Concordant Metabolite Differences Across Cecal Content, Liver, and Plasma in Mice Fed an Obesogenic Diet Containing Common Bean (Phaseolus vulgaris L.): A Multi-Compartment Metabolomics Study

Background/Objectives: How metabolic changes accompanying common bean (Phaseolus vulgaris L.) consumption are organized across gut, liver, and blood, and which are separable from lower adiposity, is unclear. Methods: Male C57BL/6J mice were fed an obesogenic diet with or without cooked bean powder supplying 70% (nominal) of total dietary protein for 14 weeks; diets were assigned to whole cages (4 per diet). Diets were formulated to nominally match macronutrient proportions (w/w) from ingredient and supplier values; sources, and hence estimated protein, amino acid, α-linolenate, and fiber supply, differed; intake was not measured. Cecal content, liver, and plasma were profiled by untargeted metabolomics and liver by RNA-seq for selected gene sets. Metabolite classes selected by three-compartment concordance were examined against prior predictions. Results: Diet groups separated in every compartment, most strongly in cecal content. In all three, 36 host metabolites differed concordantly and 22 KEGG pathways were significant. Five classes met the selection criteria: branched-chain amino acids and their keto acids accumulated in cecal content despite lower estimated supply; long-chain n-3 and n-6 fatty acids shifted reciprocally while hepatic Fads2, Elovl5, and Fads1 transcripts were 45–61% of control; and glutathione-related metabolites, ceramides, and bile acids differed as predicted. Total depot mass was significantly lower only in the full cohort (−12%). After MANCOVA adjustment for depot redistribution and hepatic lipid, the branched-chain class retained its diet association (q-value = 0.012), the glutathione-related (q-value = 0.047) and n-3 (q-value = 0.049) classes did so provisionally, the former and the hepatic component of the latter depending on one cage in cage-aware models, and ceramides (q-value = 0.20) and bile acids (q-value = 0.15) did not. Conclusions: Whether these differences drive or follow the leaner phenotype requires time-course, single-constituent designs in both sexes.