Increased Neutrophil-Platelet Adhesion in Patients With Unstable Angina

Patient Selection

The study included 25 consecutive patients with unstable angina and 25 consecutive patients with stable angina. All patients had angiographically proven coronary artery disease. Patients with stable angina had typical angina of Canadian Cardiovascular Society functional class II. Unstable angina was defined by the presence of at least one of the following features: (1) angina of recent onset, that is, within 6 weeks, brought on by minimal exertion; (2) crescendo angina, that is, more severe, prolonged, or more frequent anginal attacks, superimposed on chronic effort angina; or (3) angina at rest lasting for ≥15 minutes. Exclusion criteria were: (1) interfering noncardiac disease, that is, anemia, infection, malignoma, collagen disease, diabetes mellitus, thyrotoxicosis; (2) cardiac disease other than coronary artery disease, except for minor mitral regurgitation; (3) overt right or left ventricular failure; (4) myocardial infarction, coronary artery bypass surgery, or balloon angioplasty within 3 months preceding the study; or (5) myocardial infarction evolving within 8 hours after withdrawal of blood. Baseline characteristics of the study patients are shown in Table 1⇓.

The study was approved by the institutional ethics committee for human subjects. Informed consent was obtained from all patients.

Immunofluorescence

Neutrophil staining was performed in whole-blood samples by use of dual-color immunofluorescence.^{14 15} Blood samples were obtained before the initiation of in-hospital therapy once the diagnosis of unstable or stable angina was established. Samples were collected in 1:10 (vol/vol) CPDA (Greiner). Staining was performed within 20 minutes of blood sampling with the use of FITC-conjugated mAbs, anti-CD11b (Bear1), anti-GP IIb/IIIa (P2), and anti–L-selectin (DREG56), as well as PE-conjugated anti-CD45 (KC56, Coulter Electronics) and negative isotype control (Immunotech). Twenty-five microliters of blood and an equal volume of Tyrode's buffer (0.1% BSA, 0.1% glucose, 2 mmol/L MgCl₂, 137.5 mmol/L NaCl, 12 mmol/L NaHCO₃, 2.6 mmol/L KCL, pH 7.4) were incubated with saturating concentrations of FITC- and PE-conjugated mAbs for 30 minutes at room temperature (anti-GP IIb/IIIa) or at 4°C (anti-CD11b and anti–L-selectin). Thereafter, erythrocytes were lysed and neutrophils were fixed with commercially available solutions (Immunolyse and Fixative, Coulter Electronics). After being washed three times, the cells were stored in 1% paraformaldehyde at 4°C until flow cytometric analysis was performed within 24 hours.

Platelet immunostaining was performed as previously described.^{16 17 18 19} Briefly, platelet-rich plasma was obtained after centrifugation at 100g for 20 minutes and incubated with saturating concentrations of the following FITC-conjugated mAbs: anti-GP IIb/IIIa, anti–P-selectin (CBL-thrombo/6, Immunotech), and anti-LIBS1 directed against LIBS on GPIIIa (kindly provided by Dr M. Ginsberg, The Scripps Research Institute, La Jolla, Calif) in a total volume of 50 μL Tyrode's buffer. After 30 minutes' incubation at room temperature in the dark, samples were fixed with 0.2% paraformaldehyde. Flow cytometric analysis of the fixed samples was performed within 24 hours after sampling.

For flow cytometry, we used a FACSscan (Becton-Dickinson) equipped with a 488-nm argon laser at 500 mW. The day-to-day reproducibility was assured by a mixture of monosized, fluorescent beads (CaliBRITE, Becton-Dickinson). The histogram generated by the PE fluorescence served to identify leukocytes as positive for anti-CD45, because this panleukocyte antigen is not present on erythroid cells or platelets. Fluorescence intensity of 10 000 leukocytes and 5000 platelets was recorded as a mean channel number over a logarithmic scale. To discriminate neutrophils from monocytes and lymphocytes, a gate was set in the forward angle and the right-angle scatter plot. Platelets were gated according to their specific light-scatter properties. More than 96% of this population was GP IIb/IIIa positive. Data were stored in list mode files and processed on a Hewlett Packard computer equipped with Consort 30 software. The mean channel of fluorescence intensity was a measure for antibody binding and thus antigen surface expression. Reference ranges for antibody binding were obtained from 25 healthy control subjects.

Neutrophil Separation

Mixed leukocyte suspensions were prepared from CPDA–anticoagulated blood samples taken from healthy volunteers. After centrifugation at 100g for 10 minutes, platelet-rich plasma was removed. Mixed leukocyte suspensions were obtained after sedimentation in the presence of 3% dextran in normal saline (wt/vol) (MW 266 000 g per mol/L) and hypotonic lysis of erythrocytes at 4°C. Cells were washed twice in Tyrode's buffer.

Purified neutrophils were obtained from the mixed leukocyte suspensions by two-step density gradient centrifugation (Histopaque 1077 and 1119, Sigma).^{20 21 22} The neutrophil fraction was harvested, washed twice, and resuspended in Tyrode's buffer. After this separation procedure was performed, the neutrophil content was >99% with >95% viable cells (trypan blue exclusion).

Platelet Membrane Preparation

Platelet membranes were isolated from freshly prepared platelet concentrates as previously described.^{16 23} After preparation of the platelet concentrate, an antiplatelet cocktail that contained prostaglandin E₁ (200 μg/mL), theophylline (1 mmol/L), indomethacin (1 mmol/L), and adenosine (5 mmol/L)¹⁶ was added to prevent in vitro platelet activation. Platelets were isolated by centrifugation at 700g for 20 minutes, washed twice with Tyrode's buffer supplemented with the antiplatelet cocktail, and resuspended in 150 mmol/L NaCl, 10 mmol/L Tris buffer, pH 7.4 containing the proteinase inhibitors aprotinin (10 g/L), leupeptin (0.1 g/L), and PMSF (0.5 mmol/L). The suspension was placed in the pressure chamber of a cell disruption bomb (Parr Instruments Co), and the chamber was filled with nitrogen to achieve a pressure of 10 000 kPa. After 1 hour at room temperature, the suspension was centrifuged at 100 000g for 2 hours, and the pellet was resuspended in 150 mmol/L NaCl, 10 mmol/L Tris, pH 7.4 and layered over a 27% sucrose gradient prepared in the same buffer. After centrifugation at 30 000g for 3 hours, the platelet membranes were harvested; resuspended in 150 mmol/L NaCl, 10 mmol/L Tris, pH 7.4; homogenized with a Dounce homogenizer; and passed thorough a PD10 column (Pharmacia). The eluted fraction containing the isolated platelet membranes was collected and stored in aliquots at −70°C before use. The platelet membrane preparation used in the present study was characterized as described previously.¹⁶ SDS-PAGE and immunoblotting revealed GP IIb/IIIa as the most prominent protein along with small amounts of P-selectin and thrombospondin.¹⁶ In addition, flow cytometric analysis of membrane-coated beads suggested that the glycoprotein content of the membrane preparation resembled that of partially degranulated platelets.¹⁶ Agglutination assays of membrane-coated beads confirmed the functional integrity of the adhesion proteins.¹⁶

Inactivation of platelet membranes was performed by alkaline heat denaturation. Platelet membranes were incubated at 100°C for 20 minutes in the presence of 1 mol/L NaOH. After TCA precipitation, protein was washed with acetone to remove TCA, and additional washing steps were performed three times in PBS. Denatured platelet membranes were neutralized and resuspended in Tyrode's buffer and used in the same manner as intact membranes.

Coincubation of Neutrophils With Platelet Membranes

All coincubations were performed in Tyrode's buffer supplemented with 2 mmol/L CaCl₂ and 5 mmol/L HEPES. Leukocyte suspensions or purified neutrophils were added to suspensions with various concentrations of platelet membranes or 10⁻⁷ mol/L FMLP with a final concentration of 5000 cells/μL in a total volume of 400 μL. Incubation was performed in polystyrene tubes under constant shaking at 37°C. Time-course experiments showed that neutrophil-platelet aggregation was maximal after 10 minutes' incubation (data not shown). Therefore, dose-response experiments were performed at 10 minutes. To remove nonadherent platelet membranes, cell suspensions were then washed twice and resuspended in 150 μL Tyrode's buffer. Fluorescent labeling and flow cytometry were performed as described above.

Superoxide Anion Production

Superoxide anion production by neutrophils was determined by the superoxide dismutase–inhibited reduction of cytochrome C as previously described.²⁴ The standard reaction mixture contained 1.5×10⁵ neutrophils, 1000 U/L dismutase (30 000 U/mg, Sigma), 45 μmol/L cytochrome C, and platelet membranes or 10⁻⁷ mol/L FMLP at a volume of 500 μL. When the basal superoxide anion production was tested, FMLP and platelet membranes were omitted from the reaction mixture. After incubation for 15 minutes at 37°C under constant shaking, we determined the dismutase-inhibited reduction of cytochrome C by measuring the absorbance at 550 nm (E_{550 nm}=2.11×10⁴ mol/L⁻¹·cm⁻¹; Twinreader, Flow Laboratories). Results are expressed as superoxide anions (in nanomoles per liter)/5×10⁵ neutrophils/15 minutes.

Confocal Scanning Laser Microscopy

To investigate the nature of the neutrophil-platelet adhesion, images of the immunofluorescent samples from both patient groups and the control subjects were obtained with a confocal scanning laser microscope (Axiovert 35, Zeiss). In addition to the patient samples, the binding of platelet membranes to control neutrophils was investigated. Cellular structures were identified in the confocal mode and related to the bound antibody that was detected by use of fluorescence microscopy.

Other Methods

Cell counts were performed with a Sysmex Counter, model F-800 (Digitana GmbH).

No endotoxin contamination of platelet membrane preparations or of buffers was detected (E-toxate, Sigma).

Statistical Analysis

Results are expressed as mean±SEM unless otherwise indicated. Differences in frequencies between two groups were tested by χ² test with continuity correction. Because the Kolmogorov-Smirnov test showed that the data were not normally distributed, we chose the Mann-Whitney-White U test for comparison of the two groups. Linear correlations were obtained from least-squares analysis. A value of P<.05 in the two-tailed test was regarded as significant.

Study Population

The baseline characteristics of the patients with stable and unstable angina are summarized in Table 1⇑. Both groups were comparable with respect to age, sex, risk factor profile, and medication (Table 1⇑). We did not find significant differences in leukocyte or platelet counts (data not shown).

Neutrophil-Platelet Adhesion Ex Vivo

As shown in Fig 1⇓, circulating neutrophil-platelet conjugates in patients with unstable angina were significantly increased compared with patients with stable angina (P=.0001). Fig 2⇓ shows representative confocal and fluorescent images of neutrophils from a patient with unstable angina. Confocal scanning laser microscopy revealed that the increased anti-GP IIb/IIIa fluorescence of the neutrophils corresponded to adhering platelets.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Individual values and mean±SEM of anti-GP IIb/IIIa mean fluorescence intensity of neutrophils from patients with unstable angina (UA) and those with stable angina (SA). The gray bar represents the interquartile range of healthy control subjects.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Neutrophils from a patient with unstable angina after staining with FITC/anti-GP IIb/IIIa. In the confocal mode (A), adhering platelets are seen that correspond to GP IIb/IIIa-positive areas in the fluorescent mode (B).

The surface expression of CD11b was increased significantly by 36% in patients with unstable angina compared with those with stable angina (Fig 3⇓). Concomitantly, L-selectin surface expression in unstable angina decreased by 31% compared with stable angina (P=.0062; Fig 4⇓). In the study groups, the surface expression of CD11b and L-selectin was significantly correlated with anti-GP IIb/IIIa fluorescence on neutrophils (Fig 5⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Individual values and mean±SEM of anti-CD11b mean fluorescence intensity of neutrophils from patients with unstable angina (UA) and those with stable angina (SA). The gray bar represents the interquartile range of healthy control subjects.

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Individual values and mean±SEM of anti–L-selectin mean fluorescence intensity of neutrophils from patients with unstable angina (UA) and those with stable angina (SA). The gray bar represents the interquartile range of healthy control subjects.

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Relationship between anti-CD11b (A) or anti–L-selectin (anti-L-sel) mean fluorescence intensity (B) and anti-GP IIb/IIIa mean fluorescence intensity in the study groups.

In contrast to the data on neutrophils, we did not find significant differences in platelet immunofluorescence between the patient groups. Platelets from patients or healthy control subjects had similar surface GP IIb/IIIa expression. LIBS1 and P-selectin platelet surface expression were increased by trend (LIBS1, P=.07; P-selectin, P=.09) in the patients with unstable angina compared with the patients with stable angina (Table 2⇓). In the entire group of patients with unstable angina, the surface expression of LIBS1 and P-selectin on platelets was increased by trend compared with patients with stable angina (Table 3⇓). However, within the group of patients with unstable angina, those who presented with crescendo angina (n=14) showed a significantly higher LIBS1 and P-selectin surface expression on platelets than either patients with stable angina (P<.01) or the rest of the patients with unstable angina (LIBS1: 667±30 versus 452±50 mean fluorescence intensity, P=.003; P-selectin: 408±23 versus 337±23 mean fluorescence intensity, P=.001).

Coincubation of Leukocytes With Platelet Membranes

Neutrophils in mixed leukocyte suspensions contained low baseline levels of anti-GP IIb/IIIa fluorescence intensity, comparable to control samples obtained after whole-blood staining (Figs 1 and 6A⇑⇓). On the other hand, purified neutrophils showed negligible platelet binding (Fig 6A⇓). Confocal scanning laser microscopy of neutrophils in mixed leukocyte suspensions and of purified neutrophils confirmed these flow cytometry findings.

• Download figure
• Open in new tab
• Download powerpoint

Figure 6.

Effect of platelet membranes (filled symbols) or 10⁻⁷ mol/L FMLP (cross-hatched symbols) on anti-GP IIb/IIIa (A), anti-CD11b (B), and anti–L-selectin (C) mean fluorescence intensity of control neutrophils in mixed leukocyte suspensions (▪ and □) or of purified neutrophils (○ and ). Results are mean±SEM of six separate experiments.

In both cell preparations, fluorescent lining for GP IIb/IIIa was not observed under baseline conditions but was induced by increasing doses of purified platelet membranes (Fig 6A⇑). Stimulation with FMLP did not alter GP IIb/IIIa fluorescence of neutrophils in mixed leukocyte suspensions or of separated neutrophils (data not shown). In the presence of purified platelet membranes, however, the GP IIb/IIIa fluorescence intensity of neutrophils increased in a dose-dependent manner. Neutrophil-platelet membrane binding was more pronounced on neutrophils in mixed leukocyte suspensions than on purified neutrophils (Fig 6A⇑).

Increasing concentrations of platelet membranes induced an increase in CD11b surface expression on neutrophils. The extent of neutrophil surface CD11b upregulation was comparable to stimulation with 10⁻⁷ mol/L FMLP (Fig 6B⇑).

Neutrophils in mixed leukocyte suspensions showed high levels of anti–L-selectin binding, suggesting that they had not undergone major activation. Density gradient separation, however, caused some shedding of L-selectin. Increasing concentrations of platelet membranes induced a reduction in L-selectin surface expression of neutrophils in mixed leukocyte suspensions and purified neutrophils (Fig 6C⇑). The degree of L-selectin shedding after incubation with maximal concentrations of platelet membranes was comparable to that observed after stimulation with FMLP (Fig 6C⇑).

In addition to the changes in surface expression of adhesion receptors on neutrophils, platelet membranes induced a dose-dependent increase in superoxide anion production of purified neutrophils. This effect was observed at similar protein concentrations as for the effect on CD11b surface expression or L-selectin shedding (Fig 7⇓).

• Download figure
• Open in new tab
• Download powerpoint

Figure 7.

Effect of platelet membranes (filled symbols) or 10⁻⁷ mol/L FMLP (cross-hatched symbols) on superoxide anion production of control neutrophils. Results are mean±SEM of five separate experiments.

Denaturation of platelet membranes abolished neutrophil-platelet interaction as well as neutrophil activation (Table 3⇑).

Findings in Patients

Neutrophils from patients with unstable angina showed a substantially higher anti-GP IIb/IIIa binding than neutrophils from patients with stable angina. Neutrophils do not expose the GP IIb/IIIa complex themselves, and they can be clearly distinguished from potentially contaminating platelet aggregates by their anti-CD45 immunofluorescence signal. The anti-GP IIb/IIIa fluorescent signal detected on neutrophils, therefore, must be attributed to adhering platelets or microparticles. Confocal fluorescence laser microscopy confirmed that platelet adhesion was the source of the increased anti-GP IIb/IIIa immunofluorescence of neutrophils of patients with unstable angina.

The increase in neutrophil-platelet adhesion in patients with unstable angina was associated with an increased neutrophil anti-CD11b and a decreased anti-L-selectin immunofluorescence compared with patients with stable angina. After chemotactic stimulation, neutrophils translocated the β₂-integrin CD11b/CD18 from cellular stores to the plasma membrane and shed L-selectin.^{25 26 27} The observed changes in neutrophil surface receptors in unstable angina, therefore, indicated neutrophil activation. Hence, the positive correlation between anti-GP IIb/IIIa and anti-CD11b binding and the inverse correlation between anti–L-selectin and anti-GP IIb/IIIa binding suggests that neutrophil-platelet adhesion and neutrophil activation are interrelated.

There is increasing evidence that outside-in signaling by adhesion receptors can modulate cell function.²⁸ We therefore speculated that neutrophil-platelet adhesion might induce neutrophil activation. To test this hypothesis, we performed in vitro experiments.

In Vitro Stimulation of Neutrophils by Isolated Platelet Membranes

Coincubation experiments with neutrophils and platelets are difficult to interpret because putative changes in neutrophil function may be induced by the release of platelet mediators rather than by heterotypic adhesion. For this reason, we coincubated functionally intact isolated platelet membranes with neutrophils in mixed leukocyte suspensions and neutrophils purified by density gradient separation. Neutrophils in mixed leukocyte suspensions had similar surface properties as neutrophils stained in whole blood and showed residual platelet binding. On the other hand, purified neutrophils were essentially devoid of adhering platelets. This was achieved at the expense of altered surface properties, evidenced by the substantial shedding of L-selectin during separation.

Platelet membranes bind in a dose-dependent manner to both neutrophils in mixed leukocyte suspensions and purified neutrophils. As evidence of neutrophil activation induced by platelet membrane binding, we showed a dose-dependent increase in CD11b surface expression and shedding of L-selectin in both cell preparations. In support of this finding, we further demonstrated a dose-dependent induction of the respiratory burst reaction by platelet membrane binding on purified neutrophils. Alterations in neutrophil function were only induced by functionally intact platelet membranes, whereas denatured platelet membranes had no effect. Thus, independently of platelet mediator release, the interaction of neutrophils with platelet membranes carrying functionally intact surface receptors¹⁶ induced neutrophil activation. The receptors involved in this interaction remain to be clarified. Previous studies suggested that interactions between P-selectin and sialyl-Lewis^x binding sites on neutrophils play a major role in the adhesion of activated platelets and neutrophils.^{29 30 31} Other potential links are fibrinogen bridges between GP IIb/IIIa and CD11b/CD18^{32 33 34} and thrombospondin bridges between GP IV receptors on both cell lines.³⁵

Platelet membrane binding and the accompanying cellular responses were less pronounced in purified neutrophils than in neutrophils of mixed leukocyte suspensions. Altered neutrophil surface properties in the former preparation can largely explain this difference. Moreover, in mixed leukocyte suspensions, the interaction of platelet membranes with residual adherent platelets cannot be excluded and may present an additional mechanism for neutrophil activation. Such secondary binding of platelets or platelet membranes to platelets or platelet membranes already attached to neutrophils, however, may also play a role in vivo. Moreover, in vivo, release products from adherent platelets may augment neutrophil activation by adhesion receptors. Our experiments with functionally intact platelet membranes lend strong support to the concept that the interaction of platelet membrane receptors with neutrophils induces neutrophil activation.^{36 37} In demonstrating the differential modulation of Mac-1 and L-selectin by platelet membrane binding, the present study elucidates a potentially important new aspect of platelet-neutrophil interaction.

Limitations of the Study

Previous studies suggested that platelet activation is an essential prerequisite for leukocyte platelet adhesion.³⁸ In the present study, surface expression of the activated fibrinogen receptor and of P-selectin on platelets obtained from patients with unstable angina were only elevated by trend compared with stable angina. Nevertheless, we were unable to conclusively show platelet activation in unstable angina by immunologic markers. In contrast, previous studies revealed platelet activation in unstable angina, which was closely related to episodes of myocardial ischemia.^{10 12 39 40 41} There are several potential reasons for this discrepancy. First, substantial platelet activation may be present only in the patients with the most advanced disease. Consistent with this concept, platelet membrane markers showed considerable scatter within the entire group of patients with unstable angina. This scatter could be related in part to the variation in the patients' clinical status. Accordingly, the subgroup with crescendo angina, an entity closely related to plaque lability, showed significant elevations in activation-dependent platelet membrane markers. This finding concurs with previous studies by other investigators.^{13 40 41} Second, it has been shown that platelet activation in unstable angina is intermittent and related to episodes of ischemia. We did not take serial blood samples and, hence, might have missed temporary changes in relation to intermittent cardiac ischemia. Third, investigation of platelet function by analysis of cells rather than release products is always hampered by the rapid removal of activated single platelets from the circulation. Thus, we might have underestimated platelet activation by single-platelet flow cytometry. Platelet-neutrophil conjugates, however, may be more stable than activated single platelets, which may explain the difference in reliability between the assays. Thus, evaluating neutrophil-platelet adhesion may be a better way of investigating platelet activation in unstable angina than analysis of single platelets.

In studies of platelet function, there is always concern about changes occurring during and after blood sampling. Therefore, we took care to keep the sampling conditions constant between the patients. Moreover, neutrophil-platelet adhesion is known to depend on extracellular calcium.^{9 38 42} Because we used citrated blood samples, we can exclude substantial ex vivo neutrophil-platelet adhesion. Furthermore, previous studies^{5 7 38 42} have shown that anti-GP IIb/IIIa immunofluorescence represents in vivo neutrophil-platelet adhesion.

Pathophysiological Considerations and Clinical Perspectives

The present study demonstrated a functionally relevant interaction between neutrophils and platelets in unstable angina. It thereby reveals that neutrophil-platelet adhesion contributes to the activation of inflammatory cells and presents a novel link between thrombotic processes and inflammation. The findings of the present study, therefore, may help in the explanation of SIRS in acute coronary syndromes. In the future, neutrophil-platelet adhesion may become a target for pharmacological interventions to inhibit potentially harmful pathways of intercellular activation in thrombosis and inflammation.