Immune Escape and Metabolic Reprogramming in Colon Cancer: Insights from Endocytosis-Related Genes

Abstract

Colon cancer (COAD) is a common malignancy, yet its etiology is not fully elucidated. This study gathered endocytosis-related genes, using gene expression profiles from TCGA databases to categorize molecular subtypes of COAD into Cluster1 and Cluster2 based on genes related to endocytosis, and further explored the connection between the two molecular subtypes and prognostic characteristics. Differential expression analysis of the two subtypes revealed 3412 differentially expressed genes (DEGs), whose functions were remarkably enriched in the cellular metastasis and oxidative phosphorylation in Cluster1 by fGSEA. Additionally, Cluster1 exhibited higher expression levels of DEGs associated with immune infiltration and metabolism in comparison to Cluster2 by GSVA, and TIDE scores indicated patients with Cluster2 may benefit more from immunotherapy. Based on the DEGs, we utilized univariate Cox regression to identify 759 prognostic genes, which were then screened by three machine learning models (Lasso, RF, SVM-RFE) simultaneously, resulting in four feature genes: NEK4, MED13, OXSR1, and SLAIN2. Moreover, in Cluter1, these feature genes displayed consistent positive or negative correlations with immune escape-related and metabolic reprogramming-related pathways and genes in Pearson heatmap. There results suggesting that there are significant differences in immune escape and metabolic reprogramming between colon cancer subtypes Cluster1 and Cluster2 as determined by genes according to endocytosis.