Abstract

Despite the encouraging efficacy of anti-PD-1/PD-L1 immunotherapy in microsatellite instability-high/deficient mismatch repair (MSI-H/dMMR) advanced gastrointestinal cancer, many patients exhibit primary or acquired resistance. Using multi-omics approaches, we interrogated gut microbiome, blood metabolome and cytokines/chemokines of MSI-H/dMMR advanced gastrointestinal cancer patients (N=77) and identified a number of microbes (e.g. Alistipes putredinis) and metabolites (e.g. arginine and SCFA) highly associated with primary resistance. Fecal microbiota transplantation of patients stool replicated the clinical responsiveness as well as certain molecular signatures. Based on the clinical microbiome data, we developed a predictive machine learning model for primary resistance and achieved accuracy at 0.83 on an external validation set. Furthermore, several microbes were pinpointed which gradually changed during the process of acquired resistance. In summary, our study demonstrated the essential role of gut microbiome in drug resistance, and this could be utilized as a preventative diagnosis tool as well as therapeutic targets in the future.