Histamine alters cadherin-mediated sites of endothelial adhesion.

We tested the hypothesis that histamine alters the focal apposition of endothelial cells by acting on sites of cadherin-mediated cell-cell adhesion. Focal apposition was measured as the impedance of a cell-covered electrode, which was partitioned into a cell-matrix resistance, a cell-cell resistance, and membrane capacitance. Histamine causes an immediate, short-lived decrease in the impedance of an electrode covered with human umbilical vein endothelial (HUVE) cells. ECV304 cells are a line of spontaneously transformed HUVE cells that do not express the endothelial cadherin, cadherin-5. Histamine increased ECV304 cell calcium to 600 nM. Histamine did not increase myosin light chain phosphorylation of control or transfected ECV304 cells. ECV304 cells transfected with either E-cadherin or cadherin-5 on a dexamethasone-responsive plasmid (pLKneo) increased their cell-cell resistance when stimulated with dexamethasone, whereas ECV304 cells transfected with pLKneo-lacZ did not. Histamine did not affect the impedance of ECV304 cells transfected with pLKneo-lacZ. In contrast, histamine decreased the cell-cell resistance of ECV304 cells transfected with either pLKneo-E-cadherin or pLKneo-cadherin-5. From these data, we conclude that histamine acts on sites of cadherin-mediated cell-cell apposition.

Thrombin inhibits myosin light chain dephosphorylation in endothelial cells.

Histamine and thrombin increase myosin light-chain kinase-mediated phosphorylation of myosin light chain (MLC) in human umbilical vein endothelial cells (HUVEC). The increase in MLC phosphorylation caused by thrombin persists longer (330 min) than the increase caused by histamine (<5 min), although both increase cell calcium similarly. We hypothesized that some of the longer duration of the increase in MLC phosphorylation caused by thrombin was because of inhibition of myosin dephosphorylation by thrombin. Calyculin A, an inhibitor of type 1 and 2A protein phosphatases, caused a time-dependent increase in MLC phosphorylation in unstimulated HUVEC. As thrombin-stimulated phosphorylation approached its peak at 15 min, calyculin A caused progressively less of an increase in MLC phosphorylation in thrombin-stimulated HUVEC, and no increase at the peak of thrombin stimulation. In HUVEC in which cell calcium was maintained at 600 nM, thrombin increased MLC phosphorylation above the level caused by increased calcium alone at a time coinciding with the peak of thrombin stimulation. However, when phosphatase activity was already inhibited with calyculin A, thrombin did not further increase MLC phosphorylation in cells in which calcium was maintained at 600 nM calcium. Thrombin increases MLC phosphorylation in HUVEC not only by increasing cell calcium but also by inhibiting calyculin A-sensitive dephosphorylation of MLC.

Ionomycin and PDBU increase MDCK monolayer permeability independently of myosin light chain phosphorylation.

It has been hypothesized that modulation of epithelial paracellular permeability may be mediated by initiation of contraction of a band of actin and myosin located at the tight junction. Phosphorylation of myosin light chain (MLC) is an important determinant of actomyosin contraction. We asked if ionomycin (iono) and phorbol 12,13-dibutyrate (PDBU), which increase paracellular permeability of Madin-Darby canine kidney (MDCK) cell monolayers, increased MLC phosphorylation in MDCK cells. MDCK cell MLC was constitutively phosphorylated by myosin light chain kinase (MLCK), and after PDBU and iono > 99% of MLC continued to be phosphorylated by MLCK. Neither iono or PDBU, nor the combination of iono and PDBU, increased MLC phosphorylation. In contrast, the phosphatase inhibitor okadaic acid did increase MLC phosphorylation. Adenosine 3',5'-cyclic monophosphate (cAMP) and forskolin decreased MLC phosphorylation in control MDCK cells and in cells exposed to iono and PDBU. In contrast, cAMP and forskolin did not blunt the decrease in transepithelial resistance caused by iono and PDBU. Iono and PDBU increase MDCK monolayer permeability independently of an increase in MLC phosphorylation.

The effect of histamine and cyclic adenosine monophosphate on myosin light chain phosphorylation in human umbilical vein endothelial cells.

Histamine causes adjacent endothelial cells to retract from each another. We examined phosphorylation of the 20-kD myosin light chain (MLC20) in human umbilical vein endothelial cells (HUVECs) exposed to histamine to determine if we could find evidence to support the hypothesis that retraction of these cells in response to histamine represents an actomyosin-initiated contraction of the endothelial cytoskeleton. We found that MLC20 in HUVECs was constitutively phosphorylated with approximately 0.2 mol phosphate/mol MLC20. Histamine increased MLC20 phosphorylation by 0.18 +/- 0.05 mol phosphate/mol MLC20. This peak increase in phosphorylation occurred 30 s after initiating histamine exposure, persisted through 90s, and returned to control levels by 5 min. Agents that increase HUVEC cAMP prevent cell retraction in response to histamine. An increase in HUVEC cAMP decreased MLC20 phosphorylation by 0.18 +/- 0.02 mol phosphate/mol MLC20 and prevented the increase in MLC20 phosphorylation after exposure to histamine. Tryptic peptide maps of phosphorylated myosin light chain indicated that myosin light chain kinase phosphorylated MLC20 in HUVECs under basal, cAMP-, and histamine-stimulated conditions. Phosphoaminoacid analysis of the monophosphorylated peptide indicated that, in contrast to smooth muscle cells, ser19 and thr18 monophosphorylation occurs in HUVECs. On the basis of our results, modulation of myosin light chain kinase activity may be an important regulatory step in the control of endothelial barrier function.

Histamine, actin-gelsolin binding, and polyphosphoinositides in human umbilical vein endothelial cells.

Histamine activates inositol phospholipid metabolism, increases calcium, and causes a change in shape of human umbilical vein endothelial (HUVE) cells. Changes in endothelial cell shape are determined, in part, by changes in the actin cytoskeleton. Gelsolin is an actin-binding protein with the potential to alter the actin cytoskeleton in response to changes in cell calcium and/or changes in polyphosphoinositides. Therefore, we examined the interactions of actin and gelsolin in HUVE cells in which inositol phospholipid metabolism was activated with histamine. In HUVE cells exposed to histamine we estimated actin-gelsolin binding by quantitating actin and gelsolin, immunoprecipitated with anti-gelsolin Sepharose. We estimated the relative amount of filamentous actin in the histamine-exposed HUVE cells by quantitating the amount of actin that was Triton soluble. We also measured the amount of phosphatidylinositol 4-phosphate (PIP) and phosphatidylinositol 4,5-bisphosphate (PIP2) in the HUVE cells before and after exposure to histamine. We found that histamine decreased the amount of actin that was immunoprecipitated with gelsolin, decreased the fraction of cell actin that was Triton soluble, and increased PIP and PIP2. These results demonstrate that histamine promotes actin filament formation in HUVE cells and that histamine-mediated changes in actin-gelsolin binding in these cells are better predicted by changes in polyphosphoinositides than by increases in cell calcium.

Histamine and inositol phosphate accumulation in endothelium: cAMP and a G protein.

Histamine increases microvascular permeability through a calcium-dependent process, and histamine occupancy of the H1-receptor increases calcium in cultured endothelial cells. Agents that increase adenosine 3',5'-cyclic monophosphate (cAMP) in endothelial cells prevent the in vivo increase in microvascular permeability that follows histamine exposure. In the current experiments, histamine occupancy of the H1-receptor increased the flux of albumin across monolayers of cultured human umbilical vein endothelial cells (HUVEC). This was prevented by pretreating the cells with theophylline, forskolin, and 8-bromo-cAMP (BrcAMP), which also decreased the flux of albumin across control monolayers. Exposing the cells to histamine increased inositol phosphate accumulation in the cells, and this was prevented by the H1-antagonist pyrilamine but not by theophylline, forskolin, and BrcAMP. Exposing the cells to histamine increased intracellular calcium measured with fura-2. The increase in cell calcium was prevented by pyrilamine but not by pretreatment with theophylline, forskolin, and BrcAMP. When endogenous cell GTP was depleted by permeabilizing the membranes of the endothelial cells with Staphylococcus aureus alpha-toxin, histamine-stimulated inositol phosphate accumulation was enhanced with addition of GTP but not with addition of GDP to the buffer. Addition of GTP alone to the buffer did not increase inositol phosphate accumulation in alpha-toxin-treated cells. Histamine stimulates inositol phosphate accumulation in HUVEC via a G protein. Inhibition of the edemagenic effects of histamine by cAMP does not occur by interrupting this signal transduction pathway between the binding of histamine to its receptor and the increase in intracellular calcium.

Oxidants and conductance of cultured epithelial cell monolayers: inositol phospholipid hydrolysis.

Reactive oxidants contribute to the alterations in endothelial and epithelial permeability that characterize the inflammatory response. We previously noted that noncytolytic doses of oxidants reversibly decreased the electrical resistance across cultured monolayers of Madin-Darby canine kidney (MDCK) cells (J. Clin. Invest. 76: 1155-1168, 1985). In this investigation, we have found that similar doses of oxidants initiate inositol phospholipid hydrolysis by a phospholipase C in cultured MDCK cells, with resultant increases in inositol polyphosphates, phosphatidic acid, and 1,2 diglycerides. Activation of this pathway is linked to activation of protein kinase C in many cells. The addition of phorbol 12,13-dibutyrate (PDBU) and 1-oleoyl-2-acetyl-sn-glycerol, activators of protein kinase C, decreased the electrical resistance across MDCK monolayers cultured on micropore filters similar to the effects of hydrogen peroxide. In contrast, the addition of 4 alpha-phorbol 12,13-didecanoate, a chemically similar compound that does not activate protein kinase C, did not decrease the electrical resistance. When MDCK monolayers were exposed to PDBU, fixed, and stained with rhodamine phallicidin, the peripheral band of actin in the cells showed a loss of staining density and continuity similar to the changes in phallicidin staining we previously noted in cells exposed to hydrogen peroxide. These data are consistent with the hypothesis that some of the reversible effects of oxidants on epithelial barriers are mediated through phospholipase C hydrolysis of inositol phospholipids with consequent activation of protein kinase C.

Exogenous oxidants initiate hydrolysis of endothelial cell inositol phospholipids.

Oxidants released from inflammatory cells contribute to the pathogenesis of acute inflammatory edema in many models. Chemically produced oxidants can reversibly alter the barrier properties of cultured endothelial and epithelial monolayers. This report examines the effects of nonlytic doses of H2O2 on endothelial cell lipids. H2O2 oxidized omega-6 fatty acids in the endothelial cells and initiated hydrolysis of endothelial cell phospholipids. When endothelial cells were exposed to peroxidized linoleic acid, it caused lysis of the cells at doses 1,000-fold lower than effective doses of H2O2. The phospholipid hydrolysis was directed primarily at the inositol phospholipids and consisted of both A and C type phospholipase activity. The phospholipase A hydrolysis resulted in increases in endothelial cell free fatty acids and lysophosphatidylinositol. The phospholipase C hydrolysis resulted in increases in diglycerides, phosphatidic acid, and inositol polyphosphate levels. The phospholipase C hydrolysis of phosphatidylinositol is known to activate protein kinase C in most cells. Stimulation of protein kinase C with phorbol-12,13-dibutyrate increased albumin flux across endothelial monolayers and altered endothelial cell shape, similar to effects of oxidants. These data are consistent with the hypothesis that oxidant-initiated hydrolysis of endothelial cell inositol phospholipids contributes to oxidant-mediated reversible changes in endothelial monolayer barrier function.

Effects of calcium on transendothelial albumin transfer and electrical resistance.

Nicolaysen, and more recently Kern and Malik, reported that chelation of calcium increased microvascular hydraulic conductivity and albumin permeability in isolated perfused lungs. To begin to understand how calcium affects endothelial function we examined the effect of calcium chelation on an in vitro endothelium. Chelation of calcium with ethyleneglycol-bis(beta-aminoethylether)-N,N'-tetraacetic acid increased the rate of transendothelial albumin transfer by 125%. Reincubation of the endothelium in calcium-repleted medium restored the rate of transfer to its original value. Chelation of extracellular calcium abolished transendothelial electrical resistance. The transendothelial electrical resistance was also restored to normal by reincubation of the endothelium in calcium-repleted medium. Chelation of extracellular calcium caused adjacent endothelial cells to retract from one another, and normal apposition of adjacent cells was restored after reincubation in calcium-repleted medium. Chelation of extracellular calcium produced a centripetal retraction of the peripheral band of actin in individual endothelial cells, and the actin band resumed its normal location after reincubation in calcium-repleted medium. Calcium is an important determinant of endothelial integrity and alterations in calcium produce dynamic changes in endothelial barrier properties and in endothelial-cell shape.

Active transendothelial transport of albumin. Interstitium to lumen.

Cultured porcine pulmonary artery endothelial monolayers actively transport albumin from interstitium to lumen. The active process favors interstitial to luminal transport by a factor of 10 even in the face of a 2:1 luminal:interstitial albumin gradient. The active process is abolished by treatment with 1 mM sodium cyanide. Dextran is not actively transported. This process could be important in determining transvascular fluid balance and transfer of macromolecules across the endothelium.

Polymorphonuclear leukocyte: arachidonate edema.

Polymorphonuclear leukocytes (PMN) are important participants in many models of acute lung edema. Enhanced metabolism of arachidonate is also characteristic of many of these models. We found that PMN and arachidonate, but neither alone, increased alveolar capillary permeability of isolated perfused lungs and increased transfer of albumin across monolayers of endothelial cells cultured on micropore filters. Inhibition of PMN, but not endothelial cyclooxygenase, blunted the edematous process. Neither PMN proteases nor PMN-derived oxidants were involved. The edemagenic activity was not found in supernatants of PMN and arachidonate, and unstable prostaglandins did not alter endothelial albumin transfer. The edemagenic process was not inhibited by blocking leukotriene synthesis, and endothelial albumin transfer was not increased by direct addition of leukotrienes to endothelium. These data demonstrate that PMN and arachidonate can interact to increase endothelial permeability and that PMN cyclooxygenase activity is important for this process. This interaction is of potential significance to the acute inflammatory process in the lung vasculature.

Reversible oxidant-induced increases in albumin transfer across cultured endothelium: alterations in cell shape and calcium homeostasis.

To determine whether reactive oxygen molecules could directly and reversibly increase the transfer of albumin across an endothelial barrier, we measured albumin transfer across monolayers of endothelium cultured on micropore filters before and after exposure to xanthine and xanthine oxidase. Xanthine and xanthine oxidase increased endothelial albumin transfer in a dose-dependent fashion. Parallel phase contrast and fluorescence microscopy demonstrated retraction of adjacent cells from one another and disruption of the actin filaments. The oxidant-induced increases in albumin transfer and changes in cell shape were reversed by removing xanthine oxidase and then incubating the monolayers for 3 1/2 hours in tissue culture media enriched with fetal bovine serum. However, incubation in tissue culture media without serum resulted in progressive injury and cell death. Hence, the brief exposure to oxidants initiated a progressive injury process that was reversed by incubation in serum. Because intracellular and extracellular calcium are important determinants of cell shape, and because some oxidized membrane lipids act as calcium ionophores, we asked whether oxidants altered endothelial calcium homeostasis. Xanthine-xanthine oxidase increased release of 45Ca++ from preloaded cells. The calcium antagonist lanthanum chloride prevented xanthine-xanthine oxidase increases in endothelial albumin transfer and prevented the changes in cell shape; chelation of extracellular calcium inhibited lysis of endothelium by xanthine-xanthine oxidase; and the calcium ionophore A23187 increased endothelial albumin transfer and mimicked the oxidant-induced changes in cell shape. Lanthanum chloride inhibited these effects of A23187. These data suggest that oxygen radicals can reversibly increase endothelial permeability to macromolecules, that this is associated with reversible changes in endothelial cell shape and actin filaments, and that the changes in cell shape are related to oxidant-induced changes in endothelial calcium homeostasis.

Granulocytes and phorbol myristate acetate increase permeability to albumin of cultured endothelial monolayers and isolated perfused lungs. Role of oxygen radicals and granulocyte adherence.

Human granulocytes and phorbol myristate acetate (PMA) increased permeability to albumin of monolayers of cultured endothelial cells grown on micropore filters. Granulocytes from a patient with chronic granulomatous disease and PMA did not increase endothelial permeability to albumin, demonstrating that the increase in permeability is dependent on granulocyte-derived oxygen radicals. When granulocytes were separated from the endothelial cells by a micropore filter, granulocytes and PMA no longer increased endothelial permeability to albumin, demonstrating that PMA-stimulated granulocytes must be closely approximated to endothelial cells to increase endothelial permeability. The relevance of these in vitro findings to an intact microvasculature was confirmed by demonstrating that agents that reduce granulocyte adherence to endothelium reduce edema formed in isolated lungs by granulocytes and PMA, an oxygen radical dependent process. Pretreatment of granulocytes with cytochalasin B or addition of 2% dextran to isolated lung perfusates reduced granulocyte adherence and markedly reduced edema formation in isolated lungs. These studies demonstrate that PMA-stimulated granulocytes must be closely apposed to endothelial cells to increase endothelial permeability through an oxygen-radical-dependent mechanism, and they suggest that reduction of granulocyte adherence may protect against granulocyte-dependent edema.

Role of endothelial cell cytoskeleton in control of endothelial permeability.

Increased permeability of the pulmonary microvasculature is felt to cause acute noncardiogenic lung edema, and histological studies of edematous lungs show gaps between apparently healthy endothelial cells. To determine whether alterations in endothelial cell cytoskeletons would alter endothelial permeability, we exposed monolayers of pulmonary artery endothelial cells grown on micropore filters to cytochalasin B or D. Cytochalasin exposed monolayers demonstrated a 2- to 3-fold increase in endothelial permeability that was readily reversible by washing the monolayers free of cytochalasins. Parallel phase contrast and fluorescence microscopy demonstrated retraction of cell cytoplasm and disruption of bundles of microfilaments in cytochalasin exposed cells. These changes also were readily reversed after washing the cells free from cytochalasins. To test the relevance of these findings to an in situ microvasculature, we added cytochalasin B to the perfusate of isolated rabbit lungs and observed that cytochalasin B caused a high permeability lung edema. These studies suggest that endothelial cell cytoskeletons may be important determinants of endothelial permeability.

Granulocytes mediate acute edematous lung injury in rabbits and in isolated rabbit lungs perfused with phorbol myristate acetate: role of oxygen radicals.

Acute edematous lung injury is associated with a marked increase in the number of granulocytes in the alveoli and microvasculature of the lung. Phorbol myristate acetate (PMA) causes granulocytes to adhere, aggregate, and release oxygen radicals and granular enzymes. We found that intravenously injected PMA caused a protein-rich edema in lungs of control rabbits but not in granulocytopenic rabbits pretreated with nitrogen mustard. Specifically, control rabbits treated with PMA had higher lung weight to body weight ratios (6.4 +/- 1.0 X 10(-3)) and lung lavage albumin concentrations (190 +/- 44 mg/dl) than granulocytopenic rabbits pretreated with nitrogen mustard and then given PMA (4.74 +/- 0.23 X 10(-3) and 9.9 +/- 3.8 mg/dl, respectively). To further clarify the role of granulocytes in the production of edema, additional experiments were conducted in an isolated perfused rabbit lung. Addition of purified granulocytes and PMA to the balanced salt perfusate caused lung edema, whereas neither granulocytes nor PMA alone caused edema. Specifically, increases in lung weights (42 +/- 9.2 g) and albumin concentrations (1,182 +/- mg/dl) in lung lavages from isolated lungs exposed to granulocytes and PMA were greater than increases in lung weights or albumin concentrations in lung lavages from isolated lungs exposed to granulocytes alone (2.0 +/- 0.4 g and 15 +/- 0.6 mg/dl), or to PMA alone (6.0 +/- 0.6 g and 81 +/- 34 mg/dl). To determine the contribution of oxygen radicals to the pathogenesis of the edema, chronic granulomatous disease granulocytes, which are deficient in oxygen radical production, were added to the isolated lung perfusate. Chronic granulomatous disease granulocytes and PMA did not cause edema in isolated lungs (delta lung weight 1.0 +/- 0.2 g and lavage albumin 12 +/- 5.0 mg/dl) whereas granulocytes from normal human subjects and PMA did (delta lung weight 43 +/- 5.2 g and lavage albumin 1,120 +/- 54 mg/dl). These data suggest that oxygen radicals released from stimulated granulocytes contribute to the pathogenesis of acute edematous lung injury.

Angiotensin converting enzyme concentrations in the lung lavage of normal rabbits and rabbits treated with nitrogen mustard exposed to hyperoxia.

Increased concentrations of angiotensin converting enzyme (ACE) were found in lung lavages from rabbits exposed to hyperoxia for 72 h and the concentrations of ACE were correlated with ratios of extravascular lung water to body weight (r = 0.69, p less than 0.05) and albumin concentrations in lung lavages (r = 0.89, p less than 0.01). In parallel studies, rabbits treated with nitrogen mustard in which granulocytopenia was maintained throughout the 72-h hyperoxic exposure period had less evidence of edematous lung injury and lower concentrations of ACE in their lung lavages than similarly treated rabbits in which granulocytopenia was not maintained. The results suggested that granulocytes contribute to acute edematous lung injury from hyperoxia and that ACE concentrations in lung lavages reflect this process.