RNA-Binding Proteins Play an Important Role in the Prognosis of Patients With Testicular Germ Cell Tumor

Abstract

Testicular germ cell tumors (TGCTs) are common urological neoplasms in young adult males. The outcome of TGCT depends on pathologic type and tumor stage. RNA-binding proteins (RBPs) influence numerous cancers via post-transcriptional regulation. The prognostic importance of RBPs in TGCT has not been fully investigated. In this study, we set up a prognostic risk model of TGCT using six significantly differentially expressed RBPs, namely, TRMT61A, POLR2J, DIS3L2, IFIH1, IGHMBP2, and NPM2. The expression profiles were downloaded from The Cancer Genome Atlas (TCGA) and Genotype-Tissue Expression datasets. We observed by performing least absolute shrinkage and selection operator (LASSO) regression analyses that in the training cohort, the expression of six RBPs was correlated with disease-free survival in patients with TGCT. We assessed the specificity and sensitivity of 1-, 3-, 5-, and 10-year survival status prediction using receiver operating characteristic curve analysis and successfully validated using the test cohorts, the entire TCGA cohort, and Gene Expression Omnibus (GEO) datasets. Gene Ontology, Kyoto Encyclopedia of Genes and Genomes, and gene set enrichment analyses were carried out to seek the possible signaling pathways related with risk score. We also examined the association between the model based on six RBPs and different clinical characteristics. A nomogram was established for TGCT recurrence prediction. Consensus clustering analysis was carried out to identify the clusters of TGCT with different clinical outcomes. Ultimately, external validations of the six-gene risk score were performed by using the GSE3218 and GSE10783 datasets downloaded from the GEO database. In general, our study constructed a prognostic model based on six RBPs, which could serve as independent risk factor in TGCT, especially in seminoma, and might have brilliant clinical application value.

Keywords: RNA-binding proteins; nomogram; prognosis; risk score; testicular germ cell tumors.

FIGURE 1

The flow chart of the study design and analysis.

FIGURE 2

Heatmap and volcano plots of TGCT patients from TCGA. Heatmap (A) and volcano (B) plots were generated with FDR < 0.05 and | log2FC| > 0.5, using the data of differentially expressed RBPs in TGCT downloaded from TCGA and GTEx. TGCT, Testicular germ cell tumors; TCGA, The Cancer Genome Atlas; RBPs, RNA-binding proteins; GTEx, Genotype-Tissue Expression; N, Normal controls from healthy individuals; T, tumor samples of testicular germ cell tumors.

FIGURE 3

The most critical modules in the PPI network of differentially expressed RBPs (A–C). Significant modules were obtained from the PPI network using MCODE of Cytoscape. Modules with red color and green color relatively represent high expression and low expression levels of RBPs. The criteria for selection were set as follows: Max depth = 100, degree cut-off = 2, Node score cut-off = 0.2, MCODE scores > 5, and K-score = 2. PPI, protein–protein interactions; RBPs, RNA-binding proteins; MCODE, Molecular Complex Detection.

FIGURE 4

Prognostic value of RBPs in TCGA TGCT training cohort. (A) Cox univariate analysis of RBP genes in the training cohort. (B,C) Multivariate Cox regression via LASSO is presented, and ten candidate RBPs were selected in training cohort. (D) Forrest plot of the multivariate Cox regression analysis in TGCT. TGCT, Testicular germ cell tumors; TCGA, The Cancer Genome Atlas; RBPs, RNA-binding proteins; LASSO, last absolute shrinkage and selection operator.

FIGURE 5

Survival analysis based on the risk model in the training group. (A) Kaplan–Meier survival curve analysis of DFS in the high-risk and low-risk TGCT patients in the training group, marked as red line and blue line separately. (B–D) Risk score, distribution of survival status between the high-risk and low-risk groups, and expression heat maps of six RBPs. (E–H) Time-dependent ROC curve analyses was conducted and AUC values were calculated for 1-, 3-, 5-, and 10-year DFS in the TGCT training cohort. TGCT, Testicular germ cell tumors; RBPs, RNA-binding proteins; DFS, disease-free survival; ROC, Receiver operating characteristic curve; AUC, the area under the ROC curve; High, high-risk group; Low, low-risk group.

FIGURE 6

Validation of the prognostic value of the risk model in the test group. (A) Kaplan–Meier survival curve analysis of DFS in the high-risk and low-risk TGCT patients in the test group, marked as red line and blue line separately. (B–D) Risk score, distribution of survival status between the high-risk and low-risk groups and expression heat maps of six RBPs. (E–H) Time-dependent ROC curve analyses was conducted and AUC values were calculated for 1-, 3-, 5-, and 10-year DFS in the TGCT test cohort. TGCT, Testicular germ cell tumors; DFS, disease-free survival; RBPs, RNA-binding proteins; TGCT, Testicular germ cell tumors; ROC, Receiver operating characteristic curve; AUC, the area under the ROC curve; High, high-risk group; Low, low-risk group.

FIGURE 7

Validation of the prognostic value of the risk model in the entire TCGA cohort. (A) Kaplan–Meier survival curve analysis of DFS in the high-risk and low-risk TGCT patients in the entire TCGA cohort, marked as red line and blue line separately. (B–D) Risk score, distribution of survival status between the high-risk and low-risk groups, and expression heat maps of six RBPs. (E–H) Time-dependent ROC curve analyses was conducted and AUC values were calculated for 1-, 3-, 5-, and 10-year DFS in the entire TCGA TGCT cohort. TGCT, Testicular germ cell tumors; TCGA, The Cancer Genome Atlas; DFS, disease-free survival; ROC, Receiver operating characteristic curve; AUC, the area under the ROC curve; TGCT, Testicular germ cell tumors.

FIGURE 8

Analysis for evaluating the independent prognostic value of the risk score. Forrest plot of univariate (A) and multivariate (B) Cox regression analysis of risk score, age, serum markers, lymphovascular invasion, TNM stage, disease type, and stage. (C) Significant differences were found for the disease type, M stage, serum markers, and stage between high- and low-risk group. *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the low-risk group.

FIGURE 9

The time-dependent ROC curves for risk score, age, serum markers, lymphovascular invasion, TNM stage, disease type, and stage combining with 1- (A), 3- (B), 5- (C), and 10- (D) year DFS in TCGA TGCT cohort, respectively. DFS, disease-free survival; ROC, Receiver operating characteristic curve; TCGA, The Cancer Genome Atlas; TCGA, The Cancer Genome Atlas; DFS, disease-free survival; ROC, Receiver operating characteristic curve; TGCT, Testicular germ cell tumors.

FIGURE 10

Prognostic value of the risk score in TGCT patients classified into specific cohorts. Kaplan–Meier survival curve of DFS for patients with (A) age >36, (B) age ≤36, (C) no lymphovascular invasion, (D) lymphovascular invasion, (E) serum marker study levels within normal limits, (F) serum marker study levels beyond the normal limits, (G) non-seminoma, (H) seminoma, (I) T1 stage, (J) T2-3 stage, (K) M0 stage, and (L) M1 stage in the high-risk (red line) and low-risk (blue line) TGCT patients. TGCT, Testicular germ cell tumors.

FIGURE 11

Kaplan–Meier survival analysis and the expression profiles of the six RBPs. (A–F) Kaplan–Meier plots showed distributions in survival probabilities of six RBPs in high-risk (yellow line) and low-risk (blue line) TGCT patients. TGCT, Testicular germ cell tumors; RBPs, RNA-binding proteins.

FIGURE 12

Nomograms predicting survival probability of TGCT patients in TCGA. (A) Nomogram to predict 1-, 3-, and 5-year DFS. (B–D) Calibration plots of 1-, 3-, and 5-year DFS for nomograms. TCGA, The Cancer Genome Atlas; DFS, disease-free survival; TGCT, Testicular germ cell tumors.

FIGURE 13

Functional annotation of the differentially expressed genes between high risk and low risk groups based on six-RBP risk score of TGCT in TCGA cohort. Bubble plots showed enrichment of (A) GO terms associated with up-regulated genes and (B) GO terms associated with down-regulated genes. Bubble plots showed enrichment of (C) KEGG pathway associated with up-regulated genes and (D) KEGG pathway associated with down-regulated genes. (E) GSEA analysis showed the top ten most significantly enriched signaling pathways in high-risk score subgroup. GO, Gene Ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes.

FIGURE 14

Differential clinical characteristics of TGCT patients in the two different clusters. (A) Based on the expression similarity of RBPs, the TCGA TGCT cohort was separated into two distinct clusters when k = 2. (B,C) Consensus CDF and relative change in area under CDF curve for k = 2–9. (D) Principal component analysis based on two clusters. (E) Significant difference was observed for the type, grade, stage, and serum marker study levels between cluster 1 and cluster 2. CDF, clustering cumulative distribution function. *p < 0.05, **p < 0.01, and ***p < 0.001 compared with the cluster2 group. TGCT, Testicular germ cell tumors; RBPs, RNA-binding proteins; TCGA, The Cancer Genome Atlas.

FIGURE 15

Validation of the prognostic value of the risk model in GSE3218 and GSE10783 datasets. (A) Kaplan–Meier survival curve analysis of OS in high-risk score and low-risk score TGCT patients in GSE3218 and GSE10783 datasets, marked as red line and blue line separately. (B,C) Risk score, distribution of survival status between the high-risk and low-risk groups, and expression heat maps of six RBPs. OS, overall survival; TGCT, Testicular germ cell tumors; TCGA, The Cancer Genome Atlas; RBPs, RNA-binding proteins.