G Protein γ subunit 7 loss contributes to progression of clear cell renal cell carcinoma

Abstract Clear cell renal cell carcinoma (ccRCC) is a common urinary neoplasm, looking for useful candidates to establish scientific foundation for the therapy of ccRCC is urgent. We downloaded genomic profiles of GSE781, GSE6244, GSE53757, and GSE66271 from the Gene Expression Omnibus (GEO) database. GEO2R was used to analyze the derivative genes, while hub genes were screened by protein‐protein interactions and cytoscape. Further, overall survival, gene methylation, gene mutation, and gene expression were all analyzed using bioinformatics tools. Colony formation and cell‐cycle assay were used to detect the biological function of GNG7 in vitro. We found that GNG7 was downregulated in ccRCC tissues and negatively associated with overall survival in ccRCC patients. We also found that promoter methylation and frequent gene mutation were responsible for GNG7 gene suppression. GNG7 low expression was related to upregulation of enhancer of zeste homolog 2 and downregulation of disabled homolog 2‐interacting protein. Further, Gene Set Enrichment Analysis results showed that mTOR1, E2F, G2M, and MYC pathways were all significantly altered in response to GNG7 low expression. In vitro, A498 and 786‐O cells in which GNG7 expression was silenced, exhibited a lower G1 phase when compared to the negative control cells. Taken together, our findings suggest that GNG7 is a tumor suppressor gene in ccRCC progression and represents a novel candidate for ccRCC treatment.

With the development of bioinformatics, hub genes, relationship between genes, protein-protein interactions (PPIs), and functional pathways involved in the cancer carcinogenesis have provided valuable information about the pathogenesis of ccRCC. In 2003, Lenburg et al. (2003) compared seven cohorts of ccRCC and observed differential expression in 24.8-82.9% of genes. In 2007, Gumz et al. (2007) reported that secreted frizzled-related protein 1 is a tumor suppressor by analyzing genomic profiling of ccRCC tumors and patient-matched normal tissues. In 2014, neuronal pentraxin 2 was found to be overexpressed, specifically in ccRCC (von Roemeling et al., 2014). In 2016, Zofia et al. constructed a network of microRNA (miRNA) and messenger RNA (mRNA) involved in ccRCC progression (Wotschofsky et al., 2016). However, the core genes identified in different cohorts were different (Gumz et al., 2007;Lenburg et al., 2003;von Roemeling et al., 2014;Wotschofsky et al., 2016). Thus, further studies are needed to comprehensively and systematically identify possible prognostic, diagnostic, and therapeutic targets of differential gene expression in three or more cohorts. In our study, the differentially expressed genes (DEGs) of GSE781, GSE6244, GSE53757, and GSE66271 were obtained and the hub genes were screened on the basis of different clinical databases, PPI, and cytoscape. We found that G Protein γ subunit 7 (GNG7) plays a key role in ccRCC progression and is associated with the mTOR pathway.
Thus, GNG7 gene may be a potential target for detection of an early onset of ccRCC and may serve as target for the treatment of ccRCC.

| MATERIALS AND METHODS
2.1 | Microarray data GSE781, GSE6244, GSE53757, and GSE66271 were applied from the Gene Expression Omnibus (GEO), nine ccRCC patients-tumor tissue and matched normal tissue were obtained after surgery in GSE781 data set on a GLP96/97 platform (Lenburg et al., 2003); GSE6344 included 10 patient-matched normal renal cortex and ccRCC tissues on a GLP96/97 platform (Gumz et al., 2007); GSE53757 was performed by von Roemeling et al. (2014) and 72 ccRCC patienttumor tissue and matched normal tissue were collected. Gene expression was compared between tumor and matched normal samples on a GPL570 platform; GSE66271 was composed of 13 ccRCC patient-tumor tissue and matched normal tissue on a GPL570 platform (Wotschofsky et al., 2016). DEGs between tumor and matched normal samples were analyzed by GEO2R, an online analysis tool in the GEO website.

| Venn analysis
Log (fold-change) > 1 and adj. p ≤ 0.01 DEGs in four cohorts were considered statistically significant, and submitted to Bioinformatics & Systems Biology to obtain the statistically significant DEGs. Bioinformatics & Systems Biology is a web in the fields of gene prediction and genome annotation, comparative and evolutionary genomics, and systems biology.

| Hub gene analysis
PPI was constructed by the Search Tool for the Retrieval of Interacting Genes (STRING; http://string.embl.de/) to analyze the interaction of DEGs with the medium confidence at 0.45. Then, cytoscape (version 3.4.0), an open source software platform for data integration, analysis, and visualization, was used to analysis DEGs (Shannon et al., 2003).
The overall survival and gene expression in different stage/grade renal cell carcinoma were performed in UALCAN (http://ualcan.path.uab. edu). UALCAN is an interactive web resource for analyzing and generating graphs with The Cancer Genome Atlas (TCGA) transcriptome data using javascript and CSS (Chandrashekar et al., 2017).

| Gene Set Enrichment Analysis
The RNA-seq data of 611 ccRCC patients were download from the TCGA database (September 10, 2018) and GNG7 high-expression group (top 135 [25% of 611] high-YAP expression patients) and GNG7 low-expression group (top 134 [25% of 611] low-GNG7 expression patients) were set up. Then, mRNA expression data of the two groups were submitted to Gene Set Enrichment Analysis (GSEA) 2.0 software (Subramanian et al., 2005), and the hallmark gene sets (http:// software.broadinstitute.org/gsea/msigdb/collections.jsp#H) were selected for analysis. Hallmark gene sets summarize and represent specific well-defined biological states or processes and display coherent expression including 50 hallmarks, which condense information from over 4,000 original overlapping gene sets from v4.0 MSigDB collections C1 through C6 (Liberzon et al., 2015;Liberzon et al., 2011).

| Cancer Cell Line Encyclopedia analysis
Cancer Cell Line Encyclopedia (CCLE; https://portals.broadinstitute. org/ccle/home) database, is a publicly accessible online microarray database (Barretina et al., 2012). GNG7 was submitted to CCLE to analyze GNG7 mRNA expression in different cell lines and different cancer types.

| Oligo small interfering RNA transfection
Cells were cultured in six-well dishes for 24 hr, then transfected with oligo small interfering RNA (siRNA) against GNG7 (RiboBio, Guangzhou, China) using Roche X-tremeGENE siRNA transfection reagent (Roche Co. Ltd., Shanghai, China). The knockdown was verified by western blot analysis and real-time quantitative polymerase chain reaction.

| MTT assay
Cell proliferation in GNG7 knockdown was carried out as previously described . Cells (4,000 cells per well) were plated in a 96-well plate for 48 hr and cell viability was assessed using 3-(4,5- The growth rate was calculated as the average OD value in GNG7silenced cells group)/average OD value in the control group × 100%.

| Clone and clonogenic assay
Cells transfected with oligo siRNA were seeded in six-well plates (1,000 cells/well) and cultured for 5 days. Crystal violet (0.5% m/v) was then used to stain the cells and images were captured by a microscope.

| Cell-cycle analysis
Cell transfected with oligo siRNA A498 and 786-O were harvested, washed with phosphate-buffered saline, and fixed overnight in 70% ethanol at 20°C. Cells were incubated with propidium iodide at room temperature for 30 min as previously described (Xu et al., 2013), and were analyzed by flow cytometry using a FACSCalibur flow cytometer (Becton, Dickinson and Company, BD Biosciences, San Jose, CA). The data were analyzed using the cell fit software.

| Statistical analysis
The differences between two groups (Studentʼs t test) were analyzed by GraphPad Prism version 6.0 software (GraphPad), p < 0.05 was considered significant.

| The hub genes in clear cell renal cell carcinoma progression
To identify the hub genes involved in ccRCC progression, gene profiling was performed between tumor tissue and the matched tissue and the DEGs were extracted from each data set. We identified 740 genes in GSE781, 2,231 genes in GSE6344, 17,082 genes in GSE53757, and 5,223 genes in GSE66271 that were differentially expressed by Log (fold-change) > 1 with p ≤ 0.01. Furthermore, a fourway Venn diagram of GSE781, GSE6344, GSE53757, and GSE66271 data sets revealed that 382 explicit genes were commonly identified ( Figure 1a, Supporting Information 1).
Next, the interaction among the 382 explicit genes was constructed by the PPI network using STRING (Supporting Information 2) and the significant gene module was calculated by cytoscape ( Figure   1b). Results showed that GNG2, GNG7, KNG1, and GNB4 genes were identified as hub genes with maximal clique centrality (MCC) value > 1,000 and degree ≥ 10 (Table 1). Subsequently, overall survival analysis of hub genes was performed by UALCAN Figure 2a).

| GNG7 expression is repressed in the tumor
TCGA ccRCC patient gene array showed that GNG7 mRNA expression was significantly repressed in tumor tissue in comparison to the normal tissue ( Figure 3a). Further, we analyzed the expression level of GNG7 in normal and ccRCC tissue samples from grades 1, 2, 3, 4, and stages Ⅰ, Ⅱ, Ⅲ , Ⅳ. We found that GNG7 was significantly downregulated in ccRCC in all grades (p < 0.05; Figure 3b). However, no significant difference was found in the stages Ⅱ, Ⅲ , and Ⅳ (Figure 3c), and GNG7 was expressed in higher levels when compared to the normal tissues.
To investigate the mechanism of the low gene expression of analysis was performed using cBioPortal, which showed that nucleotide C was increased while nucleotide T was decreased in ccRCC tissues (Figure 3e). "hallmark_G2M_ckeckpoint," "hallmark_MYC_targets_v1," and "hall-mark_MTORC1_signaling" were responsible for the lower expression GNG7 in progress biological behavior of ccRCC ( Figure 5). These results reveal that repressed GNG7 was associated with increased cell proliferation and viability in ccRCC progression.

| GNG7 dampens proliferation of ccRCC cell lines
To investigate the effect of GNG7 on ccRCC cell proliferation and viability, we used a CCLE database, which demonstrated that the GNG7 mRNA expression level was low in many cancer types (Figure 6a). Meanwhile, the mRNA expression level of GNG7 in proximal tubular cell line (HK-2) was significantly higher than that in ccRCC cell lines ( Figure   6b,c). Next, we performed a series of in vitro experiments using a loss-offunction of GNG7 in ccRCC cell lines. We found that the cell growth rate

| DISCUSSION
In this study, we identified 25,276 genes that were significantly differentially expressed at Log (fold-change) > 1 and adj. p ≤ 0.01 in ccRCC tumor tissue relative to the matched tissue. A total of 382 DEGs were identified among the four cohorts (Figure 1a), cytoscape and PPI network were performed to explore the gene coexpression network among the DEGs. We selected four DEGs (GNG2, GNG7, KNG1, and GNB4) as hub genes with MCC value ≥ 1,000 combing degrees ≥10 (Figure 1b). Further validation in TCGA data set showed that low expression of GNG7 gene was related to a worse overall survival ( Figure 2b), but, GNG2, KNG1, and GNB4 genes had no significant effect on ccRCC patient overall survival (Figure 2a,c, and d).
These results suggest that GNG7 gene is a tumor suppressor in ccRCC.
GNG7 belongs to the large G protein γ family (Shibata, Mori, Tanaka, Kitano, & Akiyoshi, 1998). It was reported that GNG7 is a tumor suppressor gene in esophageal cancer, squamous cell carcinoma of head and neck, pancreatic cancer, and gastrointestinal cancer (Hartmann et al., 2012;Long, Liu, Wu, Xu, & Ge, 2016;Shibata et al., 1998). However, the role of GNG7 in cancer is poorly understood and the significance of GNG7 gene expression in ccRCC remains unknown.
Consistent with the previous studies (Long et al., 2016;Ohta et al., 2008;Shibata et al., 1998), we observed that a lower GNG7 expression was significantly associated with an overall poor survival and high grade/stage in ccRCC patients (Figure 2b, 3d, and 3e).
Regarding the clinical pathological parameters, we found that low expression of GNG7 was significantly associated with tumor grade.
These findings are consistent with the previous reports showing that loss of GNG7 was related to large tumor and tumor invasion and aggressiveness in squamous cell carcinoma of head and neck and esophageal cancer (S. Wu, F. Wu, & Jiang, 2017;Ohta et al., 2008), respectively.
GNG7 gene has a highly methylated promoter in squamous cell carcinoma of the head and neck and esophageal cancer (Hartmann et al., 2012;Ohta et al., 2008), however, the effect and underlying mechanism of GNG7 loss and the function on cancer biology in ccRCC remain unknown. In this study, we found that GNG7 gene promoter was highly methylated in tumors but unmethylated in normal tissues as shown by TCGA data set ( Figure 3b). Moreover, GNG7 gene mutation was found in almost all ccRCC patients in this study, further validating the TCGA data set (Figure 3a). This result indicates that GNG7 gene methylation and high CG sit may be responsible for GNG7 gene inactivation in ccRCC progression. The expression of GNG7 may be regulated by miR-328 and demethylation drug 5-AZAC restored GNG7 expression in other cancer cell lines (Ohta et al., 2008), but the molecular mechanism of such effects in ccRCC remain unclear and warrant further investigation.
Our GSEA analysis showed that GNG7 has a significant effect on mTOR1 signaling which is the downstream of VHL-HIFs signaling pathway in ccRCC. Further, the downregulated GNG7 gene mainly affected cell proliferation-related pathways among which the top three were "hallmark_E2F_targets," "hallmark_G2M_ckeckpoint," and "hallmark_MYC_targets_v1". As GNG7 affected mTOR1 signaling, we speculate that GNG7 is an upstream regulator of mTOR1, which may serve as a novel target for the developing diagnostic and therapeutic strategies in ccRCC. All these results predicted by bioinformatics analysis were validated in ccRCC cell lines, where we found that GNG7-silenced cells grew faster and the G2M phase was increased when compared to the negative control cells, consistent with GNG7 function in other tumors .
Meanwhile, DAP2IP, a tumor suppressor, plays an important role in drug resistant by regulating mTOR (Zhou et al., 2016). These above results indicate that GNG7 may play an important role in RCC progression, metastasis, immune control, and in drug resistance.
Conclusion, our study identified the key genes in ccRCC progression and GNG7 gene was screened from four ccRCC cohorts.
GNG7 gene was strongly suppressed in ccRCC tumor tissues as a result of promoter methylation and frequent gene mutation. GNG7 expression was negatively correlated with ccRCC patient grade and overall survival. Decreased expression of GNG7 was related to mTOR1, E2F, G2M, and MYC pathway. In addition, lower GNG7 expressing ccRCC cells showed an increase in G2/M cell-cycle phase.
These findings suggest that GNG7 is a tumor suppressor in ccRCC progression and has a potential to be a new biomarker or therapeutic target in ccRCC. 2017JM8018 to Shan Xu).