ZNF692 promotes the migration and response to immunotherapy of clear cell renal cell carcinoma cells by targeting metabolic pathway

Fifty-eight paraffin-embedded cancer and paracancerous samples were collected between January 2019 and December 2022 at The MengChao Hepatobiliary Hospital of Fujian Medical University. The samples were collected based on specific inclusion and exclusion criteria. The inclusion criteria consisted of primary clear cell renal cell carcinoma (ccRCC) cases that had not undergone any treatment prior to surgery. Conversely, the exclusion criteria encompassed other types of kidney carcinoma, non-primary ccRCC cases, and individuals who had received any form of interventional therapy, such as radiotherapy or chemotherapy. All patient is still alive at the time of writing the manuscript. All samples were obtained from patients with primary ccRCC aged 32 to 84 years. The TNM stage of the patients ranged from 1 to 4, with stage 2 being the most common stage (Table SI). The study was approved by the Institutional Review Board of MengChao Hepatobiliary Hospital of Fujian Medical University for the use of patient samples. Informed consent form was received from all participants before surgery in this study. Pathologic and clinical diagnoses of ccRCC were made using the diagnostic criteria of the International Society of Urological Pathology (ISUP) consensus. All methods were performed in accordance with the relevant guidelines and regulations of the Institutional Review Board of MengChao Hepatobiliary Hospital of Fujian Medical University.

To explore the expression profile of ZNF692 in ccRCC, we queried The University of Alabama at Birmingham Cancer data analysis Portal (UALCAN, http://​ualcan.​path.​uab.​edu/​analysis.​html) to assess the expression levels of ZNF692 in ccRCC in a dataset from The Cancer Genome Atlas (TCGA) using normal kidney tissue as a control [18].The input gene symbol "ZNF692" was utilized to query its expression within the TCGA dataset "Kidney renal clear cell carcinoma (KIRC)". To compare the expression of ZNF692 between paired tumor and adjacent normal tissues, we queried the Tumor Node Metastasis (TNM) plotter (https://​www.​tnmplot.​com) website [19]. Using “ZNF692” as keyword, we obtained mRNA expression data for ZNF692 in tumor and adjacent normal tissues and those data were analyzed by the Mann‒Whitney test.

To identify the expression pattern of ZNF692 in normal kidney tissue, we queried the scRNA section of the Human Protein Atlas website (https://​www.​proteinatlas.​org) using ZNF692 as keyword [20]. To explore the role of ZNF692 in ccRCC, we queried Single Cell Portal (https://​singlecell.​broadinstitute.​org/​single_​cell) using keywords “kidney” and “cancer” to search related studies. After identified the studies, we explored the expression pattern of ZNF692 in different cell types exposed to ICB agents using the dataset from the study by Kevin Bi et al [21] using “gene explore” tools and “ZNF692” as keyword.

To determine the prognostic value of ZNF692 expression in ccRCC, the web tool Gene Expression Profiling Interactive Analysis (GEPIA, https:// http://​gepia.​cancer-pku.​cn/​detail.​php) was used to visualize the correlation between ZNF692 expression and the overall survival (OS) of ccRCC patients, with the median expression level as the final stratification cutoff [22]. We evaluated the potential prognostic value of ZNF692 using the least absolute shrinkage and selection operator (LASSO)–Cox regression method. The receiver operating characteristic (ROC) curve was plotted using the “survival” package in R.

To explore the possible functional mechanism of ZNF693 in ccRCC, we analyzed genes coexpressed with ZNF692 and performed Gene Ontology (GO) enrichment analysis using LinkedOmics (http://​www.​linkedomics.​org) [23]. The gene associated with ZNF692 was obtained using RNA-seq data from the "sample cohort: TCGA_KIRC" by employing the "LinkFinder" tools. Subsequently, GO analysis was conducted using the ZNF692 associated genes as input through the utilization of GSEA tools. To explore the function of ZNF692 in patients treated with ICB agents, we performed correlation analysis using the “corrr” package in R to identify genes that were coexpressed with ZNF692 in tumor cells with or without exposure to ICB agents. GO enrichment analysis was performed using Metascape (http://​metascape.​org/​gp/​index.​html#/​main/​step1) [24].

The tissue microarray, containing 29 pairs of primary ccRCC and paracancerous samples, was constructed in house. Briefly, small punches from 29 paraffin-embedded (donor) tissue blocks were obtained with a manual tissue arrayer, and those tissue cores were transferred into a positionally encoded array in a recipient paraffin block. After being sealed with paraffin, the tissue blocks were fixed on a tissue slicer and were sliced into serial sections with a 5 μm thickness. After deparaffinization, IHC staining was performed using an anti-ZNF692 antibody (Bioss, Beijing, China, Cat: bs-4360R) and horseradishperoxidase (HRP) conjugated second antibody (Beyotime, Beijing, China, Cat: A0208). After IHC staining, the sections were scanned and imaged using a scanner (Motic, Xiamen, China). The first sample in the tissue microarray was used as a reference for normalization among sections. The IHCF results were presented as two parameters: the signal intensity (scored from 0 to 4) and the percent of positively stained cells (positively stained cells/total cells per field of view, scored from 0 to 4. A score of 0 indicated the absence of staining, 1 indicated that the percentage of positively stained cells was less than 25%, 2 indicated that the percentage of positively stained cells was between 25 and 50%, 3 indicated that the percentage of positively stained cells was between 50 and 75% and 4 indicated that the percentage of positively stained cells was more than 75%). The total IHC score was calculated by summing the scores of the signal intensity score and the percentage of positively stained cells.

The ccRCC cell lines Caki-1 and 786–0 were gifts from Professor Hongbo Zhao (Kunming Medical University). The cell lines were maintained in Dulbecco’s modified Eagle’s medium (DMEM, HyClone, Utah, USA) supplemented with 10% fetal bovine serum (PAN-biotech, Adenbach, Germany), 100 U/ml penicillin and 100 µg/ml streptomycin (Gibco, New York, USA) at 37 ℃ with 5% CO₂. All those cell lines were mycoplasma free as tested by PCR methods. For transfection, 2 × 10⁵ cells per well were seeded in a 6-well plate and cultured for 12–18 h. When the cells were 50–70% confluent (for siRNA transfection) or 80–90% confluent (for ZNF692 overexpression), the transfection procedure was performed using Lipofectamine 3000 (Thermo, Massachusetts, USA) according to the manufacturer's instructions.

The ZNF692 siRNAs were designed and synthesized by GenePharma (Shanghai, China). A total of three siRNAs were used to knock down ZNF692 expression. The sequences of the targeted siRNAs (si1 to si3) and the negative control siRNA (NC) were as follows: si1-sense: GGAAGUCUUUCAACUUUAAGATT, si2-sense: GGAUGAGGACACUGCACAAAUTT, si3-sense: GGAGCUUGCAGAUUUGGAAUCTT, NC-sense: UUCUCCGAACGUGUCACGUTT. The ZNF692plasmid containing a flag tag was constructed by Sunya (Fuzhou, China) using a gene synthesis method. The cDNA sequence of ZNF692 was synthesized according to CDS sequences of NCBI Reference Sequence: NM_001136036.3 and was inserted into pcDNA3.1-flag backbone by biosune company (http://​www.​biosune.​com/​).

1 μg total RNA was extracted using TransZol Up kit (Transgen, Beijing, China), and reverse-transcribed into cDNA using EasyScript® One-Step gDNA Removal and cDNA Synthesis SuperMix (transgen, Beijing, China) cDNA synthesis kit. Briefly, 1 μg total RNA and 1 nmol random hexamers primers were mixed and incubated at 65 ℃ for 5 min. After incubation, the RNA and primers mixture were put on ice immediately. Then, enzyme and buffer were added to the RNA mixture and ddH₂O was added to makeup total volume to 20 μl. The cDNAs were synthesized at 42 ℃ for 30 min and the enzyme was then inactivated at 85 ℃ for 5 s. cDNAs were quantified using Taq Pro Universal SYBR qPCR Master Mix (Vazyme, Nanjing, Chian) and a QuantStudio 3 Real-Time PCR System (ABI, California, USA) with the beta-actin (ACTB) as the reference. Gene-specific primer for qPCR were designed using NCBI “pink primer” program under default parameters with two changes: (1) primer mast span one exon-exon junction; (2) PCR product size 80 bp-300 bp. The sequences of the primer pairs were listed in Table 1. The qPCR program as following: 95 ℃ for 3min; 40 temperature cycles: 95 ℃ for 30 s,60 ℃ for 20 s,72 ℃ for 20 s; 25 ℃ to 95 ℃ by 0.1 ℃/sec for melting curve. The fluorescence signal was collected at 72 ℃ at temperature cycles stage and consecutive collected at melting curve stage.

The cells were lysed with Radio immunoprecipitation assay lysis buffer (RIPA) (Pierce, 89,900), supplemented with a protease inhibitor ‘cocktail’ (Beyotime, Beijing, China). Bicinchoninic acid assay (Pierce, Illinois, USA) was used to measure protein concentrations. The protein was separated on 10% polyacrylamide gel (PAGE) and transferred to 0.45 μm nitrocellulose (NC) filter membrane (Millipore, Massachusetts, USA). Then the NC membrane was blocked with 10% nonfat powered milk (Amresco, Washington, USA) 1 h at room temperature and then hybridized with primary antibodies (ZNF692, Bioss, Beijing, China, Cat: bs-4360R; GAPDH, Beyotime, Beijing, China, Cat: AG0122) for overnight at 4 ℃ in blocking buffer. The protein-antibody complexes were incubated with peroxidase-conjugated secondary antibodies (Beyotime, Beijing, China, Cat: A0208) and were detected using enhanced chemiluminescence (Pierce, Illinois, USA).

ccRCC cells were cultured and transfected as described above. Forty-eight hours after transfection, cell proliferation, apoptosis, and migration assays were performed. For the cell proliferation assay, transfected cells were seeded into a 96-well plate at 10,000 cells per well, with 8 wells for each condition (siRNA vs. NC). After 24 h, a CCK8 assay was used to test cell proliferation. Cell migration was tested using a wound healing assay and a transwell assay. For the wound healing assay, 1.5 × 10⁵ cells were seeded into a 24-well plate and cultured for 12–18 h. A wound was made using a white Corning pipette tip, and the wells were washed with PBS to remove cell debris and maintained with DMEM without FBS. Twenty-four hours later, the wound was photographed with a CCD camera connected to an inverted microscope (Motic, Xiamen, China), and the wound area was analyzed by ImageJ. The transwell assay was performed using transwell inserts (Corning, 3422, New York, USA). A total of 50,000 cells were seeded into the upper inserts in 100 μl DMEM plus 1% bovine serum albumin (BSA). The cells in the inserts were cultured in 24-well plates containing 600 μl DMEM supplemented with 10% fetal bovine serum (PAN-biotech, Adenbach, Germany), 100 U/ml penicillin and 100 µg/ml streptomycin. After 24 h of incubation, the inserts were washed two times with PBS, and the attached cells were fixed with 100% methanol for 10 min. Then, the inserts were washed three times with PBS, and the cells were stained using 0.1% crystal violet for at least 1 h at room temperature. Then, the inserts were photographed with a CCD camera connected to an inverted microscope, and the cells were counted using ImageJ.

Statistical analysis of tumor gene expression levels across clinicopathological feature subgroups was carried out using the UALCAN web tool. Comparisons of gene expression levels and IHC staining between tumor and adjacent normal tissues were performed using a paired t-test, otherwise unpaired t-tests were used. One-way ANOVA followed by a Turkey test was used to perform multiple comparison. p < 0.05 was set as the criterion for statistical significance. Additionally, KM survival plots with log-rank p values, hazard ratios (HRs), and 95% confidence intervals (CIs) were generated using the GEPIA web tool. p <0.05 indicated a statistically significant difference in OS.

To analyze the prognostic role of ZNF family genes in ccRCC, we identified the ZNF family genes that were differentially expressed in KIRC from GEPIA and evaluated the potential prognostic value of those genes using LASSO–Cox regression analysis. As shown in Fig. 1A, ZNF692 was the most promising gene for further investigation. We performed Kaplan‒Meier survival analysis using the GEPIA web tool. The results showed that higher ZNF692 mRNA expression was associated with poorer prognosis in patients with ccRCC (HR, 2.1; log-rank p = 3.8e−06; Fig. 1B). The mean survival time was 54.17 months in the low-expression group (lower quartile) compared to 31.53 months in the high-expression group (upper quartile). To validate the KM survival analysis results, we performed receiver operating characteristic (ROC) analysis using the “survival” package in R. As shown in Fig. 1C, the AUC was 0.675 with 1< 0.05, suggesting that ZNF692 may be a possible prognostic factor.

To investigate the role of ZNF692 in ccRCC, we analyzed expression profiles using the TCGA KIRC dataset of RNA-seq data for paired tumor and adjacent normal tissues. As shown in Fig. 2A, the expression of ZNF692 mRNA was significantly higher in the ccRCC tumor tissues than in the paired adjacent normal tissues (p < 0.001). To characterize the expression of ZNF692 in ccRCC, we examined the correlations between the expression of ZNF692 and the clinicopathological features of ccRCC. ZNF692 remained upregulated across discrete cancer stages (Fig. 2B), with stage 2, 3, and 4 tissues having higher levels of ZNF692 transcripts than stage 1 and normal tissues. Although there were no significant differences in ZNF692 expression level among stages 2, 3 and 4, there was an increase in the expression level of ZNF692 as the disease progressed (Table 2). To further explore and confirm the role of ZNF692 in ccRCC, we collected 58 pairs of cancer and paracancer samples from ccRCC patients (the patient information is listed in Table SI) and performed IHC analysis using an anti-ZNF692 antibody. Consistent with the mRNA expression data, the IHC results showed that the expression of ZNF692 was higher in tumor tissues than in paracancerous tissues (Fig. 2C, Table 3). Because most of our samples (41/58) were TNM stage 2 tissues, the effect of ZNF692 on tumor stage needs further analysis using a large sample size.

To test the hypothesis that ZNF692 is an oncogene, we performed cell proliferation and migration assays using the ccRCC cell lines Caki-1 and 786–0. We used the transfection of ZNF692 plasmid to overexpress (OE) ZNF692 and used siRNA methods to knockdown (KD) the expression of ZNF692. We designed three siRNA to KD the expression of ZNF692 and used qPCR and WB to test the effects of siRNA. QPCR and WB data (Fig. 3A–C) showing that si-2 had the best effects of KD. So si-2 was used in the flowing study to KD ZNF692. As shown in Fig. 3D, E, ZNF692 KD inhibited the proliferation of ccRCC cells; in contrast, ZNF692 OE (Fig. 3F, G), promoted the proliferation of ccRCC cells. To test the role of ZNF692 in the migration of ccRCC cells, we performed wound healing assay and Transwell assay. The results of the wound healing assay showed that the migration of ccRCC cells was inhibited when ZNF692 was knocked down, while the migration of ccRCC cells was promoted when ZNF692 was overexpressed (Fig. 3H–J). Consistent with wound healing assay, the transwell assay results exhibited that the KD of ZNF692 inhibited the migration of ccRCC cells and the overexpression of ZNF693 promoted the migration of ccRCC cells (Fig. 3K–M).

Since there are limited reports regarding ZNF692 target genes, we determined the genes whose expression correlated with ZNF692 expression and performed gene set enrichment analysis (GSEA) using those correlated genes to explore ZNF692 target genes. The “LinkFinder” and “LinkInterpreter” modules of the LinkedOmics online tool were used to identify genes whose expression was positively or negatively correlated with ZNF692 expression and were then used to perform GO biological process analysis. The positively correlated genes accounted for the majority of the correlated genes and were almost twice as numerous as the negatively correlated genes (Figure S1A). The GSEA results showed that the genes negatively coexpressed with ZNF692 were enriched in pathways associated with the immune system response or inflammation. The most enriched pathways were adaptive immune response, regulation of cytokine production, and T and B cell activation (Figure S1B, orange bars). Moreover, the genes positively coexpressed with ZNF692 were enriched in pathways associated with the energy metabolism process and protein transport and targeting (Figure S2B, yellow bars), such as mitochondrial respiratory chain complex assembly, mitochondrial gene expression, and protein transmembrane transport.

To test ZNF692 target genes, we analyzed the expression of three gene (cyclin dependent kinase 3 (CDK3), GTPase-activating protein (SH3 domain)-binding protein 2 (G3BP2), and transmembrane 9 superfamily member 2 (TM9SF2)), which involved in metabolism or cell cycle, form top 5 coexpressed genes with ZNF692 (Figure S2A). In those three genes, CDK3 was positively associated with ZNF692, G3BP2 and TM9SF2 were negatively associated with ZNF692 (Fig. 4A–C). To explore the role of ZNF692 in the regulation of the expression of CDK3, G3BP2 and TM9SF2, we knockdown the expression of ZNF692 by siRNA and the mRNA expression levels of CDK3, G3BP2 and TM9SF2 were tested using qPCR in ccRCC cell lines 786-0 and Caki-1. As showed in Fig. 4D–F, the KD of ZNF692 did not significantly change the expression of CDK3, however, the expression of G3BP2 and TM9SF2 gene was increased by KD of ZNF692. Furthermore, the prognostic analysis using GEPIA web tools showed that the high expression of G3BP2 and TM9SF2 associated with poor prognostic in the TCGA KIRC cohort (Fig. 4G–H). All those data indicated that ZNF692 negatively regulated G3BP2 and TM9SF2 expression in ccRCC.

To further explore the role of ZNF692 in ccRCC, we analyzed scRNA-seq data from normal human kidney tissue using the scRNA section of The Human Protein Atlas. As shown in Fig. 5A, B, the uniform manifold approximation and projection (UMAP) plot and heatmap showed that there are 13 cell types in normal human kidney tissue: T cells, B cells, macrophages, and various types of tubular cells. T and B cells are the main immune cells expressing ZNF692, and proximal tubular cells (c-5, glomerular parietal epithelial cells) and collecting duct cells (c-12) are the main epithelial cells that express ZNF692. Our IHC results showed that ZNF692 was expressed mainly in endothelial cells of the tubules (Fig. 5C, indicated by red arrows) and cells in the mesenchyme (Fig. 5C, indicated by red stars). The expression pattern of ZNF692 in normal kidney tissue is consistent with its function inferred from the above GSEA results.

Given that ZNF692 was primarily expressed in T and B immune cells, proximal tubular cells, and collecting duct cells, we investigated whether ZNF692 is involved in the response to ICB therapy in ccRCC. To assess the role of ZNF692 in the response to ICB therapy in ccRCC, we queried the Single Cell Portal database with the keyword “kidney”. We found a dataset containing 34326 cells in which the single-cell transcriptomes of cancer and immune cells isolated from patients with metastatic ccRCC before or after ICB therapy were characterized [21]. Fig. 6A shows the expression pattern of ZNF692 across all cell types. Upon exposure to ICB agents, some ZNF692-expressing cells disappeared from the data (Fig. 6B, red circle). UMAP clustering showed that tumor program 1 (TP1) tumor cells were the dominant cells expressing ZNF692 (Fig. 6C) and were responders to ICB therapy. Next, we explored the role of ZNF692 in TP1 cells using coexpression analysis and GSEA of the differentially expressed genes between TP1 and TP2 cells (the genes are listed in Table SII). The most enriched pathway of the positively correlated genes was the spliceosome pathway. Moreover, the pathway that positively regulates type I interferon production was the most enriched pathway of the genes with a negative correlation (Figure S2).