Author + information
- Received January 13, 2000
- Revision received April 26, 2000
- Accepted June 26, 2000
- Published online November 1, 2000.
- Dermot Cox, BSc, PhD∗,* (, )
- Richard Smith, BSc, PhD∗,
- Martin Quinn, MD∗,
- Pierre Theroux, MD†,
- Peter Crean, MD‡ and
- Desmond J Fitzgerald, MD∗
- ↵*Reprint requests and correspondence: Dr. Dermot Cox, Department of Clinical Pharmacology, Royal College of Surgeons, 123 St. Stephen’s Green, Dublin 2, Ireland.
The study was done to determine the role of partial agonist activity in the lack of effectiveness of the oral GPIIb/IIIa antagonist orbofiban.
Orbofiban, an oral GPIIb/IIIa antagonist, was found to increase the mortality of patients with acute coronary syndromes (ACS) in the OPUS-TIMI-16 trial, despite the fact that it is a very potent anti-platelet agent and that IV agents have proven very effective.
Patients (n = 520) with ACS were randomized to orbofiban 30 mg, 40 mg or 50 mg twice daily or 50 mg once daily or placebo. Platelet activity was assessed in 175 patients by examining GPIIb/IIIa receptor conformation, expression of CD63 antigen, and platelet aggregation.
Plasma concentrations of orbofiban at the highest dose (74 ± 6 ng/ml peak, 61 ± 5 ng/ml trough) exceeded the IC50 for platelet aggregation to adenosine diphosphate (ADP) (29 ± 6 ng/ml) and thrombin-activating peptide (61 ± 18 ng/ml). Orbofiban induced a conformational change in GPIIb/IIIa detected as the displacement of the monoclonal antibody mAb2; such conformational changes have been linked to partial agonist activity. Consistent with this, platelet expression of CD63 ex vivo was significantly increased at five time points during the study. In vitro, orbofiban increased platelet aggregation to a submaximal concentration of epinephrine (67 ± 19% vs. 27 ± 9%, n = 5) and increased thromboxane formation when the platelet GPIIb/IIIa were clustered using monoclonal antibodies to the receptor.
Orbofiban is both an antagonist and a partial agonist of platelet GPIIb/IIIa. At low concentrations of the drug, this partial agonist activity may enhance platelet aggregation. Along with suboptimal plasma drug levels, these findings may help explain the lack of efficacy seen with orbofiban in patients with ACS.
Glycoprotein (GP) IIb/IIIa (platelet fibrinogen receptor) antagonists are a novel class of antithrombotics that block fibrinogen binding to platelets and thereby prevent platelet aggregation (1,2). These compounds have proved effective in patients with unstable angina or myocardial infarction (MI) and following coronary angioplasty when administered as short-term infusions (3). Results of these trials are consistent with evidence of a role for platelets in unstable coronary syndromes (4–7). However, GPIIb/IIIa antagonists have proved ineffective when administered chronically as oral compounds in patients presenting with acute coronary syndromes or following coronary angioplasty. Indeed, in the OPUS-TIMI-16 Phase III trial of the GPIIb/IIIa antagonist orbofiban, the mortality was increased in patients on active treatment compared with those on placebo (8).
One factor that may limit the effectiveness of GPIIb/IIIa antagonists is their propensity to act as partial agonists. GPIIb/IIIa is one of a family of receptors, the integrins, that act as receptors for adhesive proteins (1). Upon ligand binding, integrins transmit signals into cells, which in the case of GPIIb/IIIa amplify platelet activation (9–12). Although this is most evident with large ligands such as fibrinogen, small ligands designed as integrin antagonists trigger similar signals in vitro (13,14), particularly if they induce extensive conformational changes in the receptor (15,16). We examined the effects of orbofiban on platelet function when administered for 84 days in patients presenting initially with unstable angina and MI. In addition, we looked for evidence of partial agonist activity in vitro.
The study was established as a Phase II dose-finding study of orbofiban. The trial was approved by the Irish Medicines Board, by the Drugs Directorate, Health Canada, and the institutional review bodies of the participating hospitals. All patients gave written, informed consent. Patients were considered eligible for the study if they had unstable angina (<120 h from last chest pain) or a recent MI (<120 h). Patients were randomized to placebo (n = 101) twice daily, orbofiban 30 mg twice daily (n = 109), orbofiban 40 mg twice daily (n = 104), orbofiban 50 mg twice daily (n = 104) or orbofiban 50 mg once daily with a matching placebo in place of the second dose each day (n = 102). Treatment was continued for 84 days. Detailed platelet studies were performed in 175 subjects: placebo (n = 36); orbofiban 30 mg twice daily (n = 34); orbofiban 40 mg twice daily (n = 34); orbofiban 50 mg twice daily (n = 37); and orbofiban 50 mg once daily (n = 34). These patients were all the patients enrolled in all participating Irish and Canadian sites. In addition to the study medication, patients received 162 mg aspirin daily. Samples were taken pre-dose and 4 h and 6 h post-dose on days 1, 28, and 84 (last day) of treatment.
Citrated (3.8%) whole blood was centrifuged at 850 × g for 3 min to produce platelet-rich plasma (PRP). The remaining blood was centrifuged for 2 min at 2500 × g to prepare platelet-poor plasma. After 1 min pre-incubation at 37°C, platelet aggregation was induced by 20 μmol/liter adenosine diphosphate (ADP) or 5 μmol/liter thrombin receptor activating peptide (TRAP). The extent of platelet aggregation was determined after 2 min by light transmittance (PAP-4 platelet aggregometer, BioData, Horsham, Pennsylvania).
GPIIb/IIIa receptor number and conformation
The GPIIb/IIIa receptor number and conformation were analyzed by quantifying the binding of two monoclonal antibodies to GPIIb/IIIa, mAb1 and mAb2 (BioCytex, Marseille, France), that recognize distinct epitopes on the GPIIb/IIIa receptor (17). The total number of GPIIb/IIIa receptors can be determined by mAb1, whereas mAb2 recognizes only the unbound GPIIb/IIIa. In the presence of GPIIb/IIIa antagonists, the mAb2 epitope on the receptor disappears as a result of a ligand-induced change in the conformation of GPIIb/IIIa. PRP (20 μl) was incubated with 45 μg/ml of mAb1 or mAb2 for 20 min at room temperature. Samples and a tube containing calibration beads were stained with 20 μl of fluorescein isothiocyanate-labeled anti-mouse antibody for 10 min and transferred to gel tubes containing 1 ml of 0.1% paraformaldehyde for 10 min. These gel tubes consisted of a small tube with a V-shaped bottom containing a small amount of Sephadex gel. The gel tubes were centrifuged at 850 × g for 10 min, which allowed the platelets to penetrate the gel. The supernatant was poured off, and the gel was washed twice in phosphate buffered saline (PBS). The tubes were filled with PBS and stored at 4°C until analyzed (within two weeks). These gel tubes enable the platelets to be washed and also provide a suitable storage medium for them. Prior to analyses, the tubes were centrifuged at 850 × g for 10 min, and the supernatant was poured off. The gel-platelet mix was then re-suspended in the buffer and analyzed by flow cytometry as described below.
Marker of platelet activation
Platelet activation was monitored by the expression of the protein CD63 antigen, which is released to the platelet surface after activation (18). PRP (20 μl) was incubated with CD63 (BioCytex, Marseille, France) for 20 min at room temperature. A similar tube was also prepared with the addition of TRAP (5 μmol/liter) 10 min prior to the addition of CD63, which was used to gate for CD63-positive platelets. Samples were then stained with a secondary antibody and analyzed by flow cytometry.
Quantitative fluorescence cytometry
Samples were collected using log forward scatter (FSC), side scatter (SSC) and fluorescence. A region was drawn around the peak density of platelets in the FSC versus SSC plot, and this was used as a gate for analysis of the platelet population. A histogram marker was drawn around the TRAP-activated CD63-stained platelets so as to include 95% of all positive platelets. This was defined as the positive population. Events from the sample population falling in this region were considered positive. Calibration beads (BioCytex, Marseille, France) were also analyzed, and a log plot of mean fluorescence intensity versus molecules of antibody was prepared. Flow cytometry was performed on a Becton Dickinson FACSCalibur flow cytometer and analyzed using CellQuest software (B & D, Oxford, United Kingdom).
The plasma concentration of orbofiban was measured by high-performance liquid chromatography/mass spectrometry.
In vitro studies
Platelet aggregation was induced with epinephrine, and a concentration that induced <35% platelet aggregation was identified. Platelet aggregation was performed in the absence and presence of orbofiban at a concentration (60 nmol/liter, 20 ng/ml) that alone did not inhibit ADP-induced aggregation (IC50 130 ± 10 nmol/liter, 43 ± 3 ng/ml).
Signaling through integrins is accommodated by prior clustering of the receptor (12,19). Platelet GPIIb/IIIa receptors were clustered using antibody-coated magnetic beads (20). Briefly, gel-filtered platelets (5 × 106 platelets) were incubated with 2 × 107 beads (Dynal MPC-E, Oslo, Norway) coated with CD41, a monoclonal antibody to GPIIIa, in the presence of orbofiban, abciximab or vehicle for 45 min at room temperature. They were then magnetized for 2 min, snap-frozen, and stored until the thromboxane B2 levels were measured by ELISA (R&D Systems, Oxon, United Kingdom). Alternatively, platelets were incubated with 4F8 (BioCytex, Marseille, France) without beads, followed by orbofiban, snap-frozen, and thromboxane B2 levels measured. Note that CD41 and 4F8 are directed against the GPIIIa subunit, and as they are not directed at the ligand-binding site, they do not interfere with the binding of orbofiban or abciximab.
The data were analyzed by two-way analysis of variance (ANOVA) of the log transformed data with a Bonferroni-Dunn post hoc analysis.
In vitro studies
Samples were compared with the control values using a paired t test.
Orbofiban was a gift from Dr. Bob Anders, GD Searle, Skokie, Illinois. The TRAP was from Dr. Pat Harriott, Queens University, Belfast, UK. All monoclonal antibodies and gel tubes were from BioCytex, Marseille, France.
The patient characteristics are shown in Table 1. In the patients with Q wave MI, 58% had their most recent MI in the inferior wall, 11% anterior, 8% anterolateral, 6% posterior, 14% were in another location, and 3% were unknown. There were no differences in baseline characteristics between the treatment groups. The numbers of patients completing the study were 76 of 101 on placebo, 73 of 109 on orbofiban 30 mg bid, 70 of 104 on 40 mg bid, 67 of 104 on 50 mg bid, and 73 of 102 on orbofiban 50 mg qd. The clinical outcomes of the study are shown in Table 2. Again, there were no statistical differences in the outcomes of the different groups, although MI occurred only in patients receiving orbofiban 50 mg twice daily or 50 mg daily.
Platelet aggregation and plasma drug levels
Platelet aggregation was inhibited at all doses of orbofiban (Fig. 1), although there was poor inhibition at trough (before the next dose) when the drug was given once daily. The platelet aggregation data were fitted to a hyperbolic curve and related to the plasma drug levels. This showed a dose-response relationship with an IC50 value for inhibition of ADP-induced aggregation of 29 ± 6 ng/ml with a correlation coefficient of 0.84. For TRAP-induced platelet aggregation, the IC50 value was 61 ± 18 ng/ml with a correlation coefficient of 0.75. It is worth noting that only at the highest dose was the IC50 for these agonists exceeded throughout the dosing interval (peak 74 ± 6 ng/ml, trough 61 ± 5 ng/ml).
GPIIb/IIIa receptor number and conformation
Binding of mAb1, which quantifies receptor number, was unchanged on any dose of orbofiban throughout the study (Table 3). However, binding of mAb2 was reduced at all doses of orbofiban, indicating that the drug altered the conformation of the receptor. Thus, the ratio of mAb2 to mAb1 binding fell to about 20% in patients receiving orbofiban 50 mg twice daily (Fig. 2).
A significant increase in CD63 antigen expression occurred in patients receiving active drug throughout the dosing period (Table 4), and this was dose dependent (two-way ANOVA, p = 0.015). Patients on the placebo showed no change in CD63 antigen expression. This dose-dependent increase in activation was mainly accounted for by a significant increase in CD63-positive cells at the lowest dose (30 mg bid, p = 0.003) of orbofiban by Bonferroni-Dunn post hoc test.
In vitro studies
A low level of epinephrine that induced minimal platelet aggregation (27 ± 9%, n = 5) was identified. Orbofiban at a concentration (25 ng/ml) that resulted in incomplete occupancy of platelet GPIIb/IIIa, and did not prevent platelet aggregation to ADP 20 μmol/liter, enhanced platelet aggregation to epinephrine (67 ± 16%; p = 0.04, n = 5).
Clustering of GPIIb/IIIa receptors with CD41 in the absence of receptor occupancy induced minimal thromboxane production (1.6 ± 2.9 ng/ml, n = 6). The addition of orbofiban at a concentration that will saturate the receptors (3.3 μg/ml) resulted in an increase in thromboxane to 6.6 ± 3.6 ng/ml (n = 6, p = 0.03). By contrast, abciximab, a monoclonal antibody that blocks the ligand binding site, had no significant effect (2.7 ± 0.8 ng/ml, n = 3, p > 0.05). More marked changes were seen using 4F8 to cluster the GPIIb/IIIa (Fig. 3), and the increase in thromboxane induced by orbofiban was blocked by abciximab (p < 0.05), suggesting that orbofiban must bind to GPIIb/IIIa to induce platelet activation. Neither the addition of abciximab to 4F8-bound platelets nor the addition of orbofiban without 4F8 increased thromboxane generation (data not shown).
Clinical effectiveness of orbofiban
GPIIb/IIIa is an attractive therapeutic target as antagonists of the receptor block platelet aggregation to all known agonists, in contrast to the limited activity of aspirin and clopidogrel. Administered as intravenous (IV) agents, GPIIb/IIIa antagonists have proved effective in a variety of acute coronary syndromes (3). However, oral agents have proved less effective, and in the case of orbofiban, mortality has increased in patients on active treatment (8). As shown in this study, orbofiban suppressed platelet aggregation to ADP and to TRAP, and this was correlated strongly with plasma drug levels. However, the degree of inhibition of platelet aggregation and receptor occupancy was far less than the 80% achieved with IV agents.
We also addressed the possibility that orbofiban acted as a partial agonist in vivo and increased platelet activity. Binding of fibrinogen to GPIIb/IIIa triggers several intracellular events, so-called outside-in signaling. These include an increase in intracellular Ca2+(13,21), tyrosine phosphorylation (11) and thromboxane A2 formation (10,21). Outside-in signaling is dependent on clustering of the receptor (22) (feasible only with ligands with multiple binding sites such as fibrinogen) and linkage of the cytoplasmic tail to intracellular proteins (10). Ligand binding appears to be an additional requirement. Thus, small fragments of laminin can induce PGE2 formation if the laminin receptor is first clustered (19).
Ligand-induced conformational changes
How ligand binding induces outside-in signaling is unknown, but it may involve a change in receptor conformation. Many GPIIb/IIIa antagonists induce conformational changes that are detected as the appearance of new epitopes (23,24) or, as in the case of mAb2, the disappearance of an epitope (17). We show here that orbofiban reduced the binding of mAb2, although it had no effect on the total number of receptors, measured as the binding of mAb1. The antibody mAb2 recognizes a site remote from the ligand binding site. This site disappears upon binding of small ligands, although mAb2 does not compete directly with the ligand (17). The findings suggest that mAb2 detects a region that becomes obscured as a consequence of a ligand-induced change in conformation. Thus, the disappearance of mAb2 binding in patients on orbofiban is evidence that the platelet GPIIb/IIIa has undergone a change in conformation. Ligand-induced changes in the conformation of GPIIb/IIIa have been implicated in the partial agonist activity of GPIIb/IIIa antagonists. For example, Honda and colleagues (13) found that only compounds inducing conformational changes in the GPIIIa subunit enhanced the rise in platelet calcium and aggregation response to platelet agonists.
We also looked for evidence of increased platelet activation ex vivo. Detecting subtle changes in platelet activation ex vivo is difficult. Possibly the best marker is thromboxane A2 formation (4), which was not possible in this study, because all of the patients were on aspirin. As an alternative we examined the expression of CD63 antigen, which is a protein present in the lysosomes that appears on the surface upon platelet activation (18). Unlike CD62 antigen, CD63 antigen is highly stable once expressed (25). Expression of the CD63 antigen was significantly increased throughout the treatment period in patients receiving orbofiban.
Orbofiban as a partial agonist
Together with the evidence of a change in conformation, the data suggest that orbofiban enhanced platelet activity in vivo. This hypothesis was supported by in vitro experiments demonstrating that orbofiban acted as a partial agonist. Thus, orbofiban enhanced the platelet response to epinephrine. Low concentrations of orbofiban insufficient to block all of the platelet GPIIb/IIIa receptors were used to avoid masking an effect on platelet aggregation. In a second model, the receptor was first clustered using monoclonal antibodies to mimic one of the requirements for signaling via natural ligands. The addition of the antagonist to clustered receptors resulted in an increase in thromboxane, a marker of platelet activation. It may be argued that this type of clustering does not mimic what happens in vivo. Nevertheless, the experiment demonstrates the potential of orbofiban to induce strong signaling. The increase in thromboxane in the presence of 4F8 was prevented by the monoclonal antibody abciximab, which binds to the ligand recognition site of GPIIb/IIIa and prevents orbofiban binding (17). This finding demonstrates that orbofiban-induced platelet activation is receptor mediated. However, a recent report found that orbofiban can induce apoptosis in rat cardiomyocytes (26), although it is not known if this can occur in human platelets.
Whether the increased platelet activity detected in patients on orbofiban is clinically relevant and contributed to the poor outcome of OPUS-TIMI-16 is unclear. The ability of orbofiban to enhance platelet aggregation may be important, particularly during the trough period prior to dosing, when there is insufficient drug to block platelet aggregation effectively. Platelet activation is enhanced and plays a major role in the pathogenesis of acute coronary syndromes, including unstable angina and MI (4). Platelet microemboli have also been seen in coronary microvessels of patients suffering sudden cardiac death (7). Increased platelet activation has also been reported in a phase II study of sibrafiban, another oral GPIIb/IIIa antagonist (27), and a recent report has confirmed the increase in platelet activation in the phase III orbofiban study (28). It is possible that in the setting of low-level platelet activation seen in coronary artery disease, additional activation through the GPIIb/IIIa receptor may provoke rather than prevent thrombosis, especially when drug levels are at trough.
Although there is evidence that abciximab (14,29) and eptifibatide (30), both of which are clinically effective, also activate platelets in vitro, these drugs are administered intravenously and at doses that achieve high levels of receptor occupancy. Moreover, they are combined with anticoagulants, which have been shown to enhance the response to a GPIIb/IIIa antagonist in an experimental model of coronary thrombosis (31). Thus, partial agonism may not be a problem with IV compounds, because they maintain stable plasma levels with a high level of blockade, are supplemented with anticoagulants, and are administered for only a short period of time.
Conversely, partial agonism combined with either low dosing or a low trough can lead to an increase in events in patients treated with an oral GPIIb/IIIa antagonist. The longer duration of treatment also increases the likelihood of a patient’s forgetting to take medication, thereby increasing the chance of a low trough.
In conclusion, the relatively low plasma drug levels and suboptimal inhibition of platelet aggregation along with a partial agonist effect may have contributed to the poor clinical efficacy of orbofiban.
☆ This work was funded by Searle and Company, Skokie, Illinois, and supported by grants from the Health Research Board and the Higher Education Authority of Ireland.
- adenosine diphosphate
- analysis of variance
- forward scatter
- platelet fibrinogen receptor
- myocardial infarction
- phosphate buffered saline
- platelet-rich plasma
- side scatter
- thrombin receptor activating peptide
- Received January 13, 2000.
- Revision received April 26, 2000.
- Accepted June 26, 2000.
- American College of Cardiology
- Kong D,
- Califf R,
- Miller D,
- et al.
- Fitzgerald D,
- Catella F,
- Roy L,
- Fitzgerald G
- Falk E
- Davies M,
- Thomas A,
- Knapman P,
- Hangartner J
- Cannon C.P,
- McCabe C.H,
- Wilcox R.G,
- et al.
- Stephens G,
- O’Luanaigh N,
- Reilly D,
- et al.
- Honda S,
- Tomiyama Y,
- Aoki T,
- et al.
- Peter K,
- Schwartz M,
- Ylanne J,
- et al.
- Du X,
- Plow E,
- Frelinger A,
- O’Toole T,
- Loftus J,
- Ginsberg M
- Quinn M,
- Deering A,
- Stewart M,
- Cox D,
- Foley B,
- Fitzgerald D
- Corcoran M,
- Kibbey M,
- Kleinman H,
- Wahl L
- Schmidt C,
- Pommerenke H,
- Frieda D,
- Nebe B,
- Rychly J
- Hato T,
- Pampori N,
- Shattil S
- Frelinger A,
- Cohen I,
- Plow E.F,
- et al.
- Frelinger A,
- Du X.P,
- Plow E.F,
- Ginsberg M.H
- Adderley S,
- Fitzgerald D
- Ault K,
- Cannon C,
- Mitchell J,
- et al.
- ↵Holmes M, Sobel B, Cannon C, Schneider D. Increased platelet reactivity in patients given orbofiban after an acute coronary syndrome: an OPUS-TIMI-16 substudy. Orbofiban in Patients with Unstable coronary syndromes. Thrombolysis In Myocardial Infarction. Am J Cardiol 2000;85:491–3, A10.
- Pratico D,
- Murphy N,
- Fitzgerald D