In a recent study published in the journal eBioMedicine, researchers performed a genome-wide interaction analysis to investigate whether genetic variants potentially modify the association between fiber, fruit, and vegetable intake and the risk of colorectal cancer (CRC). They successfully identified two significant loci that modify the association of fiber and fruit intake with CRC risk.

Study: Genome-wide interaction study of dietary intake of fibre, fruits, and vegetables with risk of colorectal cancer. Image Credit: Anatomy Image / Shutterstock

CRC is highly prevalent globally, with nearly two million incident cases and more than 900,000 deaths in 2020. High intake of fruits, vegetables, whole grains, and dietary fiber is reported to reduce CRC risk. While there is strong evidence for the association between CRC risk and the intake of whole grains and dietary fiber, there remains limited evidence for the association between CRC risk and the intake of fruits and vegetables.

Previous genome-wide association studies (GWAS) could identify at least 200 loci linked to CRC risk, accounting for up to 35% of heritability. Although gene-environment interactions (G × E) could potentially explain additional heritability, prior studies with small sample sizes and traditional methods found only limited significant gene-diet interactions. New statistical approaches, such as joint tests and two-step methods that prioritize single nucleotide polymorphisms (SNPs), may potentially improve the power of these analyses. Therefore, researchers in the present study aimed to identify new G × E interactions between the consumption of vegetables, fruits, and fiber and the risk of CRC by applying traditional and modern statistical approaches to a large dataset.

The analysis included up to 45 studies from three CRC genetic consortia comprising individuals of European ancestry. Cohort studies utilized nested case-control sets, while case-control studies employed cancer-free controls, mostly matched on demographic factors. Cases were confirmed as colorectal adenocarcinoma. In total, 69,599, 69,734, and 44,890 participants were analyzed for fruits, vegetables, and fiber consumption, respectively. Dietary intake was assessed via food frequency questionnaires and diet histories, typically quantified as servings per day for fruits and vegetables and grams per day for total fiber.

Data were harmonized and expressed as sex- and study-specific quartiles. Further, genotyping quality control included screening for missing call rates, Hardy-Weinberg equilibrium, and sex discordance, followed by imputation and filtering for minor allele frequency and accuracy, resulting in 7,250,911 SNPs for analysis. Genome-wide interaction scans were used to identify significant interactions using traditional logistic regression models and advanced techniques like the 3-DF joint test and the two-step approach. Additionally, interaction analyses for rare variants were conducted using the MiSTi (short for mixed effects score tests for interactions) method.

As compared to controls, participants with CRC were found to be older, had a higher body mass index and energy intake, and had a greater prevalence of risk factors like a family history of CRC and type 2 diabetes. They were also found to consume less fiber, fruits, and vegetables compared to controls. Meta-analyses revealed inverse associations between the intake of fiber (odds ratio per quartile increase (OR) = 0.79), fruits (OR = 0.79), and vegetables (OR = 0.82) and CRC risk. However, the inverse association did not vary by sex or cancer subsite. Stronger effects were observed in case-control studies as compared to cohort studies. Additionally, heterogeneity was observed in the analysis, primarily driven by case-control studies.

The 3-DF joint test identified the rs4730274 locus upstream of the SLC26A3 gene, which shows an association with fiber intake and interaction with CRC risk. Stratification by genotype revealed a stronger inverse association between fiber and CRC for each T allele copy. Functional annotation suggested enhancer activity in colon tissues, with eQTLs (short for expression quantitative trait loci) for the DLD gene. Interactions with SLC26A3 and DLD expression levels further supported these findings.

Further, the rs1620977 locus near the NEGR1 gene showed a significant association with fruit intake and a modest interaction with CRC risk. Strong inverse associations were observed with increased fruit intake for each G allele copy. However, no interaction effects were identified for vegetables using either the 3-DF or 2-step methods. Secondary analyses on rare variants and adjustments for additional confounders yielded non-significant results, indicating the robustness of the findings.

The study is strengthened by its large sample size, use of advanced statistical methods, and harmonized data quality control across all pooled studies. However, the study is limited by its reliance on single questionnaire measurements for dietary intake, inability to analyze fiber type, potential biases in case-control studies, and generalizability limited to European-ancestry populations.

In conclusion, the present study is the most extensive G × E study to date, where researchers identified two interactions between fiber, fruit intake, and CRC risk. Particularly, rs4730274 near the SLC26A3 gene suggests a significant link between fiber consumption, gut function, inflammation, and CRC. The findings warrant further research to examine the clinical implications and validate these findings across diverse populations.