Genetic Risk Factors for Gastrointestinal Cancers

Abstract

Background:Gastrointestinal cancers are comprised of a multitude of malignancies and the burden of the various sites vary greatly in incidence and mortality. Colorectal and liver cancers account for 10% and 4.7% of all incident cancer cases worldwide and rank 2nd and 3rd in terms of mortality, respectively. Precursors that can lead to colorectal cancer (CRC) and hepatocellular carcinoma (HCC) include colorectal adenomas and non-alcoholic fatty liver disease (NAFLD), respectively. Various studies have examined associations between single-nucleotide polymorphisms (SNPs) and colorectal adenomas through candidate SNP studies, but few have conducted genome-wide association studies. Many investigations have also examined genetic factors associated with NAFLD, and the various factors which comprise its diagnosis, but these studies have mainly been conducted in populations of European descent. The objectives of this dissertation were to identify genetic variants associated with colorectal adenomas, and to further examine whether the identified variants modified the relationship between selenium supplementation and adenoma development in the Selenium Trial. We also examined whether SNPs previously found to be associated with NAFLD were also associated with hepatic steatosis in a participant population of Mexican origin. Methods: To identify genetic variants associated with metachronous colorectal adenoma development we employed a genome-wide association study (GWAS) approach. 13 Participants for this GWAS were from the Selenium Trial, a phase III, double-blind, placebo controlled clinical trial designed to determine whether selenium supplementation could help prevent metachronous adenoma. All participants in the Selenium Trial were between the ages of 40 and 80 years and had at least one adenoma removed within the six months prior to study registration and had no history of hereditary syndromes which would confer increased risk to colorectal cancer. Logistic regression analyses identified variants related to metachronous adenoma. Identified variants were then validated in an independent study population of 636 patients with advanced adenoma and 855 colonoscopy-negative controls from the Colorectal Cancer Study of Austria (CORSA). Following identification of variants associated with metachronous adenoma we used logistic regression models including a multiplicative term between baseline selenium concentration tertile and the number of risk alleles in each SNP to assess effect modification. To examine whether genetic variants previously found to be associated with NAFLD were also associated with hepatic steatosis among Mexican-origin individuals, we genotyped 307 adult participants between the ages of 18 and 64 years with a body mass index (BMI) of 25 kg/m2 or higher. Participants had the following SNPs genotyped: rs12137855 (LYPLAL1), rs1260326 (GCKR), rs4240624 (PPP1R3B), rs58542926 (TM6SF2) rs641738 (MBOAT7) and rs738409 (PNPLA3). All participants had hepatic steatosis assessed through transient elastography (FibroScan®). Regression models were used to examine the association between the six SNPs and hepatic steatosis. BMI was examined as a potential effect modifier of the genetic associations. 14 Results: Participants in the Selenium trial were approximately 63 years old and were mostly male (64%). No genetic variants reached genome-wide significance for our metachronous adenoma GWAS. We did, however, identify an intron variant of FAT atypical cadherin 3 in the FAT3 gene, rs61901554, which showed a suggestive association (P=1.10x10-6); while in CORSA, a SNP within 250Kb of this same variant was associated with advanced adenoma (P=3.23x10-6; rs7944251). We also found two variants, rs12728998 and rs6699944, in NLRP3, a gene encoding cryopyrin which forms inflammasomes, which were also observed to have suggestive associations with metachronous lesions (P=2.00x10-6) in the Selenium Trial. No heterogeneity of treatment effect of selenium supplementation was observed for those who were risk allele carriers for either rs61901554 or rs12728998. For the candidate study of SNPs and hepatic steatosis, our participants of Mexican origin were, on average, 45 years old and were mostly female (63%) with a mean hepatic steatosis of 288.1 dB/m. Only PNPLA3 was statistically significantly associated with hepatic steatosis. We found no relationship between hepatic steatosis and the other SNPs and, no effect modification by BMI was observed. Conclusions: We identified two loci which were suggestively associated with metachronous adenoma. The identified loci in FAT3 and NLRP3 have biologically plausible mechanisms of action in adenoma development. The lack of heterogeneity of treatment effect of selenium supplementation for those who were at higher genetic 15 risk of adenoma development may be due to higher baseline levels of plasma selenium concentration than those observed historically in the literature. Finally, the lack of association between hepatic steatosis and five of the six SNPs previously found to be associated with NAFLD may be largely due to the inconsistent phenotypes which have been used by prior studies.