Combined tumor-associated microbiome and immune gene expression profiling predict response to neoadjuvant chemo-radiotherapy in locally advanced rectal cancer
Raffaello Roesel, Francesco Strati, Camilla Basso, Samantha Epistolio, Paolo Spina, J Djordjevic, Elisa Sorrenti, Martina Villa, Agnese Cianfarani, Francesco Mongelli, Jacopo Galafassi, Sotirios Georgios Popeskou, Federica Facciotti, Cecilia Caprera, Federica Melle, Pietro Majno‐Hurst, Alessandra Franzetti-Pellanda, Sara De Dosso, Ferdinando Bonfiglio, Milo Frattini, Dimitrios Christoforidis, Giandomenica Iezzi
Abstract
Locally advanced rectal cancer (LARC) is treated with neoadjuvant chemo-radiotherapy (nCRT) followed by surgery. A minority of patients show complete response (CR) to nCRT and may avoid surgery and its functional consequences. Instead, most patients show non-complete response (non-CR) and may benefit from additional treatments to increase CR rates. Reliable predictive markers are lacking. Aim of this study was to identify novel signatures predicting nCRT responsiveness. We performed a combined analysis of tumor-associated microbiome and immune gene expression profiling of diagnostic biopsies from 70 patients undergoing nCRT followed by rectal resection, including 16 with CR and 54 with non-CR. Findings were validated by an independent cohort of 49 patients, including 7 with CR and 42 with non-CR. Intratumoral microbiota significantly differed between CR and non-CR groups at genus and species level. Colonization by bacterial species of Ruminococcus genera was consistently associated with CR, whereas abundance of Fusobacterium, Porhpyromonas, and Oscillibacter species predicted non-CR. Immune gene profiling revealed a panel of 59 differentially expressed genes and significant upregulation of IFN-gamma and -alpha response in patients with CR. Integrated microbiome and immune gene profiling analysis unraveled clustering of microbial taxa with each other and with immune cell-related genes and allowed the identification of a combined signature correctly identifying non-CRS in both cohorts. Thus, combined intratumoral microbiome-immune profiling improves the prediction of response to nCRT. Correct identification of unresponsive patients and of bacteria promoting responsiveness might lead to innovative therapeutic approaches based on gut microbiota pre-conditioning to increase nCRT effectiveness in LARC.