Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Platelet activation occurs during coronary thrombolysis both in experimental animal models and in humans (Fitzgerald et al., 1988, 1989, 1991; Kerins et al., 1989). The increase in platelet activity delays reperfusion and induces acute reocclusion in animal models. Furthermore, antiplatelet therapy reduces the mortality in patients receiving streptokinase, although whether improved reperfusion is the cause is still unclear (ISIS-2, 1988). Several different agonists mediate the increased platelet activity, including thrombin, TXA₂ and serotonin (Fitzgerald and FitzGerald, 1989; Fitzgerald et al., 1989; Golino et al., 1988). Consequently, specific inhibitors used alone may be inadequate in preventing reocclusion. An alternative approach is antagonism of the platelet GP IIb/IIIa, the surface receptor for fibrinogen (Marguerie et al., 1979). GP IIb/IIIa is one of a series of integrins, adhesion receptors that are heterodimers of an alpha (GP IIb) and a beta (GP IIIa) subunit. Under resting conditions, this receptor has a low affinity for fibrinogen. On activation of the platelet, the receptor undergoes a conformational change and expresses a high affinity for the ligand (Marguerie et al., 1979). Activation of the platelet GP IIb/IIIa with subsequent platelet aggregation is a common response to platelet agonists. Studies with GP IIb/IIIa antagonists have demonstrated a role for this receptor in mediating much of the functional response to platelet activation during coronary thrombolysis (Gold et al., 1988; Mickelson et al., 1990; Coller et al., 1991).

The platelet GP IIb/IIIa is not a passive receptor but transduces signals into the cell during fibrinogen binding (Shattil et al., 1994). These responses include activation of phosphokinases, generation of inositol phosphates and late calcium transients and can be replicated by a peptide derived from the cytoplasmic tail of GP IIb (Stephens et al., 1998). Fibrinogen binding also provokes a conformational change in the receptor that is detected as the expression of novel epitopes, LIBS (Frelinger et al., 1991). Antagonists of the platelet GP IIb/IIIa, which are largely designed to mimic the binding region of fibrinogen, exhibit some of these effects. Several antagonists of the platelet GP IIb/IIIa induce the expression of LIBS, as seen with fibrinogen. Furthermore, the antagonists that induce LIBS provoke an "activated" state in the receptor so that it expresses a high-affinity binding site for fibrinogen (Kouns et al., 1992). Although no specific evidence of "outside-in" signaling by these compounds exists, peptide antagonists have been reported to enhance clot retraction (Cohen et al., 1989), a response subsequent to ligand binding that is absent in patients lacking the platelet GP IIb/IIIa (George et al., 1990). Thus, some compounds may behave as partial agonists. In this study, we examined the functional relevance of LIBS expression in determining the response to GP IIb/IIIa antagonists in a canine model of coronary thrombolysis. Specifically, we addressed whether such changes were associated with partial agonist activity.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Platelet GP IIb/IIIa antagonists, Ro43-5054 and Ro44-9883, were kind gifts from Dr. Sebastien Roux (Hoffman La Roche, Basel, Switzerland). The thrombolytic agent t-PA was a gift from Dr. Stuart Bunting (Genentech, San Francisco, CA). D3GP3, a monoclonal antibody to GP IIIa, was a kind gift from Dr. Lisa Jennings (University of Tennessee, Memphis, TN) (Kouns et al., 1990). FITC-labeled goat anti-mouse antibody was obtained from Becton Dickenson (Oxford, UK), ADP was obtained from Sigma (St Louis, MO) and U46619 was purchased from Cayman Chemical (Ann Arbor, MI).

In Vivo Studies

The canine model of electrically induced thrombosis. Mongrel dogs (17-30 kg) were studied with the previously described protocol (Fitzgerald et al., 1988). After administering anesthesia with pentobarbitone (2.5 mg/kg), the animal was intubated and ventilated with a Harvard respirator. A needle electrode was inserted into the lumen of the left circumflex coronary artery distal to a Doppler flow probe (Crystal Biotech, Holliston, MA). The chest was closed and the animal allowed to recover. During the first 48 h the animal was treated with heparin (10000, U), analgesics (temgesic, 1.2 mg) and broad spectrum antibiotics (penicillin, 2000 mg) per day.

Bleeding rate. The bleeding rate was assessed as the rate of blood loss from a standardized skin incision on the chest wall measured by packing and weighing surgical gauze every 30 min.

Platelet aggregation. Ex vivo platelet aggregation in response to ADP (10 µM) alone and combined with U46619 (10 µM) was determined by light transmission with a four-channel platelet aggregometer (Bio Data, Model PAP-4, Horsham, PA). Blood was collected into 0.38% sodium citrate. Platelet-rich plasma was prepared by centrifugation at 1000 rpm for 15 min and platelet-poor plasma was prepared by centrifugation at 3000 rpm for 10 min. Agonists were added in volumes of 50 µl or less to 500-µl aliquots of platelet-rich plasma.

Measurement of 2,3-dinor-TXB₂. TXA₂ biosynthesis was used as a marker of platelet activation in vivo. The major metabolite of TXA₂ in the canine is 2,3-dinor-TXB₂. Urine was collected at 30 min-intervals after occlusion and hourly after reperfusion to determine the excretion of 2,3-dinor-TXB₂ by negative-ion, chemical ionization-gas chromatography-mass spectrometry (Nowak et al., 1987).

Plasma t-PA and drug measurements. To refute any pharmacokinetic interaction between the experimental drugs and t-PA, plasma t-PA was determined by ELISA (American Diagnostics, Inc., Greenwich, CT) (Bergsdorf et al., 1983). A separate standard curve was constructed with plasma from each animal.

In Vitro Studies

LIBS expression. Expression of LIBS was determined as the binding of the monoclonal antibody, D3GP3. Platelet-rich plasma was exposed to t-PA (600 ng/ml), Ro43-5054 (1 µM), Ro44-9883 (1 µM) or EDTA for 30 min at 37°C. Treatment with EDTA results in dissociation of the receptor and maximum expression of LIBS. Samples were then fixed in an equal volume of 2% formaldehyde/PBS/0.1% BSA (FPB) for 2 h. The samples were then centrifuged at 11,000 rpm for 3 min and washed in PBS/0.1% BSA. LIBS expression was determined with the monoclonal antibody D3GP3 at a final concentration of 4 µg/ml. The samples were incubated with the antibody at room temperature for 30 min, centrifuged and washed as above. Four microliters of FITC-labeled goat anti-mouse antibody was added to the sample and incubated for 30 min at room temperature. The samples were then centrifuged and washed as above and resuspended in 1 ml of 1% FPB. A Becton Dickenson FACScan (Oxford, England) was used to analyze the samples for LIBS expression. Additionally, in three animals used for each drug we studied, LIBS expression was measured ex vivo during infusion of 5 µg/kg/min with the same procedure but without the addition of drug.

Statistical analysis. The data were analyzed by Kruskal-Wallis one-way analysis of variance with a subsequent Mann Whitney test for comparison between groups. This analysis is nonparametric and therefore makes no assumption about the distribution of the data, which is important given the sample size. Values are expressed as mean ± S.E.M. An exact Fisher test was used to compare event rates between groups.

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

In Vitro Studies

LIBS expression. To examine the effect of t-PA, Ro43-5054 and Ro44-9883 on LIBS induction, the binding of D3GP3 to its epitope on platelet GP IIb/IIIa was determined. Ro43-5054 at 1 µM (within the range of plasma concentration achieved in vivo) induced a significant increase in D3GP3 binding compared with control (P < .001) as determined by FACS analysis (fig 1). This increase in D3GP3 binding was shown to be dose-dependent in both human (EC₅₀, 100 nM) and canine (EC₅₀, 200 nM) platelets. Ro44-9883 (1 µM) and t-PA (600 ng/ml) had no effect on LIBS expression. Similarly, ex vivo determination of LIBS expression in canines during infusion of drug at 5 µg/kg/min (n = 3) showed that Ro43-5054 induced D3GP3 binding compared with control (fig. 2), whereas Ro44-9883 and t-PA had no effect.

View larger version (48K): [in this window] [in a new window]   Fig. 1.   The effect of Ro43-5054 and Ro44-9883 on LIBS exposure with flow cytometric analysis. Cells were treated with either Ro43-5054 or Ro44-9883 (1 µM). After treatment, the platelets were treated with the monoclonal antibody D3GP3 (4 µg/ml) and binding was detected with FITC-labeled goat anti-mouse antibody. Data are means ± S.E.M. of n = 5 per group.

View larger version (14K): [in this window] [in a new window]   Fig. 2.   The effect of Ro43-5054 or Ro44-9883 (5 µg/kg/min) on LIBS exposure with flow cytometric analysis (n = 3). EDTA dissociates the receptor and induces maximum expression of D3GP3 binding detected as a shift in mean fluorescence. Treatment with Ro43-5054 induces expression of D3GP3 similar to EDTA. In contrast, binding after treatment with Ro44-9883 is unchanged compared with control samples (FITC-labeled anti-mouse antibody alone or after exposure to D3GP3).

In Vivo Studies

Platelet aggregation. To examine the in vivo activity of Ro43-5054 and Ro44-9883, ex vivo platelet aggregation was performed. Platelet aggregation to ADP alone and combined with U46619, the TXA₂ analog, was abolished at 2 and 5 µg/kg/min of Ro43-5054 and Ro44-9883 (table 2).

Bleeding rate. Bleeding responses were negligible before administration of t-PA. After administration of t-PA, the rate of bleeding significantly increased with Ro43-5054 and Ro44-9883 (P < .01 for both Ro43-5054 and Ro44-9883 compared with controls). However, there were no differences between the drugs (fig. 3).

View larger version (48K): [in this window] [in a new window]   Fig. 3.   The effect of Ro43-5054 and Ro44-9883 on bleeding rate. Both drugs increased the bleeding rate significantly when compared with control (*P < .01 vs. vehicle). However there was no difference in bleeding rate between the drugs when all doses were combined. Data are means ± S.E.M of n = 10-14.

Reperfusion and reocclusion. All reperfusion and reocclusion rates and times are shown in table 3. In the Ro43-5054 experiments, three of four animals failed to reperfuse at the lowest dose of 2 µg/kg/min, in contrast to the controls, all of whom reperfused. At 5 µg/kg/min, reperfusion occurred in every case, but the time to reperfusion (49 ± 8 min) was unchanged compared with controls (52 ± 5 min). One of six animals reoccluded at 67 min, whereas the other five animals in this group sustained reperfusion. Increasing the dose to 10 µg/kg/min had no further effect with only two of four animals reperfusing. One animal demonstrated episodic reocclusion but complete reocclusion never occurred. In every case in which reperfusion failed to occur, examination of the artery showed extensive new clotting. When all three doses were combined, Ro43-5054 did not enhance reperfusion but significantly prevented reocclusion (22% vs. 100% in controls; P = .0001).

In vivo platelet activation. Urinary excretion of 2,3-dinor-TXB₂ was determined in controls and in animals treated with 5 µg/kg/min of GP IIb/IIIa antagonist. Analysis was confined to animals who reperfused, because previous experiments have shown an increase in 2,3-dinor-TXB₂ only upon reperfusion (Fitzgerald et al., 1989). In control experiments, urinary excretion of 2,3-dinor-TXB_2, the major metabolite of TXA_2, increased markedly upon reperfusion (fig. 4). Ro43-5054 had no effect on urinary 2,3-dinor-TXB₂ at reperfusion compared with controls (fig. 4). In contrast, Ro44-9883 suppressed the increase in urinary 2,3-dinor-TXB₂ (P < .05) (fig. 4).

View larger version (17K): [in this window] [in a new window]   Fig. 4.   The effect of GP IIb/IIIa antagonists, Ro43-5054 and Ro44-9883 (5 µg/kg/min), or vehicle on 2,3-dinor-TXB2 excretion before and after coronary thrombolysis induced by t-PA. Data are means ± S.E.M. of n = 4-7 per group, *P < .05 vs. vehicle.

Plasma drug levels. Plasma concentrations of both Ro43-5054 and Ro44-9883 increased in a dose-dependent manner and far exceeded the EC₅₀ for inhibition of platelet aggregation at all doses (fig. 5). To exclude a pharmacokinetic interaction between Ro43-5054 and Ro44-9883 and the thrombolytic agent, plasma t-PA levels were estimated at 30 min of t-PA and 2 h after reperfusion (1 h 50 min after discontinuing t-PA). Ro43-5054 and Ro44-9883 had no effect on t-PA levels compared with controls, either during the infusion of t-PA or during the washout phase (fig. 6).

View larger version (26K): [in this window] [in a new window]   Fig. 5.   Plasma concentrations of Ro43-5054 and Ro44-9883 after 2 h reperfusion. The plasma concentrations increased in a dose-dependent manner and well exceeded the IC50 for inhibition of platelet aggregation.

View larger version (22K): [in this window] [in a new window]   Fig. 6.   The effect of Ro44-9883 and Ro43-5054 on t-PA levels before, during and after infusion of t-PA. Data are means ± S.E.M. of n = 4 per group.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Despite their shared activitys in preventing platelet aggregation, antagonists of the platelet GP IIb/IIIa differ in some respects. Several, but not all, antagonists have been reported to induce conformational changes so that the receptor assumes an activated, ligand-bound state (Kouns et al., 1992). These changes include the appearance of neoepitopes within the active site similar to those detected after activation of platelets (Abrams et al., 1992). Furthermore, platelets that are exposed to the antagonist and washed in fixative to remove the ligand and fix the conformation of the receptor will spontaneously bind fibrinogen and aggregate (Kouns et al., 1992). Antagonists that induce a high-affinity binding site for fibrinogen also induce the expression of neoepitopes (called LIBS) in several other regions of both the alpha and beta subunits of the integrin (Steiner et al., 1993). The functional relevance of LIBS is unclear. However, antibodies interacting with LIBS have been found to provoke platelet aggregation or to interfere with clot retraction (Frelinger et al., 1991; Jennings et al., 1993). Such findings suggest that LIBS are functionally active sites involved in postreceptor occupancy or "outside-in" signaling. Thus, expression of LIBS may indicate partial agonist activity. However, evidence that antagonists can transduce "outside-in" signaling such as protein phosphorylation or Ca⁺⁺ transients is absent, although peptide antagonists of the platelet GP IIb/IIIa have been reported to enhance clot retraction (Cohen et al., 1989).

In this study, we compared the effects of two antagonists of the platelet GP IIb/IIIa, Ro43-5054 and Ro44-9883, in a canine model of coronary thrombolysis. Both compounds are potent antagonists of the platelet GP IIb/IIIa with nearly equal activity against the human receptor. Kouns (1992) demonstrated that Ro43-5054 induced the expression of a binding site for D3GP3, a LIBS antibody identified by Jennings (Kouns et al., 1990). The expression of an epitope for D3GP3 persisted and a high-affinity site for fibrinogen was exposed after the removal of the compound from purified receptor. Ro43-5054 also induced the expression of four other LIBS, two on the heavy chain of GP IIb and two that are complex specific (Steiner et al., 1993). In contrast, Ro44-9883 failed to induce LIBS in purified human receptors or in platelets even at a concentration of 10 µM. In this study, we confirmed that at concentrations achieved in vivo, Ro43-5054 but not Ro44-9883 induced expression of D3GP3 binding in canine platelets in vitro. Moreover, platelets obtained during the infusion of Ro43-5054 expressed LIBS, whereas increased D3GP3 binding was not observed during the infusion of Ro44-9883.

The model used in these experiments, the canine model of coronary thrombolysis, is associated with platelet activation after reperfusion detected as a marked rise in TXA₂ formation. The increase in platelet activity interferes with the response to the thrombolytic therapy by delaying reperfusion and inducing acute reocclusion (Fitzgerald et al., 1989). The platelet GP IIb/IIIa apparently plays a role as 7E3, an antagonist of human and canine GP IIb/IIIa, accelerates reperfusion and prevents reocclusion (Gold et al., 1988; Yasuda et al., 1988; Fitzgerald et al., 1989). Both Ro43-5054 and Ro44-9883 abolished platelet aggregation even at the lowest doses used. There was also a marked increase in bleeding from a standardized skin wound. These findings are consistent with the plasma drug levels achieved, which in vitro result in complete suppression of platelet aggregation.

The primary end-point of the experiment was the prevention of reocclusion. Once it was clear this was not possible at lower doses, we moved to the next highest, as previous studies have shown complete suppression of reocclusion with antithrombotics. Both compounds inhibited coronary reocclusion. Despite the in vivo evidence of marked antiplatelet effect and adequate plasma drug levels, neither compound accelerated reperfusion. Although there were too few experiments to address the issue, both compounds may have interfered with reperfusion in some experiments, especially at lower doses. This was not caused by a pharmacokinetic interaction with t-PA, because plasma t-PA levels were unaltered by either compound. However, the tendency to interfere with reperfusion is consistent with reports that GP IIb/IIIa antagonists increase clot retraction which impairs the ability of t-PA to induce clot lysis (Cohen et al., 1989; Kunitada et al., 1992).

The two compounds differed only in their effects on TXA₂ formation, a marker of platelet activation in vivo. As previously reported, a marked increase in the formation of TXA₂ coincided with reperfusion. Despite substantial suppression of platelet aggregation and inhibition of reocclusion, Ro43-5054 failed to alter the increase in TXA₂ formation that occurred with reperfusion. In contrast, Ro44-9883 markedly inhibited the urinary excretion of the metabolite. Ro44-9883 did not alter serum TXB₂ (data not shown) which demonstrated that it has no direct effect on the synthetic enzymes required for TXA₂ formation.

Integrins may play a role in the generation of prostaglandins by platelets in response to a primary agonist. The focal adhesion points formed with GP IIb/IIIa also bind tyrosine kinases such as pp60^src (Fox et al., 1993), FAK (Shattil et al., 1994) and Rap2b which has GTPase activity (Torti et al., 1994). This arrangement is highly reminiscent of traditional cell signaling mechanisms, so it is possible that occupancy of GP IIb/IIIa increases TXA₂ production. Shattil and colleagues (1994) have reported phosphorylation of FAK during "co-stimulation" of platelets by epinephrine and an antibody to GP IIb/IIIa that forces fibrinogen to bind to the receptor. In the absence of ligand binding, FAK phosphorylation and activation did not occur. This response was TXA₂ dependent. Therefore, occupancy of the platelet GP IIb/IIIa may be required for TXA₂ formation in response to weak agonists such as epinephrine. It is possible that Ro43-5054, which induced LIBS, provided the necessary signaling for TXA₂ formation in response to agonists in vivo. In contrast, Ro44-9883, which does not induce LIBS, failed to provide the conformational change or required signaling. The exact mechanism, however, is unclear because it was not possible to replicate the differential effects on TXA₂ formation in vitro.

In conclusion, this study shows that LIBS expression did not influence the functional response to GP IIb/IIIa antagonists in a model of coronary thrombolysis. However, a GP IIb/IIIa antagonist that failed to induce LIBS resulted in greater suppression of TXA₂ formation in vivo. The mechanism for this is unknown, but the findings suggest that LIBS expression contributes to postoccupancy signaling transmitted through GP IIb/IIIa. Consequently, antagonists of GP IIb/IIIa that induce LIBS may provide a signal which results in TXA₂ formation in vivo. Because TXA₂ has additional effects, other than platelet aggregation, including platelet activation, vascular smooth muscle contraction and mitogenesis, the findings may be relevant to the design of GP IIb/IIIa antagonists intended for clinical use.