Identification of a Programmed Cell Death-related Signature for Prediction of Osteosarcoma

Abstract

Background: This study aims to perform bioinformatics analysis of programmed cell death-related genes (PCDGs) in osteosarcoma, and to construct a multi-gene signature for predicting the prognosis.Methods: The functional enrichment analysis was applied for prognostic PCDGs, and PPI network as well as drug-gene interactions were constructed. In order to set up the prognosis evaluation system, we established a prognosis model by integrating PCDGs. The Least Absolute Shrinkage and Selection Operator (LASSO) logistic regression analysis as well as the multivariate Cox proportional hazard regression analysis were used to construct the five-genes signature (MUC1, TCF7L2, TGFB2, TRIAP1, CBS) for prediction in Gene Expression Omnibus (GEO) cohort. According to the median risk score, survival analysis was conducted to evaluate the prognostic value of the risk score in GEO cohort. Next, by combining other clinic-pathological independent prognostic factor, stage at diagnosis, a nomogram was established to predict individual survival probability. Result: GO analysis showed that the 15 prognostic PCDGs were mainly enriched in apoptotic signaling pathway, regulation of secretion and p53 signaling pathway. KEGG analysis showed that aforesaid genes were mainly related to PI3K-Akt signaling pathway, diverse neoplasms signaling pathway and hepatitis B. Drug-gene interactions displayed available drugs for influencing osteosarcoma via programmed cell death, such as adalimumab, tacrolimus and sirolimus. The risk score was constructed based on 5 genes and patients were significantly divided into high-risk and low-risk groups according to overall survival. In multivariate Cox regression analysis, risk score was still an independent prognostic factor (HR = 2.526, 95% CI = 1.597−3.994, P < 0.001). Cumulative curve showed that low-risk score patients were significantly had better prognosis than that of patients with high-risk score (P < 0.001). The external independent TARGET cohort proved the validity of risk score model and the nomograph.Conclusion: From the facet of programmed cell death, we provided a multi-gene signatureand the nomograph for the prognostic predictors of osteosarcoma patients, and available drugs displayed may provide promising treatment strategies.