DIET AND THE GUT MICROBIOME IN RELATION TO FECAL BILE ACID METABOLITES
Methods: We utilized prospective cohort of 116,429 women enrolled in the Nurses’ Health Study II who have provided validated dietary and lifestyle data every 2-4 years since 1989. Since late 2019, we started collecting stool samples from a subset of the NHSII women (the Micro-N Study). Using the integrated fecal metagenome and untargeted metabolome data of 242 women with detailed dietary history data in the Micro-N Study, we analyzed the associations of 41 food groups, 56 nutrients, and 639 gut microbial species profiled by MetaPhlAn3 with fecal BA metabolites.
Results: The median (IQR) age at stool collection was 65 (62-68) years. A total of 25 types of PBA and SBA were detected and categorized into 6 groups based on their metabolic pathways, including unconjugated PBA, conjugated PBA, reconjugated PBA, unconjugated SBA, conjugated SBA, and SBA derivatives (Fig1a-b). Correlation analysis of the metabolites demonstrated a strong clustering within each of the 6 groups and an inverse correlation of SBA derivatives with the others (Fig1c). MaAsLin2 showed that certain BA metabolites, e.g., SBA derivatives and reconjugated PBA, were more microbially associated than others. Among them, LCA derivatives, particular isoallo-LCA, showed the most microbial associations, and the associations were largely in an opposite direction to the other BA groups. As expected, certain microbes encoding BA biotransformative enzymes/genes, e.g., Clostridium leptum that carries Bai genes and 12a-hydroxysteroid dehydrogenases, were associated with higher levels of SBA derivatives (Fig1d). In terms of diet, dark-yellow vegetables (R2 =1.8%, p=0.019), tomatoes (R2 =1.6%, p=0.017), and fruit (R2 =1.2%, p=0.040) were the three food groups contributing most to the variance in overall BA community, with potential nutrient drivers to be fruit fiber (R2 =1.6%, p=0.027), fruit carbohydrate (R2 =1.5%, p=0.023), and vegetable carbohydrate (R2 =1.4%, p=0.025) (Fig2a). When we investigated the associations with individual BA metabolites, fruit showed a wide-ranging inverse association with unconjugated PBA, unconjugated SBA, and SBA derivatives; snacks, tomatoes, and French fries showed largely positive associations with conjugated and unconjugated PBA (Fig2b).
Conclusions: Different subgroups of PBA and SBA showed distinct associations with gut microbiome and dietary factors; LCA and its derivatives were the most gut microbiome-related and might be inversely modulated by fruits and vegetables.
Figure 1. Six groups of bile acid and bile acid metabolites detected in stool samples and their associations with gut microbial species. a) Categorization of the 25 types of bile acid and bile acid metabolites into 6 groups of primary bile acid (PBA) and secondary bile acid (SBA) based on metabolic pathways. b) Levels of the 25 bile acid and bile acid metabolites in 242 stool samples. c) Inter-correlations of the 25 individual bile acid metabolites and 6 bile acid groups, which suggested that SBA derivatives were distinct from other BA groups. d) Associations of the 25 individual bile acid and bile acid metabolites with gut microbial species, which suggested that Isoallo-LCA was the most microbial correlated.
Figure 2. Associations of food groups and nutrients with overall bile acidic community and individual bile acid metabolites. a) Contributions from 41 food groups and 56 nutrients to overall variance in the BA metabolites using Bray-Curtis distance. b) Associations of the food groups and nutrients with individual PBA and SBA.
Presenter
Speakers
Related Products
