We have previously shown that thrombin induces endothelial cell barrier dysfunction via cytoskeleton activation and contraction and have determined the important role of endothelial cell myosin light chain kinase (MLCK) in this process. In the present study we explored p38 MAP kinase as a potentially important enzyme in thrombin-mediated endothelial cell contractile response and permeability. Thrombin induces significant p38 MAP kinase activation in a time-dependent manner with maximal effect at 30 min, which correlates with increased phosphorylation of actin- and myosin-binding protein, caldesmon. Both SB-203580 and dominant negative p38 adenoviral vector significantly attenuated thrombin-induced declines in transendothelial electrical resistance. Consistent with these data SB-203580 decreased actin stress fiber formation produced by thrombin in endothelium. In addition, dominant negative p38 had no effect on thrombin-induced myosin light chain diphosphorylation. Thrombin-induced total and site-specific caldesmon phosphorylation (Ser789) as well as dissociation of caldesmon-myosin complex were attenuated by SB-203580 pretreatment. These results suggest the involvement of p38 MAP kinase activities and caldesmon phosphorylation in the MLCK-independent regulation of thrombin-induced endothelial cell permeability.
- p38 mitogen-activated protein kinase
- transendothelial electrical resistance
the vascular endothelium works as a semipermeable barrier between the blood and the interstitial space of all organs and participates in the regulation of macromolecule transport and blood cell trafficking through the vessel wall. The integrity of endothelial barrier is maintained by the balance between competing contractile and adhesive cell-cell and cell-extracellular matrix tethering forces. As established, the primary permeability pathway across the vessel wall occurs via a paracellular route and is closely associated with the formation of intercellular gaps between activated endothelial cells (7, 11, 39). Actin cytoskeleton rearrangement and actomyosin interaction are involved in both endothelial cell retraction and gap formation (43, 44). Reorganization of the endothelial actin cytoskeleton, such as cortical actin dissolution and an increased density of the actin stress fibers, leads to cell contraction and alteration in cell shape, providing a structural basis for increase of endothelial cell permeability. In perfused rabbit lungs, selective disruption of actin filaments leads to a significant increase in vascular permeability and marked interstitial edema formation, implicating the direct involvement of actin cytoskeleton in the regulation of permeability in vivo (8).
Multifunctional serine protease, thrombin, generated during activation of the coagulation cascade directly increases vascular permeability in vivo and in vitro (12, 31). Thrombin-induced actin cytoskeleton rearrangement and endothelial cell permeability depend on myosin light chain (MLC) phosphorylation, catalyzed primarily by MLC kinase (MLCK) (43, 44), as well as MLC phosphorylation-independent mechanisms provided by activation of other kinases (2, 3, 38). The latter may involve activation of MAP kinases (like p38 MAP kinase) followed by phosphorylation of targeted actin-binding protein, caldesmon, involved in the regulation of smooth muscle and nonmuscle contraction. Caldesmon-mediated inhibition of actin-activated myosin ATPase activity can be regulated by its phosphorylation/dephosphorylation (37). We have recently demonstrated the important role for p42/p44 MAP kinase activities in thrombin-induced endothelial cell permeability, caldesmon phosphorylation, and actin stress fiber formation (2). It was also reported that caldesmon is a substrate for p38 MAP kinase (21). However, little is known about the involvement of p38 MAP kinase activation in the regulation of endothelial cell barrier function.
In this study we explored the possible role of p38 MAP kinase activities in thrombin-induced endothelial cell barrier dysfunction, actin cytoskeleton rearrangement, and caldesmon and MLC phosphorylation. We report here that p38 MAP kinase regulates thrombin-induced endothelial cell permeability in an MLC phosphorylation-independent fashion possibly via caldesmon phosphorylation and interfering with actin stress fiber formation.
MATERIALS AND METHODS
Bovine thrombin was obtained from Sigma (St. Louis, MO); SB-203580 was purchased from Calbiochem (La Jolla, CA); antibody to MLC was produced in rabbit against purified baculovirus-expressed smooth muscle MLC by Biodesign International (Kennebunk, ME); antibodies to caldesmon and nonmuscle myosin II B were purchased from Sigma; anti-p38 MAP kinase and phosphospecific anti-p38 MAP kinase antibodies were obtained from New England Biolabs (Beverly, MA); diphosphospecific anti-MLC antibodies were raised against phosphorylated Thr18 and Ser19, MLCK sites of phosphorylation as described (32); phosphospecific anticaldesmon antibodies were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).
Bovine pulmonary artery endothelial cells (BPAEC) were obtained frozen at 16 passages from American Type Culture Collection (Manassas, VA; CCL 209) and were utilized at passages 19–24. Endothelial cells were cultured in complete media and maintained at 37°C in a humidified atmosphere of 5% CO2-95% air and grew to contact-inhibited monolayers with the typical cobblestone morphology. Cells from each primary flask were detached with 0.05% trypsin and resuspended in fresh culture medium and passaged to appropriate size flasks or dishes.
Transendothelial electrical resistance.
Endothelial cells were seeded onto evaporated gold microelectrodes and grown to confluence as we have previously described (3). Endothelial cells formed a confluent monolayer, which covered the microelectrodes connected to an electrical cell-substrate impedance system (Applied Biophysics, Troy, NY). Resistance values from each microelectrode (measured in ohms) were normalized as the ratio of measured resistance to baseline resistance and plotted versus time and statistically assessed using “Epool” software, kindly provided by Dr. K. L. Schaphorst (Johns Hopkins University School of Medicine, Baltimore, MD).
Western immunoblotting and immunoprecipitation of caldesmon.
After treatment, endothelial cell monolayers grown in 35-mm dishes were rinsed with ice-cold PBS, lysed with 2× SDS sample buffer, and boiled for 5 min. Extracts were separated on SDS-PAGE, transferred to nitrocellulose (30 V, 18 h), and reacted with antibody of interest. Immunoreactive proteins were visualized with an enhanced chemiluminescent detection system. The relative intensities of the protein bands were quantified by scanning densitometry. The comparisons of two means were performed by Student's t-test. Differences in two groups are considered statistically significant when P < 0.05.
For immunoprecipitation under either denaturing or nondenaturing conditions, confluent endothelial cells (∼106) were rinsed with PBS and then lysed with the addition of either boiling lysis buffer (1% SDS, 1 mM sodium vanadate, and 10 mM Tris·HCl, pH 7.4) or ice-cold immunoprecipitation buffer (1% Triton X-100, 150 mM NaCl, 10 mM Tris, pH 7.4, 1 mM EDTA, 1 mM EGTA, 0.5% Nonidet P-40, and protease inhibitors). A total of 20 μl protein G-agarose (Calbiochem, La Jolla, CA), 400 μl H2O, 400 μl immunoprecipitation buffer, and 100 μl total endothelial cell lysate were combined and incubated for 30 min at 4°C, followed by centrifugation for 5 min. The supernatant fraction was retrieved, and 10–20 μg of monoclonal antibody to caldesmon were added and incubated for 1 h at 4°C. Approximately 20–30 μl of protein G-agarose was added to each tube and incubated for an additional 30 min, followed by centrifugation for 1 min. Pellets were washed three times with immunoprecipitation buffer, resuspended in 2× SDS sample buffer, and boiled for 5 min. Samples were subjected onto SDS-PAGE, transferred to nitrocellulose (30 V, 18 h), and analyzed by Western immunoblotting.
Detection of caldesmon phosphorylation.
Thrombin-induced caldesmon phosphorylation was assessed either in caldesmon immunoprecipitates obtained from 32P-labeled endothelial cells or by using phosphospecific anticaldesmon antibodies raised against peptide sequences surrounding the two major MAP kinase-catalyzed phosphorylation sites: Ser759 [PDGNKS(PO4)PAPKPGC] and Ser789 [CQSVDKVTS(PO4)PTKV] (4), based on the numbering of the human high-molecular-weight caldesmon sequence (22).
Endothelial cell MLC phosphorylation was analyzed by SDS-PAGE followed by Western immunoblotting with diphosphospecific anti-MLC antibodies as we have recently reported (33).
Endothelial cell transfection with adenovirus encoding dominant negative p38α MAP kinase.
The adenoviral construct with cytomegalovirus (CMV) promoter and a cDNA for dominant negative p38α MAP kinase was generated by site-directed mutagenesis (TGY → AGF) as described (20, 41). An adenovirus with CMV promoter but without insertion was used as a control for infection. Cultured endothelial cells at 70–80% confluence were exposed to either recombinant or control adenovirus at 20–30 multiplicities of infection (moi) for 1 h in DMEM (GIBCO-BRL) containing 2% FBS. The virus-containing medium was then replaced with virus-free DMEM with 10% FBS, and endothelial cells were analyzed after 30 h.
Immunofluorescence microscopy studies were performed as we have previously described (21). After treatment, endothelial cells grown on gelatinized coverslips were rinsed with PBS, fixed in 3.7% paraformaldehyde for 10 min, and permeabilized with 0.2% Triton X-100 for 10 min. Cells were then washed with PBS, blocked with PBS-Tween 20 (0.5%) (PBS-T) containing 2% BSA for 30 min, and incubated with primary antibodies. After being washed with PBS-T, cells were incubated with corresponding fluorochrome-conjugated secondary antibodies and 1 U/ml of Texas red-X phalloidin (Molecular Probes, Eugene, OR). Coverslips were mounted on slides with SlowFade mounting medium (Molecular Probes) and analyzed under a Nikon Eclipse TE 300 microscope. Images were captured by Sony Digital Photo camera DKC 5000.
Image analysis of gap formation.
Texas red-stained endothelial cell monolayers stimulated with either thrombin or vehicle were viewed under a microscope using a 20/0.4 objective. The 16-bit images were analyzed with MetaVue 4.6 (Universal Imaging, Downington, PA). Images were differentially segmented between gaps (black) and cells (highest gray value) based on image grayscale levels. The gap formation was expressed as a ratio of the gap area to the area of the whole image. The values were statistically processed with Sigma Plot 7.1 (SPSS Science, Chicago, IL) software.
Thrombin induces p38 MAP kinase activation in BPAEC.
We have recently demonstrated that thrombin produces rapid p42/p44 MAP kinase activation with maximal effect at 5 min and returning to basal level by 30 min (2). In this study, we initially examined whether p38 MAP kinase participates in thrombin-induced endothelial cell activation. Confluent BPAEC were treated for various durations with thrombin (100 nM) in a serum-free medium and explored for p38 MAP kinase activation using phosphospecific antibodies recognizing p38 MAP kinase dually phosphorylated at TGY motif within a regulatory loop. Thrombin induces p38 MAP kinase activation in a time-dependent biphasic manner, with a modest increase at 5 min, followed by maximal effect at 30 min (Fig. 1). Thrombin-induced p38 MAP kinase phosphorylation was not blocked by pretreatment with SB-203580 (10 μM, 30 min), a specific p38α/β MAP kinase inhibitor (Fig. 2). This indicates MKK3/MKK6-dependent p38 MAP kinase activation rather than transforming growth factor-β activated protein kinase 1-binding protein-dependent autophosphorylation of p38 MAP kinase (13).
Effect of p38 MAP kinase inhibition on thrombin-induced endothelial cell barrier dysfunction, MLC phosphorylation, and actin stress fiber formation.
We have recently shown that thrombin-induced endothelial cell barrier dysfunction was attenuated by p42/p44 MAP kinase inhibition by U0126 (10 μM), a specific MEK1/2 inhibitor (2). To determine whether p38 MAP kinase activities are involved in thrombin-induced endothelial cell permeability, we used real-time measurements of electrical resistance across bovine endothelial cell monolayers. Confluent BPAEC grown on gold microelectrodes were pretreated with either vehicle or SB-203580 (10 μM) and, after stabilization, treated with either vehicle or thrombin (100 nM) in a serum-free medium. As shown in Fig. 3A, p38 MAP kinase inhibition by SB-203580 attenuates thrombin-induced declines in electrical resistance, as well as delaying the onset of barrier dysfunction. An observed decrease in electrical resistance caused by SB-203580 alone may be due SB-203580-induced Raf-1/MEK-1/ERK1/2 activation (5, 26), since the latter is proposed to be involved in thrombin-mediated barrier dysfunction in BPAEC (2). To confirm attenuation of thrombin-induced decrease in electrical resistance by SB-203580 pretreatment, BPAEC infected with either adenovirus-based dominant negative p38α MAP kinase or empty vector were prepared for transendothelial electrical resistance measurements after 30 h. Absolute values of electrical resistance between endothelial cells infected with either empty or dominant negative p38 MAP kinase adenoviral constructs did not show any statistical difference. Similarly, p38 MAP kinase inhibition dramatically attenuated thrombin-induced reduction in transendothelial electrical resistance, whereas endothelial cells treated with control vector showed no inhibitory effect (Fig. 3B).
To ectopically express dominant negative p38 MAP kinase, BPAEC were infected with either adenovirus-based dominant negative p38 MAP kinase or empty vector at 20–30 moi as described in materials and methods. Western blot analysis using anti-p38 MAP kinase antibodies demonstrated an increase in ectopic expression of p38 MAP kinase (Fig. 4).
We have previously shown that thrombin-induced endothelial cell barrier dysfunction involves MLC phosphorylation and MLCK-dependent actomyosin interaction (10). To determine the role of p38 MAP kinase activity in this mechanism, we examined the effect of p38 MAP kinase inhibition on the levels of thrombin-indexed MLC diphosphorylation. Confluent BPAEC were infected with either adenovirus-based dominant negative p38α MAP kinase or empty construct as described in materials and methods, challenged with thrombin (100 nM, 30 min) in a serum-free medium, and subjected to SDS-PAGE followed by Western immunoblotting with diphosphospecific (Thr18, Ser19) MLC antibodies. As shown in the Fig. 5, p38 MAP kinase inhibition did not alter the levels of diphosphorylated MLC, suggesting that p38 MAP kinase activity is likely to be involved in the regulation of endothelial cell barrier function in an MLCK-independent fashion. Similar experiments using SB-203580 (10 μM, 30 min) as p38 MAP kinase inhibitor also showed no significant alterations of the levels of diphosphorylated MLC within 30 min of thrombin stimulation (data not shown).
To further elucidate the role of p38 MAP kinase activation in thrombin-induced endothelial cell permeability, we next investigated the effect of p38 MAP kinase inhibition on thrombin-induced actin cytoskeleton rearrangement by immunofluorescence microscopy. Confluent BPAEC were pretreated with either vehicle or SB-203580 (10 μM, 30 min), stimulated with either vehicle or thrombin (100 nM, 30 min) in a serum-free medium, and prepared for immunofluorescence microscopy as described in materials and methods. Inhibition of p38 MAP kinase activation by SB-203580 significantly attenuates thrombin-induced actin stress fiber and paracellular gap formation (Fig. 6), which is consistent with observed decreases in thrombin-induced permeability due to p38 MAP kinase inhibition (Fig. 3). Together, these results provide evidence that thrombin-induced endothelial cell barrier dysfunction is modulated by p38 MAP kinase activities in an MLC phosphorylation-independent fashion by interfering with actin stress fiber formation.
Role of p38 MAP kinase activation in thrombin-induced caldesmon phosphorylation and dissociation of caldesmon-myosin complex.
To further elucidate the role of p38 MAP kinase activation in thrombin-induced endothelial cell permeability, we next examined whether the regulatory protein of smooth muscle and nonmuscle cell contraction, caldesmon, participates in p38 MAP kinase-triggered signaling pathways relevant to the thrombin-induced actin cytoskeletal rearrangement. Confluent BPAEC were labeled with [32P]orthophosphate for 2 h, pretreated with either vehicle or SB-203580 (10 μM, 30 min), and challenged with thrombin (100 nM, 30 min) in a serum-free medium, followed by caldesmon immunoprecipitation as described in materials and methods. As shown in Fig. 7, thrombin-induced caldesmon phosphorylation at 30 min was attenuated by p38 MAP kinase inhibition, suggesting caldesmon phosphorylation by p38 MAP kinase after thrombin challenge. To further characterize the specificity of p38 MAP kinase-induced caldesmon phosphorylation, we used phosphospecific anti-caldesmon antibodies raised against the conserved MAP kinase-dependent phosphorylation sites on mammalian caldesmon, namely Ser759 and Ser789, based on the human high-molecular-weight caldesmon sequence (22). Confluent BPAEC pretreated with either vehicle or SB-203580 (10 μM, 30 min) were stimulated with thrombin (100 nM) for various periods of time in a serum-free medium. BPAEC showed no detectable caldesmon phosphorylation at Ser759, whereas thrombin-induced caldesmon phosphorylation at Ser789 in a biphasic manner correlated with sequential ERK and p38 MAP kinase time-dependent activation and was diminished by SB-203580 pretreatment (Fig. 8). Together, these results suggest that p38 MAP kinase is involved in caldesmon phosphorylation on the MAP kinase-dependent Ser789 site.
Taking into account that caldesmon phosphorylation by p38 MAP kinase may release actin-activated myosin ATPase activity and thereby change caldesmon tethering properties, we examined caldesmon-nondenaturing immunoprecipitates for the presence of actin and myosin. Confluent BPAEC were pretreated with either vehicle or SB-203580 (10 μM, 30 min), stimulated with either vehicle or thrombin (100 nM, 30 min), and immunoprecipitated under nondenaturing conditions as described in materials and methods. As shown in Fig. 9, vehicle-treated endothelial cells contained caldesmon-bound myosin, whereas thrombin produced dissociation of this complex, suggestive of a loss of tethering forces provided by caldesmon binding. Inhibition of p38 MAP kinase abolished effect of thrombin on caldesmon-myosin interaction. Whereas we have observed changes in caldesmon binding to myosin, we did not detect significant alterations in the amount of actin immunoprecipitated with caldesmon between treated groups of cells. Collectively, these results may be an indication of the p38 MAP kinase-mediated regulation of actomyosin interaction and cell contraction via caldesmon phosphorylation.
Thrombin-induced endothelial cell barrier dysfunction occurs via a mainly paracellular route associated with formation of intercellular gaps as a result of altered equilibrium between competing tethering (cell-cell, cell-extracellular matrix contacts) and intracellular contractile forces (7). The latter (or both) is closely related to a rapid actin cytoskeleton rearrangements, which include cortical actin dissolution, stress fiber formation, and cell contraction, ultimately leading to increased vascular permeability via intercellular gaps. Given a complex process of actin stress fiber formation and the fact that thrombin induces numerous signaling cascades, it is expected that actomyosin interaction would be modulated by multiple pathways.
Several groups have previously demonstrated a critical role for MLCK activation and MLC phosphorylation in thrombin-induced actin cytoskeleton rearrangement, gap formation, and increases in endothelial cell permeability (10, 16). However, a contributory role of several kinases, other than MLCK, which represent MLC phosphorylation-independent mechanisms of endothelial cell barrier regulation also has been proposed (2, 3, 38).
We have recently showed that p42/p44 MAP kinase activities modulate thrombin-induced endothelial cell permeability in an MLCK activation-independent manner (2). In the present study, we report that thrombin-induced p38 MAP kinase activation is also involved in the regulation of endothelial cell permeability possibly via caldesmon phosphorylation and interfering with actin stress fiber formation.
p38 MAP kinase activation regulates a variety of cellular processes including inflammation, apoptosis, differentiation, cell cycle progression, and cell migration and contraction (17, 20, 32, 45). p38 MAP kinase downstream targets contain numerous transcription factors and protein kinases, as well as cytoskeletal proteins regulated by p38 MAP kinase-dependent phosphorylation, such as heat shock protein (Hsp) 27 and caldesmon. Among the p38 group of MAP kinase members including four (p38α, p38β, p38γ, and p38δ) isoforms, p38α and p38β are ubiquitously expressed and, unlike p38γ and p38δ, specifically inhibited by SB-203580 (9). Because p38α and p38β are the predominantly expressed isoforms in endothelium, they are responsible for most of the p38 MAP kinase-mediated signaling events (19). Analogously to smooth muscle (27), we report here that thrombin induces p38 MAP kinase activation in endothelium in a time-dependent manner. However, unlike in smooth muscle, the maximal effect occurs later, at 30 min. In addition, thrombin-induced p38 MAP kinase phosphorylation assessed by phosphospecific antibodies, recognizing dual phosphorylation at the TGY motif in a regulatory loop of this kinase, was not changed by SB-203580 pretreatment. As it was reported earlier (13), this indicates MKK3/MKK6-dependent p38 MAP kinase activation after thrombin challenge. Further upstream events leading to activation of this kinase in thrombin-stimulated endothelium are less understood and supposedly involve PKC, tyrosine kinase, and Ras, but not Rac, activities (27).
As we have shown here, thrombin-induced p38 MAP kinase activation is involved in the regulation of endothelial barrier function, since specific inhibition with either SB-203580 or dominant negative p38α attenuated declines in transendothelial electrical resistance after thrombin challenge. Moreover, these studies suggest an MLCK activation-independent contribution of p38 MAP kinase activities to endothelial barrier failure. This is based on several observations, including the nonsignificant effect of both SB-203580 and dominant negative p38α on the levels of MLC diphosphorylation and totally different time course of MLCK and p38 MAP kinase activation in thrombin-stimulated endothelium (Ref. 40 and present study). For this reason, some inhibition of MLC diphosphorylation detected at later time points (after 1 h of thrombin stimulation of SB-203580-treated cells) hardly can be considered as MLCK activation-dependent regulation of endothelial barrier function by p38 MAP kinase.
We have shown here that thrombin-induced actin cytoskeleton rearrangement and stress fiber formation are attenuated by specific p38 MAP kinase inhibition with SB-203580, suggesting p38 MAP kinase-dependent regulation of actin dynamics in bovine endothelium. Actin cytoskeleton targets of p38 MAP kinase-mediated activation include Hsp27 and caldesmon. Hsp27 behaves as a phosphorylation-regulated filamentous actin capping protein, capable of inhibiting actin polymerization in its nonphosphorylated form by binding to barbed ends of filamentous actin and releasing actin polymerization after phosphorylation (1, 35, 36). Phosphorylation of Hsp27 plays an important role in the regulation of actin cytoskeleton remodeling, as well as cell migration, membrane blebbing, and cell contraction (18, 23–25, 28–30, 34). p38 MAP kinase-dependent caldesmon phosphorylation is involved in smooth muscle migration (17) and may take place in the regulation of endothelial cell contraction mediated by actin stress fibers. Our data indicate that p38 MAP kinase contributes to thrombin-induced caldesmon phosphorylation, including the MAP kinase-dependent phosphorylation site (Ser789). It is reasonable to suggest that p42/p44 MAP kinase is responsible for early time points of caldesmon phosphorylation at Ser789, whereas p38 MAP kinase-dependent phosphorylation takes place later, given the different time course of p42/p44 and p38 MAP kinase activation in the endothelium after thrombin challenge. Most likely, caldesmon is phosphorylated by p38 MAP kinase directly, since it was shown in vitro (21) and, as shown in our experiments, occurs at an MAP kinase-dependent site. Caldesmon is an important regulatory protein of cell contraction, residing in stress fibers of thrombin-treated endothelial cells (37). As it was demonstrated, MAP kinase-dependent caldesmon phosphorylation interferes with smooth muscle cell contraction (6, 14, 15, 42). Consistent with these data, we observed that p38 MAP kinase inhibition preserves the caldesmon-myosin-actin complex (reflecting tethering properties of caldesmon on actomyosin interaction) in thrombin-treated endothelial cells, which may represent a p38 MAP kinase-dependent mechanism of cell contraction. Given the complex process of actin stress fiber assembly, it appears that multiple signaling pathways participate in actin cytoskeleton rearrangement. Whereas p38 MAP kinase-mediated Hsp27 phosphorylation has an impact on actin polymerization, p38 MAP kinase-dependent caldesmon phosphorylation may be a prerequisite for functional stress fiber formation.
Collectively, these results provide evidence that thrombin-induced endothelial cell barrier dysfunction is modulated by p38 MAP kinase activation in an MLC phosphorylation-independent manner. Thrombin-induced p38 MAP kinase activation leads to caldesmon phosphorylation, which by interfering with actin stress fiber formation, actomyosin interaction, and cell contraction, results in increased endothelial permeability.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-58064, HL-67307, HL-68062, the Dr. David Marine Endowment, and American Heart Association.
We gratefully acknowledge Lakshmi Natarajan and Nurgul Moldobaeva for superb technical assistance as well as Dr. K. L. Schaphorst (Johns Hopkins University School of Medicine, Baltimore, MD) for providing Epool software.
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 © 2004 the American Physiological Society