CEACAM6 cross-linking induces caveolin-1-dependent, Src-mediated focal adhesion kinase phosphorylation in BxPC3 pancreatic adenocarcinoma cells.

Despite lacking transmembrane or intracellular domains, glycosylphosphatidylinositol-anchored proteins can modulate intracellular signaling events, in many cases through aggregation within membrane "lipid raft" microdomains. CEACAM6 is a glycosylphosphatidylinositol-linked cell surface protein of importance in the anchorage-independent survival and metastasis of pancreatic adenocarcinoma cells. We examined the effects of antibody-mediated cross-linking of CEACAM6 on intracellular signaling events and anchorage-independent survival of the CEACAM6-overexpressing pancreatic ductal adenocarcinoma cell line, BxPC3. CEACAM6 cross-linking increased c-Src activation and induced tyrosine phosphorylation of p125(FAK) focal adhesion kinase. Focal adhesion kinase phosphorylation was dependent on c-Src kinase activation, for which caveolin-1 was required. CEACAM6 cross-linking induced a significant increase in cellular resistance to anoikis. These observations represent the first characterization of the mechanism through which this important cell surface oncoprotein influences intracellular signaling events and hence malignant cellular behavior.

CEACAM6 1 is a glycosylphosphatidylinositol (GPI)-anchored cell surface protein that is overexpressed in a variety of human malignancies, including pancreatic adenocarcinoma (1,2). This immunoglobulin superfamily member is emerging as an important determinant of a variety of aspects of the malignant cellular phenotype (1)(2)(3). We reported previously that levels of CEACAM6 expression modulate the susceptibility of pancreatic adenocarcinoma cells to anoikis (2). Normal epithelial cells require anchorage to an extracellular matrix for growth, survival, and differentiation. Resistance to anoikis, a subset of apoptosis induced by inadequate or inappropriate cell-substrate contact in normal cells, is a property of transformed cells that is associated with enhanced tumorigenesis and metastatic ability. Despite clear indications of its important role in cancer cell biology, the mechanisms through which CEACAM6 influences intracellular signal transduction remain poorly understood.
Despite lacking transmembrane and intracellular domains, several GPI-anchored proteins, including CEACAM family members, are able to influence intracellular events and have been implicated in transmembrane signaling via tyrosine kinases (4 -7). GPI-anchored proteins are known to exist in clusters and form microdomains at the surface of the plasma membrane, referred to as "lipid rafts" (8). These microdomains appear to be important for GPI-anchored protein signal transduction. A number of signal transduction components, including Src family tyrosine kinases, are associated with lipid rafts, and cross-linking of cell surface molecules by a variety of techniques has proven to be a useful tool for exploring the functional roles of raft-associated proteins (9 -14).
Here, we examine the effects of antibody-mediated crosslinking of CEACAM6 in pancreatic adenocarcinoma cells. We determine the effects of CEACAM6 cross-linking on the activity of c-Src and on the phosphorylation status of its substrate p125 FAK focal adhesion kinase (FAK), a tyrosine kinase of key importance in pancreatic adenocarcinoma cellular resistance to anoikis (15) and among the main targets for tyrosine phosphorylation in v-Src-transformed cells (16). We also provide evidence that modulation of c-Src tyrosine kinase activity by CEACAM6 is a caveolin-1-dependent process and illustrate the functional effects of CEACAM6 cross-linking on anchorageindependent cell survival, which correlates with tumorigenic, invasive, and metastatic potential in a range of malignancies, including pancreatic adenocarcinoma (17)(18)(19)(20)(21).

EXPERIMENTAL PROCEDURES
Cells and Cell Culture-BxPC3 human pancreatic ductal adenocarcinoma cells were obtained from the American Type Culture Collection (ATCC, Manassas, VA). The cells were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% fetal bovine serum (Invitrogen), incubated in a humidified (37°C, 5% CO 2 ) incubator, grown in 75-cm 2 culture flasks, and passaged upon reaching 80% confluence.
CEACAM6 Cross-linking-The By114 mouse monoclonal antibody was used to cross-link CEACAM6. This antibody is highly specific for CEACAM6 and is non-cross-reactive with closely related CEACAM family molecules (22,23). Cross-linking was performed in a manner similar to that described previously (12). Briefly, BxPC3 cell cultures, 3 days post-seeded in 35-mm well plates, were incubated with 50 g/ml By114 or irrelevant (control) isotype-matched mouse IgG on ice for 30 min in bovine serum albumin medium (DMEM without fetal bovine serum containing 1% bovine serum albumin and 20 mM HEPES, pH 7.4). After washing with cold bovine serum albumin medium, the cells were incubated with 50 g/ml anti-mouse IgG affinity-purified polyclonal antibody (secondary antibody) in bovine serum albumin medium at 37°C for the indicated times. After washing with ice-cold phosphatebuffered saline, lysates were prepared in 1% Triton X-100, 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride at 4°C. After centrifugation at 14,000 ϫ g for 3 min, the supernatants were used for further analysis. The cells exposed to pri-mary antibody only served as additional controls. To disrupt GPI anchor assembly, cells were cultured for 18 h in medium containing 10 mM mannosamine (2-amino-2-deoxy-D-mannose, Sigma) prior to performing antibody cross-linking (24).
Immunoprecipitation-Cells were washed extensively and lysed in buffer (10 mM Tris, pH 7.4, 100 mM NaCl, 2.5 mM MgCl 2 , 0.5% Triton X-100, 10 mM dithiothreitol) supplemented with a protease and phosphatase inhibitor mixture (Sigma) at 4°C. After standing on ice for 60 min with brief intermittent vortexing, the cell lysates were centrifuged at 23,000 ϫ g for 15 min at 4°C. Protein concentrations were determined and normalized, and crude lysates were precleared with protein-A-Sepharose (Zymed Laboratories Inc., San Francisco, CA). The lysates were subjected to immunoprecipitation assay using 25 l of an appropriate antibody. Immune complexes were precipitated with 40 l of protein A-Sepharose beads for 6 h at 4°C followed by sequential washes in immunoprecipitation buffer containing 500 mM, 300 mM, and 150 mM NaCl (twice) supplemented with 1ϫ radioimmunoprecipitation assay buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% v/v Triton X-100, 1% w/v sodium deoxycholate, and 0.1% w/v SDS). After the final wash, the pellet was resuspended in Laemmli sample buffer, and the proteins were resolved by 10% SDS-polyacrylamide gel electrophoresis followed by immunodetection by Western blotting or subjection to in vitro tyrosine kinase assay, as described below.
Western Blotting-Cells were harvested and rinsed twice with phosphate-buffered saline. The cell extracts were prepared with lysis buffer (20 mM Tris, pH 7.5, 0.1% Triton X-100, 0.5% deoxycholate, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin) and cleared by centrifugation at 12,000 ϫ g, 4°C. Total protein concentration was measured using the BCA assay kit (Sigma) with bovine serum albumin as a standard according to the manufacturer's instructions. Cell extracts containing 30 g of total protein were subjected to 10% SDS-PAGE, and the resolved proteins were transferred electrophoretically to polyvinylidene difluoride membranes (Invitrogen). Equal protein loading was confirmed by Coomassie (Bio-Rad) staining of the gels. After blocking with phosphate-buffered saline containing 0.2% casein for 1 h at room temperature, the membranes were incubated with 3-5 g/ml antibody in phosphate-buffered saline containing 0.1% Tween 20 overnight at 4°C. Anti-caveolin-1 antibody was obtained from Zymed Laboratories Inc., and anti-FAK antibody was obtained from Upstate (Waltham, MA). Anti-pTyr925-FAK, C-terminal Src kinase (Csk), c-Src, and pTyr527-c-Src antibodies were obtained from Santa Cruz (Santa Cruz, CA). Anti-actin antibody was obtained from LabVision (Freemont, CA). Chemoluminescent detection (Upstate, Lake Placid, NY) was performed in accordance with the manufacturer's instructions. The densitometric signals were quantified using ImagePro Plus software version 4.0 and normalized to those of actin. The blots were performed in triplicate.
c-Src Tyrosine Kinase Assay-c-Src tyrosine kinase activity was determined using a commercially available kinase assay kit (Sigma) according to the manufacturer's instructions. c-Src immunoprecipitates (20 g of total protein) were prepared using anti-c-Src monoclonal antibody immobilized onto protein G-Sepharose beads (Zymed Laboratories Inc.). Immunoprecipitates were washed and dissolved in tyrosine kinase buffer (final solution containing 0.3 mM ATP) and incubated for 30 min in 96-well plates coated with tyrosine kinase substrate solution (poly(Glu-Tyr)). The phosphorylated substrate was quantified by chromogenic detection using horseradish peroxidase-conjugated anti-phosphotyrosine antibody. Optical densities were determined at 492 nm using a V max microplate spectrophotometer. c-Src kinase activity was compared with an epidermal growth factor-receptor standard. Kinase assays were performed in triplicate with four determinations/condition. siRNA and Transfection-Caveolin-1-specific and scrambled sequence control siRNA were obtained from SuperArray Biosciences (Frederick, MD). siRNAs were dissolved in buffer (100 mM potassium acetate, 30 nM HEPES-potassium hydroxide, 2 nM magnesium acetate, pH 7.4) to a final concentration of 20 M, heated to 90°C for 60 s, and incubated at 37°C for 60 min prior to use to disrupt any higher order aggregates formed during synthesis. The cells were plated onto 35-mm 6-well trays and allowed to adhere for 24 h. In all, 8 l of siPORT Amine (Ambion Inc., Austin, TX) transfection reagent/well was added to serum-free medium (for a final complexed volume of 200 l), vortexed, and incubated at room temperature for 15 min. The transfection reagent-siRNA complexes were added to the wells containing 800 l of medium with 10% fetal bovine serum and incubated in normal cell culture conditions for 6 h, after which 1 ml of DMEM containing 10% fetal bovine serum was added. Assays were performed 48 h post-siRNA transfection.
Anoikis Assay-Anoikis was induced by poly(2-hydroxyethyl meth-acrylate) (poly-HEMA, Sigma) culture. A solution of 120 mg/ml poly-HEMA in 100% ethanol was made and diluted 1:10 in 95% ethanol. 0.95 l/mm 2 of this solution was pipetted into 35-mm wells and left to dry for 48 h at room temperature. Prior to use, the wells were washed twice with phosphate-buffered saline and once with DMEM. 1 ϫ 10 6 cells, suspended in 2 ml of DMEM with 10% fetal bovine serum, were incubated in the poly-HEMA-coated wells for 18 h in a humidified (37°C, 5% CO 2 ) incubator. Following the induction of anoikis, the cells were washed and resuspended in 0.5 ml of binding buffer, and annexin V/fluorescein isothiocyanate/propidium iodide labeling was performed in accordance with the manufacturer's protocol (BD Biosciences). Analysis was performed by flow cytometry (FACScan, BD Biosciences). The anoikis fraction (% cells undergoing anoikis: annexin V/fluorescein isothiocyanate-positive, propidium iodide-negative) was calculated by scoring no less than 10,000 cells. All observations were repeated in triplicate and mean values reported. Statistical Analysis-Differences between groups were analyzed using Student's t test, multifactorial analysis of variance of initial measurements, and Mann Whitney U test for nonparametric data (as appropriate) using Statistica version 5.5 software (StatSoft, Inc., Tulsa, OK).

FIG. 1. CEACAM6 cross-linking increased c-Src kinase activity in CEACAM6 immunoprecipitates.
A, c-Src activity was quantified by in vitro kinase assay following treatment with control (irrelevant) isotype-matched IgG or anti-CEACAM6 antibody either alone or in combination with the secondary cross-linking antibody. Mannosamine (MM) was used to disrupt GPI assembly. CEACAM6 cross-linking significantly increased lysate c-Src kinase activity. This effect was attenuated by prior treatment of the cells with mannosamine. Experiments were preformed three times using four determinations for each condition. *, p Ͻ 0.05 versus cells treated with control IgG and secondary antibody and those pretreated with mannosamine prior to crosslinking. B, inhibitory c-Src phosphorylation at tyrosine 527 was quantified by phospho-specific immunoblotting of the same lysates. In cases where averages were normalized to controls, the standard deviations of each nominator and denominator were taken into account in calculating the final standard deviation. p Ͻ 0.05 was considered statistically significant.

CEACAM6 Cross-linking Increases Cellular c-Src Tyrosine
Kinase Activity and Decreases Tyr-527 Phosphorylation-Tyrosine kinases play a central role in the malignant phenotype of cancer cells (25)(26)(27), and GPI-anchored proteins, including CEACAM family members, modulate intracellular signaling events via tyrosine kinases (28 -30). We examined the effect of CEACAM6 cross-linking on the tyrosine kinase activity of the prototype Src family protein tyrosine kinase, c-Src. BxPC3 cells were chosen for this study as these cells markedly overexpress CEACAM6 (2,31). Following treatment with either anti-CEACAM6 antibody or control IgG (alone or followed by secondary antibody cross-linking), cell lysates were prepared, and c-Src kinase activities in these lysates were quantified by in vitro kinase assay. CEACAM6 cross-linking resulted in a marked increase in c-Src kinase activity, whereas the control antibody did not affect c-Src activity (Fig. 1A). Total levels of c-Src did not differ among groups. Disruption of GPI assembly by pretreatment with mannosamine (24) almost completely abolished activation of c-Src by CEACAM6 cross-linking, indicating a requirement for the GPI anchor.
The activity of c-Src is modulated by the phosphorylation status of residue Tyr-527. Csk exerts important negative regulatory control on c-Src by phosphorylating this tyrosine residue (32), which impairs the ability of c-Src to phosphorylate tyrosine in downstream targets, including FAK (33). We performed phospho-specific c-Src (Tyr-527) immunoblotting on cell lysates following exposure to anti-CEACAM6 or control (irrelevant) isotype-matched IgG, alone or followed by the crosslinking secondary antibody. Lower levels of c-Src Tyr-527 phosphorylation were observed in lysates from CEACAM6 crosslinked cells, consistent with the higher levels of c-Src activity observed in these cell lysates relative to those derived from control non-cross-linked cells. Mannosamine pretreatment preserved c-Src Tyr-527 phosphorylation at a higher level following CEACAM6 cross-linking, consistent with the GPI dependence of the increased c-Src kinase activity following CEACAM6 cross-linking (Fig. 1B). is decreased by CEACAM6 cross-linking. CEACAM6 immunoprecipitates were immunoblotted for caveolin-1 (pTyr14) and total caveolin-1. CEACAM6 cross-linking inhibited caveolin-1 phosphorylation, an effect that was not observed in cells pretreated with mannosamine (MM). B, c-Src co-precipitation with CEACAM6 was not significantly affected by CEACAM6 cross-linking, whereas CEACAM6 cross-linking decreased the degree of Csk co-precipitation with CEACAM6, an effect that was also prevented by mannosamine. Lanes 1, IgG; lanes 2, By114; lanes 3, IgG ϩ secondary IgG; lanes 4, By114 ϩ secondary IgG; lanes 5, By114 ϩ secondary IgG ϩ mannosamine. Densitometric values are means (ϮS.D.) from three independent blots normalized to caveolin-1. *, p Ͻ 0.05 versus cells treated with control IgG and secondary antibody and those pretreated with mannosamine prior to cross-linking.
following CEACAM6 cross-linking, and we therefore hypothesized that cross-linking CEACAM6 may impair Tyr-527 phosphorylation by reducing phosphorylation of caveolin-1, in turn leading to decreased membrane recruitment of Csk and disinhibition of c-Src. We tested this hypothesis by examining the effect of CEACAM6 cross-linking on the phosphorylation status of caveolin-1 using phospho-caveolin-1 (Tyr-14)-specific immunoblotting. We observed that CEACAM6 cross-linking significantly decreases caveolin-1 phosphorylation ( Fig. 2A). Furthermore, when caveolin-1 immunoprecipitates were immunoblotted for Csk, we observed a lower level of Csk immunoreactivity in immunoprecipitates derived from CEACAM6-cross-linked cells compared with those from control cells (Fig. 3), supporting the hypothesis that CEACAM6 cross-linking reduces Csk recruitment through modulation of the phosphorylation status of caveolin-1. Although levels of c-Src association with CEACAM6 were relatively unchanged by CEACAM6 cross-linking, Csk co-precipitation with CEACAM6 decreased (Fig. 2B). These observations may account for the decreased c-Src Tyr-527 phosphorylation and increased c-Src kinase activity detected in lysates derived from CEACAM6-cross-linked cells. Mannosamine inhibited caveolin-1 dephosphorylation and preserved levels of Csk associated with caveolin-1, confirming the dependence of these effects on the GPI anchor.
Caveolin-1 Is Required for CEACAM6 Cross-linking-induced c-Src Activation-In view of the effects of CEACAM6 crosslinking on caveolin-1 phosphorylation, we speculated that caveolin-1 might act as an adaptor between the GPI anchor of CEACAM6 and c-Src, together with its regulator, Csk. We determined whether caveolin-1 is required for the increase in c-Src activity induced by CEACAM6 cross-linking by suppressing caveolin-1 expression using RNA interference. Suppression of caveolin-1 expression of up to 80% was confirmed by Western blotting, 48 h post-siRNA transfection (Fig. 4). Suppression of caveolin-1 expression almost completely abolished the increase in c-Src kinase activity and the decrease in Tyr-527 phosphorylation induced by CEACAM6 antibody cross-linking. c-Src kinase activity and Tyr-527 phosphorylation were unaffected by control siRNA. This observation demonstrates that caveolin-1 is required for the increase in c-Src kinase activity induced by CEACAM6 cross-linking and is consistent with a model in which caveolin-1 acts as an adaptor between GPI and c-Src signaling pathways in a manner similar to that hypothesized to occur in other model systems (36). with either anti-CEACAM6 antibody or irrelevant IgG (followed by secondary antibody), cell lysates were prepared and FAK phosphorylation was quantified by immunoblotting for phosphorylation at Tyr-925, a c-Src phosphorylation site required for Grb SH2/SH3 adaptor protein binding (39). Although irrelevant IgG had no effect on FAK Tyr-925 phosphorylation status, CEACAM6 cross-linking induced a marked increase in FAK (Tyr-925) phosphorylation (Fig. 5). Prior transfection of the Src(K296R/Y528F) dominant negative expression construct reduced cross-linking-induced FAK tyrosine phosphorylation by 73%, relative to transfection with empty vector, pUSEamp(Ϫ) (p Ͻ 0.05) (Fig. 5), which did not affect FAK phosphorylation status, indicating that FAK phosphorylation induced by CEACAM6 cross-linking is c-Src-dependent.
Antibody-mediated CEACAM6 Cross-linking Inhibits Anoikis-To define the functional implications of CEACAM6 cross-linking, we determined its effect on cellular resistance to anoikis (40). Resistance to this subset of apoptosis induced by anchorage-independent conditions is a feature of transformed cells associated with tumorigenesis, invasion, and metastasis (18 -20). Following exposure to anti-CEACAM6 or irrelevant control isotype-matched IgG (either alone or followed by crosslinking secondary antibody), BxPC3 cells were exposed to poly-HEMA culture for 18 h, and anoikis was quantified by flow cytometry. CEACAM6 cross-linking increased cellular resistance to anoikis by 66%, whereas control antibody had no effect on the anoikis fraction (Fig. 6). Treatment with mannosamine significantly attenuated the protective effect of CEACAM6 cross-linking, indicative of the GPI dependence of this effect. Transfection of cells with the Src(K296R/Y528F) dominant negative construct almost completely abolished the effects of CEACAM6 cross-linking, consistent with the c-Src dependence of CEACAM6 cross-linking-induced events in vitro.

DISCUSSION
CEACAM6 is a cell surface glycoprotein that is overexpressed in a range of gastrointestinal neoplasms (1, 3, 41). We have recently shown that modulation of CEACAM6 expression markedly alters the malignant phenotype of pancreatic adenocarcinoma cells (2). The purpose of this study was to examine how CEACAM6, a GPI-linked molecule lacking transmembrane and intracellular domains, is able to modulate intracellular signaling events and influence cellular behavior. Using the CEACAM6-overexpressing cell line, BxPC3, as a model system, we have demonstrated that CEACAM6 cross-linking promotes anoikis resistance and activates c-Src in a caveolin-1-dependent manner. Furthermore, we have shown that c-Src activation is associated with attenuated membrane recruitment of Csk, a negative regulator of c-Src. We have also demonstrated that activation of c-Src by CEACAM6 cross-linking is associated with c-Src-dependent tyrosine phosphorylation of its substrate FAK. Taken together, these observations indicate a potential mechanism through which CEACAM6 could influence important intracellular oncogenic pathways and hence the malignant phenotype.
Despite the absence of transmembrane or cytoplasmic domains, a large number of GPI-anchored proteins have the ability to transduce cellular activation signals (4). Cross-linking of GPI-anchored molecules has been shown to phosphorylate tyrosine residues on a number of intracellular substrates, leading to signal transduction (42). However, the mechanisms mediating signal transduction from an outer leaflet-associated GPI anchor and intracellular targets remain poorly understood. Cross-linking of GPI-linked molecules is a strategy commonly used to examine the effects of GPI-linked protein aggregation, an event that often affects downstream signaling targets via proteins association within lipid rafts. We used the highly specific By144 monoclonal antibody (22, 23) followed by a cross-linking secondary antibody to induce aggregation of CEACAM6. Our observation that CEACAM6 cross-linking decreases caveolin-1 phosphorylation is consistent with work conducted by Lee et al. (34) who showed that targeting GPI-linked protein to the exoplasmic leaflet of caveolar membranes disrupts caveolin-1 phosphorylation. Caveolin-1 may act as an FIG. 6. CEACAM6 cross-linking increases BxPC3 cell resistance to anoikis induced by poly-HEMA culture. Following exposure to anti-CEACAM6 or irrelevant control isotype-matched IgG, either alone or followed by cross-linking secondary antibody, BxPC3 cells were exposed to poly-HEMA culture for 18 h, and anoikis was quantified by flow cytometry after annexin V and propidium iodide staining. CEACAM6 cross-linking increased cellular resistance to anoikis by 66%, whereas the control antibody had no effect on the anoikis fraction. adaptor for the GPI anchor of CEACAM6 as it appears to do in other systems (43,44). Csk is recruited to the membrane by phosphorylated caveolin-1 (35), and Csk negatively regulates c-Src through phosphorylation of tyrosine residue 527, resulting in decreased tyrosine phosphorylation of a variety of downstream targets, including FAK (33). Our observations support this model of c-Src disinhibition leading to FAK activation.
In neutrophils, a range of GPI-anchored proteins interact with and signal via tyrosine kinases (14,30,45). The Src family of nonreceptor tyrosine kinases, most notably (although not exclusively) c-Src, is frequently overexpressed or aberrantly activated in epithelial and nonepithelial cancers and has functions that are critical in primary tumor progression and metastasis (27,46,47). Via interactions with targets such as FAK, c-Src has an important influence on the malignant phenotype (48). FAK represents an important point of convergence for a variety of signaling cascades and has been shown to be tyrosine-phosphorylated and activated by Src kinases (49 -52). Levels of FAK activity, which depend on FAK tyrosine phosphorylation status, correlate with a variety of aspects of pancreatic adenocarcinoma malignant cellular behavior (37,38). FAK Tyr-397 is a potential high affinity binding site for the SH2 domain of c-Src, and FAK Tyr-397 phosphorylation leads to the recruitment of SH2-containing proteins such as Src family kinases (51,53). Phosphorylation of this site facilitates the formation of a FAK-Src signaling complex in which both kinases are activated (54,55) and promotes FAK phosphorylation at Tyr-925 by c-Src (39). The association of Src with phosphorylated FAK has been implicated previously in integrin-receptor signaling (56) and cell spreading mediated by CD44 (57). In our study, we observed that dominant negative Src expression abolishes the increase in anoikis resistance induced by CEACAM6 cross-linking and reduces tyrosine phosphorylation of FAK induced by CEACAM6 cross-linking. Our observations are consistent with previously published evidence suggesting that v-Src induces anoikis resistance through tyrosine phosphorylation of FAK (58). A central role for FAK in anchorage-independent cell survival is further supported by our recent report that post-transcriptional silencing of FAK expression resulted in greater anoikis in pancreatic adenocarcinoma cells (15). Data in the present study support the premise that resistance to anoikis induced by CEACAM6 crosslinking may be the result of c-Src-mediated phosphorylation of FAK, leading to the transduction of cellular survival signals.
Caveolae and caveolin-containing detergent-insoluble glycolipid-enriched rafts appear to function as plasma membrane microcompartments or domains for the preassembly of signaling complexes, keeping them in a basal inactive state. Caveolin-1 is a 21-24-kDa integral membrane protein, first identified as a major substrate of v-Src, and is found to co-precipitate with c-Src (59,60). Both the N-and C-terminal domains of caveolin-1 face the cytoplasm, allowing free interaction with intracellular proteins. Caveolin-1 can be purified as part of a hetero-oligomeric complex containing c-Src, and although the functional consequences of caveolin phosphorylation remain poorly understood, increased phosphorylation of caveolin-1 has been shown to recruit Csk to the cell membrane, resulting in inhibition of c-Src activity through Tyr-527 phosphorylation (33,35). Lee et al. (34) report that targeting of GPI-anchored protein to the exoplasmic leaflet of caveolar membranes disrupts caveolin-1 phosphorylation by c-Src. Caveolin is required for the biogenesis of caveolae, although it does not result in the apical sorting of GPI-anchored proteins (61,62). Recent studies indicate that caveolin is required for the association of Srcfamily kinases with ␤ 1 integrins, and loss of caveolin/␤ 1 integrin association results in loss of ligand-induced FAK phospho-rylation and impaired development of focal adhesion sites (63). Similarly, fibronectin-dependent fyn signaling through the ␣ 5 ␤ 1 integrin leading to mitogen-activated protein kinase activation requires the presence of caveolin-1 (64). Caveolin binds Src family kinases, and the expression of the caveolin-1 c-Src binding domain inhibits c-Src autoactivation in vitro. Kinase binding to caveolin-1 appears to maintain kinases, such as c-Src, in an inactive state. Previous studies indicate that caveolin-1 acts as a membrane adaptor to couple integrins to the tyrosine kinases (43), and our observations are in keeping with this model of caveolin-1 function.
In summary, cross-linking CEACAM6 activates c-Src kinase in a caveolin-1-dependent manner. We have provided evidence that activation of c-Src is associated with dissociation of Csk from the CEACAM6-caveolin-1 membrane complex and that activation of FAK as a result of CEACAM6 cross-linking is a c-Src-dependent process. Further, cross-linking CEACAM6 influences anchorage-independent cell survival. These observations illustrate a potential mechanism through which CEACAM6 may influence intracellular events and cellular behavior.