Pleiotrophin disrupts calcium-dependent homophilic cell–cell adhesion and initiates an epithelial–mesenchymal transition

Abstract

Regulation of the levels of tyrosine phosphorylation is essential to maintain the functions of proteins in different signaling pathways and other cellular systems, but how the steady-state levels of tyrosine phosphorylation are coordinated in different cellular systems to initiate complex cellular functions remains a formidable challenge. The receptor protein tyrosine phosphatase (RPTP)β/ζ is a transmembrane tyrosine phosphatase whose substrates include proteins important in intracellular and transmembrane protein-signaling pathways, cytoskeletal structure, cell–cell adhesion, endocytosis, and chromatin remodeling. Pleiotrophin (PTN the protein and Ptn the gene) is a ligand for RPTPβ/ζ; PTN inactivates RPTPβ/ζ, leaving unchecked the continued endogenous activity of tyrosine kinases that increase phosphorylation of the substrates of RPTPβ/ζ at sites dephosphorylated by RPTPβ/ζ in cells not stimulated by PTN. Thus, through the regulation of the tyrosine phosphatase activity of RPTPβ/ζ, the PTN/RPTPβ/ζ signaling pathway coordinately regulates the levels of tyrosine phosphorylation of proteins in many cellular systems. We now demonstrate that PTN disrupts cytoskeletal protein complexes, ablates calcium-dependent homophilic cell–cell adhesion, stimulates ubiquitination and degradation of N-cadherin, reorganizes the actin cytoskeleton, and induces a morphological epithelial–mesenchymal transition (EMT) in PTN-stimulated U373 cells. The data suggest that increased tyrosine phosphorylation of the different substrates of RPTPβ/ζ in PTN-stimulated cells alone is sufficient to coordinately stimulate the different functions needed for an EMT; it is possible that PTN initiates an EMT in cells at sites where PTN is expressed in development and in malignant cells that inappropriately express Ptn.

The balanced activities of tyrosine kinases and tyrosine phosphatases dynamically regulate the steady-state levels of tyrosine phosphorylation of key proteins essential for many important cellular functions. The regulated disruption of this balance through growth factor and/or cytokine-activated receptor-transduced signals is an important mechanism of signal transduction but, when deregulated, is a mechanism frequently underlying different diseases and a major feature in the pathogenesis of many human malignancies (1). An important gap, however, is in understanding the mechanism of how different pathways and systems are coordinated to initiate the many different cellular functions required for normal cellular homeostasis, proliferation, and differentiation of cells.

The diverse substrates of the receptor protein tyrosine phosphatase (RPTP)β/ζ (2) include β-catenin (2), β-adducin (3, 4), Fyn (5), GIT1/Cat-1 (6), and P190RhoGAP (7), indicating that RPTPβ/ζ is promiscuous in substrate specificity, but through its activity, is critically positioned to coordinately regulate the steady-state levels of tyrosine phosphorylation of proteins in different signaling networks and cellular systems. Pleiotrophin (PTN the protein and Ptn the gene) is a secreted, highly conserved cytokine (8, 9) that signals through inactivation of RPTPβ/ζ; as a consequence, PTN coordinately increases tyrosine phosphorylation of the many substrates of RPTPβ/ζ through persistent activity of the tyrosine kinases that phosphorylate the same sites that are dephosphorylated by RPTPβ/ζ in cells not stimulated by PTN. The diverse substrates regulated through the PTN/RPTPβ/ζ signaling pathway thus are likely to account for the diverse functions signaled by PTN in different cellular systems and in the different malignant cell lines with inappropriate expression of Ptn (10).

In these studies, we pursued the biochemical and phenotypic consequences of the PTN-dependent inactivation of RPTPβ/ζ in PTN-stimulated U373 cells; the data demonstrate that PTN stimulates a morphological epithelial–mesenchymal transition (EMT) in U373 cells and, thus, suggest that the diversity of responses needed for an EMT are initiated coordinately through the PTN-dependent increase in tyrosine phosphorylation of substrates of RPTPβ/ζ in different signaling networks.

Results

Pleiotrophin Stimulates Increased Tyrosine Phosphorylation of β-Catenin.

Pleiotrophin stimulates increased tyrosine phosphorylation of β-catenin through inactivation of RPTPβ/ζ (2). Because tyrosine phosphorylation of β-catenin is known to perturb adherent junction protein complexes and homophilic cell–cell adhesion (11, 12), the increase in tyrosine phosphorylation of β-catenin was compared with the ability of β-catenin to associate with N-cadherin in PTN-stimulated U373 cells. U373 cells were stimulated with PTN for 60 min with increasing concentrations of PTN; β-catenin was immunoprecipitated with anti-β-catenin antibodies from lysates of control, nonstimulated cells and PTN-stimulated cells and analyzed in Western blots probed with anti-phosphotyrosine antibodies. As found in ref. 2, PTN induced a rapid increase in tyrosine phosphorylation of β-catenin when U373 cells were stimulated with PTN at concentrations from 0 to 10 ng/ml and slightly higher levels of tyrosine phosphorylation were seen as the concentration of PTN was increased to 25 ng/ml. In cells stimulated with 50 and 100 ng/ml PTN, concentrations of PTN previously found to be in excess of saturating concentrations (2, 3), the levels of tyrosine phosphorylation of β-catenin fell somewhat (Fig. 1). The reason for the decrease in tyrosine phosphorylation of β-catenin in cells stimulated with PTN in excess of saturation is unknown. A very striking increase in tyrosine phosphorylation of β-catenin also was seen in U373 cells stimulated with sodium pervanadate (50 μM) (ref. 13; Fig. 1, lane 6), suggesting that more than one tyrosine phosphatase regulates steady-state tyrosine phosphorylation levels of β-catenin.

Fig. 1.

Confluent U373 cells were not stimulated or stimulated with 10, 25, 50, or 100 ng of PTN/ml for 60 min. Confluent U373 cells also were preincubated with pervanadate or with anti-PTN antibodies (5 μg/ml) before incubation with 50 ng/ml PTN. Cell lysates were prepared and immunoprecipitated with anti-β-catenin antibodies, and the immunoprecipitates were analyzed in Western blots probed with anti-phosphotyrosine, anti-pan-cadherin, and anti-β-catenin antibodies. A statistical analysis of the inverse correlation of the levels of the phosphorylation of β-catenin with the levels of the association of β-catenin with N-cadherin is shown (Fig. 8A).

Loss of Association of β-Catenin with N-Cadherin in PTN-Stimulated Cells.

The levels of N-cadherin that coimmunoprecipitate with β-catenin from lysates of nonstimulated and PTN-stimulated cells were compared in Western blots and quantitated by scanning densitometry. An inverse linear relationship between levels of tyrosine phosphorylation of β-catenin and levels of N-cadherin that co-immunoprecipitated with β-catenin was demonstrated with a coefficient of variation of −0.98 (see Fig. 8A, which is published as supporting information on the PNAS web site). Anti-PTN antibodies (5 μg/ml) added together with PTN (50 ng/ml) effectively blocked the PTN-dependent increase in tyrosine phosphorylation of β-catenin and loss of its association with N-cadherin (Fig. 1, compare lanes 1 and 7), establishing the specificity of PTN as responsible for the increase in tyrosine phosphorylation of β-catenin and loss of association of β-catenin with N-cadherin in PTN-stimulated cells.

In recent studies, Wellstein et al. (14–16) reported that the physiological receptor of PTN is anaplastic lymphoma kinase; in the present studies, we used a short hairpin RNA to “knock down” RPTPβ/ζ and demonstrated that RPTPβ/ζ is required for the PTN-initiated responses we report (Fig. 9, which is published as supporting information on the PNAS web site). Furthermore, by using RT-PCR and Western blot analysis, we found that anaplastic lymphoma kinase is not expressed in U373 cells as reported in ref. 17. Thus, the data we provide depends on the PTN/RPTPβ/ζ signaling pathway.

Disruption of Adherent Junction Complexes in PTN-Stimulated Cells.

N-cadherin associates also with γ-catenin, P120, and IQGAP-1 in adherent junction complexes (11). β-catenin and γ-catenin are known to compete for the same site in the C-terminal region of the cadherins and through α-catenin link the cadherins to filamentous actin to stabilize cell–cell adhesion. Recent evidence, however, suggests that α-catenin regulates actin filament assembly and does not associate simultaneously with actin and β-catenin (18).

P120 binds to the juxtamembrane domain of cadherins and regulates cadherin turnover at the cell surface (19). IQGAP1 associates more strongly with cadherins and β-catenin under conditions in which cells have lost cell–cell adhesion (20).

To test the association of N-cadherin with γ-catenin, P120, and IQGAP1 in PTN-stimulated U373 cells, lysates of control (nonstimulated) and PTN-stimulated cells were prepared and immunoprecipitated with anti-pan-cadherin antibodies; the immunoprecipitates were analyzed in Western blots probed with anti-β-catenin, anti-γ-catenin, anti-P120, and anti-IQGAP-1 antibodies (Fig. 2). In the control studies, equal amounts of N-cadherin were immunoprecipitated with anti-pan-cadherin antibodies from control (nonstimulated) and PTN-stimulated cells (Fig. 2). In PTN-stimulated cells, N-cadherin was found to associate no longer with γ-catenin (Fig. 2) and to decrease the levels of association with β-catenin (≈84%). A modest reduction in the levels of association of N-cadherin with P120 (≈27%) and IQGAP-1 (≈32%) also was observed in lysates of PTN-stimulated cells (Fig. 2). An ≈2.4-fold increase in tyrosine phosphorylation of β-catenin was demonstrated in immunoprecipitates from lysates immunoprecipitated with anti-β-catenin antibodies in comparison with lysates from unstimulated cells; the increase in tyrosine phosphorylation of β-catenin was associated with a ≈59% loss of association of β-catenin with α-catenin (Fig. 2).

Fig. 2.

Western blots were prepared from immunoprecipitates of lysates incubated with anti-pan-cadherin antibodies of confluent nonstimulated control U373 cells or of U373 cells that were stimulated with 50 ng/ml PTN for 60 min. The blots were probed with anti-P120, anti-β-catenin, anti-γ-catenin, anti-IQGAP1, or anti-pan-cadherin antibodies. Lysates from the same cells were immunoprecipitated with anti-β-catenin antibodies and analyzed in Western blots probed with anti-phosphotyrosine, anti-α-catenin, or anti-β-catenin antibodies.

Loss of Adherent Junction Protein Levels in PTN-Stimulated Cells.

P120 catenin regulates turnover of cadherins at the cell surface, suggesting the possibility that N-cadherin maybe subject to proteolytic degradation through the ubiquitin-proteasome proteolysis system targeted by P120 catenin (19). To test the possibility that PTN stimulates degradation of N-cadherin in PTN-stimulated cells, lysates from PTN-stimulated and control, nontreated cells were prepared and analyzed in Western blots probed with anti-pan-cadherin antibodies. The levels of N-cadherin in lysates of U373 cells stimulated with PTN for 5, 10, and 20 min were reduced to a level ≈50% the levels of N-cadherin in U373 cells before stimulation (Fig. 3A). The same lysates were analyzed in Western blots probed with anti-α-catenin and anti-β-catenin antibodies (Fig. 3A); the levels of β-catenin decreased to ≈80% of the levels in control, nonstimulated cells 10 min after PTN stimulation and remained essentially constant for the remainder of the experiment. The levels of α-catenin rapidly and progressively decreased at 20 min to levels ≈25% of the control, nonstimulated cells. After 60 min, the levels of α-catenin returned to ≈50% of the levels observed in control, nonstimulated cells, and the levels of each protein returned to baseline in the subsequent 1–2 h (data not shown).

Fig. 3.

Degradation of adherent junction proteins in PTN-stimulated cells. (A) Lysates of confluent U373 cells stimulated with 50 ng/ml PTN for 0, 5, 10, 20, and 60 min were analyzed in Western blots probed with anti-pan-cadherin (Upper Top), anti-α-catenin (Upper Middle), and anti-β-catenin antibodies (Upper Bottom). (Lower) The results also were analyzed by using densitometry and expressed as percent of intensity compared with control nonstimulated cells. (B) Western blots of proteins captured by Rad23 from lysates of U373 cells stimulated with 50 ng/ml PTN for 30 and 60 min probed with anti-pan-cadherin antibodies and reprobed with anti-ubiquitin antibodies.

To test whether N-cadherin is ubiquitinated, Rad-23 conjugated to agarose beads was used to “capture” ubiquitinated proteins from control and PTN-stimulated U373 cells; Rad-23 binds ubiquitin, monoubiquitinated, and polyubiquitinated proteins (21) with high affinity and specificity. Lysates were incubated with Rad-23, and proteins captured by Rad-23 were analyzed in Western blots probed with either anti-pan-cadherin antibodies or with antibodies that recognize mono- or polyubiquitin covalently coupled to proteins (see Experimental Procedures). The results demonstrate that levels of ubiquitinated N-cadherin captured by Rad-23 were increased ≈4- and ≈5-fold above the levels of ubiquitinated N-cadherin in control, nonstimulated cells 30 and 60 min after stimulation with 50 ng/ml PTN (Fig. 3B, lane 1–3). The Western blots probed with anti-ubiquitin antibodies confirmed equal loading and that many ubiquitinated proteins are pulled down by Rad23 in PTN-stimulated cells. We also preincubated cells with lactacystin (10 μM) before they were stimulated with PTN; lactacystin prevented the degradation of N-cadherin seen in PTN-stimulated cells that were not incubated with lactacystin (data not shown). The data suggest that PTN rapidly targets N-cadherin for proteolysis through the ubiquitin proteasome degradation pathway in PTN-stimulated cells; however, additional experiments are needed to fully characterize the basis of N-cadherin degradation in PTN-stimulated cells.

Loss of Homophilic Cell–Cell Adhesion in PTN-Stimulated Cells.

To pursue the possibility that PTN also disrupts homophilic cell–cell adhesion, the ratio of cells dissociated in calcium-containing media (N_TC) vs. the cells dissociated in calcium-free media (N_TE) (a “dissociation index,” see Experimental Procedures) was measured in PTN-stimulated and in control, nonstimulated cells. The ratio N_TC/N_TE of U373 cells stimulated with 50 ng/ml PTN in this assay was 0.85, whereas the ratio N_TC/N_TE of nonstimulated U373 cells was 0.51, demonstrating directly that PTN stimulation decreases calcium-dependent homophilic cell–cell adhesion in PTN-stimulated U373 cells (see Fig. 8B).

Loss of Adherent Junction Complexes and Cell–Cell Adhesion in PTN-Stimulated Cells.

To directly visualize the structure of cell–cell adhesion complexes in PTN-stimulated cells, we treated control and PTN-stimulated confluent U373 cells with anti-β-catenin and anti-pan-cadherin FITC-tagged antibodies and compared the two preparations by using confocal microscopy at different planes in the z axis. Fig. 4 is representative of the different transverse planes of the z axis taken to illustrate points of cell–cell contact from untreated and PTN-stimulated cells; it demonstrates that in cells not stimulated with PTN, both β-catenin and N-cadherin are evenly distributed at juxtamembrane sites at sites of cell–cell contact (Fig. 4 A and C). In contrast, in cells stimulated with PTN for 60 min, the intensity of immunostaining of both β-catenin and N-cadherin is greatly decreased at the sites of cell–cell contact; neither immunoreactive β-catenin nor N-cadherin were consistently seen at adherent junctions, and the cell–cell adhesion complexes are disrupted. Furthermore, loss of cell–cell contact is seen at different sites (Fig. 4 B and D, arrows). The loss of cell–cell adhesion complexes and homophilic cell–cell contact at different sites occurs in parallel with loss of association of N-cadherin and β-catenin in PTN-stimulated cells illustrated above.

Fig. 4.

Confocal microscopy analysis of confluent U373 cells not stimulated (A and C) or stimulated (B and D) with 50 ng/ml PTN for 60 min were stained by using FITC-tagged anti-β-catenin (A and B) and anti-pan-cadherin antibodies (C and D) and analyzed by using confocal microscopy. The section is from a single plane in the z axis and is representative of the entire z axis from both control and PTN-stimulated cells. Loss of cell–cell contact is seen at different sites (arrows).

Pan-cadherin immunoreactive proteins are also seen scattered within the cytosolic compartment of PTN-stimulated cells in Fig. 4 B and D. The pan-cadherin immunoreactive proteins appear to be in vesicles and likely are degradation products of N-cadherin targeted for proteolysis through the ubiquitin-proteasome proteolysis pathway. However, endocytosis through the lysosomal vesicles recently has emerged as a regulatory mechanism to modulate the levels of cadherin cell surface expression in epithelial cells (22, 23), raising the possibility that loss of association of N-cadherin with P120 in PTN-stimulated cells may “uncap” N-cadherin in PTN-stimulated cells and target N-cadherin for degradation within the lysosome compartments (24). U373 cells therefore were stimulated with PTN for 60 min, and cells stained with anti-N-cadherin antibodies and anti-LAMP1 antibodies to mark lysosomal compartments were analyzed with confocal microscopy at different z axis planes through the cell. The data in a representative focal plane (Fig. 5) failed to demonstrate colocalization of N-cadherin and the lysosomal protein LAMP1 in nonstimulated or in PTN-stimulated cells.

Fig. 5.

U373 cells not stimulated (A) or stimulated (B) with 50 ng/ml PTN for 60 min were stained by using FITC-tagged anti-pan-cadherin antibodies and Texas red-conjugated anti-LAMP1 antibodies to visualize lysosomes and observed by using confocal microscopy. The sections are from a single plane in the z axis and are representative of the entire z axis (see Fig. 10) from both PTN-stimulated and control cells.

The images from the different z axis planes then were used to quantitatively analyze the height of PTN-stimulated U373 cells compared with nonstimulated cells (see Fig. 10, which is published as supporting information on the PNAS web site). The distance between the highest and the lowest focal planes in PTN-stimulated cells measured in 10 separate microscopic fields in three different slides from independent experiments was 5.31 μm, whereas in cells that were not stimulated, the average height was 3.97 μm; this difference in cell height furthermore demonstrates that PTN induces striking changes in the cytoskeletal architecture of PTN-stimulated cells (see Fig. 8C).

Pleiotrophin Stimulates a Reorganization of the Actin Cytoskeleton and Initiates an Epithelial–Mesenchymal-Like Transition.

To pursue further the ability of PTN to alter the cytoskeletal structure of PTN-stimulated cells, PTN-stimulated and control, nonstimulated cells were stained with FITC-tagged anti-tubulin antibodies and Texas red-tagged phalloidin to visualize F-actin and observed with confocal microscopy. U373 cells characteristically are rounded but relatively flat; they have a large cytosolic compartment and a centrally located nucleus. In nonstimulated cells, F-actin was localized in sites immediately beneath the plasma membrane and in stress fibers throughout the cytoplasm (Fig. 6A–C). In contrast, PTN-stimulated U373 cells for 1 h were elongated; they had a fibroblast-like shape and had numerous evenly distributed, nonpolarized filopodia and lamelipodia protruding from the cell membrane. F-actin was localized clearly in those protrusions and also in the cortical regions. The PTN-stimulated cells appear to have a markedly reduced cytosolic compartmental volume (Fig. 6 D–F). It was found that the numbers of cells with the mesenchymal phenotype increased ≈40% within 1 h in PTN-stimulated cell cultures. These data and the data presented above are consistent with the conclusion that PTN stimulates an EMT.

Fig. 6.

U373 cells were seeded in culture plates and incubated for 1 h with serum-free media (A–C) or media containing 50 ng/ml PTN (D–F). The cells were stained by using FITC-tagged anti-tubulin antibodies (B and E), Texas red-conjugated phalloidin to visualize F-actin (A and D), and DAPI to visualize nuclei and observed by using confocal microscopy. Overlay images are shown in C and F.

U373 cells treated with PTN for 60 min and control nontreated cells were compared with scanning electron microscopy. U373 cells usually are round and flat-shaped with numerous cytoplasmic extensions (Fig. 7C, arrows). The PTN-stimulated U373 cells again acquired an elongated morphology consistent with mesenchymal phenotype and enhanced motility, and the small cytoplasmic extensions used for cells to initiate cell–cell contact were blunted or lost (Fig. 7D); this observation was confirmed when the cells were stained by using anti-β-catenin antibodies and observed with confocal microscopy (Fig. 7 A and B). The data thus confirm the morphological transition of U373 cells treated with PTN for 60 min from an epithelial-like to fibroblast-like phenotype and, thus, the transition to the mesenchymal phenotype.

Fig. 7.

U373 cells not stimulated (A and C) or stimulated with 50 ng/ml PTN for 60 min (B and D) were stained by using anti-β-catenin antibodies and visualized by using confocal microscopy (A and B). In separate experiments, U373 cells not stimulated or stimulated with 50 ng/ml PTN for 60 min were visualized by using scanning electron microscopy (C and D). Cytoplasmic extensions used for cells to initiate cell–cell contact (A and C, arrows) are blunted or lost (B, arrows) in PTN-stimulated cells.

Discussion

These studies support the conclusion that PTN stimulates an EMT in PTN-stimulated U373 cells; these features of an EMT include loss of cell–cell adhesion, increased motility and invasiveness, and the morphological phenotype needed for cells to exit early developmental sites for subsequent differentiation in the process of morphogenesis and organ development (25). To initiate an EMT, many different properties of the cell need to be altered coordinately for the cell to assume the mesenchymal phenotype; in this context, for PTN to stimulate an EMT, PTN needs to coordinate different regulatory pathways and cellular systems through initiation of a single signaling mechanism. Many malignant cells with inappropriate expression of Ptn also exhibit these features of an EMT needed for the malignant cells to invade locally and metastasize effectively. In these cases, because the malignant cells constitutively express Ptn, the malignant cells have a “stable EMT” or perhaps more appropriately, an “arrested EMT.”

Pleiotrophin signals through inactivation of the endogenous protein tyrosine phosphatase activity of RPTPβ/ζ; it increases tyrosine phosphorylation of the different substrates of RPTPβ/ζ through the persistent activity of tyrosine kinases acting at the same sites as RPTPβ/ζ (2–5). Pleiotrophin was the first natural ligand to be discovered for this class of receptor-type transmembrane tyrosine phosphatases (2) and the PTN/RPTPβ/ζ signaling pathway triggered by PTN is unique. β-catenin was identified as the first substrate of RPTPβ/ζ (2), and, subsequently, β-adducin (3, 4), Fyn (5), GIT-1 (6), and P190RhoGAP (7) have been identified as substrates of RPTPβ/ζ and downstream targets of the PTN/RPTPβ/ζ-signaling pathway. Through regulation of tyrosine phosphorylation of these different substrates of RPTPβ/ζ, PTN thus regulates proteins involved in cytoskeletal stability and function, endocytosis, chromatin remodeling, and both intracellular and transmembrane receptor-type tyrosine kinase signaling pathways. RPTPβ/ζ thus is centrally positioned to coordinately regulate steady-state levels of tyrosine phosphorylation of different proteins in diverse pathways depending on the relative affinity and levels of each for RPTPβ/ζ. Pleiotrophin thus is positioned as the single mediator to up-regulate tyrosine phosphorylation of each of these proteins through its ability to inactivate the tyrosine phosphatase activity of RPTPβ/ζ (26). Because these studies were undertaken within the first hour of stimulation by PTN, it is reasonable to conclude that responses initiated by transcription activation are minimal, but it is the PTN/RPTPβ/ζ signaling pathway that directly initiates the profound changes that underlie the EMT through the coordinated increase of tyrosine phosphorylation of the different substrates of RPTPβ/ζ.

Phosphorylation of tyrosines 86, 142, and 654 in β-catenin is known to result in loss of affinity and, thus, loss of association of β-catenin with cadherins (27), and increased tyrosine phosphorylation of β-catenin is known to destabilize cytoskeletal structures and to ablate homophilic cell–cell adhesion (11) through decreased affinity of phosphorylated β-catenin for cadherins (27). β-adducin also was identified as a target of the PTN/RPTPβ/ζ-signaling pathway (3, 4), and PTN disrupts β-adducin–actin–spectrin complexes supporting cystoskeletal stability.

Finally, N-cadherin was found to be degraded in PTN-stimulated cells. Whereas the data suggest that degradation of N-cadherin is modulated through the ubiquitin-proteasome proteolytic pathway, a feature of PTN-stimulated cells likely to further destabilize the cytoskeleton, further experiments are needed to better determine the molecular mechanisms of N-cadherin degradation in PTN-stimulated cells. As cited in ref. 1, degradation of selected cytoskeletal and adhesion proteins, and particularly E-cadherin and N-cadherin, are necessary early steps in differentiation and in EMT. One candidate that may mediate degradation of N-cadherin in this study is the Cbl-like ubiquitin ligase Hakai, previously discovered to target cadherins for degradation upon dissociation of the cadherins from their binding partners (28).

In summary, this study supports the conclusion that through the regulation of the intrinsic tyrosine phosphatase activity of RPTPβ/ζ, PTN regulates the activity of different systems that are important in cytoskeletal stability and the strength of cell–cell adhesion, in stimulating degradation of cytoskeletal proteins, and potentially many more systems to initiate a morphological EMT that is similar to that in development and during tumor progression.

Experimental Procedures

Cell Lines.

U373 cells (human glioblastoma) (American Type Culture Collection, Manassas, VA) were grown in DMEM supplemented with 10% FBS and 1% penicillin/streptomycin at 37°C with 5% CO₂. U373 cells express high levels of RPTPβ/ζ and fail to express anaplastic lymphoma kinase as detected by using RT-PCR or Western blots (17). Cells were serum starved for 24 h before stimulation with PTN. PTN-Fc was prepared in the laboratory; PTN was purchased from R & D Systems (Minneapolis, MN).

Antibodies.

Anti-PTN, anti-IQGAP1, anti-actin, anti-pan-cadherin, anti-LAMP1, and anti-mouse IgG FITC-conjugated antibodies were obtained from Sigma–Aldrich (St. Louis, MO). Anti-polyubiquitin antibodies were obtained from Chemicon (Temecula, CA). Anti-β-catenin, anti-γ-catenin, anti-P120, anti-RPTPβ/ζ, and anti-phosphotyrosine antibodies were obtained from BD Biosciences (San Diego, CA). Anti-mouse IgG HRP-conjugated and anti-rabbit IgG HRP-conjugated antibodies were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Anti-Pan-Cadherin Antibodies.

In preliminary studies, N-cadherin was estimated to constitute ≈95% of the total cadherins in U373 cells. Anti-pan-cadherin antibodies were more effective in immunoprecipitation than specific anti-N-cadherin antibodies tested (data not shown) and, thus, used for immunoprecipitation and Western blots. The protein immunoprecipitated with anti-pan-cadherin antibodies is referred as N-cadherin unless otherwise noted.

Cell–Cell Aggregation Assays.

Confluent cells in 60-mm dishes were treated with 0.01% trypsin with either 1 mM calcium chloride or 1 mM EDTA for 10 min at 37°C, scraped, and dissociated by pipetting 10 times. Individual cells were counted by using a hemacytometer and the ratio N_TC/N_TE, the number of cells dissociated in presence of calcium (N_TC) divided by the number of cells dissociated in presence of EDTA (N_TE), the “dissociation index,” was determined as described in ref. 29.

Ubiquitination Assay.

Lysates of U373 cells in 50 mM Tris, pH 7.5/150 mM NaCl/1% Nonidet P-40/0.25% sodium deoxycholate/0.1% SDS/Complete EDTA-free Protease Inhibitor Mixture (Roche, Indianapolis, IN)/2 mM sodium orthovanadate were incubated overnight with Rad23 agarose-beads (Calbiochem, La Jolla, CA). Rad23 recognizes ubiquitin, monoubiquitinated, and polyubiquitinated proteins (21). The beads were washed four times in lysis buffer, boiled in loading buffer, and analyzed in Western blots.

Immunoprecipitation.

The samples were incubated with 1 μg/ml of either β-catenin or cadherin antibodies at 4°C overnight and subsequently with 50 μl of protein G conjugated to magnetic beads (Miltenyi, Auburn, CA) for 1 h at 4°C. The lysate was passed through a 1-ml column placed in a magnetic stand and washed five times with PBS. Bound proteins were eluted by using SDS-loading buffer.

Western Blots.

Protein concentrations in cell lysates were measured by using the BCA Protein Assay Kit (Pierce, Rockford, IL). Cell lysates were mixed with loading buffer, boiled for 5 min, and equal amounts of protein were run in polyacrylamide gels, transferred to nitrocellulose membranes blocked with 50 mM Tris/150 mM NaCl/0.1% Tween-20 (TBS-T)/5% nonfat milk for 1 h, and incubated with the primary antibodies at the dilutions indicated overnight in TBS-T with 5% BSA. After three washes in TBS-T, the membranes were incubated for 1 h with donkey anti-mouse secondary antibodies conjugated with horseradish peroxidase (Santa Cruz Biotechnology) diluted 1:5,000 in TBS-T with 5% nonfat milk, washed three times in TBS-T, and visualized by using the ECL Enhanced Method (Amersham, San Francisco, CA).

Immunostaining.

Cells grown on coverslips were fixed with 4% paraformaldehyde in 0.12 M phosphate buffer, pH 7.2, for 30 min, and washed three times with PBS. Nonspecific antibody binding was reduced by incubation in PBS with 1% BSA. The cells were permeabilized in PBS with 0.5% Triton X-100 for 1 h at room temperature, incubated overnight with primary antibodies at an appropriate dilution in PBS with 1% BSA and 0.5% Triton X-100, washed three times for 5 min, and incubated with the secondary antibodies conjugated with fluorescein diluted 1:100 in PBS with 1% BSA and 0.5% Triton X-100. The slides were washed in PBS, stained with Texas red-conjugated phalloidin (Molecular Probes, Eugene, OR), diluted 1:100 in PBS for 40 min, washed again in PBS, and mounted by using ProLong Antifade Kit (Molecular Probes), according to manufacturer's instructions and observed in a Nikon TE2000U microscope coupled with a confocal cell imaging CARV system.

Scanning Electron Microscopy.

U373 cells grown on culture coverslips were fixed with 2.5% glutaraldehide in 0.1 M phosphate buffer, pH 7.4, for 4 h at 4°C and washed with 0.1 M PBS, pH 7.2. Samples were dehydrated by using ethanol, treated with CO₂ by using a Balzers CPD Critical Point Dryer 030, and covered with gold by using a Balzers BAL-TEC SCD 050. Cells were observed and photographed by using a scanning electron microscope Cambridge Stereoscan 240.

Abbreviations

EMT

epithelial–mesenchymal transition

RPTP

receptor protein tyrosine phosphatase.