Integrative Analysis of Novel Ferroptosis-Related Genes Signatures as Prognostic Biomarkers in Ovarian Cancer

Abstract

Background: Ferroptosis, an iron-dependent form of cell death, has been implicated in the pathogenesis of several types of cancer. Nevertheless, the exact correlation between ferroptosis-related gene mutations and their influence on ovarian cancer (OV) diagnosis and treatment strategies remains to be fully elucidated. It is crucial to identify the ferroptosis-related gene signature in OV and elucidate the impact of these mutations and their expression on the diagnosis and treatment of OV.

Methods: In this study, we collected data from the TCGA and GEO databases. We utilized various tools and packages for data analysis, including the cBio Cancer Genomics Portal, Tumor Immune Estimation Resource (TIMER), GSVA package, and WGCNA R packages.

Results: Our results showed that ferroptosis subtypes 1 (FS1) and 2 (FS2) exhibited different levels of expression and tumor mutation burden (TMB). FS2 had a higher TMB level and survival rate compared to FS1. Furthermore, our analysis identified three ferroptosis-related genes, including IFNG, KEAP1, and PHKG2, as key biomarkers in prognosis prediction and potential targets for OV cancer therapy. The elevated expression levels of IFNG, KEAP1, and PHKG2 were found to be correlated with a good prognosis. These three genes showed a positive correlation with TMB in OV. We also observed that high TMB was more robustly associated with immune response-related gene expression, including CD28, CD40L, and type I IFN family members. Moreover, high TMB was associated with increased T cell infiltration and exhibited a distinct gene signature, which highlights the potential of IFNG, KEAP1, and PHKG2 as predictive markers for T cell infiltration and the tumor microenvironment status in OV. A significant correlation exists between the expression levels of KEAP1 and PHKG2 and TMB in OV cell lines.

Conclusion: In conclusion, our study identified KEAP1, IFNG, and PHKG2 as potential prognostic biomarkers and therapeutic targets in OV. Their expression and mutation burden were correlated with a good prognosis. The association between ferroptosis subtypes, TMB, and survival rates further supports the relevance of these biomarkers. Additionally, the positive correlation between KEAP1, IFNG, and PHKG2 with TMB and immune response-related gene expression highlights their potential as predictive markers for immunotherapy efficacy in OV. The observed association of high TMB with increased T cell infiltration and distinct gene signature further emphasizes its role as a potential biomarker for immune response. Further research is warranted to validate these findings and explore their clinical implications in OV treatment.

Keywords: ferroptosis; immunotherapy; ovarian cancer; tumor microenvironment; tumor mutation burden.

FIGURE 1

The flow chart and possible mechanisms of the analysis.

FIGURE 2

Identification of potential OV antigens. (A) The samples were collected from GTEx (n = 88) and TCGA (n = 354). (B) The Volcano Plot of gene expression. Genome segment reassortment (C, E) and the top mutation rate genes (D, F) in the two different datasets.

FIGURE 3

Identification of OV prognosis and APC infiltration‐related antigens. (A) The flow chart of screening. Kaplan–Meier curves showing the prognostic curves of IFNG, KEAP1, and PHKG2 genes in patients with OV tumors (B–D), high expression of these three genes indicates good prognosis. The correlation of IFNG, KEAP1, and PHKG2 gene expression levels with tumor purity, macrophages, dendritic cells, and B cell infiltration (E–G), p < 0.05, |cor| > 0.3 indicate a correlation. Mut, mutation; OS, overall survival; Over, overexpression.

FIGURE 4

Genomic and transcriptomic analysis of potential target genes. (A) The expression of all three genes, including IFNG, KEAP1, and PHKG2 genes was higher in the tumor tissues, compared to those in the normal tissues. (B) The mutation sites of KEAP1 are located in specific regions within the gene sequence. (C) Similarly, mutation sites within PHKG2 are identified at distinct positions in the gene sequence. (D) Changes in gene copy numbers of IFNG, KEAP1, and PHKG2 genes are observed. *** indicates p < 0.001. CNV, copy number variant.

FIGURE 5

Identification of potential ferroptosis subtypes in OV patients. (A) Cumulative distribution function curve (B) δ area of ferroptosis‐related gene in the TCGA‐OV cohort. Clustering heatmap of TCGA‐OV cohort (C) and GEO cohort (D) (k = 2). (E) Principal component analysis of ferroptosis subtypes. Kaplan–Meier curve showing OS of TCGA‐OV (F) and GEO cohorts (G) of different ferroptosis subtypes. FS2 subtype was associated with a poorer prognosis in both the TCGA‐OV and GEO cohorts. (H) Distribution proportion of stage III and IV OV patients among different ferroptosis subtypes in GEO cohort. CDF, cumulative distribution function; FS1, ferroptosis subtypes 1; FS2, ferroptosis subtypes 2.

FIGURE 6

The correlation between TMB and ferroptosis subtypes. The mutation count (A), tumor mutation burden (B) and MSI sensor score (C) in different ferroptosis subtypes. (D) The waterfall chart shows mutation genes of different ferroptosis subtypes (FS1 and FS2). Mutation burden was correlated with the ferroptosis subtype. *p < 0.05, ns indicates no significance.

FIGURE 7

The correlation of ferroptosis subtypes with ICD and ICP. Twelve ICD‐related genes, including EIF2AK3, FPR1, IFNAR1, and LRP1were identified in TCGA‐OV (A) and GEO (B) cohorts. Thirty‐six and three ICP genes were identified in TCGA‐OV (C) and GEO (D) cohorts, respectively. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

FIGURE 8

Cellular and molecular characteristics of ferroptosis subtypes. Heat map of enrichment scores for 28 immune cell markers in TCGA‐OV (A) and GEO (B) cohorts. (C) The survival curve of Type 2 T helper cell high and low groups (p = 0.018). Gene expression of Type 2 T helper cells in FS1 and FS2 subtypes in TCGA‐OV (D) and GEO (E) cohorts. The ESTIMATE score (F), immune score (G) and matrix score (H) of FS1 and FS2 subtypes in TCGA‐OV and GEO cohorts. *p < 0.05, **p < 0.01, ***p < 0.001.

FIGURE 9

The association between mutational and expression patterns of KEAP1, and PHKG2 and the genomic mutation status of cancer cells. (A) Expression levels of PHKG2 and KEAP1 across six ovarian cancer cell lines. (B) Copy number variations of PHKG2 and KEAP1 in six ovarian cancer cell lines. (C) Relative expression of KEAP1 in TOV‐112D and SKOV‐3 cell lines. (D) Relative expression of PHKG2 in TOV‐112D and SKOV‐3 cell lines. **p < 0.01.