Abstract 5734: Gut microbiota predicts response and toxicity with neoadjuvant immunotherapy

Abstract Introduction: Checkpoint blockade immunotherapy has revolutionized melanoma treatment. However, immune related adverse events (irAEs) are frequent, particularly when PD-1 and CTLA-4 blocking antibodies are combined. There is an urgent need for biomarkers to predict response as well as toxicity. The intestinal microbiome influences immune processes throughout the body and is therefore hypothesised to influence both immunotherapy response and well as development of irAEs. Methods: This study utilized sequencing to profile the fecal microbiome from Stage III melanoma patients (n= 38) treated with combination anti-PD1 and anti-CTLA4 antibodies in the neoadjuvant setting (OpacinNeo Trial; NCT02977052) at Melanoma Institute Australia. Absolute bacterial loads were determined with qPCR. Global microbial gene abundance was determined through metagenomic sequencing, allowing analysis of microbial metabolic capcity, and actual short chain fatty acid output (butyrate, acetate, propionate) was assessed by NMR. Matched PBMC were profiled in detail using mass cytometry and immune populations were correlated with microbial metrics. We tested the ability of machine learning to predict therapeutic response and irAEs from microbial data. Results: Patients who failed to respond (n=6), and those who developed severe (G3-G5) irAEs (n=11), had low microbial diversity before treatment (p=0.0234 and p=0.0226 respectively). Four of the six non-responders in this cohort failed to complete treatment and received immunosuppression before pathologic assessment of response, because of the development of severe irAEs. Thus, non-response may, in fact, have been a consequence of irAE development in those cases. High pre-treatment loads of Ruminococcaceae and methanogenic archaea, and greater relative abundance of butyrate production pathways, were significantly associated with response and protection from irAEs. Patients who had low bacterial diversity had increased proportions of effector and memory CD4 and CD8 T cells in the blood and increased numbers of circulating MAIT, NK and gdT cells prior to immunotherapy. Importantly, microbial taxa were predictive of patients who subsequently failed to respond to therapy while also developing severe irAEs with overall accuracy of 92.1%. Conclusions: A gut ecosystem enriched with beneficial microbial functions including production of butyrate is associated with protection from irAEs with neoadjuvant combination immunotherapy. The findings allow prediction of outcomes with combination immunotherapy and will inform development of therapeutic microbial interventions. Citation Format: Marcel Batten, Erin Shanahan, Rebecca Simpson, Mark Read, Ines P. Silva, Alexandra Angelatos, Jian Tan, Chandra Adhikari, Alexander M. Menzies, Robyn P. Saw, Laurence Macia, Maria Gonzalez, Kerwin Shannon, Rebecca Velickovic, Irene L. Reijers, Christian U. Blank, James S. Wilmott, Andrew J. Holmes, Richard A. Scolyer, Georgina V. Long. Gut microbiota predicts response and toxicity with neoadjuvant immunotherapy [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5734.