Loss of Receptor Protein Tyrosine Phosphatase β/ζ (RPTPβ/ζ) Promotes Prostate Cancer Metastasis*

Background: The role of pleiotrophin and its receptors RPTPβ/ζ and Syndecan-3 during tumor metastasis remains unknown. Results: RPTPβ/ζ knockdown initiates EMT, promotes pleiotrophin-mediated migration and attachment through Syndecan-3 and induces in vivo metastasis. Conclusion: RPTPβ/ζ plays a suppressor-like role in prostate cancer metastasis. Significance: Boosting RPTPβ/ζ or attenuating Syndecan-3 signaling pathways may lead to more effective therapeutic strategies in treating prostate cancer metastasis. Pleiotrophin is a growth factor that induces carcinogenesis. Despite the fact that many published reports focused on the role of pleiotrophin and its receptors, receptor protein tyrosine phosphatase (RPTPβ/ζ), and syndecan-3 during tumor development, no information is available regarding their function in tumor metastasis. To investigate the mechanism through which pleiotrophin regulates tumor metastasis, we used two different prostate carcinoma cell lines, DU145 and PC3, in which the expression of RPTPβ/ζ or syndecan-3 was down-regulated by the RNAi technology. The loss of RPTPβ/ζ expression initiated epithelial-to-mesenchymal transition (EMT) and increased the ability of the cells to migrate and invade. Importantly, the loss of RPTPβ/ζ expression increased metastasis in nude mice in an experimental metastasis assay. We also demonstrate that RPTPβ/ζ counterbalanced the pleiotrophin-mediated syndecan-3 pathway. While the inhibition of syndecan-3 expression inhibited the pleiotrophin-mediated cell migration and attachment through the Src and Fak pathway, the inhibition of RPTPβ/ζ expression increased pleiotrophin-mediated migration and attachment through an interaction with Src and the subsequent activation of a signal transduction pathway involving Fak, Pten, and Erk1/2. Taken together, these results suggest that the loss of RPTPβ/ζ may contribute to the metastasis of prostate cancer cells by inducing EMT and promoting pleiotrophin activity through the syndecan-3 pathway.

The pleiotropic functions of this growth factor are regulated primarily through the autocrine/paracrine effects of receptoractivated signaling. Receptor protein tyrosine phosphatase ␤/ (RPTP␤/, PTPRZ1) (23,24) and syndecan-3 (N-syndecan, SDC3) (25) have been characterized as pleiotrophin transmembrane receptors. Pleiotrophin is the first natural ligand discovered for any of the transmembrane tyrosine phosphatases and signals through the dimerization and inactivation of RPTP␤/ (24). Whereas most PTPs are expressed in the peripheral tissues and CNS, RPTP␤/ expression is primarily detected in the CNS (26). RPTP␤/ is highly expressed in migrating neurons in the developing brain (27). Pleiotrophin stimulates the haptotactic migration of glioma cells, and the overexpression of pleiotrophin/RPTP␤/ in human astrocytic tumor cells might create an autocrine loop that enhances cell migration (8,9). Similar to RPTP␤/, syndecan-3 is abundantly expressed in the nervous system, particularly in the migrating neurons of the olfactory placode, providing indirect evidence that syndecan-3 mediates neural migration (30).
Despite the fact that there are many published reports describing the role of pleiotrophin and its receptors during tumor development, no information is available regarding their function during tumor metastasis. Furthermore, because RPTP␤/ and syndecan-3 have not been closely studied together, it remains unclear whether their signaling cascades converge or act independently to promote the functional response of pleiotrophin. In the present study, using RNAi technology, we stably transformed prostate carcinoma cells to down-regulate RPTP␤/ or syndecan-3 expression and investigated the in vivo and in vitro molecular mechanism through which pleiotrophin regulates tumor metastasis.
Cell Culture-The human prostate cancer epithelial cell lines DU145 (ATCC) were grown in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS), 100 units/ml penicillin, and 100 g/ml streptomycin. Cultures were maintained in 5% CO 2 and 100% humidity at 37°C.
Purification of Human Recombinant Pleiotrophin-Human recombinant pleiotrophin was produced in prokaryotic expression systems, as previously described (31). Pleiotrophin was further purified to remove LPS, and its biological activity was verified by the in vitro neurite outgrowth test.
Attachment Assay-24-well culture plates were coated with 10 g/ml fibronectin, laminin, or collagen I for 1 h at 37°C. Wells were then incubated with a 0.5% solution of bovine serum albumin (BSA) for 1 h at 37°C to further block any specific adsorption of protein. 60,000 resuspended cells in RPMI 1640 medium supplemented with 2.5% FBS were then seeded. After a 10-min incubation period, unattached cells were removed by shaking the plates at 2,000 rpm for 10 s and by three washes with PBS. Attached cells were fixed with 4% paraformaldehyde and stained with crystal violet.
Crystal Violet Assay-Adherent cells were fixed with methanol and stained with 0.5% crystal violet in 20% methanol for 20 min. After gentle rinsing with water, the retained dye was extracted with 30% acetic acid and the absorbance measured at 590 nm.
Transwell Assay-Migration assays were carried out in Boyden chambers using filters (8-m pore size, Costar, Avon, France) coated with fibronectin, laminin, or collagen I (7.5 g/cm2) for 1 h at 37°C. Filters were washed, blocked with 0.5% BSA for 1 h at 37°C, and dried. Assay medium (RPMI 1640 medium supplemented with 2.5% FBS, and 0.5% BSA with or without the chemo attractant) was added to the lower compartment, and 10 4 cells were added into the insert. After incubation for 30 min at 37°C, filters were fixed. Non-migrated cells were scraped off the upper side of the filters and the filters were stained with crystal violet. The number of migrated cells was quantified by counting the entire surface area of the filter.
Tail-vein Injections of Human Prostate Cancer Cells-All in vivo experiments were carried out with ethical committee approval and under the condition established by the European Community. 1 ϫ 10 6 DU145, DU145-NC2, or DU145-RM6 cells were suspended in 0.2 ml of PBS and injected into the lateral tail vein of athymic nude mice (Janvier), using a 27-gauge needle. The mice were sacrificed 10 weeks after injection, and the lungs were perfused with 1.5 ml of 15% India ink dye in 3.7% formalin. Lungs were then removed, rinsed in water for 15 s, and bleached in Fekete' s solution. Lung surfaces were photographed and scored.
shRNA Transfection-The pSilencer 4.1-CMV expression vector and the siPORT XP-1 Transfection Agent were obtained from Ambion Inc. Based on the siRNA sequence, shRNA was designed, ligated into the pSilencer 4.1-CMV expression vector, and transfected into cells according to Ambion's instructions. Briefly, siPORT XP-1 and shRNA were mixed at a final ratio of 1:6 in OPTI-MEM media. The transfection complexes were then overlaid onto 24-well plate cultures grown in RPMI 1640 supplemented with 10% FBS. After 1 month of selection with 300 g/ml G418, clones were screened for down-regulation of RPTP␤/ expression. Double-stranded negative control shRNA from Ambion was also used.
Immunoprecipitation-Media from cell cultures grown in 60-mm plastic dishes were aspirated, cells were washed twice with ice-cold PBS, and lysed in 1 ml of buffer containing 50 mM HEPES pH 7.0, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 5 g/ml aprotinin, and 5 g/ml leupeptin. Cells were harvested, sonicated for 4 min on ice, and centrifuged at 20,000 ϫ g for 10 min at 4°C. Approximately 600 g of the supernatant was then incubated with 30 l of protein A-Sepharose bead suspension for 60 min at room temperature. Beads were collected by centrifugation and the supernatants incubated overnight at 4°C with anti-RPTP␤/ (1:200) or anti-SRC-kinase (1:1,000) primary antibodies. The mixtures were then incubated with 80 l of protein A-Sepharose beads for 3 h at 4°C. Beads and bound proteins were collected by centrifugation (10,000 ϫ g, 4°C), washed three times with ice-cold lysis buffer, and resuspended in 60 l of 2ϫ SDS loading buffer (100 mM Tris-HCl pH 6.8, 4% SDS, 0.2% bromphenol blue, 20% glycerol, 0.1 M dithiothreitol). Samples were then heated to 95-100°C for 5 min and centrifuged. 50 l of the supernatant were analyzed by Western blot.
Western Blot Analysis-Cells were starved for 4 h and then incubated with pleiotrophin for varying times. Cells were subsequently washed twice with PBS and lysed in 250 l of 2ϫ SDS loading buffer under reducing conditions. Proteins were separated by SDS-PAGE and transferred to an Immobilon-P membrane for 3 h in 48 mM Tris pH 8.3, 39 mM glycine, 0.037% SDS, and 20% methanol. The membrane was blocked in TBS containing 5% nonfat milk and 0.1% Tween 20 for 1 h at 37°C. Membranes were then probed with primary antibody overnight at 4°C under continuous agitation. The anti-RPTP␤/ antibody was used at 1:500 dilution. All other antibodies were used at 1:1,000 dilution. The blot was then incubated with the appropriate secondary antibody coupled to horseradish peroxidase, and bands were detected with the ChemiLucent Detection System Kit (Chemicon International Inc., CA) according to the manufacturer's instructions. Where indicated, blots were stripped in buffer containing 62.5 mM Tris HCl pH 6.8, 2% SDS, 100 mM 2-mercaptoethanol for 30 min at 50°C and reprobed. Quantitative estimation of band size and intensity was carried out by digital image analysis, using ImagePC image analysis software (Scion Corporation, Frederick, MD).
FACS Analysis-Cells were collected by centrifugation, fixed with 4% formaldehyde for 10 min at 4°C, chilled on ice for 1 min, and permeabilized with 90% methanol for 30 min at 4°C. Cells were subsequently blocked with 0.5% BSA and stained with monoclonal antibodies against integrin-␣5, -␤1, and -␤3 for 1 h on ice in a total volume of 100 ml binding buffer. Cells were then washed twice and incubated with FITC-conjugated secondary anti-mouse antibody for 30 min at 4°C. Cells were then washed twice and resuspended in PBS. Approximately 10,000 cells from each sample were analyzed by flow cytometry, using a Coulter EPICS-XL-MCL cytometer (Coulter, Miami, FL), and analysis was performed using the XL-2 software (Coulter).
Cell Staining and Confocal Imaging-Cells grown in 8-well tissue culture slides (Nunc) were fixed in 4% paraformaldehyde for 10 min at room temperature, rinsed three times with PBS, quenched with 50 mM Tris buffer pH 8.0 and 100 mM NaCl, permeabilized for 15 min in PBS containing 0.3% Triton X-100 and 0.5% bovine serum albumin (BSA), and blocked in PBS containing 3% BSA for 1 h at room temperature. Cells were incubated for 1 h at room temperature with Phalloidin-FITC in permeabilization buffer. The staining for E-cadherin, snail and ␤-catenin was performed for overnight at 4°C. After three rinses in PBS, cells were mounted using Sigma mounting fluid. Labeling was observed using a Nikon confocal microscope and photographed.
Statistical Analysis-Comparison of mean values among groups was done using ANOVA and the unpaired Student's t test. Homogeneity of variance was tested by Levene's test. Each experiment included at least triplicate measurements for each condition tested. All results are expressed as the mean Ϯ S.D. of at least three independent experiments. Values of p less than 0.05 were considered to be significant (*, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001).

RPTP␤/ Knockdown Induces in Vitro Migration and Invasion, as Well as in Vivo Metastasis of Prostate Cancer Cells-Ac-
cording to data from large scale, genome-wide expression analyses of prostate cancer, the expression of RPTP␤/ is reduced up to 67% in metastatic tumors (32). To understand how RPTP␤/ regulates prostate cancer metastasis, we used two androgen-independent prostate carcinoma cell lines, DU145 and PC3 which were transfected with vector coding shRNA targeting RPTP␤/ and isolated the pool of the cells and clones that stably expressed reduced levels of RPTP␤/. The pool of each cell line (DU145-RM and PC3-RM) and the clones DU145-RM2, DU145-RM6, PC3-RM1, and PC3-RM6 were used for the following experiments (Fig. 1A). The cells were then tested for their ability to migrate, invade and metastasize. As shown in Fig. 1B, the cell lines silenced for RPTP␤/ migrated more robustly (100% higher) than the parental lines or the cells transfected with control vector, DU145-NC, DU145-NC2, PC3-NC, and PC3-NC4. Furthermore, RPTP␤/ attenuation also increased the invasive capacity of the cells (130% increase) (Fig. 1C). To determine the effect of RPTP␤/ on the metastatic dissemination of human prostate cancer cells, DU145, DU145-NC2, or DU145-RM6 cells were injected into nude mice through the tail veins. The lungs were harvested after 10 weeks and scored for surface lung metastasis. As shown in Fig. 1D, the mice that were injected with DU145-RM6 cells developed lung metastasis, but not those injected with DU145 or DU145-NC2 cells. This is the first demonstration that RPTP␤/ knockdown increases the intrinsic ability of prostate cancer cells to metastasize in vivo. Taken together, these results suggest that RPTP␤/ may act as a tumor suppressor for prostate cancer metastasis.
RPTP␤/ Knockdown Induces an EMT-like Phenotype-As shown in Fig. 2A, stable down-regulation of RPTP␤/ expres-sion increased the fibroblastic morphology of the cells as evidence by their elongated shape and the formation of multiple membrane protrusions. Phalloidin staining of RPTP␤/-knockdown cells revealed increased stress fiber formation and a concomitant decrease in the number of cortical actin fibers, marker of lamellipodia formation. This phenotype, in combination with the observed increased cell motility and invasion (Fig. 1), prompted us to examine whether RPTP␤/ knockdown is associated with EMT.
The expression profile of adhesion molecules, such as integrin-␣5, -␣v, -␤1, and -␤3 was investigated by Western blot and FACS analysis. The down-regulation of RPTP␤/ induced the expression of integrin-␣5, -␣v, and -␤3 (Fig. 2, B and C). Furthermore, there was a shift in cadherin expression from the Eto the N-cadherin form, known to characterize the EMT-like phenotype and to promote cell motility (33). As shown in the immunoblots and immunofluorescence staining in Fig. 2, B and D, E-cadherin expression was clearly decreased in DU145-RM6 or PC3-RM1 cells compared with wild-type cells, while that of N-cadherin was increased. Fig. 2, B and D also shows that DU145-RM6 and PC3-RM1 cells expressed higher levels of Twist and Slug/Snail transcription factors, known to down-regulate E-cadherin (34) compared with wild-type cells. The down-regulation of RPTP␤/ also induced the translocation of ␤-catenin form the cell surface to the nucleus (Fig. 2D). Hence, down-regulation of RPTP␤/ induces an EMT-like phenotype.
The Effect of RPTP␤/ or Syndecan-3 Down-regulation on Pleiotrophin-mediated Migration and Attachment-We next examined the response of the prostate cancer cell lines silenced for RPTP␤/ to its ligand pleiotrophin. For that, we tested the effect of exogenously added pleiotrophin on cell migration using both Boyden chamber and wound closer assays. In the Boyden chamber assay, an equal number of cells from each cell line were plated in inserts coated with fibronectin and allowed to migrate toward increasing concentrations of recombinant pleiotrophin ranging from 5 to 100 ng/ml. The cells that migrated across the 8-m porous membrane were fixed, stained, and counted. As expected, pleiotrophin increased migration (40% increase) in the wild-type DU145 and PC3 cells. However, RPTP␤/ gene silencing caused a greater increase (2-fold) in these cells and when the RPTP␤/ silenced cells, DU145-RM, DU145-RM6, PC3-RM, and PC3-RM1, were treated with pleiotrophin, a further increase in cell migration was observed, which was pleiotrophin concentration dependent (data not shown), with a maximal effect (267% increase) obtained at a concentration of 50 ng/ml (Fig. 3A). These results suggest that the binding of pleiotrophin to RPTP␤/ inhibits pleiotrophin-mediated migration and that the observed induction in the RPTP␤/-silenced cells may therefore be mediated through syndecan-3, the pleiotrophin other receptor. To test this hypothesis, syndecan-3 expression was transiently downregulated in both cell lines, using two different siRNA sequences (Fig. 3B). As shown in Fig. 3A, the syndecan-3 silenced cells, DU145-SM1, DU145-SM2, PC3-SM1, and PC3-SM2, migrated less effectively than the parental lines. Moreover, treatment with pleiotrophin further decreased DU145-SM1, DU145-SM2, PC3-SM1, and PC3-SM2 migration in a concentration-dependent manner (data not shown), with a maximal effect (70% decrease) at 50 ng/ml. All migration experiments were also performed on laminin or collagen I as coating substrates and gave similar results (data not shown). The effect of pleiotrophin on the migration of DU145-NC, DU145-NC2, DU145-siNC, PC3-NC, PC3-NC4, and PC3-siNC cells (cells transfected with control vector) was similar to the effect observed on DU145 or PC3 cells. The wound-closure assays confirmed the inhibitory effect of RPTP␤/ on pleiotrophinmediated motility (Fig. 3C). These results demonstrate that the pleiotrophin/RPTP␤/ interaction inhibit cell migration whereas the pleiotrophin/syndecan-3 interaction induces cell migration.
RPTP␤/ or syndecan-3 knockdown cells. An equal number of cells were pre-incubated with increasing concentrations of pleiotrophin before seeding into culture wells coated with fibronectin and allowed to attach for 10 min. As shown in Fig.  3D, DU145-RM, DU145-RM6, PC3-RM, and PC3-RM1 cells were attached more efficiently than the parental control cells and pleiotrophin further increased DU145-RM, DU145-RM6, PC3-RM, and PC3-RM1 cell attachment, with a maximal effect (240% increase) obtained at a concentration of 50 ng/ml. However, the syndecan-3 knockdown DU145-SM1, DU145-SM2, PC3-SM1, and PC3-SM2 cells attached less effectively than the wild-type cells and their number was further decreased after pleiotrophin treatment in a concentration-dependent manner (50% inhibition at 50 ng/ml). These results indicate that pleiotrophin/RPTP␤/ interaction inhibits cell attachment, while in contrast, the pleiotrophin/syndecan-3 interaction induces cell attachment.
Opposing Effect of Pleiotrophin/RPTP␤/ and Pleiotrophin/ Syndecan-3 Pathways on Src and Fak Activation-Pleiotrophin signals through the dimerization and inactivation of RPTP␤/, and the loss of phosphatase activity results in increased tyrosine phosphorylation levels of different substrates, including ␤-catenin (24), ␤-adducin (36,37), Fyn (38), and GIT1/Cat1 (39). However, our results show that the pleiotrophin-mediated RPTP␤/ inactivation is associated with an inhibition of cell migration and attachment. To look into the mechanism involved in the inhibition of migration we searched for signal transduction molecules whose activation depends on dephosphorylation events. Src appeared as a potential candidate as its activation is strictly regulated by and depends on the dephosphorylation of Tyr-527 in its C-terminal tail, which is a prerequisite for the autophosphorylation of Tyr-416 in the activation loop and the subsequent activation of the kinase (40). After demonstrating by immunoprecipitation experiments that Src interacted with RPTP␤/ (Fig. 4A), we examined the effect of pleiotrophin on Src phosphorylation in our knockdown DU145 cells. DU145, DU145-NC2, DU145-RM6, and DU145-SM1 cells were serum-starved for 4 h and incubated with increasing concentrations of pleiotrophin for 3 to 45 min. Src activation was indirectly assessed using Western blot analysis of phosphorylated Src at site Tyr-416. Levels of HSC70 were used as loading control. Within 3 min, pleiotrophin promoted a rapid increase in Src phosphorylation in a concentration-dependent manner, with a maximal effect (150% induction relative to control) at 50 ng/ml (Fig. 4B). This was followed by a return to near basal levels at 10 to 30 min; a second, weaker increase in Src phosphorylation occurred 45 min after pleiotrophin treatment (data not shown). However, the effect of pleiotrophin on the Srcmediated phosphorylation in DU145-RM6 cells (RPTP␤/ knockdown) was markedly stronger, with a maximal induction of 350% occurring in the presence of 50 ng/ml pleiotrophin. These results indicate that the binding of pleiotrophin to RPTP␤/ inhibits Src activation in wild-type cells and that the observed induction in the RPTP␤/ knockdown cells may therefore be mediated through syndecan-3. As shown in Fig. 4, A and B, syndecan-3 interacted with Src and pleiotrophin inhibited Src phosphorylation in the syndecan-3 knockdown DU145-SM1 cells in a concentration-dependent manner, with a maximal effect (60% decrease) at 50 ng/ml. Furthermore, we observed an increased interaction between Src and syndecan-3 in the RPTP␤/ silenced DU145-RM6 cells, an increased interaction between Src and RPTP␤/ cells in the syndecan-3 silenced DU145-SM1 cells, but no interaction between the RPTP␤/ and the syndecan-3, indicating that there are two distinct signal transduction pathways (Fig. 4A). Taken together, our results suggest that the binding of pleiotrophin to RPTP␤/ inactivates Src and counterbalances the pleiotrophin/syndecan-3-mediated induction of Src phosphorylation. To study the functional impact of Src activation on pleiotrophin-mediated biological actions, we used the pharmacological inhibitor PP1 to specifically block the Src pathway. As shown in Fig. 4C, PP1 abrogated the effect of pleiotrophin on cell migration and attachment. Hence, the results suggest that Src is a key regulator of pleiotrophin-mediated migration and attachment.
We next examined the effect of pleiotrophin on other molecules known to interact with Src. We found that Fak phosphorylation was increased in a concentration-dependent manner 3 min after the incubation of DU145 cells with pleiotrophin with maximal effect (80% increase) at 50 ng/ml pleiotrophin (Fig.  4C). This activation was followed by a return to basal levels at 10 to 30 min after pleiotrophin treatment (data not shown). However, Fak phosphorylation by pleiotrophin was much greater in DU145-RM6 cells (150% increase) while it was inhibited in DU145-SM1 cells (70% decrease) (Fig. 4C). Hence, similarly to Src, the pleiotrophin/RPTP␤/ interaction inactivates Fak and counterbalances the pleiotrophin/syndecan-3-mediated Fak activation.
RPTP␤/, but Not Syndecan-3, Induces Pten and Erk1/2 Dephosphorylation after Pleiotrophin Treatment-We further tested whether pleiotrophin affects Pten or Erk1/2 activation. We found that Pten phosphorylation was decreased 3 min after the incubation of DU145 cells with increasing concentrations of pleiotrophin and had a maximal effect (60% inhibition relative to control) at 50 ng/ml. This activation was followed by a return to basal levels between 10 and 30 min after pleiotrophin treatment (data not shown). Contrary to the effect on DU145 cells, pleiotrophin had no effect on Pten phosphorylation in DU145-RM6 cells (Fig. 5A). Furthermore, the pleiotrophin effect on Pten phosphorylation in DU145-SM1 cells was similar to the observed effect on wild-type cells (65% inhibition relative to control at a concentration of 50 ng/ml) (Fig. 5A). Similar to Pten activation, the extracellular signal-regulated kinases 1 and  2 (Erk1/2) were inactivated in DU145 cells and the inactivation was sustained for up to 45 min after a 15-min incubation with pleiotrophin (data not shown). Pleiotrophin also had no significant effect on the Erk1/2-mediated phosphorylation in DU145-RM6 cells and inhibited their activation in DU145-SM1 cells in a manner similar to wild-type cells (Fig. 5B). These results suggest that Pten and Erk1/2 are implicated only in the signal transduction pathway triggered through the binding of pleiotrophin to RPTP␤/, which clearly indicate that RPTP␤/ inhibits the phosphorylation levels of these signaling molecules.
Activation of Src/Fak Pathway during EMT of RPTP␤/-silenced Cells-The activation status of Src was examined on the RPTP␤/ silenced cells, DU145-RM6 and PC3-RM1. As shown in Fig. 6A, the phosphorylation levels of Src at site Tyr-416 were increased on RPTP␤/ silenced cells compared with the parental cell lines or the cells that were transfected with control vector (DU145-NC2 and PC3-NC4). Fig. 6B also shows that the phosphorylation levels of Fak were also increased on RPTP␤/silenced cells. Hence, the results suggest that the RPTP␤/ attenuation, which induces an EMT phenotype, also activates the Src/Fak pathway.

DISCUSSION
Despite the threat that prostate cancer poses to the health of men worldwide, the molecular mechanisms involved are still poorly understood. In the initial stages, prostate cancer is dependent on androgens for growth and can be suppressed using androgen deprivation therapy. However, tumor growth eventually recur due to a transition from an androgen-dependent to an androgen-independent state, leading to a highly metastatic disease that eventually kills the patient (41). In addition to androgens, growth factors, including pleiotrophin, have been shown to play a key role in the development and progression of the disease (11,(42)(43). The present study demonstrates a role for pleiotrophin receptors, RPTP␤/ and syndecan-3, in prostate cancer metastasis and shed light on the molecular mechanism involved.
Our results show that RPTP␤/ was a negative regulator of EMT, where carcinoma cells lose polarity, cell-cell contacts and other epithelial characteristics, switching to a motile mesenchymal phenotype (45,46). Common EMT features include a loss of E-cadherin, elevated N-cadherin, integrins, Twist, Slug, and Snail expression, increased nuclear ␤-catenin, and gain of fibroblastoid morphology. Reduced E-cadherin expression has been observed in high-grade prostate cancers and is associated with poor prognosis (47,48). Prostatic epithelial cells undergo EMT in response to abnormal receptor activation, such as platelet-derived growth factor receptor (PDGFR), fibroblast growth factor receptor (FGFR), transforming growth factor-␤ receptor (TGF-␤R), and insulin-like growth factor-1 receptor (IGF-1R) (46). Previous study has shown that in glioma cells, pleiotrophin also initiates EMT (19). In the present work however, we show that the down-regulation of the expression of its receptor RPTP␤/ induced EMT and increased the ability of human prostate cancer cells to migrate and invade in vitro and to metastasize in vivo. These results are consistent with recent data available in the MSKCC database that indicate that RPTP␤/ expression is reduced up to 50% in primary prostate cancer, whereas in the case of metastatic androgen-dependent or androgen-independent prostate tumors, this percentage reaches 67 and 35%, respectively. However, it seems that the effect of RPTP␤/ on cell motility is not the same in all types of cancers. Previous studies in neuroblastoma cells have shown that down-regulation of RPTP␤/ reduces haptotactic activity (50). In gliomastoma the expression of RPTP␤/ is increased compared with normal brain tissues (8) and in human astrocytic tumor cells, the up-regulated expression of RPTP␤/ and pleiotrophin creates an autocrine loop that is important for cell migration (9). However, pleiotrophin and RPTP␤/ expression were found decreased in human colorectal cancers (51).
Our results also show that RPTP␤/ decreased pleiotrophinmediated migration and attachment through an interaction with Src, triggering a signal transduction pathway that inactivates Fak, and Erk1/2 and activates Pten, molecules frequently deregulated in prostate cancer. Specifically, Pten, a tumor suppressing molecule, is the most frequently mutated gene involved in prostate cancer metastasis. Acquired mutations in the Pten gene are associated with advanced stage and poor prognosis (52). Pten, which is activated after dephosphorylation, dephosphorylates Fak, and reduces cell migration and focal adhesion formation (53). Furthermore, Pten negatively regulates the PI3K and MAPK pathway and its loss leads to constitutive activation of the Erk1/2, which regulates motility in prostate cancer cells (54). However, unlike RPTP␤/, the interaction of pleiotrophin with its second receptor, syndecan-3, induces prostate cancer cell migration, suggesting that the effect of pleiotrophin on cell migration is the overall result of the opposing actions of its two receptors, RPTP␤/ and syndecan-3. Although pleiotrophin is known as a growth factor with anti-apoptotic and angiogenic activities, other studies, like ours, already suggested that pleiotrophin can also function as a negative regulator of biological responses. For example, it was shown to promote the apoptotic response of cardiomyocytes through the inhibition of Akt signaling (55) and to inhibit angiogenesis by interfering with VEGF165 (56). Moreover, it has been shown that pleiotrophin negatively regulates adipogenesis (35). In addition, several publications have reported contradictory results concerning the activities of pleiotrophin. While only the recombinant polypeptide produced in a mammalian system stimulated the proliferation of fibroblast, endothelial, and epithelial cells (1,29,44), pleiotrophin produced in prokaryotic expression systems had a mitogenic effect on endothelial cells only when it was immobilized on the culture plate (31). Furthermore, eukaryotic-produced pleiotrophin in solution had no significant effect on the proliferation and chemotactic migration of glioma cell lines (8,9,28), but when presented as an immobilized substrate, it strongly stimulated them (9). The explanations provided for these contradictory results focused mainly on pleiotrophin folding, modifications and on the type of expression system used. A more recent study using glioma cell lines proposed the expression of two different forms of pleiotrophin, each displaying distinct biological actions; a full-length form which induced haptotactic migration, whereas the shorter, 15 kDa form, activated a mitogenic signal transduction pathway (49). In our study however, only purified fractions of full-length pleiotrophin, produced in prokaryotic expression systems were used. Our present study showing opposing effects of the two pleiotrophin receptors provides therefore a new reasoning for pleiotrophin conflicting effects, as they would possibly depend on their relative expression levels.
In summary, we suggest that RPTP␤/ plays a tumor suppressor-like role in prostate cancer metastasis. Fig. 7 schematically describes our views on how the expression of RPTP␤/ may influence tumor progression. During early stages of prostate tumor development, both RPTP␤/ and syndecan-3 would be expressed and RPTP␤/ would counterbalance the pleiotrophin-mediated syndecan-3 pathway, reducing migration. However, inactivation of RPTP␤/ or loss of its expression which may occur during tumor progression would initiate EMT. In these cells, pleiotrophin would be left to interact with synde-can-3, increasing cell migration, invasion and metastasis. This work provides the first evidence of the molecular mechanism through which pleiotrophin and its receptors RPTP␤/ and syndecan-3 regulates prostate cancer metastasis. These results may contribute to more effective therapeutic strategies for the treatment of prostate cancer and warrant further study.