THE ROLE OF THE COLON ON THE EFFECT OF DIET ON THE HUMAN PLASMA AND GUT METABOLOME

Gut microbiome dysbiosis, defined as departure from community configurations observed in healthy control subjects, is a characteristic of some but not all patients with Inflammatory Bowel Disease patients. Metabolomic signatures of dysbiosis are more pronounced than taxonomic differences in the gut microbiome, suggesting that key metabolic roles of the gut microbiome may be disrupted in the setting of dysbiosis.

In Phase 1, we analyzed fecal metabolomic data from a controlled feeding and gut microbiome depletion study (Tanes, C. et al. Cell Host & Microbe 2021) to differentiate diet-derived fecal metabolites from those produced or consumed by the gut microbiome. Of 476 named metabolites, we identified 75 metabolites produced by the microbiome and 91 metabolites consumed by the microbiome, controlling for a false discovery rate of 10% (Figure 1). The set of microbiome-derived metabolites included widely recognized compounds such as butyrate and lithocholate, but included other molecules, such as benzoate, that are not normally considered to be microbial products in the gut. In the context of a depleted gut microbiome, we identified 56 metabolites derived from a standardized omnivore diet relative to exclusive enteral nutrition. These included a variety of carnitines and nicotinate, a B vitamin.

In Phase 2, we applied our metabolite classification results to data from the Integrative Human Microbiome Project (HMP2), to determine the diet- and microbiome-derived metabolites that were associated with episodes of dysbiosis (Figure 2). Of the metabolites linked to the microbiome in Phase 1, 74% were significantly increased or decreased in microbiome samples classified as dysbiotic relative to healthy controls. A majority of the diet-derived gut metabolites also differed in dysbiotic samples, including 8 metabolites that our analysis did not categorize as substrates or products of the microbiome.

In total, our comprehensive analysis of untargeted fecal metabolomics data across two studies revealed diet- and microbiome-derived metabolites associated with dysbiosis. Our results both reinforce our current understanding of the biological roles of many gut metabolites and offer an opportunity to re-frame our understanding of others. Future work will focus on identifying the microbes driving variation among dysbiosis-associated metabolites and quantifying the metabolomic consequences of dietary choices by patients with inflammatory bowel disease.

Background: A primary function of the colonic microbiota, due to its large biomass, is metabolism of intestinal contents. Studies in germ-free vs. conventionally housed mice show that the gut microbiota has a substantial impact on both the fecal and plasma metabolome.. Evidence for the impact of the colonic microbiota in humans is lacking. Herein, we characterized the impact of two divergent diets on the luminal and plasma metabolome in humans with and without a colon. Methods: Untargeted plasma and gut luminal metabolomics were performed by LC-MS in humans consuming an omnivore (OMV) diet followed by an exclusive enteral nutrition (EEN) liquid formula diet. Analyses compared participants with an ileostomy to those with intact colon to define colonic metabolic function. Results: The study included 10 healthy controls and 10 participants with an ileostomy. PCA analysis revealed that the ileostomy metabolome was highly divergent from the fecal metabolome at baseline on an OMV diet although both responded to EEN in a similar fashion. The plasma metabolome was also divergent from ileostomy patients compared to healthy controls at baseline on an OMV diet where intersubject variability was greater at baseline and the response to an EEN diet was more variable in patients with an ileostomy. Linear mixed effects modeling of approximately 1,000 identified metabolites comparing the two cohorts (ileostomy vs. control) and two diets (OMV vs. EEN) in both stool and plasma revealed that the cohort effect dominated diet in the luminal metabolome whereas the relative abundance of many metabolites in ileostomies relative to feces were dramatically greater despite a minimal difference in the plasma metabolome. By contrast, diet dominated the cohort effect in the plasma where the relative abundance of many metabolites were greater on the OMV than EEN diet. An example of a diet dependent host microbiota-metabolite interaction is the dietary xenobiotic piperine that undergoes host-dependent glucuronide and sulfate conjugation and are elevated in ileostomies but are undetectable in feces due to microbiota deconjugation reactions. Finally, by comparing the coefficient of variation of plasma metabolites, we show that that there is greater intersubject variability in those with an ileostomy than healthy controls due to metabolites mostly of non-microbial origin primarily on an EEN diet. Conclusions: In total, these results show that a large number of metabolites are not absorbed in the small intestine and are normally delivered into the colon where they undergo metabolism by the gut microbiota. The resulting products have a much smaller impact on plasma metabolome composition than diet. Nevertheless, by metabolizing small molecules of host and/or diet origin, the gut microbiota may play a role in regulating intersubject variability in the human plasma metabolome.