Abstract
BACKGROUND: Colon cancer is the third most common malignant tumor in the world. Although immunotherapy has been used in cancer treatment, there is still no first-line immunotherapy method for colon cancer. Therefore, it is essential to search for potential immunotherapy targets and molecular biomarkers for early diagnosis and prognosis. METHODS: In this study, we downloaded transcriptome data from The Cancer Genome Atlas (TCGA) and immune-related genes from the ImmPort database. Then we filtered genes with prognostic value and constructed an immune-related signature. Patients were classified into low- and high-risk groups, and we exerted a series of analysis between the signature and clinical phenotypes. Additionally, we used protein-protein interaction networks, gene set enrichment analysis (GSEA) and single-sample gene-set enrichment analysis (ssGSEA) to explore the underlying mechanism of this signature. Furthermore, the accuracy of this signature was verified, using two data sets from Gene Expression Omnibus (GEO). RESULTS: We selected 12 immune-related genes to construct the immune-related signature. Low-risk group had a higher level of immunity compared to high-risk group. The expression level of HLA genes and checkpoint-related genes were statistically different in low- and high-risk groups. This signature showed its prognostic value in TCGA cohort and 2 GEO data sets. The signature also had strong correlation with TNM classification, stage, survival state and lymphatic invasion. The mechanism of the signature may be related to several transcription factors and CD8+ T cell in the tumor microenvironment. CONCLUSION: In conclusion, this immune-related signature is of great prognosis value for colon cancer and its biofunction might be correlated with HLA genes, checkpoint-related genes and high-infiltrating T cells in tumor tissues.