A single amino acid substitution changes ribonuclease 4 from a uridine-specific to a cytidine-specific enzyme.

The structural features underlying the strong uridine specificity of ribonuclease 4 (RNase 4) are largely unknown. It has been hypothesized that the negatively charged alpha-carboxylate is close to the pyrimidine binding pocket, due to a unique C-terminal deletion. This would suppress the cleavage of cytidine-containing substrates [Zhou, H.-M., and Strydom, D. J. (1993) Eur. J. Biochem. 217, 401-410]. Replacement of the alpha-carboxylate by an alpha-carboxamide in a fragment complementation system decreased both (kcat/Km)CpA and (kcat/Km)UpA , thus refuting the hypothesis. However, model building showed that the deletion allowed the side chain of Arg-101 to reach the pyrimidine binding pocket. From the 386-fold reduction in (kcat/Km)UpA in RNase 4;R101N, it is concluded that this residue functions in uridine binding, analogous to Ser-123 in RNase A. In addition, it may have an effect on Asp-80. The 2-fold increase in (kcat/Km)CpA in the mutant R101N and the close proximity of the side chains of Arg-101 and Asp-80 suggested that the latter could be involved in suppressing CpA catalysis. The substrate specificity of RNase 4;D80A was completely reversed: (kcat/Km)UpA decreased 159-fold, whereas (kcat/Km)CpA increased 233-fold. The effect on CpA was unexpected, because the corresponding residue in bovine pancreatic RNase A (Asp-83) hardly affects cytidine-containing substrates. Furthermore, the residue is conserved in nearly all sequences of mammalian RNase 1. Thus, an evolutionary highly conserved residue does not necessarily function in the same way in homologous enzymes. A model, which proposes that the structure of RNase 4 has been optimized to permanently fix the position of Asp-80 and impede its movement away from the pyrimidine binding pocket, is presented.

Mechanism of tissue factor activation on HL-60 cells.

Tissue factor (TF) procoagulant activity (PCA) on the surface of intact HL-60 cells is encrypted. This latent TF PCA was activated by exposing the cells to ionomycin, a calcium ionophore. Within seconds an increase in TF PCA of greater than 100-fold was observed. The ionomycin effect was blocked by pretreating the cells with calmidazolium, a calmodulin inhibitor. Changes in TF structure and function, coincident with the ionophore-induced increase in TF PCA, were identified. TF-factor VIIa complexes formed on both untreated and ionophore-treated cells, but pseudosubstrate inhibitors only bound to TF-factor VIIa on the ionophore-treated cells. TF PCA was inhibited by reacting cells with sulfosuccinimidyl-6-(biotinamido)hexanoate, and the rate of this reaction increased twofold after cells were exposed to ionomycin. When proteins on the surface of untreated cells, expressing minimal TF PCA, were cross-linked with 3-3'-dithiobis(sulfosuccinimidylpropionate), cross-linked TF dimers were produced. TF cross-linking was prevented by first treating the cells with ionomycin. These results suggest a mechanism for the ionomycin-induced increase in TF PCA. TF activation appears to be a calmodulin-dependent process, which exposes an essential macromolecular substrate binding site on TF, possibly as the result of a change in TF quaternary structure.

Apoptosis is associated with increased cell surface tissue factor procoagulant activity.

Tissue factor (TF), the physiologic initiator of coagulation, is present on the surface of cells but is not fully active unless the cell is lysed. This phenomenon, termed TF encryption, may be regulated by changes in membrane structure. Because apoptosis is associated with cell membrane alterations and conditions associated with apoptosis have also been associated with TF de-encryption, we hypothesized that apoptosis would result in enhanced TF procoagulant activity. Cultured human fibroblasts and endotoxin-stimulated endothelial cells were treated to induce apoptosis as evidenced by morphologic and DNA changes. Under the same conditions, changes in the level of TF activity were measured. Conditions that resulted in endothelial apoptosis were associated with de-encryption of TF activity. Similar results were obtained in fibroblasts except that only the morphologic changes, not the alterations in DNA size characteristic of apoptosis in other cells, were found. The data suggest an association between apoptosis and expression of cell surface tissue factor activity. Because of the recognized linkage of the coagulation system with wound healing and neoplasia, we speculate that this association may help to regulate the flux of cells in tissues being remodelled by apoptosis.

Training internal medicine residents in the community: the Minnesota experience.

The University of Minnesota Internal Medicine Residency Program has developed ambulatory general medicine rotations in rural community settings and urban managed care settings in Minnesota. Based on what had been learned from community focus groups, from discussions with residents about what they perceived to be training holes in the traditional curriculum, and from resident evaluations of pilot rotations, an educational framework for the rotations was established. The authors describe the process of developing these rotations, their educational rationale and objectives, the structure of the rotations, teaching strategies, faculty development, and the evaluation system, and outline the important elements of the successful implementation of these new rotations. The community practice setting has offered a unique, important, and previously untapped resource for residency training, and the community rotations have been highly valued both by housestaff and by community preceptors. As residency programs begin to offer more community-based ambulatory care opportunities for their trainees, the impact of this trend on quality of training, residents' career choices, and patient outcomes will need to be evaluated.

Soluble CD14 promotes LPS activation of CD14-deficient PNH monocytes and endothelial cells.

Bacterial lipopolysaccharide (LPS) initiates the cascade of inflammatory events that, in infected patients, often result in a lethal systemic inflammatory response known as the sepsis syndrome. We studied LPS-stimulated expression of tissue factor (TF) in human peripheral blood mononuclear cells (PBMCs) and cultured endothelial cells or tumor necrosis factor-alpha (TNF-alpha) in PBMCs. CD14, a PBMC membrane protein, is involved in LPS signaling and is also present as a soluble molecule in serum. CD14 is absent from endothelial cells and, in varying degrees, from monocytes of patients with paroxysmal nocturnal hemoglobinuria (PNH). LPS stimulation of TF in normal monocytes was enhanced > 30-fold by serum at low concentrations of LPS (< or = 10 ng/ml). The serum dependence of endothelial cells was even more pronounced; a full response to LPS was not observed in endothelium under serum-free conditions, even with LPS concentrations as high as 100 ng/ml. To better define the role of CD14, CD14-deficient PBMCs from two patients with PNH were compared with normal PBMCs. Although less than 3% of PNH monocytes expressed CD14, LPS-induced synthesis of TF and TNF-alpha by PBMCs from PNH patients was inhibited by anti-CD14 antibodies. Because patient serum samples were found to contain soluble CD14, we sought to determine whether PNH monocytes might respond to LPS through an activation pathway dependent on soluble CD14. Recombinant soluble CD14 substituted for serum to enable LPS stimulation of endothelium, PNH PBMCs, and surprisingly, CD14-replete normal PBMCs. In addition, a truncated sCD14 containing the N-terminal 152 amino acids similarly enabled LPS stimulation of normal PBMCs. These data underscore the importance of soluble CD14 and suggest that CD14 present in serum enables LPS responses in PNH monocytes and endothelial cells and may even influence the effects of LPS in normal human phagocytes.

Induction of endothelial tissue factor by endotoxin and its precursors.

The structural determinants of lipopolysaccharide required for the induction of tissue factor in human umbilical vein endothelial cells were studied. Intact lipid A was essential for the induction of tissue factor whereas the incomplete lipid A precursors lipid IVA and lipid X, as well as monophosphoryl lipid A and acyloxyacyl hydrolase-treated lipopolysaccharide, were unable to induce tissue factor and tissue factor specific mRNA. However, the lipid A precursor, lipid IVA, was able to inhibit LPS-mediated induction of tissue factor; structural determinants distal to lipid A were found to be required for maximal induction of tissue factor activity and tissue factor mRNA. The presence of serum in the assay was found to amplify but was not obligate for tissue factor induction by LPS.

Herpes simplex virus type I does not require productive infection to induce tissue factor in human umbilical vein endothelial cells.

BACKGROUND: Herpes simplex virus (HSV)-infected endothelium is a model for vascular injury and possibly the development of atherosclerosis. In vitro infection of human umbilical vein endothelial cells (HUVEC) by HSV-1 results in a number of changes including the expression of a procoagulant activity (PCA) compatible with that due to tissue factor (TF) synthesis. In this study, we have further characterized this PCA using more stringent assays for TF, and examined whether virus rendered incapable of replication retains the ability to stimulate TF synthesis in HUVEC. EXPERIMENTAL DESIGN: Confluent monolayers of HUVEC were exposed to intact or ultraviolet/heat-inactivated HSV-1. At appropriate time intervals, TF PCA was assessed by clotting assays, and TF antigen by an enzyme-linked immunosorbent assay specific for TF. The appearance of mRNA specific for TF was performed by Northern blotting. RESULTS: TF activity was demonstrated by both 1-stage and 2-stage clotting assays; the dependence of the latter on factor VIIa, and the inhibition by specific blocking antibodies to human TF support the notion that the PCA is indeed due to TF. Furthermore, cellular TF antigen levels were found to parallel TF activity, and there was a transient de novo expression of TF mRNA. Tissue factor PCA in HSV-infected HUVEC remained "encrypted"; that is, full clotting activity was not expressed in the absence of cellular disruption in a situation analogous to that seen in all normal cells thus far examined that express TF PCA. However, this response did not appear to be dependent upon replicative infection of HSV-1 within the endothelial cell since a similar (although lesser) induction of TF PCA was present in cells that had been exposed to virus previously rendered incapable of replication. CONCLUSIONS: HSV-1 induces PCA in HUVEC which is clearly TF-dependent; this response does not require viral replication. These data indicate increased complexity in HSV interactions with vascular endothelium and imply induction of some procoagulant functions by nonproductive infection.

Functional neurokinin 1 receptors for substance P are expressed by human vascular endothelium.

Substance P (SP), a neurotachykinin, is important in a number of inflammatory processes in which the endothelial cell also plays a critical role. SP receptors have previously been identified only on arterial endothelium, and the scant in vitro evidence for direct effects of SP on human endothelium is based on studies using nonarterial cells. To better understand SP's role in inflammation, we sought to identify functional SP receptors on human endothelium in situ and in culture. Autoradiographic ligand binding to human umbilical cord sections demonstrates the presence of SP binding sites with characteristics of the neurokinin 1 (NK-1) receptor (displacement by GTP analogues and the NK-1 specific antagonist CP-96,345) on human umbilical arterial, but not venous, endothelium. In culture, human umbilical venous endothelial cells (HUVECs) and human aortic endothelial cells express low levels of available SP binding sites. However, HUVECs, which are serum starved and refed, undergo a dramatic increase in SP binding. SP binding to starved/refed HUVECs induces a transient increase in intracellular calcium. This calcium flux is dose dependent over appropriate SP concentrations and can be blocked by NK-1 specific antagonists. The proinflammatory effects of SP may be mediated in part through the NK-1 receptor on endothelium.

Herpes virus infection of endothelium: new insights into atherosclerosis.

Several pieces of evidence suggest that vascular endothelium may be a site of latent herpetic viral infection, and that activation of such infection might cause or aggravate atherosclerosis. The present studies which utilized HSV-1 infection of cultured endothelial monolayers, provide insights into two phenomena seemingly relevant in considerations of atherosclerosis. Thus, mechanisms are reported by which infected endothelium may be damaged by marginated inflammatory cells, and be transformed from an anticoagulant to a procoagulant tissue. First, granulocytes are attracted to, and avidly bind, endothelium infected for very brief periods. This interaction is associated with denudation of intact cells as well as actual cytolysis through release of PMN proteases and toxic oxygen species. Second, several potentially additive abnormalities of HSV-infected endothelium would seem to foster coagulation. These include: a) its loss of surface heparans and thrombomodulin; b) its inability to synthesize prostacyclin with associated incapacity to deter platelet adhesion; c) its disordered membrane lipid conformation which is likely associated with excessive surface thrombin generation; and d) its unique ability to generate and release tissue factor. We speculate that mechanical abrasion may reactivate latent herpes (HSV or CMV) infection in endothelial cells particularly those exposed to high shear forces--for instance, at vessel bifurcations. This may underlie the endothelial damage, clotting and atheroma formation commonly found at these sites.

Hydrogen peroxide alters signal transduction in human endothelial cells.

Lytic H2O2-induced injury to human umbilical vein endothelial cells provides a model for endothelial cell damage in diverse states including acute respiratory distress and septic shock. Endothelial cell lysis is an extreme result of inflammatory cell activation. Functional alterations such as responsiveness to endothelial cell agonists and eicosanoid production might be impaired by exposure to inflammatory cell products including H2O2. Soluble mediators such as thrombin or histamine cause endothelial cell activation via a signal transduction mechanism that hydrolyzes phosphatidylinositol 4,5-bisphosphate (IP), liberating inositol trisphosphate (IP3). Accordingly, pretreatment of endothelial cells with H2O2 blocked the subsequent production of IP3 in response to thrombin and histamine. H2O2 inhibition of IP3 was time- and concentration-dependent. The endothelial cells were viable by trypan blue dye exclusion and chromium release. H2O2 inhibition of signaling was completely prevented by catalase. Iron-dependent oxidant radical formation appears critical because deferoxamine (10(-4) mol/L) pretreatment of endothelial cells prevented H2O2 inhibition of IP hydrolysis. Prostacyclin and platelet activating factor production in response to thrombin have been linked to IP hydrolysis. Pretreatment of endothelial cells with H2O2 reduced prostacyclin and platelet-activating factor production by thrombin by at least 50%. It appears H2O2 can induce defects in signaling pathways with sequelae (decreased prostacyclin and platelet-activating factor) short of endothelial cell death. The possible consequences of H2O2 interaction with endothelial cells is reviewed with the aim of presenting a hypothesis to integrate these various observations.

Infection of vascular endothelial cells with herpes simplex virus enhances tissue factor activity and reduces thrombomodulin expression.

Latent infection of vascular cells with herpes-viruses may play a pathogenic role in the development of human atherosclerosis. In a previous study, we found that cultured human umbilical vein endothelial cells (HUVECs) infected with herpes simplex virus 1 (HSV-1) became procoagulant, exemplified both by their enhanced assembly of the prothrombinase complex and by their inability to reduce adhesion of platelets. We now report two further procoagulant consequences of endothelial HSV infection: loss of surface thrombomodulin (TM) activity and induction of synthesis of tissue factor. Within 4 hr of infection of HUVECs, TM activity measured by thrombin-dependent protein C activation declined 21 +/- 3% (P less than 0.05) and by 18 hr, 48 +/- 5% (P less than 0.001). Similar significant TM decrements accompanied infection of bovine aortic endothelial cells. Identical TM loss was induced with HSV-2 infection but not with adenovirus infection. Decreased surface expression of TM antigen (measured by the specific binding of a polyclonal antibody to bovine TM) closely paralleled the loss of TM activity. As examined by Northern blotting, these losses apparently reflected rapid onset (within 4 hr of HSV infection) loss of mRNA for TM. In contrast, HSV infection induced a viral-dose-dependent increase in synthesis of tissue factor protein, adding to the procoagulant state. The results indicate that loss of endothelial protein-synthetic capacity is not a universal effect of HSV infection. We suggest that the procoagulant state induced by reduction in TM activity and amplified tissue factor activity accompanying HSV infection of endothelium could contribute to deposition of thrombi on atherosclerotic plaques and to the "coagulant-necrosis" state that characterizes HSV-infected mucocutaneous lesions.

Endothelial cell platelet-activating factor primes neutrophil responses: amplification of endothelial activation by neutrophil products.

We have shown that platelet-activating factor (PAF) primes neutrophil (PMN) responses and enhances their ability to damage endothelial cells. Furthermore, thrombin-stimulated endothelial cells which produce PAF can augment and prime PMN superoxide production, elastase release and adhesion to endothelium. We wondered whether these marginated neutrophils (PMN) themselves, or their release products, might feedback and further amplify endothelial cell activation. To measure cellular activation, we assessed changes in endothelial cell intracellular calcium [( Ca2+]i) in endothelial monolayers loaded with Fura-2, and PAF production by [3H]acetate incorporation into phospholipid. Resting PMNs induced no change in [Ca2+]i, while N-formyl-L-methionine-L-leucyl-L-phenylalanine stimulated PMN and their lysosomal products caused a 25% increase in endothelial cell calcium. Sonicates of PMN produced a much larger increase in [Ca2+]i than activated PMN; the effect of PMN sonicates could in part be reduced by the serine protease inhibitor, alpha 1 antitrypsin. In contrast, purified neutrophil elastase did not alter endothelial cell [Ca2+]i. Despite hydrogen peroxide's ability to increase [Ca2+]i, catalase failed to inhibit the PMN-induced rise in [Ca2+]i. Since polyanionic heparin inhibited the PMN sonicate rise in calcium, a cationic protein released by PMN was thought to be responsible. The cationic primary granule enzyme, cathepsin G, duplicates the rise in [Ca2+]i seen with PMN sonicates. Furthermore, PAF production increased threefold in response to neutrophil sonicates. Thus, during inflammation, when coagulation and inflammatory cells are activated the endothelium responds by priming PMNs and promoting margination.(ABSTRACT TRUNCATED AT 250 WORDS)

Measurement of intracellular calcium concentration in intact monolayers of human endothelial cells.

Intracellular calcium ([Ca++]i) plays an important role in signal transduction and cell activation. The measurement of [Ca++]i in intact monolayers of human umbilical vein endothelial cells with the fluorescent calcium-sensitive probe fura 2 has been evaluated. Monolayers provide a more physiologic cell preparation than suspensions and allow a greater variety of experimental manipulation. Basal [Ca++]i was 117 +/- 5 nmol/L, with a range from 40 to 280 nmol/L that was not affected by cell age (days of primary culture) or degree of confluence. Thrombin in concentrations of 0.005 to 5 NIH units/ml produced a dose-dependent increase in [Ca++]i up to a maximum of 1500 +/- 147 nmol/L; this increase was shown to depend in part on the concentration of extracellular calcium. The presence of antithrombin III at physiologic concentrations abolished responses to 0.5 NIH units/ml thrombin but had no effect on 5 NIH units/ml. The potential of this technique was demonstrated further by our ability to examine [Ca++]i responses in endothelial cells following infection with herpes simplex virus type 1, a virus implicated in vascular injury. After 18 hours' infection, the response to both thrombin and histamine was dramatically reduced despite a normal resting [Ca++]i. It is concluded that this method may be useful for detecting early and subtle changes in endothelial cell function under a variety of physiologic and pathologic conditions.

Pulmonary hypertension and edema induced by platelet-activating factor in isolated, perfused rat lungs are blocked by BN52021.

The experimental intravenous administration of platelet activating factor (PAF) induces pulmonary hypertension and directly or indirectly increases capillary permeability. Selective PAF antagonists BN52021 and L652-731 have been shown to inhibit the action of PAF in vitro and in vivo. Using a unique isolated perfused rat lung model, we measured the effect of these PAF antagonists on PAF-induced pulmonary hypertension and edema. Isolated rat lungs were perfused with Krebs-Henseleit solution. The right and left pulmonary arteries were dissected so that they could be perfused selectively, permitting the use of one lung as an internal control for a specific pharmacologic challenge. Exposure of one lung to PAF induced an increase of perfusion pressure and wet/dry lung weight ratio in a dose-dependent manner compared with the control lung. The PAF antagonists attenuated the increase in perfusion pressure and wet/dry lung weight caused by PAF (0.75 micrograms) in a dose-dependent manner. In addition, prostaglandin F2 alpha induced an equivalent increase in pulmonary pressure without causing a similar increase in lung edema. PAF-induced pulmonary hypertension and the increase in wet/dry lung weight ratio appear to be PAF receptor-mediated processes, and the use of specific antagonists and this technique may be useful probes to determine the role of PAF in pathophysiologic states.

Endothelial cell heterogeneity: antioxidant profiles determine vulnerability to oxidant injury.

Human umbilical vein endothelial cells were more sensitive to hydrogen peroxide lysis than cow pulmonary artery endothelial cells. Conversely, activated neutrophils which utilize hydrogen peroxide-mediated cell cytotoxicity cell mechanisms were more toxic to the cow pulmonary artery cells. This discordance was not related to neutrophil adhesion to either cell type or cell passage number. The antioxidant profiles of the endothelial cells revealed that cow pulmonary artery cells were rich in catalase to consume bolus hydrogen peroxide presented to them, while human umbilical vein endothelial cells utilize glutathione peroxidase-linked mechanisms to detoxify a slower more sustained release of hydrogen peroxide generated by neutrophils. Endothelial cells from different species and sites may utilize diversified antioxidant protective mechanisms.

Interleukin 1 or endotoxin increases the release of von Willebrand factor from human endothelial cells.

von Willebrand factor (vWF), a large adhesive glycoprotein, is synthesized by vascular endothelial cells (EC). Plasma levels of vWF manifest a broad normal range, and are elevated during sepsis and in inflammatory states. Since the inflammatory mediator, interleukin 1 (IL1) and bacterial endotoxin (LPS) both initiate procoagulant changes in vascular endothelium, we investigated the effect of these substances on endothelial cell release and residual endothelial cell content of vWF-antigen (vWFAg). Cultured human EC exposed to either IL1 or LPS released greater amounts of vWFAg compared to control EC. The augmented release could be detected within 1-2 h after exposure to IL1 or LPS and was not inhibited by cycloheximide, suggesting that de novo protein synthesis was not required for release to occur. Residual cellular vWFAg was reciprocally lower in IL1- or LPS-treated EC at 24 and 48 h, indicating that compensatory increase in synthesis of vWFAg did not occur during this time interval. Released vWF contained the higher molecular weight multimers observed in normal endothelial cells, and it possessed ristocetin cofactor activity. We propose that release of functional vWF from EC exposed to inflammatory mediators may be at a mechanism for localization of platelets and enhanced thrombogenicity at inflammatory foci.

Interleukin 1 stimulates endothelial cell tissue factor production and expression by a prostaglandin-independent mechanism.

Activation of coagulation occurs at inflammatory sites following the ingress of mononuclear cells, and may result from alterations in the vessel wall. Since the monokine, interleukin 1, initiates diverse responses to inflammation, its ability to enhance vascular procoagulant activity was studied. Interleukin 1-treated cultured human endothelial cells acquired elevated levels of the procoagulant, tissue factor. This required de novo protein synthesis, was maximal at 2 h after exposure to interleukin 1, and resulted in persistently elevated cellular procoagulant activity. Tissue factor was later expressed (6-24 h) on the surface of uninjured endothelial cells. Endothelial cell procoagulant production and expression in response to interleukin 1 could be dissociated from endogenous prostaglandin metabolism, being insensitive to hydrocortisone, indomethacin, eicosatetrayionic acid and exogenous arachidonic acid. In addition, no increase in prostaglandin synthesis occurred during the interval in which tissue factor was synthesized. We therefore conclude that interleukin 1 stimulates endothelial synthesis and surface expression of tissue factor by a prostaglandin-independent mechanism.

Magnesium sulfate: rationale for its use in preeclampsia.

Preeclampsia is a disorder of pregnancy characterized clinically by hypertension, proteinuria, and edema and characterized pathologically in its late stages by widespread microvascular thrombi. There is evidence from a number of studies that production of prostacyclin (prostaglandin I2, PGI2), a potent vasodilator and inhibitor of platelet aggregation, is deficient in preeclamptic compared to normal pregnancy. Traditional therapy utilizes infusions of large amounts of MgSO4, but the physiologic basis for this is not clear. We studied the effect of MgSO4 on PGI2 release by cultured human umbilical vein endothelial cells (HUVEC) by several methods. By platelet aggregometry, the known antiaggregatory effect of intact HUVEC was enhanced by MgSO4. By radioimmunoassay for 6-keto-PGF1 alpha, the stable metabolite of PGI2, it was shown that MgSO4 amplifies release of PGI2 by HUVEC in a dose-dependent manner, with a peak occurring between 2 and 3 mM. In separate experiments, MgSO4 overcame the enhanced adherence of platelets to HUVEC exhausted by repeated exposure to thrombin. Finally, PGI2 production was 2- to 5-fold greater by HUVEC incubated with plasma obtained from preeclamptic patients undergoing MgSO4 therapy than by HUVEC incubated with pretherapy plasma. We conclude that MgSO4 mediates enhanced production of PGI2 by vascular endothelium, thereby potentially enhancing its thromboresistant properties.

Structural features of endotoxin required for stimulation of endothelial cell tissue factor production; exposure of preformed tissue factor after oxidant-mediated endothelial cell injury.

Cultured human umbilical vein endothelial cells synthesize the procoagulant, tissue factor, after exposure to bacterial endotoxin. Wild-type lipopolysaccharide from Escherichia coli 0127:B8 stimulates a five- to 20-fold increase in cellular tissue factor. Similarly, rough or incomplete lipopolysaccharide subunits from mutant bacterial strains, or lipid A prepared by mild acid hydrolysis of whole endotoxin, are also stimulatory. In addition, a lipid A biosynthetic precursor, consisting of a phosphorylated glucosamine disaccharide substituted with four beta-hydroxymyristoyl residues, is stimulatory at nanomolar concentrations. Endothelial cell tissue factor is not detectable on the surface of undisrupted cells, but can activate clotting on the cell surface after oxidant-mediated cell injury. The procoagulant, tissue factor, is synthesized by endothelial cells after stimulation mediated by a moiety contained within the lipid A region of lipopolysaccharide. Exposure of clotting factors at the endothelial cell surface after cell injury suggests a mechanism for the microvascular thrombosis associated with disseminated intravascular coagulation with sepsis.

Protection against lethal hyperoxia by tracheal insufflation of erythrocytes: role of red cell glutathione.

Intact erythrocytes placed into the tracheobronchial tree of hyperoxic rats dramatically improved their chances for survival. Over 70 percent of the animals so treated survived more than 12 days during continuous exposure to 95 percent oxygen, whereas all of the control animals died within 96 hours. Lungs from erythrocyte-protected rats showed almost none of the morphologic damage suffered by untreated animals. Erythrocytes containing cyanomethemoglobin were as beneficial as normal erythrocytes, but cells in which glutathione was partially blocked were significantly less protective. Analogous results were obtained in vitro: 51Cr-labeled target cells released 70 to 90 percent of their label when exposed briefly to hydrogen peroxide or to toxic oxygen species generated by phorbol ester-stimulated neutrophils. Addition of intact erythrocytes decreased release by approximately 75 percent, but significantly less than this if red blood cell glutathione was partially blocked. These results suggest that insufflated erythrocytes, through their recyclable glutathione, protect rats from toxic oxygen species engendered by hyperoxia.