ESPE Abstracts (2018) 89 FC3.2

Genome-Wide Meta-Analysis Identifies a Novel Low Frequency STK39 Variant of Large Effect on Risk of Type 1 Diabetes

Vincenzo Forgettaa, Despoina Manousakia,b, Stephanie Rossa,c, Marie-Catherine Tessiera, Luc Marchandd, Hui-Qi Que, Jonathan P Bradfieldf, Struan FA Grantf,g,h, Hakon Hakonarsonf,g,h, Andrew Patersoni, Ciriaco Piccirilloj,k,l, Constantin Polychronakosb,d & J Brent Richardsa,b,c,m,n


aLady Davis Institute for Medical Research, Centre for Clinical Epidemiology, Jewish General Hospital, McGill University, Montreal, Quebec, Canada; bDepartment of Human Genetics, McGill University, Montreal, Quebec, Canada; cDepartment of Epidemiology and Biostatistics, McGill University, Montreal, Quebec, Canada; dDepartment of Pediatrics, McGill University, Montreal, Quebec, Canada; eHuman Genetics Center, University of Texas School of Public Health, Houston, Texas, USA; fThe Center for Applied Genomics, The Children’s Hospital Philadelphia, Philadelphia, Pennsylvania, USA; gDepartment of Pediatrics, University of Pennsylvania School of Medicine, Philadelphia, Pennsylvania, USA; hDivision of Human Genetics, Abramson Research Center, The Children’s Hospital of Philadelphia, Philadelphia, Pennsylvania, USA; iThe Hospital for Sick Children Research Institute, The Hospital for Sick Children, Toronto, Ontario, Canada; jDepartments of Medicine and Microbiology and Immunology, McGill University, Montreal, Quebec, Canada; kProgram in Infectious Diseases and Immunology in Global Health, Centre for Translational Biology, Research Institute of McGill University Health Centre, Montreal, Quebec, Canada; lFOCiS Centre of Excellence in Translational Immunology (CETI), Montreal, Quebec, Canada; mDepartment of Medicine, McGill University, Montreal, Quebec, Canada; nDepartment of Twin Research and Genetic Epidemiology, King’s College London, London, UK


Background: The genetic etiology of Type 1 Diabetes (T1D) is well recognized, with over 60 loci being identified to date, mainly through genome-wide association studies (GWAS). Most of these genetic associations involve common variants, while a sizable portion of the missing heritability of T1D could be attributed to unidentified rare single nucleotide polymorphisms (SNPs) (minor allele frequency (MAF) < 5%). The recent availability of large human whole genome sequencing datasets has enabled the interrogation of rare genetic variation by deep imputation from directly genotyped data, in the search of large genetic effects associated with rare alleles. Here we undertook a large GWAS meta-analysis to identify rare genetic variants with large effects on T1D risk.

Methods/Results: Through deep imputation, GWAS, and meta-analysis of 12 cohorts of European ancestry totaling of 9,684 T1D cases and 15,743 controls, we identified 43 independent genome-wide significant variants outside the major histocompatibility complex, 23 of which were common, mostly at known autoimmune loci (N=14). Twenty-four of these variants had MAF < 5%, all with large effects (OR>1.5), and only five were in known autoimmune loci. Seventeen of the 24 rare variants successfully replicated in a separate cohort including 4,329 T1D cases from T1DGC and 9,543 controls from UK Biobank. After combining discovery and replication results, 14 of these 17 variants had an OR >1.5, and 10 achieved an OR > 2. Our in silico follow-up of the lead rare SNPs using a topological domain analysis prioritized a low-frequency variant at 2q24.3 (rs60587303 (C), MAF 1.8%) within the first intron of STK39, with a combined OR on T1D of 1.97 (95% CI 1.58–2.47, Pmeta 2.9 × 10−9), an effect comparable to those of the insulin (INS) and PTPN22 loci. This variant overlaps a DNase I hypersensitivity cluster of 68 cell types, and a cluster of transcription factor bindings sites including FOS, JUN and STAT3. Pharmacological blocking of Stk39 in murine T-cell lines increased activation and proliferation of effector T-cells and secretion of gamma interferon from CD4 T-cells.

Conclusion: We identified 17 novel rare variants associated with T1D, most of which have effect sizes larger than previously described common SNPs. Our findings highlight STK39 as a protein, with no previously known role in T1D, which appears to influence T-cell activation and effector functions. These findings add to the knowledge of the genetic architecture of T1D and demonstrate STK39 as a new T1D gene with potential therapeutic implications.