Caveolin-1 Expression Inhibits Wnt/β-Catenin/Lef-1 Signaling by Recruiting β-Catenin to Caveolae Membrane Domains*

Caveolin-1 is a principal component of caveolae membranes. In NIH 3T3 cells, caveolin-1 expression is dramatically up-regulated in confluent cells and localizes at areas of cell-cell contact. However, it remains unknown whether caveolin-1 is involved in cell-cell signaling. Here, we examine the potential role of caveolin-1 in regulating β-catenin signaling. β-Catenin plays a dual role in the cell, linking E-cadherin to the actin cytoskeleton and in Wnt signaling by forming a complex with members of the lymphoid enhancing factor (Lef-1) family of transcription factors. We show that E-cadherin, β-catenin, and γ-catenin (plakoglobin) are all concentrated in caveolae membranes. Moreover, we demonstrate that activation of β-catenin/Lef-1 signaling by Wnt-1 or by overexpression of β-catenin itself is inhibited by caveolin-1 expression. We also show that recombinant expression of caveolin-1 in caveolin-1 negative cells is sufficient to recruit β-catenin to caveolae membranes, thereby blocking β-catenin-mediated transactivation. These results suggest that caveolin-1 expression can modulate Wnt/β-catenin/Lef-1 signaling by regulating the intracellular localization of β-catenin.

It has been proposed that caveolin family members (caveolin-1, caveolin-2, and caveolin-3) function as scaffolding proteins (11) to concentrate and organize specific lipids (cholesterol and sphingolipids (12)(13)(14)) and lipid-modified signaling molecules within caveolae membranes. These molecules include G-protein-coupled receptors, heterotrimeric G-proteins, receptor tyrosine kinases, components of the Ras-mitogen-activated protein kinase pathway, Src family tyrosine kinases, protein kinase C isoforms, and nitric-oxide synthase (14 -19). In support of this idea, caveolin-1 was found to suppress the kinase activity of Src family tyrosine kinases (c-Src/Fyn), epidermal growth factor receptor, Neu, and protein kinase C through the caveolin-scaffolding domain, a modular protein domain that recognizes a specific caveolin binding motif in various signaling molecules (3,4,20,21). In addition, the scaffolding domain of caveolin-1 was shown to inhibit endothelial nitric-oxide synthase activity and the GTPase activity of heterotrimeric G-proteins (15,16,19,22,23).
The constitutive activation of these signaling molecules can often cause cellular transformation. Because the interaction of caveolin-1 with signaling molecules leads to their inactivation (4), it has been proposed that caveolin-1 may act as a tumor suppressor protein (24 -26). Consistent with this hypothesis, it has been shown that the level of caveolin-1 mRNA and protein expression and caveolae organelles are either lost or reduced during cell transformation by activated oncogene products such as v-Abl and Ha-Ras (G12V) (24). Induction of caveolin-1 expression in v-Abl-and Ha-Ras (G12V)-transformed NIH 3T3 cells abrogated the anchorage-independent growth of these cells in soft agar and resulted in the de novo formation of caveolae (25). More recently, we have shown that an antisensemediated reductions in caveolin-1 protein expression in NIH 3T3 cells is sufficient to drive oncogenic transformation and constitutively activate the p42/44 mitogen-activated protein kinase cascade (26). Finally, the human caveolin-1 gene is localized to a suspected tumor suppressor locus (D7S522; 7q31.1), a known fragile site (FRA7G) that is deleted in many types of cancer (27)(28)(29)(30). Thus, down-regulation of caveolin-1 expression and caveolae organelles may be critical for maintaining the transformed phenotype.
In nontransformed NIH 3T3 cells, we have previously shown that caveolin-1 levels are down-regulated in rapidly dividing cells and dramatically elevated in confluent cells (26,31). Interestingly, at low cell density when no cell-cell contacts are made, caveolin-1 is localized in a punctate distribution over the entire cell surface and within the interior of the cell. When cells start to establish cell-cell contacts, caveolin-1 redistributes and localizes to areas of cell-cell contact. These observations may be related to the ability of caveolin-1 to regulate contact inhibition and growth arrest in nontransformed cells (26,31) and raise the possibility that caveolin-1 is involved in regulating cell adhesion-mediated processes.
In addition to a structural role in mediating cell-cell adhesion, ␤-catenin is part of the Wnt signaling pathway. Activation of the Wnt pathway results in the accumulation and nuclear translocation of ␤-catenin (41)(42)(43), where in complex with transcription factors of the Lef-1 family it regulates the expression of specific target genes (44 -47). The level of ␤-catenin in cells available for signaling is regulated by its interaction with a multiprotein complex including the tumor suppressor adenomatous polyposis coli protein (48,49), glycogen synthase kinase 3␤ (GSK-3␤ (50)), 1 and axin (51,52) and involves the phosphorylation of ␤-catenin (53, 54) and its targeting for degradation via the ubiquitin-proteasome system (42,55).
Elevated ␤-catenin levels, resulting from reduced turnover, are associated with cell transformation. Mutations in the adenomatous polyposis coli gene or the ␤-catenin gene itself thus may lead to the accumulation of ␤-catenin in human colon carcinoma (56,57) and melanoma (58 -61) and other types of cancer (62). Elevation of ␤-catenin expression in such tumors is believed to induce uncontrolled activation of gene transcription by the ␤-catenin-Lef-1 complex thereby contributing to tumor progression (59,62,63).
Here, we have addressed the possible association of the E-cadherin-␤-catenin complex with caveolae and caveolin-1. We show that ␤-catenin can be sequestered away from GSK-3␤ by associating with caveolae and caveolin-1. We also demonstrate that such interactions can affect Wnt signaling/␤-catenin-mediated transactivation.
Establishment of Stable U251 Cell Lines Overexpressing Caveolin-1-The full-length untagged cDNA encoding murine caveolin-1 (71) was inserted into an expression vector driven by the ␤-actin promoter and cytomegalovirus enhancer (pCAGGS, gift of Dr. Armin Rehm, Ploegh Laboratory, Harvard Medical School, Cambridge, MA). U251 cells were co-transfected with the caveolin-1 vector (pCAGGS-Cav-1) and with a plasmid containing hygromycin resistance (pCB7) using a modified calcium phosphate precipitation protocol. U251 cells were also transfected with the empty vector pCAGGS as a critical control. Resistant clones were selected using hygromycin B (200 g/ml). Individual clones were isolated using cloning rings. Lysates from stably transfected U251 cells were prepared and assayed for caveolin-1 expression by immunoblot analysis.
Preparation of Caveolae-enriched Membrane Fractions-Cells were scraped into 2 ml of MES-buffered saline containing 1% (v/v) Triton X-100. Homogenization was carried out with 10 strokes of a loose fitting Dounce homogenizer. The homogenate was adjusted to 40% sucrose by the addition of 2 ml of 80% sucrose prepared in MES-buffered saline and placed at the bottom of an ultracentrifuge tube. A 5-30% linear sucrose gradient was formed above the homogenate and centrifuged at 39,000 rpm for 16 -20 h in a SW41 rotor (Beckman Instruments). A light-scattering band confined to the 15-20% sucrose region was observed that contained endogenous caveolin-1, but excluded most of other cellular proteins. From the top of each gradient, 1-ml gradient fractions were collected to yield a total of 12 fractions. Equal amounts of protein from each gradient fraction were separated by SDS-PAGE and subjected to immunoblot analysis.
Immunofluorescence Microscopy-Cells grown on glass coverslips were washed three times with PBS and fixed for 30 min at room temperature with 2% paraformaldehyde in PBS. Fixed cells were rinsed with PBS and permeabilized with 0.1% Triton X-100, 0.2% bovine serum albumin for 10 min. The samples were then treated with 25 mM NH 4 Cl in PBS for 10 min at room temperature to quench free aldehyde groups. The cells were rinsed with PBS and incubated with the primary antibody for 1 h at room temperature (diluted 1:1000 in PBS with 0.1% Triton X-100, 0.2% bovine serum albumin). After three washes with PBS (10 min each), cells were incubated with the secondary antibody for 1 h at room temperature: lissamine rhodamine B sulfonyl chlorideconjugated goat anti-rabbit antibody (5 g/ml)/fluorescein isothiocyanate-conjugated goat anti-mouse antibody (5 g/ml). Finally, cells were washed three times with PBS (10 min each wash), and slides were mounted with slow-Fade anti-fade reagent (Molecular Probes, Inc., Eugene, OR) and observed under a Bio-Rad MR 600 confocal microscope.
Lef-Luciferase Reporter Assay-Cells were seeded in 6-well plates at 300,000 cells/well. The following day, cells were transiently transfected with 1 g of the Lef-1 reporter (pTOP-FLASH), 0.5 g of ␤-galactosidase and 1 g of the indicated cDNA. Twelve hours post-transfection, cells were rinsed twice with PBS and incubated in 1% serum for another 24 -36 h. Cells were then lysed in 300 l of extraction buffer; 80 l was used to measure luciferase activity (72) and 50 l of which was used to measure ␤-galactosidase activity. The Lef-1 luciferase reporter activity was controlled for transfection efficiency and potential toxicity of treatments using ␤-galactosidase activity. The specificity of the effects on the Lef-1 reporter was confirmed using pFOP-FLASH, which contains mutated Lef-1 binding sites (44) and an unrelated activator protein-1 reporter (73).

E-Cadherin, ␤-Catenin, and ␥-Catenin (Plakoglobin) Are Concentrated in Caveolae Membranes and Form a Stable Complex with
Caveolin-1-To evaluate whether caveolin-1 is involved in mediating E-cadherin/␤-catenin signaling, we examined the localization of E-cadherin, ␤-catenin, and ␥-catenin (plakoglobin) in MDCK cells by laser-scanning confocal microscopy. Fig. 1 demonstrates that caveolin-1 co-localizes with Ecadherin, ␤-catenin, and ␥-catenin at the basolateral surface of MDCK cells and is predominantly excluded from the free apical plasma membrane. These results are consistent with our previous data showing that caveolin-1 is recruited from an intracellular location to areas of cell-cell contact as NIH 3T3 cells reach confluence (26).
To further investigate the co-localization of caveolin-1 with cadherin-catenin complexes, we took advantage of the known biophysical properties of caveolae domains. Caveolae membranes are resistant to extraction at 4°C by nonionic detergents such as Triton X-100 and float on bottom-loaded sucrose density gradients because of their high content of cholesterol and sphingolipids. We used a well established procedure based on the detergent resistance and low buoyant density of caveolae to separate caveolae-enriched membrane fractions from the bulk of cellular membranes and cytosolic proteins (15, 64, 74 -82). The distribution of E-cadherin, ␤-catenin, and ␥-catenin was followed by Western blot analysis. Caveolin-1, a principal structural protein of caveolar membranes, was used to track the position of caveolae-derived membranes in this fractionation scheme.
Consistent with previous reports, we found that caveolin-1 immunoreactivity is exclusively detected in the caveolae fractions (fractions 4 and 5; at a density of ϳ10 -20% sucrose) (Fig.  2). The results presented in Fig. 2 show that E-cadherin, ␤-catenin, and ␥-catenin all exhibit a distribution pattern that parallels that of caveolin-1, suggesting that these cadherincatenin complexes are enriched in caveolae membranes.
To determine whether caveolin-1 forms a stable complex with these molecules, extracts from MDCK cells were immunoprecipitated with an antibody directed against the N-terminal domain of caveolin-1 (pAb N-20, residues 2-21). The immunoprecipitates were separated by SDS-PAGE and subjected to immunoblot analysis with monoclonal antibodies directed against E-cadherin, ␤-catenin, and ␥-catenin. The results in Fig. 3A show that E-cadherin, ␤-catenin, and ␥-catenin all co-immunoprecipitate with antibodies directed against caveolin-1. Protein A-Sepharose beads alone (Fig. 3A) or an irrelevant IgG (not shown) showed no such co-immunoprecipitation. Conversely, caveolin-1 co-immunoprecipitates with antibodies directed against E-cadherin, ␤-catenin, and ␥-catenin (Fig. 3B).
Transient Expression of Caveolin-1 Blocks Wnt-1 Signaling-Wnt-1 signaling in mammalian cells results in the inac- tivation of GSK-3␤, a serine-threonine kinase that phosphorylates ␤-catenin and promotes its degradation. As a consequence of GSK-3␤ inactivation, ␤-catenin accumulates and becomes available for binding to transcription factors of the Lef family and activates transcription (44 -47).
To explore the possibility that caveolin-1 may function as a regulator of Wnt-1-mediated signal transduction involving the ␤-catenin-Lef-1 complex, Wnt-1 was transiently co-expressed in NIH 3T3 cells with a luciferase reporter plasmid containing a promoter specifically responsive to Lef-1 (TOP-FLASH). The results in Fig. 4A show that transfection with Wnt-1 activated this reporter plasmid, and co-expression of caveolin-1 dramatically inhibited Wnt-1-mediated activation of the reporter by 8 -9-fold. The empty vector used to express caveolin-1 (pCAGGS) had no inhibitory effect.
To further study the inhibitory effect of caveolin-1 on Wnt-1 signaling, we next employed an established mammary epithelial cell line, RAC311, stably transduced with the retrovirus MV7-Wnt-1 (Rac/Wnt-1) or with the empty vector MV7 (Rac/ MV7). Transient transfection of these cell lines with TOP-FLASH showed significant Lef-1-responsive transcription that was inhibited 7-8-fold by co-transfection with caveolin-1 (Fig.  4B). The empty vector used to express caveolin-1(pCAGGS) had no inhibitory effect. The specificity of the effects on the Lef-1 reporter pTOP-FLASH were confirmed using pFOP-FLASH, which contains mutated Lef-1 binding sites (44) (Fig. 4, A  and B).
To more directly evaluate whether caveolin-1 can inhibit ␤-catenin-mediated activation of transcription, we transiently transfected NIH 3T3 cells with wild-type ␤-catenin (␤-cat WT) alone or in combination with caveolin-1 and the reporter plasmid. The results presented in Fig. 5A show that ␤-cat WT Note that E-cadherin, ␤-catenin, and ␥-catenin all co-immunoprecipitate with caveolin-1. B, detergent extracts of MDCK cells were immunoprecipitated with mAbs directed against ␤-catenin, ␥-catenin, and E-cadherin bound to protein A-Sepharose beads or with beads alone. Immunoprecipitates were then subjected to SDS-PAGE and Western blotting with anti-caveolin-1 IgG (pAb N-20). Note that caveolin-1 co-immunoprecipitates with E-cadherin, ␤-catenin, and ␥-catenin.

FIG. 4. Recombinant expression of caveolin-1 blocks induction of Lef-directed transcription by Wnt-1.
A, NIH 3T3 cells. Wnt-1 and the luciferase reporter plasmid (pTOP-FLASH) were transiently transfected into NIH 3T3 cells and luciferase activity was determined. Note that co-expression with caveolin-1 inhibited the reporter activity by 8 -9-fold, whereas the empty vector used to express caveolin-1 (pCAGGS) had no inhibitory effect. B, Rac cells. Caveolin-1 and pTOP-FLASH were used to transiently co-transfect RAC311 cells stably expressing Wnt-1 (Rac/Wnt-1) or the empty vector MV7 (Rac/MV7). Note that Rac/Wnt-1 cells exhibit a significantly higher activation of the reporter compared with Rac/MV7 cells because of the stable expression of Wnt-1. Note that caveolin-1 expression blocks Wnt-1-mediated activation of the reporter by 7-8-fold. The empty vector used to express caveolin-1 (pCAGGS) had no inhibitory effect. In A and B, the specificity of the effects on the Lef-1 reporter pTOP-FLASH was confirmed using pFOP-FLASH, which contains mutated Lef-1 binding sites (44). activated the Lef-1-responsive reporter and co-expression with caveolin-1 resulted in ϳ8-fold inhibition of this activation. Thus, caveolin-1 could inhibit both Wnt-mediated and ␤-catenin-controlled activation of Lef-1-dependent transcription (Figs. 4 and 5).
We next evaluated the ability of caveolin-1 to block transactivation induced by a mutant form of ␤-catenin (␤-cat S33Y). This serine to tyrosine mutation at position 33 serves to stabilize ␤-catenin, enhancing its capacity to activate transcription. As expected, mutant ␤-catenin (S33Y) was more potent than wild-type ␤-catenin in activating the reporter (Fig. 5A). Co-expression of caveolin-1 with mutant ␤-catenin (␤-cat S33Y) inhibited its activity by 7-8-fold (Fig. 5A). These results clearly indicate that caveolin-1 expression can very potently inhibit Lef-1-driven transcription induced by either wild type or an activated form of ␤-catenin.

FIG. 6. Stable expression of caveolin-1 in caveolin-1 negative U251 cells recruits ␤-catenin to caveolae membranes. A, Western blot analysis of parental U251 cells (U251) and cells stably transfected
with caveolin-1 (U251-Cav-1), using the vector pCAGGS. B, caveolae membranes were separated from the bulk of cellular membranes and cytosolic proteins by equilibrium sucrose density gradient centrifugation (see "Experimental Procedures"). In this fractionation scheme, immunoblotting with anti-caveolin-1 IgG can be used to track the position of caveolae-derived membranes within these bottom-loaded sucrose gradient. Fractions 1-4 (lanes 1-4) are the 5% sucrose layer, and fractions 5-8 (lanes 5-8) are the 30% sucrose layer. Fractions 9 -12 (lanes 9 -12), containing 40% sucrose, represent the "loading zone" of these bottom-loaded flotation gradients, and contain the bulk of the cellular membrane and cytosolic proteins. Note that in caveolin-1 negative U251 cells harboring vector alone (pCAGGS), ␤-catenin is restricted to the bottom of the gradient (loading zone; fractions 9 -12). Identical results were obtained with parental U251 cells (not shown). In contrast, recombinant expression of caveolin-1 in U251 cells (pCAGGS-Cav-1) results in a dramatic shift of ␤-catenin to caveolae-enriched membranes (fractions 4 and 5). Each lane contains an equal amount of total protein.

Stable Expression of Caveolin-1 in the U251 Caveolin-1 Negative Cell Line Recruits ␤-Catenin to Caveolae Membranes and
Inhibits ␤-Catenin-mediated Transactivation-To assess the effects of caveolin-1 expression on the localization of ␤-catenin, we stably expressed caveolin-1 in a caveolin-1 negative cell line (U251 cells, Fig. 6A). Three independent clones expressing caveolin-1 were generated and similar results were obtained with all three clones. Figs. 6 and 7 show results obtained with one representative caveolin-1 positive clone (pCAGGS-Cav-1). As an additional control for these experiments, we also derived caveolin-1 negative U251 cells harboring the vector alone (pCAGGS).
To separate membranes enriched in caveolae from the bulk of cellular membranes and cytosolic proteins, an established equilibrium sucrose density gradient system was utilized (74). We and others have shown that these caveolae-enriched fractions exclude markers for noncaveolar plasma membrane, Golgi, lysosomes, mitochondria, and the endoplasmic reticulum (15, 64, 74 -82).
The results presented in Fig. 6B illustrate that in caveolin negative U251 cells, endogenous ␤-catenin is restricted to the loading zone of these sucrose density gradients (fractions 9 -12). In contrast, in caveolin-1 positive U251 cells, endogenous ␤-catenin was detected mainly in the caveolae-enriched membrane fractions (fractions 4 and 5). These results demonstrate that expression of caveolin-1 in U251 cells is sufficient to recruit ␤-catenin to caveolae membranes.
To examine whether caveolin-1 expression in U251 cells can effectively block ␤-catenin-mediated transcription, we transiently transfected the reporter plasmid TOP-FLASH into caveolin-1 negative and caveolin-1 positive U251 cells. The results in Fig. 7 show that Lef-1-responsive transcription is inhibited in caveolin-1 expressing U251 cells by about 3-fold, as compared with caveolin-1-deficient U251 cells. These results suggest that caveolin-1-mediated recruitment of ␤-catenin to caveolae membranes is sufficient to inhibit ␤-catenin/Lef-1directed transcription.
GSK-3␤ Is Excluded from Caveolae Membranes and Caveolin-1 Expression Blocks the Interaction between ␤-Catenin and GSK-3␤-Glycogen synthase kinase-3␤ is a key protein in the regulation of ␤-catenin turnover. Phosphorylation of ␤-catenin by GSK-3␤ promotes its degradation by the ubiquitin-protea-some system (42,55). We employed the sucrose density gradient system described above to examine if GSK-3␤ is localized to caveolae membranes or is excluded from these caveolin-rich fractions in NIH 3T3 cells. The results in Fig. 8 show that although ␤-catenin is concentrated in the caveolar fractions (fractions 4 and 5), GSK-3␤ is selectively excluded from caveo-   9. Overexpression of caveolin-1 in NIH 3T3 cells prevents the interaction between ␤-catenin and GSK-3␤. NIH 3T3 cells were transiently co-transfected with ␤-catenin and GSK-3␤, in combination with caveolin-1 or pCAGGS (the empty vector used to drive caveolin-1 expression). A, cell lysates were immunoprecipitated (IP) with anti-␤-catenin IgG, separated by SDS-PAGE, and subjected to immunoblot analysis with anti-GSK-3␤ IgG. Note that ␤-catenin does not co-immunoprecipitate with GSK-3␤ when co-expressed with caveolin-1. B, Western blot analysis of the cell extracts prior to immunoprecipitation. Note that equivalent levels of ␤-catenin are expressed with or without caveolin-1, whereas the level of GSK-3␤ was only slightly reduced by caveolin-1 expression. WB, Western blot. lae membranes and is restricted to the bottom of the gradient (loading zone, fractions 9 -12).
Because GSK-3␤ is excluded from caveolae membranes and caveolin-1 expression recruits ␤-catenin to caveolae, we asked whether the expression of caveolin-1 can inhibit the interaction of ␤-catenin with GSK-3␤. NIH 3T3 cells were transiently transfected with ␤-catenin and GSK-3␤ in combination with caveolin-1 or pCAGGS (the empty vector used to drive caveolin-1 expression). Cell lysates were immunoprecipitated with anti-␤-catenin IgG, separated by SDS-PAGE, and subjected to immunoblot analysis with anti-GSK-3␤ IgG. When co-expressed with caveolin-1, ␤-catenin did not co-immunoprecipitate with GSK-3␤ (Fig. 9A), supporting the hypothesis that caveolin-1 expression recruits ␤-catenin to caveolae, thereby reducing its interaction with GSK-3␤. The results presented in Fig. 9B show that the levels of ␤-catenin are not significantly affected by caveolin-1 expression, whereas the level of GSK-3␤ is only slightly reduced by caveolin-1 expression.
Expression of Caveolin-1 and ␤-Catenin Is Coordinately Elevated by Confluence or Serum Starvation and Is Reduced by Growth Factor Stimulation or Cell Transformation-The data presented above indicate that expression of caveolin-1 recruits ␤-catenin to caveolae membranes thereby reducing its ability to interact with GSK-3␤. As GSK-3␤ induces degradation of ␤-catenin, we would predict that an elevation in caveolin-1 expression will result in the stabilization of ␤-catenin, whereas a reduction in caveolin-1 expression will induce a decrease in ␤-catenin. Thus, caveolin-1 and ␤-catenin levels should increase or decrease in parallel in response to a variety of cellular stimuli.
In a previous study, we demonstrated that caveolin-1 expression is elevated in confluent NIH 3T3 cells and down-regulated in rapidly dividing NIH 3T3 cells (26). The results presented in Fig. 10A show that the expression of both ␤-catenin and caveolin-1 protein increases to high levels in confluent cells and that both are down-regulated in sparse, rapidly dividing NIH 3T3 cells.
We have also previously shown that caveolin-1 expression is elevated in serum-deprived NIH 3T3 cells and is reduced by growth factor stimulation of these cells (26). The results in Fig.  10B illustrate that both caveolin-1 and ␤-catenin expression are elevated in serum-deprived NIH 3T3 cells. Conversely, when serum-starved NIH 3T3 cells were incubated with platelet-derived growth factor or basic fibroblast growth factor (either alone or in combination), caveolin-1 and ␤-catenin levels are dramatically reduced (Fig. 10B).
Caveolin-1 mRNA and protein expression are significantly reduced during cell transformation by activated oncogene products, such as Ha-Ras (G12V) and v-Abl (24). We therefore examined whether ␤-catenin expression is affected in these transformed cell lines. We found that, as expected, caveolin-1 expression was dramatically reduced in Ha-Ras (G12V)-and v-Abl-transformed NIH 3T3 cells (Fig. 11A); similarly, ␤-catenin expression was significantly reduced in these transformed cells (Fig. 11A).
To further study if down-regulation of caveolin-1 in Rastransformed cells is associated with reductions in ␤-catenin expression, we also employed an NIH 3T3 cell line stably transfected with an IPTG-inducible Ha-Ras (G12V) expression vector. The results presented in Fig. 11B show that after induction of Ha-Ras (G12V) by IPTG treatment, both caveolin-1 and ␤-catenin levels are dramatically reduced, as compared with untreated control cells. Interestingly, under all these conditions, we have previously shown that caveolin-2 levels remain constant (26,31,84,85).
Taken together, these results support the hypothesis that modulation of caveolin-1 expression may be important for the regulation of ␤-catenin levels, with reduced levels of caveolin-1 favoring the interaction of ␤-catenin with GSK-3␤ leading to degradation of ␤-catenin, whereas increased levels of caveolin-1 may promote sequestration of ␤-catenin to caveolae membranes thereby protecting ␤-catenin from GSK-3␤-mediated degradation.

DISCUSSION
In recent years, several independent lines of evidence have emerged that implicate ␤-catenin as a contributing element in cancer progression. In view of the dual role of ␤-catenin in stabilizing adhesive complexes between adjacent cells and in modulating gene transcription, it is reasonable to speculate that alterations in the expression and/or the distribution of ␤-catenin are critical factors in ␤-catenin signaling. Elevated expression of ␤-catenin results in the activation of gene expression by the ␤-catenin-Lef-1 complex (59,62,63) and is associated with different forms of cancer, such as colon carcinoma (56,57) melanoma (58 -61), breast cancer (86,87), and other types of cancer (62). On the other hand, reduction in the expression of proteins that are part of the E-cadherin-adhesion complex has been found to impair cell-cell adhesion. For example, down-regulation of ␣-catenin or ␤-catenin expression (88) is associated with malignant transformation. Re-establishment of a functional cadherin-catenin complex reverts these invasive tumor cell lines to a benign cellular phenotype (89,90). These results indicate that cell-cell adhesion signals may be critical factors in carcinogenesis. However, the detailed molecular mechanism underlying the inverse relationship between cell transformation and the E-cadherin-␤-catenin complex remains unclear.
Recently, we have shown that caveolin-1 is dramatically up-regulated when NIH 3T3 cells reach confluence and is recruited to areas of cell-cell contact (26). This observation suggested that caveolin-1 may be involved in cell-cell adhesion. However, there is little information about the possible role of caveolins in regulating adhesion-mediated signaling.
In this study, we have investigated the role of caveolin-1 in modulating ␤-catenin-mediated signaling. Using several independent, yet complementary, approaches, we found that Ecadherin, ␤-catenin, and ␥-catenin co-localize with caveolin-1 at the plasma membrane, are enriched in purified caveolae membranes, and that these molecules co-immunoprecipitate with caveolin-1. We also found that whereas GSK-3␤, adenomatous polyposis coli, and axin form a multisubunit complex, GSK-3␤, which directly phosphorylates ␤-catenin, was excluded from caveolae membranes. In addition, we observed that caveolin-1 expression blocks the interaction between ␤-catenin and GSK-3␤. Because activation of GSK-3␤ induces degradation of ␤-catenin, one prediction of our results is that elevations in caveolin-1 expression will stabilize ␤-catenin, and conversely, down-regulation of caveolin-1 expression will enhance the degradation of ␤-catenin. In support of this hypothesis, we demonstrated that both caveolin-1 and ␤-catenin levels increase and decrease in parallel in response to a variety of cellular stimuli. These results are summarized schematically in Fig. 12.
Consistent with our current results, it has been shown that growth factor stimulation and Ras transformation result in adherens junction disassembly, exclusion of ␤-catenin from areas of cell-cell contact, and decreases in ␤-catenin expression (91)(92)(93). Under such conditions, ␤-catenin phosphorylation is increased. Increased phosphorylation of ␤-catenin followed by its degradation and could be because of direct interaction with GSK-3␤, as a consequence of reductions in caveolin-1 levels. Such parallel reductions in caveolin-1 and ␤-catenin expression may represent an important general mechanism to modulate cell-cell-mediated signaling, contributing to the loss of cell-cell adhesion and contact inhibition that is characteristic of transformed cells.
In this study, we have also demonstrated that caveolin-1 can efficiently inhibit ␤-catenin/Lef-1-responsive transcription induced by either Wnt-1 or by ␤-catenin overexpression. We found that expression of caveolin-1 in a caveolin-1 negative cell line recruited ␤-catenin to caveolae membranes and blocked the ␤-catenin/Lef-1 pathway. Interestingly, caveolin-1 inhibited ␤-catenin-mediated stimulation of transcription more potently than axin (Fig. 5), indicating that recruitment of ␤-catenin to caveolae membranes could represent a previously unappreciated mechanism for negatively regulating ␤-catenin/ Lef-1 signaling. The transformation of cells by Wnt-1 or by mutant ␤-catenin may possibly be inhibited by caveolin-1 by reducing the availability of ␤-catenin for the formation of a persistent, transcriptionally active complex between ␤-catenin and Lef-1. In accordance with the proposed role of caveolin-1 as a suppressor of cellular transformation (24 -29, 94, 95), our results suggest that up-regulation of caveolin-1 expression recruits ␤-catenin to caveolae membranes, inhibiting ␤-catenin/ Lef-1 signaling and thereby reinforcing cell-cell adhesion mechanisms.
In mammals, Wnt signaling has been shown to be involved in the morphogenesis of the central nervous system, kidneys, limbs, and the mammary gland (66, 67, 96). As we have shown that all three caveolin genes (Cav-1, -2, and -3) are expressed late during mouse embryogenesis (27) and that caveolin-1 in-hibits Wnt-1 signaling (this study), we suggest that caveolins may have a role in modulating embryonic and post-natal development. Future studies with caveolin-1 "knock-out" mice will be necessary to determine the possible signaling pathways affected by caveolin-1 deficiency in vivo.