Prediction of Response to Immune Checkpoint Blockade in Patients with Metastatic Colorectal Cancer with Microsatellite Instability

Tumors with microsatellite instability (MSI) represent a paradigm for the success of immune checkpoint inhibitor (ICI)-based immunotherapy, particularly in patients with metastatic colorectal cancer (mCRC). To date however, tools for predicting efficacy of these new therapies are lacking. Here we combined high-throughput DNA and RNA sequencing of tumors from 117 patients with MSI mCRC treated with anti-PD-1 +/- anti-CTLA-4 and enrolled into two cohorts, i.e., the NIPICOL clinical trial (NCT03350126) and the ImmunoMSI prospective cohort that were used as discovery and validation cohorts in this ancillary study, respectively. All analyses based on previously suggested DNA/RNA biomarkers of resistance failed to identify robust predictors of treatment response in patients, e.g., the level of MSI, mutational burden, the presence of specific somatic DNA variants as assessed by exome-sequencing and the activity of canonical signaling pathways or the presence of specific cellular contingents within the tumor bulk as estimated by RNA-sequencing. By contrast, resistance to ICI was found to depend both on a small subset of somatic DNA variants located in microsatellite-containing genes with very diverse biological functions and the expression of a stromal-oriented RNA component. Testing of these predictors in 5 large additional independent retrospective and prospective cohorts (IDEA and MOSAIC clinical trials) regrouping 446 patients with nonmetastatic or metastatic MSI CRC untreated with ICI allowed us to validate the specificity of the predictive nature of these indicators regarding ICI-treated MSI mCRC patients. The use of these DNA/RNA predictors will help to refine the ICI-based precision therapy of patients with metastatic MSI mCRC.