Abstract
BACKGROUND: Observational studies have shown that TNF-related apoptosis-inducing ligand (TRAIL) levels and vascular endothelial growth factor (VEGF) levels are associated with colon cancer risk, and TRAIL antagonizes VEGF-induced tumor angiogenesis. We aim to determine whether there is a causal relationship between TRAIL/VEGF levels and colon cancer by Mendelian randomization analysis. METHODS: Genome-wide association studies (GWAS) of 21,758 individuals of European ancestry were conducted to identify single nucleotide polymorphisms (SNPs) associated with TRAIL/VEGF. MR-Egger regression, heterogeneity test, pleiotropy test and leave-one-out sensitivity tests were used for sensitivity analysis, in addition to inverse variance weighting (IVW) as the main Mendelian-randomization (MR) analysis method. RESULTS: Univariate MR analysis using IVW indicated an inverse association between genetically predicted TRAIL levels and colon cancer risk (OR = 0.805, 95%CI = 0.681-0.953, P = 0.011), while VEGF levels showed a positive association (OR = 1.337, 95%CI = 1.163-1.536, P < 0.001). In addition, the MR Steiger test showed that the effect of TRAIL/VEGF levels on colon cancer were correct causality (P < 0.001). In multivariate MR analysis, TRAIL levels had an independent causal effect on colon cancer risk after controlling for VEGF levels (OR = 0.802, P = 0.002); likewise, the causal effect of VEGF levels on colon cancer risk remained after controlling for TRAIL levels (OR = 1.328, P < 0.001). CONCLUSIONS: There was a inverse association between genetic variants in TRAIL levels and colon cancer risk, but a positive association between genetic variants in VEGF levels and colon cancer risk. Further research is needed to explore the mechanisms behind this putative causal relationship.