Integrated bioinformatics analysis identifies a Ferroptosis-related gene signature as prognosis model and potential therapeutic target of bladder cancer

Abstract

Background: Bladder cancer (BLCA) is one of the most prevalent cancers worldwide. Ferroptosis is a newly discovered form of non-apoptotic cell death that plays an important role in tumors. However, the prognostic value of ferroptosis-related genes (FRGs) in BLCA has not yet been well studied.

Method and materials: In this study, we performed consensus clustering based on FRGS and categorized BLCA patients into 2 clusters (C1 and C2). Immune cell infiltration score and immune score for each sample were computed using the CIBERSORT and ESTIMATE methods. Functional annotation of differentially expressed genes were performed by Gene Ontology (GO) and KEGG pathway enrichment analysis. Protein expression validation were confirmed in Human Protein Atlas. Gene expression validation were performed by qPCR in human bladder cancer cell lines lysis samples.

Result: C2 had a significant survival advantage and higher immune infiltration levels than C1. Additionally, C2 showed substantially higher expression levels of immune checkpoint markers than C1. According to the Cox and LASSO regression analyses, a novel ferroptosis-related prognostic signature was developed to predict the prognosis of BLCA effectively. High-risk and low-risk groups were divided according to risk scores. Kaplan-Meier survival analyses showed that the high-risk group had a shorter overall survival than the low-risk group throughout the cohort. Furthermore, a nomogram combining risk score and clinical features was developed. Finally, SLC39A7 was identified as a potential target in bladder cancer.

Discussion: In conclusion, we identified two ferroptosis-clusters with different prognoses using consensus clustering in BLCA. We also developed a ferroptosis-related prognostic signature and nomogram, which could indicate the outcome.

Keywords: Bladder cancer; Consensus clustering; Ferroptosis; Prognostic prediction model; Risk score.

Figure 1

Consensus clustering based on Ferroptosis-related genes in bladder cancer: A) the cumulative distribution function for consensus values of K = 2–9 provides insight into the number of optimal subtypes: B) evaluation of the relative change in the area under the curve of the cumulative distribution function: C) the final classification of the samples was K = 2–9, indicating the assigned subtypes: D) Heatmap representing the consensus matrix for K = 2, illustrating the clustering patterns: E) principal component analysis (PCA) plot showing the effective separation of bladder cancer patients into two distinct subtypes using ferroptosis-related genes: F) the Kaplan–Meier curves of overall survival between C1 and C2 subtypes.

Figure 2

Somatic mutation landscape of C1 group and C2 group: A) the mutation frequency of the top 20 genes in C1 subtype: B) the mutation frequency of the top 20 genes in C2 subtype.

Figure 3

Enrichment of KEGG pathways relevant to differentially expressed genes (DEGs) in the two predicted subtypes via GSVA analysis: A) volcano plot of the differentially expressed genes B) top 10 enriched KEGG pathways of the differentially expressed genes by functional enrichment analysis: C) top 10 enriched GO terms of the differentially expressed genes by functional enrichment analysis: D) Heatmap showed KEGG pathways activated in C1 and C2 subtypes of differences.

Figure 4

Relationship between the two subtypes and the tumor immune microenvironment: A) comparison of immune cell infiltration levels between C1 and C2 subtype according to the infiltration results estimated by CIBESORT: B) comparison of immune scores between C1 and C2 subtype according to the infiltration results estimated by ESTIMATE: C) comparison of immune checkpoint gene expression between C1 and C2; ^*P < 0.05, ^**P < 0.01, ^***P < 0.001, ^****P < 0.0001.

Figure 5

Establishment of a prognostic risk model concerning bladder cancer and Ferroptosis: A) the 10-round cross-validation determined the optimal values of the penalty parameter of the proportional hazards model: B) the LASSO-cox analysis identified 25 ferroptosis related mRNAs most related to prognostics: C) distribution of riskscore, survival time, and survival status in the dataset; D) patients showed poor overall survivalin the high risk group than those in the low-risk group E) ROC curves of the risk model diagnostic performance in 1-, 3-, 5- years survival.

Figure 6

Development of the nomogram model based on risk score and clinical features: A) nomogram considering risk score and several common-used clinicopathological factors the bladder cancer population: B) calibration curve to assess concordance between predicted and actual 1-year overall survival: C) calibration curve to assess concordance between predicted and actual 3-year overall survival: D) calibration curve to assess concordance between predicted and actual 5-year overall survival.

Figure 7

Multivariable regression analysis and gene expression in the prognostic model: A) multivariable regression analysis of the risk factor: B) mRNA expression levels of 4 genes in bladder cancer: C) protein expression of SLC39A7 in HPA database; D: SLC39A7 mRNA expression level in bladder cancer cell lines; ^*P < 0.05, ^***P < 0.001.