Background
Bladder cancer (BCa) is the 10th most frequent cancer across the globe with multiple histological types, mainly including urothelial cancer and non-urothelial BCa, associated with high fatality [
1,
2]. BCa develops on the urogenital tract (papillary or nonpapillary), which correspond to clinically different kinds of the disease [
3]. The biological process from reduction of epithelial cell identity to acquirement of mesenchymal phenotyping is regarded as epithelial-mesenchymal transition (EMT) [
4]. The initiation of EMT can facilitate the angiogenesis and oncogenic phenotype of multiple malignant tumors, including BCa [
5]. Thus, finding possible molecular targets for inhibiting EMT would be of high clinical significance for the understanding of BCa pathogenesis.
Circular RNAs (circRNAs) are a class of non-coding RNAs formed by covalently closed loops through back-splicing and exon-skipping, which confer pivotal effects on plenty of biological functions, acting as microRNA (miR) sponges and reservoirs, as well as combining with RNA-binding proteins during cancer progression, including BCa [
6]. For example, circEHBP1 elevation occurs in BCa with positive correlation with lymphatic metastasis and dismal prognosis of patients with BCa [
6]. Ectopic expression of circRNA-MYLK accelerates the growth dynamics and EMT of BCa cells [
7]. In silico analysis in our work revealed the overexpression of circ_0000658 in BCa, though the role of circ_0000658 was rarely investigated. CircRNAs could exert function on by binding to miRNAs; for example, circGLIS3 was demonstrated to augment BCa cell proliferative capacity by binding to miR-1273f [
8]. miR-498 was identified to be bound to by circ_0000658 based on our bioinformatic analysis. Although rarely studied in BCa, the tumor-suppressing roles of miR-498 have been revealed in other cancers [
9,
10].
Furthermore, miR-498 has been proposed to target the 3′-untranslated region (UTR) of high mobility group AT-hook 2 (HMGA2) and inhibits its translation in non-small cell lung cancer [
11]. As a non-histone architectural transcription factor, HMGA2 regulates gene transcription by connecting AT-rich sequences in minor groove of B-form DNA and changes the chromatin structure [
12]. Upregulated HMGA2 predicts dismal overall survival for BCa patients, serving as biomarkers for cisplatin resistance [
13]. Based on the references, we posed a speculation that circ_0000658 may modulate the EMT of BCa cells via the miR-498/HMGA2 axis. Therefore, we performed loss- and gain- function assays in BCa cells and nude mouse xenografted with BCa cells to validate the possible speculation.
Material and methods
Ethical statement
The current study was approved by the Ethics Committee of The First Affiliated Hospital of Soochow University and conducted in strict accordance with the Declaration of Helsinki. All participants or their families signed informed consent documentation before sample collection. Animal experiments were performed under the approval of Animal Ethics Committee of The First Affiliated Hospital of Soochow University in accordance with the Guide for the Care and Use of Laboratory animals published by the US National Institutes of Health.
The BCa-related circRNA expression dataset GSE92675 retrieved from the Gene Expression Omnibus (GEO) database, with the platform annotation file of GPL19978, including 4 normal samples and 4 cancer tissue samples. Differential analysis was performed using R language “limma” package with |logFC| > 1 and p < 0.05 as the screening criterion. Meanwhile, the circBase database file was downloaded to convert the differentially expressed circRNA name to circRNA ID, and a heatmap of the top ten differentially expressed genes was drawn using R language “pheatmap” package. BCa-related circRNA was searched through circFunBase database, and intersection of differentially expressed circRNA and BCa-related circRNA was retrieved using the jvenn tool.
The downstream miRNAs of circRNA were predicted through RegRNA 2.0 (score ≥ 120& free_energy ≤ − 20) and circInteractome, and then intersected miRNA served as a candidate miRNA. The downstream target genes of miRNA were using bioinformatics tools StarBase and miRDB (Score > 70). BCa-related genes were analyzed using GeneCards database (Score ≥ 25) and intersected with the downstream target genes of miRNAs to identify the candidate gene. The expression of candidate genes was obtained in BCa and normal samples through the StarBase website.
Clinical sample collection
Cancer tissues and adjacent normal tissues were surgically collected from 50 patients with BCa (Table
S1) at The First Affiliated Hospital of Soochow University from January 2015 to January 2017, including 39 males and 11 females, aged between 42 to 79 years old, with an average age of 62.54 ± 11.20 years old. Four pairs of frozen samples of BCa tissues and adjacent normal tissues were selected for circRNA analysis. The samples were postoperatively pathologically confirmed. None of the patients had anti-tumor therapy prior to operation. All patients were followed up for 36 months from post-operation to January 2020. Clinical pathology referred to World Health Organization/International Society of Urological Pathology (WHO/ISUP) [
14] and International Union Against Cancer (UICC) tumor node metastasis (TNM) classification.
Immunohistochemistry
Clinical tissue specimens were collected for laser microdissection to obtain purer tumor tissues. After 3% methanol H2O2 treatment, antigen was retrieved from fixed and paraffin-embedded tissue sections. Following blocking in normal goat serum, specimens were immunostained with primary rabbit anti-human HMGA2 (ab207301, 1: 1000; Abcam) overnight at 4 °C. Incubation was further carried out with horseradish peroxidase-labeled secondary antibody goat anti-rabbit immunoglobulin G (IgG; ab6785, 1:1000, Abcam). After exposure to DAB, specimens were counterstained with hematoxylin. Observation and photographing were carried out by a microscopy with randomly selected 5 high-powered fields of view in each section and 100 cells at each field.
RT-qPCR
Total RNA was isolated from tissues and cells using TRIzol reagent (9108Q, Takara, Japan), followed by reverse transcription into cDNA using a miRNA First-Strand cDNA Synthesis Kit (Tailing Reaction) (B532451, Sangon Biotechnology, Shanghai, China) and MightyScript First-Strand cDNA Synthesis Master Mix Kit (B639251, Sangon). The RNA was diluted to 10 folds, and 2 μL of cDNA products was collected as a template for PCR amplification, which was conducted using AceQ® qPCR SYBR Green Master Mix PCR (Q111, Vazyme, China). The gene quantification was normalized to GAPDH or U6 (Table
S2) using a 2
-ΔΔCt method.
Cell culture and transfection
Normal bladder epithelial cell line (SV-HUC-1) and BCa cell lines (MGH-U3, T24, 5637, SW780) were selected for this study. MGH-U3 cell line was purchased from CoBioer Biosciences CO., LTD (Nanjing, China), while the remaining cell lines were purchased from American Type Culture Collection (Manassas, VA). Culture of cells was carried out in RPMI-1640 (Gibco, Carlsbad, CA) containing 10% FBS, 10 μg/mL streptomycin and 100 U/mL penicillin at 37 °C and 5% CO2.
Cells in the logarithmic phase were trypsinized, seeded in a 6-well plate at a density of 1 × 105 cells/well and cultured for 24 h. At 75% confluence, cells were transduced with short hairpin RNA (sh)-negative control (NC), sh-circ_0000658, overexpression (oe)-NC, oe-circ_0000658 by Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA). Next, 48 h after transfection, the transfection efficiency of sh-circ_0000658 was validated by RT-qPCR. The expression plasmid was purchased from GenePharma (Shanghai, China) at the concentration of 50 ng/mL. The core plasmid (PLKO.1) of silencing sequence of the inserted target gene and the core plasmid (pHAGE-CMV-MCS-IzsGreen) of the cDNA sequence of the inserted target gene were purchased from GenePharma (Shanghai, China), and the plasmid concentration used was 50 ng/mL.
Western blot analysis
Cell lysis was carried out in enhanced RIPA lysis buffer containing protease inhibitors. Next, the protein concentration was quantified with a BCA Kit (BOSTER, China). The isolated protein by 10% sodium dodecyl sulfate polyacrylamide gel electrophoresis was electro-transferred onto polyvinylidene fluoride membrane. The membrane was blocked with 5% BSA and probed overnight at 4 °C with diluted primary antibodies against HMGA2 (rabbit, ab97276, 1:1000, Abcam), E-cadherin (rabbit, ab40772, 1:10000, Abcam), β-catenin (rabbit, ab32572, 1:5000, Abcam), Slug (rabbit, ab27568, 1:1000, Abcam), Snail (rabbit, ab216347, 1:1000, Abcam), ZEB1 (ab203829, 1:500, Abcam), Twist (mouse, ab50581, ab175430, 1:1000, Abcam), N-cadherin (rabbit, ab76011, 1:5000, Abcam), β-actin (rabbit, ab8227, 1:5000, Abcam). The following day, the membrane was re-probed with HRP-labeled goat anti-mouse secondary antibody (ab205719; 1: 2000; Abcam) or goat anti-rabbit (ab205718, 1:2000, Abcam) at room temperature for 1 h. The immunocomplexes on the membrane were visualized using ECL reagent (EMD Millipore, Bedford, Massachusetts) and band intensities were quantified using Image J software, with β-actin as a loading control.
EdU assay
The cells to be tested were seeded in a 24-well plate. The culture medium with EdU at a concentration of 10 μmol/L was applied to incubate cells for 2 h. Following incubation with 100 μL of staining solution for 30 min, cells were subjected to DAPI development to stain the nucleus, which were observed by a fluorescence microscope .
Scratch test
At intervals of 0.5–1 cm on the bottom of the 6-well plate, horizontal lines were created, with at least five lines through each well. Cells were seeded to the 6-well plate at a density of 5 × 105 cells/well. The sterile 10 μL pipette tip was perpendicular to the horizontal line to the scratch on the back. The scratched cells were photographed with an inverted microscope at 0, 6, 12, and 24 h to observe migration of cells.
Transwell assay
The apical chamber surface of the bottom membrane was coated with Matrigel, which was polymerized into a gel. Culture medium containing 10% FBS was inserted in the basolateral chamber. Next, 100 μL of cell suspension was incubated in the chambers at 37 °C for 24 h, and cells failing to invade the surface of the Matrigel membrane were discarded. Fixed cells with 4% paraformaldehyde were stained with 1% crystal violet, which were observed and counted by an inverted light microscope using ImageJ software.
Fluorescence in situ hybridization (FISH) assay
The FITC-circ_0000658 and Cy3-miR-498 probes from Ribobio (Guangzhou, China) were used to analyze the localization in the cells or tissues. Normal bladder epithelial cells (1 × 105 cells/well) were seeded onto a 6-well culture plate and cultured for 1 day. Upon reaching 80% cell confluence, the cells were fixed with 4% paraformaldehyde. Sections were subjected to 250 μL of pre-hybridization solution at 42 °C for 1 h, and then to 250 μL hybridization solution containing probe (300 ng/mL) overnight at 42 °C. Nucleus was stained with PBST-diluted DAPI (1: 800). Finally, the sections were mounted with anti-fluorescence quencher, and observed under a fluorescence microscope.
RNA binding protein immunoprecipitation (RIP)
The binding of miR-498 and HMGA2 protein was determined using RIP kit (Millipore, Billerica, MA). The lysed cells were centrifuged to obtain the supernatant. Next, 10 μL of cell extract was taken out as input, and the other portion of the cell extract was incubated with antibody for co-precipitation. The sample and Input were digested with proteinase K, and RNA was extracted for subsequent RT-qPCR detection of HMGA2. The antibodies used were: rabbit anti-HMGA2 (1: 100, ab97276, Abcam) and rabbit anti-human IgG (1:100, ab109489, Abcam; NC).
RNA pull-down assay
BCa cells were transfected with wild type (WT) and mutant (MUT) biotinylated circ_0000658 (50 nM each). After 48 h of transfection, the cells were collected, vortexed, and incubated with cell lysis buffer (Ambion, Austin, Texas) for 10 min. Next, 50 mL of sample cell lysate was aliquoted. The remaining lysate was incubated with M-280 streptavidin magnetic beads (Sigma) pre-coated with RNase-free and yeast tRNA (Sigma) for 3 h at 4 °C, then wash twice with cold lysis buffer, three times with low-salt buffer, and once with high-salt buffer. Finally, the total protein was extracted.
Dual-luciferase reporter gene assay
The predicted fragments and MUT fragments of circ_0000658 or HMGA2 with miR-498 binding sites were inserted into the luciferase reporter vector as reporter plasmids circ_0000658-WT, circ_0000658-MUT, HMGA2-WT and HMGA2-MUT, respectively. The reporter plasmids were then co-transfected with mimic NC or miR-498 mimic into 293 T cells (Oulu Biotecnology, Guangzhou, China) to analyze whether miR-498 can bind to circ_0000658 or HMGA2. Following 48-h transfection, the cells were lysed and subjected to the luciferase detection kit (K801–200, Biovision, Milpitas CA), with luciferase activity normalized to renilla luciferase activity.
Nude mouse xenografted with BCa cells
A total of 18 healthy nude mice (Beijing Institute of Pharmacology, Chinese Academy of Sciences, Beijing, China) aged 6–8 weeks old were raised in a Specific Pathogen Free animal laboratory with humidity of 60–65% at 22–25 °C. The animals were provided with free access to food and water under a 12-h light and dark cycle. After one-week acclimatization, nude mice were subcutaneously injected with T24 cells transduced with plasmids containing oe-NC + sh-NC, oe-circ_0000658 + sh-NC or oe-circ_0000658 + sh-HMGA2 (1 × 109 pfu/100 μL) (n = 6/group). After 6 weeks, the nude mice were euthanized, with tumor volume and weight recorded. The expression of HMGA2 and tumor metastasis markers was detremined.
Statistical analysis
SPSS 21.0 software (IBM Corp., Armonk, NY) was used for data processing. Measurement data were presented as mean ± standard deviation. Paired or unpaired t-tests were applied for the data comparison between two groups. One-way analysis of variance (ANOVA) with Tukey’s post hoc tests was used to compare data among multiple groups. Bonferroni-corrected repeated measures ANOVA was used for data comparison at different time points. Kaplan-Meier analysis was used to calculate the survival rate of patients, and the survival rates were compared with Log-rank test. The correlation of indicators was observed using Pearson correlation analysis. In all statistical analysis, a value of p < 0.05 represents statistical significance.
Discussion
BCa develops on the urogenital tract (papillary or nonpapillary), which correspond to clinically different kinds of the disease [
3]. EMT develops from loss of cell polarity of epithelial cells, cell-cell adhesion to mesenchymal phenotype, which exacerbates cancer progression [
16]. The association between EMT and BCa progression and metastasis has been previously highlighted [
17]. In the present study, we identified a novel differentially expressed circ_0000658 in the BCa progression via regulation of EMT. The present study validated that circ_0000658 could attenuate miR-498 binding to HMGA2 to augment EMT, oncogenic phenotypes of BCa cells, which could exacerbate the development of BCa.
Initially, we revealed that circ_0000658 was highly expressed in BCa tissues and cells and correlated to the poor prognosis of BCa patients. In addition, circ_0000658 augmented the growth dynamics of BCa cells, which was rarely reported. Concordantly, elevation of circEHBP1 occurred in BCa and promotes the lymphatic metastasis with dismal prognosis of BCa patients [
18]. EMT is known as a process of the transdifferentiation of epithelial cells into motile mesenchymal cells, which leads pathologically to fibrosis and cancer progression [
19]. EMT occurs in both physiological and pathological conditions and can be triggered by a conserved set of inducing signals, downstream effectors and transcriptional regulators, contributing to functional changes in cell migratory and invasive potential [
20]. E-cadherin, N-cadherin, β-catenin, Slug, Snail, Twist, and ZEB2 are EMT markers, which predict intravesical recurrence in patients with non-muscle-invasive urothelial carcinoma of the bladder [
21]. Further analysis in the current study revealed that overexpression of circ_0000658 reduced the expression of β-catenin and E-cadherin, whereas elevating that of N-cadherin, Slug, Snail, ZEB1 and Twist. Conversely, a contrary result was observed in response to sh-circ_0000658. This represents the first evidence for the regulation of the EMT by circ_0000658 in BCa and may have importance in regulating the progression of BCa. Consistent with our study, knockdown of EFEMP2 in BCa cells triggered reduction in the epithelial marker E-cadherin expression, as well as increase in mesenchymal markers N-cadherin, Snail and Slug, which is associated with augmented cell proliferative, migratory and metastatic capacities [
22]. Depletion of circ_100984 retards the BCa tumor growth and migratory and invasive capacities in vitro and in vivo by inhibiting expression of EMT markers [
23]. Our in vivo experiments also demonstrated that circ_0000658 increased the tumor volume and weight, which promotes the tumorigenesis of BCa cells.
Moreover, we also revealed that circ_0000658 competitively bound to and restricted miR-498 expression, although it’s rarely documented. Evidence has suggested that lncRNAs can interact with miRNAs and regulate the expression of miRNAs as a competitive endogenous non-coding RNA [
24]. For instance, circ_GFRA1 has been proposed to competitively bind to miR-498 and negatively regulate the expression of miR-498 in the context of hepatocellular carcinoma [
25]. In addition, circ_PRMT5 can sponge miR-498 and thus inhibit the expression of miR-498 during non-small-cell lung cancer [
9]. Our results displayed that miR-498 knockdown augmented the EMT, oncogenic phenotypes of BCa cells. miRNAs exert crucial roles in the initiation and progression of cancer due to their involvement in the regulation of various biological processes, including EMT, by acting as oncomiRs or as tumor suppressors via multiple molecular mechanisms [
26]. For example, miR-485-5p has been recognized as a tumor suppressor, and its ectopic expression could restrict EMT and metastasis of bladder cancer cells through targeting HMGA2 [
15]. Meanwhile, miR-498 expression has been found to be reduced in liver cancer patient tissues and cell lines while its upregulation suppresses liver cancer cell proliferative and invasive capacities and EMT [
27]. Inhibiting miR-498 expression can induce the EMT and proliferative potentials in non-small-cell lung cancer cells [
28]. Additionally, miR-498 expression is reduced in gastric cancer associated with dismal prognosis but its abundant expression induces a reduction in EMT markers to suppress the metastatic and proliferative capacity of gastric cancer cells [
10]. However, the involvement of miR-498 in BCa confirmed by the current study warrants further investigation due to the lack of available literature to support it.
miR-498 has been proposed to target the 3′-UTR of HMGA2 and inhibits its translation in non-small cell lung cancer [
11]. A prior study supports that miR-498 targeted and negatively regulated HMGA2. Additionally, accumulating evidence has demonstrated the regulation of circRNAs on HMGA2 whereby circRNAs can act as miRNA sponges, and thus reduce their regulatory effects on the target mRNAs [
29‐
31]. The current study confirmed for the first time that circ_0000658 enhanced the expression of HMGA2 by competitively binding to miR-498. We also revealed that the inhibition of HMGA2 reversed the trends of circ_0000658 overexpression on tumor volume and EMT markers in vivo. Consistently, HMGA2 expression is elevated in the BCa tissues relative to non-cancerous tissues and its inhibition is correlated with the delayed BCa progression [
32]. N-cadherin expression is observed to be reduced in HMGA2-knocked down cells of gastric cancer, and HMGA2 elevation would exacerbate the invasive and metastatic potential of gastric cancer with the promoting effect on EMT [
33]. In the context of colorectal cancer, the depletion of HMGA2 could abrogate the promoting effects of miR-532-3p inhibitor on cell malignancy [
34].
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.