INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

A CRITICAL CHARACTERISTIC of the epithelium of the gastrointestinal (GI) mucosa is its ability to maintain a highly selective permeability barrier. This barrier normally restricts the passage of harmful proinflammatory and toxic molecules (e.g., bacterial endotoxin) into the mucosa and systemic circulation. Loss of mucosal barrier integrity, on the other hand, is characteristic of inflammatory bowel disease (IBD), necrotizing enterocolitis, and a variety of other GI disorders (e.g., gastric mucosal injury induced by ethanol) as well as several systemic disorders (e.g., alcoholic liver disease) (28, 35, 36). These disorders are also associated with oxidative tissue injury. The latter is of significant clinical importance because conditions of oxidative stress are common; they can cause mucosal barrier dysfunction and the initiation and/or perpetuation of mucosal inflammation and injury and have been implicated in the GI disorders mentioned above (28, 35, 36).

Because enhancing our knowledge of endogenous GI barrier protective mechanisms might provide insights into developing more effective treatment regimens for these oxidative inflammatory disorders, we have been investigating the protective mechanisms used by growth factors. Using monolayers of human Caco-2 cells exposed to oxidants as a model of barrier disruption, we previously found that epidermal growth factor (EGF) and transforming growth factor- protect intestinal barrier integrity by stabilizing the microtubule cytoskeleton (5-7), in large part by increasing the activity of protein kinase C (PKC) (7). In other studies (5-10), we have shown that the stability of the cytoskeleton is key in mucosal healing under in vivo as well as in vitro conditions. Because PKC consists of a family of at least 12 different serine/threonine kinases of fundamental importance in signal transduction (16, 57, 58), we investigated which isoforms mediate EGF protection in intestinal cells. Intestinal epithelial cells, including Caco-2 cells, express at least five of these isoforms: PKC-, PKC-1, PKC-2, PKC-, and PKC- (1, 7, 16, 42, 58, 64). These isoforms differ in their activation, tissue expression, intracellular distribution, and substrate specificity, suggesting that each isozyme has a unique, nonredundant role in signal transduction (1, 30, 32, 41, 44, 46, 52, 54, 55). Our recent observations of naive Caco-2 cells suggest that EGF induces the membrane association (activation) of an abundant isoform of PKC, PKC-1 (7). In the current investigation, using molecular biological approaches (i.e., transfection), we tested the hypothesis that EGF-induced protection against oxidant injury to both the microtubule cytoskeleton and the barrier integrity of epithelial monolayers depends on translocation and activation of the PKC-1 isoform.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell culture. Caco-2 cells, which were obtained from the American Type Culture Collection (ATCC, Manassas, VA) at passage 15, were chosen because they form monolayers that morphologically resemble small intestinal cells, with defined apical brush borders, junctional complexes, and a highly organized microtubule network (10, 21, 50). Cells were maintained at 37°C in complete DMEM in an atmosphere of 5% CO₂ with 100% relative humidity. Naive-type cells or transfected cells (see Stable transfection) were split at a ratio of 1:6 on reaching confluence and were set up in 6- or 24-well plates for experiments or in T-75 flasks for propagation. Cells grown for the barrier function experiments were split at a ratio of 1:2 and seeded at a density of 200,000 cells/cm² into 0.4 µM BioCoat collagen I cell culture inserts (0.3-cm² growth surface; Becton Dickinson Labware, Bedford, MA), and experiments were performed at least 7 days postconfluence. The medium was changed every 2 days. The utility and characterization of this cell line has been reported previously (10, 21, 50).

Plasmids. The sense and antisense plasmids of PKC-1 were constructed as previously described (16). Expression was controlled by the -actin promoter, which is known to be expressed in Caco-2 cells (50). The antisense PKC-1 plasmid (p-actin SP72-As-PKC-1) was constructed by ligating the 2.3-kb EcoR I fragment of PKC-1 cDNA from pJ6-PKC-1 (16) into the unique EcoR I sites of the p-actin SP72 vector. The antisense orientation of the plasmid was confirmed by SamI restriction digestion (16).

Stable transfection. Cultures of Caco-2 cells grown to 50-60% confluence were cotransfected with G418 resistance plasmid and expression plasmids encoding either PKC-1 or antisense PKC-1 with the use of LIPOFECTIN (GIBCO). Control conditions included vector alone. Briefly, cells were incubated for 16 h at 37°C with the plasmid DNA in serum-free medium in the presence of LipofectAMINE (25 µl/25-cm² flask). Subsequently, the DNA-containing solution was removed and replaced with fresh medium containing 10% fetal bovine serum (FBS) to relieve cells from the shock of exposure to the serum-free medium. After transfection, cells were subjected to G418 selection (0.6 mg/ml) over 4 wk. Resistant cells were maintained in DMEM-FBS and 0.2 mg/ml of G418 (selection medium). PKC protein expression or the lack of it was verified by Western blot analysis of cell lysates (see Fractionation and Western immunoblotting of PKC). Multiple clones stably overexpressing PKC-1 or lacking PKC-1 were assessed by immunoblotting, plated on Transwell cell culture inserts, and allowed to form confluent monolayers that were subsequently used for experiments.

Experimental design. In the first series of experiments, postconfluent monolayers of naive Caco-2 cells were preincubated with EGF (1 or 10 ng/ml) or isotonic saline for 10 min and then exposed to oxidant (0.5 mM H₂O₂) or vehicle (saline) for 30 min. As we have previously shown (5-7), H₂O₂ at 0.5 mM disrupts microtubules and barrier integrity; EGF at 10 ng/ml (but not 1 ng/ml) prevents this disruption. These experiments were then repeated with cell monolayers that were either stably overexpressing or almost completely lacking PKC-1. Reagents were applied on the apical side of the monolayers unless otherwise indicated. Because our previous studies (6, 7) showed that regardless of whether apical or basolateral exposure of oxidants was used the results were qualitatively similar, all current studies used apical application. In all experiments, barrier function, microtubule cytoskeletal stability (cytoarchitecture, tubulin assembly/disassembly), and PKC-1 subcellular distribution were assessed. Additionally, because our previous studies demonstrated that protection is observed only when the protective agent (e.g., EGF or a PKC activator) is added before exposure to the damaging agent (e.g., H₂O₂ or ethanol), all of the current experiments followed this preincubation protocol (5-7, 10).

Fractionation and Western immunoblotting of PKC. Differentiated cell monolayers grown in 75-cm² flasks were processed for the isolation of the cytosolic, membrane, and cytoskeletal fractions as previously described by others and by us (1, 7, 16). Briefly, after treatments, postconfluent monolayers were scraped and ultrasonically homogenized in Tris · HCl buffer [20 mM Tris · HCl (pH 7.5), 0.25 mM sucrose, 2 mM EDTA, 10 mM EGTA, and 2 µg/ml each of aprotinin, pepstatin, leupeptin, and phenylmethylsulfonyl fluoride (PMSF)]. The homogenates were then ultracentrifuged (100,000 g for 40 min at 4°C), and the supernatant was removed and used as a source of the cytosolic fraction. Next, pellets were washed with 0.2 ml of Tris · HCl buffer, resuspended in 0.8 ml of a buffer containing 0.3% Triton X-100, and maintained on ice for 1 h. The samples were then centrifuged (100,000 g for 1 h at 4°C), and the supernatant was used as the source of the membrane fraction. To this remaining pellet, 0.3 ml of cold (4°C) lysis buffer (150 mM NaCl, 50 mM Tris · HCl, 1 mM EDTA, 1 mM EGTA, 1% Nonidet P-40, 0.1% sodium deoxycholate, 0.1% SDS, and 2 µg/ml each of aprotinin, pepstatin, leupeptin, and PMSF) were added. The samples were then placed on ice for 1 h and ultracentrifuged as above. The remainder of the lysate or Triton-insoluble cytoskeletal fraction was then removed. Protein content of the various cell fractions was assessed by the Bradford method (15). For total PKC extraction, which provided the fraction used to assess total PKC-1 expression, scraped monolayers were placed directly in 1.5 ml of cold lysis buffer and subsequently ultracentrifuged as described above. The supernatant was used for bulk protein determination.

Immunofluorescent staining and high-resolution LSCM of microtubules. Cells from monolayers were fixed in cytoskeletal stabilization buffer and then postfixed in 95% ethanol at 20°C as we previously described (5-7, 9, 10). Cells were subsequently processed for incubation with a primary antibody, monoclonal mouse anti--tubulin antibody (IgG1, rat/human reactive; Sigma, St. Louis, MO) at 1:200 dilution for 1 h at 37°C. Slides were washed three times in Dulbecco's phosphate-buffered saline (D-PBS) and then incubated with a secondary antibody (FITC-conjugated goat anti-mouse; Sigma) at a 1:50 dilution for 1 h at room temperature. Slides were washed thrice in D-PBS, once with deionized H₂O, and subsequently mounted in Aquamount. All antibodies were diluted with D-PBS containing 0.1% BSA. After staining, cells were observed with an argon laser ( = 488 nm) with a ×63 oil immersion plan-apochromat objective (NA 1.4, Zeiss). Single cells and/or a clump of two or three cells from desired areas of the monolayers were processed with image processing software on a Zeiss ultra high-resolution laser scanning confocal microscope to create "neat black" areas surrounding the cells. The cytoskeletal elements were examined in a blinded fashion for their overall morphology, orientation, and disruption as we previously described (5-7, 9, 10, 66). The slides were decoded only after examination was complete.

Microtubule (tubulin) fractionation and quantitative immunoblotting of tubulin assembly and disassembly. Polymerized (S2) and monomeric (S1) fractions of tubulin were isolated as we previously described (5-7, 10). Cells were gently scraped and pelleted with centrifugation at low speed (700 rpm, 7 min, 4°C) and resuspended in microtubule stabilization-extraction buffer (0.1 M PIPES, pH 6.9, 30% glycerol, 5% DMSO, 1 mM MgSO₄, 10 µg/ml of anti-protease cocktail, 1 mM EGTA, and 1% Triton X-100) at room temperature for 20 min. Tubulin fractions were separated after a series of centrifugation and extraction steps. Specifically, cell lysates were centrifuged at 105,000 g for 45 min at 4°C, and the supernatant containing the soluble monomeric pool of tubulin (S1) was gently removed. The remaining pellet was then resuspended in 0.3 ml of a Ca²⁺-containing depolymerization buffer (0.1 M PIPES, pH 6.9, 1 mM MgSO₄, 10 µg/ml of an anti-protease cocktail, and 10 mM CaCl₂) and incubated on ice for 60 min. Subsequently, samples were centrifuged at 48,000 g for 15 min at 4°C, and the supernatant (S2 fraction or cold Ca²⁺-soluble fraction) was removed. To ensure the complete removal of the S2 fraction, the remaining pellet was treated with the Ca²⁺-containing depolymerization buffer twice more by resuspension and centrifugation. The "microtubules" were recovered by separately incubating (at 37°C for 30 min) the S1 and S2 fractions with the stabilizing agents taxol (10 µM) and GTP (1 mM) in microtubule stabilization buffer (0.1 M PIPES, pH 6.9, 30% glycerol, 5% DMSO, 10 µg/ml of anti-protease cocktail, 1 mM EGTA, 1 mM MgCl₂, and 1 mM GTP) to promote the polymerization of tubulin. Tubulin was then recovered by centrifugation and resuspended in the stabilization buffer. Fractionated S1 and S2 samples were then flash-frozen in liquid N₂ and stored at 70°C until being immunoblotted. For immunoblotting, samples (5 µg protein/lane) were placed in a standard SDS sample buffer, boiled for 5 min, and then subjected to PAGE on 7.5% gels. Procedures for Western blotting were performed as previously described (5-7, 10). To quantify the relative levels of tubulin, the optical density of the bands corresponding to immunoradiolabeled tubulin were measured with a laser densitometer.

Immunoprecipitation and Western blot analysis of tubulin phosphorylation. After the treatments, confluent cell monolayers were lysed by incubation for 20 min in 500 µl of cold lysis buffer (20 mM Tris · HCl, pH 7.4, 150 mM NaCl, 10 µg/ml of the anti-protease cocktail, 10% glycerol, 1 mM sodium orthovanadate, 5 mM NaF, and 1% Triton X-100). The lysates were clarified by centrifugation at 14,000 g for 10 min at 4°C. For immunoprecipitation, the lysates were incubated for 4 h at 4°C with monoclonal anti--tubulin (1:50 dilution, in excess). The extracts were then incubated with protein G plus Sepharose 4B (Zymed, South San Francisco, CA) for 2 h at 4°C. The immunocomplexes were collected by centrifugation (2,500 g for 5 min) in a microfuge tube and were washed three times with immunoprecipitation buffer containing 5 mM Tris · HCl, pH 7.4, and 0.2% Triton X-100. The resultant pellets were resuspended in a standard SDS sample buffer and boiled at 95°C for 5 min before separation by PAGE as previously described (5-7, 10). Gels were transferred to nitrocellulose membranes, blocked with 1% BSA and 0.01% Tween 20 in PBS for blotting by polyclonal anti-phosphoserine (1:3,000 dilution; Transduction Labs, Lexington, KY) and for detection of immune complexes by horseradish peroxidase-conjugated secondary antibody, incubated with chemiluminescence reagents, and autoradiographed for analysis by densitometry.

Determination of barrier permeability by fluorometry. Barrier integrity was determined by a widely used and validated technique that measures the apical-to-basolateral paracellular flux of a fluorescent marker, fluorescein sulfonic acid (FSA; 200 µg/ml, 478 Da) as we (5-8) and others (11, 12, 22, 27, 31, 34, 39, 45, 47, 49, 57, 59-61) have described. This fluorescent tracer is a lipophobic moiety and is known to be cell membrane impermeant, moving instead through the paracellular space (11, 19, 22, 27, 31, 34, 39, 42, 45, 57, 59-61). In earlier permeability studies, we found that dextrans of higher molecular mass (up to 70 kDa) yielded similar results. Briefly, fresh phenol-free DMEM (800 µl) was placed into the lower (basolateral) chamber, and phenol-free DMEM (300 µl) containing FSA was placed in the upper (apical) chamber. Aliquots (50 µl) were obtained from the upper and lower chambers at time 0 and at several subsequent time points (e.g., 0, 10, 20, 30, and 40 min) and transferred to clear 96-well plates (clear bottom, Costar, Cambridge, MA). Fluorescent signals from the samples were quantitated with a fluorescence multiplate reader (FL 600, Bio-Tek Instruments). The excitation and emission spectra for FSA were excitation = 485 nm and emission = 530 nm. The paracellular permeability of monolayers was expressed as clearance (Cl), calculated as the apical-to-basolateral flux of the FSA probe divided by the concentration of probe in the apical chamber. Clearance was expressed as nanoliters per hour per square centimeter. To calculate Cl, we used the following formula: Cl (nl · h¹ · cm²) = Fab/([FSA]_a × S), where Fab is the apical-to-basolateral flux of FSA (light units per hour), [FSA]_a is the concentration at baseline (light units per nanoliter), and S is the surface area (0.3 cm², which is the growth surface for monolayers). Because both the basal permeability of the monolayers and the magnitude of the effect of oxidants and protective agents on permeability varied from experiment to experiment, this variability was controlled by including simultaneously run controls (e.g., vehicle/isotonic saline, oxidant, and EGF-pretreated) with each experiment.

Statistical analysis. Data are presented as means ± SE. All experiments were carried out with a sample size of at least 4-6 observations/group. Statistical analysis comparing treatment groups was performed with ANOVA followed by Dunnett's multiple-range test (25). Correlational analyses were done with the Pearson test for parametric analysis or, when applicable, the Spearman test for nonparametric analysis. P values <0.05 were deemed statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Pretreatment of monolayers of naive-type Caco-2 cells with EGF and PKC activators (e.g., OAG) protected the monolayer barrier against H₂O₂-induced injury, a finding that confirmed our previous observations (6, 7). We also confirmed in these naive cells that EGF and PKC activators activated the PKC-1 isoform via its rapid translocation from the cytosol to membrane-bound fractions (data not shown). In the current investigation, with molecular interventions (transfection), the role of PKC-1 was further examined.

Stable overexpression of PKC-1 isoform after transfection of intestinal cells. Caco-2 cells were cotransfected with cDNA encoding both G418 resistance (for selection) and PKC-1. Cell lysates of confluent monolayers were prepared from these transfected cells and then analyzed by Western immunoblotting. Figure 1A shows overexpression of the PKC-1 isozyme in transfected cells (data for 4 µg of DNA plasmid shown). The PKC-1 isolated from transfected cells comigrated with a known standard (~78 kDa) of PKC-1 from rat brain lysates. The immunoblot shown in Fig. 1B demonstrates that total PKC-1 levels were elevated by 3.1-fold compared with those in naive cells.

View larger version (17K): [in this window] [in a new window]   Fig. 1.   A: stable overexpression of protein kinase C (PKC)-1 protein, as assessed by immunoblotting, in Caco-2 cells transfected with a plasmid (4 µg) containing PKC-1 cDNA. Differentiated cell monolayers were lysed, sonicated, and processed for SDS-PAGE with a monoclonal anti-PKC-1 antibody followed by a horseradish peroxidase (HRP)-conjugated secondary antibody. Total overexpressed PKC-1 protein is shown in lane a. Commercially obtained positive PKC-1 control (lane+) comigrated as a 78-kDa band. Absence of primary antibody (lane b) resulted in the disappearance of the corresponding PKC-1 band. Similarly, preincubation with the anti-peptide (lane c) to the primary antibody before incubation with monoclonal anti-PKC-1 antibody caused the disappearance of the PKC-1 band. Prestained molecular weights (Mr) 67,000 and 93,000 were also run in adjacent lanes. A representative blot is shown; n = 6 blots/group. B: comparison of total levels of PKC-1 protein expression in transfected Caco-2 cells vs. naive cells. Samples (75 µg protein/lane) were processed for Western immunoblotting with monoclonal anti-PKC-1 antibody. Quantitative analysis by densitometry showed a 3.1-fold elevation of PKC-1 protein levels in transfected cells.

1

2

1

2

1

2

1

2

1

2

1

2

1

2

View larger version (30K): [in this window] [in a new window]   Fig. 2.   Protective effect of overexpression of PKC-1 on Caco-2 cell monolayer barrier integrity as assessed by fluorescein sulfonic acid (FSA) paracellular clearance. Monolayers stably overexpressing PKC-1 were exposed to low doses of epidermal growth factor (EGF; 1 ng/ml; A) or the PKC activator 1-oleoyl-2-acetyl-sn-glycerol (OAG; 0.01 µM; B) for 10 min before exposure to oxidant (0.5 mM H2O2) for 30 min. These low doses did not protect naive (N) cells against oxidant-induced injury, but they did protect transfected (T) cells overexpressing PKC-1. Naive monolayers were protected only by high doses of EGF (10 ng/ml) or OAG (50 µM). C: time course of changes in FSA clearance. D: representative monolayer (n = 6) viewed with ultra high-resolution laser scanning confocal microscopy (LSCM) shows a complete absence of FSA probe in cytosol of a confluent monolayer of Caco-2 cells exposed to H2O2 (shown by green fluorescence in spaces between adjacent cells). Bar, 10 µm. Barrier integrity/paracellular permeability was calculated as apical-to-basolateral flux of FSA divided by the concentration of probe in the apical chamber expressed as a clearance. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. low doses of EGF or OAG before H2O2 in naive cells.

View this table: [in this window] [in a new window]   Table 1.   Effects of transfection of PKC-1 sense or antisense DNA on Caco-2 monolayer barrier integrity

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Protective effects of PKC-1 overexpression on percentage of Caco-2 cells displaying a normal microtubule cytoskeleton in the presence of a low concentration of EGF or the PKC activator OAG. Conditions and treatments were as described in Fig. 2. Cell monolayers were processed for immunofluorescent staining of microtubules by cellular fixation, incubated with a primary monoclonal anti--tubulin antibody, and subsequently incubated with a secondary FITC-conjugated antibody. Note synergy-induced protection of microtubules in PKC-1-overexpressing cells exposed to low doses of EGF or OAG. Also note that in the absence of EGF or OAG, both PKC-overexpressing cells and naive cells responded in the same manner to oxidant injury. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. low doses of EGF or OAG + H2O2 in naive cells.

View larger version (100K): [in this window] [in a new window]   Fig. 4.   Immunofluorescent staining of the microtubule network revealing intracellular distribution in intestinal cells from monolayers, as captured by ultra high-resolution LSCM. Monolayers of naive cells were incubated with vehicle (isotonic saline; a), 0.5 mM H2O2 (b), 1 ng/ml of EGF (c) and then 0.5 mM H2O2 or 0.01 µM OAG + 0.5 mM H2O2 (d). PKC-1- overexpressing cell monolayers were also exposed to low doses of 1 ng/ml EGF (e) or 0.01 µM OAG (f) and subsequently incubated with H2O2. Normal cells (a) exhibit an intact filamentous architecture of microtubules that courses in a radial and stellate fashion throughout the cytosol. Cells incubated in H2O2 (b) show a fragmented, disrupted, and collapsed organization of the microtubules. In transfected cells overexpressing PKC-1 that were exposed to low doses of EGF (e) or OAG (f) before oxidant exposure, the normal appearance of the microtubules is preserved. This protection did not occur in naive cells (those not overexpressing PKC-1) that were preexposed to the same low doses of EGF (c) or OAG (d), as shown by the disrupted organization of the microtubules. Bar, 25 µm.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   A: quantitative immunoblotting analysis of the stable polymerized tubulin fraction (S2, index of microtubule stability) and the monomeric tubulin fraction (S1, index of microtubule disassembly) in Caco-2 monolayers. Conditions were the same as in Figs. 2 and 3. %Polymerization of tubulin = [(S2)/(S2 + S1)], where S2 + S1 is the total cellular tubulin pool. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. low doses of EGF or OAG + H2O2 in naive cells. B: representative Western immunoblot photomicrograph of the S2 fractions extracted from Caco-2 cell monolayers. Conditions were similar to those in A. Tubulin fractions were processed for SDS-PAGE and Western blotting with monoclonal anti--tubulin antibody followed by HRP-conjugated secondary antibody. Lane a, vehicle; lane b, 0.5 mM H2O2 challenge in naive cells; lane c, 0.5 mM H2O2 challenge in PKC-1-overexpressing cells; lane d: 1 ng/ml EGF + 0.5 mM H2O2 in PKC-1-overexpressing cells; lane e, 1 ng/ml EGF + 0.5 mM H2O2 in naive cells; lane f: 0.01 µM OAG + 0.5 mM H2O2 in PKC-1-overexpressing cells; lane g: 0.01 µM OAG + 0.5 mM H2O2 in naive cells; lane h: 10 ng/ml EGF + 0.5 mM H2O2 in naive cells; lane i, 50 µM OAG + 0.5 mM H2O2 in naive cells; lane j, tubulin standard (50 kDa). In stably transfected cells, overexpressed PKC-1 in the presence of a low dose of EGF (1 ng/ml) or OAG (0.01 µM) enhanced the polymerized tubulin band density to control levels. In contrast, in naive cells, pretreatment with the same low dose of EGF or OAG did not increase tubulin assembly, which was comparable to that in oxidant groups. High doses of EGF (10 ng/ml) or OAG (50 µM), which increase tubulin polymerization in naive cells, are shown as additional controls (lanes h and i).

Effects of EGF and PKC activator (OAG) on the intracellular translocation and activation of overexpressed PKC-1 in transfected Caco-2 monolayers. Cytosol-, membrane- and cytoskeleton-associated fractions containing PKC-1 from transfected Caco-2 cells were isolated and assessed by Western immunoblotting. After preexposure to low doses of EGF or OAG (Fig. 6, A-E), there was a rapid redistribution of PKC-1 isoform from a mostly cytosolic distribution into both the membrane and cytoskeletal fractions. Translocation of the 78-kDa overexpressed PKC-1 protein to these particulate fractions (Fig. 6B) was readily observable as early as 2.5 min after low doses of EGF and became undetectable in the cytosolic fraction (Fig. 6A) by 10 min, indicating nearly total activation of PKC-1. OAG caused identical effects (particulate fraction, Fig. 6D; cytosolic fraction, Fig. 6C). We found rapid translocation of native PKC-1 to particulate fractions of naive cells only after higher doses of EGF or OAG (data not shown), which confirmed our previous findings (7).

View larger version (44K): [in this window] [in a new window]   Fig. 6.   Effect of EGF or OAG on distribution of overexpressed PKC-1 in both the cytosolic and particulate (membrane + cytoskeletal) fractions of Caco-2 cells. Transfected cells stably overexpressing PKC-1 were preincubated with a low dose of EGF (1 ng/ml; A and B) or OAG (0.01 µM; C and D). Cell monolayers were processed for fractionation and immunoblot detection of PKC-1. Note the gradual translocation of PKC-1 over time (0-10 min) from the cytosolic into the particulate fraction by either EGF or OAG, indicating activation of this isoform. E: intracellular distribution of overexpressed PKC-1 in cytosol-, membrane-, and cytoskeleton (Triton-X-100 insoluble)-associated fractions in stably transfected Caco-2 cells. Cells were pretreated with EGF (1 ng/ml) or OAG (0.01 µM) with or without subsequent incubation with H2O2 (0.5 mM). Note the shift in distribution of PKC-1 isoform from a mostly cytosolic pool under normal conditions into both the membrane- and cytoskeleton-associated monolayer fractions (particulate fraction) after exposure to EGF or OAG. Relative levels of PKC-1 overexpression in these fractions were quantified by measuring the optical density (OD) of the bands corresponding to anti-PKC-1 immunoreactivity with a laser densitometer. OD for the cytosolic pool in vehicle-treated cells was assigned an arbitrary value of 100, and all other data were normalized to that value. *P < 0.05 vs. corresponding fraction treated with vehicle. +P < 0.05 vs. corresponding fraction treated with H2O2. &P < 0.05 vs. corresponding fraction treated with EGF (or OAG) + H2O2. F: FSA clearance and PKC-1 particulate-associated fraction vs. time. Transfected cells were pretreated with EGF (1 ng/ml) before incubation with H2O2 (0.5 mM). Variables plotted are optical density of particulate-associated PKC-1 band (from Fig. 6B) and relative FSA clearance (expressed as %time 0 control).

View larger version (15K): [in this window] [in a new window]   Fig. 7.   Correlation of FSA clearance (marker for permeability; A), %normal microtubules (marker for status of cytoskeletal integrity; B), or S2 tubulin assembly (index of tubulin polymerization; C) vs. optical density of particulate-associated PKC-1 band (i.e., 1 activation). Data were pooled across different treatment groups for transfected cells (those overexpressing PKC-1).

Stable antisense inhibition of PKC-1 and its inhibitory effects on EGF-mediated protection. To demonstrate a key role for PKC-1 in EGF-induced protection by an independent method, naive Caco-2 cells were transfected with PKC-1 antisense plasmid (4 µg shown in Fig. 8) and cDNA encoding G418 resistance. Figure 8A, which is an immunoblot of cell lysates, shows that this manipulation substantially and stably diminished the steady-state levels of PKC-1 protein by ~90%.

View larger version (15K): [in this window] [in a new window]   Fig. 8.   A: stable antisense inhibition of PKC-1 protein expression, as assessed by immunoblotting, in Caco-2 cells transfected with antisense cDNA (4 µg) to PKC-1. Whole cell lysates of differentiated monolayers were processed for immunoblotting with monoclonal anti-PKC-1 antibody and HRP-conjugated-secondary antibody. B: stable antisense inhibition of PKC-1 protein expression prevents the protective effects of high doses of EGF or the PKC activator OAG on intestinal cell monolayer barrier function. Differentiated cells almost completely lacking PKC-1 protein (90% reduction, see A) were treated for 10 min with a high dose of EGF (10 ng/ml) or OAG (50 µM) before exposure to H2O2. FSA clearance was assessed as described in Fig. 2. A, antisense inhibition of PKC-1 protein. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. high doses of EGF or OAG + H2O2 in naive cells.

View larger version (23K): [in this window] [in a new window]   Fig. 9.   Stable antisense inhibition of PKC-1 protein prevents the protective effects of EGF (10 ng/ml) or OAG (50 µM) on the microtubule cytoskeleton as assessed by the percentage of Caco-2 cells displaying a normal microtubule cytoskeleton. Conditions were as described in Fig. 8. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. high dose of EGF (10 ng/ml) + H2O2 or high dose of OAG (50 µM) + H2O2.

View larger version (24K): [in this window] [in a new window]   Fig. 10.   Antisense inhibition of PKC-1 protein prevents the protective effects of EGF or OAG on the enhancement of tubulin assembly as determined by quantitative immunoblotting analysis. Quantitative analysis of the polymerized tubulin (S2, index of microtubule assembly) and monomeric tubulin (S1, index of microtubule disruption) were performed under conditions similar to those shown in Fig. 9. %Polymerization of tubulin = [(S2)/(S2 + S1)]. *P < 0.05 vs. vehicle. +P < 0.05 vs. H2O2. &P < 0.05 vs. high doses of EGF or OAG + H2O2.

View larger version (11K): [in this window] [in a new window]   Fig. 11.   Effects of EGF and OAG on the phosphorylation of tubulin and its prevention by antisense inhibition of PKC-1 in Caco-2 cells. Conditions were as described in Fig. 8B. The relative levels of tubulin phosphorylation in Caco-2 cell extracts were quantified by measuring the OD of the bands corresponding to anti-phosphoserine immunoreactivity for immunoprecipitated tubulin with a laser densitometer. OD for the tubulin phosphoserine levels was normalized to the corresponding vehicle/control (7% for naive cells; 7.9% for antisense-transfected cells). Lane a, vehicle in naive cells (7%); lane b, vehicle in PKC-1 antisense-transfected cells (7.9%); lane c, 0.5 mM H2O2 challenge in naive cells (5%); lane d, 0.5 mM H2O2 challenge in antisense-transfected cells (4%); lane e, EGF (10 ng/ml) alone in naive cells (89%); lane f, EGF (10 ng/ml) + 0.5 mM H2O2 in naive cells (92%); lane g, EGF (10 ng/ml) + 0.5 mM H2O2 in PKC-1 antisense-transfected cells (10%); lane h, OAG (50 µM) alone in naive cells (94%); lane i, OAG (50 µM) + 0.5 mM H2O2 in naive cells (87%); lane j, OAG (50 µM) + 0.5 mM H2O2 in antisense-transfected cells (11%).

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

With the use of monolayers of intestinal epithelial cells as a model of GI barrier integrity, we conclude that the 1 isoform of PKC is critical to EGF-induced protection against oxidant-induced damage to the microtubule cytoskeleton and to barrier integrity. This suggests that the PKC-1 isoform is a key intracellular regulator of epithelial barrier integrity. Several independent lines of evidence support these conclusions.

Transfected cells that overexpress PKC-1 are severalfold more sensitive to protection by EGF. They are equally more sensitive to the PKC activator OAG. These observations are consistent with our previous findings in naive cells (5-7) that showed that increases in overall PKC activity mediate EGF-induced protection. PKC-1 overexpression synergized with the addition of EGF or OAG to enhance the stability of polymerized tubulin, reduce the unstable monomeric tubulin, maintain a significantly higher percentage of Caco-2 cells with intact microtubules, and stabilize monolayer barrier integrity at almost normal levels. This increased sensitivity appears to require not only overexpression, which by itself is not protective, but also activation/translocation of PKC-1. As in naive cells, protection required activation through the translocation of PKC-1 from the cytosolic to the particulate fractions. The increased sensitivity did not appear to be due to a decrease in the ability of oxidants to cause damage because transfected cells (in the absence of EGF or OAG) were damaged to the same extent as naive cells; they showed similar loss of barrier integrity, as measured by FSA clearance, and similar degrees of damage and disruption of the microtubule cytoskeleton, as assessed by examination of microtubule cytoarchitecture and tubulin polymerization.

Robust and statistically significant correlations between several measures of outcome further supported our conclusion. These included associations between increasing PKC-1 translocation/activation on the one hand and, on the other hand, increasing protection against barrier disruption, tubulin disassembly, and microtubule disruption. These findings are consistent with our model showing that increased translocation/activation of PKC-1 leads to increased tubulin assembly, which leads to increased microtubule stability, which, in turn, leads to increased monolayer integrity. For example, this conclusion is supported by a correlation (r = 0.91; P < 0.05) that shows that increases in PKC-1 translocation decrease FSA clearance (Fig. 6F). The same conclusion is reached when either tubulin assembly (r = 0.92; P < 0.05) or the percentage of normal microtubules (r = 0.95; P < 0.05) is used as a marker of integrity. The high strength of these correlations, which explains 80-95% of the variance, suggests that PKC-1 activation is critical to the protective effects of EGF (and OAG) on microtubule and barrier function. These results also confirmed our earlier reports (5-7) in naive cells in which we found a significant correlation between the integrity of barrier permeability and microtubule stability (r = 0.98; P < 0.05) and between the integrity of barrier permeability and total PKC activation (r = 0.94; P < 0.05). This mechanism is also consistent with previous studies that have documented that PKC translocation/activation is necessary for the observed effects of specific PKC isotypes (13, 14, 23, 42, 58, 64).

Another major result that corroborates our conclusion is our finding that cells that are transfected with antisense to PKC-1 and that underexpress PKC-1 (at 10% of normal levels) were rendered severalfold less sensitive to the protective effects of EGF and OAG. In these cells, EGF and OAG were unable to enhance tubulin assembly, stabilize the microtubule cytoskeleton, or maintain monolayer barrier integrity. The reduced sensitivity did not appear to be a result of a loss of the ability of oxidants to cause damage because transfected cells (without EGF or OAG) were damaged to the same extent as naive cells.

Other quantitative considerations further support our conclusion that activation of the 1 isoform of PKC can explain, at least in large part, EGF-induced protection. First, the maximum protection afforded by EGF in PKC-1-overexpressing cells (90%) is essentially the same as the maximum protection afforded by EGF in naive cells (94%). Second, as we found in a previous study (7), OAG was slightly less effective as a protective agent (84%). Third, pretreatment with a different PKC activator [30 nM 12-O-tetradecanoylphorbol 13-acetate (TPA), a phorbol ester] elicited approximately the same level of protection (85%) (7). Fourth, protection of a similar magnitude was observed when protection of the microtubule cytoskeleton was the measure of outcome. When microtubule integrity was measured as the percentage of cells with normal microtubules, protection by EGF and OAG in PKC-1-overexpressing cells was, respectively, 88 and 83%. This is comparable to protection of microtubules by high doses of EGF (10 ng/ml) and OAG (50 µM) in naive cells, which was 89 and 82%, respectively. A similar parallelism was found when tubulin assembly [(S2)/(S2 + S1)] was the measure of outcome. These data indicate that a significantly large portion of protection against oxidant insult is mediated through PKC-1. It should be noted, however, that other PKC isoforms may also contribute to the protective effects of EGF, a question that merits further study.

Our previous studies showed that protection against damage and disruption to the cytoskeleton protects barrier integrity. However, the mechanism through which PKC-1 protects the cytoskeleton is not known. Our previous and present studies suggest three possibilities: 1) decreasing oxidative stress, 2) normalizing cytosolic calcium, and 3) phosphorylation of tubulin. Activation of the 1 isoform of PKC may trigger one or more of these mechanisms. Regarding the first mechanism, we have reported that EGF/PKC prevents oxidant-induced upregulation of an inducible nitric oxide synthase-driven pathway (5-8). Pretreatments such as EGF and OAG that increase overall PKC activity are associated with 1) inhibition of the ability of oxidants to upregulate this inducible nitric oxide synthase pathway and its reaction products, nitric oxide and peroxynitrite; and 2) prevention of nitration, carbonylation, and disassembly of tubulin, three oxidative mechanisms that are required for oxidant-induced disruption of microtubules and monolayer barrier integrity (5-8). Regarding the second mechanism, we have shown that activation of PKC by known PKC activators (e.g., OAG and TPA) attenuates oxidant-induced increases in cytosolic calcium through stimulation of calcium efflux. This, in turn, leads to the normalization of intracellular calcium and prevention of oxidant-induced loss of cytoskeletal and barrier integrity (7). These effects on cell calcium trafficking are likely to be important because we (7) and others (2, 40) have shown that the cytoskeleton is exquisitely sensitive to alterations in intracellular calcium homeostasis and that it can be extensively disrupted by oxidant-induced increases in intracellular calcium.

Our current data suggest a possible third mechanism: protein phosphorylation of tubulin. These data show that EGF and the PKC activator OAG cause an increase in serine phosphorylation of the tubulin subunits of the microtubules. This increase was prevented by PKC-1 antisense DNA, suggesting that PKC may be acting directly or indirectly on these cytoskeletal protein subunits. This phosphorylation mechanism is consistent with previous studies. For example, PKC has been implicated in rearrangement of the cytoskeleton (2, 17, 23, 26), although it is not clearly known which PKC isoforms are key in this process. Recent reports have proposed that PKC is capable of phosphorylating the cytoskeletal proteins talin and vinculin (23). Furthermore, a major specific substrate for PKC, myristoylated alanine-rich PKC substrate (MARCKS), has been proposed as an actin cytoskeletal remodeler (26). Specifically, the actin cytoskeletal organizing activity of MARCKS is inhibited by PKC-mediated phosphorylation. Alternatively, PKC-1 may phosphorylate one of the tubulin-associated capping proteins (e.g., microtubule-associated proteins). Further studies will be needed to explore the nature of the interactions between PKC-1 and the cytoskeleton in GI epithelial cells.

Our findings on PKC-1 are consistent with previous reports by other investigators. EGF activates constitutive PKC in many naive nonepithelial and epithelial cell types (4, 13, 56, 64, 65), including canine and human gastric cells (64, 65) and human colonic epithelial cells (4, 7). Additionally, the translocation of PKC isoforms from the cytosolic to the particulate fraction of the cell is a key step in their activation (13, 14, 23, 42, 58, 64). Furthermore, OAG induces activation of constitutively expressed PKC-1 in non-GI cellular models such as fibroblasts (23).

It should be noted, however, that the effects of PKC activation in cellular models can sometimes be complex and may vary with different experimental settings and cell types (3, 20, 38, 48). For instance, a previous study reported that overexpression of PKC- caused disruption of pig kidney epithelial (LLC-PK1) monolayers (48). To the best of our knowledge, our current findings are the first to report that a specific PKC isoform plays an essential role in protection of GI cells.

Although our studies were designed to investigate possible beneficial effects of EGF, PKC, and PKC-1 activation on protection of the GI tract and our findings are consistent with many other published studies, there do exist reports that PKC may have other effects that are not beneficial. These include the suspected role of PKC and tumor promoters (e.g., phorbol esters) in carcinogenesis (20, 33, 48, 63) as well as in barrier hyperpermeability (48). The existence of a wide array of effects of PKC, some noxious, is one reason that we focused our studies on specific PKC isoforms that mediate EGF protection because it is possible, as our data suggest, that activating or mimicking just one or a few PKC isoforms will have a much higher beneficial/therapeutic index than activating and/or mimicking total PKC activity.

It should be noted that in the current investigation we chose to use well-established and widely used fluorescent probes such as FSA for assessing intestinal monolayer permeability for several reasons. Fluorescently labeled dextrans (DF) and sulfonic acid (FSA) compounds (or radiolabeled agents) offer appropriate choices for epithelial flux assays. First, these probes are convenient because they come in a range of sizes, are nontoxic to cells, are membrane impermeable, and are relatively inexpensive (5-8, 11, 22, 27, 31, 34, 39, 45, 47, 57, 59-62). Second, both in vivo and in vitro studies have revealed that these probes move through the paracellular space (47, 57, 59, 60, 62). Third, the probes are nontoxic lipophobic moieties and are considered to be cell membrane impermeable so that their permeation of cell monolayers must be via the paracellular route. Indeed, our current data (Fig. 2D) support the aforementioned points. Accordingly, these probes can and have been used as tracers to examine many aspects of endothelial and epithelial permeability (5-8, 11, 22, 27, 31, 34, 39, 45, 47, 57, 59-62). Indeed, it is not surprising that the FSA and DF probes have been widely used by numerous groups studying GI inflammation (5-8, 18, 24, 33, 51, 62). For example, these probes were used to assess barrier integrity against lipopolysaccharide-induced mucosal inflammation (62). Similarly, in other studies, these fluorescent probes were used to assess Caco-2 monolayer hyperpermeability (5-8, 22, 31, 34, 45, 59-61).

We recognize that Caco-2 cells are a transformed cell line and that tumor cells may respond differently to PKC-1 than do nontransformed cells, including enterocytes, in native tissue. Nonetheless, our findings now provide a rationale for conducting studies that test whether these same protective mechanisms occur in vivo in animal models and humans. They also provide a rationale for considering a strategy in which agents that target PKC-1 (e.g., PKC-1 mimetics or targeted gene therapy such as the delivery of sense vector for the PKC-1 isoform) might lead to development of novel and improved therapies for GI disorders related to free radical damage, such as IBD. These approaches might maintain epithelial integrity under conditions of oxidative stress and prevent the episodic attacks that recur under proinflammatory oxidative conditions in ulcerative colitis (28, 29, 36, 37, 43, 53). Further studies will be needed to assess the importance of PKC-1 in GI tissues as well as in other cell models.

In summary, our data provide strong support for the idea that PKC-1 can influence the dynamics of microtubule cytoskeleton assembly and intestinal barrier integrity and that it is key for the occurrence of growth factor protection under in vitro conditions in Caco-2 monolayers.