Identification of Ancestry Specific Alleles in a Genome-Wide Association Study (GWAS) for Metformin (MET) Response in the Diabetes Prevention Program (DPP)

Genome-wide significant (GWS) loci for glycemic response to MET in type 2 diabetes (T2D) reported elsewhere have not replicated in the DPP. Thus, to assess whether pharmacogenetic interactions may differ in impaired glucose tolerance (IGT), we conducted a GWAS in the DPP for MET response, defined by diabetes incidence and change in quantitative traits (fasting glucose, 2-hour glucose on oral glucose tolerance test, hemoglobin A1c [HbA1c], fasting insulin, insulin sensitivity index, and weight). Samples were genotyped on the Illumina HumanCore Exome array and imputed on the 1000 Genomes Phase3 panel. Cox proportional hazards models tested associations with diabetes incidence in MET (n=876) and placebo (PBO, n=887) arms. Multiple linear regression using an additive model evaluated for association with 1-year change in quantitative traits, adjusted for baseline trait, age, sex, and 10 ancestry principal components. We tested for gene-by-environment (G×E) interaction between MET and PBO. Results were filtered to minor allele frequency (MAF) >1% and imputation score >0.7. We identified 3 GWS variants after correcting for correlated traits (P<9×10-9). In MET, a variant near ENOSF1 was associated with an increase in % HbA1c (rs144322333, β=0.39 [95% CI 0.28, 0.50], P=2.8×10-12) and was mainly found in African ancestry participants (MAFAFR=0.07, MAFEUR=0.002). A variant near OMSR was associated with greater weight loss (kg) in MET (rs145591055, β=-7.55 [95% CI -9.88, -5.22], P=3.2×10-10) and found in 10% of Native Americans. Neither variant had an impact in PBO; the interaction with treatment arm was significant for rs144322333 (G×E P=1.4×10-6) and rs145591055 (G×E P=1.5×10-5). No GWS associations emerged for diabetes incidence. In summary, a GWAS for MET response in participants with IGT revealed novel associations that require validation in ancestry-specific populations and may have implications for tailored therapy. Disclosure J. H. Li: None. R. L. Hanson: None. S. E. Kahn: Advisory Panel; Self; Bayer AG, Boehringer Ingelheim International GmbH, Eli Lilly and Company, Merck & Co., Inc., Novo Nordisk, Pfizer Inc. W. C. Knowler: None. T. I. Pollin: None. J. C. Florez: Consultant; Self; Goldfinch Bio, Inc., Other Relationship; Self; Novo Nordisk. D. Research group: None. J. A. Perry: None. K. A. Jablonski: None. L. Chen: None. S. Srinivasan: None. J. N. Todd: None. M. Harden: None. J. M. Mercader: None. P. W. Franks: Advisory Panel; Self; Zoe Global Limited, Consultant; Self; Eli Lilly and Company, Novo Nordisk. Funding National Institute of Diabetes and Digestive and Kidney Diseases (U01DK048489)