MicroRNA‐148a‐3p suppresses epithelial‐to‐mesenchymal transition and stemness properties via Wnt1‐mediated Wnt/β‐catenin pathway in pancreatic cancer

Abstract Although miR‐148a‐3p has been reported to function as a tumour suppressor in various cancers, the molecular mechanism of miR‐148a‐3p in regulating epithelial‐to‐mesenchymal transition (EMT) and stemness properties of pancreatic cancer (PC) cells remains to be elucidated. In the present study, we demonstrated that miR‐148a‐3p expression was remarkably down‐regulated in PC tissues and cell lines. Moreover, low expression of miR‐148a‐3p was associated with poorer overall survival (OS) in patients with PC. In vitro, gain‐of‐function and loss‐of‐function experiments showed that miR‐148a‐3p suppressed EMT and stemness properties as well as the proliferation, migration and invasion of PC cells. A dual‐luciferase reporter assay demonstrated that Wnt1 was a direct target of miR‐148a‐3p, and its expression was inversely associated with miR‐148a‐3p in PC tissues. Furthermore, miR‐148a‐3p suppressed the Wnt/β‐catenin pathway via down‐regulation of Wnt1. The effects of ectopic miR‐148a‐3p were rescued by Wnt1 overexpression. These biological functions of miR‐148a‐3p in PC were also confirmed in a nude mouse xenograft model. Taken together, these findings suggest that miR‐148a‐3p suppresses PC cell proliferation, invasion, EMT and stemness properties via inhibiting Wnt1‐mediated Wnt/β‐catenin pathway and could be a potential prognostic biomarker as well as a therapeutic target in PC.

through EMT, and CSCs exhibit an EMT phenotype for metastasis. 19 Overlapping of these two properties implies that they may share similar molecules/pathways. To date, multiple miRNAs have been proved to be involved in EMT and the maintenance of CSC-like properties. Yan et al 20 reported that miR-148a inhibits the metastasis of hepatocellular carcinoma by blocking the EMT process and CSC-like properties. Also, miR-148a expression is significantly attenuated in cancer stem cell-like hepatocellular carcinoma subtype. 21 However, the role of miR-148a-3p in EMT and the maintenance of stemness properties in PC cells have not yet been clarified.
As a member of the Wnt ligand gene family, Wnt1 aberrantly activates the Wnt/β-catenin pathway in many human tumours, regulates the transcription of downstream genes and then affects cell proliferation, EMT, metastasis and the maintenance of stemness properties. [22][23][24] In our study, We firstly demonstrated that Wnt1 was a direct functional target of miR-148a-3p, and miR-148a-3p suppressed cell proliferation, invasion, EMT and stemness properties via Wnt1-mediated Wnt/β-catenin pathway in PC.

| Cell lines and cell culture
The normal human pancreatic ductal epithelial cell line HPDE and human PC cell lines Panc-1, SW1990, Mia PaCa-2 and BxPC-3 were obtained from the Type Culture Collection of the Chinese Academy of Sciences. The human PC cell line Capan-2 was obtained from the ATCC. These cell lines were routinely maintained in DMEM (Invitrogen) supplemented with 10% FBS in a humidified atmosphere containing 5% carbon dioxide at 37°C.

| RNA extraction and quantitative real-time PCR
RNA was isolated from snap-frozen fresh specimens and cell lines using TRIzol Reagent (Thermo Fisher Scientific) and reverse-transcribed using the PrimeScript RT Reagent Kit (Takara), according to the manufacturer's instructions. Then, quantitative PCR was performed using SYBR Premix Ex Taq (Takara), following the manufacturer's instructions. U6 and GAPDH were regarded as the internal control, and the relative gene expression was determined using the 2 −ΔΔC t method. The primer sequences are listed in Table 1.

| Lentivectors and plasmid transfection
Lentivirus encoding miR-148a-3p, anti-miR-148a-3p and their respective negative control were synthesized by GenePharma and then infected Mia PaCa-2, BxPC-3 and Capan-2 cells. The infection efficiency was confirmed by quantitative real-time PCR (qRT-PCR). For generation of stably transfected cells, the cells were treated with 2 μg/mL of puromycin (Sigma) for 2 weeks, and then, GFP-positive cells were selected for subsequent assays. Special siRNA against Wnt1 (si-Wnt1), Wnt1 overexpression plasmid (Wnt1-pcDNA3.1-EGFP) and their respective negative control vector were purchased from GenePharma and then transfected using Lipofectamine 3000 (Invitrogen). The transfection efficiency was confirmed by qRT-PCR. In the specified group, cells were exposed to 10 ng/mL TGF-β (R&D Systems) and then incubated at 37°C for 48 hours.

| CCK-8 assay and colony formation assay
For CCK-8 assay, cells with different treatments were seeded in 96-well plate with a density of 5 × 10 3 cells per well. At the indicated time-points (1, 2, 3, 4 and 5 days), 10 μL CCK-8 reagent was added to each well and incubated in dark at 37°C for another 1 hour. Finally, the spectrophotometric absorbance at 450 nm was measured for each well. For colony formation assay, cells with different treatments were seeded in 6-well plate with a density of 1 × 10 3 cells per well. After incubation for 14 days, the cells were fixed with 4% paraformaldehyde and then stained with 0.5% crystal violet. Finally, individual clones were counted to evaluate cell proliferation. Individual clones (>50 cells/clone) were counted.
At least three wells were assessed for each group, and the mean was calculated.

| Wound healing assay
Cells with different treatments were seeded in a 6-well plate and cultured to grow to confluence. Then, the monolayer was scraped with a pipette tip (200 μL) to produce an artificial wound gap. Finally, the wound closure was photographed after 48-h incubation using an inverted microscope (Olympus), and the wound closure rate was calculated by using ImageJ software.

| Cell migration and invasion assay
The migration and invasion capacities of cells were detected by Transwell assay according to manufacturer's instruction as described previously. 10 Briefly, in the invasion assay, cells were added to the Transwell chambers pre-coated with Matrigel (BD Biosciences). Then, the chambers were placed in a 24-well plate for incubation 24 hours.
Fixed with 4% paraformaldehyde and stained with 0.1% crystal violet, the number of invaded cells was counted under a microscope (Olympus) in five fields with random choice. The same steps were performed in the cell migration assay; however, Matrigel was not used.
The experiment was repeated three times.

| Flow cytometry analysis and magnetic cell sorting
Cells with different treatments were resuspended at 1 × 10 6 cells per mL in PBS containing 2% FBS and then subsequently stained  (Gibco) and 5 μg/mL insulin (Sigma). After two weeks, the size and number of spheres were counted.

| Xenograft experiments
Male BALB/c nude mice aged 5-6 weeks were purchased from Hunan SJA Laboratory Animal Co. Ltd. Cells (1 × 10 7 cells per mice) that stably expressed miR-148a-3p, anti-miR-148a-3p or corresponding controls were subcutaneously injected into the mice, with six mice in each group. Tumour size was recorded weekly using the formula: width 2 × length/2. Mice were killed 5 weeks after inoculation, and tumours were weighed. All animal experiments were performed in accordance with the experimental animal use guidelines of the National

Institutes of Health and approved by the Ethics Committee for Animal
Experiments of the First Affiliated Hospital of Nanchang University.

| Immunohistochemistry assays
The formalin-fixed paraffin-embedded tissues of xenograft tumours were used for immunohistochemistry (IHC) analysis and performed as described previously. 26 The antibodies against β-catenin (#8480,

| Statistical analysis
Measurement data were shown as mean ± SD and analysed using SPSS 24.0 statistical software. Values were analysed using two-tailed Student's t test or one-way ANOVA for most of the experiments. Mann-Whitney U test was used to compare the expression of miR-148a-3p between PC tissues and their corresponding adjacent tissues. Spearman's correlation test was used for correlation analyses. Survival analysis was performed using the Kaplan-Meier method, and differences were assessed with the log-rank test. The experimental data were representative of at least three independent experiments and were considered statistically significant at P < .05.

| Low expression of miR-148a-3p is correlated with poor prognosis in patients with PC
Our previous study has shown that miR-148a was down-regulated in 33 PC tissues, and miR-148a expression was significantly associated with histological grade, tumour size, lymph node status and TNM stage. 9 In the present study, we expanded these samples to 61 pairs of PC tissues. The qRT-PCR results confirmed that miR-148a-3p was significantly decreased in PC tissues as compared with their corresponding adjacent tissues ( Figure 1A). Consistently, miR-148a-3p was markedly down-regulated in five PC cell lines (Panc-1, SW1990, Mia Paca-2, BxPC-3 and Capan-2) relative to normal HPDE cell line ( Figure 1B). By dividing all patients into high-and low-expression group based on the median relative expression level of miR-148a-3p, we further found that patients with lower expression of miR-148a-3p exhibited significantly poorer overall survival (OS; Figure 1C).
Consistently, data derived from the KM Plotter online database (http://kmplot.com/analy sis/) were in agreement with this result ( Figure 1D). Together, these data revealed that miR-148a-3p was down-regulated in PC, which was correlated with poor prognosis in patients with PC.

| miR-148a-3p inhibits malignant behaviour of PC cells
As

| miR-148a-3p inhibits the EMT process of PC cells
It is well-known that EMT plays an important role in promoting cell migration and invasion. 28 In the present study, we further evaluated the expression of EMT markers (E-cadherin, N-cadherin and vimentin) by Western blotting in PC cells. Our results showed that the levels of mesenchymal markers N-cadherin and vimentin were significantly up-regulated, whereas epithelial marker E-cadherin was down-regulated in Capan-2, BxPC-3 and Mia PaCa-2 cells following miR-148a-3p depletion ( Figure 3D and Figure S1A). Conversely, miR-148a-3p overexpression increased the expression of E-cadherin, accompanying decreased N-cadherin and vimentin expression ( Figure 3D and Figure S1A). Immunofluorescence analysis further confirmed those findings ( Figure S2A). Additionally, PC cell morphology was drastically altered by miR-148a-3p overexpression or depletion: Mia PaCa-2 cells in miR-148a-3p group had a round or cobblestone-like appearance, whereas Capan-2 cells in anti-miR-148a-3p group had a spindle shape as compared with the control cells ( Figure S2B). On the basis of these findings, we concluded that miR-148a-3p suppresses cell invasion and migration of PC cells in vitro through inhibiting the EMT process.

| miR-148a-3p inhibits stemness properties of PC cells
As CD133 was well-recognized as a putative CSC marker for most prevalent solid human cancers including PC, 29 we detected the proportion of CD133 + cells in normal HPDE cell line and three PC cell lines (Capan-2, BxPC-3 and Mia PaCa-2) by using flow cytometry. As shown in Figure 4A, no significant expression of CD133 was seen in normal HPDE cells, but an elevated proportion of CD133 + cells was seen in Capan-2, BxPC-3 and Mia PaCa-2 cells. As Capan-2 cells exhibited higher CD133 + proportion cells, we detected the expression of miR-148a-3p and stemnessassociated markers in CD133 + and CD133 − cells isolated from Capan-2 cells. As shown in Figure 4B, compared to CD133 − cells, miR-148a-3p was markedly down-regulated in CD133 + cells, whereas the stemness-associated markers CD133, Nanog, Oct4 and Sox2 were up-regulated. Furthermore, miR-148a-3p overexpression markedly down-regulated the expression of Nanog, Oct4 and Sox2 in Capan-2, BxPC-3 and Mia PaCa-2 cells, while miR-148a-3p depletion up-regulated these stemness-associated markers ( Figure 4C and Figure S1B). Flow cytometry results also showed that the CD133 + proportion cells in Capan-2, BxPC-3 and Mia PaCa-2 cells were significantly reduced by miR-148a-3p overexpression, while miR-148a-3p depletion caused an opposite effect ( Figure 4D). In addition, functional assays demonstrated F I G U R E 1 The expression of miR-148a-3p is down-regulated in pancreatic cancer (PC) tissues and cell lines. A, miR-148a-3p expression was assessed in 61 pairs of PC and corresponding ANT tissues using RT-qPCR. B, RT-qPCR was used for the determination of miR-148a-3p expression in five PC cell lines (Panc-1, SW1990, Mia Paca-2, BxPC-3 and Capan-2) and the normal HPDE cell line. C, Kaplan-Meier analysis of the correlation between miR-148a-3p expression and overall survival of PC patients (log-rank test, P = .011). D, Data derived from the KM Plotter online database showed that low expression of miR-148a-3p was associated with poorer OS in patients with PC. Data were expressed as means ± SD of three independent experiments. **P < .01, ***P < .001 F I G U R E 2 miR-148a-3p inhibits the proliferation of PC cells. A, After infecting lentivirus encoding miR-148a-3p or anti-miR-148a-3p, the expression level of miR-148a-3p in Capan-2, BxPC-3 and Mia PaCa-2 cells was confirmed using RT-qPCR analysis. B,C, The regulatory roles of miR-148a-3p in PC cell proliferation and colony formation were evaluated using CCK-8 assay and colony formation assay. Data were expressed as means ± SD of three independent experiments. *P < .05, **P < .01, ***P < .001
Furthermore, the expression level of miR-148a-3p was inversely associated with Wnt1 expression in PC tissues ( Figure 6E). These data made it evident that Wnt1 is a direct target of miR-148a-3p in PC.
In addition, the ectopic Wnt1 expression effectively reversed the inhibition of proliferation ( Figure

| miR-148a-3p functions in vivo
To verify the above findings in vivo, we performed studies using a nude mouse subcutaneous xenograft model. We observed that miR-148a-3p-stable-expressing Mia PaCa-2 cell-derived xenografts grew at a significantly slower rate, with lower tumour volumes and weight, compared with miR-con group ( Figure 9A). In contrast, the volume and weight of tumours formed by anti-miR-148a-3p-stable-expressing Capan-2 cells were elevated compared with those in anti-miR-con group ( Figure 9B). qRT-PCR results showed a significantly elevated expression level of miR-148a-3p in miR-148a-3p-stable-expressing Mia PaCa-2 cell-derived xenografts and a markedly decreased expression level of miR-148a-3p in anti-miR-148a-3p-stable-expressing Capan-2 cell-derived xenografts ( Figure 9C). IHC analysis showed the expression level of Ki-67 in miR-148a-3p group was lower than that in miRcon group, whereas it was higher in anti-miR-148a-3p group than that in anti-miR-con group ( Figure 9D). We further evaluated the expression of Wnt1, β-catenin, EMT markers (E-cadherin and vimentin) and stemness-associated markers (CD133, Nanog and Oct4) in xenograft tumours. Both the results of qRT-PCR and IHC analysis confirmed that the expression levels of Wnt1, β-catenin, vimentin, CD133, Nanog and Oct4 were significantly down-regulated, whereas the expression F I G U R E 7 miR-148a-3p inhibits PC cell proliferation through targeting Wnt1. A, The mRNA and protein levels of Wnt1 in Capan-2, BxPC-3 and Mia PaCa-2 cells after Wnt1 and/or miR-148a-3p up-regulation were detected using RT-qPCR and Western blot analysis, respectively. B, C, The proliferation of Capan-2, BxPC-3 and Mia PaCa-2 cells after Wnt1 and/or miR-148a-3p up-regulation was detected using CCK-8 assay and colony formation assay. Data were expressed as means ± SD of three independent experiments. *P < .05, **P < .01, ***P < .001 level of E-cadherin was up-regulated in miR-148a-3p group, compared with miR-con group; conversely, the anti-miR-148a-3p group showed a significantly elevated expression level of Wnt1, β-catenin, vimentin, CD133, Nanog and Oct-4, and a markedly decreased expression level of E-cadherin, compared with anti-miR-con group ( Figure 9E-H).
Collectively, our subcutaneous xenograft studies provided further evidence that miR-148a-3p suppresses the growth, EMT and stemness properties of PC via inhibiting Wnt1/β-catenin pathway in vivo. CSCs play critical roles in tumour progression, distant metastasis and therapeutic resistance. 15,17,18 42,43 Overlapping of these two properties implies that they may share similar molecules/pathways. In the present study, TGF-β promoted the stemness properties of PC cells, partly restored the inhibitory effect of miR-148a-3p on the stemness properties of PC cells. These findings confirmed that the EMT process is involved in the regulation of miR-148a-3p on stemness properties of PC cells.
It is well-known that miRNAs perform the specific functions through inhibiting the target mRNAs expression. 44 More than 800 genes were predicted as potential targets gene of miR-148a-3p F I G U R E 1 0 The schematic illustration of the potential molecular mechanism of miR-148a-3p as a tumour suppressor in PC. miR-148a-3p was remarkably downregulated in PC and negatively regulated Wnt1, which was a direct functional target of miR-148a-3p. miR-148a-3p suppressed EMT and stemness properties as well as the proliferation, migration and invasion via inhibiting Wnt1-mediated Wnt/βcatenin pathway in PC by bioinformatical prediction tools TargetScan and miRBase.
In summary, our results show that miR-148a-3p expression is remarkably reduced in PC, and its low expression is associated with poorer OS in patients with PC. Furthermore, we confirm that miR-148a-3p suppresses proliferation, invasion, EMT and stemness properties of PC cells via directly targeting Wnt1-mediated Wnt/βcatenin pathway. These findings provide new evidence for understanding the function and molecular mechanism of miR-148a-3p in PC, highlighting the potential of miR-148a-3p as a useful prognostic biomarker as well as a therapeutic target for PC.

ACK N OWLED G EM ENTS
The present study was supported by the National Natural Science

CO N FLI C T O F I NTE R E S T
The authors declare that there is no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.