p120ctn and P-Cadherin but Not E-Cadherin Regulate Cell Motility and Invasion of DU145 Prostate Cancer Cells

Background Adherens junctions consist of transmembrane cadherins, which interact intracellularly with p120ctn, ß-catenin and α-catenin. p120ctn is known to regulate cell-cell adhesion by increasing cadherin stability, but the effects of other adherens junction components on cell-cell adhesion have not been compared with that of p120ctn. Methodology/Principal Findings We show that depletion of p120ctn by small interfering RNA (siRNA) in DU145 prostate cancer and MCF10A breast epithelial cells reduces the expression levels of the adherens junction proteins, E-cadherin, P-cadherin, ß-catenin and α-catenin, and induces loss of cell-cell adhesion. p120ctn-depleted cells also have increased migration speed and invasion, which correlates with increased Rap1 but not Rac1 or RhoA activity. Downregulation of P-cadherin, β-catenin and α-catenin but not E-cadherin induces a loss of cell-cell adhesion, increased migration and enhanced invasion similar to p120ctn depletion. However, only p120ctn depletion leads to a decrease in the levels of other adherens junction proteins. Conclusions/Significance Our data indicate that P-cadherin but not E-cadherin is important for maintaining adherens junctions in DU145 and MCF10A cells, and that depletion of any of the cadherin-associated proteins, p120ctn, ß-catenin or α-catenin, is sufficient to disrupt adherens junctions in DU145 cells and increase migration and cancer cell invasion.


Introduction
During metastasis, cancer cells commonly gain the ability to migrate and invade tissues by regulating their interaction with other cells and their extracellular environment. Many studies show that alterations in cell-cell adhesion correlate with epithelial tumour progression and metastasis [1]. Cancer cells can invade either as single cells or collectively as groups of cells [2], [3], [4]. Both types of invasion involve disruption of the epithelium which usually requires a weakening of cell-cell contacts and a change in cell shape. Cadherins and catenins form adherens junctions, which are central mediators of cell-cell adhesion. Expression of adherens junction proteins is often decreased in tumours, and reconstitution of functional adherens junctions can revert the invasive phenotype of cancer cells [5], [6], [7].
The classical cadherins (E-, VE-, N-and P-cadherin) are the central transmembrane proteins of the adherens junction complex and mediate Ca 2+ -dependent homophilic intercellular interactions. b-catenin and plakoglobin/c-catenin bind to cadherin intracellular domains in a mutually exclusive manner and acatenin in turn binds to b-catenin [8]. p120-catenin (p120ctn) on the other hand binds to a different region of cadherins to b-catenin [9], [10], [11], [12], [13], [14]. The composition of proteins of the adherens junction complex depends on the cell type and cadherin expression. a-catenin provides an important link between adherens junctions and the actin cytoskeleton. Its interaction with actin filaments however, appears to be mutually exclusive to its binding to b-catenin suggesting an indirect link between junctions and the actin cytoskeleton [15], [16]. p120ctn has been found to regulate adherens junction stability in vitro [17], [18], [19] and in vivo [20], [21], [22]. Its depletion by RNAi leads to a decrease in adherens junction proteins and abolishes cell-cell adhesion [17], in part by increasing cadherin endocytosis and degradation [19], [23]. p120ctn also affects the activity of the small GTPases RhoA, Rac1 and Cdc42 in some cell types [24], [25], [26], [27], which play a central role in regulating cytoskeletal dynamics, the formation and maintenance of adherens junctions and cell migration [28], [29]. p120ctn can therefore link adherens junctions and Rho GTPases.
The effects of p120ctn on cell migration and invasion vary depending on the cell type, assay conditions and types of cadherins expressed [30], [31], [32]. It is not clear to what extent the effects of p120ctn are mediated through adherens junctions or Rho GTPases. To address this, we have knocked down p120ctn in DU145 prostate cancer cells and MCF10A mammary epithelial cells, both of which normally have adherens junctions containing E-cadherin and P-cadherin. In both cell types depletion of p120ctn leads to disruption of cell-cell contacts, downregulation of adherens junction proteins and increased cell motility. Interestingly, the knockdown of p120ctn did not affect activity levels of Rho GTPases Rac1 and RhoA but led to an increase in active Rap1. For the first time we have knocked down cadherins as well as catenins and show that P-cadherin, b-catenin and a-catenin but not E-cadherin depletion mimic the effects of p120ctn suppression and lead to a loss of cell-cell contacts and enhanced invasion.

Results
Suppression of p120ctn expression leads to downregulation of cell-cell adhesion proteins and disruption of cell-cell contacts To investigate the effects of p120ctn on cell-cell adhesion, we compared the effects of p120ctn depletion in two different cell lines that have adherens junctions, the human prostate cancer cell line DU145 and the human mammary epithelial cell line MCF10A. DU145 and MCF10A cells formed cell-cell contacts and expressed both E-and P-cadherin, which localized to cell-cell contacts (Fig. 1A, B). DU145 and MCF10A cells predominantly express two p120ctn isoforms that, based on their molecular weight, are predicted to be isoforms 1 and 3 [33] with isoform 3 (lower band) being expressed at slightly higher levels ( Fig. 1).
Knockdown of p120ctn with two different siRNAs in DU145 and MCF10A cells induced disruption of cell-cell adhesion resulting in a 'scattered' phenotype ( Fig. 1, Fig. 2). In addition, DU145 cells stably depleted of p120ctn did not have stable adherens junction (Fig. S1). This phenotype was specifically due to p120ctn depletion, since expression of shRNA-resistant mouse p120ctn (GFP-fusion) rescued junction formation, as determined by localization of p120ctn and E-cadherin (Fig. S1).
Both E-cadherin and P-cadherin as well as band a-catenin were downregulated in p120ctn-depleted cells. Downregulation of cadherins correlated with the extent of p120ctn depletion: at 72 h after transfection with siRNAs targeting p120ctn, p120ctn levels were lower than at 48 h, and E-cadherin levels were lower at 72 h than 48 h (data not shown).
p120ctn-depleted cells were however not likely to have undergone epithelial to mesenchymal transition (EMT) as levels of EMT marker proteins such as vimentin were not increased compared to control-transfected DU145 cells (data not shown).

p120ctn regulates migration of DU145 and MCF10A cells
To test whether cell migration was affected by p120ctn depletion, cells were monitored by time-lapse microscopy over 16 h ( Fig. 2A, B) and tracked to determine their migration speed. p120ctn-depleted DU145 and MCF10A cells both migrated faster than control cells (Fig. 2, Movies S1, S2, S3 and S4). By contrast, depletion of E-cadherin did not alter cell migration speed (Fig. 2C), even though its expression level was reduced by p120ctn knockdown (Fig. 1). The migration speed of control cells was determined from cells localised within cell colonies (Fig. 2C), but some cells in the population migrated as single cells (Movies S1 and S3). However, there was no difference in the migration speed of DU145 cells in colonies or single cells (Fig. 2D), ruling out the possibility that p120ctn-depleted cells migrated faster because they were predominantly single cells rather than in colonies.
Cell migration was also analysed in scratch-wound assays on confluent monolayers of DU145 cells. Most p120ctn-depleted cells migrated as individual cells into the wound, whereas control cells migrated as a sheet, maintaining cell-cell contacts (Fig. 3A, Movies S5, S6). p120ctn-depleted cells migrated faster than control cells (Fig. 3B). However, p120ctn-depleted cells showed a decrease in persistence (Fig. 3B), which indicates that they change direction more frequently. Thus, despite the increased migration speed, wound closure time was similar to control-transfected cells (Fig. 3C).

Suppression of p120ctn expression leads to increased Rap1 activity
Regulation of Rho GTPase activities by p120ctn has previously been correlated with increased cell motility [25,26]. However, no differences in the levels of active RhoA, Rac1 or Cdc42 were observed between p120ctn-depleted and control-transfected DU145 cells (Fig. 4A, B; data for Cdc42 not shown). Since p120ctn knockdown induced a loss of cell-cell contacts and the GTPase Rap1 has previously been associated with stabilisation of E-cadherin at the plasma membrane and to be activated following E-cadherin disengagement [34], levels of active Rap1 were analysed. p120ctn-depleted DU145 cells showed increased levels of Rap1 activity (Fig. 4A, B).
Depletion of cadherins, ß-catenin or a-catenin does not affect expression levels of other adherens junction proteins Since p120ctn depletion reduced the levels of the adherens junction proteins E-cadherin, P-cadherin, ß-catenin and a-catenin ( Fig. 1), we investigated the effects of depleting each of these proteins on adherens junctions. DU145 cells were transfected with 2 or 3 different siRNA oligos for E-cadherin, P-cadherin, bcatenin, a-catenin and p120ctn (Fig. 5A).
Expression levels of each of these proteins were then determined by quantification of immunoblots. Levels of P-cadherin, b-catenin, a-catenin and p120ctn proteins were not affected by E-cadherin depletion (Fig. 5A, B). Knockdown of P-cadherin slightly reduced levels of E-cadherin, ß-catenin and a-catenin but not p120ctn, but this was only observed for siRNA oligo (1) that gave the strongest (.80%) depletion of P-cadherin. Two other oligos that reduced Pcadherin levels by ,50% did not affect the levels of other junctional proteins. ß-catenin or a-catenin depletion did not alter the levels of other adherens junction proteins significantly (Fig. 5). It is thus only p120ctn depletion that results in a decrease in the levels of all the other cell-cell adhesion molecules (Fig. 5B).
Suppression of p120ctn, P-cadherin, b-catenin and acatenin but not E-cadherin leads to disruption of cell-cell contacts and enhances cancer cell invasion Since downregulation of p120ctn led to a loss of cell-cell junctions and an increase in cell migration speed, the effects of depleting other adherens junction proteins on cell-cell adhesion and cell motility were investigated. Knockdown of E-cadherin, Pcadherin, ß-catenin and a-catenin by RNAi was highly efficient both within colonies and in single cells, as observed by immunofluorescence, but did not discernibly affect p120ctn levels (Fig. 6).
Cells depleted of each adherens junction protein were tracked from timelapse movies to determine their migration speed (Fig. 7). Like p120ctn depletion, knockdown of P-cadherin, b-catenin and a-catenin led to an increase in cell migration speed, similar to results obtained with p120ctn depletion, whereas E-cadherin depletion did not alter migration speed (Fig. 7). Similarly Pcadherin but not E-cadherin depletion in MCF10A cells led to disruption of cell-cell contacts and increased migration speed (Fig. 8).
Based on the changes in cell-cell adhesion and migration speed observed, the effects of p120ctn, E-cadherin, P-cadherin, band acatenin on invasion through Matrigel was investigated. Knockdown of p120ctn with two different siRNAs increased invasion through Matrigel whereas E-cadherin depletion did not affect invasion (Fig. 9A, B). The knockdown of P-cadherin on the other hand led to an increase in invasion similar to p120ctn depletion (Fig. 9). The depletion of b-catenin as well as a-catenin induced an increase in invasion similar to p120ctn and P-cadherin. To rule out the possibility that differences in the number of cells on the bottom of Matrigel-coated Transwells were due to altered cell proliferation, cells were counted 48 h after they had been transfected with siRNAs. No changes in cell numbers compared to control siRNA-transfected cells were observed (Fig. 9C). In  Discussion p120ctn expression is frequently lost or downregulated in a large variety of human cancers [35] and decreased p120ctn levels correlate with the tumour grade [36], [37], [38], [39]. We have shown here that the depletion of p120ctn in E-cadherin and Pcadherin-expressing DU145 and MCF10A cells disrupts cell-cell contacts, and increases migration and invasion of DU145 cells, supporting a model where p120ctn acts as an invasion or metastasis suppressor [48]. These effects correlated with decreased expression of the adherens junction proteins E-cadherin, P-cadherin, ß-catenin and a-catenin. However, only downregulation of P-cadherin and not E-cadherin was able to induce a similar phenotype to p120ctn depletion, indicating that E-cadherin is not required to maintain cell-cell adhesion in DU145 or MCF10A cells.
In contrast to our results, p120ctn has been reported to be required for invasion of E-cadherin-deficient cells in part by stabilising the mesenchymal cadherins N-cadherin or cadherin-11, which are known to promote invasion [32]. On the other hand the collective invasion of E-and P-cadherin-expressing A431 cells was inhibited when cell-cell contacts were disrupted either by p120ctn knockdown or the simultaneous knockdown of E-cadherin and Pcadherin [35]. In these studies HGF or EGF, respectively, were used as chemoattractants in invasion and migration assays, whereas we used serum. One possibility is that p120ctn could specifically contribute to HGF/EGF-induced migration or invasion as it has previously been linked to HGF-or EGF-induced cell scattering [40].
p120ctn has been shown to stimulate Rac and/or Cdc42 activity and/or inhibit RhoA activity in cells lacking E-cadherin, including fibroblasts and cancer cell lines [32], [41]. In contrast, we found that in DU145 cells that express P-and E-cadherin but not N-cadherin, the depletion of p120ctn did not affect activity of RhoA and Rac1. Similarly, in E-cadherin/P-cadherin-expressing A431 cells, p120ctn depletion did not alter RhoA/Rac1 activity [31].
Taken together, these results indicate that the effects of p120ctn on Rho GTPase activities are cell type-specific. This could be due to differences in the composition of p120ctn complexes, for example complexes containing GEFs that could activate Rac1/ Cdc42 or GAPs to downregulate RhoA [26], [41]. Interestingly, in MDA-MB-231 breast cancer cells, activation of Rac by p120ctn requires cadherin binding [32], suggesting that mesenchymal cadherins could specifically contribute to Rac activity. Although Rho GTPase activity was not affected by p120ctn, we found that Rap1 activity is increased in p120ctn-depleted DU145 cells. Rap1 is an important regulator of cell-cell junctions, and is activated by E-cadherin engagement [42,43]. Disruption of adherens junctions by extracellular calcium depletion also induces an increase in Rap1 activity [44]. Our results with p120ctn depletion are consistent with the hypothesis that endocytosis of E-cadherin increases Rap1 activity [44].
Surprisingly, the knockdown of P-cadherin, band a-catenin but not E-cadherin led to the disruption of cell-cell contacts. Presumably P-cadherin substitutes for E-cadherin to maintain cell-cell contacts, but not E-cadherin for P-cadherin. This hypothesis is supported by studies in MDCK cells, where it has been shown that E-cadherin is important only for the establishment but not maintenance of cell-cell contacts, whereas a-catenin was required to maintain cell-cell contacts [45]. In addition, P-cadherin depletion alone was reported to induce loss of cell-cell contacts and downregulation of E-cadherin levels in MCF10A cells [46] and overexpression of P-cadherin in MDA-MB-231 breast cancer cells reduced cell migration and invasion [47]. In cancer most studies have concentrated on the link between E-cadherin downregulation and increased cancer aggressiveness and invasion, although there are also indications that P-cadherin affects tumour progression. For example, P-cadherin levels are downregulated during melanoma progression [48] and P-cadherin is often lost in prostate cancers [49]. In DU145 cells, we found that P-cadherin depletion induces loss of cell-cell junctions without significantly affecting the levels of other adherens junction proteins, indicating that it is the major cadherin required for cell-cell adhesion in these cells.
The knockdown of band a-catenin also induced disruption of cell-cell contacts and an increase in invasion but did not change Pcadherin or p120ctn protein levels. It is possible that they affect the localization of P-cadherin and/or p120ctn rather than levels. band a-catenins have been reported to affect cadherin-mediated adhesion [15], [16], and the depletion of b-catenin in an E-cadherin-deficient cell line led to a decrease in invasion [50], but so far little is known about their roles in migration and invasion. b-catenin also has an important role in Wnt-signalling and cancer cell proliferation which is thought to be independent of its cadherin function [51].
Our results demonstrate that p120ctn regulates the expression levels of ß-and a-catenin as well as cadherins in DU145 cells. Since we have shown that P-cadherin levels are more important than E-cadherin in maintaining cell-cell adhesion in DU145 and MCF10A cells, we postulate that the increased migration and invasion we observe following p120ctn, ß-catenin or a-catenin knockdown is due either to decreased P-cadherin stability or changes in its localization.

Cell culture and transfections
DU145 prostate cancer cells were grown in RPMI supplemented with 10% FCS, 100 IU/ml penicillin and 100 mg/ml streptomycin. MCF10A mammary epithelial cells were grown in DMEM/F12 supplemented with 5% horse serum, 100 IU/ml (MCF10A) for immunofluorescence and time-lapse microscopy and at 1610 5 per well (24-well plate) for immunoblotting and scratch-wound assays. For creation of stable p120ctn-depleted DU145 cells, the pSuperior vector system (Oligoengine, Seattle, WA) was used. The sequence of the shRNA oligo targeting human p120ctn was designed as described [52]. A stable cell pool containing the shRNA was selected in medium containing  [26], which is not sensitive to the shRNA targeting human p120ctn, was transiently transfected into DU145 cells using Amaxa nucleofection (Lonza, Cologne, Germany) according to the manufacturer's protocol (www.lonzabio.com).

RhoA, Rac1 and Rap1 activity assays
GST-Rhotekin-RBD, GST-PAK-PBD and GST-RalGDS-RBD were expressed in Escherichia coli and purified as previously described on glutathione beads [53]. Cells were transfected with siRNAs on 10-cm dishes. After 72 h, cells were lysed in 50 mM Tris pH 7.5, 1% Triton-X-100, 0.5% sodium deoxycholate, 0.1% SDS, 500 mM NaCl, 10 mM MgCl 2 , 2 mM EDTA, 10% glycerol, 1 mM Na 3 VO 4 , 25 mM NaF, complete mini EDTAfree protease inhibitor (Roche, Welwyn Garden City, UK), PhosSTOP phosphatase inhibitor (Roche). Lysates were cleared by centrifugation (17,000 g, 30 min, 4uC). A 50 ml aliquot was retained for determination of total levels of each GTPase, and GTPases in 500 ml of control lysate were loaded with GTPcS by incubation with 10 mM EDTA and 100 mM GTPcS for 15 min, followed by 60 mM MgCl 2 to stop the reaction. This and remaining lysates were incubated with 20 ml of GST-fusion proteins on beads (1 h, 4uC, rotating). Proteins were eluted by boiling in 26 SDS sample buffer, resolved by SDS-PAGE analysed by immunoblotting as described above.

Immunofluorescence
Cells were seeded on glass coverslips, fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 and blocked with 2% BSA before staining. Primary antibodies (as listed above; Immunoblotting) were used at a dilution of 1:200. Alexa Fluor 488, or 546 or 647-conjugated goat anti-mouse IgG or goat antirabbit IgG (Invitrogen) were used at a dilution of 1:1000. F-actin was detected by staining with phalloidin conjugated to Alexa Fluor 488, 546 or 647 (Invitrogen) at a dilution of 1:1000. Cells were visualised using a Zeiss LSM510 confocal laser-scanning microscope with an EC Plan-Neofluar 406/1.3 oil DIC M27 objective

Time-lapse microscopy
Time-lapse movies were acquired over a period of up to 16 h using a Nikon TE2000-E microscope with a Plan Fluor 106 or 206objective (Nikon, Kingston, UK) and a Hamamatsu Orca-ER digital camera. Image series were captured at 37uC and 5% CO 2 at 1 frame/10 min using Metamorph software (Molecular Devices, Wokingham, UK). Cells were tracked and migration speed (in mm/min) and persistence (displacement/track length, where displacement is the distance from the start to end point for each cell) were determined using ImageJ analysis software (http:// rsb.info.nih.gov/ij) and ibidi chemotaxis and migration tool (www. ibidi.com). For scratch-wound assays, area occupied by cells in scratch wounds was determined from phase-contrast images taken at 0 h, 4 h, 12 h and 16 h after wounding using Photoshop (CS4, Adobe).

Matrigel invasion assays
Matrigel-coated transwell filters with a PET membrane containing 8 mm pores (BD Biosciences) were rehydrated for 2 h at 37uC in medium without supplements. Transfected cells were washed once in PBS, and 10 5 cells were seeded into the upper chamber of the transwells in antibiotic-free medium containing 0.1% FCS and 0.1% BSA. Medium containing 2% FCS and 0.1% BSA was placed in the bottom chamber as a chemoattractant. Cells were allowed to invade for 17 h. Remaining cells and matrix were then removed from the upper side of the membrane facing the upper chamber. The cells on the bottom part of the membrane were fixed in methanol containing 0.1% crystal violet. Eight separate bright-field images were taken of each transwell filter using a Nikon TE2000-E microscope with a Plan Fluor 106 objective. The cells per image were counted and analysed compared to control transfected cells.

Statistics
Statistical analysis was carried out where indicated using data from three or more independent experiments. Statistical significance was calculated in Excel (Microsoft) using an unpaired Student's t-test. Figure S1 p120ctn expression rescues the disruption of adherens junctions by p120ctn depletion. Stable p120ctn-knockdown DU145 cells were transiently transfected with plasmids encoding GFP or GFP-p120ctn (wild-type murine cDNA). After 24 h, cells were fixed and stained for E-cadherin. The boxed region (left panel) is shown enlarged in the magnified panels. Scale bars = 50 mm. Found at: doi:10.1371/journal.pone.0011801.s001 (1.93 MB TIF) Figure S2 Effects of adherens junction protein depletion using additional siRNA oligos. DU145 cells were transfected with the indicated siRNAs for E-cadherin, P-cadherin, a-catenin, b-catenin or p120ctn. After 72 h, cells were fixed and stained for p120ctn and F-actin. Found at: doi:10.1371/journal.pone.0011801.s002 (2.73 MB TIF) Figure S3 Disruption of cell-cell adhesion following depletion of P-cadherin using 3 different siRNA oligos. DU145 cells were transfected with the three different siRNAs targeting P-cadherin used in Figure 5. After 72 h, cells were fixed and stained for Pcadherin and F-actin.