Stool Microbiome Signature Associated with Response to Neoadjuvant Pembrolizumab in Patients with Muscle-invasive Bladder Cancer

Neoadjuvant pembrolizumab has been shown to be a valid treatment for patients affected by muscle-invasive bladder cancer (MIBC), as demonstrated in the PURE-01 clinical trial (NCT02736266). Among the tumor-extrinsic factors influencing immunotherapy efficacy, extensive data highlighted that the microbiome is a central player in immune-mediated anticancer activity. This report aimed to investigate the composition and role of stool microbiome in patients enrolled in the PURE-01 clinical trial. An orthotopic animal model of bladder cancer (MB49-Luc) was used to support some of the findings from human data. An analysis of stool microbiome before pembrolizumab was conducted for 42 patients, of whom 23 showed a pathologic response. The information in the preclinical model of orthotopic bladder cancer treated with anti–PD-1 antibody or control isotype was validated. Linear discriminant analysis effect size and linear models were used to identify the bacterial taxa enriched in either responders or nonresponders. The identified taxa were also tested for their association with event-free survival (EFS). Survival at 31 d after tumor instillation was used as the study endpoint in the preclinical model. Responders and nonresponders emerged to differ in terms of enrichment for 16 bacterial taxa. Of these, the genus Sutterella was enriched in responders, while the species Ruminococcus bromii was enriched in nonresponders. The negative impact of R. bromii on anti–PD-1 antibody activity was also observed in the preclinical model. EFS and survival of the preclinical model showed a negative role of R. bromii. We found different stool bacterial taxa associated with the response or lack of response to neoadjuvant pembrolizumab. Moreover, we provided experimental data about the negative role of R. bromii on immunotherapy response. Further studies are needed to externally validate our findings and provide mechanistic insights about the host-pathogen interactions in MIBC. Using prepembrolizumab stool samples collected from patients enrolled in the PURE-01 clinical trials, we identified some bacterial taxa that were enriched in patients who either responded or did not respond to immunotherapy. Using an animal model of bladder cancer, we gathered further evidence of the negative impact of the Ruminococcus bromii on immunotherapy efficacy. Further studies are needed to confirm the current findings and test the utility of these bacteria as predictive markers of immunotherapy response.