Construction of a prognostic risk model for clear cell renal cell carcinomas based on centrosome amplification-related genes

Abstract

Clear cell renal cell carcinoma (ccRCC) is the urological malignancy with the highest incidence, centrosome amplification-associated genes (CARGs) have been suggested to be associated with carcinogenesis, but their roles in ccRCC are still incompletely understood. This study utilizes bioinformatics to explore the role of CARGs in the pathogenesis of ccRCC and to establish a prognostic model for ccRCC related to CARGs. Based on publicly available ccRCC datasets, 2312 differentially expressed genes (DEGs) were identified (control vs. ccRCC). Disease samples were classified into high and low scoring groups based on CARG scores and analysed for differences to obtain 345 DEGs associated with CARG scores (S-DEGs). 137 candidate genes were obtained by taking the intersection of DEGs and S-DEGs. Six prognostic genes (PCP4, SLN, PI3, PROX1, VAT1L, and KLK2) were then screened by univariate Cox, LASSO, and multifactorial Cox regression. These genes exhibit a high degree of enrichment in ribosome-associated pathways. Both risk score and age were independent prognostic factors, and the Nomogram constructed based on them had a good predictive performance (AUC > 0.7). In addition, immunological analyses identified 6 different immune cells and 23 immune checkpoints between the high- and low-risk groups, whereas mutational analyses identified frequent VHL mutations in both high- and low-risk groups. Finally, 93 potentially sensitive drugs were identified. In conclusion, this study identified six CARGs as prognostic genes for ccRCC and established a risk model with predictive value. These findings provide insights for prognostic prediction of ccRCC, optimisation of clinical management and development of targeted therapeutic strategies.

Keywords: Centrosome amplification-related genes; Clear cell renal cell carcinomas; Prognostic genes; Risk model.

Fig. 1

A total of 137 candidate genes were characterized. a Volcano plot of differential genes between ccRCC and control groups. b Heat map of differential gene expression between ccRCC and control groups. c Box plots of differences in CARGs scores between ccRCC and control groups. d CARGs scores in high and low scoring groups survival differences. e Differential gene volcano plots between high and low scoring groups. f Obtain a Venn diagram of the candidate gene

Fig. 2

Functional analyses of candidate genes. a GO enrichment analysis of candidate genes. Yellow, blue, and green represent biological processes, cellular components, and biological functions, respectively. b Candidate gene KEGG enrichment analysis. Redder colours indicate greater significance

Fig. 3

A total of 6 prognostic genes were obtained. a Univariate Cox regression analysis based on candidate genes and forest plotting. b Plot of gene coefficients for LASSO analysis, c Cross-validation error plot for LASSO analysis. d Screening of prognostic genes by multivariate Cox regression and mapping of forests

Fig. 4

Validation of the risk model. a-b Risk profiles of the training set TCGA-KIRC and validation set GSE29609 samples.c-d Survival state distribution of the training set TCGA-KIRC and validation set GSE29609 samples. e-f KM curves for high and low risk groups in the training set TCGA-KIRC and validation set GSE29609. g-h ROC curves of the model in the training set TCGA-KIRC and the validation set GSE29609. i-j Heatmap of prognostic gene expression in high and low risk groups (training set TCGA-KIRC and validation set GSE29609)

Fig. 5

Screening for independent prognostic factors. a Risk score analysis of K-M survival curves for different clinical characteristics. b Forest plot of Univariate Cox regression analysis of clinical data information and risk scores for all ccRCC patients in the training set TCGA-KIRC. c Multivariate Cox regression screening for independent prognostic factors

Fig. 6

Nomogram constructed based on independent prognostic factors. a Nomogram constructed based on independent prognostic factors. b The 1-year, 3-year, and 5-year calibration curves for the nomogram. c The 1-year, 3-year, and 5-year ROC curves for the nomogram.d The DCA for the nomogram

Fig. 7

GSEA based on prognostic genes a PCP4, b SLN, c KLK2, d PL3, e VAT1L, f PROX1

Fig. 8

Immune infiltration and differential immune checkpoint analysis. a The xCell algorithm assesses the distribution of immune cells in high and low risk groups. b Differences in the distribution of immune cell infiltration between high and low risk groups. c Correlation analysis of differential immune cells and prognostic genes between high and low risk groups. d Boxplot of difference analysis of common immune checkpoints between groups for high and low risk groups. e Correlation of prognostic genes with differential immune checkpoints. ns: not significant, *: p < 0.05, **: p < 0.01, ***: p < 0.001, ****: p < 0.0001

Fig. 9

Mutation analysis based on high and low risk groups. a Waterfall plot of tumour somatic cell mutations in the high risk group. b Waterfall plot of tumour somatic mutations in the low risk group. c Survival differences in TMB scores between high and low risk groups

Fig. 10

Drug sensitivity analysis a-b Differences in IC50 of drugs between high and low risk groups. ***: p < 0.001, ****: p < 0.0001

Fig. 11

TF regulatory network of prognostic genes. Prognostic markers in red and TF in blue