Role of AP-2α and MAPK7 in the regulation of autocrine TGF-β/miR-200b signals to maintain epithelial-mesenchymal transition in cholangiocarcinoma

The HCCC and RBE cells were cultured to about 80% confluence in 6-well plates and, using Lipofectamine 2000 (Invitrogen, USA), the cells were transfected with miR-200b, pcDNA3.0-TFAP2A, and MAPK7 and pGL-3-TGFB1 according to the manufacturer’s instructions. After transfection for the indicated time, the cells were harvested for further experiments.

After the transfection for indicated time, HCCC and RBE cells were harvested and washed with PBS and cell counting kit-8 (Kumamoto, Japan) mixed with DMEM was used for cell viability assay. The absorbance was measured at 450 nm by a microplate reader (BioTek, Synergy™ HTX, USA).

In the migration assay, 2.5 × 10⁴ HCCC and RBE cells were cultured in the upper chamber of a non-coated Transwell insert. Medium (600 μL) supplemented with 10% bovine serum was used as a chemo-attractant to encourage cell migration in the lower chamber. For invasion assay, the upper chamber of the Transwell inserts were coated with 50 μL of 2.0 mg/mL Matrigel, and 5 × 10⁴ HCCC and RBE cells were plated simultaneously in the upper chamber of the Matrigel-coated Transwell insert. The cells were incubated for 24 h; cells that did not migrate or invade were removed using a cotton swab. Using crystal violet staining, the cells were stained and counted under an inverted microscope. Five random views were selected to count the cells. All the independent experiments were repeated thrice.

Quantitative real-time RT-PCR (qRT-PCR) was performed according to the standard protocol, and the expression levels of miR-200b, E-cadherin, TTF-1, fibronectin, and a-SMA were normalized to GAPDH for gene expression. The primers used are listed below:

The total proteins were extracted and dissolved in SDS-PAGE loading buffer, were separated on 4–15% polyacrylamide gel, and were transferred to nitrocellulose membranes (Amersham Biosciences) according to the manufacturer’s instructions. The membranes were blocked in 5% non-fat milk in tris-buffer saline containing 0.1% Tween-20 (TBST) buffer for 1 h at room temperature and were subsequently incubated with primary antibodies overnight at 4 °C. After washing with TBST buffer, the blots were incubated with HRP-conjugated secondary antibody for 1 h at room temperature and were visualized using the ECL-Plus reagent (Millipore, Billerica, MA, USA). GAPDH was used as the loading control in the Western blotting.

To investigate the targets of miR-200b or the role of AP-2α on the transcription of TGFB1, the HCCC cells were co-transfected with 200 ng firefly Luciferase vector, 40 ng Renilla luciferase pRL-TK vector (Promega, USA) and psiCHECK2-UTR (wild-type or mutant) of TFAP2A and MAPK7 and pGL-3-TGFB1 (wild-type or mutant), pcDNA3.0-TFAP2A or miR-200b. Luciferase activity was determined using the dual-luciferase reporter assay system (Promega, USA). Firefly luciferase acted as a reporter gene and Renilla luciferase as a normalized control. The primers used are listed below:

The DNA-binding activity of AP-2α was determined by EMSA using nuclear extracts prepared by a nuclear extract kit (Active Motif; Rixensart, Belgium) from different groups of cells according to the manufacturer’s instructions. The related probes were synthesized (Invitrogen, USA) and incubated with nuclear extracts in the reaction solution. Unlabeled wild-type (site: GCCCCCTTTCCCCCAGGGCTGAAGGGACC) and mutant (site: GCCCCCTTAGGGGGTCCCGAGAAGGGACC) double-stranded competitor oligonucleotides were added to the respective reactions.

To investigate the tumor suppressive role of miR-200b in vivo, 3 × 10⁶ HCCC cells were subcutaneously injected in rear flank of nude mice (male BALB/c-nu/nu mice, four per group). The miRNA delivery system (micrON™ miRNA agomir and micrOFF™ miRNA antagomir) was purchased from RiboBio (Guangzhou, China). For delivery of cholesterol-conjugated RNA, the miR-200b mimics or negative control (10 nM per injection) were delivered via intra-tumoral injection for six times, 3 days apart.

The Statistical Package for Social Sciences version 16.0 (SPSS 16.0, SPSS Inc., Chicago, IL, USA) and the Prism statistical software package (Version 5.0, Graphpad Software Inc.) were used to performed the statistical analyses. Unpaired t tests or Mann-Whitney U tests were used to compare the two groups, and multiple group comparisons were analyzed with one-way ANOVA. P < 0.05 was considered statistically significant. All experiments were performed in triplicates.

Since the majority of patients with CCA are ineligible for curative surgery due to the presence of metastases at the time of diagnosis, it is imperative to confirm the status of EMT in CCA samples prior to further investigations on the mechanism. CCA tissues and corresponding non-neoplastic tissues (n = 20) were collected, and the expression of EMT markers, including epithelial-specific marker, E-cadherin, thyroid transcription factor 1 (TTF-1), and mesenchymal markers, including fibronectin and α-SMA, were determined. The results from Q-PCR and western blot indicated that the mRNA expression of E-cadherin and TTF-1 was downregulated, while the expression of fibronectin and α-SMA was increased in CCA tissues (Fig. 1a, b), which indicated that during the progression of CCA, the tumor cell exhibited activated MET phenotype. TGF-β/miR-200b-dependent signaling is a pivotal inducer of EMT in various cancers, such as CCA. It was observed that miR-200b expression was decreased in CCA tissues as compared to non-neoplastic tissues and, conversely, the TGF-β expression was significantly increased in CCA tissues (Fig. 1c, d). Further, correlation analysis indicated that the expression of miR-200b was negatively correlated with the expression of TGF-β, and the decreased miR-200b was associated with the phenotype of EMT by the switch of related markers (Fig. 1e). These findings suggested that the altered expression of miR-200b might be regulated by TGF-β in order to interfere the balance of miR-200b/EMT signals in CCA.

To clarify the role of miR-200b in EMT in CAA, two cell-lines, HCCC and BRE, were used in this study. After the stimulation of TGF-β to CCA cell-lines for 2 days, the level of miR-200b was found to be remarkably inhibited (Fig. 2a). Also, a shift from epithelial appearances to fibroblastic-like cell morphology in HCCC and BRE cell lines was observed (Fig. 2b), which significantly increased the ability of migration and invasion of tumor cells. Moreover, we analyzed the mechanism of miR-200b in TGF-β-induced migration and invasion. It was observed that TGF-β repressed the level of E-cadherin and TTF-1 and promoted the expression of fibronectin and α-SMA, which enhanced the ability of migration and invasion of HCCC and BRE cells (Fig. 2c). Overexpression of miR-200b in two cells was demonstrated to abrogate the switch to fibroblastic-like cell phenotype by repressing the expression of fibronectin and α-SMA induced by TGF-β (Fig. 2c). Besides, overexpression of miR-200b inhibit the TGF-β-induced upregulation of migration and invasion (Fig. 2d). The results indicated that high level of TGF-β in tumor microenvironment inhibited the level of miR-200b in CCA tumor cells in order to promote EMT and allow tumor cell metastases.

Further, the HCCC and BRE cells transfected with miR-200b showed decreased cell vitality and proliferation (Fig. 2e), despite the treatment of TGF-β, which suggested that miR-200b could not only inhibit the EMT of tumor cell, but also impair the cell proliferation in CCA.

We determined the potential targets of miR-200b in TGF-β-induced EMT. According to TargetScan (http://​www.​targetscan.​org/​vert_​71/​), the 3′-UTR of TFAP2A, which encoded AP-2α, harbored a binding site of miR-200b (Fig. 3a). To conform this, a luciferase reporter vector, with full-length 3′-UTR (wild-type or mutant) of TFAP2A, was transfected into HCCC cell-line. The results showed that miR-200b mimics could significantly impair the luciferase activity of wild-type TFAP2A −3’UTR, but not the mutant-type (Fig. 3a). The clinical expression of AP-2α in CCA was also estimated, and upregulated levels of AP-2α was found in tumor samples when compared to normal tissues (Fig. 3b), thereby suggesting that miR-200b could directly target TFAP2A, and downregulation of miR-200b might participate in the tumor metastasis via TFAP2A.

Interestingly, using luciferase reporter assay, we found that AP-2α could promote the transcription of TGFB1 in HCCC cells (Fig. 3c). The data indicated that AP-2α could directly bind to the promoter motif of TGFB1 to induce its expression in autocrine manner, which was also confirmed by EMSA (Fig. 3d).

To provide the evidence that AP-2α participated in miR-200b-regulated EMT in CCA, we transfected miR-200b mimics with AP-2α into the HCCC cells. The results showed that, while miR-200b could prevent the EMT, overexpression of AP-2α significantly increased the level of TGF-β to restore the EMT phenotype of HCCC cell (Fig. 3e, f).

MAPK7 belongs to the mitogen-activated protein kinase (MAPK) family of protein kinases and is involved in cell survival, angiogenesis, cell motility, cell proliferation, and EMT [23]. We investigated the function of MAPK7 in TGF-β/miR-200b pathway in CCA. Upon analyzing the 3′-UTR of MAPK7, it was observed that MAPK7 harbored a binding site of miR-200b. Luciferase reporter assays indicated that miR-200b mimics could significantly decrease the luciferase activity of wild-type MAPK7–3′UTR, but not the mutant-type (Fig. 4a), which indicated a direct relationship between miR-200b and MAPK7 in the HCCC cells. Further, overexpression of MAPK7 in HCCC cells could upregulate the TGF-β expression, repress the expression of E-cadherin and TTF-1, and promote the expression of fibronectin and α-SMA in order to re-activate the miR-200b-induced inhibition of EMT (Fig. 4b, c), which was confirmed by the upregulation of migration and invasion (Fig. 4d, e). The results demonstrated a positive regulatory loop between MAPK7 and TGF-β, which regulates EMT via miR-200b in CCA.

We also determined the role of MAPK7 in proliferation and cell cycle. It was observed that MAPK7 played a pivotal role in the cell vitality and proliferation of tumor cells. Forced expression of miR-200b inhibited cell proliferation and vitality (Fig. 4d, e). However, MAPK7 rescued the miR-200b-induced inhibition and increased cell vitality of HCCC cells and promoted the proliferation (Fig. 4f). Mechanistically, overexpression of miR-200b in HCCC cells decreased the expression of cyclin D1 and Cdk2 in order to induce cell cycle arrest and was able restore their expression for the growth of tumor cells (Fig. 4g).

To determine the role of miR-200b in tumor progression in vivo, two CCA cell-lines, HCCC and BRE, were subcutaneously injected in rear flank of nude mice. The miR-200b mimics, inhibitor or negative control were delivered via intratumoral injection for six times, 3 days apart. The results showed that the mice treated with miR-200b mimics showed a delayed tumor growth, with less tumor weight and volume, than those treated with PBS or negative control (Fig. 5a, b). The status of EMT in tumor samples was analyzed. It was observed that the treatment of miR-200b significantly abrogated the AP-2α/MAPK7/TGF-β expression (Fig. 5c–e), leading to increased expression of E-cadherin and TTF-1 and decreased expression of fibronectin and α-SMA in tumor tissues (Fig. 5f–i). However, the treatment of miR-200b inhibitor promoted the tumor progression with enhanced TGF-β-induced EMT. These findings showed that miR-200b could target AP-2α/MAPK7 to interfere the regulatory loop involving miR-200b and TGF-β in order to repress EMT for cancer metastasis in CCA.