Modulation of pp60^c-src activity and cellular localization during differentiation of human trophoblast cells in culture

Most of the specialized functions of the human placenta are carried out by a deep layer of epithelium formed by multinucleated cells (syncytiotrophoblasts) that arise from the fusion of mononuclear cytotrophoblasts as shown for term placenta (Lobo et al., 1985; Loke, 1983) and for first trimester placenta (Contractor and Soorana, 1985). This morphogenetic process results in the functional differentiation of these cells, which, following fusion, can synthesize and secrete the characteristic protein hormones of pregnancy.This in vivo differentiation process is mimicked in vitro during the primary culture of extravillous cytotrophoblasts as described by Kliman et al. (1986). In vitro, cytotrophoblasts aggregate within 24 hours and then fuse together to form syncytia.

In vitro morphological differentiation is a time-dependent phenomenon characterized by an increase in epidermal growth factor receptor (EGF-R) expression (Alsat et al., 1993). EGF-R is a transmembrane glycoprotein of 170 kDa with an extracellular EGF-binding domain, a single hydrophobic membrane-spanning stretch and an intracellular tyrosine kinase domain. Binding of EGF to the receptor leads to dimerization of the receptors and subsequent activation of the tyrosine kinase, thereby catalysing tyrosine phosphorylation of several protein substrates including the receptor itself. Activation of the receptor kinase results in a cascade of biochemical and physiological responses. Morrish et al. (1987) reported that EGF induced morphological differentiation of cytotrophoblast cells in vitro.

Cytotrophoblast interaction is a Ca²⁺- and protein synthesis-dependent process and is associated with desmosome formation (Babalola et al., 1990). As the cells aggregate in culture, E-cadherin protein levels gradually increase and reach a maximum, which coincides with the maximum aggregate formation (Coutifaris et al., 1991). The localization of E-cadherin at points of cell-cell contact indicates a close temporal relationship between E-cadherin expression and cell aggregation.

The proto-oncogene product, pp60^c-src, is a tyrosine kinase with a still unresolved cellular function although it has been endowed with morphogenetic properties (Warren et al., 1988). Moreover, v-src-dependent tyrosine kinase activity was previously shown to cause disruption of cadherin-mediated junctions (Warren and Nelson, 1987; Volberg et al., 1991, 1992), to pertub cadherin function (Matsuyoshi et al., 1992) and to down-regulate EGF-R expression (Wasilenko et al., 1990). These observations led us to examine whether pp60^c-src is expressed in placenta cells and to investigate the possible involvement of pp60^c-src in the process of trophoblast differentiation, particularly with respect to EGF-R and E-cadherin protein whose expression is modulated during syncytial formation. Our data show that pp60^c-src has the cellular distribution and biochemical activity to function as a selective regulator of placental cell differentiation.

Term placentas were obtained after elective caesarian section from pregnant women near term with uncomplicated pregnancies. Villous tissue was dissected free of membranes and vessels, rinsed and minced in calciumand magnesium-free HBSS and subjected to trypsin-DNase digestion as previously described (Kliman et al., 1986). Briefly, minced tissue was dissociated at 37°C in HBSS containing 0.25% trypsin, 50 Kings units/ml DNase, 25 mM HEPES, 4.2 mM MgSO₄, pH 7.4, for 20 min. The cell suspension was harvested and layered on 5 ml fetal calf serum (FCS) in a 50 ml centrifuge tube to inhibit trypsin action and centrifuged at 800 g for 5 min at room temperature. The cell pellets were resuspended in DMEM containing 25 mM HEPES(DMEM-H) and carefully layered on a discontinuous Percoll gradient (5% to 70% in 5 steps). After centrifugation at 1,200 g for 20 min, the middle layer cells were removed (density: 1.048-1.062 g/ml) washed once with DMEM-H and diluted to a concentration of 1×106 cells/ml in DMEH-H supplemented with 2 mM glutamine, 20% heat-inactivated FCS and antibiotics(100 mg/ml streptomycin, 100 i.u. penicillin and 2.5 μg/ml amphotericin B). The cells were plated in 30 mm culture dishes (1×106 cells in 1.5 ml) and incubated in humidified 5% CO₂/95% air at 37°C. The medium was changed every 2 days.

More than 95% of the isolated cells were cytotrophoblasts as estimated by cytokeratin staining. Cell viability, as assessed by trypan blue exclusion was 95% or more. Mononucleated cytotrophoblasts exhibited little if any cell division but were converted to multinucleated cells.

The mouse monoclonal antibody mAb 327, Oncogene Sciences, commercialized by CliniSciences (Paris, France), was used to detect pp60c-src. The rat monoclonal antibody anti-E-cadherin (DECMA-1), used for immunohistochemistry, was purchased from Sigma. It was used at 1/100 dilution.

The anti-E-cadherin antibody used in immunoblotting was the polyclonal antibody directed against the extracellular domain of the molecule; it was a kind gift from Dr Benjamin Geiger, Weizman Institute of Sciences, Israel.

The rabbit polyclonal antibody against the carboxy-terminal portion of EGF-R was a generous gift from the Rorer Society (Pr. Schlessinger).

Herbimycin A was purchased from Kamiya’s Research Biochemical Company. It was reconstituted in DMSO.

Labeled IgG were from Amersham.

For immunofluorescence staining, trophoblast cells cultured on glass slides were first fixed in 3% paraformaldehyde, rinsed in 0.1 M glycine in PBS and permeabilized in 0.1% Triton X-100 in PBS for 10 min. They were incubated with mAb 327 monoclonal antibody against pp60c-src for 2 h. Bound antibody was visualized by incubation with a biotinylated second-step antibody against mice, followed by incubation with a fluorescent streptavidin conjugate.

Double immunofluorescence labeling for pp60 and E-cad-herin was performed using mAb 327 and a DECMA-1 rat monoclonal anti-E-cadherin antibody. Second antibodies were, respectively, the Fab fragments from sheep IgG against mouse IgG conjugated with tetramethyl rhodamine isothiocyanate (TRITC) and a rabbit IgG against rat IgG, followed by an anti-rabbit IgG conjugated with FITC.

Double immunofluorescence labeling for pp60c-src and EGF-R was performed using mAb 327 and the polyclonal anti-EGF-R antibody. The second antibody used for EGF-R detection was a biotinylated anti-rabbit antibody revealed by fluorescent streptavidin.

Fluorescence microscopy was performed with a Leitz microscope equipped with the filter setting for fluorescence and for phase-contrast, which allowed the identification of mononuclear cells and syncytia. Photographs were taken on a Kodak Tmax film. Nuclei were labeled with fluorescent Hoechst reagent. Control experiments were performed using a non-relevant antibody or by omitting the first antibody.

Cells in monolayer cultures in 100 mm dishes were placed on ice, washed 3 times with PBS containing 1 mM phenylmethylsulfonyl fluoride. After complete removal of the washing solution the cells were lysed in 400 to 600 μl of cell lysis buffer consisting of either boiling Laemmli solution for c-src, or RIPA buffer for E-cadherin and EGF-R (10 mM Tris-HCl, pH 7.5 containing 1% Triton X-100, 1% sodium deoxycholate, 150 mM NaCl, 0.5 mM CaCl2, 1 mg/ml aprotinin and leupeptin). The cell lysates were scraped off the plate and centrifuged at 10,000 g for 15 min to yield a supernatant, which was used for western blot analysis.

Samples of cell lysates corresponding to equivalent amounts of protein were subjected to SDS-polyacrylamide (10% gel) electrophoresis under reducing conditions. The proteins were electrophoretically blotted onto nitrocellulose filters (Schleider & Schuell, 0.45 μm) using a semi-dry apparatus (Bio-Rad). Unspecific binding sites were blocked by incubation in 1% non-fat dry milk dissolved in 10 mM Tris-HCl, pH 7.5, containing 0.17 M NaCl, for 1 h. The filters were then incubated overnight in the same buffer containing 0.5% Tween-20 with the appropriate dilutions of the antibodies. Filters were washed 3 times in PBS containing 0.5% Tween-20. When appropriate, the filters were incubated with a rabbit anti-mouse antibody (1/500 dilution) and washed before reaction with a peroxidase-labeled anti-rabbit antibody; the peroxidase reaction was developed with 4-chloronapthol as a substrate or using the luminescent ECL reagent (Amersham), which was subsequently visualized by autoradiography.

Dishes (10 cm) were washed twice with cold PBS and cells were lysed with a minimum volume of RIPA solution containing 200 mM sodium orthovanadate and 10 μg/ml aprotinin, antipain and leupeptin. Cell lysates were scraped, centrifugated to eliminate insoluble materials and stored at −20°C until use. Fresh vanadate was added after thawing. Protein content was measured on an aliquot by the Bio-Rad protein assay and protein content was normalized to 50 μg/sample. Lysates were incubated with antibodies or specific control for 1 h on ice; 50 μl Protein A-Sepharose (10% in RIPA), preincubated with anti-mouse immunoglobulin anti-serum, was added for half an hour under constant rotation at 4°C. Samples were centrifuged for 1 min at 3500 r.p.m. and, after 3 washes with RIPA, precipitates were dissolved in kinase buffer for in vitro phosphorylation. The c-src kinase assay was performed by adding an exogenous substrate according to the method of Piwnica et al. (1987), i.e. 10 μl of 100 μg/ml acid-denatured rabbit muscle enolase (Sigma) in 10 mM Tris-HCl, pH 7.2, containing 5 mM MgCl2, 100 μM sodium orthovanadate, 0.5 μM ATP and 2 μCi [γ-32P]ATP; after 15 min incubation at 30°C the reaction was stopped by adding the same volume of 2× sample buffer and heating at 100°C for 5 min. After 7.5% SDS-PAGE, gels were fixed, dried and exposed to Amersham films with an intensifying screen. Proteins were identified by their molecular mass.

At the appropriate time of culture the cells were washed twice with ice-cold PBS and once with homogenization buffer (0.25 M sucrose, 10 mM Tris-HCl, pH 7.8, 1 mM MgCl₂ and 10 μg/ml leupeptin, antipain, aprotinin). Cells were scraped off into 1 ml of buffer and homogenized with a Dounce homogenizer. The homogenate was centrifuged at 800 g for 10 min and the supernatant was further centrifuged at 12,000 g for 15 min. The crude membranes contained in the resulting pellet were resuspended in buffer containing 20 mM Tris-HCl, 50 mM NaCl, 1 mM CaCl₂, 2mM MgCl2, pH 7.4. Crude membrane proteins (5 mg) were incubated in the presence or absence of 15 μM EGF with 5 μM [γ-32P]ATP in a final volume of 50 μl of 16 mM HEPES buffer (pH 7.4) containing 0.1 mM AMP-PNP, 1 mM MnCl₂ and 0.1% BSA. After 10 min at 4°C, the reaction was stopped by adding 10 ml of Laemmli sample buffer and heating at 100°C for 5 min. The samples were then processed for SDS-polyacrylamide gel electrophoresis and autoradiographed.

In culture, cytotrophoblasts aggregate within 24 hours and then fuse to form syncytia This in vitro differentiation is associated with a modulation of both EGF-R and E-cadherin expression. To test our experimental conditions, western blot analyses were performed to establish the time course of EGF-R and E-cadherin expression after isolation of trophoblast cells and during their in vitro differentiation (Fig. 1A,B,C). As previously described (Alsat et al., 1993) the level of EGF-R increased at day 2 of culture when cells were aggregated. This increase was accompanied by an ele -vation in EGF-dependent phosphorylation of membrane proteins that remained stable thereafter. Among the phosphorylated species a prominent band of 170 kDa was revealed that corresponded to EGF-R as identified by immunoprecipitation. Similarly, as previously established (Coutifaris et al., 1991), we found that the expression of E-cadherin increased during the first 48 hours of culture, corresponding to the formation of cellular aggregates, and decreased concomitantly with the formation of multinucleated syncytia. Expression of EGF-R and E-cadherin was therefore confirmed as being involved in cell aggregation and fusion, and thus enabled us to test later the effect of a tyrosine kinase inhibitor, herbimycin A, on cell differentiation.

Immunocytochemical studies performed with a monoclonal antibody against pp60^c-src were done at days 1, 2 and 4 of culture. At day 1, most cells were still mononucleate and isolated, at day 2 they aggregated and began to fuse, at day 4 almost all cells were in syncytia. Immunolabeling with mAb 327 revealed several characteristic features of pp60^c-src localization: at day 1, in mononuclear cells, pp60^c-src was concentrated at sites of cell-substratum adhesion, resembling adhesion plaques on the plasma membrane and around the nucleus (Fig. 2A). At day 2, in aggregated cells pp60^c-src was partly located in the plasma membrane but also accumulated in regions of cell-to-cell contacts (Fig. 2B). In addition to the plasma membrane-associated staining, pp60^c-src immunoreactivity was located in many cells over the cell nucleus or in tight contact with the nuclear envelope.

In syncytia, the localization of pp60^c-src was diffuse in the cytoplasm and around nuclei. A syncytium was seen that was still aggregated with the remaining mononuclear cells (Fig. 2C,C′). When cell fusion occurred, the membranous cellular localization of pp60^c-src disappeared.

Double immunofluoresence labeling for pp60^c-src and E-cadherin revealed that both molecules are located at cell-to-cell contacts in aggregated trophoblast cells (Fig. 3A,B, respectively).

Double immunofluorescence staining for pp60^c-src and EGF-R revealed a colocalization in aggregated cells at sites of cell-substratum adhesion and also at cell-to-cell contacts (Fig. 4A,B, respectively).

Western blot analysis with mAb 327 antibody revealed a major protein with an apparent molecular mass of 60 kDa. The level of expression was tested at different time points during in vitro differentiation of the trophoblast cells and appeared to be constant throughout the period of cell culture (Fig. 5).

Cellular lysates were immunoprecipitated by the mAb 327 monoclonal antibody. Tyrosine kinase activity was assessed using the exogenous substrate enolase. The phosphorylated proteins were identified by their molecular mass. The phosphorylation of immunoprecipitated pp60^c-src and enolase increased dramatically at day 2 when cells undergo the fusion process and remained at this high level at day 4. The specificity of the immunoprecipitation was checked by omitting the pp60^c-src-specific antibody (Fig. 6A,B). To determine the alkali stability of the radiolabeled phosphate, the same SDS-polyacrylamide gel was incubated in hot alkali, which selectively hydrolyses phosphoserine and phosphothreonine residues. The ³²P-labeled proteins were alkali-stable, suggesting that they contained phosphotyrosine (results not shown). Autoradiographs were scanned to quantify c-src protein autophosphorylation and the ratio of densities between J2 and J1 of the 60 kDa band was 4.2 ±2.5 (mean ± s.d.) as measured by 3 different experiments.

We first investigated whether herbimycin A affected both the expression and the activity of pp60^c-src in trophoblast cells. Western blot analysis was performed as described above. Herbimycin treatment did not change the protein expression (Fig. 7A). Immunoprecipitations were done on equal amounts of total cellular protein from control cultures with DMSO and cells treated with 250 ng/ml, 500 ng/ml or 1 μg/ml for 24 hours at day 1 of culture and kinase assays were performed. Treatment with herbimycin reduced the autophosphoryation of src protein (Fig. 7B), which was visualized as a band running with an apparent molecular mass of 60 kDa as described previously, except that the exogenous substrate was omitted.

To determine whether herbymicin affects the expression of EGF-R and its tyrosine kinase activity, western blot analysis and phosphorylation of membrane protein were performed after treatment. Trophoblast cells were treated for 24 hours with 0.25, 0.5 and 1 μg/ml herbimycin. This treatment resulted in a marked reduction of EGF-R expression as assessed by immunoblotting and in the parallel decrease in EGF-dependent tyrosine kinase activity of membrane proteins (Fig. 8A,B).

In contrast, herbimycin treatment of trophoblast cells done under the same experimental conditions as those described above, resulted in an increase in E-cadherin expression as assessed by immunoblotting (Fig. 8C).

In the present study we tried to assess the possible involvement of pp60^c-src during in vitro syncytial formation. The present study is the first to follow the distribution of pp60^c-src in isolated cytotrophoblasts during their morphological differentiation into aggregates and syncytia. In a very few cell types, the cellular src gene product has easily detectable levels, i.e. in neurons (Maness et al., 1988) and blood platelets (Golden et al., 1989). In this study, we showed that in trophoblast cells c-src was largely expressed and that its cellular distribution varied with trophoblast differentiation. In both mononuclear and aggregated cells, src immunoreactivity was located in the plasma membrane, in structures resembling adhesion plaques; such a distribution disappeared in syncytia. In aggregated cells, pp60^c-src was found located at cell-to-cell contacts. In this context, it is noteworthy that c-src is concentrated in adherens junctions (Warren et al., 1988) and that several src family members have been copurified with adherens junctions (Tsukita et al., 1991).

In our experimental model of in vitro differentiation of syncytiotrophoblast cells, a balance seems to exist between pp60^c-src at cell-matrix junctions in isolated cells and its localization at cell-cell contacts in aggregated cells.

In isolated, aggregated and, to a lesser extent, syncytial cells, pp60^c-src was visualized by immunohistochemistry in the perinuclear region as previously described for other cells (Reish and Eriksson, 1985; Redmond et al., 1992). Several hypotheses may explain such a localization: an intermediate processing form of pp60^c-src in transit to the plasma membrane, or a physiological site of action for pp60^c-src differentiating activity. In addition, David-Pfeuty and Nouvian-Dooghe (1990) have presented evidence suggesting that pp60^c-src is associated with endocytotic vesicles.

Interestingly, pp60^c-src was colocalized with EGF-R in aggregated cells at day 2 of culture. In human placenta the EGF-R is found predominantly on syncytiotrophoblasts (Maruo et al., 1987), but it has also been identified in cultured cytotrophoblasts (DePalo and Das, 1988). Our data are in favor of a close relationship between the EGF-R and pp60^c-src proteins. In aggregated placental cells, immunohistochemical studies demonstrated that pp60^c-src was found associated with E-cadherin at cell-cell contacts. No cadherin was observed in multinucleated syncytial trophoblasts.

Despite the important changes observed in pp60^c-src localization during the aggregation and fusion of trophoblast cells, there were no changes in protein levels. However, activation of pp60^c-src kinase was observed at day 2 of culture, which corresponded to the time of cell aggregation; thereafter the levels of pp60^c-src activity remained elevated. As a consequence of pp60^c-src activation, different substrates could be phosphorylated, especially cytoskeletal proteins, leading to changes in cell-matrix interactions and cytoskeletal reorganization (Kellie et al., 1991). The localization of pp60^c-src to the plasma membrane places it in a position to associate with and transduce signals from transmembrane proteins.

The effect of inhibition of pp60^c-src kinase activity was investigated by the use of herbimycin A, which was shown to reverse the transformed characteristics of cells expressing the src oncogene (Murakami et al., 1988). We have demonstrated that herbimycin A treatment of trophoblast cells at day 1 of culture not only reduced src tyrosine kinase activity but also both EGF-R expression and EGF-R tyrosine kinase activity. EGF-R is considered to be implicated in the mechanisms of trophoblast cell differentiation; in particular, the aggregation of trophoblast cells is accompanied by an increase in the level of EGF-R expression concomitant with an increase in its EGF-dependent tyrosine kinase activity (Alsat et al., 1993). Addition of EGF in the cell culture results in a significant enhancement of the differentiation and transformation of cytotrophoblasts to syncytiotrophoblasts (Qu et al., 1992; Morrish et al., 1987). However, the role of EGF as an inducer of syncytiotrophoblast formation is debated and, according to Douglas and King (1990), EGF stimulates chorionic gonadotropin secretion without leading to cell fusion. The decrease in both expression and tyrosine kinase activity of EGF-R by herbimycin A suggests that it may delay syncytial formation; nevertheless, we cannot exclude a direct effect of herbimycin on EGF-R tyrosine kinase activity. Previous studies on the regulation of EGF receptor gene expression have revealed that activation of EGF-R by ligand binding causes an increase in the cellular level of EGF-R mRNA (Clark et al., 1985; DePalo and Das, 1988). This increase presumably occurs via tyrosine phosphorylation induced by the tyrosine kinase activity of the receptor. Our finding that herbimycin causes an inhibition of EGF-R gene expression points to a complex role for tyrosine phosphorylation in the regulation of EGF-R synthesis as previously described for the v-src protein (Wasilenko et al., 1990).

In contrast, herbimycin A treatment of trophoblast cells as described above induced an increase in E-cadherin expression at day 2. The effect of herbimycin on E-cadherin expression could be the complex result of inhibition of the tyrosine kinase activities of both c-src and EGF-R. As it was previously demonstrated (Coutifaris et al., 1991) and reconfirmed here that high levels of E-cadherin coincide with an increase in aggregation of cells, and low levels with syncytial formation, the increase in E-cadherin expression induced by herbimycin suggests that cell aggregation might be reinforced or syncytial formation delayed by herbimycin. As previously described by Volberg et al. (1992) we speculated that specific tyrosine phosphorylation of adherens type junction components is involved in the down-regulation of these cellular contacts.

In conclusion, the relocation of src protein at cell-to-cell contacts in aggregated cells was shown to coincide with an increase in pp60^c-src tyrosine kinase activity. Colocalization of pp60^c-src and EGF-R immunoreactivities points to a close interaction between the two proteins. Herbimycin A, a tyrosine kinase inhibitor, was shown to reduce both c-srcand EGF-dependent tyrosine kinase activities. In addition, inhibition of tyrosine kinase activities resulted in an increase in E-cadherin expression, which could be implicated in placenta cell aggregation and fusion. The cellular distribution and biochemical activity of pp60^c-src in trophoblast cells implies a role for this proto-oncogene product as a specific regulator of placental cell differentiation.

We thank Mina Bissel and Brigitte Boyer for critical reading of the manuscript and Dominique Morineau for help with photography. This work was supported by a grant from le Comité de Paris de la Ligue Nationale contre le Cancer and by an EC grant (Science JU-TL 910004).