Requirement of Caveolae Microdomains in Extracellular Signal-Regulated Kinase and Focal Adhesion Kinase Activation Induced by Endothelin-1 in Primary Astrocytes


  • Alexandra Teixeira,

  • Nathalie Chaverot,

  • Christian Schröder,

    1. Institute for Physiological Chemistry and Pathobiochemistry, Johannes Gutenberg University at Mainz, Mainz, Germany
    Search for more papers by this author
  • A. Donny Strosberg,

  • Pierre-Olivier Couraud,

  • Sylvie Cazaubon

  • Abbreviations used : ERK, extracellular signal-regulated kinase ; ET-1, endothelin-1 ; ETB-R, endothelin receptor subtype B ; FAK, focal adhesion kinase ; FITC, fluorescein isothiocyanate ; PBS, phosphate-buffered saline ; SDS, sodium dodecyl sulfate.

Address correspondence and reprint requests to Dr. S. Cazaubon at CNRS UPR 0415, Institut Cochin de Génétique Moléculaire, 22 Rue Méchain, 75014 Paris, France.


Abstract : Endothelin-1 (ET-1) mitogenic activity in astrocytes is mediated by the activation of the extracellular signal-regulated kinase (ERK) pathway together with the Rho-dependent activation of the focal adhesion kinase (FAK) pathway. To clarify the mechanisms responsible for the coordinate activation of both pathways in the ET-1 signal propagation, the involvement of caveolae microdomains, suggested to play a role in signal transduction, was evaluated. In this study, it is reported that caveolae of primary astrocytes are enriched in endothelin receptor (ETB-R). Furthermore, signaling molecules such as the adaptor proteins Shc and Grb2, and the small G protein Rho, also reside within these microdomains. Selective disassembly of caveolae by filipin III impairs the ET-1-induced tyrosine phosphorylation of proteins including ERK and FAK. In agreement with these observations, astrocytes pretreated with filipin III also failed to form stress fibers and focal adhesions and did not undergo the associated morphological changes in response to ET-1. This study reveals that structural integrity of caveolae is necessary for the adhesion-dependent mitogenic signals induced by ET-1 in astrocytes, through compartmentation of ETB-R with the upstream signaling molecules of the ERK and FAK pathways.

Endothelin-1 (ET-1) has been shown to be produced by the microvessel endothelial cells that constitute the blood-brain barrier (Durieu-Trautmann et al., 1993) and may act as a growth factor for the surrounding astrocytes, inducing DNA synthesis and proliferation (Couraud et al., 1991). Primary astrocyte responses to ET-1 are mediated by the ETB-receptor subtype (ETB-R) (Lazarini et al., 1996) that belongs to the superfamily of seven-transmembrane-domain receptors coupled to heterotrimeric G proteins. The up-regulation of ETB-R expression observed during astroglia differentiation, and the dramatic deficiencies in the development of neural crest-derived tissues observed in knockout mice for ET-1 or its receptors, suggest an important role for ET-1 during brain development (Baynash et al., 1994 ; Kurihara et al., 1994). In addition, both increased ET-1 production and astrocyte proliferation have been shown to occur during brain injury or inflammatory situations such as cerebral focal ischemia, subarachnoid hemorrhage, or human immunodeficiency virus type 1 infection (Ehrenreich et al., 1993 ; Yamashita et al., 1994).

In agreement with the mitogenic effect of ET-1, it has been previously shown that stimulation of ETB-R in primary astrocytes induces the activation of the extracellular signal-regulated protein kinase (ERK) (Lazarini et al., 1996). This activation requires the tyrosine phosphorylation of Shc and its subsequent association with Grb2, the adaptor protein for the Ras guanine nucleotide exchange factor SOS. The Raf-1 protein serine/threonine kinase is also involved in the ET-1 response, indicating that ETB-R is coupled to the Ras/Raf/ERK pathway (Cazaubon et al., 1994). Furthermore, ETB-R activation in primary astrocytes promotes stress fiber and focal adhesion formation associated with tyrosine phosphorylation of multiple proteins, such as focal adhesion kinase (FAK) and the cytoskeletal protein paxillin, elements of the so-called FAK pathway (Cazaubon et al., 1997). In contrast with the ERK pathway, the FAK pathway is dependent on activation of the small G protein Rho and on cell adhesion to the extracellular matrix. Nevertheless, coordinated activation of both ERK and FAK pathways is necessary for proliferation of ET-1-stimulated astrocytes (Cazaubon et al., 1997).

The recent observation that specialized microdomains, called caveolae, are rich in G proteins and other signaling molecules in most differentiated cell types suggested the involvement of these structures in signal transduction (Anderson, 1993 ; Lisanti et al., 1994). Caveolae are flask-shaped invaginations of the plasma membrane, rich in cholesterol, sphingolipids, and a 22-kDa-integrated membrane protein called caveolin (Rothberg et al., 1992 ; Anderson, 1993). It is interesting that tyrosine kinase activity has been observed in these microdomains, which may support the involvement of caveolae in transmembrane signaling events (Sargiacomo et al., 1994 ; P. Liu et al., 1997).

As regulated signal transduction in discrete microdomains of the cell surface could facilitate the coordinated activation of signaling events induced by ET-1, the contribution of caveolae in ERK and FAK pathway activation in astrocytes was evaluated.



Mouse monoclonal antibodies specific to phosphotyrosine (4G10) and to ERK-2 and rabbit polyclonal antibodies specific to FAK, Shc and Grb2 were purchased from UBI (Lake Placid, NY, U.S.A.). Rabbit polyclonal antibodies specific to Raf-1 and Gβ were from Santa Cruz Biotechnology (Santa Cruz, CA, U.S.A.). Peroxidase-conjugated anti-mouse or anti-rabbit IgG antibodies and enhanced chemiluminescence reagents were from Amersham (Les Ulis, France). Mouse monoclonal and rabbit polyclonal antibodies specific to caveolin or to paxillin were from Transduction Laboratory (Lexington, KY, U.S.A.). Mouse monoclonal antibody specific to vinculin, fluorescein isothiocyanate (FITC)-conjugated phalloidin, filipin, and polymyxin B were from Sigma (St. Louis, MO, U.S.A). Rabbit polyclonal anti-ezrin antibodies were kindly provided by Dr. P. Mangeat (U.S.T.L., Montpellier, France). Polyclonal antibodies specific to ETB-R were produced by immunizing female New Zealand white rabbits with the peptide CGLSRIWGEERGFPPDRATPLLQTAE, which corresponds to the sequence 19-44 of the human ETB-R.

Cell culture

Primary cultures of astrocytes were prepared as described previously (Lazarini et al., 1996). Striata and cortices were removed from brains of 17-day-old CD rat embryos and dissociated mechanically in serum-free medium. Cells were plated on poly-l-ornithine (1.5 μg/ml)-precoated dishes in Dulbecco's modified Eagle's medium containing 1 g/L glucose supplemented with 10% fetal calf serum and 10 mM HEPES, pH 7.4. Under these conditions, >95% of the cells stained positively by the immunofluorescence technique using antibodies specific for glial fibrillary acid protein (Amersham). Treatments were performed on 18-20-day-old cultures. Astrocytes were maintained in serum-free medium for 24-48 h before incubation with effectors.

Detergent-free purification of caveolae

Primary astrocytes (35 × 106 cells) grown to confluence were used to prepare caveolae fractions as described previously (Song et al., 1996). In brief, cells were washed in ice-cold phosphate-buffered saline (PBS) and scraped into 2 ml of 500 mM sodium carbonate, pH 11.0. Homogenization was performed by using a loose-fitting Dounce homogenizer (20 strokes) and a sonicator (three 20-s bursts). The homogenate was then adjusted to 40% sucrose by addition of 2 ml of 80% sucrose solution prepared in MBS buffer [25 mM 2-(N-morpholino)ethanesulfonic acid (MES), pH 6.5, 0.15 M NaCl] and placed in the bottom of an ultracentrifuge tube. A 5-30% discontinuous sucrose gradient was formed above, both in MBS containing 250 mM sodium carbonate, and centrifuged at 39,000 rpm (200,000 g) for 16-20 h in an SW41 rotor (Beckman Instruments, Palo Alto, CA, U.S.A). A light-scattering band confined to the 5-30% sucrose interface was observed that contained caveolin. From the top of each gradient, twelve 1-ml fractions were collected. Protein concentration of gradient fractions was determined by Bradford assay (Bio-Rad, Hercules, CA, U.S.A.).

Immunoprecipitation of FAK

Cells were pretreated with 5 μg/ml of filipin III for 10 min, where indicated. After treatment with 50 nM ET-1 for 10 min, cells (7 × 106) were lysed in MBST buffer [400 μl MBS buffer containing 1% Triton X-100, 1 mM orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 2 μg/ml aprotinin, 2 μg/ml pepstatin, and 2 μg/ml leupeptin]. Cell lysates were incubated with 1 μg of anti-FAK antibody for 15 h at 4°C, then with 10 μl of protein A-agarose for another 1 h. Immunoprecipitates were extensively washed in MBST buffer and resuspended in 125 mM Tris-HCl (pH 6.8), 4% sodium dodecyl sulfate (SDS), 5% glycerol, 50 mM dithiothreitol, 1 mM orthovanadate, and 0.05 μg/ml bromophenol blue (SDS sample buffer) and analyzed by immunoblotting.

Immunoblot analysis

Immunoprecipitated proteins, gradient fractions, and cell lysates were analyzed by immunoblotting, using the indicated antibodies, i.e., anti-caveolin monoclonal antibody (0.5 μg/ml), anti-Gβ subunits (1 μg/ml), anti-ETB-R (1:500), anti-ezrin (1:3,000), anti-Shc (1 μg/ml), anti-Grb2 (1 μg/ml), anti-Raf-1 (1 μg/ml), anti-ERK-2 (0.5 μg/ml), anti-RhoA (1 μg/ml), anti-FAK (1 μg/ml), anti-paxillin (0.25 μg/ml), and anti-phosphotyrosine antibodies (0.5 μg/ml). For serial incubations of membranes, bound antibodies were stripped out by 0.1 M glycine, pH 2.5, for 10 min, and the membrane was then incubated with different antibodies as described above.

Immunofluorescence analysis

Astrocytes of 18-20-day-old cultures were replated on poly-l-ornithine (1.5 μg/ml)-precoated coverslips. After 15 h in culture, cells were starved in serum-free medium for 24 h, pretreated with filipin III for 10 min (where indicated), and treated with 50 nM ET-1 for 10 min. After washes with PBS, the cells were fixed with paraformaldehyde (4%) in PBS for 15 min, protected with 0.1 M glycine for 15 min, and blocked with PBS containing bovine serum albumin (2%) and saponin (0.05%) for 1 h. The cells were incubated for 1 h with FITC-conjugated phalloidin (1:1,000) for F-actin labeling, with FITC-conjugated vinculin (1:400), or with polyclonal antibodies specific to caveolin (0.25 μg/ml). Indocarbocyanine-conjugated anti-rabbit antibodies (1:150) were used as secondary antibodies. Negative controls were obtained by omission of primary antibodies. Immunofluorescence images were collected in a scanner confocal microscope (MRC 1000, Bio-Rad).


ETB-R localizes in the caveolae of primary astrocytes

To explore the possibility that ET-1 signaling in astrocytes is caveolae dependent, we first looked for the presence of the ETB-R in these microdomains. For this purpose, we separated caveolin-enriched membrane fractions from the bulk of astrocyte membrane and cytosolic proteins using an established protocol (Song et al., 1996). This detergent-free method allows the enrichment of caveolae on a sucrose gradient based on the buoyantdensity properties of these structures. Equal volumes of each of the 12 collected fractions were submitted to immunoblot analysis. Using anti-caveolin antibodies, we identified caveolin, the major component of caveolae, in the different fractions of the sucrose gradient. As previously reported for several cell types, caveolin was almost exclusively recovered in gradient fractions 4 and 5 (Fig. 1A), here referred to as caveolae fractions, which contain ~2-5% of total cellular proteins (Fig. 1B). The majority of cellular proteins was recovered in fractions 9-12 (Fig. 1B), here referred to as caveolae-free fractions, which contain virtually no caveolin (Fig. 1A). Caveolae-free fractions contain structural proteins such as the cytoskeleton-associated protein ezrin, which is absent from caveolae fractions (Fig. 1A). It is interesting that ETB-R localized mainly in caveolae fractions (Fig. 1A). Stimulation of ETB-R induces the activation of heterotrimeric G proteins, which in other cellular systems have been found in caveolae. In agreement with these reports, we observed that caveolae fractions from primary astrocytes were enriched in the β subunits of G proteins (Fig. 1A). These data suggest that in astrocytes, the signal transduction induced by ET-1 through its specific G-coupled receptor, ETB-R, may occur at the level of caveolae microdomains.

Figure 1.

Isolation of caveolae microdomains from primary astrocytes. A : Colocalization of ETB-R and β subunits of heterotrimeric G proteins (Gβ) in caveolae microdomains. Quiescent astrocytes were submitted to homogenization and subcellular fractionation in the presence of sodium carbonate as described in Materials and Methods. Fractions (1 ml) were collected from the top of the gradient and equal volumes (20 μl) were separated by SDS-polyacrylamide gel electrophoresis and processed for immunoblot analyses. Membranes were incubated with anticaveolin, anti-Gβ, and anti-ETB-R antibodies. After stripping off the bound antibodies, the same membrane was incubated with anti-ezrin antibodies. Fraction numbers are indicated horizontally. B : Distribution of proteins along the sucrose density gradient. Protein concentration was determined in each fraction by Bradford assay (Bio-Rad) and expressed as a percentage of total protein determined. The results are representative of three independent experiments.

Fig. 1.

Localization of ET-1 signaling molecules in caveolae

ET-1 acts as a growth factor for astrocytes inducing DNA synthesis and cell proliferation (Couraud et al., 1991). Stimulation of ETB-R triggers the coordinate activation of ERK and FAK pathways, necessary for adhesion-dependent proliferation of astrocytes (Cazaubon et al., 1997). To clarify the functional importance of caveolae microdomains in ET-1 signaling, we have determined the presence of signaling proteins involved in these pathways in caveolae fractions (4 and 5) compared with the caveolae-free fractions (9-12). Analysis of equal protein amounts (10 μg) in each sample revealed the presence of the adapter proteins Shc and Grb2 in caveolae fractions as well as the protein kinase Raf-1. This protein kinase was detected in equivalent amounts in both fractions (Fig. 2A), whereas Shc and Grb2 were present in caveolae fractions in smaller amounts compared with caveolae-free fractions. In contrast, ERK-2 (the main ERK isoform expressed in primary astrocytes) was only detected in caveolae-free fractions (Fig. 2A). The distribution of FAK pathway components was determined under the same conditions. It is interesting that the small G protein Rho A, known to be involved in cytoskeletal rearrangements and focal adhesion assembly (Hotchin and Hall, 1995 ; Cazaubon et al., 1997), was concentrated in caveolae fractions, whereas the focal adhesion-associated proteins FAK and paxillin were absent from these fractions (Fig. 2B). These data indicate that the signaling molecules initiating either the ERK or FAK cascades are present in the caveolae microdomains of primary astrocytes, and may propagate the ET-1-induced signals to downstream elements in the cytosol (ERK) and at focal adhesions (FAK).

Figure 2.

Localization of ET-1-signaling pathway molecules in caveolae. Quiescent astrocytes were fractionated on sucrose gradient as described in Materials and Methods. The caveolae fractions (4 and 5) and the caveolae-free fractions (9-12) were pooled separately. Total proteins were concentrated by precipitation with 10% trichloroacetic acid and equal amounts of proteins (10 μg) were loaded on a SDS-polyacrylamide gel electrophoresis gel and submitted to immunoblot analysis, using the indicated antibodies. The results are representative of two independent experiments.

FIG. 2.

Filipin III pretreatment of astrocytes impairs ET-1-induced tyrosine phosphorylation

In primary astrocytes, ET-1 was previously reported to induce tyrosine phosphorylation of several cellular proteins, with a maximal response at 10 min of ET-1 treatment (50 nM) (Cazaubon et al., 1993). To evaluate the contribution of caveolae microdomains to the propagation of the signal induced by ET-1, astrocytes were pretreated with filipin III, a polyene antibiotic that binds to cholesterol and removes it from the membrane, causing reversible caveolae disassembly (Bradley et al., 1980 ; Schnitzer et al., 1994). The effect of filipin III on the cells was first evaluated by immunofluorescence analysis of caveolin distribution (see Figs. 5 and 6). Filipin III pretreatment of astrocytes (5 μg/ml, for 10 min) largely reduced the cell surface density of caveolin expression, strongly suggesting a decrease in the number of functional caveolae (see Fig. 5). Under these conditions, the level of tyrosine phosphorylation of astrocyte proteins after ET-1 stimulation was assessed by immunoblot analysis using anti-phosphotyrosine antibodies. ET-1 induced a substantial increase in the level of tyrosine phosphorylation of proteins with apparent molecular masses of 65-70 and 100-115 kDa, as previously shown (Cazaubon et al., 1993), whereas filipin III pretreatment strongly attenuated this response (Fig. 3). Pretreatment of astrocytes with polymyxin B, a drug that interacts with lipids in the cellular membranes but not with cholesterol (Bradley et al., 1980), did not affect the immunofluorescence pattern of caveolin staining (data not shown) or the ET-1-induced tyrosine phosphorylation (Fig. 3). Similar results were obtained in astrocytes stimulated with IRL1620, a selective agonist for ETB-R (data not shown), confirming that filipin III pretreatment impairs the ET-1 response mediated by ETB-R in astrocytes. Taken together, these results clearly suggest that disassembly of caveolae structures by filipin III affects the ET-1-induced tyrosine phosphorylation in astrocytes.

Figure 5.

Confocal analysis of caveolin distribution and F-actin-based cytoskeleton organization. Quiescent astrocytes (Co) or astrocytes treated with 50 nM ET-1 for 10 min (ET-1), submitted to filipin III pretreatment (+ filipin) or not (- filipin), were double-stained by using FITC-conjugated phalloidin (for F-actin staining) and anti-caveolin plus anti-rabbit antibodies conjugated to indocarbocyanine. The results are characteristic of several cells and representative of five independent experiments.

Figure 6.

Confocal analysis of caveolin distribution and focal adhesion formation. Quiescent astrocytes (Co) or astrocytes treated with 50 nM ET-1 for 10 min (ET-1), submitted to filipin III pretreatment (+ filipin) or not (-filipin), were double-stained using FITC-conjugated vinculin and anti-caveolin plus anti-rabbit antibodies conjugated to indocarbocyanine. The results are characteristic of several cells and representative of five independent experiments.

Figure 3.

ET-1-induced tyrosine phosphorylation is impaired by filipin III pretreatment. Quiescent astrocytes were pretreated (+) with filipin III (5 μg/ml) or polymyxin B (50 μg/ml) for 10 min or not submitted to pretreatment (-). After ET-1 stimulation (50 nM, 10 min) cell lysates were analyzed by immunoblotting with antiphosphotyrosine antibodies [anti-P(Y)]. After stripping off bound antibodies, the same membrane was reincubated with anti-ezrin antibodies, showing that similar amounts of proteins were loaded in all lanes. Molecular mass markers (kDa) are shown on the left side. The results are representative of four independent experiments.

FIG. 5.

FIG. 6.

FIG. 3.

Filipin III pretreatment of astrocytes prevents ERK-2 and FAK phosphorylation induced by ET-1

In previous studies it has been reported that ET-1 induces tyrosine phosphorylation of several proteins involved in ERK and FAK pathways (Cazaubon et al., 1994, 1997). To define the importance of caveolae microdomains in each of these signaling pathways, the activation of ERK and FAK was evaluated in astrocytes pretreated with filipin III. The activation of ERK-2 was analyzed by immunoblotting cell extracts with anti-ERK-2 antibody and visualized by a decrease in electrophoretic mobility, known to reflect the phosphorylated and activated state of this enzyme. As shown in Fig. 4A, ET-1 stimulated the phosphorylation of ERK-2 in a concentration-dependent manner ; i.e., a slight effect was observed with 1 nM ET-1 and was maximal at 50 nM ET-1. However, when astrocytes were pretreated with filipin III, ERK-2 phosphorylation was only detected after a stimulation with 5 nM ET-1 and was still largely impaired after stimulation with 50 nM ET-1. A complete inhibition of this kinase could be difficult to obtain because the ERK pathway is a cascade of several interacting signaling proteins, providing a great amplification of the signal. The contribution of caveolae in the ET-1-induced FAK-activation pathway was investigated by assessing the ability of ET-1, after filipin III pretreatment, to promote tyrosine phosphorylation of FAK, which was previously shown to correlate with its activation (Cazaubon et al., 1997). Immunoblot analysis with anti-phosphotyrosine antibodies of immunoprecipitated FAK revealed that the level of tyrosine phosphorylation was increased fivefold after ET-1 stimulation, and that filipin III pretreatment of astrocytes completely prevented this response. These data indicate that in astrocytes, the ET-1-induced activation of ERK and FAK pathways is dependent on the structural integrity of caveolae microdomains.

Figure 4.

ET-1-induced ERK-2 and FAK phosphorylation is impaired by filipin III pretreatment. A : Quiescent astrocytes were either not pretreated (control) or pretreated with filipin III (5 μg/ml) for 10 min (filipin), before addition of ET-1 (1, 5, and 50 nM) for 10 min. Cell lysates were submitted to immunoblotting with anti-ERK-2 antibodies. B : Cell lysates from quiescent astrocytes pretreated (+) or not (-) with filipin III (5 μg/ml) for 10 min before addition of ET-1 (50 nM) for 10 min were submitted to immunoprecipitation with anti-FAK antibodies. Immunoprecipitated proteins were analyzed by immunoblotting with anti-phosphotyrosine antibodies [anti-P(Y)]. After stripping off bound antibodies, the same membrane was reincubated with anti-FAK antibodies, showing that similar amounts of protein were immunoprecipitated. These results are representative of two independent experiments.

FIG. 4.

Filipin III pretreatment of astrocytes impairs stress fiber and focal adhesion formation induced by ET-1

As previously shown, morphological changes and cytoskeletal reorganization are associated with the Rho-dependent activation of FAK on ET-1 stimulation of astrocytes (Cazaubon et al., 1997). Because FAK activation is impaired by filipin III pretreatment, we tested the effect of filipin III in both caveolin expression and ET-1-induced cytoskeletal rearrangements. Confocal microscopy analysis revealed that caveolin is abundant in quiescent astrocytes and is distributed in a very organized manner, with a strong punctate signal at the level of the plasma membrane (Figs. 5 and 6). Similar results were obtained with astrocytes maintained in medium containing 10% serum, and no detectable variation on the levels of caveolin expression was observed during cell proliferation (data not shown). However, after filipin III pretreatment, caveolin staining is weakened and its pattern is much less organized. These changes are consistent with the disassembly of caveolae structures at the level of the plasma membrane. The organization of F-actin after filipin III pretreatment was assessed by using FITC-conjugated phalloidin, and no effect of this pretreatment on F-actin-based cytoskeleton was observed in nonstimulated astrocytes (Figs. 5 and 6). As previously shown (Cazaubon et al., 1997), stimulation of quiescent astrocytes with ET-1 caused a marked increase in stress fiber formation, which is accompanied by the characteristic morphological changes (Fig. 5). It is interesting that filipin III-pretreated cells failed to respond to ET-1 ; i.e., under these conditions, stress fibers were not assembled and cell morphology remained reminiscent of that of quiescent astrocytes with long processes (Fig. 5). Immunostaining against vinculin, a marker for focal adhesions, showed that filipin III also prevented the ET-1-induced formation of focal adhesions at the level of the stress fiber/plasma membrane contacts (Fig. 6). Consistent with the inhibitory effect of filipin III on the ET-1-induced tyrosine phosphorylation of cellular proteins (see Fig. 3), including FAK (see Fig. 4B), the level of tyrosine phosphorylation at focal adhesion sites was also largely impaired by filipin III pretreatment (data not shown).

These results clearly suggest that caveolae are actively involved in cytoskeletal reorganization, leading to stress fiber and focal adhesion formation, both events necessary for ET-1-induced FAK activation.


The aim of the present study was to evaluate the contribution of caveolae in the signaling pathways induced by ET-1 in primary astrocytes. It is noteworthy that we observed that ETB-R, the ET-1 receptor subtype expressed in astrocytes, is localized almost exclusively in caveolae microdomains and that disassembly of these microdomains impairs the astrocytic response to ET-1. Both ERK and FAK signaling pathways, necessary for the transmission of the ET-1 mitogenic effect, as well as the tyrosine phosphorylation of endogenous cellular proteins, were largely affected. Moreover, the marked increase of stress fiber and focal adhesion formation associated with modifications of cell morphology observed in ET-1-stimulated astrocytes was also found to require caveolae structural integrity. In primary astrocytes, caveolae may therefore constitute a membrane compartment of signalization, necessary for the coordinated activation of the intracellular pathways involved in ET-1-induced proliferation.

Caveolae are small flask-shaped microdomains of the plasma membrane that are present to some degree in most of differentiated cell types and are particularly abundant in endothelial cells, adipocytes, fibroblasts, and smooth muscle cells (Parton, 1996). Until recently, it was proposed that neural cells did not possess such compartments, as mRNA for caveolin, the principal structural component of caveolae, was not detected in brain (Lisanti et al., 1994 ; Olive et al., 1995 ; Bouillot et al., 1996 ; Scherer et al., 1996). Nevertheless, it should be noted that plasma membrane invaginations that exhibit “caveolae-like” properties have been observed in neurons (Wu et al., 1997). Moreover, a recent report has provided strong evidences that astrocytes contain morphologically identifiable caveolae expressing caveolin (Cameron et al., 1997). Our study confirms that astrocytes are very rich in caveolae and further highlights the functions of these microdomains in the propagation of the ET-1 signal in astrocytes.

Initially identified as a compartment of transcytosis in endothelial cells, caveolae appear to possess a large number of functions (Okamoto et al., 1998). In nonneural cells, caveolae have been proposed to participate in transmembrane signal transduction, because they may compartmentalize several classes of signaling molecules, such as G proteins and protein kinases. The observation that caveolin-rich membrane domains are enriched in β subunits of heterotrimeric G proteins in astrocytes supports a role for caveolae in the signal transduction associated with G protein-coupled receptors in these cells. We have determined that ETB-R, the subtype of ET-1 receptor present in astrocytes, is mainly expressed in caveolae. Our results are consistent with a previous report showing the localization of ETA receptor, the other mammalian ET-1 receptor subtype, in caveolae of transfected COS cells (Chun et al., 1994), and extend these observations to endogenously expressed ET-1 receptors in CNS cells. Due to the localization of ETB-R in caveolae, one may predict that intracellular signalization induced by ET-1 should be largely initiated at the level of these compartments. We have previously determined that ET-1-induced proliferation in primary astrocytes is mediated by the concomitant activation of the ERK and FAK pathways via heterotrimeric G proteins (Cazaubon et al., 1997). Several intermediates of the ERK pathway, such as the adaptor proteins Shc and Grb2 and the protein kinase Raf-1, were found to be present, although not enriched, in the caveolae microdomains of astrocytes. These observations suggest that other intracellular pools of signaling molecules may exist and that post-translational modifications would explain their intracellular localizations. Recent studies propose that both myristoylation and palmitoylation of the proteins are important for their correct localization in caveolae (Okamoto et al., 1998). It should be noted that ERK-2 was only detected in noncaveolae fractions, in contrast with a recent study performed in human fibroblasts where ERK-2 was found to be concentrated in caveolae (P. Liu et al., 1997). These observations indicate that, in terms of caveolae composition, some differences may exist depending on the cell type analyzed. Whereas the presence of signaling molecules of the ERK pathway has been described in nonneural cell types (P. Liu et al., 1996, 1997 ; Mineo et al., 1996), nothing was known about the intermediaries of the FAK pathway. In primary astrocytes, ET-1-induced FAK activation involved a Rho-dependent mechanism that is independent of ERK activation. In nonneural cells, activation of Rho A induced by growth factors has been shown to play an important role in the assembly of F-actin filaments, promoting stress fiber and focal adhesion formation (Craig and Johnson, 1996). It is interesting that caveolae from astrocytes contain the small G protein Rho, but not FAK or paxillin, both proteins known to accumulate at focal adhesions. Altogether, these data indicate that the upstream intermediaries of ERK as well as FAK pathways colocalize with ETB-R in caveolae, providing a compartmental basis for their coordinated activation on ET-1 stimulation of astrocytes.

Further evidence for the involvement of caveolae in the propagation of the signal induced by ET-1 was obtained by using sterol binding agents such as filipin III, which disrupts caveolae microdomains by removing cholesterol from the membrane (Schnitzer et al., 1994). Action of filipin III on caveolae structures has been well documented in several cell systems (Smart et al., 1996 ; Orlandi and Fishman, 1998). In endothelial cells, disassembly of the caveolar compartment by filipin III has been shown to specifically inhibit caveolae-mediated transcellular transport (Schnitzer et al., 1994) and to affect the platelet-derived growth factor-induced signaling cascade (J. Liu et al., 1997). In this study, we show that filipin III pretreatment of astrocytes significantly reduces the amount of caveolin present at the plasma membrane and, under these conditions, the ET-1-induced tyrosine phosphorylation of cellular proteins is inhibited. These effects appear to be specific, as polymyxin B, another agent known to interact with lipids but not with cholesterol in cell membranes (Bradley et al., 1980 ; Schnitzer et al., 1994), was ineffective. Moreover, we report for the first time that filipin III pretreatment of cells prevents the activation of both ERK and FAK, the key enzymes in the ET-1-signaling pathways leading to the proliferation of astrocytes. In a previous study we have determined that ET-1-induced FAK activation is associated with stress fiber and focal adhesion formation (Cazaubon et al., 1997). Consistent with the consequence of caveolae disruption on FAK activation, the filipinpretreated astrocytes are unable to undergo the morphological changes associated with ET-1-induced cytoskeletal rearrangements. Under these conditions, focal adhesion formation is also impaired. FAK inhibition has been shown to be sufficient to block DNA synthesis (Cazaubon et al., 1997), which strongly suggested that cell proliferation would not occur in filipin III-pretreated cells. Altogether, our results indicate that caveolae may play a fundamental role in the regulation of normal adhesion-dependent proliferation of astrocytes. Consistent with these findings, it has been observed that many transformed cell lines lack caveolin and caveolae, and that recombinant expression of caveolin in oncogenically transformed 3T3 cells abrogated adhesion-independent growth (Koleske et al., 1995 ; Engelman et al., 1997).

In conclusion, we have demonstrated that caveolae structural integrity is necessary for the coordinated activation of ERK and FAK pathways, both intracellular events known to regulate, in concert, the adhesion-dependent proliferation of astrocytes on stimulation with ET-1. Moreover, our study suggests that caveolae may play an important role in the regulation of normal cell proliferation.