1469-P: Gene-Level Gene–Mediterranean Diet Interaction Detection Influencing Inflammatory and Glycemic Biomarkers in 137,000 UK Biobank Participants

Abstract

Adherence to a Mediterranean diet pattern strongly reduces type 2 diabetes risk on average, but there is substantial heterogeneity in the long-term glycemic and inflammatory response to dietary changes, likely due in part to genetic factors. Using a genome-wide interaction association analysis in the UK Biobank (UKB) cohort, we sought to prioritize and characterize genes contributing to this heterogeneity in the Mediterranean diet-risk factor relationship. We studied an unrelated, multi-ancestry UKB subset who completed at least one 24-hour diet questionnaire, had available glycemic (hemoglobin A1c; HbA1c) and inflammatory (high-sensitivity C-reactive protein; hsCRP) biomarkers measured in blood, and were free of major disease at baseline (N = 136,880). A nine-item, point-based Mediterranean diet score (MDS) based on food group and nutrient intakes was calculated. We conducted a series of genome-wide interaction studies (GWIS) to test for significant variant-diet interactions: one for the MDS itself and one for each sub-component. Variant-level results were aggregated by gene using the MAGMA tool. Gene-level MDS interaction results revealed two genes with false discovery rate < 0.1 for HbA1c: LRRC24 (p = 3.2×10−6) and CCDC40 (p = 3.3×10−6). Gene-level sub-component-specific analysis showed that these two interactions were primarily driven by alcohol and fish intake, respectively, and further prioritized an interaction between intake of nuts and LIN9 gene variants impacting HbA1c (p = 8.8×10−8). Follow-up analysis at the top variant in LIN9, rs9729447, showed a substantial attenuation of the negative nut intake-HbA1c association in major allele carriers and revealed a stronger interaction effect size in individuals who completed at least two dietary questionnaires. Our GWIS analysis suggests that there are locus-specific interactors with Mediterranean diet patterns influencing glycemic and inflammatory health. Disclosure K.E.Westerman: None. J.B.Meigs: Consultant; Quest Diagnostics. A.Manning: None. Funding National Institutes of Health (K01DK133637)