Mitochondrial metabolic reprogramming-mediated immunogenic cell death reveals immune and prognostic features of clear cell renal cell carcinoma

Abstract

Background: Mitochondrial metabolic reprogramming (MMR)-mediated immunogenic cell death (ICD) is closely related to the tumor microenvironment (TME). Our purpose was to reveal the TME characteristics of clear cell renal cell carcinoma (ccRCC) by using them.

Methods: Target genes were obtained by intersecting ccRCC differentially expressed genes (DEGs, tumor VS normal) with MMR and ICD-related genes. For the risk model, univariate COX regression and K-M survival analysis were used to identify genes most associated with overall survival (OS). Differences in the TME, function, tumor mutational load (TMB), and microsatellite instability (MSI) between high and low-risk groups were subsequently compared. Using risk scores and clinical variables, a nomogram was constructed. Predictive performance was evaluated by calibration plots and receiver operating characteristics (ROC).

Results: We screened 140 DEGs, including 12 prognostic genes for the construction of risk models. We found that the immune score, immune cell infiltration abundance, and TMB and MSI scores were higher in the high-risk group. Thus, high-risk populations would benefit more from immunotherapy. We also identified the three genes (CENPA, TIMP1, and MYCN) as potential therapeutic targets, of which MYCN is a novel biomarker. Additionally, the nomogram performed well in both TCGA (1-year AUC=0.862) and E-MTAB-1980 cohorts (1-year AUC=0.909).

Conclusions: Our model and nomogram allow accurate prediction of patients' prognoses and immunotherapy responses.

Keywords: TME; clear cell renal cell carcinoma; immunogenic cell death; immunotherapy; mitochondrial metabolic reprogramming.

Figure 1

Screening process of target genes. (A) Determine the appropriate soft threshold power equal to 11. (B) Generate a hierarchical clustering tree of genes based on TOM, with different colors representing different modules. (C) The modules are closely related to the clinical features and TME of ccRCC, with the red module being the most important. (D) Differential ranking map of DEGs between tumor and normal tissues. (E) Venn diagram of DEGs, MMR, and ICD-related genes taking the intersection. (F) Volcano plot of 140 target genes. (G) Mutation waterfall plot of target genes. H Enrichment bubble map of biological functions and pathways of target genes.

Figure 2

Development of risk model. (A) Prognostic genes were screened based on univariate COX regression and K-M survival analysis. (B) LASSO regression was used to construct a 12-gene-based risk model. (C) PCA analysis showed that the model genes had a good discriminatory ability. (D, E) Risk scores of TCGA and E-MTAB-1980 cohorts were correlated with survival status and model genes. (F, G) K-M survival analysis for high- and low-risk groups.

Figure 3

Relationship of risk scores to clinical variables. (A, B) K-M survival analysis and comparison of risk scores for high and low-risk groups in patients <=65 years and >65 years. (C, D) K-M survival analysis of high and low-risk groups of male and female patients with comparison of risk scores by gender. (E, F) K-M survival analysis of high- and low-risk groups of early and late-stage patients, with a comparison of risk scores by stage.

Figure 4

Immune landscapes in different risk groups. (A) Estimated scores, immune scores, stromal scores, and tumor purity for high- and low-risk groups. (B) 22 levels of immune cell infiltration. (C, D) Expression levels of immunoinhibitor and immunostimulator agents in TCGA and E-MTAB-1980 cohorts. “*”<0.5, "**" <0.01, and "***"<0.001.

Figure 5

Functional and pathway enrichment analysis. (A) GO and KEGG enrichment results of DEGs. (B) GSVA in high and low-risk groups. (C) GSEA of high and low-risk groups. (D) Comparison of differences in immune-related functions between high and low-risk groups in TCGA and E-MTAB-1980 cohorts. “*”<0.5, "**" <0.01, and "***"<0.001.

Figure 6

Immunotherapy and drug prediction. (A) TMB levels in high and low-risk groups, correlation of risk scores with TMB. (B) Levels of risk scores in MSI and MSS groups, correlation of risk scores with MSI scores. (C) Levels of easier scores in high and low-risk groups, correlation of risk scores with easier scores. (D) Differences in sensitivity to different drugs in high and low-risk groups. “*”<0.5, "**" <0.01, and "***"<0.001

Figure 7

Comprehensive analysis of key model genes. (A) Pan-cancer analysis of gene expression levels. (B) Pan-cancer survival analysis. (C) Methylation differences between tumor and normal tissues. (D) Relationship between methylation levels and gene expression. (E) Single-cell analysis of key model genes. “****” <0.0001.

Figure 8

Clinical features and immunological correlates. (A) Relationship between gene expression levels and stage. (B) Relationship between gene expression levels and grade. (C) Key model genes classify ccRCC into 6 immune subtypes. (D) The relationship between gene expression levels and the abundance of immune cell infiltration.

Figure 9

Construction of Nomogram. (A, B) Univariate and multivariate COX regression analysis. (C) Nomograms predict 1-, 3-, and 5-year survival probabilities. (D) Calibration plots for the TCGA and E-MTAB-1980 cohorts. (E, F) ROC curves of nomograms and risk scores in the TCGA and E-MTAB-1980 cohorts.

Figure 10

Expression validation of MYCN. (A) Expression levels of MYCN in TCGA unpaired and paired samples. (B) Expression levels of MYCN in HK-2, ACHN, and OSRC cell lines.