Potent noncovalent thrombin inhibitors that utilize the unique amino acid D-dicyclohexylalanine in the P3 position. Implications on oral bioavailability and antithrombotic efficacy.

In an effort to prepare orally bioavailable analogs of our previously reported thrombin inhibitor 1, we have synthesized a series of compounds that utilize the unique amino acid D-dicyclohexylalanine as a P3 ligand. The resulting compounds are extremely potent and selective thrombin inhibitors, and the N-terminal Boc derivative 8 exhibited excellent oral bioavailability and pharmacokinetics in both rats and dogs. The des-Boc analog 6 was not orally bioavailable in rats. The high level of oral bioavailability observed with 8 appears to be a direct function of its increased lipophilicity versus other close analogs. Although increased lipophilicity may serve to increase the oral absorption of tripeptide thrombin inhibitors, it also appears to have detrimental effects on the antithrombotic properties observed with the compounds. Compound 6 performed extremely well in our in vivo antithrombotic assay, while the much more lipophilic but essentially equipotent analog 8 performed poorly. We have found that in general with this series of thrombin inhibitors as well as with other unreported series, increased lipophilicity and the associated increases in plasma protein binding have detrimental effects on 2X APTT values and subsequent performance in in vivo antithrombotic models.

Discovery and development of the novel potent orally active thrombin inhibitor N-(9-hydroxy-9-fluorenecarboxy)prolyl trans-4-aminocyclohexylmethyl amide (L-372,460): coapplication of structure-based design and rapid multiple analogue synthesis on solid support.

Early studies in these laboratories of peptidomimetic structures containing a basic P1 moiety led to the highly potent and selective thrombin inhibitors 2 (Ki = 5.0 nM) and 3 (Ki = 0.1 nM). However, neither attains significant blood levels upon oral administration to rats and dogs. With the aim of improving pharmacokinetic properties via a more diverse database, we devised a resin-based route for the synthesis of analogues of these structures in which the P3 residue is replaced with a range of lipophilic carboxylic amides. Assembly proceeds from the common P2-P1 template 7 linked via an acid-labile carbamate to a polystyrene support. Application of the methodology in a repetitive fashion afforded several interesting analogues out of a collection of some 200 compounds. Among the most potent of the group, N-(9-hydroxy-9-fluorenecarboxy)-prolyl trans-4-aminocyclohexylmethyl amide (L-372,460 8, Ki = 1.5 nM), in addition to being fully efficacious in a rat model of arterial thrombosis at an infusion rate of 10 micrograms/kg/min, exhibits oral bioavailability of 74% in dogs, and oral bioavailability of 39% in monkeys with a serum half-life of just under 4 h. On the basis of its favorable biological properties, inhibitor 8 has been subject to further evaluation as a possible treatment for thrombogenic disorders.

Discovery of a novel, selective, and orally bioavailable class of thrombin inhibitors incorporating aminopyridyl moieties at the P1 position.

A novel class of thrombin inhibitors incorporating aminopyridyl moieties at the P1 position has been discovered. Four of these thrombin inhibitors (13b,c,e and 14d) showed nanomolar potency (Ki 0.8-12 nM), 300-1500-fold selectivity for thrombin compared with trypsin, and good oral bioavailability (F = 40-76%) in rats or dogs. The neutral P1 was expected to increase metabolic stability and oral absorption. Identification of this novel aminopyridyl group at P1 was a key step in our search for a clinical candidate.

Synthesis of a series of potent and orally bioavailable thrombin inhibitors that utilize 3,3-disubstituted propionic acid derivatives in the P3 position.

As part of an effort to prepare efficacious and orally bioavailable analogs of the previously reported thrombin inhibitors 1a, b, we have synthesized a series of compounds that utilize 3,3-disubstituted propionic acid derivatives as P3 ligands. By removing the N-terminal amino group, the general oral bioavailability of this class of compounds was enhanced without excessively increasing the lipophilicity of the compounds. The overall properties of the molecules could be drastically altered depending on the nature of the groups substituted onto the 3-position of the P3 propionic acid moiety. A number of the compounds exhibited good oral bioavailability in rats and dogs, and numerous compounds were efficacious in a rat FeCl3-induced model of arterial thrombosis. Compound 7, the 3,3-diphenylpropionic acid derivative, showed the best overall profile of in vivo and in vitro activity. Molecular modeling studies suggest that these compounds bind in the thrombin active site in a manner essentially identical to that previously reported for compound 1a.

Design and synthesis of a series of potent and orally bioavailable noncovalent thrombin inhibitors that utilize nonbasic groups in the P1 position.

As part of an ongoing effort to prepare therapeutically useful orally active thrombin inhibitors, we have synthesized a series of compounds that utilize nonbasic groups in the P1 position. The work is based on our previously reported lead structure, compound 1, which was discovered via a resin-based approach to varying P1. By minimizing the size and lipophilicity of the P3 group and by incorporating hydrogen-bonding groups on the N-terminus or on the 2-position of the P1 aromatic ring, we have prepared a number of derivatives in this series that exhibit subnanomolar enzyme potency combined with good in vivo antithrombotic and bioavailability profiles. The oxyacetic amide compound 14b exhibited the best overall profile of in vitro and in vivo activity, and crystallographic studies indicate a unique mode of binding in the thrombin active site.

Efficacious, orally bioavailable thrombin inhibitors based on 3-aminopyridinone or 3-aminopyrazinone acetamide peptidomimetic templates.

We have addressed the key deficiency of noncovalent pyridinone acetamide thrombin inhibitor L-374,087 (1), namely, its modest half-lives in animals, by making a chemically stable 3-alkylaminopyrazinone bioisostere for its 3-sulfonylaminopyridinone core. Compound 3 (L-375,378), the closest aminopyrazinone analogue of 1, has comparable selectivity and slightly decreased efficacy but significantly improved pharmacokinetics in rats, dogs, and monkeys to 1. We have developed an efficient and versatile synthesis of 3, and this compound has been chosen for further preclinical and clinical development.

Vampire bat salivary plasminogen activator promotes robust lysis of plasma clots in a plasma milieu without causing fluid phase plasminogen activation.

Vampire bat salivary plasminogen activator (BatPA), human tissue-type plasminogen activator (tPA) or streptokinase (SK) were incubated in human citrated plasma containing a plasma clot that was radiolabelled with iodine-125 fibrin(ogen). Complete clot dissolution by BatPA (30 nM) was associated with slight activation of "fluid phase" plasminogen; the plasma levels of functional fibrinogen and alpha 2-antiplasmin decreased by only 8 and 19%, respectively. Addition of SK (3,600 IU/ml) to the clot-containing plasma caused complete clot lysis and massive activation of the "fluid phase" plasminogen, leading to greater than 60 and 96% decreases of the functional levels of fibrinogen and alpha 2-antiplasmin, respectively. Incubation of tPA (30 nM) in clot-containing plasma caused complete clot lysis as well as substantial activation of "fluid phase" plasminogen; the plasma levels of functional fibrinogen and alpha 2-antiplasmin decreased by 45 and 79%, respectively. The profound degradation of fibrinogen in the SK and tPA but not BatPA-containing samples was confirmed by immunoblot analysis. Additional experiments showed that the presence of soluble clot lysate in plasma containing tPA enhanced the extent of fibrinogen degradation from 25% to greater than 60%; the addition of soluble clot lysate to the plasma containing BatPA did not prompt further fibrinogen degradation. Finally, studies using exogenous alpha 2-antiplasmin suggested that plasmin generated via tPA-mediated activation of "fluid phase" plasminogen does not play an important role in clot dissolution.

Assessment of thrombin inhibitor efficacy in a novel rabbit model of simultaneous arterial and venous thrombosis.

The importance of thrombin in arterial and venous thrombosis renders thrombin inhibition an important therapeutic target. Identification of novel inhibitors requires an appropriate animal model. We modified a previously reported rat arterial thrombosis model to allow simultaneous assessment of the arterial and venous antithrombotic efficacies of heparin, hirudin, hirulog, a novel thrombin inhibitor H-(N-Me-D-Phe)-Pro-L-trans-4-aminocyclohexyl-Gly-[CO-CO]-NHCH3+ ++ (L-370,518) and the factor Xa inhibitor tick anticoagulant peptide in rabbits. Thrombosis was induced through application of 70% ferric chloride to the femoral artery and jugular vein. Incidence of occlusion, thrombus weight, aPTT and plasma inhibitor concentrations were determined. Heparin was efficacious in preventing arterial and venous occlusive thrombosis but at a dose that profoundly elevated aPTT. On a molar dosing basis, the approximate order of potency of the thrombin and factor Xa inhibitors was similar in artery and vein: hirudin>tick anticoagulant peptide>hirulog> or =L-370,518. Data suggested that compounds tended to be more potent in preventing venous thrombosis than arterial. This thrombin-dependent model is an economical and efficient approach to arterial and venous antithrombotic efficacy screening that eliminates variabilities encountered when multiple model/multiple animal strategies are employed.

Vampire bat salivary plasminogen activator evokes minimal bleeding relative to tissue-type plasminogen activator as assessed by a rabbit cuticle bleeding time model.

Cuticle bleeding time (CBT) measurements in anesthetized rabbits were performed to assess the potential bleeding risks which may accompany the administration of tissue-type plasminogen activator (tPA) or vampire bat salivary plasminogen activator (BatPA). The dose of BatPA or tPA used in this study, 42 nmol/kg, was previously shown to be efficacious using a rabbit femoral artery thrombosis model (Gardell et al, Circulation 84:244, 1991). CBT was determined by severing the apex of the nail cuticle and monitoring the time to cessation of blood flow. CBT was minimally elevated (1.6-fold, p = NS) following bolus intravenous administration of BatPA; in contrast, bolus intravenous administration of tPA dramatically elevated CBT (6.2-fold, p < 0.05). Rabbits treated with tPA, but not BatPA, displayed profound activation of systemic plasminogen and consequent degradation of Factor VIII and fibrinogen. Elevations in CBT after the administration of tPA were reversed by the replenishment of plasma Factor VIII activity to 40% of control, but were unaffected by complete replenishment of plasma fibrinogen. The results of this study suggest that the administration of BatPA, at a dose that promotes thrombolysis, may evoke a minimal bleeding risk, relative to an equi-efficacious dose of tPA. In addition, the tPA-provoked proteolytic consumption of Factor VIII may be a key contributor to the heightened bleeding risk.

Human erythroleukemic (HEL) cells express a platelet P2T-like ADP receptor.

Treatment of Fura-2 loaded HEL cells, a human megakaryocyte-like cell line, with P2-purinoceptor nucleotide ligands (ADP, ATP, UTP, 2-methylthio-ATP) evoked a rise in cytosolic calcium. Homologous- and cross-desensitization studies using sequential addition of nucleotides showed that the ADP-induced calcium response in HEL cells was mediated mainly by purinoceptor(s) for which ADP but not ATP was an agonist. There were also minor contributions from purinoceptors for which ATP and ADP are both agonists (probably P2U and P2Y). ATP inhibited the ADP-induced calcium response in HEL cells. This inhibition was overcome by raising the ADP concentration, thus indicating that ATP was an antagonist for the HEL cell ADP receptor. AMP was also an antagonist, albeit weak, for the HEL cell ADP receptor. Antagonism of the ADP-induced calcium response by ATP was similarly observed in MEG-01 cells, another human megakaryocyte-like cell line, but not in 293 cells, a nonhematopoietic cell line. These studies suggest that HEL cells express an ADP receptor for which ATP and AMP are antagonists. This characteristic of the HEL cell ADP receptor is also displayed by the platelet P2T receptor. Thus, HEL cells appear to be a surrogate source of the platelet ADP receptor.

Effect of the small molecule plasminogen activator inhibitor-1 (PAI-1) inhibitor, PAI-749, in clinical models of fibrinolysis.

BACKGROUND: The principal inhibitor of fibrinolysis in vivo is plasminogen activator inhibitor-1 (PAI-1). PAI-749 is a small molecule inhibitor of PAI-1 with proven antithrombotic efficacy in several preclinical models. OBJECTIVE: To assess the effect of PAI-749, by using an established ex vivo clinical model of thrombosis and a range of complementary in vitro human plasma-based and whole blood-based models of fibrinolysis. METHODS: In a double-blind, randomized, crossover study, ex vivo thrombus formation was assessed using the Badimon chamber in 12 healthy volunteers during extracorporeal administration of tissue-type plasminogen activator (t-PA) in the presence of PAI-749 or control. t-PA-mediated lysis of plasma clots and of whole blood model thrombi were assessed in vitro. The role of vitronectin was examined by assessing lysis of fibrin clots generated from purified plasma proteins. RESULTS: There was a dose-dependent reduction in ex vivo thrombus formation by t-PA (P < 0.0001). PAI-749 had no effect on in vitro or ex vivo thrombus formation or fibrinolysis in the presence or absence of t-PA. Inhibition of PAI-1 with a blocking antibody enhanced fibrinolysis in vitro (P < 0.05). CONCLUSIONS: Despite its efficacy in a purified human system and in preclinical models of thrombosis, the current study suggests that PAI-749 does not affect thrombus formation or fibrinolysis in a range of established human plasma and whole blood-based systems.

Effect of tiplaxtinin (PAI-039), an orally bioavailable PAI-1 antagonist, in a rat model of thrombosis.

OBJECTIVE: To assess the antithrombotic and profibrinolytic effects of tiplaxtinin (PAI-039), an orally bioavailable antagonist of PAI-1, in rat models of thrombosis. METHODS AND RESULTS: Carotid artery and vena cava vascular injury was produced by application of FeCl3 and blood flow was monitored using ultrasonic technology. To assess efficacy in a thrombosis prevention paradigm, PAI-039 was administered orally 90 min before injury (1-30 mg kg(-1)). To assess efficacy in a thrombosis treatment paradigm, vascular injury and stable thrombus formation were followed 4 h later by recovery and PAI-039 administration. PAI-039 prevented carotid artery occlusion in 20, 68 and 60% of animals pretreated with 0.3, 1.0 and 3.0 mg kg(-1), respectively. Time to occlusive thrombosis was increased from 18.2 +/- 4.6 min in controls to 32.5 +/- 8.7 (P = ns), 46.1 +/- 7.0 (P < 0.05), and 41.6 +/- 11.3 min (P < 0.05) in the respective PAI-039 treatment groups. In the vena cava protocol, PAI-039 pretreatment significantly reduced thrombus weight at PAI-039 doses of 3, 10 and 30 mg kg(-1). When PAI-039 was dosed in a treatment paradigm 4 h after stable arterial and venous thrombosis, a significant reduction in thrombus weight was observed 24 h later at PAI-039 doses of 3, 10 and 30 mg kg(-1). PAI-039 (10, 30 and 100 mg kg(-1)) had no effect on platelet aggregation in response to ADP or collagen and was not associated with increased bleeding or prolonged prothrombin time. In animals bearing no vascular injury, PAI-039 had no effect on circulating, low-levels of PAI-1 activity. In contrast, circulating PAI-1 activity increased 5-fold following the induction of vascular injury, which was completely neutralized by PAI-039. CONCLUSIONS: PAI-039 exerts antithrombotic efficacy in rat models of arterial and venous vascular injury without effecting platelet aggregation.

The search for the ideal thrombolytic agent: maximize the benefit and minimize the risk.

Streptokinase, acylated plasminogen streptokinase complex, and tissue-type plasminogen activator (tPA) are widely used for the treatment of acute myocardial infarction. These thrombolytic agents restore blood flow in occluded coronary arteries, salvage myocardial function, and decrease mortality. The success of thrombolytic therapy depends on the prompt and stable recanalization of the occluded blood vessel. Hemorrhagic events undermine the safety of fibrinolytic agents. Most bleeding complications occur at vascular puncture sites and can be avoided by limiting invasive procedures; however, an increased incidence of devastating intracranial bleeds is also encountered. The cause of thrombolytic-induced intracranial hemorrhage is unknown but may be due to lysis of protective hemostatic plugs. Another potential contributor to bleeding is activation of circulating plasminogen by the exogenous plasminogen activator to yield fluid-phase plasmin. Plasmin is a promiscuous proteinase that degrades many proteins that play pivotal roles in hemostasis and the integrity of the microvasculature. Tissue-type plasminogen activator activity is depressed in the absence of fibrin; hence, its use as a thrombolytic agent was anticipated to avoid the untoward consequences of plasminemia. However, the administration of pharmacologic doses of tPA can result in the activation of appreciable quantities of circulating plasminogen. In contrast to currently available thrombolytic agents, the activity of the vampire bat salivary plasminogen activator (BatPA) displays a strict requirement for the presence of fibrin. Assessment of BatPA using animal models of arterial thrombosis has demonstrated efficacy but without activation of systemic plasminogen. The avoidance of plasminemia when using BatPA as a fibrinolytic agent may circumvent bleeding complications that can compromise the safety of thrombolytic therapy.

Vampire bat salivary plasminogen activator exhibits a strict and fastidious requirement for polymeric fibrin as its cofactor, unlike human tissue-type plasminogen activator. A kinetic analysis.

The vampire bat salivary plasminogen activator (BatPA) is virtually inactive toward Glu-plasminogen in the absence of a fibrin-like cofactor, unlike human tissue-type plasminogen activator (tPA) (the kcat/Km values were 4 and 470 M-1 s-1, respectively). In the presence of fibrin II, tPA and BatPA activated Glu-plasminogen with comparable catalytic efficiencies (158,000 and 174,000 M-1 s-1, respectively). BatPA's cofactor requirement was partially satisfied by polymeric fibrin I (54,000 M-1 s-1), but monomeric fibrin I was virtually ineffective (970 M-1 s-1). By comparison, a variety of monomeric and polymeric fibrin-like species markedly enhanced tPA-mediated activation of Glu-plasminogen. Fragment X polymer was 2-fold better but 9-fold worse as cofactor for tPA and BatPA, respectively, relative to fibrin II. Fibrinogen, devoid of plasminogen, was a 10-fold better cofactor for tPA than fibrinogen rigorously depleted of plasminogen, Factor XIII, and fibronectin; the enhanced stimulatory effect of the less-purified fibrinogen was apparently due to the presence of Factor XIII. By contrast, the two fibrinogen preparations were equally poor cofactors of BatPA-mediated activation of Glu-plasminogen. BatPA possessed only 23 and 4% of the catalytic efficiencies of tPA and two-chain tPA, respectively, in hydrolyzing the chromogenic substrate Spectrozyme tPA. However in the presence of fibrin II, BatPA and tPA exhibited similar kcat/Km values for the hydrolysis of Spectrozyme tPA. Our data revealed that BatPA, unlike tPA, displayed a strict and fastidious requirement for polymeric fibrin I or II. Consequently, BatPA may preferentially promote plasmin generation during a narrow temporal window of fibrin formation and dissolution.

Effects of bile acids and their glycine conjugates on gamma-glutamyl transpeptidase.

Glycine and taurine conjugates of bile acids modulate gamma-glutamyl transpeptidase by interacting with the cysteinylglycine binding site (acceptor site) of the enzyme. These compounds stimulate hydrolysis of glutamine and S-methylglutathione and the rate of the inactivation of the enzyme by the gamma-glutamyl site-directed reagent, AT-125 (L-(alpha S,5S)-alpha-amino-3-chloro-4,5-dihydro-5-isoxazoleacetic acid). Transpeptidation between S-methylglutathione and methionine was inhibited by these compounds. These effects resemble those caused by hippurate; the glycine derivatives of bile acids, however, exhibit a much greater affinity for transpeptidase than hippurate. Cholate, as shown previously for benzoate, also seems to bind to a portion of the acceptor site as indicated by its effects on S-methylglutathione utilization and AT-125-dependent inactivation of the enzyme. The Kd values for cholate and benzoate are, however, at least one order of magnitude larger than those for their respective glycine derivatives. The acceptor site-directed modulators increase the affinity of the enzyme for AT-125 and kinetic and binding studies show that binding of gamma-glutamyl site-directed reagents increases the affinity of the enzyme for cholate. These results thus indicate cooperative interactions between the gamma-glutamyl donor and acceptor binding domains of the transpeptidase active center.

Latent proteinase activity of gamma-glutamyl transpeptidase light subunit.

gamma-Glutamyl transpeptidase, which is composed of two unequal subunits, exhibits proteinase activity when treated with agents such as urea and sodium dodecyl sulfate. The heavy subunit is preferentially and rapidly degraded. The enzyme also degraded bovine serum albumin in the presence of urea; however, several other proteins and model proteinase substrates were not cleaved. Treatment of the enzyme with 6-diazo-5-oxo-L-norleucine, a gamma-glutamyl analog, results in parallel loss of transpeptidase and proteinase activities indicating that the site at which gamma-glutamylation of the enzyme occurs (presumably a hydroxyl group on the light subunit) is also involved in proteinase activity. The purified light subunit, but not the heavy subunit, exhibits proteinase activity even in the absence of urea. Results suggest that dissociation of the enzyme unmasks the proteinase activity of the light subunit involving the site at which gamma-glutamylation of the enzyme occurs, and that the heavy subunit may impose transpeptidase reaction specificity by contributing the binding domains for gamma-glutamyl substrates.

Combinatorial mutagenesis of the reactive site region in plasminogen activator inhibitor I.

Plasminogen activator inhibitor (PAI-I) rapidly inactivates tissue plasminogen activator (t-PA) and urokinase (UK) with nearly identical association rate constants. The contributions of Ser344, Ala345, and Arg346 (P3, P2, and P1 residues, respectively) in PAI-I to inhibition of UK and t-PA were evaluated using combinatorial mutagenesis of the human PAI-I cDNA. A bacteriophage lambda expression library potentially encoding the 8000 unique PAI-I species were screened for inhibitory activity against UK using a fibrin indicator gel. 390 plaques demarcated by zones of retarded fibrinolysis were analyzed to determine the DNA sequences of their associated active PAI-1 species. We found 134 unique PAI-1 variants that retained inhibitory activity towards UK; they contained a variety of amino acids in their P3 and P2 positions but only Arg or, infrequently, Lys in their P1 position. Each of the unique active PAI-1 were assayed for inhibitory activity towards UK or t-PA; many substitutions differentially affected the ability of the inhibitor to inactivate UK and t-PA. For example, replacement of Ser344 and Ala344 with Val and Pro, respectively, yielded a PAI-1 variant exhibiting an association rate constant that was unchanged for t-PA but decreased 23-fold for UK, relative to native PAI-1. In general, the PAI-1 variants were more potent inhibitors of t-PA than UK. Hence, t-PA appears more tolerant than UK of structural diversity present in the P3 and P2 positions of the PAI-1 variants.