Abstract
Background: Tumor microenvironment (TME) takes a non-negligible role in the progression and metastasis of bladder urothelial carcinoma (BLCA) and tumor development could be inhibited by macrophage M1 in TME. The role of macrophage M1-related genes in BLCA adjuvant therapy has not been studied well. Methods: CIBERSOR algorithm was applied for identification tumor-infiltrating immune cells (TICs) subtypes of subjects from The Cancer Genome Atlas (TCGA) and Gene Expression Omnibus (GEO) data sets. We identified potential modules of M1 macrophages by weighted gene co-expression network analysis (WGCNA). Nomogram was determined by one-way Cox regression and lasso regression analysis for M1 macrophage genes. The data from GEO are taken to verify the models externally. Kaplan-Meier and receiver operating characteristic (ROC) curves validated prognostic value of M1 macrophage genes. Finally, we divided patients into the low-risk group (LRG) and the high-risk group (HRG) based on the median risk score (RS), and the predictive value of RS in patients with BLCA immunotherapy and chemotherapy was investigated. Bladder cancer (T24, 5637, and BIU-87) and bladder uroepithelial cell line (SV-HUC-1) were used for in vitro validation. Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) was employed to validate the associated genes mRNA level. Results: 111 macrophage M1-related genes were identified using WGCNA. RS model containing three prognostically significant M1 macrophage-associated genes (FBXO6, OAS1, and TMEM229B) was formed by multiple Cox analysis, and a polygenic risk model and a comprehensive prognostic line plot was developed. The calibration curve clarified RS was a good predictor of prognosis. Patients in the LRG were more suitable for programmed cell death protein 1 (PD1) and cytotoxic T lymphocyte associate protein-4 (CTLA4) combination immunotherapy. Finally, chemotherapeutic drug models showed patients in the LRG were more sensitive to gemcitabine and mitomycin. RT-qPCR result elucidated the upregulation of FBXO6, TMEM229B, and downregulation of OAS1 in BLCA cell lines. Conclusion: A predictive model based on M1 macrophage-related genes can help guide us in the treatment of BLCA.