Differential susceptibility to colorectal cancer due to naturally occurring gut microbiota

Abstract

Recent studies investigating the human microbiome have identified particular bacterial species that correlate with the presence of colorectal cancer. To evaluate the role of qualitatively different but naturally occurring gut microbiota and the relationship with colorectal cancer development, genetically identical embryos from the Polyposis in Rat Colon (Pirc) rat model of colorectal cancer were transferred into recipients of three different genetic backgrounds (F344/NHsd, LEW/SsNHsd, and Crl:SD). Tumor development in the pups was tracked longitudinally via colonoscopy, and end-stage tumor burden was determined. To confirm vertical transmission and identify associations between the gut microbiota and disease phenotype, the fecal microbiota was characterized in recipient dams 24 hours pre-partum, and in Pirc rat offspring prior to and during disease progression. Our data show that the gut microbiota varies between rat strains, with LEW/SsNHsd having a greater relative abundance of the bacteria Prevotella copri. The mature gut microbiota of pups resembled the profile of their dams, indicating that the dam is the primary determinant of the developing microbiota. Both male and female F344-Pirc rats harboring the Lewis microbiota had decreased tumor burden relative to genetically identical rats harboring F344 or SD microbiota. Significant negative correlations were detected between tumor burden and the relative abundance of specific taxa from samples taken at weaning and shortly thereafter, prior to observable adenoma development. Notably, this naturally occurring variation in the gut microbiota is associated with a significant difference in severity of colorectal cancer, and the abundance of certain taxa is associated with decreased tumor burden.

Keywords: Pirc; colorectal cancer; gut; microbiota; rat.

Figure 1. The gut microbiota of Lewis rats differs qualitatively from that of F344 and SD rats

Bar charts showing the relative abundance of microbes in the gut microbiota of F344 (n = 6), SD (n = 2), and Lewis (n = 5) surrogate dams 24 hours pre-partum, annotated to the level of phylum (A. legend at right) and operational taxonomic unit B.; Principal component analysis of the samples depicted in A and B, showing complete separation of Lewis dams from F344 and SD dams along principal component 1 C.

Figure 2. Rederived pups acquire the gut microbiota of their surrogate dam

Bar charts showing the OTU relative abundance of representative F344, SD, and Lewis dams and three F344-Pirc pups born to those dams at weaning (21 days), 1 month (1m), 1.5m, 3m, 4.5m, and 6m of age A. Principal component analysis (PCA) of all surrogate dams (n = 13) and all male F344-Pirc pups born to those dams (n = 21) at 1.5 months of age (B. legend at right).

Figure 3. Rats colonized with GM_Lewis develop less severe disease than rats harboring GM_SD or GM_F344

Dot plots demonstrate the mean (± SD) small intestinal (SI) tumor count (A., E.), colonic tumor count (B., F.), total colonic area affected by tumors (C., G.), and mean colonic area of individual tumors (D., H.) in male (A-D) and female (E-H) F344-Pirc rats colonized with the GM acquired from their cognate surrogate dam. P values denote results of one way ANOVA (males) or t-test (females) and asterisks indicate p ≤ 0.05.

Figure 4. Contribution of individual OTUs to the variation between GM profiles

Loadings plot of the PCA data depicted in Figure 2, showing the relative contribution of all operational taxonomic units (OTU) to principal component 1 (PC1) and PC2, OTUs located greater than 0.1 eigenvalue from the origin are labeled with font size corresponding to overall abundance in the dataset.

Figure 5. The Lewis GM undergoes a progression distinct from that of F344 or SD

Bar charts demonstrating the progression in terms of relative abundance (mean ± SEM) of Prevotellaceae A., Ruminococcaceae B., Lachnospiraceae C., Lactobacillaceae D., and Enterobacteriaceae E. at 21 days (21d: weaning), 1 month (1m), 1.5m, 3m, 4.5m, and 6m of age in isogenic male F344-Pirc rats colonized with GM_F344 (n = 8), GM_SD (n = 5), and GM_Lewis (n = 7 or 8). Like letters indicate significant (p ≤ 0.05) differences; 2-way ANOVA with post hoc pairwise comparisons via the Holm-Sidak method. There was a significant main effect of time on the relative abundance of all taxa analyzed; only significant differences within GM are denoted for the sake of clarity. No significant interactions were detected between GM and age for any taxon analyzed.

Figure 6. Correlations between relative abundance of select OTUs early in life and terminal tumor burden

Scatter plots showing significant negative correlations detected between the relative abundance of Lactobacillus sp. at 1m A. or 3m B. of age and total number of colonic tumors A. or total colonic area affected by tumors B. at necropsy, or the relative abundance of P. copri at 1.5 C. or 3m D. of age and mean area occupied by individual colonic tumors at necropsy. P-values determined via Spearman rank order correlation.