Requirement for Ras and Phosphatidylinositol 3-Kinase Signaling Uncouples the Glucocorticoid-induced Junctional Organization and Transepithelial Electrical Resistance in Mammary Tumor Cells*

In Con8 rat mammary epithelial tumor cells, the synthetic glucocorticoid dexamethasone stimulates the remodeling of the apical junction (tight and adherens junctions) and the transepithelial electrical resistance (TER), which reflects tight junction sealing. Indirect immunofluorescence revealed that dexamethasone induced the recruitment of endogenous Ras and the p85 regulatory subunit of phosphatidylinositol (PI) 3-kinase to regions of cell-cell contact, concurrently with the stimulation of TER. Expression of dominant-negative RasN17 abolished the dexamethasone stimulation in TER, whereas, dexamethasone induced the reorganization of tight junction and adherens junction proteins, ZO-1 and β-catenin, as well as F-actin, to precise regions of cell-cell contact in a Ras-independent manner. Confocal microscopy revealed that RasN17 and the p85 regulatory subunit of PI 3-kinase co-localized with ZO-1 and F-actin at the tight junction and adherens junction, respectively. Treatment with either of the PI 3-kinase inhibitors, wortmannin or LY294002, or the MEK inhibitor PD 098059, which prevents MAPK signaling, attenuated the dexamethasone stimulation of TER without affecting apical junction remodeling. Similar to dominant-negative RasN17, disruption of both Ras effector pathways using a combination of inhibitors abolished the glucocorticoid stimulation of TER. Thus, the glucocorticoiddependent remodeling of the apical junction and tight junction sealing can be uncoupled by their dependence on Ras and/or PI 3-kinase-dependent pathways, implicating a new role for Ras and PI 3-kinase cell signaling events in the steroid control of cell-cell interactions.

The apical junctional complex, which consists of the tight junction and the adherens junction, controls intercellular adhesion and the permeability properties involved in epithelial cell-cell interactions. The tight junction, a specialized structure located at the apex of the junctional complex, restricts the lateral diffusion of lipids and membrane proteins, and thereby physically defines the border between the apical and basolateral compartments (1). Moreover, tight junctions form a regu-lated barrier for the diffusion of solutes through the paracellular pathway to control the microenvironment on each side of the epithelium (2). Immediately basal to the tight junction is the adherens junction that is responsible for intercellular adhesion between neighboring cells, a process critical for the proper organization and physiological function of the tissue (3). Both these intercellular junctions have been proposed to associate with the perijunctional actin cytoskeleton and signaling molecules through multiprotein complexes to form an integrated functional unit.
Assembly of the tight junction requires the initial engagement of cell-cell contacts at the adherens junction, a process that is mediated by the calcium-dependent intercellular adhesion between E-cadherin molecules and the formation of an intracellular protein complex that includes E-cadherin, ␣-catenin, ␤-catenin, ␥-catenin (plakoglobin), and the actin cytoskeleton (4,5). Activation of cadherin-mediated cell-cell adhesion triggers a series of molecular events that can lead to the recruitment of tight junction components to the points of cell-cell contact and assembly into a complex to form a functional unit capable of providing a tight paracellular seal to the epithelium (6). The tight junction complex consists of several classes of protein components that includes transmembrane proteins, intracellular peripheral membrane proteins, and potential cell signaling molecules. At present, 3 transmembrane protein families, occludin (7), 8 members of the claudin family (8,9), and junctional adhesion molecule (10), have been identified to reside at the tight junction. Although little is known about the claudin protein family or junctional adhesion molecule, occludin is thought to provide paracellular barrier function through its extracellular domain (11). The cytoplasmic tail of occludin interacts with a complex of related peripheral membrane proteins, ZO-1, ZO-2, and ZO-3, three members of the membraneassociated guanylate kinase (MAGUK) 1 protein family (12). Each of these proteins contain three PDZ (PSD95/SAP90, discslarge, ZO-1) domains (13), a Src homology (SH3) domain, and a region similar to guanylate kinase. It has been proposed that these protein-binding modules allow MAGUK members to coordinate the localization and clustering of transmembrane and peripheral membrane proteins. In doing so, MAGUK proteins at the tight junction may provide a bridge connecting the cytoskeleton or intracellular signaling pathways to transmembrane proteins, such as occludin, thereby regulating the extracellular tight junction seal (14). In this regard, several potential cell signaling molecules, such as PKC- (15), atypical PKC isotype-specific interacting protein (16), the heterotrimeric G protein subunit (17), and the Ras target AF-6 (18), have been localized to the tight junction. Although conceptually intriguing, their functional role in regulating tight junction assembly and/or permeability properties has not been characterized.
The physiological plasticity and tissue-specific regulation of assembly and function of epithelial cell tight junctions implicate a complex set of signal transduction pathways that likely target and control the apical junctional complex. For example, in various cell types the permeability properties of the tight junction can be influenced by growth factors, intracellular calcium, calmodulin, protein kinase C, receptor and nonreceptor tyrosine kinases, heterotrimeric G proteins, lipid second messengers, and phospholipase C (19 -24). Regulatory proteins belonging to the Ras superfamily of small GTPases consisting of the Ras, Rho, and Rab subfamilies, which transduce intracellular signals from a variety of extracellular stimuli, have been proposed to play a role in cell-cell interactions. Recently, Rac and Rho have been shown to be critical for the establishment and maintenance of intercellular adhesion (25,26), and have been implicated as regulators of tight junction assembly and permeability properties (27)(28)(29). Rho is also involved in the sphingosine 1-phosphate induction of cadherins and the formation of well developed adherens junctions in HEK293 fibroblast cells (30). Studies in Drosophila have attributed a role for Rac1 in organizing perijunctional actin at the adherens junction of the wing disc epithelium (31). Moreover, Rac-dependent signaling at cell junctions appears to be cell-type specific, because activated Rac, or its exchange factor, Tiam1, induce invasion of T lymphoma cells but suppress invasion in epithelial cells by increasing its adhesive properties (32). Rab proteins, such as rab13 (33) and rab3B (34), have been found to be concentrated at the tight junctions, although they are generally thought to function in the control of vesicle targeting to the plasma membrane. Aberrant activation of oncogenic Ras proteins in epithelial cells is characterized by mesenchymal/fibroblastic morphology with a perturbation of the adherens junction (35,36). However, little is known about the role of cellular Ras in the regulation of epithelial junctional complex or whether the physiological stimuli that control cell-cell interactions can exert their effects through Ras signaling pathways.
To further elucidate the mechanisms of signal transmission required for the regulation of the tight junction, we have been utilizing the Con8 mammary tumor epithelial cell line to investigate the hormonal control of cell-cell interactions (22). Con8 cells grow as poorly differentiated monolayers, exhibiting deficient cell adhesion and poor tight junction organization (37). Our previous studies have shown that treatment with the synthetic glucocorticoid dexamethasone induces the assembly and function of the tight junction, concurrent with an induction of a G 1 cell cycle arrest (22,38). Dexamethasone stimulates an increase in the transepithelial electrical resistance (TER) of the epithelial monolayer, which directly correlated with a decrease in paracellular permeability to radioactive tracers across the epithelium (22), thus verifying the steroid induction of tight junction barrier properties. In addition, glucocorticoids induce the reorganization of the apical junction leading to the recruitment of tight junction and adherens junction proteins to their respective location (22,37). Given the intimate association of certain growth factor signaling components with structural proteins at the cell junction, we investigated the potential roles of cellular Ras and PI 3-kinase in the cell signaling pathways by which glucocorticoids regulate mammary tumor cell-cell interactions. Our results demonstrate that Ras and PI 3-kinase are recruited to regions of cell-cell contact as a consequence of the glucocorticoid-induced membrane reorganization event. Moreover, inhibition of Ras and PI 3-kinase pathways abolishes or attenuates the glucocorticoid-mediated enhancement of the tight junction seal without altering the remodeling of the cell junction, which suggests that these two events can be uncoupled by their requirements for growth factor signaling pathways. Cell Culture and Measurement of Transepithelial Electrical Resistance-Con8 rat mammary epithelial cells were routinely grown to 100% confluency on Nunc permeable supports in Dulbecco's modified Eagle's medium/Ham's F-12 supplemented with 10% calf serum and penicillin/ streptomycin and maintained at 37°C in a humid atmosphere of air/ CO 2 (95:5). Cell culture medium was routinely changed every 24 h. To generate the control C7 and dominant-negative RasN17 cell lines, the plasmid pMMrasDN was transfected into Con8 cells, together with a plasmid encoding neomycin resistance (pRSV-neo). After selection with the neomycin analogue G418 for 2 weeks, 50 clones were selected, expanded, and tested for their induction of RasN17 by dexamethasone. The formation of tight junctions was monitored by measuring TER using an EVOM Epithelial Voltohmmeter (World Precision Instruments, Sarasota, FL) as described previously (22). Calculations for ohms cm 2 were determined by subtracting the resistance measurement of a blank filter and multiplying by the area of the monolayer (0.49 cm 2 for the 10-mm filters).

DNA Synthesis by [ 3 H]Thymidine
Incorporation-To quantitate relative rates of DNA synthesis, triplicate samples of Con8 cells were grown on 24-well plates and treated with the indicated combinations of dexamethasone, LY294002, and/or wortmannin. The media were replaced with fresh media containing 6 Ci/ml [ 3 H]thymidine (90 Ci/ mmol) and the cells were incubated for 2 h. The cells were washed three times with cold 10% trichloroacetic acid and lysed with 300 l of 0.3 N NaOH. Radioactivity was quantified on a Beckman LS 1801 liquid scintillation counter.
Immunofluorescence Microscopy and Confocal Microscopy-Con8 mammary cells were grown on Nunc filters and incubated with the indicated combinations of dexamethasone, PD 098059, LY294002, and/or wortmannin. The cell monolayers were washed three times with Dulbecco's phosphate-buffered saline (PBS) containing 130 mg/liter CaCl 2 ⅐2H 2 O and 100 mg/liter MgCl 2 ⅐6H 2 O and fixed with 1.75% formaldehyde in PBS for 15 min at room temperature. After three additional washes with PBS, the plasma membrane was permeabilized with 0.5% Triton X-100 in PBS for 10 min. Following three washes, nonspecific areas were blocked with TBST (50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 0.05% Tween 20) with 3% nonfat dry milk. All subsequent incubations with antibodies and washes were performed with this buffer. Cells were incubated with primary antibodies against ZO-1 (1:350 dilution), ␤-catenin (1:500), Ras (Y13-258, 1:300), and p85 (1:200) at room temperature for 2 h and then washed three times. For the secondary reaction, goat fluorescein-or Texas Red-conjugated antibodies were incubated for 1 h at a 1:100 dilution and stained cells were mounted with SlowFade™Light Antifade reagent (Molecular Probes, Inc.). 100 M 4Ј,6-diamidino-2-phenylindole was added to the secondary reaction to visualize nuclear morphology. Stained and mounted cells were then processed with a Zeiss Axioplan epifluorescence microscope using a Zeiss ϫ 40 Plan-Neofluar multi-immersion objective (0.9 NA) or analyzed with a Bio-Rad MRC 600 confocal system. A series of x-y optical sections was collected for each specimen in 0.5-m increments. A Kalman average of 8 frames/image was obtained for each section using a zoom setting of 1.5. A x-z view was acquired by averaging sections over a line at each z position in 0.2 M steps.
Gel Electrophoresis and Immunoblotting-Protein samples were fractionated on a 7.8% SDS-polyacrylamide gel. Proteins were electrophoretically transferred from the gel to nitrocellulose membrane (Micron Separations Inc., Westboro). Blots were blocked in TBST (5% nonfat dry milk) for 2 h at room temperature and incubated in Y13-259 anti-Ras or p85 antibodies (1:1000) overnight at 4°C. After three washes, 15 min each, in TBST (1% non-fat dry milk), the blots were incubated with anti-rat antibodies conjugated to horseradish peroxidase (Bio-Rad) for 1 h at room temperature and washed as above. The blots were developed by NEN Life Science Products chemiluminescence reagent kit. ) and total Ras protein was immunoprecipitated with monoclonal Y13-259 antibody prebound to protein G-Sepharose for 1 h at 4°C. Immunoprecipitates were washed five times with 50 mM Hepes, pH 7.4, 500 mM NaCl, 5 mM MgCl 2 , 0.1% Triton X-100, 0.005% SDS and GTP-bound Ras was eluted in 2 mM EDTA, 2 mM 1,4-dithiothreitol, 0.2% SDS, 0.5 mM GTP, 0.5 M GDP at 68°C for 20 min. The radioactivity released, representing the total initial GTP bound to Ras (GTP ϩ GTP hydrolyzed to GDP prior to immunoprecipitation) was quantitated by scintillation counting.
DNA Laddering to Monitor Apoptosis-Cells were grown to confluence in 100-mm tissue culture dishes and treated with combinations of 1 M dexamethasone, 50 M PD 098059, and/or 50 M LY294002 for 24 h. Cells were then scraped into the media in which they had been incubated (floating cells were combined with the attached cells) and samples were normalized for protein content. Low molecular weight genomic DNA was extracted with 0.5% Triton X-100, 10 mM EDTA, and 10 mM Tris, pH 7.4. After three phenol-chloroform extractions and ethanol precipitation, low molecular weight DNA was analyzed on a 1.5% agarose gel in TAE buffer.

Glucocorticoid Signaling Recruits Ras and the PI 3-Kinase Subunit, p85, to Regions of Cell-Cell Contact-In
Con8 mammary tumor cells, the glucocorticoid stimulation of tight junction formation involves a morphological remodeling of the apical junction, as well as an enhancement of the electrical tightness of the junctional barrier (22). In addition, we have observed that serum stimulates both the basal and glucocorticoid-regulated TER, suggesting that a serum/growth factor signaling pathway can facilitate the establishment of a tight paracellular seal. 2 Given that the tight junction consists of several multifunctional adapter proteins implicated in organizing signaling components into macromolecular assemblies, it is conceivable that the glucocorticoid-induced organization of the cell junction includes the recruitment of cell signaling molecules to the junctional complex. Two such signaling molecules, whose activation and function depend on their associations with the plasma membrane, are the small GTPase Ras and PI 3-kinase, which is comprised of a p85 regulatory and a p110 catalytic subunit (39,40). To determine whether the distribution of Ras or PI 3-kinase is altered during the stimulation of tight junction formation, Con8 cells were grown on filter inserts and treated with the synthetic glucocorticoid dexamethasone over a 4-day time course. During this time course, dexamethasone stimulated the monolayer TER, reflecting the formation of functional tight junctions (Fig. 1A). At the end of the time course, the glucocorticoid-treated and untreated cell monolayers were fixed, permeabilized, and immunostained for Ras localization using the Y13-238 monoclonal antibody and for PI 3-kinase localization using the polyclonal antibody to the p85 subunit of PI 3-kinase (Fig. 1B). Similar to the reorganization of tight junction and adherens junction structural proteins (22,37), dexamethasone induced the translocation of both Ras and PI 3-kinase to the cell periphery, concurrent with an increase in tight junction integrity as indicated by the stimulation of TER. In contrast, in the absence of dexamethasone, Ras and PI 3-kinase exhibited a predominately diffuse staining pattern under conditions in which the monolayer TER remained at 2 P. L. Woo and G. L. Firestone, unpublished data.

FIG. 1. Glucocorticoid stimulation of TER and recruitment of
Ras and the p85 regulatory subunit of PI 3-kinase to regions of cell-cell contact. A, Con8 cells were plated at confluency on filter inserts and cultured in the presence (ϩDex) or absence (ϪDex) of 1 M dexamethasone for 4 days. At the indicated time points, the TER was measured and the ohms cm 2 calculated as described under "Experimental Procedures." B, after fixation and permeabilization, the cells were incubated with either the Y13-238 monoclonal anti-Ras (upper panels) or polyclonal anti-p85 antibodies (lower panels), which were visualized with FITC-labeled goat anti-rat and FITC-labeled goat anti-rabbit antibodies, respectively. As a control, dexamethasone-treated samples were stained with the secondary antibodies alone. Bar, 10 M. C, post-confluent Con8 cells were cultured in the presence or absence of dexamethasone in a 5-day time course. Total cell lysates were subjected to SDS-polyacrylamide gel electrophoresis (6% polyacrylamide), and Western blots were analyzed using the polyclonal anti-p85 antibody. basal levels. The elevated level of junctional p85 staining was not due to an increase in the abundance of p85 since the protein levels did not change in response to dexamethasone (Fig. 1C). Similarly, dexamethasone has no effect on the level of endogenous Ras protein (see next section). These results demonstrate that glucocorticoid treatment recruits both Ras and PI 3-kinase to the cell periphery, suggesting that these cell signaling molecules potentially function as regulators of tight junction dynamics.
Establishment of a Stably Transfected Mammary Tumor Cell Line That Expresses Dominant Negative RasN17-To directly test whether cellular Ras plays a role in the glucocorticoid regulation of tight junction formation, Con8 cells were stably co-transfected with an expression vector encoding the dominant-negative Asn-17 mutant of Ha-ras (RasN17) under the control of the dexamethasone-inducible mouse mammary tumor virus-long terminal repeat promoter along with a plasmid encoding the neomycin resistance gene to select for transfection competent cells. Stable transfectants were selected by their resistance to the cytotoxic effects of G418, and the resulting cell clones were screened for dexamethasone induction of RasN17 protein expression. One such transfected cell clone, DN5, as well as a control clone, C7, were utilized for our studies. Western blot analysis of dexamethasone-treated and untreated cells revealed that DN5 cells rapidly produce high levels of RasN17 by 4 h of glucocorticoid treatment (Fig. 2A). The level of RasN17 continued to increase throughout the 72-h time course in dexamethasone. The C7 cells did not produce detectable levels of RasN17, even after 72 h in dexamethasone, and represents a RasN17-transfected control cell line for many of the subsequent experiments. In addition, the level of endogenous Ras did not significantly change with dexamethasone treatment during this time course.
To confirm that the RasN17 inhibits endogenous Ras function by competing for its GTP exchange factors (41), a GTP binding assay was employed to assess the total level of GTPbound Ras. DN5 cells and the control C7 cells were treated with dexamethasone for 6 h to induce expression of RasN17 protein or were left untreated and the plasma membranes were permeabilized with streptolysin-O to allow [␣-32 P]GTP to enter the cytoplasm. The cells were lysed in the presence of non-radioactive GTP and GDP and then both cellular Ras and RasN17 were immunoprecipitated with the Ras monoclonal antibody Y13-259. The total binding of [␣-32 P]GTP to the immunoprecipitated Ras protein from each cell line is shown in Fig. 2B. After dexamethasone treatment, DN5 cells produced a significant reduction in the level of Ras that is bound to [␣-32 P]GTP which is consistent with the known properties of the dominantnegative RasN17. C7 cells showed only a minor reduction in [␣-32 P]GTP-bound Ras after glucocorticoid treatment. These data indicate that DN5 cells produce a glucocorticoid-regulated form of RasN17 that inhibits the capacity of endogenous Ras to bind GTP, and therefore, represents an effective tool to investigate the functional role of cellular Ras in cell-cell interactions.
Glucocorticoid Stimulation of TER, but Not the Regulated Organization of the Junctional Complex, Is a Ras-dependent Process-To determine whether Ras signaling is required for the formation of functional tight junctions, TER was monitored over a 5-day time course of dexamethasone treatment in monolayers of DN5 and control C7 cells grown to confluency on filter inserts. As shown in Fig. 3, dexamethasone significantly induced the TER in the control C7 cells. By day 5, the dexamethasone-treated C7 cells displayed an approximately 7-fold greater TER value compared with untreated cells. In contrast, in the RasN17-expresssing DN5 cells, dexamethasone failed to induce a significant increase in TER levels. These results implicate a crucial and selective role for Ras signaling in the glucocorticoid-induced paracellular seal of mammary tumor cell tight junctions.
We have previously established that the glucocorticoid stimulation of TER in Con8 mammary tumor cells coincides with the steroid-induced remodeling of the cell junction, resulting in the distribution of the tight junction protein, ZO-1, to precise contact points along the cell periphery (22). One possible explanation for the inhibition of tight junction sealing by domi- nant-negative Ras is that the inhibition of Ras activity may prevent the glucocorticoid-dependent reorganization of the apical junction. To this end, we tested whether the expression of dominant-negative Ras can preclude the glucocorticoid-induced junctional organization. C7 and DN5 cells were treated in the presence or absence of dexamethasone for 4 days and the localization of ZO-1, ␤-catenin, and F-actin was examined by indirect immunofluorescence microscopy. Similar to nontransfected Con8 cells (data not shown), dexamethasone induced the rearrangement of the apical junction by recruiting tight junction and adherens junction proteins in the C7 cells to their respective locations (Fig. 4, left set of panels). In the absence of dexamethasone these junctional proteins were distributed in a disorganized manner, residing discontinuously along cell boundaries. Upon dexamethasone treatment, the junctional proteins were localized precisely at cell-cell contact sites in a smooth, continuous band with a honeycomb-like staining pattern. Surprisingly, even though dexamethasone failed to induce TER levels in the DN5 cells, ZO-1, F-actin, and ␤-catenin still localized to cell junctions in response to dexamethasone (Fig. 4, right set of panels). It is important to note that under these conditions, glucocorticoids had no significant effect on the level of ZO-1, ␤-catenin, or F-actin protein expression in either C7 or DN5 cells (data not shown). Moreover, our results suggest that Ras signaling is not required for the maintenance of intercellular adhesion since ␤-catenin and F-actin are localized to the adherens junction. The proper localization of junctional plaque proteins to cell-cell contacts in the presence of dominant-negative RasN17 suggests that the glucocorticoid-induced membrane reorganization and the subsequent induction of tight junction barrier function can be uncoupled by their differential requirements for Ras signaling.
RasN17 and p85 Co-localizes with ZO-1 and F-actin at the Tight Junction and Adherens Junction-Since the localization of Ras to the inner face of the plasma membrane is mediated by farnesylation at its carboxyl-terminal end (42), presumably, based on its mutation, RasN17 should retain its ability to be lipid modified and localize to the plasma membrane. The inhibition of tight junction sealing by the expression of RasN17 also predicts that a pool of the mutant protein resides and acts in a subcellular compartment at cell-cell contact sites similar to that of endogenous Ras. To investigate this issue further, highdensity cultures of DN5 and C7 cells were grown on filter inserts and treated in the absence or presence of dexamethasone for 72 h. The cells were then fixed, permeabilized, and immunostained for Ras localization using the Y13-238 Ras monoclonal antibody. Consistent with the recruitment of endogenous Ras to regions of cell-cell contact in the Con8 cells, glucocorticoids induced the translocation of Ras in C7 cells to the cell periphery (Fig. 5A). As also shown in Fig. 5A, in DN5 cells, a significantly larger portion of the exogenous RasN17 displayed junctional staining after dexamethasone treatment. These results indicate that as part of the overall membrane reorganization induced by glucocorticoids, both endogenous Ras and exogenous RasN17 is recruited to the lateral junction, which is consistent with our finding that inhibition of Ras activity does not alter junctional remodeling.
Due to the high expression of dominant-negative Ras in the dexamethasone-treated DN5 cells, the precise localization of Cells were fixed, permeabilized, and processed for immunostaining using antibodies to ZO-1 (upper panel) and ␤-catenin (middle panel). F-actin was visualized by incubation with the rhodamine-phalloidin conjugate (lower panels). ZO-1 and ␤-catenin antibodies were visualized with Texas Red-labeled goat anti-rabbit or anti-mouse antibodies, respectively. Bar, 10 m.

FIG. 5. Dominant-negative RasN17 is recruited to cell junctions and co-localizes with ZO-1 and F-actin.
A, filter-grown C7 and DN5 cells were treated with (ϩDex) or without (ϪDex) 1 M dexamethasone for 4 days. After fixation and permeabilization, cells were incubated with the Y13-238 anti-Ras monoclonal antibody. Endogenous Ras (control C7 cells) and dominant-negative RasN17 (DN5 cells) were visualized with the FITC-labeled goat anti-rat antibody. B, DN5 cells were treated with dexamethasone for 4 days and processed for double immunofluorescence for RasN17 and ZO-1 (upper panels) or RasN17 and F-actin (lower panels). Y13-238 anti-Ras and polyclonal anti-ZO-1 antibodies were visualized with FITC-labeled goat anti-rat and Texas Red-labeled goat anti-rabbit antibodies, respectively. F-actin was visualized with the rhodamine-phalloidin conjugate. Confocal images were acquired along the x-y axis (en face view) and overlaying sections stained positive for either ZO-1 or F-actin are shown in the upper panels, representing the tight junction and adherens junction, respectively. The x-z view, in the lower strips, were constructed by averaging sections over a line at each z position in 0.2-m steps. Bar, 10 m RasN17 along the lateral junction could be analyzed by laserscanning confocal microscopy. This approach requires a higher level of protein expression than that observed with endogenous Ras (Fig. 2A). Filter-grown DN5 cells were treated with dexamethasone for 4 days and co-stained for RasN17 and either ZO-1 or F-actin in order to examine the junctional location of the dominant-negative Ras protein. Confocal images were acquired along the x-y axis of the tight junction and adherens junction, as indicated by positive staining for ZO-1 and F-actin, respectively. As shown in Fig. 5B, RasN17 co-localized with both ZO-1 and F-actin along the x-y axis of the cell monolayer. Construction of the x-z views in ZO-1-and F-actin-stained cells (narrow strip below each panel) show that RasN17 co-localized more closely with F-actin along the length of the lateral membrane. Both RasN17 and F-actin displayed a broad staining pattern with enriched areas at the entire lateral junction. Moreover, a subset of RasN17 co-distributed with ZO-1 at the apex of the junctional complex. This data suggests that Ras can associate with tight junction and adherens junction components, independent of its activation state, and transduce signals crucial for regulating permeability properties of the tight junction.
Ras functions as a positive regulator for multiple signaling pathways and one potential mechanism by which Ras can propagate intracellular signals required for tight junction barrier function by glucocorticoids could be through the activation of one or more Ras effectors. For example, PI 3-kinase is a downstream target of Ras implicated in mediating Ras-dependent actin rearrangements, which could potentially be involved in the regulation of cell-cell interactions. In order to assess the localization of PI 3-kinase along the lateral junction, Con8 cells were treated with dexamethasone for 72 h, and then one set of cells co-stained for p85 and ZO-1, and a parallel set of cells were co-stained for p85 and F-actin. As shown in Fig. 6A, optical sectioning of x-y planes showed that p85 co-localizes with both ZO-1 and F-actin at the junctional complex, which suggests that, similar to RasN17, p85 resides at both the tight junction and adherens junction. Construction of the x-z plane shows the precise co-localization of p85 with ZO-1 and F-actin at the tight junction and adherens junction, respectively. One potential mechanism by which PI 3-kinase can be activated by Ras is through the targeting of PI 3-kinase to the plasma membrane, where it can interface with additional activators and phosphorylate its substrates. In dexamethasone-treated DN5 cells, p85 also specifically co-localizes with RasN17 (Fig. 6B), which is unable to interact with Ras downstream effectors, indicating that the recruitment of PI 3-kinase to intercellular junctions is not due to binding of GTP-bound Ras proteins to the p110 catalytic subunit of PI 3-kinase. Taken together, these results show that PI 3-kinase is recruited to the junctional complex by glucocorticoid signaling through a Ras-independent mechanism.
Inhibitors of PI 3-Kinase, Wortmannin, and LY294002, Prevent Optimal Activation of Tight Junctions by Glucocorticoids-A potential role for the association of p85 with the intercellular junctions may be to recruit p110 in proximity to tight junction and the adherens junction proteins whose activities may be affected by the lipid products of PI 3-kinase. To explore this possibility, Con8 cells were treated with two structurally dissimilar inhibitors of PI 3-kinase, wortmannin and LY294002, which compete for the lipid-or the ATP-binding sites on the catalytic p110 subunit, respectively, and thereby inhibit its enzymatic activity. To assess the role of PI 3-kinase in the sealing properties of glucocorticoids at the tight junction, filter grown cells were treated daily with varying doses of LY294002 and wortmannin for 4 days. As shown in Fig. 7A, both PI 3-kinase inhibitors disrupted the ability of dexametha-sone to stimulate TER, with no significant effect on the basal TER levels. Throughout a 96-h time course in 50 M LY294002, dexamethasone failed to stimulate TER in the Con8 cells, while, treatment with 10 M LY294002 reduced the dexamethasone-induced TER by approximately 50%. Treatment with either 10 or 100 nM wortmannin also reduced the dexamethasone-stimulated TER by approximately 50%, although, the basal TER levels were slightly increased by wortmannin over the 96-h time course. Importantly, consistent with our observations using the dominant-negative RasN17, the glucocorticoid induction of ZO-1 localization to cell junctions was still observed in the presence of either PI 3-kinase inhibitor (Fig. 7B), although stimulation of TER was impaired. These results suggest a role for lipid products of PI 3-kinase in potentiating the tight junction response by glucocorticoids. Incubation of these mammary tumor cells with varying concentrations of either PI 3-kinase inhibitor for 48 h inhibited DNA synthesis as monitored by the incorporation of [ 3 H]thymidine (Fig. 7C), thus showing that these inhibitors were effectively blocking a PI 3-kinase proliferative pathway. We have previously established that glucocorticoids suppress the growth of Con8 mammary tumor cells under conditions in which tight junctions are

Inhibition of Both the MEK/MAPK and PI 3-Kinase Pathways Disrupt the Glucocorticoid-induced Tight Junction Seal-
The best characterized Ras-mediated pathway is the protein kinase cascade that ultimately leads to the activation of mitogen-activated protein kinase (MAPK). Ras, in its GTP-bound state, recruits the Raf kinase to the plasma membrane where it is activated to phosphorylate and activate MEK (MAPK/ERK kinase), which in turn phosphorylates and activates the ERK1 and ERK2 members of the MAPK family (39). The specific pharmacological inhibitor of MEK, PD 098059 was employed to assess directly the role of the MAPK pathway in the glucocorticoid-enhanced tight junction function. Con8 cells were cultured on filters and treated with or without dexamethasone in the presence or absence of PD 098059. Treatment with the MEK inhibitor limited the dexamethasone stimulation of TER to approximately 50% the level observed in the absence of PD 098059, while treatment with the inhibitor alone resulted in slightly higher TER values (Fig. 8A). Under these conditions, production of the phosphorylated active form of MAPK was virtually abolished in the presence of PD 098059 (data not shown). Consistent with the results with the dominant-negative Ras expressing cells, PD 098059 had no effect on the dexamethasone-induced localization of ZO-1. These data indicate that the MEK/MAPK pathway represents a potential downstream component of Ras that facilitates the stimulation of the electrical tightness of the paracellular barrier by glucocorticoids.
The combined effects of the MEK inhibitor, PD 098059, with PI 3-kinase inhibitors, LY294002 or wortmannin, on the glucocorticoid-stimulated TER were examined in Con8 cells over a 96-h time course. As shown in Fig. 8B, treatment with 50 M PD 098059 with either 100 nM wortmannin or 10 M LY294002 cooperated to prevent the dexamethasone stimulation of TER, whereas, treatment with any of the inhibitors alone caused an approximate 50% reduction of TER. Moreover, when added to cells in which tight junctions have been formed after a 2-day treatment in dexamethasone, a combination of 50 M PD 098059 and 10 M LY294002 or 50 M LY294002 alone were able to reduce the TER back to basal levels (data not shown).
Recent studies have implicated PI 3-kinase as a critical component of a survival pathway downstream of cell surface receptor (43), whereas, glucocorticoids can inhibit involution and programmed cell death in the mammary gland (44). In order to rule out the possibility that the effects of the PI 3-kinase or MAP kinase inhibitors on TER was an indirect consequence of apoptosis, nuclear morphology of the Con8 cells was assessed by 4Ј,6-diamidino-2-phenylindole staining. As shown in Fig.  9A, cell monolayers that were untreated (Unt) or treated with 1 M dexamethasone (Dex) for 4 days did not contain fragmented nuclei. Similarly, treatment with 50 M PD 098059 (PD) or a combination of dexamethsone and PD 098059 (Dex ϩ PD) also did not display fragmented nuclei. In contrast, 4-day treatment with 10 M LY294002 (LY) induced chromatin condensation in a small number of cells in the monolayer whereas, a larger number of nuclei of cells treated with a combination of PD 098059 and LY 294002 (PD ϩ LY) displayed nuclear condensation and fragmentation that is indicative of apoptosis (Fig. 9A). Significantly, nuclei of cells treated with dexamethsone in the presence of LY 294002 (Dex ϩ LY and Dex ϩ PD ϩ LY) appeared normal.
Cell viability was also examined by DNA laddering analysis. 50 M LY294002 was added to cells in the absence or presence of dexamethasone for 24 h and assayed for the presence of DNA fragmentation. As shown in Fig. 9B, DNA fragmentation was only observed in cells treated with 50 M LY294002 or combinations of 50 M LY294002 and 50 M PD 098059. Importantly, dexamethasone treatment significantly diminished or inhibited the apoptotic effects caused by these inhibitors, and thus, the effects of the cell signaling inhibitors on TER is not a fortuitous consequence of general cell lysis. Taken together, our results implicate that both the PI 3-kinase and MEK/MAPKdependent pathways are required for the maximal Ras-dependent induction of TER by glucocorticoids. DISCUSSION A complex array of intracellular cell signaling pathways cooperate to regulate the adhesive and permeability properties of the adherens and tight junctions by coordinately targeting components of the apical junctional complex. As a result, dynamic changes in cell-cell interactions can occur in response to specific extracellular stimuli that are involved in the development, organization, and function of differentiated epithelia. In Con8 mammary epithelial tumor cells, glucocorticoid receptor signaling induces the remodeling of tight junction and adherens junction proteins from a disorganized distribution to an organized state typified by a continuous belt of staining surrounding each cell. Subsequently, glucocorticoids then stimulate the barrier property of the tight junction resulting in the increase in TER. Our results have uncovered a previously uncharacterized cross-talk between glucocorticoid receptor and growth factor receptor signaling pathways in which Ras-dependent signals are required for glucocorticoids to enhance the integrity of the tight junction (barrier function) at a step after the steroid-regulated remodeling of the apical junctional complex (see model in Fig. 10). Following the inhibition of cellular Ras function, glucocorticoids effectively reorganize the distribution of ZO-1, ␤-catenin, and F-actin to sites of cell-cell contact, whereas, the glucocorticoid-mediated stimulation of TER was abolished. Treatment with inhibitors of MEK or PI 3-kinase, two known downstream components of Ras effector pathways selectively attenuated the glucocorticoid enhancement of an electrically tight cell monolayer. Thus, the glucocorticoidmediated remodeling of the apical junction and the barrier function of the tight junction can be uncoupled by their dependence on Ras signaling and its downstream effector pathways. The signal transduction system that regulates tight junction permeability is relatively uncharacterized, although the presence of three members of the MAGUK protein family, ZO-1, ZO-2, and ZO-3, suggests that a complex network of small molecule-protein and protein-protein interactions occur in a sequential and regulated manner. A pressing challenge in the study of tight junction dynamics has been to understand the regulatory pathways by which the tight junction structure can respond to extracellular stimuli. Increasing evidence suggests that the specificity of distinct cellular responses is contingent upon the correct spatial organization of a defined repertoire of cell signaling components localized at the junctional complex. Our results demonstrate that Ras and PI 3-kinase are recruited to and highly concentrated at regions of cell-cell contact during the glucocorticoid-dependent reorganization of the intercellular junction in mammary tumor cells. Earlier studies have shown that Ras is localized to the cell periphery in v-Ha-Ras-transformed and v-Ki-Ras transformed Madin-Darby canine kidney cells (35,45). Our results further show that the dominant-negative RasN17 is also recruited to the junctional complex and specifically co-localizes with ZO-1 and F-actin at the tight junction and adherens junction, respectively, demonstrating that the inability of RasN17 to bind GTP and, thus its effectors, does not alter its proper distribution to the plasma membrane. That the p85 subunit of PI 3-kinase also co-localizes with ZO-1 and F-actin in dexamethasone-treated Con8 cells raises the question of whether the recruitment of PI 3-kinase to the junctional complex is through the direct binding to activated Ras. Our results demonstrate that the PI 3-kinase translocation to cellular junctions occurs in a Ras-independent mechanism, since p85 was still capable of co-localizing with dominant-negative RasN17 along the lateral junction. However, it remains possible that the mechanism by which Ras and PI 3-kinase are localized to the junction is not mutually exclusive, and that normally Ras can activate PI 3-kinase once they are near each other.
The biological activity of both Ras and PI 3-kinase is dependent on their correct recruitment to the plasma membrane. A likely explanation for the targeting of Ras and PI 3-kinase to the lateral junction is for the proper presentation of these cell-signaling molecules to their upstream activators. Consistent with this mechanism, growth factor receptors, such as the scatter factor/hepatocyte growth factor receptor and epidermal growth factor-receptor, function and are concentrated along the lateral junction (46,47). The adaptor protein, Shc, which mediates the association of tyrosine kinase receptors with the Grb-2/Sos complex involved in Ras activation can also interact with phosphotyrosyl residues on cadherins through its SH2 domain (48). In addition, activation of the Ras guanine nucleotide exchange factor, Ras-GRF2, by calcium influx caused its recruitment to intercellular junctions in kidney epithelial cells (49). PDZ-containing adaptor proteins, such as those of the ZO-1 family, may serve as molecular scaffolds to selectively assemble cell signaling molecules at specialized junctions. Evidence to support this model has been described in Caenorhabditis elegans, in which a complex of PDZ-containing proteins, one of which is a member of the MAGUK family, mediates the proper localization of the epidermal growth factor receptor LET-23 to the basolateral junction through direct protein-protein interaction (50). In addition, a recently cloned PDZ-containing protein in Drosophila, CNK, is thought to assemble signaling molecules in the Ras pathway to cell-cell contact regions (51). Thus, it is becoming increasingly apparent that diverse cellular responses affecting cell-cell interactions are mediated by distinct sets of regulatory proteins that are directly associated with the apical junction.
The requirement for Ras and PI 3-kinase activity for the glucocorticoid induction of tight junction sealing indicates that a select subset of downstream targets reside at the tight junction to regulate paracellular permeability. In this regard, PKC-, which can directly bind to GTP-bound Ras (52) and be activated by the PI 3-kinase target, 3-phosphoinositidedependent protein kinase-1 (53,54), have been shown to be localized specifically at the tight junction in Madin-Darby canine kidney and Caco-2 epithelial cells (15), as well as in the Con8 mammary tumor cells, which have established well formed tight junctions by glucocorticoid treatment. 3 However, the role of PKC-in the regulation of paracellular permeability has yet to be determined. In addition, the Ras effector AF-6 is specifically localized to the tight junction and may provide a link to the cytoskeleton by directly interacting with F-actin and ZO-1 (18). PI 3-kinase has also been shown to mediate Ras-dependent actin rearrangement (55). The potential role of lipid products of PI 3-kinase as direct activators of tight junction proteins may represent a unique function for these secondary messengers. However, the fact that inhibitors of PI 3-kinase cannot completely mimic the dominant-negative Ras repression of tight junction activation suggests that PI 3-kinase may function in concert with other signaling pathways to control permeability properties.
Inhibition of MAPK activation by treatment of the Con8 mammary tumor cells with the MEK inhibitor, PD 098059, prevented the full induction of TER by glucocorticoids. This result implicates a requirement for MAPK signaling in the glucocorticoid-stimulated tight junction sealing. It is interesting to note that other studies have shown that MAPK may be involved in the disassembly of adherens junctions by hepatocyte growth factor in Madin-Darby canine kidney cells (56) and decreased expression of adherens junction components in PC12 cells (57). One possible explanation for our observation that MAPK is required for the activation of tight junctions in mammary tumor cells by glucocorticoids is the alteration in the cellular location of MAPK. In the absence of glucocorticoids, MAPK is phosphorylated and translocated to the nucleus in serum-treated Con8 cells, whereas, dexamethasone treatment of the mammary tumor cells maintains MAPK primarily in a cytoplasmic compartment. 4 10. Model of the glucocorticoid receptor-regulated process that controls apical junction remodeling and tight junction sealing. We propose that glucocorticoid receptor signaling events that regulate tight junction dynamics induce the remodeling of the apical junction, which serves to recruit structural and cell signaling proteins to the cell junction. In a subsequent step, the tight junction sealing process requires the Ras, MEK/MAPK, and PI 3-kinase pathways.
is tempting to speculate that MAPK could play a role in regulating the phosphorylation and function of tight and/or adherens junction components. Intriguingly, MAPK has been shown to phosphorylate the gap junction protein connexin 43 (58) that interacts with the ZO-1 tight junction protein (59). In addition, it has been reported that MAPK can phosphorylate and activate the myosin light chain kinase (60), providing a possible mechanism in which MAPK can regulate the myosin ATPasemediated contraction of the perijunctional actomyosin belt to influence tight junction integrity. We have also shown that inhibition of both MAPK and PI 3-kinase pathways cooperate to prevent the glucocorticoid stimulated electrical tightness of the mammary tumor cell tight junctions. The combination of kinase inhibitors did not indirectly inhibit tight junction permeability due to apoptosis, because glucocorticoids were able to provide a protective effect from the inhibition of cell survival pathways caused by these agents. These results suggest that the requirement for Ras in the barrier function of tight junctions induced by glucocorticoids may involve at least two downstream pathways, PI 3-kinase and MAPK.
At present, it is unclear how each tight junction molecule is regulated to provide barrier function to the epithelia. One hypothesis is that the phosphorylation of the transmembrane protein, occludin, dictates the permeability properties of epithelial cells (61,62). Further support for the role of a kinase in the regulation of tight junctions is found in studies utilizing ATP-depletion experiments, which abolishes the barrier function of the tight junction without altering ZO-1 distribution (63). Consistent with our observations in dominant-negative Ras expressing cells, other studies have also found that the formation of a continuous junctional belt of tight junction proteins along the cell periphery does not always correlate with the establishment of electrically tight epithelia. For example, expression of dominant-negative forms of RhoA and Rac1 can induce a leaky tight junction without an apparent effect on the distribution of ZO-1 or occludin (28). We have previously shown that enhancement of tight junction sealing in a non-transformed mammary epithelial cell line, 31EG4, by glucocorticoids occurs without a change in ZO-1 localization (24,64,65). Evidence for a functionally defective tight junction without alterations in tight junction morphology has also been described in rat models of colitis in which tight junction permeability increased in the intestinal and biliary epithelia without structural changes in the tight junction (66). Our results have dissociated two key events involved in the glucocorticoid regulation of tight junction dynamics in mammary tumor cells, the first being the induced organization of the junctional complex that occurs in a Ras-independent manner, and the second being the Ras-dependent process that leads to an increase in barrier function of the tight junction. We are currently attempting to characterize the downstream targets of Ras signaling that mediate this novel convergence point between glucocorticoid receptor and growth factor receptor cell signaling that control cell-cell interactions of mammary epithelial cells.