Identification and validation of a novel cuproptosis-related genes signature associated with prognosis, clinical implications and immunotherapy of hepatocellular carcinoma

Abstract

Background: Cuproptosis is a novel type of regulated cell death and is reported to promote tumor occurrence and progression. However, whether a cuproptosis-related signature has an impact on hepatocellular carcinoma (HCC) is still unclear. Materials and methods: We analyzed the transcriptome data of HCC from The Cancer Genome Atlas (TCGA) and International Cancer Genome Consortium (ICGC) database, and searched for tumor types with different cuproptosis patterns through consistent clustering of cuproptosis genes. We then constructed a Cuproptosis-Related Genes (CRGs)-based risk signature through LASSO COX regression, and further analyzed its impact on the prognosis, clinical characteristics, immune cell infiltration, and drug sensitivity of HCC. Results: We identified the expression changes of 10 cuproptosis-related genes in HCC, and all the patients can be divided into two subtypes with different prognosis by applying the consensus clustering algorithm. We then constructed a cuproptosis-related risk signature and identified five CRGs, which were highly correlated with prognosis and representative of this gene set, namely G6PD, PRR11, KIF20A, EZH2, and CDCA8. Patients in the low CRGs signature group had a favorable prognosis. We further validated the CRGs signature in ICGC cohorts and got consistent results. Besides, we also discovered that the CRGs signature was significantly associated with a variety of clinical characteristics, different immune landscapes and drug sensitivity. Moreover, we explored that the high CRGs signature group was more sensitive to immunotherapy. Conclusion: Our integrative analysis demonstrated the potential molecular signature and clinical applications of CRGs in HCC. The model based on CRGs can precisely predict the survival outcomes of HCC, and help better guide risk stratification and treatment strategy for HCC patients.

Keywords: cuproptosis-related genes; drug sensitivity; hepatocellular carcinoma; prognosis model; tumor microenvironment.

FIGURE 1

Differential expression of cuproptosis-related genes in hepatocellular carcinoma (HCC) and normal tissues. (A) Volcano plot of differential expression (blue represents downregulation in HCC, red represents upregulation in HCC, gray represents insignificant, and cuproptosis-related genes are marked in the figure). (B) The expression of 10 cuproptosis-related genes in HCC and normal tissues. (C) Partial results of protein-protein interactions (PPI) network map and gene ontology (GO) enrichment of 10 cuproptosis-related genes. (D) Heatmap of expression correlation of 10 cuproptosis-related genes in TCGA.

FIGURE 2

Differential expression of cuproptosis-related genes in different clinical feature groups. (A) Expression of 10 cuproptosis-related genes in different alpha-fetoprotein groups (<100 mg/dL, 100–400 mg/dL, >400 mg/dL); (B) DLD expression differences in different alpha-fetoprotein groups; (C–E) The difference of GLS expression in different age, BMI and different tumor stage groups (ns: p >0.05; *: p <0.05; **: p <0.01; ***: p <0.001).

FIGURE 3

The cuproptosis-related genes divide hepatocellular carcinoma into two subtypes. (A) The sample squareness of the consistent clustering (number of classifications = 2); (B) The cumulative distribution map of the consistent clustering; (C) The principal component analysis graph of the two hepatocellular carcinoma subtypes; (D) The cuproptosis-related genes in the two categories Expression heatmap of hepatocellular carcinoma; (E) Boxplot of cuproptosis-related genes expression difference between two types of hepatocellular carcinoma (ns: p >0.05; *: p <0.05; **: p <0.01; ***: p <0.001; ****: p <0.0001).

FIGURE 4

Differences in immune cell infiltration among hepatocellular carcinoma subtypes. (A) ESTIMATE algorithm calculates differences in stromal and immune scores between subtypes; (B) ESTIMATE algorithm calculates differences in tumor purity scores between subtypes; (C) GSVA-cell report algorithm calculates differences in immune cell infiltration between subtypes; (D) MCP-counter calculates differences in subtypes differences in immune cell infiltration (ns: p >0.05; *: p <0.05; **: p <0.01; ***: p <0.001; ****: p <0.0001).

FIGURE 5

Functional analysis of different subtypes and the construction of cuproptosis-related genes (CRGs) signature in hepatocellular carcinoma. (A) Gene ontology (GO) enrichment analysis results in the two CRGs subtypes. (B) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analysis results in the two CRGs subtypes: (C) Lasso regression coefficients of each variable with L1 norm; (D) Lambda logarithm value in Lasso regression and the relationship with the error (the dotted line is the range that Lambda can choose); (E) The coefficient values of the Lasso regression screening variables.

FIGURE 6

The prognostic effect of cuproptosis-related genes signatures in hepatocellular carcinoma. (A) Survival analysis of patients in TCGA cohort based on cuproptosis score; (B) Time-dependent receiver operating characteristic (ROC) curve of cuproptosis score in TCGA dataset; (C) Multivariate Cox analysis results in TCGA dataset; (D) Survival analysis of patients in LIRI-JP cohort based on cuproptosis score; (E) ROC curve of cuproptosis score in the LIRI-JP dataset; (F) Multivariate Cox analysis results in LIRI-JP dataset.

FIGURE 7

Cuproptosis-related genes signature correlates with clinical features. (A–L) Cuproptosis scores among different clinical features of hepatocellular carcinoma, including alpha-fetoprotein, total bilirubin, albumin, fibrosis score, histological type, histological grade, tumor stage, sex, BMI, age and molecular subtypes in TCGA cohort; (M–O) The differences of cuproptosis scores by gender, age, and tumor stage in the LIRI-JP cohort (ns: p >0.05; *: p <0.05; **: p <0.01; ***: p <0.001; ****: p <0.0001).

FIGURE 8

Drug sensitivity analysis of cuproptosis-related genes (CRGs)signature groups in Genomicsof Drug Sensitivity in Cancer (GDSC), Cancer Cell Line Encyclopedia (CCLE) and Cancer Therapeutics Response Portal (CTRP) databases. (A) Correlations between CRGs score and drug area under the curve (AUC) in GDSC (p < 0.05 was selected); (B) CRGs signature of each cell line under different drug treatments with significant negative correlation in GDSC database (* represents p <0.05, ** represents p <0.01, *** represents p <0.001). (C) Correlations between CRGssignature and drug AUC in CTRP database (p < 0.05 and drug display with negative correlation were selected). (D) Correlations between CRGs signature and drug AUC in CTRP database (select p < 0.05 and positive correlated drugs); (E) The CRGs signature of each cell line under different drug treatments with significant negative correlations in CTRP database (ns: p >0.05; *: p <0.05); (F) The differences in theCRGs signature of each cell line under different drug treatments with significant positive correlation in CTRP (ns: p >0.05; *: p <0.05; **: p <0.01); (G) The correlations between the CRGs signature in CCLE database and the IC50 of drugs (p < 0.05 and positive correlations were selected); (H) The differences in the CRGs signature of each cell line under different drug treatments with significant positive correlations in CCLE (*represents p <0.05; *represents p <0.01; **represents p <0.001).

FIGURE 9

Cuproptosis-related genes (CRGs) signature was correlated with immune score and immune infiltrating cells. (A–C) Differences of matrix score (A), immune score (B) and ESTIMATE score (C) between the two groups with high and low CRGs signature; (D–F) Correlations of CRGs signature and matrix score (D), immune score (E) and ESTIMATE score (F); (G) Differences in immune cell score between two CRGs groups with high and low CRGs signature (calculated by Cibersort) (ns: p >0.05; ****: p <0.0001).

FIGURE 10

The relationship of immunotherapy responses and immune checkpoints in different cuproptosis-related genes (CRGs) signature groups. (A) Differences in CRGs signature of patients with different treatment outcomes (beneficial or non-beneficial) (B) the proportion of immunotherapy benefit and non-benefit between the two CRGs groups; (C) Heatmap of correlation analysis between CRGs groups and different immune checkpoints (*represents p <0.05; *represents p <0.01; **represents p <0.001; ***represents p <0.001); (D) Differences between the immune checkpoints and two CRGs groups (ns: p >0.05; *: p <0.05; **: p <0.01; ***: p <0.001; ****: p <0.0001).