Identification of Novel Characteristics in TP53-Mutant Hepatocellular Carcinoma Using Bioinformatics

Abstract

Background: TP53 mutations are the most frequent mutations in hepatocellular carcinoma (HCC) and affect the occurrence and development of this cancer type. Therefore, it is essential to clarify the function and mechanism of TP53 mutations in HCC. Methods: We performed a sequence of bioinformatic analyses to elucidate the characteristics of TP53 mutations in HCC. We downloaded the data of hepatocellular carcinoma from The Cancer Genome Atlas database and used different R packages for serial analyses, including gene mutation analysis, copy number variation analysis, analysis of the tumor mutational burden and microsatellite instability, differential gene expression analysis, and functional enrichment analysis of TP53 mutations, and performed gene set enrichment analysis. We established a protein-protein interaction network using the STRING online database and used the Cytoscape software for network visualization, and hub gene screening. In addition, we performed anticancer drug sensitivity analysis using data from the Genomics of Drug Sensitivity in Cancer. Immune infiltration and prognosis analyses were also performed. Results: Missense mutations accounted for a great proportion of HCC mutations, the frequency of single nucleotide polymorphisms was high, and C > T was the most common form of single nucleotide variations. TP53 had a mutation rate of 30% and was the most commonly mutated gene in HCC. In the TP53 mutant group, the tumor mutational burden (p < 0.001), drug sensitivity (p < 0.05), ESTIMATE score (p = 0.038), and stromal score (p < 0.001) dramatically decreased. The Cytoscape software screened ten hub genes, including CT45A1, XAGE1B, CT55, GAGE2A, PASD1, MAGEA4, CTAG2, MAGEA10, MAGEC1, and SAGE1. The prognostic model showed a poor prognosis in the TP53 mutation group compared with that in the wild-type group (overall survival, p = 0.023). Univariate and multivariate cox regression analyses revealed that TP53 mutation was an independent risk factor for the prognosis of HCC patients (p <0.05). The constructed prognostic model had a favorable forecast value for the prognosis of HCC patients at 1 and 3 years (1-year AUC = 0.752, 3-years AUC = 0.702). Conclusion: This study further deepened our understanding of TP53-mutated HCC, provided new insights into a precise individualized therapy for HCC, and has particular significance for prognosis prediction.

Keywords: TP53 mutation; bioinformatics; hepatocellular carcinoma; immune infiltration; prognostic model..

FIGURE 1

TP53 gene mutation and copy number variation in LIHC patients. (A): Summary of mutation information in LIHC patients. Missense mutations account for the majority of different mutation types; SNP frequency is higher than insertion or deletion, and C > T is the most common mutation of SNV; tumor mutation burden in specific samples and the top 10 mutated genes. (B): Distribution map of amino acid variation of TP53 protein in TCGA-LIHC. (C,D): GISTIC2.0 copy number variation analysis results, red represents copy number amplification, blue represents copy number deletion.

FIGURE 2

Analysis of biological characteristics of TP53 gene mutation level. (A): TMB in patients with TP53 mutations was significantly reduced (p <0.001). (B): There is no significant difference in MSI (p = 0.063). (C): Cosmic Signature heat map of TP53 mutation patients. (D): Cosmic Signature heat map of TP53 non-mutant patients.

FIGURE 3

Drug sensitivity analysis of TP53 gene mutation. (A): The relationship between gene mutation level in TCGA-LIHC and different kinds of drugs. (B): Changes of gene mutation levels in different carcinogenic signal pathways. (C): Mutation distribution of main genes in RTK-Ras signaling pathway. (D): Sensitivity of TP53 mutation to different chemotherapeutics and small molecule anticancer drugs.

FIGURE 4

Differential expression and functional enrichment analysis based on TP53 gene mutation. (A): There was no significant difference in the expression level between TP53 mutation and non-mutation (p = 0.077). (B): Volcano map of differentially expressed genes between TP53 mutant and non-mutant groups. GO analysis of differentially expressed genes, (C): Biological processes, (D): cellular component, (E): molecular function. (F): KEGG pathway enrichment analysis.

FIGURE 5

GSEA analysis.

FIGURE 6

PPI network construction and hub-gene screening. (A): PPI network of differentially expressed genes. (B): Get the top 10 genes according to the MCC algorithm, the redder the color, the more important.

FIGURE 7

Effect of TP53 mutation on the immunological characteristics of TCGA-LIHC patients. (A): ImmuneScore has no significant difference (p = 0.4891). (B): ESTIMATE Score decreased significantly (p = 0.038). (C): StromalScore was significantly reduced (p <0.001). (D): Cluster heat map of TP53 mutation and immune gene. (E): Cluster heat map of TP53 mutation and HLA gene family.

FIGURE 8

CIBERSORT analysis. (A): Abundance of immune cells in TP53 mutant TCGA-LIHC samples. (B): Correlation heat map between different immune cell infiltration.

FIGURE 9

Analysis of clinical correlation of TP53 mutation and construction of prognostic model. (A): There was no significant difference in age between the TP53 mutation group and the non-mutation group (p = 0.148). (B): The BMI of the TP53 mutation group was significantly low (p = 0.023). (C): There is a significant difference in gender (p = 0.008). (D,E): No significant difference in clinical T staging (p >0.05). (F): There are significant differences in pathological grading (p <0.001). (G): Survival analysis showed that OS with TP53 mutation was significantly shortened (p = 0.023). (H): nomogram. (I): Calibration curve. (J): Time-dependent ROC curve.