CA2333592A1 - Methods of inhibiting clot formation - Google Patents

Abstract

The invention provides methods of detecting blood clots in vivo and methods of inhibiting clot formation and promoting therapeutic thrombolysis using alpha-2 antiplasmin (.alpha.2AP) polypeptides.

Description

2 PCT/US99/11843 METHODS OF INHIBITING CLOT FORMATION
Statement as to Federally Sponsored Research S This invention was made with government support under National Institutes of Health grant HL-57314. The U.S. government has certain rights in the invention.
Backaround of the Invention This invention relates to inhibition and detection of blood clot formation.
The resistance of thrombi to fibrinolysis induced by plasminogen activators is an impediment to the successful treatment of thrombotic diseases. Fibrinolytic resistance is evident in patients with acute thrombotic coronary occlusion, where treatment 1.5 with plasminogen activators results in full coronary reperfusion in only 33-55% of patients at 90 minutes (Lincoff et al., 1995, Am. J.
Cardiol. 75:871-876; Karagounis et al., 1992, J. Am. Coll. Cardiol.
19:1-10). The resistance of thrombi to lysis by plasminogen activators may be even more marked in patients with venous s!0 thromboembolism. In deep venous thrombosis treated with tissue plasminogen activator (TPA), nearly two thirds of patients have minimal or no significant lysis evident on repeat venography at 24 hours (Salzman et al., 1994, The epidemiology, pathogenesis and natural history of venous thrombosis, in Hemostasis and Thrombosis:
25 Basic Principles and Clinical Practice, 3rd ed., Philadelphia, PA:
Lippincott, pp 1275-1296; Goldhaber et al., 1990, Am. J. Med. 88:235-240). In patients with pulmonary embolism, TPA restores blood flow within 24 hours to only about a third of occluded lung segments, as judged by serial perfusion scanning (The Urokinase Pulmonary Embolism 30 Trial. A national cooperative study, 1973, Circulation 47:1-108;
Goldhaber et al., 1986, Lancet 2:886-889). Improved thrombolysis may reduce mortality and morbidity associated with thrombotic disease _Summarv of the Invention The invention provides methods of improving therapeutic :35 thrombolysis, detecting blood clots in vivo, and inhibiting clot formation using alpha-2 antiplasmin (a2AP) palypeptides.
To detect blood clot formation in a mammal, a diagnostically effective amount of a detestably labeled alpha-2 antiplasmin (a2AP) polypeptide is administered to the mammal, acrd association of the 40 polypeptide with a blood clot determined. Association of the a2AP
polypeptide with a vascular obstruction, e.g, via a2AP-fibrin crosslinking, is an indication of the presence blood clot formation at the site of the obstruction. Since a2AP croselinks with fibrin in actively forming blood clots but not at the site of old (i.e., not actively forming) blood clots, the method is useful to characterize blood vessel obstructions as pre-existing or actively forming. Newly forming thrombi are detected by virtue of the formation of a2AP
S polypeptide-fibrin crosslinks mediated by activated factor XIIIa.
Thus, the present method provides an advantage over conventional visual detection techniques such as scintiphotography and angiography with which such characterization is difficult or unachievable.
The polypeptides of the invention are substantially pure.
Polypeptides or other compounds encompassed by the invention are said to be "substantially pure" when they are within preparations that are at least 60% by weight (dry weight) the compound of interest.
Preferably, the preparation is at least 75%, more preferably at least 90%, and most preferably at least 99%, by weight the compound of interest. Purity can be measured by any appropriate standard method, for example, by column chromatography, polyacrylamide gel electrophoresis, or HPLC analysis.
The a2AP polypeptide preferably contains the amino acid sequence of X1QX~X3X,XSPLX6LLK (SEQ ID NO:l) , wherein X, = N or A, Xs E or Q, X3 = Q or K, X, = V or L, XS = P or S, and X6 = T, S or A (the Q residue indicated in bold type is involved if a2AP-fibrin crosslinking). For example, the polypeptide comprises the amino acid sequence of a2APl,_za NQEQVSPLTLLK (SEQ ID N0:2) ar a2APl_"
(MEPLGWQLTSGPNQEQVSPLTLLK; SEQ ID N0:16). Polypeptides derived from 2S human a2AP include those which contain a sequence that is 80-100%
identical to the amino acid sequence of MEPLGXQLTS GPNQEQVSPL
TLLKLGNQEP GGQTALKSPP GVCSRDPTPE QTHRLARAI~ AFTADLFSLV AQT (SEQ ID
N0:3), where "X" represents a residue that can differ among human a2AP variants. Human N-terminal a2AP polypeptides include those which contain the amino acid sequence MEPLGWQLTS GPNQEQVSPL
TLLKLGNQEP GGQTALKSPP GVCSRDPTPE QTHRLARAN~I AFTADLFSLV AQT (SEQ ID
N0:4) or MEPLGRQLTS GPNQEQVSPL TLLKLGNQEP GGQTALKSPP GVCSRDPTPE
QTHRLARAN~'1 AFTADLFSLV AQT (SEQ ID N0:5). Alternatively, the polypeptide may contain a sequence that is 80-100% identical to the 3S amino acid MEPLDLQLNm GQAQQKLPPL SLLKLDNQEP GGQIAPKKAP EDCKLSPTPE
QTRRLARADM'I TFTTDLFSLV AQS (SEQ ID N0:6), corresponding to an N-terminal fragment of naturally-occurring bovine a2AP, or VDLPGQQPVS
EQAQQKLPLP ALFKLDNQDF GDHATLKRSP GHCKSVPTAE ETRRLAQAMM AFTTDLFSLV AQT
(SEQ ID N0:7), corresponding to an N-terminal fragment of naturally-occurring mouse a2AP. An a2AP polypeptide is a peptide with at least 80-100% sequence identity to a portion of a naturally-occurring a2AP
protein but having a length that is shorter than the length of the

- 3 -naturally-occurring mature full-length a2AP protein. Human a2AP
variants within the invention include those with the amino acid sequence of SEQ ID NO:11, 13, 15, or 17.
The invention also features methods of preventing the S development of clots in patients at risk for thrombosis and methods of treating patients with thrombotic conditions such as stroke, myocardial infarction, pulmonary embolism, and deep venous thrombosis. For example, a method of inhibiting blood clot formation in a mammal is carried out by administering to the mammal a therapeutically effective amount of an a2AP polypeptide. A method of preventing and lysing blood clots is carried out by co-administering to a mammal a therapeutically effective amount of an a2AP polypeptide and a thrombolytic agent such as a plasminogen activator, e.g., tissue plasminogen activator (t-PA). Thrombolytic agents such as 1S prourokinase, urokinase, streptokinase, staphylokinase, and vampire bat-derived plasminogen activator may be co-administered with an a2AP
polypeptide to increase the effectiveness of the thrombolytic agent.
a2AP polypeptides and peptide mimetics thereof are useful in the diagnostic and therapeutic methods described above. Preferably the a2AP polypeptide is derived from the N-terminus of a mature, naturally-occurring mammalian a2AP protein. For example, the polypeptides may be derived from human, bovine, or mouse a2AP
proteins, the amino acid sequences of Which are shown in Tables 1-3, respectively (the first amino acid of the mature protein is indicated 2S with an arrow).
TABLE 1: HUMAN a2AP

i81 QLFGAKPVSL TGKQEDDLAN INQWVKEATE GKIQEFLSGL PEDTVLLLLN AIHFQGFWRN

361 DLRGISEQSL WSGVQHQST LELSfiVGVEA AAATSIAMSR MSLSSFSVNR PFLPFIFEDT

481 YPQFGSPK (SEQ ID NO: B)

- 4 -TABLE 2: BOVINE a2AP

4B1 EEDYPQFGSP K (SEQ ID N0:9) TABLE 3: MOUSE a2AP

481 EEDYPQFSSP K ( SEQ ID NO : 1 O ) The length of an a2AP polypeptide is preferably 12 to 250 2S amino acids, inclusive. More preferably, the length is 10 to 75 amino acids, inclusive, e.g., a polypeptide that is 12 amino acids in length and has an amino acid sequence of NQEQVSPLTLLK (SEQ ID N0:2).
The amino acid sequence of the polypeptide preferably contains the amino acid sequence of X,QX,X,X,X5PLX6LLK (SEQ ID NO:1) , 30 wherein X~ = N or A, X~ = E or Q, X3 = Q or K, X, = V or L, X5 = P or S, and X6 =T, S or A. Examples of such polypeptides include those derived from human a2AP (e. g., SEQ ID N0:4 or 5) and those derived from bovine a2AP (e. g., SEQ ID N0:6). An a2AP polypeptide derived from mouse a2AP contains the amino acid sequence of SEQ ID N0:7.
3S Shorter polypeptides such as those containing amino acids 1-24, 1-41, 10-24, and 13-24 (e.g., SEQ ID N0:2) of each of SEQ ID N0:4, 5, 6, or 7 are also useful to detect and treat thrombotic conditions.
Preferably, the a2AP polypeptide contains an amino acid sequence with 80-100 sequence identity to SEQ ID N0: 3, 4, 5, 6, or 40 7 in which a non-identical amino acid of the polypeptide is a conservative amino acid substitution and the polypeptide functions to inhibit a2AP-fibrin crosslinking. Sequence identity is measured using standard sequence analysis software (e. g., the Sequence Analysis Software Package of the Genetics Computer Group, University

- 5 -of Wisconsin Biotechnology Center, 1710 University Avenue, Madison, WI 53705), with the default parameters as specified therein.
Peptide mimetics of an a2AP polypeptide are also within the invention. Preferably, the length of the peptide mimetic is 12 to 250 amino acids, inclusive, and contains the amino acid sequence of SEQ ID NO:1. For example, an a2AP peptide mimetic contains an amino acid sequence with 80-100% sequence identity to SEQ ID NO: 2; the non-identical amino acids of the polypeptide are conservative amino acid substitutions; and the polypeptide inhibits a2AP-fibrin crosslinking.
For diagnostic applications, the polypeptide is detectably labeled, e.g., the polypeptide is biotinylated or tagged with a radioisotope. In addition to administering a2AP polypeptides to a mammalwto prevent clot formation or promote physiologic thrombolysis or improving the effectiveness of therapeutically-induced thrombolysis, the polypeptide can be used as targeting agents to deliver therapeutic agents to the site of a actively forming blood clot. For example, any of the a2AP polypeptides or peptide mimetic are linked to a therapeutic agent (e. g., a thrombolytic agent). Such chimeric compounds are recombinantly produced or cosynthesized.
Thus, the invention includes a method of targeting a therapeutic agent to an actively developing thrombus in a mammal by administering to the mammal an N-terminal a2AP polypeptide linked to a therapeutic agent.
Other features and advantages of the invention will be apparent from the following description of the preferred embodiments thereof, and from the claims.

- 6 -Brief Description of the Drawincrs Figs. lA and 1B are bar graphs showing the effects of factor XIIIa-mediated crosslinking on the resistance of pulmonary emboli to endogenous and pharmacologic lysis. Fig. lA shows the effect of factor XIIIa activity on resistance to endogenous fibrinolysis.
Anesthetized ferrets were treated with heparin (100 units/kg), and radiolabeled plasma clots were embolized to the lungs. Factor XIIIa activity was normal in emboli in control animals or inhibited in the F13-I group. "F13-I" is a monoclonal antibody (9C11) which is capable of quenching all factor XIIIa-mediated crosslinking (i.e., both fibrin 'y-chain crosslinking and a2-antiplasmin-fibrin crosslinking are completely inhibited). Lysis was significantly higher in the F13-I group than in the control group (p<0.0001).
Fig. 1B shows the effect of a2 antiplasmin-fibrin crosslinking and total factor XIIIa activity on resistance to pharmacologic lysis of pulmonary emboli in animals treated with TPA (1 mg/kg). Lysis of pulmonary emboli in the presence of TPA (normal factor XIIIa activity), TPA+F13-I (inhibited factor XIIIa activity), and TPA +
a2AP-I (selectively inhibited a2 antiplasmin-fibrin crosslinking) is depicted. "a2AP-I" is a polypeptide with the amino acid sequence of SEQ ID N0:2; this polypeptide selectively inhibits only factor XIIIa-mediated a2AP-fibrin crosslinking. In animals with normal factor XIIIa activity, the addition of TPA significantly increased lysis (control vs TPA; p<.005). Inhibition of a2 antiplasmin-fibrin crosslinking significantly increased lysis (TPA+a2AP-I vs TPA;
p<0.0005). Full inhibition of factor XIIIa activity caused higher lysis than that obtained in both the normal factor XIIIa group (TPA +
F13-I vs TPA; p<.0001), and the group in which a2 antiplasmin-fibrin crosslinking had been inhibited (TPA+F13-I vs TPA+a2AP-I; p<.0005).
Mean lysis (tSD) and number of animals in each experimental group are shown.
Fig. 2 is a bar graph showing the effects of factor XIIIa activity and TPA on residual fibrinogen levels. Fibrinogen levels were measured in 25 ferrets by sodium sulfite precipitation and expressed as a percentage of the fibrinogen level prior to the experiments ("before study" level was defined as 100%). The percent residual fibrinogen level (mean t SD) is shown for each group of 5 ferrets at the end of the experiment. The following conditions were evaluated: Control, normal factor XIIIa activity, no TPA; F13-I, factor XIIIa inhibited, no TPA; TPA. TPA with normal factor XIIIa activity; TPA+F13-I, TPA with inhibited factor XIIIa activity;
TPA+a2AP-I, TPA with inhibited a2 antiplasmin-fibrin crosslinking.

Fig. 3A is a diagram showing the amino acid sequences of the native a2APl_z, amino-terminal peptide and related peptides containing specific amino acid substitutions. The top line shows the sequence of the native a2APl,_z, amino-terminal peptide (SEQ ID N0:2) used for FXIIIa crosslinking studies. The arrow indicates the glutamine residue in SEQ ID N0:2 which is crosslinked by FXIIIa to fibrin.
a2APl_" (SEQ ID N0:16) and a2AP variants (SEQ ID N0:11, 12, 13, 14, and 15) are also shown.
Fig. 3B is a bar graph showing crosslinking of a2APl_z,and the variant a2AP peptides from shown in Fig. 3A to fibrin clots. A
0.5 mM concentration of each peptide was clotted with human fibrinogen. After electrophoresis of the salubilized, reduced clots on 7.5% SDS-PAGE gels and transfer to PVDF membranes, the extent of crosslinking of each peptide was determined by incubating the 1$ membranes with a polyclonal antibody directed against the amino-terminus of a2AP, followed by incubation in 1'SI-protein A, and exposure in a phosphorimager. Each bar reflects phosphorimage quantification of the fibrin crosslinking a2AP1_,4 and the variant peptides. The mean t SD pixel volume of the quantified blots are shown.
Detailed Description a2AP polypeptides inhibit factor XIIIa-mediated crosslinking of endogenous a2AP with fibrin, thereby causing thrombi to undergo spontaneous physiologic lysis or accelerated lysis when administered with therapeutic thrombolytic agents.
Factor XIIIa catalyzes the formation of covalent bonds between glutamine and lysine residues in the y and a chains of adjacent fibrin molecules, which markedly increase the mechanical durability of the fibrin polymer. Factor XIIIa also rapidly crosslinks a2AP, the fast acting plasmin inhibitor, to fibrin. The relative contribution of fibrin-fibrin crosslinking, or of a2 antiplasmin-fibrin crosslinking, to fibrinolytic resistance in vitro is still debated.
In vivo, human thromboemboli show evidence of extensive a2AP-fibrin crosslinking by factor XIIIa, and highly crosslinked thrombi are more resistant to lysis in vitro. The contribution of factor XIIIa-mediated fibrin-fibrin crosslinking and a2 antiplasmin-fibrin crosslinking to the fibrinolytic resistance of experimental pulmonary emboli was examined. The data described herein indicates that factor XIIIa-mediated fibrin-fibrin and a2AP-fibrin crosslinking is the underlying mechanism of resistance of pulmonary emboli to endogenous and TPA-induced fibrinolysis and that a2AP polypeptides _ g _ crosslink with fibrin, thereby inhibiting crosslinking of fibrin with endogenous a2AP protein.
Reaa~ent s Materials were obtained from the following suppliers:
$ aprotinin, Sigma (St. Louis, MO); calcium chloride, Mallinckrodt (Paris, Kentucky); purified factor XIII and fibrinogen, American Diagnostica (Greenwich, CT); goat antimouse antibody, Cappel Organon Technika (Durham, NC); heparin (1000 units/ml), Elkins-Sinn Inc (Cherry Hill, NJ); fresh-frozen human plasma pooled from random donors; TPA with a specific activity of 580,000 IU/mg, Genentech (South San Francisco, California); normal saline for intravenous use, Travenol Laboratories (Deerfield, IL); sodium iodide, Aldrich Chemical Co (Milwaukee, WI); Na125I, Dupont-NEN (Cambridge, MA); and bovine thrombin, Parke-Davis (Morris Plains, NJ). Microcentrifuge 1$ tubes were obtained from National Scientific Supply Co (San Rafael, CA) .
Ferrets (weighing approximately 0.8 to 1 kg) were purchased from Marshall Farms (New York, NY). Ketamine (100 mg/ml) was obtained from Fort Dodge Laboratories (Fort Dodge, Iowa) and acepromazine maleate from Fermenta Animal Health Co. (Kansas City, MO). The surgical instruments were purchased from VWR (Boston) and the tubing from Namic (Glens Falls, NY). Bard Parker surgical blades were from Becton Dickinson (Franklin Lake, NJ), 4.0 silk sutures from American Cyanamid Co (Danbury, CT), Surflo IV catheter and 20 gauge 1 2$ 1/4-inch Venoject tubes with K3EDTA from Terumo Medical Corp (Elkton, MD), and sterile three-way stopcocks from Mallinckrodt Critical Care (Glens Falls, NY). An auto syringe infusion pump (Baxter Health Care Corp, Hooksett, NH) was used with tubing and a microbore 60 inch extension set obtained from McGaw of Puerto Rico (Sabana Grand, Puerto Rico).
Anti Factor XIII Monoclonal Antibody Production and Purification Hybridoma cells which produce monoclonal antibody 9C11 (which is specific for the catalytic A subunit of human factor XIII) is available from the American Type Culture Collection (ATCC
3$ Designation No. CRL 11458). The hybridoma producing 9C11 was cloned by limiting dilution and expanded into ascites in pristane-primed Balb/C mice. Antibody was purified from filtered ascites by precipitation with 40% ammonium sulfate. After resuspension and dialysis into 10 mM KH2P04, pH 7.2, proteins were absorbed on a DEAE-Affigel Blue Sepharose column and monoclonal antibody 9C11 was eluted with a linear gradient spanning 0 to 100 mM NaCl. Eluted protein was _ g _ collected in fractions and analyzed by SDS-polyacrylamide electrophoresis on 10% gels.
Inhibition of Fibrin-Fibrin Crosslinking and a2 Antinlasmin-Fibrin CrosslinkinQ in vitro $ . To determine the dose of factor XIII inhibitor or a2AP
polypeptide required to inhibit factor XIII-mediated crosslinking, various concentrations of anti-factor XIII antibody 9C11 (10 ~C1; 0-10 fig) was mixed with fresh-frozen plasma (45 ~1), bovine thrombin (100 units/ml; 3 ~1), and calcium chloride (0.4 M; 2.5 ~1). After clotting for 90 minutes at 37°C, the clots were compressed and washed three times in 500 ul of saline to remove unbound protein. The clots were solubilized in 95 ~1 of 9 M urea and 5 ~.1 of ~i-mercaptoethanol at 37°C for 30 min. Clots were then mixed in 100 ~.1 of SDS sample buffer with 20 ~.l of bromphenol blue-glycerol solution and incubated 1$ at 85°C for 5 min. Proteins were electrophoresed on 6% SDS-polyacrylamide gels and electroblotted to polyvinylidene membranes for immunoblotting with an antibody specific for the y chain of fibrin and an antibody specific for the carboxy-terminus of a2AP.
To inhibit the crosslinking of a2AP to fibrin, a polypeptide spanning the crosslinking site on the amino texminus of a2AP (SEQ ID
N0:2) was synthesized. The purity of the polypeptide was analyzed by high performance liquid chromatography, and its composition was verified by amino acid analysis on a Waters Picotag system. The a2AP
polypeptide was then solubilized in 20 mM Tris-HC1 and the pH Was 2$ adjusted to 7Ø Various concentrations of polypeptide (7.5 ~.1, 0-5 mM final) were mixed with 20 ul of plasma, 1.25 ~.1 of calcium chloride (0.4 mM), and 1.25 ~1 of thrombin (100 units/ml) and clotted for 90 minutes at 37°C as described above. The clots were solubilized and analyzed by immunoblotting as described above.
Evaluation of Fibrinolvtic Resistance of Pulmonarv Embolisms in vivo Pulmonary embolisms were induced in ferrets using standard methods, e.g., that described in Butte et al., 1997, Circulation.
95:1886-1891). Male ferrets (weighing approximately 1 kg) were anesthetized with ketamine and acepromazine. After full anesthesia 3$ had been obtained, the jugular vein and carotid artery were exposed by an anterior midline incision and cannulated with 20G catheters.
Pooled, citrated human plasma was mixed with "5I-fibrinogen to achieve a specific activity of approximately 1,000,000 cpm/ml.
Individual clots were formed by mixing 1'sI._fibrinogen-labeled plasma (45 ~.1) with 2.5 ~.1 of bovine thrombin (100 units/ml) and 2.5 ~1 of calcium chloride (0.4 M). In some experiments, antibody 9C11 (3.1 ~.1; 10 ~.g) was added to each mixture to inhibit factor XIII activity, or a2AP polypeptide (3 ~1, 1.5 mM final concentration) was added to attenuate a2AP crosslinking. After incubation at 37°C for 90 minutes, the clots Were compressed and washed three times with saline to remove unbound protein. The radioactive content of the clots was $ measured in a gamma counter immediately before injection. Blood samples were drawn at base line and at the end of the experiment.
Sodium iodide (10 mg) was injected to block thyroid uptake. Three clots were embolized into the lungs by injection through the internal jugular vein. Successful embolization was evinced by the accumulation of radioactivity in the thorax.
All animals received weight-adjusted heparin at 100 units/kg (bolus), a dose sufficient to keep the activated partial thromboplastin time (aPTT) above 150 seconds throughout the procedure. TPA was given as a continuous infusion over 2 hours (1 mg/kg in 5 ml of normal saline). Animals were observed for a total of 4 hours after pulmonary embolization and then killed by lethal injection of anesthesia or CO, inhalation. The thorax was dissected and all intrathoracic structures were removed for gamma counting to detect residual thrombi. The percentage of clot lysis was determined for each ferret by dividing the total residual radioactivity in the thorax by that in the initial thrombi. A total of 28 animals was studied; three were excluded because of anesthetic-related death, improper TPA infusion, and failed embolism.
Fibrinogen Assays 2$ Blood samples were collected on K3EDTA (0.15% solution, final concentration) with aprotinin (50 kallikrein units/ml).
Platelet-poor plasma was obtained by centrifugation of whole blood and assayed for fibrinogen by a standard sodium sulfite method such as that described by Rampling et al. 1976, Clin Chim Acta. 67:43-52.
statistical Tests The data were analyzed by a one way analysis of variance followed by the Bonferroni-Dunn procedure f:or testing multiple comparisons.
_Inhibition of Factor XIII Activity Durincr Clottinct The factor XIIIa-mediated crasslinking of the y chains of fibrin ie a rapid process, and, in the presence of fibrin, the activation of factor XIII by thrombin is accelerated. Monoclonal antibody 9C11 fully inhibits all factor XIIIa-mediated crosslinking in primate plasmas. To determine the amount of 9C11 required for inhibiting factor XIII activity in the present studies, dose-related effects of 9C11 on fibrin-fibrin y chain and a2-antiplasmin-fibrin croaslinking during clotting were examined. In comparison with clots formed in the absence of inhibitor or in the presence of the nonspecific alkylating agent, iodoacetamide, clots formed with 9C11 at doses of 5 and 10 ~g/clot showed no significant fibrin-fibrin y chain or a2AP-fibrin crosslinking. A 9C11 dose of 10 ~g was used for the experiments described below.
Selectiye Inhibition of a2AP CrosslinkinQ to Fibrin a2AP polypeptides, e.g., a polypeptide with the amino acid sequence of SEQ ID N0:2, were found to selectively inhibit factor XIIIa-mediated crosslinking of a2AP to fibrin. In comparison with clots formed in the absence of the a2AP peptide, clots formed in the presence of the polypeptide at concentrations of 0.38 mM and higher showed progressively less a2AP-fibrin crosslinking. At peptide concentrations of 1.5 mM and 3.0 mM, little a2AP crosslinking was visible. In contrast, the a2AP polypeptide had no apparent effect on the formation of fibrin 'y-y crosslinks. Because the 1.5 mM
concentration of the polypeptide strongly inhibited a2AP-fibrin crosslinking without preventing formation of the 'y dimer, it was used in studies of the fibrinolytic effects of a2AP crosslinked to fibrin.
_Role of Factor XIII Activity in Endogenous Lvsis To determine the importance of factor XIII in endogenous fibrinolysis (i.e., lysis caused by the ferret's own fibrinolytic system), the rates of dissolution of pulmonary emboli in animals treated with and without factor XIII inhibitor was measured. All animals received heparin at a weight-adjusted bolus of 100 units/kg;
this dose was sufficient to keep the aPTT above 150 seconds throughout the experiment. Lysis of pulmonary emboli in the control group was l4.lt 4.8% (mean~SD). Lysis of pulmonary emboli in the group treated with factor XIII inhibitor was three times as much (42.71 7.4%; p<0.0001). These data indicate that inhibition of factor XIII activity markedly increased endogenous lysis in these pulmonary emboli.
Role of Factor XIII Activity in Pharmacoloaic Lvsis The effect of factor XIII activity on the fibrinolytic resistance of pulmonary emboli in ferrets treated with TPA (1 mg/kg) 3$ was determined. TPA was administered over a time period of 2 hours, a regimen similar to that used to treat pulmonary embolism in humans.
As above, all animals were treated with heparin, the standard therapy for human pulmonary embolism. One experimental group received pulmonary emboli With normal factor XIII levels, another pulmonary emboli in which both factor XIIIa-mediated fibrin-fibrin and a2AP
croselinking had been quenched by 9011, and a third pulmonary emboli in which the crosslinking of a2AP to fibrin had been selectively inhibited by a2AP polypeptide.
In the group with nozmal factor XIII activity, TPA caused more lysis (TPA, 32.317.7%) than was seen in the group that received S no plasminogen activator (control, 14.114.8%; p<0.005) (Figs. 1A-18).
Also, there was a nonsignificant trend (evaluated by the Dunn-Bonferroni correction) to more fibrinolysis in the group that received the factor XIII inhibitor but no TPA (Figs, lA-1B, F13-I) in comparison with the group with normal factor XIII activity that received TPA (42.7~7.4% vs 32.317~7%. p<0.05). Overall, factor XIII
activity was an important determinant of fibrinolysis in animals treated with TPA because the factor XIII inhibitor group (Fig. 1B, TPA+F13I) showed significantly more lysis than the group with normal factor XIII activity (76.0111.9% vs 32.317~7%; p<0.0001). In 1$ particular, factor XIIIa-mediated crosslinking of a2AP to fibrin made a specific contribution to fibrinolytic resistance because selective inhibition of this crosslinking also significantly accelerated lysis by TPA (Fig. 1B, TPA+a2AP-I) in comparison with lysis in animals with normal factor XIII activity (54.713.9% vs 32.317~7%; p<0.0005).
Still, selective inhibition of a2AP-fibrin crosslinking was less effective at amplifying lysis than was inhibition of all factor XIIIa-mediated crosslinking (54.713.9% vs 76.0111.9%; p<0.0005), suggesting that fibrin-fibrin crosslinking also contributed to fibrinolytic resistance.
Effects on Fibrinogen Levels To determine whether the inhibition of total factor XIII
activity or the selective inhibition of a2AP-fibrin crosslinking enhanced the systemic degradation of the clotting factor fibrinogen during fibrinolysis, fibrinogen levels were measured for all animals before and after treatment. Fig. 2 shows a comparison of the residual fibrinogen levels at the end of the study for each group, expressed as a percentage of the initial fibrinogen value. There was no significant decrease in fibrinogen levels (below 100%) for any of the experimental groups. These data indicate that nonspecific degradation of fibrinogen did not occur when TPA was administered alone, in combination with inhibition of factor XIII, or in combination with inhibition of a2AP crosslinking.
These data indicate the a2AP polypeptides selectively inhibit endogenous a2AP-fibrin crosslinking (and therefore, enhance endogenous and therapeutic fibrinolysis of pulmonary emboli) in an established art-recognized model of pulmonary embolism in ferrets.
To simulate the standard therapy for humans with pulmonary embolism, and to inhibit the accretion of new thrombus on these emboli, all animals were given doses of heparin sufficient to significantly prolong the aPTT (> 150 seconds) throughout the experiment.
Inhibition of factor XIIIa activity tripled the rate of endogenous fibrinolysis (42.7*7.4 vs 14.119.8%). This was a profound effect because the amount of endogenous lysis in clots with inhibited factor XIIIa activity was as much as, or perhaps slightly more than, the amount induced by TPA in clots with normal factor XIIIa activity (32.2~7.7%). A parallel enhancement was seen in the lysis of pulmonary emboli by TPA (1 mg/kg): inhibition of both factor XIIIa fibrin-fibrin and a2-antiplasmin-fibrin crosslinking substantially increased lysis (76.0*11.9%) over that seen with the same dose of TPA
alone (32.317.7%). In addition, selective inhibition of a2AP-fibrin crosslinking amplified TPA-induced lyais (54.7*3.9%) in comparison with that induced by TPA alone (32.3~7.7%). This effect underscores the inhibitory role played by the crosslinking of a2AP to fibrin during initiation of fibrinolysis. That even higher fibrinolysis was achieved with TPA when factor XIIIa-mediated fibrin-fibrin crosslinking and a2AP-fibrin crosslinking were both inhibited fully (76.0111.9%) indicates that both crosslinking processes contribute to inhibition of fibrinolysis induced by TPA.
If full heparinization did not completely prevent the absorption and activation of ferret factor XIII onto these clots after embolization, some degree of crosslinking may have occurred in the thrombi of all experimental groups. The effect of this crosslinking would be to blunt the increased fibrinolysis attributed to the factor XIII inhibitors. This would not change the conclusion that factor XIIIa crosslinking is a major cause of fibrinolytic resistance, but would imply that its role is even larger than was observed in these experiments.
In vitro studies suggested that fibrin-crosslinked a2AP
plays a role in the susceptibility of plasma clots to fibrinolysis.
An N-terminal peptide of a2AP was reported to inhibit a2AP-fibrin crosslinking, but the peptide used did not compete efficiently with a2AP for specific crosslinking to fibrin. The a2AP polypeptides described herein specifically and efficiently inhibit a2AP-fibrin crosslinking. The level of inhibition achieved with the polypeptide described herein is at least 50% greater than that observed with previously-described peptides, e.g., as measured by the method of Ichinose et al., 1983, FEBS Letters 153:369-371.

The Catalytic Half-life of Activated Factor XIII in Thrombi To examine whether unremitting crosslinking by activated factor XIII (FXIIIa) contributes to the increased fibrinolytic resistance of older thrombi, the persistence of FXIIIa activity in S human clots of various ages was examined.
FXIIIa activity was measured with 1) full-length a2AP, a physiologic glutamine substrate; 2) a2APl,_" (SEQ ID N0:2); and 3) pentylamine, a nonspecific lysine substrate. The crosslinking of a2AP and a2APl3_z, (SEQ ID N0:2) into fibrin by clot-bound FXIIIa declined with half-lives of approximately 62 and 72 minutes, respectively. Mutational studies showed that glutamine 14 (but not glutamine 7 or 16) and valine 17 of a2AP were required for fibrin crosslinking. FXIIIa crosslinking of pentylamine into fibrin also declined with a half-life of 173 minutes. The loss of activity was 1S not due to FXIIIa proteolysis and was partially restored by reducing agents, indicating that oxidation inhibits the enzyme over time. The physiologic persistence of FXIIIa activity in thrombi was confirmed by the crosslinking of an infused a2APl,_" peptide into existing pulmonary emboli in vivo. This crosslinking was significantly attenuated when thrombus-associated FXIIIa was inhibited.
FXIIIa croaslinks a2AP and an a2AP peptide, in a sequence-specific manner, into formed clots with a catalytic half life of approximately 1 hour. These findings indicate a preferred therapeutic window for administration of FXIIIa inhibitors. The data 2S also indicate that the catalytic activity of FXIIIa can be exploited to specifically target newly formed thrombi.
Assays of FXIIIa catalytic half-life and substrate specificity Experiments were performed to determine how long FXIIIa in clots could crosslink various substrates such as human "5I-a2AP, a2APl~_z, (SEQ ID N0:2), and 5-(biotinamido) pentylamine, into fibrin.
In a typical experiment, clots were prepared (50 ~1 final volume) in duplicate by combining fibrinogen (2 mg/ml final), CaCh (2 mM
final), buffer (iris-buffered saline pH 7.4) and thrombin (1 U/ml) and incubating at 37°C. Synchronous with the addition of thrombin, 3S or up to 240 minutes afterwards, the FXIIIa substrates 5-(biotinamido) pentylamine (0.5 mM final), "SI-a2AP (70 ~.g/ml final), and a2APl3_1, (0 to 1 mM final) were mixed and added to the clot. In one set of control samples, FXIIIa was inhibited with iodoacetamide (10 mg/ml) prior to the addition of substrate; to another set, no 40 substrate was added. After 2 hours of incubation at 37°C, iodoacetamide (10 mg/mll was added to stop the reaction. The tubes were then centrifuged at 14,000 rpm for 2 minutes, washed and compressed in 1 ml saline to remove unbound proteins, and centrifuged again at 14,000 rpm for 2 minutes. After removal of the supernatant, the clots were solubilized in 100 ~eL of 9M urea, pH 9.0 at 37°C for 60 minutes. Then 100 ~.1 of SDS reducing sample buffer was added and the clots were placed at 85°C for 30 minutes until fully solubilized.
The samples were examined by SDS-PAGE. Crosslinked a~SI-a2AP was detected and quantitated by phosphorimaging. Samples containing crosslinked a2APla_2, and pentylamine substrates were electroblotted to PVDF membranes and detected by 1'SI-streptavidin followed by phosphorimaging or by streptavidin-peroxidase followed by the developing agents 5-bromo-4-chloro-3-indolyl phosphate-nitro blue tetrazolium using methods known in the art. In experiments examining the effect of specific amino acid residues on the crosslinking of the a2AP,a_" peptide (SEQ ID N0:2) to fibrin, crosslinking was detected by immunoblotting with an anti-peptide antibody directed against this epitope followed by 1'SI-protein A and phosphorimaging. In some experiments, clots were formed by mixing 25 ,ul pooled human fresh frozen plasma with 25 ~.1 30 mM CaCla instead of using human fibrinogen.
To examine the potential degradation of FXIIIa in clots, samples from these experiments were immunoblotted with a monoclonal antibody directed against the alpha subunit of FXIII.
Crosslinking of a2APla_14 into fibrin by FXIIIa and tissue transalutaminase Fibrin clots (50 /tI) were prepared in duplicate by mixing FXIII-free fibrinogen (2 mg/ml final), a2APaa-sa (SEQ ID N0:2; 0.5 mM
final), CaCl2 (2 mM final), human plasma FXIII (100 nM/L final) and guinea pig tissue transglutaminase (100 nM/L final) and thrombin (1 U/ml final) and incubating at 37°C for 2 hours. The crosslinking of a2APla-z, was detected as described above .
CrosslinkinQ of a2AP,a_~, into pulmonarv emboli in vivo To examine whether FXIIIa which formed thrombi in vivo retained the ability to crossiink substrates, the ferret pulmonary embolism model was used. Clots were formed by combining 45 ~I of pooled human fresh frozen plasma with 1'SI-labeled fibrinogen (-100,000 cpm/clot), 2.5 ~1 of 0.4 M CaCl" and 2.5 ~1 thrombin (1000 U/ml). The clots were incubated at 37°C for 20 minutes, washed three times in saline, and 6 were embolized into the lungs of each animal.
Three experimental groups were examined with 2 animals in each group.
One group received normal clots and a solution of a2APla-z, (SEQ ID
N0:2) calculated to produce a final peptide concentration of 0.5 mM.
Another group also received the same peptide concentration; prior to embolization, the clots in this group were washed in iodoacetamide (10 mg/ml) and EDTA (10 mM final) to inhibit clot associated FXIIIa.
The third group received the blood clots but no peptide infusion afterward. Four hours after embolization, the ferrets were $ sacrificed by COZ inhalation. The heart, lungs, and great vessels were removed and gamma-counted to locate the pulmonary emboli.
Lung tissue containing pulmonary emboli was localized with a Geiger counter, excised and weighed. Samples were homogenized in 200 ~,1 of a buffer containing 24 mM Tris Base, 476 mM Tris HC1, 50 mM
MgClz, 1 mg/ml DNase I, and 0.25 mg/ml RNAase A, allowed to sit on ice for 10 minutes, centrifuged at 12,000 rgm at 4°C for 20 minutes, the supernatant was removed and the pellet was dissolved in 200 ~1 of sample buffer, mercaptoethanol, and bromophenol blue, and, finally boiled for 20-30 minutes until dissolved. T'he samples (100 ~g/lane) 1$ were then subjected to 7.5% SDS-PAGE, transferred to PVDF membranes, blocked with milk, washed, and incubated with streptavidin horseradish peroxidase (1:1000) for 1 hour, washed, incubated with a chemiluminescent substrate and exposed to film.
Histoloav Lung tissue was immersed in 30% sucrose overnight, embedded in OCT and cut into 10 micron sections. The sections were fixed in 100% methanol for 5 minutes at room temperature. For the detection of fibrin, the samples were treated with 3% HzO, for 20 minutes at room temperature, washed in PBS 3 times, treated with i0% goat normal 2$ serum for 20 minutes at room temperature. The samples were then treated with a primary polyclonal goat anti-human fibrin antibody (1:400 dilution, 2.5 ~g/ml) at 4°C overnight, washed 2 times in high salt PBS and once in regular PBS. The sections were then incubated in a secondary goat anti-rabbit peroxidase-labeled antibody (1:100) for one hour at room temperature, washed in PBS 3 times, developed in DAB (3,3'-diaminobenzidine), counterstained in 0.1% methyl green, and mounted.
For the detection of a2APl3_~, (SEQ TD N0:2), endogenous peroxidase was quenched with 3% HzO~in 100% methanol for 20 minutes 3$ at room temperature, washed in PBS 3 times. The ABC (avidin-biotin complex) reagent was applied to the section for one hour at room temperature, washed in PBS 3 times, developed in DAB substrate, counterstained in 0.1% methyl green, and mounted.
FXIIIa half-life determination To examine the catalytic half-life of FXIIIa in thrombi, the crosslinking of a2AP (FXIIIa's physiologic macromolecular substrate) was measured. The greatest amount of crosslinking of 1'SI-a2AP

occurred when it was added synchronously with thrombin to fibrinogen to initiate clotting. Thereafter, the amount of a2AP crosslinking dropped off rapidly until almost no crosslinking was detected at 80 minutes, particularly when compared to controls in which FXIIIa has S been inhibited with iodoacetamide, or in experiments containing no 1'sI_a2AP. Semilog plots of the decay of crosslinking of 1'SI-a2AP to fibrin by FXIIIa indicated a half-life of 62 minutes (r'=0.92), consistent with an exponential decline in catalytic activity.
One potential explanation for the marked decline in the crosslinking of a2AP to fibrin was that the size of a2AP (70,000 Da) inhibited its diffusion into the assembling fibrin meshwork. To examine this possibility, a peptide that mimicked the region of a2AP
(a2APl;_s, (SEQ ID N0:2)) that contains the glutamine residue which is crosslinked to fibrin was used. To permit detection, the peptide was biotinylated at lysine 24. The small size of this peptide (1600 Da) readily permits diffusion into a developing thrombus. This peptide and the variant peptides described herein inhibit the crosslinking of a2AP to fibrin when added during clotting. The inhibitory peptides crosslinked into fibrin clots in a dose-dependent manner. The mast exuberant crosslinking of the peptide occurred to proteins with the apparent molecular sizes of 60-70 kDa, though crosslinking into proteins of 50 kDa, 90 kDa, and 120 kDa and above was also seen at high concentrations. When compared to control clots, crosslinking of the a2AP"_" (SEQ ID N0:2) peptide is seen at concentrations as low as 2$ 1.3 ~M, reaches half of maximum crosslinking at 110 ~.M and appears to saturate at concentrations of 0.33 to 1 mM. A similar dose response curve was found for the crosslinking of the a2APl3-a~ (SEQ ID N0:2) peptide to fibrin in plasma.
Although FXIIIa can crosslink a number of lysine analogs to glutamine sites in macromolecules such as fibrin, its specificity for crosslinking glutamine-containing substrates (acyl donors) such as a2AP to fibrin has not been defined. To examine specificity, a wild-type a2AP fragment spanning residues 1-24 (MEPLGWQLTS GPNQEQVSPL
TLLK; SEQ ID N0:16) and a panel of a2AP variant peptides (corresponding to residues 1-24 with the exception that one residue is altered from the wild type sequence; SEQ ID N0:11, 12, 13, 14, 15) was created with mutations of residues that represented other potential crosslinking sites or residues that are conserved among different a2APs from different species. Fig. 3A shows the sequences of the wild-type and variant peptides. By comparison to clots containing no peptide or the wild type peptide a2APl_", peptides containing mutations Q14A (SEQ ID N0:12) and V17N (SEQ ID N0:14) were not efficiently crosslinked to fibrin (Fig. 3B). In contrast, a2AP
peptides with mutations at E15A (MEPLGWQLTS GPNQAQVSPL TLLK; SEQ ID
N0:17), Q16A (MEPLGWQLTSGPNQEAVSPLTLLK; SEQ ID N0:13) and L22N
(MEPLGWQLTSGPNQEQVSPLTNLK; SEQ ID N0:15) were crosslinked into fibrin S at rates comparable to that of the wild type a2APl_" peptide (SEQ ID
N0:16). These variant peptides (as well as Q7A
(MEPLGWALTSGPNQEQVSPLTLLK; SEQ ID NO:11) and peptide mimetics thereof are suitable for use in the diagnostic and therapeutic applications described herein.
Similar results were seen when these peptides were used as inhibitors of the crosslinking of a2APl,_z, (SEQ ID N0:2) into fibrin:
all peptides but Q14A and V17N competed for crosslinking. Taken together, these results indicate that a2APl_s, (SEQ ID N0:16) and a2AP13_" (SEQ ID N0:2) are specific glutamine substrates for FXIIIa 1$ and indicate that Q14 and V17 are necessary for efficient fibrin crosslinking. While FXIIIa is the primary transglutaminase found in plasma, other transglutaminases are found in cells that do not require thrombin for activation. Additional experiments were carried out to determine whether the a2AP13_is ISEQ ID N0:2) peptide is efficiently crosslinked by tissue transglutaminase. In studies with purified, FXIIIa-deficient fibrinogen, there was little, if any, significant crosslinking of the a2APl~_" (SEQ ID N0:2) peptide to fibrin with clotting. Addition of purified tissue transglutaminase (100 nM) caused a slight increase in crosslinking, whereas addition of equivalent amounts of purified exogenous FXIIIa caused significantly more crosslinking of the a2AP13_s,peptide to fibrin.
Moreover, in experiments with purified human fibrinogen, the potent thrombin inhibitor hirudin blocked a2AP13_1, peptide-fibrin crosslinking. These data indicate that FXIIIa is responsible for the crosslinking of the a2AP13_" peptide in these fibrinogen preparations because it is the only thrombin-dependent transglutaminase.
Because the a2APl3_z,peptide was a selective FXIIIa substrate that should be readily permeable to the fibrin clot because of its small size, it was used to measure the catalytic life of FXIIIa in formed or forming clots. Consistent with experiments performed with "sI-a2AP, the crosslinking of a2APl3_" (SEQ ID N0:2) to fibrin was greatest when it was present at the initiation of clotting (t=0).
The relative amount of crosslinking fell rapidly over the course of an hour to nearly undetectable levels. The half-life calculated for a2APl,_z, crosslinking from these studies was 72 minutes (rz= 0.998) which was comparable to that found for the "sI-a2AP.

To determine whether FXIIIa showed a similar catalytic half-life for a lysine (acyl acceptor) mimic (pentylamine-biotin, an even smaller substrate than a2AP~,_" (SEQ ID N0:2)), pentylamine-biotin was crosslinked into fibrin clots. The half-life was 173 minutes (r~=
$ 0.774), which is nearly 2.3-2.8 times longer than was seen with a2AP
or a2APl3_s, (SEQ ID N0:2). Since proteolysis of FXIIIa at a second thrombin site has been described to inactivate the enzyme, studies were carried out to determine whether the proteolytic fragmentation accounts for the loss of FXIIIa activity in these clots. There was no significant increase in the relative amounts of the smaller FXIIIa fragment (56 kDa) versus FXIIIa (80 kDa). These data indicate that proteolysis alone does not explain the marked decline in FXIIIa activity. Since the FXIIIa contains multiple reduced cysteines, including the active site residue, dithiothreitol (DTT) was added to 1$ the reaction mixtures to determine whether the loss in catalytic activity of the molecule could be restored by the addition a reducing agent. Under normal circumstances, when pentylamine or a2APl,_" (SEQ
ID N0:2) was added to 2 or 4 hour old clots, there was little residual FXIIIa crosslinking. However, by comparison DTT, restored the crosslinking of pentylamine or a2APl,_" to fibrin clots. These data indicate that reversible oxidation of FXIIIa rather than proteolysis is responsible for the precipitous loss of enzymatic function.
To confirm the physiologic significance of these findings, the 2$ ability of FXIIIa to crosslink a2APl,_" (SEQ ID N0:2) in vivo to existing pulmonary emboli was examined. Three experimental groups were studied: a control group that received no peptide after embolism of preformed clots and two experimental that groups received a 0.5 mM dose of a2AP"_z, (SEQ ID N0:2) peptide after embolization of normal clots or clots that were pretreated with iodoacetamide and EDTA to inhibit residual clot-associated FXIIIa. After 4 hours, the animals were sacrificed and portions of the lung containing the lasl_ fibrin-labeled thrombi or no thrombi were isolated, weighed and analyzed. Pulmonary emboli occluded pulmonary arterioles 3$ Immunostaining with a polyclonal anti-fibrin antibody confirmed that the pulmonary arteriole occlusion was a thrombus. Histologic sections from animals that received emboli with normal FXIIIa activity with or without the a2APl,_z, (SEQ ID N0:2) peptide after embolization were also examined. These histologic sections were probed with avidin-peroxidase to detect the thrombus associated a2APl3_~, (SEQ ID N0:2) biotinylated peptide. When compared to a control thrombus (middle panel) from an animal not receiving a2APl3_sa (SEQ ID N0:2), specific staining was detected in the thrombus from an animal receiving the peptide infusion, indicating incorporation of the a2AP13_~, (SEQ ID N0:2) into the thrombus.
Lung tissue samples were subjected to SDS-PAGE and blotted with $ 1'SI-streptavidin to detect biotinylated a2AP13_" (SEQ ID N0:2) and stained by Coomassie blue to confirm that the peptide was covalently incorporated into fibrin in the thrombus. Lung tissue from control animals (normal emboli and no a2APl,.za (SEQ ID N0:2)) showed two nonspecific avidin binding bands at -70 kDa and --120 kDa in both the embolus-containing and non-embolus containing tissues. In animals receiving the a2APls-z4 (SEQ ID N0:2) peptide, there were similar non-specific staining bands at 70kDa and 120 kDa in the lung tissue without emboli. However, the lung tissue containing pulmonary emboli showed a new broad band at -65-70 kDa, and additional bands at 60 kDa, -100 kDa and ~140 kDa. The intensity of the bands was diminished in the lung tissue from an animal with a pulmonary embolus in which the thrombus associated FXIIIa had been inhibited by iodoacetamide and EDTA (lane 3). Taken together, these data indicate that a2AP fragments (e. g., amino-terminal a2AP fragments and variants thereof) are crosslinked into preexisting pulmonary embolus in vivo and that crosslinking is attenuated by inhibition of the FXIIIa in the thrombus.
The finding that FXIIIa crosslinking causes fibrinolytic resistance in new pulmonary emboli underscores the importance of 2$ FXIIIa regulation, and indicates that continued FXIIIa crosslinking contributes to the time-related increase in fibrinolytic resistance of thrombi with age. In clots, the catalytic ability of FXIIIa to crosslink a physiologic glutamine substrate (a2AP, 70kDa), or a small peptide fragment of that substrate (a2APl,." (SEQ ID N0:2), 1.6 kDa), declines in a negative exponential fashion with calculated half-lives of 62 and 72 minutes, respectively. The comparable half-lives of these two substrates, despite marked differences in size, indicates that crosslinking of a2AP was not simply limited by its permeation into the fibrin clot. The ability of FXIIIa to crosslink a lysine 3$ substrate analog pentylamine-biotin also declined in a negative exponential fashion but with a longer half-;life of 173 minutes. The decline in catalytic activity of FXIIIa could not be ascribed to proteolysis of the catalytic subunit FXIIIa, but was related to potential oxidation. FXIIIa contains an active site cysteine residue that is known to be vulnerable to inhibition by oxidation. The relative 'restoration' of FXIIIa crosslinking in older clots of the a2APl;_~, (SEQ ID N0:2) peptide by reducing agents, was significantly less marked than was the restoration of crosslinking of the pentylamine-biotin substrate, suggesting that the number of sites available for crosslinking of the a2APl,_" substrate may actually be limiting in these older clots. In vivo studies confirmed in vitro $ findings by demonstrating that formed pulmonary emboli (30 minutes old) covalently incorporated a2APl3_z, (SEQ ID N0:2) peptide. This incorporation was largely due to the enzymatic activity of the thrombus-bound FXIIIa as it was significantly attenuated in thrombi in which FXIIIIa had been inhibited immediately prior to embolization. Although previous studies have examined the clearance of circulating FXIII zymogen in deficient patients, these data represent the first estimates of the catalytic life of FXIIIa in clots.
FXIIIa makes thrombi resistant to plasmin by modifying the fibrin matrix through intermolecular crosslinks. FXIIIa mediated fibrin a-chain polymerization, and y-chain multimerization contributes to this fibrinolytic resistance. In addition, the crosslinking of the potent plasmin inhibitor a2AP to fibrin plays a clear role in neutralizing fibrinolysis. In vitro, the crosslinking of a2AP to fibrin and the formation of 'y-dimers occurs rapidly within 2-5 minutes. However, a -chain polymerization occurs more slowly and fibrin y-chain multimerization takes hours to days. Experimental studies of rabbit thrombi formed in vivo confirm that y-chain dimerization and a-chain polymerization is readily seen within 7 to 9 minutes of thrombus formation, though small increments in a-chain polymerization can be detected after 90 to 320 minutes. There are limited studies of the fibrin structure of human thrombi, but when examined ex vivo, all showed evidence of complete a-chain polymerization, though the ages of these thrombi were not well specified.
Crosslinking by FXIIIa is highly specific because only a minority of potential glutamines and lysines in proteins are actually substrates. The primary structure, charge, and conformation around the respective glutamine residues determine the suitability of a 3S peptide-bound glutamine as a substrate for FXIIIa. Of the 3 potential glutamine sites in the amino-terminus of the a2APl~_"
peptide, only glutamine 14 is required. In addition, Va117, which is highly conserved in other species a2AP also appears to be required.
Further, although purified tissue transglutaminase can crosslink the a2AP13_s, peptide into fibrin, much as it can crosslink fibrin chains together it does so at a lower rate than equivalent amounts of FXIIIa. Moreover, the crosslinking of the a2APl,_" (SEQ ID N0:2) peptide into fibrin and plasma clots requires thrombin activity, a property that signifies the involvement of FXIIIa.
These data indicate that FXIIIa, in 30 minute-old emboli, retains the catalytic ability to incorporate a specific substrate $ into fibrin. The observation that FXIIIa remains catalytically active only in recent thrombi is useful to detect or target therapies to 'new' thrombi. An advantage of the diagnostic and therapeutic approaches described herein is that they are not affected by treatment with conventional anticoagulants such as heparin or warfarin because they do not affect FXIIIa activity. The ability to distinguish new from old thrombi could significantly improves the specificity of fibrinolytic therapy and permits targeting therapeutic compositions to recently formed clots. The compositions are also useful to prevent future thrombotic events.
Therapeutic applications a2-antiplasmin causes resistance to endogenous as well as pharmacologic fibrinolysis in venous thromboemboli in vivo. It is likely that other molecular factors in the thrombus (e.g., PAI-1 and a2-antiplasmin) cause the fibrinolytic resistance seen in thrombotic diseases such as pulmonary embolism.
In humans, pulmonary emboli appear to develop from the fragmentation of propagating thrombi in the deep venous system.
Anticoagulants that interfere with the activity of thrombin, prevent thrombus propagation by inhibiting the new deposition of fibrin.
2$ Still, despite effective anticoagulation, the inherent fibrinolytic resistance of formed thrombi prevents optimal treatment of patients with thrombotic disease. In studies of heparinized animals described above, factor XIIIa-mediated crosslinking was found to play a critical role in limiting the endogenous and pharmacologic fibrinolysis of formed experimental pulmonary emboli. These data indicate that factor XIII inhibitors such as the a2AP polypeptides described herein are useful for the treatment of thrombotic disease.
a2AP polypeptides will ordinarily be at least about 12 amino acids, usually about 20 contiguous amino acids, preferably at least 40 contiguous amino acids, more preferably at least SO contiguous amino acids, and most preferably at least about 60 to 80 contiguous amino acids in length. Such peptides can be generated by methods known to those skilled in the art, including proteolytic cleavage of the protein, de novo synthesis of the fragment, or genetic engineering, e.g., cloning and expression of a fragment of a2AP-encoding cDNA. For example, chimeric therapeutic agents in which an N-terminal a2AP polypeptide is linked to a thrombolytic agent such as TPA can be recombinantly produced by cloning DNA encoding the a2AP
portion of the chimera in frame to DNA encoding TPA (or another therapeutic agent) into a suitable expression vector and expressing the recombinant chimeric protein in eucaryotic or procaryotic cells using standard expression technology. DNA sequences encoding mammalian a2AP (such as human, bovine, and mouse a2AP) and encoding such thrombolytic agents as TPA are publicly available from databases such as GENBANK"' .
Variants of a2AP polypeptides described above (e. g., those containing the amino acid sequences of SEQ xD N0:1-7) may also be used provided that they have the activity of: inhibiting a2AP-fibrin crosslinking. Variants can differ by amino acid sequence (e. g, conservative amino acid substitutions, e.g., those shown in Table 4), or by modifications which do not affect the sequence, or both.

mnuT.F a . r~c~w~FrtVATTVE AMINO ACIB SUBSTITUTIONS
For Amino Acid Code Replace With Alanine A D-Ala, Gly, Aib, /3-Ala, Acp, L-Cys, D-Cys Arginine R D-Arg, Lys, D-Lys, homo-Arg, D-homo-Arg, Met, Ile, D-Met, D-Ile, Orn, D-Orn $ Asparagine N D-Asn, Asp, D-Asp, Glu, D-Glu, Gln, D-Gln Aspartic Acid D D-Asp, D-Asn, Asn, Glu, D-Glu, Gln, D-Gln Cysteine C D-Cys, S-Me-Cys, Met, D-Met, Thr, D-Thr Glutamine Q D-Gln, Asn, D-Asn, Glu, D-Glu, Asp, D-Asp Glutamic Acid E D-Glu, D-Asp, Asp, Asn, D-Asn, Gln, D-Gln 1~ Glycine G Ala, D-Ala, Pro, D-Pro, Aib, (3-Ala, Acp Isoleucine I D-Ile, Val, D-Val, AdaA, AdaG, Leu, D-Leu, Met, D-Met Leucine L D-Leu, Val, D-Val, AdaA, AdaG, Leu, D-Leu, Met, D-Met Lysine K D-Lys, Arg, D-Arg, homo-Arg, D-homo-Arg, Met, D-Met, Ile, D-Ile, Orn, D-Orn Methionine M D-Met, S-Me-Cys, Ile, D-Ile, Leu, D-Leu, Val, D-Val 1$ Phenylalanine F D-Phe, Tyr, D-Thr, L-Dopa, His, D-His, Trp, D-Trp, Trans-3,4, or 5-phenylproline, AdaA, AdaG, cis-3,4, or 5-phenylproline, Bpa, D-Bpa Proline P D-Pro, L-I-thioazolidine-4-carboxylic acid, D-or L-1-oxazolidine-4-carboxylic acid (Kaiser, U.S. Patent (4,511,390) Serine S D-Ser, Thr, D-Thr, cello-Thr, Met, D-Met, Met(O), D-Met(O), L-Cys, D-Cys Threonine T D-Thr, Ser, D-Ser, cello-Thr, Met, D-Met, Met(0), D-Met(O), Val, D-Val Tyrosine Y D-Tyr, Phe, D-Fhe, L-Dopa, His, D-His 2~ Valine V D-Val, Leu, D-Leu, Ile, D-Ile, Met, D-Met, AdaA, AdaG

Modifications (which do not normally alter primary sequence) include in vivo or in vitro chemical derivitization of polypeptides, e.g., acetylation or carboxylation. Also included are modifications of glycosylation, e.g., those made by modifying the glycosylation 2$ patterns of a polypeptide during its synthesis and processing or in further processing steps, e.g., by exposing the polypeptide to enzymes which affect glycosylation, e.g., mammalian glycosylating or deglycosylating enzymes.
To render the therapeutic peptides less susceptible to cleavage by peptidases, the peptide bonds of a peptide may be replaced with an alternative type of covalent bond (a "peptide mimetic"). Several types of chemical bonds, e.g. ester, thioester, thioamide, retroamide, reduced carbonyl, dimethylene and ketomethylene bonds, are known in the art to be generally useful substitutes for peptide bonds in the construction of protease-resistant peptide mimetics.
Where proteolytic degradation of the peptides following injection into a mammalian subject, e.g., a human patient, is a problem, replacement of a particularly sensitive peptide bond with a noncleavable peptide mimetic bond will make the resulting peptide more stable, and thus more useful as a therapeutic agent. Such mimetics, and methods of incorporating them inta polypeptides, are well known in the art. Similarly, the replacement of an L-amino acid residue with a D-amino acid is a standard way of rendering the polypeptide less sensitive to proteolysis. Also useful are amino-terminal blocking groups such as t-butyloxycarbonyl, acetyl, theyl, succinyl, methoxysuccinyl, suberyl, adipyl, azelayl, dansyl, benzyloxycarbonyl, fluorenylmethoxycarbonyl, methoxyazelayl, methoxyadipyl, methoxysuberyl, and 2,4,-dinitrophenyl.
Peptides may be administered to the patient intravenously in a pharmaceutically acceptable carrier such as physiological saline.
Such methods are well known to those of ordinary skill in the art.
It is expected that an intravenous dosage of approximately 1 to 100 ~cmoles of the peptide of the invention would be administered per kg of body weight per day. For diagnostic pracedures, smaller peptides (e. g., 75 amino acids or less, preferably 50 amino acids or less, more preferably 25 amino acids or less, and most preferably 15 amino acids or less, in length) can be administered at mM concentrations because these shorter polypeptides are croselinked at a faster rate than larger polypeptides and are cleared rapidly from the body.
The formulations of this invention axe useful for parenteral administration, such as intravenous, subcutaneous, intramuscular, and intraperitoneal. Other methods of delivery, e.g., diffusion from a stent impregnated with therapeutic polypeptide or peptide mimetic, are known in the art.
Diagnostic av~lications Factor IIIa mediates a2AP-fibrin crosslinking in new or actively forming thrombi. As described above, the a2AP polypeptides described herein compete with endogenous a2AP, i.e., a2AP polypeptide also become incorporated into newly forming clots. Detectably-labelled a2AP polypeptides are therefore useful to non-invasively detect clot formation. The a2AP polypeptides are labeled With an imaging marker to permit detection. For in vivo imaging, the S peptides are preferably labeled with a radioisotope such as a gamma-emitter, positron-emitter, or x-ray-emitter (including, e.g., indium-111, technetium-99m, iodine-125, gallium-67, and gallium-68).
Alternatively, the polypeptide may be biotinylated (e.g., via a lysine residue of the polypeptide).
The labelled polypeptide is administered to a subject in a dose which is diagnostically effective. The term "diagnostically effective" means that the detectably labeled species is administered in sufficient quantity to enable detection of the blood clot. The concentration of detectably labeled species administered should be 1S sufficient so the binding to thrombi is detectable compared to the background. For example, for diagnostic purposes, shorter polypeptides are preferable because the signal-to-noise ratio is better than that associated with longer peptides (e. g., over 75 amino acids in length). Further, it is desirable that the delectably labeled polypeptide be rapidly cleared from the circulatory system in order to give the best target-to-background signal ratio.
As a rule, the dosage of detectably labeled species for in vivo diagnosis will vary depending on such factors as age, sex, and extent of disease of the individual. For instance, the dosage of monoclonal 2S antibody can vary from about 0.01 mg/m' to about 500 mg/m', preferably 0.05 mg/m' to about 200 mg/m', most preferably about 0.1 mg/m' to about 10 mg/m'. Such dosages may vary, for example, depending on whether multiple injections are given, progression of the disease, and other factors known to those of skill in the art.
For in vivo diagnostic imaging, the type of detection instrument available is a major factor in selecting a given radioisotope. The radioisotope chosen must have a type of decay which is detectable for a given type of instrument. Still another important factor in selecting a radioisotope fox in vivo diagnosis is 3S that the half-life of the radioisotope be long enough so that it is still detectable at the time of maximum uptake by the target, but short enough so that deleterious radiation with respect to the host is minimized. Ideally, a radioisotope used for in vivo imaging will lack a particle emission, but produce a large number of photons in the 140-250 keV range, which may be readily detected by conventional gamma cameras.

The sites of localization may be determined by standard imaging techniques, preferably planar imaging or single photon emission computed tomography (SPELT), and by gamma camera imaging. In addition, the polypeptides can also be labeled with a paramagnetic isotope for purposes of in vivo diagnosis, as in magnetic resonance imaging (MRI) or electron spin resonance (ESR). In general, any conventional method for visualizing diagnostic imaging can be utilized. Usually gamma and positron emitting radioisotopes are used for camera imaging and paramagnetic isotopes for MRI. Elements which are particularly useful in such techniques include ls'Gd, ss~, ~ssDy~
s'Cr, and s6Fe .
For in vivo diagnosis, radioisotopes may be bound to a polypeptide either directly or indirectly by using an intermediate functional group. Intermediate functional groups which often are used to bind radioisotopes which exist as metallic ions to immunoglobulins are the bifunctional chelating agents such as diethylenetriaminepentacetic acid (DTPA) and ethylenediaminetetraacetic acid (EDTA) and similar molecules.
Typical examples of metallic ions which can be bound to polypeptides are luln, s'Ru ~ s~Ga ~ seGa ~ »As ~ esZr eoY ~ and ~oiTl .
The labelled a2AP polypeptide is administered to the patient in a pharmaceutically acceptable carrier and labeled with a marker to permit in vivo detection as described above. Pharmaceutically acceptable carriers useful for imaging and therapy are well-known in the art and include, for example, aqueous solutions such as bicarbonate buffers, phosphate buffers, Ringer's solution and physiological saline supplemented with 5% dextrose or human serum albumin, if desired.
Following administration of the labelled polypeptide (e. g., intravenously) to the subject, association of the polypeptide with a vascular obstruction is determined. Association, e.g., via crosslinking to fibrin, is an indication of the presence of blood clot formation at that site. where the detecting step is quantitative, the amount of binding would correlate with and allow diagnosis of the severity of the disease. Furthermore, if the diagnostic method is carried out multiple times by repeatedly administering at spaced intervals with the administrations spaced by, e.g., a day, a week, a month, several months, or even years, the method is useful to detecting progressive thrombotic disease or responsiveness to therapeutic intervention..
Detection of thrombi with labelled a2AP polypeptides offers several advantages over convention detectian approaches such as assays using fibrinogen. For example, since fibrinogen was historically purified from human sources, the conventional fibrinogen-based assay is no longer feasible due to the danger of HIV-contamination. Since the disclosed a2AF polypeptides are smaller than fibrinogen, they cane be administered in higher concentration (e. g., mM doses} than conventional diagnostic compounds. Their relatively small size also allows quicker clearance from the subject and a more favorable signal-to-noise ratio.
Other embodiments are within the following claims.

Claims (45)

What is claimed is:

1. A method of detecting blood clot formation in a mammal, comprising (a) administering to said mammal a diagnostically effective amount of a detectably labeled alpha-2 antiplasmin (.alpha.2AP) polypeptide, and (b) determining whether said polypeptide associates with a blood clot, said association being an indication of the presence blood clot formation.

2. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of X1QX2X3X4X5PLX6LLK (SEQ ID NO:1), wherein X1 = N or A, X2 = E or Q, X2 = Q or K, X4 = V or L, X5 = P or S, and X6 =
T, S or A.

3. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of NQEQVSPLTLLK (SEQ ID NO:2).

4. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEQVSPLTLLK (SEQ ID NO:16).

5. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQAQVSPLTLLK (SEQ ID NO:17).

6. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of (MEPLGWALTSGPNQEQVSPLTLLK; SEQ ID NO:11).

7. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEAVSPLTLLK (SEQ ID NO:13).

8. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEQVSPLTNLK (SEQ ID NO:15).

9. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT (SEQ ID NO:4).

10. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLGRQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT
(SEQ ID NO:5).

11. The method of claim 1, wherein said polypeptide comprises the amino acid sequence of MEPLDLQLMD GQAQQKLPPL SLLKLDNQEP
GGQIAPKKAP EDCKLSPTPE QTRRLARAMM TFTTDLFSLV AQS (SEQ ID NO:6).

12. The method of claim 1, wherein the polypeptide comprises the amino acid sequence of VDLPGQQPVS EQAQQKLPLP ALFKLDNQDF
GDHATLKRSP GHCKSVPTAE ETRRLAQAMM AFTTDLFSLV AQT (SEQ ID NO:7).

13. A method of inhibiting blood clot formation in a mammal comprising administering to said mammal a therapeutically effective amount of an .alpha.2AP polypeptide.

14. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of X1QX2X3X4X5PLX6LLK (SEQ ID NO:1), wherein X1 = N or A, X2 = E or Q, X3 = Q or K, X, = V or L, X5 = P or S , and X6 =T, S or A.

15. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of NQEQVSPLTLLK (SEQ ID NO:2).

16. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEQVSPLTLLK (SEQ ID NO:16).

17. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQAQVSPLTLLK (SEQ ID NO:17).

18. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of (MEPLGWALTSGPNQEQVSPLTLLK; SEQ ID NO:11).

19. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEAVSPLTLLK (SEQ ID NO:13).

20. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTSGPNQEQVSPLTNLK (SEQ ID NO:15).

21. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT (SEQ ID NO:4).

22. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLGRQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT
(SEQ ID NO:15).

23. The method of claim 13, wherein said polypeptide comprises the amino acid sequence of MEPLDLQLMD GQAQQKLPPL SLLKLDNQEP
GGQIAPKKAP EDCKLSPTPE QTRRLARAMM TFTTDLFSLV AQS (SEQ ID NO:6).

24. The method of claim 13, wherein the polypeptide comprises the amino acid sequence of VDLPGQQPVS EQAQQKLPLP ALFKLDNQDF
GDHATLKRSP GHCKSVPTAE ETRRLAQAMM AFTTDLFSLV AQT (SEQ ID NO:7).

25. A method of preventing blood clot formation, comprising administering to a mammal at risk of developing thrombosis, a therapeutically effective amount of an .alpha.2AP polypeptide.

26. A method of preventing and lysing blood clots, comprising administering to said mammal a therapeutically effective amount of an .alpha.2AP polypeptide and a thrombolytic agent.

27. The method of claim 26, wherein said thrombolytic agent is a plasminogen activator.

28. A method of targeting a therapeutic agent to an actively developing thrombus in a mammal, comprising administering to said mammal an N-terminal .alpha.2AP polypeptide linked to said therapeutic agent.

29. The method of claim 28, wherein said therapeutic agent is a thrombolytic agent.

30. An .alpha.2AP polypeptide, the length of which is 12 to 250 amino acids, inclusive, wherein said polypeptide comprises the amino acid sequence of X1QX2X3X4X5PLX6LLK (SEQ ID NO:1), wherein X1 = N or A, X2 = E or Q, X3 = Q or K, X4 = V or L, X5 = P or S, and X6 = T, S or A.

31. The polypeptide of claim 30, wherein said polypeptide inhibits crosslinking of .alpha.2AP-fibrin crosslinking.

32. The polypeptide of claim 31, wherein said polypeptide is to 75 amino acids, inclusive, in length.

33. The polypeptide of claim 32, wherein said polypeptide is 12 amino acids in length.

34. The polypeptide of claim 32, wherein said polypeptide is 24 amino acids in length.

35. The polypeptide of claim 30, wherein said polypeptide comprises the amino acid sequence of NQEQVSPLTLLK (SEQ ID NO:2).

36. The polypeptide of claim 30, wherein said polypeptide comprises the amino acid sequence of MEPLGWQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT (SEQ ID NO:4).

37. The polypeptide of claim 30, wherein said polypeptide comprises the amino acid sequence of MEPLGRQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT (SEQ ID NO:5).

38. An .alpha.2AP polypeptide comprising an amino acid sequence with 80-100% sequence identity to MEPLGXQLTS GPNQEQVSPL TLLKLGNQEP
GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT (SEQ ID NO:3), wherein (a) a non-identical amino acid of said polypeptide is a conservative amino acid substitution; and (b) said polypeptide inhibits .alpha.2AP-fibrin crosslinking.

39. The polypeptide of claim 30, wherein said polypeptide comprises the amino acid sequence of MEPLDLQLMD GQAQQKLPPL SLLKLDNQEP
GGQIAPKKAP EDCKLSPTPE QTRRLARAMM TFTTDLFSLV AQS (SEQ ID NO:6).

40. The polypeptide of claim 30, wherein said polypeptide is detectably labeled.

41. The polypeptide of claim 40, wherein said polypeptide is biotinylated.

42. The polypeptide of claim 30, wherein said polypeptide is linked to a thrombolytic agent.

43. A peptide mimetic of an .alpha.2AP polypeptide, the length of which is 12 to 250 amino acids, inclusive, wherein said polypeptide comprises the amino acid sequence of X1QX2X3X4X5PLX6LLK (SEQ ID NO:1), wherein X1 = N or A, X2 = E or Q, X3 = Q or K, X4 = V or L, X5 = P or S, and X6 = T, S or A.

44. A peptide mimetic of an .alpha.2AP polypeptide comprising an amino acid sequence with 80-100% sequence identity to NQEQVSPLTLLK
(SEQ ID NO:2), wherein (a) a non-identical amino acid of said polypeptide is a conservative amino acid substitution; and (b) said polypeptide inhibits .alpha.2AP-fibrin crosslinking.

45. A peptide mimetic of an .alpha.2AP polypeptide comprising an amino acid sequence with 80-100% sequence identity to MEPLGXQLTS
GPNQEQVSPL TLLKLGNQEP GGQTALKSPP GVCSRDPTPE QTHRLARAMM AFTADLFSLV AQT
(SEQ ID NO:3), wherein (a) a non-identical amino acid of said polypeptide is a conservative amino acid substitution; and (b) said polypeptide inhibits .alpha.2AP-fibrin crosslinking.