A novel five-gene metabolism-related risk signature for predicting prognosis and immune infiltration in endometrial cancer: A TCGA data mining. (March 2023)

• Record Type: Journal Article
• Authors: Huang, Huaqing; Cai, Xintong; Lin, Jiexiang; Wu, Qiaoling; Zhang, Kailin; Lin, Yibin; Liu, Bin; Lin, Jie
• Is Part Of: Computers in biology and medicine. Volume 155(2023)
• Publication Date: 2023-03
• Subjects: Medicine -- Data processing -- Periodicals; Biology -- Data processing -- Periodicals; 610.285
• DOI: 10.1016/j.compbiomed.2023.106632
• Languages: English
• ISSNs: 0010-4825

Abstract

Abstract: Background: Metabolism dysfunction can affect the biological behavior of tumor cells and result in carcinogenesis and the development of various cancers. However, few thoughtful studies focus on the predictive value and efficacy of immunotherapy of metabolism-related gene signatures in endometrial cancer (EC). This research aims to construct a predictive metabolism-related gene signature in EC with prognostic and therapeutic implications. Methods: We downloaded the RNA profile and clinical data of 503 EC patients and screened out different expressions of metabolism-related genes with prognosis influence of EC from The Cancer Genome Atlas (TCGA) database. We first established a metabolism-related genes model using univariate and multivariate Cox regression and Lasso regression analysis. To internally validate the predictive model, 503 samples (entire set) were randomly assigned into the test set and the train set. Then, we applied the receiver operating characteristic (ROC) curve to confirm our previous predictive model and depicted a nomogram integrating the risk score and the clinicopathological feature. We employed a gene set enrichment analysis (GSEA) to explore the biological processes and pathways of the model. Afterward, we used ESTIMATE to evaluate the TME. Also, we adopted CIBERSORT and ssGSEA to estimate the fraction of immune infiltrating cells and immune function. At last, we investigated the relationship between the predictive model and immune checkpoint genes. Results: We first constructed a predictive model based on five metabolism-related genes (INPP5K, PLPP2, MBOAT2, DDC, and ITPKA). This model showed the ability to predict EC patients' prognosis accurately and performed well in the train set, test set, and entire set. Then we confirmed the predictive signature was a novel independent prognostic factor in EC patients. In addition, we drew and validated a nomogram to precisely predict the survival rate of EC patients at 1-, 3-, and 5-years (ROC1-year = 0.714, ROC3-year = 0.750, ROC5-year = 0.767). Furthermore, GSEA unveiled that the cell cycle, certain malignant tumors, and cell metabolism were the main biological functions enriched in this identified model. We found the five metabolism-related genes signature was associated with the immune infiltrating cells and immune functions. Most importantly, it was linked with specific immune checkpoints (PD-1, CTLA4, and CD40) that could predict immunotherapy's clinical response. Conclusion: The metabolism-related genes signature (INPP5K, PLPP2, MBOAT2, DDC, and ITPKA) is a valuable index for predicting the survival outcomes and efficacy of immunotherapy for EC in clinical settings. Highlights: A predictive model was constructed based on The Cancer Genome Atlas (TCGA) database, to predict the prognosis and immunotherapy effectiveness of EC patients. Five-gene (INPP5K, PLPP2, MBOAT2, DDC, and ITPKA) metabolism-related signature was the first to be found as an independent risk factor for EC. Those five genes may help improve clinical decision-making and optimize the personalized treatment of EC patients.