A Novel 7-Methylguanosine (m7G)-Related Gene Signature for Overall Survival Prediction in Patient with Clear Cell Renal Cell Carcinoma

Abstract

Clear cell renal cell carcinoma (ccRCC) is the most common pathology type of renal cancer that has an abysmal prognosis. Although a crucial role for 7-methylguanosine modification in cancer cell development has been reported, its role in ccRCC remains uncertain. This study was conducted to determine the efficacy of predictive biomarkers based on m7G-related genes in ccRCC. Firstly, we extracted clinical data and gene expression profiles of ccRCC patients from publicly accessible databases. It identified that 22 of the m7G-related 34 genes were related to overall survival, and 5 of the 22 genes were significantly expressed differently in tumor tissues. Based on Lasso regression analysis, five optimal genes (CYFIP2, EIF4A1, NUDT1, NUDT10, and NUDT4) were chosen to build a new predictive risk model in the TCGA cohort. Validation was carried out with the E-MTAB-1980 cohort. Then, a prognostic nomogram was erected, including the m7G-related gene risk score, age, histological grade, and stage status. Further studies and analysis showed that immune cell infiltration might be associated with the m7G-related risk genes. In addition, the relationship between gene expression and drug response was evaluated by the Pearson correlation test. Therefore, the risk signature with five selected m7G-related genes may be a promising prognostic biomarker and contribute to standardized prognostic assessment for ccRCC.

Figure 1

The overall workflow of this study.

Figure 2

Intersection of differentially expressed m7G-related genes and survival-associated m7G-related genes. (a) Venn diagram showing the five m7G-related genes associated with survival between differentially expressed genes and prognostic genes. The green circle represents differentially expressed genes, and the pink circle represents prognostic genes. (b) Heatmap showing the five m7G-related associated with survival. The red rectangle represents tumor tissue, and the blue rectangle represents normal tissue. (c) The forest figure of the 5 key m7G-related genes in univariate Cox regression. The blue line represents a 95% confidence interval. The position of the square represents the hazard ratio. (d) The correlation network of 5 m7G-related genes, in which different colors represent the correlation coefficients. The red represents a positive relationship, and the blue represents a negative relationship.

Figure 3

Identification of five optimal MRGs. (a) Partial likelihood deviance was plotted against log (λ). The vertical dotted lines indicate the λ value with minimum error. The largest λ value is where the deviation is within one standard error (SE) of the minimum. (b) Least absolute shrinkage and selection operator (Lasso) coefficient profiles of MRGs in ccRCC.

Figure 4

Survival analyses of the 5-gene model in the TCGA cohort. (a) The distribution of the risk scores in the TCGA cohort. (b) Principal component analysis plots display the established gene signature expression distribution in different risk groups. (c) t-Distributed stochastic neighbor embedding plots reveal the patients' distribution in different risk groups. (d) The distributions of the risk scores and corresponding survival times of all patients in the TCGA cohort. (e) OS-based K-M survival curves for the patients in the high and low-risk groups in the TCGA cohort. (f) AUC of time-dependent ROC curves verified the prognostic performance of the risk score in the TCGA cohort. (g) AUC of clinical ROC curves verified the prognostic performance of the risk score in the TCGA cohort.

Figure 5

Validation of the 5-gene model in the E-MTAB-1980 cohort. (a) The distribution of the risk scores in the E-MTAB-1980 cohort. (b) Principal component analysis plots display the established gene signature expression distribution in different risk groups. (c) t-Distributed stochastic neighbor embedding plots reveal the patients' distribution in different risk groups. (d) The distributions of the risk scores and corresponding survival times of all patients in the E-MTAB-1980 cohort. (e) OS-based K-M survival curves for the patients in the high- and low-risk groups in the E-MTAB-1980 Cohort. (f) AUC of time-dependent ROC curves verified the prognostic performance of the risk score in the E-MTAB-1980 cohort. (g) AUC of clinical ROC curves verified the prognostic performance of the risk score in the E-MTAB-1980 cohort.

Figure 6

Construction of a new prognostic nomogram. (a) Heatmap and the clinicopathologic characters of the low- and high-risk groups. ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001. (b) Univariate Cox regression analysis. (c)Multivariate Cox regression analysis. (d) Prognostic nomogram for predicting the survival of patients with ccRCC. The blue line represents a 95% confidence interval. The position of the square represents the hazard ratio. (e-g) Calibration curves of the nomogram for predicting survival at 1, 3, and 5 years. The nomogram prediction accuracy is higher if the actual curve is closer to the ideal curve. The red line represents the actual curve and the black line represents the ideal line.

Figure 7

Results of GO and KEGG analysis of the TCGA cohort. (a) GO enrichment analysis revealed the biological processes and molecular functions involved in the differentially expressed genes. The GO enrichment analysis is divided into three parts, including cellular component (CC), molecule function (MF), and biological processes (BP). (b) KEGG analysis shows the signaling pathways involved in the differentially expressed genes.

Figure 8

Results of ssGSEA immune infiltration in the TCGA cohort. The association between risk score and (a) immune score or (b) stromal score. (c) 16 immune cell ssGSEA scores between different risk groups. (d) 13 immune-related functional ssGSEA scores between different risk groups. Adjusted P values are: ns, not significant; ^∗P < 0.05, ^∗∗P < 0.01, and ^∗∗∗P < 0.001.

Figure 9

Correlation of m7G-related genes expressions with drug response. The relationship between drug sensitivity and CYFIP2, EIF4A1, NUDT1, NUDT1, and NUDT10 expression.