Metaproteomic Profile of the Colonic Luminal Microbiota From Patients With Colon Cancer

Abstract

Recent studies have provided evidence of interactions among the gut microbiota (GM), local host immune cells, and intestinal tissues in colon carcinogenesis. However, little is known regarding the functions exerted by the GM in colon cancer (CC), particularly with respect to tumor clinical classification and lymphocyte infiltration. In addition, stool, usually employed as a proxy of the GM, cannot fully represent the original complexity of CC microenvironment. Here, we present a pilot study aimed at characterizing the metaproteome of CC-associated colonic luminal contents and identifying its possible associations with CC clinicopathological features. Colonic luminal contents were collected from 24 CC tissue specimens immediately after surgery. Samples were analyzed by shotgun metaproteomics. Almost 30,000 microbial peptides were quantified in the samples, enabling the achievement of the taxonomic and functional profile of the tumor-associated colonic luminal metaproteome. Upon sample aggregation based on tumor stage, grade, or tumor-infiltrating lymphocytes (TILs), peptide sets enabling discrimination of sample groups were identified through discriminant analysis (DA). As a result, Bifidobacterium and Bacteroides fragilis were significantly enriched in high-stage and high-grade CC, respectively. Among metabolic functions, formate-tetrahydrofolate ligase was significantly associated with high-stage CC. Finally, based on the results of this pilot study, we assessed the optimal sample size for differential metaproteomic studies analyzing colonic luminal contents. In conclusion, we provide a detailed picture of the microbial and host components of the colonic luminal proteome and propose promising associations between GM taxonomic/functional features and CC clinicopathological features. Future studies will be needed to verify the prognostic value of these data and to fully exploit the potential of metaproteomics in enhancing our knowledge concerning CC progression.

Keywords: colon lumen; colorectal cancer; gut microbiota; metaproteome; tumor-infiltrating lymphocytes.

FIGURE 1

Taxonomic and functional profile of the tumor-associated colonic luminal metaproteome. (A) Tukey’s boxplots showing the top 20 microbial genera (left) and the top 20 species (right), ordered according to the median of the relative abundance (summed peptide intensity) distribution among patients. (B) Tukey’s boxplots showing the top 20 microbial functions (KOGs; left) and the top 20 pathways (right), ordered according to the median of the relative abundance (summed peptide intensity) distribution among patients.

FIGURE 2

Explained variance analysis. Bar graph reporting the percentage of metaproteome variance explained by clinical variables based on PCA results, using the abundance of microbial peptides (top) and host peptides (bottom) as input data.

FIGURE 3

Discrimination between sample groups based on clinicopathological features according to sparse partial least squares regression discriminant analyses (sPLS-DA). (A) Heatmaps illustrating hierarchical clustering of sample groups based on stage (top), grade (middle), and TILs (bottom) according to microbial discriminating peptides. (B) Heatmaps illustrating hierarchical clustering of samples groups based on stage (top), grade (middle), and TILs (bottom) according to host discriminating peptides.

FIGURE 4

Result of power analysis. Graph plotting the study power (probability to identify differential features) as a function of the number of patients analyzed per group. Three different log₂FC threshold values (corresponding to a small, medium, and high effect) were evaluated.