The Oncogenic Capacities of Long Noncoding RNA SNHG1 in Bladder Cancer by inducing proliferation and repressing apoptosis via upregulation of microRNA-9-3p-targeted MDM2


 Background

The involvement of long noncoding RNA small nucleolar RNA host gene 1 (lncRNA SNHG1) was documented in numerous cancers, including bladder, pancreatic and prostate cancers. However, the further mechanistic investigation of SNHG1 in bladder is still needed to conduct. With this purpose, tissue, cell, and animal experiments were implemented in our research to figure out the specific mechanism of SNHG1 in bladder cancer via microRNA-9-3p (miR-9-3p).
Methods

In harvested bladder cancer tissues, RNA-FISH and RT-qPCR were adopted for SNHG1 expression measurement and RT-qPCR for miR-9-3p expression determination. The impacts of SNHG1, miR-9-3p, MDM2, and PPARγ on cell viability, proliferation, and apoptosis were evaluated by gain- and loss-of-function approaches. RT-qPCR and western blot analysis were performed to detect expression of MDM2, PPARγ, and apoptosis-related factors. RNA pull-down, RIP, dual luciferase reporter gene assay, and IP experiment were utilized to assess the modulatory relationship among SNHG1, miR-9-3p, MDM2, and PPARγ. Tumorigenic ability of bladder cancer cells was measured in vivo.
Results

High SNHG1 and poor miR-9-3p expression was identified in bladder cancer tissues and cells. Mechanistically, SNHG1 bound to miR-9-3p which negatively targeted MDM2. MDM2 augmented PPARγ ubiquitination to downregulate PPARγ. Bladder cancer cell proliferation was diminished and cell apoptosis was enhanced by silencing SNHG1 or MDM2 or overexpressing miR-9-3p. Similarly, SNHG1 silencing orchestrated miR-9-3p/MDM2/PPARγ axis to depress bladder cancer cell tumorigenesis in vivo.
Conclusion

In summary, the obtained data provided the novel insight of the anti-oncogenic mechanism of silencing SNHG1 in bladder cancer by activating PPARγ via downregulation of miR-9-3p-targeted MDM2.


Conclusion
In summary, the obtained data provided the novel insight of the anti-oncogenic mechanism of silencing SNHG1 in bladder cancer by activating PPARγ via downregulation of miR-9-3p-targeted MDM2.

Background
Bladder cancer (BC) ranks 9th on the list of cancers in terms of incidence with nearly 430000 cases of annual incidence, and ranks 13th among all cancers in yearly mortality across the world [1]. A strong male predominance is observed in bladder cancer, where three-fourths of cases occur in men [2]. The increased risk for bladder cancer correlates to factors including age, smoking, and exposure to some industrial chemicals [3]. The treatment of bladder cancer depends on stage and grade to a great extent, which also strongly correlates to the prognosis of patients: the treatment of nonmuscle invasive bladder cancer is usually through resection and immunotherapy with intravesical drugs like Bacillus Calmette-Guerin, whilst more aggressive methods, like radical cystectomy coupled with chemotherapy, are needed for muscle invasive bladder cancer [4]. Unfortunately, bladder cancer is diagnosed generally the terminal stage, particularly in women, and there is little improvement in the treatment for bladder cancer with at 5-year survival rate until recently [5]. Therefore, it is urgent to get a deeper understanding of molecular mechanism underlying bladder cancer, thus exploring a novel targets for bladder cancer treatment.
It is well-established that long noncoding RNA small nucleolar RNA host gene 1 (lncRNA SNHG1) is involved in advanced tumor node metastasis (TNM) and tumor stage and size, and reduced overall survival [6]. The oncogenic role of SNHG1 has been elucidated in various cancers. For instance, a prior study also reported the tumor-promoting potential of SNHG1 in pancreatic cancer with the results that SNHG1 silencing triggered repression of cell proliferative, metastatic, and invasive capacities by inactivating Notch-1 pathway [7]. In addition, Li et al. observed the suppressive effect of SNHG1 on prostate cancer development by promoting cell proliferation [8]. These evidences indirectly supported that SNHG1 might promote development of bladder cancer. Moreover, starbase website used in our study predicted the binding relationship between SNHG1 and microRNA-9-3p (miR-9-3p).
miR-9-3p (previously known as miR-9) is widely known for its altered expression and function in multiple diseases, like Huntington's disease and cancers [9]. Interestingly, it was detected that ectopically expressed miR-9-3p possessed antitumor potential in bladder cancer by diminishing cell invasion, migration, and proliferation [10]. In our study, the binding sites between miR-9-3p and 3' untranslated region (UTR) of murine double minute 2 (MDM2) were predicted by TargetScan. MDM2 overexpression could reportedly neutralize the deressive effect of miR-379-5p on bladder cancer cell proliferative, migratory and invasive capacities [11]. In the presence of EGFR, MDM2 can bind to peroxisome proliferator-activated receptor-gamma (PPARγ) and regulate the ubiquitination of PPARγ protein in colon cancer cells [12]. More importantly, it was elaborated in a prior study that antagonist of PPARγ promoted cell cycle entry and decreased cell apoptosis in bladder cancer [13].
Taken these evidences into account, we hypothesized that SNHG1/miR-9-3p/MDM2/PPARγ axis correlated to the progression of bladder cancer. Therefore, the present study was implemented by focusing on the alteration in SNHG1 expression in bladder cancer tissues and cells and investigated the function and underlying mechanism of SNHG1 in bladder cancer progression via miR-9-3p/MDM2/PPARγ axis.

Compliance with ethical standards
The experiments involving human beings were implemented with rati cation of Ethics Committee of Suqian First Hospital by conforming to the principles outlined in the Declaration of Helsinki. Ethical agreements were obtained from the donors or their families through written informed consent. Animal experiments were rati ed by Animal Ethics Committee of Suqian First Hospital and concurred with the Guidelines for Animal Experiments of Peking University Health Science Center.

Study subject
Forty patients with bladder cancer undergoing radical cystectomy in Suqian First Hospital from June 2016 to June 2017 were enrolled. Fresh bladder cancer tissues and corresponding adjacent normal tissues were preserved in liquid nitrogen immediately subsequent to resection. None of patients received preoperative radiotherapy, chemotherapy or immunotherapy. Follow-up information was obtained from outpatient clinics and regular telephone interviews.

Bioinformatics methods
Gene Expression Pro ling Interactive Analysis (GEPIA) was adopted to analyze the BLCA dataset of The Cancer Genome Atlas (TCGA) database to obtain the genes with signi cant differences (p < 0.05), from which the genes with |logFC| > 0.5 were screened out. The Gene Expression Omnibus (GEO) database (https://www.ncbi.nlm.nih.gov/gds) was also analyzed by using "limma" package (http://www.bioconductor.org/packages/release/bioc/html/limma.html) of the R language with |logFC| > 0.5 and p < 0.05 as thresholds for differential analysis of bladder cancer microarray data GSE65635 and GSE40355. There were 12 samples in microarray data GSE65635, including 4 normal samples and 8 bladder cancer samples. There were 24 samples in microarray data GSE40355, including 8 normal samples and 16 bladder cancer samples. Human lncRNA names were obtained from GENCODE, followed by obtaining of the intersection of signi cantly differential genes and lncRNA names. Venn diagram was drawn to screen out the lncRNAs among intersection. LncRNA expression trends were collected in Ualcan, and the key lncRNA was determined by comparing the expression trends and combining with the existing literature. The possible downstream miR of the key lncRNA was discovered by starBase and their binding sites were obtained. The databases TargetScan (Cumulative weighted context++ score < 0), DIANA TOOLS (miTG score > 0.6), microRNA (conservation > 0.65, energy < -14, Mirsvr_score < -0.65), and mirDIP (Integrated Score > 0.1) was applied to predict downstream genes of miR. Intersection of downstream genes with signi cantly differential genes was taken to obtain critical downstream gene. The relevant genes of the critical downstream gene were predicted in GeneMANIA (http://genemania.org/), followed by construction of protein-protein interaction (PPI) network. The most core genes in PPI network were chosen as the key gene, and the binding sites of the miR to the gene were predicted by TargetScan.
Fluorescence in situ hybridization (FISH) SNHG1 cDNA fragments were ampli ed from the SNHG1 plasmid as templates by utilizing high delity DNA polymerase (Takara, Kyoto, Japan). Based on this template, uorescein-labeled lncTCF7 FISH probe DNA was prepared with uorescein-12-dUTP (Roche, Mannheim, Germany) and Klenow DNA polymerase as per the manufacturer's protocol. Four-µm frozen sections were made from bladder cancer tissues and adjacent normal tissues. Subsequent to 5-min immersion in proteinase K, the slides were washed twice in Page 5/24 probe DNA (Beijing Dingguo Changsheng Biotechnology Co., Ltd., Beijing, China) was dripped onto the tissue sections before 16-h incubation at 37℃. The slides were then washed in 0.4 × SSC/0.001% NP-40 for 5 min at 56℃, followed by another 2-min washing in 0.4 × SSC/0.001% NP-40. After being dripped with 4',6-Diamidino-2-Phenylindole (DAPI)-encompassing sealing agent, the slide was mounted and observed under a uorescence microscope (Olympus, Tokyo, Japan).   with 10% FBS. All media for cell lines encompassed 100 μg/mL streptomycin and 100 U/mL penicillin. Cell culture was performed at 37℃ with 5% CO 2 . The media were positioned in humid air and replaced every 2-3 days according to the growth of cells. Cells were subcultured when 80%-90% of the culture plate was covered by cells. Cells were utilized when they reached the logarithmic growth stage.
Reverse transcription quantitative polymerase chain reaction (RT-qPCR) Subsequent to isolation using RNeasy Mini Kit (Qiagen, Valencia, CA, USA), total RNA underwent reverse transcription to generate cDNA using First Strand cDNA Synthesis Kit (RR047A, Takara). For the detection of miR, the cDNA was obtained by reverse transcription using the miRNA First Strand cDNA Synthesis (Tailing Reaction) kit (B532451-0020, Sangon, Shanghai, China). RT-qPCR reactions were performed using SYBR® Premix Ex TaqTM II (Perfect Real Time) kit (DRR081, Takara) on real-time uorescence quantitative PCR instrument (ABI 7500, Applied Biosystems, Foster City, CA, USA). The universal reverse primers for miR and the upstream primers for U6 internal reference were provided in the miRNA First Strand cDNA Synthesis (Tailing Reaction) kit, and the other primers were synthesized by Sangon. (Table  3). After recording of the Ct value of each well, the relative expression of mRNAs or miR was calculated using the 2 -ΔΔCt method by normalizing to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or U6 expression.

5-ethynyl-2'-deoxyuridine (EdU) assay
The cells to be tested were seeded in 24-well plates with three duplicated wells set for cells in each group.
EdU (Invitrogen) was supplemented to the medium to achieve a concentration of 10 µmol/L. The medium was discarded subsequent to 2-h culture. Cells received 15-min phosphate buffer saline (PBS) encompassing 4% paraformaldehyde xing at ambient temperature before 20-min incubation at ambient temperature with PBS encompassing 0.5% Triton-100. Each well was supplemented with 100 µL dye solution before 30-min culture in the dark at ambient temperature. DAPI was added for 5-min nuclear staining. After sealing, 6-10 elds of view were randomly observed under a uorescence microscope (FM-600, Shanghai Pudan Optical Instrument Co., Ltd., Shanghai, China), and the number of positive cells in each eld was recorded.
Flow cytometry Subsequent to 48-h transfection, the cell concentration was changed to 1 × 10 6 cells/mL. Subsequent to cell xing with 70% precooled ethanol solution at 4℃, 100 μL cell suspension (no less than 1 × 10 6 cells/mL) was resuspended in 200 μL binding buffer. Subsequently, 15-min cells staining were implemented with 10 μL Annexin V-uoresceinisothiocyanat and 5 μL propidium iodide at ambient temperature under dark conditions. After 300 μL of binding buffer was added, apoptosis was assessed on a ow cytometer at excitation wavelength of 488 nm (2 × 10 4 cells each time).

RNA pull down
Cells were transfected with biotinylated wild type (WT) miR-9-3p and mutant type (MUT) miR-9-3p (50 nM each). After 48 h of transfection, 10-min cell incubation was implemented with speci c cell lysis (Ambion, Austin, Texas, USA). Then, 3-h lysate incubation was conducted with M-280 streptavidin magnetic beads (Sigma, St. Louis, MO, USA) pre-coated with RNase-free and yeast tRNA at 4℃ before two cell washes in cold lysis and RT-qPCR detection of SNHG1 expression.

RNA immunoprecipitation (RIP) assay
The binding of miR-9-3p to MDM2 was detected by RIP kit (Millipore, Temecula, CA, USA). Brie y, 5-min cell lysing was implemented in an ice bath with equal volume of RIPA lysis (P0013B, Beyotime, Shanghai, China), and supernatant was removed subsequent to 10-min centrifugation at 14000 rpm and 4℃. A portion of the cell extract was applied as input, and a portion was co-precipitated with antibody. RNA exaction was implemented by treating samples with proteinase K for subsequent RT-qPCR detection of MDM2. Antibodies used for RIP were as follows: rabbit anti-Argonaute 2 (AGO2) (1: 100, ab32381, Abcam) was mixed at ambient temperature for 30 min, and rabbit anti-human Immunoglobulin G (IgG; 1: 100, ab109489, Abcam) was applied as a NC.

Immunoprecipitation (IP)
Cells were lysed in lysis buffer [mixture of 50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 10% glycerol, 1 mM ethylene diamine tetraacetic acid, 0.5% NP-40, and protease inhibitor], and cell debris was cleared by centrifugation. After the concentration of lysis was measured by BCA, the same amount of protein was taken from each experimental group and replenished to the same volume with cell lysate. Afterwards, 1 μg anti-MDM2 (ab226939, 1: 100, Abcam), PPARγ (ab45036, 1: 100, Abcam) and 15 μL protein A/G beads (Santa Cruz Biotechnology, Santa Cruz, CA, USA) were added for 2-h incubation. Subsequent to three washes with cell lysis, beads were collected by centrifugation, added into an equal volume of reductive loading buffer, and boiled at 100℃ for 5 min. Subsequent to SDS-PAGE, samples were electroblotted to PVDF membranes (Millipore), and then analyzed by immunoblotting.

Subcutaneous tumorigenesis model in nude mice
Healthy nude mice aged 6-8 weeks (Beijing Institute of Pharmacology, Chinese Academy of Medical Sciences, Beijing, China) were bred in speci c pathogen-free animal laboratory with 60%-65% humidity at 22-25℃. They were fed in separate cages under 12:12-h light-dark cycle with food and water available ad libitum. The experiment was started one week after acclimation, and the health status of nude mice was observed before the experiment. Approximately 2 × 10 6 cells were suspended in 200 μL PBS, and then subcutaneously injected into the left or right hindlimbs of nude mice (10 mice/group). At 28 days subsequent to injection, mice were euthanized, followed by measurement and weighing of tumors.

Statistical analysis
All measurement data were manifested as mean ± standard deviation and analyzed by SPSS 21.0 software (IBM, Armonk, NY, USA), with p < 0.05 as a level of statistically signi cance. If data conformed to normal distribution and homogeneity of variance, data within groups were compared by paired t test, while data between two groups were compared by unpaired t test. Comparisons among multiple groups were performed using one-way analysis of variance (ANOVA) or repeated measures ANOVA. Intra-group pairwise comparison was performed using post-hoc test. Rank sum test was performed if data did not conform to normal distribution or homogeneity of variance. Kaplan-Meier was adopted to calculate patient survival curves, and log-rank was utilized to analyze patient survival differences.

Results
SNHG1 was highly expressed in bladder cancer tissues and associated with poor prognosis of patients with bladder cancer The BLCA data of TCGA database were analyzed by GEPIA to obtain 6597 signi cantly differential genes (|logFC| > 0.5, p < 0.05) (Fig. 1A). Then, R language was employed for difference analysis of microarray data GSE65635 and GSE40355 in GEO database to obtain 4283 and 8065 signi cantly differential genes, respectively (|logFC| > 0.5, p < 0.05). Then, 17937 human lncRNA names were obtained from GENCODE, which were intersected with differential genes. It was found that only DIO3OS and SNHG1 were signi cantly differential lncRNAs in bladder cancer (Fig. 1B). Analysis of TCGA database data by Ualcan revealed that DIO3OS was signi cantly underexpressed in bladder cancer (p = 1.949E-05; Fig. 1C), while SNHG1 was signi cantly overexpressed in bladder cancer (p = 3.889E-11; Fig. 1D). Moreover, the difference of SNHG1 was signi cantly higher than that of DIO3OS. There was literature indicating the upregulation of SNHG1 in thyroid cancer [14], non-small cell lung cancer (NSCLC) [15], colorectal cancer [16,17], but SNHG1 was not studied in bladder cancer. Next, RNA-FISH showed high SNHG1 expression in bladder cancer tissues compared with adjacent normal tissues (Fig. 1E). Further detection by RT-qPCR assay also found that SNHG1 was highly expressed in bladder cancer tissues (Fig. 1F), and that patients with high expression of SNHG1 had worse prognosis (Fig. 1G). Therefore, high SNHG1 expression was associated with poor prognosis of patients with bladder cancer.

SNHG1 was highly expressed in bladder cancer cells and promotes bladder cancer cell proliferation
To further examine the regulatory role of SNHG1 in bladder cancer, we selected one normal urothelial cell line SV-HUC-1 and four bladder cancer cell lines (5637, T24, SW780, and UM-UC-3). As described in Fig.   2A, SNHG1 expression was increased in cancer cells compared with SV-HUC-1 cells. Subsequent experiments were conducted on T24 cells with higher SNHG1 expression and 5637 cells with lower SNHG1 expression. After silencing SNHG1 in T24 cell (Fig. 2B), the shRNA with the highest silencing e ciency was selected for subsequent experiments. CCK-8 and EdU assays exhibited that the viability and proliferation of T24 cells were obviously inhibited by treatment with sh-SNHG1 (Fig. 2C-D). The apoptotic rate of T24 cells elevated signi cantly (Fig. 2E), accompanied by prominent increase of Cleaved caspase-3 and Bax expression and remarkable decline of Bcl-2 expression (Fig. 2F) after silencing SNHG1. It suggested that silencing SNHG1 could trigger the inhibition of cell proliferation and promotion of cell apoptosis in bladder cancer. Further experiments in 5637 cells manifested that overexpression of SNHG1 in 5637 cells (Fig. 3A) noteworthy enhanced viability (Fig. 3B) and proliferation ( Fig. 3C), diminished apoptosis (Fig. 3D), and reduced the expression of Cleaved caspase-3 and Bax but elevated Bcl-2 expression (Fig. 3E). Collectively, SNHG1 upregulation promoted the proliferation of bladder cancer cells.

SNHG1 promotes bladder cancer cell tumorigenesis in Vivo
Further, a subcutaneous tumorigenic model was established in nude mice to detect the tumorigenic ability of bladder cancer cells in vivo. RT-qPCR depicted that SNHG1 expression was appreciably decreased in mice treated with sh-SNHG1 (Fig. 4A). In addition, the growth rate and weight of tumors were appreciably decreased after silencing SNHG1 (Fig. 4B). FISH experiment illustrated that SNHG1silenced mice had distinct decline of SNHG1 expression (Fig. 4C). On the contrary, overexpression of SNHG1 (Fig. 4D) contributed to the appreciable elevation of the growth rate and weight of tumors (Fig.  4E) and SNHG1 expression in tumors (Fig. 4F). In summary, SNHG1 overexpression promoted bladder cancer cell tumorigenesis in vivo.

SNHG1 silencing suppressed bladder cancer cell proliferation and tumorigenesis by binding to MiR-9-3p
Then, we explored the downstream miR of SNHG1 in bladder cancer. RNA-FISH (Fig. 1B) showed that SNHG1 was localized in the cytoplasm, suggesting that SNHG1 may be involved in the process of bladder cancer by affecting miR. The Starbase website predicted that SNHG1 could bind to miR-9-3p (Fig.   5A). A previous study has reported that miR-9-3p expression is poor in bladder cancer [10], but the related regulatory mechanisms need further study. RT-qPCR revealed that miR-9-3p expression was signi cantly low in bladder cancer tissues (Fig. 5B) and had signi cantly inverse correlation with the expression of SNHG1 (Fig. 5C). Meanwhile, it was veri ed by RNA pull-down that SNHG1 indeed bound to miR-9-3p (Fig.  5D). In addition, silencing SNHG1 in T24 cells prominently increased miR-9-3p expression (Fig. 5E), while overexpressing SNHG1 in 5637 cells severely declined miR-9-3p expression (Fig. 5F).
The effect of SNHG1 binding to miR-9-3p on bladder cancer cells was further examined. Silencing SNHG1 alone resulted in decrease of SNHG1 expression. Besides, silencing SNHG1 alone reduced cell proliferation, increased apoptotic rate, elevated Cleaved caspase-3 and Bax expression, and declined Bcl-2 expression in T24 cells, which was opposite after treatment with miR-9-3p inhibitor alone. However, cotreatment with sh-SNHG1 and miR-9-3p inhibitor reversed the effect of sh-SNHG1 or miR-9-3p inhibitor alone (Fig. 6A-E). Simultaneously, in vivo experiments showed that the tumorigenic ability of bladder cancer cells in vivo was diminished by treatment with sh-SNHG1 alone and elevated by treatment with miR-9-3p inhibitor alone, which was neutralized by co-treatment with sh-SNHG1 and miR-9-3p inhibitor (Fig. 6F, G). Conclusively, silencing SNHG1 bound to miR-9-3p to inhibit bladder cancer cell proliferation and tumorigenesis.
Silencing SNHG1 decreased MDM2 expression through MiR-9-3p Subsequently, the downstream target genes of miR-9-3p were investigated. The 3615, 2399, 270 and 2387 downstream genes of miR-9-3p were respectively predicted in TargetScan, DIANATOOLS, microRNA and mirDIP, and then were intersected. The intersecting results were compared with the differential genes in bladder cancer obtained by GEPIA, which screened out 21 signi cantly differential downstream genes of miR-9-3p (Fig. 7A). By constructing the PPI network through GeneMANIA, we found that MDM2 had the highest core degree in the PPI network and was double the core degree of second-ranked genes (Fig. 7B, Table 1). The binding site of miR-9-3p in MDM2 3'UTR was predicted using TargetScan (Fig. 7C). Nevertheless, previous studies detected that MDM2 was highly expressed as a proto-oncogene in bladder cancer [11,18]. miR-9-3p may be involved in the progression of bladder cancer by inhibiting the expression of MDM2, which was further veri ed by experiments. Firstly, AGO2 pulled down MDM2 in RIP experiments (Fig. 7D). Dual luciferase reporter assay manifested that miR-9-3p mimic markedly inhibited the luciferase activity of MDM2 WT, but had no obvious effect on the luciferase activity of MDM2 MUT, suggesting that miR-9-3p bound to the 3'UTR of MDM2 (Fig. 7E).
After sh-SNHG1 and miR-9-3p inhibitor were co-transfected into T24 cells, MDM2 expression was evaluated by RT-qPCR and western blot analysis. As displayed in Fig. 7F, G, MDM2 expression was noticeably decreased after silencing SNHG1 alone, and observably increased after transfection with miR-9-3p inhibitor alone, which was normalized by co-treatment with sh-SNHG1 and miR-9-3p inhibitor. It was suggested that silencing of SNHG1 inhibited MDM2 expression through binding to miR-9-3p.

SNHG1 silencing led to suppression of proliferation and tumorigenesis of bladder cancer cells by downregulating MDM2
Further, the expression of MDM2 was overexpressed in SNHG1-silenced T24 cells. From RT-qPCR results, the expression of SNHG1 and MDM2 was substantially diminished and miR-9-3p expression was signi cantly enhanced after silencing SNHG1 alone. MDM2 expression was appreciably increased after overexpressing MDM2 alone, and silencing SNHG1 negated the effect of overexpressing MDM2 (Fig. 7H). As described in Fig. 7I, J, CCK-8 and EdU assays exhibited that silencing SNHG1 appreciably reduced but overexpressing MDM2 increased cell proliferation, and overexpressing MDM2 could reverse the effect of silencing SNHG1 on cell proliferation (Fig. 7I-J). To sum up, SNHG1 silencing suppressed the proliferation of bladder cancer cells by decreasing MDM2 expression through miR-9-3p.
Further in vivo validation was carried out. The expression of SNHG1, miR-9-3p, and MDM2 was detected by RT-qPCR. As presented in results, SNHG1 and MDM2 expression was strikingly declined and miR-9-3p expression was remarkably elevated after silencing SNHG1. MDM2 expression was prominently promoted after overexpressing MDM2, and overexpressing MDM2 could reverse the effect of silencing SNHG1 on MDM2 expression (Fig. 7K). Furthermore, after silencing SNHG1 alone, tumor growth was repressed, and after overexpressing MDM2 alone, tumor growth was promoted. Overexpression of MDM2 abrogated the effect of silencing SNHG1 on tumor growth (Fig. 7L). In summary, silencing SNHG1 decreased the tumorigenesis of bladder cancer cells in vivo by decreasing MDM2 expression through miR-9-3p.

Silencing SNHG1 upregulated PPARγ through MDM2
It has been documented that after addition of EGFR, MDM2 can bind to PPARγ and regulate the ubiquitination of PPARγ protein in colon cancer, and that MDM2 silencing can increase the level of PPARγ [12], which is further veri ed in bladder cancer. Firstly, through IP experiments in T24 cells, it was found that MDM2 and PPARγ combined with each other (Fig. 8A). Further, after screening out the MDM2 silencing sequence (Fig. 8B, sh-MDM2-3 was selected for subsequent experiments), we found that PPARγ ubiquitination decreased and PPARγ expression increased after silencing MDM2 with the addition of EGFR (Fig. 8C), which was opposite after overexpressing MDM2 (Fig. 8D). MDM2 was overexpressed after silencing SNHG1 in T24 cells with the addition of EGFR. Western blot analysis exhibited that MDM2 expression was prominently decreased and PPARγ expression was severely elevated after silencing SNHG1 alone, which was opposite after overexpressing MDM2 alone. Silencing SNHG1 annulled the effect of overexpressing MDM2 on MDM2 and PPARγ expression (Fig. 8E). The above results suggested that MDM2 reduced PPARγ expression by inducing PPARγ ubiquitination, while silencing of SNHG1 elevated PPARγ expression through downregulating MDM2.

Discussion
As one of the most prevalent genitourinary cancers with high mortality on a global scale, bladder cancer currently can be treated with appended by local or systemic immunotherapy, radiotherapy, chemotherapy, and endoscopic and open surgery [19]. However, the curative effect of such therapies is limited because of recurrence or distant spread [20]. Moreover, lncRNAs has been emerged as a modulator in the complexity of bladder cancer [21]. Consequently, this research was intended to rule of the mechanism of SNHG1 in bladder cancer with the involvement of miR-9-3p. Notably, the present study provided evidence that SNHG1 promotes MDM2 expression by binding to miR-9-3p to promote PPARγ ubiquitination and downregulate PPARγ expression, thereby resulting in elevation of bladder cancer cell proliferation in vitro and tumorigenesis in vivo.
Initially, data from our study unraveled that SNHG1 was highly expressed in bladder cancer tissues and cells. Additionally, when SNHG1 was silenced in bladder cancer cells and mice, cell proliferative capacity was depressed but cell apoptosis was accelerated in vitro, and tumorigenesis was inhibited in vivo. Importantly, SNHG1 has emerged as a novel oncogenic lncRNA in various cancers, including esophageal, colorectal, prostate, gastric, liver, and lung cancers by inducing cell proliferative, metastatic, migratory and invasive capacities of cancer cells [6]. Consistently, Lu et al. observed that SNHG1 expression was strikingly high in NSCLC tissues and cells, and that SNHG1 silencing decreased tumor volumes in mice and reduced NSCLC cell proliferation, invasion and migration [22]. Meanwhile, data collected by Bai et al. found that SNHG1 expression was upregulated in colorectal cancer cells, and that ectopically expressed SNHG1 could enhance cell migratory, proliferative, and invasive capacities in vitro and led to tumor growth [23,24]. Another study uncovered that SNHG1 silencing contributed to decline of tumor growth of breast cancer in vivo [25]. These ndings indirectly supported the tumor-promoting potential of SNHG1 in bladder cancer by enhancing cell proliferation and tumor growth and reducing apoptosis.
It is well-recognized that lncRNAs may function as endogenous sponges to regulate miRNA function in diseases [26]. For example, a prior research displayed that SNHG1 could bind to miR-204 to inhibit it, thus promoting migratory, and invasive abilities but repressing apoptosis in esophageal squamous cell cancer [27]. These ndings indirectly con rmed the binding relationship between SNHG1 and miR-9-3p in bladder cancer cells which observed by our study. Further investigations of our study identi ed that miR-9-3p inhibition led to increase of cell proliferation and decrease of apoptosis in vitro and promotion of tumorigenesis in vivo and reversed the effect of SNHG1 silencing in bladder cancer. Similarly, a research conducted by Cai et al. clari ed that in bladder cancer, miR-9-3p overexpression triggered repression of cell viability, migration, and invasion, induction of cell apoptosis in vitro, and inhibition in vivo tumor growth and metastasis [10]. Notably, another study illustrated that miR-9-3p exerted tumor-suppressive effect on hepatocellular carcinoma by depressing hepatocellular carcinoma cell proliferation [28], which was in line with our results. Hence, these results con rmed that SNHG1 overexpression promoted bladder cancer progression by binding to miR-9-3p.
It is well-established that miRs inhibit expression of target genes at post-transcriptional level by targeting the 3'UTR of mRNA [29]. As previously reported, miR-9-3p targeted HBGF-5 to function as a tumor suppressor in hepatocellular carcinoma [30]. Moreover, in our study, TargetScan website predicted the binding sites between miR-9-3p and MDM2 3'UTR, and then the targeting relationship between miR-9-3p and MDM2 was veri ed by dual luciferase reporter gene assay. In the subsequent experiments, we found that MDM2 bound to PPARγ and downregulated PPARγ by inducing PPARγ ubiquitination, which was similar to the results observed by Xu et al. [12]. Furthermore, our data elaborated that MDM2 ectopic expression neutralized the inhibitory effect of SNHG1 silencing on cell proliferation in vitro and tumor growth in vivo and the promoting effect of SNHG1 silencing on cell apoptosis in vitro in bladder cancer, suggesting the oncogenic role of MDM2 in bladder cancer. Consistently, a prior study uncovered that inhibition of MDM2 exerted tumor-suppressive effects on bladder cancer by decreasing cell invasive, proliferative, and migratory capacities [11]. Accordingly, PPARγ activation gave rise to inhibition of proliferation of 9 bladder cancer cell lines [31] All in all, the SNHG1/miR-9-3p/MDM2/PPARγ axis was involved in bladder cancer progression.

Conclusion
Collectively, this study provides evidence that SNHG1 upregulation promoted cell proliferation but depressed cell apoptosis in bladder cancer via MDM2-inhibited PPARγ by binding to miR-9-3p (Fig. 9). Thus, this nding offers a fresh molecular insight that might be utilized in new therapy development for bladder cancer. However, further studies are prerequisites on the mechanism of PPARγ in bladder cancer.  Tables   Table 1 sh-RNA sequences for transfection shRNAs Sequences sh-NC 5′-TTCTCCGAACGTGTCACGTTT-3′   Bcl-2 in T24 cells. * p < 0.05, ** p < 0.01. The experiment was repeated three times. Data were presented as mean ± standard deviation and compared by one-way analysis of variance (ANOVA), followed by Tukey's post hoc test.  Data were presented as mean ± standard deviation and compared by t test or repeated measures ANOVA with Tukey's post hoc test. SNHG1 binds to and downregulates miR-9-3p. A, Starbase (http://starbase.sysu.edu.cn/) website predicting that SNHG1 bound to miR-9-3p. B, RT-qPCR detecting the expression of miR-9-3p in 40 bladder cancer and adjacent normal tissues. C, Correlation analysis of the expression of SNHG1 and miR-9-3p. D, RNA pull-down assay detecting the binding relationship between SNHG1 and miR-9-3p. E, RT-qPCR detection of the expression of miR-9-3p after silencing SNHG1 in T24 cells. F, RT-qPCR detection of the expression of miR-9-3p after overexpressing SNHG1 in 5637 cells. * p < 0.05. The experiment was repeated three times. Data were presented as mean ± standard deviation and compared by t test or oneway analysis of variance (ANOVA), followed by Tukey's post hoc test.

Figure 6
SNHG1 silencing binds to miR-9-3p to repress bladder cancer cell proliferation and tumorigenesis. T24 cells were transfected with sh-NC + inhibitor NC, sh-SNHG1 + inhibitor NC, sh-NC + miR-9-3p inhibitor, and sh-SNHG1 + miR-9-3p inhibitor. A, SNHG1 and miR-9-3p expression in T24 cells measured by RT-qPCR. B, CCK-8 assay of the change of T24 cell viability. C, The changes of T24 cell proliferation detected by EdU assay. D, Flow cytometry analysis of the changes of T24 cell apoptosis. E, Western blot analysis of the expression of Cleaved caspase-3, Bax, and Bcl-2 in T24 cells. The stably transfected T24 cells were subcutaneously injected into the axilla of nude mice. F, SNHG1 and miR-9-3p expression in mice measured by RT-qPCR. G, Tumor weight. * p < 0.05. N = 10 mice in each group. The cell experiment was repeated three times. Data were presented as mean ± standard deviation and compared by one-way analysis of variance or repeated measures ANOVA with Tukey's post hoc test.   Mechanism. SNHG1 promotes MDM2 expression by binding to miR-9-3p to promote PPARγ ubiquitination and downregulate PPARγ expression, thereby resulting in elevation of bladder cancer cell proliferation in vitro and tumorigenesis in vivo.