BRIEF REPORTS
Neuregulin 3 Genetic Variations and Susceptibility to Schizophrenia in a Chinese Population Ying-Chieh Wang, Jen-Yeu Chen, Mao-Liang Chen, Chia-Hsiang Chen, I-Ching Lai, Tzu-Ting Chen, Chen-Jee Hong, Shih-Jen Tsai, and Ying-Jay Liou Background: The study investigated the possible association of NRG3 gene and schizophrenia in a Han Chinese population. Methods: Of a total of 1345, 270 unrelated schizophrenia inpatients, 235 normal control subjects, and 280 nuclear families (trios) with schizophrenia probands were studied. Nine single nucleotide polymorphisms (SNPs) spanning intron 1 to exon 9 of the NRG3 gene were analyzed, starting with the case-control samples. The SNPs showing significant association with schizophrenia in the case-control samples were subsequently studied in the independent trio samples with family-based association analysis. Results: In case-control samples, two SNPs (rs1937970 and rs677221) showed significant genotypic and allelic association with schizophrenia (all p ⬍ .05) with rs677221-C being the risk allele for schizophrenia (uncorrected p ⫽ .001, odds ratio ⫽ 1.439, 95% confidence interval ⫽ 1.115–1.858). Haplotypes GC constructed by the two SNPs was also significantly associated with schizophrenia (permutation p value ⫽ .0047). In the independent trio samples, rs1937970-A and rs677221-G consistently showed significant under-transmission to schizophrenic offspring (unadjusted p ⫽ .003 and p ⫽ .004, respectively). In the haplotype–transmission disequilibrium test (TDT) for allelic combination of rs1937970-rs677221, significant under-transmission for haplotype AG (uncorrected p ⫽ .006) and over-transmission for haplotype GC (uncorrected p ⫽ .004) to the affected schizophrenia offspring were observed. Conclusions: The result supports that the NRG3 gene is a susceptibility gene for schizophrenia. Key Words: Association study, neuregulin-3, neuregulins, NRG3, schizophrenia
T
he neuregulins (NRG1– 4) are a family of proteins that share an epidermal growth factor (EGF)-like domain and act as ligands that preferentially bind to two members of the ErbB receptor tyrosine kinase family, ErbB3 and ErbB4 (1). One of the members of neuregulin family, neuregulin-1 (NRG1), has been the subject of increased interest in psychiatry since the report of Stefansson et al. (2,3) that showed the association of one SNP (SNP8NRG221533) and a core “at-risk” haplotype (HapICE) in NRG1 gene and schizophrenia. Several recent metaanalyses further support the involvement of NRG1 gene in schizophrenia susceptibility (4 – 6). Among other members of the neuregulin family, NRG3 might also be of interest to those studying psychiatric disorders such as schizophrenia. The NRG3 has some features that distinguish it from NRG1. For example, the EGF-like domain of NRG3 has only 31% identity compared with NRG1’s EGF-like domain (7), and also in contrast to NRG1, the extracellular domain of NRG3 is devoid of Ig-like or kringle domains (7). Additionally, NRG3 is expressed at high levels in the hippocampus, amygdala, and thalamus (7). Meanwhile, NRG3 specifically binds to and activates tyrosine phosphorylation of ErbB4 (7), a receptor that evidence suggests might also be involved in schizophrenia From the Department of Psychiatry (Y-CW, J-YC, M-LC, I-CL, T-TC), Yuli Veterans Hospital; Institute of Medical Sciences (Y-CW, I-CL); Institute of Human Genetics (C-HC), Tzu-Chi University; Department of Psychiatry (C-HC), Tzu-Chi University and General Hospital, Hualien City, Taiwan, ROC. Department of Psychiatry (C-JH, S-JT, Y-JL), Taipei Veterans General Hospital; Division of Psychiatry (C-JH, S-JT); Institute of Clinical Medicine (Y-JL), School of Medicine; and the Institute of Brain Science (C-JH), National Yang-Ming University, Taipei, Taiwan, Republic of China. Address reprint requests to Ying-Jay Liou, M.D., Department of Psychiatry, Taipei Veterans General Hospital, No. 201, Shih-Pai Road, Sec. 2, 11217, Taipei, Taiwan. E-mail:
[email protected]. Received April 10, 2008; revised July 8, 2008; accepted July 9, 2008.
0006-3223/08/$34.00 doi:10.1016/j.biopsych.2008.07.012
(8 –10). Furthermore, the gene encoding NRG3 (HUGO ID: NRG3) is mapped to chromosome 10q22-q23 (11), a region that has been shown, through linkage studies, to harbor risk genes for schizophrenia (12,13). Fallin et al. (12) and Faraone et al. (13) independently reported significant linkage of D10S1774 and D10S2327 with schizophrenia, and the NRG3 gene is located in the genomic region between these two markers. We hypothesized that, because it is a member of the neuregulin family, shows specific expression in neural tissue and binding ability to ErbB4 receptor tyrosin kinase, and has been implicated in the disease by linkage studies, NRG3 gene might be a susceptibility gene for schizophrenia.
Methods and Materials Subjects Two hundred and seventy unrelated chronic inpatients (male/ female ⫽ 136/134, mean age ⫽ 44.9 ⫾ 9.0 years) with schizophrenia and 280 nuclear families including schizophrenia inpatient probands (male/female ⫽ 160/120, mean age ⫽ 40.2 ⫾ 8.7 years) and their parents were recruited into the study. The diagnosis for schizophrenia was made according to the criteria of DSM-IV by two board-certified senior psychiatrists on the basis of clinical interview and observation, chart recordings, ward observation, and interviews with patients’ families. The subjects in the control group (male/female ⫽ 101/134, mean age ⫽ 51.5 ⫾ 10.3 years) were also interviewed by board-certified psychiatrists and were free of symptoms or histories of psychiatric illness. All the participants were Han Chinese. The institutional review board approved the study, and written informed consent was obtained from the participants. Genotyping We used TaqMan assays (http://www.appliedbiosystems. com/) to genotype candidate SNPs for each individual. Nine single nucleotide polymorphisms (SNPs) with minor allele frequencies (MAF) ⬎ 30% in the Han Chinese population in Beijing (CHB) population (according to the International HapMap BIOL PSYCHIATRY 2008;64:1093–1096 © 2008 Society of Biological Psychiatry
1094 BIOL PSYCHIATRY 2008;64:1093–1096 Project) and nearly evenly distributed across the NRG3 gene were selected for study. They range from intron 1 to exon 9, cover 1044 kilobase pair (kb) and represent 94% of the length of NRG3 gene (http://www.ncbi.nlm.nih.gov/sites/entrez, NM_ 001010848). Statistical and Haplotype Analysis Cochran-Armitage trend test was used for comparing genotype frequency between groups. Deviation from Hardy-Weinberg Equilibrium for each SNP was check by a 2 goodness of fit test (df ⫽ 1). The threshold of significance was set as p ⬍ .05. The genotype data from the control group were also used for pairwise linkage disequilibrium (LD) measurement and for assessing block structure with Haploview (http:/www.broad. mit.edu/mpg/halpoview/index.plp). We also used SNPAlyzeV3.2 (http://www.dynacom.co.jp/e/products/package/snpalyze/about. html) to examine the difference in haplotype frequencies between groups, and the significance levels were set as p ⬍ .05 after 100,000 permutation tests. For family-based analysis, we used family-based association (transmission disequilibrium test [TDT]) and haplotype-based TDT modules implemented in PLINK (14), with p ⬍ .05 being regarded as a sign of significant association. We estimated the minimum false discovery rate (FDR) for each test p value with QVALUE (http://faculty.washington.edu/ ⬃jstorey/qvalue/) (15). An FDR of ⬍ .05 of its corresponding p value was regarded as a significant association between markers and disease after multiple testing corrections.
Results Single-Point and Haplotype Analyses in the Case-Control Samples The gender distribution was similar between case and control groups (p ⫽ .097), but the control groups were significantly older than the case group (control subjects vs. cases ⫽ 51.5 ⫾ 10.3 vs. 44.9 ⫾ 9.0 years, p ⬍ .001). No studied SNPs in the either group deviated significantly from Hardy-Weinberg Equilibrium. In single-point analysis, one SNP rs677221 showed significant genotypic and allelic association with schizophrenia (Table 1), which remained significant after correction for multiple testing (both genotype and allele FDR ⫽ .017). The rs677221-C allele was significantly associated with schizophrenia susceptibility (p ⫽ .001, odds ratio [OR] ⫽ 1.439, 95% confidence interval [CI] ⫽ 1.115 to approximately 1.858). The Rs1937970 was also found to be associated with schizophrenia (Table 1; genotype: p ⫽ .027; allele: p ⫽ .024). However, these significant signals could not survive after correction for multiple testing (FDR for genotype and allele comparisons ⫽ .062 and .061, respectively). Significant LD was observed between rs1937970 and rs677221 (absolute D’ ⫽ .904, r2 ⫽ .782) and between rs4933859 and rs2295933 (absolute D’ ⫽ .789, r2 ⫽ .611). As shown in Table 2, there were significant differences in the rs1937970-rs677221 haplotype frequency between the schizophrenic and control groups (global permutation p value ⫽ .0068, FDR ⫽ .019), and the haplotype GC was significantly more frequent in the schizophrenic group (permutation p ⫽ .0047), whereas haplotype AG was significantly more frequent in the control group (permutation p ⫽ .0016). Both findings remained after multiple testing correction (FDR for haplotype GC and AG were all ⫽ .017). There was no significant difference in haplotype distribution of rs4933859-rs2295933 combination between groups in global analysis (permutation p value ⫽ .959). www.sobp.org/journal
Y.-C. Wang et al. Replication of Findings in the Family-Based Samples We subsequently replicated the findings that rs677221 and rs1937970 were associated with schizophrenia in an independent sample set consisting of 280 parent-affected offspring trios. The Rs1937970-A and rs677221-G were both significantly undertransmitted to the affective offspring, even after taking multiple testing into consideration (rs1937970-A: p ⫽ .003, OR ⫽ .68, 95% CI ⫽ .52 to approximately .88, FDR ⫽ .017; rs677221-G: p ⫽ .004, OR ⫽ .69, 95% CI ⫽ .53 to approximately .89, FDR ⫽ .017). In the haplotype-TDT test for allelic combination of rs1937970rs677221, significant under-transmission for haplotype AG (transmission vs. non-transmission numbers ⫽ 91.5 vs. 132.5, p ⫽ .006, FDR ⫽ .017) and over-transmission for haplotype GC (transmission vs. non-transmission numbers ⫽ 140.5 vs. 96.5, p ⫽ .004, FDR ⫽ .017) to the affected schizophrenia offspring were observed. All the results echoed the observations in the casecontrol samples.
Discussion In the study, both case-control and family-based analyses found the same direction of association, in that rs677221-C conferred a risk for schizophrenia susceptibility and that haplotype-GC corresponded to the over-transmitted in schizophrenia groups. Our results support two previous linkage studies for schizophrenia susceptibility loci on chromosome 10q22 where NRG3 is located (12,13). In a study undertaking a screening of 440 SNPs in 64 candidate genes, NRG3 was listed as a suggestive (.01 ⬍ empirical p value ⬍ .05 for any SNP or haplotype) gene of association for schizophrenia (16). The association was later supported by the work of Benzel et al. (17) in which one SNP (rs3924461) was in highly significant allelic association with schizophrenia. The Rs3924461 resides in intron 2 of the NRG3 gene and is located in the downstream and 118 kb apart from rs677221, the significant SNP in our study. Collectively, results of the two former studies (16,17) and ours add a piece of evidence suggesting that aberrant neuregulin-ErbB signaling might be involved in the pathogenesis of schizophrenia (17). We cannot rule out that the association of rs677221 and its related haplotypes with schizophrenia in the case-control analysis could be chance findings due to hidden population-stratification factors. However, our findings in the case-control samples could be subsequently replicated in other family-based samples with a family-based association test, which is thought to have the advantage— over population-based studies— of circumventing population stratification (18). Therefore, it is unlikely that the association between rs677221 and its related haplotypes and schizophrenia could be attributed to spurious findings due to population stratification. In spite of that, there is still a 10% chance of a false replication even with this precision (19), which indicates that our results might need further replication. There are 1132 SNPs (MAF ⱖ 10%) throughout the NRG3 in the CHB population in HapMap (db126 Apr07), and among them, 55 (4.9%) could be tagged by the studied markers. Accordingly, one limitation of the study was that the marker selection could not fully represent the common variations in the NRG3 gene. Another limitation was that we did not sequence the whole coding region in NRG3. According to the SNP database (http:// www.ncbi.nlm.nih.gov/SNP/snp_ref.cgi?locusId⫽10718), five SNPs are located in the coding sequence of the NRG3 gene: rs2295934 (Arg472Ser), rs17101193 (Lys552Asn), rs17101196 (Pro590Pro), rs959317 (Gly610Val), and rs2295933 (Ser662Ser). The first four of the five SNPs have low MAF in Han population (all MAF ⬍
Y.-C. Wang et al.
Table 1. Genotype and Allele Frequencies of Each Locus and the Results of Comparisons Between Schizophrenia and Control Groups
Location
Position (bp)a
Distance (bp)b
Allele (1/2)c
rs10786781
Intron1
83691021
0
T/C
rs1764072
Intron1
83954476
263455
A/G
rs1937967
Intron1
84105102
150626
T/C
rs1937970
Intron2
84213446
108344
G/A
rs677221
Intron2
84304377
90931
C/G
rs2644205
Intron2
84486184
181807
A/G
rs2348553
Intron3
84591305
105121
A/G
rs4933859
Intron4
84680520
89215
C/T
rs2295933
Exon 9
84735236
54716
T/C
Genotype Frequency (%) 1/1
1/2
2/2
p (genotype)d
1
2
p (allele)
p (HWE)
Case Control Case Control Case Control Case Control Case Control Case Control Case Control Case Control Case Control
108 (42.0) 79 (34.8) 86 (32.5) 85 (37.1) 96 (36.4) 79 (34.2) 114 (43.5) 73 (31.7) 121 (46.0) 72 (30.9) 101 (38.5) 90 (39.5) 56 (21.1) 59 (25.8) 107 (40.7) 91 (40.3) 109 (41.1) 89 (38.5)
117 (45.5) 112 (49.3) 131 (49.4) 113 (49.3) 122 (46.2) 102 (44.2) 115 (43.9) 123 (53.5) 109 (41.1) 121 (51.9) 129 (49.2) 110 (48.2) 140 (49.3) 113 (49.3) 128 (48.7) 112 (49.6) 128 (48.3) 119 (51.5)
32 (12.5) 36 (15.9) 48 (18.1) 31 (13.5) 46 (17.4) 50 (21.6) 33 (12.6) 34 (14.8) 33 (12.5) 40 (17.2) 32 (12.2) 28 (12.3) 69 (26.0) 57 (24.9) 28 (10.6) 23 (10.2) 28 (10.6) 23 (10.0)
.087
333 (64.8) 270 (59.5) 303 (57.2) 283 (61.8) 314 (59.5) 260 (56.3) 343 (65.5) 269 (58.5) 351 (66.7) 265 (56.9) 331 (63.2) 290 (63.6) 252 (47.5) 231 (50.4) 342 (65.0) 294 (65.0) 346 (65.3) 297 (64.3)
181 (35.2) 184 (40.5) 227 (42.8) 175 (38.2) 214 (40.5) 202 (43.7) 181 (34.5) 191 (41.5) 175 (33.3) 201 (43.1) 193 (36.8) 166 (36.4) 278 (52.5) 227 (49.6) 184 (35.0) 158 (35.0) 184 (34.7) 165 (35.7)
.089
.971 .723 .878 .495 .502 .118 .635 .124 .282 .371 .342 .525 .335 .843 .254 .174 .283 .061
SNP, singe-nucleotide polymorphism; HWE, Hardy-Weinberg Equilibrium. Position: the base pair (bp) position of each SNP on chromosome 10. b Distance: relative distance in bases from previous SNP. c Allele 1 and 2 represent the major and minor alleles of each SNP. d P values were obtained by Cochran-Armitage trend test. e Odds ratio for allele 1 carrier belonging to the case group: 1.346 (1.039 to approximately 1.742). f Odds ratio for allele 1 carrier belonging to the case group: 1.439 (1.115 to approximately 1.858). a
Allele Frequency (%)
Subject Group
.136 .327 .022 .002 .887 .359 .993 .730
.140 .310 .024e .001f .890 .365 .926 .743
www.sobp.org/journal
BIOL PSYCHIATRY 2008;64:1093–1096 1095
SNP ID
1096 BIOL PSYCHIATRY 2008;64:1093–1096
Y.-C. Wang et al.
Table 2. Inferred rs1937970-rs677221 Haplotype Frequency and the Results of Comparisons Between the Schizophrenia and Control Groups Haplotype GC Case Control AC Case Control GG Case Control AG Case Control Global
Frequency (%)
Permutation p
63.0 54.5
.0047
3.99 2.26
.0976
2.81 3.42
.4931
30.2 39.8
.0016 .0068
.05). The fifth SNP rs2295933 had been investigated in the present study. However, the SNP was not associated with schizophrenia (Table 1) and is not in LD with the significant SNP rs667221 (D’ ⫽ .029, r2 ⫽ .001). Thus we could not demonstrate any functional variants near or within the NRG3 gene. The study indicates that, in addition to NRG1, the NRG3 gene might also be a susceptibility gene for schizophrenia. Further replications in other populations will be necessary.
This work was supported by Grants from the National Science Council, Taiwan, Republic of China (ROC), NSC-96-2314-B480-002-MY3, and Grants from Yuli Veterans Hospital, Hualien, Taiwan, ROC, VHYL-96-03 and VHYL-96-04. The authors reported no biomedical financial interests or potential conflicts of interest. 1. Buonanno A, Fischbach GD (2001): Neuregulin and ErbB receptor signaling pathways in the nervous system. Curr Opin Neurobiol 11:287–296. 2. Stefansson H, Sigurdsson E, Steinthorsdottir V, Bjornsdottir S, Sigmundsson T, Ghosh S, et al. (2002): Neuregulin 1 and susceptibility to schizophrenia. Am J Hum Genet 71:877– 892. 3. Stefansson H, Sarginson J, Kong A, Yates P, Steinthorsdottir V, Gudfinnsson E, et al. (2003): Association of neuregulin 1 with schizophrenia confirmed in a Scottish population. Am J Hum Genet 72:83– 87.
www.sobp.org/journal
4. Li D, Collier DA, He L (2006): Meta-analysis shows strong positive association of the neuregulin 1 (NRG1) gene with schizophrenia. Hum Mol Genet 15:1995–2002. 5. Munafo MR, Thiselton DL, Clark TG, Flint J (2006): Association of the NRG1 gene and schizophrenia: A meta-analysis. Mol Psychiatry 11:539 – 546. 6. Munafo MR, Attwood AS, Flint J (2008): Neuregulin 1 genotype and schizophrenia. Schizophr Bull 34:9 –12. 7. Zhang D, Sliwkowski MX, Mark M, Frantz G, Akita R, Sun Y, et al. (1997): Neuregulin-3 (NRG3): A novel neural tissue-enriched protein that binds and activates ErbB4. Proc Natl Acad Sci U S A 94:9562–9567. 8. Law AJ, Kleinman JE, Weinberger DR, Weickert CS (2007): Disease-associated intronic variants in the ErbB4 gene are related to altered ErbB4 splice-variant expression in the brain in schizophrenia. Hum Mol Genet 16:129 –141. 9. Nicodemus KK, Luna A, Vakkalanka R, Goldberg T, Egan M, Straub RE, et al. (2006): Further evidence for association between ErbB4 and schizophrenia and influence on cognitive intermediate phenotypes in healthy controls. Mol Psychiatry 11:1062–1065. 10. Silberberg G, Darvasi A, Pinkas-Kramarski R, Navon R (2006): The involvement of ErbB4 with schizophrenia: Association and expression studies. Am J Med Genet B Neuropsychiatr Genet 141:142–148. 11. Gizatullin RZ, Muravenko OV, Al-Amin AN, Wang F, Protopopov AI, Kashuba VI, et al. (2000): Human NRG3 gene Map position 10q22-q23. Chromosome Res 8:560. 12. Fallin MD, Lasseter VK, Wolyniec PS, McGrath JA, Nestadt G, Valle D, et al. (2003): Genomewide linkage scan for schizophrenia susceptibility loci among Ashkenazi Jewish families shows evidence of linkage on chromosome 10q22. Am J Hum Genet 73:601– 611. 13. Faraone SV, Hwu HG, Liu CM, Chen WJ, Tsuang MM, Liu SK, et al. (2006): Genome scan of Han Chinese schizophrenia families from Taiwan: Confirmation of linkage to 10q22.3. Am J Psychiatry 163:1760 –1766. 14. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. (2007): PLINK: A tool set for whole-genome association and populationbased linkage analyses. Am J Hum Genet 81:559 –575. 15. Storey JD (2002): A direct approach of false discovery rates. J R Stat Soc Ser B 64:479 – 498. 16. Fallin MD, Lasseter VK, Avramopoulos D, Nicodemus KK, Wolyniec PS, McGrath JA, et al. (2005): Bipolar I disorder and schizophrenia: A 440single-nucleotide polymorphism screen of 64 candidate genes among Ashkenazi Jewish case-parent trios. Am J Hum Genet 77:918 –936. 17. Benzel I, Bansal A, Browning BL, Galwey NW, Maycox PR, McGinnis R, et al. (2007): Interactions among genes in the ErbB-Neuregulin signalling network are associated with increased susceptibility to schizophrenia. Behav Brain Funct 3:31. 18. Spielman RS, McGinnis RE, Ewens WJ (1993): Transmission test for linkage disequilibrium: The insulin gene region and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 52:506 –516. 19. Sullivan PF (2007): Spurious genetic associations. Biol Psychiatry 61: 1121–1126.