Bacteroides intestinalis mediates the sensitivity to irinotecan toxicity via tryptophan catabolites

Background: Late-onset diarrhoea remains a poorly managed concern for clinical irinotecan therapy. Although bacterial β-glucuronidases (β-GUS) mediated SN-38 production is prevailingly thought to mediate intestinal toxicity, β-GUS inhibitors confer limited benefits in the clinic. Objective: This study aimed to explore the role and mechanism of endogenous bacterial metabolites in susceptibility to irinotecan toxicity. Design: Gut microbiota profiles and metabolites in patients with colorectal cancer (CRC) with or without diarrhoea were investigated via 16S rRNA sequencing, shotgun metagenomics and metabolomics. The role of microbial metabolites was investigated in mice by metabolic bioengineering and intestinal organoid culture. The mechanism of microbial metabolites on intestinal stem cells was investigated by transcriptional profiling and chemical intervention. Results: Gut microbial configuration was differentially remodelled in diarrhoea and non-diarrhoea patients with irinotecan therapy, and the susceptibility was transmissible to recipient mice via transplantation of baseline faecal microbiome. Bacteroides intestinalis (B. intestinalis) was notably expanded in the diarrhoea-prone cohorts as well as in irinotecan-treated mice. B. intestinalis colonisation sensitised intestinal epithelia to irinotecan-induced chemical injury, partially via tryptophan metabolite indole-3-acetate (IAA). Both B. intestinalis and bioengineered bacteria that produce IAA exacerbated irinotecan-induced intestinal epithelial injury in mice. Mechanistically, IAA suppressed PI3K-Akt signalling, thereby impairing the renewal of intestinal epithelia under the insult of irinotecan. In clinical patients receiving irinotecan therapy, faecal IAA level was closely associated with the diarrhoea severity. Conclusion: Our study uncovers the mechanism of endogenous bacterial metabolite in shaping the individual susceptibility to irinotecan toxicity and suggests IAA as a potential predictive biomarker.