Reduced type II interleukin-4 receptor signalling drives initiation, but not progression, of colorectal carcinogenesis: evidence from transgenic mouse models and human case-control epidemiological observations.

We investigated the role of interleukin (IL)-4 receptor (IL-4R) signalling during mouse carcinogen-induced colorectal carcinogenesis and in a case-control genetic epidemiological study of IL-4Rα single nucleotide polymorphisms (SNPs). Azoxymethane-induced aberrant crypt focus (ACF; 6 weeks) and tumours (32 weeks) were analysed in wild-type (WT) BALB/c mice, as well as in IL-4Rα (-) (/-) , IL-13 (-/-) and 'double-knockout' (DKO) animals. Colorectal cancer (CRC) cases (1502) and controls (584) were genotyped for six coding IL-4Rα SNPs. The association with CRC risk and CRC-specific mortality was analysed by logistic regression. Lack of IL-4Rα expression was associated with increased ACFs [median 8.5 ACFs per mouse (IL-4Rα (-/-) ) versus 3 (WT); P = 0.007], but no difference in the number of colorectal tumours [mean 1.4 per mouse (IL-4Rα (-/-) ) versus 2 (WT)], which were smaller and demonstrated reduced nuclear/cytoplasmic ß-catenin translocation compared with WT tumours. Tumour-bearing IL-4Rα (-/-) mice had fewer CD11b(+)/Gr1(+) myeloid-derived suppressor splenocytes than WT animals. IL-13 (-/-) mice developed a similar number of ACFs to IL-4Rα (-/-) and DKO mice. There was a significant increase in CRC risk associated with the functional SNP Q576R [odds ratio 1.54 (95% confidence interval 0.94-2.54), P trend 0.03 for the minor G allele]. There was no effect of IL-4Rα genotype on either CRC-specific or all-cause mortality. These combined pre-clinical and human data together demonstrate that reduced IL-4R signalling has stage-specific effects on colorectal carcinogenesis (increased CRC initiation and risk but reduced tumour progression and no effect on CRC mortality). These results should prompt evaluation of the effect of pharmacological manipulation of IL-4R signalling on future CRC risk and for CRC treatment.

Polymorphisms in xenobiotic metabolizing enzymes and diet influence colorectal adenoma risk.

OBJECTIVES: We have earlier shown that diet and xenobiotic metabolizing enzyme genotypes influence colorectal cancer risk, and now investigate whether similar associations are seen in patients with premalignant colorectal adenomas (CRA), recruited during the pilot phase of the Scottish Bowel Screening Programme. METHODS: Nineteen polymorphisms in 13 genes [cytochrome P450 (P450), glutathione S-transferase (GST), N-acetyl transferase, quinone reductase (NQ01) and microsomal epoxide hydrolase (EPHX1) genes] were genotyped using multiplex PCR or Taqman-based allelic discrimination assays and analyzed in conjunction with diet, assessed by food frequency questionnaire, in a case-control study [317 CRA cases (308 cases genotyped), 296 controls]. Findings significant at a nominal 5% level are reported. RESULTS: CRA risk was inversely associated with fruit (P=0.02, test for trend) and vegetable (P=0.001, test for trend) consumption. P450 CYP2C9*3 heterozygotes had reduced CRA risk compared with homozygotes for the reference allele [odds ratio (OR): 0.60; 95% confidence interval (CI): 0.36-0.99], whereas CYP2D6*4 homozygotes (OR: 2.72; 95% CI: 1.18-6.27) and GSTM1 'null' individuals (OR: 1.43; 95% CI: 1.04-1.98) were at increased risk. The protective effect of fruit consumption was confined to GSTP1 (Ala114Val) reference allele homozygotes (OR: 0.49; 95% CI: 0.34-0.71, P=0.03 for interaction). CRA risk was not associated with meat consumption, although a significant interaction between red meat consumption and EPHX1 (His139Arg) genotype was noted (P=0.02 for interaction). CONCLUSION: We report the novel associations between P450 genotype and CRA risk, and highlight the risk association with GSTM1 genotype, common to our CRA and cancer case-control series. In addition, we report a novel modifying influence of GSTP1 genotype on dietary chemoprevention. These novel findings require independent confirmation.

Methylene tetrahydrofolate reductase genotype modifies the chemopreventive effect of folate in colorectal adenoma, but not colorectal cancer.

Epidemiologic evidence suggests a role for folate, a critical component of the 1-carbon cycle, in colorectal adenoma and cancer pathogenesis. Low folate levels, along with genetic polymorphisms in key enzymes such as methylene tetrahydrofolate reductase (MTHFR), can cause DNA hypomethylation and aberrant CpG methylation, which have been associated with colorectal tumor development. We investigated self-reported folate and alcohol intake alongside possible modifying effects of MTHFR 677 C>T and 1298 A>C polymorphisms in UK case-control studies of colorectal adenoma (317 cases, 296 controls) and cancer (500 cases, 742 controls). A significant association between MTHFR 1298 and colorectal cancer risk was observed [odds ratio, 1.57; 95% confidence interval (95% CI), 1.05-2.37], which was more pronounced in males (odds ratio, 3.02; 95% CI, 1.63-5.62). Although we found no association between MTHFR 677 and colorectal cancer, when data were stratified by sex, an increased risk was seen in females (odds ratio, 1.96; 95% CI, 1.11-3.46) but not in males. High folate intake was associated with a decreased risk for colorectal adenoma (odds ratio, 0.47; 95% CI, 0.30-0.73; P(trend), <0.001), which was modified by MTHFR 1298 genotype (P(interaction) = 0.006). However, we found no evidence to support the hypothesis that a high-folate diet protects against colorectal cancer development. Consistent with previous studies, high alcohol intake (>or=14 U/wk) was associated with a significantly increased cancer risk (odds ratio, 2.57; 95% CI, 1.81-3.64). Our data suggest that dietary folate intake may be an important determinant for premalignant colorectal disease development but not colorectal cancer, an association that is modified by MTHFR genotype.

Recurrent Coding Sequence Variation Explains Only A Small Fraction of the Genetic Architecture of Colorectal Cancer.

Whilst common genetic variation in many non-coding genomic regulatory regions are known to impart risk of colorectal cancer (CRC), much of the heritability of CRC remains unexplained. To examine the role of recurrent coding sequence variation in CRC aetiology, we genotyped 12,638 CRCs cases and 29,045 controls from six European populations. Single-variant analysis identified a coding variant (rs3184504) in SH2B3 (12q24) associated with CRC risk (OR = 1.08, P = 3.9 × 10(-7)), and novel damaging coding variants in 3 genes previously tagged by GWAS efforts; rs16888728 (8q24) in UTP23 (OR = 1.15, P = 1.4 × 10(-7)); rs6580742 and rs12303082 (12q13) in FAM186A (OR = 1.11, P = 1.2 × 10(-7) and OR = 1.09, P = 7.4 × 10(-8)); rs1129406 (12q13) in ATF1 (OR = 1.11, P = 8.3 × 10(-9)), all reaching exome-wide significance levels. Gene based tests identified associations between CRC and PCDHGA genes (P < 2.90 × 10(-6)). We found an excess of rare, damaging variants in base-excision (P = 2.4 × 10(-4)) and DNA mismatch repair genes (P = 6.1 × 10(-4)) consistent with a recessive mode of inheritance. This study comprehensively explores the contribution of coding sequence variation to CRC risk, identifying associations with coding variation in 4 genes and PCDHG gene cluster and several candidate recessive alleles. However, these findings suggest that recurrent, low-frequency coding variants account for a minority of the unexplained heritability of CRC.