FOXM1-Activated LINC01094 Promotes Clear Cell Renal Cell Carcinoma Development via MicroRNA 224-5p/CHSY1

Clear cell renal cell carcinoma (ccRCC) is regarded as the most aggressive subtype of RCC, with high rates of metastasis and recurrence. An extensive body of studies had proved long noncoding RNAs (lncRNAs) play pivotal parts in the development and evolution of diverse malignant tumors. However, the potential of LINC01094 in ccRCC tumorigenesis is still unexplored. In the present research, with the aid of the TCGA database, we found that LINC01094 was highly expressed in ccRCC tissues. ABSTRACT Clear cell renal cell carcinoma (ccRCC) is regarded as the most aggressive subtype of RCC, with high rates of metastasis and recurrence. An extensive body of studies had proved long noncoding RNAs (lncRNAs) play pivotal parts in the development and evolution of diverse malignant tumors. However, the potential of LINC01094 in ccRCC tumorigenesis is still unexplored. In the present research, with the aid of the TCGA database, we found that LINC01094 was highly expressed in ccRCC tissues. Upregulation of LINC01094 was also confirmed in ccRCC cell lines, and functional experiments delineated that LINC01094 knockdown led to inhibition on ccRCC cell growth and metastasis. Moreover, LINC01094 was activated by FOXM1 at the transcriptional level. Further assay demonstrated that LINC01094 worked as a sponge of microRNA 224-5p (miR-224-5p) and CHSY1 was a miR-224-5p-targeted mRNA. Further, we verified that LINC01094 acted as a competing endogenous RNA in ccRCC to regulate CHSY1 expression via competitively bind to miR-224-5p. Lastly, our results expounded that LINC01094 exerted its tumor-promoting performance in ccRCC development through miR-224-5p/CHSY1 regulatory axis, which shed light on the molecular mechanism underlying LINC01094 in ccRCC and opened a new prospective for the treatment of ccRCC.

Chondroitin synthase 1 (CHSY1), located on chromosome 15q26.3, is one of glycosyltransferases responsible for the biosynthesis of glycosaminoglycan (24). Recent investigations have shown that CHSY1 exhibits carcinogenic activities in several cancers. For instance, CHSY1 activates the hedgehog signaling pathway to facilitate aggressive phenotypes of hepatocellular carcinoma cells (25). CHSY1 promotes proliferation and apoptosis resistance in colorectal cancer by targeting the NF-B and/or the caspase-3/7 signaling pathway (26). However, explorations on the potential of CHSY1 in ccRCC are scanty.
In this study, our purpose was to unveil the actual function and regulatory mechanism of LINC01094 in the progression of ccRCC. Our results validated that LINC01094 acted as an oncogene in ccRCC via regulation of miR-224-5p/CHSY1.

RESULTS
Silencing of LINC01094 retarded cell growth and metastasis of ccRCC. First, with the assistance of TCGA database, it was exposed that LINC01094 level was elevated in ccRCC tissues in comparison to nontumor samples (Fig. 1A). To validate the level of LINC01094 in ccRCC, we detected its expression pattern in cancerous cells. In contrast to normal renal tubular epithelial cells, upregulation of LINC01094 in ccRCC cells was demonstrated (Fig. 1B). Since ACHN and 769-P cells exhibited higher expression of LINC01094, we knocked down LINC01094 in these two cell lines by transfecting with sh-LINC01094#1/#2/#3 (Fig. 1C). This was followed by several functional experiments, and a CCK-8 assay illustrated that suppression of LINC01094 inhibited the viability of ACHN and 769-P cells (Fig. 1D). An EdU (5-ethynyl-2=-deoxyuridine) incorporation assay ascertained that depletion of LINC01094 led to the decreased percentage of EdUpositive cells, further suggesting the suppressed role of LINC01094 depletion in cell proliferation (Fig. 1E). Moreover, the results of a transwell assay showed that cell migratory capacity was restrained when LINC01094 was silenced (Fig. 1F). Consistently, knockdown of LINC01094 increased the E-cadherin level and decreased the N-cadherin level (Fig. 1G); in addition, LINC01094 insufficiency decreased the protein levels of Snail, fibronectin, MMP2, and MMP7 (Fig. 1H). To sum up, we concluded that LINC01094 downregulation hampered ccRCC progression by hindering cell proliferation, migration, and EMT process.
FOXM1 enhanced LINC01094 expression in ccRCC at the transcriptional level. Next, by browsing the UCSC website, we found that FOXM1 potentially bound to the LINC01094 promoter region ( Fig. 2A). Real-time quantitative PCR (RT-qPCR) and Western blotting showed that FOXM1 was upregulated in ccRCC cells compared to normal cells (Fig. 2B). As a result, we further explored the association between FOXM1 and LINC01094. A chromatin immunoprecipitation (ChIP) experiment illuminated that LINC01094 promoter was abundantly detected in the precipitated complex of anti-FOXM1 by RT-qPCR analysis, which reached approximately 27 to 31% compared to 1% in the IgG group (as a control), implying an interactivity of FOXM1 with the LINC01094 promoter in both ACHN and 769-P cells (Fig. 2C). After the efficiency of FOXM1 overexpression was confirmed, we found that the luciferase activity of LINC01094 promoter was fortified due to the upregulation of FOXM1 ( Fig. 2D and E). In agreement with these findings, we found that forced expression of FOXM1 increased the expression of LINC01094 (Fig. 2F). In short, these results revealed the binding relationship between FOXM1 and the LINC01094 promoter and the activation-transcription role of FOXM1in LINC01094.
LINC01094 absorbed miR-224-5p. In order to unveil the regulatory mechanism of LINC01094 in ccRCC, we utilized bioinformatics tool DIANA database to find out the putative miRNAs that could bind with LINC01094. Among the top six miRNAs with the highest predicted binding ability, only the expression of miR-224-5p was increased, owing to silencing of LINC01094, and no significant alterations occurred in the levels of other miRNAs (Fig. 3A). Previous investigations justified the tumor repressor role of miR-224-5p in several malignancies (23). In addition, miR-224-5p level was weakly expressed in ccRCC cells compared to control cells (Fig. 3B). Thus, we chose miR-224-5p as the object of subsequent researches. An RT-qPCR assay showed that miR-224-5p was overexpressed in ACHN and 769-P cells using a miR-224-5p mimic (Fig. 3C). Then, a luciferase reporter experiment showed that ectopic miR-224-5p only impaired the luciferase activity of LINC01094-WT, which unraveled the interaction of miR-224-5p to LINC01094. Meanwhile, the schematic exhibited the main components of luciferase reporter assay (Fig. 3D). Concordantly, an RNA immunoprecipitation (RIP) experiment validated that miR-224-5p and LINC01094 were enriched by Ago2 antibody in contrast to IgG antibody (Fig. 3E). Moreover, an RNA pulldown assay demonstrated great enrichment of LINC01094 in a bio-miR-224-5p-WT group (Fig. 3F). In addition, we observed that the expression of LINC01094 was remarkably lessened due to miR-224-5p upregulation (Fig. 3G). These results provide strong evidence that miR-224-5p directly interacted with LINC01094 in ccRCC.
CHSY1 was targeted via miR-224-5p. Chondroitin sulfate synthase 1 (CHSY1) was proven to work as an oncogene in colorectal cancer (26) and hepatocellular carcinoma (25), and our findings showed that the CHSY1 level in ccRCC cells was higher than in normal cells (Fig. 4A). By using the starBase database, we discovered that there were putative binding sites between CHSY1 and miR-224-5p (Fig. 4B). Hence, CHSY1 was selected for the in-depth study. A luciferase reporter experiment showed that only the CHSY1-WT luciferase activity was weakened by overexpression of miR-224-5p and that of CHSY1-Mut had no variation under miR-224-5p improvement. In addition, corresponding schematic explained the major constituents of this luciferase reporter assay (Fig. 4C). Likewise, RT-qPCR analysis and Western blotting confirmed that the CHSY1 mRNA and protein levels were diminished via upregulation of miR-224-5p ( Fig. 4D and E). Furthermore, suppression of LINC01094 attenuated CHSY1 mRNA and protein expression, and miR-224-5p inhibitor reversed this repressive function of LINC01094 knockdown ( Fig. 4F and G). Furthermore, as shown in Fig. 4H and I, overexpressed wild-type LINC01094 could enhance the expression of CHSY through regulating miR-224-5p, whereas mutant LINC01094 showed no significant influence on the level of CHSY. Furthermore, RT-qPCR measured the expression levels of miR-224-5p, LINC01094, and CHSY1 in ACHN and 769-P cells (1:14.45:52.73 and 1:15.62:55.94) (Fig. 4J). Taking the results together, LINC01094 mediated CHSY1 through sponging miR-224-5p in ccRCC.
LINC01094 contributed to the development of ccRCC by sponging miR-224-5p to regulate CHSY1 expression. Finally, rescue assays were carried out to investigate the role of LINC01094, miR-224-5p, and CHSY1 in ccRCC progression. Overexpression of CHSY1 was certified through RT-qPCR and Western blot assays after transfection in 769-P cells (Fig. 5A). The CCK-8 experiment and EdU incorporation assay demonstrated that the inhibiting influence on cell proliferation caused by LINC01094 depletion was abrogated by miR-224-5p silencing or CHSY1 overexpression ( Fig. 5B and C). The transwell assay showed that cell migration hampered by LINC01094 suppression was recovered by miR-224-5p knockdown or CHSY1 augmentation (Fig. 5D). In addition, Western blot showed that miR-224-5p inhibitor or upregulation of CHSY1 abolished the impact of LINC01094 silencing on the protein levels of E-cadherin, N-cadherin, Snail, fibronectin, MMP2, and MMP7 ( Fig. 5E and F). These results suggested that LINC01094 executed its carcinogenic function in ccRCC through regulation of miR-224-5p/CHSY1. LINC01094 promoted in vivo tumor growth and metastasis. To further research the effects of LINC01094 on ccRCC progression, we constructed an animal model to show that knockdown of LINC01094 obstructed tumor growth and metastasis in vivo (Fig. 6). These data also demonstrated that LINC01094 promoted ccRCC tumor growth and metastasis both in vitro and in vivo.

DISCUSSION
Renal cell carcinoma (RCC), the primary cause of cancer death worldwide, represents one of the most universal urological malignant neoplasms (27,28). Among the prevailing RCC histological subtypes, clear cell renal cell carcinoma (ccRCC) is the most aggressive subtype and accounts for almost 80% of RCC diagnoses (29). Although surgical excision is the curative treatment for the majority of ccRCC patients, approximately 30% of new ccRCC cases present an extremely low survival rate due to local or distant metastasis (30,31). Accordingly, it is imperative to gain a deeper understanding of the molecular mechanism underlying ccRCC and to investigate novel therapeutic strategies for the treatment of ccRCC.
In recent years, the roles of lncRNAs in various malignant tumors have drawn widespread attention. lncRNAs participate in the initiation and development of human tumors through RNA decoy, epigenetic modification, alternative splicing, and transcriptional or posttranscriptional regulation (32)(33)(34). Numerous studies have indicated that lncRNAs regulate cell processes to function as tumor boosters or repressors in a wide range of malignancies, including ccRCC (35)(36)(37). In the present study, data from the TCGA database suggested that LINC01094 was strikingly upregulated in ccRCC specimens compared to noncancer tissues. To probe the specific role of LINC01094 in ccRCC, loss-of-function assays were conducted, and our findings affirmed that depletion of LINC01094 inhibited cell proliferation, migration, metastasis, and EMT of ccRCC.
Forkhead Box M1 (FOXM1), a member of the FOX protein family, is immensely important in cell cycle progression, mitosis, and tissue homeostasis (38,39). Recent research has shown that transcription factor FOXM1 executes oncogenic activities in a variety of cancers. For example, STAT1-inhibited FOXM1 strengthens the gemcitabine sensitivity of pancreatic cancer cells (40). FOXM1 regulates KIF4A expression to promote hepatocellular carcinoma progression (41). High expression of FOXM1 is a predictor for poor clinical outcomes of prostate cancer (42). Nevertheless, the association between LINC01094 and FOXM1 in ccRCC has not been explored yet. Our study revealed that FOXM1 was present at significantly high levels in ccRCC cells compared to normal cells and that FOXM1 activated LINC01094 expression at the transcriptional level.
A growing body of research has shown that lncRNAs may interact with miRNAs to play regulatory roles in target mRNAs by serving as competing endogenous RNAs or RNA molecular sponges (43,44). In the present study, we found that LINC01094 sponged miR-224-5p and that CHSY1 was downstream of miR-224-5p. Mechanically, LINC01094 modulated CHSY1 by competitively binding to miR-224-5p. Furthermore, our results showed that LINC01094 accelerated ccRCC deterioration by targeting the miR-224-5p/CHSY1 pathway.
Here, we first investigated the exact function and latent mechanism of LINC01094 in ccRCC progression. We show that FOXM1-induced LINC01094 exhibited its oncogenic function in ccRCC through sponging miR-224-5p to mediate CHSY1 expression, which we used to explore the role of LINC01094 in ccRCC and which might provide a novel and additional therapeutic strategy for ccRCC patients.  mimic and NC inhibitor), were obtained from GenePharma (Shanghai, China). For the overexpression of CHSY1 or FOXM1, full-length CHSY1 or FOXM1 was ligated into pcDNA3.1 plasmid (Clontech Laboratories, San Francisco, CA), and empty vector served as the NC. Cells were transfected with indicated plasmids, miRNA mimic or miRNA inhibitor by using Lipofectamine 2000 (Invitrogen) as recommended by the product protocol. The experiments described above for stable overexpression, including the miR-224-5p mimic, pcDNA3.1/FOXM1, were also used for the luciferase reporter assay performed in ACHN and 769-P cells. After 24 h, pools of stably transfected cells were selected with puromycin.

MATERIALS AND METHODS
Animal model. Female BALB/c-nu mice (6 weeks old) were provided by Vital River (China). The cell lines (1 ϫ 10 6 ) with sh-LINC01094 plasmids subjected to subcutaneous injection in the right flanks of six mice in each group. We measured the tumor size every week and calculated it by using a "volume ϭ (length ϫ width 2 )/2" formula. Eventually, mice were sacrificed after 4 weeks for tumor weight detection. The animal assays were carried out with the approval of Chongming Branch, Shanghai Tenth People's Hospital, Tongji University School of Medicine.
Western blotting. After the extraction of total protein by employing radioimmunoprecipitation assay buffer containing protease inhibitor, the protein concentrations were determined using a BCA detection kit (Beyotime Institute of Biotechnology, Haimen, China) based on the producer's protocol. The protein samples were detached by SDS-10% PAGE and subsequently transferred onto polyvinylidene difluoride (PVDF) membranes. After sealing using 5% nonfat milk, the PVDF membranes were incubated overnight with primary antibodies at 4°C, probed by secondary antibodies at indoor temperature for 2 h, and detected with a chemiluminescence detection kit (Donghuan Biotech, Shanghai, China). The primary antibodies (E-cadherin, N-cadherin, CHSY1, FOXM1, Snail, fibronectin, MMP2, MMP7, and GAPDH) were procured from Cell Signaling Technology, Inc. (Danvers, MA).
Cell viability. Cell viability was evaluated with a CCK-8 detection kit (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) according to the manufacturer's protocol. A total of 3.0 ϫ 10 3 transfected cells were inoculated into each pore of 96-well dishes. After incubation for 0, 24, 48, 72, or 96 h, each pore was supplemented with 10 l of CCK-8 reagent. After an additional incubation for 2 h, the optical density at 450 nm was determined under a microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA).
EdU incorporation assay. An EdU incorporation assay was conducted to test the proliferative capability with the EdU detection kit (RiboBio, Guangzhou, China) according to the manufacturer's guidelines. After transfection, ACHN and 769-P cells (5 ϫ 10 3 cells/well) were placed in 96-well plates and maintained with 50 M EdU at 37°C for about 2 h. The cells were then immobilized by 4% paraformaldehyde for 30 min, permeated by using 0.5% Triton X-100 for 20 min, treated with Apollo Staining reaction reagent for 30 min, and finally stained by DAPI (4=,6=-diamidino-2-phenylindole) for nearly 10 min. Cells were visualized with an Olympus ZKX53 microscope (Olympus, Tokyo, Japan).
Cell migration. Transwell assay was applied to estimate cell migration potential with the transwell chambers (Corning, Cambridge, MA) in light of the product manual. ACHN and 769-P cells in 200 l of serum-free medium were placed in the upper chambers, and in the lower chambers 600 l of medium with plus 10% FBS was added. After 24 h of incubation at 37°C, the cells on the upper surface were discarded, and migrated cells on the bottom were immobilized in 4% formaldehyde, stained by 0.1% crystal violet, and observed by using an Olympus ZKX53 microscope. Next, the cells were counted in five randomly selected fields, and each assay was repeated three times.
Luciferase reporter assay. The full-length of LINC01094/FOXM1 containing wild-type or mutant miR-224-5p/LINC01094 promoter binding sites was cloned to pGL3 luciferase vector (Promega, Madison, WI). Similarly, to construct wild-type or mutant CHSY1 plasmids, 3=-UTR sequences with or without miR-224-5p binding sites were ligated into the pGL3 vector. ACHN and 769-P cells were cotreated with corresponding reporter plasmids and miR-224-5p mimic or NC mimic by utilizing Lipofectamine 2000 (Invitrogen). At 48 h posttransfection, the luciferase activity was tested with a dual-luciferase reporter assay system (Promega) based on the vendor's recommendations.
RIP assay. The RIP assay was carried out with a EZ-Magna RIP kit (Millipore, Billerica, MA) according to the product specifications. In short, harvested cells were first lysed with complete RIP lysis buffer and then treated with RIP buffer and magnetic beads coated with antibody targeting Ago2 or IgG (Millipore). Thereafter, the magnetic beads were cultured with proteinase K, and the precipitated RNA was eluted and purified for RT-qPCR.
Chromatin immunoprecipitation. Once they reached 90% confluence, the cells were cross-linked using 1% formaldehyde, and chromatins were sheared into fragments with a Bioruptor sonicator (Diagenode, Denville, NJ). After preclearing with protein A/G beads, the samples were precipitated by using a specific FOXM1 antibody at 4°C overnight. IgG antibody was used as a negative control. After washing, the expression of LINC01094 promoter in immunoprecipitate complexes was determined by RT-qPCR.
Immunohistochemical staining. Paraffin embedded ccRCC samples was sliced and rehydrated through a xylene and ethanol series, followed by hematoxylin (Gill's, 1ϫ) staining for 5 min. Slides were washed for 5 min and soaked in acid alcohol (1% HCl in 70% ethanol) two or three times until the sections became pink. The slides were then washed in water again for 3 to 5 min and slowly soaked in ammonia water (1 liter of H 2 O with 1 ml of NH 4 OH) five or six times. The slides were washed again in water, which lasted 3 to 5 min. An eosin Y solution was then used to counterstain the slides for 1 min. The samples were dehydrated with an ethanol series and washed with a xylene series. Cover slides were installed using ProLong Gold (Invitrogen, Carlsbad, CA) and held for one night at room temperature.
Statistical analysis. SPSS 17.0 software (SPSS, Chicago, IL) was used for all statistical analyses. Each experiment was assayed in triplicate, and all results are displayed as means Ϯ the standard deviations (SD). Comparisons between two groups were made using a Student t test, and comparison among three or more groups were estimated using one-way or two-way analysis of variance (ANOVA). A P value of Ͻ0.05 denoted a statistically significant difference.