The association between alpha4-integrin, P-selectin, and E-selectin in an allergic model of inflammation.

In this study, we examined the relationship between the endothelial selectins (P-selectin and E-selectin) and whether they are critical for alpha4-integrin-dependent leukocyte recruitment in inflamed (late phase response), cremasteric postcapillary venules. Animals were systemically sensitized and 2 wk later challenged intrascrotally with chicken ovalbumin. Leukocyte rolling flux, adhesion, and emigration were assessed at baseline and 4 and 8 h postantigen challenge. There was a significant increase in leukocyte rolling flux, adhesion, and emigration in sensitized and challenged mice at both 4 and 8 h. At 8 h, the increase in leukocyte rolling flux was approximately 50% inhibitable by an anti-alpha4-integrin antibody, 98% inhibitable by fucoidin (a selectin-binding carbohydrate), and 100% inhibitable by an anti-P-selectin antibody. P-selectin-deficient animals displayed no leukocyte rolling or adhesion at 8 h after challenge. However, at 8 h there were many emigrated leukocytes in the perivascular space suggesting P-selectin-independent rolling at an earlier time point. Indeed, at 4 h postantigen challenge in P-selectin-deficient mice, there was increased leukocyte rolling, adhesion, and emigration. The rolling in the P-selectin-deficient mice at 4 h was largely alpha4-integrin dependent. However, there was an essential E-selectin-dependent component inasmuch as an anti-E-selectin antibody completely reversed the rolling, and in E-selectin and P-selectin double deficient mice rolling, adhesion and emigration were completely absent. These results illustrate that P-selectin underlies all of the antigen-induced rolling with a brief transient contribution from E-selectin in the P-selectin-deficient animals. Finally, the antigen-induced alpha4-integrin-mediated leukocyte recruitment is entirely dependent upon endothelial selectins.

Absence of trauma-induced leukocyte rolling in mice deficient in both P-selectin and intercellular adhesion molecule 1.

Leukocyte recruitment during inflammation is achieved through a multistep paradigm that includes margination, selectin-mediated rolling, beta 2 integrin-mediated firm adhesion, emigration, and migration into the site of inflammation. We have used the mouse cremaster muscle as a model of trauma- and cytokine-induced inflammation to study the possible role of intercellular adhesion molecule (ICAM) 1 in leukocyte rolling using gene-targeted mice deficient in ICAM-1, P-selectin, and a combination of P-selectin and ICAM-1. Rolling flux and average leukocyte rolling velocity in ICAM-1-deficient mice was not different from wild-type mice, but P-selectin/ICAM-1-deficient mice showed a total absence of rolling for at least 2 h after surgical trauma. Rolling in both wild-type and ICAM-1-deficient mice 60-120 min after trauma was significantly inhibited by a P-selectin monoclonal antibody (mAb) (RB40.34). In contrast, an mAb (KAT-1) blocking ICAM-1 binding to leukocyte function-associated antigen 1 did not block residual rolling in P-selectin-deficient mice. TNF-alpha induced leukocyte rolling in P-selectin/ICAM-1-deficient mice, but the rolling flux fraction was significantly lower than in TNF-alpha-treated ICAM-1-deficient mice. Leukocyte rolling in P-selectin/ICAM-1-deficient mice treated with TNF-alpha for 3 h was completely blocked by an E-selectin mAb (9A9E3), and partially by an L-selectin mAb (MEL-14). This clearly demonstrates E-selectin-dependent rolling in vivo. Leukocyte rolling velocities were significantly reduced after TNF-alpha treatment and were similar in wild-type and gene-targeted strains. We conclude that the residual trauma-induced leukocyte rolling seen in P-selectin-deficient mice is completely abolished by concomitant ICAM-1 deficiency. This severe defect in leukocyte rolling may explain the absence of leukocyte recruitment into the inflamed peritoneal cavity of P-selectin/ICAM-1-deficient mice at early time points (< or = 4 h).

P- and E-selectin use common sites for carbohydrate ligand recognition and cell adhesion.

The selectins are a family of three calcium-dependent lectins that mediate adhesive interactions between leukocytes and the endothelium during normal and abnormal inflammatory episodes. Previous work has implicated the carbohydrate sialyl Lewis(x) (sLe(x); sialic acid alpha 2-3 galactose beta 1-4 [Fucose alpha 1-3] N-acetyl glucosamine) as a component of the ligand recognized by E- and P-selectin. In the case of P-selectin, other components of the cell surface, including 2'6-linked sialic acid and sulfatide (galactose-4-sulfate ceramide), have also been proposed for adhesion mediated by this selectin. We have recently defined a region of the E-selectin lectin domain that appears to be directly involved with carbohydrate recognition and cell adhesion (Erbe, D. V., B. A. Wolitzky, L. G. Presta, C. R. Norton, R. J. Ramos, D. K. Burns, R. M. Rumberger, B. N. N. Rao, C. Foxall, B. K. Brandley, and L. A. Lasky. 1992. J. Cell Biol. 119:215-227). Here we describe a similar analysis of the P-selectin lectin domain which demonstrates that a homologous region of this glycoprotein's lectin motif is involved with carbohydrate recognition and cell binding. In addition, we present evidence that is inconsistent with a biological role for either 2'6-linked sialic acid or sulfatide in P-selectin-mediated adhesion. These results suggest that a common region of the E- and P-selectin lectin domains appears to mediate carbohydrate recognition and cell adhesion.

Identification of an E-selectin region critical for carbohydrate recognition and cell adhesion.

E-selectin elicits cell adhesion by binding to the cell surface carbohydrate, sialyl Lewis X (sLe(x)). We evaluated the effects of mutations in the E-selectin lectin domain on the binding of a panel of anti-E-selectin mAbs and on the recognition of immobilized sLe(x) glycolipid. Functional residues were then superimposed onto a three-dimensional model of the E-selectin lectin domain. This analysis demonstrated that the epitopes recognized by blocking mAbs map to a patch near the antiparallel beta sheet derived from the NH2 and COOH termini of the lectin domain and two adjacent loops. Mutations that affect sLe(x) binding map to this same region. These results thus define a small region of the E-selectin lectin domain that is critical for carbohydrate recognition.

Active site-blocked factor IXa prevents intravascular thrombus formation in the coronary vasculature without inhibiting extravascular coagulation in a canine thrombosis model.

To assess the contribution of Factor IX/IXa, to intravascular thrombosis, a canine coronary thrombosis model was studied. Thrombus formation was initiated by applying current to a needle in the circumflex coronary artery. When 50% occlusion of the vessel developed, the current was stopped and animals received an intravenous bolus of either saline, bovine glutamyl-glycyl-arginyl-Factor IXa (IXai), a competitive inhibitor of Factor IXa assembly into the intrinsic Factor X activation complex, bovine Factor IX, or heparin. Animals receiving saline or Factor IX developed coronary occlusion due to a fibrin/platelet thrombus in 70 +/- 11 min. In contrast, infusion of IXai prevented thrombus formation completely (greater than 180 min) at doses of 460 and 300 micrograms/kg, and partially blocked thrombus formation at 150 micrograms/kg. IXai attenuated the accumulation of 125I-fibrinogen/fibrin at the site of the thrombus by approximately 67% (P less than 0.001) and resulted in approximately 26% decrease in serotonin release from platelets in coronary sinus (P less than 0.05). Hemostatic variables in animals receiving IXai, remained within normal limits. Animals given heparin in a concentration sufficient to prevent occlusive thrombosis had markedly increased bleeding, whereas heparin levels that maintained extravascular hemostasis did not prevent intracoronary thrombosis. This suggests that Factor IX/IXa can contribute to thrombus formation, and that inhibition of IXa participation in the clotting mechanism blocks intravascular thrombosis without impairing extravascular hemostasis.

Hypoxia-mediated induction of endothelial cell interleukin-1 alpha. An autocrine mechanism promoting expression of leukocyte adhesion molecules on the vessel surface.

Tissue injury that accompanies hypoxemia/reoxygenation shares features with the host response in inflammation, suggesting that cytokines, such as IL-1, may act as mediators in this setting. Human endothelial cells (ECs) subjected to hypoxia (PO2 approximately 12-14 Torr) elaborated IL-1 activity into conditioned media in a time-dependent manner; this activity was completely neutralized by an antibody to IL-1 alpha. Production of IL-1 activity by hypoxic ECs was associated with an increase in the level of mRNA for IL-1 alpha, and was followed by induction of endothelial-leukocyte adhesion molecule-1 and enhanced expression of intercellular adhesion molecule-1 (ICAM-1) during reoxygenation. During reoxygenation there was a three- to five-fold increased adherence of leukocytes, partly blocked by antibodies to endothelial-leukocyte adhesion molecule-1 and ICAM-1. Suppressing endothelial-derived IL-1, using either antibodies to IL-1 alpha, specific antisense oligonucleotides or the IL-1 receptor antagonist, decreased leukocyte adherence to reoxygenated ECs, emphasizing the integral role of IL-1 in the adherence phenomenon. Mice subjected to hypoxia (PO2 approximately 30-40 Torr) displayed increased plasma levels of IL-1 alpha, induction of IL-1 alpha mRNA in the lung, and enhanced expression of ICAM-1 in pulmonary tissue compared with normoxic controls. These data suggest that hypoxia is a stimulus which induces EC synthesis and release of IL-1 alpha, resulting in an autocrine enhancement in the expression of adhesion molecules.

Turnover of myofibrillar proteins in cultured muscle cells from normal and dystrophic chick embryos.

Rates of protein degradation and synthesis were determined in myofibrillar and non-myofibrillar fractions and in myofibrillar subunits of cultured muscle cells from normal and dystrophic chick embryos of genetically closely related lines. Growth characteristics of normal and dystrophic cells were identical as measured by DNA, RNA and protein accumulation. Degradation rates of myofibrillar and non-myofibrillar protein determined from label-chase experiments were the same in normal and dystrophic cultures. In similar experiments in which degradation rates of eight different components of the myofibrillar fraction were measured, a spectrum of degradation rates was obtained indicating that myofibril components turn over individually or in groups, rather than as an intact unit. Myosin heavy chain was the slowest turning over component of those measured, while actin was among the most rapid. No differences were found between normals and dystrophics in the turnover of any of the components. Fractional rates of protein synthesis were also determined for the myofibrillar and non-myofibrillar fractions of cultured cells. The rates obtained with normal and dystrophic cells were indistinguishable and were the same for the myofibrillar and non-myofibrillar fractions. Initial amino acid incorporation rates into 17 separate components of the myofibrillar fractions were compared in normal and dystrophic cells. No significant differences were found.

Characterization of a novel adhesion function blocking monoclonal antibody to rat/mouse P-selectin generated in the P-selectin-deficient mouse.

P-selectin is an important adhesion molecule involved in leukocyte migration. However, to date, no monoclonal antibodies (MAb) generated against rat P-selectin have been identified which block P-selectin mediated leukocyte adhesion. Most studies in the rat have utilized crossreacting antibodies generated against P-selectin in higher species. In a P-selectin deficient mouse we generated an anti-rat/mouse P-selectin MAb, designated RMP-1, by immunization with activated rat platelets. This IgG2a MAb immunoprecipitates a 140 kDa protein under reducing conditions from rat platelet lysate. By ELISA and immunofluorescence flow cytometry, MAb RMP-1 reacts with thrombin-activated but not unactivated rat platelets. In addition, by ELISA MAb RMP-1 binds to activated mouse platelets and recombinant rat and mouse P-selectin. MAb RMP-1 inhibited adhesion of HL-60 myeloid cells to immobilized mouse P-selectin by 97% and to activated rat and mouse platelets by 100% under static conditions, confirming the adhesion function blocking activity of MAb RMP-1. This novel MAb should be useful for studying P-selectin function in vitro and in vivo in both rat and mouse inflammation models.

Generation and characterization of a novel adhesion function blocking monoclonal antibody recognizing both rat and mouse E-selectin.

The prerequisite for the recruitment of circulating leukocytes to sites of inflammation is adhesion to vascular endothelial cells. Selectins play a significant role in the initiation of this multistep process by mediating "rolling" of the leukocytes on the endothelium. Investigation of selectin-dependent cell interactions using function blocking monoclonal antibodies (MAb) provides insights into the mechanisms involved in leukocyte migration into inflammation. Until now most studies in inflammation models in rats have relied on cross-reactive or polyclonal antibodies against rat E-selectin. In an E-selectin knockout mouse, we aimed to generate an adhesion function blocking MAb to rat E-selectin by immunization with rat E-selectin transfected Chinese hamster ovary cells (RESEC). An IgG1 kappa MAb was identified that reacts with RESEC but not with untransfected Chinese hamster ovary cells, as well as with recombinant mouse E-selectin protein as assessed by ELISA. This MAb is designated RME-1. It does not cross-react with rat L-selectin or rat P-selectin or E-selectin expressed on human umbilical vein endothelium. Adhesion of the HL-60 myeloid cells to immobilized mouse E-selectin was completely inhibited by MAb RME-1 under static conditions and adhesion of rat polymorphonuclear leukocytes to recombinant mouse E-selectin was blocked under rotation condition. This novel antibody thus recognizes a function-related epitope on rodent E-selectin.

Regulation of the (Na+ + K+)-ATPase in cultured chick skeletal muscle. Modulation of expression by the demand for ion transport.

The levels of (Na+ + K+)-ATPase expression during muscle development and in response to modulation of demand for ion transport were studied in chick skeletal muscle cells in culture. The number of (Na+ + K+)-ATPase molecules on the myogenic cell surface, quantified with 125I-labeled monoclonal antibodies, increased 20-fold during muscle differentiation, with a substantial increase in (Na+ + K+)-ATPase molecules/unit area of membrane. The demand for sodium ion transport by the (Na+ + K+)-ATPase was modulated by activating voltage-sensitive sodium channels with veratridine or exposing cultures to low [K+]o (0.5 mM). Exposure to veratridine (10 microM) resulted in a 60-100% increase in cell surface and a smaller increase in intracellular (Na+ + K+)-ATPase over a 24-36-h period. Neither high [K+]o (50 mM) nor Ca2+ ionophore A23187 (1 microM) produced any such change, suggesting that neither membrane depolarization nor elevated cytosolic calcium was mediating the effect of veratridine. Veratridine stimulated up-regulation was specific for the (Na+ + K+)-ATPase, blocked by tetrodotoxin, and completely reversible. The kinetics of the reversal (down-regulation) process were much faster (t1/2 = 3 h) than those of up-regulation (t1/2 = 18 h). Up-regulation of the (Na+ + K+)-ATPase by veratridine occurred by a combination of two mechanisms: the first an early phase involving a stimulated biosynthesis of the (Na+ + K+)-ATPase and a later phase in which the biosynthetic rate returned to approximately control levels while the degradation rate slowed (t1/2 control = 31 h, t1/2 veratridine = 64 h).

Imide and lactam derivatives of N-benzylpyroglutamyl-L-phenylalanine as VCAM/VLA-4 antagonists.

A series of imides and lactams derived from 4-amino-N-benzylpyroglutamyl-L-phenylalanine was prepared and evaluated for activity as VCAM/VLA-4 antagonists. Imides were more potent than the corresponding lactams; several had subnanomolar IC50s in an ELISA based assay and were also highly effective at blocking VLA-4 expressing Ramos cell binding to VCAM coated plates.

Interleukin-12-induced adhesion molecule expression in murine liver.

Systemically administered interleukin (IL)-12 causes liver inflammation in mice characterized by Kupffer cell proliferation and hypertrophy, hepatocyte necrosis, and multifocal accumulations of leukocytes in the hepatic parenchyma and around portal tracts and central veins. We have used both immunohistochemical staining and radiolabeled antibody quantitation to examine adhesion molecule expression in the livers of mice dosed daily with murine IL-12. Cells infiltrating livers of IL-12-treated mice were primarily mononuclear leukocytes expressing LFA-1, VLA-4, MAC-1, and CD18 adhesion molecules but little L-selectin. Kupffer cells constitutively expressed LFA-1 and smaller amounts of MAC-1, and high levels of ICAM-1 were constitutively expressed by liver sinusoidal lining cells, portal tract, and central vein endothelia. With IL-12 treatment, existing ICAM-1 expression was up-regulated and de novo expression occurred along bile duct epithelia. VCAM-1 levels were dramatically increased, with induced expression occurring along portal tract and central vein endothelia and scattered bile duct epithelial cells and in aggregations of cells in perivascular areas and the liver parenchyma. Although constitutive expression of E- and P-selectin was negligible, Il-12 induced a moderate rise in E-selectin levels. These increases in adhesion molecule expression may have implications for the therapeutic use of IL-12, especially in patients with liver disease or autoimmune conditions where augmented adhesion molecule expression may be critical to disease pathogenesis.

Characterization of murine E-selectin expression in vitro using novel anti-mouse E-selectin monoclonal antibodies.

A recombinant phage containing the structural exons for mouse E-selectin has been isolated and characterized. Utilizing PCR techniques the lectin and egf domains were fused to form an artificial cDNA for expression in eukaryotic cells. Transient expression in COS cells demonstrated the lectin and egf domains were sufficient to mediate the binding of mouse and human neutrophils as well as HL60 cells. Recombinant soluble mouse E-selectin was purified and used to immunize rats to generate mAbs specific to mouse E-selectin. A panel of mAbs directed against mouse E-selectin was characterized including five that inhibit the adhesion of HL60 cells or mouse neutrophils to COS cells expressing the mouse lectin/egf domains. These mAbs have been used to characterize the expression and function of E-selectin on cytokine stimulated eEnd.2 murine endothelial cells.

Mouse bone marrow-derived mast cells roll on P-selectin under conditions of flow in vivo.

Increased numbers of mast cells are noted at sites of wound healing and inflammation. These mast cells are either recruited from the bone marrow or proliferate locally under cytokine stimulation. However, the molecular mechanisms mediating initial adhesive interactions between mast cell precursors and vascular endothelial cells are not well understood. We have used a syngeneic dorsal skinfold chamber model of microcirculation to study early events of mast cell-endothelial cell interactions by intravital fluorescence microscopy. Because "rolling" represents the earliest step of granulocyte adhesion under conditions of flow, our objective was to determine whether vascular selectins promote rolling of immature mouse bone marrow-derived mast cells (MBMMC) on endothelial cells lining murine blood vessels in vivo. In this study, titanium window chambers were implanted on the dorsal skinfolds of BALB/c mice. The passage of injected fluorescently labeled MBMMC within blood vessels of the striated skin muscle was observed by stroboscopic epi-illumination. As previously determined for other leukocytes, MBMMC were observed to roll in venules but not in arterioles or capillaries. Mice were also treated with neutralizing anti-E-selectin (mAb 9A9) and anti-P-selectin (mAb 5H1) antibodies and tested for their ability to block MBMMC rolling on venular endothelial cells. Intravenous administration of mAb 5H1 resulted in a marked decrease in MBMMC rolling, whereas mAb 9A9 and isotype matched control antibodies had no effect on the rolling flux of MBMMC. These studies represent the first identification of P-selectin as a rolling receptor for MBMMC, and demonstrate the use of a dorsal skinfold technique to study MBMMC-endothelial cell interactions under conditions of physiologic flow. Further studies will determine whether vascular selectins participate in the rolling and tissue recruitment of true circulating immature mast cell precursors in vivo.

Blocking both E-selectin and P-selectin inhibits endotoxin-induced leukocyte infiltration into the eye.

The initial contact between leukocytes and the vascular endothelium at sites of inflammation is mediated by selectins. The purpose of this study was to investigate the role of the two selectins expressed on the vascular endothelium, E-selectin and P-selectin, in the pathogenesis of endotoxin-induced uveitis. Endotoxin-induced uveitis was produced in female C3H/HeN mice using Salmonella typhimurium endotoxin injected into one hind footpad. At the time of endotoxin injection mice were treated with an intraperitoneal injection of a monoclonal antibody against E-selectin or P-selectin, a combination of both anti-selectin antibodies, or isotype-matched control antibodies. In a second set of experiments, antibody treatment was administered 6 hr after endotoxin injection, when inflammatory cells are already entering the eye. Ocular inflammation was graded histologically by a masked observer. When administered at the time of endotoxin injection, anti-P-selectin antibody decreased ocular inflammation by 37% compared to control animals (P = 0.05). There was no statistical decrease in ocular inflammation in animals treated with anti-E-selectin antibody. The combination of anti-P-selectin and anti-E-selectin antibodies decreased infiltrating inflammatory cells by 61% (P < 0.01). When treatment was delayed until 6 hr after endotoxin injection, the combination of anti-P-selectin and anti-E-selectin antibodies again decreased ocular inflammation by 60% (P < 0.01). Immunohistochemical staining showed decreased ICAM-1 expression in the eyes of animals treated with the combination of anti-P-and anti-E-selectin antibodies. Blocking both P-selectin and E-selectin resulted in a significant decrease in endotoxin-induced intraocular inflammation.

Endothelial cell adhesion molecule expression in gene-targeted mice.

Gene-targeted mice are now routinely employed as tools for defining the contribution of different leukocyte and endothelial cell adhesion molecules to the leukocyte recruitment and tissue injury associated with acute and chronic inflammation. The objective of this study was to determine whether gene-targeted mice that are deficient in CD11/CD18, intracellular adhesion molecule-1 (ICAM-1), or P-selectin exhibit an altered constitutive or induced expression of the endothelial cell adhesion molecules E- and P-selectin. The gene-targeted mice were all developed in the 129Sv mouse strain and backcrossed into C57B1/6J mice. The number of backcrosses ranged between 8 (P-selectin) and 10 (CD18 and ICAM-1) generations. The dual-radiolabeled monoclonal antibody technique was used to quantify E- and P-selectin expression in different vascular beds. In the unstimulated state, E-selectin expression was significantly elevated (relative to wild-type mice) in the stomach, large intestine, and brain of mutants deficient in ICAM-1. In general, constitutive expression of P-selectin did not differ between wild-type, ICAM-1-deficient, and CD11/CD18-deficient mutants. In CD11/CD18-deficient mice, tumor necrosis factor-alpha (TNF-alpha) administration elicited a more profound upregulation of P-selectin in several vascular beds, compared with wild-type and ICAM-1-deficient mice. E-selectin expression in brain of TNF-alpha-stimulated, ICAM-1-deficient, and P-selectin-deficient mice was attenuated compared with wild-type mice. These findings indicate that chronic deficiency of some of the adhesion glycoproteins that mediate leukocyte recruitment alters basal and induced surface expression of other adhesion molecules on endothelial cells.

Carbacyclic peptide mimetics as VCAM-VLA-4 antagonists.

Substitution of carbon for sulfur in a potent 13-membered cyclic disulfide containing peptide was accomplished via an intramolecular Wittig reaction and resulted in a series of 'carba' analogues. Potency in the VCAM-VLA-4 assay was sensitive to ring size and lower than that of the parent disulfide.

Cyclic thioether peptide mimetics as VCAM-VLA-4 antagonists.

Selective substitution of a sulfur atom by carbon in a highly potent 13-membered cyclic disulfide was accomplished by intramolecular displacement of a bromide. The potency of the resulting thioethers in the VCAM/VLA-4 assay was dependent on ring size and the position of the sulfur atom.

Heterogeneity of expression of E- and P-selectins in vivo.

A novel technique involving radiolabeled monoclonal antibodies was used to characterize and compare the expression of E- and P-selectin on unstimulated, histamine-challenged, and endotoxin-challenged endothelial cells in various tissues of the mouse. Under unstimulated conditions, E-selectin was absent in all organs, but significant expression of P-selectin was observed in several organs. Histamine induced a rapid time-dependent upregulation of P-selectin, with the largest responses observed in mesentery and lung. Significant fold elevations in P-selectin expression occurred as early as 5 minutes after the histamine injection and remained elevated up to 1 hour. Histamine-induced P-selectin upregulation was inhibited by the H1 receptor antagonist diphenhydramine, whereas the H2 receptor antagonist cimetidine had no effect. Endotoxin (lipopolysaccharide [LPS]) also induced a time-dependent expression of P-selectin that reached a maximum between 4 and 8 hours after endotoxin administration. LPS-induced upregulation of P-selectin was greatest in heart and stomach, which exhibited insignificant constitutive expression of P-selectin. LPS also induced a time-dependent upregulation of E-selectin, with maximal expression occurring 3 to 5 hours after intraperitoneal administration. The lung and small intestine exhibited the largest responses to LPS challenge. Histamine administration did not affect E-selectin expression in any tissue. E- and P-selectin-deficient mice were used to test the specificity of monoclonal antibody binding in unstimulated, histamine-challenged, and LPS-stimulated tissues. Vascular binding of the radiolabeled E-selectin and P-selectin monoclonal antibodies was not observed in the respective deficient mice. These findings suggest that P-selectin is constitutively expressed on vascular endothelium in some tissues of the mouse and that there are significant regional differences in the magnitude and time course of histamine- and endotoxin-induced P-selectin expression. In contrast, E-selectin appears to be absent on unstimulated vascular endothelium but is upregulated within 3 hours after the administration of endotoxin in most tissues.