Interaction networks of lithium and valproate molecular targets reveal a striking enrichment of apoptosis functional clusters and neurotrophin signaling.

The overall neurobiological mechanisms by which lithium and valproate stabilize mood in bipolar disorder patients have yet to be fully defined. The therapeutic efficacy and dissimilar chemical structures of these medications suggest that they perturb both shared and disparate cellular processes. To investigate key pathways and functional clusters involved in the global action of lithium and valproate, we generated interaction networks formed by well-supported drug targets. Striking functional similarities emerged. Intersecting nodes in lithium and valproate networks highlighted a strong enrichment of apoptosis clusters and neurotrophin signaling. Other enriched pathways included MAPK, ErbB, insulin, VEGF, Wnt and long-term potentiation indicating a widespread effect of both drugs on diverse signaling systems. MAPK1/3 and AKT1/2 were the most preponderant nodes across pathways suggesting a central role in mediating pathway interactions. The convergence of biological responses unveils a functional signature for lithium and valproate that could be key modulators of their therapeutic efficacy.

A systems approach to the biology of mood disorders through network analysis of candidate genes.

Meta analysis of association data of mood disorders has shown evidence for the role of particular genes in genetic risk. Integration of association data from meta analysis with differential expression data in brains of mood disorder patients could heighten the level of support for specific genes. To identify molecular mechanisms that may be disrupted in disease, a systems approach that involves analysis of biological networks created by these selected genes was employed.Interaction networks of hierarchical groupings of selected genes were generated using the Michigan Molecular Interactions (MiMI) software. Large networks were deconvoluted into subclusters of core complexes by using a community clustering program, GLay. Network nodes were functionally annotated in DAVID Bioinformatics Resource to identify enriched pathways and functional clusters. MAPK and beta adrenergic receptor signaling pathways were significantly enriched in the ANK3 and CACNA1C network. The PBRM1 network bolstered the enrichment of chromatin remodeling and transcription regulation functional clusters. Lowering the stringency for inclusion of other genes in network seeds increased network complexity and expanded the recruitment of enriched pathways to include signaling by neurotransmitter and hormone receptors, neurotrophin, ErbB and the cell cycle. We present a strategy to interrogate mechanisms in the cellular system that might be perturbed in disease. Network analysis of meta analysis- generated candidate genes that exhibited differential expression in mood disorder brains identified signaling pathways and functional clusters that may be involved in genetic risk for mood disorders.

Two variants in Ankyrin 3 (ANK3) are independent genetic risk factors for bipolar disorder.

Two recent reports have highlighted ANK3 as a susceptibility gene for bipolar disorder (BD). We first reported association between BD and the ANK3 marker rs9804190 in a genome-wide association study (GWAS) of two independent samples (Baum et al., 2008). Subsequently, a meta-analysis of GWAS data based on samples from the US and the UK reported association with a different ANK3 marker, rs10994336 (Ferreira et al., 2008). The markers lie about 340 kb apart in the gene. Here, we test both markers in additional samples and characterize the contribution of each marker to BD risk. Our previously reported findings at rs9804190, which had been based on DNA pooling, were confirmed by individual genotyping in the National Institute of Mental Health (NIMH) waves 1-4 (P=0.05; odds ratio (OR)=1.24) and German (P=0.0006; OR=1.34) samples. This association was replicated in an independent US sample known as NIMH wave 5 (466 cases, 212 controls; P=0.017; OR=1.38). A random-effects meta-analysis of all three samples was significant (P=3 x 10(-6); OR=1.32), with no heterogeneity. Individual genotyping of rs10994336 revealed a significant association in the German sample (P=0.0001; OR=1.70), and similar ORs in the NIMH 1-4 and NIMH 5 samples that were not significant at the P<0.05 level. Meta-analysis of all three samples supported an association with rs10994336 (P=1.7 x 10(-5); OR=1.54), again with no heterogeneity. There was little linkage disequilibrium between the two markers. Further analysis suggested that each marker contributed independently to BD, with no significant marker x marker interaction. Our findings strongly support ANK3 as a BD susceptibility gene and suggest true allelic heterogeneity.

Detection, simultaneous display and direct sequencing of multiple nuclear hormone receptor genes using bilaterally targeted RNA fingerprinting.

We have developed a PCR-based method to detect, display and directly sequence multiple members of the nuclear hormone receptor (NHR) gene family. Our approach employs the basic concepts of RNA fingerprinting (Welsh et al. (1992) Nucleic Acids Res. 20, 4965-4970; Stone, B. and Wharton, W. (1994) Nucleic Acids Res. 22, 2612-2618) and differential display PCR (Liang, P. and Pardee, A.B. (1992) Science 257, 967-971), with modifications. In contrast to the previous methods, two conserved regions within the gene family were targeted to derive primers for PCR amplification. One of the conserved sites was used to deduce primers for cDNA synthesis. We believe that this strategy led to increased specificity. The use of degenerate primers with low redundancy in both reverse transcription and PCR steps also contributed to enhanced signal-to-noise ratio. The ability to directly sequence the amplified fragments constitutes a vast improvement over the previous methods. This method permitted the successful identification and simultaneous display of six different NHR genes, which included the previously unreported rat homolog of COUP-TFI and a recently described orphan receptor. We believe that this approach provides a convenient and rapid screening method for detecting and characterizing members of a gene family.

A linkage study of bipolar illness.

BACKGROUND: Although genetic epidemiological studies of bipolar (BP) illness are consistent with a heritable component, inherited risk factors remain unknown. The goal of the present study is to describe the localization of BP susceptibility loci through linkage strategies, including a genome-wide search. METHODS: A linkage study of 22 BP families has been performed. These BP families include almost 400 persons, 173 of whom have been diagnosed as having BP I, schizoaffective, BP II with major depression, or recurrent unipolar illness. Using an autosomal dominant disease model with 85% or 50% age-dependent penetrance, and a recessive model with 85% penetrance, linkage analyses were performed assuming a narrow (BP and schizoaffective) or a broad (BP, schizoaffective, or unipolar) definition of the BP spectrum. Affected sibling pairs and affected pedigree member analyses were performed when positive lod scores were observed in multiple pedigrees. The present article describes linkage analysis of 310 DNA markers on chromosomes 1, 5p, 6, 8, 10q, 11q, and 12 to 18. RESULTS: None of the loci examined disclosed compelling evidence for linkage using lod score analyses. Model-independent analysis by multilocus affected pedigree member method in the pericentromeric chromosome 18 region disclosed statistically significant evidence (P < .0001) for a BP susceptibility gene in this region. Multilocus analysis by affected sibling pair method also disclosed evidence for linkage (P < .00008). CONCLUSIONS: Our results imply that a BP susceptibility gene exists near the centromere of chromosome 18. Confirmation of this finding (by independent investigators studying different pedigrees) has been published, suggesting that a valid BP disease linkage may have been discovered.

Refinement of regions with allelic loss on chromosome 18p11.2 and 18q12.2 in esophageal squamous cell carcinoma.

Esophageal cancer ranks among the 10 most common cancers worldwide and is almost invariably fatal. The detailed genetic repertoire involved in esophageal carcinogenesis has not been defined. We have shown previously that the esophageal squamous cell carcinoma genome exhibits a frequent loss of heterozygosity (LOH) in the pericentromeric region of chromosome 18. To construct a fine deletion map, we screened 76 new samples composed of microdissected esophageal squamous cell carcinoma and matched morphologically normal epithelial cells using closely spaced markers. Maximal LOH frequency (54%) was displayed by D18S542 on 18p11.2. The pattern of LOH in selected patients indicated that the short region of overlap extends 3 cM on either side of D18S542. On the long arm of chromosome 18, the highest frequency of allelic loss (42%) was detected by D18S978 on 18q12.2-q21.1. This analysis revealed a short region of overlap of approximately 0.8 cM. These findings further implicate unreported tumor suppressor genes encoded by 18p11.2 and 18q12.2 in esophageal squamous cell carcinogenesis and they indicate a refinement of their map location.

Expression of the smoothelin gene is mediated by alternative promoters.

OBJECTIVE: Two major isoforms of smoothelin have been reported, a 59-kDa smoothelin-A in visceral smooth muscle cells and a 110-kDa smoothelin-B in vascular smooth muscle cells. The present study was undertaken to investigate the expression of these smoothelin isoforms in different smooth muscle tissues and to determine how they are generated. METHODS: Western blotting with a new, well-defined, smoothelin antibody was used to confirm the existence of two major smoothelin isoforms. Northern blotting, RT-PCR, primer extension and 5'RACE were applied to analyse the expression of these isoforms in human and mouse. Promoter reporter assays were carried out to establish the existence of a dual promoter system governing the expression pattern of the gene. RESULTS: Antibody C6G confirmed the existence of two smoothelin proteins. Northern blotting showed that in vascular tissues a larger smoothelin transcript is generated than in visceral tissue. The cDNA of this larger smoothelin-B was cloned. Computer analysis of the open reading frame suggests an alpha-helical structure of 130 amino acids at the amino terminus of smoothelin-B. The smoothelin gene was cloned and sequenced. It comprises about 25 kb and contains 21 exons. The translational start of smoothelin-B is located in exon 2, whereas transcription and translation of the previously described smoothelin-A starts inside exon 10. Smoothelin-A and -B were demonstrated to be generated by two physically separated promoters. Splice variants within the calponin homology domain at the 3' end of the gene were found for both isoforms. CONCLUSIONS: Two major smoothelin isoforms are generated from a single gene by a dual promoter system in a tissue specific manner. Further variation in the smoothelin proteins is achieved by alternative splicing in the calponin homology domain.

Splice variants of rat TR4 orphan receptor: differential expression of novel sequences in the 5'-untranslated region and C-terminal domain.

The use of rapid amplification of 5'-cDNA ends-PCR yielded two novel sequences for the rat orphan receptor, TR4, representing heterogeneity on the 5'-untranslated region. Genomic structure analysis revealed that the 5'-untranslated region of the longer messenger RNA fragment, rTR4-1, contained three exons, alpha, beta, and gamma. The skipping of exon gamma gave rise to rTR4-2, indicating that rTR4-1 and rTR4-2 are products of alternative splicing. We isolated another novel rat TR4 splice variant, rTR4-NS, which was found to diverge from rTR4-2 at codon 504. rTR4-NS contained an unspliced intronic sequence with in-frame codons for eight amino acids followed by a termination codon. The three TR4 messenger RNA variants were differentially expressed. rTR4-NS appeared to be a rare transcript found in limited areas of the brain. In situ hybridization detect prominent TR4 signals in brain areas known to be involved in stress response. In cerebellar granule cells, the rise in TR4 expression correlated with the progression of neuronal maturation. N-Methyl-D-aspartate treatment triggered a marked increase in TR4 expression. These results suggest a possible role for TR4 in neuronal differentiation.

Familial glucocorticoid resistance caused by a splice site deletion in the human glucocorticoid receptor gene.

The clinical syndrome of generalized, compensated glucocorticoid resistance is characterized by increased cortisol secretion without clinical evidence of hyper- or hypocortisolism, and manifestations of androgen and/or mineralocorticoid excess. This condition results from partial failure of the glucocorticoid receptor (GR) to modulate transcription of its target genes. We studied the molecular mechanisms of this syndrome in a Dutch kindred, whose affected members had hypercortisolism and approximately half of normal GRs, and whose proband was a young woman with manifestations of hyperandrogenism. Using the polymerase chain reaction to amplify and sequence each of the nine exons of the GR gene alpha, along with their 5'- and 3'-flanking regions, we identified a 4-base deletion at the 3'-boundary of exon 6 in one GR allele (delta 4), which removed a donor splice site in all three affected members studied. In contrast, the sequence of exon 6 in the two unaffected siblings was normal. A single nucleotide substitution causing an amino acid substitution in the amino terminal domain of the GR (asparagine to serine, codon 363) was also discovered in exon 2 of the other allele (G1220) in the proband, in one of her affected brothers and in her unaffected sister. The functional importance of this mutation was tested in a cotransfection study using the recombinant expression vector pRShGR-Ser363 and the glucocorticoid responsive vector mouse mammary tumor virus-chloramphenicol transferase. This amino acid substitution did not alter the function of the glucocorticoid receptor. Using reverse transcription-polymerase chain reaction we could only identify messenger RNA transcripts of the G1220-allele but not of the delta 4-allele in the affected members of this family who were heterozygous for the G1220 mutation. This deletion in the glucocorticoid receptor gene was, thus, associated with the expression of only one allele and a decrease of GR protein by 50% in affected members of this glucocorticoid resistant family. The mutation identified in exon 2 did not segregate with the disease and appears to be of no functional significance. The presence of the null allele was apparently compensated for by increased cortisol production at the expense of concurrent hyperandrogenism.

Mutation analysis of oculopharyngeal muscular dystrophy in Hispanic American families.

BACKGROUND: Oculopharyngeal muscular dystrophy (OPMD) is a late-onset autosomal dominant muscular dystrophy characterized by progressive ptosis, swallowing difficulties, and proximal limb weakness. Recently, the genetic basis of this disease has been characterized by mutations in the PABP2 gene that involve short expansions of the trinucleotide repeat GCG. OBJECTIVES: To independently confirm the presence and study the meiotic stability of the GCG expansion mutations in a distinct ethnic population with OPMD. SETTINGS: Hospital and university research laboratories in Los Angeles, Calif. SUBJECTS AND METHODS: Three unrelated families of Hispanic American descent were identified in whom OPMD was transmitted in an autosomal dominant pattern. All of these families can trace affected ancestors to the southwestern United States or to the bordering states of Mexico. In these families, 14 persons with OPMD were identified and studied. RESULTS: Our results confirm that in these families, expansion mutations characterized by a gain of 3 GCG repeats in the wild-type allele result in an abnormal nucleotide length of 9 GCG repeats in the PABP2 gene. In these families, these mutations are associated with the OPMD phenotype. The identical repeat mutation ([GCG]9) is found in all affected members of these unrelated families and shows relative meiotic stability. CONCLUSIONS: These results support and extend our study of haplotype analysis and suggest that a founder effect may have occurred for OPMD in this Hispanic American population.

Closing in on genes for manic-depressive illness and schizophrenia.

Advances in the human genetic map, and in genetic analysis of linkage and association in complex inheritance traits, have led to genetic progress in the major psychoses. For chromosome 6 in schizophrenia, and chromosomes 18 and 21 in manic-depressive illness, there are reports of linkage in several independent data sets. These are small effect genes, best detected with affected-relative-pair linkage methods. Association with candidate genes is an alternative strategy to uncovering susceptibility genes for these illnesses, but convincing associations remain to be demonstrated. New clinical and laboratory investigation methods are being developed. Testing every gene in the human genome for association with illness has recently been proposed (Risch and Merikangas 1996). This would require further progress in characterizing the genome and in automated large-scale genotyping. The best type of pedigree sampling for common disease studies, whether for linkage or association, is not yet established. An endophenotype hybrid strategy can combine genetic linkage, association, and pathophysiologic studies. As clinical molecular investigation methods advance, identification of disease susceptibility mutations and delineation of their pathophysiological roles may be expected.

A systematic search for a bipolar predisposing locus on chromosome 5.

Chromosome 5 markers spanning the pter to the qter were used to examine linkage to bipolar illness in 14 pedigrees. Twenty-four loci were examined in 237 individuals, of whom 69 were either bipolars or schizoaffectives. Marker genotypes were determined for each individual and lod scores were calculated under a dominant disease model with a maximum penetrance of 85%, a disease gene frequency of 0.015, a variable age of onset, and a phenocopy rate of 0.001. Under the assumption that bipolar illness is genetically homogeneous, the total lod scores from all pedigrees with each marker were uniformly lower than -2.0, suggesting the absence of linkage to disease at any of these loci. Multipoint analysis allowed exclusion of intervals between markers. When lod scores were calculated allowing for heterogeneity, no subset of linked families was found. These results indicate that in our pedigree series almost the entire mapped region of chromosome 5 can be excluded for linkage to bipolar illness.

Chromosomes 12 and 16 workshop.

Recent linkage results independently derived from a large French Canadian pedigree and Danish kindreds coupled with supportive data from other studies provide compelling evidence for a bipolar disorder susceptibility locus on chromosome 12q23-q24. The idea is further strengthened by the finding that Darier's disease, which maps to this region, has been shown to cosegregate with affective disorder in a family. This linkage finding, however, was not supported in other independent genome scans. On chromosome 16, bipolar families from Denmark exhibited suggestive linkage with D16S510, on 16p13. Multipoint nonparametric analysis on the NIMH Genetics Initiative bipolar pedigrees yielded increased allele sharing that maximized approximately 18 cM proximal to the latter locus. In contrast, evidence of linkage was not detected in other panels of bipolar families that were presented. At 16p13, a maximum multipoint lod score of 4 for a latent class-derived phenotype that has aspects of alcohol dependence was found in a genome scan of 105 families from the Collaborative Study of the Genetics of Alcoholism, identifying a potential vulnerability locus.

Maternal inheritance and chromosome 18 allele sharing in unilineal bipolar illness pedigrees.

We have replicated the observation of McMahon et al. [1995] that there is excess maternal transmission of illness in a series of previously described unilineal Bipolar manic-depressive illness extended pedigrees [Berrettini et al., 1991]. ("Transmission" is defined for any ill person in a pedigree when father or mother has a personal or immediate family history of major affective disorder.) We divided our pedigrees into exclusively maternal transmission (Mat) and mixed maternal-paternal transmission (in different pedigree branches) (Pat). Using affected sib-pair-analysis, linkage to a series of markers on chromosome 18p-cen was observed in the Pat but not the Mat pedigrees, with significantly greater identity by descent (IBD) at these markers in the Pat pedigrees. As compared with the pedigree series as a whole, the proportion of alleles IBD in the linkage region is much increased in the Pat pedigrees. As shown by Kruglyak and Lander [1995], as the sharing proportion of alleles in affected relative pairs increases, the number of such pairs needed to resolve the linkage region to a 1 cM interval becomes smaller. Genetic subdivision of an illness by clinical or pedigree configuration criteria may thus play an important role in discovery of disease susceptibility mutations.

Serotonin transporter (5-HTT) gene and bipolar affective disorder.

Interactions with antidepressants, as well as other biochemical evidence, implicate the serotonin transporter 5-HTT in the etiology of affective disorders. However, genetic studies have produced conflicting results concerning an association of 5-HTT with bipolar disorder. We examined a variable number tandem repeat in the regulatory region of this gene to investigate the possible contribution of 5-HTT to bipolar disorder susceptibility in a 22-pedigree series. By affected-sib-pair analysis and the transmission/disequilibrium test, we found no significant linkage or association of 5-HTT to bipolar disorder. During the course of this study, we adapted a PCR technique designed to amplify long templates to replicating long GC stretches with complex structure. We also refined the location of 5-HTT by radiation hybrid mapping, placing the locus between D17S1294 and SHGC11022 on 17q11.2.

Isolation of chromosome 18-specific brain transcripts as positional candidates for bipolar disorder.

Several studies have proposed the existence of susceptibility loci for bipolar disorder on chromosome 18. To identify possible candidate genes for this disease, we isolated brain-expressed transcripts by direct cDNA selection on chromosome 18-specific biotinylated cosmid clones. Longer cognate cDNA clones of the selected cDNAs were isolated from a normalized infant brain cDNA library. Physical mapping by PCR on a panel of somatic cell hybrids was conducted by the use of primers derived from partial sequences on either the 5' or 3' ends of the clones. In our initial analysis, 48 cDNA clones were found to be chromosome 18-specific, mapping to different subchromosomal regions. Sequence redundancy among these clones yielded 30 unique transcripts, five of which were represented in previously known genes. Further sequencing of the remaining 25 unique cDNA clones confirmed the absence of significant homology to known genes, indicating that these transcripts represented novel genes. Mapping with the use of a radiation hybrid panel positioned the brain cDNAs to within = 100 to 1100 kb from reference sequence tag sites (STSs) and assembled them into six high resolution linkage groups. The majority of the transcripts were found to cluster to discrete locations on 18p and 18q, previously hypothesized as susceptibility regions for bipolar disorder, identifying them as positional candidate genes.

Fine mapping supports previous linkage evidence for a bipolar disorder susceptibility locus on 13q32.

A region between D13S71 and D13S274 on 13q32 showed linkage to bipolar disorder (BP) based on a genome scan using markers with an average spacing of approximately 6 cM and an average heterozygosity of approximately 60% [Detera-Wadleigh et al., 1999: Proc Natl Acad Sci USA 96:5604-5609]. In an attempt to confirm this finding and achieve fine mapping of the susceptibility region, nine additional microsatellite markers with average heterozygosity of approximately 86%, located between D13S71 and D13S274, were typed in the same sample. The strongest linkage evidence was detected by multipoint linkage analysis (ASPEX program) around D13S779-D13S225 with maximum LOD score of 3.25 under Affection Status Model II (ASM II; P = 0.0000546). Data from additional nine markers resulted in a decrease of the 95% confidence interval of the linkage region. Association analyses with GASSOC TDT and ASPEX/sib_tdt detect potential linkage disequilibrium with several markers, including D13S280 (ASPEX TDT P = 0.0033, ASM I). These data generated using a higher marker density within the proposed susceptibility region strengthen the validity of our previous findings and suggest a finer localization of the susceptibility gene(s) on 13q32.

Genetic linkage mapping for a susceptibility locus to bipolar illness: chromosomes 2, 3, 4, 7, 9, 10p, 11p, 22, and Xpter.

We are conducting a genome search for a predisposing locus to bipolar (manic-depressive) illness by genotyping 21 moderate-sized pedigrees. We report linkage data derived from screening marker loci on chromosomes 2, 3, 4, 7, 9, 10p, 11p, 22, and the pseudoautosomal region at Xpter. To analyze for linkage, two-point marker to illness lod scores were calculated under a dominant model with either 85% or 50% maximum penetrance and a recessive model with 85% maximum penetrance, and two affection status models. Under the dominant high penetrance model the cumulative lod scores in the pedigree series were less than -2 at theta = 0.01 in 134 of 142 loci examined, indicating that if the disease is genetically homogeneous linkage could be excluded in these marker regions. Similar results were obtained using the other genetic models. Heterogeneity analysis was conducted when indicated, but no evidence for linkage was found. In the course of mapping we found a positive total lod score greater than +3 at the D7S78 locus at theta = 0.01 under a dominant, 50% penetrance model. The lod scores for additional markers within the D7S78 region failed to support the initial finding, implying that this was a spurious positive. Analysis with affected pedigree member method for COL1A2 and D7S78 showed no significance for linkage but for PLANH1, at the weighting functions f(p) = 1 and f(p) = 1/sqrt(p) borderline P values of 0.036 and 0.047 were obtained. We also detected new polymorphisms at the mineralocorticoid receptor (MLR) and calmodulin II (CALMII) genes. These genes were genetically mapped and under affection status model 2 and a dominant, high penetrance mode of transmission the lod scores of < -2 at theta = 0.01 were found.

Adrenocorticotropin receptor/melanocortin receptor-2 maps within a reported susceptibility region for bipolar illness on chromosome 18.

We have examined the possible linkage of adrenocorticotropin receptor/melanocortin receptor-2 (ACTHR/MC-2) to a reported putative susceptibility locus for bipolar illness (BP) in 20 affected pedigrees. Initially, allelic variants of the gene were identified by polymerase chain reaction-single stranded conformation polymorphism (PCR-SSCP) and the gene was genetically mapped using both the Centre d'Etudes du Polymorphisme Humain (CEPH) pedigrees and the BP pedigrees used in this study. We found that the ACTHR/MC-2 gene maps between D18S53 and D18S66. These loci span a region of chromosome 18 which, in a previous study [Berrettini et al.: Proc Natl Acad Sci USA 91:5918-5921, 1994) revealed a putative predisposing locus to BP through nonparametric methods of linkage analysis. Linkage of ACTHR/MC-2 to BP was not demonstrable under parametric and nonparametric methods of analyses, although affected sib-pair (ASP) method revealed an increase in allele sharing among ill individuals, P = 0.023. Since this receptor is within a potential linkage region, ACTHR/MC-2 could be considered a candidate gene for BP.

Initial genomic scan of the NIMH genetics initiative bipolar pedigrees: chromosomes 3, 5, 15, 16, 17, and 22.

As part of the four-center NIMH Genetics Initiative on Bipolar Disorder we carried out a genomic scan of chromosomes 3, 5, 15, 16,17, and 22. Genotyping was performed on a set of 540 DNAs from 97 families, enriched for affected relative pairs and parents where available. We report here the results of the initial 74 markers that have been typed on this set of DNAs. The average distance between markers (theta) was 12.3 cM. Nonparametric analysis of excess allele sharing among affected sibling pairs used the SIBPAL program of the S.A.G.E. package to test three hierarchical models of affected status. D16S2619 gave some evidence of linkage to bipolar disorder, with P = 0.006 for Model II (in which bipolar 1, bipolar 2 and schizoaffective-bipolar type individuals are considered affected). Nearby markers also showed increased allele sharing. A second interesting region was toward the telomere of chromosome 5q, where D5S1456 and nearby markers showed increased allele sharing; for D5S1456, P = 0.05, 0.015 and 0.008 as the models of affected status become more broad. MOD score analysis also supported the possible presence of a susceptibility locus in this region of chromosome 5. A pair of adjacent markers on chromosome 3, D3S2405 and D3S3038, showed a modest increased allele sharing in the broad model. Several isolated markers had excess allele sharing at the P < 0.05 level under a single model. D15S217 showed a MOD score of 2.37 (P < 0.025). Multipoint analysis flagged the region of chromosome 22 around D22S533 as the most interesting. Thus, several regions showed modest evidence for linkage to bipolar disorder in this initial genomic scan of these chromosomes, including broad regions near previous reports of possible linkage.