rid="B10" ref-type="bibr">10</xref>).</p><p>Those <italic>in vitro</italic> skin equivalents are generally composed of freshly isolated cells from human skin specimens. However, more preferable tools for large scale examinations - either the molecular analysis of regulatory processes or pharmaco-toxicological studies - are skin equivalents composed of cell lines, which are easier to handle and to standardize. In our experiments, we chiefly utilize the spontaneously transformed human keratinocyte cell line HaCaT derived from normal human trunk skin (<xref rid="B11" ref-type="bibr">11</xref>). Despite the altered karyotype and unlimited growth potential, HaCaT cells are not tumorigenic and form an orderly structured and differentiated epidermis with essentially all structural and functional features after transplantation onto athymic nude mice (<xref rid="B11" ref-type="bibr">11</xref>-<xref rid="B13" ref-type="bibr">13</xref>). In organotypic cocultures in culture medium supplemented with TGF-α HaCaT cells form well organised and differentiated epithelia in the presence of human or mouse fibroblasts, respectively (<xref rid="B14" ref-type="bibr">14</xref>).</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><sec><title> <bold>Cell culture</bold> </title><p>Human dermal fibroblasts (HDF) were derived from adult skin by trypsinization as described (<xref rid="B8" ref-type="bibr">8</xref>, <xref rid="B9" ref-type="bibr">9</xref>). HDF obtained from outgrowth of explant cultures, were grown in DMEM (Dulbecco’s modified Eagle's medium; Bio Whittaker) supplemented with 10% fetal calf serum (FCS), and cells from passages 4 to 8 were used. Mouse embryonic fibroblasts (MEF) were isolated from mouse embryos (day 9.5; 19), immortalised according to the 3T3 protocol (<xref rid="B15" ref-type="bibr">15</xref>), and grown in DMEM (Bio Whittaker) supplemented with 10% FCS (<xref rid="B16" ref-type="bibr">16</xref>, <xref rid="B17" ref-type="bibr">17</xref>). </p><p> Normal epidermal keratinocytes (NEK) derived from adult skin (<xref rid="B8" ref-type="bibr">8</xref>, <xref rid="B9" ref-type="bibr">9</xref>) were plated on X-irradiated fibroblast feeder cells (HDF, 70 Gy; MEF, 20 Gy) in FAD medium (DMEM:Ham’s F-12 / 3:1) with 100 U/ml penicillin, 100 µg/ml streptomycin and supplemented with 5% FCS, 5 µg/ml insulin, 1 ng/ml recombinant human EGF, 10<sup>-10 </sup>M cholera toxin, 10<sup>-4 </sup>M adenine, and 0.4 µg/ml hydrocortisone (Sigma) as described (<xref rid="B8" ref-type="bibr">8</xref>). </p><p>Cells of the immortalized human skin keratinocyte cell line HaCaT (<xref rid="B11" ref-type="bibr">11</xref>) were subcultivated with a split ratio of 1:10 every 10 days in DMEM with 10% FCS.</p></sec><sec><title> <bold>Organotypic cocultures</bold> </title><p>Lyophilized Collagen type I was resolubilised at the desired concentration (2-4 mg/ml; dry weight) in 0.1% acetic acid and kept at 4°C. (Suppliers e.g. (a) Vitrogen-100, bovine dermal collagen (type I) (Collagen Corp., Palo Alto, California); (b) Type I collagen, calf skin (Cerard, Lyon, France); (c) Type I collagen, rat tail tendons (SIGMA); (d) Type I collagen, calf skin (IBFB, Leipzig, Germany). The ice cold collagen solution (80% of total volume, 4 mg/ml) was mixed with Hanks salt 10x with phenol red (10% of total volume) and adjusted to pH 7.4 by adding about 80 µl of 2 M NaOH per 12 ml collagen mixture while gently stirring on ice to avoid air bubbles and premature gelation. The fibroblast number necessary for the desired concentration in the gel (1x10<sup>5</sup>-5x10<sup>5</sup> per ml) was resuspended in FCS (10% of total volume) and added to the gel solution on ice under cautious stirring. With cooled pipettes aliquots of the collagen gel mixture were poured into filter inserts (2.5 ml/ filter insert; one insert per well of a 6-well plate) (Falcon cell culture inserts (3.0 µm pores) and Biocoat 6-fold Deep Well Plates from Becton Dickinson; see also <xref rid="B9" ref-type="bibr">9</xref>) (Fig. <xref rid="F1" ref-type="fig">1</xref>). For gelation, the collagen solution in filter inserts is incubated for 1 h at 37°C in a humified incubator. Thereafter, glass rings (inner diameter 18 mm, wall thickness 2 mm, 80 mm high) were placed on the gel and gently pushed down by mild pressure with forceps in order to confine the area for epithelial cell growth. The gels with the glass rings are placed for 1 h at 37°C in a humified incubator. The excess liquid pressed out of the gel was gently and carefully aspirated. Eventually, the gels were equilibrated by complete immersion in culture medium for 24 h.</p><fig id="F1"><label>Fig. 1</label><caption><title>Schematic illustration of the organotypic culture system.</title><p> Keratinocytes grow air exposed on a fibroblast containing collagen gel. The filter inserts are in contact with the culture medium.</p></caption><graphic xlink:href="bpo_v6_p55_m72f1lg"/></fig><p>Thereafter keratinocytes were plated inside the glass ring (after removing the medium) at a density of 1x10<sup>6</sup> in 1 ml FAD medium for each 2.5 cm insert. They attach within 12-24 h and form a nearly confluent layer on top of the collagen gel. 24-30 h after seeding the glass rings were removed, thereby avoiding any mechanical distortion of the epithelial cell sheet.</p><p>Culture medium was changed, and, by lowering the medium level to the lower part of the gels, the cultures were raised to the air-liquid interphase thus restricting nourishment to diffusion from below. This air-lift procedure is defined as the start of the culture time of organotypic cocultures. Organotypic cocultures with primary keratinocytes were grown in FAD-Medium or DMEM with 10% FCS and 50 µm ascorbic acid, those with HaCaT cells were supplemented with 2 ng/ml TGF-α. Medium was changed every 2-3 days. Depending on the purpose of the respective study, alternative media might be used, such as serum free formulations or media with specific supplementations such as inhibitors, neutralizing antibodies, growth factors, et cetera (described in detail in <xref rid="B1" ref-type="bibr">1</xref>-<xref rid="B2" ref-type="bibr">2</xref>, <xref rid="B8" ref-type="bibr">8</xref>-<xref rid="B9" ref-type="bibr">9</xref>).</p></sec><sec><title> <bold>Analysis</bold> </title><p>For histological or immunohistochemical analysis, organotypic cocultures were fixed in formaldehyde (3.7%) for at least 24h. Thereafter each culture was covered by a drop of hand-warm agar (2%) to prevent dislodgement of the epithelium during further preparation, and then the whole specimen was processed for paraffin embedding following protocols for routine histology. Alternatively, for cryosectioning specimens were embedded in Tissue Tek-OTC-compound and subsequently snap frozen in liquid nitrogen vapour.</p><p>The most important quality parameter of epidermal tissue reconstitution is its morphologic appearance. In addition , tissue maturation is readily evaluated by <italic>in situ</italic>-analytical techniques on frozen or fixed tissue sections determinig expression (<italic>in situ</italic>-hybridization) and distribution of specific epidermal differentiation products (immunohistochemistry) such as involucrin, keratin 1/10, transglutaminase, filaggrin, loricrin (<xref rid="B2" ref-type="bibr">2</xref>, <xref rid="B4" ref-type="bibr">4</xref>, <xref rid="B9" ref-type="bibr">9</xref>-<xref rid="B10" ref-type="bibr">10</xref>). Usually, the formation of a regular epidermal tissue architecture is paralleled by typical, in vivo-like expression patterns of differentiation products and the development of a stratum corneum.</p></sec></sec><sec><title>Results and Discussion</title><p>For generating organotypic cocultures, epidermal keratinocytes were plated onto the upper surface of collagen gels containing embedded fibroblasts, where they attached rapidly and formed confluent layers within 1-2 days. Subsequently, these keratinocytes reconstituted an epithelial tissue architecture resembling native epidermis and expressing characteristic epidermal differentiation markers (<xref rid="B18" ref-type="bibr">18</xref>-<xref rid="B19" ref-type="bibr">19</xref>, <xref rid="B1" ref-type="bibr">1</xref>, <xref rid="B9" ref-type="bibr">9</xref>). In the absence of fibroblasts, only thin epithelia developed with rapid loss of proliferation within 2 to 3 days (<xref rid="B20" ref-type="bibr">20</xref>). However, the addition of fibroblasts into the collagen matrix - either proliferative or postmitotic cells at a density of at least 2x10<sup>4</sup>/ml – caused a sustained keratinocyte growth accompanied by regular differentiation. It could be demonstrated by means of this <italic>in vitro</italic>-system, that the cocultured fibroblasts produce growth factors which are essential for epidermal morphogenesis (<xref rid="B8" ref-type="bibr">8</xref>, <xref rid="B20" ref-type="bibr">20</xref>). In general, using normal human epidermal keratinocytes (1x10<sup>6</sup> per culture, 3x10<sup>5</sup> per cm<sup>2</sup>) and human dermal fibroblasts (2.5-5x10<sup>5</sup> per culture, 1-2x10<sup>5</sup> per cm<sup>2</sup>), multilayered and differentiated epithelia had formed after 7-10 days comprising all characteristic epidermal layers including an orthokeratotic stratum corneum. Beside the dependency on fibroblasts, variations in epithelial quality can be observed due to the interindividual variability of primary epithelial cells (different donors, body sites, ages) and differences in the isolation procedures as well as culture conditions (<xref rid="B21" ref-type="bibr">21</xref>).</p><p>Depending on the aim of the studies different fibroblast numbers or genotypes have been used. For example, heterologous skin equivalents consisting of mouse embryonic fibroblasts (MEF) and human keratinocytes are an excellent tool to identify the source of paracrine factors contributing in dermal-epidermal interactions (<xref rid="B22" ref-type="bibr">22</xref>-<xref rid="B24" ref-type="bibr">24</xref>). In both, homologous as well as heterologous organotypic cocultures, a typical epidermal tissue is formed within one week suggesting that responsible factors are effective also beyond species barriers (Fig. <xref rid="F2" ref-type="fig">2</xref>a and b).</p><fig id="F2"><label>Fig. 2</label><caption><title>Epithelial architecture of organotypic cocultures of different epithelial and mesenchymal cell types.</title><p> Epidermal tissue morphology of 7 day organotypic cocultures of human keratinocytes with human dermal fibroblasts (a) and human keratinocytes with mouse embryonic wild type fibroblasts (b). HaCaT cells formed epithelia on human dermal fibroblasts (c) and mouse embryonic wild type fibroblasts (d). (H and E staining; same magnification; bar: 100 µm).</p></caption><graphic xlink:href="bpo_v6_p55_m72f2lg"/></fig><p>In coculture with mouse fibroblasts, growth of human keratinocytes is clearly delayed resulting after 7 days in thinner epithelia as compared to cocultures with human fibroblasts. This is indicative for a reduced efficacy of these cells in supporting the growth of human keratinocytes. Nevertheless, this can be compensated by a prolonged culture time of 10 to 14 days. However, heterologous skin equivalent models have the unique advantage, that genetically modified mouse cells can be exploited such as embryonic fibroblasts from knock-out or transgenic mice. In recent studies we have analyzed the distinct effects of different embryonic mouse fibroblast lines on epidermal morphogenesis of human keratinocytes, thus establishing a sensitive and efficient read out system for fibroblast derived mediators (<xref rid="B10" ref-type="bibr">10</xref>, <xref rid="B16" ref-type="bibr">16</xref>, <xref rid="B25" ref-type="bibr">25</xref>).</p><p>To further standardise the in vitro skin equivalent model and to avoid donor-dependent variations, the epidermal cell line HaCaT has been used instead of primary keratinocytes. The HaCaT cell line is a spontaneously transformed keratinocyte line (<xref rid="B11" ref-type="bibr">11</xref>) with sustained genetic alterations indicative of transformed but not tumorigenic cells (<xref rid="B26" ref-type="bibr">26</xref>). When grown on plastic the HaCaT cells show continuous proliferation, independence of feeder-cells, and a rather typical epithelial morphology with expression of a large panel of keratins (<xref rid="B12" ref-type="bibr">12</xref>-<xref rid="B13" ref-type="bibr">13</xref>, <xref rid="B26" ref-type="bibr">26</xref>). When transplanted onto the back of nude mice, HaCaT cells regenerate and differentiate to well structured epithelia expressing the characteristic epidermal markers (<xref rid="B12" ref-type="bibr">12</xref>). On the other hand, formation of multilayered epithelia by HaCaT cells in organotypic cocultures is delayed and require an increased number of fibroblasts (<xref rid="B13" ref-type="bibr">13</xref>). This was found to be due to certain deficiencies of HaCaT cells in the production of IL-1 as well as a low expression of the receptors for KGF and GM-CSF what can be compensated by application of TGF-α (for details see <xref rid="B14" ref-type="bibr">14</xref>). In organotypic cocultures of HaCaT cells either with human or mouse fibroblasts well structured and differentiated stratified epithelia developed, when supplemented with TGF-α , although the stratum corneum remained parakeratotic indicating incomplete keratinization (Fig. <xref rid="F2" ref-type="fig">2</xref>c and d). Clearly resembling normal human keratinocytes, cocultures of HaCaT cells with mouse mesenchymal cells show a diminished stratification within 7 days. In conclusion, their reproducible production and standardized quality renders HaCaT organotypic cocultures an excellent tool for large scale examinations of mechanisms regulating skin reepithelialization and homoeostasis in a tissue-type context.</p></sec>
Serum-Free Cryopreservation of Five Mammalian Cell Lines in Either a Pelleted or Suspended State	<p>Herein we have explored two practical aspects of cryopreserving cultured mammalian cells during routine laboratory maintenance. First, we have examined the possibility of using a serum-free, hence more affordable, cryopreservative. Using five mammalian lines (Crandell Feline Kidney, MCF7, A72, WI 38 and NB324K), we found that the serum-free alternative preserves nearly as efficiently as the serum-containing preservatives. Second, we compared cryostorage of those cells in suspended versus a pellet form using both aforementioned cryopreservatives. Under our conditions, cells were in general recovered equally well in a suspended versus a pellet form.</p>	<contrib contrib-type="author" corresp="yes"><name><surname>Corsini</surname><given-names>Joe</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Hacker</surname><given-names>Christy</given-names></name><xref rid="O1" ref-type="aff">1</xref></contrib><contrib contrib-type="author"><name><surname>Bare</surname><given-names>Charles</given-names></name><xref rid="O2" ref-type="aff">2</xref></contrib>	Biological Procedures Online	<sec><title>Introduction</title><p>The primary causes of cell death during cryopreservation appear to be formation of both ice crystals and osmotic gradients across the cell membrane (<xref rid="B1" ref-type="bibr">1</xref>-<xref rid="B5" ref-type="bibr">5</xref>). As such, cryopreservatives have been employed to retain viability during the freezing and thawing processes. Although numerous agents have been observed to promote cell viability during the freezing process (<xref rid="B6" ref-type="bibr">6</xref>), historically, dimethyl sulfoxide (DMSO) and glycerol have been used as preservatives for storing cells in a frozen state (<xref rid="B1" ref-type="bibr">1</xref>-<xref rid="B2" ref-type="bibr">2</xref>, <xref rid="B4" ref-type="bibr">4</xref>, <xref rid="B6" ref-type="bibr">6</xref>). During routine maintenance in many cell biology, virology, and molecular biology laboratories, mammalian cell lines are commonly stored in fetal bovine serum (FBS) plus 10% DMSO or growth medium containing serum plus 10%DMSO. Herein we have explored two practical aspects of cryopreserving cultured mammalian cells during routine laboratory maintenance. First, we have examined the possibility of using a serum-free, hence more affordable, cryopreservative. Lim <italic>et al</italic>. (<xref rid="B7" ref-type="bibr">7</xref>) have preserved bovine oocytes for 2-3 weeks in a serum-free PBS/DMSO cryopreservative, and Hubel <italic>et al</italic>. have employed a PBS/DMSO cryopreservative for short term cryo-studies of B lymphoblasts (<xref rid="B8" ref-type="bibr">8</xref>). We hypothesized that phosphate buffered saline (PBS) containing 10% DMSO would effectively preserve a variety of mammalian cell lines. To test this, we compared recovery of five cell lines after 30 days storage in either FBS+10%DMSO, RPMI1640+10%DMSO, or PBS+10%DMSO.</p><p>Second, we compared cryostorage of those same cells in suspended versus a pellet form in both aforementioned cryopreservatives. It has been reported that cell to cell contact influences survival during cryo-storage (<xref rid="B9" ref-type="bibr">9</xref>-<xref rid="B11" ref-type="bibr">11</xref>). In particular, membrane integrity of monolayered Chinese Hamster fibroblasts cells was more resistant to intracellular ice formation than that of non-monolayered controls (<xref rid="B9" ref-type="bibr">9</xref>, <xref rid="B10" ref-type="bibr">10</xref>). While we do not expect the cells packed in a pellet to be gap- or tight-junctioned, we predicted that the packing of cells into a pellet would influence both the rate and extent of of intracellular ice formation, promoting cell survival. In addition, we also predicted that that the damage caused by ion partitioning (occurring during formation of the H<sub>2</sub>O crystal lattice) would be mitigated if cells were tightly packed in a pellet that excluded buffer. To test these predictions, we compared five mammalian cell lines, cryopreserved as either a pellet or a suspension, in each of the three cryopreservatives.</p></sec><sec sec-type="materials\|methods"><title>Materials and Methods</title><p>Figure <xref rid="F1" ref-type="fig">1</xref> depicts the basic experiment carried out on each of the cell lines. This experiment was repeated three times for each cell line. In conducting this basic experiment, NB324K, A72, Crandell Feline kidney, MCF-7, and WI 38 cells were grown to 70%-90% confluency in 25cm<sup>2 </sup>tissue culture dishes in RPMI1640 (SIGMA, Sigma – St. Louis, MO, USA) containing 10% FBS (SIGMA, Sigma – St. Louis, MO, USA), in the absence of antibiotics. They were detached with trypsin/EDTA (SIGMA, St. Louis, MO, USA), the trypsin solution removed via centrifugation in a 15ml polypropylene conical tube (VWR International, West Chester, PA, USA), and then resuspended in 1 ml of fetal bovine serum (FBS), phosphate buffered saline (PBS), or RPMI1640 containing 10% dimethyl-sulfoxide (DMSO-SIGMA, St. Louis, MO, USA). Cells cryo-stored in a suspended state were immediately placed at -80ºC in a foam freezer box (18x18x9cm, with wall thickness of 5cm) packed with paper towels. Those cryo-stored in a pellet state were first spun for 5 minutes at 200xg. Cells were thawed 30-90 days later by placing the 15ml polypropylene conical tubes in a 37ºC water bath, and then spun at 200xg for 5 minutes. Cryopreservative was drained and residual solution removed with sterile transfer pipet and cells resuspended in 4mls fresh growth medium (above) followed by seeding into 25cm<sup>2</sup> flasks.</p><fig id="F1"><label>Fig. 1</label><caption><title>Cells were grown to 50-100% confluency in a 25cm<sup>2</sup> flask.</title><p> They were next trypsinized, and upon detachment, 3ml of growth medium were added. Cells were then split equally into three 15ml conical polycentrifuge tubes, followed by centrifugation and removal of trypsin solution. Each pellet was resuspended in 1 ml of the idicated cryopreservative (each with 10%DMSO: FBS=fetal bovine serum, PBS=phosphate-buffered saline, and RPMI = RPMI1640 with bicarbonate). One half of the cells from each treatment were removed to a new tube and placed in centrifuge to generate a pellet. Pelleted and suspended cells were then inserted into a foam box containing paper towels and placed in a -80°C freezer.</p></caption><graphic xlink:href="bpo_v6_p61_m73f1lg"/></fig><p>Cell viability was measured using attachment and subsequent ‘spreading’ as a readout. Live cells were easily identified by the conversion from a spherical morphology in suspension to a ‘spread’ morphology characteristic for each individual cell line. Validation of this scoring method was carried out by monitoring cells post attachment. In all treatments for all cell lines, attached and spread cells proved viable by dividing themselves. This was ascertained by monitoring doubling time over several days. A72 and WI 38 were observed to double in confluency (substratum coverage) after 1-2 days, and, being at lower densities, attached NB324K, CRFK, and MCF-7 cells were located within hand-marked circular fields and observed to double within 1-3 days. This scoring method was used instead of dye exclusion because of the possibility that, during freezing and thawing, cells might sustain lethal damage to internal membranes but not external membranes; such cells would exclude dye and be scored as viable when in fact they were dead. Within 18-24 hours after seeding, flasks were rocked briefly to resuspend dead cells. Relative cell recovery (viability) for each treatment was then estimated by determining the average number (5 replicate windows) of attached and spread cells in a window at 100x.</p><p>We evaluated the resulting data with two approaches. Firstly, we compared the relative performance of PBS and RPMI treatment to that of the FBS treatment by calculating, for each replication, the ratio of recovered cells in either the PBS or the RPMI1640 treatment to that in the FBS treatment. The estimator of the population ratio r is</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p61_eqn01cor.jpg"/> </inline-formula> </p><p>with the variance of the ration estimator being calculated by</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p61_eqn02cor.jpg"/> </inline-formula> </p><p>The individual ratios for each rep were then averaged to obtain a mean ratio for the three reps, and the accompanying pooled standard error, S<sub>p</sub>, was calculated for three replications and five observations per replication</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p61_eqn03cor.jpg"/> </inline-formula> </p><p>and pooled standard error</p><p> <inline-formula> <inline-graphic xlink:href="bpo_v6_p61_eqn04cor.jpg"/> </inline-formula> </p><p>These ratios were calculated for cells cryo-stored in both the suspended and the pellet state, and the two states compared using a t-test. These results are presented in Table <xref rid="T1" ref-type="table">1</xref>. Secondly, for each cell line we calculated the weighted mean and pooled standard deviation of cells recovered from each of the three repetitions for each cryopreservative, in either the suspended or the pellet state. For each cell line we then compared recovery in the three cryopreservatives using analysis of variance (ANOVA). If ANOVA indicated a significant difference in means, Fisher’s least significant difference (LSD) statistic was calculated. LSD is a comparison criterion used to determine where differences in means occur. We also compared recovery in each cryopreservative in the suspended state to that in the pellet state using the standard t-test. These results are summarized in Table <xref rid="T2" ref-type="table">2</xref>.</p></sec><sec><title>Results</title><p> <bold>Crandell Feline Kidney</bold> (CRFK) cells were isolated from the kidney of a domestic cat (<xref rid="B12" ref-type="bibr">12</xref>-<xref rid="B14" ref-type="bibr">14</xref>). </p><p> <underline>Suspended</underline>: The mean ratio of suspended cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. According to the ratios, cells in the suspended state the PBS preservatives appear to perform somewhat less efficiently than the RPMI and the FBS treatments. However, the F statistic generated from the mean number of cells recovered from each treatment suggests that these differences are not likely to be real (Table <xref rid="T2" ref-type="table">2</xref>).</p><p> <underline>Pellet</underline>: The mean ratio of pelleted cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. Both serum-free media perform somewhat less efficiently than the FBS preservative, with the RPMI1640 cryopreservative supporting recovery at 0.5x that of the FBS preservative in the pellet state. The F statistic generated from the means of each cryopreservative in the pellet state suggests that these differences in ratios can be attributed to a large sampling variance rather than real differences between the preservatives (Table <xref rid="T2" ref-type="table">2</xref>).</p><p> <underline>Suspended vs. Pellet</underline>: A t-test comparing the ratios of RPMI/FBS in pellet vs. suspension indicates that there is significant difference between relative recovery of RPMI-preserved cells stored a suspended or a pellet state (pellet state were recovered at 0.65x relative to those in suspended state). While this suggests that CRFK cells were recovered less efficiently from the pellet state if stored in RPMI, t-tests on the mean cell number recovered from each state suggests that there is no significant difference between recovery of cells from the RPMI preservative in either state. The t-tests on mean cell number also suggest that CRFK were more effectively (1.8-fold) recovered from the PBS cryopreservative when stored in a pellet form (Table <xref rid="T2" ref-type="table">2</xref>).</p><p> <bold>MCF7</bold> (<xref rid="B15" ref-type="bibr">15</xref>) is a human mammary epithelial line derived from a metastatic site. </p><p> <underline>Suspended</underline>: The mean ratio of suspended cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. The observed serum-free to FBS ratios were all within 25% of 1, and the F statistic generated from the mean number of cells recovered from the three cryopreservatives shows a significant difference only between the PBS and the FBS cryopreservatives. The difference amounts to 0.58x cells recovered from recovered from PBS versus FBS cryopreservative. </p><p> <underline>Pellet</underline>: The mean ratio of pelleted cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. The ratios of cells recovered from serum-free to those from the FBS cryopreservative are within 25% of 1, with the RPMI preservative slightly outperforming the FBS preservative. However, the F statistic generated from the means of the cells recovered from the three cryopreservatives suggests that there is no significant difference between the cryopreservatives if MCF7 cells are stored in the pelleted state.</p><p> <underline>Suspended vs. Pellet</underline>: t-tests indicate that pelleting cells had no influence on the PBS/FBS or the RPMI/FBS ratios, suggesting that all of the cryopreservatives performed equally in either the suspended or pelleted state (Table <xref rid="T1" ref-type="table">1</xref>). t-tests comparing means number of cells recovered from each cryopreservative in each state also indicates that the suspension state (suspended or pelleted) does not influence recovery of MCF7 cells (Table <xref rid="T2" ref-type="table">2</xref>).</p><p> <bold>A72 </bold>cells were derived from a canine fibroma and have been utilized in studies of canine parvovirus infection (<xref rid="B13" ref-type="bibr">13</xref>-<xref rid="B14" ref-type="bibr">14</xref>, <xref rid="B16" ref-type="bibr">16</xref>-<xref rid="B17" ref-type="bibr">17</xref>).</p><p> <underline>Suspended</underline>: The mean ratio of suspended cells surviving in serum-free treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. These ratios indicate that both serum-free preservatives supported somewhat less recovery than the FBS preservative (both within 20% of 1), but ANOVA with the means of the cells recovered from the three cryopreservatives suggests that there is no significant difference between the cryopreservatives with cells stored in the suspended state.</p><p> <underline>Pellet</underline>: The mean ratio of pelleted cells surviving in serum-free treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. Both serum-free preservatives supported somewhat less recovery than the FBS preservative, and the LSD statistic generated from the means of the cells recovered from the three cryopreservatives suggests that recovery from the RPMI preservative promotes recovery of pelleted A72 cells at 0.76x that of FBS preservative. </p><p> <underline>Suspended vs. Pellet</underline>: A t-test suggests that pelleting cells had no influence on the PBS/FBS ratio (Table <xref rid="T1" ref-type="table">1</xref>), suggesting that the suspension state does not influence the relative recovery of cells in these two preservatives. A t-test also suggests that the RPMI/FBS recovery ratio was 0.8x in the pellet state vs. in the suspended state. t-tests comparing mean numbers of recovered cells for each each cryopreservative (in each state) suggest that the suspension state (suspended or pelleted) does not influence recovery of A72 cells stored in either PBS or FBS cryopreservatives, but that A72 in RPMI in the suspended state were recovered 1.47x more efficiently than those in pellet state (Table <xref rid="T2" ref-type="table">2</xref>). These results indicate efficient recovery of A72 in serum free cryopreservatives from either the pelleted or the suspended state.</p><p> <bold>NB324K</bold> is an SV/40 transformed human kidney cell line (<xref rid="B18" ref-type="bibr">18</xref>-<xref rid="B20" ref-type="bibr">20</xref>).</p><p> <underline>Suspended</underline>: The mean ratio of cells surviving in the two serum-free treatments, PBS or RPMI based, to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. Both serum-free cryopreservatives supported only half the recovery of the FBS preservative, but the LSD statistic on mean number of recovered cells suggests that only the difference between the PBS and FBS treatments is real (Table <xref rid="T2" ref-type="table">2</xref>). This difference is 0.60x cells recovered from the PBS versus the FBS treatment. </p><p> <underline>Pellet</underline>: The mean ratio of cells surviving in serum-free treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. Ratios indicate somewhat lower recovery from the serum-free preservatives with cells stored in a pelleted state. However, the LSD statistic from mean number of NB324K cells recovered from the three cryopreservatives indicate that the RPMI treatment likely supported a 1.3x higher rate of recovery than either the PBS or the FBS preservatives (Table <xref rid="T2" ref-type="table">2</xref>) when these cells were stored in a pellet state.</p><p> <underline>Suspended vs. Pellet</underline>: t-tests indicate that the ratio of recovered cells for PBS/FBS and RPMI/FBS is 1.3x higher with cells stored in the pellet state. Comparisons of the mean number of recovered cells (Table <xref rid="T2" ref-type="table">2</xref>) reveals that cells stored in PBS preservative were recovered with 1.4x greater efficiency in the pellet state. Those in the RPMI preservative were recovered with 1.5x greater efficiency in the pellet state. These results indicate that the serum-free preservatives were slightly more effective at preserving pelleted NB324K cells (versus suspended cells) in a frozen state.</p><p> <bold>WI 38 </bold>is a human diploid fibroblast line with a finite life of 50+/-10 divisions (<xref rid="B21" ref-type="bibr">21</xref>-<xref rid="B23" ref-type="bibr">23</xref>), that has been utilized in production of poliovirus vaccine (<xref rid="B22" ref-type="bibr">22</xref>). These cells arose from a primary explant of human fetal lung tissue, the chromosomes are considered normal, and the cells have undergone no known transformation events.</p><p> <underline>Suspended</underline>: The mean ratio of suspended cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. According to the ratios, both serum free preservatives appear to perform somewhat less efficiently than the FBS treatment with cells in the suspended state. The F statistic generated from the mean number of cells recovered from each treatment suggests that the PBS treatment truly did not support recovery as well as did the FBS or RPMI-based preservatives.</p><p> <underline>Pellet</underline>: The mean ratio of pelleted cells surviving in PBS or RPMI treatment to those surviving FBS treatment are shown in Table <xref rid="T1" ref-type="table">1</xref>. Both serum-free media perform approximately 0.60x less efficiently than the FBS preservative. The F statistic generated from the means of each cryopreservative in the pellet state suggests that these differences in ratios are real (Table <xref rid="T2" ref-type="table">2</xref>).</p><p> <underline>Suspended vs. Pellet</underline>: A t-test indicates that there is a significant difference in the ratio of cells recovered from serum-free versus serum containing after storage in a pellet versus a suspended state (Table <xref rid="T1" ref-type="table">1</xref>). Analysis of the mean number of recovered cells shows that for the FBS treatment, WI 38 cells stored in a pellet state were recovered at 0.67x relative to those in suspended state, and those from the RPMI cryopreservative in pellet state were recovered at 0.48x relative to those in suspended state (Table <xref rid="T2" ref-type="table">2</xref>).</p><table-wrap id="T1"><label>Table 1</label><caption><title>Comparison of cell recovery in phosphate-buffered saline or RPMI1640 based cryopreservative with recovery after storage in fetal bovine serum based cryopreservative.</title></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left" colspan="1"> <bold>Cell Line</bold> </td><td align="left" colspan="1"> <bold>PBS/FBS suspended</bold> </td><td align="left" colspan="1"> <bold>PBS/FBS pellet</bold> </td><td align="left" colspan="1"> <bold>t-test</bold> </td><td align="left" colspan="1"> <bold>RPMI/FBS suspended</bold> </td><td align="left" colspan="1"> <bold>RPMI/FBS pellet</bold> </td><td align="left" colspan="1"> <bold>t-test</bold> </td></tr><tr><td align="left" colspan="1">Crandell Feline Kidney </td><td align="left" colspan="1">0.62 s<sub>p</sub> =0.25 </td><td align="left" colspan="1">0.79 s<sub>p</sub> =0.35 </td><td align="left" colspan="1">t=1.53<break/>P=0.137 </td><td align="left" colspan="1">0.75 s<sub>p</sub> = 0.38 </td><td align="left" colspan="1">0.49 s<sub>p</sub> =0.21 </td><td align="left" colspan="1">t=2.32<break/>P=0.028 </td></tr><tr><td align="left" colspan="1">MCF7 </td><td align="left" colspan="1">0.84 s<sub>p</sub> =0.34 </td><td align="left" colspan="1">0.90 s<sub>p </sub>=0.38 </td><td align="left" colspan="1">t=0.46<break/>P=0.652 </td><td align="left" colspan="1">0.92 s<sub>p</sub> =0.29 </td><td align="left" colspan="1">1.25 s<sub>p</sub> =0.85 </td><td align="left" colspan="1">t=1.42<break/>P=0.166 </td></tr><tr><td align="left" colspan="1">A72 </td><td align="left" colspan="1">0.79 s<sub>p</sub> =0.15 </td><td align="left" colspan="1">0.76 s<sub>p</sub> =0.13 </td><td align="left" colspan="1">t=0.58<break/>P=0.563 </td><td align="left" colspan="1">0.87 s<sub>p</sub> =0.14 </td><td align="left" colspan="1">0.67 s<sub>p</sub> =0.09 </td><td align="left" colspan="1">t=4.65<break/>P=0.000 </td></tr><tr><td align="left" colspan="1">NB324K </td><td align="left" colspan="1">0.58 s<sub>p</sub> =0.12 </td><td align="left" colspan="1">0.75 s<sub>p</sub> =0.21 </td><td align="left" colspan="1">t=2.72<break/>P=0.011 * </td><td align="left" colspan="1">0.50 s<sub>p</sub> =0.07 </td><td align="left" colspan="1">0.64 s<sub>p</sub> =0.09 </td><td align="left" colspan="1">t=3.88<break/>P=0.001 </td></tr><tr><td align="left" colspan="1">WI 38 </td><td align="left" colspan="1">0.65 s<sub>p</sub> =0.10 </td><td align="left" colspan="1">0.62 s<sub>p</sub> =0.22 </td><td align="left" colspan="1">t=0.48<break/>P=0.634 </td><td align="left" colspan="1">0.87 s<sub>p</sub> =0.13 </td><td align="left" colspan="1">0.61 s<sub>p</sub> =0.18 </td><td align="left" colspan="1">t=4.54<break/>P=0.000 </td></tr><tr><td align="left" colspan="7">Storage was carried out in either a suspended or a pellet state. Shown is a <underline>weighted mean</underline> and <underline>pooled standard error</underline> of the ratio of recovered cells in PBS/10%DMSO or RPMI1640/10%DMSO treatment to the same in accompanying FBS/10%DMSO treatment. Recovery was scored by counting five windows at 100x and then averaging for each rep. The weighted mean was calculated based on data from three repetitions for all treatments with all lines except for the RPMI/FBS ratio of NB324K, which is the average of two repetitions. s<sub>p</sub> is the pooled standard error of the replications. The t-tests are two tailed. significant at the 0.05 level. significant at the 0.01 level.</td></tr></tbody></table></table-wrap><table-wrap id="T2"><label>Table 2</label><caption><title>The weighted mean of three reps for each treatment in each state.</title></caption><table frame="hsides" rules="groups"><tbody><tr><td align="left" colspan="1"> <bold>Treatment</bold> </td><td align="left" colspan="1"> <bold>Suspension</bold> </td><td align="left" colspan="1"> <bold>Pellet</bold> </td><td align="left" colspan="1"> <bold>t-test suspension vs pellet</bold> </td></tr><tr><td align="left" colspan="1">FBS/10%DMSO(1)</td><td align="left" colspan="1"> <underline>CRFK            66</underline> <bold>s<sub>P</sub> </bold> <underline>=38</underline> <break/> <underline>MCF7           12</underline> <bold>s<sub>P</sub> </bold> <underline>= 7</underline> <break/> <underline>A72              234</underline> <bold>s<sub>P</sub> </bold> <underline>=69</underline> <break/> <underline>NB324K        83</underline> <bold>s<sub>P</sub> </bold> <underline>=26</underline> <break/> <underline>WI 38            29</underline> <bold>s<sub>P</sub> </bold> <underline>=9</underline> </td><td align="left" colspan="1"> <underline>CRFK       89</underline> <bold>s<sub>P</sub> </bold> <underline>=54</underline> <break/> <underline>MCF7         8</underline> <bold>s<sub>P</sub> </bold> <underline>=  4</underline> <break/> <underline>A72         191</underline> <bold>s<sub>P</sub> </bold> <underline>=58</underline> <break/> <underline>NB324K   76</underline> <bold>s<sub>P</sub> </bold> <underline>=22</underline> <break/> <underline>WI 38        20</underline> <bold>s<sub>P</sub> </bold> <underline>= 4</underline> </td><td align="left" colspan="1">t=1.35   p=0.188<break/>t=1.92   p=0.065<break/>t=1.85   p=0.075<break/>t=0.79   p=0.433<break/>t=3.544  p=0.001</td></tr><tr><td align="left" colspan="1">PBS/10%DMSO(2) </td><td align="left" colspan="1"> <underline>CRFK            40</underline> <bold>s<sub>P</sub> </bold> <underline>=11</underline> <break/> <underline>MCF7              7</underline> <bold>s<sub>P</sub> </bold> <underline>=  2</underline> <break/> <underline>A72              190</underline> <bold>s<sub>P</sub> </bold> <underline>=49</underline> <break/> <underline>NB324K       50</underline> <bold>s<sub>P</sub> </bold> <underline>=31</underline> <break/> <underline>WI 38            16</underline> <bold>s<sub>P</sub> </bold> <underline>= 5</underline> </td><td align="left" colspan="1"> <underline>CRFK      75</underline> <bold>s<sub>P</sub> </bold> <underline>=16 </underline> <break/> <underline>MCF7        8</underline> <bold>s<sub>P</sub> </bold> <underline>=  3</underline> <break/> <underline>A72        160</underline> <bold>s<sub>P</sub> </bold> <underline>=37</underline> <break/> <underline>NB324K   72</underline> <bold>s<sub>P</sub> </bold> <underline>=26</underline> <break/> <underline>WI 38        12</underline> <bold>s<sub>P</sub> </bold> <underline>= 8</underline> </td><td align="left" colspan="1">t=6.98   p=0.000<break/>t=1.07   p=0.292<break/>t=1.89   p=0.069<break/>t=2.10   p=0.044<break/>t=1.64   p=0.112</td></tr><tr><td align="left" colspan="1">RPMI1640/10%DMSO(3) </td><td align="left" colspan="1"> <underline>CRFK           69</underline> <bold>s<sub>P</sub> </bold> <underline>=73</underline> <break/> <underline>MCF7           10</underline> <bold>s<sub>P</sub> </bold> <underline>=  4</underline> <break/> <underline>A72              213</underline> <bold>s<sub>P</sub> </bold> <underline>=31</underline> <break/> <underline>NB324K        65</underline> <bold>s<sub>P</sub> </bold> <underline>=14</underline> <break/> <underline>WI 38            25</underline> <bold>s<sub>P</sub> </bold> <underline>= 6</underline> </td><td align="left" colspan="1"> <underline>CRFK       70</underline> <bold>s<sub>P</sub> </bold> <underline>=55</underline> <break/> <underline>MCF7       10</underline> <bold>s<sub>P</sub> </bold> <underline>=  5</underline> <break/> <underline>A72        145</underline> <bold>s<sub>P</sub> </bold> <underline>=29</underline> <break/> <underline>NB324K    95</underline> <bold>s<sub>P</sub> </bold> <underline>=19</underline> <break/> <underline>WI 38        12</underline> <bold>s<sub>P</sub> </bold> <underline>= 6</underline> </td><td align="left" colspan="1">t=0.00  p=0.967<break/>t=0.00  p=1.000<break/>t=6.20  p=0.000 <break/>t=4.92  p=0.000 <break/>t=5.93  p=0.000 ** </td></tr><tr><td align="left" colspan="1"> <bold>ANOVA FBS vs PBS vs RPMI</bold> </td><td align="left" colspan="1">CRFK   F=1.66, P=0.202<break/>MCF7   F=4.13, P=0.023, LSD=3.5<break/>A72    F=2.68, P=0.080<break/>NB324K F=6.70, P=0.003, LSD=18.2<break/>WI 38  F=14.05, P=0.000, LSD=5.07* </td><td align="left" colspan="1">F=0.7,  P=0.5<break/>F=1.2,  P=0.3<break/>F=4.4,  P=0.02, LSD=31.8<break/>F=4.5,  P=0.02, LSD=16.6 <break/>F=8.28, P=0.001, LSD=4.58*</td><td align="left" colspan="1"> </td></tr><tr><td align="left" colspan="4">The mean is followed by <bold>s<sub>P</sub> </bold>, the pooled standard deviation of the three reps. The t-tests are two tailed with 28 degrees of freedom. Analysis of Variance: The critical F value = 3.22 for α =0.05 and 2, 42 degrees of freedom. LSD=Fisher’s least significant difference for α =0.05. The critical F value = 5.15 for α = 0.01 and 2, 42 degrees of freedom.significant at the 0.05 level. **significant at the 0.01 level.</td></tr></tbody></table></table-wrap></sec><sec><title>Discussion</title><p>Based upon the PBS/FBS and RPMI/FBS ratios, it appears that in most cases survival in the serum-free preservatives is somewhat lower than that in serum-containing conditions. The largest lowering observed was a halving of the recovery rate, which occurred in two cases: NB324K in RPMI1640 preservative in the suspended state and CRFK in RPMI preservative in the pellet state (see Table <xref rid="T2" ref-type="table">2</xref>). In some cases, the serum-free preservatives performed somewhat better than the serum-containing counterpart. In general, we feel that the differences we observed are of no practical consequence when cryostoring cells during routine maintenance because typically millions of cells are frozen during these procedures. In laboratories that store large volumes of cultured cells, this ability to eliminate expensive serum from cryopreserving solutions may represent a significant cash savings. Collectively, the data also indicate that there are not large differences in recovery rate of cells stored in pellet versus suspended form. This suggests that packing the cells together did not influence damage due to ice crystal formation or osmotic stresses. We note, however, we were working with fairly low numbers of cells that produced pellets on the order of 0.1-0.5 mm<sup>3</sup>, and it is possible that the above predictions would not manifest themselves until the packing volume reached an as yet undetermined threshold.</p></sec>