Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Neutrophils and platelets are the predominant cell types found in the microcirculation of the lung after an acute inflammatory injury. In animal models of acute lung injury (1), or patients with the adult respiratory distress syndrome (ARDS) (2), electron microscopy of the lung reveals discrete sites of membrane-membrane contact between neutrophils and platelets, suggesting an active adhesive process between these two cell types. The significance of such an interaction in the pathogenesis of lung disease has been speculated upon (3), but with incomplete understanding of the mechanisms involved, though this interaction is known to generate the pathophysiologically important transcellular metabolic products, such as leukotriene C4 (4).

In vitro evidence points to direct receptor-ligand interactions as being responsible for adhesive interactions between neutrophils and platelets (7). Such heterotypic cell-cell interaction likely alters both neutrophil and platelet functional responses, since in vitro evidence suggests that the neutrophil can affect platelet function (8, 9) and conversely that the platelet can modify neutrophil behavior (10).

The adhesion of fixed, activated platelets to neutrophils has been shown to be dependent, in part, upon platelet surface expression of P-selectin (7). P-selectin (CD62P) is a 140 kD platelet alpha granule membrane protein expressed on the surface of activated platelets (13) and endothelial cells. This cell adhesion molecule is able to interact with its ligand P-selectin glycoprotein ligand-1 (PSGL-1) on the surface of neutrophils, and likely on other leukocytes (14), and cell lines (15). However, in vitro adhesion of activated platelets to neutrophils is often incompletely inhibited by antibodies to P-selectin or a previously defined P-selectin ligand, sLex, alone (16).

While the role of the activated platelet in this interaction has been extensively evaluated, the role of the stimulated neutrophil has received much less attention, despite the active participation of neutrophils in a variety of other cell-cell interactions. In both its homotypic aggregation response and its heterotypic adhesive interaction with the endothelial cell, the neutrophil utilizes cell adhesion molecules from both the selectin family and the ₂ integrin family (17).

We therefore investigated the potential involvement, and relative contribution to the adhesive interaction, of the neutrophil ₂ integrins when the neutrophil is primarily stimulated with a chemoattractant and/or the platelet is primarily activated with thrombin. Using an in vitro system that allows for the analysis, by flow cytometry, of neutrophil-platelet adhesion in the presence of active mixing, we show that the adhesive interaction between live activated neutrophils and platelets is dependent upon both the expression of P-selectin (CD62P) on the surface of the activated platelet and the CD11b/CD18 (a_M2, Mac-1) ₂ integrin complex on the activated neutrophil.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Reagents

All chemicals were from Sigma Chemical Co. (St. Louis, MO), unless otherwise indicated. Salts for KRPD (Kreb's Ringer phosphate-buffer supplemented with 0.2% dextrose, obtained from Abbott Laboratories) were purchased from Mallinckrodt (Paris, KY). Percoll (colloidal silica coated with polyvinylpyrrolidone) and 6% Dextran T500 were obtained from Pharmacia Fine Chemicals (Piscataway, NJ). All solutions were freshly made on each experimental day and pH adjusted to 7.40 ± 0.05. Human thrombin and the chemotactic peptide N-Formyl-Met-Leu-Phe (fMLP) were obtained from Sigma. The vital nucleic acid stain LDS-751 was obtained from Exciton Chemical Co. Inc. (Dayton, OH). Neutrophil inhibitory factor (NIF) derived from the hookworm Ancylostoma caninum was expressed and purified as previously described (20). Denatured NIF was generated by boiling for 4 min in 10 mM DTT. Peptides RGDS, RGES, and GPRP were synthesized by Molecular Resource Center at National Jewish Center for Immunology and Respiratory Medicine.

Antibodies

FITC-labeled monoclonal antibody to human CD45 (anti- HLe-1, clone 2D1) was obtained from Becton-Dickinson. Biotinylated monoclonal antibody to human CD41a (GP IIb/IIIa, clone P2) and anti-CDw49b (VLA2, GPIa, clone Gi 9), anti-CD18 (clone 7E4), anti-CD11c (clone BU15), anti-CD11a (clone 25.3.1) were obtained from Immunotech, Inc. Anti-CD18 (clone MHM23) and anti-CD11b (clone 2LPM19c) were obtained from Dako. Phycoerythrin-avidin was obtained from Molecular Probes (Eugene, OR). Anti-CD11b (clone LM2) was from the ATCC. Monoclonal antibody G1 and non-blocking control antibody S12 directed against human P-selectin were kind gifts of Dr. Rodger McEver (University of Oklahoma, Oklahoma City, OK). All blocking antibodies were used at saturating concentrations (10 µg/ml) as determined by preliminary experiments.

Neutrophil Isolation

Human blood neutrophils were prepared by a plasma- Percoll gradient method as described previously (21) to minimize exposure to endotoxin, with slight modifications. Briefly, 40 ml of whole blood was centrifuged at 175 × g for 15 min in a 50-ml sterile, polypropylene centrifuge tube containing 4.4 ml of 3.8% citrate (in H₂O) as anticoagulant. The upper platelet-rich plasma layer was removed, part of which was underlain with 90% Percoll and centrifuged at 1,000 × g for 15 min to produce platelet-poor plasma. Following removal of the upper layer, 5 ml of 6% dextran T500 and 0.9% sterile saline were added to the sedimented cells to bring the volume to 50 ml. The solution was mixed thoroughly, and allowed to stand at room temperature for 30 min. The leukocyte-rich supernatant was then collected and centrifuged at 112 × g for 8 min at room temperature. Pelleted cells were resuspended in platelet-poor plasma and then separated through 42% and 51% plasma-Percoll gradients at 180 × g for 10 min. The lower neutrophil band was collected and washed in platelet-poor plasma by centrifuging 112 × g for 7 min. As much platelet-poor plasma was removed as possible, and neutrophils were washed twice in KRPD by centrifuging at 112 × g for 7 min. The neutrophils isolated in this manner were > 98% pure and > 99% viable as determined by trypan blue exclusion.

Platelet Isolation

Venous blood was obtained from adult human volunteers free of medication for at least one week prior to the blood draw. Forty milliliters of whole blood was collected in a 50-ml sterile polypropylene centrifuge tube containing 4.4 ml of 3.8% sodium citrate and centrifuged for 15 min at 175 × g at room temperature to obtain platelet rich plasma (PRP). The PRP was transferred to a polypropylene tube with a plastic pipette and further acidified to pH 6.5 by addition of 1/10 vol/ vol acid-citrate-dextrose (ACD). The cells were then centrifuged for 15 min at 1,000 × g and the platelets washed after the method of Patscheke (22) in 36 mmol/L citric acid, 5 mmol/L glucose, 5 mmol/L KCl, 2 mmol/L CaCl₂, 1 mmol/ L MgCl₂, 103 mmol/L NaCl, 100 nmol/L PGE₁ (pH 6.5). The platelet pellet was resuspended in Krebs' buffer (pH 7.4).

Cell Incubation

All incubations were performed at 37°C. Just prior to incubation, the individual cell populations were resuspended in Krebs' buffer. The isolated neutrophils were placed in a polypropylene tube containing an 8-mm magnetic stir bar and incubated with or without blocking antibodies (saturating concentrations 10 µg/ml) or for 20 min at 37°C. Isolated platelets were added (final cell concentrations 1.25 × 10⁶/ml neutrophils and 3.125 × 10⁷/ml platelets), immediately followed by Ca²⁺ and Mg²⁺ (1.27 and 0.8 mM final concentration, respectively), thrombin (1 U/ml) and fMLP (10⁷ M final concentration). The test tube was placed in a 37°C water bath on top of a magnetic stir plate set at 1,000 revolutions per minute (RPM) to maintain thorough mixing throughout the experiment. In time course experiments over 30 min, each tube was repeatedly sampled. All experiments were repeated at least three times.

Assessment of Adhesion

Evaluation of platelet/leukocyte adhesion was performed using a modification of a flow cytometric method recently described (23). While the cell mixture was actively stirring, 400 µl containing 5 × 10⁵ leukocytes and 12.5 × 10⁶ platelets was removed and added to an equal volume of 2% paraformaldehyde for 60 min, followed by a wash in 0.1 N glycine 8:1 vol:vol. The cells were then washed with Krebs' buffer, and incubated with saturating concentrations of biotinylated anti-CD41a (anti-IIbIIIa) to stain the platelets and FITC-anti-CD45 to stain the neutrophils, for 20 min at 23°C. In some experiments, to confirm platelet fluorescence, an anti-CDw49b (GPIa) was used to stain platelets. The results with anti-CDw49b were always similar to those obtained with the anti-CD41a antibody. After washing, the cells were incubated with phycoerythrin (PE)-avidin for 20 min, washed, and then resuspended in 400 µl of Krebs buffer.

Samples were analyzed on a FACScan flow cytometer (Becton Dickinson) with data stored in list mode files. The percentage of leukocytes binding platelets was evaluated by performing live gating on leukocyte-sized events using neutrophil (CD45-FITC) fluorescence and forward or side scatter, and subsequently the combination of CD45-FITC fluorescence and platelet (PE) fluorescence. The threshold for identification of neutrophil fluorescence was defined by unstimulated neutrophils stained in an identical fashion to the heterotypic cell mixtures. No differences in the identification of neutrophil-platelet aggregates were found when acquisition was performed at two different speeds suggesting an insignificant effect of coincident isolated neutrophil and platelet events. Ten thousand events were measured in the leukocyte gate. The data was analyzed using LYSYS software (Becton Dickinson). An isotype-matched control antibody with PE-avidin, and PE-avidin alone without a primary antibody were used to set the threshold for positive platelet fluorescence on the neutrophils. Both the percentage of neutrophils with platelet fluorescence, and the median platelet fluorescence of those platelets bound to neutrophils under each condition were generated.

Evaluation of neutrophil aggregation was performed both with and without the addition of platelets using the methods of Simon and associates (24). LDS-751, a vital nucleic acid stain, was mixed with isolated neutrophils (20 × 10⁶/ml) at a concentration of 0.2 µg/ml for 5 min at 37°C prior to stimulation. At appropriate time points after stimulation, 400 µl of sample was removed and fixed in an equal volume of 2% paraformaldehyde. After washing in Krebs' buffer, the cells were evaluated on the FACScan and analysis of aggregation was quantified using forward scatter, side scatter, and LDS-751 fluorescence (FL-3). Ten thousand events were counted, stored in list mode files, and evaluated in linear mode.

Electron Micrography

Ice-cold fixative containing 1.5% glutaraldehyde, 1% acrolein, 1% osmium tetroxide in 0.1 M sodium cacodylate, pH 7.3, was added 1:1 (v/v) to the cell suspension. The cells were fixed for 10 min, washed with 0.1 M sodium cacodylate buffer, and stained with 3% aqueous uranyl acetate. Cells were dehydrated through a graded acetone series and infiltrated with Luft's 3:7 embedding resin. Thin sections were stained with uranyl acetate and Reynold's lead, evaluated and photographed using a Philips 400T electron microscope at an accelerating voltage of 60 kV.

Statistics

Comparisons were performed using Dunnett's method. Significance was measured at the P = 0.05 level. The data are presented as the standard error of the mean (SEM).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

Analysis of Neutrophil-Platelet Adhesion

Mixtures of neutrophils and platelets, allowed to interact, can result in neutrophil-platelet, neutrophil-neutrophil, and platelet-platelet adhesive interactions. By labeling the cellular aggregates after fixation (neutrophils with FITC-labeled anti-CD45 and platelets with PE-labeled anti-IIb/ III or anti-Ia) the heterotypic interactions between neutrophils and platelets were distinguished. A small (< 5%) but consistent percentage of neutrophils isolated from normal donors had detectable platelets bound to them. Incubation of neutrophils and platelets at a ratio of 1:25 and stimulated with fMLP and thrombin, led to large numbers of platelets associating with cells bearing the neutrophil marker. Figure 1A shows the dot plot of an experiment in which neutrophils and platelets were mixed and stimulated with fMLP and thrombin. R1 defines a region of interest that identifies neutrophils based on their fluorescent label. The panel B shows the histogram of region R1, identifying the neutrophil-associated platelet fluorescence. "Threshold" refers to the upper limit (99%) of nonspecific (control) platelet fluorescence. The percentage of neutrophils associated with platelet fluorescence above this level was identified as the percentage of neutrophils binding platelets. As a contrast, panel C shows the neutrophil- associated platelet fluorescence when neutrophil-platelet cell mixtures were mixed without added stimuli.

View larger version (18K): [in this window] [in a new window]   Figure 1.   Analysis of neutrophil-platelet adhesion. (A) shows a dot plot of a labeled neutrophil platelet mixture that utilizes neutrophil fluorescence (FL1, FITC-labeled anti-CD45) and platelet fluorescence (FL2, phycoerythrin-labeled anti-GPIIbIIIa) as one way to separate the cells of interest (only the first 2,000 events are shown for clarity). R1 defines the region of interest, "Neutrophil region," that captures all neutrophils and excludes individual and aggregate platelets not associated with neutrophils. Under each condition, 10,000 events were collected within this region. (B) shows the platelet fluorescence histogram of region R1, identifying the neutrophil-associated platelet fluorescence when the isolated cell mixtures were mixed with a stir bar and stimulated with the neutrophil-specific stimulus fMLP, and the platelet-specific stimulus thrombin. "Threshold" refers to the upper limit (99%) of nonspecific (control) platelet fluorescence. The number of neutrophils with associated platelets was measured from this level. For comparative purposes, (C) shows a histogram of region R1 from a separate experiment identifying the neutrophil-associated platelet fluorescence when the isolated cell mixtures were solely mixed with a stir bar. Note that in this analysis program the fluorescence axes begin at 10°, such that all of those neutrophils with 0-1 relative platelet fluorescence, while still measured, are off the graphical scale to the left. In these two examples the relative number of neutrophils were: for fMLP/Thrombin, free neutrophils = 2,868, neutrophils bound to platelets = 6,680; for mixing alone, free neutrophils = 7,747, neutrophils bound to platelets = 1,827.

A relative platelet fluorescence index was also calculated as follows: A median platelet fluorescence found on those neutrophils binding platelets was measured from the platelet fluorescence threshold. On each experimental day the median fluorescence of single platelets was measured by labeling isolated fixed platelets in an identical fashion to the heterotypic cell mixtures. The median platelet fluorescence found on neutrophils under each condition was then divided by the fluorescence of single platelets to generate the relative fluorescence index. As GPIIb/IIa complex fluorescence can increase slightly after stimulation (70% of the complexes are randomly distributed on the platelet surface while 30% are internal at rest) the relative platelet fluorescence index can only be considered a relative number of platelets per neutrophil, slightly overestimating the number of platelets per neutrophil after stimulation.

Adherence was enhanced by mixing of the cells with a small magnetic stir bar (data not shown). Accordingly, all further experiments were performed with mixing. When the platelet:neutrophil ratio was increased to 50:1 or greater, the resulting cellular aggregates were too large for measurement in the flow cytometer; therefore all further experiments were performed at a platelet:neutrophil ratio of 25:1, a ratio similar to that seen in normal blood.

Ultrastructure

In order to confirm that the association detected by flow cytometric methods reflected interactions between neutrophil and platelet surface membranes, we examined the ultrastructural appearance of the fixed neutrophil platelet mixtures after stimulation. In Figure 2A, a representative neutrophil from a mixture incubated at a neutrophil:platelet ratio of 1:25 is shown. Four platelets are in close approximation to the neutrophil surface. In a higher magnification view of a different neutrophil, the complex area of contact between neutrophil and platelet membranes is evident (Figure 2B).

View larger version (130K): [in this window] [in a new window]   View larger version (168K): [in this window] [in a new window]   Figure 2.   Electron micrograph of activated neutrophil-platelet adhesion. (A) Four platelets adhere to a central neutrophil after active mixing and activation with fMLP (107 M) and thrombin (1 U/ml) (bar = 1.5 µm). (B) Tight interdigitating adherence relationship between activated neutrophils and platelets (bar = 0.5 µm).

Time Course of Heterotypic Adhesion During Cell-Specific Stimulation

Previous studies have shown that stimulation of the platelet leads to their enhanced adherence to neutrophils (7, 23, 25). In order to determine whether stimulation of the neutrophil alone can regulate platelet adherence, and whether activation of both the neutrophil and platelet further increases this adhesive interaction, we compared the mean percentage of neutrophils binding platelets and the relative number of platelets per neutrophil after stimulation with either fMLP or thrombin or both over time. Evidence for the cellular specificity of these agonists comes from the observation that fMLP did not induce Ca²⁺ mobilization in isolated platelets while thrombin did not induce an elevation in cytosolic free calcium in isolated neutrophils (data not shown).

The stimulation of neutrophils and platelets leads to a heterotypic adhesive interaction that is time-dependent. As shown in Figure 3, unstimulated mixing of neutrophils and platelets led to ~ 15% of neutrophils binding platelets by 15 s, and this level of adhesion was unchanged over 30 min. In contrast, adhesion induced by thrombin showed a gradual increase beginning at 15 s, peaking at 2 min and then decreasing at 5 min, when the number of neutrophils binding platelets was similar to that of mixing alone. The neutrophil-specific stimulant fMLP showed a similar time course with peak adhesion at 2 min (being significantly greater than that of thrombin), and with disaggregation evident by 5 min. The combination of thrombin and fMLP induced enhanced neutrophil-platelet adhesion at 15 s; in contrast to either stimulant alone, the combination led to peak adhesion by 60 s, but disaggregation still occurred by 5 min. At every time point beyond 15 s, the adhesion induced by the combination of thrombin and fMLP was significantly greater than that generated by fMLP or thrombin alone.

View larger version (14K): [in this window] [in a new window]   Figure 3.   Time course of neutrophil-platelet adhesion. Neutrophils and platelets were mixed in a ratio of 1:25 and stimulated with either thrombin, fMLP, or both, with adhesion evaluated at selected time points.

The time course of heterotypic adhesion was similar to that of neutrophil homotypic aggregation. Accordingly, to determine what role neutrophil homotypic aggregates might play in the formation of neutrophil-platelet aggregates, we measured the occurrence of neutrophil aggregates both in the presence and absence of platelets when either the neutrophil alone or both the neutrophils and the platelets were stimulated. As seen in Table 1, at 2 min, more than 89% of the neutrophils were still in singlet form when in the presence of platelets and both cells types were stimulated (the conditions operative in our evaluation of the heterotypic adhesion). In contrast, when stimulation occurred in the presence of a > 10-fold higher concentration of neutrophils and platelets the number of singlet neutrophils dropped to ~ 60% at maximal aggregation. Neutrophil aggregation occurs over a similar time course to heterotypic adhesion and is more intense in the presence of activated platelets. As approximately 10% of singlet neutrophils are recruited into aggregates at the cell concentrations and with the stimuli used in our system, this may well increase the calculated percentage of neutrophils binding platelets and the relative number of platelets per neutrophil by approximately 10-15% beyond that which would be seen if all neutrophils remained in singlet form. While this 10-15% increase in heterotypic binding percentage might account for a portion of the difference noted between fMLP-alone- and thrombin-alone-induced adhesion, it is not enough to explain the overall differences identified between the specific stimuli. And, as the number of neutrophil aggregates remains relatively constant over the measured 30 s to 5 min time period it cannot explain the differences noted over time.

Figure 4 reveals the time course of the relative platelet fluorescence index after cell-specific stimulation. As shown, with mixing alone only single platelets were bound and this did not change over time; for fMLP alone, peak fluorescence occurred by 30 s and remained constant until a subsequent fall at 5 min; for thrombin, a progressive increase in the number of platelets per neutrophil over 2 min occurred followed by a significant decrease by 5 min; for the combination, the peak of relative platelet fluorescence occurred much earlier, by 60 s, remained stable through 2 min and declined at 5 min to a level similar to thrombin alone.

View larger version (13K): [in this window] [in a new window]   Figure 4.   Time course of relative platelet fluorescence on neutrophils binding platelets.

Role of P-Selectin

Previous studies have proposed a primary role for P-selectin-mediated adhesion of stimulated platelets to neutrophils (7, 23, 25). In order to evaluate this interaction, we used the antibody G1, a blocking monoclonal antibody directed against P-selectin (CD62P). G1 had no effect on unstimulated adherence (data not shown), but it significantly inhibited thrombin-induced adherence (Figure 5A) by limiting the percentage of neutrophils binding platelets to the level seen in unstimulated cells. Interestingly, G1 had a similar inhibitory effect upon neutrophil-platelet adhesion stimulated by fMLP (Figure 5B). When both cell types were stimulated, the inhibitory potency of G1 was time-dependent, blocking 85% of the agonist-stimulated adhesion at 30 s, and only 32% at 120 s (Figure 6A). In contrast, a nonblocking antibody against P-selectin (S12) had no effect on the interaction (Figure 7).

View larger version (27K): [in this window] [in a new window]   View larger version (31K): [in this window] [in a new window]   Figure 5.   (A) Role of P-selectin and CD11b/18 in adhesion induced by thrombin over time. Neutrophils and platelets were actively mixed and stimulated with the cell-specific stimulus thrombin in the presence of a blocking anti P-selectin antibody (G1), the CD11b blocking protein NIF, or both. All blocking conditions significantly lower than thrombin alone at 30, 60, and 120 s (P < 0.05). (B) Role of P-selectin and CD11b/18 in adhesion induced by fMLP over time. Neutrophils and platelets were actively mixed and stimulated with the cell-specific stimulus fMLP in the presence of a blocking anti P-selectin antibody (G1), the CD11b blocking protein NIF, or both. All blocking conditions significantly lower than thrombin alone at 30, 60, and 120 s (P < 0.05).

View larger version (15K): [in this window] [in a new window]   View larger version (14K): [in this window] [in a new window]   Figure 6.   (A) Role of P-selectin and CD11b/18 in neutrophil-platelet adhesion induced by the combination of fMLP and thrombin over time. Neutrophils and platelets were actively mixed and stimulated with the cell-specific stimuli fMLP and thrombin in the presence of a blocking anti P-selectin antibody (G1), the CD11b blocking protein NIF, or both. Adhesion was evaluated at selected time points. (B) Role of P-selectin and CD11b/18 in the relative platelet fluorescence induced by the combination of fMLP and thrombin over time. Neutrophils and platelets were actively mixed and stimulated with the cell-specific stimuli fMLP and thrombin in the presence of a blocking anti P-selectin antibody (G1), the CD11b blocking protein NIF, or both. The relative platelet fluorescence was evaluated at selected time points.

View larger version (16K): [in this window] [in a new window]   Figure 7.   Effect of selected antibodies or NIF on the percentage of neutrophils binding platelets. Neutrophils and platelets were mixed and stimulated with the cell-specific stimuli fMLP and thrombin in the presence of various blocking antibodies or NIF and adhesion was evaluated at 2 min after stimulation. * Signifi- cantly different from fMLP/thrombin alone (P < 0.05).

We also evaluated the effect of blocking P-selectin on the median platelet fluorescence on those neutrophils binding platelets. As shown in Figure 6B, G1 significantly decreased the relative platelet fluorescence index at 30, 60, and 120 s, the times during which the greatest number of platelets bound to each neutrophil.

Role of Neutrophil ₂ Integrins

To evaluate the role of neutrophil ₂ (CD18) integrins in neutrophil-platelet adhesion, we used blocking antibodies to CD11a, CD11b, CD11c, CD18, and a novel CD11b blocking protein, neutrophil inhibitory factor (NIF). NIF is a 41 kD glycoprotein derived from the canine hookworm Ancylostoma caninum that binds to, and blocks adhesion mediated through CD11b (20).

NIF had no effect on the adhesion generated by mixing alone (data not shown). As CD11b is upregulated and activated on stimulated neutrophils (26), NIF might be expected to inhibit fMLP stimulated platelet-neutrophil adherence, but to have little or no effect on that induced by thrombin. However, Figures 5A and 5B show that incubation of the neutrophils with NIF significantly inhibited both fMLP and thrombin stimulated adhesion at 30, 60, and 120 s. When used in the presence of both stimuli, NIF-induced inhibition of adhesion was time-dependent, similar but more potent that G1, blocking 90% of the agonist-stimulated adhesion at 30 s and 61% at 120 s. In contrast, denatured NIF (see MATERIALS AND METHODS) had no effect (Figure 7).

NIF binds to the I-domain of CD11b and blocks neutrophil function presumably by interfering with interactions between CD11b/CD18 and some of its cognate ligands (27). Antibodies known to block some CD11b functions (LM2, 2LPM19c) were also examined. In contrast to NIF neither of these antibodies had an effect on the adhesive interaction (Figure 7). Neither an anti-CD11c (BU15) nor an anti-CD11a (25.3.1) consistently inhibited neutrophil-platelet interaction in this system (Figure 7). Similarly, while MHM23, an antibody to CD18, the common (₂) subunit to the neutrophil CD11/CD18 integrins had no effect on adhesion, incubation of neutrophils with 7E4, another anti-CD18 antibody, led to a 45% inhibition of the stimulated neutrophil-platelet adhesion (Figure 7). When the relative platelet fluorescence index was examined (Figure 6B), NIF blocked the number of platelets per neutrophil when both stimuli were combined with a time course similar to the anti P-selectin antibody, though with less potency. The effect of the anti-CD18, 7E4, on the median platelet fluorescence mirrored the effect of NIF (data not shown).

Anti-P-Selectin/Anti-₂ Integrin Combinations

We next investigated the effect of combined blockade of P-selectin and CD11b on neutrophil platelet adhesion and the relative number of platelets per neutrophil. The combination of G1 and NIF had no effect on unstimulated adhesion (data not shown). When evaluated after stimulation of the mixtures with single cell stimuli (Figures 5A and 5B), the combination of blockade was not significantly greater than either individual blocking agent alone. In contrast, as seen in Figure 6A, when the neutrophil-platelet mixtures were stimulated with both fMLP and thrombin the combination of G1 and NIF was significantly more potent at inhibiting adhesion than either alone at both the 60 and 120 s time points.

The combination of P-selectin and CD11b blockade also lowered the relative number of platelets per neutrophil to the level seen in unstimulated mixtures (Figure 6B).

Effect of RGDS Peptide

RGDS peptides have previously been reported to inhibit neutrophil-platelet interactions through a glycoprotein IIb/IIIa-mediated mechanism when platelets are fixed prior to interaction (16). In our system, RGDS at 1 mg/ml had no effect on the number of neutrophils binding platelets (data not shown); however, it did inhibit the number of platelets per neutrophil as measured by a 40% reduction in the relative platelet fluorescence index as compared with the control peptide RGES (Figure 8).

View larger version (68K): [in this window] [in a new window]   Figure 8.   Effect of RGDS on the relative number of bound platelets per neutrophil (median platelet fluorescence). Platelets and neutrophils were actively mixed at a ratio of 25:1, and stimulated with fMLP plus thrombin in the presence or absence of RGDS (1 mg/ml) or RGES (1 mg/ml). * Significantly different from fMLP/ thrombin and RGES (P < 0.05).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Clinically-apparent neutrophil-platelet interaction was reported in the early 1960s (28). Subsequent evaluation has revealed such interactions in the adult respiratory distress syndrome (ARDS) (2) and in animal models of acute lung injury (1). In the clinical setting, the mechanisms responsible for the initiation and maintenance of this adhesive interaction are unclear, as the complex in vivo environment precludes a precise delineation of the events.

We, as well as others (7, 14, 23, 25), have developed in vitro systems to assess the responsible mechanisms. These studies and others have identified and expanded upon the important contributory role of platelet P-selectin and its glycoprotein counterreceptors on the neutrophil in these adhesive interactions. In the majority of these models of adhesion, either the neutrophil, the platelet, or both have undergone an initial fixation step prior to mixing, while our model of platelet-neutrophil interaction uses live cells in a dynamic system with fixation only after the interaction is complete. This system has permitted new insights into neutrophil-platelet interactions.

Our data indicates that specific stimulation of either the neutrophil or the platelet leads to the formation of neutrophil-platelet aggregates, and that there are characteristic differences between the resulting adhesive events. Stimulation of the platelet leads to a significant increase in the number of neutrophils binding platelets, and to a marked increase in the number of platelets bound per neutrophil. In contrast, specific neutrophil stimulation increases the number of neutrophils binding platelets beyond that seen with thrombin alone, but with fewer platelets bound per neutrophil. Stimulation of both cell types simultaneously prompts the largest numbers of neutrophils to bind large numbers of platelets (Figure 9).

View larger version (33K): [in this window] [in a new window]   Figure 9.   Proposed scheme for the heterotypic adhesive interaction between neutrophils and platelets after mixing alone, stimulation with thrombin, fMLP, or the combination (see text for further details).

When neutrophils and platelets were mixed and exposed to thrombin, the resulting stimulated adhesion was largely (though not exclusively) dependent upon platelet P-selectin, as judged by its inhibition by the anti-P-selectin antibody G1. These data are thus consistent with considerable data in the literature on the important role of platelet P-selectin in adherence of stimulated platelets. Surprisingly then, the novel anti-CD11b protein NIF (20, 27) also significantly inhibited neutrophil-platelet interaction induced by thrombin. The inhibitory properties of NIF appear to be a function of its ability to bind to the I-domain of the CD11b subunit, and its inhibitory actions have been shown to mimic those of some anti-₂ integrin antibodies (27). In a novel finding, neutrophil-platelet mixtures stimulated solely by fMLP also resulted in significant neutrophil-platelet adhesion, and this fMLP-stimulated adhesion was also blocked not only by blocking P-selectin mediated adhesion, but also CD11b blockade.

When neutrophils and platelets were mixed and exposed to the combination of cell-specific stimuli, the greatest number of neutrophil-platelet adhesive events occurred, large numbers of platelets were adherent to neutrophils, and the dependence on specific adhesion systems changed. Although anti-P-selectin inhibited the stimulated interaction when thrombin or fMLP were used singly, it inhibited only 30% of the adherence at 2 min when both stimuli were used. The inhibition attributable to NIF, or to the blocking anti-CD18 antibody 7E4, was significantly greater at 2 min (60%) than that induced by blocking P-selectin, but again only partial. The combination of anti-P-selectin and the CD11b antagonist or anti-CD18 antibody essentially abrogated the stimulated interaction between neutrophils and platelets supporting separable adhesive mechanisms when maximal stimulation was used.

Time course experiments support the notion that both P-selectin- and ₂-integrin-mediated adhesion occurs and is separable when both cell types are stimulated by direct agonists. While stimulated adhesion appears both P-selectin-and ₂-integrin-dependent early, the P-selectin dependency is lost at a more rapid rate than that due to integrins over time. This is in contrast to sole platelet or neutrophil stimulation when both P-selectin and integrin dependence remain at 2 min, suggesting that while qualitatively similar, primary stimulation of both cells together invokes adhesive mechanisms that may differ quantitatively from those invoked by single cell stimulation. Interestingly, the heterotypic adhesive interaction under all conditions appears to disassemble by 5 min, suggesting a transient interaction in the face of continuous mixing and in contrast to those studies performed with fixed platelets or in whole blood (29, 30).

These data suggest that both ₂ integrin and selectin adhesion molecule families are responsible for mediating the adhesive interaction of live neutrophils and platelets. Selectin and ₂ integrins have been shown to be functionally interrelated in a number of systems, including neutrophil aggregation and neutrophil-endothelial interactions (17, 31). Our data also suggest that the live platelet is able to actively participate in heterotypic adhesion in a way that fixed platelets cannot (29), and that this participation involves the expression of a ligand for neutrophil CD11b/ CD18. We hypothesize that stimulation of the neutrophil and platelet, either primarily with specific stimuli or through transcellular mechanisms, with resultant upregulation of adhesion molecules, allows CD11b/CD18-mediated adhesion to maintain adhesion in the face of transient P-selectin mediated binding. This is supported by a recent study revealing that stimulation of the neutrophil reduces the extent of adhesion mediated through P-selectin through a redistribution of P-selectin glycoprotein ligand-1 (PSGL-1) on the neutrophil and that inhibition of CD18 magnifies this decrease in P-selectin-dependent adhesion (32). Other studies have also noted that P-selectin and neutrophil ₂ integrins appear to be functionally related in a defined (33) or undefined way (34), both in vitro and in vivo (38).

PSGL-1 appears to be the dominant P-selectin counterreceptor found on the neutrophil (39). The counterreceptor(s) for neutrophil CD11b/CD18 on the platelet, however, is unclear. Platelet expression of ICAM-1 has not been described. While ICAM-2, a known counterreceptor for CD11a (LFA-1, _a2), has recently been shown to be constitutively expressed on the surface of the platelet (40) we were unable to inhibit adhesion with a blocking antibody to CD11a. Fibrinogen, a known ligand of CD11b and CD11c (41), is bound to the surface of activated platelets, and is likely available for interaction with surface expressed neutrophil integrins (42). However, the lack of consistent effect of anti-CD11c antibodies and RGDS peptides on heterotypic adhesion makes it unlikely that fibrinogen is a dominant mechanism.

In those conditions that used thrombin, large numbers of platelets bound to each neutrophil. The number of platelets bound per neutrophil, but not the number of neutrophils binding platelets, was decreased by RGDS. This suggests that some platelet-platelet adhesion occurs on the surface of singlet neutrophils. The time course of the relative platelet fluorescence shows that this platelet-platelet interaction occurs over time and that it is delayed slightly compared with the neutrophil platelet curve. This may suggest a slight competitive advantage of neutrophil-platelet interaction over platelet-platelet adhesion. The lack of change in the number of neutrophils binding platelets in the presence of RGDS also suggests that platelet-platelet interaction neither limits nor promotes the number of neutrophil-platelet adhesive interactions.

Interestingly, there appear to be a small but significant number of neutrophils whose adherent platelets are independent of both P-selectin and Cd11b/CD18, as adhesion induced by mixing alone is unaffected by blocking either of these two cell adhesion molecules. This adhesion is characteristically single platelets on a small fraction of neutrophils. This low level of adhesion appears to remain even after the neutrophil-platelet disassembly noted at 5 min after stimulated adhesion. The mechanism of this adhesion and its significance awaits exploration.

In summary, the major findings of this study are: the heterotypic adhesive interaction between live neutrophils and platelets can be initiated by activation of either cell type, this cell-specific activation results in characteristic adhesion patterns, and stimulated adhesion involves both platelet P-selectin and the neutrophil CD18 (₂) integrins. The findings in this study broaden our understanding of the adhesive repertoire used to form neutrophil-platelet interactions, support a close association between the thrombotic and inflammatory systems, and suggests that therapies directed at either P-selectin or CD11b may modify both thrombosis and inflammation.

Since this data was first published in abstract form (43) and submitted for publication, two other papers (44, 45) have been published supporting our observations.