A novel inhibitor of activated thrombin-activatable fibrinolysis inhibitor ...

© 2007 Schattauer GmbH, Stuttgart

Blood Coagulation, Fibrinolysis and Cellular Haemostasis

A novel inhibitor of activated thrombin-activatable fibrinolysis inhibitor (TAFIa) – Part I: Pharmacological characterization Yi-Xin Wang, Lei Zhao, Mariko Nagashima, Jon Vincelette, Drew Sukovich, Weiwei Li, Babu Subramanyam, Shendong Yuan, Kumar Emayan, Imadul Islam, Paul Hrvatin, Judi Bryant, David R. Light, Ronald Vergona, John Morser, Brad O. Buckman Berlex Bioscience, Richmond, California, USA

Summary We have discovered a novel small-molecule (3-phosphinoylpropionic acid) inhibitor of activated thrombin activatable fibrinolysis inhibitor (TAFIa), BX 528, which had an IC50 of 2 nM in an enzymatic assay and 50 nM in an in-vitro clot lysis assay, with 3,500- to 35,000-fold selectivity against other carboxypeptidases, such as CPN, CPZ and CPD, and 5- and 12-fold selectivity against CPE (CPH) and CPB, respectively. At 10 µM, BX 528 had no significant activity (< 50% inhibition or antagonism) in a panel of 137 enzymes and receptors. It had no effects on blood coagulation and platelet aggregation up to 300 and 10 µM, respectively. The plasma half-life following intravenous administration was 0.85 hours in rats and 4.5 hours in dogs. No signifi-

cant metabolism was detected in human, dog or rabbit hepatic microsomes, and no significant inhibition of cytochrome P450 3A4 and 2D6 up to 30 µM. No cytotoxic or cell proliferative effects were found in three hepatic and renal cell lines up to 300 µM and no mutagenic activity was seen in the Ames II screen. There were no significant hemodynamic effects in rats and dogs up to 100 and 30 mg/kg with peak plasma drug concentrations of ~1,000 and 300 µM, respectively. In an in-vivo complement activation model in guinea pigs,BX 528 showed minimal inhibition of plasma CPN activity up to 60 mg/kg with peak plasma concentrations up to 250 µM.Thus, these data demonstrate that BX 528 is a novel,potent,selective and safeTAFIa inhibitor.

Keywords TAFI, carboxypeptidases, potency, selectivity, safety, pharmacokinetics

Thromb Haemost 2007; 97: 45–53

Introduction One of the major pathologic events in thrombotic diseases is the formation of a fibrin clot, during which the endogenous fibrinolytic system is also activated. The first step in activation of the fibrinolytic system is generation of plasmin by tissue-type plasminogen activator (t-PA), thus, initiating clot lysis by proteolytic cleavage of internal lysine residues in the Aα-chain of fibrin. Thrombin-activatable fibrinolysis inhibitor (TAFI), also known as plasma carboxypeptidase B, carboxypeptidase R and carboxypeptidase U is an endogenous inhibitor of plasminogen activation that removes the binding sites for plasminogen and t-PA (1, 2). When cleaved by thrombin, plasmin or by the thrombin/ thrombomodulin complex, TAFI zymogen is converted to activated TAFI (TAFIa). Accumulating evidence shows that the elevation of TAFIa levels plays an important pathogenic role in thrombosis formation and in retardation of endogenous and exogenous thrombolysis in both animal models and patients with thrombotic diseases (1, 3–5). Therefore, inhibition of TAFIa is a Correspondence to: Yi-Xin (Jim) Wang Berlex Biosciences 2600 Hilltop Drive Richmond, CA 948047, USA Tel.: +1 510 669 4489, Fax: +1 510 669 4247 E-mail: [email protected]

promising therapeutic target for treatment of thrombotic diseases by enhancing both endogenous and exogenous thrombolysis. Recently, several groups have reported on small-molecule TAFIa inhibitors (6–8). Here, we report the discovery, pharmacological profile and characterization of a novel potent, safe and highly selective TAFIa inhibitor, BX 528. Since carboxypeptidases are a large family of proteins with a broad range of physiopathological functions, the specificity of a TAFIa inhibitor is extremely important in determining its safety. Therefore, in profiling the current compound, we not only characterized its potency and selectivity, but also further tested it in an in-vivo complement activation model in guinea pigs to investigate its effects on plasma CPN inhibition and on efficacy.

Materials and methods All animal experimental procedures were approved by the Institutional Animal Care and Use Committee and were in accordance with the Guide for the Care and Use of LaboratoryAnimals (NIH). Received September 29, 2006 Accepted after resubmission November 20, 2006 Prepublished online December 8, 2006 doi:10.1160/TH06–09–0551

45

Wang et al. Pharmacological profile of a TAFIa inhibitor

Chemical structure BX 528, with the chemical structure shown in Figure 1, was synthesized as described previously (see ref. 9–11). Enzymatic assay measuring inhibition of activated TAFI and other carboxypeptidases TAFIa activity was measured in a 96-well format based on a modification of the published method (12, 13). Purified TAFI (1.67 µM) was activated with 10 nM thrombin and 50 nM thrombomodulin (TM) in 20mM HEPES pH 7.4, 150 mM NaCl, 5 mM CaCl2 for 10 minutes (min) at room temperature (RT). The activation was terminated by addition of 1 µM PPACK and stored on ice before use. The TAFIa (final concentration 9.2 nM) was incubated with BX 528 for 2 min prior to the addition of 1 mM hippuryl-L-arginine substrate. After 30 min at RT, the amount of hydrolysed substrate was determined by conversion of the product, hippuric acid, to a chromophore with 3% w/v cyanuric chloride/ dioxane in 0.2M NaHPO4 buffer. Absorbance (O.D./min x 1,000) of the supernatant was read at 382 nm. The IC50 of the compound was determined from an 8-point concentration-response curve. Other related carboxypeptidases, either human, such as CPN, CPZ, CPE (also known as CPH), or non-human, such as duck CPD and porcine CPB, were used to determine the selectivity of BX 528. The optimum pH of the pro-hormone carboxypeptidases, human CPE and CPD, occurs at an acid pH, consistent with an activity associated with secretory granules. Porcine pancreatic CPB and recombinant duck CPD were included as surrogates for their respective human enzyme counterparts. The selectivity of BX 528 was calculated as the ratio of its IC50 against other carboxypeptidases (n=2–8) compared to its IC50 against TAFIa. Human CPN was purified by Enzyme Research Laboratories (South Bend, IN, USA) by modifying a published procedure (14). Both human CPN and porcine pancreatic CPB (Sigma, St. Louis, MO, USA) were assayed in a 96-well format adapted from a published method (9, 10). Either CPN (~1.56 units) or CPB (2 nM) were incubated with TAFI inhibitor in 20 mM HEPES pH 7.4, 150 mM NaCl, 5 mM CaCl2 for 2 min prior to adding hippuryl-L-arginine substrate at concentrations near Km (0.8 mM for CPN and 0.6 mM for CPB). After incubation at RT (90 min for CPN and 30 min for CPB), the amount of hydrolysed substrate was determined by conversion of the hippuric acid product to a chromogen as described above. The purified recombinant carboxypeptidases, human CPE expressed in Sf9 cells, human CPZ expressed in AtT-20 cells (15), and duck CPD expressed in Sf9 cells were provided by Dr. Lloyd Fricker, Albert Einstein College of Medicine (Bronx, NY,

Figure 1: Chemical structure of BX 528, (S)-2-[3-(aminomethyl) phenyl]-3-{hydroxyl [(R)-2-methyl-1-{[(3-phenylpropyl) sulfonyl] amino propyl] phosphoryl propanoic acid.

46

USA). Assays were performed using an HPLC step developed to replace the chloroform extraction originally described (25). Purified CPD (1 nM) was assayed both at low and neutral pH by incubation of TAFI inhibitor in either 50 mM acetate, 0.05% BSA, pH 5.5 or 50 mM TRIS, 0.05% BSA, pH 7.5 for 5 min prior to addition of substrate. The substrate, dansyl-Phe-Ala-Arg was added to a concentration equal to Km (11 µM at pH 5.5 or 240 µM at pH 7.5) and incubated for 30 min at 37°C. Purified CPZ (0.04 or 0.3 nM) was treated with BX 528 in 50 mM TRIS, pH 7.5, 0.05% BSA for 10 min prior to dansyl-Phe-Ala-Arg addition at Km (3,000 µM) and incubated 60 min at 37°C. Purified CPE (0.5 nM) was incubated with BX 528 in 50 mM acetate, pH 5.5 for 2 min prior to dansyl-Phe-Ala-Arg addition at Km (10 µM) and incubated 10 min at RT. After the indicated incubation times all reactions were stopped by addition of TFA to 0.5%. The amount of substrate hydrolysed was determined by separation of substrate from the product, dansyl-Phe-Ala, by RP-HPLC. Integrated peaks were quantified by comparison to a standard curve of dansyl-Phe-Ala in the same concentration range. The IC50 of BX 528 was determined from an 8-point dose-response curve, tested in duplicate. In addition, the effect of BX 528 was tested at a concentration of 10 µM on 137 enzymes and receptors, including proteases and zinc metalloproteases, by MDS Pharma Services (Taipei, Taiwan). In-vitro plasma clot lysis assay As described in detail in our previous publications (16–18), a clot was rapidly formed in 30 µl plasma by addition of 25 nM thrombin. Clot lysis was initiated by adding 67 to 83 ng/ml t-PA in a total volume of 120 µl. Clot lysis time was defined as the time when turbidity, monitored by the absorbance at 405 nm, was one-half the difference between the plateau reached after clotting and the baseline value achieved at complete lysis. Thrombomodulin (TM, 1 µg/ml) was added to enhance the activation of endogenous TAFI, which leads to a dramatic increase in clot lysis time (Fig. 2, left). Different concentrations of BX 528 were added to shorten TAFIa-induced lysis time prolongation. The concentration response curve of BX 528 was plotted and an IC50 value was estimated by curve fitting (n= 2–6). In-vitro coagulation and platelet aggregation assays The methods of both assays have be described in detail in our previous publication (19). Briefly, the prothrombin time (PT), activated partial thromboplastin time (aPTT) and thrombin clot time (TT) were measured using reagents from Sigma and a programmable coagulometer (MLA Electra 1400C). Lyophilized human plasma (S.A.R.P.; Helena Laboratories) or fresh rabbit plasma were spiked with BX 528 (0.01 – 300 µg/ml). Fold increase in clot time in drug exposed plasma was calculated relative to control plasma without drug. Whole blood was collected from human, rabbit, dog, or rat, and 10 units/ml heparin was added as anticoagulant. ADP (5 µM), collagen (3 µg/ml), or arachidonic acid (300 µM) was used to induce a 90 to 95% of the maximum response in control blood. Impedance of blood samples was measured as the extent of platelet aggregation using a whole blood aggregometer (Model 592, Chrono-log). The area under the curve (AUC) of


Concentration of BX 528 (µM) Figure 2: Effect of BX 528 on reducing clot lysis time (left and middle) and enzymatic activity of CPN (right) in plasma from human (top), dog (middle) and rabbit (bottom). Left: Representative original spectrophotometer absorbance changes over time representing the levels of clot formation and lysis. Clot formation of the plasma samples was initiated by thrombin and clot lysis by tPA. Thrombomodulin (TM) was added to activate endogenous TAFI in the plasma that

significantly prolonged clot lysis time. Different concentrations of BX 528 (cpd) was added to inhibit TAFIa activity measured by a concentration-dependent shortening of clot lysis time. Middle: Curve fitting of the concentration-responses of BX 528 in reducing the time required for 50% clot lysis in plasma, from which an average IC50 was calculated from 2 – 6 experimental results. Right: Enzymatic activity of CPN at elevated concentrations of BX 528.

platelet aggregation over 15 min in the absence or presence of BX 528 was calculated as the percent inhibition of platelet aggregation. Pharmacokinetics in conscious rats and dogs Conscious, pre-catheterized male Sprague-Dawley rats or Beagle dogs were fasted overnight. BX 528, dissolved in 10% DMSO in physiological saline (v/v), was either injected intravenously (i.v.) as a bolus (i.v. 2 or 10 mg/kg for rats, and 4 mg/kg for dogs) or given orally (p.o.) by gavage in rats (p.o., 10 mg/kg). Blood samples were drawn at pre-designated time intervals, and plasma was prepared and kept frozen at –20°C until analysis of drug concentrations by LC/MS/MS.

was expressed as a percentage based on the area under the parent compound peak at 1 h relative to the 0-h sample. Recombinant cytochrome P450 CYP3A4 and CYP2D6 expressed by baculovirus-infected insect cells (Gentest Corporation, Woburn, MA, USA) was used to catalyze a novel non-fluorescent substrate to a fluorescent metabolite by which the P450 activity can be measured as a function of metabolite formation. BX 528 (9 nM to 20 µM) was incubated with the substrate and cytochrome P450 at 37°C in the presence of the NADPH cofactor system. An inhibition curve was constructed to determine the IC50 value for BX 528 against each isoform. Ketoconazole and quinidine served as positive inhibitor controls for CYP3A4 and CYP2D6, respectively.

In-vitro hepatic metabolism and cytochrome P450 inhibition A standard protocol using characterized hepatic microsomes of known P-450 content/mg protein was used to evaluate oxidative metabolic pathways (20). BX 528 was incubated with the microsomal enzymes in the presence and absence of an NADPH generating system at 37°C for 1 hour (h). The extent of metabolism

Plasma protein binding BX 528 (1, 10 and 30 µM) was incubated with rat, dog, rabbit, and human plasma (Calbiochem, LaJolla, CA, USA) at 37°C for 20 min. After the incubation, the samples were filtered through an Amicon CentrifreeTM (Millipore Corp.) by centrifugation at 1,000g for 10 min. BX 528 was quantified by LC/MS-MS methodology to determine the binding coefficient. Concurrently, non-

47


specific binding of the drug substance to the filtration device was determined. In-vitro cytotoxicity assays and Ames II screen The potential toxicity of BX 528 was evaluated in cell-based assays designed to measure cytotoxicity (based on intracellular ATP levels) or inhibition of cell proliferation (by [14C] thymidine incorporation). As described in detail in our previous publication (21), liver (human hepatoma HepG2 cells) and kidney (normal porcine kidney LLC-PK1 cell line and human primary renal proximal tubule epithelial cells RPTEC) cells were exposed to BX 528 at six serial dilutions (ranging from 330 to 1.4 µM) for 24 h. An ATP bioluminescence assay (luciferin-luciferase catalyzed luminescence) was used in which reduction of cellular ATP is a marker of cytotoxicity. The effect of BX 528 on cell proliferation was quantified by the incorporation of [14C] thymidine during DNA synthesis. Toxicity to cells is presented as the effective concentration of test compound to cause a 50% reduction (EC50) in ATP level or [14C] thymidine incorporation relative to the solvent control. BX 528 was also tested by Xenometrix, Inc. (Boulder, CO, USA) for mutagenic activity in an Ames II Salmonella reverse mutation assay (Ames II assay), using the Ames II Mix and strain TA98 in both the presence and absence of S9 (4-Nitroquinoline N-oxide, 2-nitroflourene and 2-aminoanthracene) to serve as positive controls. Hemodynamics in anesthetized rats and dogs Rats Male Sprague-Dawley Rats (350–400 g) were anesthetized by inhalation of 2% isoflurane (IMPAC 6, VetEquip, Pleasanton, CA, USA). A PE50 catheter was introduced into the femoral artery for blood sampling, and a second catheter into the jugular vein for drug administration. A 1.4 F Millar micro-tip catheter was implanted into the carotid artery for arterial pressure measurement. BX 528 was prepared in saline + NaOH (pH 7.0 – 7.2) at concentrations of 0.3, 1, 3, 10 and 30 mg/ml and 1 ml/kg was administered. Following surgery and a period of stabilization, ascending doses of BX 528 (0 for vehicle, 0.3, 1, 3, 10, and 30 mg/kg) were injected i.v. as a bolus over 1 min at intervals of 15 min for each consecutive dose. Mean arterial pressure and heart rate were monitored continuously, and 0.4 ml blood samples were collected for measurement of plasma drug concentration 1 min after each dose. In a separate experiment, a high dose of BX 528 (100 mg/ kg) was injected as a bolus followed by a 60 min-infusion (60 mg/ kg/h) via the jugular vein catheter. Blood samples were taken at 1, 5, 10, 30, and 60 min after drug administration for measurement of plasma drug concentration and blood chemistry. Dogs Male Beagle dogs (12–14 kg) were anesthetized with pentobarbital sodium and instrumented for measurements of arterial blood pressure, left ventricular pressure, aortic flow velocity (cardiac output) and lead II electrocardiograph (ECG). Derived hemodynamic parameters included: heart rate, left ventricular end diastolic pressure (EDP), dP/dt (index of myocardial contractility), and total peripheral vascular resistance. QTc, defined as the duration from the start of QRS wave to the end of T wave

corrected for heart rate, was determined from the ECG. Vehicle (0.9% saline) or three ascending doses of 3, 10, and 30 mg/kg BX 528 were infused intravenously for 10 min at 60–200 µl/kg per min with 20 min intervals between successive doses. Blood samples (1 ml) were withdrawn at selected time-points for measurement of plasma drug concentrations. In-vivo complement activation model in guinea pigs Male Hartley guinea pigs (400–500 g) were anesthetized by a bolus intramuscular injection of a mixture of 60 mg/kg ketamine and 7 mg/kg zylazine, placed on a warm water circulating blanket to maintain body temperature, and catheterized (20 gauge, 1.5 inch) in the right jugular vein and carotid artery. The animals were allowed to stabilize for 10 min after surgery and divided into Vehicle (phosphate buffered saline, pH 7.5, 1 ml/kg + 1 ml/ kg/h), MERGEPTA (DL-mercaptomethyl-3-guanidinoethylthiopropanoic acid, 10 mg/kg + 10 mg/kg/h), and BX 528 (0.1, 1, 10, 30, and 60 mg/kg + 0.047, 0.47, 4.7, 14 and 28 mg/kg/h, respectively) groups. Drugs or vehicle were administered through the jugular vein catheter 15 min prior to a 30 second bolus injection of 125 µg cobra venom factor (CVF) given via the carotid artery catheter. Animals were monitored for signs of respiratory distress and/or death. Surviving animals were euthanized 30 min after application of CVF. Blood samples (0.9 ml) were taken before drug infusion, 15 min after drug infusion but before injection of CVF, and 30 min after injection of CVF, placed in an Eppendorf tube containing 0.1 ml 3.8% sodium citrate as an anticoagulant and centrifuged at 1,000 x g for 10 min at 4oC. The plasma samples were divided into samples for measuring drug concentration by LC/MS/MS, stored at –20oC, or for measuring CPN activity and ex-vivo clot lysis, stored at –80oC.

Results Potency and selectivity Enzymatic inhibition of TAFIa with high selectivity BX 528 inhibited human TAFIa with an IC50 of 2 ± 0.7 nM (n=8). The IC50 of BX 528, when compared with other related carboxypeptidases, showed a selectivity of 35,000-fold against human CPN (IC50 = 69,000 nM), 25,000-fold against human CPZ (IC50 = 57,000 nM), 4.5-fold against human CPE (CPH, IC50 = 9 nM), and 12-fold against porcine pancreatic CPB (IC50 = 24 nM). The selectivity to duck CPD was 3,500-fold at pH 5 (IC50 = 7,000 nM) and increased to >50,000 at pH 7.5 as no inhibition was observed (IC50 > 100,000 nM). The effect of BX 528 on human CPN activity was also compared to that on dog and rabbit CPN in plasma milieu showing a selectivity of several thousand folds with an IC50 in the 50–100 µM range, which is consistent with the selectivity in humans (Fig. 2, right). In addition, BX 528 did not show significant activity against any of the 137 receptors and enzymes when tested at 10 µM. Thus, BX 528 is a potent and highly selective TAFIa inhibitor. Functional inhibition of TAFIa-induced retardation of clot lysis in plasma BX 528 was tested in an in-vitro clot lysis assay in order to show that the enzymatic inhibition of TAFIa in a purified system can be translated to functional inhibition in a plasma-based system.

48


After adding thrombin and CaCl2 to the plasma to cleave fibrinogen to form fibrin, clot formation occurred rapidly as measured by an increase in turbidity (Fig. 2, left). Clot lysis was initiated by t-PA, which can be detected by a subsequent rapid reduction in turbidity. Addition of TM enhanced thrombin-induced activation of endogenous TAFI, leading to a prolongation of clot lysis time, which could be blocked by addition of BX 528 in a concentration-dependent manner. The potency of BX 528 in reducing plasma clot lysis time (IC50), calculated by curve fitting of the concentration-response curves, was 45 ± 2.6, 23 ± 0.4 and 15 ± 3.7 nM (n=2–6), for human, dog and rabbit, respectively, under the conditions tested (Fig. 2, middle). No effects on in-vitro coagulation and platelet aggregation Prolongation of PT and aPTT indicates inhibition of clotting factors that comprise the extrinsic and intrinsic coagulation pathway, respectively. Prolongation of TT indicates inhibition of thrombin, abnormal fibrinogen (dysfibrinogenemia) or presence of substances interfering with normal fibrin polymerization. BX 528 had minimal effects on PT, aPTT and TT in human and rabbit plasma at concentrations up to 200 µM, nor did it affect platelet aggregation in response to ADP, arachidonic acid or collagen when tested against rat, rabbit, dog and human platelets up to 10 µM (data not shown).

Figure 3: Pharmacokinetic profiles of BX 528 in conscious rats (top) and dogs (bottom).

Absorption, distribution, metabolism and elimination Pharmacokinetics in conscious rats and dogs I.v. administration of BX 528 resulted in a plasma concentration curve that fitted a two-compartment distribution model (Fig. 3). The plasma half-life was moderate in rats (0.28 to 0.85 h), but much longer in dogs (4.5 h). Comparing the AUC between the 2 and 10 mg/kg i.v. doses in rats, the compound appeared to have linear kinetics, 4.5 ± 0.6 and 19 ± 1.1 µg *h/ml, respectively. The AUC following an oral dose of 10 mg/kg was 0.23 ± 0.04 µg *h/ml. Thus, the oral bioavailability was approximately 1%. In-vitro hepatic metabolism The susceptibility of BX 528 to undergo hepatic Phase-I oxidative metabolism was evaluated utilizing liver microsomes from rat, dog, and human. After incubation of 10 µM BX 528 with liver microsomes, more than 90% of the intact compound was detected in rat and dog, and 80% in human (Fig. 4, top). Thus, BX 528 was not metabolized in the microsomal metabolic system to any appreciable extent in all three species examined and no significant metabolites were identified. Therefore, oxidative metabolic pathways are not expected to play a major role in the disposition of BX 528. Cytochrome P450 inhibition Cytochrome P450 enzymes are an important superfamily of enzymes responsible for the metabolism of a variety of xenobiotics. Any interaction of novel chemical entity with the cytochrome P450 system has the potential to elicit clinical drug-drug interactions. Since CYP3A4 and CYP2D6 inhibition are responsible for the metabolism of roughly 85% of marketed drugs, we evaluated the potential of BX 528 to inhibit these two human cytochrome P450 enzymes. BX 528 did not inhibit either CYP3A4 or CYP2D6 at a concentration up 30 µM (Fig. 4, bottom).

49

Figure 4: Effects of BX 528 on in-vitro hepatic metabolism (top) and cytochrome P450 inhibition (bottom).

Plasma protein binding BX 528 was moderately bound to plasma proteins in all of the four species tested, indicating that a significant portion of drug was free and available in the plasma. At the concentrations tested of 1, 3 and 10 µM, approximately 50% was bound in human, dog and rabbit plasma, and 70% in rat plasma. The proportion bound was not affected by the concentration of BX 528 in the range tested.


Figure 5: Hemodynamic effects in anesthetized rats following intravenous injection of BX 528 as a bolus at ascending doses (0.3 – 30 mg/kg, left) or a single dose (100 mg/kg, right).

Safety pharmacology Cytotoxicity and mutagenic activity BX 528 had no effects on either the cytotoxicity and cell proliferation assays at concentrations up to 300 µM in all three cell types tested (HepG2, LLC-PK1, RPTEC), and no mutagenic activity when tested with the Ames II Mix or strain TA98 assays, either in the absence or presence of up to 5 mg/ml S9. Hemodynamic effects in anesthetized rats and dogs BX 528 did not affect blood pressure, heart rate, cardiac function and ECG parameters at doses up to 100 mg/kg in rats (Fig. 5) and 30 mg/kg in dogs (Fig. 6). In these experiments the peak plasma drug concentration reached approximately 1,000 and 300 µM, respectively. No abnormalities were found in standard blood chemistry assays. Thus, BX 528 was well tolerated hemodynamically in both rats and dogs.

(1 out of 12 animals had continued labor breathing and was cyanotic throughout the experiment). Pretreatment with a CPN inhibitor, MERGEPTA, at a dose that completely inhibited CPN activity, exacerbated CVF-induced respiratory distress and resulted in the death of all animals within 8 min. In contrast, pretreatment with BX 528, which completely inhibited TAFIa-induced clot lysis prolongation at 0.1 mg/kg and 0.42 µM, but only minimally attenuated plasma CPN activity at 60 mg/kg and 251 µM (Fig. 7), did not exacerbate symptoms, but completely prevented CVF-induced respiratory distress. Thus, the safety margin with regard to CPN inhibition was estimated to be ~600 by comparing the highest safe dose tested and plasma concentration over the lowest effective dose tested and plasma concentration.

Discussion

In-vivo complement activation model in guinea pigs Since inhibition of CPN causes acute respiratory distress during complement activation, we tested BX 528 during complement activation in a guinea pig model in order to demonstrate its invivo specificity. Treatment with CVF immediately caused respiratory distress (labored breathing, shallow, rapid breathing, gasping) in all animals in the vehicle group. This response was selflimiting with the majority of animals appearing normal in 30 min

50

We have discovered a novel small-molecule TAFI inhibitor that is potent, highly selective and safe. BX 528 inhibited human TAFIa in purified enzyme assays with an IC50 of 2 nM. This effect translated to a functional inhibition in plasma of TAFIa-induced retardation of clot lysis with an IC50 of 45 nM. Reduced potency in plasma clot lysis assay compared to purified enzymatic assay is a common phenomenon for most small-molecule compounds. In this case, it could be partially due to protein bind-


Figure 6: Hemodynamic effects in anesthetized dogs following intravenous infusion of BX 528 at ascending doses (3–30 mg/kg). The duration of the infusion is indicated by a short bar below each dose.

ing (~50% in human) or that the effect of TAFIa is also affected to some degree by t-PA concentration and other parameters in plasma. Since carboxypeptidases are a large family of proteins with broad range of physiological functions, the specificity of a TAFI inhibitor is extremely important regarding its safety. There was a relatively low (4.5-fold) selectivity of BX 528 against CPE when measured at pH 5.5, but not at pH 7.5. Similar to CPD and CPZ, CPE (CPH) is found in tissues and plays a role in peptide hormone activation by cleaving C-terminal basic residues from intermediate forms. The phenotype of the CPE (fat/fat) mouse suggests what the potential adverse effects due to long-term inhibition of CPE might be (22). Inhibition of CPE is unlikely to result in these side effects during transient, acute administration. BX 528 also had a relatively low (12-fold) selectivity against porcine pancreatic CPB, an exopeptidase that processes peptide fragments exposed during digestion by digestive endopeptidases, including pepsin, trypsin and chymotrypsin. However, short-term and transient inhibition of pancreatic CPB alone may not lead to intestinal malabsorbtion due to the presence of alternate active proteases available for digestion. In addition, CPB is secreted into the intestine (23), and will therefore only minimally interact with a TAFIa inhibitor in the circulation, thus, further reducing the potential risk.

51

TAFI, the current therapeutic target, shares the same physiological compartment in blood as plasma CPN, a basic carboxypeptidase involved in the degradation of anaphylatoxin generated during activation of the complement system and acting as a pleiotropic regulator of inflammation (24). In a guinea pig model, CVF is infused to activate the alternative pathway of complement, thereby generating the anaphylatoxins C3a and C5a, causing peripheral small airway constriction, bronchial constriction, vasoconstriction of small muscular arteries, cell aggregates in blood vessels, severe congestion, pulmonary edema, and an interstitial infiltrate of mononuclear cells, etc. (25). These effects are usually non-lethal, because the generated anaphylatoxins are rapidly converted to inactive forms of C3a des Arg and C5a des Arg by CPN. Indeed, in the present study, after initial respiratory distress induced by CVF, most of the animals with functioning CPN recovered within 30 min. However, in animals receiving a non-selective CPN inhibitor, MERGEPTA, at a dose that completely abolished CPN activity, CVF-induced respiratory distress was exacerbated to an extent that caused death in all animals. Thus, because CPN is a negative modulator of the complement system, it is critical to maintain its activity. This is particularly important in acute myocardial infarction patients whose complement system has previously been shown to be activated (26). In contrast, BX 528 which has a 35,000-fold selec-


Figure 7: Plasma concentrations of BX 528 before and 30 min after CVF administration (top) and inhibition of CPN activity and prolongation of clot lysis (bottom) in an in-vivo complement activation model in guinea pigs.

tivity over CPN in the enzyme assay, only minimally inhibited CPN activity at 60 mg/kg (the highest dose tested), which gives a plasma concentration of approximately 250 µM. BX 528 at all doses tested (10–60 mg/kg) completely prevented CVF-induced respiratory distress, an outcome that was significantly better than the vehicle. This suggests that in this model TAFI is activated and plays a role in the pathophysiology, confirming the observation that TAFI-deficient mice are susceptible to CVF when concurrently treated with LPS (27). Although MERGEPTA also weakly inhibits TAFIa (8), due to its relatively selective inhibition of CPN, any beneficial effects of TAFIa inhibition on prevention of CVF-induced respiratory distress are overcome by the primary effects caused by CPN inhibition. Thus, the extremely high selectivity away from CPN inhibition by BX 528

provides a great safety margin. Furthermore, the present results also suggest that inhibition of TAFIa is not only useful as a thrombolytic enhancer, but also as an anti-inflammatory agent, since accumulating evidence shows that TAFIa is also proinflammatory (28, 29). In addition, thromboembolism may be an extraintestinal manifestation of inflammatory bowel disease and an important cause of mortality, in which TAFIa levels have been linked with thrombophilia (30). Although the risk of thromboembolism is multifactorial, it appears to be related to mucosal inflammatory activity in most patients. Thus, like activated protein C, TAFIa may plays a pivotal role in regulating the crosstalk between coagulation, fibrinolysis, and inflammation. Systemic hemodynamic effects are a general safety concern for small-molecule compounds, especially cardiovascular agents. We demonstrated that in-vivo administration of BX 528 did not affect cardiac function, systemic blood pressure or electrocardiographic parameters in both rats and dogs at doses up to 100 and 30 mg/kg, which gave peak plasma concentrations of up to 1,000 and 300 µM, respectively. This result not only indicates that BX 528 is free from general hemodynamic toxicity, but also demonstrates that inhibition of endogenous TAFIa does not affect cardiovascular hemodynamics under these conditions. This data is in contrast to the effects on blood pressure observed when bradykinin was shown to be degraded by TAFIa following its infusion into TAFI deficient mice (31). The apparent difference between the two studies may be that normally bradykinin is cleared by another mechanism than TAFIa, but when infused at a concentration sufficient to perturb blood pressure, then TAFIa plays a role in its clearance. Thus, we have developed a potent, highly selective smallmolecule TAFIa inhibitor, BX 528. In separate studies, we have demonstrated that BX 528 enhanced both exogenous (t-PA induced) and endogenous thrombolysis, thus, producing therapeutic efficacy without increasing bleeding risk in four different animal models of thrombosis. Acknowledgements We are grateful to Dr. Lloyd Fricker, Albert Einstein College of Medicine (Bronx, NY, USA) for providing enzyme and advice for CPE, CPZ and CPD assays; and to Kieu Chu, Galina Rumennik, Christopher West, Joseph Post, Mark E. Sullivan, William P. Dole, for their contribution and support for this project. This research was part of the discovery and development of BX 528 carried out in Berlex Biosciences, USA, an affiliation of Schering AG, Germany, by which all authors were employed.

References 1. Nesheim M, et al. Thrombin, thrombomodulin and TAFI in the molecular link between coagulation and fibrinolysis. Thromb Haemost 1997; 78: 386–91. 2. Bouma BN, et al. Thrombin-activatable fibrinolysis inhibitor (TAFI, plasma procarboxypeptidase B, procarboxypeptidase R, procarboxypeptidase U). J Thromb Haemost 2003; 1: 1566–74. 3. Leurs J, et al. Carboxypeptidase U (TAFIa): a metallocarboxypeptidase with a distinct role in haemostasis and a possible risk factor for thrombotic disease. Thromb Haemost 2005; 94: 471–87. 4. Nesheim M. Myocardial infarction and the balance between fibrin deposition and removal. Ital Heart J 2001; 2: 641–5.

5. Boffa MB, et al. Thrombin activable fibrinolysis inhibitor (TAFI): molecular genetics of an emerging potential risk factor for thrombotic disorders. Curr Drug Targets Cardiovasc Haematol Disord 2001; 1: 59–74. 6. Polla MO, et al. Design and synthesis of potent, orally active, inhibitors of carboxypeptidase U (TAFIa). Bioorg Med Chem 2004; 12: 1151–75. 7. Barrow JC, et al. Synthesis and evaluation of imidazole acetic acid inhibitors of activated thrombin-activatable fibrinolysis inhibitor as novel antithrombotics. J Med Chem 2003; 46: 5294–7. 8. Suzuki K, et al. Enhancement of fibrinolysis by EF6265 [(S)-7-amino-2-[[[(R)-2-methyl-1-(3phenylpropanoylamino)propyl]hydroxypho sphinoyl]

52

methyl]heptanoic acid], a specific inhibitor of plasma carboxypeptidase B. J Pharmacol Exp Ther 2004; 309: 607–15. 9. Boyd EA, et al. Synthesis of gamma-keto-substituted phosphinic acids from bis(trimethylsilyl)phosphonite and alpha,beta-unsaturated ketones. Tetrahedron Lett 1992; 33: 813–16. 10. Baylis EK, et al. 1-Aminoalkylhosphonous acids. Part 1. Isosteres of the protein amino acids. Chem Soc, Perkin Trans 1984; 1: 2845–53. 11. Islam I, et al. 3-Mercaptopropionic acids as efficacious inhibitors of activated thrombin activatable fibrinolysis inhibitor (TAFIa). Bioorg Med Chem Lett 2006; dpo:10.1016/j.bmcl.2006.11.08.

Wang et al. Pharmacological profile of a TAFIa inhibitor 12. Suzuki S, et al. Spectrophotometric Determination of Glycine with 2,4,6 Trichloro-s-Triazine. Anal Chem 1970; 42: 101–3. 13. Hendricks D, et al. Colorimetric assay for carboxypeptidase N in serum. Clinca Chimica Acta 1986; 157: 103–8. 14. Levin Y, et al. Isolation and characterization of the subunits of human plasma carboxypeptidase N (kininase i). Proc Natl Acad Sci USA 1982; 79: 4618–22. 15. Novikova EG, et al. Purification and characterization of human metallocarboxypeptidase Z. Biochem Biophys Res Commun 1999; 256: 564–8. 16. Nagashima M, et al. An inhibitor of activated thrombin-activatable fibrinolysis inhibitor potentiates tissue-type plasminogen activator-induced thrombolysis in a rabbit jugular vein thrombolysis model. Thromb Res 2000; 98: 333–42. 17. Zhao L, et al. Mutations in the substrate binding site of thrombin-activatable fibrinolysis inhibitor (TAFI) alter its substrate specificity. J Biol Chem 2003; 278: 32359–66. 18. Wu C, et al. Activated thrombin-activatable fibrinolysis inhibitor attenuates spontaneous fibrinolysis of

batroxobin-induced fibrin deposition in rat lungs. Thromb Haemost 2003; 90: 414–21. 19. Post JM, et al. Human in vitro pharmacodynamic profile of the selective Factor Xa inhibitor ZK-807834 (CI-1031). Thromb Res 2002; 105: 347–52. 20. Subramanyam B, et al. Studies on the in vitro conversion of haloperidol to a potentially neurotoxic pyridinium metabolite. Chem Res Toxicol 1991; 4: 123–8. 21. Li W, et al. Use of cultured cells of kidney origin to assess specific cytotoxic effects of nephrotoxins. Toxicol In Vitro 2003; 17: 107–13. 22. Udupi V, et al. Effect of carboxypeptidase E deficiency on progastrin processing and gastrin messenger ribonucleic acid expression in mice with the fat mutation. Endocrinology 1997; 138: 1959–63. 23. Borgstrom A, et al. Active carboxypeptidase B is present in free form in serum from patients with acute pancreatitis. Pancreatology 2005; 5: 530–6. 24. Matthews KW, et al. Carboxypeptidase N: a pleiotropic regulator of inflammation. Mol Immunol 2004; 40: 785–93. 25. Huey R, et al. Potentiation of the anaphylatoxins in vivo using an inhibitor of serum carboxypeptidase N

53

(SCPN). I. Lethality and pathologic effects on pulmonary tissue. Am J Pathol 1983; 112: 48–60. 26. Ren G, et al. Inflammatory mechanisms in myocardial infarction. Curr Drug Targets Inflamm Allergy 2003; 2: 242–56. 27. Asai S, et al. Absence of procarboxypeptidase R induces complement-mediated lethal inflammation in lipopolysaccharide-primed mice. J Immunol 2004; 173: 4669–74. 28. Bouma BN, et al. Thrombin activatable fibrinolysis inhibitor (TAFI) at the interface between coagulation and fibrinolysis. Pathophysiol Haemost Thromb 2003; 33: 375–81. 29. Bajzar L, et al. Thrombin activatable fibrinolysis inhibitor: not just an inhibitor of fibrinolysis. Crit Care Med 2004; 32: S320–4. 30. Quera R, et al. Thromboembolism--an important manifestation of inflammatory bowel disease. Am J Gastroenterol 2004; 99: 1971–3. 31. Myles T, et al. Thrombin activatable fibrinolysis inhibitor, a potential regulator of vascular inflammation. J Biol Chem 2003; 278: 51059–67.