Excess of allele1 for α3 subunit GABA receptor gene (GABRA3) in bipolar patients: a multicentric association study

Article metrics


The available data from preclinical and pharmacological studies on the role of gamma amino butyric acid (GABA) support the hypothesis that a dysfunction in brain GABAergic system activity contributes to the vulnerability to bipolar affective disorders (BPAD). Moreover, the localization of the α3 subunit GABA receptor GABRA3 gene on the Xq28, a region of interest in certain forms of bipolar illness, suggests that GABRA3 may be a candidate gene in BPAD. In the present study, we tested the genetic contribution of the GABRA3 dinucleotide polymorphism in a European multicentric case-control sample, matched for sex and ethnogeographical origin. Allele and genotype (in females) frequencies were compared in 185 BPAD patients and 370 controls. A significant increase of genotype 1–1 was observed in BPAD females compared to controls (P = 0.0004). Furthermore, when considering recessivity of allele 1 (females with genotype 1–1 and males carrying allele 1), results were even more significant (P = 0.00002). Our findings suggest that the GABRA3 polymorphism may confer susceptibility to or may be in linkage disequilibrium with another gene involved in the genetic etiology of BPAD.


The hypothesis that bipolar affective disorder (BPAD) is transmitted through a gene on the X-chromosome was first proposed by Rosanoff et al in 1935.1 Several linkage studies have investigated the region Xq26–28 in BPAD2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28 (see Table 1). Among them, the largest pedigree series reported a high Lod score (z) confirming linkage (z = 8.1 at θ = 0.15) to color blindness (CB) in 17 North American families.4 This finding was further supported by additional studies in different populations showing linkage to CB2,3,4,5,11,23 and glucose-6-phosphate dehydrogenase (G6PD).9,11,23 In more recent studies, genome-wide scans using highly polymorphic DNA markers in large numbers of pedigrees have provided conflicting evidence for X-linkage in BPAD. Negative linkages,26,28 and modest but not significant evidence for linkage to Xq26–28, may be indicative of loci of modest susceptibility.25,27 Conversely, conclusive evidence for linkage from a Finnish pedigree24 was reported with a maximum Lod score (Zmax) of 3.54 at θ = 0, for DXS994 at Xq25–q26. These controversial results support the notion that genetic heterogeneity and ethnic differences may be a possible source of divergent results.

Table 1 Linkage studies for Xq26–28 region in BPAD

Replication was reported in different populations, that makes the Xq26–X28 region particularly interesting for applying other methodologies. Because the two regions of positive linkage (Xq26 and Xq28) are separated by a genetic distance of 15–30 cM, there may be more than one susceptibility locus for BPAD in that area. More recent hypotheses of mode of inheritance of BPAD focus on complex heredity.29 Consequently, true positive linkage is difficult to replicate in disorders that are predisposed to by the combined action of several genes of minor or moderate effect.29 The allelic association strategy offers a powerful means of identifying such genes in realistically sized samples of unrelated cases by the study of polymorphisms within candidate genes and/or regions. To further explore the presence of a vulnerability gene for BPAD in the Xq28 region, we selected the dinucleotide repeat (CA) polymorphism of the α3 subunit of gamma-aminobutyric acid receptor gene (GABRA3), located near the 3′ end of the gene in the 8 intron for carrying out a case-control association study in a large European multicentric sample. A total of 555 subjects were recruited, including 185 BPAD patients, each matched with two controls (n = 370) for sex and ethnogeographical origin (see Table 2). To our knowledge, only three association studies have examined the role of GABRA3 in BPAD, and didn't report significant findings30,31,32 (Table 3). Three other genes have been investigated in the same region: FMR1 implicated in the fragile X syndrome in Xq27.333 reported negative findings, G6PD in Xq2834,35,36,37 reported positive results in two different populations. Finally, a recent preliminary study38 reported a trend of association between BPAD males and the SYBL1 gene (Xq28) involved in synaptic vesicle function.

Table 2 Contribution of each center to the recruitment of case controls sample
Table 3 Case control association studies for the Xq27–28 region in BPAD

Pathophysiological arguments for selecting GABRA3 as a candidate gene are summarized as follows: gamma amino-butyric acid (GABA) is the principal inhibitory aminoacid in the vertebrate brain. Emrich et al39 in 1980 formulated the so-called ‘GABA hypothesis of affective disorder’ in which valproic acid was presumed to correct a GABA deficit in BPAD. Since then, several lines of evidence from preclinical and clinical studies have lent support to the involvement of GABA in AD (see Massat et al40 for a review). GABA (type A) receptor is an ionotropic receptor with a hetero-oligomeric structure composed of five distinct classes α, β, γ, δ and ρ (divided into subunits) organized as a channel. These subunit types differ with respect to their sensitivity for GABA and recognition sites for drugs, that is relevant to the pharmacological differences observed between drugs, such as benzodiazepines (BZDs), which interact with GABAA receptors. Receptors that are composed of α3 subunits yield much greater responses to BZD than do receptors, which contain α1 or α2 subunits.41 Mutations are also known to alter the binding characteristics of the GABAA receptor. Especially for the α3 subunit, a single amino acid substitution increases by 10 the affinity of certain compounds for the BZD site.42

No significant deviation from HWE was observed in female BPAD patients and control populations. When considering centers separately, a heterozygote deficit was noted in the Bonn control sample (P = 0.008) and Brussels BPAD sample (P = 0.02).

The initial focus of our analysis was to detect differences between BPAD patients and controls in the distribution of genotypes, and alleles (see Table 4 for a description). Results are shown in Table 5 and Figure 1. The genotype distribution (after collapsing genotypes with a frequency less than 3%) showed an overall significant difference between BPAD and control females (P = 0.0024), with a level of significance of 0.03 for the genotype 1–1. When the presence or absence of genotype 1–1 was tested separately, significance increased (P = 0.0004, OR = 2.5 (1.5–4.3)). The relatively narrow confidence interval suggested precision in the estimate of the magnitude of the association. Comparison of allelic distribution (after collapsing allele 2 and allele 4 in the same group) between BPAD and matched controls showed an overrepresentation of allele 1 in BPAD patients (P = 0.002, OR = 1.8 (1.2–2.7)). In order to examine the allele 1 recessive hypothesis, we carried out an analysis that compared groups of BPAD patients and controls, according to genotype 1–1 homozygote females and allele 1 hemizygote males. Robust statistical differences emerged (P = 0.00002, OR = 2.5 (1.6–3.9)). When comparing the number of copies of allele 1 in BPAD females, we found that there was a significant difference between 1 and 2 copies (two copies of allele 1 were present in 78.8% of BPAD patients and 58.7% of controls, P = 0.0012 after bonferroni corrections). However, there was no statistical difference between 0 and 2 copies (two copies of allele 1 were present in 96.7% of BPAD patients and 93.6% of controls) and 0 and 1 copy (one copy was present in 88.9% of BPAD patients and 91.1% of controls): power was too low to detect significant results.

Table 4 Genotype and allelic distributions between BPAD patients and controls, by centers (A1 = 169 bp, A2 = 167 bp, A3 = 165 bp, A4 = 163 bp)
Table 5 Comparison of genotypes, genotype 1–1 distributions, and alleles, allele 1 recessive hypothesis at the GABRA3 locus, between bipolar patients (BPAD), and controlsConditional logistic regression (dependant variable: diagnosis)
Figure 1

Allele 1 recessivity hypothesis (%) in BPAD and controls: stratification by centers and by sex. Stratification analysis: (a) Strata = centers. Breslow Day (Homogeneity test) χ2 = 5.9, df 6, P = 0.43. (b) Strata = sex. Breslow Day (Homogeneity test) χ2 = 0.002, df 1, P = 0.97.

Effect modification of center and sex (tested separately) didn't show any significant difference (see Figure 1). The allelic distribution of our overall control population did not differ significantly from data in ‘Genome Database-http://www.gdb.org/’.

In this multicentric European study, a robust association was found in favor of allele 1 recessive hypothesis of a (CA)n polymorphism43 in GABRA3, in BPAD patients. The involvement of GABRA3 may be related to a direct effect of the polymorphism on the disease. Microsatellites may alter DNA structure, which has been implicated in the regulation of transcription and translation. Altered binding of transcription and translation factors may compromise RNA editing.44,45 Another possibility is that the association may reflect an indirect effect of another gene close to GABRA3, in linkage disequilibrium. Our results are not in line with some previous association studies testing the same GABRA3 (CA)n repeat polymorphism.30,31,32 However, samples used in previous studies didn't have a sufficient statistical power for detecting association, if we consider a minor or moderate effect of genetic variant at this polymorphism. Furthermore, our present investigation supports DNA studies that have reported positive association findings for G6PD enzyme deficiency (Xq28) in BPAD patients in two populations.34,35,36,37 In light of previous positive linkage findings in this region,2,3,4,5,7,9,11,23 together with the present results implicating the GABRA3 candidate gene in BPAD, we suggest further investigation of the GABRA3 (CA)n repeat in other populations, using family-based association methods, to further explore the genetics of BPAD. Additionally, it would be of interest to investigate other GABA receptor genetic markers, such as Snps in the vicinity of GABRE (GABA receptor epsilon) and GABRβ4 (GABA β4 subunit) in Xq28 region.

Finally, the positive association found in this multicentric European study between BPAD patients and allele 1 of GABRA3 (CA)n repeat polymorphism, underlines the interest of investigating further this genetic marker in controlled pharmacogenetic studies evaluating the treatment response of BPAD patients to anticonvulsants with primary activity at the GABA system.



The present sample has been recruited within the Biomedical European (BIOMED 1) European Collaborative Project on Affective Disorders (ECPAD) (European Community Grant CT92–1217, project leader J Mendlewicz) and BIOMED 2 (European Collaborative Study on Molecular Genetics in Affective Disorders, Contract No: BMH4-CT-97-2307). This network was established within the framework of the European Commission. The objectives and the detailed methodology of the project were described previously.46 A total of seven clinical centers participated in the present study (Brussels, Sofia, Zagreb, Bonn, Athens, Milan and Edinburgh, see Table 2). All subjects were interviewed using standard diagnostic interviews, such as the Schedule for Affective Disorders and Schizophrenia–Lifetime Version (SADS-LA47) and the schedule for the Clinical Assessment of Neuropsychiatry (SCAN48). One of the two diagnostic interviews was used for all patients and controls recruited for the project. The decision to adopt two instruments arose out of different research experience within individual research teams having their own preferences and expertise with the two interviews. Comparability between SADS and SCAN instruments is inferred from data published by the European Science Foundation ESF showing good concordance between the two instruments.49 Patients met the diagnosis of BPAD according to RDC, DSM-III-R and DSM-IV classification systems.

Within the control group, subjects with a positive personal or familial history of psychiatric disorder were excluded. Informed consent was obtained from patients and controls.


Standard procedures were used for DNA extraction. The GABRA3 (CA)n polymorphism was identified using hot-start polymerase chain reaction (PCR). The primers employed were those described by Hicks et al.43 Cross-examination of DNAs ensured identical allele assignments between centers. The 169-bp dinucleotide was named allele 1, 167-bp allele 2, the 165-bp allele 3 and the 163-bp allele 4 in our analyses.

Statistical analysis

The overall sample, as well as the samples of all centers separately, were tested for Hardy–Weinberg equilibrium (HWE) in female patients and control populations by using the exact HW test and the GENEPOP, 3.d program-updated version of GENEPOP.50 Matching in the design was utilized to control for potential confounding factors, sex and especially ethnogeographical origin, a complex nominal variable which represents a wide and undefinable range of environmental and genetic factors difficult to quantify and thus to control by other means. Because of the matched design with two controls individually matched to cases, conditional logistic regression was used to assess the association between BPAD and some predictors, and to derive odds ratios and 95% confidence intervals after adjustment for potential confounding factors. However, the limitation of matching is the inability to evaluate the effect of matching factor on risk of the outcome. To evaluate the possible effect modification of center, a stratified analysis was performed (match ignoring) and odds ratios (OR) and confidence intervals (95% CI) were calculated. The Breslow Day test was applied to test the homogeneity of the stratum-specific estimates (center and sex for strata and allele 1 recessivity as exposition). Bonferroni corrections were applied when necessary.


  1. 1

    Rosanoff AJ, Handy LH, Plesset IR . The etiology of manic depressive syndromes with special reference to their occurrence in twins Am J Psychiatry 1935 91: 725–740

  2. 2

    Winokur G, Tanna VL . Possible role of X-linked dominant factor in manic depressive disease Dis Nerv Syst 1969 30: 89–94

  3. 3

    Reich T, Clayton PJ, Winokur G . Family history studies: V. The genetics of mania Am J Psychiatry 1969 125: 1358–1369

  4. 4

    Mendlewicz J, Fleiss JL . Linkage studies with X-chromosome markers in bipolar (manic-depressive) and unipolar (depressive) illnesses Biol Psychiatry 1974 9: 261–294

  5. 5

    Baron M . Linkage between an X-chromosome marker (deutan color blindness) and bipolar affective illness. Occurrence in the family of a lithium carbonate-responsive schizo-affective proband Arch Gen Psychiatry 1977 34: 721–725

  6. 6

    Gershon ES, Targum SD, Matthysse S, Bunney WE Jr . Color blindness not closely linked to bipolar illness. Report of a new pedigree series Arch Gen Psychiatry 1979 36: 1423–1430

  7. 7

    Mendlewicz J, Linkowski P, Guroff JJ, Van Praag HM . Color blindness linkage to bipolar manic-depressive illness. New evidence Arch Gen Psychiatry 1979 36: 1442–1447

  8. 8

    Reading CM . X-linked dominant manic-depressive illness: linkage with Xg blood group, red-green color blindness, and vitamin B12 deficiency Orthomolecular Psychiatry 1969 8: 69–77

  9. 9

    Mendlewicz J, Linkowski P, Wilmotte J . Linkage between glucose-6-phosphate dehydrogenase deficiency and manic-depressive psychosis Br J Psychiatry 1980 137: 337–342

  10. 10

    Del Zompo M, Bocchetta A, Goldin LR, Corsini GU . Linkage between X-chromosome markers and manic-depressive illness. Two Sardinian pedigrees Acta Psychiatr Scand 1984 70: 282–287

  11. 11

    Baron M, Risch N, Hamburger R, Mandel B, Kushner S, Newman M et al. Genetic linkage between X-chromosome markers and bipolar affective illness Nature 1987 326: 289–292

  12. 12

    Mendlewicz J, Simon P, Sevy S, Charon F, Brocas H, Legros S et al. Polymorphic DNA marker on X chromosome and manic depression Lancet 1987 1: 1230–1232

  13. 13

    Berrettini WH, Goldin LR, Gelernter J, Gejman PV, Gershon ES, Detera-Wadleigh S . X-chromosome markers and manic-depressive illness. Rejection of linkage to Xq28 in nine bipolar pedigrees Arch Gen Psychiatry 1990 47: 366–373

  14. 14

    Gejman PV, Detera-Wadleigh S, Martinez MM, Berrettini WH, Goldin LR, Gelernter J et al. Manic depressive illness not linked to factor IX region in an independent series of pedigrees Genomics 1990 8: 648–655

  15. 15

    Neiswanger K, Slaugenhaupt SA, Hughes HB, Frank E, Frankel DR, McCarty MJ et al. Evidence against close linkage of unipolar affective illness to human chromosome 11p markers HRAS1 and INS and chromosome Xq marker DXS52 Biol Psychiatry 1990 28: 63–72

  16. 16

    Mendlewicz J, Sevy S, Charon F, Legros S . Manic depressive illness and X chromosome Lancet 1991 338: 1213

  17. 17

    Lucotte G, Landoulsi A, Berriche S, David F, Babron MC . Manic depressive illness is linked to factor IX in a French pedigree Ann Genet 1992 35: 93–95

  18. 18

    Gill M, Castle D, Duggan C . Cosegregation of Christmas disease and major affective disorder in a pedigree Br J Psychiatry 1992 160: 112–114

  19. 19

    Bredbacka PE, Pekkarinen P, Peltonen L, Lönngvist J . Bipolar disorder in an extended pedigree with a segregation pattern compatible with X-linked transmission: exclusion of the previously reported linkage to F9 Psychiatr Genet 1993 3: 79–87

  20. 20

    Baron M, Freimer NF, Risch N, Lerer B, Alexander JR, Straub RE et al. Diminished support for linkage between manic depressive illness and X-chromosome markers in three Israeli pedigrees Nat Genet 1993 3: 49–55

  21. 21

    Jeffries FM, Reiss AL, Brown WT, Meyers DA, Glicksman AC, Bandyopadhyay S . Bipolar spectrum disorder and fragile X syndrome: a family study Biol Psychiatry 1993 33: 213–216

  22. 22

    De bruyn A, Raeymaekers P, Mendelbaum K, Sandkuijl LA, Raes G, Delvenne V et al. Linkage analysis of bipolar illness with X-chromosome DNA markers: a susceptibility gene in Xq27–q28 cannot be excluded Am J Med Genet 1994 54: 411–419

  23. 23

    Mendelbaum K, Sevy S, Souery D, Papadimitriou GN, De Bruyn A, Raeymaekers P et al. Manic-depressive illness and linkage reanalysis in the Xq27–Xq28 region of chromosome X Neuropsychobiology 1995 31: 58–63

  24. 24

    Pekkarinen P, Terwilliger J, Bredbacka PE, Lonnqvist J, Peltonen L . Evidence of a predisposing locus to bipolar disorder on Xq24–q27.1 in an extended Finnish pedigree Genome Res 1995 5: 105–115

  25. 25

    Stine OC, McMahon FJ, Chen L, Xu J, Meyers DA, Mackinnon DF, Simpson S 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

  26. 26

    Smyth C, Kalsi G, Brynjolfsson J, O'Neill J, Curtis D, Rifkin L et al. Test of Xq26.3–28 linkage in bipolar and unipolar affective disorder in families selected for absence of male to male transmission Br J Psychiatry 1997 171: 578–581

  27. 27

    Liu J, Aita VM, Wang Z, Knowles JA, Terwilliger J, Matis TC et al. Progress in a genome-side search for genetic factors for bipolar disorder Am J Med Genet 1997 74: 590

  28. 28

    Vallada HP, Vasques L, Curtis D, Zatz M, Kirov G, Lauriano V et al. Linkage analysis between bipolar affective disorder and markers on chromosome X Psychiatr Genet 1998 8: 183–186

  29. 29

    Comings D . Polygenic inheritance in psychiatric disorders. In: Blum K, Noble EP (eds) Handbook of Psychiatric Genetics CRC Press: Boca Raton 1997 pp 235–260

  30. 30

    Walsh C, Hicks A, Sham P, Castle D, Hunt N, Clements A et al. GABAA receptor subunit genes as candidate genes for bipolar affective disorder—an association analysis Psych Genet 1992 2: 239–247

  31. 31

    Puertollano R, Visedo G, Saiz-Ruiz J, Llinares C, Fernandez-Piqueras J . Lack of association between manic-depressive illness and a highly polymorphic marker from GABRA3 gene Am J Med Genet 1995 60: 434–435

  32. 32

    Oruc L, Furac I, Croux C, Jakovljevic M, Kracun I, Folnegovic V et al. Association study between bipolar disorder and candidate genes involved in dopamine-serotonin metabolism and GABAergic neurotransmission: a preliminary report Psychiatr Genet 1996 6: 213–217

  33. 33

    Craddock N, Daniels J, McGuffin P, Owen M . Variation at the fragile X locus does not influence susceptibility to bipolar disorder Am J Med Genet 1994 54: 141–143

  34. 34

    Bocchetta A, Piccardi MP, Del Zompo M . Is bipolar disorder linked to Xq28? Nat Genet 1994 6: 224

  35. 35

    Bocchetta A, Piccardi MP, Martinelli V, Quesada G, Del Zompo M . Maternal inheritance of manic depression in hemizygotes for the G6PD-Mediterranean mutation. Indirect evidence for Xq28 transmission in Sardinia Psychiatr Genet 1999 9: 63–68

  36. 36

    Kalsi G, McQuillin A, Lauwrence J, Curtis D, Gurling HMD . Positive allelic association between bipolar affective disorder and G6PD polymorphisms on chromosome Xq28 Mol Psychatry 1999 4: S73

  37. 37

    Bocchetta A, Piccardi M, Palmas MA, Pedditzi M, Chillotti C, Del Zompo M . Family-based association study between bipolar disorder and the G6PD-Mediterranean mutation Am J Med Genet 2000 4: 494

  38. 38

    Saito T, Guan A, Nishimatsu Y, Lachman H . Analysis of the pseudoautosomal X-linked gene SYBL1 in bipolar affective disorder: description of a new candidate allele for psychiatric disorders Am J Med Gen 2000 96: 317–323

  39. 39

    Emrich HM, Zerssen DV, Kissling W, Möller HJ, Windorfer A . Effect of sodium valproate on mania: the GABA-hypothesis of affective disorders Arch Psychiatr Nervenkr 1980 229: 1–16

  40. 40

    Massat I, Souery D, Papadimitriou GN, Mendlewicz J . The GABAergic hypothesis of mood disorders. In: Soares JC, Gershon S (eds) Bipolar Disorders Marcel Dekker: New York 2000 pp 143–165

  41. 41

    Luddens H, Pritchett DB, Kohler M, Killisch I, Keinanen K, Monyer H et al. Cerebellar GABAA receptor selective for a behavioural alcohol antagonist Nature 1990 346: 648–651

  42. 42

    Hommer DW, Skolnick P, Paul SM . The benzodiazepine GABA receptor complex and anxiety. In: Meltzer HY (ed). Psychopharmacology: The Third Generation of Progress Raven Press: New York 1987

  43. 43

    Hicks AA, Johnson KJ, Barnard EA, Darlison MG . Dinucleotide repeat polymorphism in the human X-linked GABAA receptor alpha 3-subunit subunit gene Nucleic Acids Res 1991 19: 4016

  44. 44

    Epplen JT, Kyas A, Maueler W . Genomic simple repetitive DNAs are targets for differential binding of nuclear proteins FEBS Lett 1996 389: 92–95

  45. 45

    Comings DE . Polygenic inheritance and micro/minisatellites Mol Psychiatry 1998 3: 21–31

  46. 46

    Souery D, Lipp O, Serretti A, Mahieu B, Rivelli SK, Cavallini C et al. European Collaborative Project on Affective Disorders: interactions between genetic and psychosocial vulnerability factors Psychiatr Genet 1998 8: 197–205

  47. 47

    Endicott J, Spitzer RL . A diagnostic interview: the schedule for affective disorders and schizophrenia Arch Gen Psych 1978 35: 837–862

  48. 48

    Wing JK, Babor T, Brugha T, Burke J, Cooper JE, Giel R et al. SCAN: Schedules for Clinical Assessment in Neuropsychiatry Arch Gen Psychiatry 1990 47: 589–593

  49. 49

    Farmer AE, Cosyns P, Leboyer M, Maier W, Mors O, Sargeant M et al. A SCAN–SADS comparison study of psychotic subjects and their first degree relatives Eur Arch Psychiatry Clin Neurosci 1993 242: 352–357

  50. 50

    Raymond M, Rousset F . Genepop (ver1.2): a population genetic software for exact test and ecuniscism J Hered 1995 95: 248–249

Download references


This work was supported by the Association for Mental Health Research (AESM), the European Community Biomed Grant (Grant No. CT 92–1217), the National Fund for Scientific Research (NFSR), and the Fund for Scientific Research Flanders-Belgium (FWO).

Author information

Correspondence to I Massat.

Rights and permissions

Reprints and Permissions

About this article


  • GABRA3
  • association study
  • candidate gene
  • chromosome X
  • bipolar disorder

Further reading