Risk Predictive Model Based on Three DDR-Related Genes for Predicting Prognosis, Therapeutic Sensitivity, and Tumor Microenvironment in Hepatocellular Carcinoma

Abstract

Hepatocellular carcinoma (HCC) is the seventh most common malignancy and the second most common cause of cancer-related deaths. Tumor mutational load, genomic instability, and tumor-infiltrating lymphocytes were associated with DNA damage response and repair gene changes. The goal of this study is to estimate the chances of patients with HCC surviving their disease by constructing a DNA damage repair- (DDR-) related gene profile. The International Cancer Genome Consortium (ICGC) and The Cancer Genome Atlas (TCGA) provided us with the mRNA expression matrix as well as clinical information relevant to HCC patients. Using Cox regression and LASSO analysis, DEGs strongly related to general survival were discovered in the differentially expressed gene (DEG) study. In order to assess the model's accuracy, Kaplan-Meier (KM) and receiver operating characteristic (ROC) were used. In order to compute the immune cell infiltration score and immune associated pathway activity, a single-sample gene set enrichment analysis was performed. A three-gene signature (CDC20, TTK, and CENPA) was created using stability selection and LASSO COX regression. In comparison to the low-risk group, the prognosis for the high-risk group was surprisingly poor. In the ICGC datasets, the predictive characteristic was confirmed. A receiver operating characteristic (ROC) curve was calculated for each cohort. The risk mark for HCC patients is a reliable predictor according to multivariate Cox regression analysis. According to ssGSEA, this signature was highly correlated with the immunological state of HCC patients. There was a significant correlation between the expression levels of prognostic genes and cancer cells' susceptibility to antitumor therapies. Overall, a distinct gene profile associated with DDR was identified, and this pattern may be able to predict HCC patients' long-term survival, immune milieu, and chemotherapeutic response.

Figure 1

A list of possible DDR-related genes identified in the TCGA cohort. DEGs between nearby normal specimens and HCC specimens are calculated using a Venn diagram (a). (b) Expression of nine genes that overlap between neighboring normal tissues and HCC tissues. (c) Forest plots showing the associations between OS and the expression of 9 overlapping genes. (d) Correlation network of candidate genes.

Figure 2

Gene signatures associated with DDR were identified in TCGA datasets using LASSO regression analysis. (a) Choosing the optimal LASSO model parameter (lambda). (b) LASSO coefficient profiles of the nine prognostic DDR genes.

Figure 3

The performance of DDR-related gene signature in TCGA and ICGC datasets. Based on Kaplan-Meier analysis of the (a) TCGA and (c) ICGC datasets, patients with lower risk ratings had greater overall survival than those with higher risk scores. ROC curves were used to assess the prognostic signature's accuracy in the (b) TCGA and (d) ICGC datasets.

Figure 4

The OS by Cox regression model's univariate and multivariate evaluations. Datasets (a, b) from the TCGA. Datasets (c, d) from the ICGC.

Figure 5

The risk score in different groups divided by clinical factors. TCGA cohort (a–d) and ICGC cohort (e, f): (a) age; (b) gender; (c) grade; (d) clinical stage; (e–g) age, gender, and clinical stage.

Figure 6

A correlation between the tumor microenvironment and risk markers. The characteristics of 16 immune cells (a, c) and 13 immune-related activities (b, d) were illustrated in boxplots. (e) Comparison of risk scores across several subtypes of immune infiltration.

Figure 7

The scatter plot showed the relationship between prognostic gene expression and medication sensitivity.