Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Platelet secretion of vasoactive factors plays an important role in the development of atherosclerosis (Ross, 1999) and restenosis after coronary interventions (Chandrasekar and Tanguay, 2000). A major mitogenic compound released by aggregating platelets is platelet-derived growth factor (PDGF), which markedly stimulates smooth muscle cell proliferation and migration (Heldin and Westermark, 1999). Neutralizing antibodies to PDGF, and competitive PDGF receptor blocking agents have been shown to inhibit neointimal formation in animal and human studies (Serruys et al., 1997; Bilder et al., 1999; Waltenberger et al., 1999). PDGF exists in three different and biologically active isoforms (AA, BB, AB). Platelets contain mainly PDGF_AB and small amounts of PDGF_BB, which are stored in -granules (Hart et al., 1990; Heldin and Westermark, 1999). Activation of platelets, e.g., by thrombin or ADP, is associated with the translocation of CD62 (P-selectin) from the -granule membrane to the outer surface (Leytin et al., 2000a, b). Once exposed at activated platelets, CD62 allows the interaction of leukocytes with platelets by interacting with leukocyte PSGL-1, thereby triggering inflammatory responses (Furie and Furie, 1995; Evangelista et al., 1996; Zahler et al., 1999).

Flow cytometric determination of CD62 is commonly used to quantify platelet activation status (Hagberg and Lyberg, 2000; Leytin et al., 2000b; Zeiger et al., 2000). The expression of CD62 on the surface of platelets is considered to be an indicator for platelet degranulation and secretion (Gawaz et al., 1996; Michelson et al., 1996; Neumann et al., 1997) and a predictor of acute coronary events (Hollander et al., 1999). It has been shown that upon in vitro activation of platelets, CD62 is detected at the platelet surface, and both -granule-derived products, like -thromboglobulin, and dense granule-derived products, like serotonin, are released (Rand et al., 1996). However, the correlation between the surface expression of CD62 and granule content released is still unclear. Moreover, it remains to be determined whether changes in CD62 expression induced by either drugs or diseases are paralleled by changes in the secretion of granule-derived products. Therefore, the aim of the present study was to characterize the relationship between platelet activation status (translocation of CD62) and PDGF secretion from -granules in response to different concentrations of thrombin-receptor activating peptide (TRAP), and to determine whether such a relationship is maintained under antiplatelet drugs for which it is known that they either reduce CD62 expression (namely, the thienopyridine clopidogrel) (Rupprecht et al., 1998; Klinkhardt et al., 2000) or do not influence CD62 expression (namely the GPIIb/IIIa-inhibitor abciximab) (Fredrickson et al., 2000; Graff et al., 2001), or in clinical conditions that are associated with platelet hyper-reactivity (e.g., diabetes).

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Chemicals

TRAP (H-Ser-Phe-Leu-Leu-Arg-Asn-Pro-OH) was obtained from Bachem (Heidelberg, Germany). FITC-anti-CD62 antibody, PE-anti CD42b antibody (IgG1-mouse) and FACS solution for analysis in a FACScan cytometer was obtained from Becton Dickinson (Heidelberg, Germany). PDGF_AB was determined by an immunoassay (human PDGF_AB; Quantikine, R & D Systems, Wiesbaden, Germany), which has a 10% cross-reactivity with PDGF_AA and a 2% cross-reactivity with PDGF _BB and a limit of sensitivity of 8.4 pg/ml. In several samples, we also determined PDGF_BB with a newly available immunoassay (human PDGF_BB; Quantikine, R & D Systems), which has a 0.1% cross-reactivity with PDGF_AB and a limit of sensitivity of 15 pg/ml. Clopidogrel (Plavix) was obtained from Sanofi-Synthelabo (Berlin, Germany), and abciximab (Reopro) was a gift from Eli Lilly Inc. (Indianapolis, IN).

Experimental Protocols

The following experimental protocols were performed during this study.

Protocol 1. Relationship between platelet secretion of PDGF_AB and CD62 expression in healthy subjects. Blood was drawn from the antebrachial vein of male volunteers (n = 6, age 25-41 years) into 3.18% sodium citrate.

Protocol 2. Effect of ex vivo treatment in whole blood with the GPIIb/IIIa-antagonist abciximab on platelet secretion of PDGF_AB and expression of CD62. Blood was drawn from healthy subjects (n = 6, age 22-38 years) as indicated above and spiked with abciximab (5 µg/ml final concentration). This concentration of abciximab has been proven to confer >80% inhibition of platelet aggregation and GPIIb/IIIa-receptor activation (Klinkhardt et al., 2000).

Protocol 3. Effect of in vivo treatment with clopidogrel on platelet secretion of PDGF_AB and expression of CD62 in healthy subjects (n = 8, age 25-41 years). Platelet secretion of PDGF_AB and expression of CD62 was determined before and after oral administration of clopidogrel (loading dose of 2 × 75 mg/day, followed by 75 mg/day for 6 days). Blood was drawn as indicated above.

Procotol 4. Relationship between platelet secretion of PDGF_AB and CD62 expression in diabetic patients and age-related healthy control subjects. Patients with noninsulin-dependent diabetes mellitus (n = 8, age 56-71 years) were selected from the diabetes day clinic of the Deutsche Klinik of Diagnostik (Wiesbaden, Germany) and compared with a group of healthy control patients (n = 8, age 59-69 years). Blood was drawn as indicated above.

Flow Cytometry Analysis

Citrated whole blood (250 µl) was diluted 1:1 in Hepes buffer (20 mM Hepes, 137 nM NaCl, 2.7 mM KCl, 1 mM MgCl₂, 5.6 mM glucose and 1 g l¹ bovine serum albumin, pH 7.4) and carefully mixed. TRAP at a final concentration of 0 (i.e., unactivated), 1, 2, 5, 10, 20, or 30 µM in the case of protocol 1, and 5 µM for all other protocols, was added to blood samples to activate platelets. Thereafter, the sample (30 µl) was washed with Hepes buffer by centrifugation for 5 min at 750g. Platelet sediment was resuspended in Hepes buffer (200 µl) and incubated with saturating concentrations of FITC-anti-CD62 (10 µl) at room temperature for 30 min in darkness. Subsequently samples were incubated with saturating concentration of PE-anti-CD42b antibody (20 µl), which is used to set a gate for platelet events during the analysis. After incubation with labeled antibodies, samples were diluted with 1 ml of sodium citrate solution (3.8%) in Dulbecco's phosphate-buffered saline, and centrifuged at 750g for 5 min. The labeled platelet pellets were resuspended in 300 to 400 µl of the FACS solution. Acquisition and processing of data from 5,000 platelets were carried out with CONSORT software (Becton Dickinson). Fluorescence channels were set on logarithmic scales. Expression of CD62 was quantified either as mean fluorescence intensity (MFI, given in arbitrary units) or as CD62 positive platelets (%+) identified by binding of FITC-labeled CD62 antibody to the surface of platelets (Evangelista et al., 1996; Hagberg and Lyberg, 2000).

Determination of PDGF (Enzyme-Linked Immunosorbent Assay). Platelet rich plasma (PRP) was prepared by centrifugation of citrated blood at 754g for 5 min at 24°C. Subsequently, 1 ml of PRP from the upper part of the tube content was transferred into a new tube, and platelet count was obtained. These conditions have been proven to confer an optimal yield of platelets (400-500 · 10⁴ per µl) with a minimum contamination by leukocytes (0.01-0.03 · 10⁴ per µl). The platelet suspension was incubated at 37°C for 2 min and for protocol 1 subsequently stimulated with TRAP at concentrations of 0 (i.e., unactivated), 1, 2, 5, 10, 20, and 30 µM and incubated for 20 min at room temperature. In samples obtained during protocols 2 to 4, 0 and 5 µM TRAP was used. Thereafter, platelet aggregates were removed by centrifugation at 754g for 5 min, and 500 µl of the supernatant were transferred into new tubes, which were centrifuged for another 30 min with 48,000g. The supernatant was kept and stored at 70°C before use. Furthermore, one sample of platelet poor plasma was obtained from each subject by centrifugation of citrated whole blood with 2500g for 15 min. For the determination of the total amount of PDGF in PRP, platelets were lysated according to the following methods (Hart et al., 1990): 1 ml of PRP was mixed with 1 ml of lysate buffer (0.5% Triton X-100, 50 nM triethanolamine, diluted in 10 ml of Dulbecco's phosphate-buffered saline), then frozen and thawed three times before centrifuged as described above. The amount of PDGF_AB (and in some samples also of PDGF_BB) in the aliquots was determined by the use of a commercially available immunoassay. The concentration of PDGF in each PRP aliquot was adjusted to the actual platelet count of each sample and is quoted as nanograms per 10⁹ platelets.

Statistical Evaluation. Results are presented as mean values ± S.D. Concentration-response curves for the increase in CD62 expression or secretion of PDGF in response to TRAP (protocol 1) were established on the basis of a sigmoidal E_max model [E = (E_max · C)/(EC₅₀ + C)]. The minimization of model parameters by least squares based on the Newton-Gauss algorithm was carried out using a computer program (Scientist, MicroMath Inc., Salt Lake City, UT). The correlation between CD62 expression and PDGF_AB release following activation by increasing TRAP concentration has been determined using a second order polynomial function, and alternative curve functions were rejected on the basis of statistical comparison of the curve fit according to the Akaike criterion. The relationship between PDGF_AB and PDGF_BB release in the same sample was assessed by linear regression analysis and Spearman's rank order correlation coefficient. Differences between observations in samples spiked or not spiked with abciximab (protocol 2) and in diabetic patients and their controls (protocol 4) were assessed by the U test (Mann-Whitney). The differences before and after clopidogrel treatment (protocol 3) were assessed by paired t test.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

PDGF Response to Activation with TRAP (Protocol 1). Platelets activated by TRAP release PDGF in a concentration-dependent manner, and the concentration-response curve for TRAP yielded an EC₅₀ of 7.9 ± 2.0 µM. Lower concentrations of TRAP (1 and 2 µM) only slightly provoked PDGF_AB release, compared with unactivated platelets. Maximal PDGF_AB release was seen at 20 µM TRAP (Fig. 1A). The maximal PDGF_AB content as determined from platelet lysate (131 ng/10⁹ PLT) was only less above the maximum of the estimated TRAP concentration-response curve. In platelet poor plasma, the amount of PDGF was 4 ± 1 ng/ml. In a subset of samples of protocol 1 and 3, PDGF_BB release was also investigated. The ratio between the BB and the AB isoform, calculated from all samples was 0.09 ± 0.04. Linear regression analysis showed that the ratio was constant over all concentration levels of PDGF, and therefore seems not to depend on the strength of TRAP as inducer of PDGF release (Fig. 2). Furthermore, the ratio in those samples obtained in protocol 3 under clopidogrel were not different from the samples obtained during protocol 1. 

View larger version (13K): [in this window] [in a new window]   Fig. 1.   PDGFAB-release (ng/109 platelets) in PRP and CD62 expression (MFI) in whole blood after activation with TRAP (protocol 1). Mean values ± S.D. (n = 6).

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Linear regression analysis (dashed line, 95% confidence interval) of PDGFAB release and corresponding PDGFBB release in a subgroup of subjects from protocol 1.

CD62 Expression to Activation with TRAP and Correlation to PDGF Release (Protocol 1). CD62 expression was determined in all samples that were stimulated with TRAP. Differences in the shape of the concentration-response curve were dependent from the analysis of the flow cytometric data. After stimulation with 5 µM TRAP, the expression of CD62 quantified as CD62 positive (%+) platelets showed 76%+ platelets, suggesting a larger effect as when the response is given as MFI, where the increase in MFI from baseline approximated 50% of the maximal response. The EC₅₀ of TRAP was 2.4 ± 1.0 µM for the CD62 response in %+ and 4.3 ± 0.7 for MFI, in closer agreement to the EC₅₀ for release of PDGF (Fig. 1B). Furthermore, since the concentration-response curve for MFI showed a better correlation to the PDGF_AB release than expression of CD62%+ platelets after various TRAP concentrations (r² = 0.82 versus 0.74) (Fig. 3), the MFI of CD62 expression has been regarded for reporting the subsequent experiments. From protocol 1, we chose a TRAP concentration of 5 µM as stimulus for platelet activation in the subsequent experiments. At this concentration effects in CD62 expression or PDGF release induced by disease or drugs might be detected in either direction.

View larger version (27K): [in this window] [in a new window]   Fig. 3.   Correlation between CD62 expression (MFI) and PDGFAB release. Data have been fitted to a second order polynomial function, and the mean curve (solid line) and 95% confidence interval (dashed lines) are given. The inset shows the curve based on determination of the fraction of CD62 positive platelets in percentage (%).

Effect of Platelets Spiked in Vitro with Abciximab (Protocol 2). CD62 expression (MFI) was not affected by spiking the blood with abciximab (5 µM/ml) both in unactivated as well as in TRAP-activated samples. PDGF_AB secretion even showed a slight but not significant increase under abciximab after TRAP activation (29 to 42 ng/10⁹ PLT) (Table 1).

Effect of Clopidogrel on Activation of Platelets (Protocol 3). Baseline values of CD62 expression and PDGF_AB release were similar at days 1 and 6. After activation with 5 µM of TRAP, a significant decrease (p < 0.02) in CD62 expression (238 to 158 MFI), as well as PDGF_AB (23 to 12 ng/10⁹ PLT), from day 1 to day 6 could be demonstrated (Fig. 4). When the CD62-PDGF data pairs were superimposed to the curve obtained during protocol 1 from a different group of subjects, almost all data were within the 80% prediction interval (Fig. 5).

View larger version (54K): [in this window] [in a new window]   Fig. 4.   CD62 expression (MFI) (top) and PDGFAB release (ng/109 platelets) (bottom) in healthy volunteers (n = 8) before (day 1) and under (day 6) clopidogrel intake at baseline (BL) and after TRAP (5 µM) stimulation (protocol 3). Results are presented as means ± S.D.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   Baseline and TRAP-induced CD62 expression and PDGFAB release obtained in subgroups (diabetics, elderly controls, and healthy volunteers [before and under clopidogrel intake]) superimposed on the curve and 80% prediction interval (dashed line) of protocol 1 from a different subgroup of healthy subjects.

Platelet Activation in PRP from Patients with Diabetes Mellitus (Protocol 4). Compared with age-matched controls, platelets from diabetic patients showed a significantly elevated release of PDGF_AB in unactivated PRP (16.4 ng/10⁹ PLT versus 9.4 ng/10⁹ PLT, p < 0.01) as well as after stimulation with 5 µM of TRAP (45.8 versus 28.7 ng/10⁹ PLT, p < 0.03) (Table 1). CD62 expression in unactivated samples was similar in both groups but was significantly enhanced after TRAP stimulation in the diabetes group (276 to 217 MIF, p < 0.03). When the CD62-PDGF data pairs were superimposed to the data obtained from protocol 1, only a few data pairs fell of the 80% prediction interval of the reference curve (Fig. 5).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The principal findings presented in this paper are: 1) expression of CD62 as a constituent of platelet -granular membrane and secretion of PDGF, an important ingredient of -granules, can be stimulated by TRAP-induced activation in a dose-dependent fashion; 2) the activation marker and the secretion product are closely correlated with each other; and 3) changes in the CD62 expression induced by a drug, namely clopidogrel, or a disease, namely diabetes, are paralleled by changes in PDGF secretion. Although CD62 is perceived as an activation marker of platelets indicating enhanced aggregability and secretion of -granular content, the proof that the CD62 status and its modifications reflect directly the actual secretion of the most important platelet mitogen, PDGF, has not been given so far. This ex vivo-in vitro study shows that at least for the activation pathway provided by the PAR-1 thrombin receptor, for which TRAP is the selective agonist (Kahn et al., 1999), CD62 expression on platelets is a surrogate for their secretory activity. From the stimulus-response curves, we established a concentration of 5 µM TRAP that allows detection of drug- or disease-induced modifications in both directions. On the other hand, patients or treatment groups in our study were not distinguishable on the basis of their baseline secretion of PDGF or their baseline expression of CD62. Samples collected independently from the TRAP dose-response curve obtained from protocol 1 were mostly within the prediction interval (Fig. 5) and suggested a robust and reproducible relationship.

Flow cytometric detection of CD62 expressed on the platelet surface after -granule release is a common method used to characterize platelet activation in various experimental and clinical conditions (Evangelista et al., 1996; Leytin et al., 2000b). Parameters widely used to describe the activation status are either percentage of CD62 positive platelets in the total platelet population (%+) or the MFI of CD62 positive cells expressed in arbitrary units. Assaying %+ platelets will quantify the proportion of activated cells but will disregard the quantity of expressed platelet surface antibody CD62 (Michelson et al., 2000). MFI represents the mean epitope density of CD62 molecules on the average platelet surface and therefore reflects the activity of the single platelet but not their quantity (Evangelista et al., 1996). In our study, we found a better correlation to PDGF secretion when using the MFI to describe CD62 expression instead %+ platelets. The curve describing the relationship between CD62%+ platelets and PDGF release (Fig. 3, inset) starts extremely shallow, which implies that even upon mild activation a large proportion of the platelet population is detected as %+, but the overall secretion levels remains relatively low. In contrast, at levels of CD62%+ over 80%, PDGF secretion increases 2-fold despite only minor variations in CD62 expression.

PDGF is one of the major mediators for vascular smooth muscle cell proliferation and migration that occur in initial hyperplasia in the process of restenosis and atherosclerosis (Ross 1999). Platelet granules are reported to contain mainly PDGF_AB and PDGF_BB (Hammacher et al., 1988; Hart et al., 1990). Animal experiments confirmed the role of the B-chain in intimal hyperplasia, especially in preventing vascular smooth muscle cells from apoptosis (Leppanen et al., 2000). Accordingly, the growth activating potency of PDGF_BB is reported to be about 4-fold larger than that reported for PDGF_AB (Hart et al., 1990). From our data, based on enzyme-linked immunosorbent assay technique, the ratio of the PDGF isoforms AB and BB was approximately 1:10 and not 1:3 as quoted from a recent report employing reversed phase HPLC technique for differential PDGF assay (Hart et al., 1990).

We tried to prove whether the relationship between CD62 expression and secretion of PDGF is maintained under modifications induced by pharmacological or clinical conditions. The thienopyridine P2Y₁₂ receptor antagonists are the only antiplatelet drugs that are known to reduce CD62 expression (Rupprecht et al., 1998; Klinkhardt et al., 2000), whereas treatment with GPIIb/IIIa inhibitors (Fredrickson et al., 2000; Klinkhardt et al., 2000), or aspirin (Rinder et al., 1993; Fredrickson et al., 2000; Klinkhardt et al., 2000) did not. We could demonstrate a significant decrease in the platelet activation marker CD62 and the secretion product PDGF after 6 days of clopidogrel treatment. Since it has been observed that via stimulation of phosphatidylinositol-3 kinase the P2Y₁₂ receptor is a necessary constituent for sustained aggregation induced by TRAP (Trumel et al., 1999; Storey et al., 2000), a cross talk between signaling pathways of the PAR-1 thrombin receptor and purinergic receptors is likely and might explain why clopidogrel effects TRAP-induced degranulation.

In diabetic patients, platelet hyper-reagibility is implicated as a risk factor for both microvascular and macrovascular disease (Bern 1978; Barnett 1991) and is associated with an enhanced platelet aggregation response at sites of vascular injury (Winocour 1992) and enhanced CD62 expression on the platelet surface (Tschoepe et al., 1991; Rauch et al., 1999). However, it has not been determined whether platelets from diabetic subjects release more of the content of their granules in response to activation. Although our group of diabetic patients (noninsulin-dependent diabetes mellitus) was small, we observed a significantly higher PDGF-release from platelets after stimulation with TRAP than in age-matched controls. However, CD62 expression and PDGF release did not match in unactivated platelets, which might indicate that a separate mechanism must be taken into account to explain the dissociation between the parameters, at least in patients with noninsulin-dependent diabetes mellitus.

Thrombus-bound platelets at sites of vascular injury could act as a depot for -granule release of factors like PDGF with sustained effects on the remodeling process (Schini-Kerth et al., 1997). Therefore investigation of platelets activated by a defined stimulus, mimicking the local thrombin response might be closer to the situation of platelet depots in a thrombus than the determination of systemic serum growth factor levels. However, some limitations of our study results need attention. For platelet activation, we used TRAP (SFLLRNP), which is a synthetic peptide and stimulates the PAR-1 receptor without proteolytic cleavage like the physiological activator thrombin, which also activates PAR-4 and GPIb-receptors (Coughlin 1999; Kahn et al., 1999; Ofosu and Nyarko, 2000). On the other hand, it has been demonstrated that PAR-1 is the primary binding site for human thrombin at platelets, and PAR-1 binding contributes to almost all platelet-activating effects mediated by thrombin (Ofosu and Nyarko, 2000). PDGF itself interferes with platelet activation, and it has been shown in heparinized blood that PDGF_BB in concentrations of 100 ng/ml reduced TRAP-induced platelet aggregation and platelet microparticle formation by approximately 20% (Selheim et al., 1999). From our study, the steep shape of the curve describing the relationship between PDGF release and CD62 is determined by data pairs obtained after strong TRAP stimulation. In this part of the curve, the increase in the MFI seems to be terminated already despite a further increase in PDGF, which indeed might indicate a negative feedback of PDGF on platelet activation. Nevertheless, in the above-mentioned study (Selheim et al., 1999), PDGF although reducing platelet aggregation did not affect the expression of CD62.

In conclusion, our data support the utility of flow cytometric determination of CD62 as a target parameter during clinical studies with antiplatelet agents and in acute or chronic vascular diseases. The use of a defined stimulus, as used in common platelet aggregation tests, might allow for detection of drug- or disease-induced modifications and is also closely related to the secretion of the potent mitogen PDGF. However, marked discrepancies in the level of CD62 expression based on the flow cytometric protocol or device used have been described (Serebruany et al., 1999), and the need for standardization of the methodology must be emphasized.