Neutrophil-mediated injury to endothelial cells. Enhancement by endotoxin and essential role of neutrophil elastase

The disturbance of hemostasis in septic shock: role of neutrophil elastase and thrombin, effects of antithrombin III and plasma substitution

In 42 patients with septic shock, 29 of whom underwent substitution with antithrombin III concentrate and fresh frozen plasma for coagulation disorders, the proteinase-inhibitor complexes thrombin-antithrombin III and neutrophil elastase-alpha 1 proteinase inhibitor, were elevated on admission. On admission, the elastase complex was significantly higher in the patients receiving substitution (p = 0.0039), but at the endpoint it was higher in the non-survivors (p = 0.0040). The elastase decrease was confined to the substitution group with the thrombin complex decreasing in both groups. Initially the thrombin complex correlated with prothrombin times and factor XIII, while the elastase complex correlated with creatinine, thrombocyte count and prothrombin times in the late stages. Hemostatic disturbance, thrombin generation and neutrophil elastase release were favorably influenced by substitution. Furthermore, in this uncontrolled pilot study, the survival rate was higher in the treated (16 of 29) than in the untreated (1 of 13) patients, although the treated patients initially had pronounced hemostatic disturbances.

Endotoxin stimulates endothelin-release in vivo and in vitro as determined by radioimmunoassay

A marked increase in immunoreactive endothelin was observed in rat serum collected within 10-15 min after infusion of endotoxin. Endothelin level was 117 +/- 11.5 pg/ml (mean +/- S.E., N = 4) in rats exposed to endotoxin as compared with undetectable levels (less than 2 pg/ml, N = 4) in controls. We have also observed a significant stimulation of endothelin-release by endotoxin from cultured bovine transformed thoractic aortic endothelial cells at concentrations of endotoxin ranging between 0.1 and 10.0 micrograms/ml. Serum was indispensable for the stimulating effect of endotoxin, although serum itself did not show any effect at the concentration used (1%). These results suggest that endothelin plays an important role in mediation of pathophysiological responses caused by endotoxin. The levels of endothelin were measured by radioimmunoassay with high sensitivity.

Promotion and subsequent inhibition of plasminogen activation after administration of intravenous endotoxin to normal subjects

To evaluate the effect of endotoxin on the fibrinolytic response, we administered Escherichia coli endotoxin (4 ng per kilogram of body weight) intravenously to 19 healthy volunteers and measured fibrinolytic proteins, protease inhibitors, neutrophil elastase, and von Willebrand factor in serial blood samples obtained over 24 hours. One hour after endotoxin administration, the level of tissue plasminogen activator (t-PA) antigen rose from 10 to 23 ng per milliliter, peaking at 52 ng per milliliter at three hours. The level of alpha 2-plasmin inhibitor-plasmin complexes increased sevenfold, peaking at three hours. Plasminogen-activator inhibitor-1 activity rose more slowly, from 7 U per milliliter to a maximum of 49 U per milliliter at five hours. The concentrations of neutrophil elastase and von Willebrand antigen were unchanged at one hour, increased approximately threefold by 3 hours, and remained elevated at 24 hours. None of these measures changed in a control group (n = 5) given intravenous saline instead of endotoxin. We studied t-PA functional activity in four subjects. The level of activity rose rapidly, from 1.2 ng per milliliter at base line to 8.3 ng per milliliter at one hour and 13.9 ng per milliliter at two hours; it was undetectable at three hours. This increase in plasminogen activator activity was abolished in vitro by incubation of t-PA with an antiserum specific for human t-PA, suggesting that t-PA may be directly responsible for plasmin generation in the response to endotoxin. We conclude from this study of healthy subjects that endotoxin activates the fibrinolytic system, beginning with release of t-PA in the blood within one hour. The early activation of plasmin by endotoxin may prevent thrombosis, and the increase in fibrinolysis is then offset by the release of plasminogen activator inhibitor.

The polymorphonuclear leucocyte count in childhood haemolytic uraemic syndrome

Review of data from 79 children with the haemolytic uraemic syndrome (HUS) showed that the polymorphonuclear leucocyte (PMN) count at presentation in childhood HUS predicts outcome. Logistic regression analysis of several features at presentation identified only the PMN count and the presence of a diarrhoeal prodrome as having a significant effect on the outcome (P less than 0.01 and P less than 0.001 respectively). The geometric mean PMN count was significantly raised in 70 children who had typical HUS following a diarrhoeal prodrome (D+ cases) compared with that of 9 children who had atypical disease without diarrhoea (D- cases) (t-test on log-transformed data, P less than 0.005). Fifty-seven children with D+ HUS who recovered completely had a significantly lower geometric mean PMN count than D+ cases with a bad outcome (P less than 0.001). Four of these patients, who died in the acute stage of the disease, had a significantly higher mean count than the rest of the D+ patients (P less than 0.001). Multiple regression analysis demonstrated that the PMN count in D+ cases was not significantly influenced by haemoglobin concentration, platelet count, length of the prodrome, or the administration of antibiotics in the prodromal period. A high PMN count at presentation in D+ HUS indicates a poor prognosis. The data emphasise the heterogeneity of HUS and suggest that PMN participate in the pathogenesis of the disorder in typical D+ cases but not in atypical D- cases.

Limitation of myocardial reperfusion injury by intravenous perfluorochemicals. Role of neutrophil activation

Neutrophil activation and infiltration into the ischemic myocardium after reperfusion may limit the amount of salvageable myocardium (reperfusion injury). The effects of intravenous perfluorochemicals (Fluosol-DA) on infarct size, ventricular contractility, and neutrophil function were assessed in an occlusion-reperfusion canine model. Closed-chest dogs were subjected to 90 minutes of left anterior descending artery occlusion followed by 24 hours of reperfusion. Animals were randomized to receive either Fluosol-DA (FDA, n = 8) or Ringer's lactate (CONT, n = 10) intravenously over 30 minutes just before left anterior descending artery reperfusion. Neutrophil demargination and infiltration into the myocardium were assessed in vivo with In111. Neutrophil chemotaxis, superoxide radical production, and lysozyme degranulation were evaluated ex vivo at baseline, 1 hour after occlusion, and 1 hour after reperfusion. Perfluorochemicals significantly reduced infarct size expressed as percent of area at risk (FDA, 7 +/- 4%; CONT, 24 +/- 6%; p less than 0.01). This was associated with positive wall motion in the jeopardized zone of Fluosol-DA animals compared with dyskinesis in control animals (FDA, +4.4 +/- 2.1%; CONT, -1.1 +/- 1.5%; p less than 0.05). Electron microscopy showed reduced neutrophil and erythrocyte plugging of capillaries with relative preservation of endothelial cells in the Fluosol-DA animals. Myocardial blood flow was greater in the ischemic endocardium of Fluosol-DA animals 1 hour after reperfusion (FDA, 1.23 +/- 0.21; CONT, 0.62 +/- 0.08 ml/g/min; p less than 0.01). Neutrophil demargination and infiltration into the ischemic myocardium was reduced in the animals treated with Fluosol-DA. (FDA, 2.5 +/- 0.7 x 10(3); CONT, 14.1 +/- 2.7 x 10(3) neutrophils/g; p less than 0.01). Neutrophil chemotaxis and lysozyme release were also markedly suppressed in the Fluosol-DA groups ex vivo. These results show that intravenous Fluosol-DA significantly reduces reperfusion injury with greater salvage of myocardium and improved left ventricular function. The chief mechanism of action of Fluosol-DA appears to be the suppression of neutrophil function.

Adherence of neutrophils to cultured human microvascular endothelial cells. Stimulation by chemotactic peptides and lipid mediators and dependence upon the Mac-1, LFA-1, p150,95 glycoprotein family

The process of neutrophil adhesion to and migration through the microvascular endothelium, an early event in the induction of the acute inflammatory response, has been attributed to the generation of extravascular chemoattractants. Although both chemotactic peptides and lipid mediators enhance neutrophil adherence in vitro and in vivo, the mechanism(s) involved in the interaction between circulating neutrophils and microvascular endothelial cells is still not completely understood. In a microtiter well adherence assay, the chemotactic peptides, FMLP and C5a, and the lipid mediators, leukotriene B4 (LTB4) and platelet activating factor (PAF), enhanced human neutrophil adherence to cultured human microvascular endothelial cells as well as to human umbilical vein endothelial cells in a dose-dependent manner with a rapid time course. This stimulated adhesive interaction between neutrophils and cultured human endothelial cells was dependent on the expression of the Mac-1, LFA-1, p150,95 glycoprotein family on the neutrophil surface since neutrophils from patients with leukocyte adhesion deficiency, lacking surface expression of the adhesive glycoproteins, exhibited markedly diminished adherence to human endothelial cells in response to stimulation with chemotactic factors compared to normal control neutrophils. All four mediators enhanced expression of the glycoprotein family on the surface of normal neutrophils as determined by flow cytofluorimetry using a monoclonal antibody (TS1/18) to the glycoprotein common beta subunit. In addition, TS1/18 inhibited up to 100% the adherence of normal neutrophils to endothelial cells stimulated by maximal concentrations of FMLP, C5a, LTB4, or PAF. Moreover, HL-60 cells, human promyelocytic leukemia cells, neither increased glycoprotein surface expression nor adherence in response to stimulation. Thus, peptide and lipid mediators of the acute inflammatory response appear to enhance adherence of circulating neutrophils to the microvascular endothelium by a mechanism dependent on expression of the Mac-1, LFA-1, p150,95 glycoprotein family on the neutrophil surface.

Endothelial and myocardial injury during ischemia and reperfusion: pathogenesis and therapeutic implications

Early reperfusion remains the most effective way of limiting myocardial necrosis and improving ventricular function in experimental models and human patients. However, the introduction of oxygen and cellular elements, especially the neutrophil, into the ischemic zone may initiate a deleterious cascade of events that limits myocardial salvage after reperfusion. Although the pathogenesis of reperfusion injury remains controversial, recent studies have suggested that the endothelium may play a critical role. Endothelial cells maintain flow in the microcirculation by secreting a number of vasodilatory compounds and substances that prevent plugging of capillaries by inhibiting neutrophil adherence and platelet aggregation. Reperfusion of ischemic myocardium accelerates structural and functional changes in endothelial cells, resulting in a progressive decrease in microcirculatory flow ("no reflow" phenomenon). Numerous studies suggest that activated neutrophils mediate vascular damage by releasing reactive oxygen species and potent proteolytic enzymes. The administration of therapeutic agents that limit endothelial disruption and neutrophil plugging has shown promising results in limiting myocardial reperfusion injury in experimental models.

Diagnostic criteria for adult respiratory distress syndrome: time for reappraisal

To assess whether current diagnostic criteria for the adult respiratory distress syndrome (ARDS) may limit research to patients with a range of severity of respiratory failure and risk factors for ARDS were studied. In 28 patients ARDS did not develop; in 9 it did; and the other 13 met the diagnostic criteria at their first assessment. Plasma concentration of neutrophil elastase-alpha 1-antitrypsin complex was higher in all groups than in controls. For all patients, plasma elastase-complex was related to worsening hypoxia and to increased bronchoalveolar lavage protein content. Elastase-complex was present in all bronchoalveolar lavage samples and was related to protein content and differential neutrophil counts. Patients at risk of or with ARDS had a spectrum of respiratory failure to which intravascular and intra-alveolar neutrophil elastase release and capillary permeability were related. This suggests that the state recognised as ARDS is not a distinct pathophysiological entity.

Acute microvascular effects of the chemotactic peptide N-formyl-methionyl-leucyl-phenylalanine: comparisons with leukotriene B4

Acute effects of the chemotactic peptide N-formyl-methionyl-leucyl-phenylalanine (fMLP) on microvessels of the hamster cheek pouch were studied using intravital and electron microscopy. FMLP (1.0 microM) applied topically to the pouch microvasculature produced transient vasoconstriction of second- and third-order arterioles, which was accompanied by modest narrowing of pericytic venules, leukocyte adherence, diapedesis, and macromolecular extravasation. Contrary to observations of leukotriene B4 (LTB4)-exposed venules in the same animal model, electron micrographs of fMLP-treated venules supported junctional, vesicular, and potential transcellular modes of transendothelial protein transport. Endothelial junctions were not characterized by large gaps, but failure of endothelial processes to effectively bridge diapedesing cells and reestablish junctional contact resulted in substantial deposition of subendothelial reaction product. Venous microvessels 18-21 micron in diameter were particularly susceptible to the deleterious effects of fMLP and revealed the highest percentage of total junctions or endothelial cells labeled by reaction product. Implied barrier dysfunction was more prevalent in fMLP-treated vessels exhibiting adherence and/or diapedesis; however, the presence of barrier defects was not correlated with leukocyte margination and/or emigration (Marg/Emig). The absence of simple correlation, without persuasive evidence for direct vasopermeability effects, recognizes that some Marg/Emig events may not have been present at the time of fixation and suggests the potential for neutrophil-mediated endothelial dysfunction, governed by multiple determinants, is often unrealized.

In vitro effects of endotoxin on bovine and sheep lung microvascular and pulmonary artery endothelial cells

A single infusion of Escherichia coli endotoxin into sheep results in structural evidence of pulmonary endothelial injury, increases in both prostacyclin and prostaglandin E2 (PGE2) in lung lymph, and an increase in pulmonary microvascular permeability. Endotoxin-induced lung endothelial damage can also be induced in vitro, but to date these studies have utilized endothelium from large pulmonary vessels. In the present study, we have grown endothelial cells from peripheral lung vessels of cows and sheep and exposed these microvascular endothelial cells to endotoxin. Controls included lung microvascular endothelium without endotoxin and endothelial cells from bovine and sheep main pulmonary artery with and without addition of endotoxin. We found that endotoxin caused significant increases in release of prostacyclin and PGE2 from both bovine and sheep lung microvascular and pulmonary artery endothelium. Normal bovine and sheep pulmonary artery and bovine lung microvascular endothelium released greater levels of prostacyclin than PGE2 (ng/ng); release of PGE2 from the microvascular cells was greater than from the pulmonary artery endothelium in both species. Exposure of endothelial cells from cow and sheep main pulmonary artery to endotoxin results in endothelial cell retraction and pyknosis, a loss of barrier function, increased release of prostacyclin and PGE2 and eventual cell lysis. In lung microvascular cells, the increases in prostanoids were accompanied by changes in cell shape but occurred in the absence of either detectable alterations in barrier function or cytolysis. Thus, while endotoxin causes alterations to endothelial cells from both large and small pulmonary vessels, the effects are not identical suggesting site specific phenotypic expression of endothelial cells even within a single vessel. To determine whether the response of either the large or small pulmonary vessel endothelial cells in culture mimics most closely the in vivo response of the lung to endotoxin requires further study.

Elastase-like activity in cultured aortic endothelial cells

Cultured porcine aortic endothelial cells were studied for cellular and secreted elastase activity. We describe an activity hydrolyzing the synthetic elastase substrate, succinyl(alanine)3 nitroanilide, but not elastin, which was shown to be membrane located and was not secreted to the culture medium. A different neutral proteinase activity degrading insoluble elastin was demonstrated in the culture medium following its fractionation by gel filtration high performance liquid chromatography (HPLC). Since no elastinolytic activity could be directly detected in the conditioned medium, it is likely that the chromatographic separation removed an endogenous inhibitor.

Hydrogen peroxide stimulates the synthesis of platelet-activating factor by endothelium and induces endothelial cell-dependent neutrophil adhesion

Oxidant-induced damage to the intima of pulmonary and systemic vessels is thought to be an important mechanism of injury in a variety of syndromes of vascular damage. Hydrogen peroxide (H2O2) is an active oxygen metabolite that may induce intimal injury by cytolytic attack or by inducing biochemical and functional alterations in the endothelial cells (EC); however, mechanisms involved in noncytolytic perturbation of EC are largely unknown. We found that H2O2 stimulated the synthesis of platelet-activating factor (PAF) by primary cultures of bovine pulmonary artery endothelium (BPAEC) and by human umbilical vein endothelium (HUVEC). In each cell type the incorporation of [3H]acetate into [3H-acetyl]PAF was concentration- and time-dependent and was temporally dissociated from severe plasma membrane disruption and cytolytic cell injury; the newly synthesized PAF remained associated with the EC. H2O2 caused permeabilization of EC to 45Ca2+ and an increase in intracellular Ca2+, suggesting that a transmembrane Ca2+ flux is the signal that initiates PAF synthesis. H2O2 also induced the endothelial cell-dependent adhesion of neutrophils to HUVEC monolayers. This response was rapid, with an onset within minutes and a subsequent time course that paralleled the time course of PAF accumulation, and was dependent on extracellular Ca2+ but not on de novo protein synthesis. These studies demonstrate that H2O2 can induce two rapid activation responses of endothelium, PAF synthesis and EC-dependent neutrophil adhesion, events that may be important in physiologic and pathologic inflammation.

Reduced free radical activity detected by decreased diene conjugates in insulin-dependent diabetic patients

Free radicals are unstable chemical species which react with and oxidize adjacent molecules, particularly polyunsaturated lipids. The diene-conjugated non-peroxide isomer of linoleic acid (PL-9,11-LA') has been identified as the main diene-conjugated compound in plasma, and is a probable marker of free radical activity. The aim of the current study was to determine whether the level of PL-9,11-LA', measured by HPLC, is altered in insulin-dependent diabetes, and to investigate whether any abnormality demonstrated correlated with microvascular disease in the form of retinopathy. There was no difference in the concentrations of linoleic acid between the diabetic and control groups (422(129) vs 402(81) (SD) mumol l-1). However, the concentration of PL-9,11-LA' was significantly reduced in the diabetic group compared with control group (15.6(6.7) vs 19.3(3.9) mumol l-1, p less than 0.01), with the molar ratio of PL-9,11-LA':linoleic acid x 100 similarly reduced (3.8(1.3) vs 5.0(1.6)%, p less than 0.005). This study does not support the concept that free radicals play a significant role in the development of diabetic vascular disease.

Serine proteinase esterolytic activity as an assay of cytotoxic reactions

A method for assessing cytotoxicity in cell-mediated cytolysis as well as in reactions of natural killing and antibody-dependent cellular cytotoxicity through assaying the total esterolytic activity of serine proteinases (TEASP) is both sensitive and specific. TEASP levels in all cytotoxic tests correlated with the percentage of lysis of 51Cr-labelled target cells. The method permits the differential assessment of endoproteases in both the effector cell and the target cell.

Distribution and function of alveolar cells in multiply injured patients with trauma-induced ARDS

To determine whether alveolar cells are involved in the pathogenesis of adult respiratory distress syndrome (ARDS), we assessed the distribution and function of alveolar cells from 30 polytraumatized patients with trauma-induced respiratory failure, 5 of whom also had lung contusion. Cells were obtained by bronchoalveolar lavage performed daily beginning on the day of trauma and continuing for 14 days. Neutrophils constituted about 60% of lavage cells in ARDS patients with lung contusion 0-2 days after polytrauma and about 50% in ARDS patients without lung contusion. In ARDS patients with lung contusion the neutrophil fraction decreased to 52% 3-6 days after trauma and to 40% 7-14 days after trauma. In patients without lung contusion the neutrophil fraction increased to 77% and then decreased to 60% at these times. Total cell counts in ARDS patients with lung contusion were more than twice as high as in patients without lung contusion 0-2 days after trauma. The difference in total cell counts decreased during days 3-6 and disappeared by day 7. In all patients morphologically altered alveolar cells were observed 4 days and more after trauma. In non-survivors significantly more altered cells were found. The chemiluminescence-response pattern of the alveolar cells was enhanced throughout the study and correlated with the neutrophil fraction (r = 0.6). The neutrophil fraction also correlated with the pulmonary vascular resistance during the first two days after trauma (r = 0.53). We conclude that alveolar cells are involved in the pathogenesis of trauma-induced ARDS and that the alveolar cell distribution is different in patients with and without lung contusion during the development of posttraumatic respiratory failure.

Neutrophil degranulation and increased pulmonary capillary permeability following oesophagectomy: a model of early lung injury in man

In nine subjects undergoing oesophagectomy a transient but significant increase in pulmonary vascular permeability to transferrin was detected in both lungs 8 h after surgery. It coincided with a significant increase in plasma lactoferrin, evidence of intravascular and intra-alveolar neutrophil elastase release and with arterial hypoxaemia. The postoperative state in these patients had intravascular and pulmonary features often associated with the adult respiratory distress syndrome, and might be a useful model for the study of the early inflammatory events underlying lung injury in man.

Corticosteroids inhibit endotoxin-induced lung lymph neutrophil stimulating activity in sheep

The Adult Respiratory Distress Syndrome (ARDS) most frequently is the result of sepsis. Accumulation of neutrophils in lung interstitium is a well-documented phenomenon, but the nature of their presence remains obscure. We hypothesized that endotoxin causes the release of substances into lung lymph that activate neutrophils and that methylprednisolone may prevent sequestration and activation of neutrophils. We used the sheep lung lymph fistula-endotoxin model of ARDS to test this hypothesis. Unanesthetized animals were given either 0.5 microgram/kg of E. coli endotoxin intravenously alone or, on a different experimental day, an identical dose of endotoxin preceded by a 1 gm bolus of methylprednisolone plus a 1 gm/hr continuous infusion. Endotoxin infusion caused the release of substances into lung lymph that were capable of stimulating normal sheep neutrophils to aggregate, migrate, and release superoxide. This activity appeared within 1 hour of endotoxin and persisted for at least 4 hours. Pretreatment by methylprednisolone did not prevent the early activity but did significantly reduce such activity 3-4 hours after endotoxin, when the permeability defects caused by endotoxin are most pronounced. We speculate that endotoxin-stimulated production of humoral neutrophil-activating substances in the lung may play a role in the pathogenesis of acute lung injury.

Microbicidal/cytotoxic proteins of neutrophils are deficient in two disorders: Chediak-Higashi syndrome and "specific" granule deficiency

Although several genetic defects are known to impair oxidative microbicidal/cytotoxic mechanisms in human PMN, no deficiencies of PMN granule components that mediate oxygen-independent microbicidal activity have yet been reported. We analyzed PMN from patients with various granulocyte disorders for their content of two azurophil granule constituents, defensins and cathepsin G, that exert microbicidal/cytotoxic activity in vitro, and one component, elastase, that has ancillary microbicidal/cytotoxic activity. PMN from two (of two) patients with specific granule deficiency (SGD) displayed an almost complete deficiency of defensins, which in normal cells constitute greater than 30% of the protein content of azurophil granules. The SGD PMN contained normal or mildly decreased amounts of cathepsin G and elastase. Conversely, the PMN of three (of three) patients with Chediak-Higashi syndrome (CHS) substantially lacked cathepsin G and elastase, but their defensin content was normal or mildly decreased. Both CHS and SGD patients suffer from frequent and severe bacterial infections, and CHS patients frequently develop an atypical lymphoproliferative syndrome. The profound deficiency of PMN components with microbicidal/cytotoxic activity in SGD and CHS may contribute to the clinical manifestations of these disorders.

Allopurinol and lodoxamide in complement-induced hepatic ischemia

Intravascular complement activation with either zymosan or cobra venom factor (CVF) impairs hepatic blood flow. Oxygen radical scavengers given at the time of complement activation attenuate the resulting hepatic ischemia. It is not clear whether complement-stimulated phagocytes or transiently ischemic then reperfused endothelial and parenchymal cells generated the toxic oxygen radicals. In this study, a group of rats were given allopurinol (50 mg/kg/day postoperatively X 3 days plus 100 mg/kg iv at t = 0), a specific inhibitor of xanthine oxidase, prior to complement activation with CVF (20 units/kg iv at t = 30 and 60 min) to determine whether xanthine oxidase-derived oxygen radicals contributed significantly to the hepatic perfusion abnormalities. Additional rats received lodoxamide tromethamine (10 mg/kg iv bolus at t = 0 followed by 20 mg/kg/hr iv infusion), a novel and potent inhibitor of mast cell release and inhibitor of xanthine oxidase, prior to the same CVF challenge to determine whether mast cell mediators were involved in the flow disturbance. Thermodilution cardiac output, mean arterial pressure, heart rate, hematocrit, and effective hepatic blood flow (EHBF) by galactose clearance were determined at t = 2 hr. The percentage change in total hemolytic complement activity (% delta CH50) was determined between serum obtained prior to sacrifice and at t = 0. Systemic hemodynamics and HCT were for the most part unaffected regardless of pretreatment group or challenge with CVF or saline. CVF challenge produced a 25% reduction (P less than 0.05) in EHBF in vehicle-pretreated rats compared to saline challenge. Neither allopurinol nor lodoxamide tromethamine significantly improved EHBF when given prior to CVF challenge.(ABSTRACT TRUNCATED AT 250 WORDS)

Synergistic cytolysis mediated by hydrogen peroxide combined with peptide defensins

Possible cytolytic interactions between hydrogen peroxide (H2O2) and neutrophil granule proteins were studied. Preliminary experiments demonstrated synergistic cytolysis when erythro-leukemia targets were exposed to H2O2 combined with a low molecular weight (approximately 3900) granule extract that was predominantly composed of peptide defensins. The synergistic interaction was confirmed when sublytic concentrations of H2O2 were combined with defensin preparations that had been purified to homogeneity. Synergy was concentration dependent in regard to both molecules and could not be explained by trace contamination of defensin preparations with myeloperoxidase. Sequential addition experiments suggested that synergistic lysis required a simultaneous exposure to both cytotoxins. In the presence of sublytic concentrations of H2O2, the binding of iodinated defensin to targets was significantly increased, providing a possible explanation for the observed synergy. Since both molecules are concurrently secreted by activated neutrophils, this interaction may be important during leukocyte-mediated anti-tumor effects or inflammatory tissue injury.

Tumor necrosis factor causes increased pulmonary permeability and edema. Comparison to septic acute lung injury

Tumor necrosis factor alpha (TNF), a monokine produced by mononuclear cells in response to bacterial endotoxin (LPS), creates a syndrome similar to septic shock in animal models. To study whether TNF could induce acute lung injury similar to that seen in gram-negative sepsis, we injected recombinant human TNF (rHuTNF alpha) into guinea pigs and monitored arterial blood gases, leukocyte counts, and left atrial (Pla), pulmonary artery (Ppa), and mean arterial pressures (MAP) serially for 8 h. Pulmonary histopathology was assessed microscopically, and cell counts and 125I-labeled albumin (125I-albumin) in bronchoalveolar lavage (BAL) fluid and lung wet/dry weight ratios were determined. Five groups of animals were studied; the 2 TNF groups received high (1.4 X 10(6) U/kg) or low (1.0 X 10(6) U/kg) doses of rHuTNF alpha, the sepsis group received 2 X 10(9) Escherichia coli/kg intravenously, and the control group received saline. An LPS control group receiving 40 ng/kg E. coli LPS was also included because the rHuTNF alpha contained a small amount of LPS as a contaminant. Pulmonary permeability was assessed by studying the Pla and the BAL fluid/plasma 125I-albumin ratio (permeability index). The permeability index was significantly increased in the high-dose TNF (0.0408 +/- 0.0041, p less than 0.05) and sepsis groups (0.0466 +/- 0.0068, p less than 0.01) relative to controls (0.0215 +/- 0.0028). The wet/dry lung weight ratios were also significantly increased in the high-dose TNF (6.07 +/- 0.29, p less than 0.05) and sepsis groups (6.22 +/- 0.30, p less than 0.05) relative to the control group (5.18 +/- 0.20).(ABSTRACT TRUNCATED AT 250 WORDS)

In vitro toxicity of polymorphonuclear neutrophils to rat hepatocytes: evidence for a proteinase-mediated mechanism

Human polymorphonuclear neutrophils, when exposed to soluble or particulate stimuli, can destroy various types of cells. The aim of this study was to investigate their toxicity against hepatocytes. Human polymorphonuclear neutrophils were incubated in basal conditions and after stimulation with 5 mg per ml opsonized zymosan in the presence of rat hepatocytes isolated by collagenase digestion. Cytotoxicity was quantified by the percentage of ALT activity released by hepatocytes in culture medium. Whereas unstimulated neutrophils exhibited only minor effects, opsonized zymosan-stimulated neutrophils induced, after 16 hr incubation, a 24.0 +/- 4.1% (mean +/- 1 S.E.) ALT activity release at a neutrophil/hepatocyte ratio of 5, and a 51.7 +/- 6.8% ALT activity release at a ratio of 20. At this ratio of 20, the ALT activity release was 9.0% at 1 hr and 24.0% at 4 hr. Three proteinase inhibitors (i.e., soybean trypsin inhibitor, alpha 1-proteinase inhibitor and fetal calf serum) decrease cytotoxicity by 78, 76 and 78%, respectively. The protective effect of proteinase inhibitors was not due to a nonspecific effect of proteins, since bovine serum albumin did not decrease the toxicity of stimulated polymorphonuclear cells. The supernatant of stimulated neutrophils was also found to be toxic against hepatocytes, and again, this effect was inhibited by soybean trypsin inhibitor, alpha 1-proteinase inhibitor and fetal calf serum. Finally, the role of proteinases was supported by the demonstration of a cytotoxic effect of two purified proteinases: porcine pancreatic elastase and human neutrophil cathepsin G. The toxicity of these proteinases was also markedly reduced by the specific inhibitors used in the study.(ABSTRACT TRUNCATED AT 250 WORDS)

Lipopolysaccharide-dependent enhancement of adherence-mediated chemiluminescence response of polymorphonuclear leukocytes

Adherence of resting polymorphonuclear leukocytes to nylon fibre increased the chemiluminescence response (CL) from 99,400 to 910,300 cpm/25,000 PMNL. This effect could be amplified by lipopolysaccharide (LPS) priming of granulocytes in a dose-dependent fashion. The results of nylon fibre adherence experiments suggest an in vitro model that might approximate certain conditions of in vivo PMNL-endothelial adherence and respiratory burst activation, and these reactions of polymorphonuclear leukocytes may contribute to the pathomechanisms of the Adult Respiratory Distress Syndrome.

Oxygen radicals, inflammation, and tissue injury

Inflammatory reactions often result in the activation and recruitment of phagocytic cells (e.g., neutrophils and/or tissue macrophages) whose products result in injury to the tissue. In killing of endothelial cells by activated neutrophils as well as in lung injury produced by either activated neutrophils or activated macrophages there is evidence that H2O2 and iron play a role. HO. may be a key oxygen product related to the process of injury. Endothelial cells in some vascular compartments may be susceptible to neutrophil mediated injury in a manner that is independent of oxygen radicals. On the basis of in vitro observations, a synergy exits between platelets and neutrophils, resulting in enhanced oxygen radical formation by the latter. Finally, the cytokines, interleukin 1 and tumor necrosis factor, released from macrophages have both direct stimulatory effects on oxygen radical formation in neutrophils and can "prime" macrophages for enhanced oxygen radical responses to other agonists. Cytokines may also alter endothelial cells rendering them more susceptible to oxygen radical mediated injury by neutrophils. This suggests a complex network of interactions between phagocytic cells and peptide mediators, the result of which is acute, oxygen radical mediated tissue injury.

Modulation of human neutrophil and monocyte oxidative burst by Legionella pneumophila sonic extract

The effect of Legionella pneumophila sonic extract on human neutrophil and monocyte oxidative burst was studied by superoxide anion release and luminol-enhanced chemiluminescence assays. Legionella pneumophila sonic extract by itself did not stimulate neutrophils and monocytes. The sonic extract at 8-2000 micrograms/ml primed neutrophils for enhanced superoxide release and, at 8-62.5 micrograms/ml, for enhanced chemiluminescence. Monocytes were only primed for enhanced chemiluminescence at very low extract concentrations (below 16 micrograms/ml). Monocyte superoxide release was suppressed by extract concentrations higher than 2000 micrograms/ml and the chemiluminescence response of neutrophils and monocytes by concentrations higher than 250 and 125 micrograms/ml, respectively. The priming activity was heat stable and present in fractions below 5 kDa. On the basis of these findings it is suggested that enhanced production of oxygen metabolites by neutrophils in contact with legionella components at low concentrations could contribute to the lung tissue damage seen in Legionnaires' disease, whereas the suppression of phagocyte oxidative burst by higher extract concentrations may be one of the mechanisms by which Legionella pneumophila survives intracellularly.

Neutrophil activation on biological surfaces. Massive secretion of hydrogen peroxide in response to products of macrophages and lymphocytes

Recombinant tumor necrosis factor alpha (rTNF alpha) and beta (rTNF beta) did not trigger H2O2 release from PMN in suspension. However, when PMN were plated on polystyrene surfaces coated with serum, fibronectin, vitronectin, laminin, or human umbilical vein endothelial cells (HUVEC), rTNFs induced a massive, prolonged secretory response, similar to that elicited by phorbol myristate acetate (PMA) or bacteria. On serum-coated plates, the maximum sustained rate of H2O2 release in response to rTNF alpha was 2.6 +/- 0.2 nmol/min per 10(6) PMN, the same as that with PMA; release continued for 73 +/- 4 min. On laminin-coated surfaces or HUVEC, release of H2O2 in response to rTNFs was slower, but lasted approximately 3.5 h, reaching the same total (greater than 100 nmol/10(6) PMN). Not only was this response far longer and larger than for other soluble stimuli of the respiratory burst studied with PMN in suspension, but the concentration necessary to elicit a half-maximal response (EC50) for rTNF alpha was orders of magnitude lower (55 pM). Responses were similar with FMLP, but ranged from zero to small with recombinant IFN alpha, recombinant IFN beta, recombinant IFN gamma, platelet-derived growth factor, recombinant IL-1 beta, or bacterial lipopolysaccharide. Adherent monocytes did not secrete H2O2 in response to rTNFs. H2O2 secretion by adherent PMN was first detectable 15-90 min after addition of rTNFs or FMLP. This lag period was unaffected by prior exposure of PMN to rTNF alpha in suspension, by allowing PMN to adhere before adding rTNF alpha, or by incubating adherent PMN in medium conditioned by rTNF alpha-treated PMN. Cytochalasins abolished H2O2 secretion in response to rTNFs, but not FMLP, if added during, but not after, the lag period. Thus, H2O2 secretion from rTNF alpha-treated PMN appears to be a direct but delayed response that requires assembly of microfilaments during exposure to the cytokine. These results suggest that PMN adherent to intra- or extravascular surfaces may undergo a massive, prolonged respiratory burst at the command of macrophages and lymphocytes reacting to microbial products and antigens.

Effects of aprotinin on hemostatic mechanisms during cardiopulmonary bypass

Cardiopulmonary bypass (CPB) is associated with activation of humoral systems, which results in the release of proteases. These proteases may affect platelets and stimulate granulocytes. In the present study, the protease inhibitor aprotinin was given in high doses to 11 patients to achieve plasma concentrations of more than 150 kallikrein inactivator units per milliliter during CPB. At such concentrations, kallikrein and plasmin are effectively inhibited. This treatment resulted in platelet preservation during CPB. Platelet numbers were virtually unaffected, and thromboxane release was prevented in the aprotinin-treated group in contrast to the control group. Postoperatively, hemostasis was significantly better preserved after aprotinin treatment (blood loss of 357 ml in the treated group versus 674 ml in the untreated group; p less than 0.01). Since tissue-plasminogen activator activity was similar in both groups, the improved hemostasis most likely should be attributed to platelet preservation. Furthermore, aprotinin lessened neutrophilic elastase release, which might contribute to decreased pulmonary dysfunction in patients at risk.

The effect of neutrophil migration and prolonged neutrophil contact on epithelial permeability

The effect of neutrophil migration and prolonged neutrophil contact on epithelial permeability was examined. Although neutrophil migration was not associated with a change in epithelial permeability, prolonged neutrophil-epithelial contact following migration resulted in an increase in epithelial permeability. These results were not altered by catalase, a specific neutrophil elastase inhibitor, methoxysuccinyl-Ala-Ala-Pro-Val-chloromethyl ketone or cyclohexamide. This suggests that neutrophil migration does not occur via an H2O2-induced reversible mechanism of junctional opening, which we describe herein.

Reduction of reperfusion injury in the canine preparation by intracoronary adenosine: importance of the endothelium and the no-reflow phenomenon

We hypothesized that the endogenous coronary vasodilator adenosine may reduce infarct size by progressively increasing reflow in a preparation of coronary occlusion-reperfusion. After 90 min of proximal left anterior descending artery occlusion, 20 dogs were randomized to blood reperfusion with (n = 10) or without (n = 10) adenosine into the proximal left anterior descending vessel at 3.75 mg/min for 60 min after reperfusion. Regional myocardial blood flow was determined serially with microspheres and regional ventricular function was assessed by a computerized radial shortening method. At 24 hr, the area at risk was defined in vivo with monastral blue dye and area of necrosis was determined after incubation of left ventricular slices in triphenyltetrazolium chloride. Hemodynamic variables were similar in the two groups during the experimental protocol. Infarct size was significantly reduced in treated animals, both when expressed as a percentage of the area at risk (9.9 +/- 2.8% vs 40.9 +/- 6.6%, p less than .001) and as a percentage of the left ventricle (4.6 +/- 1.3% vs 18.0 +/- 3.4%, p = .002). This was associated with significant improvement in radial shortening in the ischemic zone 24 hr after reperfusion (10.1 +/- 2.5 vs -2.8 +/- 2.2%, p less than .01). Regional myocardial blood flow was significantly increased in endocardial and epicardial regions from the lateral ischemic zone 1 hr after reperfusion in adenosine-treated animals. Light microscopy demonstrated decreased neutrophil infiltration in the ischemic zone and electron microscopy showed relative preservation of endothelial structure in the subendocardium with reduced neutrophil and red cell stagnation of capillaries in the treated group. These findings suggest that intracoronary administration of adenosine after reperfusion significantly reduces infarct size and improves regional ventricular function in the ischemic zone in the canine preparation.

Inhibition of endotoxin-induced priming of human neutrophils by lipid X and 3-Aza-lipid X

Lipid X, a precursor of lipid A (the toxic moiety of endotoxin), has been shown to protect animals from the lethal effects of endotoxin challenge. We investigated the mechanism of action of lipid X and 3-aza-lipid X, a diamino-analogue, in vitro, using the ability of lipopolysaccharide (LPS) to prime neutrophils for an enhanced release of toxic oxygen radicals. Lipid X and 3-aza-lipid X inhibited LPS-induced neutrophil priming in a concentration-dependent manner. At high concentrations, 3-aza-lipid X was a partial agonist of priming. Lipid X was found to inhibit LPS-induced priming by directly interacting with the neutrophil in contrast to polymyxin B, which neutralized LPS by binding to it. Increasing concentrations of lipid X shifted the LPS dose response curve of neutrophils rightward but did not prevent maximum priming at higher LPS concentrations, a finding consistent with competitive inhibition. These results suggest that lipid X, a compound structurally related to lipid A, may block neutrophil priming by competing with LPS for cellular binding sites. Lipid X appears to have a novel mechanism of inhibiting LPS effect and may have efficacy in the treatment of gram-negative sepsis.

Endotoxin alters arachidonate metabolism in pulmonary endothelial cells

Endotoxin and lipid A dose dependently (1 ng/ml to 10 micrograms/ml) and time dependently (6-24 h) stimulated the generation of large amounts of prostacyclin in cultured pig pulmonary artery endothelial cells. This effect occurred in the absence of cell detachment and overt cell damage. The presence of at least 1% serum was required but the activation of the complement cascade was not. Endotoxin-treated endothelial cells generated increased amounts of prostacyclin upon stimulation with A23187 and arachidonic acid. Endotoxin-induced activation of arachidonate metabolism could be reduced by 10(-10) M glucocorticoids but not by progesterone. It was further affected by inhibitors of protein and RNA synthesis and calmodulin function. In addition, exposure of endothelial cells to endotoxin resulted in an enhanced synthesis of cyclooxygenase and in a higher enzymatic capacity of prostacyclin synthase. The data indicate that endotoxin in concentrations occurring in the plasma of patients profoundly alters arachidonic acid metabolism in endothelial cells.

Preservation of endothelial cell structure and function by intracoronary perfluorochemical in a canine preparation of reperfusion

To determine the effect of intracoronary perfluorochemical on endothelial cell structure and function, 16 dogs were randomized to receive either low-dose (15 ml/kg) intracoronary perfluorochemical (Fluosol-DA) or saline after 90 min of proximal occlusion of the left anterior descending coronary artery (LAD). The animals underwent reperfusion for 60 min with the introduction of perfluorochemical or saline 5 to 10 min after the onset of reperfusion. Endothelium-dependent coronary vasodilatory reserve was determined in vivo both at baseline and 1 hr after reperfusion by infusion of acetylcholine and then serotonin into the distal LAD bed in 12 animals (six in each group). Both agonists significantly increased regional flow measured by 133Xe washout in the two groups before occlusion, but at 1 hr after reperfusion only animals given perfluorochemical demonstrated a significant increase in flow. Vasodilatory reserve was assessed in vitro with cumulative dose-response curves to acetylcholine on LAD rings proximal and distal to the snare in all animals. These studies demonstrated a significant reduction in endothelial cell-mediated relaxation of epicardial arterial segments in the ischemic segment of control but not treated animals. Light microscopy revealed the presence of neutrophils within vessels in the ischemic zones in control animals only. Electron microscopy showed capillary obstruction by endothelial cell protrusions and neutrophil and red cell plugging in control animals in the ischemic region but an intact endothelium and predominantly unobstructed capillaries in treated animals. These findings suggest that the structural and functional endothelial damage after reperfusion may be prevented by the administration of intracoronary perfluorochemical after the onset of reperfusion.

Differences and similarities between human and rabbit neutrophil granulocyte responses in vitro: the effects of zymosan-activated plasma, phorbol myristate acetate and n-formyl-methionyl-leucyl-phenylalanine

The amount of reactive oxygen intermediates (ROI) generated by activated polymorphonuclear neutrophils (PMN), as well as the closeness of contact between PMN and vessel wall, may determine whether PMN activators will induce the adult respiratory distress syndrome. We examined the ROI-generating and aggregating effects of zymosan activated plasma (ZAP), phorbol myristate acetate (PMA) and n-formyl-methionyl-leucyl-phenylalanine (FMLP), on isolated human and rabbit PMN. PMA, after a short lag phase, induced a large and long-lasting increase in ROI generation. The initial peak response was higher and more rapid in human than in rabbit cells. The reaction to FMLP occurred almost instantaneously, but was much weaker than that to PMA, and ROI generation returned to near baseline in less than 10 min. No species difference was seen. ZAP caused an FMLP-like ROI response in human cells, whereas no response was observed in rabbit PMN. PMN aggregation was induced by all three activators, most markedly by PMA. No species difference was detected for PMA; FMLP gave a stronger aggregation of rabbit than of human PMN, however, while the opposite was true for ZAP. In conclusion, ZAP was a potent stimulus for PMN aggregation, but had modest (or no) effects on the production of ROI. Marked differences between human and rabbit PMN responses were observed.

The pulmonary vascular sequestration of neutrophils in endotoxemia is initiated by an effect of endotoxin on the neutrophil in the rabbit

Endotoxemia causes neutrophil sequestration in the pulmonary vascular bed. Such sequestration may be a critical initiating event in the generation of microvascular injury, although the mechanisms that lead to this localization are not understood. To investigate these phenomena, the following study employed intravenous pulses of 111Indium-tropolonate-labeled neutrophils (111In-neutrophils), which circulated in the rabbit with normal kinetics and responded in a manner indistinguishable from unlabeled, circulating neutrophils in response to an intravenous injection of purified endotoxic lipopolysaccharide (LPS) or epinephrine. Pulmonary sequestration of 111In-neutrophils was assessed by quantitative external gamma camera scintigraphy of a lung suprahilar region of interest. Noninvasive assessment of radioactivity by this method accurately reflected total lung radioactivity, which was shown by autoradiography to be confined to the injected 111In-neutrophils. Intravenously administered LPS caused a marked, dose-dependent sequestration of 111In-neutrophils in the pulmonary vasculature, and exhaustive ultrastructural autoradiography showed discretely radiolabeled neutrophils located within pulmonary capillaries. A distinct effect was seen with an intravenous injection of as little as 100 ng per rabbit (i.e., 500 pg/ml blood). A 5-min ex vivo pretreatment of 111In-neutrophils with 10 ng to 10 micrograms/ml LPS in heat-inactivated plasma (which resulted in retention of as little as 500 pg LPS per 10(7) neutrophils) also caused dose-dependent pulmonary sequestration of the pretreated 111In-neutrophils but did not cause generalized neutropenia in recipient rabbits. There was no evidence of complement activation on the surface of pretreated neutrophils.(ABSTRACT TRUNCATED AT 250 WORDS)

Microvascular injury and repair in acute human bacterial pyelonephritis

Acute inflammatory cell-capillary endothelial cell interactions, related to injury and repair, were investigated light and electron microscopically in acute human bacterial pyelonephritis. In inflammatory infiltrate-adjacent microvessels, the small capillaries were completely occluded by leukocyte plugs and the large capillaries were densely filled with acute inflammatory cells adhering to the endothelium. Severe damage to small and large capillaries was observed around endothelium adherent, degranulated neutrophil granulocytes containing phagocytosed bacteria. There were spaces in the endothelium, degradation of the vascular basement membrane, of the perivascular interstitial matrix and of collagen fibrils, with fibrin deposition and vessel wall fragmentation. In the small capillaries relatively distant from the interstitial infiltrates, emigration of leukocytes was frequently seen. Around the escaping cells the endothelial lining displayed occasional discontinuities, allowing leakage of vascular fluid into the interstitial space. Some small capillaries not related to the infiltrate were occluded by fibrin thrombi with apparent damage to the endothelial cells and disruption of the capillary wall. Various reparative changes were noticed in association with this change including capillary neovascularization. The findings confirm the existence of polymorphonuclear leukocyte-mediated injury of capillaries during the development of inflammatory responses in acute pyelonephritis.

Endotoxin enhances tissue factor and suppresses thrombomodulin expression of human vascular endothelium in vitro

Endotoxemia is frequently associated clinically with disseminated intravascular coagulation (DIC); however, the mechanism of endotoxin action in vivo is unclear. Modulation of tissue factor (TF) and thrombomodulin (TM) expression on the endothelial surface may be relevant pathophysiologic mechanisms. Stimulation of human umbilical vein endothelial cells with endotoxin (1 microgram/ml) increased surface TF activity from 1.52 +/- 0.84 to 11.89 +/- 8.12 mU/ml-10(6) cells at 6 h (n = 11) which returned to baseline by 24 h. Repeated stimulation at 24 h resulted in renewed TF expression. Endotoxin (1 microgram/ml) also caused a decrease in TM expression to 55.0 +/- 6.4% of control levels at 24 h (n = 10) that remained depressed at 48 h. Both effects were dose and serum dependent. A temporary rise in TF expression accompanied by a sustained fall in TM expression comprise a shift in the hemostatic properties of the endothelium that would favor intravascular coagulation and may contribute to the pathogenesis of DIC in gram-negative septicemia.

Neutrophil-induced injury of rat pulmonary alveolar epithelial cells

The damage to pulmonary alveolar epithelial cells that occurs in many inflammatory conditions is thought to be caused in part by phagocytic neutrophils. To investigate this process, we exposed monolayers of purified rat alveolar epithelial cells to stimulated human neutrophils and measured cytotoxicity using a 51Cr-release assay. We found that stimulated neutrophils killed epithelial cells by a process that did not require neutrophil-generated reactive oxygen metabolites. Pretreatment of neutrophils with an antibody (anti-Mo1) that reduced neutrophil adherence to epithelial cells limited killing. Although a variety of serine protease inhibitors partially inhibited cytotoxicity, we found that neutrophil cytoplasts, neutrophil lysates, neutrophil-conditioned medium, purified azurophilic or specific granule contents, and purified human neutrophil elastase did not duplicate the injury. We conclude that stimulated neutrophils can kill alveolar epithelial cells in an oxygen metabolite-independent manner. Tight adherence of stimulated neutrophils to epithelial cell monolayers appears to promote epithelial cell killing.

Anthrax toxin blocks priming of neutrophils by lipopolysaccharide and by muramyl dipeptide

We studied the pretreatment of human polymorphonuclear neutrophils (PMN) with purified preparations of the anthrax toxin components--protective antigen (PA), edema factor (EF), and lethal factor (LF)--and their effects on release of superoxide anion (O-2) after stimulation with the chemotactic peptide N-formyl-methionyl-leucyl-phenylalanine (FMLP). PMN isolated in the absence of lipopolysaccharide (LPS) (less than 0.1 ng/ml) released only small amounts of O-2 after FMLP stimulation; pretreatment with anthrax toxin had little effect. The release of O-2 was increased fivefold by prior treatment with 3 ng/ml LPS for 1 h at 37 degrees C, an effect referred to as priming. PMN were primed to an equivalent extent by treatment with 100 ng/ml N-acetyl-muramyl-L-alanyl-D-isoglutamine (muramyl dipeptide [MDP]). Pretreatment of PMN with anthrax toxin components PA plus EF or PA plus LF inhibited priming by LPS or MDP, as shown by the reduction in the release of O-2 up to 90% relative to controls not treated with toxin; single toxin components were inactive. The inhibition was markedly reduced when priming with LPS or MDP was carried out before exposure to toxin. O-2 release after stimulation by phorbol myristate acetate was not increased by priming, and pretreatment with toxin did not inhibit O-2 release after this stimulus. Evidently, anthrax toxin inhibits the priming that is normally induced in PMN by bacterial products and is necessary for the full expression of antimicrobial effects.

Adenosine: an endogenous inhibitor of neutrophil-mediated injury to endothelial cells

Since adenosine and its analogue 2-chloroadenosine prevent neutrophils from generating superoxide anion in response to chemoattractants, we sought to determine whether these agents could inhibit neutrophil-mediated injury of endothelial cells. The chemoattractant N-formyl-methionyl-leucyl-phenylalanine (FMLP, 0.1 microM) enhanced the adherence of neutrophils to endothelial cells twofold (18 +/- 2% vs. 39 +/- 3% adherence, P less than 0.001) and caused substantial neutrophil-mediated injury to endothelial cells (2 +/- 2% vs. 39 +/- 4% cytotoxicity, P less than 0.001). 2-Chloroadenosine (10 microM) not only inhibited the adherence of stimulated neutrophils by 60% (24 +/- 2% adherence, P less than 0.001) but also diminished the cytotoxicity by 51% (20 +/- 4% cytotoxicity, P less than 0.002). Furthermore, depletion of endogenously released adenosine from the medium by adenosine deaminase-enhanced injury to endothelial cells by stimulated neutrophils (from 39 +/- 4% to 69 +/- 3% cytotoxicity, P less than 0.001). Indeed, in the presence of adenosine deaminase, even unstimulated neutrophils injured endothelial cells (19 +/- 4% vs. 2 +/- 2% cytotoxicity, P less than 0.001). These data indicate that engagement of adenosine receptors prevents both the adhesion of neutrophils and the injury they cause to endothelial cells. Adenosine inhibits injury provoked not only by cells that have been stimulated by chemoattractants but also by unstimulated cells. Based on this model of acute vascular damage we suggest that adenosine is not only a potent vasodilator, but plays the additional role of protecting vascular endothelium from damage by neutrophils.