Leukocyte Microparticles Stimulate Endothelial Cell Cytokine Release and Tissue Factor Induction in a JNK1 Signaling Pathway*

A role of membrane microparticles (MP) released by vascular cells in endothelial cell (EC) activation was investigated. Flow cytofluorimetric analysis of blood samples from normal volunteers revealed the presence of an heterogeneous MP population, which increased by ∼2-fold after inflammatory stimulation with the chemotactic peptide, N-formyl-Met-Leu-Phe (2,799 ± 360 versus 5241 ± 640, p < 0.001). Blood-derived MP stimulated release of EC cytokines interleukin (IL)-6 (377 ± 68 pg/ml) and MCP-1 (1, 282 ± 79) and up-regulatedde novo expression of tissue factor on the EC surface. This was associated with generation of a factor Xa-dependent procoagulant response (2.28 ± 0.56 nm factor Xa/min/104 cells), in a reaction inhibited by a monoclonal antibody to tissue factor. Fluorescent labeling with antibodies to platelet GPIbα or leukocyte lactoferrin demonstrated that circulating MP originated from both platelets and leukocytes. However, depletion of platelet MP with an antibody to GPIbα did not reduce EC IL-6 release, and, similarly, MP from thrombin-stimulated platelets did not induce IL-6 release from endothelium. EC stimulation with leukocyte MP did not result in activation of the transcription factor NF-κB and was not associated with tyrosine phosphorylation of extracellular signal-regulated protein kinase, ERK1. In contrast, leukocyte MP stimulated a sustained, time-dependent increased tyrosine phosphorylation of ∼46-kDa c-Jun NH2-terminal kinase (JNK1) in EC. These findings demonstrate that circulating leukocyte MP are up-regulated by inflammatory stimulation in vivo and activate a stress signaling pathway in EC, leading to increased procoagulant and proinflammatory activity. This may provide an alternative mechanism of EC activation, potentially contributing to dysregulation of endothelial functions during vascular injury.

Considerable interest has recently focused on alternative mechanisms of EC activation by vascular cells. Recent work has suggested that released membrane microparticles (MP) from platelets (14,15) or leukocytes (16) may provide such an alternative pathway of EC activation. In these studies, platelet or leukocyte MP stimulated increased expression of various adhesion molecules on EC, up-regulation of inflammatory and chemotactic cytokines, and increased monocyte adhesiveness (14 -16). For platelet MP, this pathway was recapitulated by arachidonic acid, consistent with the presence of bioactive lipids in platelet MP, and their ability to influence gene expression in target cells (14,15). Although the existence of platelet MP in vivo and their potential contribution to procoagulant and/or anticoagulant responses have long been established (17)(18)(19)(20), little is known about leukocyte MP or their potential ability to stimulate EC in vivo.
In this study, we sought to investigate a potential role of leukocyte MP in EC responses and to identify signaling requirements involved in gene expression. We found that although both platelet and leukocyte MP are present in the normal circulation in vivo, only the leukocyte fraction initiates signal transduction and stimulates inflammatory and procoagulant responses in the endothelium.

MATERIALS AND METHODS
Cells and Cell Cultures-Polymorphonuclear leukocytes (PMN) were isolated from heparin sodium-anticoagulated blood drawn after informed consent from normal healthy volunteers by differential centrifugation on Ficoll/Hypaque gradient and dextran sedimentation as described (16). Human umbilical vein EC were prepared by collagenase treatment, maintained in medium 199 (BioWhittaker, Walkersville, MD) supplemented with 20% heat-inactivated fetal bovine serum (Bio-Whittaker), L-glutamine (2 mM), and 1% endothelial cell growth supplement, pH 7.4, and used between passages 2 and 4.
Flow Cytofluorimetric Detection of MP in the Normal Circulation-Heparin sodium-anticoagulated blood was drawn from normal healthy volunteers after informed consent. Aliquots (1.5 ml) of undiluted blood or samples diluted 1:5 in PBS, pH 7.4, were incubated with the fluorescent labeling dye, quinacrine mustard (mepacrine, 0.1 mM; Sigma), in the presence or in the absence of fMLP (1 M) for 2 h at 37°C. Blood samples were centrifuged at 1500 ϫ g for 20 min at 22°C, and the cell-free supernatant was collected and analyzed by flow cytometry. In some experiments, cell-free supernatants prepared as described above were passed through 100-kDa cut-off nanospins (Millipore, Bedford, MA) according to the manufacturer's specifications. Aliquots of 0.5 ml were analyzed on a Becton-Dickinson (Mountain View, CA) flow cytometer as described (16). To identify the MP population, samples were gated according to their fluorescence (mepacrine) labeling. With the light scatter and fluorescence channels set at logarithmic gain, samples were analyzed for 6-s intervals for forward light scatter, right angle light scatter, and mepacrine fluorescence. The light scatter profile of mepacrine-positive MP typically demonstrated one single low light scatter population, in agreement with previous observations (16). Indistinguishable results were obtained when blood samples were anticoagulated with EDTA or acid citrate dextrose. In some experiments, aliquots of diluted blood were incubated with 5 g/ml FITC-conjugated anti-GPIb␣ mAb (the generous gift of Dr. Z. M. Ruggeri, Scripps Research Institute), anti-lactoferrin antibody (ICN Biomedicals, OH) or control mouse IgG (Roche Molecular Biochemicals, IN), for 30 min at 22°C in the dark. FITC conjugation was carried out according to the manufacturer's specifications (Roche Molecular Biochemicals). Samples were then stimulated or not with 200 nM fMLP and 2.5 mM CaCl 2 for 1 h at 22°C and centrifuged at 1500 ϫ g for 20 min to remove the various cellular fractions, and the resulting supernatants were collected and ultracentrifuged at 100,000 ϫ g for 1 h. Pellets containing MP were washed once with PBS, pH 7.4, suspended in 0.5 ml of PBS, and analyzed by flow cytometry. Samples were analyzed for a total of 2000 events for forward, side, and fluorescein light scatters as described above. In parallel experiments, supernatants collected after incubation with the various unconjugated antibodies as described above were tested for induction of endothelial IL-6 release by ELISA (see below).
In Vitro MP Analysis-PMN-derived MP were quantitated by flow cytometry using a fluorescent lipid intercalating dye, PKH26-GL (Sigma). This aliphatic chromophore partitions into lipid bilayers and confers a red fluorescence. PMN (2 ϫ 10 7 cells/ml) were labeled with PKH26-GL (4 M) according to the manufacturer's specifications. Labeled cells were suspended in serum-free 199 medium and preincubated with or without 100 ng/ml pertussis toxin (Calbiochem, San Diego, CA) for 2 h at 37°C. Cells were stimulated with fMLP (10 nM, Sigma) or recombinant human IL-8 (Endogen, Woburn, MA; 100 ng/ml) for a further 2 h at 37°C in the presence of 2.5 mM CaCl 2 . The cell-containing supernatants were isolated and analyzed by flow cytometry, as described above. Each sample was analyzed for a total of 10,000 events. A gate was chosen to include particles distinctly positive for red fluorescence.
Preparation of Platelet-rich Plasma (PRP) and Platelet MP-Blood was collected from healthy volunteers into a plastic syringe and anticoagulated with acid citrate dextrose. PRP was prepared by centrifugation at 120 ϫ g for 15 min at 22°C. Platelet MP were isolated after platelet aggregation induced by 1 unit/ml thrombin (Sigma) and 2.5 mM CaCl 2 with gentle shaking (14). After a 10-min incubation at 37°C, large platelet aggregates were sedimented at 1500 ϫ g for 20 min, and the MP-containing supernatants were collected, diluted 1:5 in PBS, pH 7.4, and analyzed by flow cytometry, as described above.
EC Cytokine Release by MP-200-l aliquots of unstimulated, fMLP (1 M), or thrombin (1 unit/ml)-stimulated cell-free supernatants from whole blood or PRP samples were added to quiescent EC monolayers grown to confluency in 48-well plates. After a 12-h incubation at 37°C, EC supernatants were collected, centrifuged at 200 ϫ g for 10 min, and analyzed for released IL-6 and IL-8 by ELISA (Endogen), as described previously. In other experiments, aliquots of whole blood were incubated in the absence or presence of 20 g/ml anti-GPIb␣ mAb, antilactoferrin polyclonal antibody, or control 14E11 mAb, for 30 min at 22°C. After fMLP stimulation, cellular fractions of the samples were removed by centrifugation at 1500 ϫ g for 20 min, and 200-l aliquots of the resulting supernatants were added to EC monolayers for determination of cytokine release, as described above. In control experiments, cell-free supernatants from fMLP (1 M)-stimulated PMN (3 ϫ 10 6 /ml) were preincubated with neutralizing anti-TNF mAb (Sigma) or control mAb 14E11 (25 g/ml) for 30 min at 22°C before addition to EC for and determination of IL-6 release, as described above. In other experiments, cell-free supernatants from unstimulated or fMLP-stimulated blood samples were collected and sterile-filtered through 100-kDa cut-off filter membrane of nanospin units (Millipore), before addition to EC and determination of released IL-6 and MCP-1.
TF Expression in MP-stimulated EC-For cytofluorimetric analysis of TF induction in EC, cell-free supernatants from resting or fMLP (1 M)-stimulated PMN were added to confluent EC monolayers grown in a 6-well plate for 6 h at 37°C. Unstimulated or MP-stimulated EC were harvested, washed, and blocked with 50% human serum for 30 min on ice followed by washing and staining with 20 g/ml anti-TF mAb 5G9 (the generous gift of Dr. T. S. Edgington, The Scripps Research Institute, La Jolla, CA) or control mAb 14E11. After washing, EC were incubated with a 1:20 dilution of FITC-conjugated goat anti-mouse IgG and immediately analyzed by flow cytometry. In control experiments, EC were stimulated with 10 ng/ml TNF␣ and analyzed for TF induction as described above. In parallel experiments, EC in 6-well plates at 90% confluence were incubated with medium alone or with 10 ng/ml TNF or cell-free supernatant from fMLP-stimulated PMN (4 ϫ 10 7 /ml) for 4 h at 37°C. Total RNA was extracted by RNAzol B method (TEL TEST, Friendswood, TX) and reverse transcribed with 1 ϫ 10 4 unit/ml Super-Script II RNase H Ϫ reverse transcriptase (Life Technologies, Inc.) using oligo(dT) or gene-specific primer (5Ј-CACTCCTGCCTTTCTACAC-3Ј). The reverse transcription reaction containing 5 g of total EC RNA or 50 ng of control RNA, 0.5 g of oligo(dT) or 2 M GSP, 1 ϫ 10 4 unit/ml reverse transcriptase in the presence of 25 mM MgCl 2 , 10 mM dNTP mix, and 0.1 M dithiothreitol was incubated for 50 min at 42°C. At the end of the incubation, samples were heated for 15 min at 70°C, chilled on ice, mixed with 1 l of RNase H for 20 min at 37°C, and amplified by polymerase chain reaction with oligonucleotides 5Ј-GTCAGAAGGAA-CAACACT-3Ј (forward) and 5Ј-CACTCCTGCCTTTCTACAC-3Ј (reverse) derived from the sequence of human TF. Thirty-five cycles of amplification were carried out in a Perkin-Elmer 480 thermal cycler with denaturation for 30 s at 94°C, annealing for 40 s at 55°C, and extension for 1 min at 72°C. MgCl 2 was used at a final concentration of 1.5 mM. Polymerase chain reaction products were analyzed on 1% agarose gels by ethidium bromide staining.
In another series of experiments, EC in a 48-well plate were washed and incubated in the presence or absence of 10 ng/ml TNF␣, cell-free supernatant from fMLP-stimulated PMN (4 ϫ 10 7 /ml), or equivalent purified MP prepared as described above for 6 h at 37°C. After washes, EC were incubated with 100 l of phenol red-free RPMI 1640 medium containing 150 nM human factor X (Alexis), 5 nM activated factor VII (Alexis), and 2 mM CaCl 2 for 20 min at 37°C. In some experiments, EC monolayers were preincubated with anti-TF mAb 5G9 or control mAb 14E11 (20 g/ml) for 30 min at 22°C before addition of factors X and VIIa. Generation of activated factor X (factor Xa) was stopped by addition of 10 mM EDTA to each incubation reaction. Samples were transferred to a 96-well plate, and factor Xa activity was determined by hydrolysis of the chromogenic substrate S-2222 (Chromogenix, Moelndal, Sweden) at a final concentration of 0.2 mM. The optical densities were read at 405 nm using a plate-reader spectrophotometer (Thermomax, Molecular Devices). The amount of factor Xa generated under the various conditions tested was calculated by comparison with a standard curve constructed with serial increasing concentrations of factor Xa (Alexis).
For electrophoretic mobility shift assays, confluent EC monolayers were either left untreated (control) or incubated with PMN (4 ϫ 10 7 /ml) supernatant or TNF␣ (10 ng/ml) for 90 min at 37°C. Nuclear extracts were prepared as described previously (21). Nuclear extracts were normalized for protein concentration, and 10 g were incubated with 4 g of poly(dI⅐dC) (Amersham Pharmacia Biotech) for 20 min at 22°C and then for another 15 min with a 32 P-labeled probe in a final volume of 15 l. The oligonucleotide used in these studies for the B binding was 5Ј-AGTTGAGGGGACTTTCCCAGGC-3Ј. Double-stranded oligonucleotide probe was end-labeled with [␥-32 P]ATP (Amersham Pharmacia Biotech) using T4 polynucleotide kinase (New England Biolabs, Beverly, MA). DNA-protein complexes were resolved on a 6% nondenaturing polyacrylamide gel containing 7.5% glycerol in 0.25% Tris borate-EDTA buffer, pH 8.3 (1ϫ TBE: 89 mM Tris borate, 89 mM boric acid, 20 mM EDTA). Dried gels were exposed to x-ray film (Kodak Co., Rochester, NY), and relevant bands were visualized by autoradiography.
Detection of MP in Vivo-Flow cytofluorimetric analysis of whole blood supernatants revealed the constitutive presence of a discrete population with low forward scatter consistent with MP ( Fig. 2A, top left quadrant). Stimulation of blood samples with fMLP resulted in an ϳ2-fold increase in the MP population ( Fig. 2A, top right quadrant), in agreement with previous observations (16). In parallel experiments, fMLP stimulation increased the number of fluorescent MP following mepacrine labeling of whole blood samples ( Fig. 2A, lower left and right  quadrants). Consistent with these in vivo data, cytofluorimetric analysis of cell-free supernatants collected from 23 normal volunteers revealed the constitutive presence of an MP population in nonmanipulated blood samples (events, 2799 Ϯ 360). Stimulation of whole blood samples with fMLP resulted in a ϳ2-fold increase in the number of MP detected by flow cytometry (events, 5241 Ϯ 640; p ϭ 0.001; Fig. 2B).
Modulation of EC Activation by Blood-derived MP-Incubation of quiescent EC with cell-free supernatants from unstimulated whole blood samples did not result in release of cytokines, IL-6, and MCP-1 (Fig. 3A). In contrast, cell-free supernatants from fMLP-stimulated whole blood supernatants caused a 2-4-fold increased release of cytokines MCP-1 (1282 Ϯ 79 pg/ml) and IL-6 (377 Ϯ 68 pg/ml), respectively (Fig. 3A). A 2-fold increase in IL-8 release was also observed under the same experimental conditions (not shown). Filtration of cell-free supernatants through 100-kDa filters prior to incubation with EC completely abolished the MP-stimulated release of IL-6 and MCP-1 by EC (Fig. 3A), in agreement with previous observations (16). Consistent with these findings, filtration of cell-free supernatants resulted in complete depletion of the blood MP population, identified by forward and side scatter parameters and quantitated by flow cytometry (Fig. 3B). In control experiments, preincubation of PMN-derived MP with a neutralizing mAb to TNF␣ did not significantly reduce EC IL-6 release (unstimulated, 28 pg/ml; MP-stimulated plus control mAb 14E11, 422 pg/ml; MP-stimulated plus anti-TNF␣ mAb, 378 pg/ml). In contrast, the anti-TNF␣ antibody inhibited by ϳ50% EC IL-6 release induced by TNF␣ stimulation (TNF␣ alone plus control mAb 14E11, 615 pg/ml; TNF␣ plus anti-TNF␣ mAb, 305 pg/ml).
Induction of EC Procoagulant Activity by Circulating MP-A 6-h exposure of EC to supernatants from fMLP-stimulated PMN resulted in moderate but consistent increased surface expression of TF, as determined by flow cytofluorimetric staining with mAb 5G9 (Fig. 4A). Similar results were obtained using whole blood-containing MP (not shown). EC stimulation with blood-derived MP also resulted in de novo expression of TF mRNA, as determined by appearance of a 352-base pair TF RNA transcript detected by reverse transcriptase-polymerase chain reaction amplification of EC RNA (Fig. 4A, inset). In control experiments, EC stimulation with TNF␣ resulted in maximal increase in TF surface expression and TF RNA (Fig.  4A), in agreement with previous observations (23). In parallel experiments, EC stimulation with MP-containing PMN supernatant, or purified MP resulted in the generation of 2.28 Ϯ 0.56 and 2.75 Ϯ 0.07 nM of factor Xa/min/10 4 EC, respectively (Fig.  4B). This response was abolished in the absence of factor VIIa or by addition of anti-TF mAb 5G9 (0.51 Ϯ 0.07 nM of factor Xa/min/10 4 EC) (Fig. 4B). In contrast, EC preincubation with control mAb 14E11 did not reduce EC procoagulant activity (Fig. 4B). In control experiments, EC stimulation with 10 ng/ml TNF␣ resulted in the formation of 6 Ϯ 1.5 nM factor Xa/min/10 4 EC (not shown), in agreement with previous observations (23).
Characterization of Circulating MP by Flow Cytometry-To identify the potential cell(s) of origin of released MP in vivo, flow cytofluorimetric experiments were carried out with antibodies to platelet GPIb␣ or leukocyte lactoferrin. Tryptic digestion and ionization mass spectrometry had previously identified the main ϳ85-kDa component in purified MP (16) as lactoferrin. 2 In these experiments, unstimulated blood-derived MP reacted with antibodies to GPIb␣ and lactoferrin, as compared with control nonbinding antibody (Fig. 5). Furthermore, fMLP stimulation increased by ϳ10-fold the reactivity of the MP population with the lactoferrin antibody, whereas no significant differences in the binding of anti-GPIb␣ mAb were observed with unstimulated or fMLP-stimulated samples (Fig. 5).
Functional Dissociation between Platelet and Leukocyte MP in EC Stimulation-Preincubation of blood samples with 20 g/ml anti-GPIb␣ mAb resulted in a significant reduction in the amount of MP released after fMLP stimulation (Fig. 6A). In contrast, no differences in fMLP-induced MP release were observed in the presence of control mAb 14E11 or the antibody to lactoferrin (Fig. 6A). In parallel experiments, preincubation of cell-free supernatants with antibodies to lactoferrin or GPIb␣ did not reduce EC IL-6 release, as compared with untreated samples or treated with control mAb 14E11 (Fig. 6B). A potential differential ability of platelet or leukocyte MP to induce EC IL-6 release was further investigated. Thrombin stimulation of PRP resulted in MP release, as determined by flow cytometry (Fig. 6C), in agreement with previous observations. However, thrombin-stimulated platelet MP or unstimulated PRP supernatant did not stimulate EC IL-6 release ( Fig. 6D and data not  shown). In contrast, TNF␣ stimulation of EC resulted in the generation of 599 Ϯ 142 pg/ml IL-6 after a 12-h culture at 37°C (Fig. 6D).
Signal Transduction Initiated by PMN-derived MP in EC-EC stimulation with TNF␣ resulted in strong activation of NF-B, as compared with unstimulated EC extracts, as determined by electrophoretic mobility shift assay (Fig. 7A). In contrast, leukocyte MP did not induce NF-B activation ( 7A). Similarly, no significant differences in tyrosine phosphorylation of ERK1 were observed in EC treated with leukocyte MP, as compared with unstimulated or serum-containing cultures (Fig. 7B). In contrast, immunoprecipitation and Western blot of EC extracts with anti-phosphotyrosine antibody Py20 revealed a time-dependent and sustained increased tyrosine phosphorylation of a ϳ46-kDa band in MP-stimulated EC, as compared with untreated cultures (Fig. 7B). The identity of the ϳ46-kDa band was further investigated. Immunoprecipitation of EC lysates with an antibody to JNK1 followed by Western blot with Py20 antibody revealed a Ͼ2-fold increase in tyrosine phosphorylation of JNK1 in MP-stimulated cultures but not in control untreated EC (Fig. 7C). In contrast, EC serum treatment did not result in tyrosine phosphorylation of immunoprecipitated JNK1 (Fig. 7C). In control experiments, Western blot of JNK1 immunoprecipitates demonstrated a comparable amount of JNK1/lane, under the various conditions tested (Fig.  7C).

DISCUSSION
In this study, we have shown that leukocyte-derived MP circulate in the bloodstream under normal conditions and are rapidly up-regulated by inflammatory stimulation. Secondly, leukocyte, but not platelet, MP stimulate release of inflammatory/chemotactic cytokines and expression of functional tissue factor in EC, in a pathway associated with sustained phosphorylation of ϳ46-kDa JNK1.
The notion that vascular cells release MP in response to disparate environmental stimuli is well established and has been experimentally validated for platelets (24,25), monocytes (26), and endothelium (27). Through their ability to assemble a functional prothrombinase complex, platelet (26,28) and monocyte (29) MP may amplify cellular procoagulant responses, thus potentially contributing to aberrant fibrin deposition in vivo (30,31). However, it has also been recently proposed that platelet (14,15), and leukocyte (16), MP may function as gen- FIG. 4. TF induction in MP-stimulated EC. A, EC were stimulated with cell-free supernatants from fMLP (1 m)-stimulated PMN (2 ϫ 10 7 /ml) or 10 ng/ml TNF␣ for 6 h at 37°C. EC were harvested, washed, and blocked with 50% human serum followed by staining with 20 g/ml anti-TF mAb, 5G9. After washes, EC were incubated with FITC-conjugated goat anti-mouse IgG and analyzed by flow cytometry. Data are representative of one experiment out of three independent determinations. Inset, TF RNA transcripts in unstimulated cultures or EC treated with 10 ng/ml TNF␣ or leukocyte MP (PMN Sup.) were detected by reverse transcriptase-polymerase chain reaction amplification of EC RNA with analysis of amplified bands by ethidium bromide staining. B, EC were incubated in the presence or in the absence of cell-free supernatant from fMLP-stimulated PMN (5 ϫ 10 7 /ml) and purified MP and with or without preincubation with control mAb 14E11 or anti-TF mAb 5G9 for 6 h at 37°C. EC were washed and incubated with 150 nM factor X, 5 nM factor VIIa, and 2 mM CaCl 2 for 20 min at 37°C. The reaction was terminated by addition of 10 mM EDTA, and factor Xa activity was determined by hydrolysis of the chromogenic substrate S-2222 at 405 nm. Data are the means Ϯ SD of three independent experiments. uine cellular agonists and stimulate complex EC responses. In this context, platelet MP induced COX-2 and prostacyclin production in EC (14) and stimulated increased monocyte adherence through up-regulation of adhesion molecule ICAM-1 (15). A similar paradigm has been also proposed for leukocyte MP for their ability to stimulate EC expression of adhesion molecules ICAM-1, E-selectin, and VCAM-1 and promote release of cytokines IL-6 and IL-8 (16). In expanding these observations, leukocyte MP released by inflammatory/chemotactic mediators fMLP or IL-8 in a pertussis toxin-insensitive pathway mediate a procoagulant response in EC by increasing the expression of TF mRNA and up-regulating functional TF at the cell surface. This is consistent with the ability of leukocyte MP to directly affect EC gene expression, as reflected by the ϳ18-fold upregulation of IL-6 mRNA, under these experimental conditions (16). Combined with the presence of leukocyte MP in vivo, and their rapid quantitative up-regulation by inflammatory stimuli, these data suggest that leukocyte MP may cooperate with locally released cytokines (2) and leukocyte-EC intercellular signaling (7, 32) to stimulate a broad proadhesive, procoagulant, and proinflammatory phenotype in EC. This pathway may be potentially relevant to EC dysfunction during vascular diseases, invariably associated with increased leukocyte recruitment, platelet activation, and fibrin deposition at the site of vascular injury (13).
As shown here, this mechanism of EC stimulation appears selective for leukocyte MP, because depletion of platelet MP did not decrease cytokine induction in EC, and thrombin-stimulated platelet MP did not stimulate EC IL-6 release. Although induction of EC ICAM-1 by platelet MP required treatment with phospholipase A 2 (14, 15), a potential role of platelet MP on EC cytokine release has not been previously investigated (15). In contrast, EC stimulation by leukocyte MP did not require phospholipase A 2 treatment, and similarly, arachidonic acid failed to stimulate IL-6 release, 2 whereas it recapitulated EC activation by platelet MP (15). Altogether, these data suggest that EC activation by platelet (14,15) or leukocyte (16) MP may involve separate or only partially overlapping signaling pathways. On the other hand, fMLP stimulation was also associated with platelet MP release in vivo, in a reaction inhibited by an antibody to GPIb␣. This suggests that inflammatory challenges, directly or through intercellular collaboration (33), FIG. 6. Differential EC activation by leukocyte or platelet MP. A, diluted blood samples were incubated with 20 g/ml aliquots of control mAb 14E11 or antibodies to GPIb␣ or lactoferrin for 30 min at 22°C before stimulation with 1 M fMLP for 1 h at 22°C. Samples were centrifuged at 1500 ϫ g for 20 min, and the cell-free supernatants were analyzed for MP content by flow cytometry. B, the experimental conditions are the same as those in A except that antibody-treated supernatants from blood samples were added to EC monolayers for 12 h before determination of IL-6 release by ELISA. C, diluted PRP samples were incubated in the absence or presence of 1 unit/ml thrombin and 2.5 mM CaCl 2 for 10 min at 37°C with gentle shaking. After platelet aggregation, PRP samples were centrifuged at 1500 ϫ g for 20 min, and platelet-free supernatants were analyzed for MP content by flow cytometry. D, the experimental conditions are the same as those in C except that supernatants from thrombin-stimulated PRP were added to EC monolayers for 12 h before determination of IL-6 release by ELISA. may result in both platelet and leukocyte MP release, thus further amplifying EC procoagulant and proinflammatory gene expression (14 -16).
In investigating potential downstream signals of this EC activation pathway, we found that leukocyte MP did not stimulate NF-B activation or promote ERK1 tyrosine phosphorylation. In contrast, MP stimulated a prominent and sustained tyrosine phosphorylation of c-Jun NH 2 -terminal kinase, JNK1, in EC. A member of the mitogen-activated kinase gene family, JNK1 phosphorylation through upstream activators SEK and MEKK has been observed in response to growth factors, cytokines, and stress signals, including ultraviolet lights or alkylating agents (34 -37). JNK1 phosphorylation has also been implicated in apoptosis of neuronal PC12 cells after nerve growth factor deprivation, TNF␣ treatment, or ischemia (34,38) and in stimulatory phosphorylation of c-Jun, a component of the transcription factor, AP-1 (39). This suggests that MP may target JNK1 activation in the context of a broad EC stress signaling response, culminating with proinflammatory, proadhesive, and procoagulant gene expression through AP-1 activation. In this context, it is of interest that AP-1 has been implicated in regulated transcription of the TF gene in endothelium (40), and in platelet-derived growth factor-dependent IL-6 gene transcription in osteoblasts (41). Whether this pathway may also influence EC apoptosis (42), thus further exacerbating vessel wall dysfunction and procoagulant activity (43), is cur- FIG. 7. Stress signal transduction pathway initiated by leukocyte MP in EC. A, EC were either left untreated or treated with PMN supernatant or TNF␣ (10 ng/ml) for 90 min at 37°C. 10-g nuclear extracts were incubated with 4-g poly(dI⅐dC) for 20 min at 22°C and for an additional 15 min with a 32 P-labeled NF-B oligonucleotide probe in a final volume of 15 l. DNA-protein complexes were resolved on a 6% nondenaturing polyacrylamide gel and exposed to autoradiography. B, EC in 6-well plates were stimulated with medium (Control), serum, or PMN supernatants for the indicated increasing time intervals at 37°C, before detergent solubilization and centrifugation at 14,000 ϫ g for 30 min. EC extracts were precleared with protein A-Sepharose and immunoprecipitated (IP) with antibody to ERK1 (left panel) or to phosphotyrosine proteins Py20 for 2 h at 4°C (right panel). The immune complexes were washed six times in lysis buffer, electrophoresed on a 10% SDS gel, electroblotted to nylon membranes, and immunoblotted with Py20 antibody followed by alkaline phosphatase-conjugated goat anti-mouse IgG and visualization of relevant bands by chemiluminescence. F-PMN Sup., filtered supernatant through a 100-kDa cut-off nanospin. C, the experimental conditions are the same as those in B, except that EC stimulated with medium (Control), serum, or PMN supernatant were immunoprecipitated with an antibody to JNK1. The immune complexes were immunoblotted with Py20 antibody to phosphotyrosine proteins (left panel) or with JNK1 antibody (right panel) as a control for protein loading. Relevant bands were detected by chemiluminescence. Densitometric quantitation of JNK1 tyrosine phosphorylation under the various conditions tested is shown at the bottom. For all panels, data are form a representative experiment out of at least three independent determinations. rently unknown and may depend on the differential expression of AP-1 components or activation of other modulatory transcription factors (34,39).
In summary, we have identified a pathway of EC signaling and gene expression centered on the release of leukocyte MP. Future studies will dissect the molecular requirement(s) of this stress response mechanism and its potential contribution to EC dysfunction in vascular diseases (13).