Homologous Recombination Related Signatures Predict Prognosis and Immunotherapy Response in Metastatic Urothelial Carcinoma

Pan Li; Chaohu Chen; Jianpeng Li; Li Yang; Yuhan Wang; Zhilong Dong; Jun Mi; Yunxin Zhang; Juan Wang; Hanzhang Wang; Ronald Rodriguez; Junqiang Tian; Zhiping Wang

doi:10.3389/fgene.2022.875128

Homologous Recombination Related Signatures Predict Prognosis and Immunotherapy Response in Metastatic Urothelial Carcinoma

Front Genet. 2022 Apr 26:13:875128. doi: 10.3389/fgene.2022.875128. eCollection 2022.

Authors

Pan Li¹, Chaohu Chen¹, Jianpeng Li¹, Li Yang^{1

2

3}, Yuhan Wang¹, Zhilong Dong^{1

2

3}, Jun Mi^{1

2

3}, Yunxin Zhang^{1

2

3}, Juan Wang¹, Hanzhang Wang⁴, Ronald Rodriguez⁴, Junqiang Tian^{1

2

3}, Zhiping Wang^{1

2

3}

Affiliations

¹ Department of Urology, Lanzhou University Second Hospital, Lanzhou, China.
² Key Laboratory of Gansu Province for Urological Diseases, Lanzhou, China.
³ Clinical Center of Gansu Province for Nephron-Urology, Lanzhou, China.
⁴ Department of Urology, University of Texas Health Science Center at San Antonio, San Antonio, TX, United States.

Abstract

Objective: This study used homologous recombination (HR) related signatures to develop a clinical prediction model for screening immune checkpoint inhibitors (ICIs) advantaged populations and identify hub genes in advanced metastatic urothelial carcinoma. Methods: The single-sample gene enrichment analysis and weighted gene co-expression network analysis were applied to identify modules associated with immune response and HR in IMvigor210 cohort samples. The principal component analysis was utilized to determine the differences in HR-related module gene signature scores across different tissue subtypes and clinical variables. Risk prediction models and nomograms were developed using differential gene expression analysis associated with HR scores, least absolute shrinkage and selection operator, and multivariate proportional hazards model regression. Additionally, hub genes were identified by analyzing the contribution of HR-related genes to principal components and overall survival analysis. Finally, clinical features from GSE133624, GSE13507, the TCGA, and other data sets were analyzed to validate the relationship between hub genes and tumor growth and mutation. Results: The HR score was significantly higher in the complete/partial response group than in the stable/progressive disease group. The majority of genes associated with HR were discovered to be involved in the cell cycle and others. Genomically unstable, high tumor level, and high immune level samples all exhibited significantly higher HR score than other sample categories, and higher HR scores were related to improved survival following ICIs treatment. The risk scores for AUNIP, SEPT, FAM72D, CAMKV, CXCL9, and FOXN4 were identified, and the training and verification groups had markedly different survival times. The risk score, tumor neoantigen burden, mismatch repair, and cell cycle regulation were discovered to be independent predictors of survival time following immunotherapy. Patients with a high level of expression of hub genes such as EME1, RAD51AP1, and RAD54L had a greater chance of surviving following immunotherapy. These genes are expressed at significantly higher levels in tumors, high-grade cancer, and invasive cancer than other categories, and are associated with TP53 and RB1 mutations. Conclusion: HR-related genes are upregulated in genomically unstable samples, the survival time of mUC patients after treatment with ICIs can be predicted using a normogram model based on HR signature.

Keywords: genomically unstable; homologous recombination; immune checkpoint inhibitors; metastatic urothelial carcinoma; predict prognosis.