Articles |
From the Gaubius Laboratory TNO-PG, Leiden, and the Institute for Cardiovascular Research, Free University, Amsterdam, the Netherlands.
Correspondence to Prof V. van Hinsbergh, Gaubius Laboratory TNO-PG, Zernikedreef 9, PO Box 2215, 2301 CE Leiden, Netherlands.
Key Words: endothelium permeability phosphorylation
| Introduction |
|---|
|
|
|---|
| Nature of the Endothelial Barrier |
|---|
|
|
|---|
Electron microscopic evidence strongly suggests that in tissue capillaries, hormones and macromolecules are shuttled across the endothelial barrier via vesicles, in which these biomolecules may accumulate owing to receptor binding.1 By such mechanisms can an adequate supply of hormones and albumin be provided to the different tissues. However, according to Starling's law, the extravasation of fluid and macromolecules is directly related to hydrostatic and osmotic pressure. Only when vesicles fuse into cell-spanning pores2 is it understandable how an increase in hydrostatic pressure gradient increases vesicle-mediated exchange. However, it is plausible that the flux of macromolecules from the blood to the interstitium that contributes to the lymphatic fluid also involves other structures, probably located in another segment of the vascular tree. The postcapillary venules are good candidates for this because their ECs have rather simple intercellular junctions, and minute gaps within them that cause vascular leakage are formed on exposure of a postcapillary venule to vasoactive agents.
In the early 1960s, Majno and Palade3 showed that after exposure of tissue to histamine, carbon particles injected into the blood compartment left it selectively via postcapillary venule ECs, between which gaps were occasionally seen. In a subsequent study, Majno et al4 observed that the nucleus of these ECs had a wrinkled appearance and postulated that EC contraction was the basis of the increased extravasation of macromolecules. This concept has now been generally accepted, and subsequent biochemical and cell studies have indicated that both actin-myosin interaction5 6 7 8 9 and reversible disintegration of intercellular tight and adherens junctions10 11 are involved. The formation of small intercellular gaps is shared by many other vasoactive agents and is particularly important in edema formation during acute inflammation and the capillary leakage syndrome.
In addition, other forms of vascular leakage may occur. After prolonged inflammation, capillaries may become leaky, a phenomenon that is probably related to an increase in angiogenesis.12 The angiogenic vascular endothelial growth factor causes serious edema when applied to the skin of animals and can induce clusters of connected vesicular structures, the so-called vesicular vacuolar organelles, and pores within EC (ie, VEGF). It makes the endothelial monolayer leaky13 14 and alters protein phosphorylation in adherens junctions.15 Impairment of endothelial barrier function can also occur in arterioles after prolonged exposure to histamine.16 It is not yet known whether the loss of tight junction integrity recently demonstrated in cultured ECs after prolonged exposure to histamine11 contributes to this phenomenon. Furthermore, the endothelial permeability of arteries and veins can occasionally increase. Massive endothelial injury and desquamation occur only in exceptional cases, eg, after introduction of a bacterial toxin, such as in pig edema disease. More frequently, focal leaky spots are encountered in the endothelium of arteries and veins. They are found after exposure of the vessel to injurious conditions, such as hypercholesterolemia17 and atherosclerosis.18 These focal leaky spots are frequently associated with EC division19 or occasionally with leaky junctions.20
| Actin-Myosin Interaction and Endothelial Gap Formation |
|---|
|
|
|---|
The mechanism of contraction in ECs is comparable to that in SMCs, although many of the proteins involved in the contraction process and its regulation are different in the two cell types (see Reference 2222 for a review). As in SMC contraction, the interaction between actin and nonmuscle myosin in ECs is regulated by the phosphorylation status of MLC7 8 9 23 (Fig 1
). Two MLCs, one structural and one regulatory, are bound to the head part of myosin heavy chain. By the action of MLCKs, the regulatory MLC is monophosphorylated or diphosphorylated. The phosphorylated MLC activates the myosin molecule, after which it interacts with F-actin filaments and can then cause contraction by sliding along the actin filament.
|
When ECs in vitro are stimulated by histamine or thrombin, which induces a transient or prolonged increase in endothelial permeability, respectively, monophosphorylated and diphosphorylated MLCs are generated. Within 1 to 2 minutes phosphorylation reaches a maximum, which follows an increase in the cytoplasmic Ca2+ concentration. The induced increase in permeability can be partially inhibited by intracellular Ca2+ chelators or inhibitors of calmodulin and MLCK.6 7 9 23 24 Therefore, it is generally believed that Ca2+/calmodulindependent PK I (the classic MLCK) plays a central role in the onset of endothelial gap formation after cell exposure to vasoactive agents. In close agreement with these data, it has been observed in vivo that a transient increase in cytoplasmic Ca2+ concentration parallels an increase in endothelial permeability after exposure of frog mesenteric microvessels to histamine.25 In this case the endothelial barrier recovers within several minutes.
| MLCKs and Phosphatases |
|---|
|
|
|---|
210 kD but represents an alternatively spliced product of the same gene as that of smooth muscle MLCK (Genbank access No. U 48959, 1996). In other types of mesenchymal cells, it was recently demonstrated that Rho kinase can also act as an MLCK28 and thus can contribute to stress fiber contraction29 and tethering of actin fibers to the cell membrane.30 It is likely that Rho kinase is also present in ECs. This brings the number of MLCK (iso)forms to four. It has been demonstrated that the classic MLCK and its endothelial splice variant are activated by Ca2+/calmodulin, whereas activation of Rho kinase depends on RhoA and tyrosine phosphorylation. No information is yet available about the presence of embryonic MLCK in other types of ECs than those from the umbilical cord or about its regulation. Also unknown is the distribution of the various potential MLCKs in ECs. Future studies have yet to reveal where these MLCKs are located and which of them are involved in the organization of the actin cytoskeleton, formation of filopodia and membrane ruffles, stress fiber tension, and induction of gaps in EC junctions. The phosphorylation state of MLCs depends not only on its phosphorylation by MLCKs but also on its dephosphorylation by MLC phosphatase. Indeed, ample evidence exists that in SMCs, inactivation of MLC phosphatase also can enhance the phosphorylation state of MLC and the contractile state of the cell.31 The involvement of a type 1 myosinassociated protein phosphatase in the regulation of endothelial barrier function was recently demonstrated in bovine pulmonary artery ECs.32 In this context, it is also of interest that RhoA not only activates MLCK but also binds to the regulatory subunit of MLC phosphatase and that this binding reduces dephosphorylation of MLC.33
Experiments with calcium ionophore7 have shown that elevation of cytoplasmic Ca2+ concentration per se is insufficient to induce endothelial contraction. However, these data should be interpreted with caution, because it is known from studies in SMCs that the high concentration of Ca2+ generated by the ionophore activates Ca2+/calmodulindependent PK II, which interferes with MLC phosphorylation (see Reference 3131 ).
| Additional Mechanisms Contributing to Prolonged Permeability |
|---|
|
|
|---|
Recent measurements of isometric tension in ECs demonstrated that the thrombin-increased isometric tension was accompanied by MLC phosphorylation to a considerable extent, whereas histamine had a much smaller effect8 9 The effects of thrombin were inhibited by cytochalasin D and an MLCK inhibitor, suggesting that actin and MLC phosphorylation play a role in thrombin-induced isometric tension. These data strongly suggest that an increase in actin-myosindependent isometric tension contributes to the prolonged increase in endothelial permeability. However, two aspects remain to be clarified.8 9 First, it is possible that thrombin acts on another MLCK in addition to the Ca2+/calmodulindependent MLCK that is also expected to be activated by histamine. This may explain the striking difference between the effects of thrombin and histamine. Alternatively, histamine may activate MLCK in only selected areas of the cell. Second, the involvement of isometric contraction in thrombin-induced permeability does not exclude the simultaneous involvement of other mechanisms. In particular, signal transduction via PKC and protein tyrosine phosphorylation has been implicated in the regulation of endothelial permeability.
Several authors have reported that activation of PKC contributes to the thrombin-induced increase in permeability of bovine ECs.34 35 In human EC monolayers, platelet-activating factorinduced permeability was increased by PKC activation,36 but direct activation of PKC by phorbol ester reduced endothelial permeability.37 Overexpression of PKC ß-I isotype in ECs by transfection of cDNA increased the permeability in these cells, further supporting the role of PKC in the regulation of endothelial permeability.38 However, in contrast to direct MLC phosphorylation by PKC in SMCs, activation of PKC by phorbol ester did not directly cause phosphorylation of MLC in ECs.7 Equally well, the pattern of phosphorylation of individual amino acids in MLC by thrombin indicated phosphorylation by MLCK but not by PKC.8 This finding suggests that PKC either indirectly facilitates the increase in permeability induced by vasoactive agents or acts on other factors that contribute to the maintenance of endothelial barrier function, such as the integrity of cell-cell junctions and cell-matrix contacts. In line with this latter suggestion, Rabiet et al10 reported that thrombin disrupted the VE-cadherin-catenin complex in adherens junctions and that this effect could be prevented by inhibition of PKC or the tyrosine PK inhibitor herbimycin A. This suggests that in addition to the contraction mechanism that involves actinnonmuscle myosin interaction, disintegration of adherens junctions also contributes to the increased permeability observed in vitro. This disintegration is reversible after several hours.10 The time course of its induction and reversal is on par with that of thrombin-induced increase in permeability. These data suggest that disintegration of adherens junctions between cells and also the possible loss of focal contacts between the cell and its matrix may contribute largely to the prolonged leakage induced by thrombin in vitro.
Further studies are needed to verify whether these mechanisms also act in vivo or are overactive in vitro. Published studies on the microcirculation have demonstrated a short-lasting transient leakage in postcapillary venules after stimulation with a vasoactive agent. These observations seem to give little support to the concept of prolonged leakage sites. However, these observations were made in healthy tissues and with a single stimulus. It is indeed conceivable that disintegration of cell-cell contacts and prolonged leakage occurs, particularly in areas of inflammation where leukocytes adhere to ECs and act on them by producing platelet-activating and other factors.39 If this is true, then it is important to elucidate its mechanism and regulation, because a different approach may be required to treat such cases of vascular leakage.
| Stabilization of the Endothelial Barrier In Vivo |
|---|
|
|
|---|
Other experiments have indicated that platelet-release products are important for maintaining endothelial barrier function.41 Vascular leakage is associated with gray platelet syndrome, a disease in which platelets do not release their contents. Thrombocytopenic animals suffer from vascular leakage and microvascular bleeding, but endothelial barrier function can be restored in these animals by exogenous serotonin. Factors that stabilized endothelial barrier function were also demonstrated in the blood of hemorrhaged animals40 ; such factors included catecholamines, vasopressin, and serotonin.
| Improvement of Endothelial Barrier Function by cAMP |
|---|
|
|
|---|
It should be noted that elevations of cAMP affect endothelial barrier function not only directly but also indirectly by reducing the interaction between leukocytes and the endothelium.47 A number of compounds, such as complement factors 3a and 5a, increase endothelial permeability not directly but via interaction with polymorphonuclear leukocytes, which then bind to and activate the endothelium of postcapillary venules in particular. Indeed, interference of leukocyte-endothelium interaction by antiICAM-1 and anti-CD18 integrin antibodies reduced experimental lung edema in animals.48 This suggests that selective elevation of cAMP in ECs and leukocytes may be a target for reducing vascular leakage. Cell typeselective elevation of cAMP is essential to avoid the unwanted effects of treatment on the heart and kidney. Furthermore, it should be realized that cAMP also acts on SMCs and induces smooth muscle relaxation. This may distally cause an increase in surface area of the capillary bed that is being perfused. As a consequence, cAMP may increase vascular leakage in selected areas of the vascular bed, despite the fact that overall endothelial barrier function improves.49
| Effects of cGMP and NO on Endothelial Permeability |
|---|
|
|
|---|
In cultured aortic ECs, an elevation in cGMP reduced thrombin-enhanced permeability but did not affect basal permeability.53 cGMP also counteracted the H2O2-induced increase in hydraulic conductivity through monolayers of pulmonary artery ECs.54 Subsequent studies have shown that cGMP affects endothelial permeability both directly by activating cGMP-dependent PK24 55 and indirectly by inhibiting and activating cAMP-degrading phosphodiesterases (Fig 2
). The presence of active cGMP-dependent PK I was demonstrated in ECs of large arteries and veins55 56 but was absent in those of the umbilical vein and kidney glomeruli.55 57 Selective activation of cGMP-dependent PK reduced thrombin-induced accumulation of cytoplasmic Ca2+ ions as well as the passage of macromolecules through human aortic EC monolayers, whereas it had no effect on umbilical vein ECs.24
|
The fact that cGMP counteracts increases in cytoplasmic Ca2+ concentration has an interesting consequence. The increase in cytoplasmic Ca2+ concentration generated by vasoactive agents not only affects endothelial permeability but also activates endothelial NO synthase. NO increases the cGMP concentration in the EC, which then again counteracts Ca2+ accumulation and the increase in endothelial permeability. Studies with NO synthase inhibitors have demonstrated that thrombin-induced NO indeed counteracts endothelial permeability increases in aorta and pulmonary artery ECs. Limitation of the extent of Ca2+-activated processes in these cells by NO also explains the observation that NO modulates its own Ca2+-dependent formation.58
cGMP also affects endothelial permeability indirectly by influencing the cAMP concentration via cAMP-degrading phosphodiesterases. In umbilical vein ECs, cGMP reduced thrombin-enhanced endothelial permeability by inhibiting a cGMP-inhibited phosphodiesterase (PDE III). This effect on the permeability of umbilical vein cells was mimicked by synthetic PDE III inhibitors, whereas these inhibitors had little effect in human aorta and pulmonary artery ECs.24 In addition to PDE III, a cGMP-stimulated phosphodiesterase (PDE II) can be involved, which reduces the concentration of both cAMP and cGMP. Inhibition of PDE II prevents the loss of barrier function induced by H2O2, as shown in pulmonary artery ECs.54 59 This suggests that PDE II can play a significant role in the regulation of endothelial permeability. This may be important, because NO and cGMP frequently increase the permeability of mesenteric and peripheral microvessels in vivo.50 52 60 Unfortunately, little information is available regarding the presence of PDE II and PDE III in different types of ECs in vivo. Therefore, one can only speculate at present whether a difference in the amount of PDE II in ECs of different vascular beds contributes to the difference in the effect of cGMP-elevating agents on vascular leakage in the lung and peripheral vascular beds.
| Perspective |
|---|
|
|
|---|
It is anticipated that in the coming years, the structures and biochemical pathways will be elucidated that underlie differences in the regulation of barrier function of the endothelium in different vascular beds, not only between different tissues such as the brain, heart, and lung, but also within tissues at the level of capillaries, arterioles, and postcapillary venules. More important for clinical practice, the new insights obtained will provide new leads for developing drugs to combat vascular leakage.
| Selected Abbreviations and Acronyms |
|---|
|
| Acknowledgments |
|---|
Received January 28, 1997; accepted February 10, 1997.
| References |
|---|
|
|
|---|
-trinositol exerts its edema-preventing effects through modulation of ß1 integrin function. Circ Res. 1994;75:942-948.This article has been cited by other articles:
![]() |
J. Surapisitchat, K.-I. Jeon, C. Yan, and J. A. Beavo Differential Regulation of Endothelial Cell Permeability by cGMP via Phosphodiesterases 2 and 3 Circ. Res., October 12, 2007; 101(8): 811 - 818. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. Predescu, S. Predescu, J. Shimizu, K. Miyawaki-Shimizu, and A. B. Malik Constitutive eNOS-derived nitric oxide is a determinant of endothelial junctional integrity Am J Physiol Lung Cell Mol Physiol, September 1, 2005; 289(3): L371 - L381. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Cao, A. Eriksson, H. Kubo, K. Alitalo, Y. Cao, and J. Thyberg Comparative Evaluation of FGF-2-, VEGF-A-, and VEGF-C-Induced Angiogenesis, Lymphangiogenesis, Vascular Fenestrations, and Permeability Circ. Res., March 19, 2004; 94(5): 664 - 670. [Abstract] [Full Text] [PDF] |
||||
![]() |
P. Kouklis, M. Konstantoulaki, S. Vogel, M. Broman, and A. B. Malik Cdc42 Regulates the Restoration of Endothelial Barrier Function Circ. Res., February 6, 2004; 94(2): 159 - 166. [Abstract] [Full Text] [PDF] |
||||
![]() |
S. M. Stamatovic, R. F. Keep, S. L. Kunkel, and A. V. Andjelkovic Potential role of MCP-1 in endothelial cell tight junction `opening': signaling via Rho and Rho kinase J. Cell Sci., November 15, 2003; 116(22): 4615 - 4628. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. K. Kiemer, N. C. Weber, R. Furst, N. Bildner, S. Kulhanek-Heinze, and A. M. Vollmar Inhibition of p38 MAPK Activation via Induction of MKP-1: Atrial Natriuretic Peptide Reduces TNF-{alpha}-Induced Actin Polymerization and Endothelial Permeability Circ. Res., May 3, 2002; 90(8): 874 - 881. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. Kanthou and G. M. Tozer The tumor vascular targeting agent combretastatin A-4-phosphate induces reorganization of the actin cytoskeleton and early membrane blebbing in human endothelial cells Blood, March 15, 2002; 99(6): 2060 - 2069. [Abstract] [Full Text] [PDF] |
||||
![]() |
V. W. M. van Hinsbergh NO or H2O2 for Endothelium-Dependent Vasorelaxation : Tetrahydrobiopterin Makes the Difference Arterioscler. Thromb. Vasc. Biol., May 1, 2001; 21(5): 719 - 721. [Full Text] [PDF] |
||||
![]() |
G. P. van Nieuw Amerongen and V. W.M. van Hinsbergh Cytoskeletal Effects of Rho-Like Small Guanine Nucleotide-Binding Proteins in the Vascular System Arterioscler. Thromb. Vasc. Biol., March 1, 2001; 21(3): 300 - 311. [Abstract] [Full Text] [PDF] |
||||
![]() |
D. S. A. Majid, K. E. Said, S. A. Omoro, and L. G. Navar Nitric Oxide Dependency of Arterial Pressure-Induced Changes in Renal Interstitial Hydrostatic Pressure in Dogs Circ. Res., February 16, 2001; 88(3): 347 - 351. [Abstract] [Full Text] [PDF] |
||||
![]() |
B Wojciak-Stothard, S Potempa, T Eichholtz, and A. Ridley 9Rgr; and Rac but not Cdc42 regulate endothelial cell permeability J. Cell Sci., January 4, 2001; 114(7): 1343 - 1355. [Abstract] [PDF] |
||||
![]() |
G. P. v. N. Amerongen, M. A. Vermeer, P. Negre-Aminou, J. Lankelma, J. J. Emeis, and V. W. M. van Hinsbergh Simvastatin Improves Disturbed Endothelial Barrier Function Circulation, December 5, 2000; 102(23): 2803 - 2809. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. P. v. N. Amerongen, M. A. Vermeer, and V. W. M. van Hinsbergh Role of RhoA and Rho Kinase in Lysophosphatidic Acid-Induced Endothelial Barrier Dysfunction Arterioscler. Thromb. Vasc. Biol., December 1, 2000; 20 (12): e127 - e133. [Abstract] [Full Text] [PDF] |
||||
![]() |
X. Gao, P. Kouklis, N. Xu, R. D. Minshall, R. Sandoval, S. M. Vogel, and A. B. Malik Reversibility of increased microvessel permeability in response to VE-cadherin disassembly Am J Physiol Lung Cell Mol Physiol, December 1, 2000; 279(6): L1218 - L1225. [Abstract] [Full Text] [PDF] |
||||
![]() |
G. P. v. N. Amerongen, S. v. Delft, M. A. Vermeer, J. G. Collard, and V. W. M. van Hinsbergh Activation of RhoA by Thrombin in Endothelial Hyperpermeability : Role of Rho Kinase and Protein Tyrosine Kinases Circ. Res., August 18, 2000; 87(4): 335 - 340. [Abstract] [Full Text] [PDF] |