RhoA GTPases modulate endothelial permeability. We have previously shown that adenosine and homocysteine enhance basal barrier function in pulmonary artery endothelial cells by a mechanism involving diminution of RhoA carboxyl methylation and activity. In the current study, we investigated the effects of adenosine and homocysteine on endothelial monolayer permeability in cultured monolayers. Adenosine and homocysteine significantly attenuated thrombin-induced endothelial barrier dysfunction and intercellular gap formation. We found significantly diminished RhoA associated with the membrane subcellular fraction in endothelial cells pretreated with adenosine and homocysteine, compared with vehicle-treated endothelial cells. Additionally, adenosine and homocysteine significantly blunted RhoA activation following thrombin exposure. Incubation with adenosine and homocysteine also enhanced in vitro interactions between RhoA and RhoGDI, as well as subcellular translocation of p190RhoGAP to the cytosol. These data demonstrate that elevated intracellular concentrations of homocysteine and adenosine enhance endothelial barrier function in cultured endothelial cells isolated from the main pulmonary artery and lung microvasculature, suggesting a potentially protective effect against pulmonary edema in response to lung injury. We speculate that homocysteine and adenosine modulate the level of endothelial barrier dysfunction through modulation of RhoA posttranslational processing resulting in diminished GTPase activity through altered interactions with modulators of RhoA activation.
- vascular permeability
- lung edema
- RhoA GTPase
- guanine nucleotide dissociation inhibitors
- GTPase-activating proteins
increased endothelial monolayer permeability is observed in inflammatory states, leading to tissue edema and organ dysfunction. Identification of agents protecting against endothelial barrier dysfunction could have therapeutic applications against increased vascular edema caused by sepsis and acute respiratory distress syndrome. Studies of changes in endothelial monolayer permeability in vitro are useful for investigating molecular mechanisms important in modulating barrier function, since such studies lend insight into vascular changes in vivo.
Rho GTPases are a family of enzymes within the Ras superfamily. Of the Rho GTPases, RhoA, Rac-1, and cdc42 have been best characterized with regard to modulation of endothelial monolayer permeability. Although cdc42 has been shown not to be required in endothelial barrier function modulation (25), evidence has accumulated showing requirements for RhoA in basal and thrombin- and histamine-induced increases in monolayer permeability (5, 18, 25). RhoA has been shown to regulate endothelial barrier dysfunction through signaling pathways both dependent on and independent of Rho-kinase. The RhoA/Rho-kinase-dependent pathway modulates the state of myosin light chain phosphorylation and formation of stress fibers (5, 7, 18, 22–24). However, signaling involved in the RhoA/Rho-kinase-independent pathway regulating endothelial barrier dysfunction is not yet well described.
Thrombin-mediated activation of protease-activated receptors causes increased permeability pulmonary edema (13). The intracellular signaling mechanisms by which thrombin promotes endothelial monolayer permeability include the generation of second messengers and contractile forces, with concomitant disruption of intercellular and extracellular adhesive forces (4, 17). We have previously shown that adenosine and homocysteine enhance pulmonary artery endothelial barrier function through a signaling mechanism involving inhibition of isoprenylcysteine-O-carboxyl methyltransferase (ICMT) with subsequent diminution of RhoA GTPase carboxyl methylation and activity (12). Because RhoA GTPases are known regulators of thrombin-induced endothelial barrier dysfunction (5, 18, 25), in the current study we investigated the effects of adenosine and homocysteine concentrations on endothelial monolayer permeability in cultured endothelial monolayers isolated from lung macrovasculature and microvasculature.
MATERIALS AND METHODS
Cell lines, reagents, and animals.
Endothelial cells were obtained from bovine main pulmonary arteries (PAEC) as previously described (3), and those from human lung microvasculature were purchased from Cambrex (Walkersville, MD).
Adenosine and dl-homocysteine were purchased from Sigma (St. Louis, MO). Antibodies directed against RhoA and p190Rho GTPase-activating protein (GAP) were purchased from Santa Cruz Biotechnologies (Santa Cruz, CA) and BD Biosciences (San Diego, CA), respectively. Thrombin, horseradish peroxidase (HRP), lysophosphatidic acid (LPA), and O-phenylenediamine HCl were supplied by Sigma. β-Catenin-specific antibodies were purchased from Transduction Laboratories (San Diego, CA).
The pGST-C21 construct was a generous gift from Dr. John G. Collard (The Netherlands Cancer Institute) (15). The glutathione S-transferase-Rho guanine-nucleotide dissociation inhibitor (GST-RhoGDI) plasmid was a generous gift from Dr. Allan Hall (University College London) (16).
Endothelial monolayer permeability assay.
Endothelial monolayer permeability was assayed by the rate of diffusion of HRP through endothelial monolayers grown on collagen-coated polycarbonate membranes and by changes in resistance of endothelial monolayers grown on collagen-coated gold electrodes using the electrical cell impedance system (ECIS; Applied Biophysics, Troy, NY), as previously described (8, 12). HRP transport was assessed using 0.4-μm pore polycarbonate Transwell supports with confluent endothelial cell monolayers. Thrombin and HRP were placed in the upper chamber, and the appearance of HRP in the lower chamber was determined by retrieving aliquots of medium from the lower compartment every 30 min over 3 h. The HRP concentration was determined by spectrophotometric assay, and the data are presented as the number of moles of HRP that diffused to the lower chamber over time. The second method was performed using ECIS with confluent PAEC monolayers on gold electrodes. When resistance across the endothelial monolayer had reached ≥500 ohms, the cells were incubated as described in RhoA GTPase activation and subcellular localization assays.
Immunofluorescence analyses and intercellular gap determination.
Endothelial cells grown on coverslips were treated as described in the legend for Fig. 3. The cells were fixed with 4% paraformaldehyde and rendered permeable with Triton X-100. The cells were immunofluorescently stained for β-catenin, as previously described (8, 12). Images were viewed at ×1,000 magnification under a Nikon Eclipse E400 fluorescence microscope and recorded. Images were converted to grayscale in Adobe Photoshop 7.0, and intercellular gaps were identified and highlighted. The intercellular gaps were then analyzed with Optimas 6.5 image processing program to obtain the area of each gap, and the data are presented as the percentage of area taken up by intercellular gaps relative to the total area of the field examined.
RhoA GTPase activation and subcellular localization assays.
The endothelial cells were grown to confluence. The cells were preincubated for 4 h with vehicle or adenosine and homocysteine and then treated with or without thrombin for the indicated times. Endothelial cells were washed once with PBS and lysed in FISH buffer (10% glycerol, 50 mM Tris, pH 7.4, 100 mM NaCl, 1% Nonidet P-40, and 2 mM MgCl2) as previously described (8, 12, 15). Cell lysates were incubated on ice for 10 min and insoluble debris was removed by centrifuging at 13,000 g for 5 min at 4°C. Equivalent volumes of supernatants were incubated with 50 μg of bacterially produced GST-C21 or GST-RhoGDI bound to glutathione agarose beads for 1 h at 4°C. The beads were washed with FISH buffer and suspended in Laemmli buffer. Protein complexes bound to the beads were resolved on 15% SDS-PAGE. Parallel gels were run with corresponding crude cell lysates. All gels were transferred to Immobilon-P and probed with antibodies directed against RhoA.
Subcellular localization analyses of RhoA or p190RhoGAP were done as previously described (9). Cells were lysed in 20 mM Tris·Cl, pH 7.5, 3 mM MgCl2, 1 mM EGTA, 1 mM DTT, 1 mM PMSF, 10 μg/ml leupeptin, and 10 μg/ml aprotinin, sonicated, and centrifuged. The supernatant (cytosolic fraction) was removed, and the insoluble pellet (membrane/cytoskeletal fraction) was solubilized by sonication in the lysis buffer supplemented with 2% Triton X-100. The samples were then centrifuged. The supernatant (membrane fraction) was removed, and the pellet (cytoskeletal fraction) was solubilized by sonication in the lysis buffer supplemented with 4% Triton X-100. Equivalent amounts of cytosolic, membrane, and cytoskeletal proteins were resolved on SDS-PAGE and immunoblotted for RhoA or p190RhoGAP.
For permeability experiments, linear regression analysis was performed to determine the rate of HRP diffusion in individual wells with Statview 4.0. The means and SE were determined from these slopes. 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 or an unpaired t-test. Significance was reached when P < 0.05. All data are presented as means ± SE. The n is indicated for each set of data.
Adenosine and homocysteine attenuated thrombin-induced endothelial barrier dysfunction.
We have previously shown that elevated concentrations of adenosine and homocysteine enhance basal barrier function of PAEC monolayers (12). In the current studies, we investigated the effects of adenosine and homocysteine on thrombin-induced barrier dysfunction in endothelial cells derived from bovine pulmonary artery (Fig. 1) and in human lung microvascular endothelial cells (Fig. 2). Thrombin significantly enhanced monolayer permeability across both bovine pulmonary artery and human lung microvascular endothelial monolayers, as demonstrated by a decreased electrical resistance across the monolayer (Figs. 1 and 2). Preincubation of the endothelial cells with adenosine and homocysteine significantly blunted thrombin-induced barrier dysfunction in both endothelial cell types (Figs. 1 and 2). Parallel studies measured the rate of HRP diffusion, correlating with paracellular flux, through bovine pulmonary artery monolayers by the Transwell system. Monolayers preincubated with 50 μM adenosine and 50 μM homocysteine and exposed to thrombin had a significantly lower rate of HRP diffusion (8.09 ± 0.6 mol HRP/min), compared with monolayers treated with thrombin only (12.5 ± 0.7 mol HRP/min) (n = 4, P < 0.0001).
Immunofluorescence staining for β-catenin demonstrated intercellular gap formation following 10 min of thrombin exposure (Fig. 3), indicating disruption of adherens junctions. PAEC preincubated with adenosine and homocysteine before thrombin exposure displayed fewer interendothelial cell gaps (Fig. 3). Quantitation of the immunofluorescence images demonstrated a significant reduction in thrombin-induced intercellular gaps at 10 and 30 min in endothelial cell monolayers pretreated with adenosine and homocysteine (Fig. 3B). Thus adenosine and homocysteine blunted thrombin-induced barrier dysfunction by inhibiting adherens junction disruption and intercellular gap formation.
Adenosine/homocysteine attenuates thrombin-induced RhoA activation through changes in RhoA interactions with modulators of activation.
RhoA GTPases are known key regulators of thrombin-induced endothelial barrier dysfunction (5, 26). In previous work, we showed that inhibition of ICMT with adenosine and homocysteine resulted in diminution of RhoA GTPase methylation and activity (12); thus we next investigated if adenosine and homocysteine modulated endothelial monolayer permeability by altering RhoA GTPase subcellular localization, activity, and/ or its association with modulators of activation. Activated RhoA has been shown to localize to the plasma membrane of cells. We therefore analyzed the level of RhoA in the membrane fraction of the cell lysates compared with total RhoA protein (Fig. 4A). Densitometric analyses reveal that PAEC incubated with adenosine and homocysteine for 4 h had a significantly diminished amount of RhoA protein associated with the membrane fraction compared with vehicle-treated cells (Fig. 4A). No detectable differences were noted in the amount of RhoA in the membrane fractions of cells exposed to thrombin for 1, 10, or 30 min; however, a 10-min exposure to 2 μg/ml LPA increased RhoA translocation to the membrane fraction of these cells (data not shown). Affinity precipitation assays of GTP-bound RhoA, however, demonstrated a biphasic response with significant increases in RhoA activity at 1 and 30 min following thrombin exposure (Fig. 4B). Preincubation of the endothelial cells with adenosine and homocysteine inhibited thrombin-induced RhoA GTPase activity (Fig. 4B). It is possible that the discrepancy between the subcellular localization results and affinity precipitation data lies in differences in the level of sensitivity of the two techniques for detecting activated RhoA.
Studies assessing the effects of adenosine and homocysteine exposure on RhoA modulator molecules demonstrated an increase in RhoA association with RhoGDI, indicating an enhanced level of GDP-bound RhoA (Fig. 5A). In addition, 4-h incubation with adenosine and homocysteine promoted the subcellular translocation of the Rho GAP, p190RhoGAP, from the cytoskeleton fraction to the cytosolic fraction (Fig. 5, B and C), correlating with the subcellular relocalization of RhoA in response to adenosine and homocysteine incubation (Fig. 4A). Thus it is possible that adenosine and homocysteine modulate thrombin-induced RhoA activation through effects on RhoA carboxyl methylation and subsequent association with modulators of activation.
We have shown previously that adenosine and homocysteine diminish RhoA carboxyl methylation and activity and enhance basal barrier function in PAEC (12). In the current study, we show that adenosine and homocysteine significantly attenuated thrombin-induced endothelial monolayer permeability, RhoA activation, and intercellular gap formation. In addition, adenosine and homocysteine exposure enhanced the binding of RhoA to RhoGDI and promoted relocalization of both RhoA and p190RhoGAP to the cytosol. Our results suggest that homocysteine and adenosine may modulate agonist-induced barrier dysfunction through inhibition of RhoA GTPase via effects on RhoA carboxyl methylation and its subsequent association with modulators of activation.
RhoA GTPases have been shown to regulate endothelial barrier dysfunction (2, 5, 7, 18, 19, 22, 24, 25). RhoA posttranslational processing (prenylation, proteolysis, and carboxyl methylation) is important in modulating subcellular localization, enzyme activation, and protein stability (1, 10, 12). Farnesyl pyrophosphate and geranylgeranyl pyrophosphate are prenyl moieties covalently bound to RhoA GTPase and are intermediates in the conversion of mevalonic acid into cholesterol. Inhibitors of 3-hydroxy-3-methylglutaryl coenzyme A (HMG CoA) reductase, statins, block the conversion of HMG CoA to mevalonic acid. The statins are clinically effective in reducing cardiovascular diseases by lowering cholesterol synthesis. These drugs also alter endothelial cell function by inhibition of posttranslational prenylation, hence inactivating Ras, RhoA, Rac-1, cdc42, and other small GTPases (11, 20). Indeed, preincubation with simvastatin inhibited thrombin-induced RhoA activation (6, 21). Simvastatin also diminished stress fiber and focal contact formation and enhanced barrier function in thrombin-exposed endothelial cells (21). Thus inhibition of small GTPase prenylation alters thrombin-induced changes in monolayer permeability.
RhoA is also carboxyl methylated posttranslationally by S-adenosylmethionine (SAM)-dependent methyltransferases. In previous studies, we have shown that overexpression of ICMT, a SAM-dependent methyltransferase, enhanced the level of endogenously methylated RhoA and activated RhoA and increased basal monolayer permeability in endothelial cells (12). SAM-dependent methyltransferases may be inhibited by enhancing intracellular S-adenosylhomocysteine by increasing adenosine and homocysteine intracellular concentrations or by an isoprenylated competitive inhibitor, N-acetyl-geranylgeranyl-l-cysteine (AGGC). We previously showed that ICMT inhibition with adenosine and homocysteine or AGGC attenuated the RhoA methylation and activation and increased endothelial basal barrier function (12). In the current study, adenosine and homocysteine significantly attenuated thrombin-induced endothelial monolayer permeability, RhoA activation, and intercellular gap formation.
Small GTPases act as molecular switches, being active when bound to GTP and localized to the cell membrane (14). GTPase activation is regulated by a number of molecular modulators, such as RhoGDI and p190RhoGAP. Thus there are several possible mechanisms by which RhoA carboxyl methylation might alter GTPase activation. Ras GTPase methylation has been shown to enhance hydrophobicity and thereby increase membrane localization (9). Our results indicate that adenosine and homocysteine inhibited both Ras (9) and RhoA membrane localization, and we have previously reported that adenosine and homocysteine decrease RhoA carboxyl methylation (12). RhoA carboxyl methylation also alters the half-life of this GTPase (1). However, diminished degradation of undercarboxyl-methylated RhoA does not explain our results, since we found less activated RhoA, when normalized to the total level of RhoA. Carboxyl methylation could also alter RhoA activation by changing interactions with modulator molecules. We found that adenosine and homocysteine enhanced RhoA association with RhoGDI in vitro. This is likely due to enhanced cytosolic subcellular localization of RhoA in the GDP-bound state. Furthermore, adenosine and homocysteine increased cytosolic subcellular localization of p190RhoGAP, a finding that is also consistent with cytosolic subcellular localization of RhoA.
Thus enhanced RhoA interaction with RhoGDI upon adenosine and homocysteine exposure correlates with increased cytosolic levels of RhoA. We speculate that increased intracellular concentrations of homocysteine and adenosine modulated endothelial barrier dysfunction resulting from exposure to thrombin by inhibiting ICMT-mediated RhoA carboxyl methylation, hence activation. These effects may be mediated by modulation of RhoA interactions with modulators of activation, such as the GDI or GAP molecules. These studies do not, however, exclude other potential mechanisms of effects of adenosine plus homocysteine, such as effects mediated by adenosine receptors.
In summary, these results demonstrate that increased homocysteine and adenosine enhanced endothelial barrier function in cultured monolayers isolated from lung macrovasculature and microvasculature. The effects of adenosine and homocysteine on thrombin-induced barrier dysfunction and RhoA activation suggest that inhibition of posttranslational carboxyl methylation of RhoA modulates endothelial barrier function. We speculate that carboxyl methylation of RhoA also alters lung microvascular permeability in vivo.
This material is the result of work supported with resources and the use of facilities at the Providence VA Medical Center and supported with National Heart, Lung, and Blood Institute Grants HL-64936 to S. Rounds and HL-67795 to E. O. Harrington.
The authors thank Drs. J. Collard and A. Hall for the GST-C21 and GST-RhoGDI constructs, respectively. Some of these results were presented at the American Thoracic Society international conference and are published in abstract form in the American Journal of Respiratory and Critical Care Medicine 167: A565, 2003.
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