Association of intra-tumoral microbiome and response to immune checkpoint inhibitors (ICIs) in patients with metastatic renal cell carcinoma (mRCC).

Abstract

372 Background: Multiple studies have suggested that the gut microbiome plays a modulatory role in ICI activity and that specific bacteria and/or cumulative microbial diversity may drive response in patients (pts) with mRCC (Routy et al Science 2018; Salgia et al Eur Urol 2020). Even though the tumor microenvironment has substantial bacterial proliferation (Heymann et al Cancer L. 2021), there is a paucity of data assessing the impact of intra-tumoral microbiota in response to ICI therapy. In this study, we sought to explore this association in pts with mRCC. Methods: Pts diagnosed with mRCC who had available RNA sequencing (RNA-seq) data collected in the course of routine clinical care and who were treated with ICIs were retrospectively identified.Intra-tumoral microbiome analysis was performed on formalin-fixed paraffin-embedded samples. Following quality and adapter trimming, RNA-seq reads were mapped to a human genome to filter host reads using the Burrows-Wheeler alignment (BWA) tool. Taxonomic classification was performed using Kraken2 and the absolute abundances of species were estimated using Bracken. The relative abundances among all non-human species were calculated. Statistical testing with Student’s t-test was performed to compare the relative abundance for all species seen within pts who responded to ICIs and those who did not. Results: Among the 28 pts (22:6, M:F) included in this analysis, 24 (86%) had clear cell histology and 20 (71%) were IMDC intermediate/poor risk. All of the samples were collected prior to starting treatment with ICIs and the majority of these (57%) were collected from the primary site. 11 pts (39%) received ICIs as first line treatment and 17 (61%) as second line. Clinical response was seen in 50% of pts included in the study and the most common rendered treatment was nivolumab (17 pts). In the overall cohort, Cutibacterium acne, Moraxella osloensis, and Pasteurella multocida had the highest relative abundances. Additionally, significant differences in relative abundances of specific bacteria were found between ICI responders and non-responders. Among these, Stenotrophomonas maltophilia (p = 0.037) and Corynebacterium sp. zg-917 (p = 0.035) had significantly higher relative abundances in pts who responded to ICIs. Conclusions: This is the first study evaluating the association between intra-tumoral microbiome and response to ICIs in pts with mRCC. Among bacteria associated with response, several have particular relevance – for instance, Corynebacterium spp. have been studied for decades as a possible adjunct to immunotherapeutic agents such as BCG. Efforts are ongoing to validate these findings in a larger cohort.