Highly selective mechanism-based thrombin inhibitors: structures of thrombin and trypsin inhibited with rigid peptidyl aldehydes

The crystal structures of three highly potent and selective low-molecular weight rigid peptidyl aldehyde inhibitors complexed with thrombin have been determined and refined to R values 0.152-0. 170 at 1.8-2.1 A resolution. Since the selectivity of two of the inhibitors was >1600 with respect to trypsin, the structures of trypsin-inhibited complexes of these inhibitors were also determined (R = 0.142-0.157 at 1.9-2.1 A resolution). The selectivity appears to reside in the inability of a benzenesulfonamide group to bind at the equivalent of the D-enantiomorphic S3 site of thrombin, which may be related to the lack of a 60-insertion loop in trypsin. All the inhibitors have a novel lactam moiety at the P3 position, while the two with greatest trypsin selectivity have a guanidinopiperidyl group at the P1 position that binds in the S1 specificity site. Differences in the binding constants of these inhibitors are correlated with their interactions with thrombin and trypsin. The kinetics of inhibition vary from slow to fast with thrombin and are fast in all cases with trypsin. The kinetics are examined in terms of the slow formation of a stable transition-state complex in a two-step mechanism. The structures of both thrombin and trypsin complexes show similar well-defined transition states in the S1 site and at the electrophilic carbon atom and Ser195OG. The trypsin structures, however, suggest that the first step in a two-step kinetic mechanism may involve formation of a weak transition-state complex, rather than binding dominated by the P2-P4 positions.     

Potent and selective thrombin inhibitors featuring hydrophobic, basic P3-P4-aminoalkyllactam moieties

Crystal structure and evolving SAR considerations of potent, selective benzylsulfonamide lactam thrombin inhibitors and related serine protease inhibitors have led to the design of novel thrombin inhibitors 1a-g, featuring hydrophobic, basic, P4-alkylaminolactam scaffolds that serve as novel types of P3-P4 dipeptide mimics. The design, synthesis, and biological activity of these targets is presented.     

Benzylamine-based selective and orally bioavailable inhibitors of thrombin

A series of p-aminomethylphenylalanine derivatives were investigated as novel thrombin inhibitors. This study led to potent inhibitors of thrombin (Ki up to 3.3 nM) that are trypsin-selective, highly orally bioavailable in rats, and highly permeable across Caco-2 cells. The P1 benzylamine binding mode in the thrombin active site was identified by X-ray crystallographic analysis.     

Identification of tissue-type plasminogen activator-specific plasminogen activator inhibitor-1 mutants. Evidence that second sites of interaction contribute to target specificity

Plasminogen activator inhibitor-1 (PAI-1) is the primary inhibitor of the plasminogen activators (PAs), tissue-type plasminogen activator (tPA), and urokinase-type plasminogen activator (uPA). A library of PAI-1 mutants containing substitutions at the P1 and P1' positions was screened for functional activity against tPA and thrombin. Several PAI-1 variants that were inactive against uPA in a previous study (Sherman, P. M., Lawrence, D. A., Yang, A. Y., Vandenberg, E. T., Paielli, D., Olson, S. T., Shore, J. D., and Ginsburg, D. (1992) J. Biol. Chem. 267, 7588-7595) had significant inhibitory activity toward tPA. This set of tPA-specific PAI-1 mutants contained a wide range of amino acid substitutions at P1 including Asn, Gln, His, Ser, Thr, Leu, Met, and all the aromatic amino acids. This group of mutants also demonstrated a spectrum of substitutions at P1'. Kinetic analyses of selected variants identified P1Tyr and P1His as the most efficient tPA-specific inhibitors, with second-order rate constants (ki) of 4.0 x 10(5) M-1s-1 and 3.6 x 10(5) M-1s-1, respectively. Additional PA-specific PAI-1 variants containing substitutions at P3 through P1' were constructed. P3Tyr-P2Ser-P1Lys-P1'Trp and P3Tyr-P2Ser-P1Tyr-P1'Met had ki values of 1.7 x 10(6) M-1s-1 and 2.5 x 10(6) M-1s-1 against tPA, respectively, but both were inactive against uPA. In contrast, P2Arg-P1Lys-P1'Ala inhibited uPA 74-fold more rapidly than tPA. The mutant PAI-1 library was also screened for inhibitory activity toward thrombin in the presence and absence of the cofactor heparin. While wild-type PAI-1 and several P1Arg variants inhibited thrombin in the absence of heparin, a number of variants were thrombin inhibitors only in the presence of heparin. These results demonstrate the importance of the reactive center residues in determining PAI-1 target specificity and suggest that second sites of interaction between inhibitors and proteases can also contribute to target specificity. Finally, the PA-specific mutants described here should provide novel reagents for dissecting the physiological role of PAI-1 both in vitro and in vivo.     

Novel amidine-containing peptidyl phosphonates as irreversible inhibitors for blood coagulation and related serine proteases

A series of new peptidyl (alpha-aminoalkyl)phosphonate diphenyl esters containing the 4-amidinophenyl group were synthesized and tested as irreversible inhibitors for thrombin and other trypsin-like enzymes. These phosphonates irreversibly inhibited several coagulation enzymes and trypsin. Boc-D-Phe-Pro-(4-AmPhGly)P(OPh)2 is the best human thrombin inhibitor in the series with a k(obs)/[I] value of 11,000 M-1 s-1, and it inhibits thrombin more than 5-fold more effectively than the other enzymes tested. Z-(4-AmPhGly)P(OPh)2 is the best inhibitor for plasma kallikrein with a k(obs)/[I] value of 18,000 M-1 s-1. Generally, the (4-AmPhGly)P(OPh)2 derivatives are better inhibitors of thrombin and trypsin than the corresponding (4-AmPhe)P(OPh)2 derivatives which contain an extra CH2 separating the amidinophenyl group from the peptide backbone. The amidino phosphonates did not inhibit acetylcholinesterase and were chemically stable in neutral buffers. In addition, the inhibited trypsin derivative did not regain any enzyme activity after removal of excess inhibitor and incubation in a pH 7.5 buffer for 1 day. Boc-D-Phe-Pro-(4-AmPhGly)P(OPh)2 and D-Phe-Pro-(4-AmPhe)P(OPh)2 prolonged the prothrombin time ca. 2-fold and prolonged the activated partial thromboplastin time ca. 3-4-fold in human plasma at concentrations of 63 and 125 microM, respectively. The novel amidine-containing peptidyl phosphonates reported here are thus effective anticoagulants in vitro, and they may have utility for use in vivo.     

The importance of enzyme inhibition kinetics for the effect of thrombin inhibitors in a rat model of arterial thrombosis

The relation between the antithrombotic effect in vivo, and the inhibition constant (Ki) and the association rate constant (k(on)) in vitro was investigated for eight different thrombin inhibitors. The carotid arteries of anaesthetized rats were exposed to FeCl3 for 1 h, and the thrombus size was determined from the amount of incorporated 125I-fibrinogen. The thrombin inhibitors were given intravenously, and complete concentration- and/or dose-response curves were constructed. Despite a 50,000-fold difference between the Ki-values comparable plasma concentrations of hirudin and melagatran were needed (0.14 and 0.12 micromol l(-1), respectively) to obtain a 50% antithrombotic effect (IC50) in vivo. In contrast, there was a comparable in vitro (Ki-value) and in vivo (IC50) potency ratio for melagatran and inogatran, respectively. These results can be explained by the concentration of thrombin in the thrombus and improved inhibition by the low-molecular-weight compounds. For all eight thrombin inhibitors tested, there was an inverse relationship between k(on)-values in vitro and the slope of the dose response curves in vivo. Inhibitors with k(on)-values of < 1 x 10(7) M(-1) s(-1) gave steep dose response curves with a Hill coefficient > 1. The association time for inhibition of thrombin for slow-binding inhibitors will be too long to give effective antithrombotic effects at low plasma concentrations, but at increasing concentrations the association time will decrease, resulting in a steeper dose-response curve and thereby a more narrow therapeutic interval.     

Diarylsulfonamides as selective, non-peptidic thrombin inhibitors

Based on the structures of aminopyridine thrombin inhibitors (1), a series of aminoalkyl- and guanidinoalkyl-substituted diarylsulfonamides were prepared. The most potent derivative, N-[3-(4-guanidinobutoxy)-5-methyl-phenyl]-benzenesulfonamide (6c) had Ki = 0.18 microM for thrombin and did not inhibit trypsin, plasmin, or factor Xa. Comparison of the X-ray structures of the thrombin/1b and the thrombin/6c complexes revealed important aspects which govern the binding of such diarylsulfonamides to thrombin.     

Structural modification of an orally active thrombin inhibitor, LB30057: replacement of the D-pocket-binding naphthyl moiety

An amidrazonophenylalanine derivative LB30057 (2) was identified as a potent (Ki = 0.38 nM), selective, and orally active thrombin inhibitor. As a continuation of studies into benzamidrazone-based thrombin inhibitors, we have structurally modified compound 2 by replacing the naphthyl group with a variety of hydrophobic moieties. This study led to discovery of several compounds with significantly enhanced potency in thrombin inhibition without sacrificing selectivity against trypsin and oral absorption. The highest activity was obtained with compound 23 (Ki = 0.045 nM).     

Inhibition of thrombin-induced contractile responses by protein kinase inhibitors in porcine pulmonary arteries

The clotting enzyme thrombin is known to cause receptor-mediated contractile effects in isolated blood vessels. In the present studies the influence of protein kinase inhibitors on the contractile response of porcine pulmonary arteries to thrombin (3 U/ml) was investigated. Endothelium-denuded rings (2-3 mm) from small arteries were placed in organ baths for isometric tension recording. The vessels were preincubated for 30 min with the inhibitors before inducing contractions. In the presence of the protein kinase C (PKC)-inhibitors staurosporine, BIM I (bisindolyl-maleimide I), chelerythrine and Ro 31-8220 (1 microM each), the contractile responses to the PKC activator phorbol 12,13-dibutyrate (PDBu; 50 nM) were diminished by 70-100%. However, for inhibition of thrombin-induced contractions generally higher concentrations of the inhibitors were required. Only staurosporine at 1 microM inhibited the thrombin effect by about 75%. The tyrosine kinase inhibitor erbstatin (30 microM) did not significantly alter the thrombin effect, whereas genistein at 10 microM caused a significant inhibition of contractile responses to both thrombin and PGF2alpha. At 100 microM genistein also inhibited the contractile effects of PdBu and KCl. These studies suggest that activation of both PKC and non-receptor tyrosine kinases seems to be involved in the signal transduction pathways of thrombin-induced contractile effects in isolated vessels.     

Potent and selective bicyclic lactam inhibitors of thrombin: Part 2: P1 modifications

The synthesis and antithrombotic activity of a series of nonpeptide bicyclic thrombin inhibitors is described. We have explored the SAR with modifications to the P1 site. The introduction of arginine mimetics at the P1 site led to potent and selective thrombin inhibitors.     

The venous antithrombotic profile of naroparcil in the rabbit

The venous antithrombotic profile of naroparcil or (4-[4-cyanobenzoyl]-phenyl)-1.5-dithio-beta-D-xylopyranoside was investigated in the rabbit following single i.v. and oral administration. Naroparcil attenuated thrombus development in a Wessler stasis model of venous thrombosis (jugular vein) employing bovine factor Xa as a thrombogenic stimulus giving ED50 values of 21.9 mg/kg and 36.0 mg/kg after respectively i.v. and oral administration. Venous antithrombotic activity was maximal 2-3 h after i.v. administration and 4-8 h after oral administration. Four hours after the oral administration of maximal antithrombotic (Wessler model, factor Xa) doses (100 and 400 mg/kg), naroparcil had no significant effect on bleeding time. In platelet poor plasma obtained from animals treated 4 h previously with various doses (25 to 400 mg/kg) of naroparcil, there was no detectable anti-factor Xa nor antithrombin activity. Similarly, naroparcil had no effect on APTT nor on thrombin time. A sensitized thrombin time (to about 35 s) was modestly but significantly increased following oral administration of the compound at 400 mg/kg. However, thrombin generation by the intrinsic pathway was reduced in a dose-related manner, maximal reduction being 65% at 400 mg/kg. The same dose of naroparcil enhanced the formation of thrombin/heparin cofactor II complexes at the expense of thrombin/antithrombin III complexes in plasma incubated with (125I)-human alpha-thrombin and induced the appearance of dermatan sulfate-like material in the plasma of treated rabbits, as measured by a heparin cofactor II-mediated thrombin inhibition assay.(ABSTRACT TRUNCATED AT 250 WORDS)     

Studies on 5-lipoxygenase inhibitors. II. Discovery, optical resolution and enantioselective synthesis of FR110302, a highly potent non-redox type 5-lipoxygenase inhibitor

A novel series of 2,2-dialkyl-5-(2-quinolylmethoxy)-1,2,3, 4-tetrahydro-1-naphthols was synthesized and evaluated as 5-lipoxygenase (5-LO) inhibitors. Systematic optimization led to identification of several highly potent non-redox type 5-LO inhibitors with nanomolar IC50s as racemic mixtures. Optical resolution of racemate 50 indicated that its 5-LO inhibitory activity was enantiospecific and due to the (+)-enantiomer. An efficient synthetic route to the (+)-enantiomers via asymmetric reduction of tetralone intermediates was established. The best compound, (+)-2,2-dibutyl-5-(2-quinolylmethoxy)-1,2,3,4-tetrahydro-1-naphtho l (FR110302, (+)-50), showed potent inhibitory activity against leukotriene (LT) biosynthesis by intact neutrophiles in rats (IC50 4.9 nM) and in humans (IC50 40 nM). Furthermore oral administration of FR110302 significantly inhibited neutrophil migration in the rat air pouch model at 1 mg/kg.     

5,5-trans lactone-containing inhibitors of serine proteases: identification of a novel, acylating thrombin inhibitor

Synthesis of a variety of 5,5-trans fused lactones, related to compounds found in extracts of Lantana camara, has provided a series of novel acylating inhibitors of human thrombin, trypsin, chymotrypsin and human leucocyte elastase. The most effective thrombin inhibitor is 7 with an IC50 of 130 nM and a Kobs/[1] of 4,000 M-1 s-1.     

Inhibition of inositol(1,4,5)Trisphosphate generation by endothelin-1 during postischemic reperfusion: A novel antiarrhythmic mechanism

BACKGROUND: Reperfusion of ischemic rat hearts in the presence of thrombin or norepinephrine but not endothelin-1 causes the generation of inositol 1,4,5-trisphosphate (Ins 1,4,5P3) and arrhythmias. The present study investigates the effect of endothelin-1 on these responses. METHODS AND RESULTS: Ins 1,4,5P3 generation was quantified by use of [3H] labeling and high-performance liquid chromatography as well as by mass analysis. Twenty minutes of global ischemia followed by 2 minutes of reperfusion increased [3H]Ins 1,4,5P3 from 2828+/-265 to 5033+/-650 cpm/g tissue in the presence of thrombin 2.5 IU/mL and to 4561+/-286 cpm/g tissue in response to release of norepinephrine (n=4, P<0.01) in both cases. Reperfusion in the presence of endothelin-1 alone caused no change in Ins 1,4,5P3 (2762+/-240 cpm/g tissue), but when added together with thrombin or norepinephrine, endothelin-1 reduced the Ins 1,4,5P3 responses to 2313+/-197 and 1764+/-168 cpm/g tissue, respectively (n=4, P<0.01 in both cases). Similar inhibitory interactions between endothelin-1 10 nmol/L and thrombin 2.5 IU/mL were observed under normoxic conditions in nonperfused ventricle, eliminating the possibility that excessive vasoconstriction was responsible. In parallel studies, endothelin-1 suppressed the development of reperfusion arrhythmias initiated by either thrombin (ventricular fibrillation, 75% to 39%, n=16 to 18) or norepinephrine (83% to 8%, n=12 to 22) (P<0.01 in both cases). CONCLUSIONS: Inhibition of Ins 1,4,5P3 generation during myocardial reperfusion by endothelin-1 represents a novel antiarrhythmic mechanism.     

The ability of thrombin inhibitors to reduce the thrombin activity generated in plasma on extrinsic and intrinsic activation

In a thrombin generation test with continuous registration of thrombin activity in plasma we studied the ability of a variety of thrombin inhibitors of different type and mechanism of action of influence the activity of thrombin after activation of the coagulation system. Depending on the inhibitor, the peak of thrombin activity is delayed and/or reduced. By blocking the active site of generated thrombin inhibitors cause a concentration dependent reduction of the thrombin peak and inhibit feed-back reactions of thrombin resulting in a delay of thrombin generation. Highly potent synthetic active-site directed inhibitors (Ki < or = 20 nM) reduce the thrombin activity formed in plasma after extrinsic or intrinsic activation with the same efficiency (IC50 0.1-0.6 microM) as hirudin. The delay and reduction of thrombin generation by inhibitors of the anion-binding exosite 1 of thrombin is only attributed to an inhibition of feed-back reactions of thrombin. For a 50% reduction of thrombin activity in plasma by this type of inhibitors relatively high concentrations were determined.     

Protein engineering of chimeric Serpins: an investigation into effects of the serpin scaffold and reactive centre loop length

The exposed Serpin reactive centre loop controls the specificity of the serpin proteinase interaction. Mutations within this region have been used to generate novel potentially therapeutic inhibitors. In this study we examine the effect of the serpin scaffold and reactive centre loop length upon the generation of such inhibitors. The reactive centre loop regions, P7-P3', of alpha1-antitrypsin and alpha1-antichymotrypsin were replaced by the corresponding residues of the viral serpin, Serp1, to form AT/Serp1 and ACT/Serp1, respectively. AT/Serp1 formed SDS stable complexes with a range of proteinases with association rate constants for plasmin, tissue plasminogen activator, urokinase, thrombin and factor Xa of approximately 10(4) M(-1)s(-1) and a stoichiometry of inhibition of approximately 1 for all of them. ACT/Serp1, however, formed SDS-stable complexes with only plasmin and thrombin with association rate constant 100-fold slower than AT/Serp1 and an increased stoichiometry of inhibition. The reactive centre loop of ACT/Serp1 is four amino acid residues longer than AT/Serp1. These four additional residues (VETR) were inserted into AT/Serp1 to form AT/Serp1(VETR). AT/Serp1(VETR) formed SDS stable complexes with plasmin, thrombin and tissue plasminogen activator similar to AT/Serp1, however, the association rate constants were 10-fold slower than those observed with AT/Serp1, while the stoichiometry of inhibition remained around 1. These results suggest that the additional reactive centre loop residues effect the rate of initial complex formation by placing the reactive centre loop in a non-ideal conformation. This study demonstrates that both reactive centre loop length and serpin scaffold are important in defining the inhibitory characteristics of a serpin.     

Heterogeneity of human platelet density subpopulations in aggregation, secretion of ATP, and cytosolic-free calcium concentration

AIM: To investigate thrombin (500 U.L-1)-, ADP (0.1-30 mumol.L-1)-, and 5-hydroxytryptamine (5-HT, 3 mumol.L-1)-induced aggregation, secretion of ATP and cytosolic-free calcium mobilization in density subpopulations of human washed platelets. METHODS: Using Percoll discontinuous gradient. RESULTS: The human platelets were separated into high density (HD), intermediate density (ID), and low density (LD) subpopulations, and their sizes were diminished with decreasing density (r = 0.978, P < 0.01). The magnitude of aggregations by thrombin, ADP, and 5-HT was more significant in HD platelets than that in LD platelets (P < 0.01). The amount of secretion of ATP induced by thrombin and ADP in HD platelets was also much higher than that in LD platelets (P < 0.01), except for 5-HT which did not cause the ensuring release reaction in any subpopulation of human platelets. Thrombin (1500 U.L-1)-, ADP (mumol.L-1)-, and 5-HT (3 mumol.L-1)-induced cytosolic-free calcium mobilization was evaluated as well. Results showed that the resting level of cytosolic-free calcium concentration ([Ca2+]i) was the same in all subpopulations, about 80-90 nmol.L-1. However, the level of [Ca2+]i mobilization was entirely different, heightened with increasing density. CONCLUSION: The function of HD platelets was much stronger and more active than that of LD platelets in human.     

Dibasic benzo[b]thiophene derivatives as a novel class of active site directed thrombin inhibitors. 2. Exploring interactions at the proximal (S2) binding site

In an effort to increase the thrombin inhibitory activity of a novel series of inhibitors (i.e., 1a), substituents were incorporated at the C-3" position of the C-3 aryl ring (2). Consistent with the X-ray crystallography studies, small hydrophobic groups at the C-3" site (Br and Me) enhanced thrombin inhibitory activity by 8-fold. However, a few more hydrophilic substituents (NO2 and OMe) also enhanced the potency of the series. The biological results are discussed in terms of molecular modeling studies.     

Vasoflux, a new anticoagulant with a novel mechanism of action

BACKGROUND: Heparin and direct thrombin inhibitors, such as hirudin, have limitations in the treatment of acute coronary syndromes. Heparin does not inactivate fibrin-bound thrombin, whereas hirudin fails to block thrombin generation. In contrast, Vasoflux is a novel anticoagulant that inactivates fibrin-bound thrombin and attenuates factor Xa generation. METHODS AND RESULTS: Vasoflux is prepared by depolymerization of heparin, restricting molecular size to between 3000 and 8000 Da, and reducing antithrombin affinity by periodate oxidation. Vasoflux catalyzes fibrin-bound thrombin inactivation by heparin cofactor II (HCII) and inhibits factor IXa activation of factor X independently of antithrombin and HCII. Compared with other anticoagulants in a thrombogenic extracorporeal circuit, Vasoflux maintains filter patency at concentrations that produce an activated clotting time (ACT) of 220 seconds. In contrast, to maintain filter patency, heparin, low-molecular-weight heparin (LMWH), and hirudin require concentrations that produced an ACT of 720, 415, and >1500 seconds, respectively, whereas dermatan sulfate was ineffective at concentrations that produced an ACT of 360 seconds. CONCLUSIONS: Vasoflux is more effective than heparin and LMWH because it inactivates fibrin-bound thrombin and is superior to hirudin and dermatan sulfate because it also blocks factor Xa generation.     

Inhibition of thrombin generation in plasma by inhibitors of factor Xa

A series of inhibitors of factor Xa (FXa) were investigated using the thrombin generation assay to evaluate the potency and specificity needed to efficiently block thrombin generation in activated human plasma. By inhibiting FXa the generation of thrombin in plasma is delayed and decreased. Inhibitor concentrations which cause 50 percent inhibition of thrombin generation (IC50) correlate in principle with the Ki values for inhibition of free FXa. Recombinant tick anticoagulant peptide (r-TAP) is able to inhibit thrombin generation with considerably low IC50 values of 49 nM and 37 nM for extrinsic and intrinsic activation, respectively. However, the potent synthetic, low molecular weight inhibitors of FXa (Ki values of about 20 nM) are less effective in inhibiting the generation of thrombin with IC50 values at micromolar concentrations. The overall effect of inhibitors of FXa in the thrombin generation assay was compared to that of thrombin inhibitors. On the basis of similar Ki values for the inhibition of the respective enzyme, synthetic FXa inhibitors are less effective than thrombin inhibitors. In contrast, the highly potent FXa inhibitor r-TAP causes a stronger reduction of the thrombin activity in plasma than the most potent thrombin inhibitor hirudin.