LncRNA GAS5 regulates the Wnt/β-catenin pathway through the miR-18a-5p/AXIN2/GSK3β axis to inhibit the proliferation and migration of bladder cancer cells

Abstract

LncRNA growth arrest specific 5 (GAS5) has been confirmed to play an essential role in a number of biological processes, such as tumor regulation and gene transcription. GAS5 has been shown to be a tumor suppressor gene in many types of cancer, but its specific mechanism of action in bladder cancer (BC) remains to be elucidated. In this study, we explored the biological properties of GAS5 in BC and its mechanism of action in BC. We analyzed the expression of GAS5 in 50 pairs of BC tissues and found that GAS5 was low expressed in BC tissues compared with normal mucosal tissues. In vitro and in vivo experiments showed that GAS5 could affect the proliferation and migration of BC cells. Nucleoplasmic isolation assays and fluorescence in situ hybridization (FISH) assays demonstrated the localization of GAS5 in cell cytoplasm. Chromatin isolation by RNA purification (ChIRP), RNA immunoprecipitation (RIP) and luciferase assay demonstrated the target binding relationship of GAS5 with miR-18a-5p. Rescue experiments demonstrated that GAS5 promoted the proliferation and migration of BC cells through target binding of miR-18a-5p. Moreover, miR-18a-5p bound to its targets AXIN2 and GSK3β, which in turn affected the expression of Wnt/β-catenin pathway-related proteins. Our findings demonstrate that GAS5 regulates Wnt/β-catenin pathway activity by regulating the miR-18a-5p/AXIN2/GSK3β axis to modulate BC progression, providing a new potential therapeutic strategy for the treatment of BC.

Graphical abstract

Graphical Abstract.

Introduction

Bladder cancer (BC) is a common malignancy worldwide, with an estimated 380 000 new cases and 15 000 related deaths each year (1). BC is mainly classified into non-muscle-invasive bladder cancer (NMIBC) and muscle-invasive bladder cancer (MIBC) based on pathology reports (2). NMIBC (stage Ta, T1) and MIBC (stage T2 or above) have a significant impact on the prognosis of patients with bladder cancer (BC) due to the different depths of invasion (3); therefore, the treatment varies greatly. The incidence and mortality rates of BC have been increasing rapidly in recent years. BC is usually characterized by a high recurrence rate and the ability to invade and migrate, resulting in a poor prognosis (4). It has been found that the different phenotypes of patients with low-grade NMIBC, highly malignant MIBC, or advanced BC are based on genetic alterations in DNA and subsequent RNA expression level changes, creating distinct molecular subtypes with prognostic, predictive, and therapeutic implications (5). Therefore, exploring the molecular mechanisms of BC progression will help to identify potential therapeutic targets and develop more effective molecular biomarkers.

Non-coding RNAs (ncRNAs) were initially thought to be transcriptional noise; however, it is becoming increasingly apparent that they may play critical roles in a variety of cellular processes ranging from normal development to disease processes (6). Based on their size, ncRNAs include small RNAs (miRNAs) and long non-coding RNAs (lncRNAs), where lncRNAs are greater than 200 nt in length (7). Gene expression regulation by lncRNAs is complex and often involves multiple mechanisms (8). LncRNAs exert biological functions mainly through three modes of action: (i) RNAs are directly involved in regulation as functional molecules, interacting with other molecules in the cell, such as DNA, proteins or RNA; (ii) lncRNAs bind to gene regulatory elements, and the activity of lncRNA genes directly affects the activity of regulatory elements; and (iii) lncRNAs in the transcription process affect the genome and thus gene activity. Recent studies have shown that the sub-localization of lncRNAs in the cell has a great impact on the function of lncRNAs. By interacting with different functional molecules and targets of action in different subcellular microenvironments, the same lncRNA can have different functions, and thus, lncRNAs act differently in the nucleus than in the cytoplasm (9). LncRNA growth arrest specific 5 (GAS5), a classical oncogene suppressor, has been demonstrated in numerous studies to be expressed at low levels in a variety of cancer types (10).

In this study, we aimed to investigate the molecular mechanisms of the role of GAS5 in the development and progression of BC. We collected BC tissue samples for analysis, and we found that GAS5 expression was downregulated in 50 pairs of BC tissues versus normal tissues. Both in vitro and in vivo experiments demonstrated that GAS5 plays an inhibitory role in the biological progression of BC. In in vitro experiments, we found that GAS5 inhibited BC progression through the miR-18a-5p/AXIN2/GSK3β axis by bioinformatics analysis, AGO2 RNA immunoprecipitation (RIP), chromatin isolation by RNA purification (ChIRP) and dual-luciferase reporter assays. Our study may provide new insights and a potential therapeutic target for lncRNA studies in the context of BC development.

Methods

Clinical samples

Fifty pairs of tissue samples (tumor tissue and normal bladder mucosa tissue adjacent to cancer) were obtained from the First Affiliated Hospital of Wannan Medical College (Yiji Shan Hospital) via total cystectomy from 2019 to 2021. All normal bladder mucosa tissue specimens were at least 3 cm from the tumor. All samples were confirmed as normal bladder epithelial tissue or BC tissue by the pathology department. All patients provided written informed consent to participate in the study. The ethics committee of the First Affiliated Hospital of Wannan Medical College approved the use of these tissues for the study.

Cell culture

SV-HUC-1, EJ, T24, BIU-87 and 5637 cells were purchased from the Shanghai Cell Bank of the Chinese Academy of Sciences. Cell lines were characterized and authenticated by the supplier. F-12 K medium (Gibco, United States) and RPMI-1640 medium (HyClone, China) were supplemented with 10% fetal bovine serum (FBS, Yeasen, China) and 1% penicillin–streptomycin solution (Gibco) to form complete media, which were then used to culture normal human urothelial cells and BC cell lines, respectively. Cell culture was performed under standard cell culture conditions: 37°C, 5% CO₂.

Cytoplasmic and nuclear RNA fractionation

Nuclear-cytoplasmic isolation was performed using nuclear and cytoplasmic extraction reagents (Thermo Fisher, CA, USA), and RNA was extracted for RT–PCR. GAPDH and U6 were used as positive controls for cytoplasmic and nuclear RNA, respectively.

Fluorescent in situ hybridization (FISH)

We inoculated appropriate amounts of cells into confocal dishes and incubated them for 24 h. After cell fixation, we added pre-cooled permeabilization solution for 5 min. Next, we performed a PBS wash, added pre-hybridization solution, and incubated the sample for 30 min. We then preheated the hybridization solution at 37°C and added lncRNA FISH Probe Mix storage solution. We discarded the prehybridization solution, added the mixed hybridization solution, and allowed the samples to hybridize at 37°C overnight in the dark. The cells were washed with hybridization wash at 42°C for 5 min at room temperature and PBS for 5 min at room temperature. LncRNA distribution was examined using laser confocal microscopy. GAS5, 18S RNA and U6 RNA probes were synthesized by RiboBio. RNA localization was detected using the Ribo^TM Fluorescent In Situ Hybridization Kit (RiboBio, Guangzhou, China). DAPI was used for cell staining.

Plasmid construction and cell transfection assay

A GAS5 over-expression plasmid, null control, small interfering RNA (siRNA) for GAS5 and negative control (siNC) were designed by GenePharma (Shanghai, China). The siRNA sequences are shown in Table 1 of the Supplementary File, available at Carcinogenesis Online. We stably overexpressed GAS5 in T24 cells by transducing the cells with lentivirus (GenePharma, Shanghai, China). The cells were screened with puromycin for 4 weeks. Stable transfectants were used for animal experiments. The miR-18a-5p mimic, mimic control, miR-18a-5p inhibitor and inhibitor control were purchased from RiboBio (Guangzhou, China). Transfection of T24 and EJ cells was performed with Lipo8000™ transfection reagent (Biyuntian Biotechnology, Shanghai, China). After 48 h of transfection, T24 and EJ cells were collected for further experiments.

Real-time quantitative polymerase chain reaction (RT–qPCR)

Cells and tissues were lysed using TRIzol (Invitrogen) to obtain total RNA. A GAS5 cDNA reverse transcription kit (RR037A) and real-time PCR kit (RR820B) were obtained from TaKaRa. The miRNA was reverse transcribed using the miRcute miRNA First Strand cDNA Synthesis Kit (Tiangen Biotech). miRNA was subsequently amplified by a SYBR Green miRNA qPCR detection kit (Tiangen Biotech). mRNA reversal was performed using the SuperMix for qPCR kit from Yeasen Biotechnology (China), followed by mRNA PCR amplification using the qPCR SYBR Green Master Mix kit (Yeasen). GAS5, GSK3β and GAPDH primers were synthesized by Biotech (Biotech, China). GAPDH and U6 were used as normalization controls for quantification. The 2^−ΔΔCt method was used to assess gene expression levels. The primer sequences are shown in Table 1 of the Supplementary file, available at Carcinogenesis Online.

Western blotting

RIPA lysis buffer (Biyuntian) was used to extract the total protein of the cells and tissues. Electrophoretic separation was performed using a 10% polyacrylamide gel (Cat# PG112) provided by EpiZyme Biotechnology, and the proteins were transferred onto polyvinylidene difluoride (PVDF) membranes (Invitrogen, CA). After blocking with skim milk for 2 h, the membrane was incubated with anti-GSK3β (ab32391, 1:500, Abcam), anti-AXIN2 (ab109307, 1:150, Abcam) anti-c-myc (ab32072, 1:1000, Abcam), anti-β-catenin (ab32572, 1:10000, Abcam) and anti-β-actin (1:5000, Proteintech) antibodies at 4°C overnight. After washing the membrane three times with TBST, the membrane was incubated with the secondary antibody at room temperature for 1 h, and the bands were detected with an enhanced chemiluminescence (ECL) system. All antibodies were diluted with TBST.

EdU assay for proliferation

For EdU labeling (Hoechst staining), an EdU labeling kit purchased from Guangzhou RiboBio Co., Ltd. (Guangzhou, China) was used. After 24 h of T24 and EJ transfection, cells were collected. After resuspend by complete media, the cells were inoculated into a 96-well plate (4 × 10³/well). After the cells adhered to the wells (24 h), 100 µL of RPMI-1640 containing 50 μM EdU was added to each well and then incubated in an incubator at 37°C for 4 h. The subsequent EdU labeling step was carried out in accordance with the instructions provided by the EdU kit.

Colony formation assay

After 24 h of T24 and EJ transfection, cells were collected, inoculated into six-well plates (500 cells per well) and cultured in DMEM with 10% FBS. The culture was terminated when the cells formed colonies of more than 50 cells. Then, the cells were washed three times with PBS. Fixed cells were incubated with methanol (Beyotime) and stained with 0.1% crystal violet (Invitrogen) for 20 min. Colonies with >50 cells were counted, and the experiment was repeated three times.

Wound healing assay

Transfected T24 and EJ cells were cultured in six-well plates, and when the cells were confluent and had formed a monolayer of cells, the cell layer was scraped vertically with the fine end of a 200-µL pipette tip and washed three times with PBS to ensure that the suspended cells were removed; then, fresh 1640 medium containing 10% FBS was added. For the wound healing experiments, the cells were photographed with an inverted microscope at 0 h and 24 h after injury. Remodeling was measured as the decrease in distance across the induced injury. Wound healing rate = (0-h wound area − 24-h wound area)/0-h wound area × 100%.

Transwell cell migration assay

The migration ability of T24 and EJ cells after transfection was tested in a Transwell chamber (8 µm, Millipore, Boston, MA, USA). After 48 h of T24 and EJ transfection, cells were collected, RPMI-1640 was used to resuspend the cells, and 200 µL of cell suspension was inoculated into the chamber (1 × 10⁴/well). Then, 500 µL of complete medium (10% FBS) was added to the lower chamber. After culturing for 24 h, the cells were fixed in 4% paraformaldehyde for 30 min, stained with 0.1% crystal violet (crystal, Sigma–Aldrich, USA) for 15 min, dried thoroughly and observed under a microscope.

Luciferase reporter assay

All dual-luciferase experiments were performed in 293T cell lines. GAS5-WT/GAS5-Mut (GenePharma), GSK3β-WT/GSK3β-Mut (GenePharma) and AXIN2-WT/AXIN2-Mut (GenePharma) were cotransfected with pmirGLO as a vector and miR-18a-5p mimics/mimic NC. After 48 h of transfection, the assay was performed using the Dual-Luciferase Reporter Gene Assay Kit (GenePharma) according to the kit instructions.

ChIRP assay with a biotinylated GAS5 probe

The ChIRP kit was provided by Saicheng Biotech (Guangzhou, China). The probes and operation methods for this assay were provided by RiboBio (China). Cells were cultured in tissue culture plates or flasks until they reached confluence. The cells were rinsed once with phosphate-buffered saline (PBS), and trypsin digestion was performed. Medium was added to terminate digestion, and the cells were resuspended to form a single-cell suspension. Twenty million cells were required for each pull-down sample. The cells from step one were crosslinked with glutaraldehyde to maintain RNA-chromatin interactions, and cell precipitation was performed. The crosslinked cells were lysed, and cell lysates were prepared. The crosslinked cell lysates were sonicated to shear the DNA. Then, the biotinylated DNA probe was hybridized to the RNA, and the bound chromatin was isolated. RNA fragments were extracted from pull-down samples and quantified by qRT–PCR. The amplified DNA products were subjected to 1.5% agarose gel electrophoresis.

RNA immunoprecipitation (RIP)

A total of 1 × 10⁷ cells were washed with 1 mL of PBS, and the cells were collected by centrifugation at 800–1000g for 5 min at room temperature. The cells were lysed by adding cell lysis solution for 10 min (1% protease inhibitor and 1% RNase inhibitor were added before use), and the supernatant was removed after centrifugation for 10 min (4°C, 14 000g). One hundred microliters of supernatant was used as an input control, and the remaining 900 μL was placed in RNase-free centrifuge tubes. The samples and beads were pretreated. The pretreated beads were divided into two tubes, 1 mL of buffer was added, and 5 μg of RIP antibody (Abcam, ab186733) and IgG (Abcam, ab172730) was added. Next, 350 μL of buffer and 450 μL of sample were added to the magnetic beads conjugated with RIP antibody or IgG, and the sample was rotated for 2 h (4°C, 10 rpm). The RNA was extracted after 5 repeated washes using wash buffer.

Immunohistochemistry (IHC)

Human tumor tissue and mouse transplanted tumor tissues were excised, fixed with tissue fixative and routinely embedded in paraffin; sectioning and immunohistochemical staining were performed by Servicebio (Wuhan, China). Immunostaining scan images were also provided by Servicebio. At least three representative stained areas were selected for each IHC-stained section and analyzed at ×400 magnification. The results were calculated based on a randomly selected field of view. The integrated optical density (IOD) value for each slice was measured using Image-Pro Plus 6.0, and the average density (IOD/AREA) was calculated based on this value.

Tumor xenografts in nude mice

The study was approved by the Animal Ethics Committee of Wannan Medical College. Four-week-old female thymus-free BALB/c nude mice were randomly divided into the following four groups with three mice in each group to examine tumorigenicity: (i) agomir control group, (ii) GAS5 overexpression + agomir control group, (iii) miR-18a-5p agomir group, and (iv) GAS5 overexpression + miR-18a-5p agomir group. First, T24 cells (stably overexpressing GAS5 or empty vector) were transfected with miR-18a-5p agomir (GenePharma, China) or control agomir (GenePharma) for 48 h. After treatment, the cells were collected and resuspended in PBS, and a single-cell suspension of 2 × 10⁶ cells was inoculated subcutaneously into the right axilla of mice. In addition, to ensure the expression of miR-18a-5p in tumors, we also performed experiments in which agomir-18a-5p NC/agomir-18a-5p was injected directly into implanted tumors at a dose of 10 nmol per mouse (in 20 μL of PBS) twice. The width (W) and length (L) of the tumors were measured every 7 days. Tumor volume was expressed as L × W²/2. The mice were killed after 4 weeks. All tumors were debrided, weighed, and fixed.

Bioinformatics analysis

The relevant comprehensive database we used was The Cancer Genome Atlas (TCGA) database (https://portal.gdc.cancer.gov/), and we used the TCGA database to analyze the expression of miR-18b-5p in BC tissues, as well as receiver operating characteristic (ROC) curves. We used the ENCORI (https://starbase.sysu.edu.cn), miRcode (http://www.mircode.org/), and lncBase (http://www.microrna.gr/LncBase) databases to predict the downstream target genes of GAS5. The ENCORI database was used to predict relevant pathways downstream of miR-18a-5p.

Statistical analysis

For analysis of the results, data are expressed as the mean ± SD unless otherwise stated; Student’s t test and χ² test were used to compare means between two groups, and P < 0.05 was considered to indicate a significant difference. All experiments were repeated more than three times.

Results

LncRNA GAS5 is expressed at low levels in BC, and can regulate the ability of BC to proliferate and migrate

GAS5 has been previously reported to be a cancer suppressor gene in several cancer types. To verify its role in BC, we examined the expression of GAS5 in 50 pairs of BC tissues and found that GAS5 was expressed at low levels in BC (Figure 1A). To further verify this finding, we examined GAS5 expression in several BC cell lines and found that GAS5 was expressed at lower levels in BC cell lines than in normal bladder epithelial SV-HUC-1 cells (Figure 1B), suggesting the possibility that the GAS5 gene acts as a tumor suppressor gene in BC. Through nucleoplasmic separation experiments, we found that GAS5 is mainly present in the cytoplasm (Figure 1C), and we used FISH experiments to further verify the localization of GAS5 in the cell (Figure 1D). This information suggests that GAS5 exerts regulatory effects, probably mainly through a competing endogenous RNA (ceRNA) mechanism. Statistical results of clinical information showed a significant correlation between GAS5 expression levels and the pathological grading, staging and lymphatic metastasis of clinical patients. Decreased expression levels of GAS5 promote BC progression in clinical patients (Table 1).

Figure 1.

LncRNA GAS5 is expressed at low levels in BC, and can regulate the ability of BC to proliferate and migrate. (A) The expression of lncRNA GAS5 was detected by real-time PCR in 50 pairs of BC and adjacent normal tissues. GAPDH was used as a control. Data are the mean ± SEM. (n = 3, Student’s t test). (B) The expression of lncRNA GAS5 was tested by real-time PCR in SV-HUC-1, BIU-87, 5637, T24 and EJ cells. GAPDH was used as a control. Data are the mean ± SEM, n = 3. (C) The relative expression of nuclear and cytosolic lncRNA GAS5. D RNA-FISH images showing that lncRNA GAS5 was predominantly localized in the cytoplasm. 18S was mainly localized in the cytoplasm and was used as a positive control. circHIPK3 and 18S probes were labeled with Cy3, and nuclei were stained with DAPI. *P < 0.05, **P < 0.01. (E, F) The EdU assay showed that lncRNA GAS5 inhibited the proliferation of T24/EJ cells. G-H Transwell assays showed that lncRNA GAS5 inhibited the migration of T24 and EJ cells. **P < 0.01.

Table 1.

Clinicopathological features of 50 bladder cancer patients and the expression of lncRNA GAS5 and miR-18a-5p

Parameters	Group	Cases	lncRNA GAS5 expression				P-value	miR-18a-5p expression				P-value
			Low	%	High	%		Low	%	High	%
Gender	Male	32	27	84	5	16	0.66	7	22	25	78	0.639
	Female	18	16	89	2	11		5	28	13	72
Age at surgery	<60	22	18	82	4	18	0.776	5	23	17	77	0.852
	≥60	28	22	79	6	21		7	25	21	75
Pathological	pTa-T1	12	4	33	8	67	0.000	3	25	9	75	0.000
	pT2-T4	38	36	95	2	5		9	24	29	76
Grade	Low	14	7	50	7	50	0.001	8	57	6	43	0.001
	High	36	33	92	3	8		4	11	32	89
Lymph node metastasis	Absent	30	21	70	9	30	0.03	11	37	19	63	0.01
	Present	20	19	95	1	5		1	5	19	95
Total		50	40		10			12		38

To investigate the role of GAS5 in BC, we designed a GAS5 over-expression plasmid and small interfering RNA. To ensure that the siRNA successfully knocked down the expression of GAS5, we designed two siRNAs (siGAS5-1 and siGAS5-2), and to prevent off-target effects, we mixed siGAS5-1 and siGAS5-2 for GAS5 interference, denoted as siGAS5, in our experiments. After testing the transfection efficiency (Supplementary Figure 1A, B is available at Carcinogenesis Online), we performed a series of functional experiments. EdU assays and colony formation assays revealed that overexpression of GAS5 inhibited the proliferation of BC cells; in contrast, knocking down GAS5 expression promoted the proliferation of BC cells (Figure 1E, F and Supplementary Figure 1C is available at Carcinogenesis Online). Transwell and wound healing assays revealed that GAS5 overexpression suppressed the migration ability of BC cells, while knocking down GAS5 expression improved the migration ability of BC cells (Figure 1G, H and Supplementary Figure 1D is available at Carcinogenesis Online). This finding demonstrates that GAS5 regulates the progression of BC as a cancer suppressor gene in BC.

LncRNA GAS5 inhibits miR-18a-5p expression through direct interactions

To investigate the mechanism of action of GAS5 in BC in-depth, we obtained nine downstream miRNAs of GAS5 using the ENCORI, miRcode, and lncBase databases (Figure 2A). To screen the downstream miRNAs, we designed ChIRP probes and validated the specificity and efficiency of the probes in the T24 and EJ cell lines (Figure 2B, C). We mixed the probes to prevent off-target effects. The ChIRP results showed that miR-18a-5p and miR-18b-5p had the highest pull-down ploidy (Figure 2D, E). Furthermore, the RT–PCR results showed that only miR-18a-5p and miR-18b-5p were more highly expressed in BC cells (EJ and T24) than in SV-HUC-1 cells, while the remaining seven miRNAs were expressed at low levels (Figure 2F). Therefore, we selected miR-18a-5p and miR-18b-5p for further study.

Figure 2.

LncRNA GAS5 inhibits miR-18a-5p expression through direct interactions. (A) Prediction of downstream genes of lncRNA GAS5 using the ENCORI database, lncBase database, and MiRcode database. (B, C) Expression of lncRNA GAS5 detected by qRT–PCR analysis and agarose gel electrophoresis of samples obtained by RNA pull-down assay. GAPDH was used as a control. (D, E) The expression of miRNAs in the biotinylated lncRNA GAS5 probe group and the control probe group was detected by qRT–PCR. U6 was used as a control. F qRT–PCR to detect the differential expression of genes downstream of miR-18a-5p in cell lines. U6 was used as a control. (G, H) Luciferase reporter assay to detect the interaction between lncRNA GAS5 and miR-18a-5p in T24 cells. I A RIP assay was performed to confirm the interaction between lncRNA GAS5 and AGO2, miR-18a-5p and AGO2, AXIN2 and AGO2, and GSK3β and AGO2 in the T24 cell line. **P < 0.01.

We found that miR-18b-5p was highly expressed in EJ, T24, BIU-87 and 5637 cells (Supplementary Figure 2A is available at Carcinogenesis Online). We also found that miR-18b-5p was highly expressed in BC tissues in the BioSign database (Supplementary Figure 2B is available at Carcinogenesis Online). We performed analysis using the TCGA database and found that the ROC curve of miR-18b-5p had an area under the curve (AUC) value of 0.765, indicating that miR-18b-5p may have the ability to act as a molecular marker (Supplementary Figure 2C is available at Carcinogenesis Online). To investigate the role of miR-18b-5p in BC, after successful overexpression of miR-18b-5p (Supplementary Figure 2D is available at Carcinogenesis Online), we examined the effect of miR-18b-5p on the proliferation ability of BC cells using an EdU assay (Supplementary Figure 2E is available at Carcinogenesis Online) and the effect of miR-18b-5p on the migration ability of BC cells using a Transwell assay (Supplementary Figure 2F is available at Carcinogenesis Online). MiR-18b-5p over-expression did not have a significant effect on the proliferation or migration of BC cells. Therefore, we focused our research on miR-18a-5p. Luciferase experiments showed that GAS5 and miR-18a-5p have a direct regulatory relationship (Figure 2G, H). Next, a RIP assay of Ago2 showed that GAS5 and miR-18a-5p were mainly enriched on Ago2, indicating that GAS5 was recruited to RISC associated with Ago2 to interact with miR-18a-5p (Figure 2I).

MiR-18a-5p acts as an oncogene regulating BC cell proliferation and migration

To investigate the role of miR-18a-5p in BC, we examined the expression of miR-18a-5p in 50 pairs of BC tissues and found that miR-18a-5p was highly expressed in BC tissues (Figure 3A). Observation of changes in cell proliferation and migration ability after miR-18a-5p overexpression in BC cells (Figure 3B). EdU and colony formation assays confirmed that miR-18a-5p overexpression could enhance the proliferation of BC cells, and knockdown of miR-18a-5p expression induced the opposite effects (Figure 3C, D and Supplementary Figure 3A is available at Carcinogenesis Online). It was demonstrated by wound healing and Transwell assays that miR-18a-5p overexpression enhanced the migration ability of BC cells (Figure 3E, F and Supplementary Figure 3B is available at Carcinogenesis Online). Statistical results of clinical information showed a significant correlation between miR-18a-5p expression levels and the pathological grading, staging and lymphatic metastasis of clinical patients. Elevated expression levels of miR-18a-5p promote the progression of BC in clinical patients (Table 1).

Figure 3.

MiR-18a-5p acts as an oncogene regulating BC cell proliferation and migration. (A) qRT–PCR revealed that miR-18a-5p was upregulated in BC tissues (n = 50). The error bars represent the SEM (n = 3, Student’s t test). U6 was used as a control. B miR-18a-5p was over-expressed using miRNA mimic transfection. U6 was used as a control. (C, D) EdU assay to detect the effect of miR-18a-5p on the proliferative capacity of T24 and EJ cells. (E, F) Transwell assay to detect the effect of miR-18a-5p on the migration ability of T24 and EJ cells. **P < 0.01.

MiR-18a-5p targets AXIN2 and GSK3β and regulates Wnt/β-catenin pathway activity

We used the ENCORI BioSign database to explore miR-18a-5p-related pathways and found that miR-18a-5p is closely related to the Wnt/β-catenin pathway (Figure 4A). MiRNAs generally act by binding to the 3ʹ end of mRNAs, thereby inhibiting mRNA expression. Therefore, we used the BioSign database to predict whether miR-18a-5p can bind to key proteins of the Wnt/β-catenin pathway and thus affect the activity of the Wnt/β-catenin pathway. The database results showed that miR-18a-5p could bind to AXIN2 and GSK3β (Supplementary Figure 4A is available at Carcinogenesis Online). We assayed the differences in AXIN2 and GSK3β expression in normal bladder epithelial cells and BC cells, as well as the differences in Wnt/β-catenin pathway activity, and showed that AXIN2 and GSK3β expression levels were significantly lower and that the Wnt/β-catenin pathway was activated in BC cells (Supplementary Figure 5A is available at Carcinogenesis Online). We found that AXIN2 and GSK3β expression levels were significantly higher in the paracancerous tissues of BC patients than in BC tissues by immunohistochemistry (Supplementary Figure 5B is available at Carcinogenesis Online). To further confirm this relationship, we designed dual-luciferase assays, and the results demonstrated that miR-18a-5p can bind directly to AXIN2 and GSK3β (Figure 4B–E). We overexpressed and knocked down miR-18a-5p in the T24 and EJ cell lines and detected the expression of AXIN2 and GSK3β. The results showed that overexpression of miR-18a-5p inhibited the expression of AXIN2 and GSK3β; conversely, knockdown of miR-18a-5p promoted the expression of AXIN2 and GSK3β (Figure 4F, G). We obtained the same results using Western blotting (Figure 4J, Supplementary Figure 6A is available at Carcinogenesis Online). We further explored the relationships among GAS5, AXIN2 and GSK3β and found that overexpression of GAS5 led to elevated AXIN2 and GSK3β expression; conversely, GAS5 knockdown suppressed AXIN2 and GSK3β expression (Figure 4H, I). We verified the results using Western blotting and reached the same conclusion (Figure 4K, Supplementary Figure 6B is available at Carcinogenesis Online).

Figure 4.

MiR-18a-5p targets AXIN2 and GSK3β and regulates Wnt/β-catenin pathway activity. (A) The ENCORI BioSign database predicts miRNA-18a-5p-mediated downstream pathways. (B, C) Designing a luciferase assay to detect the relationship between miR-18a-5p and the regulation of AXIN2 and GSK3β. (D, E) The luciferase assay confirms the direct regulation by miR-18a-5p of AXIN2 and GSK3β. (F, G) Variation in the expression of AXIN2 and GSK3β after overexpression and interference with miR-18a-5p was detected by qRT–PCR. GAPDH was used as a control. (H, I) Changes in AXIN2 and GSK3β expression after overexpression and knockdown of lncRNA GAS5 were detected by qRT–PCR. GAPDH was used as a control. (J) Over-expression of miR-18a-5p in T24 cell lines downregulated AXIN2 and GSK3β expression and activated Wnt/β-catenin pathway activity, while interference with miR-18a-5p expression had the opposite effect. (K) Overexpression of lncRNA GAS5 in the T24 cell line upregulated AXIN2 and GSK3β expression and inhibited Wnt/β-catenin pathway activity, while interference with lncRNA GAS5 expression had the opposite effect. *P < 0.05, **P < 0.01.

GAS5 regulates Wnt/β-catenin signaling through the GAS5/miR-18a-5p/AXIN2/GSK3β axis

We designed a rescue assay to further validate the regulatory relationship of the GAS5/miR-18a-5p/AXIN2/GSK3β axis in BC. EdU (Figure 5A) and Transwell (Figure 5B) assays demonstrated that simultaneous transfection of GAS5 and miR-18a-5p reversed the increase in proliferation and migration capacity caused by over-expression of miR-18a-5p. Western blotting assays also demonstrated that simultaneous transfection of GAS5 and miR-18a-5p could reverse the activation of the Wnt/β-catenin pathway induced by the overexpression of miR-18a-5p (Figure 5C, D).

Figure 5.

LncRNA GAS5 regulates Wnt/β-catenin signaling through the GAS5/miR-18a-5p/AXIN2/GSK3β axis. (A) LncRNA GAS5 counteracted the enhanced effect of the miR-18a-5p mimics on cell proliferation. (B) LncRNA GAS5 counteracted the enhanced effect of the miR-18a-5p mimics on cell migration. (C) In BC cell lines, lncRNA GAS5 counteracted the downregulation of AXIN2 and GSK3β by miR-18a-5p while counteracting the enhanced Wnt/β-catenin pathway activity caused by miR-18a-5p. **P < 0.01.

GAS5 reversed the activation of the Wnt/β-catenin pathway caused by miR-18a-5p in vivo

In vitro experiments demonstrated the regulatory relationship of the GAS5/miR-18a-5p/AXIN2/GSK3β axis in BC. We designed in vivo experiments to further confirm its function. The results showed that miR-18a-5p overexpression significantly increased tumor volume (Figure 6A) and weight (Figure 6B) compared to those of controls, while simultaneous over-expression of GAS5 and miR-18a-5p reversed the increase in tumor volume and weight caused by miR-18a-5p. We collected tissues, extracted proteins, examined Wnt/β-catenin pathway-related protein expression by Western blotting assay, and found that simultaneous overexpression of GAS5 and miR-18a-5p in vivo also reversed the activation of the Wnt/β-catenin pathway caused by overexpression of miR-18a-5p (Figure 6C). We subjected some tissues from the four groups of tumors to IHC, and the results showed that overexpression of GAS5 reversed the decreased expression levels of GSK3β and AXIN2 as well as the increased expression levels of Ki67, E-cadherin and N-cadherin caused by miR-18a-5p over-expression (Figure 6D, Supplementary Figure 7A–E is available at Carcinogenesis Online). It was shown that over-expression of GAS5 in BC could inhibit activation of the Wnt/β-catenin pathway and thus suppress the proliferation and migration abilities of BC cells.

Figure 6.

LncRNA GAS5 reversed the activation of the Wnt/β-catenin pathway caused by miR-18a-5p in vivo. (A) Tumor growth curves of different groups of mice. (B) Tumor weight statistics in mice. (C) Western blot of mouse tumor tissues showed that lncRNA GAS5 counteracted the downregulation of AXIN2 and GSK3β by miR-18a-5p while decreasing the enhanced Wnt/β-catenin pathway activity induced by miR-18a-5p. (D) Immunohistochemical staining for AXIN2, GSK3β, Ki67, E-cadherin and N-cadherin in mouse tumor tissue. **P < 0.01.

Discussion

Bladder cancer is a common malignancy of the urinary system (11,12). BC is usually characterized by a high recurrence rate, invasiveness and migration ability, leading to a poor prognosis (4). Approximately 80% of BC cases are NMIBCs (13). Most NMIBCs can be detected and treated at an early stage, and their 5-year survival rate exceeds 90% (14). However, approximately 70–80% of patients with NMIBC develop tumor recurrence after the first surgical treatment, and 10–20% of patients develop MIBC (15). For patients with MIBC, the 5-year survival rate is less than 50% and drops to 20% if metastasis occurs (16). For non-metastatic MIBC, radical cystectomy is the gold standard for treatment. For patients with metastatic BC, gemcitabine (GEM) and cisplatin (CDDP) [GC] combined with chemotherapy are effective first-line treatments (17). The incidence and mortality rates of BC have been increasing rapidly. Therefore, an in-depth study of the molecular mechanisms of bladder tumor progression is closely related to improving the prognosis of patients (18).

Accumulating studies suggest that lncRNAs are involved in the formation and progression of multiple cancers (19). LncRNAs can regulate cancer development at different levels (8). For instance, lncRNA BCRT1 promotes breast cancer progression by targeting the miR-1303/PTBP3 axis (20). LncRNA PTENP1 inhibits cervical cancer progression by suppressing miR-106b (21). GAS5 is a lncRNA located on chromosome 1q25 and plays an important role as a tumor suppressor (22). For example, GAS5 regulates ovarian cancer progression via miR-196-5p/HOXA5 (23). GAS5 regulates ABCB1-mediated adriamycin resistance in breast cancer through the Wnt/β-catenin signaling pathway (24). GAS5 has also been reported in BC, and GAS5 promotes apoptosis in BC cells by inhibiting EZH2 transcription (25). GAS5 regulates the proliferative capacity of BC cells by regulating CDK6 (26) as well as the miR-21/PTEN axis (27). GAS5 may serve as an independent molecular cancer biomarker for short-term recurrence and progression in BC patients. We demonstrated that GAS5, which can act as an endogenous sponge of miR-18a-5p, exerts oncogenic effects in BC. MiR-18a-5p can downregulate AXIN2 and GSK3β expression and activate the Wnt/β-catenin pathway, thus exerting pro-oncogenic effects. These findings provide new clues regarding the novel tumor suppressor mechanism of GAS5 in BC.

In the present study, we confirmed that GAS5 was expressed at low levels in both BC tissues and cells by PCR assay, and RNA FISH showed that GAS5 was mainly located in the cytoplasm of T24 cells. This finding suggests that it may function as a miRNA sponge. Through functional assays, we confirmed that GAS5 could regulate the proliferation and migration ability of BC cells. However, the specific mechanism by which GAS5 regulates BC progression is not clear. We performed prediction via the BioSign database and predicted 9 miRNAs as possible target genes of GAS5. We used ChIRP experiments and qRT–PCR for the selection of miRNAs. We found that only miR-18a-5p and miR-18b-5p were highly expressed in BC. We searched PubMed and found that miR-18a-5p and miR-18b-5p were not correlated with BC, so we performed gain- and loss-of-function assays for miR-18a-5p and miR-18b-5p in BC. We found that overexpression of miR-18b-5p did not have an effect on the proliferation and migration ability of BC cells. In contrast, overexpression of miR-18a-5p promoted the proliferation and migration of BC cells. Therefore, we investigated the mechanism by which miR-18a-5p regulates BC in-depth. Using the BioSign database, we analyzed the pathways influenced by miR-18a-5p and found that miR-18a-5p has a significant effect on the Wnt/β-catenin pathway. We used the Biosign database to explore whether miR-18a-5p has a targeting relationship with Wnt/β-catenin pathway-related proteins, and the results showed that miR-18a-5p can bind to the Wnt/β-catenin pathway regulatory proteins AXIN2 and GSK3β. MiR-18a-5p was further demonstrated to bind to AXIN2 and GSK3β proteins using a dual-luciferase assay.

The classical Wnt/β-catenin signaling pathway is a highly conserved signaling pathway in species evolution that regulates the balance of cell growth, migration and differentiation in embryonic development as well as in adults (28); it is also involved in regulating various pathophysiological processes, such as cell proliferation, differentiation, invasion, apoptosis, organismal immunity and tissue repair. A multitude of studies have confirmed that the abnormal activation of various components of the classical Wnt/β-catenin signaling pathway is closely related to the development, infiltration and metastasis of BC, which affects the invasion and metastasis of the tumor, patient survival and prognosis, and β-catenin accumulation in the nucleus of cancer cells, suggesting the involvement of the classical Wnt/β-catenin signaling pathway in bladder carcinogenesis (29). Infiltration and metastasis are the most important biological features of BC and are the main causes of morbidity and mortality in patients; these are also important factors in assessing the effectiveness of BC treatment. Mutations and abnormal expression of members of the classical Wnt/β-catenin signaling pathway are the basis for BC infiltration and metastasis. LncRNA CASC9 promotes BC tumor metastasis by regulating the FZD6/Wnt/β-catenin signaling pathway (30); EFEMP2 inhibits epithelial-mesenchymal transition (EMT) in human BC via the Wnt/β-catenin signaling pathway (31).

It has been widely reported that GSK3β and AXIN2 can inhibit the expression of Wnt/β-catenin pathway components and are key regulatory proteins of the Wnt/β-catenin pathway, especially in the contexts of cell proliferation, cell mutation, cell migration, cell death and other cell functions (32). For example, genetic variants of AXIN can activate the Wnt/β-catenin signaling pathway and thus promote BC development (33); GSK3β negatively regulates and inhibits the Wnt/β-catenin signaling pathway by phosphorylating β-catenin, leading to its proteasomal degradation (34). Decreased expression of AXIN2 and GSK3β has been shown to be associated with the development of BC and to be correlated with poor prognosis of the disease. The findings show that the GSK3β and AXIN2 proteins are mainly involved in the formation of the axin/GSK3β/APC/β-catenin complex, which promotes β-catenin protein degradation (35). In breast cancer cells, the Wnt-Axin2-GSK3β cascade regulates Snail1 activity and thus EMT, which represents an aggressive phenotype of the tissue (36). In this study, we found that GSK3β and AXIN2 protein expression levels decreased after overexpression of miR-18a-5p and increased after miR-18a-5p knockdown, which in turn resulted in corresponding changes in Wnt/β-catenin pathway-related proteins.

In conclusion, our study shows that GAS5 is downregulated in human BC and that it can efficiently bind miR-18a-5p to promote AXIN2 and GSK3β expression. We also demonstrated that overexpression of GAS5 can effectively inhibit the classical Wnt/β-catenin signaling pathway by targeting the miR-18a-5p/AXIN2/GSK3β axis, thereby suppressing BC cell proliferation and metastasis. Our results may provide ideas for the early diagnosis of BC and new targets for anti-BC therapy.

Conflict of Interest Statement

The authors declare that they have no conflict of interest.

Author contributions

Z.Z., T.L. and Y.L. conceived and designed the study and critically revised the manuscript. H.H. provided funds and ­technical support. C.C., J.W., C.W., and H.W. performed the experiments and collected the specimens. All the authors read and approved the final manuscript.

Data Availability Statement

All data are fully available without restrictions.