Characterization of a ferroptosis-related gene signature predicting survival and immunotherapeutic response in lung adenocarcinoma

Abstract

Lung cancer remains the leading cause of cancer-related death worldwide, and drug resistance represents the main obstacle responsible for the poor mortality and prognosis. Here, to identify a novel gene signature for predicting survival and drug response, we jointly investigated RNA sequencing data of lung adenocarcinoma patients from TCGA and GEO databases, and identified a ferroptosis-related gene signature. The signature was validated in the validation set and two external cohorts. The high-risk group had a reduced survival than the low-risk group (P < 0.05). Moreover, the established gene signature was associated with tumor mutation burden, microsatellite instability, and response to immune checkpoint blockade. In addition, four candidate oncogenes (RRM2, SLC2A1, DDIT4, and VDAC2) were identified to be candidate oncogenes using in silico and wet experiments, which could serve as potential therapeutic targets. Collectively, this study developed a novel ferroptosis-related gene signature for predicting prognosis and drug response, and identified four candidate oncogenes for lung adenocarcinoma.

Keywords: drug response; ferroptosis; gene signature; lung adenocarcinoma; prognosis.

Figure 1

Establishment of the prognostic ferroptosis-related gene signature. (A) Volcano plot showed differentially expressed genes in lung adenocarcinoma. Blue dots represent downregulated genes, red dots represent upregulated genes, and grey dots represent unchanged genes. (B) There were 20 overlapping genes among 20205 DEGs, 5905 prognostically relevant genes, and 259 ferroptosis genes. (C) 11 eligible genes were obtained in the univariate Cox regression model. (D, E) Seven genes (RRM2, IL33, SLC2A1, PEBP1, DDIT4, VDAC2, and FLT3) were acquired in LASSO regression model.

Figure 2

Validation of the predictive ability of the ferroptosis-related gene signature. (A–C) A receiver operating characteristic (ROC) curve was performed and area under the curve (AUC) was calculated in the validation set and two external test sets (GSE72094 and GSE8894). (D–F) Survival analysis was performed in the TCGA cohort, the GSE72094 cohort, and the GSE8894, respectively. (G–I) Principal component analysis of genes consisting the prognostic ferroptosis-related signature revealed a distinct expression pattern between the low- and the high-risk groups in dimensionality.

Figure 3

Comparison of the predicting capacity between the signature and TNM staging. (A) Five-year AUC value for the ferroptosis-related gene signature was 0.655. (B) Five-year AUC value for T staging was 0.603. (C) Five-year AUC value for N staging was 0.634. (D) Five-year AUC value for M staging was 0.540. (E–G) The risk score was significantly augmented as TNM staging increased.

Figure 4

Investigation of the signature-related biological function. (A) Enriched terms of biological process in gene ontology analysis. (B) Enriched terms of cell component in gene ontology analysis. (C) Enriched terms of molecular function in gene ontology analysis. (D) Enriched terms of KEGG pathway in gene ontology analysis. (E, F) Enriched terms of KEGG pathway in gene set enrichment analysis (GSEA).

Figure 5

Investigation of tumor immune microenvironment of lung adenocarcinoma. (A) Comparison of tumor-infiltrating immune cells between cancerous and normal tissue. (B) Heatmap analysis of tumor-infiltrating immune cells in lung adenocarcinoma. (C) Correlations analysis of tumor-infiltrating immune cells in lung adenocarcinoma. (D–F) The risk score based on the ferroptosis-related gene signature was negatively correlated with CD8 T cell, CD4 T cell, and follicular helper T cell. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001.

Figure 6

Profiling of somatic nucleotide variation for lung adenocarcinoma patients. (A) The waterfall plot showed that the high-risk group had a nucleotide variation rate of 95.93%. (B) The waterfall plot showed that the low-risk group had a nucleotide variation rate of 80.83%. (C) The bar plot showed that the risk score was critically increased in the high-mutation group than in the low-mutation group. (D) The box plot demonstrated that the risk score was critically increased in the high-mutation group than in the low-mutation group. (E) Tumor mutation burden (TMB) for lung adenocarcinoma patients. (F) The high-risk group had an increased TMB level as compared with the low-risk group. (G) The TMB levels were also positively correlated with the risk score. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001.

Figure 7

Effects of the ferroptosis-related gene signature on response to ICB. (A) The high-risk patients had a higher MSI level than the low-risk patients based on ssGSEA approach. (B) The risk score was positively correlated with MSI level based on ssGSEA approach. (C) MSI-H patients had a higher risk score level than MSI-L/MSS patients based on UCSCXenaShiny approach. (D) The risk score was positively correlated with MSI level based on UCSCXenaShiny approach. (E) The high-risk group had a higher proportion of stable disease (SD) and progressive disease (PD) than the low-risk group. (F, G) The SD/PD patients had an increased risk score than the CR/PR group. MSI: microsatellite instability; MSS: microsatellite stability, MSI-L: microsatellite instability-low, MSI-H: microsatellite instability-high, SD: stable disease, PD: progressive disease, CR: complete response, PR: partial response. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001.

Figure 8

Identification of the hub genes in lung adenocarcinoma. (A) Survival analysis of seven candidate hub genes in lung adenocarcinoma. (B) Expression level analysis of seven candidate hub genes in lung adenocarcinoma. (C) The association of the hub gene expression levels and the levels of proliferation, invasion, metastasis, cell cycle and EMT in lung adenocarcinoma. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001.

Figure 9

Validation of the expression levels of seven genes in lung adenocarcinoma. (A) RT-qPCR assay detected mRNA levels of the seven genes related with ferroptosis in lung adenocarcinoma. (B) Immunohistochemistry assay detected protein levels of the seven genes related with ferroptosis in lung adenocarcinoma. * represents P < 0.05, ** represents P < 0.01, *** represents P < 0.001.