Immune infiltration and a ferroptosis-associated gene signature for predicting the prognosis of patients with endometrial cancer

Abstract

Ferroptosis, a form of programmed cell death induced by excess iron-dependent lipid peroxidation product accumulation, plays a critical role in cancer. However, there are few reports about ferroptosis in endometrial cancer (EC). This article explores the relationship between ferroptosis-related gene (FRG) expression and prognosis in EC patients. One hundred thirty-five FRGs were obtained by mining the literature, retrieving GeneCards and analyzing 552 malignant uterine corpus endometrial carcinoma (UCEC) samples, which were randomly assigned to training and testing groups (1:1 ratio), and 23 normal samples from The Cancer Genome Atlas (TCGA). We established a signature using eight screened FRGs (MDM2, GPX4, PRKAA2, PRNP, SLC11A2, ATP5MC3, PHKG2 and ACO1) related to overall survival using LASSO regression analysis. The samples were divided into low- and high-risk subgroups according to the median risk score. Kaplan-Meier survival curves showed that the low-risk group had better OS. ROC curves showed that this signature performed well in predicting OS (1-, 2-, 3-, and 5-year AUCs of 0.676, 0.775, 0.797, and 0.826, respectively). We systematically analyzed the immune infiltrating profile in UCEC samples from TCGA. Overall, our study identified a novel prognostic signature of 8 FRGs that can potentially predict the prognosis of EC.

Keywords: endometrial cancer; ferroptosis; prognosis.

Figure 1

Construction of the signature. (A) The prognostic analyses for eight ferroptosis-related genes in the training cohort of endometrial cancer using a univariate Cox regression model. Hazard ratio >1 represented risk factors for survival and hazard ratio <1 represented protective factors for survival. (B) Heatmap of eight ferroptosis-related genes in 23 normal samples and 272 endometrial cancer samples. (C) Optimal parameter (λ) selected in the LASSO Cox regression model based on the minimum criteria. (D) The LASSO coefficient profiles of the eight ferroptosis-related genes signature. (E) The distribution and median value of the risk scores in the training cohort. (F) The distribution and median value of the risk scores in the training cohort. (G) Survival statuses of endometrial cancer patients in the training cohort. (H) Survival statuses of endometrial cancer patients in the testing cohort.

Figure 2

Validation of the signature. (A) PCA plot in the training cohort. (B) PCA plot in the training and testing cohorts. (C) K-M survival curve of endometrial cancer patients in the training group. (D) K-M survival curve of endometrial cancer patients in the testing group. (E) Time-dependent ROC curve of endometrial cancer patients in the training group. (F) Time-dependent ROC curve of endometrial cancer patients in the testing group.

Figure 3

Independent prognostic analysis of risk scores and clinical parameters. (A) Univariate Cox regression analysis in the training cohort. (B) The multivariate Cox regression analysis in the training cohort. (C) The univariate Cox regression analysis in the testing cohort. (D) The multivariate Cox regression analysis in the testing cohort.

Figure 4

(A) PPI of eight ferroptosis-related genes (FRGs). (B) Correlation network of the eight FRGs. (C) KEGG analysis of the differentially expressed genes (DEGs) between the low- and high-risk groups in the training cohort. (D) KEGG analysis of the DEGs between the low- and high-risk groups in the testing cohort. (E) GO analysis of the DEGs between the low- and high-risk groups in the training cohort. (F) GO analysis of the DEGs between the low- and high-risk groups in the testing cohort.

Figure 5

Tumor-infiltrating immune cells (TIICs) analysis of 243 endometrial cancer (EC) patients (133 low-risk patients and 110 high-risk patients) (CIBERSORT: P < 0.05). (A, B) Composition of 22 TIICs. (C) Wilcoxon test analysis of 22 TIICs between low- and high-risk EC patients. (D, E) Single-sample gene set enrichment analysis (ssGSEA) of specific immune cells and immune functions (^*p < 0.05; ^**p < 0.01; ^***p < 0.001).

Figure 6

(A–H) The TIMER database results of the correlations between the expression of eight ferroptosis-related genes and immune infiltrating cells in endometrial cancer patients, showing the purity-corrected partial Spearman’s rho value and statistical significance.

Figure 7

Landscape of genetic and expression variation of eight ferroptosis-related genes (FRGs) in endometrial cancer (EC) samples. (A) The alteration frequency of 8 FRGs in 529 EC samples. Each column represented individual patients. The upper bar plot showed TMB. The number on the right indicated the alteration frequency in each regulator. The right bar plot showed the proportion of each variant type. The stacked bar plot below showed fraction of conversions in each sample. (B) The CNV variation frequency of FRGs in EC samples. The height of the column represented the alteration frequency. The deletion frequency, blue dot; The amplification frequency, red dot. (C) The location of CNV alteration of FRGs on 23 chromosomes. (D) The expression of 8 FRGs between normal tissues and EC tissues. Tumor, red; Normal, blue. The upper and lower ends of the boxes represented interquartile range of values. The lines in the boxes represented median value, and black dots showed outliers. The asterisks represented the statistical p value (^*P < 0.05; ^**P < 0.01; ^***P < 0.001).