DataSheet1.docx

Abstract

Background: The unfolded protein response (UPR) played a significant role in maintain protein hemostasis in tumor cells, which were crucial for tumor growth, invasion and resistance to therapy. This study aimed to develop a UPR-related signature and explore its correlation with immunotherapy and chemotherapy in bladder cancer. Methods: The differentially expressed UPR-related genes were put into Lasso regression to screen out prognostic genes, which constituted the UPR signature and were incorporated into multivariate Cox regression to generate risk scores. Subsequently, the predictive performance of this signature was estimated by receiver operating characteristic (ROC) curves. The CIBERSORTx, the maftool, and Gene set enrichment analysis (GSEA) were applied to explore infiltrated immune cells, tumor mutational burden (TMB), and enriched signaling pathways in both risk groups, respectively. Moreover, The Cancer Immunome Atlas (TCIA) and Genomics of Drug Sensitivity in Cancer (GDSC) databases were used predict the responses of chemotherapy and immunotherapy. Results: Twelve genes constituted the UPR-related signature. Patients with higher risk scores had worse overall survival (OS) in training and three validation sets. The UPR-related signature closely correlated with clinicopathologic parameters and could serve as an independent prognostic factor. M0 macrophages showed significantly infiltrated difference in both risk groups. TMB analysis showed that the risk score in the wild type and mutation type of FGFR3 was significantly different. GSEA indicated that immune-, extracellular matrix-, replication and repair associated pathways belonged to the high risk group and metabolism-related signal pathways were enriched into the low risk group. Prediction of immunotherapy and chemotherapy revealed that patients in the high risk group might benefit from chemotherapy, but had a worse response to immunotherapy. Finally, we constructed a predictive model with age, stage and UPR-related risk score, which had a robustly predictive performance and were validated in GEO datasets. Conclusion: We successfully constructed and validated a novel UPR-related signature in bladder cancer, which could robustly predict survival outcomes and closely correlate with response of immunotherapy and chemotherapy in bladder cancer.