|
|
Mol Psychiatry. Author manuscript; available in PMC 2009 December 14. Published in final edited form as: | PMCID: PMC2793269 NIHMSID: NIHMS79546 |
Two variants in Ankyrin 3 (ANK3) are independent genetic risk factors for bipolar disorder TG Schulze,1,2* SD Detera-Wadleigh,1 N Akula,1 A Gupta,1 L Kassem,1 J Steele,1 J Pearl,1 J Strohmaier,2 R Breuer,2 M Schwarz,3 P Propping,4 MM Nöthen,4,5 S Cichon,4,5 J Schumacher,1,4 NIMH Genetics Initiative Bipolar Disorder Consortium,§ M Rietschel,2,6 and FJ McMahon1 1 Genetic Basis of Mood and Anxiety Disorders, NIMH, NIH, U.S. Department of Health and Human Services, Bethesda, MD, USA 2 Department of Genetic Epidemiology in Psychiatry, Central Institute of Mental Health, Mannheim, Germany 3 Psychiatrisches Zentrum Norbaden, Wiesloch, Germany 4 Institute of Human Genetics, University of Bonn, Bonn, Germany 5 Department of Genomics, Life & Brain Center, University of Bonn, Bonn, Germany 6 Department of Psychiatry, University of Bonn, Bonn, Germany 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 NIMH Waves 1-4 (p=0.050; 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 × 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 × 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 × marker interaction. Our findings strongly support ANK3 as a BD susceptibility gene and suggest true allelic heterogeneity. Keywords: Genome-wide association study, GWAS, allelic heterogeneity, interaction, meta-analysis, ankyrins, ion channels, SNP, linkage disequilibrium, manic depressive illness The gene ANK3 encodes ankyrin-G, a large protein whose neural-specific isoforms, localized at the axonal initial segment and nodes of Ranvier, may help maintain ion channels and cell adhesion molecules 1. ANK3 was first implicated in the etiology of bipolar disorder by a genome-wide association study carried out in a US discovery and a German replication sample, totaling about 1,000 cases and controls 2. A SNP in an intron of ANK3 (rs9804190) was one of 88 that met criteria for replication in that study. Recently, Ferreira et al. 3 reported that several SNPs spanning a ~145 kb region in and around ANK3 were associated with bipolar disorder in a meta-analysis of some 4,000 cases and 6,000 controls from the US and the British Isles. Among these, rs10994336, whose genotypes were imputed, showed the strongest association signal. The report of Ferreira et al. 3 prompted us to revisit our earlier findings. In the present study, we address 3 questions: 1) Do our published results, which were based on pooled DNA, hold up after individual genotyping? 2) Do the association signals at rs9804190 and rs10994336 replicate in independent samples? 3) Do these two ANK3 markers, separated by approximately 340 kb, act independently to influence risk for bipolar disorder? Samples German case-case control sample This sample was used in our previous GWAS analysis. Detailed ascertainment and phenotype information was presented in Baum et al. 2 For the present study, 745 cases (401 males, 344 females; mean age 44 ± 13) with DSM-IV-defined bipolar 1 disorder from consecutive hospital admissions and 830 population-based control individuals (447 males, 383 females; mean age 49 ± 16 ) were available. The study protocol was approved by the Ethics Committees of the Faculties of Medicine of the Universities of Bonn and Heidelberg. NIMH 1–4 case-control sample This sample was used in our previous GWAS analysis 2. Cases were drawn from waves 1 through 4 of the NIMH Genetics Initiative ( http://nimhgenetics.org), a sample of multiplex families ascertained through sibling pairs with bipolar 1 or schizoaffective-bipolar disorder. One proband per family was selected for study, as detailed in Baum et al. 2 The control sample was also ascertained by the NIMH Genetics Initiative with the help of a marketing firm. For the present analysis, 457 cases (166 males, 291 females; mean age 41 ± 11) with bipolar I disorder (n=451) or schizoaffective-bipolar disorder (SABP; n=6), and 562 screened control individuals (315 males, 247 females; mean age 57 ± 17) were available. All subjects were of Eurpoean descent and were collected under protocols approved by the local Institutional Review Boards. NIMH 5 case-control sample This constitutes an additional sample that has previously not been used in our studies. Both cases and controls were recruited within the framework of wave 5 of the NIMH Genetics Initiative. The ascertainment scheme allowed for the inclusion of unrelated subjects with a diagnosis of DSM-IV bipolar I disorder or SABP, without regard to family history. Controls consisted of volunteers collected by the NIMH Genetics Initiative who were not available at the time of our initial study. The sample included 466 cases (237 males, 229 females; mean age 43 ± 13) with DSM-IV-defined bipolar I disorder (n=438) and SABP (n=28), and 212 control individuals (119 males, 93 females; mean age 51 ± 17). All subjects were of European descent and were collected under protocols approved by the local Institutional Review Boards. Markers and genotyping In the German, NIMH 1–4, and NIMH 5 samples, we studied the following ANK3 markers: rs9804190 (C/T) located in intron 36 between exons 36 and 37 (bp position 61509837) and rs10994336 (C/T) located 30 kb downstream of ANK3 (bp position 61849818). Location assignment is based on RefSeq NM_020987 of the UCSC Genome Browser ( http://genome.ucsc.edu). SNPs were individually genotyped using a modification of the 5′ nuclease (Taqman) assay, scored on a fluorescent plate reader by use of a clustering algorithm (kindly provided by Sam Chen, Virginia Commonwealth University). Genotyping accuracy was ensured by genotyping of duplicate samples on each plate. Analysis PLINK (v. 1.03) ( 4; http://pngu.mgh.harvard.edu/~purcell/plink/) was used for primary association testing and to assess deviation from Hardy-Weinberg equilibrium (HWE). Analyses were performed for each sample individually. LD between the two SNPs was calculated using the UNPHASED 3.08 software 5. A Mantel-Haenszel random effects meta-analysis, implemented in Review Manager6, was used to asses association between BD and the two ANK3 SNPs in multiple samples. Heterogeneity between the samples was assessed using the I 2-statistic 7,8. To explore further the individual contribution of each SNP to BD susceptibility in the combined German, NIMH 1–4, and NIMH 5 samples, we tested for independent effects of each SNP and potential interaction between them with PLINK version 1.03 4. Independence of the two SNPs was tested using the --logistic and --condition functions. We tested the effect of rs9804190 while adding the allelic dosage for rs10994336 as a covariate, and vice versa. Interaction was tested using the --interaction function, which is identical to the interaction term in a standard logistic regression analysis. Genotyping call rates ranged from 98 % (NIMH 1–4 controls) to 100 % (NIMH 5 cases) for rs9804190, and from 93 % (NIMH 1–4 cases) to 96 % (German cases) for rs10994336. Both SNPs were within Hardy-Weinberg expectations (p>0.05) in all samples. rs9804190 in the German, NIMH 1–4, and NIMH 5 samples The association between rs9804190 and BD that we observed in our earlier study 2 held up after individual genotyping in the German and NIMH 1–4 samples. The C allele was consistently more frequent in cases than in controls (German sample: p=0.0006, OR=1.34; NIMH 1–4 sample: p=0.05, OR=1.24). In the independent NIMH 5 sample, we also observed a significant excess of the C allele in cases (p=0.017, OR=1.38). The meta-analysis supported the observations in the individual samples (p=0.000003; OR=1.38), with an I 2 heterogeneity score of 0. ( Table 1; ) | Table 1Results from the single marker case-control association analyses and the meta-analysis |
rs10994336 in the in the German, NIMH 1–4, and NIMH 5 samples The SNP rs10994336, which was imputed in the Ferreira et al. 3 study, was also associated with BD in the German data set. The T allele was more frequent in cases than in controls (p=0.0006, OR=1.70). No significant differences in allele frequencies were observed in either of the two smaller NIMH samples (p=0.112 in NIMH 1–4 and p=0.241 in NIMH 5), but an overall association was supported by meta-analysis of the combined samples (p=0.0000170, OR=1.54, I 2=0; Table 1; ). Dissecting the relationship between rs9804190 and rs10994336 There was no detectable LD between the two markers. In the combined German, NIMH 1–4, and NIMH 5 samples (n=3272), the r 2 value was 0.003 in controls and 0.002 in cases. Each SNP was a significant, independent predictor of affection status (t=3.987, p=0.00007 for rs9804190, conditioned on rs10994336; t=3.352, p=0.0008 for rs10994336, conditioned on rs9804190). No evidence of interaction between rs9804190 and rs10994336 was detected (p=0.93). These results, in a combined set of 1668 cases and 1604 controls, provide strong support for association of two markers in ANK3 with BD across several independent samples, comprising populations from the United States and Germany. Both rs9804190 and rs10994336 were significantly and consistently associated with BD in most of the samples we studied, and both yielded significant overall association by meta-analysis. Although rs10994336 was not significantly associated with BD in either NIMH sample, the consistency of the odds ratios suggests that the lack of significance in the NIMH samples is attributable to low statistical power, rather than true heterogeneity. These findings demonstrate that replicable association findings are attainable in bipolar disorder, with sufficient sample sizes, but universal association of individual SNPs across all samples may be an unrealistic goal. The NIMH control samples we used overlap with those in the Ferreira et al. 3 study. However, the largest effects we observed were in the German sample, which was completely independent from that of Ferreira et al. 3 Thus, spurious association due to atypical allele frequencies in the NIMH controls seems highly unlikely. Since rs9804190 was one of 88 SNPs initially implicated in an earlier GWAS 2, any one of which may be the subject of future studies, the present results may need to be corrected for multiple testing. The meta-analysis p-value of 3 × 10 −6 remains significant at the p<0.01 level, whether corrected by a factor of 88 or by the factor of 1877 we adopted in the earlier study 2. Our analyses suggest that each of the two ANK3 markers we studied is an independent risk factor for bipolar disorder. The two markers are located over 340 kb apart and there is little detectable LD between them. Furthermore, each marker was a significant, independent predictor in a logistic regression analysis that included both markers. We could not detect any evidence of marker × marker interaction. Our findings may be attributable to true allelic heterogeneity, where more than one functional variant in the same locus contributes to disease risk. Allelic heterogeneity is the rule for Mendelian disorders, but to our knowledge has not yet been demonstrated for a common disease such as bipolar disorder. Taken together, the findings of Baum et al. 2, Ferreira et al. 3, and the present study strongly suggest that ANK3 is one of the many genes that play a role in the etiology of BD. Other genes have been implicated in previous meta-analyses. 3,9 It seems likely that additional genes will be identified as sample sizes grow and complementary approaches – such as those focused on rare alleles and sub-phenotypes – are pursued. Although bipolar disorder is a polygenic disease, the identification of individual risk genes may still reveal important aspects of etiology. Ankyrins belong to a family of proteins that connect integral membrane proteins to the spectrin-actin cytoskeleton. Animal studies have shown that ANK3 is essential for both normal clustering of voltage-gated sodium channels at axon initial segments and normal action potential firing 1,10–11. These are very basic mechanisms of brain function, only distantly related to the widely-held theories that focus on neurotransmitters or other signaling pathways. Our finding that two SNPs act independently within the same gene to influence risk underscores the importance of scrutinizing a broad range of genetic variation in a potential susceptibility gene highlighted by GWAS. Such scrutiny should be a goal of future studies of ANK3. Additional important markers may yet be found, but the regions near rs9804190 and rs10994336 are logical foci for future work aimed at defining the functional variants that account for these association findings. Supported by the NIMH Intramural Research Program, Deutsche Forschungsgemeinschaft, the National German Genome Research Network of the Federal Ministry of Education and Research, the National Alliance for Research on Schizophrenia and Depression (FJM, TGS), and the Alfried Krupp von Bohlen und Halbach-Stiftung. We thank the study participants who made this research possible. Data and biomaterials for the NIMH samples were collected as part of 10 projects that participated in the NIMH Bipolar Disorder Genetics Initiative. From 1991–98, the Principal Investigators and Co-Investigators were: Indiana University, Indianapolis, IN, U01 MH46282 – John Nurnberger, MD, PhD, Marvin Miller, MD and Elizabeth Bowman, MD; Washington University, St Louis, MO, U01 MH46280 – Theodore Reich, MD, Allison Goate, PhD and John Rice, PhD; Johns Hopkins University, Baltimore, MD U01 MH46274 – J. Raymond DePaulo Jr, MD Sylvia Simpson, MD, MPH and Colin Stine, PhD; NIMH Intramural Research Program, Clinical Neurogenetics Branch, Bethesda, MD – Elliot Gershon, MD, Diane Kazuba, BA and Elizabeth Maxwell, M.S.W. From 1999–03, the Principal Investigators and Co-Investigators were: Indiana University, Indianapolis, IN, R01 MH59545 – John Nurnberger, MD, PhD, Marvin J Miller, MD, Elizabeth S Bowman, MD, N Leela Rau, MD, P Ryan Moe, MD, Nalini Samavedy, MD, Rif El-Mallakh, MD (at University of Louisville), Husseini Manji, MD (at Wayne State University), Debra A Glitz, MD (at Wayne State University), Eric T Meyer, MS, Carrie Smiley, RN, Tatiana Foroud, PhD, Leah Flury, MS, Danielle M Dick, PhD and Howard Edenberg, PhD; Washington University, St Louis, MO, R01 MH059534, John Rice, PhD, Theodore Reich, MD, Allison Goate, PhD and Laura Bierut, MD; Johns Hopkins University, Baltimore, MD, R01 MH59533 – Melvin McInnis, MD, J Raymond DePaulo Jr, MD, Dean F MacKinnon, MD, Francis M Mondimore, MD, James B Potash, MD, Peter P Zandi, PhD, Dimitrios Avramopoulos and Jennifer Payne; University of Pennsylvania, PA, R01 MH59553 – Wade Berrettini, MD, PhD; University of California at Irvine, CA, R01 MH60068 – William Byerley, MD and Mark Vawter, MD; University of Iowa, IA, R01 MH059548 – William Coryell, MD and Raymond Crowe, MD; University of Chicago, IL, R01 MH59535 – Elliot Gershon, MD, Judith Badner, PhD, Francis McMahon, MD, Chunyu Liu, PhD, Alan Sanders, MD, Maria Caserta, Steven Dinwiddie, MD, Tu Nguyen, Donna Harakal; University of California at San Diego, CA, R01 MH59567 – John Kelsoe, MD, Rebecca McKinney, BA; Rush University, IL, R01 MH059556 – William Scheftner, MD, Howard M Kravitz, DO, MPH, Diana Marta, BA, Annette Vaughn-Brown, MSN, RN and Laurie Bederow, MA; NIMH Intramural Research Program, Bethesda, MD, 1Z01MH002810-01, Francis J. McMahon, MD, Layla Kassem, PsyD, Sevilla Detera-Wadleigh, PhD, Lisa Austin, PhD, Dennis L Murphy, MD. The NIMH control subjects were collected by the NIMH Schizophrenia Genetics Initiative ‘Molecular Genetics of Schizophrenia II’ (MGS-2) collaboration. The investigators and coinvestigators are: ENH/Northwestern University, Evanston, IL, MH059571 – Pablo V. Gejman, MD (Collaboration Coordinator; PI), Alan R. Sanders, MD; Emory University School of Medicine, Atlanta, GA, MH59587 – Farooq Amin, MD (PI); Louisiana State University Health Sciences Center; New Orleans, LA, MH067257 –Nancy Buccola APRN, BC, MSN (PI); University of California-Irvine, Irvine, CA, MH60870 – William Byerley, MD (PI); Washington University, St Louis, MO, U01, MH060879 – C. Robert Cloninger, MD (PI); University of Iowa, Iowa, IA, MH59566 –Raymond Crowe, MD (PI), Donald Black, MD; University of Colorado, Denver, CO, MH059565 – Robert Freedman, MD (PI); University of Pennsylvania, Philadelphia, PA, MH061675 – Douglas Levinson, MD (PI); University of Queensland, QLD, Australia, MH059588 – Bryan Mowry, MD (PI); Mt Sinai School of Medicine, New York, NY, MH59586 – Jeremy Silverman, PhD (PI). We thank the following clinician colleagues for help in collecting German patients: Margot Albus, Margitta Borrmann-Hassenbach, Ernst Franzek, Jürgen Fritze, Magdalena Gross, Thilo Held, Roland Kreiner, Mario Lanczik, Dirk Lichtermann, Wolfgang Maier, Jürgen Minges, Stephanie Ohlraun, Ulrike Reuner, Monja Tullius, Bettina Weigelt. We thank Jay Tischfield and Douglas Fugman of the Rutgers Cell and DNA Repository for banking blood samples and providing DNA for the NIMH collections. 1. Kordeli E, Lambert S, Bennett V, Ankyrin G. A new ankyrin gene with neural-specific isoforms localized at the axonal initial segment and node of Ranvier. J Biol Chem. 1995;270:2352–9. [PubMed] 2. Baum AE, Akula N, Cabanero M, Cardona I, Corona W, Klemens B, et al. A genome-wide association study implicates diacylglycerol kinase eta (DGKH) and several other genes in the etiology of bipolar disorder. Mol Psychiatry. 2008;13:197–207. [PMC free article] [PubMed] 3. Ferreira MA, O’Donovan MC, Meng YA, Jones IR, Ruderfer DM, Jones L, et al. Collaborative genome-wide association analysis supports a role for ANK3 and CACNA1C in bipolar disorder. Nat Genet. 2008 Aug 17; [Epub ahead of print] 4. Purcell S, Neale B, Todd-Brown K, Thomas L, Ferreira MA, Bender D, et al. PLINK: a tool set for whole-genome association and population-based linkage analyses. Am J Hum Genet. 2007;81:559–575. [PMC free article] [PubMed] 5. Dudbridge F. Likelihood-based association analysis for nuclear families and unrelated subjects with missing genotype data. Hum Hered. 2008;66:87–98. [PMC free article] [PubMed] 6. Review Manager (RevMan) v 5.0. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration; 2008. 7. Higgins JP, Thompson SG. Controlling the risk of spurious findings from meta-regression. Stat Med. 2004;23:1663–1682. [PubMed] 8. Trikalinos TA, Salanti G, Zintzaras E, Ioannidis JP. Meta-analysis methods. Adv Genet. 2008;60:311–334. [PubMed] 9. Baum AE, Hamshere M, Green E, Cichon S, Rietschel M, Noethen MM, Craddock N, McMahon FJ. Meta-analysis of two genome-wide association studies of bipolar disorder reveals important points of agreement. Mol Psychiatry. 2008;13:466–467. [PMC free article] [PubMed] 10. Zhou D, Lambert S, Malen PL, Carpenter S, Boland LM, Bennett V. AnkyrinG is required for clustering of voltage-gated Na channels at axon initial segments and for normal action potential firing. J Cell Biol. 1998;143:1295–1304. [PMC free article] [PubMed] 11. Pan Z, Kao T, Horvath Z, Lemos J, Sul JY, Cranstoun SD, et al. A common ankyrin-G-based mechanism retains KCNQ and NaV channels at electrically active domains of the axon. J Neurosci. 2006;26:2599–2613. [PubMed] Click on image to enlarge  Figure 1
|
The publisher's final edited version of this article is available at 
