Establishment of a novel prognostic signature based on an identified expression profile of integrin superfamily to predict overall survival of patients with colorectal adenocarcinoma

The abnormal expression of integrin superfamily members commonly related to kinds of malignancies. However, the role of integrins in predicting the prognosis of cancers is still little known, especially for colorectal cancer that is one of the leading causes of cancer-related death. RNA-seq data and clinical features of colorectal adenocarcinoma (COAD) patients were derived from The Cancer Genome Atlas (TCGA), used to analyze the expression pattern and genomic alterations of integrin genes in the COAD cohort. Unsupervised hierarchical clustering divided COAD patients into two clusters (clusters 1 & 2), and we observed that patients in cluster 2 with high expressions of most integrin genes had worse clinical features and shorter overall survival (a median OS: 67.25 months vs 99.93 months, p = 0.012), compared to those in cluster 1. Combined with univariate Cox regression analysis, Pearson Correlation Coefficients (PCC), and Principal Component Analysis (PCA), an integrin-related signature was established, including ITGA1, ITGA5, ITGA7, ITGA11, ITGAX, ITGAM, ITGB1, and ITGB5. And the AUC values for OS at 1, 3, and 5 years was 0.61, 0.59, and 0.56, further demonstrating the predicting capacity of our signature. Furthermore, overexpression of which also significantly correlated with poorer prognosis of colon cancer patients in a separate validation cohort, GSE17536 (p < 0.05). Meanwhile, the AUC values for OS in the validation cohort at 1, 3, and 5 years was 0.62, 0.59, and 0.59. Additionally, enrichment analysis indicated significant differences between cluster 1 and cluster 2 in the biological processes of cell adhesion, signal transduction, extracellular matrix, immune system, and in tumor microenvironment (TME), which were crucial to the progression of tumor. The findings supplied compelling evidence that our signature could be a novel prognostic biomarker for COAD patients, and these genes had the potential to be therapeutic targets.