Genome-wide association studies (GWAS) have successfully identified numerous genetic loci for colorectal cancer. However, GWAS loci explain only a subset of the estimated heritability and interactions with environmental risk factors have been discussed as a potential source of the missing heritability. To explore this further we conducted a genome-wide interaction analyses between genetic variants and alcohol consumption, a frequently observed risk factor for colorectal cancer. We included 8,058 cases and 8,765 controls from the Genetics and Epidemiology of Colorectal Cancer Consortium. Individual-level alcohol consumption data were harmonized across studies and participants were grouped as nondrinkers (drinking < 1 g of alcohol per day); light-moderate drinkers (drinking 1-28 g of alcohol per day); and heavy drinkers (drinking >28 g of alcohol per day). We used standard logistic regression (including an interaction term) and our newly developed Cocktail method, which includes a screening step based on: a) marginal associations between genetic variants and colorectal cancer and b) correlation between genetic variants and an environmental factor, followed by a testing step for interactions while correcting for multiple testing using weighted hypothesis testing. As the association between alcohol intake and colorectal cancer was inverse in light/moderate drinkers (OR=0.91, P=0.005) but positive in heavy drinkers (OR=1.22, P=0.0005) compared with non-drinkers, we conducted the interaction analyses separately. Using standard logistic regression to test for interaction, we identified two potential novel loci associated with colorectal cancer, 9q22.32/FBP1 (the most significant SNP rs9409565 with PG×E =1.76×10-8) and 11q22.1/CNTN5 (the most significant SNP rs6590690 with PG×E=4.95×10-8). At 9q22.32/FBP1, the common allele of rs9409565 was associated with decreased risk of colorectal cancer in all nondrinkers (OR=0.87, 95% Confidence interval [CI] 0.82-0.99, P=6.4×10-6) but was associated with increased risk in all light-moderate drinkers (OR=1.16, 95% CI 1.10-1.22, P=8.2×10-8). Similarly, at 11q22.1/CNTN5 the common allele of rs6590690 was associated with reduced risk of colorectal cancer in all nondrinkers (OR=0.76, 95% CI 0.69-0.85, P=1.2×10-6) but with increased risk in light-moderate drinkers (OR=1.32, 95% CI 1.20-1.45, P=2.5×10-8). We did not find any interaction for heavy drinkers; however, only 12 % of the participants were heavy drinkers thus limiting power. We also did not find genome-wide significant interactions using the Cocktail method. FBP1 functions as a tumor suppressor and has been shown to be involved in gastric, liver and colon carcinogenesis. Our findings demonstrate the utility of genome-wide gene-environmental interaction analyses to identify novel interactions in colorectal cancer, contributing to our understanding of colorectal cancer etiology.

Citation Format: Jian Gong, Carolyn Hutter, Jenny Chang-Claude, Polly Newcomb, Sonja Berndt, Hermann Brenner, Andrew T. Chan, Loic Le Marchand, Tabitha Harrison, Yi Lin, Martha L. Slattery, Emily White, John Potter, Shuo Jiao, Mathieu Lemire, Li Hsu, Ulrike Peters. Genome-wide interaction analyses between genetic variants and alcohol consumption and risk of colorectal cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 2195. doi:10.1158/1538-7445.AM2014-2195