SERPINH1 overexpression in clear cell renal cell carcinoma: association with poor clinical outcome and its potential as a novel prognostic marker

Abstract

Precision therapy for clear cell renal cell carcinoma (ccRCC) requires molecular biomarkers ascertaining disease prognosis. In this study, we performed integrated proteomic and transcriptomic screening in all four tumour-node-metastasis stages of ccRCC and adjacent normal tissues (n = 18) to investigate differentially expressed genes. Most identified differentially expressed genes revealed a strong association with transforming growth factor-β level and the epithelial-to-mesenchymal transition process. Of them, Serpin peptidase inhibitor clade H member 1 (SERPINH1) revealed the strongest association with poor prognosis and regulation on the expression levels of epithelial-to-mesenchymal transition markers. Subsequently, two independent sets (n = 532 and 105) verified the high level of SERPINH1 in ccRCC tissues and its association with reduced overall survival and disease-free survival in all tumour-node-metastasis stages and patients with von Hippel-Lindau wild-type (VHL-WT). SERPINH1 was an independent predictor of poor overall survival (hazard ratio 0.696 for all patients) and disease-free survival (hazard ratio 0.433 for all patients and 0.362 for patients with VHL-WT) in ccRCC. We have thus shown for the first time that SERPINH1 is an independent precision predictor for unfavourable prognosis in ccRCC. This could assist in identifying patients who need early aggressive management and deepen our understanding of the pathogenesis of VHL-WT ccRCC.

Keywords: SERPINH1/HSP47; prognostic marker; proteomics; renal cancer; transcriptomics.

Figure 1

SERPINH1 is identified as potential prognostic marker candidate in ccRCC. Integrated analyses of proteomic and transcriptomic results between paired ccRCC and adjacent normal tissues at four TNM stages identified 35 common DEGs. Integrated univariate Cox analysis, prognosis analysis and GSEA were further used to screen for potential prognostic markers. A Venn diagram showed the overlap between genes that predicted meaningful OS risk ratio (UP, up‐regulated; DN, down‐regulated) and genes associated with OS poor prognosis (FC, fold change) from the TCGA_KIRC data set. The ccRCC GSEA result revealed that the expression of SERPINH1 got the highest ranking metric score among 35 DEGs enriched in poor OS prognosis group.

Figure 2

SERPINH1 level is abnormally up‐regulated in ccRCC tissues. (A) The mRNA level of SERPINH1 was up‐regulated in ccRCC tissues of the unpaired (Left) and paired (Right) TCGA_KIRC data sets. ***P < 0.001. (B, C) Abnormal up‐regulation of SERPINH1 level in ccRCC tissues was verified by WB and immunohistochemical analysis. **P < 0.01, ***P < 0.001.

Figure 3

SERPINH1 is correlated with poor clinical outcome of ccRCC patients. (A) Stacked bar graphs showing differential mRNA levels of SERPINH1 in patients with good and poor prognosis. Higher SERPINH1 levels were associated with poor prognosis of patients. Expression level was quantified in a four‐tier scale by the K‐means cluster method (from A with the lowest level to D with the highest level). The table shows the details (right). (B) Scatter plot displaying the level of SERPINH1 in patients with/without recurrence or metastasis. ***P < 0.001. (C) Kaplan–Meier (KM) curves indicated a shorter OS and DFS time with a high SERPINH1 mRNA level. (D) The KM curve indicated a shorter OS time with a high SERPINH1 protein level; P values were calculated with a log‐rank test.

Figure 4

SERPINH1 is a potential independent prognostic marker in ccRCC. SERPINH1 was an independent OS (A) and DFS (B) prognostic factor in ccRCC. U (univariate) and M (multivariate) Cox regression analyses. *P < 0.05, **P < 0.01, ***P < 0.001.

Figure 5

SERPINH1 predicts the DFS prognosis of VHL‐WT ccRCC. (A) The SERPINH1 mRNA level predicted the DFS prognosis of patients with VHL‐WT ccRCC better than that of patients with VHL‐MT. (B) The PDZK1 mRNA level also better predicted the DFS prognosis of patients with VHL‐WT. P values were calculated with a log‐rank test. (C) Receiver operator characteristic (ROC) curve results revealed that the SERPINH1 mRNA level had a stronger DFS prognosis‐predicting ability for VHL‐WT ccRCC than PDZK1. The area under curve (AUC) and the corresponding 95% CI are shown in the plots.