Original Research Article

Molecular Psychiatry (2004) 9, 603–608. doi:10.1038/sj.mp.4001461 Published online 16 December 2003

Association of SNPs and haplotypes in GABAA receptor bold italic beta2 gene with schizophrenia

W-S Lo1,*, C-F Lau1,2,*, Z Xuan1, C-F Chan1, G-Y Feng3, L He3, Z-C Cao4, H Liu4, Q-M Luan4 and H Xue1

  1. 1Department of Biochemistry and Applied Genomics Laboratory, The Hong Kong University of Science & Technology, Hong Kong, China
  2. 2PharmacoGenetics Limited, Hong Kong, China
  3. 3Bio-X Life Science Research Centre, Jiaotong University, Shanghai, China
  4. 4An Kang Hospital, Shangdong, China

Correspondence: H Xue, Department of Biochemistry, The Hong Kong University of Science & Technology, Clear Water Bay, Hong Kong, China. E-mail: hxue@ust.hk

*These authors contributed equally to this work

Received 28 May 2003; Revised 29 September 2003; Accepted 14 October 2003; Published online 16 December 2003.

Top

Abstract

Disturbances in GABAergic system have been observed in schizophrenics.1,2,3 In the present study, population association analysis was performed on 19 SNPs in the alpha1, beta2, italic gamma2, alt epsilon and pi subunit genes of GABAA receptor. Five SNPs in GABRB2, namely B2I7G1584T, rs1816071, rs194072, rs252944 and rs187269, were found to be significantly associated, and their haplotypes in linkage disequilibrium, with schizophrenia. This represents the first report on any disease association of SNPs in the human GABAA receptor genes, and focuses attention on the GABAergic hypothesis of schizophrenia etiology.3,4

Keywords:

complex disease, genotyping, population association, psychiatric disorder, SNP

Schizophrenia is a debilitating mental illness that affects 1% of the world's population, with a heritability of 0.70–0.85 attributable to complex inheritance.5 Despite extensive studies, its molecular causes have remained unidentified. Like many common diseases, it is believed to be multifactorial in origin, with both genetic and environmental contributions playing important roles in determining the symptoms. Several neurotransmitters, including dopamine, glutamate and serotonin are thought to be important to the disease, and significant associations with genes of the glutamate neurotransmitter and related systems have recently been reported.6,7 Although a defect in neurotransmission involving italic gamma-amino butyric acid (GABA) in schizophrenia was first proposed in the early 1970s, followed by extensive investigations on the GABAergic system in schizophrenic subjects,1,2,3,4 the lack of molecular genetic evidence has left the GABAergic system out of the roster of leading candidate genes for schizophrenia.

The type A italic gamma-amino butyric acid receptors (GABAA) are the major inhibitory receptors in the CNS.8,9 Among different subunit combinations, the alpha1/beta2/italic gamma2-containing heteropentamer is the dominant subtype in mammalian brains.10,11 To test the potential involvement of GABAA receptor genetics in the molecular etiology of schizophrenia, single-nucleotide polymorphisms (SNPs) and SNP-based haplotypes in the genes of GABAA receptors have been examined in the present study. In all, 19 SNPs in the GABAA alpha1, beta2, italic gamma2, alt epsilon and pi subunit genes were assessed for possible association with schizophrenia (Table 1). Five of them were discovered, and the others confirmed, through sequencing of PCR-amplified DNA samples. None of the nine SNPs identified in the alpha1, italic gamma2, alt epsilon and pi subunit genes was associated with schizophrenia (P>0.05; Table 1). On the other hand, five of the 10 SNPs in the beta2 subunit gene showed a significant association. Accordingly, a detailed investigation of the SNPs in the GABAA receptor beta2 subunit gene (GABRB2) was carried out.


Human GABRB2 is widely expressed in different brain regions, especially the cerebellum.12 It is located on chromosome 5 at 5q34 close to 5q33.2, which has been identified through genome-wide linkage analysis as one of the susceptibility loci for schizophrenia.13,14 In the present study, population association and linkage disequilibrium (LD) analyses were carried out on schizophrenic patients and unaffected subjects of Han Chinese origin for 10 SNPs within GABRB2, spanning approximately 13 kb from approx1 kb upstream of exon 6 to approx200 bp downstream of exon 9. Five of them, all located in introns 7 and 8, showed significant association with susceptibility to schizophrenia both in an initial screening, and upon further testing with increased sample sizes of schizophrenics and controls (Table 1). These five positive SNPs include one novel SNP at base 1584 of intron 7 within the sequence TTGTATC(G/T)ATTACAG, which is designated B2I7G1584T, and four previously reported SNPs, rs1816071, rs194072, rs252944 and rs187269 (dbSNP database). The genotype frequencies of all the five SNPs did not deviate from Hardy–Weinberg equilibrium for either the control or the patient group when tested with the program GENEPOP.15

At the allele level, the minor allele of each of the five positive SNPs was found to increase schizophrenia risk (Table 1). Strong associations with schizophrenia were found for rs252944 (P=0.0003; OR=2.34, 95% CI 1.46–3.76), rs187269 (P=0.0009; OR=1.93, 95% CI 1.31–4.85), rs1816071 (P=0.0008; OR=2.05, 95% CI 1.34–3.13) and rs194072 (P=0.0005; OR=2.50, 95% CI 1.48–4.23), and moderate association found for B2I7G1584T (P=0.0046; OR=2.02, 95% CI 1.23–3.31).

Genotype frequency differences between the schizophrenic and control groups were also analyzed. As shown in Table 1, highly significant overall genotype association with schizophrenia was observed for rs252944 (P=0.0018) and rs187269 (P=0.0031), and significant association observed for B2I7G1584T (P=0.0195), rs1816071 (P=0.0096) and rs194072 (P=0.0044).

The effect of each of the five positive SNPs on schizophrenia susceptibility was further assessed in Table 2. The risk of schizophrenia for individuals having one or two minor alleles (m/m or m/M) at any of the five SNP sites is estimated to be about two times higher (OR=1.89–2.61) than individuals with two copies of the major allele (M/M; P=0.0010–0.0079). A homozygous-minor-allele (m/m) genotype increases schizophrenia susceptibility over the majorallele-containing genotypes (M/m+M/M) in the cases of rs187269 (P=0.0046; OR=4.60), rs1816071 (P=0.0139; OR=2.86) and rs252944 (P=0.0267; OR=5.12). These results thus point to a significant enhancement of the risk of schizophrenia by the minor alleles. To compare the minor-allele homozygote and the heterozygote, association analysis was performed separately for m/m vs M/M and M/m vs M/M. The results suggest that the homozygous minor-allele genotypes tend to favor susceptibility to schizophrenia over the heterozygous genotypes, but an increased sample size would be needed before a firm conclusion can be drawn in this regard.


To further define the contribution of the five possible SNPs to schizophrenia susceptibility, haplotype associations with the disease were evaluated for all the possible two-SNP combinations of these five SNPs (Table 3). Among the two-loci haplotypes, rs1816071–rs187269, rs252944–rs187269 and rs194072–rs187269 yielded asymptotic P-values, derived from the model-free chi2 statistics,16,17 that were less than 0.00001 for chi2 distribution with four degrees of freedom (Table 3). None of the 10 000 replicates in the permutations computed by the EHPLUS program produced any empirical P-value that exceeded the asymptotic P-values from the observed data, thus pointing to a strong association between these three two-loci haplotypes and schizophrenia. The remaining two-SNP combinations also showed significant association with schizophrenia, with P-value ranging from 0.0003 to 0.0149. In addition, pair-wise linkage disequilibrium analysis (Table 3) showed that the five positive SNPs were highly linked with D' value ranging between 0.6645 and 0.9687, further supporting the haplotype associations.


The frequencies of the three-SNP haplotypes were likewise found to be in LD with schizophrenia, especially for the haplotype rs194072–rs252944–rs187269 (chi8df2=38.72, P<1.0 times 10-5). Upon the permutation of 10 000 replicates, the empirical P-value was smaller than the asymptotic P-value in this case. The empirical evidence thus further supported its disease association.

Four of the other five SNPs tested in GABRB2, namely rs2303055, rs967771, rs1644522 and rs1644436, were not significantly associated with the disease at either the allele or the genotype level (P>0.05; Table 1). However, the association of rs1816072 with the disease, while insignificant (P=0.1044) at the genotype level, was marginally significant (P=0.0266; OR=1.55, 95% CI 1.05–2.29; Table 1) at the allele level. Two haplotypes formed by the four negative SNPs were estimated and found not to be significantly associated with schizophrenia (rs2303055–rs967771, P=0.1741; rs1644522–rs1644436, P=0.1557).

The present study demonstrated that the five SNPs B2I7G1584T, rs1816071, rs194072, rs252944 and rs187269 in introns 7 and 8 of GABRB2 (Figure 1) were significantly associated with susceptibility to schizophrenia in Han Chinese. Our results suggest that the GABRB2 gene might be responsible for the previously reported positive linkage of chromosomal locus 5q33.2 with schizophrenia.13 So far, no other gene has been found in the neighboring sequences up to 351 960 bp upstream (GABRA6) and 927 375 bp downstream (LOC63920) of the block of five positive SNPs (Figure 1). Although crossethnic multicenter studies are needed before the observed associations can be generalized, the fact that our association analysis and the earlier linkage study were based on samples from different ethnic groups increases the likelihood that the observed association might be multiethnic in nature.

Figure 1.
Figure 1 - Unfortunately we are unable to provide accessible alternative text for this. If you require assistance to access this image, please contact help@nature.com or the author

Positions of 10 SNPs in GABRB2 gene. *Cluster ID numbers of SNPs are as designated by the dbSNP of NCBI except B2I7G1584T, which was discovered in this study.

Full figure and legend (14K)

The potential functional consequences of the minor alleles of the five positive SNPs, as well as interactions between GABRB2 and other plausible candidate genes in determining schizophrenia susceptibility,6,7,18,19 require further investigation in order to better understand the network of genetic factors underlying schizophrenia. It is known that there are at least two isoforms of GABRB2 expression products,20 and the possibility exists that the five positive SNPs may modulate transcriptional regulation, such as alternative splicing and mRNA expression, of GABRB2, thereby influencing schizophrenic phenotype expression,21 as has been suggested for the italic gamma2 subunit of GABAA receptor.22 While the minor alleles at the five loci increase disease susceptibility, they are neither necessary nor sufficient for the development of schizophrenia. The present study thus supports a multifactorial causation. By linking a GABAA receptor gene for the first time to schizophrenia, however, it has provided important evidence for the involvement of GABAergic processes in the etiological mechanism of the disease.

Top

Materials and methods

Study subjects

Peripheral blood samples were collected from Shanghai Mental Health Center with informed consent from unrelated Han Chinese subjects. Schizophrenics (mean age of onset 26.7 years; mean age 47.9) were diagnosed according to the criteria in the Diagnostic and Statistical Manual of Mental Disorders, fourth edition (DSM-IV).23 All patients displayed at least two of the following symptoms for a period of 1 month or more (except where successfully treated): delusion, hallucination, disorganized speech, grossly disorganized or catatonic behavior. Unrelated healthy Han Chinese (mean age 33.5 years) served as controls. All control individuals were interviewed to exclude any history of psychiatric disorders. All successfully genotyped results were included for disease association studies.

SNP discovery and genotyping

Genomic DNA was prepared from blood samples by using the DNA purification kit from Amersham, and diluted to 100 ng/ml. Primers were designed in an adjacent-primer-overlapping-200-base-pairs manner as described by Primer3 (Primer3 website). The specificity of each potential primer was checked through nBlast of the National Center for Biotechnology Information (Blastn Home Page). Only pairs of both forward and reverse primers with less than five hits to the human genome were accepted as specific primers. PCR was performed in a final volume of 20 mul containing 10 ng sample DNA, 75 nmol/l of each primer, 50 nmol/l of each dNTP, 2.5 mmol/l MgCl2 and 1 U Taq DNA polymerase (Amersham). PCR amplification consisted of denaturation at 94°C for 5 min, followed by 40 cycles of 1 min at 94°C, 1 min at the annealing temperature optimum for each pair of primers, 90 s at 72°C, followed by a final extension step at 72°C for 5 min. PCR products were resolved on 1.5% agarose gel electrophoresis, and stained with ethidium bromide to confirm the specificity of the products generated. PCR products were purified by Montage™ PCR96 Purification kit (Millipore). Each sequencing reaction contained 2 mul ABI PRISM® BigDye™ Terminator (v.2.0), 5 mul purified PCR products and 1 mmol/l of primer. Sequencing cycling conditions consisted of 1 min denaturation at 96°C, followed by 34 cycles of 96°C for 30 s, 50°C for 30 s and 60°C for 3 min. Sequencing products were purified by AutoSeq96 Plates containing DNA Grade Sephadex G-50 (Amersham). The purified products were then denatured at 95°C for 5 min with addition of 5 mul hi-deionized formamide, and run on the ABI PRISM® model 3100 capillary DNA sequencer. Sequence chromatograms were aligned and analyzed for SNPs by the PolyPhred software.24 The SNPs suggested by PolyPhred based on automated sequencer traces were double checked manually by two independent researchers to ensure the accuracy of the genotype calls. All of the SNPs analyzed were located within the high-quality region of the chromatogram, from base 200 to base 400, and occasional low-quality sequencing passes were re-run. Samples from both patient and control groups were included in the same experimental batches to minimize batch-to-batch variations in genotyping.

Study design

The first stage of the present study was directed to the discovery of SNPs in the coding regions of various subunits of GABAA receptors by direct DNA sequencing of PCR-amplified exons and of intronic sequences flanking each exon, with about 48 samples from each of the two subject groups. When the SNP rs252944, about 500 bp upstream to exon 8 of GABRB2 gene, showed promising association with schizophrenia, the study proceeded to the second stage. PCR was performed on genomic DNA flanking rs252944, spanning approximately 13 kb and consisting mostly of introns, on an expanded number of samples (Table 1). The PCR products were completely sequenced to: (a) discover any SNPs in this region and (b) genotype these SNPs for potential association with schizophrenia.

Statistical analysis

All the SNPs analyzed were tested for Hardy–Weinberg equilibrium by means of the program GENEPOP v. 1.2 (GENEPOP website).15 The complete enumeration method was chosen as suggested by GENEPOP.

Possible SNP association with susceptibility to schizophrenia was assessed by comparing the allelic and genotypic frequencies of schizophrenics and controls, using the standard chi2-test under normal approximation. The level of significance was indicated by P-values, as implemented by the CLUMP program.25 A value of P<0.05 was considered to be indicative of a statistically significant effect.

The two-loci and three-loci haplotype associations were carried out using the permutation and model-free analysis (PMPLUS v.1.1) program16,17 in the estimated haplotype-frequencies (EHPLUS) software package. In view of the incomplete penetrance and uncertain mode of inheritance of schizophrenia, model-free analysis and permutation test were performed. PMPLUS implemented the model-free statistics by maximizing log-likelihood ratio statistic over a range of parameter values, instead of fixed values as in Mendelian dominant or recessive models. A recessive model with disease allele frequency 0.1 and penetrances 0.005, 0.005, 0.5 was initially assumed and inputted as the user-specified model.26 PMPLUS prepared input files based on user-inputted data with different inheritance models, including the user-specified, recessive, dominant, model-free and heterogeneity models for EH. Maximum likelihood estimates of all the 10 two-loci haplotype frequencies were then performed by employing the expectation-maximization (EM) algorithm,27 as implemented in EH. The hypotheses concerning different disease-model chi2-statistics were performed by EH. In order to resolve the potential inaccurate significance level caused by asymptotic theory in association analyses of two-loci haplotypes, additional permutation procedures of 10 000 replicates were followed to obtain empirical P-values by means of EHPLUS.16

LD between each pair of loci in the control group was tested using the standard LD coefficient28 and the likelihood-ratio test, as implemented by the ARLEQUIN program (ARLEQUIN website). For the standard LD coefficient calculation, two-loci haplotype frequencies were estimated using a maximum-likelihood approach from unphased diploid genotype data, as computed in the ARLEQUIN program. All measures of LD were based on D=x11-p1q1, where x11 is the estimated frequency of haplotype A1B1, and p1 and q1 are the frequencies of alleles A1 and B1 at loci A and B, respectively. Measures of complete association, D', and absolute association, r2, were used in LD analyses. D' is given by D/Dmax, where Dmax=min[p1q1, p2q2] when D<0, or Dmax=min[p1q2, p2q1] when D>0. r2 is given by D/(p1p2q1q2)1/2. D' varies between 0 (no association) and 1 (maximum disequilibrium). For the likelihood-ratio tests, likelihood under the hypothesis of LD between loci is compared to the likelihood of the experimental data under LD.29 Permutation procedure was performed in order to better approximate the underlying distribution of the likelihood ratio under the null hypothesis of linkage equilibrium. The significance level of LD was represented as P-value of the chi2-test.

Top

Notes

Electronic-database information

URLs for databases and program used herein are as follows:

Top

References

  1. Dean B, Hussain T, Hayes W, Scarr E, Kitsoulis S, Hill C. et al. Changes in serotonin2A and GABAA receptors in schizophrenia: studies on the human dorsolateral prefrontal cortex. J Neurochem 1999; 72: 1593–1599. | Article | PubMed | ChemPort |
  2. Guidotti A, Auta J, Davis JM, Giorgi-Gerevini V, Dwivedi Y, Grayson DR. et al. Decrease in reelin and glutamic acid decarboxylase67 (GAD67) expression in schizophrenia and bipolar disorder: a postmortem brain study. Arch Gen Psychiatry 2000; 57: 1061–1069. | Article | PubMed | ISI | ChemPort |
  3. Benes FM, Berretta S. GABAergic interneurons: implications for understanding schizophrenia and bipolar disorder. Neuropsychopharmacology 2001; 25: 1–27. | Article | PubMed | ISI | ChemPort |
  4. Roberts E. Prospects for research on schizophrenia. A hypothesis suggesting that there is a defect in the GABA system in schizophrenia. Neurosci Res Program Bull 1972; 10: 468–482. | PubMed |
  5. Tsuang MT, Stone WS, Faraone SV. Genes, environment and schizophrenia. Br J Psychiatry Suppl 2001; 40: s18–s24. | Article | PubMed |
  6. Stefansson H, Sigurdsson E, Steinthorsdottir V, Bjornsdottir S, Sigmundsson T, Ghosh S. et al. Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet 2002; 71: 877–892. | Article | PubMed | ISI |
  7. Makino C, Fujii Y, Kikuta R, Hirata N, Tani A, Shibata A. et al. Positive association of the AMPA receptor subunit GluR4 gene (GRIA4) haplotype with schizophrenia: linkage disequilibrium mapping using SNPs evenly distributed across the gene region. Am J Med Genet 2003; 116B: 17–22.
  8. Barnard EA, Skolnick P, Olsen RW, Mohler H, Sieghart W, Biggio G. et al. International Union of Pharmacology. XV. Subtypes of gamma-aminobutyric acid A receptors: classification on the basis of subunit structure and receptor function. Pharmacol Rev 1998; 50: 291–313. | PubMed | ISI | ChemPort |
  9. Mehta AK, Ticku MK. An update on GABAA receptors. Brain Res Rev 1999; 29: 196–217. | Article | PubMed | ISI | ChemPort |
  10. Benke D, Fritschy JM, Trzeciak A, Bannwarth W, Mohler H. Distribution, prevalence, and drug binding profile of gamma-aminobutyric acid type A receptor subtypes differing in the beta-subunit variant. J Biol Chem 1994; 269: 27100–27107. | PubMed | ISI | ChemPort |
  11. Nusser Z, Sieghart W, Somogyi P. Segregation of different GABAA receptors to synaptic and extrasynaptic membranes of cerebellar granule cells. J Neurosci 1998; 18: 1693–1703. | PubMed | ISI | ChemPort |
  12. Wisden W, Seeburg PH. GABAA receptor channels: from subunits to functional entities. Curr Opin Neurobiol 1992; 2: 263–269. | Article | PubMed | ChemPort |
  13. Gurling HM, Kalsi G, Brynjolfson J, Sigmundsson T, Sherrington R, Mankoo BS. et al. Genome-wide genetic linkage analysis confirms the presence of susceptibility loci for schizophrenia, on chromosomes 1q32.2, 5q33.2, and 8p21–22 and provides support for linkage to schizophrenia, on chromosomes 11q23.3–24 and 20q12.1–11.23. Am J Hum Genet 2001; 68: 661–673. | Article | PubMed | ISI | ChemPort |
  14. Lewis CM, Levinson DF, Wise LH, DeLisi LE, Straub RE, Hovatta I. et al. Genome scan meta-analysis of schizophrenia and bipolar disorder, Part II: Schizophrenia. Am J Hum Genet 2003; 73: 34–48. | Article | PubMed | ISI | ChemPort |
  15. Raymond M, Rousset F. GENEPOP (version 1.2): population genetics software for exact tests and ecumenicism. J Heredity 1995; 86: 248–249. | ISI |
  16. Zhao JH, Curtis D, Sham PC. Model-free analysis and permutation tests for allelic associations. Hum Hered 2000; 50: 133–139. | Article | PubMed | ISI | ChemPort |
  17. Xie X, Ott J. Testing linkage disequilibrium between a disease gene and marker loci. Am J Hum Genet 1993; 53: 1107. | PubMed |
  18. Chumakov I, Blumenfeld M, Guerassimenko O, Cavarec L, Palicio M, Abderrahim H. et al. Genetic and physiological data implicating the new human gene G72 and the gene for D-amino acid oxidase in schizophrenia. Proc Natl Acad Sci USA 2002; 99: 13675–13680. | Article | PubMed | ChemPort |
  19. Shifman S, Bronstein M, Sternfeld M, Pisante-Shalom A, Lev-Lehman E, Weizman A. et al. A highly significant association between a COMT haplotype and schizophrenia. Am J Hum Genet 2002; 71: 1296–1302. | Article | PubMed | ISI | ChemPort |
  20. McKinley DD, Lennon DJ, Carter DB. Cloning, sequence analysis and expression of two forms of mRNA coding for the human beta 2 subunit of the GABAA receptor. Brain Res Mol Brain Res 1995; 28: 175–179. | Article | PubMed | ChemPort |
  21. Abe S, Suzuki T, Ito T, Baba A, Hori T, Kurita H. et al. Differential expression of GABAA receptor subunit mRNAs and ligand binding sites in rat brain following phencyclidine administration. Synapse 2000; 38: 51–60. | Article | PubMed |
  22. Huntsman MM, Tran BV, Potkin SG, Bunney Jr WE, Jones EG. Altered ratios of alternatively spliced long and short gamma2 subunit mRNAs of the gamma-amino butyrate type A receptor in prefrontal cortex of schizophrenics. Proc Natl Acad Sci USA 1998; 95: 15066–15071. | Article | PubMed | ChemPort |
  23. American Psychiatric Association. Schizophrenia and other psychiatric disorders. In: Diagnostic and Statistical Manual of Mental Disorders: DSM-IV-TR. American Psychiatric Press: Washington, DC, 2000 pp 155–165.
  24. Nickerson DA, Tobe VO, Taylor SL. PolyPhred: automating the detection and genotyping of single nucleotide substitutions using fluorescence-based resequencing. Nucleic Acids Res 1997; 25: 2745–2751. | Article | PubMed | ISI | ChemPort |
  25. Sham PC, Curtis D. Monte Carlo tests for associations between disease and alleles at highly polymorphic loci. Ann Hum Genet 1995; 59: 97–105. | Article | PubMed | ISI | ChemPort |
  26. Murray R. The essentials of postgraduate psychiatry. In: Murray R, Hill P, McGuffin P (eds) Schizophrenia. Cambridge University Press: London, 1997 pp 281–309.
  27. Excoffier L, Slatkin M. Maximum-likelihood estimation of molecular haplotype frequencies in a diploid population. Mol Biol Evol 1995; 12: 921–927. | PubMed | ISI | ChemPort |
  28. Lewontin RC. The interaction of selection and linkage. I. General considerations: heterotic models. Genetics 1964; 49: 49–67. | PubMed | ISI | ChemPort |
  29. Slatkin M, Excoffier L. Testing for linkage disequilibrium in genotypic data using the expectation-maximization algorithm. Heredity 1996; 76: 377–383. | Article | PubMed | ISI |
Top

Acknowledgements

We are grateful to Professor Laszlo Endrenyi and Professor J Tze-Fei Wong for helpful discussion, Professor Huimin Zhu for assistance in sample collections, and Miss Peggy Lee for technical support. We wish to thank the Innovation and Technology Fund of the Government of Hong Kong, and PharmacoGenetics Limited, for financial support.

Competing interests statement: We declare that we have no competing financial interests.

Extra navigation

.
ADVERTISEMENT