Identification of angiogenesis-related subtypes, the development of prognostic models, and the landscape of tumor microenvironment infiltration in colorectal cancer
Background: Angiogenesis is one of the most prominent markers of cancer progression and contributes to tumor metastasis and prognosis. Anti-angiogenic drugs have proven effective in treating metastatic colorectal cancer. However, there is some uncertainty regarding the potential role of angiogenesis-related genes in the tumor microenvironment. Methods: We analyzed 1,214 colorectal cancer samples to identify alterations in angiogenesis-related genes (ARGs), and then correlated angiogenesis with clinical features, prognosis, and TME. The ARGs expression profiles in colorectal cancer were analyzed using three computational methods (CIBERSORT, ssGSEA, and MCPcounter) and provided a systematic immune landscape. Patients with CRC were classified into two subtypes based on consensus clustering analysis of angiogenesis-related genes. The revealed differentially expressed genes between the two subtypes were used to create and validate ARGscore prognostic models. In addition, we collected eight colorectal cancer patient specimens and performed RT-qPCR to validate the signature gene expression. Results: We assessed the expression patterns of ARGs in colorectal cancer. We identified two molecular subtypes and confirmed that the expression of ARGs was associated with prognosis and TME characteristics. Based on differentially expressed genes between subtypes, we constructed ARGscore and evaluated their predictive power for the survival of colorectal cancer patients. We also developed an accurate nomogram to make the ARGscore more clinically useful. In addition, ARGscore was significantly correlated with microsatellite instability, cancer stem cells, and chemotherapeutic drug sensitivity. Patients with ARGscore-low characterized by immune activation and microsatellite instability high had a better prognosis. Conclusion: ARGs expression influenced the prognosis, clinicopathological features, and tumor stromal immune microenvironment in colorectal cancer. We developed a new risk model, ARGscore, for the treatment and prognosis of CRC patients and validated its promising predictive power. These findings will enable us to understand colorectal cancer better, assess prognoses, and develop more effective treatment options.