In the present study, we assessed the effects of chemical inhibitors shown to be selective for protein kinase C (PKC) isoforms on lung barrier function both in vitro and in vivo. Rottlerin, a purported inhibitor of PKCδ, but not other chemical inhibitors, dose dependently promoted barrier dysfunction in lung endothelial cells in vitro. This barrier dysfunction correlated with structural changes in focal adhesions and stress fibers, which were consistent with functional changes in cell stiffness. To determine whether the effects noted in vitro correlated with changes in intact lungs, we tested the effects of rottlerin in the formation of pulmonary edema in rats using both ex vivo and in vivo models. Isolated, perfused lungs demonstrated a significant increase in filtration coefficients on exposure to rottlerin, compared with vehicle-treated lungs, an effect that correlated with increased extravasation of Evan's blue dye (EBD)-conjugated albumin. Additionally, compared with vehicle, the ratio of the wet lung weights to dry lung weights was significantly greater on exposure of animals to rottlerin; rottlerin also produced a dose-dependent increase in EBD extravasation into the lungs. These effects on lung edema occurred without any increase in right ventricular pressures. Microscopic assessment of edema in the ex vivo lungs demonstrated perivascular cuffing, with no evidence of septal capillary leak, in rottlerin-exposed lungs. Taken together, rottlerin increases barrier dysfunction in pulmonary endothelial cell monolayers and causes pulmonary edema in rats; results suggestive of an important role for PKCδ in maintaining lung endothelial barrier function.
- protein kinase C
protein kinase c (PKC) C has long been thought to be involved in the regulation of endothelial monolayer permeability. Activation of PKC, using phorbol esters, promoted endothelial monolayer permeability in vitro (27, 33, 49) and lung edema in vivo (21), whereas inhibition of PKC blunts endothelial barrier dysfunction induced by various agonists, including bradykinin (5), platelet-activating factor (7), thrombin (5, 27), vascular endothelial growth factor (50), hydrogen peroxide (22), glucose (17), tumor necrosis factor-α (TNF-α) (11), and neutrophils (39). As such, PKC-mediated microvascular barrier dysfunction has been implicated in the pathogenesis of diabetes (6, 20, 51), atherosclerosis (30), preeclampsia (13), and acute lung injury (37).
PKC is a family of serine/threonine kinases, consisting of 10 isoforms classified according to cofactor requirements (reviewed in Ref. 32). The conventional class of PKCs (cPKCs) encompasses the PKCα, βI, βII, and γ isoforms. The cPKCs require calcium, diacylglycerol, and phosphatidylserine or phorbol esters for activation. The novel PKCs (nPKCs) and atypical PKCs (aPKCs) are calcium independent. The nPKCs, PKCδ, ε, η, and θ, are activated by diacylglycerol and phosphatidylserine or phorbol ester. Finally, the aPKCs, PKCι/λ and ζ, are not activated by phorbol esters, but they require phosphatidylserine or phosphoinositides for activation. Activation of the PKC isoforms involves autophosphorylation, cofactor binding, and enzyme translocation (32). Each PKC isoform is thought to have its own discrete activators, cofactors, and substrates; however, few have been identified.
Studies have shown that infusion of PKC-activating agents, such as phorbol ester or diacylglycerol, increases the lung wet-to-dry weight ratio in guinea pig lungs ex vivo (21). Additionally, infusion of general chemical inhibitors of PKC blocked the effects of H2O2 (22)- or pertussis toxin (42)- induced pulmonary edema in lungs ex vivo or in vivo, respectively. However, in other modes of injury, such as ischemia-reperfusion, inhibitor studies demonstrated PKC not to be necessary for promoting lung edema (23). Work in isolated, perfused coronary venules using chemical inhibitors suggested a role for cPKCs in glucose-induced hyperpermeability (51).
Contrary to results found with nonselective PKC stimulation, our laboratory has previously shown that overexpression of PKCδ enhanced barrier function in epididymal microvascular endothelial cells (14), whereas chemical, with rottlerin, or molecular inhibition, with dominant negative constructs, of PKCδ caused barrier dysfunction in endothelial cells in vitro (16). These effects were selective for PKCδ and not for the cPKCs. We hypothesized that PKCδ plays a facilitory role in endothelial barrier function. In the present study, we tested this hypothesis by investigating the role of chemical inhibitors reported to be selective for PKC isoforms in regulating pulmonary endothelial monolayer permeability in vitro and in regulating pulmonary edema formation. Monolayer permeability was significantly increased in unstimulated lung microvascular and macrovascular endothelial cells in vitro on exposure to rottlerin in a dose-dependent manner. These effects were not noted in monolayers exposed to Ro-31-7549 or Gö-6976. These changes in monolayer permeability correlated with disruption of stress fibers and focal adhesions, as well as attenuation of cytoskeletal (CSK) stiffness in lung endothelial cells. We also demonstrated that rottlerin dose dependently promoted lung edema in Sprague-Dawley rats in vivo and in isolated, perfused lungs ex vivo within 1 h of exposure to the inhibitor, without effecting pulmonary hemodynamics. These effects were not seen with other PKC inhibitors. Furthermore, microscopy analyses identified perivascular cuffing in the extra-alveolar vessels in rottlerin exposed lungs. These data suggest an important function for PKCδ for maintaining lung vascular endothelial barrier function in vivo.
MATERIALS AND METHODS
Cell lines and reagents.
Rat lung microvascular endothelial cells (LMVEC), purchased from Vec Technologies (Rensselaer, NY), were obtained from the periphery of the lung, consisting of pulmonary capillary endothelial cells. Rat primary pulmonary artery endothelial cells (PAEC) were a gift from Dr. T. Stevens (Center for Lung Biology, University of South Alabama College of Medicine, Mobile, AL). The LMVEC and PAEC were propagated in MCDB-131 and DMEM media, respectively, and used between passages 6 and 13.
The PKC inhibitors myelin basic protein (MBP), rottlerin, Gö-6976, Ro-31-7549, and phosphatase inhibitor cocktail set I were purchased from Sigma Chemical (St. Louis, MO), Biomol Research Laboratories (Plymouth Meeting, PA), and Calbiochem (La Jolla, CA), respectively. PKC isoform-specific antibodies were obtained from Santa Cruz Technologies (Santa Cruz, CA) and BD Biosciences (San Jose, CA). Focal adhesion kinase (FAK)-, vinculin-, and tensin-specific antibodies were purchased from Santa Cruz Biotechnologies, Sigma Chemical, and BD Biosciences (San Jose, CA), respectively. Texas red-conjugated phalloidin was obtained from Molecular Probes (Eugene, OR). Phosphatidylserine and diacylglycerol were obtained from Avanti Polar Lipids (Alabaster, AL).
Endothelial monolayer permeability assay.
Changes in endothelial monolayer permeability were assayed using the electrical cell impedance sensor technique (Applied Biophysics, Troy, NY). Equivalent numbers of endothelial cells were seeded on collagen-coated gold electrode (8W10E) arrays to confluence and permitted to adhere overnight. Experiments were performed, and changes in electrical resistance were measured over time (14, 16, 26).
For immunofluorescence studies, endothelial cells grown on coverslips were treated as described. The cells were fixed with 4% paraformaldehyde, and they were rendered permeable with Triton X-100. The cells were immunofluorescently stained for vinculin or filamentous actin, as previously described (12, 13). Images were viewed and recorded at ×1,000 magnification with a Nikon Eclipse E400 fluorescence microscope interfaced with a SPOT Diagnostics Instruments digital camera.
In the studies whereby focal adhesions were quantitated, we measured regions within immunofluorescently stained endothelial cells where colocalization of FAK and tensin occurred. LMVEC were simultaneously stained with fluorescently labeled antibodies directed against FAK and tensin, and images were acquired over the same area, using the appropriate filter settings. Using the ImageJ “colocalization threshold” plug-in [analysis software developed by W. S. Rasband and offered by the National Institutes of Health (NIH), Bethesda, Maryland, USA, http://rsb.info.nih.gov/ij/; plug-in authors, T. Collins (Wright Cell Imaging Facility, Toronto, Canada), W. S. Rasband, and K. Baler (NIH)] the number of pixels colocalized with both FAK and tensin immunofluorescence staining were calculated and normalized to the number of pixels of FAK immunofluorescence staining in each image. Samples were done in duplicate in each experiment, and multiple (5–6) images were acquired for each sample.
Magnetic twisting cytometry with optical detection.
Stiffness of each individual LMVEC was measured as described previously (4, 10). A ferrimagnetic microbead (4.5 μm in diameter) was coated with a synthetic peptide containing the sequence Arg-Gly-Asp (RDG) and was then allowed to bind to LMVEC. Such RGD-coated bead binds avidly to cell surface integrin receptors (46) and form focal adhesions (29); it becomes well-integrated into the cytoskeletal (CSK) scaffold (10, 46) and displays tight functional coupling to stress-bearing CSK structures and the contractile apparatus (3, 18, 47). Briefly, an RGD-coated ferrimagnetic microbead bound to adherent cell was first magnetized horizontally (parallel to the surface on which cells were plated) with a brief 1,000-G pulse and then twisted in a vertically aligned homogenous magnetic field (20 G) that was varying sinusoidally in time at a frequency of 0.75 Hz. The sinusoidal twisting magnetic field caused both a rotation and a pivoting displacement of the bead: as the bead moves, the cell develops internal stresses that in turn resist bead motions (10). Lateral bead displacements in response to the resulting oscillatory torque were detected optically (in spatial resolution of ∼5 nm), and the ratio of specific torque to bead displacements was computed and expressed as the cell stiffness in units of pascals per nanometer.
PKC activity assays.
Confluent LMVEC were exposed to vehicle (DMSO), or indicated amounts of rottlerin, Ro-31-7549, or Gö-6976 in serum-free medium for 30 min. The cells were immediately washed in ice-cold PBS, scraped into equivalent volumes of PKC homogenization buffer (50 mM Tris·HCl, pH 7.5, 5 mM EDTA, 10 mM EGTA, 0.3% β-mercaptoethanol, 500 μM PMSF, 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1× phosphatase inhibitor cocktail set I), and sonicated, as previously described (15). Equivalent amounts of lysate protein (10 μg) were then assayed for PKC activity by incubating with 10 μg MBP in PKC reaction buffer [20 mM Tris·Cl, pH 7.5, 5 mM MgCl2, 200 μM CaCl2, 400 ng phosphatidylserine, 40 ng diacylglycerol, 5 μM ATP (with 1 μCi γ-32P-labeled ATP, 3,000 Ci/mmol)] for 15 min at 30°C. The reaction was blotted onto P81 phosphocellulose membranes, and membranes were washed three in 5% acetic acid. The amount of radioactivity remaining on the membranes was determined by a scintillation counter. PKC activity is expressed as the picomoles of phosphate transferred to MBP per minute per micrograms of lysate protein.
Parallel experiments examined PKCδ activity in LMVEC exposed to vehicle or PKC inhibitors by scraping in lysis buffer (50 mM HEPES, 150 mM NaCl, 200 μM sodium orthovandate, 10 mM sodium pyrophosphate, 100 mM NaF, 1 mM EDTA, 1.5 mM MgCl2, 10% glycerol, 1% Trition X-100, 1 mM PMSF, and 10 μg aprotinin) and immunoprecipitating equivalent amounts of lysate for PKCδ overnight, as previously described (14, 35). The immunoprecipitates were washed 3× with lysis buffer and 2X with 25 mM Tris·HCl, pH 7.4, 5 mM MgCl2, 0.5 mM EDTA, and 1 mM DTT. The immunoprecipitates were suspended in PKC activity buffer [25 mM Tris·Cl, pH 7.4, 50 mM MgCl2, 5 mM EDTA, 1 mM DTT, 2 μg diacylglycerol, 200 μg phosphatidylserine, 10 μM ATP (with 1 μCi γ-32P-labeled ATP, 3,000 Ci/mmol), and 5 μg MBP] and incubated at for 30 min at 30°C, shaking. The reaction was terminated with Laemli buffer. The reactions were subsequently resolved by SDS-PAGE. In parallel, phosphorylation levels of MBP were determined by autoradiography, and immunoprecipitation of equivalent amounts of PKCδ was confirmed by immunoblot analysis.
For the ex vivo lung edema studies, lungs were isolated from anesthetized (pentobarbital sodium 50 mg/kg ip) adult male Sprague-Dawley rats (300–500 g) and perfused as previously described (1, 2). Briefly, following a tracheotomy, the animals were ventilated at 6 ml/kg with 5% CO2-balance air at 60 breaths/min, and peak airway pressure was recorded at 12 cmH2O with positive end-expiratory pressure at 2 cmH2O. The heart and lungs were exposed via a subdiaphragmatic incision and the pulmonary artery (PA) and left atrium were cannulated and perfused with modified Earle's buffer containing 4% albumin maintained at 37°C, pH 7.4. The pH of the perfusate was monitored before and after each experiment. The heart and lungs were then removed en bloc and suspended on a force transducer (Grass FT03C). Arterial (Pa), venous (Pv), and airway pressures were monitored using a Grass model 78D polygraph (Grass Instrument, Quincy, MA). Perfusate flow was adjusted to 0.04 ml/g body weight. Basal capillary filtration coefficient (Kf) was measured, following lung recruitment to reach zone III conditions and after achieving isogravimentric state. Kf was determine by the rate of weight gain during the final 2 min following an increase in Pv pressure by ∼8 cmH2O (high hydrostatic challenge) for 15 min divided by the change in capillary pressures (Pc) taken by the double-occlusion technique. Kf was normalized to 100 g wet lung mass (WLM), which was empirically derived [WLM = 0.00472 body mass (44)]. After baseline Kf was taken, the lungs were then returned to basal pressures, and either vehicle (DMSO) or rottlerin (50 μM final) was added to the reservoir and allowed to circulate for 45 min, after which a second Kf measurement was taken.
For the in vivo lung edema studies, adult, male Sprague-Dawley rats were anesthetized and the right internal jugular vein was dissected and cannulated. A catheter was advanced into the right ventricle and connected to a Grass 78D polygraph to measure pulmonary intravascular pressure for both pre- and posttreatment. Right ventricle systolic and diastolic pressures were measured and are presented as mean pressures. Assuming a total blood volume in the rats of 25 ml, and no other volume of distribution, the rats were randomly given doses of the PKC inhibitors approximate to the IC50 concentrations for rottlerin, Gö-6976, Ro-31-7549 or corresponding volume of vehicle (65 μl DMSO) in 1 ml of 0.9% saline through the right ventricular catheter.
For measuring wet and dry lung weights, the animals were killed after 50 min and the lungs were removed. Lungs were blotted on gauze to remove excess fluid, and wet weights were recorded immediately. The lungs were then dried for 48 h at 90°C and weights of the dried lungs were recorded. Data are presented as the ratio of wet lung weight relative to dry lung weight.
For the EBD extravasation experiments, 5 min after infusion of the vehicle or PKC inhibitor, 1.25 ml of 0.5% EBD in saline was injected at the same site. The animals were killed after an additional 45 min, the pulmonary vasculature was perfused with 5 ml PBS via the right ventricle, and the lungs were removed and homogenized in 4 ml formamide and incubated at 60°C overnight. The amount of dye in the lungs was then determined spectrophotometrically and extrapolated from a standard curve.
All animal protocols were approved by the Providence Veterans Affairs Medical Center and Brown University Institutional Animal Care and Use Committee and comply with the Health Research Extension Act and the Public Health Service policy.
Microscopic assessment of lung.
For identifying segment-specific leak sites within the lungs, the lungs were perfused with 3% glutaraldehyde in cacodylate buffer at high hydrostatic challenge following the final Kf measurements, until the fixative was recovered from the venous outflow. The lungs were then immersed in the same fixative. Specimens were postfixed in 1% osmium tetroxide, dehydrated with graded alcohol series, and embedded in PolyBed 812 resin. Semithin sections (1 μm) were cut with a glass knife, stained with 1% toluidine blue, and examined by light microscopy. Representative regions of both the extravasular and the septal compartment of the lungs were then selected and processed for transmission electron microscopy (TEM). Thin sections (80 nm) were cut using a diamond knife. The thin sections were stained with uranyl acetate and counterstained with Reynolds's lead citrate. The sections were examined for TEM (Philips CM 100, FEI), and representative images were selected for presentation.
For three or more groups, differences among the means were tested for significance in all experiments, using ANOVA with Fisher's least significance difference test. For two groups, differences among the means were tested for significance using Students’ unpaired t-test or paired t-test. Significance was reached when P < 0.05. All data are presented as means ± SE; n is indicated for each set of data.
Effects of inhibitors of PKC on lung endothelial barrier function.
Utilizing immunoblot analysis, we determined that endothelial cell monolayers isolated from rat lung microvasculature (LMVEC) and macrovasculature (PAEC) express PKC isoforms α, δ, ε, η, and μ (Fig. 1). In addition, the LMVEC also expressed PKC ι/λ, whereas the PAEC did not. To determine whether selective PKC isoforms play a role in maintaining endothelial barrier dysfunction in the cells that regulate pulmonary edema formation, we examined the effect of rottlerin, a chemical inhibitor purported to be selective for PKCδ, in LMVEC and PAEC. We found that rottlerin significantly increased basal endothelial monolayer permeability in both rat LMVEC and PAEC in a dose-dependent manner (Fig. 2, A and B). A significant increase in permeability was seen at concentrations below the reported PKCδ IC50 of 5.3μM (12). Chemical inhibitors of other PKC isoforms (α, β, γ, and ε), with concentrations of Ro-31-7549 and Gö-6976 above reported IC50 values (28, 48), did not significantly alter basal monolayer permeability in rat LMVEC (Fig. 2C) or PAEC (data not shown). These findings corroborate our laboratory's earlier studies in which our group demonstrated increased barrier dysfunction in endothelial cells isolated from the epididymis on either chemical or molecular inhibition of PKCδ (16).
Rottlerin causes actomyosin filament and focal adhesion disruption.
The integrity of both contractile forces (i.e., actomyosin filaments and microtubules) and adhesive forces (i.e., focal adhesions) have been shown to be important in maintaining endothelial barrier function. To determine whether endothelial barrier dysfunction occurred by disrupting filamentous actin and focal adhesions, we examined the effects rottlerin and additional PKC chemical inhibitors on structural changes in stress fibers and focal adhesions. Compared with vehicle-treated LMVEC, cells exposed to rottlerin demonstrated fewer stress fibers and focal adhesions (Fig. 3A), as demonstrated by simultaneous immunofluorescence staining for filamentous actin and vinculin, respectively. These structural changes in LMVEC were consistent with our laboratory's previous data in rat epididymal and pulmonary artery endothelial cells (16). In contrast, neither of the other PKC chemical inhibitors tested, Ro-31-7549 or Gö-6976, significantly altered the stress fiber or focal adhesion structures as identified by immunofluorescence microscopy (data not shown). To quantitate these changes in the CSK structures, we next probed functional changes in CSK stiffness of LMVEC using optical magnetic twisting cytometry. Cell stiffness was measured before treatment (i.e., baseline) and in the same cultures at 30 min following exposure to vehicle, Ro-31-7549, Gö-6976, or rottlerin. Exposure to rottlerin significantly decreased the CSK stiffness in LMVEC, relative to baseline cell stiffness, whereas no significant effects on stiffness were noted in cells exposed to vehicle or the other PKC inhibitors (Fig. 3B).
Focal adhesions and focal complexes are integrin-mediated adhesive structures, each defined as multimeric protein complexes incorporating distinct subsets of proteins (8, 36, 43). Whereas focal adhesions are larger adhesive structures associated with the ends of bundles of actin stress fibers, the focal complexes are described as nascent cell-extracellular matrix adhesions that are small in size and typically found at the edge of migrating cell protrusions, such as lamellipodia and filopodia (8, 36, 43). To further characterize the effect of PKCδ inhibition on the adhesive structures, LMVEC were immunofluorescently stained for a protein found in both focal adhesions and focal complexes, FAK, and for a protein localized in focal adhesions only, tensin (52). We noted colocalization of tensin with FAK in focal adhesions in LMVEC exposed to vehicle (Fig. 4A). On PKCδ inhibition, however, tensin staining remained strong within the nucleus, but it was less noted in focal adhesions. Quantitation of FAK and tensin colocalization relative to total FAK demonstrated fewer focal adhesions (23 ± 5% reduction) in cells treated with rottlerin, compared with vehicle (Fig. 4B). Parallel immunoblot analyses showed no significant effect of rottlerin on the protein levels of tensin or FAK in LMVEC lysates, compared with vehicle-treated endothelial cells (data not shown), suggesting that PKCδ inhibition promotes the disassembly of focal adhesions and possible formation of focal complexes. These results correlated with the data demonstrating increased monolayer permeability.
Rottlerin attenuated PKC activities.
To confirm that the chemical inhibitors did indeed attenuate PKC activity, in vitro PKC activity assays were performed. In vitro assays on total protein lysates of LMVEC demonstrated a significant attenuation of PKC activity on exposure to Ro-31-7549 (250 nM) and rottlerin (10 and 100 μM) (Fig. 5A), relative to vehicle-exposed LMVEC lysates. Additionally, we noted a 15.3 ± 6.7% reduction in PKC activity in Gö-6976 treated lysates using this experimental approach. Furthermore, immunoprecipitation experiments demonstrated diminished PKCδ activity in LMVEC lysates exposed to 1 μM rottlerin (Fig. 5B). Thus the data suggest that rottlerin promoted LMVEC barrier dysfunction through the inhibition of PKCδ.
Rottlerin causes pulmonary edema.
To determine whether the effects of rottlerin on endothelial barrier function in vitro correlated with changes in lung edema, we tested the effects rottlerin on pulmonary edema formation both in isolated, perfused lungs and in intact animals. Rat lungs were isolated and perfused at a constant flow, as previously described (1, 2). Kf values were determined at baseline and on exposure to vehicle or 50μM rottlerin for 45 min. Rottlerin promoted a twofold increase in Kf, compared with baseline Kf, whereas vehicle had no effect on the Kf (Fig. 6).
To determine whether the increased lung permeability was unique to rottlerin, additional experiments were done with inhibitors of other PKC isoforms in vivo. For these experiments, rats were given a bolus of vehicle, 0.125 mmol rottlerin, 6 nmol Ro-31-7549, or 2 nmol Gö-6976 in 1 ml of saline via the right ventricle, which correlated with approximate concentrations of 0, 5 μM rottlerin, 250 nM Ro-31-7549, or 10 nM Gö-6976, respectively (assuming 25-ml volume of distribution per animal). After 5 min, an additional bolus of 0.5% EBD in saline was given via the right ventricle catheter. These studies demonstrated while rats administered rottlerin became edematous, no significant extravasation of EBD was noted in animals administered other PKC chemical inhibitors, Gö-6976 or Ro-31-7549, agents selective against PKCα, β, γ, or ε (Fig. 7A).
Further studies were conducted in vivo where rats were given varying doses of rottlerin, 0 (vehicle), 0.0625, 0.125, or 0.25 μmol (Fig. 7B) in 1 ml of saline via the right ventricle, which correlated with approximate concentrations of 0, 2.5, 5, or 10 μM rottlerin, respectively (assuming 25-ml volume of distribution per animal). We noted a dose response in EBD extravasation in animals treated with varying concentrations of rottlerin, with significant extravasation at 5 and 10 μM rottlerin. In additional experiments, rats were randomly given 0 (vehicle) or 5 μM of rottlerin in the right ventricle. After 50 min, the lungs were harvested and lung wet to dry weights were determined. These studies confirmed increased edema formation in animals given 5 μM, compared with vehicle-treated animals (Fig. 7C).
The increase in pulmonary edema caused by rottlerin did not appear to be associated with increases in hydrostatic pressure. Measurements were made at baseline, before administration of vehicle or rottlerin, and again at 50 min after administration, before the animal was killed. These measurements showed no change in right ventricular pressures (vehicle: 13.2 ± 1.63 mmHg vs. rottlerin: 11.2 ± 1.08 mmHg, n = 5; P = 0.30), compared with baseline (vehicle: 15.3 ± 1.63 mmHg vs. rottlerin: 14.4 ± 1.17 mmHg, n = 5; P = 0.66). Thus rottlerin appears to cause pulmonary edema in anesthesized rats via disruption of endothelial barrier function independent of hydrostatic changes, results suggestive that PKCδ activity is important in maintaining endothelial barrier function in vitro and in vivo.
Lung edema in extra-alveolar vessels on rottlerin exposure.
In Fig. 8, we next analyzed ex vivo prepared lungs exposed to vehicle (left) or rottlerin (right) by light microscopy or TEM. We noted normal parenchymal architecture in vehicle-exposed lungs, where arterial vessels are adjacent to a bronchus with a minimal distance observed between the two structures (Fig. 8A). Normal appearing capillaries and alveolar spaces were also visualized in these lungs. Conversely, we noted an accumulation of fluid in the perivascular space separating the vascular and airway structures in the rottlerin-exposed lungs (Fig. 8B). We further noted a preservation of the architecture in the distal parenchyma in rottlerin-exposed lungs, with no evidence of injury and/or edema accumulation in capillaries or alveolar spaces.
Ultrastructural analyses with TEM of extra-alveolar vessels in vehicle-exposed lungs demonstrated an intact intimal layer with normal apposition of interendothelial junctions (Fig. 8C). A typical smooth muscle and adventitia layers were visualized, with the capillaries subjacent to the vessel wall. The integrity of the endothelial barrier is intact as evidenced by the adjunct interendothelial junctions (Fig. 8C, inset). Rottlerin promoted the disruption of the endothelial barrier, as noted by the formation of gaps at the endothelial cell-cell borders in rottlerin exposed lungs (Fig. 8D, inset), leading to movement of fluid to the vessel wall and the development of perivascular cuffs. Interestingly, analyses of septal capillaries by TEM revealed normal morphology and the absence of injury in the endothelial or the alveolar aspect of the septum in both vehicle-exposed and rottlerin-exposed lungs (Fig. 8, E and F, respectively).
In the present study, we present novel findings demonstrating an important role for a selective PKC isoform, PKCδ, in maintaining endothelial barrier function in the pulmonary microvasculature. Pulmonary microvascular and extra-alveolar endothelial cells control pulmonary edema formation by regulating transudation of fluid and protein across the pulmonary vascular membrane and into the interstitium. Previous studies have suggested a requirement for PKC in agonist-induced lung edema in in vivo and ex vivo models (21, 22, 42); however, the specific PKC isoform important for the maintenance of lung barrier function in vivo is not known. In the present study, we found that rottlerin dose dependently promoted barrier dysfunction in lung microvascular and macrovascular endothelial cells in vitro; effects that were not induced by cPKC inhibitors, Gö-6976 or Ro-31-7549. Rottlerin-induced barrier dysfunction correlated with a decrease in focal adhesions and stress fibers in the pulmonary endothelial cells, as well as a decrease in cell stiffness. We also showed that the PKC inhibitor rottlerin, but not Gö-6976 or Ro-31-7549, dose dependently promoted lung edema in ex vivo and in vivo models, which correlated with interendothelial cell gapping and resultant perivascular cuffs. These effects on lung edema occurred independently of any significant alteration in right ventricular hemodynamics, suggesting that the increased edema formation was not due to an increase in pulmonary hydrostatic pressure. Thus we show that rottlerin promotes barrier dysfunction in endothelial monolayers isolated from the lung microvasculature and macrovasculature (16) and causes pulmonary edema in rats. These findings are consistent with our hypothesis that PKCδ activity is important in maintaining endothelial barrier function in vitro and in vivo.
Previous investigations have yielded data elucidating the role of PKC isoforms in endothelial barrier function; however, these studies have been limited to in vitro endothelial monolayer permeability models. For example, work by Tinsley et al. (41) using antisense oligonucleotides selective for distinct PKC isoforms demonstrated a requirement for PKCδ and μ, but not PKCα, βI, or ε, in phorbol ester-induced rat lung microvascular endothelial monolayer permeability (41). Others have demonstrated increased barrier dysfunction in endothelial cells isolated from human umbilical veins on transient expression of dominant negative PKCζ protein (25). Stable overexpression of PKCβI in human dermal microvascular endothelial cells promoted increased basal and phorbol ester-induced barrier dysfunction (31). Also, using a series of chemical, molecular, and peptide inhibitors, PKCα was shown to be involved in increases in bovine pulmonary microvascular endothelial monolayer permeability by lysophosphatidylcholine and TNF-α (11, 19). Our laboratory has previously shown that PKCα overexpression increased thrombin-induced barrier dysfunction through destabilization of adherens junctions, whereas PKCδ overexpression enhanced basal barrier function in rat epididymal endothelial cells by augmenting focal adhesion formation (14). Our laboratory has also shown that inhibition of PKCδ using either rottlerin, or transient overexpression of dominant negative PKCδ cDNA, caused barrier dysfunction in unstimulated endothelial cells isolated from rat epididymus and proximal pulmonary artery (16). In the present study, we provide new evidence suggesting that basal PKCδ activity is vital for maintenance of endothelial barrier function in LMVEC and PAEC in vitro, and for the first time we extend these observations to both ex vivo and in vivo models measuring lung edema. However, while we demonstrate attenuation of PKCδ activity in endothelial cells by rottlerin, because rottlerin has been shown to inhibit other enzymes (9, 12, 24) and to promote mitochondrial uncoupling in liver cells (38), we cannot exclude the possibility that this chemical inhibitor is also acting through another signaling pathway, independent of or indirect to PKCδ. Although further studies are needed to better decipher the role of the PKC isoforms in lung function in vivo, our study suggests a potential role for PKCδ in modulating pulmonary endothelial barrier function both in vitro and in vivo.
Our finding that rottlerin augmented lung Kf values, increased EBD extravasation, and elevated lung wet weight-to-dry weight ratios in ex vivo and in vivo models of lung edema is the first demonstration suggesting that inhibition of basal PKCδ activity causes pulmonary edema. The rottlerin-induced increases in EBD extravasation were dose dependent and were not simulated by other PKC inhibitors. EBD is a large molecule that binds to albumin. The increase in EBD extravasation from the pulmonary intravascular space demonstrates disruption of the pulmonary endothelial barrier large enough to allow passage of macromolecules in to the alveolar space. We also observed interendothelial cell gap formation and fluid leak from extra-alveolar vessels in rottlerin-exposed lungs, data suggestive of capillary barrier dysfunction. Although we have no direct evidence of septal capillary leak within these lungs, we noted occasional leukocytes and red blood cells within alveoli in several rottlerin-exposed lungs (data not shown). Further work is needed to determine whether PKCδ chemical inhibition promoted lung edema within the capillary bed. Rottlerin has been reported to uncouple mitochondrial respiration/oxidation resulting in the depletion of intracellular ATP levels (38, 40) and to inhibit the activities of other proteins, including p38-regulated kinase and MAPK-activated protein kinase 2 (9). Although we cannot exclude the possibility that rottlerin caused pulmonary edema by mechanisms other than PKCδ inhibition, our findings that rottlerin attenuated LMVEC PKCδ activity in vitro, caused a similar increase in monolayer permeability in lung microvascular and macrovascular (16) endothelial cells in vitro, and disruption in stress fiber and focal adhesion formation (16) strongly implicate a role of PKCδ in maintenance of normal barrier function of the pulmonary vasculature.
Increased pulmonary edema in the rats given rottlerin could not be attributed to increased hydrostatic pressures because the right ventricular pressures were within normal limits throughout the experiment and did not change significantly in response to rottlerin. Although right ventricular pressure is not a direct measurement of Pc, it serves as a useful surrogate to exclude hydrostatic pulmonary edema in this animal model. Of note, during the performance of the isolated lung preparations, the rottlerin had a negligible effect on Pa and Pv. Also, it had no discernible effect on Pc as estimated by the double-occlusion method, thereby supporting the argument that the increased edema formation was not due to increased hydrostatic pressure.
Much of the data, thus far, have suggested that PKC associates with the protein components of the cytoskeleton or adherens junction and that these proteins may serve as substrates for PKC and thus may be important in the modulation of agonist-induced changes in endothelial monolayer permeability. Indeed, we have shown that PKCδ, but not PKCα, enhances barrier function in unstimulated endothelial monolayers by stabilizing actomyosin filament and focal adhesion formation through regulating RhoA GTPase activity, possibly involving upstream modulators of RhoA GTPase (16). We hypothesize that PKCδ may be protective against basal endothelial barrier dysfunction by maintaining a level of active RhoA possibly through modulation of p190RhoGAP. Future studies are needed to investigate the connection between PKCδ and the RhoA signaling pathways in pulmonary edema in vivo.
In summary, we show that rottlerin promoted barrier dysfunction and focal adhesion and stress fiber disruption with concomitant diminution of cytoskeletal stiffness in lung endothelial cells in vitro. In addition, rottlerin, but not Gö-6976 nor Ro-31-7549, induced significant pulmonary edema in unchallenged, normal lungs. We further show that rottlerin promoted gapping of endothelial cells within extra-alveolar vessels and subsequent perivascular cuffing. These effects on lung edema on rottlerin exposure occurred independent of significant alterations in right heart pressures. These findings, together with our previous findings, support the hypothesis that PKCδ activity is important in maintaining endothelial barrier function in the pulmonary circulation and suggest that maintenance of normal PKCδ activity is important in protecting against pulmonary edema formation. We speculate that maintenance of normal PKCδ activity may be important in limiting pulmonary edema formation in acute lung injury.
This material is the result of work supported with resources and the use of facilities at the Providence Veterans Affairs Medical Center and supported with Veterans Affairs Merit Review grants and National Heart, Lung, and Blood Institute Grant HL-67795 (to E. O. Harrington); American Heart Association (AHA) Grant EIG 0240190N (to J. R. Klinger); and AHA Predoctoral Award 0615676T (to A. Owusu-Sarfo).
We thank S. Rounds, M. I. Townsley, M. Jian, and S. Barnes for their assistance, training, and advice with the ex vivo lung edema methodology. We thank H. Duong, E.-B. Kwon, N. Morin, H. Jackson, and C. Murphy for helpful technical assistance. Some of these results were presented at the 2005 American Thoracic Society international meeting and were published in abstract form in Proceedings of the American Thoracic Society 2: A753, 2005.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society