Protein-based prognostic signature for predicting the survival and immunotherapeutic efficiency of endometrial carcinoma

Abstract

Abstract Background Endometrial cancer (EC) is the most frequent malignancy of the female genital tract worldwide. Our study aimed to construct an effective protein prognostic signature to predict prognosis and immunotherapy responsiveness in patients with endometrial carcinoma. Methods Protein expression data, RNA expression profile data and mutation data were obtained from The Cancer Proteome Atlas (TCPA) and The Cancer Genome Atlas (TCGA). Prognosis-related proteins in EC patients were screened by univariate Cox regression analysis. Least absolute shrinkage and selection operator (LASSO) analysis and multivariate Cox regression analysis were performed to establish the protein-based prognostic signature. The CIBERSORT algorithm was used to quantify the proportions of immune cells in a mixed cell population. The Immune Cell Abundance Identifier (ImmuCellAI) and The Cancer Immunome Atlas (TCIA) web tools were used to predict the response to immunochemotherapy. The pRRophetic algorithm was used to estimate the sensitivity of chemotherapeutic and targeted agents. Results We constructed a prognostic signature based on 9 prognostic proteins, which could divide patients into high-risk and low-risk groups with distinct prognoses. A novel prognostic nomogram was established based on the prognostic signature and clinicopathological parameters to predict 1, 3 and 5-year overall survival for EC patients. The results obtained with Clinical Proteomic Tumor Analysis Consortium (CPTAC), Human Protein Atlas (HPA) and immunohistochemical (IHC) staining data from EC samples in our hospital supported the predictive ability of these proteins in EC tumors. Next, the CIBERSORT algorithm was used to estimate the proportions of 22 immune cell types. The proportions of CD8 T cells, T follicular helper cells and regulatory T cells were higher in the low-risk group. Moreover, we found that the prognostic signature was positively associated with high tumor mutation burden (TMB) and high microsatellite instability (MSI-H) status in EC patients. Finally, ImmuCellAI and TCIA analyses showed that patients in the low-risk group were more inclined to respond to immunotherapy than patients in the high-risk group. In addition, drug sensitivity analysis indicated that our signature had potential predictive value for chemotherapeutics and targeted therapy. Conclusion Our study constructed a novel prognostic protein signature with robust predictive ability for survival and efficiency in predicting the response to immunotherapy, chemotherapy and targeted therapy. This protein signature represents a promising predictor of prognosis and response to cancer treatment in EC patients.