Prognostic modeling of hepatocellular carcinoma based on T-cell proliferation regulators: a bioinformatics approach

Abstract

Background: The prognostic value and immune significance of T-cell proliferation regulators (TCRs) in hepatocellular carcinoma (HCC) have not been previously reported. This study aimed to develop a new prognostic model based on TCRs in patients with HCC.

Method: This study used The Cancer Genome Atlas-Liver Hepatocellular Carcinoma (TCGA-LIHC) and International Cancer Genome Consortium-Liver Cancer-Riken, Japan (ICGC-LIRI-JP) datasets along with TCRs. Differentially expressed TCRs (DE-TCRs) were identified by intersecting TCRs and differentially expressed genes between HCC and non-cancerous samples. Prognostic genes were determined using Cox regression analysis and were used to construct a risk model for HCC. Kaplan-Meier survival analysis was performed to assess the difference in survival between high-risk and low-risk groups. Receiver operating characteristic curve was used to assess the validity of risk model, as well as for testing in the ICGC-LIRI-JP dataset. Additionally, independent prognostic factors were identified using multivariate Cox regression analysis and proportional hazards assumption, and they were used to construct a nomogram model. TCGA-LIHC dataset was subjected to tumor microenvironment analysis, drug sensitivity analysis, gene set variation analysis, and immune correlation analysis. The prognostic genes were analyzed using consensus clustering analysis, mutation analysis, copy number variation analysis, gene set enrichment analysis, and molecular prediction analysis.

Results: Among the 18 DE-TCRs, six genes (DCLRE1B, RAN, HOMER1, ADA, CDK1, and IL1RN) could predict the prognosis of HCC. A risk model that can accurately predict HCC prognosis was established based on these genes. An efficient nomogram model was also developed using clinical traits and risk scores. Immune-related analyses revealed that 39 immune checkpoints exhibited differential expression between the high-risk and low-risk groups. The rate of immunotherapy response was low in patients belonging to the high-risk group. Patients with HCC were further divided into cluster 1 and cluster 2 based on prognostic genes. Mutation analysis revealed that HOMER1 and CDK1 harbored missense mutations. DCLRE1B exhibited an increased copy number, whereas RAN exhibited a decreased copy number. The prognostic genes were significantly enriched in tryptophan metabolism pathways.

Conclusions: This bioinformatics analysis identified six TCR genes associated with HCC prognosis that can serve as diagnostic markers and therapeutic targets for HCC.

Keywords: GEO; T-cell proliferation regulators; bioinformatic; hepatocellular carcinoma; prognostic model.

Figure 1

Identification and functional enrichment analysis of differentially expressed T-cell proliferation regulators (DE-TCRs). (A) The heatmap of differentially expressed genes. In the middle annotation bar, non-cancerous and HCC samples are indicated in blue and red colors, respectively. The intensity of the color in the heatmap signifies gene expression density per sample, with darker colors indicating higher density. The y-axis of the lower heatmap represents genes (red and blue colors indicate upregulation and downregulation, respectively). (B) Volcano plot of differentially expressed genes. The x-axis shows the log₂ fold-change (FC) values, while the y-axis shows the −log₁₀ (adjusted p-values). Each point represents a gene. The horizontal reference line represents −log₁₀ (0.05) = 1.3, while the vertical reference line represents log₂ (FC) = ± 0.5. Divided by the reference line, upper right genes are upregulated genes (red), and upper left genes are downregulated genes (blue). Gray represents genes with non-significant differences in expression levels. The top 10 upregulated genes and the top 10 downregulated genes with the largest log₂ (FC) values are marked. (C) Venn diagram for DE-TCR identification. Pink and blue represent genes unique to the differentially expressed gene set and the TCR set, respectively. (D) Gene Ontology (GO) enrichment (Top 5). Each column represents a GO term. The color of the column represents different GO categories. The length represents the number of genes enriched in the term. (E) Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment. The color transition from red (high) to blue (low) indicates log₂ (FC) values of the genes. Each pathway is represented by a distinct color.

Figure 2

Prognostic gene screening and analysis. (A) Forest plot of Cox univariate analysis results. The leftmost side lists the prognostic genes. The three columns of numbers on the right represent the hazard ratio (HR) values corresponding to the gene, the 95% confidence interval of the HR value, and the p-value. (B) Least absolute shrinkage and selection operator (LASSO) analysis. The left figure demonstrates cross-validation with the middle line marking the minimum error. The optimal log (lambda) value is determined. Key genes and their coefficients are then identified in the right figure based on the lambda value. (C) Gene set enrichment analysis (GSEA) (Top 5) of characteristic genes. The figure has three sections. The top section displays the enrichment score calculated for each gene, linked into a line graph. The middle section visualizes the rank of each gene within the set. The bottom section depicts the overall rank distribution of all genes. (D) Analysis of the Expression of characteristic genes in hepatocellular carcinoma (HCC) versus control samples.

Figure 3

Predictive performance of the risk model. (A, E) Risk curve and survival status distribution of hepatocellular carcinoma (HCC) samples in the training and validation sets. The x-axis represents the risk score, increasing from left to right. In the upper part, red and blue points indicate high-risk and low-risk patients, respectively. In the lower part, red and blue dots denote deceased and surviving patients, respectively. (B, F) Characteristic gene expression analysis in the training and validation sets. The y-axis lists six characteristic genes. Red and blue colors indicate upregulation and downregulation, respectively. High-risk and low-risk groups are shown in red and blue colors, respectively. (C, G) Kaplan-Meier survival curves of patients in the training and validation sets. Red and blue represent high-risk and low-risk groups, respectively. (D, H) Receiver operating characteristic (ROC) curve of the training and validation sets.

Figure 4

Application of the nomogram model for clinical hepatocellular carcinoma (HCC) cases. (A) Results of univariate and multivariate Cox regression analysis: The leftmost column shows the risk score and clinical characteristics, while the two right columns show the corresponding p-values and hazard ratio (HR) values. (B) Nomogram. (C) Nomogram calibration curves for predicting 1-year, 2-year, and 3-year survival rates. The x-axis shows the predicted event rate, while the y-axis shows the observed event rate (both ranging from 0 to 1). (D) The receiver operating characteristic (ROC) curves for predicting the 1-year, 2-year, and 3-year survival rates. (E) The upper panel shows the pathways activated and inhibited in the high-risk and low-risk groups (Top 5). The horizontal axis represents the samples (pink and blue colors indicate the high-risk and low-risk groups, respectively). The vertical axis represents the pathway (red and blue colors indicate the enriched and suppressed pathways, respectively). The lower panel shows the gene set variant analysis (GSVA) results.

Figure 5

Correlation between risk scores and the tumor microenvironment of hepatocellular carcinoma (HCC). (A) The Tumor Immune Estimation Resource (TIMER) algorithm assesses the relationship between the risk score and immune cell infiltration. Yellow and blue represent the high-risk and low-risk groups, respectively. Each image is divided into two parts: correlation analysis (left) and differential analysis (right). (B) Analysis of differential immune cell infiltration levels using the single-sample gene set enrichment analysis (ssGSEA) algorithm. (C) Analysis of differential immune-related pathways using the ssGSEA algorithm. * p< 0.05, ** p< 0.01, *** p< 0.001, **** p< 0.0001.

Figure 6

Prognostic genes can affect the immunotherapy efficacy in patients with hepatocellular carcinoma (HCC). (A) Expression of immune checkpoints in the high-risk and low-risk groups. (B) Correlation between characteristic genes and differential immune checkpoints (Dot size represents significance; red and blue colors indicate positive and negative correlations, respectively; darker shades indicate stronger correlations). (C) Immunephenoscore (IPS) of the high-risk and low-risk groups (red and green colors indicate high-risk and low-risk, respectively). Each image features a density plot at the top and a scatter plot with a box plot at the bottom. (D) Tumor immune dysfunction and exclusion (TIDE) scores of the high-risk and low-risk groups (yellow and blue colors indicate the high-risk and low-risk groups, respectively). (E) Immune response ratio of patients in the high-risk and low-risk groups. * p< 0.05, ** p< 0.01, *** p< 0.001, **** p< 0.0001.

Figure 7

Differential characteristics of different subtypes of hepatocellular carcinoma (HCC). (A, B) Consistent clustering of characteristic genes in patients with HCC. (C) Kaplan-Meier (K-M) curves of HCC subtypes. The x-axis represents overall survival time in days, while the y-axis shows survival probability. Different colors represent different HCC subtypes. (D) Differential immune cell infiltration levels in HCC subtypes. * p< 0.05, ** p< 0.01, **** p< 0.0001.

Figure 8

Differential drug sensitivity and mutation profiles between the high-risk and low-risk groups. (A) The half-maximal inhibitory concentration (IC50) values of the top 10 drugs in patients belonging to the high-risk and low-risk groups (Top 10). (B) Mutation profiles of the high-risk and low-risk groups (the top and bottom sections represent the low-risk and high-risk groups, respectively). (C) Analysis of characteristic gene mutations and Kaplan-Meier (K-M) curves based on tumor mutational burden (TMB) and risk score. (D) Copy number variation of characteristic genes (red and blue represent amplification and deletion, respectively). **** p< 0.0001.

Figure 9

Validation of prognostic genes. The mRNA expression levels of five prognostic genes in cancer and para-cancerous tissues. * p< 0.05.