Construction of a risk model based on exosome-related genes predict clinical prognosis and therapeutic response and revealing TIMP1 as a promising target in colorectal cancer.

Various exosome-derived proteins have been reported to play essential roles in regulating colorectal cancer progression and affecting the prognosis of cancer patients. It is necessary to explore the critical exosome-related genes in colorectal cancer. In this study, 23 differentially expressed exosome-related genes associated with prognosis in colorectal cancer (CRC) were identified based on two datasets, The Cancer Genome Atlas (TCGA) and exoRbase. Based on machine learning-Boruta and lasso-Cox regression-nine essential genes were finally identified, and a risk model was constructed. The risk model was able to predict the prognosis of the patients well. Specifically, the prognosis of high-risk patients was worse, and the prognosis of low-risk patients was better. Multivariate Cox regression revealed that the risk model was an independent prognostic factor. Mechanism studies showed that pathways such as MYOGENESIS, APICAL JUNCTION, Epithelial-Mesenchymal Transition (EMT), ANGIOGENESIS, and KRAS SIGNALING DN were highly enriched in the high-risk group. In addition, tumor-promoting immune cells, such as Treg cells and macrophages, exhibited increased activity in the high-risk group, suggesting that high-risk patients may be less responsive to immunotherapy. Furthermore, in multiple external immunotherapeutic and chemotherapeutic datasets, we found that high-risk patients are less sensitive to therapy than low-risk patients, suggesting that this risk score may predict immune and chemotherapy response. Scoring the importance of nine genes, we found that TIMP1 was the most critical exosome-related gene in colorectal cancer patients. Knockdown of TIMP1 in colorectal cancer cells significantly inhibited the proliferation and migration of colorectal cancer cells. In conclusion, we identified several crucial colorectal cancer exosome-associated genes using public datasets and machine learning, and constructed risk models to predict prognosis and response to immunotherapy. TIMP1 was further identified as a critical oncogene in patients with colorectal cancer. Our results provide a theoretical basis for subsequent exosome-based preclinical trials.