KCNQ1OT1 aggravates cell proliferation and migration in bladder cancer through modulating miR-145-5p/PCBP2 axis

From June 2012 to January 2018, 70 bladder cancer specimens and corresponding normal tissues were gathered from Affiliated Hospital of Beihua University. Patients who provided these tissues did not receive any treatment before being operated. Every patient signed informed consent, based on the review of Ethics of committee of Affiliated Hospital of Beihua University.

Human bladder epithelial immortalized cell (SV-HUC-1) as well as bladder cancer cells (UM-UC-3, T24, HT-1376 and HT-1197) were provided by Chinese Academy of Sciences (Beijing, China). The above cells were cultured at 37 °C in a moist incubator containing 5% CO₂ in RPMI-1640 (Invitrogen, Waltham, MA, USA) supplied with 10% FBS (Invitrogen), 1% penicillin (Sigma-Aldrich, Milan, Italy) or streptomycin (Sigma-Aldrich).

Transfected cells were put into a 6-well plate until cells was 80% confluence. T24 and HT-1197 cells were co-transfected with shRNAs targeting KCNQ10T1 (sh-KCNQ10T1#1/2), PCBP2 (sh-PCBP2#1/2) and shNC. The pcDNA3.1 vector for PCBP2 was used for overexpression studies. The miR-145-5p mimics, NC mimics and miR-145-5p inhibitor were gained from Genechem (Shanghai China). All plasmids were transfected into cells using Lipofectamine 2000 (Invitrogen).

Total RNA was isolated from BC cells (T24 and HT-1197) or BC patient tissues according to the reference method TRIzol Reagent (Invitrogen). PrimeScript RT reagent Kit (TaKaRa, Tokyo, Japan) was applied to synthesize cDNA, which was used for qPCR analysis. SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) was utilized on 7500 Real-time PCR System (Applied Biosystems), with cDNA as template. The gene expression was calculated by the 2^−ΔΔCt method. GAPDH or U6 was the internal control [1].

Cells were placed into 6-well plates at first and cultivated in normal cell growth fluid. Sterile pipette was applied for scratching the cell layer to made scratches. After that, the cells were cleaned and cultured in another medium. Eventually, wound closure was collected at different times.

Total proteins were extracted through a Total Protein Extraction Kit (KeyGen Biotech, Nanjing, China). Concentrations of protein were examined under BCA commercial kit (KeyGen). Proteins were separated using SDS-PAGE and then moved onto PVDF membranes (Millipore, Carlsbad, USA). Membranes were sealed in non-fat milk and then incubated with given primary antibodies: anti-MMP2 (1:1000, ab37150, Abcam, Cambridge, USA), anti-MMP7 (1:1000, ab5706, Abcam), anti-Bcl-2 (1:1000, ab32124, Abcam), anti-bax (1:1000, ab32503, Abcam), anti-caspase 3 (1:1000, ab13847, Abcam), anti-cleaved caspase-3 (1:1000, ab2302, Abcam), anti-caspase 8 (1:1000, ab25901, Abcam), anti-cleaved caspase-8 (1:1000, ab25901, Abcam), anti-caspase 9 (1:1000, ab32539, Abcam), anti-cleaved caspase-9 (1:1000, ab2324, Abcam), anti-E-cadherin (1:1000, ab15148, Abcam), anti-N-cadherin (1:1000, ab76057, Abcam) and anti-GAPDH (1:1000, ab8245, Abcam). GAPDH was used as a measurement control for other proteins. Moreover, the membranes were co-cultured with goat anti-mouse IgG H&L (Cy3^®) preadsorbed (1:2000, ab97035, Abcam) secondary antibodies for 1 h darkness. Chemiluminescence system (Invitrogen) was employed to observe the protein bands.

About CCK-8 assay, cultured T24 and HT-1197 cells were collected to 6-well plates with the density of 1 × 10⁴ cells/well, and then which were mixed with CCK8 reagent (Dojindo, Kumamoto, Japan) for 2 h. The absorption at 450 nm was evaluated at the indicated time points by utilizing microplate reader (EL340; Bio-Tek Instruments, Hopkinton, MA, USA). As for colony formation assay, (800–1000) cells were plated in the six-well plates. Change the medium every 3 days. After 14 days of incubation, PBS (Solibao technology, Shanghai, China) was used to wash and methanol (Grass biotechnology, Nanjing, China) was applied to wash and fix it for 10 min, and then stained with crystal violet (Beyotime Biotechnology, Nantong, China) for 5 min. The visible colony numbers then were counted [19].

T24 and HT-1197 cells in apical compartment were cultured with serum-free medium, and the basolateral chamber was filled with 10% FBS medium. Cells were plated on the top compartment of membrane pre-coated with Matrigel for invasion assay. 24 h later, invaded cells were fixed in methanol (Solarbio, Beijing, China) and dyed in crystal violet (Solarbio). The cell numbers were recorded and photographed randomly from 5 chosen fields by an inverted microscope (Olympus, Tokyo, Japan).

Cell apoptosis was evaluated by an annexin/PI Kit (KeyGen). T24 and HT-1197 cells were washed by using PBS (Solarbio) and rehanged, then fixed in 70% cooled ethanol (Solarbio). Ultimately, the rate of apoptosis was analyzed by Flow Cytometer (BD Bioscience, Massachusetts, USA).

TUNEL assays were carried out conforming to supplier’s protocols. After rising twice by PBS, cells of T24 or HT-1197 were fixed and permeabilized by paraformaldehyde (Solarbio) and 0.25% Triton-X 100, respectively. In short, the above cells were stained in DAPI (Invitrogen) or Merge (Invitrogen). Then, cells were surveyed and photographed by fluorescence microscopy (Olympus).

Subcellular isolation of RNAs in T24 and HT-1197 cells was carried out via Cytoplasmic and Nuclear RNA Purification Kit (Norgenbiotek Corporation, Thorold, Canada). Then the relative expression of KCNQ1OT1 was determined by qRT-PCR. Besides, GAPDH/U6 was regarded as cytoplasmic and nuclear control, separately.

GenePharma was used to design the FISH probes of KCNQ1OT1. T24 and HT-1197 cells were immobilized for 10 min in 4% PFA and washed with PBS (Sigma-Aldrich) for three times. Subsequently, PBS containing with 0.5% Triton X-100 was applied for cell permeabilization. Following washing, pre-hybridization buffer (Sigma-Aldrich) was added into each well to incubate, which lasted for half an hour. The slide was hybridized with KCNQ1OT1 FISH probe, and then was washed three times with 2× Saline Sodium Citrate (SSC; Sigma-Aldrich) at 42 °C. Hoechst 33342 solution (Invitrogen) was used to stain cells. Finally, fluorescent signal was observed by using the microscope [20].

The wild-type and mutant binding sites of miR-145-5p in KCNQ1OT1 sequence or PCBP2 3′UTR were sub-cloned into pmirGLO dual-luciferase vector to construct plasmids, named KCNQ1OT1-Wt/Mut or PCBP2-Wt/Mut. Later, the plasmids were co-transfected with miR-145-5p mimics into T24 or HT-1197 cells, respectively. At last, the activity of luciferase was examined by Dual-luciferase Reporter System (Promega, Massachusetts, USA). Relative firefly luciferase activity was normalized to Renilla luciferase activity.

The miR-145-5p-Wt, miR-145-5p-Mut and NC were biotin labeled into Bio-miR-145-5p-Wt, Bio-miR-145-5p-Mut or Bio-NC. Then, cell lysates were cultivated after 48 h and co-incubated with the biotinylated probe and M-280 streptavidin magnetic beads (Sigma). And the level of KCNQ10T1 was analyzed using qRT-PCR subsequently.

RIP assays were employed for investigating the potential interaction under a Magna RNA-Binding Protein Immunoprecipitation Kit (Millipore). RNA was extracted via the Total RNA Isolation Kit (Invitrogen), and qRT-PCR was carried out. The lysates were co-cultured in magnetic beads covering anti-Ago2 and IgG in RIP buffer. Eventually, the relative expression quantity was measured by qRT-PCR.

Approximately 1 × 10⁷ cells were injected into the nude mice subcutaneously. Tumor growth was recorded to measure the width or length, and the volume and weight of the tumor was calculated at indicated times. The experiment was permitted by the ethics of committee.

T24 or HT-1197 cells were place in 6-well plates and incubated on sterilized coverslips. After cells were treated under different conditions separately, then 4% paraformaldehyde was used to fix cells. Afterwards, 0.5% Triton X-100 was used to make cell permeabilization, which lasted for 10 min at room temperature. Subsequently, 5% BSA (Amersco) was employed to block cells in PBST for one hour, which were incubated with primary antibodies against E-cadherin and N-cadherin at 4 °C for one whole night, then cultured with fluorochrome-labeled anti-rabbit secondary antibody (MultiSciences), which lasted for one hour at room temperature. Additionally, DAPI (1:5000, Beyotime) was applied to stain the coverslips, which were observed by a fluorescence microscopy (Nikon). Adobe Photoshop 6.0 software was utilized to merge the images.

Data were showed as mean ± SD. SPSS (Chicago, IL, USA) and GraphPad Prism 5 software (San Diego, CA) were applied to make statistical analysis. Significance of the variance between two or several groups was assessed by Student’s t test or ANOVA. Gene expression correlation was conducted by Pearson’s correlation analysis. P value less than 0.05 had statistically significance. The above experiments were made at least thrice.

To determine the expression status of KCNQ1OT1 in bladder cancer, 70 BC tissue and adjacent normal tissue specimens were gathered together. Then qRT-PCR result revealed that the expression of KCNQ1OT1 was notably up-regulated in BC tissues (Fig. 1a). Additionally, Further, BC patients with advanced stage (III/IV) possessed more expression than that those with early stage (I/II) (Additional file 1: Figure S1). QRT-PCR also manifested an aberrant overexpression of KCNQ1OT1 in BC cell lines (UM-UC-3, T24, HT‐1376 and HT‐1197) comparing with that in SV-HUC-1 cell (Fig. 1b). Hence, we supposed that it might serve as an oncogene in BC.

To verify the hypothesis, loss-of-function experiments were done in T24 and HT‐1197 cell lines, as these two cell lines contained the most obvious expression of KCNQ1OT1 among the above BC cell lines. Firstly, sh-NC, sh-KCNQ1OT1#1 and sh-KCNQ1OT1#2 plasmids were transfected into T24 and HT-1197 to ensure the transfection efficiency. And qRT-PCR detected that sh-KCNQ1OT1#1 exhibiting a better knockdown effect (Fig. 1c). Therefore, it was chosen for all subsequent assays in need for KCNQ1OT1 inhibition plasmid. CCK-8 and colony formation assays were carried out to detect cell proliferation ability. We observed that knockdown of KCNQ1OT1 significantly crippled BC cell viability (Fig. 1d). The decreased number of colonies illustrated that sh-KCNQ1OT1 inhibited cell proliferation (Fig. 1e). TUNEL examined the influence of sh-KCNQ1OT1 in cell apoptosis. After calculating TUNEL positive rate, we found significantly increased TUNEL positive cells caused by KCNQ1OT1 depletion (Fig. 1f). Also, flow cytometry result validated this conclusion (Fig. 1g). Then wound healing and transwell assays were applied to evaluate the effects brought by sh-KCNQ1OT1 on cell migration and invasion, respectively. We observed that knockdown of KCNQ1OT1 shrank wound closure region and decreased migrated cells dramatically (Fig. 1h, i). Moreover, western blot measured relevant protein expressions. According to result, there was an increased level of Bax and cleaved caspase-3/8/9, while a reduction of MMP2, MMP7 and Bcl-2 after silencing KCNQ1OT1 (Fig. 1j). This further proved that KCNQ1OT1 knockdown promoted cell apoptosis, while curbing cell migration. In addition, KCNQ1OT1 deficiency resulted in rising expression of E-cadherin, and declining expression of N-cadherin, which was illustrated by western blot and immunofluorescence staining assays (Additional file 1: Figure S1B-C). From these findings, we preliminarily judged that KCNQ1OT1 is an oncogene in BC progression.

Given that lncRNAs usually regulates cancer development via acting as (microRNAs) miRNA sponge [21]. Accordingly, we browsed in Starbase (http://​starbase.​sysu.​edu.​cn/​) to find combinable miRNAs with KCNQ1OT1. We narrowed scope of the numerous combinable miRNA and found six miRNAs qualified (strict stringency ≥ 5 in Clip data, low stringency in degradome data ≥ 1, and 2 cancer types in Pan-cancer). And qRT-PCR measured that knockdown of KCNQ1OT1 could up-regulate the expression of miR-145-5p most significantly (Fig. 2a), thus we selected miR-145-5p as candidate miRNA for present study. Moreover, significant down-regulated expression of miR-145-5p in BC tissue and cell lines was detected by qRT-PCR (Fig. 2b, c). We also noticed that the expression of miR-145-5p was negatively correlated with that of KCNQ1OT1 from Pearson correlation analysis, which proved the sponge role of KCNQ1OT1 on miR-145-5p (Fig. 2d). Next, subcellular fractionation and FISH assays were employed to determine the location of KCNQ1OT1. Both results manifested that it was principally located in cytoplasm (Fig. 2e). Then, we obtained potential binding sites between miR-145-5p and KCNQ1OT1 from Starbase (Fig. 2f). Dual luciferase reports verified the potential combination between KCNQ1OT1 and miR-145-5p. MiR-145-5p mimics could remarkably attenuate the luciferase activity of KCNQ1OT1 WT compared with negative control groups (Fig. 2g). RNA pull down assay further confirmed this conclusion, as enriched expression of KCNQ1OT1 was pulled down by biotin-miR-145-5p WT group (Fig. 2h). From above findings, we assumed that miR-145-5p could be a tumor suppressor in BC. It was also the upstream gene of KCNQ1OT1. To determine the joint effect on pathological activities, miR-145-5p inhibitor was co-transfected with sh-KCNQ1OT1 in BC cells. Besides, the knockdown efficiency of miR-145-5p inhibitor was ensured by qRT-PCR (Fig. 2i). On one hand, after silencing miR-145-5p, cell viability was significantly increased compared with sh-KCNQ1OT1 group from CCK-8 assay (Fig. 2j). This conclusion was further confirmed by colony formation assay (Fig. 2k). MiR-145-5p inhibitor reversed the pro-apoptosis role of sh-KCNQ1OT1 in BC cells (Fig. 2l). Furthermore, miR-145-5p inhibitor offset the suppressing effects of sh-KCNQ1OT1 on BC cell migration and invasion (Fig. 2m, n). Similarly, western blot assay was measured relevant protein expression. KCNQ1OT1 silencing could increase the level of cell-apoptosis-promoting proteins (Bax and cleaved caspase-3/8/9), but reduced that of cell-apoptosis-suppressing protein (Bcl-2) and cell-migration-promoting proteins (MMP2, MMP7). However, miR-145-5p inhibitor reversed this protein-level-promoting (Bax and cleaved caspase-3/8/9) inhibiting (Bcl-2, MMP2 and MMP7) role of sh-KCNQ1OT1 (Fig. 2o). In a word, KCNQ1OT1 realizes its regulatory effects on BC progression via sponging miR-145-5p.

In present study, we proved that KCNQ1OT1 could bind with miR-145-5p, whether this sponge can affect downstream gene of miR-145-5p still needs to be explored. Hence we screened out twelve qualified messenger RNAs (mRNAs) from Starbase (chose strict stringency ≥ 3 on clip data, high stringency on degradome data, 2 cancer types in Pan-cancer and one program for program number). Then we detected which mRNA expression level was influenced most dramatically by miR-145-5p mimics. And the satisfactory overexpression efficiency of miR-145-5p mimics was ensured by qRT-PCR. Subsequently, qRT-PCR measured that over-expression of miR-145-5p caused significant depletion of PCBP2 only, which suggested that PCBP2 could be the most potential target gene (Fig. 3a). To verify this, qRT-PCR detected the relative expression of PCBP2 in BC tissue and cell lines later, and PCBP2 exhibited obviously high expression (Fig. 3b, c). QRT-PCR analysis also showed that knockdown of KCNQ1OT1 can down-regulate the expression of PCBP2 (Fig. 3d). Given this phenomenon, we anticipated that KCNQ1OT1 could regulate the expression of PCBP2 via miR-145-5p. To prove this assumption, we examined whether a chemical interaction existed between miR-145-5p and PCBP2. Hence, the Starbase-predicted binding sites between them were verified by dual luciferase assays result (Fig. 3e, f). We observed miR-145-5p mimics dramatically impaired the luciferase activity of PCBP2 3′-UTR WT compared with negative control groups. After further verification of RIP assay, we obtained the conclusion that PCBP2 physically bound with miR-145-5p (Fig. 3g). Therefore, we proved that KCNQ1OT1 overexpression could upregulate PCBP2 expression via sequestering miR-145-5p in BC.

To elucidate the cellular function of PCBP2, loss-of-function experiments were conducted in T24 and HT-1197 cell lines. The knockdown efficiency of sh-PCBP2 was guaranteed by qRT-PCR assay (Fig. 3h). According to the results of functional experiment, knockdown of PCBP2 repressed cell proliferation, migration and invasion ability, while enhancing cell apoptosis (Fig. 3i–m). Additionally, western blot tested the expression of proteins participating in cell migration and apoptosis, which confirmed the pro-apoptosis and anti-migration role of sh-PCBP2 in BC cells (Fig. 3n). In a summary, PCBP2, the target gene of miR-145-5p, serves as an oncogene in BC progression.

To understand the effect of PCBP2 expression on regulatory mechanism of KCNQ1OT1 in BC, rescue experiments were conducted. The transfection efficiency of pcDNA3.1/PCBP2 was guaranteed by qRT-PCR (Fig. 4a). According to the results of CCK-8, up-regulation of PCBP2 reversed the suppression on cell viability of sh-KCNQ1OT1 (Fig. 4b). In addition, TUNEL assay examined that pcDNA3.1/PCBP2 co-transfected with sh-KCNQ1OT1 could greatly reduce the number of apoptotic cells compared with group treated with sh-KCNQ1OT1 (Fig. 4c). Besides, wound healing and transwell assays measured cell migration and invasion, separately. The wound closure rate of sh-KCNQ1OT1 was obviously lower than negative control group, while pcDNA3.1/PCBP2 remedied this inhibiting effect on cell migration (Fig. 4d). Transwell assay showed that invaded cells decreased remarkably after sh-KCNQ1OT1 transfection compared with negative control group. However, when co-transfecting with pcDNA3.1/PCBP2, the invaded cell number recovered noticeably again (Fig. 4e). The relative proteins expression detected by western blot assay further validate above conclusion (Fig. 4f). In general, PCBP2 upregulation can rescue tumor-inhibitor role of KCNQ1OT1 depletion in BC.

Xenografts with sh-KCNQ1OT1#1 subcutaneously injected on nude mice slowed down the tumor growth compared with negative control (Fig. 5a). The xenograft tumor volume was measured after 4 weeks of injection, and the tumor volume was smaller in sh-KCNQ1OT1 group compared with group injected with sh-negative control (NC) group (Fig. 5b). The tumor weight was also found lighter in sh-KCNQ1OT1 group compared with NC group at the last day of inoculation (Fig. 5c). Western blot also measured that KCNQ1OT1 knockdown inhibited the expression of N-cadherin and Bcl-2, but stimulated that of E-cadherin and Bax in BC tissues (Fig. 5d). These experiments in vivo further proved that KCNQ1OT1 plays an oncogenic role in BC progression.