Novel driver genes and genomic predictors of immune infiltrates in colorectal cancer.

Abstract

557 Background: Despite progress that has been made in the molecular characterization of colorectal cancer (CRC), prior studies have had limited power to detect, as significantly mutated, less-frequently altered CRC genes. In addition, most large-scale sequencing studies have generally lacked clinicopathologic annotations to link genomic features with important CRC characteristics such as immune infiltration. Methods: We performed Whole-Exome Sequencing on 619 archival primary CRCs from two large prospective cohorts, the Nurses’ Health Study and the Health Professionals Follow-Up Study. We identified CRC driver genes using significance-calling algorithms and employed a novel computational pipeline to calculate tumor neoantigens (peptides resulting from somatic mutations and recognized by the immune system as foreign). The neoantigen load and altered genes/pathways were correlated to immunohistochemically determined immune infiltrates and survival data linked to these specimens. Results: We identified 90 significantly mutated genes in non-hypermutated CRC. These include previously underappreciated genes that participate in pathways such as WNT-signaling and RNA processing. Among all samples, the neoantigen load was associated with increased overall lymphocytic infiltration (P = 2.6e-11), tumor-infiltrating lymphocytes (TILs) (P = 2.0e-19), memory T-cells (P = 0.091) and CRC-specific survival (P = 0.025). Neoantigen load was also associated with higher TILs in microsatellite-stable tumors (P = 0.015). We found that HLA and Antigen Processing Machinery (APM) mutations have undergone positive selection in tumors with TILs. Conclusions: In the largest exome-sequencing study of CRC to date, we leveraged the increased number of samples and their associated clinicopathologic annotations to identify new driver genes and genomic predictors of immune infiltrates in this disease. Neoantigen load is prognostic in CRC and predictive of immune infiltration, even among microsatellite-stable tumors. We also find evidence of positive selection for HLA and APM mutations in immune-cell infiltrated tumors. These results have implications for patient selection in CRC immunotherapy trials.