Measurement of albumin permeability across endothelial monolayers in vitro

Cytokine-induced increases in endothelial permeability occur after adhesion molecule expression

BACKGROUND

Transmigration of neutrophils (PMNs) through endothelial cell tight junctions is a critical stage in the tissue injury of ischemia-reperfusion (I/R). Although cytokines are released in I/R, it is unclear whether cytokines directly increase permeability or this phenomenon requires both expression of cell adhesion molecules and PMN adhesion-activation.

METHODS

We exposed confluent monolayers of human umbilical vein endothelial cells to physiologic concentrations of interleukin-1 (10 pg/ml) and tumor necrosis factor-alpha (10 pg/ml) in the absence of PMNs. Tight junction permeability was quantified with both transendothelial electrical resistance and albumin flux, whereas expression of endothelial-leukocyte adhesion molecule-1 was measured by flow cytometry (t test p < 0.05).

RESULTS

Stimulation with tumor necrosis factor-alpha or interleukin-1 produced maximal transendothelial electrical resistance decreases at 12 hours with return to baseline at 24 hours. Increases in albumin flux began at 6 hours, with maximum effects at 24 hours. These changes occurred soon after maximal expression of endothelial-leukocyte adhesion molecule-1 at 4 hours.

CONCLUSIONS

Cytokines induced increases in both cell adhesion molecule expression and endothelial permeability. This sequence of events is consistent with direct cytokine effects on cytoarchitecture, because it occurred without the adhesion-activation of PMNs.

Protein kinase C beta 1 overexpression augments phorbol ester-induced increase in endothelial permeability

We studied the postulated involvement of the protein kinase C beta 1 (PKC beta 1) isoform in the regulation of endothelial permeability using human dermal microvascular endothelial cell line (HMEC-1). We overexpressed the recombinant PKC beta 1 gene via retroviral-mediated transduction in these cells. PKC beta 1 gene transfer was stable, and PKC beta 1 protein production was persistent for at least 1 month posttransduction. Addition of 2 x 10(-9) M and 2 x 10(-8) M phorbol 12-myristate 13-acetate (PMA) to the control (nontransduced) HMEC-1 cells increased the transendothelial 125I-albumin clearance rate (an index of endothelial permeability) from 2.5 +/- 0.2 x 10(-2) microliters/min to 5.4 +/- 1.2 x 10(-2) microliters/min and 16.8 +/- 3.1 x 10(-2) microliters/min, respectively. However, addition of 2 x 10(-9) M PMA to PKC beta 1-overexpressing HMEC-1 cells produced a maximal increase in the transendothelial 125I-albumin clearance rate of 15.9 +/- 2.0 x 10(-2) microliters/min. Challenge of these cells with 2 x 10(-8) M PMA did not further augment the increase in permeability. Activation with PMA was associated with the translocation of the PKC beta 1 from the cytosol to the membrane. These data show that PKC beta 1 overexpression augments the increase in endothelial permeability in response to PKC activation, suggesting an important function for the PKC beta 1 isoform in the regulation of endothelial barrier.

Soluble CD14 in serum mediates LPS-induced increase in permeability of bovine pulmonary arterial endothelial cell monolayers in vitro

Lipopolysaccharide (LPS) is a mediator of septic shock and acute respiratory distress syndrome (ARDS), conditions which are characterized by high-permeability pulmonary edema. LPS increases endothelial permeability both directly and indirectly via the pro-inflammatory cytokines produced by monocytes and macrophages. We investigated the role of soluble CD14 in serum in the increased endothelial permeability induced by LPS. Bovine pulmonary artery endothelial cells were grown to confluence on a microporous filter and the 125I-albumin clearance rate across the monolayer was determined. Even a high concentration of LPS (1 microgram/ml) did not increase endothelial permeability under a serum-free condition. In the presence of more than 3% normal human serum, LPS increased endothelial permeability. The presence of neutralizing anti-CD14 monoclonal antibody eliminated the serum-dependent effect of LPS. The addition of recombinant sCD14 completely replaced the requirement for serum. LPS-binding protein (LBP) did not enhance the rsCD14-mediated LPS effect, and anti-LBP antibody did not attenuate the serum-dependent LPS effect. These findings suggest that sCD14 in serum mediates the permeability-increasing effect on LPS on endothelial cells but that LBP is not necessary for this effect.

Beta 2-adrenoceptors mediate a reduction in endothelial permeability in vitro

The permeability of bovine pulmonary artery endothelial (CPAE) monolayers to Evans blue-labelled albumin (Evans blue-albumin) has been measured in vitro. Thrombin caused a concentration-dependent increase in Evans blue-albumin clearance across endothelial monolayers. Isoprenaline inhibited thrombin-induced Evans blue-albumin clearance in a concentration-dependent manner (EC50 21 nM). This effect was mimicked by the selective beta 2-adrenoceptor agonists salbutamol (EC50 64 nM) and salmeterol (EC50 2.7 nM), but not by the selective beta 1-adrenoceptor agonist, RO-363 ((1-[3',4'-dihydroxyphenoxy]-2-hydroxy-[3",4"- dimethoxyphenethylamino]-propane)oxalate), nor by the selective beta 3-adrenoceptor agonist, CL-316,243 (disodium (R,R)-5-[2-[[2-(3-chlorophenyl)-2-hydroxyethyl]-amino]propyl]-1,3- benzodioxole-2,2-dicarboxylate). Isoprenaline, salbutamol and salmeterol, but not RO-363 or CL-316,243 produced small, but significant reductions in Evans blue-albumin clearance across unstimulated endothelial monolayers. Inhibition of the response to thrombin by isoprenaline was antagonised by the selective beta 2-adrenoceptor antagonist, ICI-118,551 ((erythro-DL-1(7-methylindan-4- yloxy)3-isopropylaminobutan-2-ol), pKB 8.4). Salmeterol also inhibited hydrogen peroxide-stimulated Evans blue-albumin clearance. Hence, the widely used beta 2-adrenoceptor agonists, salbutamol and salmeterol, are able to reduce endothelial permeability at nanomolar concentrations.

Heparin induces endothelial extracellular matrix alterations and barrier dysfunction

We investigated the influence of heparin on the composition of the subendothelial matrix and on endothelial permeability to elucidate the structure-function relationship of matrix composition and permeability. Albumin flux across the confluent endothelial monolayers was used to assess the macromolecular permeability. Lowest values were obtained when 100% serum was used as medium for permeability studies. The endothelial matrix components, fibronectin and basement membrane-associated heparan sulfate proteoglycan (HSPG), were measured by enzyme immunoassay. Treatment of proliferating endothelial cells with heparin (0-900 micrograms/ml) induced a dose-dependent decrease in endothelial HSPG content, whereas the fibronectin content was unaltered. This structural change was accompanied by an increase in albumin permeability. Both heparin effects exhibited similar dose-response curves with half-maximal effects at approximately 5 micrograms/ml heparin. Acute addition of 300 micrograms/ml heparin had no effect on permeability or HSPG content. When endothelial cells were preincubated with an HSPG antiserum, the endothelial permeability increased nearly threefold. Our results indicate that heparin-induced loss of HSPG may cause the increase in endothelial permeability. The data underline the importance of HSPG for the integrity of the endothelial barrier.

Carrier-mediated transport of H1-antagonist at the blood-brain barrier: mepyramine uptake into bovine brain capillary endothelial cells in primary monolayer cultures

The transport mechanism of the H1-antagonist mepyramine at the blood-brain barrier (BBB) was studied by using primary cultured monolayers of bovine brain capillary endothelial cells (BCEC). The initial uptake of [3H]mepyramine into the BCEC showed strong temperature and concentration dependency, indicating that it involves both saturable and nonsaturable processes. Transport at the luminal membrane may be the rate-limiting process in the transcellular transport, since the values of the uptake coefficient of [3H]mepyramine at the luminal membrane (609 microliters/mg protein/min) and the transcellular permeability coefficient (488 microliters/mg protein/min) are very similar. The initial uptake of [3H]mepyramine was not affected by metabolic inhibitors, but was stimulated by preloading with the drug. Mepyramine appears to be transported into the BCEC by a carrier-mediated transport system which does not require metabolic energy, probably via a facilitated diffusion mechanism.

Atherogenic levels of low density lipoprotein alter the permeability and composition of the endothelial barrier

In the present study we investigated the influence of elevated low density lipoprotein (LDL) concentration on endothelial permeability. Endothelial cells were cultured on microporous membranes until confluence and albumin, dextran and LDL transfer across endothelial monolayers was determined to assess macromolecular permeability. Exposure of proliferating aortic endothelial cells to LDL levels of more than 1 mg/ml LDL-cholesterol induced a concentration-dependent exponential increase in the permeability of confluent endothelial monolayers. Acute addition of high LDL concentration did not alter macromolecular permeability. Once elevated permeability was induced, it persisted. It was not readily reversible after addition of normal LDL levels. Change in permeability was accompanied by a selective decrease in basement membrane associated heparan sulfate proteoglycan (HSPG) content. The apparent parallel between the loss in endothelial barrier function and HSPG decrease implicates a connection between the two events. Prolonged, but not acute, incubation with antiserum directed against the core-protein of HSPG also led to increased permeability, suggesting a causal role of HSPG for the proper function of endothelium. The fact that non-atherogenic LDL-cholesterol levels had no effect indicates that a 'threshold' concentration for LDL-cholesterol may exist, leading to nondenuding injury in the endothelial barrier as an early event in development of atherosclerosis.

In vitro and in vivo transport of zidovudine (AZT) across the blood-brain barrier and the effect of transport inhibitors

The transport of the antiviral nucleoside analogue zidovudine (3'-azido-3'-deoxythymidine; AZT) into the central nervous system (CNS) was characterized in vitro and in vivo. The in vitro model consisted of primary cultures of isolated bovine capillary endothelial cells. The transport rate of AZT across the monolayer, expressed as endothelial permeability P, was determined following luminal and abluminal administration. P did not differ between the two administration sites (luminal, 1.65 +/- 0.44 cm/min/10(3); abluminal, 1.63 +/- 0.28 cm/min/10(3)). The transport of AZT across the endothelial cell monolayer was found to be concentration independent in the range between 0.4 and 50 micrograms/mL. AZT transport was not affected by pretreatment of the cells with either metabolic inhibitors (DODG and DODG/NaN3) or probenecid. This suggests that AZT passes the monolayer mainly by passive diffusion. The in vivo transport of AZT across the blood-brain barrier and the blood-CSF barrier was studied in male Wistar rats after coadministration of potential inhibitors of active transport of AZT: probenecid (organic anion transport) and thymidine (nucleoside transport). Intracerebroventricular and intravenous coadministration of probenecid caused a significant (P < 0.001) increase in the CSF/plasma concentration ratio compared to the control phase, indicating that the organic anion carrier is involved in AZT transport from CSF to blood. Since there was no effect of probenecid on the transport of AZT in vitro, it is suggested that this carrier is located at the choroid plexus. Coadministration of thymidine did not affect the CSF/plasma concentration ratio, suggesting that a nucleoside carrier system is not involved in AZT transport into or out of the CNS.

In vivo and in vitro blood-brain barrier transport of 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase inhibitors

Among the HMG-CoA reductase inhibitors, lovastatin and simvastatin have central nervous system (CNS) side effects, such as sleep disturbance, whereas pravastatin does not. This difference in CNS side effects may be due to a difference in blood-brain barrier (BBB) permeability among these inhibitors. To test this hypothesis, we compared the BBB transport ability of HMG-CoA reductase inhibitors by using an in vivo brain perfusion technique in rats and an in vitro culture system of bovine brain capillary endothelial cells. The in vivo BBB permeability coefficients of the lipophilic inhibitors, [14C]lovastatin and [14C]simvastatin, were high. In contrast, that of the hydrophilic inhibitor, [14C]pravastatin, was low and not significantly different from that of [14C]sucrose, an extracellular space marker. Similarly, the in vitro BBB permeability coefficients of [14C]lovastatin and [1C]simvastatin were high, while that of [14C]-pravastatin was low. The in vivo and in vitro transcellular permeabilities obtained for HMG-CoA reductase inhibitors were comparable. This study shows that the BBB permeability correlates with the CNS side effects of the HMG-CoA reductase inhibitors.

Neutrophil adhesion to endothelial cells impairs the effects of catalase and glutathione in preventing endothelial injury

We studied the effects of the CuZn superoxide dismutase (SOD), catalase (CAT), and glutathione (GSH) on endothelial permeability to 125I-albumin after activation of neutrophils (PMN) with phorbol 12-myristate-13-acetate (PMA; 10(-8) M). PMN were either in direct contact with the endothelial cell monolayer grown on a porous gelatin-coated microporous 10-microns-thick polycarbonate filter (upright system) or separated from the endothelium by a similar filter (inverted system). Transendothelial 125I-albumin clearance rates were measured as an index of endothelial permeability. In the absence of antioxidants, activation of PMN increased transendothelial 125I-albumin clearance rates in both systems from 0.041 +/- 0.006 microliters/min (baseline) to 0.262 +/- 0.18 microliters/min (upright system) and from 0.063 +/- 0.02 microliters/min to 0.244 +/- 0.06 microliters/min (inverted system). PMA induced 80-90% of PMN to adhere to either gelatin-coated filters or to endothelial cells, from the basal PMN adhesion value of 5.3 +/- 2.2% and 4.3 +/- 1.1%, respectively. SOD, which dismutases superoxide anion to hydrogen peroxide (H2O2), did not alter the transendothelial 125I-albumin clearance rates in either system at any concentration from 10-300 U/ml. CAT (100-1,000 U/ml) and GSH (0.5-10 mM), which remove the H2O2 generated during PMN activation, did not alter the increase in transendothelial 125I-clearance rates after PMN activation in the upright system, but both agents prevented the increase in transendothelial 125I-clearance rates in the inverted system. We conclude that PMN activation with PMA causes endothelial injury irrespective of PMN contact to the endothelial monolayer. Moreover, H2O2, a release product of PMN activation, is a critical mediator of PMN-dependent endothelial injury. Finally, the results indicate that CAT and GSH prevent endothelial injury only in the absence of direct PMN contact with endothelial cells, suggesting that antioxidants such as GSH and CAT are excluded from sites of PMN-endothelial contact and thus are ineffective antioxidants.

Thrombin receptor peptide inhibits thrombin-induced increase in endothelial permeability by receptor desensitization

Thrombin, a potent activator of cellular responses, proteolytically cleaves, and thereby activates its receptor. In the present study, we compared the effects of the thrombin receptor 14-amino acid peptide (TRP-14; SFLLRNPNDKYEPF), which comprises the NH2 terminus after cleavage of the thrombin receptor, and of the native alpha-thrombin on endothelial monolayer permeability. Addition of TRP-14 (1-200 microM) to bovine pulmonary artery endothelial cells increased [Ca2+]i in a dose-dependent manner. The peak increase in [Ca2+]i in response to 100 microM TRP-14 or 0.1 microM alpha-thrombin was similar (i.e., 931 +/- 74 nM and 1032 +/- 80 nM, respectively), which was followed by a slow decrease with t1/2 values of 0.73 and 0.61 min, respectively. Extracellular Ca2+ chelation with 5 mM EGTA abolished the sustained increases in [Ca2+]i induced by either TRP-14 or alpha-thrombin. alpha-thrombin (0.1 microM) increased transendothelial [125I]albumin permeability, whereas TRP-14 (1-100 microM) had no effect. Coincubation of 100 microM TRP-14 with 1 microM DIP-alpha-thrombin also did not increase permeability over control values. Stimulation of BPAEC with 0.1 microM alpha-thrombin induced translocation of protein kinase C (PKC) from the cytosol to the plasma membrane indicative of PKC activation, whereas TRP-14 had no effect at any concentration. TRP-14 at 100 microM desensitized BPAEC to thrombin-induced increases in [Ca2+]i and transendothelial permeability. The Ca2+ desensitization was reversed after approximately 60 min, and this recovery paralleled the recovery of the permeability response. These findings indicate that the TRP-14-induced Ca2+ mobilization in the absence of PKC activation is insufficient to increase endothelial permeability. In contrast, the increase in endothelial permeability after alpha-thrombin occurred in conjunction with Ca2+ mobilization as well as PKC activation. TRP-14 pretreatment prevented the alpha-thrombin-induced increase in endothelial permeability secondary to desensitization of the Ca2+ signal. The results suggest that combined cytosolic Ca2+ mobilization mediated by TRP-14 and PKC activation mediated by a TRP-14-independent pathway are dual signals responsible for the thrombin-induced increase in vascular endothelial permeability.

Culture and characterization of pulmonary microvascular endothelial cells

Surface proteins were compared in endothelial cells (EC) obtained from bovine peripheral lung, pulmonary artery and vein, and dorsal aorta using sodium dodecyl sulfate (SDS) polyacrylamide gel electrophoresis. Galactose-containing glycoproteins [molecular weight (M(r)) 160-220 and 40 kDa] binding to the Ricinus communis agglutinin (RCA) and peanut agglutinin (PNA) were selectively observed on pulmonary microvessel EC as compared to EC from pulmonary artery, pulmonary vein, and dorsal aorta. The unique RCA- and PNA-binding profiles of EC from the pulmonary artery and microvessels may be important in characterizing EC from different sites in the pulmonary circulation. The pulmonary microvessel EC monolayer was also 15-fold more restrictive to transendothelial flux of [14C]sucrose (M(r) = 342 Da) than the pulmonary artery EC monolayer. In contrast, the microvessel EC were only six- and twofold more restrictive to the flux of larger tracer molecules, ovalbumin (M(r) 43 kDa) and albumin (M(r) = 69 kDa) than pulmonary artery EC. The greater restrictiveness of pulmonary microvessel EC monolayer indicates a major phenotypic difference in the cultured pulmonary microvessel EC barrier function.

Interleukin-2 directly increases albumin permeability of bovine and human vascular endothelium in vitro

The direct effects of interleukin-2 (IL-2) on albumin permeability of cultured bovine pulmonary artery endothelial cell (BPAEC) and human arterial endothelial cell (HAEC) monolayers were studied. BPAEC were exposed to IL-2 (500 to 25,000 U/ml) for 4 h. The steady-state transfer rate of [125I]albumin across the BPAEC monolayer was 3.3 +/- 0.4%/h (n = 10) in control BPAEC (diluent alone), was significantly increased in BPAEC exposed to 500 U/ml of IL-2 (72 +/- 3% above control values, n = 6, P less than 0.02), and further increased in BPAEC exposed to 5,000 U/ml (60 +/- 2% increase above 500 U/ml values, n = 5, P less than 0.02). No further increase was noted after exposure to 25,000 U/ml of IL-2. Additionally, no further increase in [125I]albumin transfer rates was noted in BPAEC exposed to 5,000 U/ml of IL-2 for 24 versus 4 h. Similar changes were found using HAEC. Preincubation of HAEC with an anti-IL-2 low-affinity receptor antibody (anti-IL-2R alpha) inhibited the IL-2-induced permeability increase. Expression of IL-2R alpha receptors in HAEC incubated with 5,000 U/ml of IL-2 for 4 h was also found. Thus, IL-2 appears to have a direct effect on cultural arterial endothelial monolayers not requiring the presence of other cell types or serum proteins. IL-2-induced increases in endothelial macromolecular permeability may play an important role in the pathogenesis of the IL-2-induced vascular leak syndrome seen in vivo.

Reoxygenation of endothelial cells increases permeability by oxidant-dependent mechanisms

We investigated the effects of hypoxia/reoxygenation exposure on the barrier function of endothelial cell monolayers. Bovine pulmonary microvessel endothelial cells were grown to confluence on microporous filters (0.8-microns pore diameter) and exposed to hypoxia (0.1% O2 or PO2 approximately 1 mm Hg) for 2, 4, 12, or 24 hours, followed by reoxygenation with room air for a period ranging from 16 seconds to 2 hours. The transendothelial clearance rate of 125I-albumin was measured to determine the permeability of endothelial monolayers. Permeability increased twofold or fivefold over control values after 1 hour of reoxygenation in monolayers that had been exposed to either 12 or 24 hours of hypoxia. The response occurred within 5 minutes of reoxygenation, increased maximally by 40 minutes, and remained elevated with continuous reoxygenation for up to 2 hours. The increase in permeability was associated with F-actin reorganization, a change to spindlelike cells, and injured mitochondria. Immunoblot analysis indicated that neither hypoxia alone nor reoxygenation changed CuZn superoxide dismutase (SOD), MnSOD, and catalase levels. However, release of superoxide anions (O2-) into the extracellular medium increased by twofold within 40-60 minutes of reoxygenation. Treatment of endothelial cells with CuZnSOD (100 units/ml) for the 24-hour hypoxia period prevented O2- generation and approximately 50% of the increase in permeability. Higher CuZnSOD concentrations (greater than or equal to 200 units/ml) were not protective. Treatment with catalase (100-1,000 units/ml) inhibited the reoxygenation-induced increase in permeability at the highest catalase concentration (1,000 units/ml), suggesting a critical role of hydrogen peroxide in mediating the response.(ABSTRACT TRUNCATED AT 250 WORDS)

Fibrin contact increases endothelial permeability to albumin

We studied the effects of contact of bovine pulmonary artery endothelial cell monolayers with fibrin on the endothelial barrier function. Fibrin formed by clotting purified fibrinogen (0.5 to 3.0 mg/ml) with alpha-thrombin (1 U/ml) was added to endothelial monolayers and permeability measurements were made after fibrin removal. Fibrin incubation for 3 hours resulted in 2- to 5-fold increases in transendothelial 125I-albumin permeability. Permeability returned to baseline value within 3 hours after fibrin removal. Direct contact with fibrin was necessary for the response, since fibrin separated from the endothelium did not increase permeability. Contact with agarose (2 mg/ml) or fibrinogen (0.5 to 3.0 mg/ml) also did not increase endothelial permeability. Transmission electron microscopic examination indicated normal appearance of interendothelial junctions at a time when albumin permeability was increased and no overt evidence of endothelial injury. Incubation of fibrin with endothelial monolayers at 4 degrees C prevented the increase in albumin permeability. We examined the possibility that increased albumin transcytosis was responsible for fibrin's effect using 14C-sucrose (Mr = 342D), a lipid insoluble tracer. Fibrin increased sucrose flux by 1.5-fold compared to 2- to 5-fold increases in albumin flux. The results indicate that fibrin contact with the endothelial cell increases endothelial permeability. The effect of fibrin may involve activation of temperature-sensitive bulk phase transcytosis of albumin.

Tumor necrosis factor-alpha-mediated decrease in glutathione increases the sensitivity of pulmonary vascular endothelial cells to H2O2

We examined the effects of tumor necrosis factor-alpha (TNF alpha) stimulation of endothelial cells on the increase in endothelial permeability induced by H2O2. Bovine pulmonary microvascular endothelial cells (BPMVEC) were grown to confluence on a microporous filter and the 125I-albumin clearance rate across the monolayer was determined. Pretreatment with TNF alpha (100 U/ml) for 6 h had no direct effect on transendothelial 125I-albumin permeability. However, TNF alpha pretreatment enhanced the susceptibility of BPMVEC to H2O2; that is, H2O2 (10 microM) alone had no direct effect, whereas H2O2 increased 125I-albumin permeability more than threefold when added to monolayers pretreated for 6 h with TNF alpha. Determination of lactate dehydrogenase release indicated that increased permeability was not due to cytolysis. We measured the intracellular contents of GSH and catalase to determine their possible role in mediating the increased susceptibility to H2O2. TNF alpha treatment (100 U/ml for 6 h) decreased total GSH content and concomitantly increased the oxidized GSH content, but did not alter the cellular catalase activity. The role of GSH was examined by pretreating endothelial cells with 2 mM GSH for 3 h, which produced an 80% increase in intracellular GSH content. GSH repletion inhibited the increased sensitivity of the TNF alpha-treated endothelial cells to H2O2. We tested the effects of xanthine oxidase (XO) inhibition since XO activation may be a source of oxidants responsible for the decrease in cellular GSH content. Pretreatment with 0.5 mM oxypurinol attenuated the synergistic effect of TNF alpha and H2O2 on endothelial permeability. The results indicate that decreased oxidant buffering capacity secondary to TNF alpha-induced reduction in intracellular GSH content mediates the increased susceptibility of endothelial cells to H2O2. This mechanism may contribute to oxidant-dependent vascular endothelial injury in septicemia associated with TNF alpha release.

Hypoxia-induced increased permeability of endothelial monolayers occurs through lowering of cellular cAMP levels

Prolonged exposure to hypoxia, as at high altitude, results in increased vascular permeability that may be ameliorated by administration of glucocorticoids. To understand mechanisms underlying these observations, cultured bovine aortic and pulmonary artery endothelial cells (ECs) were subjected to hypoxia, and changes in monolayer permeability and adenosine 3',5'-cyclic monophosphate (cAMP) levels were assessed. Exposure of both types of cultured ECs to hypoxia (PO2 approximately 14 Torr) led to a time- and dose-dependent increase in monolayer permeability, as measured by diffusion of radiolabeled solutes, which was associated with a progressive decrease in EC cAMP levels from 60 to 15 pmol/mg protein, and a decrease in EC adenylate cyclase activity. The change in endothelial barrier function was prevented by addition of cAMP analogues. Pertussis toxin protected EC monolayers from hypoxia-mediated increase in permeability while maintaining cAMP levels and adenylate cyclase activity. Addition of dexamethasone to EC monolayers before or simultaneously with their incubation under hypoxic conditions blocked the hypoxia-mediated increase in monolayer permeability. Dexamethasone pretreatment also prevented the decline in cAMP and adenylate cyclase levels in oxygen-deprived cultures. These data indicate that hypoxia decreases EC barrier function by lowering adenylate cyclase activity and cellular cAMP levels. They suggest that dexamethasone may exert its protective effect, in part, by preventing the hypoxia-induced decline in adenylate cyclase activity, leading to an increase in cellular cAMP and maintenance of EC barrier function.

In vitro penetration of des-tyrosine1-D-phenylalanine3-beta-casomorphin across the blood-brain barrier

The blood-brain barrier transport and metabolism of the synthetic beta-casomorphin (beta CM) derivative des-tyrosine1-D-phenylalanine3-beta-casomorphin (DT-D-Phe3-beta CM) were investigated using an in vitro model consisting of primary cultures of bovine cerebrovascular endothelial cells. DT-D-Phe3-beta CM was transported across the endothelial monolayer without significant metabolism. The endothelial permeability expressing the transport rate ranged between 1.4 and 2.2 cm x 10(-3)/min and was neither affected by luminal concentration changes (1 nM and 1 microM) nor different after luminal and abluminal administration. The metabolic inhibitor 2-desoxy-D-glucose did not affect the permeability of DT-D-Phe3-beta CM. These results suggest that DT-D-Phe3-beta CM is able to cross the blood-brain barrier by paracellular transport without using a carrier system.

The cerebral microvessels in culture, an update

Recent advances in our knowledge of the blood-brain barrier (BBB) have in part been made by studying the properties and function of cerebral endothelial cells in vitro. After an era of working with a fraction, enriched in cerebral microvessels by centrifugation, the next generation of in vitro BBB model systems was introduced, when the conditions for routinely culturing the endothelial cells were established. This review summarizes the results obtained from this rapidly growing field. It can be stated with certainty that, in addition to providing a better insight into the chemical composition of cerebral endothelial cells, much has been learned from these studies about the characteristics of transport processes and cell-to-cell interactions during the last 12 years. With the application of new technologies, the approach offers a new means of investigation, applicable not only to biochemistry and physiology but also to the drug research, and may improve the transport of substances through the BBB. The in vitro approach has been and should remain an excellent model of the BBB to help unravel the complex molecular interactions underlying and regulating the permeability of the cerebral endothelium.

Lectin binding to gp60 decreases specific albumin binding and transport in pulmonary artery endothelial monolayers

The effect of albumin binding to cultured bovine pulmonary artery endothelial cell (BPAEC) monolayers on the transendothelial flux of 125I-labelled bovine serum albumin (BSA) was examined to determine its possible role on albumin transcytosis. The transport of 125I-BSA tracer across BPAEC grown on gelatin- and fibronectin-coated filters (0.8 microns pore diam.) was affected by the presence of unlabelled BSA in the medium in that transendothelial 125I-BSA permeability decreased, reaching a 40% reduction at BSA concentrations equal to or greater than 5 mg/ml. BSA binding to BPAEC monolayers was saturated at concentration of 10 mg/ml with an apparent binding affinity of 6 x 10(-7) M. In contrast, gelatin added to the medium altered neither 125I-BSA binding nor transport. Several lectins were tested for their ability to inhibit 125I-BSA binding and transport. One lectin, Ricinus communis (RCA), reduced 125I-BSA binding by 70% and transport by 40%. Other lectins, Ulex europaeus, Triticum vulgare, and Glycine max decreased neither 125I-BSA binding nor transport. The reduction of 125I-BSA transport by RCA was not observed in the presence of saturating levels of BSA, indicating that RCA influenced only the albumin-dependent component of transport. RCA, but not other lectins, precipitated a 60 kDa plasmalemmal glycoprotein from cell lysates of surface radioiodinated BPAEC monolayers. This 60 kDa glycoprotein appears to be the equivalent of gp60 identified previously as an albumin binding glycoprotein in rat microvascular endothelium. In summary, approximately 40% of albumin transport across BPAEC monolayers is dependent on albumin binding. This component of albumin transport is inhibited by 80% by the binding of RCA to gp60. These results suggest that binding of albumin to gp60 on pulmonary artery endothelial cell membrane is a critical determinant of transendothelial albumin flux involving mechanisms such as plasmalemmal vesicular transcytosis.

Permeability of human venous endothelial cell monolayers perfused in microcarrier cultures: effects of flow rate, thrombin, and cytochalasin D

We have applied a multiple isotope dilution technique to examine junctional permeability of human umbilical vein endothelial cells (HUVEC) in vitro. Primary cultures were grown to confluence on porous Cytodex-3 microcarrier beads, packed into 0.3 ml columns (3 x 10(6) cells) and perfused at varying flow rates (0.3-1.2 ml/min) with HEPES-buffered Tyrodes solution containing unlabeled cyanocobalamin, insulin, and albumin. Columns were challenged periodically with mixtures of radioactive tracers of different molecular size. Permeability to 22Na+, [57Co]cyanocobalamin (1.3 kD), [125I]insulin (6 kD) or [125I]albumin (66 kD) was assessed relative to [131I]IgG (160 kD, impermeant reference tracer) by comparing column elution profiles. Although the single passage extraction of [125I]albumin by beads alone approximated 40%, the presence of confluent HUVEC rendered these beads effectively impermeable to albumin. High junctional extractions were measured for cyanocobalamin (0.79 +/- 0.02, n = 28) and insulin (0.51 +/- 0.05, n = 14) in cultures perfused at 0.3-0.4 ml/min, and tracer extraction decreased as perfusion rates increased. Permeability coefficients for cyanocobalamin (9.66 x 10(-5) cm/s) and insulin (4.18 x 10(-5) cm/s) increased significantly during perfusion with thrombin (10 U/ml) or cytochalasin D (1 microgram/ml), whereas permeability to albumin (0.39 x 10(-5) cm/s) remained unchanged. Morphological studies, using the glycocalyx stain ruthenium red, revealed that thrombin or cytochalasin D increased the penetration of the stain into junctions between endothelial cells.

Serum albumin decreases transendothelial permeability to macromolecules

We examined the effects of serum albumin and other serum proteins on the fluxes of tracer 125I-albumin (MW 69 kDa) and 125I-haptoglobin (MW 100 kDa) across the pulmonary artery endothelial monolayer in vitro to test the role of serum proteins in modulating the endothelial barrier function. Replacement of control complete culture medium (20% fetal calf serum in DMEM) with DMEM alone increased the transendothelial 125I-albumin clearance rate (a measure of 125I-albumin permeability) by 83% of the control value. Repletion with 50% calf serum or with 2.0 g% albumin (i.e., the albumin concentration in 50% serum) decreased 125I-albumin permeability to the control value. This effect of serum or albumin was concentration-dependent since neither 12.5% serum nor 0.5 g% albumin (i.e., albumin concentration in 12.5% serum) altered 125I-albumin permeability from control values. The ammonium sulfate-precipitated serum protein fraction rich in albumin decreased 125I-albumin permeability from the control DMEM value, whereas serum fractions containing predominantly gamma-globulin or depleted of protein did not significantly alter 125I-albumin permeability. Other serum proteins that have been proposed to reduce endothelial permeability, alpha 1-acid glycoprotein (0.035-0.14 g/100 ml) and fibronectin (5 mg/100 ml), did not decrease 125I-albumin permeability from DMEM values. The endothelial permeability of 125I-haptoglobin of 4.63 +/- 0.53 x 10(-6) cm/sec in the presence of DMEM was 30% of the 125I-albumin permeability value. The addition of 2.0 g% albumin or 50% serum decreased 125I-haptoglobin permeability to 57 and 31%, respectively, of the DMEM value. These results indicate the critical role of serum albumin in regulating the restrictiveness of the endothelial barrier to macromolecules.

Mechanism of endothelin-1-induced pulmonary vasoconstriction

Endothelins are endothelial cell-derived peptides with potent vasoconstrictor properties. We investigated the actions of porcine/human endothelin-1 (ET-1) on the microvasculature of the guinea pig lung perfused at constant flow with Ringers-albumin. We measured the perfusion pressure, distribution of pulmonary vascular resistance (using the double occlusion method), lung weight change, and the pulmonary capillary filtration coefficient. At concentrations of greater than or equal to 10(-10) M, ET-1 produced dose-dependent increases in mean pulmonary artery pressure (EC50, approximately 10(-9.5) M), which were rapid in onset and biphasic (first phase peaking at 1-2 minutes; second phase peaking at 10-15 minutes) up to 60 minutes of the perfusion period. The vasoconstrictor response was sustained for the 60-minute perfusion period. The pulmonary vasoconstriction was inhibited by pretreatment with indomethacin (10(-5) M), the thromboxane A2 receptor antagonist SQ-29,548 (4 x 10(-6) M), or papaverine (10(-5) M). Nifedipine (10(-5) or 10(-7) M) had no effect on the first phase but prevented the second phase of the vasoconstriction. The vasoconstriction was primarily the result of a 10-fold increase in pulmonary venous resistance. Pulmonary edema developed after ET-1 challenge because of the venoconstriction and the resultant pulmonary capillary hypertension. However, the pulmonary capillary filtration coefficient was unchanged, indicating that pulmonary vascular permeability did not increase. ET-1 also had no effect on transendothelial 125I-albumin flux. The results indicate that ET-1 is a potent thromboxane-dependent venoconstrictor in the guinea pig lung.(ABSTRACT TRUNCATED AT 250 WORDS)

Tumor necrosis factor enhances the neutrophil-dependent increase in endothelial permeability

We examined the effect of tumor necrosis factor alpha (TNF alpha) on the increase in pulmonary microvascular endothelial monolayer permeability induced by activated neutrophils (PMN). Layering of PMN onto endothelial monolayers followed by activation of PMN with phorbol 12-myristate 13-acetate (PMA) increased 125I-albumin clearance rate across the monolayers. Pretreatment of endothelial monolayers for 6 hr with TNF alpha (200 U/ml) potentiated the PMN-dependent increase in endothelial permeability, whereas 1 hr or 6 hr pretreatment of endothelial monolayers with 200 U/ml and 100 U/ml, respectively, TNF alpha did not enhance the response. Adherence of PMN to the endothelial cells was increased at 1 and 6 hr after TNF alpha (200 U/ml) treatment, but the adherence response was markedly greater following 6 hr of TNF alpha. The TNF alpha treatment of endothelial cells did not enhance neutrophil activation responses to PMA. Pretreatment of PMN with IB4, a MAb to the CD18 integrin, the common beta subunit of the adhesion proteins LFA-1, Mac-1, and p150,95 of PMN, reduced the increases in PMN adherence and the endothelial monolayer permeability induced by the 6 hr TNF alpha treatment. In contrast, pretreatment of PMN with OKM-1, a MAb to the CD11b epitope (alpha-subunit), had no effect on the adherence and the potentiation of the increase in permeability. The potentiation of the PMN-dependent permeability increase and enhanced endothelial adhesivity at 6 hr after TNF alpha priming of endothelial cells was dependent on protein synthesis. The results indicate that protein synthesis-dependent expression of an endothelial ligand for CD18 and resultant endothelial hyperadhesiveness potentiates the PMN-mediated increase in endothelial permeability after TNF alpha activation of endothelial cells. The priming of endothelial cells by TNF alpha may be a critical step in the mediation of endothelial injury.

Pulmonary edema induced by phagocytosing neutrophils. Protective effect of monoclonal antibody against phagocyte CD18 integrin

We studied the changes in pulmonary hemodynamics and lung wet weight induced with opsonized zymosan (OZ) in isolated guinea pig lungs perfused with Ringer-albumin solution containing neutrophils (PMNs). Addition of OZ to the PMN-perfused lungs caused pulmonary vasoconstriction and weight gain; neither OZ nor PMNs added individually to the perfusate altered pulmonary vasomotor tone or wet weight. The steady gain in lung weight by 1,588 +/- 464 mg over the 45-minute study period was associated with pulmonary capillary hypertension and an increase in the capillary filtration coefficient, indicative of increased lung vascular permeability. These responses may not be due to generation of oxygen radicals, because the alterations in pulmonary hemodynamics and lung weight were not reduced by addition of superoxide dismutase, catalase, or superoxide dismutase plus catalase. We examined the basis of the PMN-mediated effects by layering PMNs on bovine pulmonary artery endothelial monolayers. Challenge with OZ resulted in increased endothelial permeability to 125I-albumin. The monoclonal antibody IB4 (directed against CD18, the common beta-subunit of structurally related adhesion receptors on phagocytes, LFA-1, Mac-1, and P150,95) prevented the OZ-mediated increase in PMN adherence to endothelial cells and the increase in endothelial permeability to 125I-albumin. IB4 also inhibited the lung weight gain mediated by the OZ-stimulated PMNs in intact lungs. The protective effect of IB4 could be ascribed neither to inhibition of uptake of OZ by PMNs nor to inhibition of release of oxygen radicals, myeloperoxidase, and elastase.(ABSTRACT TRUNCATED AT 250 WORDS)

Increased endothelial albumin permeability mediated by protein kinase C activation

We examined the effects of activation of endothelial protein kinase C (PKC) of the endothelial barrier function. Exposure of confluent bovine pulmonary artery endothelial cell monolayers to phorbol 12-myristate 13-acetate (PMA) resulted in concentration-dependent (10(-8)-10(-6) M) increases in PKC activity and in the transendothelial flux of 125I-albumin. Exposure of the endothelium to 1-oleoyl 2-acetyl glycerol (OAG) also increased the transendothelial flux of 125I-albumin in a concentration-dependent manner. Neither 4 alpha-phorbol didecanoate nor 1-mono-oleoyl glycerol, which do not activate PKC, altered permeability. The increase in 125I-albumin permeability induced by PMA was inhibited by 25 microM H7 (a PKC inhibitor), but not by the control compound HA1004 (25 microM). After 16 h of exposure to PMA, 125I-albumin permeability returned to baseline and a significant reduction in cytosolic PKC activity was noted. Further challenge with PMA at this time resulted in no significant increase in PKC activity indicating downregulation of the enzyme; moreover, this PMA challenge did not increase endothelial permeability. Exposure of endothelial monolayers to phospholipase C (PLC), which increases membrane phosphatidylinositide turnover, or to alpha-thrombin also induced concentration-dependent activation of PKC and increases in 125I-albumin endothelial permeability. The thrombin- and PLC-induced permeability increases were inhibited by H7, but not by HA1004. The activation of endothelial PKC directly by PMA or OAG and by PLC and alpha-thrombin increases the transendothelial albumin permeability, indicating that PKC activation is an important signal transduction pathway by which extracellular mediators increase endothelial macromolecular transport.

Hypoxia/reoxygenation increases the permeability of endothelial cell monolayers: role of oxygen radicals

We assessed the effect of hypoxia/reoxygenation on 14C-albumin flux across endothelial monolayers. Cultured bovine pulmonary artery endothelial cells were grown to confluence on nitrocellulose filters (pore size 12 microns). The endothelialized filters were mounted in Ussing-type chambers which were filled with cell culture medium (M 199). Equimolar amounts (33 nM) of 14C-labeled and unlabeled albumin were added to the "hot" and "cold" chambers, respectively. The monolayers were then exposed to successive periods (90 min) of normoxia (pO2 145 mmHg), hypoxia (pO2 20 mmHg), and reoxygenation (pO2 145 mmHg). A gas bubbling system was used to control media pO2 and to ensure adequate mixing. Four aliquots of culture media were taken during each period in order to calculate the 14C-albumin permeability across the endothelialized filter. In some experiments, either the xanthine oxidase inhibitor, oxypurinol (10 microM), or superoxide dismutase (600 U/mL), was added to the media immediately prior to the experiments. As compared to the normoxic control period, albumin permeability was 1.5 times higher during hypoxia (p less than 0.01) and 2.3 times higher during reoxygenation (p less than 0.01). The reoxygenation-induced increase in albumin permeability was prevented by either oxypurinol or superoxide dismutase. These data indicate that xanthine oxidase-derived oxygen radicals contribute to the hypoxia/reoxygenation-induced endothelial cell dysfunction. The altered endothelial barrier function induced by hypoxia/reoxygenation is consistent with the microvascular dysfunction observed following reperfusion of ischemic tissues.

Role of albumin arginyl sites in albumin-induced reduction of endothelial hydraulic conductivity

We determined the effect of albumin on endothelial hydraulic conductivity (Lp) and the contributions of the positively charged arginyl and lysinyl residues of albumin in mediating the effect. Studies were made using monolayers of cultured sheep pulmonary artery endothelial cells grown to confluence on polycarbonate filters. Water flux was measured as transendothelial hydrostatic pressure was varied from 5 to 20 cm H2O. Lp was calculated from the slope of the relationship of water flux versus pressure. The Lp of endothelial monolayers perfused with albumin-free Hanks Balanced Salt Solution (HBSS) was compared to perfusion with HBSS containing either native albumin, or albumin in which the arginyl residues were modified by a condensation reaction with 1,2-cyclohexanedione (CHD-albumin), or albumin in which the lysinyl residues were modified by a substitution reaction with succinic anhydride (SC-albumin). Baseline Lp at 2.5 mg/ml native albumin was 1.6 +/- 0.1 X 10(-6) cm/s/cm H2O compared to the filter Lp after removing cells of 3.0 +/- 0.3 X 10(-4) cm/s/cm H2O. Endothelial Lp increased by 60% when albumin concentration was decreased from 2.5 mg/ml to 0.5 mg/ml (P less than 0.05), but did not change with an increase in concentration to 10 mg/ml. Albumin-free buffer and CHD-albumin increased endothelial Lp by 2.2 +/- 0.3-fold and 1.9 +/- 0.3-fold, respectively (P less than 0.05). All endothelial Lp values were restored to baseline when the native albumin concentration was returned to 2.5 mg/ml. Excess l-arginine (2 X 10(-3) M) inhibited the effect of native albumin and increased endothelial Lp 1.5 +/- 0.02-fold (P less than 0.05), but excess l-lysine (4 X 10(-3) in the presence of native albumin had no effect on Lp. None of the perfusates altered the filter Lp value. Neutral dextran (70 kD), in contrast to native albumin, had no effect on endothelial Lp. These results indicate that albumin reduces the hydraulic conductivity of endothelial monolayers in a concentration-dependent fashion and that the arginyl residues of albumin are required for the response. The effect of albumin may be mediated by a charge interaction of albumin with the endothelium.

Phallacidin prevents thrombin-induced increases in endothelial permeability to albumin

Calf pulmonary artery endothelial monolayers cultured on polycarbonate filters were utilized to study 125I-labeled albumin permeability and actin filament distribution in response to thrombin challenge. Thirty-minute exposure to alpha-thrombin (10(-7) M) significantly increased albumin clearance rates. These changes were associated with marked alterations in actin filament distribution, resulting in loss of peripheral actin bands and an increase in the number of cytoplasmic stress fibers. Because the actin peripheral filaments are thought to play an important role in junctional stability, we postulated that stabilization of actin filaments should protect against thrombin-induced barrier disruptions. Pretreatment of cells with 0.3 microM 7-nitrobenz-2-oxa-1,3-diazole (NBD)-phallacidin, a specific actin-stabilizing agent, prevented the changes in actin filament distribution and markedly attenuated the increase in albumin permeability. Because of the potential toxicity of phallatoxins, we evaluated the effects of pretreatment on cell viability and growth parameters. There were no differences in viability, seeding efficiency, or doubling times in cells treated with 0.3 microM NBD-phallacidin in comparison to controls. Our data support the hypothesis that actin filaments, particularly peripheral bands, contribute significantly to the maintenance of barrier function in cultured endothelial cells.

Endothelial barrier function

The endothelial barrier in all organ beds allows the free exchange of water, but is restrictive to varying degrees to the transport of solutes such as albumin. For example, in the brain microvessels, the endothelial barrier restricts the transport of protein, whereas in fenestrated and continuous endothelial cells of the renal and lung endothelial cells, the endothelial barrier is semipermeable. The endothelial monolayer demonstrates selectivity, i.e., the permeation of molecules is inversely related to the molecular weight. Although the "pore" theory has been used to describe the transport across the endothelial barrier, the transport of solutes is also dependent on the charge of solutes and the endothelial cell, and the ability of the solute to bind to or be taken up by endothelial cells. Receptor-mediated trancytosis of albumin may contribute to albumin transport in addition to transport by paracellular pathways (i.e., through a so-called pore). Water permeability across the endothelium is determined by the interaction of albumin with glycocalyx and interstitial components of the endothelium (the "fiber matrix"). Ambient concentration of albumin serves to lower endothelial hydraulic conductivity. Increased endothelial permeability to solutes and water in inflammatory states is dependent on the shape and configuration of endothelial cells as determined by alterations in cytoskeletal elements, such as f-actin, and as regulated by intracellular second messengers such as free cytosolic calcium.

Permeability and mechanism of albumin, cationized albumin, and glycosylated albumin transcellular transport across monolayers of cultured bovine brain capillary endothelial cells

We have measured the permeability and binding characteristics of bovine serum albumin (BSA), cationized BSA (cBSA), and glycosylated BSA (gBSA) to primary cultures of bovine brain capillary endothelial cells (BBCEC). These endothelial cells serve as an in vitro model to study the binding, uptake, and transcellular transport of small and large molecule flux across the blood-brain barrier. The rate of [3H]BSA flux across the cultured BBCEC monolayers grown onto polycarbonate membranes (5-microns pore size) was linear with increasing BSA concentration and the flux could be inhibited by temperature reduction to 0-4 degrees C. The maximal binding of [3H]BSA was 0.04 fmol/mg total cell protein and could not be inhibited by nonradiolabeled BSA. The binding of cBSA and gBSA was rapid and could be inhibited by nonradiolabeled cBSA or gBSA, respectively. The maximal amount bound was 1.8 fmol/mg total cell protein for cBSA and 17.4 fmol/mg total cell protein for gBSA. The dissociation constants (Kd's) were 27 +/- 13 and 3.7 +/- 1.1 nM for cBSA and gBSA, respectively. The flux rates of cBSA and gBSA across the endothelial cell monolayers were linear with respect to concentration and they were approximately seven times greater than those observed for BSA. Each of the proteins appeared on the antiluminal side of the endothelial cell monolayers primarily (90%) as intact protein as determined by trichloroacetic acid (TCA) precipitations and sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE). The results for BSA are similar to those observed for lucifer yellow, a fluid-phase endocytic marker.(ABSTRACT TRUNCATED AT 250 WORDS)

Platelets decrease albumin permeability of pulmonary artery endothelial cell monolayers

Since platelets may modulate endothelial cell permeability, we examined the effects of platelets on 125I-albumin permeability of cultured bovine pulmonary artery endothelial cell monolayers. The experimental system consisted of endothelial cells grown to confluence on a gelatinized polycarbonate filter. We quantified the diffusive flux of 125I-albumin from luminal chamber to the abluminal chamber. Washed human platelets added to the monolayers decreased the albumin flux in a concentration-dependent manner, with a 65% decrease occurring at the highest concentration of platelets (5 x 10(7) platelets) added to the 700-microliters luminal chamber. In contrast, neither paraformaldehyde-fixed platelets nor fresh red blood cells changed 125I-albumin permeability. Platelets had no effect on 125I-albumin permeability across the gelatinized filters without endothelial cells present. Supernatants of platelet lysates also reduced albumin flux. The effect produced by intact platelets or platelet lysate was not influenced by the presence of ketanserin (a serotonin receptor antagonist), propranolol (a beta-adrenergic receptor antagonist), or aspirin (an inhibitor of cyclooxygenase). Platelets activated by thrombin did not produce an effect that was different from the effect produced by intact platelets. The activity of the supernatant of platelet lysate remained in the aqueous phase after ether extraction. The results indicate that the platelet-mediated decrease in endothelial cell permeability to 125I-albumin is the result of a hydrophilic platelet-derived factor(s) and not secondary to mechanical obstruction of endothelial "leaks" by the platelets.

Role of cyclic adenosine monophosphate in the induction of endothelial barrier properties

Cyclic adenosine monophosphate (AMP) has numerous important effects on cell structure and function, but its role in endothelial cells is unclear. Since cyclic AMP has been shown to affect transmembrane transport, cell growth and morphology, cellular adhesion, and cytoskeletal organization, it may be an important determinant of endothelial barrier properties. To test this we exposed bovine pulmonary artery endothelial cell monolayers to substances known to increase cyclic AMP and measured their effect on endothelial permeability to albumin and endothelial cell cyclic AMP concentrations. Cholera toxin (CT), a stimulant of the guanine nucleotide binding subunit of adenylate cyclase, led to a concentration-dependent 2-6-fold increase in cyclic AMP which was associated with a 3-10-fold reduction in albumin transfer across endothelial monolayers. The effect was not specific to albumin as similar barrier-enhancing effects were also noted with an unrelated macromolecule, fluorescein isothiocyanate (FITC)-dextran (MW 70,000). Barrier enhancement with cyclic AMP elevation was also observed with forskolin, a stimulant of the catalytic subunit of adenylate cyclase. The temporal pattern of barrier enhancement seen with these agents paralleled their effects on increasing cyclic AMP, and the barrier enhancement could be reproduced by incubation with either dibutyryl cyclic AMP or Sp-cAMPS, cyclic AMP-dependent protein kinase agonists. Furthermore, the forskolin effect on barrier enhancement was partially reversed with Rp-cAMPS, an antagonist of cyclic AMP-dependent protein kinase. Since endothelial actin polymerization may be an important determinant of endothelial barrier function, we sought to determine whether the cyclic AMP-induced effects were associated with increases in the polymerized actin pool (F-actin). Both cholera toxin and forskolin led to apparent endothelial cell spreading and quantitative increases in endothelial cell F-actin fluorescence. In conclusion, increased endothelial cell cyclic adenine nucleotide activity was an important determinant of endothelial barrier function in vitro. The barrier enhancement was associated with increased endothelial apposition and increases in F-actin, suggesting that influences on cytoskeletal assembly may be involved in this process.

The transport of albumin across the ferret in vitro whole trachea

1. The whole trachea of the ferret has been isolated in vitro in an organ bath and used to study the transport of bovine serum albumin (BSA) and two dextrans (70,000 and 9000 Da) from external buffer solution to air-filled lumen, assessed by fluorescent-labelled tracers. 2. In control conditions, when mucus secretion was not stimulated by drugs, the concentration of albumin in the lumen was over half that in the buffer, and about six times greater than those of the two dextrans. 3. Methacholine and phenylephrine caused large increases in mucus secretion and albumin output and decreases in albumin concentration. The responses were proportional to drug concentration. We concluded that albumin output is increased but diluted with submucosal gland secretion. 4. Salbutamol caused a small increase in mucus secretion and large increases in output and concentration of albumin. The concentration of albumin became greater than that in the external buffer medium. The responses were proportional to concentration of salbutamol. 5. Histamine increased mucus secretion and albumin output and concentration. 6. None of the four drugs increased the output of dextran-70,000. Methacholine and phenylephrine increased the output of dextran-9000, but to a far less extent than for albumin. 7. Cooling the trachea and buffer to 4 degrees C almost abolished the stimulation of mucus and albumin outputs due to methacholine. 8. Increasing the concentration of albumin external to the trachea did not proportionally increase albumin secretion, the logarithmic relationship suggesting saturation of an active transport system. 9. We conclude that albumin is secreted by active transport into the tracheal lumen, and that the rate of transport can be augmented by salbutamol to build up a higher concentration in the lumen than in the external buffer.

Lipoxygenase products induce neutrophil activation and increase endothelial permeability after thrombin-induced pulmonary microembolism

We examined the mechanism of the neutrophil (PMN)-dependent increase in pulmonary vascular permeability to protein after thrombin-induced pulmonary microembolism. Humoral factors that activate PMNs after thrombin-induced pulmonary microembolism were characterized in pulmonary lymph obtained from unanesthetized sheep challenged with intravenous infusion of alpha-thrombin. Time-dependent increases in PMN migration, aggregation, and superoxide anion (O2-) generation were induced by the pulmonary lymph obtained within 20 minutes after thrombin infusion. The pulmonary lymph neutrophil activating factors present in ether extracts of lymph had retention times of leukotriene B4 (LTB4) and monohydroxyeicosatetraenoic acids (HETEs) by high-performance liquid chromatography. The postthrombin lymph samples containing the LTB4 and HETEs increased PMN O2- generation and endothelial monolayer permeability to 125I-albumin in the presence of PMNs layered on the endothelial monolayers. Control lymph samples replete with LTB4, 5-HETE, and 15-HETE induced increases in PMN O2- generation and endothelial monolayer permeability to 125I-albumin in the presence of PMNs layered on the endothelial monolayers. Maximal increases in PMN O2- production and endothelial permeability occurred when LTB4, 5-HETE, and 15-HETE were coincubated with PMNs, indicating a synergistic action of these mediators in inducing PMN activation. Endothelial monolayer permeability to 125I-albumin did not increase with postthrombin lymph samples obtained after pretreatment with the 5-lipoxygenase inhibitor, L-651,392. The results indicate that lipoxygenase products generated in the lungs after thrombin-induced microembolism contribute to increased endothelial permeability secondary to PMN activation.

Recombinant tumor necrosis factor increases pulmonary vascular permeability independent of neutrophils

We studied the effects of intravenous infusion of recombinant human tumor necrosis factor type alpha (rTNF-alpha; 12 micrograms/kg) on lung fluid balance in sheep prepared with chronic lung lymph fistulas. The role of neutrophils was examined in sheep made neutropenic with hydroxyurea (200 mg/kg for 4 or 5 days) before receiving rTNF-alpha. Infusion of rTNF-alpha resulted in respiratory distress and 3-fold increases in pulmonary arterial pressure and pulmonary vascular resistance within 15 min, indicating intense pulmonary vasoconstriction. Pulmonary lymph flow (i.e., net transvascular fluid filtration rate) and transvascular protein clearance rate (a measure of vascular permeability to protein) increased 2-fold within 30 min. The increased permeability was associated with leukopenia and neutropenia. The pulmonary hypertension and vasoconstriction subsided but fluid filtration and vascular permeability continued to increase. Sheep made neutropenic had similar increases in pulmonary transvascular fluid filtration and vascular permeability. rTNF-alpha also produced concentration-dependent increases in permeability of 125I-labeled albumin across ovine endothelial cell monolayers in the absence of neutrophils or other inflammatory mediators. The results indicate that rTNF-alpha increases pulmonary vascular permeability to protein by an effect on the endothelium.

Selectivity of the endothelial monolayer: effects of increased permeability

We investigated the mechanism of thrombin-induced increases in endothelial monolayer permeability by examining the effect of thrombin on the molecular sieving characteristics of the endothelial monolayer and comparing the responses of arterial- and venous-derived endothelial cell lines. Bovine pulmonary artery (BPA) and pulmonary vein (BPV) endothelial cells were similarly harvested and cultured. The endothelial cells were grown to confluence on gelatinized polycarbonate filters and the permeabilities to sucrose, albumin, and IgG were measured and corrected for effects of unstirred layers. The control permeabilities of BPA and BPV were similar with both monolayers, demonstrating selectivity to different sized tracer molecules. alpha-Thrombin (10(-6) M) increased the permeability of both BPA and BPV to albumin and sucrose. The permeability of BPA was increased to a greater extent than BPV, perhaps due to phenotypic differences. In both cell lines, the permeability increase was most pronounced for albumin, which by pore theory is best described by an increase in the radius of the small pore pathway for diffusion.

Effects of alterations in endothelial cell volume on transendothelial albumin permeability

We examined the effects of alterations in endothelial cell volume on transendothelial albumin permeability. Studies were done using a confluent monolayer of bovine pulmonary artery endothelial cells grown on gelatinized microporous filters. When endothelial cells were exposed to media made hypertonic with 200 mM mannitol, the intracellular volume (measured with 14C-urea) decreased twofold and remained decreased over a 30-minute time-span, thus showing no significant regulatory volume increase (RVI) within this time period. When endothelial cells were exposed to hypotonic media, intracellular volume rapidly doubled within 2 minutes, and then decreased to baseline values within 10 minutes in spite of the sustained hypotonic environment, a process known as regulatory volume decrease (RVD). We also measured the transendothelial flux of 125I-albumin with the cells exposed to the same osmotic changes. We observed that only under hypertonic conditions was there a significant change in the 125I-albumin permeability. These results indicate that the pulmonary artery endothelial cells in culture alter their cell volume when exposed to variations in the osmotic environment, and also show RVD in response to hypotonic conditions but no RVI within 40 minutes after exposure to hypertonic conditions. The transendothelial albumin permeability did not change under hypotonic conditions but increased under hypertonic conditions. Thus, endothelial cells shrinkage may be an important mechanism of increased endothelial macromolecule permeability. These volume changes may occur in endothelial cells in situ and have a role in inducing alterations in the transendothelial permeability to proteins.