Identification of a Novel Ferroptosis-Related Gene Prognostic Signature in Bladder Cancer

Abstract

Background: Ferroptosis is a newly found non-apoptotic forms of cell death that plays an important role in tumors. However, the prognostic value of ferroptosis-related genes (FRG) in bladder cancer (BLCA) have not been well examined.

Methods: FRG data and clinical information were collected from The Cancer Genome Atlas (TCGA). Then, significantly different FRGs were investigated by functional enrichment analyses. The prognostic FRG signature was identified by univariate cox regression and least absolute shrinkage and selection operator (LASSO) analysis, which was validated in TCGA cohort and Gene Expression Omnibus (GEO) cohort. Subsequently, the nomogram integrating risk scores and clinical parameters were established and evaluated. Additionally, Gene Set Enrichment Analyses (GSEA) was performed to explore the potential molecular mechanisms underlying our prognostic FRG signature. Finally, the expression of three key FRGs was verified in clinical specimens.

Results: Thirty-two significantly different FRGs were identified from TCGA-BLCA cohort. Enrichment analyses showed that these genes were mainly related to the ferroptosis. Seven genes (TFRC, G6PD, SLC38A1, ZEB1, SCD, SRC, and PRDX6) were then identified to develop a prognostic signature. The Kaplan-Meier analysis confirmed the predictive value of the signature for overall survival (OS) in both TCGA and GEO cohort. A nomogram integrating age and risk scores was established and demonstrated high predictive accuracy, which was validated through calibration curves and receiver operating characteristic (ROC) curve [area under the curve (AUC) = 0.690]. GSEA showed that molecular alteration in the high- or low-risk group was closely associated with ferroptosis. Finally, experimental results confirmed the expression of SCD, SRC, and PRDX6 in BLCA.

Conclusion: Herein, we identified a novel FRG prognostic signature that maybe involved in BLCA. It showed high values in predicting OS, and targeting these FRGs may be an alternative for BLCA treatment. Further experimental studies are warranted to uncover the mechanisms that these FRGs mediate BLCA progression.

Keywords: bioinformatics analysis; bladder cancer; ferroptosis; gene signature; tumor tissue microarray.

Figure 1

Identification of differentially expressed FRGs in TCGA–BLCA cohort. (A) Heatmap of FRGs. Green represents downregulation, and red represents upregulation of genes. (B) Volcano plot of FRGs. Green dots represent 10 downregulated genes; red dots represent 22 upregulated genes. FRG, ferroptosis−related gene; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer.

Figure 2

Functional enrichment analysis of the 32 differentially expressed FRGs in TCGA-BLCA cohort. Top 10 enriched biological processes, molecular functions, cellular components (A), and KEGG pathways (B) terms of 32 differentially expressed FRGs are shown in this study. FRG, ferroptosis−related gene; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer; KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 3

Development of FRG prognostic signature in TCGA cohort. (A) Results of the univariate cox analysis of the OS in TCGA-BLCA cohort. (B) Ten-time cross-validation for tuning parameter selection in the LASSO Cox regression model. (C) LASSO coefficient profiles of the seven ferroptosis-related genes. FRG, ferroptosis−related gene; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer; LASSO, least absolute shrinkage and selection operator.

Figure 4

Evaluation of the constructed FRG prognostic signature in TCGA–BLCA cohort. (A) Heatmap of seven genes expression between high- and low-risk score group in TCGA–BLCA cohort. (B) Kaplan–Meier survival curve. (C) Risk score curve plot. The dotted line indicates the individual distribution of risk score, and the patients are categorized into low-risk (green) and high-risk (red) groups. (D) Risk score scatter plot. Red dots indicate the dead patients, and green dots indicate the alive. With the increase in risk score, more patients died. FRG, ferroptosis−related gene; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer.

Figure 5

Validation of the constructed FRG prognostic signature in GSE13507. (A) Heatmap of seven genes expression between high- and low-risk score group in GSE13507 cohort. (B) Kaplan–Meier survival curve. (C) Risk score curve plot. The dotted line indicates the individual distribution of risk score, and the patients are categorized into low-risk (green) and high-risk (red) groups. (D) Risk score scatter plot. Red dots indicate the dead patients, and green dots indicate the alive. With the increase in risk score, more patients died. FRG, ferroptosis−related gene.

Figure 6

Establishment of nomogram. Univariate (A) and multivariate (B) analyses assessing relationship between risk scores and relevant clinical parameters and OS in TCGA–BLCA cohort. (C) Establishing of a signature-based prognostic nomogram predicting OS in BLCA. OS, overall survival; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer.

Figure 7

Validation of nomogram. Calibration curves of the nomogram prediction of 1-year (A), 3-year (B), and 5-year (C) OS of patients in TCGA–BLCA cohort. (D) ROC curve of the risk score and other relevant clinical parameters. OS, overall survival; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer; ROC, receiver operating characteristic.

Figure 8

Gene Set Enrichment Analysis of our prognostic FRG signature in TCGA-BLCA cohort. FRG, ferroptosis−related gene; TCGA, The Cancer Genome Atlas; BLCA, bladder cancer.

Figure 9

Immunohistochemistry results of three key FRGs. (A) Representative immunohistochemistry images of PRDX6, SCD, and SRC in different T stage of BLCA. Comparisons of IRS of PRDX6 (B, E), SCD (C, F), and SRC (D, G) between NMIBC and MIBC or in different T stage of BLCA. *p < 0.05, **p < 0.01, ns means no significant difference. FRG, ferroptosis−related gene; BLCA, bladder cancer; IRS, immunoreactive score; NMIBC, non-muscle invasive bladder cancer; MIBC, muscle-invasive bladder cancer.

Figure 10

Flow chart of data collection and analysis. TCGA, The Cancer Genome Atlas; BLCA, bladder cancer; FRG, ferroptosis−related gene; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; OS, overall survival; LASSO, least absolute shrinkage and selection operator; GSEA, Gene Set Enrichment Analysis; IHC, immunohistochemistry; TMA, tissue microarray.