A positive feed-forward loop between LncRNA-URRCC and EGFL7/P-AKT/FOXO3 signaling promotes proliferation and metastasis of clear cell renal cell carcinoma

Microarray analysis for the expression of lncRNAs was performed by Shanghai Gminix Biological Information Company (Shanghai, China), applying a pipeline as previously described [26, 27] to identify the probe sets uniquely mapped to lncRNAs from the Affymetrix array, in which special way can evaluate the lncRNA expressions in RCC gene expression data. The accession numbers for the microarray data are Gene Expression Omnibus database GEO: GSE46699 and GSE53757 [28, 29]. Gene expression profiles of the 786-O cells with or without URRCC downregulation were determined using Human Cancer Focus mRNA PCR Array following the manufacturer’s instructions. Microarray experiments were performed by KangChen Bio-tech, Shanghai, China. The differentially expressed genes with statistical significance were identified using volcano plot filtering. The threshold we used to screen upregulated or downregulated genes is fold change > = 2.0 and a p-value <= 0.05.

A total of 116 tumor samples and paired non-cancerous renal tissues were obtained from patients who underwent partial or radical nephrectomy in Department of Urology, Shanghai Tenth People’s Hospital, Tongji University (Shanghai, China). The fresh tissues were frozen in liquid nitrogen to protect the protein or RNA away from degradation. The use of human tissues was approved by the ethics committee of Shanghai Tenth People’s Hospital.

The human RCC cell lines A498, 786-O, CaKi-1, ACHN, 769-P and OSRC-2 and human normal renal tubular epithelial cell line HK-2 were obtained from Cell Bank of the Chinese Academy of Sciences (Shanghai, China). A498 was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco, USA) plus 10% Fetal Bovine Serum (FBS, Hyclone, USA) with 1% Penicillin/Streptomycin (P/S, Gibco, USA). 786-O, CaKi-1, ACHN, 769-P and OSRC-2 cells were cultured in RMPI 1640 (Gibco, USA) plus 10% Fetal Bovine Serum (FBS, Hyclone, USA) with 1% Penicillin/Streptomycin (P/S, Gibco, USA). HK-2 cells were cultured in Keratinocyte Medium (KM, ScienCell, USA) plus 1% Keratinocyte Growth Supplement (KGS, ScienCell, USA) with 1% Penicillin/Streptomycin (P/S, ScienCell, USA). All cells described above were cultured at 37° in 5% CO₂.

According to the manufacturer’s protocol, short interfering RNA (siRNA) si-EGFL7/si-FOXO3/si-NC (IBS Solutions Co. Ltd., China) were transiently transfected in A498 cells using Lipofectamine 3000 (Invitrogen, USA) at a final concentration of 100 nM. Lentiviral sh-control/sh-URRCC and oe-URRCC/oe-FOXO3 were purchased from Shanghai Integrated Biotech Solutions Co., Ltd. Lentiviral infection was performed as previously described [30, 31]. In order to avoid the off-target effects of shRNA, we designed two different shRNAs for lncRNA-URRCC and two different siRNAs for EGFL7 and FOXO3. The sequences of oligonucleotides used in this experiment were listed in Additional file 1: Table S1 and Additional file 2: Table S2.

FISH was performed as previously described [14]. It is a powerful technique that uses non-toxic fluorescent DNA probes to target any given sequence within a nucleus, resulting in colored signals that are detected with a fluorescence microscope and was performed at Biosense Co. Ltd. Paraffin-embedded tissue blocks were retrospectively retrieved from renal cancer patients. Quantum dot fluorescent in situ hybridization (QD-FISH) was performed to detect the presence of URRCC using a digoxin-labeled oligonucleotide probe indirectly labeled with digoxin-antibody-conjugated quantum dots.

Cytoplasmic and nuclear RNA were isolated and purified using the Cytoplasmic & Nuclear RNA Purification Kit (Norgen, Belmont, CA) according to the manufacturer’s instructions.

RCC cells were seeded in 24-well plates (3000 cells/well) and cultured for 0, 2, 4, or 6 days. Cells were harvested and cell numbers were calculated using MTT agent. DMSO was used as the control. We added 250 μl of 5 mg/ml MTT to each well, incubated for 2 h in incubator at 37 °C, removed the media and added 150 μl DMSO. We then covered with tinfoil and agitated cells on orbital shaker for 15 min and then read the absorbance at 570 nm.

The invasive capability of RCC cells was determined by the Transwell assay. The upper chamber of Transwll was pre-covered with Matrigel (BD356230, CORNING, USA). RCC cells were harvested and seeded with serum-free DMEM into the upper chambers at 5 × 10⁴ cells/well, and the bottom chambers contained DMEM with 10% FBS, and then transwells incubated for 24 or 36 h at 37 °C. Following incubation, the invasive cells attached to the lower surface of the membrane were fixed by 4% paraformaldehyde and stained with 1% toluidine blue. Cell numbers were counted in five randomly chosen microscopic fields (100×) per membrane.

Six-week-old male BALB/c nude mice were purchased from ShanghaiSipper-BK laboratory animal Company (Shanghai, China). After 5 days of adapting to the environment, a total of 1 × 10⁶ A498 cells (stably transfected with sh-URRCC and sh-control) and OSRC-2 cells (stably transfected with oe-URRCC and mock) were injected hypodermically into the right armpit or tail vein of each mouse, respectively (n = 8 mice/group). Cells were also transduced with luciferase for the non-invasive in vivo imaging system (IVIS) (NightOWL II, LB983, Berthold Technologies, Germany) that was performed once a week. The tumor volume and the weight of each mouse were measured per week. The mice were killed on the 60th days after injection. We also measured and recorded the weight of each tumor. All animal studies were approved by the Institutional Animal Care and Use Committee of the Shanghai Tenth People’s hospital.

Total RNA was extracted from frozen tissues or cells using Trizol reagent (Invitrogen, USA) according to the manufacturer’s instructions. Reverse transcription was performed using a PrimeScript RT reagent kit (TaKaRa, Japan), and qRT–PCR was performed with KAPA SYBR FAST qPCR Kit (Kapa Biosystems, USA) using a 7900HT Fast Real-Time PCR System (Applied Biosystems, Japan). The mRNA and lncRNA levels were normalized against β-actin in cell and tissue lysates. The miRNA levels were normalized against U6. Data were analyzed using the 2^-ΔΔCt method. The primer sequences were listed in the Additional file 1: Table S1.

Protein extracts were separated from cells or human tissues by RIPA buffer containing protease inhibitors. 30 μg of protein extracts were loaded to 8–10% sodium dodecylsulfate–polyacrylamide gel electrophoresis gels and transferred onto nitrocellulose membranes. The membranes were hybridized with a primary antibody at 4 °C overnight, and incubated with a secondary antibody in 1 h at room temperature. The expression of β-actin was used as loading control. The intensity of the fluorescence was scanned by the Odyssey scanner (LI-COR Biosciences, USA). The information of antibodies was listed as follow: EGFL7 (1:500, proteintech, 19,291–1-AP), AKT (1:10000, Abcam, ab179463), P-AKT (1:1000, Abcam, ab131443), FOXO3 (1:500, Abcam, ab12162), H3K27ac (1:1000, Abcam, ab4729), H3 (1:1000, Abcam, ab1791), β-actin (11,000, Abcam, ab8226).

Immunohistochemistry for the target molecules was performed on paraffin sections using a primary antibody against EGFL7 (1:50, proteintech, 19,291–1-AP), FOXO3 (1:500, Abcam, ab12162), P-AKT (1:50, Abcam, ab131443), Ki67 (1:500, Abcam, ab92742) and the proteins in situ were visualized with 3, 3-diaminobenzidine. For histological scoring, 3 high power fields (magnification, × 400) were randomly selected from renal cancer tissues and normal renal tissues. The reactivity degree was assessed by at least two pathologists without knowledge of the clinicopathological features of the tumors. The degree of positivity was initially classified according to scoring both the proportion of positive staining tumor cells and the staining intensities. Scores representing the proportion of positively stained tumor cells were graded as: 0 (< 10%); 1 (11–25%); 2 (26–50%); 3 (51–75%) and 4 (> 75%). The intensity of staining was determined as: 0 (no staining); 1 (weak staining = light yellow); 2 (moderate staining = yellow brown); and 3 (strong staining = brown). The staining index (SI) was calculated as the product of staining intensity × percentage of positive tumor cells, resulting in scores of 0, 1, 2, 3, 4, 6, 8, 9 and 12. Only cells with clear tumor cell morphology were scored.

EZ ChIP™ Chromatin Immunoprecipitation Kit (Millipore, Bedford, MA, USA) and the EpiQuik Tissue Acetyl-Histone H3 and H4 CHIP Kit (Epigentek Group Inc., NY, USA) were used to perform ChIP assays. In brief, cells were crosslinked with 4% formaldehyde for 10 min followed by cell collection and sonication with a predetermined power to yield genomic DNA fragments of 300–1000 bp long. The chromatin was immunoprecipitated using an anti-acetyl-histone H3 or anti-acetyl-histone H4 antibody, or anti-FOXO3 antibody. Normal mouse IgG or normal rabbit IgG were used as the negative control. qRT-PCR was conducted using KAPA SYBR FAST qPCR Kit (Kapa Biosystems, USA). PCR products were analyzed by agarose gel electrophoresis. Specific primer sequences are listed in Additional file 1: Table S1.

Cells were co-transfected with psiCHECK2 dual-luciferase reporter and pcDNA3.1-URRCC-promoter (wt or mt) or pcDNA3.1-FOXO3 using Lipofectamine 3000 (Invitrogen, USA). All groups were run in triplicate in 48-well plates. Luciferase activity was measured by Dual-Luciferase Assay (Promega, Madison, WI, USA) according to the manufacturer’s manual after 48 h of transfection. Renilla luciferase activity was normalized against Firefly luciferase activity.

The preprocessed level 3 RNA-seq data and corresponding clinical information of clear cell renal cell carcinoma patients were collected from The Cancer Genome Atlas (TCGA) database (http:// cancergenome.​nih.​gov/). All statistical analysis was done using IBM SPSS Statistics (Version 20.0, IBM) and Graphpad Prism V6 (GraphPad Software, Inc., 2012). Student’s t-test, one-way ANOVA, LSD-t test, Pearson chi-square test, Log-rank test, linear regression, and Cox Regression Analyses were performed as indicated. All data was considered significantly when *p<0.05.

To search for the potential lncRNAs involved in ccRCC progression, we systematically analyzed the dis-regulated lncRNAs expression profiles of ccRCC tissues versus matched normal kidney tissues using two published GEO DataSets (GSE46699 and GSE53757). The detailed program was shown in Additional file 3: Figure S1A.

We computed the intersection between the result sets of these two datasets through heatmap analysis and focused on 3 potential lncRNAs (AK026225, BX649059 and AK055783) whose expression increased among top 50 lncRNAs that up-regulated in ccRCC tissues compared to paired normal renal tissues (Additional file 3: Figure S1B and C). The expression of these 3 lncRNAs was validated using qRT-PCR analysis and the outcome showed that only BX649059 expression was dramatically higher in 20 ccRCC tissues compared to paired noncancerous renal tissues (Fig. 1a), which was consistent with the result of datasets and the analysis of ccRCC patients from TCGA KIRC dataset (Fig. 1b and Additional file 4: Table S3). However, the other two lncRNAs (AK026225 and AK055783) seemed to be no differences between ccRCC tissues and paired noncancerous renal tissues (Additional file 3: Figure S1D). Moreover, the transcript level of BX649059 was reproducibly upregulated in 96 ccRCC tissues compared to 25 unpaired normal renal tissues (Fig. 1c). Therefore, we named BX649059 as LncRNA-URRCC (up-regulation in clear cell renal cell carcinoma). In addition, the qRT-PCR analysis revealed remarkably higher URRCC expression in a series of human renal cancer cell lines than that in human normal renal tubular epithelial cell line, HK2 cells (Fig. 1d). By contrast, such a change was not observed in the expression of the other two lncRNAs (AK026225 and AK055783) (data no shown).

Using Ensemble software, we further confirmed the character of URRCC was from clone DKFZp779P0730, position within chromosome 12: 100,623,715-100,628,286 (Additional file 3: Figure S1E). The full sequence of URRCC, with transcript length 3967 bp, was presented in Additional file 3: Figure S1F. Furthermore, RNA fluorescence in situ hybridization (FISH) on ccRCC tissues detected that URRCC (green) was mainly located in the cytoplasm while the nucleuses were stained with DAPI (blue). Subcellular fraction assay also found that URRCC was mainly located in the cytoplasm, consistent with online software lncLocator (http://​www.​csbio.​sjtu.​edu.​cn/​bioinf/​lncLocator/​) predicting the location of URRCC (Fig. 1e, f and Additional file 3: Figure S1G).

To further explore the relationship between URRCC expression and the clinicopathological characteristics of 116 patients with ccRCC, we analyzed URRCC expression in ccRCC tissues. We ranked URRCC mRNA level and selected the median (4.215) of the whole set of data as the cutoff point between the high URRCC group and the low URRCC group (Fig. 1g). Clinicopathological analysis confirmed that high level of URRCC in tumors was associated with tumor size, T stage and metastasis (Additional file 5: Table S4). In the univariate analysis, high URRCC expression in tumors (hazard ratio, HR = 2.59; 95% confidence interval, CI = 1.46–4.33; p = 0.001), Tumor size (HR = 1.59; 95%CI = 1.44–2.74; p = 0.035), T stage (HR = 1.42; 95%CI = 1.25–2.68; p = 0.038) and metastasis (HR = 1.47; 95%CI = 1.05–1.86; p = 0.026) was remarkably correlated with overall survival (Additional file 6: Table S5). Multivariate analysis showed that a higher URRCC expression (HR = 2.44; 95%CI = 1.36–4.40; p = 0.003) were significantly associated with overall survival (Additional file 6: Table S5). In addition, Kaplan-Meier analysis in the 116 patients with ccRCC demonstrated that high URRCC expression level in ccRCC tissues dramatically correlated with a reduction in 5 years overall survival rates (Fig. 1h).

The result above demonstrated that URRCC overexpressed in ccRCC tissues and potentially used as diagnostic or prognostic markers. To ascertain whether URRCC could impact RCC progression, we chose the shRNA with the best interference from two shRNAs, which were confirmed in 786-O and OSRC-2 cells (Additional file 7: Figure S2A and B). Preliminary experiments showed that knocking down the expression of URRCC could inhibit the proliferation and invasion of 786-O cells (Additional file 7: Figure S2C and D). Then we established to knock down URRCC with sh-URRCC in human A498 cells. As shown in Fig. 2a and b, knocked-down URRCC dramatically decreased the growth speed of A498 cells, when compared with control cells through MTT assay. In addition, transwell invasion assay also demonstrated that sh-URRCC attenuated invasive capability of A498 cells than that of control cells (Fig. 2c). The effect of sh-URRCC was also confirmed in 786-O cells (Additional file 7: Figure S2C and D).

Conversely, we examined the tumor-inducing phenotype in OSRC-2 cells with overexpression of URRCC (oe-URRCC) (Fig. 2d). As expected, upregulating URRCC expression obviously enhanced proliferation of OSRC-2 cells compared with the mock group via MTT assay (Fig. 2e). Consistently, overexpressed URRCC also induced invasion of OSRC-2 cells through transwell invasion assay (Fig. 2f).

Together, results from Fig. 2 substantiated that URRCC could promote cell proliferation and invasion in ccRCC cells.

The entire above in vitro data demonstrated that URRCC might play a positive role to enhance ccRCC cell proliferation and invasion. We next explored the proliferative role of URRCC in vivo. The A498/sh-URRCC cells as well as their corresponding control group cells were inoculated into the right flank of nude mice. As Fig. 3a showed, URRCC knocked-down effectively inhibited tumor proliferation, as revealed by measuring tumor mass. On the contrary, stable transfection of overexpressing URRCC (oe-URRCC) into OSRC-2 cells resulted in dramatically induced tumor growth (Fig. 3b).

We next generated both cells labeled with firefly luciferase and then injected into tail vein of nude mice to monitor tumor metastasis. A dramatic reduction of metastatic luciferase expression in lungs of A498/sh-URRCC mice was detected by In Vivo Imaging Systems (IVIS) compared with sh-control group (Fig. 3c and d). HE analysis also showed a decrease of pulmonary metastasis in A498/sh-URRCC group compared with sh-control group (Fig. 3f). Contrarily, URRCC overexpression yielded an enhanced pulmonary metastasis of OSRC-2 cells in vivo (Fig. 3c, e and g). In addition, IHC demonstrated the decreased expression of Ki67 from sh-URRCC group compared with sh-control group, while the increased expression of Ki67 from oe-URRCC group compared with the mock group in pulmonary metastasis (Fig. 3h and i).

Together, results from Fig. 3 demonstrated that URRCC functioned as a critical tumor enhancer in vivo by promoting tumor proliferation and metastatic colonization.

Recently, lncRNA has been reported to regulate mRNA and was independent of functioning as competing endogenous RNAs (ceRNA), although the lncRNA was mainly located in cytoplasm [14]. To investigate the potential gene related with URRCC-driven proliferative and metastatic phenotype in ccRCC, we used “Human Cancer Focus mRNA PCR Array” from KangCheng Bio-tech Inc. to compare dysregulated mRNAs that shown more than 2 fold change difference in expression between sh-control and sh-URRCC in the treated A498 cells (Additional file 8: Figure S3A, Additional file 9: Table S6). Volcano plots indicated six down-regulated mRNAs and four up-regulated mRNAs (Fig. 4a and Additional file 8: Figure S3B). We focused on genes that have the same changing trend with URRCC. Among them, JMY and EGFL7 were markedly decreased by knocking down URRCC at mRNA expression (Fig. 4a). We chose EGFL7 as our candidate for next experiments due to the tumor suppressor role of JMY (Additional file 8: Figure S3C and D, Additional file 10: Table S7 and Additional file 11: Table S8) in ccRCC. qRT-PCR and WB analysis revealed that sh-URRCC dramatically decreased the mRNA and protein levels of EGFL7 in A498 and OSRC-2 cells (Fig. 4b), whereas upregulated mRNA and protein levels of EGFL7 were identified in oe-URRCC group of both cells (Fig. 4c). Meanwhile, qRT-PCR assays showed there is a positive correlation between URRCC and EGFL7 in mRNA level in cell lines (Additional file 8: Figure S3E). In addition, The TCGA KIRC dataset analysis and IHC analysis substantiated that EGFL7 mRNA levels and protein levels were much higher in the ccRCC tissues than their noncancerous tissues (Fig. 4d and e, Additional file 12: Table S9). Moreover, subcutaneous xenograft tumors from sh-URRCC group demonstrated a decrease in EGFL7 expression by IHC compared with those from the sh-control group, while subcutaneous xenograft tumors from oe-URRCC group exhibited an increase in EGFL7 expression by IHC compared with those from the mock group (Fig. 4f and g).

To investigate the mechanism responsible for the upregulation of EGFL7 in ccRCC, we probed for EGFL7 could be regulated by inhibitors of histone deacetylases according to previous studies revealing that the histone acetylation was the important regulation approach for EGFL7 [25]. qRT-PCR and WB assays indicated that EGFL7 was upregulated via the histone deacetylase inhibitor trichostatin A (TSA) in A498 and OSRC-2 cells (Fig. 4h). Meanwhile, ChIP assay confirmed that sh-URRCC caused a significant decreased level of histone H3 acetylation, but not histone H4 across EGFL7 promoter region (− 2000 to + 50 bp relative to Transcriptional start site (TSS)) in A498 cells (Fig. 4i). Additional ChIP assay indicated that sh-URRCC could not affect the H3 and H4 acetylation of GAPDH promoter in A498 cells (Fig. 4j). Conversely, oe-URRCC increased histone H3 acetylation level across EGFL7 promoter region in OSRC-2 cells (− 2000 to + 50 bp relative to TSS) (Fig. 4k). To further investigate the mechanism that why EGFL7 expression was regulated by histone H3 acetylation, we explored the potential histone acetylation sites from the promoter of EGFL7 via UCSC browser. Interestingly, aberrant enrichment of the H3K27ac histone mark were found across the promoter region of EGFL7 (Fig. 4l). Consistently, we found TSA could increase expression of H3K27ac in protein level in A498 and OSRC-2 cells, which may explain why H4 acetylation was not affected (Fig. 4m). Moreover, ChIP assay showed that TSA could reverse the down-regulation of EGFL7 caused by sh-URRCC both on acetylation and expression of EGFL7 (Additional file 8: Figure S3F-H).

Taken together, all results above illustrated that URRCC enhanced EGFL7 expression by mediating histone H3 acetylation across EGFL7 promoter region.

Aberrant AKT signaling pathway played an important role in ccRCC and our group constantly concentrated on the studies of AKT signaling pathway [30, 32]. To investigate whether URRCC could regulate AKT signaling pathway in ccRCC, we performed WB assays and the result indicated that sh-URRCC could blunt P-AKT signaling pathway in both A498 and OSRC-2 cell lines (Fig. 5a). Concomitant with the decrease of P-AKT, we observed an increase of FOXO3 (Fig. 5a), which was reported in previous studies that downregulation of FOXO3 expression is due to increased AKT activity [33‐35]. Contrarily, oe-URRCC could enhance EGFL7/P-AKT signaling pathway in both cell lines while decrease the FOXO3 expression (Fig. 5b). Furthermore, we identified that the phenotypes caused by oe-URRCC could be partly rescued by si-EGFL7 using CCK8 and invasion assays in A498 cells (Fig. 5c and d). Similarly, the increase of EGFL7 and the decrease of FOXO3 caused by oe-URRCC also could be rescued by si-EGFL7 through WB assays in A498 cells (Fig. 5e). IHC of subcutaneous xenograft tumors showed a decrease of P-AKT expression from the sh-URRCC group compared with the sh-control group, while an increase of P-AKT expression from the oe-URRCC group compared with the mock group (Additional file 8: Figure S3I).

Altogether, our data suggested that URRCC could enhance renal cancer cell proliferation and invasion through EGFL7/P-AKT/FOXO3 signaling.

To investigated the potential mechanism for URRCC upregulation in ccRCC cells, bioinformatics analysis of the promoter region (− 1572 to − 1 bp) of URRCC was performed by online software RegRNA 2.0 (http://​regrna2.​mbc.​nctu.​edu.​tw/​detection.​html) [36]. Notably, one FOXO3 binding element (− 176 to − 163 bp) was predicted (Additional file 13: Figure S4A). Moreover, ChIP assay confirmed the enrichment of FOXO3 on the predicted region (− 176 to − 163 bp) of URRCC promoter, but not on the NC site (unpredicted, − 300 to − 280) (Fig. 6a). Meanwhile, a decrease of luciferase activity on the wild type (Wt) URRCC promoter was observed in FOXO3-overexpressed A498 cells but no change on the mutant type (Mut) URRCC promoter (Fig. 6b and c). Here, IHC also confirmed that FOXO3 expression was lower in ccRCC tissues than in normal renal tissues (Fig. 6d), which was correspond with the Kaplan–Meier survival analysis from TCGA KIRC datasets that ccRCC patients with high FOXO3 expression had longer survival times than patients with low FOXO3 levels (Log-rank test, p < 0.001, Fig. 6e, Additional file 14: Table S10). In addition, qRT-PCR assays showed there is a negative correlation between URRCC and FOXO3 in mRNA level in cell lines (Additional file 13: Figure S4B) and knockdown of FOXO3 could enhance the mRNA expression of URRCC and EGFL7, while over-expression of FOXO3 could down-regulate the mRNA expression of URRCC and EGFL7 (Additional file 13: Figure S4C-F). Given that FOXO3 affects anti-apoptotic signal, we tried to explore the effect of URRCC on cell apoptosis in A498 cells. As figures shown, knocked-down URRCC dramatically induced the apoptosis of A498 cells, while overexpressed URRCC decreased apoptosis of A498 cells (Additional file 13: Figure S4G and S4H). Overall, these data suggested that FOXO3 silenced URRCC transcription by binding to its specific promoter in ccRCC.