Oral Presentation The 13th International Congress of the Immunology of Diabetes Society 2013

Non-HLA SNPs predicting early childhood auto-antibodies and diabetes: Teddy Study (#14)

Carina Torn 1 , David Hadley 2 , Hye_Seung Lee 2 , William Hagopian 3 , Jin-Xiong She 4 , Ake Lernmark 1 , Olli Simell 5 , Marian Rewers 6 , Andrea Steck 6 , Anette G Ziegler 7 , Desmond Schatz 8 , Beena Akolkar 9 , Suna Onengut-Gumuscu 10 , Wei-Min Chen 10 , Jorma Toppari 5 , Stephen S Rich 10 , Jeffrey Krischer 2 , TEDDY STUDY GROUP
  1. Lund University, Malmo, Sweden
  2. Pediatric Epidemiology Center, University of South Florida, Tampa, FL, US
  3. Pacific Northwest Diabetes Research Institute, Seattle, WA, US
  4. Medical College of Georgia, Atlanta, Georgia, US
  5. University of Turku, Turku, Finland
  6. Barbara Davis Center for Childhood Diabetes, Denver, CO, US
  7. Diabetes Research Institute, Munich, Germany
  8. University of Florida, Gainesville, FL, US
  9. National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, US
  10. University of Virginia, Charlottesville, VA, US

Forty-one SNPs reported as associated with Type 1 diabetes (T1D) from GWAS meta-analysis by the Type 1 Diabetes Genetics Consortium (T1DGC) were interrogated in The Environmental Determinants of Diabetes in the Young (TEDDY) study using time-to-event analysis. The HLA-categories that were used as criteria for eligibility into the TEDDY-study mainly identified high-risk subjects within the Caucasian population. Therefore, analyses were restricted to Caucasians from the US sites and all subjects from European sites where a homogenous population was demonstrated using principal components analysis. Of 5,163 TEDDY children prospectively followed from birth to a median 53 months (range 6 to 101), versus a median 8 years at diagnosis in T1DGC, 300 children attained persistent islet autoantibody (IA) positivity, 74 transitioned to T1D, while 4853 remained autoantibody negative and non-diabetic. Six SNPs reported by the T1DGC were significantly (P<0.05) associated with time-to-occurrence of IAs; rs2476601 (PTPN22) HR=1.52, rs2816316 (RGS1) HR=1.24, rs10517086 (chr4p15.2) HR=1.26, rs2292239 (ERBB3) HR=1.30, rs3184504 (SH2B3) HR=1.40 and rs1004446 (INS) HR=0.77. Five SNPs were replicated by time-to-T1D analysis, rs2476601 (PTPN22) HR=2.42, rs2292239 (ERBB3) HR=1.74, rs7111341 (INS) HR=0.53, rs1004446 (INS) HR=0.60 and rs3825932 (CTSH) HR=0.68. These SNPs replicated directions as reported by T1DGC, except rs7111341where only genome wide association has been reported. Despite limited power, time-to-event analyses successfully detect associations of T1D risk loci with both the appearance of autoantibodies as well with the subsequent development of T1D.