Large-scale genome-wide interaction analyses on multiple cardiometabolic risk factors to identify age-specific genetic risk factors

Abstract

The genetic landscape of cardiometabolic risk factors has been explored extensively. However, insight in the effects of genetic variation on these risk factors over the life course is sparse. Here, we performed genome-wide interaction studies (GWIS) on different cardiometabolic risk factors to identify age-specific genetic risks. This study included 270,276 unrelated European-ancestry participants from the UK Biobank (54.2% women, a median age of 58 [interquartile range (IQR): 50, 63] years). GWIS models with interaction terms between genetic variants and age were performed on apolipoprotein B (ApoB), low-density lipoprotein-cholesterol (LDL-C), log-transformed triglycerides (TG), body mass index (BMI) and systolic blood pressure (SBP). Replication was subsequently performed in the Copenhagen General Population Study (CGPS) and the Estonian Biobank (EstBB). Multiple lead variants were identified to have genome-wide significant interactions with age (Pinteraction < 1e − 08). In detail, rs429358 (tagging APOE4) was identified for ApoB (Pinteraction = 9.0e − 14) and TG (Pinteraction = 5.4e − 16). Three additional lead variants were identified for ApoB: rs11591147 (R46L in PCSK9, Pinteraction = 3.9e − 09), rs34601365 (near APOB, Pinteraction = 8.4e − 09) and rs17248720 (near LDLR, Pinteraction = 2.0e − 09). Effect sizes of the identified lead variants were generally closer to the null with increasing age. No variant-age interactions were identified for LDL-C, SBP and BMI. The significant interactions of rs429358 with age on ApoB and TG were replicated in both CGPS and EstBB. The majority of genetic effects on cardiometabolic risk factors remain relatively constant over age, with the noted exceptions of specific genetic effects on ApoB and TG.