Author contributions: D.K.: conception and design, execution, collection of data and data analysis, manuscript writing; Y.L.: experimental design, data analysis, manuscript writing; Z.W.: experimental design and execution, data analysis; S.B.: experimental design, execution, and data analysis; A.A.: experimental execution and collection of data; H.-R.C.K.: development of LNCaP PDGF-D and PC3 PDGF-D cell lines and experimental design; F.H.S.: principal investigator, conception and design, laboratory facility and financial support, experimental design, data collection and interpretation, manuscript writing, final approval of the manuscript.
First published online in STEM CELLS EXPRESS April 23, 2009.
MicroRNAs have been implicated in tumor progression. Recent studies have shown that the miR-200 family regulates epithelial–mesenchymal transition (EMT) by targeting zinc-finger E-box binding homeobox 1 (ZEB1) and ZEB2. Emerging evidence from our laboratory and others suggests that the processes of EMT can be triggered by various growth factors, such as transforming growth factor β and platelet-derived growth factor-D (PDGF-D). Moreover, we recently reported that overexpression of PDGF-D in prostate cancer cells (PC3 PDGF-D cells) leads to the acquisition of the EMT phenotype, and this model offers an opportunity for investigating the molecular interplay between PDGF-D signaling and EMT. Here, we report, for the first time, significant downregulation of the miR-200 family in PC3 PDGF-D cells as well as in PC3 cells exposed to purified active PDGF-D protein, resulting in the upregulation of ZEB1, ZEB2, and Snail2 expression. Interestingly, re-expression of miR-200b in PC3 PDGF-D cells led to reversal of the EMT phenotype, which was associated with the downregulation of ZEB1, ZEB2, and Snail2 expression, and these results were consistent with greater expression levels of epithelial markers. Moreover, transfection of PC3 PDGF-D cells with miR-200b inhibited cell migration and invasion, with concomitant repression of cell adhesion to the culture surface and cell detachment. From these results, we conclude that PDGF-D-induced acquisition of the EMT phenotype in PC3 cells is, in part, a result of repression of miR-200 and that any novel strategy by which miR-200 could be upregulated would become a promising approach for the treatment of invasive prostate cancer. STEM CELLS 2009;27:1712–1721
Epithelial–mesenchymal transition (EMT) is a process that is reminiscent of “cancer stem-like cell” characteristics, whereby epithelial cells with a cobblestone phenotype acquire mesenchymal cell characteristics with a spindle-shaped fibroblast-like morphology. This process involves a disassembly of cell–cell junctions, including downregulation and relocation of E-cadherin and zonula occludens (ZO)-1 as well as downregulation and translocation of β-catenin from the cell membrane to the nucleus, actin cytoskeleton reorganization, and upregulation of mesenchymal molecular markers such as vimentin, fibronectin, and N-cadherin . Typically, epithelial cells form clusters mediated through the regulation of cell–cell junctions and adhesion, such as tight junctions, adherens junctions, desmosomes, and gap junctions, and thereby inhibit the movement of individual cells. In contrast, mesenchymal cells have less adhesion between cells than their epithelial counterparts, allowing for more motile and invasive characteristics, which contribute to cancer cell invasion and metastasis .
During the acquisition of EMT, loss of epithelial markers is a critical process that is regulated by important transcription repressors. In the last few years, several transcription repressors have been identified, including zinc-finger E-box binding homeobox 1 (ZEB1), ZEB2/SIP1, a member of the δEF-1 family of two-handed zinc finger nuclear factors, Snail1, Snail2/Slug, Twist, and E47. ZEB1 has been shown to be a critical mediator of EMT induced by various inducers in different cell lines [3–6]. Transcription factor ZEB1 has been shown to regulate expression of genes by binding to ZEB-type E-boxes (CACCTG) within the promoter region of target genes, resulting in chromatin condensation and gene silencing . The expression of E-cadherin is negatively regulated by ZEB1, which is fundamental for the processes of EMT . Recent studies have also demonstrated that ZEB1 promotes cell migration and tumor metastasis by repressing the expression of cell polarity factors [9, 10]. ZEB2/SIP1 has been shown to downregulate the expression of many genes coding for crucial proteins of the epithelial phenotype, including E-cadherin , and to upregulate the expression of vimentin . Snail2/Slug has been shown to play a central role in the induction of EMT by growth factors [13, 14], suggesting that ZEB1, ZEB2, and Snail2 are important regulators in the induction of EMT.
The processes of EMT could be triggered by many growth factors, including transforming growth factor β and platelet-derived growth factor (PDGF)-A, PDGF-B, and PDGF-D [15–19]. It is also known that PDGF-D could regulate many cellular processes, including invasion and angiogenesis, by activating its cognate receptor PDGFR-β [20, 21]. A recent study has shown increased expression of PDGF-D in human prostate carcinoma samples, suggesting that PDGF-D could play an important role in the progression of human prostate cancer . We have recently shown that stable transfection of PC3 cells with PDGF-D cDNA led to the acquisition of the EMT phenotype, and this process was consistent with increased invasiveness and in vivo tumor growth rate of PC3 PDGF-D cells . This model offers an opportunity for further investigation of the molecular interplay between PDGF-D signaling and EMT and the precise mechanism by which PC3 PDGF-D cells acquire the EMT phenotype.
Emerging evidence suggests that the miR-200 family could regulate the processes of EMT by targeting ZEB1 and ZEB2 [4, 5, 24]. MicroRNAs (miRNAs) are small (19-24 nucleotides) noncoding RNA molecules that downregulate gene expression by interacting with sequences located in the 3′ untranslated region (UTR) of multiple target mRNAs, resulting in either translational repression or degradation of mRNAs . It is known that miRNAs are involved in embryonic development and in cancer progression, a process that is known to be associated with the acquisition of the EMT phenotype of epithelial tumor cells . Moreover, recent studies showed that miR-200 could repress ZEB1/SIP1 expression through binding to a sequence of the 3′-UTR of ZEB1/SIP1 mRNA and that ZEB1 and ZEB2 could directly bind to the promoter of the miR-200 gene cluster, resulting in the repression of miR-200 expression and establishing a double-negative feedback loop controlling the expression of ZEB1/SIP1 and the miR-200 family during EMT [4, 24, 27]. Under these conditions, discovery of factors that regulate the expression of miR-200 or ZEB1/SIP1/ZEB2 could be very important for controlling EMT, and such knowledge could be useful for designing strategies for the treatment of aggressive prostate cancer. Therefore, in the current study, we sought to test our hypothesis of whether or not the loss of miR-200 could contribute to the acquisition of the EMT phenotype observed in PC3 PDGF-D cells and whether or not re-expression of miR-200 could lead to the reversal of the EMT phenotype in PC3 PDGF-D cells. We further hypothesized that the above-mentioned processes could be mediated by the deregulated expression of ZEB1/ZEB2 and Snail2 transcription factors.
We found that miR-200 expression was significantly lower in PC3 PDGF-D cells as well as in PC3 cells exposed to purified active PDGF-D protein than in parental PC3 cells, which was associated with the overexpression of ZEB1, ZEB2, and Snail2, and that the transfection of PC3 PDGF-D cells with miR-200b led to the downregulation of ZEB1, ZEB2, and Snail2, with a corresponding upregulation of epithelial markers. Overexpression of PDGF-D in LNCaP cells resulted in lower expression of miR-200 and induced EMT. From these results, we concluded that the loss of miR-200 plays an important role in the acquisition of the EMT phenotype of LNCaP and PC3 cells induced by PDGF-D, and that the re-expression of miR-200 could cause reversal of the EMT phenotype to the mesenchymal–epithelial transition (MET) phenotype. These results provide a mechanistic role of miR-200 in the processes of EMT/MET in PDGF-D-overexpressing prostate cancer cells.
MATERIALS AND METHODS
Cell Lines and Culture Conditions
Generation of stable cell lines overexpressing PDGF-D was accomplished by transfection of PC3 and LNCaP cells with the corresponding empty vector pcDNA3 Neo or pcDNA3-PDGF-D:His, as previously described elsewhere , which are referred to as PC3 Neo or PC3 PDGF-D cells and LNCaP Neo or LNCaP PDGF-D cells, respectively. The PC3, LNCaP, and resultant transfected cell lines were cultured in RPMI 1640 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 5% or 10% fetal bovine serum (FBS), 2 mmol/l glutamine, 10 μmol/l HEPES, 50 units/ml penicillin, and 50 μg/ml streptomycin. All cells were maintained in a 5% CO2-humidified atmosphere at 37°C.
Research Reagents and Antibodies
Recombinant human PDGF-D was purchased from R&D Systems (Minneapolis, http://www.rndsystems.com). Antibodies against Snail2, N-cadherin, and vimentin were purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com), BD Biosciences (Bedford, MA, http://www.bdbiosciences.com), and Abcam (Cambridge, MA, http://www.abcam.com), respectively. Antibodies against ZEB1, ZEB2, Twist, Snail1, fibronectin, retinoblastoma protein, and E-cadherin were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com). Antibodies against PDGF-D or ZO-1, Alexa Fluor 594 goat anti-rabbit IgG or Alexa Fluor 594 goat anti-mouse IgG, and Alexa Fluor 594 phalloidin for F-actin staining were purchased from Invitrogen. Goat anti-rabbit IgG (H + L)-horseradish peroxidase (HRP) conjugate and goat anti-mouse IgG (H + L)-HRP conjugate were obtained from Bio-Rad (Reinach, Switzerland, http://www.bio-rad.com). Antibody to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was purchased from Affinity BioReagents (Golden, CO, http://www.bioreagents.com).
Small Interfering RNA and Transfection
PC3 PDGF-D cells were transfected with 100 nmol/l ZEB1 small interfering (si)RNA or control siRNA (Santa Cruz) using DharmaFECT3 siRNA transfection reagent (Dharmacon, Lafayette, CO, http://www.dharmacon.com). The media were removed after a 24-hour transfection and then the cells were incubated in media containing 5% FBS for another 24 hours. Cell lysates were prepared for Western blot analysis and total RNA was extracted for real-time reverse transcription-polymerase chain reaction (RT-PCR) assay.
Western Blot Analysis
Western blot analysis was performed using cytoplasmic or nuclear extract or total cell lysates. Total cell lysates from different experiments were obtained by lysing the cells in RIPA buffer containing 50 mM Tris–HCl, 150 mM NaCl, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate, 2 mM sodium fluoride, 2 mM Na3VO42, 1 mM EDTA, 1 mM EGTA, and 1× protease inhibitor cocktail. Nuclear and cytoplasmic extracts were prepared according to the method described by our laboratory previously , and Western blotting was performed as previously described elsewhere .
Transfection of miRNA Precursors (pre-miRNA) and Specific Anti-miRNA (Inhibitors)
PC3 PDGF-D cells were seeded at 3 × 105 cells per well in six-well plates and transfected with pre-miR-200b or miRNA-negative control #1 (Ambion, Austin, TX, http://www.ambion.com) at a final concentration of 20 nM using DharmaFECT3 transfection reagent (Dharmacon). LNCaP cells were transfected with specific anti-miRNA (miRNA inhibitors) or anti-miRNA control at a final concentration of 600 nM—200 nM each of miR-200a, miR-200b, and miR-200c anti-miRNAs (Ambion)—using DharmaFECT3 transfection reagent (Dharmacon). After 3 days of transfection, cells were split and transfected repeatedly with pre-miR-200b, miRNA inhibitors, or control every 3-4 days for the indicated times.
Total RNA was isolated using Trizol reagent. One microgram of RNA was reverse transcribed using a reverse transcription system (Invitrogen) according to the manufacturer's instructions. Real-time PCR was used to quantify mRNA expression. Sequences of primers for ZEB1, ZEB2, Snail2, vimentin, epithelial cell adhesion molecule (EpCAM), sciellin, stratifin, crumbs homologue 3 (CRB3), connexin 26, F11 receptor (F11R, junctional adhesion molecule 1), and GAPDH are shown in supporting information Table S. The primer sequence used for E-cadherin in this study is described elsewhere  and the relative amount of RNA was normalized to the expression of GAPDH. For miRNA analysis, total RNA was isolated using the mirVana miRNA isolation kit (Ambion) according to the manufacturer's instructions. miRNA- and RNU6B-specific cDNA was produced from 10 ng of total RNA samples using the Taqman MicroRNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, http://www.appliedbiosystems.com) with miRNA-specific RT primer from the Taqman MicroRNA Assay (Applied Biosystems). The levels of miRNAs were determined using miRNA-specific Taqman MGB probes from the Taqman MicroRNA Assay (Applied Biosystems). The relative amount of miRNA was normalized to RNU6B.
Cell Invasion and Migration Assay
The cell invasion and migration assay was performed using 24-well Transwell permeable supports with 8-μm pores (Corning, Lowell, MA, http://www.corning.com) and is shown in supporting information data.
Immunofluorscence staining was performed as described previously by our laboratory . Briefly, cells were fixed with 4% paraformaldehyde and permeabilized in 0.5% Triton X-100, then blocked with 10% goat serum. The cells were incubated for 1 hour with antibodies against E-cadherin (1:20), ZO-1 (1:50), or vimentin (prediluted) in 5% goat serum, and were stained for 1 hour with Alexa Fluor 594-conjugated secondary antibody (1:250). For F-actin staining, cells were incubated with 0.33 μM of Alexa Fluor 594 phalloidin for 1 hour at 4°C. The slides were mounted with mounting medium containing antifade reagent and 4′,6-diamidino-2-phenylindole. Cells were viewed by fluorescence microscopy and images were analyzed using Advanced Sport software (Diagnostic Instruments, Sterling Heights, MI, http://www.diaginc.com).
Cell Attachment and Detachment Assay
The cell attachment and detachment assay is shown in supporting information data.
Experiments presented in this study are representative of three or more repetitions. data are shown as the mean value ± standard error. A two-tailed Student's t-test was used for comparisons between groups. Values of p < .05 were considered to be statistically significant.
PDGF-D Regulates Expression of the Transcription Factors ZEB1, ZEB2, and Snail2 in PC3 Cells
Because it is known that transcriptional repressors such as ZEB1, ZEB2, Snail1, and Snail2 are the key regulators in inducing the processes of EMT, we first determined the expression status of these transcription repressors in our cell culture model. The results showed that the expression of ZEB1, ZEB2, and Snail2 was dramatically upregulated in PC3 PDGF-D cells, concomitant with the loss of E-cadherin and gain of vimemtin and N-cadherin expression. Moreover, we found that the expression level of Snail1 was significantly lower and the expression level of Twist was not significantly different in PC3 PDGF-D cells compared with PC3 Neo cells (Fig. 1A). To further determine subcellular localization of ZEB1 expression, we tested ZEB1 levels in whole cell lysates and cytoplasmic and nuclear extracts. We found that the levels of ZEB1 were significantly higher in whole cell lysates and in the nuclei from PC3 PDGF-D cells than in PC3 Neo cells (Fig. 1B). These results suggest that ZEB1 is mainly located in the nuclear compartment. Most importantly, PC3 PDGF-D cells displayed elongated/irregular fibroblastoid morphology, compared with PC3 Neo cells, which showed an epithelial cobblestone appearance (Fig. 1C). These results are consistent with the expression status of PDGF-D in PC3 PDGF-D and PC3 Neo cells (Fig. 1A).
Overexpression of PDGF-D Represses Expressions of Epithelial-Specific Genes
In this study, we found that overexpression of PDGF-D significantly downregulated expression of epithelial-specific genes such those encoding as E-cadherin, EpCAM, sciellin, stratifin, CRB3, connexin 26, and F11R (Fig. 2A), concomitant with greater mRNA levels of ZEB1 and ZEB2 (Fig. 2B).
To further verify whether ZEB1 could repress the expression of these epithelial markers, we knocked down ZEB1 by transfection of PC3 PDGF-D cells with ZEB1-specific siRNA. Our results showed that knockdown of ZEB1 significantly increased the mRNA levels of E-cadherin, F11R, CRB3, and sciellin (Fig. 2C), suggesting that ZEB1 is responsible for the regulation of genes that are associated with the acquisition of the EMT phenotype in PC3 PDGF-D cells. Moreover, Figure 2D and 2E show that transfection of cells with ZEB1 siRNA was effective in reducing the expression of ZEB1 at both the mRNA and protein levels; however, for reasons that are presently unknown, we found that the knockdown of ZEB1 decreased mRNA expression of ZEB2, although this finding is consistent with the results shown by other investigators .
miR-200 Contributes to the Regulation of Epithelial Marker Genes in PC3 PDGF-D Cells, Partly by Regulation of ZEB1, ZEB2, and Snail2
To test whether miR-200 contributes to the regulation of epithelial marker genes in PC3 PDGF-D cells, we first determined the expression levels of miR-200 in PC3 Neo and PC3 PDGF-D cells. We found that the expression of miR-200a, miR-200b, and miR-200c was significantly downregulated in PC3 PDGF-D cells compared with PC3 Neo cells, whereas there was no significant difference in the expression level of miR-16 (Fig. 3A).
In order to mechanistically investigate the role of miR-200 in the regulation of ZEB1 and ZEB2 during the processes of EMT, we transfected PC3 PDGF-D cells with miR-200b and subsequently determined the expression status of transcriptional repressors and epithelial marker genes. We found that the expression levels of not only ZEB1 and ZEB2 but also Snail2 were markedly lower in PC3 PDGF-D cells transfected with miR-200b than in cells transfected with control miRNA, which is consistent with the enhanced expression of epithelial marker genes, such as those encoding E-cadherin, stratifin, EpCAM, F11R, and connexin 26 (Fig. 3B, 3C). These results clearly suggest that loss of miR-200b contributes to the induction of ZEB1, ZEB2, and Snail2, resulting in negative regulation of epithelial marker factors during acquisition of the EMT phenotype in PDGF-D overexpressing PC3 cells.
Re-expression of miR-200b Induces MET in PC3 PDGF-D Cells
Because transfection of miR-200b into PC3 PDGF-D cells resulted in upregulation of the expression of epithelial marker genes, we sought to assess whether the EMT phenotype was reversed in these cells. As expected, we found that PC3 PDGF-D cells transfected with miR-200b displayed a round-like morphology and adhered together after 3 days of transfection (Fig. 4A). The results from immunofluorescence staining showed that the expression of E-cadherin was significantly greater and mainly located on the cell membrane at cell–cell junctions (Fig. 4B). Concomitantly, we found that ZO-1 was localized to the cell membrane at tight junctions in miR-200b-transfected PC3 PDGF-D cells. In contrast, ZO-1 was disrupted from tight junctions in PC3 PDGF-D cells transfected with control miRNA, as shown in Figure 4C. In addition, the patterns of expression and distribution of mesenchymal markers were different in PC3 PDGF-D cells transfected with miR-200b. We found a dramatically lower expression level of vimentin in PC3 PDGF-D cells transfected with miR-200b (Fig. 4D). We also found differences in the expression patterns of F-actin. A cortical actin pattern, characteristic of the epithelial phenotype, was observed in PC3 PDGF-D cells transfected with miR-200b, whereas actin stress fiber, consistent with characteristics of the mesenchymal phenotype, was found in PC3 PDGF-D cells transfected with control miRNA (Fig. 4E). These results suggest that the loss of miR-200 expression found in PC3 PDGF-D cells contributes to the upregulation of ZEB1, ZEB2, and Snail2, resulting in downregulated expression of epithelial marker genes and acquisition of the EMT phenotype, and that this process could be reversed (acquisition of the MET phenotype) by the re-expression of miR-200b in PC3 PDGF-D cells.
miR-200b Inhibits the Adhesion and Invasion of PC3 PDGF-D Cells
In order to demonstrate whether or not miR-200b could affect the behaviors of PC3 PDGF-D cells, we tested cell migration and invasion and found that transfection of PC3 PDGF-D cells with miR-200b significantly inhibited the migration and invasion of PC3 PDGF-D cells (Fig. 5A). It is well known that cell detachment from the matrix in the growth environment and attachment to a secondary site is the “hallmark” of cell migration and invasion during the metastatic process. Interestingly, we found that PC3 PDGF-D cells displayed enhanced detachment and attachment (Fig. 5B, 5C). More importantly, transfection of PC3 PDGF-D cells with miR-200b markedly reduced cell detachment and attachment, compared with transfection with control miRNA (Fig. 5D). These results clearly suggest that re-expression of miR-200b inhibits migration and invasion of PC3 PDGF-D cells through reversal of the EMT phenotype to the MET phenotype. Thus, we believe that a novel therapeutic strategy by which miR-200b could be re-expressed in invasive human prostate cancer would become a useful approach for the treatment of invasive and metastatic prostate cancer in the future. However, the questions remain whether PDGF-D is in fact responsible for the acquisition of EMT characteristics in PC3 cells and whether or not the repression of miR-200 is mechanistically linked with this process. Therefore, we assessed the effect of purified active PDGF-D protein treatment on PC3 cells.
PDGF-D Protein Treatment Represses the Expression of miR-200 Family Members and Induces EMT Characteristics
In order to further verify whether PDGF-D could indeed be responsible for induction of the EMT phenotype of PC3 cells through regulating expression of the miR-200 family, we determined the expression levels of miR-200 family members and assessed the expression of ZEB1, ZEB2, Snail2, vimentin, and E-cadherin in PC3 cells chronically treated with a purified active form of PDGF-D. Treatment of cells with purified PDGF-D protein resulted in significant repression of the expression of miR-200a, miR-200b, and miR-200c (Fig. 6A), which is consistent with findings that were obtained from PC3 PDGF-D cells. Moreover, we also found that PDGF-D protein treatment significantly increased the expression of ZEB2, Snail2, and vimentin at the mRNA level, with a concomitantly lower expression level of E-cadherin at both the mRNA and protein levels (Fig. 6B–6D). Most importantly, PDGF-D treatment dramatically enhanced cell detachment and migration of PC3 cells (Fig. 6E, 6F), suggesting that PDGF-D is indeed responsible for induction of the EMT phenotype in PC3 cells, which is in part mediated via downregulation of miR-200 expression and regulation of its target genes.
Overexpression of PDGF-D in LNCaP Cells Induces EMT
To demonstrate whether PDGF-D could induce EMT in different prostate cancer cell lines, we established stable LNCaP cell lines overexpressing PDGF-D through transfection of LNCaP cells with PDGF-D plasmid. We found that overexpression of PDGF-D in LNCaP cells resulted in a significantly lower expression level of ZO-1 (an epithelial marker) and greater expression levels of vimentin and fibronectin (Fig. 7A), with concomitantly lower levels of miR-200b and miR-200c (Fig. 7B). More interestingly, we found that mRNA levels of PDGF-D in LNCaP cells were significantly lower than those of PC3 cells (Fig. 7C), whereas levels of miR-200, especially miR-200c, were dramatically higher in LNCaP cells than in PC3 cells (Fig. 7D). It is well known that LNCaP cells are less invasive than PC3 cells. To investigate whether miR-200 could contribute to the regulation of invasion of LNCaP cells, we transfected LNCaP cells with anti-miR-200a, anti-miR-200b, and anti-miR-200c and performed an invasion assay. The results show that LNCaP cells transfected with anti-miR-200a, anti-miR-200b, and anti-miR-200c had a significantly greater cell invasion capacity than cells transfected with anti-miRNA control (Fig. 7E). These results clearly suggest that overexpression of PDGF-D could induce EMT, mediated in part by regulating the expression of miR-200 family members not only in PC3 cells but also in LNCaP cells.
PDGF-D has been demonstrated to be expressed in many tumor cell lines [30, 31] and in prostate tumor tissues [21, 22], suggesting that PDGF-D could play an important role in the development and progression of prostate cancer. In our previous studies, we reported that the sustained overexpression of PDGF-D was responsible for acquisition of the EMT phenotype in PC3 cells, with greater invasive characteristics and a greater tumor growth rate in a xenograft model [23, 32]. However, the mechanistic role of PDGF-D in the greater context of tumorigenesis and in prostate cancer progression remains to be elucidated. Acquisition of the EMT phenotype of epithelial tumor cells is believed to play critical roles in the increased invasion and metastatic behaviors of tumor cells during tumor progression, and this process is reminiscent of “cancer stem-like cell” characteristics. Transcription repressors, including the ZEB and Snail families, are strongly linked to induction of the EMT phenotype induced by various factors. In our current study, we found that the expression levels of ZEB1, ZEB2, and Snail2/Slug were significantly greater in PC3 PDGF-D cells, which is consistent with the EMT characteristics of these cells, concomitant with lower expression levels of epithelial markers, such as E-cadherin, EpCAM, sciellin, stratifin, CRB3, connexin 26, and F11R.
To gain further insight into whether ZEB1 could repress the expression of the above-mentioned epithelial markers, we knocked down the expression of ZEB1 by transfection of PC3 PDGF-D cells with ZEB1-specific siRNA. Transfection of PC3 PDGF-D cells with ZEB1 siRNA led to higher expression levels of E-cadherin, CRB3, sciellin, and F11R, all of which are markers of epithelial cells. However, knockdown of ZEB1 by specific siRNA not only reduced the expression of ZEB1 but also downregulated the mRNA level of ZEB2, which is consistent with findings reported by other investigators [24, 29]. Taken together, these results suggest that the transcription factors ZEB1, ZEB2, and Snail2/Slug can mediate the induction of the EMT phenotype observed in PC3 PDGF-D cells and that the downregulation of specific transcription factors may prove to be useful for reversal of the EMT phenotype.
It is also tempting to speculate that these transcription factors could be regulated in cells by novel mechanisms. Indeed, recent findings suggest that miRNA, especially the miR-200 family, could regulate the processes of EMT by repressing the expression of ZEB1 and ZEB2 through targeting the 3′-UTR of ZEB1 and ZEB2 mRNA [4, 5, 24, 33]. Levels of mature miRNAs within cells are regulated by many complex steps, wherein p68 RNA helicase is required for processing primary miRNAs into miRNA precursors . Interestingly, phosphorylation of p68 RNA helicase at Y593 by PDGF-B has been shown to mediate the induction of EMT , suggesting that PDGF-B could induce EMT through regulating the expression of miRNAs by phosphorylating p68 RNA helicase.
This tantalizing and emerging evidence showing regulation of mRNAs by miRNAs in the processes of EMT prompted us to investigate whether the miR-200 family could play a role in PDGF-D-induced EMT of prostate cancer cells and further assess whether the processes of EMT in our cell culture model system could also be regulated by the ZEB1, ZEB2, and Snail2 transcription factors. We found that the expression levels of miR-200a, miR-200b, and miR-200c were significantly lower in LNCaP PDGF-D and PC3 PDGF-D cells as well as in PC3 cells exposed to PDGF-D protein than control cells, suggesting that the miR-200 family may regulate the EMT process in our model. ZEB1 contains three putative binding sites for miR-200a and five for miR-200b/miR-200c [4, 33], and ZEB2 harbors three putative binding sites for miR-200a and six for miR-200b . In addition, we found that the level of miR-200b was eightfold higher than that of miR-200c in PC3 cells, suggesting that miR-200b could play a central role in the negative regulation of ZEB1 and ZEB2 expression, and thus the loss of miR-200b leads to the upregulation of ZEB1 and ZEB2, resulting in the acquisition of EMT induced by PDGF-D. Therefore, we elected to re-express miR-200b in PC3 PDGF-D cells using transfection studies in order to gain further mechanistic insight into the regulation of ZEB1 and ZEB2 and the biological consequence of re-expression of miR-200b on the EMT phenotype. The transfection of PC3 PDGF-D cells with miR-200b significantly reduced the expression of ZEB1, ZEB2, and Snail2 at both the mRNA and protein levels, with concomitantly greater expression levels of epithelial markers such as E-cadherin, stratifin, CRB3, EpCAM, F11R, and connexin 26. In addition, PC3 PDGF-D cells transfected with miR-200b displayed an epithelial phenotype, suggesting that the loss of miR-200b contributes to acquisition of the EMT phenotype and that the re-expression of miR-200b could result in reversal of the EMT phenotype to the MET phenotype.
The processes of EMT have been linked to cell migration and invasion, and the re-expression of miR-200b in PC3 PDGF-D cells led to reversal of the EMT phenotype. In this study, we found that transfection of PC3 PDGF-D cells with miR-200b significantly inhibited cell migration and invasion of PC3 PDGF-D cells. It is well known that the metastatic process of cancer cells requires cell detachment from the site of origin, intravasation, translocation through blood and lymphatic vessels, extravasation, attachment to the secondary site, and colonization. Moreover, cell detachment from the basement membrane and reattachment play critical roles during cell migration and invasion as well as tumor cell metastasis. Interestingly, we found that PC3 PDGF-D cells exhibited a significant enhancement in cell detachment from the culture surface and attachment to the culture surface. More importantly, transfection of PC3 PDGF-D cells with miR-200b remarkably reduced the ability of PC3 PDGF-D cells to attach to and detach from the culture surface, and our results are consistent with the role of miR-200b re-expression in PC3 PDGF-D cells, with reversal of EMT to MET characteristics with a less invasive phenotype. More interestingly, LNCaP cells with a lower invasive capacity possessed a lower expression level of endogenous PDGF-D and higher levels of miR-200 family members, and overexpression of PDGF-D in LNCaP cells resulted in the downregulation of miR-200 expression, which was associated with EMT characteristics.
Our earlier results showed that the mammalian target of rapamycin (mTOR) and nuclear factor (NF)-κB pathways play critical roles in the induction of EMT induced by overexpression of PDGF-D in PC3 cells . In the current study, we show that the loss of expression of miR-200 family members, especially miR-200b, is in part responsible for the induction of EMT in PC3 PDGF-D cells. It is well known that NF-κB plays an important role in mediating the processes of EMT induced by different factors through upregulation of the transcription repressor function of ZEB1 and ZEB2 [29, 35], which in turn repress the expression of miR-200 family members by binding to the E-box sequence of the miR-200 promoter [27, 33]. More interestingly, miR-200 could downregulate the expression of ZEB1 and ZEB2 by interacting with the 3′-UTR of ZEB1 and ZEB2 mRNA [4, 27]. Moreover, the mTOR pathway could indirectly regulate NF-κB activity by regulating glycogen synthase kinase (GSK)-3β phosphorylation . These results also signify a double-negative feedback loop between miR-200 and ZEB1/ZEB2 that allows the maintenance of the EMT phenotype, even after cessation of the inducing signal, which could become a critical target for the reversal of EMT. Together, our results are consistent with findings that the re-expression of miR-200b could reverse the EMT phenotype induced by PDGF-D, and in this process the mTOR pathway could indirectly regulate NF-κB activity by regulating GSK-3β phosphorylation, thereby establishing a mechanistic link in the regulatory function of miR-200, although further mechanistic studies are warranted.
In order to further demonstrate whether PDGF-D could indeed be responsible for induction of the EMT phenotype of PC3 cells by regulating the expression of miR-200 family members, we analyzed the expression levels of miR-200 family members and transcription repressors as well as EMT molecular markers in PC3 cells chronically treated with a purified active form of PDGF-D. We found that treatment of cells with purified PDGF-D protein resulted in significantly lower expression levels of miR-200a, miR-200b, and miR-200c, and significantly greater expression levels of ZEB2, Snail2, and vimentin at the mRNA level, with a concomitant lower expression level of E-cadherin. Most importantly, PDGF-D treatment dramatically enhanced cell detachment and migration of PC3 cells, suggesting that PDGF-D is indeed responsible for induction of the EMT phenotype in PC3 cells, which is in part mediated via the downregulation of miR-200 expression and the regulation of target genes. However, we found that chronic exposure of PC3 cells to PDGF-D protein was not sufficient to induce significant morphological differences, compared with those seen in PDGF-D-transfected PC3 cells. This observation is consistent with our previous findings showing that PDGF-D-transfected cells in early passages retain epithelial characteristics, whereas EMT characteristics could be seen only in later passages (passage >7), suggesting that chronic exposure to a high concentration of PDGF-D is required for induction and maintenance of the EMT phenotype in PC3 cells. Another explanation could be that an intracrine mechanism exists for PDGF-D action, which may not function when cells are exogenously treated with purified PDGF-D, in contrast to intracellularly synthesized PDGF-D in PC3 PDGF-D cells. Interestingly, recent studies have shown that many growth factors, such as vascular endothelial growth factor, fibroblast growth factor, epidermal growth factor, and PDGF, have an intracrine mechanism of signaling [36–39], which could be one of the mechanism by which PDGF-D overexpression contributes to EMT, as seen in our study.
In conclusion, we have developed a cell culture model for understanding the molecular regulation of EMT and its reversal to the MET phenotype. More importantly, we found that PDGF-D could induce EMT through regulating the expression of miR-200 family members, which was associated with the deregulation of ZEB1, ZEB2, and Snail2. Based on our novel findings, we believe that the re-expression of miR-200b in human prostate cancer by innovative approaches could be useful for designing strategies for the treatment of invasive and metastatic prostate cancer.
This work was partly funded by grants from the National Cancer Institute, NIH (5R01CA108535 to F.H.S.). This work was also partly funded by the Puschelberg foundation.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.