Biochemical properties of conjugates of urokinase-type plasminogen activator with a monoclonal antibody specific for cross-linked fibrin

Conjugation of an antibody to cross-linked fibrin for targeted delivery of anti-restenotic drugs

There is an urgent need to treat restenosis, a major complication of the treatment of arteries blocked by atherosclerotic plaque, using local delivery techniques. We observed that cross-linked fibrin (XLF) is deposited at the site of surgical injury of arteries. An antibody to XLF, conjugated to anti-restenotic agents, should deliver the drugs directly and only to the site of injury. An anti-XLF antibody (H93.7C.1D2/48; 1D2) was conjugated to heparin (using N-succinimidyl 3-(2-pyridyldithio)propionate), low molecular weight heparin (LMWH) (adipic acid dihydrazide) and rapamycin (1-ethyl-3-(3-dimethylaminopropyl)carbodiimide/N-hydroxysuccinimide), and the conjugates purified and tested for activity before use in vivo. Rabbits had their right carotid arteries de-endothelialised and then given a bolus of 1D2-heparin, 1D2-LMWH or 1D2-rapamycin conjugate or controls of saline, heparin, LMWH, rapamycin or 1D2 (+/-heparin bolus) and sacrificed after 2 or 4 weeks (12 groups, n=6/group). Rabbits given any of the conjugates had minimal neointimal development in injured arteries, with up to 59% fewer neointimal cells than those given control drugs. Rabbits given 1D2-heparin or 1D2-LMWH had an increased or insignificant reduction in luminal area, with positive remodelling, while the medial and total arterial areas of rabbits given 1D2-rapamycin were not affected by injury. Arteries exposed to 1D2-heparin or 1D2-rapamycin had more endothelial cells than rabbits given control drugs. Thus, XLF-antibodies can site-deliver anti-restenotic agents to injured areas of the artery wall, where the conjugates can influence remodelling, re-endothelialisation and neointimal cell density, with reduced neointimal formation.

Targeted delivery of heparin and LMWH using a fibrin antibody prevents restenosis

This study investigates a stent-less local delivery system for anti-restenotic agents utilizing antibodies to cross-linked fibrin (XLF). Heparin and low molecular weight heparin (LMWH) were conjugated to an antibody to cross-linked fibrin D-dimer (1D2). Rabbit right carotid arteries were injured with a balloon catheter, then the animals were given a bolus injection of 40 microg/kg 1D2-heparin (26-70 microg/kg heparin) or 1D2-LMWH (29-80 microg/kg LMWH) conjugates or controls of saline (0.5 ml/kg), heparin (150 U/kg), LMWH (2 mg), or 1D2 (40 microg/kg), with or without a heparin bolus and sacrificed after 2 weeks (8 groups, n = 6/group). The injured artery of rabbits given 1D2-heparin or 1D2-LMWH conjugates had reduced neointimal development, with decreased luminal narrowing and positive remodelling compared with animals given control drugs. Animals given 1D2-heparin conjugate (with a heparin bolus) had three to five times more endothelial cells than the rabbits given saline or unconjugated heparin, while rabbits given 1D2-LMWH conjugate had up to 59% fewer neointimal cells than those given unconjugated drugs. There was little difference in extracellular matrix organization or composition. Thus cross-linked fibrin-antibodies can site-deliver anti-restenotic agents to injured areas of the artery wall where they influence wall remodelling and endothelial and neointimal cell number, reducing neointimal formation without systemic complications. Local delivery of anti-restenotic agents should minimise systemic effects, bleeding complications and potentially the cost of treatment due to a single, lower dose.

A fast-acting, modular-structured staphylokinase fusion with Kringle-1 from human plasminogen as the fibrin-targeting domain offers improved clot lysis efficacy

To develop a fast-acting clot dissolving agent, a clot-targeting domain derived from the Kringle-1 domain in human plasminogen was fused to the C-terminal end of staphylokinase with a linker sequence in between. Production of this fusion protein in Bacillus subtilis and Pichia pastoris was examined. The Kringle domain in the fusion protein produced from B. subtilis was improperly folded because of its complicated disulfide-bond profile, whereas the staphylokinase domain produced from P. pastoris was only partially active because of an N-linked glycosylation. A change of the glycosylation residue, Thr-30, to alanine resulted in a non-glycosylated biologically active fusion. The resulting mutein, designated SAKM3-L-K1, was overproduced in P. pastoris. Each domain in SAKM3-L-K1 was functional, and this fusion showed fibrin binding ability by binding directly to plasmin-digested clots. In vitro fibrin clot lysis in a static environment and plasma clot lysis in a flow-cell system demonstrated that the engineered fusion outperformed the non-fused staphylokinase. The time required for 50% clot lysis was reduced by 20 to 500% under different conditions. Faster clot lysis can potentially reduce the degree of damage to occluded heart tissues.

A tissue plasminogen activator/P-selectin fusion protein is an effective thrombolytic agent

BACKGROUND

P-selectin is expressed on the surface of activated endothelial cells and platelets. We hypothesized that a tissue plasminogen activator (TPA)/P-selectin fusion protein would have not only thrombolytic activity but also might target TPA to the thrombi. In addition, it seemed possible that this chimeric protein would competitively inhibit the binding of native P-selectin on endothelial cells and platelets to leukocytes and thus further promote thrombolysis.

METHODS AND RESULTS

The full-length, plasminogen activator inhibitor-1-resistant form of TPA (TPAIR) together with two TPAIR/P-selectin fusion constructs (P280IR and P121IR) were expressed with the use of baculovirus vectors. After infection of Sf21 cells with the recombinant baculovirus, recombinant TPAIR and P-selectin/TPAIR fusion proteins were purified with the use of metal ion chromatography. The intact protease activity of TPAIR and the ligand binding capability of P-selectin were confirmed through indirect chromogenic and cell binding assays, respectively. These molecules were assessed both in vitro and in vivo for thrombolytic activity. In vitro clot lysis assays indicated equal efficacy of TPAIR, P280IR, and P121IR (P > .5). The in vivo efficacy was tested in a cyclic flow variation model with the use of the rat mesenteric artery. Compared with saline control treatment, reduction in cyclic flow variations was significant (P < .05) and similar (P > .5) among TPAIR, P280IR, and P121IR. No significant bleeding was noted among treated animals.

CONCLUSIONS

Chimeric proteins P280IR and P121IR have clot lysis activities that are similar to TPAIR both in vitro and in vivo. These chimeric proteins also bind to P-selectin ligand in vitro. Thus, these proteins may provide an efficient method of targeting TPA to the thrombotic region. Further experimental analysis with the use of larger animal coronary occlusion models should help determine the future value of these proteins as clinical therapeutic agents.

Thrombolysis with prourokinase versus urokinase: an in vitro comparison

PURPOSE

Despite advantages demonstrated in vitro, no single thrombolytic agent has been clearly shown to be superior to another in the clinical setting. Prourokinase has recently received attention as a new thrombolytic agent with higher fibrin specificity. The thrombolytic activity of prourokinase, however, remains ill defined. The purpose of this study was to evaluate thrombolysis with prourokinase in comparison to urokinase in vitro.

METHODS

We used an in vitro parallel channel perfusion model that simulates catheter-directed thrombolysis in the peripheral arterial system. Radiolabeled thrombi were subjected to 90 minutes of endhole catheter-directed infusion with either prourokinase 5000 IU/ml, urokinase 5000 IU/ml; or 5% dextrose in water at 4 ml/hr.

RESULTS

Prourokinase and urokinase were found to be equivalent with respect to thrombolytic effect. Percent lysis was maximal at 90 minutes in both the urokinase and prourokinase groups. Prourokinase and urokinase were found to be equally effective in restoring flow through thrombosed graft segments.

CONCLUSION

Prourokinase appears to offer little benefit over urokinase with respect to thrombolytic activity in an in vitro model that closely resembles the clinical setting. If prourokinase is to be accepted as an alternative to urokinase, advantages must relate to differences in fibrin specificity.

Preparation and characterization of a disulfide-linked bioconjugate of annexin V with the B-chain of urokinase: an improved fibrinolytic agent targeted to phospholipid-containing thrombi

A conjugate of annexin V and the B-chain of urokinase was prepared and its fibrinolytic properties were studied. First, a mutant of annexin V was constructed with an N-terminal extension of six amino acids (Met-Ala-Cys-Asp-His-Ser) and with Cys316 mutated to Ser; this molecule was expressed in Escherichia coli. The urokinase B-chain was prepared by limited reduction of the interchain disulfide bond between the A- and B-chains of urokinase. These two molecules were then then connected by a disulfide bond and purified to yield a 1:1 stoichiometric conjugate. The conjugate had the same catalytic activity as urokinase against a synthetic substrate, Glt-Gly-Arg-MCA, and a similar plasminogen activating activity. The conjugate showed the same binding affinity for phosphatidylserine-containing membranes as annexin V. The in vitro fibrinolytic activity of the conjugates on clots prepared from platelet-rich plasma was comparable to that of urokinase. However, the conjugate showed 3-4-fold stronger in vivo thrombolytic activity than urokinase in a rat pulmonary embolism model, while having essentially the same plasma clearance rate as urokinase or B-chain. These results show that annexin V is a useful agent for targeting plasminogen activators to phospholipid-containing thrombi.

Thrombolytic profiles of clot-targeted plasminogen activators. Parameters determining potency and initial and maximal rates

BACKGROUND

Targeting of plasminogen activators to the thrombus by means of fibrin-specific monoclonal antibodies may enhance their thrombolytic potency. The kinetics of clot binding of two human fibrin-specific monoclonal antibodies (MA-12B3 and MA-15C5) and of clot lysis with their chemical 1:1 stoichiometric complexes with recombinant single-chain urokinase-type plasminogen activator (rscu-PA) (rscu-PA/MA-12B3 and rscu-PA/MA-15C5) were determined in hamsters and rabbits. Thrombolytic potencies, maximal rates of clot lysis, and the duration of the lag phases before clot lysis of the antibody/rscu-PA conjugates were compared with those of rscu-PA and tissue-type plasminogen activator (rt-PA).

METHODS AND RESULTS

Bolus injection of 7.5 micrograms of 125I-labeled antibody in rabbits with an extracorporeal arteriovenous loop containing a 0.3-mL human plasma clot produced clot-to-blood ratios of 6.6 +/- 1.0 (mean +/- SEM) for MA-12B3 and 1.1 +/- 0.15 for MA-15C5 (p < 0.001 versus MA-12B3) within 6 hours. Progressive digestion of the clot did not alter the binding of MA-12B3 but resulted in as much as a 10-fold increase of the binding of MA-15C5. The conjugates infused intravenously over 90 minutes in hamsters with a human plasma clot in the pulmonary artery produced dose-related in vivo clot lysis. Thrombolytic potencies (maximal slope of the percent lysis versus dose in milligrams of u-PA equivalent per kilogram body weight) were 2,500 +/- 440 for rscu-PA/MA-12B3, 3,600 +/- 640 for rscu-PA/MA-15C5 (p = NS vs. rscu-PA/MA-12B3), 60 +/- 8 for rscu-PA (p < 0.001 versus both conjugates), and 380 +/- 66 for rt-PA (p < 0.001 versus both conjugates). The plasma clearances of the conjugates were fourfold to sixfold slower than those of rscu-PA and rt-PA. Maximal rates of clot lysis, determined by continuous external radioisotope scanning over the thorax, were 0.90 +/- 0.13%, 0.91 +/- 0.17%, 0.84 +/- 0.12%, and 1.1 +/- 0.16% lysis per minute for rscu-PA/MA-12B3, rscu-PA/MA-15C5, rscu-PA, and rt-PA, respectively; these maximal rates were obtained with 0.016, 0.016, 1.0, and 0.25 mg/kg, respectively, and were associated with minimal lag phases of 18 +/- 3.2, 28 +/- 4.9, 34 +/- 3.7, and 25 +/- 3.9 minutes, respectively.

CONCLUSIONS

The thrombolytic potency of the rscu-PA/antifibrin conjugates is determined by their clearance, as well as by rate and extent of initial binding to clots and by changes in binding during clot lysis. Clot targeting of rscu-PA with fibrin-specific antibodies increases its thrombolytic potency but does not alter the maximal rate or the minimal lag phase of clot lysis. These parameters appear to be independent of the nature of the plasminogen activator and of targeting.

Biochemical characterization of single-chain chimeric plasminogen activators consisting of a single-chain Fv fragment of a fibrin-specific antibody and single-chain urokinase

K12G0S32 is a 57-kDa recombinant single-chain chimeric plasminogen activator consisting of scFv-K12Go, a single-chain variable-region antigen-binding fragment (Fv) of the monoclonal antibody MA-15C5, which is specific for fragment D-dimer of human cross-linked fibrin, and a low-molecular-mass (33 kDa) urokinase-type plasminogen activator (u-PA-33k) containing amino acids Ala132-Leu411 (Holvoet, P., Laroche, Y., Lijnen, H. R., Van Cauwenberghe, R., Demarsin, E., Brouwers, E., Matthyssens, G. & Collen D. (1991) J. Biol. Chem. 266, 19717-19724). In addition, the Arg156-Phe157 thrombin-cleavage site in the u-PA moiety of K12G0S32 is removed by substitution of Phe157 with Asp. In the present study, the fibrinolytic potency of K12G0S32, determined in a system composed of a 125I-fibrin-labeled human plasma clot submerged in citrated plasma, was found to be only twofold higher than that of intact single-chain u-Pa (rscu-PA), but 17-fold higher than that of rscu-PA(M), a variant of rscu-PA in which the thrombin-cleavage site was removed by substitution of Phe157 with Asp. The fibrinolytic potency of K12G0S32T, with an intact thrombin-cleavage site, was 6-15-fold higher than that of rscu-PA. Conversion of 1 microM single-chain K12G0S32 or rscu-PA(M) into their two-chain derivatives with plasmin occurred at a rate of 1.0 +/- 0.15 nmol.min-1.nmol plasmin-1 and 0.85 +/- 0.074 nmol.min-1.nmol plasmin-1, compared to 14 +/- 2.3 nmol.min-1.nmol plasmin-1 and 18 +/- 2.6 nM.min-1.nmol plasmin-1 for K12G0S32T and rscu-PA, respectively. Purified fragment D-dimer of human cross-linked fibrin inhibited the fibrinolytic potency of single-chain K12G0S32T, but not of two-chain K12G0S32T, in a dose-dependent manner. Furthermore, the fibrinolytic potencies of two-chain K12G0S32 and K12G0S32T were not significantly higher than those of recombinant two-chain u-PA (rtcu-PA) or of rtcu-PA(M). These findings suggest that the 59-fold increase in fibrinolytic potency of K12G0S32T, relative to that of rscu-PA(M), is due both to targeting of the activator to the clot via the single-chain Fv fragment (sixfold increase) and to a more efficient conversion of single-chain K12G0S32T to its two-chain derivative (eightfold increase). Thus, targeting to clots by means of fibrin-specific antibodies results in a significant increase of the fibrinolytic potency of single-chain but not of two-chain u-PA.(ABSTRACT TRUNCATED AT 400 WORDS)

Remaining perspectives of mutant and chimeric plasminogen activators

Potential approaches to improve thrombolytic agents comprise the construction of mutants and variants of tissue-type plasminogen activator (tPA) or of single chain urokinase-type plasminogen activator (scuPA, pro-urokinase), of chimeric plasminogen activators and of conjugates of plasminogen activators with monoclonal antibodies. tPA mutants have been constructed with altered pharmacokinetic properties or altered functional properties, including binding to and stimulation by fibrin, resistance to plasmin and to protease inhibitors. Mutants of tPA described to date, obtained by deletion/substitution of functional domains or of single amino acids, have markedly reduced clearances, but usually also reduced specific thrombolytic potencies. Mutants of scuPA with improved thrombolytic potencies have thus far not been reported. Chimeric molecules containing functional domains of both tPA and scuPA have intact enzymatic properties of uPA and some fibrin affinity of tPA. Surprisingly, chimeras endowed with fibrin affinity usually have unaltered or reduced thrombolytic potencies. However, a chimera consisting of amino acids 87-274 of tPA and amino acids 138-411 of scuPA, with negligible fibrin affinity, has a 10-fold higher thrombolytic potency than scuPA in animal models of venous thrombosis, as a result of a delayed in vivo clearance and a relatively maintained specific thrombolytic activity. Plasminogen activators conjugated with antifibrin or antiplatelet monoclonal antibodies, either chemically or by recombinant DNA technology, are targeted to blood clots, resulting in a 5- to 10-fold increased thrombolytic potency. Thus, it is possible to develop plasminogen activators with improved thrombolytic potency. Whether such agents will be clinically useful remains to be established.

Characterization of a recombinant chimeric plasminogen activator composed of a fibrin fragment-D-dimer-specific humanized monoclonal antibody and a truncated single-chain urokinase

A recombinant chimeric plasminogen activator, MA-15C5Hu/scu-PA-32k, composed of a humanized fibrin fragment-D-dimer-specific monoclonal antibody (MA-15C5Hu) and a recombinant low-molecular-mass single-chain urokinase-type plasminogen activator, comprising amino acids Leu144-Leu411 (scu-PA-32k), was produced by cotransfecting Chinese hamster ovary (CHO) cells with the cDNA encoding the MA-15C5Hu light-chain sequence and the cDNA encoding the MA-15C5Hu heavy-chain sequence fused with the cDNA encoding scu-PA-32k. Purified MA-15C5Hu/scu-PA-32k migrated as a 215-kDa band on non-reducing SDS/PAGE, which is consistent with a molecule composed of one antibody and two scu-PA-32k moieties. However, the chimera was obtained as a mixture of single-chain u-PA-32k (37%) and amidolytically inactive (50%) and active (13%) two-chain u-PA-32k, the latter of which was removed by immunoadsorption on a monoclonal antibody specific for two-chain urokinase. The fragment-D-dimer affinity and enzymatic properties of MA-15CHu/scu-PA-32k were similar to those of MA-15C5Hu or of scu-PA-32k. In an in vitro system composed of a 125I-fibrin-labeled human plasma clot submerged in citrated human plasma, MA-15C5Hu/scu-PA-32k had a 12-fold higher fibrinolytic potency than scu-PA-32k: 50% lysis in 2 h required 0.43 +/- 0.12 micrograms u-PA-32k equivalent of the chimera/ml versus 5.4 +/- 0.3 micrograms/ml of scu-PA-32k (mean +/- SEM, n = 4). Addition of purified fibrin fragment-D dimer reduced the fibrinolytic potency of MA-15C5Hu/scu-PA-32k in a concentration-dependent way, indicating that the increased potency is the result of antibody targeting. Thus, a recombinant humanized antifibrin antibody/u-PA chimera has been obtained in which only the variable domains of the antibody moiety are of non-human origin. The chimera has intact antigen-binding capacity, u-PA enzymatic activity and a significantly increased fibrinolytic potency in a plasma medium in vitro.

Advances in thrombolytic therapy

Alteplase and saruplase are more fibrin-specific thrombolytic drugs than anistreplase. These and the thrombolytic drugs of the first generation (streptokinase and urokinase) have shortcomings and limitations. The prolonged intravenous maintenance infusions have been replaced by a bolus injection, accelerated infusions, or the combined intravenous administration of thrombolytic agents. Numerous truncated alteplase or saruplase molecules have been constructed by deletion and domain substitution or hybrids made of the two molecules without gaining in thrombolytic potency. Recombinant staphylokinase and plasminogen activator from bat saliva have some interesting properties and are being investigated. Thrombus-targeted thrombolytic drugs were constructed using monoclonal antibodies against fibrin fragments or against epitopes of activated platelets. Fibrin-specific thrombolytic drugs require the concomitant use of a potent antithrombotic drug to prevent reocclusion. Whether hirudin or synthetic thrombin inhibitors are superior to heparin and whether novel antiplatelet agents, including monoclonal antibodies to platelet receptors and disintegrins, are more effective than aspirin is under clinical investigation. The place of stable analogues of prostacyclin during thrombolytic treatment is still unsettled.

Designing thrombolytic agents: focus on safety and efficacy

Thrombolytic therapy for evolving acute myocardial infarction (AMI) reduces infarct size, preserves ventricular function, and reduces mortality. Intravenous streptokinase is commonly followed by approximately 50% patency of coronary arteries within 90 minutes and by reduction of mortality by 25%. Recombinant tissue plasminogen activator (rt-PA) is more potent for coronary arterial thrombolysis, producing both more rapid and more frequent recanalization (approximately 75% patency at 90 minutes) with a dose of 100 mg given over 3 hours. Side effects (mainly bleeding) associated with the use of streptokinase and rt-PA are not markedly different. That the higher efficacy of rt-PA would translate into a larger reduction of mortality is suggested by the results of several small trials but remains to be confirmed in well-designed comparative clinical trials. This question has not been adequately answered by the recent International rt-PA/streptokinase mortality trial and the International Study on Infarct Survival (ISIS-3) study, because of concerns with respect to the role of conjunctive intravenous heparin administration and the dose of rt-PA used in ISIS-3. All available thrombolytic agents still have significant shortcomings, including the need for large doses to be maximally efficient, a limited fibrin specificity, and a significant associated bleeding tendency. New developments toward improved efficacy and fibrin-specificity of thrombolytic agents include the use of mutants of rt-PA, chimeric rt-PA or single chain urokinase plasminogen activator molecules, and antibody-targeted thrombolytic agents. Some of these artificial plasminogen activators have a 5- to 10-fold increased potency (thrombolytic activity per unit dose), but whether they are safe enough to be clinically useful remains to be established. The conjunctive use of anticoagulants and antiplatelet agents with thrombolytic agents increases their efficacy to an extent that monotherapy with a plasminogen activator alone is no longer tenable. Heparin and aspirin are only moderately efficient for acceleration of lysis and prevention of reocclusion, but are relatively safe. More selective thrombin inhibitors and antiplatelet agents are more potent, but their safety remains to be confirmed. Continued investigation in this area will provide new insights and promote progress toward the development of the ideal thrombolytic therapy, characterized by maximized stable coronary arterial thrombolysis with minimal bleeding.

A recombinant chimeric plasminogen activator with high affinity for fibrin has increased thrombolytic potency in vitro and in vivo

A recombinant plasminogen activator with high fibrin affinity and specificity was expressed by transfecting hybridoma cells with a plasmid that combines sequence coding for low molecular mass (32 kDa) single-chain urokinase-type plasminogen activator [scuPA(32kDa)] and anti-fibrin monoclonal antibody 59D8. The expression of the recombinant molecule [r-scuPA(32kDa)-59D8] was optimized by replacing the 3' untranslated region (initially that of high molecular mass scuPA) in the plasmid with the 3' untranslated region of either beta-globin or mouse immunoglobulin. This modification resulted in a greater than 100-fold improvement in the level of protein expression. The 103-kDa r-scuPA(32kDa)-59D8 protein displayed catalytic activity indistinguishable from that of high molecular mass scuPA and fibrin binding comparable to that of native antibody 59D8. r-scuPA(32kDa)-59D8 was 6 times more potent than high molecular mass scuPA in lysing a human plasma clot in vitro and was 20 times more potent than high molecular mass scuPA in the rabbit jugular vein model of thrombolysis. Molecules of this type may serve as prototypes for highly specific, antibody-targeted enzymes suitable for human use.

Thrombolytic and pharmacokinetic properties of chimeric tissue-type and urokinase-type plasminogen activators

BACKGROUND

Chimeric molecules comprising the A-chain of tissue-type plasminogen activator (t-PA) and the catalytic domain of urokinase-type plasminogen activator (u-PA) have intact enzymatic characteristics of u-PA, partial fibrin-binding properties of t-PA, and thrombolytic properties in animal models comparable with but not superior to those of single-chain u-PA (scu-PA). Deletion of the finger and growth factor domains (t-PA-delta FE/scu-PA-e) in such chimeras further reduces their affinity for fibrin.

METHODS AND RESULTS

A detailed investigation of the thrombolytic potency and the pharmacokinetics of t-PA and u-PA chimeras was performed in quantitative animal models for thrombolysis. In hamsters with pulmonary embolism, in rabbits with jugular vein thrombosis, and in baboons with femoral vein thrombosis, the thrombolytic potency (percent lysis per milligram of compound administered per kilogram of body weight) of t-PA-delta FE/scu-PA-e was significantly higher than that of recombinant scu-PA (rscu-PA, Saruplase) as shown by a maximal rate of 720 +/- 170% versus 45 +/- 5% lysis per milligram of compound per kilogram of body weight (mean +/- SEM, p less than 0.01) in hamsters, 210 +/- 18% versus 49 +/- 3% lysis per milligram of compound per kilogram of body weight (mean +/- SEM, p less than 0.01) in rabbits, and 310 +/- 73% versus 90 +/- 0.3% lysis per milligram of compound per kilogram of body weight (p less than 0.01) in baboons. However, the specific thrombolytic activity (percent lysis per microgram per milliliter steady-state plasma antigen level) of t-PA-delta FE/scu-PA-e was not significantly different from that of rscu-PA in hamsters (210 +/- 57% versus 160 +/- 27% lysis per microgram per milliliter antigen level) and was lower than that of rscu-PA in rabbits (37 +/- 4% versus 130 +/- 5% lysis per microgram per milliliter antigen level; p less than 0.01). In dogs with a combined femoral vein blood clot and a platelet-rich femoral arterial eversion graft thrombosis, 0.25 mg/kg body wt bolus injections of t-PA-delta FE/scu-PA-e produced significantly more venous clot lysis (90 +/- 5%, n = 10) than 0.25 mg/kg rscu-PA (26 +/- 3%, n = 10) (p less than 0.001) and, at the arterial side, more frequent (10 of 10 dogs versus three of 10 dogs) and more persistent (six of 10 dogs versus none of 10 dogs) recanalization (p = 0.002). After bolus injection in hamsters, rabbits, or baboons, t-PA-delta FE/scu-PA-e had a fourfold to sixfold longer initial half-life than rscu-PA and a slower plasma clearance of sixfold in hamsters, 10-fold in rabbits, and more than 10-fold in baboons.

CONCLUSIONS

These results indicate that t-PA-delta FE/scu-PA-e has a markedly enhanced thrombolytic potency toward venous and arterial thrombi caused by a delayed in vivo clearance with relatively maintained specific thrombolytic activity. These properties suggest that the chimera may be clinically useful for thrombolytic therapy by bolus administration in patients with thromboembolic disease.

Construction and characterization of a functional chimeric murine-human antibody directed against human fibrin fragment-D dimer

Fibrin-directed monoclonal antibodies may be clinically useful for in vitro thrombus imaging and for the targeting of fibrinolytic agents to blood clots. One such murine monoclonal antibody, (mAb-15C5), raised against the fragment-D dimer epitope of cross-linked human fibrin, was previously characterized [Holvoet, P., Stassen, J. M., Hashimoto, Y., Spriggs, D., Devos, P. & Collen, D. (1989) Thromb. Haemostasis 61, 307-313] has recently been cloned and expressed [Vandamme, A.-M., Bulens, F., Bernar, H., Nelles, L., Lijnen, H. R. & Collen, D. (1990) Eur. J. Biochem. 192, 767-775]. In order to reduce the immunogenicity of the murine mAb-15C5 in man, we have now constructed a murine--human chimera of mAb-15C5, by substituting the cDNA sequences encoding the constant regions of the murine kappa light chain and gamma 1 heavy chain by the corresponding human genomic sequences. Both chimeric murine--human Ig chains were cloned into two separately selectable expression vectors, which were contransfected into Chinese hamster ovary (CHO) cells. Murine--human chimeric mAb-15C5 (mAb-15C5Hu) was purified from the conditioned medium of selected cell lines by chromatography on Zn-chelating Sepharose, protein-A-Sepharose and on insolubilized antigen (fragment-D dimer), with a final yield of 29 micrograms/l and a recovery of 33%. SDS/PAGE without reduction revealed a homogeneous band with a mobility similar to that of natural mAb-15C5, whereas after reduction, both the heavy and the light chains had slightly slower mobilities than their natural counterparts. Expression in the presence of tunicamycin suggested that the differences in gamma 1-chain mobility were due to different N-glycosylation patterns. Immunoblotting of proteins from SDS gels showed immunological reactivity of recombinant mAb-15C5Hu with goat anti-(human IgG) IgG and of recombinant and natural murine mAb-15C5 with goat anti-(mouse IgG) IgG. Competitive binding revealed a comparable affinity of recombinant murine mAb-15C5, recombinant mAb-15C5Hu and natural mAb-15C5, for fragment-D dimer, indicating that recombinant mAb-15C5Hu was obtained in a functionally intact form. Thus, mAb-15C5Hu may constitute a useful alternative to mAb-15C5 for in vivo use in man.

Thrombolytic and pharmacokinetic properties of a conjugate of recombinant single-chain urokinase-type plasminogen activator with a monoclonal antibody specific for cross-linked fibrin in a baboon venous thrombosis model

Chemical conjugates between recombinant single-chain urokinase-type plasminogen activator (rscu-PA) and a murine monoclonal antibody directed against fragment D-dimer of cross-linked human fibrin (MA-15C5), rscu-PA/MA-15C5, and between rscu-PA and a control monoclonal antibody (MA-1C8), rscu-PA/MA-1C8, were produced by cross-linking with N-succinimidyl 3-(2-pyridyldithio) propionate (SPDP). In an in vitro system composed of a [125 I]fibrin-labeled baboon plasma clot immersed in autologous citrated plasma, dose- and time-dependent lysis was obtained with a ratio of the potencies of free and conjugated rscu-PA similar to that in human plasma: 50% lysis in 2 hours required 4.3 micrograms/ml rscu-PA, 1.0 microgram/ml urokinase-type plasminogen activator (u-PA) equivalent rscu-PA/MA-15C5, or 15 micrograms/ml u-PA equivalent rscu-PA/MA-1C8. The thrombolytic and pharmacokinetic properties of rscu-PA and of rscu-PA/MA-15C5 were compared in baboons with a 0.8-1.0 ml [125 I]fibrin-labeled autologous blood clot produced in a femoral vein. Continuous intravenous infusion of these compounds during a 2-hour period resulted in dose- and time-dependent lysis. The thrombolytic potency of rscu-PA/MA-15C5 was 3.0 +/- 0.5 times higher (50% lysis with 0.3 +/- 0.02 mg u-PA equivalent/kg body wt) than that of rscu-PA measured by ex vivo isotope recovery from the femoral vein segment (p less than 0.001) and was 2.7 +/- 0.5 times higher (50% lysis with 0.35 +/- 0.02 mg/kg rscu-PA/MA-15C5) by external radioisotope counting (p less than 0.001). A dose of 0.5 mg/kg of rscu-PA/MA-1C8 was much less active than rscu-PA. After the end of the infusion, u-PA-related antigen disappeared from plasma in a biphasic manner with an initial half-time of 2.7 +/- 0.5 for rscu-PA, 24 +/- 1.2 for rscu-PA/MA-15C5, and 21 +/- 0.5 minutes for rscu-PA/MA-1C8 with corresponding plasma clearances of 340 +/- 40, 20 +/- 3, and 24 +/- 2 ml/min, respectively. In conclusion, the increased thrombolytic potency of rscu-PA/MA-15C5 is the result of a reduction of the thrombolytic potency due to coupling of rscu-PA to the antibody molecule, which is counter-balanced by an enhancement of the thrombolytic potency due to fibrin targeting by the specific idiotype.

Construction and characterization of a recombinant murine monoclonal antibody directed against human fibrin fragment-D dimer

cDNA libraries in lambda phage were generated from the murine hybridoma secreting mAb-15C5, a monoclonal antibody directed against fragment-D dimer of crosslinked human fibrin [Holvoet et al. (1989) Thromb. Haemostasis 61, 307-313], and clones encoding fragments of the heavy (gamma 1) and the light (kappa) chain were isolated. The kappa-chain cDNA was reconstructed from two overlapping clones encoding 20 amino acids of signal sequence and the 214 amino acids of the mature protein chain. The gamma 1-chain cDNA was reconstructed from the mAb-15C5 kappa-chain signal sequence, the mAb-15C5 gamma 1 variable-domain coding sequence and murine gamma 1-gene and gamma 1-chain cDNA fragments encoding the constant domains. These cDNAs were expressed in Chinese hamster ovary cells, selected cell lines were scaled up in roller bottle culture, and recombinant mAb-15C5 was purified from the conditioned medium by chromatography on Zn-chelate - Sepharose, protein-A - Sepharose and insolubilized fragment-D dimer, with a yield of 50 micrograms/l and a recovery of 20%. SDS-gel electrophoresis without reduction revealed a homogeneous band, and after reduction a light-chain band with identical and a heavy-chained band with a somewhat slower mobility than that of the natural mAb-15C5. Competitive binding revealed a comparable affinity of natural and recombinant mAb-15C5 for fibrin fragment-D dimer. Thus recombinant mAb-15C5, obtained by co-expression of the reconstructed cDNAs of the kappa and gamma 1 chain in Chinese hamster ovary cells, has very similar properties to natural mAb-15C5. These recombinant mAb-15C5 cDNAs may be useful for the construction of a humanized monoclonal antibody for thrombus imaging, and for targeting of thrombolytic agents to fibrin.

A bispecific antibody enhances the fibrinolytic potency of single-chain urokinase

A monoclonal antibody specific for an epitope at the amino terminus of the beta chain of fibrin and a monoclonal antibody that binds both one- and two-chain high molecular weight urokinase were chemically cross-linked [using N-succinimidyl 3-(2-pyridyldithio)propionate and 2-iminothiolane]. The chemically modified material was heterogeneous, ranging in molecular size from tetramers to monomers and containing the two antibodies in various ratios. Nevertheless, fractions of a molecular size larger than a monomer were capable of binding fibrin and urokinase simultaneously in a radioimmunoassay. These fractions also enhanced fibrinolysis by high molecular weight single-chain urokinase (scuPA) by 50-fold and plasma clot lysis by 5-fold. Whereas scuPA significantly decreased the concentration of fibrinogen in plasma clot assay supernatants, scuPA in association with the bispecific antibody did not.

Effect of fibrin-targeting on clot lysis with urokinase-type plasminogen activator

The effect of fibrin-targeting of urokinase-type plasminogen activator (u-PA) on its fibrinolytic potency was studied using recombinant fusion proteins of u-PA with the NH2-terminal region of tissue-type plasminogen activator (t-PA/u-PA) and chemical complexes of u-PA with F(ab')2 fragments of a fibrin specific monoclonal antibody (u-PA/MA-15C5-F(ab')2). Two chain derivatives of a low Mr variant of u-PA comprising amino acids Leu144-Leu411 (tcu-PA-32k), obtained by cleavage of recombinant single-chain u-PA (rscu-PA-32k) with thrombin (rtcu-PA-32k/T) or plasmin (rtcu-PA-32k/P) were investigated. The plasmin-derived two chain u-PA moieties, rtcu-PA-32k/P, rt-PA/tcu-PA-32k/P and rtcu-PA-32k/MA-15C5-F(ab')2/P had high specific activities in amidolytic and fibrin plate assays (130,000 and 150,000 IU/mg u-PA, 43,000 and 71,000 IU/mg u-PA and 32,000 and 56,000 IU/mg u-PA respectively). The thrombin-derived two chain u-PA moieties had a very low amidolytic activity, corresponding to less than or equal to 1 percent of that of their plasmin-derived counterparts. On fibrin plates, however, rtcu-PA-32k/T had a negligible activity, whereas rt-PA/tcu-PA-32k/T and rtcu-PA-32k/MA-15C5-F(ab')2/T had specific activities of 12,000 and 25,000 IU/mg u-PA respectively. The catalytic efficiency for plasminogen activation of rtcu-PA-32k/MA-15C5-F(ab')2/T is 4,000-fold lower than that of rtcu-PA-32k/MA-15C5-F(ab')2/P, but its concentration required for 50 percent lysis in 2 hours of a 125I-fibrin labeled plasma clot in human plasma (C50) is only 25-fold higher. The catalytic efficiency of rt-PA/tcu-PA-32k/T is 1,600-fold lower and the C50 100-fold higher than that of rt-PA/tcu-PA-32k/P. The catalytic efficiency and the fibrinolytic potential of rtcu-PA-32k/T are negligible as compared to that of rtcu-PA-32k/P. These observations may be explained by conversion of the thrombin derived two chain u-PA moieties to their plasmin-derived analogues at the fibrin surface. This conversion appears to be most efficient for the antibody conjugate which has a high fibrin-affinity, less efficient for the t-PA/u-PA chimera which has only moderate fibrin-affinity, and negligible for the unconjugated u-PA moiety which has no fibrin-affinity. These findings illustrate the importance of plasmin-mediated positive feedback mechanisms in u-PA mediated clot lysis.