A ferroptosis associated gene signature for predicting prognosis and immune responses in patients with colorectal carcinoma

Abstract

Background: Colorectal carcinoma (CRC) is one of the most prevalent malignancies globally. Ferroptosis, a novel type of cell death, is critical in the development and treatment of tumors. Objective: This study was designed to establish a genetic signature for ferroptosis which has a predictive effect on the outcomes and immunotherapeutic response of CRC. Methods: Data of CRC patients were retrieved from TCGA and GEO databases. The genes associated with ferroptosis were obtained from GeneCards. The genetic signature for ferroptosis was identified by performing Cox regression analysis. Kaplan-Meier and ROC analysis were performed to assess the prognosis role of the genetic signature. CIBERSORT tool was used to identify a potential association of the genetic signature with the immune cells. The potential immunotherapeutic signatures and drug sensitivity prediction targeting this signature were also discussed. Immunohistochemistry was used to detect expression of ferroptosis-associated genes in CRC tissues and adjacent tissues. Results: A ferroptosis-associated gene signature comprised of three genes (CDKN2A, FDFT1, and ACSL6) was developed for prediction of prognosis and evaluation of immune responses in CRC. Patients in the high-risk group tended to have a poor prognosis. In CRC, the ferroptosis-associated gene signature may function as independent predictors. Additionally, the expressional levels of the immune checkpoint proteins PD-L1 and CTLA-4 were substantially increased in the high-risk group. Moreover, we can distinguish between patients based on their immunotherapeutic responses more effectively if we categorize them by this signature. Additionally, candidate compounds were identified for the differentiation of CRC subtypes. Conclusion: The ferroptosis-associated gene signature identified in this study is effective in predicting the prognosis and evaluating immunotherapeutic response in CRC patients, and provides us with novel insights into the potential effect of ferroptosis targeted treatment on CRC.

Keywords: colorectal carcinoma; ferroptosis; gene signature; immune response; prognosis.

FIGURE 1

Identification of different expressional levels of ferroptosis-associated genes and their prognostic significance in CRC. (A) Different expressional levels of ferroptosis-associated genes in TCGA cohort were displayed in the heatmap and (B) the volcano map; (C) PPI network indicated the interactions among the candidate genes from the STRING; (D,E) a ferroptosis-associated gene signature was constructed by using univariate and multivariate Cox regression analyses to exert a predictive effect on the prognosis of CRC.

FIGURE 2

Prognostic effect of the ferroptosis-associated gene signature in CRC patients. (A) Heatmaps showed the expressional levels of 3 ferroptosis-associated genes respectively in low-and high-risk groups of TCGA and GEO cohorts; (B) the patients were grouped according to the ferroptosis-associated risk score. (C) The scatter plot demonstrated a difference in the survival status of CRC patients between low- and high-risk groups. The dot indicates the survival status of CRC patient, which ranked according to risk score in ascending order. (D) Mortality rates of the low- and high-risk groups; (E) Kaplan-Meier curves revealed a survival difference between two risk groups in TCGA and GEO cohorts.

FIGURE 3

Prognostic significance of the ferroptosis-associated gene signature in CRC patients from the TCGA and GEO cohorts. (A,B) ROC curves indicated the accuracy of the ferroptosis-associated gene signature in the prediction of survival rates at 1-, 3-, and 5-years; (C–F) The independent prognostic significance of the ferroptosis-associated gene signature in OS in CRC patients using univariate and multivariate Cox analyses.

FIGURE 4

Constructing and verifying a nomogram. (A,C) The prognostic nomogram developed according to the risk scores of ferroptosis-associated genes and clinicopathological features predicted the 1‐, 3‐, and 5-year OS of CRC patients in the TCGA and GEO cohorts. (B,D) Calibration curves of nomogram on consistency between predicted and observed 1‐, 3‐, and 5-year survival in the TCGA and GEO cohorts.

FIGURE 5

Correlation between ferroptosis-associated gene expressions and clinicopathological features in CRC patients. (A,B) Expression patterns of 3 ferroptosis-associated genes in different stages in TCGA and GEO cohorts; (C,D) Expression levels of 3 ferroptosis-associated genes in CRC at different stages in TCGA and GEO cohorts. *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 6

A difference in immune cell landscape between low and high ferroptosis-associated risks in CRC patients. (A) Relative distribution of 22 immune cells in all samples from TCGA and GEO cohorts; (B) The contents of immune cells in low- and high-risk groups. The low-risk group is indicated in green, the high-risk group is indicated in red.

FIGURE 7

Correlation of ferroptosis-associated gene signature with immunity microenvironment. (A) Heatmaps of gene profiles of the cancer-immunity cycle in two risk groups in the TCGA and GEO cohorts; (B) comparison of the common differential immune gene expression between two risk groups in the TCGA and GEO cohorts; (C) comparison of immune checkpoint expression between two risk groups. *p < 0.05, **p < 0.01, and ***p < 0.001.

FIGURE 8

GSEA for identification of ferroptosis-associated signaling pathways. (A) GSEA of related signaling pathways in the high-risk group in TCGA cohort. (B) GSEA of related signaling pathways in the high-risk group in GEO cohort.

FIGURE 9

Estimation of the tumor immune microenvironment and cancer immunotherapy response using the ferroptosis-associated gene signature in the TCGA entire set. (A) TIDE prediction difference in the high- and low-risk patients. (B,C) Waterfall plot displays mutation information of the genes with high mutation frequencies in the high-risk group (B) and low-risk group (C). (D) TMB difference in the high- and low-risk patients. (E) Kaplan-Meier curve analysis of OS is shown for patients classified according to the TMB status. (F) Kaplan-Meier curve analysis of OS is shown for patients classified according to the TMB status and ferroptosis-associated gene signature. (G) Comparison of Stromal_score, Immune_score and ESTIMATE_Score between two groups.

FIGURE 10

(A–E) Identification of novel candidate compounds targeting the ferroptosis-associated gene signature.

FIGURE 11

Verification of the expressions of ferroptosis-associated genes in CRC and normal colorectal tissues using HPA database.

FIGURE 12

Verify the translational expression of ferroptosis-associated in CRC and normal tissues. (A) The representative images of ACSL6 in the adjacent tissues and tumor tissues. (B) The representative images of CDKN2A in the adjacent tissues and tumor tissues.