An Immune Subtype Classification System Enables the Development of Strategies to Predict and Enhance Immunotherapy Responses in Colorectal Cancer

Abstract

Colorectal cancer (CRC) is the second leading cause of cancer-related mortality globally. While immunotherapeutic approaches are effective in a subset of CRC patients, the majority of CRC cases receive limited benefits from immunotherapy. This study developed an immune subtype classification system based on diverse immune cells and pathways. A model constructed through machine learning based on immune subtypes could accurately predict the sensitivity of CRC patients to immunotherapy. Validation of this model across public datasets and clinical samples confirmed its high precision and reliability. Furthermore, drug screening based on the immune subtypes identified the IGF1R inhibitor I-OMe-AG-538 (AG-538) as a potent enhancer of antitumor immunity. Mechanistic investigations revealed that AG-538 induced reactive oxygen species (ROS)-dependent DNA damage and downregulated the expression of multiple repair genes, triggering cGAS/STING-mediated type I IFN signaling within tumor cells. This signaling cascade increased tumor immunogenicity and refined the tumor immune microenvironment, thereby enhancing efficacy of immune checkpoint blockade treatment. In summary, these findings present a predictive model for immune response and highlight the potential of AG-538 combined with anti-PD1 antibodies as a chemoimmunotherapeutic strategy.