INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

AN IMPORTANT CHARACTERISTIC of the epithelium is that it forms a functional barrier to prevent the permeation of noxious agents to the internal milieu and, at the same time, to allow passive vectorial transepithelial transport of nutrients, electrolytes, and small solutes. Several disease-associated states, such as ischemic and hypoxic injuries, effectively disrupt the permeability barrier as well as enhance the movement of ions, large solutes, and inflammatory cells across tight epithelial structures (7, 13, 17, 24, 48). Hence, a better understanding of the mechanism underlying the regulation of epithelial barrier function is of potential physiological and pharmacological significance. Although the structure and function of junctional barriers have been extensively studied, the molecular signaling pathways that modulate assembly, disassembly, and maintenance of junctional integrity under various health and disease states appear to be multifactorial and rather complex.

The epithelium contains at least four classes of intercellular junctions: tight junctions (TJs), adherens junctions (AJs), desmosomes, and gap junctions. Among these classes, TJs are the most critical in forming barriers to the diffusion of solute through the paracellular pathway, whereas AJs play a key role in the formation of tight junctions (10). TJs are also essential for the polarization of epithelial cells, since they form a boundary between apical and basolateral plasma membrane domains. Structurally, TJs comprise several transmembrane proteins, such as occludin and members of the claudin family, and peripheral membrane proteins, such as zonula occludens (ZO)-1, ZO-2, ZO-3, cingulin, 7H6, and symplekin (15). The peripheral membrane protein ZO-1 binds to actin filaments, directly or through a linking protein, serving thus as a candidate for coupling perijunctional actin to the paracellular barrier (15). The basic components of AJs in epithelial cells include transmembrane protein E-cadherin and the cytoplasmic proteins -, -, and -catenins, which link E-cadherin to the actin cytoskeleton (9, 21). Among them, E-cadherin is responsible for the correct establishment and maintenance of AJs through a Ca²⁺-dependent homophilic interaction with adjacent cells. Blocking the function of E-cadherin results in a destruction of AJs and subsequent disassembly of TJs (31). On the other hand, catenins tether the cadherin complexes to actin cytoskeleton and have often been investigated as potential cytoplasmic targets for regulation of AJs (19). Together, the expression and modification of these TJ and AJ molecules determine the permeability properties of epithelial barriers toward hydrophilic solutes and plasticity of intercellular junctions.

In addition to the structural interdependence between TJs and AJs, agents that disrupt the actin cytoskeleton can also lead to the disassembly of TJs (2, 22). This observation is consistent with a model in which the establishment of appropriate actin cytoarchitecture is a key factor in the formation of TJs (15). This notion is further supported by the observation that TJs appear to be tethered to the actin filaments (12). Hence, both actin filaments and AJs appear to participate directly or indirectly in the formation and/or maintenance of TJs. While AJ complexes are primarily involved in maintaining cell-cell adhesions between adjacent epithelial cells, TJ structures modulate epithelial barrier function and paracellular permeability.

The results from many laboratories have suggested that a complex set of signal transduction pathways is likely to target and control the junctional properties of epithelial cells. The barrier function of TJs is reportedly influenced by growth factors, extra- and intra-cellular Ca²⁺ levels, protein kinase C, receptor and non-receptor tyrosine kinases, and phospholipase C in different types of epithelial cells (3, 4, 6, 16, 19, 27, 28, 42, 54, 55). In terms of AJs, significant tyrosine phosphorylation of -catenin, -catenin, and p120-catenin is detected in proliferating epithelial cells (35). As epithelial cells reach confluence and undergo the process of contact inhibition, tyrosine phosphorylation of catenins decreases. This observed decrease in tyrosine phosphorylation is correlated with an increased tyrosine phosphatase activity (5, 11). Many receptor and non-receptor tyrosine phosphatases have been coimmunoprecipitated with cadherin-catenin complexes. Thus components of TJs, AJs, plasma membranes, and cytoskeleton are all potential targets for these already-identified or to-be-identified kinases and phosphatases. The biochemical basis of these modulations by signaling molecules is only beginning to be unraveled. The focus of this study is the characterization of the effects of a novel epithelial tyrosine kinase, Etk, on paracellular permeability and junctional protein complexes of a model epithelial barrier, Pa-4.

Etk, also named Bmx, belongs to a new class of cytoplasmic non-receptor tyrosine kinases, Btk/Tec, members of which are expressed in both hematopoietic and nonhematopoietic cells. This family consists of Btk (46, 50), Itk (18, 39), Tec (30), and Etk/Bmx (34, 43) tyrosine kinases, which share homologous structures, including the NH₂-terminal pleckstrin homology (PH) domain, followed by Tec homology (TH), Src homology (SH) 3, SH2, and tyrosine kinase domains. These Tec kinases have been demonstrated to participate in signaling pathways involving a variety of cytokine receptors and antigen receptors. Etk, unlike other members of the Tec family kinases that are mostly hematopoietic cell specific, is preferentially expressed in epithelial cells (34). We have previously demonstrated that Etk directly activates signal transducer and activator of transcription (STAT) 1, STAT3, and STAT5 in salivary epithelial cells by using an estrogen (E₂)-inducible Etk construct (52). To date, the precise biological function of Etk in epithelial cells has not been defined.

To better understand the molecular nature of salivary epithelial cellular responses to the activation of Etk, we have established a model system by stably transfecting rat salivary epithelial Pa-4 cells with an inducible Etk-estrogen receptor (ER) chimeric construct (Etk:ER), which is activated in cells by the ER ligand, -estradiol. Through a combined approach in search of Etk-mediated biological events, we identify here a series of effects of Etk on transepithelial resistance (TER) in parallel with components of AJs and TJs. We have demonstrated that Etk activation is sufficient to elicit an increase in TER, a common gauge of junctional tightness between epithelial cells, and have further shown that this increase was sustained in response to hypoxic challenge. The increased TER elicited by Etk activation was accompanied by changes in actin filaments organization and in the recruitment and/or changes in protein properties of several peripheral and integral membrane proteins involved in TJ and AJ regulation. Together, we postulate that the functional and biochemical modulation of TJ and AJ during normoxia and hypoxia appears to be dependent on an Etk-mediated signaling cascade. Moreover, our results shown herein suggest that the cells with stable expression of Etk:ER represent a unique and useful tool to elucidate the role of Etk in modulating TJ/AJ assembly/disassembly and subsequent responses to hypoxic challenge in salivary epithelial cells as well as in other cell types. These observations are significant in understanding the physiological regulation of epithelial permeability and discerning the mechanism leading to epithelial permeability dysfunction(s) associated with many disease states.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reagents. Rhodamine-phalloidin, phenylmethylsulfonyl fluoride (PMSF), aprotinin, pepstatin A, N-tosyl-L-phenylalanine chloromethyl ketone, leupeptin, N--p-tosyl-L-lysine chloromethyl ketone, and N--p-tosyl-L-arginine methyl ester were all obtained from Sigma (St. Louis, MO). Goat anti-mouse horseradish peroxidase-conjugated secondary antibody and enhanced chemiluminescence (ECL) reagents were from Amersham (Arlington Heights, IL). The cell culture media, sera, and antibiotics were from Life Technologies (Rockville, MD).

Cell culture. The rat parotid epithelial cell line Pa-4, also known as parotid C5 cells (26), was plated on Primaria culture dishes (Falcon) in Dulbecco's modified Eagle's/Ham's F-12 (1:1) medium supplemented with 2.5% fetal calf serum, insulin (5 µg/ml), transferrin (5 µg/ml), epidermal growth factor (25 ng/ml), hydrocortisone (1.1 µM), glutamate (5 mM), and kanamycin monosulfate (60 µg/ml) and was maintained in a humidified atmosphere of 5% CO₂-95% air at 35°C. The Pa-4Etk:ER cells were established by stably transfecting Pa-4 cells with Etk:ER. The tyrosine kinase activity of Etk:ER in Pa-4Etk:ER cells can be further induced by the addition of 1 µM estrogen receptor agonist, -estradiol, to the culture medium, as demonstrated by the autophosphorylation of Tyr-566 of Etk (52). The Pa-4Etk:ER cells were maintained with geneticin (G418; 600 µg/ml) and Dulbecco's modified Eagle's/Ham's F-12 (1:1, phenol red free) medium supplemented with 2.5% charcoal-stripped fetal calf serum plus the aforementioned ingredients. Madin-Darby canine kidney (MDCK) Etk:ER cell clones were established and screened as described previously (52).

Measurements of TER. Epithelial cells were grown on permeable membranes (Clearwell; Costar-Corning, San Francisco, CA) that allow visual monitoring growth of polarized epithelial cells to confluence. Bioelectric parameters of cell monolayers were monitored at predesignated time intervals with a MilliCell ERS screening device (Millipore, Bedford, MA) that can measure spontaneous potential difference (SPD; expressed in mV, taking the apical aspect as reference) and TER (expressed in k · cm²) with chopstick-style electrodes. Background potential difference (PD) arising from the asymmetry of voltage-sensing electrodes and electrical resistance contributed by both the bathing fluids and the filter membrane were measured and averaged by using the values observed at the beginning and end of each set of SPD and TER measurements from two blank filters bathed with the same medium utilized in cultivation of epithelial cells. These background PD and electrical resistance values were subtracted from the raw data of SPD and TER, respectively. With the use of the corrected SPD and TER, equivalent active ion transport rate (I_eq; equivalent short-circuit current) is estimated as SPD/TER, assuming the prevalence of Ohm's law for a given epithelial cell monolayer system.

Confocal fluorescence microscopy. Pa-4 and Pa-4Etk:ER cells were cultured and exposed to 1 µM E₂ for 4 h before they were rinsed with Dulbecco's PBS (DPBS). For -catenin detection, cells were processed by following the procedures reported by Woo et al. (55). Briefly, cells were fixed with 2% paraformaldehyde in PBS, followed by exposure to PBS supplemented with 50 mM NH₄Cl for 5 min, and then permeabilized for 10 min in PBS supplemented with 0.5% Triton X-100 (Tx-100) before being blocked and exposed to anti--catenin antibody and an appropriate secondary antibody. Rhodamine-phalloidin was used to detect actin filaments. For analysis of the effects of hypoxia on F-actin and -catenin, confluent Pa-4 and Pa-4Etk:ER cells were exposed to 1 µM estradiol for 4 h before being exposed to hypoxia for 16 h. Cells were then fixed and processed as described above for detection of F-actin and -catenin.

Isolation and Western blot analysis of soluble and insoluble protein fractions. Cells were cultured on petri dishes, exposed to 1 µM E₂ for 4 h, and then washed twice with warm DPBS. These washed cells were incubated in a cell lysis buffer solution (pH 6.75) containing 0.1 M PIPES, 1 mM EGTA, 1 mM MgSO₄, 2 M glycerol, 1% Tx-100, 1 mM PMSF, 1 µg/ml pepstatin A, 10 µg/ml N-tosyl-L-phenylalanine chloromethyl ketone, 1 µg/ml leupeptin, 1 mM sodium orthovanadate, 10 µg/ml N--p-tosyl-L-lysine chloromethyl ketone, and 10 µg/ml N--p-tosyl-L-arginine methyl ester for 10-15 min. Some experiments utilized the buffer above containing 0.1% Nonidet P-40 (NP-40) in place of Tx-100 to isolate detergent-soluble fractions (51). However, results were comparable when either detergent was used to isolate cytosol and detergent-soluble membranes. After the soluble (cytosolic and detergent-soluble membrane proteins) fraction of cell lysates was harvested by pipetting, the remaining cellular materials on the dish representing the insoluble fraction were scraped into RIPA composed of 1% NP-40, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate (SDS), 1 mM PMSF, 20 µg/ml aprotinin, and 1 mM sodium orthovanadate in PBS. Unless indicated, equal amounts (between 10 and 30 µg) of soluble and insoluble fractions of cell lysates were diluted with 2× SDS sample buffer, resolved on 7.5% polyacrylamide gels, and electroblotted onto Immobilon-P (Millipore) or nitrocellulose membranes. Immunoprecipitation of Etk:ER with an anti-ER antibody (HC-20; Santa Cruz Biotechnology, Santa Cruz, CA) was carried out as described previously (25), followed by immunoblot analyses using anti-phosphotyrosine antibody (4G10; Upstate Biotechnology). Blots were probed with appropriate primary antibodies (occludin, phosphotyrosine, or estrogen receptor) and goat anti-rabbit or anti-mouse secondary antibody, where appropriate, conjugated to horseradish peroxidase before visualization with an ECL detection system.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Activated Etk:ER is translocated to a detergent insoluble pool. The Btk/Tec family of kinases is so far the only known tyrosine kinase family to carry an NH₂-terminal PH domain. The versatile roles of Btk/Tec kinases are reflected by their protein-protein and protein-lipid interaction through the PH domain (for reviews, see Refs. 29 and 58). Accumulating evidence has suggested that PH domains of Btk/Tec family members are able to interact with F-actin (59). The actin cytoskeleton plays an essential role in a variety of cellular processes including cell division, shape, and motility, to name a few. Hence, we explored the possibility of translocation of Etk:ER from the cytoplasm to the membrane upon its activation to further study the role of Etk activation in epithelial cell signaling.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Activation of an inducible Etk-estrogen receptor (ER) chimeric construct (Etk:ER) promotes translocation of the activated chimera to detergent-insoluble fractions. Cells treated with a vehicle () or 1 µM estradiol (E2) (+) were fractionated into detergent-soluble (S) and -insoluble (I) fractions as indicated. Equal amounts of fractionated lysates prepared from the same numbers of cells were immunoprecipitated with an anti-ER antibody, separated by SDS-PAGE, electroblotted to polyvinylidene difluoride membranes, and immunostained by respective antibodies against ERs (top) and phosphotyrosine (pTy; middle). The immunoblot with the actin antibody is also shown (bottom) as a loading control. Similar results were obtained from 6 independent experiments; 1 representative result is shown.

Phenotypic manifestation of epithelial cells that express Etk. To investigate the role of Etk in the regulation of epithelial cell physiology, both Pa-4Etk:ER and parental Pa-4 cells were grown to form confluent monolayers on polyester Clearwell, and their TER values were measured. As shown in Fig. 2, TER in stably transfected and E₂-stimulated Pa-4Etk:ER as well as MDCKEtk:ER epithelial cells reached a higher level than that in corresponding parental cells. The corresponding I_eq for E₂-stimulated parental and Pa-4Etk:ER cell monolayers were 1.85 ± 0.08 and 1.08 ± 0.09 µA/cm², respectively. Because the measurements of both potential difference (SPD) and TER in MDCK cells were too low to give reproducible calculations of I_eq, the calculated I_eq values from MDCK and MDCKEtk:ER cells are not shown. The observed difference in TER between parental and Etk-activated cells was persistent throughout an 8-day period of measurement (data not shown). This suggests tightening of the paracellular seals upon Etk activation. The TER of parental Pa-4 and MDCK cells are quite different in that Pa-4 cells exhibit intrinsically higher TER than do the MDCK cells (Fig. 2). Thus Etk activation enhances TER in epithelial barrier of either leaky or tight nature. The epithelial barrier to the diffusion of hydrophilic solutes through the paracellular pathway is afforded by TJ. Permeation across TJ is not static but is dynamically regulated under physiological environment and under pathophysiological conditions, such as hypoxia. In particular, epithelial cells have been reported to respond to hypoxic stress, rendering the depletion of ATP and causing the loss of TER (12).

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Etk activation increases the transepithelial resistance (TER) of Pa-4 and Madin-Darby canine kidney (MDCK) cell monolayers. Cells were grown on semipermeable polyester Clearwell membranes until a confluent monolayer was established. E2 was added to the culture media at a final concentration of 1 µM 4 h before TER measurements were performed. Results represent means ± SE of 4 independent measurements performed in triplicate of E2-treated parental and Etk chimera stably transfected cells, respectively.

View larger version (20K): [in this window] [in a new window]   Fig. 3.   Relative changes in TER and equivalent active ion transport rate (Ieq) of Pa-4 and Pa-4Etk:ER cell monolayers under hypoxic conditions. The Pa-4 and Pa-4Etk:ER cells were cultured and treated as described in Fig. 1. TER (A) was measured, and the corresponding Ieq (B) was estimated (see text for details) at 0, 4, 8, and 24 h after the beginning of hypoxia treatment (1% O2). Data represent percentage changes compared with TER obtained from time 0, which is designated as 100%, normalized by the TER measured in the normoxia control. Results represent means ± SE of 4 independent measurements of E2-treated parental and Etk chimera stably transfected cells, respectively.

Etk-induced enhancement of TER in response to hypoxia involves regulation of the actin cytoskeleton. One of the injurious effects of hypoxia on cells is to induce actin depolymerization (23, 37). Because the ability of the TJ to form a seal is dependent on the actin filaments organization (for a review, see Ref. 15), we investigated whether the hypoxia-induced reduction in TER might be mimicked by disassembly of actin and, further, whether Etk activation could preserve TER under conditions of actin filaments loss. Latrunculin B, an actin-depolymerizing agent (20), was utilized for this purpose.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   Differential effect of latrunculin B on TER of Pa-4 and Pa-4Etk:ER cell monolayers. The Pa-4 and Pa-4Etk:ER cells were grown and treated with estradiol as described in Fig. 1. Latrunculin B (48 nM) was instilled into the culture media 4 h before the first measurement was performed. TER was measured at 4, 8, 12, and 24 h after latrunculin B was administered. Results represent means ± SE of 4 independent experiments performed in triplicate.

View larger version (113K): [in this window] [in a new window]   Fig. 5.   Etk activation elicits recruitment of -catenin and F-actin to the cell periphery. Pa-4 and Pa-4Etk:ER cells were cultured and processed for immunofluorescence study as described in MATERIALS AND METHODS. After fixation, cells were probed with a mouse monoclonal antibody to -catenin, followed by a goat anti-mouse secondary antibody conjugated to FITC. Rhodamine-phalloidin was used to label F-actin. The distribution of -catenin (green) (top) and the organization of F-actin (red) (middle), as well as the colocalization of -catenin (green) and actin (red) (dual, bottom), are shown in Pa-4 and Pa-4Etk:ER cells. Bar, 10 µm.

View larger version (115K): [in this window] [in a new window]   Fig. 6.   Etk activation prevents the fragmentation of peripheral actin and -catenin elicited by hypoxia. Pa-4 and Pa-4Etk:ER cells were exposed to hypoxia and then fixed and processed as described in MATERIALS AND METHODS for imaging of actin (red) and -catenin (green). After fixation, cells were probed with a mouse monoclonal antibody to -catenin, followed by a goat anti-mouse secondary antibody conjugated to FITC. Rhodamine-phalloidin was used to label F-actin. The distribution of these markers is shown in Pa-4 and Pa-4Etk:ER cells exposed to hypoxia in the absence of estradiol (E2) (top) and those exposed to 1 µM estradiol (+E2) for 4 h before hypoxia (bottom). Arrows indicate regions of discontinuity or disorganization of -catenin and F-actin at the cell peripheries. Bar, 10 µm.

Etk activation reduces the mobility of occludin on SDS-PAGE. To investigate whether comparable changes in elements of TJs were also elicited by Etk activation, we next examined the cellular distribution of tight junctional proteins, occludin, claudin, and ZO-1 in Etk-activated Pa-4Etk:ER and parental Pa-4 cells. Immunofluorescence studies of these cell monolayers demonstrated that occludin was localized at the cell membrane in a comparable manner in both monolayers (Fig. 7). The tight junctional peripheral protein ZO-1 and the transmembrane protein claudin also exhibited a similar distribution in both Pa-4 and Pa-4Etk:ER cells, where no marked changes in ZO-1 or claudin localization were detected as a result of Etk activation (data not shown). Hence, the localization of occludin, ZO-1, and claudin did not appear to be significantly modulated by Etk activation.

View larger version (64K): [in this window] [in a new window]   Fig. 7.   Etk activation does not alter occludin localization. Pa-4 and Pa-4Etk:ER cells were probed with a mouse monoclonal antibody to occludin, followed by a goat anti-mouse secondary antibody conjugated to FITC. Rhodamine-phalloidin was used to label F-actin. The distribution of occludin alone (top) as well as the colocalization of occludin (green) and F-actin (red) in Pa-4 and Pa-4Etk:ER cells (bottom) is shown. Bar, 15 µm.

View larger version (28K): [in this window] [in a new window]   Fig. 8.   Occludin mobility on SDS-PAGE is increased as a result of Etk activation. Equal volumes of fractionated Nonidet P-40 (NP-40)-soluble (S) and -insoluble (I) lysates from the same numbers of cells were loaded onto a 7.5% SDS-PAGE gel, blotted onto nitrocellulose membranes, and probed with an antibody against occludin. One low-molecular-weight (LMW) form of occludin is present predominantly in Pa-4, but not Pa-4Etk:ER, cells. Clusters of intermediate-molecular-weight (IMW) and high-molecular-weight (HMW) forms of occludin are present in detergent-soluble and -insoluble pools of lysates from Pa-4 and Pa-4Etk:ER cells. Similar results were obtained from 6 independent experiments; 1 representative result is shown. Comparable effects were seen when detergent-soluble and -insoluble pools were isolated using Tx-100 rather than NP-40 (data not shown) and when samples were loaded according to equivalent protein content.

Increased TER after Etk activation is dependent on tyrosine kinase activity. Etk is a non-receptor tyrosine kinase. To establish the mechanistic role of Etk in the demonstrated enhancement of TER in Etk:ER-transfected cells, as shown in Fig. 2, we utilized a widely used tyrosine kinase inhibitor, genistein (Fig. 9). Treatment of genistein caused decreases of TER in Pa-4, Pa-4Etk:ER, and E₂-stimulated Pa-4Etk:ER cell monolayers over the 24-h measurements. The rate and extent of genistein-induced TER decreases were barely distinguishable between Pa-4 and Pa-4Etk:ER monolayers in the absence of E₂-treatment. However, decrease of TER in E₂-activated Pa-4Etk:ER cell monolayers in response to genistein at both 1 µM and 50 µM was more pronounced than in those cell monolayers without E₂-activation (Fig. 9). Moreover, genistein-elicited TER decreases in both Pa-4 and Pa-4Etk:ER cell monolayers, except for exposure to extremely high concentrations (e.g., 200 µM) of genistein, were reversible after 16 h of treatment (data not shown). Although genistein is not a specific Etk inhibitor , our results demonstrate that the catalytic activity of tyrosine kinase(s) is necessary to maintain TER in both Pa-4 and Pa-4Etk:ER cell monolayers. The observation that Pa-4Etk:ER cells (with E₂) are more sensitive to genistein than the parental Pa-4 and Pa-4Etk:ER cells (without E₂) is also consistent with our hypothesis that Etk activation enhances epithelial barrier function and that Etk may be the critical tyrosine kinase involved in this signaling pathway in epithelial cells.

View larger version (24K): [in this window] [in a new window]   Fig. 9.   Increased TER observed with Etk activation is diminished by pretreatment with 1 µM (A) and 50 µM (B) genistein. Pa-4 and Pa-4Etk:ER cell monolayers were grown and treated with E2 as described in Fig. 1. TER was measured at 4, 8, and 12 h after the indicated concentrations of genistein were administered. Data represent percentage changes compared with TER obtained at time 0, which is designated as 100%, normalized by the TER measured in corresponding cells maintained in genistein-free cultures. Similar results were obtained from 3 independent experiments; 1 representative result is shown. Values are means ± SE calculated from triplicate data.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

We sought to develop a cell system that could serve as a useful model to provide insights into the function of Etk and to determine cellular events following Etk activation in both physiological and pathophysiological conditions. To this end, we demonstrated conclusively that Etk activation increases TER in both Pa-4 and MDCK epithelial cells under normoxia (Fig. 2). We also showed that prolonged hypoxia without reoxygenation results in a loss of TER and I_eq in the parental Pa-4 epithelial cells (Fig. 3). By contrast, Etk activation resulted in an enhanced TER and, thus, cell protection against injury from prolonged hypoxia, allowing the Pa-4Etk:ER cells to maintain their TER and I_eq under hypoxic stress (Fig. 3).

The increased TER shown in Pa-4Etk:ER cells was associated with changes in the organization and localization of components of AJs, including actin filaments and -catenin (Fig. 5). Moreover, the hypoxia-induced changes in F-actin and -catenin were prevented by prior activation of Etk. Although no prominent changes in cellular localization of constituents of TJs, such as occludin and ZO-1, were detected in Pa-4Etk:ER (Fig. 7 ), the mobility of occludin prepared from Pa-4Etk:ER cells was markedly reduced on SDS-PAGE (Fig. 8), suggesting that Etk activation renders increased occludin phosphorylation.

We also demonstrated that elements involved in the maintenance of TER in both Pa-4 and Pa-4Etk:ER cells are sensitive to the treatment of a known tyrosine kinase inhibitor, genistein (Fig. 9). The literature on the role of tyrosine phosphorylation in TJ and AJ assembly/disassembly is somewhat controversial and inconclusive. For example, as epithelial cells reach confluence and undergo the process of contact inhibition, tyrosine phosphorylation of catenins decreases. This observed decrease in tyrosine phosphorylation in catenins is correlated with an increased association of tyrosine phosphatase activity (5, 11). Considerable circumstantial evidence also implicates tyrosine phosphorylation in the disassembly of cadherin-mediated cell-cell adhesion. Specifically, expression of constitutively activated v-Src oncoproteins, which induce tyrosine phosphorylation of - and p120-catenins and E-cadherin, leads to the loss or weakening of AJs (32, 41). However, very little information is available on the regulation of AJs by other tyrosine kinases such as Etk that may be more involved in the modulation of cell function, rather than proliferation and differentiation afforded by Src tyrosine kinase activation.

Occludin has been a prime target for a number of signaling pathways involved in the regulation of TJs, and the level of occludin tyrosine phosphorylation has been reported to be correlated with the TER level (8, 47). In these reports, increased occludin tyrosine phosphorylation is shown to be associated with the reassembly of TJs after ATP depletion and TJ restoration following "rescue" from Ras-mediated transformation in MDCK cells. We have now demonstrated by functional, immunocytochemical, and biochemical analyses that Etk-activated cascades enhance TJ/AJ assembly under normoxia and hypoxia. However, it is plausible that other phosphorylation targets in addition to occludin exist in the AJ and TJ complexes that may also be putative. The positioning of these junctions is coordinated and stabilized through an association with a continuous band of bundled actin filaments, known as an adhesion belt (27). However, it is unclear how the expression and function of each TJ and/or AJ molecule(s) are regulated to confer the overall epithelial barrier function. A profound Etk-induced reorganization of actin cytoskeleton into bundles of peripherally localized filaments may influence TER, as has been previously suggested (12, 22). In fact, we propose that reorganization and stabilization of actin filaments may be one of the principal functions of Etk, whether or not additional direct regulation of AJ or TJ components occurs. The proposal that Etk serves as a major regulator of actin cytoskeleton is derived from the observations of 1) the dramatically improved barrier function against hypoxic stress (Fig. 3), 2) the observed reorganization of F-actin and -catenin by Etk (Fig. 5), 3) the blockage of effects on F-actin and -catenin in cells exposed to hypoxia (Fig. 6), and 4) the preservation of TER from latrunculin B-treated Pa-4Etk:ER cells (Fig. 4). Moreover, Btk, closely related to Etk, has been shown to colocalize with actin filaments upon stimulation (49, 59). In a separate study, the activated Bmx-green fluorescent protein (GFP) has been demonstrated to be localized in the membrane, while the resting Bmx-GFP is restricted to the cytoplasm (14), further supporting our notion.

Several studies have explored the relationship between hypoxia and disassembly of actin filaments. For instance, hypoxia induces dephosphorylation and activation of actin depolymerization factor (ADF)/cofilin, an actin regulatory protein that mediates cellular actin dynamics (37). Cofilin dephosphorylation is associated with acceleration of actin exchange on filament polymerization as well as loss of F-actin. Enhancing or preserving ADF/cofilin phosphorylation is, therefore, one way of preserving cellular F-actin. Recent work has implicated two LIM kinases in regulation of ADF/cofilin phosphorylation and actin dynamics: LIM- kinase 1 via a Rac-mediated pathway (1, 57) and LIM-kinase 2 via a Rho and/or Cdc42-mediated pathway (40). If Etk-induced protection from hypoxic injury involves prevention of ADF/cofilin-induced actin disassembly, this could occur either by direct activation of LIM-kinases or indirectly through actions on Rho-, Rac- or Cdc42-based signaling pathways. Etk may also act to alter effectors of actin filaments assembly other than ADF/cofilin. The membrane association of activated Etk also implicates the possibility that Etk is involved in Rho/Rac/Cdc42-mediated signaling pathways.

Hypoxic injury results in a disturbance in intra- and intercellular homeostasis ranging from the depletion of intracellular ATP to development of intracellular acidosis, decreased redox buffer capacity, and rearrangement of actin-based cytoskeleton (37). Theoretically, such processes could compromise the integrity of many epithelia, rendering an inability to function as an effective barrier to prevent leakages of small solutes to HMW proteins and to modulate the passage of inflammatory cells. For example, ATP depletion resulting from hypoxia or through usage of metabolic inhibitors in polarized epithelial cells has been demonstrated to cause perturbations in the actin cytoskeleton as well as disruption of TJ structures (38). Although several candidate genes and their regulatory mechanisms identified in response to oxidative stresses induced by intracellular oxygen radical production or cell penetration by oxidants are known, eukaryotic signal transduction and gene regulatory pathways that respond to hypoxia have only begun to unfold. In addition to altering epithelial cell behavior and function, hypoxia initiates a novel gene expression program (38) that culminates in a variety of changes in downstream events. Presumably, the balance between phosphorylation and dephosphorylation status is one of the means to render the ultimate biological manifestation as a result of hypoxia. However, little information on the kinase(s) or phosphatase(s) involved in hypoxia-related responses is available to date.

On the basis of our results, we hypothesize that the Etk activation in Pa-4 and also possibly in MDCK cells may "prime" these cells against hypoxic injury. In particular, Etk activation may upregulate the ATP-producing pathway(s) by improving their stoichiometric efficiency via phosphorylation/activation modalities or by inducing the expression of genes for ATP-supplying glycolytic enzymes. Alternatively, Etk activation may repress the activity or expression of the less required enzymes or pathways that consume ATP. These working hypotheses are consistent with our observations that 1) I_eq was sustained under prolonged hypoxic conditions in cells with activated Etk (Fig. 3B), 2) the injurious effect on TER elicited by latrunculin B was rapidly ameliorated by Etk activation (Fig. 4), and 3) the organization of actin-based cytoskeleton and the assembly of TJ were altered concomitantly with the augmentation of sealing function of TJs in the stimulated Pa-4Etk:ER cells (Figs. 2, 3, and 8). Moreover, our preliminary results demonstrated that Etk activation enhances hypoxia-response element-dependent gene activation (C. Chau and D. K. Ann, unpublished observation), further supporting our notion. The reported role of Etk/Bmx protein in cell survival and apoptosis suggests rather complex functions for this protein. For example, Etk was shown to be essential for interleukin-6-induced neuroendocrine differentiation (34) and for protection against therapy-induced apoptosis (56) in prostate cancer cells. On the other hand, overexpression of Bmx in 32D myeloid progenitor cells resulted in apoptosis in the presence of granulocyte colony-stimulating factor, while cells expressing a kinase dead mutant of Bmx differentiated into mature granulocytes (14). Moreover, Bmx was shown to reconstitute apoptosis in Btk-deficient chicken B cells (44) and to link Src to STAT3 activation in Src-mediated cell transformation (45). Together, these findings imply that Etk/Bmx is more than likely to have distinct functions in different cell types, even in the same cell lineage during stages of development and differentiation or exposure to external environmental stimuli. In accordance with our present results, Etk/Bmx activation may be involved in modulating cell response to extreme environmental conditions, such as maintaining TER during prolonged hypoxic condition, in addition to regulating cell proliferation and differentiation.

In summary, this study provides the first direct evidence demonstrating that activated Etk enhances TER under resting conditions and that it sustains TER as well as maintains barrier function under prolonged hypoxia. On the basis of the data presented herein, we envision two potential signaling pathways by which Etk activation and its downstream events enhance and stabilize epithelial barrier function. As depicted in Fig. 10, we postulate that Etk restores epithelial TER properties by preventing the disassembly of intercellular TJ through a novel signaling pathway, even in the face of hypoxic insult. In the first pathway, tight junctional seal and, thus, TER are enhanced and/or maintained by Etk activation via stabilization of the actin cytoskeleton, as noted above. In the second pathway, in response to hypoxia, Etk acts directly on elements of TJs such as occludin to enhance TJ integrity. To our knowledge, this is the first report on a non-receptor tyrosine kinase, Etk, that enhances epithelial barrier function by biochemical modulation of TJ/AJ components in epithelial cells. Specific mechanisms and target proteins involved in the possible regulation of actin filaments and in the regulation of TJ by Etk activation remain future challenges.

View larger version (11K): [in this window] [in a new window]   Fig. 10.   A putative model for protection against the injurious effect of prolonged hypoxia on TER by Etk activation. A schematic presentation shows possible roles of Etk signaling in the enhancement of TER. At least 2 major pathways are involved: Etk activation leads to actin polymerization and/or prevents the disassembly of intercellular tight junctions under hypoxic conditions. The exact molecular mechanism by which Etk enhances TER and the downstream effector(s) of Etk remains to be established. TJ, tight junction.