Abstract A124: Genome-wide association identified susceptibility loci as modifiers of colorectal cancer risk in Lynch syndrome

Abstract

Abstract Purpose: The purpose of this study was to determine whether 5 common low-penetrance variants identified in recent genome-wide association studies (GWAS) for colorectal cancer modify colorectal cancer risk in people with Lynch syndrome. People with Lynch syndrome are genetically predisposed to colorectal cancer owing to an inherited DNA mismatch repair (MMR) gene mutation; however, disease expression varies, pointing to the role of modifying factors. A study of Dutch Lynch syndrome families (Wijnen et al., 2009) examined 6 GWAS-identified markers and found that 2 of them—rs16892766 (8q23.3) and rs3802842 (11q23.1)—were associated with colorectal cancer risk. However, other risk variants have yet to be characterized in Lynch syndrome. A replication study of 3 markers in subjects enrolled irrespective of whether they were MMR gene mutation carriers (Poynter et al., 2007) found that 2 risk variants in the 8q24 region (rs10505477 and rs6983267) and 1 variant in the 9p24 region (rs719725) were associated with colorectal cancer risk. Methods: In a retrospective cohort study, we analyzed 278 people with Lynch syndrome and a confirmed MMR gene mutation from the M. D. Anderson Lynch syndrome registry. We genotyped the 5 aforementioned risk variants—rs16892766, rs3802842, rs10505477, rs6983267, and rs719725—and analyzed the association between each risk variant and colorectal cancer using Cox proportional hazards regression while adjusting for sex, MMR gene mutation status, and familial correlation. Results: Of the 278 subjects from 128 families with 11,702 person years of follow-up, 146 subjects had colorectal cancer, and the remaining subjects were unaffected. In a codominant model, rs16892766 was significantly associated with colorectal cancer (hazard ratio [HR]: 1.61; 95% confidence interval [CI]: 1.10–2.35). Similarly, rs3802842 exhibited a marginally significant trend in an additive model (Ptrend = 0.07), and the association was most clearly evident among carriers of mutations in the MLH1 MMR gene in a codominant model (HR: 1.98; 95% CI: 1.18–3.33). Conclusion: Our study replicates the previous finding of an association between 2 GWAS-identified susceptibility loci and risk of colorectal cancer in people with Lynch syndrome. Although people with Lynch syndrome have a strong predisposition to colorectal cancer due to the underlying MMR mutation, these common genetic variants may further modify their colorectal cancer risk, particularly in carriers of multiple risk alleles. These and other low-penetrance susceptibility loci are likely to have future application in personalized risk prediction in identifying individuals who are most susceptible so that they can be provided more rigorous cancer-prevention strategies. Citation Information: Cancer Prev Res 2010;3(1 Suppl):A124.