Abstract 21: Accounting For Smoking Behavior In Genome-wide Analysis Of Obesity Phenotypes: The Giant (genetic Investigation Of Anthropometric Traits) Consortium

Obesity, especially central obesity, and cigarette smoking (SMK) are both important risk factors for cardiovascular disease (CVD). Yet, smokers often exhibit lower body mass index (BMI) and higher waist circumference (WC) compared to non-smokers; also, smoking cessation leads to weight gain. Genome-wide association (GWA) studies have identified genetic loci that are associated with risk of overall and central obesity; yet little is known about whether and how SMK influences the genetic susceptibility to obesity. This study aims to discover genetic loci that are associated with obesity, measured as BMI and WC adjusted for BMI (WC a ), and where the effects are hidden by SMK. We used results from 42 studies including up to 126,767 subjects with GWA data available through the GIANT Consortium. Each study employed three association models while considering smoking status (current vs. not current smokers): 1) SNP effects adjusted for SMK (SNPadjSmk), 2) SNP by SMK interaction effects (SNPxSMK), and 3) joint effect of SNP and the SNP x SMK interaction effect (SNPxSMK J ). Study specific results were combined by inverse-variance weighted fixed-effects meta-analyses of the study specific results in men and women separately and then combined sexes. A total of 53 SNPs for WC a reached genome-wide significance (GWS) (p<1E-8) by adjusting for and allowing for interaction with SMK (SNPadjSMK and SNPxSMK J ), 27 of these loci are novel for WC a . Additionally, SNPadjSMK and SNPxSMK J models identified a total of 49 SNPs associated with BMI that reached GWS, including 15 novel loci. While no loci reached GWS for SNPxSMK on BMI, one novel SNPxSMK association with WC a near PRNP reached GWS in women. PRNP is highly expressed throughout the CNS and especially the hippocampus, as are many obesity-related genes. PRNP also plays a role in the body’s response to oxidative stress, making it a likely candidate gene for interaction with smoking as well. Of the GWS SNPs, 47 WC a and 41 BMI loci are nearby (<500 kb) previously established loci for traits of interest (e.g. SMK, obesity, CVD, VAT, lipids, T2D, T1D and other anthropometrics). We identify one locus with strong evidence of interaction with SMK on obesity, and notably, we find several new loci associated with obesity by accounting for the influence of SMK. Our results highlight the importance of appropriately modeling genetic associations by considering known biological relationships between phenotypes and the environment.