ORIGINAL RESEARCH ARTICLE

(2002) 7, 594 603 2002 Nature Publishing Group All rights reserved 1359-4184/02 $25.00 www.nature.com/mp ORIGINAL RESEARCH ARTICLE A genome screen of 13 bipolar affective disorder pedigrees provides evidence for susceptibility loci on chromosome 3 as well as chromosomes 9, 13 and 19 RF Badenhop 1,2, MJ Moses 1, A Scimone 1, PB Mitchell 2, KR Ewen-White 3, A Rosso 1, JA Donald 4, LJ Adams 1 and PR Schofield 1 1 Garvan Institute of Medical Research, Sydney, Australia; 2 School of Psychiatry, University of New South Wales and Mood Disorders Unit, Prince of Wales Hospital, Sydney, Australia; 3 Australian Genome Research Facility, Melbourne, Australia; 4 School of Biological Sciences, Macquarie University, Sydney, Australia Bipolar affective disorder is a severe mood disorder that afflicts approximately 1% of the population worldwide. Twin and adoption studies have indicated that genetic factors contribute to the disorder and while many chromosomal regions have been implicated, no susceptibility genes have been identified. We undertook a combined analysis of 10 cm genome screen data from a single large bipolar affective disorder pedigree, for which we have previously reported linkage to chromosome 13q14 (Badenhop et al, 2001) and 12 pedigrees independently screened using the same 400 microsatellite markers. This 13-pedigree cohort consisted of 231 individuals, including 69 affected members. Two-point LOD score analysis was carried out under heterogeneity for three diagnostic and four genetic models. Non-parametric multipoint analysis was carried out on regions of interest. Two-point heterogeneity LOD scores (HLODs) greater than 1.5 were obtained for 11 markers across the genome, with HLODs greater than 2.0 obtained for four of these markers. The strongest evidence for linkage was at 3q25 26 with a genome-wide maximum score of 2.49 at D3S1279. Six markers across a 50 cm region at 3q25 26 gave HLODs greater than 1.5, with three of these markers producing scores greater than 2.0. Multipoint analysis indicated a 20 cm peak between markers D3S1569 and D3S1614 with a maximum NPL of 2.8 (P = 0.004). Three other chromosomal regions yielded evidence for linkage: 9q31 q33, 13q14 and 19q12 q13. The regions on chromosomes 3q and 13q have previously been implicated in other bipolar and schizophrenia studies. In addition, several individual pedigrees gave LOD scores greater than 1.5 for previously reported bipolar susceptibility loci on chromosomes 18p11, 18q12, 22q11 and 8p22 23. (2002) 7, 594 603. doi:10.1038/sj.mp.4001025 Keywords: bipolar disorder; manic depression; genome screen; genetic linkage Introduction Bipolar affective disorder (MIM 125480) affects approximately 1% of the population worldwide 1 and is the sixth highest cause of disability globally. 2 The etiology of bipolar disorder remains unknown, but family, twin and adoption studies have indicated that genetic factors contribute significantly to the disorder. 3,4 Family studies indicate that there is a seven-fold increase in the risk to first-degree relatives while twin studies show that there is an average four-fold increase in risk for monozygotic vs dizygotic twins. 3,4 The mode of inheritance remains unclear but it is likely that the disorder is oligogenic and heterogeneous with multiple loci of modest effect. Correspondence: Professor PR Schofield, Garvan Institute of Medical Research, 384 Victoria Street, Sydney 2010, Australia. E- mail: p.schofield garvan.org.au Received 12 July 2001; revised 16 October 2001; accepted 16 October 2001 While initial genetic linkage studies of bipolar disorder were fraught with inconsistencies and failures to replicate, more recently the accumulation of multiple studies on large pedigree cohorts has led to reproducible identification of several susceptibility loci. These include regions on chromosomes 4, 12q, 18, 21q and 22q. 5 10 We have previously reported evidence for linkage to chromosome 4q35 11 and 13q14. 12 Both of these findings were generated from genome screen analyses of single large pedigrees. Several groups have since reported support for the 4q35 locus 13 15 in other bipolar cohorts and the 13q14 locus in studies of both bipolar disorder 16 18 and schizophrenia 19 21 pedigrees. We have performed a genome screen analysis on a further 12 medium to large bipolar affective pedigrees. As these 12 pedigrees were analysed with the same marker set as the pedigree for which we reported linkage to 13q14, 12 we have combined the analysis of the genome screen data. In the present paper, we present the results of the combined analysis of these 13 bipolar affective disorder pedigrees.

Subjects and methods Family ascertainment The pedigrees used in this study were ascertained as part of an ongoing bipolar genetics study via the Mood Disorders Unit, Prince of Wales Hospital/University of New South Wales, Sydney, Australia. Medium to large multigenerational pedigrees with illness over at least two generations and containing a minimum of three affected individuals (at least two of whom were diagnosed with bipolar I) were recruited. Families were Caucasian and almost entirely of British or Irish descent. The families were assessed using the Diagnostic Interview for Genetic Studies (DIGS) 22 and best-estimate Research Diagnostic Criteria (RDC) diagnoses were made after independent evaluation of interviews and medical records. All marrying-in individuals were routinely questioned about any family history of psychiatric illness to ensure unilateral descent of bipolar disorder in the pedigrees. Individuals who participated in the study provided appropriate informed written consent. The study was approved by the Ethics Committee of the University of New South Wales. The pedigree cohort in the present study was comprised of 13 pedigrees made up of 293 members of whom 231 were available for analysis, including 69 affected individuals (Figure 1) of which Ped 04 (40 individuals including six affecteds) has previously been reported. 12 In total there were 33 individuals diagnosed with bipolar I (BPI), seven with schizoaffective disorder manic type (SZ/MA), four with bipolar II (BPII) and 25 with recurrent unipolar depression (UP). The average number of individuals per pedigree was 17.7 with the average number of affected individuals per pedigree being 5.3. Genotyping DNA was extracted from whole blood using a standard salting-out method. All samples were genotyped with 400 microsatellite markers from the ABI PRISM Linkage Mapping Set Version 2 (Applied Biosystems, Foster City, CA, USA). The markers in this mapping set had an average heterozygosity of 0.79 and were spaced at approximately 10 cm intervals across the genome. Genotyping was carried out at the Australian Genome Research Facility, Melbourne, Australia (www.agrf. org.au). Some additional microsatellite markers were analysed on chromosome 3q25. The markers at 3q25 were identified using the chromosome 3 summary map from The Genetic Location DataBase (LDB). Primer sequences, expected-size information and known allele frequencies for most markers were obtained from The Genome Database (GDB) (http://gdbwww.gdb.org/). The markers and their GDB accession numbers are D3S1309 (GDB:188504), D3S1593 (GDB:199848), D3S3606 (GDB:609402), D3S1555 (GDB:199255) and D3S3637 (GDB:610608). PCR of the additional markers was carried out in a 15 l volume containing 60 ng DNA, 250 M dntps, 0.33 M of each primer (the forward primer labelled with 6-FAM fluorescent dye), 2.5 mm MgCl 2,1 PCR buffer and 0.6 units AmpliTaq Gold polymerase Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 (Applied Biosystems). PCR reactions were carried out on a Hybaid OmniGene thermal cycler (Hybaid, Middlesex, UK) using the following protocol; 95 C for 12 min, 30 cycles of 95 C for 30 s, either 58 C or60 C for 30 s, 72 C for 30 s, 72 C for 5 min. One and a half microlitres of the PCR product were mixed with loading dye (2.5 l formamide, 0.5 l of50mgml 1 blue dextran : 50 mm EDTA, 0.5 l GeneScan500 Tamra size standard). Samples were loaded onto a 4% polyacrylamide gel and electrophoresis was carried out on an ABI PRISM 377 DNA Sequencer. Products were detected using the GeneScan Analysis program, version 3.1 and alleles were assigned using the Genotyper program, version 2.5 (Applied Biosystems). Marker typing incompatibilities were detected using the Ped- Check program, version 1.0. 23 Statistical analysis Two-point LOD score analysis was carried out using the ANALYZE package. 24 LOD score analysis was carried out under heterogeneity to take into account the possibility of interfamilial heterogeneity. Three disease models were used in the analysis. In disease model I, individuals diagnosed with either BPI or SZ/MA were classified as affected and all other family members were classified as unaffected. In the disease model II, individuals diagnosed with either BPI, SZ/MA or BPII were classified as affected and all other family members were considered unaffected. In the disease model III, individuals diagnosed with either BPI, SZ/MA, BPII or UP were classified as affected and all other family members were considered unaffected. Four liability classes (class 1, 20 years; class 2, 20 29 years; class 3, 30 39 years and class 4, 40 years) were used in the analysis with maximum age-specific penetrance levels of either 60% or 90%. In the 90% model, liability classes were defined with penetrances of 0.18, 0.45, 0.68 and 0.9; those in the 60% model were defined with penetrances of 0.12, 0.30, 0.45 and 0.6. The data were analyzed under both dominant and recessive inheritance models. The disease-allele frequency was set at 0.035 for the dominant model and 0.2 for the recessive model, and a phenocopy rate of 5% was used in all analyses. Marker allele frequencies were calculated from the data using the DOWNFREQ program from the ANALYZE package 24 using all individuals. Recombination fractions were converted from centimorgan distances provided on the ABI/CHLC integrated map (http://lpg.nci.nih.gov/html-chlc/chlc Maps.html). Simulation analyses using SLINK and MSIM 25,26 were performed on these pedigrees to aid in the interpretation of results. Simulation analyses were performed under both dominant and recessive models at both 60% and 90% maximum age-dependent penetrance and for the three different disease models using a 6 allele marker and 1000 replicates. In order to determine the minimum LOD score required to select markers for further analysis, we determined the maximum expected LOD scores of the cohort under the assump- 595

Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 596 Figure 1 Pedigrees used in the genome screen. There were 231 members available for genotypic analysis including 69 affected individuals. tion of no linkage. Under the assumption of no linkage the maximum expected LOD score ranged from 0.9 to 1.8 (depending on the model) with an average of 1.3, and no power to detect a LOD score greater than 3.0. In comparison, the maximum expected LOD score for the entire pedigree cohort under the assumption of homogeneity ranged from 1.8 to 8.5 with an average of 4.7, and 40% power to detect a LOD score greater than 3.0. Under the assumption of 50% heterogeneity the maximum expected LOD score ranged from 1.4 to 5.3 with an average of 2.9, and 5% power to detect a LOD score greater than 3.0. Based on the results of the simulations, markers giving LOD scores greater than 1.5 from the analysis of the genome data were subjected to an initial review. These selected markers were discarded from any further analysis if they had flanking

markers which gave negative LOD scores, or the positive LOD scores obtained were only under a single disease or genetic model. After this initial review of the data three regions were selected for analysis using multipoint analysis in order to define any potential regions of linkage. Non-parametric multipoint analyses were performed using GENEHUNTER. 27 Results The results of the two-point score analysis under heterogeneity are shown in Figure 2. Two-point heterogeneity LOD scores (HLODs) greater than 1.5 were obtained for 11 markers across the genome, with HLOD scores greater than 2.0 obtained for four of these markers (Table 1). The strongest evidence for linkage was on chromosome 3q25 q26. A genome-wide maximum HLOD score of 2.49 ( = 0, = 1) was obtained at marker D3S1279 under a 60% dominant model, for the disease model I. A HLOD of 1.5 ( = 0, = 0.23) was obtained at the adjacent marker, D3S1614, under a 90% dominant model and the disease model I. Five additional markers at 3q25 q26 were genotyped in the 13 pedigrees (Figure 3). HLODs greater than 1.0 were obtained for four of these markers spanning 50 cm. Of the additional markers genotyped, HLODs greater than 2.0 were obtained for D3S1593 (2.3, = 0, = 0.23) and D3S1555 (2.5, = 0, = 0.63) and HLODs greater than 1.5 were obtained for D3S3637 (1.9, = 0, = 0.58) and D3S1309 (1.6, = 0.12, = 1.0). The positive HLODs for the chromosome 3 markers were largely obtained under the narrower disease models I or II and a dominant model. While HLODs for the chromosome 3 markers were also positive under diagnostic model III, no marker gave a score greater than 1.5 under this model. As the results for marker D3S1614 were at an alpha value of 0.23, LOD scores for individual pedigrees were inspected at this marker. A single large pedigree (pedigree 05) was found to contribute to the positive score at D3S1614 with a maximum LOD score of 2.38 ( = 0). Positive HLODs of all other markers were obtained at higher alpha values with several families contributing to the positive scores at these markers. Non-parametric multipoint analysis of the ABI PRISM Linkage Mapping Set Version 2 markers was carried out in order to define the potential disease region. From this analysis, a maximum NPL score of 2.8 (P = 0.004) was obtained between D3S1569 and D3S1614 under a diagnostic model II (Figure 4a). Multipoint analysis using the five additional markers genotyped in this cohort did not further refine the region. Table 1 shows other regions on chromosomes 9, 13 and 19 which yielded evidence for linkage. There are two regions on chromosome 9 for which HLODs greater than 1.5 were obtained; D9S285 (1.64, = 0, = 0.4) at 9p22 and markers D9S1776 (1.74, = 0.16, = 1) and D9S1690 (1.60, = 0.0, = 0.81) at 9q31 q33 (Table 1). At 9p22 there were several markers which gave HLODs greater than 1.0 across a 30 cm region, including mark- Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 ers D9S157 and D9S171. All positive HLODs at 9p22 were obtained under disease model III. Similarly multiple markers at 9q31 q33 gave HLODs greater than 1.0 across a 50 cm region, including D9S287 and D9S290. The maximum HLODs for the 9q31 q33 markers were all obtained under disease model III, although small positive scores were also seen for all these markers under the narrower disease models I and II. The regions on chromosome 9 are separated by approximately 60 cm. Non-parametric multipoint analysis of chromosome 9 shows a peak between D9S1690 and D9S1677 at 9q31 q33 with a maximum NPL of 2.5 (P = 0.02) under diagnostic model III (Figure 4b). There were two markers at 13p11 q14 which gave HLODs greater than 1.5; D13S153 (2.29, = 0, = 0.26) at 13q14 under a 90% recessive model and disease model III and D13S175 (1.9, = 0, = 0.74) at 13p11 under a 60% recessive model and disease model I. The positive score at D13S153, obtained at an alpha value of 0.26, can be attributed to the single large pedigree (pedigree 04) for which we have previously reported linkage to this locus. 12 Several other markers spanning a 80 cm region on 13q gave HLODs greater than 1.0 in the cohort, including D13S218 (1.14, = 0, = 0.21), D13S263 (1.15, = 0.14, = 1.0), D13S156 (1.28, = 0, = 0.37) and D13S173 (1.43, = 0.18, = 1.0). The maximum HLODs for the 13q markers were all obtained under a recessive model and disease model III, although small positive scores were also seen for all these markers under the narrower disease models I and II. Inspection of the LOD scores for individual pedigrees at the 13q markers revealed that pedigree 04 was the only pedigree contributing to the positive scores at markers D13S218 and D13S156, while there were several positive families for markers D13S175 and D13S263. Analysis of the chromosome 13 markers in the cohort excluding pedigree 04, gave HLODs greater than 1.0 for two markers: D13S175 (1.15, = 0.0, = 0.63) and D13S173 (1.04, = 0.2, = 1.0). There are several markers spanning chromosome 19 which gave HLODs greater than 1.0. A marker flanking D19S216 (1.55, = 0.18, = 1.0), D19S209, gave a HLOD of 1.12 also under a recessive model and disease model III while three markers: D19S220, D19S420 and D19S571 some 65 cm from D19S216 also gave HLODs greater than 1.0 but under a dominant model and disease model III. Small positive HLOD scores were also seen for the chromosome 19 markers under the narrower disease models I and II. Non-parametric multipoint analysis of chromosome 19 gave a peak of 4.55 (P = 0.0002) between markers D19S414 and D19S220 at 19q12 q13 under diagnostic model III. There is a smaller positive peak (NPL = 1.33, P = 0.09) observed between D19S216 and D19S84 at 19p13 (Figure 4c). While the NPL score at 19q12 q13 is significant, the HLODs are modest and therefore inconsistent with the significance indicated by the multipoint analysis. Multipoint analysis is not as robust to the presence of errors as two-point analysis, and can result in an inaccurate estimation of the location of the disease-prediposing gene. 28 Therefore we have taken a more con- 597

Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 598 Figure 2 The genome screen consisted of 400 microsatellite markers spaced at approximately 10 cm intervals across the genome. Each point is the maximum heterogeneity LOD score (HLOD) obtained at each marker under all models tested (see Methods). Markers that produced HLODs greater than 1.5 are indicated by their chromosomal location. servative approach in the interpretation of these results and consider the evidence for linkage to 19q based on the HLODs to be weak in comparison to other loci in our study. HLODs greater than 1.5 were also obtained for markers D7S507, D8S514 and DXS986 (Table 1). However the absence of supporting data from flanking markers or positive scores under the multiple models tested, indicated that these are unlikely to be true positive results.

Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 Table 1 Two-point HLODs greater than 1.5 in all 13 pedigrees 599 Chromosome Marker HLOD a Theta ( ) Alpha ( ) Genetic model b Diagnostic model c 3q25 D3S1279 2.49 0.0 1.0 60% dominant BPI, SZ/MA 3q26 D3S1614 1.50 0.0 0.23 90% dominant BPI, SZ/MA 7p21.1 D7S507 1.94 0.1 0.53 90% dominant BPI, BPII, SZ/MA, UP 8q24 D8S514 2.08 0.0 0.66 90% dominant BPI, SZ/MA 9p22 D9S285 1.64 0.0 0.40 90% recessive BPI, BPII, SZ/MA, UP 9q32 q33 D9S1776 1.74 0.16 1.0 90% recessive BPI, BPII, SZ/MA 9q31 D9S1690 1.60 0.0 0.81 60% dominant BPI, BPII, SZ/MA 13p11 D13S175 1.92 0.0 0.74 60% recessive BPI, SZ/MA 13q14 D13S153 2.29 0.0 0.26 90% recessive BPI, BPII, SZ/MA, UP 19p13 D19S216 1.55 0.0 1.0 90% recessive BPI, BPII, SZ/MA, UP Xq21.1 DXS986 2.48 0.14 1.0 90% X-linked BPI, BPII, SZ/MA, UP a HLOD; maximum two-point HLOD score under heterogeneity. b Either 60% or 90% maximum age-dependent penetrance. c BPI: Bipolar I; BPII: Bipolar II; SZ/MA: schizoaffective disorder manic type; UP: unipolar depression. Figure 3 Chromosome 3q markers and the corresponding maximum heterogeneity LOD (HLOD) scores obtained in the 13 pedigrees. For each marker, the genetic position and the diagnostic and genetic models that yielded the maximum score are indicated. In addition to the findings in the entire cohort, there were individual pedigrees which gave LOD scores greater than 1.5 at loci previously reported in other genome screens of bipolar cohorts (Table 2). These included regions on chromosomes 18, 22 and 8. 7,8,10,29 Pedigree 17 had a genome-wide maximum LOD score of 1.82 at D18S53 (18p11) under a 90% dominant gen- etic model and diagnostic model III while pedigree 22 gave a genome-wide maximum LOD score of 1.68 at D18S1102 (18q12) under a 90% dominant genetic and diagnostic model III. Pedigree 03 gave a genome-wide maximum LOD score of 2.0 at D22S420 (22q11) under a 90% recessive genetic model and diagnostic model III. Pedigrees 21 and 26 both had their genome-wide

Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 600 Figure 4 Nonparametric multipoint analysis in the 13 pedigrees of chromosome 3 under diagnostic model II, chromosome 9 under diagnostic model III and chromosome 19 under diagnostic model III.

Table 2 Two-point HLODs greater than 1.5 in individual pedigrees Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 601 Pedigree Chromosome Marker HLOD Theta ( ) Genetic model a Diagnostic b Previous reports 17 18p11 D18S53 1.82 0 90% dominant BPI, BPII, SZ/MA, UP Berretini et al, 1994 8 22 18q12 D18S1102 1.68 0 90% dominant BPI, BPII, SZ/MA, UP Stine et al, 1995 7 3 22q11 D22S420 2.0 0 90% recessive BPI, BPII, SZ/MA, UP Kelsoe et al, 2001 10 21 8p22 23 D8S264 1.7 0 90% recessive BPI, BPII, SZ/MA, UP Foroud et al, 1998 28 26 8p22 23 D8S264 1.84 0 60% dominant BPI, BPII, SZ/MA, UP Foroud et al, 1998 28 a Either 60% or 90% maximum age-dependent penetrance. b BPI: Bipolar I; BPII: Bipolar II; SZ/MA: schizoaffective disorder manic type; UP: unipolar depression. maximum LODs at D8S264 on 8p22 23, with pedigree 21 having a maximum LOD of 1.7 under a 90% recessive genetic model and diagnostic model III, and pedigree 26 having a maximum of 1.84 under the 60% dominant genetic model and diagnostic model III. The LOD scores obtained at 18p11, 18q12, 22q11 and 8p22 23 in the above pedigrees were the genome-wide maximum scores obtained for that pedigree. There were, however, several other markers across the genome for which LOD scores greater than 1.0 were obtained in each pedigree. This indicates that there is not a single major locus segregating for bipolar disorder within an individual pedigree but rather multiple susceptibility loci of smaller effect. Discussion We have examined genome screen data for 400 markers in 13 medium to large bipolar pedigrees. While we did not identify any regions of significant linkage, there were several regions of interest for which suggestive linkage, according to the criteria established by Lander and Kruglyak, 30 is indicated. The genome-wide maximum was a HLOD of 2.49 at D3S1279 on chromosome 3q25 q26 under a narrow diagnostic model and an autosomal dominant inheritance model. Other potential regions included loci on chromosomes 9q, 13q and 19q and support for existing bipolar loci on chromosomes 8p, 18p, 18q and 22q. From a previous genome screen analysis of a single large pedigree (pedigree 04), 12 LOD scores greater than 1.0 were obtained for four markers on chromosome 3q25 q26, including D3S1292, D3S1278, D3S1279 and D3S1569. In the present study, we provide additional support for this locus in a larger pedigree cohort and with additional markers. At chromosome 3q25 q26 there were six markers across 50 cm which gave HLODs greater than 1.5, with three markers giving LOD scores greater than 2.0. Nonparametric multipoint analysis indicated a peak between markers D3S1569 and D3S1614, a region of approximately 20 cm. The region on 3q25 q26 has also been implicated in a recent study by Kelsoe et al. 10 In a genome screen of 20 North American pedigrees, Kelsoe et al 10 identified two regions on chromosome 3, at 3p21 and 3q27, with LOD scores greater than 2.0. The region implicated by Kelsoe et al 10 overlaps the region identified in our analysis. The maximum LOD scores reported by Kelsoe et al 10 were found under a narrow diagnostic model, including only BPI individuals as affected, with a dominant mode of inheritance, as was found in our analysis. Chromosome 13q has been implicated in multiple studies of both bipolar disorder 16 18 and schizophrenia 19 21 cohorts. In our study, the positive scores for the 13q markers appear to be largely due to a single pedigree (pedigree 04) in which we have previously reported linkage. 12 Analysis of pedigree 04 gave a maximum two-point LOD score of 2.91 at D13S153 and a maximum NPL score of 4.09 (P = 0.008) between markers D13S1272 and D13S153. There is also some evidence for linkage to this region in the other pedigrees analysed, although this is at D13S175 (13p11), approximately 40 cm from the region implicated in pedigree 04. It is of interest that there are now several overlapping loci implicated in both bipolar disorder and schizophrenia, including loci on 22q, 10,31 18p, 32 10p 29 as well as 13q. 16 21 The overlap in loci indicated by linkage studies raises the possibility that there are common susceptibility genes for these disorders, suggesting that the disorders may be related with heterogeneous underlying etiologies due to different combinations of common genetic and environmental factors. Few studies have implicated chromosome 9q31 q33 in bipolar disorder. The exception is Detera-Wadleigh et al 33 who reported evidence for increased allele sharing at D9S302 (P = 0.004) on chromosome 9q33 in the 97 pedigree cohort from the National Institute of Mental Health Genetics Initiative. This region is of interest as it contains the candidate gene, dopamine betahydroxylase, an enzyme involved in dopamine neurotransmission. There have been no reports of linkage to chromosome 19 which could represent a novel locus for susceptibility to bipolar disorder. Clearly the number of possible susceptibility loci identified to date provide strong evidence that bipolar disorder is multigenic and subject to genetic heterogeneity. If there is significant heterogeneity, this could explain the failure of genome screens of large cohorts of smaller sized pedigrees 13,34 to identify loci that meet the criteria for an initial finding of significant linkage.

602 Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 This may be due to the presence of multiple independently segregating susceptibility loci. Conversely, genome screens of large multigenerational pedigrees have provided the most compelling evidence for linkage of bipolar disorder to a locus. 5,6,9,11 The most likely explanation is that within a single large pedigree, where the locus is of modest effect, the other susceptibility loci could be largely homozygous such that one is able to observe evidence for linkage for that single segregating locus. In this context, it is of interest that there were several individual pedigrees in our cohort which provided support for linkage to known bipolar susceptibility loci at 18p11, 8 18q12, 7 22q11 10 and 8p22 23. 29 In each of these individual pedigrees, the LOD scores obtained at these known loci were the maximum obtained for that pedigree across the genome. In our analysis of a genome screen of 13 medium to large bipolar pedigrees the strongest evidence for linkage was at 3q25 q26, with three other chromosomal regions, 9q31 q33, 13q14 and 19q12 13 yielding evidence for linkage and individual pedigrees providing support for existing bipolar loci at 18p11, 18q12, 22q11 and 8p22 23. The regions on 3q25 q26, 9q31 q33 and 13q14 have previously been implicated in other bipolar disorder genome screens, while the region on 19q12 13 may represent a novel bipolar susceptibility locus. Although several of these regions satisfy the criteria for suggestive linkage, none are significant. The strength of the results, however, must be interpreted relative to other regions of the genome for which no evidence for linkage was found and also in the context of continuing support for loci identified in genome screen of other bipolar cohorts. Acknowledgments We thank Jim McBride for development of a query database to aid in analysis of the linkage data, John Barlow and staff of the Australian Genome Research Facility, Melbourne, Australia and the family members who participated in this study. This study was supported by the National Health and Medical Research Council (Australia) via grants to the Garvan Institute of Medical Research (Block grant 993050), the Mood Disorders Unit, Prince of Wales Hospital (Program Grant 993208) and the Network of Brain Research into Mental Disorders (Unit Grant 983302). References 1 Weissman MM, Bland RC, Canino GJ, Faravelli C, Greenwald S, Hwu HG et al. Cross-national epidemiology of major depression and bipolar disorder. J Am Med Assoc 1996; 276: 293 299. 2 Murray CJT, Lopez AD. Global mortality, disability and the contribution of risk factors: Global Burden of Disease Study. Lancet 1997; 349: 1436 1442. 3 Kelsoe JR. The genetics of bipolar disorder. Psychiatr Ann 1997; 27: 285 292. 4 MacKinnon DE, Jamison KR, DePaulo JR. Genetics of manic depressive illness. Ann Rev Neurosci 1997; 20: 355 373. 5 Blackwood DH, He L, Morris SW, McLean A, Whitton C, Thomson M et al. A locus for bipolar disorder on chromosome 4p. Nat Genet 1996; 12: 427 430. 6 Craddock N, Owen M, Burge S, Kurina B, Thomas P, McGuffin P. Familial cosegregation of major affective disorder and Darier s Disease (keratosis follicularis). Br J Psychiatry 1994; 164: 355 358. 7 Stine OC, Xu J, Kosela R, McMahon FJ, Geschwind M, Friddle C et al. Evidence for linkage of bipolar disorder to chromosome 18 with a parent-of-origin effect. Am J Hum Genet 1995; 57: 1384. 8 Berretini WH, Ferraro TN, Goldin LR et al. Pericentromeric chromosome 18 DNA markers and manic-depressive illness: evidence for a susceptibility gene. Proc Natl Acad Sci 1994; 91: 5918 5921. 9 Straub RE, Leher T, Luo Y, Loth JE, Shao W, Sharpe L et al. A possible vulnerability locus for bipolar affective disorder on chromosome 21q22.3. Nat Genet 1994; 8: 291 296. 10 Kelsoe JR, Spence MA, Loetscher E, Foguet M, Sadovnick AD, Remick RA et al. A genome survey indicates a possible susceptibility locus for bipolar disorder on chromosome 22. Proc Natl Acad Sci 2001; 98: 585 590. 11 Adams LJ, Mitchell PB, Fielder SL, Rosso A, Donald JA, Schofield PR. A susceptibility locus for bipolar affective disorder on chromosome 4q35. Am J Hum Genet 1998; 62: 1084 1091. 12 Badenhop RF, Moses MJ, Scimone A, Mitchell PB, Ewen KR, Rosso A et al. A genome screen of a large bipolar affective disorder pedigree supports evidence for a susceptibility locus on chromosome 13q. Mol Psychiatry 2001; 6: 396 403. 13 Friddle C, Koskela R, Ranade K, Hebert J, Cargill M, Clark CD et al. Full-genome scan for linkage in 50 families segregating the bipolar affective disease phenotype. Am J Hum Genet 2000; 66: 205 215. 14 Bennett P, Bort S, Seguardo R, Middle F, Jones I, Heron J et al. The Wellcome Trust UK-Irish bipolar sib-pair genome screen: complete first stage data and second stage progress report. Am J Med Genet 2000; 96: 467. 15 Ginns EI, St Jean P, Philibert RA, Galdzicka M, Damschroder-Williams P, Thiel B et al. A genome-wide search for chromosomal loci linked to mental health wellness in relatives at high risk for bipolar affective disorder among the Old Order Amish. Proc Natl Acad Sci 1998; 95: 15531 15536. 16 Detera-Wadleigh SD, Badner JA, Berrettini WH, Yoshikawa T, Goldin LR, Turner G et al. A high-density genome scan detects evidence for a bipolar-disorder susceptibility locus on 13q32 and other potential loci on 1q32 and 18p11.2. Proc Nat Acad Sci USA 1999; 96: 5604 5609. 17 Ginns EI, Ott J, Egeland JA, Allen CR, Fann CSI, Pauls DL et al. A genome-wide search for chromosomal loci linked to bipolar affective disorder. Nat Genet 1996; 12: 413 435. 18 Stine OC, McMahon FJ, Chen L, Xu J, Meyers DA, MacKinnon DF et al. Initial genome screen for bipolar disorder in the NIMH genetics initiative pedigrees: chromosomes 2, 11, 13, 14, and X. Am J Med Genet 1997; 74: 263 269. 19 Brzustowicz LM, Honer WG, Chow EW, Little D, Hogan J, Hodgkinson K et al. Linkage of familial schizophrenia to chromosome 13q32. Am J Hum Genet 1999; 65: 1096 1103. 20 Lin MW, Curtis D, Williams N, Arranz M, Nanko S, Collier D et al. Suggestive evidence for linkage of schizophrenia to markers on chromosome 13q14.1 Psychiatr Genet 1995; 5: 117 126. 21 Blouin JL, Dombroski BA, Nath SK, Lasseter VK, Wolyniec PS, Nestadt G et al. Schizophrenia susceptibility loci on chromosome 13q32 and 8p21. Nat Genet 1998; 20: 70 73. 22 Nurnberger JI Jr, Blehar MC, Kaufman CA, York-Cooler C, Simpson SG, Harkavy-Friedman J et al. Diagnostic interview for genetic studies: rationale, unique features, and training: NIMH Genetics Initiative. Arch Gen Psychiatry 1994; 51: 849 859. 23 O Connell JR, Weeks DE. PedCheck: a program for identifying marker typing incompatibilities in linkage analysis. Am J Hum Genet 1998; 63: 259 266. 24 Terwilliger J. ANALYZE 1996 ftp://linkage.cpmc.columbia.edu/ software/analyze 25 Ott J. Computer-simulation methods in human linkage analysis. Proc Natl Acad Sci USA 1989; 86: 4175 4178. 26 Weeks DE, Ott J, Lathrop GM. SLINK: a general simulation program for linkage analysis. Am J Hum Genet 1990; 47: A204 (supplement). 27 Kruglyak L, Daly MJ, Reeve-Daly MP, Lander ES. Parametric and nonparametric linkage analysis: a unified mulipoint approach. Am J Hum Genet 1996; 58: 1347 1363. 28 Risch N, Giuffra L. Model misspecification and multipoint analysis. Hum Hered 1992; 42: 77 92.

29 Foroud T, Castelluccio PF, Koller DL, Edenberg HJ, Miller M, Bowman E et al. Suggestive evidence of a locus on chromosome 10p using the NIMH genetics initiative bipolar affective disorder pedigrees. Am J Med Genet 2000; 96: 18 23. 30 Lander E, Kruglyak L. Genetic dissection of complex traits: guidelines for interpreting and reporting linkage results. Nat Genet 1995; 11: 241 247. 31 Pulver AE, Karayiorgou M, Wolyniec PS, Lasseter VK, Kasch L, Nestadt G et al. Sequential strategy to identify a susceptibility gene for schizophrenia: report of potential linkage on chromosome 22q12 q13.1. Am J Med Genet 1994; 54: 36 43. Bipolar disorder susceptibility loci on chromosomes 3, 9, 13 and 19 32 Berrettini WH. Are schizophrenic and bipolar disorders related? A review of family and molecular studies. Biol Psychiatry 2000; 48: 531 538. 33 Detera-Wadleigh SD, Badner JA, Yoshikawa T, Sanders AR, Goldin LR, Turner G et al. Initital genome scan of the NIMH genetics initiative bipolar pedigrees: chromosomes 4, 7, 9, 18, 19, 20, and 21q. Am J Med Genet 1997; 74: 254 262. 34 Nurnberger JI, DePaulo JR, Gershon ES, Reich T, Blehar MC, Edenberg HJ et al. Genomic survey of bipolar illness in the NIMH genetics initiative pedigrees: a preliminary report. Am J Med Genet 1997; 74: 227 237. 603