Gut microbiota ablation depletes secondary bile acids and shifts the cytokine balance in BCG-infected mice

Abstract Background. Antibiotic ablation is the standard test of whether a host phenotype depends on the gut microbiota, but it removes the community and perturbs the host at once, so depletion alone cannot assign a metabolite shift to the community. We built a four-arm mouse model on a single BCG infection background: untreated controls (NC), BCG alone, BCG after mixed-antibiotic ablation (ABX_BCG), and BCG followed by faecal microbiota transplantation with no antibiotic exposure at any point (BCG_FMT). The ablated and transplanted arms were independent branches, not a depletion-then-restoration sequence. Results. Dual-polarity untargeted LC-MS (9,706 negative-mode and 11,589 positive-mode features) and shotgun metagenomics showed that ablation left a block of 726 negative-mode and 313 positive-mode features absent from all three ABX_BCG samples but detected in 88–94% of the other nine; total ion signal in the ablated arm exceeded that of the quality-control injections, excluding dilution and matrix loss. Twenty-seven block features were annotated, resolving to 17 unique InChIKeys, 21 of them microbial compound classes including secondary bile acids. Deoxycholic acid eluted at 437.93 s in negative mode and 437.28 s in positive mode; deoxycholic and lithocholic acid were undetected in the ablated samples. Both omics layers resolved the same four groups (PERMANOVA R² = 0.562 for the genus table and 0.635 for the metabolome) yet failed to couple at sample level; exhaustive within-group re-pairing excluded sample mismatch, identifying a shared treatment rather than an individual axis. Ablation accompanied higher TNF-α and IL-6 and lower IFN-γ and IL-10 in bronchoalveolar lavage fluid and serum alike. Conclusions. The gut community contributes a set of metabolite features whose absence is specific to the ablated state rather than shared with infection or transplantation, and the same arm occupies the extreme of the cytokine readout in both compartments. Running the ablated and transplanted arms as independent branches off the infected baseline is what makes this an antibiotic-specific control, and the two omics layers carry a shared treatment structure rather than a shared individual axis.