Development of a TGF-β signaling-related genes signature to predict clinical prognosis and immunotherapy responses in clear cell renal cell carcinoma

Abstract

Background: Transforming growth factor (TGF)-β signaling is strongly related to the development and progression of tumor. We aimed to construct a prognostic gene signature based on TGF-β signaling-related genes for predicting clinical prognosis and immunotherapy responses of patients with clear cell renal cell carcinoma (ccRCC).

Methods: The gene expression profiles and corresponding clinical information of ccRCC were collected from the TCGA and the ArrayExpress (E-MTAB-1980) databases. LASSO, univariate and multivariate Cox regression analyses were conducted to construct a prognostic signature in the TCGA cohort. The E-MTAB-1980 cohort were used for validation. Kaplan-Meier (K-M) survival and time-dependent receiver operating characteristic (ROC) were conducted to assess effectiveness and reliability of the signature. The differences in gene enrichments, immune cell infiltration, and expression of immune checkpoints in ccRCC patients showing different risks were investigated.

Results: We constructed a seven gene (PML, CDKN2B, COL1A2, CHRDL1, HPGD, CGN and TGFBR3) signature, which divided the ccRCC patients into high risk group and low risk group. The K-M analysis indicated that patients in the high risk group had a significantly shorter overall survival (OS) time than that in the low risk group in the TCGA (p < 0.001) and E-MTAB-1980 (p = 0.012). The AUC of the signature reached 0.77 at 1 year, 0.7 at 3 years, and 0.71 at 5 years in the TCGA, respectively, and reached 0.69 at 1 year, 0.72 at 3 years, and 0.75 at 5 years in the E-MTAB-1980, respectively. Further analyses confirmed the risk score as an independent prognostic factor for ccRCC (p < 0.001). The results of ssGSEA that immune cell infiltration degree and the scores of immune-related functions were significantly increased in the high risk group. The CIBERSORT analysis indicated that the abundance of immune cell were significantly different between two risk groups. Furthermore, The risk score was positively related to the expression of PD-1, CTLA4 and LAG3.These results indicated that patients in the high risk group benefit more from immunotherapy.

Conclusion: We constructed a novel TGF-β signaling-related genes signature that could serve as an promising independent factor for predicting clinical prognosis and immunotherapy responses in ccRCC patients.

Keywords: TGF-β signaling; biomarkers; clear cell renal cell carcinoma; immune infiltration; prognosis signature.

Figure 1

Flow chart of the analysis process in our study.

Figure 2

Identification of the prognostic TGF-β signaling-related genes in the TCGA cohort. (A) Venn diagram to identify DEGs between normal and tumor tissue. (B) The 29 overlapping genes were differently expressed in normal and tumor tissue. (C) Forest plots showing the significantly prognostic genes identified with univariate Cox regression analysis based on OS. (D) The PPI network downloaded from the STRING database indicated the interactions among candidate genes.

Figure 3

Construction of a prognostic model based on TGF-β signaling-related genes in the TCGA cohort. (A, B) LASSO Cox regression analysis was applied to screen the key genes. (C) The median value and distribution of the risk score. (D) The distribution of survival status. (E) Expression of seven prognostic genes. (F) K-M curves for the OS. (G) ROC curve of the prognostic signature.

Figure 4

Validation of the prognostic signature in the E-MTAB-1980 dataset. (A) Distribution of patients’ risk score, (B) Survival status, (C) Expression of seven prognostic genes, (D) K-M curves for the OS, and (E). ROC curve for evaluating the performance of the prognostic signature in the E-MTAB-1980 dataset.

Figure 5

Development of a nomogram predicting OS in ccRCC. (A, B) Univariate and multivariate cox regression for risk score and clinical features, including age, gender, stage, and risk score in the TCGA cohort. (C, D) Univariate and multivariate cox regression for risk score and clinical features, including age, gender, stage, and risk score in the E-MTAB-1980 cohort. (E) Nomogram integrated age, stage, and riskscore. (F, H) Calibration curve for predicting OS at 1, 3 and 5 years.

Figure 6

Relationship between riskscore and clinicopathological parameters in the TCGA cohort (A, B) and E-MTAB-1980 cohort (C, D). P values were shown as: ns, not significant; *P< 0.05; **P< 0.01; ***P< 0.001.

Figure 7

GO and KEGG analysis in the TCGA cohort. (A, B) GO enrichment analysis. (C, D) KEGG enrichment analysis.

Figure 8

Immune infiltration pattern analysis in the TCGA cohort. (A) Relationship heatmap of the riskscore and ssGSEA scores. (B) Box plots presenting the scores of immune cells. (C) Box plots presenting the scores of immune function. (D) CIBERSORT algorithm analysis on correlations between 22 immune cell types. (E) CIBERSORT algorithm analysis the distribution of the abundance of immune cell infiltration between the high and low risk score groups. P values were shown as: ns, not significant; *P< 0.05; **P< 0.01; ***P< 0.001.

Figure 9

The correlations between riskscore and expression of immune checkpoint molecules in the TCGA cohort. (A) Heatmap of immune checkpoint molecules expression, including PD-1, PD-L1, LAG3 and CTLA4. (B–E)The relevance between the risk score and the expression of immune checkpoints.