p42/44 MAP Kinase-dependent and -independent Signaling Pathways Regulate Caveolin-1 Gene Expression

Caveolin-1 is a principal component of caveolae membranes in vivo. Caveolin-1 mRNA and protein expression are down-regulated in NIH 3T3 cells in response to transformation by activated oncogenes, such as H-Ras(G12V) and v-Abl. The mechanisms governing this down-regulation event remain unknown. Here, we show that caveolin-1 gene expression is directly regulated by activation of the Ras-p42/44 MAP kinase cascade. Down regulation of caveolin-1 protein expression by Ras is independent of (i) the type of activating mutation (G12V versus Q61L) and (ii) the form of activated Ras transfected (H-Ras versus K-Rasversus N-Ras). Treatment of Ras or Raf-transformed NIH 3T3 cells with a well characterized MEK inhibitor (PD 98059) restores caveolin-1 protein expression. In contrast, treatment of v-Src and v-Abl transformed NIH 3T3 cells with PD 98059 does not restore caveolin-1 expression. Thus, there must be at least two pathways for down-regulating caveolin-1 expression: one that is p42/44 MAP kinase-dependent and another that is p42/44 MAP kinase-independent. We focused our efforts on the p42/44 MAP kinase-dependent pathway. The activity of a panel of caveolin-1 promoter constructs was evaluated using transient expression in H-Ras(G12V) transformed NIH 3T3 cells. We show that caveolin-1 promoter activity is up-regulated ∼5-fold by inhibition of the p42/44 MAP kinase cascade. Using electrophoretic mobility shift assays we provide evidence that the caveolin-1 promoter (from −156 to −561) is differentially bound by transcription factors in normal and H-Ras(G12V)-transformed cells. We also show that activation of protein kinase A (PKA) signaling is sufficient to down-regulate caveolin-1 protein expression and promoter activity. Thus, we have identified two signaling pathways (Ras-p42/44 MAP kinase and PKA) that transcriptionally down-regulate caveolin-1 gene expression.

The subcellular distribution of several signaling molecules is restricted and regulated by association with scaffolding proteins (Ste5p, AKAPs (protein kinase A anchor proteins), and 14-3-3) (1, 2), forming a signaling pathway or module. Accumulating evidence suggests that caveolins possesses all the qualities of scaffolding proteins. We and other investigators have proposed the "caveolae signaling hypothesis," which states that caveolar localization of certain inactive signaling molecules could provide a compartmental basis for their regulated activation and explain cross-talk between different signaling pathways (3). In support of this idea, caveolin-1 binding can functionally suppress the GTPase activity of heterotrimeric G-proteins and inhibit the kinase activity of Src family tyrosine kinase through a common caveolin domain, termed the caveolin-scaffolding domain (4). Thus, we have suggested that caveolin may function as a negative regulator of many different classes of signaling molecules through the recognition of specific caveolin-binding motifs (4).
Caveolins form multivalent homo-and heter-oligomers and each caveolin-interacting protein binds to the same cytosolic membrane-proximal region of caveolin (5,6). Domain-mapping studies have revealed that the interaction of caveolin-1 with signaling molecules is mediated via a membrane proximal region of caveolin, termed the caveolin-scaffolding domain (residues 82-101). Through this domain, caveolin-1 interacts with G-protein ␣ subunits, H-Ras, Src family tyrosine kinases, protein kinase C isoforms, epidermal growth factor receptor, Neu, and eNOS (see Ref. 7; reviewed in Ref. 8). In many cases, it has been shown that mutational activation of these signaling molecules (G-proteins, H-Ras, or Src family kinases) prevents regulated interaction with the caveolin-scaffolding domain. These activating mutations include H-Ras(G12V) and G␣ s (Q227L) that are found in human cancers.
The caveolin-scaffolding domain recognizes a well defined caveolin-binding motif that includes several crucial aromatic amino acid residues (4,9,10). This motif was identified by using the caveolin-scaffolding domain to select random peptide ligands from phage display libraries (4,9,10). The relevance of the motif we identified was stringently evaluated using a well characterized caveolin-binding protein, namely a G-protein ␣ subunit (G␣ i2 ). Since the identification of the caveolin-scaffolding domain (6) and caveolin-binding sequence motifs (4,9,10), these observations have been extended to other caveolin-interacting proteins. Functional caveolin-binding motifs have been deduced in both tyrosine and serine/threonine kinases, as well as eNOS (reviewed in Ref. 8). In all cases examined, the caveolin-binding motif is located within the catalytic domain of a given signaling molecule. For example, in the case of tyrosine and serine/threonine kinases, a kinase domain consists of 11 conserved subdomains (I-XI), and the caveolin-binding motif occurs within subdomain IX (4,9,10). Caveolin-binding via the scaffolding domain is sufficient to inhibit the enzymatic activity of these kinases in vitro. Indeed, in many cases, a synthetic peptide corresponding to this caveolin domain is the most potent peptide inhibitor known for these enzymes. Agents that mimic the interaction with caveolins are potentially useful as general kinase inhibitors, and possibly as anti-tumor drugs.
Modification and/or inactivation of caveolin-1 appears to be a common feature of the transformed phenotype. Historically, caveolin was first identified as a v-Src substrate (11). Thus, caveolin may represent a critical target during cell transformation (11). In support of this notion, caveolin-1 mRNA and protein expression are reduced or absent in NIH 3T3 cells transformed by a variety of activated oncogenes (v-Abl, Bcr-Abl, H-Ras(G12V)), and caveolae are missing from these transformed cells (12); caveolin-2 protein is not down-regulated in response to oncogenic transformation (7,13). In addition, caveolin-1 expression levels correlated inversely with the ability of these cells to grow in soft agar, i.e. cells expressing the smallest amount of caveolin-1 and lacking detectable caveolae formed the largest colonies in soft agar. Furthermore, our laboratory and other investigators have demonstrated that recombinant expression of caveolin-1 in transformed NIH 3T3 cells and mammary carcinoma cell lines abrogates their growth in soft agar (14,15). These results suggest that down-regulation of caveolin-1 protein expression and caveolae organelles may be critical to maintaining the transformed phenotype.
The mechanisms that govern caveolin-1 down-regulation remain largely unknown. Here, we have identified two signaling pathways (Ras-p42/44 MAP kinase 1 and PKA) that can transcriptionally down-regulate caveolin-1 promoter activity.

EXPERIMENTAL PROCEDURES
Materials-Anti-caveolin-1 IgG (monoclonal antibody 2297 (16)) and anti-caveolin-2 IgG (monoclonal antibody 65 (13)) were the gifts of Dr. Roberto Campos-Gonzalez, Transduction Labs. Antibodies against GDP-dissociation inhibitor were the generous gift of Dr. Perry Bickel, Washington University, St. Louis, MO. Other reagents were purchased commercially: anti-activated ERK-1/2 IgG (p42/44 MAP kinase; New England Biolabs, Inc), fetal bovine serum (JRH Biosciences), and prestained protein markers (Life Technologies, Inc.). PD 98059 was purchased from Calbiochem and dissolved in dimethyl sulfoxide at a concentration of 50 mM and used at a final concentration of 50 M. Forskolin and IBMX were purchased from Sigma and used at final concentrations of 10 and 500 M, respectively.
Caveolin-1 Promoter Constructs-pA3Luc (19,20) was the luciferase reporter plasmid into which the caveolin promoters were cloned. A ϳ13-kb NotI genomic clone containing murine caveolin-1 exons 1 and 2 was isolated from a murine genomic library and cloned into Bluescript SK. An ϳ3-kb fragment of genomic sequence upstream of the starting ATG was isolated by PCR. This fragment was subcloned into pZero Blunt (Invitrogen). Subsequently, a SacI fragment containing the entire PCR product was subcloned into pXP2 (a generous gift from Dr. Susan Horwitz, The Albert Einstein College of Medicine, NY). For construction of Pr-3kb, a 2.2-kb KpnI-HindIII fragment from pXP2, and a ϳ750-bp HindIII-HindIII fragment from pZero Blunt were combined in a three part ligation into the KpnI-HindIII site of pA3Luc. For construction of Pr-750bp, the 750-bp HindIII-HindIII fragment was subcloned into the HindIII site of pA3Luc. For construction of Pr-3kb and Int1, we created a fusion protein that contains ϳ3 kb of promoter, the first exon of caveolin-1, the first intron, and the beginning of the second exon (extending to caveolin-1 protein sequence NIYKP) fused in-frame with the starting methionine of the luciferase gene. We took advantage of unique NarI sites at position Ϫ750 of the promoter of caveolin-1 and another located 32 bp downstream of the starting ATG in the luciferase gene. We used a primer, AGGATAGAATGGCGCCGGG-CCTTTCTTTATGTTTTTGGCGTCTTCCATGGGCTTGTAGATGTTG-CCCTGTTC, to generate the 3Ј end of the PCR product. This primer encodes for NIYKP(cav-1 exon 2)-MEDAK (luciferase) and extends past the NarI site in luciferase. The PCR product was cloned into Pr-3kb which had been digested with NarI. Constructs expressing ERK-2, epidermal growth factor receptor, and Raf were as we described previously (21). The PKA expression vector (encoding the murine ␣ catalytic subunit) was obtained from the PathDetect CREB trans-Reporting System (Stratagene, Inc).
Luciferase Assays-Transient transfections (using calcium phosphate precipitation) and luciferase assays were performed essentially as described previously (7,21). Briefly, 300,000 cells (NIH 3T3 cells or CHO cells as specified) were seeded in six-well plates 12-24 h before the transfection. Each point was transfected with either 2 g of reporter for experiments in which only one plasmid was transfected; or 1 g of each plasmid plasmid for experiments in which two plasmids were co-transfected. 12 h post-transfection, the cells were rinsed twice with phosphate-buffered saline and incubated in fetal bovine serum for another 24 -36 h. This incubation was done in the presence of 50 M PD 98059 or 1 mM IPTG when applicable. The cells were then in lysed in 200 l of extraction buffer, 75 l of which was used to measure luciferase activity, as described (22). For experiments assessing promoter activity in response to the p42/44 MAP kinase activators and PKA, after washing with phosphate-buffered saline, the cells were incubated in the presence of 1% fetal bovine serum for 24 -36 h as described previously (21). These assays were made possible through the use of a special CHOderived cells line, called GRC ϩ LR-73. Unlike parental CHO cells, GRC ϩ LR-73 cells are a non-transformed growth control revertant that has normal fibroblastic morphology, does not grow in suspension, requires high serum concentrations for growth, and undergoes synchronized growth arrest in low concentrations of serum (1-2%) without a loss of viability (18). Also, these cells have a much higher transfection efficiency (ϳ10-fold) than parental CHO cells.
Immunoblotting-Samples were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose. After transfer, nitrocellulose sheets were stained with Ponceau S to visualize protein bands and subjected to immunoblotting. For immunoblotting, incubation conditions were as described by the manufacturer (Amersham Pharmacia Biotech), except we supplemented our blocking solution with both 1% bovine serum albumin and 2% non-fat dry milk (Carnation). Bound antibodies were visualized using ECL (Amersham Pharmacia Biotech). Quantitation of Western blot films was performed using an AlphaInnotech ChemiImager 4000 low-light imaging system (San Leandro, CA) using the AlphaEase software package.
Immunoblotting with Phospho-specific Antibody Probes-To investigate the activation state of p42/44 MAP kinase, we employed a phosphospecific antibody probe that has been generated against the activated form of ERK-1/2 (New England Biolabs, Inc). It has previously been shown that this antibody can be used to selectively detect activated p42/44 MAP kinase by Western blotting. Cells were lysed in boiling sample buffer, as suggested by the manufacturer of phospho-specific antibody probes (New England Biolabs, Inc.). Samples were then collected and boiled for a total of 5 min. Samples were homogenized using a 26-g needle and a 1-ml syringe. After SDS-polyacrylamide gel elec-trophoresis and transfer to nitrocellulose, blots were probed with primary antibodies (dilution of 1:500; New England Biolabs, Inc.) and the appropriate horseradish peroxidase-conjugated secondary antibody (dilution of 1:5000; Transduction Laboratories). Bound antibodies were visualized using ECL (Amersham Pharmacia Biotech).
Northern Analysis-Total RNA was extracted and purified according to the manufacturer's instructions (Qiagen, Inc.). Ten micrograms of total cellular RNA was denatured with formaldehyde and subjected to Northern blot analysis with 32 P-labeled probes for the mouse caveolin-1 mRNA (2.4 kb). The 28 S and 18 S rRNA were visualized by ethidium bromide staining.
Electrophoretic Mobility Shift Assay (EMSAs)-Electrophoretic mobility shift assays were performed as described (23,24), with minor modifications. Briefly, nuclear extracts were prepared by the method described by Schreiber et al. (25). Extracts were isolated from ϳ10 8 cells, aliquoted, and frozen immediately. Concentrations were determined using the BCA Protein Assay Reagent (Pierce Chemical Co.). DNA probes for the EMSA were constructed by PCR using the mouse caveolin-1 genomic clone described above. The four overlapping probes (each ϳ250 bp) that spanned 750 bp of the promoter were generated as follows: Probe A was amplified with 5Ј-AGACCCGGCGCAGAG-CACGTCCTAG-3Ј and 5Ј-TCGGAGTCCAC GTATTTGCCC-3Ј primers; Probe B was amplified with 5Ј-CCTCCACCCCTGCTGAGATGATG-3Ј and 5Ј-GTTCTGCTCTCAGTTGGC TAGGAC-3Ј primers; Probe C was amplified with 5Ј-GGTTCCCAGCCATCTCGCTTCTATATC-3Ј and 5Ј-AACCTACAGAGAGGCATCCAGGG-3Ј primers; and Probe D was amplified with 5Ј-CTCTCTAGTAACAAGCTTGAACCTC-3Ј and 5Ј-TCT-GTCTCCTTGTTTCACAGAG-3 primers. Approximately 200 ng of purified PCR product was end-labeled with [␥-32 P]ATP (NEN Life Science Products Inc.). EMSA was performed by the method of Singh et al. (23,24), with minor modifications. Briefly, 15 g of nuclear extracts was incubated with 5 g of poly(dI-dC) (Amersham Pharmacia Biotech) in binding buffer (12% glycerol, 12 mM HEPES, pH 7.9], 4 mM Tris, pH 8.0, 50 mM KCl, 1 mM EDTA, and 1 mM dithiothreitol) on ice for 15 min. Approximately 40,000 cpm of end-labeled probe was added and incubated for an additional 30 min on ice. Protein-DNA complexes were separated on a 5% polyacylamide gel in 1 ϫ TBE at 20 mA. The gels were dried and complexes were visualized by autoradiography.
Determination of the Transcription Start Site-The murine transcriptional start site was determined by 5Ј-rapid amplification of cDNA ends analysis using a previously described 3T3-L1 adipocyte library cloned into pCDNA1 (26). Briefly, PCR products were amplified using an anchor primer from pCDNA1 and an oligonuceotide primer that is antisense to nucleotides 9 -26 of murine caveolin-1. PCR products were cloned into pCR-Blunt (Invitrogen). The transcriptional start site was determined to be at Ϫ63 by direct sequencing of the subcloned inserts. Thus, sequence of the 5Ј-untranslated region is: CAGTTCTC-TTAAATCACAGCCCAGGGAAACCTCCTCAGAGCCTGCAGCCAGCC-ACGCGCCAGC.

Down-regulation of Caveolin-1 Protein Expression by p42/44 MAP Kinase-dependent and -Independent
Signaling Pathways-Expression of caveolin-1 mRNA and protein are downregulated in H-Ras(G12V) and v-Abl transformed NIH 3T3 cells (12,14). In contrast, expression of the caveolin-2 protein is largely unaffected in these cells (13). The mechanism by which these transforming oncogenes down-regulate caveolin-1 expression remains unknown. One possibility is that caveolin-1 gene expression is negatively regulated by constitutive activation of the Ras-p42/44 MAP kinase cascade.
To investigate this hypothesis further, we examined the expression of caveolin-1 protein in a number of other Ras-transformed NIH 3T3 cells. Fig. 1A shows that caveolin-1 protein expression was down-regulated in NIH 3T3 cells transformed by H-Ras(Q61L), K-Ras(G12V), N-Ras(Q61K), and v-Raf. These results indicate that Ras-induced down-regulation of caveolin-1 is (i) independent of the type activating mutation (G12V versus Q61L), (ii) independent of the type of Ras transfected (H-Ras versus K-Ras versus N-Ras), and (iii) also occurs if transformation is mediated by an element of the Ras-MAP kinase pathway that is directly downstream of Ras itself (v-Raf). The expression of caveolin-2 was unaffected by these activated oncogenes. The expression of both caveolin-1 and -2 in v-Abl and v-Src transformed NIH 3T3 cells is shown for comparison. Down-regulation of caveolin-1 protein expression by v-Src is also secondary to down-regulation of the caveolin-1 mRNA, as seen by Northern analysis (Fig. 1B). In accordance with these observations, we have shown that caveolin-1 mRNA levels are also dramatically down-regulated in Ras-and v-Abl transformed NIH 3T3 cells (12,14).
We previously demonstrated that treatment of H-Ras (G12V)-transformed NIH 3T3 cells with a well characterized MEK inhibitor (PD 98059) is sufficient to restore expression of the caveolin-1 protein product in these cells (14). , v-Raf, v-Abl, and v-Src. A lane containing normal NIH 3T3 cells is shown for comparison (wt). Caveolin-1 and caveolin-2 protein expression was detected by immunoblotting with isoform-specific monoclonal antibody probes. Two exposures for caveolin-1 expression are shown to better illustrate the relative level of caveolin-1 down-regulation. Each lane contains equal amounts of total protein. Quantitation revealed that caveolin-1 protein levels were down-regulated ϳ50 -100-fold by v-Abl and all the Ras isoforms tested, while only ϳ3-4-fold down-regulation of caveolin-1 was observed during v-Raf and v-Src induced transformation. B, Northern analysis of total RNA for the expression of caveolin-1 mRNA (2.4 kb message) in normal and v-Src transformed NIH 3T3 cells. Note that the caveolin-1 message is down-regulated. The levels of 18 S rRNA are shown for comparison and serve as control for equal loading. In accordance with our current results with v-Src, we have previously shown that transformation of NIH 3T3 cells by H-Ras(G12V) or v-Abl down-regulates caveolin-1 mRNA levels (12).
hypothesis that the down-regulation of caveolin-1 expression in these cells is due to constitutive activation of the p42/44 MAP kinase cascade. Interestingly, treatment of v-Src and v-Abl transformed NIH 3T3 cells with PD 98059 did not restore caveolin-1 expression. Thus, there must be at least two or three independent pathways for down-regulating caveolin-1 expression: one that is p42/44 MAP kinase-dependent and others that are p42/44 MAP kinase-independent (i.e. activated by v-Abl or v-Src). In accordance with the observation that caveolin-1 protein expression is restored by treatment of H-Ras(G12V) transformed cells with PD 98059, we also observed that the same treatment up-regulated expression of the caveolin-1 mRNA by ϳ5-fold (Fig. 2B). Activation of the p42/44 MAP Kinase Cascade Transcriptionally Down-regulates Caveolin-1 Promoter Activity-As the down-regulation of caveolin-1 protein expression is strictly correlated with a loss of caveolin-1 mRNA expression (12), this event may be governed by transcriptional control. To test this hypothesis directly, we identified and cloned the murine caveolin-1 gene. The DNA sequence of the murine caveolin-1 promoter region has been deposited in GenBank under accession number AF124227. The murine transcriptional start site was determined to be at Ϫ63 (see "Experimental Procedures").
We next used this murine genomic clone to generate three caveolin-1 promoter constructs that use luciferase expression as the reporter. These three constructs are illustrated schematically in Fig. 4A. Briefly, the first construct contains the 750 bp of sequence upstream of the caveolin-1 ATG (Pr-750bp), the second construct contains ϳ3 kb upstream of the caveolin-1  ATG (Pr-3kb), and the third construct contains ϳ3 kb upstream of the caveolin-1 ATG, caveolin-1/exon 1, caveolin-1 intron 1, and a portion of caveolin-1/exon 2 (Pr-3 kb and Int 1).
The promoter activity of these constructs was next examined by transient transfection of normal NIH 3T3 cells. Fig. 4B shows that all three constructs contained promoter activity, as compared with the empty vector control (pA3 Luc). It should be noted that the 3-kb promoter in the antisense orientation did not produce any luciferase activity above the empty vector control (pA3 Luc) (data not shown). Interestingly, Pr-750 and Pr-3kb behaved similarly, while Pr-3kb and Int 1 had about twice the promoter activity. These results indicate that additional positive regulatory elements may be present within caveolin-1 coding sequences or caveolin-1/intron 1.
The activity of these caveolin promoter constructs was next evaluated using transient expression in H-Ras(G12V)-transformed NIH 3T3 cells. In addition, these cells were treated with PD 98059 or left untreated. Our results indicate that all three promoter constructs were stimulated by treatment with PD 98059, but that Pr-3kb and Int 1 showed the largest response ( Fig. 5). In addition, the empty vector alone (pA3 Luc) showed no activity either in the presence or absence of PD 98059. These results indicate that caveolin-1 promoter activity can be upregulated by ϳ5-fold through inhibition of the p42/44 MAP kinase cascade in Ras-transformed cells. This finding is in agreement with the up-regulation of caveolin-1 mRNA observed in H-Ras(G12V)-transformed NIH 3T3 cells treated with PD 98059 (Fig. 2B).
Conversely, we examined if caveolin-1 promoter activity could be down-regulated by constitutive activation of the p42/44 MAP kinase cascade. For this purpose, we utilized normal NIH 3T3 cells that harbor H-Ras(G12V) under the control of a lac z inducible promoter. Thus, these cells can be induced to express H-Ras(G12V) and undergo cell transformation in the presence of IPTG. Fig. 6A shows that addition of IPTG to the medium is sufficient to cause down-regulation of the caveolin-1 protein product and that this reduction in caveolin-1 protein expression is blocked by treatment with PD 98059. Similarly, addition of IPTG (i.e. induction of H-Ras(G12V) expression) was sufficient to down-regulate caveolin-1 promoter activity under the same conditions (Fig. 6B). Interestingly, Pr-3kb and Int 1 again showed the largest response.
To evaluate if the p42/44 MAP kinase cascade is directly involved in controlling caveolin-1 promoter activity, we transiently co-transfected CHO cells with either vector alone, a constitutively active form of Raf, or ERK-2 itself and the caveolin-1 promoter (Pr-3kb and Int 1). Under these conditions, co-transfection with either constitutively active Raf or ERK-2 was sufficient to dramatically down-regulate caveolin-1 promoter activity (Fig. 7). These results clearly indicate that ERK itself can down-regulate caveolin-1 promoter activity. Conversely, we have previously shown that these Raf and ERK-2 constructs dramatically up-regulate a p42/44 MAP kinase-sensitive Elk-reporter (21).
The Caveolin-1 Promoter Region from Ϫ156 to Ϫ561 Is Differentially Bound by Transcription Factors in Normal and H-Ras(G12V) Transformed Cells-As caveolin-1 mRNA is down-regulated in response to activation of the p42/44 MAP

FIG. 4. Caveolin-1 promoter activity in normal NIH 3T3 cells.
A, three caveolin-1 promoter constructs that use luciferase expression (pA3 Luc) as the reporter are illustrated schematically. The first construct contains 750 bp of sequence upstream of the caveolin-1 ATG (Pr-750bp), the second construct contains ϳ3 kb upstream of the caveolin-1 ATG (Pr-3kb), and the third construct contains ϳ3 kb upstream of the caveolin-1 ATG, plus caveolin-1/exon 1, caveolin-1 intron 1, and a portion of caveolin-1/exon 2 (Pr-3kb and Int 1). B, NIH 3T3 cells were transiently transfected with each of the caveolin-1 promoter constructs or empty vector (pA3 Luc) alone. Relative luciferase activity is shown. Note that all three caveolin-1 promoter constructs contained promoter activity, as compared with the empty vector control (pA3 Luc). Interestingly, Pr-750 and Pr-3kb behaved similarly, while Pr-3kb and Int 1 had ϳ2 times the promoter activity.

FIG. 5. Treatment of H-Ras(G12V)-transformed NIH 3T3 cells with PD 98059 up-regulates caveolin-1 promoter activity.
H-Ras(G12V)-transformed NIH 3T3 cells were transiently transfected with the three caveolin-1 promoter constructs and then treated in the presence (ϩ) or absence (Ϫ) of the MEK inhibitor (PD 98059; 50 M for 2 days). Relative luciferase activity is shown. Note that all three promoter constructs were stimulated by treatment with PD 98059, but that Pr-3kb and Int 1 showed the largest response (ϳ5-fold). Empty vector alone (pA3 Luc) showed no activity either in the presence or absence of PD 98059. kinase pathway, we wished to determine if different complexes are formed on the caveolin-1 promoter in normal NIH 3T3 cells as compared with Ras(G12V)-transformed NIH 3T3 cells. To this end, we conducted a series of EMSAs covering the smallest promoter region that was responsive for regulation by activation of the p42/44 MAP kinase cascade. Four overlapping fragments (staggered by ϳ50 bp) were end-labeled and incubated with nuclear extracts from either normal NIH 3T3 cells or H-Ras(G12V)-transformed NIH 3T3 cells. After incubation, samples were run on nondenaturing acrylamide gels and subjected to autoradiography. Fig. 8 shows that no differences were observed with fragment A (Cav-1/exon 1 plus Ϫ1 to Ϫ197). However, dramatic differences were noted with fragments B (Ϫ156 to Ϫ401) and C (Ϫ344 to Ϫ561). No differences were observed with fragment D (Ϫ509 to Ϫ736; data not shown). Thus, our results provide evidence that the caveolin-1 promoter region from Ϫ156 to Ϫ561 is differentially bound by transcription factors in normal and H-Ras(G12V)-transformed cells.
Down-regulation of Caveolin-1 Protein and Promoter Activity by Activation of PKA-As caveolin-1 expression can be downregulated by constitutive activation of the p42/44 MAP kinase cascade through transcriptional regulation, we next evaluated the effects of another well established signaling cascade on caveolin-1 promoter activity. We transiently co-transfected CHO cells with either vector alone or the catalytic subunit of protein kinase A (PKA) and the caveolin-1 promoter (Pr-3kb and Int 1). Our results indicate that overexpression of PKA was sufficient to down-regulate caveolin-1 promoter activity (Fig.  9A). In support of these observations, treatment of CHO cells with agents that elevate cellular cAMP and activate the PKA pathway (either IBMX (a PDE inhibitor) or forskolin (an activator of adenylyl cyclase)) dramatically down-regulates caveolin-1 protein expression (Fig. 9B). Similarly, treatment with IBMX or forskolin also dramatically down-regulated caveolin-2  7. The p42/44 MAP kinase cascade is directly involved in controlling caveolin-1 promoter activity. CHO cells were transiently transfected with either vector alone, a constitutively active form of Raf, or ERK-2 and the caveolin-1 promoter (Pr-3kb and Int 1). Note that co-transfection with either constitutively active Raf or ERK-2 was sufficient to dramatically down-regulate caveolin-1 promoter activity.

FIG. 8. The caveolin-1 promoter region from ؊156 to ؊561 is differentially bound by transcription factors in normal and H-Ras(G12V)-transformed cells.
A series of EMSA were performed covering the smallest promoter region that was responsive to regulation by activation of the p42/44 MAP kinase cascade. Four overlapping fragments (staggered by ϳ50 bp) were end-labeled and incubated with nuclear extracts from either normal NIH 3T3 cells or H-Ras(G12V)transformed NIH 3T3 cells. A minus (Ϫ) indicates that no nuclear extract was added as a negative control. Samples were run on nondenaturing acrylamide gels and subjected to autoradiography. Note that no differences were observed with fragment A (exon 1 plus Ϫ1 to Ϫ197). In contrast, dramatic differences were noted with fragments B (Ϫ156 to Ϫ401) and C (Ϫ344 to Ϫ561). No differences were observed with fragment D (Ϫ509 to Ϫ736; data not shown). protein expression. The expression of another cellular protein (GDP-dissociation inhibitor) is shown as an additional control for equal protein loading. Fig. 9C shows that the effects of PKA activation and p42/44 MAP kinase activation on caveolin-1 expression are independent. Note that addition of the MEK inhibitor (PD 98059) to forskolin or IBMX-treated fibroblasts does not restore caveolin-1 expression (upper panel), although we show that PD 98059 effectively inhibits the activation of p42/44 MAP kinase under these conditions (lower panel).
In the case of PKA activation, both caveolin-1 and caveolin-2 protein are down-regulated (Fig. 9B), while in the case of v-Raf and various forms of Ras, caveolin-1 levels are down-regulated and caveolin-2 levels are relatively unaffected (Figs. 1 and 2). Thus, these data independently support the idea that the effects of PKA and p42/44 MAP kinase activation on caveolin-1 protein expression are separate and independent. DISCUSSION Down-regulation of the caveolin-1 protein is a direct consequence of the oncogenic stimulus as it can be reversed by employing a temperature-sensitive form of v-Abl or by treating Ras(G12V)-transformed 3T3 cells with an inhibitor of the p42/44 MAP kinase pathway (PD 98059) (14). Re-introduction of caveolin-1 under control of an inducible expression system is sufficient to block the anchorage-independent growth of these transformed cells in soft agar and restore the formation of morphologically detectable caveolae (14). Consistent with its antagonism of Ras-mediated cell transformation, caveolin-1 expression dramatically inhibited both Ras/MAPK-mediated and basal transcriptional activation of a mitogen-sensitive promoter (14). Taken together, these results indicate that downregulation of caveolin-1 expression and caveolae organelles may be critical to maintaining the transformed phenotype in certain cell populations (14).
Recently, we have employed an antisense approach to derive stable NIH 3T3 cell lines that contain normal amounts of caveolin-2, but express dramatically reduced levels of caveolin-1 (27). NIH 3T3 cells harboring antisense caveolin-1 spontaneously formed foci, exhibited anchorage-independent growth in soft agar, formed tumors in immunodeficient mice, and appeared morphologically transformed as seen by scanning electron microscopy (27)). Biochemically, these cells also showed increased levels of activated MEK and ERK (27). Taken together, these results suggest that down-regulation of caveolin-1 expression is sufficient to drive oncogenic transformation and constitutively activate the p42/44 MAP kinase cascade (27). Importantly, cell transformation induced by targeted down-regulation of caveolin-1 expression was completely reversed when caveolin-1 protein levels were restored to normal by loss of the caveolin-1 antisense vector (27). Thus, caveolin-1 behaves as would be expected for a tumor suppressor.
Here, we have examined the signaling pathways that govern caveolin-1 gene expression. We show that caveolin-1 gene expression is directly regulated by activation of the p42/44 MAP kinase cascade. Treatment of Ras(H-, K-, and N-Ras) or v-Raftransformed NIH 3T3 cells with a well characterized MEK inhibitor (PD 98059) restores the expression of the caveolin-1 protein. However, treatment of v-Src and v-Abl transformed NIH 3T3 cells with PD 98059 has no effect on caveolin-1 expression. Thus, there are at least two or three pathways for down-regulating caveolin-1 expression: one that is p42/44 MAP kinase dependent and others that are p42/44 MAP kinase independent and depend on the activation of non-receptor tyrosine kinases (such as Src or Abl).
The activity of caveolin-1 promoter constructs was evaluated using expression in H-Ras(G12V)-transformed NIH 3T3 cells. Caveolin-1 promoter activity was up-regulated by ϳ5-fold through inhibition of the p42/44 MAP kinase cascade with PD 98059. In addition, transient transfection of CHO cells with ERK-2 dramatically down-regulates caveolin-1 promoter activity. To determine if different complexes form on the caveolin-1 promoter in normal and Ras(G12V)-transformed NIH 3T3 cells, we performed electromobility shift assays. Our results provide evidence that the caveolin-1 promoter from Ϫ156 to Ϫ561 is differentially bound by transcription factors in normal and H-Ras(G12V)-transformed cells.
We also evaluated the effects of the PKA pathway on caveolin-1 gene expression. Activation of the PKA pathway by pharmacological agents (IBMX and forskolin) or by overexpression of the PKA catalytic subunit was sufficient to down-regulate caveolin-1 promoter activity and caveolin-1 protein expression. Thus, there may be three "independent" signaling pathways (Ras-p42/44 MAP kinase, NRTKs, and PKA) that can transcriptionally down-regulate caveolin-1 gene expression.
These findings may have relevance to human cancers. 1) Using differential display and subtractive hybridization techniques, Sager and co-workers (34) have identified a number of "candidate tumor suppressor genes"; these are genes whose mRNAs are down-regulated in human mammary carcinomas. In this screening approach, caveolin-1 was independently identified as one of 26 gene products down-regulated during human mammary tumorigenesis. In addition, caveolin-1 expression was absent in several transformed cell lines derived from human mammary carcinomas including: MT-1, MCF-7, ZR-75-1, T47D, MDA-MB-361, and MDA-MB-474 (34). In contrast, caveolin-1 mRNA was abundantly expressed in normal mammary epithelium.
2) Human tumor cytogenetic data are also consistent with this proposal. Loss of heterozygosity analysis implicates 7q31.1 in the pathogenesis of multiple types of cancer, including breast, ovarian, prostate, and colorectal carcinomas, as well as uterine sarcomas and leiomyomas. The locus of the presumed 7q31.1 tumor suppressor gene has been narrowed to a ϳ1 mega-base region that includes the highly polymorphic marker D7S 522. Zenklusen and colleagues (see references cited in Refs. 35 and 36) have shown that the D7S 522 locus is the most commonly deleted marker in primary breast cancers, and they note that loss of heterozygosity at this site is strongly associated with systemic progression and death in prostate cancers. D7S 522 also spans the aphidicolin-induced fragile site FRA7G at 7q31. Given the usefulness of 7q31.1 and D7S 522 loss of heterozygosity as markers for carcinogenesis, many laboratories are currently searching this chromosomal region for a novel tumor suppressor gene. Recently, we have shown that CAV1 and CAV2 map within 100 kb of D7S 522, in the middle of the 1 Mb smallest common deleted region for the presumed tumor suppressor gene (35,36). Evidence that caveolin-1 can suppress cell transformation in murine fibroblasts and human breast cancer cell lines provides independent support for the model that CAV1 is the missing tumor suppressor gene (14,15).
3) Neu, c-erbB2, is a proto-oncogene product that encodes an epidermal growth factor-like receptor tyrosine kinase. Amplification of wild-type c-Neu and mutational activation of Neu (Neu T) have been implicated in oncogenic transformation of cultured fibroblasts and the pathogenesis of human breast cancers in vivo. Recently, we examined the relationship between Neu tyrosine kinase activity and caveolin-1 protein expression in vitro and in vivo. These studies demonstrated that mutational activation of c-Neu down-regulated caveolin-1 protein expression, but not caveolin-2, in cultured NIH 3T3 and Rat 1a cells (7). Conversely, recombinant overexpression of caveolin-1 blocked Neu-mediated signal transduction in vivo.
These results indicate that a negative reciprocal relationship exists between c-Neu tyrosine kinase activity and caveolin-1 protein expression. In accordance with these in vivo studies, a 20-amino acid peptide derived from this region (the caveolin-1 scaffolding domain) was sufficient to inhibit Neu-autophosphorylation in an in vitro kinase assay (7). Based on these studies, caveolin-1 expression also inhibits the function of c-Neu, suggesting that caveolin-1 based mimetic peptides or drugs that up-regulate caveolin-1 gene expression would provide an independent and valid approach for the treatment of human breast cancers.
In conclusion, as caveolin-1 down-regulation appears to be involved in mammary and fibroblastic cell transformation (14,15,27), an understanding of the signaling pathways that control caveolin-1 expression may ultimately yield novel cancer treatments. For example, our results suggest that the caveolin-1 promoter may be useful in identifying compounds that reverse oncogenic transformation of Ras-transformed cells. We show here that PD 98059, a well characterized inhibitor of MEK, clearly up-regulates caveolin-1 promoter activity in Rastransformed cells. This MEK inhibitor is also known to revert the phenotype of Ras-transformed cells (37). Thus, screening assays employing the caveolin-1 promoter could provide a general strategy to identify novel inhibitors of the p42/44 MAP kinase cascade, the PKA cascade, or other signaling cascades whose inhibition up-regulates caveolin-1 gene expression.