Background
Bladder cancer (BC) is the ninth most common cancer and is among the most frequent types of urinary malignancies worldwide [
1]. In 2012, 429,793 patients were diagnosed with BC, and 165, 084 deaths occurred globally [
2]. Approximately 25% of newly diagnosed patients present with muscle-invasive BC (MIBC) [
3] or metastatic disease [
4]. Lymph node (LN) metastasis is a crucial and powerful prognostic factor in BC [
5]. Therefore, a profound understanding of the detailed mechanisms underlying LN metastasis in BC is essential for improving the treatment strategies for BC.
Tumor progression is a complex, multistage process. Many oncogenes or pathways have been reported to be involved in cell invasion processes. G3BP2, a member of the Ras-GTPase-activating protein (RasGAP) SH3 domain-binding protein (G3BP) family, is remarkably overexpressed in various human cancers [
6,
7] and contributes to tumor invasion [
8]. Focal adhesion kinase (FAK) has been confirmed to play a vital role in tumor invasion and metastasis by binding to steroid receptor coactivator (SRC) [
9]. Knockdown of G3BP2 limited the migratory and invasive abilities of human lung cancer cells by inhibiting the activity of SRC and FAK [
8]. However, the upstream regulatory mechanism of G3BP2 remains elusive.
Circular RNAs (circRNAs) have been identified as a kind of endogenous noncoding RNA [
10] and are characterized by covalently closed loop structures without 5′ caps or 3′ polyadenylated tails [
11]. In recent years, large amounts of circRNAs have been successfully discovered and identified in different types of cell lines and species by high-throughput sequencing analysis [
12]. Many circRNAs have been proven to be dysregulated in various human cancers [
13,
14] and have been confirmed to regulate gene expression by sponging miRNAs [
13‐
15]. Although several circRNAs have been reported in BC [
16,
17], few tumor invasion-related circRNAs and their involved signaling pathways have been elucidated.
In this study, we identified a novel invasion-related circRNA, circFNDC3B, from published RNA-Seq data of human BC tissues and normal bladder tissues [
13] through our established cell invasion model. We further demonstrated that circFNDC3B, which originated from exons 5 and 6 of the FNDC3B gene, was markedly downregulated in BC tissues and cell lines. Low expression of circFNDC3B was significantly associated with pathological stage, grade, lymphatic metastasis and poor survival. Subsequently, we discovered that overexpression of circFNDC3B dramatically inhibited the proliferation, migration and invasion of BC cells via sponging miR-1178-3p to suppress the expression of G3BP2 and inhibit the phosphorylation of SRC/FAK.
Methods
Ethics statement and tissue collection
Fresh BC tissues were obtained from patients who were diagnosed with bladder urothelial cancer at Sun Yat-sen Memorial Hospital, Sun Yat-sen University between 2010 and 2016. Eighty-two cases of fresh BC tissues and fifty-six paired adjacent noncancerous tissues (≥3 cm away from the tumor) were immediately frozen in liquid nitrogen and stored at − 80 °C until further use. The experiments were conducted in accordance with the Declaration of Helsinki and approved by the Ethics Committee of Sun Yat-sen Memorial Hospital, Sun Yat-sen University. Written informed consent was obtained from all patients before participation in this study.
Cell culture
The human invasive BC cell lines T24 and UM-UC-3, and the human immortalized uroepithelium cell line SV-HUC-1 were purchased from ATCC. T24 cells were cultured in RPMI 1640 medium (Gibco, USA); UM-UC-3 cells were cultured in DMEM, and SV-HUC-1 cells were cultured in F-12 K medium (Gibco, USA) supplemented with 10% fetal bovine serum (BI, Israel); and 1% penicillin/streptomycin (Gibco, USA) in a humidified atmosphere of 5% CO2 at 37 °C.
Isolation of invasive and noninvasive BC cell sublines
As previously reported [
18], six-well polycarbonate transwell membranes with 8-mm pore inserts (Corning, USA) were used to isolate BC cell T24 sublines. First, T24 cells were serum-starved for 24 h. Then, 1 mL of cell suspension (5 × 10
5 cells/ml) with serum-free RPMI 1640 medium was seeded into the upper chamber, which was coated with 200 mg/mL of Matrigel (BD Biosciences, USA), and 2.5 mL of RPMI-1640 medium supplemented with 20% FBS was placed into the lower well. The highly invasive cells on the bottom of the membrane and the poorly invasive cells on the top of the membrane were harvested aseptically after incubation for 24 h. Next, the harvested cells were cultured and selected for ten rounds. The cell subline that failed to invade through the membranes in all selection rounds was designated as poorly invasive T24 and the subline that succeeded migrating through the membranes was designated as highly invasive T24. The invasive abilities of the two sublines were then confirmed again by wound healing assay and transwell Matrigel invasion assay.
Actinomycin D assay
T24 and UM-UC-3 cells were seeded at 1 × 10
5 cells per well in a 6-well plate overnight and then exposed to 2 mg/L actinomycin D (Sigma, USA) for 4, 8, 12 and 24 h [
13]. The cells were harvested at the indicated time points and the stability of circFNDC3B mRNA was analyzed using qRT-PCR. Three independent experiments were performed in triplicate.
Fluorescence in situ hybridization
When T24 cells seeded in confocal dishes were cultured to 80–90% confluence, the cells were fixed, prehybridized, and hybridized in hybridization buffer [
13] with a Cy3-labeled circFNDC3B probe (GenePharma, China) at 37 °C overnight. For the double FISH assay, the Cy3-labeled circFNDC3B probe and Cy5-labeled miR-1178-3p probe were used for hybridization. The signals of the probe were detected by a Fluorescent In Situ Hybridization Kit (GenePharma, China) according to the manufacturer’s protocol. Nuclei were counterstained with Hoechst 33342. The images were captured on ZEISS LSM800 confocal microscope (Carl Zeiss AG, Germany).
CircRNA plasmid construction and stable transfection
To construct circFNDC3B-overexpressing plasmids, human circFNDC3B cDNA was synthesized and cloned into the plenti-ciR-GFP-T2A vector (IGE Biotech Co, China). Plasmids were transfected into 293 T cells to package lentivirus using X-treme (Sigma, USA) according to the manufacturer’s instructions. T24 and UM-UC-3 cells were infected with the packaged lentivirus and selected with 2 μg/ml puromycin for 3 days. Surviving cells were then used to confirm overexpression efficiency.
Oligonucleotide transfection
T24 and UM-UC-3 cells were seeded in 6-well plates and cultured to 60–70% confluence before transfection. siRNAs, miRNA mimics, or inhibitors (GenePharma, China) were transiently transfected using Lipofectamine RNAiMax (Invitrogen, USA) according to the manufacturer’s protocol.
RNase R treatment
RNase R (Epicentre Technologies, USA) was used to degrade linear RNA. Briefly, RNAs extracted from T24 and UM-UC-3 cells were split to two aliquots: one for RNase R digestion and another for control with digestion buffer only [
19]. For RNase R digestion, 2 μg of total RNA was mixed with 0.6 μl 10 × RNase R Reaction Buffer and 0.2 μl RNase R (20 U/μl); for control, 2 μg of total RNA was mixed with 0.6 μl 10 × RNase R Reaction Buffer and 0.2 μl DEPC-treated water. Then, the samples were incubated at 37 °C for 30 min [
13]. GAPDH in the control group was used as an internal control [
19]. Three independent experiments were performed in triplicate.
RNA preparation and qRT-PCR
RNA isolation of nuclear and cytoplasmic fractions was performed with NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, USA) according to the manufacturer’s protocol. Total RNA from tissues and cells was extracted with RNAiso Plus (TaKaRa, Japan). Reverse transcription was performed with Prime Script RT Master Mix (TaKaRa, Japan) for circular RNA and mRNA. For miRNA, cDNA was synthesized using a miRNA First Strand cDNA Synthesis Kit (Sangon Biotech, China). Subsequently, the cDNA was subjected to real-time PCR on a Quantstudio™ DX system (Applied Biosystems, Singapore). GAPDH was used as an endogenous control for circRNA and mRNA. The expression of miRNA was normalized to small nuclear U6B (RNU6B). The 2-∆∆CT method was used to calculate relative expression.
Cell proliferation assay
For the cell viability assay, T24 and UM-UC-3 cells transfected with si-circFNDC3B- or circFNDC3B-overexpression vector were reseeded into 96-well plates (1 × 103 cells per well), and cell viability was assessed by MTS assay (Promega, USA). The absorbance of each well was read at a wavelength of 492 nm on a SPARK 10 M spectrophotometer (Tecan, Austria). For cell colony formation ability, the treated T24 and UM-UC-3 cells were seeded into 6-well plates (500 cells per well). After incubation at 37 °C in a 5% humidified atmosphere for 10 days, colonies were fixed in 4% paraformaldehyde, stained with 0.1% crystal violet, counted and photographed. Three independent experiments were performed in triplicate.
Cell migration and invasion assays
For wound healing assays, T24 and UM-UC-3 cells were seeded in 6-well plates, and a straight scratch was made using 200 μL pipette tips. Images of wounds were captured at indicated time at three different positions using 10 high-power fields after scratching with a sterile 200 μL pipette. Three independent experiments were performed in triplicate.
For transwell migration and invasion assays, a 24-well transwell chamber (Costar, USA) with or without precoated Matrigel was used to detect cell migratory or invasive abilities according to the manufacturer’s protocol. Cells suspended in 0.2 ml serum-free medium (8 × 104 cells/well for migration, 1 × 105/well for invasion) were added to the upper chambers, and medium supplemented with 10% FBS was applied to the lower chambers. After incubating the cells for 11 h (for T24) and 24 h (for UM-UC-3) at 37 °C; and 5% CO2, the cells that migrated to the lower membrane surface were fixed with 4% paraformaldehyde and stained with 1% crystal violet in PBS. The migrated and invaded cells were counted in three randomly selected fields. Three independent experiments were performed in triplicate.
Biotin-labeled probe pull-down assay
Pull-down assay was performed as previously described [
16,
17]. First, to prepare probe-coated beads, the biotinylated circFNDC3B probe or oligo probe (GenePharma, China) was incubated with Streptavidin-Dyna beads M-280 (Invitrogen, USA) at room temperature for 2 h. Then, appoximately 1 × 10
7 stably overexpressing circFNDC3B or control BC cells were fixed with 1% formaldehyde, lysed in lysis buffer and incubated with probe-coated beads at 4 °C overnight. The probe-Dyna bead-circRNA mixture was washed, followed by incubation with 200 μl lysis buffer and proteinase K at room temperature for 2 h to reverse the formaldehyde crosslinking. Then, the RNA complexes were extracted with RNAiso Plus (TaKaRa, Japan) and detected by qRT-PCR. The results are presented as the means ± SEM from three independent experiments.
Biotin-labeled miRNA capture
Stably overexpressing circFNDC3B BC cells were transfected with biotin-labeled miRNA mimics or nonsense control (GenePharma, China) for 48 h. Streptavidin-Dyna beads M-280 were washed with lysis buffer; and blocked with yeast tRNA on a rotator at 4 °C for 2 h. The cells were harvested, lysed and incubated with the blocked beads at 4 °C overnight. The abundance of circFNDC3B in bound fractions was tested by qRT-PCR and agarose gel electrophoresis. Three independent experiments were performed in triplicate.
Dual luciferase reporter assay
Cells were seeded into 24-well plates with 3 × 104 cells per well. After transient transfection with constructed luciferase plasmids and miRNA mimics for 48 h, Rluc activity was measured with a dual-luciferase reporter assay system (Promega, USA) according to the manufacturer′s protocol. Renilla luciferase activity was normalized to the luminescence of firefly luciferase. Three independent experiments were performed in triplicate.
Western blot analysis
Cells were lysed in RIPA buffer (CWBIO, China) with protease and phosphatase inhibitors (CWBIO, China). Identical quantities of proteins were electrophoresed by SDS-PAGE, transferred onto PVDF membranes and incubated with primary antibodies specific for G3BP2 (1:1000 dilution, Proteintech, USA), SRC (1:800 dilution, ABclonal, China), p-SRC (1:1000 dilution, ABclonal, China), FAK (1:800 dilution, ABclonal, China), p-FAK (1:1000 dilution, ABclonal, China), β-actin (1:1000 dilution, Proteintech, USA), GAPDH (1:1000 dilution, ABclonal, China) at 4 °C overnight, followed by incubation with appropriate HRP-conjugated secondary antibodies at room temperature for 1 h. Signals were detected by Immobilon ECL substrate (Millipore, Germany), and the images were acquired using an Optimax X-ray Film Processor (Protec, Germany).
HE and immunohistochemistry
Immunohistochemical staining was performed according to published methods [
5]. First, 5-μm paraffin sections of tissue samples were stained with HE and immunohistochemistry. The primary antibodies specific for G3BP2 (Proteintech, USA), SRC (ABclonal, China), p-SRC (ABclonal, China), FAK (ABclonal, China), and p-FAK (ABclonal, China) were used at a 1:100 dilution in the experiments. Images were captured using a Nikon Eclipse 80i system with NIS-Elements software (Nikon, Japan).
Animal experiments
All animal care and experiments were performed according to the guidelines of the National Institutes of Health; and were approved by the Animal Ethics Committee of Sun Yat-sen University. To study the effect of circFNDC3B on tumor growth, circFNDC3B-overexpressing or control UM-UC-3 cells were subcutaneously injected into the upper back of 4-week-old female BALB/c nude mice (5 × 10
6 cells per mouse). Tumors were measured with a caliper every week. The tumor volume was calculated as (length× width
2)/2. One month later, the mice were executed, and the excised tumor tissues were further used for evaluation of tumor weight and pathologic examination. To explore the effect of circFNDC3B on lymphatic metastasis, lentivirus-transduced UM-UC-3 cells that stably expressed firefly luciferase were first constructed; subsequently, approximately 5 × 10
5 of these cells were inoculated into the footpads [
5]. After 4 weeks, the popliteal LNs were captured using an in vivo bioluminescence imaging system, then enucleated, weighed and embedded in paraffin.
Sequences used in this study
The sequences of the primers, oligonucleotides and probes used in this study are listed in Additional file
1: Table S1.
Statistical analysis
Statistical analyses were carried out using GraphPad Prism 7.0. Chi-square test was performed to analyze the relationship between circFNDC3B levels and clinicopathological characteristics. Kaplan-Meier method and log-rank test were used to calculate overall survival rates. Two-tailed Student’s t-test, Wilcoxon rank-sum test, or Mann-Whitney U-test were used to determine statistically significant differences between two groups, as appropriate. The Data are presented as the mean ± standard error of the mean (SEM). P < 0.05 was considered statistically significant.
Discussion
CircRNAs are a large class of widespread and highly stable endogenous noncoding RNAs that have been identified by high-throughput sequencing and bioinformatics analysis in recent years [
15,
28]. An increasing number of circRNAs have been confirmed to be dysregulated in various human cancers [
13,
14,
16,
17]. In the present study, we identified a novel circular RNA termed circFNDC3B that was aberrantly downregulated in BC tissues. Low expression of circFNDC3B was correlated with high grade, lymphatic metastasis and poor prognosis.
Invasion and metastasis are the greatest obstacles to successful tumor treatment and are the leading reason for the resultant mortality of patients with BC [
29], which urges us to identify invasion-related genes and elucidate the molecular mechanisms resulting in tumor invasion and metastasis. In our study, we constructed an invasion model by isolating highly and poorly invasive cell sublines from T24 using the repeated transwell method, which has been successfully used to investigate tumor metastasis [
18]. We successfully identified invasion-related circRNAs from RNA-Seq data of BC tissues through the cell invasion model. CircFNDC3B, one of the most differentially expressed circRNAs between highly and poorly invasive T24 sublines, was further confirmed to inhibit proliferation, migration and invasion in vitro and to suppress tumor growth and lymphatic metastasis in vivo, suggesting its tumor-suppressive effect.
A general phenomenon was discovered: circRNAs were described as pivotal gene regulators in humans mainly due to their posttranscriptional function. CircRNAs contain multiple miRNA-binding sites or miRNA response elements (MREs) that can function as miRNA sponges [
15,
28]. For example, circHIPK3 could bind to a host of miRNAs and act as a miRNA sponge in human cancers [
13,
16,
30]. hsa_circ_0000977 upregulates oncogene PLK1 by sponging miR-874-3p in pancreatic ductal adenocarcinoma [
31]. In our study, circFNDC3B was confirmed to be mainly located in the cytoplasm; and to harbor numerous miRNA-binding sites predicted by CircInteractome, suggesting that circFNDC3B may also function as miRNA sponge in BC. A series of experiments such as biotin-labeled probe pull-down assay, dual- luciferase reporter assay and biotin-labeled miRNA capture assay, confirmed that circFNDC3B directly binds to miR-1178-3p. Subsequent “rescue” experiments confirmed that circFNDC3B could reverse the oncogenic roles of miR-1178-3p, suggesting that circFNDC3B acts as a miR-1178-3p sponge in BC.
Mature miRNAs have been reported to bind to the 3’UTR of target mRNAs, leading to the repression of translation or induction of degradation of target mRNAs [
21]. To date, most studies have revealed that circRNAs positively regulate target genes of miRNAs. For example, circular RNA_LARP4 expression positively correlates with the tumor suppressor LATS1, the direct target of miR-424-5p in gastric cancer [
14]. Circ-ITCH, as a tumor suppressor in BC, could function as a ceRNA for miR-17 and miR-224 to promote p21 and PTEN expression [
17]. However, miRNAs also function to positively regulate gene expression by targeting the 5’UTR [
24]. In this study, miR-1178-3p was predicted to bind to the 5’UTR of G3BP2 by RegRNA2.0; A dual-luciferase reporter assay confirmed that miR-1178-3p could enhance G3BP2 expression via binding to its 5’UTR. Further study showed that overexpression of circFNDC3B significantly decreased the expression of G3BP2 in vitro and in vivo. G3BP2 has been reported to be overexpressed in BC cell lines [
6] and to promote migration and invasion, as confirmed in our study. We speculated that regulatory networks of circFNDC3B/miR-1178-3p/G3BP2 could be involved in invasion and metastasis. First, miR-1178-3p was overexpressed in BC tissues and promoted the proliferation, migration and invasion of BC cells, which functionally exerted a negative correlation with circFNDC3B. Second, G3BP2 was validated as the direct target of miR-1178-3p. Third, overexpression of circFNDC3B partially abolished the effect of miR-1178-3p on G3BP2, implying a novel regulatory axis formed by circFNDC3B/miR-1178-3p/G3BP2 in BC.
Nonreceptor protein tyrosine kinases, such as FAK and SRC protooncogenes, play vital roles in integrin-associated signal transduction [
32]. Aberrant activation of SRC/FAK signaling leads to enhanced migratory and invasive capabilities in many human tumors [
33] and plays a pivotal role in tumor metastasis [
8]. G3BP2 has been reported to activate the SRC/FAK signaling pathway. Therefore, during further studies, we observed that overexpression of circFNDC3B also inhibited the activity of SRC and FAK. Additionally, the effects of miR-1178-3p on SRC and FAK, which increased their activity, were abrogated upon overexpressing circFNDC3B simultaneously in BC cells compared with the control group. Furthermore, knockdown of G3BP2 could reverse sh-circFNDC3B-induced increase on p-SRC and p-FAK expression. These results indicate that circFNDC3B inactivates the SRC/FAK signaling pathway via miR-1178-3p/G3BP2.