High VSX1 expression promotes the aggressiveness of clear cell renal cell carcinoma by transcriptionally regulating FKBP10

RNA-sequence data were downloaded from TCGA and one Gene Expression Omnibus (GEO) dataset. Clinical data were downloaded from the TCGA database (https://​portal.​gdc.​cancer.​gov/​). The survival prognosis of related genes was investigated via the Kaplan–Meier method using 539 samples of ccRCC in TCGA, with analsyses including overall survival (OS), disease-specific survival (DSS), and progression-free interval (PFI). Kaplan–Meier curves of survival analyses with log-rank P values < 0.05 were drawn. The group cutoff was set as 50% of VSX1 expression. Using a hypothetical two-tailed test, differentially expressed genes (DEGs) between the high- and low-expressed VSX1 groups were identified using the R package, “DESeq2” [20]. Thresholds were set as a log-fold change > 1.5 and an adjusted P-value < 0.05. Based on Spearman’s test, a preranked list of DEGs was ordered by the relationship with VSX1 using the R package “ggplot2” [21]. In addition, an enrichment analysis was conducted using the R package, “ClusterProfiler” to identify VSX1-related molecular mechanisms in DEGs [22]. Gene Ontology (GO) functions and the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways were analyzed to determine the top 100 up- and down-regulated DEGs. The R package, “pheatmap”, was applied to present gene expression (BEST4, LMO1, ARID3C, TEME44, FKBP10, TRIB3) in all ccRCC samples as heatmaps. Pearson’s test was used to verify the relationship of these six DEGs with VSX1, and the results were visualized as scatter plots. To investigate the prognostic benefits of six co-expressed DEGs (BEST4, LMO1, ARID3C, TEME44, FKBP10, TRIB3) in OS, DSS, and PFI, univariate Cox-regression analyses were employed based on the “survival” and “survminer” R packages. A forest plot was used to examine the correlation between the six co-expressed DEGs and prognosis in T stage subgroups and pathological stage subgroups of ccRCC using the R package “forestplot”. To validate the diagnostic efficiency of these six DEGs, receiver operating characteristic (ROC) curves were built and the area-under-the-curve (AUC) value of each gene was calculated by using the R package “pROC” [23]. The prognostic significance of VSX1 in ccRCC was revealed, and the expression level of VSX1 was positively associated with the T stage, lymph node stage, and pathological stage of the disease. Similarly, the expression level of VSX1 was also investigated in the pan-cancer dataset from TCGA using an online tool, Xiantao Academy (https://​www.​xiantao.​love), which consists of multiple kinds of tumor and adjacent tissue samples. The mRNA relative expression level of VSX1 was comparatively analyzed between the adjacent and tumoral tissues based on GSE40435 (http://​ncbi.​nlm.​nih.​gov/​geo/​).

Frozen ccRCC tissues and matched normal adjacent tissues for quantitative real-time PCR (n = 20), and paraffin-embedded cancerous tissues (n = 40) and peritumoral kidney tissue microarrays (n = 19) for immunohistochemistry (IHC), were collected from patients undergoing surgery in the Affiliated Drum Tower Hospital, Medical School of Nanjing University. All patients enrolled in the study were diagnosed with ccRCC according to the World Health Organisation (WHO) classification system, and pathological diagnoses were reviewed retrospectively. Clinical and pathological data were collected from patients and their informed consent to participate in the study was obtained.

Human RCC cell lines in the study included the renal adenocarcinoma cell line 786-O (ATCC, CRL-1932), renal adenocarcinoma cell line 769-P (ATCC, CRL-1933), renal adenocarcinoma cell line ACHN (ATCC, CRL-1611), renal carcinoma cell line A498, clear cell renal carcinoma cell line Caki-1, kidney cortex/proximal tubule cell line HK-2 (ATCC, CRL-2190) and embryonic kidney cell line HEK-293T (ATCC, CRL-3216). The human renal carcinoma cell line A498 was purchased from the National Infrastructure of Cell Line Resource (Beijing, China). The ccRCC cell line Caki-1 was obtained from the Cell Bank of the Chinese Academy of Science (Shanghai, China). These cells, except for Caki-1, were cultured in DMEM (Wisent, 31900608) supplemented with 10% fetal bovine serum (FBS) (Gibco, 2232510). Caki-1 cells were cultured in McCoy’s 5A medium (Gibco, 16600082), containing 10% FBS (Gibco, 2232510). All media contained 1% penicillin and streptomycin (Gibco, 15140122). All cell lines were incubated at 37 °C with 5% CO₂.

Total protein was extracted from cells for western blotting. Cells were lysed with moderate RIPA lysis buffer (Beyotime, P0013C) after rinsing three times in PBS at 4 °C and were centrifuged. The resulting supernatant was mixed with loading buffer and heated at 100 °C for 10 min. Proteins were separated by sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and transferred to a polyvinylidene fluoride (PVDF) membrane (Roche, Basel, Switzerland) and blocked in 5% bovine serum albumin (BSA) (SigmaAldrich) for 1 h at room temperature. Blots were then incubated with primary antibodies at 4 °C overnight, and secondary antibodies conjugated with horseradish peroxidase were incubated for 1 h at room temperature. Protein signals were detected with ECL solution (Millipore) and quantified with Image J software (National Institutes of Health). β-actin (ACTB) was used as an internal control.

Total RNA was isolated using the RNA-easy isolation reagent (Vazyme, R701-01) according to the manufacturer’s instructions and then reverse-transcribed into cDNA with HiScript Q RT SuperMix for qPCR (Vazyme, R122-01). Subsequently, quantitative real-time PCR was performed to quantify cDNA using Taq Pro Universal SYBR qPCR Master Mix (Vazyme, Q712-02). For normalization, 18S rRNA was used as an internal reference, and 2^−ΔΔCt was used to analyze the relative expression levels of target genes. Primer sequences for target genes were listed in Additional file 1: Table S1.

ParafFin-embedded tissue microarrays were dewaxed using dimethyl benzene and rehydrated with gradient ethanol solutions. Then, 3% H₂O₂ was used to inhibit the endogenous peroxidase for 15 min, and the antigens were retrieved by microwave heating. The antigens were blocked with 10% BSA for 1 h. The tissue microarrays were incubated with the primary antibody against VSX1 (1:300; LS-C829216, Life science) at 4 °C overnight and then incubated with the secondary antibody in the dark for 2 h. Diaminobenzidine was used to visualize immunostaining. The staining intensity of VSX1 was quantified with Image J software.

The indicated cells were seeded in the 96-well plates (3 × 10³ cells/well) and cultured overnight. The next day, cell proliferation was evaluated using the Cell Counting Kit-8 reagent (CCK-8, MCE), and the absorbance values were measured at 450 nm. Thereafter, 1 × 10³ RCC cells were seeded into 6-well plates and cultured for 2 weeks. The number of cell clones formed was assessed after crystal violet (Beyotime, C0121) staining, which was imaged and counted under a microscope.

Cell migration and invasion experiments were performed using a Transwell chamber (Corning, USA). After serum-starving the cells for 24 h, 1 × 10⁵ cells with 200 μL serum-free medium were seeded into the upper chamber, and the lower chamber was filled with 500 μL DMEM (Wisent, 31900608) supplemented with 10% FBS (Gibco, 2232510). The migrated cells were imaged and counted under a microscope after crystal violet (Beyotime, C0121).

Flow cytometry was conducted based on the manufacturer’s instructions. Cells were measured with a BD Beckman cytometer (BD Biosciences) and analyzed using FlowJo software after incubating with the Annexin V-PE/7-ADD Apoptosis Detection Kit (Vazyme, A213-01). Similarly, cells were incubated with the PI/RNase Staining Kit (BD Biosciences) to analyze the cell cycle and were quantified with a BD Beckman cytometer.

The pGL3-Basic plasmids-containing promoter region of the target gene and VSX1 overexpression plasmid was constructed and transfected into HEK-293T cells. The pRL-TK-Renilla luciferase plasmid was transfected into the same cells and used as an internal reference. The dual-Luciferase Reporter Assay Kit (Vazyme, DL101-01) was used to evaluate the activities of firefly luciferase and Renilla luciferase. The pGL3-Basic vector plasmid was used as an internal control for reporter genes.

For the tumor sphere formation assay, 1 × 10³ Caki-1 or 786-O cells were seeded into U-bottom ultra-low attachment 96-well plates (Corning, cat. no. 174925) and cultured in DMEM supplemented with 10% FBS for 2 weeks. Tumor sphere images were captured under a microscope and the area was measured with Image J software (National Institutes of Health).

The VSX1 overexpression plasmid ligated with the linearized TK-PCDH-copGFP-T2A-Puro vector plasmid was constructed by Tsingke Technology (Beijing, China). Promoter sequences for the target genes—TMEM44, FKBP10, and TRIB3—were acquired from the National Center for Biotechnology Information (NCBI; https://​www.​ncbi.​nlm.​nih.​gov/​), and the purified promoter fragments were ligated with the linearized pGL3Basic vector plasmid. The sequences of VSX1-shRNA and FKBP10-shRNA were obtained from Sigma (https://​www.​sigmaaldrich.​cn/​CN/​zh) and ligated with pLVshRNA-EGFP-Puro vector plasmid. All targeted sequences were synthesized by Tsingke Technology (Beijing, China). Cells were transfected with shRNA or plasmids as per the manufacturer’s protocols of jetPRIME (Polyplus-transfection, Shanghai, Illkirch, France). Cells were harvested 48 h after transfection. The shRNA sequences are listed in Additional file 1: Table S1.

Data were expressed as the means ± standard deviation. SPSS 23.0 software (SPSS Inc., Chicago, IL) was used for statistical analyses of the data. GraphPad Prism 8.0 (GraphPad Software, San Diego, CA, USA) was used to depict Kaplan–Meier survival curves and compare between subgroups using the log-rank test. Spearman’s correlation coefficient was calculated to analyze the connection between VSX1 expression and that of other target genes. Student’s t-test and one-way analysis of variance (ANOVA) were used to assess the significance of differences; P < 0.05 was considered statistically significant (*P < 0.05, ** P < 0.01, and *** P < 0.001).

To search for novel prognostic biomarkers of ccRCC, the survival prognosis, including OS, DSS, and PFI, of related genes was investigated for patients with ccRCC. Differentially expressed target genes were identified between ccRCC and peritumoral tissues. A four-way Venn diagram showed there were 131 DEGs in ccRCC tissues compared with peritumoral tissues (Fig. 1a and Additional file 2: Table S2). Among these genes, 80 were potential protein-coding genes. A total of 56 candidate upregulated genes were identified and were listed in Additional file 2: Table S2. However, some of these genes, such as RNASET2, TF, and ZIC2, were already known to have an oncogenic role in the tumorigenesis, progression, and metastasis of renal cancer [24‐26]. Among these candidate upregulated genes, VSX1 was selected and further analysis focused on the oncogenic role of VSX1 in ccRCC. Figure 1b showed that VSX1 mRNA was not only upregulated in cancerous tissues of ccRCC, but was also upregulated in pan-cancer tissues. In addition, the expression level of VSX1 in ccRCC tissues was higher than that in the matched normal adjacent tissues (Fig. 1c). Clinical correlation studies of data revealed that upregulated VSX1 mRNA was strongly associated with advanced T stages, distant metastasis, and high pathological stages. However, no such correlation was found with lymphatic metastasis (Fig. 1d). Similarly, the VSX1 mRNA level and an advanced T stage or a high pathological stage were associated with papillary RCC (pRCC) (Fig. 1d). Additionally, the survival analyses revealed that OS, DSS, and PFS were shortened for ccRCC patients with a high expression of VSX1 (Fig. 2a). However, no prognostic difference was observed in pRCC patients (Fig. 2b). These data implied that the upregulation of VSX1 might play an essential role in the progression of ccRCC.

After screening VSX1 as a potential prognostic biomarker by using bioinformatics methods, VSX1 mRNA levels were detected in 20 pairs of ccRCC tissues and adjacent nontumoral tissues by qRT-PCR. Of the 20 cancerous samples, 19 (95%) had upregulated VSX1 mRNA levels compared with adjacent noncancerous tissues (P = 0.0278, Fig. 3a and b). The upregulation of VSX1 was consistently confirmed between the adjacent and tumoral tissues based on dataset GSE40435 from the GEO database (P = 0.0023, Fig. 3c). Furthermore, the upregulation of VSX1 was also confirmed in cancerous tissues by the IHC assay. IHC scoring showed that the expression of VSX1 increased in cancerous tissues compared with normal kidney tissues (Fig. 3d and e).

The expression of VSX1 was further detected in HEK-293 T cells, the immortal kidney cortex/proximal tubule cell line HK-2, and a panel of human RCC cell lines by qRT-PCR. Human RCC cells (786-O, 769-P, ACHN, A498, and Caki-1) exhibited higher expression of VSX1 compared with HEK-293 T and HK-2 cells, with A498 and Caki-1 cells showing the highest expression (Fig. 3f). The Caki-1 cell line was acquired from metastatic sites of the skin, while the A498 cell line was initially obtained from the primary tumor of a female patient.

To reveal the oncogenic role of VSX1 upregulation in ccRCC progression, VSX1 was overexpressed in the ccRCC cell line 786-O and knocked down in the Caki-1 cell line. Stable VSX1 overexpression or knockdown cell lines were constructed using a lentivirus system and confirmed by qRT-PCR and western blotting analyses (Fig. 3g and h).

In the CCK-8 assay, 786-O cells with stable VSX1 overexpression displayed significantly enhanced growth compared with that in the empty vector control group (Fig. 3i). In contrast, VSX1 knockdown in Caki-1 cells markedly suppressed cell proliferation compared with the shRNA negative control (shNC) group. Moreover, colony formation assays indicated that VSX1 knockdown in Caki-1 cells inhibited colony formation; the opposite results were revealed in 786-O cells overexpressing VSX1 compared with the empty vector control group (Fig. 3j). Furthermore, a cell cycle assay showed a marked reduction of cell populations in the S phase induced by the knockdown of VSX1 in Caki-1 cells, and the opposite results were obtained in 786-O cells with overexpression of VSX1 in comparison with the shNC group (Fig. 3k). This result was also confirmed by the CCK-8 assay.

Flow cytometry analyses implied that the knockdown of VSX1 expression significantly increased the apoptosis rate in ccRCC cells (Fig. 3l), and opposing results were obtained in 786-O cells overexpressing VSX1 compared with the empty vector control group. Moreover, the tumor sphere formation assay revealed that knockdown of VSX1 expression in Caki-1 cells significantly inhibited the tumor sphere formation capacity of the cells compared with that of the shNC group (Additional file 3: Fig. S1a). Conversely, compared with the empty vector control group, the overexpression of VSX1 in 786-O cells markedly enhanced tumor sphere formation ability of the cells (Additional file 3: Fig. S1b). The Transwell assay, performed to evaluate the migration and invasion abilities of ccRCC cells, revealed that upregulation of VSX1 promoted the migration and invasion abilities of Caki-1 cells and 786-O cells, and vice versa (Additional file 3: Fig. S1c and d). Collectively, these findings suggested that VSX1 acted as a tumor activator gene in ccRCC.

Gene correlation analyses were performed between VSX1 and other DEGs in ccRCC from TCGA using the Spearman’s correlation coefficient. As shown in Fig. 4a, many upregulated or downregulated genes in ccRCC were correlated with the expression of VSX1. KEGG analysis via a bubble chart of the top 100 up- or down-regulated genes indicated that these genes showed good correlation with VSX1 and might play a leading role in the transcriptional misregulation of cancer and organic acid transmembrane transporter activity (Fig. 4b).

Since VSX1 is essentially a transcription factor [27], we focused on the top upregulated genes relevant to transcriptional dysregulation, such as BEST4, LMO1, ARID3C, TMEM44, FKBP10, and TRIB3. Pearson’s correlation analyses suggested that the expression levels of BEST4, LMO1, ARID3C, TMEM44, FKBP10, and TRIB3 were positively correlated with upregulated VSX1, which was also observed in the heatmap analysis (Fig. 4c and d). Examination of the expression of these genes in ccRCC samples revealed that mRNA expression of these genes was higher in cancerous tissues than in the normal adjacent tissues, and this finding was confirmed in the matched normal and cancerous tissues (Fig. 4e and f).

The expression of these genes in ccRCC samples was further investigated in relation to the T stage, N stage, M stage, and pathological stage. As shown in Fig. 4g–j, upregulated mRNA levels of BEST4, LMO1, and ARID3C were associated with an advanced T stage and a high pathological stage; however, the association was not found in the N stage and M stage. In addition, upregulated mRNA levels of TMEM44, FKBP10, and TRIB3 were strongly associated with advanced T stage, lymphatic metastasis, distant metastasis, and a high pathological stage. Survival analyses implied that OS, DSS, and PFI were shortened for ccRCC patients with high expression of these genes (Fig. 5a). Moreover, high expression of TMEM44, FKBP10, and TRIB3 was an unfavorable prognostic risk factor for ccRCC patients in advanced T and high pathological stages (Fig. 5b). These genes might therefore have potential as diagnostic biological markers for ccRCC (Fig. 5c). These data led to the reasonable assumption that VSX1 might transcriptionally regulate these tumor-related genes either directly or indirectly, and this hypothesis warranted further experimental verification.

The upregulation of TMEM44, FKBP10, and TRIB3 was confirmed between the adjacent and tumoral tissues based on dataset GSE40435 from the GEO database (Fig. 6a). Gene correlation analyses between VSX1 and TMEM44, FKBP10, or TRIB3 were also performed based on dataset GSE40435. TMEM44, FKBP10, and TRIB3 were positively correlated with VSX1 (Fig. 6b). Next, the expression of these genes was detected by qRT-PCR in 20 pairs of ccRCC tissues and adjacent nontumoral tissues. There was no significant difference in the expression of TMEM44 between cancerous samples and adjacent noncancerous tissues (Fig. 6c), but both FKBP10 and TRIB3 mRNA were upreglated in most cancerous samples (Fig. 6d and e).

The dual-luciferase reporter gene assay revealed that the knockdown of VSX1 expression significantly decreased the relative luciferase activities of TMEM44, FKBP10, and TRIB3 (Fig. 6g), while the overexpression of VSX1 increased the relative luciferase activities of these genes (Fig. 6h). Next, VSX1 protein levels were changed by either knockdown or overexpression in vitro, and the expression levels of TMEM44, FKBP10, and TRIB3 in Caki-1 and 786-O cells were analyzed via qRT-PCR, respectively. The expression of TMEM44, FKBP10, and TRIB3 was downregulated by the knockdown of VSX1 compared with the negative control in Caki-1 cells (Fig. 6i), and similar results were confirmed in 786-O cells (Fig. 6j). The CCK-8 assay showed that significantly enhanced growth in 786-O cells with stable VSX1 overexpression was almost restored to normal by the knockdown of FKBP10 (Fig. 6k).

The CCK-8 assay showed significantly inhibited growth in Caki-1 cells with stable FKBP10 knockdown (Additional file 4: Fig. S2a), while the Transwell assay demonstrated that the knockdown of FKBP10 inhibited the migration and invasion abilities of Caki-1 cells (Additional file 4: Fig. S2b and c). Colony formation assays revealed that FKBP10 knockdown in 786-O cells inhibited colony formation (Fig. 6l and Additional file 4: Fig. S2d). In addition, the tumor sphere formation assay indicated that the ability to inhibit the tumor sphere formation in 786-O cells with the knockdown of FKBP10 after stable VSX1 overexpression was stronger than that in the shNC group (Additional file 4: Fig. S2e and f). A Transwell assay was performed to illuminate the impact of FKBP10 on the behavior of 786-O cells and showed that downregulation of FKBP10 inhibited the migration and invasion ability of the cells compared with the shNC group (Additional file 4: Fig. S2g and h). These findings suggested that VSX1 affected tumor invasiveness via transcriptionally regulating FKBP10.