Tumor Microenvironment Analysis Identified Subtypes Associated With the Prognosis and the Tumor Response to Immunotherapy in Bladder Cancer

Hongxian Zhang; Jiwen Song; Junqiang Dong; Zhuo Liu; Lixuan Lin; Bing Wang; Qiang Ma; Lulin Ma

doi:10.3389/fgene.2021.551605

Tumor Microenvironment Analysis Identified Subtypes Associated With the Prognosis and the Tumor Response to Immunotherapy in Bladder Cancer

Front Genet. 2021 Mar 1:12:551605. doi: 10.3389/fgene.2021.551605. eCollection 2021.

Authors

Hongxian Zhang¹, Jiwen Song², Junqiang Dong³, Zhuo Liu¹, Lixuan Lin⁴, Bing Wang⁵, Qiang Ma⁶, Lulin Ma¹

Affiliations

¹ Department of Urology, Peking University Third Hospital, Beijing, China.
² Department of Urology, Shanxi Provincial Cancer Hospital, Taiyuan, China.
³ Department of Urology, Heze Municipal Hospital, Heze, China.
⁴ Department of Biotechnology, College of Life Sciences, Sichuan University, Chengdu, China.
⁵ Department of Oncology, Wuhan No.1 Hospital, Wuhan, China.
⁶ Department of Oncology, People's Hospital of Xintai City, Xintai, China.

Abstract

Background: The efficiency of immune checkpoint inhibitors (ICIs) in bladder cancer (BLCA) treatment has been widely validated; however, the tumor response to ICIs was generally low. It is critical and urgent to find biomarkers that can predict tumor response to ICIs. The tumor microenvironment (TME), which may play important roles to either dampen or enhance immune responses, has been widely concerned. Methods: The cancer genome atlas BLCA (TCGA-BLCA) cohort (n = 400) was used in this study. Based on the proportions of 22 types of immune cells calculated by CIBERSORT, TME was classified by K-means Clustering and differentially expressed genes (DEGs) were determined. Based on DEGs, patients were classified into three groups, and cluster signature genes were identified after reducing redundant genes. Then TMEscore was calculated based on cluster signature genes, and the samples were classified to two subtypes. We performed somatic mutation and copy number variation analysis to identify the genetic characteristics of the two subtypes. Correlation analysis was performed to explore the correlation between TMEscore and the tumor response to ICIs as well as the prognosis of BLCA. Results: According to the proportions of immune cells, two TME clusters were determined, and 1,144 DEGs and 138 cluster signature genes were identified. Based on cluster signature genes, samples were classified into TMEscore-high (n = 199) and TMEscore-low (n = 201) subtypes. Survival analysis showed patients with TMEscore-high phenotype had better prognosis. Among the 45 differentially expressed micro-RNAs (miRNAs) and 1,033 differentially expressed messenger RNAs (mRNAs) between the two subtypes, 16 miRNAs and 287 mRNAs had statistically significant impact on the prognosis of BLCA. Furthermore, there were 94 genes with significant differences between the two subtypes, and they were enriched in RTK-RAS, NOTCH, WNT, Hippo, and PI3K pathways. The Tumor Immune Dysfunction and Exclusion (TIDE) score of TMEscore-high BLCA was statistically lower than that of TMEscore-low BLCA. Receiver operating characteristic (ROC) curve analysis showed that the area under the curve (AUC) of TMEscore and tumor mutation burden (TMB) is 0.6918 and 0.5374, respectively. Conclusion: We developed a method to classify BLCA patients to two TME subtypes, TMEscore-high and TMEscore-low, and we found TMEscore-high subtype of BLCA had a good prognosis and a good response to ICIs.

Keywords: bladder cancer; copy number variation; differentially expressed genes; immune cells; immune checkpoint inhibitors; overall survival; somatic mutation; tumor microenvironment.