Abstracts for the XIIth World Congress of Psychiatric

0 downloads 0 Views 2MB Size Report
Hennessy RJ,1 Lane A,2 Kinsella A,1 Larkin C,2 O'Callaghan E,2 ...... factor analysis of the two temperament and character scales together based on the item ...... Curran SR2, Roberts S,3 Thomas S,3 Veltman M,7 Browne J,4 Medda E,5.
Abstracts for the XIIth World Congress of Psychiatric Genetics Burlington Hotel, Dublin, Ireland 9th–13th October 2004 Sponsored by the International Society of Psychiatric Genetics

Program Committee Chairperson: Michael Gill

Abstracts Organized and Edited by: Michael Gill, Eimeir Burton, Derek Morris, Katie Armstrong, Aiden Corvin, Martin Schalling, Nicholas Barden, Mike Owen, Nick Craddock, Lynn DeLisi, Douglas Blackwood, Marcus Noethen, and Kieran Murphy

ß 2004 Wiley-Liss, Inc.

2

Abstracts

Outline Programme Saturday October 9th 08:00–18:00 12:00–15:00 16:00–17:00 17:00–18:00 10:00–12:00 10:00–12:00 12:00–14:00 14:00–16:00 14:00–17:00 19:00

Registration ISPG Board Meeting Programme Committee Meeting WCPG 2005 Programme Committee Meeting WCPG 2006 Educational Workshop E1 Part 1, Overview of the Genetics of Neuropsychiatric Disorders Educational Workshop E2, Epidemiology for the Neurosciences. Lunch Break Educational Workshop E1 Part 2, Overview of the Genetics of Neuropsychiatric Disorders Educational Workshop E3, Pharmacogenetics. Welcome Reception, Burlington Hotel

Sunday October 10th 08:00–18:00 10:00–18:00 08:30–09:30 09:30–10:00 10:00–12:00 10:00–12:00 12:00–13:30 13:30–15:30 15:30–16:00 16:00–18:00

19:00

Registration Exhibitions Plenary Session PL1, Insights into the control of body weight from human and mouse genetics Coffee Break Symposium S1, Finding Genes for Dementia Symposium S2, Association Mapping in Complex Phenotypes Lunch Break Slide Sessions 01, Affective Disorders; 02, Dissecting the Phenotype; 03, Neurodegenerative Disorders Coffee Break Poster Sessions P1, Affective Disorders; P2, Bipolar Disorder; P3, Neuropsychological Traits/ Endophenotypes; P4, Neurodegeneration/Ethical Issues; P5, Anxiety/Eating Disorders; P6, Cytogenetics Studies State Reception, Dublin Castle

Monday October 11th 08:00–18:00 10:00–18:00 08:30–09:30 09:30–10:00 10:00–12:00

Registration Exhibitions Plenary Session PL2, Interaction between measured genes and measured environments: A research strategy Coffee Break Symposia S3, Genes and the Neurodevelopmental Origins of Schizophrenia; S4, Industrial-Academic

12:00–13:30 13:30–15:30 15:30–16:00 16:00–18:00

18:00–19:00

Collaborations in Molecular Genetics: Lessons from Past Experience and Future Needs Lunch Break Oral Sessions 04, Schizophrenia; 05, Childhood Onset Disorders; O6, Addiction and Behaviour Coffee Break Poster Sessions P7, Schizophrenia (I); P8, Schizophrenia (II); P9, Childhood Disorders (I); P10, Childhood Disorders (II); P11, Alcohol/Substance abuse disorders; P12, Behavioural Genetics ISPG General Assembly

Tuesday October 12th 08:00–18:00 10:00–18:00 08:30–09:30 09:30–10:00 10:00–12:00 12:00–13:30 13:30–15:30 15:30–16:00 16:00–18:00

19:30

Registration Exhibitions Plenary Session PL3, From Genes To Drugs For Cognitive Dysfunction Coffee Break Symposia S5, Statistical Methods for Complex Traits; S6, Genetic Basis for Sex Differences in Behaviour. The Role of Sex Chromosomes Lunch Break Oral Sessions 07, Neurobiology; O8, The Mood Psychosis Interface; 09, Bioinformatics and Molecular Methods Coffee Break Poster Sessions P13, Mood-Psychosis Interface; P14, Schizophrenia (II); P15, Gene Expression Studies; P16, Neurobiology/Animal Studies; P17, Bioinformatics/Population Genetics; P18, Pharmacogenetics Congress Party, Guinness Storehouse

Wednesday October 13th 08:00–10:00 08:30–08:45 08:45–09:45 09:45–10:15 10:15–12:15

12:30–14:00

Registration Closing Ceremony Plenary Session PL4, Psychiatric Genetics: From Early Achievements to Historical Burden, from an Anxious Society to Critical Geneticists Coffee Break Symposia S7, Developmental Biology and Molecular Genetics of Craniofacial Dysmorphogenesis in Psychosis; S8, Gene-Environment Coaction and Interaction. Farewell Reception, Burlington Hotel

Abstracts PL1 INSIGHTS INTO THE CONTROL OF BODY WEIGHT FROM HUMAN AND MOUSE GENETICS O’Rahilly S University of Cambridge, Cambridge, United Kingdom Considerable public attention has been drawn to the increasing prevalence of obesity in most Western and developing countries and to the relative role of increased food intake and decreased physical activity in producing this trend. While the recent rise in average fatness of human populations is notable, it has tended to obscure the fact that adiposity (the amount of fat stored) is one of the most heritable of human traits. Inherited factors predisposing to obesity have usually been thought of in terms of familial variation in metabolic rate and efficiency. In our recent studies of humans with severe obesity resulting from mutations in single genes we have shown that, in all cases, those genes are, in fact intimately involved with the mediation of appetite and satiety in the brain. Thus, these affected individuals are obese, not because of a defective metabolic rate, but because they eat too much. However, they eat too much because of faulty biochemical machinery in the brain centres controlling appetite. It is very likely that this concept is applicable, to some degree, to a broader range of people with obesity. This work provides some of the clearest examples to date of the molecular determinants of a specific human behaviour. PL2 INTERACTION BETWEEN MEASURED GENES AND MEASURED ENVIRONMENTS: A RESEARCH STRATEGY Moffit T1 and Caspi A2 1 Institute of Psychiatry, King’s College, London, United Kingdom 2 University of Wisconsin, Madison, United States Interactions between genes and environments (GxE) were thought to be rare in psychiatry, but empirical findings of measured GxE are emerging. To encourage GxE studies, we describe strategic elements of seven steps for testing GxE. Step 1, consulting quantitative behavioralgenetic research on a disorder to ascertain the likelihood of GxE. Step 2, identifying a candidate environmental pathogen. Step 3, optimizing environmental risk measurement. Step 4, selecting a good candidate gene. Step 5, testing the hypothesized GxE interaction. Step 6, testing if the finding extends beyond the initial triad of gene, environmental pathogen, and disorder. Step 7, replication. We explain potential benefits of the measured GxE approach for basic neuroscientists, for interventionists, and for gene hunters. We argue that, in psychiatric genetics, ignoring nurture handicaps the field’s capacity to make new discoveries about nature.

PL3 FROM GENES TO DRUGS FOR COGNITIVE DYSFUNCTION Tully T St. Giles Foundation Professor of Neuroscience, Cold Spring Harbor Laboratory, United States, Acting Chief Scientific Officer, Helicon Therapeutics, Inc., United States Behavioral and biochemical properties of simple forms of (Pavlovian) learning in Drosophila show properties conserved across the animal kingdom. Given this observation, we designed automated ‘‘Robotrainers’’ and screened 6700 randomly generated mutants for defects in long-term memory of an odor-shock association. From this ‘‘forward’’genetic strategy, we have identified 60 new memory mutants. Using a ‘‘reverse’’-genetic strategy, we also have begun to identify genes that are transcriptionally regulated during long-term memory formation— using DNA microarray technology. We have developed a novel statistical approach to analyze DNA chip data, yielding more than 3900 candidate memory genes (CMGs) genome-wide. This method has been validated by establishing significant overlap between CMGs from the DNA chip and molecular lesions associated with the memory mutants. These results suggest new biological pathways involved in memory formation, one of which has been biologically validated in vivo. Given the well-established molecular homology in neuronal function among invertebrates and vertebrates, such gene discovery leads to drug discovery ultimately to develop effective therapies for human cognitive dysfunction. A high-throughput cell-based screen for modulators of the CREB pathway has yielded several novel drug compounds. The most advanced of these, a phosphodiesterase inhibitor, has been shown in animal models to facilitate several forms of memory loss, including a genetic model of mental retardation.

3

PL4 PSYCHIATRIC GENETICS: FROM EARLY ACHIEVEMENTS TO HISTORICAL BURDEN, FROM AN ANXIOUS SOCIETY TO CRITICAL GENETICISTS Propping P Institute of Human Genetics, University of Bonn, Bonn, Germany When the study of human inheritance became a scientific topic more than 100 years ago, the dimension of brain function and mental disease immediately attracted the particular interest of researchers. From its early roots psychiatric genetics was dominated by the question of nature and nurture. Today this question can be pinned down to odds ratio and attributable risk of a certain genotype. The false doctrine of eugenics and its practical application in the Nazi system paved the way for a constant anxiety of society that psychiatric genetics might lead to stigmatisation or even a revitalization of eugenics. The major challenge of the field, however, comes from fellow geneticists who doubt that genetics can ever contribute to understanding of brain function and mental disease. Whereas complex traits in other fields of medicine are being successfully pinned down to the molecular level, psychiatric genetics still awaits a major breakthrough. It has to be analysed why mental disorders are obviously harder to tackle. Is it only the difficulty of the field? E1.1 THE GENETICS OF SCHIZOPHRENIA Owen MJ Cardiff University, Cardiff, United Kingdom The high heritability of schizophrenia has stimulated much work aimed at identifying susceptibility genes using positional genetics. However, difficulties obtaining clear replicated linkages has led to scepticism that such approaches would ever be successful. Fortunately, there are now signs of real progress. Several strong and wellestablished linkages have emerged. Four of the best-supported regions are 6p24-22, 1q21-22, 13q32-34 and 10q25.3-26.3. In these cases single studies achieved genome-wide significance at P < 0.05 and suggestive positive findings have also been reported in other samples. Other promising regions include 8p21-22, 6q21-25, 22q11-12, 5q21-q33, 10p15-p11 and 1q42. The study of chromosomal abnormalities in schizophrenia has also added to the evidence for susceptibility loci at 22q11 and 1q42. Recently, evidence implicating individual genes within some of the linked regions has been reported and more importantly replicated. The weight of evidence now favours NRG1 and DTNBP1 as susceptibility loci, though work remains before we understand precisely how genetic variation at each locus confers susceptibility and protection. The evidence for RGS4 and G72 is promising but not yet persuasive. While further replications remain the top priority, the respective contributions of each gene, relationships with aspects of the phenotype, the possibility of epistatic interactions between genes and functional interactions between the gene products will all need investigation. The ability of positional genetics to implicate novel genes and pathways will open up new vistas for neurobiological research, and all the signs are that it is now poised to deliver crucial insights into the nature of schizophrenia.

E1.2 REVIEW OF THE GENETICS OF THE AFFECTIVE DISORDERS Craddock N Cardiff University, Cardiff, United Kingdom A robust body of evidence from classical family, twin and adoption studies demonstrates the importance of genes in the pathogenesis of both bipolar disorder (BPD) and unipolar major depression (UP). Occasional families may exist in which a single gene plays a major role in determining susceptibility, but the majority of affective disorder involves more complex genetic mechanisms such as the interaction of multiple genes and environmental factors. Recent advances in molecular genetics have made it possible to identify these susceptibility genes and interesting findings are beginning to emerge—although the pace of advance has not been as rapid as in schizophrenia, probably, at least in part, because fewer studies have been undertaken to date in BPD—and far fewer in UP. Molecular genetic positional and candidate gene approaches are being used for the genetic dissection of BPD. Several genomic regions of interest have been identified in linkage studies although the pattern of results is complex and the 2 metaanalyses conducted to date have not shown consistent findings.

4

Abstracts

Candidate gene association studies using both case-control and familybased paradigms are ongoing. Findings with ‘‘traditional’’ candidates have been inconsistent but some, including the genes encoding COMT, MAOA and the serotonin transporter, find some support from metaanalyses for the existence of modest influences upon illness susceptibility but await robust scrutiny in large samples. Recent reports arising from positional and candidate studies implicate several genes, including BDNF, G72/G30, XBP1 and GRK3. These are interesting and await robust replication. There is increasing use of sub-phenotypes of BPD in an attempt to refine the process of genetic dissection. Examples include Rapid Cycling, puerperal triggering of episodes, lithium response and presence of psychosis. The latter has particular relevance to exploring the overlap in genetic susceptibility across the traditional Kraeplinian dichotomy. In this regard it is important to explore the extent to which genes implicated in the pathogenesis of schizophrenia (e.g. Dysbindin, neuregulin, DAAO) contribute also to the pathogenesis of BPD. The first genome scan of UP have only recently been published but findings are encouraging with some regions of overlap and reports of locus refinement. This presentation will give a critical overview of the current state of knowledge and the directions in which the field is moving. E1.3 THE GENETICS OF ALZHEIMER’S DISEASE AND OTHER DEMENTIAS Hardy J Laboratory of Neurogenetics, National Institute on Aging, National Institutes of Health, Bethesda, USA All autosomal dominant Alzheimer’s disease is caused by mutations in the amyloid precursor protein gene (APP) or in the presenilin 1 and 2 (PS1 and PS2) genes. All of these mutations have in common the fact that they alter APP processing such that more Ab42 is produced from APP. These genetic data have led to the formulation for the amyloid cascade hypothesis and have allowed the production of mice which show prominent Ab42 deposition, but none of the downstream pathologies of Alzheimer’s disease. Mutations in the tau gene lead to the syndrome, frontal temporal dementia with parkinsonism linked to chromosome 17. This complex entity is characterized by a dementing syndrome with tau pathology: sometimes tangles, sometimes Pick bodies and sometimes discrete pathology only visible with immunocytochemistry. Mice with tau mutations driven by the prion promoter develop tangles and cell loss, but only in the midbrain and spinal cord, presumably relating to the expression patterns of the promoter. However, mice with both amyloid and tau transgenes develop increased tangle pathology in cortical regions demonstrating directly a link between Ab and tangles, and thus offering experimental proof of the amyloid cascade hypothesis and an experimental system in which therapies against many aspects of the cascade can be tested. E2.1 THE DESIGN OF GENETIC EPIDEMIOLOGY STUDIES Merikangas K Chief, Senior Investigator, Section on Developmental Genetic Epidemiology, National Institute of Mental Health, Maryland, USA This workshop will describe the role of the designs and analytic methods of the field of epidemiology in gaining understanding of the role of genes in complex diseases. Sources of complexity in mental disorders will be described and exemplary models from other complex diseases will be used to illustrate methods for reducing such sources of complexity. Detailed sampling methods and population-based strategies in genetic epidemiology will be presented. A range of analytic methods for complex disorders will be presented. Criteria that may be used to establish priorities for genomics research and steps for translation of advances in molecular genetics to public health will be discussed.

E2.2 STATISTICAL GENETICS STRATEGIES FOR COMPLEX DISEASES McGinnis RE GlaxoSmithKline, Harlow, United Kingdom Unlike Mendelian diseases which are caused by a single defective (i.e. mutated) gene, complex genetic diseases such as diabetes and schizophrenia appear to be caused by combinations of genes that jointly

increase susceptibility to disease. In addition, environmental factors sometimes also increase susceptibility to complex diseases. In this session, I will describe basic statistical approaches for localizing and identifying genes that cause complex diseases. I will cover study designs in which subjects are from families (with or without DNA from some or all parents) or in which some or all subjects are unrelated individuals (disease cases and unrelated controls). Relative statistical power, advantages and properties of these designs will be examined as a function of underlying genetic and clinical parameters that affect their performance. This overview will also use key examples of genes now known to cause complex disease to illustrate use of the strategies for identifying disease-causing genes, the role of environmental factors, and some of the challenges and complexities that pertain to successful use of the methods surveyed. E3.1 PHARMACOGENETICS OF ANTIDEPRESSANTS EFFICACY Serretti A Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy Pharmacogenetic studies in mood disorders mainly involved short term antidepressant response. Antidepressant drugs are the first line treatment for major depression but the therapeutic response in clinical practice is expected in about two thirds of patients. The large interindividual variability in the pharmacological response pattern has been partially ascribed to heritable factors. During the recent years the possible influence of a set of candidate genes as possible genetic predictors of antidepressant response efficacy were investigated. The functional polymorphism in the upstream regulatory region of the serotonin transporter gene (5-HTTLPR), the A218C gene variant on the tryptophan hydroxylase gene (TPH), the 102TC variant in the 5HT2A receptor, the G-protein beta3-subunit (Gbeta3) C825T gene variant and the Circadian Locomotor Output Cycles Kaput (CLOCK) gene variants were independently associated with short term SSRIs antidepressant efficacy. The effects of 5-HTTLPR and TPH polymorphisms were more pronounced in subjects not taking pindolol. Marginal associations were reported for ADRB1, ACE I/D, BDNF and IL-1beta. DRD2, DRD4, Mao-A, SERT-STin2, 5HT6, NOS gene variants were not associated with outcome. Although in its preliminary phase, the results obtained in the pharmacogenetics of antidepressants is promissing for an individualized therapy.

E3.2 NEW DEVELOPMENTS IN THE PHARMACOGENETICS OF ANTIPSYCHOTIC DRUG RESPONSE Malhotra AK The Zucker Hillside Hospital, Glen Oaks, United States Pharmacogenetic approaches to antipsychotic drug response offer the prospect of the identification of biological predictors of clinical efficacy and drug-induced adverse events, as well as the potential to detect the molecular substrates of antipsychotic drug action. Initial data in this regard, however, have been inconsistent, perhaps secondary to relatively small sample sizes, the use of retrospective, non-specific assessments of outcome, and the problem of ethnic stratification within study groups. Moreover, these initial studies have only included a limited number of single nucleotide polymorphisms (SNPs) primarily due to the paucity of data on genetic variation and limitations in genotyping capabilities. More recent studies however, have begun to suggest that pharmacogenetic studies of antipsychotic drug response may provide informative data. Several groups have identified a relationship between genetic variation in the dopamine D2 receptor gene (DRD2) and clinical response to treatment, and follow-up studies are now underway to comprehensively assess DRD2 in atypical antipsychotic drug response. Moreover, new phenotypes are now being considered beyond simple assessment of interview—based rating scales, including neurohormone plasma levels, neurocognitive indices, and structural and functional brain imaging measures. These endophenotypic measures may provide greater power to detect the relatively subtle effects of gene variants on the complex phenotype of drug response. Finally, the utilization of antipsychotic drug induced adverse events such as typical antipsychotic drug-induced tardive dyskinesia, clozapine-induced agranulocytosis, and atypical antipsychotic-induced weight gain has shown particular promise—with a number of groups now reporting replication of results across indepen-

Abstracts dent cohorts. In this section of the educational workshop, we will review the first generation of genetic studies of antipsychotic drug response, examine new pharmacogenetic data from our group and others suggesting replicability of results, and then discuss strategies for utilizing pharmacogenetic and pharmacogenomic techniques in new antipsychotic drug development. E3.3 PHARMACOGENETICS DISSECTION OF SUSCEPTIBILITY OF TARDIVE DYSKINESIA Segman RH Department of Psychiatry, Hadassah Hebrew University, Jerusalem, Israel Tardive dyskinesia (TD) is an involuntary choreoathetotic movement disorder observed in a significant minority of schizophrenia patients exposed to dopamine D2 receptor antagonist drugs. Genetically determined individual variability in compensatory responses to chronic dopaminergic antagonism, as well as in factors affecting drug levels, may account for a key portion of the variance in the incidence of TD. In recent years a number of groups have reported small effect contribution of candidate genes to risk for TD. These findings support a polygenic multi factorial inheritance for TD vulnerability. Implicated genes offer a renewed exploration of the molecular basis of TD, and open a window for improved understanding of drug induced extra pyramidal reactions. S1.1 THE MOLECULAR GENETICS OF EARLY-ONSET ALZHEIMER’S DISEASE Van Broeckhoven C, Rademakers R, Dermaut B, Theuns J, Kumar-Singh S, van Duijn C, and Cruts M Department of Molecular Genetics VIB8, Flanders Inter-University Institute for Biotechnology, University of Antwerp, Antwerp, Belgium Alzheimer’s disease is a neurodegenerative disease of CNS and is clinically characterized by progressive loss of memory and cognitive functions. The major pathological hallmarks of disease are neuronal cell loss, parenchymal deposition of amyloid Aaˆ in senile plaques, intraneuronal accumulation of the paired helical filaments of tau protein and aggregation of amyloid Ab in meningeal vessel walls. Alzheimer’s disease predominantly affects elderly people but in 1–2% of patients the disease symptoms appear at early age (25–65 years). In the majority of patients the disease is multifactorial with both genetic and environmental factors contributing to the expression of dementia. In the young patients the genetic factors are predominant and in 10% the disease is inherited as an autosomal dominant trait. In the latter rare families positional cloning efforts have identified 3 genes: the amyloid precursor protein gene (APP) and the presenilin 1 and 2 genes (PS1 and PS2). The mutations are mostly single base changes predicting amino acid substitutions. Currently 196 mutations in the 3 genes have been documented in 434 families worldwide. However, not all families are explained by mutations in known genes, our data and that of others suggest that 30–50% of families segregate mutations in currently unknown genes. We recently performed a genome-wide linkage analysis in 2 informative multigenerational Alzheimer families and were able to identify the chromosomal locus in both families. In family 1083 the locus overlaps with the 17q21 locus for FTD at 17q21, however a mutation in the tau gene (MAPT), was excluded by both exonic and genomic sequencing. We are currently investigating other candidate genes as well as the possibility of genomic abnormalities in the MAPT region. In family 1270 we identified a novel locus at 7q36. Analysis of the coding exons of the 29 known genes in the candidate region identified a single base change in exon 10 of PTIP coding for Pax transactivation domain-interaction protein. PTIP, originally identified as a regulator of PAX transcription factors, plays also a role in the DNA damage response pathway.

S1.2 FINDING GENES FOR LATE ONSET ALZHEIMER’S DISEASE Williams J UWCM, Cardiff, United Kingdom The financial costs of Alzheimer’s disease (AD) are measured in billions and are estimated to be higher than those of stroke, heart disease and cancer combined. The identification of genes for early onset forms of AD have provided valuable insight into AD development, but it remains to

5

be seen whether these translate to common, late onset forms of the disease. The association of allele e4 of the Apolipoprotein gene with increased susceptibility to AD stands as one of the most consistently replicated findings in complex genetics, but accounts for only a proportion of the genetic contribution to late onset AD (LOAD). Molecular genetic studies of LOAD have, with a few exceptions notably APOE4, been characterised by a few large collaborative, appropriately powered linkage studies and large numbers of small, generally underpowered and, for the most part, un-replicated association studies. The most compelling evidence of linkage to LOAD is observed on chromosomes 10, 12 and 9. Recent co-variate linkage analysis has also highlighted regions on chromosomes 6 for LOAD with psychosis and 21 for LOAD with a late age at onset. Positional/functional candidate genes have been tested for association in many of these regions but none appear to account for the linkage signals, convincingly. Our UK/US AD collaborative group have performed a low density genome screen for linkage disequilibrium spanning the regions of linkage on 10, 12 and 9 and have re-sequenced and tested many functional candidate genes in these regions. These novel data will be presented and interpreted in the context of previous findings. S1.3 THE ROLE OF TAU IN FAMILIAL AND SPORADIC TAUOPATHIES Goate A,1 Pastor P,1 Norton J,1 Charraverty S,1 Morris J,1 McKeel D,1 Racette B,1 Perlmutter J,1 Esquerra M,2 and Tolosa E2 1 Washington University, St Louis, United States, 2 Movement Disorders Unit, Neurology Service, Barcelona, Spain Tau inclusions have been reported in a large number of familial and sporadic neurodegenerative disorders, A causal role for tau in the development of disease was first established in 1998 with the report of tau gene mutations that cause an autosomal dominant form of familial frontotemporal dementia. Since this time more than thirty missense and splice site mutations have been reported. However, not all familial frontotemporal dementia families carry mutations in the tau gene. Several families, including the HDDD2 family, studied by our group, show evidence of linkage to markers on 17q21 but do not carry mutations in the tau coding sequence. Fine mapping studies are underway in these kindreds to identify the disease locus. In addition, several FTD kindreds show linkage to markers on chromosomes 3 and 9. Sporadic tauopathies include progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD). Genetic studies in these two disorders have reported consistent evidence for association with polymorphisms within and flanking the tau locus. Our recent studies have identified a region of one megabase, containing three haplotype blocks that shows strong linkage disequilbrium with PSP and CBD. Analyses in a Spanish sample identified a specific haplotype (H1E’A), which was present in 16% of PSP patients, but was not observed in controls. Furthermore, the opposite haplotype to H1E’ A, the H2E’A haplotype was rarely present in the disease group suggesting that it plays a protective role. These strongly support the hypothesis that variation within or close to the tau locus modifies risk for sporadic tauopathies.

S1.4 GENETIC DISSECTION OF THE ETIOLOGIES AND PATHOGENESES OF ALZHEIMER’S AND PARKINSON’S DISEASES AND RELATED DISORDERS Hardy J NIH, Bethesda, United States Genetic analyses have implicated mutations in the APP gene as one cause of early onset, autosomal dominant, Alzheimer’s disease. These mutations, like mutations in the presenilin genes, alter the amount of the peptide, Ab42 produced during APP processing. Genetic analysis of a recent genome screen for late onset Alzheimer’s disease, has implicated the APP locus as risk factor locus, for late onset disease although frank mutations have not been found: this suggests that genetic variability in APP expression contributes to the risk of this form of the disorder: not surprising, perhaps, given the longstanding association between trisomy 21 and AD. Similarly, mutations in the tau gene cause autosomal dominant tangle disease (FTDP-17), and genetic variation at the tau locus, but not coding changes, is associated with the sporadic tangle diseases, progressive supranuclear palsy and corticobasal degeneration: this suggests that genetic variability in either tau expression or in tau splicing contributes to the risk of these

6

Abstracts

diseases. Finally, mutations in the a-synuclein gene cause autosomal dominant Lewy body disease, and genetic variability (haplotypic association) at the a-synuclein locus contributes to sporadic disease: our recent demonstration that one cause of autosomal dominant disease is a triplication of the a-synuclein locus, indicates that the most likely explanation of this observation is that this haplotypic association reflects the fact that genetic variability in the control of a-synuclein contributes to disease risk. These observations have in common support for the notion that these common diseases are initiated by overexpression of key pathogenic proteins which are close to their threshold of solubility.

S2.2 HAPLOTYPE MAPPING FOR MEDICAL GENETICS Daly MJ Broad Institute, Cambridge, United States, Whitehead Institute, Cambridge, United States The Human Haplotype Map is a variation catalog, which is designed to accelerate the discovery of genes involved in complex disease risk, severity and response to therapy. Remarkable progress has been made in the last year in the development and interpretation of this resource. The primary focus of this talk will be to highlight what has been learned about human genetic variation as a result of this work and, most importantly, how this is translating into more efficient and thorough studies of genetic variation and its association to disease. In addition, the prospects for whole-genome scans for association in neuropsychiatric disease will be discussed as well as how analagous work in mice may enable more effective use of that model system in this discovery process.

S2.3 DEFINITION, DETECTION AND INTERPRETATION OF GENE-GENE INTERACTION (EPISTASIS) IN COMPLEX DISEASE Cordell HJ University of Cambridge, Cambridge, United Kingdom Detection and interpretation of gene-gene interactions (epistasis) has been considerably confused in the literature by differing (usually unstated) definitions and assumptions and by the use of the same terminology to apply to rather different statistical and biological concepts. Classical concepts of epistasis as applied to to mutually exclusive phenotype events are not easily generalized to apply to dichotomous outcomes such as presence or absence of disease, particularly for complex diseases where we expect to see reduced penetrance and the presence of phenocopies. In this presentation I will survey some of the most commonly used statistical and biological definitions of epistasis and point out some problems with and differences between them. The detection of (various definitions of) epistasis in the context of linkage and association analysis will be discussed, and the impact of epistasis on the power of population and family-based association studies will be considered.

S2.4 HAPLOTYPE MAPPING IN PHARMACOGENETICS Goldstein D University of London, London, United Kingdom There is considerable enthusiasm for haplotype mapping in pharmacogenetics, but most studies reported to date have been haphazard and insufficient to comprehensively represent variation in the genes most likely to be relevant to drug safety and efficacy. Here I report on an analysis of patterns of genetic variation in 56 genes that metabolise or transport prescription medicines. Detailed analyses of 754 single nucleotide polymorphisms (SNPs) genotyped in two population samples (European and Japanese) provide a set of haplotype tagging SNPs that economically represent variation in most of the major enzymes that act on prescription drugs. These analyses provide a framework for systematic association studies in pharmacogenetics, and address a number of outstanding questions relating to haplotype mapping. Following this, I provide a number of applications of haplotype mapping of variable drug response, emphasizing the work that needs to be done to translate genotype-phenotype correlations into clinically useful diagnostics, and clinically useful leads concerning new therapeutic targets.

S3.1 DISC1: PROGRESS FROM DISCOVERY TOWARDS AN UNDERSTANDING Millar K,1 Mackie S,1 Pickard B,1 James R,1 Buchanan S,1 Christie S,1 Malloy P,1 Blackwood D,2 Muir W,2 and Porteous D1 1 University of Edinburgh, Edinburgh, United Kingdom, 2 Royal Edinburgh Hospital, Edinburgh, United Kingdom Disrupted In Schizophrenia 1 and 2 (DISC1 & DISC2) were identified as likely genetic risk factors for schizophrenia and affective disorders through study of a large Scottish family. In this family a balanced translocation between chromosomes 1 and 11 co-segregates with psychiatric illness generating a maximum LOD score of 7.1 for a broad model including schizophrenia, bipolar affective disorder and severe recurrent depression. Inheritance of the translocation clearly causes the psychiatric illness in this family although not all translocation carriers are affected. However P300 event-related potential measurements indicate that inheritance of the translocation is likely to affect brain function in all individuals. Consequently all translocation carriers are likely to be at increased risk of developing psychiatric symptoms. There is no evidence that the translocation affects expression of any genes on chromosome 11, however on chromosome 1 the translocation directly disrupts DISC1 and DISC2, indicating that altered expression of one or both of these genes underlies the psychiatric symptoms suffered by translocation carriers. Several subsequent genetic studies have provided independent supporting evidence for an involvement of the DISC locus in causing psychiatric illness in many populations. DISC1 encodes a large protein of unknown function and has been the focus of many recent studies aimed at identifying its role in the central nervous system. We and others have used yeast two-hybrid screens to identify potential DISC1-binding proteins. These indicate a novel role for DISC1 in cell signalling pathways. The effect of the translocation upon DISC1 expression and these interactions will be discussed. S3.2 DISC1: BASIC BIOLOGY AND ITS IMPLICATION IN GENERAL SCHIZOPHRENIA AND BIPOLAR DISORDER Kamiya A and Sawa A Department of Psychiatry, Johns Hopkins University School of Medicine, Baltimore, United States Although schizophrenia is one of the severest psychiatric illnesses, its pathophysiology remains unclear. Disrupted-in-Schizophrenia-1 (DISC1), identified at the breakpoint of a chromosome (1;11)(q42.1; q14.3) translocation tightly linked to schizophrenia in a large Scottish pedigree, is also associated with schizophrenia in other populations, suggesting that it is a model for the pathogenesis of schizophrenia. The translocation causes C-terminal-truncation of DISC1 protein (mutDISC1). Here, we present evidence that normal DISC1 is a component of the dynein/microtubule complex, which regulates microtubule dynamics and neuronal migration. Cellular and biochemical analysis revealed that mutDISC1 acts as a dominant-negative, by binding to wild-type DISC1 and disturbing its normal subcellular distribution. Loss of DISC1 function impaired neurite outgrowth in vitro as well as neuronal migration in developing cerebral cortex in vivo. The defects induced by mutDISC1 support the idea that schizophrenia is a disorder of neurodevelopmental origin.

S3.3 DYSBINDIN—HAPLOTYPES, INTERMEDIATE PHENOTYPES AND DYSREGULATION OF TRANSCRIPT AND PROTEIN EXPRESSION Straub R National Institute of Health, Bethesda, MD, United States We have been pursuing the leads provided by the identification of dysbindin as a susceptibility gene for schizophrenia at a number of different levels. We have resequenced portions of the gene using DNA from at least 90 affected individuals and have identified and genotyped many novel SNPs. Many haplotypes are evident, and allele and haplotype frequencies often differ dramatically between Caucasians and African Americans. We have examined the relationship between SNP alleles and haplotypes and 1) intermediate phenotypes such as executive function, working memory, verbal fluency, and intelligence, 2) functional magnetic resonance imaging (fMRI) measures of activation of the dorsolateral prefrontal cortex (DLPFC) during cognitive tasks, and 3) transcript and protein expression levels in controls versus

Abstracts schizophrenics. We have also tested SNPs in genes coding for proteins in two macromolecular complexes that may be compromised by dysbindin dysfunction. We have examined genes coding for other members of the BLOC (biogenesis of lysosome-related organelles) complexes, as well as genes for proteins present in the DPC (dystrophin protein complex). We have also examined a number of genes that may be functionally related to the BLOC complexes, especially those that regulate synaptic vesicle formation, filling, docking, fusion and recycling. Finally, we have begun using statistical methods to test for the interaction between dysbindin and other susceptibility genes, and have generated some preliminary evidence that the dysbindin TDT signal with the clinical phenotype is far stronger in families where the proband has a COMT met/met genotype than in other families. Some or our recent findings will be presented, along with an overview of the haplotype association results and expression findings from other groups. S3.4 VARIATION IN DISC1 AFFECTS HIPPOCAMPAL STRUCTURE AND FUNCTION AND RISK FOR SCHIZOPHRENIA Weinberger DR, Callicott JH, Pezawas L, Egan MF, Hariri AR, Mattay VS, Goldberg TE, Verchinski BA, and Straub RS National Institutes of Health, Bethesda, United States Disrupted-in-schizophrenia-1 (DISC1) is a promising schizophrenia candidate gene expressed with greatest abundance within the hippocampus—a brain region strongly implicated in the pathogenesis of the illness. We studied genetic variation in this gene and its relationship to schizophrenia and to biologic phenomena associated with schizophrenia in a family based study of 320 Caucasion families, in two other smaller trio datasets, and in a case-control dataset. A threemarker single nucleotide polymorphism (SNP) haplotype across 83 kb of the gene was associated with schizophrenia in the family dataset (P ¼ 0.002). Weak association to this haplotype was also found in the NIMHGI Caucasion dataset. A common nonconservative SNP (ser704cys) within this haplotype also was associated with schizophrenia in the large family sample (P ¼ 0.004). A six marker haplotype spanning the gene also showed association in the case control analysis (P < .001). In addition to overtransmission to probands, the ser allele was associated with subtle hippocampal abnormalities even in healthy subjects, including: 1) reduced hippocampal volume measured with MRI and voxel-based morphometry, 2) reduced hippocampal synaptic activity as measured by n-acetylaspartate (NAA) levels and 3) abnormal engagement of the hippocampus during several cognitive tasks assayed with fMRI. These convergent data further implicate DISC1 as a schizophrenia risk gene and suggest that the mechanism of this effect involves structural and functional alterations in the hippocampal formation.

S4.1 INDUSTRIAL/ACADEMIC RELATIONSHIPS IN PSYCHIATRIC GENETICS: ONE ACADEMIC’S EXPERIENCE DeLisi L New York University School of Medicine, New York, United States Finding genes that contribute to psychiatric illness is an arduous and monumental task, particularly for academic researchers with limited physical and financial resources. With these problems, the time to obtain even a partial result can be years, and the room for laboratory error large. Thus collaborations with commercial facilities are particularly welcome, providing quality is maintained, and hardware are state-of-the-art. More than a decade ago, this academic university professor had amassed a large collection of multiplex families with schizophrenia and was in a search for a laboratory large and expert enough to be able to perform a genome-wide screen for linkage to genes for schizophrenia. She began with a colleague with his own laboratory, who spent 6 months putting together and purifying the DNA, after which only 10 markers were eventually run in more than a year later, and with much difficulty, on only some of the samples. Once one realizes how many samples need to be run with about 400 markers, the organization of a normal academic laboratory does not seem sufficient to carry this out. One small genomics company was found with enthusiastic investigators interested in the problem of schizophrenia that had impressive rooms filled with PCR equipment connected to a robotic system to manipulate and organize samples, free from technician error. However, through the years this company changed

7

investigators, was bought by another larger enterprise, directions changed and they were sold to a pharmaceutical company, which a year later was taken over by another pharmaceutical company. The project then stood in limbo for 1 more year while the company reorganized and ultimately terminated the project without ever accomplishing its goal. There are many ‘‘pro’s’’ to academic collaborations with companies, such as obtaining a large amount of expensive data produced in the highest quality. The question is whether the ‘‘pro’s’’ outweigh the ‘‘con’s’’? Some of the ‘‘con’s’’ from an academic prospective include, the restriction to secrecy surrounding data produced, the conflicting goals (academic: scientific progress; company: financial), ownership of data and ownership of samples, and the reluctance to return material to investigators when projects are terminated. These are but a few of the issues that need resolution before such liaisons are established and the agreements need to be explicit in writing. Ultimately, the end-point is to find a new treatment for schizophrenia (or other serious psychiatric disorder) based on knowledge gained through new genetic technology. Thus, we must overcome the problems to find a way to work together toward this important goal. S4.2 AUTOMATED DNA ANALYSIS: LIFE AT THE UNIVERSITY-INDUSTRY INTERFACE Cantor C SEQUENOM, United States SEQUENOM was founded in 1994 based on technology licensed from three academic groups. It operated as a virtual company in these three laboratories until 1996 when it transitioned to a real company, first in Hamburg, Germany, and now principally in San Diego. The company is a world leader in the provision of automated technology for precise analysis of DNA and RNA. While the core technology was developed in house, many of the most useful applications have been developed independently by academic collaborators or not-for-profit customers. SEQUENOM is also the largest single user of its own technology. The company has completed whole genome genetic association scans in twelve disease areas including various cancers, metabolic diseases and cardiovascular risk factors, and schizophrenia. Academic collaborations have been key, both in providing populations for genetic discovery or confirmation studies and for downstream biology studies to characterize the genes found associated with disease. S5.1 DNA POOLING ANALYSIS METHODS Knight J and Sham PC Institute of Psychiatry, London, United Kingdom Although a number of the genes that influence complex traits have been identified there are many yet to be discovered hence laboratory methods and statistical techniques are being continually developed. In part due to the large number of SNPs that have been identified over the past few years the use of pooled DNA has become increasingly popular. This approach allows the preliminary investigation of markers to be undertaken in less time and with less resources. The methodology, concerns and power of pooled DNA analysis will be outlined in this talk. Analysis of pooled data using an alternate formulation of the chi-squared test that incorporates measurement error is very common. After providing details of this method I will present a novel technique involving meta-regression. This technique can be applied to many pooling designs but is particularly useful as it can be applied to multiple pools made up of individuals with different phenotypic values, for example 5 pools containing the individuals from each quintile of the distribution of a quantitative trait. There are still some concerns relating to the analysis of pooled DNA such as the effect that differential amplification and mis-specification of measurement error. An overview of some of the literature and recent work investigating such concerns will be given. The talk will conclude with a comparison of the power of the method with the power of individual genotyping as pooling will only be useful if this ratio is high. S5.2 USE OF COVARIATES IN MODEL-FREE LINKAGE ANALYSIS Holmans P University of Wales College of Medicine, Cardiff, United Kingdom Genetic susceptibility to complex traits is likely to involve a number of factors, with the effects due to any one factor being small. It is therefore

8

Abstracts

important to increase the power of linkage studies by refining the phenotype and/or reducing genetic heterogeneity. A promising way to achieve this is to incorporate covariates into the analysis. These may include clinical data (e.g. disease severity, age at onset), environmental risk factors (e.g. smoking) or data from other susceptibility loci (genotypes associated with disease or linkage peaks). We present an overview of the existing methods for including covariates in model-free linkage analyses, together with the strengths and weaknesses of each approach. S5.3 DETECTING MULTIPLE ASSOCIATIONS IN GENOMEWIDE STUDIES Dudbridge F Medical Research Council, Cambridge, United Kingdom With recent advances in high throughput sequencing and genotyping, association studies are becoming feasible on a genome-wide scale. The large number of tests performed in these scans has led to calls for more efficient statistical methods to detect multiple associations from exploratory analyses. A recently proposed method is to combine the evidence from the most significant subset of tests. This approach has higher power than single-test adjustments and identifies a candidate set of loci for follow-up. Analytic distributions can be fitted to correlated tests occurring in genome scans, allowing standard marker panels to be calibrated for repeated use. Another recent approach, the false discovery rate, is less suitable for genome scans on account of its high variance when the number of true associations is small, and its modest reduction in required sample sizes. S5.4 FINE-SCALE LINKAGE DISEQUILIBRIUM MAPPING Morris A Wellcome Trust Centre for Human Genetics, Oxford, United Kingdom In the hunt for disease genes, initial screens of the genome may identify large regions with a positive signal, containing many potential candidates for further investigation. The aim of fine-scale mapping is to refine the location of the disease locus within these regions using high density maps of markers. The key principle underlying all linkage disequilibrium (LD) mapping methods is that, in the vicinity of the disease locus, a sample of case chromosomes will tend to have more recent shared ancestry than do control chromosomes, because they tend to share a recent disease-predisposing mutation. Consequently, a sample of case chromosomes is expected to display excess sharing of marker alleles over control chromosomes, the excess decaying with distance from the disease locus. However, this simple situation is often complicated by multiple disease mutations, sporadics among the sample of case chromosomes, mutations at marker loci, and allele sharing due to population substructure or founder effects. In this presentation, we will review fine-scale LD mapping methods that model the shared ancestry of case and control chromosomes in the candidate region. By modelling ancestry, we can more efficiently detect excess allele sharing due to shared inheritance of a disease mutation and distinguish it from background patterns of LD.

S6.1 CAUSAL ROUTES OF GENETIC INFLUENCES ON SOCIOSEXUAL BEHAVIORS IN ANIMALS Pfaff D, Ogawa S, Easton A, and Dwyer E The Rockefeller University, New York, United States Success in analyzing hormone-responsive gene expression in hypothalamic neurons as they contribute to simple sexual behaviors (‘‘Drive,’’ The MIT Press, 1999) led to an attempt to conceptualize the causal routes for these genetic influences. They are arranged in modules. In females, estrogens cause nerve cells to Grow; they Amplify their own effects through transcriptional facilitation of the progesterone receptor; they foster Preparatory behaviors; they cause Permissive actions by hypothalamic neurons on the rest of the female sex behavior circuitry; and they help Synchronize sex behavior with ovulation (GAPPS, Mong and Pfaff, Molecular Psychiatry, 2004). A closely related set of functions permits social recognition and the corresponding reduction of aggression among mice. These require normal gene expression for Estrogen Receptors alpha and beta, oxytocin and the oxytocin receptor expressed variously in the amygdala and the paraventricular nucleus of the hypothalamus (Choleris et al., PNAS 2003).

Finally, some of the genetic influences on these sociosexual behaviors are indirect, routed through alterations in generalized arousal of the CNS, itself regulated by more than 100 gene products (Brain Arousal and Information Theory, Harvard University Press, 2005). S6.2 SEX CHROMOSOMES AND BRAIN GENDER Arnold A Department of Physiological Science and Laboratory of Neuroendocrinology of the Brain Research Institute, Los Angeles, United States A male child is born with brain cells that express Y chromosome genes not present in females. A female child has two copies of some X chromosome genes expressed in each brain cell, twice as many as in the male. Do these differences in gene expression induce sex differences in brain function? The dominant theory of mammalian sexual differentiation holds that sex differences in neural phenotype are attributed exclusively to the action of gonadal secretions during early periods of development, not to the direct action of sex chromosome genes in the brain itself. We have compared the neural and behavior phenotype of mice in which the testis determination gene Sry is moved from the Y chromosome to an autosome, so that testis determination is inherited independently from the sex chromosomes. This allows comparison of mice that have the same gonadal type but differ in their complement of sex chromosomes (XY females (possessing ovaries) vs. XX females; XX males (possessing testes) vs. XY males). Most sexual dimorphisms in these mice are masculine in animals with testes, and feminine in animals with ovaries, confirming the dominant hormonal theory. However, XY males are more masculine than XX males, and XY females are more masculine than XX females, in the density of vasopressin fibers in the lateral septum. When cells are cultured from the midbrain of embryonic mice, XX and XY cultures differ in the number of dopamine neurons, independent of the gonadal type of the mouse. Thus the sex chromosomes can also influence sexual differentiation via mechanisms that are probably not mediated by gonadal secretions. Male (but not female) zebra finches sing a courtship song, and the male’s neural song circuit is much larger than the female’s. Gonadal secretions are probably not the primary cause of sexually dimorphic brain development. Females induced to develop testes have a feminine song circuit. Moreover, a lateral gynandromorphic finch, which was genetically male on the right half of its body, and female on the left half, had a brain that was more masculine on the right than left. In this case genetically male and female brain cells developed in an identical gonadal hormonal milieu, so that the differences between the two sides are attributable to lateral differences in the complement of sex chromosomes. The results suggest that the sex difference is caused by brain-autonomous action of sex chromosome genes.

S6.3 WHAT CAN STUDIES ON TURNER SYNDROME TELL US ABOUT THE ROLE OF X-LINKED GENES IN SOCIAL COGNITION? Skuse DH,1 Purcell S,2 Daly MJ,2 Dolan RJ,3 Morris JS,1 Lawrence K,1 Lander ES,2 and Sklar P4 1 Institute of Child Health, London, United Kingdom 2 Whitehead Institute for Biomedical Research, Cambridge, United States 3 Institute of Neurology, London, United Kingdom 4 Broad Institute, Cambridge, United States 45,X (Turner syndrome) females have specific social-cognitive deficits including impairments of gaze monitoring, face recognition memory, and Theory of Mind skills. Facial emotion recognition is especially poor and of a similar pattern to patients with bilateral amygdalectomy, although amygdala volume in 45,X is increased. We have shown, in fMRI experiments, that impairment in the recognition of fear is associated with decreased functional connectivity between amygdala and fusiform gyrus compared to 46,XX controls, implying X-linked genes play a critical role in the development and function of the social brain. We predicted, and confirmed that normal (46,XY) males with exceptionally poor fear recognition skills (selected from population screening) would show equivalent functional amygdala-cortical disconnection. Taken together, these findings imply an X-linked QTL influences the integrity of circuits in the ‘social brain’ not only in Turner syndrome but also in the general male population. We mapped the locus in females with variably sized deletions of the X-chromosome to a region

Abstracts of 4.96 Mb at Xp11.3/4. Females with terminal or interstitial Xdeletions of this critical locus had normal intelligence, but socialcognitive deficits and brain structural anomalies similar to 45,X females. A simple facial fear recognition endophenotype was measured in 93 45,X individuals and the resulting quantitative trait coded as a standardized Z-score. 252 SNPs were genotyped across the 4.96 Mb region with an average inter-SNP distance of 21 kb. Quantitative trait haplotype analysis was performed using a maximum-likelihood approach, regressing the quantitative trait onto haplotype copy number. Haplotype analysis revealed 3 initial loci of interest, one of which contained a 4-SNP block spanning a 40 kb region in a novel and uncharacterised gene, which is strongly expressed in brain. Allelic variation in this block had a substantial effect on the endophenotype (i.e. fear recognition and by implication, amygdala-fusiform connectivity). These initial findings are currently the subject of a replication study. If supported, they are likely to have major implications for our understanding of male vulnerability to disorders of social cognition. S6.4A A ROLE FOR GENOMIC IMPRINTING IN MALE SEXUAL ORIENTATION Bocklandt S,1 Roselli CE,2 DuPree MG,3 Vilain E,1 and Hamer DH3 1 Department of Human Genetics, David Geffen School of Medicine, University of California Los Angeles, Los Angeles, CA, United States, 2 Department of Physiology and Pharmacology, Oregon Health and Science University, Portland, OR, United States, 3 Laboratory of Biochemistry, National Cancer Institute, National Institutes of Health, Bethesda, MD, United States Genomic imprinting of X-chromosome genes could lead to sex-specific gene expression, and is therefore a candidate mechanism for the regulation of sex specific traits, such as sexual orientation. Since X-inactivation and genomic imprinting share several molecular mechanisms, we measured X-inactivation in mothers of gay men to test the role of imprinted X genes in sexual orientation. We analyzed X chromosome inactivation in 96 mothers of gay men compared to 105 control women matched for age and ethnicity. Skewing was assayed in white blood cell DNA by an androgen receptor gene assay and confirmed at the fragile X locus. The results of the two assays were highly correlated (Pearson correlation ¼ 0.74). We found a significant difference (P ¼ 0.004) in X chromosome skewing between the two groups. The mothers of gay men showed a bimodal distribution in which 14% of the cases displayed extreme skewing (more than 90% inactivation of one chromosome). The control women showed a normal distribution in which less than 3% of the sample had extreme skewing. Analysis of mouthwash samples showed that individuals with extreme skewing in blood DNA also displayed higher skewing in buccal cell DNA (Pearson correlation ¼ 0.75). Analysis of 47 members of a large family suggests that the extreme skewing is shared by sisters of the mother, suggesting that the effect is genetic or epigenetic. However, the skewing is passed on to only 1 out of 24 daughters of the mothers with extreme skewing, indicating that it is not caused by a mutation on the X chromosome, or by specific alleles of the X-Inactivation Center. In order to identify the exact mechanisms and genes involved in genomic imprinting in male sexual orientation, we undertook two complimentary approaches, based on linkage and expression profiling. A full genome linkage scan was performed on 146 gay brother pairs. Significant linkage to 10q26 was found on the maternal allele; however, there was no linkage on the paternal allele, indicating the presence of a maternally expressed, paternally silenced, imprinted gene involved in male sexual orientation. Extensive micro array expression analysis of cell lines derived from gay and straight men and their families was performed. In addition, cross species micro array hybridization was performed on the anterior hypothalamus and amygdala of male oriented and female oriented rams. Candidate genes derived from a combination of human and sheep data have been identified. Altogether, this data suggests that male sexual orientation in mammals is influenced by genomic imprinting. S6.4B PROGRESS IN THE SEARCH FOR X-LINKED BEHAVIOURAL QTLS Craig IW and Loat CS SGDP Centre, Institute of Psychiatry, Denmark Hill, London, United Kingdom The X chromosome has been a focus for research on genes affecting cognition and behaviour. We have explored the hypothesis that this

9

chromosome may harbour quantitative trait loci for a range of behaviours by studying the correlations between both monozygous (MZ) and dizygous (DZ) females and males. The X-linked QTL hypothesis predicts that among MZ twins, females should be more discordant than males (as a result of discordant skewing) and the reverse should obtain for DZ twins. Studies on twins at ages 2-4 supports this concept. Given also that females possess two X chromosomes to the male’s one, genes which escape from dosage compensation may be expressed at higher levels in their tissues including brain. Micro array approaches have detected a range of Xlinked genes, whose expression levels suggest that they are not inactivated and whose functions suggest them as possible candidates underpinning at least some of the behavioural differences between males and females. S7.1 MAKING FACES: THE MOLECULAR CONTROL OF CRANIOFACIAL DEVELOPMENT Sharpe PT Kings College London, London, United Kingdom The mammalian head is the most complex structure in the body. It contains the brain and provides the framework to house the sense organs and feeding apparatus. The organisation of the hard tissues of the head that provide this framework begins early in embryonic development when mesenchyme cells acquire their positional information. The molecular mechanisms that determine the patterning of hard tissues that direct their subsequent morphogenesis are starting to be elucidated. Focussing our attention on the development of the jaws and teeth, we have identified a hierarchy of temporo-spatial cell interactions that regulate patterning. Details of these interactions, together with how this information is used to generate ‘‘form’’ will be presented. The prospects of harnessing this basic understanding to develop new cell-based approaches for the repair and replacement of craniofacial hard tissues will be discussed.

S7.2 EMBRYOLOGICALLY DERIVED MEASURES OF CRANIOFACIAL AND BRAIN DYSMORPHOLOGY IN SCHIZOPHRENIA Deutsch CK Shriver Cr, Harvard, Waltham, United States Using an embryological approach that relates craniofacial dysmorphology to brain pathology, we have identified a craniofacial phenotype associated with schizophrenia. Both the face and brain derive from common embryonic primordia and are shaped by shared forces. Therefore, a genetic or environmental insult that disrupts early development could manifest itself both as brain pathology and craniofacial dysmorphology. Indeed, medical genetics and teratology atlases abound with examples of neuroembryological development that has gone awry. Our craniofacial studies conform to the embryologic convention of demarcating facial prominences, which are outgrowths from neural crest ectomesenchyme that arise as recognizable, discrete masses. By the 19th day of embryogenesis, three primary craniofacial primordial regions (Anlagen) have formed: the single median frontonasal and the paired maxillary and mandibular prominences. Development of these prominences progresses from the embryonic to the mature face, and anomalies can be classified by the specific primordia from which they derive. Our quantitative dysmorphology studies of schizophrenia have found prominent asymmetries at the interface of the frontonasal and maxillary anlagen. The derivatives of this interface are exemplified at its upper and lower boundaries by the orbital and palatal regions, respectively. We have also found these patterns of dysmorphology to be overrepresented among the first-degree relatives of schizophrenic probands. Yet they are absent among bipolar probands and their first-degree relatives. Further, we will discuss an embryologically derived prediction of brain midline dysmorphogenesis based on our craniofacial findings. To test this hyptothesis, we have studied brain morphology in schizophrenia in collaboration with colleagues at the MRI Division, Surgical Planning Laboratory, Brigham & Women’s Hospital (Boston). The prediction has been borne out—there was increased brain midline skewing in schizophrenia, arising at a locus consistent with the embryologic model. Moreover, these patterns of maldevelopment cohered in schizophrenia: there was a significant within-subject correlation between brain and craniofacial dysmorphology.

10

Abstracts

S7.3 FACIAL ASYMMETRY IN PATIENTS WITH SCHIZOPHRENIA AND THEIR BIOLOGICAL RELATIVES: AN INFORMATIVE PHENOTYPE? Gourion DG, Goldberger CG, and Krebs MOK INSERM 0117, PARIS, France The frequency of minor physical anomalies (MPAs) in patients with schizophrenia suggests an early disturbance in the development of the neuroectoderm. To improve the phenotypic delimitation of this disorder, we used a comprehensive scale of MPAs (41 items) in patients with schizophrenia and their first-degree relatives. This scale, adapted from a revised version of the Waldrop Scale with new items assessing facial and limbs asymmetry. Patients with schizophrenia and their non-psychotic parents had significantly more MPAs than healthy comparison subjects. A logistical regression model showed the ability of several items to predict group status, including facial asymmetry, cleft palate, hair whorls and abnormal palmar crease. Biological relatives were further classified as ‘‘presumed carriers’’ of the genetic loading if they had a second relative with schizophrenia in their ascendants and/ or collaterals (first or second degree) or as ‘‘presumed noncarriers.’’ MPAs but also Neurological Soft Signs (NSS) were compared. A discriminant function analysis based on total NSS and MPA scores correctly classified 71% of nonpsychotic parents in presumed carriers or presumed noncarriers. Conclusions: Neurological impairments and slight morphological anomalies seem to be associated with the genetic risk for schizophrenia, even when the disease itself is absent. Their presence might be a valuable composite intermediate phenotype for genetic studies. The high prevalence of facial asymmetry in patients with schizophrenia and their first-degree relatives provides new insights into the underlying dysembryogenic processes.

S7.4 3D MORPHOMETRICS REVEALS GENDER-RELATED DIFFERENCES IN FACIAL SHAPE AND ASYMMETRY IN SCHIZOPHRENIA Hennessy RJ,1 Lane A,2 Kinsella A,1 Larkin C,2 O’Callaghan E,2 and Waddington JL1 1 Royal College of Surgeons in Ireland, Dublin, Ireland 2 St John of God Hospital, Dublin, Ireland Over early fetal life cerebral and craniofacial morphogenesis proceed in embryological intimacy. Therefore, craniofacial shape differences between schizophrenia patients and controls are informative of developmental disturbance(s) in cerebral-craniofacial morphogenesis. 3D craniofacial coordinates were calculated from interlandmark distances for 169 patients with DSM-III-R schizophrenia and 78 matched normal controls. These were analysed using geometric morphometrics with visualisation of the resultant statistical models. Patients of both sexes were characterised by an intricate topography of 3D shape change involving lengthened lower mid-facial height, shortened upper mid-facial height, nasion located posteriorly and a wider face posteriorly; there was sex-specific rotation of the midface such that the base of the nose is more anterior in female patients but more posterior in male patients. Importantly, there were sex-specific asymmetries: in males, controls evidenced marked directional asymmetry while patients showed reduced directional asymmetry; conversely, in females controls evidenced little directional asymmetry while patients showed marked directional asymmetry. In schizophrenia, the topography of craniofacial dysmorphology appears to reflect subtle disruption to a critical 3D trajectory of embryonic-fetal craniofacial growth, particularly along the midline, with disturbance to the establishment of normal asymmetries in a sex-related manner. These studies were supported by the Stanley Foundation.

S8.1 FUNCTIONAL POLYMORPHISMS AND G-E INTERACTIONS Sugden K, McGuffin P, Caspi A, Moffitt T, and Craig IW SGDP Centre, IOP, London, United Kingdom Twin and adoption studies have established that there is a substantial genetic component in most behavioural and psychiatric disorders. However, identifying the molecular factors responsible for such influences has proven difficult. Associations with promising candidate

genes have often proved difficult to replicate. An explanation for this could be that other non-genetic factors that may modify the effect of a gene have been overlooked. By examining functional polymorphisms of genes, where the consequence of variation can be predicted, the extent of the influence of these individual differences in environment can be tested. The benefit of this type of approach has been demonstrated in two recent reports by Caspi et al. (Science (2002): 297(5582):851–4; Science (2003): 301(5631):386–9.). Here it was suggested that functional polymorphisms of the MAOA and 5-HTT genes could interact with environmental factors to predict a behavioural outcome (antisocial behaviour and depression respectively). Interestingly, no main effect associations were observed between the traits and polymorphisms. Other studies are currently underway to assess the extent of gene-environment interactions in other disorders. These findings might help to explain the contradiction in classical association study findings, and go someway to resolving the ongoing nature-nuture debate.

S8.2 ADVERSITY, PERSONALITY AND GENETIC VULNERABILITY TO DEPRESSION Farmer AE Institute of Psychiatry, London, United Kingdom Aspects of personality may be associated with the vulnerability to develop depression. A sib-pair method has been used to examine the familiality of the 7 scales of the Temperament and Character Inventory (TCI)1 and whether this could be related to the genetic vulnerability to develop depression. Depressed probands and their nearest aged siblings from the UK, were compared with healthy control probands and their nearest aged siblings on the TCI and measures of depressed mood and life events. All 7 scales of the TCI were familial, and scores on 6 of the scales were similar to US population scores. However, the UK subjects’ scores on the Self-transcendence scale were markedly lower than the US mean, suggesting strong cultural/national influences on this measure. Harm avoidance scores were substantially influenced by current and past depression, but the scale also showed stable trait-like characteristics that appear to be related to the genetic vulnerability to depression. Novelty seeking and self-directedness were also partly state-dependent and negatively correlated to low mood. High reward dependence scores may also protect against the development of depression and is unrelated to mood-state. The Cooperativeness, Persistence and Self-transcendence scales appear to have only a limited relationship to the development of depression. The 7 scales of the TCI were also examined in relation to the type and severity of life events that may have occurred in the 12 months prior to the onset of depression or prior to interview in the non-depressed subjects. Fiftyfive control subjects who had shown resilience by not developing depression despite experiencing an adverse event had significantly more non-severe events than subjects with depression. This suggests that experiencing such events may provide some protection against the depressogenic effects of severe threatening events when these occur.

S8.3 GENE-ENVIRONMENT INTERACTIONS IN PSYCHIATRIC DISORDERS: SUBSTANTIVE FINDINGS AND METHODOLOGICAL CHALLENGES Prescott CA, Kuhn JW, Vittum J, Riley BP, and Kendler KS Virginia Institute for Psychiatric & Behavioural Genetics, Virginia Commonwealth University, Richmond, VA, United States It has become increasingly clear that the etiology of psychiatric disorders can only be elucidated by identifying genetic and environmental contributing factors and understanding how they combine. In our work with the Virginia Adult Twin Study of Psychiatric and Substance Use Disorders (VATSPSUD, Kendler et al., 1992; Prescott et al., 1999), we have used two approaches to investigate geneenvironment interaction and coaction. First, we have studied how the occurrence and severity of recent stressful life events (SLE) interact with inferred genetic risk (based on cotwin and parental psychiatric history) to affect onsets of depressive and anxiety disorders. Second, we have examined the role of measured genes and SLE in risk for depression, anxiety, and alcoholism. We replicated the recent findings of Caspi et al. (2004) of an interaction between a functional variant in the serotonin transporter (5-HTT) and SLE in risk for major depression

Abstracts and extended the analysis in several ways: by using life event and onset information dated at the month level, by examining the role of the severity and type of stressful life event, and by studying whether these findings are specific to depression and to this genetic variant. The results underscore the importance of incorporating both measured genes and measured environments in genetic epidemiological models and point to some methodological challenges for future work of this kind. Supported by NIH grants MH-R01-40828, MH/AA/DA-R0149492, AA-R01-09095 and AA-K01-00236. Subject access was provided by the Mid-Atlantic Twin Registry, directed by Drs. J. Silberg, L. Eaves and L. Corey.

S8.4 GENE-ENVIRONMENT INTERACTION AND ASSOCIATION ANALYSIS Purcell SM Whitehead Institute, Cambridge, United States Gene-environment interaction (GxE) (genetic control of sensitivity to the environment, or, conversely, environmental control of gene effect) is an area of current focus in complex trait genetics. Consideration of GxE could potentially help to map genes and elucidate biological pathways. However, as has long been appreciated in epidemiology, the analysis of interactions is a difficult area harboring unresolved problems: a central problem when dealing with quantitative traits is that of scale dependence. Interactions can be induced or attenuated by transformation of the trait variable. For a botanical example: the length of a leaf might appear to be additive with respect to a certain genotype; if, however, one were to measure leaf surface area, one might conclude dominance effects instead. More subtle effects are to be expected with polygenic behavioral measures that have no absolute unit of phenotypic measurement. With these well-known issues in mind, two novel likelihood-based methods are developed, both of which model genotype conditional on trait value (which is generally more robust to mild nonnormality and sample selection). The performance of the likelihoodbased models and two standard regression-based models is compared, in unselected, selected and trait-transformed samples. One likelihood approach appears to be the most generally robust. This approach has been extended to nuclear families, in the software package ‘cafe’. Furthermore, some interesting differences in the performance of this likelihood method versus the standard regression approach hint at the possibility of new, more robust methods. Finally, a set of simulations performed within a developmental framework is presented. These simulations show why we should expect to encounter nonlinear scaling effects in complex behavioral variables. The consequence of this is that, if naı¨ve statistical methods are applied, we should expect to see many ‘statistical’ interactions that do not represent true interaction on any ‘biological’ level.

O1.1 GENOME-WIDE GENETIC LINKAGE ANALYSIS OF THE NIMH GENETICS INITIATIVE ‘‘WAVE 4’’ BIPOLAR DISORDER PEDIGREES McMahon FJ,1 Austin L,1 Steele CJM,1 Foroud T,2 Meyer ET,2 Chen YS,3 Cox NJ,3 Nurnberger, Jr JI,2 and the NIMH Genetics Initiative Bipolar Disorder Group4 1 Genetic Basis of Mood & Anxiety Disorders, National Institute of Mental Health, Bethesda, United States 2 University of Indiana, Indianapolis, United States 3 University of Chicago, Chicago, United States 4 University of California, Johns Hopkins, University of Iowa, University of Pennsylvania, Washington University, United States The NIMH Genetics Initiative for Bipolar Disorder, a multi-center genetic study of bipolar affective disorder (BPAD), has previously reported genetic linkage studies in 3 waves of families. The results have implicated regions on chromosomes 1q, 6q, 7p, 10p, 16p, 16q, 17q, and 22q. We now report a genome-wide linkage analysis in a new sample of families (‘‘Wave 4’’), the largest to date. A total of 1058 persons in 309 families were ascertained through a sibling pair affected with bipolar I (BPI) or schizoaffective-bipolar disorder (SABP), evaluated with a semi-structured interview, and diagnosed in a best-estimate procedure. DNA samples were genotyped at a mean resolution of 9 cM by the Center for Inherited Disease Research. Genotype data was cleaned of monozygotic twins, family structure, and Mendelian errors. Nonparametric linkage analysis was performed with Genehunter-Plus.

11

Linkage evidence was assessed under 3 affection status models: ASM I—BPI and SABP; ASM—II ASM I plus bipolar II with recurrent major depression; and ASM—III ASM II plus recurrent major depression. Overall genetic information content was 55%, due to missing parents. The strongest linkage signal was observed at 16p (Zlr 2.8, ASM III), near the region reported in the previous samples. Gene-dropping simulations indicate that a Zlr of 3.5 is equivalent to a genome-wide P < 0.05 in this dataset. Zlr scores >2 were detected near several other previously-reported loci, including 1q, 13q, 16q, and 18q, but no signals reached genome-wide significance. Suspecting large-sample heterogeneity, we undertook a series of exploratory analyses aimed at detecting linkage in clinically-defined subsets. We used 4 clinical features known to be familial in BPAD pedigrees: psychosis, age at onset, comorbid alcoholism, and episode frequency. A significant increase in the Zlr score on 16q was observed in relative pairs with episode frequency below the median. Pairs with episode frequency above the median were not linked to 16q, but in these pairs we observed a significant increase in the Zlr score on 4q. Similarly, significant increases in the Zlr score were observed at 5q and 20q in relative pairs with non-psychotic BPAD. This genome scan, the largest to date in BPAD, provides supportive evidence for several previously-reported linkages. Episode frequency and psychosis may influence linkage to some of these loci, but this requires replication in additional samples. Clinical heterogeneity may reflect genetic heterogeneity in large bipolar disorder samples. O1.2 SEX-SPECIFIC ASSOCIATION BETWEEN WOMEN WITH BIPOLAR AFFECTIVE DISORDER AND GPR50, AN X-LINKED ORPHAN G PROTEIN-COUPLED RECEPTOR Thomson PA,1 Wray NR,1 Thomson AM,2 Dunbar DR,2 Grassie MA,2 Condie A,1 Pulford DJ,2 Muir W,3 Blackwood DHR,3 and Porteous DJ1 1 University of Edinburgh, Edinburgh, United Kingdom 2 Organon Laboratories Ltd, Newhouse, United Kingdom 3 Royal Edinburgh Hospital, Edinburgh, United Kingdom GPR50 is an orphan G-protein coupled receptor (GPCR) located on Xq28 in a region previously implicated in multiple genetic studies of bipolar affective disorder (BPAD). Expression of the protein is restricted in human brain to the hypothalamus, pituitary and infundibular stalk, regions increasingly studied in mood disorders. Allele frequencies of three polymorphisms were compared in casecontrol studies between 250 samples each of recurrent major depression (MDD), bipolar affective disorder (BPAD), and schizophrenic (SCZ) patients, and 550 ethnically matched controls. Significant associations were found for both MDD (P ¼ 0.01), and BPAD (P ¼ 0.007) with an insertion/deletion polymorphism in exon 2. Analysis of the allele frequencies of the insertion/deletion in females only, further increased the associations with MDD and BPAD (P ¼ 0.006 and P ¼ 0.0002 respectively). Clinical re-examination of the MDD sample identified a group meeting the criteria for bipolar spectrum disorder. Analysis of this subgroup indicated that the association within MDD was largely limited to bipolar spectrum (P ¼ 0.002) and not the remaining patients in this group (P ¼ 0.04). The two other polymorphisms tested in this gene also detected associations within our sample set. A non-synonymous SNP in exon 2 was associated with MDD in females only (P ¼ 0.01), and an intronic SNP was associated with SCZ (P ¼ 0.001) again only in females. No association was detected between any of the markers and male case groups. The boundaries of the haplotype block containing the insertion/deletion polymorphism have been defined. It spans 4.5 kb covering part of the intron, all of exon 2 and approximately 2 kb of the 30 , and includes the non-synonymous SNP in exon 2. The results suggest that the GPR50 insertion/deletion polymorphism, or a locus in tight LD with this polymorphism, is a sexspecific marker for familial susceptibility to bipolar disorder.

O1.3 ASSOCIATION OF DISC1 AND BIPOLAR DISORDER Maeda K,1 Chang J,1 Youn R,1 Kamiya A,1 Kawahara R,2 McInnis M,1 and Sawa A1 1 Johns Hopkins University, Baltimore, United States 2 Tottori University, Tottori, Japan Disrupted in Schizophrenia-1 (DISC1) was identified at the breakpoint of a balanced translocation segregating with schizophrenia and mood disorder in a large Scottish family. Additional findings of association between DISC1 and schizophrenia have been reported, but not with

12

Abstracts

mood disorder. We conducted a family-based association study of DISC1 and bipolar disorder in 57 bipolar pedigrees. We performed direct sequences with 10 samples from the patients with bipolar disorder (BP) and identified a novel (unreported) SNP. Twelve SNPs were genotyped at the DISC1 locus, including the novel SNP, using the Taqman assay. The computer program ‘‘FBAT’’ was used for familybased TDT and haplotype analysis. Association at individual SNP was observed at rs3737597 in 30 -UTR of DISC1 gene (P < 0.05). Haplotype analysis identified one haplotype (HP1) that was overtransmitted to the BP phenotype (P ¼ 0.01) and a second haplotype that was undertransmitted (HP2). There was evidence of gender influence in the transmission distortion, with overtransmission of HP1 to affected females (P ¼ 0.004). Expression levels of DISC1 in lymphoblasts were compared between affected subjects with HP1 and unaffected subjects with HP2. There was a significant decrease in DISC1 expression in affected HP1 group compared to cell lines derived from unaffected subjects with the HP2 (P ¼ 0.006). This difference was more pronounced in females (P ¼ 0.001). These results suggest an association between DISC1 and bipolar disorder, and a decreased level of expression of DISC1 in lymphoblastoid cell lines from affected female subjects with the risk haplotype. O1.4 ASSOCIATION OF GENETIC VARIATION AT THE TRYPTOPHAN HYDROXYLASE 2 (TPH2) GENE WITH BIPOLAR AFFECTIVE DISORDER Van Den Bogaert A,1 Ohlraun S,2 De Zutter S,1 Schumacher J,3 Schulze TG,2 Kovalenko S,4 Maier W,4 Propping P,3 No¨then MM,6 Rietschel M,2 and Cichon S6 1 Molecular Genetics, VIB8, University of Antwerp, Antwerp, Belgium 2 Central Institute of Mental Health, Division Genetic Epidemiology in Psychiatry, Mannheim, Germany 3 Institute of Human Genetics, University of Bonn, Bonn, Germany 4 Department of Psychiatry, University of Bonn, Bonn, Germany 5 Institute for Medical Biometry, Informatics and Epidemiology University of Bonn, Bonn, Germany 6 Life & Brain Center, University of Bonn, Bonn, Germany Recently, Walther and colleagues (2003) identified the gene for a novel tryptophan hydroxylase (TPH) isoform, TPH2, that is encoded on chromosome 12 and is exclusively expressed in the brain. TPH is the rate limiting enzyme for the production of serotonin from tryptophan. Serotonin is a key neurotransmitter in the central nervous system, and dysfunction of the serotonergic system has been implicated in several psychiatric diseases. We have investigated the potential influence of polymorphisms at the TPH2 locus on the development of bipolar disorder. In a first step, we re-sequenced the coding region and exon-intron boundaries of the gene in a representative, European population-based sample of 80 healthy individuals to determine the naturally occurring genetic variability. We identified a total of 13 SNPs with a minor allele frequency of >1%. 5 of these identified SNPs plus additional non-coding SNPs derived from the public databases were genotyped in a sample of 300 subjects with DSM-diagnosed bipolar affective disorder and 300 unaffected control subjects, all of German descent. Out of 12 genotyped SNPs, we identified four SNPs that were significantly associated with bipolar disorder in single marker analyses. Two SNPs located in the putative promoter region and 50 UTR showed P-values of 0.004 (SNPprom) and 0.031 (SNP1), respectively. Significance was also observed for a rare non-synonymous SNP in exon 6 (SNP3) and a non-synonymous SNP in exon 9, SNP6 (P ¼ 0.003 and P ¼ 0.038, respectively). Haplotype analyses provided supportive evidence for an involvement of genetic variation at the TPH2 locus and bipolar affective disorder. Replication of these findings in independent samples is currently underway.

O1.5 MOLECULAR GENETIC POSITIONAL AND CANDIDATE GENE STUDIES OF BIPOLAR AFFECTIVE PUERPERAL PSYCHOSIS Jones I,1 Hamshere M,1 Nangle JM,2 Middle F,3 Green E,1 Raybould R,1 Robertson E,3 Jones L,3 Heron J,3 Corvin A,2 Lambert D,2 Gill M,2 Owen M,1 and Craddock N1 1 Department of Psychological Medicine, UWCM, Cardiff, United Kingdom 2 Department of Psychiatry, Trinity College, Dublin, Ireland 3 Department of Psychiatry, University of Birmingham, Birmingham, United Kingdom

Women with bipolar disorder are at very high risk of suffering episodes of mania or psychosis in the immediate postpartum period (puerperal psychosis—PP). The available evidence suggests that in bipolar women, genetic factors influence vulnerability to puerperal episodes and, moreover, those women who have suffered such an episode, represent a more familial form of bipolar illness. For complex genetic disorders there are marked benefits in focusing on a more homogenous subtype and molecular genetic studies of bipolar affective puerperal psychosis provides such an opportunity as well as offering specific biological hypotheses to test. In our recruitment of families and individuals with Bipolar I Disorder we have obtained clinical information regarding the relationship of episodes of illness to childbirth allowing us to identify the subgroup of women who have suffered an episode of puerperal psychosis—over 200 women to date. We are using both positional and candidate gene approaches in the genetic investigation of PP. Positional: We have identified the subset of families in the Wellcome Trust UK-Irish bipolar sib-pair genome screen that include at least one female who has suffered an episode of puerperal psychosis. A maximum of 48 pedigrees contribute 71 affected sibling pairs to the sample. A genome scan was performed with MAPMAKER/SIBS, using allele frequencies estimated from the original Bipolar sample with SPLINK. Analysis of this more homogenous subgroup of families gives two LOD scores above 2—a maximum LOD score of 2.88 on 16p and of 2.03 on 2q. Candidate gene studies: We have undertaken studies at candidate genes of interest in Bipolar Disorder with an emphasis on those for which the expression is known to be influenced by steroid hormones. These include COMT, HSERT, MAOA, BDNF, G72 and DAAO. We find evidence for linkage disequilibrium between PP and several alleles and haplotypes at G72 (e.g. polymorphism M23: P < 0.001) with an effect size greater than that in our unselected Bipolar sample. There is also evidence for a linkage disequilibrium signal at DAAO.

O1.6 GENETICS OF RECURRENT EARLY-ONSET MAJOR DEPRESSION: GENOME SCAN LINKAGE ANALYSIS OF THE FULL GENRED STUDY SAMPLE Levinson DF7 Holmans P,1 Weissman MM,2 Zubenko GS,3 DePaulo JR,4 Crowe RR,5 Scheftner WA,6 Adams P,2 Knowles JA,2 March D,2 Zubenko WN,3 MacKinnon D,4 McInnis MG,4 Murphy-Eberenz K,7 Gladis M,7 Boutelle S,5 Marta DH,6 Thomas J,4 and Miller E4 1 Biostatistics and Bioinformatics Unit, University of Wales College of Medicine, Cardiff, UK, and MRC Biostatistics Unit, Cambridge, Cardiff, United Kingdom 2 Department of Psychiatry, College of Physicians and Surgeons, Columbia University, and NY State Psychiatric Institute, New York, NY, New York, United States 3 Department of Psychiatry, University of Pittsburgh, Pittsburgh, PA, Pittsburgh, United States 4 Department of Psychiatry, Johns Hopkins University, Baltimore, MD, Baltimore, United States 5 Department of Psychiatry and Mental Health CRC, University Iowa, Iowa, United States 6 Department of Psychiatry, Rush University Medical Center, Chicago, IL, Chicago, United States 7 Department of Psychiatry and Center for Neurobiology and Behavior, University of Pennsylvania, Philadelphia, PA., Philadelphia, United States The GenRED (Genetics of Recurrent Early-Onset Depression) project has collected 680 families containing 927 independent affected sibling pairs with the early-onset recurrent major depression phenotype, and a total of 971 ASPs when siblings with bipolar-II disorder are counted as affected. The full sample includes 1,791 interviewed MDD cases with DNA available, and an additional 47 BP-II relatives. DNA is available from both parents in 89 families, from one parent in 212, and from neither parents in 379 families. Permanent cell lines have been created at the NIMH Cell Repository at Rutgers University, and DNA specimens and blinded clinical data will be made available to qualified scientists by the NIMH Human Genetics Initiative program (http:// zork.wustl.edu/nimh/NIMH_initiative/NIMH_initiative_link.html) in mid-2004. Linkage analysis of the first 297 families (MDD phenotype only) produced genome-wide significant evidence for linkage on chromosome 15q (AJHG 74:1154–1167, 2004). Genotyping of the remaining families has been completed, and linkage analyses are

Abstracts currently in progress. Data will be presented for the primary allelesharing linkage analysis of the MDD phenotype (ALLEGRO), and for secondary analyses of the broader phenotype and of MDD when taking additional covariates (sex, clinical factor scores, age at onset) into account.

O1.7 PHARMACOGENOMICS AND PLACEBO RESPONSE: RESULTS FROM RECENT MAJOR DEPRESSION TRIALS AT PFIZER Sakul H Director and Sitehead, Clinical Pharmacogenomics, Pfizer Global R&D, Groton, United States Although the development of antidepressant therapeutics represents an important area of public health impact, the high rate of failed clinical trials in depression, even in the case of proven drugs, remains a major concern. The measurement of individual response to active treatment is complex and, coupled with high placebo response rates, has created the need for multiple large expensive clinical trials to evaluate new compounds for efficacy. There has been growing evidence that the insertion/deletion polymorphism, 5HTTLPR, located in the gene encoding the serotonin transporter, is associated with response to SSRIs, but to date there has been little evidence to suggest its role in placebo response in large clinical trials. We examined the effect of the 5HTTLPR polymorphism on both active and placebo treatment response rates in subjects with major depressive disorder. This presentation will summarize our findings from six different industry-sponsored, multi-center, and randomized clinical trials. While our findings concentrate on the relationship between 5HTTLPR polymorphisms and placebo response rates, trends for onset of response to SSRI treatment based on 5HTTLPR genotype will also be presented.

O1.8 POLYMORPHISMS IN FKBP5, A CO-CHAPERONE OF THE GLUCOCORTICOID RECEPTOR ARE ASSOCIATED WITH RESPONSE TO ANTIDEPRESSANT DRUGS Binder EB,1 Salyakina D,1 Wochnik G,1 Ising M,1 Pu¨tz B,1 Kern N,1 Lucae S,1 Mueller JC,2 Lohmussaar E,2 Meitinger T,2 Baghai T,3 Bondy B,3 Rupprecht R,3 Uhr M,1 Rein T,1 Holsboer F,1 and Mu¨ller-Myhsok B1 1 Max-Planck Institute of Psychiatry, Munich, Germany 2 Institute for Human Genetics, Technical University and GSF Research Centre, Neuherberg, Germany, 3 Department of Psychiatry, Ludwig-Maximilians-University, Munich, Germany It is hypothesized that antidepressants, independently from their initial mode of action, might all target a common final signaling pathway. Due to consistent dysregulations of stress hormones in the brain and peripheral circulation of depressed patients, the hypothalamic-pituitary-adrenal (HPA) system is considered a strong candidate. We previously reported a strong association of single nucleotide polymorphisms (SNPs) in FKBP5, a glucorticoid receptor (GR)regulating co-chaperone of hsp90 with response to antidepressant drugs. By fine-mapping about 300kb around this gene, we could show that only SNPs located within the major linkage dysequilibrium block containing FKBP5 were highly associated with response to antidepressant treatment (smallest P-value ¼ 0.00003). We genotyped the three SNPs with the strongest association in our sample (N ¼ 233) in a replication sample (N ¼ 85) recruited at three different hospitals in Germany. Two of the three SNPs showed a significant association with response to antidepressant treatment in the replication sample. In both samples, the association of response reflects an overrepresentation of homozygotes of the minor allele of the associated SNPs among the responders. Homozygocity for the minor alleles was also correlated with a higher protein expression of FKBP5 in lymphocytes. Previous studies have described that increased expression of FKBP5 lead to an enhanced GR resistance. The association of FKBP5 polymorphisms with response to antidepressant drugs appears to be independent of the primary pharmacological profile of antidepressants, as similar associations were obtained when dividing patients into three groups treated with different classes of antidepressants (selective serotonin reuptake inhibitors, tricyclic antidepressants and mirtazapine, a compound targeting serotonin 1A and alpha2A adrenergic receptors). These

13

results support that FKBP5 may be involved in the common final signaling pathway of antidepressant drugs.

O2.1 IDENTIFICATION OF GENES INVOLVED IN PREPULSE INHIBITION, AN ENDOPHENOTYPE OF SCHIZOPHRENIA, IN CONSOMIC MOUSE STRAINS USING AN INNOVATIVE HAPLOTYPE AND EXPRESSION MAPPING APPROACH Petryshen TL,1 Rockwell G,1 Aldinger KA,1 Kirby AN,2 Wade CM,2 Waggoner SG,1 Daly MJ,2 and Sklar P3 1 Broad Institute of MIT and Harvard, Cambridge, United States 2 Whitehead Institute for Biomedical Research, Cambridge, United States 3 Broad Institute of MIT and Harvard, and Massachusetts General Hospital, Harvard Medical School, Charlestown, United States Prepulse inhibition of startle (PPI) is a neurophysiological trait that is impaired in schizophrenia patients and varies significantly among inbred mouse strains. PPI is considered an endophenotype of schizophrenia based on extensive evidence for PPI deficits in patients and in unaffected family members. We are studying mouse PPI to identify genes involved in this trait, which will then be translated into schizophrenia association studies to determine whether these genes are also risk factors for schizophrenia. We are utilizing chromosome substitution mouse strains (CSS), also known as consomic strains. Each CSS has a donor A/J chromosome substituting for the corresponding chromosome in a host C57BL/6J (‘‘B6’’) strain, thus elegantly partitioning the genome into a panel of chromosome-specific strains. We are able to rapidly identify chromosomes harboring PPI genes by identifying CSSs that differ in PPI from the parental B6 strain. To date, we have determined that CSS-11 and CSS-16 have significantly elevated PPI compared to B6 (P ¼ 4  106 and P ¼ 0001, respectively), whereas the PPI phenotypes of CSS-2 and CSS-18 do not significantly differ from that of B6. These data indicate the presence of a PPI gene(s) on chromosomes 11 and 16, but not on chromosomes 2 and 18. Genetic intercross mapping of CSS-11 is currently underway to refine the chromosome 11 PPI locus. Intercross mapping of CSS-16 has been completed and has identified two highly significant quantitative trait loci (QTLs) with LOD ¼ 3.8 and LOD ¼ 4.6 (significant LOD ¼ 2.7) that exhibit distinct modes of inheritance (heterozygote only and recessive). We are currently fine-mapping each QTL to refine the location, and are performing backcrosses with mice segregating each locus to thoroughly characterize their independent effects. Since the causal PPI genes underlying each of the chromosome 16 QTLs may be cisacting (a variant in the gene alters expression of the gene itself), we have performed microarray expression studies of CSS-16 and B6 medial prefrontal cortex, a brain region known to regulate PPI, to determine transcript levels in each strain. We have identified several genes within ancestrally divergent segments in each QTL that have significantly altered expression between CSS-16 and B6. These genes are strong candidates for the PPI gene underlying each locus, and are being investigated for functional evidence that would suggest a role in PPI.

O2.2 WHAT IS FAMILIAL ABOUT FAMILIAL BIPOLAR DISORDER, PART II: PSYCHOSOCIAL FUNCTIONING AND COMORBID CONDITIONS ARE AT LEAST AS FAMILIAL AS CLINICAL FEATURES Schulze TG,1 Ohlraun S,1 DePaulo JR,3 Maier W,4 Rietschel M,1 and McMahon FJ2 1 Division of Genetic Epidemiology, Central Institute of Mental Health, Mannheim, Germany 2 Genetic Basis of Mood and Anxiety Disorders, NIMH, NIH, Bethesda (MD), United States 3 Department of Psychiatry, The Johns Hopkins University, Baltimore (MD), United States 4 Department of Psychiatry, University of Bonn, Bonn, Germany Progress in mapping genes that play a role in bipolar affective disorder (BPAD) has been hampered by definitions of the phenotype that may be heterogeneous or correspond only weakly to the underlying biology. We have set out to identify the most promising phenotypic variables for use in linkage and association mapping of BPAD. The first task is to reduce the amount of available clinical data to a reasonable set of items

14

Abstracts

for statistically robust analyses. Familiality (i.e., resemblance among relatives) is a good criterion for such data reduction. While everything familial is not genetic, variables that are not familial are unlikely to be genetically informative. We used 2 large samples of BPAD families that were originally ascertained for genetic linkage studies: the Johns Hopkins/Dana Pedigree series (n ¼ 197 families, 1271 individuals) and the University of Bonn pedigrees (n ¼ 76 families, 901 individuals). We investigated 3 types of variables: (1) clinical symptoms, (2) comorbid conditions, and (3) variables reflecting psychosocial functioning, all collected by retrospective self-report in a semistructured clinical interview. Clinical features are scored only for those with a diagnosis of BPAD. The other variables are available for all individuals, regardless of affection status. Familiality was assessed using a mixed-effects logistic regression approach that takes into account the dependence among relatives. Among clinical features, episode frequency and psychotic features were strongly familial (P < 0.0001, intraclass correlation [ICC]>0.2), diagnostic subtype and age at onset were moderately familial (P < 0.05, ICC>0.05), and individual symptoms of mania and depression were not familial. Among comorbid conditions, substance abuse and alcoholism were strongly familial, while panic disorder and suicidal behavior were moderately familial. Among the functional variables we tested, ‘‘social relations’’ was strongly familial, while employment status and healthiest level of functioning were moderately familial. Homogenous, biologically meaningful subsets of BPAD would greatly enhance efforts at gene mapping. Such subsets should be defined by phenotypic variables showing strong familiality. Our results show that psychosocial functioning and comorbid conditions are at least as familial as clinical features of affective episodes in families with BPAD. Genotypephenotype correlation studies in BPAD may need to look beyond classical clinical variables to discover genetically more homogeneous subsets for use in linkage and association studies.

O2.3 EVIDENCE FOR A ROLE OF THE DYSBINDIN GENE IN NEUROCOGNITIVE FUNCTION Burdick KE,1 Funke B,2 Finn CT,2 Dillon K,1 Bates J,1 Kucherlapati R,2 and Malhotra AK1 1 The Zucker Hillside Hospital-LIJMC, Glen Oaks, United States 2 Harvard Partners Center for Genetics and Genomics, Boston, United States Dysbindin (DTNBP1) is an evolutionarily conserved protein, which binds alpha and beta dystrobrevin, components of the dystrophin complex (DPC) in muscle and brain. Dystrophin is known to be defective in a group of patients with Duchenne muscular dystrophy (DMD), a common lethal, chromosome X-linked disease. Overlapping phenotypes are seen in DMD and schizophrenia, including ocular control abnormalities and cognitive dysfunction, suggesting that the DPC might be involved in the etiology of both DMD and schizophrenia. In fact, recent evidence from our group has replicated previous reports implicating the gene for dysbindin in the etiology of schizophrenia. Patients with DMD have been shown to have specific cognitive impairment, raising the possibility that the cognitive abnormalities are attributable to synaptic dysfunction associated with deficits in brain dystrophin molecules. However, the effects of DTNBP1 have not been systematically studied across multiple domains of cognition in a healthy control sample. 154 healthy Caucasian volunteers (Age ¼ 53.6  17.5, 60.3% Female) were genotyped across a cassette of single nucleotide polymorphisms (SNPs) within the DTNBP1 gene, and administered a cognitive battery assessing multiple domains of cognition. We detected significant associations between DNTBP1 genotype and specific domains of neurocognition. DNTBP1 1578 C/T genotype was associated with poorer performance on attention measures as compared with C/C genotype (Digit SpanTotal, P ¼ .001, Digit Span Back, P ¼ .006). In addition, DTNBP1 1325 C/C genotype, compared to DTNBP1 1325 C/T genotype, was associated with poorer performance on measures of verbal learning (CVLT Trials 1–5, P ¼ .011), sustained attention (CPT d’, P ¼ .004), and executive functioning (Trails B, P ¼ .042). Finally, subjects with DTNBP1 1635 G/ A and G/G genotypes performed significantly worse than those with A/ A genotype on measures of attention (Digit Span Total, P ¼ .007, Digit Span Back, P ¼ .030). There were no significant differences between groups on estimates of premorbid overall intellectual functioning. These data suggest that the effect of DTNBP1 genetic variation may be associated with specific domains of neurocognitive functioning in

healthy subjects. Studies are underway to assess the relationship of DTNBP1 to cognitive function in patients with schizophrenia. O2.4 ENDOPHENOTYPES OF ATTENTION DEFICIT HYPERACTIVITY DISORDER AND THE DOPAMINE SYSTEM Maher BS, Ferrell RE, Devlin B, Kirillova GP, Tarter RE, and Vanyukov MM University of Pittsburgh, Pittsburgh, United States The risk for attention deficit hyperactivity disorder (ADHD) and its symptom-based activity and attention dimensional measures have been shown to be highly heritable. ADHD is also frequently premorbid to substance use disorders (SUD). The dopaminergic system (DS) has been implicated in the etiology of both disorders, and the genes encoding its components may contribute to the high heritability of ADHD and SUD liabilities and the association between them. Using the family-based association test (FBAT), we investigated the association between single nucleotide polymorphisms (SNPs) in six DS genes and a battery of neuropsychological tests in a sample of 267 10–12 year old males and their parents ascertained in a SUD research program. The test battery was designed to measure mental processes related to ADHD liability, and included several commonly used tasks. The Stroop Color Word Task measures attentional control directed to a response requirement (e.g. color naming) while suppressing a more dominant response (word rendering). The Vigilance task is a continuous performance test that provides measures of attention and impulsiveness. The qualitative score of the Porteus Maze test provides a measure of planning ability and impulse control. The Actigraph measures total motor activity over a two day period during which the subjects undergo a standardized protocol. A total of 19 SNPs in the dopamine transporter (DAT1) and each of the five dopamine receptors (DRD1-DRD5) genes were genotyped. Robust associations were found between the 521 C/T SNP of the DRD4 gene and each of the measures: a modest association for mean total motor activity (P ¼ 0.02) and a strong association for the Porteus Maze impulsive errors score (P ¼ 0.0019), Stroop Interference (P ¼ 0.00016), and Vigilance Task Omission Error Rate (P ¼ 0.00006). No strong associations were observed for the SNPs in the other genes.

O2.5 A NOVEL IMPRINTED LOCUS ON CHROMOSOME 2P12 ASSOCIATED WITH RELATIVE HAND SKILL IN HUMANS Francks C,1 Maegawa S,2 McAuley EZ,1 Richardson AJ,3 Stein JF,3 Oshimura M,2 and Monaco AP1 1 Wellcome Trust Centre for Human Genetics, University of Oxford, Oxford, United Kingdom 2 Department of Molecular and Cell Genetics, Tottori University, Tottori, Japan 3 Department of Physiology, University of Oxford, Oxford, United Kingdom Cerebral functional asymmetry is a key feature of complex human cognition. A range of neuropsychiatric conditions, including schizophrenia, autism, dyslexia, and language impairment, show evidence for associations with non-right-handedness and/or abnormal asymmetrical brain structure/function. Data from twins suggest that many of these associations are genetically mediated in part. We have previously used linkage analysis to localize a quantitative-trait-locus for relative left-versus-right hand skill to a region of human chromosome 2 (2p12q11, Francks et al., Am J Hum Genet 2002, 2003). We have shown previously that only paternal inheritance of 2p12-q11 is linked to relative hand skill in our samples (linkage P ¼ 0.000004, Francks et al. HMG 2003). We have also shown previously, in a separate sample of schizophrenic sibling pairs (244 pairs), that 2p12-q11 is paternally linked to schizophrenia (linkage P ¼ 0.000002, Francks et al. HMG 2003). We have now identified strong paternal-specific associations, with relative hand skill, of polymorphisms in the region of two brainexpressed candidate genes on 2p12 (haplotype association P ¼ 0.0004). We have also shown paternal-only expression and differential methylation of at least one gene at this locus in human brain tissue, lymphocytes, and mouse-human hybrid cells. No imprinted genes were known previously on the short arm of chromosome 2. Since less than 1% of genes are imprinted, our findings indicate that we may have identified the locus influencing relative hand skill on 2p12-q11. We are currently testing the locus in direct relation to schizophrenia by mutation screening, genetic association, epigenetic methylation, and quantitative expression analysis.

Abstracts O2.6 NEUROPSYCHOLOGICAL FUNCTION AT AGE 13 AND LATER SCHIZOPHRENIFORM DISORDER: A LONGITUDINAL BIRTH COHORT STUDY Cannon M,1 Moffitt TE,2 Caspi A,2 Harrington H-L,3 Murray RM,5 and Poulton R4 1 Royal College of Surgeons in Ireland, Dublin, Ireland 2 Social, Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, London, United Kingdom 3 Department of Psychology, University of Wisconsin, Madison, United States 4 Dunedin School of Medicine, Dunedin, New Zealand 5 Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom There is compelling evidence that individuals with schizophrenia exhibit impairment on a wide range of neuropsychological tasks notably those involving attention, language and memory. However it is not known whether such deficits predate the diagnosis. There is little information on specific neuropsychological functioning in childhood or adolescence in individuals who later develop psychosis. The Dunedin Multidisciplinary Health and Development Study is an unselected prospective birth cohort of 1037 individuals born in Dunedin, New Zealand between April 1 1972 and March 31, 1973. Study members have participated in ten assessments between ages 3 and 26 including a neuropsychological test battery at age 13. Of the 850 cohort members who participated in the study at age 13, complete neuropsychological assessment data is available on 710 subjects. Psychiatric diagnostic interviews conducted at age 26 revealed that 21 of these individuals fulfilled diagnostic criteria for schizophreniform disorder, 10 for a mania and 198 for depressive disorder or anxiety disorder. Study members with a subsequent diagnosis of schizophreniform disorder performed significantly worse than the other groups at age 13 on the Trails B test (time to completion), Trails B-A, the Grooved Pegboard test (both left and right hands),Verbal Fluency and the Arithmetic and Coding subtypes of the Weschler Intelligence Scale for ChildrenRevised. No significant differences were noted between the groups on the Rey Osterreith Copy score, the Wisconsin Card Sort Test or the Rey Auditory Verbal Learning Test. These results indicate that deficits in executive functioning/attention, and psychomotor function precede the onset of schizophrenia in a representative birth cohort and may represent an endophenotype for the disorder.

O2.7 UFD1L—277A/G PROMOTER POLYMORPHISM IN SCHIZOPHRENIA: COGNITIVE AND GREY MATTER MORPHOMETRIC CHARACTERISTICS Spalletta G,1 Bonaviri G,4 Tomaiuolo F,2 De Luca A,3 Sinibaldi L,3 Dallapiccola B,3 and Caltagirone C1 1 Department of Neuroscience, Tor Vergata University and IRCCS Fondazione Santa Lucia, Rome, Italy 2 IRCCS Fondazione Santa Lucia, Rome, Italy 3 IRCCS-CSS, San Giovanni Rotondo and CSS-Mendel Institute and Department of Experimental Medicine and Pathology, La Sapienza University, Rome, Italy 4 ASL Frosinone, Frosinone, Italy Chromosome 22q11.2 deletion is associated with Di George syndrome and velo-cardio-facial syndrome (VCFS). The prevalence of psychiatric disorders is high in VCFS and this can be explained by the presence, in the deleted region, of many interesting candidate genes, including UFD1L, a developmentally expressed ubiquitination gene. UFD1L is expressed in the medial telencephalon during the mouse development and in humans it is expressed starting from week 10 of gestational age. Recently, our research group described an UFD1L promoter polymorphism in the noncoding region upstream the first exon, at 277A/G, as a susceptibility gene for schizophrenia (SCZ). In particular, SCZ was associated with genotypes containing the 277A allele. This polymorphism has a functional role because in vitro it influences the transcriptional activity of UFD1L gene promoter with a significant upregulation of the gene expression by constructs containing the— 277A allele compared to the 277G variant. Here we describe cognitive and morphometric (VBM) correlates of patients with SCID-P–DSM-IV SCZ carrying the 277A allele (n ¼ 30) compared to patients with the 277G/G genotype (n ¼ 16). The Mental Deterioration Battery was used to provide information about the functionality of different areas of cognition: language, verbal memory (immediate and delayed recall), visual memory, logical reasoning, and constructional praxis. Magnetic

15

resonance T1-weighted images were obtained with a Siemens 1.5 tesla MR system. The MPRAGE brain volume images underwent a nonuniformity correction. Subsequently, MR images were transformed linearly to standardised stereotaxic space. The tissue-classification algorithm (INSECT) was applied to classify each voxel into one of three classes, namely grey matter (GM), white matter or cerebral-spinal fluid. A binary volume consisting of GM voxels was extracted from the classified image and smoothed using a Gaussian smoothing kernel of 10-mm full-width at half-maximum. Using a t test, we compared the cognitive level and the GM density map of patients carrying the 277A allele and patients with 277G/G genotype. We found that patients carrying the 277A allele were more impaired in verbal memory (delayed recall) (P < 0.02). Furthermore, the GM signal intensity of the left amygdala (x ¼ 28.1, y ¼ 2.7, z ¼ 20.2) was significantly lower (P < 0.001) in patients carrying the 277A allele than in patients with the 277G/G genotype. Considering that: 1) verbal memory deficit is a marker of risk for schizophrenia, 2) amygdala may be of pivotal importance in the pathogenesis of schizophrenia and 3) delayed verbal memory activity is correlated with amygdala-anterior hippocampus volume, these results are of importance for studies aiming at investigating the prenatal-developmental roots of schizophrenia. O2.8 A QUANTITATIVE TRAIT MEASURING COGNITIVE ANXIETY AND REACTIVITY TO CHANGE SHOWS LINKAGE TO LOCI ON CHROMOSOMES 15 AND 19 IN A GENOME-WIDE LINKAGE ANALYSIS OF 28 BIPOLAR FAMILIES Evans L,1 Akiskal H,2 Shekhtman T,1 Gaucher M,1 and Kelsoe J2 1 University of California, San Diego, United States 2 UCSD & VA Healthcare System, San Diego, United States Despite the preponderance of evidence for the genetic transmission of bipolar disorder, finding bipolar genes has proven difficult. One potential reason there have been problems with finding bipolar genes is phenotype definition. Previously, to address this problem, new quantitative traits were derived from the TEMPS-A and TCI-125 and a factor analysis of the two temperament and character scales together based on the item responses of 383 subjects from 85 bipolar families and 63 controls. Each of the temperament and character subscales and extracted factors were compared across four subject groups: bipolar subjects, major depressive disorder subjects, unaffected relatives, and controls. Several of these traits showed significant differences across groups when using two-way ANOVA with subject group and sex being the two factors. However, in the post-hoc Tukey analyses, the first extracted factor was the only trait examined that showed a significant difference between unaffected relatives and controls, one of the hallmarks of a good quantitative trait. This new quantitative trait is characterized by cognitive anxiety and reactivity to change. Based on this previous body of research we decided to use this anxious-reactive extacted factor as a quantitative trait in genome-wide linkage analysis in 151 subjects from 28 bipolar families. Using MERLIN’s variance components nonparametric multipoint linkage analysis and gene dropping simulations to derive P-values, a maximum LOD score of 1.83 (chromosome-wide simulated P ¼ 3.9  1017) on chromosome 19 at 109.38 cM was detected. The second highest peak was found on chromosome 15 at 44.33 cM (LOD ¼ 1.75, chromosome-wide simulated P ¼ 2.4  1014) also using the extracted anxious/reactive trait. Both of these linkage peaks have interesting genes located under them that should be further explored in relation to bipolar disorder. These analyses confirm the idea that anxiety is an important feature of bipolar disorder and may identify a genetically distinct dimension of illness. It also validates the use of temperament measures as quantitative traits in bipolar disorder.

O3.1 GENETIC VARIATION IN THE RENIN-ANGIOTENSIN SYSTEM AND RISK OF ALZHEIMER’S DISEASE Gustafson D, Eriksson E, Blennow K, and Skoog I 1 Go¨teborg University, Go¨teborg, Sweden 2 Medical College of Wisconsin, Milwaukee, United States Vascular factors, such as high blood pressure and obesity, are risk factors for Alzheimer’s disease (AD). The renin-angiotensin system (RAS) plays a fundamental role in blood pressure regulation and vascular integrity, and therefore may be involved in AD pathogenesis.

16

Abstracts

Three major proteins involved in the RAS include angiotensinogen (AGT), angiotensin-converting enzyme (ACE), and angiotensin receptor 1 (AGTR1). Objective: We examined the relationship between AGT, ACE, and AGTR1 gene variants and dementia occurrence over a 32-year follow-up period. A population-based sample of 594 Swedish women aged 38–60 years in 1968 was followed for 32 years. Neuropsychiatric, laboratory, and other assessments were conducted in 1968, 1974, 1980, 1992, and 2000. Analyses of three gene variants related to the RAS were performed, and included: AGTR1 A1166C, AGT T235C, and ACE I/D. Dementia from 1992–2000 was diagnosed according to DSM-III-R. AD was diagnosed according to NINCDSADRDA-criteria and vascular dementia (VAD) according to NINDSAIREN criteria. Information on those lost to follow-up was collected from case records, the hospital linkage system, and death certificates. Logistic regression analyses, adjusted for age, were used to evaluate the relationship between RAS genotypes and dementia. Between 1968 and 2000, 55 women became demented (age of onset, 64–89 years), 28 of whom were diagnosed with AD. The AGTR1 CC and ACE II alleles were significantly related to dementia risk. The odds of dementia were 4.44 (95% CI 1.98–9.92) among women homozygous for AGTR1 CC allele, and 2.22 (95% CI 1.19–4.12) among women homozygous for ACE II. In particular, the odds of having an AD diagnosis were almost 5-fold higher (OR 4.92, 95% CI 1.79–13.52) among women homozygous for ATR1 CC. AGT was not related to dementia occurrence. Gene variants related to the RAS may be related to dementia occurrence in women, and help to explain the vascular etiology of AD. (Supported by the Alzheimer’s Association Zenith Award, ZEN-01-3151).

O3.2 FAMILY-BASED ASSOCIATION OF UBIQUILIN 1 (UBQLN1) AND ALZHEIMER’S DISEASE IN TWO INDEPENDENT SAMPLES Bertram L,1 Hiltunen M,1 Parkinson M,1 Mullin K,1 Menon R,1 Elliott KJ,2 Wagner SL,2 Blacker D,1 Becker KD,2 and Tanzi RE1 1 Massachusetts General Hospital, Charlestown, United States 2 Neurogenetics, Inc., La Jolla, United States Several studies have implicated chromosome 9 as harboring at least one major Alzheimer’s disease (AD) locus. Our own recently completed full genome screen on multiplex families of the NIMH AD Genetics Initiative Study sample (Blacker et al., Hum Mol Genet, 2003) maps this locus to 9q22, between 80 and 105 Mb. This study was aimed at the identification and characterization of the underlying putative AD gene(s) in this chromosomal interval. Candidate genes are tested for genetic association using single-locus and haplotype analyses in the NIMH AD Genetics Initiative study sample, consisting of 1439 individuals from 437 multiplex AD families. Positive findings were subsequently tested in an independent collection of AD families (the Consortium on Alzheimer’s Genetics [CAG]), currently 344 subjects from 150 sibships discordant for AD. Our analyses reveal significant evidence of association with adjacent single-nucleotide polymorphisms (SNP)s in the gene encoding ubiquilin-1 (UBQLN1), located at 82 Mb. Ubiquilin 1 is involved in protein degradation, and has recently been shown to interact with presenilins 1 and 2, and to colocalize with neurofibrillary tangles in the brains of AD patients (Mah et al., J Cell Biol, 2000). After genotyping five SNPs in the NIMH cohort, we were able to delineate a risk haplotype that was significantly overtransmitted to affected individuals (P ¼ 0.02). The same haplotype was associated with disease risk in the smaller, and independently collected CAG sample (P ¼ 0.05), and showed its strongest association after combining both samples (P ¼ 0.002). Importantly, this association accounts for nearly half of the linkage signal observed on chromosome 9q22 in the NIMH families. Meanwhile, no consistent genetic association was observed with a total of 14 SNPs tested in two other compelling AD candidate genes on chromosome 9, i.e. APBA1 (encoding the amyloid beta precursor protein [APP] binding protein 1 [X11a]) and ABCA1 (encoding the ATP-binding cassette A1 transporter). Taken together, our results obtained in two differently ascertained AD family samples strongly support UBQLN1 as a novel and major putative AD gene underlying the well established linkage signal on chromosome 9q22. These data, together with the evidence from previous studies suggesting an involvement of ubiquilin 1 in protein degradation and presenilin metabolism, warrant further investigation of the role of this protein and its homologues in the neuropathogenesis of AD. Along these lines, potential effects of the AD-associated UBQLN1 SNPs and SNPhaplotypes on gene expression are currently being assessed in our laboratory.

O3.3 INVESTIGATION OF THE MICROTUBULE ASSOCIATED PROTEIN TAU LOCUS Myers AJ,1 Pittman A,2 Duckworth J,1 Fung P,1 Evans J,1 Evans W,1 Bryden L,1 Hanson M,1 Abou-Sleiman P,3 Wood NW,3 Lees A,2 de Silva R,2 and Hardy JA1 1 Laboratory of Neurogenetics, National Institute on Aging, National Institutes of Health, Bethesda, United States 2 Reta Lila Weston Institute of Neurological Studies, University College London, London, United Kingdom 3 Department of Molecular Neuroscience, Institute of Neurology, Queen Square, London, United Kingdom The group of neurodegenerative diseases collectively known as tauopathies are characterised by hallmark lesions consisting of fibrillar aggregates of the microtubule associated protein, tau. Mutations of the tau gene (MAPT) are the cause of frontotemporal dementia with parkinsonism linked to chromosome 17 (FTDP-17) giving tau a central role in neurodegenerative aetiology. The chromosomal region containing MAPT has been shown to have evolved into two major haplotypes, H1 and H2 which are defined by linkage disequilibrium (LD) between several polymorphisms over the entire MAPT gene. Studies to date suggest a complete absence of recombination between these two haplotypes. The more common haplotype (H1) is over-represented in patients with progressive supranuclear palsy (PSP) and corticobasal degeneration (CBD). By linkage disequilibrium mapping with single nucleotide polymorphisms (SNPs), we have delineated the maximum extent of the MAPT haplotype block as a region covering 2 Mb at 17q21.31. There are two possibilities as to why such a large region of linkage disequilibrium exists in this region. The first possibility is that there is selection at this locus. The second is that there is some architectural and/or sequence rearrangement of the chromosomes in this region making these two haplotypes unable to recombine. We are performing fibre FISH to investigate whether there is an inversion in this region. To look at factors affecting selection, we are typing SNPs within expressed regions of the haplotype block in a series of primates to determine the ancestral allele. We are also typing a series of individuals from the CEPH diversity panel series to examine the distribution and recombination of this locus worldwide. O3.4 RISK INTERPRETATION AND ATTITUDES TOWARDS PREDICTIVE GENETIC TESTS FOR NEUROPSYCHIATRIC DISEASES Illes F,1 Rietz C,2 Prell K,3 Rudinger G,2 Maier W,3 and Rietschel M1 1 Central Institute of Mental Health, Mannheim, Germany 2 University of Bonn, Institute of Psychology, Bonn, Germany 3 University of Bonn, Department of Psychiatry, Bonn, Germany If predictive genetic testing for neuropsychiatric disorders becomes available, the ability to accurately interpret relevant risk information will be crucial. We have therefore investigated attitudes towards predictive genetic testing for neuropsychiatric diseases, and the perception and interpretation of information regarding the probability of risk. In addition, we have assessed whether genetic counseling has an influence on attitudes and risk perception, using Alzheimer’s disease as an example. In our ongoing study, a representative sample of the German general population (n ¼ 2000) and relatives of patients with Alzheimer’s disease (n ¼ 100) as well as physicians (n ¼ 43) are being interviewed about their attitudes towards genetic tests for Alzheimer’s disease and their perception and interpretation of risk information. To assess the influence of genetic counseling on these factors, 200 people are to be questioned before, and four weeks after, they have been provided with detailed relevant information. Results show that 43% of the general population (representative sample) as well as 50% of the relatives and 30% of the physicians expressed an interest in being tested for APOE4. 34% of the general population, about 40% of the relatives and 20% of the physicians would consider it appropriate for their partner to have predictive testing. While 25% of the general population were in favor of prenatal testing only about 10% of the relatives and 2% of the physicians expressed the same opinion. Furthermore our results suggest that the majority had difficulties in interpreting risk information: in our representative sample of the general population only 34% of the probands understood the meaning of risk estimates given as percentages. Results are available at present for 65 of the group that received detailed information, and these show that, after counseling, understanding of risk increased from 49% to 61% (w2 ¼ 6.7, df ¼ 1, P ¼ .009). Results from the whole data set will be available at the end of August.

Abstracts O3.5 RESISTANCE TO ß-AMYLOID AGGREGATION AND Aß-INDUCED CELL DEATH IN MICE LACKING GD3 SYNTHASE McDonald MP, Bernardo A, Oaghere-DaSilva UB, Bruchey A, and Zhao J Vanderbilt University, Nashville, United States Previous studies have shown that gangliosides are necessary for ßamyloid (Aß) binding and aggregation in vesicle preparations. In addition, the ganglioside GD3, which is increased in Alzheimer’s disease, has been shown to be necessary for Aß-induced apoptosis. Knocking out the GD3 synthase (GD3S) gene effectively eliminates half of the major brain gangliosides, including GD3. Knockout mice lacking GD3S are viable, grossly normal, and have a normal lifespan. The present studies investigated the relationship between membrane gangliosides, Aß aggregation, and cell death, using mice null for GD3S. Primary neuronal and astrocyte cultures were established from 1-day-old mouse pups lacking GD3S, or wild-type controls. Aß binding was visualized using fluorescent-labeled antibodies to Aß after application of exogenous Aß 1–40. DNA damage was assessed using triphosphate-biotin nick end labeling (TUNEL) after 24-hr. incubation with Aß 1–42. GD3S/ mice and wild-type littermate controls were assessed on a battery of memory tests at young and old ages. In astrocytes and neurons from GD3S knockout mice, there was a genedose-dependent reduction in exogenous Aß binding, with a greater than 50% reduction in the homozygous knockout mice. Neurons from GD3S null mice were completely resistant to apoptosis after exposure to 10 mM Aß 1–42, a concentration that induced widespread cell death in wildtype neurons. Ongoing studies are addressing the role of gangliosides in Aß aggregation, apoptosis, and memory in vivo, in GD3S knockout mice cross-bred with transgenic mice over-expressing amyloid precursor protein (APP). Our results suggest that strategic reduction of gangliosides may be a novel therapeutic strategy to combat both the formation of amyloid plaques as well as the widespread cell death that afflicts Alzheimer’s patients.

O3.6 PARENT OF ORIGIN EFFECT AND LINKAGE TO CHROMOSOME 10 FOR ALZHEIMER’S DISEASE, REPLICATION IN AN INDEPENDENT SAMPLE AND GENOME-WIDE SIGNIFICANCE Avramopoulos D,1 Fallin MD,1 Perry RT,2 Wiener H,2 Watson B,2 Go RCP,2 and Bassett SS1 1 Johns Hopkins University, Baltimore, United States 2 University of Alabama at Birmingham, Birmingham, United States In a previous publication we reported the observation of a parent of origin effect in late onset Alzheimer’s disease (LOAD) that produced a significant linkage finding on the proximal long arm and pericentromeric region of chromosome 10. The same region has also been found to be linked to LOAD in previous genome scans and it has shown strong linkage to plasma A-beta42 levels. In that paper we reported evidence of linkage in our subset of the NIMH genetics initiative pedigrees that approached genome-wide significance. With our most up to date information, including genotypes on additional markers and new autopsy results, the non-parametric LOD score is 2.96 when the JHU maternal families are analyzed separately, up from 1.79 in the overall JHU sample. This rise in LOD score after stratification by parental origin is statistically significant, as we show empirically by performing multiple random stratifications (P ¼ 0.014). We now have also acquired parent of origin information from an independent subset of the NIMH genetics initiative pedigrees, those collected by the University of Alabama (UAB). Like in our sample, maternal pedigrees from the UAB set (N ¼ 21) show a statistically significant (P ¼ 0.037) increase of linkage evidence by one LOD unit in the previously defined 1-LOD interval, thus replicating the observation of a parent of origin effect. When we combine all maternal pedigrees (N ¼ 21 þ 48 ¼ 69) the nonparametric LOD score is 3.53. In order to assess the significance of this score at the genome-wide level we have performed 1,000 simulated genome scans using our family structure and marker properties and we determined an empirical genome wide significance of 0.004. Given the strong significance and the prior positive findings in the region, as well as the replication of the observed parent of origin effect in a different family set we conclude that this region is very likely to contain a gene or genes that increase the risk for LOAD through a mechanism that produces a parent of origin effect, therefore it warrants further investigation.

17

O3.7 PLEIOTROPIC GENETIC EFFECTS ON VERBAL, SPATIAL, AND FACE MEMORY Almasy L,1 Gur RC,2 Pogue-Geile M,3 Blangero J,1 Nimgaonkar V,3 and Gur RE2 1 Southwest Foundation for Biomedical Research, San Antonio, United States 2 University of Pennsylvania, Philadelphia, United States 3 University of Pittsburgh, Pittsburgh, United States Patients with schizophrenia have been shown to have impairments in performance on a variety of neurocognitive tests, as compared to healthy controls. These include measures of verbal, spatial, and face memory. Their unaffected first degree relatives show similar, but milder, impairments for the same traits, suggesting that these measures may represent endophenotypes correlated with underlying genetic liability to schizophrenia. We are studying the genetics of these neurocognitive endophenotypes in families ascertained through relative pairs with DSM-IV schizophrenia. To date, our analyses include over 250 individuals in 31 nuclear and extended families, with relatives as distantly related as second cousins. Verbal memory was assessed using the Penn Word Memory Test, face memory with the Penn Face Memory Test, and spatial memory with the Visual Object Learning Test. All tests were administered using a standardized, computerized platform. Although data were collected on both accuracy and response time, bivariate genetic analyses were conducted only with accuracy measures as the response times showed heritabilities not significantly different from zero. Bivariate variance component methods, implemented in SOLAR, were used to decompose the phenotypic correlations among memory test scores into genetic and environmental components. Likelihood ratio tests were used to assess whether these correlations were significantly different from zero. Heritabilities for accuracy scores on all three memory traits were statistically significant (P < 0.01) and ranged from 0.36 to 0.50. Genetic correlations were 0.62 (P ¼ 0.01) for spatial and verbal memory, 0.94 (P ¼ 5  106) for facial and verbal, and 1 (P ¼ 0.00008) for spatial and facial. No significant environmental correlations were observed between these traits. These results suggest that there is substantial overlap in the genes that influence variation in accuracy on spatial, verbal, and face memory tasks and that bivariate linkage analyses may therefore improve power to localize genes influencing these phenotypes.

O3.8 DIRECT GENOMIC PCR SEQUENCING OF THE HIGH AFFINITY THIAMINE TRANSPORTER (SLC19A2) GENE ON CHROMOSOME 1 AND OF THE OGDH GENE ON CHROMOSOME 7 IN ALCOHOLICS WITH WERNICKE KORSAKOFF SYNDROME (WKS) Guerrini I,1 Thomson AD,1 Cook CCH,2 McQuillan A,1 Sharma V,1 Kopelman M,3 Reynolds G,4 Jahaur P,4 and Gurling HMD1 1 University College London, London, United Kingdom 2 Chad College, Durham, United Kingdom 3 Kings College, London, United Kingdom 4 University of Glasgow, Glasgow, United Kingdom The Wernicke Korsakoff Syndrome (WKS) is a complication of alcohol misuse caused by an absolute or relative deficiency of dietary vitamin B1 (thiamine). Post mortem studies have shown that 1.5% of all random general hospital post mortem cases specifically examined for damage in the brain regions affected by WKS have the neuropathological lesions of WKS (Harper et al., 2003, Prog Neuropsychopharmacol Biol Psychiatry 27:951–61). However in alcoholic patients this rate is said to be much higher at 12.5% (Torvik et al., 1982, J Neurol Sci 56:233–48). Evidence from animal studies report that alcohol intake and thiamine deficiency interact to increase cognitive impairment (Homewood et al., 1999 Alcohol 19:75–84). Neuropsychological studies show that 50–75% of detoxified alcoholics develop cognitive impairment, mostly memory deficits, which may be due to thiamine deficiency. Other factors, such as diet, may also be involved in the pathogenesis of this condition (Heap et al., 2002 Psychiatr Genet 12:217–24). It has been hypothesized that there is a genetic susceptibility in the pathogenesis of Wernicke Korsakoff Syndrome (WKS). Therefore thiamine transporter proteins or enzymes that require thiamine as a co-factor are candidate genes for WKS. We report the screening of the high affinity thiamine transporter gene, SLC19A2 which is localised to chromosome 1q23.2-q23.3, and the screening of the oxoglutarate dehydrogenase gene, OGDH, mapped to 7p13-p11.2, in 32 cases of alcoholism with WKS and in 32 healthy

18

Abstracts

controls. We carried out semi-automated direct genomic PCR sequencing in 30 alcoholics patients affected by WKS and in 30 super-normal controls We sequenced a) the promoter region, the six exons and the 30 untranslated region (UTR) region of the SLC19A2 gene and b) the promoter region and exon 9 of the OGDH gene. We identified three genetic variants in the 30 UTR region of the SLC19A2 these were: (1) A single base pair insertion/deletion in one patient with WKS which was not found in 30 controls. (2) A base pair substitution (T/C) found in one WKS case but not in the normal controls. (3) A second base pair substitution (T/C) which was found in three WKS subjects and five controls. A base pair change (T/C) was detected in 5 patients and 5 healthy controls at position 42 of the promoter region of the OGDH gen. This genetic variant has not been described before in the common polymorphism databases. The small sample of 30 cases and 30 controls that we have studied precludes any firm inferences about the possible functional significance of these genetic variants in WKS. The fact that an insertion/deletion and a base pair substitution were only found in cases warrants further investigation in larger samples of WKS cases. Funded by the Brain Damage Research Trust and by the Neuroscience Research Charitable Trust.

O4.1 THE CHROMOSOME 5 GABA A RECEPTOR GENES GABRA1 AND GABRB2 CONFER RISK OF SCHIZOPHRENIA AND ARE CORRELATED WITH ALTERED GABA PATHWAY TRANSCRIPT LEVELS Petryshen TL,1 Kirby AN,2 Pato CN,3 Tahl AR,1 Middleton FA,4 Rockwell G,1 Schwab SG,5 Wildenauer DB,5 Aldinger KA,1 Waggoner SG,1 Morley CP,4 McGann L,3 Gentile KL,3 Medeiros H,4 Azevedo MH,4 Daly MJ,2 Pato MT,3 and Sklar P6 1 Broad Institute of MIT and Harvard, Cambridge, United States 2 Whitehead Institute for Biomedical Research, Cambridge, United States 3 Veterans Administration Medical Center, Washington, United States 4 State University of New York, Syracuse, United States 5 University of Western Australia, Perth, Australia 6 Broad Institute of MIT and Harvard, and Massachusetts General Hospital, Harvard Medical School, Charlestown, United States We previously reported linkage between schizophrenia and chromosome 5q31-q35 in a genome-wide scan of families of Portuguese descent (Sklar et al., Mol Psychiatry 2004). Our finding supports a published meta-analysis of 20 other genome-wide scans that identified 5q23.2-q34 as the second most significant schizophrenia susceptibility locus in the genome. Based on evidence for GABA-ergic system defects in the frontal cortex of schizophrenia patients, we performed a haplotypebased association study of the GABA A receptor genes GABRA1, GABRA6, GABRB1, GABRG2, and GABRP located in a cluster within 4–13 Mb of our linkage peak to determine whether one or more of these genes confers risk of schizophrenia. We genotyped 132 SNPs spanning these five genes in a sample of 111 parent-proband trios and 320 unrelated cases/240 controls of Portuguese descent. We identified significantly altered frequencies of two GABRA1 haplotypes in patients: an over-represented ‘risk’ haplotype (P ¼ 0.0019) and an under-represented ‘protective’ haplotype (P ¼ 0.027). Both of these associations were replicated in an independent German sample of 240 parent-proband trios (P ¼ 0.037 and 0.0036, respectively). Our data also indicated a higher frequency in Portuguese and German patients of a GABRB2 risk haplotype (P ¼ 0.10) that was previously reported to be associated in a Chinese schizophrenia sample. We also investigated whether the associated GABRA1 and GABRB2 haplotypes were related to altered transcript levels of the 5q GABA cluster genes, and of genes that are co-expressed with the GABA cluster genes. Correlation of haplotypes with microarray transcript expression in peripheral leukocytes of 76 individuals from our Portuguese linkage families found intriguing relationships between the associated haplotypes and gene expression. Individuals carrying the GABRA1 protective haplotype had significantly upregulated expression of genes that are coexpressed with GABRA1 (P ¼ 0.001), as well as genes that are coexpressed with GABRB2 (P ¼ 0.025). Conversely, individuals carrying the GABRB2 risk haplotype had significant downregulation of GABRA1 co-expressed genes (P ¼ 0.045). Thus, schizophrenia-associated haplotypes of GABRA1 and GABRB2 appear to have reciprocal effects on the expression of genes related to the GABA-ergic system. Taken together, these results provide substantial evidence that the GABA A receptor genes GABRA1 and GABRB2 on chromosome 5q confer risk for schizophrenia.

O4.2 INTERACTION BETWEEN NEUREGULIN 1 AND ITS RECEPTOR ERBB4 INCREASES SUSCEPTIBILITY TO SCHIZOPHRENIA Norton N,1 Moskvina V,1 Morris DW,2 Bray NJ,1 Zammit S,1 Williams NM,1 Williams HJ,1 Preece AC,1 Dwyer SL,1 Wilkinson JC,1 Kirov G,1 Buckland P,1 Waddington JL,3 Gill M,2 Corvin AP,2 Owen MJ,1 and O’Donovan MC1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Neuropsychiatric Genetics Group, Department of Psychiatry, Trinity College, Dublin, Ireland 3 Stanley Research Unit, St. Davnet’s Hospital, Monaghan, Ireland There is now strong evidence that Neuregulin 1 (NRG1) is a susceptibility gene for schizophrenia. NRG1 mediates some of its effects through the tyrosine kinase receptor erbB4, and analysis of gene knock-out animals suggests that the functional interaction of NRG1 and erbB4 mediates behaviours that may model some aspects of the schizophrenia phenotype in mice. Given these findings, we have sought evidence for association between schizophrenia and erbB4. Mutation screening of erbB4 in 14 DSMIV schizophrenics revealed 14 SNPs, none of which were nonsynonymous. Analysis of the allele frequencies of each SNP in pools of 368 DSMIV schizophrenics and 368 controls provided modest evidence for association with two of the SNPs, although individual genotyping in an extended sample of 680 cases subjects did not confirm this. However, we did find evidence for a significant interaction between the NRG1 schizophrenia risk haplotype and erbB4 (P ¼ 0.019). The NRG1 and erbB4 interacting marker was further genotyped in an independent sample of 299 cases and 645 controls from Dublin. Interaction between NRG1 and erbB4 was significant in the combined sample of 860 cases and 1243 controls, (P ¼ 0.01). Given previous evidence in our sample for increased frequency of the NRG1 haplotype in cases with a first degree relative with schizophrenia (Williams et al., 2003), we also tested the interaction in familial cases (n ¼ 158 in the combined Cardiff and Dublin samples). Our data show an increased effect size in the familial cases, OR 11.71 (CI: 1.56, 87.8), P ¼ 0.017. Pooled genotyping of public database SNPs flanking the interacting SNP identified significant differences in the Cardiff sample, which replicated with individual genotyping, P ¼ 0.019. Given the absence of changes affecting erbB4 protein structure, we postulated the existence of unknown cis-acting variants affecting erbB4 expression that may provide the basis for the interaction. Expression analysis confirmed this hypothesis although the specific variants(s) responsible remain unknown. Our data suggest that NRG1 may mediate its effects on schizophrenia susceptibility through its functional interaction with erbB4, and that genetic interaction between variants at the 2 loci increases susceptibility to schizophrenia. O4.3 POSITIVE ASSOCIATION BETWEEN SYNAPSIN II AND SCHIZOPHRENIA Chen Q,1 He G,1 Wang XY,2 Chen QY,2 Liu XM,1 Gu ZZ,3 Liu J,5 Li KQ,5 Wang SJ,6 Zhu SM,7 Feng GY,4 and He L1 1 Shanghai Jiaotong university, Shanghai, China 2 Chinese Academy of Sciences, Shanghai, China 3 Huzhou Third People Hospita, Zhejiang, China 4 Shanghai Institute of Mental Health, Shanghai, China 5 Hebei Institute of Mental Health, Hebei, China 6 Fuyang Mental Hospital, Anhui, China 7 Jilin Institute of Mental Health, Jlin, China Synapsin II encodes a neuron-specific phosphoprotein that selectively binds to small synaptic vesicles in the presynaptic nerve terminal. The expressions of mRNA and protein of synapsin II have been reported significantly reduced in the brain of schizophrenia patients. The synapsin II gene is located on 3p25, a region that has been implicated to be associated with schizophrenia by genetic linkage. All these suggest that synapsin II is a candidate gene for schizophrenia. In this work, we studied 4 markers (two single nucleotide polymorphisms (SNPs): rs308963 and rs795009, two insertion/deletion polymorphisms: rs2307981 and rs2308169) covering 144.2 kilo base pair (kb) with average interval of 38 kb in synapsin II in a sample of 654 schizophrenia cases and 628 normal controls to explore the mechanism underlying schizophrenia. We found significant differences in allele frequency distribution of SNP rs795009 (P ¼ 0.000018, OR 1.405, 95% Confidence Interval 1.202–1.641) between cases and controls. The T allele was

Abstracts significantly higher in cases than in controls. Moreover, the overall frequency of haplotype showed significant differences between cases and controls (P < 0.000001). This study suggests a positive association between synaspin II and schizophrenia, implying that synapsin II is involved in the etiology of schizophrenia. O4.4 CHROMOSOME 6Q .27 HOSTS A NEW POTENTIAL CANDIDATE GENE FOR SCHIZOPHRENIA Aberg K,1 Lindholm E,1 Saetre P,1 Pettersson U,2 Adolfsson R,3 and Jazin E1 1 Uppsala University, Dep. of Evolution, Genomics and Systematics, Uppsala, Sweden 2 Uppsala University, Dep. of Genetics and Pathology, Uppsala, Sweden 3 Umea˚ University, Dep of Clinical Sciences, Umea˚, Sweden Chromosome region 6q.27 has been proposed as a susceptibility locus for schizophrenia by previous linkage analysis. In this study we have fine mapped the region in 17 individuals diagnosed with schizophrenia and 17 parents. All individuals belong to one large pedigree, and descend from a common ancestor 12 generations back. Individuals were genotyped with 42 microsatellite and SNP markers, spanning 10.7 Mb, and haplotypes were constructed with the computer software Simwalk2. We found one haplotype of 2.4 Mb that is common in the family. To identify regions within this haplotype that were inherited identical by decent (IBD) we searched for similarities in ancestral haplotypes using in-married haplotypes as controls. A region of 0.3 Mb within the 2.4 haplotype showed a genetic polymorphism. In the 17 affected individuals the frequency of this polymorphism differed between chromosome-segments inherited from the large family compared to segments inherited from in-married parents. Therefore this 0.3 Mb region is likely to be inherited IBD. To investigate if this 0.3 Mb region is associated to schizophrenia, we genotyped three SNP markers in 178 unrelated nuclear families originating from the same geographical area in Northern Sweden as the large family. The haplotype that was likely to be inherited IBD in the large family was more frequent in individuals diagnosed with schizophrenia then in their healthy siblings. The 0.3 Mb region, identified in the two studies, hosts only one known gene. The function of this candidate gene in humans is unknown, but the homologous mouse gene is involved in neural development and myelination, functions that have been suggested to be involved in the development of schizophrenia. We have compared the mRNA levels of the candidate gene in a case-control study including brain autopsy samples from 55 schizophrenic and 55 control individuals. Expression levels were quantified with real-time PCR and preliminary results show that the expression level of the gene is strongly affected by the neuroleptics used to treat the schizophrenic patients. That is, the gene is up regulated in patients treated with typical neuroleptics, but down regulated in patients treated with atypical neuroleptics. Taken together, our results suggest that we have identified a candidate gene that may be associated with schizophrenia in a geographical area in Northern Sweden. Moreover, the regulation of the gene expression levels appears to be affected by neuroleptic drugs, suggesting that it may be involved in molecular pathways that are generally disturbed in schizophrenia. To identify mutations in the candidate gene we are presently sequencing affected individuals and controls.

O4.5 EXPOSURE TO CYTOMEGALOVIRUS AND POLYMORPHISMS IN TWO GENES ON CHROMOSOME 6P21-23 AS JOINT RISK FACTORS FOR SCHIZOPHRENIA Kim JJ,2 Dayal M,1 Bacanu S,1 Shirts BH3, Wood J,1 Xie W,1 Zhang X,4 Chowdar KV,1 Yolken R,5 Devlin B,1 and Nimgaonkar VL1 1 Department of Psychiatry, University of Pittsburgh School of Medicine, Pittsburgh, United States 2 Department of Psychiatry, The Catholic University of Korea, College of Medicine, Seoul, Korea, South 3 Department of Human Genetics, University of Pittsburgh, Graduate School of Public Health, Pittsburgh, United States 4 Department of Statistics, Carnegie Mellon University, Pittsburgh, United States 5 Stanley Laboratory of Developmental Neurovirology, Department of Pediatrics, Johns Hopkins School of Medicine, Baltimore, United States Liability to schizophrenia may be due to genetic and environmental factors. One potential environmental risk factor is viral infection, but

19

results remain inconsistent. Part of the inconsistency could be explained if interactions between infectious agents and host genetic variation were to play a role in susceptibility. We used a novel combination of familial aggregation and linkage/association studies to investigate the connection of four putative infectious etiological agents, cytomegalovirus (CMV) herpes simplex viruses 1 and 2 (HSV1)(HSV2) and Toxoplasma gondi (TOX), with liability to schizophrenia. We evaluated exposure among simplex and multiplex families. If infection increases liability to schizophrenia, we reasoned that greater exposure would lead to greater clustering of schizophrenia in families. We also evaluated the role of genetic variation in enhancing risk in CMV seropositive patients. We surveyed chromosome 6p21-23, which harbors a large number of immune-related genes, using 22 short tandem repeat polymorphisms (STRPS). We additionally investigated eleven genes localized to this chromosomal region, selected because they encode proteins mediating CMV infectivity using 46 single nucleotide polymorphisms (SNPs). Exposure to all these agents, evaluated using specific antibodies in the serum, was increased amongst multiplex families (more than one affected child) versus simplex families (one affected child), significantly so for CMV, HSV1, and HSV2. The difference was most consistent for CMV and supported CMV exposure as a risk factor for schizophrenia. Initial transmission disequilibirium tests (TDT) and case-control comparisons using STRPS revealed specific linkage/association with M6S125 among patients showing CMV exposure (CMVþ). Consistent with the STRP results, we detected significant linkage and association at two SNPs at MICB, and association with one TNF SNP among CMVþ patients. Notably, M6S125 is localized only 6kb from MICB and 71kb from TNF. The interaction of CMV exposure in schizophrenic patients and genetic variation at MICB and TNF is biologically plausible. Our analyses suggest an intriguing interaction between exposure to CMV and variation at MICB and TNF or other linked loci in the etiology of schizophrenia. If replicated, these findings would open new vistas of research into gene-environment interactions and schizophrenia pathogenesis.

O4.6 POLYMORPHISMS IN TRACE AMINE RECEPTOR 4 (TRAR4) ARE ASSOCIATED WITH SUSCEPTIBILITY FOR SCHIZOPHRENIA ON CHROMOSOME 6Q23.2 Duan J,1 Martinez M,2 Sanders AR,1 Hou C,1 Saitou N,3 Kitano T,3 Mowry BJ,4 Crowe RR,5 Silverman JM,6 Levinson DF,7 and Gejman PV1 1 Center for Psychiatric Genetics, Department of Psychiatry and Behavioral Sciences, Evanston Northwestern Healthcare & Feinberg School of Medicine, Northwestern University, Evanston, United States 2 Me´thodologie Statistique et Epide´miologie Ge´ne´tique des Maladies Multifactorielles, Institut National de la Recherche et de la Sante´ Me´dicale, Evry, France 3 Division of Population Genetics, National Institute of Genetics, Mishima, Japan 4 Queensland Centre for Schizophrenia Research, The Park, Centre for Mental Health (Wacol), and Department of Psychiatry, University of Queensland, Brisbane, Australia 5 Mental Health Clinical Research Center and Department of Psychiatry, University of Iowa College of Medicine, Iowa City, United States 6 Department of Psychiatry, Mount Sinai School of Medicine, New York, United States 7 Department of Psychiatry, University of Pennsylvania, Philadelphia, United States Several linkage studies across multiple population groups provide convergent support for chromosome 6q13-q26 as containing a susceptibility locus for schizophrenia, and more recently for bipolar disorder. We genotyped 192 linked schizophrenia pedigrees, both European and African American (AA) ancestry, across the MOXD1-STX7-TRARs gene cluster at 6q23.2, which contains a number of prime candidate genes for schizophrenia. Thirty-one screening SNPs were selected, providing a minimum coverage of at least one SNP per 20 kb. The association observed with rs4305745 (P ¼ 0.0014) within the TRAR4 (trace amine receptor 4) gene remained significant after correction for multiple testing. Evidence for association was proportionally stronger in the smaller AA sub-sample. Through database searching and sequencing genomic DNA in a thirty probands sub-sample, we obtained a high-density map of twenty-three SNPs spanning 21.6 kb of this gene. Single SNP and also haplotypic analyses revealed that rs4305745, and/

20

Abstracts

or two other SNPs in perfect linkage disequilibrium (LD) with rs4305745, appears the most likely mutation underlying the association of the TRAR4 region with schizophrenia. Comparative genomic analyses further revealed that rs4305745, and/or the associated polymorphisms in complete LD, can affect gene expression. Moreover, RT-PCR studies of various human tissues including brain confirm that TRAR4 is preferentially expressed in those brain regions implicated in the pathophysiology of schizophrenia. These data provide strong preliminary evidence that TRAR4 is a candidate gene for schizophrenia and await replication attempts in additional clinical samples. O4.7 AN EXAMINATION OF THE G72/G30 AND DAO LOCI IN SCHIZOPHRENIA Williams NM,1 Dwyer S,1 Norton N,1 Williams H,1 Kirov G,1 Green E,1 Dimitrova A,2 Grozeva D,1 McGregor S,3 Craddock N,1 Owen MJ,1 and O’Donovan MC1 1 University of Wales College of Medicine, Cardiff, United Kingdom 2 Department of Medical Genetics, Medical University Sofia, Sofia, Bulgaria 3 Biostatistics & Bioinformatics Unit, Cardiff, United Kingdom A number of studies have reported positive linkage signals from chromosome 13q34 in both schizophrenia and bipolar disorder. In following up one such positive linkage study, Chumakov and colleagues (2002) reported highly significant results for SNP’s spanning a 65 kb region in which two overlapping genes termed G72 and G30 are located. Yeast-2-Hybrid analysis demonstrated that G72 interacts with the enzyme D-amino acid oxidase (DAO). The gene for DAO is located at 12q24 and was also found to be associated with schizophrenia. These findings were subsequently replicated in an independent study (Schumacher et al 2004). We have performed mutation screening to identify new sequence variants across the coding sequence and adjacent flanking intronic sequence of G72, G30 and DAO. We identified 20 SNP’s spanning the G72/G30 locus and 8 SNP’s spanning the DAO locus. All were genotyped in 542 schizophrenia cases and 552 controls by DNA pooling. Two SNPs, one at the DAO locus and one at the G72/G30 locus were associated with schizophrenia (P < 0.05), these were confirmed by individual genotyping. In addition we tried to establish the LD structure of the G72/G30 and DAO loci by genotyping a 7.9 kb and 10.8 kb density grid of SNPs spanning each gene, which included the SNP’s reported to be associated to schizophrenia in the original publication, in 96 control individuals. Of these, 9 and 8 SNPs spanning G72/G30 and DAO respectively were selected as our most informative set of SNPs. Each were genotyped in 709 schizophrenia cases and 1349 controls. No single marker or haplotypic associations with schizophrenia were identified at the G72/G30 locus. At the DAO locus only rs2111902 (MDAAO4) was associated with schizophrenia (P ¼ 0.02) while the most significant haplotypic association was with markers rs2111902 and rs3918346 (MDAAO5) (global P ¼ 0.008). We were unable to replicate the reported genetic interaction between rs778293 (M22) and rs3741775 (DAAO6) at the G72/G30 and DAO loci respectively (P ¼ 0.12). Our data provides modest support for the association between DAO and schizophrenia. O4.8 KNOWLEDGE AND PERCEPTIONS REGARDING GENETICS AND MENTAL ILLNESS: A SURVEY OF INDIVIDUALS AFFECTED BY MENTAL ILLNESS AND THEIR FAMILIES Austin JC University of British Columbia, Vancouver, Canada The field of genetic research into mental illness is now advancing at an exciting rate. Despite the rapid advance in the understanding of the nature of the genetic contribution to mental illness within the academic community, the understanding and perceptions of affected individuals and their families regarding the role of genetics in mental illness remains relatively unexplored. OBJECTIVE: To assess perceptions of families affected by mental illness with respect to: potential impact of predictive genetic testing, perceptions of genetic risk, the associated level of concern and effect on family planning, need for education regarding the genetic contribution to mental illness. In general, individuals affected by mental illness and their families: 1) overestimate genetic recurrence risks and have fewer or no children as a result of perceived risk, 2) view the possibility of predictive genetic testing for mental illness positively, 3) feel there is a need for education regarding the genetic contribution to mental illness. A web-based

questionnaire was posted on the British Columbia based South Fraser Early Psychosis Intervention (EPI) team website (www.psychosissucks.ca) on October 1 2003. The questionnaire comprised both checkbox items and items where a full answer could be provided. Quantitative data was assessed using descriptive statistics. All 220 completed surveys submitted between October 1 2003 and May 1 2004 were analysed. 91% of respondents either had a family history of mental illness, or a personal history, or both. Of these individuals, 81% said they were concerned about the chances for other family members to develop mental illness, and 24% reported fewer children or no children due to family history. The most significantly affected group were those who had an affected—sibling in this group 36% reported fewer or no children as a result of family history. Overall 25% of respondents indicated that they did not know what the chance would be for the sib of an affected individual to be affected too. The mean risk overall was perceived to be 26% (actual 8–17%), and 56% of respondents thought the chance for a sib to be affected was >/ ¼ 25%. 50% of respondents were not sure if there is currently clinical genetic testing for mental illness. Predictive genetic testing was viewed positively (82%), as was the possibility of prenatal genetic testing for mental illness (50%). 92% of respondents reported a desire for education regarding the genetic contribution to mental illness. Families are concerned about the perceived possible role of genetics in mental illness, and overestimate genetic risks. Genetic testing is regarded positively. There is a need and desire for education about the genetic contribution to mental illness, indicating that genetic counselling should be routinely offered to affected families.

O5.1 HAPLOTYPE BLOCK ANALYSIS FINDS ASSOCIATION BETWEEN THE DOPAMINE TRANSPORTER GENE AND COMORBID ADHD AND BIPOLAR DISORDER BUT NOT ADHD WITHOUT BIPOLAR DISORDER Faraone SV,2 Arbeitman L,1 Fagerness H,1 Kirby A,1 Daly M,4 Doyle A,3 Smoller JW,3 and Sklar P3 1 Massachusetts General Hospital, Boston, United States 2 Harvard School of Public Health, Boston, United States 3 Harvard Medical School, Boston, United States 4 Whitehead Institute for Biomedical Research, Cambridge, United States Studies of individuals with a diagnosis of Attention Deficit Hyperactivity Disorder (ADHD) show a 10-fold increase risk for bipolar disorder (Biological Psychiatry 2001 49(12):1062–70). Additionally, both of these disorders tend to co-occur in families (J Affect Disord. 2001 Apr,64(1):19–26), suggesting a common genetic etiology. No studies to date have sought to identify genes implicated in comorbid ADHD and bipolar disorder. However, individual markers within genes have been associated with ADHD including those in SLC6A3, DRD4, DRD5, SNAP-25 and 5HT1B. We have developed detailed linkage disequilibrium maps across these genes. For each gene an initial map was generated by selecting SNPs at a density of 1 every 4 kb. To increase the likelihood of a SNP being informative we chose SNPs that were found independently in two databases (dbSNP and Celera). SNPs were genotyped using mass spectrometry in a sample of 12 multigenerational CEPH pedigrees. Pairwise measures of linkage disequilibrium were calculated and haplotype blocks constructed according to Gabriel et al. (Science 2002 296:2225-9). Additional SNPs were added to blocks where the haplotype diversity was less than 90% until a maximum of 6 SNPs per block was reached. Additional SNPs were added between blocks until the interblock distance was less than 3 kb. The structure of linkage disequilibrium for each gene is visualized using LocusView. We performed association analyses of a subgroup of ADHD patients with bipolar disorder. We genotyped 152 SNPs in 6 genes in 280 trios in (199 families) in which probands were ascertained for DSM-IIIR or DSM_IV. Fifty-four trios comprised offspring affected with comorbid bipolar disorder. Quality control metrics included analysis of SNPs for which 90% of the genotypes were available and meeting criteria for Hardy-Weinberg equilibrium and fewer than 3 Mendel errors. Analyses of individual SNPs and haplotypes were performed using the transmission-disequilibrium test as implemented in the software TDT-permute. No associations were detected for HTR1B, BDNF, SNAP25, DRD5. In contrast, for SLC6A3, 35 SNPs were tested and 7 SNPs were associated at the P < 0.05 level. Four of these SNPs were in strong linkage disequilibrium and defined a haplotype block that was overtransmitted to bipolar probands (nominal P ¼ 0.00067). This block encompasses the first three exons, promoter as well as a portion of the

Abstracts hypothetical gene FLJ12443. We did not find a significant association between ADHD without BP and SLC6A3. Ongoing work is aimed at more full definition of these haplotype blocks and replication in additional samples. O5.2 EVIDENCE OF ASSOCIATION FOR A SUSCEPTIBILITY REGION FOR AUTISM ON CHROMOSOME 2Q Conroy J,1 Segurado R,1 Meally E,2 Stallings R,3 Green A,3 Fitzgerald M,2 Gill M,2 Ennis S,3 and Gallagher L2 1 Department of Genetics, Trinity College Dublin, Dublin, Ireland 2 Dept of Psychiatry, Trinity College Dublin, Dublin, Ireland 3 National Centre for Medical Genetics, Dublin, Ireland Autism, a neurodevelopmental disorder of childhood, has a strong genetic component to its aetiology. Molecular genetic studies have highlighted several genomic regions of interest, with strong evidence for LD on chromosome 2q (Buxbaum et al., 2001, IMGSAC 2001, Shao et al., 2002). We have described an individual with a cytogenetic abnormality on chromosome 2q32 (Gallagher et al., 2002). This region was tested for linkage disequilibrium in a sample of parent-child trios with autism recruited in the Irish population. A microsatellite marker (D2S2077) located at the centromeric end of the region was in LD with autism (w2 ¼ 5.8, P ¼ 0.013, OR ¼ 1.75). A SNP based fine-mapping experiment was conducted around marker D2S2077 and it’s two flanking markers (D2S2310 and D2S364). One SNP (dbSNP: rs155149) showed significant transmission in the autism sample based on HHRR association testing (w2 ¼ 5.16, P ¼ 0.023, RR ¼ 1.48, 1.01 < RR < 2.16). Association testing of haplotypes was performed using TRANSMIT. Association with autism was detected with several haplotypes that contained the over-transmitted alleles of the D2S2077 marker and the SNP (rs155149). A six-marker haplotype was found to have an odds ratio of 6.71 (w2 ¼ 12.06, P ¼ 0.0005). This haplotype was located around a known gene, Integrin Alpha 4 (ITGA4), and two of the SNPs on the haplotype were located within introns in the gene. ITGA4 is involved in cell-adhesion and is expressed in the limbic brain, a region postulated to be involved in the pathological process of autism. ITGA4 has been prioritized for follow up association testing using further SNPs, and the results of these experiments will also be presented. O5.3 LINKAGE AND ASSOCIATION OF AN ASPARTATE/ GLUTAMATE CARRIER WITH AUTISM: GENETIC VARIANTS ARE ASSOCIATED WITH A SEVERAL-FOLD INCREASED RISK Buxbaum JD, Ramoz N, Reichert J, Smith CJ, Silverman J, Bespalova I, and Davis KL Mt Sinai School of Medicine, New York, United States A major susceptibility locus for autism was mapped to 2q24-q33. We analyzed genes across this interval to identify an autism susceptibility gene. First, we identified genetic variants in exons and flanking sequence in unrelated subjects showing linkage to 2q24-q33, and compared frequencies between subjects and controls. Second, variants showing nominal association were further genotyped in 411 autistic families, and linkage and association tests were carried out. Linkage and association were observed between autism and two SNPs, both within a gene encoding a mitochondrial aspartate/glutamate carrier (AGC1) that is involved in adenosine triphosphate (ATP) synthesis. Using a single affected per family, evidence for excess transmission was found by the TDT and by TRANSMIT for AGC1-4 (P ¼ 0.001 by both tests), AGC1-9 (P ¼ 0.01 and 0.007), and a two-locus G*G haplotype (P ¼ 0.000003 and 0.004). Linkage analysis with the two SNPs demonstrated a maximal multipoint NPL score of 1.57 and a maximal multipoint heterogeneity LOD score of 2.11. Genotype relative risk (GRR) was estimated at 2–5 depending on number of alleles, although GRR is underestimated in studies such as this. Further studies are needed to confirm this association and to decipher any potential etiological role of AGC1 and mitochondrial function in autism. O5.4 WHOLE GENOME LINKAGE ANALYSIS OF WORKING MEMORY COMPONENT PHENOTYPES OF DYSLEXIA Brkanac Z, Chapman NH, Igo RP, Thomson JB, Matsushita M, Holzman T, Berninger VW, Wijsman EM, and Raskind WH University of Washington, Seattle, United States Dyslexia is a common disorder that has evidence for a genetic component. The genetics is complex, as phenotypes do not show Mendelian

21

patterns of inheritance. This may be due to underlying genetic and diagnostic heterogeneity. To reduce heterogeneity and increase the chances for linkage detection, we are using careful evaluation of phenotypic components of dyslexia coupled with univariate, componentbased linkage analyses. We have previously shown through family aggregation and segregation analyses that working memory phenotypes have a genetic basis in the families under study. We have assessed working memory with the Digit Span (DS) subtest of WAIS-R and WISC-III. Phonological working memory was assessed with the Non Word Memory (NWM) task of the Comprehensive Test of Phonological Awareness. These phenotypes were chosen for initial studies as there is evidence for a genetic basis of performance on phonological short-term memory tasks, as well as for their role in dyslexia. Families were collected through probands who had a Verbal IQ (VIQ)  90 and performance on at least one of ten dyslexia measures both below the population mean and at least one Standard Deviation below the VIQ. A genome scan was performed on 444 people from 52 families selected on the basis of power to detect linkage for NWM. Families were genotyped at an average 10 cM resolution. Linkage analyses were performed using multipoint Variance Component (VC) and Markov chain Monte Carlo (MCMC) multipoint joint linkage and segregation analysis methods. These methods are complimentary as MCMC methods allow simultaneous QTL localization and model parameter estimation, while the VC approach allows for easier interpretation of the results and examination of the contribution of individual families to the signal. Evaluation of the contribution of individual families might be invaluable for subsequent molecular genetics studies. For NWM we have observed numerous regions where VC dominance lod scores exceeded 1.5 and MCMC intensity ratios (IR) exceeded 5.0. These cutoffs were chosen to distinguish signals from background noise. Lod scores >2.0 that were supported by IR scores >5 were observed on chromosomes 2q, 4p, 6q, 17q, and 22q. DS lod scores >1.5 were also observed on chromosomes 4q, 13p and 22q. MCMC IR results >5 for DS were detected on chromosomes 5q and 14q. Our previous segregation analysis suggested that some genetic factors might contribute jointly to DS and NWM, and that additional factors might contribute to NWM only. Data analysis of NWM adjusted for DS reduced the number of regions with VC lod >1.5 and identified an additional location on chromosome 7q with VC lod >2.0 and MCMC IR>5. Our study has identified novel loci potentially linked to NWM in families with dyslexia. Linkage was supported by two independent methods of analysis. DS has not shown strong evidence for the linkage, although previous aggregation and segregation studies implicated a genetic component. Analysis of additional families will be needed to further refine these findings.

O5.5 IDENTIFICATION OF SIGNIFICANT AUTISM ASSOCIATION ON GABAA RECEPTOR SUBUNIT GENES IN CHROMOSOME 4 AND 5 Ma DQ,1 Whitehead PL,1 Menold MM,1 Martin ER,1 Ashley-Koch AA,1 DeLong GR,1 Abramson RK,2 Wright HH,2 Cuccaro ML,1 Hussman JP,1 Gilbert JR,1 and Pericak-Vance MA1 1 Duke Center for Human Genetics, Durham, NC, United States 2 University of South Carolina, Columbia, SC, United States GABA is the primary inhibitory neurotransmitter in adult brain. Eight GABA classes (a, b, d, e, g, p, y and r) and twenty receptor subunit genes have been characterized in mammals. Multiple lines of evidence implicate GABAA receptor systems in autism. We have undertaken ‘‘The GABA Project’’ to test all known GABR subunits and selected GABAergic related genes for association with autism. To date we have tested 9 GABAA receptor subunit genes located in chromosome 4 and 5. They are RG1| LOC16, GABRA2, GABRA 4 and GABRB1 located in 4p12 and GABRB2, GABRA6, GABRA1, GABRG2 and GABRP in 5q34 to q35.1. One to five intronic and silent mutation SNPs within each gene have been analyzed in 589 families. 301 families from Duke/USC and 288 from AGRE were analyzed for association using the family-based Pedigree Disequilibrium Test (PDT), and geno-PDT (PDT extension that examines genotypic association), as well as multi-locus geno-PDT to test for gene-gene interactions. Significant association was found for rs1912960 (GABRA4) (P ¼ 0.02) in the overall dataset and (P ¼ 0.0005) in the AGRE. rs2280073 (GABRA4) was also significant (P ¼ 0.007) in the AGRE. In the Duke/USC, a single SNP hcv8262290 (GABRA2) (P ¼ 0.013) was significant. Geno-PDT gave significant associated with rs1912960 and rs2280073 (GABRA4) and RS6289 (GABRB1) as well as rs2617503 and rs2962407 (GABRB2). Multi-locus geno-PDT showed a significant joint effect between these genes (Global P from 0.01 to

22

Abstracts

0.002). These data support the hypothesis that GABAergic genes are involved in autism. O5.6 RARE VARIATIONS AND STOP MUTATIONS IN THE NEUROLIGIN GENES IN AUTISM AND ASPERGER SYNDROME Jamain S,1 Quach H,1 Betancur C,2 Goubran-Botros H,1 Durand C,1 Melke J,1 Delorme R,1 Rastam M,3 Colineaux C,2 Gillberg IC,3 Soderstrom H,3 Bonneau D,5 Firth H,4 Leboyer M,2 Gillberg C,3 and Bourgeron T1 1 Institut Pasteur, Paris, France 2 INSERM U513, Creteil, France 3 Goteborg University, Goteborg, Sweden 4 Addenbrooke’s Hospital, Cambridge, United Kingdom 5 CHU Angers, Angers, France Autism is characterised by qualitative impairments in social interactions and communication and associated with restrictive interests and repetitive behaviours. Autistic disorders affect approximately one in 1000 children, with a higher prevalence among boys. In a recent paper, we reported mutations in X linked neuroligins (NLGN3 and NLGN4) in subjects with autism and Asperger syndrome. Both the NLGN3 and NLGN4 mutations result in intracellular retention of the mutant proteins. The association of neuroligins with autism was recently also supported by the identification of a frame-shift mutation of the NLGN4 gene in one large family with mental retardation (10 males), autism (2 males) and Pervasive Development Disorder (1 male). Neuroligins (NLGNs) are cell adhesion molecules that contribute to the formation of synapses. They are located at the postsynaptic membrane of glutamatergic synapses, where they interact with b-neurexins to form a heterophylic complex that is probably involved in synaptogenesis. In humans, five neuroligins have been described and all of them are localised in chromosomal regions previously associated with autism. In addition to the X-linked genes NLGN3 and NLGN4, NLGN1 is located at 3q26, which corresponds to the most significant linkage identified in a recent study from Finland. NLGN2 is located at 17p13, close to two translocation breakpoints reported in two independent individuals with Asperger syndrome. Finally, NLGN4Y is located on the Y chromosome at Yq11, where a number of chromosomal abnormalities and aneuploidies have been described in individuals with autism. Here, we report the analysis of the five NLGN genes in autistic subjects. To investigate whether these genes may have a role in autistic disorder, we systematically screened all neuroligin genes for mutations in the coding sequence in 90 (NLGN1, NLGN2) to 290 (NLGN3, 4 and 4Y) individuals affected with autism or Asperger syndrome. In addition, single nucleotide polymophisms (SNPs) identified in this study were analysed for association with autism. O5.7 AN INTERSTITIAL DELETION OF CHROMOSOME REGION 21Q22.3 IN FOUR RELATIVES, INDICATING S100B AND GRIP1B AS DYSLEXIA CANDIDATE GENES: IMPLICATIONS FOR THE GENETICS OF DYSLEXIA Poelmans G,1 Engelen J,2 Van Lent J,2 Schoenmaekers E,3 Wuisman-Frerker M,4 Erens W,4 Schrander-Stumpel C,2 Steyaert J1 1 Department of Child and Adolescent Psychiatry, Catholic University of Leuven, Leuven, Belgium 2 Department of Clinical Genetics, Academic Hospital Maastricht, Maastricht, Netherlands 3 Department of Cytogenetics, University Hospital Nijmegen, Nijmegen, Netherlands 4 Department of Neuropsychology, Academic Hospital Maastricht, Maastricht, Netherlands Dyslexia is a significantly heritable learning disorder. At least five chromosomal loci have been repeatedly linked to dyslexia, and additional dyslexia susceptibility loci on other chromosomes have been suggested. A first candidate gene for dyslexia (DYX1C1) was recently identified through the molecular analysis of the 15q21 translocation breakpoint in four members of the same family cosegregating dyslexia and a t(2,15). At the 2003 WCPG in Que´bec, we have reported on a family cosegregating dyslexia and an interstitial deletion of chromosome region 21q22.3. In our family, a father and his three sons mainly show deficits in phonological decoding and single-word reading, two genetically correlated components of the reading process. FISHanalysis with subtelomeric probe CTB-63H24 showed a fluorescent signal on only one of the chromosomes 21 in all four relatives, indicating

an interstitial deletion in chromosome band 21q22.3. Because linkage between a marker in chromosome region 21q22.3 and dyslexia has been reported, our first objective was to try to refine the deleted region by further cytogenetical analysis, in order to find one or more possible dyslexia candidate genes. Secondly, we wanted to find out whether the thus identified candidate genes would be correlated with each other or with other positional candidate genes. To determine the size of the deleted region further, FISH was performed with additional smaller probes covering the subtelomeric probe CTB-63H24. To determine which genes are located within the deleted region and whether they are correlated with each other or with other positional dyslexia candidate genes, we used information from the Ensembl Genome Browser and the UCSC Genome Browser. By performing FISH analysis with the additional probes, we were able to refine the deletion to a region of approximately 230 kb in (maximal) size on the far end of chromosome band 21q22.3. Within this region, only three known genes are located, of which two are expressed in the brain, namely GRIP1b (or DIP2_human) and S100B. By combining data from the literature concerning these two genes and information from the Ensembl and UCSC Genome Browsers, we have found some evidence that S100B, GRIP1b, but also some other positional dyslexia candidate genes belong to the same neurobiochemical pathway. We have identified S100B and GRIP1b as dyslexia candidate genes by the thorough cytogenetical analysis of a very small interstitial deletion of chromosome region 21q22.3, which is cosegregating with dyslexia in a father and his three sons. Moreover, these two genes and other positional dyslexia candidate genes might belong to the same neurobiochemical pathway. O5.8 EVIDENCE THAT VARIATION AT BRAIN DERIVED NEUROTROPHIC FACTOR (BDNF) INFLUENCES SUSCEPTIBILITY TO ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD) Kent L,1 Green E,2 Hawi Z,3 Kirley A,3 Lowe N,3 Dudbridge F,1 Fitzgerald M,3 Gill M,3 O’Donovan M,2 Thapar A,2 and Craddock N2 1 University of Cambridge, Cambridge, United Kingdom 2 Division of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 3 Department of Psychiatry and Genetics, Trinity College, Dublin, Ireland Attention Deficit Hyperactivity Disorder (ADHD) is a highly heritable, common, psychiatric disorder of childhood which probably involves several genes. There are many plausible candidate genes systems which may be involved in ADHD. Brain derived neurotropic factor (BDNF) is a member of the neurotrophin family and is involved in the survival and differentiation of dopaminergic neurons in the developing brain. Only one frequent non-conservative polymorphism in the human BDNF gene (dbSNP number rs 6065) has been identified, a SNP at nucleotide 196 (G/A). This produces an amino acid substitution (valine to methionine) at codon 66. In a large collaborative family based sample of 326 ADHD probands and their parents we found evidence for the involvement of BDNF in ADHD. In the entire sample, preferential transmission of the G allele was nominally significant (OR ¼ 1.61, P ¼ 0.02) as was the HHRR (OR ¼ 1.54, P ¼ 0.03). There was highly significant overtransmission of the G allele from paternal transmissions only (OR ¼ 3.2, P ¼ 0.0005). Transmission was even more significant after estimation of missing parental genotypes (HHRR OR ¼ 3.02, P ¼ 0.00017). The over transmission of paternal alleles was demonstrated in each study site’s sample separately. These findings suggest a possible role for imprinting mechanisms in ADHD susceptibility. O6.1 LINKAGE OF ALCOHOL DEPENDENCE SYMPTOMS TO CHROMOSOME 4 IN THE IRISH AFFECTED SIB PAIR STUDY OF ALCOHOL DEPENDENCE Prescott CA,1 Sullivan PF,2 Webb BT,1 Vittum J,1 Patterson DG,3 Neale MC,1 van den Oord EJ,1 Walsh D,4 Riley BP,1 and Kendler KS1 1 Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, Virginia, United States 2 Department of Psychiatry, University of North Carolina, Chapel Hill, North Carolina, United States 3 Shaftsbury Square Hospital, Belfast, Northern Ireland, United Kingdom 4 Health Research Board, Dublin, Ireland A genome scan was conducted using data from the Irish Affected Sib Pair Study of Alcohol Dependence (IASPSAD, Prescott et al., 2002). The

Abstracts majority of probands were ascertained through alcoholism treatment programs and have severe clinical histories. Probands and affected siblings were evaluated using the SSAGA interview and parents were evaluated using the SCID. A 4 cM genome scan was conducted using 474 families of which most (96%) were affected sib pairs. Nonparametric and quantitative linkage analysis were conducted using DSM-IV alcohol dependence (AD) and number of DSM-IV AD criteria, respectively. Preliminary QTL results indicate strong linkage (multipoint QTL LOD ¼ 4.59, P < 0.0000021) for number of AD criteria to a region of chromosome 4 which includes the ADH cluster. NPL analysis showed weak linkage nearby (maximum nearby NPL LOD < 1). The chromosome 4 results are consistent with prior linkage studies (Long et al., 1998; Foroud et al., 2000) and a number of association studies. These results await further analysis and follow-up genotyping, but suggest the importance of variation in the ADH cluster in the etiology and severity of alcoholism in Caucasian populations. This work was supported by NIH grant AA-11408. O6.2 A GENOME-WIDE SCAN OF 250 SWEDISH AFFECTED SIBPAIR FAMILIES WITH ALCOHOL ADDICTION ¨ sby U,1 Kockum I,1 Liljenberg J,1 Bransome M,1 Schumacher J,2 O Wire´n K,1 No¨then M,2 Propping P,2 and Terenius L1 1 Karolinska Institutet, Stockholm, Sweden 2 Friedrich-Wilhelms-Universita¨t, Bonn, Germany The risk for alcohol dependency is both genetically and environmentally influenced. Evidence from family, twin and adoption studies indicate an important genetic contribution. Having an alcoholic parent increases the risk fivefold, and for siblings the risk is 3–8 fold increased. Twin studies of alcoholism and alcohol consumption find approximately twice the concordance rate for monozygotic compared to dizygotic twins. Heritability is estimated to 50–60%. Molecular genetic studies, mainly the Collaborative Study on the Genetics of Alcoholism (COGA), in a study including totally 264 affected sibpair families, have found evidence for linkage to several chromosomal regions. Within the Swedish Alcoholism in siblings study, we are recruiting a sample of Swedish families with Nordic origin, with at least two siblings with alcohol dependency. Parents are also included, and other siblings if both parents are not available. Probands and affected siblings are identified from several clinical treatment centers for alcoholism, and parents an other siblings are contacted with the consent of the probands. Phenotypic assessment is based on a structured psychiatric interview using SCAN together with information from medical records. Individuals are considered to be affected if their diagnosis were confirmed in a structured psychiatric interview or if they had been in inpatient treatment for alcohol addiction. The study is approved by the IRB at Karolinska Institutet. The first 250 families have been genotyped by deCode with 500 microsatellite markers evenly spaced throughout the genome. The statistical linkage analyses are being performed with Allegro, after the family structures have been checked with Siberror. So far, only ch 1–3 have been analyzed. In the first results, there is a locus with a LOD of 0.90 on ch 1, a locus with a LOD of 3.58 on ch 2, and a locus with a LOD of 0.83 on ch 3. The findings of the linkage analyses will be presented.

O6.3 GENETIC HETEROGENEITY FOR SOCIAL AND NON-SOCIAL BEHAVIOURS CHARACTERISTIC OF AUTISM SPECTRUM DISORDERS IN THE GENERAL POPULATION AND AT THE EXTREME Ronald A,1 Happe F,1 Bolton P,1 Wheelwright S,2 Baron-Cohen S,2 and Plomin R1 1 SGDP Centre, Institute of Psychiatry, London, United Kingdom 2 Autism Research Centre, Cambridge University, Cambridge, United Kingdom This study explored the extent to which genes and environment influence behaviours characteristic of autism spectrum disorders in the general population and at the increasingly impaired extreme. Data came from the Twins Early Development Study (TEDS), which is representative of the UK population and includes children diagnosed with autism spectrum disorders. Parents of 3,400 pairs of 8-year-old twins completed the Childhood Asperger Syndrome Test (CAST, Scott et al 2002), a screening measure of autism spectrum disorders designed for non-clinical settings and with a cut-off for identifying children at risk. As well as looking at the ‘at risk’ group (>98% on scale measuring

23

impairment), analyses were also run for the children in the top 95%, 90% and 85% of the distribution. Twin correlations suggested that the whole scale was highly heritable in the general population (correlations: MZ ¼ .81, DZ ¼ .31) as did group correlations for the ‘at risk’ group (MZ ¼ .80, DZ ¼ .14) and the other extreme groups. Variance components and liability threshold model fitting confirmed that behaviours characteristic of the autism spectrum are highly heritable in the general population and at the increasingly severe extremes (heritabilities ¼ 80–90%). At the extreme, DZ twin concordances were lower than for the general population, suggesting there might be increasing epistatic effects occurring in the extreme groups. Most importantly, when the items relating to social impairments and nonsocial obsessive repetitive behaviours were analysed separately, they were found to be highly heritable but largely independent of each other in terms of genetic influences (genetic correlation ¼ .24–.36). In conclusion, behaviours characteristic of the autism spectrum are highly heritable both in the general population and at the impaired extreme (which included diagnosed children). Second, there appears to be genetic heterogeneity in the etiology of behaviours characteristic of the autism spectrum: the genes that influence social impairments and nonsocial obsessive repetitive behaviours are largely independent of each other. This latter finding suggests a preliminary hypothesis that molecular genetic research might profit from considering these two aspects of the autism spectrum separately. O6.4 LONGITUDINAL GENETIC ANALYSIS OF ANXIETY/ DEPRESSION DURING CHILDHOOD: A STUDY FROM THE NETHERLANDS TWIN REGISTER Boomsma DI,1 Van Beijsterveldt CEM,1 Derks EM,1 Bartels M,1 and Hudziak JJ2 1 Dept Biological Psychology, VU, Amsterdam, Netherlands 2 Department Psychiatry and Medicine, Burlington, United States In our longitudinal study of the development of behavioral and emotional problems in young children we studied Anxiety/Depression (A/D) at ages 3, 5, 7, 10 and 12 years. The children, mono- and dizygotic twin pairs, were registered at birth with the Netherlands Twin Register. Indices of Anxiety/Depression (A/D) were obtained from maternal and paternal ratings on the Child Behavior Checklist (CBCL). At age 5, anxiety was obtained from the Devereux Child Behavior (DCB) rating scale. At ages 3 and 5 data were available for around 9000 twin pairs, at 7, 10 and 12 years for around 7300, 4400 and 2400 pairs. When the reliable variance of A/D was analyzed, heritability became smaller as children grew older (around 70% at age 3 and around 40% at age 12) but remained the most important factor in explaining the variance of A/D at each age. Across age, the stability coefficient for A/D was relatively low between age 3 and later ages (0.31–0.23), but higher for the shorter intervals (up to 0.66 between ages 10 and 12 years). We applied longitudinal genetic models to investigate whether genetic or environmental factors underlie the stability of A/D. On average, about 50% of the covariance across time could be explained by genetic factors and about 40% of the stability of A/ D was accounted for by shared environmental influences. Non-shared environmental factors operated mainly in a time-specific manner.

O6.5 ASSOCIATION OF CHRNA4 SNPS AND HAPLOTYPES WITH SMOKING AND ADHD Todd RD, Lobos EA, Sun LW, and Neuman RJ Washington University School of Medicine, Departments of Psychiatry and Genetics, St. Louis, United States Smoking and its health related consequences represent a major source of morbidity, mortality and economic loss on a worldwide basis. Though a variety of cultural, social, legal and personality factors have been proposed as contributing to the etiology or development of smoking problems, twin studies support that regular smoking and/or nicotine dependence are moderately heritable behaviors. Recent clinical and community-based studies have also demonstrated an association between cigarette smoking and inattentive ADHD symptoms. Using population-based definitions of attention deficit/hyperactivity disorder (Todd et al., 2001; Rasmussen et al., 2004) we have reported on a genetic association between an intron 2 polymorphism of the high affinity nicotinic acetylcholine receptor sub unit gene (CHRNA4) and inattentive ADHD (Todd et al., 2003). Similarly, in a large population-based twin sample, we found that inattentive ADHD was a strong predictor of

24

Abstracts

regular smoking in youth (odds ratio equal 4.7, 95% CI 1.8–12.4). In the current study, we analyzed common polymorphisms of the CHRNA4 gene for association with regular smoking and inattentive ADHD in the same population-based twin sample. Analyses were restricted to those youth who had tried a cigarette in order to increase the likelihood of significant exposure to nicotine. Using case-control logistic regression analysis with covariates, we found significant associations between several CHRNA4 snps and related haplotypes and progression to regular smoking in population-defined attentive ADHD (odds ration 2.7, 95% CI, 1.1–6.3). There was no similar association for combined subtype ADHD or DSM-IV defined ADHD subclasses. These findings are currently being extended to population-defined groups of nonADHD youth who have progressed to regular smoking. O6.6 A FAMILY-BASED STUDY SHOWS ASSOCIATION BETWEEN THE INSULIN-LIKE GROWTH FACTOR 2 (IGF2) GENE AND EATING BEHAVIOR (EAT-26) IN A NON-CLINICAL POPULATION Bachner-Melman R,1 Gritsenko I,2 Zohar A,3 and Ebstein RP1 1 Hebrew University, Jerusalem, Israel 2 Herzog Hospital, Jerusalem, Israel 3 Rupin Academic Center, Emek Hefer, Israel An interesting candidate gene that makes ’biological sense’ for a role in eating disorders is insulin-like growth factor 2 (IGF2). IGF2 is a member of the insulin family of polypeptide growth factors that is involved in development and growth. Consistent with its profile of metabolic actions, an association has been reported between a SNP in the 30 untranslated region of the IGF2 gene (ApaI) and BMI. Towards examining the role of the IGF2 gene in eating behavior, we tested for association between three IGF2 SNPs and scores on the EAT-26 instrument, a self-report questionnaire widely used as a screening instrument for eating disorders, in a non-clinical sample of 376 families. A highly significant association was observed between the IGF2 ApaI G allele and EAT-26 scores overall (z ¼ 4.126, P ¼ 0.000037) as well as for each of the EAT-26 subscales (Bulimia: z ¼ 4.149, P ¼ 0.000033, Dieting: z ¼ 4.149, P ¼ 0.000033, Oral Control: z ¼ 3.237, P ¼ 0.0012). Additionally, significant association was observed between this polymorphism, BMI and menarche. The current findings that the IGF2 ApaI polymorphism, that predisposes to weight gain, may also contribute to the pathology of eating disorders as evidenced by higher EAT-26 scores, is intriguing. Neurotransmitter modulation of appetitive behavior is the focus of most hypotheses regarding the etiology of severe eating disorders. However, the current results to some measure challenge this view and we hypothesize that inborn metabolic tendencies to weight gain in some women may trigger constant dieting that in predisposed individuals eventually leads to severe eating disorders including AN, BN and BED.

O6.7 ASSOCIATION OF THE CORTICOTROPIN RELEASING HORMONE GENE WITH ANXIOUS TEMPERAMENT IN CHILDREN AT RISK FOR PANIC DISORDER Smoller JW,1 Yamaki LH,1 Fagerness J,1 Racette S,1 Biederman J,1 Laird NM,2 Tsuang MT,3 Rosenbaum JF,1 and Sklar P1 1 Department of Psychiatry, Massachusetts General Hospital, Boston, United States 2 Harvard School of Public Health, Boston, United States 3 Department of Psychiatry, University of California at San Diego, San Diego, United States Behavioral inhibition to the unfamiliar (BI) is a heritable temperamental phenotype involving the tendency to display fearful, avoidant, or shy behavior in novel or unfamiliar situations. Biological features include evidence of hypothalamic-pituitary-adrenal axis activation and hyperreactivity of the amygdala to novelty. BI has been shown to be a familial and developmental risk factor for panic and phobic anxiety disorders and may be a useful intermediate phenotype for genetic studies of these disorders. Corticotropin releasing hormone (CRH) is a key mediator of the stress response through its control of the hypothalamic-pituitary-adrenal axis and effects on limbic brain systems. Evidence from experimental animal models has implicated CRH in the development of BI-like behaviors. We recently observed an association between BI and a microsatellite marker linked to the corticotropin releasing hormone (CRH) gene in children at risk for panic disorder (Smoller et al., Biol Psychiatry 2003). To evaluate this

further, we genotyped additional families for this marker and a panel of markers encompassing the CRH locus. We first examined the haplotype structure of the CRH locus by determining linkage disequilibrium relationships among multiple SNP markers in and around the gene in a sample of 12 multigenerational CEPH pedigrees.. Association analyses using FBAT were then conducted in 62 families that included parents with panic disorder and children who underwent laboratory-based behavioral observations for BI. We examined a set of markers which we found to reside in a block of strong linkage disequilibrium encompassing the CRH locus. The BI phenotype was significantly associated with the microsatellite marker (P ¼ 0.0016) and three of the eight single nucleotide polymorphisms (SNPs) tested (P < 0.05), including a SNP in the coding sequence of the gene. Haplotype-specific tests revealed association for a haplotype comprising all of the markers (P ¼ 0.008), and a global multiallelic test of the haplotypes was also significant (P < 0.05). These results suggest that the CRH gene influences inhibited temperament, a risk factor for panic and phobic anxiety disorders. Future research will be required to identify the specific variants that confer susceptibility. Genetic studies of anxiety-related temperament represent an important strategy for identifying the genetic basis of anxiety disorders. O6.8 REPLICATION STUDY SUPPORTS EVIDENCE FOR LINKAGE TO 9P24 IN OBSESSIVE-COMPULSIVE DISORDER Willour VL,1 Shugart YY,1 Samuels J,1 Grados M,2 Cullen B,1 Bienvenu OJ,1 Wang Y,1 Liang K-Y,1 Valle D,1 Hoehn-Saric R,1 Riddle M,1 and Nestadt G1 1 Johns Hopkins University, Baltimore, MD, United States 2 Kennedy Krieger Institute, Baltimore, MD, United States Obsessive-compulsive disorder (OCD) is a severe psychiatric illness that is characterized by intrusive and senseless thoughts and impulses (obsessions) and repetitive behaviors (compulsions). Family, twin, and segregation studies support the presence of both genetic and environmental susceptibility factors. The only published genome scan for OCD identified a candidate region on 9p24 at marker D9S288 that met criteria for suggestive significance (Hanna et al., 2002). In an attempt to replicate this finding, we genotyped 50 OCD pedigrees using microsatellite markers spanning the 9p24 candidate region and analyzed the data using parametric and nonparametric linkage analyses under both a narrow phenotype model (DSM IV OCD definite, 41 affected sib pairs) and a broad phenotype model (DSM IV OCD definite and probable, 50 affected sib pairs). Similar to what was described by Hanna et al. (2002), our strongest findings came with the dominant parameters and the narrow phenotype model: the parametric signal peaked at marker D9S1792 with an HLOD of 2.26 (a ¼ 0.59), and the nonparametric linkage signal peaked at marker D9S1813 with an NPL of 2.52 (P ¼ 0.006). These findings are striking in that D9S1813 and D9S1792 lie within 0.5 cM (20%), or the expected effect size is very small or the total cost of genotyping controls in a study is large compared with the cost of screening the controls.

O9.5 GENE-GENE AND GENE-ENVIRONMENT INTERACTIONS IN BASIC BIOLOGICAL NETWORKS McClay JL and van den Oord EJCG Medical College of Virginia, Virginia Commonwealth University, Richmond, United States There is much current interest in gene-gene (GxG) and geneenvironment (GxE) interactions in the development of complex traits and diseases. The nature of biological processes giving rise to such interactions on a molecular level remains an unresolved conundrum. For example, it has been asserted that due to the ubiquity of biomolecular interactions in gene regulation and metabolic systems, GxG interaction will be a common feature of complex phenotypes (Moore J, Hum Hered, 56:73–82, 2003). Yet molecular interaction is very different, in many ways, to statistical interaction. To examine this issue, we simulated generic metabolic networks specifying linear enzyme cascades and examined if GxG interactions arose from their architectures. Models included all relevant cellular processes, i.e. transcription, mRNA transport and decay, enzyme translation and decay, and metabolite formation and decay. Diploidy was also explicitly modelled. The first simulated enzyme cascade was under No Control (NC), whilst the derivative was under the control of Competitive Inhibition (CI), a common form of negative feedback in metabolism. The quantitative outcome variable was the steady state concentration of the enzyme cascade product, which can be considered as a metabolic biomarker. Functional genetic variation was introduced at fixed frequencies into ‘‘populations’’ of models by altering rate constants for particular processes, such as transcription. Simulations suggested that GxG interaction is negligible for both NC and CI architectures. Negative feedback (CI), whilst increasing interaction on a molecular level, actually serves to reduce the modest GxG interaction across the range of ‘‘mutant’’ gene effects, relative to the NC model. GxG is at a maximum as the effect conferred by ‘‘mutant’’ alleles approaches null. However, multiple null alleles, present at high frequencies, are unlikely to be observed in viable populations. This implies that GxG interaction is unlikely to be a prevalent feature arising from metabolic pathways of the types described. Our results suggest that systems which are biologically complex, with multiple interacting components, may behave in a simple, tractable fashion on an aggregate level. We are also currently examining the effect of ‘‘environmental’’ influences by varying across genotypes the rate at which material flows into systems from external sources. From initial simulations, it would appear that statistical interaction is more likely to arise in our models

by this mechanism. Implications for complex trait research will be discussed.

O9.6 UNIVARIATE AND MULTIVARIATE QTL LINKAGE ANALYSIS OF BIPOLAR AFFECTIVE DISORDER DIMENSION SCALE (BADDS) SCORES IN BIPOLAR DISORDER Macgregor S,1 Jones I,4 Segurado R,3 Kirov G,4 Corvin A,3 Jones L,2 Hamshere M,4 Gill M,3 and Craddock N4 1 Bioinformatics and Biostatistics Unit, University of Wales College of Medicine, Cardiff, United Kingdom 2 Division of Neurosciences, University of Birmingham, Birmingham, United Kingdom 3 Trinity College, Dublin, Ireland 4 University of Wales College of Medicine, Cardiff, United Kingdom Genetic analysis of dimensional and quantitative phenotypic measures may provide useful additional information over and above that obtained from conventional categorical analyses. We have previously reported the results of a categorical affected only sib-pair linkage analysis of a set of 169 families multiply affected by bipolar disorder using a 10cM map of microsatellite markers (stage 1 of the Wellcome Trust UK-Irish sib-pair linkage study). Here we extend the analysis by considering dimensional measures of three domains of the phenotype using variance components analysis. Affected individuals were rated for 3 quantitative dimension scores: Mania (M), Depression (D) and Psychosis (P) (Craddock, 2000). Heritability estimates in this dataset (calculated using variance components analysis) were: M 0.14 (95% confidence interval 0,0.36), D 0.03 (C.I. 0,0.25) and P 0.52 (CI 0.28,0.76). Univariate variance components quantitative trait locus (QTL) analysis was performed for markers across the genome. Gene-dropping simulations were used to account for the non-normal distribution of the phenotypes. The maximum (simulation corrected) LOD scores for the three traits were M 1.05 (chromosome 5, 68 cM), D 1.56 (chromosome 1, 142 cM) and P 1.53 (chromosome 10, 100 cM). We conducted multivariate QTL analysis at two regions indicated by the univariate analysis. Multivariate models which considered polygenic and QTL specific effects were applied in a mixed model maximum likelihood framework. All possible covariances between the different traits were fitted for the polygenic effects whilst a range of simplifying assumptions were applied for the QTL effects. We considered bivariate analysis of the most heritable dimensions (M and P) as well as trivariate analysis of all three dimensions. The univariate results gave weak evidence (P0.05) for a QTL in all 3 traits at chromosome 19 (56 cM), prompting us to look at all 3 traits together. Fitting this trivariate model yielded an asymptotic P-value of 0.0025 for the test of QTL plus polygenic effects versus polygenic only effects at this position. On chromosome 5 (60 cM), univariate peaks were found for M and P, a bivariate analysis yielded an asymptotic P-value of 0.001 for the QTL test. We are currently conducting gene-dropping simulations to investigate further the significance of these multivariate results. We plan to follow up these results in the stage 2 UK-Irish linkage sample.

O9.7 RECURSIVE PARTITIONING: A DATA MINING STRATEGY TO DETECT GENOTYPE-PHENOTYPE CORRELATIONS IN LINKAGE SAMPLES Schulze TG,1 Xu W,3 DePaulo JR,4 McMahon FJ,2 and Greenwood C3 1 Division of Genetic Epidemiology, Central Institute of Mental Health, University of Heidelberg, Mannheim, Germany 2 Genetic Basis of Mood and Anxiety Disorders, NIMH, NIH, Bethesda (MD), United States 3 Genetics and Genomic Biology, Hospital for Sick Children, Toronto (ON), Canada 4 Department of Psychiatry, The Johns Hopkins University, Baltimore (MD), United States Linkage analyses in psychiatric phenotypes have delineated several candidate regions for fine-mapping endeavors, some of which have already yielded promising susceptibility genes. Still the progress is slow and unpredictable. One reason for this may be the difficulty in choosing among a range of possible phenotype definitions, especially when several covariates may alter disease risk. Innovative analytical strategies to disentangle potential genotype-phenotype correlations

Abstracts are needed. Here we present a novel approach that uses a recursive partitioning algorithm (RPA) to search for phenotypic features that correlate with linkage evidence. We combined a standard linkage analysis tool (GENEHUNTER) with a RPA that uses IBD sharing and covariate information to identify combinations showing strong evidence for linkage. Our method can be thought of as a data-mining approach which may not only be used to assess the influence of clinical covariates but also to detect gene-environment interactions for complex diseases. This tree-based model uses all the phenotype information entered to find homogeneous, covariate-defined subgroups of the affected pairs, where the total evidence for linkage including covariate splits is significantly greater than in the model without the covariate splits. If a covariate helps define homogeneous subgroups, then the pairs will be ‘‘split’’ into two groups. The algorithm will then recursively try to split each subgroup again until no more covariates are helpful. The algorithm is implemented in a program which first generates the tree-structure of each marker (splitting), then uses a cross-validation algorithm to choose the best tree-size at each marker (pruning). A bootstrap method is then used to get the best posterior tree-structure according to the chosen tree-size and to provide the final tree-structure for each marker. To test our algorithm, we used the data on 58 families from the Johns Hopkins/Dana bipolar pedigrees, for which linkage to chr. 18 has previously been reported. In our data set, we identified several interesting regions on chr. 18q where covariates defined subgroups with stronger evidence for linkage than a standard linkage analysis. For some chromosomal locations, evidence of linkage was detected only when covariates were included. Covariates that showed strong interactions with genotype were age at onset, suicidal behavior, and alcohol/substance abuse. RPA may be a valuable tool to delineate the impact of clinical covariates or unveil gene-environment interactions. RPA can make use of the entire available sample and of a multitude of phenotypic variables at the same time, in contrast to stratification procedures that inherently suffer from limited power and the problem of multiple-testing.

O9.8 QTLS FOR ANXIETY AND DEPRESSION USING EXTREME DISCORDANT AND CONCORDANT TWINS AND SIBS IN AUSTRALIA AND THE NETHERLANDS Birley AJ,1 Beem AL,2 Neale BM,3 Sullivan PF,4 Boomsma DI,2 and Martin NG1 1 Queensland Inst Med Res, Brisbane, Australia 2 Biological Psychology, Vrije Universiteit, Amsterdam, Netherlands 3 SGDP, Inst of Psychiatry, London, United Kingdom 4 Genetics, Univ North Carolina, Chapel Hill, United States There is ample evidence from twin studies that genetic factors predispose to the common psychiatric diagnoses of anxiety and depression. Furthermore, these genetic influences are largely shared in common with the personality trait neuroticism. We have conducted extreme discordant and concordant sib pair (EDAC) studies in both Australia and Holland to search for linkage with QTLs influencing anxiety and depression. Around 10,000 sib pairs from twin families in each country were screened for neuroticism and extreme pairs, their sibs and parents were selected for genome scan. Maximum likelihood multivariate variance components approaches were used to detect QTLs linked to markers from the genome scan (using full multipoint information), in addition to nonparametric approaches. Preliminary results for both studies were presented at the WCPG in Brussels, with suggestive linkage in regions on chromosomes 1, 2, 4, 8, 9, 17 and 19 where others have also reported linkage. We have just obtained a 5 cM genome scan on a further 1000 individuals from our EDAC study representing a fourfold increase in information. New analyses are underway and results will be presented at the congress.

P1.1 HARM AVOIDANCE AND SELF-DIRECTEDNESS AS INTERMEDIATE PHENOTYPES FOR DEPRESSION Smith DJ,1 Duffy L,1 Stewart M,2 Muir WJ,1 and Blackwood DHR1 1 University of Edinburgh, Edinburgh, United Kingdom 2 Herriot-Watt University, Edinburgh, United Kingdom The personality dimension of harm avoidance (HA), as measured by the Temperament and Character Inventory (TCI), has been widely

31

associated with depression and there is preliminary evidence that high HA may represent a trait marker for depression. However, many of the studies performed in this area so far are limited by the use of heterogeneous samples of depressed patients and by the confounding effect of depressed mood during personality testing. In the current study, we use the TCI-125 to compare personality dimension scores between a group of euthymic young adults with a strongly genetic depressive sub-type (recurrent early-onset major depressive disorder, RE-MDD) and a group of well-matched controls. Fifty two euthymic young adults with a past history of RE-MDD were recruited from consecutive referrals to a psychiatric clinic at a university health service. Eighty nine controls were also recruited. Euthymia was established in subjects by a score of less than 9 on the Hamilton Rating Scale for Depression (HRSD) and in controls by a Becks Depression Inventory (BDI) score of less than 10. All participants completed the TCI-125. Subjects and controls were well matched in terms of sociodemographic profile and all were clinically euthymic at the time of personality testing. RE-MDD subjects scored significantly higher than controls on the temperament dimension of harm avoidance (HA, mean score 14.5 versus 7.8, P < 0.0001) and significantly lower than controls on the character dimension of self directedness (SD, mean score 14.1 versus 19.9, P < 0.0001). High HA and low SD may represent intermediate phenotypes for depressive disorder. Further research is needed to replicate these findings and to examine the usefulness of these personality traits in genetic studies of mood disorders.

P1.2 A GENOME WIDE SCAN FOR LINKAGE IN FOUR FAMILIES WITH DEPRESSION AND CO-MORBID UNEXPLAINED SWELLING Anderson C,1 Maclean A,2 Dunnigan M,3 Pelosi A,4 Murray V,5 McKee I,4 McDonald G,6 Visscher P,1 Muir W,2 and Blackwood D2 1 School of Biological Sciences, Ashworth Laboratories, University of Edinburgh, Edinburgh, United Kingdom 2 School of Molecular and Clinical Medicine, University of Edinburgh, Edinburgh, United Kingdom 3 University Department of Nutrition, Glasgow Royal Infirmary, Glasgow, United Kingdom 4 Department of Psychiatry, Hairmyres Hospital, East Kirkbride, United Kingdom 5 Yorkhill Hospital, Glasgow, United Kingdom 6 Dykebar Hospital, Paisley, United Kingdom Major depressive disorder (unipolar depression) is a complex heterogeneous condition and when attempting to identify genetic variation contributing to the phenotype it may be advantageous to focus on clinically homogenous sub groups of the disorder. Here we describe families with a clinical phenotype characterised by major depression comorbid with unexplained swelling (also known as idiopathic oedema or the fluid retention syndrome). Unexplained swelling affects several sites including fingers, hands, breasts, abdomen, legs and feet, and is more common in women than men. Unexplained swelling is often familial, occurs at all ages, and is regularly comorbid with depression. The cause is unknown and sufferers often present to primary care physicians and gynaecologists rather than psychiatrists. We describe four unrelated families from central Scotland with this phenotype. A genomewide scan for linkage was performed using 405 microsatellite markers. Of the 116 individuals within the 4 families, 66 were genotyped, 39 of which were classified as ‘‘affected’’ with a diagnosis of unexplained swelling and a life-time history of recurrent major depression often with marked features of anxiety. Initial two-point parametric linkage analysis was carried out using LINKAGE, revealing 32 markers with a LOD score greater than 1. No marker reached genomewide level of significance, LOD > 3.40. Marker D8S277 produced suggestive linkage under both the dominant and recessive models (LOD ¼ 2.14 and 2.15 respectively). 10 other markers displayed suggestive linkage, with LOD > 1.80. A further 26 markers were genotyped in regions of previous suggestive linkage. Markers D3S1263, D4S2952, D14S275 all produced LOD scores above 2.00. Whilst these loci of suggestive linkage should be interpreted cautiously, these regions clearly deserve further study. LOD score statistics can be sensitive to changes in model parameters and it is proposed to carry out non-parametric and multipoint analysis to elucidate the genetic architecture of disease in this cohort of families with depression.

32

Abstracts

P1.3 ASSOCIATION OF CYTOKINE POLYMORPHISMS WITH DEPRESSIVE MOOD IN OLDER PERSONS IN A COMMUNITY-BASED SAMPLE: A POSSIBLE ROLE FOR IL4R-503 T/C, ILR4-576 A/G AND IL6-174 G/C PROMOTER POLYMORPHISMS AS INFLAMMATORY MARKERS IN THE PATHOPHYSIOLOGY OF DEPRESSION Baune BT,1 Funke H,2 Junker R,3 Arolt V,1 and Berger K4 1 Department of Psychiatry, Universityof Muenster, Muenster, Germany 2 Institute of Vascular Biology, University of Jena, Jena, Germany 3 Institute of Arteriosclerosis Research, University of Muenster, Muenster, Germany 4 Institute of Epidemiology and Social Medicine, University of Muenster, Muenster, Germany Inflammation has been suggested to play an important role in the pathophysiology of major depression. Immune activity in the brain and the endocrine system are linked by systemic cytokines. High levels of cytokines (e.g. IL-1, IL-4, IL-6, IL-10, TNF-alpha) may induce symptoms (e.g. fatigue, decreased psychomotor activity, sleep disturbances) leading to depression. We studied the association of systemic inflammatory markers (CRP, fibrinogen) and several cytokine polymorphisms with depressed mood in older people. Cross-sectional study examining the association between cytokine polymorphisms and depressive mood in an elderly general population in Germany. 259 participants were studied for depressed mood applying the CES-Dscale, blood was genotyped for inflammatory markers, C-Reactive Protein and fibrinogen were analysed with standard laboratory methods. Differences in proportions were tested with Pearsons Chi2test, and associations were estimated with Odds Ratios in logistic regression models. The prevalence of depressed mood (CES-D>16) was 12.7% (n ¼ 33) in this population, significantly more often in female participants (P ¼ 0.007). Significant dose-response relationships were found between the number of variant allels for IL4R-503 T/C, IL4R-576 A/G, IL6-174 G/C promoter polymorphisms and depressed mood. The age and gender adjusted Odds Ratio associated with depressed mood was significantly increased for the homozygous genotype in IL-6-174G/ C (OR ¼ 3.5, 95% CI 1.6-8.1, P ¼ 0.003). High CRP (>3.5 mg/dl) levels were significantly associated with IL6_572 G/C (OR ¼ 3.5, 95% CI 1.2– 9.8, P ¼ 0.017) and TGFB1_509 C/T (OR ¼ 2.0, 95% CI 1.1–3.8, P ¼ 0.035) promoter polymorphisms. Our results support the supposed role of cytokines in depression. Moreover, increased CRP is also associated with cytokines polymorphisms. Our results need to be replicated in independent larger samples of people with depressed mood.

P1.4 AFFECTIVE DISORDERS AND THE HAPLOTYPES OF SEROTONIN TRANSPORTER GENE (SLC6A4): A FAMILY-BASED ASSOCIATION STUDY Zhang X-N1 and Gao S-G2 1 Department of Medical Genetics, School of Medicine, Zhejiang University, Hangzhou, China 2 Ningbo Kangning Mental Hospital, Ningbo, Zhejiang Province, Ningbo, China Serotonin (5-hydroxytryptamine, 5-HT) is a neurotransmitter in the central and peripheral nervous systems. Following release, 5-HT is actively cleared from synaptic spaces by a high-affinity, Na(þ)- and Cl ()-dependent transporter localized in presynaptic neuronal membranes. The brain 5-HT transporter appears to be a principal site of action of many tricyclic antidepressants (such as impramine) and may mediate behavioral and/or toxic effects of cocaine and amphetamines. The SLC6A4 gene, also 5-HTT, localized to chromosome 17q11.1-q12, spans 31 kb and consists of 14 exons. The serotonin transporter encoded by SLC6A4 is the target of an important class of antidepressant drugs, the serotonin selective reuptake inhibitors. To investigate the association between affective disorder and SLC6A4 gene. We selected 72 Southern Han Chinese core pedigrees, sum to 222 family members (including 56 bipolar disorder and 34 major depression individuals) for the analyzing the three polymorphisms: a deletion/insertion of 44bp in the promoter region (5-HTTLPR), a variable number tandom repeat located in intron 2(5-HTTVNTR) and a PstI digestion polymorphism AG/TTAC in the 30 untranslated region (30 UTRG/T) of SLC6A4 gene. Genotyping were carried out using polymerase chain reaction and restriction fragment length polymorphism (PCR-RFLP). The transmission disequilibrium test (TDT) of multiple alleles showed an association between affective disorder and 5-HTTVNTR-5-HTT30 UTRG/T haplo-

type in SLC6A4 gene (TDT-chi2 ¼ 4.083, P ¼ 0.0434.p was corrected by Monte Carlo simulation after 1000 replications). There were no linkage disequilibrium between the other multi-markers (5-HTTLPR-5HTTVNTR, 5-HTTLPR-5-HTT 30 UTRG?T and 5-HTTLPR-5-HTT VNTR-5-HTT30 UTRG-T) and the disease (P > 0.05). The results suggested that SLC6A4 gene may influence clinical features of affective disorders, as other researchers reported previously. P1.5 SNP AND HAPLOTYPE ANALYSIS OF A NOVEL TRYPTOPHAN HYDROXYLASE ISOFORM (TPH2) GENE PROVIDE EVIDENCE FOR ASSOCIATION WITH MAJOR DEPRESSION Zill P,1 Baghai TC,2 Schu¨le C,2 Zwanzger P,2 Eser D,2 Rupprecht R,2 Mo¨ller HJ,2 Bondy B,1 and Ackenheil M1 1 Department of Psychiatry of the LMU Munich, Section for Psychiatric Genetics and Neurochemistry, Munich, Germany 2 Department of Psychiatry of the LMU Munich, Munich, Germany Several lines of evidence indicate that disturbances of the central serotonergic system are involved in the pathophysiology of affective disorders. Tryptophan hydroxylase (TPH), the rate limiting enzyme in the synthesis of the neurotransmitter serotonin (5-HT) is widely discussed as a major candidate gene in many psychiatric disorders, especially depression and suicidal behavior. Numerous studies about genetic variants or different expression levels of the TPH gene related to depression, schizophrenia, alcoholism, drug abuse, aggression and sucidality have been reported, but none of them have led to convincingly confirmed results up to now. Recently a second TPH isoform (TPH2) was identified in mice which was exclusively present in the brain. In a previous post-mortem study of our own group we could demonstrate that TPH2 is also expressed in the human brain, but not in peripheral tissues. A previous screening approach in our own group with a few validated SNPs in the TPH2 gene detected an association between MD and the SNP rs1386494. Therefore, the aim of the present study was to investigate a probable role of polymorphisms around this significant SNP in the TPH2 gene in the etiology of major depression.This is the first report of an association study between polymorphisms in the TPH2 gene and major depression (MD). We performed SNP, haplotype and linkage disequlibrium studies on 300 depressed patients and 265 healthy controls with 10 SNPs in the TPH2 gene. Significant association was detected between one SNP (P ¼ 0.0012, global P ¼ 0.0051) and MD. Haplotype analysis produced additional support for association (P < 0.0001, global P ¼ 0.0001). Our finding provide evidence for an involvement of genetic variants of the TPH2 gene in the pathogenesis of MD and might be a hint on the repeatedly discussed duality of the serotonergic system. These results may open up new research strategies for the analysis of the observed disturbances in the serotonergic system in patients suffering from several other psychiatric disorders. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine.’’

P1.6 DISTRIBUTION OF DEPRESSIVE SYMPTOMS IN A POPULATION-BASED CONTROL SAMPLE Ohlraun S,1 Hoefels S,2 Eckermann A,2 Pottgiesser S,2 Weber S,2 Neidt H,3 Braeman C,1 Schumacher J,3 Cichon S,3 Maier W,2 Propping P,3 Schulze TG,1 Noethen MM,3 and Rietschel M1 1 Central Institute of Mental Health, Division of Genetic Epidemiology in Psychiatry, Mannheim, Germany 2 University of Bonn, Dept of Psychiatry, Bonn, Germany 3 University of Bonn, Institute of Human Genetics, Bonn, Germany To identify susceptibility genes for affective disorders and schizophrenia, large samples of controls, as well as patients, are necessary In case-control studies of disorders such as depression, that have a high frequency in the general population, it is essential to assess the prevalence of the disorder in the control sample. Working within the framework of the National Genome Research Network, we contacted 6000 people from the general population. After giving informed consent, probands were rated with standardized phenotyping instruments for psychiatric, neurological and somatic disease. Of the 1195 who participated, 49% were male and 51% female with an average age of 48 years (SD 15,4), ranging from 19 to 80. 84% were of German decent, defined as having German-only grandparents. 11% fulfilled DSM-IV criteria for a lifetime diagnosis of major depression, 25% had

Abstracts sub-threshold depression. The latter is defined as the presence of 1-9 depressive symptoms during the same 2-week period-one symptom being either depressed mood or loss of interest or pleasure-where the symptoms do not cause impairment, and are not due to the effect of substances, a general medical condition, or bereavement. 66% of the population had never had any depressive symptoms. The ratio of males to females was 1:2.2 in the major depression group, 1:1.5 in the subthreshold depression group and 1:0.8 in the healthy. While the groups with major depression and sub-threshold depression differed significantly from the healthy (P ¼ 0.000), the difference between the groups with major and sub-threshold depression did not reach significance (P ¼ 0.118). In addition we tested whether the three groups varied in reported family history (FH) of depression in first or second degree relatives. 17% of all participants reported a positive FH: 29% in the major depression group, 21% in the sub-threshold depression group and 14% in the healthy. While the groups with major depression and sub-threshold depression differed significantly from the healthy (P ¼ 0.000 and P ¼ 0.007 respectively) the difference between the groups with major and sub-threshold depression did not reach significance (P ¼ 0.75). Our control sample allows us to match for not only sociodemographic, but also for differentiated psychopathological variables ˆ(super-normals’’). In addition, our data suggest that female to male ratio and positive FH both increase with the severity of depression. P1.7 THE ROLE OF SOCIAL SUPPORT IN FAMILIES AT HIGH RISK OF DEPRESSION Reynolds NJ,1 Wilkie E,1 Katz R,2 and McGuffin P1 1 Institute of Psychiatry, London, United Kingdom 2 University of Toronto, Toronto, United States There is consistent evidence that unipolar depression is familial. Nevertheless, most relatives of depressed probands do not suffer from the disorder. There is also some evidence from community-based studies that social support may have a protective role. We postulated that there would be differences in the social networks of those members of high-risk families who suffer from depression compared with those who do not. The subjects were 246 first or second-degree relatives of depressed probands between the ages of 18 to 83. They were interviewed using the Past History Schedule, the Present State Examination and the Interview Measure of Social Relationships. Logistic regressions were used to assess the aspects of social support that predicted current depressive disorder or a past history of depression. The only gender difference in social support was that men had higher levels of social contact in the week before interview. Women had significantly more current depression and lifetime ever depression than men. Overall, the only predictor of lifetime depression was marital status and this remained so for women when the sample was divided by gender. However lifetime depression in men was significantly associated with having fewer primary support group members. Overall current depression was predicted by number of children at home and marital status and was inversely related to number of grown-up children and amount of contact with friends and relatives. In women marital status and children at home were the only significant predictors but in men current depression was associated with lower social contact in the past week. The relatives of depressed patients show differences in their degree of social support and this is associated with whether they themselves suffer from current depression. Lifetime depression was associated with marital status in women and a small primary support group in men. The direction of cause is difficult to establish.

P1.8 SOCIAL ADJUSTMENT COULD BE ASSOCIATED WITH THE SEROTONIN TRANSPORTER GENE IN REMITTED PATIENTS WITH MOOD DISORDERS AND HEALTHY SUBJECTS Mandelli L,1 Serretti A,1 Calati R,2 Massa R,2 and Smeraldi E1 1 Vita-Salute University, Milan, Italy 2 San Raffaele Hospital, Milan, Italy Overall studies suggest mainly disadvantageous effects of the short 5HTTLPR variant on human behavior. The 5-HTTLPR short variant has been associated with anxiety-related traits in normals, poor antidepressant response in mood disorders and other mainly negative traits. We hypothesized a possible counterbalancing effect of this

33

polymorphism on other human features to explain its evolutionary meaning. For this aim we investigated the social functioning of both healthy subjects and patients affected by Mood Disorders. For the present study we collected a sample of 379 subjects: 285 were patients affected by Mood disorder (MOOD), both major depressives (MDD ¼ 99) and bipolars (BD ¼ 186), 94 were healthy controls. Individual degree of social adjustment was evaluated using the Social Adjustment Scale (SAS). At the evaluation, all patients were in full remission from manic or depressive symptomatology for at least 3 months. The 5-HTTLPR s/s genotype was related with a better working functioning for controls, a better financial adjustment for MOOD, and with more careful behaviors towards children for males. On the other hand, those subjects homozygotes for the short 5-HTTLPR variant, showed an impaired functioning in the social activities and leisure time. Among BD, the 5-HTTLPR s/s genotype was also associated with impaired functioning in the ‘‘extended family’’ area. All results should be considered with cautions since no multiple comparison correction was performed. In conclusion, 5-HTTLPR variants have a complex interaction of favourable and negative effects on different levels of human functioning, including the social adjustment. P1.9 TEMPERAMENT AND CHARACTER IN MOOD DISORDERS: INFLUENCE OF DRD4, SERTPR, TPH AND MAO-A POLYMORPHISMS Calati R,1 Serretti A,2 Mandelli L,2 Landoni S,1 and Smeraldi E1 1 San Raffaele Hospital, Milan, Italy 2 Vita-Salute University, Milan, Italy Several studies indicated that genetic mechanisms underlie at least a portion of individual differences in personality traits. In this paper we investigated possible associations between temperamental traits in Mood disorders patients and polymorphisms in the dopamine receptor D4 gene (DRD4), serotonin transporter gene (SERTPR), tryptophan hydroxylase gene (TPH) and monoamine oxidase A gene (MAO-A). Two hundred and seven euthymic subjects, affected by Major depressive disorder (n ¼ 73) and Bipolar disorder (n ¼ 134), were assessed by the Cloninger’s Temperament and Character Inventory (TCI) and they were typed using PCR techniques. TCI scores of depressed patients were also compared to those obtained on a sample of healthy controls. Possible stratification factors such as demographic, clinical and other temperamental factors were systematically evaluated. Analyses evidenced that homozygosity for the short DRD4 variants reduced Harm avoidance scores (P ¼ 0.05), while homozygosity for the short SERTPR allele reduced Novelty seeking (P ¼ 0.006) and Self trascendence scores (P ¼ 0.009). Reward dependence scores were lowered by homozygosity for the long SERTPR allele (P ¼ 0.03) and by the C TPH allele (P ¼ 0.048). Reward dependence and Cooperativeness were lowered by homozygosity for the short DRD4 allele (P ¼ 0.003, P ¼ 0.0001). Finally, the long MAO-A allele decreased Persistence scores among females (P ¼ 0.02). In conclusion, polymorphisms of the SERTPR, DRD4, TPH and MAO-A genes may explain a portion of personality differences among individuals and temperamental vulnerabilities to develop a depressive disorder.

P1.10 INSOMNIA AND MOOD DIURNAL VARIATIONS IN MOOD DISORDER PATIENTS: THE INVOLVEMENT OF CLOCK AND PER3 GENES Massa R,1 Serretti A,2 Benedetti F,1 Artioli P,1 Mandelli L,2 Ploia C,1 Pirovano A,2 and Smeraldi E2 1 Ville turro San Raffaele, Milan, Italy 2 Vita-Salute University, Milan, Italy Sleep disturbances are commonly observed in mood disorders, and sleep manipulations can influence clinical manifestations of depressive and bipolar illness. In the present study we investigated the possible effect of the 3111T/C CLOCK gene polymorphism and three PER3 gene polymorphisms on insomnia symptomatology in mood disorder patients. Patients included into the study were genoyped for the CLOCK gene (n ¼ 620) and for three PER3 gene polymorphisms (Per3ex15T1940G, Per3ex18ins/del and Per3ex18T3110C) (n ¼ 638). All subjects genotyped for the CLOCK gene were evaluated for the lifetime symptomathology with the OPCRIT checklist, all subjects genotyped for the PER3 gene and 178 subjects genotyped for the CLOCK gene were observed during antidepressant treatment with SSRIs for six weeks and evaluated with the Hamilton Depression

34

Abstracts

Rating Scale. We found a significantly higher recurrence of initial (P ¼ 0.0001), middle (P ¼ 0.0009) and early (P ¼ 0.0008) insomnia among homozygotes for the CLOCK*C variant, a similar trend was found regarding decreased need of sleep among bipolars (P ¼ 0.0074). Moreover, the CLOCK*C variant was significantly associated with a more severe insomnia during antidepressant treatment throughout the six weeks (P ¼ 0.026). All three Per3 gene polymorphisms were not associated with insomnia severity during SSRIs treatment, however the Per3ex18TC*C/C genotype was found to be significantly associated with the non-melancholic mood diurnal variation: ‘‘worse mood in the evening’’ at the baseline (P ¼ 0.03) till the second week of treatment (P ¼ 0.03). This preliminary observations supports the hypothesis of an involvement of the CLOCK gene polymorphism in the sleep disregulations in mood disorders, while the Per3 gene polymorphims seems to be linked to mood diurnal variations. P1.11 IDENTIFICATION OF TWO NEW VARIANTS IN THE CIRCADIAN CLOCK GENE Pirovano A, Serretti A, Fontana V, Ploia C, Tubazio V, Catalano M, and Smeraldi E Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy Circadian (24 h) rhythm of biological processes is an important feature of most living organism and it has been conserved during the evolution. This activity is controlled by endogenous self-sustaining oscillators that represent the core of biological clock. At the cellular level, the circadian clock system is due to expression of different genes that include the Circadian Locomotor Output Cycles Kaput (Clock) gene. The Clock gene, which is the first essential component of the mammalian clock, codifies for a protein that acts as a transcriptional activator. In humans, clock genes play a central role in generating and regulating circadian rhythms, therefore, a number of researcher hypothesized that mutation in this genetic system could be related to observed circadian dysfunctions in mood disorders. In particular, the hClock gene seems to be a potential candidate for susceptibility to circadian rhythm sleep disorders, which are often observed in major depressive episode. Recently, our research group validated the possible relationship between C allele and evening preference for usual activity reported by Katzenberg and collaborators (Katzenberg D et al., Sleep 1998, 21: 569–576). In fact, studying the T3111C polymorphism, we demonstrated that Clock gene has an effect on circadian rhythms in subjects affected by mood disorders, resulting in sleep disturbances. In order to investigate the possible effect of the T3111C Clock gene polymorphism (Human Genome Database accession number GDB: 9785615) on sleep disregulations, a sample of 1113 subjects (479 major depressive disorder, 558 bipolar disorder and 76 healthy controls) was genotyped. During genetic analysis, the PCR product of two patients shown an intriguing pattern on Denaturant Gradient Gel Electrophoresis (DGGE) migration, related to a new mutations in the Clock gene. Sequencing analysis identified the specific DNA alterations: the first patient showed a G/T point mutation at the 3117 nucleotide of Clock gene, while in the second patient a A/G substitution was localized at nucleotide 3125 (Gene Bank Accession Number AF011568). In both patients we observed interesting sleep abnormalities. Such single nucleotide mutation (3117 G/T and 3125 A/G) could be responsible for alteration in Clock protein translation. In fact, sequence modification in the 30 flanking region have been often shown to affect mRNA stability and half-life. The low frequency of these new mutations could suggest the functional importance of this untranslated sequence, whose integrity is probably necessary to obtain a right level of protein translation.

P1.12 SYMPTOMATOLOGIC ANALYSIS OF MOOD DISORDERS: POSSIBLE CANDIDATE GENES Fontana V, Serretti A, Pirovano A, Tubazio V, Smeraldi E Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy The use of traditional psychiatric diagnoses as phenotype definition of major psychosis is one of the main problem that hampers psychiatric research, since it does not guarantee biological homogeneity. Five symptomatology factors were extracted by statistical analyses from the Operational Criteria Checklist for Psychotic Illness (OPCRIT): Excitement, Depression, Delusion, Disorganisation and Negative. We tested

the possible involvement of individual genetic pattern in symptomatology, in a sample of inpatients affected by bipolar and unipolar disorders. We focused on some gene variants involved both in metabolic Serotonin pathway like MAO-A (Monoamine Oxydase A), COMT (catechol o-methyltransferase) and 5-HT2A (serotonin receptor 2A: 102T/C variant and 1420C/T polymorphism) 5-HT1A (serotonin receptor 1A, and in second messengers signalling pathway Gb3 (G Beta 3 subunit) and GSK3b (Glycogen synthase kinase: 50 T/C). In particular, for each polymorphism we studied samples of subjects affected by mood disorders as follow: MAO-A (n ¼ 748), COMT (n ¼ 458), 5-HT2A (n ¼ 468 and n ¼ 404), 5-HT1A (n ¼ 170), Gb3 (n ¼ 418), GSK3bn ¼ 335). We observed an association between 5HT2A 102T/C CC genotype and depression and disorganization symptomatology (P ¼ 0.016), particularly in bipolar subjects, the GSK3b (50 T/C) genotype within the affective group was associated with delusion symptomatology (P ¼ 0.00442, this association was particularly marked in the group of BP subjects separately analyzed, in which both genotype and allelic variants were associated with delusion symptomatology (P ¼ 0.0069, P ¼ 0.0031). The other studied gene variants showed no significant association with any factors. Our analysis did not evidence strong liability factors for affective disorders, even if further studies on GSK3b role in symptomatology could be necessary.

P1.13 SEX DIFFERENCES IN SYMPTOMS IN SIBLINGS WITH MAJOR DEPRESSION Moskvina V,2 Brewster S,3 Ferrero F,6 Gill M,8 Jones I,2 Jones L,4 Maier W,9 Mors O,11 Owen M,2 Preisig M,7 Reich T,12 Reitschel M,10 Farmer A,5 McGuffin P,5 Craddock N,2 and Korszun A1 1 Department of Psychiatry, Barts and the London Queen Mary School of Medicine and Dentistry, London, United Kingdom 2 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 3 GlaxoSmithKline Medical Genetics, Middlesex, United Kingdom 4 Department of Psychiatry, University of Birmingham, Birmingham, United Kingdom 5 Social Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, London, United Kingdom 6 Department of Psychiatry, University of Geneva, Geneva, Switzerland 7 Department of Psychiatry, University of Lausanne, Lausanne, Switzerland 8 Trinity Centre for Health Sciences,Trinity College, Dublin, Ireland 9 Department of Psychiatry, University of Bonn, Bonn, Germany, 10 Division of Genetic Epidemiology in Psychiatry, University of Mannheim, Mannheim, Germany 11 Department of Psychiatric Demography, University of Aarhus, Aarhus, Denmark 12 Department of Psychiatry, Washington University School of Medicine, St. Louis, United States Women have a higher prevalence of depression and there is debate as to whether the same genetic and aetiologic risk factors for depression are operative in men and women. The aim of this study was to explore sex differences in depressive symptom patterns in a large group of siblings with depression. A total of 1034 subjects, including 475 sibling pairs, with a past history of at least two depressive episodes, were recruited from the depression network study (DeNT), a large scale multi-centre collection of families affected by recurrent unipolar depression (Korszun A et al. (2004) Familiality of Symptom Dimensions in Depression. Arch Gen Psychiatry 61:468–474). Subjects were interviewed using the SCAN interview and diagnosed according to DSM-IV and ICD-10 using a computerized scoring programme (CATEGO5). Sex differences in depressive symptoms were assessed using analysis of variance (ANOVA) and each symptom was first adjusted for age and family effect. The SAS PROC MIXED procedure was used to fit a mixed model with age as a covariate and family effect as a random effect. The residuals were considered as continuous variables and used for further analysis. Familial effect was identified by calculating intraclass correlations (ICC) for depressive symptoms (adjusted by age only) between affected sibling pairs. Women showed an earlier onset of depression compared to men (23.2 (CI: 22.3–24.1) vs. 25.2 (CI: 23.7– 26.7) P ¼ 0.005). Also, women had a significantly higher frequency of the reversed vegetative symptoms of ‘‘atypical’’ depression: appetite gain and hypersomnia. Intraclass correlations were significant between female siblings for appetite gain (0.23, P < 0.001) whereas

Abstracts male siblings showed a positive correlation for hypersomnia (0.30, P ¼ 0.016). There was no correlation for either of these symptoms between siblings of different sex. These data support previous findings that atypical depression is more common in women (Matza LS et al. (2003). Depression with Atypical Features in the National Comorbidity Survey. Arch Gen Psychiatry 60:817–826) and demonstrate that sex significantly modifies some clinical features of depression. This may result from a modification of the effect of predisposing genetic factors or from the presence of different genetic risk factors for depression in the two sexes. P1.14 RESEQUENCING OF THE TRPM2 AND TSPEAR GENES PROVIDES ADDITIONAL SNPS IN THE 21Q22.3 REGION OF CHROMOSOME 21 THAT SHOW STRENGTHENED ALLELIC AND HAPLOPTYPIC ASSOCIATION WITH BIPOLAR DISORDER McQuillin A,1 Bass N,1 Kalsi G,1 Lawrence J,1 Smyth C,1 Curtis D,2 and Gurling HMD1 1 London, United Kingdom 2 Ireland, London, United Kingdom Linkage analysis with bipolar families has identified the 21q22.3 region of chromosome 21 as a genetic susceptibility locus for bipolar affective disorder. We previously reported evidence for allelic association with bipolar disorder using the marker D21S171 at 21q22.3 (P ¼ 0.0024). Since then we have found two SNPs in the region, rs1556314, rs1785467 also showing evidence for allelic association (P ¼ 0.0192 and P ¼ 0.0303 respectively). Flanking microsatellite markers only showed significant evidence for allelic association if a correction for multiple alleles was not made. We then sequenced the coding regions of the TRPM2 and TSPEAR genes in cases of bipolar disorder who have inherited susceptibility alleles and have found additional SNPs. Haplotype analyses using GENECOUNTING with these nine new SNPs gave further independent evidence for association with bipolar disorder with an empirical significance of P < 0.00001 from a permutation test. The GENECOUNTING data indicates that a limited number of haplotypes show association with bipolar disorder. The estimated frequency of two haplotypes was approximately 5% higher in bipolar cases relative to controls. The frequency of haplotype 3 is 35.6% in bipolar cases and 30.3% in controls. The frequency of haplotype 16 is 8.3% in bipolar cases and 3.4% in controls. No genetic variants have yet been found that could have an aetiological role in increasing genetic susceptibility to bipolar disorder. The research was funded by Medical Research Council grant G9623693N and by a research lectureship from the Priory Hospital, London, to Dr. J Lawrence and by the Neuroscience Research Charitable Trust. We would like to acknowledge the help of Hamish Scott and Stylianos Antonarakis for providing unpublished mapping information and primer sequences.

P1.15 ASSOCIATION STUDY OF A BDNF PROMOTER POLYMORPHISM AND AFFECTIVE DISORDERS IN BULGARIAN SAMPLE Chorbov V,1 Milanova V,2 Kostov C,2 Nikolova-Hill A,4 Stoyanova V,2 Krastev S,2 Onchev G,2 Mitev V,3 Kremensky I,1 and Kaneva R1 1 Laboratory of Molecular Pathology, University Hospital of Obstetrics, Medical University, Sofia, Bulgaria 2 Department of Psychiatry, Medical University, Sofia, Bulgaria 3 Department of Chemistry and Biochemistry, Medical University, Sofia, Bulgaria 4 Bunbary Regional Hospital, SW Mental Health Service, Bunbary, Australia Mood disorders fall into two main groups: Major depressive disorder (MDD)) and bipolar affective disorder (BPAD) where depression alternates with mania. Despite the strong genetic component, no vulnerability or susceptibility genes have been found yet and the mechanisms of pathogenesis are not well defined. Brain-derived neurotrophic factor is a neuroprotectant, induced by cortical neurons and necessary for the survival of striatal neurons in the brain. Molecules that are critical in the development and maintenance of cortical neurons and cortical synapses, such as BDNF, may play a role in the neuropathology of Affective Disorders. These suggestions prompted us to examine BDNF as a candidate gene for AD. Following previous reports we tested for association a dinucleotide polymorphism situated at about 1kb 50 of the transcription initiation site in a case-

35

control study. A group of 94 patients with BPAD (73 female and 21 male), 51 patients with Unipolar Affective Disorder (38f and 13m) were compared with 96 healthy controls (73f and 23m), all of them of Bulgarian origin. Ten alleles, four with frequency more than 1% and six rare alleles were found. No significant allele-wise or genotype-wise association was found between the two groups of patients and the controls. When divided by sex, statistically significant association was obtained with allele 5 in the female bipolar patients—Chi-square 5.52 (df ¼ 1), P ¼ 0.015, Yates corrected P ¼ 0.0285. Allele 5 was about three times more frequent in BPAD females (9.6%) compared to healthy controls (2.7%). A similar trend was observed in the MDD females (7.9%), but could not reach statistical significance. The nominally significant results for this BDNF polymorphism are in line with previous association studies in BPAD. The functional significance of this extragenic polymorphism is unknown. Further studies with functional polymorphisms in BDNF are warranted to elucidate the potential impact of this gene to the etiology of the affective disorders. P1.16 META-ANALYSIS OF ASSOCIATION BETWEEN THE 48-BASE-PAIR VARIABLE-NUMBER TANDEM-REPEAT (VNTR) POLYMORPHISM IN EXON 3 OF DOPAMINE D4 RECEPTOR GENE AND MOOD DISORDERS Yeh Y,1 Faraone SV,1 Glatt SJ,4 and Tsuang MT1 1 Harvard Institute of Psychiatric Epidemiology and Genetics, Harvard Departments of Epidemiology and Psychiatry, Boston, United States 2 Department of Psychiatry, Harvard Medical School at the Massachusetts Mental Health Center, Boston, United States 3 Department of General Psychiatry, Bali Mental Hospital, Department of Health of Taiwan, Taipei County, Taiwan 4 Institute of Behavioral Genomics, Department of Psychiatry, University of California, San Diego, United States The dopamine D4 receptor gene was considered a candidate for association with mood disorders based on the pharmacological properties and the linkage study finding of an Old Order Amish pedigree. From all 9 case-control studies of bipolar disorder, 3 family-based studies of bipolar disorder, and 6 case-control studies of unipolar disorder, we calculated pooled estimates of the association between 48base-pair variable-number tandem-repeat (VNTR) polymorphism in exon 3 of dopamine D4 receptor gene and mood disorders separately. The meta-analysis sequentially considered the two-, three-, four-, and seven-repeat alleles as risk alleles, and then considered repeat length of the 48-base-pair segment as a risk factor. The pooled odds ratios from each analysis approximated 1.0 and were not statistically significant except for 2-repeat allele. The pooled odds ratio of 2-repeat allele in family-based studies of bipolar disorder was 0.5 with statistically significant (P ¼ .013) but not stable since there were only 3 studies and omission of single study would produced pooled ORs with 95% CIs that encompassed 1.0. The pooled odds ratio of 2-repeat allele with casecontrol studies of unipolar disorder was 1.5, which was marginally significant (P ¼ .04). We found more significant heterogeneity in the meta-analysis of unipolar disorder, which was at least partially explained by different diagnostic systems applied. Otherwise, the odds ratio derived from each study was unrelated to the ethnicity or gender composition of the sample, or the age of the control group. There was no evidence of publication bias. Our results provide limited evidence that DRD4 influence susceptibility to mood disorders, although more family-based studies and genomic-control studies are needed to elucidate this relationship. P1.17 NEURONAL TRYPTOPHAN HYDROXYLASE (TPH2) GENE AND SUICIDAL BEHAVIOUR Buresi C,1 Courtet P,2 Bellivier F,3 Jollant F,2 Leboyer M,3 and Malafosse A1 1 Hoˆpitaux Universitaires de Gene`ve, Gene`ve, Switzerland 2 Centre Hospitalier Universitaire, Montpellier, France 3 Centre Hospitalier Universitaire, Cre´teil, France A genetic contribution to suicidal behaviour has been demonstrated in family, twin and adoption studies. There is compelling evidence for a specific genetic component of susceptibility, which may be independent of or additive to the genetic transmission of the psychiatric disorders strongly related to suicidal behaviour (bipolar disorder, schizophrenia, alcoholism). Since deviations in the serotonin system are among the most robust findings in suicidal behaviour and in impulsive aggression, we hypothesised that DNA polymorphisms affecting serotonin-related

36

Abstracts

genes may be the molecular basis of the genetic vulnerability to suicidal behaviour (SB). We previously reported an association between a gene coding for tryptophan hydroxylase (TPH1) and SB. Two recent metaanalyses confirmed this association. However, recently a second isoform (called neuronal TPH or TPH2) has been discovered. This isoform has been shown to predominate in the rat and mouse central nervous system. The goal of the present study is to test the association between three TPH2 polymorphisms and SB. Seventy hundred and two patients consecutively admitted to our psychiatric departments after a suicide attempt and three hundred and twenty seven controls were included in this study. Suicide attempters and controls were administrated standard psychiatric interviews as well as self-report questionnaires. Three single nucleotide in introns 5, 7 and 9 respectively were genotyped. The variations in the TPH2 gene polymorphisms were not associated with SB. P1.18 POLYMORPHISMS IN THE SEROTONIN TRANSPORTER, MAOA AND COMT GENES AND SUSCEPTIBILITY TO UNIPOLAR DEPRESSION Grozeva DV,1 Raybould R,1 Dunn E,1 Cooper C,2 Swainson V,1 Lyon L,2 Jones L,2 Nikolov I,1 Green E,1 Jones I,1 Sham P,3 Farmer I,3 Craig I,3 McGuffin P,3 Kirov G,1 Owen MJ,1 and Craddock N1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Division of Neuroscience, University of Birmingham, Birmingham, United Kingdom 3 SDGP, Institute of Psychiatry, London, United Kingdom Genes involved in the dopaminergic and serotoninergic systems have been implicated in the aetiology of depression. The genes coding for the serotonin transporter (5-HTT), MAOA and COMT have been investigated in a number of studies, but the findings have been inconclusive. We conducted a case-control study for the most widely investigated polymorphisms in these genes: a VNTR in intron 2 (5-HTT VNTR) and a functional 44 bp deletion/insertion in the promoter region of 5-HTT (5HTT LPR), a 30 bp repeat in the promoter of MAOA (MAOA VNTR), 941 T to G substitution in MAOA gene, and a functional polymorphism in COMT (Val158Met). The sample set is a large newly collected series of Caucasian individuals of UK origin meeting criteria for DSMIV recurrent major depression (N ¼ 1023). The controls are UK Caucasians screened to exclude psychiatric illness and affective morbidity (N ¼ 745). For genotyping we used either ABI Prism 3100 genetic analyser (for VNTRs) or the AmplifluorTM SNP Genotyping System (for SNPs). We found no evidence for association between any of the polymorphisms studied and susceptibility to unipolar depression. The frequencies of alleles and genotypes were closely similar between patient and control groups. 5-HTT LPR: For the short allele the allele frequencies in the cases are 0.428 and in the controls- 0.425, for the long allele they are respectively 0.573 and 0.575. (P > 0.8). 5-HTT VNTR: For allele 9 the allele frequencies are 0.015 for the cases and 0.013 for the controls, for allele 10-0.383 for the cases and 0.378 for the controls, for allele 12- 0.602 for the cases and 0.609 for the controls. (P > 0.9). MAOA VNTR: For allele 5-0.002 and 0.001, for allele 6-0.332 and 0.309, for allele 7-0.021 and 0.021, for allele 8-0.632 and 0.658 and for allele 100.012 and 0.012 for the cases and controls respectively. (P > 0.6). MAOA 941T > G: For the T-variant the allele frequencies in the cases and in the controls are: 0.703 and 0.717 and for the G-variant they are respectively: 0.297 and 0.283 (P > 0.9) COMT Val158Met: For the Tvariant the allele frequencies are 0.484 and 0.481 and for G-variant they are respectively 0.516 and 0.519 for the cases and the controls. (P > 0.8). It is clear from our results that none of these polymorphisms, on its own, exerts a major influence on susceptibility to unipolar depression. We are currently exploring our data in more detail to test for possible gene-gene and gene-environment interactions and the possible influence of clinical covariates and subtypes. P1.19 VULNERABILITY TO POSTPARTUM DEPRESSION IN WOMEN WITH MAJOR DEPRESSIVE DISORDER: DO FAMILIAL FACTORS PLAY A ROLE? Forty L,1 Jones LA,2 Caesar ES,2 Dave S,2 Hough A,2 Dean L,2 Craddock M,2 Farmer A,3 McGuffin P,3 Craddock N,1 Jones I. R.1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Department of Psychiatry, Division of Neuroscience, University of Birmingham, Birmingham, United Kingdom 3 SGDP Research Centre, Institute of Psychiatry, London, United Kingdom

We have previously reported compelling evidence for the familial aggregation of episodes of puerperal (postpartum) psychosis in families multiply affected with bipolar disorder. We here describe a similar analysis of postpartum triggering in female—female sib-pairs with DSMIV recurrent major depressive disorder recruited as part of an international multi-centre unipolar sib-pair genetic study (the Depression Network (DeNT) study funded by GlaxoSmithKline Research and Development). Of the 120 afftected sib-pairs recruited by the Birmingham site, 61 were female—female and in 45 of these pairs both sisters had given birth and therefore had an opportunity to express the postpartum depression phenotype. On the basis of a modified SCAN interview and case note review, consensus ratings were made both of diagnosis and of a range of other clinical variables, including the occurrence of episodes of illness in relationship to childbirth. Employing a narrow definition of postnatal (postpartum) depression in which DSMIV criteria for a major depressive episode are met and with a clear onset of the episode within 4 weeks of delivery we found that 33 of the 45 pairs were concordant for puerperal status (kappa ¼ 0.44, P ¼ 0.003). Excluding women with episodes of perinatal illness but not meeting this narrow criterion gives 25 sibling pairs which show similar evidence for familial aggregation (kappa ¼ 0.53, P ¼ 0.003). Interestingly, if the criterion is widened to include all perinatal episodes in pregnancy or within 6 months of delivery then the familiality is considerably reduced (kappa ¼ 0.175, P ¼ 0.23). Our results suggest that familial factors play a role in determining vulnerability to narrowly defined postnatal (postpartum) depression but have either less or no effect in determining vulnerability to perinatal depressive episodes defined more widely. P1.20 INTERACTION BETWEEN LIFE EVENTS AND 5-HTT GENOTYPE IN DETERMINING THE LIKELIHOOD OF DEPRESSION AND ANXIETY IN A 25-YEAR LONGITUDINAL STUDY OF AUSTRALIAN TEACHERS Mitchell PB, Wilhelm K, and Parker G 1 University of New South Wales, Sydney, Australia 2 Garvan Institute for Medical Research, Sydney, Australia Caspi et al (2003) have reported that subjects with at least one copy of the short allele of the 5-HTT promoter polymorphism are more likely to exhibit depressive symptoms after exposure to life events in the prior five years. In this study we examine whether a similar relationship can be demonstrated in a 25-year longitudinally assessed cohort of Australian teachers. This sample was originally recruited in 1978 and has been extensively described by Wilhelm and Parker in a series of publications. The cohort originally comprised 170 individuals studying at the same teachers college in Sydney, and has been followed up with detailed structured interviews each subsequent 5 years. The most recent publications have focussed on the 20 year follow-up, for which 155 subjects were evaluable. Measures performed at each 5-year assessment have included a structured evaluation of life events, history of depressive and anxiety syndromes and subsyndromal presentations, neuroticism, childhood parental bonding, social support, treatments, substance abuse and medical history. One-hundred-and-twenty three subjects have now provided samples for genotyping. Results of analyses examining for the relationship between life events, 5-HTT promoter region polymorphisms, and likelihood of depressive and anxiety symptoms will be presented. Furthermore, the relationship of these variables to neuroticism will be examined.

P1.21 NEUROPSYCHIATRIC PHENOTYPE IN DARIER’S DISEASE Gordon-Smith K,1 Jones L A,2 Burge S,3 Munro C,4 Owen M,1 and Craddock N1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Department of Psychiatry, University of Birmingham, Birmingham, United Kingdom 3 Department of Dermatology, Churchill Hospital, Oxford, United Kingdom 4 Department of Dermatology, Southern General Hospital, Glasgow, United Kingdom Darier’s Disease is a rare autosomal dominantly inherited skin disorder. The disease maps to chromosome 12q23-q24.1 and has been shown to be caused by mutations in a single gene, ATP2A2. This gene encodes SERCA2, a sarcoplasmic/endoplasmic reticulum calcium

Abstracts pump which has a pivotal role in intracellular calcium signalling. Calcium is an important second messenger that controls many aspects of neuronal function therefore it is plausible that mutations in ATP2A2 could be involved in neuropsychiatric disorders. Clinical experience and reports of cases and families in the literature suggests that neuropsychiatric abnormalities, including affective disorders, suicidal ideation and epilepsy, occur at an increased rate in Darier’s Disease compared to other skin disorders of similar severity. We are currently conducting the first systematic study to investigate the neuropsychiatric phenotype in Darier’s Disease. A battery of neuropsychiatric tests is being administered to at least 100 unrelated individuals with Darier’s Disease and their family members. This battery contains measures assessing the presence of mood disorders and other psychiatric disorders, neurological abnormalities and features of epilepsy, personality traits and intellectual functioning. Second stage analysis will be to look for possible relationships between the type and locations of mutations in the ATP2A2 gene and neuropsychiatric abnormalities. Preliminary results, including a report of a new family where there appears to be co-segregation between affective disorders and Darier’s Disease, will be presented. The implications of understanding neuropsychiatric abnormalities that occur in Darier’s Disease will be discussed. P1.22 GENOTYPE-PHENOTYPE ASSOCIATIONS IN UNIPOLAR MAJOR DEPRESSION Mazei-Robison M, Hazelwood L, Myers R, Prasad H, Ritchie M, Sanders-Bush E, Shelton R, and Blakely R Vanderbilt University, Nashville, United States Unipolar major depression (MDD) appears to have genetic antecedents. However, specific genotype-phenotype relationships remain obscure. We have investigated the relationship between specific genotypes and clinical characteristics in MDD. We have established a mood disorders subject collection for the analysis of the associations between genetic variants and clinical characteristics and treatment response information in MDD. In the current study, we have ascertained in a group of 110 subjects (mean age 43, 60% female) with MDD using the Hamilton Rating Scale for Depression (HAM-D) and the NEO-Five Factor Inventory. Polymorphic variants of the serotonin transporter (SERT, 5HTTLPR), dopamine transporter (DAT), serotonin (5HT) 2A and 2C receptors were characterized. Binomial and multinomial logistic regression analysis, ANOVA, and Chi-square analyses were used to evaluate the relationships between genotypes and phenotypic expression of depression. The homozygous 5HTTLPR Ls þ ss genotypes were significantly associated with the HAM-D depressed mood item (a trend for association of hypersomnia-oversleeping, anxiety psychic, and anxiety somatic items were also found). The LL genotype associated significantly on the suicide item and at a trend level on the energy item. For the 5-HT2A 1438A/G, the AA þ AG genotypes associated with work/activities, agitation, and somatic (physical) anxiety, while the GG loaded on psychic anxiety (worry/rumination), and appetite change (increase and decrease). For the 5-HT2C Ser23Cys substitution, Cys was significantly associated with early insomnia (with a trend for middle insomnia), while Ser associated with somatic anxiety. The 9/10 DAT 30 UTR VNTR was associated with hypersomnia-daytime napping and agitation (with a trend towards NEO neuroticism), the 10/10 genotype was exhibited a trend to associate with NEO openness. Specific genotypes appear to be associated with selected clinical features of depression.

P1.23 POLYMORPHISM AND LINKAGE DISEQUILIBRIUM PATTERNS AT THE SEROTONIN TRANSPORTER GENE LOCUS (SLC6A4) Mansour HA, Talkowski ME, Chowdari KV, Wood JA, Bamne MN, Fagiolini A, Frank E, Thase ME, Kupfer DJ, and Nimgaonkar VL Departments of Psychiatry and Human Genetics, University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, United States During the course of association studies involving mood disorders and SLC6A4 (5HTT, ch 17q11.1-q12), we have analyzed 11 polymorphisms and their linkage disequilibrium patterns. Our sample consists of 135 US Caucasian nuclear families having a proband with Bipolar I Disorder (BD1), available parents and 73 patients with Major Depressive Disorder (MDD, DSM IV criteria). Cord blood samples from local

37

Caucasian live births serve as unrelated, unscreened controls (n ¼ 182). An insertion/deletion polymorphism (5-HTTLPR) in the promoter region, as well as ten SNPs downstream to 5-HTTLPR are being investigated. Variation in 5-HTTLPR is being analyzed by re-sequencing, as well as a PCR based assay (Heils et al., 1996). Genotype assays for the SNPs are based on multiplex PCR followed by single base extension analysis (SnaPshot assays, ABI Inc). We have evaluated pair-wise linkage disequilibrium (LD) patterns between the SNPs using our BD1 case sample. Significant LD was observed among the first 6 SNPs, thus comprising one ‘‘haplotype block’’ (rs140701, rs2020942, rs6354, rs2020939, rs2020937, rs2066713). Significant LD was also noted among the remaining SNPs (rs2020935, rs2020934, rs2020933 and rs2020930). We are evaluating LD patterns further using family-based analyses. We are also re-sequencing a panel of cases at 5-HTTLPR. We thus hope to identify additional polymorphisms in the insertion, as well as flanking sequences. Once these variations are resolved, we will evaluate complete LD across the region using unrelated cases and family-based analyses. We will also present our association analyses involving the BD1 patients. These analyses may help resolve some of the persistent controversies involving the etiological role of SLC6A4. P1.24 GLICOGEN SYNTHASE KINASE (GSK) GENE PROMOTER POLYMORPHISMS: A TDT AND HAPLOTYPE APPROACH IN MOOD DISORDERS Tubazio V, Benedetti F, Serretti A, Plioa C, Pirovano A, Smeraldi E, Milan, Italy The molecular mechanisms driving the biological clock in the suprachiasmatic nucleus of the hypothalamus may play a role in mood disorders. Two single nucleotide polymorphisms (SNPs) (1727 A/T and 50 T/C) located into the promoter region and the promoter sequence itself of the for glycogen synthase kinase-3b coding gene respectively, have been identified. GSK3b is a central enzyme in many intracellular signaling systems. The enzyme is directly regulated by lithium, that is the primary therapeutic agent for bipolar mood disorders, and it is also known to regulate circadian rhythms in Drosophila. We investigated the possible association between these two polymorphisms variants and the presence of affective disorders. We recruited 237 trios diagnosed for mood disorders and investigated the association between the two polymorphisms of GSK3b and affected status by a family-based association study approach (Maximum Likelihhod of Haplotype Frequencies—HAPMAX and transmission disequilibrium test TDT). There were no differences in the transmission of alleles both for glycogen synthase kinase-3b—50 T/C (chi square ¼ 0.06, P ¼ 0.19) and glycogen synthase kinase-3b—1727 A/T (chi square ¼ 0.56, P ¼ 0.54), Haplotype maximum likelihood frequencies analysis did not lead to significant results (P ¼ 0.99). Neither the two genotypes nor haplotypes resulted significantly associated with mood disorder affection status. A limitation of the study could be considered the relative little presence of informative trios in our sample and the relative small sample size.

P1.25 THE BRAIN DERIVED NEUROTROPHIC FACTOR (BDNF) VAL66MET POLYMORPHISM AND RECURRENT UNIPOLAR DEPRESSION Cohen S,1 Rosa A,1 Corsico A,1 Sterne A,1 Owen M,2 Korzsun A,2 Craddock N,2 Craig I,1 Farmer A,1 McGuffin P1 1 Institute of Psychiatry, London, United Kingdom 2 Cardiff University, Cardiff, United Kingdom 3 University of Birmingham, Birmingham, United Kingdom A substantial genetic contribution to the liability to depression has been demonstrated, with heritabilities ranging from around 30% in community twin samples to 70% in clinically ascertained twins. However specific genes have yet to be consistently identified. Brain Derived Neurotrophic Factor (BDNF) has been implicated in the pathogenesis of major depressive disorder in preclinical and clinical studies and has lead to the neuroptrophic hypothesis of depression. The BDNF-gene Val66Met polymorphism has been associated with bipolar disorder. Furthermore an association has been reported with neuroticism, a risk factor for depression. We carried out a case control association study to determine if the Val66Met coding variant contributes to liability to depression. The sample, collected at 3 UK

38

Abstracts

Centres, consists of 602 cases (413 women, 189 men) who have had 2 or more episodes of ICD10/DSMIV depression of at least moderate severity assessed by SCAN interview and 718 healthy controls (396 women, 322 men) who were screened by questionnaire and telephone interview for absence of psychiatric disorder. No significant differences were found in the genotype or allele frequencies between the depressed and the control subjects either in the sample as a whole or when it was divided by sex. These findings suggest that the BDNF-gene Val66Met polymorphism does not contribute to the pathogenesis of recurrent depression. The depression case-control collection (DeCC) was supported by a grant from the Medical Research Council (UK) and SC is in receipt of an MRC studentship.

P1.26 AGE OF ONSET IN DEPRESSION: EVIDENCE FOR TWO SUBGROUPS SHOWING FAMILIALITY Tandon K,1 Aitchison KJ,2 Craddock N,3 Ferrero F,7 Gill M,4 Jones IR,1 Jones LA,1 Maier W,5 Mors O,8 Owen MJ,3 Preisig M10, Reich T,9 Rietschel M,6 Farmer A,1 and McGuffin P1 1 SGDP Centre, Institute of Psychiatry, London, United Kingdom 2 SGDP Centre and Psychological Medicine, Institute of Psychiatry, London, United Kingdom 3 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 4 Trinity College, Dublin, Ireland 5 Department of Psychiatry, University of Bonn, Bonn, Germany 6 Division of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, Mannheim, Germany 7 Department of Psychiatry, University of Geneva, Geneva, Switzerland 8 Department of Psychiatric Demography, Psychiatric Hospital in Aarhus, Aarhus, Denmark 9 Department of Psychiatry, Washington School of Medicine, St Louis, United States 10 Departement Universitaire de Psychiatrie Adult, Prilly-Lausanne, Switzerland, 11 Department of Psychiatry, University of Birmingham, Birmingham, United Kingdom Early onset depression has been used to define a subtype of depression with increased heritability (Holmans P et al., 2004, Am J Hum Genet, 74, 1154–1167), although the evidence for this has been contradictory (McGuffin P et al., 1996, Arch Gen Psych, 53, 129–136). The aim of the study was to use admixture analysis to more fully examine the age of onset in depression using a cohort of patients with recurrent DSM-IV depression, and to examine the familiality of any subgroups delineated using a separate sibling pair sample. The distribution in age of onset of depression in subjects (N ¼ 775 women, N ¼ 366 men) interviewed using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) and diagnosed with recurrent DSM-IV depression was examined using admixture analysis, which was performed using the program SKUMIX (Mclean CJ et al., 1976, Biometrics, 32, 695-9, Morton NE 1982, Outline of Genetic Epidemiology. Basel: Karger). Two separate distributions compared to one distribution best described the ages of onset in both the men and women (P < 0.05, w2 ¼ 6.3, df ¼ 2 in women, and P < 0.001, w2 ¼ 18.3, df ¼ 2 in men). The mean age of onset for the first distribution in women was 21.9 years and in men was 26.5 years. The mean age of onset for the second distribution in women was 36.8 years in women and 47.9 years in men. Two standard deviations from the mean of the first distribution in men and women was used to define the ages of onset in subgroup 1. Subgroup 2 was defined as ages of onset over and above this. The familiality of age of onset of depression was examined using 356 sibling pairs with a history of DSM-IV recurrent depression interviewed using the SCAN. There was a significant correlation between the ages of onset of depression in the proband and in the sibling (Pearson’s correlation coefficient ¼ 0.297, 2-tailed P < 0.0001). The familiality of the two subgroups of depression, as defined by the admixture analysis, was also examined using the sibling pair sample. The distribution of probands and siblings in the two age of onset subgroups was tabulated in a 2 by 2 table, and the chi square test used to determine dependency between the proband and sibling subgroups. This gave a w2 ¼ 11.49 (df ¼ 1, P ¼ 0.007), suggesting that the two subgroups of age of onset of depression showed familiality. This is the first study to use admixture analysis to explore the ages of onset in depression to determine if there is more than one distribution in the age of onset present. The results indicate the presence of two distributions in the ages of onset in depression, and that these two

distributions show familiality in their aetiology. The genetic contribution towards this familiality needs now to be explored.

P1.27 BRAIN DERIVED NEUROTROPHIC FACTOR (BDNF) VAL66MET POLYMORPHISM IN MAJOR DEPRESSION. RELATION TO CLINICAL VARIABLES, RESPONSE TO TREATMENT AND CORTISOL LEVELS Bondy B, Baghai TC, Schulle C, Eser D, Zill P, and Rupprech R Psychiatric Clinic, Munich, Germany Several lines of evidence suggest that brain-derived neurotrophic factor (BDNF) is not only involved in mood disorders but also a downstream target of a variety of antidepressant treatment. BDNF might provide against stress induced neuronal damage. Despite several clinical studies demonstrating that serum BDNF levels are significantly decreased in patients with major depression, recent studies have reported associations of the Va66Met polymorphism in the coding region of the BDNF gene with cognitive functions, personality traits, bipolar and eating disorders. With respect to the increasing importance of neuroplasticity mechanisms in mood disorders we have investigated the BDNF Val196Met polymorphism in patients with major depression. We included 331 hospitalized patients with major depression (125 males, 206 females, age 49,2  13,7 years) and related the genotypes and alleles to clinical variables as well as to the results of the Dexamethasone-CRH Test. The majority of the patients were on a combined drug regimens, only 50 were on monotherapy and 35 patients on ECT (without additional drugs). Severity of depressive symptoms was assessed before treatment, after 4 weeks and at discharge using the HAMD rating scale. The genotype or allele frequencies did not differ from those of 335 controls. The HAMD rating scale revealed a 25% reduction in Met/Met homozygotes within the first week of treatment (only 10% in Val/Val and Val/Met genotypes). Later on no differences between the genotypes could be observed. Concerning the data of the combined Dex-CRH test, we found no relation between basal cortisol levels and the genotypes. However, after stimulation with CRH the cortisol levels increased more pronounced in Met/Met genotypes (P ¼ 0,002 at peak, AUC: P ¼ 0.012) as compared to Val/Val homozygotes and heterozygotes. In conclusion, our results have shown that there is no direct relationship between the BDNF Val66Met polymorphism and severity of depression or the response to treatment. However, our results could indicate that the homozygous 66Met genotype might be a subtype of major depression, which is more susceptible to stress adaptation mechanisms. Other modulatory systems, including receptors or neurotransmitters might enhance or decrease this effect. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine.’’

P1.28 ASSOCIATION OF SNP WITHIN THE SEROTONIN TRANSPORTER GENE WITH DEPRESSIVE DISORDER AND ASSOCIATED FUNCTIONAL PARAMETERS Lucae S, Binder EB, Pu¨tz B, Salyakina D, Mueller-Myhsok B, Ising M, Horstmann S, Unschuld PG, and Holsboer F Max Planck Institute of Psychiatry, Munich, Germany The serotonin transporter gene promotor polymorphism (5-HTTLPR) has been described to be associated with response to treatment with SSRIs in patients with affective disorders. In this study we analyzed 7 SNPs throughout the SLC6A4 gene in 253 patients suffering from depressive disorders. We investigated gender differences, case/control associations, associations with response to antidepressant treatment, associations with neuropsychological items, as well as with neuroendocrinological data. We found no case/control association in the overall sample. However, when male patients with recurrent unipolar disorder were analyzed seperately against male controls, we found a small effect in 4 SNPs (P ¼ 0.01). Furthermore, we found an association regarding divided attention at admission and discharge in some SNPs independent of gender differences. We found no association with response independent of the class of antidepressant treatment. 96 patients were treated with SSRI, 93 with mirtazapine and 64 patients were treated with tricyclic antidepressants. We could not find any significant effects regarding neuroendocrinological data emerging from combined dexamethasone-CRH tests. We are currently on the way to investigate more SNPs, especially in the promotor region. We are going to

Abstracts determine 5-HTTLPR genotypes as well as single nucleotide polymorphisms (SNPs) adjacent to the 5-HTTLPR region. P1.29 PATTERNS OF COMORBIDITY OF MIGRAINE AND MENTAL DISORDERS Low NCP,1 Angst J,2 Cui L,1 and Merikangas KR1 1 National Institute of Mental Health, Bethesda, United States 2 University of Zurich, Zurich, Switzerland Migraine is a complex disorder characterized by phenotypic heterogeneity, complex patterns of inheritance and a high magnitude of comorbidity with other disorders. This poster will present data from three clinical and population-based studies: (1) a family study of the coaggregation of migraine and mental disorders, (2) a prospective study of youth at high risk for the development of psychopathology, and (3) a twenty-year prospective community of young adults, and (4) a large cross-sectional study of a clinical sample of bipolar subjects. Each study examines the pattern of comorbidity between migraine and mental disorders. Results reveal a strong association between migraine and mood disorders, particularly with the atypical and bipolar subtypes of depression. We found that the clinical course of bipolar disorder differed among those with and without migraine. Corroboration of findings from clinical samples in community studies suggests that the association is not attributable to treatment-seeking bias. Inspection of prospective data reveals that the onset of anxiety disorders predates the onset of migraine in children followed by the onset of mood disorders in late adolescence and adulthood. Patterns of expression of migraine and mood disorders in families are consistent with the explanation that migraine may be an index of a discrete subtype of bipolar spectrum disorder. P1.30 ASSOCIATION OF SUICIDAL BEHAVIOR WITH THE TPH2 GENE Dietrich I, Giegling I, Hartmann AM, Mo¨ller H-J, and Rujescu D Division of Molecular and Clinical Neurobiology, Munich, Germany Risk of suicide-related behavior is supposed to be determined by a complex interplay of sociocultural factors, traumatic life experiences, psychiatric history, personality traits, and genetic vulnerability. This view is supported by adoption and family studies indicating that suicidal acts have a genetic contribution that is independent of the heritability of Axis I and II psychopathology. The heritability for serious suicide attempts was estimated to be 55%. Neurobiological studies have shown that serotonergic dysfunction is implicated in suicidal behaviors. These findings stimulated the investigation of variations in serotonergic genes in this context. Tryptophan hydroxylase (TPH) is the rate-limiting enzyme of serotonin biosynthesis and regulates serotonin levels. Recently, a second TPH isoform (called TPH2) encoded by an additional gene on human chromosome 12 has been discovered. We and others have shown that isoform predominates in the central nervous system. In this study we examined three single nucleotide polymorphisms (SNPs, one in the promoter region and two in introns 2 and 5, respectively) in the TPH2 gene with regard to suicidal behavior and to anger as an intermediate suicide-related phenotype. A sample of 250 suicide attempters and 500 healthy volunteers were examined. Suicide attempters and controls were administered standard psychiatric interviews as well as self-report questionnaires. These variations in TPH2 were not associated with suicidal behavior per se. Interestingly, anger-related traits were associated with variations in this gene. The implication of these findings will be discussed.

P1.31 INTERACTION OF CHILDHOOD TRAUMA AND TPH1 GENE IN SUICIDAL BEHAVIOUR Courtet P,1 Bellivier F,3 Baud P,2 Buresi C,2 Jollant F,1 Leboyer M,3 and Malafosse A2 1 Centre Hospitalier Universitaire, Montpellier, France 2 Hoˆpitaux Universitaires de Gene`ve, Gene`ve, Switzerland 3 Centre Hospitalier Universitaire, Cre´teil, France A stress-diathesis model has been proposed for suicidal behaviour (SB). This predict that individuals’ sensitivity to stressful events depends on their genetic makeup. Both non-genetic and genetic familial factors have been shown to play a role in the susceptibility to SB. Adults

39

reporting childhood physical and sexual abuse experiences have a higher rate of SB and greater impulsivity. However, not all people who encounter such experiences succumb to their suicidogenic effect. A genetic component, suggested by family, twin and adoption studies, might explain this variability. Recent meta-analyses yielded statistically significant replication of first reports of association between the tryptophan hydroxylase gene (TPH1) and suicide behaviour (SB). The goal of the present study is to determine whether intron 7 A218C TPH1 polymorphism moderate the influence of childhood trauma on SB. Using the Childhood Trauma Questionnaire we evaluated patients with (N ¼ 403) and without (N ¼ 57) a history of SB. Diagnostic assessment was done according to DSM IV criteria with the French version of the Mini International Neuropsychiatric Interview (MINI 5.0.0). SB was characterized clinically (violent/nonviolent, number of attempts) and using the French versions of the Risk Rescue Rating Scale (RRRS), and of the Suicidal Intent Scale (SIS). We tested intercorrelations, fitted a common factor model to this different measures of SB and created a composite index of SB. Preliminary results suggest that subjects carrying the at risk TPH1 genotype AA, are more sensitive to childhood trauma in the intermediary range (CTQ scores between 34 and 54). Genotypes have no effect on the impact of severe childhood trauma. Interaction study between childhood trauma and other serotonin- related genes will be done. P1.32 DECISION-MAKING AND ITS GENETIC MODULATION IN SUCIDIAL BEHAVIOUR Jollant F,1 Buresi C,2 Baud P,2 Bellivier F,3 Leboyer M,3 Malafosse A,2 and Courtet P1 1 Centre Hospitalier Universitaire, Montpellier, France 2 Hoˆpitaux Universitaires de Gene`ve, Gene`ve, Switzerland 3 Centre Hospitalier Universitaire, Cre´teil, France The aggregation of suicide and suicide attempts in the families of suicide victims is well known and has been demonstrated in the relatives of both adolescent and adult suicides. Several recent metaanalyses yielded statistically significant replication of first reports of association between the tryptophan hydroxylase gene (TPH1) and the serotonin transporter gene (5-HTT) and suicide behaviour (SB). These results suggest TPH1 and 5-HTT variants have real but modest effects on SB. These variants may have a more obvious impact at the level of intermediary traits. Such a trait may be decision-making (DM), a cognitive function related to choices in complex and uncertain situations, that provides a neuroanatomical and cognitive framework for SB. Our goal is to measure the DM in patients with a history of suicidal behaviour using the Gambling Task and to assess the influence of TPH1and 5-HTT variants. We evaluated patients with a history of violent (N ¼ 45) or non violent (N ¼ 115) SB healthy controls (N ¼ 82) and controls with a history of affective disorders (N ¼ 25). Intron 7 A218C TPH1 and promoter S/L 5-HTT polymorphisms were tested. Both groups of suicide attempters scored significantly lower than healthy controls, and violent suicide attempters performed significantly worse than affective controls (P < 105). No significant differences were observed between the groups of suicide attempters or between the two control groups. The differences in performances could not be accounted for by age, premorbid level, educational level, number of suicide attempts, age at first suicide attempt, history of Axis-I disorder or medication use. While subjects with AC or CC TPH1 genotypes and those with SL or LL genotypes improve their scores during the course of the test, individuals with the two at risk genotypes to SB, i.e. AA TPH1 and SS 5-HTT, do not. These results suggest that DM could be an intermediary phenotype in genetic studies of SB. P2.1 A48G POLYMORPHISM IN THE D1 RECEPTOR GENES ASSOCIATED WITH BIPOLAR I DISORDER IN A SARDINIAN POPULATION Severino G, Congiu D, Serreli C, De Lisa R, Chillotti C, Piccardi MP, and Del Zompo M Section of Clinical Pharmacology, Department of Neurosciences, University of Cagliari, Cagliari, Italy Several lines of evidence point to a role of dopamine in mood disorders, in particular in bipolar disorders. With regard to DA receptors, a considerable body of evidence suggests that many of the acute behavioural effects depending on the activity of D2-class dopamine receptors, are known to require simultaneous activation of D1-class receptors. Thus, the dopamine D1 receptor gene (DRD1) is considered a

40

Abstracts

good candidate gene for bipolar disorders. Several studies did not find any association between bipolar I patients and DRD1. In this study we investigate a possible association between BP disorder and 48A/G polymorphism of the DRD1. We genotyped 107 bipolar I patients and 129 healthy control subjects of exclusively Sardinian descent. A statistically significant difference in genotype (w2 ¼ 6.29 df ¼ 2 P ¼ 0.042) and allele (w2 ¼ 5.46 df ¼ 1 P ¼ 0.019, OR 1.53, 95% IC 1.08–2.16) frequencies was found, suggesting an association between the DRD1 gene and bipolar I disorder in the Sardinian population. In conclusion, in spite of the small size of the sample examined and the need of replication in other types of population, our study suggests that the D1 receptor gene may play a role in the etiology of BP I disorder. P2.2 IS AGE AT ONSET A USEFUL MARKER FOR SUB-TYPING BIPOLAR DISORDER? Coelho I,1 Dourado A,1 Valente J,1 Macedo A,1 Soares M,1 Pato M,2 Pato C,2 Palha J,3 and Azevedo MH1 1 Faculdade de Medicina de Coimbra, Coimbra, Portugal 2 Center for Psychiatric and Molecular Genetics, SUNY Upstate Medical University, Syracuse, NY, United States 3 Universidade do Minho, Braga, Portugal To compare two groups of bipolar patients defined by age at onset of illness. Several clinical aspects were analysed including psychotic features, course, medical/psychiatric comorbidity and female reproductive events. The sample comprises 150 Portuguese bipolar patients (92 females), aged 16–89 years (mean ¼ 44.5, SD ¼ 15.9), 48.7% were married, 32% single and 18.7% were divorced/separated/widowed. The data for this study were drawn from ongoing research projects on Genetic Analysis of Psychosis. Most of the patients (n ¼ 118, 78.7%) had a family history of schizophrenia and/or affective disorder. All were interviewed using the Portuguese version of the Diagnostic Interview for Genetic Studies (DIGS) and diagnoses were based on the OPCRIT Polydiagnostic System (selected according DSM-III-R and/or CID-10 Diagnostic Criteria definitions). The age at onset was defined according to the OPCRIT definition-the earliest age at which medical advice was sought for psychiatric reasons or at which symptoms began to cause subjective distress or impair functioning—(mean ¼ 25.05, SD ¼ 9.25, range ¼ 12–59). Comparing the two groups of patients (age onset G and the promoter VNTR polymorphisms of MAOA, whereas other studies have failed to replicate these results. For COMT, the functional Val/Met polymorphism has been reported to influence susceptibility to rapid cycling in bipolar disorder and Shifman et al. (2002) recently reported a significant association between three COMT SNPs (rs737865, rs165599 and rs4680) and their haplotypes and schizophrenia. We have carried out association studies using these polymorphisms in a large, well-powered, clinically and genetically homogeneous sample including approximately 900 UK Caucasian DSMIV bipolar 1 patients and 680 UK caucasian controls group-matched for age and gender. We found no evidence for significant main effects for any of the polymorphisms (or for the Shifman haplotype with COMT) and are currently undertaking interaction and sub-phenotype analyses to seek evidence for more complex contributions of these genes to susceptibility and modification of the course of Bipolar Disorder. P2.7 ASSOCIATION ANALYSIS OF POLYMORPHISMS WITHIN GCHI WITH BIPOLAR DISORDER IN IRISH FAMILIES Kealey C,1 Roche S,1 Cassidy F,1 and McKeon P2 1 Trinity College, Dublin, Ireland 2 St. Patrick’s Hospital, Dublin, Ireland To identify genes contributing susceptibility to bipolar disorder (BPD) in the Irish population, a modified genome scan was performed. Microsatellite markers concentrated in gene-rich regions of the genome were chosen for analysis on a collection of sib-pairs affected by BPD (Murphy et al., 2000). Interesting results were obtained on chromosomes 5, 14 and 17, with location 14q22-24 showing the strongest

41

linkage to BPD in a follow-up analysis. Multiple adjacent markers in this region showed suggestive linkage to BPD type I indicating excess allele sharing between sibs in this region. The most significant linkage score occurred at marker D14S281 (NPL ¼ 2.7, P ¼ 0.003). Other groups have also found interesting results in this region (McInnis et al., 2003; Liu et al., 2003; Cichon et al., 2001). The linkage peak at 14q22-24 is quite large and spans over 35 cM (30 Mb, >185 genes). Interestingly, a gene involved in neurotransmitter synthesis, GTP Cyclohydrolase I (GCHI), is located approximately 200 Kb 30 of D14S281. GCHI catalyzes the first, rate-limiting step in the conversion of GTP to tetrahydrobiopterin (BH4). BH4 is a cofactor for amino acid hydroxylases involved in neurotransmitter synthesis, such as tyrosine (dopamine) and tryptophan hydroxylases (serotonin). Numerous studies suggest that the clinical symptoms of depression, and possibly BPD, may be related to a deficiency of BH4. Furthermore, mutations in the GCHI gene have been identified in a large Texan family suffering from psychiatric disorders such as anxiety and depression (Hahn et al., 2001). We previously observed that a novel polymorphism within the promoter region of GCHI was linked and associated with BPD in Irish families (X2 ¼ 5.26, P ¼ 0.0218) suggesting that variants within GCHI may contribute susceptibility to BPD in this population. We are extending our association analysis to include additional polymorphisms within GCHI and will present data on linkage disequilibrium patterns within this gene with a view to localising the disease-related variant. P2.8 ASSOCIATION OF HSPA5 WITH BIPOLAR DISORDER IN JAPANESE POPULATION Kakiuchi C,1 Nanko S,2 Kunugi H,3 and Kato T1 1 Brain Science Institute, RIKEN, Wako, Japan 2 Teikyo University School of Medicine, Itabashi, Japan 3 National Institute of Neuroscience, Kodaira, Japan We previously reported the pathophysiological role of endoplasmic reticulum (ER) stress response signaling in bipolar disorder based on multiple evidences. Comparing gene expression in the lymphoblastoid cells of monozygotic twins discordant for bipolar disorder, ER stress response related genes, XBP1 and HSPA5 (GRP78/BiP) were downregulated in the affected co-twin. Induction of XBP1 and HSPA5 mRNA by thapsigargin was reduced in the patients’ cell lines, and functional polymorphism of XBP1 (116C/G) was associated with bipolar disorder. Valproate induced ATF6 mRNA expression and enhanced the ER stress response. To further examine the pathophysiological role of ER stress response signaling in bipolar disorder, we investigated whether or not the HSPA5 is associated with this illness. By screening all exons and the upstream region (1 kb) of HSPA5 in 24 Japanese patients with bipolar disorder, entire HSPA5 gene was found to be in one haplotype block. By genotyping 3 key polymorphisms (hcv1754746 (C/T, Celera ID), rs3216733 (G/del), and rs12009 (C/A)) in Japanese case control samples (194 patients and 254 controls), C-del-T haplotype among 4 haplotypes (C-del-C, C-del-T, C-G-C and T-del-T) was significantly associated with bipolar disorder (frequency in control, 0.076 and in the patients, 0.128, P ¼ 0.011. P-values were calculated by COCAPHASE), which was stronger in patients with familial history (frequency in the patients, 0.201. P ¼ 0.000085). The C-del-T haplotype was extremely rare in the 88 NIMH trio samples, and no association was found. Our result supported that ER stress response signaling contribute to the pathophysiology of bipolar disorder. Recently, antimalarial drug mefloquine, which often causes an episode of depression or mania in susceptible individuals, was reported to cause ER stress in the brain. Further investigations of ER stress response signaling system in the pathophysiology of bipolar disorder is warranted. P2.9 MITOCHONDRIAL CALCIUM LEVELS IN CYBRIDS OF CONTAINING MITOCHONDRIAL DNA DERIVED FROM PATIENTS WITH BIPOLAR DISORDER AND HEALTHY CONTROLS Kazuno A,1 Munakata K,1 Nagai T,2 Tanaka M,3 Yoneda M,4 Kato N,5 Miyawaki A,2 and Kato T1 1 Lab for Molecular Dynamics of Mental Disorders, BSI, RIKEN, Wako, Japan 2 Lab for Cell Function and Dynamics, BSI, RIKEN, Wako, Japan 3 Department of Gene Therapy, Gifu International Inst. of Biotechnol., Kagamihara, Japan 4 2nd Department of Int. Med., Fukui University, Fukui, Japan 5 Department of Neuropsychiatry, University of Tokyo, Bunkyo, Japan

42

Abstracts

In many complex diseases, pathophysiological roles of mitochondrial DNA (mtDNA) have been recently suggested. We previously examined the energy metabolism of the brains of patients with bipolar disorder using phosphorus-31 magnetic resonance spectroscopy and demonstrated the possibility of mitochondrial dysfunction in the brain. In addition, we found that the mitochondrial DNA (mtDNA) 10398A polymorphism was significantly associated with bipolar disorder in Japanese. It is generally accepted that bipolar disorder is accompanied by altered intracellular calcium signaling. The intracellular calcium signaling is regulated by the calcium channels, endoplasmic reticulum, and other signal transduction pathways. In addition, recent cell biological studies have revealed an important role of mitochondria in intracellular calcium signaling. We therefore proposed that abnormal calcium signaling in bipolar disorder might be due to mitochondrial dysfunction. We started this study to test our working hypothesis that bipolar disorder is caused in part by the alteration of calcium signaling linked with the mtDNA polymorphisms. We generated Rho zero cell line, lacking mtDNA, stably transfected with Pericam, a fluorescent calcium indicator protein, which was targeted to mitochondria and nuclei. And we fused the stable cell line and platelets derived from bipolar patients and healthy controls. These hybrid cells (cybrids) have a nuclear DNA of common origin and mtDNA derived from platelet of bipolar patients and controls. We measured intracellular calcium responses evoked by 10 uM histamine in these cybrids. Mitochondrial and cytosolic calcium levels reached the peak level within several seconds after the stimulation and returned to the previous level within one minute. Basal mitochondrial calcium levels of 7 cybrids were extremely different from those of the other 28cybrids. To identify mtDNA polymorphisms affecting mitochondrial calcium regulation, we analyzed entire mtDNA sequence in all 35 cybrids. We identified several mtDNA polymorphisms altering mitochondrial calcium levels.

P2.10 IDENTIFICATION OF A LOCUS INCREASING GENETIC SUSCEPTIBILITY TO BIPOLAR AFFECTIVE DISORDER ON CHROMOSOME 12Q24.3 IN BOTH BRITISH/IRISH AND DANISH CASE-CONTROL ASSOCIATION SAMPLES Kalsi G,1 McQuillan A,1 Degn B,2 Dybro Lundforf M,2 Lawrence J,1 Curtis D,3 Kruse T,2 Gurling H,1 and Mors O2 1 Royal Free and University College Medical School, London, United Kingdom 2 Psykriatrisk Hospital Aarhus, Aarhus, Denmark 3 Royal London School of Dentistry & Medicine, London, United Kingdom Several genetic linkage studies of multiply affected bipolar families in Canada, Denmark, the UK and Iceland (Curtis et al., 2003, Psychiatr Genet 13:77-84) have found evidence for linkage to a bipolar affective disorder susceptibility locus on chromosome 12q24.3. In order to identify which mutations or genetic variants are responsible for this genetic subtype of bipolar disorder we have carried out linkage disequilibrium mapping in a genetically and ancestrally matched sample of bipolar cases and matched supernormal controls. 381 UK/ Irish and Danish bipolar patients and 434 UK/Irish and Danish supernormal controls were genotyped with a total of 21 new and previously identified polymorphic microsatellite markers and several single nucleotide polymorphisms. Statistically significant differences in allele frequencies were found for eight closely linked microsatellite markers spanning a region of about 300 kb. The pattern of allelic association between bipolar disorder and specific markers overlapped in the Danish and UK samples when analysed separately and when pooled. The most significant result in the UK sample was obtained with the novel marker, D12SDK1, which gave an empirically derived Monte Carlo significance of P ¼ 0.0004 using CLUMP. The most positive result for the Danish sample was with D12S1634, which was significant at P ¼ 0.016. The most significant result from the combined samples was also obtained with D12SDK1, P ¼ 0.003. When these eight microsatellite markers were combined into three and four marker haplotypes together with eight SNPs empirically derived significances for haplotype association ranged from P ¼ 0.04 to P ¼ 0.00024 thus confirming the positive fine mapping of a bipolar susceptibility locus by the single locus allelic associations. There are four putative genes in the implicated region and many other expressed sequences. One of the potential candidate genes with human brain expression shows extensive alternative splicing but has no known function. It is known to be homologous to a gene expressed in the macaque brain. Most of the

expressed sequences in the candidate region have been sequenced in a subset of about 30 bipolar cases who have inherited susceptibility haplotypes or alleles. We identified several sequence variants in the genes and expressed sequences in the candidate regions but none had an obvious role in the aetiology of bipolar disorder. Funded by the Neuroscience Research Charitable Trust and by Lundbeckfonden . We wish to acknowledge the contribution of Professor Henrik Ewald, who died on 20th Jan 2004, to this research. P2.11 BIPOLAR DISORDER AND POLYMORPHISMS IN THE DYSTOBREVIN BINDING PROTEIN GENE Raybould RE,1 Green EK,1 Macgregor S,2 Gordon-Smith K,3 Hyde S,3 Heron J,3 Middle FA,3 Lendon C,3 Jones L,3 Nikolov I,2 O’Donovan M,1 Owen MJ,1 Jones I,1 Kirov G,1 and Craddock N1 1 Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Biostatistics and Bioinformatics Unit, University of Wales College of Medicine, Cardiff, United Kingdom 3 Molecular Psychiatry Group, University of Birmingham, Birmingham, United Kingdom The dysbindin gene (DTNBP1) located on 6p22.3 has been implicated by several studies as influencing susceptibility to schizophrenia. It is expressed ubiquitously and the product is a component of the dystrophin protein complex, which plays a structural role in synapse formation and maintenance. Our group recently reported that variation at a haplotype defined by 3 ploymorphisms, P1635 (rs3213207), P1655 (rs2619539) and SNP A (rs2619538), influence susceptibility to schizophrenia in a large UK caucasian case-control sample. We have investigated this same 3 locus haplotype in a large, well-characterized Bipolar sample ascertained from the same population and recruited and assessed using similar methods to those in our schizophrenia study. Our sample included 726 Caucasian DSMIV bipolar 1 patients and 1407 controls group-matched for age, sex and ethnicity. Comparing the full Bipolar I sample with controls, no significant differences were found in allele or genotype frequencies for any of the polymorphisms individually, nor for any of the 2- or 3-locus haplotypes. However, when considering Bipolar I cases with psychosis against controls we did find nominally significant support for association at one of the polymorphisms, SNP A (rs2619538), individually, with a pattern of findings similar to that previously observed in our schizophrenia sample. We found no evidence for association with any other sub-phenotypes examined. Our data do not suggest that variation at this locus makes a major contribution to influencing susceptibility to Bipolar Disorder in general but suggest the possibility that DTNBP1 may influence susceptibility to psychotic Bipolar Disorder in particular. This hypothesis requires testing in independent samples—which are likely to need to be large to provide sufficient power to replicate. If confirmed, this will advance our understanding of the pathophysiology of Bipolar Disorder and contribute to dissection of the interface between Bipolar illness and schizophrenia. P2.12 IDENTIFICATION OF BIPOLAR RISK VARIANTS IN THE GENE ENCODING THE WINGED-HELIX TRANSCRIPTION FACTOR RFX4 ON CHROMOSOME 12Q23 Glaser B, Bray N, O’Donovan MC, Kirov G, Craddock N, and Owen MJ Department of Psychological Medicine, UWCM, Cardiff, United Kingdom Evidence for the involvement of susceptibility variants on chromosome 12q23-24 conferring risk to Major affective disorder has been provided by numerous linkage studies. Investigating two large pedigrees segregating Bipolar disorder, members of our department delineated a haplotype in this area (Green et al., In Press, Am J Psychiatry), which is likely to harbour variants of major effect, which operate according to a highly penetrant autosomal-dominant mode of inheritance in these families. It is thus possible that polymorphisms of smaller effect in the same locus might increase susceptibility to Bipolar disorder in the general population as well. The gene encoding the winged-helix transcription factor RFX4 resides in our broader linkage region on chromosome 12q23 and seems to be involved in the regulation of circadian rhythms due to its expression in the suprachiasmatic nucleus and photic induction (Araki et al., 2004). Because of the strong functional and positional candidacy of the RFX4 gene a comprehensive haplotype-map-based association study was undertaken. The gene was screened for mutations and the LD structure of the region determined

Abstracts through genotyping of 22 common SNPs/InsDels in 95 individuals (average inter-marker distance of 10.0 kb). Analysis of 10 haplotypetagging markers in up to 676 Bipolar patients and 706 controls identified a haplotype of two novel sequence variants associated with Manic-depressive illness (P ¼ 0.042), which appeared to be strongest in the male subpopulation (P ¼ 0.007). Evidence for potential disease association with mutations across the RFX4 region came also from the analysis of the nearby microsatellite D12S2072 (P ¼ 0.009). Allelespecific expression analyses identified a strong cis-regulatory variation that was in high LD with one of the polymorphisms of the significant haplotype. Our findings require replication in independent samples. P2.13 COMPARISON OF MICROSATELLITE MARKERS (MSM) AND HIGH DENSITY SNP GENOTYPING ARRAYS FOR ANALYSIS OF COMPLEX DISEASE: INCREASED POWER TO DETECT LINKAGE AND PREDICTION OF GAPS IN MSM-BASED GENOME SCANS Middleton FA,2 Pato MT,1 Pathak V,2 Morley CP,2 Daly MJ,3 Sklar P,4 and Pato CN1 1 VA Medical Center, Washington, DC, United States 2 SUNY Upstate Medical Center, Syracuse, NY, United States 3 Broad Institute, Cambridge, MA, United States 4 Harvard Medical School, Charlestown, MA, United States In the analysis of complex diseases, whole genome linkage scans using microsatellite markers (MSM) at 10 cM intervals have traditionally been used for defining regions of interest in the search for candidate genes. Recent technologies now permit the simultaneous genotyping of thousands of SNPs in a manner that can also be used for whole genome linkage scans. To compare the power and results obtainable with both approaches, we performed a non-parametric whole genome linkage analysis (using MERLIN) on 12 extended multiplex families segregating for bipolar disorder from a Portuguese population using the Affymetrix GeneChip Human Mapping 10K Array (HMA10K). These families included 107 subjects (36 founders, 71 non-founders) with a mean family size of 8.92 (range 5–17, mode 7), spanning 2.75 generations (range 2–4, mode 3) with a mean of 2.91 affecteds per family. A total of 68 subjects were used for the comparisons of the two platforms. The HMA10K gave consistently high call rates (97%) with very few detectable genotyping errors. Moreover, our comparative analysis indicated the presence of significant linkage peaks in the SNP data in chromosomal regions characterized by poor coverage and low information content on the microsatellite assays. A subsequent analysis extended throughout the entire genome identified numerous regions on other chromosomes where the information content and/or coverage of the 10 cM MSM platform may be largely inadequate for determining the candidate regions to focus on in follow up studies. The combination of increased informativeness and coverage in our 12 family data set specifically led to the identification of two chromosomal regions with genome-wide significant linkage (6q22 and 11p11) that failed to meet this threshold with the microsatellite assays. The ease, speed, and accuracy with which these assays can be performed and analyzed on high quality DNA samples should serve as an impetus to other researchers searching for clues regarding the pathogenesis of complex human diseases. P2.14 FINE MAPPING OF A SUSCEPTIBILITY LOCUS FOR BIPOLAR DISORDER ON CHROMOSOME 4P IN BULGARIAN FAMILIES Kaneva R,1 Chorbov V,1 Milanowa V,2 Kostov C,2 Evans KL,3 Schumacher J,4 Stoyanova V,2 Nikolova-Hill A,5 Onchev G,2 Gaidarova R,6 Krastev S,2 Porteous DJ,3 Kremensky I,1 Propping P,4 and Noethen MM7 1 Laboratory of Molecular Pathology, Medical University of Sofia, Sofia, Bulgaria 2 First Psychiatric Clinic, Department of Psychiatry, Medical University of Sofia, Sofia, Bulgaria 3 Medical Genetics Section, Molecular Medicine Centre, The University of Edinburgh, Edinburgh, United Kingdom 4 Institute of Human Genetics, University of Bonn, Bonn, Germany 5 Bunbary Regional Hospital, South West Mental Health Service, Bunbary, Australia 6 Department of Psychiatry, Medical University of Pleven, Pleven, Bulgaria 7 Department of Medical Genetics, University of Antwerp, Antwerp, Belgium

43

The current study aimed at fine mapping of a region on 4p, where suggestive results were previously reported by our group. The genomewide search for loci linked to bipolar affective disorder (BPAD) in the same Bulgarian sib pair sample produced modest supportive results (MLS ¼ 1.81, NPL ¼ 2.84, P ¼ 0.004 at marker D4S2639 and narrow model, recessive) in the same region on 4p, where significant Lod scores were obtained in a single Scottish family (Blackwood et al., 1996). The area showing positive results spans over about 30 cM, therefore fine mapping to further narrow it down was warranted. Fifteen additional STR markers were genotyped and the newly generated data were added to the initial 6 markers. Parametric and nonparametric multipoint analysis was performed on the complete set of 21 markers. Two phenotype models (‘‘narrow’’: BPI, BPII and schizoaffective manic, ‘‘broad’’: including recurrent major depression in addition to the ‘‘narrow’’ phenotype) were used. The best results were obtained under a narrow diagnostic definition and dominant model close to the peak from the genome scan and D4S419 (MLS ¼ 1.94, NPL ¼ 2.99, P ¼ 0.00037). All seven families contributed to this finding and there was no evidence of genetic heterogeneity. The fine mapping allowed us to define better the haplotypes segregating with the disorder and select the target areas for further LD analysis. The reported results surpass the genome-wide threshold for suggestive linkage and further strengthen the evidence for a BPAD susceptibility locus on 4p.

P2.15 SIB-PAIR STUDY OF EARLY ONSET BIPOLAR AFFECTIVE DISORDER: PRELIMINARY RESULTS OF A GENOME SCAN Etain B,1 Mathieu F,2 Bellivier F,1 Rietschel M,3 Albus M,4 Blackwood D,5 Gurling H,6 Mc Keon P,7 Malafosse A,8 Preisig M,8 Marusic A,9 Noethen M,10 Betard C,11 Henry C,12 Bourgeron T,12 Giros B,2 and Leboyer M1 1 Hoˆpitaux Henri Mondor et Albert Chenevier—Assistance Publique des Hoˆpitaux de Paris, Cre´teil, France 2 Laboratoire INSERM U513, Cre´teil, France 3 Department of psychiatry, University of Bonn, Bonn, Germany 4 District Hospital, Haar, Germany 5 Royal Edinburgh Hospital, Edinburgh, United Kingdom 6 University College London Medical School, London, United Kingdom 7 Department of psychiatry, St Patrick’s Hospital, Dublin, Ireland 8 Division de neuropsychiatrie, Hoˆpital Belle Ide´e, Geneva, Switzerland 9 Institute of Public Health of the Republic of Slovenia, Ljubljana, Slovenia 10 Markus University of Antwerp, Antwerp, Belgium 11 Centre National de Ge´notypage, Evry, France 12 Hoˆpital Charles Perrens, Bordeaux, France Initial enthusiasm for the genetic investigation of bipolar affective disorder (BPAD) has been tempered by a series of failures to replicate linkage results. Conflicting results may be due to the lack of consensus regarding the proper definition of the affected phenotype and the questionable homogeneity of BPAD. Thus, attention is shifting towards the study of specific clinical indicators, which are familial and may be useful for identifying heritable forms of the illness. In BPAD, we and others have reported clinical, familial and biological differences according to age at onset (AAO). Clinical profile have been studied extensively demonstrating that early-onset BPAD is associated with specific clinical features and particuliar comorbid pattern. Evidence of the existence of different subgroups according to age at onset is provided by admixture analysis that identified three subgroups (group 1 (16.9  2.7 years), group 2 (26.9  5 years), group 3 (46.2  8 years)). These three subgroups as well as their thresholds have been replicated in different populations (Bellivier et al., 2001 and 2003). The familial nature of AAO is shown by the similarity in AAO among affected siblings. Relatives of early onset bipolar probands have a higher risk of affective disorders than those of late-onset probands. Preliminary genetic association also suggested differences between early and late onset BPAD. An association between the apolipoprotein E polymorphisms and early-onset BPAD and between late-onset BPAD and the tyrosine hydroxylase gene polymorphisms have been reported. Recently, a segregation analysis showed that the underlying genetic model is significantly different between early-onset BPAD and lateonset BPAD, the best model being one with a non-mendelian major gene and a polygenic component for the early-onset group and a multifactorial model for the late-onset group. To clarify the genetic control of early onset BPAD, a genome wide search was conducted in 71 nuclear families with at least one early-onset bipolar proband (age at onset

44

Abstracts

lower than 21 years old), as part of the European Collaborative Study of Early Onset BPAD (France, UK, Germany, Ireland, Scotland, Switzerland). As a first step towards gene identification, we conducted a genome-wide search using model free methods, as implemented in the GENEHUNTER 2.1 program, on a sample of 87 BPAD sib-pairs genotyped for a panel of 384 well-characterized microsatellites in the Centre National de Ge´notypage (France). 5 regions were suggestive for linkage (0.0002 < P-value < 0.01) in multi-point analysis (2p, 3p, 7q, 10q, 20p). Of special interest are the significant differences in linkage between paternal and maternal alleles for early-onset BPAD (maternal linkage on 3p and paternal linkage on 10q). All results will be discussed regarding results of previous genome scan in bipolar disorder. P2.16 BIPOLAR DISORDER AND GENE EXPRESSION IN LYMPHOBLASTOID CELL LINES: THE TRANSCRIPTOME SCREEN McInnis MG, Chen H, DePaulo JR, and Ross CA Johns Hopkins University, Baltimore, United States Lymphoblastoid cell lines offer a readily available source of RNA for complex disorders affecting the brain, where tissue is limited to postmortem samples. The rapid progress in the development of highdensity gene arrays provides for comparing the gene expression of the available transcriptome, comparing the expression pattern between affected and unaffected subjects. We selected 11 subjects from 10 BP pedigrees with evidence of linkage to chromosome 18q21-22 (in an effort to increase homogeneity) and 11 unaffected 1st degree relatives from the same families. The unaffected relatives were considered beyond the age of risk for BP (average age at interview was 62 yr), the mean age of onset of the affected subjects was 25 yr (average age at interview and sampling was 42 yr). RNA from these 22 subjects was standardized and hybridized on the GeneChip U133 plus 2.0 (Affimetrix) at our Genetics Core facility. Data were analyzed using the Spotfire DecisionSite and Functional Genomics packages. Typically, approximately 40% of genes were expressed in these cell lines. The mean normalized values of the affected group were compared with unaffected relatives and the significance of the fold changes estimated using the Student’s t test. We selected 10 genes for verification using real-time quantitative RT-PCR (Taqman). There were 70 genes showing a 1.3 fold change or greater associated with P < 0.05. Taqman analysis indicated 9 of the 10 genes with a similar fold change in expression, 3 were significant and 3 marginally significant. Two genes, MARCKS and homer1B, have been previously reported in BP or schizophrenia and were down-regulated in our sample. These findings are considered very preliminary and await further study in brain tissue. Study of gene expression in lymphoblastoid cell lines is very attractive in view of the availability, however, caution in interpretation of the findings is warranted.

P2.17 IS COGNITIVE STYLE THE KEY TO MANIA?: RESPONSE TO INDUCED POSITIVE MOOD IN EUTHYMIC BIPOLAR INDIVIDUALS COMPARED WITH HEALTHY CONTROLS Burke A,1 Farmer A,2 Lam D,3 and Sahakian B4 1 Institute of Psychiatry, London, United Kingdom 2 Institute of Psychiatry, London, Ireland 3 Institute of Psychiatry, London, United Kingdom 4 Cambridge University, Cambridge, United Kingdom Bipolar affective disorder is a comparatively common, highly heritable disorder that contributes substantially to the Global Burden of Disease. Although there is compelling evidence of a strong genetic contribution and indications that cognitive therapy may be beneficial, neither the cognitive schemata that underlay bipolar disorder, nor the triggering of manic episodes are well understood. Cognitive factors may influence susceptibility to mood disorders, and both the depressed and manic mood states in bipolar disorder are associated with altered cognitions. Emerging evidence suggests that there may be stable patterns of cognitive response that underlay the manic syndrome. In this study we investigated the nature and extent of cognitive behavioural changes associated with induced positive mood in 30 bipolar individuals who had been euthymic for at least two months, and an equal number of healthy controls. All participants were screened for present mood to ensure they were euthymic at the time of testing. Participants completed questionnaires assessing cognitive style and self-percep-

tions at baseline. A positive mood induction task, the Go Task in a reward condition, was used to induce an elevated mood, after which participants were asked to estimate their success rate and make attributions as to the cause of their success. Participants then completed two neuropsychological tasks, a computerised Gambling task and an Affective Go-NoGo task. Analyses were conducted between cases and controls to assess whether euthymic bipolar individuals, relative to controls: 1. over-estimated their proficiency on the Go task, 2. showed greater increases in positive affect (interested, excited etc.) relating to perceived success, 3. made more risky, reward-oriented decisions on the computerised gambling task, 4. demonstrated biases for processing positive and negative stimuli on the affective go-nogo task. Findings indicated significant differences between cases and controls on these measures, as well as on the baseline measures of general cognitive style. A detailed exploration of the changes in mood, behaviour and cognition that occur in an elevated mood state, will improve our understanding of how mania unfolds and have important implications for the development of psychological treatments for bipolar disorder. P2.18 ESTIMATING THE LOCATION OF BIPOLAR DISORDER SUSCEPTIBILITY GENES BY INCORPORATING PARENT-OF-ORIGIN EFFECTS INTO AFFECTED SIB PAIR ANALYSES Chen WM,1 Fallin DM,2 Lasseter VK,3 Wolyniec PS,3 McGrath JA,3 Nestadt G,3 Liang KY,1 and Pulver AE3 1 The Johns Hopkins Bloomberg School of Public Health department of Biostatistics, Baltimore, Maryland, United States 2 The Johns Hopkins Bloomberg School of Public Health, Department of Epidemiology, Baltimore, Maryland, United States 3 The Johns Hopkins University School of Medicine, Department of Psychiatry and Behavioral Sciences, Baltimore, Maryland, United States We propose incorporating parent-of-origin effects into the assessment of linkage evidence in targeted regions when mapping complex diseases. Parent-of-origin effects, or imprinting effects, are hypothesized epigenetic effects in multiple complex diseases including Alzheimer’s disease, diabetes, and bipolar disorder. Genomic regions with known imprinting effects have been identified. Over 60 imprinted genes have been identified in mice, with the majority of these also imprinted in humans (Wilkins and Haig, 2003, Nat Rev Genet 4:359– 368). A new method for taking into account the possibility of imprinting in estimating the location of a susceptibility gene is a modification of the GENEFINDER software (Liang et al., 2001, Hum Hered 51:64–78), which computes a statistic based on the simultaneous use of the number of alleles shared identical-by-descent at all available markers in affected sib pairs, considering paternal and maternal chromosomes separately. We apply the method to data from a genome scan for bipolar disorder (BP) (Fallin et al., in press, Am J Hum Genet) using 34 multiplex BP pedigrees (38 affected sib pairs) from a genetically more homogeneous population (Ashkenazi Jewish). The genome scan nonparametric analysis (GeneHunter 2.0) suggests four potential BP susceptibility regions on chromosomes 1, 3, 11, and 18. Further analysis with GENEFINDER: 1) suggests that the chromosome 18 evidence is limited to paternally derived alleles (maternal imprinting), 2) allows for a more refined estimate of the 95% confidence interval for the location of a BP susceptibility locus in the chromosome 18 region, and 3) gives slightly narrower 95% confidence intervals and similar estimates for the location of BP susceptibility loci on chromosomes 1, 3, and 11.

P2.19 POLYMORPHISMS ON CHROMOSOME 21 ASSOCIATED WITH SUSCEPTIBILITY TO BIPOLAR AFFECTIVE DOSIRDER Grozeva DV,1 Dimitrova A,2 Georgieva L,1 Nikolov I,1 Hoogendoorn B,1 O’Donovan M,1 Buckland P,1 Owen MJ,1 Craddock N,1 and Kirov G1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Department of Medical Genetics, Medical University Sofia, Sofia, Bulgaria Several linkage studies implicate the distal long arm of chromosome 21 in the aetiology of Bipolar Affective Disorder (BP). We have previously identified an association between BPI with markers within 3 adjacent

Abstracts genes on this chromosome: SH3BGR, WRB and MGC3329 (Kirov et al., in press). We investigated the observed association in a large bipolar sample comprised of 972 cases and 699 matched controls of UK origin. We chose the marker that gave the best P-value in the previous study: rs2837009, centromeric to MGC3329. For genotyping technology we used AmplifluorTM SNP Genotyping System. The products were visualized on Analyst AF fluorescence reader (LJL Biosystems, CA, USA). We replicated our previous results: Putative risk allele frequency: cases 79%, controls 75% (combined P-value for the full case-control sample is 0.007). The signal is strongest among the BP individuals with mood congruent psychotic symptoms (allele frequency 85% in cases, 75% in controls, P < 0.001). This marker rs2837009 is in almost full linkage disequilibrium (LD) with two aminoacid changes: one in the SH3BGR gene (þ69C/G Asp23Gly) and one in the MGC3329 gene (rs2837029 þ49G/A Gly17Ser). Due to the high LD between these 3 markers and the long region of strong LD, it is not possible to conclude which of the SNPs represent disease variants. More replications of independent samples with different LD structure are needed to resolve this issue. P2.20 IDENTIFICATION AND ANALYSIS OF CANDIDATE GENES FROM THE BIPOLAR DISORDER SUSCEPTIBILITY LOCUS ON CHROMOSOME 4Q35 Blair IP,1 Badenhop RF,2 Scimone A,1 Moses MJ,1 Kerr NP,1 Donald JA,3 Mitchell PB,4 and Schofield PR1 1 Garvan Institute of Medical Research, Sydney, Australia 2 University of NSW, Sydney, Australia 3 Macquarie University, Sydney, Australia 4 Black Dog Institute, Sydney, Australia We previously identified a bipolar disorder susceptibility locus on chromosome 4q35 and refined the interval harboring this susceptibility gene to approximately 5Mb, a size that is amenable to positional cloning. Several independent studies have now been described that support the presence of a susceptibility gene at this locus. In order to identify candidate genes for testing for association with bipolar disorder, we previously established a comprehensive transcript map that encompasses the chromosome 4q35 susceptibility locus implicated in our linkage analysis. This transcript map includes at least 22 genes and comprises about equal numbers of well characterized genes and poorly characterized genes. Transcripts within this interval are currently being investigated for association with bipolar disorder. In this study, we selected a subset of genes from this transcript map, based upon tissue expression in the brain, for investigation for association with bipolar disorder. The expression of mRNA was investigated by RT-PCR using cDNA derived from a range of brain tissues. The presence of mRNA in one or more of these brain tissues was confirmed for 18 of 22 genes. These genes were therefore selected for association analysis. We initially determined the genomic structure of each characterized gene. We then identified informative, intragenic single nucleotide polymorphisms (SNPs) by screening all exons and flanking intron sequences in affected individuals from seven bipolar pedigrees that we previously reported as showing evidence for linkage to chromosome 4q35. Analysis of these SNPs was then extended to our unrelated bipolar case-control cohort to test for association with the disorder. Our data suggests that most of the well characterized genes that were analyzed can be excluded from direct involvement in the disorder. We have therefore, excluded over half the genes within the chromosome 4q35 candidate interval from playing a direct pathogenic role in bipolar disorder. Of the remaining genes to be investigated, full-length cDNA and genomic structure has been determined and association analysis is underway. P2.21 ASSOCIATION BETWEEN COMT GENE MET108(158)VAL POLYMORPHISM IN MALES WITH BIPOLAR AFFECTIVE DISORDER Kapelski P,1 Skibinska M,2 Czerski PM,2 Leszczynska-Rodziewicz A,1 Dmitrzak-Weglarz M,2 Rybakowski JK,1 and Hauser J1 1 Department of Adult Psychiatry academy of medical sciences, Poznan, Poland 2 Laboratory of Psychiatric genetics Departemnt of Adult Psychiatry, Poznan, Poland Catechol-O-methyltransferase (COMT) inactivates catecholamine neurotransmitters such as noradrenaline, adrenaline and dopamine. The COMT gene that maps to the chromosome 22q11.1-q11.2. may be considered as a candidate gene in bipolar affective disorder. A common

45

Val/108(158)/Met polymorphism of this gene in humans determines different enzyme activity. COMT gene polymorphism in bipolar patients has been frequently examined, but the studies yielded inconsistent results. The study was performed on 343 patients with bipolar affective disorder (141 males, 202 females), diagnosed according to DSM-IV and ICD-10 criteria, and 349 control subjects (138 males, 211 females). A polymorphism of COMT gene was analysed by PCRRFLP method. Comparing the group of bipolar affective patients with the control subjects we did not find differences in the genotype distribution for COMT polymorphism (P ¼ 0.166), dividing the groups according to gender we also did not find any difference either for males (P ¼ 0.131) or for females (P ¼ 0.075). Comparing allele frequencies we did not find significant difference between patients and controls (P ¼ 0.306), dividing the groups according to gender we find significant difference for males (P ¼ 0.041) but not for females (P ¼ 0.728). Affected males have Met allele more often than control subjects. In our study we find an association between Met allele of the studied polymorphisms of COMT gene in males with bipolar affective disorder. Our results may provide a preliminary evidence of a possible small contribution of the COMT gene Val108(158)Met polymorphism to the bipolar affective disorder susceptibility in Polish population. P2.22 COGNITIVE STYLE AND PERSONALITY IN EUTHYMIC PATIENTS WITH BIPOLAR DISORDER Hyde SA,1 Jones LA,1 Jones IR,2 Haque S,1 Gordon-Smith K,2 Heron J,1 Caesar ES,1 Cooper C,1 Forty L,2 Lyon L,1 Greening J,1 Sham P,3 Farmer A,3 McGuffin P,3 Scott J,4 and Craddock N2 1 Department of Psychiatry, Division of Neuroscience, University of Birmingham, Birmingham, United Kingdom 2 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 3 SGDP Research Centre, Institute of Psychiatry, London, United Kingdom 4 Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom Abnormalities of cognitive style and personality in bipolar disorder are of theoretical and clinical importance. If aspects of cognition and personality are shown to be markers of underlying liability to bipolar disorder, they could be useful in aetiological research, such as molecular genetic studies. We have examined aspects of cognitive style (selfesteem and dysfunctional attitudes) and personality (extraversion, neuroticism and psychoticism) in large, representative, narrowly defined and well characterised samples of euthymic individuals with bipolar I disorder (N ¼ 118) and unipolar major recurrent depression (N ¼ 265), and healthy controls at low genetic risk for developing a mood disorder (N ¼ 268). Our data suggest that: i) measures of cognitive style and personality are substantially affected by current levels of depressive symptomatology, ii) bipolar and unipolar patients demonstrate abnormal patterns of cognitive style and personality, iii) these abnormal patterns are similar in both disorders, and iv) cognitive style is significantly associated with neuroticism, and when neuroticism is taken into account, adds very little power in predicting patient versus control status. Our study design cannot determine whether elevated neuroticism is a cause or a consequence of affective illness. Prospective longitudinal studies beginning in the pre-morbid state are required.

P2.23 ‘‘RAPID CYCLING’’ MAINLY DETERMINES SIGNIFICANT FINDINGS BETWEEN THE BDNF GENE AND BIPOLAR DISORDER Mu¨ller DJ, De Luca V, Sicard T, King N, Lanktree M, Strauss J, Mundo E, Macciardi F, Muglia P, and Kennedy JL Centre for Addiction and Mental Health, Neurogenetics Section, Toronto, Canada Our group has previously reported significant associations between the Val66Met and GT(n) repeat polymorphisms of the BDNF gene with bipolar disorder (BD). However, these findings have not consistently been replicated which may be explained by considerable methodological heterogeneities among studies. In our sample of 312 nuclear families (with at least one offspring affected with BD) we have added four novel SNPs (rs3763965, hCV11592756, rs2049045, rs2140887) located within or in the vicinity of the BDNF gene and performed TDT and haplotype analyses. Furthermore, we investigated whether our findings may be determined by the phenomenon of rapid cycling (RC)

46

Abstracts

that has been present in 58 patients (19%). According to DSM-IV, RC is defined by the occurrence of four or more (depressive and/or manic) mood episodes within 12 months. Our analyses revealed that both the hCV11592756 and the rs2049045 SNPs were significantly associated with BD using TDT analyses (P ¼ 0.02 and P ¼ 0.009, respectively). These two SNPs appear to belong to a haplotype block flanking the Val66Met and GT(n) repeat polymorphisms. The haplotype including the hCV11592756 (A-allele), Val66Met (G-allele), GT(n) repeat (allele 3) and rs2049045 (G-allele) proved to be over transmitted (49:25) in patients with BD (P ¼ 0.008). Interestingly, and despite the reduced sample size and reduced power, TDT analyses revealed significant associations with similar or higher likelihood ratio scales with all four SNPS in the sample of patients with RC as compared to the total sample. Conversely, removing patients with RC from the total sample led to non-significant findings with all four SNPs. Our findings suggest that the BDNF gene is likely to be associated with a specific homogeneous subgroup of patients with BD, namely RC that is supposed to occur in 5%–15% of persons with BD. Thus, nonreplications in studies focusing on the BDNF gene and bipolar disorder may either be due to a non-distinction of this particular phenotype or due to limited samples and power. Unless our findings are false positive ones, our observations should be taken into account in past and future genetic studies dealing with the BDNF gene in mood disorders. P2.24 EVOLUTIONARY-BASED TRANSMISSION DISEQUILIBRIUM TESTING FOR BIPOLAR DISORDER-ASSOCIATED HAPLOTYPES IN PARENT-CHILD TRIOS Nicodemus KK,1 Fallin MD,1 Lasseter VK,2 McGrath JA,2 Wolyniec PS,2 Valle D,3 Liang KY,4 and Pulver AE2 1 Department of Epidemiology, Johns Hopkins School of Public Health, Baltimore, MD, United States 2 Department of Psychiatry & Behavioral Sciences, Johns Hopkins School of Medicine, Baltimore, MD, United States 3 Department of Pediatrics and of Molecular Biology, Johns Hopkins School of Medicine, Baltimore, MD, United States 4 Department of Biostatistics, Johns Hopkins School of Public Health, Baltimore, MD, United States The recent advent of haplotype analysis, which takes into account a series of alleles on each chromosome, has led to a quandary. Although haplotype analysis uses more information than single marker analysis, the necessary increase in degrees of freedom can reduce power. Therefore, an a priori method of combining haplotypes would be an ideal way to reduce the number of parameters needed and increase the power of the test statistic. In 2001, Seltman et al. proposed a method that exploits the evolutionary relationships between the haplotypes to determine how they may be combined, which they called the Evolutionary Tree Transmission Disequilibrium Test (ET-TDT)(Seltman et al., 2001, Hum Genet 68:1250–1263). The ET-TDT provides an intuitive and meaningful method of data reduction while increasing power versus alternatives such as collapsing haplotypes together based on lack of association with disease, by frequency less than some arbitrary cut-off value such as 5%, or by choosing one ‘‘at-risk’’ haplotype to examine versus all other haplotypes. In addition, the ET-TDT may help localize causal variants because haplotypes with evidence of increased risk should be evolutionarily more similar to one another than to non-disease haplotypes, especially among populations that are relatively recently founded, such as the Ashkenazi Jewish population. We applied the ET-TDT to data from a 63-candidate gene association study for bipolar I disorder among 374 case-parent trios of Ashkenazi Jewish descent and compared the results obtained using the ET-TDT against a conditionally-weighted haplotype association test approach that does not use any data reduction techniques (HBAT, (Horvath et al., 2004, Genet Epidemiol 26:61–69), see abstract by Fallin in this issue). P2.25 POPULATION-BASED AND FAMILY-BASED STUDIES ON THE SEROTONIN TRANSPORTER GENE POLYMORPHISMS AND BIPOLAR AFFECTIVE DISORDER: A SYSTEMATIC REVIEW AND META-ANALYSIS Cho HJ,1 Meira-Lima I,2 Michelon L,2 Cordeiro Q,2 and Vallada H2 1 Institute of Psychiatry, King’s College London, London, United Kingdom 2 Department of Psychiatry, University of Sa˜o Paulo Medical School, Sa˜o Paulo, Brazil

The serotonin transporter (5-HTT) is a candidate gene for bipolar disorder (BPD). It has been investigated for association with the illness in a series of studies but overall results have been inconsistent and its role in the disorder remains controversial. Systematic reviews using meta-analytical techniques are a reliable method for objectively and reproducibly assessing individual studies and generating combined results. We performed a meta-analysis of published studies investigating the association between BPD and the 5-HTT promoter 44 bp insertion/deletion and the intron 2 VNTR polymorphisms (both population-based and family-based studies). The literature was searched using Medline and Embase to identify studies for inclusion. For population-based studies our review revealed the pooled odds ratio for 5-HTTLPR as significant (OR 1.16 95% CI 1.05–1.27). However, our meta-analysis of family-based studies did not support this finding (OR 1.20 95% CI 0.93–1.54). The pooled odds ratio for the 5-HTT intron 2 VNTR was not significant for population-based studies (OR 1.13 95% CI 0.97–1.33), but significant for family-based studies (OR 1.54 95% CI 1.10–2.17). Despite of the opposite results, the trend to a significant association between the 5-HTT gene and the susceptibility for BPD should be taken into account in future research. P2.26 TDT ANALYSES SUGGEST TWO DISTINCT HAPLOTYES IN G PROTEIN RECEPTOR KINASE-3 (GRK3) ARE ASSOCIATED WITH BIPOLAR DISORDER Barrett TB, Emberton JE, Liang SG, Nievergelt CM, Schork NJ, and Kelsoe JR University of California at San Diego and San Diego VA Healthcare System, San Diego, United States In a genome-wide linkage survey we have previously shown evidence indicating chromosome 22q12 contains a susceptibility locus for bipolar disorder (BPD). Two independent family sets gave lod scores of 2.19 and 2.72 at or near marker D22S419 which lies 20 kb centromeric to the gene G protein receptor kinase-3 (GRK3). GRK3 is an excellent candidate susceptibility gene for BPD since GRKs play key roles in the homologous desensitization of G protein-coupled receptor signaling. In 2002 we reported an association study indicating GRK3 may be a risk gene for BPD (Barrett TB, et al. Molecular Psychiatry). In that study we sequenced the putative promoter region, exons, and intron flanking each exon, in 14 affected individuals selected from families showing evidence for linkage between the GRK3 region and BPD. We found six variants in the 50 -UTR and promoter region, but no coding or obvious splice variants. Transmission disequilibrium data from two triad sets indicated a variant in the promoter region (P-5) was associated with BPD. In 329 northern European Caucasian triads the transmission to non-transmission ratio was 26:7.7, chi-square ¼ 9.6, P ¼ 0.0019. However, these six promoter SNPs were present at only low to moderate frequency. Altogether they ‘‘tagged’’ only 25% of the founding chromosomes, and, hence, provided only modest power to detect overall association between GRK3 and BPD. To identify additional variants for use in association studies we have now sequenced a 30 kb genomic segment of GRK3, centered on the promoter and first exon, in 25 individuals with BPD. 50 variants were identified, half of which were present in three or more subjects. A subset of these variants was then genotyped in 181 Caucasian nuclear families. TDT analyses using Unphased identified an overtransmitted haplotype, defined by two SNPs located upstream of the promoter region, for which the transmission to non-transmission ratio was 121:74, chisquare ¼ 7.6, P ¼ 0.0057. This haplotype was found on an entirely different set of chromosomes from the previous set of haplotypes defined by SNPs in the promoter region. This haplotype was present on 18% of the founding chromosomes in these families. The P-5 haplotype, by comparison, was present on 2% of founding chromosomes. The data from these studies provide evidence that at least two distinct haplotypes in the gene GRK3 are associated with BPD. One of these haplotypes is relatively common. These new findings add support for the hypothesis that a dysregulation in GRK3 expression alters signaling desensitization and thereby predisposes to the development of bipolar disorder. P2.27 HIGH RESOLUTION LINKAGE DISEQUILIBRIUM STUDY OF BIPOLAR DISORDER AND SNPS ON A 5-MB INTERVAL OF 22Q Liang SG,1 Schork NJ,1 Gaucher MA,2 and Kelsoe JR1 1 UCSD Dept of Psychiatry, La Jolla, United States 2 San Diego VA Healthcare System, La Jolla, United States

Abstracts We have previously reported results of a genome scan of bipolar disorder in which we found evidence of a susceptibility locus on chromosome 22. Twelve markers yielded Lod scores greater than 1.0 over a 20 cM interval. The maximum lod score was 3.8 at D22S278. We have now examined 189 SNPs in a 5-MB region flanking D22S278 in a sample of 499 individuals from 137 nuclear families in order to conduct linkage disequilibrium analyses to fine map the bipolar susceptibility gene. Single-marker TDT statistics were calculated, and several SNPs were found to be in linkage disequilibrium with bipolar disorder. The maximum evidence for linkage disequilibrium in the region was 684KB telomeric of D22S278 (P ¼ 0.0048). We then performed the TDT test using a 3-SNP haplotype moving window approach using the program Unphased. Phase-uncertain haplotypes were generated using an EM algorithm. Several haplotypes were also found to be in linkage disequilibrium. The maximum evidence for linkage disequilibrium was generated by a haplotype 684mb telomeric of D22S278 and resulted in a nominal P-value of 0.0016. However, these results were not significant when corrected for the number of tests by using a permutation analysis across all markers in the region. Our results prioritize this region for further study and indicate the need for a larger sample size. P2.28 LINKAGE HETEROGENEITY AND MARKER INFORMATIVITY ANALYSIS OF THE NIMH BIPOLAR FAMILIES FROM WAVES 3 AND 4 Greenwood TA,1 Schork NJ,1 Dick DM,2 Reich T,2 Foroud T,3 Nurnberger JI,3 Berrettini W,4 Byerley W,5 Coryell W,6 DePaulo R,7 McInnis M,7 Gershon ES,8 McMahon FJ,9 Scheftner W,10 and Kelsoe JR1 1 University of California, San Diego, La Jolla, United States 2 Washington University, St. Louis, United States 3 Indiana University, Indianapolis, United States 4 University of Pennsylvania, Philadelphia, United States 5 University of California, Irvine, Irvine, United States 6 University of Iowa, Iowa City, United States 7 Johns Hopkins University, Baltimore, United States 8 University of Chicago, Chicago, United States 9 Mood and Anxiety Program, National Institute of Mental Health, National Institutes of Health, U.S. Department of Health and Human Services, Bethesda, United States 10 Rush-Presbyterian Medical, Center, Chicago, United States Locus heterogeneity is known to plague the identification of disease susceptibility loci via linkage mapping of complex traits and diseases. As part of the wave 3 data release, the NIMH Genetic Initiative for Bipolar Disorder has collected 250 families from 10 different field centers. The combined family set is thus likely to contain individuals with varying genetic backgrounds and genetic susceptibility profiles. We have applied a novel mixture model in a reanalysis of these families that accommodates both locus heterogeneity and marker informativity. Our results provide information on not only regions of the genome likely to harbor susceptibility loci but also information on the degree to which heterogeneity influences the linkage signal at that locus. Ultimately, we find that previous linkage analyses that did not accommodate locus heterogeneity could have been underpowered or inappropriate for large data sets gathered on families with different origins. For example, the results of the previous analyses of the wave 3 indicated suggestive evidence of linkage to chromosomes 6q and 17q. Our findings also suggest evidence for linkage to 6q with a linkage peak at 108 cM, but we were unable to replicate previous findings of linkage to 17q in our analyses. Additionally, we found suggestive evidence for linkage on several other chromosomes that was not previously reported in this data set. Similar analyses of the wave 4 data will also be presented. These results suggest the value of incorporating marker informativity and locus heterogeneity in analysis of large datasets. P2.29 ASSOCIATION STUDY BETWEEN 1727 A/T, 50 C/T AND (CAA)N REPEAT GSK-3BETA GENE POLYMORPHISMS AND BIPOLAR DISORDER Bocchio-Chiavetto L,1 Scassellati C,1 Bonvicini C,1 Bignotti S,1 Rotondo A,2 Carlini M,2 Perez J,1 and Gennarelli M1 1 IRCCS S. Giovanni di Dio FBF, Brescia, Italy 2 Dept Psychiatry, Neurobiology, Pharmacology and Biotecnology, University of Pisa, Pisa, Italy Structural and neuroimaging studies have reported that Bipolar Disorder (BD) may be caused by early alterations in the normal

47

program of brain maturation. The WNT/Wingless family of genes plays critical regulatory functions in brain development and could be implicated in producing some of the abnormalities observed in developmental disorders. The glycogen synthase kinase-3 (GSK-3) is a downstream component of the Wnt pathway. Moreover, the gene localizes in a region of chromosome 3 (q13.3) that has been indicated as susceptibility locus for Bipolar Disorder (Badenhop et al., 2002). On this basis, we analyzed, in a sample of 174 bipolar patients and 502 healthy individuals, two common SNPs at position 1727 A/T and 50 C/T and a (CAA)n repeat polymorphism localized in intron 1 of the gene. The allele and genotype distributions for the three polymorphisms investigated are not different between patients and controls, suggesting that GSK3beta gene does not play a major role in susceptibility to Bipolar Disorder. This results confirms the negative findings reported previously for the 50 C/T polymorphism (Benedetti et al., 2004). Further association analyses between GSK3beta haplotypes and different clinical subgroups of bipolar patients are needed to define the exact role of this gene in the pathophysiology of BD. P2.30 REPLICATION OF LINKAGE OF BIPOLAR DISORDER TO CHROMOSOME 17 IN A NEW INDEPENDENT PEDIGREE SERIES Greenwood TA,1 Schork NJ,1 Sadovnick AD,2 Remick RA,3 Keck PE,4 McElroy S,4 Shekhtman T,1 Gaucher M,1 and Kelsoe JR1 1 La Jolla, United States 2 University of British Columbia, Vancouver, Canada 3 St. Paul’s Hospital, Vancouver, Canada 4 University of Cincinnati, Cincinnati, United States A sample of 34 pedigrees segregating bipolar disorder was ascertained through a collaboration among three sites (UCSD, U. Cincinnati, U. British Columbia). This represents the second set of families from this consortium following a previous report on an initial set of 20 pedigrees. Families were ascertained through a bipolar I or II proband for the presence of bipolar I, bipolar II or recurrent major depression in at least two other family members. All subjects were interviewed using the SCID, and diagnoses were determined by a consensus best estimate procedure. A total of 440 markers at an average spacing of 8 cM were genotyped in 229 family members using fluorescent methods. Initial nonparametric analyses of chromosome 17 provide evidence for a modest replication of linkage to this chromosome previously reported in the NIMH Genetic Initiative for Bipolar Disorder wave 3 family set with a LOD of 2.1 at 34 cM. A LOD of 1.6 was also observed on chromosome 6q in a region that showed evidence for linkage in the NIMH data. The results of the complete genome scan analysis will be presented.

P2.31 GENOME-WIDE SIGNIFICANT EVIDENCE FOR LINKAGE ON CHROMOSOME 9Q31-Q33 IN NORTH SWEDISH BIPOLAR FAMILIES Venken T,1 Claes S,1 Sluijs S,1 Adolfsson R,2 Van Broeckhoven C,1 and Del-Favero J1 1 Department of Molecular Genetics (University of Antwerp), Antwerp, Belgium 2 Department of Psychiatry (University of Umea), Umea, Sweden A genome-wide scan was performed in 9 multiplex bipolar families (BP), ascertained in the Va¨sterbotten province in the North of Sweden. The north Swedish population is an old population with geographical isolation and low immigration rate and can therefore be considered as an isolated population. Since isolated populations are likely to have both a reduced number of genetic vulnerability factors and reduced environmental background compared to outbred populations, its use maximizes the probability of detecting genes by genetic approaches. Initial genome-wide linkage analysis revealed 3 loci with suggestive evidence for linkage located on chromosomes 9q31.1-q34.1 (MPLOD 3.24), 6q22.2-q24.2 (MPLOD 2.48) and 2q33-q36 (MPLOD 2.26) under a recessive affected-only model, with genetic intervals of respectively 26, 29 and 32 cM. Fine mapping of the 9q candidate region at 200 kb STR spacing resulted in genome-wide significant linkage at 9q31-q33 with a max MPLOD of 3.70 at 122.9 cM from 9pter under a recessive affectedonly model. Additionally, we confirmed linkage at 6q23-q24 and 2q33q36 applying a 2 cM density STR map under a recessive affected-only model. Evidence for linkage increased on chromosome 6 to a max MPLOD of 3.25 at 146 cM from 6pter, while the max MPLOD on

48

Abstracts

chromosome 2 decreased to 2.16 at 200 cM from 2pter. The 95% confidence intervals of the 3 candidate regions extend over 4, 3 and 22 cM, respectively. Genealogical studies showed that 3 9q linked families descend from a common founder couple about 10 generations ago. Interestingly, these families enabled the identification of an identical by descent (IBD) haplotype in patients allowing further reduction of the candidate region to a 1.6 Mb interval. P2.32 TUMOR NECROSIS FACTOR-ALPHA (TNFALPHA) GENE -308 G/A POLYMORPHISM IN BIPOLAR AFFECTIVE DISORDER Leszczynska-Rodziewicz ALR,1 Dmitrzak-Weglarz MDW,2 Czerski PC,2 Kaczmarkiewicz-Fass MKF,1 and Hauser JH2 1 Department of Adult Psychiatry academy of medical sciences, Poznan, Poland 2 Laboratory of Psychiatric genetics Departemnt of Adult Psychiatry, Poznan, Poland Tumor Necrosis Factor alpha (TNFa) is a cytokine involved in immunological response, the abnormal immunological response and disturbances of cytokine system are though to be involved in pathogenesis of bipolar disorder. There was only a single case/control association study of TNFa promoter polymorphism (-G308A) in bipolar disorder with negative results (Meira-Lima et al. Mol Psychiatry. 2003 Aug,8(8):718-20). There are ambiguous reports of the role of this polymorphism in the susceptibility to major depressive disorder. In the present work, promoter -G308A TNFa polymorphism was studied in 364 patients with bipolar affective disorder, and in 351 healthy controls. In highly homogenous sample drawn from Polish population of Wielkopolska the case/control study design was used. The analysis of TNFa polymorphisms was performed by PCR-RFLP method (Wilson et al. Hum Mol Genet. 1992 Aug,1(5):353). We observed an association of TNFa -308G allele (P ¼ 0.031) with bipolar affective disorder. When analyzing both genders separately the differences in the frequency of alleles and genotypes did not reach the level of statistical significance. Our results may provide a preliminary evidence of a possible small contribution of the TNFa gene -G308A polymorphism to the bipolar affective disorder susceptibility in Polish population.

P2.33 VARIATION AT BRAIN-DERIVED NEUROTROPHIC FACTOR (BDNF) INFLUENCES RAPID CYCLING IN A LARGE UK CAUCASIAN BIPOLAR DISORDER CASE-CONTROL SAMPLE Green EK,1 Raybould R,1 Nikolov I,2 McGregor S,2 Jones L,3 Jones I,1 Young A,4 Owen MJ,1 Kirov G,1 and Craddock N1 1 Department of Psychologcial Medicine, University of Wales College of Medicine, Heath Park, Cardiff, United Kingdom 2 Biostatistics and Bioinformatics Unit, University of Wales College of Medicine, Heath Park, Cardiff, United Kingdom 3 Molecular Psychiatry Group, Division of Neuroscience, University of Birmingham, Birmingham, United Kingdom 4 Department of Psychiatry, University of Newcastle upon Tyne, Newcastle, United Kingdom BDNF (Brain-derived neurotrophic factor), a member of the neurotrophin superfamily, plays important roles in neuronal survival, proliferation and plasticity. Over recent years both neurotrophic and synaptic plasticity models of pathogenesis for mood disorders have been advanced. Thus, BDNF can be considered a plausible candidate gene for Bipolar disorder, as well as for other mood disorder phenotypes. Two recent family-based association studies have provided evidence for association between DNA variation at BDNF and susceptibility to Bipolar illness. Sklar et al. (2002) reported a significant association for the allele G at the val66met SNP (P ¼ 0.042) in 136 parent-proband trios, which was replicated in a further 334 parent-proband trios (P ¼ 0.066). This was supported by Neves-Pereira et al (2002) (P ¼ 0.0064) in 283 nuclear families. We have studied variation at the BDNF gene in our UK Caucasian case-control sample (DSMIV Bipolar I Disorder cases n ¼ 848, Bipolar II Disorder cases n ¼ 96, controls n ¼ 685). We have undertaken genotyping of 4 polymorphisms within the gene: (a) Val66Met, (b) a microsatellite (1040 bp upstream of the transcription initiation site) that, together with Val66Met, formed a significant 2 marker haplotype in the study of Neves-Pereira et al., (c) SNP a20 (Sklar et al., 2002), and (d) a 50 SNP rs2049046. Within our sample there was no overall evidence of allele, genotype or haplotype

association—and relevant frequencies were closely similar between cases and controls. However, there was evidence that variation at Val66Met SNP influences susceptibility to rapid cycling in our sample (frequencies of alleles: rapid cycling patients G ¼ 88.4%, A ¼ 11.6%, controls G ¼ 81.2%, A ¼ 18.8%, P ¼ 0.005). We conclude that the BDNF gene does not play a major role in influencing susceptibility to Bipolar Disorder as a whole but may influence susceptibility to, or influence the course of illness in, a subset of the disorder. P2.34 NO ASSOCIATION BETWEEN POLYMORPHISMS OF DRD2, DRD3, DRD4, DAT, COMT, BDNF GENES AND PROPHYLACTIC LITHIUM RESPONSE IN BIPOLAR DISORDER Kosmowska M,1 Suwalska A,1 Lesczynska-Rodziewicz A,1 Dmitrazak-Weglarz M,2 Skibinska M,2 Czerski PM,2 Szczepankiewicz A,2 Rybakowsijk,1 and Hauser J1 1 Dept of Adult Psychiatry, University of Medical Sciences, Poznan, Poland 2 Laboratory of Psychiatric Genetics, Department of Adult Psychiatry, University of Medical Sciences, Poznan, Poland The experimental and clinical data suggest that the dopamine system can be implicated in the mechanism of the therapeutic activity of lithium, in case of manic state as well as in the prevention of relapse of bipolar disorder. Genes coding for dopamine receptors are among candidates for genetic studies on prophylactic lithium response. COMT (catechol-O-methyltransferase) is the enzyme which plays a main role in catecholamine degradation. BDNF (brain-derived neurotrophic factor) is the main neurotrophin in central nervous system. Dysfunction in the regulation of BDNF can be implicated in pathophysiology of affective disorders. The neurotrophic activity may determine one of the mechanisms of the therapeutic effect of antidepressant and normothymic drugs such as: lithium and valproate. The study was performed on 87 patients with bipolar affective disorder, diagnosed according to DSM IV and ICD-10 criteria. All the patients were divided into two groups of lithium responders (71 patients) and non-responders (16 patients). Responders are defined as persons who have at least 50% reduction of number of episodes of bipolar affective disorder per year. Polymorphisms of DRD2 (-141C ins/del), DRD3 (Ser9Gly), DRD4 (-521 C/T), COMT (Val108(158)Met) and BDNF (Met66Val) were analyzed by PCR-RFLP method, DAT 30 UTR VNTR polymorphism was analysed by PCR-VNTR method. There were no statistically significant differences in the frequency of alleles and genotypes between responders and nonresponders. The trend toward higher frequency of BDNF gene Met allele in responders group was observed (P ¼ 0.061). In our study we have not confirmed an association between the studied polymorphisms of DRD2, DRD3, DRD4, COMT and DAT genes and prophylactic lithium response in bipolar affective disorder. Positive treatment response effect may be related to BDNF gene Met66Val polymorphism.

P2.35 VARIATIONS IN THE NMDA RECEPTOR SUBUNIT 2B GENE (GRIN2B) AND BIPOLAR DISORDER: A CASE-CONTROL STUDY Di Maria E,1 Gulli R,1 Ducci F,2 Carlini M,2 Bianchi C,2 Ciotti P,1 Bellone E,1 Gennarelli M,3 Rotondo A,2 Ajmar F,1 and Mandich P1 1 Department of Neuroscience, Ophthalmology and Genetics, University of Genova, Genova, Italy 2 Department of Psychiatry, Neurobiology, Pharmacology and Biotechnology, University of Pisa, Pisa, Italy 3 Genetics Unit, IRCCS San Giovanni di Dio-Fatebenefratelli and University of Brescia, Brescia, Italy A well established model for the pathophysiology of major psychoses postulates a role for the NMDA-mediated glutamate transmission. Lithium and valproate (commonly used to treat Bipolar illness) were found to stimulate glutamate release via the N-Methyl-D-aspartate receptor (NMDAR). Functional NMDARs are assembled from two ubiquitous NR1 subunits and two of a family of four NR2 subunits (NMDAR2A-D) which are expressed in different areas of the brain. Recent studies suggested a possible involvement of the genes coding the NR1 and the 2A subunits in the pathogenesis of Bipolar Disorder (BP). The 2B subunit, coded by the human GRIN2B gene, can be considered a candidate gene based on its functional properties and its selective expression in cerebral cortex, hippocampus and basal ganglia. We tested the hypothesis that single nucleotide polymorphisms within the

Abstracts GRIN2B gene are associated with BP. A cohort of 103 individuals diagnosed as having DSM-IV Bipolar I disorder (mean age ¼ 34.3, males/females ¼ 0.56) was included in the study, and compared with 145 normal subjects. The case-control sample was genotyped for three SNPs, located in the 50 UTR region (T-200G), in the exon 13 (SNP2664), and in the 30 UTR (SNP890) of the gene, respectively. The chi-squared analysis on the case-control sample did not demonstrate any difference in allele and genotype frequency of each polymorphism, in patients as compared with normal controls. The haplotype analysis showed that SNP890 and SNP2664 are in disequilibrium with each other, allowing to carry out the case-control match on the frequencies of the two polymorphism haplotype. Again, the analysis did not provide evidence of association. In order to improve the statistical power and evaluate these results with more confidence, we are currently extending the probands cohort. P2.36 ATTEMPTED CONFIRMATION OF THE FINE MAPPING OF A PUTATIVE BIPOLAR SUSCEPTIBILITY LOCUS ON CHROMOSOME 4P16.1 AT THE PPP2R2C GENE LOCUS Choudhury K,1 Puri V,1 McQuillin A,1 Kalsi G,1 Bass N,1 Lawrence J,1 Curtis D,2 and Gurling HMD1 1 University College London, London, United Kingdom 2 St Bartholomew’s and Royal London School of Medicine and Dentistry, London, United Kingdom

49

mitochondria related genes with bipolar disorder. We quantified the mRNA expression level of five nuclear-encoded mitochondrial complex I subunit genes other than NDUFV2, three complex IV subunit genes, and four mitochondrial transcription related genes using a real-time quantitative reverse transcription polymerase chain reaction (RTPCR) method in the lymphoblastoid cell lines from patients with bipolar disorder and controls. The trends of decreased expression levels in patients with bipolar I disorder compared with control subjects were found in most of complex I subunit genes. Decreased expression levels of these genes were well correlated with that of NDUFV2. No statistically significant alterations of mRNA expression levels were found between bipolar patients and controls among two of three complex IV subunit genes and all transcription related genes. These findings suggested that the decreased expression of NDUFV2 have a considerable effect on other subunit genes in the mitochondrial respiratory chain and presented the further evidence of biological significance of NDUFV2 in bipolar disorder. P2.38 REPLICATION IN AN INDEPENDENT SAMPLE OF A LINKAGE FINDING FOR BIPOLAR DISORDER IN 16P12.3 ´ mond C, Chagnon Y, Me´rette C, Cayer M, Fournier A, E Roy M-A, and Maziade M Centre de recherche Universite Laval Robert-Giffard, Beauport, Canada

Benoit et al. (Patent number WO2004001068, http://gb.espacenet. com/) describe single nucleotide polymorphisms showing allelic and haplotypic association with bipolar disorder in a British and Irish sample of 315 bipolar research subjects and 300 supernormal controls made available by the UCL Molecular Psychiatry Laboratory. SNP 9924169/139 showed allelic association with bipolar disorder with a significance of P ¼ 0.00036. When this marker was combined with SNP 24-2417/216, the base pair GA haplotype was associated with bipolar disorder with a significance of P ¼ 0.0005. The protein phosphatase 2A, regulatory subunit B, gamma isoform gene (PPP2R2C) is localised at this chromosome 4p16.1 locus and is a candidate gene for increasing susceptibility to bipolar disorder. Northern blot analysis has shown a 4.7-kb transcript expressed at high levels in the brain. The product of the PPP2R2C gene is a subunit of the protein phosphatase 2A (PP2A), a heterotrimeric serine/threonine phosphatase enzyme which regulates the initiation of DNA replication through negative control of cell growth and division (Hu et al., 2000: Genomics 67: 83–86). In order to confirm the findings of Benoit et al., we genotyped two polymorphic microsatellite markers, D4S431 and D4S2366, which are situated within and also 5 prime to the PPP2R2C gene, respectively. Both markers showed a trend towards a significant association (D4S431, P ¼ 0.072, D4S2366, P ¼ 0.079) with bipolar disorder in the UCL case-control samples. Haplotype analysis with these two microsatellite markers using GENECOUNTING showed a nominally significant (P ¼ 0.03) global association with schizophrenia but this result was not confirmed with a permutation test. Further data, including additional SNP associations and haplotype association analyses on the fine mapping of this putative bipolar affective disorder locus in the UCL sample will be presented. Funded by the Neuroscience Research Charitable Trust.

We recently completed a dense genome scan in a first sample of 21 multigenerational families from Quebec/Canada densely affected by schizophrenia (SZ) or bipolar disorded (BP) (Maziade et al., submitted). Several interesting results emerged form such a full scan including a lod score of 4.05 for BP in 16p12.3, according to a model-based multipoint analysis, and a corresponding proportion of identical by descent (IBD) allele sharing of 0.66 (P ¼ 3  104) at marker D16S499, according an ASP analysis performed with SAGE. We now report the replication of this original finding in an independent sample of 27 new families from Quebec involving 535 family members of whom 133 were affected by a broad phenotype definition of bipolar disorder which includes BP I, BP II and recurrent major depression. Model-based (using FASTLINK) and model-free (using SAGE) linkage analysis in the new sample of 27 families as well as in the combined sample of 48 families involving 194 affected sib-pairs. At the exact same location as in sample 1 (D16S499), the proportion of IBD allele sharing in affected pairs in sample 2 was estimated to be 0.59 (P ¼ 6  104). When samples 1 and 2 were combined, the evidence for linkage further improved to reach a level of significance of 8.25  106 corresponding to an estimated proportion of IBD sharing of 0.60. This positive finding is clearly supported by the surrounding markers that also showed an overproportion of allele sharing among ASP with P-values smaller than 0.03. The model-based analysis failed to detect the replication and we are in the process of investigating the methodological reasons for such a discrepancy between the two approaches. Chromosome 16p12.3 have previously been reported to be a site of positive linkage findings for BP by our own group but also by several others. Our replication study allowed to further strengthen the evidence for a BP gene in 16p12.3. Other regions of interest identified through the genome scan of our first sample of families are being investigated for replication both for SZ and BP.

P2.37 EXPRESSION OF MITOCHONDRIA RELATED GENES IN LYMPHOBLASTOID CELLS FROM PATIENTS WITH BIPOLAR DISORDER Washizuka S,1 Kakiuchi C,2 Mori K,2 Tajima O,3 Akiyama T,4 and Kato T2 1 Department of Psychiatry, Shinshu University School of Medicine, Matsumoto, Japan 2 Laboratory for Molecular Dynamics of Mental Disorders, Brain Science Institute, RIKEN, Saitama, Japan 3 Kyorin University School of Health Sciences, Tokyo, Japan 4 Department of Neuropsychiatry, NTT East Kanto Medical Center, Tokyo, Japan

P2.39 ROLE OF VAL66MET POLYMORPHISM AT THE BRAIN DERIVED NEUROTROPHIC FACTOR GENE (BDNF) IN MAJOR AFFECTIVE DISORDERS: ASSOCIATION WITH PSYCHOTIC SYMPTOMATOLOGY IN BIPOLAR DISORDER Arias B,1 Catala´n R,2 Gasto´ C,2 Gutie´rrez B,1 Fatjo´-Vilas M,1 Aguilera M,1 and Fan˜ana´s L1 1 Unitat d’Antropologia. Dept de Biologia Animal. Facultat de Biologia. Universitat de Barcelona., Barcelona, Spain 2 CSM Esquerre de l’Eixample. Hospital Clı´nic de Barcelona, Barcelona, Spain

Several studies have suggested mitochondrial abnormality in bipolar disorder. We reported the association of mitochondrial complex I subunit gene, NDUFV2 at 18p11, with bipolar disorder. However, it was unclear whether only NDUFV2 gene exhibited the decreased expression level in bipolar disorder. Our study was to clarify the association of other nuclear-encoded complex I subunit genes and

BDNF belongs to the nerve growth factor family and it has been shown to play an important role in the growth, survival, differentiation and maintenance of several central nervous system neurons. Moreover, data derived from animal studies suggest that BDNF may be involved on the pathophysiology of mood disorders. Particularly, it has been shown that this neurotrophic factor could promote the function and growth of neurons belonging to the serotonergic system, which

50

Abstracts

alterations has been suggested to be on the basis of the mood disorders aetiology. The BDNF gene (chromosome 11p14) encodes for a precursor peptide (proBDNF) which is proteolytically cleaved to form the mature protein BDNF. Recently, Egan et al. (2003) have reported that the SNP 196 G/A (Val66 Met) located in the 50 pro-BDNF sequence may affect intracellular processing and secretion to the mature protein. We have analysed the Val66 Met polymorphism in a Spanish sample of patients with affective disorders in order to test the contribution of this genetic variable to the risk for these disorders. The sample consisted of 288 patients with major depression (DSM-IV) (MD), 74 bipolar patients (DSM-III-R) (BP) and 134 healthy controls, all of Spanish origin. Patients have been explored for several clinical features including age of onset, seasonality, melancholy (only MD), psychotic symptomatology or suicidal behaviour. The Val 66 Met genotype was determined using the Taqman 50 exonuclease assay. No overall differences in allele or genotype frequencies were found between patients with major depression, bipolar disorder or controls. However, we found that being a carrier of the Val/Val genotype seem to increase the risk of presenting psychotic symptomatology in a bipolar disorder episode (chisquare ¼ 9.28, P ¼ 0.005, OR ¼ 5.36 CI 95% [1.32-22.71]). Previous studies have shown the possible role of the Val66 allele in schizophrenia (Egan et al., 2003, Rosa et al., 2004). In this sense, the association that we found between psychotic symptomatology in bipolar disorder might support in terms of genetic background the hypothesis proposing that schizophrenia and bipolar disorder are on a continuum of liability. This work has been supported by a grant from the Ministerio de Ciencia y Tecnologı´a (SAF 2001-3400)

P2.40 COMBING DATASETS: GENOME-WIDE ANALYSIS OF 733 BP PEDIGREES McInnis MG,1 Zandi P,1 Edenberg H,2 Foroud T,2 Moe PR,2 Berrettini W,3 Beirut L,4 Gershon ES,5 Badner J,5 Scheftner W,6 Byerley W,7 Kelsoe J,8 MacKinnon DF,1 Potash JB,1 Willour VL,1 Coryell W,9 Rice J,4 Depaulo JR,1 and Nurnberger JL2 1 Johns Hopkins University, Baltimore, United States 2 Indiana University, Indianapolis, United States 3 University of Pennsylvania, Philadelphia, United States 4 Washington University, St. Louis, United States 5 University of Chicago, Chicago, United States 6 Rush University, Chicago, United States 7 University of California, Irvine, United States 8 University of California, San Diego, United States 9 University of Iowa, Iowa City, United States Genetic data from two large studies in familial BP disorder have been merged for a genome-wide nonparametric analysis. This included data from the Hopkins/Dana pedigree set and waves 1-4 from the NIMH Genetics Initiative for BP disorder. These pedigrees have been ascertained over a 15-year period at 10 sites in the US. The common ascertainment criterion was a BPI proband, the NIMH sample required an additional BPI sib, while the Hopkins sample required two additional 1st degree relatives with any major mood disorder. Data were merged using two standards, primarily the CEPH control subject typed across samples was used to standardize the sample, second the distribution of allele frequencies were compared. Markers with disparate allele frequency patterns were not transformed, rather suffixed and a map distance of 0.001 cM assigned. Three affection status models were analyzed, model 1 included BPI and schizo-affective manic (1,771 affected, 1,203 relative pairs), model 2 added BPII (2,078 affected, 1,753 relative pairs), and model 3 added recurrent major depression (2,448 affected, 2,491 relative pairs). Preliminary analyses have been performed using Genehunter 2.1, as it allows for the analysis of chromosomes with 50 or greater markers. There were 10 regions associated with a P < 0.01. These included 2p25 (D2S2952 NPL ¼ 2.3), 4q32 (D4S1629, NPL ¼ 2.2), 6q21 (D6S474, NPL ¼ 2.2), 7q32 (D7S559, NPL ¼ 2.0), 9p22 (D9S925, NPL ¼ 2.2), 16p13 (D16S748, NPL ¼ 2.8), 16q24 (D16S539, NPL ¼ 2.3), 17q25 (D17S785, NPL ¼ 2.7), 18q21 (D18S64, NPL ¼ 2.6), and 20p13 (D20482, NPL ¼ 2.5). Simulations are ongoing to determine the genome-wide significance of these findings. The strength of this analysis is the large number of pedigrees that have been combined and will ultimately allow for the planned and ongoing covariate analyses in the sample. The weaknesses in the sample include the fact that not all markers were typed in all pedigrees, genotype data from several methods have been included and the fact that the ascertainment criteria differ. The Hopkins data set and the initial data set (waves 1 & 2) of the NIMH were primarily multiplex

families, and the later (waves 3 & 4) focused primarily on ascertaining sibling pairs. Finally, these preliminary analyses were performed assuming homogeneity, which explains why the NPL scores are relatively low given the large sample size. P2.41 HAPLOTYPIC VARIATION AT THE 50 -FLANKING REGION OF THE HUMAN DOPAMINE TRANSPORTER GENE IN MANIC-DEPRESSION Sto¨ber G,1 Sprandel J,1 Schmidt F,1 Knapp M,2 and Jabs B1 1 Department of Psychiatry, University of Wu¨rzburg, Wu¨rzburg, Germany 2 IMBIE, University of Bonn, Bonn, Germany The human dopamine transporter (hDAT, SLC6A3) plays a central role in buffering of released dopamine. The SLC6A3 gene region has been implicated in bipolar disorder by linkage studies in multiplex pedigrees and by linkage disequilibrium studies on parent-offspring triads. In a previous study (Rubie et al., 2000) we had performed a systematic mutation screening of the DAT-promoter region with overlapping PCRfragments encompassed the entire promoter region and the non-coding exon 1 of the hDAT gene between 1586 basepair (bp) and þ97 bp relative to the transcription start site. Five diallelic polymorphisms were found, which were shown to be due to single base substitutions: 67T > A, 660G > C, 839C > T, 1169C > G, 1476T > G. In the present study, we analyzed genotype and haplotype frequencies in a case-control association study of 105 index cases with manic depression, and 199 controls. At the five polymorphic sites at the 50 -flanking region of the SLC6A3 (hDAT) gene, we observed no significant allelic or genotypic differences between index cases and controls. However, distinct 5-locus-genotypes accumulated in bipolar affective disorder compared to controls at P-value 0.029. Mainly, the 5-locus-genotype of 67TA—660GG—839CT—1169CG—1476TT was overrepresented in manic depression (41.9%) compared to controls (28.1%), indicating a potential role of the SLC6A3 locus in manic depression. P2.42 GENOME-WIDE SCAN OF THE BIPOLAR II PHENOTYPTE IN BIPOLAR DISORDER: FURTHER SUPPORT FOR BPII ON 18Q21 Nwulia EA, Zandi PP, and McInnis MG Johns Hopkins University, Baltimore, United States It has been suggested in earlier reports that linkage to chromosome 18q in bipolar disorder was primarily due to BPII (McMahon 2000). The BPII phenotype has not been examined genome-wide in these families. We have performed genome wide nonparametric analysis on 76 pedigrees using BPII affection model. This sample consists of the 65 Hopkins/DANA pedigrees and an additional 11 pedigrees, 9 of which had BPII exclusively, as the affected phenotype. These 11 pedigrees have not been previously reported or published. In the entire sample, there were 80 BPII sib-pairs, 9 cousins, and 56 avuncular relatives. Analysis was performed using the allele sharing model (ASM) of GENEHUNTER PLUS. We found a maximum peak NPL score of 3.14 between two markers, D18S814 and D18S1270 located on 18q21. A smaller peak with NPL of 2.07 occurred between D9S915 and D9S2157, located on 9q34. This finding strengthens previously reported evidence of linkage between BPII and 18q21-23 region, and suggests that 9q34 may harbor susceptibility genes of modest effect for BPII.

P2.43 SUGGESTIVE EVIDENCE OF LINKAGE TO CHROMOSOMES 1 AND 22 IN A FAMILY COSEGREGATING FOR BIPOLAR DISORDER AND AN AUTOSOMAL DOMINANT KIDNEY DISORDER Kimmel RJ,1 Kovacs I,1 Vrabel C,1 Wood B,2 Shekhtman T,4 Alexander M,4 Schalling M,3 and Kelsoe JR1 1 University of California San Diego, La Jolla, United States 2 St. Luke’s Hospital, Kansas City, United States 3 Karolinska Institutet and Karolinska Sjukhuset, Stockholm, Sweden 4 San Diego VA Healthcare System, La Jolla, United States We have previously described a large family where an autosomaldominant kidney disorder, Medullary Cystic Kidney Disease (MCKD), appears to cosegregate with Bipolar Affective Disorder—Type I (BADI). MCKD has been narrowed to the uromodulin gene on chromosome 16p (MCKD2), as well as to a 650 kb region of 1q23 (MCKD1). Such

Abstracts cosegregation is consistent with a possible chromosomal rearrangement that could facilitate mapping of susceptibility genes. In order to identify genomic regions for more detailed study, we have conducted a linkage genome scan. We performed a 9-cM screen on this family, using 440 markers. Multipoint, non-parametric analyses were used to determine linkage for bipolar disorder, a broader category of affective disorders, and for MCKD. MCKD linkage analysis on this family yields a long, flat, peak (P ¼ 0.05) on 1q. This peak encompasses the known MCKD1 region, as well as regions previously identified in schizophrenia and bipolar linkage studies. When analyzing a broad diagnostic category that includes all affective disorders and the hyperthymic phenotype, the data yields a peak (P ¼ 0.05) on 1q, as well as a peak (P ¼ 0.03) over a large region of 22q11. The 22q11 region has been previously identified in schizophrenia and bipolar studies, and includes the GRK3 gene. Analysis of BAD-I alone suggests (P ¼ 0.06) linkage to regions of chromosomes 1q and 22q, as well as weaker association to chromosomes not previously strongly linked to bipolar disorder. The analysis links this family’s MCKD to a region of 1q that is well documented in the renal literature. One interpretation of the data is that this family shows multiple susceptibility loci for affective disorders. One of the regions may be on 1q, segregating with MCKD. The addition of other susceptibility loci, suggested by the peak in 22q11, may produce a cumulative vulnerability that explains the high incidence of bipolar disorder in this family. P2.44 THE WELLCOME TRUST IRELAND/UK BIPOLAR SIBLING PAIR STUDY: STAGE II ANALYSIS Lambert DBJ,1 Hamshere ML,2 Segurado R,1 Bennett P,3 Jones I,2 McCandless F,3 Jones L,3 Raybould R,2 Heron J,3 Green E,2 Nokolov I,2 Corvin A,1 O’Mahony E,1 Owen M,2 Middle F,3 Kirov G,2 Holmans P,2 Craddock N,2 and Gill M1 1 Neuropsychiatric Genetics Laboratory, Departments of Psychiatry and Genetics, Trinity College, Dublin, Ireland 2 Neuropsychiatric Genetics Unit, Department of Psychological Medicine, University of Wales College of Medicine, Heath Park, Cardiff, United Kingdom 3 Division of Neuroscience, University of Birmingham, Queen Elizabeth Psychiatric Hospital, Birmingham, United Kingdom The presented genome scan, has been undertaken using a two-stage, grid tightening and sample splitting strategy, with a total sample of 395 affected sibling pairs (ASP). The Stage I sample for this genome scan has been described in Bennett et al. (2002). An additional sample of ASPs has been collected in the UK and Irish Republic for Stage II of our genome scanning strategy. The follow-up stage has completed genotyping of all 17 of the chromosomal regions that achieved the Stage I nominally significant linkage threshold maximum likelihood score (MLS) of 0.74 for autosomes, or 1.18 for the X chromosome. The average inter-marker centiMorgan distance in these follow-up regions is 4.4 cM. Non-parametric linkage analysis was conducted using MAPMAKER/SIBS, with allele frequencies derived from our observed data. Results show continued support for 10 of the original 17 regions implicated. The strongest evidence of linkage derived from the narrow (BPI only) diagnostic model was on chromosome 9p21-p12 (43 cM) with an MLS of 1.99 (57% IBD sharing), and chromosome 4q12-q21 (77 cM) with an MLS of 1.83 (56.4% IBD sharing). While chromosomes 10p1412 (MLS ¼ 1.65), 2q37-2qTel (MLS ¼ 1.48), 18q22 (MLS ¼ 1.44), 4q26q28 (MLS ¼ 1.43), 7q21 (MLS ¼ 1.37), 18q12 (MLS ¼ 1.36), 10p11.2 (MLS ¼ 1.28) and 7q11.2 (MLS ¼ 1.22) all show increases over their Stage I linkage statistics. The highest linkage score detected was under the intermediate diagnostic model on chromosome 2q37 (MLS ¼ 2.42) between markers D2S125 and the q-terminus. Under the broadest diagnostic model an MLS of 2.39 was detected on chromosome 4q12-q21 between microsatellite markers D4S398 and D4S392. P2.45 MOLECULAR GENETIC POSITIONAL AND CANDIDATE GENE STUDIES OF BIPOLAR AFFECTIVE PUERPERAL PSYCHOSIS Jones I,1 Hamshere M,1 Nangle JM,2 Middle F,3 Green E,1 Raybould R,1 Robertson E,3 Jones L,3 Heron J,3 Corvin A,2 Lambert D,2 Gill M,2 Owen M,1 and Craddock N1 1 Deparment of Psychological Medicine, UWCM, Cardiff, United Kingdom 2 Deparment of Psychiatry, Trinity College, Dublin, Ireland 3 Department of Psychiatry, University of Birmingham, Birmingham, United Kingdom

51

Women with bipolar disorder are at very high risk of suffering episodes of mania or psychosis in the immediate postpartum period (puerperal psychosis-PP). The available evidence suggests that in bipolar women, genetic factors influence vulnerability to puerperal episodes and, moreover, those women who have suffered such an episode, represent a more familial form of bipolar illness. For complex genetic disorders there are marked benefits in focusing on a more homogenous subtype and molecular genetic studies of bipolar affective puerperal psychosis provides such an opportunity as well as offering specific biological hypotheses to test. In our recruitment of families and individuals with Bipolar I Disorder we have obtained clinical information regarding the relationship of episodes of illness to childbirth allowing us to identify the subgroup of women who have suffered an episode of puerperal psychosis-over 200 women to date. We are using both positional and candidate gene approaches in the genetic investigation of PP. Positional: We have identified the subset of families in the Wellcome Trust UK-Irish bipolar sib-pair genome screen that include at least one female who has suffered an episode of puerperal psychosis. A maximum of 48 pedigrees contribute 71 affected sibling pairs to the sample. A genome scan was performed with MAPMAKER/SIBS, using allele frequencies estimated from the original Bipolar sample with SPLINK. Analysis of this more homogenous subgroup of families gives two LOD scores above 2-a maximum LOD score of 2.88 on 16p and of 2.03 on 2q. Candidate gene studies: We have undertaken studies at candidate genes of interest in Bipolar Disorder with an emphasis on those for which the expression is known to be influenced by steroid hormones. These include COMT, H-SERT, MAOA, BDNF, G72 and DAAO. We find evidence for linkage disequilibrium between PP and several alleles and haplotypes at G72 (eg. polymorphism M23: P < 0.001) with an effect size greater than that in our unselected Bipolar sample. There is also evidence for a linkage disequilibrium signal at DAAO. P2.46 A POSSIBLE SUSCEPTIBILITY LOCUS FOR BIPOLAR AFFECTIVE DISORDER IN CHROMOSOMAL REGION 13Q12-Q13 Abou Jamra A,1 Schumacher J,1 Ohlraun S,2 Fu¨rst R,3 Windemuth-Kieselbach C,3 Maier W,4 Wienker TF,3 Schulze TG,2 Strauch K,3 Propping P,1 Cichon S,5 Rietschel M,2 and No¨then MM5 1 Institute of Human Genetics, University of Bonn, Bonn, Germany 2 Central Institute of Mental Health, Division Genetic Epidemiology in Psychiatry, Mannheim, Germany 3 Institute for Medical Biometry, Informatics and Epidemiology, University of Bonn, Bonn, Germany 4 Department of Psychiatry, University of Bonn, Bonn, Germany 5 Life & Brain Center, University of Bonn, Bonn, Germany We have recently performed a complete genome scan for linkage in families with bipolar affective disorder (Cichon et al., 2001). One of the most promsing chromosomal regions was located on 13q12-q13. Using a dominant model of transmission, a 2-Point LOD-Score of 2,34 and HLOD-Score of 3,43 were observed at STR marker D13S175 (18 cM) (Cichon et al., 2001). Further support for chromosomal region 13q12q13 comes from two independent linkage studies (Badenop et al., 2002; McInnis et al., 2003). In order to narrow down the linked region and facilitate the finding of susceptibility genes, we performed fine mapping linkage analysis on chromosome 13q. In 66 German families with BPAD, 13 high polymorphic STR-markers were genotyped covering a region of 49 cM. Parametric and non-parametric affecteds-only linkage analyses were performed using three affection status models. Parametric MOD-Score analysis revealed a maximum MOD-Score of 4.5 at 22 cM using a narrow disease model and a dominant transmission model. Our data provide further support for the presence of a susceptibility gene in chromosomal region 13q12-q13. P3.1 FINE MAPPING OF A REGION ON CHROMOSOME 8P GIVES EVIDENCE FOR A QTL CONTRIBUTING TO INDIVIDUAL DIFFERENCES IN AN ANXIETY-RELATED PERSONALITY TRAIT, TPQ HARM AVOIDANCE Dina C,1 Zohar A,1 Gritsenko I,1 and Ebstein RP4 1 Institut de Biologie de Lille, Lille, France 2 Rupin Academic Center, Emek Hefer, Israel 3 Herzog Hospital, Jerusalem, Israel 4 Hebrew University, Jerusalem, Israel The chromosome 8p region is of interest in human behavioral genetics since it harbors a susceptibility region not only for schizophrenia but

52

Abstracts

also for anxiety-related personality traits such as harm avoidance and neuroticism. Towards verifying our preliminary linkage finding of a QTL for TPQ Harm Avoidance at chromosome 8p, we have now genotyped altogether 24 micro-satellite markers in 377 families. Using three methods (maximum likelihood binomial or MLB, MERLIN & an associated one parameter model) we observed significant results (P-values from 0.002 to 0.0004) for linkage to harm avoidance in this region. A peak multipoint lod score of 2.76 (P-value 0.0002) was obtained with the MLB method. The region-wide empirical P-value was 0.002 [0.001–0.0046]. Although the peak position varied somewhat according to the method (D8S1048 for MLB, D8S1463 for the two other methods), for three methods D8S1810 (60 cM) is within 1–2 cM of the peak for Harm Avoidance. This marker is of particular interest since it is proximate ( 0.1) since there were no control subjects with the ChAT AA and APOE e4 allele simultaneously. However, we could find a significant interactive effect of the ChAT AA and the APOE e4 for AD in the best model selected with highest posterior probability by SSVS. The highest ‘‘peak’’ of the posterior probability of the optimal subset was appeared under the choice of (sbk/t, c) ¼ (10, 500) with a value of 89.9% for the model with the interaction of the APOE e4 and the ChAT AA. The ChAT AA is a novel genetic risk factor for AD, and the SSVS is a useful approach for analyzing multiple candidate genes simultaneously.

P4.5 ALLELIC EXPRESSION OF APOE IN BRAIN Bray NJ, Jehu L, Moskvina V, Buckland PR, Williams J, Owen MJ, and O’Donovan MC Dept of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom Apolipoprotein E (ApoE) serves a central role in lipid metabolism and is the major apolipoprotein synthesized in brain. Polymorphisms in the coding sequence of the APOE gene are strongly associated with risk for late onset Alzheimer’s disease, with possession of the epsilon4 allele (encoding ApoE4) increasing risk in a dose-dependent manner, whereas the epsilon2 allele (encoding ApoE2) confers a protective effect. Variants in regulatory regions of the APOE gene might also contribute to disease susceptibility, and modulate the impact of structural changes in the ApoE protein, by altering its expression. In order to assess cis-acting influences on APOE expression in human brain, whilst controlling for trans-acting confounders, highly quantitative measures of allelic representation were applied to cortical RNA from individuals heterozygous for the epsilon4 and/or epsilon2 allele. A small, but significant (P ¼ 0.001), increase in expression of the epsilon4 allele was observed, relative to that of the epsilon3 and epsilon2 alleles. Similar differences in allelic expression were observed in AD subjects heterozygous for the epsilon 4 allele. The influence of known promoter variants was assessed by stratification of observed allele ratios on the basis of heterozygosity at these sites and calculation of diplotype probabilities. Preliminary results indicate a significant (P ¼ 0.008) association with the 219T/G polymorphism, with lower relative expression of haplotypes containing the 219T allele, consistent with in vitro data and in vivo measures of total expression. These results indicate that common cis-acting influences on APOE expression are modest, but detectable, in human cerebral cortex.

62

Abstracts

P4.6 POSSIBLE ROLE OF SEROTONIN TRANSPORTER GENE IN ALZHEIMER’S DISEASE AND OTHER DEMENTIAS: A PILOT STUDY Marino E, Lorenzi C, Artioli P, Landoni S, and De Ronchi D Department of Psychiatry, Alma Mater Studiorum, University of Bologna, Bologna, Italy Alzheimer’s Disease (AD) is a major cause of disability and mortality in the elderly and represents a relevant public health problem, considering also the increasing life duration in last decades. Genetics could play a relevant role in AD and in other dementias. Mutations in amyloid precursor (APP), presenilin-1 (PS-1), presenilin-2 (PS-2) and apolipoprotein E (ApoE) seem to be implicated in AD. Evidences suggested that also serotonergic system could be involved in dementias. The brain serotonin transporter (SERT), besides being the principal site of action of Selective Serotonin Reuptake Inhibitors (SSRI), was hypothesized to be involved in the onset of neurodegenerative mental diseases. Several studies tried to validate an association between the 44-bp insertion/ deletion (long/short) polymorphism within the promoter of the serotonin transporter gene (SERTPR) and AD, independently from APOE genotype. However, since the result was not univocal, we performed a pilot study to test the association between SERT polymorphism and AD or vascular dementia. The sample consisted of 34 dements (female/ male ¼ 24/10, age ¼ 66.56  4.83), 14 (40%) were affected by AD and 20 (60%) by vascular dementia. It was compared to 64 healthy controls, matched for sex and age (female/male ¼ 31/23, age ¼ 78.33  8.03). No significant difference between the SERTPR polymorphism genotype or allele distribution was found, when comparing demented patients and healthy controls, even when analysing the two diagnostic subgroups separately. We detected a higher frequency of the short/short genotype among elderly patients and controls (F ¼ 32.40, df ¼ 2, P < 0.0000000001). Studies in different populations underlined that the short variant may represent a risk factor for late onset AD, so we test the association between genotype and allele frequencies and the diagnosis of dementia in individuals older than 70. The small numbers did not allow us to detect any statistical difference, even if the *l allele resulted to be more frequently represented in healthy controls.

P4.7 IS THE SEROTONIN TRANSPORTER GENE A PREDICTOR FOR LONGEVITY? Lorenzi C, Marino E, Artioli P, Landoni S, and De Ronchi D Department of Psychiatry, Alma Mater Studiorum, University of Bologna, Bologna, Italy Serotonergic system plays a major role in many physiological functions (sleep, appetite, pain, sexual behaviour) and in a number of neuropsychiatric disorders (anxiety, mood disorders, schizophrenia) also associated with aged old (dementia and Alzheimer Disease). Nevertheless, the exact role of serotoninergic dysfunction in relation to old age was not completely cleared. Among genes involved in this neurotransmitter pathway, serotonin transporter (SERT) is the major determinant of serotonin (5HT) inactivation following release at synapses, it is also the site of action of Selective Serotonin Reuptake Inhibitors (SSRIs) antidepressants. Results from post mortem and in vivo imaging human studies, about age-related number of presynaptic serotonin transporters (SERT) are controversial. Nevertheless, recently, some researches, studying subjects without neuropsychiatric impairment, by means of high-resolution Single-Photon Emission Tomography (SPET) technique, identify a significant age-related decline in brainstem, SERT density, which was more pronounced than that previously reported in literature. These studies demonstrate reductions in the availability of central SERT binding sites related to age in living human subjects. In the present pilot research, we tried to investigate, a possible association between a SERT polymorphism and old age. This polymorphism (which consist of an 44-bp insertion/ deletion causing long/short alleles within the promoter of the SERT gene) could be relevant in age dependent SERT availability, since short allele seems to be related to a reduced level of gene and protein production of SERT. The sample consisted of 2408 subjects affected by: major depression (n ¼ 809), bipolar disorder (n ¼ 826), schizophrenia (n ¼ 258), obsessive compulsive disorder (n ¼ 19), dementia (n ¼ 14), otherwise specified psychoses (n ¼ 43), and control subjects (n ¼ 439). We analyzed SERTPR polymorphism in our sample on the whole, independently from diagnosis, but in relation to old age. We found out that s/s genotype is more frequent among elderly, particularly in subjects older than 70 (chi square ¼ 9.54 df ¼ 2 P ¼ 0.00846) or 75 (chi

square 12.12 df ¼ 2 P ¼ 0.00234). Though our sample was not specifically recruited for this aim, these results are intriguing, since they may suggest a relationship between SERTPR short variant and longevity.

P4.8 RESULTS OF AN ASSOCIATION STUDY BETWEEN PHOSPHOLIPASE A2 POLYMORPHISMS AND ALZHEIMER’S DISEASE Cordeiro Q, Bottino C, Almeida ACG, Noguti R, Menezes PR, Scazufca M, and Vallada H School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil Phospholipase A2 (PLA2) is a key enzyme in the metabolism of membrane phospholipids. PLA2 influences the processing and secretion of the amyloid precursor protein, which gives rise to the betaamyloid peptide, the major component of the amyloid plaque in Alzheimer’s disease (AD). Hence, we performed a study of association between two polymorphisms of PLA2 enzymatic system and AD comparing their frequencies in 59 AD patients, diagnosed according to the NINCDS-ADRDA criteria, and 106 healthy matched controls. There was no difference in the allelic (X2 ¼ 0.02, 1 d.f., P ¼ 0.87) neither in the genotypic (X2 ¼ 1.48, 2d.f., P ¼ 0.47) distributions of the AvrII polymorphism of the independent-calcium PLA2 (iPLA2) gene located at the 22q13.1. However, the BanI polymorphism of the cytosolic PLA2 (cPLA2) gene, located at the 1q25, presented a statistical difference in the genotypic distribution between the groups (X2 ¼ 8.96, 2d.f., P ¼ 0.01). The allele 2 (the BanI digested fragment) showed a trend to distribute more in the AD group (X2 ¼ 2.87, 1 d.f., P ¼ 0.09). These results suggest that the cPLA2 may play a role in the susceptibility to AD in our Brazilian sample.

P4.9 THE 22 C/T AND 2797 I/D POLYMORPHISM IN PRESENILIN 1 GENE ARE NOT CONNECTED WITH LATE-ONSET AND EARLY-ONSET FAMILIAL ALZHEIMER’S DISEASE Gacia M, Religa D, Safranow K, Maruszak A, Dziedziejko V, Zekanowski C, Chlubek D, Barcikowska M Polish Academy of Sciences Medical Research Center, Warsaw, Poland The 22c/t polymorphism in the promoter of the presenilin 1 gene (PSEN1) is associated with increased risk for Alzheimer‘s disease (AD) in some populations. It was shown that 22c allele is connected with two-fold decrease in promoter activity. We analyzed the impact of the polymorphism in groups of Polish late-onset, early-onset AD patients and control subjects. Our results suggest that 22 c/t polymorphism is not connected with AD in Polish population. The 22t allele showed a high degree of linkage disequillibrium with 2797 insertion of 13 bp. An additional 2897g/t polymorphism is also not connected with 22c/t one and is not a risk factor for AD.

P4.10 APOLIPOPROTEIN E E4 ALLELE INFLUENCES AGGRESSIVE BEHAVIOUR IN ALZHEIMER’S DISEASE Craig D, Hart DJ, Carson R, McIlroy SP, and Passmore AP Queen’s University, Belfast, United Kingdom The rising number of people with cognitive impairment is placing health care budgets under significant strain. Dementia-related behavioural change is a major independent risk factor for admission to expensive institutional care and aggressive symptoms in particular are poorly tolerated by carers and frequently precipitate the collapse of home coping strategies. Aggressive change may result from known genetic risk factors for Alzheimer’s disease (AD) and therefore accompany conventional markers such as apolipoprotein E (APOE) e4. We tested this hypothesis in 400 moderately to severely affected AD patients who were phenotyped for the presence of aggressive or agitated behaviour during the month prior to interview using the Neuropsychiatric Inventory (NPI-D). The proportion of subjects with aggression/agitation in the month prior to interview was 51.8%. A significantly higher frequency of the e4 allele was found in individuals recording aggression/agitation in the month prior to interview (w2 ¼ 6.69 df ¼ 2 P ¼ 0.03). The additional risk for aggression/agitation conferred by e4 was also noted when econtaining genotypes were compared against non-e4 containing genotypes (w2 ¼ 5.45 df ¼ 1

Abstracts P ¼ 0.02 OR ¼ 1.60 (1.06–2.43)). These results indicate that advanced Alzheimer’s disease patients are at greater risk of aggressive symptoms because of their APOE status. P4.11 FOLATE DEFICIENCY IMPAIRS WATER MAZE PERFORMANCE IN APP TRANSGENIC MICE Bernardo A,1 Galindo L,1 Da-Silva U,1 Jablonsk M,3,4 Zhao, J,2 and McDonald M1,2,4 1 Program in Neuroscience 2 Department of Pharmacology, Vanderbilt University, Nashville, TN, USA 3 Department of Ophthalmology, University of Tennessee, Memphis, TN, USA 4 Tennessee Mouse Genomics Consortium The ß-amyloid precursor protein (APP) over-expressing mouse (Tg2576) is a model of the ß-amyloid (Aß) aggregation and plaque formation of Alzheimer’s disease. Previous studies have shown that these mice do not exhibit Aß-related cell death, which is an important feature of Alzheimer’s disease. Recent studies have demonstrated that diets deficient in folic acid and containing increased levels of homocysteine can sensitize hippocampal neurons to Aß-induced toxicity. Three months of this experimental diet induces CA3 hippocampal cell death in APP transgenics but not in wild-type controls. Increased levels of homocysteine have also been proposed as a mechanism for cell death in Alzheimer’s disease. We were interested in whether this increase in toxicity will also increase the memory deficits of APP transgenic mice, and lead to greater ß-amyloid aggregation. Twenty animals (16–18 month old) were divided into two groups. One group was fed a diet that was deficient in folic acid with increased levels of homocysteine. The control group was fed a standard mouse diet with defined levels of folic acid and choline and lacking homocysteine. After 6 months on the experimental diets, APP transgenics under the folate-deficient diet exhibited deficits on a water maze memory task requiring diffuse, visuospatial recognition, however, they were unimpaired on a spatial working memory task requiring discrimination of discrete visual stimuli. Analyses of cell death and amyloid burden are ongoing. These results demonstrate that increased homocysteine in APP transgenic mice may induce cell death and memory deficits that parallel those observed in Alzheimer’s patients. Supported by NIA (1R01AG02243901), NIMH (5U01MH061971-04) and VUMC Discovery grant.

P4.12 GENETIC RISK FACTORS FOR ALZHEIMER’S DISEASE IN AFRICAN AMERICANS Murray MM and Clark RF Meharry Medical College, Nashville, TN, United States The incidence of Alzheimer’s disease (AD) varies considerably among ethnic groups. First-degree relatives of African Americans (AA) with AD have a higher cumulative risk of dementia than do those of whites with AD. Furthermore, the AD incident rates between AA is double that for a Nigerian population. There is a slightly higher AD incidence among AA compared to white Americans but it hasn’t been solely explained by the presence of more APOE-e4 in the AA population. The search for genes responsible for the increase in susceptibility for AD has resulted in only one consistently replicated marker, APOE. The remainder of AD risk genes have not proven to be reliable markers for risk of developing AD. APOE itself is not sufficient to predict an individual’s risk of developing AD. In fact, the association between APOE is questionable in the AA population. We hypothesize genetic interactions involving combinations of alleles from multiple loci are important in the development of AD. Furthermore, we theorize the loci involved in these interactions may also be associated with other cerebrovascular disease processes specifically hypertension (HTN). To test this hypothesis we selected 8 loci to perform a case/control association study of AD using 197 AA subjects. We choose as markers 4 loci reported to be associated with AD in at least one study (APOE, LRP, A2M, and PS) and 4 loci reported to be associated with HTN in at least one study (ACE, AGT, AGTR, and B2AR). The selection of these markers was based on the possible role each might play in AD etiology. Four of the loci in our study have been individually associated with risk for developing AD in AA. In addition four of the loci involved in our study had significant effects on the age of onset of AD in AA. Using MDR analysis, the identification of a highly predictive multilocus genotype associated with AD in AA was by far the most important finding of

63

our study. We believe this is the first report of such genetic interaction in AD. P4.14 DIFFERENTIAL FMRI ACTIVATION PATTERNS IN OFFSPRING OF LATE-ONSET AD CASES CORRESPOND TO PREVIOUSLY IDENTIFIED LINKAGE HETEROGENEITY ACCORDING TO PARENTAL AFFECTION STATUS Bassett SS,1 Kusevic I,1 Cristinzio C,1 Avramopoulos D,1 Fallin D,2 and Yousem DM1 1 Johns Hopkins University, School of Medicine, Baltimore, United States 2 Johns Hopkins University, Bloomberg School of Public Health, Baltimore, United States The purposeof this work was to examine brain activation patterns during memory encoding using fMRI in offspring of familial AD cases contrasting those from maternal families linked to chromosome 10 with those from paternal families unlinked to this region. Participants included female offspring of autopsy-confirmed familial AD cases, 9 from families with maternal transmission and 6 from families with paternal transmission. All were right handed, cognitively normal and ApoE e4 negative. Ages ranged from 50–75. They were administered a verbal word pair memory test both before and during fMRI scanning. The functional paradigm was a modified block design with six repetitions of intervals in which seven word pairs were presented for memorization, followed by presentation of the first word of the word pair with the subject told to internally speak the second word of the pair, followed by a rest period. The six-minute 10 sec. paradigm was repeated twice using two distinct word pair sets. Scanning was performed on a Phillips 1.5 T scanner. Mean scores for word pair recall for both groups obtained before scanning and during imaging were compared by means of 2 sample t-test. The fMRI data were processed using SPM99 with the SPM{t}maps transformed to Talairach space. The volumes of activation (k value) of both groups were compared by Welch modified 2 sample t-test for each location studied using an uncorrected height threshold (P < 0.01) and a minimum voxel cluster size of 20. Performance on word-pair recall, both before (Mat, 13  2, Pat, 11  4) and during scanning (Mat, 3.6  1.6, Pat, 2.8  1.8), did not differ. The fMRI results, when the samples were investigated separately, showed different patterns, with maternal having more extensive and bilateral temporal and frontal activation, whereas the paternal showed only unilateral activation (right frontal, left temporal). When the samples were subjected to a 2 sample, one-tailed t-test within statistical parametric mapping for regions of interest that showed greater volume of activation in the maternal than paternal sample, bilateral thalamus and temporal gyrus as well as right parahippocampal gyrus were found to be more activated in the maternal cases. Importantly, when the same areas of interest were analyzed for voxels where the paternal sample activation was greater than the maternal, increased activation was found only in the right frontal lobe within the paternal sample. These imaging results provide preliminary evidence of functional brain differences related to genetic linkage findings and suggest the possibility that activation patterns may prove useful as a quantitative trait to examine genetic heterogeneity in late-onset AD.

P4.16 THE ROLE OF HEPARIN-MODIFIYING GENES IN ALZHEIMER’S DISEASE Hazeel C,1 Lendon CL,2 Mann D,3 Turnbull J,4 and Talbot CJ1 1 University of Leicester, Leicester, United Kingdom 2 University of Birmingham, Birmingham, United Kingdom 3 University of Manchester, Manchester, United Kingdom 4 University of Liverpool, Liverpool, United Kingdom Heparan sulphate (HS) proteoglycans play an important but underrecognised role in Alzheimer’s disease. They are a major component of amyloid plaques where they colocalise with Ab and promote its aggregation. It has been shown that HS binds to the BACE1 enzyme and modifies the amount of Ab that it produces. This binding is crucially dependent on the type of modifications on the HS saccharide sidechains. HS that have been modified by N-desulfation and N-reacetylation have increased BACE1 binding activity. There are at least twenty known genes that add or modify the sugar chains of HSPGs. The four NDST (N-deacetylase/N-sulfotransferase) genes though are responsible for analagous in vivo modifications making them plausible biological candidate genes for AD. The NDST2 gene is on chromosome 10 between two regions that have been strongly linked with AD,

64

Abstracts

meaning it is also a positional candidate (VR22 68 Mb, NDST2 75 Mb, IDE 93 Mb). We have confirmed four SNPs in the gene by sequencing and established their degree of linkage disequilibrium in a panel of controls. These SNPs have been typed in a large cohort of early and lateonset AD patients and age-matched controls from the Birmingham and Manchester samples. This resource has the added advantage that 125 of the AD brains have neuropathological measurements including Ab load. Haplotypes derived from the SNPs have been used in an association study. P5.1 DUAL EFFECT OF POLYMORPHISMS IN NR2B AND SK3 GENES ON PREDISPOSITION TO ANOREXIA NERVOSA IN FAMILY TRIOS STUDY Koronyo-Hamaoui M,2 Frisch A,4 Stein D,3 Leor S,2 Apter A,2 Goldman B,1 Weizman A,5 and Gak E1 1 The Danek Gertner Institute of Human Genetics, Sheba Medical Center, Tel Hashomer, Israel 2 Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel 3 Eating Disorders Unit, Sheba Medical Center, Tel Hashomer, Israel 4 Felsenstein Medical Research Center, Belinson Campus, Petach Tikva, Israel 5 Research Unit, Geha Psychiatric Hospital, Petach Tikva, Israel Epidemiological and familial studies strongly support the notion that genetic factors play a role in aetiology of anorexia nervosa (AN). Heterogeneity of AN phenotypes and high incidence of co-morbidity with major depression, generalized anxiety and obsessive compulsive disorders, further suggest that AN may share common genetic factors with other psychiatric disorders. We investigated possible involvement of N-methyl-D-aspartate type glutamate receptor (NMDA-R) signalling pathway in anorexia nervosa (AN) on the basis of distribution of polymorphisms in candidate genes encoding its various components in 90 AN family trios and case-control series. We presently report an association of two meaningful polymorphisms to AN, a multiallelic coding CAG repeat in the small-conductance Ca2þ-activated Kþ (SK3) channel gene and a biallelic 5073T > G in the 30 UTR of the NR2B subunit of NMDA-R. Longer SK3 alleles, containing above the population modal of 19 CAG repeats, were significantly over-represented in AN patients, as was revealed in the case-control analysis (P < 0.001) and by the more robust HHRR (haplotype-haplotype-relative-risk) and TDT (transmission disequilibrium test) analyses on AN family trios (w2 ¼ 11.75 and P < 0.001). Further using TDT analysis in NR2B subunit, revealed that the 5073G allele was preferentially transmitted to AN (w2 ¼ 5.01, P ¼ 0.025). Composite analysis that includes both, NR2B and SK3, suggests that these factors independently and additively contribute to the risk of AN [OR ¼ 2.44 and OR ¼ 3.01, respectively for NR2B and SK3 genotypes and OR ¼ 6.8 for the combined NR2B/SK3 genotype that are prevalent in AN]. Our findings suggest possible consequential link between SK3 and NR2B, as two factors driving towards the same psychopathological outcome. As SK channels hyperpolarize membrane potential, they may impose inhibition of the NMDA-R. Hypofunction of the NMDA-R has been implicated in the aetiology of schizophrenia, however, has not been sufficiently substantiated in the context of eating disorders. Implication of NMDAR signalling pathway in AN, might be meaningful in order to elucidate its underlying cause and appropriate treatment.

P5.2 THE DRD4 GENE IS ASSOCIATED WITH OBSESSIVE COMPULSIVE DISORDER: A CONFIRMATION STUDY Camarena B,1 Aguilar A,2 Loyzaga C,3 Graff A,3 Weissbecker K,3 and Nicolini H3 1 Instituto Nacional de Psiquiatrı´a Ramo´n de la Fuente, Me´xico city, Mexico 2 Tulane University, New Orleans, United States 3 Mexico city University, Me´xico city, Mexico Pharmacological and neuroanatomical data suggests that dopaminergic system could be involved in the development of obsessivecompulsive disorder (OCD). We previously reported an increased prevalence of 7R variant of the DRD4-48 bp repeat polymorphism in twelve OCD patients with tics. Therefore, we decided to re-analyse this polymorphism in a sample of 211 OCD patients (with and without tics) compared with 187 healthy control subjects. A sub-sample of 91 families was analysed using a family-based association method. Eight different repeat alleles were detected in Mexican population (2,3,4,5,

6,7,8 and 10). Allele analysis of OCD patients and controls showed a significant association (w2 ¼ 29.13, df ¼ 7, P ¼ 0.0003). Also, analysis by gender between OCD and control group showed significant association between male groups (w2 ¼ 22.74, df ¼ 7, P ¼ 0.0024). In contrast, there was no statistical significance in OCD females compared with control females, after Bonferroni correction (w2 ¼ 13.13, df ¼ 6, P ¼ 0.0414). Finally, analysis between OCD with tics (n ¼ 38) and without tics (n ¼ 130) did not show significant association (w2 ¼ 5.9, df ¼ 6, P ¼ 0.575). Additionally, allele-wise TDT in the total sample showed a trend for significance in 67 heterozygous parents (w2 ¼ 14.05, df ¼ 7, P ¼ 0.0531). In conclusion, we found significant association between OCD and DRD4 gene. This association was obtained for the high frequency of long alleles in the OCD patients compared with the control group. Also, we observed a sexually dimorphic effect in OCD males. However, we did not replicate our findings of association between 7R allele and OCD patients with tics. Therefore it is possible that the differences reported in our previous study could be obtained for the excess of males in the OCD with tics group, and for the small size of the sample. Therefore, the findings reported in the present study support the hypothesis of gender differences in clinical and molecular data in OCD. Finally, it is important the replication of this study using a family-based association study in a larger sample of informative families. P5.3 GENETICALLY DETERMINED DOPAMINE LEVELS ARE ASSOCIATED WITH OBESITY Need AC,1 Ahmadi KR,1 Spector TD,2 and Goldstein DB1 1 University College London, London, United Kingdom 2 St Thomas’s Hospital, London, United Kingdom Obesity, generally defined as a body mass index (BMI) of 30 or above, is an important risk factor for a range of common diseases including coronary heart disease, type 2 diabetes, osteoarthritis, and some forms of cancer. Although the current epidemic of obesity is clearly driven by lifestyle factors, family and twin studies have shown that BMI has a strong genetic component. Human and animal studies have implicated dopamine (DA) in appetite regulation. Four genes are responsible for the major part of DA reuptake and metabolism-the DA metabolising genes: COMT, MAOA and MAOB, and the dopamine transporter DAT1. Each of these genes has a well-characterized functional variant influencing its activity or expression. Here we look at these four functional polymorphisms as a set in order to investigate how heritable variation in dopamine levels influences the risk of obesity. The COMT and DAT1 polymorphisms showed no association with either weight or BMI as quantitative traits, or with obesity in a case/control design. We found, however, that both MAOA and MAOB show an excess of the low-activity genotypes in obese individuals (MAOA: w2 ¼ 15.45, P ¼ 0.004, MAOB: w2 ¼ 8.05, P ¼ 0.018). Additionally, the MAOA genotype was significantly associated with both weight (P ¼ 0.0005) and BMI (P ¼ 0.001) in a multiple regression model. When considered together, the ‘at risk genotype’—low activity genotypes at both the MAOA and MAOB loci—shows a relative risk for obesity of 5.01. This is the first association study that directly implicates dopamine levels in human obesity.

P5.4 INTERACTIONS OF POLYMORPHISMS IN CORTICOTROPIN RELEASING HORMONE RECEPTOR 1 (CRHR1) AND VASOPRESSIN RECEPTOR 1B (V1B) GENES PREDISPOSE FOR PANIC DISORDER Keck ME, Kern N, Erhardt A, Ising M, Unschuld PG, Knorr CC, Poluda L, Thoeringer CK, Salyakina D, Pu¨tz B, Holsboer F, Mu¨ller-Myhsok B, and Binder E Max Planck Institute of Psychiatry, Munich, Germany Lifetime prevalence of anxiety disorders disorders is estimated to be as high as 25%. Twin and family studies have demonstrated that most anxiety disorders have a genetic component, but yet very few genetic factors are known. Accumulating evidence from animal studies suggests hypothalamic-pituitary-adrenocortical (HPA) system-related genes as promising candidates. Our aim was to study whether the susceptibility to develop panic disorder is related to polymorphisms in HPA system-related genes in patients suffering from an anxiety disorder associated with panic attacks (n ¼ 180) and healthy controls (n ¼ 380). We performed an association study candidate gene approach with 141 single nucleotide polymorphisms (SNPs) in 11 HPA system-

Abstracts related genes as genetic markers. All subjects were unrelated and ascertained from the anxiety disorders outpatient clinic. The patient sample included 180 caucasian subjects (males/females) from Germany who met DSM-IV criteria for panic disorder with or without agoraphobia, social phobia and specific phobia and were individually matched for ethnicity, age and sex to controls. We find significant evidence for an association between SNPs in CRHR1, most prominently SNP1 (P < 0.01) and in AVPR1B (most prominently SNP1, P < 0.01). In addition, a highly significant interaction was identified between genotypes at these SNPs with P < 0.0004. In summary, the genotypic association with SNPs in CRHR1 and AVPR1B demonstrates that both the corticotropin releasing hormone receptor 1 (CRHR1) and vasopressin receptor 1b (V1b) are involved in the genetic vulnerability for panic disorder. Moreover, we are able to show that these genes do not act independently but rather jointly on the phenotype. P5.5 ASSOCIATION OF A FUNCTIONAL POLYMORPHISM IN THE SEROTONIN TRANSPORTER GENE WITH NEUROTICISM IN A POLISH POPULATION Dragan WL and Oniszczenko W Department of Psychology, University of Warsaw, Warsaw, Poland A large number of studies have investigated an association between the polymorphism in the serotonin transporter-linked promoter region (5-HTTLPR) and personality trait of neuroticism. The results, however, have been inconsistent. We studied an association between the 5HTTLPR and neuroticism in female healthy subjects (n ¼ 200) of Polish origin. Personality trait of neuroticism was measured by the NEO-FFI. Results showed significant association between the polymorphism and neuroticism. Thus, our findings provide evidence for association of the 5-HTTLPR with neuroticism. P5.6 TNFA (308G/A) AND BDNF VAL66MET POLYMORPHISM IN ANOREXIA NERVOSA Slopien AS, Dmitrzak-Weglarz MDW, Rybakowski FR, Skibinska MS, Czerski PC, Hauser JH, and Rajewski AR Department of Child and Adolescent Psychiatry, Poznan, Poland Previous studies have shown that tumor necrosis factor (TNFa) gene polymorphism (Monteleone et al., 1999) and brain-derived neurotrophic factor (BDNF) Val66Met polymorphism (Ribas at al 2002) have been implicated in the pathogenesis of anorexia nervosa (AN). 91 female patients with ICD-10 and DSM-IV diagnosis of AN (mean age 18.22, SD ¼ 3.13) and 144 healthy female control subjects (mean age 33.97, SD ¼ 9.62) were included in the study. The frequency of genotypes and alleles were compared with chi2 test, with significance level P ¼ 0.05. A functional polymorphism G/A at 758 position of BDNF gene (Val66Met substitution at amino acid level) and polymorphism of TNFa gene (substitution, 308G/A, at promoter region) were analyzed by PCR-RFLP method. There were no statistically significant differences in the frequency of TNFa genotypes and alleles between AN patients and healthy controls (genotypes: P ¼ 0.084, alleles: P ¼ 0.076). We did not observed any difference in the distribution of alleles and genotypes of this polymorphism in patients with different subtypes of AN (genotypes: P ¼ 0.700, alleles: P ¼ 0.305). There were no statistically significant differences in the frequency of BDNF genotypes and alleles between AN patients and healthy controls (genotypes: P ¼ 0.855, alleles: P ¼ 0.895). We did not observed any difference in the distribution of alleles and genotypes of this polymorphism in patients with different subtypes of AN (genotypes: P ¼ 0.218, alleles: P ¼ 0.344). There was a statistical trend for the association between TNFa polymorphism and AN. We did not find association between risk of AN and studied polymorphism of BDNF gene. These results may support the association between TNFa polymorphism and AN. However, they require confirmation in a larger sample size. P5.7 AN ASSOCIATION STUDY OF THE DOPAMINE RECEPTOR (DRD1 AND DRD4) GENE IN ANOREXIA NERVOSA Dmitrzak-Weglarz MDW, Rybakowski FR, Slopien AS, Czerski PC, Hauser JH, and Rajewski AR Department of Child and Adolescent Psychiatry, Poznan, Poland Anorexia Nervosa is a disease of complex etiopathogenesis. Several lines of evidence suggest that a disturbance of dopamine neuronal

65

pathway may contribute to the pathogenesis of Anorexia Nervosa. We investigated polymorphisms of two genes: DRD1 mapped to 5q35.1 and DRD4 localized to region 11p15.5. The polymorphisms of this genes were described as: RLFP polymorphisms (48A/G) in the 50 untranslated (50 UTR) region of the DRD1 gene and functional polymorphisms (521 C/T) in the 50 flanking promoter fragment of the DRD4 gene. These polymorphisms may have potential functional importance because it occurs in a region regulating transcriptional activity. Okuyama et al. (2000) was shown that 521C allele is 40% more active then the allele 521T in an expression system. We investigated the association between these two polymorphisms and anorexia nervosa and we also performed an interaction analysis between them. We studied both polymorphisms in 91 females psychiatrically diagnosed with anorexia nervosa (both restricting and bulimic type) according to DSM—IV criteria and 93 control female subjects. The analysis of both polymorphisms of the (48A/G) DRD1 and (-521C/T) DRD4 genes was performed by PCRRFLP method. Informed consent from individuals participating in the study was obtained. For the DRD1 polymorphisms there were no statistically significant differences in the frequency of the alleles (P ¼ 0,408) and genotypes (P ¼ 0,373) between the group of patients and controls. For the DRD4 polymorphisms there were no statistically significant differences in the frequency of the alleles (P ¼ 0,371) and genotypes (P ¼ 0,362) between the group of patients comparing to the controls. For the interaction between DRD1 and DRD4 polymorphisms we did not observed statistically significant differences (P ¼ 0,737). We conclude that neither DRD1 nor DRD4 polymorphisms are associated with anorexia nervosa. The interaction between studied polymorphisms also did not increase the risk of anorexia nervosa. P5.8 APOE GENOTYPE AND PANIC DISORDER Martinez-Barrondo S, Saiz PA, Garcia-Portilla MP, Morales B, Coto E, Alvarez V, Bascaran MT, Bousono M, and Bobes J Department of Psychiatry, University of Oviedo, Oviedo, Spain To investigate the association of apolipoprotein E (ApoE) genotype and panic disorder (PD). We genotype 92 PD out-patients (DSM-IV criteria) and 174 healthy volunteers from Asturias (Northern Spain). The ApoE genotypes (e2, e3, e4 alleles) were determined after polymerase chain reaction (PCR) amplification, followed by digestion with the restriction enzyme CƒoI and electrophoresis on a 4% agarose gel. PD patients.Mean age (SD): 35,87 (12,38) years, 30,4% males (28 patients). Healthy volunteers.- Mean age (SD): 38,40 (8,94) years, 38,5% males (67subjects). ApoE genotypes (PD vs controls): e2e3: 14,1%, 12,1%, e2e4: 0%, 0,6%, e3e3: 65,2%, 75,3%, e3e4: 19,6%, 12,1%, e4e4: 1,1%, 0% (P ¼ 0,216). ApoE allele frequencies (PD vs controls): e2: 7,1%, 6,3%, e3: 82,1%, 87,4%, e4: 10,9%, 6,3% (P ¼ 0,161). An analysis according to gender only shows that differencies occur in males in genotype (P ¼ 0,021) and allele (P ¼ 0,044) frequencies. Genotype frequencies in males (PD vs controls): e2e3: 21,4%, 11,9%, e2e4: 0%, 1,5%, e3e3: 46,4%, 76,1%, e3e4: 32,1%, 10,4%, e4e4: 0%, 0% (P ¼ 0,021). Allele frequencies in males (PD vs controls): e2: 10,7%, 6,7%, e3: 73,2%, 87,3%, e4: 16,1%, 6,0% (P ¼ 0,044). Polymorphic variations in ApoE might influence susceptibility to PD in males. However, larger samples are needed to confirm or reject the current data.

P5.9 CONTRIBUTION OF HERITABILITY TO LEANNESS AND ADIPOSITY IN INBRED STRAINS OF MICE Gelegen EC,1 Collier DA,1 Kas MJ,2 and Campbell IC1 1 Institute of Psychiatry, London, United Kingdom 2 Rudolf Magnus Institute of Neuroscience, Utrecht, Netherlands There is evidence of a genetic component to the Eating Disorders (EDs), anorexia nervosa and bulimia nervosa and to obesity from twin and family studies. It is important identify genes that are involved in susceptibility to EDs and obesity as a better understanding of aetiology will facilitate improvement in treatment and prevention. Environmental factors such as psychosocial stress is also very important in the development of EDs and obesity, and they are best regarded as complex disorders in which genes and environment interact. There are probably several genes of small effect and consequently it is advantageous to examine the nature of weight regulation in a model organism such as the mouse. The first aim of this study was to screen the 50 and 30 regulatory and coding regions of genes involved in weight regulation for Single Nucleotide Polymorphism (SNP) and Simple Sequence Repeats in normal, lean and obese mice strains using denaturing HPLC and

66

Abstracts

DNA sequencing. We found a SNP (G/A) in the coding region of Melanocortin 4 Receptor (MC4R) gene at 73 bp position causing a change Gly ! Ser in C57Bl/6 and DBA/2 mice, which differ from each other in that DBA/2 is more resistant to the development of obesity in response to a high-fat diet than C57Bl/6J. Secondly, we found a SNP (G/ C) in the ghrelin gene, at 173 bp upstream of the transcription start site in C57Bl/6 and SWR mice, SWR being resistant to the development of dietary obesity. Thirdly, we found a SNP (C/G) in the first exon of the gene coding for Neuropeptide Y (NPY), at 59 bp position causing a change Pro ! Arg in C57Bl/6 and KK mice, which differ in body weight, C57Bl/6 being lean and KK being obese. Finally, we found a deletion in 30 UTR region of NPY gene, at 31 bp position in NZO mice, which develop severe obesity even when fed with a standard diet. As for simple repeats, we found a variation in the number of ‘‘GA’’ and ‘‘GT’’ repeat in 50 UTR region of the gene coding for Cholecystokinin A Receptor (CCKAR) in C57BL/6J and A/J, A/J being more resistant to diet-induced obesity than C57BL/6J and a variation in the number of ‘‘TG’’ repeat in 30 UTR region of the gene coding for NPY Receptor I in C57BL/6J, A/J and DBA mice. These results suggest that polymorphisms found in regulatory or coding regions of these genes may have an effect on the susceptibility to the development of EDs. We have also examined Chromosome Substitution Strains of mice generated from crossing C57Bl/6 and A/J strains to identify loci which confer genetic susceptibility to behaviours related to eating disorders and obesity. P5.10 POSSIBLE ASSOCIATION OF OXYTOCIN AND PROLACTIN POLYMORPHISMS WITH PANIC AND ANXIETY DISORDERS Erhardt A, Kern N, Knorr CC, Poluda L, Thoeringer CK, Unschuld G, Salyakina D, Pu¨tz B, Holsboer F, Binder EB, Mu¨ller-Myhsok B, and Keck ME Max-Planck Institute of Psychiatry, Munich, Germany Anxiety disorders are among the most common psychiatric disorders with a high number of hospital admissions and a life time prevalence up to 25%. Anxiety disorders and panic disorder are complex diseases proposed to be caused by a combination of psychological, physical stressors and genetic predisposition. Preclinical studies point to a series of candidate genes, including several neuropeptides, that are related to anxiety behavior in animals, which could also play an important pathophysiological role in patients. The nonapeptide oxytocin (OXT), for example, is one of the hypothalamic neuropeptides which shows anxiolytic-like effects when administered centrally or peripherally in rats Another important neuropeptide in modulating behavioral expression of anxiety in animals is prolactin (PRL). PRL is released from the anterior pituitary in response to exposure to different stressors and has anxiolytic effects after intracerebral infusion. Downregulation of the long form of the prolactin receptor on the other hand leads to an increase in anxiety-related behavior. In order to investigate possible case/control associations of the genes with anxiety disorder, we genotyped 53 SNPs in prolactin and oxytocin and their respective receptors (PRLR and OXTR) in 160 patients with syndromal panic attacks (including panic disorder, social phobia, specific phobia) and in 320 controls individually matched for ethnicity, age and sex. Genotyping was performed using the MassArray1 system (SequenomTM, CA). This system uses primer extension and matrix-assisted laser desorption/ionization time of flight mass spectrometry for genotype assignment. SNPs were selected from public (dbSNP) and private (Celera) databases. All patients underwent a thorough clinical and laboratory examination and the Structured Clinical Interview for DSM-IV (SKID l and II) and were only included when meeting DSM-IV criteria for panic disorder with or without agoraphobia, social phobia and specific phobia. We found an association of 2 SNPs in the OXTR (P ¼ 0.01 and P < 0.05) and of one SNP in the PRLR (P < 0.05). Haplotype analysis is currently performed to further investigate the role of the OXTR and PRLR variants in the genetic vulnerability for anxiety disorders.

P5.11 THE SHARED GENETICS OF TRAIT NEUROTICISM AND MOOD AND ANXIETY DISORDERS Hettema JM, Neale MC, Myers JM, Prescott CA, and Kendler KS Virginia Commonwealth University, Richmond, United States Neuroticism (N) has long been associated with symptoms of depression and anxiety, and some studies have begun to model the structure

underlying the relationship between neuroticism and internalizing disorders. Similarly, high rates of comorbidity have been consistently observed among the internalizing disorders. In addition, linkage and association studies have begun to identify putative susceptibility loci separately for N and these disorders, so a better understanding of their genetic relationship is indicated. In this study, we have used the twin design to examine the joint genetic and environmental risk factors of major depression (MD), generalized anxiety disorder (GAD), panic disorder (PD), agoraphobia (AG), and social phobia (SP) as indexed by the liability to N. We included N from the short form of the EPQ and DSM-III-R based psychiatric diagnoses from over 7,000 same- and opposite sex twin subjects from the population-based Virginia Adult Twin Study of Psychiatric and Substance Use Disorders (VATSPSUD) in multivariate structural equation models using common and traitspecific genetic and environmental factors. Analytic results found a moderate-to-high degree of shared genetic risk between N and all of the disorders studied, with genetic correlations between them estimated to be in the range 0.5–0.8. The environmental correlations were small-tomodest, in the range 0.1–0.3. No significant gender differences were observed. Thus, shared genetic risk factors between N and these disorders account for a substantial proportion of the phenotypic correlation seen between N and the individual disorders and as well as the high rates of comorbidity among the disorders themselves. These findings suggest a substantial overlap of genetic risk factors between N and the internalizing disorders, which may have important implications for gene-finding studies of these conditions. P5.12 BDNF AND MOOD DISORDERS: A NOVEL FUNCTIONAL PROMOTER POLYMORPHISM AND VAL66MET ARE ASSOCIATED WITH ANXIETY BUT HAVE OPPOSING EFFECTS Lipsky RH,1 Jiang X,2 Hobermann J,1 Waheed JF,1 Harris CR,1 Marini AM,2 Xu K,1 and Enoch MA1 1 NIAAA, NIH, Rockville, United States 2 USUHS, Bethesda, United States Because of its critical role in neuronal function and survival, the brainderived neurotrophic factor gene (bdnf) is a suitable candidate for vulnerability to addictions and other psychiatric disorders. We report a novel functional single nucleotide polymorphism (SNP), 281 in promoter 1 of human bdnf with decreased DNA binding in vitro and decreased basal reporter gene activity in transfected rat hippocampal neurons. Genotyping was performed across U.S. Caucasian, American Indian, and African American populations. The frequency of the 281A allele was 0.03 in the Caucasian sample but was virtually absent in the other populations. Single allele-based association studies were performed in this community-based Caucasian sample for both the novel promoter polymorphism and for the functional Val66Met polymorphism. Lifetime DSM-III-R psychiatric diagnoses were assigned and the Tridimensional Personality Questionnaire (TPQ) was administered to measure anxious temperament (harm avoidance [HA]) and novelty seeking (NS). Individuals with the low activity 281A allele (13 heterozygotes) had lower TPQ HA (F ¼ 4.8, P < 0.05). These heterozygous individuals were also less likely to have a psychiatric diagnosis (0.69 v 0.47, w2 ¼ 2.4, P ¼ 0.12). In contrast, the Met 66 allele was associated with increased HA (F ¼ 4.1, P ¼ 0.02) and was most abundant in individuals with both anxiety disorders and major depression (P < 0.05). Our results suggest that in this population, the low activity 281A allele may be protective against anxiety and psychiatric morbidity whereas Met 66 may be a risk allele.

P5.13 POSSIBLE ASSOCIATION OF ANGIOTENSIN 1 CONVERTING ENZYME POLYMORPHISMS WITH SYNDROMAL PANIC ATTACKS Kern N, Binder EB, Pfennig A, Erhardt A, Ising M, Lucae S, Salyakina D, Putz B, Holsboer F, Muller-Myhsok B, and Keck ME Max-Planck Institute of Psychiatry, Munich, Germany Panic disorder is a common anxiety condition with a lifetime prevalence of 4%. A genetic etiology has been suggested as panic disorder is often familial. Indeed family and twin findings indicate that genes contribute to both panic disorder and panic attacks as a symptom of other anxiety disorders i.e. social phobia and specific phobia. There is both preclinical and clinical evidence that neuropeptides such as cholecystokinin (CCK) and natriuretic peptides such as atrial natriuretic peptide (ANP) and C-

Abstracts type natriuretic peptide (CNP) are involved in the regulation of panic attacks. Generally, CCK has been shown to induce, whilst ANP is thought to attenuate panic attacks. Neuropeptides are substrates of exopeptidases, the most important of which is angiotensin I- converting enzyme (ACE). Dysregulation of ACE modulates the interaction of neuropeptides which could possibly cause vulnerability for panic attacks. Hence neuropeptides and their main exopeptidase ACE are considered as strong candidates. We focused on ACE, CCK, CCKreceptor A and B, natriuretic peptide precursor (NPP) A, NPPC as well as their receptors, NPR1 and NPR3, respectively. In order to investigate possible case/control associations we genotyped 70 SNPs in 8 candidate genes in 160 patients with syndromal panic attacks (including panic disorder, social phobia, specific phobia) and in 320 controls individually matched for ethnicity, age and sex. Genotyping was performed using the MassArray1 system (SequenomTM, CA). This system uses primer extension and matrix-assisted laser desorption/ ionization time of flight mass spectrometry for genotype assignment. SNPs were selected from public (dbSNP) and private (Celera) databases. All patients underwent a thorough clinical and laboratory examination and the Structured Clinical Interview for DSM-IV (SKID l and II) and were only included when meeting DSM-IV criteria for panic disorder with or without agoraphobia, social phobia and specific phobia. Preliminary data reveal a case/control association with a set of intronic SNPs in the ACE gene (P  0.01). There is no evidence, however, of association with the investigated neuropeptides and their receptors. Haplotype analysis is currently performed to further investigate the role of the ACE variant in the genetic vulnerability for panic attacks.

P5.14 POLYMORPHISMS IN THE PROMOTER OF THE NORADRENALIN TRANSPORTER (NAT) GENE AND PANIC DISORDER Lee YJ,1 Hohoff C,1 Sand PG,2 Kuhlenbaeumer G,1 Schirmacher A,1 Freitag C,5 Meyer J,2 Stoeber G,2 Franke P,3 Noethen MM,3 Fritze J,4 Fimmers R,3 Garritsen H,1 Stoegbauer F,1 and Deckert J1 1 Department of Psychiatrie, Institute of Molecular Psychiatrie, University of Muenster, Muenster, Germany 2 Department of Psychiatrie, University of Wuerzburg, Wuerzburg, Germany 3 Department of Psychiatrie, University of Bonn, Bonn, Germany 4 Department of Psychiatrie, University of Frankfurt, Frankfurt, Germany 5 Department of Medizinischeinformatik, University of Homburg, Homburg, Germany Panic disorder is characterised by recurrent, unexpected and sudden attacks of intense anxiety or fear. It is often accompanied by agoraphobia, an avoidance behaviour of public places or situations. Panic disorder patients exhibit increased plasma levels of adrenaline and noradrenaline and panic attacks are shown to be triggered by noradrenergic substances like yohimbine. Anxiolytic psychopharmaca like tricyclic antidepressants have an inhibitory effect on the noradrenaline transporter (NAT). Therefore a role for the NAT in the development of panic disorder is supposed. In the present study the relevance of four polymorphisms in the promoter and 50 -untranslated region of the NAT gene for the development of panic disorder was investigated. Altogether, 115 German patients with panic disorder (DSM-III-R and DSMIV, respectively) and 115 German gender- and age-matched healthy controls (anonymous blood donors) were genotyped by means of polymerase-chain-reaction based restriction-fragment-length-polymorphism analyses. Statistical analyses of allele and genotype frequencies (Chi Square and Armitage Trend) failed to show an association of the investigated polymorphisms with panic disorder in the total sample. An analysis of subgroups differentiated by the presence of agoraphobia revealed an association of three of the four investigated polymorphisms with panic disorder without agoraphobia (1070A/G: P < 0.01, 467G/C: P < 0.05, 872T/C: P < 0.01). This was particularly true for male patients (all three polymorphisms: P < 0.05). Our study, in line with a previous study investigating rare, coding exon polymorphisms (Sand et al., 2002, Neuroscience Letters 333, 41–44) thus failed to provide evidence for a major role of the NAT gene in the pathogenesis of panic disorder. As opposed to the rare coding polymorphisms the promoter polymorphisms though may play a role in a panic disorder subgroup without agoraphobia. They might be associated with decreased expression rates of NAT, increased catecholamine levels and consequently increased panic reactions. Replication

67

studies in independent panic disorder samples as well as a functional characterisation of the promoter polymorphisms are needed.

P5.15 POLYMORPHISMS OF DOPAMINE RECEPTOR GENES (DRD1 AND DRD4) AND TEMPERAMENT DIMENSIONS IN ANOREXIA NERVOSA Rybakowski F,1 Dmitrzak-Weglarz M,1 Slopien A,1 Hauser J,2 and Rajewski A1 1 Department of Child and Adolescent Psychiatry, University of Medical Sciences, Poznan, Poland 2 Department of Adult Psychiatry, Poznan, Poland Some evidence points to the role of temperament dimensions in vulnerability to Anorexia Nervosa (AN). These traits are in part genetically influenced and regulated by dopaminergic system. We investigated polymorphisms of two dopamine receptor genes: (48A/G) in DRD1 mapped to 5q35.1 and (521 C/T) in DRD4 localized to region 11p15.5. We assessed the association between these two polymorphisms and temperament dimensions of Temperament and Character Inventory (TCI) in anorexia nervosa patients and age-matched healthy controls. We recruited 75 females with anorexia nervosa (DSM-IV criteria) and 70 matched control subjects. The genotyping of both polymorphisms was performed by PCR-RFLP method. Informed consent from individuals participating in the study was obtained. In General Linear Model (GLM) analysis, genotypes were entered as independent variables, temperament dimensions as dependent variables, with age and ill/healthy status as covariates. Possible associations between genotype and temperament (from GLM analysis) were further explored with multiple linear regression. A statistical trend for the association of DRD1 genotype with harm-avoidance and DRD4 genotype with novelty seeking was observed. In linear regression, which included genotype, diagnosis and age as independent variables, DRD1 A/A and A/G genotypes were associated with lower levels of harm-avoidance than genotype G/G in combined group of patients and controls (odds ratios, respectively: 0.84 and 0.90). No significant differences were found for DRD4 genotypes and dimension of novelty seeking. The genotype of DRD1 receptor might be associated with personality dimension of harm avoidance in AN patients and controls.

P5.16 POLYMORPHISMS IN SEROTONIN RECEPTOR GENES HTR1B, HTR2A, HTR2C AND HTR4 ARE POSSIBLY ASSOCIATED WITH SPECIFIC PERSONALITY TRAITS IN PATIENTS SUFFERING FROM ANXIETY DISORDERS Unschuld PG, Ising M, Erhardt A, Kern N, Lucae S, Horstmann S, Salyakina D, Pu¨tz B, Binder EB, Mu¨ller-Myhsok B, Holsboer F, and Keck ME Max-Planck-Institut fu¨r Psychiatrie, Munich, Germany Serotonin (5-hydroxytryptamine, 5-HT) is a neurotransmitter which plays an eminent role in complex physiologic functions such as sleep, appetite, thermoregulation, pain perception, hormone secretion and sexual behaviour. It binds to different receptors, which can be divided in seven main groups, HTR1 to HTR7. Genetic studies could show an involvement of serotonin metabolism in human neuropsychiatric disorders and suggest a participation in etipathogenesis of anxiety, depression and schizophrenia. We analyzed a sample of 180 caucasian patients suffering from anxiety disorders and 380 healthy controls, matched for age, sex and ethnicity. All patients were not related and had been treated in our anxiety outpatient department for agoraphobia, social or specific phobia (diagnosed according to DSM-IV). Additionally each patient underwent a structured clinical interview for DSM-IV (SKID l and II) and tests for personality traits (TPQ, EPQ-RK, PSSI). An association study candidate gene approach was used with 102 single nucleotide polymorphisms (SNPs) within the genes coding for HTR1A, HTR1B, HTR2A, HTR2C, HTR3A, HTR4 and HTR5AGenotyping was performed by combining primer extension and matrixassisted laser desorption/ionization time of flight mass spectrometry (MALDI-TOF—MassArray1 system, SequenomTM, CA). We find significant evidence, that particular personal traits, such as psychoticism, novelty seeking, harm avoidance and reward dependence are associated with certain polymorphisms within the HT-receptor genes. Interestingly, this genetic association was not present in a population of 800 patients suffering from unipolar depression, suggesting an effect specific for anxiety-patients. The here presented preliminary data sug-

68

Abstracts

gests, that within this disease there are genetically distinct subgroups characterized by uniform psychometrically observable traits. P5.17 A GENETIC SYNDROME WITH CHILDHOOD MORBID OBESITY Agranat-Meged AN, Gandri Y, and Galili-Weisstub E Hadassah Medical Center, Jerusalem, Israel Purpose: 1) to present a genetic syndrome not previously described characterized by abnormal eating behaviour resulting in morbid obesity. 2) to highlight in the context of the syndrome, the interaction between a genetic predisposition and environmental factors in the development of an eating disorder. Following a suspected syndrome diagnosis in a morbid obese child, extended family members with a similar clinical picture were allocated. One suspected female child died. The remaining children: one female and five males underwent genetic, medical and psychological evaluations and their syndrome was characterized. The family’s coping with the eating habits of the child was evaluated. A newly described genetic syndrome was identified in four nuclear families of a multiple consanguineous family. The syndrome consists of dysmorphic features, orthopedic abnormalities and morbid obesity due to increased intake. Intelligence is within the normal range. An endocrine/metabolic etiology was excluded. Although the children had a prominent genetic tendency for increased intake, the environmental determinants influenced greatly their outcome. When provided with support and limit setting, they were able to adhere to a weight reducing balanced diet. To illustrate our findings one case will be presented in detail, a case of a child with extraordinary morbid obesity that was neglected by his family who could not cope with his eating behaviour. When treated, his life threatening medical complications, behavioural problems and severe developmental delay (secondary to neglect) all had a very favourable outcome. In conclusion: this newly defined syndrome sheds light on the interaction between genetic predisposition and environmental factors. Further study of the syndrome may enhance our knowledge on the biologic etiology of eating behaviours. P5.18 A FAMILY GENETIC STUDY OF CLINICAL SUBTYPES OF OCD Bhattacharyya S,1 Prasanna CLN,2 Reddy YCJ,2 Sheshadri S,2 and Khanna S2 1 Maudsley Hospital, London, United Kingdom 2 National Institute of Mental Health & Neurosciences (NIMHANS), Bangalore, India The results of various family studies of OCD have varied widely, possibly because of heterogeneity in the OCD phenotype. The current study was conducted to determine familiality of checker and washer subtypes of OCD compared to normal controls. In this study, all available first-degree relatives of 25 checker, 34 washer and 40 psychiatrically normal control probands were interviewed directly using structured interviews. The morbid risk of syndromal and subsyndromal OCD were significantly greater among the relatives of checker probands (19.4%) compared to the relatives of washer (8.7%) or control probands (5.4%), while the morbid risk for the relatives of washer and control probands were not significantly different. 67% of the checker relatives with syndromal and subsyndromal OCD were checkers, while 54% of the washer relatives with syndromal and subsyndromal OCD were washers. Our study provides preliminary evidence of familiality of checker subtype of OCD. Future, studies using patients sub-typed on the basis of symptom dimensions in larger OCD samples are needed, to further validate the genetic basis of OCD.

P5.19 IS THE GENETIC POOL OF OCD AND TIC DISORDER ABSENT IN INDIA? Bhattacharyya S,1 Reddy YCJ,2 Khanna S,2 and Prasanna CLN1 1 Maudsley Hospital, London, United Kingdom 2 National Institute of Mental Health & Neurosciences (NIMHANS), Bangalore, India Contrary to western reports, we observed the striking rarity of tic disorders comorbid with OCD, during our clinical work. We hypothesized that the genetic pool of OCD and tic disorders is absent in India.

We examined our hypothesis by first studying the prevalence of tic disorders among OCD patients and then also studying familiality of the two disorders in an enriched sample of OCD patients. We evaluated charts of 218 OCD patients seen in the OCD clinic between January to December 2001, using the OPCRIT criteria for ICD-10 for presence of co-morbid tic disorders. We also selected another sample of 59 consenting OCD probands based on whether they had predominant washing or checking compulsions, from those attending the OCD clinic at NIMHANS and, age and sex matched psychiatrically normal controls and interviewed all available first-degree relatives of checker, washer and psychiatrically normal control probands directly using structured interviews. Out of 218 OCD patients, 12 (5.5%) had comorbid tic disorder. None satisfied criteria for Tourette syndrome. Only 2 of the checker subtype of OCD probands, while none of the washer probands, had comorbid tic disorder. Only 3 of the checker relatives and none of the washer or control relatives had Tic disorder. Only one of the tic disorder positive OCD probands had a positive family history of tic disorder. The probable absence of genetic pool of tic-related OCD in our sample is an interesting aspect in the study of transcultural biology of mental illnesses.

P5.20 ASSOCIATION ANALYSIS OF THE SEROTONIN RECEPTOR TYPE 1D BETA (5HT1DB), NOREPINEPHRINE TRANSPORTER (NET) AND DOPAMINE TRANSPORTER (DAT) GENE POLYMORPHISMS WITH OBSESSIVECOMPULSIVE DISORDER Miguita K, Cordeiro Q, Destro HM, Shavitt RG, Miguel EC, and Vallada H School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil Family and twin studies have supported a genetic factor in the etiology of obsessive-compulsive disorder (OCD), although the precise mechanism of inheritance is still unclear. Clinical and pharmacological studies have shown the contribution of the serotonergic, noradrenergic and dopaminergic systems in the pathogenesis of this disorder. In this cross-sectional study, we have examined the allelic and genotypic frequencies of the G861C polymorphism of the serotonin receptor type 1D beta (5HT1DB), the A1287G variant in the norepinephrine transporter (NET) gene and the VNTR polymorphism of the dopamine transporter (DAT1) gene in 91 OCD patients and 317 healthy matched control subjects. There were no observed statistical differences in the frequencies of alleles and genotypes between patients and controls groups for the 5HT1DB (alleles: X2 ¼ 0.03, 1 d.f., P ¼ 0.85, genotypes: X2 ¼ 0.17, 2 d.f., P ¼ 0.91), for the NET (alleles:, X2 ¼ 0.38, 1 d.f., P ¼ 0.53, genotypes: X2 ¼ 0.51, 2 d.f., P ¼ 0.77), and for the DAT1 (alleles:, X2 ¼ 12.21, 6 d.f., P ¼ 0.057, genotypes, X2 ¼ 14.20, 10 d.f., P ¼ 0.16) gene polymorphisms. These results suggest that the investigated variants are not a genetic risk factor for the OCD in our Brazilian sample. However, further studies using larger samples and family based methods are recommended to confirm these findings.

P5.21 PSYCHOPATHOLOGICAL TRAITS IN PATIENTS WITH BULIMIA NERVOSA ARE ASSOCIATED TO SEROTONIN RECEPTOR 5HT2C AND SEROTONIN TRANSPORTER 5HTT ALLELES Ribase´s M,1 Mercader J,1 Grataco`s M,1 Badı´a A,2 Jime´nez L,2 Solano R,2 Vallejo J,2 Ferna´ndez-Aranda F,2 and Estivill X1 1 Genes and Disease Program, Center of Genomic Regulation, Barcelona, Spain 2 Department of Psychiatry, Hospital Princeps d’Espanya, Barcelona, Spain Several lines of investigation suggest that the serotoninergic system has an essential role in eating behaviour and weight regulation. Moreover, alterations in this neurotransmitter system could participate in the aetiology of eating disorders (ED) and in some associated psychopathological traits. To test this hypothesis, we analyzed the potential involvement of the 5HT2C and 5HTT genes in different psychiatric symptoms in 86 patients with ED. Forty-six patients with anorexia nervosa (AN) and 36 with bulimia nervosa (BN) completed the Symptom Checklist 90-revised (SCL90-R) questionnaire and were genotyped for the Cys23Ser, 995G/A, 759C/T and the 697G/C SNPs within 5HT2C and the 5HTTLPR SNP in the 5HTT gene. No

Abstracts effect of the Cys23Ser SNP was detected, however, those bulimic patients carrying the 995A/759T/967C haplotype showed increased symptomatology for seven of the nine subescales of the SCL90R, that include somatisation (P ¼ 0.029), obsessive-compulsiveness (P ¼ 0.021), depression (P ¼ 0.032), anxiety (P ¼ 0.004), hostility (P ¼ 0.028), phobic anxiety (P ¼ 0.029) and paranoid ideation (P ¼ 0.008). Moreover, although not epistatic effects were observed, we detected that increased mean scores in the anxiety subescale were significantly associated to the 5HTTLPR-S allele in the BN group (P ¼ 0.004). We did not observe a significant effect of the 5HT2C and 5HTT genes in any of the psychopathological traits in the AN patients. These results suggest that 5HT2C and 5HTT could participate in the physiopathology of BN not only through its direct effect on eating behaviour but also in the psychiatric symptoms related to the anxiety and depressive traits associated to ED.

P5.22 INCREASED BLOOD LEVELS OF BDNF IN PATIENTS WITH EATING DISORDERS IS INDEPENDENT FROM THE BDNF VALMET66 GENOTYPE BUT IS ASSOCIATED TO THE 278C ALLELE Ribase´s M, Mercader J, Grataco`s M, Badı´a A, Jime´nez L, Solano R, Vallejo J, Ferna´ndez-Aranda F, and Estivill X Barcelona, Spain Anorexia nervosa (AN) and bulimia nervosa (BN) are complex psychiatric conditions where genetic and environmental factors are involved. Murine models and association studies in patients with eating disorders (ED) suggest a potential role of brain-derived neurotrophic factor (BDNF) in eating behaviour. We have reported a consistent association of BDNF single nucleotide polymorphisms (SNPs) and ED. The 278C/T BDNF SNP is associated to BN, and the Val66Met variant to both AN and BN in six different European populations. We present here the analysis of BDNF levels in blood of patients with AN and BN that have been genotyped for BDNF SNPs. We assessed the BDNF blood concentration in a clinical sample of 118 Spanish Caucasoid ED patients (54 AN and 64 BN) and 68 sex-matched unrelated controls. BDNF levels were significantly higher in both AN (46.7  25.9 ng/ml) and BN (41.0  23.4 ng/ml) patients than in control subjects (16.4  5.7 ng/ml, P < 0.0005). We did not find a significant effect of the Met66 functional variant on BDNF levels, as both Met66 carriers and non-carriers showed increased BDNF concentrations. In contrast, mean BDNF levels were significantly higher in ED patients homozygous for the 278C variant (49  24.2, P ¼ 0.012) than those carrying the 278T allele (26.9  17.245), which showed BDNF levels similar to controls. The altered BDNF plasma levels observed in this study provide physiological evidence of a possible role of this neurotrophic factor in AN and BN, and strongly argues for its involvement in eating behaviour and body weight regulation.

P5.23 ASSOCIATION TO MARKERS NEAR THE SLC6A4 GENE IN PANIC DISORDER SAMPLES FROM THE GENETIC ISOLATE OF FAROE ISLANDS Dahl HA,1 Wang AG,2 Als TD,3 Lauridsen MB,3 Ewald H,3 Mors O,3 and Kruse TA1 1 Odense University Hospital, Odense, Denmark 2 Copenhagen University Hospital, Copenhagen, Denmark 3 Psychiatric Hospital in Aarhus, Risskov, Aarhus, Denmark 13 patients, meeting the ICD-10 and DSM-IV criteria for panic disorder with agoraphobia, from a sub-isolate of the Faroe Islands are used in a genome-wide search for candidate loci. In the search for possible shared segments inherited identical by descent, the patient haplotypes were reconstructed from the parental genotypes and compared to haplotypes from 43 healthy individuals from the same region on the Faroe Islands. Although sampled independently, a subsequent genealogical assessment revealed a common ancestor 12 generations back for all 13 patients, indicating possible founder effect. Preliminary analysis of data from the ongoing genome scan revealed association to a two-locus segment D17S1293-D17S1842 on chromosome 17q12 (P ¼ 0.006). A closer look at the region showed LD extending both proximal and distal to the two-loci, indicating probable IBD inherited regions. Interestingly, proximal and right next to the two-loci, is the SLC6A4 gene situated on 17q11.1. The SLC6A4 gene, encoding the serotonin tran-

69

sporter (SERT), the primary target for SSRI’s, makes this gene an obvious candidate gene for further studies.

P5.24 A FAMILY-BASED ASSOCIATION STUDY OF NOVEL SEROTONIN POLYMORPHISMS IN OCD AND OCD SYMPTOM SUBGROUPS Arnold PD,1 Summerfeldt LJ,2 Sicard T,1 Geronimo J,1 Trakalo J,1 Zai G,1 Hu X,3 Goldman D,3 Hanna GL,4 Pato M,5 Richter MA,1 and Kennedy JL1 1 Centre for Addiction and Mental Health, Toronto, Canada 2 Trent University, Peterborough, Canada 3 NIAAA/NIH, Rockville, United States 4 University of Michigan, Ann Arbour, United States 5 VA Medical Center, Washington, United States Investigations of the serotonin 1B (5HT1B) and serotonin transporter (5HTT) genes in obsessive-compulsive disorder (OCD) have produced variable findings. Interpretation of these findings is limited as most studies have only reported: 1) analysis of a single informative polymorphism per gene, and 2) examination of OCD as a unitary phenotype, although factor analyses have repeatedly demonstrated distinct symptom factors that likely have distinct genetic correlates. In particular, it has been suggested that symmetry/ordering obsessions and compulsions may be more highly heritable than other symptom subtypes, such as contamination symptoms, aggressive/sexual symptoms and hoarding (Alsobrook et al., 1999). We therefore set out to examine novel candidate gene polymorphisms in 5HT1B and 5HTTLPR for their association with OCD and with symmetry/ordering symptoms. Four single nucleotide polymorphisms in 5HT1B were genotyped in 271 nuclear families: the commonly studied 861G/C, a promoter polymorphism shown to affect transcriptional activity (161A/T, Duan et al., 2003), and two additional SNP’s located at the 50 (rs1213371) and 30 (rs2000292) ends of the gene. In 180 of these families, we genotyped an insertion/deletion polymorphism (5HTTLPR) in the promoter region of 5HTT, in addition to a novel A/ G SNP contained within the 5HTTLPR long allele that has been shown to alter transcriptional activity (Goldman et al., 2004). Subjects were classified as having the symmetry/ordering phenotype based on target symptoms endorsed on the Yale-Brown Obsessive-Compulsive Scale (YBOCS). The analysis for genetic association was conducted using the Family Based Association Test (FBAT) and included both single locus and haplotype analyses. For the phenotype of OCD diagnosis, no statistically significant associations were identified with either 5HT1B or 5HTT, although there was a trend towards association with increased transmission of the 861G allele (P ¼ 0.07). Secondary analyses based on quantitative YBOCS severity score, sex, and age of onset were negative for all polymorphisms tested. For symmetry/ordering symptoms, however, there were significant associations with increased transmission of the 861G allele (P ¼ 0.03) under the additive model of inheritance, and increased transmission of the 161T allele (P ¼ 0.04) under the recessive model. No statistically significant haplotype associations were identified. These results were not corrected for multiple comparisons. Overall, these findings provide preliminary evidence that the phenotype of symmetry/ordering may be associated with 5HT1B and that future candidate gene studies may benefit from examining symptom subtypes in OCD. Future research, currently underway in our laboratory, will included genotyping of additional polymorphisms in 5HT1B.

P5.25 ASSOCIATION STUDY OF CORTICOTROPIN RELEASING HORMONE RECEPTOR 2 (CRHR2) IN PANIC DISORDER Tharmalingam S,1 King N,1 De Luca V,1 Rothe C,3 Koszycki D,2 Bradwejn J,2 Macciardi F,1 and Kennedy JL1 1 Neurogenetics Section, Centre for Addiction and Mental Health and Department of Psychiatry, University of Toronto, Toronto, Canada 2 Department of Psychiatry, University of Ottawa and the University of Ottawa Institute of Mental Health Research, Royal Ottawa Hospital, Ottawa, Canada 3 Department of Psychiatry, University of Munster, Munster, Germany Panic disorder is classified as an anxiety disorder and affects 1–3% of the population. An individual suffering from such a disorder may

70

Abstracts

experience unexpected recurrent panic attacks and fear of future attacks. Twin and family studies have pointed towards a strong heritability of the disorder. Stress response and anxiety are thought to be mediated, at least in part, by the corticotrophin-releasing hormone (CRH), which is known to be a regulator of the hypothalamic-pituitaryadrenal (HPA) pathway (Coste et al., 2000). To search for markers conferring genetic susceptibility to panic disorder, we genotyped three polymorphisms of the CRHR2 gene, CRHR2(CA)n, CRHR2(GT)n, and CRHR2(GAT)n, in 466 individuals, 108 of whom had DSM-IV panic disorder. There were 75 case-control pairs and 101 triad families. Casecontrol association analyses, using chi-square tests, yielded no difference in the distribution of the alleles. Linkage analysis using the Transmission Disequilibrium Test (TDT) showed no preferential transmission of alleles for any of the three markers. Haplotype analysis indicated that allele 7 of CRHR2 (GAT)n and 8 of CRHR2 (GT)n are in almost complete linkage disequilibrium (P ¼ 0.0000001). Although both neurobiology and chromosomal location point to the CRHR2 gene as a candidate for panic disorder, our study indicates that the CRHR2 polymorphisms examined do not confer susceptibility to panic disorder. Further studies investigating additional polymorphisms in this gene and other components of the CRH signalling system may prove useful.

P5.26 INTERACTION OF GENES OF THE CATECHOLAMINERGIC AND THE SEROTONERGIC SYSTEM IN PANIC DISORDER Freitag CM,1 Rothe C,2 Domschke K,2 Gutknecht L,3 Franke P,4 Fritze J,5 Kuhlenbaeumer G,6 Schirmacher A,6 Nyhuis P,2 Noethen MM,9 Garittsen H,7 Fimmers R,8 Stoegbauer F,6 Lesch KP,3 and Deckert J2 1 Department of Child and Adolescent Psychiatry, University Hospital Saarland, Homburg, Germany 2 Department of Psychiatry, University Hospital Muenster, Muenster, Germany 3 Department of Psychiatry, University Hospital Wuerzburg, Wuerzburg, Germany 4 Department of Psychiatry, University Hospital Bonn, Bonn, Germany 5 Department of Psychiatry, University Hospital Frankfurt, Frankfurt, Germany 6 Department of Neurology, University Hospital Muenster, Muenster, Germany 7 Insitute of Transfusion Medicine, University Hospital Muenster, Muenster, Germany 8 Insitute of Medical Biometry, University Bonn, Bonn, Germany 9 Department of Medical Genetics, University Antwerp, Antwerp, Belgium Panic disorder is an anxiety disorder with an estimated heritability of up to 48%. We have recently confirmed that the disorder is associated with the more active allele of the 472G/A ¼ V158M polymorphism in the catechol-O-methyltransferase (COMT) gene. In addition, we had investigated the role of a 30-bp insertion-deletion polymorphism in the monoamine oxidase A gene promoter (MAOA-LPR) and a 1019C > G promotor polymorphism in the serotonin receptor 1 A (5-HT1A) gene. Here, we report an analysis of the interaction between these polymorphisms. 113 cases and 117 age- and sex-matched controls had been genotyped for all three polymorphisms according to the published protocols. We first assessed the best model of the four possible (genotypes-categorical, multiplicative, dominant and recessive model for the risk alleles) for the association of each single polymorphism with the disorder. We then calculated the interaction by a logistic regression procedure. For the COMT 472G/A ¼ V158M polymorphism, a multiplicative model showed the best fit to the data. For MAOA-LPR (long alleles) and 5-HT1A 1019C > G, a recessive model was chosen. We observed a nominally significant (P ¼ 0.040) interaction between the COMT and the 5-HT1A polymorphisms for the total panic disorder sample (n ¼ 113). Surprisingly, the interaction of the risk alleles was consistent with a protective effect. This effect was not observed for the subgroup of patients with agoraphobia (n ¼ 81). For the MAOA-LPR polymorphism we failed to find an interaction with either the 5-HT1A or the COMT polymorphism. In addition to the previously reported results, we found evidence for an interaction of genes of the catecholaminergic and the serotonergic system in panic disorder. The tentative protective effect of the interaction of the two risk alleles of the COMT and 5-HT1A polymorphisms in the total panic disorder

sample needs replication in a bigger sample with more power to elucidate the effect in different subgroups of panic disorder. The MAOLPR—seems to be a risk factor for panic disorder independent from the 5-HT1A- and COMT-polymorphisms. P5.27 ASSOCIATION OF THE V158M CATECHOL O-METHYLTRANSFERASE GENETIC POLYMORPHISM WITH PANIC DISORDER Rothe C,1 King N,1 Koszycki D,3 Bradwejn J,3 Deluca V,1 Tharmalingam S,1 Macciardi F,1 Deckert J,2 and Kennedy JL1 1 Clarke Division, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada 2 Department of Psychiatry, University of Mu¨nster, Mu¨nster, Germany 3 Department of Psychiatry, University of Ottawa and the University of Ottawa Institute of Mental Health Research, Royal Ottawa Hospital, Ottawa, Canada Genetic as well as clinical data suggest that Catechol O-methyltransferase (COMT) is involved in multiple complex psychiatric conditions. Recently, the valine allele of the Val158Met COMT polymorphism has been reported to be associated with panic disorder in two studies.To evaluate the impact of the Val158Met COMT polymorphism in panic disorder we genotyped the single nucleotide polymorphism in the coding region of COMT in two independent Canadian samples, 79 triad families (panic disorder patient and parents), and 89 cases with matched controls. In the triads, significant transmission disequilibrium for the valine allele was observed between the alleles of the Val158Met COMT polymorphism and panic disorder (P ¼ 0.002). A significant excess of the valine allele was found in the case-control analysis (P ¼ 0.00019). In addition to these findings, data of two more single nucleotide polymorphisms (rs 737865, rs 165599) will be presented. These results support the hypothesis that the valine allele of the Val158Met COMT polymorphism or a nearby locus is involved in the aetiopathogenesis of panic disorder.

P5.28 OBSESSIVE-COMPULSIVE FACTORS IN SIBLING PAIRS DIAGNOSED WITH OBSESSIVE-COMPULSIVE DISORDER: A PILOT STUDY Chacon P,1 Rosario-Campos MC,1 Hounie AG,1 Ackerman F,1 Shimabokuro F,1 de Mathis MA,1 Curi M,1 Pauls DL,2 and Miguel EC1 1 University of Sa˜o Paulo, Sa˜o Paulo, Brazil 2 Harvard, Boston, United States Obsessive-Compulsive Disorder (OCD) is a heterogeneous disorder. Many studies have tried to identify more homogeneous subgroups. Factor analyses of obsessive-compulsive symptoms (OCS) have consistently found from 3 to 5 factors, which seem to be helpful in sub-grouping OCD. Leckman et al. (1997) presented a solution of 4 factors accounting for more than 60% of the variance. Factor 1 consists in aggressive, sexual, and religious obsessions and checking compulsions. Factor 2 is characterized by symmetry and ordering obsessions and compulsions. Factor 3 is characterized by contamination obsessions and cleaning compulsions. And finally, Factor 4 includes hoarding obsessions and compulsions. The main goal of this study was to assess whether siblings affected with OCD resemble one another in each of the four factors cited above. Forty OCD siblings (from 18 families) were recruited and had their symptoms evaluated by the Yale Brown Obsessive-Compulsive Scale (Y-BOCS). Diagnoses were made according to DSM-IV. Data of all 40 siblings were used in a principal components factor analysis with Varimax Rotation to check if we would find a similar solution to Leckman et al. (1997). Once a similar solution was confirmed, factors scores were then generated using an algorithm derived from Leckman’s study. Correlation analyses were performed for all 4 factors. Considering that some families had more than 2 affected siblings, the analyses were repeated considering: 1) the eldest siblings (ES)-18 pairs, 2) the youngest siblings (YS)-18 pairs, and 3) all possible sib pair (AS)-28 pairs. There were no correlations in any of the 4 factors, except for factors 3 and 4 when we selected sibling pairs according to concordance of gender, using all different analyses mentioned above (ES, YS and AS). Significant correlation for Factor 3 (contamination obsessions and cleaning compulsions) was found in those sib pairs concordant for the male gender (ESP ¼ 0,065, YSP ¼ 0,055 and ASP ¼ 0,007). Factor 4 (hoarding obsessions and compulsions) was significantly correlated in those sib pairs concordant for the female gender (ES P ¼ 0,021, YS P ¼ 0,011 and AS P ¼ 0,002).

Abstracts These preliminary results suggest that the phenotype of OCD may be partially influenced by the gender of the affected relative. These findings need to be replicated in larger samples. P5.29 OBSESSIVE-COMPULSIVE SPECTRUM DISORDERS AND RHEUMATIC FEVER: A FAMILY STUDY Hounie AG,1 Chacon P,1 Mercadante MT,1 Rosario-Campos MC,1 Shavitt RG,1 De Mathis MA,1 Alvarenga PG,1 Curi M,1 Pauls DL,2 and Miguel EC1 1 University of Sa˜o Paulo, Sa˜o Paulo, Brazil 2 Harvard, Boston, United States Studies support the notion that rheumatic fever (RF), an autoimmune disease triggered by a streptococcus infection, is associated with obsessive-compulsive disorder (OCD). This is a family study that assessed the frequency of OCD and obsessive-compulsive spectrum disorders (OCSD) in first-degree relatives (FDR) of RF probands (with and without SC) to verify if there was aggregation of those disorders in RF families. We assessed 98 probands and their 381 FDR. Of the 98 probands, 31 had RF without SC, 28 had RF with SC and 39 were controls without RF recruited in an orthopedic clinic. Trained and blinded interviewers assessed subjects with semi-structured interviews (KSADS and SCID) and best estimate diagnosis. Comparisons between frequencies of disorders were done with Fisher and chi-square tests. Kaplan-Meier survival analysis was used to calculate agecorrected morbid risks. Generalized estimated equations (GEE) were used to compare frequencies of OCSD. OCSD combined were more frequent in FDR of RF probands (with or without SC) than in FDR of controls (P ¼ 0.03). Chronic tic disorder (CTD) was more frequent in FDR of SC probands than FDR of controls (P ¼ 0.05). Generalized anxiety disorder (GAD) was more frequent in FDR of RF probands than in FDR of controls (P ¼ 0.008). Controlling for the presence of an OCSD in the probands, GAD (x2 ¼ 0.004), tic disorders combined (GEE, P ¼ 0.01), body dysmorphic disorder (BDD) (GEE, P ¼ 0.02) and OCSD combined (GEE, P ¼ 0.02) were more frequent in FDR of RF probands with an OCSD than in FDR of controls. OCSD were more frequent (nonsignificantly, P ¼ 0.09) in FDR of RF probands without an OCSD than in FDR of controls. RF in the probands increased the risk for family members presenting an OCSD. This risk was even higher if the proband had RF plus an OCSD. It is possible that a subgroup of RF is associated to OCSD. RF may be a risk factor for OCSD.

P5.30 MUTATION SCREENING OF NRCAM AND IMMP2L, TWO CANDIDATE GENES FOR TOURETTE’S SYNDROME AND OBSESSIVE COMPULSIVE DISORDER ON CHROMOSOME 7Q31 Delorme R,1 Leboyer M,2 and Bourgeron T3 1 Institut Pasteur, Paris, France 2 INSERM U513, Creteil, France 3 Hopital Robert Debre´, Paris, France Tourette’s Syndrome (TS) is characterized by involuntary motor and vocal tics, whereas Obsessive-Compulsive Disorder (OCD) is characterized by recurrent, distressing unwanted thoughts and repetitive ritualistic behaviour. Both disorders have phenomenological and familial-genetic overlaps. Although no susceptibility genes for TS or OCD have been clearly identified, positive linkage and association studies have been reported. Chromosome 7q31 has been implicated in both disorders by the presence of a de novo duplication [46, XY,dup(7)(q22.1-q31.1)] in a proband with TS and the segregation of a translocation t(7, 18) (q31, q22.3) in a family with OCD and TS. Within 7q31, the inner mitochondrial membrane peptidase 2 like (IMMP2L) and the neural cell adhesion molecule (NrCAM) may represent suitable candidates for both disorders. Indeed, IMMP2L, disrupted in the duplication, has recently been shown to be associated with TS. NrCAM is an important mediator with multiple effects on the development of the neurons such as guidance cues. Here, we report the mutation screening of the IMMP2L and NrCAM genes (promoter region and coding exons) in 24 individuals (16 diagnosed with TS and OCD and 8 with OCD and family history of TS). No variations were identified in the IMMP2L gene. In NrCAM, thirteen single nucleotide polymorphisms (SNPs) were identified: five in the intronic sequences and seven in coding regions. Of the exonic polymorphisms, only one changed the amino acid sequence (A526P). Linkage disequilibrium was calculated and the most informative SNPs were genotyped in 100

71

additional OCD patients and 95 controls. None of the SNPs showed association with OCD. In conclusion, this analysis strongly suggests that coding variations of IMMP2L and NrCAM do not play a major role in TS and OCD etiology. These results also delineate the screening of additional candidate genes in the 7q31 region. P5.31 NO ASSOCIATION BETWEEN NORADRENALINE TRANSPORTER GENE AND ANOREXIA NERVOSA Hu X,1 Karwautz A,3 Li T,2 Wagner G,3 Holliday J,4 Treasure J,4 and Collier D2 1 Social, Genetic and Developmental Psychiatry Research Centre, The Institute of Psychiatry, London, United Kingdom 2 Section of Genetics, Department of Psychological Medicine, Institute of Psychiatry, London, United Kingdom 3 University Clinic of Neuropsychiatry of Childhood and Adolescence, University of Vienna, Vienna, Austria 4 Eating Disorders Unit, The Institute of Psychiatry, London, United Kingdom Because of the possible functional relevance of the noradrenaline to eating disorders and the association observed between the noradrenaline transporter and anorexia nervosa by Urwin et al. (2002) we analysed a promoter polymorphism (NETpPR) in 142 family trios from London, UK and Vienna, Austria, consisting of 67 with restricting anorexia nervosa (RAN), 48 with binge-purging anorexia nervosa (ANBP) and 27 unclassified AN patients. In the sample overall we und that there was no significant transmission distortion for any of the three alleles we detected with the putative L4 risk transmitted 41 times and not transmitted 40 times (w2 1.41, ns). When we looked at RAN only, allele L4 was transmitted 20 times and not transmitted 22 times (w2 ¼ 0.09, ns). For BP-AN only allele L4 was 17 times and not passed 9 times (w2 ¼ 2.46, P ¼ 0.12). Thus we failed to replicate the association Urwin et al. (2002) saw between alleles of the noradrenaline transporter polymorphism, NETpPR.

P5.32 ASSOCIATION OF COMT ALLELES IN SPANISH PATIENTS WITH PANIC DISORDER Gratacos M,1 Martin-Santos R,2 Cabre´ M,2 Navine´s R,2 Morell M,1 and Estivill X1 1 Centre de Regulacio´ Geno`mica, Barcelona, Spain 2 Unitat de Recerca en Psiquiatria, Barcelona, Spain COMT modulates one of the major degrading pathways of the catecholamine transmitters, including dopamine, epinephrine, and norepinephrine. COMT is widely expressed in the mammalian brain and, in humans, a common functional polymorphism is associated with a three- to four-fold variation in the enzyme activity. This variation in activity is due to a G to A transition at codon 158 that results in a Valine (high-activity allele) to a Methionine (low-activity allele) change. Recent studies support the hypothesis of a role for COMT in PD, with a higher prevalence of the low activity allele in PD patients, and the frequency of the M/M genotype significantly higher in patients with PD and a poorer treatment response than those with the other genotypes. However, another association analysis revealed a significant excess of the more active COMT allele in female patients with PD in a german sample. Moreover, linkage and genetic association in 70 multiplex PD families and 83 triads of American origin, found significative values analyzing several intragenic SNPs and microsatellites located near the COMT gene. The aim of our study was to replicate these results in a sample of 93 outpatients of Spanish origin, with or without agoraphobia (DSM-IV). The sample consisted of 69 women and 24 men, all diagnosed by the SCID-DSM-IV, and they were compared to a set of 98 unrelated sex-matched controls (70 women and 28 men).There were no overall allele differences between patients and controls, but a higher prevalence of genotypes composed of high activity alleles was observed when only women were considered (P ¼ 0.003, OR ¼ 5.6, 95% CI ¼ 1.6– 19.8). Conversely, women with PD had a higher prevalence of H alleles (P ¼ 0,02, OR ¼ 1.6, 95% CI ¼ 1.05–2.6). There were no differences between patients with PD and the control group when we stratified the patient’s sample in one group, including those with PD with agoraphobia (n ¼ 60) and the other PD without agoraphobia (n ¼ 31). These data do not suport a major contribution of this SNP to the subgroup with agoraphobia. Moreover, we also stratified patients by age of onset, harm avoidance trait (HA-TCI) and total phobias score (F100), and only HA showed a significant association to the high activity

72

Abstracts

genotype of this COMT polymorphism (P ¼ 0.008). In conclusion, our study confirms a potential role of COMT in PD in female patients, replicates a previous observation in a German sample (Domschke et al., 2004), and points out these gene as related to personality harm avoidance temperament trait in panic disorder patients. Supported by Psychiatric Genetics Network (Spanish Ministry of Health), Genome Spain, Generalitat de Catalunya and Fundacio´ La Marato´ de TV3.

P5.33 COMPARISION OF MEXICAN AND LOUISIANA OBSESSIVE COMPULSIVE DISORDER (OCD) PATIENTS ASCERTAINED THROUGH A FAMILY STUDY Weissbecker K,1 Nicolini H,2 Camarena B,2 and Winstead D1 1 Tulane University Health Science Center, New Orleans, United States 2 Child Psychiatric Hospital, Mexico City, Mexico There have been very few cross-cultural studies of symptomology of Obsessive Compulsive Disorder (OCD). The few cross-national epidemiologic studies that have been done, found consistency in rates and comorbidity. We compared age-of-onset, severity, comorbidity and types of obsessions and compulsions reported by 40 Mexican patients and 35 patients from Louisiana, USA. Diagnoses were done by direct interview in both countries using either the English or Spanish versions of the Diagnostic Interview Schedule (DIS), the Y-BOCS and the Y-BOCS checklist. There was a significant difference in the age-ofonset of OCD with the Mexican patients reporting an older age. There were also significant differences in severity. The Mexican patients had higher Y-BOCS scores. However, in the Mexican group some of the patients were assessed prior to drug therapy whereas all the Louisiana patients were medicated before joining the study. There was no difference in the rate of ‘‘familial’’ OCD between the two groups. The Mexican patients had significantly higher rates of comorbid tic disorders (not including TS), while the Louisiana patients had higher rates of alcoholism, drug abuse and major depression. There were no differences in rates of other psychiatric disorders. Interestingly, a significantly higher percentage of Mexican patients endorsed aggressive obsessions compared to the Louisiana patients (95% versus 60%). The Mexican patients also endorsed a higher percentage of symmetry obsessions (70% versus 17%), somatic obsessions (82% versus 51%), and ordering compulsions (57% versus 43%). Although the Y-BOCS checklist records both past and present symptoms, it is possible that recall of symptoms in medicated patients is lower than recall in patients who have not yet received treatment, and are therefore more severely affected at the time of the assessment. This may explain the overall higher rates of obsessions and compulsions reported by the Mexican patients compared to Louisiana patients. These findings suggest a mixture of cultural and diagnostic differences between Mexican and Louisana OCD patients that require further exploration.

P5.34 GENETIC VARIATION AT THE G72/G30 LOCUS IS ASSOCIATED WITH PANIC DISORDER Schumacher J,1 Abou Jamra R,1 Becker T,9 Klopp N,2 Franke P,3 Jacob C,4 Sand P,4 Fritze J,5 Ohlraun S,6 Schulze TG,6 Rietschel M,6 Illig T,2 Propping P,1 Cichon C,7 Deckert J,8 and No¨then MM7 1 Institute of Human Genetics, University of Bonn, Bonn, Germany 2 Genome Analysis Center (GAC), GSF-National Research Center for Environment and Health, Neuherberg, Germany 3 Department of Psychiatry, University of Bonn, Bonn, Germany 4 Department of Psychiatry, University of Wu¨rzburg, Wu¨rzburg, Germany 5 Department of Psychiatry, University of Frankfurt, Frankfurt, Germany 6 Central Institute of Mental Health, Division Genetic Epidemiology in Psychiatry, Mannheim, Germany 7 Life & Brain Center, University of Bonn, Bonn, Germany 8 Department of Psychiatry, University of Mu¨nster, Mu¨nster, Germany 9 Institute for Medical Biometry, Informatics and Epidemiology, University of Bonn, Bonn, Germany The G72/G30 locus on chromosome 13q33 can be considered a strong positional and functional candidate for panic disorder (PD). In the present study, we assessed whether genetic variation at this locus contributes to the development of PD using a case-control association study. We analyzed five SNPs in a sample of 152 German patients and

208 controls. In the single-locus analyses, three of five tested SNPs were significantly associated with PD at the P < 0.05 level. Further support was obtained at the haplotypic level suggesting risk-conferring as well as protective haplotypes. A stringent adjustment for testing multiple loci through a robust permutation procedure yielded an empirical P-value of 0.11 for the single marker analysis, suggesting the possibility of a chance finding. However, given that we previously observed the very same alleles and haplotypes to be associated with disease in bipolar affective disorder and schizophrenia in the German population, we believe that our findings are of importance and warrant further replication. Our results suggest that G72/G30 contributes to psychiatric diseases beyond diagnostic boundaries. It will be a matter of future studies to fully understand the phenotypic expression of the G72/G30 locus.

P6.1 CYTOGENETIC ABNORMALITIES IMPLICATE GLUTAMATE NEUROTRANSMISSION, CAMP REGULATION AND AXONAL GUIDANCE PATHWAYS IN PSYCHIATRIC ILLNESS Pickard BS,1 Malloy MP,1 Birtley JR,1 LeHellard S,1 Hampson M,1 Ewald HL,3 Mors O,3 Porteous DJ,1 Blackwood DHR,2 and Muir WJ2 1 Medical Genetics, Molecular Medicine Centre, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh, United Kingdom 2 Psychiatry, University of Edinburgh, Kennedy Tower, Royal Edinburgh Hospital, Morningside Park, Edinburgh, United Kingdom 3 Department of Psychiatric Demography, Institute for Basic Psychiatric Research, Psychiatric Hospital in Aarhus, Aarhus University Hospital, Risskov, Denmark Determining the molecular basis for the well-characterised heritability of psychiatric illnesses such as bipolar disorder and schizophrenia has proved a considerable challenge. Several aetiological hypotheses such as deficits in monoamine and glutamatergic neurotransmission, glial cell dysfunction and failures during neurodevelopment have been suggested. Fluorescence in situ Hybridisation (FISH) approaches have already identified candidate genes in subjects where chromosomal abnormalities are associated with neuropsychiatric disorders. Here we identify a novel set of genes disrupted by, or proximal to, chromosomal breakpoints in five subjects with psychiatric conditions, most commonly schizophrenia. These include two ionotropic glutamate receptors, GRIK4 and GRID2, a component, SEMCAP3, of semaphorinmediated axonal guidance, a cell adhesion molecule, CDH8, a transcription factor, PKNOX2/PREP2 and a cAMP phosphodiesterase, PDE4B, the target of the antidepressant rolipram. These candidate genes and their biological pathways add weight to the involvement of particular processes in mental illness aetiology and present opportunities for future functional and population-based studies.

P6.2 SEMCAP3: A CANDIDATE GENE FOR SCHIZO-AFFECTIVE DISORDER McLaren LJ,1 Pickard BS,1 Malloy MP,1 Ewald HL,2 Mors O,2 Porteous DJ,1 Blackwood DHR,3 and Muir WJ3 1 Medical Genetics, Molecular Medicine Centre, University of Edinburgh, Edinburgh, United Kingdom 2 Department of Psychiatric Demography, Institute for Basic Psychiatric Research, Aarhus University Hospital, Risskov, Denmark 3 Psychiatry, University of Edinburgh, Edinburgh, United Kingdom A patient diagnosed with schizo-affective disorder was found to have a balanced translocation between chromosomes 3 and 8: t(3,8)(p13,p22). The breakpoint on chromosome 3 lies between exons 3 and 4 of the gene SEMCAP3 (semaphorin cytoplasmic domain-associated protein) resulting in one inactivated allele. The breakpoint on chromosome 8 lies downstream of the coding region of the TUSC3/N33 gene involved in the N-linked glycolsylation pathway. Therefore, it is proposed that the SEMCAP3 disruption is the most likely cause of the schizo-affective disorder. SEMCAP3 was originally identified in mouse from a yeast two-hybrid screen using the C-terminal region of the semaphorin protein, Sema4C, as bait. Semaphorin proteins are involved in neurite outgrowth and patterning and also in neural crest migration and immune system modulation. All the identified SEMCAP proteins contain PDZ domains suggesting they may act to organize protein structures and regulate transmembrane signalling complexes. How-

Abstracts ever, unlike SEMCAP1 and 2, SEMCAP3 also contains a RING (really interesting new gene) finger, a motif present in (E3) ubiquitin ligases, constituents of the proteosome degradation pathway. Current work involves analysis of SEMCAP3 RNA and protein expression in human cells lines and mouse embryos, and in silico analysis of homologues of SEMCAP3 present on human chromosomes 12 and X. Future work will include identification of SEMCAP3 interacting proteins and association studies to determine whether SEMCAP3 contributes to psychiatric illness in karyotypically normal patients.

P6.3 EVALUATION OF PARENTAL REPORTING OF MEDICAL HISTORY AND PHENOTYPIC FEATURES FOR A RARE GENETIC AND BEHAVIOURAL DISORDER, RING CHROMOSOME 22 Jeffries AR, Curran SR, and Powell JF Institute of Psychiatry, London, United Kingdom Phenotypic studies on rare genetic disorders can prove logistically difficult and costly, especially when carried out on a worldwide scale. One solution to this problem is to use questionnaires which the parents can complete for phenotypic collection. This minimises the cost and increases the speed of the phenotypic collection. However, the accuracy of such phenotypic data may be called into question. This study addresses the accuracy of such parental completed surveys by application to a rare genetic disorder—ring chromosome 22 or r(22). Phenotypic data was previously collected from the parents through use of postal or web based questionnaires which included medical history, basic anthropometric measurements, developmental charts and behavioural scales. Photographs of the child were also taken by the parents to enable dysmorphology assessment by a clinical geneticist. In order to evaluate the accuracy of our phenotypic collection, we also contacted the patient’s primary care physician to collect the phenotypic data through similar questionnaires. We compare the accuracy of parental report against the medical records held by their physician. Here we present the findings of this study, together with recommendations on the design of parental questionnaires for similar phenotypic collection studies.

P6.4 MOLECULAR CYTOGENETIC EVIDENCES FOR THE PRESENCE OF CHROMOSOMAL MOSAICISM IN THE HUMAN BRAIN Yurov YB,1 Monachov VV,1 Iourov IY,1 Vostrikov VM,1 and Vorsanova SG2 1 National Research Center of Mental Health, RAMS, Moscow, Russia 2 Institute of Pediatrics and Children Surgery, Russian Ministry of Health, Moscow, Russia The human brain is the control center that stores, computes, integrate and transmit information contains about one trillion (1012) neurons, each forming as many as a thousand connections with other neurons. Up to date, there were no direct studies of the large scale genomic variations and chromosomal complement in the central nervous system in man. It was postulated (without experimental proofing) that the neuronal cells of the normal brain contain normal (diploid) chromosomal complement. However, direct proof of chromosomal mutations, leading to large-scale genomic alterations in neuronal cells, has been missing in the human brain. The aim of present study was the analysis of chromosomal complement in neuronal cells of the adult brain tissues using interphase multicolor fluorescence in situ hybridization (MFISH). DNA probes for different chromosomes (1, 9, 13, 14, 15, 16, 18, 21, 22, X, Y) and individual chromosome scoring up to 10 000 neuronal cells per sample were applied. The level of aneuploidy was in range of 0.002–0.08% for different chromosomes. We could propose that several billions of aneuploid neurons could exist in the adult brain, and therefore, affect the functioning of the enormously large amount of normal neurons. ‘‘Criptic’’ chromosomal mosaicism of the adult brain has no phenotypic appearance, but should have substantial negative effect on normal brain development and functions. Together with the evidences concerning occurrence of chromosomal abnormalities in neuronal cells of the brain in Alzheimer disease (Yang et al., 2002) and schizophrenia (Yurov et al., 2002), our data also indicate that instability of genome due to chromosomal variations may have relevance to the psychiatric diseases. Supported in parts by grant s Copernicus 2 and INTAS 03-51-4060.

73

P6.5 RING 14 STABLY INHERITED FROM MOTHER TO DAUGHTER BUT WITH AN APPARENT CHANGE IN ASSOCIATED CLINICAL PHENOTYPE Malloy MP, Pickard BS, MacIntyre D, Porteous DJ, Blackwood DHR, and Muir WJ University of Edinburgh, Edinburgh, United Kingdom Ring chromosome 14 is a rare abnormality (under 100 cases described) and typically associated with growth impairment, microcephaly, a distinct facies and learning disability (US Mental retardation). The phenotypic features are usually associated with a partial deletion of euchromatin on chromosome 14 as part of the postulated ring formation process. A previous case of inherited ring 14 was associated with mosaicism of ring 14 and a t(14q,21q) in the mother, and one possible mechanism to explain this involves a balanced rearrangement between both anomalies in the mother, with the sons who had learning disability and seizures, receiving only the ring chromosome itself containing a partial deletion of 14. We describe a separate family where ring 14 is stably inherited. The mother is not mosaic and the ring chromosome 14 is the only anomaly present. She does not have learning disability or any of the other described phenotypic features. She has passed the ring 14 to her daughter who does have severe learning disability. We have been able to study the ring 14 from mother using FISH markers mapping to 14 telomere and into 14 euchromatin. No loss of chromosome 14 euchromatin has been detected and we conclude that euchromatic deletion on chromosome 14 is not a necessary prerequisite for ring chromosome 14 formation: telomere fusion seems to be adequate in this case. However passage through meiosis might not be stable such that the phenotype in daughter may be due to further re-arrangement of the ring. We are attempting to obtain a sample from the daughter to verify this latter hypothesis but the lack of deletion in the mother is of considerable interest as regards chromosomal abnormality formation and it would be of interest to directly sequence the telomeric region in this person. P6.6 AUTISM PHENOTYPE IN CHILDREN WITH 22Q11 DELETION SYNDROME Vorstman JAS,1 Morcus MEJ,1 Swaab-Barneveld H,1 Heineman-de Boer JA,2 Duijff SN,2 Jansen P,2 Beemer FA,2 and van Engeland H2 1 University Medical Center, Utrecht, Netherlands 2 Wilhelmina Children’s Hospital, Utrecht, Netherlands The present study is part of a large ongoing research project on the molecular genetic characteristics and the cognitive and psychiatric phenotype of children with 22q11 deletion syndrome (22q11DS). The question was addressed whether children with 22q11DS show specific child psychiatric problems. It was expected that these children would show behavioural and emotional problems that result in social problems and therefore would result in diagnosis of autism spectrum disorders. If social problems are found, we expected that these problems were explained partially by their difficulty in visual emotion recognition. 50 Children with 22q11DS, aged 10 to 20, were thoroughly assessed following standardised child psychiatric examination. The psychiatric examination included a semi-structured DSM-IV interview, the ADI-R and the K-SADS sections ‘‘psychosis,’’ ‘‘bipolar disorder’’ and ‘‘mood disorder,’’ as well as several parent and teacher questionnaires. In addition, the children were neuropsychologically extensively examined regarding their ability to process social information. Their intelligence level was measured by using the WISC-III. Results will be reported of the outcome of the child psychiatric examination in this group of children. Moreover their ability to recognise human faces will be evaluated in and will be compared to their ability to recognise facial emotions. Preliminary results indicate that a diagnosis of an autism spectrum disorder is highly common in the 22q11DS group. The relation between the social problems and the facial and visual emotion recognition will be elaborated.

P6.7 PREVALENCE OF 22Q11 DELETIONS IN A DUTCH POPULATION OF PATIENTS WITH DEFICIT SCHIZOPHRENIA Hoogendoorn MLC, Vorstman JAS, Bakker SC, Sinke RJ, Beemer FA, and Kahn RS 1 University Medical Center, Utrecht, Netherlands 2 Wilhelmina Children’s Hospital, Utrecht, Netherlands

74

Abstracts

The 22q11-deletion syndrome (22q11DS) is a congenital malformation syndrome with characteristic physical manifestations including typical facial appearance, cleft palate, pharyngeal hypotonia, lymphoid tissue hypoplasia and heart defects. Besides the physical defects, cognitive and psychiatric disorders can be part of the phenotype as well. Several studies found a high rate, -( 20–30%)-, of psychotic disorders, most frequently schizophrenia, in their samples of 22q11DS patients. The concurrence of a well-described cytogenetic defect with an increased risk of schizophrenia provides a unique opportunity to study genetic mechanisms underlying this psychiatric illness. From a broader clinical perspective, assessment of the prevalence of 22q11DS in the general schizophrenia population is important, guiding clinicians in decisions with regard to genetic testing of patients with schizophrenia. To date, data regarding the prevalence of 22q11DS among patients with schizophrenia is somewhat conflicting with earlier studies reporting rates ranging from 2% in an adult sample to 6% in a childhood-onset schizophrenia sample, while in recent studies lower rates (1/300–500 (Arinami et al., 2001; Ivanov et al., 2003)) in adult patients with schizophrenia are reported. In the current study we examined the prevalence of 22q11DS in a sample of adult patients with schizophrenia with a poor prognostic outcome. Diagnostic assessment of patients with schizophrenia was performed by trained clinicians using the Comprehensive Assessment of Symptoms and History (CASH) and the Schedule for the Deficit Syndrome (SDS), as part of a larger ongoing study on the genetics of schizophrenia. Within this sample a subgroup of 130 patients were diagnosed with the deficit syndrome, which is characterised by persistent, idiopathic negative symptoms. Screening for the 22q11DS was performed using a multi-probe PCR variant, Multiplex Ligation-dependent Probe Amplification (MLPA). With this method 11 DNA sequences within and flanking the 22q11 region were quantified in each patient. So far 75 patients have been screened, yielding no 22q11 deletion. These preliminary results indicate that 22q11DS is not a major factor contributing to deficit schizophrenia. Screening of the remaining 55 patients will be finished prior to the conference, and results of the total sample (n ¼ 130) will be presented. P6.8 IDENTIFICATION OF RISK GENES AND BIOLOGICAL PATHWAYS: WHAT WE CAN LEARN FROM CYTOGENETIC ABNORMALITIES ASSOCIATED WITH THE AUTISM PHENOTYPE Vorstman JAS, Staal WG, van Daalen E, Hochstenbach R, and van Engeland H 1 University Medical Center, Utrecht, Netherlands 2 Wilhelmina Children’s Hospital, Utrecht, Netherlands Cumulative evidence from family, twin- and adoption studies suggests that genetic factors play an important role in the pathology of autism. The susceptibility for autism is estimated to be 90% genetic, most likely through a polygenetic mode of inheritance. Through linkage and association studies several loci have been connected with autism. However, most results have been difficult to replicate. Presumably this could be due to considerable biological, hence genetic, heterogeneity of the autism phenotype. However, when the autism phenotype associated with a specific cytogenetic abnormality is studied, a common biological cause is likely, thereby reducing the problem of biological heterogeneity. Subsequently, the actual autism phenotype may be refined with this type of information (e.g. ‘‘dup 15q11-14 autism’’ or ‘‘del 22q11 autism’’). Moreover, studying cytogenetic abnormalities associated with an autism phenotype is one possible way to reduce the total number of potentially implicated genes to be considered. The Englishlanguage literature was searched with the Pubmed and Medline libraries. Studies that related cytogenetical abnormalities with autism spectrum diagnosis were investigated and summarized. The information resulting from this search was used to define genomic regions of interest. Consequently, all genes within the boundaries of these regions were listed. The next step consisted of making a selection of potential candidate genes from these lists. Finally, using an extensive bio informatics search strategy shared biological functions/biological pathways were identified. An overview of all (cytogenetic) regions of interest on the human genome is presented. Specific regions on the human genome are clearly more often implicated in cytogenetic disorders concurring with the autism phenotype. The next step after the identification of regions of interest is using these data to search for biological pathways. The knowledge of possibly implicated biological pathways may contribute significantly to the development of research strategies aiming for the identification of genes responsible for autism.

P7.1 AGE OF PSYCHOTIC ONSET IN SCHIZOPHRENIA: A GENETIC NETWORK INVOLVING BDNF AND DOPAMINE D3 RECEPTOR GENES? Gourion DG, Goldberger CG, and Krebs MOK INSERM, Paris, France Brain-derived neurotrophic factor (BDNF) plays a major role in neuronal proliferation, survival and differentiation. This factor is involved in the development of meso-limbic dopaminergic-related system and regulates the dopamine D(3) receptor (DRD3) expression (Guillin et al., 2001). Both BDNF and DRD3 have been implicated in the emergence of psychotic symptoms in patients with schizophrenia. We tested the hypothesis of a gene-gene interaction between DRD3 and BDNF associated with the age of psychotic onset in patients with schizophrenia. A sample of 198 patients with DSM-IV schizophrenia were genotyped for the Val66Met polymorphism in the BDNF gene and for the Ser9gly polymorphism in the DRD3 gene. We performed a survival analysis in order to detect an effect of the genotype on the age of schizophrenic onset. There was a significant effect of the epistatic interaction between val66met and ser9gly genotypes on the age of psychotic onset in patients with schizophrenia (P < 0.017). These findings suggest that variation in the BDNF and D3 genes and their interaction may modify the onset of the illness, thus contributing to interindividual differences in course and severity of the disease. P7.2 ALLELES AT THE DOPAMINE D4 RECEPTOR LOCUS COULD CONTRIBUTE TO THE GENETIC SUSCEPTIBILITY IN FIRST PSYCHOTIC EPISODE Aguirre Ana Julia AJ,1 Cruz Carlos CC,2 Aguilar-Garcı´a Alejandro AA,2 Hernandez Sandra SH,2 Gomez Ariadna AG,2 Garcia Marı´a MG,2 Fresan Ana AF,2 and Apiquian Rogelio RA2 1 Universidad Nacional Autonoma de Mexico, DF, Mexico 2 Instituto Nacional de Psiquiatrı´a, DF, Mexico The dopamine D4 receptor gene (DRD4) has been suggested as a candidate gene for schizophrenia. Recently, one study found an association between a putative functional promoter polymorphism 521 C/T (DRD4) and schizophrenia. However, no information is available for association studies with 521 C/T polymorphism and first psychotic episode (FPE). Here, we analyzed 113 unrelated Mexicans with first psychotic episode, evaluated according to the DSM IV criteria and with respect to the DRD4 521 C/T polymorphism. The sample included to patients with major depression and bipolar disorder with psychotic symptoms, schizoaffective disorder, schizophrenia, brief psychotic disorder, delusional disorder and NOS disorder. We consider that FPE diagnosis guaranties a more homogeneous way to classify a psychotic group. 43 healthy subjects were recruited. Genotyping was performed using PCR-RFLP. Allele frequencies of the 521 C/T (0.6/ 0.4) of controls were in Hardy-Weinberg equilibrium (P ¼ 0.545, X2 ¼ 0.46). The heterocigocity were 0.5, which indicates moderate heterozygotes. The polymorphism information content (PIC) was high (0.7). The marker 521 C/T is informativeness, it is the probability that the parent is heterozygous by the probability that the offspring is informative for this locus. Subsequent analyses revealed that association was mainly due to T allele and TT genotype (P ¼ 0.01, X2 ¼ 6.605, df ¼ 1, with correction of Yates). This result supports previous postulation that DRD4 gene plays a role in the etiology of the psychotic symptoms, and this case, DRD4 could be important among FPE Mexican patients.

P7.3 A CASE-CONTROL STUDY ON THE RELATIONSHIP BETWEEN METABOTROPIC GLUTAMATE RECEPTOR 3 (GRM3) AND SCHIZOPHRENIA IN THE CHINESE POPULATION Chen Q,1 He G,1 Chen QY,2 Wu SN,1 Xu YF,3 Feng GY,3 Li YC,4 Wang LJ,5 and He L1 1 Shanghai Jiaotong University, Shanghai, China 2 Chinese Academy of Sciences, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China 4 Chenzhou mental hospital, Hunan, China 5 Tiexi neuro-mental hospital of Shenyang, liaoning, China The recent studies on the association between metabotropic glutamate receptor 3 (GRM3) and schizophrenia by Fujii et al and Marti et al.,

Abstracts produced conflicting results although GRM3 was thought to be a promising candidate gene for schizophrenia. Fujii et al. found rs1468412 to have a positive association with Japanese schizophrenics. To investigate this further, we genotyped 7 SNPs around GRM3 including rs1468412, in 752 Chinese schizophrenics and 752 controls using Taqman technology. We did not detect any association between rs1468412 and schizophrenia, however we found differences in the allele frequency distribution of SNP rs2299225 (P ¼ 0.0297, OR ¼ 1.44, 95% Confidence Interval 1.05–1.99) between cases and controls. Moreover the overall frequency of haplotypes constructed from three SNPs including rs2299225 showed significant differences between cases and controls (P ¼ 0.0017). Our results partially support the previous studies based on other ethnic groups and indicate that the GRM3 gene may play an important role in the etiology of schizophrenia in the Han Chinese.

P7.4 THE HUMAN FRIZZLED-3 (FZD3) GENE ON CHROMOSOME 8P21, A RECEPTOR GENE FOR WNT LIGANDS, IS ASSOCIATED WITH THE SUSCEPTIBILITY TO SCHIZOPHRENIA Katsu T,1 Ujike H,2 Nakano T,1 Tanaka Y,1 Nomura A,1 Nakata K,1 Takaki M,1 Sakai A,1 Uchida N,1 Imamura T,1 and Kuroda S1 1 Department of Neuropsychiatry, Eiko hospital, Mitoyo-gun, Japan 2 Department of Neuropsychiatry, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan Neurodevelopmental abnormalities have been reported in studies on the pathogenesis of schizophrenia. The Wnt-signaling pathway has been implicated in a variety of processes in neurodevelopment, and the frizzled proteins have been identified as receptors for Wnt ligands. Of the frizzled proteins, frizzled-3 (FZD3) is required for formation of the neural crest and for development of major fiber tracts in the CNS. The human FZD3 gene is located on chromosome 8p21, a positive linkage locus for schizophrenia. We analyzed polymorphisms of the FZD3 gene in patients with schizophrenia and control subjects in the Japanese population. We found a significant association between schizophrenia and the FZD3 gene in single nucleotide polymorphisms and haplotype analyses. Our data suggest that dysregulation of the Wnt-signaling pathway may be involved in the susceptibility to schizophrenia.

P7.5 TUMOR SUPPRESSOR GENE P53 IS GENETICALLY ASSOCIATED WITH SCHIZOPHRENIA IN THE CHINESE POPULATION Yang Yifeng,1 Xiao Zeping,3 Chen WX,4 Sang Hong,5 Peng Yanwei,6 Zhang Diran,7 Gu ZZ,8 He Guang,1 Qin Wei,2 Li Dawei,1 Gu NiuFan,3 and He Lin1 1 No.3 People’s Hospital of Huzhou, Zhejiang, China 2 Institute for Nutritional Sciences, SIBS, Chinese Academy of Sciences, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China 4 Pingshan Hospital, Hannan Province, Pinshan, China 5 Kaixuan Hospital of Changchun, Jilin Province, Chuangchun, China 6 Mental Hospital of Xiangxi Municipality, Hunan Province, Xiangxi, China 7 Anning Hospital, Guizhou Province, Anning, China Schizophrenia is a debilitating mental disorder. The p53 tumorsuppressor gene, encoding a phosphoprotein, is a key element in maintaining genomic stability and cell apoptosis. Recently, reduced risk of cancer in patients of schizophrenia has been reported. Some evidence also suggests the possible implication of p53 in the neurosystem development. In order to examine the role of the p53 gene for the pathogenesis of schizophrenic disorders, we investigated the genetic association between a functional polymorphism rs1042522 and schizophrenia by directly sequencing the fragment covering 72Pro> Arg in 701 cases and 695 controls in this work. In addition, we studied two other SNPs rs2078486 and rs8064946 by allele-specific PCR in the same samples. Though rs1042522 did not show positive association with schizophrenia, we did observe statistically significant differences on SNP rs2078486 (P ¼ 0.029, OR ¼ 1.21, 95% CI 1.02–1.42) and on haplotype CAC (P ¼ 0.007, OR ¼ 1.36, 95% CI 1.09–1.70). These results demonstrated p53 might play a role in susceptibility to schizophrenia.

75

P7.6 TRANSMISSION DISEQUILIBRIUM ANALYSIS OF THE MAG GENE IN A COHORT OF FAMILY TRIOS WITH SCHIZOPHRENIA Yang Yifeng,1 Wei Qin,2 He Guang,1 Zhou Jian,1 Xu Yifeng,3 Chen Qi,1 Liu Xinmin,1 Gu Niufan,1 Feng Guoyin,3 Sang Hong,4 Wang Peng,5 and He Lin1 1 Shanghai, China 2 Institute for Nutritional Sciences, SIBS, Chinese Academy of Sciences, 320 Yueyang Road, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China 4 Changchun Kaixun Hospital, Changchun, China 5 Wuhu No.4 People’s Hospital, Wuhu, China Recently, oligodendrocyte and myelin dysfunction has been found in schizophrenia. For example, MAG, myelin-associated glycoprotein, has shown a significant reduction in the schizophrenia patient groups compared to that in the control groups. These observations are changing the thinking way accumulated in the past decades of research on neuron, synapse, and neurotransmitter-based hypotheses produced by only moderately effective therapeutic approaches. In order to test this finding, we genotyped four markers around the MAG locus in 413 trios sample of the Han Chinese using allele-specific PCR. None of the four markers revealed significant allelic significance. However, we identified one common haplotype TA (BLOCK B, P value ¼ 0.00007145, df ¼ 1) where the schizophrenia susceptibility variant(s) likely reside. The results demonstrated MAG might play a role in susceptibility to schizophrenia. P7.7 NO ASSOCIATION BETWEEN THE GENETIC POLYMORPHISMS WITHIN RTN4 AND SCHIZOPHRENIA IN THE CHINESE POPULATION Chen WY,1 Gu NF,3 Duan SW,1 Sun Y,2 Zheng YL,2 Li C,1 Pan YX,1 Xu YF,3 Feng GY,3 and He L1 1 Bio-X Life Science Research Center, Shanghai Jiao Tong University, Shanghai, China 2 Institute for Nutrition Sciences, SIBS, Chinese Academy of Sciences, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China The RTN4 gene on chromosome 2p13-14 has been reported to be overexpressed in schizophrenia by a cDNA subtractive hybridization experiment between postmortem human frontal cerebral cortices from Canadian schizophrenia individuals and neurological controls. The same study also reported a high prevalence of homozygous CAA insertion in the schizophrenia. In a replication attempt to investigate the role of RTN4 in the etiology of schizophrenia, we genotyped the CAA insertion polymorphism and other three genetic polymorphisms (a TATC deletion in the 30 -untranslated region and two single nucleotide polymorphisms in the 50 region) within RTN4 and conducted a casecontrol study in the Chinese population. There were no significant discrepancies in allele and genotype frequencies of the four polymorphisms individually and haplotype distribution between the cases and the controls. Our current data suggest that the genetic polymorphisms within RTN4 are unlikely to confer an increased susceptibility to schizophrenia in the Chinese population.

P7.8 METHYLENETETRAHYDROFOLATE REDUCTASE VARIANT IN SCHIZOPHRENIA AND MOOD DISORDERS Tan EC,1 Chong SA,2 Lim LCC,3 Chan AOM,4 and Tan CH5 1 Defence Medical and Environmental Research Institute, Singapore, Singapore 2 Institute of Mental Health, Singapore, Singapore 3 LP Clinic, Singapore, Singapore 4 Changi General Hospital, Singapore, Singapore 5 National University of Singapore, Singapore, Singapore Severe deficiency of the MTHFR enzyme is the most common inborn error of folate metabolism. It can result in either hyperhomocysteinaemia, homocystinuria, or hypomethioninaemia. Patients with these metabolic disorders have been reported to present with features of mental and physical disorders including psychiatric symptoms. Elevated homocysteine level has been reported for patients with schizophrenia and depression. We investigated the frequency of the common C667T variant of the enzyme methylenetetrahydrofolate reductase in controls and patients of Chinese descent. Controls with no

76

Abstracts

history of mental disorder and patients who met the DSM IV criteria for schizophrenia, bipolar or uniploar disorders were recruited. Genomic DNA from all were genotyped for the C667T polymorphism by PCRRFLP. There was no significant difference in genotype distributions or allele frequencies between controls and any of the diagnostic groups although the frequency of allele T was higher for all diagnostic groups and for both the male and female genders. When data was analyzed with the minor T allele as dominant, there was an excess of the T-containing genotypes in each of the patient groups compared to controls. For the difference between controls and all cases combined it almost reached statistical significance (P ¼ 0.077) with the odds ratio of 1.46 (95% CI 0.96–2.22). Our results showed that although there was no significant association as measured by P-value, the odds ratio and confidence interval provided some evidence of increased risk for individuals with the T-containing genotypes. As there was more than one positive association reported for both schizophrenia and depression, a minor role for this polymorphism in the pathogenesis of these two disorders could not be ruled out and would warrant further investigation. P7.9 AN ASSOCIATION BETWEEN POLYMORPHISMS OF THE INTERLEUKIN-10 GENE PROMOTER AND SCHIZOPHRENIA IN THE CHINESE POPULATION Yu L,1 Yang MS,1 Zhao J,2 Shi YY,2 Zhao XZ,2 Yang JD,1 Liu ZJ,2 Gu NF,3 Feng GY,3 and He L2 1 Institute of Nutrition Science, Shanghai Institutes for Biological Science, Chinese Academy of Sciences, Shanghai, China 2 Bio-X Life Science Research Center, Shanghai Jiaotong University, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China Immunological abnormalities have been found to be association with schizophrenia for decades. Cytokines are key proteins involved in the immune system activation. Interleukin-10 (IL-10), an important immunoregulatory cytokine, is located on chromosome 1q31-32, a region previously reported to be linked to schizophrenia in genetic studies. Thus, a study was carried out on the association between schizophrenia and three singly nucleotide polymorphisms in the promoter region of IL-10 gene. A total of 341 patients and 334 controls of Chinese descent were analysed. Statistically significant differences were observed in both allelic and genotypic frequencies of the 592A/C polymorphism (Allele, w2 ¼ 4.43, df ¼ 1, P ¼ 0.032, odds ratio (OR) ¼ 1.26, 95% CI 1.02–1.56, Genotype, w2 ¼ 8.18, df ¼ 2, P ¼ 0.044) while the other two polymorphisms did not show such differences. The observed haplotype distributions revealed a significant association with schizophrenia (P ¼ 0.0008). These data suggest that the IL-10 gene may confer susceptibility to the development of schizophrenia in the Chinese Population.

P7.10 TRANSMISSION DISEQUILIBRIUM ANALYSIS OF THE MAG GENE IN A COHORT OF FAMILY TRIOS WITH SCHIZOPHRENIA Yang YF,1 Qin W,2 He G,1 Zhou J,1 Xu YF,3 Chen Q,1 Liu XM,1 Gu NF,3 Feng GY,3 Sang H,4 Wang P,5 and He L1 1 Shanghai, China 2 Institute for Nutritional Sciences, SIBS, Chinese Academy of Sciences, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China 4 Changchun Kaixun Hospital, 74 Beihuan Road, Changchun, China 5 Wuhu No.4 People’s Hospital, Jiulong, Matang District, Wuhu, China Recently, oligodendrocyte and myelin dysfunction has been found in schizophrenia. For example, MAG, myelin-associated glycoprotein, has shown a significant reduction in the schizophrenia patient groups compared to that in the control groups. These observations are changing the thinking way accumulated in the past decades of research on neuron, synapse, and neurotransmitter-based hypotheses produced by only moderately effective therapeutic approaches. In order to test this finding, we genotyped four markers around the MAG locus in 413 trios sample of the Han Chinese using allele-specific PCR. None of the four markers revealed significant allelic significance. However, we identified one common haplotype TA (BLOCK B, P value ¼ 0.00007145, df ¼ 1) where the schizophrenia susceptibility variant(s) likely reside. The results demonstrated MAG might play a role in susceptibility to schizophrenia.

P7.11 A CASE CONTROL AND FAMILY BASED ASSOCIATION STUDY OF NEUREGULIN1 GENE AND SCHIZOPHRENIA Shi XZ,1 Shi YY,1 Tang JX,2 Tang RQ,1 Yu L,2 Gu NF,3 Feng GY,3 Zhu SM,4 Liu HJ,4 Xing YL,5 Zhao SM,5 Sang H,6 Guan YS,6 St. Clair D,7 and He L1 1 Bio-X Life Science Research Center, Shanghai Jiao Tong University, Shanghai, China 2 Institute of Nutrition Science, Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China 3 Shanghai Institute of Mental Health, Shanghai, China 4 JiLin Institute of Mental Health, Jinlin, China 5 Xi’an Institute of Mental Health, Xi’an, China 6 Changchun Kaixuan Hospital, Changchun, China 7 Dept of Mental Health, University of Aberdeen, Medical School, Foresterhill, Aberdeen, United Kingdom Recently a series of studies have identified Neuregulin 1 (NRG1) as a potential susceptibility gene for schizophrenia and a seven-marker atrisk haplotype has been reported to be associated with schizophrenia in different population. We have genotyped 5 of the 7 markers and performed case control and TDT analyses in total of 1724 Han Chinese individuals, including 369 cases, 299 controls and 352 family trios. We failed to detect association between the reported at-risk haplotype and schizophrenia due to the low frequency of this haplotype in our samples ( 3) and six genomic regions displayed suggestive linkage (LOD > 2). The two significant linkage peaks occurred at 10q, a region which has previously been linked to schizophrenia in an Ashkenazi Jewish sample, and a mixed sample from the UK, USA, and Sweden. The other peak is at 16p, near to the NDE1 gene, coding for a molecule suggested to be interacting with the DISC1 protein. Interestingly, among the six suggestive linkage peaks, five represent genomic regions, also previously linked to schizophrenia, including the loci for the Dysbindin, Neureglin 1, and Reelin genes. Linkage analysis of the families without the HEP3 haplotype revealed only two suggestive linkage regions. These findings would support the true role of several previously identified loci and highlight three additional loci for further investigation for predisposing to familial schizophrenia.

P7.21 ASSOCIATION STUDY OF RXR-BETA UTR376A > T BB GENOTYPE AND HEAD DISMORPHISM AND THOUGHT DISORDER IN A SAMPLE OF BRAZILIAN DSM-IV SCHIZOPHRENICS OF EUROPEAN ETNICITY DIAGNOSED BY OCRIT SYSTEM Lobato MI,1 Gama CS,1 Hutz M,2 Palha J,3 Goodman AB,4 Azevedo MH,5 and Belmonte-de-Abreu PS1 1 Universidade Federal do Rio Grande do Sul-Faculdade de Medicina, Porto Alegre, Brazil 2 Universidade Federal do rio Grande do Sul- Departamento de Gene´tica, Porto Alegre, Brazil 3 Universidade do Minho, Braga, Portugal 4 Harvard University, Boston, United States 5 Universidade de Coimbra, Coimbra, Portugal Despite several technical and medical advances over the last decade, schizophrenia is still a personal and family disaster. It is a chronic disabling disease affecting about 1% of the world’s population, associated to mutations in several small effect genes but with no major effect gene. In additon to the association of Dopaminergic system in the

Abstracts illness, recent models mechanisms of neurodevelopment and neuroregulation by the interaction of hormones and vitamins. Among these, the retinoids and thyroid hormones have been proposed as good candidates, because of clinical evidence of retinoid toxity or deficit resulting in symptom presentations resembling schizophrenia, eg., thought disorder, mental deficit, enlarged ventricules, agenesis of the corpus callosum, microcephaly and congenital malformations. Craniofacial and digital anomalies are prominent and frequently reported among schizophrenics. A sample of 87 brazilian from european origin with DSM-IV diagnosis of schizophrenia obtained by the OPCRIT system (age range 18–63, mean 34.3 years) were studied in a protocol approved by the ethics commitee the major University Hospital of the State of Rio Grande do Sul, Hospital de Clı´nicas de Porto Alegre. Venous blood was drawn from all subjects and DNA extracted using standard salting-out procedures. Genotyping of TTR c.-245 T > A, c.289 A > g, c.308 A > g, RXRBeta UTR376A > T and UTR 243G > T mutations and the deletion TTR c.-469delG followed PCR conditions previously described (Leˆ et al., 2003; Buervenich et al., 2000; Chen et al., 2001). Antropometric measures were obtained by the Complete form of Waldrop Scale. Minor Physical Abnormalities (MPA) were assessed by 2 trained researchers from the Genetics Service from the same Hospital. The study assessed the association of Transtyretin (TTR) and Nuclear Receptor X Beta (RXRaˆ) polimorphisms with symptoms and dismorphysms. RXRbeta UTR376A > T BB genotype was associated to the presence of thought racing (Opcrit # 31, P ¼ 0.009) and Decreased Head circunference (P ¼ 0,007). There was no association of TTR mutations and deletion with MPA and dismorphism. The study gives partial support to the theory of association of retinoid cascade gene mutations (RXRB UTR376A > T BB genotypes) with developmental abnormalities in schizophrenia (microcephaly) and with specific psychotic symptoms (thought racing). BB genotype was associated to decreased head and thought racing. The sample size does not allow the stratification and multivariate analysis, to check interactions and specific endophenotype studies. The group is increasing sample size through the International Consortium Brazil-US-Portugal to test further associations and confounding effects.

P7.22 CHARACTERISATION OF DISC2, A GENE DISRUPTED BY A TRANSLOCATION THAT SEGREGATES WITH SCHIZHPHRENIA Chubb J, Porteous DJ, and Millar JK University of Edinburgh, Edinburgh, United Kingdom Using a cytogenetic approach we have identified two genes that are disrupted by a reciprocal chromosomal translocation that segregates with major psychiatric illness in a large Scottish family. This is a balanced translocation between the long arms of chromosomes 1 and 11 that directly disrupts two genes on chromosome 1 which have been named Disrupted-In-Schizophrenia 1 & 2. Further support for the involvement of this region in psychiatric illness has come from numerous independent reports of linkage to this region on chromosome 1. DISC1 encodes a large protein that in recent years has been under intense scrutiny regarding its involvement in schizophrenia and affective disorders. To date however there has been little research into the potential role of DISC2 in the manifestation of schizophrenia and affective disorders, although linkage and association data do not discriminate between the two genes. DISC2 is located on the noncoding strand partially antisense to DISC1. Antisense genes are attracting much attention in recent years due to the widespread occurrence of transcription from the antisense strand within the human genome and also due to their potential involvement in regulation of sense gene expression. We aim to answer questions relating to the structure of DISC2, its function(s) and its conservation throughout evolution, all with a view to evaluating the potential role of DISC2 in the pathology of schizophrenia. This work will comprise 1) identification of the full length gene, building on the 15kb of unspliced DISC2 sequence currently known and evaluation of its coding capacity, 2) evaluation of DISC2 as an antisense regulator of DISC1, 3) investigation of DISC2 as a candidate gene which functions independently of DISC1 and 4) investigation of the hypothesis that DISC2 transcription is driven by a retroviral insertion upstream of the known sequence and that the function of DISC2 may therefore be specific to primates.

79

P7.23 FURTHER EVIDENCE FOR THE 5HTR2A GENE CONTRIBUTION TO VERBAL MEMORY: A STUDY OF SCHIZOPHRENIC POPULATION Golimbet VE, Alfimova MV, Lyachenko GL, Kopaneva NA, and Borozdina SA Mental Health Research Center Russian Academy of Medical Sciences, Moscow, Russia Memory is one of the most impaired cognitive domains in schizophrenia and heritability of short-term memory in the patients is considered to be essential. Recently, an impact of serotonin receptor type 2A (5HTR2A) His/Tyr polymorphism on episodic memory has been reported in psychiatrically well subjects (Quervain et al. Nat Neurosci. 2003 Nov,6(11):1141). The present study aimed at searching for an association between another 5HTR2A polymorphism (T102C) and verbal memory. The A2A2 genotype of T102C 5HT2RA polymorphism located in exon 1 was reportedly found to be associated with schizophrenia and its clinical features. Genotyping and neuropsychological assessment have been carried out in 135 patients with ICD-10 diagnosis of schizophrenia (90 male, 45 female, mean age 34.4  13.0 years, illness duration 10.610.1 years). Patients were given a word-list recall task: to listen to a list of 10 semantically unrelated nouns and recall them immediately, the procedure being performed two times. An averaged number of correct words recalled was taken as a short-term memory measure. Patients with the A2A2 genotype performed poorer (4.3  1.3) on the words recalling compared to those with the A1A1 genotype (4.9  1.3) (t ¼ 2.6, df ¼ 74, P ¼ 0.01). Correlation analysis revealed significant reverse correlation between short-term memory and both positive (r ¼ 0.35, P ¼ 0.00*) and negative symptoms (r ¼ 0.52, P ¼ 0.00*) measured with the PANSS as well as the illness duration (r ¼ 0.26, P ¼ 0.009). However, when adjusted for these traits as covariates, the difference between short-term memory and genotypes remained significant (P ¼ 0.02 and P ¼ 0.01 respectively). Regression analysis revealed that genotype was a significant predictor of memory (P ¼ 0.004), contributing to 5% of the total variance. Thus, our results support a role of 5HTR2A gene in verbal memory. This work was supported, in part, by the Russian Fundamental Research Fund N 04-04-48588.

P7.24 ADRA1A PROMOTER POLYMORPHISMS ASSOCIATED WITH SCHIZOPHRENIA IN SPANISH ISOLATE Clark DA,1 Arranz MJ,1 Mata I,2 Lopez-Ilundain J,2 Perez-Nievas F,2 and Kerwin RW1 1 Institute of Psychiatry, London, United Kingdom 2 Fundacion Argibide, Pamplona, Spain The status of noradrenergic receptor populations in psychiatric patients remains unclear, but evidence suggests changes in expression levels in comparison to unaffected individuals. We focused on one receptor in this family, the a1A-adrenoceptor and proposed that SNPs in the promoter region of this receptor gene might alter its expression levels leading to the change in receptor populations observed in psychiatric patients. We tested for association between schizophrenia and a1A-adrenoceptor promoter SNPs in a case-control study. The sample consisted of 101 schizophrenia and 14 schizoaffective disorder patients diagnosed to DSM-IV criteria and 174 controls, all of Basque origin. A 10 kb region upstream of the translation initiation site was searched for polymorphisms using a combination of bioinformatics and laboratory-based mutation screening. Of the 8 polymorphisms that were confirmed at a rare allele frequency of greater than 5% (9625-G/ A, 7255-A/G, 6274-C/T, 4884-A/G, 4155-C/G, 2760-A/C, 1873G/A and 563-C/T), the 9625-G/A, 4155-C/G and 563-C/T were found to be associated with schizophrenia and schizoaffective disorder (P ¼ 0.003, P ¼ 0.02, and P ¼ 0.00003) for allele distribution and (P ¼ 0.002, P ¼ 0.04 and P ¼ 0.0004) for genotype distribution. Several haplotypes were also associated, the most significant of which was a combination of the 9625-G/A, 7255-A/G, 4884-A/G, 4155-C/G, 1873-C/T and 563-C/T SNPs. (LRT ¼ 78.36, d.f. ¼ 26, P < 0.000001). Even after multiple testing, the association with the 9625-G/A and 563-C/T SNPs remained significant (P ¼ 0.0005 and P ¼ 0.05 for allele, P ¼ 0.02 and P ¼ 0.006 for genotype, respectively) providing evidence that genetic alterations in the promoter region of the a1Aadrenoceptor might contribute to schizophrenia by influencing expression levels.

80

Abstracts

P7.25 CAN A SNP ON SPLICING VARIANT OF HTR4 PREDICT POST PSYCHOTIC DEPRESSION OR ANXIETY? Yamanouchi Y,1 Iwata N,1 Suzuki T,1 Ikeda M,2 and Ozaki N2 1 Departmentof Psych. Fujita Health University, School of Medicine, Toyoake, Japan 2 Nagoya University Graduate School of Medicine, Nagoya, Japan Post psychotic depression and anxiety are obstacles during schizophrenic treatment, caused to incompliance and suicide. 5-HT4 were reported had relation with mood disorder and anxiety. On the other hand, antipsychotics cause post psychotic depression and risperidone may not have affinity with 5-HT4 receptor. Therefore we expect risperidone treated schizophrenic patients can pure evaluate post psychotic depression or anxiety without drug action. 94 Japanese schizophrenic patients were received written informed consent. They were evaluated positive and negative syndrome scale (PANSS) and isolated genomic DNA after eight weeks risperidone single dosage treatment. Two candidate HTR4 SNPs were selected and genotyped PCR-RFLP and primer extension methods referred to Suzuki, et al. On each several symptomatic factors,were extracted from PANSS scores by principal component analysis, two SNPs were compared to each genotype using multiple regression analysis. Analyzing results were corrected due to multiple analysis. SNP on splicing variant associated with depression/anxiety factor (F ¼ 6.2, corrected P ¼ 0.03, statistical power ¼ 0.88). The other SNP or other factors did not have association. This SNP may be able to be biological marker that predict post psychotic depression or anxiety. Nevertheless this conclusion have weaknesses about multiple analysis and symptom evaluation method. Moreover the function of this SNP and pharmacological confirmations will be needed.

P7.26 NO ASSOCIATION BETWEEN GABAA RECEPTOR GENE CLUSTER ON 5Q AND JAPANESE SCHIZOPHRENIA Ikeda M,1 Iwata N,2 Suzuki T,2 Yamanouchi Y,2 and Ozaki N1 1 Nagoya University Graduate School of Medicine, Nagoya, Japan 2 Fujita Health University, Toyoake, Japan Several lines of evidence suggest that abnormalities of gamma amino butyric acid (GABA) neurotransmission systems are considered to be the candidate factors related to pathophysiology of schizophrenia. GABAA receptor genes cluster on 5q31-35 (composed of b2(GABRB2), a6(GABRA6), a1(GABRA1), and g2(GABRG2) subunit genes) is one of the most attracting candidate regions of schizophrenia-suscepitibility gene due to these products’ abundance in brain, pharmacological profiles, and results from positional cloning. In this study, we performed the first fine linkage disequilibrium (LD) mapping within this cluster after systematic polymorphism search of GABRB2, GABRA6, and GABRA1, in addition to our colleague’s study of GABRG2. Then we picked up and genotyped ‘representative SNPs’ for 288 Japanese schizophrenics and 288 controls. We found unique LD patterns in GABRA1 and positive association of SNP and haplotypes in promoter of GABRA1 by conventional genotypic and haplotypic analyses. However, we could not support this positive result by adjustments of correlations between markers and multiple testing. Furthermore we confirmed this preliminary positive association was false positive from new haplotypic analysis based on genealogical estimation of ancestral haplotypes and molecular biological method. These results suggest that this cluster may not play a major role in Japanese schizophrenia, and it is important to consider especially multiple testing. Our empirical data raise alerts of preliminary positive genetic association without fine LD mapping, consideration of problems in genetic statistics and confirming biological evidence.

P7.27 LACK OF ASSOCIATION OF THE PROMOTER REGION OF THE MONOAMINE OXIDASE-A GENE (MAOA-UVNTR) WITH VIOLENT BEHAVIOR IN SCHIZOPHRENIA Urraca N,1 Fresa´n A,1 Camarena B,1 Aguilar-Garcı´a A,1 Apiquian R,1 De la Fuente-Sandoval C,2 Meyenberg N,2 Nicolini H,2 Escamilla M,3 Raventos H,3 Medina R,3 Mendoza R,3 Mun˜oz R,3 Ontiveros A,3 Can˜ive J,3 and Jerez A3 1 National Institute of Psychiatry, Mexico, Mexico 2 Caracci Medical Group, Mexico, Mexico 3 The Genetics of Schizophrenia in Latino Populations Group, San Antonio, United States

The monoamine oxidase A (MAO-A) is a degradative enzyme that catalyzes the deamination of serotonin and norepinephrine. Studies of aggressive behavior have implicated altered metabolism of serotonin and to a lesser extent dopamine and noradrenaline. Based on these findings MAO-A gene has been proposed as a candidate gene involved in violent behavior. It has been postulated that allele 1 or 4 of the MAOA-uVNTR gene polymorphism are associated with interindividual variability in aggressive disposition. The aim of the present research was to examine the association between the MAOA-uVNTR gene polymorphism and violent behavior in a schizophrenic sample of Mexico City. Method: We recruited 46 schizophrenic patients and their parents. Diagnoses were based on the Spanish version of the DIGS. A score of 7 in the total score of the Overt Aggression Scale was used as a cutoff point to classify the patients as violent (n ¼ 31, 67.4%) or nonviolent (n ¼ 15, 32.6%). From the sample, a total of 30 trios (violent ¼ 11, 36.6%, non-violent ¼ 19, 63.4%) were obtained for TDT analysis. The MAOA-uVNTR polymorphism was genotyped by DNA amplification from genomic DNA according to conditions previously described. Results: There were no differences in allele 1 frequency distribution between violent (n ¼ 15, 34.1%) and non-violent patients (n ¼ 6, 33.3%) (P ¼ 0.95). TDT analysis provided no indication for preferential transmission in allele 1 in violent trios (P ¼ 0.15). Conclusion: Our results do not support the association between low transcription alleles of the MAOA-uVNTR with violence in schizophrenia. Confirmatory studies are needed using larger samples to increase the power of the study.

P7.28 ALLELIC VARIATION OF GENES ON 7Q22 IN FINNISH SCHIZOPHRENIA FAMILIES Peltonen JO,1 Ekelund J,1 Loukola A,1 Silander K,1 Paunio T,1 Varilo T,1 Partonen T,2 Lo¨nnqvist J,2 and Peltonen L1 1 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland 2 Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland The prevalence of schizophrenia in Finland is similar to other populations, excluding some regional isolates with an exceptionally high frequency of familial schizophrenia. We have carried out nationwide collection of families with several affected members, implying a high genetic load of schizophrenia. In these families we have performed several genome wide scans, both in study samples collected from the whole population and from an internal subisolate. Our linkage studies have revealed evidence of loci on chromosomal regions 1q, 4q, 5q, and 7q. Such findings are in line with the linkage data from other populations: Several genetic studies have shown evidence of linkage to numerous different loci, but only some of these findings have been replicated in different study samples and only very few of them have resulted in the initial identification of a specific allelic variant associated with the trait. The current study focuses on the characterization of the allelic diversity and potential transmission disequilibrium of four candidate genes on 7q22: GRM3, RELN, SEMA3A, and VGF. The initial evidence for linkage on 7q22 was obtained in a sib pair analysis of 52 Finnish families, the multiallelic marker D7S486 produced a two point lod score of 3.18 under dominant model. The evidence for linkage in this region was increased in a further study of 175 families (Ekelund et al., submitted) with a two point lod score of 5.29. Genome-wide scans in other populations have not revealed a linkage to this same locus. However, some regional genes on 7q have been associated with schizophrenia, making this particular genome area worth of additional analyses. The regional candidate genes are located over a 20 cM region of 7q22, and have been selected not only based on their location, but also on their known function, and reported suggestive evidence of association. Our study sample consists of 1338 individuals from 279 Finnish schizophrenia families, which significantly expands our previous study demonstrating linkage to 7q22. We use Sequenom’s MassARRAY technology to genotype intragenic SNPs, with an even distribution and sufficient information content (the minor allele frequency over 5%). To establish the LD and haploblocks in the Finnish population, the SNPs and constructed haplotypes are first analyzed in 60 Finnish trios to establish major haplotypes and haploblock boundaries. Typically 3–9 haplotypes with a minor allele frequency >3% can be constructed and MDL BlockFinder (Koivisto et al., 2002) predicts 1–4 major haploblocks in Finnish families. Our aim is to characterize the contribution of these genes and their specific alleles in the genetics predisposition of schizophrenia in this nationwide family collection.

Abstracts P7.29 CONVERGENT FUNCTIONAL GENOMICS, ASSOCIATION AND LINKAGE ANALYSIS SUGGESTS 20 , 30 -CYCLIC NUCLEOTIDE 30 -PHOSPHODIESTERASE (CNP) AS A POTENTIAL SUSCEPTIBILITY GENE FOR SCHIZOPHRENIA Peirce TR,1 Bray NJ,1 Williams N,1 Haroutunian V,2 Buxbaum J,2 Buckland P,1 Owen M,1 and O’Donovan M1 1 Dept of Psychological Medicine, UWCM, Cardiff, United Kingdom 2 Mount Sinai School of Medicine, New York, United States Reduced brain expression of 20 , 30 -Cyclic nucleotide 30 -phosphodiesterase (CNP) has been reported in patients with schizophrenia at both the mRNA and protein level. CNP-deficient mice display neuropathological features such as enlarged ventricles that are consistent with observations made in schizophrenic patients. Moreover, we have recently reported significant genome-wide evidence for linkage to schizophrenia in a single pedigree to 17q21.2, a region of the genome to which CNP maps. In order to screen indirectly for cis-acting sequence variants in CNP that might alter its expression, and that might account for the original finding of altered expression in schizophrenic brains, we have measured the relative expression of mRNA originating from the paternal and maternal copies of CNP in brain mRNA from 25 informative individuals. A transcribed allele of CNP showed highly significant differences in its expression (average expression difference between alleles ¼ 31%, P < 0.0001). In keeping with previous studies showing reduced CNP expression in schizophrenic brain, the allele that is consistently under-expressed in this assay was associated with schizophrenia in a case (n ¼ 709) control (n ¼ 710) sample (P ¼ 0.03). In the family showing linkage to 17q21.2, all six affecteds were homozygous for the under expressed allele, a finding that while not readily explained by chance (P ¼ 0.02), falls short of convincingly implicating CNP as a susceptibility gene for schizophrenia. We have followed up these findings by sequencing the complete CNP genomic sequence in eighteen schizophrenic individuals including four individuals from the linked family. All SNPs have then been genotyped in 96 subjects in order to determine a comprehensive LD map of the gene. The non-redundant SNPs are currently being genotyped at a level that will extract 95% of the available haplotype information by haplotype analysis, and we are currently performing further analyses of CNP expression in an extended sample of brains from schizophrenic patients and controls.

P7.30 MOLECULAR CHARACTERIZATION OF A PATIENT WITH UNUSUAL 22Q11 DELETION AND SLIGHT PSYCHIATRIC PHENOTYPES Zhang Y, Murphy BC, Siu V, Fan YS, O’Reilly R, and Singh SM University of Western Ontario, London, Canada Microdeletion of 22q11 is associated with Velocardiofacial syndrome (VCFS). Apparently 30% of these patients have a 1.5 or 3 Mb deletion and show psychiatric abnormality, particularly schizophrenia. Deletion mapping of novel patients with 22q11 deletion and schizophrenia has the potential to further narrow our search for candidate genes involved in this common disease. Clinical report: This study deals with a patient with VCFS and slight psychiatric symptoms. This 17year-old young man is the second of two children born to nonconsanguineous parents at 36 weeks gestation weighing 6 pounds. Pregnancy was uncomplicated with no known exposure to teratogens. Delivery was by Cesarean section following prolonged labour and breech presentation. In early infancy there was frequent nasopharyngeal reflux. Speech delay was evident by 3 years of age. He began walking at 16 months. Velopharyngeal insufficiency and submucous cleft palate were identified at nasal endoscopy. He underwent pharyngoplasty. Carotid arteries were noted to be medially deviated. Cytogenetic and molecular study: Fluorescence in situ hybridization allowed confirmation of a 22q11 deletion using the DNA probe N25 (Oncor). Molecular analysis of the proximal and distal breakpoints was determined by semi-quantitative PCR using 40 primer sets specific to the 22q11 region. The proximal breakpoint was established between RH48663 and RH 48348, while the distal breakpoint was between D22S1138 and SHGC-145314. The size of the deletion in this patient is a novel 2.3 Mb. Here although the proximal breakpoint coincides with the LCR common in 3 Mb and 1.5 Mb deletions, the distal breakpoint represents a novel site. This hemizygous deletion spans a region containing 65 genes, including the popular schizo-

81

phrenia candidate gene COMT. At COMT Val/Met genetic polymorphic site, this patient has a Met allele. In a region of COMT’s putative promoter region, a cytosine in a CpG dinucleotide is unusually fully methylated. Interestingly, both the genetic and epigenetic alteration may theoretically lower the COMT protein activity. Conclusion: These results support the suggestion that LCRs play an important role in causing breaks and genes contained within the deleted area remain prime candidates for genetic and epigenetic studies. Supported by Schizophrenia Society of Ontario and Ontario Mental Health Foundation.

P7.31 ANALYSIS OF GENETIC VARIATION AND EXPRESSION OF THE DYSTROBREVIN BINDING PROTEIN GENE (DTNBP1) IN SCHIZOPHRENIA Schwab SG,1 Mondabon S,2 Knapp M,2 Albus M,3 Borrmann-Hassenbach M,3 Lerer B,4 Hallmayer J,5 Maier W,2 Schmitt A,6 Timmermann B,7 Hoehe MR,7 Reinhardt R,7 and Wildenauer DB1 1 University of Western Australia, Nedlands, Australia 2 University of Bonn, Bonn, Germany 3 State Mental Hospital, Haar, Haar, Germany 4 Hadassah Medical School, Jerusalem, Israel 5 University of Stanford, Palo Alto, United States 6 ZI fuer Seelische Gesundheit, Mannheim, Germany 7 Max-Planck-Institut fuer Molekulare Genetik, Berlin, Germany Single nucleotide polymorphism marker within the dystrobrevin binding protein (DTNBP1) gene has been repeatedly reported to be associated with schizophrenia. However, since different markers as well as different alleles have been involved in association, allelic heterogeneity has been suggested. We have reported association of a set of markers with schizophrenia in two independent family samples (Schwab et al., 2003, Am J Hum Genet 72:185–190) one sample comprising 79 families with at least two affected siblings and both parents available for genotyping and the second sample comprising 125 triad families (affected offspring with parents). None of the reported markers are located in exons or are known to cause functional or structural changes. Sequence analysis has been initiated to analyze the genomic region for further variants. 96 patients from our family sample have been included in this analysis. Until now, 79 variants have been identified, 29 of them having a minor allele frequency of >0.10 in the analyzed sample. So far, no exonic variation has been detected in the investigated region. In addition, we have started expression analysis using real time technology for relative quantification of RNA levels. 62 RNA samples isolated from immortalized lymphoblast cell lines from the patients of our family samples and RNA samples isolated from post mortem brain of 10 schizophrenic patients and 10 controls were available. Overall, no differences in RNA levels of DTNBP1 have been identified, however, if genotype information was used as a co-variant in the statistical analysis, preliminary results indicate statistical significant differences for the expression levels of DTNBP1 in the lymphoblasts of our patients.

P8.1 LINKAGE OF SCHIZOPHRENIA TO CHROMOSOME 8P21-12 IN KOREAN SCHIZOPHRENIA FAMILIES WITH AUDITORY HALLUCINATION Hong KS,1 Lee YS,2 Cho EY,1 Choi KS,3 Jang YL,4 Cho SH,5 Jeun HO,1 and Kim JW1 1 Department of Psychiatry, Sungkyunkwan University School of Medicine, Samsung Seoul Hospital, Seoul, Korea, South 2 Yong-In Mental Hospital, Yong-in, Korea, South 3 Department of Psychiatry, Eulji University School of Medicine, Daejeon, Korea, South 4 National Chuncheon Hospital, Chuncheon, Korea, South 5 Department of Laboratory Medicine, Sungkyunkwan University School of Medicine, Samsung Seoul Hospital, Seoul, Korea, South Several genomewide linkage scans with Caucasian families have identified chromosome 8p21-12 as a susceptibility locus for schizophrenia. The purpose of this study is to investigate the linkage of this locus to schizophrenia in Korean multiplex families with schizophrenia. Seven microsatellite markers of this region with 3cM intervals were genotyped for 146 family members, ninety of whom were affect-

82

Abstracts

ed individuals. Highest non-parametric linkage (NPL) score (1.51, P ¼ 0.06), was observed at D8S1769 which is located at 0.35Mb upstream of the first exon of Neuregulin 1. To identify genetically more homogenous subgroup of schizophrenia, we evaluated the intrafamilial correlation for various clinical subtypes and symptoms of schizophrenia. We performed the same non-parametirc linkage analysis in each subgroup of families classified by the symptoms or subtypes which showed significant correlations among the relative pairs. For the twenty-three families in which all of the affected individuals showed auditory hallucination, statistically significant allele sharing was observed at D8S1769 (NPL ¼ 2.18, P ¼ 0.015). This study supports the previous evidence from Caucasian families for a locus predisposing to schizophrenia at 8p21-12. It also suggests that the susceptiblilty gene at this locus might contribute to the development of the auditory hallucination in schizophenia.

P8.2 CONFIRMATION OF ASSOCIATION AT THE MRDS1 (OFCC1) LOCUS IN JAPANESE SCHIZOPHRENIA SAMPLES Yamada K,1 Detera-Wadleigh SD,2 Iwayama-Shigeno Y,1 Takao H,1 Toyota T,1 Hattori E,1 and Yoshikawa T1 1 Lab. for Molecular Psychiatry, RIKEN Brain Science Institute, Wako-City, Saitama, Japan 2 Mood and Anxiety Disorders Program, National Institute of Mental Health, Bethesda, MD, United States Reanalysis of 262 Irish schizophrenic pedigrees suggested that the MRDS1 gene located at 6p24.3 is a novel susceptibility gene for schizophrenia [Straub et al., 2003, Am J Med Genet 122B, p18]. This follows from a previous study that identified an association between DTNBP1 at 6p22.3 and schizophrenia in the same family samples [Straub et al., 2002, Am J Hum Genet 71, 337–348]. We tested the hypothesis of association between the MRDS1 gene and schizophrenia using family-based association analyses and a case-control approach in Japanese cohorts. We analyzed 124 nuclear families, 360 cases and 360 controls with 25 SNPs encompassing the gene. Five SNPs, and surprisingly, haplotypes constructed from 18 SNPs that span a 279 kb region of the gene showed significant transmission distortion in a family-based association study (P ¼ 0.0017 for SNP, P ¼ 0.0019 for haplotype). In addition, the results for three individual SNPs as well as multilocus haplotypes were successfully replicated in case-control samples (P ¼ 0.012 for SNP, P ¼ 0.017 for haplotype). In our Japanese samples, we found that the boundary of linkage disequilibrium associated with a genetic risk for disease differed from, but overlapped with that of European samples. In contrast to Irish families that showed an association of minor haplotypes with schizophrenia, in Japanese families, common variants were associated with disease. Our findings highlight the existence of a genetic variation within MRDS1 that confers a susceptibility to schizophrenia in more than one ethnic group, and suggests that the associated variants may be located within a haplotype block shown in this study.

P8.3 NO ASSOCIATION FOR POLYMORPHISMS IN THE PROMOTER REGION OF DRD4 IN JAPANESE SCHIZOPHRENIA Nakajima M,2 Yamada K,1 Iwayama-Shigeno Y,1 Aoki M,1 Takao H,1 Toyota T,1 Hoshika A,2 and Yoshikawa T1 1 Laboratory for Molecular Psychiatry, RIKEN Brain Science Institute, Wako-City, Saitama, Japan 2 Department of Paediatrics, Tokyo Medical University Hospital, Shinjuku, Tokyo, Japan Studies suggest that polymorphisms of the dopamine D4 receptor gene (DRD4) may be implicated in the pathogenesis of schizophrenia, although many conflicting results have been reported. In metaanalysis, there is evidence for significant association between the 521C/T promoter variant and schizophrenia. A functional role for this variant has been determined by transcriptional analysis. However, ethnic heterogeneity of allele frequencies is a confounding factor and therefore further independent association studies are required. The aim of this study is to investigate the possible involvement of the DRD4 gene in the etiology of Japanese schizophrenia, by analyzing polymorphisms in the promoter region. We performed a mutation screen of the promoter region using 30 unrelated Japanese schizo-

phrenics, confirming ten reported SNPs and identifying a novel but rare SNP, 929C/G. We failed to detect five previously reported SNPs, probably due to their rarity in the Japanese population. Genotyping of 7 common SNPs (1105T/C, 905T/C, 808G/A, 615G/C, 602G/T, 521T/C and 376C/T) and three repeat polymorphisms (120 bp tandem duplication in the 50 upstream region, 12 bp repeat in exon 1 and 48 bp VNTR in exon 3) was performed on 124 Japanese schizophrenic pedigrees consisting of 373 subjects. We examined linkage disequilibrium structures within the gene and assessed associations using the pedigree disequilibrium test (PDT). The analyzed polymorphisms were in relatively weak linkage disequilibrium to each other, with no significant transmission in our samples. Consequently, our analysis does not indicate a predisposing role for the DRD4 gene in Japanese cohorts.

P8.4 AN ASSOCIATION STUDY BETWEEN DYSTROBREVIN BINDING PROTEIN 1 (DTNBP1), REGULATOR OF G-PROTEIN SIGNALLING 4 (RGS4) AND DISRUPTEDIN-SCHIZOPHRENIA 1 (DISC1) AND SCHIZOPHRENIA IN SCOTTISH POPULATION Zhang F,1 Sarginson J,1 Crowbie C,1 Walker N,4 Li T,2 Collier D,2 SinClair M,1 Yates P,3 Shaw D,1 and St. Clair D1 1 Aberdeen University, Aberdeen, United Kingdom 2 Institute of Psychiatry, London, United Kingdom 3 Aberdeen Royal Infirmary Hospital, Aberdeen, United Kingdom 4 Ravens Craig Hospital, Greenock, United Kingdom Recent researches have suggested the variants in several candidate genes increase the susceptibility to schizophrenia. The results to date for each gene have not been consistent among different populations examined. Dystrobrevin binding protein 1 (DTNBP1), Regulator of Gprotein signalling 4 (RGS4) and Disrupted-In-Schizophrenia 1 (DISC1) are three genes that have been reported with significant results in various populations. We try to investigate the potential association between DISC1, DTNBP1 and RGS4 and schizophrenia in 626 Scottish Schizophrenia case and 600 control samples. We genotyped polymorphic markers spanning all three genes. We genotyped six SNPs and three microsatellites in DISC1, eleven SNPs in DTNBP1 and four SNPs in RGS4. The SNP genotyping was carried out by using Amplifour and microsatallites genotyping was carried out by GENESCAN on an ABI 310. We have found significant association between Schizophrenia and three genes, the respective results are shown. DTNBP1: For eleven SNPs, there are no significant results on allele frequencies. But twolocus, three locus and four locus haplotype analysis produced the numerous significant results (P < 0.05 and the best P < 0.00001). RGS4: For four SNPs, three of them are significant on allele frequencies (P < 0.05). Two locus and three locus haplotype results are also significant. DISC1: For all nine markers, two of them show significant results on allele frequencies (P < 0.005). Four and six locus haplotypes also show the significant results. Our results suggest that these three genes are implicated in the pathogeneses of schizophrenia in Scottish population. We are now examining these genes for gene-gene interactions both with each other and Neuregulin-1 (NRG).

P8.5 METHYLATION AND VAL/MET POLYMORPHISM OF COMT GENE IN PATIENTS WITH SCHIZOPHRENIA Murphy BC, O’Reilly R, and Singh SM University of Western Ontario, London, Canada Schizophrenia is a complex multi-factorial disorder involving environmental and genetic determinants, but neither of these factor(s) has been clearly established. One of the genes of particular interest in schizophrenia is the COMT gene on chromosome 22q11, but with contradictory results. We screened 51 blood samples (20 schizophrenia patients DSM IV criteria and 31 matched controls) for cytosine DNA methylation in the COMT promoter as well as the Val/Met polymorphism that has been implicated in schizophrenia. The analysis included the polymorphism, methylation specificity and the resulting haplotype.Cytosine DNA Methylation: We assessed the cytosine methylation in the S-COMT promoter using sodium bisulfite followed by direct sequencing of the PCR product. Of 68 cytosines present in this region only the cytosines within CpG dinucleotides were methylated. Of the

Abstracts th

6 CpG’s present 5 of them were fully methylated. The 6 CpG site was found to show individual specific variability in the degree of methylation. Of the 51 samples tested, an unusual complete methylation was found in one schizophrenia patient with prominent negative schizophrenia symptoms and a Val/Met genotype. All other samples irrespective of the disease showed a pattern that approaches 66% methylation and 33% unmethylation at this site-specific cytosine. Val/ Met Polymorphism: The schizophrenia and control data followed Hardy-Weinberg equilibrium (P ¼ 0.05, df ¼ 1). The genotypic frequencies between the control and schizophrenia groups were not significantly different (calculated w2 ¼ 3.002 < critical w2 ¼ 5.99, P ¼ 0.05, df ¼ 2). As well, the allelic frequencies of Val and Met were not significantly different between the 2 groups (calculated w2 ¼ 2.91 < critical w2 ¼ 3.84, P ¼ 0.05, df ¼ 1). Although the Met allele was higher in the schizophrenia group a larger sample size is required to determine its relevance. The site-specific epigenetic modification (cytosine DNA methylation) being reported here for the first time is expected to affect the level of expression of this gene and thereby COMT protein activity. Further, given that the difference in COMT protein activity is associated with the Val/Met polymorphism and the possibility that methylation could affect gene expression, the results included in this presentation favour studies reflecting functional aspects rather than the reliance on markers and polymorphisms.

P8.6 ASSOCIATION STUDY OF NEUREGULIN 1 GENE AND SCHIZOPHRENIA IN A POPULATION ISOLATE FROM THE NORTH OF SPAIN Mata I,1 Arranz MJ,2 Staddon S,2 Beperet M,1 Lopez-Ilundain JM,1 Murray RM,2 and Kerwin RW2 1 Fundacion Argibide, Pamplona, Spain 2 Institute of Psychiatry, London, United Kingdom Neuregulin 1 is a cytoskeletal synaptic protein which plays a central role in neural development, most likely involved in regulating synaptic plasticity, or in how the brain adapts to the environment. It also appears to have a marked impact on the expression of several neurotransmitter receptors, including NMDA glutamate receptors. Neuregulin 1 gene is localized in chromosome 8p12-21, region where suggestive linkage to schizophrenia has already been shown. A ‘‘core haplotype’’ of this gene has been associated with schizophrenia in Iceland and Scotland, while a different haplotype has also been associated with schizophrenia in a Han Chinese population. In the present study we investigated the possible association between two microsatellites in the neuregulin 1 gene and schizophrenia in a sample of 103 patients and 106 control subjects from a population isolate in the north of Spain (Navarra) and in a replication sample of 53 patients and 68 controls from surrounding areas. For the first microsatellite (Nr2 ¼ 478B14-848) a significantly different allele distribution was found between schizophrenia patients and control subjects, both in the isolate population (X2 ¼ 24.0, P < 0.0005), and in the replication sample (X2 ¼ 13.3, P ¼ 0.02). For the second microsatellite (Nr3 ¼ 420M91395), a trend towards association was observed in the isolate population (X2 ¼ 13.6, P ¼ 0.06), reaching marginal statistical significance in the replication sample (X2 ¼ 13.8, P ¼ 0.03). A larger sample is being compiled to test the validity of these results, but overall, our findings provide further support for the implication of the neuregulin 1 gene in the aetiology of schizophrenia.

P8.7 SCHIZOTYPAL TRAITS IN THE RELATIVES OF SCHIZOPHRENIA PATIENTS: DOES THE GENDER OF THE PATIENT INFLUENCE ITS EXPRESSION? Lopez-Ilundain JM,1 Mata I,1 Beperet M,1 Albeniz A,1 Perez-Nievas F,1 and Murray RM2 1 Fundacion Argibide, Pamplona, Spain 2 Institute of Psychiatry, London, United Kingdom The finding of a higher rate of schizotypal personality disorder (SPD) in the relatives of schizophrenia patients when compared to the relatives of control subjects, led to the conceptualisation of SPD as a phenotypic expression of familial-genetic liability to schizophrenia. Although a number of family studies have shown a higher familial liability for schizophrenia among the relatives of female probands than among those of male probands, a re-analysis of the Iowa 500 and non-500

83

family studies found that relatives of male schizophrenia probands had a significantly increased risk of SPD as well as a significantly higher risk of expressing flat affect than relatives of male schizophrenia probands. These findings suggest that males could be at a higher risk of a genetic subtype of schizophrenia with a lower penetrance for psychosis than other subtypes, but expressed among relatives as rather schizotypal or negative symptoms. In the present study, a sample of 152 first-degree relatives (43% males) of 58 unrelated schizophrenia probands (70% males) were assessed with the Schizotypal Personality Questionnaire (SPQ). This scale gives a total score, as well as a score for each of its three factors (cognitive-perceptual, interpersonal, disorganized), and each of its nine sub-scales (which are the same as the nine SPD symptoms listed in DSM-IV criteria). We found that, regardless of their own gender, relatives of male schizophrenia probands had higher scores on the SPQ total scale (P ¼ 0.03) and interpersonal factor (P ¼ 0.01), as well as on the social anxiety (P ¼ 0.01) and constricted affect (P ¼ 0.04) sub-scales. These findings support the notion, mentioned above, that males could be at a higher risk of a genetic subtype of schizophrenia expressed among relatives as mainly negative schizotypal symptoms.

P8.8 NO EVIDENCE OF ASSOCIATION BETWEEN POLYMORPHISMS AT THE SEROTONIN TRANSPORTER GENE (SLC6A4) AND SCHIZOPHRENIA Preece AC, Williams NM, Norton N, Williams HJ, O’Donovan MC, and Owen MJ Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom Altered serotonergic neurotransmission has been proposed to play a role in the pathogenesis of schizophrenia (Meltzer, Br J Psych Suppl. 8, 1989). The serotonin transporter has a major role in regulating extracellular 5-HT levels and appears to be the site of action of some antidepressants (Ramamoorthy et al., PNAS USA, 90, 1993). The serotonin transporter gene (Slc6A4, OMIM #182138) has been mapped to chromosome 17q11.1 q12, a region of the genome in which we have recently reported genome wide significant linkage to schizophrenia. Slc6A4 has two well characterised polymorphisms: a 44 bp insertion/ deletion located in the promoter region which has been reported to be functional and a variable number of tandem repeat (VNTR) located in intron 2. Both polymorphisms were genotyped in a large UK casecontrol sample of 709 DSM-IV schizophrenia cases and 710 matched controls. Neither polymorphism was significantly associated with schizophrenia (w2 ¼ 0.05, P ¼ 0.82, 1 df and w2 ¼ 0.21, P ¼ 0.9, 2 df respectively). Sequence analysis of the linked family revealed no rare non-synonymous changes that might be responsible for the linkage finding. Our data provide no support that the Slc6A4 locus is associated with schizophrenia, or is responsible for our linkage to 17q.

P8.9 IDENTIFICATION OF SCHIZOPHRENIA GENES IN AN ANIMAL MODEL FOR PSYCHOSIS Hartmann AM, Giegling I, Beissbarth T, Schmidt E, and Rujescu D Division of Molecular and Clinical Neurobiology, Department of Psychiatry, University of Munich, Munich, Germany The psychotomimetic effects of noncompetitive N-methyl-D-aspartate (NMDA) receptor antagonists such as PCP and ketamine in healthy humans and their ability to exacerbate several psychotic symptoms in schizophrenic patients have promoted a view of schizophrenia as being related to an altered glutamatergic neurotransmission. Attempts to mimic these effects in rats has lead to the recognition of parallels between schizophrenia and molecular, cellular, functional and behavioral abnormalities in animals chronically treated with NMDA receptor antagonist MK801 in a low dosage. We performed cDNA arrays analyses comparing the expression of 1175 genes between MK801 treated rats and saline treated controls in order to identify candidate genes contributing to schizophrenia. We found several genes to be differentially expressed in cortex and in cerebellum. Our findings demonstrate that a functional genomics approach can be applied in the identification of new and unexpected candidate genes for psychosisrelated traits. These genes are currently under investigation in our genetic analyses.

84

Abstracts

P8.10 NO EVIDENCE FOR ASSOCIATION BETWEEN COMT GENOTYPES AND SCHIZOPHRENIA IN TWO LARGE SAMPLES Williams HJ,1 Glaser B,1 Williams NM,1 Zammit S,1 Mcgregor S,1 Murphy KC,2 Kirov G,1 Owen MJ,1 and O’Donovan MC1 1 UWCM, CARDIFF, United Kingdom 2 Royal College of Surgeons in Ireland, DUBLIN, Ireland A functional valine/methionine polymorphism in the COMT gene has been proposed to influence susceptibility to schizophrenia, particularly in individuals of European ancestry. More convincingly, so has a functional COMT haplotype in Ashkenazi Jews. We have attempted to replicate each of these findings. We genotyped more than 2500 individuals from two relatively homogeneous large schizophrenia association samples, a case (N ¼ 709) control (N ¼ 710) sample from the UK and a family based association sample (N ¼ 488) from Bulgaria. We also typed a sample of VCFS probands with (N ¼ 12) and without (N ¼ 32) schizophrenia. None of the three samples showed evidence for association with schizophrenia for any COMT marker or haplotype. SNP rs737865 UK P ¼ 0.99, Bulgarian P ¼ 0.803 and VCFS P ¼ 0.24. SNP rs165688 UK P ¼ 0.99, Bulgarian P ¼ 0.75 and VCFS P ¼ 1.00. SNP rs165599 UK P ¼ 1.0, Bulgarian P ¼ 0.086 and VCFS P ¼ 0.30. For the schizophrenia risk haplotype described by Shifman and colleagues (2002) (1) UK P ¼ 0.78. Bulgarian P ¼ 0.96 and the VCFS P ¼ 0.45. In Europeans, the high risk haplotype described by Shifman and colleagues (2002) (1) does not increase susceptibility to schizophrenia, nor, despite its considerable attractions as a candidate, does the valine allele of COMT. Ethnic variation in the linkage disequilibrium structure at COMT means that our haplotype data may not generalize beyond outbred people of European ancestry. However, as the valine allele is itself the proposed susceptibility risk factor, with a particularly strong effect in Europeans, no such caveat applies.

P8.11 GENETIC ASSOCIATION OF PPP3CC WITH SCHIZOPHRENIAASSOCIATION BETWEEN SCHIZOPHRENIA AND PPP3CC GENE IN A CAUCASIAN SAMPLE Thierfelder K, Giegling I, Mu¨hlenhoff L, Mu¨ller J, Bettecken T, Meitinger T, Mo¨ller H-J, and Rujescu D 1 Division of Molecular and Clinical Neurobiology, Department of Psychiatry, University of Munich, Munich, Germany 2 Institute of Human Genetics, GSF, Neuherberg, Germany Calcineurin (protein phosphatase 2B) is a heterodimeric Ca2þ/ calmodulin-dependent serine/threonine protein phosphatase composed of CNB regulatory and CNA catalytic subunits. Forebrainspecific CNB knockout mice display behavioral abnormalities related to altered behaviors observed in schizophrenia patients. A recent study in a large sample of affected families detected association of the PPP3CC gene, which encodes the calcineurin gamma catalytic subunit, with schizophrenia. The aim of this study is to replicate the association of variations in the PPP3CC gene in an independent case-control sample of Caucasian origin. Calcineurin is a calcium-dependent serine/ threonine protein phosphatase and plays a significant role in the central nervous system. Its activity plays a key role in the downstream regulation of dopaminergic signal transduction and in the induction of certain forms of NMDA receptor-dependent synaptic plasticity. There is the hypothesis that calcineurin function is involved in dopaminergig and glutamatergig signaling. It is assumed that alterations affecting signaling could comprise a contributing factor in schizophrenia pathogenesis. Several calcineurin-related genes were already investigated. Association could only be identified in the PPP3CC gene, a positional canditate gene, located at 8p21.3, which encodes the calcineurin g catalytic subunit. The aim of this study is to verify and replicate sequence variations in PPP3CC gene associated with susceptibility to schizophrenia in Caucasian individuals. A sample of 368 schizophrenic patients according to DSM IV (295.x), and 368 community-based healthy volunteers entered the study. A single centre, case-pedigree-design was chosen. About 368 patients are ascertained. The case group includes persons, diagnosed with Schizophrenia (DSM IV: 295.x). The diagnosis wasnclusion criteria were interviewed verified by SCID (Structural clinical Interview of Disorders). We analysed several SNPs (e.g. rs2449348, rs2461491, rs7821470) and haplotypes in the PPP3CC gene. A control-group of 368, including healthy persons, also was interviewed. For the genetic

analysis a subset of SNPs were investigated. A PCR amplification was accomplished and polymorphisms were genotyped. Nonparametric analysis methods were used. The frequencies of alleles and genotypes on particular gene marker between cases and controls were compared by Chi-Square statistics. A subset of SNPs (e.g. rs2449348, rs2461491, rs7821470) and haplotypes in PPP3CC gene were analysed. Preliminary results analyses show a trend of an associationconfirm the association with schizophrenia in our sample. The next step is conducting quantitative measures to search for association between the phenotype and gene marker and understanding the mechanism of calcineurin function in the schizophrenia pathogenesis.

P8.12 EVIDENCE FOR SCHIZOPHRENIA (SC) SUBTYPES IN THE CENTRAL VALLEY OF COSTA RICA (CVCR): LD ANALYSIS OF CHROMOSOMES 8, 13, 18 AND 22 USING THE MANIA SYNDROME TO SUBDIVIDE THE PHENOTYPE OF SC Walss-Bass C,1 Escamilla MA,1 Raventos H,2 Montero AP,2 Medina R,1 Contreras S,1 Armas R,3 Dassori A,1 Liu W,1 Levinson D,4 Pereira R,2 Balderas TG,1 Pereira M,2 Atmella I,2 NeSmith L,1 Castro R,2 Salazar R,2 Hernandez M,2 Saenz J,2 Garbarz JJ,5 Katz I,6 Leach R,1 and Almasy L7 1 University of Texas Health Science Center San Antonio, San Antonio, United States 2 Universidad de Costa Rica, San Jose, Costa Rica 3 Langley Porter Psychiatric Institute, University of California, San Francisco, United States 4 University of Pensylvania School of Medicine, Pensylvania, United States 5 Department of Mental Health, San Francisco, United States 6 San Francisco General Hospital, University of California, San Francisco, United States 7 South West Foundation for Biomedical Research, San Antonio, United States The long-standing concept that SC and bipolar disorder (BP) represent two distinct illnesses has been challenged by findings of overlap of genetic susceptibility loci for these two diseases. To address this issue, we hypothesized that the presence or absence of mania in the course of chronic psychotic illness might be a marker for differences in the set of contributory susceptibility loci. To test this hypothesis, we have performed a linkage disequilibrium analysis, using micro-satellite markers, of four chromosomes that have been previously reported to harbour overlapping susceptibility loci for both SC and BP (8, 13, 18 and 22). We used a sample of 105 SC patients from the isolated founder population of the CVCR. Analysis of the overall sample showed evidence of association (P < 0.05) on all four chromosomes: Ch 8 (8p22, 8p12, 8q13.3 and 8qter), Ch 13 (13p11.2 and 13q33.2), Ch 18 (18p11.3, 18q12.3, 18q23qter) and Ch. 22 (22p11.1). Stratification of the sample by the presence or absence of mania showed clear segregation of association in several of the chromosomal regions. The regions of 8p12, 8q13.3, 18p11.3, and 18q23qter appear to be associated with non-manic psychosis. The regions of 13p11.2, 13q33.2, 18q12.3 18q21.33, 18q22.3 and 22p11.1 appear to be associated with manic psychosis. Two of these regions, 18q21.33 and 18q22.3, did not appear to be associated with the overall SC sample, but where associated with the SC sample in which symptoms of mania were present. Furthermore, these two regions plus the regions of 13q33.2 and 18q12.3, all of which are associated with manic psychosis, have also been reported to be associated with BP. Our results provide evidence for the existence of SC subtypes and point to etiologic overlap between the subtypes in which manic symptoms are present and BP. Furthermore, our results suggest that the use of the mania syndrome to refine the SC phenotype may provide a powerful methodology for identifying genes which underlie the psychotic spectrum of psychiatric illnesses.

P8.13 META-ANALYSIS SHOWS ASSOCIATION BETWEEN SEROTONIN TRANSPORTER GENE INTRON 2 VNTR AND SCHIZOPHRENIA Fan JB and Sklar P 1 Department of Psychiatry, Harvard Medical School, and Psychiatric and Neurodevelopmental Genetics Unit, Massachusetts General Hospital, Charlestown, MA, United States 2 Broad Institute of Harvard and MIT, Cambridge, MA, United States

Abstracts The serotonin transporter gene (SLC6A4) has long been considered a promising candidate gene for schizophrenia based on serotonin transporter’s crucial role in serotonergic neurotransmission. However, association studies have produced conflicting results regarding the association between two common SLC6A4 gene polymorphisms, the promoter insertion/deletion (5-HTTLPR) and the intron 2 VNTR (STin2 VNTR) polymorphisms, and schizophrenia susceptibility. To elucidate the putative association between the two common SLC6A4 gene polymorphisms and schizophrenia more comprehensively, we performed a meta-analysis based on all original published association studies between the two polymorphisms and schizophrenia up to April 2004. Our analyses showed no statistically significant evidence for the association between the Short allele of the 5-HTTLPR polymorphism and schizophrenia (pooled odds ratio (OR) ¼ 0.99, 95% CI ¼ 0.92–1.07, Z ¼ 0.23, P ¼ 0.82) from 19 population-based association studies consisting of 2,959 cases and 3,875 control subjects. However, highly significant evidence for association between the STin2.12 allele of the STin2 VNTR polymorphism (Pooled OR ¼ 1.21, 95% CI ¼ 1.08–1.36, Z ¼ 3.25, P ¼ 0.0012) was found from 12 population-based association studies consisting of 2,162 cases and 2,458 control subjects. Our metaanalysis suggests that the STin2.12 allele of the STin2 VNTR polymorphism is likely a risk factor for schizophrenia susceptibility. Our data imply that following completion of the International HapMap Project a comprehensive evaluation of a set of markers that fully characterize the linkage disequilibrium relationships at the SLC6A4 gene should be tested in large, well-characterized clinical samples in order to understand the role of this gene in schizophrenia susceptibility.

P8.14 POSSIBLE ASSOCIATION OF THE GAMMA-AMINOBUTYRIC ACID B RECEPTOR 1 (GABBR1) GENE IN SCHIZOPHRENIA Zai G,1 King N,1 Wong G,1 Barr CL,2 and Kennedy JL1 1 Department of Psychiatry, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada 2 Department of Psychiatry, The Toronto Western Hosptial, University of Toronto, Toronto, Canada Schizophrenia (SCZ) is a severe neuropsychiatric illness and genetic factors are believed to be important etiologically. Although historically genetic studies have examined dopamine and other neurotransmitter systems, there is increasing evidence that the major inhibitory neurotransmitter, gamma-aminobutyric acid (GABA), may also be functionally involved. Furthermore the GABAB receptor 1 (GABBR1) gene has been localized to the chromosome 6p21.3 region, which has shown linkage to SCZ. We investigated five polymorphisms (7265A/G substitution, 10497C/G substitution, 33795A/G substitution in the 30 UTR, Ser491Ser-a T to C transition at position 1473, Phe659Phe-a T to C transition at position 1977) in the GABBR1 gene, in a sample of 101 DSM-IV SCZ probands and their families, 150 unrelated affected individuals matched individually for gender, age, and ethnicity with 150 healthy controls. We performed the transmission disequilibrium test (TDT) and case-control analysis. We did not observe biased transmission of alleles in any of the polymorphisms individually or haplotypes within the gene, to SCZ probands. However, a weak significant difference was observed for the allele frequencies (chisquare ¼ 4.310, P ¼ 0.038) and a trend was observed for the genotype distributions (chi-square ¼ 4.970, 2 df, P ¼ 0.083) between SCZ individuals and controls. The observed trends suggest that further investigations of the role of the GABBR1 gene in SCZ are warranted.

P8.15 NMDA RECEPTOR SUBUNIT NR2B GENE (GRIN2B) IS ASSOCIATED WITH SCHIZOPHRENIA IN A CASE CONTROL SAMPLE Martucci L, De Luca V, Trakalo J, and Kennedy JL Centre for Addiction and Mental Health, Toronto, Canada The ability of N-methyl-D-aspartate receptor (NMDAR) antagonists to induce a syndrome closely resembling schizophrenia suggests that dysfunction or dysregulation of NMDA receptor mediated neurotransmission occurs in schizophrenia. The NR2B subunit appears to be critical for a number of the basic structural and functional attributes associated with the NMDA receptor. Several investigators have specifically been looking at the role of the NR2B subunit in

85

schizophrenia. We studied three polymorphisms located on GRIN2B: G200T, located in the 50 UTR, the A5806C and T5988C, located in the 30 UTR. Our sample consisted of 200 unrelated patients and 200 controls, matched by age, gender, and ethnical background. All probands met the DSM-IV diagnostic criteria for schizophrenia. Differences in the allele frequencies between the patients and healthy controls were tested using the chi-square association test. Haplotype distributions in our case-control sample were obtained using Helix Tree. The results are as follows: for the G-200T marker: w2 ¼ 14.121, P ¼ 0.0027. For the A5806G marker: w2 ¼ 3.258, P ¼ 0.196. For the T5988C marker: w2 ¼ 0.691, P ¼ 0.708. The results were corrected for multiple testing. The haplotype analysis showed a preferential distribution of the G C C haplotype (P ¼ 0.031). These results support the hypothesis of a role of GRIN2B in the pathogenesis of schizophrenia. The promoter region appears to be of most interest. P8.16 ROLE OF DISC1, A POTENTIAL CANDIDATE GENE PRODUCT FOR SCHIZOPHRENIA, IN THE NEUROTRANSMITTER SYSTEMS Ogawa F, Kohu K, and Akiyama T Laboratory of Molecular and Genetic Information, Institute of Molecular and Cellular Biosciences, The University of Tokyo, Tokyo, Japan Disrupted-In-Schizophrenia 1 (DISC1) was identified as a potential candidate gene for schizophrenia. DISC1 is disrupted by a balanced (1,11)(q42.1,q14.3) translocation segregating with schizophrenia and related psychiatric disorders in a large Scottish family. Molecular and functional analyses of the gene product of DISC1 will contribute to understanding the molecular basis of schizophrenia and other major mental illnesses. To elucidate the function of DISC1, we attempted to identify DISC1-associated proteins using the yeast two-hybrid system. We screened a human brain library using the full-length DISC1 as bait, and obtained several cDNA fragments encoding potential partners for DISC1, including NUDE, NUDEL and ATF4/5. Furthermore, we isolated proteins that are involved in neurotransmission. We confirmed the interaction between DISC1 and these proteins by in vitro and in vivo pull-down assays and immunostaining experiments. In addition, although mouse DISC1 possesses low-homology to human DISC1, mouse DISC1 was also found to interact with these proteins. Our findings are consistent with previous reports showing that DISC1 is a multifunctional protein, and involved in microtubule dynamics, intracellular transport, neurite outgrowth and signal transduction. Of particular interest is the fact that DISC1 is associated with proteins involved in neurotransmission, since schizophrenia is known to be associated with dysfunctions of several neurotransmitter systems such as dopaminergic, glutamatergic, serotoninergic and GABAergic systems.

P8.17 EXPANSION OF DISRUPTED IN SCHIZOPHRENIA 1 (DISC1) NETWORK OF PROTEIN INTERACTIONS IMPLICATES DISC1 IN BOTH THE GLUTAMATE AND NEURODEVELOPMENTAL HYPOTHESES OF SCHIZOPHRENIA Camargo LM,1 Brandon NJ,1 Schurov I,1 Whiting PJ,1 Bonnert TP,1 Colloura V,2 Rian JC,2 Hermkajakob H,3 kerrien S,3 Montecchi-Palazzi L,4 and Cesareni G4 1 Merck Sharpe & Dohme, Harlow, United Kingdom 2 HYBRIGENICS, Paris, France 3 European Bioinformatics Institute, Hinxton, United Kingdom 4 University of Rome Tor Vergata, Rome, Italy Disruption of the DISC1 gene by a balanced (1, 11) (q42,q14) translocation has been reported to co-segregate with major psychiatric illness (schizophrenia, bipolar, and other major affective disorders in a Scottish family, Millar et al., 2000). Interestingly, all carriers of the translocation have cognitive impairments independent of the presence of psychiatric phenotype (Blackwood et al., 2001). The biological corollary of expressing a truncated from of DISC1 is unclear, however several in vitro studies have shown that truncation results in the loss of the C-terminus mediated interactions with other proteins, namely to NUDEL and PAFAH1B (Lis-1) (Brandon et al., 2004; Ozeki et al., 2003). DISC1’s interactions with NUDEL and PAFAH1B, suggest a role for DISC1 in neurodevelopment. Indeed, we have recently demonstrated that the interactions of DISC1, PAFAH1B and NUDEL

86

Abstracts

are developmentally regulated (Brandon et al., 2004). Here we describe a systematic expansion of the network of DISC1 interactions by yeasttwo-hybrid screens. By screening DISC1, and selected DISC1-interactors such as NUDEL, against both human foetal and adult brain expression libraries we were able to derive a network of proteininteractions consisting of more than one hundred proteins. Through such interactions, we were able to link DISC1 and its interacting partners to several processes crucial to dendrite formation, stability, and the proper formation of synapses, namely anterograde protein transport (kinesin motor complex), cell organisation and biogenesis, cytoskeletal/microtubule processes, neurotrophic receptor signalling and the glutamate type NMDA receptor complex. Additionally, processes related to cell division, such as chromosome organization and biogenesis as well as mitotic cell division were over-represented. These putative biological roles for DISC1 are consistent with both the neurodevelopmental and glutamate hypofunction hypotheses of schizophrenia and therefore support a role for this gene in the pathophysiology of this disorder. P8.18 FURTHER GENETIC ANALYSES OF THE PCM1 GENE ASSOCIATION WITH SCHIZOPHRENIA ON CHROMOSOME 8P21 AND TESTS OF THE G72, DYSBINDIN, RGS4, CALCINEURIN, COMT, FRIZZLED 3, MRDS1, AKT1 AND CAPON ASSOCIATIONS Datta SR, Mcqullin A, Rizig MA, Thirumalai S, Pimm J, Moorey H, Quested D, Kalsi G, Bass N, Lawrence J, Choudhury K, Puri V, Curtis D, and Gurling HMD University College London, London, United Kingdom We have found that the markers D8S261, D8S2615 and D8S2616 at the Pericentriolar Material 1 (PCM1) gene locus on chromosome 8 were significantly associated with schizophrenia in a London UK and Scottish case control sample (Gurling et al., 2002). These markers were also found to predict MRI volumetric changes in the orbitofrontal cortex of subjects with schizophrenia compared to controls. We have now carried out a haplotype analysis in the London UK sample and have shown positive haplotypic association with schizophrenia with a significance of P ¼ 0.0001 using a permutation test. Attempted replication of the association in Scotland and USA samples found significant association in a USA trio sample with the marker D8S261 but not in Scotland. We have now sequenced the coding and locus control regions of the PCM1 gene in DNA from selected PCM1 associated cases. We found 15 new genetic variants or SNPs. These new SNPs have been genotyped in the original sample and in a further 300 cases and 300 controls from London UK. Further tests of association at the PCM1 locus in this enlarged sample (440 cases and 440 controls) will be reported. The enlarged sample has now also shown positive allelic or haplotypic association at the RGS4, Dysbindin-1, and G72/G30 loci. The most significant haplotype at G72 was associated with an empirical significance of P ¼ 0.0.006. Dysbindin-1 haplotypes showed strong association (P < 0.0001) also using a permutation test and RGS4 haplotypes were associated (Permutation P ¼ 0.00001). Tests of the reported associations with schizophrenia at the calcineurin (Gerber et al., 2003), frizzled 3, (Yang et al., 2003), D-amino acid oxidase, Catechol-o-methyl transferase (Shifman et al.), MRDS1 (Shifman, 2002), (Straub, 2003), AKT1 protein kinase (Emamian, 2004) and carboxyl-terminal PDZ ligand of neuronal nitric oxide synthase, CAPON (Brzustowicz, 2004) will also be reported.

P8.19 THE G72/G30 REGION IS IMPLICATED IN SCHIZOPHRENIA SUSCEPTIBILITY IN PALESTINIAN FAMILIES FROM THE NORTH PART OF ISRAEL Cholostoy A,1 Korostishevsky M,1 Kremer I,2 Kaganovich M,1 Murad I,3 Muhaheed M,3 Bannoura I,4 Rietschel M,5 Dobrusin M,6 Belmaker RH,6 Maier W,5 Bening Abu-Shach U,1 Ebstein RP,7 and Navon R1 1 Department of Human Genetics and Molecular Medicine, Tel Aviv University, Tel Aviv, Israel 2 Emek Hospital, Afula, Israel 3 Dr. Kemal Psychiatric Hospital, Bethlehem, Israel 4 The Palestinian Research Center for Genetics of Mental Disorders, Bethlehem, Israel 5 Department of Psychiatry, University of Bonn, Bonn, Germany 6 Beersheva Mental Health Center, Beersheva, Israel 7 Herzog Hospital, Jerusalem, Israel

Association of the G72/G30 region in chromosome 13q32-33 with schizophrenia was reported in several case-control studies: French Canadian, Russian and Ashkenazi populations. In an effort to establish the involvement of G72/G30 in schizophrenia susceptibility, we genotyped 223 Palestinian Arab family trios, using 11 SNPs in the vicinity of the G72/G30 genes. The families were recruited from three different regions of Israel: 56 from the North (Afula), 136—from Bethlehem (Palestine authority) and 31—Bedouins from the South (Beersheba). The patients were diagnosed according to SCID interview and medical records. None of the SNPs demonstrated transmission disequilibrium in the total of the Arab family trios tested. However when the families were divided according to place of inhabitancy, transmission disequilibrium was observed in Afula sample for the rs778294 SNP (TDT ¼ 5.59, P < 0.02). This SNP is located close to the 50 end of G30 and the 30 of G72. It is 12 kb apart from the block of the tightly linked SNPs, which was found to be associated with schizophrenia in Ashkenazi Jews. None of the SNPs demonstrated transmission disequilibrium in the Bethlehem and Beersheba samples. Our results support the involvement of G72/G30 region with susceptibility to schizophrenia in Palestinian Arabs of the North area of Israel. P8.20 ASSOCIATION AND EXPRESSION ANALYSIS OF NEUROGULIN-1 REGION: IMPLICATION TO ITS ROLE IN SUSCEPTIBILITY TO SCHIZOPHRENIA Navon R, Shamir E, Zerah G, Swartz M, Kaganovich M, and Korostishevsky M Department of Human Genetics and Molecular Medicine, Tel Aviv University, Tel Aviv, Israel Genome-wide scan studies of affected families with schizophrenia showed suggestive linkage to chromosome 8p12-22 in several different populations. Neurogulin1 (NRG1) is considered a candidate gene for schizophrenia susceptibility because of its function and location in 8p21. Recent studies on Iceland and Scottish families as well as on Chinese families implicate that NRG1 contribute to the etiology of schizophrenia. In an attempt to award more power to the role of NRG1 region to schizophrenia we genotyped five SNPs upstream the NRG1 in 60 unrelated Ashkenazi schizophrenic patients and 130 matched controls: The five SNPs were: SNP8NRG221132, SNP8NRG221533, SNP8NRG241930, SNP8NRG243177, SNP8NRG433E1006. Significant difference in the genotype frequency of SNP8NRG221533 between the patient and control groups was found (P < 0.03). Exact test of sample differentiation based on haplotype frequency estimation revealed a borderline P-value of 0.052  0.002. Expression analysis of NRG1 in DLPFC of 43 patients and 40 controls measured in a region common for the two main isoforms (type I and type II) of the gene was performed using real-time quantitative RT-PCR. The analysis disclosed an over-expression in the DLPFC schizophrenic patients (Mann-Whitney test revealed P ¼ 0.063  0.005). Our results strengthen findings on the role of NRG1 region in susceptibility to schizophrenia.

P8.21 AN ASSOCIATIVE STUDY ON DRD4 GENE POLYMORPHISMS IN PROMOTER REGION AND CLINICAL FEATURES OF SCHIZOPHRENIC PATIENTS Gritsenko IK,1 Golimbet VG,2 Abramova LI,2 Kaleda VG,2 and Ebstein RP1 1 Scheinfeld Center of Human Genetics for the Social Sciences Psychology Department Hebrew University, Jerusalem, Israel 2 Mental Health Research Center Russian Academy of Medical Sciences, Moscow, Russia Results of association studies between the polymorphic dopamine D4 receptor gene and schizophrenia are inconsistent but might be explained by an association with clinical presentations of schizophrenia rather than DSM categorical definitions of illness. To test this notion, we examined three promoter regions SNPs at 809, 616 and 521 for association with clinical symptoms of schizophrenic patients. 151 patients with ICD-10 diagnosis of schizophrenia (78 males & 73 females, mean age 34.5  12.9 years, mean age at disease onset 23.9  8.3 years, illness duration 10.6  6.1 years) were genotyped. Clinical symptoms were inventoried using the PANSS. An association was observed between 521C/T and positive symptoms (F ¼ 2.9, df ¼ 145,2, P ¼ 0.05). When patients were stratified by gender, significance (F ¼ 4.2, df ¼ 67,2, P ¼ 0.019) was confirmed only in females.

Abstracts Female patients with the TT genotype had higher scores compared to patients with the TCþCC genotypes (t ¼ 2.9, df ¼ 67, P ¼ 0.005). For 809 G/A, positive symptoms were also more pronounced in female patients with GG genotype (t ¼ 2.1, df ¼ 62, P ¼ 0.036) comparing to GA and AA subjects. We also analysed PANSS scores using the pentagonal model and tested association for both 809 G/A and 521 C/T. Female patients with the TT genotype had higher scores for Disorganization Factor (t ¼ 1.8, df ¼ 35, P ¼ 0.008) and Excitement Factor (t ¼ 2.3, df ¼ 35, P ¼ 0.02) when compared with TC & CC patients. Similarly, patients with the 521TT/809GG genotypes were found to have the highest total positive symptoms scores (t ¼ 2.9, df ¼ 25, P ¼ 0.007). They similarly differed by the PANSS items: Conceptual disorganization (t ¼ 3.8, df ¼ 25, P ¼ 0.0002), Excitement (t ¼ 3.8, df ¼ 25, P ¼ 0.001), Unusual thought content (t ¼ 3.5, df ¼ 25, P ¼ 0.002), Mannerism and posturing (t ¼ 2.9, df ¼ 25, P ¼ 0.007) as well as by Excitement Factor (t ¼ 2.5, df ¼ 25, P ¼ 0.02). In summary, we show that 521T/C and 809G/A DRD4 gene polymorphisms contribute to positive symptoms in female schizophrenic patients, especially to disorganization and excitement, which were most pronounced in subjects with the TT and GG genotypes. This work was supported, in part, by the RFRF grant N 04-04-48588.

87

the genome scans of largest sample size and whether the peaks relate to the genes currently under investigation in association studies. We selected the three studies with a sample size of over 200 families and have compared the distribution of lod scores reported by the authors relating to the peaks in each study (Figure illustrates lod scores in the three studies for chromosomes 2, 10 and 17 that include the major findings of the first two studies).

P8.22 ASSOCIATION OF LATE-ONSET SCHIZOPHRENIA WITH THE 32-NUCLEOTIDE DELETION ALLELE OF THE GENE ENCODING THE CHEMOKINE RECEPTOR 5 (CCR5) Rasmussen HB,1 Wang AG,2 and Werge T1 1 Research Institute of Biological Psychiatry, Sct. Hans Hospital, Roskilde, Denmark 2 Department of Psychiatry, Amager Hospital, Copenhagen, Denmark Cytokines are messengers of the immune system. In addition they act as neuromodulators. This has inspired a number of studies to investigate the possible roles of cytokines in major psychiatric diseases, including schizophrenia. Chemokines are proinflammatory cytokines with the ability to attract and activate leucocytes. The chemokine receptor 5 (CCR5) is capable of binding several chemokines besides being a major co-receptor for HIV. There is a common mutant allele of the CCR5 gene characterised by a lack of a 32-nucleotide segment. This allele confers protection to infection with HIV. In addition, it probably modulates the susceptibility to a variety of immune-mediated diseases. We examined whether the 32-nucleotide deletion allele of the CCR5 gene could affect susceptibility to schizophrenia. A possible role of this allele in age at onset of the illness was also investigated. Included in our study were 268 patients with F20 schizophrenia and 323 healthy control subjects. The genotype proportions in patients as well as control subjects corresponded to those expected under conditions of HardyWeinberg equilibrium (P > 0.05). The null hypothesis of no association between CCR5 genotype and patient-control status could not be rejected (P ¼ 0.16). The frequencies of the CCR5 32-nucleotide deletion allele did not differ significantly between patients and control subjects (P ¼ 0.17). Treated as continuous variable, age at first admission (AFA) of the schizophrenic patients was marginally associated with CCR5 genotype (P ¼ 0.05). Using AFA as a measure of age of onset of illness, we distinguished between onset below and onset above 40 years of age with the latter group representing late-onset schizophrenia. With this dichotomised approach we a revealed an overrepresentation of genotypes homozygous or heterozygous for the 32-nucleotide deletion allele in the late-onset group (P ¼ 0.04). The frequency of the deletion allele was 0.22 in the late-onset group and 0.13 in the other group, reflecting a relatively large difference in allele distributions between the two groups (P ¼ 0.03). Our observations implicate the CCR5 32-nucleotide deletion allele in genetic susceptibility to late-onset schizophrenia.

P8.23 ARE THE LINKAGES FOR SCHIZOPHRENIA CONSISTENT ACROSS GENOME SCANS? Crow TJ1 and DeLisi LE2 1 SANE POWIC, Oxford, United Kingdom 2 New York University, New York, United States Several genes (e.g. dysbindin, neuregulin, DISC1 & 2, COMT) have recently been claimed to have a relationship with the aetiology of schizophrenia on the basis of linkage and association studies. Two meta-analyses 1,2 have coordinated the genome scan findings. Here we ask the question of how consistent the peak findings are across

The findings show little consistency across studies and no concensus with respect to the regions that contain candidate genes. We conclude that the support for linkage from the genome scans (e.g. in the metaanalyses) is not matched by agreement across individual scans, and therefore that the positive claims from association studies (particularly those based upon multiple haplotype associations in different populations) lack the foundation in linkage that they would be expected to have if they represented real associations with disease status.

P8.24 THE HISTAMINE H1 RECEPTOR GENE POLYMORPHISMS AND SCHIZOPHRENIA Godlewska B,1 Limon J,2 and Landowski J1 1 Medical University of Gdansk, Department of Psychiatry, Gdansk, Poland 2 Medical University of Gdansk, Department of Biology and Genetics, Gdansk, Poland There is growing evidence suggesting involvement of the histaminergic system in the patophysiology of schizophrenia, especially contribution of the H1 receptors to deficit symptomps. We have therefore selected H1 receptor gene as a cadidate gene and examined its five previously described polymorphisms (17C/T, Asp349Glu, 1068A/G, Phe358, Leu449Ser) in the population of the Northern Poland. The study was performed on 204 probands (102 schizophrenic patients and 102 matched healthy controls from the general population). All schizophrenic patients met DSM-IV criteria assessed with the structured interview SCID-I. Clinical symptoms and their intensity were assessed using the Positive and Negative Syndrome Scale (PANSS). Genomic DNA was extracted from leukocytes using a standard phenol-chloroform method and the polymorphisms were analysed using PCR-SSCP and PCR-RFLP methods. Automated sequencing was performed to confirm the presence of each polymorphism. There was a weak trend towards the more frequent incidence of Leu449Ser heterozygotes (w2 ¼ 2.01, df ¼ 1, P ¼ 0.15) and Ser449 allele (w2 ¼ 2.01 df ¼ 1 P ¼ 0.15) among schizophrenic patients than in healthy controls, whereas there were no differences in genotype and allele distribution in case of other polymorphisms analysed. There was no association between the above described polymorphisms and the age of onset of schizophrenia ( T that was associated with either BD or SZ in three separate cohorts of Caucasian patients from the United States, Czech Republic and Israel. The 432 promoter variant affects an 8-base motif that resembles the core consensus of the POU/Oct-1 family of transcription factors. Electromobility gel shift assays (EMSA) show that oligonucleotides containing the 432T allele bind with greater affinity to several unknown proteins derived from brain and other tissues. The DNA/protein complexes are competitively inhibited by an oligonucleotide containing an exact POU/Oct-1 consensus. We have now extended our analysis to include zed DNA samples from African American patients SZ. Similar to our previous findings, an increase in 432T was seen in patients compared with controls. Although there was no difference detected in the overall allele frequency (‘‘T’’ allele frequency ¼ 0.25 controls and 0.26 patients), a statistically significant increase in TT homozygotes was detected in patients (frequency TT ¼ .04 controls, .13 patients; TT vs other genotypes, Fisher Statistic ¼ 5.1, P ¼ 0.024, genotype distribution, chi square test for independence, P ¼ 0.016). The SZ sample genotype distribution was not in Hardy-Weinberg equilibrium (Pearson’s goodness of fit chi square P ¼ 0.002). The data provide additional support for PIK3C3, and its regulation by members of the POU-family of transcription factors, in the pathogenesis of SZ.

TABLE III. Case control comparison of PIK3C3 432 C>T SNP

CONT SZ

CC

CT

TT

C

T

.54 (54) .60 (66)

.42 (42) .27 (30)

.04 (4) .13 (14)

.75 (150) .74 (162)

.25 (50) .26 (58)

Frequency with number of subjects in parenthesis. Chi square test for independence ¼ 8.3, P ¼ 0.016. Fisher exact test TT vs CT þ CC. Fisher Statistic ¼ 5.1, P ¼ 0.024.

HML and DFP are recipients of NARSAD Independent Investigator Awards and an award from The Juvenile Bipolar Research Foundation.

P8.32 ASSOCIATION OF SCHIZOPHRENIA IN AFRICAN AMERICANS TO POLYMORPHISM IN SYNAPSIN III GENE Lachman H,1 Stopkova P,2 Aghalar-Rafael M,1 and Saito T1 1 Department of Psychiatry and Behavioral Sciences, Division of Psychiatry Research, Albert Einstein College of Medicine, Bronx, New York 2 Psychiatric Clinic, First Medical Faculty, Charles University, Prague, Czech Republic

90

Abstracts

Linkage studies in families with schizophrenia (SZ) have pointed to chromosome 22q12-q13 as one of several regions of the genome that may contain a susceptibility gene. The gene coding for synapsin III, an intrinsic synaptic vesicle membrane protein, maps to this target region. Two tightly linked single nucleotide polymorphisms (SNP), were recently found in a small subset of patients with SZ, a nonsynonymous variant, L469L, and S470N, which results in the loss of a mitogen-activated protein (MAP) kinase serine phosphorylation site (Porton et al., Biological Psychiatry 55:118–125). We too found a slight increase in 470N in Caucasian patients with SZ from the United States. But, the sample size and allele frequency were too small draw definitive conclusions. However, the frequency of both SNPs was much higher in African American controls than Caucasian controls, thereby providing a better sample cohort to analyze for SZ involvement. For L469L, there was a 50-fold increase in the frequency of the minor allele in African Americans compared with Caucasians (minor allele frequency codon L469L SNP: African American controls ¼ 0.24; Caucasian controls U.S. ¼ 0.004; Czech Republic controls ¼ 0.005: minor allele frequency codon S470N SNP: African American controls ¼ 0.05; Caucasian controls U.S. ¼ not detected in 123 subjects; Czech Republic controls ¼ not detected in 95 subjects). Furthermore, there was an increase in the percentage of African American patients with SZ who were homozygous for the 469 SNP compared with African American controls who were homozygous (11% vs 5%, AA genotype vs GA and GG, Fisher exact statistic ¼ 3.08, P ¼ .04, one tailed, allele distribution, w2 ¼ 1.65, P ¼ 0.1 one-tailed; N ¼ 126 patients with SZ, 106 controls). An increase in 470N heterozygotes was also found, but the results fell short of being statistically significant. The functional significance of the 469 SNP has not been determined. However, the 469 SNP does create a 5/7 match for a lariat consensus, which is necessary for proper alignment of the 30 splice acceptor site which might disturb the regulation of exon 12 splicing (lariat consensus is YNYYRAY, 469 SNP is CCCCTG/AA). The findings support a role for synapsin III in a subset of African American patients with SZ and raises questions about selective pressure in Africa to account for the extraordinary disparity of the 469 and 470 SNPs between African and Caucasian populations. If these findings are confirmed, the stark differences in allele frequencies of SZ candidate genes across ethnic groups could account for ethnic differences in symptoms and response to medications. TABLE -SYNAPSIN III SNPs-AFRICAN AMERICANS L469L GENOTYPES

CONT SZ

GG

GA

AA

GG

GA

60 (.57) 66 (.52)

41 (.39) 46 (.37)

5 (.05) 14 (.11)

96 (.91) 108 (.86)

9 (.08) 17 (.13)

L469L ALLELES

CONT SZ

S470N GENOTYPES AA 1 (.01) 1 ( A and A218C polymorphisms of TPH1 and the hCV245410, rs1487280 and rs1872824 variants of TPH2 for

Abstracts association with suicide attempts in a sample of 176 subjects with onset of major depressive disorder or dysthymic disorder prior to the age of 14 (97 with attempts, 79 without). Allelic chi-square results indicated no association of the three TPH2 markers in COD probands with at least one lifetime suicide attempt: hCV245410 (P ¼ 0.92), rs1487280 (P ¼ 0.96) and rs1872824 (P ¼ 0.70). At the TPH1 locus, the 6526G > A polymorphism was not associated with lifetime history of suicide attempt (P ¼ 0.80), though the A218C variant exhibited a trend towards association (chisq ¼ 3.03, d.f. ¼ 1, P ¼ 0.082). In conclusion, we found no association between TPH2 alleles and COD subjects with suicide attempts. Rather, we found a trend for the TPH1 A218C marker to be associated with suicide attempts in this sample. Although this observation was in contrast to results expected given that TPH1 is not expressed in mammalian brain, it was convergent with several previous genetic association studies of TPH1 and suicidal behaviour phenotypes. P9.19 ANALYSIS OF DUPLICATIONS IN SAMPLES OF AUTISM USING MAPH Grataco`s M,1 Armengol LL,1 Cerrato C,1 Herva`s A,2 Sua´rez M,1 and Estivill X1 1 Centre de Regulacio´ Geno`mica, Barcelona, Spain 2 Hospital Mutua de Terrassa, Terrassa, Spain Autism is characterized by impairments in reciprocal social interaction and communication, restricted and stereotyped patterns of interests and activities, and the presence of developmental abnormalities by three years of age. Several genes are likely to be responsible for the autistic phenotype, and susceptibility loci have already been identified. 10–15% of patients with an autistic phenotype have Rett syndrome, tuberous sclerosis or fragile X syndrome. Moreover, a small proportion of autistic patients have mutations in the neuroligins 3 and 4 genes. Chromosomal rearrangements have consistently been detected in patients with autistic spectrum phenotype, mainly involving chromosome 7 and 15. In this line, additional 15q11-13 maternal copies of the Prader Willi/Angelman critical region (PWACR) produce an abnormal phenotype that is distinct from both PWS and AS and includes developmental delay, particularly with regard to speech and language, and behavioural problems sometimes falling within the autistic spectrum, with only mild or absent dysmorphic features. Because of cytogenetic abnormalities in proximal 15q in some cases of autism and evidence of linkage to this region, we have studied the variability of the genomic organization of the PWACR in patients with autism and their parents with the MAPH technology. We have designed a set of 15q-1113 test probes covering a 4 Mb region, selected from the BAC tiling path of the region. A probe for each overlapping BAC was prepared, as well as a set of control probes corresponding to unique loci in the genome and from all chromosomes. To perform a MAPH analysis, all probes were mixed and hybridized with genomic DNA of patients and their parents. Probes were further amplified with universal fluorescent primers and resolved in an automated capillary analyzer to compare area and height of picks between test and control probes. This approach should permit the study of this target region to test if there is variable number of segmental copies and to determine the extension of the polymorphic region and its relationship with phenotypic changes in patients and control samples. Supported by Psychiatric Genetics Network (Spanish Ministry of Health), Genome Spain, and Generalitat de Catalunya.

P9.20 FINE MAPPING OF CHROMOSOMAL REGIONS IMPLICATED IN ADHD Kim JW,1 Fagerness J,1 Arbeitman L,1 Doyle A,2 Petty C,2 Biederman J,2 Faraone SV,2 and Sklar P1 1 Department of Psychiatry, Psychiatric and Neurodevelopmental Genetics Unit, Massachusetts General Hospital, Harvard Medical School, Charlestown, United States 2 Department of Psychiatry, Pediatric Psychopharmacology Unit, Massachusetts General Hospital, Harvard Medical School, Boston, United States Attention deficit hyperactivity disorder (ADHD) is a common childhood-onset psychiatric disorder with strong genetic effect. Previously published linkage results from several research groups revealed candidate chromosomal regions which showed significant or suggestive evidence of linkage for ADHD. These regions are interesting because

95

they have been also implicated in other childhood-onset disorders such as autism and dyslexia. Replication of these linkage results with other independent patient samples are required in order to confirm the true loci. We used a dense set of SNP (single nucleotide polymorphism) markers chosen from the public SNP database (dbSNP) to fine map these regions. SNPs were selected for five distinct chromosomal areas and were spaced at an average inter-marker distance of 500 kb. Our patient sample consisted of 102 families with a total of 145 affected sib pairs available for analysis. SNP genotyping was done using a modified single nucleotide extension reaction with allele discrimination by mass spectrometry. Nonparametric linkage analysis of the data were performed using MERLIN (Multipoint Engine for Rapid Likelihood Inference) (Abecasis et al., 2002). Marker allele frequencies were estimated by use of the default in MERLIN, which estimates allele frequencies by counting alleles across all individuals. Five candidate chromosomal regions were screened using the SNP markers: these areas include chromosome 5p13, 7p13, 15q15, 16p13, and 17p11. Due to the dense inter-marker spacing, we were able to obtain high levels of information content in our analysis. For each chromosome, the information content was: 0.86, 0.87, 0.83, 0.83, and 0.85. Several regions yielded positive NPL scores, including chromosomal regions 7p13, 15q15 and 16p13. We did not find evidence of significant or suggestive linkage. This is likely due to a combination of small sample size and the heterogeneity of complex genetic disorders like ADHD, both of which make replication a challenge. Future studies will investigate linkage to quantitative phenotypic traits, and will include genotyping in larger samples.

P9.21 A GENETIC INVESTIGATION OF AUTISM USING EXTREMELY DISCORDANT SIB PAIRS Haddad SA,1 Fagerness J,2 Yamaki LH,2 Andresen KA,2 and Santangelo SL3 1 Harvard School of Public Health, Massachusetts General Hospital, Boston, United States 2 Massachusetts General Hospital, Boston, United States 3 Harvard Medical School, Harvard School of Public Health, Massachusetts General Hospital, Boston, United States Autism is a neurodevelopmental disorder of childhood characterized by impairments in social interaction, language development, and patterns of behavior. Models implicate anywhere from 2 to  15 susceptibility genes, none of which have yet been identified. Several studies have shown evidence for linkage to chromosomes 2q and 7q. Several reports have documented the existence of a broad autism phenotype (BAP) consisting of several sub-clinical autistic-like traits that are common in relatives of autistic children. A framework was developed for measuring the BAP that allowed for the treatment of autism as a quasicontinuous phenotype and the use of the Extremely Discordant Sib Pair (EDSP) method for linkage analysis. We ascertained families with one or more autistic children and at least one non-autistic child aged 16 or older. Unaffected siblings were assessed for broad autism phenotype (BAP) characteristics, and were included if they displayed no BAP traits. The sample is comprised of 30 extremely discordant sib pairs, 29 of which were included in the analyses. Single nucleotide polymorphism (SNP) markers at roughly 500 kilobase intervals were selected to cover a 40 cM long segment of chromosome 2q and an 85 cM long segment of chromosome 7q where prior work suggested the presence of autism susceptibility loci. Microsatellite maps were used to estimate the genetic locations of the selected SNP markers using linear regression. IBD sharing was assessed using Aspex. Family-based association analyses were run using the Whap program developed by Shaun Purcell and Pak Sham. There was no evidence for linkage to the regions examined on chromosomes 2q or 7q. Global family-based association tests also returned non-significant results. However, nominally significant allelic associations were found at three markers on chromosome 2q and at eight markers on chromosome 7q. We genotyped additional SNPs at roughly 20 kilobase intervals around two of the nominally significant markers on chromosome 7q. Family-based haplotype association tests showed evidence of association with a four-SNP haplotype in this region (empirical P-value 0.010), with two of the four SNPs comprising this haplotype showing significant individual allelic associations. We tested the effect of each SNP within this haplotype conditional on the other three markers and found that each of the four SNPs appear to be in linkage disequilibrium with the other three markers in the haplotype. Additional followup work is in progress.

96

Abstracts

P9.22 INVESTIGATION OF THE CALCYON GENE IN ATTENTION DEFICIT/HYPERACTIVITY DISORDER Laurin N,1 Ickowicz A,2 Pathare T,2 Roberts W,2 Malone M,2 Tannock R,2 Schachar R,2 Kennedy JL,3 and Barr CL1 1 Toronto Western Research Institute, Toronto, Canada 2 The Hospital for Sick Children, Toronto, Canada 3 University of Toronto, Toronto, Canada Attention-deficit hyperactivity disorder (ADHD) is a neurodevelopment disorder characterized by impaired attention, excessive motor activity, and impulsivity. Family, twin and adoption studies support a genetic component in ADHD. Pharmacology, neuroimaging, and animal models studies point toward a dopamine system dysfunction in the pathogenesis of ADHD. Positive association and/or linkage have been found for several genes in the dopamine pathways, including the dopamine receptor D1 (DRD1), suggesting that other genes involved in D1 pathways may also contribute to ADHD. The Calcyon gene encodes a brain-specific DRD1-interacting protein with a potential role in cognitive, integrative and reward processes, and in motor control. It maps to 10q26.3, a region that has shown linkage to ADHD. We thus tested Calcyon as a candidate for ADHD susceptibility. Four bi-allelic polymorphisms in the Calcyon gene (rs11101694, rs10857701, rs2298122 and rs2275723) were genotyped in 186 small nuclear families ascertained through a proband with ADHD (236 affected children). Association was assessed with the transmission/disequilibrium test (TDT), using ETDT for individual alleles and TRANSMIT for haplotypes. Quantitative trait analyses were performed using FBAT in relation to ADHD symptom scores. No significant association was found with any of these polymorphisms. However, a trend for biased transmission was observed for rs2298122 (w2 ¼ 2.65, 1 df P ¼ 0.103). In addition, one haplotype (C-A-G-G) demonstrated a marginal significant bias for non-transmission (w2 ¼ 3.88, 1 df P ¼ 0.049), and, a second haplotype (C-A-T-G) was slightly over-transmitted by heterozygous parents to affected offspring (w2 ¼ 3.08, 1 df P ¼ 0.079). Quantitative traits analyses show significant relationships between the C-A-G-G haplotype and the hyperactive/impulsive symptoms according to parents’ report (P ¼ 0.041), but not by teachers’ report (P ¼ 0.703) No relationship was found between between the CAGG haplotype and inattentive symptoms reported by either parents or teachers (P ¼ 0.137 and 0.119, respectively). These modest positive results, support the need for further studies of Calcyon gene and ADHD. Since the four markers used in these analyses are not very polymorphic in our sample, assays for three additional polymorphisms have been developed and are currently underway to increase the power of the study. P9.23 SUPPORT FOR EKN1 AS THE SUSCEPTIBILITY LOCUS FOR READING DISABILITIES AND ATTENTION DEFICIT HYPERACTIVITY DISORDER ON 15Q21 Wigg KG,1 Couto JM,1 Feng Y,1 Crosbie J,2 Anderson B,2 Cate-Carter T,2 Tannock R,2 Lovett M,2 Humphries T,2 Kennedy JL,3 Ickowicz A,2 Pathare T,2 Roberts W,2 Malone M,2 Schachar R,2 and Barr CL1 1 The Toronto Western Hospital, Toronto, Canada 2 The Hospital for Sick Children, Toronto, Canada 3 The Centre for Addiction and Mental Health, Toronto, Canada Reading disabilities (RD) have been linked to a number of chromosomal regions including 15q. Recently a gene in the 15q region, EKN1, was identified via a translocation breakpoint in a family segregating with RD. This gene was further supported in that study as a susceptibility locus by association studies in a small sample of families from Finland. This gene, if confirmed, would be the first gene identified as contributing to reading disabilities. In our sample of 148 families identified through a proband with reading difficulties, we found evidence for allelic association of polymorphisms in EKN1 to the phenotype of RD identified as a categorical trait marker (w2 ¼ 5.586, 1 d.f., P ¼ 0.018) and to the reading component processes of phonological awareness, word identification, decoding, and spelling, and to the reading-related phenotypes of rapid automatized naming, expressive and receptive language ability, verbal short-term and working memory, using quantitative analysis. Our findings, therefore, further support this locus as contributing to reading disabilities and as contributing to important reading-related processes. However, association was observed with different alleles and haplotypes than those reported to be associated in the Finnish sample. Because of the evidence from twin

studies for shared genetic factors contributing to RD and attentiondeficit hyperactivity disorder (ADHD), particularly inattention symptoms, we investigated the relationship of genetic markers in this gene to ADHD and ADHD symptoms in a sample of 186 nuclear families collected through a proband with ADHD. We found a trend for linkage to the ADHD phenotype as a categorical trait using the transmission disequilibrium test, and significant evidence for biased transmission of a haplotype containing this marker (w2 ¼ 6.926, 1 d.f., P ¼ 0.009). Using quantitative analysis, we observed evidence for association of this haplotype to the inattention and hyperactive/impulsive symptom dimensions as reported by parents, and inattention symptoms as reported by teachers, as well as a trend for association with the reading phenotypes of word identification and decoding. Our results provide preliminary support for the role of the EKN1 locus in RD and ADHD, suggesting that this locus may be pleiotrophic, contributing to both phenotypes. P9.24 THE INVOLVEMENT OF GABA-A RECEPTOR GENES IN CONDUCT DISORDER AND OTHER BEHAVIORAL PHENOTYPES IN CHILDREN Dick DM,1 Bierut L,1 Fox L,1 Bucholz K,1 Kramer J,2 Kuperman S,2 Hesselbrock V,3 Schuckit M,4 Porjesz B,5 Nurnberger, Jr. J,6 Xuei X,6 Foroud T,6 and Edenberg HJ6 1 Washington University at St. Louis, St. Louis, United States 2 University of Iowa, Iowa City, United States 3 University of Connecticut, Farmington, United States 4 University of California-San Diego, San Diego, United States 5 SUNY Downstate Medical Center, Brooklyn, United States 6 Indiana University, Indianapolis, United States The Collaborative Study on the Genetics of Alcoholism (COGA) is a long-term research program designed to identify genetic factors that contribute to the development of alcoholism and related disorders. We have recently published evidence of significant associations between two GABA-A receptor genes, GABRA2 and GABRG3, and alcohol dependence in the COGA sample. In addition to this genetically informative adult sample, COGA has collected information on children ages 7–17 from these alcoholic families. These children were interviewed using the Child Semi-Structured Assessment for the Genetics of Alcoholism (C-SSAGA), and for a subset of these children, one of the parents was also interviewed about the child’s behavior. This child sample makes it possible to test hypotheses about how genes predisposing to alcohol dependence in adulthood may be involved in related behavioral phenotypes across different periods of development. For example, previous research has documented a significant association between childhood conduct disorder and adult alcohol dependence. Twin studies have suggested that this overlap is largely due to shared genetic factors, indicating that some of the genes that influence alcohol dependence in later life may be contributing to conduct disorder in childhood and adolescence. Accordingly, we have genotyped in the sample of COGA children a subset of the snps in the GABA-A receptor genes that were associated with alcohol dependence in the adults, to test for association with related behavioral phenotypes earlier in development. Preliminary analyses suggest that GABRA2 is associated with conduct disorder in children, with the risk of conduct disorder among individuals who carry a copy of the high risk allele approximately double that of individuals who carry no copies of the high risk allele. Additionally, using survival analysis techniques, we find suggestive evidence that GABRA2 may also be associated with the age of onset of alcohol and drug dependence among adolescents. These analyses demonstrate the utility of using a developmental perspective to understand genetic contributions to behavioral outcome across the lifespan.

P9.25 A POTENTIAL ROLE FOR BRAIN DERIVED NEUROTROPHIC FACTOR (BDNF) IN ADULT ADHD Lanktree M, Muglia P, Squassina A, Krinsky M, Jain U, Macciardi F, and Kennedy JL Neurogenetics Section, Centre for Addiction and Mental Health, Toronto, United Kingdom Attention Deficit Hyperactivity Disorder (ADHD) is a common psychiatric disorder with a large genetic component. It has been shown that ADHD symptoms persist into adulthood in 30 to 60% of children diagnosed with ADHD. Adult ADHD shows an increased prevalence in

Abstracts families, possibly due to a greater genetic liability than the childhood form. Neuroimaging and pharmacological data indicate that the dopamine system plays a crucial role in the pathophysiology of ADHD. The Brain Derived Neurotrophic Factor (BDNF) is a neurotrophin involved in survival, differentiation, and synaptic plasticity of several neuronal systems including the dopaminergic system and is expressed throughout the life of the neurons. In the present study we tested the association between six BDNF polymorphisms (five SNPs and one STR) and ADHD. The sample consisted of 119 trios comprised of an adult ADHD proband and their biological parents and an independent sample of 121 ADHD cases and a corresponding number of sex, age and ethnically matched controls. Haplotype Relative Risk (HRR) was conducted generating ‘‘pseudo’’-controls to combine with the casecontrol sample. Association was found at each marker (P < .05) with the strongest linkage at the STR (P < .005). Quantitative trait analysis was performed comparing core symptom Wender and BADDS mean scores across genotype groups. No significant differences were found in symptom scores between the genotype groups within cases. Subjects who met the criteria for depression were removed from the study, however, we examined the affective symptom scores on the BADDS scale to determine if our positive results were artifacts of comorbid depression and no association was found. Strong linkage disequilibrium was found among all our markers (D0 > .85), therefore only two haplotypes were observed with a frequency greater than 10%. Of the two common haplotypes, Haplotype Trend Regression (HTR) analysis of the six polymorphisms showed one was a risk haplotype and one was a protective haplotype (P < 0.005). While BDNF has been strongly implicated in affective disorders (Strauss et al., 2004; Sklar et al., 2002; Neves-Pereira et al., 2002) no previous studies have investigated for the presence of association between ADHD and BDNF. Our results suggest that the BDNF gene may play a role in increasing the susceptibility to ADHD. P9.26 EVIDENCE FOR LINKAGE AND ASSOCIATION TO CHROMOSOME 19P IN AUTISM McCauley JL, Olson LM, Amin T, Delahanty R, Haines JL, and Sutcliffe JS. Center for Human Genetics Research, Vanderbilt University, Nashville, TN, United States Autism is a complex genetic neuropsychiatric disorder in which affected individuals have life-long disabilities in three core phenotypic domains: language, social relationships, and patterns of compulsive and stereotyped behaviors. Several genomic screens in multiplex autism families have identified numerous loci across the genome, and a few of these (e.g. 7q, 2q) have been replicated across multiple studies. We recently completed a genomic screen in 144 multiplex autism families and identified suggestive evidence for linkage on chromosome 19p13 exhibiting a multipoint heterogeneity LOD (HLOD) of 2.55 at 40 cM using a recessive model. To refine linkage information and narrow the linked interval, we genotyped and analyzed single nucleotide polymorphism (SNP) markers at 1 cM intervals flanking the peak marker D19S930. In addition to follow-up linkage analysis resulting in a narrowed peak, association analysis in 178 combined multiplex and trio families showed that a marker 5 kb 50 to the CRSP7 gene revealed significant association (P ¼ 0.001) using the Pedigree Disequilibrium Test (PDT). CRSP7 is subunit 7 of a cofactor required for Sp1 transcriptional activation. To date we have examined 18 SNPs across a 163-kb region inclusive of CRSP7. Several nominally-significant single marker associations have been identified which lie in one large (>80kb) linkage disequilibrium block as measured by D’. This block spans the CRSP7, SLC35E1, and CHERP genes. Although potential biological relevance of these genes to autism is unclear, they are positional candidates that may play a role in autism susceptibility.

P9.27 SPEECH AND LANGUAGE CHARACTERISTICS ASSOCIATED WITH A 7/13 TRANSLOCATION INVOLVING FOXP2 Tomblin J,1 Shriberg L,2 Murray J,1 Patil S,1 and Williams C3 1 University of Iowa, Iowa City, United States 2 University of Wisconsin-Madison, Madison, United States 3 University of Florida, Gainsville, United States Mutations in the FOXP2 gene have recently been found in a large family (KE) in the U.K. who have significant speech and language

97

problems. The FOXP2 gene has been found to be a gene coding for a transcription factor that has been shown to be expressed in mouse during neural development. The affected individuals of the KE family have been described as presenting developmental apraxia of speech. To date, only one other instance of a disruption of FOXP2 has been found. This study will describe the speech, language, and nonverbal intellectual skills of two family members with a 7/13 translocation that affects FOXP2. A mother and daughter were administered a speech, language, and cognitive protocol. Cytogenetic studies showed that these individuals had a chromosome 7/13 translocation and FISH studies revealed that the breakpoint on chromosome 7 was either in or immediately adjacent to FOXP2. Speech measures emphasized a range of simple single word tasks in association with demanding speech and nonspeech oral production tasks. Language measures consisted of standardized receptive and expressive tasks and a nonstandarized grammatical judgment task. Nonverbal IQ was measured using the Picture Completion, Block Design and Coding. Speech: Both the mother and daughter presented speech that was marked with errors in both speech sound production and in several domains of prosody and voice Language: The language skills of the mother and daughter revealed receptive and expressive vocabulary in the low normal range (Standard Scores 79-83). However, their ability to accurately repeat sentences and make grammaticality judgments was very poor. Nonverbal IQ: Performance on the Picture Completion and Block Design were in the low to normal range. In contast, coding performance was poor for each of these individuals. Two individuals with involvement of the FOXP2 gene demonstrated marked impairments of speech and grammatical aspects of language while showing low normal to near normal levels of vocabulary and nonverbal IQ. This profile of performance is similar to that described in the KE family. However, unlike the involvement reported for affected members of the KE family the motor speech impairment appears to include elements of both apraxia and ataxic dysarthria. P9.28 INVESTIGATION OF DNA VARIANTS IN THE 50 REGULATORY REGION OF THE DOPAMINE RECEPTOR D1 GENE IN ATTENTION-DEFICIT/HYPERACTIVITY DISORDER Misener VL,1 Capano L,1 Luca P,1 Roberts W,2 Malone M,2 Tannock R,2 Schachar R,2 Ickowicz A,2 Kennedy JL,3 and Barr CL3 1 Toronto Western Research Institute-University Health Network, Toronto, Canada 2 The Hospital for Sick Children, Toronto, Canada 3 Centre for Addiction and Mental Health, Toronto, Canada Attention-deficit/hyperactivity disorder (ADHD) is a prevalent childhood-onset psychiatric disorder characterized by age-inappropriate and impairing levels of inattention, hyperactivity and impulsivity. ADHD has a strong genetic basis, and since medications effective in the treatment of ADHD are known to influence the dopamine system, many molecular genetic studies have focused on dopamine system genes as candidates for involvement. Among these, our laboratory has recently reported evidence for association between ADHD and the dopamine receptor D1 gene (DRD1). Using the transmission/disequilibrium test to analyse the inheritance of four DRD1 polymorphisms in a large sample of ADHD families, we found significant evidence for association of a particular haplotype, designated Haplotype 3. Since none of the marker alleles comprising Haplotype 3 (D1P.5 (1251 HaeIII) Allele 1 [G], D1P.6 (800 HaeIII) Allele 1 [T], D1.1 (48 DdeI) Allele 1 [G] and D1.7 (þ1403 Bsp1286I) Allele 2 [C]) is predicted to alter DRD1 function, we hypothesized that a functional DRD1 variant, conferring susceptibility to ADHD, is associated with this haplotype. As a first test of this hypothesis, we carried out a sequencing screen of the DRD1 coding region, focusing on the ADHD-affected children who had shown preferential transmission of Haplotype 3. No DNA sequence variations were identified, indicating that the putative Haplotype 3-associated risk variant resides outside of the coding region of the gene. In the present study, we have expanded this search for a functional variant, by investigating the 50 regulatory region of the gene. A 2468 bp region, extending upstream of the DRD1 translational start site (þ1), was screened by direct sequencing of PCR products amplified from genomic DNA of children who had shown preferential transmission of Haplotype 3. Among 41 children analysed, two were found to be heterozygous for novel DNA variants. One was found to have a T to C change at position 2115, and a T to C change at 2031. The other was found to

98

Abstracts

have a C to T change at 1785. Interestingly, a search for potential transcription factor binding sites, using the MATCHTM program, showed that the 2115C and 2031C variants might alter transcription factor binding, thus possibly affecting DRD1 expression. Since the newly identified variants are present in only two of the children selected for analysis on the basis of Haplotype 3 transmission, none can be the putative Haplotype 3-associated risk variant related to our original linkage finding. However, given that at least two of these DNA changes might alter gene expression, and that allelic heterogeneity in ADHD is possible, we have developed PCR/restriction enzyme-based genotyping assays to continue to investigate these variants as potential risk factors for ADHD in our sample of ADHD families.

P9.29 INVESTIGATION OF IL-1 SYSTEM GENES AS RISK FACTORS FOR CHILDHOOD-ONSET MOOD DISORDERS Misener VL,1 Luca P,1 Wigg K,1 King N,2 Vetro A,3 Kiss E,3 Katay L,3 Kennedy JL,2 Kovacs M,4 and Barr CL1 1 Toronto Western Research Institute-University Health Network, Toronto, Canada 2 Centre for Addiction and Mental Health, Toronto, Canada 3 Szeged University, Szeged, Hungary 4 University of Pittsburgh, Pittsburgh, United States Pro-inflammatory cytokines, such as Interleukin-1 (IL-1), act on the brain to induce a set of behavioural changes known as ‘‘sickness behaviour,’’ a syndrome characterized by fatigue, sleep alterations, anhedonia, psychomotor retardation, social withdrawal and impaired cognitive functioning. Notably, these symptoms overlap considerably with symptoms typically seen in depression, and thus the question as to whether IL-1 activity plays a causal role in ‘‘endogenous’’ depression is currently of great interest. A number of findings support a role for cytokine activity in depression, including the depressogenic effects of cytokine therapy in cancer patients, increased incidences of depression in patients with chronic inflammation, and evidence of dysregulated cytokine activity in depression patients even in the absence of overt inflammatory disease. Given that genes are strongly implicated in the etiology of depression, a logical extension of the cytokine theory of depression is the hypothesis that variations in cytokine system genes, leading to dysregulated cytokine activity, confer risk for the disorder. To test this, we are investigating IL-1 system genes as candidates for involvement in childhood-onset mood disorder (COMD), a highly familial and heritable form of depressive disorder. Currently, our study sample consists of 183 nuclear families identified through probands with onset of depression in childhood, 31 of whom also have affected siblings. Two genes have been analysed so far. The first is IL1B, which encodes IL-1b, one of two IL-1 isoforms. The second is IL1RN, which encodes the IL-1 receptor antagonist, IL-1Ra, a negative regulator of IL-1 activity. For IL1B, we analysed a C/T single nucleotide polymorphism (SNP) located at position 511 in the promoter region. Notably, the T allele of this SNP has recently been found to be associated with depressive symptoms in Alzheimer’s disease and schizophrenia. Consistent with those findings, using the transmission/disequilibrium test (TDT) we found a trend for biased transmission of the T allele from heterozygous parents to their affected children (91 times transmitted vs. 72 times not transmitted, P ¼ 0.137) By contrast, TDT analysis showed no evidence for biased transmission of the common alleles of an intronic 86-bp VNTR polymorphism of the IL1RN gene (4-repeat allele: 68 times transmitted vs. 70 times not transmitted, P ¼ 0.865, 2-repeat allele: 71 times transmitted vs. 64 times not transmitted, P ¼ 0.547). Based on our findings for IL1B 551(T), we are currently investigating a functional promoter region variant at position 31 that is reported to be in linkage disequilibrium with this allele. Additional polymorphisms of IL1RN, as well as other IL-1 system genes, including those for IL-1a and the IL-1RI and IL-1RII receptors, are also being tested for association with depression in our sample.

P9.30 EARLY ONSET ANTISOCIAL BEHAVIOUR IN CHILDREN WITH ADHD IS ASSOCIATED WITH A FUNCTIONAL VARIANT IN THE CATECHOL-O-METHYLTRANSFERASE (COMT) GENE Fowler TA,1 Thapar A,1 Langley K,1 Turic D,1 Whittinger N,1 Van den Bree M,1 Aggleton J,2 Owen M,1 and O’Donovan M1

1

University of Wales College of Medicine, Cardiff, United Kingdom Cardiff University, Cardiff, United Kingdom

2

Neuroimaging, head injury and neuropsychological research all suggest that prefrontal cortical functioning is implicated in antisocial behaviour. Degradation by the enzyme COMT is the main mechanism for removal of dopamine in the prefrontal cortex. A functional polymorphism in COMT (Val158Met) has been found to be associated with prefrontal cortical functioning (PFC). As there is evidence that childhood onset conduct disorder with ADHD has a stronger neurobiological aetiology than broader classes of conduct disorder, we examined whether this type of antisocial behaviour is associated with the functional COMT variant. Participants were 227 British Caucasian children, referred from clinics, who met DSM-IV criteria for ADHD or ICD-10 criteria for Hyperkinetic Disorder. Assessments for ADHD and conduct disorder were undertaken using the Child and Adolescent Psychiatric Assessment- Parent version (CAPA). Child onset (13 years or less) DSM-IV conduct symptoms were coded as present or absent and summed to yield a total antisocial symptom score. Analytical method was regression. The high activity Val allele was significantly associated with conduct disorder symptoms (P ¼ 0.006) and conduct disorder (P ¼ 0.017). This association was independent of the effects of age, gender and verbal and performance IQ. These findings add to the evidence that this subtype of childhood onset antisocial behaviour accompanied by ADHD has neurobiological origins and that prefrontal cortical dopamine mechanisms are involved.

P9.31 GENETIC STUDY OF A DE NOVO TRANSLOCATION IN AUTISM Neves-Pereira M, Massie D, Williams J, Miedzybrodzka Z, and St Clair D University of Aberdeen, Aberdeen, United Kingdom Autism is the most common of the pervasive developmental disorders and is characterized by impairments in reciprocal social interaction, repetitive and ritualistic behaviors and the presence of development abnormalities. The prevalence of Autism has been estimated to be 5.5/ 10,000 and it is more prevalent in boys than girls in a ratio of 4 to 1. Autism is almost certainly a multifactorial disorder with environmental and genetic components. Monozygotic concordance rate approaching 90% are observed and several chromosomal regions thought to harbor candidate genes are reported using linkage studies. However because of genetic heterogeneity and reduced fecundity in autism it is likely that linkage studies will only be able to detect a proportion of disease loci. Chromosomal anomalies associated with unusual phenotypes offers an alternative strategy for identifying predisposing genes. We have identified a de novo balanced translocation 46,XY,t(4:5)(q25,q33) in a 6 year old boy with an unambiguous diagnosis of severe autistic disorder assessed using standardized diagnostic research instruments. We have prepared metaphase preps using both lymphocytes from fresh blood and from a lymphoblastoid cell line. We performed FISH using Bacterial Artificial Chromosomes (BAC) containing the insert DNA mapping to the region. BAC probes were labeled with fluorochromes and hybridized to metaphase chromosomes. The region is narrowed depending on the relative position of the two distinct fluorochrome-labelled probes per slide. In order to confirm that the general region identified by cytogenetic analysis was the correct one, we used a set of BAC flanking the band in each chromosome. The general region of the breakpoint is more proximal than initially assigned and rest now in t(4:5)(q22,q32). Further examination of the probes covering the new interval has allowed us to obtain probes flanking the breakpoint. The region of interest has been narrowed in both chromosomes to about 1 MB. In parallel we plan to use flow activated chromosome (FAC) sorted derivative DNA chromosomes to screen gridded microarrays for the bacterial artificial chromosomes that will span the translocation breakpoint. It will also allow us to walk to the breakpoints using PCR.Although these chromosomal regions have not been previously reported as strong candidates for genes responsible for Autism, we believe that it is nevertheless important to determine whether the translocation disrupts a gene not previously identified. Chromosomal breakpoints have been instrumental in assigning genes to particular cytogenetic loci and have resulted in the identification of a number of disease-causing genes. The breakpoints may directly disrupt a key gene or they may affect gene expression even at distances of up to 900 kb from the site of the breakpoint.

Abstracts P9.32 NO EVIDENCE OF ASSOCIATION BETWEEN N-METHYL D-ASPARTATE 2A (GRIN2A) GENE AND ADHD IN THE IRISH POPULATION Lowe N,1 Hawi Z,1 Kirley A,2 Mullins C,3 Sheehan K,1 Fitzgerald M,3 and Gill M2 1 Dept of Genetics, Trinity College, Dublin, Ireland 2 Dept of Psychiatry, Trinity College Dublin, Ireland 3 Dept of Psychology, Trinity College, Dublin, Ireland The N-methyl D-aspartate 2A (GRIN2A) gene, which encodes for the 2A subunit of the NMDA receptor, has recently been proposed as a possible candidate gene for attention deficit hyperactivity disorder (ADHD). Glutamate transmission and NMDA receptors have been shown to play a role in motor over activity in animal studies. The GRIN2A gene is mapped to 16p13 region, which has been implicated as a potential susceptibility locus for ADHD by a recent genome scan (Fisher et al., 2002). An association between a polymorphism located to exon five of this gene and ADHD was first reported in a UK sample (chi square ¼ 5.7, P ¼ 0.01). However, this finding was not replicated in a subsequently tested Canadian ADHD sample. We examined this marker in 179 Irish ADHD nuclear families and found no evidence of the initially reported association. A number of additional polymorphisms of the GRIN2A gene have also been examined. However, no further association with ADHD has been reported in the literature. These findings collectively follow the typical pattern of failure to replicate association findings in psychiatric disorders. However, further exploration into this gene and the glutamatergic system is warranted before concluding its role in the aetiology of ADHD. P10.1 ASSOCIATION OF MAO-A WITH ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD) IN AN IRISH SAMPLE Domschke K, Sheehan K, Lowe N, Kirley A, Mullins C, O’Sullivan R, Conroy J, Fitzgerald M, and Gill M, Hawi Z Departments of Psychiatry and Genetics, Trinity College Dublin, Dublin, Ireland Attention Deficit Hyperactivity Disorder (ADHD) is a highly heritable neuropsychiatric behavioral disorder affecting school-aged children. Imbalances in the neurotransmission of dopaminergic, noradrenergic and serotonergic systems are involved in the development of ADHD. The monoamine oxidase A (MAO-A) is known to play a key role in the modification of signal transduction in these neurotransmitter systems. Additionally, MAO-A inhibitors have been shown to be effective in the pharmacological treatment of ADHD. The MAO-A 941G allele was reported to influence MAO-A activity more efficiently than the 941T variant. This more active MAO-A variant might be responsible for the hypodopaminergic state suggested for ADHD. Previous studies have reported conflicting association results of ADHD with markers in the MAO-A gene. In the present study, we examined four markers at the MAO-A gene including the 30bp VNTR in the MAO-A promoter region, the CA(n) microsatellite in intron 2, the 941G/T SNP in exon 8 and an A/ G SNP in intron 12 in 179 Irish ADHD nuclear families. Linkage/ Association analysis of individual markers using the TDT revealed a significant association of the more active 941G allele with the disorder (TDT: 5.08, w2 ¼ 4.56, P ¼ 0.03). Additionally, a trend towards an association between the 941G allele and a positive family history for ADHD was observed (TDT: 4.33, w2 ¼ 3.69, P ¼ 0.06). Analysis of all other MAO-A markers did not yield significant results. Linkage disequilibrium measured as D’ showed that the examined markers are in significant LD with each other. However, haplotype analysis could not be performed at present due to technical problems associated with the HGMP program TRANSMIT. These data suggest a possible role of the 941G/T MAO-A polymorphism in the pathogenesis of ADHD.

P10.2 MULTI MARKER ANALYSIS OF SNAP-25 PROVIDED FURTHER EVIDENCE OF ASSOCIATION WITH ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD) IN THE IRISH POPULATION Hawi Z, Lowe N, Kirley A, Mullins C, Fitzgerald M, and Gill M Departments of Psychiatry and Genetics, Trinity College Dublin, Dublin, Ireland Animal models such as the dopamine transporter knockout mouse (DAT1-KO) and Snap-25 have provided valuable information on the

99

importance of catecholamine dysregulation in the development of ADHD. Coloboma is a semi dominant mutation (cm/þ) which was shown to be the result of a 2 cM deletion encompassing the Snap-25 gene (Maglott,1996). It has a phenotype similar to that of ADHD with 3–10 fold hyperactivity compared to the control littermates. Molecular (association/linkage) analysis conducted by Barr et al. (2000) on two SNPs mapped to the 30 untranslated region of the gene showed a biased transmission of a haplotype made of these two polymorphisms. Two other studies (Mill et al., 2002, 2004) have implicated two SNPs, a microsatellite and few Snap-25 haplotypes in the development of ADHD. In this study, we investigated 7 markers mapped to positions 2015 (A-T), 900 (G-T), (TAAA)n at intron 1, 80609 (G-A), 87610 (TG), 87614 (T-C) and 88268 (T-C) in an effort to replicate the above findings in 179 Irish ADHD nuclear families. TDT analysis of individual markers showed increased transmission of allele T of 2015, allele G of the 87610, allele T of the 87614, and allele T of 88268 to ADHD, however none of these transmissions was significant. Haplotype analysis using the program TRANSMIT showed a significant association of a haplotype made of all the examined markers. This association was further enhanced when analysis was restricted to markers 900(G-T), (TAAA)n at intron 1, 80609 (G-A), 87610 (T-G), and 87614 (T-C). Although not highly significant, these findings provided further support for the importance of Snap-25 in the predisposition to ADHD. P10.3 NO EVIDENCE OF ASSOCIATION BETWEEN MAO-B GENE POLYMORPHISMS AND ADHD IN THE IRISH POPULATION Sheehan K, Domschke K, Lowe N, Kirley A, Mullins C, O’Sullivan R, Conroy J, Fitzgerald M, Gill M, and Hawi Z Departments of Psychiatry and Genetics, Trinity College Dublin, Dublin, Ireland Attention Deficit Hyperactivity Disorder (ADHD) is a highly heritable neuropsychiatric disorder. Several studies suggest that the dopaminergic, serotonergic and noradrenergic system genes play a pivotal role in the pathogenesis of the disease. The genes coding for the monoamine oxidases A and B (MAO-A and MAO-B) are mapped to chromosome Xp11.23-22.1. They are functionally related to each other and have identical exon-intron organization as well as high sequence similarity. While MAO-A has been analyzed for association/linkage with Attention Deficit Hyperactivity Disorder (ADHD) in several studies, MAO-B was examined in only one linkage study with ADHD. MAO-B catalyzes the oxidative deamination of neurotransmitters such as dopamine and tyramine. Pharmacological studies provide preliminary evidence for a beneficial effect of monoamine oxidase B inhibitors such as deprenyl in the treatment of ADHD (Feigin et al., 1996). Administration of deprenyl has also been shown to significantly reduce impulsiveness in an animal model of ADHD (Boix et al., 1998). In the present study, we genotyped a CA(n) microsatellite in intron 2 as well as a T/C SNP in intron 13 of the MAO-B gene in 179 Irish ADHD nuclear families. Linkage/Association analysis using TDT revealed no significant association of either of the two markers with the disorder. LD analysis using the program GOLD showed no significant relation between the two examined markers. These data confirm previous findings suggesting that the examined markers in the MAO-B gene do not play a major role in the pathogenesis of ADHD at least in the Irish population.

P10.4 FINE MAPPING AND ASSOCIATION ANALYSIS OF CHROMOSOME 19 AND AUTISM Rabionet R,1 Ma DQ,1 Martin ER,1 Ashley-Koch AA,1 DeLong GR,1 Abramson RK,2 Wright HH,2 Cuccaro ML,1 Gilbert JR,1 and Pericak-Vance MA1 1 Duke University Medical Center, Durham, NC, United States 2 University of South Carolina, Columbia, SC, United States Genome-wide screens for autism (AUT) have provided suggestive evidence for linkage to Chromosome 19. We previously presented the results of our linkage study on 210 multiplex families ascertained through Duke and AGRE (AGRE1)(maximum heterogeneity linkage score (HLOD) of 2.38 at marker D19S593 (17.17 Mb)). The majority of the linkage signal came from AGRE1. Subsequently, we examined 14 SNPs flanking D19S593 (13.23–21.38 Mb) in these data as well as in 200 new families from AGRE (AGRE2). The peak HLOD in the overall dataset was for D19S593 (0.98 (DOM model), 0.78 (REC model)). Again

100

Abstracts

the linkage information came from AGRE1 (HLOD ¼ 3.6 (REC model)). Using the M-test we found significant evidence for heterogeneity showing AGRE1 was significantly different from both Duke and AGRE2 with respect to these data (P ¼ 0.001). Fine mapping of the region using association studies (Pedigree Disequilibrium Test (PDT) and Geno-PDT (genotypic associations)) with SNPs located every 100 Kb flanking D19S593 showed evidence for significant association for AGRE1 for rs901792 (P < 0.005, 16.65 Mb)). Other markers that showed significant association (P < 0.05) were rs1870071 (16.37 Mb), rs2305777 (13.90 Mb) and rs7125 (18.15 Mb). Rs901792 also showed significant genotypic association (geno-PDT P < 0.0005). Functional candidate genes located around these markers are CRSP7, a transcription factor located between rs901792 and rs1870071, EPS15L1, a receptor for the epidermal growth factor near rs1870071, and various predicted genes, like FLJ20241, a putative NFkB activating factor, located near rs2305777. Their involvement if any in AUT susceptibility is currently unknown.

P10.5 IDENTIFICATION OF SIGNIFICANT AUTISM ASSOCIATIONS ON CHROMOSOME 7Q Skaar DA,1 Jaworski JM,1 Martin ER,1 Ashley-Koch AA,1 Benton J,1 McCauley JL,2 Haines JL,2 Sutcliffe JS,2 DeLong GR,1 Wright HH,3 Abramson RK,3 Cuccaro ML,1 Gilbert JR,1 and Pericak-Vance MA1 1 Duke University Medical Center, Durham, NC, United States 2 Vanderbilt University Medical Center, Nashville, TN, United States 3 University of South Carolina, Columbia, SC, United States Multiple studies indicate an area on chromosome 7q as potentially contributing to autism susceptibility. Linkage peaks have previously been found for microsatellite markers d7s1813 (102.82 cM), d7s821 (109.1 cM), d7s477 (111.79 cM), and d7s496 (119 cM). In addition, haplotype and recombination analysis has identified the region 119.8– 122.5 cM as being predominantly shared among families with autistic individuals. Given these indications, as well as the location of the candidate gene RELN in this area of 7q, we undertook a fine-mapping study of this region. SNPs spaced approximately 100kb apart over an area centered on RELN (111.3–121.4 cM/98.8 Mb-108.5 Mb) were analyzed in 407 autism families. Both family history (fam hx)þ and fam hx- families were included. Also tested for association was a GGC repeat in the 50 untranslated region (50 UTR) of RELN (103.2 Mb), which was previously reported as associated with autism. Analyses of the RELN gene markers included an additional 68 fam hx þ families from Vanderbilt University. Significant association was seen for the RELN 50 UTR (P ¼ 0.002) and the RELN Exon 44 SNP (P ¼ 0.03) in the overall dataset. Analysis of haplotypes of the 50 UTR and RELN SNPs 44, 45, 50 and 59 found that the most common haplotype in these families was also significantly associated with autism (P ¼ 0.002). SNP rs925720, at 98.3 Mb, showed significant association (P ¼ 0.004) both in the overall data set as well as in sets stratified by family type (fam hxþ or ). There is no evidence for significant linkage disequilibrium between this SNP and the RELN polymorphisms. However, rs925720 is in the region spanning the proximal breakpoint of a previously reported paracentric inversion in an autism family ((inv (7) (q22-q31.2)). These results continue to support one or more regions on chromosome 7q as areas contributing to autism susceptibility, with RELN again indicated as a strong candidate gene for autism risk.

P10.6 LINKAGE DISEQUILIBRIUM ANALYSES OF CANDIDATE LOCI FOR ADHD: ASSOCIATION BETWEEN THE 5HT1B RECEPTOR GENE AND THE INATTENTIVE SUBTYPE OF ADHD Smoller JW, Arbeitman L, Fagerness J, Doyle AE, Perlis R, Petty C, Faraone SV, and Sklar P Massachusetts General Hospital, Boston, United States Attention deficit hyperactivity disorder (ADHD) is a highly heritable disorder thought to be influenced by multiple genes. Several specific loci have been strongly implicated by previous genetic and pharmacologic studies in both animal models and clinical samples. However, previous association studies have typically examined a small number of markers in or around candidate genes, limiting the interpretation of both positive and negative findings. We have been attempting to evaluate the association between leading candidate genes and ADHD using a more comprehensive linkage disequilibrium (LD) mapping

approach. Here we report results for 4 genetic loci that have been associated with ADHD but have yielded conflicting results in previous studies: 5HT1B, SLC6A3 (DAT), SNAP-25, and DRD5. We examined the LD structure of these genes by genotyping a dense map of SNPs using mass spectrometry in a sample of 12 multigenerational CEPH pedigrees. To optimize the selection of informative markers, we chose ‘‘double-hit’’ SNPs found independently in two databases (dbSNP and Celera). Pairwise measures of LD in the CEPH pedigrees were calculated and haplotype blocks constructed according to the method of Gabriel et al. (Science 2002, 296:2225–9). We performed SNP-wise and haplotype association analyses in up to 229 families of probands ascertained for either DSM-IIIR or DSM-IV ADHD using FBAT. Because previous studies have suggested that susceptibility loci may influence specific ADHD subtypes, we also examined the inattentive and combined ADHD subtypes. At the 5HT1B locus, no association was seen with the ADHD diagnosis, however, 3 individual SNPs were associated (nominal P < 0.05) with the inattentive subtype. These included a synonymous cSNP G861C (rs6296) that has previously been associated with ADHD. The associated SNPs spanned two haplotype blocks, both of which were associated with the inattentive subtype. A 4marker haplotype including the G861C SNP was overtransmitted to ADHD probands with the inattentive subtype (P < .001). Of 35 SNPs tested at the SLC6A3 locus, none were individually associated with ADHD, although a 4-marker haplotype encompassing exons 10 and 11 was associated (P < 0.05). Three of 63 SNPs spanning the large SNAP25 gene were nominally associated with ADHD, including an intron 7 SNP (P < 0.0025). A 4-marker haplotype including exon 8 markers was also associated with ADHD (P < 0.01). For DRD5, none of the 13 SNPs tested around the gene were associated with ADHD or the two subtypes. In sum, these analyses support previous studies implicating the 5HT1B receptor, SLC6A3 and SNAP-25 in the etiology of ADHD. The strongest evidence implicates markers encompassing the 5HT1B gene with the inattentive subtype.

P10.7 DOPAMINE AND NOREPINEPHRINE CANDIDATE GENES IN BRAZILIAN ATTENTION-DEFICIT/ HYPERACTIVITY DISORDER INATTENTIVE TYPE PROBANDS Roman T,1 Schmitz M,2 Guimara˜es APM,2 Denardin D,2 da Silva T,2 Pianca T,2 Hutz MH,2 and Rohde LA2 1 Federal School of Medical Sciences of Porto Alegre, Porto Alegre/RS, Brazil 2 Federal University of Rio Grande do Sul, Porto Alegre/RS, Brazil Attention-deficit/hyperactivity disorder (ADHD) is a highly heterogeneous disorder whose core symptoms include marked, persistent inattention and/or excessive motor activity and impulsivity. Family, adoption and twin studies have supported the role of a strong genetic component in its etiology. In fact, several genes have already been suggested as possible susceptibility genes in ADHD, including the dopamine D4 receptor gene (DRD4), the dopamine transporter gene (DAT1) and the dopamine beta hydroxylase gene (DBH). However, most of the molecular studies investigated ADHD as a whole (inattentive þ hyperactive þ combined types) or limited to patients with combined type of the disorder. The aim of this study is to investigate the DRD4, DAT1, DBH and alfa-2A adrenergic receptor (ADRA2A) genes in ADHD predominantly inattentive type (ADHD-IT). For this purpose, 77 school-age probands, with their biological parents, were ascertained from 12 state schools located in 3 poor neighborhoods in Porto Alegre, the capital of Brazil’s southernmost state. The diagnoses of ADHD according to DSM-IV criteria were performed using semi-structured (K-SADS-E) and clinical interviews with both the subjects and the parents. To date, all these individuals were genotyped for the DAT1 and DBH gene, while 52 and 23 families could be genotyped for ADRA2A and DRD4 genes, respectively. The haplotype relative risk (HRR) statistics was applied. Information on allele transmission could be obtained unambiguously from 70, 63, 48 and 21 families, for DBH, DAT1, ADRA2A and DRD4 genes, respectively. The HRR analyses suggested no association, since all the P values were non-significant (P ¼ 0.52 for DBH, P ¼ 0.66 for DAT1, P ¼ 0.65 for ADRA2A, P ¼ 0.80 for DRD4). Our results suggest that the dopaminergic and noradrenergic genes investigated herein are not contributing to ADHD development in this sample of ADHD-IT patients. However, its putative high etiological heterogeneity does not exclude an involvement of any of these genes in a more specific, homogeneous aspect of the disorder.

Abstracts P10.8 THE REPETITIVE BEHAVIOR PHENOTYPE IN AUTISM: AGE-RELATED CHANGES Cuccaro ML,1 Cope H,1 Donnelly S,1 Wolpert C,1 Gabriels R,3 Wright HH,2 Abramson RK,2 Koch AA,1 Gilbert J,1 and Pericak-Vance MA1 1 Duke University Medical Center, Durham, United States 2 University of South Carolina, Columbia, United States 3 University of Colorado Health Sciences Center, Denver, United States Restricted and repetitive behaviors and interests (RBs) are a defining feature of autism. A repetitive behavior phenotype has been used successfully in a number of studies to create phenotypic subgroups for genetic analyses. The current study examined changes in RBs across development (chronological age and developmental level) using a crosssectional design. 162 participants with autism between the ages of 3 and 21 years were drawn from a larger group of individuals involved in an autism genetics study. Participants consisted of 130 males (mean age ¼ 94 months, sd ¼ 45) and 32 females (mean age ¼ 105 months, sd ¼ 44) from both family history positive and negative families. A standard protocol of clinical diagnostic and behavioral assessments including the Autism Diagnostic Interview-Revised (ADI-R), Aberrant Behavior Checklist (ABC), and Vineland Adaptive Behavior Scales (VABS) was completed for all participants. Using different combinations the 16 ADI-R Restricted and Repetitive Behaviors or Interests domain items we constructed three repetitive behavior scores. The first score, ADI-R Repetitive Behaviors (ARB), was based on items used in the Restricted and Repetitive Behaviors or Interests domain ADI-R algorithm. The other two scores, Repetitive Sensory Motor Behavior (RSMB) and Insistence on Sameness (IS), were based on a previous factor extraction of ADI-R Restricted and Repetitive Behaviors or Interests domain items. The ABC Stereotypy Factor was also examined. Analyses indicated that the RSMB factor was significantly negatively correlated with both age (r ¼ .16, P ¼ 0.04) and developmental level (r ¼ .20, P ¼ 0.01). The IS factor was significantly positively correlated with age ((r ¼ .16, P ¼ 0.05). While significant, these correlations suggest weak relationships. The ABC Stereotypy factor score failed to attain significant correlations with either age or developmental level. Multiple regression analyses were conducted using age, developmental level, and RB measures as predictors. For the RSMB and IS factors, predictor variables account for a limited amount of the variance (RSMB R2 ¼ 14%, IS R2 ¼ 11%). Both age at exam and the VABS Adaptive Behavior Composite score contributed significantly to the dependent variables. Group comparisons (based on age) revealed that younger children showed significantly more RBs based on the ADIR Repetitive Domain score (P ¼ 0.037). The results suggest that select RBs may change with age and merits further study. These results will be discussed in the context of the evolving repetitive behavior phenotype and its use in autism genetic studies.

P10.9 REFINING THE INATTENTIVE SUBTYPE OF ATTENTION-DEFICIT/HYPERACTIVITY DISORDER: AN EXAMINATION OF EXTERNAL VALIDITY USING TWO DOPAMINE GENES Harrington KM,1 Gizer IR,1 Abramowitz A,1 and Stever C2 1 Emory University, Atlanta, GA, United States 2 University of Arizona, Tucson, AZ, United States Ongoing debate regarding the optimum classification and diagnosis of the Inattentive subtype of Attention-Deficit/Hyperactivity Disorder (ADHD) has stimulated an extensive research literature, calling the distinct validity of this subtype into question. Recent research has suggested that one characteristic of DSM-III Attention Deficit Disorder Without Hyperactivity, labeled sluggish cognitive tempo (SCT), may be highly relevant for discriminating among the ADHD subtypes. SCT symptoms include forgetfulness, daydreaming, and sluggishness/ drowsiness, and have been associated with a relatively pure form of the ADHD Inattentive subtype which is unaccompanied by hyperactive-impulsive symptoms. The inclusion of SCT symptoms in the ADHD diagnostic criteria may thus be useful in distinguishing a more homogeneous subgroup within the Inattentive subtype. The primary goal of this study was to evaluate how theoretically guided refinements to the ADHD Inattentive subtype affect its relations with external validity indicators, in particular dopaminergic candidate genes, in an effort to develop a more valid conceptualization of the Inattentive subtype. In the present study, we examined whether the association and linkage between two dopamine genes (i.e., the dopamine

101

transporter and dopamine receptor D4 genes, DAT1 and DRD4) and the ADHD subtypes differed as a function of refinements in the Inattentive subtype. We predicted that selecting children who were high on SCT symptoms and low on hyperactive-impulsive symptoms (e.g., the SCT group) would serve to strengthen the association between DRD4 and the Inattentive subtype. Additionally, we predicted that refining the Inattentive subtype would result in a stronger association between DAT1 and the Combined subtype versus the Inattentive subtype. Our sample included 154 families who were referred to psychiatric clinics for an identified child’s behavioral and learning problems. Logistic regression analyses did not demonstrate strengthened association between DRD4 and the Inattentive subtype or between DAT1 and the Combined subtype, when the Inattentive subtype was restricted to the SCT group. Furthermore, a logistic regression-based extension of the Transmission Disequilibrium Test (TDT) revealed that the Inattentive-SCT group was not associated with a higher proportion of DRD4 high-risk allelic transmissions compared to the standard Inattentive subtype. Thus, these results do not support the reinstatement of SCT symptoms as a component of the diagnostic criteria for the Inattentive subtype of ADHD. P10.10 ASSOCIATION ANALYSIS OF DOPA DECARBOXYLASE, A FUNCTIONAL AND POSITIONAL CANDIDATE GENE FOR ATTENTION-DEFICIT HYPERACTIVITY DISORDER Bakker SC,1 Van der Meulen EM,1 Kusters K,2 Koeleman BPC,2 Verduijn W,3 Buitelaar JK,4 Pearson PL,2 and Sinke RJ2 1 Department of Psychiatry, University Medical Center, Utrecht, Netherlands 2 Department of Biomedical Genetics, University Medical Center, Utrecht, Netherlands 3 Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, Netherlands 4 Department of Psychiatry, University Medical Center, Nijmegen, Netherlands ADHD is a childhood psychiatric disorder characterized by inattention and hyperactivity. Segregation studies have indicated a substantial hereditary susceptibility to the disease (60–80%), but the responsible genes remain largely unknown. Dopa decarboxylase (DDC) could contribute to ADHD and other psychiatric disorders, since this enzyme mediates the final step in the synthesis of dopamine. Genes involved in dopamine neurotransmission have received much attention, since methylphenidate, the most widely used drug to treat ADHD, is known to block the dopamine transporter (DAT-1). Recently, association was reported with a haplotype of a four base pair (bp) insertion/deletion (in/ del) polymorphism and microsatellite marker D7S2422, which is located at 0.5 megabase (Mb) distance from DDC. Moreover, in a recent genome scan in 164 sib pairs with ADHD, we found DDC to be located in a linkage region on chromosome arm 7p, with a maximum multipoint LOD score of 3.04. The marker with the highest single point LOD score in the genome scan was D7S2422, mentioned above. The fact that DDC is both a functional and a positional candidate gene for ADHD, prompted us to investigate the possible involvement of the gene in both family-based and case-control association studies. The first sample consisted of the 195 children with full ADHD according to DSMIV criteria and their parents, analyzed in our previous genome scan. Five single nucleotide polymorphisms (SNPs), the four bp in/del polymorphism and two microsatellite markers, together spanning DDC and its promoter region, were genotyped. Patterns of linkage disequilibrium throughout the gene were investigated and marker haplotypes were constructed for association analysis in a family-based Transmission Disequilibrium Test (TDT). In addition, the five SNPs were genotyped in a case-control study, involving one affected child from each sib pair family plus additional trio families (100 cases), and 252 unrelated controls. All markers except D7S2422 were in significant LD, but no evidence for association of single markers or haplotypes with ADHD was found. Our results indicate that DDC is not likely to play a major role in ADHD in Dutch families, and that other genes in the linkage region on chromosome 7 merit further investigation.

P10.11 ASSOCIATION STUDY OF GILLES DE LA TOURETTE SYNDROME WITH VARIANTS OF THE IL-10 RECEPTOR 1 Schosser A,1 Aschauer HN,1 Schloegelhofer M,1 Leisch F,2 Stamenkovic M,1 Hornik K,3 and Gasche C4

102

Abstracts

1

Department of General Psychiatry, Medical University Vienna, Vienna, Austria 2 Department of Statistics and Probability Theory, Vienna University of Technology, Vienna, Austria 3 Department of Statistics and Mathematics, Wirtschaftsuniversita¨t, Vienna, Austria 4 Department of Medicine 4, Div. of Gastroenterology & Hepatology, Medical University, Vienna, Austria Since there are several reports of linkage of Gilles de la Tourette Syndrome (GTS) to chromosome 11q23-24, and the gene of the IL-10 receptor 1 (IL-10R1) is situated within this region, this gene could be an interesting candidate. In addition, an involvement of autoimmunity and possible beneficial effects of immunomodulatory intervention have been reported. Recently, we have identified two single nucleotide polymorphisms (SNPs) of the IL-10 receptor 1 (IL-10R1) causing a substitution of glycine 330 to arginine (G330R, rs2229113) and of serine 138 to glycine (S138G, rs3135932). The SNPs render cells IL-10 insensitive (Gasche et al. 2003). This project aims to investigate on a possible genetic association between IL-10R1 cSNPs and GTS. A total of 77 unrelated Austrian Caucasians with a DSM-IV (Diagnostic and Statistical Manual of Mental Disorders) consensus diagnosis of GTS, and in addition 250 unrelated healthy controls were examined in order to apply the association strategy. Two allele-specific multiplex PCRs were used for detection of S138G and G330R in genomic DNA. All statistical analyses (Fisher´s exact test, standard X2 test) were performed using the statistical computing environment R version 1.6.2 (http://www.R-project.org). Deviation from Hardy-Weinberg equilibrium was not significant both with the G330R (P ¼ 0.08137, X2 ¼ 3.0374) and the S138G allele (P ¼ 0.1052, X2 ¼ 2.6244). No difference was seen in the allele frequency of patients (G330R: 28.3% and S138G: 15.6%) and controls (G330R: 30.0%, S138G: 16.6%) by Fisher´s exact test (G330R: P ¼ 0.76098, S138G: P ¼ 0.80463). Similarly, there was no significant difference between both the genotype distribution of G330R (P ¼ 0.56224) and S138G (P ¼ 0.36573) between GTS and controls. In conclusion, within our genotyped sample we could not find an association of GTS and the IL10R1. P10.12 MULTIPLE MARKER ANALYSIS OF 5-HT1B IN AN IRISH POPULATION: NO EVIDENCE OF HAPLOTYPE ASSOCIATION WITH ADHD Sheehan K, Hawi Z, Lowe N, Kirley A, Fitzgerald M, and Gill M Departments of Genetics and Psychiatry, Trinity College, Dublin, Ireland Attention deficit hyperactivity disorder (ADHD) is a common neuropsychiatric condition, which affects more males than females. Pharmacological studies using the 5-HT1B agonist RU24969 suggested that the activation of 5-HT1B receptors in mice leads to increased anxiety and locomotion. This hyperlocomotion effect was absent in mice lacking the 5-HT1B receptor indicating that this receptor is involved in mediating the agonist effect. Hawi et al. (2002) reported significant evidence of linkage/association between allele G of the G861C polymorphism of 5HT1B and ADHD in a multi-centre study of 273 nuclear families. In the current investigation we sought to replicate this finding through expansion of the original Irish sample and to perform linkage disequilibrium (LD) mapping in an attempt to narrow down the region of association. Five markers spanning the coding and flanking region of the gene were examined including dbSNP: rs1228814 (A-G), 261 (T-G), 161 (T-A), 861 (G-C) and dbSNP: rs6297 (A-C) in a sample of 179 Irish ADHD trios. The previous finding of significant transmission of 861G was not replicated by this study (TDT: w2 ¼ 0, P ¼ 1). In addition none of the newly examined markers showed significant distortion in the transmission of any particular allele to the ADHD cases. Linkage disequilibrium measured as D’ showed strong LD among the examined markers. Haplotype analysis using the program TRANSMIT failed to show significant association with of any one particular haplotype with ADHD.

P10.13 COMORBIDITY OF AARSKOG SYNDROME AND CHILD PSYCHIATRIC PATHOLOGY: CASE REPORTS, LITERATURE REVIEW AND IMPLICATIONS FOR THE GENETICS OF (DEVELOP)MENTAL RETARDATION Poelmans G,1 Nilius-Hoffmann E,1 Schrander-Stumpel C,2 Danckaerts M,1 Devriendt K,3 and Steyaert J1

1

Department of Child and Adolescent Psychiatry, Catholic University of Leuven, Leuven, Belgium 2 Department of Clinical Genetics, Academic Hospital Maastricht, Maastricht, Netherlands 3 Center for Human Genetics, Catholic University of Leuven, Leuven, Belgium The Aarskog syndrome consists of short stature, associated with typical facial, digital and genital anomalies. Aarskog syndrome has an Xlinked recessive mode of inheritance, and the gene responsible for the syndrome (the FGD1-gene) is located at Xp11.22. To date, only a limited number of mutations in this gene have been reported in Aarskog syndrome patients, and this is probably due to the possible genetic heterogeneity of the disorder and to the considerable overlap between the clinical features of Aarskog syndrome and other syndromes. Consequently, the diagnosis of Aarskog syndrome remains essentially a clinical one and should therefore be based on a thorough clinical examination. We have identified three patients with the clinical diagnosis of Aarskog syndrome, who also have (develop)mental retardation and/or one or more child psychiatric conditions. Our first objective was to review the existing literature on Aarskog syndrome to find out whether there is a comorbidity between Aarskog syndrome and child psychiatric pathology. Moreover, a mutation of the FGD1-gene was recently reported to be associated with nonspecific X-linked mental retardation (or MRX) in one family. For this reason, our second objective was to formulate a hypothesis on how a mutation in the FGD1gene would lead to (develop)mental retardation, and this based on information about the function of the other identified MRX-genes.We have searched PubMed for the phrase ‘‘Aarskog syndrome.’’ We have retrieved 26 articles, in which there were case reports with sufficient information on either psychomotor and language development, IQ, and/or child psychiatric comorbidity. Psychomotor and language developmental delay is rather frequent in Aarskog syndrome and approximately one third of the boys with Aarskog syndrome may have an IQ in the borderline or mentally retarded range (IQ below 85). In 51 of the 164 reported cases of Aarskog Syndrome in the literature, one or more comorbid child psychiatric conditions other than (develop)mental retardation were diagnosed, and especially ADHD and autism spectrum disorders seem to be frequent in Aarskog syndrome. However, since biased ascertainment could not be excluded, further studies with larger groups of patients are needed to confirm these findings. Furthermore, the literature shows that the FGD1-gene and three other MRX-genes are involved in the same biochemical pathway, namely the Rho GTPase signaling cascade in the brain. However, further studies are needed to find out whether mutations in the FGD1-gene could explain for the MRX-phenotype in other families and which effect these mutations would have on the Rho GTPase cascade.

P10.14 ANOMALOUS ASYMMETRY IN TURNER’S AND KLINEFELTER’S SYNDROMES—FURTHER EVIDENCE FOR X-Y LINKAGE OF THE CEREBRAL DOMINANCE GENE Rezaie R,1 Roberts N,1 Cutter WJ,2 Murhpy DGM,2 Robertson DMW,2 Daly EM,2 Maurizio A,3 Delisi LE,3 and Crow TJ4 1 MARIARC, University of Liverpool, Liverpool, United Kingdom 2 Brain Maturation Unit, Institute of Psychiatry, London, United Kingdom 3 Deparment of Psychiatry, New York University, New York, United States 4 SANE POWIC, Warneford Hospital, Oxford, United Kingdom Turner’s syndrome (XO) individuals have deficits of spatial and Klinefelter’s individuals (XXY) have deficits of verbal ability. On this basis it was suggested that a determinant of cerebral asymmetry (the ‘‘cerebral dominance gene’’) is located in a region of X-Y homology (Crow, 1993). Here we investigated groups of individuals with Turner’s and Klinefelter’s syndromes by magnetic resonance imaging with a new technique for assessing anatomical asymmetry (the cerebral ‘‘torque’’ from right frontal to left occipital across the antero-posterior axis, Mackay et al., 2003). The main findings are that relative to same sex controls. 1) Turner’s syndrome individuals have exaggerated occipital asymmetry to the left, compatible with loss of tissue on the right. Turner’s individuals also have L > R asymmetry of the posterior superior temporal gyrus while controls have L < R. 2) Klinefelter’s syndrome individuals have loss of the rightward asymmetry in the frontal lobes, and diminished asymmetry to the left in white matter in a region of the occiptal lobe. They also have changes in asymmetry in grey

Abstracts and white matter in the anterior superior temporal gyrus. These changes in anatomical asymmetry are correlates of the reciprocal neuropsychological impairments in these syndromes. In that deficits in sex chromosome aneuploidies are attributable to genes that are protected from X inactivation (generally those that are present in regions of X-Y homology) the findings add significantly to the case that the genetic determinant of the Homo sapiens-specific characteristic of cerebral asymmetry is X-Y homologous. Given its evolutionary history ProtocadherinX/Y is the obvious candidate (Crow & Williams, 2004). P10.15 THE GENETICS OF AD/HD: SUBTYPING, COMORBIDITY AND DEVELOPMENTAL CONSIDERATIONS Schug MD,1 Anastopoulos A,2 Kollins S,3 Hennis L,2 Nelson S,1 Mehltretter L,4 Gilbert J,4 Cuccaro M,4 and Ashley-Koch A4 1 Department of Biology, University of North Carolina Greensboro, Greensboro, NC, United States 2 Department of Psychology, University of Noth Carolina Greensboro, Greensboro, NC, United States 3 Department of Psychiatry, Duke University Medical Center, Durham, NC, United States 4 Center for Human Genetics, Duke University, Durham, NC, United States Attention-Deficit/Hyperactivity Disorder (AD/HD) is a complex genetic disorder with an extremely variable clinical presentation. First, within the AD/HD diagnosis, children are subtyped according to the relative distribution of the primary features, resulting in one of three major classifications: Combined, Predominantly Inattentive, and Predominantly Hyperactive-Impulsive subtypes. Second, children with AD/HD display additional variability in terms of the presence or absence of comorbid externalizing and internalizing psychiatric conditions. Finally, the clinical presentation of AD/HD with respect to subtyping and comorbidity often changes within an individual over time. For example, a child may exhibit primarily hyperactive symptoms in early years and primarily inattentive symptoms later in development. To date, over 30 candidate genes and chromosomal regions have been implicated in the etiology of AD/HD. Most candidate gene and genomic screen studies have defined the AD/HD phenotype as a dichotomous variable from clinical assessments performed at a single time point according to the DSM-IV criteria. While there has been success towards identifying potential genetic influences using this approach, Faraone and colleagues (2000) have observed that the sibling recurrence risk (ls) for AD/HD increases from 4.0 among AD/HD probands ascertained at a single time point to 17.2 among probands whose AD/HD persists into adolescence, suggesting that the temporal persistence of AD/HD is a ‘‘more genetic’’ form of the condition. Moreover, the sibling recurrence risk (ls) in families with probands that not only have persistence of AD/ HD but also comorbid conduct disorder is even higher (ls ¼ 26.2). Thus, we should expect to find molecular genetic variation related to the temporal persistence of AD/HD, not only when broadly defined but also when defined in terms of its subtyping and comorbidity features. In support of this approach, we present evidence from a study of 25–30 AD/ HD children followed from age 2 to 7 that the developmental trajectory of AD/HD subtypes and comorbid risk factors is extremely variable and thus may impact the likelihood of finding a genetic association. Based on these data, the Duke Center for Human Genetics and University of North Carolina Greensboro have initiated a longitudinal study of AD/ HD families to identify genetic factors that may contribute to these variable developmental trajectories.

P10.16 INVESTIGATION OF MID-FRONTAL EEG ALPHA ASYMMETRY AND GENE VARIANTS AT THE BRAIN-DERIVED NEUROTROPHIC FACTOR (BDNF) GENE LOCUS Bulgin N,1 Strauss J,1 Shaikh S,1 King N,1 George C,2 Fox N,2 Kovacs M,2 Barr C,3 and Kennedy J1 1 Centre for Addiction and Mental Health, Clarke Institute, Toronto, Canada 2 University of Pittsburgh School of Medicine, Western Psychiatric Institute and Clinic, Pittsburgh, United States 3 Toronto Western Research Institute, University of Toronto, Toronto, Canada A recent examination of interhemispheric brain activity in adults with a history of childhood-onset mood disorders (COMD) found significant

103

differences in mid-frontal EEG alpha asymmetry between affected and control subjects. Our previous genetic studies of this sample demonstrated significant differences in allele and haplotype frequencies of the brain-derived neurotrophic factor (BDNF) gene, a gene involved in neuronal survival and plasticity, between demographically matched COMD cases and control populations (Strauss et al. 2004). Since alpha power has been shown to be moderately heritable, a logical extension of these two observations would be to examine the association of midfrontal alpha asymmetry and BDNF gene variants. In this study we genotyped 172 adults with a history of COMD at 7 biallelic singlenucleotide polymorphisms (SNPs) within and flanking the BDNF gene and tested for association of genotypes and alleles with mid-frontal EEG alpha asymmetry using ANOVA. No significant associations were observed when genotype was considered (P > 0.05). However, we did observe a moderate (0.1 > P > 0.05) association between mean asymmetry and the presence of specific intragenic BDNF alleles. These results suggest that the subphenotype asymmetry patterns observed in adults with a history of COMD may in part be associated with specific gene variants at the BDNF locus, providing a potential link in the understanding of the complex neurobiology of COMD. P10.17 AUTISTIC TRAITS IN THE SIBLINGS OF AFFECTED PROBANDS Todd RD, Constantino JN, and Abbacchi AM Washington University, St. Louis, United States The recurrence risk for autism spectrum disorders in the siblings of affected probands is known to be on the order of 10 per cent. The primary purpose of this study is to investigate the aggregation and distribution of autistic traits in the siblings of probands with autism spectrum conditions, in order to gain a better understanding of whether family liability for autism involves discrete recurrence risk (to that minority of family members who become affected) or a continuous distribution of elevated risk, for which the tail of that distribution represents affected status. We obtained quantitative assessments of autistic symptomatology in the brothers of 64 subjects with clinically diagnosed autism spectrum conditions. We also obtained assessments in an age-and gender-matched (1:2) contrast group of 32 brothers of children with non-autism-related child psychiatric conditions. In order to minimize measurement error incurred by rater contrast effects (which have been observed when parents complete ratings on both their affected and ‘‘unaffected’’ children), the assessments were conducted by teacher-report (independent raters for proband and sib), using the Social Responsiveness Scale (SRS). The SRS is an established quantitative measure of autistic traits, which has been validated against the Autism Diagnostic Interview-Revised (the research standard for a clinical diagnosis of autistic disorder). The SRS generates a singular scale score that serves as an index of severity of social impairment in the autism spectrum: higher scores on the SRS indicate greater severity of social impairment. Scores on the SRS are highly heritable, do not exhibit age-effects in school-aged children, are generally unrelated to I.Q., and are continuously distributed in the general population. Sib SRS scores were continuously distributed in both groups—there was no evidence of bimodality in the distribution of scores for siblings of affected probands. When comparing mean teacher-report SRS scores between groups, the siblings of autistic probands had significantly higher scores than did siblings of non-PDD probands (51.6  40.3 verses 22.9  35.0, t ¼ 3.95, df ¼ 94, P < .001). As predicted, when comparing these independent teacher reports to maternal SRS reports on the same subjects, rater contrast effects were evident in the maternal report data on the order of an effect size of 0.5. By October 2004, data on an additional 150 families will be available for incorporation into these analyses. Genetic studies of autism may benefit from quantitative trait analyses involving data from affected individuals and their family members. The results to date indicate that genetic liability for autism appears to be manifested by a relative aggregation of autistic traits, which are continuously distributed (in severity) among the relatives of autistic probands.

P10.18 EVALUATING ALTERNATIVE CONTINUOUS PERFORMANCE TASK ERROR INDICES AS ENDOPHENOTYPES FOR MOLECULAR GENETIC STUDIES OF ADHD Gizer I and Waldman ID Emory University, Department of Psychology, Atlanta, United States

104

Abstracts

Researchers studying the genetic influences underlying psychiatric disorders have hypothesized that biological markers of a disorder might be more useful in identifying susceptibility genes for that disorder than the diagnostic category itself. Such markers, termed endophenotypes, are thought to represent quantitative traits indicating disease liability that more directly reflect the action of individual genes that confer risk for developing the disorder of interest. As such, a valid endophenotype should, at minimum, satisfy three necessary criteria. First, the putative endophenotype must represent a heritable trait. Second, given their genetic relation, it should be present in the unaffected siblings of affected children to a greater degree than is found in the general population. Third, the putative endophenotype should show evidence of association and linkage with some of the same genes previously shown to be related to the diagnosis. For ADHD, sustained attention and executive functions have been suggested as putative endophenotypes because children with ADHD often show deficits in these areas. In the current study, we used two of the three criteria described to critically evaluate ‘A-X’ Continuous Performance Task (AX CPT) indices of sustained attention and executive functions as endophenotypes for ADHD. In this task, subjects attend to a series of letters presented one at a time and respond whenever an ‘A’ followed by an ‘X’ appears. Researchers have presumed that missed targets index inattention, whereas incorrect responses to non-target stimuli index impulsivity. A second scoring method, which relies on more finegrained analyses of these errors to yield indices of inattention and impulsivity indices, has been proposed by Halperin and colleagues [Halperin, J. M., et al. (1991). Subtype analysis of commission errors on the continuous performance test in children. Developmental Neuropsychology, 7, 207–217]. Currently, the superiority of one scoring method over the other has not been demonstrated. Thus, the present study utilized both scoring methods to evaluate A-X CPT indices of sustained attention and executive functions as putative endophenotypes for ADHD. As a first step, we compared performance on the A-X CPT indices of each scoring method among clinic-referred probands, their unaffected siblings, and unrelated control children to determine whether probands’ unaffected siblings showed greater inattention and/ or impulsivity than population controls. Second, we examined the relation between the A-X CPT indices from each scoring method and two genes, the dopamine transporter gene (DAT1) and the dopamine D4 receptor gene (DRD4), previously shown to be associated with ADHD in clinic-referred samples. As a final step, we tested the extent to which the A-X CPT indices mediated and moderated the observed associations between the candidate genes and ADHD. P10.19 SCREENING FOR MUTATIONS IN EXON 1 OF MECP2, AMONG INDIVIDUALS WITH EITHER RETT SYNDROME, AUTISM OR MENTAL RETARDATION Harvey C,1 Alfred SE,1 Mnatzakanian G,2 Roberts WS,2 Schanen NC,3 Scherer SW,2 Minassian BA,2 and Vincent JB1 1 Centre for Addiction and Mental Health, Toronto, Canada 2 Hospital for Sick Children, Toronto, Canada 3 University of Delaware, Wilmington, United States We recently identified a new isoform of the Rett syndrome gene, MECP2, which results in a new coding region within the first exon. We have identified a mutation in a single female Rett patient within this new coding sequence- an 11 base pair deletion that results in a frameshift, causing a premature truncation of the protein. Previously, mutations in the MECP2 gene (exons 3 and 4) have also been identified in individuals with other neuropsychiatric disorders, including autism, mental retardation and Angelman syndrome. We hypothesized that, since this new coding region is clearly important for the correct function of MECP2 during neurodevelopment, this coding region would also be relevant for screening in cases of autism and mental retardation (and non-15q11 cases of Angelman). We have now screened this region in 20 Rett patients, 100 autism probands, and in 100 patients with mental retardation. Direct DNA sequencing was used for mutation detection. No mutations were identified among the autism screening set, however 4 sequence variants were identified among the MR cases. We report here on the sequence variants identified, and their possible relevance to neurodevelopment, autism and mental retardation.

P10.20 ARE MINOR PHYSICAL ANOMALIES AN ENDOPHENOTYPE OF AUTISM? Malhotra T,1 Srinath S,2 Girimaji SC,2 and Seshadri SP2

1 Department of Psychiatry, Hinchingbrooke Hospital, Huntingdon, United Kingdom 2 Department of Psychiatry, National Institute of Mental Health and Neurosciences, Bangalore, India

Family history data suggests that not only Autism but also its lesser variant referred to as the Broad Autism Phenotype (BAP) is inherited. However, the exact patterns of inheritance have still not been clearly delineated, owing to the stumbling blocks of genetic and phenotypic heterogeneity. It may be necessary to investigate different variables to see whether they reflect familial component of autism. The aim of this study was to examine the relation between familial loading of BAP and minor physical anomalies (MPAs) in 37 autistic probands. MPAs were assessed using the Waldrop Minor Physical anomaly scale and the familial loading of BAP was assessed using the Family History Interview for Developmental Disorders of Cognition and Social Functioning across three generations. Statistical analysis was carried out on SPSS 10.0, using the Chi-square test, with a threshold of significance of P < 0.05. The results of this study indicate that the risk to relatives of BAP is significantly greater if the autistic proband has low rate of MPAs rather than higher rate of MPAs. Probands with low MPA score also had significantly higher loading for social and communication deficits in their relatives as compared to probands with high MPA score. This study supports the previous finding that the autistic subjects with low MPA score are less retarded and have more relatives with psychiatric illness. One hypothesis suggested from our results could be that the autistic subjects with high MPA score represent sporadic or ‘nonfamilial’’ autism due to single environmental insult or non-transmitted genetic events. The autistic subjects with low MPA score could be representative of ‘‘familial’’ autism where familial loading for the lesser variant of autism reflects variable expression of the underlying genotype. MPAs could be used as an indicator to index the degree of genetic liability among both affected and non-affected family members in molecular genetic studies.

P10.21 ETIOLOGICAL HETEROGENEITY IN AUTISM: IMPLICATIONS FOR LINKAGE AND ASSOCIATION STUDIES Betancur C,1 He´ron D,2 Verloes A,3 Philippe A,4 Munnich A,4 Depienne C,5 Brice A,5 and Leboyer M6 1 INSERM U513, Cre´teil, France 2 Unite´ de Ge´ne´tique Clinique, Hoˆpital Pitie´-Salpeˆtrie`re, Paris, France 3 Unite´ de Ge´ne´tique Clinique, Hoˆpital Robert Debre´, Paris, France 4 INSERM U393 and De´partement de Ge´ne´tique, Hoˆpital Necker-Enfants Malades, Paris, France 5 INSERM U289 and De´partement de Ge´ne´tique Hoˆpital Pitie´-Salpeˆtrie`re, Paris, France 6 De´partement de Psychiatrie, Hoˆpital Albert Chenevier et Henri Mondor, Cre´teil, France There is compelling evidence from family and twin studies that genetic factors play a major role in the development of autism. The mechanism of transmission is assumed to be complex and is likely to involve multiple susceptibility loci. Despite numerous genome-wide screens and association studies performed in the past years, no susceptibility genes have been identified. Etiological heterogeneity is a major obstacle to the identification of susceptibility genes, and is likely to be considerably greater in autism than in other complex disorders. Thus, we recently began a study to search for genetic disorders associated with autism in the cohort of patients recruited by the Paris Autism Research International Sibpair (PARIS) study for linkage and association studies. At the time of ascertainment, patients had been evaluated by a psychiatrist and underwent a routine medical screening, including karyotype, fragile X testing and metabolic studies. Patients demonstrated to suffer from organic conditions or chromosomal anomalies were excluded. The present study was thus performed on the group of patients with presumed ‘‘idiopathic’’ autism. All families from our cohort living in France (n ¼ 252 subjects belonging to 212 families) were asked to participate in this new study. Those who accept are evaluated by a clinical geneticist and a systematic screening for chromosomal and metabolic abnormalities is performed. Cytogenetic studies include high-resolution karyotype and FISH analysis of 15q11-q13, 22q13.3 and 22q11.2, subtelomeric rearrangements are investigated by FISH. Various metabolic pathways involved in mental retardation and/or autism are examined using mass spectrometry. We also search for microdeletions/microduplications in eight chromosomal regions previously associated with mental retardation and autistic behaviors:

Abstracts 2q37.3, 5q35, 7q11.23, 15q11.2, 16q13.3, 17q11.2, 22q11.2 and 22q13.3. Approximately 80 families have already been evaluated. Among the multiplex families seen so far, we have identified a 22q11 deletion, an ATRX (X-linked alpha-thalassemia and mental retardation) syndrome, and a mutation in the neurotrypsin gene (PRSS12), recently implicated in autosomal recessive nonsyndromic mental retardation. Among the sporadic cases, we have identified two with chromosome 15q11-q13 duplications, one with a 22q13 deletion, one with Sotos syndrome, and a girl with a MECP2 mutation. These findings show the existence of a significant number of diagnosable medical conditions in autism, including single-gene defects and cytogenetic abnormalities associated with well-defined syndromes. Our results strongly support the need for the systematic search for genetic conditions associated with autism in other populations recruited for linkage and association studies. P10.22 MITOCHONDRIAL DYSFUNCTION AND AUTISM: LACK OF ASSOCIATION OF THE ASPARTATE/GLUTAMATE CARRIER SLC25A12 GENE WITH AUTISM OR HYPERLACTACIDEMIA IN PORTUGUESE PATIENTS Correia C Inst Gulbenkian de Ciencia, Oeiras, Portugal A growing body of evidence suggests the involvement of mitochondrial dysfunction in the pathophysiology of autism spectrum disorders. In a previously reported population-based study of autistic children in Portugal, we have found hyperlactacidemia in 20.3% of the tested patients. High lactate levels are suggestive of the occurrence of mitochondrial dysfunction, which results in defects in the brain bioenergetic metabolism and in impaired central nervous system functioning. Further evaluation of the children with hyperlactacidemia confirmed diagnosis of definite mitochondrial disease in 7.2% of the autistic patients. As mtDNA mutations were not detected in any of the subjects, involvement of nuclear genes was suggested. Recently, a strong association of autism with two SNPs within the SLC25A12 gene, a nuclear gene on chromosome 2 that encodes a mitochondrial aspartate/ glutamate carrier (AGC1), was demonstrated. In the present study, we have examined the association with autism of two SLC25A12 polymorphisms, rs2056202 and rs11757, in a sample of 241 autistic nuclear families. Both polymorphisms are SNPs, with rs11757 localized in the 50 UTR region and rs2056202 in intron 3-4. In our sample we found no transmission disequilibrium of either SNP alleles (Z ¼ 0.873, P ¼ 0.383 for rs2056202, Z ¼ 0.816, P ¼ 0.414 for rs11757) or haplotypes of the two polymorphisms (w2 ¼ 3.05, 3 df, P ¼ 0.384). AGC1 plays an important role in the malate/aspartate shuttle, which regulates the cytosolic redox state controlling lactate to pyruvate conversion. Given that hyperlactacidemia was present in 17% of cases in this population, we examined whether these markers could be associated with the lactate level distribution, and therefore contribute to the determination of hyperlactacidemia. The association of lactate levels with these markers was tested in 199 autistic subjects, using the non-parametric Kruskal-Wallis test, no significant association was found (H ¼ 1.68, P ¼ 0.4318 for rs2056202, H ¼ 1.766, P ¼ 0.4135 for rs11757). We therefore do not replicate the recent findings of an association of rs2056202 with autism, possibly reflecting the known genetic heterogeneity in autism. We further show that SLC25A12 polymorphisms are not associated with hyperlactacidemia in our sample. Other genes encoding enzymes from the malate/aspartate-shuttle or from the mitochondrial respiratory chain complexes, important for the maintenance of the cytosolic redox states, are potential candidates for a role in the etiology of autism spectrum disorders.

P10.23 MUTATION SCREENING AND SNP ANALYSIS OF GABRB3 IN A SUBSET OF AUTISTIC INDIVIDUALS Menold MM,1 Jaworski JM,1 Dementieva Y,2 Wolpert CM,1 Donnelly S,1 Abramson RK,3 Wright HH,3 Ashley-Koch A,1 Cuccaro ML,1 Pericak-Vance MA,1 and Gilbert JR1 1 Duke Center for Human Genetics, Durham NC, United States 2 Marshall University, Huntington WV, United States 3 University of South Carolina, Columbia SC, United States Autistic disorder is a neurodevelopmental disorder with a strong genetic component. Genetic linkage studies, as well as cytogenetic evidence suggest the presence of an autism susceptibility locus on chromosome 15q11-q13. Association of autistic disorder with the GABRB3 gene further supported this as a candidate region. Ordered

105

subset analysis (OSA) is a method that can be used to identify a particular set of families that are contributing to linkage or association at a given locus. Using this technique, Shao et al. (2003) demonstrated significantly increased LOD scores in the region of the GABRB3 gene by using a factor derived from repetitive and stereotyped patterns of behavior as the covariate for OSA. This analysis strengthened previous results of linkage and association at this locus for autistic disorder. Using the DNA from the probands of the 23 multiplex families with the highest OSA repetitive and stereotyped behavior scores, we have performed a screen for sequence variations in the promoter, the 10 known exons, and the untranslated regions of GABRB3. We have also analyzed the sequence of the 4 exons of a novel gene (NT_026446.71) that is predicted to be located 30 to GABRB3. Nineteen primer sets that spanned the exons and their splice regions were used to generate PCR products for sequencing. Analysis of the GABRB3 sequence data confirmed 4 single nucleotide polymorphisms (SNPs) and identified 1 that was novel. The new SNP was located within intron 3 (ex4-57 G ! A, gb:AC009698.10). No new coding mutations were found in the exons of GABRB3. In addition to the sequence analysis, we analyzed 17 SNPs or deletion/insertion polymorphisms (DIPs) within GABRB3, extending to 100 kb beyond the 30 end of the gene. Six of the SNPs showed positive linkage (>1.0) with a peak LOD score of 1.7 for SNP rs2017247, located 3.6 kb beyond the end of the gene. Association analysis was not significant (P > 0.05) for any of the SNPs using either the Pedigree Disequilibrium Test (PDT) or Transmit. While these results provide further support for the presence of an autism susceptibility locus in this region of chromosome 15q, autistic disorder does not appear to be associated with a particular SNP haplotype or a coding mutation within the GABRB3 gene in these 23 families. Examination of the cDNA sequence of GABRB3 to look for splicing variations in the area of the novel SNP is ongoing.

P10.24 EVIDENCE FOR GENE-GENE INTERACTIONS INFLUENCING SUSCEPTIBILITY TO AUTISTIC DISORDER Ashley-Koch AE,1 Jaworski J,1 Mei H,1 Menold M,1 Rabionet R,1 Skaar D,1 Ritchie M,2 Abramson RK,3 Wright HH,3 DeLong GR,1 Cuccaro ML,1 Gilbert JR,1 Martin ER,1 and Pericak-Vance MA1 1 Duke University Medical Center, Durham, United States 2 Center for Human Genetics Research, Vanderbilt University Medical School, Nashville, United States 3 Department of Neuropsychiatry, University of South Carolina School of Medicine, Columbia, United States Autistic disorder (AD) has a complex genetic architecture with evidence for susceptibility loci on several chromosomes. We, and others, have hypothesized that some of these genes may act epistatically to contribute to AD susceptibility. To explore this hypothesis, we have examined 241 multiplex (including 99 from AGRE) and 210 singleton AD families for gene-gene interactions among AD candidate genes on chromosomes 2, 7, 15, 17 and 19. All AD affected individuals met ADI-R criteria. SNP genotypes in candidate genes were identified via public databases and generated using Assays on Demand or Assays by Design from Applied Biosystems. All markers were tested for deviations from Hardy-Weinberg Equilibrium and for linkage disequilibrium with other SNPs in the respective candidate gene. We used a novel statistical method, the multi-locus genotype-PDT, to detect multi-locus associations, as well as the pattern recognition approach of the multifactor dimensionality reduction (MDR) method to detect such epistatic interactions. The first hypothesis that we tested was that the cluster of three GABA receptor subunit genes on chromosome 15 (GABRb3, GABAa5, GABAg3) interact to influence AD susceptibility, due to their positional proximity and functional similarity. In the subset of singleton families, we observed evidence for a multi-locus effect of 4 SNPs involving all three GABA genes in the MDR analysis (P ¼ 0.02). The second hypothesis we examined was that RELN on chromosome 7 interacts with APOE on chromosome 19 as these genes compete for binding to the VLDL/APOE receptor. While the MDR analysis did not detect a significant multi-locus effect between these two genes, the genotype-PDT gave a significant global result (P ¼ 0.003) for a C/T substitution at position 427 in APOE with an A/G substitution in exon 45 of RELN. The final hypothesis we examined was whether RELN interacts with the SLC6A4 gene on chromosome 17. In the subset of singleton families, we observed evidence for a multi-locus effect of several SNPs involving combinations of 2 and 3 SNPs from both genes in the MDR analysis (P < 0.001). This effect was supported in the MDR analysis of the entire dataset (P ¼ 0.02), as well as in the genotype-PDT

106

Abstracts

analysis (P < 0.001). We are currently exploring these multi-locus effects through more traditional statistical models, such as logistic regression, to disintangle main effects and interactions. We are encouraged by these preliminary results as they support our hypothesis that epistatic interactions do indeed contribute to susceptibility to autistic disorder. P10.25 EXAMINATION OF A POSSIBLE INFLUENCE OF GENETIC VARIATION AT THE DYX1C1 LOCUS IN THE DEVELOPMENT OF DYSLEXIA Dahdouh F,1 Schumacher J,1 Koenig IR,2 Lindgren C,3 Peyrard-Janvid M,3 Anthoni H,3 Zucchelli M,3 Grimm T,4 Warnke A,5 Roth E,5 Cichon S,6 Remschmidt H,7 Propping P,1 Ziegler A,2 Kere J,3 Schulte-Ko¨rne G,7 and No¨then MM6 1 Institute of Human Genetics, University, Bonn, Germany 2 Institute of Medical Biometry and Statistics, University, Lu¨beck, Germany 3 Department of Biosciences at Novum, Karolinska Institute, Stockholm, Sweden 4 Department of Medical Genetics, University, Wu¨rzburg, Germany 5 Department of Child and Adolescent Psychiatry, University, Wu¨rzburg, Germany 6 Life & Brain Centre, University, Bonn, Germany 7 Department of Child and Adolescent Psychiatry, University, Marburg, Germany Dyslexia is the most common neurobehavioral disorder in childhood, with a prevalence of 5–10% of school-age children. A number of studies have previously identified chromosomal regions likely to contain genes contributing to Dyslexia. Most recently, Taipale et al. (PNAS 2003, 100: 11553–11558) investigated the gene DYX1C1 on chromosome 15q21 as a positional candidate gene for the developmental of dyslexia. Single locus and haplotype analysis revealed variants within DYX1C1 to be associated in a case-control sample with dyslexia of Finish descent. We attempted to replicate these findings and genotyped 4 significantly associated SNPs from the Finish study in a sample comprising 171 trios with dyslexia of German descent. TDT analyses provided no evidence for association on the single marker level. However, frequencies of two-, three-, and four-marker haplotypes were significantly different between patients and controls (P ¼ 0.038, P ¼ 0.018, and P ¼ 0.018 respectively). Currently, we are sequencing all 10 exons and the putative promoter region of DYX1C1 in 10 German patients with dyslexia carrying the risk haplotype. Identified variants will be genotyped in an extended trio sample with dyslexia in order to identify the disease causing variant/haplotype. Our association data will be presented at the conference.

P10.26 AUTISTIC TRAITS IN THE SIBLINGS OF AFFECTED PROBANDS Todd R Washington University School of Medicine, Washington, United States The recurrence risk for autism spectrum disorders in the siblings of affected probands is known to be on the order of 10 per cent. The primary purpose of this study is to investigate the aggregation and distribution of autistic traits in the siblings of probands with autism spectrum conditions, in order to gain a better understanding of whether family liability for autism involves discrete recurrence risk (to that minority of family members who become affected) or a continuous distribution of elevated risk, for which the tail of that distribution represents affected status. We obtained quantitative assessments of autistic symptomatology in the brothers of 64 subjects with clinically diagnosed autism spectrum conditions. We also obtained assessments in an age-and gender-matched (1:2) contrast group of 32 brothers of children with non-autism-related child psychiatric conditions. In order to minimize measurement error incurred by rater contrast effects (which have been observed when parents complete ratings on both their affected and ‘‘unaffected’’ children), the assessments were conducted by teacher-report (independent raters for proband and sib), using the Social Responsiveness Scale (SRS). The SRS is an established quantitative measure of autistic traits, which has been validated against the Autism Diagnostic Interview-Revised (the research standard for a clinical diagnosis of autistic disorder). The SRS generates a singular scale score that serves as an index of severity of social impairment in the autism spectrum: higher scores on the SRS indicate greater severity

of social impairment. Scores on the SRS are highly heritable, do not exhibit age-effects in school-aged children, are generally unrelated to I.Q., and are continuously distributed in the general population. Sib SRS scores were continuously distributed in both groups—there was no evidence of bimodality in the distribution of scores for siblings of affected probands. When comparing mean teacher-report SRS scores between groups, the siblings of autistic probands had significantly higher scores than did siblings of non-PDD probands (51.6  40.3 verses 22.9  35.0, t ¼ 3.95, df ¼ 94, P < .001). As predicted, when comparing these independent teacher reports to maternal SRS reports on the same subjects, rater contrast effects were evident in the maternal report data on the order of an effect size of 0.5. By October 2004, data on an additional 150 families will be available for incorporation into these analyses. Genetic studies of autism may benefit from quantitative trait analyses involving data from affected individuals and their family members. The results to date indicate that genetic liability for autism appears to be manifested by a relative aggregation of autistic traits, which are continuously distributed (in severity) among the relatives of autistic probands. P10.27 DYSBINDIN (DTNBP1, 6P22.3) IS ASSOCIATED WITH CHILDHOOD ONSET PSYCHOSIS AND ENDOPHENOTYPES MEASURED BY THE PREMORBID ADJUSTMENT SCALE (PAS) Gornick MC National Institutes of Health, NIMH, Child Psychiatry Branch, Bethesda, United States Straub et al. (2002) recently identified the 6p22.3 gene dysbindin (DTNBP1) as a schizophrenia susceptibility gene through a positional cloning effort. We studied a rare cohort consisting of 102 patients with onset of psychotic symptoms before the age of 13. Information on these subjects included standardized ratings of early development, history of medication response, neuropsychological and cognitive test performance, smooth pursuit eye movements, and premorbid dysfunction along with 2–6 year clinical follow-up status. We genotyped 14 SNPs in the DTNBP1 gene and performed family-based pairwise and haplotype transmission disequilibrium test (TDT) analyses with the clinical phenotype. We also used a quantitative transmission disequilibrium test (QTDT) to explore the relationship between genotype and endophenotypes. Only one SNP, located in the 50 untranslated region, was associated with diagnosis (TDT P ¼ 0.01). However, the QTDT analyses showed several significant results. In particular, 4 adjacent SNPs were strongly associated (empirical P values ¼ 0.0009–0.003) with poor premorbid functioning as measured by the Premorbid Adjustment Scale (PAS). These findings provide additional support for the genetic continuity between childhood-onset and adult-onset forms of the disorder. It is hypothesized that other schizophrenia susceptibility genes may also relate to these nonspecific measures in other developmental disorders, reflecting early neurodevelopmental impairment. We are currently testing these same SNPs for association with cognitive measures in both normative and ADHD samples.

P10.28 SUGGESTIVE EVIDENCE FOR LINKAGE ON ODD SYMPTOM SEVERITY IN A GENOMEWIDE QUANTITATIVE TRAIT LOCUS ANALYSIS OF DUTCH AFFECTED SIBPAIRS WITH ADHD van der Meulen EM,1 Bakker SC,2 de Kovel CGF,3 Pauls DL,1 Sinke RJ,2 and Buitelaar JK4 1 Unit of Psychiatric and Neurodevelopmental Genetics/Massachusetts General Hospital-Harvard Medical School, Boston, United States 2 Department of Medical Genetics, University Medical Center Utrecht, Utrecht, Netherlands 3 Department of Human Genetics, University Medical Center Nijmegen, Nijmegen, Netherlands 4 Department of Psychiatry, University Medical Center Nijmegen, Nijmegen, Netherlands A previous genome-wide scan based on 402 microsatellite markers with an average distance of 10 cM revealed three peak regions located at chromosome 15q, (MLS ¼ 3.54 under a broad phenotype definition) and chromosome regions 7p and 9q (MLSs of 3.04 and 2.05 respectively, in a narrow -ADHD only- phenotype). This study focused on the genetic basis of ADHD and Oppositional Defiant Disorder (ODD) symptom severity in the original sibpair sample. The sample consisted of a group

Abstracts of 164 affected sibpairs (N ¼ 238). All children had been diagnosed with Attention-Deficit Hyperactivity Disorder (ADHD), (N ¼ 199), subthreshold ADHD (N ¼ 13) or autistic spectrum disorder and major comorbid ADHD symptomatology (N ¼ 26) using a best estimate procedure based on DSM-IV criteria. A traditional Haseman-Elston analyis on ADHD and ODD symptom severity was used in GENEHUNTER to test for linkage in the original genome-wide marker data. The parent SWANscore (without medication) was used to measure ADHD symptom severity. ODD symptom severity data were derived from parent interviews using the Diagnostic Interview Schedule for Children (DISC). A scale ranging from 0-8 was used, which was directly based on the 8 DSM-IV criteria for ODD. In the QTL analysis of ODD symptom severity the highest peak corresponding with a LOD-score of 2.23—was located on chromosome 11p. Other peaks with LOD-scores of 1.71, 1.93, 1.04, 1.20 and 1.62 were found on chromosome 5q, 6q, 12q, 13q and 17p, respectively. The traditional Haseman-Elston analysis of ADHD symptom severity resulted in only one LOD-score > 1, which was located on chromosome 17q (LOD ¼ 1.45). No clear overlap was found with the peak regions located at chromosome 7p, 9q and 15q in the original genome scan, nor with the regions described in an American genomewide scan for ADHD (Ogdie et al., 2003). Suggestive evidence for linkage with ODD symptom severity was found on chromosome 11p, a region not linked with ADHD in our previous genome scan. Further fine mapping and replication in different samples is needed to confirm that this region is linked with specific clinical features in ADHD.

P10.30 FEWER BEHAVIOUR PROBLEMS ASSOCIATED WITH THE T.7 DRD4 HAPLOTYPE AT 6 YEARS Lakatos K,1 Birkas E,2 Nemoda Z,2 and Gervai J1 1 Institute of Psychology, Hungarian Academy of Sciences, Budapest, Hungary 2 Semmelweis University, Budapest, Hungary In a subsample of 75 typically developing children participating in the longitudinal Budapest Infant-Parent Study, effects of early relationship and genetic factors were investigated on behaviour problems at the age of 6 years. Behaviour problems were assessed via maternal questionnaires. The SWAN (Swanson et al., ADHD.net) reports on inattention and hyperactivity/impulsivity. The Strengths and Difficulties Questionnaire (Goodman, 1997) provides information on hyperactivity, emotional, conduct and peer problems and also on prosocial behaviour. Using the standard cutoff points, children can be classified as ‘normal’, ‘borderline’ and ‘abnormal’. Mother-infant attachment was assessed at 12 months in the standard Strange Situation Test and served as a relationship characteristic. Infants’ D4 dopamine receptor (DRD4) genotype was determined from non-invasively collected buccal epithelial cells. The haplotype defined by the 521 C/T promoter SNP (genotypes: C or T) and the exon III 48 bp VNTR polymorphism (genotypes 7-repeat present or absent) was identified as T.7 present or absent. Multivariate analysis of variance revealed no influence of infant attachment on 6-year behavioural problems. However, there was a significant multivariate main effect of haplotype (P ¼ 0.045). Univariate effects were significant for emotional problems (P ¼ 0.048) and conduct problems (P ¼ 0.033). Children, who lacked the T.7 haplotype were reported as showing more of the above problem behaviours. For conduct problems, children with T.7 haplotypes were less likely to fall into the ‘borderline’ or ‘abnormal’ groups, than children of the T.7-absent group. No significant multivariate effect was found for the SWAN scales. In this subsample, a significant association was identified among the T.7 haplotype, emotional and conduct problems in preschool children at the age 6 years. Since these behavioural dimensions were not correlated (r ¼ 0.032), the results reflect indepent contributions of the genotype.

P10.31 ATTENTION DEFICIT HYPERACTIVITY DISORDER (DSM-IV COMBINED SUBTYPE) SHOWS A FAMILIAL RELATIONSHIP TO CONTINUOUS MEASURES OF ADHD SYMPTOM SCORES AMONG SIBLINGS Chen Wai,1 Knight Jo,1 Sham Pak,1 Banaschewski Tobias,2 Buitellar Jan,3 Ebstein R,3 Eisenberg J,3 Manor Iris,4 Miranda Ana,6 Oades B,7 Roeyers Herber,8 McGuffin Peter,1 Plomin Robert,1 Rothenberger Ariber,2 Sergeant Jo,9 Steinhausen H,11 Faraone S,10 Asherson P,1 and Gill M5

107

1

Institute of Psychiatry, London, United Kingdom University of Goettingen, Goettingen, Germany University of Nymegen, Nymegen, Netherlands 4 Univserity of Jerusalem, Jerusalem, Israel 5 Trinity College, Dublin, Ireland 6 University of Valencia, Valencia, Spain 7 University of Essen, Essen, Germany 8 University of Ghent, Ghent, Belgium 9 University of Amsterdam, Amsterdam, Netherlands, 10 Harvard University, Boston, United States 11 University of Zurich, Zurich, Switzerland 2 3

ADHD is a common, complex genetic disorder. Genetic risk factors are expected to be multiple, have small effect sizes when considered individually and to interact with each other and with environmental factors. Two major linkage scans using an affected sibling pair strategy have been completed and identify a number of potential loci containing ADHD susceptibility genes, although the two studies show only one region of partial overlap on chromosome 5. This almost certainly reflects the low power of linkage analysis to detect genetic variation conferring relatively small risks to the disorder. Further linkage studies are required to resolve the issues of power and the requirement for replication. Here we describe a quantitative trait locus (QTL) linkage study, the International Multi-center ADHD Gene project (IMAGE), which aims to take advantage of QTL methods to potentially increase power over more traditional study designs. ADHD probands are selected if they have the DSM-IV combined subtype and one or more siblings unselected for phenotype. Probands are diagnosed following structured parent interview (Parental Account of Childhood Symptoms, PACS) and completion of the Conners’ Teacher Rating Scale (DSM-IV checklist items). Other rating scales include Conners’ Parent Rating Scale and parent and teacher Strengths and Difficulties Questionnaires (SDQ). A requirement for QTL linkage analysis using this dataset is that continuous ratings of ADHD symptoms among siblings show a familial relationship to the DSM-IV combined subtype diagnosis. Although this may be inferred from previous twin studies, this has not been explicitly demonstrated. Using the first set of families for whom we have completed ascertainment (n ¼ 129), we estimated the correlation (r) between trait liability and the clinical disorder, (r ¼ [population mean  sibling mean]/[population mean  proband mean]). We found the following: Teacher Conners’ scale (r ¼ 0.3), Teacher SDQ (r ¼ 0.34), Parent Conners’ scale (r ¼ 0.19), Parent SDQ (r ¼ 0.01). These findings are comparable to published sibling correlations in dizygotic twin pairs and demonstrate the familial relationship between sibling ADHD symptom scores and combined type diagnosis in probands (twin studies using group heritability suggest this is predominantly due to shared genes). Near-zero sibling correlations are commonly seen for the parent-rated SDQ scale that uses a few global questions, explained by sibling contrast effects due to rater bias. Teacher measures show higher correlations and suggest that teacher measures may be more informative for QTL linkage. Twin data, however, suggest that parents and teachers may be using the same scales in different ways, representing only partially overlapping phenotypes and genotypes.

P10.32 GENETIC TESTING FOR AUTISM: CURRENT STATE OF CLINICAL PRACTICE IN THE SOUTHEASTERN UNITED STATES Wolpert CM, Boisen K, Cope H, Donnelly S, Cuccaro ML, and Pericak-Vance MA Duke University Med. Center, Durham, United States Genetic testing is available for several inherited disorders that demonstrate clinical overlap with autistic disorder (autism). However, it is not known what percentage of individuals with a confirmed diagnosis of autistic disorder will yield positive results for the currently available genetic tests. While it is estimated that 10 to 15 percent of individuals with autism have an identifiable genetic disorder, this estimate has never been studied in a formal manner. Currently, recommendations for genetic testing in children with autism are limited to those individuals who also have mental retardation, dysmorphic features, or both. The current study is a preliminary effort to examine the frequency with which genetic testing was conducted in a sample of individuals with autism. This sample consisted of 625 families with one or more children with a confirmed diagnosis of autism, which are enrolled in an autism genetic study. While the focus of this genetic

108

Abstracts

study is to identify currently unknown genetic etiology for autism, many of the study participants had genetic testing as part of clinical evaluations prior to entering the study. In this data set, less than 20 percent of individuals with confirmed autism have had at least one type of genetic testing. The most common genetic tests ordered in our sample were as follows: Fragile X testing (18%), chromosome analysis (14%), FISH (5.2%) amino acids (9%), and organic acids (8%). Some individuals had more than one genetic test. Two comparisons were conducted to examine potential differences related to genetic testing. In the first analysis, we compared adaptive functioning scores in participants with autism who had any genetic testing vs. those who did not have any genetic testing. The group who had genetic testing had significantly lower mean Vineland Adaptive Behavior Scale composite scores (nominal P ¼ 0.0001). Lower scores reflect poorer level of functioning. In the second analysis, the data set was grouped on the basis of family history for autism. The family history positive group had a significantly higher number of individuals who had any genetic test (nominal P ¼ 0.004). These data suggest that genetic testing is more likely to be ordered for individuals with lower adaptive functioning or in the event of a positive family history. Although less than 20% of this data set have had genetic testing, 72% of the participants have mental retardation, which is one of the indications for the currently recommended genetic testing for Fragile X syndrome, chromosome analysis, or both. Therefore, more than 50% of eligible individuals did not receive genetic testing suggesting that in this data set genetic testing is underutilized for individuals with autism.

P11.1 GENETIC ANALYSIS OF CIRCADIAN RHYTHM GENES PER1 AND PER2 IN PATIENTS WITH ALCOHOL DEPENDENCE Schumann G,1 Dahmen N,2 Soyka M,3 Rujescu D,2 Lathrop M,4 and Mann K1 1 Central Institute of Mental Health, Mannheim, Germany 2 University of Mainz, Mainz, Germany 3 University of Munich, Munich, Germany 4 Centre National de Ge´notypage, Evry, France 1 Molecular Genetics Laboratory, Central Institute of Mental Health (CIMH), 68159 Mannheim, Germany Alcohol intake and alcohol dependence are thought to be linked to circadian rhythmicity in humans and animal models. Recently, an interaction of circadian rhythm genes and alcohol drinking behaviour was described. Per-1 knock out mice showed an increased loss of righting reflex (LORR) as a sign of increased alcohol sensitivity, while in wild type mice LORR was inversely correlated with Per-1 expression. Per-2 knock out mice showed increased alcohol intake as compared to wild type mice. We performed a SNP-discovery and frequency analysis of the human Per1 and Per2 genes and conducted association studies and haplotype analysis of the informative SNPs identified in a total of 990 subjects. 18 genetic variations of the Per1 gene were identified in the regulatory domains, exons and exon-intron boundaries. Genotyping and haplotype analysis of 6 informative SNPs revealed no stable association with alcohol dependence or other relevant phenotypes. Negative results were confirmed using equivalence testing. In the case of Per2 we identified 11 gene variations, 6 of which were selected for genotyping in a sample of 215 patients with a detailed assessment of alcohol intake. We found a significant association of high vs. low alcohol intake with one SNP (P ¼ 0.020, OR 2.155), and identified haplotypes consisting of 4 SNPs with a significant association between high vs. low amount of alcohol intake (global P-value by permutation test ¼ 0.0075). Thus, our findings support the results of the Per2 mutant mouse studies and suggest a role of the hPer2 gene in the regulation of alcohol consumption in human alcohol dependent patients.

P11.2 INTERACTION OF POLYMORPHISMS IN THE DOPAMINE D4 RECEPTOR AND SEROTONIN TRANSPORTER GENES IN HEROIN ADDICTION Barta C,1 Szilagyi A,2 Boor K,2 Gaszner P,3 Katai-Fodor I,4 and Sasvari-Szekely M1 1 Department of Medical Chemistry, Molecular Biology and Pathobiochemistry, Semmelweis University, Budapest, Hungary 2 Institute of Pharmacology and Pharmacotherapy, Semmelweis University, Budapest, Hungary 3 National Institute of Psychiatry and Neurology, Budapest, Hungary

4 Dr. Tere´zia Farkasinszky’’ Centrum for Young Drug Addicts of Local Council of the City of Szeged, Szeged, Hungary

Among the numerous genetic risk factors contributing to the development of drug addiction a key role of the reward system has been implicated through the dopaminergic mesocorticolimbic pathway, which is under modulation by the serotonin system. The presented analysis focuses on polymorphisms of the dopamine D4 receptor (DRD4) and the serotonin transporter (SERT) gene in a case-control study of 120 drug abusing subjects (100 heroin abusers) and 384 healthy Caucasian (Hungarian) controls. The investigated polymorphic sites include the coding region (exon 3 VNTR) and the 50 upstream region (521 CT and 120 bp duplication) of the DRD4 gene, the 50 upstream region (5-HTTLPR), and the intron 2 polymorphism (STin2) of the SERT gene. A significant interaction was found between the 521 C/C genotype in the promoter of the DRD4 gene and the 5HTTLPR s/s genotype (P 0.01, OR: 3.17) underlining the cross-talk between the two neurotransmitter systems. P11.3 GENETIC INVESTIGATIONS IN POLISH FAMILIES WITH ALCOHOL PROBLEMS—PRELIMINARY RESULTS. Samochowiec J,1 Kucharska-Mazur J,1 Pelka-Wysiecka J,1 Syrek S,1 Grzywacz A,1 Rommelspacher H,2 Samochowiec A,3 and Horodnicki J1 1 University MSKP Szczecin, Szczecin, Poland 2 Department of Clinical Neurobiology, FU, Berlin, Germany Alcohol dependence is a multifactorial disorder caused by complex interaction of genetic and environmental factors. To avoid the problem of population’s stratification we used the affected children/parents samples (trios). We have examinated 85 whole families (trios: proband þ parents) and 22 incomplete families (proband þ one of parents). Using CIDI (Composite Interntional Diagnostic Interview) evaluation of ICD 10 alcohol was done. Control group consisted of ethnically, gender and aged matched 183 healthy volunteers. We have analysed by using PCR methods seven polymorphisms and estimated differences in alleles transmission by TDT test and calculated associations by w2-test by using SPSS programme: DRD2 gene (ins/ del 141C in promoter region, Taq IA, Exon 8), TaqIA A1–A2 allele was found to be significantly more frequent transmitted from alcohol parents (TDT: P ¼ 0,001) to their alcohol dependent sons. We found, that the presence of the 10 copy (A10) allele of the dopamine transporter gene (DAT1) was associated with more severe form of alcohol dependence and TDT analysis proved significant increase in each transmission from parents to alcoholics (P ¼ 0,033). No significantly TDT and associations were found when examining 5 HTT_LPR (44 bp del in promoter), MAO-A (30 bp VNTR in promoter region) and COMT (A158G) genes polymorphisms. More frequent transmission of DAT (A10) and DRD2 Taq I (A2) alleles was found in affected families. Due to still rising sample of alcohol families these results should be regarded as preliminary. Supported by grants: KBN 3PO 5D 14622 and BMBF/ POL 01/063.

P11.4 RELATIONSHIPS BETWEEN SOME CANDIDATE GENES POLYMORPHISMS AND CLONIGER’S TYPOLOGY OF ALCOHOLISM IN THE SAMPLE OF POLISH FAMILIES WITH ALCOHOL PROBLEMS-PRELIMINARY RESULTS Kucharska-Mazur J,1 Pelka-Wysiecka J,1 Syrek S,1 Grzywacz A,1 Horodnicki J,1 Rommelspacher H,2 Samochowiec A,3 and Samochowiec J1 1 Department Psychiatry PAM, Szczecin, Poland 2 Department Clinical Biology FU, Berlin, Germany 3 University MSKP, Szczecin, Poland There are reports of the genetic background of Cloninger’s alcoholism typology. We have examinated 107 trios of families with alcohol problem. Entire group of alcohol dependent subjects was divided in to type I and type II alcoholism according to Cloninger’s typology. SAS (Self Rating Anxiety Scores) and the SDS (Self Rating Depression Scores) was measured. Control group consisted of ethnically, gender and aged matched 183 healthy volunteers. 7 polymorphisms were investigated: DRD2 gene (ins/del 141C in promoter region, Taq IA, Exon 8), DAT1 gene (40 bp VNTR), 5-HTT_LPR (del 44 bp in promoter), MAO-A (30 bp VNTR in promoter), COMT (A158G). We found no associations between the above mentioned genes polymorphisms and the Cloninger’s type I and II subgroups.There were statistically

Abstracts significant differences in SDS and SAS total scores between subgroups of type I and type II alcoholics: DRD2 Taq IA polymorphism: type I subjects with A2/A2 vs. A1/A2 genotype scored in SDS: 46 vs. 49 points (P ¼ 0,02), DRD2 Exon 8 polymorphism: type II subjects with G/G vs. A/ A vs. A/G genotypes scored in SDS: 54 vs. 48 vs. 42 points (P ¼ 0,002), subjects type II with G/G vs. A/A vs. A/G genotypes scored in SAS: 55 vs. 41 vs. 40 points (P ¼ 0,036), 5 HTT_LPR polymorphism: subjects type I and type II with s/s genotypes scored significantly higher in SAS (P ¼ 0,042 and P ¼ 0,022, respectively). There were no significant differences observed between type I and II alcoholics in MAO - A, COMT and DRD2 ins/del 141C genes polymorphisms. Our results did not confirm previously reported genetic background of Cloninger’s alcoholism typology. We found that genetic factors may play some role in development of the anxiety and depresiveness in alcoholics subgroups divided by Cloninger’s typology. Supported by grants: KBN 3PO 5D 14622 and BMBF/POL 01/063. P11.5 ASSOCIATION OF THE SEROTONIN TRANSPORTER WITH SMOKING BEHAVIOR Gritsenko I,1 Kremer I,2 Bachner-Melman R,3 and Ebstein RP3 1 Herzog Hospital, Jerusalem, Israel 2 Emek Hospital, Afula, Jamaica 3 Hebrew University, Jerusalem, Israel The serotonin transporter (SERT) promoter region polymorphism, 5HTTLPR and an intronic VNTR, were genotyped in families recruited in an ongoing molecular genetic study of temperament to examine association with smoking behavior.Full information regarding smoking and genotype was available in 330 families and 247 ever smokers were identified (past ¼ 55, present ¼ 192). Subjects also completed the TPQ personality inventory. Association between genotype, personality and smoking phenotype was tested using a robust family design working in a variance-components framework (QTDT) as well as by case control analysis. There was a significant excess of the 5-HTTLPR long allele and the 12 VNTR repeat in both current and past smokers compared to never smokers. When ever smokers (past and current) were compared to never smokers there was a significant excess of the 5-HTTLPR long allele and the 12 repeat VNTR (chi-square ¼ 14.89, P ¼ 0.0001) and the estimated risk was 1.37 (95% CI ¼ 1.17 to 1.61). The results from the population design were confirmed in the family based analysis for both current smokers (P ¼ 0.0004), ever smokers (P ¼ 0.00001) and past smokers (P ¼ 0.04). No association was observed between two quantitative measures of the smoking phenotype (Fagerstrom scores and number of cigarettes smoked per day) and SERT. A weak (P ¼ 0.04), association is observed between Novelty Seeking (NS) and the VNTR polymorphism and between 5-HTTLPR and TPQ Reward (P ¼ 0.03). Smokers regardless of gender score significantly higher on NS and do not differ on Harm Avoidance or Reward. The main finding of this investigation was a highly significant association between two SERT polymorphisms and smoking phenotype, most likely initiation, employing a robust family as well as population design. Although smokers scored significantly higher on Novelty Seeking, and SERT was weakly associated with this temperament trait, mediation analysis failed to substantiate the hypothesis that Novelty Seeking partially mediates the effect of SERT on smoking. SERT appears to independently contribute to both Novelty Seeking and smoking.

P11.6 IONOTROPIC GLUTAMATE RECEPTOR GRIK3 SER310ALA FUNCTIONAL POLYMORPHISM IS RELATED TO DELIRIUM TREMENS IN ALCOHOL-DEPENDENT INDIVIDUALS Preuss UW,1 Zill P,2 Koller G,2 Bondy B,2 and Soyka M2 1 Johanna-Odebrecht-Stiftung, Greifswald, Germany 2 Ludwigs-Maximilians-Universita¨t, Mu¨nchen, Germany Up regulation of the glutamatergic neurotransmission resulting from chronic ethanol intoxication may cause a hyperexcitable state during alcohol withdrawal which may lead to seizures and delirium tremens. The aim of our study was to evaluate the association between a history of alcohol withdrawal-induced seizures and delirium tremens, and a functional polymorphism (Ser310Ala) of the GRIK3 gene coding for the glutamatergic kainate receptor subunit GlurR7 in a sample of wellcharacterized alcoholics compared to controls. 233 patients meeting DSM IV alcohol dependence criteria and 309 controls, all of German

109

descent, were investigated. GRIK3 functional polymorphism was determined using PCR (Polymerase Chain Reaction) of lymphocyte DNA. History of alcohol withdrawal-induced delirium tremens and seizures were obtained using the SSAGA (Semi-Structured Assessment of Genetics in Alcoholism). Data were cross-checked with inpatients’ clinical files. While a significant relationship between history of delirium tremens and the Ser310 allele was detected, no significant results were obtained for alcohol withdrawal-related seizures. Though this result is suggestive for a significant role of this polymorphism in the pathogenesis of delirium tremens in alcoholdependent individuals, further investigation and replication is warranted. P11.7 ASSOCIATION BETWEEN TAQI A POLYMORPHISM OF DRD2 GENE AND CEREBRAL FORM IN METHAMPHETAMINE PSYCHOSIS Harano M,1 Uchimura N,1 Ueno M,1 Abe H,2 Ishibashi M,3 Iida N,4 Tanaka T,5 Maeda H,1 Sora I,6 Iyo M,6 Komiyama T,6 Yamada M,6 Sekine Y,6 Inada T,6 Ozaki N,6 Iwata N,6 and Ujike H6 1 Department of Neuropsychiatry, Kurume University School of Medicine, Fukuoka, Japan 2 Department of Radiology, Kurume University School of Medicine, Fukuoka, Japan 3 Jyuzen Hospital, Fukuoka, Japan 4 Kaisei Hospital, Fukuoka, Japan 5 Taihozan Hospital, Fukuoka, Japan 6 JGIDA (Japanese Genetics Initiative for Drug Abuse), Okayama, Japan Methamphetamine is taken into cytosol by dopamine transporters on the synaptic terminals of dopamine neurons, and endogenous dopamine is concurrently released through the transporters by carrierexchange mechanisms, which results in a superfluous increase in the dopamine concentration in the synaptic clefts, and brings abusers into over arousal and euphoria. Certain methamphetamine abusers fall into psychosis, and show symptoms very similar to schizophrenia. Recently, the existence of gene polymorphisms that affect the susceptibility to methamphetamine psychosis has been proposed. Of these, the TaqI A polymorphism of the dopamine receptor D2 (DRD2) gene is the polymorphism that has attracted most interest. For the purpose of clarifying the relation between the TaqI A polymorphism and the disease, we performed association analysis between the TaqI A polymorphisms and cerebral form as an endophenotype. The cerebral form was analyzed with Voxel-Based Morphometry using the data of a brain MRI. The results indicate that there is a positive correlation (P ¼ 0.027) between the disease and A2-allele of the TaqI A polymorphism, and that in patients left temporal lobe (midfrontal gyrus and frontal part of superior temporal gyrus) and the putamen are narrow compared with a healthy person. Moreover, in patients with the A2/A2genotype the putamen and left frontal lobe undersurface (orbital gyrus) are narrow compared with patients with other genotypes. These results suggest that the A2-allele of the TaqI A polymorphism is a strong risk factor for methamphetamine psychosis, and that the left temporal lobe (the midfrontal gyrus and frontal part of the superior temporal gyrus) and the putamen is an etiologic site of the disease. These results suggest further the possibility that the orbital gyrus is an etiologic site of aggravation of the disease and is closely involved with the TaqI A polymorphism. This study was approved by the Ethics Committee of Kurume University. After a complete description of the study was provided, written informed consent was obtained from all subjects before study entry.

P11.8 THE DRD4 GENE, DRINKING AND URGE TO DRINK AFTER NALTREXONE: PRELIMINARY RESULTS FROM ECOLOGICAL MOMENTARY ASSESSMENT McGeary JE, Monti PM, Rohsenow DJ, Tidey JW, and Swift RM Brown University, Providence, United States Numerous studies have implicated the VNTR (variable number of tandem repeats) polymorphism in the D4 dopamine receptor (DRD4) gene in alcohol dependence and as a moderator of alcohol urge in a laboratory setting. Drinkers with alcohol abuse or dependence (N ¼ 72) entered daily drinking-related data on palmtop computers for 4 weeks (3 weeks randomized to medication or placebo) in a study of naltrexone (NTX) vs. placebo. Participants were genotyped for the DRD4 VNTR to

110

Abstracts

examine whether this genetic variant would also predict drinking levels, urge to drink, or alcohol’s subjective effects in the natural environment. Participants who had at least one copy of the long allele (7þ repeats) were found to be significantly more likely to smoke (P < .01) at baseline and reported a greater number of drinking days (in the previous 90 days) at a trend level (P ¼ .07) with no difference in quantity of alcohol consumed. Urge during a placebo lead-in week as measured by Ecological Momentary Assessment (EMA) did not differ by DRD4 variant. However, participants on placebo with the long DRD4 variant reported significantly higher urge compared to those with the short version and to those with the long variant DRD4 on NTX (P ¼ .05). After the first drink of the day in the natural environment, participants with the short variant of the DRD4 did not differ in stimulation while those with the long variant reported more stimulation after NTX than placebo participants. Despite significant moderating effects on the subjective reports, the DRD4 gene did not significantly predict the number of alcoholic drinks consumed after medication or placebo within the assessment period. If replicated in the larger sample, these findings suggest a differential response to naltrexone treatment due to genotype and may help explain conflicting results in alcoholism clinical trials with naltrexone.

drinking risk pattern was found (OR: 1.8, CI95%: 1.2–2.9, Fisher’s P < 0,01). The second study investigated paternal alcoholism by Brazilian schizophrenics (N ¼ 40) and found a significant higher proportion of female schizophrenics with a family history of paternal alcoholism (RR ¼ 2.7, CI95%: 1.3–5.9, Fisher’s P < 0,01). The third study investigated longitudinally the presence of psychiatric symptoms and the cognitive development of German children of alcoholics (COAs) from birth to eleven years of age and found a distinct pattern of development of psychiatric symptoms by the female children of alcoholics. Despite of a general risk pattern for both male and female offspring of alcoholic fathers there would be gender differences in the types of psychiatric problems in childhood and adulthood. In childhood female children of alcoholic fathers show a age related tendency to more emotional problems within the dimensions of anxiety and depression. In adulthood the pattern of more emotional problems remains identifiable within the group of pregnant women of the COAs group and if we look more close to a diagnose group, like schizophrenic women, we found again a distinct pattern, whereby paternal alcoholism emerge as a discriminative antecedent. The mechanisms underlying the clinical findings should be more investigated as well ethnic differences in a multiethnic population, as in the case of the Brazilian samples.

P11.9 THE DRD4 GENE IN A NALTREXONE LAB STUDY: PRELIMINARY RESULTS ON CUE-INDUCED ALCOHOL CRAVING McGeary JE, Monti PM, Rohsenow DJ, Tidey J, and Swift R Brown University, Providence, United States

P11.11 INVESTIGATION OF THE CATECHOL-OMETHYLTRANSFERASE (COMT) GENE POLYMORPHISM IN A SAMPLE OF COCAINE DEPENDENTS Cunha N,1 Cordeiro Q,1 Guindalini C,1 Messas GP,2 Castelo A,2 Laranjeira R,2 and Vallada H1 1 School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil 2 School of Medicine, Federal University of Sa˜o Paulo, Sa˜o Paulo, Brazil

Efforts to identify individual genes responsible for alcohol dependence have led to inconsistent results. Previous studies have implicated the VNTR (variable number of tandem repeats) polymorphism in the D4 dopamine receptor (DRD4) gene in alcohol dependence and as a moderator of alcohol craving induced with priming doses. To examine whether this genetic variant would also predict drinking levels and differential cue-induced craving among heavy social drinkers (some of whom were alcohol dependent), 50 male and female participants in a larger study of the mechanisms of naltrexone’s action were genotyped for the DRD4 VNTR and were presented with two trials of alcohol cues. Participants who had at least one copy of the long allele (7þ repeats) were found to report significantly more drinking occasions in the last month but not more heavy drinking occasions. Suggesting that, contrary to expectation, the DRD4 gene may moderate the likelihood of drinking rather than the likelihood of drinking heavily. Urge for alcohol before alcohol exposure did not vary by genotype. During the first alcohol exposure, there was a trend for a gene by medication interaction. During the second alcohol exposure, there was a significant gene by medication interaction with a medium effect size. Post hoc tests showed that in the placebo condition those with long versions of the DRD4 gene reported increased urge compared to those with only the shorter version, but within the naltrexone condition there was no effect of genotype on urge. If replicated in the larger sample, these findings may suggest a differential response to naltrexone treatment due to genotype that may help explain conflicting results in clinical trials.

P11.10 PATERNAL ALCOHOLISM, FEMALE GENDER AND PSYCHIATRIC RISK Furtado EF,1 Webster CMC,1 Pinheiro SN,1 Laucht M,2 and Schmidt MH2 1 University of Sao Paulo, Ribeirao Preto, SP, Brazil 2 University of Heidelberg, Mannheim, Germany Paternal alcoholism has been reported as a risk factor for alcohol related problems in the offspring. Some findings support the assumption of a hereditary transmission of risk. There are evidences of a relationship between dopamine receptors polymorphisms and a psychopathological developmental pathway. Most findings point out the higher risk for the male offspring of alcoholics. Fewer studies have been dedicated to investigate the role of paternal alcoholism in the psychiatric risk of the female offspring. To report preliminary results from three different research settings that offer clinical evidences of gender-related pathway with some specificity for the female offspring of alcoholic fathers. The first study investigated alcohol risky use by Brazilian pregnant women (N ¼ 450) and related factors. A strong relationship between paternal alcohol problems and psychiatric symptoms (OR: 1.6, CI95%: 1.1–2.3, X2: 4.8, df: 1, P < 0,05) and alcohol

Drug dependence is due to a complex interaction between genetic and environmental factors. Much evidence of the reinforcing properties of cocaine is associated with the dopaminergic system. Thus catechol-Omethyltransferase (COMT), which is an enzyme that metabolizes dopamine, is a candidate target for genetic investigations in the susceptibility to develop cocaine dependence. The present study examines the distribution of the Val-158-Met variant in the COMT gene in a Brazilian sample consisting of 462 cocaine dependents and 237 matched healthy controls. There was no difference in the allelic (Val ¼ 62.98% and Met ¼ 37.01% for patients, Val ¼ 62.86% and Met ¼ 37.13% for controls, X2 ¼ 0.00, 1 d.f., P ¼ 0.96) neither in the genotypic (X2 ¼ 1.30, 2 d.f., P ¼ 0.52) distributions between cocaine dependents and controls. In conclusion, the Val-158-Met polymorphism is not a genetic risk factor for cocaine dependence in our Brazilian sample.

P11.12 TESTING GENETIC SUSCEPTIBILITY FOR ALCOHOL INDUCED SERIOUS MEDICAL CONDITIONS Bondy B, Koller G, Lieb M, Zill P, and Soyka M Psychiatric Clinic, Munich, Germany Alcohol abuse has been linked to multiple medical complications, such as nonischemic, dilated cardiomyopathy or intracranial hemorrhage. Especially the intracranial hemorrhage was related to the dosage of alcohol misuse and acute arterial hypertension was proposed as likely mechanism for this complication. However, although many heavy alcohol users have only subclinical signs of medical complications and only part of them develop deleterious organic disorders, individual susceptibility differences may exist due to genetic liability factors. The aim of the present study was to investigate several candidate genes being under debate in relation to cardiac disease and hypertension in patients with alcohol abuse during withdrawal. Enrolled were 275 chronic alcoholic patients (214 m, 61 f, age 41,3  9,3 years) being hospitalized for withdrawal under clinical conditions. 181 of them were acutely intoxicated (0.8–3.75). Blood pressure and heart rate were monitored at the time of hospitalization and during withdrawal. Additionally, cholesterol, HDL, LDL, triglycerides were measured as cardiovascular risk factors. We have genotyped the patient’s DNA for functionally active polymorphisms within the following candidate genes: the ACE I/D, G-proteinß3-C825T, GNAS1-T393C, beta2adrenergic-Arg16Gly, beta2-adrenergic- Glu27Gln and the HTTLPRL/S. At hospitalization we observed a weak correlation between the daily alcohol consumption and systolic (P ¼ 0.05) and diastolic (P ¼ 0.01) blood pressure, which was more pronounced in relation to the actual alcohol concentration (P ¼ 0.001). We have not found a

Abstracts relation between the ACE I/D, G-proteinß3 C825T, beta2-adrenergic Glu27Gln, the HTTLPR L/S variants and any of the clinical parameters investigated. Concerning the beta2-adrenergic Arg16Gly polymorphism, higher Rrsys and Rrdias values were observed in Gly-allele carriers (P ¼ 0.029) than in Arg homozygotes, but the increase in blood pressure during detoxication was not further related to its genotypes or alleles. In contrast to that a relation between the GNAS1 T393C polymorphism and the increase in Rrsys and Rrdias during detoxication (P ¼ 0.002) was observed. These findings suggest that the beta-adrenergic stimulatory guanine-binding-protein system is playing a role in cardiovascular function in alcoholic patients. Functionally active variants of both genes, the beta2-adrenoceptor and the the G-protein-stimulatory unit might thus be candidate determinants for hypertension or hypertensive reactions in alcoholic patients and thus be prone to identify persons at risk for severe medical complications. during intoxication a. P11.13 THE PRODYNORPHIN GENE PROMOTER IS ASSOCIATED WITH METHAMPHETAMINE DEPENDENCE IN JAPANESE POPULATION Nomura A,1 Ujike H,1 Tanaka Y,1 Hrano M,2 Inada T,2 Yamada M,2 Komiyama T,2 Sekine Y,2 Iwata N,2 Sora I,2 Iyo M,2 Ozaki N,2 and Kuroda S1 1 Department of Neuropsychiatry, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan 2 Japanese Genetics Initiative for Drug Abuse, Okayama, Japan Methamphetamine (METH), belongs to amphetamines, has been the most popular abused drug in Japan. Patients with METH dependence tend to commit serious crimes, and METH dependence has been severe social problems in Japan. Family, twin and adoption studies have indicated that genetic factors substantially influence the vulnerability to develop psychoactive substance use disorder including psychostimulants such as amphetamines and cocaine. Several candidates of genetic factors were found to associate significantly with vulnerability to psychoactive substance use disorder, but the precise involvement in development and the genetic mechanisms remain to be elusive. Prodynorphin (PDYN) belongs to the family of opioid peptide precursor proteins and yields dynorphin peptides. There are lines of evidence that dynorphin peptides interact with psychostimulants. It was reported that the PDYN gene promoter had 1 to 4 repeats of a 68-bp element and the allele with 3 or 4 repeats have significantly higher promoter activity of the PDYN gene. We hypothesized that this functional polymorphism was associated with the vulnerability to METH depnendence. We investigated the allelic and genotypic frequencies of a polymorphism in the PDYN gene promoter on 143 patients with METH dependence and 209 healthy controls in Japanese population. This study was performed after obtaining approval from the ethics committees of each institutes of JGIDA, Japanese Genetics Initiative for Drug Abuse, and all subjects provided written informed consent for the use of their DNA samples for this research. A 3 or 4 repeat-allele in the PDYN gene promoter, which has significantly higher promoter activity, was significantly higher in patients with METH dependence than controls (w2 ¼ 9.45, P ¼ 0.0021). A 3 or 4 repeat-allele in the PDYN gene promoter, which was shown to produce significantly higher transcription activity of the PDYN gene than that with 1 or 2 repeat-allele, is a genetic risk factor for development of METH dependence (odds ratio: 1.83).

P11.14 ASSOCIATION STUDY OF APOLIPOPROTEIN E POLYMORPHISM WITH SUBTYPES OF ALCOHOLISM Parsian A,1 Zhang ZH,1 and Cloninger CR2 1 University of Arkansas for Medical Sciences, Little Rock, United States 2 Washington University, St. Louis, United States Apolipoprotein E (ApoE) is a polymorphic protein with three major isoforms (E2, E3, and E4). It is present in high concentration in brain and synthesized by actrocytes and can be internalized by nerve cells utilizing a receptor mediated pathway. To determine the role of ApoE in the development of alcoholism, especially type II and severe form, we screened a sample of alcoholics (N ¼ 131) and normal controls (N ¼ 87) with polymorphism of this gene. The alcoholic sample was categorized into type I and II based on antisocial personality and into mild, moderate, and severe based on the medical complications. Comparison of allele and genotype frequencies between total alcoholics, type I and II, and normal controls was not statistically significant. Similarly,

111

allele and genotype frequencies in mild, moderate, and severe groups were not different from controls. Allele and genotype frequencies of ApoE in subtypes of alcoholics were very similar and their differences were not significant. We concluded that the ApoE gene is not playing any major role in the development of alcoholism or subtypes of alcoholism. P11.15 ALCOHOL DEPENDENCE AND POLYMORPHISMS IN ALCOHOL DEHYDROGENASE 4 Guindalini C2, Scivolleto S,3 Zibermann M,3 Breen G,1 and Zatz M2 1 Section of Genetics, Institute of Psychiatry, King’s College London, London, United Kingdom 2 Human Genome Research Center, Institute of Bioscience, University of Sao Paulo, Sao Paulo, Brazil 3 GREA, Institute of Psychiatry, University of Sao Paulo, Sao Paulo, Brazil Alcohol metabolism strongly influences drinking behavior and the likelihood to be dependent. Most ethanol is oxidised to acetaldehyde and acetate, mainly by alcohol dehydrogenases (ADHs types 1–4) and aldehyde dehydrogenase (ALDH). Human ADH4 enzyme is found mainly in the liver and at intoxicating levels of alcohol it may account for as much as 40% of the total ethanol oxidation rate. Three polymorphisms were reported in the promoter of ADH4. The authors analyzed the effects of 4 different haplotypes upon gene expression. Two promoters with A at 75 bp had relatively high activity in comparison to the promoters with C at the same position. They hypothesized that the 75A allele, should be protective against alcoholism similarly to the protective effect of the higher Vmax alleles of ADH1B and ADH1C genes. An association study was performed to evaluate the influence of the three functional promoter polymorphisms of the ADH4 gene on alcoholism. DNA from alcohol-dependent patients (n ¼ 92) and healthy controls (n ¼ 92) was sequenced and genotyped for all three polymorphisms. Variants at the 75 bp (C allele, OR ¼ 1.6, 95% C.I.: 2.4–1.05, P < 0.05) and 159 bp (A allele, OR ¼ 2.2, 95% C.I.: 3.3–1.4, P < 0.001) positions were associated with alcoholism. Individuals with haplotypes carrying both risk alleles were three times more likely to develop alcoholism (OR ¼ 2.9, 95%CI: 4.73–1.89, P < 0.00001). As suggested, the 75A allele showed to be protective against alcoholism. Moreover it was reported that the T,A,C and T,A,A promoters had lower and higher levels of expression, respectively. Our results agree with the authors as the first haplotype, found mainly among patients was a risk factor in our study while the second, significantly over-represented in the control group, was protective. The possibility that the association of these ADH4 polymorphisms could be due to linkage disequilibrium with the ADH1B locus was investigated and shown not to be the case in our sample. These preliminary results which require further confirmation suggest that ADH4 may play a major role in the etiology of alcoholism. The association requires further study and replication but is functionally plausible and has a large effect size.

P11.16 SH-SY5Y CELLS ARE A MODEL TO STUDY DOPAMINE D2 RECEPTOR (DRD2) AND TRANSPORTER (DAT) ADAPTATION DURING ETHANOL AND DRUG EXPOSURE Wernicke C, Finckh U, and Rommelspacher H Charite´ Campus Benjamin Franklin, Clinical Neurobiology, Berlin, Germany Mesolimbic dopamine activity is conceived to underlie the transition from controlled to uncon-trolled use of alcohol and other drugs. Further studies demonstrated that coactivation of the dopa-mine D1 and D2 receptors is necessary to regulate alcohol self-administration. The DRD2 is ex-pressed in two isoforms generated by alternative splicing. The long isoform (DRD2L) is localized postsynaptically, the short one (DRD2S) presynaptically. In dopaminergic neurons, dopamine re-lease is downregulated by the stimulation of presynaptic DRD2 (autoinhibition). The action is ter-minated by the dopamine transporter. To investigate adaptive changes during ethanol and drug ex-posure, a suitable model, expressing both isoforms of DRD2 and the DAT is required. The human neuroblastoma cell line SH-SY5Y consists of a mix of neuron-like and epithelium-like cell popu-lations. Retinoic acid (RA) treatment induces transdifferentiation and consecutive treatment with RA and brain-derived neurotrophic factor (BDNF) maintains suppression of epithelium-like cells in longterm culture.

112

Abstracts

Expression analysis of DRD2 and DAT was performed by Western Blot (WB) and real-time RT-PCR (rtPCR) in undifferentiated, RA and RA/ BDNF transdifferentiated cells. rtPCR was done using FRET probes in a LightCycler with the h-PBGD housekeeping gene set (Roche) as standard. Additionally, [3H]dopamine uptake was tested. WB showed a DAT-specific band at 73 kDa in undifferentiated, RA and RA/BDNF treated cells, in the latter ones to a lower extent correlating to preliminary results in rtPCR. High-affinity uptake of [3H]dopamine was ob-served in undifferentiated cells. The DRD2 signal at 65 kDa was enhanced after RA/BDNF treat-ment. mRNAs of both isoforms were detectable at a ratio DRD2L/DRD2S  100:1. rtPCR revealed a RAinduced increase of mRNA of DRD2L(1.4) and DRD2S(2.4) and a further increase by RA/BDNF exposure of 16- and 30-fold, respectively. Thus, RA/BDNF treatment promotes the ex-pression of some features of dopaminergic neurons in SH-SY5Y cells. They are a promising model to study the adaptive regulation during ethanol and drug exposure specifically the expression of DRD2 isoforms and DAT, being major functional determinants of brain dopaminergic neurons. P11.17 THE ROLE OF DOPAMINE DRD2 TAQ1A VARIANTS IN INFLUENCING TREATMENT RESPONSE TO NICOTINE REPLACEMENT THERAPY Stapleton J, O’Gara C, Ball D, and Sutherland G Institute of Psychiatry, London, United Kingdom Most recent initiatives in the field of pharmacogenetics and nicotine replacement therapies have focused on the development of targeted therapies based on genotype. Historically, the dopamine system has been a popular candidate for influencing smoking initiation and persistence due to its role in reward processes. Recent evidence has suggested that genetic influences on treatment outcome may be different between the sexes, and that allelic variants of the dopamine receptor DRD2 Taq1A may be associated with increased efficacy of the nicotine patch in females within a randomised trial setting. We examined the effect of genotype on nicotine patch treatment outcome in females in a naturalistic smoking cessation clinic setting. 230 patients who chose to attend a smoking cessation clinic and use the NRT patch were interviewed and agreed to donate a cheek swab sample from which DNA was extracted. Participants were genotyped using the Taqman method for the presence of the variant T allele of the dopamine receptor DRD2 32806 (CT or TT genotype). Abstinence from smoking was confirmed at week 1, 2, 3 and 4 weeks post quit date by self report and expired carbon monoxide concentration of less than 5 parts per million. The interaction of treatment outcome, genotype and sex was examined in a logistic regression model. Treatment outcome did not differ with genotype at any of the time points in the 4 week post treatment period. Our results do not support the recent interest in the dopamine receptor DRD2 as a candidate for predicting treatment outcome in smoking cessation.

P11.18 GENOME-WIDE EXPRESSION ANALYSIS IN A MOUSE MODEL OF ADDICTION Grice DE,1 Golden GT,2 Smith GG,2 and Berrettini WH2 1 New Jersey Medical School—UMDNJ, Newark, United States 2 University of Pennsylvania, Philadelphia, United States C57BL/6J (C57) and DBA/2J (DBA) inbred mouse strains have been extensively characterized and have significant differences in their behavioral and pharmacological responses to morphine. For example, in a two-bottle choice paradigm, C57 mice will voluntarily consume 200–300 mg/kg/day of a morphine and saccharin mixture while DBA mice consume 10–20 mg/kg/day. Thus, these inbred strains provide a model for the study of morphine consumption and addiction. Quantitative trait loci (QTL) analysis indicates that a major locus underlying the difference in morphine consumption between the two strains is on murine chromosome 10, at the site of the mu opiate receptor (MOR) locus. Our studies on the primary coding sequence of the major isoforms of MOR gene demonstrate no coding variation between these two strains, suggesting that differences in gene expression of MOR isoforms may contribute to differences in morphine-seeking behavior in these two strains. Remarkably, C57 and DBA strains differ in substanceseeking behavior for several drugs of abuse, consistent with the polysubstance nature of human substance use disorders. Surprisingly, QTL analyses in mice for various substances do not show shared QTL loci. This can either indicate that each of the substances has unique, non-

overlapping loci, or that the QTL analyses fail to identify susceptibility loci that are shared across substances. Genome-wide expression studies may identify genes that are expressed differently between the two strains and also underlie their divergence in substance-seeking behavior. We have carried out genome-wide microarray analysis in C57 and DBA under basal conditions and following chronic (4 day) treatment with morphine. Several genes were identified that show altered expression in the two strains in the absence and/or presence of morphine. Some of these genes, such as COMT, may contribute to substance-seeking behavior in the C57 strain. Validation of the microarray results by quantitative, real-time PCR is now being carried out. P11.19 ASSOCIATION BETWEEN THE SEROTONIN TRANSPORTER GENE AND ALCOHOL DEPENDENCE IN THE IRISH AFFECTED SIB PAIR STUDY OF ALCOHOL DEPENDENCE Prescott CA,1 Sullivan PF,2 Vittum J,1 Patterson DG,3 van den Oord EJ,1 Walsh D,4 Kendler KS,1 and Riley BP1 1 Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, Virginia, United States 2 Department of Psychiatry, University of North Carolina, Chapel Hill, North Carolina, United States 3 Shaftsbury Square Hospital, Belfast, Northern Ireland, United Kingdom 4 Health Research Board, Dublin, Ireland The short variant of the serotonin transporter (httlpr-ins/del) has been associated with alcohol dependence in several association studies (Hammoumi et al., 1999; Lichterman et al., 2000). Other studies have found an association only for subgroups, such as alcoholics with antisocial behavior (Hallikainen et al., 1999; Ishiguro et al., 1999; Sander et al., 1998) or suicide attempts (Gorwood et al., 2000). We investigated this issue using data from the Irish Affected Sib Pair Study of Alcohol Dependence (IASPSAD, Prescott et al., 2002). Based on power analyses, we selected a sample size of 328 cases and 328 controls. Cases met criteria for lifetime DSM-IV alcohol dependence and were randomly selected from affected individuals in participating families. Controls included 72 individuals screened for problem alcohol use and 256 unscreened samples. The majority of probands were ascertained through alcoholism treatment programs and have severe clinical histories. Usable genotypes were obtained for 323 cases and 328 controls. Individuals with alcohol dependence were significantly more likely than controls to have a short allele (75.0% vs 66.8%, w12 ¼ 4.83, P < .03). There was no difference in between the SS and LS genotypes. Follow-up analyses within the alcohol dependent group to test for associations between httlpr and comorbid disorders suggest no significant associations between allele variation and DSM-IV conduct disorder, antisocial personality, major depression, or nicotine dependence, but there was a trend for individuals with the SS genotype to be less likely to have substance dependence (on a substance other than alcohol or nicotine, w12 ¼ 3.08, P < .08). These results replicate prior studies of an association between httlpr and alcohol dependence, but are not consistent with some studies suggesting that the association exists only among individuals with both alcoholism and antisocial behavior. The absence of association with antisocial behavior and major depression suggests httlpr may be directly involved in the etiology of alcoholism rather than acting indirectly through comorbid disorders. This work was supported by NIH grant AA-11408.

P11.20 ASSOCIATION BETWEEN POLYMORPHISMS OF SEROTONIN 1B RECEPTOR GENE, PURE ALCOHOLISM AND ALCOHOLISM COMORBID WITH ANXIETY/DEPRESSIVE DISORDER IN HAN CHINESE MEN Lin W-W,1 Huang S-Y,1 Lu R-B,2 and Ko H-C3 1 Dept of Psychiatry, Tri-Service General Hospital, National Defense Medical Center, Taipei, Taiwan 2 Dept of Psychiatry, National Cheng Kung University, Medical College, Tainan, Taiwan 3 Institute of Behavioral Medicine, National Cheng Kung University, Medical College, Tainan, Taiwan From previous studies concerning humans and animals, the 5-HT1B receptor (HTR1B) gene has been proposed as a candidate gene in alcohol dependence. But it appears that the inconsistencies do exist

Abstracts regarding the possible implication of the HTR1B gene in alcohol dependence, which might be due to mixtures of different clinical subtypes of alcohol dependence and the problem of normal control. We therefore investigated the association between polymorphisms of HTR1B gene, different clinical subtypes of alcoholism, i.e. pure alcohol dependence (Pure ALC) and alcohol dependence comorbid with anxiety/depressive disorder (Anx/Dep ALC), and ‘‘super-normal’’ controls (NC), i.e., exclusion of major or minor mental illness. We recruited a total of 506 Han Chinese male subjects, i.e., 108 Pure ALC, 105 Anx/Dep ALC, 113 anxiety-depressive disorder without alcohol use disorder (Anx/Dep) and 180 NC. The subjects of ALC and Anx/Dep were recruited from hospital and NC form community. All subjects were interviewed with Chinese version of modified SADS-L and DSM-IV to obtain their lifetime psychiatric diagnoses. There was only significant difference between Pure ALC and NC in HTR1B G861C genotypes (P ¼ 0.025) and with ALDH2*2 allele (P ¼ 0.018). There was no significant difference in both genotype and allele frequencies among Anx/Dep ALC, Anx/Dep and NC. In conclusion, our results suggest that HTR1B gene might be associated with pure alcohol dependence, especially with inactive ALDH2*2, and also implicate that the ‘‘supernormal’’ controls and the specific subtypes of alcohol dependence might be the important factors to determine whether the HTR1B gene is associated with alcohol dependence.

P12.1 FAMILIAL SUICIDE BEHAVIOR: ASSOCIATION WITH PROBANDS SUICIDE ATTEMPT CHARACTERISTICS AND 5-HTTPRL POLYMORPHISM Correa H, Romano-Silva MA, De Marco LA, Boson W, Campi-Azevedo AC, Machado M, and Marques D Universidade Federal de Minas Gerais, Belo Horizonte, Brazil There is compelling evidence that a serotonergic dysfunction may play a major role in suicide behavior and it has also been demonstrated that suicide is, at least partially, genetically determined. Thus, the serotonin-related genes are major candidates. Previously, a functional polymorphism in the promoter region of the serotonin transporter gene (5-HTTLPR) was identified and the presence of the short allele (S) was found to be associated with a lower level of expression of the gene and lower levels of 5-HT uptake when compared to the long allele (L). The purpose of this study was to evaluate the association between family history of suicide behavior and probands’ suicide attempt history, suicide attempt characteristics and 5-HTTLPR genotype. We genotyped 237 probands (major depressed or schizophrenic patients) and used a semi-structured interview to determine probands’ suicide attempt characteristics and first and second degree suicidal behavior. An association between suicidal family history and proband’s suicide attempt, but not with suicide attempt characteristics and probands genotype, was found. Our results suggest that multiple biological and environmental factors underlie familial transmission of suicidal behavior.

P12.2 LACK OF ASSOCIATION BETWEEN THE TPH A218C POLYMORPHISM AND SUICIDAL BEHAVIOR IN BRAZIL Correa H, Machado M, De Marco LA, Boson W, and Romano-Silva MA Uninersidade Federal de Minas Gerais, Belo Horizonte, Brazil Central serotonergic dysfunction and genetic factors are associated with suicidal behavior in psychiatric patients. The gene that codes for tryptophan hydroxylase (TPH), the rate-limiting enzyme in the biosynthesis of serotonin, is a major candidate. The goal of this study was to examine the association between a polymorphism in the intron 7 of this gene (A218C) and suicidal behavior in a Brazilian sample of psychiatric patients. We genotyped 279 subjects. They were schizophrenic (n ¼ 99), major depressed (n ¼ 80), or alcoholic (n ¼ 40), which diagnosis was conducted using a structured instrument (MINI-PLUS) according DSM-IV criteria, and 60 healthy controls. Patients were assessed as regard of their suicide history by means of a semistructured instrument. The genotypes (AA, AC CC) were not statistically different across individuals following diagnosis (x2 ¼ 1.7, d.f ¼ 6, P ¼ 0.94).. The genotype was also not associated with the suicide attempt history (x2 ¼ 1.59, d.f ¼ 2 P ¼ 0.45) nor with suicide attempt characteristics in suicide attempter patients. The authors conclude that the A218C polymorphism of the TPH gene was not a susceptibility factor for suicidal behavior in this group of psychiatric patients.

113

P12.3 THE ROLE OF THE DOPAMINE D4 RECEPTOR POLYMORPHISMS IN PERSISTENCE AS AN ENDOPHENOTYPE OF HUMAN TEMPERAMENT Szekely A,1 Tarnok Z,2 Kiraly O,3 Ronai Z,3 and Sasvari-Szekely M3 1 Institute of Psychology, Eotvos Lorand University, Budapest, Hungary 2 Foundation for Mentally Ill Children, Vadaskert Institute of Child and Adolescent Psychiatry, Budapest, Hungary 3 Department of Medical Chemistry, Molecular Biology and Pathobiochemistry, Semmelweis University, Budapest, Hungary Genetic association studies linked measures from self-report questionnaires to candidate genes to reveal inheritable components of human personality traits. ‘Novelty Seeking’ was the first temperament scale to be associated with the exon III repeat polymorphism of the dopamine D4 receptor gene (DRD4 VNTR) by Ebstein and Benjamin in 1996. Replication studies in various ethnic groups revealed contradictory results, and meta-analytic reviews of these studies showed a lack of consistent association between the DRD4 VNTR and ‘Novelty Seeking’. Further association studies focused on the upstream regulatory region of the DRD4 gene. Among these the 521 C/T single nucleotide polymorphism (SNP) of the promoter region was shown to be associated with ‘Novelty Seeking’ scores of the Temperament and Character Inventory (TCI) in a Japanese male population (Okuyama, et al., 2000). These results were replicated in a Hungarian population (Ronai et al., 2001), however, investigation of male and female subgroups revealed a more pronounced association for females, and showed no genotypic effect for males. Results from a Hungarian study did not find association between ‘Novelty Seeking’ and the DRD4 VNTR, but male individuals with a 7-repeat allele exhibited significantly lower ‘Persistence’ scores (Szekely et al., 2004). This association can be theoretically linked to the 7-repeat allele as a risk factor for attention deficit hyperactivity disorder (ADHD), since family and twin studies demonstrated a significant contribution of genetic factors in the etiology of ADHD and Tourette syndrome (TS). Here we present an association study of a Caucasian clinical sample (N ¼ 90) of unrelated children with ADHD and TS. We tested the DRD4 VNTR and the 521 C/T polymorphisms in relation to the child temperament measures obtained by the Junior Temperament and Character Inventory (JTCI). Similarly to previous results of the adult sample, significant association (P ¼ 0.03) of the seven-repeat allele and the ‘Persistence’ scores was found. Moreover, an interaction effect of the two DRD4 polymorphisms was detected (P ¼ 0.01). The clinical subgroup with a 7-repeat allele and the 521 CC genotype showed lower mean ‘Persistence’ scores than any other combination of these DRD4 genotypes. Further evidence from a recent genetic association study of adult behaviour in a sustained attention task also supports the role of the DRD4 VNTR 7-repeat allele in decreasing persistence. These results are in accordance with our previous findings from studies on a non-clinical adult population, and suggest that the human personality trait of ‘Persistence’ may be considered as an endophenotype in a child psychiatric sample.

P12.4 DOPAMINERGIC POLYMORPHISMS ASSOCIATED WITH A SELF-REPORT MEASURE OF HUMAN ALTRUISM Ebstein RP,1 Bachner-Melman R,1 Gritsenko I,2 Dina C,3 and Zohar A4 1 Hebrew Univeristy, Jerusalem, Israel 2 Herzog Hospital, Jerusalem, Israel 3 Institut de Biolgie de Lille, Lille, Afghanistan 4 Rupin Academic Center, Emek Hefer, Israel Although human altruism has confounded evolutionary biologists since the days of Darwin, recent research has provided a theoretical basis for the maintenance of this trait despite its liability in reducing individual fitness. However, specific genes associated with human altruism have yet to be identified. We here show that three genes important for dopaminergic neurotransmission (dopamine D4 & D5 receptors and insulin like growth factor 2) are associated in a familybased study with a self-report selflessness scale or altruism. We examined 354 non-clinical families with multiple siblings inventoried for scores on the Selflessness Scale. This questionnaire measures the propensity to ignore ones own needs and serve the needs of others, or in other words altruism. We examined two dopaminergic genes (DRD4 & DRD5) in these subjects that we hypothesized might contribute to prosocial or altruistic traits based on the role a single variant of these genes plays in a childhood behavioral disorder, attention deficit hyperactivity disorder (ADHD). This disorder is often comorbid with anti-social behavior such as conduct disorder and oppositional defiant disorder.

114

Abstracts

We reasoned that if one variant contributes to anti-social traits then conversely the absence of this variant or the presence of other variants might contribute to altruistic behavior. Additionally, we examined three DRD4 promoter region SNPs. We also genotyped 3 SNPs in the insulin like growth factor 2 gene (IGF2), an imprinted gene on chromosome 11p15.5. IGF2 that was an attractive candidate due to its proximity to DRD4 as well as studies that connect this class of growth factors to survival of dopamine neurons. We used FBAT to test several traits simultaneously using a multivariate version of the statistic. Significant association was observed between scale scores (KIN, NON-KIN and NON-CARING) and the DRD4 exon III repeat (chi-square ¼ 16.99, P ¼ 0.002, 4 df) in the multiallelic mode. Employing the biallelic mode, that compares each allele individually against all others collapsed into a single category to determine which specific alleles determine association, significant association was observed only with the most common D4.4 allele (chi-square ¼ 14.06, P ¼ 0.0008, 2 df) and not with the second most common allele in this population, the D4.7 repeat, the risk allele in ADHD. Evidence was also observed showing a negative correlation between the DRD5 148 bp microsatellite repeat, the risk allele in ADHD, and scores on the Selfless Scale. Significant association was also observed the IGF2 ApaI ‘‘G’’ allele and all three measures of altruism, KIN (P ¼ 0.002), NON-KIN (P ¼ 0.001) and NON-CARING (P ¼ 0.002). The demonstration that common polymorphisms pertinent to dopamine pathways are associated with altruistic behavior in man is consistent with a central role of the brain reward center in mediating this phenotype. P12.5 LINKAGE BETWEEN VASOPRESSIN RECEPTOR AVPR1A PROMOTER REGION MICROSATELLITES AND MEASURES OF SOCIAL BEHAVIOR IN HUMANS Ebstein RP, Bachner-Melman R, Elizur Y, and Zohar A 1 Hebrew University, Jerusalem, Israel 2 Rupin Academic Center, Emek Hefer, Israel Although progress has been made in unraveling the molecular genetic architecture of individual personality traits, little is known regarding the genetic basis of social behaviors (measures of the interaction between at least two individuals) in humans. Much more has been elucidated on the genetic control of social behavior in animals, including insects and lower vertebrates. In particular, research over the last two decades has revealed the molecular mechanisms by which two peptide hormones, vasopressin and oxytocin, shape social behavior from fish to rodents. However, despite speculation, little evidence has been forthcoming linking these hormones to corresponding human behaviors. In the present report we examined two markers near the vasopressin receptor (AVPR1A) gene and tested for linkage to two complex social behaviors in humans, sibling relationships and selfpresentation style. Self-report questionnaires were administered to 552 same-sex siblings from 247 families. Suggestive linkage was observed between both microsatellites (RS1 and RS3) and the Sibling Relationship Questionnaire ‘Conflict’ scale (RS1 245 families: chisquare ¼ 12.64, LOD ¼ 2.74, P < 0.001, RS3: chi-square ¼ 13.43, LOD ¼ 2.92, P < 0.001) and the Concern for Appropriateness Scale ’self-presentational style’ (RS1 245 families: chi-square ¼ 8.76, LOD ¼ 1.90 P < 0.01, RS3: chi-square ¼ 9.28, LOD ¼ 2.02, P < 0.01). The current results provide the first provisional evidence that the vasopressin receptor mediates social behavior in humans and links a specific genetic element to perceived sibling interactions. Although social behavior and the acquisition of social skills in humans constitute the cornerstone of society and culture, few if any genetic studies have attempted to relate individual differences in social skills to specific genes. Moreover, social skills are relevant to a variety of psychiatric disorders including autism, schizophrenia and externalizing behavior problems in children. From an evolutionary perspective it is not surprising that a hormone that almost universally affects a spectrum of social and affiliative behaviors in lower animals has a parallel role in man, a social animal highly dependent on group interactions for individual and species survival. P12.6 FMRI STUDY OF VISUOSPATIAL WORKING MEMORY IN CHILDREN AND ADOLESCENTS WITH VELO-CARDIO-FACIAL SYNDROME Azuma R,1 Campbell L,1 Daly E,1 Murphy D,1 and Murphy K2 1 Institute of Psychiatry, KCL, London, United Kingdom 2 Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin, Ireland

Velo-cardio-facial syndrome (VCFS) is a genetic disorder that is associated with a microdeletion in the Chromosome 22q11. Children with VCFS typically have mild learning disabilities and show deficiencies in specific cognitive areas such as visuospatial, working memory, mathematics abilities and executive functions. Recent imaging studies showed that the VCFS patients have structural abnormalities in various brain regions, including parietal and frontal lobes that are known to be involved in visuospatial and mathematical processing. In this study, we investigated visuospatial working memory processing in VCFS, using functional Magnetic Resonance Imaging (fMRI), to investigate how cognitive deficits are related to the structural brain abnormalities. Twelve children and adolescents with VCFS and twelve control children (aged 8–17 for both groups) were scanned while performing a visuospatial working memory task. Both groups showed task related activation in dorsolateral prefrontal cortex (DLPFC) and bilateral parietal association cortices that are associated with working memory. However, the normal group showed gigantically higher levels of activation in these areas. In addition, the normal group showed activation in insula, putamen, premotor cortex and SMA that are related to decision making processes. Previous studies suggested structural abnormalities in these areas in the VCFS population. The results from the current study suggest that genetically determined functional abnormalities in DLPFC may underlie cognitive disorder in VCFS. P12.7 GENETIC VARIATIONS IN THE DOPAMINERGIC SYSTEM AFFECT PREFRONTAL CORTEX ACTIVITY DURING MEMORY FORMATION IN HUMANS Schott BH,1 Seidenbecher CI,2 Sellner DB,1 Lauer CJ,1 Guderian S,1 Frey JU,2 Gundelfinger ED,2 Heinze HJ,1 and Duezel E1 1 Department of Neurology II, Otto von Guericke University, Magdeburg, Germany 2 Leibniz Institute for Neurobiology, Magdeburg, Germany We have recently shown that human midbrain regions, including mesolimbic dopaminergic structures, show higher hemodynamic responses (functional magnetic resonance imaging, fMRI) to novel stimuli that are successfully encoded into episodic memory (later recalled) when compared to novel stimuli that are later forgotten. We now investigated the effects of common polymorphisms in the dopamine D2 and D3 receptors (DRD2, DRD3) on this brain activity pattern. 44 young, healthy subjects studied visually presented words at two levels of processing (deep vs. shallow) and were tested for free recall performance after a distractor task. Irrespective of genotype, subsequently recalled words showed an increased activation in bilateral, predominantly left, prefrontal regions, in the medial temporal lobes bilaterally, in the basal ganglia and in the midbrain. The DRD3 Ser9Gly polymorphism has previously been associated with altered dopamine binding capacity. Subjects homozygous for the (high affinity) Ser allele (22) showed a higher bilateral activation of a medial prefrontal region (Brodmann area, BA, 6), bilateral striatum and in the left nucleus accumbens as compared to DRD3 Ser9Gly heterozygous subjects (17) (Gly/Gly homozygous subjects were excluded from analysis due to their small number). The DRD2 TaqIA polymorphism has previously been associated with decreased D2 receptor binding in subjects carrying the A1 allele. In our study DRD2 A2 homozygous subjects (25) showed a higher activation of bilateral, predominantly right, prefrontal regions including BA 6, as compared to subjects carrying a copy of the A1 allele (18 A1/A2 heterozygous subjects) (a single A1/A1 homozygous subject was excluded from analysis). Neither of the two polymorphisms affected memory performance or reaction times. Our results suggest that genetic variations in dopamine D2/D3 receptor subtypes do neither affect episodic memory performance nor the differential activation of midbrain dopaminergic brain regions during successful memory formation. Activation of medial prefrontal and subcortical target regions, on the other hand, are affected, and these are possibly related to the working memory component of episodic encoding.

P12.8 CANDIDATE GENE ANALYSIS USING DNA POOLING TO SCREEN FOR ASSOCIATION WITH ATTENTION DEFICIT HYPERACTIVITY DISORDER (ADHD)—ANALYSIS OF STX1A, VAMP2, SYT1, VMAT2 AND SYP Brookes KJ, Knight J, Xu X, Plomin R, and Asherson P Institute of Psychiatry, London, United Kingdom

Abstracts ADHD is one of the most prevalent, and heritable behavioural disorders in childhood, often persisting into adult life. Previous genetic associations have been reported with genes in the dopaminergic system, namely the dopamine Transporter (DAT1) and the dopamine receptors (DRD4 and DRD5). Recently SNAP-25 has also shown association with ADHD in several datasets. With this in mind, genes that produced proteins that interact with SNAP-25 in the procedure of releasing of neurotransmitters at the synapse were investigated. Sixty-one single nucleotide polymorphisms (SNP’s) with minor allele frequencies >5%, were screened in DNA pools made up from an ADHD clinical sample of DSM-IV combined type probands (n ¼ 180), a control sample 90 males and 90 females who fell within the bottom 20% of hyperactivity scores at ages 2, 3 and 4 years and high score pools of 90 males and 90 females who fell within the top 20% of the distribution. High and low score samples were selected from unrelated individual within the Twins Early Development Study (TEDS). On initial screening we looked for allele frequency differences of 5% or more. Those that were found to show allele frequency differences of more than 5% between the case and controls were analysed for significance using a meta-regression method prior to selection for individual genotyping. We found no significant association of DSM-IV ADHD combined type with. One significant association was found in SYP, which retained significance when individually genotyped and analysed under TDT. No significant findings were found in the high score versus low score comparisons. P12.9 MOLECULAR GENETIC DETERMINANTS OF PSYCHOLOGICAL MORBIDITY IN CANCER PATIENTS Edwards LJR,1 Sanders R,2 and Selby P1 1 Cancer Research UK, Leeds, United Kingdom 2 University of Leeds, Leeds, United Kingdom Complex disorders such as major depression have been demonstrated to have a genetic basis. Genetic research in psychiatric populations has made tentative links between candidate genes and depression and anxiety. Candidate genes have been selected on the basis of their role in pathways implicated in the mediation of depression and anxiety, for example through pathways involved in antidepressant action such as the serotonin system. This study aims to investigate the relationship between candidate genes and self-reports of psychological morbidity in the form of an association study in a population of cancer patients. Each of the patients share the same adverse life event of a cancer diagnosis suggesting that differences between patients who go onto develop psychological morbidity and those that do not, may in part be explained by genetics. A cross section of 500 cancer patients completed depression and anxiety inventories (HADS, BDI, STAI and MHI-5). Venous blood samples were collected. DNA was extracted from blood using the Nucleon genomic DNA extraction protocol. PCR was conducted on the extracted DNA and genotyping analysis carried out. Results will be reported with particular attention to the clinical utility of the results with a cancer centre.

P12.10 ASSOCIATION STUDY OF SEROTONIN TRANSPORTER GENE POLYMORPHISMS IN ATTENTION DEFICIT HYPERACTIVITY DISORDER Xu X, Mill J, Chen C-K, Brookes K, Knight J, and Asherson P Institute of Psychiatry, London, United Kingdom Attention deficit hyperactivity disorder (ADHD) is a common highly heritable neuropsychiatric disorder characterized by hyperactivity, inattention, and impulsivity. Current estimates indicate that 3–6% of school age children are diagnosed with ADHD. ADHD is a complex, multifactorial disorder and the precise aetiology is not yet known. Identification of polymorphic variation within serotonergic system genes has been associated with clinical cases of ADHD. Three common polymorphisms of serotonin transporter (5-HTT) gene have been reported to be associated with ADHD: the long allele of a 44 base pair insertion/deletion polymorphism (5-HTTLPR), variable number tandem repeat (VNTR) polymorphism and the T allele of the 30 UTR single nucleotide polymorphism. In this study we further investigated the relationship between polymorphisms in the 5-HTT gene and ADHD in UK and Taiwanese samples. The three polymorphisms in the 5-HTT gene were genotyped in 197 UK children and 212 Taiwanese children who met diagnostic criteria for ADHD and their parents and siblings. No association was found between the 5-HTTLPR or VNTR or 30 UTR polymorphisms in either the UK or Taiwanese ADHD samples: ETDT

115

P ¼ 0.521, P ¼ 0.497 and P ¼ 0.524 in UK ADHD samples, respectively, P ¼ 0.497, P ¼ 0.440 and P ¼ 0.345 in Taiwanese ADHD samples, respectively). There was no significant finding by pooling UK and Taiwanese samples using ETDT, HHRR and TRANSMIT analysis. However, ETDT analysis of haplotypes showed significant preferential transmission of a haplotype containing the long allele of promoter polymorphism-the 9 repeat VNTR-the T allele of the 30 UTR SNP (w2 ¼ 4.5, P ¼ 0.034). Our results show there is no evidence of association of three 5-HTT gene polymorphisms and ADHD in either UK and Taiwanese samples, but there is a trend of significant association between the long allele of the promoter, the 9 repeat VNTR and the T allele of the 30 UTR SNP haplotype transmission. This study suggests that there may be the possible involvement of 5-HTT in susceptibility to ADHD, although further work needs to be done before any firm conclusions can be made. P12.11 DNA POOLING IDENTIFIES ASSOCIATIONS BETWEEN SNPS IN NET1, 5HT1B AND ATTENTION DEFICIT HYPERACTIVITY DISORDER Xu X, Knight J, Brookes K, Mill J, and Asherson P Institute of Psychiatry, London, United Kingdom Attention deficit hyperactivity disorder (ADHD) is one of the most common childhood behavioural disorders. Genetic factors are known to make an important contribution to the underlying liability to develop ADHD. Identification of polymorphic variation within dopaminergic and serotonergic system genes has been associated with ADHD. To date replicated associations have been described with several genes that regulate monoamine neurotransmission (DRD4, DRD5, DAT1, SNAP25, 5HT1B). This study set out to identify association between ADHD and common polymorphisms of candidate genes. Several candidate genes were selected in our study (DAT1, DRD1, DRD4, 5HT1B, DBH, NET1, SERT, and ADRA). We screened for association with multiple single nucleotide polymorphisms (SNPs) using a DNA pooling approach in DSM-IV ADHD combined subtype samples and controls. We identified available SNPs from publicly available databases that have minor allele frequencies 10% and span the genomic region of the 8 genes (DAT1: n ¼ 5, DRD1: n ¼ 7, DRD4: n ¼ 14, 5HT1B: n ¼ 7, DBH: n ¼ 18, NET1: n ¼ 21, SERT: n ¼ 11 and ADRA: n ¼ 5). Heteroyzgosity rates were validated in a test pool of 40 control DNAs. DNA pooling was used to screen for association analysis using SNaPshot method to estimate allele frequencies of SNPs. Pooled DNA samples consisted of 2 case pools n1 ¼ 90, n2 ¼ 90) and 4 control pools (n1 ¼ 90, n2 ¼ 88, n3 ¼ 77 and n4 ¼ 79). We identified 3 SNPs in NET1 that showed significant allele frequency differences between the case and control pools and confirmed the findings using conventional TDT analysis in 180 cases. We also found allele frequency differences in 2 SNPs of serotonin receptor 1B between the case and control pools, confirmed by individual genotyping. For both genes significant linkage disequilibrium was observed across the regions of the associated markers. The preliminary data provides some evidence for association between ADHD and SNPs within NET1 and 5HT1B although the nominal levels of significance are not robust to adjustment for multiple testing.

P12.12 SNP-, HAPLOTYPE AND LD-ANALYSIS OF THE NOVEL NEURONAL TRYPTOPHAN HYDROXYLASE GENE (TPH2) IN SUICIDE VICTIMS Zill P,1 Bu¨ttner A,2 Eisenmenger W,2 Mo¨ller HJ,3 Bondy B,1 and Ackenheil M1 1 Department of Psychiatry of the LMU Munich, Section for Psychiatric Genetics and Neurochemistry, Munich, Germany 2 Institute for legal medicine of the LMU Munich, Munich, Germany 3 Department of Psychiatry of the LMU Munich, Munich, Germany The involvement of genetic risk factors in suicidal behaviour is supported by family, twin and adoption studies. Several studies have suggested that the genetic factors are independent of the genetic transmission of other psychiatric disorders which show co morbidity in suicidal behaviour but specific genes and relevant DNA sequence variations involved in the pathogenesis of suicidal behaviour have not yet been identified. There is great evidence that disturbances of the central serotonin (5-HT) system are involved in the neurobiology of suicide, but numerous genetic studies with genes of the serotonergic system yielded non convincingly confirmed results. Tryptophan hydroxylase (TPH), as the rate limiting enzyme in the synthesis of

116

Abstracts

the neurotransmitter serotonin is widely discussed as a major candidate gene for suicidal behavior. Recently a new TPH isoform which is designated as TPH2 or neuronal TPH was identified. In our own group we were recently able to demonstrate that in human brain TPH2 is predominantly expressed in the brain stem, the major locus of the serotonin producing neurons. The aim of the present study was to investigate whether there is an association between TPH2 gene polymorphisms and completed suicide. We performed association and linkage disequilibrium studies on 263 individuals who committed suicide and 265 healthy controls with 10 SNPs in the TPH2 gene spanning approximately 28 kb between exon 5 and exon 7, a region which was shown to be associated also with major depression by our group. To enhance the detection power of association, we compared also estimated haplotypes frequencies deduced from genotype frequencies of the 10 SNPs. Significant association was detected between one SNP (P ¼ 0.004, global P ¼ 0.01) and suicide. Additionally haplotype analysis produced also support for association (P < 0.0001, global P ¼ 0.0001). Our finding provide evidence for an involvement of genetic variants of the new TPH2 gene in the pathogenesis of suicidal behavior and might be a hint on the repeatedly discussed duality of the serotonergic system. The potential functional consequences of the investigated SNPs, as well as possible interactions between the TPH2 gene and other plausible candidate genes in determining the susceptibility to suicide require further studies in order to understand the genetic factors underlying suicidality. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine.’’ P12.13 TEST-RETEST RELIABILITY OF ERP COMPONENTS OF P300, P50, AND DURATION MISMATCH NEGATIVITY IN MONOZYGOTIC TWINS Hall M,1 Schulze K,1Rijsdijk F,1 Picchioni M,1 Ettinger U,1 Bramon E,1 Freedman R,2 Murray R,1 and Sham P1 1 Institute of Psychiatry, London, United Kingdom 2 University of Colorado Health Sciences Centre, Denver, United States Event-related potentials (ERPs) have been widely studied as a possible endophenotype of schizophrenia. We examined the test-retest reliability of three ERP components (P3, P50 and MMN) and estimated their heritabilities and measurement errors using model fitting analyses. Methods: ERPs were recorded with a 19-channel electroencephalograph (EEG) from 26 monozygotic twin pairs, 15 of them on two occasions with an average interval of 17.8 days. P3 was assessed using an auditory oddball paradigm. P50 was recorded with a conditioningtesting paradigm and MMN was elicited by a duration auditory oddball task. Zygosity was determined using DNA genotype. Results: significant high reliabilities were found for the P3 amplitude and P3 latency (ICC ¼ 0.88 and 0.9 respectively) as well as high twin correlations (ICC ¼ 0.65 and 0.87 respectively). Reliability for the MMN amplitude was also high (ICC ¼ 0.64) with a moderate twin correlation (ICC ¼ 0.47). MMN latency however, was modest (ICC ¼ .33) with a low twin correlation (ICC ¼ 0.26). Reliability of P50’s T/C ratio was modest (ICC ¼ 0.34) as well as the twin correlation (ICC ¼ 0.35). Model fitting analyses, assuming the absence of shared environment, indicated a significant presence of genetic variance components (a2) for all ERP measures. We demonstrated generally good reliabilities of the ERP components. The high reliability and heritability of ERP components supports their use as endophenotype for schizophrenia research.

P12.14 GENOTYPE ANALYSES OF THE ESTROGEN RECEPTOR ALPHA AND BETA GENES IN GENDER IDENTITY DISORDER Otani K,1 Ujike H,1 Sato T,1 Nakatsuka M,2 Tanaka Y,1 Morita Y,1 Kishimoto M,1 Nomura A,1 Ota J,1 Okabe N,1 and Kuroda S1 1 Department of Neuropsychiatry, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan 2 Department of Gynaecology, Okayama University Graduate School of Medicine and Dentistry, Okayama, Japan Gender Identity Disorder is a biological psychiatric disorder, which is defined by presence of wish to be accepted for an opposite sex and feeling of uncomfortable with assigned sex. Several studies have reported this disorder may result from aberration in the early sexual differentiation of the brain structure by sexual hormones. Estrogens, most important sexual hormone, are well known to affect early brain development.

There are two subtypes of the receptor, estrogen receptor(ER) alpha and beta, for estrogens in the brain. The ER alpha gene is located on chromosome 6q25.1, containing (TA)n dinucleotide polymorphism in the promoter region. The ER beta gene is located on chromosome 14q2324, containing a CA dinucleotide repeat in 30 flanking region. We examined genetic association between these two ER genes and gender identity disorder by case-control study. There were no significant differences between the ER genes and gender identity disorder, even when divided into 2 subgroups, male to female and female to male. The results suggest that ER gene may not confer major susceptibility to gender identity disorder. P12.15 MOLECULAR DATING OF THE XQ21/YP11 HOMOLOGIES CONTAINING A CANDIDATE GENE PAIR FOR PSYCHOSIS TO 4MYA Close JP and Crow TJ Department of Psychiatry, University of Oxford, Oxford, United Kingdom Protocadherin X/Y (PCDH11XY) has been highlighted as a functional candidate for the observed relationships between cerebral asymmetry, psychosis and sex. Such observations include sex-dependant anomalies of cerebral asymmetry in schizophrenia and relative hemispheric deficits associated with sex chromosome aneuploidies. The brainspecific expression and putative biological function of PCDH11XY is consistent with a role in the neurodevelopmental establishment of cerebral structure. The gene pair is located in a homologous section of the X-Y chromosomes thought to have been generated by an X to Y translocation occurring 2–3 million years ago, followed by a paracentric inversion which split the homologous Y region. It has been argued that these events relate to the evolution of language and the speciation of Homo sapiens (Brain Res. Rev. 31:118–29) and subsequently the ubiquity of psychosis across all human populations.To investigate the role of PCDH11XY in human speciation, we have clarified the order and timing of the events leading to the establishment of the gene pair. This was achieved through comparison of publicly available X/Y and chimp sequences, revealing that the date of the X-Y translocation was earlier than previously thought (approximately 4 million years ago). Furthermore, evidence is presented that the two X-Y homologous regions appear to have been generated by two paracentric inversions (rather than one), with the final inversion being polymorphic within modern human populations. We also suggest a somewhat tentative date for the first paracentric inversion using a novel method, which - to the authors’ knowledge—is the first attempt at the molecular dating of a nonduplicative event. This was achieved by using the gene conversion properties of inverted repeats on the Y-chromosome, with sequence convergence (rather than divergence) used to date a chromosomal rearrangement. The date of the initial reduplicative transposition has implications for the role of the translocation in hominid speciation and suggests that this primary event was not itself directly involved in the evolution of language, although it may relate to the Australopithecus/ Homo boundary. The duplicated sequence would therefore be a hominid specific characteristic, providing the landscape for later positive selection on the Protocadherin XY genes. P12.16 A NOVEL METHOD FOR DISENTANGLING THE EFFECTS OF GENES, INTRAUTERINE AND ENVIRONMENTAL RISK FACTORS Rice F,1 Goody A,2 Boivin J,2 Hay DF,2 Harold GT,2 and Thapar A1 1 University of Wales College of Medicine, Cardiff, United Kingdom 2 Cardiff University, Cardiff, United Kingdom Genes play an important role in influencing health, behaviour and disease. However, even after specific genetic variants are identified it will be necessary to understand how genes interact and co-act with environmental factors. In addition, environmental factors contribute towards most common disorders and traits and there is increasing evidence that some of these may originate in intrauterine and very early life. Twin and adoption studies are used to disentangle the effects of genes and environment. However, using these methods, it is difficult to distinguish between the effects of genes from intrauterine and very early environmental influences. For example, children who have been adopted not only share genes with their biological parents but will also have spent nine months in the intrauterine environment provided by the biological mother. Moreover, monozygotic twin pairs may well share environment to a greater extent that dyzygotic twin pairs,

Abstracts particularly in-utero and in early life. There is consistent evidence that many important environmental risk factors (e.g. smoking in pregnancy, family adversity) are not independent of genetic predisposition. Strategies are needed to look at the association of environmental risk factors independent of genetic influences. If an association between risk factor and outcome is ‘truly’ environmental mediated, the association will remain the same regardless of whether or not the parent and offspring are biologically related. We propose a novel research strategy to address these limitations, examining 5 groups of children conceived by in vitro fertilisation (IVF). 1) IVF with parent’s own gametes (genetically related to mother and father), 2) IVF with sperm donation (related to mother only), 3) IVF with egg donation (related to father only), 4) IVF with embryo donation (related to neither parent but intrauterine environment provided by social mother), 5) IVF with host surrogacy (related to both parents but intrauterine environment provided by an unrelated surrogate). To assess the feasibility of this method and to examine the contribution of intrauterine factors independent of genetic risk. Data collection is via 15 IVF clinics across the U.K. 250 families have participated by completing postal questionnaires. Mothers and fathers reported on their own health and behaviour as well as their children’s. There were no significant differences between the groups in terms of antenatal risk factors. The contribution of genes, intra-uterine and environmental factors to childhood depressive symptoms will be presented as an example of the feasibility of this method. P12.17 HOW DO GENES EXERT THEIR ROLE? PER3 VARIANTS AND POSSIBLE INFLUENCES ON MOOD DISORDER PHENOTYPES Serretti A, Artioli P, Mandelli L, Lorenzi C, Benedetti F, and Smeraldi E Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy Complex disorders recognize an influence of multiple liability genes, but on the other hand liability genes exert their influence of a wide range of phenotypes and according to the condition of individuals (healthiness, disease). Human PER3 gene (hPER3) is a clock-relevant member of the PER gene family, located on the human chromosome 1 and it is expressed in the hippocampus, piriform cortex, and cerebellum, mPER3 RNA levels exhibit a robust circadian variation in the SCN, similarly to mPER1 and mPER2, but differently from what happens for the former two, light pulses do not affect mPer3 RNA levels during subjective night. Some genetic variants were found to be associated in healthy subjects with the preference for and the concentration of activity in the evening (eveningness) than in the morning (morningness) and to Delayed Sleep Phase Syndrome. According to this rationale we decided to investigate PER3 variants association with various aspects of symptomatology and character in mood disorder patients. We found an association between Per3ex18T/C and age of onset of the disease (P < 0.005). According to item 18A and 18B scores of HAMD-21, related to diurnal variation of the symptomatology during antidepressant treatment, we found that Per3ex18TC*CC subjects showed a worse symptomatology in the evening more frequently than TT homozygous and heterozygous, at baseline and at the second week of treatment (Chi-sq ¼ 10.27, P ¼ 0.03 and Chi-sq ¼ 10.2, P ¼ 0.03, respectively). Moreover, we also found an association between Per3ex18ins/ del polymorphism and ‘‘novelty seeking’’ as measured by the Temperament and Character Inventory (TCI) (P < 0.05), especially for the subscale ‘‘extravagance’’ (P < 0.05). Interestingly, healthy people showing higher ‘‘novelty seeking’’ were previously found to show ‘‘eveningness.’’ Similarly to the HTTLPR one, a single polymorphism may influence a range of aspects: we may hypothesize that subjects with PER3 variants are characterised by eveningness and temperamental novelty seeking, if affected by mood disorders, this features have an impact as earlier onset of illness and more diurnal variation during illness episodes. These results could shed further light on the involvement of genes controlling the circadian rhythms in various aspects of physiological and pathological mechanisms of the brain.

P12.18 THE PROMOTER OF THE VERVET SEROTONIN TRANSPORTER GENE: SEQUENCING AND ASSOCIATION OF A NOVEL POLYMORPHISM WITH ANXIETY-RELATED BEHAVIOURAL DIMENSIONS Fainman J,1 Desbiolles C,2 Ervin FR,3 and Palmour RM1

117

1

McGill University, Department of Human Genetics, Montreal, Canada 2 McGill University, Department of Biology, Montreal, Canada 3 Behavioural Sciences Foundation, Estridge, Saint Kitts and Nevis Serotonin (5-hydroxytryptamine) has long been considered a key modulator of mood, with deficits of central nervous system serotonin implicated in many psychiatric disorders (anxiety, depression, alcoholism, etc.). The serotonin transporter (human slc6a4, 5-HTT) is responsible for the high-affinity reuptake of serotonin from the synapse, effectively terminating the serotonergic signal. This work examines the genetics of the vervet 5-HTT gene and its relationship to the behavioural dimension ‘fearfulness’ which is the broad analogue of human anxiety-related personality dimensions (Desbiolles et al., 2004, submitted). Sequence of the vervet 5-HTT gene promoter was obtained via direct sequencing of a clone obtained from a vervet BAC library (CHORI 252). In all some 2.8 kB of previously unknown promoter sequence was generated including exon 1. Characterization and mutation analysis of the promoter yielded an Alu retrotransposon insertion of 267 bp (2 kb upstream of the transcription initiation site) which was not present in the human sequence, investigation yielded a size variant some 100 bp smaller than the wild-type. Four successive cohorts of 199 monkeys from the Behavioural Science Foundation (St. Kitts, Eastern Caribbean) were genotyped for the polymorphism (named V-SERT-Alu). All monkeys studied had been behaviourally phenotyped in a previous study. Five of the 199 monkeys studied were heterozygous for the variant allele, no rare homozygotes were observed. A positive correlation (P < 0.05) between the rare allele and the behavioural dimension fearfulness emerged upon analysis (ANOVA F1,213 ¼ 3.922). Mean trait values for monkeys with 0 or 1 copy of the variant allele were 0.02 and 0.975 respectively (a difference of nearly 1 standard deviation). The extent to which this polymorphism influences 5-HTT gene expression is under investigation. P12.19 PSYCHIATRIC PROFILE OF CHILDREN WITH VELO-CARDIO-FACIAL SYNDROME (VCFS) Stevens AF,1 Campbell LE,1 Morris R,1 Karmiloff-Smith A,3 Simonoff E,1 Owen MJ,4 Murphy DGM,1 and Murphy KC2 1 Institute of Psychiatry, London, United Kingdom 2 Royal College of Surgeons, Dublin, Ireland 3 Institute of Child Health, London, United Kingdom 4 University of Wales College of Medicine, Cardiff, United Kingdom The genetic disorder, Velo-cardio-facial syndrome, is associated with variably sized deletions in the q11 region of chromosome22. Phenotypic expression is extremely variable and includes congenital heart disease, palatal abnormalities, renal abnormalities, borderline to mild learning disability, behavioural difficulties and mental health problems (e.g. schizophrenia and mood disorders). In this study, we aimed to characterize the mental health, behaviour and cognitive profile (behavioural phenotype) of children and adolescents with VCFS. We recruited 50 children (age 6–16) with VCFS and 30 age matched sibling controls. The participants then underwent a battery of psychiatric and behavioural questionnaires (including the Vineland Adaptive Behaviour Questionnaire, The Child and Adolescent Psychiatric Assessment and The Strengths and Difficulties Questionnaire). We found that emotional symptoms were found in 53% of the VCFS cohort (10% of controls), conduct problems in 40% (20% of controls), attentional/ hyperactivity problems in 46% (20% of controls) and peer problems in 73% (0% of controls). Individuals with VCFS, when compared to their siblings, have more difficulties in areas of socialization, hyperactivity/ attention/concentration, prosocial behaviour, emotional symptoms, peer problems and autistic spectrum disorder traits. We intend to prospectively follow-up these individuals and identify predictors of future mental health problems in later life.

P12.20 GENES, ENVIRONMENTS AND SEXUAL EXPERIENCE IN YOUNG ADULTS: A POPULATION-BASED TWIN STUDY Rose RJ,1 Viken R,1 Mustanski B,1 Kaprio J,2 and Winter T2 1 Indiana University, Bloomington, United States 2 University of Helsinki, Helsinki, Finland Questionnaire data from the fourth-wave of FinnTwin16-25, when twins were ages 23.1 to 25.8, were analysed for the role of genes and

118

Abstracts

environments in 3 dimensions of sexual experience. We distinguished initiation of coitus from number of sexual partners, post-initiation, because their causal factors may differ, for the adolescent experiences of smoking and drinking, initiation is influenced largely by shared environmental factors, while frequency of use, once initiated, is increasingly heritable over time. Age at first sexual intercourse was the third variable included in our analyses. Fourth-wave data were available from 4,548 twins of known zygosity. Earlier studied at ages 16, 17, and 18.5, these twins are from 5 consecutive Finnish birth cohorts (1975–79), ascertained from a population registry, effectively all twin pairs born in the five cohorts, alive and resident in Finland at baseline. Response rates were very high. Sex limitation models were fit to the data, allowing for estimation of two types of sex-limitation effects, sex-specificity of additive genetic (A) or common environmental (C) effects and different magnitudes of these effects across sex. Because initiation data were dichotomous and the number of sexual partners was ordinal, models were fit to polychoric correlations and asymptotic covariance matrices using weighted least squares. Age at first intercourse was continuous, and a model was fit to its covariance matrices using maximum likelihood. Descriptive results: 88.6% of males and 91.8% of females reported initiation of coitus, with a difference in age at first sexual intercourse, a mean for females about 4 months earlier. Correlations for the three measures of sexual experience are about twice as high among MZ twin pairs, suggesting genetic effects, and those for opposite-sex DZ twins approximate those for same-sex DZ pairs, offering no evidence of sex-specificity in genetic or environmental effects. Model Fitting: Number of sexual partners and age at first intercourse were fit by AE models, with no significant sex difference in magnitudes of genes (A) or unshared environments (E). Common environmental effects were necessary in models fit to data on initiation, and about 64% of A effects and 48% of the C effects appear to be sexspecific. Confidence intervals for initiation parameters are large, and these results must be interpreted with caution. Conclusions: Shared environmental experience significantly contributes to individual differences in initiation of coitus by age 25, at that age, most adults are not abstinent and variation in their age at first coitus and the number of later partners is explained largely by genetic influences. Thus, the causes of initiation are partially distinct from those that influence other aspects of sexual experience, so models of independent liabilities, like those for early experience with smoking and drinking, apparently apply to sexual experience, as well.

that increased CGG repeat lengths, which are still within the normal range, are associated with decreased cognitive ability, there is some evidence that longer alleles may be a limiting factor in the cognitive ability of individuals who nevertheless have superior IQ. There was also a tendency for those with longer repeat lengths to be more likely to show high levels of anxiety.

P12.21 INVESTIGATION OF THE RELATIONSHIP BETWEEN FMR1 ALLELE LENGTH AND COGNITIVE/ BEHAVIOURAL PHENOTYPES Loat CS, Plomin R, and Craig IW Institute of Psychiatry, London, United Kingdom

P12.23 REWARD MECHANISMS IN PATHOLOGICAL GAMBLING AND GENES INVOLVED IN DOPAMINE TRANSMISSION Lobo DSS,1 Martins SS,1 Miguita K,1 Knight J,2 Gentil V,1 and Vallada HP1 1 Institute of Psychiatry—university of Sao Paulo, Sao Paulo, Brazil 2 Institute of Psychiatry—Kings College of London, London, United Kingdom

Fragile X syndrome is the most common inherited cause of mental retardation and results from a CGG trinucleotide repeat expansion in the 50 UTR of the FMR1 gene, from a normal size of around 30 repeats to a full mutation of 200þ repeats. Premutations (55–200 repeats) were until recently thought to have no phenotypic consequences. However, there is now evidence that male and female premutation carriers can suffer from late-onset tremor/ataxia, and females have an increased risk of premature ovarian failure. There are also reports of premutations being in excess in samples with special educational needs, and being associated with anxiety and social phobia. We investigated the hypothesis that FMR1 may be a QTL for cognitive ability, and explored the relationship between CGG repeat length and various cognitive/ behavioural measures in 2 case groups of 4 year old children one group with low nonverbal ability (n ¼ 112) and one with low verbal and nonverbal ability (n ¼ 214), a control group representative of the normal distribution of cognitive ability (n ¼ 308), and a separate sample of individuals with high IQ, at least 2 standard deviations above the mean (n ¼ 122). The distribution of allele sizes in the total study population was similar to that reported in other studies, and the two case groups showed no difference from controls in allele distribution and no increased incidence of longer alleles. Only one repeat in the premutation range was found, and this was in a female control. There was no correlation between cognitive ability and repeat length in the cases or controls. However, in the high IQ group, there was a significant negative correlation between repeat length and IQ (r ¼ .229, P ¼ .011), and this correlation became more significant when considering males only (r ¼ .302, P ¼ .006), as would be expected for an Xlinked effect. Therefore, whilst our data do not support the hypothesis

P12.22 X INACTIVATION AS A POTENTIAL SOURCE OF BEHAVIOURAL DIFFERENCES IN MONOZYGOTIC FEMALE TWINS Loat CS, Craig G, Plomin R, and Craig IW Institute of Psychiatry, London, United Kingdom We have previously published quantitative genetic data from a large sample of monozygotic and dizygotic male and female twin pairs (aged 2–4 yrs), which supported the hypothesis that differentially skewed X inactivation renders MZF twins less similar on average (within-pair) than MZM twins on a range of cognitive and behavioural measures. We also found that DZM twins were less similar than DZF twins, which would be consistent with the X inactivation hypothesis. This was proposed to be a result of differences in X-linked gene expression and protein production. We found particular evidence for X-linked QTLs affecting verbal ability, prosocial behaviour and peer problems. Subsequently, we have carried out X inactivation assays on genomic DNA from MZF twins who were particularly dissimilar in their scores on these 3 phenotypes of interest (n ¼ 42 pairs), as well as on DNA from MZF twins who were totally concordant in their scores (n ¼ 36 pairs). Currently, we have found no evidence that the phenotypically discordant twins are significantly less similar in their inactivation profiles than the phenotypically concordant twins. These results do not necessarily disprove our previous hypothesis, since it is possible that some of the phenotypically concordant twins who are differentially skewed are actually homozygous for the hypothetical QTL of interest, and thus confounding our results. It is also known that X inactivation patterns can differ between tissues of the same individual. Brain would be the tissue of preference for analysis when considering cognitive and behavioural traits, but practical difficulties in obtaining brain tissue from individuals for whom we have relevant phenotypic information has prevented such studies.

Pathological gambling (PG) is a psychiatric disorder characterized by an impaired sense of control and also by an inability to delay rewards. Reward mechanisms in the brain have been studied and the role of the dopaminergic systems is been investigated due to the evidence of the involvement of nucleus accumbens. The long form of DRD4 exon III polymorphism (7 repeat allele) and the COMT gene have been previously reported as associated to impulsive behavior, which has been replicated by some investigators. The objective was to investigate the association of functional polymorphisms of a dopamine receptor (DRD4) and COMT and the diagnosis of pathological gambling. 140 sibpairs discordant for pathological gambling diagnosis were investigated and genotyped for DRD4 exon III polymorphism and COMT val/met substitution. All gamblers were interviewed and diagnosed according to DSM-IV criteria by a trained psychiatrist, who also excluded the diagnosis of PG in the siblings who were investigated. Transmission disequilibrium test (TDT) was performed through the program UNPHASED. A t-test was performed for continuous variables and a chi-square test was used for categorical variables. COMT and DRD4 genotypes were grouped as follows: 1. met/met and 7 repeat allele present, 2. met/met and 7 repeat allele absent, 3. met/val and 7 repeat allele present, 4. met/val and 7 repeat allele absent, 5. val/val and 7 repeat allele present, and 6. val/val and 7 repeat allele absent. Sib-pairs were 37 male gambler/male sibling, 33 male gambler/female sibling, 37 female gambler/female sibling, 33 female gambler/male sibling. No significant differences were found regarding age (P ¼ 0.77) and years of

Abstracts education (P ¼ 0.31), although siblings presented better economic status (P ¼ 0.02). No significant differences were found when DRD4 (P ¼ 0.64) and COMT (P ¼ 0.23) were analyzed separately neither when grouped as above (groups: 1. P ¼ 0.45, 2. P ¼ 0.50, 3. P ¼ 0.23, 4. P ¼ 0.37, 5. P ¼ 0.76, 6. P ¼ 0.20). Our results show no evidence for the involvement of these polymorphisms in pathological gambling. It is also important to consider that, due to family problems caused by pathological gambling, it is possible that siblings presented an aversive behavior towards gambling, meaning that they might also have vulnerability factors for this disorder but prevented exposure. Some studies have shown that availability of gambling venues is a factor that contributes significantly to increase the frequency of pathological gambling.

P12.24 PR MISSENSE VARIANTS: A POSSIBLE ASSOCIATION WITH PUERPERAL PSYCHOSIS Sommer S,1 Feng J,1 Yan J,1 Chen J,1 Glanzmann C,1 Craddock N,2 and Jones I2 1 City of Hope National Medical Center, Duarte, United States 2 University of Wales College of Medicine, Cardiff, United Kingdom Steroid and thyroid hormone administration can alter behavior, in occasional individuals, psychosis can result. The levels of progesterone and other hormones change dramatically after birth, suggesting that mutations in the steroid receptor gene family may predispose to puerperal psychosis. To investigate this possibility, the progesterone receptor gene was scanned by DOVAM-S (Detection of Virtually All Mutations-SSCP) in 146 patients with puerperal psychosis. Seven missense variants, three silent changes and two intronic changes were identified. Three missense variants at highly conserved amino acids were found in nine patients with puerperal psychosis while none were found in 100 ethnically matched healthy controls (P ¼ 0.008). A possible association between PR and puerperal psychosis may warrant further study.

P12.25 IMPULSIVITY IN PATHOLOGICAL GAMBLERS: NO ASSOCIATION WITH DRD5 Lobo DSS,1 Wilson D,1 Martins SS,1 Gentil V,1 Kennedy JL,2 and Vallada HP1 1 Institute of Psychiatry—University of Sao Paulo, Sao PAulo, Brazil 2 Centre for Addiction and Mental Health—University fo Toronto, Toronto, Canada Personality studies have consistently demonstrated that impulsiveness is an important feature in pathological gamblers (PG). Other psychiatric disorders present impulsiveness as a core feature, such as attention-deficit hyperactivity disorder (ADHD). Co-morbidity studies have shown that 20% of PG have co-morbid with ADHD. Recent studies suggest a possible association of DRD5 (CAn) and ADHD. Since impulsiveness is a common feature in both pathological gambling and ADHD, we investigated if DRD5 (CA)n is also associated to pathological gambling. 140 discordant sib-pairs for pathological gambling were investigated and genotyped for DRD5 (CA)n repeats. All patients were interviewed and diagnosed according to DSM-IV criteria by a trained psychiatrist, who also excluded the diagnosis of PG in the siblings who were investigated. Impulsiveness was evaluated through the Barrat Impulsiveness Scale (BIS). Transmission Disequilibrium Test (TDT) was performed through the program UNPHASED. A t-test was performed for continuous variables and a chi-square test was used for categorical variables. No significant differences were found regarding age (P ¼ 0.77) and years of education (P ¼ 0.31), although siblings presented better economic status (P ¼ 0.02). Pathological gamblers were significantly more impulsive than their siblings in all BIS subscales (inattention, P < 0.0001, motor impulsiveness, P < 0.0001, non-planning, P < 0.00001). The UNPHASED showed no significant transmission of any DRD5 (CA)n alleles (P ¼ 0.592). Although both PG and ADHD are impulsive disorders, it is possible that DRD5 is associated with another clinical feature of ADHD rather than impulsiveness. Also, impulsiveness is a complex phenotype with many definitions and evaluation instruments. Further investigation is needed to verify if this polymorphism is related to impulsiveness itself, to a specific characteristic of impulsiveness or to another aspect of ADHD.

119

P12.26 ASSOCIATION OF DRD4 AND 5HTT WITH THE EAS TEMPERAMENT SCALES Singh AG, Rowe D, Stever C, and Waldman ID Emory University, Atlanta, United States Behavior genetic studies have consistently yielded genetic influences on both adult personality traits and childhood temperament characteristics. Furthermore, molecular genetic techniques have been used to examine the role of specific genes in the development of such characteristics. The bulk of this research has been in the adult personality literature, with a focus on genes in the dopaminergic and serotonergic neurotransmitter systems. For example, studies have shown associations of the dopamine receptor D4 gene (DRD4) with novelty seeking, and associations of the serotonin transporter gene (5HTT) with Neuroticism. Despite the abundance of behavior genetic research on temperament characteristics in children, there have been few studies of the role of specific genes underlying childhood temperamental traits. This area of research is crucial not only because several childhood characteristics parallel those that have been studied in adults, but also because it is often hypothesized that childhood temperamental characteristics may confer risk for later psychopathology. It is thus important to also understand the role of genes such as DRD4 and 5HTT in the development of childhood temperamental traits. In this study, we investigated the association of DRD4 and 5HTT with parent ratings of the childhood temperament dimensions of Emotionality, Activity, Sociability, and Shyness, the four scales of the EAS Temperament Survey. We tested these associations in a sample of clinic-referred children and their siblings who ranged in age from 6 to 18. We hypothesize that 5HTT and DRD4 will demonstrate differential associations with particular EAS dimensions. Specifically, it is predicted that Emotionality and Shyness, characteristics similar to Neuroticism and related to anxiety and depression, will be associated with 5HTT. In contrast, consistent with its association with ADHD, we predict that DRD4 will be associated with Activity. Exploratory analyses will also investigate associations between Sociability and 5HTT and DRD4. P12.27 GENETIC ASSOCIATION ANALYSES OF GAD2 (GABA PATHWAY) WITH FOOD BEHAVIOUR SCORES AND LATENT OBESITY Dina C,1 Boutin P,1 and Froguel P2 1 Institut Biologie Lille CNRS, Lille, France 2 Imperial College, Hammersmith Hospital, London, United Kingdom During the past two decades, the prevalence of obesity has greatly risen worldwide. These changes have occurred among genetically stable populations, indicating that modifications of behavioural factors like dietary and physical activity must underlie the recent obesity epidemic. However, genetic factors undoubtedly have a great effect on individual predisposition, since 25 to 80 percent of the variation in body-mass index is heritable. Eating behaviour, disordered eating behaviours and food preferences are in part heritable, and a family history of overeating is more frequent in obese individuals with binge-eating disorder than in the general population. The identification of genes causing both syndromic (Prader Willi) and rare monogenic (leptin/melanocortin pathways) forms of obesity, highlight the importance of genetic determinants in obesity. The common forms of obesity are, however, polygenic, and could modestly modulate behaviour component. The recent identification of GAD2 gene variations modulating food intake and increasing the risk for severe obesity highlights the role of the GABA pathway in eating behaviour. A case-control study (575 morbidly obese and 646 control subjects) analyzing GAD2 variants identified an at-risk SNP (243 A > G) for morbid obesity (OR ¼ 1.3, 95% CI [1.053–1.585], P ¼ 0.014). The 243 A > G SNP was associated with higher hunger scores (P ¼ 0.007) and disinhibition scores (P ¼ 0.028), as assessed by the Stunkard Three-Factor Eating Questionnaire. These data support the hypothesis of the orexigenic effect of GABA in humans and of a contribution of genes involved in GABA metabolism in the modulation of food intake and in the development of morbid obesity.

P12.28 D4 DOPAMINE RECEPTOR GENE POLYMORPHISMS AND SOCIAL BEHAVIOR IN ADHD Tarnok Z, Birkas E, Nemoda Z, Gadoros J, Sasvari M, and Gervai J Vadaskert Child Psychiatric Clinic, Budapest, Hungary

120

Abstracts

The association between the 7-repeat allele of D4 dopamine receptor gene (DRD4) 48 bp VNTR polymorphism and ADHD has been confirmed by a recent meta-analysis. There are inconsistent results regarding an association with childhood and adult temperament/ personality. Modifying effect of other polymorphisms in the DRD4 gene—e.g. the 521 C/T promoter SNP—was suggested as one explaining factor. The purpose of our study was to examine the effects of DRD4 polymorphisms on behaviour problems in ADHD-children. Inclusion criteria for the 111 Hungarian children (18 girls, 93 boys) were DSM-IV diagnosis of ADHD. The diagnostic procedure was conducted by two independent psychiatrists. Comorbid diagnosis was measured by standard psychiatric interviews and questionnaires. Mean age of the patient group was 10.11 years (SD: 3.03). Parent reported behaviour problems were assessed by the Child Behaviour Checklist (CBCL). Children were genotyed for DRD4 48 bp VNTR and 521 C/T polymorphisms from non-invasively collected buccal cells. MANOVAs were computed with DRD4 48 bp VNTR and 521 C/T genotypes and gender as between-subjects factors on CBCL scales as dependent variables. The analysis revealed a significant interaction effect for gender and DRD4 48 bp VNTR on withdrawal: girls with the 7-repeat genotype showed elevated withdrawal scores. Modest interaction effect was found between the two polymorphisms on social problems. In the presence of the 7-repeat genotype, chidren with the CC promoter genotype had the lowest social problem score, whereas in the absence of the 7-repeat an opposite pattern was observed, with the TT group scoring lowest. Our results suggest that DRD4 polymorphisms have a modest impact on social adjustment in ADHD-children. Possible explainations of 2-way interactions are discussed. Our further aim is to continue data collection and include DRD4 haplotype in the analysis as well. P12.29 EXTERNALIZING BEHAVIOUR PROBLEMS IN EARLY ADOLESCENTS ARE MODULATED BY DRD4 EXON III AND 5-HTTLPR VARIANTS Nobile M,1 Giorda R,1 Marino C,1 Battaglia M,2 Pastore V,1 Baschirotto C,1 Vanzin L,1 Frigerio A,1 Rucci P,3 Carlett O,4 Martinuzzi A,4 and Molteni M1 1 The Eugenio Medea Scientific Institute, Bosisio Parini, Italy 2 The University Vita-salute San Raffaele & The San raffaele Scientific Institute, Milan, Italy 3 The University of Pisa, Pisa, Italy 4 The Eugenio Medea Scientific Institute, Conegliano, Italy A combined effect of the DRD4 and the 5-HTTLPR polymorphisms was found to be associated with increased avoidant-like behaviour in early childhood (Lakatos et al., 2003; Auerbach et al., 2001; Auerbach et al., 1999; Ebstein et al., 1998). In adults, there is some evidence that Novelty Seeking is modulated by interactions between 5-HTTLPR, DRD4 exon3 and COMT val/met variants (Benjamin et al., 2000; Lesch et al., 2003). The aim of this study was to explore the possible association between the DRD4 and 5-HTTLPR polymorphisms and quantitative measures of internalising and externalising behaviour problems, as assessed by the Child Behaviour Check-List 6–18 (CBCL 6–18, Achenbach et al., 2001), in early-adolescents. The sample consisted of 606 unrelated early-adolescents, 298 girls and 308 boys, aged 10–14, (mean age 12.7  0.9). After written informed consent was obtained, for each subjects CBCL 6–18 was completed by parents and DNA was extracted and genotyped from buccal epithelial cells. Significant interaction was observed by multivariate analyses only on Externalising Behaviour Problem scale: 5-HTTLPR  DRD4 (F: 5.095, P: 0.006). In the absence of the short 5-HTTLPR allele Externalising Behaviour scores are higher in the presence of the DRD4 seven-repeat allele.A possible role for the interaction between DRD4 and 5-HTTLPR polymorphisms on externalising behaviour problems in early adolescents is suggested. This result is consistent with earlier reports on an influence of the interaction between these two genes both in adult and in child behaviour. P12.31 PATTERNS OF COMORBIDITY OF MIGRAINE AND MENTAL DISORDERS Low N, Angst J, Cui L, and Merikangas KR 1 National Institute of Mental Health, Bethesda, MD, United States 2 University of Zurich, Zurich, Switzerland Migraine is a complex disorder characterized by phenotypic heterogeneity, complex patterns of inheritance and a high magnitude of

comorbidity with other disorders. This poster will present data from three clinical and population-based studies: (1) a family study of the coaggregation of migraine and mental disorders, (2) a prospective study of youth at high risk for the development of psychopathology, and (3) a twenty-year prospective community of young adults, and (4) a large cross-sectional study of a clinical sample of bipolar subjects. Each study examines the pattern of comorbidity between migraine and mental disorders. Results reveal a strong association between migraine and mood disorders, particularly with the atypical and bipolar subtypes of depression. We found that the clinical course of bipolar disorder differed among those with and without migraine. Corroboration of findings from clinical samples in community studies suggests that the association is not attributable to treatment-seeking bias. Inspection of prospective data reveals that the onset of anxiety disorders predates the onset of migraine in children followed by the onset of mood disorders in late adolescence and adulthood. Patterns of expression of migraine and mood disorders in families are consistent with the explanation that migraine may be an index of a discrete subtype of bipolar spectrum disorder. P12.32 A LARGE-SCALE SEARCH FOR GENES UNDERLYING BEHAVIOUR, COGNITION AND DEPRESSION van Belzen MJ, Posthuma D, Hoogendijk WJ, Boomsma DI, and Heutink P VU University Medical Center and VU University, Amsterdam, Netherlands Traits like behaviour, cognition and depression are at least partly influenced by genetic factors. They are of complex inheritance and are likely to be caused by many different genes which, each by itself, contribute only little to the total genetic background of the trait. These susceptibility genes may be involved in one of several processes underlying the complex phenotype. Therefore, studying so-called endophenotypes, as opposed to studying the complex phenotype itself, may ease isolation and identification of each of these subsets of genes. The new Center for Neurogenomics and Cognitive Research (CNCR) initiated two programs in which large patient cohorts are established in search for genes underlying behaviour, cognition and depression. The first program aims at identifying genes involved in attention, memory and cognition. For this purpose, 500 twin pairs phenotyped for the presence of attention problems have been collected. In addition, detailed phenotypic cognition data is available for 2400 subjects consisting of twin pairs and siblings. The investigated endophenotypes have been selected from three major functional domains of cognition: speed of information processing, attention and working memory capacity. The aim of the second program is the identification of genes involved in the development of major depression. For this program, approximately 1500 twin pairs and 2400 single patients are available, with information on drug and endocrine response. Genome-wide screens will be conducted in DZ twin pairs from both programs. In addition, polymorphisms in candidate genes will be tested for association with selected (endo)phenotypes. This will allow combined linkage and within/between family association studies in order to assess the contribution of these polymorphisms to the variation in the endophenotypes. Pilot studies have already started for both programs, in which the serotonin transporter is tested in relation with depression and the dopamine D4 receptor in relation with attention.

P13.1 HOW SYMPTOMATIC ARE NON-PSYCHOTIC-FIRST DEGREE RELATIVES OF PATIENTS WITH SCHIZOPHRENIA OR OTHER PSYCHOSIS? Valero JS,1 Mulet B,4 Montserrat C,4 Pe´rez M,1 Gutie´rrez A,2 Rodrı´guez T,2 Jariod M,1 Martorell L,4 and Labad A1 1 HPU Institut Pere Mata/FMCS-URV, Reus, Spain 2 HPUInstitut Pere Mata, Reus, Spain 3 FMCS-URV, Reus, Spain 4 Fundacio´ Institut de Recerca en Cie`ncies de la Salut-IRCIS, Reus, Spain Establishing that normal and disease states represent end points of a single continuum of liability has important implications for understanding the genetic architecture of complex disorders like schizophrenia. A recent paper (1) has shown that positive and negative symptoms in schizophrenia (SZ) predict corresponding schizotypal symptoms in relatives. But positive and negative symptoms also appear

Abstracts in other non-schizophrenic psychotic (P) disorders. The aim of this study is to clarify if these symptoms are expressed along a schizotaxic phenotype dimensional axis, along a psychotic one or on both. We studied 66 patients (31 with SZ), 95 first degree relatives (54 of SZ patients) and 14 normal control subjects. All the individuals were interviewed using the Schedules for Clinical Assessment in Neuropsychiatry (SCAN) and diagnosed according to the DSM-IV and the International Classification of Diseases, 10th Revision, using a computerized scoring program. Symptoms were also measured using the Positive and Negative Symptom Scale (PANSS). Data were analysed comparing medians with the Mann-Whitney test. Both SZ and P patients had significatively more symptoms than their relatives (P < 0.001). Differences between relatives of SZ patients vs. relatives of P patients and healthy controls were also significative for the positive syndrome of PANSS. Of these symptoms, conceptual disorganization marked the gap. Differences in negative symptoms were found between patients (Sz vs P), but not between their relatives. Positive syndrome appear to relate with a schizotaxic dimension phenotype rather than with a psychotic continuum. The marker of this syndrome would be the conceptual disorganization symptom. Non differentiation in the negative syndrome between SZ and P relatives must be furtherly studied splitting primary and secondary negative symptoms. Supported in part by grants FIS 01/1605, Marato´TV3/010510 and Marato´TV3/010610. P13.2 EVIDENCE OF LINKAGE AND ASSOCIATION ON 18P11.2 FOR PSYCHOSES Mukherjee O,1 Purushottam M,1 Kubendran S,1 Kiran HGK,1 Brahamachari SK,2 Mazumder PP,3 Ghosh S,3 and Jain S1 1 National Institute of Mental Health and Neurosciences, Bangalore, India 2 Institute of genomics and Integrative Biology, Delhi, India 3 Indian Statistical Institute, Kolkata, India Bipolar disorder (BPAD) and schizophrenia (SCZ) are common, severe psychiatric illnesses increasingly seen to share many epidemiological, genetic and clinical features. Several positive linkage findings for BPAD and SCZ are reported to the pericentromeric region of chromosome 18. As these findings are not statistically significant, and often not reproducible, it is hypothesized that the pericentromeric region of chromosome 18 may harbor gene of modest affect for mental illness. Recent studies have used composite phenotypes like ‘psychosis’ to localize susceptibility regions for these disorders. We investigated the long and the short arm of chromosome 18 for any putative loci for the phenotype psychosis, that consists of schizophrenia, bipolar disorder (BP1), psychosis NOS and schizoaffective disorder. Two point Lod score analysis performed on a series of 52 multiplex extended pedigrees with 23 polymorphic markers yielded a Lod support of 2.02 at D18S37. An independent set of 159 parent offspring trios was used to confirm this suggestive finding. The TDT analysis yielded a strong support for association between the marker D18S453 and the disease allele (w2 ¼ 2.84, P < 0.028). The close proximity of the parametric LOD report and the TDT analysis suggests that this region harbors putative loci for psychotic symptoms. An independent result of this magnitude validates previous reports and warrants further screening of the region to identify causal genes for psychoses.

P13.3 META-ANALYSIS OF THE ASSOCIATION BETWEEN TWO POLYMORPHISMS IN THE SEROTONIN TRANSPORTER GENE AND SCHIZOPHRENIA, BIPOLAR DISORDER, AND UNIPOLAR DISORDER Su JLS, Faraone SVF, Glatt SJG, and Tsuang MTT Harvard University, Boston, United States Family, twin, and adoption studies show that psychiatric diseases including schizophrenia (SZ), bipolar disorder (BP), and unipolar disorder (UP) have a substantial genetic component. For these illnesses, both positive and negative associations have been reported for two polymorphisms located in the serotonin transporter gene (5HTT) on chromosome 17: a 17-base-pair (bp) variable-number tandemrepeat (VNTR) in intron 2 and a 44-bp insertion/deletion in the promoter region. Thus, associations between these 5-HTT polymorphisms and SZ, BP, and UP remain unclear. The present work investigates these potential associations in meta-analyses that maximize the power to find associations between each disease and the two 5-HTT

121

polymorphisms. We applied meta-analysis techniques to case-control studies of two 5-HTT polymorphisms and three psychiatric disorders (BP, UP, and SZ), resulting in six meta-analyses. For each polymorphism, we assessed the evidence for allelic associations, heterogeneity among studies, the influence of individual studies, and the potential for publication bias. The short allele(s) of the 44-bp insertion/deletion polymorphism showed a significant association for BP (OR ¼ 1.13, P ¼ 0.001) but not UP and SZ. For the 17-bp VNTR, an increase in the number of tandem repeats had no significant association with any of the disorders. The small but significant effects of the 44-bp insertion/ deletion polymorphism for BP is consistent with being one of many genes that contributes to the multi-factorial nature of these psychiatric disorders.

P13.4 PROTEOMIC ANALYSIS OF THE DORSO-LATERAL PREFRONTAL CORTEX IN SCHIZOPHRENIA AND BIPOLAR DISORDER Pennington K,1 Beasley CL,2 Hudson L,1 English J,1 Dunn M,2 and Cotter D1 1 RCSI, Dublin, Ireland 2 Institute of Psychiatry, London, United Kingdom The dorso-lateral prefrontal cortex (dlPFC) is located within the frontal lobe of the brain and has been strongly implicated in the pathophysiology of schizophrenia (SCZ) and bipolar disorder (BPD). Its abnormality in both size and activity throughout the course of these disorders is well-documented and thus it is considered a candidate region for revealing potential causative factors and markers of the disease process. The holistic approach offered by proteomic methods enables the comparison of hundreds of protein profiles expressed within a complex brain tissue sample. In this study we have combined 2D gel electrophoresis and mass spectrometry (MALDI-TOF-TOF) in order to isolate and identify differentially expressed proteins in schizophrenia and bipolar disorder. Post-mortem brain samples comprising of 35 controls, 35 schizophrenics and 35 bipolar disorder were obtained from the Stanley Foundation Brain Consortium. Consequently, dissected cortical grey matter was used (100 ug protein) to run analytical gels with pH range 4–7 in the 1st dimension and 12% SDS in the 2nd dimension. All 105 gels were scanned and the images imported into Progenesis Workstation 3.0. Following editing, matching and warping of these images multivariate statistical analysis will be performed. Protein profiles differentially expressed within disease groups will be removed and following digestion, identified through the generation of a peptide mass fingerprint using MALDI-TOF-TOF. The results from this analysis will provide important insights into the molecular mechanisms underlying SCZ and BPD in the dlPFC, helping further to elucidate the causes of abnormality in these disorders.

P13.5 PROTEOMIC ANALYSIS OF THE ANTERIOR CINGULATE CORTEX IN SCHIZOPHRENIA, BIPOLAR DISORDER AND MAJOR DEPRESSIVE DISORDER: EVIDENCE FOR METABOLIC DISTURBANCES Pennington K,1 Beasley CL,2 Wait R,3 Dunn M,2 and Cotter D1 1 RCSI, Dublin, Ireland 2 Institute of Psychiatry, London, United Kingdom 3 Imperial College, London, United Kingdom Abnormalities of the anterior cingulate cortex have previously been described in patients with schizophrenia, major depressive disorder and bipolar disorder. In this study 2-dimensional gel electrophoresis (2DGE) was performed, followed by mass spectrometric sequencing of proteins to identify disease-specific protein changes within the anterior cingulate cortex in major psychiatric disorders. Samples were obtained from subjects with schizophrenia, bipolar disorder, major depressive disorder and controls (15 subjects per group). The 2DGE system comprised immobilised pH gradients 4–7 and 6–9 in the first dimension and polyacrylamide gel electrophoresis in the second dimension. Protein spots were analysed quantitatively and compared between control and disease groups by analysis of covariance. 33 spots were differentially expressed in one or more of the disease groups. Proteins within 23 of these spots were identified by mass spectrometry as: aconitate hydratase, fructose bisphosphate aldolase A, glyceraldehyde3-phosphate dehydrogenase, carbonyl reductase, carbonic anhydrase, alpha and beta tubulin, malate dehydrogenase, glutamate dehydro-

122

Abstracts

genase, glutamate synthetase, neuronal protein 25, dihydropyrimidinase-related protein-2, vacuolar ATP synthase, ATP synthase beta chain, sorcin, creatine kinase, succinyl CoA and guanine nucleotide binding protein G(i). We have identified 18 proteins with altered expression in one or more disorders. Many of these proteins are involved in energy metabolism, glutamate neurotransmission and the cytoskeleton. These findings replicate and extend previous observations, and provide important evidence for metabolic disturbances in the post-mortem brain of subjects with these disorders.

P13.6 LINKAGE DISEQUILIBRIUM IN THE DYSBINDIN (DTNBP1) GENE REGION AMONG PSYCHOTIC PATIENTS FROM A GENETIC ISOLATE IN ISRAEL Kohn Y,1 Danilovich E,2 Karni O,3 Kanyas K,3 Lerer B,3 and Lifton RP1 1 Department of Genetics, Yale University School of Medicine, New Haven, United States 2 Eitanim Mental Health Center, Jerusalem, Israel 3 Biological Psychiatry Laboratory, Hadassah-Hebrew University Medical Center, Jerusalem, Israel SNP’s and SNP haplotypes within the Dysbindin (DTNBP1) gene have been found to be associated with schizophrenia. Similar to findings in other genes, the significance of these associations is debated. Two main limitations of the findings are the inconsistency of replications, and the lack of clear functional significance of the associated variants. More replications as well as identification of other candidate genes are clearly needed. We have recently demonstrated increased sharing of microsatellite haplotypes at the Dysbindin region among psychotic patients from a genetically isolated population in Israel (Kohn et al., AJMG, Neuropsychiatric Genetics, in press). This isolated population is thought to descend from a small group of Moslem immigrants who came to Israel from the Caucasus 500 years ago. A bottle neck event occurred 250 years ago and reduced the number of founders even further. The 5000 inhabitants of this isolated community consider themselves to be descendents of one common paternal ancestor who lived at that time. Y chromosome haplotypes strongly support this oral tradition. We have recruited 52 affected individuals from this isolate with any major psychiatric disorder (mostly psychotic). 42 first degree relatives were also recruited for phase determination of haplotypes and for construction of a control group of non-transmitted haplotypes. In our previous study, we investigated haplotype sharing of 359 microsatellite markers, 4–5 cM apart, on 9 candidate chromosomes, and identified significant evidence for linkage disequilibrium at the region of the Dysbindin gene as well as on chromosome 1p36. In this study we have further investigated linkage and linkage disequilibrium in this sample by studying a whole genome dense map of SNPs, the 10,000 SNP chip by Affymetrix. Further evidence for haplotype sharing in the Dysbindin gene region was observed, as well as suggestive linkage and haplotype sharing in other genomic regions. We have also sequenced the Dysbindin gene exons and flanking intronic regions in patients with the shared haplotype. Even though no obvious amino-acid-changecausing mutation was found, we did identify an association with a SNP in the 30 UTR region of the last exon of Dysbindin. The rare allele of this gene (15% in other populations) was found in 9/10 of patients with the shared haplotype, and in 20% of all affected individuals from the sample. TDT in 9/11 informative trios was also significant. Thus far, we were able to replicate the association of SNPs in the Dysbindin region with a broad phenotype of psychosis. As in other studies, we also find a complex nature of these associations. We are currently pursuing the study of variations within the Dysbindin region as well in other regions of suggestive linkage and linkage disequilibrium in this sample.

P13.7 ADDITIONAL SUPPORT FOR SCHIZOPHRENIA AND BIPOLAR DISORDER LINKAGE TO CHROMOSOME 3Q29 Schosser A,1 Fuchs K,2 Leisch F,3 Bailer U,1 Kasper S,1 Sieghart W,2 Hornik K,4 and Aschauer HN1 1 Department of General Psychiatry, Medical University Vienna, Vienna, Austria 2 Brain Research Institute of the Medical University Vienna, Division of Biochemistry and Molecular Biology, and Division of Biochemical Psychiatry, Medical University Vienna, Vienna, Austria 3 Institut fu¨r Statistik und Wahrscheinlichkeitstheorie, University of Technology, Vienna, Austria 4 Department of Statistics, Wirtschaftsuniversita¨t, Vienna, Austria

We recently published a follow-up linkage study (Schosser et al., 2004) within the chromosome 3q29 region in schizophrenia and bipolar affective disorder families (highest NPL [non parametric lod] score Zall ¼ 1.93296, highest P-value ¼ 0.032166) after conducting a genome scan (Bailer et al., 2002), resulting in evidence for linkage of both disorders to this region (highest NPL score Zall ¼ 3.74, highest Pvalue ¼ 0.0003). Using the same family sample (five pedigrees with schizophrenic index patients and three pedigrees with index bipolar disorder patients N ¼ 86, 50 of them were available for genotyping), we now genotyped five additional markers, spanning 2.6 cM (centiMorgan) within the region of highest linkage of our recently published follow-up linkage analysis. That is to say, we conducted an additional fine-mapping, narrowing down our newly identified candidate region. Linkage analysis was performed using the GENEHUNTER program version 2.1_r3 beta. The highest NPL score Zall observed was 4.03484 with SNP rs225, corresponding to P ¼ 0.000096. Doing subset analyses of the bipolar disorder and schizophrenia families separately, linkage signals arose substantially from bipolar disorder families (the highest P value was 0.000061), with strong contribution from schizophrenia families (the highest P value was 0.031433), too. In conclusion, our new linkage results strongly support the previous linkage finding of schizophrenia and bipolar affective disorder to chromosome 3q29.

P13.8 IDENTIFICATION AND CHARACTERISATION OF A GENE PREDISPOSING TO BOTH BIPOLAR AND UNIPOLAR AFECTIVE DISORDERS Barden N,1 Harvey M,1 Shink E,1 Tremblay M,1 Gagne´ B,1 Raymond C,1 Labbe´ M,1 St-Onge L,2 Stadler H,2 Sillaber I,2 Paez-Pereda M,2 Mu¨ller-Mysok B,3 Binder E,3 and Holsboer F3 1 Neuroscience, CRCHUL Research Centre, Sainte-Foy, Quebec, Canada 2 NeuroNova Pharmaceuticals, Munich, Germany 3 Max-Planck Intitute of Psychiatry, Munich, Germany We have previously reported linkage analysis results that pointed to the localisation of a gene for susceptibility to bipolar disorder (BP) in the chromosome 12q23-24 region. These results were supported by followup analysis using 20 polymorphic microsatellite markers distributed over 7.7 Mb of genomic sequence on 12q24.31. Moreover, association studies with these markers revealed significant allelic associations with BP for two loci, NBG6 (P ¼ 0.008) and NBG12 (P < 0.001). We identified and genotyped polymorphisms in close to 40 genes around these markers. We were lead to focus on three adjacent candidate genes that overlapped NBG6 and had similar functional roles in cell signalling mechanisms. Mutation analysis in regulatory and coding regions of these three candidate genes allowed us to identify 105 variants, of which twenty-seven are found in coding regions. Forty-four SNPs were genotyped in the Quebec case/control sample (BP affected ¼ 213, control ¼ 214) to perform association studies. We identified five SNPs that gave significant association with either allelic association, genotypic association or both. In one of the genes, three of these polymorphisms are located in the intracellular C-terminal domain of the protein product, a domain known to be critical for protein function. The two other associated SNPs were found in the two adjacent genes and it is to be expected that SNPs in these proximal genes would be in LD. Using a sliding-window strategy with 15 successive SNPs, we observed significant haplotypic association (P < 0.001) overlapping this genomic region. Further SNP-based association studies on this gene with a larger case/control sample (largely unipolar, BP/UP affected ¼ 315, control ¼ 220) of European origin demonstrated significant associations as well. Together, our data strongly suggests that, alone or in combination, polymorphisms in this gene can confer predisposition to BP/unipolar affective disorders. To confirm the role of this gene in affective disorders, we have performed functional and biological studies. The gene product is localised in hippocampus and is down regulated by stress and up regulated by antidepressant therapy. A reduction of gene expression in mice by small interfering RNA lead to a significant increase in passive behaviour in the Porsolt forced swim test, which is generally construed as depressive-like behaviour. Gene activity can also be regulated by known agonists/antagonists and while antagonists have no effect on behaviour, treatment of mice with agonists reverses depressive-like behaviour. Together these results clearly indicate the identification of a novel gene involved in the susceptibility to affective disorders and indicates new therapeutic approaches.

Abstracts P13.9 COMMON GENETIC BASIS OF SCHIZOPHRENIA, SCHIZOAFFECTIVE DISORDER AND BIPOLAR ILLNESS Stassen HH,1 Bridler R,1 Hell D,1 Nurnberg P,3 Sander T,3 Szegedi A,2 Thiele H,3 Weisbrod M,4 and Scharfetter C1 1 Psychiatric University Hospital, Zurich, Switzerland 2 Psychiatric University Hospital Charite, Berlin, Germany 3 Max-Delbruck-Center, Berlin, Germany 4 Psychiatric University Hospital, Heidelberg, Germany The functional psychoses ‘‘schizophrenia,’’ ‘‘schizoaffective disorders,’’ and ‘‘bipolar illness’’ are diagnostic entities that exhibit a considerable overlap in terms of underlying clinical syndromes. The utility of current diagnostic entities for genetic studies seems limited, since psychiatric diagnoses do not offer much information about a patient’s response to treatment and prognosis. We have introduced (1) quantitative, syndrome-oriented approaches to psychopathology which extend the DSM-IV or ICD-10 phenotype definitions by replacing the dichotomy of the diagnostic schema by dimensional quantities, and (2) genotype-tophenotype strategies which allow one to correlate multilocus deviations in the genome sequence with quantitative scores on the phenotype level, so that the subjects’ individual characteristics of the disorder, their lifestyles, and other environmental details can be included in the genetic model as well. Based on a test-retest design and genome scan data of 71 nuclear families (468 genotyped subjects, 451 marker locations), we treated the families with index cases suffering from bipolar illness as ‘‘training’’ samples, and used the families with index cases suffering from schizophrenia or schizoaffective disorder as independent ‘‘test’’ samples. During our systematic search for oligogenic marker configurations we evaluated the between-sib similarities, which were expected to deviate from ‘‘0.5’’ in affected sib pairs if the region of interest contained markers close to vulnerability genes. The reference value ‘‘0.5’’ was derived from parent-offspring similarities which are always ‘‘0.5,’’ irrespective of the parent and offspring affection status. Preliminary results suggested several vulnerability loci on chromosomes 1, 5, 6, 13 and 18, which were reproducible across the two samples under comparison and displayed some remarkable overlap with previous findings in the literature.

P13.10 GLUTAMATE DECARBOXYLASE 1 GENE: A POTENTIAL CANDIDATE GENE FOR BIPOLAR AFFECTIVE DISORDER AND SCHIZOPHRENIA? Buttenschøn HN,1 Lundorf MD,1 Kruse TA,5 Blackwood DHR,2 Foldager L,1 Muir WJ,2 Murray V,3 Pelosi AJ,4 Ewald HL,1 and Mors O1 1 Institute of Basic Psychiatric Research, Psychiatric hospital in Aarhus, Aarhus, Denmark 2 Department of Psychiatry, University of Edinburgh, Edinburgh, United Kingdom 3 Department of Adolescent Psychiatry, Gartnavel Royal Hospital, Glasgow, United Kingdom 4 Department of Psychiatry, Hairmyres Hospital, Glasgow, United Kingdom 5 Department of Clinical Biochemistry and Genetics, Odense, Denmark Evidence from several studies suggests that the Glutamate decarboxylase gene (GAD1) on chromosome 2 is a functional candidate susceptibility gene for bipolar affective disorder (BPAD) and schizophrenia. The gene encodes the gamma-aminobutyric acid synthetic enzyme, GAD67. Several studies have supported the importance of this gene by showing decreased levels of GAD67 mRNA and protein levels in patients with BPAD and schizophrenia. Previous studies furthermore suggest linkage between D2S326 near GAD1 and BPAD. We systematically screened GAD1 exons, flanking intronic sequences and promoter sequence for polymorphisms in 16 BPAD patients and 5 controls from Denmark. We identified eight single nucleotide polymorphisms (SNPs) including two in the promoter sequence. To improve coverage of the GAD1 gene and in an attempt to include nonsynonymous SNPs we also validated public GAD1 database SNPs. In total eight SNPs were included in the association analysis. They span the GAD1 gene from the promoter to the 30 untranslated region. An association study of SNPs covering GAD1 was performed in a Danish sample of 82 BPAD subjects and 120 controls and in a Scottish sample of 193 individuals with schizophrenia, 200 BPAD subjects and 203 controls. Linkage disequilibrium (LD) and haplotype frequencies were estimated from genotype data from eight SNPs. Strong pairwise LD was observed among all pairs of neighbouring markers. In the Danish sample we found weak association between BPAD and two promoter

123

SNPs spaced 1 kb apart. Furthermore, one, two and three loci haplotype analysis showed weak association with BPAD in the Danish sample. P13.11 A SEARCH FOR ASSOCIATION OF MARKERS AND SEGMENTS ON CHROMOSOME 2 WITH BIPOLAR AFFECTIVE DISORDER AND SCHIZOPHRENIA IN PATIENTS FROM THE FAROE ISLANDS Als TD,1 Dahl HA,2 Flint TJ,1 Wang AG,3 Vang M,4 Kruse TA,2 Ewald H,1 and Mors O1 1 Center for Basic Psychiatric Research, Aarhus University Hospital, Aarhus, Denmark 2 Department of Clinical Biochemistry and Genetics, Odense University Hospital, Odense, Denmark 3 Department of Psychiatry, Copenhagen University Hospital, Copenhagen, Denmark 4 Department of Psychiatry, National Hospital Faroe Islands, Torshavn, Denmark Chromosome 2 may potentially harbor genes involved in bipolar affective disorder and schizophrenia, as suggested by several linkage studies. There are, however, also negative reports of linkage to bipolar disorder and schizophrenia within these regions. Distantly related patients from genetically isolated populations may be valuable in the mapping of complex traits, as they potentially share identical disease mutations and alleles of markers surrounding the risk loci. Patients with schizophrenia (n ¼ 11) and bipolar affective disorder (n ¼ 17) from the relatively isolated population of the Faroe Islands were initially genotyped for 36 polymorphic markers (with an average distance of 7.3 cM) on chromosome 2 in a search for allelic association and haplotype sharing. Forty-four unaffected individuals from the same panmictic population were genotyped for the same set of markers. Association tests were performed for both single marker and twomarker segments, considering the two disorders separately, and combined. For bipolar affective disorder a single marker on 2q31.1 turned out to have alleles occurring with different frequencies among cases and controls, supporting this previously identified candidate region. There was no support for markers in the previously identified 2p13-16 and 2q21-24 regions. When considering schizophrenia there was evidence for association with a single marker and a two-marker segment on 2q31. In addition there was evidence for association between schizophrenia and a two-marker segment on 2p22.3. There was, however no support for the previously identified regions on 2p1214 and 2q37. When combining the two disorders, 2q31.1 received additional support from two single markers and one two-marker segment. In conclusion the present study identified 2q31.1 as a candidate region not only for bipolar affective disorder but also for schizophrenia, suggesting that the two disorders may have at least some risk genes in common. Follow-up analysis with seven additional microsatellite markers flanking the interesting markers will be presented as single marker and multi-marker analyses.

P13.12 AN INTEGRATED APPROACH TOWARDS THE IDENTIFICATION OF A SUSCEPTIBILITY GENE FOR PSYCHOSIS Evans KL,1 Le Hellard S,1 Underwood S,1 Thomson PA,1 Morris SW,1 Adams R,1 Lee AJ,1 Wray NR,1 Ballereau S,1 Torrance HS,1 Anderson S,1 Beekman J,1 Muir WJ,2 Blackwood DHR,2 and Porteous DJ1 1 Medical Genetics, University of Edinburgh, Edinburgh, United Kingdom 2 Department of Psychiatry, University of Edinburgh, Edinburgh, United Kingdom We have an integrated programme of clinical and molecular genetic studies of a large number of Scottish families and individual patients with BPAD and related phenotypes. Genetic analysis of this resource has lead to the identification of a vulnerability locus for affective disorder at human chromosome 4p within an extended Scottish pedigree F22 (max multipoint LOD ¼ 4.8, Blackwood et al., 1996). Subsequently, a number of independent groups have provided confirmatory evidence for linkage to the region. Currently, we have access to biological material from four families with chromosome 4p linked psychosis. We have defined at high resolution the haplotype that segregates with illness in each of the four families. Comparison of these overlapping disease haplotypes implicates two 4 Mb sub regions of the

124

Abstracts

F22 candidate region. We are now attempting to prioritise these candidate regions further by testing for allele sharing between families. This ongoing work has involved the construction of transcript maps of the region, which allow the identification of SNPs from candidate genes amplified from family members and their localisation with respect to publicly available SNPs. To date, we have scored 100 SNPs on a panel of family DNAs and identified a 120 kb region with preliminary evidence for allele sharing above that seen in control chromosomes. We are also using these SNPs and those HapMap SNPs that tag haplotypes at frequencies of greater than 10% within haplotype blocks in a case control association study. A strength of this work is that the samples used in the linkage and the association studies are both drawn from the Scottish population. Finally, we are initiating studies of differential RNA expression between cases and controls using Taqman based allele-specific quantitative RT-PCR. We believe that the combination of genetic analysis of both family and population resources with functional analyses provides the best means of identifying the susceptibility gene for psychosis from this region of chromosome 4. P13.13 FAMILIAL RESEMBLANCE OF SCHIZOTYPAL TRAITS IN UNAFFECTED SIBLINGS OF PSYCHOSIS PATIENTS Mata I,1 Lopez-Ilundain JM,1 Perez-Nievas F,1 Beperet M,1 Albeniz A,1 and Murray RM2 1 Fundacion Argibide, Pamplona, Spain 2 Institute of Psychiatry, London, United Kingdom The heritability of schizotypy has been addressed by several sibling and twin studies. While sibling studies have reported familial correlations for Magical Ideation and Social Anhedonia, twin studies using such psychosis proneness scales have yielded contradictory results. Twin studies using schizotypy scales have reported heritability estimates of 40% for the Schizotypal Personality Questionnaire (SPQ) and 50% for the Oxford-Liverpool Inventory of Feelings and Experiences (O-LIFE). Moreover, it has been suggested that, while it appears that no familial association exists between positive and negative schizotypal traits, disorganized schizotypy may be genetically related to both types of schizotypal traits, thus possibly reflecting the core genetic liability to schizotypy (Linney et al., 2003). In the present study, a sample of 26 unaffected sibling pairs of psychosis probands were assessed with the SPQ. In order to study the intrafamilial correlation for schizotypy, the intraclass correlation method, with the SPQ total score and each of the three (cognitive perceptual, interpersonal, disorganized) factor scores as continuous variables, was used. Scores on the SPQ total score showed a clear intrafamilial resemblance (P ¼ 0.0005). The same occurred for the SPQ interpersonal (P ¼ 0.003) and disorganized factor score (P ¼ 0.004), with a trend for the cognitive-perceptual factor score (P ¼ 0.06). These results would imply that schizotypy, as measured with the SPQ, show a significant degree of familiality. Apparently, the interpersonal (negative) and disorganized dimensions of schizotypy may be more familial than the cognitive-perceptual (positive) dimension. Overall, it can be suggested that schizotypy, and specially its negative and disorganized dimensions, may be a useful phenotype for molecular genetic studies.

P13.14 CAN THE PATTERN OF SCHIZOTYPAL TRAITS IN PSYCHOSIS PATIENTS PREDICT THE ONE EXPRESSED BY THEIR FIRST-DEGREE RELATIVES? Lopez-Ilundain JM,1 Mata I,1 Albeniz A,1 Beperet M,1 Perez-Nievas F,1 and Murray RM2 1 Fundacion Argibide, Pamplona, Spain 2 Institute of Psychiatry, London, United Kingdom The finding of a higher rate of schizotypal personality disorder (SPD) in the relatives of schizophrenia patients when compared to the relatives of control subjects, led to the conceptualisation of SPD as a phenotypic expression of familial-genetic liability to schizophrenia. However, as the specificity of SPD to schizophrenia has not been demonstrated, it has also been suggested that SPD might be a phenotypic expression of familial-genetic liability to psychoses. Factor analyses have consistently shown that schizotypy breaks down into three factors of cognitive-perceptual (positive), interpersonal (negative), and disorganized symptoms, these factors bearing considerable resemblance to those reported in schizophrenia patients, namely positive, negative,

and disorganized. Studies on the familial resemblance of symptoms shown by schizophrenia patients and their relatives have reported a positive correlation for negative symptoms between patients and their relatives, and between positive symptoms in schizophrenia probands and general schizotypy in their relatives. Although a positive correlation between schizophrenia patients and their relatives has been reported for Physical Anhedonia and Perceptual Aberration, to our knowledge, no previous study has used a scale based on DSM criteria of SPD to address this issue. In the present study, 75 unrelated probands with psychoses and 201 of their first-degree relatives were assessed with the Schizotypal Personality Questionnaire (SPQ). Linear regression analyses, with age and sex of both probands and relatives also entered as predictor variables, were carried out to examine relationships between scores on the SPQ in the probands and their respective first-degree relatives in order to test whether the scores in the probands could predict the scores in their relatives. We found that the SPQ disorganized factor score was the only one that significantly predicted the SPQ total score in the relatives. Moreover, the SPQ disorganized factor score predicted higher scores on the SPQ cognitive-perceptual and interpersonal factors in the relatives. Our finding of higher degree of schizotypal traits among relatives of patients who themselves have a high degree of disorganized schizotypal traits would support the hypothesis that the disorganized dimension of psychoses increases the familial liability to psychotic disorders.

P13.15 PREMORBID ADJUSTMENT IN SCHIZOPHRENIA, BIPOLAR AFFECTIVE DISORDER AND MAJOR DEPRESSION—A NEGLECTED FACTOR IN PHENOTYPE DEFINITION Braeman C,1 Ohlraun S,1 Hoefels S,2 Kovalenko S,2 Gross M,2 Tullius M,2 Ruether E,3 Schumacher J,4 Cichon S,4 Propping P,4 Schulze TG,1 Noethen MM,4 and Rietschel M1 1 Central Institute of Mental Health, Division of Genetic Epidemiology in Psychiatry, Mannheim, Germany 2 University of Bonn, Department of Psychiatry, Bonn, Germany 3 University of Goettingen, Department of Psychiatry, Goettingen, Germany 4 University of Bonn, Institute of Human Genetics, Bonn, Germany Phenotype definition is a crucial aspect of genetic studies. Research has suggested that differences in onset and course of disorder may provide a measure for homogenising samples. Whereas a relationship between poor Premorbid adjustment (PMA) and more negative symptoms and poor outcome has been shown in schizophrenia, scant research examining PMA in affective disorders exists. Further investigation in large, well characterised samples is warranted. We measured PMA in a sample of German patients who were systematically recruited from consecutive hospital admissions. They were assigned DSM IV diagnoses after administration of a SCID interview: schizophrenia (SCZ n ¼ 316), bipolar affective disorder (BPAD n ¼ 346), or major depression (MD n ¼ 110). PMA was measured only up to age 18 to avoid confusion between premorbid and prodromal phases and patients with age of onset before 18 were excluded. Average age of onset was for SCZ ¼ 26, BPAD ¼ 29, and MD ¼ 36. The German version of the PMA Scale was used. Subscale scores were obtained for sociability, peer relationships,scholastic performance, adaptation to school and sociosexual aspects of adolescent life. Lower scores indicate better adjustment. There is a significant difference between total PMA scores for patients with SCZ and those with MD (P ¼ 0.004) and BPAD (P ¼ 0.001), the SCZ patients showing higher scores. Bipolar patients and MD patients do not differ significantly in the total score and scores on peer relationships, scholastic performance, adaptation to school and sociosexual aspects, but do differ significantly on the measure of sociability (P ¼ 0.001), with BPAD patients showing lower scores. MD patients have significantly lower scores than SCZ patients across all subscales except for that of sociability, where MD and SCZ individuals do not differ. Our results confirm previous suggestions of differences in PMA between schizophrenia and bipolar disorder, and suggest that several dimensions of premorbid adjustment, in particular sociablity,are discriminating factors amongst patients with illness sufficiently severe to require hospitalisation. We will present data from assessment of PMA in healthy controls, the relationship of PMA to family history of psychiatric illness, PMA within diagnostic categories and against individual psychopathology items (OPCRIT). PMA is potentially useful in defining more homogenous samples for genetic studies.

Abstracts P13.16 SUPPORT FOR THE INVOLVEMENT OF NEUREGULIN 1 AS A SUSCEPTIBILITY GENE IN MAJOR PSYCHIATRIC DISORDERS IN 820 TRIOS FROM BULGARIA Georgieva LNG,1 Dimitrova AD,2 Ivanov DKI,1 Nikolov INN,1 Grozeva DG,1 Vladimirova RV,2 Janakiev NJ,2 Toncheva DIT,2 Owen MJO,1 O’Donovan MCO’D,1 and Kirov GK1 1 Department of Psychological Medicine, University of Wales College of Medicine, Cardiff, United Kingdom 2 Department of Medical Genetics, Medical University Sofia, Sofia, Bulgaria The Neuregulin 1 gene (NRG1) has been implicated as a susceptibility factor in schizophrenia in association studies from Iceland and Scotland. We wanted to examine this gene for association with illness in our parent-offspring trios sample from Bulgaria. We first genotyped one SNP (SNP8NRG221533) and two microsatellite markers (NRG1_478b14834, NRG1_420m9-1395) that define the ‘‘core haplotype’’ implicated in the original studies in 288 trios affected with schizophrenia (SZ) or schizoaffective (SA) disorder. The result for the ‘‘core haplotype’’ was negative, however the SNP was significant on its own, with the trend in the opposite direction to that reported in the previous studies. We therefore genotyped this marker in a total of 539 trios affected with SZ/SA and 236 trios affected with BP1. Both results were significant: for SZ/SA the common allele was transmitted 272 and not transmitted 213 times from heterozygous parents to affected offspring, P ¼ 0.007. In the BP1 trios it was the minor allele that was overtransmitted: 120:86, P ¼ 0.018. We then genotyped 5 additional SNPs around this marker and established that the signal was coming only from the original marker, which was in a haplotype block of around 50kb in the 50 flanking region of NRG1. We screened 12 fragments of this region with DHPLC in 24 SZ cases but most of the polymorphisms identified were already present in public databases. We genotyped individually 5 new SNPs in a subset of 150-300 trios and established that only one of them, rs7014762, contributed to the signal. We therefore typed it in the full sample of 820 trios. In the SZ/SA sample it produced a significant two-marker haplotype with SNP8NRG221533 but was not significant on its own. In the BP1 sample it was significant on its own: P ¼ 0.02. This result was more impressive when we restricted the analysis to the 139 bipolar cases that had the most typical bipolar presentation: T:NT ¼ 79:39, P ¼ 0.00023. P13.17 SCHIZOPHRENIA AND BIPOLAR DISORDER SHARE AT LEAST ONE COMMON GENETIC RISK LOCUS AT G72/G30 REGION Navon R, Cholostoy A, Kaganovich M, Bening-Abu-Shach U, and Korostishevsky M Department of Human Genetics and Molecular Medicine, Tel Aviv University, Tel Aviv, Israel Schizophrenia and bipolar disorder are major psychotic disorders with strong genetic components. Although they are classified as different entities there are emerging evidence that the two disorders share some common loci. Genome wide scan performed on both bipolar and schizophrenic pedigrees showed that chromosome 13q32-33 is a potential susceptibility region for those two diseases. Two overlapping genes G72 and G30 that are transcribed from opposite DNA strands were identified in this chromosome region. Recently, an association of the G72/G30 gene region with schizophrenia as well as with bipolar disorder was observed in number of independent studies. In the present study, genomic DNA from 35 schizophrenic, 35 bipolar patients and 35 normal controls obtained from Stanley Foundation Brain Collection were genotyped with seven SNPs at the G72/G30 gene region. Linkage disequilibrium analysis revealed two tightly linked marker blocks including 4 and 3 SNPs respectively. Significant association with schizophrenia and bipolar disorder was observed for the SNPs of the first block that overlap both genes. No difference between schizophrenic and bipolar samples was found regarding genotype and allele frequencies. The combined sample of schizophrenic and bipolar patients versus the controls strengthens the significance of the result. Likelihood ratio (LR) test based on the first block haplotypes reveals significant difference between total patient sample and controls (LR ¼ 16.91, d.f. ¼ 6, P-value < 0.01). Gene expression analysis of G72 and G30 performed on DLPFC samples demonstrated statistically significant correlations between the two gene expressions (r ¼ 0.64, P < 0.01) as well as an over expression of the G72 gene in the schizophrenic patients versus controls (P < 0.04). Significant correlation

125

between G72 expression and SNP genotypes was found in the control group for two of the SNPs (r ¼ 0.40, P < 0.02, r ¼ 0.34, P < 0.05). In contrast no correlation was observed in the patient group for any of the SNPs. P13.18 TESTING THE MITOCHONDRIA HYPOTHESIS IN BIPOLAR DISORDER AND SCHIZOPHRENIA BY LARGE-SCALE DNA MICROARRAY ANALYSIS OF POSTMORTEM BRAINS Iwamoto K, Bundo M, and Kato T Brain Science Institute, RIKEN, Wako, Japan We have been performing the comprehensive microarray analyses of postmortem brains of various mental disorders including bipolar disorder, schizophrenia, major depression, and alcoholism. To date, about 180 individual brain samples, which were mainly donated from the Stanley Foundation, were processed for analysis. Among them, expression profiles were successfully obtained from 161 samples. By using these data, we examined the possible expression changes of mitochondria-related genes in bipolar disorder and schizophrenia. Among the mitochondria-related probes on Affymetrix HU133Achip, 15.1% and 11.3% of the probes showed altered expressions in bipolar disorder and schizophrenia, respectively. In consistent with the previous microarray reports (Konradi et al., 2004, Prabakaran et al., 2004), most of them showed down-regulation compared to controls. We also found the pH-dependent down regulation of mitochondria-related probes, which was originally reported by Li (Li et al., 2004). Even after the effect of pH was controlled in data analysis, mitochondria-related probes including the component of respiratory chains still showed altered expressions in bipolar disorder or schizophrenia. Our results suggest that altered expression of mitochondria-related genes is involved in the pathophysiology of bipolar disorder and schizophrenia. P13.19 SEROTONIN TRANSPORTER (5-HTT) POLYMORPHISM ASSOCIATED WITH EMOTIONAL RESPONSE TO AUDITORY HALLUCINATIONS IN PSYCHOTIC PATIENTS Rivero O,1 Na´jera C,1 Molto´ MD,1 Gonza´lez JC,2 Aguilar E,2 de Frutos R,1 and Sanjua´n J2 1 Departamento de Gene´tica. Facultad de Biologı´a. Universitat de Vale`ncia., Valencia, Spain 2 Departamento de Psiquiatrı´a. Facultad de Medicina. Universitat de Vale`ncia., Valencia, Spain The serotonin transporter (5-HTT) is implicated in the recovering of the serotonin (5-HT) from the synaptic cleft to the presynaptic neuron, regulating the length of the serotoninergic response in the nervous system. Thus, this gene appears to be an interesting study target with influence in different psychiatric disorders. In humans, transcriptional activity of the 5-HTT gene is modulated by a polymorphic repetitive element (5-HTTLPR), with two principal alleles, long (l) and short (s). The l allele has greater transcriptional activity. Malhotra et al (1998) found a specific significant increase in intensity of auditory hallucinations (AH) in schizophrenic patients with the ll genotype providing preliminary evidence for a role of serotonin in the physiopathology of hallucinations and Pae et al (2003) found significant differences in the scores of positive and negative symptoms in schizophrenic patients according to 5-HTTLPR genotypes and alleles. In our association study, 200 schizophrenic patients with a clinical history of AH evaluated by means of the PSYRATS scale, and 136 blood donors as controls were investigated. A significant excess of the s allele in the patients was observed. Moreover, a highly significant correlation between the s allele and the emotional response to auditory hallucinations was also observed. These results are coherent with previous reports of significant associations between the allele s of this polymorphism and anxiety-related traits. P13.20 FOXP2 POLYMORPHISMS AND AUDITORY HALLUCINATIONS IN PSYCHOTIC PATIENTS Tolosa A,1 Na´jera C,1 Molto´ MD,1 Gonza´lez JC,2 Aguilar E,2 de Frutos R,1 and Sanjua´n J2 1 Departamento de Gene´tica. Facultad de Biologı´a. Universitat de Vale`ncia., Valencia, Spain 2 Departamento de Psiquiatrı´a. Facultad de Medicina. Universitat de Vale`ncia., Valencia, Spain

126

Abstracts

FOXP2 was the first gene characterized in which a mutation affects human speech and language. Several studies related impairments of this human ability to auditory hallucinations, one of the core symptoms of schizophrenia. In order to look for possible associations between these polymorphisms and auditory hallucinations, we compared the distribution of FOXP2 polymorphisms between psychotic patients and healthy controls. A sample of 200 patients with history of auditory hallucinations and 170 controls were studied. Several SNPs have been analysed in FOXP2: Intron3a, SNP 923875, the transition in exon 14 that causes speech deficits in the family KE and the following polymorphisms located at the putative regulatory region of the gene and described in NCBI database, SNP1358278, SNP 2396722, SNP 1597548, and SNP 7803667. Although, no significant differences were found between controls and schizophrenic patients, the SNP 923875 (A) allele was significantly associated with the extreme phenotype (chronicity of voices). We concluded that implication of this gene in the vulnerability for the persistence of auditory hallucinations can not be excluded. P13.21 ASSOCIATION AND LINKAGE ANALYSIS OF RGS4 POLYMORPHISMS WITH SCHIZOPHRENIA AND BIPOLAR DISORDER IN BRAZIL Cordeiro Q,1 Talkowski ME,2 Chowdari KW,2 Wood J,2 Nimgaonkar V,2 and Vallada H1 1 School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil 2 School of Medicine, University of Pittsburgh, Pittsburgh, United States Linkage and association studies have suggested that polymorphisms of the regulator of G-protein signaling 4 (RGS4) may confer risk for schizophrenia (SCZ). Suggestive evidence for association with bipolar disorder (BD) has also been presented. However, the associated alleles and haplotypes have differed among the samples. Data from other independent samples may clarify the putative associations. Hence, we investigated a Brazilian population comprising patients with SCZ (n ¼ 271) and BD (n ¼ 306), who were contrasted with 576 controls. Parents of 49 SCZ cases and 44 BD cases were available for transmission disequilibrium tests (TDT). Four RGS4 SNPs (1, 4, 7, 18) putatively associated with SCZ were investigated. In the SCZ samples, significant case-control differences were not observed for individual SNPs or haplotypes, though the TDT suggested transmission distortion similar to that observed in the initial report. For the BD sample, casecontrol comparisons revealed no significant differences for individual SNPs, but an omnibus test suggested differences in the overall distribution of haplotypes bearing all four SNPs (SNP-EM Omnibus likelihood ratio test, P ¼ 0.003). The TDT revealed over-transmission of allele A at SNP 7 (P ¼ 0.016), as well as haplotypes incorporating this allele. However, global tests incorporating all haplotypes yielded only suggestive trends for association (P ¼ 0.19). In conclusion, association with SCZ was not detected in the present analyses. Suggestive associations with BD detected here need to be investigated in a larger sample. P13.22 ASSOCIATION STUDY BETWEEN DRD2 TAQI A POLYMORPHISM AND FIRST-EPISODE PSYCHOSIS IN A BRAZILIAN SAMPLE Vasconcellos K, Cordeiro Q, Jardim D, Scazufca M, Busatto G, Menezes PR, Krieger JE, and Vallada H School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil Dopaminergic signaling has been involved both in the pathogenesis and clinical treatment of psychotic disorders. The present study investigates the TaqI A polymorphism of the dopamine receptor D2 gene (DRD2) as a susceptibility gene for first-onset psychotic subjects. The sample is part of an ongoing large epidemiological neuroimage project, sponsored by the Wellcome Trust, on first-onset psychosis. We analyzed the distribution of DRD2 TaqI A polymorphism in a sample of 92 individual cases of first-episode psychosis and 781 controls. No significant differences were found in allelic (psychosis: A1 ¼ 25.86%, A2 ¼ 74.13%, controls: A1 ¼ 28.23%, A2 ¼ 71.76%, X2 ¼ 0.44, 1 d.f., P ¼ 0.5) and genotypic (psychosis: A1/A1 ¼ 0.02%, A1/A2 ¼ 47.12%, A2/ A2 ¼ 50.57%, controls: A1/A1 ¼ 0.08%, A1/A2 ¼ 39.56%, A2/A2 ¼ 51.98%, X2 ¼ 4.92, 2d.f., P ¼ 0.08) distributions. No evidence of association between the DRD2 TaqI A polymorphism and our first-episode psychosis sample was found.

P13.23 BRAIN-DERIVED NEUROTROPHIC FACTOR (BDNF) AND SUSCEPTIBILITY TO FUNCTIONAL PSYCHOSIS Cordeiro Q, Michelon L, Sartorelli A, Destro HM, Miguita K, and Vallada H School of Medicine, University of Sa˜o Paulo, Sa˜o Paulo, Brazil Brain-derived neurotrophic factor (BDNF), plays an important role in neuronal survival, proliferation and plasticity. Recent studies have suggested its contribution in the pathogenesis of psychotic disorders. The present study investigated the distribution of the BDNF Val-66Met gene polymorphism in 341 patients with functional psychosis [223 bipolar disorder (BD) and 118 with schizophrenia (SCZ)] and in 252 healthy controls. There was an allelic association (P ¼ 0.0009, X2 ¼ 10.83, OR ¼ 1.75 1.23 < OR < 2.48) and a trend for genotypic association (P ¼ 0.051, X2 ¼ 10.5, 2d.f.), when we compared the group of patients with functional psychosis with the control group. Comparing only the SCZ group with the control group, there was no difference in the allelic (P ¼ 0.069, X2 ¼ 3.30, OR ¼ 1.51 0.94 < OR < 2.40) or genotypic (P ¼ 0.11, X2 ¼ 4.36, 2d.f.) distribution. However, in the analyses of the BD sample, there were allelic (P ¼ 0.0005, X2 ¼ 12.06, OR ¼ 1.88 1.29 < OR < 2.73) and genotypic (P ¼ 0.0025, X2 ¼ 11.91, 2d.f.) differences when compared to the controls. These results suggest that the Val-66-Met polymorphism of the BDNF gene may be associated to BD in our Brazilian sample.

P13.24 IDENTIFICATION OF VARIANTS IN 15Q13-Q14 PREDISPOSING TO SCHIZOPHRENIA AND BIPOLAR AFFECTIVE DISORDER Flomen RH, Collier D, and Makoff AJ Institute of Psychiatry, London, United Kingdom The alpha-7 nicotinic cholinergic receptor subunit (CHRNA7) gene is widely expressed in the central nervous system and maps to 15q13-q14, which has been implicated in several neuropsychiatric disorders, including schizophrenia and bipolar disorder. This region contains segmental duplications, including two 300 kb duplicons which contain exons 5–10 of CHRNA7, as either part of the complete CHRNA7 gene or as part of a hybrid gene, CHRFAM7A. The presence of CHRFAM7A is itself polymorphic and it additionally carries a common 2bp deletion polymorphism within exon 6. The existence of these variants in the human population may be relevant to the involvement of 15q13-q14 in psychiatric disorders. In order to investigate this hypothesis, we have studied the frequency of these polymorphisms in individuals with either schizophrenia or bipolar disorder and from the general population. Two approaches were used: Taqman quantitative PCR to determine CHRFAM7A copy number and fluorescent PCR to characterise the 2 bp deletion polymorphism by peak height ratio. These methods were first validated in two families previously characterised for copy number of CHRFAM7A and in DNA from unaffected individuals, before extending the study to greater numbers of cases and controls.

P13.25 DIFFERENTIALLY SKEWED X INACTIVATION AS A POSSIBLE SOURCE OF THE GENETIC CONTRIBUTION TO DISCORDANCE TO PSYCHOSIS IN MONOZYGOTIC FEMALE TWINS Rosa A, Picchioni M, Toulopoulou T, Loat C, Murray RM, and Craig IW SGDP. Institute of Psychiatry, London, United Kingdom Monozygotic twins discordant for psychiatric disorders have provided an intriguing paradigm for investigating the aetiology of these complex conditions and are thought to frequently reflect the influence of environmental factors. That some monozygotic female twin pairs are discordant for X linked disorders believed to result from differentially skewed X-inactivation, such as for Duchenne muscular dystrophy and haemophilia, however, raises the possibility that female twins discordant for behavioural disorders may reflect the existence of Xlinked quantitative trait loci (QTLs), contributing to the phenotypes. We tested this hypothesis by investigating concordant and discordant schizophrenic female twin pairs from the Maudsley Hospital Twin studyof psychosis. To this end, we employed an inactivation assay based on the discrimination between active and inactive X chromosomes as reflected by their methylation patterns at the human

Abstracts androgen receptor locus (Allen et al., 1992. J. Med. Genetic, 51:1229– 39). To date, we have examined mouth swab and/or blood samples from four twin pairs discordant and five pairs concordant for schizophrenia and compared their inactivation skewing with a series of 21 healthy controls pairs. The results obtained so far although generally reflecting a consistency in skewing estimates obtained with blood and with buccal swabs, do not suggest a significant contribution from X–linked loci to the differences in behaviour observed for the discordant twin pairs. We are, however, expanding the number of twin pairs investigated, which will enable a more robust investigation into the hypothesis of discordant skewing in X chromosome inactivation contributing to differences in female twin pair phenotype. This work was supported by the Wellcome Trust and the Stanley Medical Research Institute. We also thank AGAUR (Generalitat de Catalunya) and Fundacio´ Seny (Barcelona) for post-doctoral fellowship to A. Rosa. P13.26 EVIDENCE FOR LINKAGE (LOD ¼ 3.54) AT 1Q42 CLOSE TO DISC1 IN A GENOME SCAN OF FUNCTIONAL PSYCHOSIS PEDIGREES INCLUDING AT LEAST ONE MEMBER WITH SCHIZOAFFECTIVE DISORDER OF BIPOLAR TYPE Hamshere ML,1 Williams NM,1 Segurado R,2 Lambert D,2 Jones LA,3 Holmans P,1 Kirov G,1 Corvin A,2 Jones I,1 O’Donovan MC,1 Gill M,2 Owen MJ,1 and Craddock N1 1 Department of Psychological Medicine, University of Wales College of, Cardiff, United Kingdom 2 Departments of Psychiatry and Genetics, Trinity College, Dublin, United Kingdom 3 Division of Neuroscience, University of Birmingham, Birmingham, United Kingdom Traditionally the search for genes involved in predisposition to functional psychoses has proceeded with implicit acceptance of the Kraeplinian dichotomy with separate studies of schizophrenia and Bipolar phenotypes. However, twin data suggest that, in addition to genes specific for the Kraeplinian extremes, at least some genes contribute susceptibility across the Kraeplinian divide—i.e. genes that simultaneously influence susceptibility to Schizophrenia, Schizoaffective disorder and Bipolar disorder. We have undertaken a search by systematic genome scan for such loci. The sample comprises Schizophrenia and Bipolar affected sibling pair (ASP) series collected in the UK and Ireland that we have previously evaluated in systematic genome screens for either schizophrenia or Bipolar Disorder. Pedigrees were selected for the current analysis where there was at least one member in the pedigree with DSMIV Schizoaffective bipolar type (SABP). Within these pedigrees, individuals were coded as affected if they had been diagnosed with DSMIV Schizophrenia, SABP or Bipolar I Disorder. A total of 24 pedigrees contribute 35 ASPs to the sample. A genome scan was performed with MAPMAKER/SIBS, using sample specific allele frequencies estimated from the original Schizophrenia and Bipolar samples with SPLINK. A maximum LOD score of 3.54 was observed on chromosome 1q42 at 257.22 cM from p-tel (5.1 cM distal to D1S2800). At this peak the expected proportion of alleles shared IBD are 0.78, 0.81, and 0.77 in the SABP, Schizophrenia and Bipolar samples, respectively. Linkage signals were also observed in our SABP sample on chromosomes 19 (LOD ¼ 1.85) and 22p (LOD ¼ 1.96). The DISC1 candidate gene lies within 10 cM of our maximum peak. Our linkage evidence near DISC1 is an independent replication of other linkage findings in Scottish and Finnish samples. No linkage was observed to this region in our original Schizophrenia or Bipolar scans. Our findings suggest that the DISC1 locus may influence susceptibility across the functional psychosis spectrum and that evidence to support the candidacy of DISC1 should be sought in association samples comprising the Schizophrenia, Schizoaffective disorder and Bipolar disorder phenotypes. P13.27 SEROTONIN TRANSPORTER (5-HTT) POLYMORPHISM ASSOCIATED WITH EMOTIONAL RESPONSE TO AUDITORY HALLUCINATIONS IN PSYCHOTIC PATIENTS Rivero O,1 Na´jera C,1 Molto´ MD,1 Gonza´lez JC,2 Aguilar E,2 de Frutos R,1 and Sanjua´n J2 1 Departamento de Gene´tica. Facultad de Biologı´a. Universitat de Vale`ncia, Valencia, Spain 2 Departamento de Psiquiatrı´a. Facultad de Medicina. Universitat de Vale`ncia., Valencia, Spain

127

The serotonin transporter (5-HTT) is implicated in the recovering of the serotonin (5-HT) from the synaptic cleft to the presynaptic neuron, regulating the length of the serotoninergic response in the nervous system. Thus, this gene appears to be an interesting study target with influence in different psychiatric disorders. In humans, transcriptional activity of the 5-HTT gene is modulated by a polymorphic repetitive element (5-HTTLPR), with two principal alleles, long (l) and short (s). The l allele has greater transcriptional activity. Malhotra et al. (1998) found a specific significant increase in intensity of auditory hallucinations (AH) in schizophrenic patients with the ll genotype providing preliminary evidence for a role of serotonin in the physiopathology of hallucinations and Pae et al. (2003) found significant differences in the scores of positive and negative symptoms in schizophrenic patients according to 5-HTTLPR genotypes and alleles. In our association study, 200 schizophrenic patients with a clinical history of AH evaluated by means of the PSYRATS scale, and 136 blood donors as controls were investigated. A significant excess of the s allele in the patients was observed. Moreover, a highly significant correlation between the s allele and the emotional response to auditory hallucinations was also observed. These results are coherent with previous reports of significant associations between the allele s of this polymorphism and anxiety-related traits. P13.28 IDENTIFICATION OF AN OVERLAPPING SUSCEPTIBILITY REGION FOR SCHIZOPHRENIA AND BIPOLAR AFFECTIVE DISORDER ON 22Q13.3 IN THE INDIAN POPULATION Sanjev J National Institute of Mental Health and Neuorsciences, Bangalore, Karnatka, India We have identified an overlapping susceptibility region for schizophrenia (SCZ) and bipolar affective disorder (BPAD) on 22q13.3 in the Indian population. A dinucleotide repeat marker, D22S1161, displayed significant association of allele 2 with SCZ (P ¼ 0.0004) as well as BPAD (P ¼ 0.0009) and also excess transmission (P < 105) of the same allele in both the disorders. Sequencing of a non-selective cation channel gene, MLC1, present 2 Mb downstream of this marker in 216 BPAD, 205 SCZ and 116 controls revealed two novel missense mutations: Leu308Gln in one BPAD familial sample and Arg328His in two BPAD sporadic cases, and 31 other polymorphisms spanning nearly the entire gene. Further we observed two LD blocks: one encompassing 7 kb defined by 11 SNPs along with a 33 bp insertion-deletion polymorphism (block1) associated with BPAD familial (P ¼ 0.0052), and another comprising two SNPs in intron7 (block2) that displayed modest association with SCZ sporadic cases (P ¼ 0.0485). Haplotype studies with eight SNPs spanning nearly the entire gene using EHþ program revealed five major haplotypes with one of the eight-SNP haplotype (HAP4) and another seven-SNP subset haplotype (HAP30 ) overrepresented in BPAD familial (P ¼ 0.0019) and SCZ sporadic (P ¼ 0.0251) cases respectively. Further haplotype analysis using 3, 4 and 5-SNP combinations derived from the same 8-SNP set displayed significant association for most of the subset haplotypes of HAP4 with BPAD familial and a few derived from HAP3 with SCZ sporadic samples. This was further substantiated by the excess transmission of subsets derived from the same extended haplotypes, HAP4 and HAP3, in 113 BPAD and 107 SCZ parent-offspring trios respectively. However, most of the haplotypes with excess transmission in SCZ families were anchored to one of the SNPs with significant individual transmission disequilibrium. Analysis of MLC1 missense mutations and exonic SNPs using ESEfinder program indicated the disruption and creation of new SR motifs that could effect exon splicing. Our findings, therefore, provide evidence for the existence of a common susceptibility region for schizophrenia and bipolar affective disorder on 22q13.3 in the Indian population and also suggest the potential role of MLC1 gene in the etiology of these disorders. P14.1 VARIATION OF THE 120BP DUPLICATION IN THE UPSTREAM REGION OF THE DOPAMINE D4 RECEPTOR AND SCHIZOPHRENIA AMONG DANES Olsen L, Werge T, and Rasmussen HB Institute of Biological Psyciatry, Sct. Hans Hospital, Roskilde, Denmark Several lines of evidence suggest that the dopamine D4 receptor is a potential candidate for involvement in schizophrenia. The basis for this

128

Abstracts

idea rests on pharmacological findings that the atypical antipsychotic drug clozapine displays a high affinity for the dopamine D4 receptor as compared to other dopamine receptors. Variations in the coding region of the dopamine D4 receptor gene (DRD4) have been studied in relation to schizophrenia but only week associations have been found. However, a recent study by Xing et al. (2003) reported a significant association between a 120 bp tandem duplication in the upstream region of the DRD4 gene and schizophrenia among Chinese. Since this segment contains putative binding sites for several transcription factors, allelic variation due to duplication of this region could affect expression of the DRD4 gene. Elevated levels of DRD4 mRNA have been reported in the frontal cortex of schizophrenic patients compared to normal control subjects. This has been taken as indication for high expression levels of DRD4 to increase the susceptibility to schizophrenia. Genetic variants may contribute differently to the etiology of schizophrenia in different populations. We therefore examined whether there is a similar association between the 120 bp tandem duplication and schizophrenia in 242 patients with a strict schizophrenic diagnosis (WHO ICD10) and 178 matched control subjects. Only participants with Danish origin (i.e. both their parents were born in Denmark) were included in the study. We detected allele and genotype frequencies in cases and controls and used X2 statistics to detect differences between the two groups. P14.2 ASSOCIATION BETWEEN DISC1 POLYMORPHISMS AND MAJOR MENTAL ILLNESS IN THE SCOTTISH POPULATION Thomson PA,1 Wray NR,1 Millar JK,1 Condie A,1 Walker MT,2 Muir WJ,2 Blackwood DHR,2 and Porteous DJ1 1 University of Edinburgh, Edinburgh, United Kingdom 2 Royal Edinburgh Hospital, Edinburgh, United Kingdom We are studying a large Scottish family, in which a balanced translocation t(1:11)(q42,q14) co-segregates with major psychiatric illness, with a maximum LOD score of 7.1. The translocation directly disrupts two novel genes on chromosome 1: DISC1, a large alternatively spliced gene containing 13 exons, and DISC2 a gene antisense to DISC1. The translocation may also affect the expression of TRAX (Translin-associated factor X) a gene which undergoes intergenic splicing with DISC1. The TRAX/DISC region spans approximately 600 kb. This region has shown independent evidence for linkage to schizophrenia in the Finnish, and Taiwanese populations, as well as to BPAD in the Scottish population. We have constructed a linkage disequilibrium map of the TRAX/DISC region consisting of 39 markers, approximately one marker every 15 kb. The LD map of the TRAX/DISC locus suggests that this region is split into 4 haplotype blocks of between 250 kb and 105 kb, interspersed with regions of low LD. This allowed us to select 17 SNPs that represented the region for use in case-control association studies. Case—control association studies were performed on individuals from the Scottish population with diagnoses of schizophrenia (SCZ) and bipolar affective disorder (BPAD). Two regions gave global P-values < 0.05. The first region is defined by a two SNP haplotype in intron 6 of DISC1, association was detected with a haplotype containing these SNPs in SCZ (global P-value ¼ 0.004), and in BPAD (global P-value ¼ 0.006). The second region showed association between a two SNP haplotype and BPAD only (global P-value ¼ 0.023). This region extends from approximately 10 kb 50 to DISC1 exon 1 into intron1. Only the SNP 50 of DISC1 gave a nominally significant single marker association (global P-value ¼ 0.03) again with BPADs only. No association was detected using the SNPs within TRAX with either SCZ or BPAD, with sexes either combined or analysed independently. Eleven additional markers have been added to the TRAX/DISC LD map, allowing us to define the LD regions involved in our associations. We have also replicated 15 SNPs identified in previous studies in other laboratories within our samples sets. This should aid the comparison of haplotypes between populations.

P14.3 CHARACTERIZATION OF THE DISC1-NUDEL/ ENDOOLIGOPEPTIDASE A PROTEIN COMPLEX Hayashi MAF,1 Portaro FCV,1 Bastos MF,1 Guerreiro JR,1 Camargo ACM,1 Camargo M,2 and Brandon NJ2 1 Butantan Institute, Sa˜o Paulo, Brazil 2 Merck Sharp and Dohme, Harlow, United Kingdom Disrupted in Schizophrenia 1 (DISC1) has recently suggested to be a schizophrenia risk factor. We have shown that DISC1 interacts with

the molecule NUDEL, a molecule which has been shown to be essential for brain development and cortical organization. DISC1 and Nudel interact in a neurodevelopmentally regulated manner involving discrete protein motifs on each protein. Interestingly, NUDEL is the same protein as the endooligopeptidase A (EOPA), a thiol-activated peptidase involved in the conversion and the inactivation of a number of bioactive peptides. We have characterized the activity of EOPA in the rat brain and determined the functional significance of the interaction with DISC1. In this study rat brain cytosol was fractionated by gelfiltration chromatography and the proteolytic activity of NUDEL/ EOPA was assayed fluorimetrically. We have been able to show that the monomeric form of NUDEL/EOPA, which represents less than 2% of the immunoreactive EOPA/NUDEL in the cytosol of rat brain, is responsible for most of its oligopeptidase activity. The majority of NUDEL is found in high molecular weight protein complexes, including DISC1 complexes, which show weak enzymatic activity. We have analysed the effect of DISC1 on the activity of EOPA in purified protein assays and transient-transfection cell based assays and have been able to show that DISC1 competitively inhibits NUDEL/EOPA activity. Furthermore we have undertaken various proteomic approaches (yeast 2 hybrid and mass spec based analyses) to identify other molecules which interact with NUDEL/EOPA to increase our understanding of its regulation and function. The significance of the DISC1-NUDEL/EOPA activity for schizophrenia and other psychiatric illnesses will be discussed. P14.4 APOD POLYMORPHISMS AS A POSSIBLE DIAGNOSTIC MARKERS FOR SCHIZOPHRENIC DISORDERS Hansen T and Werge T Sct. Hans Hospital, Research Institute of Biological Psychiatry, Roskilde, Denmark Schizophrenia is a serve severe mental disorder that affects approximately 1% of the world population and is estimated to be the 7th leading non-fatal burden to the world population (WHO). It mainly emerges in the adolescence or early adulthood and several studies indicate that schizophrenia has a significant genetic component. ApoD is a protein of the lipocalin superfamily and is primarily observed in human plasma high-density lipoprotein particles. It is also widely expressed in various tissues including the brain. It has a putative role in neuroregenerative and neurodegenerative process, and in the lipid metabolism. There are strong indications that Apolipoprotein D (ApoD) is involved in arachidonic acid signalling and its metabolism, which is seen deficient, in patients with a schizophreniac diagnosis. Furthermore, elevated levels of ApoD are foundhaves been observed in the plasma and in the brain of never-medicateddrug-naive and clozapine-treated schizophrenic patients. This raises the possibility suggests that ApoD polymorphisms are may acts as a disease markers for schizophrenia or may are is directly be involved in the aetiology of schizophrenia. Only few studies have compared polymorphism in apolipoproteins with schizophrenia, even though these proteins play an important role in neuropsychiatric disorders. ApoD is a protein of the lipocalin superfamily and is primarily observed in human plasma high-density lipoprotein particles, but it is also widely expressed in various tissues including the brain. It has a putative role in neuroregenerative and neurodegenerative process, together with lipid metabolism. Only few studies have compared polymorphism in apolipoprotein with the schizophrenic diagnosis, even though they play an important role in neuropsychiatric disorders. This study resolves the geno- and haplotype of 3 single nucleotide polymorphism in ApoD and the potential association to schizophrenia. A cohort of 353 patients with schizophrenia (WHO IDC10) and 353 controls, is used in a case-control study. All samples are of Caucasian origin from the Danish Psychiatric Biobank. Here we investigate if genetic variations of 3 single nucleotide polymorphism in ApoD are associated with a schizophrenic diagnose (WHO ICD 10). In a case-control study of 389 patients, from the Danish Psychiatric Biobank, we determine the distribution of the 3 polymorphisms.

P14.5 ASSOCIATION ANALYSIS OF NINE CANDIDATE GENES OF SCHIZOPHRENIA OF RGS4, MRDS1, DTNBP1, TNF-A, NOTCH4, PPP3CC, NRG1, DAAO, AND G72 IN TAIWANESE FAMILY SAMPLE Hwu HG,1 Liu CM,1 Fann CSJ,2 Liu YL,3 Lin CY,2 Chen WJ,4 Chuang MJ,5 Faraone SV,6 Ou-Yang WC,7 Jann HY,8 Chen JJ,8 and Wu JY9

Abstracts 1

Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan Institute of Biomedical Science, Academia Sinica, Taipei, Taiwan 3 Department of Medical Research, National Taiwan University Hospita, Taipei, Taiwan 4 Institute of Epidemiology, College of Public Health, National Taiwan University, Taipei, Taiwan 5 Institute of Behavioral Genomics, Department of Psychiatry, University of California at San Diego, San Diego, United States 6 Harvard Institute of Epidemiology and Genetics, and Department of Epidemiology, Harvard School of Public Health, and Department of Psychiatry, Harvard Medical School, Boston, United States 7 Institute of Public Health, National Yang-Ming University, Chia-Nan Psychiatric Center, Department of Health, Tainan, Taiwan 8 Taoyuan Psychiatric Center, Taoyuan, Taiwan 9 National Genotyping Center, Academia Sinica, Taipei, Taiwan 2

Several potential candidate genes were found to be associated with schizophrenia in recent years. Some replication studies confirmed the association evidence, however, some did not. We conducted an association analysis of nine candidate genes of schizophrenia, RGS4, MRDS1, DTNBP1, TNF-a, NOTCH4, PPP3CC, NRG1, DAAO, and G72, in a total of 218 co-affected sib-pair families of schizophrenia in Taiwan. The family samples of the Han Chinese ethnic background came from two sampling collection projects, MPSS and TSLS. 86 families from MPSS and 132 from TSLS were genotyped. The sampling strategy of the two projects is similar and both assumed DSM-IV criteria as the phenotype definition, but the two sub- samples came from different districts of Taiwan. A two-stage SNP mapping was performed. In the first stage of SNPs validation, we selected 125 SNPs located in the nine genes from previous reports and public SNP databank. The SNP genotyping was done, initially, in a small testingsample. 68 SNPs were validated with the criteria of minor allele frequency above 0.1 and missing rate below 10%. We genotyped the 68 SNPs in a total of 218 families, and conducted single locus and haplotype association analysis using Transmit v2.5.4 program. In single locus association analysis, we found significant association evidence of the SNPs of MRDS1 and G72 in the MPSS sub-sample, of MRDS1, NOTCH4, PPP3CC, NRG1 in the TSLS sub-sample, and of NOTCH4, PPP3CC, NRG1 when pooling together. Using the haplotype association analysis, we found significant haplotype transmission distortion of the SNP haplotype of RGS4, MRDS1, TNF-a, G72 in the MPSS sub-sample, of MRDS1, NOTCH4, NRG1, and DAAO in the TSLS sub-sample, and of NOTCH4, NRG1, and DAAO when pooling together. Our results revealed association evidence of NOTCH4, NRG1, DAAO, MRDS1and G72 in Taiwanese family sample. The different pattern of association in these two sub-samples remained to be discussed.

P14.6 ANALYSIS OF IL-1ALPHA, IL-1BETA AND IL-RA POLYMORPHISMS IN SCHIZOPHRENIA Muhlenhoff L,1 Giegling I,1 Gonnermann C,1 Dahmen N,2 Thierfelder K,1 Mu¨ller J,3 Bettecken T,3 Meitinger T,3 Mo¨ller HJ,1 and Rujescu D1 1 Division of Molecular and Clinical Neurobiology, Department of Psychiatry, University of Munich, Munich, Germany 2 Department of Psychiatry, University of Mainz, Mainz, Germany 3 Institute of Human Genetics, GSF, Neuherberg, Germany Previous studies have suggested that dysfunctions in cytokines could be involved in pathophysiology of schizophrenia. Independent studies reported that functional polymorphisms in some cytokine genes may have regulatory effects on the cytokine system. IL-1alpha, IL-1 beta and IL-1RA are not only functional but also positional candidate genes due to their chromosomal location on chromosome 2q14, a region with highly positive linkage to schizophrenia. We therefore examined the possible role of the IL-1alpha C-889T, the IL-1beta C-511T and the IL1RA (86bp)(n) repeats polymorphisms in 355 Caucasian schizophrenic patients and 506 unrelated healthy volunteers. The IL-1alpha C-889T and IL-1beta C-511T single nucleotide polymorphisms (SNPs) were not associated with schizophrenia. Our IL-1RA data showed a trend towards a protective effect of allele (86bp)(2) against schizophrenia (w2 ¼ 2.841, OR ¼ 1.154 (0.976–1.364), df ¼ 1, P ¼ 0.08). This result prompted us to evaluate additional SNPs covering the entire IL-1RA gene in another independent sample of 368 German schizophrenic patients and 368 German controls. The results on allele, genotype and

129

haplotype frequencies in this case control association study will be presented. P14.7 HUMAN 5HT2C RECEPTOR: RNA EDITING AND ALTERNATIVE SPLICING IN SCHIZOPHRENIA VS CONTROL POST-MORTEM BRAIN Flomen RH, Papapetrou D, Ly L, Kerwin R, and Makoff AJ Institute of Psychiatry, London, United Kingdom In the precursor RNA of the 5HT2C receptor gene (HTR2C), a stable stem-loop structure is predicted between exon 5 and the 50 end of intron 5. This stem-loop is flanked by two alternative splice sites (GU1 and GU3) and contains the preferred donor splice site (GU2). The stem also includes five exonic adenosines (A-E) and a sixth intronic adenosine (F). All six adenosines may undergo RNA editing, modifying the amino acid sequence. We have recently shown that RNA editing also modulates donor splice site selection.We describe an extensive analysis of RNA editing at these six sites and of GU1/GU2 alternative splicing in choroid plexus, thalamus and prefrontal cortex from 5 schizophrenic patients compared to five control individuals. Chi squared analyses have been performed to compare relative frequencies of editing isoforms present in the control and schizophrenic groups. Relative splicing at the preferred (GU2) and upstream (GU1) alternative donor splice sites has also been quantified in schizophrenia as compared to control samples. Finally, we have studied regional differences in editing and alternative splicing in controls and schizophrenic individuals. The results of this study are presented and their implications discussed. P14.8 FINE MAPPING OF THE CHROMOSOME 1Q23-24.2 SCHIZOPHRENIA SUSCEPTIBILITY REGION NEAR THE RGS4 D1S196 AND CAPON LOCI Puri V,1 Rizig MA,1 Datta SR,1 McQuillin A,1 Thirumalai S,3 Pimm J,1 Moorey H,1 Quested D,4 Kalsi G,1 Bass N,1 Lawrence J,1 Curtis D,2 and Gurling HMD1 1 University College London, London, United Kingdom 2 St Bartholomew’s and Royal London School of Medicine and Dentistry, London, United Kingdom 3 Berkshire Mental Health Trust, Prospect Park Hospital, Reading, United Kingdom 4 Warneford Hospital, University of Oxford, Oxford, United Kingdom Linkage to schizophrenia on chromosome 1q21-24.2 has been confirmed by several family linkage studies. Subsequently an allelic and haplotypic association study of schizophrenia detected significant evidence of association near the candidate gene RGS4 (Chowdari et al 2002). Akabame et al (2003) reported significant allelic association to schizophrenia on chromosome 1 with the marker D1S196 (P ¼ 0.0009). A third locus near the carboxyl-terminal PDZ ligand of neuronal nitric oxide synthase gene (CAPON) also shown apparent haplotype and allelic association with schizophrenia using a family based design (Brzustowicz et al 2004). The distance between the RGS4 and CAPON genes is approximately 700 Kb, making it unlikely that there is a single schizophrenia susceptibility locus can explain both associations. However the region between RGS4 and CAPON includes several other potential candidate genes. D1S196 is 4 megabases distal to RGS4. We have carried out a replication of the D1S196, RGS4 and CAPON associations in ancestrally matched cases of schizophrenia and supernormal controls from the South of England. We genotyped the microsatellite marker D1S196 but could not replicate the association found by Akabame in Japan. We are studying the markers D1S1677, D1S2675, D1S377, D1S2768, D1S2434, D1S2420, D1S2100, D1S2133 and D1S2844 that span the region between CAPON and RGS4. SNPs at the RGS4 and CAPON locus have also been genotyped. We found strong evidence for haplotypic association at the RGS4 locus but not at D1S196. Further data on the putative CAPON gene association will also be presented.

P14.9 A META-ANALYSIS OF ASSOCIATION BETWEEN THE SER9GLY POLYMORPHISM IN THE DRD3 GENE AND SCHIZOPHRENIA Shahsavand-Ananloo E, Breen G, Collier D, and Sham P Institute of Psychiatry, London, United Kingdom Schizophrenia is a major psychiatric illness that has evidence for a significant genetic contribution toward its development. In recent

130

Abstracts

years, the Ser9Gly in the dopamine D3 receptor gene has been examined as a possible susceptibility factor for schizophrenia. To further evaluate the controversial putative association between a Ser9Gly variant in the first exon of the dopamine D3 receptor gene (DRD3) and schizophrenia, we performed a meta-analysis from 53 independent samples, from case-control association studies, allowing the examination of the combined sample of 10,188 individuals (5,028 schizophrenic patients and 5,160 controls). The association studies were between a biallelic, Ser9Gly polymorphism in exon 1 of the dopamine D3 receptor gene and schizophrenia. No significant allelic association was observed (P ¼ 0.87, odds ratio (OR) ¼ 1.0047, 95% Confidence Interval (CI) ¼ 0.94–1.06). No significant differences in genotype frequency were detected using either dominant (P ¼ 0.52, OR ¼ 0.95, 95% CI ¼ 0.84– 1.09) or recessive (P ¼ 0.56, OR ¼ 1.02, 95% CI ¼ 0.94–1.10) between samples respectively. No significant difference in heterogeneity was detected between samples (P ¼ 0.36, OR ¼ 0.96, 95% CI ¼ 0.88–1.04). The results of the meta-analysis provide independent support for a relationship between schizophrenia and homozygosity at the Ser9Gly polymorphism of the D3 receptor gene. P14.10 ASSOCIATION ANALYSIS IN A LARGE IRISH CASE-CONTROL SAMPLE OF HSPA8 AND GSTM3, TWO GENES PUTATIVELY INVOLVED IN SCHIZOPHRENIA McGhee KA,1 Morris DW,1 Schwaiger S,1 Nangle J,1 Murphy K,1 Donohoe G,1 Clarke S,1 Baldwin P,2 Scully P,2 Quinn J,2 Meagher D,2 Waddington JL,3 Gill M,1 and Corvin AP1 1 Neuropsychiatric Genetics Group, Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland 2 Stanley Research Unit, St. Davnet’s Hospital, Monaghan, Ireland 3 Department of Clinical Pharmacology, Royal College of Surgeons, 123 St. Stephen’s Green, Dublin, Ireland Prabakaran et al. 2004 combined transcriptomic, proteomic and metabolomic approaches to study post mortem human brain tissue in order to explore the molecular disease pathways of schizophrenia (SZ). Almost half the proteins identified as either over-expressed or underexpressed in SZ brains were associated with mitochondrial function and oxidative stress responses. This finding was supported by transcriptional and metabolite perturbations. These results suggest that oxidative stress and the ensuing cellular adaptations are linked to the SZ disease process. Therefore genes encoding proteins involved in these pathways are putative candidate genes for SZ. In an initial screen, we compared 26 of the differentially expressed proteins identified by Prabakaran et al. with their corresponding chromosomal gene positions. Four oxidative protein genes are positioned within significant bin regions identified by the metanalysis of SZ genome scans (Lewis et al 2003): HSPA1A, HSPA1L, HSPA8, and GSTM3. Subsequent bioinformatics identified that pseudogenes existed for two of the genes: HSPA1A and HSPA1L. We subjected the two remaining genes, HSPA8 AND GSTM3, to candidate gene association analysis. Novel SNPs were identified by direct sequencing of 15 case samples. These data supplemented existing data on the dbSNP database (NCBI). Haplotypic/LD structure across the two gene determined using a sample of 100 cases. Tagging SNPs were identified and analysed in our full association sample (cases n ¼ 299, controls n ¼ 645). Results of the study will be presented. P14.11 PERSONALITY, SOCIAL FUNCTIONING, AND PSYCHOPATHOLOGY OF ADOLESCENT FIRST-DEGREE RELATIVES OF INDIVIDUALS WITH SCHIZOPHRENIA Faraone SV,1 Seidman LJ,2 Glatt SJ,3 Stone WS,2 Gabel LB,2 and Tsuang MT3 1 Department of Epidemiology, Harvard School of Public Health, Boston, MA, United States 2 Department of Psychiatry, Harvard Medical School, Boston, MA, United States 3 Department of Psychiatry, University of California, San Diego, La Jolla, CA, United States Schizotaxia is associated with negative symptoms, neuropsychological dysfunction, and structural and functional brain abnormalities in 20– 50% of non-psychotic, adult relatives of patients with schizophrenia. Our ongoing work is focused on understanding the divergent paths that are associated with this condition (e.g., stable deficit, descent into prodrome and psychosis, or transient deficit) in order to refine early intervention and prevention efforts. The prevalence and nature of

schizotaxic deficits among adolescent first-degree relatives of patients with schizophrenia is not well characterized. Thus, we are conducting a longitudinal study of adolescent children and siblings of schizophrenic patients and controls. Our ongoing characterization of this sample involves comprehensive analysis at multiple levels, including physical development, social functioning, personality, psychopathology, substance use, neuropsychology, and brain structure and function. Assessments of social functioning, personality, and psychopathology use a wide range of instruments, including the Temperament and Character Inventory, the Youth Self-Report, a Life Event Scale, a Family Environment Scale, the Symptom Checklist-90, a Puberty and Development Questionnaire, the Perceptual Aberration Scale, the Magical Ideation Scale, the Revised Physical Anhedonia Scale, the Social Adjustment Inventory for Children and Adolescents, and the Epidemiologic Version of the Schedule for Affective Disorders and Schizophrenia for School-Age Children. These assessments are now complete for 40 children of patients and 52 children of controls. The cross-sectional data support the presence of hypothesized deficits in non-psychotic adolescents at high risk for schizophrenia prior to the potential onset of illness. For example, compared to children of controls, children of patients exhibit more delinquency, externalizing behavior, and physical anhedonia, and less social competence. In conclusion, adolescent children and siblings of patients with schizophrenia exhibit social dysfunction comparable to those of adult relatives with schizotaxia. These children are at high genetic and clinical risk for the subsequent development of schizophrenia, and must be monitored carefully. Longitudinal tracking of these high-risk individuals will allow us to determine which measures are the most reliable markers of the genetic liability toward schizophrenia and the best predictors of subsequent transition to full illness.

P14.12 PARENTAL HOMOZYGOSITY MASKS THE LINKAGE SIGNAL OF THE REELIN REGION IN SCHIZOPHRENIA FAMILIES Ekelund J, Hennah W, Turunen JA, Lo¨nnqvist J, and Peltonen L 1 National Public Health Institute, Helsinki, Finland 2 Columbia University, New York, United States Genetic mapping of susceptibility genes for common, complex disorders has been notoriously difficult. According to the common-variant/ common-disease hypothesis, frequent polymorphisms would underlie such disorders. Traditional ascertainment strategies for linkage studies are often designed to collect as large and disease-loaded pedigrees as possible. This strategy also maximizes parental homozygosity for any common susceptibility allele, which could lead to a decrease in statistical power for such variants. We have here analyzed the genome-wide linkage information separately in families with ‘‘low’’ and ‘‘high’’ loading of schizophrenia (number of affected children versus non-affected children). We present evidence that markers in the Reelin gene on chromosome 7q22 show strong linkage (lod ¼ 5.29) to schizophrenia that would have been missed in the total data set of 312 families, not ascertained for genetic loading. We suggest that this strategy of ‘‘ascertainment by genetic loading’’ should be more efficiently utilized in common disorders for which the susceptibility variants can be expected to show high prevalence in the population. We also performed simulations of linkage analysis for genotype data adapting a disease model with low penetrance and high allele frequency. These simulations showed that a sample of 200 families with 2 affected children was clearly more powerful than 200 families with 6 affected children, in contrast to general assumptions (100 replicates, mean lod 7.08, SD 3.41, versus mean lod 0.91, SD 0.49). In addition, we developed a simulation program that takes the entire disease model as input. The program simulates a population of specified size to estimate sibling recurrence risks and also draws a sample for genotyping. There is epidemiologic data available for the recurrence risks of siblings of schizophrenic subjects in Finland, so any potential model must be in line with that data to be plausible for schizophrenia. Disease models with high gene frequencies best fitted our data. Our simulations taken together with the novel analysis of the actual data show that for common susceptibility variants, traditional ascertainment strategies can be inadequate. Specifically, there is now evidence that the Reelin gene on chromosome 7q22 is a susceptibility variant with high allele frequency and low penetrance. This should stimulate other research groups to re-analyze their samples using this novel approach, potentially leading to the identification of other common genes for common disorders.

Abstracts P14.13 USE OF THE DIAGNOSTIC INTERVIEW FOR GENETIC STUDIES (DIGS) IN AN INTERNATIONAL GENETICS STUDY OF LATINO SUBJECTS FROM THE US, MEXICO AND COSTA RICA: TRAINING AND RELIABILITY ISSUES Salazar R,1 Escamilla MA2, Mendoza R,3 Esparza E,4 Montero P,1 Ontiveros A,5 Nicolini H,6 Medina R,2 Saucedo E,5 Herrera C,7 Lizano A,1 Contreras S,2 and Dassori A2 1 Universidad de Costa Rica, San Jose, Costa Rica 2 The University of Texas Health Science Center at San Antonio, San Antonio, United States 3 Harbor-University of California at Los Angeles Research and Education Institute, Los Angeles, United States 4 Logan Heights Family Clinic, San Diego, United States 5 Instituto de Informacion e Investigacion en Saludad Mental, Monterrey, Mexico 6 Grupos de Estudios Medicos y Familiares Carracci, Mexico City, Mexico 7 Centro de Salud Mental del Estado de Baja California, Mexicali, Mexico The DIGS was developed for the National Institute of Mental Health (NIMH) Genetics Initiative, and was designed to ensure reliable diagnoses of major mental illnesses. An initial study conducted in Enlish-speaking subjects from the United States suggested that the DIGS is reliable for Schizophrenia, Bipolar disorder and Major Depression and less so for Schizoaffective Disorder (Nuremberg et al., 1994). We have developed a Spanish version of the DIGS for use in Mexico, the USA, and Central America, we tested for reliability using DIGS (English and Spanish versions, as appropriate per patient choice) at three different training sessions (San Antonio, Monterey Mexico, and San Diego) involving a total of 12 psychiatrists from the United States, 22 psychiatrists from Mexico, and 5 psychiatrists from Costa Rica. We tested for accuracy of diagnosis (percentage of agreement between the expert and the trainees) between psychiatrists. For each training session, several patients with different clinical diagnoses were interviewed by small groups of psychiatrists. Within each group, a patient was interviewed by one psychiatrist using the DIGS, with the others observing and performing ratings at the same time. At the end of the interview, each rater was asked to fill out sheets coding for overall DSM IIIR diagnosis, as well as for whether symptoms and syndromes of depression, mania or psychosis had been present during the course of the subject’s illness. Combining data across all of the training sessions, we observed high rates of agreements for all of the above items (Depression Syndrome (DSMIIIR) 94.14%, Mania Syndrome (DSMIIIR) 90.34%, Psychosis 91.65%, Bipolar Disorder 91.40%, Schizophrenia 88.20%, Schizoaffective Disorder 87.01%). There were no biases related to country of origin of the interviewers. Each group had one member who was a researcher with extensive experience in using the DIGS for clinical diagnoses. We found a high percentage of agreement with the ‘‘expert’’ rater for diagnoses and for lifetime prevalence of mood syndromes and psychosis using DSMIIR, for each case. We conclude that the DIGS in Spanish is reliable when used to diagnose Latino subjects with Bipolar Disorder, Schizophrenia, Schizoaffective Disorder, or history of psychosis, depression and mania. Psychiatrists from Central America, Mexico, and the United States have high rates of agreement when the DIGS is used to gather information and they are asked to arrive at diagnoses. Finally, the DIGS (English and Spanish versions) can be taught over a brief (two day) period and reliably used by clinically experienced psychiatrists from the US, Mexico and Central America. P14.14 A SCREEN OF CANDIDATE GENES (NEUREGULIN, CALCINEURIN, AKT1) WITH SNPS FOR ASSOCIATION/ LINKAGE DISEQUILIBRIUM WITH SCHIZOPHRENIA USING TWO INDEPENDENT FAMILY SAMPLES Hoefgen B,1 Schwab SG,2 Borrmann Hassenbach M,3 Albus M,3 Trixler M,4 Lerer B,5 Hallmayer J,6 Maier W,1 and Wildenauer DB2 1 University of Bonn, Bonn, Germany 2 University of Western Australia, Perth, Australia 3 Mental State Hospital, Haar, Germany 4 University of Pecs, Pecs, Hungary 5 Haddassah Medical School, Jerusalem, Israel 6 Stanford University, Stanford, United States Availability of the genomic sequence as well as SNP databases facilitate the design of maps with tightly spaced SNPs for screening candidate

131

genes for linkage disequilibrium with schizophrenia. We used 13 SNPs for investigation of Neuregulin (covering 160 kb of the genomic sequence), 9 SNPs for Calcineurin (92 kb of the genomic sequence) and 5 markers for AKT1 as described recently by Emamian et al. (Nat. Genet. 2004, 36:131). Allelic discrimination assays using Taqman probes designed by ABI were used for genotyping. Two independent family samples were available: 79 sib-pair families from Germany and Israel, and 125 trios from Germany and Hungary. TDT analysis was performed for single markers and haplotypes. Results for markers within the genomic region of Neuregulin and Calcineurin did not reach statistical significance after adjusting for multiple comparisons. Evidence for association has been obtained for SNPs located in the genomic region of the gene for AKT1. SNP2 and SNP3 (Emamian et al) as well as the three marker haplotype 2-3-4 revealed statistically significant P-values in the sib-pair sample, while marker 4 showed statistically significant biased transmission in the trio sample. Our results add further evidence for involvement of AKT1 in development of schizophrenia. P14.15 NO ASSOCIATION BETWEEN POLYMORPHISMS OF METHYLENETETRAHYDROFOLATE REDUCTASE GENE AND SCHIZOPHRENIA IN BOTH CHINESE AND SCOTTISH POPULATIONS Yu L2, Li T,5 Robertson Z,1 Dean J,1 Gu NF,6 Feng GY,6 Yates P,1 Sinclair M,1 Crombie C,1 Collier DA,4 Walker N,7 He L,3 and St Clair D1 1 Department of Mental Health and Medical School,University of Aberdeen, Aberdeen, United Kingdom 2 Institute of Nutrition Science, Shanghai Institutes for Biological Sciences,Chinese Academy of Sciences, Shanghai, China 3 Bio-X Life Science Research Center, Shanghai Jiaotong University, Shanghai, China 4 Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom 5 West China Hospital, Sichuan University, Chengdu, China 6 Shanghai Insitute of Mental Health, Shanghai, China 7 Ravenscraig Hospital, Greenock, United Kingdom Evidence from case reports, increased homocysteine levels in patients with schizophrenia and famine studies has indicated that Methylenetetrahydrofolate reductase (EC 1.5.1.20, MTHFR) is possibly associated with schizophrenia. Several studies have examined the common MTHFR variations as a risk genetic factor for the pathogenesis of schizophrenia. However, discrepant results have been observed. In the present study, we investigated the two common variations (C677T and A1298C) in MTHFR in DNA samples from Chinese and Scottish populations. In the Han Chinese population, we used both a casecontrol design and a family trio design. Two hundred and thirty patients with schizophrenia and 251 controls as well as 267 family trios were examined. In the Scottish population, we examined a sample of 426 patients with schizophrenia and 628 controls. There were no significant allele, genotype and haplotype frequency differences between cases and controls in either Chinese or Scottish populations. Transmission ratios were also normal in Chinese schizophrenia trios. Our results suggest that variations in MTHFR gene do not predispose to increased risk of schizophrenia in either Chinese or Scottish populations.

P14.16 ASSESSMENT OF SCHIZOPHRENIA GENOMES FOR MICROINSERTIONS AND MICRODELETIONS Sachs NA,1 Bachani A,1 Cowell JK,2 Conroy J,2 Nowak NJ,2 DeLisi LE,3 Ross CA,1 and Margolis RL1 1 The Johns Hopkins University School of Medicine, Department of Psychiatry, Baltimore, United States 2 Roswell Park Cancer Institute Department of Cancer Genetics, Buffalo, United States 3 New York University Department of Psychiatry, New York, United States Comparative genome hybridization array (CGHa) was used to determine if the genomes of individuals with schizophrenia contain microdeletions and/or microinsertions. Polymerase chain reaction representations of 6,116 bacterial artificial chromosomes (BAC) spanning the human genome at 0.5–1 mb intervals were arrayed in triplicate. Subject genomic DNA was derived from lymphoblastoid cell lines of 27 individuals with schizophrenia and 6 individuals with

132

Abstracts

schizoaffective disorder, all from families with multiple members affected with schizophrenia or other major mental illness originally ascertained as part of a genetic study of sibpairs. Comparison of control and subject genomic DNA hybridization signals revealed 114 abnormally high or low hybridization signals in 28 patients. Two reduced hybridization signals occurred at adjacent BACs located at 5p15.1, in one patient. Four patient genomic DNA samples exhibited increased hybridization at 22q11.21, a previously reported linkage region in schizophrenia. The largest number of altered hybridization signals, detected in ten patients, occurred near the 8p23.1 schizophrenia susceptibility locus. CGHa analysis of genomic DNA isolated from control lymphoblastoid cell lines will be used to determine the baseline rate of abnormal hybridization signals, and fluorescence in situ hybridization will be used to determine which, if any, of the altered hybridization signals observed using CGHa represent actual microdeletions or microinsertions. CGHa represents a novel approach for detecting mutations that contribute to the genetic risk of schizophrenia.

P14.17 VARIATION SCREENING OF A HIGH RISK HAPLOTYPE IN THE DYSTROBREVIN BINDING PROTEIN 1 (DTNBP1) GENE FROM THE IRISH STUDY OF HIGH DENSITY SCHIZOPHRENIA FAMILIES Riley B,1 Thiselton DL,1 Ribble RC,1 Wormley BK,1 Frank G,1 van den Oord EJ,1 Walsh D,2 O’Neill FA,3 and Kendler KS1 1 Virginia Commonwealth University, Richmond, VA, United States 2 Health Research Board, Dublin, Ireland 3 Queen’s University, Belfast, Ireland We have previously shown association of markers in the dystrobrevin binding protein 1 (DTNBP1) gene with schizophrenia, identified a haplotype block of 30.1 Kb covering exons 1 to 5 of this gene, and identified one high risk haplotype (HRH) from the six common haplotype structures within this block in the Irish Study of High Density Schizophrenia Families (ISHDSF). Numerous positive replications using the same marker set have been reported, as well as a number using different marker sets. Sequencing of the 30.1 Kb region in all genetically independent, narrowly-defined ISHDSF cases with the HRH is complete. We identified 60 confirmed SNPs, of which 21 are novel and 39 known, an additional 12 rare SNPs which could not be confirmed by sequence overlap and 12 other polymorphisms (including duplications, INS/DELs and STRs). 7 of the confirmed SNPs are specific to our identified HRH, and 2 others appear to exist on additional haplotype backgrounds as a result of recombination. Bioinformatic analysis does not suggest any functional role for the HRH-specific SNPs and genotyping in our full sample suggests that none of these markers provide any greater evidence of association than those typed previously. However, sequencing of all exons of the gene revealed additional SNPs 34–95 Kb outside the boundaries of the identified block that unambiguously map to our HRH. These data strongly suggest that haplotype blocks are aggregate structures, and that individual haplotypes within them may vary considerably in length as a function of age and frequency. PHASE reconstructions of the sequence data support our original haplotype structures, and identify the positions of historical recombinations. The most common haplotype extends 27.6 Kb with no evidence of historical recombination. By contrast, our HRH extends for 67.8 Kb with no evidence of historical recombination, and 130.1 Kb excluding one rare recombinant with estimated population frequency 0.0027. We have therefore extended our planned sequencing to cover this greatly extended HRH. Sequencing of the whole 130 Kb segment will be complete for presentation. Including current results from this extended region, we have identified a total of 87 SNPs, of which 33 are novel and 54 known, and a total of 13 rare SNPs unconfirmed by sequence overlap. 8 of the confirmed SNPs are specific to our identified HRH, and 7 others appear to exist on additional haplotype backgrounds as a result of recombination. This variation in the length of haplotypes within a block may resolve the differences between haplotypes identified as associated in different samples. Genotyping is underway currently in our recently completed trio and case/control collection, and these data may clarify the overall pattern of risk and the contribution of particular variants to disease liability. P14.18 POSSIBLE ROLE OF DISC1 INTERACTING PROTEINS IN SCHIZOPHRENIA Sachs NA,1 Bachani A,1 Leister F,2 Callahan C,1 Ross CA,1 DeLisi LE,3 Yolken RH,2 Holmes SE,1 and Margolis RL1

1 The Johns Hopkins University School of Medicine, Department of Psychiatry, Baltimore, United States 2 The Johns Hopkins University, Stanley Division of Developmental Neurovirology, Baltimore, United States 3 New York University Department of Psychiatry, New York, United States

A translocation in Disrupted in Schizophrenia 1 (DISC1) may be a major risk factor for schizophrenia in a large Scottish pedigree, and several other genetic studies of schizophrenia have suggested linkage to the DISC 1 locus. The DISC1 protein product appears to be involved in neurodevelopment, and interacts with a series of other proteins that also regulate neurodevelopment. We are investigating this pathway to determine if genes other than DISC1 may play a role in the etiology or pathogenesis of schizophrenia. In our first set of experiments, we are testing the hypothesis that genes encoding DISC1 interacting proteins, including 14-3-3e, NUDEL, and LIS1, may be abnormally regulated in schizophrenia. Real-time PCR was utilized to determine the relative level of expression of these genes in RNA isolated from prefrontal cortex of 35 individuals each with schizophrenia, bipolar disorder, or no psychiatric disorder (Stanley Medical Research Institute array collection). No significant differences were found among the three groups, although a few individuals had expression levels that were markedly divergent from mean levels. In a second set of experiments, we are sequencing this group of genes in 34 individuals with schizophrenia or schizoaffective disorder to search for rare mutations of major effect. Thus far we have identified 2 promoter and 2 non-coding SNPs, along with 1 silent mutation in 14-3-3e, a previously reported G/A SNP (rs1049583) in the 30 untranslated region of 14-3-3Z, 10 allelic variants of Nudel, and 16 allelic variants of LIS1. Analysis is in progress to determine if any of these polymorphisms are limited to individuals with schizophrenia or schizoaffective disorder and segregate with major mental illness in the affected pedigrees.

P14.19 SIGNIFICANT ASSOCIATION EVIDENCE OF POLYMORPHISMS OF PRODH/DGCR6 WITH NEGATIVE SYMPTOMS OF SCHIZOPHRENIA Liu CM,1 Liu YL,2 Lin CY,3 Fann CSJ,3 Hwu HG,1 Jann HY,4 Chen JJ,4 and Ou-Yang WC1 1 Department of Psychiatry, National Taiwan University Hospital and College of Medicine, National Taiwan University, Taipei, Taiwan 2 Department of Medical Research, National Taiwan University Hospital, Taipei, Taiwan 3 Institute of Biomedical Science, Academia Sinica, Taipei, Taiwan 4 Taoyuan Psychiatric Center, Taoyuan, Taiwan 5 Institute of Public Health, National Yang-Ming University, Chia-Nan Psychiatric Center, Department of Health, Tainan, Taiwan PRODH and DGCR6 genes were reported to be associated with schizophrenia previously. We genotyped two SNPs and one insertion/ deletion (DGCRID) polymorphism of DGCR6 and 6 SNPs of PRODH in 124 schizophrenic cases and 112 normal controls in the first stage. We found five polymorphisms, DGCRID (P ¼ 0.03, 0.04), PRO2026 (P ¼ 0.02, 0.003), PRO1945 (P ¼ 0.06, 0.03), PRO1766 (P ¼ 0.04, 0.04), PRO757 (P ¼ 0.002, 0.006) were significantly associated with schizophrenia either through allelewise or genotypewise comparison. Then we genotyped the five SNPs in 93 co-affected sib-pair families of schizophrenia. Two phenotype models (narrow: DSM-IV schizophrenia only, and broad: including schizophrenia, schizoaffective, and other non-affective psychotic disorders) were used to define the disease phenotype. We conducted single locus and haplotype association analysis using Transmission v2.5.4 program. We found the polymorphisms of DGCRID and PRO2026 were borderline significantly associated with schizophrenia either in narrow or broad model (P ¼ 0.05–0.1) in the single locus analysis. The haplotype 2-1 composed of DGCRIDPRO2026 was significantly over-transmitted to affected individuals (P ¼ 0.006, P ¼ 0.01 in the narrow and broad model, respectively). To clarify whether the PRODH and DGCR6 gene are associated with specific subtypes of schizophrenia, we defined schizophrenic individuals with global negative symptom rating score above 2 (prominent negative symptoms) as affected, otherwise unaffected. We found the polymorphisms of DGCRID, PRO2026, and PRO1945 were significantly associated with schizophrenia (P ¼ 0.03, 0.002, 0.02, respectively). The haplotype 2-1 composed of DGCRID-PRO2026 was more significantly over-transmitted to schizophrenic individuals with prominent negative symptoms (P ¼ 0.0007). Our results revealed the

Abstracts PRODH/DGCR6 locus is significantly associated with schizophrenia, especially those with prominent negative symptoms. P14.20 FAMILIAL VULNERABILITY TO SCHIZOPHRENIA INCREASES THE RISK OF SUBSTANCE USE: A TWIN STUDY Leung YJ,1 Sannesy S,1 Lyons MJ,1 and Tsuang MT2 1 Boston University, Boston, United States 2 University of California—San Diego, San Diego, United States The use of illicit substances is significantly associated with schizophrenia. However, it is unclear as to whether the use of these substances is related to the illness itself or to an underlying vulnerability to schizophrenia. We studied 3362 male twin pairs from the Vietnam Era Twin (VET) Registry, including 12 MZ and 12 DZ pairs discordant for schizophrenia. Patterns of substance use were assessed with the Diagnostic Interview Schedule Version III for marijuana, stimulants, and ‘‘any illicit drug,’’ a category that included marijuana, stimulants, opiates, sedatives, and PCP. Five subjective effects of marijuana use (paranoia, confusion, anxiety, feeling ‘‘out of control,’’ and hallucinations) were also examined. For both schizophrenic probands and their nonaffected cotwins, regular use of marijuana, stimulants, and any illicit drug was significantly higher than for controls (P < .05). Use did not differ between the probands and cotwins themselves. In schizophrenic probands (but not their cotwins), marijuana and stimulant abuse/dependence was diagnosed significantly more often than in controls. Symptom counts did not differ between the probands and cotwins. Symptoms counts for cotwins versus controls differed as a function of zygosity for stimulants but not for marijuana or any illicit drug. None of the marijuana subjective effects differentiated probands or cotwins from controls. These results suggest that schizophrenia increases the risk of marijuana and stimulant abuse/dependence and of regular use of marijuana, stimulants, and any illicit drug. However, the differences between the cotwins and controls indicate that a familial vulnerability for schizophrenia is sufficient for the increased risk of regular use of these substances. This is supported and extended by the finding that probands and their unaffected cotwins do not differ significantly in regular use or in abuse/dependence for these three classes of substances. Familial vulnerability includes influences from both genes and the shared environment. The abuse/dependence of stimulants among cotwins of schizophrenics reflects genes associated with vulnerability to schizophrenia. We could not distinguish between genes and the family environment as the source of the elevated risk for abuse/dependence of marijuana and any illicit drug observed among the cotwins of schizophrenics compared to controls.

P14.21 RISK REGIONS FOR SCHIZOPHRENIA SITUATED ON CHROMOSOMES 6, 8 AND 10 Corder EH1 and Jablensky A2 1 Duke University, Durham, United States 2 University of Western Australia, Perth, Australia Regions of interest for schizophrenia have been mapped to chromosomes 6p, 6q, 8p, 8q, and 10q among others. We investigated 73 sibling pairs from 61 well-characterized families where one or more members was affected with schizophrenia or schizophrenia spectrum disorder. Information on the number of alleles shared by state for each pair was available for 20 marker loci over chromosome 6, 14 on chromosome 8, and 20 loci on chromosome 10. This information was entered into a grade-of-membership model (GoM, Woodbury 1978) together with measure of attention, verbal memory, executive function, handedness, IQ, and aspects of personality of proband and sib. Specifically, a first stage-model was constructed where the sole predictor was affectedness for the pair, the number of alleles shared on chromosome 6 provided starting values. A second-stage model was then constructed where affectedness for the pair and IBS status for variables demonstrating weak association with disease status in the first stage-model (H entropy score>0.1) were used predictively to identify risk sets: These were D6S470 at 18 cM, D6S422 at 36 cM, D6S292 at 137 cM, D6S281 at 190 cM, D8S1771 at 50 cM, D8S505 at 61 cM, D10S547 at 29 cM, D10S1693 at 137 cM, and D10S212 at 171 cM. Starting values were provided by CPT- IP scores for the pair. The five identified model-based groups were distinct. They were labeled ‘Schizophrenia’ (I), ‘Other— D10S212’ (II), ‘Other—6p-6q’ (III), ‘Unaffected’ (IV), ‘Other—D10S1850

133

(V). I: The core (50%) and spectrum (50%) phenotypes were found at similar frequency, the risk set consisted of chromosome 6 at 18 cM and 190 cM, chromosome 8 at 50 cM and 61 cM and chromosome 10 at 29 cM and 137 cM, both proband and sib had low CPT- IP score, the sib had intermediate laterality. II: Mapped to 10q at 171 cM. III: Mapped to 6p at 36 cM and 6q at 190 cM [shared with I]. IV: Unaffected status mapped to 10q at 116 cM. V: There was low allele sharing except at 116 cM on 10q. These results define a risk set for schizophrenia on chromosomes 6, 8 and 10. Both schizophrenia and depression may share a risk locus on chromosome 6q at 190 cM. P14.22 ASSOCIATION OF THE DTNBP1 LOCUS WITH SCHIZOPHRENIA IN A US POPULATION Funke B,1 Finn CT,1 Plocik AM,1 Lake S,3 DeRosse P,2 Kane JM,2 Kucherlapati R,1 and Malhotra AK2 1 HPCGG, Boston, United States 2 The Zucker Hillside Hospital, Glen Oaks, United States 3 Brigham and Women’s Hospital, Boston, United States Linkage and association studies have recently implicated the dystrobrevin-binding protein 1 (DTNBP1) in the etiology of schizophrenia. We analyzed seven previously tested DTNBP1 single nucleotide polymorphisms in a cohort of 563 cases with schizophrenia or schizoaffective disorder and 649 control subjects. The minor alleles of three SNPs (P1578, P1763, P1765) were positively associated with the diagnosis of schizophrenia or schizoaffective disorder in the Caucasian subset of the study cohort (258 cases, 467 controls) with P1578 showing the most significant association (P ¼ 0.0026). The same three SNPs were also associated in a smaller Hispanic subset (51 cases, 32 controls). No association was observed in the African American cohort (215 cases, 74 controls). A stratified analysis including the Caucasian and Hispanic subsets showed association with the minor alleles of four SNPs (P1578, P1763, P1320 and P1765). Again, the most significant association was observed for P1578 (P ¼ 0.0006). Haplotype analysis supported these findings with a single risk haplotype significantly overrepresented in the Caucasian sample (P ¼ 0.005). Our study provides further evidence for a role of the DTNBP1 gene in the genetic etiology of schizophrenia. P14.23 AN ANALYSIS OF DTNBP1 (DYSBINDIN) IN AUSTRALIAN AND INDIAN POPULATIONS Handoko HY,2 Nyholt DR,2 Thara R,3 Srinivasan TN,3 Hayward NK,2 James MR,2 Nertney DA,1 Smith HJ,2 Filippich CF,2 Hannah DE,1 McGrath JJ,1 Sujit KJ,3 Giri KP,3 and Mowry BJ1 1 Queensland Centre for Mental Health Research, Brisbane, Australia 2 Queensland Institute of Medical Rearch, Brisbane, Australia 3 Schizophrenia Research Foundation (India), Chennai, India Straub et al. (2002) found evidence for association between DTNBP1 (dysbindin, dystrobrevin binding protein) and schizophrenia, although no specific susceptibility variants were identified. This study was followed by haplotype analyses in the same population, with typing of several novel SNPs in addition to the markers used by Straub and colleagues. Replication rapidly followed in another family-based study of German and Israeli families albeit with a different haplotype (Schwab et al., 2003). Failure to find evidence of association in another Irish sample (Morris et al., 2003) which utilised similar family history based stratifications to that of Straub, might be due to lack of family history in this mostly case/control sample, as suggested by van den Bogaert and others (2003) in their study using German, Polish, and Swedish populations. Results will be presented from an Australian sample consisting of 50 ASP families, 53 cases (with a family history of SZ) and 191 controls (without a family history of SZ), and an Indian sample consisting of 21 ASP families and 65 trios. Linkage and LD analysis of 16 SNPs indicate that genetic variation in the DTNBP1 gene is involved in the development of SZ in the Australian and Indian populations and confirm the peak association of Straub et al. for SNP P1635. P14.24 ASSOCIATION BETWEEN SCHIZOPHRENIA WITH OCULAR MISALIGNMENT AND POLYALANINE LENGTH VARIATION IN PMX2B Toyota T, Ebihara M, Yamada K, Ohba H, Fukasawa M, Minabe Y, Nakamura K, Sekine Y, Takei N, Suzuki K, Itokawa M, Meerabux J.M.A, Iwayama-Shigeno Y, Tomaru Y, Shimizu H, Hattori E, Mori N, and Yoshikawa T

134

Abstracts

1

Riken Brain Science Institute, Saitama, Japan The Japan Foundation for Aging and Health, Aichi, Japan

2

The increased incidence of minor physical anomalies (MPAs) in schizophrenia is the fundamental basis for the neurodevelopmental hypothesis of schizophrenia etiology. Ocular misalignment, or strabismus, falls into the category of MPAs, but this phenotype has not been assessed in schizophrenia. This study reveals that a subtype of strabismus, constant exotropia, displays marked association with schizophrenia (P ¼ 0.00000000906). To assess the genetic mechanisms, we examined the transcription factor genes ARIX (recently identified as a causative gene for syndromic strabismus) and its paralogue, PMX2B. We identified frequent deletion/insertion polymorphisms in the 20-alanine homopolymer stretch of PMX2B, with a modest association between these functional polymorphisms and constant exotropia in schizophrenia (P ¼ 0.029). The polymorphisms were also associated with overall schizophrenia (P ¼ 0.012) and more specifically with schizophrenia manifesting strabismus (P ¼ 0.004). These results suggest a possible interaction between PMX2B and other schizophrenia-precipitating factors, increasing the risk of the combined phenotypes. This study also highlights the unique nature of the polyalanine length variations found in PMX2B. In contrast with other transcription factor genes, the variations in PMX2B show a high prevalence, with deletions being more common than insertions. Additionally, the polymorphisms are of ancient origin and stably transmitted, with mild phenotypic effects. In summary, our study lends further support to the disruption of neurodevelopment in the etiology of schizophrenia, by demonstrating the association of a specific MPA, in this case, constant exotropia with schizophrenia, along with molecular variations in a possible causative gene.

P14.25 CONFIRMATION OF THE CATECHOL-OMETHYLTRANSFERASE GENE CONTRIBUTION TO SCHIZOPHRENIA Handoko HY2, Nyholt DR,2 Hayward NK,2 James MR,2 Hannah DE,1 Nertney DA,1 Smith HJ,2 Filippich CF,2 and Mowry BJ1 1 Queensland Centre For Mental Health Research, Brisbane, Australia 2 Queensland Institute of Medical Research, Brisbane, Australia Several lines of evidence have implicated the gene for catecholamine metabolizing enzyme catechol-O-methyltransferase (COMT) as a candidate for schizophrenia susceptibility, not only because it encodes a key dopamine catabolic enzyme, but also because the COMT gene maps to the velocardiofacial syndrome region of chromosome 22q which has long been associated with schizophrenia predisposition. We have investigated the genetic contribution of COMT to schizophrenia susceptibility in a family-based sample. Fifty Australian Caucasian families were genotyped, utilising 4 single-nucleotide polymorphisms (SNPs), effectively representing 3 independent SNPs. Two separate and interacting effects within the COMT gene were associated with schizophrenia susceptibility. Haplotype analyses showed strong evidence of an association, the most significant in this 20-kb region being the three-marker haplotype rs737865-rs4680-rs165599 (global P ¼ 0.0022), the most aberrant haplotype for these loci being 1-1-2 (T-GG) (one transmitted vs 18 non-transmitted)(P ¼ 0.0012).

P14.26 RELATIONSHIP BETWEEN A HIGH-RISK HAPLOTYPE IN THE DYSTROBREVIN BINDING PROTEIN-1 GENE (DTNBP1) AND CLINICAL FEATURES OF SCHIZOPHRENIA Fanous AH,1 van den Oord EJ,2 Riley BP,2 Aggen SH,2 Neale MC,2 O’Neill AF,4 Walsh D,3 and Kendler KS2 1 Washington VA Medical Center-Georgetown University Medical Center Schizophrenia Research Program, Washington, D.C., United States 2 Virginia Commonwealth University, Richmond, United States 3 St. Loman’s Hospital, Dublin, Ireland 4 Queen’s University, Belfast, United Kingdom We sought to determine whether the putative mutation in DTNBP1 not only increases the susceptibility to psychotic illness, but also to a more or less clinically specific form of illness. In the Irish Study of High Density Schizophrenia Families, subjects with psychotic illness

(n ¼ 755) were given lifetime ratings of clinical features using the Operational Criteria Checklist for Psychiatric Illness (OPCRIT). Using exploratory and confirmatory factor analysis, five factors, including hallucinations, delusions, negative, manic, and depressive symptoms were extracted and subsequently used to create factor-derived scores. Using TRANSMIT, we determined whether a high-risk haplotype (HRH) in the DTNBP1 gene was over-transmitted to subjects in the upper 20 and 40 percentiles of each factor. This was compared to baseline over-transmission by examining the empirical distribution of chi-squared statistics in groups of 5000 replicates in which 20% and 40% of ill subjects were randomly selected. This was done for both narrow and broad definitions of illness. The upper 40% of subjects on the negative symptom factor inherited the HRH more than would be expected by chance in both the narrowly (P ¼ .008) and broadly (P ¼ .01) defined groups. No other relationships between clinical features and HRH transmission were observed. Therefore, psychotic subjects with prominent negative symptoms may be more likely to inherit risk alleles in DTNBP1 than are other patients. This supports previous evidence that the clinical heterogeneity of schizophrenia is due to etiological heterogeneity.

P14.27 POSITIVE ASSOCIATION BETWEEN G72/G30 GENE MARKERS ON 13Q33 AND SCHIZOPHRENIA IN THE IRISH STUDY OF HIGH DENSITY SCHIZOPHRENIA FAMILIES (ISHDSF) Thiselton DL,1 Wormley B,1 Webb BT,1 Neale B,2 Ribble RC,1 O’Neill FA,3 Walsh D,4 Kendler KS,1 and Riley BP1 1 Virginia Commonwealth University, Richmond, United States 2 Institute of Psychiatry, London, United Kingdom 3 The Queen’s University, Belfast, Ireland 4 The Health Research Board, Dublin, Ireland A growing number of reports are providing strong evidence for involvement of the nested antisense genes G30 and G72 in the etiology of schizophrenia and bipolar disorder. Both genes are expressed in brain, where the G72 protein interacts with D-amino acid oxidase (DAAO), a key component in the activation of NMDA glutamate receptors. As part of a multi-centre collaborative candidate gene study for schizophrenia, we have genotyped 7 SNPs within the G72/G30 gene, spanning a region of 30 kb, and 2 SNPs in DAAO, in the ISHDSF. This sample consists of 265 families (1410 individuals) with a high density of schizophrenia, and was analysed under a range of diagnostic categories from a stringent DSMIII-R criteria (d2) to a broader phenotype (d5 then d8). The chosen SNPs showed most significant levels of association in previous studies and/or were putative functional variants. Eight are from prior reports (G72/G30: M19, M18, rs1935062, M15, rs1341402, rs1935058 and DAAO: MDAAO5, MDAAO6) whilst 1 is novel (identified at the University of Wales College of Medicine, Cardiff, UK: G72-IVS3). Analysis of our data using Haploview and GOLD showed all 7 markers to be in strong LD with one another in a single haplotype block. We tested for association between single marker genotypes, then haplotypes with disease, by TDT using TRANSMIT. All individuals within each family were included in the analysis. This non-conservative ‘first-pass’ test of association showed consistently positive results across all diagnostic categories. Two single markers, M18 and rs1935058, showed evidence for association with schizophrenia in the ISHDSF (M18: global P ¼ 0.04-0.002, rs1935058: global P ¼ 0.040.02 from diagnostic category d2-d8). Several haplotypes showed stronger evidence of association with schizophrenia, possibly indicating unknown disease-specific sequence variations contained therein. The most significant results were obtained for 4-SNP haplotype rs1935062/G72-IVS3/M15/rs1341402 (global P ¼ 0.000050.0007 across diagnostic categories d2-d8, restricted to common haplotypes). Three further SNPS, M12, M22 and M23 are currently being added to the dataset for complete gene coverage. We found no evidence for association with disease for the DAAO SNPs. The lack of linkage signal on 13q33 in prior whole genome scan data from the ISHDSF is interesting in light of these results. Although this is not unprecedented, we are incorporating the new genotypes into our existing data and will present the new linkage findings. In addition to other association tests including FBAT and PDT, consecutive runs of more conservative TRANSMIT analyses, with 1 nuclear family/ pedigree and 1 affected/nuclear pedigree/family, are underway. We thereby hope to more definitively confirm these preliminary findings which further support a role for G72 gene variation in schizophrenia.

Abstracts P14.28 ASSOCIATION STUDY OF THE G72 AND D-AMINO-ACID OXIDASE (DAAO) GENES IN SCHIZOPHRENIA Shinkai T,1 De Luca V,1 Lanktree M,1 Hwang R,1 Zai G,1 Sicard T,1 Shaikh S,1 Wong G,1 Wong AHC,1 Arnold P,1 Potapova N,1 King N,1 Trakalo J,1 Mu¨ller D,1 Nakamura J,2 Kennedy JL,1 and Macciardi F1 1 Neurogenetics Section, CAMH, Department of Psychiatry, University of Toronto, Toronto, Canada 2 Department of Psychiatry, University of Occupational and Environmental Health, Kitakyushu, Japan A number of linkage studies have reported positive signals from the region of chromosome 13q34 in schizophrenia. Chumakov and colleagues (2002) have identified a gene called G72 in this region and found an association between this gene, as well as the D-amino-acid oxidase (DAAO) gene and schizophrenia. DAAO oxidizes D-serine, a potent activator of an N-methyl-D-aspartate (NMDA) type glutamate receptor in the human brain. The G72 protein is considered an activator of DAAO. The interaction of these two genes has therefore been implicated in the NMDA receptor regulation pathway in schizophrenia. Schumacher et al. (2004) have also found a positive association between these two genes and schizophrenia. Our aim was to replicate these findings in another population in order to further evaluate the contributions of these genes. Our sample consists of 192 matched case-control sets as well as a set of small families (N ¼ 110). Although preliminary data for M-15 (rs2391191) in the case-control sample gave negative results (genotypic distribution: w2 ¼ 3.16, d.f. ¼ 2, P ¼ 0.21, allele frequency: w2 ¼ 2.22, d.f. ¼ 1, P ¼ 0.14), individual genotyping is now being performed for 8 more single nucleotide polymorphisms (SNPs) in the G72 gene and 4 SNPs in the DAAO gene. Analyses to test the association include the transmission disequilibrium test (TDT) in the family sample. Details of association analyses for single markers as well as marker haplotypes will be presented.

P14.29 ASSOCIATION STUDIES OF THE NEUREGULIN 1, DYSBINDIN AND G72 GENES IN DUTCH SCHIZOPHRENIA PATIENTS WITH AND WITHOUT PROMINENT NEGATIVE SYMPTOMS Bakker SC,1 Hoogendoorn MLC,1 Selten J-P,1 Verduijn W,3 Pearson PL,2 Kahn RS,1 and Sinke RJ2 1 Department of Psychiatry, University Medical Center, Utrecht, Netherlands 2 Department of Biomedical Genetics, University Medical Center, Utrecht, Netherlands 3 Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, Netherlands A considerable number of schizophrenia patients show a rapid clinical deterioration during the first years of illness, accompanied by persisting negative symptoms (e.g. flat affect, loss of initiative, and social withdrawal). In this group with chronic schizophrenia, progressive loss of grey matter volume on follow-up brain scans has been reported. These findings suggest subtypes of schizophrenia with a different biological basis. There is increasing evidence for the involvement of several genes in the susceptibility to schizophrenia, which raises the question whether these genes are equally involved in schizophrenia with permanent deficit symptoms or with a more relapsing and remitting disease course. Previously, we have replicated the reported association of the neuregulin 1 (NRG1) gene, but not of the dysbindin and G72 genes, with schizophrenia in the Dutch population (Hoogendoorn et al., presented at the WCPG 2003). This sample was enlarged, and enriched for chronic schizophrenia by including patients mainly from long-stay wards of psychiatric hospitals. All patients had at least three Caucasian grandparents of Dutch ancestry. DSM-IV diagnosis of schizophrenia was made using the Comprehensive Assessment of Symptoms and History (CASH, Andreasen, 1987) and information from medical records. In addition, patients were diagnosed using the Schedule for the Deficit Syndrome (SDS, Kirkpatrick, 1989), which measures idiopathic, enduring negative symptoms. The sample consisted of 301 unrelated schizophrenia patients, of whom 143 fulfilled the criteria for the deficit syndrome, while 137 were classified as non-deficit schizophrenia. No SDS was obtained from 21 patients (7%). Control DNA was provided by 571 anonymous donors. Markers were selected based on published associations and our own previous results for 32 Single Nucleotide Markers (SNPs) in the three genes. In total eight SNPs were typed in NRG1 (n ¼ 1), dysbindin

135

(n ¼ 3) and G72 (n ¼ 4) on a Taqman system. For NRG1, marker SNP8NRG221533 was associated with non-deficit schizophrenia (P ¼ 0.009), but not with the deficit syndrome (P ¼ 0.14). Association with deficit or non-deficit schizophrenia was detected for neither of the markers in dysbindin and G72, nor for marker haplotypes (P > 0.24). This suggests that these genes do not play a major role in our sample, which is enriched for deficit schizophrenia. Recent results, however, indicate that different SNPs may be associated in different populations, and therefore, our results are not sufficient to exclude these genes as schizophrenia susceptibility genes. In conclusion, our results further support NRG1 as a schizophrenia susceptibility gene across populations. Interestingly, it is suggested that the course of illness in schizophrenia is influenced by different sets of genes. TM

P14.30 ASSOCIATION BETWEEN POLYMORPHISM OF THE SNAP29 GENE PROMOTER REGION AND SCHIZOPHRENIA Wonodi IW,1 Avila MTA,1 Hong LEH,1 Mitchell BDM,2 Stine OCS,3 and Thaker GKT1 1 Maryland Psychiatric Research Center, University of Maryland School of Medicine, Baltimore, United States 2 Department Medicine, University of Maryland School of Medicine, Baltimore, United States 3 General Clinical Research Center (GCRC) Genomics Core Facility, University of Maryland School of Medicine, Baltimore, United States Schizophrenia is a complex disorder with a heritability of around 80%. Identifying disease-related genes has proven to be particularly challenging, as the syndrome is likely to be heterogeneous with a multifactorial aetiology and many susceptibility genes, each exerting a small effect. Linkage studies implicate chromosome 22q as a locus for a schizophrenia-related gene (or genes). Additionally, a high prevalence of schizophrenia and bipolar disorder is found in individuals with the velo-cardio-facial syndrome (VCFS), which is commonly associated with small interstitial deletions of 22q11. A recent report suggests an association between the 849A/G single nucleotide polymorphism (SNP) of the SNAP29 gene promoter region, which maps to the same locus as the VCFS microdeletion, and schizophrenia. SNAP29 is a member of the SNAP-25 (synaptosome associated protein-25 kD) family of SNARE proteins, which play a crucial role in synaptic vesicle function, and synaptic transmission. Interestingly, the catechol-omethyl transferase (COMT) gene, which has been implicated in the pathophysiology of schizophrenia by its unique impact on prefrontal dopaminergic signalling, also maps to this region. In the current study, we report a replication of the finding of association between the 849A/ G SNP of SNAP29 and schizophrenia. We compared allele and genotype frequencies of the 849A/G SNP in 101 schizophrenia cases, and 71 normal controls. Schizophrenia cases met SCID criteria, and normal controls had no family history of Axis I disorders. Controls were selected from a similar pool to cases in ethnicity and gender. There was a significant difference in the frequency of alleles and genotypes in cases compared to controls (allele frequency: w2 ¼ 10.2, 1 df, P ¼ 0.0014, OR ¼ 2.03, 95% CI [1.3–3.1], genotype: w2 ¼ 11.25, 2 df, P ¼ 0.0036). These results support an earlier finding suggesting that a mutation in the promoter region of the SNAP29 gene, or a mutation in linkage disequilibrium with the 849A/G SNP may play a role in the etiopathophysiology of schizophrenia.

P14.31 CANDIDATE GENE STUDY ON CHROMOSOME 1Q32 IN FINNISH SCHIZOPHRENIA FAMILIES Pylkko¨ E,1 Kestila¨ M,1 Varilo T,1 Silander K,1 Loukola A,1 Kilpikari R,1 Hennah W,1 Partonen T,2 Paunio T,1 Lo¨nnqvist J,2 and Peltonen L1 1 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland 2 Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland Schizophrenia is a complex psychiatric disorder with a worldwide prevalence of 1%. The prevalence varies among populations and is significantly higher in certain isolated sub-populations, such as an internal isolate in North-Eastern Finland. The use of isolates can greatly facilitate the study of complex disorders due to their more homogeneous genetic and environmental background. We have previously identified a region on chromosome 1q32-q41 that shows significant evidence for linkage (Zmax ¼ 3.82) to schizophrenia in the Finnish sub-isolate (Hovatta et al., 1999). Linkage to chromosome 1q

136

Abstracts

has also been observed in other populations (Brzustowicz et al., 2000; Millar et al., 2000). Here we aim to investigate association on the 1q32 locus. We ascertained our study sample based on parents originating from the internal isolate. This provided an initial data set consisting of 360 individuals from 88 families, of which167 individuals are classified as affected. We genotyped intragenic SNPs to characterize the allelic diversity of the genes on the 13.4 Mb candidate region. The region contains approximately 80 known genes, of which we selected 18 based on their putative role in neuronal functions or their expression level in the brain. A total of 37 single nucleotide polymorphisms (SNPs) were selected from public databases. Selection criteria for SNPs included sufficient information content (minor allele frequency over 5%) in the Finnish population, and the aim to provide an even distribution of SNPs across these genes. The SNPs were genotyped using Sequenom’s MassARRAY technology. Preliminary statistical analyses indicate evidence for association to schizophrenia, especially for the genes in the vicinity of the multiallelic marker D1S2757, that initially provided the best evidence for linkage. P14.32 ASSOCIATION STUDY OF SCHIZOPHRENIA, AGE AT ONSET OF SCHIZOPHRENIA AND VAL158MET COMT GENOTYPE IN A FRENCH CAUCASIAN SAMPLE Meary A,1 Schurhoff F,2 Szoke A,2 Roy I,2 Chevalier F,3 Matheui F,3 Giros B,3 and Leboyer M2 1 Service de Pharmacologie Clinique, hoˆpital Henri Mondor, Cre´teil, France 2 Service de Psychiatrie Adulte, hoˆpital Albert Chenevier et Henri Mondor, Cre´teil, France 3 Unite´ INSERM 513, Faculte´ de me´decine, Cre´teil, France Several lines of evidence suggest that abnormal dopaminergic function in the prefrontal cortex may be a key factor in the pathophysiology of schizophrenia (Weinberger et al., 2001). The main inactivation pathway for dopamine is enzymatic cleavage by the Carboxy-O-MethylTransferase (COMT), thus this gene is an outstanding candidate gene for schizophrenia. A common polymorphism at codon 158 in the gene coding for COMT (COMT Val158Met) results in substantial effects, with homozygosity for the low activity (Met) allele leading to a three- to fourfold reduction in enzymatic activity compared with the high activity (Val) allele. There have been many conflicting reports of associations, which culminated in a recent meta-analysis of data (Glatt et al., 2003). The authors performing a meta-analysis of existing casecontrol and family-based association studies could not confirm the association between COMT genotype and schizophrenia. They concluded that the association could be different in Europe and Asia and needed more studies to be confirmed. Later, Tsai et al., in 2004 found an association between age at onset of schizophrenia and COMT genotype in an Asiatic population. Our study was designed to assess the association between COMT genotype and schizophrenia and to assess the association between COMT genotype and age at onset of schizophrenia in a European Caucasian population. Therefore 142 schizophrenic patients and 100 normal controls of Caucasian French origin were consecutively recruited. Clinical data were assessed using the Diagnostic Interview for Genetic Studies and the COMT Val158Met polymorphism was analysed. The association between schizophrenia and COMT genotype was analysed using a Chi-square test and the association with age at onset using an ANOVA test including genotype and age on onset. Results failed to show a significant association between the Val158Met polymorphism and schizophrenia or age at onset of schizophrenia in this population. In conclusion, our findings do not support a major role for this polymorphism in schizophrenia or age at onset of schizophrenia. Nevertheless, it is still possible that other polymorphisms in COMT gene confer risk for schizophrenia or that the tested COMT polymorphism is associated with another phenotype of interest.

P14.33 ASSOCIATION BETWEEN DOPAMINE TRANSPORTER GENE 30 UTR VNTR POLYMORPHISM AND SCHIZOPHRENIA—RESULTS OF MULTICENTER STUDY Czerski PM,1 Kapelski P,1 Godlewska B,3 Grzywacz A,2 Dmitrzak-Weglarz M,1 Parus M,2 Chmara M,4 Tybura P,2 Leszczynska-Rodziewicz A,1 Syrek S,2 Skibinska M,1 Samochowiec J,2 Landowski J,3 Rybakowski JK,1 and Hauser J1 1 Poznan University of Medical Sciences, Department of Psychiatry, Poznan, Poland

2 Pomeranian Academy of Medicine, Department of Psychiatry, Szczecin, Poland 3 Medical University of Gdansk, Department of Psychiatry, Gdansk, Poland 4 Medical University of Gdansk, Department of Medical Genetics, Gdansk, Poland

Dopamine hypothesis of schizophrenia is well established. One of the proteins involved in dopaminergic functions is dopamine transporter (DAT)—for which the coding gene displays a VNTR polymorphism in the 30 UTR region (Vandenbergh et al., 1992). Persico and Macciardi (1997) reported a statistically significant difference in genotype distribution of this polymorphism in patients with schizophrenia and controls. We investigated 30 UTR VNTR polymorphism of DAT gene in patients with the diagnosis of schizophrenia (n ¼ 518) and in control subjects (n ¼ 663), all from Polish population, recruited in three centres (Poznan, Szczecin, Gdansk). Subgroups of patients according to gender and the onset before the age of 18 years (n ¼ 69) were analysed. DAT polymorphism was genotyped by PCR-VNTR method (Vandenbergh et al., 1992). There were statistically significant differences in the frequency of the alleles (P ¼ 0.011 for whole group) and genotypes (P ¼ 0.033 for whole group) comparing patients with schizophrenia to the controls. When subgroups were analysed, the statistically significant differences were observed for males (P ¼ 0.005 for alleles and P ¼ 0.012 for genotypes), but not for females (P ¼ 0.580 for alleles and P ¼ 0.490 for genotypes). An association was also present in the subgroup of patients with onset of schizophrenia at the age of 18 or earlier. We found a positive result on both allele (P ¼ 0.004) and genotype (P ¼ 0.002) level. We observed an association between allele A9 (containing 9 repeats of 40 bp motif) and genotypes A9/A9 and A9/ A10 of DAT gene and schizophrenia. Of interest is also an observation of this effect in males, and in the subgroup of patients with early onset. Due to negative results in the most of studies of this polymorphism in schizophrenia (except for the study of Persico and Macciardi—1997) our observation requires an independent replication. The possibility that another polymorphism of DAT gene is responsible for the observed association remains possible. P14.34 ANTICIPATION AND DNA METHYLATION IN SCHIZOPHRENIA Imamura A,1 Jinno Y,2 Shimabukuro M,2 Hashida A,1 Matsumoto S,1 Ozawa H,1 and Okazaki Y3 1 Division of Neuropsychiatry, Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan 2 Department of Molecular Biology, Ryukyu University School of Medicine, Okinawa, Japan 3 Department of Neuropsychiatry, Mie University School of Medicine, Mie, Japan Several researchers have reported on the phenomenon of anticipation (a decrease in age of onset and an increase in severity over successive generations) in schizophrenia, but the biological entity of the phenomenon has been unclear. The aim of this study was to explore the molecular mechanisms including epigenetic DNA modification underlying genetic anticipation in schizophrenia. Twelve two-generation pairs with schizophrenia, who showed prominent anticipation, were examined. The subjects consisted of 13 members in the lower generation (6 male and 7 female, mean age (SD) was 35.6 (10.7) years old, mean age at onset was 20.4 (2.75) years old, all members were diagnosed with DSM-IV schizophrenia) ]and 12 members in the upper generation (2 male and 10 female, mean age (SD) was 64.3 (9.93) years old, mean age at onset was 48.5 (13.9) years old, subjects included 4 with delusional disorder and 1 with schizo-affective disorder).We measured 5-methyl-20 -deoxycytidine (mC) contents by a high performance liquid chromatography (HPLC) method in peripheral leukocytes DNA obtained from eight two-generation pairs with schizophrenia, to compare the global methylation levels between the generations. The proportion of (mC/(mC þ dC)) in the upper generation was 3.98, and that in the lower generation was 3.96. We found no evidence of a relationship between genetic anticipation and DNA methylation in schizophrenia. In future studies we will aim to collect data from more subjects. P14.35 SUICIDAL BEHAVIOUR AND THE SEROTONIN PATHWAY AMONG SCHIZOPHRENICS Wang AGW, Rasmussen HBR, and Werge TW University Hospital, Copenhagen, Denmark

Abstracts Many studies have shown that there may be a relation between the serotonin pathway and suicidal behaviour, especially with violent means. Schizophrenics have one of the highest frequencies of suicidal behaviour, especially in the first years of illness. Establishment of the Danish Psychiatric Biobank has made it possible to include schizophrenic patients with a history of years of illness and also information from clinical records, interviews, etc. Lifetime suicidal behaviour is seen as a behavioural endophenotype and related to the illness, some clinical aspects and SERT polymorphism. Results so far have shown age of onset as of greater importance for suicidal behaviour than polymorphism in the SERT gene.

P14.36 POSITIONAL CANDIDIATE GENE ANALYSIS IN SCHIZOPHRENIA USING HAN CHINESE AND SCOTTISH POPULATIONS: CONFIRMATION OF ASSOCIATION WITH NEUREGULIN AND DYSBINDIN AND ANALYSIS OF NOVEL CANDIDATES Collier DA,1 Li T,1 Yu Lan,2 Hu X,1 Sun X,3 Liu X,3 and StClair D2 1 Institute of Psychiatry, London, United Kingdom 2 University of Aberdeen, Aberdeen, United Kingdom 3 Sichuan University, Chengdu, China Individual and meta analysis linkage studies have identified several genetic loci for schizophrenia, including chromosome 8p, 2p, 13q, 22q and 6p. For several of these loci, putative susceptibility genes have been identified including neuregulin on chromosome 8p and dysbindin on chromosome 6p, whereas others such as 2p remain orphan loci. Consensus on associated haplotypes and the identity of pathogenic polymorphisms is still emerging. We have analysed a series of genes from these loci, including dysbindin, RGS4, neuregulin 1, proline dehydrogenase, brain-derived neurotrophic factor, D-amino acid oxidase and G72, focussing on both the replication of existing haplotypic/individual marker association in addition to the analysis of novel genes from orphan loci. We used a set of Han Chinese family trios and sibling pairs from Sichuan, SW China, and Caucasian cases and controls from Scotland, in order to compare associated haplotype structure between the two ethnic groups. We found association with neuregulin 1 and dysbindin in both ethnic groups, but only observed a direct replication of the original risk haploptype in the Scottish sample. These finding are supported by functional genomic analysis of gene expression in human brain tissue. Other genes showed less significant or no evidence for association. We have also identified new associations both with novel polymorphisms from previously assoicated genes and new candidates from orphan loci. These findings indicate that considerable mapping efforts plus functional genomics is required before a full understanding of the role of these various genes in psychosis is attained.

P14.37 TRANSMISSION DISEQUILIBRIUM ANALYSIS OF DOPAMINE D3 AND D4 RECEPTORS POLYMORPHISM Li T,1 Liu X,3 Sham P,2 Sun X,3 Hu X,1 Murray R,1 and Collier D1 1 Division of Psychological Medicine, Institute of Psychiatry, London, United Kingdom 2 SGDP Centre, Institute of Psychiatry, London, United Kingdom 3 West China Hospital, Sichuan University, Chengdu, China The dopamine hypothesis of schizophrenia has catalysed the analysis of dopamine receptors as candidate genes. A variety of studies have analysed dopamine receptor genes for association with schizophrenia, including polymorpshims in DRD2, DRD3 and DRD4. In previous study, we found no significant evidence for linkage or linkage disequilibrium between the DRD2 polymorphism -1416C and schizophrenia. In order to examine the relationship between DRD3, DRD4 and schizophrenia, we have examined DRD3 Ser-9-Gly polymorphism and DRD4 exon III VNTR polymorphism using a family-based association analysis in a sample of 229 schizophrenia family trios from SW China. We found no significant evidence for linkage or linkage disequilibrium between the DRD3 Ser-9-Gly polymorphism and schizophrenia. There was also no evidence of transmission disequilibrium either allele-wise (P ¼ 0.06) or genotype-wise (P ¼ 0.07) between the DRD4 exon III VNTR polymorphism and schizophrenia after 10,000 simulations for correction of multiallelic marker. However there was a trend of deficit in transmission of 2 repeats (x2 ¼ 5.65, df ¼ 1, P ¼ 0.02), though the significance became much weaker after 10,000 simulation

137

for correction of multiallelic marker (P ¼ 0.05). The association must be regarded as preliminary. Further studies on other DRD4 variant are warranted. P14.38 A POSITIVE ASSOCIATION BETWEEN HOMA-IR SCORE AND THE TRP64ARG POLYMORPHISM OF THE BETA3-ADRENERGIC RECEPTOR GENE IN CHRONIC SCHIZOPHRENIC PATIENTS Lan T,1 Lan T,2 Chu F,3 Liu Y,3 Chiu J,4 and Sun H1 1 Department of Adult Psychiatry,TYPC, Tao-Yuan, Taiwan 2 Institute of Genetics, Yang-Ming Univeristy, Taipei, Taiwan 3 Department of Clinical Laboratory, Tao-Yuan Hospital, Tao-Yuan, Taiwan 4 Yu-Li Hospital, Hualien, Taiwan The glucose intolerance has been identified as a major healththreatening issue in chronic schizophrenic patients whether they were exposed to conventional or atypical antipsychotics previously. We recruited 203 chronic schizophrenic patients who met the DSM-IV criteria in a psychiatric center in Taiwan with fulfilled consent forms. The HOMA-IR score was calculated by the equation, (Fasting blood sugar in mg/dL) X (Serum insulin level in mu/ml)/405, to indicate the degree of insulin intolerance for a subject. The Trp64Arg polymorphism of the beta3-adrenergic receptor gene has been quoted to have a relationship with obesity and type II DM in the past publications. 200 subjects were successfully genotyped in this region by using PCR and specified restriction enzyme RFLP method. After adjusting for each participant’s sex, age, and BMI status, the association between HOMAIR score and the Trp64Arg polymorphism of the beta3-adrenergic receptor gene was still held (regression coefficient ¼ 0.65, P-value ¼ 0.033). It suggested that the polymorphism of the beta3-adrenergic receptor gene may be related to the development of insulin resistance in chronic schizophrenic patients in Taiwan.

P14.39 EXPRESSION OF PROTOCADHERINX/Y IN THE DEVELOPING HUMAN BRAIN Priddle TH,1 Cranfield M,2 Hewitson KS,1 Groome NP,2 Schofield CJ,1 Esiri MM,1 and Crow TJ1 1 University of Oxford, Oxford, United Kingdom 2 Oxford Brookes University, Oxford, United Kingdom Protocadherin X (PCDHX, PCDH11X) and Protocadherin Y (PCDHY, PCDH11Y) encode cell adhesion molecules and are found on the X and Y chromosomes respectively. PCDHX is present in placental mammals, where one allele is thought to undergo X inactivation. However, PCDHY is found exclusively in humans as a result of a reduplicative transposition from the hominid X to the hominid Y after separation of the lineages that gave rise to the great apes and Homo sapiens. Human PCDHX is also noteworthy as it appears to escape X inactivation. As such this human specific gene pair, PCDHX and PCDHY, could account for uniquely human traits such as language and have been suggested as candidate genes for schizophrenia. There are six X linked isoforms and three Y linked isoforms generated by alternative splicing. We have used oligonucleotide probes and novel antibodies to determine the location of PCDHX/Y mRNAs and proteins in the developing human brain (tissue provided by the MRC/Wellcome Developmental Biology Resource). Both PCDHX/Y mRNA and immunoreactivity were detected in the developing cerebral cortex. Other areas of expression included the hippocampus, cerebellum and brainstem.

P14.40 LACK OF SUPPORT FOR A NOVEL SCHIZOPHRENIA RISK HAPLOTYPE OF THE DYSTROBREVIN BINDING PROTEIN GENE (DTNBP1) IDENTIFIED IN THE UK/IRISH POPULATION IN SAMPLES FROM GERMANY, POLAND AND SWEDEN Van Den Bogaert A,1 Schumacher J,2 Schulze TG,3 De Zutter S,1 Otte AC,2 Ohlraun S,3 Kovalenko S,4 Becker T,5 Jo¨nsson EG,6 Mattila-evenden M,6 Sedvall GC,6 Czerski PM,7 Kapelski P,7 Hauser J,7 Maier W,4 Propping P,2 Rietschel M,3 No¨then MM,8 and Cichon S8 1 Molecular Genetics, VIB8, University of Antwerp, Antwerp, Belgium 2 Institute of Human Genetics, University of Bonn, Bonn, Germany 3 Central Institute of Mental Health, Division Genetic Epidemiology in Psychiatry, Mannheim, Germany

138

Abstracts

4

Department of Psychiatry, University of Bonn, Bonn, Germany Institute for Medical Biometry, Informatics and Epidemiology, University of Bonn, Bonn, Germany 6 Department of Clinical Neuroscience, Psychiatry Section, HUBIN project, Karolinska Institute and Hospital, Stockholm, Sweden 7 Department of Psychiatry, University of Poznan, Poznan, Poland 8 Life & Brain Center, University of Bonn, Bonn, Germany 5

We have recently attempted to replicate previously reported associations of genetic variants in the schizophrenia susceptibility gene dystrobrevin-binding protein 1 (DTNBP1), or dysbindin, in three schizophrenia case-control samples of German, Polish, and Swedish descent. We identified significant evidence for association in the Swedish, but not in the German and Polish sample (Van Den Bogaert et al., Am J Hum Genet 2003, 73:1438–43). Interestingly, Williams et al. (Arch Gen Psychiatry 2004, 61:336–44) also failed to replicate originally implicated dysbindin haplotypes in a large schizophrenia sample from the UK/Ireland, However, after screening for novel variants and including them in the analysis, they obtained highly significant evidence for association with specific haplotypes which subsequently could be replicated in an independent sample from Ireland. The most significant haplotype was formed by a novel promoter SNP (SNP A), P1635, and P1655. We investigated this haplotype in our large case-control samples from Germany, Poland, and Sweden. Disappointingly, we do not find support for association of this haplotype in any of the samples tested. Possible explanations include a small proportion of patients linked to chromosomal region 6p22.

P14.41 POSITIVE ASSOCIATION OF G72-G30 AND DAAO WITH SCHIZOPHRENIA IN AN IRISH CASE-CONTROL SAMPLE McGhee KA,1 Morris DW,1 Murphy K,1 Schwaiger S,1 Nangle J,1 Donohoe G,1 Clarke S,1 Scully P,2 Quinn J,2 Meagher D,2 Baldwin P,2 Waddington JL,3 Gill M,1 and Corvin AP1 1 Neuropsychiatric Genetics Group, Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland 2 Stanley Research Unit, St. Davnet’s Hospital, Monaghan, Ireland 3 Department of Clinical Pharmacology, Royal College of Surgeons, 123 St. Stephen’s Green, Dublin, Ireland Dysfunction of the NMDA receptor regulation pathway may be involved in the pathogenesis of schizophrenia (SZ). Chumakov et al. (2002) reported association with two genes that are involved in NMDA regulation, G72 and DAAO. These findings were in the Canadian and Russian populations. Our study aimed to replicate these findings in an SZ association sample taken from the Irish population (cases n ¼ 299, controls n ¼ 645). Four markers were analysed at the G72 locus: rs3916965, rs2391191, rs3918342 and rs1421292. A further four markers were studied at the DAAO locus: rs2111902, rs3918346, rs3741775 and rs888531. These markers were chosen based on results reported by additional studies independent of the original findings (Williams et al., Schumacher et al., Addington et al., Datta et al., Bass et al., Ashkenazi et al.). We identified single marker associations at G72 with rs3916965 (P ¼ 0.005) and rs2391191 (P ¼ 0.01), and at DAAO with rs2111902 (P ¼ 0.02) and rs3918346 (P ¼ 0.003). Haplotype analysis using FastEHplus gave a significant association with G72 for a three marker haplotype: rs3916965, rs2391191 and rs1421292 (sim. P ¼ 0.003), and a significant association with DAAO for the four marker haplotype (sim. P ¼ 0.004). Our findings add further support for the role of G72 and DAAO in the pathogenesis of SZ. P14.42 NEW EVIDENCES FOR A HIGH RISK HAPLOTYPE FOR PSYCHOSES IN THE SYNAPTIC VESICLE MONOAMINE TRANSPORTER GENE (SVMT) Gutierrez B,1 Rosa A,1 Peralta V,2 Cuesta MJ,2 and Fan˜anas L1 1 University of Barcelona, Barcelona, Spain 2 Hospital Virgen del Camino, Pamplona, Spain The synaptic vesicular monoamine transporter (SVMT) plays a key role in the synaptic acummulation and release of monoamines such as dopamine and serotonin. In a recent case-control association study, we identified two high risk haplotypes for schizophrenia and bipolar disorder in SVMT gene (Gutie´rrez et al., Am J Med Genet 122B:35, 2003). Particularly, the

haplotype combinations 2666C-2683C-745A and 2666T-2683A-745G were estimated to be significantly more frequent in patients than in controls (P < 0.0000). The aim of the present work was to replicate in a family-based study the association previously described between psychoses and SVMT gene. Methods: Five different SVMT SNPs (440-T/G, 1368-C/T, 2666-T/C, 2683-A/C and 745-A/G) were analysed in a sample of 94 quadruplets consisting of discordant sibs for schizophrenia and their parents. Molecular analyses were performed using the ABI PRISM SNaPshot multiplex kit which allowed us to analyze all 5 markers at a time. To estimate genetic associations from allele and haplotype transmission from parents to affected offspring we used TDT and TRANSMIT programs. The independent statistical analysis of each SVMT marker showed a trend for preferential transmission of 2666-C allele (P ¼ 0.06) and 2683-C variant (P ¼ 0.07) from heterozygous parents to the affected offspring. No overall associations with the rest of polymorphisms were detected. To further validate the possible 2666-T/C and 2683-A/C associations, we analysed a haplotype system composed of 3 SNPs at SVMT locus (2666-T/C, 2683-A/C and 745-A/G). These three markers had previously been reported to form the core of the two high risk haplotypes mentioned above. In the present study, results confirmed that the 2666C-2683C-745A combination (one of the two putative haplotypes of risk) was excessively transmitted by parents to affected offspring (chi-squared ¼ 3.718, P ¼ 0.0212). The present family based association study replicates our previous findings in an independent sample of psychotic patients and controls and strongly supports that SVMT gene or certain regions of it could constitute a genetic substrate of susceptibility for schizophrenia and functional psychoses. This study was supported by Ministerio de Ciencia y Tecnologı´a (SAF2001-3400) and by the Vada and Stanley Foundation. Dr. Blanca Gutie´rrez is a researcher from the Spanish ‘‘Ramo´n y Cajal’’ program. P14.43 ASSOCIATION ANALYSIS OF VARIATIONS IN GABRG2 AND AKT1 GENES IN FINNISH SCHIZOPHRENIA SAMPLE Turunen JA,1 Peltonen JO,1 Hennah W,1 Loukola A,1 Paunio T,1 Sinsheimer JS,3 Ekelund J,1 Varilo T,1 Partonen T,2 Lo¨nnqvist J,2 and Peltonen L1 1 Department of Molecular Medicine, National Public Health Institute, Helsinki, Finland 2 Department of Mental Health and Alcohol Research, National Public Health Institute, Helsinki, Finland 3 Department of Human Genetics, University of California, Los Angeles, United States We have previously identified a region on chromosome 5q as a potential susceptibility locus for schizophrenia in a large Finnish family sample. Furthermore, several other studies have also displayed evidence for linkage within the vicinity of this locus. In order to further track putative schizophrenia predisposing genes at 5q31-34, we performed a SNP-based analysis of positional candidate genes. The first stage analysis was performed using 61 nuclear families from the late settlement region of North-Eastern Finland, which had primarily contributed to the linkage finding of the original study. Two SNPs within the gamma-aminobutyric acid receptor A subunit gamma 2 gene (GABRG2) provided suggestive evidence of association with the family based association test (FBAT). The analysis was expanded to cover our complete schizophrenia sample from Finland containing 449 nuclear families with 967 affected individuals. Multiple SNPs within the GABRG2 gene provided evidence for association, which was further strengthened by the observed overtrasmittance of the 3 SNP haplotype to affected family members (P ¼ 0.000079). Recently, a putative predisposing 3-SNP haplotype was identified in the AKT1 gene (Emamian et al., 2004), evoking high interest due to its association with decreased AKT1 protein levels in lymphocytes and brains of schizophrenic individuals. Furthermore, AKT1 has also been linked to the GABAsignalling pathway speculated to be involved in the etiology of schizophrenia. We are now studying the role of the AKT1 gene in the nationwide Finnish schizophrenia sample. We are genotyping AKT1 variations in the same family sample that we used for the GABRG2 study. We utilize Sequenom’s MassARRAY technology and target AKT1 with 8 SNPs. The SNPs are selected based on their reported association, sufficient minor allele frequency (>3%) and an even distribution of SNPs across the genes. Joint association analyses of SNPs in both GABRG2 and AKT1 will be performed to address the role of these two genes in the genetic aetiology of schizophrenia in Finnish families.

Abstracts P14.44 NEUROLOGICAL ABNORMALITIES IN SCHIZOPHRENIC TWINS: EXPLORING AETIOLOGICAL HYPOTHESES Picchioni MM,1 Toulopoulou T,1 Davies N,2 Landau S,1 and Murray RM1 1 Institute of Psychiatry, London, United Kingdom 2 Camden and Islington Mental Health NHS Trust, London, United Kingdom To explore the prevalence of neurological abnormalities in twins with schizophrenia, their non-schizophrenic co-twins, and control twins. The aim of the study was to clarify the aetiological significance of neurological abnormalities, their relevance as proposed endophenotype markers and to explore which subject characteristics modify the expression of the neurological abnormalities detected. Sixty five twin pairs varying in their zygosity and concordance for schizophrenia and seventy three control twin pairs were compared for total, primary and integration subscale scores of neurological abnormalities using the Neurological Evaluation Scale. There were no significant differences between probands from concordant and discordant pairs. Neurological abnormalities were increased in patients with schizophrenia compared to their non-schizophrenic co-twins and control twins. Neurological abnormalities in the non-schizophrenic co-twins from discordant pairs were increased compared to control twins, but there were no significant differences between the MZ and DZ non-schizophrenic co-twins. Neurological abnormalities were not related to levels of psychotic symptoms but were increased by anti-psychotic medication. Neurological abnormalities were additionally modified by pre-morbid schizotypal traits in all groups, particularly in the probands from discordant pairs. Neurological abnormalities in schizophrenia are a relatively stable characteristic, determined by a number of interacting factors. These factors include the genetic risk for the illness. However these data suggest that the presence of the illness and premorbid schizotypal traits confer additional risk for the development of neurological abnormalities, possibly by a separate mechanism to the genetic risk.

P14.45 INCREASED FREQUENCY OF A D18S70 ALLELE IN A FACTOR-DERIVED CLUSTER OF PATIENTS WITH NON-AFFECTIVE PSYCHOSIS Medina RA,1 Escamilla MA,1 Walss-Bass C,1 Ravento´s H,3 Montero AP,3 Dassori A,1 Armas R,4 and Contreras S1 1 University of Texas Health Science Center, San Antonio, United States 2 South Texas Veterans Health Care System, San Antonio, United States 3 Universidad de Costa Rica, San Jose´, Costa Rica 4 University of California at San Francisco, San Francisco, United States Emerging genetic information provides support to the notion that Schizophrenia (SC) and Bipolar Disorder (BD) are not two distinct and separate clinical entities, but two poles of a continuum of symptoms, in which Schizoaffective Disorder (SCA) typically shares attributes of both disorders. In an attempt to correct the limitations of current diagnostic systems, statistical approaches (factor, cluster and latent class analyses) have suggested alternative naturalistic phenotypes, which probably are closer to the biological underpinnings of these complex disorders. Using these methods, it has been reported that one subset of patients with SC clustered with BD Type I. From the genetic perspective, several chromosomal regions have been identified as susceptibility loci for both BP and SC: 10p13, 13q32, 18p, 18 q, and 22q11-q13. Particularly, D18S70 (on 18q23) has been found associated with severe BD in two families of the Central Valley of Costa Rica (CVCR). In the present study, we explore the same marker in 145 patients clinically diagnosed with SC. We recruited probands independently from one another from psychiatric hospitals and clinics in the CVCR. All probands had disease onset by age 40, a history of at least one psychiatric hospitalization, a discharge diagnosis of SC, bilineal Spanish surnames and a current residence in the CVCR. Each subject was interviewed by a bilingual psychiatrist using the DIGS. Medical records were abstracted. An interview with a close relative, using a Spanish version of the FIGS was completed for each subject. A best estimate consensus process was used to arrive at final diagnoses using three systems (DSM-III-R, DSM-IV, and RDC). The Lifetime Dimensions of Psychosis Scale was scored (Levinson) using all information

139

gathered. Using the LDPS information of 200 cases, we performed a principal component factor analysis, with varimax rotation. Using the factor ratings for each individual, we performed a hierarchical cluster analysis. We used a ABI Prism fluorescently labeled microsatellite marker. Genotypes were assigned using GENESCAN and GENOTYPER software. Principal components extracted were: 1) Non-affective psychosis, 2) negative symptoms and disorganization, 3) psychotic mania component, and 4) psychotic depression component. Compared with Cluster 1, Cluster 2 included cases with higher non-affective component, lower psychotic mania component and lower psychotic depression component, The negative and disorganization component was similar in the two clusters. D18S70 allele 14 was more frequently found in cluster 2 compared with cluster 1 (P ¼ 0.03). Our results suggest that D18S70 is also a susceptibility locus for non-affective psychosis.

P14.46 GAP JUNCTION BETA 5 GENE (GJB5) IS ASSOCIATED WITH SCHIZOPHRENIA IN PORTUGUESE NUCLEAR FAMILIES Ni X,1 Valente J,2 Azevedo M,2 Pato M,3 Pato C,3 and Kennedy J1 1 Centre for Addiction and Mental Health, Toronto, Canada 2 University of Coimbra, Coimbra, Portugal 3 Veterans Administration Hospital, Washington DC, United States Using differential screening of gene initiation sequences (GIS), the DNA sequences that surround the gene start codon, we identified a product, showing weaker signal in a pool of probands with schizophrenia compared with their parents. The sequence of this DNA fragment shows high homology with the gap junction beta 5 gene (GJB5). In a case-control study, we found a significant association between GJB5 and schizophrenia in Toronto sample. To confirm this finding in an independent population we used 163 Portuguese trios and nuclear families with DSM-IV schizophrenia probands and genotyped three polymorphisms of GJB5, 77A/G, 15bpIns/Del and 1117A/G. Transmission disequilibrium test (TDT) was examined using FBAT (version 150) and the transmission disequilibrium between haplotypes of the three GJB5 polymorphisms and schizophrenia was performed with TRANSMIT (version 2.5.2). In TDT analysis using each allele we found a trend to significance for allele transmission of the 77A/G (P ¼ 0.0796). TDT analysis using haplotypes showed a significant association between GJB5 and schizophrenia (P ¼ 0.0483). These results suggest GJB5 may play a role in the pathogenesis of schizophrenia.

P14.47 A FAMILY-BASED ASSOCIATION STUDY OF THE BDNF AND DRD3 GENES IN SCHIZOPHRENIA SPECTRUM DISORDERS Fatjo´-Vilas M,1 Rosa A,1 Arias B,1 Peralta V,2 Cuesta MJ,2 Miret S,3 Navarro ME,4 Martı´n M,4 Mun˜oz MJ,4 and Fan˜ana´s L1 1 Universitat de Barcelona, Barcelona, Spain 2 Virgen del Campo Hospital, Pamplona, Spain 3 Hospital Santa Maria, Lleida, Spain 4 Sant Boi de Llobregat, Barcelona, Barcelona, Spain Brain-derived neurotrophic factor (BDNF) promotes neuron proliferation, differentation and survival. Among others, BDNF is involved in the neurodevelopment of dopaminergic-related systems, and interacts with the mesolimbic dopamine systems. Specifically, a recent study has shown that BDNF synthesized by dopaminergic neurons is required for the normal expression of the dopamine D3 receptors in nucleus accumbens both during development and adulthood (Guillin et al., 2001). In view of the neurodevelopmental hypothesis and the dopamine theory in the etiology of schizophrenia, BDNF and DRD3 genes are therefore strong candidate genes for the illness. Furthermore, due to the BDNF role in D3 receptor expression, the study of their interaction appears to be of great interest. The sample consisted of 456 subjects from 114 Spanish nuclear families (proband with schizophrenia spectrum disorder according to DSM-IV and their healthy parents and sib) and 176 healthy individuals with similar ethnic and sociodemographic characteristics. We examined two polymorphisms: i) a Val66Met change in the coding region of the BDNF gene (11p13), ii) a Ser9Gly substitution located in exon1 of DRD3 gene (3q13.3). Both polymorphisms were determined using a Taqman 50 exonuclease assay. The transmission disequilibrium test (TDT) showed a significant

140

Abstracts

differential transmission of the BDNF Val allele from heterozygous parents to patients (w2 ¼ 9.00, P ¼ 0.002). Moreover, this preferential transmission was not observed in the group of healthy siblings (w2 ¼ 0.86, P ¼ 0.35) which constitute a control for a generalised segregation distortion at the examined locus. However, the DRD3 polymorphism analysed did not show preferential allelic transmission neither in patients (w2 ¼ 1.08, P ¼ 0.29) or siblings (w2 ¼ 2.38, P ¼ 0.12). When an epistatic analysis was conducted, we found no effect of BDNF and DRD3 genotypes within the patients (w2 ¼ 1.31, P ¼ 0.85), siblings (w2 ¼ 4.46, P ¼ 0.34) and controls (w2 ¼ 5.48, P ¼ 0.24) groups. Although no epistatic effect has been detected with the BDNF and DRD3 genes, our results suggest that the BDNF Vall66Met polymorphism is involved in the genetic vulnerability for schizophrenia. This work was supported by the Vada and Stanley Foundation and the Fundacio´ La Marato´ de TV3 (014430/31). P14.48 NEUROLOGICAL SOFT SIGNS IN SCHIZOPHRENIA SPECTRUM DISORDERS AS AN INTERMEDIATE PHENOTYPE: A FAMILY-BASED STUDY Fatjo´-Vilas M,1 Miret S,2 Campanera S,2 Navarro ME,3 Martı´n M,3 Mun˜oz MJ,3 Guerra A,3 Illa JM,3 Miralles ML,3 Barrantes-Vidal N,4 and Fan˜ana´s L1 1 Universitat de Barcelona, Barcelona, Spain 2 Hospital Santa Maria, Lleida, Spain 3 Complex Assistencial en Salut Mental, Barcelona, Spain 4 Universitat Autonoma de Barcelona, Barcelona, Spain Dissecting schizophrenia into its neurological component and using biological traits as intermediate phenotypes can dramatically improve the power of genetic studies. In this sense, neurological soft signs (NSS) have been pointed out to be a valuable intermediate phenotype for genetic studies (Egan et al., 2001; Gourion et al., 2003, 2004). The aim of this study was to analyse the genetic component of this putative indice of neurodevelopmental abnormality in schizophrenia studing: i) its familial distribution, ii) its relation with severity markers such as the age at onset and the premorbid adjustment in patients. The sample consisted of 140 individuals from 33 nuclear families (proband with schizophrenia spectrum disorder -DSM-IV-, and their healthy parents and sib) and 25 control subjects. Ratings on NSS were obtained with the Neurological Evaluation Scale (Buchanan & Heinrichs, 1988) and the premorbid ajustment was evaluated with the Premorbid Adjustment Scale (Cannon-Spoor, 1982). In order to test the genetic loading effect parents were classified into presumed carriers (one first or second degree relative with psychosis in their ascendant or collateral pedigree) and presumed non-carriers (no family history of psychosis other than their offspring) (FH-RDC, Endicott et al., 1975). First: we compared proband, sibling and control groups scores on NSS (ANOVA test) and the status group appeared to affect the NSS total score (P ¼ 0.00). Specifically, probands scored significantly worse than siblings (P ¼ 0.00) and controls (P ¼ 0.00). Second: a comparison between probands and their parents was conducted adjusting for age (ANCOVA test). This analysis showed no significant relation between the group status and the NSS total score (P ¼ 0.34). When the same statistical approach was performed considering the parents classification in presumed carriers (PC) and presumed non-carriers (PNC), probands and PC presented similar NSS scores both differing from the PNC group (P ¼ 0.02 and P ¼ 0.05 respectively). Third: a correlation analysis between the NSS total score and age at onset showed no significant values. When the relation between the NSS total score and the general premorbid adjustment was explored, significant results were found (P ¼ 0.04), suggesting that a poor premorbid adjustment could be a predictor of later neurological abnormalities. Despite the size of the sample included in this preliminar analysis, our results seem to indicate that neurological soft signs are an adequate intermediate phenotype in the research of schizophrenia etiology. This work was supported by the Fundacio´ La Marato´ de TV3 (014430/31). P14.49 COMT: EVIDENCE THAT VAL158MET IS THE FUNCTIONAL VARIANT AFFECTING RISK FOR SCHIZOPHRENIA Egan MF, Cheng J, Kolachana BS, Straub RE, Kleinman JE, and Weinberger DR NIMH, NIH, Bethesda, United States Prior studies suggest that the COMT val allele at codon 158 increases risk for schizophrenia by reducing prefrontal dopamine. More recently,

several studies (e.g. Schifman et al., 2002; Bray et al., 2003) suggest that other functional SNPs in COMT may increase risk, perhaps by reducing COMT expression and, presumably, increasing prefrontal dopamine. We examined these hypotheses by testing the impact of these SNPs on 1) association with diagnosis, and 2) COMT enzyme activity in lymphoblasts and brain tissue from human dorsolateral prefrontal cortex DLPFC. Because estrogen regulates COMT expression and sex has been hypothesized to interact with genotype, we stratified analyses when possible by sex. We genotyped 11 SNPs in 3 family samples, the largest of which included 320 Caucasian families (216 added to 104 previously reported ‘‘CBDB families’’) with at least one proband with a schizophrenia spectrum disorder. Two other cohorts came from the NIMH Human Genetics Initiative consortium. The val allele was overtransmitted to probands in each sample, reaching nominal significance only in females from 2 of the 3 samples. A putative promoter SNP (287) also showed nominally significant association in males from the CBDB families. A 3 marker haplotype (AG-G) containing val and including the SNPs from Schifman et al showed slight overtansmission in each cohort, reaching significance only in the CBDB sample females (P ¼ .04). While underpowered, these results do not show that the addition of the Schifman SNPs improves power or are more strongly associated with diagnosis relative to val158met. In vitro analysis of COMT enzyme activity was performed using a tritiated dopamine methylation assay. In samples of human DLPFC from 166 subjects, the val/val group had approximately 48% higher activity compared to met/mets, while val/met subjects were 22% higher than met/mets. Other SNPs had little effect on enzyme activity in either tissue type, except for 287, where the more common A allele was associated with slightly higher enzyme activity in lymphoblasts. This effect remained significant when samples were stratified by val/ met genotype. Similarly, val/met effects remained significant when conditioned on 287 genotype. These results demonstrate that val158met polymorphism exerts a substantial effect on enzyme activity in human dorsolateral prefrontal cortex and that this is not due to the effects of other COMT SNPs. It therefore likely that COMT genotype effects on prefrontal function are primarily due to the val158met polymorphism. While a weak effect of the 287 promoter or other functional elements cannot be excluded, these data also suggest that the reported associations between COMT genotype and schizophrenia are likely due at least in part to val158met.

P14.50 A 63-GENE ASSOCIATION SCREEN FOR SCHIZOPHRENIA AMONG 290 SCHIZOPHRENIA TRIOS OF ASHKENAZI JEWISH DESCENT Fallin MD,1 Nicodemus KK,1 Lasseter VK,2 McGrath JA,2 Wolyniec PS,2 Nestadt G,2 Liang KY,4 Valle D,3 and Pulver AE2 1 Dept of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States 2 Dept of Psychiatry and Behavioral Sciences, Johns Hopkins School of Medicine, Baltimore, United States 3 Dept of Pediatrics and Molecular Biology, Johns Hopkins School of Medicine, Baltimore, United States 4 Dept of Biostatistics, Johns Hopkins Bloomberg School of Public Health, Baltimore, United States Schizophrenia is a common complex psychiatric disorder that is highly heritable. However, few genes have been consistently identified as causal or as risk factors for this illness. Genetic heterogeneity has been cited as one of the primary culprits for such inconsistencies. For this reason, we have focused on gene identification among schizophrenic families of Ashkenazi Jewish descent, in hopes of reducing genetic heterogeneity across families and increasing the detectable effects of any particular allele/haplotype. We have performed a 63-gene screen (435 SNPs) of candidate genes (located in linkage regions or previously reported to be associated with schizophrenia) among 290 Ashkenazi Jewish case-parent trios with a DSM-IV diagnosis of schizophrenia or schizoaffective disorder. Although this was a pilot study, we achieved good coverage of many genes, with an average of 6.9 SNPs/gene (54 genes with at least 3 SNPs) and an average density of 1 SNP per 11.8 kb across genes (range 1/.6 kb 1/86 kb). We report descriptive analyses of LD patterns in these genes among our Ashkenazi sample. We also show single-SNP and haplotype-based TDT analyses for these trios, following both windowing and haplotype block structure-based approaches. Nine of the 54 genes for which haplotype analyses could be performed showed nominally significant haplotype associations. Our strongest haplotype findings are in the DPYSL2 gene on chromosome 8,

Abstracts the NOS1 gene on chromosome 12 and in genes located under our recently reported chromosome 10q linkage signal (Fallin et al. 2003). Further, given the increasing evidence for overlap in genetic etiology between schizophrenia and bipolar disorder, we have also genotyped a set of 374 bipolar I disorder case-parent trios of Ashkenazi descent on the same set of SNPs and have identified regions of overlap among these association studies. We note overlap in association on chromosomes 10 and 12.

P14.51 LINKAGE AND ASSOCIATION WITH SCHIZOPHRENIA ON CHROMOSOME 3Q13 AT THE DRD3 LOCUS AND EXTENSIVE FLANKING REGIONS Talkowski ME, Mansour H, Wood J, Chowdari KV, and Nimgaonkar VL University of Pittsburgh, Pittsburgh, United States The dopamine receptor D3 (DRD3) has been implicated in the pathogenesis of schizophrenia (SZ) following post-mortem studies as well as genetic association studies, particularly with a Ser9Gly coding polymorphism in the first exon. Despite intensive investigation and multiple meta-analyses, evidence for association with this gene is not persuasive. We hypothesized that the disparate results may reflect a primary association with SZ in a flanking genomic region. Hence, we investigated linkage and associations with 15 single nucleotide polymorphisms (SNPs) localized over a 365 kb region within and flanking DRD3. We genotyped 331 probands with SZ or schizoaffective disorder (SZA, DSM-IV), including 151 cases with available parents (case-parent trios). Cord blood samples from local live births served as unrelated, unscreened controls (n ¼ 279) for case-control comparisons. We conducted family based linkage and association analyses to evaluate transmission distortions. We also estimated differences in genotype and haplotype frequencies between cases and controls. TDT analyses revealed significant associations with 5 SNPs. Transmission distortions for multiple haplotypes including these SNPs were also evident, particularly haplotypes spanning intron 2-3 to exon 4 of DRD3. Significant case-control differences were found for the ser9gly variant as well as four additional SNPs, including SNPs localized 51 kb upstream (rs905568) and 255 kb downstream (rs1025398) of the ser9gly variant. Haplotypes spanning these SNPs were also significantly associated. An assessment of linkage disequilibrium (LD) and haplotype block structure across the entire 365 kb indicates a minimum of four distinct regions at chr 3q linked and associated with SZ. Our results continue to support an association of SZ/SZA with DRD3. Further, our results suggest an association may exist with additional variants over an extensive genomic region not previously investigated. To determine the implications of these initial results further analyses using additional SNPs (n ¼ 53) are underway, along with replicate studies. P14.52 GENETIC INVESTIGATION OF THE DOPAMINE HYPOTHESIS IN SCHIZOPHRENIA: COMPREHENSIVE LINKAGE/ASSOCIATION ANALYSIS OF 23 GENES Talkowski ME, Mansour H, Wood J, Chowdari KV, and Nimgaonkar VL University of Pittsburgh, Pittsburgh, United States A role of the dopaminergic (DA) dysfunction in the pathogenesis of schizophrenia (SZ) has long been postulated. Credible associations with several DA genes have been reported, e.g., COMT, DRD2, DRD3, yet consistent replications are lacking. This may partly be due to the fact that most association studies have analysed 1–2 polymorphisms/ gene. Further, credible linkage information for the appropriate genomic regions has been uncertain. We report here a comprehensive investigation of DA gene polymorphisms using a combined linkage/ association approach. All analyses are being conducted systematically in one sample, thus enabling further analyses of gene-gene interactions. We used available genome databases to mine twelve 30 cM regions of the human genome identified as having a high probability of linkage to SZ (Lewis et al., 2003). From 6,127 annotated genes, we selected seven positional DA candidate genes. We then selected 16 additional ‘biological’ candidate genes not localized to linked regions. We used public database information to identify 389 SNPs in exonic, intronic, 50 upstream, and 30 downstream regions. Genotype assays were designed using multiplexed single base extension reactions. We genotyped 151 probands diagnosed with SZ or schizoaffective disorder

141

(DSM-IV criteria) as well as their parents (case-parent trios). Cord blood samples from local live births serve as unrelated, unscreened controls (n ¼ 279) for case-control comparisons. Analyses of transmission distortions to affected probands as well as genotype, allele, and haplotype frequency estimations between cases and controls are underway.

P14.53 EVIDENCE FOR A CHROMOSOME 3Q LOCUS IN A COMPLETELY ASCERTAINED SAMPLE OF FAMILIAL SCHIZOPHRENICS FROM PALAU, MICRONESIA Byerley W,1 Klei L,2 Roeder K,2 Myles-Worsley M,3 Tiobek S,4 Otto C,4 Galke B,1 Xie W,5 Bacanu S-A,5 and Devlin B1 1 University of California, Irvine, United States 2 Carnegie Mellon, Pittsburgh, United States 3 University of Utah, Salt Lake City, United States 4 Belau National Hospital, Korror, Palau 5 University of Pittsburgh, Pittsburgh, United States We report genome wide linkage analyses of a completely ascertained sample of schizophrenics derived from Palau, a geographically isolated oceanic nation in Micronesia. Our sample includes 60 families and 649 subjects, of whom 154 are schizophrenics. Of the 60 families, 7 are large extended pedigrees with 8 to 14 samples cases per family, 18 are moderately sized with 4 to 7 samples cases per kindred, 23 are small nuclear-type families with 2 to 3 sampled cases per family and 12 are simplex families. The 649 subjects were genotyped with microsatellite markers every 10 cM by the Center for Inherited Disease research. We used FASTLINK 4.0 software package for testing two point parametric models and Simwalk 2 for nonparametric and multipoint analyses. The simplex families were not included in the linkage analyses. Genome wide, the highest lod score was observed with D3S2418 (lod 3.2), a telomeric marker on the long arm of chromosome 3 (216 cM from the pter). This region has not been implicated by other studies. Two markers yielded lod scores of almost 2, D8S136 (lod ¼ 1.99) and D4S2394 (lod ¼ 1.97, theta ¼ 0.25). The area surrounding D8S136 (44 cM from the pter) has been implicated by numerous studies and it contains the Neuregulin 1 gene, a possible susceptibility allele for schizophrenia.

P14.54 SCHIZOPHRENIA LINKAGE ON 5Q AFTER STRATIFICATION BY DTNBP1 HIGH RISK HAPLOTYPE AND ASSOCIATION WITH HINT1 IN THE ISHDSF Webb BT,1 Van den Oord EJCG,1 Kendler KS,1 Riley BP,1 O’Neill FA,2 Walsh D,3 Chen X,1 and Neale MC1 1 Virginia Institute for Psychiatric and Behavioral Genetics, Virginia Commonwealth University, Richmond, United States 2 Department of Psychiatry, The Queens University, Belfast, Ireland 3 The Health Research Board, Dublin, Ireland The evidence for linkage on 5q in the Irish Study of High Density Schizophrenia Families (ISHDSF) was reexamined in light of the association evidence with DTNBP1 on 6p. We previously showed that when the families containing the DTNBP1 high-risk haplotype (HRHþ) are removed, linkage evidence on 8p increased in the remaining families (HRH-). The pattern of linkage across 5q is also altered by this stratified analytic approach and the results suggest that the broad region of linkage on 5q may be a mixture of distinct susceptibility loci. Our original linkage results highlighted a 10 cM region at 135 cM, although a broad region with Z-scores >1 extends across more than 50 cM. Stratified analysis of this data set appears to produce two separate peaks with differential contribution from the HRH and HRHþ subgroups. In the HRH subgroup, the maximum difference is observed using the intermediate (D5) diagnostic category at the original peak (135 cM) where the NPL Z scores are 2.13 (HRH), 0.95 (HRHþ), and 1.37 (combined). By contrast, at 122 cM the pattern of difference is reversed, particularly under a narrow (D2) diagnostic model. At this position, NPL Z scores are 0.52 (HRH), 2.6 (HRHþ), and 1.81 (combined). Although none of the stratified results reached empirical genome-wide levels of significance based on simulated null genome scans, the differences observed in the stratified analyses at 122 and 135 cM were unlikely to occur by chance ( A and hCV245410 (Intron 1) polymorphisms may be associated with the suicide attempts in schizophrenic patients. The promoter variant (473T > A) was previously identified in the novel TPH2 gene (M. Kennedy, 2003). TPH2 genotypes were determined in DNA samples from 372 schizophrenic patients. Among the schizophrenic patients, 116 had a history of suicide attempt. Genomic DNA was amplified by PCR and typed by ABI7000. Both markers were found to be in Hardy-Weinberg Equilibrium and are in strong linkage disequilibrium (chi-squared ¼ 24.8, 1 df P ¼ 6.15  107). A trend towards association with history of suicide attempt in schizophrenia was found in the hCV245410 polymorphism (P ¼ 0.101), but there was no significance with the 473T > A polymorphism (P ¼ 0.351). When we analyzed the combination of the two markers using haplotype analysis, we found no association between suicide attempt and haplotype distribution (P ¼ 0.64). This finding suggests that these 50 marker haplotypes in TPH2 gene may not influence suicidal behaviour in schizophrenia, however, different markers in the TPH2 30 region should be investigated regarding this phenotype.

P14.57 EVALUATING AN ASSOCIATION OF RGS4 POLYMORPHISMS WITH SCHIZOPHRENIA: ‘MEGA-ANALYSIS OF OVER 12,000 INDIVIDUALS FROM 11 INTERNATIONAL SITES

Talkowski M, Bassett AS,10 Brzustowicz LM,9 Chen X,5 Chowdari KV,1 Collier DA,12 Cordeiro Q,6 Corvin AP,4 Deshpande, S,7 Egan MF,13 Gill M,4 Kendler KS,5 Kirov G,3 Levitt P,14, Lewis DA,1 Li,12 Mirnics K,1 Morris DW,4 O’Donovan MC,3 Owen MJ,3 Seltman H,2 Semwall P,8 Sobell JL,11 Thelma BK,8 Weinberger DR,13 Williams NM,3 Wood J,1 and Nimgaonkar VL1 1 University of Pittsburgh, Pittsburgh,United States 2 Carnegie Mellon University, Pittsburgh,United States 3 University of Wales College of Medicine, Cardiff, United Kingdom 4 Trinity College, Dublin, Ireland 5 Virginia Commonwealth University, Richmond, United States 6 University of Sao Paulo, Sao Paulo,Brazil 7 Dr. RML Hospital, New Delhi, India 8 University of Delhi South Campus, New Delhi, India 9 Rutgers University, Newark, United States 10 University of Toronto, Toronto, Canada 11 University of Southern California, Los Angeles, United States 12 Institute of Psychiatry, London, United Kingdom 13 National Institute of Mental Health, Bethesda, Maryland 14 Vanderbilt University, Nashville, Tennessee RGS4 was first proposed as a candidate gene for schizophrenia (SZ) following a report of specific, decreased expression of RGS4 in the prefrontal cortex from post-mortem SZ samples (Mirnics et al., 2001). We initially reported an association of a four SNP haplotype (SNPs 1, 4, 7, and 18, spanning the 50 untranslated region and first intron of RGS4) with SZ in three independently ascertained samples (Chowdari et al., 2002). Subsequently, three additional published replicate studies have reported associations (Morris et al., 2004) (Williams et al., 2004)(Chen et al., 2004 in press). However, the associated alleles and haploytpes have differed between populations within our own study and in the replicate samples. Further, the effect size for the putative association is modest (odds ratio 1.2, Morris et al.). Thus, the observations may reflect stochastic variation and/or allelic heterogeneity between populations. To test these hypotheses, we have initiated a multi-site collaborative meta-analysis. We compiled genotype data for RGS4 SNPs 1, 4, 7, and 18 independently ascertained by 11 research groups to enable familybased and case-control analyses. These data are available from four major ethnic groups: Caucasians, Latin Americans, Asian Indians and Chinese. The family based sample is comprised of 7,810 individuals from 2,160 families collected at 9 sites (1,716 case-parent trios). Unrelated control individuals were available from 7 sites, enabling case-control comparisons of 2,906 cases and 3,159 unrelated controls. We intend to evaluate transmission distortions in case-parent trios and extended familial pedigrees. We will also estimate genotype, allele, and haplotype differences between individual cases and unrelated controls. These analyses will be grounded in an understanding of the evolutionary history of the RGS4 region for the different populations.

P14.58 EXAMINATION OF THE D-AMINO-ACID OXIDASE LOCUS AS GENETIC RISK FACTOR FOR SCHIZOPHRENIA Suliman H,1 Schumacher J,1 Abou Jamra R,1 Becker T,2 Ohlraun S,3 Klopp N,4 Freudenberg J,1 Schulze TG,3 Tullius M,5 Illig T,4 Maier W,5 Rietschel M,3 Propping P,1 No¨then MM,6 and Cichon S6 1 Institute of Human Genetics, University of Bonn, Bonn, Germany 2 Institute for Medical Biometry, Informatics and Epidemiology, University of Bonn, Bonn, Germany 3 Central Institute of Mental Health, Division Genetic Epidemiology in Psychiatry, Mannheim, Germany 4 Genome Analysis Center (CAG), GSF-National Research Center for Environment and Health, Mu¨nchen, Germany 5 Department of Psychiatry, University of Bonn, Bonn, Germany 6 Life & Brain Center, University of Bonn, Bonn, Germany D-amino-acid oxidase (DAAO) catalyzes the oxidative deamination of D-amino-acids. Particularly the neurotransmitter D-serine, acting on the gycine binding site of NMDA-receptors (NMDA-R) is a specific substrate of DAAO and several lines of evidence suggest an involvement of D-serine as well as NMDA-R in the pathogenesis of schizophrenia. Most interestingly, genetic studies point to an association between variants at the DAAO locus and schizophrenia. Genotyping of 3 SNP markers within the DAAO gene in a German sample with schizophrenia confirmed a French Canadian study, in which 4 SNP markers were analyzed and gave significant association results. To detect additional genetic variants, we re-sequenced all 11 exons of the

Abstracts DAAO gene and putative regulatory regions in patients with schizophrenia and controls. We identified 22 variants that were not known so far and validated 17 SNPs already described in public databases. Based on our sequencing results we are currently genotyping 18 SNP markers in a case-control sample comprising 530 patients with schizophrenia. Results of the association study will be presented at the conference. P14.59 GENETIC ANALYSIS OF DARPP-32 IN SCHIZOPHRENIA Carroll LS, Williams NM, O’Donovan MC, and Owen MJ University of Wales College of Medicine, Cardiff, United Kingdom The dopamine- and cAMP-regulated phosphoprotein (DARPP-32) protein (OMIM #604399) is a common signalling target for neurotransmitter and neurohumeral pathways implicated in schizophrenia. DARPP-32 is expressed in neurones of the caudatoputamen, nucleus accumbens, amygdala and prefrontal cortex, regions believed to be dysfunctional in schizophrenia. DARPP-32 knockout mice exhibit sensorimotor gating deficits, attenuation of dopamine sensitivity and other phenotypes potentially related to schizophrenia. In addition a reduction in DARPP-32 protein expression has been reported in the dorsolateral prefrontal cortex of schizophrenics (Albert et al 2002 Arch Gen Psychiatry 59(8)). The DARPP-32 gene (Locuslink ID: 84152) maps to chromosome 17q11.1–q12, a region of in which we have recently reported genome wide significant linkage to schizophrenia (Williams et al 2003 Am J Hum Gen 73(6)). We have screened exonic and adjacent intronic sequences of DARPP-32 for sequence variation in 14 schizophrenic individuals. In total we have identified 10 SNPs, 7 of which are intronic, 2 are non-synonymous, and one is in the 30 UTR. All SNPs have been tested for association with schizophrenia in our sample of 709 DSM-IV schizophrenia cases and 710 matched controls by a quantitative DNA pooling allele frequency estimation approach.

P14.60 AN UPDATE ON ASSOCIATION STUDIES OF DTNBP1, NRG-1, RGS4 AND MRDS1 IN AN EXTENDED IRISH SCHIZOPHRENIA CASE—CONTROL SAMPLE Morris D,1 McGhee K,1 Schwaiger S,1 Nangle J,1 Murphy K,1 Donohoe G,1 Clarke S,1 Baldwin P,2 Scully P,2 Quinn J,2 Meagher D,2 Waddington JL,2 Gill M,1 and Corvin AP1 1 Neuropsychiatric Genetics Group, Institute of Molecular Medicine, Trinity College Dublin, Dublin, Ireland 2 Stanley Research Unit, St. Davnet’s Hospital, Monaghan, Ireland 3 Department of Clinical Pharmacology, Royal College of Surgeons, 123 St. Stephen’s Green, Dublin, Ireland We have previously published association analysis of dysbindin (DTNBP1, Morris et al. Schizophr Res 2003, Williams et al. Arch Gen Psychiatry 2004), neuregulin-1 (NRG-1, Corvin et al. Mol Psychiatry 2004) and regulator of G-protein signalling 4 (RGS4, Morris et al. Am J Med Genet 2004) in a schizophrenia case—control sample taken from the Irish population. The sample consisted of 249 cases and 231 controls. We have since extended this sample to consist of 299 cases and 645 controls. A sample of this size provides 80% power with 95% confidence to detect association with an odds ratio of 1.6. We have used this more powerful sample to test associations we previously reported at the DTNBP1, NRG-1 and RGS4 loci. In addition we have tested for evidence of association at the MRDS1 (OFCC1) locus at 6p24.3. The results in brief are, DTNBP1—association with the P1655-P1635SNPA haplotype (sim P ¼ 0.006). NRG-1—no association with the HapBIRE haplotype or the ‘core’ haplotype first reported in the Icelandic population. RGS4—association with RGS4-4—RGS4-7—RGS4-18 haplotype (P ¼ 0.04). MRDS1—no evidence of association. P14.61 EFFECT OF MATERNAL LOADING IN FAMILIAL SCHIZOPHRENIA Velayudhan L,1 Gangadhar BN,2 Janakiramaiah N,2 and Murali N2 1 St Pancras Hospital, London, United Kingdom 2 National Institute of Mental Health & Neurosciences (NIMHANS), Bangalore, India Familial and nonfamilial schizophrenias have differed in several aspects. We examined if differences can be explained on the basis of origin of genetic loading. We examined 48 first episode DSM-IV schizophrenia patients who were consecutively recruited for an ongoing study. Family history of schizophrenia in first and/or second

143

degree relative(s) was absent in 29 (NFH) and present in 19 (FH) patients. Diagnosis in the index patient as well as the affected family member was confirmed by clinical interview by two clinicians. Age at onset (AAO) was assessed using Interview for Retrospective Assessment of Onset of Schizophrenia. Psychopathology was rated on Positive and Negative Syndrome Scale. Familial schizophrenia had higher (P ¼ 0.0008) negative symptom score [mean (SD) 32.2 (10.5)] compared to NFH schizophrenia [mean (SD) 24.4 (8.5)]. There were more females in NFH group (M:F—NFH, 22:7, FH, 7:12, P ¼ 0.025). AAO was not different between the groups [NFH, 26.4 (6.6) yrs, FH, 25.1 (8.1) yrs, P ¼ 0.54]. The family loading came from the maternal side (MO) for eight and from non-maternal side (NMO) for eleven. MO group had significantly higher preponderance of females, younger AAO and more negative symptom score. (M:F—NFH, 22:7, MO, 1:7, NMO, 6:5) (AAO, mean (SD) yrs: NFH, 26.4 (6.57), MO, 20.5 (4.9), NMO, 28.4 (8.5)) (Negative syndrome score, mean (SD)—NFH, 24 (9), MO 50 (11), NMO, 44 (7)). We suggest the sex-wise origin of family loading should be considered in family studies of schizophrenia. P14.62 PRENATAL FAMINE AND SCHIZOPHRENIA: A COMPLEX GENE/ENVIRONMENT INTERACTION St Clair DM,1 Xu M,2 Wang P,3 Yu Y,3 Fang Y,4 Zhang F,1 Gu N,5 Feng G,5 and Sham P6 1 University of Aberdeen, Aberdeen, United Kingdom 2 Bio_X Life Science Reasearch Center, Jiao Tong University, SHANGHAI, China 3 Fourth People’s Hospital,Wu Hu, Anhui, China 4 Second People’s Hospital, Huang Shan, Anhui, China 5 Shanghai Institute of Mental Health, Shanghai, China 6 Institute of Psychiatry, London, United Kingdom 7 Institute of Nutritional Sciences, SIBS, Shanghai, China An earlier report that pre-natal exposure to malnutrition during the short-lived Dutch hunger winter of 1944–1945 produced an increased risk of schizophrenia in adult life, prompted us to examine this hypothesis in the context of a major famine. During 1959 -1961 China experienced one of the biggest famines in recorded history. Estimates suggest over 30 million Chinese died. Anhui was one of the worst affected provinces with a mortality rate of over 12 per cent. We have examined the risk of schizophrenia in a population of 3 million people living in Wuhu and surroundings counties in Anhui province and compared rates among those born before, during and after the famine years. Mortality adjusted rates for schizophrenia were roughly comparable among those born prior to the famine and in the years following the famine. However, the most exposed birth cohort conceived at the height of the famine showed a mortality adjusted two fold and highly statistically significant (P < 0.0001) increased risk (Odds ratios 2.33 and 1.96 for 1960 and 1961 respectively) for schizophrenia. The proportion of familial cases remained unchanged during the famine years. Different models of gene/environment interaction make differing predictions of effect on proportion of familial cases. Our data argue against an independent effect for famine regardless of genetic liability to schizophrenia and supports the idea that prenatal exposure to famine lowers the genetic risk threshold and/or increases survival of carriers of schizophrenia risk alleles. A modest selection advantage in times of famine may account for the high prevalence of schizophrenia in all human populations. P14.64 HAPLOTYPIC ASSOCIATION SPANNING THE 22Q11.21 GENES COMT AND ARVCF WITH SCHIZOPHRENIA Sanders AR,1 Rusu I,1 Duan J,1 Vander Molen JE,2 Hou C,1 Schwab SG,3 Wildenauer DB,4 Martinez M,5 and Gejman PV1 1 Center for Psychiatric Genetics, Evanston Northwestern Healthcare Research Institute, and Northwestern University, Evanston, United States 2 Department of Human Genetics, University of Chicago, Chicago, United States 3 School of Psychiatry and Clinical Neurosciences, and Centre for Medical Research, University of Western Australia, and Department of Psychiatry, University of Bonn (Germany), Perth, Australia 4 School of Psychiatry and Clinical Neurosciences, and Centre for Clinical Research in Neurosciences, University of Western Australia, Perth, Australia 5 Me´thodologie Statistique et Epide´miologie Ge´ne´tique des Maladies Multifactorielles, Institut National de la Recherche et de la Sante´ Me´dicale, Evry, France

144

Abstracts

Catechol-O-methyltransferase (COMT) has been implicated in schizophrenia by its function through its roles in monoamine neurotransmitter metabolism and its impact on prefrontal cognition, and also by its position through linkage scans and a strong cytogenetic association. Further support comes from association studies, especially family based ones examining the COMT variant, Val108/158Met. We have studied eight markers spanning COMT and including portions of the two immediately adjacent genes, thioredoxin reductase 2 (TXNRD2) and armadillo repeat deleted in velocardiofacial syndrome (ARVCF), using association testing in 136 schizophrenia families. We found nominal evidence for association of illness to rs165849 (P ¼ 0.051) in ARVCF, and a stronger signal (global P ¼ 0.0019–0.0036) from threemarker haplotypes spanning the 30 portions of COMT and ARVCF, including Val108/158Met with Val108/158 being the over-transmitted allele, consistent with previous studies. We also find Val108/158Met to be in linkage disequilibrium with the markers in ARVCF. These findings support previous association signals of schizophrenia to COMT markers, and suggest that ARVCF might contribute to this signal. ARVCF, a member of the catenin family, besides being a positional candidate, is also one due to its function, i.e., its potential role in neurodevelopment, which is implicated in schizophrenia pathogenesis by several lines of evidence.

P14.65 ASSOCIATION OF SER311CYS DRD2 POLYMORPHISM IN PATIENTS WITH SCHIZOPHRENIA FROM SOUTH INDIA Neetha NV, Nair C, and Banjeree M Rajiv Gandhi Centre for Biotechnology, Trivandrum, India Schizophrenia is a severe debilitating psychiatric disorder afflicting 0.5 to 1% of the world population. Nearly 6–7 million Indians suffer from this disorder variously described as the ‘‘Cancer of the mind’’ and the ‘‘Greatest disabler of the youth.’’ Schizophrenia has a proven genetic but complex mode of inheritance. The D2 subtype of dopamine receptor has been widely implicated in the pathogenesis of schizophrenia. The pharmacological observations reveal that excess dopaminergic activity lead to the psychotic symptoms of schizophrenia. In this study we have investigated the allele and genotypic frequency for Ser311/Cys polymorphism in Schizophrenia and control subjects from South India using the PCR-RFLP technique. Ser311/Cys polymorphism in the DRD2 gene is a C-G conversion, which changes the amino acid Serine to Cysteine in the third cytoplasmic loop of the receptor, which is responsible for G protein activation and is therefore a crucial component in the signal transduction cascade. Subjects of the present study included 70 schizophrenic patients (32 Males and 37 Females) with mean age of 33 from South India. Control samples included 103 ethnically matched individuals recruited from the staff and students of Rajiv Gandhi Centre. This study in Indian population reveals a nearly significant association of the Cys311Ser polymorphism to the disease. The allele frequency for the Cys311 is 15% in the schizophrenics and 8%in controls, which is the highest reported frequency. Both the Cys 311 allele frequency (P value ¼ 0.054) and the heterozygous genotype was found to have a very significant association with the disease (P value ¼ 0.02).There also exists a significant excess of Cys 311 allele in male schizophrenics when compared to controls. Current data shows no correlation of the heterozygous genotype to either, age of onset (P value ¼ 0.909.), use of antipsychotics (P value ¼ 0.84) positive family history (P value ¼ 0.679), or to the positive and negative symptoms. P15.1 THE MET ALLELE OF THE BDNF (VAL66MET) POLYMORPHISM IS ASSOCIATED WITH FRONTAL EXECUTIVE IMPAIRMENT, BUT NOT WITH VERBAL MEMORY IMPAIRMENT, IN EUTHYMIC YOUNG ADULTS WITH RECURRENT, EARLY-ONSET MAJOR DEPRESSION Smith DJ,1 Wilkie M,2 Smith G,2 Wolff R,2 Muir WJ,1 and Blackwood DHR1 1 University of Edinburgh, Edinburgh, United Kingdom 2 University of Dundee, Dundee, United Kingdom Brain-derived neurotrophic factor (BDNF) plays an important role in synaptic transmission and plasticity, and is involved in cognitive processes such as memory and learning. Bipolar patients with the homozygous Met/Met or heterozygous Val/Met polymorphism of BDNF have worse prefrontal cognitive function than Val/Val homogygotes. In this study, we tested for association between the Met allele and

cognitive function in a group of euthymic young adults with a history of recurrent, early-onset major depressive disorder (RE-MDD). Forty clinically euthymic young adults with a history of RE-MDD were genotyped for the Val66Met polymorphism of BDNF and assessed with a neuropsychological battery consisting of the California Verbal Learning Test (CVLT) and three separate tests of frontal executive function: the Brixton Spatial Anticipation Test, the Trail-making test, and the Stroop test. One of the subjects had the Met/Met genotype, 15 were Val/Met and 24 were Val/Val. Subjects were divided into two groups based on their BDNF genotype: 24 Val/Val subjects, and 16 Val/ Met or Met/Met subjects. Both groups were well matched in terms of current medications, sociodemographic characteristics and premorbid IQ. Compared to the homozygous Val/Val genotype, subjects with one or more Met alleles performed significantly worse on both the Trailmaking test (P < 0.04) and on the Stroop test (P < 0.02). There were no differences between groups on the CVLT or on the Brixton test. The Met allele of the BDNF Val66Met polymorphism is significantly associated with frontal executive impairment, but not with verbal memory impairment, in euthymic young adults with RE-MDD. This extends to unipolar depression the finding in bipolar disorder that the Met allele of this polymorphism is associated with frontal executive dysfunction.

P15.2 THE SEROTONIN TRANSPORTER GENE: INVESTIGATION OF EXPRESSION PATTERNS IN HUMAN BRAIN TISSUE AND CELL LINES Sugden K, D’Souza UM, and Craig IW SGDP Centre, IOP, London, United Kingdom The serotonin transporter has an important role in serotonergic neurotransmission, as it is responsible for termination of the action of synaptic serotonin. Expression of the serotonin transporter gene, 5HTT, in cell lines has been demonstrated to be modified by the presence of different variants of the 5-HTTLPR, a VNTR within the promoter region of the gene. This region exists in two main forms, the long form (L, 16 repeats) and the short (S, 14 repeats), although an additional 5 rare length variants have been described. Furthermore, there are a number of single nucleotide polymorphisms and single base insertion/ deletions within the common L and S variants, adding to the complexity within the region. Additionally, a VNTR within intron 2 of the gene has been shown to have an enhancer-like activity in mouse embryos and embryonic stem cells. In order to investigate the role of these variations in moderating 5-HTT gene expression, we performed real-time PCR on RNA extracted from human post-mortem brain tissue (15 individuals), with up to seven different brain regions assayed in each case (total samples ¼ 88). The brain regions examined were cerebellum, striatum, Midbrain, thalamus, Superior Frontal Gyrus, Superior Temporal Gyrus and hippocampus. Expression level was correlated to genotype at the 5-HTTLPR and intron 2 VNTR, and to brain region. Furthermore, we assessed the ability of these regions to support differential reporter gene expression in various neuroblastoma cell lines dependent upon presence of either an L or S allele of the 5-HTTLPR or 10 or 12 copies of the intron 2 VNTR. The implications of these investigations will be discussed in relation to incongruity of 5-HTT expression in brain tissue and cell lines. P15.3 ALTERED EXPRESSION OF BRAIN RELATED GENES IN LYMPHOCYTES IN SCHIZOPHRENIA Bowden NA,1 Weidenhofer JC,1 Scott RJ,3 Todd J,1 Case V,1 Schall U,1 and Tooney PA3 1 University of Newcastle, Newcastle, Australia 2 NISAD, Newcastle, Australia 3 Hunter Area Pathology Service, Newcastle, Australia In recent years changes in gene expression have been observed in postmortem brain tissue from individuals with schizophrenia using microarrays. In particular, significant changes have been reported in neuronal myelination, cellular signalling and presynaptic pathways. This study aimed to determine if similar changes in gene expression could be detected in peripheral blood lymphocytes from 14 individuals with schizophrenia and 14 non-psychiatric controls, matched for age and gender. cDNA microarrays consisting of 6,000 genes were used to identify genes with a fold change greater than 1.5 in peripheral blood lymphocytes. The level of expression for 123 genes was significantly altered in schizophrenia for more than half of the matched pairs, 5 of which were confirmed using Real-time PCR. Confirmed changes were

Abstracts found in genes associated with myelination, metabolism, neurotransmitter receptor regulation and damage recognition in the CNS. Some of these results are consistent with previous reports from post-mortem brain tissue studies suggesting these pathways may be involved in the pathogenesis of schizophrenia. Further studies of these genes are required in individuals with schizophrenia, including investigation of the effects of anti-psychotic drugs. In conclusion, this study shows peripheral blood lymphocytes can be used to indicate gene expression changes in schizophrenia that may be useful in the development of a biological basis for diagnosing this disorder.

P15.4 GENE PROFILING IN THE AMYGDALA IN SCHIZOPHRENIA Weidenhofer J,1 Bowden NA,1 Scott RJ,2 and Tooney PA1 1 Neuroscience Institute of Schizophrenia and Allied Disorders (NISAD)/University of Newcastle, Newcastle, Australia 2 Hunter Area Pathology Services/University of Newcastle, Newcastle, Australia There is a genetic component involved in the development of schizophrenia, however the genes involved remain unknown. The amygdala is one brain region that has been implicated in the pathogenesis of schizophrenia. To investigate gene expression changes in the amygdala in schizophrenia, microarray studies were performed on post-mortem tissue from the NSW Tissue Resource Centre. Arrays consisting of DNA oligos from 19,000 genes were used to identify genes whose expression were consistently and significantly altered in the amygdala from seven subjects with schizophrenia when compared to seven subjects with no history of psychiatric illness, matched for age, gender and post-mortem interval. This study identified 132 genes whose expression was consistently up- or down- regulated (fold change greater than 1.5) in at least four of the seven matched pairs. Analysis of 5 genes by Realtime PCR has confirmed these array results. Western Blotting is now being used to identify whether the alterations in gene expression also effect protein expression. Some of the genes with altered expression have functions in cellular signaling and myelination of axons. The effect of antipsychotic drug treatments on these genes needs to be determined to ascertain if the dysregulation is due to the pathogenesis of schizophrenia or its treatment. In conclusion, these findings support arrays studies performed by other investigators showing changes in other brain regions of the expression of genes involved in cellular signaling and myelination of axons, implicating these processes in the pathophysiology of schizophrenia. P15.5 ALLELIC EXPRESSION PATTERNS OF PSYCHIATRIC CANDIDATE GENES IN A MOUSE MODEL Stuffrein S, Rogers GR, Cameron V, Eccles MR, Joyce PR, and Kennedy MA Christchurch School of Medicine & Health Sciences, University of Otago, Christchurch, New Zealand Genetic factors contribute to the risk of developing psychiatric disorders, yet despite extensive efforts, identifying molecular origins and mechanisms for mental illnesses continue to be a major challenge due to the complexity of these disorders. Experimental data accumulated by genetic linkage and association studies are often inconsistent and ambiguous. In recent years, clinical and molecular investigations have established that epigenetic phenomena are an important component of human disorders and are likely to contribute to the observed human phenotypic diversity. Epigenetic mechanisms, such as genomic imprinting and X chromosome inactivation, influence the level and/or pattern of expression of the involved gene. One specific epigenetic phenomenon is monoallelic expression in which only one of the two alleles is expressed. To date, most psychiatric genetic studies are based on the assumption that both alleles of a given gene are equally expressed. Yet monoallelic gene expression would profoundly influence the outcome and interpretation of association studies. We are investigating mental disorders from an epigenetic perspective by systematically screening a range of psychiatric candidate genes for their allelic expression pattern in the brains of hybrid mice. We hypothesise that some candidate genes will show monoallelic expression in the brain. A secondary hypothesis is that monoallelic expression will occur in a tissue specific fashion. We are testing these hypotheses by examining the expression patterns of specific genes in the brains of hybrid offspring of two inbred mouse strains (BALB/c and C57BL/6). These

145

offspring potentially possess different alleles at every gene locus. Key differences between the alleles are being identified using a combination of bioinformatics, DNA sequencing and automated mutation detection (WAVE DHPLC). Genetic assays for these differences will allow us to distinguish the RNA expressed from the two different alleles and thus analyse the expression pattern of any given gene.It has been previously shown that there is significant conservation of imprinting and monoallelic expression between rodents and humans. If any of the psychiatric candidate genes show monoallelic expression in the mouse, we envisage testing human brain RNA to confirm whether the same pattern holds in humans. Therefore, the outcome of this study may have substantial implications for the research into the genetics of human mental health.

P15.6 SEARCHING FOR NEW CANDIDATE GENES IN MAJOR DEPRESSION: DIFFERENT PROTEIN EXPRESSION LEVELS AFTER ANTIDEPRESSANT TREATMENT ON EBV CELLS OF DEPRESSED PATIENTS Zill P,1 Baghai TC,2 deJonge S,1 Neumeier K,1 Ackenheil M,1 Mo¨ller HJ,2 and Bondy B1 1 Department of Psychiatry of the LMU Munich, Section of Psychiatric Genetics and Neurochemistry, Munich, Germany 2 Department of Psychiatry of the LMU Munich, Munich, Germany In recent years it became obvious that the targets of antidepressant drug action are more likely to lie beyond the receptor levels. Although, the molecular loci of antidepressant drug action have not yet been fully established, it is known that antidepressant treatment modulate gene expression on the genomic level, respectively protein expression.The aim of the present study was to investigate the effects of mirtazapine (NASSA), reboxetine (SNRI) and fluoxetine (SSRI) on the protein expression levels on EBV cells of depressed patients as peripheral model using antibody microarrays (‘‘protein chips’’). Lymphocytes were prepared from whole blood and immortalized via EBV transformation. The cells were incubated with the medium plasma level concentration of each antidepressant drug for 48 hours and compared with an untreated sample of the same cell line. The protein chips contained 380 monoclonal antibodies on glass slides (proteins of signal transduction, cell-cycle regulation, gene transcription, apoptosis, oncogenesis). First results show clear differences in the protein expression. Especially the levels of three proteins (neurogenin 3, K-channel a, JNK1), which play a role in signal transduction processes were altered after incubation with each antidepressant. These results could be supported in single experiments on mRNA level with quantitative PCR. Our data suggest that mirtazapine, reboxetine and fluoxetine treatment modulate to a great extent the expression of proteins involved in signal transduction and neuronal plasticity. The corresponding genes of such proteins could serve as new candidate genes in affective disorders. These results have to be validated in further studies. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine’’ P15.7 THE HUMAN SEROTONIN TRANSPORTER INTRON 2 POLYMORPHISM WHICH HAS BEEN POSTULATED TO CORRELATE WITH AFFECTIVE DISORDERS IS A TARGET FOR LITHIUM MODULATION Quinn JP, Bubb VJ, and Scott A University of Liverpool, Liverpool, United Kingdom In a number of genes variable number tandem repeats (VNTR) sequences within non-coding regions have been associated with susceptibility to a variety of diseases. The serotonin transporter (5HTT) gene contains such a VNTR within intron 2. The implications of this polymorphism are not yet understood, however we have demonstrated that the 5-HTT VNTR is a transcriptional regulatory domain in vitro and in vivo and the allelic variation correlates with differential transcriptional regulation. Analysis of reporter gene expression supported by the different intron 2 VNTRs was addressed in stable transfection in the presence and absence of either chronic or acute exposure of the cells to lithium. The human cell, Jar, which express the endogenous 5HTT gene was used. The different copy number variants of the VNTRs exhibit a differential response to both chronic and acute exposure to lithium. We demonstrate the mechanisms by which these VNTR domains may in part correlate with progression of affective disorders as their transcriptional properties are specifically

146

Abstracts

regulated both in transient and stable reporter gene constructs by exposure of the cells to lithium chloride. Further the different copy numbers of the VNTR have significantly differential responses to lithium whether the exposure is chronic or transient. That these domains are a target for Lithium regulation indicates that they could also be considered drug targets for future therapeutic intervention.

P15.8 EXPRESSION ANALYSIS OF FIVE SCHIZOPHRENIA CANDIDATE GENES IN POST-MORTEM CEREBELLUM SAMPLES USING QUANTITATIVE REAL TIME RT-PCR Dempster EL, Mill JS, Murray R, and Collier DA Institute of Psychiatry, London, United Kingdom There are now compelling genetic associations between several genes and schizophrenia, such as neuregulin and dysbindin. Although these genes are not novel, relatively little is known about their function and the mechanism through which they cause vulnerability to schizophrenia is unknown. To examine the effects of schizophrenia associated polymorphisms and haplotypes on the mRNA expression of these genes we have utilised quantitative real time RT-PCR on RNA extracted from post-mortem cerebellum samples. Genes with candidate polymorphisms hypothesised to alter expression levels (Neuregulin-1 (NR-1), Brain-derived neurotrophic factor (BDNF), catechol-O-methyltransferase COMT), or implicated by protein expression studies, where they were found to be down regulated in schizophrenic brain (synaptosomeassociated protein 25 kDa (SNAP-25) and Reelin (RELN)). In appropriate genes we attempted to examine different splice variants to determine if the genotype of certain polymorphism altered the abundance of a particular splice variant. 60 cerebellum post-mortem samples were obtained from the Stanley Brain Foundation. RNA and DNA were simultaneously extracted from the brain tissues using the Trizol reagent (Life technologies). The TaqMan Real-time RT-PCR system (Applied Biosystems 3700, Foster City, USA), which uses a nonextendable 50 fluorescently labelled probe, was used to measure gene expression. The relative expression of specific candidate genes was determined by comparing their expression to a set of two housekeeping control genes: b-actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) utilising the comparative CT method (~~CT). Preliminary results examining expression versus ‘at risk’ genotype show, the val allele of the COMT val156met polymorphism confers lower expression of mRNA P ¼ 0.02, the presence of the genotypes that compile the ‘schizophrenia at risk haplotype’ in neuregulin-1 predicts lower expression of the NR-1 type 3 transcript P ¼ 0.01, and the val allele of the BDNF val66met polymorphism predicts increased expression, however, this observation only reaches significance in the BDNF 5U variant P ¼ 0.04 (non-parametric test). To date, no effect of genotype on expression of RELN or SNAP-25 has been found, but the ‘at risk’ variants in these genes are less well described. Subsequent analysis will concentrate on expression versus phenotype information, which should become available for this sample. P15.9 X INACTIVATION OF PCDH11X IN HUMAN AND NON-HUMAN PRIMATES Ross NA,1 Close J,1 Lopes A,2 and Crow TJ1 1 SANE, OXFORD, United Kingdom 2 IPATUMUP, OPORTO, Portugal PCDH11X/Y are human-specific homologous genes located in Xq21.3/ Yp11.2 respectively. The transposition from the X to the Y occurred in hominids about 4 Myrs ago. Consequently human males have 2 copies of PCDH11 whereas non-human male primates have only one copy. In view of the relatively recent acquisition of a Y copy in humans, it is predicted but not established that PCDH11X escapes inactivation in humans but is inactivated in female primates. We have investigated the methylation status of CpG island 1 of PCDH11X in human and gorilla lymphocytes using the restriction enzyme McrBC. This is a novel enzyme which cuts methylated DNA but fails to cut unmethylated DNA. Digests of genomic DNA from human and gorilla females were amplified by PCR using Cy5-labelled primers within CpG island 1 of PCDH11X and the products (representing unmethylated DNA) were separated by electrophoresis in acrylamide gels and quantified. The results suggest that there are approximately twice as many unmethylated CpG sites at the 50 end of the human CpG island compared to the gorilla homologue indicating that PCDH11X is probably subject to X inactivation in the gorilla but escapes inactivation in humans. These

findings are consistent with chromatin immunoprecipitation tests which have shown that lysine 4 methylated histone H3 and lysine 5 methylated histone H4 (which bind to active chromatin) are bound to approximately twice as much human PCDH11X sequence than the corresponding sequence in gorilla. Work is in progress to confirm these results by bisulphite sequencing of the promoter/CpG island region. P15.10 A MICROARRAY GENE EXPRESSION STUDY OF THE SIDE EFFECTS OF THE ANTIPSYCHOTICS DRUGS CLOZAPINE AND HALOPERIDOL: ANTIPSYCHOTIC METABOLIC SIGNATURES IN THE RIGHT SIDE OF THE MOUSE BRAIN AS MODEL FOR THE TREATMENT OF SCHIZOPHRENIA Rizig MA,1 McQuillin A,1 Hubank M,3 Hunt SP,2 and Gurling HMD1 1 Molecular Psychiatry Laboratory, Dept of Mental Sciences, Windeyer Institute, University College London, 46 Cleveland Street, London W1T 4JF, London, United Kingdom 2 Department of Anatomy and Developmental Biology, UCL, Gower Street, London WC1E 6BT, London, United Kingdom 3 Molecular Haematology Unit, Institute of Child Health, 30 Guilford Street, London WCIN 1EH, London, United Kingdom Clozapine has a superior clinical effect compared to any other antipsychotics. Metabolic abnormalities such as diabetes, hyperlipidemias and weight gain are troublesome side effects, which can limit clozapine’s superiority. In order to understand the cause of these side effects we have carried out a comparison of mRNA expression in the right side of the mouse brain in response to the antipsychotic drugs Clozapine and Haloperidol, compared to normal controls after three months of oral treatment. A metabolic transcriptional signature was derived from gene expression data on the 22690 transcripts in the affymetrix Mouse 430A chip and using 10 replicates in each of the three conditions studied (Clozapine, Halperidol, no treatment). Genes with gene ontology within the domain of a metabolic function, which showed more than a 1.5 fold change in expression in each condition, were identified. Haloperidol was found to have affected a larger number of transcripts (99 up regulated as opposed 1176 down regulated) compared to Clozapine (315 up regulated and 456 down regulated). Nevertheless the density ratio of the transcripts with a metabolic gene ontology classification was the same for both drugs (30%, 32% respectively). However, the pattern of the transcripts, which were affected by each drug individually, was different. The majority of the metabolic transcripts affected by Clozapine related to nucleotides, amino acid metabolism, phosphorus metabolism, lipid and carbohydrate metabolism. Haloperidol was found to have a similar effect to Clozapine with respect to the number of transcripts involved in nucleotide metabolism, phosphorus and lipid metabolism, but, interestingly, not a single gene having a function in carbohydrate metabolism was identified. On the other hand genes encoding proteins involved in sulphur and heterocyclic metabolism were affected by Haloperidol only. Clozapine, but not Haloperidol changed expression of three metabolic transcripts in the Wnt/beta-catenin and GSK3b pathways (i.e. Thymoma viral protooncogen AKT1, Mitogen activated kinase 9 MapK9 and Protein Kinase C gamma Prkcc) which suggests that the antipsychotic and metabolic effects of clozapine could be mediated by a similar mechanism as those of lithium and insulin which also affect the GSK3b pathway. Interestingly AKT1 has recently been reported as a candidate gene for schizophrenia and Clozapine has been reported to down regulate GSK3b in vitro. Several genes involved in Ubiquitin metabolism including the Ubiquitin specific protease 9 usp9, which has been previously reported to be differentially expressed in the post mortem schizophrenic brain (Vawter et al., 2001) also showed changes in response to antipsychotics in our study. These data should draw the attention to the fact that post mortem gene expression studies should be interpreted with caution. P15.11 EXPRESSION ANALYSIS OF THE TRYPTOPHAN HYDROXYLASE GENES TPH1 AND TPH2 IN POST MORTEM HUMAN BRAIN SAMPLES Zill P,1 Bu¨ttner A,2 Eisenmenger W,2 Ackenheil M,1 and Bondy B.1 1 Psychiatric Clinic of University, Munich, Germany 2 Institute of Legal Medicine, Munich, Germany The Tryptophan hydroxylase (TPH) gene is widely discussed as a candidate gene in many psychiatric disorders, which implicate disturbances of the serotonergic system. Numerous studies about genetic

Abstracts variants or different expression levels of the TPH gene in depression, schizophrenia, alcoholism, drug abuse, aggression and sucidality have been reported, but non of them have led to convincingly confirmed results. Until recently only one gene encoding TPH (TPH1) was described for vertebrates. Recently a second TPH isoform was identified in mice, designated as TPH2 or neuronal TPH, which was shown to be exclusively present in the mouse brain. We were recently able to demonstrate that TPH2 is also expressed in the human brain, but not in peripheral tissues such as heart, lung, kidney, duodenum, liver and adrenal gland. The aim of the present study was to compare the expression levels of the TPH1 and TPH2 genes in the human brain. We analyzed TPH1 and TPH2 mRNA levels in frontal cortex, thalamus, hippocampus, amygdala, cerebellum, hypothalamus, median raphe nuclei and pituitary by quantitative real time RT-PCR. The probes were obtained from 10 individuals, who died on sudden death due to cardiovascular failure (n ¼ 5), on drug abuse (n ¼ 4) and due to suicide (n ¼ 1). The results of our study show that the mRNA expression of TPH1 and TPH2 is nearly identical except in the raphe nuclei and the pituitary. In raphe nuclei, the major locus of the serotonin producing neurons, TPH2 was the predominately gene (up to 10 fold increase), whereas in the pituitary TPH1 mRNA was expressed at up to 20 fold higher levels. Although the tissues were obtained from individuals who died on different causes of death, respectively suffered from different disorders there were no interindividual differences. Our finding suggest a priority role of the TPH2 isoform in 5-HT related behavior and opens up new research strategies for the analysis of psychiatric disorders with altered serotonergic function. Additionally, numerous genetic studies with the TPH1 gene in relation to psychiatric disorders, especially in sucidality have to be re-evaluated using genetic variants within the TPH2 gene. In conclusion further studies of this new TPH2 isoform in psychiatric are needed to clarify the role of TPH2 in the pathophysiology of psychiatric diseases. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine.’’

P15.12 SCANNING OF GENOMIC DIFFERENCES IN MONOZYGOTIC TWINS DISCORDANT FOR SCHIZOPHRENIA, USING CA REPEAT MARKERS Tasaki S,1 Imamura A,2 Hirayama R,2 Kikuchi T,2 Tsujita T,3 Okazaki Y,4 and Ozawa H2 1 Goto-Tyuou Hospital, Nagasaki, Japan 2 Nagasaki University Graduate School of Biomedical Sciences, Nagasaki, Japan 3 Hironaka Hospital, Nagasaki, Japan 4 Department of Neuropsychiatry, Mie University School of Medicine, Mie, Japan Monozygotic (Mz) twins are thought to be genetically identical, so if they are discordant for schizophrenia, it is considered to be due to environmental factors. Recently several researchers reported Mz twins discordant for some phenotype caused by genomic differences or epigenetic modification, such as triplet repeat expansion, X-inactivation, and uniparental disomy (UPD). UPD is the genetic phenomenon in which both alleles are derived from same parents and caused genetic imprinting-related disease. To search for these differences in Mz twins, we analyzed their genome using 305 CA repeat markers, which covered all autosomal and X chromosomes. The distance between markers was about 10 cM. The subjects were five pairs of monozygotic twins discordant for DSM-IV schizophrenia (all pairs were male, mean age 42.6 years, all twins had been discordant more than ten years) and 6 normal control monozygotic twins (4 male twins and 2 female twins, mean age 36.0 years). The genome DNA samples were extracted from their lymphocytes by the phenol method. PCR products were analyzed by Pharmacia ALF DNA sequencer. Although we obtained no differences between co-twins, we found that the occurrence of epigenetic events such as UPD can be detected by this method, and we conclude that further study is needed using other markers.

P15.13 SIMILAR PROFILES IN HUMAN PERIOD1 GENE EXPRESSION IN PERIPHERAL MONONUCLEAR AND POLYMORPHONUCLEAR CELLS Mishima K, Kusanagi H, Satoh K, Echizenya M, and Katoh T Div of Neuropsychiatry, Dep of Neuro and Locomotor Science, Akita Univ School of Medicine, Akita, Japan

147

The suprachiasmatic nucleus (SCN) is the site of the circadian pacemaker in mammals, which generates circadian fluctuation of various physiological functions including mood, sleep-waking, and psychoneuroimmunological functions. Clock genes compose the translation-transcription feedback loop in the SCN and also in many peripheral tissues/cells, including liver, heart, kidney, and skin, suggesting that peripheral clocks are entrained by a ‘‘master’’ clock localized in the SCN. In other words, changes in transcription of clock genes could influence the circadian phase and/or amplitude of a physiological process at the single-cell level. In the present study, we examined expression of the human homolog of period1 (hPer1) in peripheral mononuclear cells (MNCs) and polymorphonuclear neutrophils (PMNs) in seven healthy young male volunteers (mean age, 21.0 yrs, range, 19–24 yrs) under modified constant routine conditions in which the masking effects of physical movement, posture, calorie intake, clothing, ambient temperature, and environmental light intensity were controlled. hPer1 expression and serum melatonin secretion levels were determined every 2 hours for 26 hours (13 timepoints) by real-time PCR with gene-specific hybriprobes in MNCs and PMNs and radioimmunoassay. Significant daily variations in hPer1 expression were observed in MNCs (F ¼ 4.451, df ¼ 12, P < 0.001) and PMNs (F ¼ 5.336, df ¼ 12, P < 0.001) and peak transcription occurred between 07 h and 09 h. We found a significant positive correlation between the acrophase of the melatonin rhythm, a reliable marker of SCN oscillatory phase, and hPer1 expression in MNCs (r ¼ 0.939, P < 0.01) and a weak correlation with PMNs (r ¼ 0.657, P ¼ 0.15). In addition, acrophases of hPer1 expression in MNCs and PMNs had a significant reciprocal positive relation (r ¼ 0.803, P < 0.03). Despite the similar acrophase of hPer1 expression rhythm, the present subjects showed significant daily variation of peripheral MNCs and PMNs distribution with opposite acrophases. Various peripheral WBC subsets have been reported to have widely diverse circadian phases in cellular functions, including circulating cell counts, cytokine production, lymphocyte blastogenesis, and phagocytic activity. We expected that MNCs and PMNs would exhibit different acrophases of clock gene expression rhythms given that these two fractions of peripheral blood cells have divergent timing in cellular functions. The present findings indicate that clock gene activity could be preserved across different peripheral blood cell types, cell functions and support the assumption that peripheral clocks are entrained by the SCN. P15.14 CANDIDATE GENES, PATHWAYS AND MECHANISMS FOR BIPOLAR (MANIC-DEPRESSIVE) AND RELATED DISORDERS: AN EXPANDED CONVERGENT FUNCTIONAL GENOMICS APPROACH Ogden CA, Rich ME, Schork NJ, Paulus MP, Geyer MA, Lohr JB, Kuczenski R, and Niculescu AB Department of Psychiatry, UC San Diego/VA Medical Center San Diego, La Jolla, United States Identifying genes for bipolar mood disorders through classic genetics has proven difficult. Here we present a comprehensive convergent approach that translationally integrates brain gene expression data from a relevant pharmacogenomic mouse model (involving treatments with a stimulant- methamphetamine, and a mood stabilizer—valproate), with human data (linkage loci from human genetic studies, changes in post-mortem brains from patients), as a bayesian strategy of cross-validating findings. Topping the list of candidate genes we have DARPP-32 (dopamine- and cAMP- regulated phosphoprotein of 32 kilodaltons) located at 17q12, PENK (preproenkephalin) located at 8q12.1, and TAC1 (tachykinin 1, substance P) located at 7q21.3. These data suggest that more primitive molecular mechanisms involved in pleasure and pain may have been recruited by evolution to play a role in higher mental functions such as mood. The analysis also revealed other high probability candidates genes (neurogenesis, neurotrophic, neurotransmitter, signal transduction, circadian, synaptic, and myelin related), pathways and mechanisms of likely importance in pathophysiology. P15.15 AFFECTIVE DISORDERS, THE STATE OF MIND: EXPRESSION PROFILING IN POST-MORTEM TISSUE Lefebvre R, Weckx S, Van Broeckhoven C, and Del-Favero J Molecular Genetics (VIB8), University of Antwerp, Antwerp, Belgium Bipolar disorder (BP) is a severe psychiatric condition characterized by recurring episodes of mania and depression and is devastating for both the patient and his/her environment. The lifetime prevalence of BP is

148

Abstracts

between 0.5–1.5%. Schizophrenia (SZ) is like BP a severe psychiatric condition, schizophrenic patients follow an unpredictable course of subsequent psychotic seizures, possibly combined with delusions, hallucinations, disorganized speech, and grossly disorganized or catatonic behaviour. The global lifetime prevalence for SZ is 0.55%. Worldwide, a large effort is made to identify the disease causing genes and to clarify the aetiology for both inflictions. For both disorders a broad spectrum of associated loci are published and just recently, a number of robustly reproduced genes were published. Nevertheless, many questions concerning functional candidates and origin of disease pathways remain. To provide more functional related data, we are employing a strategy to discover differential expressed genes for BP and SZ in a tandem research approach on post-mortem brain tissue. Herein the powerful Suppression Subtractive Hybridization (SSH) technique is combined with the efficient micro-array technique to elucidate differential expressed genes between patients and controls. Whereas micro-arrays allows the parallel analysis of ten thousands of genes, SSH enables the detection of differentially expressed low abundant and unknown genes. SSH was performed on RNA pools derived from hippocampal and frontal lobe tissue. The micro-array analysis was performed on 105 prefrontal cortex tissue samples comprising 35 BP, 35 SZ and 35 control individuals. The Stanley Brain Foundation provided both samples. The SSH results from post-mortem brain tissue between BP and unaffected control individuals confirm the newly proposed pathways concerning calcium regulation and glutamergic pathways. Next to these results, a set of genes, such as NUCKS, RyR3, ANT2 and SERPINA1 and SERPINA3, not previously reported in the light of BP showed a differential expression. For SZ we could confirm previously reported differential expressed genes and discovered other new candidates.Overall, these analyses will allow us to identify functional candidate genes, determine BP and SZ related disease pathways, provide us with valuable biomarkers and will render information that can be used for development of new drug targets.

P15.16 NEUROLEPTIC-INDUCED DIFFERENTIAL GENE EXPRESSION IN HUMAN AND RAT NEURONAL CELL-LINES INVESTIGATED BY MICROARRAY TECHNOLOGY AND REAL-TIME PCR Fernø J, Ræder MB, Skrede S, Vik-Mo AO, Glambek M, Løvlie R, Stansberg C, and Steen VM Center for Medical genetics and Molecular Medicine, University of Bergen, Bergen, Norway Neuroleptic-induced differential gene expression in human and rat neuronal cell-lines investigated by microarray technology and real-time PCR Fernø J, Ræder MB, Skrede S, Vik-Mo AO, Glambek M, Løvlie R, Stansberg C, and Steen VM Dr. Einar Martens Research Group for Biological Psychiatry and Bergen Mental Health Research Center, Center for Medical Genetics and Molecular Medicine, University of Bergen, Haukeland University Hospital Helse Bergen HF, Norway Antipsychotic drugs are used in the treatment of psychiatric illnesses, especially schizophrenia and other psychotic disorders. These drugs are known to block D2-like dopamine receptors in various brain regions, but the precise mechanisms involved in the therapeutic effect and development of side-effects are still to be established. To further elucidate the biological effects of antipsychotic drugs, we treated cultured human and rat neuronal tissue-related cell-lines with either typical (haloperidol and chlorpromazine) or atypical (clozapine) antipsychotics for different time intervals (1 hr-7 days). Drug-induced changes in the global geneexpression were investigated by microarray technology, using the FairPlayTM aminoallyl labeling kit, followed by hybridization of the labeled samples on human and rat cDNA and oligo glass-slides (provided by the The Norwegian Microarray Consortium, www. mikromatrise.no). The scanning was performed on Axon GenePix 4000B or Agilent Microarray scanners, and the image- and data analysis was done by GenePix Pro and J-Express software. Real-time PCR technology (ABI PrismTM 7700 Sequence Detector) was used for independent verification of subsets of up- and down-regulated genes. The differentially expressed genes, which will be presented at the congress, are involved in several biological systems including signal transduction and membrane-structure. Our data may improve the understanding of both the therapeutic action and adverse effects related to the use of antipsychotic drugs.

P15.17 ANTIPSYCHOTIC-INDUCED CHANGES IN GLOBAL GENE EXPRESSION: DEMONSTRATION OF NOVEL MECHANISMS OF DRUG ACTION BASED ON CLUSTERING AND VERIFICATION OF DIFFERENTIALLY EXPRESSED GENES Fernø J, Ræder MB, Skrede S, Vik-Mo AO, Glambek M, Løvlie R, Jørgensen HA, Stansberg C, and Steen VM Center for Medical Genetics and Molecular Medicine, University of Bergen, Bergen, Norway Treatment of schizophrenia with antipsychotic drugs may reduce symptoms and increase cognitive functions and quality of life. Classical antipsychotics (e.g., chlorpromazine, haloperidol) mainly reduce positive symptoms, whereas atypical drugs (e.g., clozapine) also seem to improve negative symptoms and cognition. In general, the antipsychotic effect correlates with blockage of dopamine D2-like (DRD2, DRD3 and DRD4) receptors in mesolimbic-frontal brain regions, but other targets are also likely to be involved, especially with respect to the atypical effects. Unfortunately, most classical antipsychotic drugs are hampered by a marked tendency to induce extrapyramidal side-effects. Atypical antipsychotics generally have fewer adverse effects, but marked weight gain, metabolic syndrome, cardiovascular disorders and diabetes have become a large problem for some of the drugs. In order to further explore the mechanisms of action underlying the therapeutic efficacy and side-effects of antipsychotic drugs, we have examined and identified drug-induced changes in global gene expression, using DNA microarray technology in various neuronal- and nonneuronal cell cultures and living rats. In short, isolated RNA samples were labeled with FairPlayTM aminoallyl kit, followed by hybridization of the samples on human and rat cDNA and oligo glass-slides (provided by the The Norwegian Microarray Consortium, www.mikromatrise.no). The scanning was performed on Axon GenePix 4000B or Agilent Microarray scanners, and the image- and data analysis was done by GenePix Pro and J-Express software. Real-time PCR technology (ABI PrismTM 7700 Sequence Detector) was used for independent verification of subsets of up- and down-regulated genes. Based on the clustering and verification of differentially expressed genes (detailed data will be presented at the congress), we have identified a novel mechanism of drug action which may be of importance for both the therapeutic efficacy and adverse effects related to the use of antipsychotic drugs.

P15.18 APPLICATION OF EMERGING GENOMICS AND EPIGENOMICS TECHNOLOGIES TO THE ANALYSIS OF 22Q11 DELETION SYNDROME Urban AE, Pan X, Tang Y, Lian J, Szekely A, Lian Z, Euskirchen G, Nelson K, Goetsch-Martone R, Rinn J, Bertone PN, Gerstein M, Snyder M, Weissman SM, and Cubells JF Yale University, New Haven, United States Schizophrenia is a complex behavioral disorder believed to result from abnormal neural development. The etiology of this condition is unknown, but a major role for genetic factors can be assumed. The 22q11 deletion syndrome (22q11DS) is considered to be a promising point of entry in the study of schizophrenia on the molecular level. The syndrome, which is characterized by various developmental abnormalities, results from loss of chromosomal material in the q11 region of a single copy of chromosome 22, with 25–30% of patients developing schizophrenia. Several excellent studies have led to great progress in understanding the mechanisms leading to disease in 22q11DS. But many questions still remain. For example there is no apparent correlation between the highly variable phenotype and the size of the deletion. And for most of the many genes that might be affected by the deletion there is very little known about what role they might play in neural development. We plan to take advantage of new and emerging methods from the fields of genomics and epigenomics address some of these questions. We have constructed a genomic tiling microarray covering all non-repetitive genomic sequence of human chromosome 22q, consisting of approximately 20 000 PCR fragments (averaging 800 bp), representing 17 Mb of chromosomal sequence. We have already used this tiling microarray for experimental gene annotation by expression analysis (Rinn et al., Genes Devel., 2003, vol. 17, p. 529) and for transcription factor binding site mapping by chromatinimmunoprecipitation (ChIP) followed by array hybridization (ChIPchip, Martone et al., PNAS, 2003, vol. 100, p. 12247, Euskirchen et al.,

Abstracts MCB, 2004, vol. 24, p. 3804). Now we are planning to use this technology to analyze the molecular genetics and epigenetics of 22q11DS on several levels. The microarray analysis should allow us to study changes in gene activity and gene-dosage compensation on chromosome 22 in 22q11DS. It should also be possible to carry out highresolution fine-mapping of deletion breakpoints in an efficient manner. Furthermore we want to test the hypothesis that deletions such as those found in 22q11DS can cause changes in gene activity outside of the deleted region via epigenetic mechanisms. To this end we are currently extending the ChIP-chip approach to map the binding of methyl-cytosine-binding proteins and the modification of histones along 22q. When combined with classical methods of detecting DNA methylation this should allow us to analyze the chromatin status along the chromosome as well as to make inferences about DNA methylation.

P15.19 GLYCOGEN SYNTHASE KINASE 3 BETA AS A LIKELY TARGET FOR THE ACTION OF LITHIUM ON CIRCADIAN CLOCKS Jain S1 and Padiath Q2 1 National Institue of Mental Health and Neurosciences, Bangalore, India 2 National Center for Biological Sciences, Bamgalore, India Lithium is one of the most commonly used drugs in the prophylaxis and treatment of bipolar disorder, but the mechanisms underlying its therapeutic action are still unclear. Together with its mood stabilizing effects, lithium is also known to influence the circadian clocks of several organisms including man. Circadian rhythms are altered in patents with bipolar disorder and it is believed that these rhythms may play an important role in disease mechanisms. It is therefore possible that some of the therapeutic actions of lithium may be related to its effect on circadian clocks. Identifying the targets for lithium’s action on circadian clocks would therefore be important both for understanding the mechanisms of its therapeutic effect and also in further understanding disease mechanisms in bipolar disorders. Using Drosophila melanogaster as a model system, we show that the long-term administration of lithium results in lengthening of the free-running period (t) of circadian locomotor activity rhythm of flies in constant darkness (DD). This effect occurs at concentrations similar to the plasma levels of lithium used in the treatment of bipolar disorder. The lithium treated flies also show reduced activity of one of the previously reported targets of lithium action, Glycogen Synthase Kinase 3b (GSK 3b). GSK 3b has been shown to be involved in the regulation of circadian clocks, as the down regulation of this protein results in an elongation of t. The t elongation resembles the effect seen with lithium administration in a number of organisms including man, and taken together with the earlier observations, our results suggest that lithium inhibits the activity of GSK 3b to produce its effect on circadian clocks. P16.1 X-LINKED GENETIC MECHANISMS INFLUENCING COGNITION IN MICE Davies W,1 Isles AR,1 Burgoyne PS,2 and Wilkinson LS1 1 The Babraham Institute, Cambridge, United Kingdom 2 MRC National Institute for Medical Research, London, United Kingdom Genomic imprinting is a parent-of-origin dependent epigenetic mechanism through which a subset of mammalian genes are expressed preferentially from one parental allele. Imprinted gene dysfunction has been implicated in a number of diverse disorders associated with aberrant cognition including Angelman and Prader-Willi syndromes, autism, schizophrenia and bipolar disorder. We tested the hypothesis that X-linked imprinted genes (and by default X-monosomy) may affect cognition in a 39,XO mouse model. Our data show that mice inheriting their single X chromosome maternally (39,XmO) are impaired on a task of behavioural flexibility (reversal learning) relative to mice inheriting their single X chromosome paternally (39,XpO), they are also impaired relative to 40,XX females. This suggests the presence of one or more Xlinked genes that are only (or predominantly) expressed from one parental allele and that influence behavioural flexibility. We also show that 39,XO mice display impaired visuospatial attention relative to their 40,XX counterparts on a 5-choice serial reaction time task, we propose that this is due to haploinsufficiency for one or more X-linked genes that escape X inactivation. The identification of candidate genes underlying the aforementioned phenotypes is ongoing through micro-

149

array analysis and molecular genetics techniques respectively. Our data will have relevance for determining the molecular bases of disorders of behavioural flexibility and attention such as autism, schizophrenia and ADHD, and may partially explain the sex bias in prevalence and severity of these disorders.

P16.2 WISTAR-ZAGREB 5HT RAT: POSSIBLE ANIMAL MODEL FOR AUTISM Hranilovic D, Bordukalo-Niksic T, Cicin-Sain L, and Jernej B Rudjer Boskovic Institute, Zagreb, Croatia Autistic disorder is a behavioural sindrom characterized by impaired communication, impaired social interactions and repetitive interests and behaviours. Increased platelet serotonin level, as well as positive effect of serotonin reuptake inhibitors on some symptoms in autistic patients, have implicated the role of serotonergic neurotransmission in etiopathogenesis of autism. There are several animal models for autistic disorder, obtained by physical lesions of brain regions, by early intoxication/infection, or by gene knock-out. However, more specific model for this disorder would be overexpression of 5HT transporter, mimicking more closely the situation in hyperserotonemic patients. By breeding selection for the extreme values of platelet 5HT level and platelet 5HT uptake, two sublines of Wistar-derived rats, with constitutionally high or low values of both parameters, have been developed (termed ‘‘high 5HT’’ subline and ‘‘low 5HT’’ subline, respectively). Molecular genetic studies demonstrated that the observed differences were underlain by differences in expression of platelet 5HT transporter between the sublines at both, protein and mRNA levels. Earlier behavioral tests indicated higher level of anxiety-related behavior in ‘‘high 5HT’’ subline in comparison to ‘‘low 5HT’’ subline. In this work we investigated social interaction as a measure of social anxiety. Tendency to explore novel animal was measured in open field as time spent in active contact within 10 minutes. Animals from ‘‘high5HT’’ subline showed significantly lower social interaction than the animals from ‘‘low-5HT’’ subline: 216  36 vs. 277  39 seconds spent in active contact (P < 0.001, N ¼ 16 pairs per subline). The results demonstrate higher social anxiety in hyperserotonemic rat and implicate the use of Wistar-Zagreb 5HT rat as an original animal model of autistic disorder. P16.3 MICROARRAY ANALYSIS OF ALTERED GENE EXPRESSION IN THE MOUSE BRAIN AFTER TREATMENT WITH LITHIUM CHLORIDE OR SODIUM VALPROATE Chetcuti A,1 Adams L,2 and Schofield PR2 1 Neuroscience Institute of Schizophrenia and Allied Disorders, Sydney, Australia 2 Garvan Institute of Medical Research, Sydney, Australia Bipolar disorder is a common psychiatric disorder that affects an estimated 0.8–1.6% of the population. Patients experience mood swings, full-blown manic episodes and syndromal major depressions. In addition, patients also have a substantial lifetime risk of suicide. Breakthroughs in the treatment of bipolar disorder have progressed slowly. To date limited medications are used in the treatment of bipolar disorder. Lithium chloride has been used successfully since the 1950s, yet the underlying mechanism of its action is still not known. Although lithium has remained the primary treatment of bipolar disorder, newer drugs such as valproate are frequently prescribed. The fact that these two chemically distinct drugs display similar anti-manic effects on bipolar patients has prompted us to determine whether these drugs act via overlapping biochemical pathways. Our hypothesis is that genes commonly altered by lithium and valproate would be likely to have an important role in bipolar disorder. Our study involved the use of a mouse model of drug action. Either lithium chloride or sodium valproate were administered daily via intraperitoneal injection. Control mice received daily injections of saline. Following seven days of treatment, blood and brain tissue were collected. The serum concentration of lithium or valproate were measured. Whole brain total RNA from control or drug treated animals were processed, pooled and hybridised to Affymetrix Murine Genome U74Av2 GeneChip microarrays. The expression profile for 12,488 transcripts was determined using Affymetrix Microarray Analysis Suite (MAS) software. Data analysis was undertaken and differentially expressed transcripts were determined. Using 2-fold difference relative to control mice, 5 and 372 transcripts were up-regulated after valproate or

150

Abstracts

lithium treatment, respectively. A total of 1 and 237 transcripts were down-regulated after valproate or lithium treatment, respectively. A subset of genes was selected for confirmation using Rotor-gene 3000 (Corbett) quantitative PCR. Animal models are being used as tools to further understand the biology and biochemistry of bipolar disorder. The identified genes may represent novel targets for the development of new compounds for the treatment of bipolar disorder.

P16.4 SEQUENCE ANALYSIS OF ADHD CANDIDATE GENES IN THE SHR AND WKY RAT STRAINS Mill J,1 Sagvolden T,2 and Asherson P1 1 Institute pf Psychiatry, London, United Kingdom 2 University of Oslo, Oslo, Norway The Spontaneously Hypertensive Rat (SHR) shows a number of behaviours that closely parallel those seen in children with ADHD. These include motor hyperactivity, excessive responses under a fixedinterval/extinction schedule, difficulty in acquiring operant tasks, and increased sensitivity to immediate behavioural reinforcement. As in children with ADHD, the behavioural and cognitive deficits in the SHR are responsive to stimulants, including d-amphetamine and d,Imethylphenidate. Numerous studies have shown that dopaminergic neurotransmission is altered in the SHR, strongly implicating this system in the aetiology of ADHD. Human genetic studies have found associations between several dopaminergic genes and ADHD. We sequenced three candidate dopaminergic genes (DRD2, DRD4, and DAT1) in the SHR and a non-hyperactive control strain, the WKY. No between-strain sequence differences were found in either DRD2 or DRD4, but several variations were found in the DAT1 gene. Future work will examine whether these variations affect levels of DAT1 expression, and whether they contribute to the SHR phenotype.

P16.5 SEQUENCE AND POLYMORPHISMS OF THE REGULATOR OF G PROTEIN SIGNALING 4 (RGS4) GENE IN VERVET MONKEYS, A CANDIDATE GENE FOR QUANTIFIABLE BEHAVIORAL DIMENSIONS ASSOCIATED WITH PSYCHOPATHOLOGY Trakadis YJ,1 Desbiolles C,2 Ervin FR,3 and Palmour RM1 1 Department of Human Genetics, McGill University, Montreal, Canada 2 Department of Biology, McGill University, Montreal, Canada 3 Behavioural Sciences Foundation, Estridge Estate, Saint Kitts and Nevis Regulators of G-protein coupled signaling (RGS) accelerate GTP hydrolysis and consequently influence signal termination. The RGS-4 gene has recently been reported to be implicated in schizophrenia, Alzheimer’s disease and addictions. Because of the central role of these proteins as regulators of cell signaling, RGS-4 and other RGS genes may be significantly over- or under-expressed in a wide range of neuropsychiatric disorders. Non-human primates are the closest phylogenetic relatives to humans and have been reported to resemble humans in terms of development, neurophysiology, neuroanatomy and behavior. Consequently, the vervet monkey, a non-endangered species with substantial genetic and behavioral documentation, would appear to be an excellent model in which to explore the role of RGS4 gene and especially its potential role in the regulation of dopamine, serotonin and glutamate networks. In this study, the vervet RGS-4 gene was cloned and characterized using molecular and bioinformatic tools. Both the coding and non-coding regions of the RGS4 gene in vervet monkeys, along with 2,500 bases in the promoter region and the long 30 UTR (1200 bases) were sequenced on a CEQ 8000 genetic analysis system (Beckman Coulter). When compared to the known human gene, the obtained vervet sequence overall showed great similarity {>90%}. Thereafter, SNPs in the proximal promoter, exon 1 and the first 450 bp of intron 1 were identified by direct seqencing of 8 unrelated individuals. Two of the identified SNPs (-14A/G, 632 A/T) were genotyped in 100 juvenile vervets phenotyped for personality traits, including social isolation (Desbiolles et al., 2004, submitted). Although preliminary association analysis fails to attain statistical significance, the sample size is small. Additional genotyping of phenotypically defined individuals is in progress.

P16.6 HYPERACTIVITY, IMPAIRED LEARNING, AND AN ABNORMAL RESPONSE TO METHYLPHENIDATE IN MICE WITH A MUTATION IN THE THYROID RECEPTOR ß GENE Siesser WB, Cheng S-Y, and McDonald MP 1 Vanderbilt University, Nashville, United States 2 National Cancer Institute, Bethesda, United States The thyroid receptor ß PV knock-in (TRßPV KI) mice express a mutant TRß allele derived from a patient with the rare genetic disorder Resistance to Thyroid Hormone (RTH). The mutant TRß protein has reduced binding to thyroid hormone, resulting in an impairment of thyroid hormone’s negative feedback loop. Among other features, the majority of RTH patients meet the diagnostic criteria for Attention Deficit-Hyperactivity Disorder (ADHD). The TRßPV KI mice are hyperactive in the open field, and this hyperactivity persists over multiple sessions. On a measure of sustained attention, the TRßPV KI mice have a learning deficit but are ultimately able to reach the same baseline performance as wild-type littermates. The TRßPV KI mice are not impaired on learning tasks that do not tax sustained attention, suggesting that the observed impairment may be the result of abnormal sustained attention. Treatment with methylphenidate (MPH) improves the performance of the TRßPV KI mice at doses that impair the performance of wild-type littermates. This effect is dependent in part on their baseline performance. On a measure of motivation, the TRßPV KI mice perform similarly to wild-type littermates, but MPH increases the motivation of the TRßPV KI mice at doses that reduce the motivation of the wild-type mice. This finding suggests that MPH is improving the performance of the TRßPV KI mice on the measure of attention by enhancing their motivation, rather than their attention, a hypothesis that has been proposed for ADHD children. Future studies will assess changes in the catecholaminergic systems that may be responsible for their behavioral phenotype. The TRßPV KI mice provide a useful model for studying hyperactivity and the effects of MPH on sustained attention.

P16.7 TRANSGENIC MICE EXPRESSING A HUMAN MUTANT ß1 THYROID RECEPTOR ARE HYPERACTIVE, IMPULSIVE, AND INATTENTIVE McDonald MP,1 Siesser WB,1 and Cheng SY2 1 Vanderbilt University Medical Center, Nashville, United States 2 National Cancer Institute, Bethesda, United States Attention deficit hyperactivity disorder (ADHD) is the most commonlydiagnosed childhood psychiatric disorder. We have found that a transgenic mouse bearing a human mutant thyroid receptor (TRß1) expresses all of the defining symptoms of ADHD—inattention, hyperactivity, and impulsivity—as well as a ‘‘paradoxical’’ response to methylphenidate. As with ADHD, the behavioral phenotypes expressed by the TRß transgenic mice are dynamic, and sensitive to changes in environmental conditions, stress, and reinforcement. TRß transgenic mice are euthyroid except for a brief period during postnatal development, but the behavioral phenotypes and paradoxical response to methylphenidate persist into adulthood. Thus, like the vast majority of children with ADHD, the TRß transgenic mice exhibit the symptoms of ADHD in the complete absence of thyroid abnormalities. This demonstrates that even mild, transient perturbations in developmental thyroid homeostasis can have long-lasting behavioral and cognitive consequences, including producing the full spectrum of symptoms of ADHD. P16.8 IDENTIFICATION OF INFORMATIVE LINES AND PROVISIONAL QTL MAPPING OF AMPHETAMINE INDUCED LOCOMOTION IN RECOMBINANT CONGENICS LINES OF MICE Torkamanzehi A, Boksa P, Ayoubi M, Fortier M-E, Skamene E, Rouleau G, and Joober R Douglas Hospital Research Centre, Montreal, Canada Genetic predisposition plays a major role in the etiology of schizophrenia (SCZ). It is likely that multiple genes, each with small effects, coupled with environmental factors, mediate the disorder. The aim of this study was to map and identify, in a mouse model, markers associated with the behavioral trait amphetamine (AMPH) induced locomotor activity that is relevant to schizophrenia. AMPH-induced

Abstracts hyperlocomotion is a rodent model for the hyperdopaminergic state postulated in SCZ. We used two sets of recombinant congenic lines (RCLs) of mice derived from A/J and C57BL/6J and were informative for 620 microsatellite markers covering the whole genome. We tested 433 animals from 38 congenic and parental lines. Locomotion data was recorded electronically for various aspects of locomotion behavior including total distance traveled (TDIST) which was used in phenotypic and QTL analysis by appropriate statistical methods. We identified two lines in the AcB background (AcB52 and AcB63) which showed significantly elevated TDIST compared to their parental A/J strain. These lines behaved similar to C57BL/6J and showed high TDIST, high peak values and long time to peak. Markers on chromosome 1, 9, 10 and 20 were significantly associated with TDIST in these lines. Within the BcA background, 3 lines significantly deviated from the parental strain. Markers associated with TDIST in these lines mapped to chromosomes 5 and 20. The informative lines harboring genes from either A/J or C57BL/6J, which are implicated in determining variation for the AMPH-induced hyperlocomotion, provide the basis for identifying QTLs involved in this trait. Combining data from this approach and other genetic (mapping data in humans) and functional sources may help to identify suitable candidate genes involved in specific behavioral traits relevant to SCZ.

P16.9 EPIGENETIC DISCORDANCE IN MONOZYGOTIC TWINS WITH RETT SYNDROME: EVIDENCE FOR CRITICAL ROLE OF X CHROMOSOME INACTIVATION IN RETT SYNDROME PHENOTYPE Iourov IY,1 Vorsanova SG,2 Ulas VY,2 Villard L,3 and Yurov YB1 1 National Center of Mental Health RAMS, Moscow, Russia 2 Institute of Pediatrics and Children Surgery Russian Ministry of Health, Moscow, Russia 3 INSERM U491, Faculte´ de Me´decine, Marseille, France Rett syndrome (RTT, MIM 312750) is a severe neuropsychiatric disorder affecting principally females with the incidence of 2.5% in mentally retarded girls in Russia. We describe a monozygotic twin pair being concordant for MECP2 mutation (R255X) and demonstrating a different clinical manifestation of RTT. We have studied X chromosome inactivation patterns in this twin pair. It has been revealed that one girl has a random X chromosome inactivation (activity ratio 56:44) and another skewed X chromosome inactivation (activity ratio 80:20) with the preferential inactivation of paternally derived chromosome X. Analysis of RTT clinical manifestation in these girls has shown that a girl with random X-inactivation has a more severe RTT manifestation comparing to girl with skewed X-inactivation. The main differences of phenotypic RTT features were the disease onset age and presence of seizures in a girl with random X-inactivation being not observed in girl with skewed X-inactivation. The seizure occurring was critical for girl with random X-inactivation and have lead to the death at the age of 8 years. Thus, we conclude X chromosome inactivation to have an influence on the clinical manifestation of RTT. It should be emphasized that mutations in X-linked gene MECP2 associated with RTT are sporadic in 99% of cases being located on paternally derived chromosome X. Therefore in the present case the X chromosome with mutation were inactivated in 80% of cells. It is in accordance with the clinical manifestations in a girl with skewed X-inactivation comparing to girl with random X-inactivation. In conclusion it should be noticed that in some cases of RTT X chromosome inactivation patterns could play a significant role in survivability of RTT females and the study of X chromosome inactivation patterns is of importance for genetic investigation of neuropsychiatric diseases. Supported in parts by INTAS grant.

P16.10 THE ROLE X CHROMOSOME INACTIVATION AND MECP2 MUTATIONS PLAY IN PATHOGENESIS OF RETT SYNDROME: EPIGENETIC AND MOLECULAR GENETIC STUDIES OF RUSSIAN COHORT Vorsanova SG,1 Iourov IY,2 Ulas VY,1 Villard L,3 Gianti L,4 Giovanucci-Uzielli ML,4 and Yurov YB2 1 Institute of Pediatrics and Children Surgery Russian Ministry of Health, Moscow, Russia 2 National Center of Mental Health RAMS, Moscow, Russia 3 INSERM U491, Faculte´ de Me´decine, Marseille, France 4 University of Florence, Florence, Italy

151

Rett syndrome (RTT) is a severe neurodevelopmental disorder with the incidence of 2.5% in mentally retarded girls in Russia. MECP2 gene (Xq28) is considered as a primary cause of the disease. We have screened 40 patients (39 females and 1 male with classical form of RTT) for the MECP2 mutations and 84.6% (33 from 39) of females were found to have a MECP2 mutation. We have studied X-inactivation patterns in 70 girls with RTT. The incidence of skewed X-inactivation (degree of skewing >80%) was 37% (26 from 70). This value were significantly higher than in control group (6.5%, P < 0,001). Thus we have concluded the skewed X-inactivation to be considered as a common feature of RTT. The most common hypothesis of X chromosome inactivation skewing is that cells having a chromosome X with mutation being active had a selective disadvantage comparing to cells with active chromosome X without mutation. For proving this mechanism taking place in RTT we have carried out the study of preferentially inactivated chromosome X determination by assessing the X-inactivation in RTT females and their mothers. About 33% (informative for X-inactivation assay pairs) had a maternal X chromosome to be preferentially inactivated. The mechanism proposed is not describing this epigenetic phenomenon in RTT. Basing on previous studies of MeCP2 protein functions we have proposed two hypotheses that could lead to the explanation of skewed X-inactivation higher incidence in RTT. First hypothesis is based on the transcriptional regulation function of MeCP2 protein and explains the occurrence of skewed X-inactivation in RTT by the unstable transcription caused by MECP2 mutations probably leading to biallelic expression of some X-linked genes as well as to skewed X-inactivation. Second hypothesis is based on the fact that MeCP2 represses LINE-1 expression therefore the abnormal expression of these elements of chromosome X (being expressed while it have to be silenced) may lead to skewed X-inactivation. The analysis of preferentially inactivated chromosome X (paternal or maternal origin) allowed us to make genotype and phenotype correlation. We have observed the skewed Xinactivation to occur as in more severe forms of RTT as in milder ones. This is principally due to skewing X-inactivation in both directions (against maternal or paternal chromosome X with mutated MECP2 gene). We have concluded that different skewing direction is compensating the effect of this epigenetic phenomenon in the whole group studied. Thus the individual approach of the describing genetic and clinical peculiarity of RTT females should be recommended for appropriate genotype-phenotype correlations. Supported in part by INTAS grant.

P16.11 INVESTIGATION OF THE TRYPTOPHAN HYDROXYLASE 2 (TPH2) GENE IN BIPOLAR AFFECTIVE DISORDERS Harvey M,1 Shink E,1 Tremblay M,1 Gagne´ B,1 Raymond C,1 Labbe´ M,1 Walther D,2 Bader M,2 and Barden N1 1 CHUL Research Centre, Sainte-Foy, Canada 2 Max Delbru¨ck Center for Molecular Medicine, Berlin, Germany Bipolar affective disorders (BP) are defined as a common mood disorder characterized by manic and depressive episode. It is now well accepted that this pathology occurs from complex interactions between environmental and genetic factors. To date, genetic studies on BP strongly suggest the involvement of several loci with minor to moderate effects instead of a single major susceptibility locus. Hypotheses of dysfunctional neurotransmission and/or neurotransmitter metabolism remain among the most relevant. Serotonin is a major monoaminergic neurotransmitter involved in several brain functions including mood control, sleep, thermoregulation, drug abuse, learning and memory. The serotonergic system has also been implicated in psychiatric disorders, such as affective and anxiety disorders. Although many sites in the serotonergic system may affect its homeostasis the most pertinent targets are its receptors, transporter, and enzymes of its biosynthetic pathway. The first step, and also the rate-limiting step, of serotonin biosynthesis is tryptophan hydroxylation catalyzed by the tryptophan hydroxylase enzyme (TPH). There are two characterized TPH isoforms, TPH1 and TPH2. TPH1 isoform is present in pineal gland, thymus, spleen and gut while TPH2 was detected exclusively in brain. A gene on chromosome 12q21 encodes the human TPH2. The long arm of chromosome 12 has been reported as a serious BP disorder susceptibility locus. In this work we report SNP-based association study using CLUMP in a case-control sample (bipolar affected ¼ 213, controls ¼ 214) from the Saguenay-Lac-St-Jean region of Quebec. The analyzed polymorphisms were selected following a mutation analysis

152

Abstracts

on coding sequences and intron-exon junctions using 16 affected individuals. We found and genotyped five SNPs, of which two were synonymous coding SNPs. Individual SNP analysis did not suggest any significant allelic or genotypic association at the 5% level. However, haplotype analysis gave some support (P ¼ 0.018) for association between TPH2 and BP in the Quebec population. Haplotype frequencies and effectives were estimated with the EM algorithm implanted in the cocaphase module of UNPHASE. Further association studies, using other polymorphisms distributed over this genomic area (particularly in regulatory regions of TPH2), as well as in a second population, will be useful to confirm this result.

P16.12 THE EFFECT OF DELETION 22Q11 ON BRAIN ANATOMY: A STRUCTURAL IMAGING STUDY OF CHILDREN AND ADOLESCENTS WITH VELO-CARDIO-FACIAL SYNDROME (VCFS) Campbell LE,1 Stevens AF,1 Daly EM,1 Azuma R,1 Morris RG,1 Karmiloff_Smith A,2 Owen MJ,3 Murphy DGM,1 and Murphy KC4 1 Institute of Psychiatry, King’s College London, London, United Kingdom 2 Institute of Child and Health, University College London, London, United Kingdom 3 University of Wales College of Medicine, Cardiff, United Kingdom 4 Royal College of Surgeons in Ireland, Beaumont Hospital, Dublin, Ireland VCFS is a genetic disorder characterised by borderline learning disability and psychosis, and associated with variably sized deletions of 22q11. While VCFS individuals have high rates of learning disability and psychiatric disorder, there have been few large quantitative studies of the effects of the deletion on brain anatomy. Magnetic resonance imaging of the brain was performed on a GE Signa 1.5 Tesla system to study 39 children (6–16 years) and 26 non-VCFS siblings. Pre-statistical image processing was carried out with Statistical Parametric Mapping (SPM99) software using study specific customised prior probability maps. Between-group statistical inferences of grey and white matter differences was performed using Brain Activation and Morphological Mapping (XBAMM), with an analysis of variance modelling at each intracerebral voxel. A manual tracing method (MEASURE) was used to calculate lobar and regional volumes. Subjects with VCFS had significantly more midline abnormalities such as cavum septum pellucidum/vergae and peripheral white matter hyperintensities. In addition, they had smaller volume of grey matter bilaterally in the cerebellum, brainstem and temporal lobe and indeed smaller volume of white matter in the brainstem, and bilaterally in the cerebellum, temporal, occipital and parietal lobes. Deletion of chromosome 22q11 is associated with brain abnormalities that are most likely neurodevelopmental in origin and may partially explain the high prevalence of learning disability and psychiatric disorder in VCFS. We aim to follow up these children as part of a larger cohort to identify early precursors for the development of schizophrenia as they progress into adulthood.

P16.13 CNS LOCALIZATION OF A NEW PEPTIDE ENCODED BY A CANDIDATE GENE FOR PSYCHIATRIC DISORDERS Bjarkam CR,1 Olsen IM,2 Lunddorf MD,2 Gurling HMD,3 and Mors O2 1 Department of Neurobiology, Institute of Anatomy, University of Aarhus, Aarhus, Denmark 2 Institute for Basic Psychiatric Research, Aarhus University Hospital, Risskov, Denmark 3 Molecular Psychiatry Laboratory, University College of London, London, United Kingdom Recent genetic mapping studies have indicated that chromosome 12q23-q24 may contain risk genes for bipolar affective disorder, attention deficit hyperactivity disorder and Alzheimer disease. Further examination of genes in the chromosome region of interest revealed a new gene expressed in brain-tissue. The aim of the current study was to develop antibodies towards the putative encoded polypeptide and examine its localization in the mammalian CNS by immunohistochemistry. A 12 amino acid sequence was synthesized based on the original polypeptide sequence. The synthesized peptide was coupled to Freund’s adjuvant and used for immunization in two rabbits. After complete immunization serum was extracted, and the derived primary poly-

clonal antibody was immunopurified. Brain tissue from human, pig, rat and mouse was then vibratome sectioned into 50 mm coronal sections and stained using the generated polyclonal primary antibody along with a biotinylated secondary antibody. Specific staining was seen in sections from all species. Staining of neuronal cell bodies was observed in the brainstem raphe nuclei and the hippocampal region, while intense staining of nerve fiber boutons was seen in the basal forebrain.This study indicates that the identified gene associated with psychiatric disease may be of major importance as the derived polypeptide sequence is located in a subpopulation of neurons and nerve fibers, which have a specific distribution in anatomical regions previously correlated to psychiatric disorders. We wish to acknowledge the contribution of Professor Henrik Ewald, who died 20th of January 2004, to this research.

P16.14 CHARACTERISING A GENE PRODUCT FROM A BIPOLAR CANDIDATE REGION ON CHROMOSOME 12 Olsen IM,1 Nyegaard M,1 Bjarkam CR,2 Lundorf MD,1 Børglum AD,3 and Mors O1 1 Centre for Basic Psychiatric Research, Psychiatric Hospital in Aarhus, Aarhus University Hospital, Aarhus, Denmark 2 Department of Anatomy, University of Aarhus, Aarhus, Denmark 3 Department of Human Genetics, University of Aarhus, Aarhus, Denmark A region on chromosome 12q24.3 was previously identified by linkage and association studies as being a susceptibility locus for bipolar affective disorder. Using the UCSC genome browser the region was found to contain several putative genes. Based on the finding that one of the putative genes was found to be expressed in the human brain as well as that of the macaque monkey and that sequencing of three independent human cDNA libraries revealed no longer open reading frame (ORF), we have selected one ORF called ORF48 for our continued studies. Poly- and monoclonal antibodies have been raised against a short peptide sequence encoded by ORF48. Using these antibodies in immunostainings of brain sections from a variety of species including rat, mouse, pig, rabbit and human revealed an interesting and specific staining of brain regions known to be involved in patients with psychiatric diseases. However, it is not known whether the antibodies are specific and whether ORF48 is in fact expressed as a protein. We are characterising the putative protein, ORF48, with respect to expression both at the level of specific brain regions and at the cellular level. Additionally, we are attempting to verify the specificity of the antibodies. This includes analyses by Western blotting, immunoprecipitation, protein identification by sequencing and establishing of exact molecular weight by mass spectrometry, cloning and recombinant expression in human cell lines. Preliminary Western blotting experiments indicate the presence of a protein in rat brain stem of the approximate size expected for ORF48. This suggests that ORF48 is not only a putative protein but an expressed protein. We wish to acknowledge the contribution of Professor Henrik Ewald, who died 20th of January 2004, to this research. P16.15 SEQUENCE VARIATIONS IN HUMAN CASEIN KINASE1EPSILON GENE AND CIRCADIAN RHYTHM SLEEP DISORDERS Takano A,1 Uchiyama M,2 Kajimura N,3 Mishima K,4 Inoue Y,5 Kamei Y,6 Kitajima T,7 Shibui K,2 Katoh M,3 Watanabe T,3 Hashimotodani Y,1 Nakajima T,8 Ozeki Y,9 Hori T,3 Yamada N,9 Toyoshima R,11 Ozaki N,10 Okawa M,9 Nagai K,1 Takahashi K,3 Isojima Y,1 Yamauchi T,11 and Ebisawa T12 1 Division of Protein Metabolism, Institute for Protein Research, Osaka University, Osaka, Japan 2 Department of Psychophysiology, National Center of Neurology and Psychiatry (NCNP), Chiba, Japan 3 Musashi Hospital, NCNP, Tokyo, Japan 4 Department of Psychiatry, Akita University School of Medicine, Akita, Japan 5 Japan Somnology Center, Neuropsychiatric Research Institute, Tokyo, Japan 6 Kohnodai Hospital, NCNP, Chiba, Japan 7 Department of Psychiatry, Fujita Health University School of Medicine, Aichi, Japan 8 Department of Neuropsychiatry, Kyorin University, School of Medicine, Tokyo, Japan

Abstracts 9

153

Department of Psychiatry, Shiga University of Medical Science, Shiga, Japan 10 Department of Psychiatry, Nagoya University Graduate School of Medicine, Aichi, Japan 11 Department of Neuropsychiatry, Saitama Medical School, Saitama, Japan 12 Project Research Division, Research Center for Genomic Medicine, Saitama Medical School, Saitama, Japan

the 14 coding amplicons have had at least one SNP identified. However, only one nonsynonymous SNP has been identified, A559V, which was previously identified in a subject with bipolar disorder. A559V was identified in siblings with ADHD and additional family members are currently being genotyped and ascertained. Supported by: P01 HL56693 (RDB) and 1 F31 MH067472 (to MSM).

Recent studies revealed that functional polymorphisms in human clock genes confer susceptibility to circadian rhythm sleep disorders, such as delayed sleep phase syndrome (DSPS), advanced sleep phase syndrome (ASPS), and non-24-hour sleep-wake syndrome (N-24). A missense mutation in Period2 (Per2) gene, which reduces casein kinase1epsilon(CK1e)-induced phosphorylation of PER2 protein, reportedly causes familial ASPS. We have already reported that a missense variation in Per3 gene, which presumably affect phosphorylation of PER3 protein, increases the risk for the development of DSPS. Another candidate gene for rhythm disorder susceptibility is CK1e, which is known to phosphorylate clock proteins such as PER1, PER2, and PER3, and play a pivotal role in regulating circadian rhythm. In this study, we have analyzed all of the coding exons of human CK1e gene in circadian rhythm sleep disorder patients and control subjects. Using SSCP analysis and D-HPLC analysis followed by direct sequencing analysis of the PCR products, we have found four sequence variations, one of which encoded a missense variation. Allele frequency of the missense variation in sleep disorder patients was significantly lower than that in the healthy control subjects. To analyze the functional alteration induced by the missense polymorphism, the CK1e cDNA with the variation was introduced into E.coli and the expressed enzyme was used for in vitro kinase assay. CK1e with the variation showed increased enzyme activity for various substrates, including PER proteins, compared with the wild-type enzyme. These results indicate that the missense variation of CK1e we found reduces susceptibility to the development of circadian rhythm sleep disorders through alteration of the enzyme activity.

P17.1 CANDIDATE GENE SELECTION FOR DEVELOPMENTAL DYSLEXIA: A BIOINFORMATICS APPROACH Orourke J, Platko JV, and Pauls D 1 Massachusetts General Hospital, Charlestown, United States 2 Tufts University, Medford, United States 3 Harvard University, Cambridge, United States

P16.16 IDENTIFICATION AND CHARACTERIZATION OF SINGLE NUCLEOTIDE POLYMORPHISMS IN THE HUMAN DOPAMINE TRANSPORTER GENE IN CHILDREN WITH ATTENTION DEFICIT HYPERACTIVITY DISORDER Mazei-Robison MS and Blakely RD Vanderbilt University, Nashville, United States Alterations in dopamine transporter (DAT) function have long been implicated as a potential source of pathophysiology in the central nervous system. DAT dysregulation has been postulated to underlie many neuropsychiatric diseases ranging from bipolar disorder to drug abuse to attention deficit hyperactivity disorder (ADHD). To this end, many groups have attempted to genetically link DAT to these conditions via a 30 -untranslated region variable number tandem repeat (30 UTR VNTR). While the functional relevance of the VNTR remains to be established, very little work has focused on the identification of other coding variations in the human dopamine transporter. To date, only three published studies have examined the coding region of DAT for polymorphisms. These studies resulted in the identification of five nonsynonymous single nucleotide polymorphisms (SNPs) V55A, R237Q, V382A, A559V and E602G, the functional consequences of which have not been thoroughly studied. The goals of this project have been to characterize the effect of these SNPs on hDAT function using heterologous cell lines and to identify novel SNPs in hDAT in children with ADHD. Wild-type and mutant hDAT DNA was transiently transfected into COS-7 cells and substrate transport, cell surface expression, pharmacological sensitivity and regulation of transport was examined. The only mutation that differed significantly from wild-type hDAT was V382A, which exhibited a decreased Vmax for DA transport and decreased expression at the plasma membrane as wells as a significantly increased sensitivity to regulation by the phorbol ester PMA. In order to meet the goals of the second aim, DNA was extracted from buccal cells of 45 subjects with either primarily hyperactive-impulsive or combined inattentive hyperactive-impulsive ADHD. The 14 coding exons and splice junctions of hDAT were screened for polymorphisms using the Reveal Mutation Discovery System. Polymorphic samples identified by the Reveal system were subsequently sequenced using dideoxysequencing methods to determine the polymorphism(s) present. To date, 9 of

Developmental dyslexia is one of the most common neurological childhood disorders affecting 5–10% of the population. It is generally defined by a significant discrepancy between general IQ and reading skills. Reading deficits are highly heritable as demonstrated by a concordance rate of 68% in MZ twins and 38% in DZ twins. Transmission of dyslexia tends to be complex suggesting the influence of multiple genes on the phenotype. Although linkage studies have identified several regions linked to dyslexia, 1p, 2p, 3, 6p, 15q, and 18p, these regions contain hundreds of genes among which only a few are likely to contribute to the disorder. Here, we propose utilizing bioinformatics methods to select candidate genes from the regions of significant linkage. We will begin by constructing a list of all genes within the 1-LOD interval of significant linkage scores utilizing results from published and internal studies. As this list will contain hundreds of genes, we will select candidate genes based on their known function, gene ontology searches, participation in neurological pathways, brain expression, and co-expression with known candidate genes. In addition to manual information gathering through publicly available sources, we will utilize bioinformatics programs, such as PathwayAssist and GeneWays, which search the PubMed database and predict pathways of interacting genes. All of these sources of information will be organized into a database. The information can then be weighted and genes for further analysis prioritized.

17.2 AN INTEGRATIVE DATABASE FOR GENETIC EPIDEMIOLOGICAL STUDIES Potapova N, Taratine B, Trakalo J, Lanktree M, Kennedy J, and Macciardi F Neurogenetics Section, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada We have developed a database for genetic epidemiological studies that allows for the integration of genetic, phenotypic and clinical data as well as for developing analysis workflows. The software is built on a multi-tiered web-based (HTML, XML) platform and uses MS SQL Server. User interfaces are generated using ASP and ASP.net. This format results in highly increased stability over conventional platforms such as MS Access, while at the same time allowing for increased integration with other network/web based services and applications. The size of the database is chiefly limited only by the ability of the users to organize their research projects. Scripting and coding capabilities allow for most manual operations (such as error checking) to be automated. Chief features of the database include import functions for genetic, phenotypic and clinical data that incorporate measures to ensure data integrity and allow for importing from common spreadsheet programs such as MS Excel or directly from instrumentation. Security and access to the database can handle complexities that restrict user access to critical database functions, and allow for targeted, remote collaborator access while protecting key elements of the data. The database permits user-defined queries to be generated under a ‘wizard’ model, thus reducing the database administrator’s workload. The database is versatile enough to interface with most statistical and office packages, as well as with programs for associationbased analyses (TDT, FBAT, Transmit) and can incorporate automated ‘assessment’ tools such as tests for mendelian inheritance errors and LD in the pre-formatted dataset. Analysis workflow can be customized for an entire lab, or for individual projects. A tool for automation of

154

Abstracts

interviewing schedule is being developed that will incorporate wireless technology for real-time input of interview results. Because the format is HTML-based, links to text documents, reference tables, and web sites can be embedded in the database interface itself.

Estrada JK,1 Nicolini H,2 Vallejo E,1 Fresan A,3 De la Fuente C,3 and Meyemberg N3 1 Monterrey Institute of Technology, Mexico City, Mexico 2 Mexico City University, Mexico City, Mexico 3 Carracci Medical Group, Mexico City, Mexico

P17.3 DEVELOPMENT OF A NETWORKED DATABASE SERVICE FOR GENETIC ASSOCIATION STUDIES Nikolov I, Ivanov D, Moskvina V, Williams J, Craddock N, and Kirov G UWCM, Cardiff, United Kingdom

The use of machine learning techniques for detecting epigenetic effects of several loci in multigenetic diseases has been poorly, explored, particularly in schizophrenia. In addition, the use of simulation data for designing the best methodological approach has not been widely discussed in the scientific literature. Most studies in schizophrenia have focused on collecting extended pedigrees, sib-pairs, trios and population case-control subjects. From the clinical perspective, it would be much easier to collect discordant sibs where one has schizophrenia and the other one has not. In this study, we simulate data of discordant sibs with schizophrenia using the epistasis model described by Li and Reich (2000). Similar studies using other computational methods have found a four-locus interaction associated with Sporadic Breast Cancer, whereas individual studies of each locus have produced inconsistent results. We use data produced in simulation to validate the efficiency of the proposed model. Data is generated using two parameters for schizophrenia: a two loci interaction with a penetrance of 60% and a heritability of 90%. Further, data is evaluated using different sample size and number of markers to detect the best fit model. We propose a computational model based on genetic programming for inducting decision trees from this data. We estimate the prediction error based on the average predictive ability across 10 fold cross-validations for 10 different data sets simulated for each different sample size and number of markers. Experimental results showed an average error of 25% in the 10 datasets generated under the epistasis model with a penetrance of 60%, a heritability of 90%. On the other hand, the average error estimates obtained by genetic programming using 300 discordant sib pairs across the 10 times 10 fold cross-validation was 27.083%, that is just 2.083 points above the true error rate. Furthermore, the resulting decision tree can be directly interpreted as rules to predict affection status, this is very important since previously explored methods such as Artificial Neural Networks have good prediction capabilities but interpretation of the model obtained is rather difficult. Simulation results showed that by using 300 discordant sib pairs with schizophrenia we are able to detect epistatic effects among ten markers when the epistasis model is given by two biallelic markers that exhibit interaction effects in the absence of main effects. However, further studies are required to attest the potentials of our model. We propose the use of ten of the most consistent candidate genes associated with the disorder.

We have developed a network-based database system for the Neuropsychiatric Genetics Unit at UWCM. It stores extensive phenotype data for association samples for a variety of phenotypes: Schizophrenia, Bipolar Affective Disorder, Unipolar Depression, Autism, Alzheimer’s Disease and Dyslexia. Currently there are more than 2000 cases, and more than 2000 unrelated controls and 1200 parent-offspring trios. All our data from individual and pooled genotyping on those sample sets is stored in a relational database. Information on the markers and genotyping assays is also available. The main purpose of the database is to assist the work of a number of research groups by providing a convenient access to, and manipulation of data, and the opportunity to perform some common statistical analyses via user friendly interface and in an automated way. The service is powered by MySQL server. Web based Graphical User Interface is used through an application server and an HTTP server. The application is scripted in Visual Basic Script and JavaScript. All modules for statistical analysis are developed in house using either original or published algorithms. Connection to and simultaneously searching public databases are common features. The following functions have been implemented so far: manipulation of phenotypic data, automated bulk import of data from files containing genotypes, customised data output in several standard file formats, case/control Chi squared test, allelic and genotype frequencies with HWE statistics, TDT, Linkage Disequilibrium (D0 and R squared) measures from biallelic markers with the option of using an EM algorithm for estimation of phase uncertain haplotypes, haplotype association tests for trios data (haplotype TDT), Likelihood ratio test for haplotypes in case/control data, error checks on the genotypic data. The analysis is performed in real time, reflecting any newly updated information rather than using stored statistics. The database was designed according to widely accepted conventions in database design and file formats for association analysis. Our system can easily be adapted to serve other research groups with data hosting and web based access. An example service can be accessed at http:// w011.pc.uwcm.ac.uk/example_db/default.asp P17.4 DESIGN OF OPTIMAL PRIMERS FOR SNP GENOTYPING ASSAYS BASED ON PRIMER EXTENSION Ivanov DK, Moskvina V, Owen MJ, and Kirov G University of Wales College of Medicine, Cardiff, United Kingdom A number of methods for genotying SNPs have been developed in recent years. In our work with designing primers for SnapShotTM, extension of fluorescently labelled primers and Flourescence Polarisation, we observed a number of primers that did not extend well. We frequently noticed that such primers had low minimum free energy of the secondary structure of the primer, low melting temperature or both. In order to assist our own work and automate the process of designing extension primers we developed our own programme for primer design. The input is a flanking DNA sequence with the SNP in brackets, e.g. [a/ g] and the output is a list of the sequences of all possible forward and reverse extension primers for given parameters, their main characteristics (dG, Tm, GC content) and an image of the optimal secondary structure, for which we designed our own algorithm. In order to set optimal primer parameters which can be used for most of the assays we analysed the performance of 131 primers used in our laboratory. We set up cut-off point for the minimum free energy to be -3.00. Most primers below this cut-off point have very strong secondary structure, which compromises the efficiency of the extension reaction. The optimal melting temperature of the extension primer is set between 60 and 808C. The programme is web based and freely available at: http://m034.pc.uwcm.ac.uk/FP_Primer.html P17.5 A NOVEL COMPUTATIONAL APPROACH FOR DETECTING EPISTASIS IN A SIMULATION MODEL OF MULTIPLE LOCI USING DISCORDANT SIBLINGS: A PROPOSAL FOR THE STUDY OF SCHIZOPHRENIA

P17.6 POSITIVE ASSOCIATION OF 5HTR2C SER23 ALLELE IN SOUTH INDIAN YOUNG AGE AT ONSET BIPOLAR PATIENTS Jain S,1 Purushottam M,1 Mukherjee O,1 Kubendran S,1 Dash D,2 and Brahmachari SK2 1 National Institute of Mental Health and Neurosciences, Bangalore, India 2 Institute of Genomics and Integrative Biology, Delhi, India Serotonergic dysfunction is implicated in the biology of mood disorders. Functional consequences caused by variations in these genes may thus provide evidence to the genetic underpinnings of these disorders. We studied the Cys23Ser polymorphism in the third exon of theHTR2C gene in a sample set of 269 individuals which included 144 bipolar subjects (F ¼ 69, age at onset (AAO) with mean 17.9 years and s.d. 6.4 years) and 125 age, sex and ethnicity matched controls. We observed a significant increase in the Ser23 allele frequency in the patient group compared to the control (w2 ¼ 8.611, P < 0.003) population. When we analyzed the data in a subset of very young symptom onset (AAO < 17 years, N ¼ 90, F ¼ 45), we found that the difference in the Ser23 allele frequency was more pronounced (w2 ¼ 10.729, P < 0.001) than when the entire BPAD sample was considered. However, we did not observe a significant difference in the Ser23 allele frequency between the late onset patients (AAO > 17 years, N ¼ 54, F ¼ 25) and the control population (w2 ¼ 2.469, P > 0.116). Homology modelling of 5HTR2C using bovine rhodopsin suggested possibility of a disulfide bridge between the four conserved Cys residues at positions 127, 207, 337, and 341 and the Cys 23 residue which occurs only in humans. Substitution of Cys by Ser could influence the functional characteristics of the various edited isoforms, and thus the properties of 5HTR2c

Abstracts receptor function. The findings confirm the previously reported association between the polymorphism and bipolar disorder, and suggest a model by which it could influence susceptibility to BPAD.

P17.7 MODEL-FREE LINKAGE MAPPING OF X-LINKED QUANTITATIVE TRAITS VIA MATERNAL ALLELE SHARING Ghosh S and Datta S Indian Statistical Institute, Kolkata, India Although allele-sharing methods for mapping autosomal traits, both qualitative and quantitative, have been extensively proposed and compared, very few attempts have been made to develop similar methods for mapping X-linked traits. While some methods have been suggested for X-linked binary traits, they have essentially been direct applications of existing statistics for autosomal loci. However, an interesting feature of X-linked traits is that the sharing of a paternal allele at the trait locus by a sib-pair is deterministic (i.e., it is irrelevant for male sibs, always shared by female sibs and never shared by two sibs of opposite sexes). This implies that the mode of inheritance of the paternal allele does not carry any information on linkage for an Xlinked locus. Thus, a test for linkage on the X-chromosome utilizes information only on the nature of transmission of maternal alleles to the offspring (Ho and Bailey-Wilson 2000). However, to our knowledge, there have not been studies on model-free statistical methods for mapping X-linked quantitative traits. In this study, we explore three methods for linkage analyses of X-linked quantitative traits based on sib-pair data. The first method is a non-parametric variance components approach using the squared difference in sib-pair trait values and a random effect of maternal allele sharing at a marker locus. The second method is a suitable modification of the classical Haseman-Elston linear regression (1972). The third method uses a logistic model linking the number of maternal alleles shared identical-by-descent and the squared difference in sib-pair quantitative trait values. The models in these methods differ from those for analyzing autosomal trait loci because the additive allele effect model for autosomal loci may not hold good for X-linked loci. We assess and compare the powers of our proposed methods using extensive Monte-Carlo simulations and find that the logistic model, in general, yields the maximum power among the three methods.

P17.8 MOLECULAR EVOLUTION OF NR2E1: A CANDIDATE FOR PATHOLOGICAL VIOLENCE AND 6Q-LINKED BRAIN-BEHAVIOUR DISORDERS Kumar RA,1 Leach S,2 Chen J,3 Yokom DW,3 Abrahams BS,4 Brooks-Wilson A,5 and Simpson EM1 1 Centre for Molecular Medicine & Therapeutics and the Department of Medical Genetics, University of British Columbia, Vancouver, Canada 2 Genome Sciences Centre, Vancouver, Canada 3 Centre for Molecular Medicine and Therapeutics and University of British Colubmia, Vancouver, Canada 4 Centre for Molecular Medicine and Therapeutics, Graduate Program in Neuroscience and Department of Medical Genetics, University of British Columbia, Vancouver, Canada 5 Genome Sciences Centre, British Columbia Cancer Agency, and the Department of Medical Genetics, University of British Colubmia, Vancouver, Canada Mice deleted for the brain-specific nuclear receptor Nr2e1 (‘fierce’) are ‘hard to handle’ and pathologically violent, such that males will bite, wound, and kill their siblings and intended mate. Increased aggression is also evident in females, who also lack maternal instinct. In addition, ‘fierce’ mice present with abnormalities of the forebrain, retina, and olfactory lobes. Human and mouse NR2E1 proteins are 99% identical, and we have recently corrected the abnormal brain-behaviour phenotype in ‘fierce’ mice with the human NR2E1 gene. In light of these molecular and functional similarities between mouse and human NR2E1, together with the mapping of putative schizophrenia and bipolar loci to the NR2E1 region, we hypothesize that humans with ‘fierce-like’ features will harbor functional variants in their NR2E1 gene. Critical to the testing of this hypothesis is the prerequisite to understand the extent of natural variation in human NR2E1, and to identify single-nucleotide polymorphisms with which to assess the role of NR2E1 in human mental health and disease. We sought to comprehensively and systematically evaluate DNA variation at this locus

155

by high-throughput sequencing of all coding, untranslated, and 6 putative regulatory regions in 95 ethnically diverse individuals, as well as in chimpanzee, gorilla, orangutan, and the rhesus and Japanese macaques. We demonstrate that over the course of approximately 15 million years of evolution, NR2E1 has not tolerated a single aminoacid replacement change. We have also tested the neutrality assumption of molecular evolution (i.e., that sequence variation at NR2E1 is selectively neutral). Our analyses indicate a significant departure from neutrality. The general absence of common variation and a relative excess of rare variants suggest that selection or demographic factors have shaped the evolution of this highly conserved locus. Our DNA sequencing efforts, together with evolutionary-based studies, will shed light onto the molecular origins and history of NR2E1. Importantly, our findings complement ongoing work aimed at evaluating the role of NR2E1 in disorders of human brain and behaviour.

P17.9 STUDY OF THE DRD2 GENE 30 REGION HAPLOTYPE IN SPANISH PSYCHIATRIC PATIENTS Hoenicka J,1 Ponce G,1 Jime´nez-Arriero MA,1 Aragues M,1 Rodrı´guez-Jime´nez R,1 Cubero P,1 Diez J,1 Holgado P,1 Martı´nez I,2 Mun˜iz MJ,1 Rubio G,2 Vega S,1 and Palomo T1 1 Hospital 12 de Octubre, Madrid, Spain 2 Centro de Salud Mental del Retiro, Madrid, Spain The dopamine receptor (DRD2 gene, chromosome 11q23) has been extensively involved in the schizophrenia and substance abuse disorder pathogenesis. Because of that, a number of molecular case-control studies have investigated whether there is any association with different DRD2 polymorphisms. These study results have not been conclusive, at least in part, due to heterogeneity ethnicity of the sample and also that in most of them only one polymorphism is analyzed. It has been found that C957T, that has functional consequences, is in linkage disequilibrium in a European-American population with TaqI ‘A’ variants. In this study we have genotyped three DRD2 single nucleotide polymorphism (SNPs), two of them in the coding region (C957T, dbSNP rs6277, G1101A) and the other one in the 30 UTR (TaqI A, dbSNP rs1800497) in Spanish patients and control population free of psychiatric illnesses The aim of this study was to compare the frequency of the DRD2 variants among the different groups in order to investigate the role of this gene in the vulnerability of these disorders. The study included three Spanish populations: 164 schizophrenics, 149 alcoholics and 66 healthy controls and 40 trios. These populations were recruited from the ‘‘Hospital 12 de Octubre,’’ Madrid, Spain. Genotype for each SNP was determined using Taqman 50 exonuclease assay, and afterwards haplotypes were defined. The molecular analysis of the DRD2 gene in our samples revealed no differences regarding to the frequency of the alleles defined by the three SNPs. In addition we found that the Spanish population has a different allelic distribution for this gene when compare to European-American and African-American populations Currently we are extending our analysis to study if there is any association between any specific trait and DRD2 gene. P17.10 ANALYSIS OF MAO-A GENE IN INDIGENOUS POPULATION OF MEXICO Camarena B,2 Serrano C,1 Cruz C,2 and Gonza´lez-Sobrino BZ1 1 Instituto de Investigaciones Antropologicas UNAM, Mexico, Mexico 2 Instituto Nacional de Psiquiatria Ramon de la Fuente, Mexico, Mexico The monoamine oxidase A (MAOA) is an important enzyme in human physiology and behavior. The promoter region uVNTR which consists of a 30-bp repeated in 3, 3.5, 4 or 5 copies have been used as functional genetic marker. Have been reported variants in allele frequencies between human populations. In Mexico there is great variability in the distinct indigenous groups. In the present paper we show the case of one group of 236 individuals of different linguistic group: a) Huicholes and Coras of the Sierra Madre Occidental of Mexico, which speak the corachol language b) nahuas of the Sierra Madre Oriental, which speak one type of nahuatl in the State of Veracruz, and c) one group of Mexico City. The heterozygosity value is 0.36 to 0.48. We calculated the statistics FST which measure the endogamy between populations. In this preliminary study FST is of 0.1 between the both woman of the city and indigenous groups whose distance is of 0.02 to 0.05. In contrast, the distance between men (indigenous/men of city) is of 0.06. This event can be result of the recurrent migration of the indigenous men to the city

156

Abstracts

because the economic situation and commercial activities. Also, it suggests that there are similarities among the Coras, Huicholes and Nahuas whose history, traditions and language are different.

P17.11 STUDIES OF THE 19-BP INSERTION/DELETION POLYMORPHISM IN THE 50 FLANK OF THE DOPAMINE B-HYDROXYLASE GENE Togsverd M,1 Werge T,1 Guerini FR,2 and Rasmussen HB1 1 Research Institute of Biological Psychiatry, Sct. Hans Hospital, Roskilde, Denmark 2 Laboratory of Biology, Don C. Gnocchi Foundation, ONLUS, IRCCS, Milan, Italy Dopamine b-hydroxylase (DBH), which converts dopamine to norepinephrine, is the only enzyme of the catecholamine pathway present within synaptic vesicles. Upon stimulation of dopaminergic neurons it is released into the synaptic cleft. This appears to be the major source of DBH in plasma. Most likely, the activity of DBH in plasma is under strong genetic control. Upstream of the gene encoding DBH is a functional 19-bp insertion/deletion polymorphism, which may influence promoter activity and account for individual variations in the activity of the enzyme. In this study we determined the ancestral allele of the 19-bp insertion/deletion polymorphism. We also compared the frequencies of the two alleles of this polymorphic site in populations of different ethnicity. Finally, we examined whether the 19-bp insertion/ deletion polymorphism could influence susceptibility to schizophrenia or modify the course of this disease. Examination of samples from gorilla and chimpanzee revealed the so-called insertion allele to be the ancestral form. On comparison of three populations deriving from Northern Europe, Southern Europe and China we did not find significant differences in distribution of genotypes (P ¼ 0.51) or alleles (P ¼ 0.31). The frequency of the so-called deletion allele ranged from 0.40 to 0.46 in the three populations. The null hypothesis of no association between 19-bp DBH genotype and patient-control status was not rejected (P ¼ 0.32). Neither did the distribution of alleles between the two groups differ significantly (P ¼ 0.15). We did not find a relationship between age at first admission and DBH genotype status. In summary, we found that the so-called 19-bp insertion allele of in the 50 end of the DBH gene is the ancestral form and that the deletion allele is the derived and novel variant. The derived allele is associated with low plasma activity of DBH and is present in different human populations at a frequency above 0.40. Whether positive selection has been involved in the establishment of this allele at such high frequency remains unknown. In agreement with most other studies we did not find that the 19-bp deletion allele was associated with schizophrenia. Neither could we associate it with onset of this disease.

P17.12 RELATIVE EFFICIENCY OF THE LINKAGE DISEQUILIBRIUM MAPPING APPROACH TO DETECT CANDIDATE GENES FOR SCHIZOPHRENIA IN DIFFERENT EUROPEAN POPULATIONS Costas J, Torres M, Quintela I, Cristobo IJ, Phillips C, and Carracedo A Unidade de Medicina Molecular, Complexo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain With the availability of millions of single nucleotide polymorphisms (SNPs), association mapping by linkage disequilibrium appears as a powerful approach to unveil the genetic basis of complex diseases. In such indirect association studies, whether susceptibility genes can be detected is dependent not only on the degree of linkage disequilibrium between the disease variant and the SNP marker but also on the difference in their allele frequencies. Although it is expected that both the extent of linkage disequilibrium and the allele frequencies are quite similar in related populations, very little is known about how small variations in these parameters may affect the power to detect a disease variant by genotyping a linked SNP in different european populations. To this goal, we genotype around 50 SNPs in COMT, G72/G30, DTNBP1, and NRG1 in three european populations, Galician (NW of Spain), Greek, and Norwegian. All the four genes are involved in susceptibility to schizophrenia and, perhaps other psychiatric disorders. First, we analyze the pairwise linkage disequilibrium (LD) between SNPs using the E-M algorithm to infer haplotypes. Then, we investigate the power of indirect association studies by assuming each one of the SNPs to be the functional one, and calculating the required sample size to reach a 80% power to detect association, given an OR ¼ 2

for the functional variant, using each one of the other SNPs as markers. Our results show that, in general, the haplotype-block structure is shared by the three populations. Overall, common haplotypes are the same in all populations and the haplotype frequencies and patterns of LD are similar in the three populations. As a result, most SNPs present an equivalent power to detect the disease allele in the three populations. Nevertheless, several SNPs are in intermediate but significant LD with some others in only one or two populations. As a consequence of this change, as well as the differences in allele frequencies, the relative efficiency of these SNPs as markers of the disease allele is clearly different in different populations. In the most extreme cases, some SNPs are appropiate markers in some populations but totally ineffective in others. These findings have important implications for studies attempting to replicate previous associations in different populations. In general, these studies test the same subset of markers showing first association. Taking into account our results, the lack of association in this type of studies may be related to the different efficiency of a SNP as marker of the disease variant in different populations, indeed if the populations are closely related, as is the case for the European ones.

P17.13 STRUCTURAL VARIANTS OF THE NEUROLIGIN 4 GENE IN US AND PORTUGUESE PATIENTS WITH AUTISM Yan J,1 Oliveira G,2 Coutinho A,3 Yang C,1 Feng J,1 Cook E,4 Vicente A,3 and Sommer S1 1 City of Hope National Medical Center, Duarte, United States 2 Hospital Pediatrico de Coimbra, Coimbra, Portugal 3 Instituto Gulbenkian de Ciencia, Oeiras, Portugal 4 University of Chicago, Chicago, United States Neuroligins are postsynaptic membrane cell-adhesion molecules. Recently, mutations in the neuroligin 4 gene (NLGN4), located in the X-chromosome, were found in patients with autism and with mental retardation. To better understand the relationship between the neuroligin gene and autism, we used DOVAM-S (Detection of Virtually All Mutations-SSCP) to scan the coding regions and associated splice junctions of NLGN4 in 96 unrelated Caucasian patients with autism (72 Portuguese and 24 Midwest US Caucasian), 48 Portuguese healthy controls and 48 patients with ADHD or BPD. Putative missense mutations were identified in NLGN4 in three patients. V403M was found in a US male patient of Irish descent, and G99S and K378R were found in two Portuguese patients, a female and a male, respectively. Each missense mutation altered an amino acid conserved in all known members of neuroligin 2, 3, 4, and 4Y, which encompass an evolutionary divergence of over 1.5 billion years. In addition, K378 and V403 are also identical in neuroligin 1. Each structural variant was present in the asymptomatic mothers of the patients, and mutation V403M was also present in one affected sibling of the proband. No missense changes were found during comprehensive scanning of 96 individuals without autism. In addition, sequencing of the appropriate exons in 96 Portuguese and 96 US Caucasian controls (276 X-chromosomes total) did not reveal G99S, K378R, or V403M or any other missense variants. While functional studies are important to unequivocally prove causation, the data herein support the role of neuroligin 4 mutations in the etiology of autism in some affected families.

P17.14 STUDY OF FUNCTIONAL PATHWAYS IN PSYCHIATRIC DISORDERS USING A COMBINED THE SNPLEX AND MASSARRAY TECHNOLOGY Gratacos M,1 Costas J,2 Baye´s M,1 and Cristobo J2 1 Barcelona Genotyping Center, Genes and Disease Program, Center for Genomic Regulation, Barcelona, Spain 2 Mollecular Medicine Unit, University of Santiago de Compostela, Santiago de Compostela, Spain 3 Spanish Psychiatric Genetics Network, Spain In order to understand the genetic architecture of prevalent psychiatric disorders through association studies we have developed an efficient and cost effective strategy for SNP genotyping of candidate genes. Most of the SNPs will be genotyped using SNPlex, a technology recently developed by Applied Biosystems which has a current plexing level of 45–48 SNP loci in a single reaction. SNPlex uses oligonucleotide ligation assay (OLA) combined with multiplex PCR technology to achieve allelic discrimination and target amplification. The use of mobility modifiers that are incorporated into the primers in conjunction

Abstracts with differential dye labelling allows the detection of each allele by capillary electrophoresis on a 3730 XL DNA Analyzer. In our hands, conversion rate for SNPlex is around 70%, and thus other technologies such us Sequenom MassArray Spectrometry is also used in our strategy. We will explore genes involved in different pathways in the central nervous system: noradrenaline (i.e. TH, MAO, COMT, ADRA2C, SLC6A2), dopamine (i.e. DRD1, DRD2, SLC6A3, PPP1R1B), inhibitory neurotransmitters (i.e. GAD1, GABRA1, GABBR1), glutamate (i.e. GRIN2B, NOS1, GRIA2, SNAP25, GRM3, SLC1A2), serotonin (i.e. TPH, HTR2A, HTR6, SLC6A4) peptidergic systems (i.e. CRH, PENK, PDYN, CCK), neuronal differentiation and growth (i.e. BDNF, EGF, NLG4), steroid hormones (i.e. ESR1, REA), adrenal steroid receptors (i.e. GRLL1, SERPINA6), opioid receptors (i.e. OPRD1, OPRKL1), cannabinoid receptors (i.e. CNR1, FAAH) purinergic receptors (i.e. ADORA1, ADORA2A) and cholinergic system (i.e. CHRNA7, CHMR2, ACHE). A total of 366 genes have been included in SNPlexes. The project has been organized in two phases. For the first phase we have selected 166 SNPs with putative functional consequences (missense or nonsense mutations, and those affecting splicing consensus sequences). These SNPs will be genotyped in 3 SNPlex assays, with 44–48 SNPs per reaction and 5 Sequenom assays, each of them with of 6–8 SNPs per reaction. In a second phase, we will focus on the haplotype diversity at the 50 cis-regulatory regions, selecting about 1500 SNPs, priorizing those that may have influence on the levels of gene expression: SNPs located on promoters and that alter the consensus sequence of transcription binding factors, and SNPs located in gene regions that are highly evolutionary conserved regions. We plan to study, with this approach, several psychiatric disorders, such as opiate dependence, panic disorder, schizophrenia, eating disorders, pathological gambling, suicide attempts in depressive patients, ADHD, OCD, autism and a set of non-psychiatric controls. All the disorders will be included in first phase, while second phase will be restricted to those disorders with enough samples to detect moderate association by linkage disequilibrium mapping. The study provides a comprehensive approach to elucidate the contribution of functional candidate variants to participate in different psychiatric phenotypes.

P17.15 ANALYSIS OF LINKAGE DISEQUILIBRIUM AND HAPLOTYPE DIVERSITY ACROSS THE NRG1 GENE IN A WORLDWIDE POPULATION SET Gardner M, Gonzalez-Neira A, Bertranpetit J, and Comas D Pompeu Fabra University, Barcelona, Spain Neuregulin 1 (NRG1) was first identified as being associated with Schizophrenia in an Icelandic population, a finding which has since been supported by replication studies in several different populations, Scottish, British, Irish and Chinese. To date no study has examined the diversity of both the haplotype composition of this gene or the Linkage Disequilibrium (LD) patterns within it across worldwide populations. We present such a study in the large 1.4 Mb NRG1 gene region from a global perspective. 15 SNPS were typed across the gene in a set of 39 worldwide populations (1,080 individuals) representing each continent in the world. Differences among populations were found in both haplotype composition and LD patterns, with populations found to group according to geographic region. Observations on both these parameters in this gene region from a global perspective are fundamental in a number of respects, for example, to understand the extent to which the finding of association of a haplotype in one population is applicable to another, for facilitating gene mapping studies in general and from a broader view to infer population history, among others. These results may also be extended to include implications for other gene regions related to psychiatric diseases.

P17.16 HAPLOTYPE-BLOCK ASSOCIATION MAPPING IN PSYCHIATRIC COMPLEX TRAITS: CONSISTENCY OF FINDINGS ACROSS DIFFERENT POPULATIONS Macciardi F,1 Elias-Billon I,2 Chumakov I,2 Abderrahim H,2 Tanaka H,2 and Cohen D1 1 Dept of Biology & Genetics, U of Milano, Milano, Italy 2 Serono Genetic Institute—SGI, Paris, France Our current understanding of Linkage Disequilibrium (LD) in humans is best described by a block-like pattern, where SNPs show a high pairwise disequilibrium in regions of variable size, ranging from a few up to several hundred kilobases and thus making up a discrete block,

157

and where these blocks are interspersed with shorter regions with lower or no LD across SNPs. Under the block-like hypothesis, the small inter-block regions with low or absent LD correspond to meiotic recombination spots, while the set of SNPs included within a block constitutes a haplotype with a limited variability, due to the highly conserved disequilibrium across all of them, making a haplotype mapbased LD approach to find genes for complex traits a feasible strategy. In fact, given the low variability of the intrablock LD, it is possible to select and genotype only those SNPs, among all included in the block itself, that best identify the block-based haplotype (tagging SNPs) without appreciable loss of information, but with a considerable reduction of genotypings and consequently with a reduction of costs. Initial applications of haplotype-block mapping showed positive identifications of genes related to complex traits and prompted the development of the HapMap project which aims at identifying the complete LD block structure of the human genome, eventually leading to a complete map of the human haplotypes. However, several issues remain to be clarified, before considering a haplotype-block map-based approach the best strategy to identify genes in complex traits. In this respect, we do not yet know if a given haplotype-block pattern from a particular population is generalizable to other populations. The data available at present mostly derive from Gabriel et al. (2002) who showed a similar block pattern in Asian and Caucasian populations, both being different from Africans who presented a mean shorter block structure, due to the older ancestry of African populations compared to either the Asian or the Caucasian. Only a diversity in haplotype frequency and not in the block structure is the major difference between Asian and Caucasian populations, thus making a case in favour of the generalization of the haplotype block approach. Nonetheless, we know that there are large population differences also within the Caucasian or the Asian subdivisions, due to evolution forces at population level that may shape differently also the LD block structure. To evaluate this effect, we analyzed the block structure of the gene G72 in 5 different Caucasian populations: our results show that there are qualitative differences in block structure and not only quantitative variations, but even so we can map the gene G72 with a haplotype block strategy, providing that we detect the proper set of tagging SNPs that appear to be population dependent and thus can not be generalized. P17.17 JUSTIFICATION AND TECHNIQUES FOR GENE-BASED ASSOCIATION ANALYSIS Neale BM,1 Visscher PM,3 and Sham PC2 1 King’s College, London, United Kingdom 2 Hong Kong University, Hong Kong, China 3 University of Edinburgh, Edinburgh, United Kingdom Association studies were previously limited to single variants, so the allele was considered the basic unit for association testing. With the increase in marker density and the utilization of indirect approaches of assessment of association signal via linkage disequilibrium (LD), geneticists have shifted focus to haplotypes, rather than single variants. We suggest that there are a number of problems when attempting replication of association at the single variant or haplotypic level, and argue for a move to a gene-based approach in which the entire set of common variation across a candidate gene is considered jointly. Population differences leading to non-replication are more readily overcome when approaching association from a gene-based perspective rather than either a SNP-based or haplotype-based perspective. A gene-based approach analyzes all potential risk-conferring variants, therefore power is the only source of negative findings. Additionally, multiple testing issues are more easily considered by utilizing a genewide significance level, yielding a convenient two-stage procedure. Eventually, we will have mapped all variation across all the genes in the genome. When complete knowledge of gene variation is achieved, the gene-based approach is the natural endpoint for association analysis, and will serve to inform our search for functional variants relevant to disease-aetiology. We also propose some methodological approaches for dealing with such gene-based analysis. These include techniques for approaching both common variation and multiple rare variants. P18.1 PREDICTING CLOZAPINE-INDUCED WEIGHT GAIN: ASSOCIATION WITH THE 759 C/T POLYMORPHISM OF THE 5HT2C RECEPTOR Buckley PFB, Miller DDM, and Ellingrod VLE Medical College of Georgia, Augusta, United States

158

Abstracts

Weight gain has been documented as a significant adverse effect of many of the atypical antipsychotic medications. Several recent reports have linked a 759C/T polymorphism of the 5HT2C receptor and obesity as well as to weight gain from chlorpromazine, risperidone and clozapine. To determine the association between changes in BMI during clozapine treatment and the 759 C/T polymorphism of the 5HT2C receptor gene. This study included 41 patients with treatmentrefractory schizophrenia (DSM-IV) who were followed prospectively during treatment with clozapine. Weight and height measurements were obtained prior to starting clozapine and after 6-months of treatment. Clozapine doses were determined based on clinical response. Genomic DNA was isolated from a whole blood sample and analyzed for the 759C/T polymorphism of the 5HT2C receptor. A Chi Squared analysis comparing the distribution of T and C alleles in subjects grouped as having an increase of more or less than 7% of their baseline BMI during treatment with clozapine found that the distribution of C alleles was significantly higher in subjects with an increase in baseline BMI of 7% or more compared to those with less than a 7% increase BMI A multiple linear regression analysis showed that both baseline BMI and 759 C/T genotype had significant effects on 6-month BMI. The T allele may have a protective function in preventing significant weight gain from clozapine. Subjects with a C allele were at a greater risk for weight gain from clozapine over 6-months compared to those with the T allele.

P18.2 MUTATION ANALYSIS OF THE GENE ENCODING THE RYANODINE RECEPTOR GENE ISOFORM 3 (RYR3) IN RECURRENT NEUROLEPTIC MALIGNANT SYNDROME Opgen-Rhein C,1 Steinlein O,4 Sander T,2 Neuhaus AH,1 Weber M,3 and Dettling M1 1 Department of Psychiatry, Charite´, University Medicine Berlin, Berlin, Germany 2 Department of Neurology, Charite´—University Medicine Berlin, Berlin, Germany 3 Department of Psychiatry, Clinic Holstein, Lu¨beck, Germany 4 Institute of Human Genetics, University Hospital Bonn, Bonn, Germany Neuroleptic malignant syndrome (NMS) is a rare and potentially lethal complication of antipsychotic medication (1,2). The striking symptomatic and pathophysiological similarities with malignant hyperthermia (MH) suggest that NMS might be a neurogenic form of MH. MH can be caused by mutations in the ryanodine receptor (RYR1) gene. The present pilot study tested whether mutations in conserved coding sequences of the brain-expressed RYR3 gene confer vulnerability to NMS (3). Direct DNA sequencing was performed for 40 RYR3 exons and adjacent intron/exon boundaries, covering all conserved sequence motifs, in two patients with recurrent NMS induced by different antipsychotics. Mutation analysis of the conserved RYR3 sequence motifs revealed no sequence variants that are likely to alter protein structure or gene expression. Several intronic single nucleotide polymorphisms (SNPs) were identified, including IVS9 þ 98A ! G, IVS1452C ! T, IVS16 þ 70T ! C, IVS17 þ 16T ! C, IVS18 þ 8C ! T, and IVS24 þ 55G ! A. The present pilot study of two patients with recurrent NMS induced by different antipsychotics revealed no RYR3 sequence variants that are likely to confer vulnerability to NMS. However, these results do no exclude the possibility that so far undetected RYR3 mutations, in particular in regulatory elements, may contribute to the aetiology of NMS. Several intronic SNPs identified here provide valuable tools to search for vulnerability alleles by linkage disequilibrium mapping.

P18.3 EMBRYONIC STEM CELL-DERIVED NEURONS: A MODEL FOR ANTIDEPRESSANT PHARMACOGENOMICS McHugh P, Rogers G, Allington M, Cameron V, Begg E, Joyce P, and Kennedy M Christchurch School of Medicine & Health Sciences, Christchurch, New Zealand Genetic variation in individuals may affect their ability to respond to antidepressant drug treatment. We are currently affecting a mouse an embryonic stem (ES) cell-derived neuronal culture model to understand the variable antidepressant drug responses that occur in individual humans. Microarray and proteomic assays have been

developed to look at genes and proteins that show altered expression after exposure of neuronal cells to antidepressants. These genes will be reasonable candidates for harbouring variability that might account for inter-individual differences in antidepressant drug response. Proteomics is generally less sensitive, and therefore will sample a smaller range of expressed genes, although it has the advantage of identifying translated products that are likely to be functionally relevant in each cell. The transcriptome analysis is more sensitive but less discerning, as it interrogates any transcript that is present in the cell whether or not it is functionally important. However, the combined information from both approaches should point to a set of genes or pathways that undergo changes in expression in response to antidepressants. The ultimate goal of this project is to develop a model system to identify genes that may be relevant to clinical responses to antidepressant treatment. Genes identified in the model system may play a role in cellular and molecular responses to antidepressants, and therefore inherited variation in these genes may contribute to inter-individual variability in antidepressant response. As candidate genes are derived from the model system, polymorphic variants will be identified in the human equivalents. These will then be used in allelic association studies in a clinical research setting to test their validity and relevance to antidepressant drug response.

P18.4 POLYMORPHISM ANALYSIS OF SEROTONERGIC TRANSCRIPTION FACTORS Kennedy MA and Joyce P Dept of Pathology and Dept of Psychological Medicine, Christchurch School of Medicine and Health Sciences, University of Otago, Christchurch, New Zealand The 5-HT neurotransmitter system influences a wide range of behavioural and physiological processes including cognition, circadian rhythms and mood, and genetic polymorphisms of the serotonergic system have been widely examined for association with these and other phenotypes. However, the transcriptional mechanisms controlling these pathways have been largely unexplored with regards to their role as possible new gene candidates in affective disorders. Recently, several genes encoding transcription factors which are activated downstream of Sonic hedgehog (Shh), have been shown to be necessary and sufficient for development of the specific subset of 5-HT neurons which project to the forebrain. These genes include, in their proposed order of activation, Nkx2.2, Nkx6.6, Gata2, Lmx1B and Fev. Furthermore, in knock-out mice lacking Fev, the resultant defective development of the serotonergic system leads to a heightened anxiety-like and aggressive behavioural phenotype in adults, suggesting an important role for this class of genes in behaviour. The aim of our research is to establish the level of natural variability in these genes by using denaturing high performance liquid chromatography (DHPLC) and DNA sequencing to analyse coding and promoter regions from a panel of individuals. The functional relevance of any SNP’s will be explored as well as any association of such variants with onsett, symptom patterns or treatment outcomes in patients with affective disorders.

P18.5 ASSOCIATION STUDY OF MYO-INOSITOL MONOPHOSPHATASE 2 (IMPA2) POLYMOPHISMS WITH BIPOLAR AFFECTIVE DISORDER AND RESPONSE TO LITHIUM TREATMENT Dimitrova A,1 Georgieva L,2 Nikolov I,2 Toncheva D,1 Owen MJ,2 and Kirov G2 1 Depart. of Medical Genetics, Medical University, Sofia, Bulgaria 2 Depart. of Psychologica Medicine, Neuropsychiatric Genetics Unit, UWCM, Cardiff, United Kingdom Bipolar disorder (BP) is a complex genetic disorder with a lifetime prevalence of 1%. Lithium is the most effective mood-stabilizing drug in the therapy of BP that is thought to exert its effect via the phosphatidylinositol signaling system. IMPA2 codes for an enzyme in this system which is inhibited by lithium. It is located on 18p11.2, a region implicated as a BP susceptibility locus. We examined seven Single Nucleotide Polymorphisms (SNPs) identified within this gene (97-15G > A, 443G > A(R > Q), 490 þ 13-14insA, 558C > T, 599 þ 97G > A, 599 þ 99G > A and rs3786282) for association with BP using 245 parents-offspring trios and in 174 cases and 170 controls. SNPs genotyping was performed by using two methods: Extension of fluorescently labelled primers and AMPLIFLUORTM allele-specific

Abstracts PCR. Allele frequencies distribution did not reveal an association with BP. When good responders to lithium treatment were compared with the poor responders, some statistically significant differences emerged for two SNPs (P ¼ 0.004 and P ¼ 0.03). We cannot find support for the involvement of variation in IMPA2 in susceptibility to BP, but it might be more important in influencing response to Lithium treatment. However our sample is too small to draw definitive conclusions. Future studies in larger samples are needed to confirm the role of allelic variants in IMPA2 for the efficacy in Lithium treatment.

P18.6 ADRA1A PROMOTER SNP ASSOCIATED WITH NEGATIVE SYMPTOM IMPROVEMENT IN OLANZAPINE RESPONSE Clark DA,1 Arranz MJ,1 Mata I,2 Beperet M,2 and Kerwin RW1 1 Institute of Psychiatry, London, United Kingdom 2 Fundacion Argibide, Pamplona, Spain The ADRA1A gene codes for the a1A-adrenergic receptor (adrenoceptor) which has been implicated in antipsychotic treatment by combining the a1-antagonist prazosin with the typical antipsychotic haloperidol to produce effects similar to those shown by atypical antipsychotics. We hypothesised that genetic mutations in the promoter region of the ADRA1A gene may alter the processes and mechanisms of receptor expression contributing to inter-individual variations in antipsychotic response. A 10kb region upstream of the translation initiation site was searched for polymorphisms using a combination of bioinformatic and laboratory-based mutation detection techniques. A total of 7 SNPs (9625-G/A, 7255-A/G, 6274-C/T, 4884-A/G, 4155-C/G, 2760A/C, and 563-C/T) where found above a minimum rare allele frequency of 5%. 79 subjects (62 with schizophrenia, 9 with schizoaffective disorder, 3 with bipolar disorder, 4 with major depression and 1 with atypical psychosis) all of Basque origin from northern Spain, were genotyped for these polymorphisms. All subjects were receiving olanzapine antipsychotic treatment that was assessed prospectively using the Positive and Negative Symptoms Scale (PANSS) and the Global Assessment Scale (GAS). The 4155-A/G polymorphism was found to be associated with response according to GAS (P ¼ 0.05) and further to the change in negative symptoms according to PANSS (P ¼ 0.03). P18.7 CLOZAPINE TREATMENT AND HLA GENETIC SYSTEM: BIOLOGICAL MARKERS OF FAMILIAL TRANSMISSION FOR THE LIABILITY TO SCHIZOPHRENIA Cavallini MC,1 Lattuada E,1 Cocchi F,1 Rossattini M,1 Lorenzi MC,1 Scorza R,2 and Cavallaro R1 1 Fondazione Centro San Raffaele del Monte Tabor, Milano, Italy 2 Clinical Immunology, University of Milano, Milano, Italy Previous studies suggest that partitioning schizophrenic patients according to their HLA structure (i.e. those with HLA-A1 or CRAGA1 antigens and those with HLA-non-CRAG-A1 antigens, respectively), revealed different illness distribution in the two groups (Smeraldi et al., 1986), while the HLA-A1 antigens was associated to chlorpromazine response. Furthermore a specific HLA haplotype (A2/ B35) is associated to the outcome of clozapine treatment, while in the Finnish patients the responsiveness to clozapine is related only to HLA A1 allele. Our hypothesis is that the response to clozapine and the associated HLA structure could identify a strongest/specific biological vulnerability to psychosis and that this vulnerability could be reflected by the distribution familial risk to schizophrenia spectrum disorders. We recruited 45 patients affected with schizophrenia and treated with clozapine: these patients have been treated with clozapine because they are resistant to traditional treatments, assuming the Kane criteria of resistance. In this study we used the clozapine treatment like the first level of biological selection of schizophrenic/psychotic patients. Then, the patients have been typed for HLA A genetic system and reclassified into two groups: 29 subjects having at least a HLA A2 allele and 17 patients non HLA A2. In this latter group we included also patients with genotype HLA A1/A2. Allele A1 is in linkage disequilibrium with HLA B35 that was indicated as a predictor of response to clozapine, independently from HLA A2. The first degree relatives of these probands have been collected and a complex segregation analysis has been performed on the familial samples. 43.7% of patients with HLA A2 have a postive family history for schizophrenia spectrum disorders (Schizophrenia, PPD, delusional disorders), versus the 23% of non HLA

159

A2 patients, while no differences between the two groups have been observed for family history of schizophrenia. Nevertheless, the segregation analysis supports the hypothesis that a Single Major Locus may account for the transmission of schizophrenia spectrum disorders in families of patients treated with clozapine.

P18.8 DOPAMINE D3 RECEPTOR GENE POLYMORPHISMS INFLUENCE OLANZAPINE AND RISPERIDONE RESPONSE Staddon S,1 Arranz MJ,1 Munro J,1 Mata I,2 and Kerwin RW1 1 Clinical Neuropharmacology, Institute of Psychiatry, London, United Kingdom 2 Fundacion Argibide, Pamplona, Spain The dopamine D3 receptor may play an important part role in the mediation of antipsychotic action. It has been suggested that D3 targeting is responsible for the superior efficacy of atypical antipsychotics in the treatment of positive and negative symptoms of schizophrenia. The Ser9Gly polymorphism of the dopamine D3 receptor gene (DRD3) has been extensively investigated in connection with the disease and antipsychotic response. Recent reports have described novel polymorphisms in the promoter region of the gene that may influence drug efficacy. We have investigated DRD3 polymorphisms (Ser9Gly, 205-G/A, 7685-G/C) in a sample of schizophrenic patients of Basque or Spanish Caucasian origin from Navarra (northern Spain) treated with olanzapine (n ¼ 110) or risperidone (n ¼ 107). Response was assessed prospectively using PANSS scores after a minimum of three months of treatment. Association between genotype and symptom improvement was determined by regression analysis with variation in PANSS scores as the dependent variables and genotypes as the predictor variables. Ethnicity was a co-variant in with the predictor variables because of population stratification. Marginal associations were found between the Ser9Gly and 7685-G/C variants and improvement in positive symptoms in risperidone-treated patients. In the olanzapine sample we found a trend of association between the Ser9Gly polymorphism and total PANSS scores (P ¼ 0.07). The three polymorphisms combined showed strong association with total improvement in symptoms (P < 0.01). These results suggest that genetic variation in the dopamine D3 receptors is influential in mediating response to both olanzapine and risperidone response.

P18.9 IS SER9GLY POLYMORPHISM OF DOPAMINE D3 RECEPTOR RELATED TO TARDIVE DYSKINESIA? McGuinness B, McIlroy S, Carson R, Doherty F, Miller P, McHugh S, Nolan A, and O’Neill FA Queen’s University Belfast,United Kingdom Tardive dyskinesia is an involuntary movement disorder of the orofacial musculature and may involve trunk and extremities. It occurs in 20–30% of schizophrenic patients. Risk factors include age, duration of antipsychotic treatment, gender and length of diagnosis. Several groups have shown that a Ser9Gly DRD3 gene polymorphism is associated with risk for tardive dyskinesia where either the G/G genotype or glycine allele has conferred elevated risk. Other studies, however, have not replicated this finding. To determine if the presence of the Ser9Gly DRD3 gene polymorphism was associated with increased risk of tardive dyskinesia in schizophrenic patients. From 580 cases of schizophrenia, a subsample of 87 individuals were analysed, including 79 trios, 6 quads and 2 quins. Blood was taken for DNA extraction and genotyped for the presence of allele 1 (Ser-9) and allele 2 (Gly-9) of the DRD-3 Ser-9-Gly polymorphism. The transmission disequilibrium test was then applied to look for preferential transmission. A research diagnosis of tardive dyskinesia was determined using criteria outlined by Schooler and Kane. A logistic regression analysis was then carried out on 338 singletons using TD as the dependent variable and age, gender, duration of illness and DRD3 genotype as the covariates. Average age of patients- 38.25 years (s.d 7.6) M: F, 3.35:1. 23 (29%) patients were TD positive, 56 (71%) patients were TD negative. We found no significant genotypic association of the Gly/Gly genotype with TD (P ¼ 0.51, Fisher’s exact test). Patients with TD and those without did not differ significantly in the distribution of allele frequencies (w2 ¼ 1.39, d.f. ¼ 2, P ¼ 0.50). The transmission disequilibrium test was not significant (P ¼ 0.84). With regard to singletons, (mean age 42.98 years, s.d.13.4) only age contributed significantly to presence of tardive dyskinesia, (P ¼ 0.05) after adjusting for the confounding variables by logistic regression. There was no positive

160

Abstracts

correlation between genotype and presence of orofacial (w2 ¼ 0.42, d.f ¼ 2, P ¼ 0.81) or limb truncal dyskinesia (w2 ¼ 1.25, d.f ¼ 2, P ¼ 0.54). We found no significant association between DRD3 genotype and presence of dyskinesia in our sample of patients. Age was the only positive predictor of presence of tardive dyskinesia.

P18.10 MOOD DISORDERS AND ANTIDEPRESSANTS TREATMENT: HAPLOTYPE ANALYSIS OF HUMAN PERIOD 3 GENE Ploia C, Serretti A, Benedetti F, Fontana V, Colombo C, Catalano M, and Smeraldi E Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy There is increasing evidence that genetic factors are contributing to the inter-individual variability in antidepressant drug response. Genetic variability of drug action, described by pharmacogenetic, is partially explained by gene variants usually involved in neurosinaptic transmission. Further evidences underline a possible involvement, in this variance, of clock genes implicated in circadian rhythms which are altered in mood disorders. In the present study, we applied an Haplotype analysis to three polymorphisms of Human Period 3 gene (T1940G exon 15, ins/del of 54 bp exon 18, T3110C exon 18) to test the possible association with antidepressant response (particularly SSRIs). We analyzed a sample of 738 subjects affected by Major Depressive Episode (Major Depression n ¼ 421, Bipolar Disorder n ¼ 317) according to DSM-IV criteria, treated with fluvoxamine (300 mg/day) or paroxetine (20–40 mg/day) for six weeks. The severity of depressive symptoms was weekly assessed by the Hamilton Rating Scale for Depression (HAM-D). Haplotype analysis showed a marginal association between the haplotype including the SNPs 1940 TG and 3110 TC, and drug response. In particular, patients affected by Bipolar Disorder showing genotype TG (SNPs T1940G) and CC (SNPs T3110C) are more frequently responders at sixth week of antidepressant treatment (T5 Heterogeneity permutation model: P-value ¼ 0.05). These results support the hypothesis that hPer 3 gene could be involved in antidepressant response, even if the small number of subject with this haplotype did not allow us to detect a more significant result.

P18.11 ASSOCIATION OF A COMMON FUNCTIONAL POLYMORPHISM OF THE 5-HT1A RECEPTOR WITH NEGATIVE SYMPTOM RESPONSE TO ANTIPSYCHOTIC DRUG TREATMENT Reynolds GP,1 Fertuzinhos S,1 Templeman LA,1 Arranz B,2 and San L2 1 Dept of Mental Health, Queen’s University, Belfast, United Kingdom 2 Benito Menni, Dept of Psychiatry, Sant Boi de Llobregat, Barcelona, Spain The 5-HT1A receptor gene contains a common promoter region polymorphism that is thought to have functional effects on receptor protein expression. This 1019C/G polymorphism is also reported to be associated with depression and suicide (Lemonde et al., 2003). We have investigated the association of this receptor with response to antipsychotic drug treatment of schizophrenia. A group of 54 first-episode drug-naı¨ve subjects with schizophrenia had symptoms assessed by the positive and negative syndrome scale (PANSS) at baseline and following 3 months treatment with antipsychotic drugs, mainly risperidone and olanzapine. Blood samples from these subjects were collected after obtaining ethical approval and informed consent, genotyping was performed for the 1019C/G polymorphism following two primer PCR in two tubes employing reverse primers specific for one or other allele. There were 22 subjects with the CC genotype and 32 with one or two copies of the G allele. Percentage improvement in total PANSS was significantly associated with genotype, whereby the G allele predicted poorer response to treatment (P ¼ 0.013). Within the PANSS subscores, no difference was observed in changes in positive symptom scores. However, a highly significant association with percentage improvement in negative symptoms (P ¼ 0.003), with mean improvements of 11.0% for GG/GC and 42.2% for CC genotypes, and a significant effect on general psychopathology symptoms (P ¼ 0.016) was identified. Although no significant association with initial PANSS score was found, there was a significant association with negative symptom sub-score at baseline (P ¼ 0.028). However, including this as a covariate in the analysis, the association with negative symptom response remained significant (P ¼ 0.024). These data demonstrate

that a functional polymorphism of the 5-HT1A receptor is strongly associated with the response of the negative syndrome to treatment with antipsychotic drugs, indicating the role played by 5-HT system and the 5-HT1A receptor in negative symptoms and their treatment. Study partially supported by a grant from Fundacio´ La Marato´ (01/ 5330).

P18.12 ADENOSINE A2A RECEPTOR GENE POLYMORPHISMS ARE ASSOCIATED WITH INCREASED ANXIETY AFTER ACUTE AMPHETAMINE MEDICATION Hohoff C,1 McDonald JM,2 Baune BT,1 Cook EH,2 Deckert J,1 and de Wit H2 1 Department of Psychiatry, Molecular Psychiatry, University of Mu¨nster, Mu¨nster, Germany 2 Department of Psychiatry, The University of Chicago, IL, Chicago, United States Amphetamine is a psychostimulant drug that increases feelings of energy, well-being and alertness. Some individuals report increased anxiety or nervousness after amphetamine. These effects are assumed to be mainly mediated by the dopamine and norepinephrine systems. Dopamine D2 and adenosine A2a receptors form a functional heterodimer complex in cell membranes. In the present study we therefore studied the relevance of two adenosine A2a receptor gene polymorphisms for the interindividual variability in subjective responses to amphetamine. A group of 99 healthy young volunteers received placebo or d-amphetamine (10 mg or 20 mg, respectively) double-blind, in randomised order and under standardised conditions. Self-report questionnaires on subjective mood states using the Profile of Mood States (POMS) were obtained. The A2a polymorphisms 1976C/T (formerly known as 1083C/T) and 2592C/Tins were genotyped by means of PCR-based restriction fragment length polymorphism (RFLP)assays or single strand conformation polymorphism (SSCP)-analysis. Both polymorphisms were in nearly complete linkage disequilibrium. For both polymorphisms an association with increases in the moods anxiety and arousal was detected, in anxiety at both amphetamine doses and in arousal at the 10 mg amphetamine dose. In contrast, no significant association could be detected for the other moods. No difference in the demographic measures age, gender, race/ethnicity, BMI, education and current drug use was observed between the genotypic groups. These findings are consistent with observations indicating a role for A2a receptor gene polymorphisms in panic disorder (Deckert et al. 1998) and variability in anxiety after acute caffeine administration (Alsene et al., 2003). The polymorphisms might influence transcription efficiency or mRNA stability. Alternatively, they may reflect linkage disequilibrium with other functionally relevant variants. Replication studies in independent samples and functional characterization of the two polymorphisms are necessary.

P18.13 GENOTYPING OF THE CYTOCHROME-P450 (CYP) ENZYMES CYP2D6 AND CYP2C19 FOR INDIVIDUALISATION OF TREATMENT IN A LARGE PSYCHIATRIC HOSPITAL Rasmussen HB, Jacobsen KD, and Werge T Research Institute of Biological Psychiatry, Sct. Hans Hospital, Roskilde, Denmark A large number of antidepressants and antipsychotics are metabolised by the cytochrome-P450 (CYP) enzymes, CYP2D6 and CYP2C19. These enzymes are genetically polymorphic resulting in large individual variations in metabolic capacity. To individualise and improve treatment with psychotropic drugs genotyping of CYP2C6 and CYP2C19 has been carried out for more than one year in our hospital. This hospital is a large and specialised psychiatric unit with about 400 patients many of whom suffer from schizophrenia in a severe form. Patients who do not response properly to medical treatment or develop adverse drug reactions as well as all newly admitted patients are offered a genetic screening for inactive variants of CYP2D6 and CYP2C19. We screen for the four most common defective alleles of CYP2D6 (*3, *4, *5 and *6), duplication of the CYP2D6 gene and the three most common defective alleles of CYP2C19 (*2, *3 and *4). Blood samples are delivered to the laboratory once a week and the results are available two or three days later. Based upon CYP2D6 genotypes patients are classified as ultrarapid metabolisers (three or more active alleles), extensive metabolisers (two active alleles), intermediate metabolisers (one active allele) or poor metabolisers (no active alleles).

Abstracts With respect to CYP2C19 activity patients are classified into extensive/ intermediate metabolisers (one or two active allele) and poor metabolisers (no active alleles). All laboratory answers are accompanied by a set of proposals to the clinicians, including dose adjustments of CYP2D6 or CYP2C19 dependent drugs, switch to drugs not metabolised by these enzymes and monitoring of serum drug levels. We are currently evaluating the possible advantages of CYP2D6 and CY2C19 genotyping by review of the pharmacy records of the patients in question. Various measures are used for this purpose, such as frequency of prescriptions of drugs metabolised by CYP2D6 or CYP2C19 and use of drugs against Parkinsonian-like side effects. So far, we have genotyped more than 200 patients. Experiences with CYP2D6 and CYP2C19 genotyping to improve psychiatric treatment are to be given. P18.14 ASSOCIATION STUDY OF INPP1, 5HTT, BDNF, AP-2B AND GSK-3B GENE VARIANTS AND RESPONSE TO LITHIUM PROPHYLAXIS IN BIPOLAR DISORDER Michelon L,1 Meira-Lima I,1 Miguita K,1 Ikenaga E,1 Breen G,2 Collier D,2 and Vallada H1 1 University of Sao Paulo—Institute & Department of Psychiatry, Sao Paulo, Brazil 2 King’s College London—Institute of Psychiaty, London, United Kingdom Lithium is the most widely used mood stabilizing agent and its efficacy in preventing relapses in Bipolar Disorder (BD) is well established. Most patients, however, do not show a full response to lithium. Despite the fact that some clinical features can predict a favorable long-term prophylaxis, genetic factors have emerged as potentially strong predictors of lithium prophylactic efficacy. In the present study we investigated the influence of a series of candidate genes thought to be involved in lithium’s actions in the CNS [the serotonin transporter (5-HTT), glycogen synthase kinase-3 beta (GSK-3b), inositol polyphosphatase 1-phosphate (INPP1), brain-derived neurotrophic factor (BDNF) and activatior protein 2 isoform b (AP-2b)] on the efficacy of long-term lithium response in bipolar patients. Response criteria were defined according to clinical parameters and lithium levels. Of the 134 unrelated patients with DSM-IV BD type I patients in the study, sixtyone patients were considered full responders, forty-nine non-responders and twenty-four partial responders. The following polymorphisms were analysed: the 44 bp insersion/deletion in the promoter region of 5HTT, the 1727A/T for GSK-3b, the C973A for INPP1, the G196A for BDNF, the VNTR [CAAA]4/5 for AP-2b. No significant difference was observed in the allelic or genotypic distribution between the three response groups for all the polymorphisms analysed in the five candidate genes. Therefore, we conclude that these variants are not major predictive factors for lithium response in our sample. P18.15 INVESTIGATING THE INFLUENCE OF CHAT GENE POLYMORPHISMS IN SCHIZOPHRENIA AND OLANZAPINE DRUG THERAPY AMONG BASQUE AND SPANISH PATIENTS Mancama D,1 Arranz MJ,1 Munro J,1 Mata I,2 and Kerwin R1 1 Institute of Psychiatry, Kings College, London, United Kingdom 2 Fundacion Argibide, Navarra, Spain Choline acetyltransferase (ChAt) is responsible for synthesizing acetylcholine, the neurotransmitter that mediates muscarinic and nicotinic cholinergic activity. Widely distributed throughout the brain, ChAt activity modulates levels of acetylcholine to play an important role in sleep, movement, and cognitive functions that include attention, learning and memory. Recent studies have identified salient evidence pointing to gross alterations in the physiological integrity of ChAt among patients with schizophrenia. Most notably, these include significant reductions in levels of ChAt within the pontine tegmentum of such subjects, further to which these changes have been found to correlate with measures of cognitive decline in the disorder. The mechanisms underlying these changes are unknown, and the extent to which they influence the disorders susceptibility and treatment remain uninvestigated. Given the existing evidence, we hypothesized that polymorphisms within ChAt may be important to determining susceptibility to schizophrenia, and furthermore may influence the outcome of such patients to antipsychotic therapy. To investigate this potential we screened the ChAt gene for polymorphisms of potential interest, and investigated two (designated ChAtPP1/RsaI and ChAtPP3/TaqI) for influence on schizophrenia susceptibility and

161

treatment outcome in Basque (n ¼ 245) and Spanish (n ¼ 108) patients and controls. All individuals undergoing olanzapine treatment were prospectively rated for response according to GAF and PANSS scales, and statistical analyses performed using chi-square tests and regression analysis. Upon examining each polymorphism in relation to olanzapine outcome among Basque patients, no evidence for association was found between genotype status and drug response (ChAtPP1 chi square ¼ 0.07, P ¼ 0.97, regression F ¼ 0.05, P ¼ 0.83, ChAtPP3 chi square ¼ 0.8, P ¼ 0.5 regression F ¼ 0.30, P ¼ 0.60). This lack of association was also observed for combined (stratified) Basque and Spanish groups (ChAtPP1 P ¼ 0.67, ChAtPP3 P ¼ 0.59). Similarly, no evidence was found for an influence of ChAt variants on improvement in positive symptoms, negative symptoms, and general psychotic symptoms during treatment, further to which the polymorphisms did not appear to significantly influence susceptibility to schizophrenia. These observations were confirmed upon analysis of haplotypes. We conclude from our findings that the ChAt polymorphisms we investigated do not significantly influence susceptibility to schizophrenia or olanzapine outcome in our sample population of Basque and Spanish patients. Further studies however are required to investigate these findings in other populations, while the potential presence of other polymorphisms in this gene of more significant importance cannot be ruled out. P18.16 ALLELIC VARIATIONS IN CYP450 ENZYMES AND OLANZAPINE RESPONSE Staddon S,1 English J,1 Lopez-Ilundain JM,2 Mata I,2 Arranz MJ,1 and Kerwin RW1 1 Clinical Neuropharmacology, Institute of Psychiatry, London, United Kingdom 2 Fundacion Argibide, Pamplona, Spain The atypical antipsychotic olanzapine is metabolized in phase I by members of the CYP450 family, including CYP1A2 and CYP2D6. In this study, we have investigated the influence of polymorphisms in these enzymes in a group of psychotic patients of Basque origin diagnosed with schizophrenia (n ¼ 70). After a minimum of three months of olanzapine treatment patients were classified as responders or nonresponders by differences in GAS scores and improvement in symptoms was assessed with the Positive and Negative Syndrome scale (PANSS). Two polymorphisms of CYP1A2 were investigated, 1A2*C and 1A2*F. CYP1A2 *C is located in the 50 flanking region of the gene and codes for a G ! A point mutation. This variant was found to cause decreased enzyme inducibility in a Japanese population (Nakajima et al., 1999). In our sample, one patient was a G ! C heterozygote, the remainder were homozygous for the G allele. The CYP1A2* F variant, located in the first intron of the gene, involves an A ! C substitution. The A allele was found to be associated with increased enzyme activity in a study carried out in a Caucasian population (Sachse et al., 1999). In our sample the A/A genotype was more frequent in the olanzapine nonresponders than responders (46% versus 36%) but this difference was not significant when genotype or allele frequency were considered (2 ¼ 2.95, P ¼ 0.23 for genotype, ¼ 2.35, P ¼ 0.13 for allele). No association was found between genotype and improvement in symptoms (total PANSS, positive PANSS, negative PANSS, general psychosis PANSS: F < 2.5, P ¼ ns). CYP2D6 is the most extensively investigated CYP450 enzyme and is involved in the metabolism of almost all antipsychotic drugs. CYP2D6*4 is the most common variant, coding for a G ! A base change causing a splice site mutation resulting in a truncated protein and decreased enzyme activity. The frequency of olanzapine nonresponders versus nonresponders carrying the rare A allele was virtually equal (35% in nonresponders versus 31% in responders), However, a trend of association was found between genotype and total improvement in symptoms (totPANSS: F ¼ 2.8, P ¼ 0.10) and marginal association with improvement in positive symptoms (PPANSS: F ¼ 4.12, P ¼ 0.05). A less frequent polymorphism is the CYP2D6*3A variant with a base pair deletion resulting in a frameshift mutation. No rare alleles for this polymorphism were found in the patient sample. Our results suggest that CYP2D6*4 polymorphism may exert a mild effect on olanzapine response. P18.17 EXPLORING THE ROLE OF DOPAMINE D1 AND D2 RECEPTOR GENE POLYMORPHISMS IN ANTIPSYCHOTIC TREATMENT RESPONSE Hwang RW,1 Shinkai T,1 Deluca V,1 Meltzer H,2 Lieberman J,3 Volavka J,4 Czobor P,4 Ni X,1 and Kennedy JL1

162

Abstracts

1

Neurogenetics Section, Centre for Addiction and Mental Health, University of Toronto, Toronto, Canada Case Western Reserve University, Cleveland, United States 3 University of North Carolina at Chapel Hill, Chapel Hill, United States 4 Nathan S. Kline Institute for Psychiatry Research, Orangeburg, United States 2

The purpose of this study is to explore the role of dopamine D1 and D2 receptor gene variants on antipsychotic treatment response, which has been observed to be highly variable in patients with schizophrenia (SCZ). Four single nucleotide polymorphisms (SNPs) in DRD1 and 11 SNPs spanning DRD2 were studied. These SNPs were genotyped in a combined sample of 3 patient populations: 1) 82 DSM-IV SCZ subjects recruited by JL and JV given either clozapine, haloperidol, olanzapine or risperidone and assessed using the Brief Psychiatric Rating Scale (BPRS), 2) 105 treatment refractory DSM-IV SCZ subjects recruited by HM treated with clozapine and assessed using the BPRS, 3) 100 treatment refractory DSM-III-R SCZ subjects recruited by JL treated with clozapine and assessed using the BPRS. This study used a dichotomous responder/non-responder variable defined using a 20% improvement on the BPRS as the criterion (Kane, 1989). SPSS, EH and COCAPHASE statistics programs were used to identify SNPs and haplotypes that may predict treatment outcome. We found the 48 A/G SNP in DRD1 as well as a new SNP and the well-known TaqIA polymorphism of DRD2 to be significantly associated with treatment response. Two areas within the DRD2 gene were also identified that contain haplotypes strongly associated with response. An interaction was also found between DRD1 and DRD2. These exploratory results support the role of DRD1 and especially DRD2 in contributing to the observed variability in antipsychotic treatment response as well as illuminating exciting areas for future psychopharmacogenetic studies to focus on.

P18.18 MULTIFACTOR DIMENSIONALITY REDUCTION REVEALS GENE-GENE INTERACTION ASSOCIATED WITH RESPONSE TO TREATMENT OF UNIPOLAR MAJOR DEPRESSION DISORDER Ritchie MD, Mazei M, Hazelwood LA, Myers RL, Prasad HC, Sanders-Bush E, and Shelton R Vanderbilt University, Nashville, United States Unipolar major depression disorder (MDD) affects more than 9.9 million Americans each year. The causal mechanism of MDD is unknown, but we do know it is a result of a combination of factors related to brain chemistry, family history, and psychosocial environment. While we are not certain which factors dominate, we do know that MDD is treatable. Numerous antidepressant therapies are currently available, many of which work quite well. Unfortunately, some individuals with MDD do not respond to certain antidepressant drugs. A variety of factors may contribute to this lack of response, however it has been hypothesized that genetic variations may contribute treatment response. There is evidence that an individuals’ response to drug treatment can be explained, in part, by their genetic variation in certain areas of the genome. Pharmacogenomics holds the promise for individualized medicine, where drugs might one day be tailor-made for individuals and adapted to each person’s own genetic makeup. The key is to identify the genetic factors that contribute to drug response so that we can screen individuals before treatment is administered and prescribe the appropriate treatment based on genotype. To detect genetic variation associated with treatment response, we have used the Multifactor Dimensionality Reduction (MDR) method. In brief, MDR is a method that reduces the dimensionality of multilocus information to identify gene-gene and gene-environment interactions associated with an increased risk of disease. In the current study, we have ascertained a group of 80 subjects (mean age 43, 60% female) with MDD using the Hamilton Rating Scale for Depression (HAM-D) and the NEO-Five Factor Inventory. Individuals were treated for MDD with serotonergic antidepressants and response was measured at 8 weeks by repeating the HAM-D. Polymorphic variants of the serotonin transporter (SERT, 5HTTLPR), dopamine transporter (DAT), serotonin (5HT) 2A and 2C receptors were characterized. Using MDR, we detected a twopolymorphism interaction between 5HT2A and 5HT452HY polymorphisms. This model correctly predicted who would respond to serotonergic treatment with 70% accuracy (P < 0.06). This model is marginally significant and with an increased sample size is likely to reach highly

significant levels. Thus, using methods such as MDR we can uncover gene-gene and gene-environment interactions associated with drug treatment response. Once genetic variations that contribute to treatment response are identified, appropriate treatment regimens can be prescribed and individualized medicine will emerge.

P18.19 THE CYTOKINE NETWORK GENES AND RESISTANCE TO ANTIPSYCHOTICS IN SCHIZOPHRENIA Bignotti S, Bocchio-Chiavetto L, Zanardini R, Bonvicini C, Randazzo R, Tura GB, Rossi G, Perez J, and Gennarelli M 1 IRCCS S.Giovanni di Dio FBF, Brescia, Italy 2 Division of Biology and Genetics, Dept Biomed Biotech, University of Brescia, Brescia, Italy Cytokines are involved in normal brain development acting directly on neural cells or modulating neurotransmitter systems and neurotrophins. Dysregulation of the inflammatory response system has been related to pathophysiology of schizophrenia and a growing number of studies have associated both proinflammatory (IL-1, TNFa) and antiinflammatory (IL-1RA, IL-10) cytokine genes with susceptibility to this illness. Several evidences have involved cytokines in the mechanism of action of antipsychotic drugs. Atypical antipsychotics as clozapine, risperidone and olanzapine have shown an immunomodulatory effect on cytokine peripheral levels and on mononuclear cells cytokine secretion. Moreover, abnormalities in plasma cytokines are more pronounced in treatment resistant schizophrenia. All these evidences together suggest that cytokine genes could be first choice candidates for pharmacogenetics studies on antipsychotics. In order to verify if allelic variants in cytokine genes may influence the therapeutic response, we enrolled in the study 150 schizophrenic inpatients. All the patients were recruited from the Psychiatric Rehabilitation Centre of IRCCSFBF (Brescia, Italy) and met the diagnostic criteria of DSM-IV for schizophrenia. Schizophrenic patients were defined drug resistants (61/150) according to the criteria described by Kane et al., 1988 and Dixon et al., 1995. We genotyped the patient sample for seven different polymorphisms that had showed a functional significance: A-308G in TNF-alpha, G-174C in IL-6, C-889T in IL-1alpha, C-511T in IL-1beta, VNTR in IL-1RA, G-1082A in IL-10 and C-592A in IL-10. Our results do not support associations of different gene polymorphisms independently. However, we found an epistatic interaction of TNF-alpha IL-6, IL-1beta, IL-1RA and IL-10 genes on antipsychotic resistance.

P18.20 GENETIC PREDICTORS OF RESPONSE TO LITHIUM IN BIPOLAR DISORDER Kelsoe J, McKinney R, Shekhtman T, Gaucher M, and Smith G Univeristy of California, San Diego and San Diego VA Healthcare System, San Diego, United States Lithium is the first mood stabilizer and the one for which there exists the most efficacy data. Though most patients with bipolar disorder respond well, others fail to respond or to tolerate its side effects. Largely because of side effects, its use has been supplanted in some countries by other agents. It has been argued that genetic factors may be important in determining response. A pharmacogenetic panel of DNA tests capable of predicting response would be of great clinical value in optimizing the treatment of bipolar disorder. Towards this, we have identified 92 patients with bipolar disorder that had a good response to lithium and 92 with a poor or incomplete response. These patients were identified largely from probands of families ascertained for linkage studies. Lithium response was assessed retrospectively by the lifechart method, chart review and interview of patients. A variety of clinical parameters regarding course of illness were also extracted. Response was determined by consensus review of all data. 86 SNP markers and one repeat marker were selected in nine candidate genes. SNPs were genotyped using Taqman and the repeat by fluorescent methods. Categorical data was analyzed by Chi square contingency table and continous variables by t test. Consistent with previous literature patients with a predominance of dysphoric mania or rapid cycling did not respond well to lithium. A novel finding was that patients with comorbid PTSD also did not respond well to lithium. Contrary to previous reports the 50 repeat in the serotonin transporter gene (HTTLPR) did not predict response. SNPs in three genes were associated with lithium response with nominal significance. These include two inositol monophosphatases, (IMPA1, P ¼ 0.02) and (IMPA2, P ¼ 0.04) and the gene for neurotrophin receptor kinase 2 (NTRK2, 2 SNPs, P ¼ 0.006

Abstracts and P ¼ 0.04). Variation in these genes may play a role in the response to lithium.

P18.21 HOW DIFFERENT GENE VARIANTS ARE INVOLVED WITH THE DIFFERENT HAMILTON FACTORS OF ANTIDEPRESSANT RESPONSE? Insacco C, Artioli P, Mandelli L, Pirovano A, Serretti A, and Smeraldi E Department of Psychiatry, Vita-Salute University, San Raffaele Institute, Milan, Italy Antidepressant response is considered an important evaluation target in psychiatric research. Patients show a great variability in clinical improvement after treatment. Antidepressant response could not be considered a univocal variable: in some patients, in fact, it is limited to specific symptom clusters. Many efforts have been tried to identify the variables influencing this complex parameter. Among the polymorphic variables explaining these differences, genetic factors could play an important role. The aim of the present study is to evaluate if gene influence could be exerted on specific symptom clusters instead of on the entire symptomatology improvement after antidepressant treatment. 683 mood disorder patients were evaluated weekly for 6 weeks during antidepressant treatment, with HAMD-21 rating scale. They were genotyped for some polymorphisms (SERTPR, TPH-1 A218C, CLOCK T3111C, GSK3b T-50C, PER ex15 T1940G, PER ex18 ins/del, PER ex18 T3110C, Gb3 C825T, 5HT1a C1019G) which have previously been related to, or are supposed to be involved in specific depressive symptom clusters. PER ex15 T1940G *TG (P < 0.05), subjects showed a worse course in HAMD-21 psychic anxiety items (9-10), and similarly GSK3aˆ T-50C *CC (P < 0.05) subjects, also 5HT1a C1019G *GG (P < 0.05) was found to be associated to a worse course concerning the psychic anxiety items, while the presence of PER ex18 ins/del *55 genotype was associated to a worse course and a slower decrease in HAMD-21 core items (1-2-7-8-10-13). These results represent a confirmation of genetic influences on specific symptom clusters instead of on the entire symptomatology, and could be useful to better understand the influence of genetics in antidepressant response.

P18.22 COMT GENE (VAL 158 MET POLYMORPHISM) AND CLINICAL RESPONSE TO SSRIS IN DEPRESSIVE PATIENTS OF EUROPEAN ORIGIN Arias B,1 Serretti A,2 Lorenzi C,2 Gasto´ C,3 Catala´n R,3 Gutie´rrez B,1 and Fan˜ana´s L1 1 Unitat d’Antropologia. Dept Biologia Animal. Facultat de Biologia. Universitat de Barcelona., Barcelona, Spain 2 Department of Psychiatry, Vita-Salute University, San Raffaele Institute., Milan, Italy 3 CSM Esquerre de l’Eixample. Hospital Clı´nic de Barcelona., Barcelona, Spain SSRIs exert their antidepressive effect by blocking the neuronal serotonin transporter and increasing the availability of extracellular serotonin (5-HT), and in consequence 5-HT neurotransmission. However, clinical response to antidepressant treatment has been shown to be a complex phenomenon in which more than one neurotransmission system is involved. In this sense, there is convincing evidence of interactions between serotoninergic and dopaminergic systems. In fact, it seems that serotonergic projections inhibit dopamine function (release and probably synthesis) in the midbrain, striatum and cortex (Kapur and Remigton, 1996; Bonhomme and Esposito, 1998). The aim of our study was to analyse genetic the functional polymorphism on COMT gene and clinical response (4 and 6 weeks) and clinical remission (12 weeks) in depressive patients treated with SSRIs. The Val 158 Met polymorphism was analysed in two different samples, a) 259 major depressive inpatients (DSM-IV), consecutively admitted to the Mood Disorder Centre, Department of Psychiatry at the Institute H. San Raffaele, Milan, and 130 depressive outpatients (DSM-IV)from the Centre de Salud Mental of the Hospital Clinic de Barcelona. All patients were evaluated at baseline and along the follow up using the 21-item Hamilton Rating Scale for Depression (HAM-D). Italian patients were treated with paraxotine (20–40 mg/day) and fluvoxamine (300 mg/day) along 4 and 6 weeks (assessed weekly), respectively, Spanish patients were treated with citalopram (20–40 mg/day) and followed along 12 weeks (assessed monthly). Genotype distribution of the Val158Met polymorphism was analysed using an ANOVA repeated measures based on the HAM-D scores along the three follow-ups (4, 6 and

163

12 weeks). When we analyse genotype effect in short term response (four weeks -paroxetine- and six weeks-fluvoxamine-) we could not find any effect of genotype or genotype x time interaction. No differences were found when we explore this polymorphism in response (6 weeks) and remission (12 weeks) of the episode (citalopram) either. Although COMT gene (Val 66 Met) seems to have a role in neurocognitive functions and maybe pathophysiological aspects of schizophrenia. The preliminary study in Italian and Spanish samples does not seem to support its role in clinical response to SSRIs. However the existence of interactions between serotoninergic and dopaminergic systems should maintain the interest of this gene in depression or in clinical and neurocognitive aspects of the illness. This work has been supported by a grant from the Ministerio de Ciencia y Tecnologı´a (SAF 2001-3400).

P18.23 GENETIC VARIANTS IN THE ANGIOTENSIN-1CONVERTING ENZYME (ACE) AND THE ANGIOTENSIN II RECEPTOR (AT1) GENES INFLUENCE THERAPEUTIC OUTCOME IN MAJOR DEPRESSION Baghai TC, Zill P, Schule C, Eser D, Rupprecht R, and Bondy B Psychiatric Clinic of University, Munich, Germany Angiotensin converting enzyme (ACE) is expressed in the central nervous system, where it catalyses the synthesis of angiotensin II and cleaves many other substrates as e.g. the multi-functional substance P. Despite these direct effects there are a number of modulatory interactions between angiotensin II and other neurotransmitter systems which are known to be involved in a range of behavioural or cognitive processes and in the pathophysiology of psychiatric disorders. The broad array of effects of angiotensin II are mediated by a number of specific receptors, as e.g. the AT1 receptor, which are expressed in many brain regions, as in cortex, limbic and hypothalamic regions. Thus the central renin-angiotensin system is subject of both, pathophysiology of psychiatric disorders and response to treatment. Functionally active polymorphisms were identified for both genes. The I/D polymorphism of the ACE gene and a common polymorphism with an A ! C transition at position 1166 of the AT1 receptor gene were subject of several investigations. In our study with 313 depressed patients who were treated with various antidepressants, we have evaluated a possible impact of these polymorphisms on treatment response. We could show a divergent clinical outcome in relation to different ACE genotypes, with lower HAMD scores after 4 weeks of treatment in D/D- and I/D- carriers in comparison to I/I-genotypes (P ¼ 0.01) and also a shorter duration of hospitalization in D-allele carriers (P ¼ .001). After subdivision of the patients according to their gender only female patients contributed significantly to the genotype dependent therapeutic outcome. A similar gender effect was not observed for the AT1 A1166C polymorphism, where CC carriers showed a slightly better treatment response than the combined AC and AA genotypes. Our data further suggest that patients with a haplotype combining the CC and DD/ID genotype respond better to treatment than those with either single allele. Although the haplotype results are preliminary and have to be handled with care, this study could suggest a gene-gene interaction in the response to treatment. This project is supported by the German Federal Research Ministry within the promotional emphasis ‘‘Competence Nets in Medicine.’’

P18.24 ANALYSIS OF CRHR1, CRHR2 AND CRHBP GENES IN DEPRESSION AND THEIR ROLE IN THE OUTCOME OF DEPRESSIVE EPISODES TREATED WITH SSRIS Papiol S,1 Gutie´rrez B,1 Arias B,1 Catala´n R,2 Gasto´ C,2 Gonza´lez N,1 and Fan˜ana´s L1 1 Unitat d’Antropologia. Departament de Biologia Animal. Facultat de Biologia. Universitat de Barcelona., Barcelona, Spain 2 Centre de Salut Mental Esquerre de l’Eixample. Hospital Clı´nic i Provincial de Barcelona., Barcelona, Spain Hypothalamic-pituitary-adrenal (HPA) axis has been suggested to play a pivotal role in the aetiological basis of affective disorders, specially in the heterogeneous group of depressive and anxiety-related disorders. Likewise, some authors have pointed out that normalization of HPA disturbances is a prerequisite of stable remission of depressive symptoms after pharmacological treatment (Holsboer, 1996). Corticotropin-releasing hormone (CRH) system is involved in the modulation of the response to stress through HPA axis and other pathways. Knockout mice, clinical and association studies in major depression have

164

Abstracts

highlighted the role of genes coding for proteins that modulate this system, such as CRH receptors and CRHBP, in the etiopathology of affective disorders (Zobel, 2000; Villafuerte, 2002; Claes, 2003). The aim of the present study was to analyze the effect of the genetic variability of genes of the CRH system in the risk to develop major depression and in the response to SSRIs treatment. For this purpose we analyzed seven SNPs in CRHR1, CRHR2 and CRHBP genes using the ABI PRISM SNaPshot multiplex kit. Our sample consisted of 159 DSMIV major depressive patients and 96 healthy controls, all of them of Spanish origin. The patient sample (for further details see Arias, 2003) was recruited in the context of a long term pharmacogenetic follow-up study based on the SSRI citalopram. The severity of symptoms was evaluated during the follow-up (weeks 0, 4, 8 and 12) using a 21-item Hamilton Depression Rating Score (HDRS). HDRS scores at 4th and 12th week were used for determination of clinical response and remission, respectively. We were not able to detect any effect of the genetic variability of these genes on both the clinical response and remission. Additionally, there was no significant effect of any analyzed SNP in the outcome of the episode. Neither of the SNPs showed significant differences in genotype or allele frequencies between patients and controls. However, we found that a SNP(C/T, rs110402) of the CRHR1 gene was associated with severity traits of the illness. Specifically, patients T/T homozigous (18.6% of our sample) showed nearly 3 times more risk (OR ¼ 2.89 95% IC [1.07–8.32], P ¼ 0.02) to develop illness with a seasonal pattern, in addition, this group of patients were significantly younger when they suffered their fist episode (age at onset TT homoz. [25.9  10.1 years] vs. C/C or C/T [32.3  10.9 years] (P ¼ 0.007)). Our results suggest that variability on CRHR1 gene might contribute to define a subgroup of depressive patients in which genes of HPA axis would participate in the origin and course of the illness. This work was supported by a grant from Spanish Ministerio de Ciencia y Tecnologı´a (SAF 2001-3400).

P18.26 IN VIVO EXAMINATION OF THE INTRONIC CYP1A2 *1F VARIANT DOES NOT SUPPORT FUNCTIONAL EFFECTS OF THE ‘‘HIGH INDUCIBILITY’’ ALLELE ON GENE EXPRESSION Sand PG, Wechsung V, Stoertebecker P, Hajak G, and Eichhammer P University of Regensburg, Regensburg, Germany The cytochrome P450 enzyme CYP1A2 plays an important role in the metabolism of atypical antipsychotic drugs, including olanzapine and clozapine. As a result, sequence variation in the corresponding gene on chr15q22 has received considerable attention in pharmacogenetic studies. While functional effects of genetic variation are often inferred from metabolizer phenotypes, the true contribution of most individual variants remains unproven. A CYP1A2 polymorphism commonly investigated is the intronic *1F variant, which has been implicated in a number of pathologies, e.g. in tardive dyskinesia. We used transient transfection experiments to assess the impact of the CYP1A2*1F variant on gene expression in vivo in three mammalian cell lines. While no significant differences were noted in baseline activities of the two alleles under investigation, stimulation with PMA resulted in an increase in expression of the C-allele, but not the A-allele, previously termed the high inducibility allele. Overall, however, effect size was moderate and did not reach significance. The present findings are in line with a previous investigation which failed to show any in vivo effects of the *1F variation site using a different cell line (Aklillu et al., 2003). Together, these data are in contrast to earlier assessments of CYP1A2*1F alleles by parameters of caffeine metabolism and call for a cautious interpretation of such measures. With a growing number of pharmacogenetic studies and an increasing number of DNA variants under study, the use of cell-based reporter systems is encouraged to minimize the statistical noise from non-functional SNPs.

P18.27 ASSOCIATION OF A D2 RECEPTOR GENE POLYMORPHISM WITH SCHIZOPHRENIA AND TREATMENT RESPONSE IN THE BASQUE AND SPANISH POPULATIONS Parsons MJ,1 Mata I,2 Beperet M,2 Iribarren-Iriso F,2 Arranz MJ,1 and Kerwin RW1 1 Institute of Psychiatry, London, United Kingdom 2 Fundacion Argibide, Navarra, Spain

A number of converging methodologies including findings from postmortem, PET and pharmacology studies have highlighted the D2 receptor as a possible candidate gene in the pathophysiology of the schizophrenia. The involvement of the DRD2 in schizophrenics may be directly related to functional polymorphisms within the DRD2 gene that affect overall expression of the receptor protein such as a 30 flanking region, referred to as TaqIA, and a single nucleotide deletion located 141 bp from the 50 end of the encoded gene, referred to as 141C (Ins/Del). A recent study observed an excess of transmission of the TaqI A2 allele in schizophrenic patients, which we also observed in a Basque population isolate. The TaqI-A1 allele has also be associated with both nemonapride, a neuroleptic and hyperprolactinemia response and is associated with diminished DRD2 density and dopaminergic activity compared to the TaqI-A2 allele. The 141C Ins/Del, a functional polymorphism, has also been associated with schizophrenia in a number of studies, though these findings have not been replicated in all studies. Previous reports have failed to find an association between the 141C Ins/Del polymorphism and clinical response to antipsychotic treatment. Together these findings suggest that these D2 receptor polymorphisms may be associated with schizophrenia and may influence antipsychotic response. We tested these hypotheses by investigating the TaqI-A1/A2 and 141C Ins/Del polymorphisms in a population consisting of 266 controls and 188 schizophrenic patients of either a Spanish Caucasian or Basque ethnicity. Subsets of the above patients were evaluated for olanzapine response (N ¼ 107) and risperidone response (N ¼ 77). Regression analyses were preformed in the total sample using variation in PANSS values after treatment as the dependent variable and the genotypes and ethnicity as the covariant variables. There was a non-significant trend for an excess of the TaqIA2 allele in schizophrenic patients in the Spanish Caucasian sample (w2 ¼ 1.0, df ¼ 1, P ¼ 0.31). This excess was statistically significant in the Basque sample (w2 ¼ 7.0, df ¼ 1, P ¼ 0.008) and in a stratified analysis of both samples (w2 ¼ 6.8, df ¼ 1, P ¼ 0.00913.766, df ¼ 2, P ¼ 0.001). No significant associations were found between the 141C Ins/Del polymorphism and schizophrenia. Neither polymorphism was associated with treatment response following olanzapine or risperidone treatment. Following risperidone treatment, symptom variation was significantly associated with the TaqI A polymorphism covariant for the TOTPANSS (P ¼ 0.001), NPANSS (P ¼ 0.007) and PGPANSS (P ¼ 0.006), but not for the PPANSS (P ¼ 0.11). These results support the involvement of the D2 receptor in the pathology of schizophrenia and in symptom variation following risperidone treatment.

P18.28 DBH GENOTYPE IN DISULFIRAM TREATMENT FOR COCAINE DEPENDENCE Cubells JF, Chawarski MC, George T, and Schottenfled RS Dept of Human Genetics, Emory University, Atlanta, United States Background: Disulfiram inhibits dopamineb-hydroxylase (DBH), which converts dopamine to norepinephrine in brain reward and other pathways. We identified a single nucleotide polymorphism (SNP), 1021C > T, at the DBH locus accounting for almost 50% of the variance in plasma DBH, the T allele associates with low DBH activity. We hypothesized that 1021C > T genotype may affect response to disulfiram for treating cocaine dependence. Aims: To evaluate pharmacogenetic interactions between 1021C > T and disulfiram. Methods: Subjects dependent on cocaine and opioids (N ¼ 128) completed 2-week buprenorphine induction, were then maintained on buprenorphine 24 mg SL daily, provided weekly group drug counseling, and randomized to double-blind disulfiram 250 mg daily or placebo for 12 weeks. Primary outcomes: proportion cocaine-positive urine tests (COCþ) and treatment efficacy score (TES-number of cocaine-negative tests), based on 3x/wk urine toxicology testing. Subjects with CC genotype (high DBH, N ¼ 78) were compared with carriers of the T allele (low DBH: N ¼ 50). Because % COCþ during induction was correlated with % COCþ during treatment (r ¼ .73, P < .001), and differed between disulfiram vs. placebo groups within genotype cluster (high DBH: 48% vs. 60%, low DBH: 57% vs, 47%), subjects were stratified into high (>50% COCþ, N ¼ 64) and low cocaine severity during induction (50% COCþ, N ¼ 64) for data analyses. Results: In high- severity subjects, disulfiram vs. placebo was associated with greater reductions in % COCþ (mean ( SD): 69  26 vs. 82  23, P ¼ 0.21) and higher TES (7.9  7.3 vs. 3.7  5.0, P ¼ 0.13) in T carriers but not CC subjects (% COCþ: 88  22 vs. 83  25, TES: 2.3  3.7 vs. 4.2  7.5). For low severity subjects, cocaine use remained low and was

Abstracts not reduced by disulfiram or placebo in either genotype group. Conclusions: The findings suggest differential efficacy of disulfiram 250 mg daily in high-severity, T-carrier subjects, higher disulfiram dose may be needed for high DBH subjects, low baseline severity of cocaine use in some subjects creates difficulties for detecting medication effects. Supported by: R01 DA12422, K02 DA015766 and K24 DA00445, and the VA Connecticut MIRECC. P18.29 INVESTIGATING THE INFLUENCE OF CHAT GENE POLYMORPHISMS IN SCHIZOPHRENIA AND OLANZAPINE DRUG THERAPY AMONG BASQUE AND SPANISH PATIENTS Mancama D,1 Arranz MJ,1 Munro J,1 Mata I,2 and Kerwin R1 1 Institute of Psychiatry, Kings College, London, United Kingdom 2 Fundacion Argibide, Navarra, Spain Choline acetyltransferase (ChAt) is responsible for synthesizing acetylcholine, the neurotransmitter that mediates muscarinic and nicotinic cholinergic activity. Widely distributed throughout the brain, ChAt activity modulates levels of acetylcholine to play an important role in sleep, movement, and cognitive functions that include attention, learning and memory. Recent studies have identified salient evidence pointing to gross alterations in the physiological integrity of ChAt among patients with schizophrenia. Most notably, these include significant reductions in levels of ChAt within the pontine tegmentum of such subjects, further to which these changes have been found to correlate with measures of cognitive decline in the disorder. The mechanisms underlying these changes are unknown, and the extent to which they influence the disorders susceptibility and treatment remain uninvestigated. Given the existing evidence, we hypothesized that

165

polymorphisms within ChAt may be important to determining susceptibility to schizophrenia, and furthermore may influence the outcome of such patients to antipsychotic therapy. To investigate this potential we screened the ChAt gene for polymorphisms of potential interest, and investigated two (designated ChAtPP1/RsaI and ChAtPP3/TaqI) for influence on schizophrenia susceptibility and treatment outcome in Basque (n ¼ 245) and Spanish (n ¼ 108) patients and controls. All individuals undergoing olanzapine treatment were prospectively rated for response according to GAF and PANSS scales, and statistical analyses performed using chi-square tests and regression analysis. Upon examining each polymorphism in relation to olanzapine outcome among Basque patients, no evidence for association was found between genotype status and drug response (ChAtPP1 chi square ¼ 0.07, P ¼ 0.97, regression F ¼ 0.05, P ¼ 0.83, ChAtPP3 chi square ¼ 0.8, P ¼ 0.5 regression F ¼ 0.30, P ¼ 0.60). This lack of association was also observed for combined (stratified) Basque and Spanish groups (ChAtPP1 P ¼ 0.67, ChAtPP3 P ¼ 0.59). Similarly, no evidence was found for an influence of ChAt variants on improvement in positive symptoms, negative symptoms, and general psychotic symptoms during treatment, further to which the polymorphisms did not appear to significantly influence susceptibility to schizophrenia. These observations were confirmed upon analysis of haplotypes. We conclude from our findings that the ChAt polymorphisms we investigated do not significantly influence susceptibility to schizophrenia or olanzapine outcome in our sample population of Basque and Spanish patients. Further studies however are required to investigate these findings in other populations, while the potential presence of other polymorphisms in this gene of more significant importance cannot be ruled out.

Author Index Abbacchi, A.M., P10.17, 103 Abderrahim, H., P17.16, 157 Abe, H., P11.7, 109 Aberg, K., O4.4, 19 Abou Jamra, A., P2.46, 51 Abou Jamra, R., O8.3, 27; O8.5, 28; P5.34, 72; P14.58, 142 Abou-Sleiman, P., O3.3, 16 Abrahams, B.S., O7.1, 24; P17.8, 155 Abramova, L.I., P8.21, 86 Abramowitz, A., P10.9, 101 Abramson, R.K., P10.24, 105; O5.5, 21; P10.4, 99; P10.5, 100; P10.8, 101; P10.23, 105 Ackenheil, M., P1.5, 32; P12.12, 115; P15.6, 145; P15.11, 146 Ackerman, F., P5.28, 70 Adams, L., P16.3, 149 Adams, P., O1.6, 12 Adams, R., P13.12, 123 Adolfsson, R., O4.4, 19; P2.31, 47 Aggen, S.H., P14.26, 134 Aggleton, J., P9.30, 98 Aghalar, R.M., P8.31, 89 Aghalar-Rafael, M., P8.32, 89 Agranat-Meged, A.N., P5.17, 68 Aguilar, A., P5.2, 64 Aguilar, E., P13.19, 125; P13.20, 125; P13.27, 127 Aguilar-Garcı´a, A., P7.2, 74; P7.27, 80 Aguilera, M., P2.39, 49 Aguirre, A.J., P7.2, 74 Ahmadi, K.R., P5.3, 64 Aitchison, K.J., P1.26, 38 Ajmar, F., P2.35, 48 Akiskal, H., O2.8, 15 Akiyama, T., P2.37, 49; P8.16, 85 Aksenova, M.G., P8.29, 89 Akula, N., O8.4, 27 Albeniz, A., P8.7, 83; P13.13, 124; P13.14, 124 Albus, M., P2.15, 43; P7.31, 81; P14.14, 131 Aldinger, K.A., O2.1, 13; O4.1, 18 Alexander, M., P2.43, 50 Alexopoulus, J., P3.16, 55 Alfimova, M.V., P3.2, 52; P7.23, 79 Alfred, S.E., P10.19, 104 Allington, M., P18.3, 158 Almasy, L., O3.7, 17; P3.20, 57; P8.12, 84 Almeida, A.C.G., P4.8, 62 Als, T.D., P2.3, 40; P5.23, 69; P9.14, 93; P13.11, 123 Alvarenga, P.G., P5.29, 71 Alvarez, V., P5.8, 65 Amin, T., P9.26, 97 Anastopoulos, A., P10.15, 103 Anderson, B., P9.12, 93; P9.23, 96 Anderson, C., P1.2, 31 Anderson, S., P13.12, 123 Andreou, P., P3.13, 55 Andresen, K.A., P9.21, 95 Angst, J., P1.29, 39; P12.31, 120 Anthoni, H., P10.25, 106 Aoki, M., P8.3, 82 Apiquian, R., P7.2, 74; P7.27, 80 Apter, A., P5.1, 64 Aragues, M., P17.9, 155 Arbeitman, L., O5.1, 20; P9.20, 95; P10.6, 100 Ardekani, B., O8.6, 28

Arias, B., P2.39, 49; P14.47, 139; P18.22, 163; P18.24, 163 Armas, R., P3.21, 57; P8.12, 84; P14.45, 139 Armengol, L.L., P9.19, 95 Arnold, A., S6.2, 8 Arnold, P., P14.28, 135 Arnold, P.D., P5.24, 69 Arolt, V., P1.3, 32; P3.7, 53 Arranz, B., P18.11, 160 Arranz, M.J., P7.24, 79; P8.6, 83; P18.6, 159; P18.8, 159; P18.15, 161; P18.16, 161; P18.27, 164; P18.29, 165 Artioli, P., P1.10, 33; P4.2, 61; P4.6, 62; P4.7, 62; P12.17, 117; P18.21, 163 Aschauer, H.N., P10.11, 101; P13.7, 122 Asherson, P., P3.13, 55; P10.31, 107; P12.8, 114; P12.10, 115; P12.11, 115; P16.4, 150 Ashley-Koch, A., P10.15, 103; P10.23, 105 Ashley-Koch, A.A., P10.4, 99; P10.5, 100 Ashley-Koch, A.E., O5.5, 21; P10.24, 105 Atmella, I., P8.12, 84 Austin, J.C., O4.8, 20 Austin, L., O1.1, 11 Avila, M.T., P7.12, 76; P14.30, 135 Avramopoulos, D., O3.6, 17; P4.14, 63 Ayoubi, M., P16.8, 150; P14.46, 139 Azevedo, M.H., O4.1, 18; P2.2, 40; P7.21, 78; P14.41, 138 Azuma, R., P12.6, 114; P16.12, 152 Baas, F., P3.30, 59 Babb, C.M., P3.16, 55 Bacanu, S., O4.5, 19 Bacanu, S.-A., P14.53, 141 Bachani, A., P14.16, 131; P14.18, 132 Bachner-Melman, R., O6.6, 24; P11.5, 109; P12.4, 113; P12.5, 114 Badenhop, R.F., P2.20, 45 Bader, M., P16.11, 151 Badı´a, A., P5.21, 68; P5.22, 69 Badner, J., P2.40, 50 Bagger, Y.Z., P3.28, 59 Baghai, T., O1.8, 13; P15.6, 145 Baghai, T.C., P1.5, 32; P1.27, 38; P18.23, 163 Bailer, U., P13.7, 122 Bakker, S.C., P10.10, 101; P14.29, 135; P3.22, 57; P6.7, 73; P10.28, 106 Balderas, T.G., P8.12, 84 Baldwin, P., P14.10, 130; P14.41, 138; P14.60, 143 Ball, D., P11.17, 112 Ballereau, S., P13.12, 123 Bamne, M.N., P1.23, 37 Banaschewski, T., P10.31, 107 Banjeree, M., P14.65, 144 Bannoura, I., P8.19, 86 Barcikowska, M., P4.9, 62 Barden, N., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Baron, C., P9.8, 92 Baron-Cohen, S., O6.3, 23 Barr, C., P10.16, 103 Barr, C.L., P8.14, 85; P9.12, 93; P9.17, 94; P9.22, 96; P9.23, 96; P9.28, 97; P9.29, 98 Barrantes-Vidal, N., P14.48, 140 Barrett, T.B., P2.26, 46 Barta, C., P11.2, 108

In this index, the numbers in italics are the pages in which the abstracts appear.

166

Bartels, M., O6.4, 23 Bascaran, M.T., P5.8, 65 Baschirotto, C., P12.29, 120 Bass, N., O7.7, 26; P1.14, 35; P2.36, 49; P8.18, 86; P14.8, 129 Bassett, S.S., O3.6, 17; P4.14, 63; P14.57, 142 Bastos, M.F., P14.3, 128 Bates, J., O2.3, 14 Battaglia, M., P12.29, 120 Bau, C.H.D., P9.11, 93 Baud, P., P1.31, 39; P1.32, 39 Baune, B.T., P1.3, 32; P18.12, 160 Baye´s, M., P17.14, 156 Beasley, C.L., P13.4, 121; P13.5, 121 Becker, K.D., O3.2, 16 Becker, T., O8.3, 27; O8.5, 28; P5.34, 72; P14.40, 137; P14.58, 142 Beckmann, H., P3.17, 56 Beekman, J., P13.12, 123 Beem, A.L., O9.8, 31 Beemer, F.A., P6.6, 73; P6.7, 73 Begg, E., P18.3, 158 Beirut, L., P2.40, 50 Beissbarth, T., P8.9, 83 Bellgrove, M.A., P3.5, 53 Bellivier, F., P1.17, 35; P1.31, 39; P1.32, 39; P2.15, 43; P3.26, 58 Bellone, E., P2.35, 48 Belmaker, R.H., P8.19, 86 Belmonte-de-Abreu, P.S., P7.21, 78; P9.11, 93 Ben Amor, L., P9.8, 92 Benedetti, F., P1.10, 33; P1.24, 37; P12.17, 117; P18.10, 160 Bening Abu-Shach, U., P8.19, 86; P13.17, 125 Benner, A., P9.4, 91; P9.7, 91 Bennett, P., P2.44, 51 Benton, J., P10.5, 100 Beperet, M., P8.6, 83; P8.7, 83; P13.13, 124; P13.14, 124; P18.6, 159; P18.27, 164 Berger, K., P1.3, 32 Bernardo, A., O3.5, 17; P4.11, 63 Berninger, V.W., O5.4, 21 Berrettini, W., P2.28, 47; P2.40, 50 Berrettini, W.H., P11.18, 112 Bertone, P.N., P15.18, 148 Bertram, L., O3.2, 16 Bertranpetit, J., P17.15, 157 Bespalova, I., O5.3, 21 Betancur, C., O5.6, 22; P10.21, 104 Betard, C., P2.15, 43 Bettecken, T., P8.11, 84; P14.6, 129 Beyer, K.S., P9.4, 91 Bhattacharyya, S., P5.18, 68; P5.19, 68 Bian, L., P9.1, 90 Bianchi, C., P2.35, 48 Biason, L., P9.6, 91 Biederman, J., O6.7, 24; P9.20, 95 Bienvenu, O.J., O6.8, 24 Bierut, L., P9.24, 96 Bignotti, S., P2.29, 47; P18.19, 162 Bilder, R.M., O7.3, 25 Binder, E., O8.3, 27; P5.4, 64; P13.8, 122 Binder, E.B., O1.8, 13; P1.28, 38; P5.10, 66; P5.13, 66; P5.16, 67 Birkas, E., P3.33, 60; P10.30, 107; P12.28, 119 Birley, A.J., O9.8, 31 Birtley, J.R., P6.1, 72

Author Index Bjarkam, C.R., P16.13, 152; P16.14, 152 Blacker, D., O3.2, 16 Blackwood, D., S3.1, 6; P1.2, 31; P2.15, 43 Blackwood, D.H.R., O1.2, 11; P1.1, 31; P6.1, 72; P6.2, 72; P6.5, 73; P13.10, 123; P13.12, 123; P14.2, 128; P15.1, 144 Blair, I.P., P2.20, 45 Blakely, R., P1.22, 37; P16.16, 153 Blanco, L., P2.3, 40 Blangero, J., O3.7, 17 Blennow, K., O3.1, 15 Bobes, J., P5.8, 65 Bocchio-Chiavetto, L., P2.29, 47; P18.19, 162 Bocklandt, S., S6.4A, 9 Boisen, K., P10.32, 107 Boivin, J., P12.16, 116 Boksa, P., P16.8, 150 Bolton, P., O6.3, 23; P9.5, 91; P9.9, 92 Bonaviri, G., O2.7, 15 Bondy, B., O1.8, 13; P1.27, 38; P1.5, 32; P11.6, 109; P11.12, 110; P12.12, 115; P15.6, 145; P15.11, 146; P18.23, 163 Bonneau, D., O5.6, 22 Bonnert, T.P., P8.17, 85 Bonvicini, C., P2.29, 47; P18.19, 162 Boomsma, D.I., O6.4, 23; O9.8, 31; P12.32, 120 Boor, K., P11.2, 108 Bordukalo-Niksic, T., P16.2, 149 Børglum, A., P2.3, 40 Børglum, A.D., P16.14, 152 Borkowska, A., P3.27, 59; P3.6, 53 Borozdina, S.A., P7.23, 79 Borrmann-Hassenbach, M., P7.31, 81; P14.14, 131 Boson, W., P12.1, 113; P12.2, 113 Botteron, K.N., P3.16, 55 Bottino, C., P4.8, 62 Bourgeron, T., O5.6, 22; P2.15, 43; P5.30, 71 Bousono, M., P5.8, 65 Boutelle, S., O1.6, 12 Boutin, P., P12.27, 119 Bowden, N.A., P15.3, 144; P15.4, 145 Bradwejn, J., P5.25, 69; P5.27, 70 Braeman, C., P1.6, 32; P13.15, 124 Brahamachari, S.K., P13.2, 121; P17.6, 154 Bramon, E., P3.8, 53; P3.9, 54; P12.13, 116 Brandon, N.J., P8.17, 85; P14.3, 128 Brandt, B., P3.7, 53 Bransome, M., O6.2, 23 Bray, N., P2.12, 42 Bray, N.J., O4.2, 18; O7.5, 25; P4.5, 61; P7.29, 81 Breen, G., P14.9, 129; P11.15, 111; P18.14, 161 Brewster, S., P1.13, 34 Brice, A., P10.21, 104 Bridler, R., P13.9, 123 Brkanac, Z., O5.4, 21 Brookes, K.J., P12.8, 114; P12.10, 115; P12.11, 115 Brooks, P., O9.1, 29 Brooks-Wilson, A., P17.8, 155 Browne, J., P9.9, 92 Bruchey, A., O3.5, 17 Bryden, L., O3.3, 16 Brzustowicz, L.M., P14.57, 142 Bubb, V.J., O7.4, 25; P15.7, 145 Buchanan, R.W., P7.12, 76 Buchanan, S., S3.1, 6 Bucholz, K., P9.24, 96 Buckland, P., P2.19, 44; P7.29, 81 Buckland, P.R., O4.2, 18; O7.5, 25; P4.5, 61 Buckley, P.F.B., P18.1, 157

Budaghzadeh, S., O7.1, 24 Buervenich, S., O8.4, 27 Buitelaar, J.K., P10.10, 101; P10.28, 106 Buitellaar, J., P10.31, 107 Bulgin, N., P9.18, 94; P10.16, 103 Bullmore, E., P3.3, 52 Bundo, M., P13.18, 125 Burdick, K.E., O2.3, 14; O7.3, 25 Buresi, C., P1.17, 35; P1.31, 39; P1.32, 39 Burge, S., P1.21, 36 Burgoyne, P.S., P16.1, 149 Burke, A., P2.17, 44 Busatto, G., P13.22, 126 Butcher, L.M., O9.2, 29; P3.23, 58 Buttenscho¨n, H.N., P2.3, 40; P13.10, 123 Bu¨ttner, A., P12.12, 115; P15.11, 146 Buxbaum, J., O5.3, 21; P7.29, 81 Byerley, W., P2.28, 47; P2.40, 50; P14.53, 141; P14.55, 142 Cabre´, M., P5.32, 71 Caesar, E.S., P1.19, 36; P2.22, 45; P3.25, 58 Caesar, S., O8.2, 27; P2.6, 41 Calati, R., P1.8, 33; P1.9, 33 Callahan, C., P14.18, 132 Callicott, J.H., S3.4, 7 Caltagirone, C., O2.7, 15 Camacho, A., P3.21, 57 Camarena, B., P5.2, 64; P5.33, 72; P7.27, 80; P17.10, 155 Camargo, A.C.M., P14.3, 128 Camargo, L.M., P8.17, 85 Camargo, M., P14.3, 128 Cameron, V., P15.5, 145; P18.3, 158 Campanera, S., P14.48, 140 Campbell, I.C., P5.9, 65 Campbell, L., P12.6, 114 Campbell, L.E., P12.19, 117; P16.12, 152 Campi-Azevedo, A.C., P12.1, 113 Can˜ive, J., P7.27, 80 Cannon, M., O2.6, 15 Cano, E., P3.21, 57 Cantor, C., S4.2, 7 Capano, L., P9.28, 97 Cardno, S., O8.7, 28 Carlett, O., P12.29, 120 Carlini, M., P2.29, 47; P2.35, 48 Carracedo, A., P17.12, 156 Carroll, L.S., P14.59, 143 Carson, R., P4.10, 62; P18.9, 159 Case, V., P15.3, 144 Caspi, A., PL2, 3; S8.1, 10; O2.6, 15 Cassidy, F., P2.7, 41 Castelo, A., P11.11, 110 Castermans, D., O7.2, 25 Castillo, A.R.G.L., P9.6, 91 Castillo, J.C.R., P9.6, 91 Castro, R., P8.12, 84 Catala´n, R., P2.39, 49; P18.22, 163; P18.24, 163 Catalano, M., P1.11, 34; P18.10, 160 Cate-Carter, T., P9.12, 93; P9.23, 96 Cavallaro, R., P18.7, 159 Cavallini, M.C., P18.7, 159 Cayer, M., P2.38, 49 Cerrato, C., P9.19, 95 Cesareni, G., P8.17, 85 Chacon, P., P5.28, 70; P5.29, 71 Chagnon, Y., P2.38, 49 Chan, A.O.M., P7.8, 75 Chan, R.C.K., P7.13, 77 Chang, J., O1.3, 11 Chant, D.C., P3.15, 55 Chapman, N.H., O5.4, 21 Charraverty, S., S1.3, 5

167

Chau, T.W.Y., P7.13, 77 Chawarski, M.C., P18.28, 164 Chen, C.-K., P12.10, 115 Chen, E.Y.H., P7.13, 77 Chen, H., P2.16, 44 Chen, J., P12.24, 119; P17.8, 155 Chen, J.J., P14.5, 128; P14.19, 132 Chen, Q., O4.3, 18; P7.3, 74; P7.6, 75; P7.10, 76 Chen, Q.Y., O4.3, 18; P7.3, 74 Chen, R.Y.L., P7.13, 77 Chen, W., P3.13, 55; P10.31, 107 Chen, W.J., P14.5, 128 Chen, W.M., P2.18, 44 Chen, W.X., P7.5, 75 Chen, W.Y., P7.7, 75; P7.19, 78 Chen, X., P14.54, 141; P14.57, 142 Chen, Y.S., O1.1, 11 Cheng, J., P14.49, 140 Cheng, S.Y., P16.6, 150; P16.7, 150 Chetcuti, A., P16.3, 149 Chevalier, F., P3.26, 58; P14.32, 136 Chillotti, C., P2.1, 39 Chitnis, X., P3.3, 52 Chiu, J., P14.38, 137 Chlubek, D., P4.9, 62 Chmara, M., P14.33, 136 Cho, E.Y., P8.1, 81 Cho, H.J., P2.25, 46 Cho, S.H., P8.1, 81 Choi, K.S., P8.1, 81 Cholostoy, A., P8.19, 86; P13.17, 125 Chong, S.A., P7.8, 75 Chorbov, V., P1.15, 35; P2.14, 43 Choudhury, K., P2.36, 49; P8.18, 86 Chowdari, K.V., O4.5, 19; P1.23, 37; P13.21, 126; P14.51, 141; P14.52, 141; P14.57, 142 Christiansen, C., P3.28, 59 Christie, S., S3.1, 6 Chu, F., P14.38, 137 Chuang, M.J., P14.5, 128 Chubb, J., P7.22, 79 Chumakov, I., P17.16, 157 Cichon, C., P5.34, 72 Cichon, S., O1.4, 12; O8.3, 27; O8.5, 28; P1.6, 32; P2.46, 51; P10.25, 106; P13.15, 124; P14.40, 137; P14.58, 142 Cicin-Sain, L., P16.2, 149 Ciotti, P., P2.35, 48 Claes, S., P2.31, 47 Clair, D.S., P9.1, 90 Clark, D.A., P7.24, 79; P18.6, 159 Clark, R.F., P4.12, 63 Clarke, S., P14.10, 130; P14.41, 138; P14.60, 143 Cloninger, C.R., P11.14, 111 Close, J., P15.9, 146 Close, J.P., P12.15, 116 Cocchi, F., P18.7, 159 Coelho, I., P2.2, 40 Cohen, D., P17.16, 157 Cohen, S., P1.25, 37 Colineaux, C., O5.6, 22 Collier, D., P3.3, 52; P3.9, 54; P5.31, 71; P7.13, 77; P8.4, 82; P13.24, 126; P14.9, 129; P14.37, 137; P18.14, 161 Collier, D.A., P5.9, 65; P14.15, 131; P14.36, 137; P14.57, 142; P15.8, 146 Colloura, V., P8.17, 85 Colombo, C., P18.10, 160 Comas, D., P17.15, 157 Condie, A., O1.2, 11; P14.2, 128 Congiu, D., P2.1, 39

168

Author Index

Conroy, J., O5.2, 21; P10.1, 99; P10.3, 99; P14.16, 131 Constantino, J.N., P10.17, 103 Contreras, S., P3.21, 57; P8.12, 84; P14.13, 131; P14.45, 139 Cook, C.C.H., O3.8, 17 Cook, E., P17.13, 156 Cook, E.H., P18.12, 160 Cooper, C., P1.18, 36; P2.22, 45 Cooper, C.A., P3.25, 58 Cope, H., P10.32, 107; P10.8, 101 Cope, N., P9.16, 94 Cope, N.A., P9.13, 93 Cordeiro, Q., P2.25, 46; P4.8, 62; P5.20, 68; P11.11, 110; P13.21, 126; P13.22, 126; P13.23, 126; P14.57, 142 Cordell, H.J., S2.3, 6 Corder, E.H., P14.21, 133 Correa, H., P12.1, 113; P12.2, 113 Correia, C., P10.22, 105 Corsico, A., P1.25, 37 Corte´s, M.J., P3.24, 58 Corvin, A., O1.5, 12; O4.2, 18; O9.6, 30; P2.44, 51; P2.45, 51; P3.19, 56; P3.31, 60; P13.26, 127; P14.60, 143 Corvin, A.P., P14.10, 130; P14.41, 138; P14.57, 142 Coryell, W., P2.28, 47; P2.40, 50 Costas, J., P17.12, 156; P17.14, 156 Coto, E., P5.8, 65 Cotter, D., P13.4, 121; P13.5, 121 Courtet, P., P1.17, 35; P1.31, 39; P1.32, 39 Coutinho, A., P17.13, 156 Couto, J.M., P9.12, 93; P9.23, 96 Cowell, J.K., P14.16, 131 Cox, N.J., O1.1, 11 Craddock, M., P1.19, 36; P3.25, 58 Craddock, N., E1.2, 3; O1.5, 12; O4.7, 20; O5.8, 22; O8.1, 26; O8.2, 27; O8.7, 28; O9.4, 29; O9.6, 30; P1.13, 34; P1.18, 36; P1.19, 36; P1.21, 36; P1.25, 37; P1.26, 38; P2.6, 41; P2.11, 42; P2.12, 42; P2.19, 44; P2.22, 45; P2.33, 48; P2.44, 51; P2.45, 51; P3.19, 56; P3.25, 58; P12.24, 119; P13.26, 127; P17.3, 154 Craig, D., P4.10, 62 Craig, G., P12.22, 118 Craig, I., O9.2, 29; P1.18, 36; P1.25, 37 Craig, I.W., S6.4B, 9; S8.1, 10; P12.21, 118; P12.22, 118; P13.25, 126; P15.2, 144 Cranfield, M., P14.39, 137 Creemers, J.W., O7.2, 25 Cristinzio, C., P4.14, 63 Cristobo, I.J., P17.12, 156; P17.14, 156 Crombie, C., P14.15, 131 Crosbie, J., P9.23, 96 Crossland, N., P8.25, 88 Crow, T.J., P3.12, 54; P7.15, 77; P8.23, 87; P10.14, 102; P12.15, 116; P14.39, 137; P15.9, 146 Crowbie, C., P8.4, 82 Crowe, R.R., O1.6, 12; O4.6, 19 Cruts, M., S1.1, 5 Cruz, C., P17.10, 155 Cruz Carlos, C.C., P7.2, 74 Ct Clair, D.M., P14.62, 143 Cubells, J.F., P15.18, 148; P18.28, 164 Cubero, P., P17.9, 155 Cuccaro, M., P10.15, 103 Cuccaro, M.L., O5.5, 21; P10.4, 99; P10.5, 100; P10.8, 101; P10.23, 105; P10.24, 105; P10.32, 107 Cuesta, M.J., P14.42, 138; P14.47, 139 Cui, L., P1.29, 39; P12.31, 120

Cullen, B., O6.8, 24 Cunha, N., P11.11, 110 Curi, M., P5.28, 70; P5.29, 71 Curran, S.R., P6.3, 73; P9.5, 91; P9.9, 92 Curtis, D., O7.7, 26; P1.14, 35; P2.10, 42; P2.36, 49; P8.18, 86; P14.8, 129 Cusack, B., P3.19, 56 Cutter, W.J., P10.14, 102 Czerski, P.C., P2.32, 48; P5.6, 65; P5.7, 65 Czerski, P.M., O8.5, 28; P2.21, 45; P2.34, 48; P3.27, 59; P3.6, 53; P14.33, 136; P14.40, 137 Czobor, P., P3.34, 60; P18.17, 161 da Silva, T., P10.7, 100 Dahdouh, F., P10.25, 106 Dahl, H.A., P5.23, 69; P9.14, 93; P13.11, 123 Dahmen, N., P11.1, 108; P14.6, 129 Dale, P.S., P3.23, 58 Dallapiccola, B., O2.7, 15 Daly, E., P12.6, 114 Daly, E.M., P10.14, 102; P16.12, 152 Daly, M., O5.1, 20 Daly, M.J., S2.2, 6; S6.3, 8; O2.1, 13; O4.1, 18; P2.13, 43 Danckaerts, M., P10.13, 102 Danilovich, E., P13.6, 122 Dash, D., P17.6, 154 Da-Silva, U., P4.11, 63 Dassori, A., P3.21, 57; P8.12, 84; P14.13, 131; P14.45, 139 Datta, S., P17.7, 155 Datta, S.R., O7.7, 26; P8.18, 86; P14.8, 129 Dave, S., P1.19, 36 Davies, N., P3.18, 56; P14.44, 139 Davies, W., P16.1, 149 Davis, K.L., O5.3, 21 Dayal, M., O4.5, 19; P3.20, 57 de Frutos, R., P13.19, 125; P13.20, 125; P13.27, 127 de Kovel, C.G.F., P10.28, 106 De la Fuente, C., P17.5, 154 De la Fuente-Sandoval, C., P7.27, 80 De Lisa, R., P2.1, 39 De Luca, A., O2.7, 15 De Luca, V., P2.23, 45; P3.34, 60; P5.25, 69; P8.15, 85; P8.26, 88; P14.28, 135; P14.56, 142 De Marco, L.A., P12.1, 113; P12.2, 113 de Mathis, M.A., P5.28, 70; P5.29, 71 De Mille, M.C., P9.17, 94 De Ronchi, D., P4.6, 62; P4.7, 62 de Silva, R., O3.3, 16 de Wilde, O., P3.30, 59 de Wit, H., P18.12, 160 De Zutter, S., O1.4, 12, P14.40, 137 Dean, J., P14.15, 131 Dean, L., P1.19, 36 Deckert, J., P3.7, 53; P5.14, 67; P5.26, 70; P5.27, 70; P5.34, 72; P18.12, 160 Degn, B., P2.10, 42 deJonge, S., P15.6, 145 Del Zompo, M., P2.1, 39 Delahanty, R., P9.26, 97 Del-Favero, J., P15.15, 147; P2.31, 47 DeLisi, L., S4.1, 7 Delisi, L.E., O8.6, 28; P7.16, 77; P8.23, 87; P10.14, 102; P14.16, 131; P14.18, 132; P14.55, 142 DeLong, G.R., O5.5, 21; P10.4, 99; P10.5, 100; P10.24, 105 Delorme, R., O5.6, 22; P5.30, 71 Deluca, V., P5.27, 70; P18.17, 161 Dementieva, Y., P10.23, 105 Dempster, E., P3.3, 52 Dempster, E.L., P3.9, 54; P15.8, 146

Denardin, D., P10.7, 100 Denton, M., P7.16, 77 DePaulo, J., P2.28, 47 DePaulo, J.R., O1.6, 12; O2.2, 13; O8.4, 27; O9.7, 30; P2.16, 44; P2.40, 50 Depienne, C., P10.21, 104 Derks, E.M., O6.4, 23 Dermaut, B., S1.1, 5 DeRosse, P., P14.22, 133 Desbiolles, C., P12.18, 117; P16.5, 150 Deschner, M., O8.5, 28 Deshpande, S., P14.57, 142 Destro, H.M., P5.20, 68; P13.23, 126 Detera-Wadleigh, S.D., O8.4, 27; P8.2, 82 Dettling, M., P18.2, 158 Deutsch, C.K., S7.2, 9 Devlin, B., O2.4, 14; O4.5, 19; P14.53, 141 Devriendt, K., O7.2, 25; P10.13, 102 Di Maria, E., P2.35, 48 Dick, D.M., P2.28, 47; P9.24, 96 Dietrich, I., P1.30, 39 Diez, J., P17.9, 155 Dillon, K., O2.3, 14 Dimitrova, A., O4.7, 20; O8.1, 26; P2.19, 44; P13.16, 125; P18.5, 158 Dina, C., P3.1, 51; P12.4, 113; P12.27, 119 Dmitrzak-Weglarz, M., P2.21, 45; P2.34, 48; P5.15, 67; P14.33, 136 Dmitrzak-Weglarz, M.D.W., P2.32, 48; P5.6, 65; P5.7, 65 Dobrusin, M., P8.19, 86 Doherty, F., P18.9, 159 Dolan, R.J., S6.3, 8 Domschke, K., P3.5, 53; P5.26, 70; P10.1, 99; P10.3, 99 Donald, J.A., P2.20, 45 Donnelly, S., P10.8, 101; P10.23, 105; P10.32, 107 Donohoe, G., P14.10, 130; P14.41, 138; P14.60, 143 Dourado, A., P2.2, 40 Doyle, A., O5.1, 20; P9.20, 95 Doyle, A.E., P10.6, 100 Dragan, W.L., P5.5, 65 D’Souza, U.M., P15.2, 144 Duan, J., O4.6, 19; P14.64, 143 Duan, S.W., P7.19, 78; P7.7, 75; P9.1, 90 Dube, M.-P., P7.16, 77 Ducci, F., P2.35, 48 Duckworth, J., O3.3, 16 Dudbridge, F., S5.3, 8; O5.8, 22 Duezel, E., P12.7, 114 Duffy, L., P1.1, 31 Duijff, S.N., P6.6, 73 Dunbar, D.R., O1.2, 11 Dunn, E., P1.18, 36 Dunn, M., P13.4, 121; P13.5, 121 Dunnigan., M, P1.2, 31 DuPree, M.G., S6.4A, 9 Durand, C., O5.6, 22 Dwyer, E., S6.1, 8 Dwyer, S., O4.7, 20; O8.1, 26 Dwyer, S.L., O4.2, 18 Dybro Lundorf, M., P2.10, 42 Dziedziejko, V., P4.9, 62 Easton, A., S6.1, 8 Ebihara, M., P14.24, 133 Ebisawa, T., P16.15, 152 Ebstein, R., P10.31, 107 Ebstein, R.P., O6.6, 24; P3.1, 51; P8.19, 86; P8.21, 86; P11.5, 109; P12.4, 113; P12.5, 114 Eccles, M.R., P15.5, 145 Echizenya, M., P15.13, 147 Eckermann, A., P1.6, 32

Author Index Edenberg, H., P2.40, 50 Edenberg, H.J., P9.24, 96 Edwards, L.J.R., P12.9, 115 Egan, M.F., S3.4, 7; P14.49, 140; P14.57, 142 Eichhammer, P., P18.26, 164 Eisenberg, J., P10.31, 107 Eisenmenger, W., P12.12, 115; P15.11, 146 Ekelund, J., P3.32, 60; P7.20, 78; P7.28, 80; P14.12, 130; P14.43, 138 Ekici, A., P3.17, 56 El Daoud, A., P2.3, 40 Elias-Billon, I., P17.16, 157 Elizur, Y., P12.5, 114 Ellingrod, V.L.E., P18.1, 157 Elliott, K.J., O3.2, 16 Emberton, J.E., P2.26, 46 ´ mond, C., P2.38, 49 E Engelen, J., O5.7, 22 English, J., P13.4, 121; P18.16, 161 Ennis, S., O5.2, 21 Enoch, M.A., P5.12, 66 Erens, W., O5.7, 22 Erhardt, A., P5.4, 64; P5.10, 66; P5.13, 66 Eriksson, E., O3.1, 15 Ervin, F.R., P12.18, 117; P16.5, 150 Escamilla, M., P7.27, 80 Escamilla, M.A., P3.21, 57; P8.12, 84; P14.13, 131; P14.45, 139 Eser, D., P1.5, 32; P1.27, 38; P18.23, 163 Esiri, M.M., P14.39, 137 Esparza, E., P3.21, 57; P14.13, 131 Esquerra, M., S1.3, 5 Estivill, X., P5.21, 68; P5.22, 69; P5.32, 71; P9.19, 95 Estrada, J.K., P17.5, 154 Estrada, L., P3.4, 52 Etain, B., P2.15, 43 Ettinger, U., P12.13, 116 Euskirchen, G., P15.18, 148 Evans, J., O3.3, 16 Evans, K.L., P2.14, 43; P13.12, 123 Evans, L., O2.8, 15 Evans, W., O3.3, 16 Ewald, H., P2.3, 40; P5.23, 69; P6.1, 72; P6.2, 72; P9.14, 93; P13.10, 123; P13.11, 123 Fagerness, H., O5.1, 20 Fagerness, J., O6.7, 24; P9.20, 95; P9.21, 95; P10.6, 100 Fagiolini, A., P1.23, 37 Fainman, J., P12.18, 117 Fallin, D., P4.14, 63 Fallin, D.M., P2.18, 44 Fallin, M.D., O3.6, 17; P2.24, 46; P3.11, 54; P14.50, 140 Fan, J.B., P8.13, 84 Fan, Y.S., P7.30, 81 Fan˜ana´s, L., P2.39, 49; P14.42, 138; P14.47, 139; P14.48, 140; P18.22, 163; P18.24, 163 Fang, Y., P14.62, 143 Fann, C.S.J., P14.5, 128; P14.19, 132 Fanous, A.H., P14.26, 134 Faraone, S.V., O5.1, 20; P1.16, 35; P8.27, 88; P9.20, 95; P10.6, 100; P10.31, 107; P13.3, 121; P14.5, 128; P14.11, 130 Farmer, A., P1.13, 34; P1.19, 36; P1.25, 37; P1.26, 38; P2.17, 44; P2.22, 45 Farmer, A.E., S8.2, 10 Farmer, I., P1.18, 36 Fathalli, F., P7.16, 77 Fatjo´-Vilas, M., P2.39, 49; P14.47, 139; P14.48, 140

Feinberg, A.P., O8.4, 27 Felder, B., P9.4, 91; P9.7, 91 Feng, G.Y., O4.3, 18; P14.15, 131; P7.10, 76; P7.11, 76; P7.3, 74; P7.9, 76; P9.1, 90; P7.19, 78; P7.7, 75 Feng, G., P7.6, 75; P14.62, 143 Feng, J., P12.24, 119; P17.13, 156 Feng, Y., P9.12, 93; P9.17, 94; P9.23, 96 Fernandes, C., O9.2, 29 Ferna´ndez-Aranda, F., P5.21, 68; P5.22, 69 Fernø, J., P15.16, 148; P15.17, 148 Ferrell, R.E., O2.4, 14 Ferrero, F., P1.13, 34; P1.26, 38 Fertuzinhos, S., P18.11, 160 Filippich, C.F., P14.23, 133; P14.25, 134 Fimmers, R., P5.14, 67; P5.26, 70 Finckh, U., O7.6, 25; P11.16, 111 Finn, C.T., O2.3, 14; P14.22, 133 Finucane, C., P3.31, 60 Firth, H., O5.6, 22 Fischer, A., P9.11, 93 Fitzgerald, M., O5.2, 21; O5.8, 22; P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99; P10.12, 102 Flake, L., P3.16, 55 Flibotte, S., O9.3, 29 Flint, T.J., P2.3, 40; P9.14, 93; P13.11, 123 Flomen, R.H., P13.24, 126; P14.7, 129 Foldager, L., P13.10, 123 Fontana, V., P1.11, 34; P1.12, 34; P18.10, 160 Foroud, T., O1.1, 11; P2.28, 47; P2.40, 50; P9.24, 96 Fortier, M.-E., P16.8, 150 Forty, L., P1.19, 36; P2.22, 45; P3.25, 58 Fournier, A., P2.38, 49 Fowler, T.A., P9.30, 98 Fox, L., P9.24, 96 Fox, N., P10.16, 103 Francks, C., O2.5, 14 Frangou, S., P3.8, 53 Frank, E., P1.23, 37 Frank, G., P14.17, 132 Franke, P., P5.14, 67; P5.26, 70; P5.34, 72 Fraser, C., P2.6, 41 Freedman, R., P12.13, 116 Freitag, C., P5.14, 67 Freitag, C.M., P5.26, 70 Fresa´n, A., P7.2, 74; P7.27, 80; P17.5, 154 Freudenberg, J., P14.58, 142 Frey, J.U., P12.7, 114 Frigerio, A., P12.29, 120 Frisch, A., P5.1, 64 Fritze, J., P5.14, 67; P5.26, 70; P5.34, 72 Froguel, P., P12.27, 119 Fryns, J.P., O7.2, 25 Fuchs, K., P13.7, 122 Fujimaru, K., P3.14, 55 Fukasawa, M., P14.24, 133 Fung, P., O3.3, 16 Funke, B., O2.3, 14; P14.22, 133 Funke, H., P1.3, 32 Fu¨rst, R., P2.46, 51 Furtado, E.F., P11.10, 110 Gabel, L.B., P8.27, 88; P14.11, 130 Gabriels, R., P10.8, 101 Gacia, M., P4.9, 62 Gadoros, J., P9.10, 92; P12.28, 119 Gagne´, B., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Gaidarova, R., P2.14, 43 Gak, E., P5.1, 64 Galili-Weisstub, E., P5.17, 68 Galindo, L., P4.11, 63 Galke, B., P14.53, 141

169

Gallagher, L., O5.2, 21 Gama, C.S., P7.21, 78 Gamliel, I., P9.3, 90 Gandri, Y., P5.17, 68 Gangadhar, B.N., P14.61, 143 Gao, J.J., P9.1, 90 Gao, S.-G., P1.4, 32 Gao, X.C., P9.1, 90 Garbarz, J.J., P8.12, 84 Garcia Marı´a, M.G., P7.2, 74 Garcia-Portilla, M.P., P5.8, 65 Gardner, M., P17.15, 157 Garittsen, H., P5.14, 67; P5.26, 70 Gasche, C., P10.11, 101 Gasto´, C., P2.39, 49; P18.22, 163; P18.24, 163 Gaszner, P., P11.2, 108 Gaucher, M., O2.8, 15; P2.30, 47; P18.20, 162 Gaucher, M.A., P2.27, 46 Gejman, P.V., O4.6, 19; P14.64, 143 Gelegen, E.C., P5.9, 65 Gennarelli, M., P2.29, 47; P2.35, 48; P18.19, 162 Gentil, V., P12.23, 118; P12.25, 119 Gentile, K.L., O4.1, 18 George, C., P9.18, 94; P10.16, 103 George, T., P18.28, 164 Georgieva, L., P2.19, 44; P18.5, 158 Georgieva, L.N.G., P13.16, 125 Geronimo, J., P5.24, 69 Gershon, E.S., O8.4, 27; P2.28, 47; P2.40, 50 Gerstein, M., P15.18, 148 Gervai, J., P3.33, 60; P10.30, 107; P12.28, 119 Geyer, M.A., O7.8, 26; P15.14, 147 Ghosh, S., P13.2, 121; P17.7, 155 Gianti, L., P16.10, 151 Giegling, I., O7.6, 25; P1.30, 39; P8.9, 83; P8.11, 84; P14.6, 129 Gilbert, J., P10.8, 101; P10.15, 103 Gilbert, J.R., O5.5, 21; P10.4, 99; P10.5, 100; P10.23, 105; P10.24, 105 Gill, M., O1.5, 12; O4.2, 18; O5.2, 21; O5.8, 22; O9.6, 30; P1.13, 34; P1.26, 38; P2.44, 51; P2.45, 51; P3.5, 53; P3.19, 56; P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99; P10.12, 102; P10.31, 107; P13.26, 127; P14.10, 130; P14.41, 138; P14.57, 142; P14.60, 143 Gillberg, C., O5.6, 22 Gillberg, I.C., O5.6, 22 Giorda, R., P12.29, 120 Giouzeli, M., P7.15, 77 Giovanucci-Uzielli, M.L., P16.10, 151 Giri, K.P., P3.15, 55; P14.23, 133 Girimaji, S.C., P10.20, 104 Giros, B., P2.15, 43; P14.32, 136 Gizer, I.R., P10.9, 101; P10.18, 103 Gladis, M., O1.6, 12 Glambek, M., P15.16, 148; P15.17, 148 Glanzmann, C., P12.24, 119 Glaser, B., P2.12, 42; P8.10, 84 Glatt, S.J., P1.16, 35; P8.27, 88; P13.3, 121; P14.11, 130 Go, R.C.P., O3.6, 17 Goate, A., S1.3, 5 Godlewska, B., P8.24, 87; P14.33, 136 Goetsch-Martone, R., P15.18, 148 Gohlke, H., P9.4, 91 Goldberg, T.E., S3.4, 7 Goldberger, C.G., S7.3, 10; P7.1, 74; P7.18, 78 Golden, G.T., P11.18, 112

170

Author Index

Goldman, B., P5.1, 64 Goldman, D., O7.3, 25; P5.24, 69 Goldstein, D.B., S2.4, 6; P5.3, 64 Golimbet, V.E., P7.23, 79; P8.29, 89 Golimbet, V.G., P8.21, 86 Gomez Ariadna, A.G., P7.2, 74 Gonnermann, C., P14.6, 129 Gonzalez, J.R., P3.21, 57 Gonza´lez, J.C., P13.19, 125; P13.20, 125; P13.27, 127 Gonza´lez, N., P18.24, 163 Gonzalez-Neira, A., P17.15, 157 Gonza´lez-Sobrino, B.Z., P17.10, 155 Goodman, A.B., P7.21, 78 Goody, A., P12.16, 116 Gordon-Smith, K., O8.2, 27; P1.21, 36; P2.6, 41; P2.11, 42; P2.22, 45 Gordon-Smith, K.M., P3.25, 58 Gornick, M.C., P10.27, 106 Goubran-Botros, H., O5.6, 22 Gourion, D.G., S7.3, 10; P7.1, 74; P7.18, 78 Grados, M., O6.8, 24 Graff, A., P5.2, 64 Grassie, M.A., O1.2, 11 Grataco`s, M., P5.21, 68; P5.22, 69; P5.32, 71; P9.19, 95; P17.14, 156 Green, A., O5.2, 21 Green, E., O1.5, 12; O4.7, 20; O5.8, 22; O8.2, 27; P, P1.18, 36; P2.44, 51; P2.45, 51 Green, E.K., O8.1, 26; P2.6, 41; P2.11, 42; P2.33, 48 Greening, J., P2.22, 45 Greenwood, C., O9.7, 30 Greenwood, T.A., P2.28, 47; P2.30, 47 Grevet, E.H., P9.11, 93 Grice, D.E., P11.18, 112 Grimm, T., P10.25, 106 Gritsenko, I., O6.6, 24; P3.1, 51; P11.5, 109; P12.4, 113 Gritsenko, I.K., P8.21, 86 Grizenko, N., P9.8, 92 Groome, N.P., P14.39, 137 Gross, M., O8.5, 28; P13.15, 124 Grozeva, D., O4.7, 20; O8.1, 26; O8.2, 27; P2.6, 41; P13.16, 125 Grozeva, D.V., P1.18, 36; P2.19, 44 Grzywacz, A., P11.3, 108; P11.4, 108; P14.33, 136 Gu, N.F., P7.5, 75; P7.6, 75; P7.7, 75; P7.9, 76; P7.10, 76; P7.11, 76; P7.19, 78; P14.15, 131; P14.62, 143 Gu, Z.Z., O4.3, 18; P7.5, 75 Guan, Y.S., P7.11, 76 Guderian, S., P12.7, 114 Guerini, F.R., P17.11, 156 Guerra, A., P14.48, 140 Guerreiro, J.R., P14.3, 128 Guerrini, I., O3.8, 17 Guimara˜es, A.P.M., P10.7, 100 Guindalini, C., P11.11, 110; P11.15, 111 Gulli, R., P2.35, 48 Gundelfinger, E.D., P12.7, 114 Guo, T.W., P9.1, 90 Guoyin, F., P4.3, 61 Gur, R.E., O3.7, 17; P3.20, 57 Gur, R.C., O3.7, 17 Gurling, H., P2.10, 42; P2.15, 43 Gurling, H.M.D., O3.8, 17; O7.7, 26; P1.14, 35; P2.36, 49; P14.8, 129; P15.10, 146; P8.18, 86; P16.13, 152 Gustafson, D., O3.1, 15 Gutierrez, B., P14.42, 138 Gutie´rrez, A., P13.1, 120 Gutie´rrez, B., P2.39, 49; P18.22, 163; P18.24, 163

Gutie´rrez-Zotes, A., P3.24, 58 Gutknecht, L., P5.26, 70 Haddad, S.A., P9.21, 95 Hager, J., O9.1, 29 Haines, J.L., P9.26, 97; P10.5, 100 Hajak, G., P18.26, 164 Hall, M., P12.13, 116 Hallmayer, J., P7.31, 81; P14.14, 131 Hamer, D.H., S6.4A, 9 Hampson, M., P6.1, 72 Hamshere, M., O1.5, 12; O8.7, 28; O9.6, 30; P2.45, 51 Hamshere, M.L., P2.44, 51; P3.19, 56; P13.26, 127 Handoko, H.Y., P14.23, 133; P14.25, 134 Hanna, G.L., P5.24, 69 Hannah, D.E., P14.23, 133; P14.25, 134 Hansen, T., P14.4, 128 Hanson, M., O3.3, 16 Happe, F., O6.3, 23 Haque, M.S., P3.25, 58 Haque, S., P2.22, 45 Harano, M., P11.7, 109 Hardy, J., E1.3, 4; S1.4, 5 Hardy, J.A., O3.3, 16 Hariri, A.R., S3.4, 7 Harold, D., P9.16, 94 Harold, G.T., P12.16, 116 Haroutunian, V., P7.29, 81 Harrington, H.-L., O2.6, 15 Harrington, K.M., P10.9, 101 Harris, C.R., P5.12, 66 Harrison, P.J., P8.25, 88 Hart, D.J., P4.10, 62 Hartmann, A.M., P1.30, 39; P8.9, 83 Hartwig, C., O7.6, 25 Harvey, C., P10.19, 104 Harvey, M., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Harvey, P., O8.6, 28 Hashida, A., P3.14, 55; P14.34, 136 Hashimotodani, Y., P16.15, 152 Hattori, E., P8.2, 82; P14.24, 133 Hauser, J., O8.5, 28; P2.21, 45; P2.34, 48; P3.6, 53; P3.27, 59; P5.6, 65; P5.7, 65; P5.15, 67; P14.33, 136; P14.40, 137 Hauser, J.H., P2.32, 48 Hawi, Z., O5.8, 22; P3.5, 53; P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99; P10.12, 102 Hay, D.F., P12.16, 116 Hayashi, M.A.F., P14.3, 128 Hayward, N.K., P14.23, 133; P14.25, 134 Hazeel, C., P4.16, 63 Hazelwood, L., P1.22, 37 Hazelwood, L.A., P18.18, 162 He, G., O4.3, 18; P7.3, 74; P7.5, 75; P7.6, 75; P7.10, 76 He, L., O4.3, 18; P, P14.15, 131; P7.3, 74; P7.5, 75; P7.6, 75; P7.7, 75; P7.9, 76; P7.10, 76; P7.11, 76; P7.19, 78; P9.1, 90 Heath, A.C., P3.16, 55 Heidmann, V., O8.5, 28 Heineman-de Boer, J.A., P6.6, 73 Heinze, H.J., P12.7, 114 Hell, D., P13.9, 123 Henn, F.A., O8.5, 28 Hennah, W., P3.32, 60; P7.20, 78; P14.12, 130; P14.31, 135; P14.43, 138 Hennessy, R.J., S7.4, 10 Hennis, L., P10.15, 103 Henry, C., P2.15, 43 Hermkajakob, H., P8.17, 85

Hernandez, M., P8.12, 84 Hernandez Sandra, S.H., P7.2, 74 Heron, H., O8.2, 27 Heron, J., O1.5, 12; P2.6, 41; P2.11, 42; P2.22, 45; P2.44, 51; P2.45, 51 Heron, J.C., P3.25, 58 He´ron, D., P10.21, 104 Herrera, C., P14.13, 131 Herva`s, A., P9.19, 95 Hesselbrock, V., P9.24, 96 Hettema, J.M., P5.11, 66 Heutink, P., P12.32, 120 Hewitson, K.S., P14.39, 137 Hiekkalinna, T., P7.20, 78 Hill, G., P9.13, 93; P9.16, 94 Hiltunen, M., O3.2, 16 Hirayama, R., P3.14, 55; P15.12, 147 Hobermann, J., P5.12, 66 Hochstenbach, R., P6.8, 74 Hoefels, S., P1.6, 32; P13.15, 124 Hoefgen, B., P14.14, 131 Hoehe, M.R., P7.31, 81 Hoehn-Saric, R., O6.8, 24 Hoenicka, J., P17.9, 155 Hohoff, C., P3.7, 53; P5.14, 67; P18.12, 160 Holgado, P., P17.9, 155 Holliday, J., P5.31, 71 Holmans, P., S5.2, 7; O1.6, 12; O8.7, 28; O9.4, 29; P2.44, 51; P3.19, 56; P13.26, 127 Holmans, P.A., P9.15, 94 Holmes, S.E., P14.18, 132 Holsboer, F., O1.8, 13; O8.3, 27; P1.28, 38; P5.4, 64; P5.10, 66; P5.13, 66; P5.16, 67; P13.8, 122 Holt, R.A., O9.3, 29 Holzman, T., O5.4, 21 Honer, W.G., O9.3, 29 Hong, K.S., P8.1, 81 Hong, L.E., P7.12, 76 Hong, L.E.H., P14.30, 135 Hoogendijk, W.J., P12.32, 120 Hoogendoorn, B., P2.19, 44 Hoogendoorn, M.L.C., P3.22, 57; P6.7, 73; P14.29, 135 Hori, H., P8.26, 88 Hori, T., P16.15, 152 Hornik, K., P10.11, 101; P13.7, 122 Horodnicki, J., P11.3, 108; P11.4, 108 Horstmann, S., P1.28, 38; P5.16, 67 Hosakare, M., P3.16, 55 Hoshika, A., P8.3, 82 Hossain, S.M., O7.1, 24 Hou, C., O4.6, 19; P14.64, 143 Hough, A., P1.19, 36 Hounie, A.G., P5.28, 70; P5.29, 71 Hranilovic, D., P16.2, 149 Hrano, M., P11.13, 111 Hu, X., P5.24, 69; P5.31, 71; P14.36, 137; P14.37, 137 Huang, S.Y., P11.20, 112 Hubank, M., O7.7, 26; P15.10, 146 Hudson, L., P13.4, 121 Hudziak, J.J., O6.4, 23 Hui, T.C.K., P7.13, 77 Humphries, T., P9.12, 93; P9.23, 96 Hunt, S.P., O7.7, 26; P15.10, 146 Hussman, J.P., O5.5, 21 Hutz, M., P7.21, 78; P10.7, 100 Hwang, R., P3.34, 60; P14.28, 135 Hwang, R.W., P18.17, 161 Hwu, H.G., P14.5, 128; P14.19, 132 Hyde, S., P2.6, 41; P2.11, 42 Hyde, S.A., P2.22, 45; P3.25, 58 Ickowicz, A., P9.22, 96; P9.23, 96; P9.28, 97

Author Index Igo, R.P., O5.4, 21 Iida, N., P11.7, 109 Ikeda, M., P7.25, 80; P7.26, 80 Ikenaga, E., P9.6, 91; P18.14, 161 Illa, J.M., P14.48, 140 Illes, F., O3.4, 16 Illig, T., O8.3, 27; O8.5, 28; P5.34, 72; P9.4, 91; P14.58, 142 Imamura, A., P3.14, 55; P14.34, 136; P15.12, 147 Imamura, T., P7.4, 75 Inada, T., P11.7, 109; P11.13, 111 Inoue, Y., P16.15, 152 Insacco, C., P18.21, 163 Iourov, I.Y., P6.4, 73; P9.2, 90; P16.9, 151; P16.10, 151 Iribarren-Iriso, F., P18.27, 164 Ishibashi, M., P11.7, 109 Ising, M., O1.8, 13; P1.28, 38; P5.4, 64; P5.13, 66; P5.16, 67 Isles, A.R., P16.1, 149 Isojima, Y., P16.15, 152 Itokawa, M., P14.24, 133 Ivanov, D., P9.16, 94; P17.3, 154 Ivanov, D.K., P17.4, 154 Ivanov, D.K.I., P13.16, 125 Iwamoto, K., P13.18, 125 Iwata, N., P7.25, 80; P7.26, 80; P11.7, 109; P11.13, 111 Iwayama-Shigeno, Y., P8.2, 82; P8.3, 82; P14.24, 133 Iyo, M., P11.7, 109; P11.13, 111 Jablensky, A., P14.21, 133 Jablonsk, M., P4.11, 63 Jabs, B., P2.41, 50 Jacob, C., P5.34, 72 Jacobsen, K.D., P18.13, 160 Jaghori, B., P9.17, 94 Jahuar, P., O3.8, 17 Jain, S., P13.2, 121; P15.19, 149; P17.6, 154 Jain, U., P9.25, 96 Jamain, S., O5.6, 22 James, M.R., P14.23, 133; P14.25, 134 James, R., S3.1, 6 Janakiev, N.J., P13.16, 125 Janakiramaiah, N., P14.61, 143 Jang, Y.L., P8.1, 81 Jann, H.Y., P14.5, 128; P14.19, 132 Jansen, P., P6.6, 73 Jardim, D., P13.22, 126 Jariod, M., P3.4, 52; P3.24, 58; P13.1, 120 Jaworski, J., P10.24, 105 Jaworski, J.M., P10.5, 100; P10.23, 105 Jazin, E., O4.4, 19 Jeffries, A.R., P6.3, 73 Jehu, L., P4.5, 61 Jerez, A., P7.27, 80 Jernej, B., P16.2, 149 Jeun, H.O., P8.1, 81 Jhoo, J.H., P4.4, 61 Jiandong, Y., P4.3, 61 Jiang, X., P5.12, 66 Jime´nez, L., P5.21, 68; P5.22, 69 Jime´nez-Arriero, M.A., P17.9, 155 Jinno, Y., P14.34, 136 Jollant, F., P1.17, 35; P1.31, 39; P1.32, 39 Jones, G., O8.7, 28 Jones, I., O1.5, 12; O8.1, 26; O8.2, 27; O9.6, 30; P1.13, 34; P1.18, 36; P1.19, 36; P2.6, 41; P3.19, 56; P2.11, 42; P2.33, 48; P2.44, 51; P2.45, 51; P12.24, 119; P13.26, 127 Jones, I.R., P1.26, 38; P2.22, 45; P3.25, 58 Jones, L., O1.5, 12; O8.1, 26; O8.2, 27; O9.6, 30; P1.13, 34; P1.18, 36;

P2.6, 41; P2.11, 42; P2.33, 48; P2.44, 51; P2.45, 51 Jones, L.A., O8.7, 28; P1.19, 36; P1.21, 36; P1.26, 38; P2.22, 45; P3.25, 58; P13.26, 127 Jo¨nsson, E.G., P14.40, 137 Joo, Y.H., P3.20, 57 Joober, R., P7.16, 77; P9.8, 92; P16.8, 150 Jorgensen, T.H., P2.3, 40 Jørgensen, H.A., P15.17, 148 Joyce, P., P18.3, 158; P18.4, 158 Joyce, P.R., P15.5, 145 Junker, R., P1.3, 32 Kachalin, E.U., P8.29, 89 Kaczmarkiewicz-Fass, M.K.F., P2.32, 48 Kaganovich, M., P8.19, 86; P8.20, 86; P13.17, 125 Kahn, R.S., P3.22, 57; P6.7, 73; P14.29, 135 Kajimura, N., P16.15, 152 Kakiuchi, C., P2.8, 41; P2.37, 49 Kaleda, V.G., P8.21, 86 Kalsi, G., O7.7, 26; P1.14, 35; P2.10, 42; P2.36, 49; P8.18, 86; P14.8, 129 Kamei, Y., P16.15, 152 Kamiya, A., S3.2, 6; O1.3, 11 Kane, J.M., P14.22, 133 Kaneva, R., P1.15, 35; P2.14, 43 Kanyas, K., P13.6, 122 Kapelski, P., P2.21, 45; P3.27, 59; P14.33, 136; P14.40, 137 Kaprio, J., P12.20, 117 Karmiloff_Smith, A., P12.19, 117; P16.12, 152 Karni, O., P13.6, 122 Karwautz, A., P5.31, 71 Kas, M.J., P5.9, 65 Kasahara, M., P9.5, 91 Kasper, S., P13.7, 122 Kassem, L., O8.5, 28 Katai-Fodor, I., P11.2, 108 Katay, L., P9.29, 98 Kato, N., P2.9, 41 Kato, T., P2.37, 49; P2.8, 41; P2.9, 41; P13.18, 125 Katoh, M., P16.15, 152 Katoh, T., P15.13, 147 Katsu, T., P7.4, 75 Katz, I., P8.12, 84 Katz, R., P1.7, 33 Kawahara, R., O1.3, 11 Kazuno, A., P2.9, 41 Kealey, C., P2.7, 41 Keck, M.E., P5.4, 64; P5.10, 66; P5.13, 66; P5.16, 67 Keck, P.E., P2.30, 47 Kelsoe, J., O2.8, 15; P2.40, 50; P18.20, 162 Kelsoe, J.R., P2.26, 46; P2.27, 46; P2.28, 47; P2.30, 47; P2.43, 50 Kendler, K.S., P14.17, 132; P14.57, 142; P5.11, 66; S8.3, 10; O6.1, 22; P11.19, 112; P14.26, 134; P14.27, 134; P14.54, 141 Kennedy, J., P10.16, 103; P17.2, 153; P14.46, 139 Kennedy, J.L., P2.23, 45; P3.34, 60; P5.24, 69; P5.25, 69; P5.27, 70; P8.14, 85; P8.15, 85; P8.26, 88; P9.18, 94; P9.22, 96; P9.23, 96; P9.25, 96; P9.28, 97; P9.29, 98; P12.25, 119; P14.28, 135; P14.56, 142; P18.17, 161 Kennedy, M.A., P15.5, 145; P18.3, 158; P18.4, 158 Kent, L., O5.8, 22 Kere, J., P10.25, 106

171

Kereszturi, E., P7.14, 77; P9.10, 92 Kern, N., O1.8, 13; P5.4, 64; P5.10, 66; P5.13, 66; P5.16, 67 Kerr, N.P., P2.20, 45 Kerrien, S., P8.17, 85 Kerwin, R., P14.7, 129; P18.15, 161; P18.29, 165 Kerwin, R.W., P7.24, 79; P8.6, 83; P18.6, 159; P18.8, 159; P18.16, 161; P18.27, 164 Keshavan, M., P3.20, 57 Kestila¨, M., P7.20, 78; P14.31, 135 Khanna, S., P5.18, 68; P5.19, 68 Kidd, J.R., P9.17, 94 Kidd, K.K., P9.17, 94 Kikuchi, T., P15.12, 147 Kilpikari, R., P14.31, 135 Kim, C., P7.17, 78 Kim, J.J., O4.5, 19 Kim, J.W., P8.1, 81; P9.20, 95 Kim, K.W., P4.4, 61 Kimmel, R.J., P2.43, 50 King, N., P2.23, 45; P3.34, 60; P5.25, 69; P5.27, 70; P9.18, 94; P9.29, 98; P14.28, 135 King, N., P8.14, 85; P10.16, 103 Kinsella, A., S7.4, 10 Kiraly, O., P9.10, 92; P12.3, 113 Kiran, H.G.K., P13.2, 121 Kirby, A., O5.1, 20 Kirby, A.N., O2.1, 13; O4.1, 18 Kirillova, G.P., O2.4, 14 Kirley, A., O5.8, 22; P3.5, 53; P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99; P10.12, 102 Kirov, G., O4.2, 18; O4.7, 20; O8.1, 26; O8.2, 27; O9.6, 30; P1.18, 36; P2.6, 41; P2.11, 42; P2.12, 42; P2.19, 44; P2.33, 48; P2.44, 51; P8.10, 84; P13.26, 127; P14.57, 142; P17.3, 154; P17.4, 154; P18.5, 158 Kirov, G.K., P13.16, 125 Kishimoto, M., P12.14, 116 Kiss, E., P9.29, 98 Kitajima, T., P16.15, 152 Kitano, T., O4.6, 19 Klauck, S.M., P9.4, 91; P9.7, 91 Klei, L., P14.53, 141 Kleinman, J.E., P14.49, 140 Klenova, E., O7.4, 25 Klopp, N., O8.3, 27; O8.5, 28; P5.34, 72; P14.58, 142 Knapp, M., P2.41, 50; P7.31, 81 Knight, J., S5.1, 7; P10.31, 107; P12.8, 114; P12.10, 115; P12.11, 115; P12.23, 118 Knorr, C.C., P5.4, 64; P5.10, 66 Knowles, J.A., O1.6, 12 Ko, H.-C., P11.20, 112 Koch, A.A., P10.8, 101 Kockum, I., O6.2, 23 Koeleman, B.P.C., P10.10, 101 Koenig, I.R., P10.25, 106 Koh, J., P7.17, 78 Kohn, Y., P13.6, 122 Kohu, K., P8.16, 85 Kolachana, B.S., P14.49, 140 Koller, G., P11.6, 109; P11.12, 110 Kollins, S., P10.15, 103 Komiyama, T., P11.7, 109; P11.13, 111 Kopaneva, N.A., P7.23, 79 Kopelman, M., O3.8, 17 Koronyo-Hamaoui, M., P5.1, 64 Korostishevsky, M., P8.19, 86; P8.20, 86; P13.17, 125 Korszun, A., P1.13, 34; P1.25, 37 Kosmowska, M., P2.34, 48

172

Author Index

Kostov, C., P1.15, 35; P2.14, 43 Koszycki, D., P5.25, 69; P5.27, 70 Kovacs, I., P2.43, 50 Kovacs, M., P9.18, 94; P9.29, 98; P10.16, 103 Kovalenko, S., O1.4, 12; O8.5, 28; P13.15, 124; P14.40, 137 Kramer, J., P9.24, 96 Krastev, S., P1.15, 35; P2.14, 43 Krebs, M.O.K., S7.3, 10; P7.1, 74; P7.18, 78 Kremensky, I., P1.15, 35; P2.14, 43 Kremer, I., P8.19, 86; P11.5, 109 Krieger, J.E., P13.22, 126 Krinsky, M., P9.25, 96 Kruse, T., P2.10, 42 Kruse, T.A., P2.3, 40; P5.23, 69; P9.14, 93; P13.10, 123; P13.11, 123 Kubendran, S., P13.2, 121; P17.6, 154 Kucharska-Mazur, J., P11.3, 108; P11.4, 108 Kucherlapati, R., O2.3, 14; P14.22, 133 Kuczenski, R., O7.8, 26; P15.14, 147 Kuhlenbaeumer, G., P5.14, 67; P5.26, 70 Kuhn, J.W., S8.3, 10 Kumar, R.A., P17.8, 155 Kumar-Singh, S., S1.1, 5 Kuntsi, J., P3.13, 55 Kunugi, H., P2.8, 41 Kuperman, S., P9.24, 96 Kupfer, D.J., P1.23, 37 Kurlan, R., P9.17, 94 Kuroda, S., P7.4, 75; P11.13, 111; P12.14, 116 Kusanagi, H., P15.13, 147 Kusevic, I., P4.14, 63 Kusters, K., P10.10, 101 Kwok, M.C.H., O7.1, 24 Labad, A., P3.4, 52; P3.24, 58; P13.1, 120 Labbe´, M., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Lachman, H., P8.31, 89; P8.32, 89 Laegsgaard, M.M., P4.1, 60 Lageix, P., P9.8, 92 Laguerre, A., P3.10, 54 Laird, N.M., O6.7, 24 Lakatos, K., P3.33, 60; P10.30, 107 Lake, S., P14.22, 133 Lam, D., P2.17, 44 Lambert, D., O1.5, 12; P2.45, 51; P13.26, 127 Lambert, D.B.J., P2.44, 51; P3.19, 56 Lan, T., P14.38, 137 Landau, S., P14.44, 139 Lander, E.S., S6.3, 8 Landoni, S., P1.9, 33; P4.6, 62; P4.7, 62 Landowski, J., P8.24, 87; P14.33, 136 Lane, A., S7.4, 10 Langley, K., P9.15, 94; P9.30, 98 Lanktree, M., P2.23, 45; P9.25, 96; P14.28, 135; P17.2, 153 Laranjeira, R., P11.11, 110 Larkin, C., S7.4, 10 Lasseter, V.K., P2.18, 44; P2.24, 46; P3.11, 54; P14.50, 140 Lathrop, M., P11.1, 108 Lattuada, E., P18.7, 159 Laucht, M., P11.10, 110 Lauer, C.J., P12.7, 114 Lauridsen, M.B., P5.23, 69 Laurin, N., P9.22, 96 Lauritsen, M.B., P9.14, 93 Lawrence, K., S6.3, 8 Lawrence, J., O7.7, 26; P1.14, 35; P2.10, 42; P2.36, 49; P8.18, 86; P14.8, 129 Le Hellard, S., P13.12, 123

Leach, R., P8.12, 84 Leach, S., P17.8, 155 Leboyer, M., O5.6, 22; P1.17, 35; P1.31, 39; P1.32, 39; P2.15, 43; P3.10, 54; P3.26, 58; P5.30, 71; P10.21, 104; P14.32, 136 Lee, A.J., P13.12, 123 Lee, D.Y., P4.4, 61 Lee, Y.J., P5.14, 67 Lee, Y.S., P8.1, 81 Lees, A., O3.3, 16 Lefebvre, R., P15.15, 147 LeHellard, S., P6.1, 72 Leisch, F., P10.11, 101; P13.7, 122 Leister, F., P14.18, 132 Lencer, R., P3.7, 53 Lendon, C., P2.11, 42 Lendon, C.L., P4.16, 63 Leonard, J., O8.6, 28 Leor, S., P5.1, 64 Lerer, B., P7.31, 81; P13.6, 122; P14.14, 131 Lesch, K.P., P5.26, 70 Leszczynska-Rodziewicz, A., O8.5, 28; P2.21, 45; P2.34, 48; P3.6, 53; P14.33, 136 Leszczynska-Rodziewicz, A.L.R., P2.32, 48 Leung, Y.J., P14.20, 133 Levitt, P., P14.57, 142 Levinson, D., P8.12, 84 Levinson, D.F., O1.6, 12; O4.6, 19 Lewis, D.A., P14.57, 142 Li, Bian, P4.3, 61 Li, C., P7.7, 75 Li, Dawei, P7.5, 75 Li, K.Q., O4.3, 18 Li, T., P5.31, 71; P7.13, 77; P8.4, 82; P14.15, 131; P14.36, 137; P14.37, 137; P14.57, 142 Li, Y.C., P7.3, 74 Lian, J., P15.18, 148 Lian, Z., P15.18, 148 Liang, K.Y., O6.8, 24; P2.18, 44; P2.24, 46; P3.11, 54; P14.50, 140 Liang, S.G., P2.26, 46; P2.27, 46 Lichter, P., P9.7, 91 Lieb, M., P11.12, 110 Lieberman, J., P18.17, 161 Lieberman, J.A., P3.34, 60 Lifton, R.P., P13.6, 122 Liljenberg, J., O6.2, 23 Lim, L.C.C., P7.8, 75 Limon, J., P8.24, 87 Lin, C.Y., P14.5, 128; P14.19, 132 Lin, He, P4.3, 61 Lin, W.W., P11.20, 112 Lindgren, C., P10.25, 106 Lindholm, E., O4.4, 19 Linszen, D., P3.30, 59 Lipsky, R., O7.3, 25 Lipsky, R.H., P5.12, 66 Liu, C.M., P14.5, 128; P14.19, 132 Liu, H.J., P7.11, 76 Liu, J., O4.3, 18 Liu, L., O9.2, 29 Liu, W., P8.12, 84 Liu, X., P7.6, 75; P14.36, 137; P14.37, 137 Liu, X.M., O4.3, 18; P7.10, 76 Liu, Y., P14.38, 137 Liu, Y.L., P14.19, 132; P14.5, 128 Liu, Z.J., P7.9, 76 Lizano, A., P14.13, 131 Loat, C.S., S6.4B, 9; P12.21, 118; P12.22, 118; P13.25, 126 Lobato, M.I., P7.21, 78 Lobo, D.S.S., P12.23, 118; P12.25, 119 Lobos, E.A., O6.5, 23

Lohmussaar, E., O1.8, 13 Lohr, J.B., O7.8, 26; P15.14, 147 Lo¨nnqvist, J., P3.32, 60; P7.20, 78; P7.28, 80; P14.12, 130; P14.31, 135; P14.43, 138 Lopes, A., P15.9, 146 Lopez-Ilundain, J.M., P7.24, 79; P8.6, 83; P8.7, 83; P13.13, 124; P13.14, 124; P18.16, 161 Lorenzi, C., P4.6, 62; P4.7, 62; P12.17, 117; P18.7, 159; P18.22, 163 Loukola, A., P7.28, 80; P14.31, 135; P14.43, 138 Lovett, M., P9.23, 96 Løvlie, R., P15.16, 148; P15.17, 148 Low, N., O5.8, 22; P12.31, 120 Low, N.C.P., P1.29, 39 Lowe, N., P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99; P10.12, 102 Loyzaga, C., P5.2, 64 Lu, R.B., P11.20, 112 Luca, P., P9.28, 97; P9.29, 98 Lucae, S., O1.8, 13; P1.28, 38; P5.13, 66; P5.16, 67 Lunddorf, M.D., P13.10, 123; P16.13, 152; P16.14, 152 Ly, L., P14.7, 129 Lyachenko, G.L., P7.23, 79 Lyon, L., P1.18, 36; P2.22, 45; P3.25, 58 Lyons, M.J., P14.20, 133 Ma, D.Q., O5.5, 21; P10.4, 99 Macciardi, F., P2.23, 45; P5.25, 69; P5.27, 70; P9.25, 96; P14.28, 135; P17.2, 153; P17.16, 157 Macedo, A., P2.2, 40 Macgregor, S., O8.2, 27; O9.6, 30; P2.6, 41; P2.11, 42 Machado, M., P12.1, 113; P12.2, 113 MacIntyre, D., P6.5, 73 Mackie, S., S3.1, 6 MacKinnon, D., O1.6, 12 MacKinnon, D.F., P2.40, 50 Maclean, A., P1.2, 31 Maeda, H., P11.7, 109 Maeda, K., O1.3, 11 Maegawa, S., O2.5, 14 Maher, B.S., O2.4, 14 Maier, W., O1.4, 12; O2.2, 13; O3.4, 16; O8.3, 27; O8.5, 28; P1.6, 32; P1.13, 34; P1.26, 38; P2.46, 51; P7.31, 81; P8.19, 86; P14.14, 131; P14.40, 137; P14.58, 142 Makoff, A.J., P13.24, 126; P14.7, 129 Malafosse, A., P1.17, 35; P1.31, 39; P1.32, 39; P2.15, 43 Malhotra, A.K., E3.2, 4; O2.3, 14; O7.3, 25; P14.22, 133 Malhotra, T., P10.20, 104 Malloy, M.P., P6.1, 72; P6.2, 72; P6.5, 73 Malloy, P., S3.1, 6 Malone, M., P9.22, 96; P9.23, 96; P9.28, 97 Mancama, D., P18.15, 161; P18.29, 165 Mandelli, L., P1.8, 33; P1.9, 33; P1.10, 33; P12.17, 117; P18.21, 163 Mandich, P., P2.35, 48 Mann, D., P4.16, 63 Mann, K., P11.1, 108 Manor, I., P10.31, 107 Mansour, H., P14.51, 141; P14.52, 141 Mansour, H.A., P1.23, 37 March, D., O1.6, 12 Marcheco Teruel, B., P2.3, 40 Margolis, R.L., P14.16, 131; P14.18, 132 Marini, A.M., P5.12, 66 Marino, C., P12.29, 120

Author Index Marino, E., P4.6, 62; P4.7, 62 Marques, D., P12.1, 113 Marques, F.Z.C., P9.11, 93 Marta, D.H., O1.6, 12 Martin, E.R., O5.5, 21; P10.4, 99; P10.5, 100; P10.24, 105 Martin, N.G., O9.8, 31 Martı´n, M., P14.47, 139; P14.48, 140 Martinez, M., O4.6, 19; P14.64, 143 Martı´nez, I., P17.9, 155 Martinez-Barrondo, S., P5.8, 65 Martins, S.S., P12.23, 118; P12.25, 119 Martin-Santos, R., P5.32, 71 Martinuzzi, A., P12.29, 120 Martorell, L., P3.24, 58; P13.1, 120 Martucci, L., P8.15, 85 Marusic, A., P2.15, 43 Maruszak, A., P4.9, 62 Massa, R., P1.8, 33; P1.10, 33 Massie, D., P9.31, 98 Mata, I., P7.24, 79; P8.6, 83; P8.7, 83; P13.13, 124; P13.14, 124; P18.6, 159; P18.8, 159; P18.15, 161; P18.16, 161; P18.27, 164; P18.29, 165 Mateo, P., P3.4, 52 Mathieu, F., P2.15, 43; P14.32, 136 Matsumoto, C., P8.26, 88 Matsumoto, S., P3.14, 55; P14.34, 136 Matsushita, M., O5.4, 21 Mattay, V.S., S3.4, 7 Mattila-evenden, M., P14.40, 137 Maurizio, A., P10.14, 102 Maurizio, A.M., O8.6, 28 Mazei, M., P18.18, 162 Mazei-Robison, M., P1.22, 37 Mazei-Robison, M.S., P16.16, 153 Maziade, M., P2.38, 49 Mazumder, P.P., P13.2, 121 Mbekou, V., P9.8, 92 Mc Keon, P., P2.15, 43 McAuley, E.Z., O2.5, 14 McCandless, F., P2.44, 51 McCarthy, M.Y., O8.7, 28 McCauley, J.L., P9.26, 97; P10.5, 100 McClay, J.L., O9.5, 30 McDonald, C., P3.3, 52; P3.8, 53; P3.9, 54 McDonald, G., P1.2, 31 McDonald, J.M., P18.12, 160 McDonald, M., P4.11, 63 McDonald, M.P., O3.5, 17; P16.6, 150; P16.7, 150 McElroy, S., P2.30, 47 McGann, L., O4.1, 18 McGeary, J.E., P11.8, 109; P11.9, 110 McGhee, K., P14.60, 143 McGhee, K.A., P14.10, 130; P14.41, 138 McGinnis, R.E., E2.2, 4 McGrath, J.A., P2.24, 46; P3.11, 54; P14.50, 140; P2.18, 44 McGrath, J.J., P14.23, 133 McGregor, S., O4.7, 20; O8.1, 26; P2.33, 48; P8.10, 84 McGuffin, P., S8.1, 10; P1.7, 33; P1.13, 34; P1.18, 36; P1.19, 36; P1.25, 37; P1.26, 38; P2.22, 45; P10.31, 107 McGuinness, B., P18.9, 159 McHugh, P., P18.3, 158 McHugh, S., P18.9, 159 McIlroy, S., P18.9, 159 McIlroy, S.P., P4.10, 62 McInnis, M., O1.3, 11; P2.28, 47 McInnis, M.G., O1.6, 12, P2.16, 44; P2.40, 50; P2.42, 50 McKee, I., P1.2, 31 McKeel, D., S1.3, 5 McKeon, P., P2.7, 41

McKinney, R., P18.20, 162 McLaren, L.J., P6.2, 72 McMahon, F.J., O1.1, 11; O2.2, 13; O8.4, 27; O8.5, 28; O9.7, 30; P2.28, 47 McQuillan, A., O3.8, 17; O7.7, 26; P1.14, 35; P2.10, 42; P2.36, 49; P8.18, 86; P14.8, 129; P15.10, 146 Meaburn, E., P3.23, 58 Meaburn, E.L., O9.2, 29 Meagher, D., P14.10, 130; P14.41, 138; P14.60, 143 Meally, E., O5.2, 21 Meary, A., P3.26, 58; P14.32, 136 Me´ary, A., P3.10, 54 Medda, E., P9.9, 92 Medeiros, H., O4.1, 18 Medina, R., P7.27, 80; P8.12, 84; P14.13, 131 Medina, R.A., P3.21, 57; P14.45, 139 Meerabux, J.M.A., P14.24, 133 Mehltretter, L., P10.15, 103 Mei, H., P10.24, 105 Meira-Lima, I., P18.14, 161; P2.25, 46 Meisenzahl, E.M., O7.6, 25 Meitinger, T., O1.8, 13; P8.11, 84; P14.6, 129 Melke, J., O5.6, 22 Meltzer, H., P18.17, 161 Meltzer, H.Y., P3.34, 60 Mendoza, R., P3.21, 57; P7.27, 80; P14.13, 131 Menezes, P.R., P4.8, 62; P13.22, 126 Menold, M., P10.24, 105 Menold, M.M., O5.5, 21; P10.23, 105 Menon, R., O3.2, 16 Mercadante, M.T., P5.29, 71 Mercader, J., P5.21, 68; P5.22, 69 Me´rette, C., P2.38, 49 Merikangas, K., E2.1, 4 Merikangas, K.R., P1.29, 39; P12.31, 120 Mesen-Fainrdi, A.G., P14.55, 142 Messas, G.P., P11.11, 110 Meyemberg, N., P7.27, 80; P17.5, 154 Meyer, E.T., O1.1, 11 Meyer, J., P5.14, 67 Michelon, L., P2.25, 46; P13.23, 126; P18.14, 161 Middle, F., O1.5, 12 Middle, F.A., P2.11, 42; P2.44, 51; P2.45, 51 Middleton, F.A., O4.1, 18; P2.13, 43 Miedzybrodzka, Z., P9.31, 98 Miguel, E.C., P5.20, 68; P5.28, 70; P5.29, 71 Miguita, K., P5.20, 68; P9.6, 91; P12.23, 118; P13.23, 126; P18.14, 161 Milanova, V., P1.15, 35; P2.14, 43 Mill, J., P12.10, 115; P12.11, 115; P16.4, 150 Mill, J.S., P15.8, 146 Millar, J.K., P7.22, 79; P14.2, 128 Millar, K., S3.1, 6 Miller, D.D.M., P18.1, 157 Miller, E., O1.6, 12 Miller, M.I., P3.16, 55 Miller, P., P18.9, 159 Minabe, Y., P14.24, 133 Minassian, B.A., P10.19, 104 Mincheva, A., P9.7, 91 Miralles, M.L., P14.48, 140 Miranda, A., P10.31, 107 Miret, S., P14.47, 139; P14.48, 140 Mirnics, K., P14.57, 142 Misener, V.L., P9.28, 97; P9.29, 98 Mishima, K., P3.29, 59; P15.13, 147; P16.15, 152 Mishima, Y., P3.29, 59 Mitchell, B.D., P7.12, 76 Mitchell, B.D.M., P14.30, 135 Mitchell, K.J., P3.31, 60

173

Mitchell, P.B., P1.20, 36; P2.20, 45 Mitev, V., P1.15, 35 Miyawaki, A., P2.9, 41 Mnatzakanian, G., P10.19, 104 Moe, P.R., P2.40, 50 Moffit, T.E., PL2, 3; O2.6, 15 Moffitt, T., S8.1, 10 Mo¨ller, H.J., P1.5, 32; P1.30, 39; P8.11, 84; P12.12, 115; P14.6, 129; P15.6, 145 Molteni, M., P12.29, 120 Molto´, M.D., P13.19, 125; P13.20, 125; P13.27, 127 Monachov, V.V., P6.4, 73 Monaco, A.P., O2.5, 14 Mondabon, S., P7.31, 81 Montecchi-Palazzi, L., P8.17, 85 Montero, A.P., P3.21, 57; P8.12, 84; P14.45, 139 Montero, P., P14.13, 131 Monti, P.M., P11.8, 109; P11.9, 110 Montserrat, C., P3.24, 58; P3.4, 52; P13.1, 120 Moorey, H., O7.7, 26; P8.18, 86; P14.8, 129 Morales, B., P5.8, 65 Morcus, M.E.J., P6.6, 73 Morell, M., P5.32, 71 Mori, K., P2.37, 49 Mori, N., P14.24, 133 Morita, Y., P12.14, 116 Morley, C.P., O4.1, 18; P2.13, 43 Morris, A., S5.4, 8 Morris, D., P14.60, 143 Morris, D.W., O4.2, 18; P14.10, 130; P14.41, 138; P14.57, 142 Morris, J., S1.3, 5 Morris, J.S., S6.3, 8 Morris, R., P12.19, 117 Morris, R.G., P16.12, 152 Morris, S.W., P13.12, 123 Mors, O., P1.13, 34; P1.26, 38; P2.3, 40; P2.10, 42; P4.1, 60; P5.23, 69; P6.1, 72; P6.2, 72; P9.14, 93; P13.10, 123; P13.11, 123; P16.13, 152; P16.14, 152 Moses, M.J., P2.20, 45 Moskvina, V., O4.2, 18; O7.5, 25; O9.4, 29; P1.13, 34; P4.5, 61; P9.13, 93; P17.3, 154; P17.4, 154 Mowry, B.J., O4.6, 19; P3.15, 55; P14.23, 133; P14.25, 134 Mueller, D.J., P3.34, 60 Mueller, J.C., O1.8, 13 Mueller-Myhsok, B., P1.28, 38 Muglia, P., P2.23, 45; P9.25, 96 Muhaheed, M., P8.19, 86 Muhlenhoff, L., P8.11, 84; P14.6, 129 Muir, W., S3.1, 6; O1.2, 11; P1.2, 31 Muir, W.J., P1.1, 31; P6.1, 72; P6.2, 72; P6.5, 73; P13.10, 123; P13.12, 123; P14.2, 128; P15.1, 144 Mukherjee, O., P13.2, 121; P17.6, 154 Mulet, B., P3.4, 52; P3.24, 58; P13.1, 120 Mu¨ller, D., P14.28, 135 Mu¨ller, D.J., P2.23, 45 Mu¨ller, J., P8.11, 84; P14.6, 129 Muller-Myhsok, B., O1.8, 13; P5.4, 64; P5.10, 66; P5.13, 66; P5.16, 67; P13.8, 122 Mullin, K., O3.2, 16 Mullins, C., P9.32, 99; P10.1, 99; P10.2, 99; P10.3, 99 Munakata, K., P2.9, 41 Mundo, E., P2.23, 45 Mun˜iz, M.J., P17.9, 155 Munnich, A., P10.21, 104 Mun˜oz, M.J., P14.47, 139; P14.48, 140

174

Author Index

Mun˜oz, R., P3.21, 57; P7.27, 80 Munro, C., P1.21, 36 Munro, J., P18.8, 159; P18.15, 161; P18.29, 165 Murad, I., P8.19, 86 Murali, N., P14.61, 143 Murphy, B.C., P7.30, 81; P8.5, 82 Murphy, D., P12.6, 114 Murphy, D.G.M., P10.14, 102; P12.19, 117; P16.12, 152 Murphy-Eberenz, K., O1.6, 12 Murphy, K., P12.6, 114; P14.10, 130; P14.41, 138; P14.60, 143 Murphy, K.C., O8.7, 28; P12.19, 117; P16.12, 152; P8.10, 84 Murray, J., P9.27, 97 Murray, M.M., P4.12, 63 Murray, R., P3.3, 52; P3.9, 54; P12.13, 116; P14.37, 137; P15.8, 146 Murray, R.M., O2.6, 15; P3.8, 53; P3.18, 56; P8.6, 83; P8.7, 83; P13.13, 124; P13.14, 124; P13.25, 126; P14.44, 139 Murray, V., P1.2, 31; P13.10, 123 Mustanski, B., P12.20, 117 Myers, A.J., O3.3, 16 Myers, J.M., P5.11, 66 Myers, R., P1.22, 37 Myers, R.L., P18.18, 162 Myles-Worsley, M., P14.53, 141 Nagai, K., P16.15, 152 Nagai, T., P2.9, 41 Nair, C., P14.65, 144 Na´jera, C., P13.19, 125; P13.20, 125; P13.27, 127 Nakajima, M., P8.3, 82 Nakajima, T., P16.15, 152 Nakamura, J., P8.26, 88; P14.28, 135 Nakamura, K., P14.24, 133 Nakano, T., P7.4, 75 Nakata, K., P7.4, 75 Nakatsuka, M., P12.14, 116 Nangle, J., P14.10, 130; P14.41, 138; P14.60, 143 Nangle, J.M., O1.5, 12; P2.45, 51; P3.19, 56 Nanko, S., P2.8, 41 Navarro, M.E., P14.47, 139; P14.48, 140 Navine´s, R., P5.32, 71 Navon, R., P8.19, 86; P8.20, 86; P13.17, 125 Neale, B.M., O9.8, 31; P14.27, 134; P17.17, 157 Neale, M.C., O6.1, 22; P5.11, 66; P14.26, 134; P14.54, 141 Need, A.C., P5.3, 64 Neetha, N.V., P14.65, 144 Neidt, H., P1.6, 32 Nelson, K., P15.18, 148 Nelson, S., P10.15, 103 Nemoda, Z., P3.33, 60; P7.14, 77; P9.10, 92; P10.30, 107; P12.28, 119 Nertney, D.A., P14.23, 133; P14.25, 134 Nerung, L., P9.11, 93 NeSmith, L., P8.12, 84 Nestadt, G., O6.8, 24; P2.18, 44; P3.11, 54; P14.50, 140 Neuhaus, A.H., P18.2, 158 Neuman, R.J., O6.5, 23 Neumeier, K., P15.6, 145 Neves-Pereira, M., P9.31, 98 Newell, F.N., P3.31, 60 Ni, X., P3.34, 60; P14.46, 139; P18.17, 161 Nicastro, I., O7.8, 26 Nicodemus, K., P3.11, 54

Nicodemus, K.K., P2.24, 46; P14.50, 140 Nicolini, H., P14.13, 131; P17.5, 154; P3.21, 57; P5.2, 64; P5.33, 72; P7.27, 80 Niculescu, A.B., O7.8, 26; P15.14, 147 Nierenberg, J., O8.6, 28 Nievergelt, C.M., P2.26, 46 Nikolov, I., O8.1, 26; O8.2, 27; P1.18, 36; P2.6, 41; P2.11, 42; P2.19, 44; P2.33, 48; P17.3, 154; P18.5, 158 Nikolov, I.N.N., P13.16, 125 Nikolova-Hill, A., P1.15, 35; P2.14, 43 Nilius-Hoffmann, E., P10.13, 102 Nimgaonkar, V., O3.7, 17 Nimgaonkar, V.L., O4.5, 19; P1.23, 37; P3.20, 57; P13.21, 126; P14.51, 141; P14.52, 141; P14.57, 142 Nishino, T., P3.16, 55 Nobile, M., P12.29, 120 Noethen, M., P2.15, 43 Noethen, M.M., O1.4, 12; O8.5, 28; P1.6, 32; P2.14, 43; P2.46, 51; P5.14, 67; P5.26, 70; P5.34, 72; P10.25, 106; P13.15, 124; P14.40, 137; P14.58, 142 Noguti, R., P4.8, 62 Nokolov, I., P2.44, 51 Nolan, A., P18.9, 159 Nomura, A., P7.4, 75; P11.13, 111; P12.14, 116 Norton, J., S1.3, 5 Norton, N., O4.2, 18; O4.7, 20; O7.5, 25; O8.1, 26; O8.2, 27; O8.7, 28; P8.8, 83 Nosikov, V.V., P8.29, 89 No¨then, M., O6.2, 23; O8.3, 27 Nowak, N.J., P14.16, 131 Nurnberg, P., P13.9, 123 Nurnberger, J., P9.24, 96 Nurnberger, J.I., O1.1, 11; P2.28, 47 Nurnberger, J.L., P2.40, 50 Nwulia, E.A., P2.42, 50 Nyegaard, M., P16.14, 152 Nyholt, D.R., P14.23, 133; P14.25, 134 Nyhuis, P., P5.26, 70 O’Donovan, M., O5.8, 22; P2.6, 41; P2.11, 42; P2.19, 44; P7.29, 81; P9.13, 93; P9.30, 98 O’Donovan, M.C., O4.2, 18; O4.7, 20; O7.5, 25; O8.1, 26; O8.2, 27; O8.7, 28; P2.12, 42; P4.5, 61; P8.8, 83; P8.10, 84; P9.15, 94; P9.16, 94; P13.16, 125; P13.26, 127; P14.57, 142; P14.59, 143 O’Reilly, R., P7.30, 81 O’Neill, F.A., P14.54, 141 Oades, B., P10.31, 107 Oaghere-DaSilva, U.B., O3.5, 17 O’Callaghan, E., S7.4, 10 O’Gara, C., P11.17, 112 Ogawa, F., P8.16, 85 Ogawa, S., S6.1, 8 Ogden, C.A., O7.8, 26; P15.14, 147 Ohba, H., P14.24, 133 Ohlraun, S., O1.4, 12; O2.2, 13; O8.3, 27; O8.5, 28; P1.6, 32; P2.46, 51; P5.34, 72; P13.15, 124; P14.40, 137; P14.58, 142 Ohmori, O., P8.26, 88 Okabe, N., P12.14, 116 Okawa, M., P16.15, 152 Okazaki, Y., P3.14, 55; P14.34, 136; P15.12, 147 Oliveira, G., P17.13, 156 Olsen, I.M., P16.13, 152; P16.14, 152 Olsen, L., P14.1, 127 Olson, L.M., P9.26, 97

O’Mahony, E., P2.44, 51; P3.19, 56 Onchev, G., P1.15, 35; P2.14, 43 O’Neill, A.F., P14.26, 134 O’Neill, F.A., P14.17, 132; P14.27, 134; P18.9, 159 Oniszczenko, W., P5.5, 65 Ono, S., P3.14, 55 Ontiveros, A., P3.21, 57; P7.27, 80; P14.13, 131 Opgen-Rhein, C., P18.2, 158 O’Rahilly, S., PL1, 3 O’Reilly, R., P8.5, 82 Orntoft, T.F., P2.3, 40 Orourke, J., P17.1, 153 O¨sby, U., O6.2, 23 Oshimura, M., O2.5, 14 O’Sullivan, R., P10.1, 99; P10.3, 99 Ota, J., P12.14, 116 Otani, K., P12.14, 116 Oton, A., P3.4, 52 Otte, A.C., P14.40, 137 Otto, C., P14.53, 141 Ou-Yang, W.C., P14.19, 132; P14.5 Owen, M., O1.5, 12; P1.13, 34; P1.21, 36; P1.25, 37; P2.44, 51; P2.45, 51; P3.19, 56; P7.29, 81; P9.13, 93; P9.30, 98 Owen, M.J., E1.1, 3; O4.2, 18; O4.7, 20; O7.5, 25; O8.1, 26; O8.2, 27; O8.7, 28; P1.18, 36; P1.26, 38; P2.6, 41; P2.11, 42; P2.12, 42; P2.19, 44; P2.33, 48; P4.5, 61; P8.10, 84; P8.8, 83; P9.15, 94; P9.16, 94; P12.19, 117; P13.16, 125; P13.26, 127; P14.57, 142; P14.59, 143; P16.12, 152; P17.4, 154; P18.5, 158 Ozaki, N., P7.25, 80; P7.26, 80; P11.13, 111; P11.7, 109; P16.15, 152 Ozawa, H., P3.14, 55; P14.34, 136; P15.12, 147 Ozeki, Y., P16.15, 152 Padiath, Q., P15.19, 149 Padmavati, R., P3.15, 55 Paez-Pereda, M., P13.8, 122 Pakstis, A.J., P9.17, 94 Palha, J., P2.2, 40; P7.21, 78 Palmour, R.M., P12.18, 117; P16.5, 150 Palomo, T., P17.9, 155 Pan, X., P15.18, 148 Pan, Y.X., P7.7, 75 Papapetrou, D., P14.7, 129 Papiol, S., P18.24, 163 Paracchini, S., P7.15, 77 Park, W.Y., P4.4, 61 Parker, G., P1.20, 36 Parkinson, M., O3.2, 16 Parsian, A., P11.14, 111 Parsons, M.J., P18.27, 164 Partonen, T., P3.32, 60; P7.28, 80; P14.31, 135; P14.43, 138 Parus, M., P14.33, 136 Paschou, P., P9.17, 94 Passmore, A.P., P4.10, 62 Pastor, P., S1.3, 5 Pastore, V., P12.29, 120 Pathak, V., P2.13, 43 Pathare, T., P9.22, 96; P9.23, 96 Patil, S., P9.27, 97 Pato, C., P2.2, 40; P14.46, 139 Pato, C.N., O4.1, 18; P2.13, 43 Pato, M., P14.46, 139; P2.2, 40; P5.24, 69 Pato, M.T., O4.1, 18; P2.13, 43 Patterson, D.G., O6.1, 22; P11.19, 112 Pauls, D., P17.1, 153

Author Index Pauls, D.L., P5.28, 70; P5.29, 71; P9.17, 94; P10.28, 106 Paulus, M.P., O7.8, 26; P15.14, 147 Paunio, T., P3.32, 60; P7.20, 78; P7.28, 80; P14.31, 135; P14.43, 138 Pearson, P.L., P3.22, 57; P10.10, 101; P14.29, 135 Peirce, T.R., P7.29, 81 Pelka-Wysiecka, J., P11.3, 108; P11.4, 108 Pelosi, A.J., P1.2, 31; P13.10, 123 Peltonen, J.O., P7.28, 80; P14.43, 138 Peltonen, L., P3.32, 60; P7.20, 78; P7.28, 80; P14.12, 130; P14.31, 135; P14.43, 138 Peng, Y., P7.5, 75 Pennington, K., P13.4, 121; P13.5, 121 Peralta, V., P14.42, 138; P14.47, 139 Pereira, M., P8.12, 84 Pereira, R., P8.12, 84 Perez, J., P2.29, 47; P18.19, 162 Pe´rez, M., P13.1, 120 Perez-Nievas, F., P7.24, 79; P8.7, 83; P13.13, 124; P13.14, 124 Pericak-Vance, M.A., O5.5, 21; P10.4, 99; P10.5, 100; P10.8, 101; P10.23, 105; P10.24, 105; P10.32, 107 Perlis, R., P10.6, 100 Perlmutter, J., S1.3, 5 Perry, R.T., O3.6, 17 Petryshen, T.L., O2.1, 13; O4.1, 18 Pettersson, U., O4.4, 19 Petty, C., P9.20, 95; P10.6, 100 Peyrard-Janvid, M., P10.25, 106 Pezawas, L., S3.4, 7 Pfaff, D., S6.1, 8 Pfennig, A., P5.13, 66 Philippe, A., P10.21, 104 Philippi, A., O9.1, 29 Phillips, C., P17.12, 156 Pianca, T., P10.7, 100 Piccardi, M.P., P2.1, 39 Picchioni, M.M., P3.18, 56; P12.13, 116; P13.25, 126; P14.44, 139 Pickard, B., S3.1, 6 Pickard, B.S., P6.1, 72; P6.2, 72; P6.5, 73 Pickles, A., P9.9, 92 Pimm, J., O7.7, 26; P8.18, 86; P14.8, 129 Pinheiro, S.N., P11.10, 110 Pirovano, A., P1.10, 33; P1.11, 34; P1.12, 34; P1.24, 37; P18.21, 163 Pittman, A., O3.3, 16 Platko, J.V., P17.1, 153 Plocik, A.M., P14.22, 133 Ploia, C., P1.10, 33; P1.11, 34; P1.24, 37; P18.10, 160 Plomin, R., O6.3, 23; O9.2, 29; P3.23, 58; P10.31, 107; P12.8, 114; P12.21, 118; P12.22, 118 Poelmans, G., O5.7, 22; P10.13, 102 Pogue-Geile, M., O3.7, 17 Poluda, L., P5.10, 66; P5.4, 64 Ponce, G., P17.9, 155 Porjesz, B., P9.24, 96 Portaro, F.C.V., P14.3, 128 Porteous, D., S3.1, 6 Porteous, D.J., S3.14, O1.2, 11; P2.14, 43; P6.1, 72; P6.2, 72; P6.5, 73; P7.22, 79; P13.12, 123; P14.2, 128 Posthuma, D., P12.32, 120 Potapova, N., P17.2, 153; P14.28, 135 Potash, J.B., O8.4, 27; P2.40, 50 Pottgiesser, S., P1.6, 32 Poulton, R., O2.6, 15 Poustka, A., P9.4, 91; P9.7, 91 Poustka, F., P9.4, 91; P9.7, 91 Powell, J., P9.5, 91 Powell, J.F., P6.3, 73

Prasad, H., P1.22, 37 Prasad, H.C., P18.18, 162 Prasanna, C.L.N., P5.18, 68; P5.19, 68 Preece, A., O7.5, 25; O8.1, 26 Preece, A.C., O4.2, 18; P8.8, 83 Preisig, M., P1.13, 34; P1.26, 38; P2.15, 43 Prell, K., O3.4, 16 Prescott, C.A., O6.1, 22; S8.3, 10; P11.19, 112; P5.11, 66 Preuss, U.W., P11.6, 109 Priddle, T.H., P14.39, 137 Propping, P., PL4, 3; O1.4, 12; O6.2, 23; O8.3, 27; O8.5, 28; P1.6, 32; P2.14, 43; P2.46, 51; P5.34, 72; P10.25, 106; P13.15, 124; P14.40, 137; P14.58, 142 Pulford, D.J., O1.2, 11 Pulver, A.E., P2.18, 44; P2.24, 46; P3.11, 54; P14.50, 140 Purcell, S.M., S6.3, 8; S8.4, 11 Puri, V., P2.36, 49; P8.18, 86; P14.8, 129 Purmann, S., P3.7, 53 Purushottam, M., P13.2, 121; P17.6, 154 Pu¨tz, B., O1.8, 13; P1.28, 38; P5.4, 64; P5.10, 66; P5.13, 66; P5.16, 67 Pylkko¨, E., P14.31, 135 Qasim Brohi, M., P7.16, 77 Qin, W., P7.5, 75; P7.10, 76 Quach, H., O5.6, 22 Quested, D., O7.7, 26; P8.18, 86; P14.8, 129 Quinn, J., P14.10, 130; P14.41, 138; P14.60, 143 Quinn, J.P., O7.4, 25; P15.7, 145 Quintela, I., P17.12, 156 Rabionet, R., P10.4, 99; P10.24, 105 Racette, B., S1.3, 5 Racette, S., O6.7, 24 Rademakers, R., S1.1, 5 Radlwimmer, B., P9.7, 91 Ræder, M.B., P15.16, 148; P15.17, 148 Rafail, E., P7.16, 77 Rajewski, A., P5.15, 67 Rajewski, A.R., P5.6, 65; P5.7, 65 Ramirez, M., P3.21, 57 Ramoz, N., O5.3, 21 Randazzo, R., P18.19, 162 Raskind, W.H., O5.4, 21 Rasmussen, H.B., P3.28, 59; P8.22, 87; P14.1, 127; P14.35, 136; P17.11, 156; P18.13, 160 Rastam, M., O5.6, 22 Rathe, J.A., O8.4, 27 Ratnanather, J.T., P3.16, 55 Ravento´s, H., P3.21, 57; P7.27, 80; P8.12, 84; P14.45, 139 Raybould, R., O1.5, 12; P2.44, 51; P2.45, 51 Raybould, R.E., O8.2, 27; P, P1.18, 36; P2.6, 41; P2.11, 42; P2.33, 48 Raymond, C., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Reddy, Y.C.J., P5.18, 68; P5.19, 68 Reich, T, P1.13, 34; P1.26, 38; P2.28, 47 Reichert, J., O5.3, 21 Rein, T., O1.8, 13 Reinhardt, R., P7.31, 81 Reis, A., P3.17, 56 Reitschel, M., P1.13, 34 Religa, D., P4.9, 62 Remick, R.A., P2.30, 47 Remschmidt, H., P10.25, 106 Reynolds, G.P., O8.8, 28; P18.11, 160 Reynolds, N.J., P1.7, 33 Reynolds, G., O3.8, 17 Rezaie, R., P10.14, 102

175

Rian, J.C., P8.17, 85 Ribase´s, M., P5.21, 68; P5.22, 69 Ribble, R.C., P14.17, 132; P14.27, 134 Ribchester, T., P3.18, 56 Rice, F., P12.16, 116 Rice, F.J., P9.15, 94 Rice, J., P2.40, 50 Rich, M.E., O7.8, 26; P15.14, 147 Richardson, A.J., O2.5, 14 Richter, M.A., P5.24, 69 Riddle, M., O6.8, 24 Rietschel, M., O1.4, 12; O2.2, 13; O3.4, 16; O8.3, 27; O8.5, 28; P1.6, 32; P1.26, 38; P2.15, 43; P2.46, 51; P5.34, 72; P8.19, 86; P13.15, 124; P14.40, 137; P14.58, 142 Rietz, C., O3.4, 16 Rijsdijk, F., P12.13, 116 Riley, B., P14.17, 132 Riley, B.P., S8.3, 10; O6.1, 22; P, P11.19, 112; P14.26, 134; P14.27, 134; P14.54, 141 Rinn, J., P15.18, 148 Ritchie, M., P1.22, 37; P10.24, 105 Ritchie, M.D., P18.18, 162 Rivero, O., P13.19, 125; P13.27, 127 Rizig, M.A., O7.7, 26; P8.18, 86; P14.8, 129; P15.10, 146 Roberts, N., P10.14, 102 Roberts, J., O7.4, 25 Roberts, S., P9.9, 92 Roberts, W., P9.22, 96; P9.23, 96; P9.28, 97 Roberts, W.S., P10.19, 104 Robertson, D.M.W., P10.14, 102 Robertson, E., O1.5, 12; P2.45, 51 Robertson, I.H., P3.5, 53 Robertson, Z., P14.15, 131 Roche, S., P2.7, 41 Rockwell, G., O2.1, 13; O4.1, 18 Rodrı´guez, T., P13.1, 120 Rodrı´guez-Jime´nez, R., P17.9, 155 Roeder, K., P14.53, 141 Roeyers, H., P10.31, 107 Rogers, C., P3.16, 55 Rogers, G., P18.3, 158 Rogers, G.R., P15.5, 145 Rohde, L.A., P10.7, 100 Rohsenow, D.J., P11.8, 109; P11.9, 110 Roman, T., P10.7, 100 Romano-Silva, M.A., P12.1, 113; P12.2, 113 Rommelspacher, H., P11.3, 108; P11.4, 108; P11.16, 111 Ronai, Z., P7.14, 77; P12.3, 113 Ronald, A., O6.3, 23 Rosa, A., P1.25, 37; P3.18, 56; P13.25, 126; P14.42, 138; P14.47, 139 Rosario-Campos, M.C., P5.28, 70; P5.29, 71 Roschmann, E., O9.1, 29 Rose, R.J., P12.20, 117 Roselli, C.E., S6.4A, 9 Rosenbaum, J.F., O6.7, 24 Ross, C.A., P2.16, 44; P14.16, 131; P14.18, 132 Ross, N.A., P15.9, 146 Rossattini, M., P18.7, 159 Rossi, G., P18.19, 162 Roth, E., P10.25, 106 Rothe, C., P5.25, 69; P5.26, 70; P5.27, 70; P14.56, 142 Rothenberger, A., P10.31, 107 Rotondo, A., P2.29, 47; P2.35, 48 Rouleau, G., P16.8, 150 Rouleau, G.A., P7.16, 77 Rousseau, F., O9.1, 29 Rowe, D., P12.26, 119

176

Author Index

Roy, M.-A., P2.38, 49 Roy, I., P3.26, 58; P14.32, 136 Rubio, G., P17.9, 155 Rucci, P., P12.29, 120 Rudinger, G., O3.4, 16 Ruether, E., P13.15, 124 Rujescu, D., O7.6, 25; P1.30, 39; P8.9, 83; P8.11, 84; P11.1, 108; P14.6, 129 Rupprecht, R., O1.8, 13; P1.5, 32; P1.27, 38; P18.23, 163 Rust, S., P3.7, 53 Rusu, I., P14.64, 143 Rybakowski, J.K., P2.21, 45; P2.34, 48; P3.6, 53; P3.27, 59; P14.33, 136 Rybakowski, F., P5.15, 67 Rybakowski, F.R., P5.6, 65; P5.7, 65 Sachs, N.A., P14.16, 131; P14.18, 132 Sadovnick, A.D., P2.30, 47 Saenz, J., P8.12, 84 Saetre, P., O4.4, 19 Safranow, K., P4.9, 62 Sagvolden, T., P16.4, 150 Sahakian, B., P2.17, 44 Saito, T., P8.31, 89; P8.32, 89 Saitoh, H., P3.29, 59 Saitou, N., O4.6, 19 Saiz, P.A., P5.8, 65 Sakai, A., P7.4, 75 Sakai, J., P7.17, 78 Sakul, H., O1.7, 13 Salazar, R., P3.21, 57; P8.12, 84; P14.13, 131 Salgado, C.A.I., P9.11, 93 Salyakina, D., O1.8, 13; P1.28, 38; P5.4, 64; P5.10, 66; P5.13, 66; P5.16 Samochowiec, A., P11.3, 108; P11.4, 108 Samochowiec, J., P11.3, 108; P11.4, 108; P14.33, 136 Samuels, J., O6.8, 24 San, L., P18.11, 160 Sand, P., P5.34, 72 Sand, P.G., P5.14, 67; P18.26, 164 Sander, T., P13.9, 123; P18.2, 158 Sanders, A.R., O4.6, 19; P14.64, 143 Sanders, R., P12.9, 115 Sanders, R.D., O8.7, 28; P3.20, 57 Sanders-Bush, E., P1.22, 37; P18.18, 162 Sandor, P., P9.17, 94 Sang, H., P7.5, 75; P7.6, 75; P7.10, 76; P7.11, 76 Sanjev, J., P13.28, 127 Sanjua´n, J., P13.19, 125; P13.20, 125; P13.27, 127 Sannesy, S., P14.20, 133 Santangelo, S.L., P9.21, 95 Sanz, I., P3.24, 58 Sarginson, J., P8.4, 82 Sartorelli, A., P13.23, 126 Sasvari-Szekely, M., P7.14, 77; P9.10, 92; P11.2, 108; P12.3, 113; P12.28, 119 Sato, T., P12.14, 116 Satoh, K., P3.29, 59; P15.13, 147 Saucedo, E., P14.13, 131 Sawa, A., S3.2, 6; O1.3, 11 Scassellati, C., P2.29, 47 Scazufca, M., P4.8, 62; P13.22, 126 Schachar, R., P9.22, 96; P9.23, 96; P9.28, 97 Schalkwyk, L., O9.2, 29; P3.23, 58 Schall, U., P15.3, 144 Schalling, M., P2.43, 50 Schanen, N.C., P10.19, 104 Scharfetter, C., P13.9, 123 Scheftner, W., P2.28, 47; P2.40, 50 Scheftner, W.A., O1.6, 12

Scherer, S.W., P10.19, 104 Schibler, U., O7.8, 26 Schirmacher, A., P5.14, 67; P5.26, 70 Schloegelhofer, M., P10.11, 101 Schmidt, E., P8.9, 83 Schmidt, F., P2.41, 50 Schmidt, M.H., P11.10, 110 Schmitt, A., P7.31, 81 Schmitz, M., P10.7, 100 Schoenberg, P., P3.8, 53 Schoenmaekers, E., O5.7, 22 Schofield, C.J., P14.39, 137 Schofield, P.R., P2.20, 45; P16.3, 149 Schork, N.J., P2.26, 46; P2.27, 46; P2.28, 47; P2.30, 47; P15.14, 147 Schosser, A., P10.11, 101; P13.7, 122 Schott, B.H., P12.7, 114 Schottenfled, R.S., P18.28, 164 Schrander-Stumpel, C., O5.7, 22; P10.13, 102 Schrander-Stumpel, C.T., O7.2, 25 Schuckit, M., P9.24, 96 Schug, M.D., P10.15, 103 Schule, C., P18.23, 163 Schu¨le, C., P1.5, 32 Schulle, C., P1.27, 38 Schulte-Ko¨rne, G., P10.25, 106 Schulze, T.G., O1.4, 12; O2.2, 13; O8.3, 27; O8.5, 28; O9.7, 30; P1.6, 32; P2.46, 51; P5.34, 72; P13.15, 124; P14.40, 137; P14.58, 142 Schulze, K., P12.13, 116 Schumacher, J., O1.4, 12; O6.2, 23; O8.3, 27; O8.5, 28; P1.6, 32; P2.14, 43; P2.46, 51; P5.34, 72; P10.25, 106; P13.15, 124; P14.40, 137; P14.58, 142 Schumann, G., P11.1, 108 Schu¨rhoff, F., P3.10, 54; P3.26, 58; P14.32, 136 Schurov, I., P8.17, 85 Schuster, C., P9.4, 91 Schwab, S.G., O4.1, 18; P7.31 81; P14.14, 131; P14.64, 143 Schwaiger, S., P14.10, 130; P14.41, 138; P14.60, 143 Schwartz, G., P9.8, 92 Schwinger, E., P3.7, 53 Scimone, A., P2.20, 45 Scivolleto, S., P11.15, 111 Scorza, R., P18.7, 159 Scott, A., O7.4, 25; P15.7, 145 Scott, J., P2.22, 45 Scott, R.J., P15.3, 144; P15.4, 145 Scully, P., P14.10, 130; P14.41, 138; P14.60, 143 Sedvall, G.C., P14.40, 137 Segman, R.H., E3.3, 5 Segurado, R., O5.2, 21; O9.6, 30; P2.44, 51; P3.19, 56; P13.26, 127 Seidenbecher, C.I., P12.7, 114 Seidman, L.J., P8.27, 88; P14.11, 130 Sekine, Y., P11.13, 111; P11.7, 109, P14.24, 133 Selby, P., P12.9, 115 Sellner, D.B., P12.7, 114 Selten, J.P., P3.22, 57; P14.29, 135 Seltman, H., P14.57, 142 Semwall, P., P14.57, 142 Sengupta, S., P9.8, 92 Seo, J.S., P4.4, 61 Sergeant, J.O., P10.31, 107 Serrano, C., P17.10, 155 Serreli, C., P2.1, 39 Serretti, A., E3.1, 4; P1.11, 34; P1.8, 33; P1.9, 33; P1.10, 33; P1.12, 34;

P1.24, 37; P4.2, 61; 12.17, 117; P18.10, 160; P18.21, 163; P18.22, 163 Seshadri, S.P., P10.20, 104 Severino, G., P2.1, 39 Sheehan, K., P9.32, 99 Shahsavand-Ananloo, E., P14.9, 129 Shaikh, S., P9.18, 94; P10.16, 103; P14.28, 135 Shaked, M., P9.3, 90 Sham, P., O9.2, 29; P1.18, 36; P2.22, 45; P3.23, 58; P3.3, 52; P3.8, 53; P3.9, 54; P7.13, 77; P9.9, 92; P12.13, 116; P14.37, 137; P14.62, 143; P14.9, 129; P10.31, 107 Sham, P.C., S5.1, 7; P17.17, 157 Shamir, E., P8.20, 86 Sharma, V., O3.8, 17 Sharpe, P.T., S7.1, 9 Shavitt, R.G., P5.20, 68; P5.29, 71 Shaw, D., P8.4, 82 Sheehan, K., P9.32, 99; P10.1, 99; P10.3, 99; P10.12, 102 Shekhtman, T., O2.8, 15; P2.30, 47; P2.43, 50; P18.20, 162 Shelton, R., P1.22, 37; P18.18, 162 Sheshadri, S., P5.18, 68 Shi, X.Z., P7.11, 76 Shi, Y.Y., P7.9, 76; P7.11, 76 Shibui, K., P16.15, 152 Shimabokuro, F., P5.28, 70 Shimabukuro, M., P14.34, 136 Shimizu, H., P14.24, 133 Shink, E., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Shinkai, T., P8.26, 88; P14.28, 135; P14.56, 142; P18.17, 161 Shirts, B.H., O4.5, 19 Shiwei, D., P4.3, 61 Shriberg, L., P9.27, 97 Shugart, Y.Y., O6.8, 24 Sicard, T., P2.23, 45; P3.34, 60; P5.24, 69; P14.28, 135 Sieghart, W., P13.7, 122 Siesser, W.B., P16.6, 150; P16.7, 150 Silander, K., P7.28, 80; P14.31, 135 Sillaber, I., P13.8, 122 Silverman, J., O5.3, 21 Silverman, J.M., O4.6, 19 Simonoff, E., P12.19, 117 Simpson, E.M., O7.1, 24; P17.8, 155 Sinclair, M., P8.4, 82; P14.15, 131 Sindhi, M., P7.16, 77 Singh, A.G., P12.26, 119 Singh, S.M., P7.30, 81; P8.5, 82 Sinibaldi, L., O2.7, 15 Sinke, R.J., P6.7, 73; P10.10, 101; P10.28, 106; P14.29, 135 Sinke, R.S., P3.22, 57 Sinsheimer, J.S., P14.43, 138 Siu, V., P7.30, 81 Skaar, D., P10.24, 105 Skaar, D.A., P10.5, 100 Skamene, E., P16.8, 150 Skibinska, M., P2.21, 45; P2.34, 48; P14.33, 136 Skibinska, M.S., P5.6, 65 Sklar, P., S6.3, 8; O2.1, 13; O4.1, 18; O5.1, 20; O6.7, 24; P10.6, 100; P2.13, 43; P8.13, 84; P9.20, 95 Skoog, I., O3.1, 15 Skrede, S., P15.16, 148; P15.17, 148 Skuse, D.H., S6.3, 8 Slopien A., P5.15, 67 Slopien A.S., P5.6, 65; P5.7, 65 Sluijs, S., P2.31, 47

Author Index Smeraldi, E., P1.8, 33; P1.9, 33; P1.10, 33; P1.11, 34; P12.17, 117; P18.10, 160; P18.21, 163; P1.12, 34; P1.24, 37 Smith, C.J., O5.3, 21 Smith, D.J., P1.1, 31; P15.1, 144 Smith, G., P15.1, 144; P18.20, 162 Smith, G.G., P11.18, 112 Smith, H.J., P14.23, 133; P14.25, 134 Smoller, J.W., O5.1, 20; O6.7, 24; P10.6, 100 Smyth, C., P1.14, 35 Snyder, M., P15.18, 148 Soares, M., P2.2, 40 Sobell, J.L., P14.57, 142 Soderstrom, H., O5.6, 22 Sodhi, M.S., P8.25, 88 Soerensen, M., P8.30, 89 Solano, R., P5.21, 68; P5.22, 69 Sommer, S., P12.24, 119; P17.13, 156 Sora, I., P11.7, 109; P11.13, 111 Soyka, M., P11.1, 108; P11.6, 109; P11.12, 110 Spalletta, G., O2.7, 15 Spector, T.D., P5.3, 64 Speed, W.C., P9.17, 94 Sprandel, J., P2.41, 50 Squassina, A., P9.25, 96 Srinath, S., P10.20, 104 Srinivasan, T.N., P3.15, 55; P14.23, 133 St. Clair, D., P7.11, 76; P8.4, 82; P9.31, 98; P14.15, 131; P14.36, 137 Staal, W.G., P6.8, 74 Staddon, S., P8.6, 83; P18.8, 159; P18.16, 161 Stadler, H., P13.8, 122 Stallings, R., O5.2, 21 Stamenkovic, M., P10.11, 101 Stansberg, C., P15.16, 148; P15.17, 148 Stapleton, J., P11.17, 112 Stassen, H.H., P13.9, 123 Steele, C.J.M., O1.1, 11; O8.4, 27 Steen, V.M., P15.16, 148; P15.17, 148 Stein, D., P5.1, 64 Stein, J.F., O2.5, 14 Steinhausen, H., P10.31, 107 Steinlein, O., P18.2, 158 Sterne, A., P1.25, 37 Stevens, A.F., P12.19, 117; P16.12, 152 Stever, C., P10.9, 101; P12.26, 119 Stewart, M., P1.1, 31 Steyaert, J., O5.7, 22; O7.2, 25; P10.13, 102 Stine, O.C., P7.12, 76; P14.30, 135 Sto¨ber, G., P2.41, 50; P3.17, 56 Stoddart, S.D.R., P3.25, 58 Stoeber, G., P5.14, 67 Stoegbauer, F., P5.14, 67; P5.26, 70 Stoertebecker, P., P18.26, 164 Stoll, M., P3.7, 53 Stone, W.S., P8.27, 88; P14.11, 130 St-Onge, L., P13.8, 122 Stopkova, P., P8.32, 89 Stoyanova, V., P1.15, 35; P2.14, 43 Straub, R., S3.3, 6 Straub, R.S., S3.4, 7 Straub, R.E., P14.49, 140 Strauch, K., P2.46, 51 Strauss, J., P2.23, 45; P9.18, 94; P14.56, 142; P10.16, 103 Stuffrein, S., P15.5, 145 Su, J.L.S., P13.3, 121 Sua´rez, M., P9.19, 95 Sugden, K., S8.1, 10; P15.2, 144 Suh, Y.J., P4.4, 61 Sujit, K.J., P3.15, 55; P14.23, 133 Suliman, H., P14.58, 142 Sullivan, P.F., O6.1, 22; O9.8, 31; P11.19, 112

Summerfeldt, L.J., P5.24, 69 Sun, H., P14.38, 137 Sun, L.W., O6.5, 23 Sun, X., P14.36, 137; P14.37, 137 Sun, Y., P7.7, 75; P7.19, 78 Sutcliffe, J.S., P9.26, 97; P10.5, 100 Sutherland, G., P11.17, 112 Suwalaska, A., P2.34, 48 Suzuki, K., P14.24, 133 Suzuki, T., P7.25, 80; P7.26, 80 Svetina, C., O8.6, 28 Swaab-Barneveld, H., P6.6, 73 Swainson, V., P1.18, 36 Swartz, M., P8.20, 86 Swift, R.M., P11.8, 109; P11.9, 110 Syrek, S., P11.3, 108; P11.4, 108; P14.33, 136 Szantai, E., P7.14, 77 Szczepankiewicz, A., P2.34, 48 Szegedi, A., P13.9, 123 Szekely, A., P12.3, 113; P15.18, 148 Szeszko, P.R., O7.3, 25 Szilagyi, A., P11.2, 108 Szo¨ke, A., P3.10, 54; P3.26, 58; P14.32, 136 Szulc, K., O8.6, 28 Tabbane, K., P7.16, 77 Tahl, A.R., O4.1, 18 Tajima, O., P2.37, 49 Takahashi, K., P16.15, 152 Takaki, M., P7.4, 75 Takano, A., P16.15, 152 Takao, H., P8.2, 82; P8.3, 82 Takei, N., P14.24, 133 Talbot, C.J., P4.16, 63 Talkowski, M.E., P1.23, 37; P13.21, 126; P14.51, 141; P14.52, 141; P14.57, 142 Tan, C.H., P7.8, 75 Tan, E.C., P7.8, 75 Tanaka, H., P17.16, 157 Tanaka, M., P2.9, 41 Tanaka, T., P11.7, 109 Tanaka, Y., P7.4, 75; P11.13, 111; P12.14, 116 Tandon, K., P1.26, 38 Tang, J.X., P7.11, 76 Tang, R.Q., P7.11, 76 Tang, Y., P15.18, 148 Tanko, L.B., P3.28, 59 Tannock, R., P9.12, 93; P9.22, 96; P9.23, 96; P9.28, 97 Tanzi, R.E., O3.2, 16 Taratine, B., P17.2, 153 Tarnok, Z., P9.10, 92; P12.3, 113; P12.28, 119 Tarter, R.E., O2.4, 14 Tasaki, S., P3.14, 55; P15.12, 147 Templeman, L.A., O8.8, 28; P18.11, 160 Terenius, L., O6.2, 23 Ter-Stepanian, N., P9.8, 92 Thaker, G.K., P7.12, 76; P14.30, 135 Thapar, A., O5.8, 22; P9.15, 94; P9.30, 98; P12.16, 116 Thara, R., P3.15, 55; P14.23, 133 Tharmalingam, S., P3.34, 60; P5.25, 69; P5.27, 70 Thase, M.E., P1.23, 37 The NIMH Genetics Initiative Bipolar Disorder Group, O1.1, 11 Thelma, B.K., P14.57, 142 Theuns, J., S1.1, 5 Thiele, H., P13.9, 123 Thierfelder, K., P8.11, 84; P14.6, 129

177

Thirumalai, S., O7.7, 26; P8.18, 86; P14.8, 129 Thiselton, D.L., P14.17, 132; P14.27, 134 Thoeringer, C.K., P5.4, 64; P5.10, 66 Thomas, J., O1.6, 12 Thomas, S., P9.9, 92 Thomson, A.D., O3.8, 17 Thomson, A.M., O1.2, 11 Thomson, J.B., O5.4, 21 Thomson, P.A., O1.2, 11; P13.12, 123; P14.2, 128 Tidey, J.W., P11.8, 109; P11.9, 110 Tidow, N., P3.7, 53 Timmermann, B., P7.31, 81 Tingwei, Guo, P4.3, 61 Tiobek, S., P14.53, 141 Todd, J., P15.3, 144 Todd, R., P10.26, 106 Todd, R.D., O6.5, 23; P3.16, 55; P10.17, 103 Togsverd, M., P17.11, 156 Tolosa, A., P13.20, 125 Tolosa, E., S1.3, 5 Tomaiuolo, F., O2.7, 15 Tomaru, Y., P14.24, 133 Tomblin, J., P9.27, 97 Toncheva, D., P18.5, 158 Toncheva, D.I.T., P13.16, 125 Tooney, P.A., P15.3, 144; P15.4, 145 Torkamanzehi, A., P9.8, 92; P16.8, 150 Torralbas, M., P2.3, 40 Torrance, H.S., P13.12, 123 Torres, M., P17.12, 156 Toulopoulou, T., P3.18, 56; P3.9, 54; P13.25, 126; P14.44, 139 Toyoshima, R., P16.15, 152 Toyota, T., P8.2, 82; P8.3, 82; P14.24, 133 Tozawa, T., P3.29, 59 Trakadis, Y.J., P16.5, 150 Trakalo, J., P5.24, 69; P8.15, 85; P17.2, 153; P14.28, 135 Trandafir, A., P3.26, 58 Treasure, J., P5.31, 71 Tremblay, M., P2.4, 40; P2.5, 40; P13.8, 122; P16.11, 151 Trinh, E., O7.1, 24 Trixler, M., P14.14, 131 Tsuang, M.T., O6.7, 24; P1.16, 35; P8.27, 88; P13.3, 121; P14.11, 130; P14.20, 133 Tsujita, T., P3.14, 55; P15.12, 147 Tubazio, V., P1.11, 34; P1.12, 34; P1.24, 37 Tullius, M., O8.5, 28; P13.15, 124; P14.58, 142 Tully, T., PL3, 3 Tunbridge, E., P8.25, 88 Tura, G.B., P18.19, 162 Turic, D., P9.15, 94; P9.30, 98 Turnbull, J., P4.16, 63 Turunen, J.A., P14.12, 130; P14.43, 138 Tuulio-Henriksson, A., P3.32, 60 Tybura, P., P14.33, 136 Tyler-Smith, C., P7.15, 77 Uchida, N., P7.4, 75 Uchimura, N., P11.7, 109 Uchiyama, M., P16.15, 152 Ueno, M., P8.28, 88; P11.7, 109 Uhr, M., O1.8, 13 Ujike, H., P7.4, 75; P11.13, 111; P11.7, 109; P12.14, 116 Ulas, V.Y., P9.2, 90; P16.9, 151; P16.10, 151 Underwood, S., P13.12, 123 Unschuld, G., P5.10, 66 Unschuld, P.G., P5.4, 64; P5.16; P1.28, 38 Urban, A.E., P15.18, 148 Urraca, N., P7.27, 80 Uvarova, L.G., P3.2, 52

178

Author Index

Valente, J., P2.2, 40; P14.46, 139 Valero, J.S., P3.4, 52; P3.24, 58; P13.1, 120 Vallada, H., P2.25, 46; P4.8, 62; P5.20, 68; P11.11, 110; P13.21, 126; P13.22, 126; P13.23, 126; P18.14, 161 Vallada, H.P., P9.6, 91; P12.23, 118; P12.25, 119 Valle, D., O6.8, 24; P2.24, 46; P3.11, 54; P14.50, 140 Vallejo, E., P17.5, 154 Vallejo, J., P5.21, 68; P5.22, 69 Van Amelsvoort, T., P3.30, 59 Van Beijsterveldt, C.E.M., O6.4, 23 Van Belzen, M.J., P12.32, 120 Van Broeckhoven, C., S1.1, 5; P15.15, 147; P2.31, 47 van Daalen, E., P6.8, 74 Van de Ven, W.J.M., O7.2, 25 Van Den Bogaert, A., O1.4, 12; O8.3, 27; P14.40, 137 van den Bree, M., P9.13, 93; P9.16, 94; P9.30, 98 van den Bree, M.B., P9.15, 94 van den Oord, E.J., O6.1, 22; P11.19, 112; P14.17, 132; P14.26, 134 van den Oord, E.J.C.G., O9.5, 30; P14.54, 141 Van der Meulen, E.M., P10.10, 101; P10.28, 106 van Duijn, C., S1.1, 5 van Engeland, H., P6.6, 73; P6.8, 74 van Haren, N.E., P3.22, 57 Van Lent, J., O5.7, 22 Vander Molen, J.E., P14.64, 143 Vang, M., P13.11, 123 Vanyukov, M.M., O2.4, 14 Vanzin, L., P12.29, 120 Varilo, T., P7.28, 80; P14.31, 135; P14.43, 138 Varnai, Z., P3.33, 60 Vasconcellos, K., P13.22, 126 Vega, S., P17.9, 155 Velayudhan, L., P14.61, 143 Veltman, M., P9.5, 91; P9.9, 92 Venken, T., P2.31, 47 Verchinski, B.A., S3.4, 7 Verduijn, W., P10.10, 101; P14.29, 135 Verloes, A., P10.21, 104 Veske, A., O7.6, 25 Veske, S., O7.6, 25 Vetro, A., P9.29, 98 Vicente, A., P17.13, 156 Victor, M.M., P9.11, 93 Viken, R., P12.20, 117 Vik-Mo, A.O., P15.16, 148; P15.17, 148 Vilain, E., S6.4A, 9 Vilella, E., P3.4, 52 Villard, L., P16.9, 151; P16.10, 151 Vincent, J.B., P10.19, 104 Visscher, P., P1.2, 31 Visscher, P.M., P17.17, 157 Vittum, J., S8.3, 10; O6.1, 22; P11.19, 112 Vladimirova, R.V., P13.16, 125 Voineskos, D.Z.A., P14.56, 142 Volavka, J., P3.34, 60; P18.17, 161 Vorsanova, S.G., P6.4, 73; P9.2, 90; P16.9, 151; P16.10, 151 Vorstman, J.A.S., P6.6, 73; P6.7, 73; P6.8, 74 Vostrikov, VM, P6.4, 73 Vrabel, C., P2.43, 50 Waddington, J.L., S7.4, 10; O4.2, 18; P14.10, 130; P14.41, 138; P14.60, 143

Wade, C.M., O2.1, 13 Waggoner, S.G., O2.1, 13; O4.1, 18 Wagner, G., P5.31, 71 Wagner, S.L., O3.2, 16 Waheed, J.F., P5.12, 66 Wait, R., P13.5, 121 Waldman, I.D., P10.18, 103; P12.26, 119 Walker, M.T., P14.2, 128 Walker, N., P8.4, 82; P14.15, 131 Walsh, D., O6.1, 22; P11.19, 112; P14.17, 132; P14.26, 134; P14.27, 134; P14.54, 141 Walss-Bass, C., P8.12, 84; P14.45, 139 Walther, D., P16.11, 151 Wang, A.G., P5.23, 69; P8.22, 87; P9.14, 93; P9.14, P13.11, 123; P14.35, 136 Wang, H.Y., P9.1, 90 Wang, L.J., P7.3, 74 Wang, P., P7.6, 75; P7.10, 76; P14.62, 143 Wang, S.J., O4.3, 18 Wang, X.Y., O4.3, 18 Wang, Y., O6.8, 24 Warnke, A., P10.25, 106 Washizuka, S., P2.37, 49 Watanabe, T., P16.15, 152 Watson, B., O3.6, 17 Webb, B.T., O6.1, 22; P14.27, 134; P14.54, 141 Weber, M., P18.2, 158 Weber, S., P1.6, 32 Webster, C.M.C., P11.10, 110 Wechsung, V., P18.26, 164 Weckx, S., P15.15, 147 Wei, Q., P7.6, 75 Weidenhofer, J.C., P15.3, 144; P15.4, 145 Weinberger, D.R., S3.4, 7; P14.57, 142; P14.49, 140 Weisbrod, M., P13.9, 123 Weissbecker, K., P5.2, 64; P5.33, 72 Weissman, M.M., O1.6, 12 Weissman, S.M., P15.18, 148 Weizman, A., P5.1, 64 Wellek, S., O8.5, 28 Werge, T., P3.28, 59; P8.22, 87; P8.30, 89; P14.1, 127; P17.11, 156; P18.13, 160 Werge, T.W., P14.4, 128; P14.35, 136 Wernicke, C., P11.16, 111 Wheelwright, S., O6.3, 23 Whitehead, P.L., O5.5, 21 Whiting, P.J., P8.17, 85 Whittinger, N., P9.30, 98 Wickham, H., P3.9, 54 Wiener, H., O3.6, 17 Wienker, T.F., P2.46, 51 Wigg, K., P9.29, 98 Wigg, K.G., P9.12, 93; P9.23, 96 Wijsman, E.M., O5.4, 21 Wikman, F., P2.3, 40 Wildenauer, D.B., O4.1, 18; P7.31, 81; P14.14, 131; P14.64, 143 Wilhelm, K., P1.20, 36 Wilkie, E., P1.7, 33 Wilkie, M., P15.1, 144 Wilkinson, J., O7.5, 25 WiIkinson, J.C., O4.2, 18 Wilkinson, L.S., P16.1, 149 Williams, C., P9.27, 97 Williams, H., O4.7, 20; O8.1, 26; O8.2, 27; O8.7, 28 Williams, H.J., O4.2, 18; O7.5, 25; P8.8, 83; P8.10, 84 Williams, J., S1.2, 5; P4.5, 61; P9.13, 93; P9.16, 94; P9.31, 98; P17.3, 154 Williams, N., O8.2, 27; P7.29, 81

Williams, N.A., P3.12, 54; P7.15, 77 Williams, N.M., O4.2, 18; O4.7, 20; O7.5, 25; O8.1, 26; O8.7, 28; P8.10, 84; P8.8, 83; P13.26, 127; P14.57, 142; P14.59, 143 Willour, V.L., O6.8, 24; P2.40, 50 Wilson, D., P12.25, 119 Wilson, G.G., O9.3, 29 Windemuth-Kieselbach, C., P2.46, 51 Winstead, D., P5.33, 72 Winter, T., P12.20, 117 Wire´n, K., O6.2, 23 Wochnik, G., O1.8, 13 Wolff, R., P15.1, 144 Wolpert, C.M., P10.8, 101; P10.23, 105; P10.32, 107 Wolyniec, P.S., P2.18, 44; P2.24, 46; P3.11, 54; P14.50, 140 Wong, A.H.C., P14.28, 135 Wong, G., P14.28, 135; P8.14, 85 Wong, G.W.H., P14.56, 142 Wonodi, I., P7.12, 76; P14.30, 135 Woo, J.I., P4.4, 61 Wood, B., P2.43, 50 Wood, J., O4.5, 19; P3.20, 57; P13.21, 126; P14.51, 141; P14.52, 141; P14.57, 142 Wood, J.A., P1.23, 37 Wood, N.W., O3.3, 16 Wormley, B., P14.27, 134 Wormley, B.K., P14.17, 132 Wray, N.R., O1.2, 11; P13.12, 123; P14.2, 128 Wright, H.H., O5.5, 21; P10.4, 99; P10.5, 100; P10.8, 101; P10.23, 105; P10.24, 105 Wu, J.Y., P14.5, 128 Wu, S.N., P7.3, 74 Wuisman-Frerker, M., O5.7, 22 Wuyan, C., P4.3, 61 Xiao, Zeping, P7.5, 75 Xie, W., O4.5, 19; P14.53, 141 Xing, Y.L., P7.11, 76 Xiong, L., P7.16, 77 Xu, K., P5.12, 66 Xu, M., P14.62, 143 Xu, W., O9.7, 30 Xu, X., P12.8, 114; P12.10, 115; P12.11, 115 Xu, Y., P7.6, 75 Xu, Y.F., P7.3, 74; P7.7, 75; P7.10, 76 Xuei, X., P9.24, 96 Yamada, K., P14.24, 133; P8.2, 82; P8.3, 82 Yamada, M., P11.7, 109; P11.13, 111 Yamada, N., P16.15, 152 Yamaki, L.H., O6.7, 24; P9.21, 95 Yamanouchi, Y., P7.25, 80; P7.26, 80 Yamauchi, T., P16.15, 152 Yan, J., P12.24, 119; P17.13, 156 Yang, C., P17.13, 156 Yang, J.D., P7.9, 76 Yang, M.S., P7.9, 76; P9.1, 90 Yang, Y., P7.5, 75; P7.6, 75 Yang, Y.F., P7.10, 76 Yates, P., P8.4, 82; P14.15, 131 Yeh, Y., P1.16, 35 Yirmiya, N., P9.3, 90 Yokom, D.W., P17.8, 155 Yolken, R., O4.5, 19 Yolken, R.H., P14.18, 132 Yoneda, H., P7.17, 78 Yoneda, M., P2.9, 41 Yoshikawa, T., P8.2, 82; P8.3, 82; P14.24, 133

Author Index Youn, J.C., P4.4, 61 Youn, R., O1.3, 11 Young, A., P2.33, 48 Yousem, D.M., P4.14, 63 Yu, L., P7.9, 76; P7.11, 76; P14.15, 131; P14.36, 137 Yu, Y., P14.62, 143 Yun, Sun, P4.3, 61 Yurov, Y.B., P6.4, 73; P9.2, 90; P16.9, 151; P16.10, 151 Yuxi, Pan, P4.3, 61 Zai, G., P5.24, 69; P8.14, 85; P14.28, 135 Zammit, S., O4.2, 18; O8.7, 28; P8.10, 84 Zanardini, R., P18.19, 162

Zandi, P., P2.40, 50; P2.42, 50 Zatz, M., P11.15, 111 Zekanowski, C., P4.9, 62 Zerah, G., P8.20, 86 Zhang, D., P7.5, 75 Zhang, F., P8.4, 82; P14.62, 143 Zhang, F.C., P9.1, 90 Zhang, X., O4.5, 19 Zhang, X.-N., P1.4, 32 Zhang, Y., P7.30, 81 Zhang, Z.H., P11.14, 111 Zhao, J., O3.5, 17; P4.11, 63; P7.9, 76 Zhao, S.M., P7.11, 76 Zhao, X.Z., P7.9, 76 Zheng, Y.L., P7.7, 75; P7.19, 78

179

Zheng, Z.J., P9.1, 90 Zhou, J., P7.6, 75; P7.10, 76; P7.19, 78 Zhu, S.M., O4.3, 18; P7.11, 76 Zibermann, M., P11.15, 111 Ziegler, A., P10.25, 106 Zill, P., P1.5, 32; P1.27, 38; P11.6, 109; P11.12, 110; P12.12, 115; P15.6, 145; P15.11, 146; P18.23, 163 Zinkstok, J., P3.30, 59 Zohar, A., O6.6, 24; P3.1, 51; P12.4, 113; P12.5, 114 Zubenko, G.S., O1.6, 12 Zubenko, W.N., O1.6, 12 Zucchelli, M., P10.25, 106 Zwanzger, P., P1.5, 32