Enhanced Thrombolytic and Antithrombotic Potency of a Fibrin-Targeted Plasminogen Activator in Baboons
Background Thrombolytic therapy reduces mortality in patients with acute myocardial infarction, but significant limitations exist with the use of currently available agents. In the present report, we describe the thrombolytic and antithrombotic potencies of a hybrid recombinant plasminogen activator consisting of an antifibrin antibody 59D8 (AFA) and low-molecular-weight single-chain urokinase-type plasminogen activator (scuPA).
Methods and Results A thrombolysis model in which thrombi are preformed in vivo in juvenile baboons was developed to compare the potencies of AFA-scuPA, recombinant tissue plasminogen activator (rTPA), and recombinant scuPA (rscuPA) in lysing nonocclusive 111In-labeled platelet-rich arterial-type thrombi and 125I-labeled fibrin-rich venous-type thrombi. Systemic infusion of 1.89 nmol/kg AFA-scuPA produced thrombolysis that was comparable to that obtained with much higher doses of TPA (14.2 nmol/kg) and rscuPA (28.5 nmol/kg). When steady-state plasma concentrations are normalized, AFA-scuPA lyses thrombi sixfold more rapidly than scuPA and TPA (P<.001) and reduces the rate of formation more than comparable doses of rscuPA (P<.0001). At equivalent thrombolytic doses, AFA-scuPA produced fewer antihemostatic effects than either rTPA or rscuPA. Template bleeding time measurements were shorter (3.5±0.12 minutes for AFA-scuPA versus 5.3±0.36 and 5.2±0.04 minutes for rTPA and rscuPA, respectively; P<.05), α2-antiplasmin consumption was less (P<.05), and d-dimer generation was lower (P<.05).
Conclusions We conclude that antibody targeting of scuPA to fibrin increases thrombolytic and antithrombotic potencies with less impairment of hemostasis compared with rTPA and rscuPA.
Thrombotic occlusion of coronary arteries leads to acute myocardial infarction in humans, the most common cause of death in the western world.1 Restoration of coronary patency with the use of thrombolysis reduces mortality but is associated with significant limitations, including failure to achieve reperfusion (15% to 20% of patients), rethrombosis after cessation of therapy (5% to 15% of patients), and bleeding, the most devastating of which is hemorrhagic stroke (≈0.5% to 1% of patients).2 3 4 5 6 7 Neither of the general approaches taken to improve clinical outcomes—optimization of plasminogen activator function with the use of molecular methods or the use of adjunctive platelet or thrombin inhibition—has demonstrated improved clinical potency.8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 For these reasons, there continues to be interest in developing improved plasminogen activators.
We focused on targeting plasminogen activators to fibrin or activated platelets in thrombi.18 19 20 26 In the present study, we compared the thrombolytic, antithrombotic, and antihemostatic potencies of AFA-scuPA with those of two clinically available plasminogen activators, rTPA and rscuPA. The baboon models used for these studies were selected to simulate thrombosis, thrombolysis, and associated hemostatic perturbations in humans.
Compared with thrombolytic agents currently in clinical use, the hybrid recombinant plasminogen activator AFA-scuPA was found to offer improved thrombolytic and antithrombotic potency without impairment of hemostasis.
Plasminogen Activators for In Vivo Use
Several recombinant hybrid plasminogen activators have been developed in our and other laboratories.11 19 20 AFA-scuPA consists of most of the heavy chain of AFA; the antifibrin monoclonal antibody 59D8, including CH1, CH2, and the majority of the CH3 domains; and scuPA. A second isoform of AFA-scuPA lacking the Fc domain of antibody 59D8 but possessing comparable thrombolytic potency was characterized and was also used in the present study. SDS-PAGE was performed according to the method of Laemmli.19 27 Proteins were either visualized with Coomassie brilliant blue or transferred through electrophoresis to a nitrocellulose filter for Western blotting.19 Modified hybridoma cells expressing the isoform of AFA-scuPA that lacks the Fc domain were used for production of large quantities of high-purity AFA-scuPA.
For both isoforms, transfection of heavy-chain hybridoma cells was performed as described previously.19 20 Cells were grown to high density (total cell mass of ≈4×1010 cells) in AIM-V medium (GIBCO) with 1% FCS and aprotinin (600 -1252858839T-1252858839IU/L medium) in the extrafiber space of a CellMax hollow fiber (type B) bioreactor. Affinity purification was performed as described previously.18 19 After final purification, both isoforms of AFA-scuPA were ≈95% single chain and had activities of 20 000 to 30 000 IU/mg (batch-to-batch variation). This material was used for the present experiments. For each isoform of AFA-scuPA, all of the material purified from individual batch productions was thawed and pooled. From these final two pools, an aliquot was used for SDS-PAGE and Western blotting to determine purity and for calculation of specific activity. The remainder of the pooled material was aliquoted and frozen for later in vivo experiments. On the basis of the concentration and specific activity, the appropriate number of aliquots of AFA-scuPA were thawed immediately before each experiment. Excess AFA-scuPA was not refrozen or reused.
Comparison of Two Isoforms of AFA-scuPA
The two isoforms of AFA-scuPA (one containing the Fc domain and one lacking the Fc domain) were compared with the use of functional and structural assays. In human plasma clot assays and in limited in vivo experiments, the isoforms were found to be equivalent (not shown). On the basis of directly measured serial plasma levels, there was no significant difference in plasma half-life (mean value, 61.3±10.2 minutes). As described above, the purity of all material used in vivo was confirmed through SDS-PAGE and Western blotting analysis. The data that we present represent individual aliquots from pooled samples. SDS-PAGE and Western blot analysis showed the predicted composition and size of the isoform of AFA-scuPA lacking the Fc domain (Fig 1⇓). The affinity-purified isoform of AFA-scuPA lacking Fc consisted of a major band of 91 kD when electrophoresed under nonreducing conditions. This band most likely represents the fusion protein AFA-scuPA because the 91-kD protein was recognized by both anti-mouse IgG and anti-human urokinase antibodies. There also was a minor band of 180 kD, which probably represents a dimer of this isoform of AFA-scuPA. Under reducing conditions, the affinity of purified AFA-scuPA preparation consists of two bands with molecular masses of 64 and 27 kD. The 27-kD protein is recognized only by the anti-IgG antibodies, whereas the 64-kD protein is recognized only by anti-urokinase antibodies. (As described previously,28 the FD domain is not recognized by anti-IgG antibodies.) Thus, these Western blot findings are consistent with the identity of the 27-kD protein as the 59D8 light chain and the 64-kD protein as a fusion protein of rscuPA and the heavy chain of 59D8 Fab (lacking the Fc domain). For the experiments described here, both isoforms of AFA-scuPA were used because no significant functional differences were present. Aliquots from a single pooled sample of the AFA-scuPA isoform lacking Fc were used for the thrombolysis experiments, and aliquots from a single pooled sample of the AFA-scuPA containing Fc were used for the inhibition-of-thrombosis experiments.
Single-chain rTPA (Genentech) and high-molecular-weight rscuPA (a gift from Farmitalia, Milan, Italy) were also used in the present study. Both rTPA and rscuPA were dissolved in the sterile aqueous buffer supplied by the manufacturer to a final concentration of 1 mg/mL (14.2 nmol/mL). Calculations of molar amounts of rTPA, rscuPA, and AFA-scuPA were performed immediately before each experiment and were based on the protein concentration and the measured amidolytic activity of the pools of material used in these experiments (S-2288 assay for rTPA or S-2444 assay for rscuPA and AFA-scuPA). The amidolytic activities of rscuPA and rTPA were 125 000±5000 and 500 000±8000 IU/mg, respectively.
Thrombolytic, Antithrombotic, and Antihemostatic Effects In Vivo
The thrombolytic model described below was developed for quantitative and reproducible determination of lytic doses for arterial and venous thrombi. Relevant features of this model are (1) thrombi are formed in vivo before the administration of a thrombolytic agent, (2) flow geometry and thrombus formation are defined and reproducible, and (3) there are marked similarities between baboon and human hemostatic mechanisms.
Eighteen normal male juvenile baboons weighing 9 to 11.5 kg and bearing chronic arteriovenous femoral shunts were used in these experiments. Different test agents and doses were studied in these animals according to random sequence. All procedures were approved by the Institutional Animal Care and Use Committee (Emory University) in compliance with National Institutes of Health guidelines (“Guide for Care and Use of Laboratory Animals,” NIH No. 80-23, revised 1985), Public Health Service policy, the Animal Welfare Act, and related university policies. Before the experiments, all animals were observed to be disease free for ≥3 months.
For surgical procedures, animals were administered ketamine hydrochloride (20 mg/kg IM) for induction, 1% halothane by endotracheal tube for anesthetic maintenance, and buprenorphine analgesic (0.01 mg/kg every 8 hours as needed) postoperatively. For subsequent short-term immobilization in performing experimental procedures postoperatively, ketamine hydrochloride (5 to 20 mg/kg IM) was used.
Chronic exteriorized AV access shunts were surgically placed between the femoral artery and vein to permit interposition of thrombogenic devices, drug infusions, and blood sampling. The chronic AV shunts were composed of silicone rubber tubing (3.0 mm ID, Silastic, Dow Corning Corp). The arterial and venous arms of the shunt were connected with a 1-cm length of polytetrafluoroethylene (Teflon) tubing (2.8 mm ID). All materials were sterilized through autoclaving before surgical placement. These chronic AV shunts do not detectably activate platelets or fibrinogen.26 29 Thrombogenic devices30 31 32 were subsequently incorporated into the exteriorized AV shunts of awake animals for 165 minutes, and blood flows in the AV shunts were measured with a C-clamp–type ultrasonic flow probe interfaced with a Transonic T206 blood-flow analyzer. Blood flows averaged 150±35 mL/min in control animals.
Blood counts and hematocrits were measured on whole blood collected in EDTA (2 mg/mL) with the use of a JT Baker model 810 whole-blood analyzer. The mean platelet count was 318±70×103/μL in the control group.29 Template bleeding time measurements were performed on the shaved volar surface of the forearm as described previously.29 33 Plasma fibrinogen concentrations were measured as thrombin clottable protein according to the method described previously.29 30 33 The activated partial thromboplastin and thrombin clotting times were determined with the use of citrated plasma (9 vol blood:1 vol 3.8% sodium citrate) from blood samples drawn 30 and 60 minutes after the beginning of each experiment.34 Activated partial thromboplastin time (activated partial thromboplastin time reagent; Ortho Diagnostic Systems) and thrombin clotting time (bovine thrombin, final concentration 1.7 U/mL) determinations were performed with a fibrometer (Fibrosystem; Becton Dickinson). The activities of rTPA and rscuPA were measured and antigen assays were performed as described previously.8 19 35 For external validation, some assays were also performed at Haemetech (Essex Junction, VT), and plasma antigen levels for AFA-scuPA were determined by measuring both scuPA antigen and mouse IgG antigen.
Model of Thrombolysis
To study thrombolysis in baboons, we compared the lytic effects of infusion of different plasminogen activators on preformed, isotopically labeled thrombi in real time with the use of gamma camera imaging. An important distinction between this model and other thrombolysis models is that these thrombi are formed in vivo. Nonoccluding platelet-rich thrombi with fibrin-rich propagated tails (see below) were preformed endogenously for 75 minutes on thrombogenic devices interposed in chronic AV shunts in nonanticoagulated animals previously labeled with autologous 111In-platelets and 125I-fibrinogen.
The thrombogenic device was composed of a 2-cm-long segment of Dacron (synthetic polyester textile fiber) vascular graft incorporated into the chronic AV shunt with blood flowing at a rate of 40 mL/min. Because of the low-flow conditions, thrombus forming on the vascular graft concurrently propagated a fibrin-rich tail downstream in the shunt (Fig 2⇓). We selected a flow rate of 40 mL/min based on the empirical observation that at higher flow rates the fibrin-rich tail occasionally formed incompletely or embolized.* Uncrimped Dacron graft (Bioknit; 4 mm ID) was obtained from CR Bard. Segments 2 cm in length were rendered impervious to blood leakage through external wrapping in Parafilm (American Can Company) and 5.3-mm-ID “heat shrink” polytetrafluoroethylene tubing. Connections were constructed to ensure that the devices were isodiametric and suitable for incorporation into the AV shunts.31 33 34 A platelet-rich thrombus formed rapidly on the segment of Dacron vascular graft, reaching a plateau by ≈75 minutes.
Autologous baboon platelets were labeled with 1 mCi 111In-oxine as described previously34 and reinjected ≥1 hour before placement of the thrombogenic devices. Labeling efficiencies averaged 90%. 111In-Labeled platelets were functionally normal.26 31 Baboon fibrinogen was purified through β-alanine precipitation and labeled with 125I according to the ICI method as described previously.29 34 Labeling efficiency averaged 70%; the labeled fibrinogen was >90% clottable. A 5-μCi dose of labeled fibrinogen was injected intravenously 10 minutes before exposure of the device.
To maintain the amount of thrombus that formed over 75 minutes as relatively constant, ie, without occlusion during the subsequent hour of study (Fig 2⇑), heparin infusions were initiated, beginning at 75 minutes and continuing for the additional 60 minutes of the experiment. Heparin was administered as a bolus of 100 U/kg, followed by continuous intravenous infusions of 100 U/kg over 60 minutes. Concurrent with heparin administration, plasminogen activators were administered as a 10% bolus followed by a continuous infusion of the remaining dose, as described below. The preexisting labeled thrombus during lysis was measured for each plasminogen activator throughout the 1-hour infusion and the subsequent 30-minute observation period. Blood tests of thrombolysis, thrombosis, and hemostatic function (see below) were also performed on blood samples collected in tubes containing 3.8% sodium citrate before placement of the thrombogenic devices (0 minutes), before test infusions began (75 minutes), at the end of the test infusions (135 minutes), and at the end of the experiments (165 minutes). Measurements were performed on blinded, coded samples for plasma rTPA and rscuPA antigen and activity levels and for plasma activity levels of plasminogen, α2-antiplasmin, and plasminogen activator inhibitor–1.
Plasminogen activators were administered intravenously in combination with heparin beginning at 75 minutes, when the thrombus was established and stabilized. For each plasminogen activator, 10% of the total dose was administered as an intravenous bolus, followed by the remainder of the dose infused over 1 hour. Dose-response studies were performed for rTPA (2.85 nmol/kg [0.2 mg/kg], 14.2 nmol/kg [1.0 mg/kg], or 28.5 nmol/kg [2.0 mg/kg]), rscuPA (2.85 nmol/kg [0.15 mg/kg], 14.2 nmol/kg [0.77 mg/kg], 28.5 nmol/kg [1.54 mg/kg], or 72.2 nmol/kg [3.9 mg/kg]), and AFA-scuPA (1.89 nmol/kg [0.15 mg/kg]). Imaging of both the platelet-rich graft thrombus and the fibrin-rich tail was performed as described below, including a final image 30 minutes after completion of the infusion. Control studies were carried out in six animals receiving no heparin and in six animals receiving heparin only.
Images of the vascular graft, including proximal and distal segments of the AV shunts, were acquired with a General Electric 400T MaxiCamera and stored and analyzed with a Medical Data Systems A3 image processing system (Medtronic) interfaced with the camera.31 The low-energy peak (172 KeV) of 111In was imaged with a 10% energy window. Dynamic images were acquired at 5-minute intervals. Immediately after each dynamic study, standards were imaged, including a syringe containing 5.0 mL of whole blood (blood standard) and an identical thrombogenic device filled with static autologous blood (device standard). The imaging routines and isotopic detection protocols for these shunt studies have been reported previously.26 30 32 33
For thrombolysis experiments, the thrombus radioactivity was counted in two regions of interest that were analyzed separately: the platelet-rich thrombus at the Dacron graft, analyzed over 2 cm (8×10 pixel region of interest), and the fibrin-rich tail, analyzed over a length of 20 cm (80×10 pixel). Images were acquired at 5-minute intervals. The total number of deposited platelets in each region (labeled plus unlabeled platelets) was calculated by dividing the deposited platelet activity (cpm) by the circulating blood activity (cpm/mL) and multiplying by the circulating platelet count (platelets/mL) as described previously.33 Radioactivity values referred to platelet activity only, with all blood measurements having been corrected for the small fraction of nonplatelet isotope in each experiment. Nonplatelet plasma activities averaged 10±1% (n=24) of whole blood activity in these studies. 111In-platelet emissions were counted to measure thrombus formation. For this measurement, emissions from both the Dacron graft and the fibrin-rich tail regions of the experimental device were counted in 2-cm-long (8×10 pixel) regions of interest. Deposited 111In-labeled platelet activity was calculated by subtracting the device standard activity from each region of interest.
Fibrin in platelet-rich and fibrin-rich thrombi was measured with homologous 125I-labeled fibrinogen. Baboon 125I-fibrinogen (5 μCi) was injected intravenously 5 minutes before incorporation of the device in the AV shunt. After completion of the experiment, the device was thoroughly washed with isotonic saline solution perfused at 20 mL/min. The vascular graft thrombus was then separated from the propagated tail for counting of emissions. The 125I activity was measured at ≥30 days after the study to allow for the decay of 111In activity (half-life, 2.8 days). Total fibrin deposition was calculated by dividing the deposited 125I-fibrin activity (cpm) by the clottable fibrinogen activity (cpm/mL) and multiplying by the plasma fibrinogen level (mg/mL). The concentration of fibrinogen in plasma was estimated spectrophotometrically with the use of a modification of the method of Jacobsson (see References 29 and 34).
Model of Thrombus Formation
In studies designed to measure the effects of lytic agents on rates of new thrombus formation, thrombogenic devices were interposed between the arms of the permanent shunt system of awake animals for 1 hour. Use of the device resulted in formation of platelet-rich arterial-type thrombus on 2-cm-long segments of Dacron vascular graft and the formation of fibrin-rich venous-type thrombus in a chamber of expanded diameter and disturbed flow placed immediately distal to the segment of vascular graft, as described previously.30 33 34 36 Dacron vascular graft segments were prepared from externally supported uncrimped knitted Dacron grafts (2 cm long with a 4.0 mm ID) prepared as described above. At the flow rate used in these studies (20 mL/min), the wall shear rate on the Dacron graft was 50 s−1. The “chamber” was 2 cm distal to the Dacron segment and consisted of an expansion of tubing diameters from 3.2 to 9.3 mm. As reported previously,30 36 this expansion region produces a complex flow pattern, with reverse flow along the wall and formation of a captive annular vortex. Before thrombus formed, shear rates along the chamber walls were quite low, with essentially no flow in the corner regions. In general, this pattern of flow recirculation results in a prolonged residence time of blood cells and procoagulant species and increases the likelihood that these elements will form thrombus. Blood flow was maintained at 20 mL/min with the use of a variable-speed peristaltic roller pump (Masterflow model 7016; Cole-Parmer Instrument Co) interposed between the device and the femoral vein (ie, distal to the device). As described for the thrombolysis studies, 111In-platelet emissions were counted to measure thrombus formation. Emissions from both the Dacron graft and chamber regions of the experimental device were counted in 2-cm-long (8×10 pixel) regions of interest. Deposited 111In-labeled platelet activity was calculated by subtracting the device standard activity from each region of interest.
Four indicators of acute thrombus formation were measured. Deposited autologous 111In-platelets were counted through scintillation camera imaging. The accumulation of 125I-fibrin was measured through gamma counting. Device patency was measured with the use of Doppler flow analysis through the shunt using a C-clamp–type flow probe interfaced with a Transonic T206 blood flow analyzer. Plasma levels of markers of thrombosis (platelet factor 4, β-thromboglobulin, fibrinopeptide A, and thrombin-antithrombin III complex)36 37 were determined at three time points: at baseline before the incorporation of thrombogenic segments into the AV shunt, after exposure of the thrombogenic segments to arterial rates of blood flow for 1 hour,30 33 34 36 and at the conclusion of the experiment.
Infusion of either rscuPA or AFA-scuPA and 111In-labeled platelet imaging was begun as soon as blood flow was established. rscuPA (total dose of 3.7 nmol/kg) or AFA-scuPA (total dose of 0.31 nmol/kg), diluted in saline solution before the experiment, was administered as a continuous infusion over 1 hour, with the goal of maintaining a constant systemic drug level in the shunt throughout the experiment. The dosing regimen designed was based on previous observations in a rabbit model19 because the plasma half-life of AFA-scuPA in the baboon was not known before these experiments. If a constant systemic drug level was not achieved because of progressive accumulation due to the prolonged plasma half-life of AFA-scuPA, the calculated thrombolytic potency of AFA-scuPA would be underestimated.
Statistical analysis of data was performed with the following methods. For comparison of a single end point after different treatments, eg, template bleeding times in animals treated with rTPA, rscuPA, or AFA-scuPA, the data were analyzed with SigmaStat. When the distribution of the mean values was equal, data were analyzed with the Student's t test for two-tailed analysis. When the mean values had unequal distribution, the Mann-Whitney rank sum test was used for analysis. Analysis of curves, such as lysis curves based on measurement of platelet deposition, was performed with the SysStat software program, in which the P values were obtained for the ANOVA through repeated measures ANOVA. For the analyses presented in Figs 3D⇓ through F, 4D through F, 5C, and 5D, the measured values of 111In-labeled platelets (or of 125I-fibrin remaining on the graft segment) were adjusted for steady-state plasma level or dose before comparison of the results by group.
Development and Use of Thrombolysis Model
The thrombolytic model is schematically shown in Fig 2⇑. In this model, thrombi (to be lysed) are preformed in vivo. The addition of systemic heparin was found to maintain stable, nonocclusive thrombi of defined morphology and geometry throughout the study period in the absence of plasminogen activator infusion. Six animals were treated with heparin but did not receive plasminogen activator therapy. In these animals, platelets accumulated rapidly, reaching a plateau by 75 minutes, with no significant changes in platelet deposition during the subsequent 90 minutes. This formation of stable platelet-rich thrombi (within the graft) and fibrin-rich thrombi (propagated tail) was quantified with the use of gamma camera imaging (Fig 2⇑, middle and bottom). As shown, sufficient numbers of platelets were present in the fibrin-rich tail to quantify the amount of thrombus with imaging. In contrast, a second group of controls consisted of six animals that received neither a plasminogen activator nor heparin (data not shown). Five of the six animals treated in this way had an occluded graft at 100±15 minutes. Thrombi formed in the absence of heparin were not suitable for thrombolytic studies because (1) the timing of occlusion was unpredictable, (2) the occluding thrombus was inaccessible to systemic plasminogen activators, and (3) when lysed, the occlusive thrombi embolized erratically, obscuring the kinetics of thrombolysis.
Thrombolytic Comparison of AFA-scuPA With rTPA and rscuPA
Four to six animals were evaluated in each control and experimental group to determine the relative thrombolytic potencies of rTPA, rscuPA, and AFA-scuPA for dissolution of platelet-rich and fibrin-rich thrombi. Based on plasma measurements, the plasma half-lives were 4.3±1.0, 5.0±1.7, and 61.3±10.2 minutes for rTPA, rscuPA, and AFA-scuPA, respectively. The same cohort of animals received each of the three different plasminogen activators on serial days in random sequence. The effects of administering a plasminogen activator on platelets or fibrin present in either the platelet-rich or fibrin-rich thrombi are shown in Figs 3 and 4⇑⇓. Dose-dependent thrombolysis was achieved for both rTPA (Fig 3⇑) and rscuPA (data not shown). As the dose of the plasminogen activator was increased, the thrombus was lysed more quickly and more completely. However, when the rate of lysis for different doses of rTPA was adjusted for the steady-state plasma level of TPA, no difference was observed, which is consistent with a dose-dependent effect (Fig 3D and 3F⇑⇑). The amount of fibrin remaining in the thrombus was inversely related to the dose (Fig 3B⇑) or plasma concentration (Fig 3E⇑), ie, as the plasma level of TPA increased, the amount of fibrin remaining decreased. Similar dose-response relations were observed for lysis of the platelet-rich thrombus present on the Dacron graft and the fibrin-rich tails, although both the rate and extent of thrombolysis were less for platelet-rich thrombi. A comparable dose-response relation was observed for rscuPA (data not shown). Fig 4A and 4C⇓⇓ shows that a low dose of AFA-scuPA (1.89 nmol/kg) produced equivalent thrombolysis to that observed for rTPA (14.2 nmol/kg) and rscuPA (28.5 nmol/kg), exhibiting 7.5- and 15-fold higher potency based on administered dose, respectively. Furthermore, when the measured amount of fibrin remaining in the graft was adjusted for the steady-state plasma concentration, AFA-scuPA achieved significantly more thrombus dissolution at the measured doses (Fig 4E⇓). Because the measured plasma half-lives of rTPA and rscuPA were much shorter than that of AFA-scuPA, a comparison of potency based on relative plasma levels is also useful. When the thrombolysis data are adjusted on the basis of steady-state plasma concentrations for equivalent thrombolytic effect, lysis with AFA-scuPA was significantly more rapid than with rTPA or rscuPA (Fig 4D and 4F⇓⇓), a clear departure from a single-dose dependence. Thus, although potency is in part related to the longer half-life of AFA-scuPA, it is also due to increased rates of lysis measured. This difference is most marked when the rates of dissolution of platelets in the fibrin-rich tail are compared.
Antihemostatic Comparison of AFA-scuPA With rTPA and rscuPA
Significant differences were observed in the measurements of template bleeding time (Table 2⇓). Bleeding times were doubled in the animals receiving 14.2 nmol/kg rTPA (5.3±0.36 minutes, P<.05 versus control) or 28.5 nmol/kg rscuPA (6.5±1.0 minutes, P<.05 versus control), whereas there was no prolongation of the bleeding time in animals receiving a comparable thrombolytic dose of AFA-scuPA (3.5±0.6 minutes in animals receiving 1.89 nmol/kg, P=NS versus control). In addition, the consumption of α2-antiplasmin and the increase in d-dimer were significantly less in animals receiving AFA-scuPA than in those receiving equivalent doses, 28.5 nmol/kg rscuPA and 14.2 nmol/kg rTPA (Table 1⇓). Because thrombin-antithrombin III complex formation was comparable for all three plasminogen activators, the amount of soluble thrombin formed during thrombogenesis appeared to be relatively constant during the experiments.
Inhibition of Thrombus Formation
AFA-scuPA inhibited thrombus formation more effectively than rscuPA (Fig 5⇓). The antithrombotic effects of rscuPA and AFA-scuPA were compared under low-flow conditions (20 mL/min) with the use of a thrombogenic device consisting of a proximal segment of vascular graft followed by an expanded chamber with static and disturbed flow. In this model, platelet-rich arterial-type thrombi form on the graft and fibrin-rich venous-type thrombi form in the expanded chamber. Each of five animals bearing segments of Dacron vascular graft and the expansion chamber was evaluated on serial days in random sequence (Fig 5⇓). Either rscuPA or AFA-scuPA was infused proximal to the thrombogenic segment for 60 minutes, and the rate and extent of thrombus formation were quantified with gamma camera imaging of 111In-labeled platelets. Thrombi formed rapidly on segments of Dacron vascular graft in untreated controls, reaching a peak value by 60 minutes. Platelet and fibrin depositions were interrupted in the chamber by both rscuPA (3.7 nmol/kg per hour, achieving a plasma concentration of 23.2±0.2 pmol/mL) and AFA-scuPA (0.31 nmol/kg per hour, with plasma concentrations below the limit of detection of the assay used). On the basis of administered doses, AFA-scuPA was ≥11-fold more potent than rscuPA in preventing platelet accumulation on the segment of vascular graft (P<.0001).
Hemostatic determinations were also measured. Fibrinogen levels and platelet counts in animals treated with either rscuPA or AFA-scuPA did not differ (P=NS) (Table 3⇓). In addition, plasminogen, α2-antiplasmin, and plasminogen activator inhibitor–1 levels were not significantly changed compared with controls for rscuPA or AFA-scuPA (Table 3⇓).
The results of the present study demonstrate that AFA-scuPA, a hybrid recombinant plasminogen activator, increased thrombolytic and antithrombotic potency and specificity in comparison with either rTPA or rscuPA. These conclusions are based on results in relevant primate models showing (1) the doses necessary to obtain comparable thrombolysis, (2) the rates of thrombolysis achieved for AFA-scuPA, rscuPA, and rTPA, and (3) that the template bleeding time was less prolonged and the degradation of α2-antiplasmin and generation of d-dimer were reduced in animals treated with AFA-scuPA.
The thrombolytic model developed in baboons for these studies (Fig 2⇑) is quantitative and reproducible, discriminates lytic doses for arterial (platelet-rich) and venous (fibrin-rich) thrombi formed in vivo, and provides concurrent evaluation of hemostasis. In this thrombolysis model, nonocclusive platelet-rich and fibrin-rich thrombi are preformed through incorporation of a segment of Dacron vascular graft into an exteriorized AV shunt and a decrease in blood flow to 40 mL/min. This model is useful for comparing different thrombolytic agents because the flow geometry is defined, thrombus formation is reproducible, and baboon and human hemostatic mechanisms are similar.31 32 33 34
The novel feature of this model compared with other thrombolysis models is that the formation of thrombus occurs in vivo before the administration of thrombolytic agent. Thus, although this model, which uses thrombus initiated and anchored by a synthetic Dacron graft, is not an accurate representation of thrombosis initiated by coronary plaque rupture in humans, subsequent mechanisms of platelet and fibrin deposition are similar to humans, thereby offering significant advantages over other reported models as a method of comparing plasminogen activators in vivo.
Four related observations from these in vivo experiments merit additional comment. First, when compared with rTPA and rscuPA, AFA-scuPA therapy exhibits greater thrombolytic potency and increased rates of thrombolysis with less impairment of hemostatic function than either TPA or scuPA. Second, the plasma half-life of AFA-scuPA is significantly longer than that of either rTPA or rscuPA. In fact, as seen in Fig 4F⇑, AFA-scuPA lyses thrombi six times as rapidly as rTPA and rscuPA when rates are adjusted for antigen blood level. Third, hemostatic plug formation is normal in animals receiving AFA-scuPA, whereas animals receiving comparable thrombolytic doses of rTPA and rscuPA showed prolongations in template bleeding time. Fourth, low doses of AFA-scuPA inhibited the rate and extent of both platelet-rich and fibrin-rich thrombus formation, indicating that thrombolytic doses of AFA-scuPA also inhibited thrombus formation.
The increase in thrombolytic potency of AFA-scuPA for fibrin-rich thrombi is attributable to the increased specificity and the reduced clearance rate of this hybrid plasminogen activator. In this regard, our findings are consistent with previous reports by our group19 and others.38 The increased thrombolytic potency observed for platelet-rich arterial-type thrombi suggests that sufficient fibrin is present, even in these thrombi, for effective targeting by AFA-scuPA. The experiments reported suggest that higher doses of AFA-scuPA might result in improved thrombolysis without unacceptable bleeding and, if so, that AFA-scuPA may be useful clinically because of improved efficacy over TPA and rscuPA, the tolerated doses of which are limited by bleeding.
The steady-state plasma levels of AFA-scuPA, rTPA, and rscuPA were measured directly, with the thrombogenic device in place, and both antigen and activity measurements were used. Independent confirmation of the plasma half-lives was obtained through (1) measurement of clearance rates for 125I-labeled rTPA, rscuPA, and AFA-scuPA in baboons with a thrombogenic device in place and (2) measurement of plasma rTPA, rscuPA antigen levels, and clearance of 125I-labeled rTPA, rscuPA, and AFA-scuPA after bolus infusion in animals without a thrombogenic device. The results obtained for rscuPA and rTPA are consistent with prior reports. The prolonged plasma half-life of AFA-scuPA is much less than that of immunoglobulin, which is ≈14 days.
An important difference in the plasminogen activators studied here is the effect on template bleeding times. Clinically, both rTPA39 and rscuPA prolong template bleeding times, and a correlation has been reported between elevation in template bleeding time and intracranial bleeding.39 The mechanism for this is not clear but may relate to systemic effects of fibrinogen breakdown products. It is not known whether the specificity of AFA-scuPA and its reduced systemic effects will also reduce intracranial bleeding.
Measurements of both AFA-scuPA plasma concentration and antithrombotic potency are relevant when differences in template bleeding time, α2-antiplasmin degradation, and d-dimer generation observed at comparable thrombolytic doses are interpreted. For example, a minimal decrease in α2-antiplasmin is expected, based on both the low plasma concentration of AFA-scuPA and its fibrin selectivity, since AFA-scuPA likely generates less systemic plasmin than either rTPA or rscuPA. The difference in the template bleeding times and d-dimer production is also intriguing. For both template bleeding time and d-dimer generation, less effect was observed in animals treated with AFA-scuPA than in animals treated with either rTPA or rscuPA. Several potential mechanisms may account for these observations. First, it is possible that the d-dimer assay used here also measures fibrinogen breakdown products. In this case, the concordance between d-dimer levels and template bleeding times would be due to the known inhibitory effect of fibrinogen breakdown products on platelet aggregation. Alternatively, AFA-scuPA may inhibit thrombosis by decreasing the total amount of thrombus lysed (because less new thrombus is formed) during the experimental period and thus less d-dimer is produced. The observation that there is no significant difference in thrombin-antithrombin III levels among the three plasminogen activators is consistent with either explanation. Further investigation will be necessary to reach a definitive conclusion about this effect.
The thrombosis model that we used was designed to simulate the process of rethrombosis occurring in 5% to 15% of patients after thrombolytic therapy.5 6 7 In this model, AFA-scuPA is 11-fold more efficient than rscuPA in preventing formation of both platelet-rich and fibrin-rich thrombi. Although we had anticipated the increased antithrombotic potency observed in the expansion regions of low shear venous-type blood flow, the enhanced potency observed in the Dacron vascular graft was of interest. It has previously been argued that fibrin targeting is not applicable to treatment of coronary artery thrombi because arterial thrombi tend to be platelet rich and fibrin poor. Our data indicate that there is sufficient fibrin present in the platelet-rich thrombi found in the vascular graft segments to effectively target rscuPA to platelet-rich thrombi using an antifibrin antibody and to inhibit subsequent thrombus formation.
In summary, these data demonstrate that the recombinant hybrid plasminogen activator AFA-scuPA combines increased thrombus specificity and prolonged plasma half-life to effect improved thrombolytic and antithrombotic potency in primate models for thrombolysis and inhibition of thrombus formation. In these models, AFA-scuPA both lyses thrombi and prevents rethrombosis more effectively and with less impairment of hemostasis than the native plasminogen activators rTPA or rscuPA.
Selected Abbreviations and Acronyms
|AFA-scuPA||=||recombinant hybrid plasminogen activator antifibrin antibody 59D8–low-molecular-weight single-chain urokinase-type plasminogen activator|
|rscuPA||=||recombinant single-chain urokinase-type plasminogen activator|
|rTPA||=||recombinant tissue plasminogen activator|
This work was supported in part by National Institutes of Health grants HL-41619, HL-31950, HL-02414, HL-44307, HL-31469, and RR-00165 and by a grant from the Bristol-Myers Squibb Pharmaceutical Research Institute. We thank Joann Aaron and Kate W. Harris for assistance with graphics and editing.
*The flow rate of 40 mL/min may be applicable to flow in a stenosed coronary artery. For example, for a 70-kg individual at rest with coronary blood flow of 250 mL/min (based on 5% of total blood) proximal to a 90% stenosis at the midpoint of an epicardial coronary artery, calculated flow rates would be ≈25 to 100 mL/min.
- Received September 21, 1995.
- Revision received March 21, 1996.
- Accepted March 26, 1996.
- Copyright © 1996 by American Heart Association
Aberg A, Bergstrand R, Johansson S, Ulvenstam G, Vedin A, Wedel H. Declining trend in mortality after myocardial infarction. Br Heart J. 1984;51:346-351.
Chesebro JH, Knatterud G, Braunwald E. Thrombolytic therapy. N Engl J Med. 1988;319:1544-1545.
Topol EJ, Califf RM, George BS, Kereiakes DJ, Abbottsmith CW, Candela RJ, Lee KL, Pitt B, Stack RS, O'Neill WW. A randomized trial of immediate versus delayed elective angioplasty after intravenous tissue plasminogen activator in acute myocardial infarction. N Engl J Med. 1987;317:581-588.
Bode C, Meinhardt G, Runge MS, Freitag M, Nordt T, Arens M, Newell JB, Kubler W, Haber E. Platelet-targeted fibrinolysis enhances clot lysis and inhibits platelet aggregation. Circulation. 1991;84:805-813.
Browne MJ, Carey JE, Chapman CG, Tyrrell AW, Entwisle C, Lawrence GM, Reavy B, Dodd I, Esmail A, Robinson JH. A tissue-type plasminogen activator mutant with prolonged clearance in vivo: effect of removal of the growth factor domain. J Biol Chem. 1988;263:1599-1602.
Jackson CV, Crowe VG, Craft TJ, Sundboom JL, Grinnell BW, Bobbitt JL, Burck PJ, Quay JF, Smith GF. Thrombolytic activity of a novel plasminogen activator, LY210825, compared with recombinant tissue-type plasminogen activator in a canine model of coronary artery thrombosis. Circulation. 1990;82:1062-1063.
Holvoet PY, Laroche Y, Lijnen HR, Van Cauwenberge R, Demarsin E, Brouwers E, Matthyssens G, Collen D. Characterization of a chimeric plasminogen activator consisting of a single-chain Fv fragment derived from a fibrin fragment D-dimer-specific antibody and a truncated single-chain urokinase. J Biol Chem. 1991;266:19717-19724.
Larsen GR, Henson K, Blue Y. Variants of human tissue-type plasminogen activator: fibrin binding, fibrinolytic, and fibrinogenolytic characterization of genetic variants lacking the fibronectin finger-like and/or the epidermal growth factor domains. J Biol Chem. 1988;263:1023-1029.
Lijnen HR, Nelles L, Holmes WE, Collen D. Biochemical and thrombolytic properties of a low molecular weight form (comprising Leul44 through Leu411) of recombinant single-chain urokinase-type plasminogen activator. J Biol Chem. 1988;263:5594-5598.
Lijnen HR, Nelles L, Van Hoef B, De Cock F, Collen D. Biochemical and functional characterization of human tissue-type plasminogen activator variants obtained by deletion and/or duplication of structural/functional domains. J Biol Chem. 1990;265:5677-5683.
Nelles L, Lijnen HR, Collen D, Holmes WE. Characterization of recombinant human single chain urokinase-type plasminogen activator mutants produced by site-specific mutagenesis of lysine 158. J Biol Chem. 1987;262:5682-5689.
Nelles L, Lijnen HR, Collen D, Holmes WE. Characterization of a fusion protein consisting of amino acids 1 to 263 of tissue-type plasminogen activator and amino acids 144 to 411 of urokinase-type plasminogen activator. J Biol Chem. 1987;262:10855-10862.
Runge MS, Quertermous T, Zavodny PJ, Love TW, Bode C, Freitag M, Shaw SY, Huang PL, Chou D, Mullins D, Schnee JM, Savard CE, Rothenberg ME, Newell JB, Matsueda GR, Haber E. A recombinant chimeric plasminogen activator with high affinity for fibrin has increased thrombolytic potency in vitro and in vivo. Proc Natl Acad Sci U S A. 1991;88:10337-10341.
Schnee JM, Runge MS, Matsueda GR, Hudson NW, Seidman JG, Haber E, Quertermous T. Construction and expression of a recombinant antibody-targeted plasminogen activator. Proc Natl Acad Sci U S A. 1987;84:6904-6908.
Heras M, Chesebro JH, Penny WJ, Bailey KR, Badimon L, Fuster V. Effects of thrombin inhibition on the development of acute platelet-thrombus deposition during angioplasty in pigs: heparin versus recombinant hirudin, a specific thrombin inhibitor. Circulation. 1989;79:657-665.
Jang I, Gold HK, Leinbach RC, McNary JE, Fallon JT, Collen D. Acceleration of reperfusion by combination of rt-PA and a selective thrombin inhibitor, argatroban. Circulation. 1989;80(suppl II):II-217. Abstract.
Rudd MA, George D, Johnstone MT, Moore RT, Collins L, Rabbani LE, Loscalzo J. Effect of thrombin inhibition on the dynamics of thrombolysis and on platelet function during thrombolytic therapy. Circ Res. 1992;70:829-834.
Yasuda T, Gold HK, Leinbach RC, Yaoita H, Fallon JT, Guerrero L, Napier MA, Bunting S, Collen D. Kistrin, a polypeptide platelet GPIIb/IIIa receptor antagonist, enhances and sustains coronary arterial thrombolysis with recombinant tissue-type plasminogen activator in a canine preparation. Circulation. 1991;83:1038-1047.
Savage B, McFadden PR, Hanson SR, Harker LA. The relation of platelet density to platelet age: survival of low- and high-density 111Indium-labeled platelets in baboons. Blood. 1986;68:386-393.
Harker LA, Hanson SR. Experimental arterial thromboembolism in baboons: mechanisms, quantitation and pharmacologic prevention. J Clin Invest. 1979;64:559-560.
Hanson SR, Kotze HF, Savage B, Harker LA. Platelet interactions with Dacron vascular grafts: a model of acute thrombosis in baboons. Arteriosclerosis. 1985;5:595-603.
Hanson S, Harker L. Interruption of acute platelet-dependent thrombosis by the synthetic antithrombin D-phenylalanyl-L-prolyl-L-arginyl chloromethyl ketone. Proc Natl Acad Sci U S A. 1988;85:3184-3188.
Kelly AB, Marzec UM, Krupski W, Bass A, Cadroy Y, Hanson SR, Harker LA. Hirudin interruption of heparin-resistant arterial thrombus formation in baboons. Blood. 1991;77:1006-1012.
Runge MS, Bode C, Matsueda GR, Haber E. Antibody-enhanced thrombolysis: targeting of tissue plasminogen activator in vivo. Proc Natl Acad Sci U S A. 1987;84:7659-7662.
Cadroy Y, Maraganore JM, Hanson SR, Harker LA. Selective inhibition by a synthetic hirudin peptide of fibrin-dependent thrombosis in baboons. Proc Natl Acad Sci U S A. 1991;88:1177-1181.
Gruber A, Hanson SR, Kelly AB, Yan BS, Bang N, Griffin JH, Harker LA. Inhibition of thrombus formation by activated recombinant protein C in a primate model of arterial thrombosis. Circulation. 1990;82:578-585.
Holvoet P, Laroche Y, Stassen JM, Lijnen HR, Van Hoef B, De Cock F, Van Houtven A, Gansemans Y, Matthyssens G, Collen D. Pharmacokinetic and thrombolytic properties of chimeric plasminogen activators consisting of a single-chain Fv fragment of a fibrin-specific antibody fused to single-chain urokinase. Blood. 1993;81:696-703.
Gimple LW, Gold HK, Leinbach RC, Coller BS, Werner W, Yasuda T, Johns JA, Ziskind AA, Finkelstein D, Collen D. Correlation between template bleeding times and spontaneous bleeding during treatment of acute myocardial infarction with recombinant tissue-type plasminogen activator. Circulation. 1989;80:581-588.