ccRCC is considered to be an immunogenic and heterogeneous tumor caused by various carcinogenic factors [29]. Although, ccRCC can be successfully treated with surgical strategies, some of patients will develop metastasis, which is the main cause of death in ccRCC patients [30]. To reduce the mortality of ccRCC patients and improve their prognosis, there is urgently required to screen significantly prognosis-related biomarkers for patients with ccRCC. For the past few years, ZNFs have been found to be closely correlated with various biological processes, and they are significantly involve in the ontogeny, progression and prognosis of many tumors, indicating that ZNFs have the potential to be used as prognosis-related biomarkers. The objective of this study is to construct a group of significantly Pr-ZRGs of ccRCC and to stratify risk of ccRCC patients by ZRGs signature, so as to achieve individual prognosis analysis and then guide early intervention and treatment of ccRCC patients.
We analysed the DE-ZRGs between tumor tissues and adjacent normal renal tissues through TCGA database and 271 DE-ZRGs were finally screened out. Then, through a series of regression analysis in the training set, we constructed a five-ZRG signature, including TRIM59, VAV3, ZNF189, AGAP9 and PYGO1, which is validated an independent prognostic factor of patients with ccRCC. According to the risk score formula based on five-ZRG signature, patients were stratified into two groups (high-risk and low-risk). KM survival analysis showed that the high-risk patients had a significantly lower OS time than that in the low-risk patients. ROC curve showed that the combination of five ZRGs signature had good prognostic accuracy. We further validated these findings in the testing set and entire set. Furthermore, to establish a more reliable and individualized clinical prediction method, we incorporated the five-ZRG signature with other prognosis-related clinical factors, including age, stage and grade, to construct a nomogram. The predicted outcome of the nomogram in 1-, 3-, and 5-year OS showed a good agreement in actual outcome in the calibration plot, and KM survival analysis and ROC curve indicated that the nomogram also had a better prognostic performance.
Herein, we identified five closely Pr-ZRGs of ccRCC patients. Interestingly, previous studies have reported that they were involved in the initiation, progression and prognosis of numerous tumors. Rongkun Li et al. found that TRIM59 could enhance glycolysis by activating the PI3K/AKT/mTOR pathway, which eventually led to the progression of pancreatic cancer [31]. Studies showed that up-expression of TRIM59 was indicated the poor prognosis of NSCLC and breast cancer [32, 33]. As a new carcinogenic EGFR/STAT3 signal regulator, TRIM59 promoted the occurrence of glioblastoma by inhibiting STAT3 dephosphorylation, which related to poor prognosis of glioblastoma [34]. In addition, TRIM59 might promote prostate cancer cell proliferation and tumor growth through its effect on cell cycle progression [35]. VAV3, a member of the VAV family of oncoproteins, is a carcinogenic gene [36]. Studies showed that overexpression of VAV3 led to abnormal expression of cell cycle control and metastasis-related molecules in colorectal cancer cells by activating PI3K/AKT signal pathway, which might be used as an independent prognosis-related biomarker of colorectal cancer patients [37]. VAV3 is also overexpressed in gastric cancer, and its overexpression is related to poor prognosis and short survival time of patients [38]. Additionally, VAV3 is correlated with the oncogenesis and development of prostate carcinoma by regulating the activity of androgen receptor [39, 40]. PYGO1 is a key regulatory gene of Wnt signaling pathway, which is correlated with the development of kidney and the transcription of colorectal carcinoma cells [41, 42]. Herein, we conducted GO and KEGG analysis of 271 DE-ZRGs. The result of GO analysis showed that DE-ZRGs were mainly associated with the regulation of GTPase activity and the ubiquitin-related proteases activity, suggesting that these ZRGs were involved in the ontogeny and progression of tumors such as migration, invasion and tumor angiogenesis [43, 44]. KEGG enrichment analysis showed ccRCC ontogency and progression may related to the signal pathways of endocytosis, phosphatidylinositol signaling system, choline metabolism in cancer and glycerolipid metabolism. Furthermore, to explore the potential biological effects of five screened ZRGs, we conducted GSEA analysis and found that they were related to intestinal immune network, natural killer cell-mediated cytotoxicity and cytokine receptor interaction, suggesting that the five Pr-ZRGs signature may be strongly correlated with the immune response process in the microenvironment of ccRCC.
In the two groups of ccRCC patients based on five-ZRG signature model classification, we found that immune infiltrating cells in the tumors of high-risk group were mainly T cells, tumor-infiltrating lymphocytes and macrophages, suggesting that chronic inflammation may promote tumor progression and reduce the survival time of patients. Further analysis of immune function in the two groups showed that CCR, check-point, cytolytic activity, para-inflammation and T cell co-stimulation were significantly increased in high-risk ccRCC patients. In the TME score, they also had significantly higher stromal score, immune score, and estimate score and lower tumor purity than that in low-risk patients. These results indicate that immune microenvironment of ccRCC plays a crucial role in the development of tumors. Immune cells and stromal cells in TME may promote the proliferation and migration of tumor cells, which resulted in the poor prognosis of ccRCC patients. Finally, we found that the different groups of patients have different sensitivities to many antineoplastic drugs. For the 28 antineoplastic drugs identified, high-risk patients had a higher IC50, suggesting that tumor resistance in the high-risk patients was stronger. However, some limitations have existed in this research. First, this study is only a mining analysis of the database based on bioinformatics, multicenter and prospective studies are needed to further confirmed these findings accurately. Second, it is quite necessary to further explore the potential molecular mechanism of the proteins encoded by the five ZRGs in ccRCC.