Background Autophagy has a critical involvement in the initiation and progression of various cancers, including colorectal cancer (CRC). The feasibility of using autophagy‐related genes (ATGs) as prognostic tools for CRC patients is yet to be determined. Methods RNA sequencing data and clinical information for CRC were obtained from The Cancer Genome Atlas (TCGA) (training set) and Gene Expression Omnibus (GEO) datasets GSE39582 and GSE44076 (validation). ARGs were retrieved from the Human Autophagy Database, and differentially expressed ARGs (DAGs) were identified using the “limma” R package. Prognostic signature DAGs were established via univariate Cox and LASSO Cox regression analyses. The obtained signature was validated through expression analysis, autophagy scoring, survival prediction, and correlation with immune status. Immunohistochemistry assays and in vitro functional experiments in SW480 cells were performed to assess the biological roles of selected DAGs, particularly WIPI2. Results We constructed an 11‐gene prognostic signature (CANX, NRG1, WIPI1, EIF2AK3, WDR45, PELP1, ULK1, WIPI2, DAPK1, ULK3, and MAP1LC3C), with high‐risk patients showing significantly reduced overall survival compared to low‐risk patients. WIPI2, highly expressed in SW480 cells, was selected for functional validation. Knockdown of WIPI2 inhibited cell proliferation, suppressed autophagy (decreased LC3B‐II, increased p62), and promoted apoptosis (increased cleaved caspase‐3). These findings confirm the pro‐survival role of WIPI2 in CRC. The prognostic signature remained independently predictive after adjusting for clinical factors and showed a strong correlation with immune infiltration in TCGA CRC samples. Conclusion The autophagy‐related signature independently predicts CRC prognosis and guides immunotherapy strategies.
Miao et al. (Thu,) studied this question.