m6A RNA methylation regulators play an important role in the prognosis of patients with testicular germ cell tumor

Abstract

Background: N6-methyladenosine (m6A) is found to be associated with promoting tumorigenesis in different types of cancers, however, the function of m6A-related genes in testicular germ cell tumors (TGCT) development remains to be illuminated. This study aimed to investigated the prognostic value of m6A RNA methylation regulators in TGCT.

Methods: We collected TGCT patients' information about clinicopathologic parameters and twenty-two m6A regulatory genes expression from The Cancer Genome Atlas (TCGA) database and Genotype-Tissue Expression (GTEx). We analyzed the differentially expressed m6A RNA methylation regulators between tumor tissues and normal tissues, as well as the correlation of m6A RNA methylation regulators. By using Cox univariate analysis, last absolute shrinkage and selection operator (LASSO) Cox regression algorithm and Cox multivariate proportional hazards regression analysis, a risk score was constructed based on a TCGA training cohort, and further verified in the TCGA testing cohort. Then, univariate and multivariate Cox regression analyses were used to evaluate the relationship between risk score and progression-free survival (PFS) in TGCT. Finally, the six-gene risk score was further verified by two gene expression profiles (GSE3218 and GSE10783) as an independent external validation cohort.

Results: Distinct expression patterns of m6A regulatory genes were identified between TGCT tissues and normal tissues in TCGA and GTEx datasets. To predict prognosis of TGCT patients, a risk score was calculated based on six selected m6A RNA methylation regulators (YTHDF1, RBM15, IGF2BP1, ZC3H13, METTL3, and FMR1). Additionally, we found significant differences between the high-risk and low-risk groups in serum marker study levels and histologic subtype. Univariate and multivariate analysis indicated that high risk score was associated with unfavorable PFS. Ultimately, the risk score was further verified by two gene expression profiles (GSE3218 and GSE10783).

Conclusions: Based on six selected m6A RNA methylation regulators, we developed a m6A methylation related risk score that can independently predict the prognosis of TGCT patients, and verify the prediction efficiency in TCGA and GEO datasets. Patients in high-risk group were associated with serum tumor marker study levels beyond the normal limits, non-seminoma, and unfavorable survival time. However, further prospective experiments should be carried out to verify our results.

Keywords: N6-methyladenosine (m6A); Testicular germ cell tumors (TGCT); m6A RNA methylation; prognosis; risk score.

Figure 1

The differential expression of m6A RNA methylation regulators between TGCT samples and normal tissues. (A) Hierarchical clustering of TGCT and normal tissues expressing the 22 m6A regulatory genes in TCGA and GTEx databases. (B) The expression of 22 m6A regulatory genes between TGCT and normal tissues. The red represents tumor group and blue represents normal tissue group. TCGA, The Cancer Genome Atlas; GTEx, Genotype-Tissue Expression. *, P<0.05 ; ***, P<0.001.

Figure 2

Correlation matrix of 22 m6A regulatory genes in the TCGA and GTEx databases. An X represents P>0.001, which means there was no statistically significant correlation between two m6A regulatory genes. TCGA, The Cancer Genome Atlas; GTEx, Genotype-Tissue Expression.

Figure 3

Overall survival of TGCT patients in the two different clusters. (A) The TCGA TGCT cohort was divided into two distinct clusters when k=2. (B) Consensus clustering cumulative distribution function (CDF) for k=2 to 10. (C) Relative change in area under CDF curve for k=2 to 10. (D) Principal component analysis of the total RNA expression profile. TGCT in cluster 1 and 2 are marked in red and blue, respectively. (E) Kaplan-Meier PFS curve for TGCT patients in cluster 1 and 2. TCGA, The Cancer Genome Atlas; PFS, progression-free survival; TGCT, testicular germ cell tumors.

Figure 4

The distribution of clinicopathological variables between different clusters. Significant difference was found for the serum marker study levels and histologic subtype between cluster 1 and cluster 2. ***, P<0.001.

Figure 5

Multivariate Cox regression via LASSO is presented, and ten candidate m6A RNA methylation regulators were selected. (A) Cross-validation for tuning parameter screening in the LASSO regression model. (B) LASSO coefficient profiles of the common genes. LASSO, last absolute shrinkage and selection operator.

Figure 6

Validation of the prognostic signature in the TCGA TGCT training cohort (A) Kaplan-Meier plot represents that patients in the high-risk group had significantly shorter PFS than those in the low-risk group. (B) Time-dependent ROC curve analysis for survival prediction by the risk score in the training test based on the TCGA dataset. (C) The risk score distribution of patients in the training cohort. (D) The distributions of risk scores and PFS status. The red and green dots indicated the progress and progress free respectively. (E) The heatmap of the six key genes expression profiles in the training cohort. TCGA, The Cancer Genome Atlas; TGCT, testicular germ cell tumors; ROC, receiver operating characteristic; PFS, progression-free survival.

Figure 7

Validation of the prognostic signature in an independent TGCT cohort. (A) Kaplan-Meier survival curve showing PFS outcomes according to relative high-risk and low-risk patients. (B) Time-dependent ROC curve analysis for survival prediction by the risk score in the testing cohort based on the TCGA dataset. (C) The risk score distribution of patients in the testing cohort. (D) The distributions of risk scores and PFS status. The red and green dots indicated the progress and progress free respectively. (E) The heatmap of the six key genes expression profiles in the testing cohort. TGCT, testicular germ cell tumors; TCGA, The Cancer Genome Atlas; PFS, progression-free survival; ROC, receiver operating characteristic.

Figure 8

Effects of the risk score and clinicopathological variables on the prognosis of TGCT patients (A) The heat map shows the expression of six m6A RNA methylation regulators and the distribution of clinicopathological variables between the high- and low-risk groups. (B) Cox univariate analyses of clinicopathological variables (including the risk score) and overall survival. (C) Cox multivariate analyses of clinicopathological variables (including the risk score) and PFS. TGCT, testicular germ cell tumors; PFS, progression-free survival. *, P<0.05.

Figure 9

Prognostic value of the risk score in TGCT patients classified into specific cohorts. Kaplan-Meier survival curve of PFS for patients with (A) seminoma, (B) non-seminoma, (C) serum marker study levels within normal limits, (D) serum marker study levels beyond the normal limits, (E) stage I, (F) stage II-III, (G) no lymphovascular invasion, (H) lymphovascular invasion. TGCT, testicular germ cell tumors; PFS, progression-free survival.

Figure 10

Enrichment plots from gene set enrichment analysis (GSEA). (A) The ten most significantly enriched signaling pathways in the high-risk score subgroup. (B) The ten most significantly enriched signaling pathways in the low risk score subgroup. ES, enrichment score; NES, normalized ES; NOM p-val, normalized P value.

Figure 11

The prognostic value of six-gene risk score of TGCT patients based on the GSE3218 and GSE10783 datasets. (A) Kaplan-Meier survival curve showing OS outcomes according to relative high-risk and low-risk patients. (B) ROC analysis of 3/5/10-year OS for risk score showed the predictive efficiency of the risk score. Kaplan-Meier curves of OS in different expression levels of (C) FMR1, (D) RBM15, (E) ZC3H13, (F) METTL3, (G) YTHDF1, and (H) IGF2BP1. TGCT, testicular germ cell tumors; OS, overall survival; ROC, receiver operating characteristic.