β2-Integrins and Acquired Glycoprotein IIb/IIIa (GPIIb/IIIa) Receptors Cooperate in NF-κB Activation of Human Neutrophils*♦

Microparticles from various cells are generated during inflammation. Platelet-derived microparticles (PMPs) harbor receptors that are not genuinely expressed by neutrophils. We tested whether or not functional glycoprotein IIb/IIIa (GPIIb/IIIa) receptors can be acquired by neutrophils via PMPs and whether these receptors participate in pro-inflammatory signaling. Surface expression was analyzed by flow cytometry and confocal microscopy. NF-κB activation was analyzed by Western blot experiments, electrophoretic mobility shift assays, and reverse transcription-PCR. Cell adhesion and spreading were estimated by myeloperoxidase assay and light microscopy. We found that PMPs transfer GPIIb/IIIa receptors to isolated and whole blood neutrophils via PMPs. We used specific antibodies in granulocyte macrophage colony-stimulating factor-treated neutrophils and observed that acquired GPIIb/IIIa receptors co-localized with β2-integrins and cooperated in NF-κB activation. We show that Src and Syk non-receptor tyrosine kinases, as well as the actin cytoskeleton, control NF-κB activation. In contrast to NF-κB, acquisition of GPIIb/IIIa receptors was not necessary to induce adhesion to fibronectin or phosphatidylinositol 3-kinase/Akt signaling. When granulocyte macrophage colony-stimulating factor-stimulated neutrophils were incubated on fibronectin, strong NF-κB activation was observed, but only after loading with PMPs. Blocking either β2-integrins or GPIIb/IIIa receptors abrogated this effect. Therapeutic GPIIb/IIIa inhibitors were similarly effective. The compounds also inhibited NF-κB-dependent tumor necrosis factor-α mRNA up-regulation. The data implicate GPIIb/IIIa receptors as new therapeutic targets in neutrophil-induced inflammation.

Cells send and receive signals to communicate with their environment. In addition to soluble mediators such as cytokines and growth factors, cells may also generate membranederived microparticles. Conditions that lead to release of microparticles include cell activation, apoptosis, oxidative stress, and shear stress (1). The composition of these microparticles varies from cell type to cell type. The microparticle cargo determines the biological consequences. For example, microparticles are able to transfer receptors to cells that do not normally express a particular receptor. The acquired receptors may be functional, as demonstrated for the CCR5 chemokine receptor released by lymphocytes. Acquisition of CCR5 receptors by endothelial cells rendered these cells susceptible to human immunodeficiency virus infection (2). Another example for the generation of microparticles is inflammatory vascular injury, including acute coronary syndromes and acute vasculitis. Both conditions are associated with an increase in circulating PMPs 2 (3,4). Circulating PMPs can be acquired by components of the vessel wall and by blood-borne cells, such as neutrophils (5).
Neutrophils participate in non-infectious inflammatory processes including the acute coronary syndrome and systemic necrotizing vasculitis (6). The nuclear factor B (NF-B) is important for neutrophil survival and the generation of numerous inflammatory mediators (7). We tested the hypothesis that GPIIb/IIIa receptors can be transferred to human neutrophils via PMPs and that these receptors participate in the NF-B activation when neutrophils are challenged by cytokines. We found that acquired GPIIb/IIIa receptors cooperate with ␤2-integrins, allowing for NF-B activation when GM-CSF-treated neutrophils interact with fibronectin, but not in suspension cells. The specific GPIIb/IIIa blockers abciximab, eptifibatide, and tirofiban abrogated NF-B activation. We suggest that GPIIb/IIIa receptors provide a new therapeutic target in neutrophil-induced inflammation.
In experiments where we saw no effect with these low concentrations, we used 1000 ng/ml abciximab, 2500 ng/ml eptifibatide, and 60 ng/ml tirofiban, respectively. These concentrations are at the upper limit of therapeutic concentrations used in vivo. Endotoxin-free reagents and plastic disposables were used in all experiments.
Preparation and Culture of Human Neutrophils-Heparinized whole blood was drawn from healthy human donors after written approval (EA 3/011/06, Charité Berlin). Neutrophils were isolated using a double gradient method. The gradient was formed by layering an equal volume of Histopaque 1077 over Histopaque 1119. Whole blood was carefully layered onto the upper gradient and centrifuged for 30 min. Cells of the granulocytic series were found at the Histopaque 1077/1119 interphase, whereas lymphocytes, other mononuclear cells, and platelets were found at the plasma/Histopaque 1077 interphase. The granulocytic cells were carefully aspirated and washed once with phosphate-buffered saline. Remaining erythrocytes were lysed by incubation with hypotonic saline for 15 s. PMN were spun down (1050 rpm, 10 min) and reconstituted in HBSS with calcium and magnesium. The final cell concentration was 5 ϫ 10 6 cells/ ml. The cell viability was detected in every cell preparation by trypan blue exclusion and found to be Ͼ99%. With this method, we were able to isolate neutrophils with a very low amount of CD41/61-positive cells.
Preparation of Platelet-rich Plasma (PRP), PMPs, and Platelet-free Plasma-PRP was obtained by centrifugation of heparinized whole blood at 200 ϫ g for 10 min at room temperature. PMPs were generated using previously reported protocols with minor modifications (15,16). Briefly, PRP was activated with 10 M ADP for 20 min at 37°C, and whole platelets were removed by centrifugation (twice at 4600 rpm, 13 min) and filtration through a 0.8-m filter. To remove ADP, this solution was centrifuged for 60 min at 20,000 ϫ g. The pellet containing PMPs was resuspended in HBSS ϩϩ . As a negative control, we also prepared platelet-free plasma that was ADP-treated, filtered, and centrifuged in the same way as described above.
Preparation of PMP-loaded Neutrophils in Vitro-PMPloaded neutrophils were prepared by co-incubation of isolated human neutrophils with PMPs for 45 min at 37°C. PMN were spun down (200 ϫ g, 7 min) and resuspended in HBSS ϩϩ with a final cell concentration of 5 ϫ 10 6 cells/ml.
Western Blot-Samples were incubated for 5 min at 95°C in loading buffer (250 mM Tris-HCl, pH 6.8, with 4% SDS, 20% glycerol, 0.01% bromphenol blue, 10% ␤-mercaptoethanol). 10 -20 g of protein was loaded per lane, electrophoresed on a 10% SDS-polyacrylamide gel, and transferred to a nitrocellulose membrane. The membrane was blocked with Tris-buffered saline with Tween, 5% skim milk for 1 h and incubated overnight with the indicated antibodies. Membranes were washed and incubated with a horseradish peroxidase-labeled secondary antibody. The blot was developed by incubation in a chemiluminescence substrate (ECL, Amersham Biosciences) and exposed to an x-ray film. We confirmed equal loading of protein in parallel experiments using ␣-actin or total Akt antibodies.
Flow Cytometry-CD41/61 receptor expression on freshly isolated neutrophils was determined by incubation with saturating concentration of the antibody (Clone A2A9/6, Santa Cruz Biotechnology) for 25 min on ice. A mouse IgG 2a antibody was used in the same concentration as a control (Santa Cruz Biotechnology). Cells were washed and resuspended in HBSS ϩϩ , and phycoerythrin-conjugated F(abЈ) 2 fragments of goat-anti mouse IgG (DAKO, Hamburg, Germany) were added in saturating concentration for a further 25 min. After washing, samples were assayed using a FACScan (BD Biosciences, Heidelberg, Germany).
CD41/61 receptor expression on whole blood neutrophils was determined by stimulation of whole blood for 20 min with 10 M ADP or phosphate-buffered saline, respectively, and further incubation with saturating concentration of the antibody for 20 min at room temperature. Erythrocytes were lysed by incubation with FACS lysing solution (BD Biosciences) for 10 min at room temperature. After centrifugation and washing, samples were assayed using a FACScan.
In experiments where CD61 receptor expression was determined on PMP-loaded neutrophils after incubation with the indicated inhibitory antibodies, we first incubated PMP-loaded neutrophils with 20 g/ml of the antibodies for 30 min at 37°C. The cells were washed and incubated with saturating concentration of a CD61-FITC antibody (Clone Y2/51, DAKO, Glastrup, Denmark) for a further 25 min on ice. After washing, samples were assayed using a FACScan (BD Biosciences).
Fluorescence Microscopy-After incubation with the respective antibody as described for flow cytometry, cells were spun down on poly-L-lysine coated coverslips and incubated with 4Ј,6-diamidino-2-phenylindole dihydrochloride (DAPI; Sigma; 0.25 g/ml, 30 s) for nuclei counterstaining. Immunostained specimen were analyzed with a Leica DMRB light microscope (Leica Microsystems, Bensheim, Germany). Images were acquired using a digital camera (SPOT) and processed with MetaVue software.
Confocal Analysis-For double-staining experiments, cells were immunolabeled with a monoclonal antibody against CD41/61, washed, and incubated with a Alexa Fluor 555 donkey anti-mouse IgG antibody (Molecular Probes, Eugene, OR). Cells were washed again and incubated with a direct-conjugated CD18-FITC antibody (Clone 7e4, Immunotech, Marseille, France). After incubation and washing, cells were spun down on coverslips. Immunofluorescence images were scanned with a three-channel confocal laser scanning microscope with a ϫ40/1.25 NA oil immersion objective (Leica TCS SP-2, Leica Microsystems). All digitized images were analyzed using the Leica confocal software.
Nuclear Extract Preparation-Nuclear extracts were prepared as described previously (7). The reaction was stopped with ice-cold HBSS ϩϩ supplemented with diisopropyl fluorophosphate (2 mM final concentration). Cells were centrifuged at 2000 ϫ g for 2 min at 4°C, and pellets were resuspended with the hypotonic buffer A (HEPES 10 mM, pH 7.5, KCl 10 mM, EGTA 0.1 mM, EDTA 0.1 mM, pH 8.0) containing an anti-protease mixture (2 mM diisopropyl fluorophosphate, 1 mM 4-(2aminoethyl) benzenesulfonyl fluoride hydrochloride, 1 mM phenylmethylsulfonyl fluoride, 1 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin A, 0.5 mM benzamidine, 1 mM dithiothreitol). Cells were kept on ice for 10 min. Cells were vortexed briefly and centrifuged (1000 ϫ g, 10 min, 4°C) to pellet the nuclei and to remove intracellular granules that may cause nuclear degradation. The pellet was washed again with buffer A, and the hypertonic, high salt buffer C was added (HEPES 20 mM, pH 7.5, NaCl 0.4 M, EDTA 1 mM, EGTA 1 mM, glycerol 20%) together with the above mentioned protease inhibitor mixture. After centrifugation (10 min at 14,000 ϫ g, 4°C), the supernatant was collected and referred to as the nuclear fraction. Protein measurements were done using the Bio-Rad assay (Bio-Rad, Munich, Germany).
Electrophoretic Mobility Shift Assay (EMSA)-EMSAs were performed as described previously (7). Briefly, nuclear extracts (5 mg of protein) were incubated with 20,000 cpm of a 32 Plabeled H 2 K probe. Incubations were performed for 30 min at room temperature in the presence of poly(dI-dC) and 20 mM HEPES, containing 60 mM KCl, 4% Ficoll, 5 mM dithiothreitol, and 0.5 mg/ml nuclease-free bovine serum albumin. Probes were subjected to electrophoresis on native 5% polyacrylamide gels and autoradiographed.
For quantification of the amount of RNA present in the various samples, the fluorescence signal was measured at each PCR cycle, and the increase in the fluorescence normalized reporter signal (R n ) was documented in an amplification plot. Using non-template controls, the threshold was set in the log phase to subtract unspecific fluorescence signals. Cycle threshold (C T ) values were determined for each sample. In short, the C T value difference (⌬C T ) was used to calculate the factor of differential expression (2 ⌬ CT ). Results were imported in an Excel spreadsheet and analyzed according to the standard curve method.
Spreading and Adhesion Assay-We used 96-well plates coated with fibronectin (10 g/cm 2 ) for the adhesion assay. 1 ϫ 10 5 neutrophils in 100 ml of HBSS ϩϩ were either left untreated or were treated with 20 ng/ml GM-CSF. Plates were incubated at 37°C in 5% CO 2 for 60 min. Wells were flicked dry and washed three times with phosphate-buffered saline, and adherent cells were estimated using the myeloperoxidase assay. Briefly, adherent cells were lysed in 100 l of 0.5% Triton-X-100 for 10 min. 100 l of substrate (2,2Ј-azino-bis 3-ethylbenzthiazoline-6-sulfonic acid, Sigma) was added, and OD was read after 10 min at 450 nm with a microtiter plate reader. Each experiment was done in triplicate. OD of the experimental sample was compared with a standard curve that showed an excellent correlation between OD and cell number. The standard curve was established over a range of 1 ϫ 10 4 -1 ϫ 10 5 cells and showed an R 2 value of 0.96.
For estimation of cell spreading, neutrophils were cultured as described above. At the indicated time point, non-adherent cells were discarded, and the percentage of spread cells was assessed. At least 100 cells were counted using phase contrast microscopy, and those cells that were phase dark, enlarged, and irregular were considered as spread.
Statistical Analysis-Results are given as mean Ϯ S.E. Comparisons between two groups were done using paired t tests. Comparisons between multiple groups were done using oneway analysis of variance. Specific differences between multiple groups were then determined by use of a Bonferroni post hoc test. Differences were considered significant if p Ͻ 0.05 and are indicated by asterisks in the figure legends.

Neutrophils Acquire GPIIb/IIIa Receptors via Transfer of
PMPs-We first tested whether or not GPIIb/IIIa receptors can be acquired by the neutrophil via PMPs. We stimulated whole blood with ADP to activate platelets and to generate PMPs. By flow cytometry, we observed that whole-blood neutrophils displayed very low amounts of GPIIb/IIIa receptors on their membranes and that this amount was significantly up-regulated after ADP whole blood treatment (Fig. 1A). Using fluorescence microscopy in DAPI-treated cells, we could detect a high amount of neutrophils with very small CD41/61-positive particles on the cell surface in the ADP-treated sample (Fig. 1B). In addition, we observed whole platelets (see inserted box) and platelet-aggregates (not shown) that were larger and showed much higher fluorescence intensity when compared with PMPs. Epinephrine treatment gave similar results to those observed with ADP (data not shown).
We next confirmed the whole blood data under defined in vitro conditions. We particularly wanted to exclude whole platelets and platelet aggregates, ensuring that platelet-derived microparticles caused the GPIIb/IIIa transfer to neutrophils. This setting would also allow us to study functional consequences of such a GPIIb/IIIa receptor transfer. Thus, we generated PMPs from PRP in vitro and transferred these particles to isolated neutrophils. Filtration through 0.8-m filters assured complete removal of whole platelets. We observed that incubation of neutrophils with PMPs resulted in acquisition of GPIIb/IIIa receptor staining (Fig. 2). By flow cytometry, both the percentage of GPIIb/IIIa ␤3-integrin receptor-positive neutrophils, as well as the mean fluorescence intensity for GPIIb/ IIIa staining, increased significantly (p Ͻ 0.05, Fig. 2A). We also observed significant transfer of GPIb (from 14.2 Ϯ 4.1 to 39.1 Ϯ 8.1% of the neutrophils; n ϭ 5, p Ͻ 0.05 and from 14.3 Ϯ 1.1 to 16.2 Ϯ 2.8 for mean fluorescence intensity, respectively, not significant). However, the amount of receptor acquisition was lower when compared with GPIIb/IIIa receptors. In contrast, the other known ␤3-integrin receptor, vitronectin (␣v␤3), was not detected on neutrophils, neither without nor with PMP loading (data not shown). Fluorescence microscopy clearly visualizes acquisition of GPIIb/IIIa receptors in isolated neutrophils after PMP incubation (Fig. 2B).
Acquired GPIIb/IIIa Receptors Are Functional and Cooperate with ␤2-Integrins to Activate NF-B after GM-CSF Stimulation-We next studied whether or not GPIIb/IIIa receptors acquired by the neutrophil were functional. We reported previously that ␤2-integrins provide co-stimulation, allowing cytokines such as GM-CSF and interleukin-8 to activate NF-B signaling (17). However, in our earlier study, we had FIGURE 1. Flow cytometry and fluorescence microscopy for assessment of GPIIb/IIIa receptor expression on neutrophils in whole blood. ADP-stimulated whole blood was incubated with an antibody to the GPIIb/IIIa receptor (CD41/61) followed by red blood cell lysis and incubation with a secondary R-phycoerythrin-labeled antibody. Based on light scatter characteristics, neutrophils were gated. The data show a low percentage of neutrophils expressing GPIIb/IIIa on the surface and show that this percentage increases after stimulation with 10 M ADP (panel A; n ϭ 6; *, p Ͻ 0.05). Using fluorescence microscopy in DAPI-treated cells, we could detect neutrophils with very small CD41/61-positive particles on the cell surface in the ADP-treated sample (PMP; low fluorescence intensity; panel B). When compared with these microparticles, whole platelets (see inserted box) and platelet-aggregates (not shown) were larger and showed much higher fluorescence intensity. not investigated the contribution of PMPs to this process. Importantly, the isolation method used in these earlier experiments yielded neutrophils contaminated with variable amounts of PMPs (data not shown). We used specific antibodies to acti-vate either the ␤2-integrin CD18 or the ␤3-integrin GPIIb/IIIa. We first studied NF-B activation by Western blot experiments assaying the inhibitory IB␣-protein. Fig. 3A indicates that, in GM-CSF-treated suspension neutrophils that had not acquired GPIIb/IIIa receptors, neither an activating CD18 (KIM127) nor an activating GPIIb/IIIa antibody (LIBS-6) induced IB␣ degradation. In contrast, GM-CSFtreated neutrophils that were loaded with PMPs showed NF-B activation in response to an activating CD18 (KIM127) or an activating GPIIb/IIIa antibody (LIBS-6), respectively (Fig. 3B). NF-B activation in PMP-loaded neutrophils was demonstrated by additional independent methods, namely a DNA binding assay (EMSA, Fig. 3C) and quantitative RT-PCR for NF-B-dependent IB␣ mRNA up-regulation (Fig. 3D). A blocking antibody to CD18 (7E4) prevented NF-B activation by the activating GPIIb/IIIa and CD18 antibodies. Similarly, the blocking antibody to GPIIb/IIIa (P2) prevented NF-B activation in response to the activating CD18 and GPIIb/IIIa antibodies (Fig. 3E). We observed similar findings when the assay was carried out on polyhemacrylate at a low cell density where we ensured by light microscopy that neutrophils had no contact to each other (data not shown). This culture condition was shown to securely exclude cell-plastic contact that may have provided a ligand for integrin receptors in the tube experiments and homotypic cellcell contact (18). As an additional control experiment, we preincubated PMP-loaded neutrophils with a blocking antibody to another platelet receptor CD42b (glycoprotein Ib␣). This blocking antibody did not prevent NF-B activation in GM-CSF-treated cells in response to the activating CD18 antibody (Fig. 3F).
Thus far, our data indicate that acquisition of GPIIb/IIIa receptors via PMPs was a prerequisite for NF-B activation in GM-CSF-treated neutrophils. We next tested whether or not PMP loading affected classical ␤2-integrin-mediated functions such as adherence to FIGURE 3. NF-B activation was studied using suspension neutrophils. In GM-CSF-treated suspension neutrophils that had not acquired GPIIb/IIIa receptors, neither an activating ␤2-integrin antibody (aCD18, KIM127, 10 g/ml) nor an activating monoclonal antibody to GPIIb/IIIa (aCD61; LIBS-6, 5 g/ml) induced IB␣ degradation (panel A). ϪPMPs indicates incubation with ADP-treated platelet-poor plasma that was processed exactly as the PRP preparation (ϩPMPs). Co-stimulation for GM-CSF-induced NF-B activation was only seen in neutrophils that were loaded with platelet-derived microparticles and subsequently treated with an activating antibody to CD18 or GPIIb/IIIa, respectively (panel B) (Buf, buffer). NF-B activation in PMP-loaded neutrophils was also shown using EMSA (panel C; n ϭ 3) and quantitative RT-PCR for IB␣ (panel D; n ϭ 3). Blocking antibodies to CD18 (bCD18, 20 g/ml) and to GPIIb/IIIa (bCD41/61, 20 g/ml), but not an isotype control (Iso, 20 g/ml), prevented NF-B activation in response to the activating CD18 antibody and to the activating GPIIb/IIIa antibody in PMP-loaded neutrophils (panel E). A blocking antibody to the platelet receptor CD42b (glycoprotein Ib␣) did not prevent NF-B activation in GM-CSF-treated cells in response to the activating CD18 antibody (panel F). The data presented here are representative Western blots from four independent experiments. fibronectin or spreading or other signaling pathways. Our data indicate that GM-CSF induced neutrophil adhesion to fibronectin and that this response was similar in neutrophils without and with PMP loading (Fig. 4, A and B). The results also document that the blocking CD18 antibody (7E4) completely abrogated adhesion and spreading, whereas the blocking GPIIb/IIIa antibody (P2) showed no effect. Furthermore, neutrophil loading with PMPs or preincubation with blocking antibodies against ␤2-integrins and GPIIb/IIIa receptors, respectively, did not affect PI3K/Akt activation, another central signaling pathway for several inflammatory neutrophil functions (Fig. 4C). These data underscore the specific significance of the GPIIb/IIIa receptor acquisition for NF-B activation in GM-CSF-treated neutrophils.

␤2-Integrins and GPIIb/IIIa Receptors Cooperate in NF-B Activation of GM-CSF-stimulated Neutrophils That
Interact with Fibronectin-During migration, neutrophils interact with extracellular matrix proteins such as fibronectin (19). We next studied whether or not signals transduced by acquired GPIIb/ IIIa receptors were involved in NF-B activation on fibronectin. We incubated PMP-loaded or not PMP-loaded control neutrophils on fibronectin and assessed NF-B activation for up to 3 h. In GM-CSF-treated neutrophils that were loaded with PMPs, NF-B activation was observed at 30 and 60 min. Expectantly, at later time points, IB␣ was already resynthesized. In contrast, in GM-CSF-treated neutrophils that had not acquired GPIIb/IIIa receptors, no NF-B activation was detected at any time point (Fig. 5A). We then selected the 30-min time point and pretreated GPIIb/IIIa receptor-loaded neutrophils with blocking antibodies to CD18 or to GPIIb/IIIa prior to GM-CSF stimulation and incubation on fibronectin. Our data indicate that blockade of either CD18 or GPIIb/IIIa completely prevented NF-B activation in these GM-CSF-stimulated neutrophils on fibronectin (Fig. 5B).  fibronectin (panel A), spreading (panel B), or PI3K/Akt activation (panel C). PMP-loaded neutrophils and not PMP-loaded control neutrophils were incubated with blocking antibodies against CD18 (bCD18), GPIIb/IIIa (bCD41/61), and an isotype control (Iso), respectively. After 30 min, cells were stimulated with GM-CSF (20 ng/ml) on fibronectin-coated (10 g/cm 2 ) wells. Based on time course studies, samples were incubated for 60 min (adherence and spreading) or 15 min (PI3K/Akt activation). A blocking CD18 antibody completely abrogated adhesion and spreading, whereas the blocking GPIIb/IIIa antibody showed no effect (n ϭ 4; *, p Ͻ 0.05). PI3K/Akt activation was done by Western blot experiments using a phospho-specific Akt antibody. The PI3K inhibitor LY294002 (10 M) was used in parallel as a positive control. Equal loading was confirmed using an antibody to total Akt (n ϭ 3).

␤2-Integrins and GPIIb/IIIa Receptors Co-localize in PMP-
loaded Neutrophils-We then studied whether or not ␤2-integrins and GPIIb/IIIa receptors co-localized in PMP-loaded neutrophils. For these experiments, we used double staining with CD41/61 and CD18 specific antibodies and confocal microscopy. The results of our experiments clearly showed that CD41/61 molecules were detected on neutrophils and co-localized with ␤2-integrins (Fig. 6). Moreover, when we quantitatively assayed ␤2-integrin fluorescence intensity, we observed an increase in ␤2-integrin staining in the presence of CD41/61 staining, suggesting that a clustering of the neutrophilic ␤2-integrins was induced after PMP loading (Fig. 6, arrow).
Specific non-receptor tyrosine kinases, such as Src and Syk, are implicated in platelet GPIIb/IIIa outside-in signaling (20 -22). We investigated whether or not specific inhibitors of these kinases as well as cytochalasin B, an inhibitor of actin polymerization, prevents NF-B activation. We observed that preincubation of PMP-loaded neutrophils with the blockers prevented degradation of the inhibitory IB␣-protein (Fig. 7A). In contrast, the PI3K inhibitor LY294002 (Fig. 7A) and the p38 MAPK SB202190 (data not shown) had no effect. However, LY294002 prevented Akt phosphorylation (data not shown). We assured in parallel experiments that these inhibitors had no effect on the percentage of neutrophils that were loaded with GPIIb/IIIa receptors (data not shown).
The Therapeutic GPIIb/IIIa Receptor Blockers Abciximab, Eptifibatide, and Tirofiban Abrogate NF-B Activation of GM-CSF-stimulated Neutrophils on Fibronectin-The specific GPIIb/IIIa receptor blockers abciximab, eptifibatide, and tirofiban are widely used in patients with acute coronary syndromes to prevent platelet aggregation. We preincubated PMPloaded neutrophils with the specific GPIIb/IIIa blockers abciximab, eptifibatide, and tirofiban, respectively, before plating cells on fibronectin. Interestingly, we found that rather low therapeutic concentrations were sufficient for NF-B blockade (100 ng/ml abciximab, 1500 ng/ml eptifibatide, and 40 ng/ml tirofiban), whereas even maximal therapeutic concentration (1000 ng/ml abciximab, 2500 ng/ml eptifibatide, and 60 ng/ml tirofiban, respectively) did not prevent activation of other pathways, as demonstrated for PI3K/Akt (Fig. 7B), p38, and ERK (data not shown). These data underscore the specificity of our findings for NF-B activation. Blockade of NF-B activation by abciximab, eptifibatide, and tirofiban was shown with additional assays. The data indicate that the drugs also prevented DNA binding (Fig. 7C) and transcription of NF-B-dependent genes as shown for TNF-␣ (Fig. 7D) and IB␣ (data not shown) in GM-CSF-treated neutrophils on fibronectin that had acquired GPIIb/IIIa receptors. Thus, three independent methods gave the same results.
We assured in parallel experiments that neither the inhibitory antibodies nor the therapeutic GPIIb/IIIa inhibitors had a significant effect on the percentage of neutrophils that were loaded with GPIIb/IIIa receptors (Fig. 7E). Furthermore, we assessed the influence of the therapeutic GPIIb/IIIa inhibitors on adhesion and on the expression of CD18 in unstimulated and TNF-␣-stimulated cells (data not shown). Neither the basal expression nor the up-regulation of CD18 expression after TNF-␣ stimulation were significantly affected by the compounds. We also found no effect on basal or GM-CSF-stimulated adhesion by the drugs (data not shown).

DISCUSSION
We provide novel data indicating that human neutrophils may acquire GPIIb/IIIa receptors via PMPs. These acquired receptors are functional, allowing for NF-B activation in GM-CSF-stimulated neutrophils that interact with fibronectin. Our experiments suggest that the underlying mechanisms include GPIIb/IIIa-␤2 integrin receptor interactions. The therapeutic GPIIb/IIIa inhibitory compounds abciximab, eptifibatide, and tirofiban all abrogate this NF-B activation, implicating that these anti-thrombotic drugs also exercise anti-inflammatory effects.
Neutrophil functions are not only important for host defense responses but also mediate undesired collateral injury. Several neutrophil functions are controlled by blood-circulating mediators, temperature, pH, oxygen tension, as well as by cell-cell and cell-matrix interactions. Recently, our understanding of circulating mediators was extended by the observation that such mediators may also include cell-derived microparticles that may contain cargo from several host cell types (1). Microparticles, also termed "cell dust," are generated from a variety of cell types, including platelets (1). In fact, in addition to the large body of evidence already documenting the importance of neutrophil interactions with the whole platelet, the functional significance of PMPs for neutrophils is increasingly recognized. Neutrophil-platelet interactions contribute to the pathophysiology of unstable angina, myocardial infarction, cardiopulmonary bypass sequellae, thrombosis, and sepsis (23). Leukocyteplatelet adhesion increases in parallel with the extent of platelet activation (24), as demonstrated in in vitro studies and in patients with coronary artery disease (25,26). However, platelets that are activated by various agonists, such as epinephrine and ADP (25,27), also generate PMPs that are released into the circulation. PMPs are increased in the blood of patients with stroke (28), acute coronary syndromes (29), peripheral vascular disease (30), acute vasculitis, and hemodialysis (4). PMPs contain several proteins and lipids similar to those found on membranes of the platelet from whence they originated. Conceivably, PMPs transfer functional receptors to other cells, including neutrophils. These platelet-derived receptors, including the most abundant GPIIb/IIIa receptor (31), may then participate in neutrophil-induced inflammation. Characterization of such events may identify novel anti-inflammatory targets.
Our study is the first to demonstrate that PMPs transfer GPIIb/IIIa receptors to neutrophils. Because we wanted to study PMPs and not platelet-neutrophil interactions, we very carefully assured ourselves that whole platelets were removed from the preparation by centrifugation and filtration. We also felt that it was important to document GPIIb/ IIIa receptor acquisition in whole blood experiments to exclude the possibility that our results merely reflect an in vitro phenomenon. Flow cytometry and confocal microscopy clearly showed that PMPs, as opposed to whole platelets, transferred GPIIb/IIIa receptors to the neutrophil. However, an important subsequent question was to show whether or not acquired GPIIb/IIIa receptors had functional consequences for the neutrophil. Recently, Lo et al. (32) described findings indicating that PMPs mediate neutrophil activation via platelet glycoprotein Ib. Blocking antibodies against GPIb␣ (CD42b) from the PMPs and against CD18 from the neutrophils abrogated respiratory burst activity and adherence, suggesting that the GPIb-␤2-integrininteraction was functionally important. However, GPIb␣ is not the PMP-derived receptor mediating NF-B activation in our experiments. Blocking antibodies did not abrogate NF-B activation. When we explored signaling, adhesion, and spreading as initial neutrophil responses when these cells are exposed to cytokines such as GM-CSF, we observed that PMP loading enabled the neutrophil to activate NF-B. We knew from previous experiments that GM-CSF does not activate this pathway in suspended neutrophils but does so when the cells interact with fibronectin (17). However, these previous studies were not designed to explore the role of PMPs in this process. In the present study, we used specific activating and blocking antibodies to show that acquisition of the GPIIb/IIIa receptor via PMPs is an underlying mechanism for this NF-B-activating effect. Interestingly, the GPIIb/IIIa receptor-mediated effect was quite specific for NF-B since ERK and PI3K/Akt did not require such a helper signal. Moreover, classical ␤2-integrin-mediated functions such as cell adhesion and spreading on fibronectin were also not affected by GPIIb/IIIa receptor acquisition. Thus, FIGURE 7. NF-B activation depends on non-receptor tyrosine kinases, an intact actin cytoskeleton, and an active GPIIb/IIa receptor. Src and Syk non-receptor tyrosine kinases as well as the actin cytoskeleton control NF-B activation (panel A). Preincubation of PMP-loaded neutrophils with the Src inhibitor PP2 (10 g/ml), the Syk inhibitor piceatannol (Pice, 10 g/ml), and the actin polymerization inhibitor cytochalasin B (Cyto B, 5 g/ml) prevented degradation of the inhibitory IB␣-protein, whereas the inhibitor of PI3K (LY294002 (LY)10g/ml) had no effect (n ϭ 3). The therapeutic GPIIb/IIIa inhibitors abciximab (Ax, 100 ng/ml), eptifibatide (Ep, 1500 ng/ml), and tirofiban (Ti, 40 ng/ml) abrogated NF-B activation by GM-CSF in neutrophils on fibronectin (panels B-E). Western blot experiments using an antibody to IB␣ (n ϭ 5) showed that all three drugs block IB␣ degradation but not PI3K/Akt activation in cells subjected to GM-CSF (20 ng/ml) (n ϭ 5) (panel B). Blockade of NF-B activation was also shown with EMSA (n ϭ 3) (panel C) and quantitative RT-PCR for TNF-␣ (n ϭ 3) (panel D). Additionally, we proved that the inhibitory substances did not influence the percentage of neutrophils with acquired GPIIb/IIIa receptors (n ϭ 4) (panel E).
the data imply that not all neutrophil responses are modified by GPIIb/IIIa receptor acquisition. However, in a GM-CSF-rich milieu, the resulting NF-B activation and subsequent generation of NF-B-mediated pro-inflammatory molecules would have significant consequences for the inflammatory process. For example, NF-B-dependent TNF-␣ generation would affect activation and survival of both neutrophils and neighboring cells.
When we explored mechanisms of the GPIIb/IIIa receptormediated NF-B activation, we found that acquired GPIIb/IIIa receptors and CD18 integrins cluster in close proximity on the neutrophil surface. On the functional level, we found that both receptors cooperate in NF-B activation. We observed that activating either GPIIb/IIIa or CD18 was sufficient to induce NF-B. No such effect was seen without prior GPIIb/IIIa receptor acquisition. We excluded cross-reactivity since the blocking antibody against CD18 abrogated adhesion and spreading, whereas the blocking GPIIb/IIIa antibody showed no effect. Interestingly, blocking platelet-derived GPIIb/IIIa abrogated CD18-induced NF-B activation and vice versa. With GPIb-CD11b/18 and the intercellular adhesion molecule ICAM-2 to leukocyte ␤2-integrin, additional examples for integrin-integrin cross-talk were described (26,(33)(34)(35). We do not yet know exactly how acquired GPIIb/IIIa receptors link to the neutrophil signaling machinery. However, similar to GPIIb/IIIa receptor signaling in platelets, Src and Syk kinases, as well as the actin cytoskeleton, seem to be involved. The exact mechanism as to how the GPIIb/IIIa receptor and the CD18 integrins cooperate to activate NF-B needs clarification in further studies.
Regardless of the exact nature of the receptor interaction, we obtained novel data using the therapeutic GPIIb/IIIa inhibitors abciximab, eptifibatide, and tirofiban. All three compounds completely prevented NF-B activation in GM-CSF-treated neutrophils on fibronectin that had acquired GPIIb/IIIa receptors. We believe that the NF-B inhibition observed in our study was due to GPIIb/IIIa inhibition rather than cross-reactivity to the CD18 ␤2-integrin. Cross-reaction with the activated form of CD11b/CD18 (␣M␤2) (36) and the ␤3 integrin CD51/61 (vitronectin receptor) (37) was reported for abciximab (38) but not shown for eptifibatide and tirofiban (39). Our assumption is further supported by the fact that the therapeutic compounds, similar to the specific GPIIb/IIIa blocking antibody P2, but unlike the blocking CD18 antibodies, showed no inhibitory effect on pure ␤2-dependent neutrophil adhesion and spreading. Thus, our data support the contention that abciximab, eptifibatide, and tirofiban provide anti-inflammatory effects that involve the neutrophil. Acute coronary syndromes are increasingly appreciated as inflammatory diseases. It may well be that GPIIb/IIIa receptor blocker-mediated antiinflammatory effects participate in the improved outcome of this disease.
In summary, we believe we have shown that GPIIb/IIIa receptors are acquired by neutrophils via PMPs. These newly acquired receptors cooperate with ␤2-integrins to activate NF-B signaling. This effect is at work when GM-CSF-treated neutrophils interact with fibronectin. Therapeutic GPIIb/IIIa inhibitory compounds, such as abciximab, eptifibatide, and tirofiban, prevent NF-B activation under these conditions and may have novel implications in anti-inflammatory treatment protocols.