rss
J Med Genet doi:10.1136/jmedgenet-2012-101411
  • Genotype-phenotype correlations
  • Communications

Genome-wide significant association of ANKRD55 rs6859219 and multiple sclerosis risk

  1. Lars Bertram1
  1. 1Neuropsychiatric Genetics Group, Department of Vertebrate Genomics, Max Planck Institute for Molecular Genetics, Berlin, Germany
  2. 2Department of Neurology, Focus Program Translational Neuroscience, University Medical Center of the Johannes Gutenberg-University Mainz, Mainz, Germany
  3. 3Max Planck Institute for Human Development, Berlin, Germany
  4. 4UPMC-INSERM-CNRS-UPMC-ICM, UMR 975–7225, Institut Cerveau Moelle Epinière (ICM), Hôpital Pitié-Salpêtrière, Paris, France
  5. 5Department of Human Genetics, Ruhr University, Bochum, Germany
  6. 6Department of Neurology, University of Rostock, Rostock, Germany
  7. 7Institute for Clinical Neuroimmunology, Ludwig Maximilian University, Munich, Germany
  8. 8Department of Neurology, University of Würzburg, Würzburg, Germany
  9. 9Centre for Research in Neuroscience at McGill University, Montreal, Canada
  10. 10Assistance Publique-Hôpitaux de Paris (AP-HP), Département de Neurologie, Hôpital Pitié-Salpêtrière, Paris, France
  11. 11Department of Neurology, Hôpital Sainte-Anne, Paris, France
  12. 12Department of Neurology, Centre Hospitalier de Versailles, Le Chesnay, France
  13. 13Department of Neurology, Centre Hospitalier Regional Universitaire, Tours, France
  14. 14Área de Neurociencias, Inst. Investigación Sanitaria Biodonostia, San Sebastián, Spain
  15. 15Servicio de Neurología, Hospital de Basurto, Bilbao, Spain
  16. 16Department of Biología Celular e Inmunología, Instituto de Parasitología y Biomedicina ‘LópezNeyra’ (IPBLN), CSIC, Granada, Spain
  17. 17Centre d'Esclerosi Múltiple de Catalunya, CEMCat, Unitat de Neuroimmunologia Clínica, Hospital Universitari Vall d'Hebron, Barcelona, Spain
  18. 18Servicio de Neurología, Unidad de Esclerosis Múltiple, Hospital Donostia, San Sebastián, Spain
  19. 19Department of Medicine, Rheumatology, and Clinical Immunology & DRFZ, Charité University Medicine, Berlin, Germany
  20. 20Department of Psychology, Technische Universität Dresden, Dresden, Germany
  21. 21Interdisciplinary Metabolic Center, Lipids Clinic, Charité University Medicine, Berlin, Germany
  22. 22Department of Neurology, St. Josef-Hospital, Ruhr-University, Bochum, Germany
  23. 23Department of Neurology, Sozialstiftung Bamberg Hospital, Bamberg, Germany
  24. 24Department of Neurology, Medical Faculty, Heinrich Heine University, Düsseldorf, Germany
  25. 25Department of Clinical Chemistry, Ludwig Maximilian University, Munich, Germany
  26. 26Institute of Human Genetics, University of Ulm, Ulm, Germany
  27. 27Neurogenomiks Laboratory, Department of Neuroscience, University of the Basque Country UPV/EHU, Leioa, Spain
  28. 28IKERBASQUE, Basque Foundation for Science, Bilbao, Spain
  1. Correspondence to Dr Christina M Lill, Neuropsychiatric Genetics Group, Department of Vertebrate Genomics, Max Planck Institute for Molecular Genetics, Ihnestr. 63-73, 14195 Berlin, Germany; lill{at}molgen.mpg.de
  • Received 9 November 2012
  • Revised 10 December 2012
  • Accepted 12 December 2012
  • Published Online First 12 January 2013

Multiple sclerosis (MS) is a genetically complex disease that shares a substantial proportion of risk loci with other autoimmune diseases.1 Along these lines, ANKRD55, originally implicated in rheumatoid arthritis, was recently reported as a potential novel MS risk gene (rs6859219, p=1.9×10−7).2 Here, we comprehensively validated this effect in independent datasets comprising 8846 newly genotyped subjects from Germany and France as well as 5003 subjects from two genome-wide association studies (GWAS). Upon meta-analysis of all available data (19 686 subjects), ANKRD55 rs6859219 now shows compelling evidence for association with MS at genome-wide significance (OR=1.19, p=3.1×10−11). Our study adds ANKRD55 to the list of established MS risk loci and extends previous evidence suggesting an overlapping genetic foundation across autoimmune diseases.

Ankyrin repeats are abundant in a large number of different proteins in humans and mediate protein–protein interactions. DNA-sequence variants in ankyrin repeat domain-containing proteins have been linked to a wide range of diseases; for example, KRIT1 mutations causative for cerebral cavernous malformations,3 NOTCH3 mutations in cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy, and RFXANK mutations in the bare lymphocyte syndrome.4 ANKRD55 (located on chromosome 5q11.2) encodes the ‘ankyrin repeat domain-containing protein 55’ the function of which is currently unknown. Single nucleotide polymorphism rs6859219 in ANKRD55 was implicated in a recent GWAS meta-analysis on rheumatoid arthritis.5 Furthermore, a joint analysis of datasets on rheumatoid arthritis and coeliac disease also indicated a role of ANKRD55 in the latter.6 Given the augmenting evidence suggesting an overlap in the genetic architecture of autoimmune diseases including MS, we have previously investigated 10 ‘autoimmune loci’ in 2895 Spanish MS cases and 2942 controls.2 In that study, rs6859219 emerged as a putative new MS locus albeit at subgenome-wide significance (p=1.9×10−7).2 Our failure to establish genome-wide significance was likely owing to the comparatively small sample size; thus, we set out to corroborate our initial association finding in additional independent datasets and to assess the overall evidence for association by meta-analysis.

We genotyped rs6859219 in 5106 MS cases and 3740 healthy control subjects of self-reported European descent from Germany and France7 ,8 (table 1) using a commercially available assay (‘TaqMan’, Applied Biosystems, Inc.). Furthermore, we obtained, reanalysed and included data on 1868 cases and 3135 controls for rs6859219 from two publicly available GWAS (‘IMSGC’9 and ‘GeneMSA’;10 in the latter, rs6859219 was analysed following imputation). GWAS quality control, imputation and analysis protocols were followed as described previously.8 Combined, these replication datasets comprised 6974 cases and 6875 controls and had ∼94% power to detect an OR of 1.20 at α=1×10−4. Power to detect association at genome-wide significance (α=5×10−8) using all available data (9869 cases and 9817 controls, ie, after including the Spanish datasets of the original study) was ∼96%.

Table 1

Demographic details of the German and French case-control datasets genotyped for ANKRD55 rs6859219

Genotyping efficiency and accuracy (based on 5% duplicate samples) in the newly genotyped datasets were 99.0% and 100%, respectively. Genotypes in controls were distributed according to Hardy–Weinberg equilibrium (p=0.209 using Pearson's χ2). Logistic regression analyses based on an additive model were adjusted for age and sex in the German and French datasets, and for principal components (PC 1–3) in IMSGC and GeneMSA to account for population substructure as previously described.8 Fixed-effect meta-analysis revealed significant association of the ANKRD55 rs6859219 C-allele with increased risk for MS across all replication datasets (OR=1.15, 95% CI 1.08 to 1.23, p=1.0×10−5, figure 1). Inclusion of the Spanish case-control datasets now exceeded the threshold for genome-wide significance by more than three orders of magnitude (OR=1.19, 95% CI 1.13 to 1.25, p=3.1×10−11, figure 1). While we found some evidence for heterogeneity of effect size estimates across datasets (I2=62, 95% CI 20 to 83, p Q statistic=0.0093; figure 1), all dataset-specific ORs suggested a risk effect for the C-allele, except for Central Spain. This indicates that heterogeneity was nearly entirely due to variance of effect size estimates at the same side of the null (OR<1).

Figure 1

Meta-analysis of datasets assessing the association between ANKRD55 rs6859219 and multiple sclerosis susceptibility in populations of European descent. Study-specific ORs (black squares) and 95% CIs (lines) were calculated using an additive model. The x-axis depicts the OR with regard to the risk allele dosage, that is, the C-allele. The summary ORs and 95% CIs (grey diamonds) were calculated based on fixed-effect meta-analysis combining all datasets as well as after stratification for the initial datasets and the validation datasets as indicated. GeneMSA, Genetic Multiple Sclerosis Associations10; GWAS, genome-wide association studies; IMSGC, International Multiple Sclerosis Genetics Consortium9; RAF, risk allele frequency in controls in the individual datasets.

Despite the compelling evidence now adding ANKRD55 to the list of established MS risk loci, the following limitations should be considered when interpreting our results. First, since determination of ethnic origin was based on self-report in the German, French and Spanish datasets, the possibility exists that results in these samples are affected by more subtle population substructure. However, appropriately adjusting for potential substructure effects in the GWAS datasets did not show any substantial change in results as compared with the unadjusted datasets. Hence, it is unlikely that population substructure has had a notable influence on our association results with rs6859219. Second, not all MS GWAS datasets published to date are publicly available and could therefore not all be included in the current study. This applies to two datasets in particular (‘Australia and New Zealand Multiple Sclerosis Genetics Consortium’ and ‘Brigham and Women's Hospital’) included in a recent GWAS meta-analysis that reported an association between rs6859219 and MS at nominal significance.11 However, combining the summary results reported in that study11 with our data (while excluding the GeneMSA and IMSGC results calculated here) does not appreciably change our overall meta-analysis results (OR=1.16, 95% CI 1.11 to 1.22, p=2.9×10−10). Finally, the pathophysiological mechanisms underlying the association between rs6859219 in ANKRD55 and MS remain elusive. Rs6859219 is located in intron 7 of the gene, which is highly expressed in CD4 effector memory cells12 but whose function remains largely unknown. As described previously, the linkage disequilibrium pattern (LD) around ANKRD55 is rather narrow,2 ,5 and rs6859219 shows noteworthy LD (r2≥0.3) only with other intronic ANKRD55 variants (based on 1000G CEU data; assessed with the SNAP software, http://www.broadinstitute.org/mpg/snap/). The LD block does not appear to extend to the two neighbouring genes (IL6ST encoding interleukin 6 signal transducer and IL31RA encoding interleukin 31 receptor A), which may be other immediate candidates interfering with the underlying autoimmune process in MS. Interestingly, recent in vitro ChIP-seq data generated by the international ENCODE project indicate that rs6859219, located ∼182 kb 5′ of the IL6ST transcription start site, lies within a target site for several transcription factors including activating enhancer binding protein 2 and early B cell factor 113 (accessible via http://genome.ucsc.edu/cgi-bin/hgGateway). Thus, functional genetic studies have to assess whether the association between rs6859219 and risk for MS and other autoimmune diseases is due to a dysfunction of ANKRD55, for example, via affecting mRNA splicing, or whether the effects may be caused by an altered transcriptional regulation ofIL6ST and/or IL31RA.

Acknowledgments

We are grateful to the individuals participating in this study. We would like to thank investigators from the International Multiple Sclerosis Genetics Consortium and from the GeneMSA project for making their GWAS data available via the dbGaP platform. We acknowledge use of the cohort of the CRB-REFGENSEP and thank ICM, CIC Pitié-Salpêtrière, Généthon and REFGENSEP's members for their help and support.

Footnotes

  • Contributors Study design and supervision: CML, FZ, KV and LB. Data acquisition and performing of experiments: CML, B-MMS, CG, VD, DAA, PB, L-AG, AK, FL, IC-R, SH, AW, ET, FP, PC, DO, AAn, AAl, MC, XM, JO, FM, TD, S-CL, ES-T, UL, AC, PR, H-PH, OA, PL, MB, TK, CK, UKZ, JTE and BF. Data analysis and interpretation of results: CML, TL, KV and LB. Writing of manuscript: CML and LB with the help of all coauthors.

  • Funding This project was funded by grants from the German Ministry for Education and Research (BMBF) and German Research Foundation (DFG; to FZ), the BMBF and the Cure Alzheimer's Fund (to LB), the Walter- and Ilse-Rose-Stiftung (to H-PH and OA), the BMBF (grant NBL3 to UKZ; grant 01UW0808 to UL and ES-T), and the Innovation Fund of the Max Planck Society (M.FE.A.BILD0002 to UL). This project was supported by INSERM, ARSEP, AFM and GIS-IBISA. CML was supported by the Fidelity Biosciences Research Initiative.

  • Competing interests LA Gerdes reports to have received travel expenses and personal compensation from Merck Serono, Teva Pharmaceutical Industries, Bayer Schering Pharma, Novartis and Biogen Idec. T Kl reports to have received travel expenses and personal compensations from Bayer Schering Pharmacy, Teva, Merck-Serono, Novartis, Sanofi-Aventis and Biogen-Idec as well as grant support from Bayer-Schering AG. None of the other authors reports any disclosures.

  • Provenance and peer review Not commissioned; externally peer reviewed.

References