Department of Pathology, Karmanos Cancer Institute, Wayne State University School of Medicine, Detroit, Michigan, USA
Department of Pathology, Barbara Ann Karmanos Cancer Institute, Wayne State University School of Medicine, 740 Hudson Webber Cancer Research Center, 4100 John R. Street, Detroit, Michigan 48201, USA. Telephone: 313-576-8327; Fax: 313-576-8389
The majority of human malignancies are believed to have epithelial origin, and the progression of cancer is often associated with a transient process named epithelial-mesenchymal transition (EMT). EMT is characterized by the loss of epithelial markers and the gain of mesenchymal markers that are typical of “cancer stem-like cells,” which results in increased cell invasion and metastasis in vivo. Therefore, it is important to uncover the mechanistic role of factors that may induce EMT in cancer progression. Studies have shown that platelet-derived growth factor (PDGF) signaling contributes to EMT, and more recently, PDGF-D has been shown to regulate cancer cell invasion and angiogenesis. However, the mechanism by which PDGF-D promotes invasion and metastases and whether it is due to the acquisition of EMT phenotype remain elusive. For this study, we established stably transfected PC3 cells expressing high levels of PDGF-D, which resulted in the significant induction of EMT as shown by changes in cellular morphology concomitant with the loss of E-cadherin and zonula occludens-1 and gain of vimentin. We also found activation of mammalian target of rapamycin and nuclear factor-κB, as well as Bcl-2 overexpression, in PDGF-D PC3 cells, which was associated with enhanced adhesive and invasive behaviors. More importantly, PDGF-D-overexpressing PC3 cells showed tumor growth in SCID mice much more rapidly than PC3 cells. These results provided a novel mechanism by which PDGF-D promotes EMT, which in turn increases tumor growth, and these results further suggest that PDGF-D could be a novel therapeutic target for the prevention and/or treatment of prostate cancer.
Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: D.K.: conception and design, execution, collection of data and data analysis, manuscript writing; Z.W.: experimental design and execution, data analysis; S.H.S.: experimental execution, collection of data; Y.L.: experimental design, data analysis, manuscript writing; S.B.: experimental design, execution of animal experiment, data analysis; A.S.: animal experiments, data collection, data analysis; H.-R.C.K.: development of PC3 PDGF-D cell line, experimental design; M.L.C.: animal experiment, data interpretation; F.H.S.: principal investigator, conception and design, laboratory facility and financial support, experimental design, data collection and interpretation, manuscript writing, final approval of manuscript.
Epithelial-mesenchymal transition (EMT) is a process by which epithelial cells undergo remarkable morphological changes characterized by a transition from epithelial cobblestone phenotype to elongated fibroblastic phenotype. The process of EMT involves loss of epithelial cell-cell junction, actin cytoskeleton reorganization, and upregulation of mesenchymal molecular markers, such as vimentin, fibronectin, and N-cadherin . A disassembly of cell-cell junction, including downregulation and relocation of E-cadherin and zonula occludens-1 (ZO-1), as well as downregulation and translocation of β-catenin from cell membrane to nucleus, results in the induction of EMT [2, –4]. Epithelial cells have a regular cell-cell junction and adhesion that inhibits cell movement of individual cells. In contrast, mesenchymal cells have less strong adhesion between cells compared with their epithelial counterparts, rendering mesenchymal cells more motile and invasive, and these processes are consistent with the acquisition of cancer stem-like cell phenotype . EMT is triggered by the interplay of extracellular signals, such as collagen and growth factors, including transforming growth factor-β (TGF-β), fibroblast growth factor (FGF), epidermal growth factor (EGF), and platelet-derived growth factors (PDGFs)-A and -B [5, , , –9]. However, the mechanism by which EMT occurs remains to be elucidated, especially what role, if any, the newly identified PDGF-D plays in the processes of EMT; this question requires in-depth investigation.
PDGF-D is known to regulate many cellular processes, including invasion and angiogenesis, by activating its cognate receptor (platelet-derived growth factor receptor-β [PDGFR-β]) [10, 11]. PDGF-D is believed to be activated by removing the N-terminal complement subcomponent C1r/C1s, Uegf, Bmp1 domain to make the C-terminal growth factor domain active [12, 13]. It is known that growth factors, such as PDGF and EGF, can activate phosphoinositide 3-kinase/Akt through activation of receptor tyrosine kinase and thereby associate the mammalian target of rapamycin (mTOR) pathway in transducing cellular signaling . Recent studies have shown that mTOR is required for cell motility, which is mediated by S6K and 4E-BP1 , downstream targets of Akt/mTOR pathway.
Increasing evidence suggests that nuclear factor-κB (NF-κB) activation by a variety of inflammatory mediators and growth factors [16, 17] is associated with the processes of angiogenesis, invasion, metastasis, and antiapoptosis [18, 19], which have been shown to contribute to the development and/or progression of prostate cancer . Interestingly, recent studies have shown that activation of the NF-κB pathway is required for the induction and maintenance of epithelial-mesenchymal transition [21, 22]. This is partly due to NF-κB-mediated repression of E-cadherin expression and induction of vimentin expression by upregulating transcription factor ZEB1 and snail, which inhibits E-cadherin expression during EMT [23, 24]. Additional studies have shown that Rel B, a subunit of NF-κB, could induce Bcl-2 expression and promote invasion of estrogen receptor-negative breast cancer cells, which are associated with EMT . Moreover, Bcl-2 has been shown to be involved in the processes of angiogenesis, invasion, and metastasis [25, 26]. However, the detailed mechanism by which PDGF-D, NF-κB, Bcl-2, and their interplay regulate EMT and thereby regulate the processes of invasion, angiogenesis, and metastasis of tumor cells has not been elucidated, especially in prostate cancer.
Therefore, in this study, we have used PDGF-D-overexpressing PC3 cells to investigate the role of PDGF-D in the induction of EMT and its underlying mechanism. We found that overexpression of PDGF-D led to the induction of EMT, as documented by the loss of E-cadherin and ZO-1 expression and the gain of mesenchymal markers such as vimentin, which was also consistent with dramatic changes in cellular morphology and growth kinetics. We also found activation of NF-κB and mTOR signaling, resulting in the increased expression of Bcl-2, which was associated with increased invasive behavior of PDGF-D-overexpressing PC3 cells. Moreover, the cells with EMT characteristics grew significantly faster in an animal model of experimental bone metastasis compared with control cells. These results suggest that overexpression of PDGF-D contributes to the acquisition of cancer stem-like cell characteristics, which is consistent with the induction of EMT, and further suggest that PDGF-D could be a potential preventive and/or therapeutic target in prostate cancer.
Materials and Methods
Cell Lines and Culture Condition
Prostate cancer cell line PC3 cells and resultant transfected cell lines were maintained in RPMI 1640 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 5% fetal bovine serum (FBS), 2 mmol/l glutamine, 10 μmol/l Hepes, 100 units/ml penicillin, and 100 μg/ml streptomycin. All cells were cultured in a 5% CO2 humidified atmosphere at 37°C.
Generation of Stable Cell Lines Overexpressing PDGF-D
Establishment of PDGF-D-overexpressing PC3 cells (pooled stable clones to rule out the possibility of artifacts that could be due to clonal selection) by transfection of PC3 cells with pcDNA3-PDGF-D:His or the corresponding empty vector pcDNA3 Neo was previously described ; these cells are referred to as PC3 PDGF-D or PC3 Neo cells.
Research Reagents and Antibodies
Antibodies against mTOR, phospho-mTOR (Ser2,448), 4E-BP1, phospho-4E-BP1 (Thr37/Thr46), β-catenin, and raptor were purchased from Cell Signaling Technology (Beverly, MA, http://www.cellsignal.com). Antibodies against vimentin and nestin were purchased from Abcam (Cambridge, MA, http://www.abcam.com). Antibodies against E-cadherin and PDGFR-β (P-20) were obtained from Santa Cruz Biotechnology Inc. (Santa Cruz, CA, http://www.scbt.com). Antibody against NF-κB (p65 antibody) was purchased from Millipore (Billerica, MA, http://www.millipore.com). The monoclonal antibody to β-actin was purchased from Sigma-Aldrich (St. Louis, http://www.sigmaaldrich.com). The monoclonal antibody to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was purchased from Affinity Bioreagents (Golden, CO, http://www.bioreagents.com). Antibody against human Bcl-2 was from Dako North America, Inc. (Carpinteria, CA, http://www.dako.com). Antibodies against PDGF-D, 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 (Hercules, CA, http://www.bio-rad.com).
Plasmid Constructs, Small Interfering RNA, and Transfection
To insert Bcl-2 into mammalian expression vector pcDNA3, pSPUTKBcl2 plasmid was used as the template. Bcl-2 coding region was amplified via polymerase chain reaction (PCR) by using the following primers to add an EcoRI site at the 5′ end and an XhoI site at the 3′ end of the Bcl-2 open reading frame: forward, 5′GGAATTCATGGCGCACGCTGGGAGAACG-3′; reverse, 5′-GGCCCTCGAGTCACTTGTGGCTCAGATAGG-3′. The PCR product was then digested with EcoRI and XhoI and inserted into the pcDNA3 vector. PC3 cells were transfected with Bcl-2 plasmid and incubated for 48 hours. Cell lysates were prepared for Western blot analysis. PC3 Neo and PC3 PDGF-D cells were transfected with mTOR, raptor small interfering RNA (siRNA), or control siRNA (100 nmol/l; Santa Cruz Biotechnology) using DharmaFECT3 siRNA transfection reagent (Dharmacon, Inc., Lafayette, CO, http://www.dharmacon.com). The media were removed after 24 hours of transfection, and then the cells were incubated in media containing 5% FBS for another 24 hours. The cells were collected for invasion assay, or cell lysates were prepared for Western blot analysis.
Preparation and Processing of Conditioned Media
PC3 Neo and PC3 PDGF-D cells were seeded in six-well plates and incubated for 24 hours in media containing 5% FBS. Cells were then washed with phosphate-buffered saline (PBS) and incubated in media containing 0.5% FBS for 48 hours. Conditioned media (CM) were collected from the culture and centrifuged at 10,000g for 10 minutes at 4°C to remove the cells or cell debris. Media were collected and used for Western blot assay.
Cell Invasion Assay
Cell invasion assay was performed using the BD BioCoat Tumor Invasion Assay System (BD Biosciences, Bedford, MA, http://www.bdbiosciences.com) according to the manufacturer's instructions. Briefly, PC3 Neo, PC3 PDGF-D cells, and PC3 PDGF-D cells transfected with mTOR and raptor siRNA or control siRNA suspended in serum-free media were seeded into the upper chamber of the system. The bottom wells were filled with media containing 10% FBS. After 24 hours of incubation, the cells were stained with 4 μg/ml Calcein AM (Invitrogen) in PBS at 37°C for 1 hour. The fluorescently labeled cells were photographed under a fluorescence microscope. The fluorescence of the invaded cells was read in an ULTRA Multifunctional Microplate Reader (TECAN, San Jose, CA, http://www.tecan.com) at excitation/emission wavelengths of 485/530 nm.
Western Blot Analysis
Western blot analysis was performed using CM or cell lysates. Cell lysates from different experiments were obtained by scraping the cells from the dishes. The cell pellet, washed twice with cold PBS, was suspended in 125 mmol/l Tris-HCl (pH 6.8) and sonicated for 10 seconds, and an equal volume of 4% SDS was added. The lysates were boiled for 10 minutes. CM or equal amounts of proteins from cell lysates were separated by SDS-polyacrylamide gel electrophoresis, transferred to a nitrocellulose membrane, blocked, and incubated with specific primary antibodies overnight at 4°C. The membrane was washed and incubated with the respective secondary antibodies conjugated with peroxidase. Protein detection was performed with a chemiluminescence detection system (Pierce, Rockford, IL, http://www.piercenet.com).
Real-Time Reverse Transcription-PCR
PC3 Neo and PC3 PDGF-D cells were grown in six-well plates with medium containing 5% FBS for 48 hours. Alternatively, PC3 Neo and PC3 PDGF-D cells were transfected with PDGF-D siRNA or control siRNA and incubated for 48 hours. The total RNA was isolated using the Trizol reagent. Two micrograms of RNA were reverse-transcribed using a reverse-transcription system (Invitrogen) according to the manufacturer's instructions. Real-time PCR was used to quantify mRNA expression. The sequence of oligonucleotide primers for PDGF-D and PDGFR-β used in this study was described by Pohlers et al. . β-Actin or GAPDH was used for the internal control to correct the potential variation in RNA loading. Targets and β-actin or GAPDH genes were run under the same conditions. All reactions were performed in 25 μl of volume containing the sample cDNA, platinum SYBR Green qPCR Supermix-UDG (Invitrogen), and primers. Before the PCR cycles, samples were incubated for 2 minutes at 50°C and 2 minutes at 95°C. Thermal cycles consisted of 45 cycles at 95°C for 30 seconds and 60°C for 30 seconds, with melting curve.
Cells were fixed with 4% paraformaldehyde for 10 minutes, permeabilized in 0.5% Triton X-100 for 10 minutes, and incubated in PBS and 10% goat serum blocking solution for 1 hour. The cells were incubated for 2 hours with anti-PDGF-D, anti-E-cadherin, anti-ZO-1, anti-nestin, anti-vimentin, anti-β-catenin, or PDGFR-β antibody diluted (1:50, 1:25, 1:100, 1:100, 1:10, 1:50, and 1:50, respectively) in PBS and 5% goat serum and were stained for 1 hour with Alexa Fluor 594-conjugated secondary antibody (1:500). 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 Adhesion Assay
PC3 Neo and PC3 PDGF-D cells were trypsinized. Cells were counted and seeded in six-well plates for 30 minutes. Unattached cells were removed by washing twice with PBS, and the attached cells were counted after trypsinization.
Electrophoretic Mobility Shift Assay for NF-κB DNA-Binding Activity
PC3 Neo and PC3 PDGF-D cells were seeded in a six-well plate. After 48 hours of incubation, nuclear extracts were prepared according to the method described by our laboratory previously . For detecting NF-κB DNA-binding activity, electrophoretic mobility shift assay was performed by incubating 3 μg of nuclear protein of each sample with IRDye 700-labeled NF-κB oligonucleotide (LI-COR, Lincoln, NE, http://www.licor.com) as described previously . Retinoblastoma protein was used as a protein loading control.
Tumor Implant and Growth in SCID Mice
Male homozygous CB-17 SCID/SCID mice (4 weeks old) were purchased from Taconic Farms (Germantown, NY, http://www.taconic.com). The mice were maintained according to the Animal Care and Use Guidelines approved by the NIH. Mice received Lab Diet 5021 (Purina Mills, Inc., Richmond, IN, http://www.purinamills.com). Human male fetal bone tissue was obtained by a third-party nonprofit organization (Advanced Bioscience Resources, Alameda, CA, http://www.abr-inc.com). After 1 week of acclimatization, the mice were implanted with a single human fetal bone fragment as described previously . After 1 month, 16 mice with the bone implants were randomized into two groups (n= 8): (a) PC3 Neo and (b) PC3 PDGF-D. PC3 Neo and PC3 PDGF-D cells were harvested and resuspended in serum-free RPMI medium. Only suspensions consisting of single cells with >90% viability were used for the injections. Cells (1 × 105) in 20 μl of serum-free RPMI medium were injected intraosseously by insertion of a 27-gauge needle through the mouse skin directly into the marrow surface of the previously implanted bone. In our previous experience with this model, we found a tumor take rate of >90%. The majority of the bone implants began to enlarge (at which point, they were called bone tumors), as determined by caliper measurements approximately 20 days after cancer cell injection. For initial comparison of growth kinetics between Neo and the PDGF-D-transfected cell line using this model system, we compared the bone tumor volume of each group by twice-weekly caliper measurements (after subtracting the initial bone volume). The body weight of mice in each group was also measured. All mice were sacrificed 5 weeks after tumor cell injection and could not be continued further because of large tumor size in the PDGF-D-transfected group of mice.
Experiments presented in the figures are representative of three or more different repetitions. The data are presented as the mean values ± SE. Comparisons between groups were evaluated by a two-tailed Student's t test. Values of p < .05 were considered to be statistically significant.
PDGF-D Induces Changes in the Morphology of PC3 Cells
To unveil the possible role of PDGF-D in the development and progression of prostate cancer, we used an overexpression model in which PC3 cells were stably transfected with PDGF-D expression plasmid or empty vector pcDNA3 plasmid. Real-time reverse transcription (RT)-PCR showed that PDGF-D mRNA was dramatically increased in PC3 PDGF-D cells compared with PC3 Neo cells (Fig. 1A), whereas there was no change in the expression of β-actin mRNA. Moreover, we observed a significantly increased full-length form, as well as active form, of PDGF-D in PC3 PDGF-D cells (Fig. 1B), whereas there was no change in the expression of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) protein. The results from immunofluorescence staining indicate that PDGF-D located mainly in the cytoplasm in PC3 PDGF-D cells (Fig. 1C). Interestingly, we found that PDGF-D-overexpressing PC3 cells displayed elongated/irregular fibroblastoid morphology (Fig. 1D, lower panel). In contrast, PC3 Neo cells had a rounded shape, typical of an epithelial cobblestone appearance (Fig. 1D, upper panel). These changes in phenotype suggested that PDGF-D could induce EMT of PC3 cells.
Changes in Markers of Epithelial and Mesenchymal Phenotypes
To further confirm whether PDGF-D could really induce EMT in PC3 cells, we determined the expression and location of E-cadherin, ZO-1, and β-catenin using immunofluorescence staining. PC3 PDGF-D cells acquired their elongated shape, which was consistent with the loss or relocation of E-cadherin at cell-cell junction (Fig. 2A). Concomitantly, ZO-1 was disrupted from tight junction (Fig. 2B), and the β-catenin was redistributed from cell membrane to the nuclear compartment in PC3 PDGF-D cells (Fig. 2C). In addition, the patterns of expression and distribution of mesenchymal markers were changed during EMT induced by PDGF-D overexpression. The results from immunostaining indicate that PC3 PDGF-D cells had increased levels of expression of vimentin and nestin (Fig. 2D, 2E). F-actin reorganization and a diffuse pattern were also observed in PC3 PDGF-D cells (Fig. 2F), which were correlated with EMT phenotype and, as such, were consistent with the acquisition of cancer stem-like cell or “stemness” characteristics.
Overexpression of PDGF-D in PC3 Cells Downregulates PDGFR-β Expression
PDGF-D has been shown to regulate many cellular processes by activating its cognate receptor PDGFR-β. Therefore, we analyzed the levels of PDGFR-β mRNA using real-time RT-PCR. We found that PDGFR-β mRNA was downregulated in PC3 PDGF-D cells compared with PC3 Neo (Fig. 3A), consistent with the results obtained from Western blot analysis showing that the expression of PDGFR-β proteins was reduced in PC3 PDGF-D cells compared with PC3 Neo cells (Fig. 3B), whereas there was no change in the expression of GAPDH protein (used as a protein loading control). To further detect the expression levels of PDGFR-β in PDGF-D-overexpressing PC3 cells, we have immunostained PC3 Neo and PC3 PDGF-D cells with anti-PDGFR-β antibody and found that expression of PDGFR-β was higher in PC3 Neo cells than in PC3 PDGF-D cells (Fig. 3C). To determine whether overexpression of PDGF-D was responsible for reduced PDGFR-β expression, we knocked down PDGF-D expression with PDGF-D siRNA in PC3 PDGF-D cells. As shown in Figure 3D and 3E, knockdown of PDGF-D expression significantly increased the expression of PDGFR-β mRNA and protein, and there was no change in the expression of GAPDH protein (used as protein loading control). These results clearly suggest that prolonged PDGF-D stimulation resulted in a decreased PDGFR-β expression in PC3 PDGF-D cells and that the knockdown of PDGF-D resulted in the restoration of PDGFR-β expression in PC3 PDGF-D cells.
Activation of mTOR Pathway Is Involved in EMT Induced by PDGF-D Overexpression
Our results showing changes in morphology suggested that PC3 cells underwent EMT induced by overexpression of PDGF-D and that the cells with EMT phenotype exhibited an increase in cell proliferation . mTOR serves as a sensor and integrator of many stimuli, such as growth factors and nutrients, and is also involved in the control of cell growth and proliferation. To determine whether mTOR pathway is involved in EMT induced by PDGF-D overexpression, we assessed the levels of phosphorylated mTOR and its downstream target 4E-BP1. PC3 PDGF-D cells displayed an increased activation of mTOR concomitant with the loss of E-cadherin and gain of vimentin and nestin (Fig. 4A). Moreover, we found that β-catenin was significantly downregulated in PDGF-D-overexpressing cells (Fig. 4B) and that there was no change in the expression of GAPDH protein (used as protein loading control). To further confirm the activity of mTOR involved in EMT induced by PDGF-D overexpression in PC3 cells, we knocked down the expression of mTOR or raptor using specific siRNA, respectively. As shown in Figure 3C, siRNA specific for mTOR or raptor was effective at reducing the expression of these two proteins without any change in the expression of GAPDH protein (used as protein loading control). Knockdown of mTOR was associated with a concomitant increase in the expression of E-cadherin, whereas knockdown of raptor expression reduced the level of vimentin in PDGF-D cells, suggesting that the activation of mTOR is associated with the acquisition of EMT phenotype induced by overexpression of PDGF-D.
PDGF-D Increases Adhesion and Invasion of PC3 Cells
Numerous observations support the concept that EMT has a central role in tumor invasion and metastasis . In this study, we found that PDGF-D could induce EMT in PC3 cells; therefore, we hypothesized that PDGF-D overexpression could result in increased adhesion and invasion of PC3 cells. To prove this hypothesis, we performed adhesion and invasion assays. We found that overexpression of PDGF-D significantly promoted invasion of PC3 cells through Matrigel (BD Biosciences) compared with PC3 Neo control cells (Fig. 5A). To further confirm whether mTOR pathway was responsible for the increased invasion of PC3 PDGF-D cells, we knocked down the expression of mTOR or raptor by siRNA as documented above. Transfection of PC3 PDGF-D cells with mTOR or raptor siRNA led to a significant decrease in mTOR or raptor expression, as shown earlier (Fig. 4C), and these cells were found to be less invasive (Fig. 5B, 5C). These results suggest that mTOR activation is mechanistically associated with increased invasion of PDGF-D-overexpressing PC3 cells, which is consistent with EMT characteristics. In addition, overexpression of PDGF-D dramatically increased the attachment of cells within 30 minutes of incubation, suggesting that PDGF-D-overexpressing PC3 cells had a greater ability of adhesion (Fig. 5D).
NF-κB and Its Downstream Target Bcl-2 Contribute to EMT Induced by PDGF-D Overexpression
To further explore the mechanism by which PDGF-D overexpression contributed to the induction of EMT, resulting in increased invasion of PC3 cells, we measured NF-κB expression and DNA-binding activity. The results from Western blot analysis show that expression of NF-κB (p65 protein) was upregulated in cell lysates of PDGF-D-overexpressing PC3 cells compared with control cells (Fig. 6A) and that there was no change in the expression of GAPDH protein (used as protein loading control). Moreover, the activated NF-κB was also increased in the nuclear compartment, which was associated with increased NF-κB DNA-binding activity (Fig. 6B) where Rb Western blot showed equal nuclear protein loading. Since Bcl-2 is one of the downstream target genes of NF-κB and is known to contribute to the induction of EMT and increased cancer cell invasion , we tested whether the activation of NF-κB could lead to the increased levels of Bcl-2 in PDGF-D-overexpressing PC3 cells. We found that the expression of Bcl-2 was remarkably upregulated in PC3 PDGF-D cells compared with PC3 Neo cells (Fig. 6A). These results suggest that the overexpression of PDGF-D results in the activation of NF-κB, thereby upregulating the expression of its target genes, especially Bcl-2, which appears to be partly responsible for the induction of EMT.
To further determine whether Bcl-2 is involved in the induction of EMT, we transfected parental PC3 cells with Bcl-2 expression plasmid or empty vector and also knocked down the expression of Bcl-2 in PDGF-D-overexpressing PC3 cells by transfection with Bcl-2 siRNA. We found that overexpression of Bcl-2 significantly decreased the levels of E-cadherin in the parental PC3 cells (Fig. 6C), whereas knockdown of Bcl-2 repressed the expression of vimentin in PDGF-D-overexpressing cells (Fig. 6D). E-cadherin and vimentin are the key markers of epithelial and mesenchymal cells, respectively. Therefore, these results suggest that Bcl-2 could partly contribute to EMT induced by PDGF-D overexpression in PC3 prostate cancer cells.
PDGF-D Overexpression Increases Tumor Growth Rate in SCID Mice
PDGF-D overexpression led to the induction of EMT in PC3 cells and promoted invasion in vitro, and, as such, it could lead to increased tumor growth rate in vivo. Thus, we have tested whether the growth rate of tumors induced by PC3 PDGF-D cells could be increased compared with PC3 Neo cells using an animal model of experimental bone metastasis . We found that injection of PC3 Neo and PC3 PDGF-D cells had virtually no effect on tumor latency period (20 days after cell implantation, when a measurable tumor volume could be detected after subtracting the implanted bone volume), but there was a significant difference in their growth kinetics. The growth kinetics of PC3 Neo cells were similar to those published earlier ; relative to PC3 Neo group, the PC3 PDGF-D group exhibited similar growth on day 20 (Neo, 61.65 ± 50.33 mm3 vs. PDGF-D, 61.8 ± 41.12 mm3), and then the tumor growth of PC3 PDGF-D cells was very rapid when measured on days 26, 29, and 35, with 8-fold (Neo, 113.83 ± 86.6 mm3 vs. PDGF-D, 952.05 ± 354.63 mm3), 10-fold (Neo, 151.88 ± 111.96 mm3 vs. PDGF-D, 1,578.85 ± 599.33 mm3), and 12-fold (Neo, 207.28 ± 78.74 mm3 vs. PDGF-D, 2,497.31 ± 957.44 mm3), respectively (Fig. 7A, 7B). These results suggest that overexpression of PDGF-D contributes to enhanced tumor growth, and, as such, it is partly due to the acquisition of EMT (cancer stem-like cell or stemness) phenotype; these results are consistent with our results on cell adhesion and invasion assays (Fig. 5).
Progression of most carcinomas toward malignancy is associated with a loss of epithelial differentiation and a switch toward a mesenchymal phenotype, which is accompanied by increased cell motility and invasion. This process is referred to as EMT and, as such, could be similar to the process known as cancer stem-like cell or stemness, which is characterized by the downregulation of molecular markers of epithelial cells together with the loss of intercellular junction, resulting in the reduction of intercellular adhesion, and a gain of mesenchymal molecules . In this study, we found that transfection of PC3 cells with PDGF-D expression plasmid induced changes in the cellular morphology that are consistent with EMT. Moreover, the results from immunostaining showed the downregulation and relocation of E-cadherin and ZO-1 and the redistribution of β-catenin from cell membrane to the nuclear compartment in PC3 PDGF-D cells. Concomitantly, PDGF-D-overexpressing cells showed upregulated expression of vimentin and nestin and induction of F-actin reorganization, all of which are consistent with EMT characteristics.
Although the PC3 cell line is a highly malignant prostate cancer cell line, the majority of EMT studies in prostate cancer have used PC3 and DU145 cells . These cell lines showed expression of molecular markers of cell-cell adhesion junction, such as E-cadherin, with concomitant epithelial-like morphology, which is consistent with the characteristics of primary epithelial tumor cells. Increasing evidence suggests that tumor progression is involved in the occurrence of EMT, which allows tumor cells to acquire the capacity to infiltrate surrounding tissue and metastasize to distant sites [2, 33, 34]. Interestingly, the disseminated mesenchymal tumor cells undergo reverse transition (i.e., mesenchymal to epithelial transition [MET]) at the site of metastasis to allow colonization of tumors in the secondary sites [35, 36]. It is highly likely that prostate cancer cells from the primary site in patients undergo EMT and may also acquire MET characteristics when they arrive at the site of metastasis (such as bone and the brain, from which PC3 and DU145 cells, respectively, were originally derived), which could also be associated with the acquisition of incomplete epithelial phenotype or mixed phenotype, typically known as fused-cell phenotype [1, 3]. Therefore, PC3 cells were chosen in our study for investigating the role of PDGF-D overexpression in the induction of EMT, because PDGF-D is known to be associated with tumor progression via regulation of cell proliferation, invasion, and angiogenesis in a variety of cancer cell lines and tumors . A recent study has shown that PDGF-A plays a critical role in the induction of EMT by TGF-β in liver cancer cells [6, 7]. Interestingly, PDGF-B has also been shown to induce EMT by mediating phosphorylation of p68 RNA helicase, which promotes nuclear translocation of β-catenin . β-Catenin is a component of cell adhesion junctions, connecting E-cadherin to α-catenin to form E-cadherin-β-catenin-α-catenin complex at adhesion junctions that are linked to the actin cytoskeleton . Downregulation of β-catenin is known to mediate the dissociation of the E-cadherin-β-catenin-α-catenin complex from membrane and has been found to be associated with EMT . A recent study has shown that nuclear translocation of β-catenin mediated by PDGF-B is required for EMT . In our study, overexpression of PDGF-D reduced the expression of β-catenin and induced nuclear translocation of β-catenin, and on the basis of our results, we suggest that the regulation of β-catenin plays an important role in PDGF-D-induced EMT. However, it is possible that either chronic exposure of PC3 cells due to overexpression of PDGF-D over a period of several passages or the selection of certain clones during serial passaging could lead to the observed EMT in PDGF-D-overexpressing PC3 prostate cancer cells, although it is highly unlikely because PC3 Neo cells did not show any changes in similar passages. These provocative results further suggest that additional mechanistic studies must be done in detail in the future; these studies are being carried out in our laboratory.
PDGFR-β is the cognate receptor for PDGF-D. In this study, we found that PDGF-D overexpression significantly downregulated the expression of PDGFR-β mRNA and protein in PC3 PDGF-D cells compared with PC3 Neo cells, which is consistent with published results showing that overexpression of PDGF-D in NIH/3T3 cells downregulated both PDGFR-α and PDGFR-β expression and their activation. Despite the observed reduction in receptor phosphorylation in the growth factor transfectants, the downstream mitogenic signaling pathway was still activated, resulting in the higher proliferation in the growth factor transfectants (PDGF-D-transfected NIH/3T3 cells) compared with the respective controls . Earlier studies have clearly shown that chronic exposure of cells to agonist resulted in downregulation of its receptors, due in part to reduced mRNA levels of the receptors [39, , –42], and these findings are consistent with our current results. However, the question was raised of how PDGF-D could sustain EMT events under the condition of lower levels of transcription of its receptor. Studies have shown that the expression of PDGFR-β is associated with the growth status of the cells. Increased expression of PDGFR-β is likely linked with the state of growth arrest at G0. In contrast, the proliferating cells and oncogene-transformed cell lines, without undergoing growth arrest following serum deprivation, express constitutively low levels of PDGFR-β [43, 44]. Additional studies have demonstrated that cell morphology and cytoskeleton elements were associated with the regulation of PDGFR-β activity in p21ras-transformed cells . These findings indicated that downstream targets of PDGF signaling pathway regulate PDGFR-β function by a feedback mechanism. Recent studies have also demonstrated that activation of mTOR pathway could downregulate the expression of PDGFR-β mRNA at the transcriptional level [46, 47]. Interestingly, PRR5, a novel component of mTOR complex 2 (mTORC2), can regulate the expression PDGFR-β mRNA and its downstream signaling, and mTORC2 has been shown to be linked with actin cytoskeleton reorganization . In our study, overexpression of PDGF-D induced hyperactivation of mTOR pathway, with concomitant changes in cellular morphology and actin cytoskeleton reorganization, which is most likely linked with the downregulation of PDGFR-β mRNA. More importantly, PDGF-D-overexpressing PC3 cells exhibited a rapid growth rate, even under serum deprivation conditions similar to those observed in p21ras-transformed cells by other investigators [43, 49]. In our study, we found that overexpression of PDGF-D induced activation of mTOR, as characterized by increased phosphorylation of 4E-BP and S6K , which, in turn, could repress PDGFR-β mRNA expression via a negative feedback regulation. On the other hand, activation of S6K increased mTOR activity via phosphorylation of mTOR at site Ser2,448 [50, 51], which is consistent with our results showing that knockdown of raptor reduced phosphorylation of mTOR at site Ser2,448 (Fig. 4C). This positive feedback loop might be one the mechanism by which PDGF-D-overexpressing PC3 cells sustain EMT events under the low levels of receptor transcription. Additional studies are ongoing to explore how mTOR could cross-talk with other signaling pathway in controlling EMT events in PDGF-D-overexpressing PC3 cells. In the present study, despite the reduction in the levels of PDGFR-β mRNA and protein in PDGF-D-overexpressing PC3 cells, we found a higher proliferation rate and PDGF-D-mediated activation of downstream signaling mTOR pathway in PDGF-D-overexpressing PC3 cells compared with control cells, suggesting that sustained exposure of PC3 cells to PDGF-D leads to PDGF-D-mediated activation of its downstream signaling, which contributes to the EMT phenotype.
Interestingly, recent studies have also shown that many growth factors, such as vascular endothelial growth factor, FGF, EGF, and PDGF, have intracrine mechanism of signaling [52, 53], which may be one of the mechanisms by which PDGF-D overexpression contributes to EMT, as seen in our study. Our results showed, for the first time, that PDGF-D could induce EMT in PC3 cells, and additional mechanistic studies showed that EMT was associated with activation of mTOR. The activation of mTOR pathway, as documented by enhanced phosphorylation of mTOR and 4E-BP1 in our study, was concomitantly associated with the loss of E-cadherin and the gain of vimentin and nestin, which is associated with increased cell adhesion and invasion, consistent with the acquisition of EMT characteristics. Interestingly, knockdown of mTOR or raptor markedly repressed the invasion of PC3 PDGF-D cells, suggesting that PDGF-D-induced activation of mTOR is mechanistically associated with EMT and that the inactivation of mTOR could lead to the reversal of EMT. In this study, however, knockdown of mTOR with siRNA produced an increase in the expression of E-cadherin protein, but not a decrease in vimentin protein expression, whereas knockdown of raptor with siRNA produced a decrease in vimentin protein expression but not a significant increase in E-cadherin protein expression, although it was marginally increased. These results could be partly explained by the following observations. Sarbassov et al. found that Drosophila S6K and raptor (dS6K and draptor) double-stranded RNA (dsRNA) significantly increased p-dAkt, resulting from abrogating the negative feedback inhibition of p-dAkt by dS6K . However, despite reducing dS6K phosphorylation to the same extent as did draptor dsRNA, the dTOR dsRNA failed to increase dAkt phosphorylation, and instead it was decreased by a small amount. Similar results were obtained in human cell lines. Increasing evidence has also shown that activation of Akt is associated with EMT. Irie et al. found that knockdown of Akt1 in insulin-like growth factor-I receptor (IGF-IR)-stimulated cells promoted EMT and enhanced cell migration, suggesting that Akt1 inhibited EMT in IGF-IR-stimulated cells . Therefore, knockdown of raptor resulted in an increased Akt phosphorylation, which, in turn, inhibited EMT by repressing vimentin expression under conditions of hyperactivation of mTOR similar to the conditions in IGF-IR-stimulated cells, whereas knockdown of mTOR failed to increase p-Akt but rather decreased it, and thus vimentin expression was not inhibited. The published data, along with our results, clearly suggest a complex regulation of these genes in the processes of EMT. The mTOR or raptor knockdown experiment by specific siRNA was associated with a significant increase in the expression of E-cadherin and the reduced expression of vimentin in PDGF-D-overexpressing cells, suggesting the phenotypic reversal of EMT, although stable cell lines with mTOR or raptor knockdown in PC3 PDGF-D cells may shed more in-depth mechanistic insight, and thus additional studies are warranted to characterize the possibility of EMT to MET transformation of PC3 PDGF-D cells in the future.
The results showing that activation of mTOR was associated with EMT raise a question of how activation of mTOR could regulate the expression and location of epithelial and mesenchymal molecular markers and whether activation of mTOR cross-talks with other factors to regulate EMT. To answer this question, we determined the expression of NF-κB and the DNA-binding activity of NF-κB in PC3 Neo and PC3 PDGF-D cells. PDGF-D significantly increased the expression of NF-κB and its DNA-binding activity. Several recent studies have shown that the activation of NF-κB is mechanistically linked with the processes of EMT via regulating the expression of transcription factors, such as snail, ZEB1, and ZEB2, resulting in reduced expression of E-cadherin [23, 24]. Bachelder et al. demonstrated that glycogen synthase kinase-3 (GSK-3) activity and function are essential for maintaining epithelial structure via maintaining the expression of E-cadherin and have also identified GSK-3 as an inhibitor of NF-κB activation . Moreover, studies have shown that S6K inactivates GSK-3 under conditions of mTOR-dependent feedback inhibition of Akt, resulting from hyperactivation of mTOR. Overall, these reports, together with our current data, clearly suggest that PDGF-D-induced EMT in PC3 cells occurs through the activation of mTOR, resulting in an mTOR-dependent feedback inhibition of Akt (data not show). Under these conditions, activation of S6K, a direct downstream target of mTOR, could inhibit GSK-3 activity, which in turn could be associated with the activation of NF-κB, and this notion is partly supported by previous findings showing that NF-κB plays a central role in the processes of EMT .
Even though the activation of NF-κB is associated with EMT, it is not known what downstream gene of NF-κB is mechanistically associated with EMT. We therefore tested the role of NF-κB downstream genes in the processes of EMT. We found a dramatic increase in the levels of Bcl-2 expression in PDGF-D-overexpressing PC3 cells compared with PC3 Neo cells. We also found that overexpression of Bcl-2 in parental PC3 cells inhibited the expression of E-cadherin, whereas the knockdown of Bcl-2 expression led to the downregulation of vimentin expression in PC3 PDGF-D cells, suggesting that the regulation of Bcl-2 expression may play an important role in PDGF-D-induced EMT. However, knockdown of Bcl-2 expression did not increase E-cadherin expression in PC3 PDGF-D cells with EMT, suggesting that Bcl-2 has a different function in regulating the expression of E-cadherin and vimentin in epithelial cells and mesenchymal cells. Bcl-2 has been shown to be associated with invasion and metastasis, which might be linked with the regulation of vimentin expression by Bcl-2 but may not be directly related to E-cadherin expression in PC3 PDGF-D cells with EMT. Recent studies have shown that poly(ADP-ribose) polymerase-1 (PARP-1) could upregulate vimentin expression [56, 57], and it is known that Bcl-2 inhibits PARP cleavage and increases the PARP-1 level. Therefore, knockdown of Bcl-2 could decrease vimentin expression by increasing PARP cleavage. In our study, we found a dramatic increase in the levels of Bcl-2 and PARP-1 expression in PDGF-D-overexpressing PC3 cells, which was consistent with lower levels of apoptotic cell death, as seen in PDGF-D-overexpressing PC3 cells compared with PC3 Neo cells (unpublished data). Knockdown of Bcl-2 expression led to the downregulation of vimentin expression in PC3 PDGF-D cells, which is believed to be partly due to regulating PARP by Bcl-2, and, as such, this pathway may not be directly linked with the regulation of E-cadherin during the acquisition of EMT phenotype. Nevertheless, our results are provocative and clearly suggest that additional in-depth investigations are needed in elucidating the complex regulation of Bcl-2 and its association with the regulation of E-cadherin and vimentin during EMT.
In this study, our results provided strong evidence, for the first time, showing that overexpression of PDGF-D could induce EMT, and, as such, it was associated with increased invasion in vitro and enhanced tumor growth rate in vivo. The PDGF-D-induced EMT was mechanistically associated with activation of mTOR and NF-κB signaling, as well as overexpression of Bcl-2. On the basis of our results, we believe that the inactivation of PDGF-D signaling in prostate cancer by novel approaches could be an important means by which prostate cancer progression could be prevented and/or treated in the near future.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.
This work was partly funded by Grant 5R01-CA108535-04 from the National Cancer Institute, NIH (to F.H.S.).