不同的肠道微生物群落与黑色素瘤基线免疫特征、肿瘤特征和ICI预后的差异相关

We have previously identified links between gut microbial community assemblages, defined by the ratio of Ruminococcaceae to Bacteroidaceae, and response to immune checkpoint inhibitor (ICI) immunotherapy, suggesting microbiome-immune interactions may prime patients for treatment outcomes. To investigate how these microbial states shape baseline immune tone and anti-tumor immunity we profiled peripheral immune and tumor-intrinsic features across these stratified patient groups. In 129 Australian and Dutch stage III melanoma patients treated on trial with neo-adjuvant anti-PD-1/anti-CTLA-4 immunotherapy (OpACIN-neo/PRADO), pre-treatment stool (microbiome; 16S rRNA gene sequencing) and PBMCs (immune profiles; mass cytometry) were analyzed. Tumor biopsies were subject to DNA/RNA sequencing, with tumor mutation burden (TMB) and IFN-gamma score assessed. Clear differences in baseline peripheral immune profiles were observed upon stratifying patients by microbial community types. Ruminococcaceae (Ru)-dominated gut microbiomes were associated with higher ICI response rates, had greater frequencies of circulating memory B and T cell subsets (CD27+ B cells, CD8+ Tem (CD45RA-CCR7-)) and higher ICOS+ Tregs. Bacteroidaceae (Ba)-dominated patients had higher frequencies of total B cells and a more naïve immune profile. Further, the pre-treatment tumors of Ru-dominated patients were enriched with pathways involved with immune signalling and antigen presentation while Ba-dominated patient tumors were enriched with pathways associated with dysregulated metabolism and cell proliferation. The distribution of tumor TMB and IFN-gamma groups between responders (R) and non-responders (NR) within each microbial community type was also significantly different; 81% of Ba-R had high TMB, compared to only 54% of Ru-R, and 72% of Ba-dominated patients belonged to the IFN-gamma low group compared to 48% of Ru-dominated patients, with all Ba-NR being IFN-gamma low. Gene ontology enrichment analysis further revealed distinct biological programs associated with response within each microbial community type. Ru-R were enriched for T, B and NK cell differentiation and cytotoxicity, while Ba-R showed enrichment for extracellular matrix remodelling. Together this demonstrates that different microbial community types are associated with altered baseline peripheral immune tone and tumor features. Furthermore, it suggests that ICI response within each microbial community type may be characterised by divergent predictive markers and mechanisms of response. This is highly relevant to optimally targeting the gut microbiome in the clinic and identifying patients who will benefit most from microbiome-modifying interventions.