- Email: [email protected]
IIIa antagonists
Thrombosis Research 108 (2003) 181 – 189
Regular Article
Regulation of clot retraction by glycoprotein IIb/IIIa antagonists Dietmar Seiffert *, Donna L. Pedicord, Cathy J. Kieras, Bokang He, Andrew M. Stern, Jeffrey T. Billheimer Chemical Enzymology, Experimental Station E400/3255, Bristol-Myers Squibb Company, P.O. Box 80400, Wilmington, DE 19880, USA Accepted 15 November 2002
Abstract Binding of fibrinogen to platelet glycoprotein (GP) IIb/IIIa induces clot retraction. Significant differences among GP IIb/IIIa antagonists were previously noted to inhibit thromboelastography in whole blood specimens. The relationship between efficacy of these agents and inhibition of clot retraction is unclear. Here, we use a plasma-free clot retraction assay to evaluate potency of GP IIb/IIIa antagonists to inhibit clot retraction and modulate platelet signaling, and to address whether these effects are realized in the clinically relevant dose range. The potencies for inhibition of clot retraction and aggregation are similar for antagonists with high affinity for resting platelets and slow off-rates, whereas lower affinity and fast off-rate antagonists are disproportionately less effective in blocking clot retraction. A positive correlation is observed between inhibition of clot retraction and inhibition of tyrosine dephosphorylation across a number of GP IIb/IIIa antagonist pharmacophores. For lower affinity and fast off-rate antagonists, the concentrations required for inhibition of clot retraction clearly exceed the clinical dose range. Site occupancy studies combined with clot retraction experiments addressed whether high affinity and slow off-rate compounds can alter clot retraction during the dosing interval. Binding studies using [3H] Roxifiban, a high affinity GP IIb/IIIa antagonist, indicate that occupancy of >95% of GP IIb/IIIa sites is required to inhibit clot retraction. This level of occupancy is not routinely achieved in the clinic and is not tolerated, at least for chronic therapy. These results suggest that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists. D 2002 Elsevier Science Ltd. All rights reserved. Keywords: Clot retraction; Platelets; GP IIb/IIIa; GP IIb/IIIa antagonists
1. Introduction Intravenous glycoprotein (GP) IIb/IIIa antagonists, including Abciximab, Eptifibatide, and Aggrastat, are widely used for the prevention of thrombotic complications in patients undergoing coronary revascularization procedures [1– 6]. In addition, a number of orally bioavailable GP IIb/ IIIa antagonists underwent clinical development but failed to demonstrate efficacy (reviewed in Refs. [7– 9]). GP IIb/IIIa antagonists block the binding of fibrinogen and other RGDcontaining adhesive glycoproteins to platelet GP IIb/IIIa, thereby inhibiting platelet aggregation. GP IIb/IIIa is present at approximately 40,000 to 80,000 copies per platelets [10 – 12]. In addition, an intracellular pool of GP IIb/IIIa becomes Abbreviations: ADP, adenosine diphosphate; GP IIb/IIIa, glycoprotein IIb/IIIa; TRAP, thrombin receptor activating peptide. * Corresponding author. Tel.: +1-302-467-5034; fax: +1-302-4676820. E-mail address: [email protected] (D. Seiffert).
surface expressed upon stimulation with strong agonists [11,13,14]. Occupancy of GP IIb/IIIa leads to a series of intracellular responses, referred to as outside-in signaling (reviewed in Ref. [15]). This process is triggered by ligand-induced oligomerization of GP IIb/IIIa [16] and followed by cytoskeletal rearrangements and a characteristic wave of tyrosine phosphorylation and dephosphorylation of a number of platelet molecules (reviewed in Refs. [17,18]). These events lead to a process commonly referred to as clot retraction [19 – 23]. The in vivo significance of clot retraction for hemostasis and efficacy of antithrombotic agents is less understood. The retracted clot may improve the mechanical stability of thrombi under conditions of high shear stress. In addition, a retracted clot may provide less surface area for fibrinolytic proteins. As a result, clot retraction may reduce susceptibility to fibrinolysis. We and others reported that GP IIb/IIIa antagonists retard clot retraction as measured by thromboelastography [19 – 22]. The relationship between efficacy of these agents and
0049-3848/02/$ - see front matter D 2002 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0049-3848(02)00395-X
182
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
inhibition of clot retraction is unclear. Recently, inhibition of clot retraction by GP IIb/IIIa antagonists was also reported to retard tyrosine dephosphorylation in a purified platelet/ fibrin clot retraction model, suggestive of a tyrosine dephosphorylation activity regulated by GP IIb/IIIa engagement during retraction [24]. Interestingly, low affinity GP IIb/IIIa antagonists like arginine– glycine –aspartic acid peptides or gamma chain peptides failed to block clot retraction and tyrosine dephosphorylation. The peptidomimetic GP IIb/IIIa antagonist Eptifibatide blocked both events in concentration ranges that are not clinically achievable. The question whether other GP IIb/IIIa antagonists in clinical use or previously in development alter clot retraction associated tyrosine dephosphorylation remains unanswered. Here, we provide evidence that high affinity and slow offrate antagonists like Abciximab and Roxifiban block clot retraction and tyrosine dephosphorylation with potencies similar to that required to block platelet aggregation. In contrast, lower affinity and fast off-rate antagonists required at least 10-fold higher concentrations to block clot retraction compared to inhibition of aggregation. Results are presented that a high level of occupancy of GP IIb/IIIa sites is required for inhibition of outside-in signaling and clot retraction. This level of occupancy is not routinely achieved in the clinically relevant dose range, suggesting that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists.
2. Materials and methods 2.1. GP IIb/IIIa antagonists Roxifiban, [3H] Roxifiban (specific activity 63 Ci/mmol) and DMP 802 were described previously [25 – 27]. Drs. Gary Cain, Doug Batt, and Joanne Smallheer (Bristol-Myers Squibb) are acknowledged for their preparation of OrbofiTable 1
ban, Lotrafiban and Sibrafiban. All small molecule GP IIb/ IIIa antagonists were characterized as the free acid form (see Table 1). Abciximab was obtained from Centocor Leiden. A summary table of platelet binding kinetic of GP IIb/ IIIa antagonists to activated and resting human platelets comprised of previously published and unpublished results was kindly provided by Drs. G. Feuerstein and S. Mousa, Bristol-Myers Squibb. Platelet aggregation and binding studies in platelet-rich plasma induced by 20 AM ADP were carried out as described [28]. The antagonists are categorized, based on the classification suggested by Mousa et al. [22], in class I compounds, characterized by high binding affinity for resting and stimulated platelets and slow offrates, and class II compounds with increased binding to stimulated platelets and fast off-rates. The target level of ADP-induced platelet aggregation inhibition during clinical trials with the orally bioavailable agents (Roxifiban, Orbofiban, Lotrafiban, Sibrafiban) was 30– 70%. Thus, the IC50 values for platelet aggregation (Table 1) can be used as a guidance for the clinically achieved plasma concentrations. 2.2. Washed platelet preparation Blood was collected from non-fasting human donors who had taken no aspirin in the preceding 2 weeks in 5 ml vacutainers filled with 500 Al of acid citrate dextrose (130 mM citric acid, 124 mM sodium citrate, 10 mM glucose, pH 4.0). After addition of prostaglandin E1 (1 AM final concentration; Sigma), the blood was incubated at room temperature for 15 min followed by centrifugation at 150 g for 10 min. The platelet-rich plasma was collected and platelets pelleted by centrifugation at 1200 g for 10 min. The platelet-poor plasma supernatant was removed and platelets resuspended in platelet resuspension buffer (10 mM Hepes, pH 7.4, 140 mM NaCl, 3 mM KCl, 0.5 mM MgCl2, 5 mM NaHCO3, 10 mM glucose) containing 4 mg/ml human fibrinogen (Cal-
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
biochem). The fibrinogen was extensively dialyzed against platelet resuspension buffer and stored at 4 jC for up to 1 week. The platelet concentration was adjusted to 6 108/ml. 2.3. Fibrin clot retraction assay The fibrin clot retraction assay was essentially performed as previously described [24]. Flat-bottom clear non-siliconized borosilicate glass vials (National Scientific) were used for retraction assays. Briefly, a 100 Al 6% (w/v) acrylamide cushion was polymerized at the bottom of the tube, and tubes were rinsed extensively with platelet resuspension buffer. A 200-Al aliquot of the above platelet suspension was added to
183
the glass tubes and pre-incubated with GP IIb/IIIa antagonists for 10 min at 37 jC. Blood clotting was initiated by quickly pipetting up and down 200 Al of 1 U/ml human thrombin (Sigma) and 2 mM CaCl2 in platelet resuspension buffer. The concentration of GP IIb/IIIa antagonists stated in the text and figures is the final concentration after 1/2 dilution of the platelets with thrombin and calcium. Clot retraction was documented by Polaroid photography for 1 h at 37 jC. Pictures were scanned and clot area relative to the 5-min time point were assessed. Data are expressed as follows: percentage of retraction=(1 area t/area t5 min) 100. Preliminary experiments revealed that clot retraction is similar between 1 and 10 mM of CaCl2 (final concentration) and
Fig. 1. Clot retraction and tyrosine phosphorylation in the presence of Roxifiban. Panel A: Washed platelets were pre-incubated with Roxifiban or vehicle control, clot formation was initiated with thrombin and calcium and followed for 1 h (see Materials and methods). The photographs of the 60-min time points are depicted in the insert. The final concentrations of Roxifiban (nM) are indicated to the right in panel A. Panel B: After 60 min, clots were extracted and analyzed for tyrosine phosphorylation after fractionation by SDS-PAGE (see Materials and methods). The relative migration of molecular weight standards is indicated. Numbering in panel A, insert, and panel B: 1, vehicle control; 2, 3 nM; 3, 10 nM; 4, 30 nM; 5, 100 nM; 6, 300 nM; 7, 1000 nM. Asterisk: Protein quantified to estimate the fold change in phosphorylation.
184
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
thrombin concentrations from 0.1 to 1 U/ml (final concentration). The results presented are representative for two independent experiments using two different donors. 2.4. Immunoblotting analysis Clots were solubilized in 400 Al of sample buffer (252 mM Tris, pH 6.8 containing 40% glycerol, 4% SDS, 0.25% Bromophenol Blue, 8 M urea and 32.5 mM dithiothreitol) and heated at 90 jC for 60 min. Samples were fractionated immediately on 4 –12% gradient gels (Novex) or stored at 80 jC. Gels were electroblotted to nitrocellulose membranes (0.45-Am pore size; Schleicher & Schuell). Complete
transfer was evaluated by use of pre-stained molecular weight markers (Gibco). The membranes were rinsed in distilled water for 2 min and blocked for 1 h in 50 mM Tris HCl, pH 7.5 containing 150 mM NaCl, 0.1% Tween 20 (Sigma) and 5% bovine serum albumin (ICN). Primary antibody (horseradish peroxidase-labeled anti-phosphotyrosine clone 4G10; 15 ng/ml; Upstate Biotechnology) was added in the above buffer containing 3% albumin for 1 h at room temperature. The blots were extensively washed in 50 mM Tris – HCl, pH 7.5 containing 150 mM NaCl and 0.1% Tween 20, developed with ECL Plus reagent (Amersham) and exposed to Kodak Biomax film. The films were scanned and band area and gray scale pixel values were quantified.
Fig. 2. Clot retraction and tyrosine phosphorylation in the presence of Orbofiban. Panel A: Kinetics of clot retraction as a function of Orbofiban concentration (compare Fig. 1). The photographs of the 60-min time points are depicted in the insert. The final concentrations of Orbofiban (nM) are indicated to the right in panel A. Panel B: Tyrosine phosphorylation pattern at the 60-min time point in the presence of Orbofiban (compare Fig. 1). Numbering in panel A, insert, and panel B, gel picture: 1, vehicle control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM; 7, 3000 nM.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
The membranes were stained with Amidoblack (Sigma) to confirm even loading and immunoblotting results were normalized based on the fibrinogen/fibrin beta chain intensity. Each experiment was performed at least twice with different donors and representative results are shown. The effects of GP IIb/IIIa on the tyrosine phosphorylation pattern are not limited to a specific phosphoprotein, but effects the overall phosphorylation pattern of multiple platelet proteins. In initial experiments, the extent of phosphorylation as a function of GP IIb/IIIa antagonist concentration was quantified for four different phosphoproteins and the fold change was similar. In subsequent studies presented here, the analysis was limited to a single protein that could be easily separated from other phosphoproteins by SDSPAGE (indicated by asterisk in Fig. 1). 2.5. Receptor binding studies [3H] Roxifiban was diluted 1/10 with unlabeled compound. Washed platelets were prepared as above and incu-
185
bated with labeled antagonists in centrifuge tubes (1.5 ml) for 5 min. The samples were split into three aliquots. The first aliquot was processed for clot retraction as described above. After 1 h, the sample (400 Al) containing the clot was transferred to centrifuge tubes (1 ml; Beckman) and centrifuged at 300,000 g for 5 min. Supernatants were removed and clot rinsed with 400 Al of phosphate buffered saline. The clot was resuspended in 250 Al of sample buffer (without Bromophenol Blue) overnight at room temperature. Samples were subsequently transferred to 1.5 ml centrifuge tubes and incubated at 90 jC for 60 min to completely dissolve the clots. Upon cooling, entire sample was added to a scintillation vial and counted in the presence of 250 Al of scintillation cocktail (Ultima Gold; Packard). The other two samples were diluted 1/2 in platelet resuspension buffer containing 10 AM thrombin receptor activation peptide (TRAP 6-mer; Peninsula Laboratories) or saline control for 15 min. The platelet pellet was isolated by centrifugation at 6000 g for 5 min, rinsed with 250 Al of saline containing 1% bovine serum albumin and re-centri-
Fig. 3. Clot retraction in the presence of the class I compounds Abciximab and DMP802. Panel A: Extent of clot retraction as a function of Abciximab and DMP802 concentration at the 60-min time point (compare Fig. 1). The extent of retraction using 100 nM of Roxifiban for the same donor is indicted by asterisk. Panel B: Tyrosine phosphorylation pattern at the 60-min time point (compare Fig. 1). Lane numbering: Abciximab: 1, saline control; 2, 2 nM; 3, 6 nM; 4, 20 nM; 5, 60 nM; 6, 200 nM. DMP802: 1, saline control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM.
186
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
fuged. The platelet pellet was solubilized in 500 Al of 1% Triton X-100, 0.1% bovine serum albumin overnight at room temperature. Specific binding was determined by adding a 50-fold excess of unlabeled compound and contributed to more than 95% of total binding. After addition of an equal volume of scintillation cocktail (Ultima Gold; Packard), radioactivity was determined by beta counting. The results presented are representative for two independent experiments using two different donors.
3. Results 3.1. Roxifiban and Orbofiban differ in blocking clot retraction and tyrosine dephophorylation in the clinically relevant dose range The kinetics of clot retraction and tyrosine phosphorylation were compared between two bioavailable small molecule GP IIb/IIIa antagonists representing class I (Roxifiban; Fig. 1) and class II (Orbofiban; Fig. 2) compounds. Roxifiban concentrations up to 10 nM were ineffective in retarding clot retraction and tyrosine phosphorylation. A
steep concentration dependency was observed, and nearly complete inhibition occurred at 30 nM (Fig. 1A). Roxifiban has a high affinity for resting platelets (compare Table 1) and titrates available GP IIb/IIIa binding sites on platelets. The titration-like binding isotherm is consistent with the steep concentration dependency (Hill slope >1) observed in these experiments. Concentration-dependency experiments using a more narrow concentration range revealed that varying Roxifiban concentrations by as little as 5 nM was sufficient to discriminate between complete inhibition and no effect (not shown). It should be noted that up to 2.5-fold differences between donors in the concentration of Roxifiban required for complete inhibition of clot retraction were observed. The effects of Roxifiban on clot retraction could be independently confirmed by analyzing the tyrosine phosphorylation pattern after 60 min of incubation (Fig. 1B). Concentrations of Roxifiban that blocked clot retraction also blocked tyrosine dephosphorylation (Fig. 1). The steady state levels of phosphoproteins were increased up to three-fold in the presence of Roxifiban compared to the no drug control based on quantification of the band indicated with asterisk in Fig. 1. This positive correlation suggests that both events are mechanistically linked.
Fig. 4. Clot retraction in the presence of the class II compounds Lotrafiban and Sibrafiban. Panel A: Extent of clot retraction as a function of Lotrafiban and Sibrafiban concentration at the 60-min time point (compare Fig. 1). The extent of retraction using 100 nM of Roxifiban for the same two donors is indicted by asterisk. Panel B: Tyrosine phosphorylation pattern at the 60-min time point (compare Fig. 1). Lane numbering: 1, vehicle control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM; 7, 3000 nM.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
Using the same platelet preparation for Orbofiban (Fig. 2), concentrations up to 3 AM were insufficient to completely inhibit clot retraction (Fig. 2A). The concentrations required for partial inhibition of clot retraction are outside the clinically achievable dose range. While no complete inhibition of clot retraction could be obtained with Orbofiban, the concentration-dependency curve appears more shallow compared to Roxifiban (compare Figs. 1A and 2A), suggesting that Orbofiban effects are not limited by the concentration of agonist relative to available GP IIb/IIIa binding sites. Again, the effects of Orbofiban on clot retraction could be independently confirmed by analyzing the tyrosine phosphorylation pattern after 60 min of incubation (Fig. 2B). To begin to understand whether the observed differences between Roxifiban and Orbofiban are related to receptor binding isotherms or chemical structure, additional members of class I (Abciximab and DMP802; Fig. 3) and class II antagonists (Sibrafiban and Lotrafiban; Fig. 4) were characterized. Slow off-rate compounds blocked clot retraction with high potency (Fig. 3A). Similarly to Roxifiban, steep concentration response curves were observed for both Abciximab and DMP802. Again, the inhibition of clot retraction paralleled inhibition of dephosphorylation (Fig. 3B). It should be noted that inhibition of clot retraction by Abciximab appears to be within the pharmacologically achievable dose range after a bolus injection. In contrast, the class II antagonists Lotrafiban and Sibrafiban were less potent inhibitors of clot retraction, the concentration response curves were shallow, and inhibition of clot retraction and inhibition of dephosphorylation appeared closely coupled (Fig. 4). It should be noted that the concentrations of both compounds required to significantly inhibit clot retraction are not within the therapeutic obtainable range. 3.2. Fractional GP IIb/IIIa receptor occupancy required for inhibition of clot retraction High affinity, slow off-rate compounds like Roxifiban are expected to ‘titrate’ available platelet GP IIb/IIIa sites. Under our experimental conditions, the molar concentration of GP IIb/IIIa binding sites under nonstimulated conditions is on average 25 nM (assuming 50,000 receptors per nonstimulated platelet and 300,000 platelets per Al). Platelets were incubated with a concentration range of [3H] labeled Roxifiban, bound and free ligand separated by centrifugation, and occupancy calculated based on the known specific activity of the ligand. [3H] Rosifiban bound to resting platelets in a concentration-dependent manner with a maximal occupancy (25 nM) occurring at approximately 40 nM of Roxifiban (Fig. 5). TRAP stimulation facilitates the translocation of the internal pool of GP IIb/IIIa to the platelet surface. Binding curves to resting and TRAP stimulated platelets were parallel up to 40 nM. However, maximal occupancy increased to 40 nM when platelets were stimulated with TRAP. At Roxifiban concentrations of 100 and 150 nM, a slight decrease in receptor occupancy was
187
Fig. 5. Inhibition of clot retraction versus receptor occupancy. Washed platelets were pre-incubated with [3H] Roxifiban for 5 min. One aliquot was processed for clot retraction. After 60 min of incubation, clots were harvested and centrifuged at 300,000 g for 20 min. After washing, clot associated radioactivity was quantified (clot bound [thrombin]; filled squares). The concentration of Roxifiban to completely inhibit clot retraction after 30 min incubation at 37 jC is indicated by the arrow. Platelets were also treated with saline (platelet bound [no TRAP]) or 10 AM TRAP (platelet bound [TRAP 10 AM]) for 15 min at 37 jC. At the end of the incubation period, platelets were collected by centrifugation, washed, and associated radioactivity quantified.
noted. The latter may be an experimental artifact due to selective recovery of small platelet aggregates compared to single platelets in the presence of high concentrations of GP IIb/IIIa antagonists. A third sample was processed in parallel for clot retraction. Complete inhibition of clot retraction in this donor was observed at 70 nM. These observations suggest that a high, sustained level of receptor occupancy by GP IIb/IIIa antagonists is required to prevent clot retraction. Indeed, quantification of clot bound Roxifiban revealed a high (>95%) occupancy of available GP IIb/IIIa sites, based on comparison to the TRAP-stimulated specimen, at concentrations required for inhibition of clot retraction (Fig. 5). Expressed in different terms, only a small number of fibrinogen/fibrin engaged GP IIb/IIIa sites are required for effective clot retraction.
4. Discussion Inhibition of clot retraction is associated with prevention of platelet tyrosine dephosphorylation and suggested that GP IIb/IIIa engagement prevents the activation of a tyrosine phosphatase activity [24]. Here, we provide further evidence that inhibition of tyrosine dephosphorylation and inhibition
188
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
of clot retraction are tightly coupled events. This conclusion is based on the characterization of five chemically distinct small molecule GP IIb/IIIa antagonists and one function blocking antibody spanning two log orders of potency with respect to inhibition of clot retraction and inhibition of tyrosine dephosphorylation. Taken together, these observations support the idea that clot retraction and tyrosine dephosphorylation are mechanistically linked. Evidence is provided that GP IIb/IIIa antagonists in clinical use or in development differ in their ability to regulate clot retraction. A number of mechanistic considerations predict that class I GP IIb/IIIa antagonists (binding to resting and stimulated platelets with similar affinity; slow off-rate) are more efficacious compared to class II antagonists in clot retraction assays. In a platelet aggregation assay, the ability of GP IIb/IIIa antagonists to compete with fibrinogen binding is typically measured over a 5-min time period. In contrast, the clot retraction assay determines the competition of GP IIb/IIIa antagonists for fibrin binding over an extended time period, typically 1 or 2 h. While the initial interaction of fibrinogen with platelets is reversible, post-occupancy changes make the fibrinogen binding practically irreversible [29 – 31]. Similar conclusions may also hold true for fibrin. As a result, the ability of a GP IIb/IIIa antagonist to compete for fibrinogen binding is expected to decrease over time. In addition, fibrin polymerization is expected to increase the avidity of the binding events. Based on these considerations, a high affinity slow off-rate compound is expected to be more efficacious in preventing clot retraction. Indeed, this concept could be confirmed experimentally by comparing three slow off-rate compounds (Abciximab, DMP 802, Roxifiban) with three fast off-rate compounds (Orbofiban, Sibrafiban, Lotrafiban). While slow off-rate compounds prevented clot retraction in concentration ranges close to those required for inhibition of aggregation, fast off-rate compounds were 10 –20 fold less potent in preventing clot retraction. These results confirm and extend previous thromboelastographic studies [22,32]. For class II compounds (Orbofiban, Sibrafiban, Lotrafiban), the EC50s for inhibition of clot retraction are outside the clinically achievable dose range and thus can not contribute to the observed antithrombotic efficacy. Initial concentration-dependency experiment using class I compounds suggests that a GP IIb/IIIa antagonist concentrations close to the expected total GP IIb/IIIa binding sites present in the reaction mixture are required to block clot retraction. This conclusion is consistent with the steep concentration response curves observed for class I compounds. The GP IIb/IIIa concentration in human donors varies as much as three-fold between blood donors [10 –12]. The high variability in receptor number implies that in vitro concentration –response curves alone are not suitable to address the question of fractional receptor occupancy required to block clot retraction and to make predictions about the clinical situation. The availability of a radiolabeled analog of Roxifiban enabled us to address receptor occupancy requirements
in more detail, removing variable receptor numbers as a confounding factor. These experiments revealed that compound concentrations close to the total available binding sites are required to effectively block clot retraction. In summary, based on the characterization of chemically distinct GP IIb/IIIa antagonists spanning two log orders of potency for platelet binding, we provide evidence that inhibition of clot retraction and inhibition of platelet tyrosine dephosphorylation are closely coupled events. In addition, we provide evidence that inhibition of clot retraction occurs at levels of receptor occupancy not routinely achieved in the clinic, suggesting that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists.
Acknowledgements The authors like to thank Drs. G. Feuerstein and S. Mousa for providing a summary of receptor binding kinetics and platelet aggregation inhibition of GP IIb/IIIa antagonists (Table 1). Dr. Rosa is acknowledged for the helpful discussion at the onset of this work.
References [1] The EPIC Investigators. Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high risk angioplasty. New Engl J Med 1994;330:956 – 61. [2] The EPILOG Investigators. Platelet glycoprotein IIb/IIIa receptor blockade and low-dose heparin during percutaneous coronary revascularization. New Engl J Med 1997;336:1689 – 96. [3] The CAPTURE Investigators. Randomized placebo-controlled trial of abciximab before and during coronary intervention in refractory unstable angina: the CAPTURE study. Lancet 1997;349:1429 – 35. [4] The PRISM PLUS Investigators. Inhibition of the platelet glycoprotein IIb/IIIa receptor with tirofiban in unstable angina and non-Qwave myocardial infarction. New Engl J Med 1998;338:1488 – 97. [5] The PURSUIT Investigators. Inhibition of platelet glycoprotein IIb/ IIIa with eptifibratide in patients with acute coronary syndromes. New Engl J Med 1998;339:436 – 43. [6] The EPISTENT Investigators. Randomized placebo-controlled and balloon-angioplasty-controlled trial to assess the safety of coronary stenting with use of platelet glycoprotein IIb/IIIa blockage. Lancet 1998;352:87 – 92. [7] Cannon CP. Learning from the recently completed oral glycoprotein IIb/IIIa receptor antagonist trials. Clin Cardiol 2000;23(Suppl VI): 14 – 7. [8] Newby LK, McGuire DK. Oral platelet glycoprotein IIb/IIIa inhibitors. Curr Cardiol Rep 2000;2:372 – 7. [9] Chew DP, Moliterno DJ. A critical appraisal of platelet glycoprotein IIb/IIIa inhibition. J Am Coll Cardiol 2000;36:2028 – 35. [10] Phillips DR, Charo IF, Parise LV, Fitzgerald LA. The platelet membrane glycoprotein IIb/IIIa complex. Blood 1988;71:831 – 43. [11] Niiya K, Hodson E, Bader R, Byers-Ward V, Koziol JA, Plow EF, et al. Increased surface expression of the membrane glycoprotein IIb/IIIa complex induced by platelet activation. Relationship to the binding of fibrinogen and platelet aggregation. Blood 1987;70: 475 – 83. [12] Wagner CL, Mascelli MA, Neblock DS, Weisman HF, Coller BS, Jordan RE. Analysis of GP IIb/IIIa receptor number by quantification of c7E3 binding to human platelets. Blood 1996;88:907 – 14.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189 [13] Gogstad GO, Hagen I, Korsmo R, Solum NO. Characterization of the proteins of isolated human platelet alpha granules. Evidence for a separate alpha-granule pool of the glycoprotein IIb and IIIa. Biochim Biophys Acta 1981;670:150 – 62. [14] Woods VL, Wolff LE, Keller DM. Resting platelets contain a substantial centrally located pool of glycoprotein IIb – IIIa complex which may be accessible to some but not other extracellular proteins. J Biol Chem 1986;261:15242 – 51. [15] Plow EF, Shattil SJ. Integrin IIb/IIIa and platelet aggregation. In: Colman RW, Hirsh J, Marder VJ, Clowes AW, George JN, editors. Hemostasis and Thrombosis. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 479 – 92. [16] Shattil SJ, Kashiwagi H, Pampori N. Integrin signaling: the platelet paradigm. Blood 1998;91:2645 – 57. [17] Clark EA, Brugge JS. Tyrosine phosphorylation. In: Authi KS, Watson SP, editors. Platelets. Oxford: IRL Press; 1995. p. 199 – 216. [18] Abrams CS, Brass LF. Platelet signal transduction. In: Colman RW, Hirsh J, Marder VJ, Clowes AW, George JN, editors. Hemostasis and Thrombosis. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 540 – 1. [19] Cohen I, Burk DL, White JG. The effect of peptides and monoclonal antibodies that bind to platelet glycoprotein IIb – IIIa complex on the development of clot tension. Blood 1989;91:2645 – 57. [20] Carr ME, Carr SL, Hantgan RR, Braaten J. Glycoprotein IIb/IIIa blockade inhibits platelet-mediated force development and reduces gel elastic modulus. Thromb Haemost 1995;82:468 – 80. [21] Jennings LLK, White MM, Mandrell TD. Interspecies comparison of platelet aggregation, LIB expression and clot retraction: observed differences in GPIIb – IIIa functional activity. Thromb Haemost 1995; 74:1551 – 6. [22] Mousa SA, Khurana S, Forsythe MS. Comparative in vitro efficacy of different platelet glycoprotein IIb/IIIa antagonists on platelet-mediated
[23] [24]
[25]
[26]
[27]
[28]
[29]
[30]
[31] [32]
189
clot strength induced by tissue factor with use of thromboelastography. Arterioscler Thromb Vasc Biol 2000;20:1162 – 7. Cohen K, Gerrard JM, White JG. Ultrastructure of clots during isometric contraction. J Cell Biol 1982;93:775 – 83. Osdoit S, Rosa J-P. Fibrin clot retraction by human platelets correlates with GP IIb/IIIa integrin-dependent protein tyrosine dephosphorylation. J Biol Chem 2001;276:6703 – 10. Olson RE, Sielecki TM, Wityak J, Pinto DJ, Batt DG, Frietze WE, et al. Orally active isoxazoline glycoprotein IIb/IIIa antagonists with extended duration of action. J Med Chem 1999;42:1178 – 92. Mousa SA, Bozarth J, Naik U, Slee A. Platelet GP IIb/IIIa binding characteristics of small molecule RGD mimetic: distinct binding profile for Roxifiban. Br J Pharmacol 2001;133:331 – 6. Mousa SA, Bozarth J, Lorelli W, Forsythe M, Thoolen M, Slee A, et al. Antiplatelet efficacy of XV459, a novel nonpeptide platelet GP IIb/IIIa antagonist: comparative platelet binding profiles with c7E3. J Pharmacol Exp Ther 1988;286:1277 – 84. Mousa SA, Bozarth J, Lorelli W, Forsythe M, Thoolen M, Slee A, et al. Discovery of novel orally active non-peptide platelet GP IIb/ IIIa receptor antagonists, DMP754: comparative platelet binding profile with c7E3. J Pharmacol Exp Ther 1998;286:1277 – 84. Marguerie GA, Edgington TS, Plow EF. Interaction of fibrinogen with its platelet receptor is part of a multistep reaction in ADP-induced platelet aggregation. J Biol Chem 1980;255:154 – 61. Marguerie GA, Plow EF. Interaction of fibrinogen with its platelet receptor: kinetics and effect of pH and temperature. Biochemistry 1981;20:1074 – 80. Peerschke EIB, Wainer JA. Examination of irreversible platelet – fibrinogen interactions. Am J Physiol 1985;248:C466 – 72. Hantgan RR, Mousa SA. Inhibition of platelet-mediated clot retraction by integrin antagonists. Thromb Res 1998;89:271 – 9.
Regular Article
Regulation of clot retraction by glycoprotein IIb/IIIa antagonists Dietmar Seiffert *, Donna L. Pedicord, Cathy J. Kieras, Bokang He, Andrew M. Stern, Jeffrey T. Billheimer Chemical Enzymology, Experimental Station E400/3255, Bristol-Myers Squibb Company, P.O. Box 80400, Wilmington, DE 19880, USA Accepted 15 November 2002
Abstract Binding of fibrinogen to platelet glycoprotein (GP) IIb/IIIa induces clot retraction. Significant differences among GP IIb/IIIa antagonists were previously noted to inhibit thromboelastography in whole blood specimens. The relationship between efficacy of these agents and inhibition of clot retraction is unclear. Here, we use a plasma-free clot retraction assay to evaluate potency of GP IIb/IIIa antagonists to inhibit clot retraction and modulate platelet signaling, and to address whether these effects are realized in the clinically relevant dose range. The potencies for inhibition of clot retraction and aggregation are similar for antagonists with high affinity for resting platelets and slow off-rates, whereas lower affinity and fast off-rate antagonists are disproportionately less effective in blocking clot retraction. A positive correlation is observed between inhibition of clot retraction and inhibition of tyrosine dephosphorylation across a number of GP IIb/IIIa antagonist pharmacophores. For lower affinity and fast off-rate antagonists, the concentrations required for inhibition of clot retraction clearly exceed the clinical dose range. Site occupancy studies combined with clot retraction experiments addressed whether high affinity and slow off-rate compounds can alter clot retraction during the dosing interval. Binding studies using [3H] Roxifiban, a high affinity GP IIb/IIIa antagonist, indicate that occupancy of >95% of GP IIb/IIIa sites is required to inhibit clot retraction. This level of occupancy is not routinely achieved in the clinic and is not tolerated, at least for chronic therapy. These results suggest that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists. D 2002 Elsevier Science Ltd. All rights reserved. Keywords: Clot retraction; Platelets; GP IIb/IIIa; GP IIb/IIIa antagonists
1. Introduction Intravenous glycoprotein (GP) IIb/IIIa antagonists, including Abciximab, Eptifibatide, and Aggrastat, are widely used for the prevention of thrombotic complications in patients undergoing coronary revascularization procedures [1– 6]. In addition, a number of orally bioavailable GP IIb/ IIIa antagonists underwent clinical development but failed to demonstrate efficacy (reviewed in Refs. [7– 9]). GP IIb/IIIa antagonists block the binding of fibrinogen and other RGDcontaining adhesive glycoproteins to platelet GP IIb/IIIa, thereby inhibiting platelet aggregation. GP IIb/IIIa is present at approximately 40,000 to 80,000 copies per platelets [10 – 12]. In addition, an intracellular pool of GP IIb/IIIa becomes Abbreviations: ADP, adenosine diphosphate; GP IIb/IIIa, glycoprotein IIb/IIIa; TRAP, thrombin receptor activating peptide. * Corresponding author. Tel.: +1-302-467-5034; fax: +1-302-4676820. E-mail address: [email protected] (D. Seiffert).
surface expressed upon stimulation with strong agonists [11,13,14]. Occupancy of GP IIb/IIIa leads to a series of intracellular responses, referred to as outside-in signaling (reviewed in Ref. [15]). This process is triggered by ligand-induced oligomerization of GP IIb/IIIa [16] and followed by cytoskeletal rearrangements and a characteristic wave of tyrosine phosphorylation and dephosphorylation of a number of platelet molecules (reviewed in Refs. [17,18]). These events lead to a process commonly referred to as clot retraction [19 – 23]. The in vivo significance of clot retraction for hemostasis and efficacy of antithrombotic agents is less understood. The retracted clot may improve the mechanical stability of thrombi under conditions of high shear stress. In addition, a retracted clot may provide less surface area for fibrinolytic proteins. As a result, clot retraction may reduce susceptibility to fibrinolysis. We and others reported that GP IIb/IIIa antagonists retard clot retraction as measured by thromboelastography [19 – 22]. The relationship between efficacy of these agents and
0049-3848/02/$ - see front matter D 2002 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0049-3848(02)00395-X
182
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
inhibition of clot retraction is unclear. Recently, inhibition of clot retraction by GP IIb/IIIa antagonists was also reported to retard tyrosine dephosphorylation in a purified platelet/ fibrin clot retraction model, suggestive of a tyrosine dephosphorylation activity regulated by GP IIb/IIIa engagement during retraction [24]. Interestingly, low affinity GP IIb/IIIa antagonists like arginine– glycine –aspartic acid peptides or gamma chain peptides failed to block clot retraction and tyrosine dephosphorylation. The peptidomimetic GP IIb/IIIa antagonist Eptifibatide blocked both events in concentration ranges that are not clinically achievable. The question whether other GP IIb/IIIa antagonists in clinical use or previously in development alter clot retraction associated tyrosine dephosphorylation remains unanswered. Here, we provide evidence that high affinity and slow offrate antagonists like Abciximab and Roxifiban block clot retraction and tyrosine dephosphorylation with potencies similar to that required to block platelet aggregation. In contrast, lower affinity and fast off-rate antagonists required at least 10-fold higher concentrations to block clot retraction compared to inhibition of aggregation. Results are presented that a high level of occupancy of GP IIb/IIIa sites is required for inhibition of outside-in signaling and clot retraction. This level of occupancy is not routinely achieved in the clinically relevant dose range, suggesting that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists.
2. Materials and methods 2.1. GP IIb/IIIa antagonists Roxifiban, [3H] Roxifiban (specific activity 63 Ci/mmol) and DMP 802 were described previously [25 – 27]. Drs. Gary Cain, Doug Batt, and Joanne Smallheer (Bristol-Myers Squibb) are acknowledged for their preparation of OrbofiTable 1
ban, Lotrafiban and Sibrafiban. All small molecule GP IIb/ IIIa antagonists were characterized as the free acid form (see Table 1). Abciximab was obtained from Centocor Leiden. A summary table of platelet binding kinetic of GP IIb/ IIIa antagonists to activated and resting human platelets comprised of previously published and unpublished results was kindly provided by Drs. G. Feuerstein and S. Mousa, Bristol-Myers Squibb. Platelet aggregation and binding studies in platelet-rich plasma induced by 20 AM ADP were carried out as described [28]. The antagonists are categorized, based on the classification suggested by Mousa et al. [22], in class I compounds, characterized by high binding affinity for resting and stimulated platelets and slow offrates, and class II compounds with increased binding to stimulated platelets and fast off-rates. The target level of ADP-induced platelet aggregation inhibition during clinical trials with the orally bioavailable agents (Roxifiban, Orbofiban, Lotrafiban, Sibrafiban) was 30– 70%. Thus, the IC50 values for platelet aggregation (Table 1) can be used as a guidance for the clinically achieved plasma concentrations. 2.2. Washed platelet preparation Blood was collected from non-fasting human donors who had taken no aspirin in the preceding 2 weeks in 5 ml vacutainers filled with 500 Al of acid citrate dextrose (130 mM citric acid, 124 mM sodium citrate, 10 mM glucose, pH 4.0). After addition of prostaglandin E1 (1 AM final concentration; Sigma), the blood was incubated at room temperature for 15 min followed by centrifugation at 150 g for 10 min. The platelet-rich plasma was collected and platelets pelleted by centrifugation at 1200 g for 10 min. The platelet-poor plasma supernatant was removed and platelets resuspended in platelet resuspension buffer (10 mM Hepes, pH 7.4, 140 mM NaCl, 3 mM KCl, 0.5 mM MgCl2, 5 mM NaHCO3, 10 mM glucose) containing 4 mg/ml human fibrinogen (Cal-
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
biochem). The fibrinogen was extensively dialyzed against platelet resuspension buffer and stored at 4 jC for up to 1 week. The platelet concentration was adjusted to 6 108/ml. 2.3. Fibrin clot retraction assay The fibrin clot retraction assay was essentially performed as previously described [24]. Flat-bottom clear non-siliconized borosilicate glass vials (National Scientific) were used for retraction assays. Briefly, a 100 Al 6% (w/v) acrylamide cushion was polymerized at the bottom of the tube, and tubes were rinsed extensively with platelet resuspension buffer. A 200-Al aliquot of the above platelet suspension was added to
183
the glass tubes and pre-incubated with GP IIb/IIIa antagonists for 10 min at 37 jC. Blood clotting was initiated by quickly pipetting up and down 200 Al of 1 U/ml human thrombin (Sigma) and 2 mM CaCl2 in platelet resuspension buffer. The concentration of GP IIb/IIIa antagonists stated in the text and figures is the final concentration after 1/2 dilution of the platelets with thrombin and calcium. Clot retraction was documented by Polaroid photography for 1 h at 37 jC. Pictures were scanned and clot area relative to the 5-min time point were assessed. Data are expressed as follows: percentage of retraction=(1 area t/area t5 min) 100. Preliminary experiments revealed that clot retraction is similar between 1 and 10 mM of CaCl2 (final concentration) and
Fig. 1. Clot retraction and tyrosine phosphorylation in the presence of Roxifiban. Panel A: Washed platelets were pre-incubated with Roxifiban or vehicle control, clot formation was initiated with thrombin and calcium and followed for 1 h (see Materials and methods). The photographs of the 60-min time points are depicted in the insert. The final concentrations of Roxifiban (nM) are indicated to the right in panel A. Panel B: After 60 min, clots were extracted and analyzed for tyrosine phosphorylation after fractionation by SDS-PAGE (see Materials and methods). The relative migration of molecular weight standards is indicated. Numbering in panel A, insert, and panel B: 1, vehicle control; 2, 3 nM; 3, 10 nM; 4, 30 nM; 5, 100 nM; 6, 300 nM; 7, 1000 nM. Asterisk: Protein quantified to estimate the fold change in phosphorylation.
184
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
thrombin concentrations from 0.1 to 1 U/ml (final concentration). The results presented are representative for two independent experiments using two different donors. 2.4. Immunoblotting analysis Clots were solubilized in 400 Al of sample buffer (252 mM Tris, pH 6.8 containing 40% glycerol, 4% SDS, 0.25% Bromophenol Blue, 8 M urea and 32.5 mM dithiothreitol) and heated at 90 jC for 60 min. Samples were fractionated immediately on 4 –12% gradient gels (Novex) or stored at 80 jC. Gels were electroblotted to nitrocellulose membranes (0.45-Am pore size; Schleicher & Schuell). Complete
transfer was evaluated by use of pre-stained molecular weight markers (Gibco). The membranes were rinsed in distilled water for 2 min and blocked for 1 h in 50 mM Tris HCl, pH 7.5 containing 150 mM NaCl, 0.1% Tween 20 (Sigma) and 5% bovine serum albumin (ICN). Primary antibody (horseradish peroxidase-labeled anti-phosphotyrosine clone 4G10; 15 ng/ml; Upstate Biotechnology) was added in the above buffer containing 3% albumin for 1 h at room temperature. The blots were extensively washed in 50 mM Tris – HCl, pH 7.5 containing 150 mM NaCl and 0.1% Tween 20, developed with ECL Plus reagent (Amersham) and exposed to Kodak Biomax film. The films were scanned and band area and gray scale pixel values were quantified.
Fig. 2. Clot retraction and tyrosine phosphorylation in the presence of Orbofiban. Panel A: Kinetics of clot retraction as a function of Orbofiban concentration (compare Fig. 1). The photographs of the 60-min time points are depicted in the insert. The final concentrations of Orbofiban (nM) are indicated to the right in panel A. Panel B: Tyrosine phosphorylation pattern at the 60-min time point in the presence of Orbofiban (compare Fig. 1). Numbering in panel A, insert, and panel B, gel picture: 1, vehicle control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM; 7, 3000 nM.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
The membranes were stained with Amidoblack (Sigma) to confirm even loading and immunoblotting results were normalized based on the fibrinogen/fibrin beta chain intensity. Each experiment was performed at least twice with different donors and representative results are shown. The effects of GP IIb/IIIa on the tyrosine phosphorylation pattern are not limited to a specific phosphoprotein, but effects the overall phosphorylation pattern of multiple platelet proteins. In initial experiments, the extent of phosphorylation as a function of GP IIb/IIIa antagonist concentration was quantified for four different phosphoproteins and the fold change was similar. In subsequent studies presented here, the analysis was limited to a single protein that could be easily separated from other phosphoproteins by SDSPAGE (indicated by asterisk in Fig. 1). 2.5. Receptor binding studies [3H] Roxifiban was diluted 1/10 with unlabeled compound. Washed platelets were prepared as above and incu-
185
bated with labeled antagonists in centrifuge tubes (1.5 ml) for 5 min. The samples were split into three aliquots. The first aliquot was processed for clot retraction as described above. After 1 h, the sample (400 Al) containing the clot was transferred to centrifuge tubes (1 ml; Beckman) and centrifuged at 300,000 g for 5 min. Supernatants were removed and clot rinsed with 400 Al of phosphate buffered saline. The clot was resuspended in 250 Al of sample buffer (without Bromophenol Blue) overnight at room temperature. Samples were subsequently transferred to 1.5 ml centrifuge tubes and incubated at 90 jC for 60 min to completely dissolve the clots. Upon cooling, entire sample was added to a scintillation vial and counted in the presence of 250 Al of scintillation cocktail (Ultima Gold; Packard). The other two samples were diluted 1/2 in platelet resuspension buffer containing 10 AM thrombin receptor activation peptide (TRAP 6-mer; Peninsula Laboratories) or saline control for 15 min. The platelet pellet was isolated by centrifugation at 6000 g for 5 min, rinsed with 250 Al of saline containing 1% bovine serum albumin and re-centri-
Fig. 3. Clot retraction in the presence of the class I compounds Abciximab and DMP802. Panel A: Extent of clot retraction as a function of Abciximab and DMP802 concentration at the 60-min time point (compare Fig. 1). The extent of retraction using 100 nM of Roxifiban for the same donor is indicted by asterisk. Panel B: Tyrosine phosphorylation pattern at the 60-min time point (compare Fig. 1). Lane numbering: Abciximab: 1, saline control; 2, 2 nM; 3, 6 nM; 4, 20 nM; 5, 60 nM; 6, 200 nM. DMP802: 1, saline control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM.
186
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
fuged. The platelet pellet was solubilized in 500 Al of 1% Triton X-100, 0.1% bovine serum albumin overnight at room temperature. Specific binding was determined by adding a 50-fold excess of unlabeled compound and contributed to more than 95% of total binding. After addition of an equal volume of scintillation cocktail (Ultima Gold; Packard), radioactivity was determined by beta counting. The results presented are representative for two independent experiments using two different donors.
3. Results 3.1. Roxifiban and Orbofiban differ in blocking clot retraction and tyrosine dephophorylation in the clinically relevant dose range The kinetics of clot retraction and tyrosine phosphorylation were compared between two bioavailable small molecule GP IIb/IIIa antagonists representing class I (Roxifiban; Fig. 1) and class II (Orbofiban; Fig. 2) compounds. Roxifiban concentrations up to 10 nM were ineffective in retarding clot retraction and tyrosine phosphorylation. A
steep concentration dependency was observed, and nearly complete inhibition occurred at 30 nM (Fig. 1A). Roxifiban has a high affinity for resting platelets (compare Table 1) and titrates available GP IIb/IIIa binding sites on platelets. The titration-like binding isotherm is consistent with the steep concentration dependency (Hill slope >1) observed in these experiments. Concentration-dependency experiments using a more narrow concentration range revealed that varying Roxifiban concentrations by as little as 5 nM was sufficient to discriminate between complete inhibition and no effect (not shown). It should be noted that up to 2.5-fold differences between donors in the concentration of Roxifiban required for complete inhibition of clot retraction were observed. The effects of Roxifiban on clot retraction could be independently confirmed by analyzing the tyrosine phosphorylation pattern after 60 min of incubation (Fig. 1B). Concentrations of Roxifiban that blocked clot retraction also blocked tyrosine dephosphorylation (Fig. 1). The steady state levels of phosphoproteins were increased up to three-fold in the presence of Roxifiban compared to the no drug control based on quantification of the band indicated with asterisk in Fig. 1. This positive correlation suggests that both events are mechanistically linked.
Fig. 4. Clot retraction in the presence of the class II compounds Lotrafiban and Sibrafiban. Panel A: Extent of clot retraction as a function of Lotrafiban and Sibrafiban concentration at the 60-min time point (compare Fig. 1). The extent of retraction using 100 nM of Roxifiban for the same two donors is indicted by asterisk. Panel B: Tyrosine phosphorylation pattern at the 60-min time point (compare Fig. 1). Lane numbering: 1, vehicle control; 2, 10 nM; 3, 30 nM; 4, 100 nM; 5, 300 nM; 6, 1000 nM; 7, 3000 nM.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
Using the same platelet preparation for Orbofiban (Fig. 2), concentrations up to 3 AM were insufficient to completely inhibit clot retraction (Fig. 2A). The concentrations required for partial inhibition of clot retraction are outside the clinically achievable dose range. While no complete inhibition of clot retraction could be obtained with Orbofiban, the concentration-dependency curve appears more shallow compared to Roxifiban (compare Figs. 1A and 2A), suggesting that Orbofiban effects are not limited by the concentration of agonist relative to available GP IIb/IIIa binding sites. Again, the effects of Orbofiban on clot retraction could be independently confirmed by analyzing the tyrosine phosphorylation pattern after 60 min of incubation (Fig. 2B). To begin to understand whether the observed differences between Roxifiban and Orbofiban are related to receptor binding isotherms or chemical structure, additional members of class I (Abciximab and DMP802; Fig. 3) and class II antagonists (Sibrafiban and Lotrafiban; Fig. 4) were characterized. Slow off-rate compounds blocked clot retraction with high potency (Fig. 3A). Similarly to Roxifiban, steep concentration response curves were observed for both Abciximab and DMP802. Again, the inhibition of clot retraction paralleled inhibition of dephosphorylation (Fig. 3B). It should be noted that inhibition of clot retraction by Abciximab appears to be within the pharmacologically achievable dose range after a bolus injection. In contrast, the class II antagonists Lotrafiban and Sibrafiban were less potent inhibitors of clot retraction, the concentration response curves were shallow, and inhibition of clot retraction and inhibition of dephosphorylation appeared closely coupled (Fig. 4). It should be noted that the concentrations of both compounds required to significantly inhibit clot retraction are not within the therapeutic obtainable range. 3.2. Fractional GP IIb/IIIa receptor occupancy required for inhibition of clot retraction High affinity, slow off-rate compounds like Roxifiban are expected to ‘titrate’ available platelet GP IIb/IIIa sites. Under our experimental conditions, the molar concentration of GP IIb/IIIa binding sites under nonstimulated conditions is on average 25 nM (assuming 50,000 receptors per nonstimulated platelet and 300,000 platelets per Al). Platelets were incubated with a concentration range of [3H] labeled Roxifiban, bound and free ligand separated by centrifugation, and occupancy calculated based on the known specific activity of the ligand. [3H] Rosifiban bound to resting platelets in a concentration-dependent manner with a maximal occupancy (25 nM) occurring at approximately 40 nM of Roxifiban (Fig. 5). TRAP stimulation facilitates the translocation of the internal pool of GP IIb/IIIa to the platelet surface. Binding curves to resting and TRAP stimulated platelets were parallel up to 40 nM. However, maximal occupancy increased to 40 nM when platelets were stimulated with TRAP. At Roxifiban concentrations of 100 and 150 nM, a slight decrease in receptor occupancy was
187
Fig. 5. Inhibition of clot retraction versus receptor occupancy. Washed platelets were pre-incubated with [3H] Roxifiban for 5 min. One aliquot was processed for clot retraction. After 60 min of incubation, clots were harvested and centrifuged at 300,000 g for 20 min. After washing, clot associated radioactivity was quantified (clot bound [thrombin]; filled squares). The concentration of Roxifiban to completely inhibit clot retraction after 30 min incubation at 37 jC is indicated by the arrow. Platelets were also treated with saline (platelet bound [no TRAP]) or 10 AM TRAP (platelet bound [TRAP 10 AM]) for 15 min at 37 jC. At the end of the incubation period, platelets were collected by centrifugation, washed, and associated radioactivity quantified.
noted. The latter may be an experimental artifact due to selective recovery of small platelet aggregates compared to single platelets in the presence of high concentrations of GP IIb/IIIa antagonists. A third sample was processed in parallel for clot retraction. Complete inhibition of clot retraction in this donor was observed at 70 nM. These observations suggest that a high, sustained level of receptor occupancy by GP IIb/IIIa antagonists is required to prevent clot retraction. Indeed, quantification of clot bound Roxifiban revealed a high (>95%) occupancy of available GP IIb/IIIa sites, based on comparison to the TRAP-stimulated specimen, at concentrations required for inhibition of clot retraction (Fig. 5). Expressed in different terms, only a small number of fibrinogen/fibrin engaged GP IIb/IIIa sites are required for effective clot retraction.
4. Discussion Inhibition of clot retraction is associated with prevention of platelet tyrosine dephosphorylation and suggested that GP IIb/IIIa engagement prevents the activation of a tyrosine phosphatase activity [24]. Here, we provide further evidence that inhibition of tyrosine dephosphorylation and inhibition
188
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189
of clot retraction are tightly coupled events. This conclusion is based on the characterization of five chemically distinct small molecule GP IIb/IIIa antagonists and one function blocking antibody spanning two log orders of potency with respect to inhibition of clot retraction and inhibition of tyrosine dephosphorylation. Taken together, these observations support the idea that clot retraction and tyrosine dephosphorylation are mechanistically linked. Evidence is provided that GP IIb/IIIa antagonists in clinical use or in development differ in their ability to regulate clot retraction. A number of mechanistic considerations predict that class I GP IIb/IIIa antagonists (binding to resting and stimulated platelets with similar affinity; slow off-rate) are more efficacious compared to class II antagonists in clot retraction assays. In a platelet aggregation assay, the ability of GP IIb/IIIa antagonists to compete with fibrinogen binding is typically measured over a 5-min time period. In contrast, the clot retraction assay determines the competition of GP IIb/IIIa antagonists for fibrin binding over an extended time period, typically 1 or 2 h. While the initial interaction of fibrinogen with platelets is reversible, post-occupancy changes make the fibrinogen binding practically irreversible [29 – 31]. Similar conclusions may also hold true for fibrin. As a result, the ability of a GP IIb/IIIa antagonist to compete for fibrinogen binding is expected to decrease over time. In addition, fibrin polymerization is expected to increase the avidity of the binding events. Based on these considerations, a high affinity slow off-rate compound is expected to be more efficacious in preventing clot retraction. Indeed, this concept could be confirmed experimentally by comparing three slow off-rate compounds (Abciximab, DMP 802, Roxifiban) with three fast off-rate compounds (Orbofiban, Sibrafiban, Lotrafiban). While slow off-rate compounds prevented clot retraction in concentration ranges close to those required for inhibition of aggregation, fast off-rate compounds were 10 –20 fold less potent in preventing clot retraction. These results confirm and extend previous thromboelastographic studies [22,32]. For class II compounds (Orbofiban, Sibrafiban, Lotrafiban), the EC50s for inhibition of clot retraction are outside the clinically achievable dose range and thus can not contribute to the observed antithrombotic efficacy. Initial concentration-dependency experiment using class I compounds suggests that a GP IIb/IIIa antagonist concentrations close to the expected total GP IIb/IIIa binding sites present in the reaction mixture are required to block clot retraction. This conclusion is consistent with the steep concentration response curves observed for class I compounds. The GP IIb/IIIa concentration in human donors varies as much as three-fold between blood donors [10 –12]. The high variability in receptor number implies that in vitro concentration –response curves alone are not suitable to address the question of fractional receptor occupancy required to block clot retraction and to make predictions about the clinical situation. The availability of a radiolabeled analog of Roxifiban enabled us to address receptor occupancy requirements
in more detail, removing variable receptor numbers as a confounding factor. These experiments revealed that compound concentrations close to the total available binding sites are required to effectively block clot retraction. In summary, based on the characterization of chemically distinct GP IIb/IIIa antagonists spanning two log orders of potency for platelet binding, we provide evidence that inhibition of clot retraction and inhibition of platelet tyrosine dephosphorylation are closely coupled events. In addition, we provide evidence that inhibition of clot retraction occurs at levels of receptor occupancy not routinely achieved in the clinic, suggesting that inhibition of clot retraction is not necessary for efficacy of GP IIb/IIIa antagonists.
Acknowledgements The authors like to thank Drs. G. Feuerstein and S. Mousa for providing a summary of receptor binding kinetics and platelet aggregation inhibition of GP IIb/IIIa antagonists (Table 1). Dr. Rosa is acknowledged for the helpful discussion at the onset of this work.
References [1] The EPIC Investigators. Use of a monoclonal antibody directed against the platelet glycoprotein IIb/IIIa receptor in high risk angioplasty. New Engl J Med 1994;330:956 – 61. [2] The EPILOG Investigators. Platelet glycoprotein IIb/IIIa receptor blockade and low-dose heparin during percutaneous coronary revascularization. New Engl J Med 1997;336:1689 – 96. [3] The CAPTURE Investigators. Randomized placebo-controlled trial of abciximab before and during coronary intervention in refractory unstable angina: the CAPTURE study. Lancet 1997;349:1429 – 35. [4] The PRISM PLUS Investigators. Inhibition of the platelet glycoprotein IIb/IIIa receptor with tirofiban in unstable angina and non-Qwave myocardial infarction. New Engl J Med 1998;338:1488 – 97. [5] The PURSUIT Investigators. Inhibition of platelet glycoprotein IIb/ IIIa with eptifibratide in patients with acute coronary syndromes. New Engl J Med 1998;339:436 – 43. [6] The EPISTENT Investigators. Randomized placebo-controlled and balloon-angioplasty-controlled trial to assess the safety of coronary stenting with use of platelet glycoprotein IIb/IIIa blockage. Lancet 1998;352:87 – 92. [7] Cannon CP. Learning from the recently completed oral glycoprotein IIb/IIIa receptor antagonist trials. Clin Cardiol 2000;23(Suppl VI): 14 – 7. [8] Newby LK, McGuire DK. Oral platelet glycoprotein IIb/IIIa inhibitors. Curr Cardiol Rep 2000;2:372 – 7. [9] Chew DP, Moliterno DJ. A critical appraisal of platelet glycoprotein IIb/IIIa inhibition. J Am Coll Cardiol 2000;36:2028 – 35. [10] Phillips DR, Charo IF, Parise LV, Fitzgerald LA. The platelet membrane glycoprotein IIb/IIIa complex. Blood 1988;71:831 – 43. [11] Niiya K, Hodson E, Bader R, Byers-Ward V, Koziol JA, Plow EF, et al. Increased surface expression of the membrane glycoprotein IIb/IIIa complex induced by platelet activation. Relationship to the binding of fibrinogen and platelet aggregation. Blood 1987;70: 475 – 83. [12] Wagner CL, Mascelli MA, Neblock DS, Weisman HF, Coller BS, Jordan RE. Analysis of GP IIb/IIIa receptor number by quantification of c7E3 binding to human platelets. Blood 1996;88:907 – 14.
D. Seiffert et al. / Thrombosis Research 108 (2003) 181–189 [13] Gogstad GO, Hagen I, Korsmo R, Solum NO. Characterization of the proteins of isolated human platelet alpha granules. Evidence for a separate alpha-granule pool of the glycoprotein IIb and IIIa. Biochim Biophys Acta 1981;670:150 – 62. [14] Woods VL, Wolff LE, Keller DM. Resting platelets contain a substantial centrally located pool of glycoprotein IIb – IIIa complex which may be accessible to some but not other extracellular proteins. J Biol Chem 1986;261:15242 – 51. [15] Plow EF, Shattil SJ. Integrin IIb/IIIa and platelet aggregation. In: Colman RW, Hirsh J, Marder VJ, Clowes AW, George JN, editors. Hemostasis and Thrombosis. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 479 – 92. [16] Shattil SJ, Kashiwagi H, Pampori N. Integrin signaling: the platelet paradigm. Blood 1998;91:2645 – 57. [17] Clark EA, Brugge JS. Tyrosine phosphorylation. In: Authi KS, Watson SP, editors. Platelets. Oxford: IRL Press; 1995. p. 199 – 216. [18] Abrams CS, Brass LF. Platelet signal transduction. In: Colman RW, Hirsh J, Marder VJ, Clowes AW, George JN, editors. Hemostasis and Thrombosis. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 540 – 1. [19] Cohen I, Burk DL, White JG. The effect of peptides and monoclonal antibodies that bind to platelet glycoprotein IIb – IIIa complex on the development of clot tension. Blood 1989;91:2645 – 57. [20] Carr ME, Carr SL, Hantgan RR, Braaten J. Glycoprotein IIb/IIIa blockade inhibits platelet-mediated force development and reduces gel elastic modulus. Thromb Haemost 1995;82:468 – 80. [21] Jennings LLK, White MM, Mandrell TD. Interspecies comparison of platelet aggregation, LIB expression and clot retraction: observed differences in GPIIb – IIIa functional activity. Thromb Haemost 1995; 74:1551 – 6. [22] Mousa SA, Khurana S, Forsythe MS. Comparative in vitro efficacy of different platelet glycoprotein IIb/IIIa antagonists on platelet-mediated
[23] [24]
[25]
[26]
[27]
[28]
[29]
[30]
[31] [32]
189
clot strength induced by tissue factor with use of thromboelastography. Arterioscler Thromb Vasc Biol 2000;20:1162 – 7. Cohen K, Gerrard JM, White JG. Ultrastructure of clots during isometric contraction. J Cell Biol 1982;93:775 – 83. Osdoit S, Rosa J-P. Fibrin clot retraction by human platelets correlates with GP IIb/IIIa integrin-dependent protein tyrosine dephosphorylation. J Biol Chem 2001;276:6703 – 10. Olson RE, Sielecki TM, Wityak J, Pinto DJ, Batt DG, Frietze WE, et al. Orally active isoxazoline glycoprotein IIb/IIIa antagonists with extended duration of action. J Med Chem 1999;42:1178 – 92. Mousa SA, Bozarth J, Naik U, Slee A. Platelet GP IIb/IIIa binding characteristics of small molecule RGD mimetic: distinct binding profile for Roxifiban. Br J Pharmacol 2001;133:331 – 6. Mousa SA, Bozarth J, Lorelli W, Forsythe M, Thoolen M, Slee A, et al. Antiplatelet efficacy of XV459, a novel nonpeptide platelet GP IIb/IIIa antagonist: comparative platelet binding profiles with c7E3. J Pharmacol Exp Ther 1988;286:1277 – 84. Mousa SA, Bozarth J, Lorelli W, Forsythe M, Thoolen M, Slee A, et al. Discovery of novel orally active non-peptide platelet GP IIb/ IIIa receptor antagonists, DMP754: comparative platelet binding profile with c7E3. J Pharmacol Exp Ther 1998;286:1277 – 84. Marguerie GA, Edgington TS, Plow EF. Interaction of fibrinogen with its platelet receptor is part of a multistep reaction in ADP-induced platelet aggregation. J Biol Chem 1980;255:154 – 61. Marguerie GA, Plow EF. Interaction of fibrinogen with its platelet receptor: kinetics and effect of pH and temperature. Biochemistry 1981;20:1074 – 80. Peerschke EIB, Wainer JA. Examination of irreversible platelet – fibrinogen interactions. Am J Physiol 1985;248:C466 – 72. Hantgan RR, Mousa SA. Inhibition of platelet-mediated clot retraction by integrin antagonists. Thromb Res 1998;89:271 – 9.