Background: Bile acids (BA) and the gut microbiome-derived BA metabolites (secondary BA [SBA]) modulate immune functions and contribute to intestinal tumorigenesis. While diet is known to influence BA production, the findings in humans remain inconsistent. We hypothesize that the impact of diet on BA, particularly SBA levels, may depend on the host’s gut microbial features.

Methods: Leveraging the integrated stool metagenomes, stool and plasma metabolome, and long-term habitual dietary data from a subset of the Microbiome Among Nurses study (Micro-N), including Micro-N-diverticulitis substudy (n=225), Micro-N-polyp substudy (n=400), and Mind-Body Study (MBS) (n=220), we explored the interplay between gut microbial species and functions, stool and blood BA metabolites, and diet (41 food groups and 51 nutrients).

Results: A total of 25 individual BA metabolites were detected in stool, while 13 of them were found in blood. Random Forest regression analysis of individual BA metabolites against 384 microbial species in our samples revealed that microbial species were most strongly associated with the microbial metabolites of dehydro-LCA and isoallo-LCA. Among the individual microbes, Methanobrevibacter smithii was the top contributor, with findings validated in independent cohorts. Functional and comparative genetic sequence analysis revealed that M. smithii might not directly biotransform dehydro-LCA and isoallo-LCA, but instead promote dehydrogenation and shift bacterial fermentation toward more oxidized products through production of methane gas via methanogenesis (METHANOGENESIS-PWY: methanogenesis from H2 and CO2). Among dietary predictors, 41 food groups accounted for only up to 9% of the variance in individual BA metabolites in stool. However, significant heterogeneity was observed between carriers and non-carriers of the methanogenesis pathway: diet explained up to 40% of the variance in BA metabolites among carriers. This interaction was consistent across analyses of both food groups and nutrients. Notably, distinct associations between certain nutrients of different sources and LCA metabolites were found among carriers of the methanogenesis pathway. For example, vegetable fat and polyunsaturated fat were negatively correlated, while animal fat and trans-fat were positively correlated with LCA metabolites. Adding the methanogenesis pathway abundance to the dietary and demographic characteristics substantially increased the predictive accuracy of high dehydro-LCA levels (top quartile), from 60% to 90%.

Conclusions: The methanogenesis pathway encoded primarily by M. smithii and the resultant production of methane gas may influence the biotransformation of dehydro-LCA and isoallo-LCA in gut microbial communities and modulate the dietary effect on BA metabolism.