Journal of Thrombosis and Haemostasis, 6: 82(F829
`
`DOI: 10.1111/j.1538-7836.2008.02939.x
`
`ORIGINAL ARTICLE
`
`Apixaban, an oral, direct and highly selective factor Xa
`inhibitor: in vitro, antithrombotic and antihemostatic studies
`
`P. C. WONG, E. J. CRAIN, B. XIN, R. R. WEXLER, P. Y. S. LAM, D. J. PINTO, J. M. LUETTGEN and
`R. M. KNABB
`
`Thrombosis Research, Bristol—Myers Squibb Company, Pennington, NJ, USA
`
`To cite this article: Wong PC, Crain EJ, Xin B, Wexler RR, Larn PYS, Pinto DJ, Luettgen JM, Knabb RM. Apixaban, an oral, direct and highly
`selective factor Xa inhibitor: in vitro, antithrombotic and antihemostatic studies. J Thromb Haemost 2008; 6: 820-9.
`
`Summary. Background: Apixaban is an oral, direct and highly
`selective factor Xa (FXa) inhibitor in late-stage clinical devel-
`opment for the prevention and treatment of thromboembolic
`diseases. Objective: We evaluated the in vitro properties of
`apixaban and its in vivo activities in rabbit models of thrombosis
`and hemostasis. Methods: Studies were conducted in arteriove-
`
`thrombosis (AVST), venous thrombosis (VT),
`nous-shunt
`electrically mediated carotid arterial thrombosis (ECAT) and
`cuticle bleeding time (BT) models. Results: In vitro, apixaban is
`potent and selective, with a Ki of 0.08 nM for human FXa. It
`exhibited species diflerence in FXa inhibition [FXa K, (nM):
`0.16, rabbit; 1.3, rat; 1.7, dog] and anticoagulation [ECZX (uM,
`concentration required to double the prothrombin time): 3.6,
`hmnan; 2.3, rabbit; 7.9, rat; 6.7, dog]. Apixaban at 10 uM did
`not alter human and rabbit platelet aggregation to ADP, 'y-
`thrombin, and collagen. In vivo, the values for antithrombotic
`ED50 (dose that reduced thrombus weight or increased blood
`flow by 50% of the control) in AVST, VT and ECAT and the
`values for BT ED3x (dose that increased BT by 3-fold) were
`0.27 as 0.03, 0.11 J: 0.03, 0.07 3: 0.02 and > 3 mg kg‘1 h"
`i.v. for apixaban, 0.05 3: 0.01, 0.05 3: 0.01, 0.27 3: 0.08 and
`> 3 mg kg_1 h"l i.v. for the indirect FXa inhibitor fondapar-
`inux,
`and 0.53 d: 0.04,
`0.27 i 0.01,
`0.08 3: 0.01
`and
`0.70 3: 0.07 mg kg_1 day_1 p.o. for the oral anticoagulant
`warfarin, respectively. Conclusions: In summary, apixaban was
`effective in the prevention of experimental thrombosis at doses
`that preserve hemostasis in rabbits.
`
`Correspondence: Pancras C. Wong, Bristol-Myers Squibb Company,
`311 Pennington-Rocky Hill Road, Pennington, NJ 08534, USA.
`Tel.: +1 609 818 5572; fax: +1 609 818 7877.
`E-mail: pancras.wong@bms.com
`
`Presented in part at the 48th Annual Meeting of the American Society of
`Hematology, %12 December, 2006, Orlando, Florida (abstract 917) and at
`the XXI International Society on Thrombosis and Haemostasis Congress,
`&12 July 2007, Geneva, Switzerland (abstract no. P-W-656).
`
`Received 31 October 2007, accepted 15 February 2008
`
`factor Xa
`Keywords: apixaban, blood coagulation, direct
`inhibitor, hemostasis, oral anticoagulant, thrombosis.
`
`I ntroduction
`
`Thrombosis is a major cause of morbidity and mortality in the
`Western world and contributes to cardiovascular disorders.
`
`Despite recent advances in interventional and medical therapy
`for
`the prevention and treatment of a wide variety of
`thromboembolic events,
`the morbidity and mortality rates
`are still high [1]. Thus, there is a real unmet medical need for
`safer and more effective antithrombotic therapies. As thrombin
`plays a key role in the generation of an occlusive thrombus,
`inhibiting thrombin activity and/or its generation becomes a
`major drug target in the search of novel antithrombotic agents.
`Because thrombin is the last serine protease in the blood
`coagulation cascade that causes fibrin clot formation, the early
`research approach has been focused on the inhibition of
`thrombin activity by direct thrombin inhibitors [2]. Another
`approach is the inhibition of thrombin production by blocking
`the upstream proteases in the blood coagulation cascade, such
`as factor Xa (FXa). Experimental evidence suggests that FXa
`inhibitors have antithrombotic efficacy with lower bleeding risk
`in animals when compared with thrombin inhibitors [3,4].
`Thus, inhibition of FXa appears to be a promising mechanism
`for anticoagulant therapy, which prompted us to initiate a drug
`discovery program on small-molecule, direct FXa inhibitors
`[5,6].
`Based on years of research and development, we recently
`identified a direct FXa inhibitor, apixaban (BMS-562247),
`which is highly potent and selective with a K, of 0.08 nM for
`human FXa and with greater than 30 000-fold selectivity over
`other coagulation proteases [7]. Preliminary studies show that
`apixaban has high oral bioavailability in rats, dogs, and
`humans [8,9]. It prevented venous thromboembolic events with
`a favorable efficacy and safety profile in patients after knee
`replacement surgery [10]. Apixaban was also shown to be
`efficacious and safe in the treatment of patients with acute
`symptomatic deep vein thrombosis [11]. Currently, apixaban is
`in late-stage clinical development for the prevention and
`treatment of thromboembolic diseases.
`
`© 2008 International Society on Thrombosis and Haemostasis BMS 2002
`CFAD
`V. BMS
`IPR2015-01723
`
`

`
`In this study, we characterized the in vitro properties of
`apixaban and its antithrombotic profile for the prevention of
`thrombosis in rabbit models of arteriovenous—shunt thrombo-
`
`(AVST), venous thrombosis (VT) and electrolytically
`sis
`mediated carotid arterial thrombosis (ECAT), in comparison
`with the indirect FXa inhibitor fondaparinux and the oral
`anticoagulant warfarin. Antihemostatic effects of these agents
`were also studied in a well-characterized rabbit cuticle bleeding
`time (BT) model.
`
`Materials and methods
`
`Reagents
`
`The following drugs and chemicals were used in this study:
`chromogenic substrates S-2222, S-2238 and S-2765 from
`Chromogenix AB (distributed by DiaPhar1na Group, Inc.,
`West Chester, OH, USA); activated partial thromboplastin
`time (APTT) reagent Alexin from Trinity Biotech (St Louis,
`MO, USA); prothrombin time G’T) reagents Thromboplastin
`C Plus from Dade-Behring (Deerfield, IL, USA) or Thromb-
`max® with calcium from Sigma Chemical Co (St Louis, MO,
`USA); HepTest reagent from American Diagnostica (Stam-
`ford, CT, USA); human FXa from Haematologic Technolo-
`gies (Essex Junction, VT, USA); sodium warfarin, ADP and
`collagen from Sigma Chemical Co.; human 'y-thrombin from
`ICN Biomedicals, Inc. (Costa Mesa, CA, USA); fondaparinux
`(Arixtra®) from GlaxoSrr1ithKline (Research Triangle Park,
`NC, USA). Apixaban and DMP802 were synthesized at
`Bristol-Myers Squibb Company.
`
`Animals
`
`Studies were conducted in male New Zealand White rabbits
`
`weighing about 2-4 kg obtained from Covance (Denver, PA,
`USA). Experiments were conducted in accordance with the
`NIH Guide for the Care and Use of Laboratory Animals, and
`the regulations of the Animal Care and Use Committee of the
`Bristol-Myers Squibb Company.
`
`In vitro and ex vivo studies
`
`assays Using established protein purification
`Enzyme
`procedures, FX was isolated from citrated plasma obtained
`from healthy dogs, rats, and rabbits [12]. Purified FXa was
`obtained after activation with Russell’s viper venom followed
`by affinity chromatography. The resulting FXa was > 95%
`pure as judged by sodirun dodecylsulfate polyacrylamide gel
`electrophoresis. The
`substrate affinity values
`for FXa,
`expressed as the Michaelis—Menten—Henri constant (Km), for
`human, rabbit, rat and dog FXa were determined using the
`chromogenic substrate S-2765, and were 36, 60, 240 and 70 ]JM,
`respectively. The substrate hydrolysis was monitored by
`measuring absorbance at 405 nm at 25 °C for up to 30 min
`using a SpectraMax 384 Plus plate reader and SoftMax
`(Molecular Devices, Sunnyvale, CA, USA). FXa activity for
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`Apixaban in vitro/in vivo studies 821
`
`each substrate and inhibitor concentration pair was determined
`in duplicate. The Ki values were calculated by non-linear least-
`squares fitting of the steady-state substrate hydrolysis rates to
`the equation for competitive inhibition (Equation 1) using
`GRAFIT (Erithacus Software Ltd., Surrey, UK), where v
`equals reactions velocity in OD min_1, Vmax equals maxiumum
`reaction velocity, S equals substrate concentration, and I equals
`inhibitor concentration.
`
`v—#maxS
`Km (1 + Kii) + S
`
`(1)
`
`Blood samples were collected in tubes
`Clotting assays
`containing 1/10 volrune of 3.2% sodirun citrate, and platelet-
`poor plasma was obtained after centrifuging at > 2000 X g for
`10 1r1in. Clotting times were measured with an automated
`coagulation analyzer (Sysmex®, Dade Behring Inc., Deerfield,
`IL, USA). PT, APTT, and HepTest reagents were reconstituted
`and assays were performed according to the manufacturer’s
`instructions. The modified PT (mPT) assay was performed by
`diluting 1mL of thromboplastin C Plus with 1.25 mL of
`100 mM calcirun chloride and using this diluted reagent in place
`of the normal PT reagent. For PT and mPT, plasma (50 uL)
`was warmed to 37 °C for 3 min before adding PT reagent
`(100 uL). For APTT, plasma (50 uL) was warmed to 37 °C for
`1 min before adding APTT reagent (50 uL). After two more
`minutes, 25 mM calcirun chloride (50 uL) was added. For
`HepTest, plasma (50 IL) was warmed to 37 °C for 2 min
`before bovine FXa (50 uL) was added. After two more
`minutes, HepTest ReCal mix
`(50 uL) was
`added.
`Determinations were performed in duplicate and expressed as
`a mean ratio of treated vs. baseline control. The concentrations
`
`required to prolong clotting time by 2-fold (ECZX) were
`expressed as total plasma concentrations, not final assay
`concentrations after addition of clotting assay reagents. For
`in vitro studies, apixaban was serially diluted into citrated
`plasma obtained from healthy dogs,
`rats, and rabbits,
`beginning with a 10 mM dimethylsulfoxide stock solution.
`
`aggregation was
`Platelet
`aggregation assays
`Platelet
`measured in citrated hmnan and rabbit platelet-rich plasma
`G’RP) in vitro with a platelet aggregometer (Model PAP-4D,
`BioData, Horsham, PA, USA). PRP was obtained from
`citrated blood after centrifuging at 250 X g for 6 min. Citrated
`PRP (250 IL) was mixed with 20 uL of vehicle, DMP802 at
`3 ]J.M or apixaban at 1—10 [JM, and incubated for 3 min at
`37 °C. DMP802,
`a glycoprotein (GP)IIb/IIIa
`receptor
`antagonist, was included as a positive control (IC50 = 29 nM
`against human platelet aggregation response to 10 [JM ADP)
`[13]. Peak platelet aggregation was determined after
`the
`addition of 20 uL of the agonist (ADP at 10 ]JM, 'y-thrombin
`at 35 nM, and collagen at 10 pg mL", final concentration).
`
`and
`assays Anti-FXa
`antithrombin
`and
`Anti-FXa
`the
`determined
`using
`antithrombin
`activities
`were
`chromogenic substrates S-2222 and S-2238,
`respectively,
`based on the method of Sato et al.
`[14] with modifications
`
`

`
`822 P. C. Wong et al
`
`[15]. The hydrolysis of the chromogenic substrates was assayed
`by measuring absorbance at 405 nm at 37 °C as described
`above. Anti-FXa or antithrombin activity was calculated by
`comparing the values of optical density from samples taken in
`the post-treatment period to those taken in the pretreatment
`period.
`
`In vivo studies
`
`The rabbit AVST model, described by Wong
`A VST model
`et al. [16], was used in this study. Briefly, male New Zealand
`White rabbits were anesthetized with ketamine (50 mg kg"l
`i.m.) and xylazine (10 mg kg"1 i.m.), and their femoral artery,
`jugular vein and femoral vein were catheterized. These
`anesthetics were supplemented as needed. Thrombosis was
`induced by an arteriovenous (AW-shunt device containing a
`silk thread. Blood flowed from the femoral artery via the AV
`shunt into the opposite femoral vein for 40 min. The shunt was
`then disconnected and the silk thread covered with thrombus
`
`was weighed.
`As apixaban has an oral bioavailability of < 5% in rabbits
`(unpublished result), it was administered intravenously for
`in vivo studies. To achieve a stable plasma level with
`minimum experimental variability, apixaban, fondaparinux
`or vehicle was given by a continuous intravenous infusion 1 h
`prior
`to shunt placement. The infusion was continued
`throughout the experiment. Warfarin or vehicle was dosed
`orally once daily for 4 days. On the fourth day after the last
`oral dose of warfarin or vehicle, rabbits were anesthetized
`1.5 h later, and the treatment effect was evaluated about 2 h
`postdose. Arterial blood samples for the determination of
`clotting times or plasma levels were collected 20 min after
`shunt placement. Plasma levels of apixaban were measured by
`a specific and sensitive liquid chromatographic mass spec-
`trometry method (LC/MS/MS).
`In rabbits treated with
`apixaban,
`fondaparinux or warfarin,
`the antithrombotic
`effects of these agents were expressed as percentage inhibition
`of
`thrombus
`formation based on the treated vs.
`the
`
`corresponding mean vehicle. The ED50 value (dose that
`produced 50% inhibition of
`thrombus
`formation) was
`determined as described below.
`
`The apixaban group treatment consisted of vehicle (10%
`N,N-dimethylacetamide; 30% 1,2-propanediol; 60% water)
`(11 2 4), and apixaban (mg kg“ h“) at 0.03 (n 2 7), 0.1
`(n = 7), 0.3 (n = 7),
`1 (n = 7), and 3 (n = 3). The fonda-
`parinux group treatment consisted of vehicle (saline) (:1 = 6),
`and fondaparinux (mg kg—‘ h'1)at0.01(n 2 5),0.03(n 2 5),
`0.1 (n = 5), 0.3 (n = 5), and 1 (n = 5). The warfarin group
`treatment consisted of vehicle (water) (11 = 6), and warfarin
`(mg kg" day”) at 0.1 (n 2 6), 0.3 (n 2 6), 1 (n 2 6), and 3
`(n = 6).
`
`The rabbit VT model, described by Hollenbach
`VT model
`et al. [17], was used in this study with modifications. Briefly,
`rabbits were anesthetized as above. The left femoral vein was
`
`catheterized, using 11-crn IntraMedic polyethylene tubing
`
`(PE-200; Becton Dickinson, Sparks, MD, USA). A prosthetic
`device was passed through the PE-200 tubing into the
`abdominal vena cava. The prosthetic device consisted of a
`single
`strand of #10 awg braided copper wire (14 cm)
`terminated with eight pieces of 4-0 silk threads 3 cm in
`length. The silk threads were positioned in the abdominal
`vena cava by advancing the copper guide wire. Thrombi were
`formed on the silk threads in a time-dependent fashion.
`Apixaban and fondaparinux were given intravenously as
`described above 1 h prior to the placement of the prosthetic VT
`device. Warfarin or its vehicle was dosed orally once daily for
`4 days in rabbits as described above, and the prosthetic VT
`device was placed 2 h after the last oral dose. Ninety minutes
`after the placement of the prosthetic device, the abdominal
`vena cava was isolated through a midline abdominal incision.
`The vena cava was ligated just above and below the prosthetic
`device with 2-0 silk. The vena cava segment between the
`ligations was excised, and the threads with associated thrombus
`were removed, blotted twice on paper, and weighed. The
`weight of thrombus formed on the threads was calculated by
`subtracting the average weight of eight pieces of 4-0 silk threads
`3 cm in length. Clotting times of apixaban, fondaparinux and
`warfarin in plasma samples collected during VT were measured
`as above.
`
`In the VT study, the apixaban group treatment consisted of
`vehicle (10% N,N-dimethylacetamide; 90% of 5% dextrose)
`(n 2 6), and apixaban (mg kg" h“) at 0.03 (n 2 6), 0.1
`(n = 6), 0.3 (n = 6), and 1 (n = 6). The fondaparinux group
`treatment consisted of vehicle (saline) (:1 = 6), and fondapar-
`inux (mg kg‘1 h") at 0.01 (n = 6), 0.03 (n = 6), 0.1 (n = 6),
`and 0.3 (n = 6). The warfarin group treatment consisted of
`vehicle (water) (11 = 6), and warfarin (mg kg_1 day_1) at 0.1
`(n = 5), 0.3 (n = 5),
`1 (n I 5), and 3 (n = 6). The ED50
`value (dose that produced 50% inhibition of mean vehicle
`thrombus weight) was determined as described below.
`
`rabbit ECAT model,
`The
`thrombosis model
`Arterial
`[15], was used in this
`study.
`described by Wong et al.
`Briefly, male New Zealand White rabbits were anesthetized
`as above. An electromagnetic flow probe was placed on a
`segment of an isolated carotid artery to monitor blood flow.
`Thrombosis was induced by electrical stimulation of the
`carotid artery for 3 min at 4 mA, using an external stainless-
`steel bipolar electrode. Carotid blood flow was measured
`continuously over a 90-min period to monitor thrombosis-
`induced occlusion.
`Integrated carotid blood flow over
`90 min was measured by the area under the flow—ti1ne
`curve, calculated using the trapezoidal rule, and expressed as
`percentage of total control carotid blood flow, which would
`result
`if
`control
`blood
`flow had
`been maintained
`
`continuously for 90 min. The administration of apixaban
`and fondaparinux was initiated intravenously as described
`above 1 h prior to the artery injury. Warfarin or its vehicle
`was dosed orally once daily for 4 days in rabbits, and
`thrombosis was initiated 2 h after the last oral dose. Clotting
`times of
`apixaban,
`fondaparinux
`and warfarin,
`and
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`

`
`concentrations of apixaban in plasma samples, taken during
`electrically induced arterial
`thrombosis, were measured as
`above. In addition, we also measured ex vivo anti-FXa and
`antithrombin activities.
`
`Results
`
`In vitro studies
`
`Apixaban in vitro/in vivo studies 823
`
`The Lineweaver—Burk plot of inhibition of
`Enzyme assays
`human FXa by apixaban indicates
`that apixaban is a
`competitive inhibitor vs. the chromogenic peptide substrate
`S-2765, with a K of 0.08 nM (Fig. 1). Apixaban also inhibited
`FXa from rats, dogs, and rabbits (Table 1). In terms of FXa Ki
`at 25 °C, apixaban has similar potency in inhibiting hmnan and
`rabbit FXa, but is 10-20 times less potent against rat and dog
`FXa (Table 1).
`
`Clotting assays As expected for an inhibitor of FXa,
`addition of apixaban to normal hmnan plasma prolonged
`clotting times,
`including APTT, PT, mPT, and HepTest.
`Among the three clotting time assays, it appears that the mPT
`and HepTest are 10-20 times more sensitive than APTT and
`PT in monitoring the in vitro anticoagulant effect of apixaban
`in human plasma (Table 1). In both the PT and APTT assays,
`apixaban had the highest potency in hmnan and rabbit plasma,
`but was less potent in rat and dog plasma (Table 1).
`
`In this study, the apixaban group treatment consisted of
`vehicle (10% N,N-dimethylacetamide; 90% of 5% dextrose),
`and apixaban (mg kg_1h"l) at 0.01, 0.03, 0.1, 0.3, and 1
`(n = 6 per group). The fondaparinux group treatment
`consisted of vehicle (saline), and fondaparinux (mg kg‘1 h")
`at 0.1, 0.3, 1, and 3 (n = 6 per group). The warfarin group
`treatment consisted of vehicle (water) (n = 6), and warfarin
`(mg kg_1 day_1) at 0.03, 0.1, 0.3, 1, and 3 (n = 6 per group).
`The ED50 (dose that increased carotid blood flow to 50% of the
`control) of compounds and the EC50 (plasma concentration
`that increased carotid blood flow to 50% of the control) of
`apixaban were estimated as described below.
`
`The rabbit cuticle BT model [4] was
`Cuticle bleeding model
`used in this study. Briefly,
`rabbits were anesthetized as
`described above. A standard cut was made at the apex of the
`cuticle with a razor blade. Blood was allowed to flow freely by
`keeping the bleeding site in contact with 37 °C lactated Ringer’s
`solution. BT was defined as the time after transection when
`
`bleeding ceased. It was measured by averaging the bleeding
`time of three nail cuticles. The maximum bleeding recorded was
`20 n1in. Apixaban,
`fondaparinux
`and warfarin were
`administered as described above.
`In rabbits treated with
`
`the BT effect was expressed as a ratio of
`anticoagulants,
`treated vs.
`the mean vehicle value. The ED3X (dose that
`increased BT 3-fold) values of compounds were estimated as
`described below.
`
`5
`E
`V
`
`43
`0'1
`
`The apixaban group treatment consisted of vehicle (10%
`N,N-dimethylacetamide; 30% 1,2-propanediol; 60% water)
`(n = 6), and apixaban (mg kg‘1h") at
`1
`(n = 6) and 3
`(n = 6). The fondaparinux group treatment consisted of
`vehicle (saline) (n = 6), and fondaparinux (mg kg] h_1) at
`0.3 (n = 6),
`1 (n = 6), and 3 (n = 6). The warfarin group
`treatment consisted of vehicle (water) (n = 6), and warfarin
`(mg kg") at 0.1 (n : 5), 0.3 (n : 5),
`1
`(n : 5), and 3
`(n = 5).
`
`Statistical analysis
`
`The statistical analyses used were analysis of variance and the
`Student—Newman—Keuls test using the SAS system (SAS for
`Windows release 8.02A; Cary, NC, USA). ED50 doses were
`determined using the
`four-parameter
`logistic
`equation,
`y: A + [(3 — A)/(1 + [(C/x)D])], where A : minimiun
`y value, B = maximmn, C = log ED50 and D = slope fac-
`tor, and the logistic fit was analyzed by XLfit© (IDBS,
`Bridgewater, NJ, USA). Antithrombotic ED50 values were
`determined using a maximmn value of 100 and a minimiun
`value of zero, whereas BT ED3X values were determined using a
`maximmn value (BT treated/BT control) of seven and a
`minimiun value of one. A value of P < 0.05 was considered
`
`statistically significant. All data are means 3: SE.
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`0 no inhibitor
`o 0.25 nM
`A 0.5 nM
`A 1 nM
`
`n 2 nM
`:1 4 nM
`
`Ono inhibitor
`00.25 nM
`A0.5 nM
`A1 nM
`-2 nM
`D 4 nM
`
`
`
`100
`
`200
`[$2765] (|JM)
`
`300
`
`400
`
`5 ,
`
`/-\ 4 —
`
`8 _
`g
`E 3
`V
`93
`2 2
`
`_
`
`1 _
`
`,
`
`+
`
`*
`
`0
`
`0.02
`
`0.04
`
`0.06
`
`0.08
`
`1/[S2765] (pm-‘)
`
`Fig. 1. Plot of apixaban inhibition of human FXa activity at different
`concentrations of the chromogenic peptide substrate S-2765. Top panel:
`Untransformed data. The solid lines are from non-linear fits of the data to
`
`the equation for competitive inhibition with K; = 0.075 :: 0.0031 nM,
`S-2765 Km = 31 :: 1.4 uM, and maximum rate = 5.6 :: 0.062
`mOD min'1 (moD = OD X 1000 where OD is optical density).
`Bottom panel: Lineweaver—Burk plot of the data as in the top panel.
`
`

`
`824 P. C. Wong et al
`
`Table 1 In vitro potency (K) of apixaban against human, rabbit, rat and dog FXa and the concentrations (ECZX) required to double the prothrombin time
`(PT), modified prothrombin time (mPT), activated partial thromboplastin time (APTT) and HepTest in human, rabbit, dog and/or rat plasma
`
`Species
`
`Human
`Rabbit
`Rat
`Dog
`
`FXa K, (nM)
`
`PT ECZX (uM)
`
`mPT ECZX (W)
`
`APTT ECZX (uM)
`
`HepTest ECZX (um)
`
`0.081
`0.16
`1.3
`1.7
`
`0.002*
`0.01 *
`0.1
`0.2
`
`36*
`2.3
`7.9
`6.7
`
`0.37
`0.6
`ND
`ND
`
`7 .4*
`4.8
`20
`> 20
`
`0.4
`1.8
`ND
`ND
`
`*Data from Pinto et al. [8]. ND, not determined.
`
`In vitro platelet aggregation responses
`Platelet aggregation
`to ADP,
`'y-thrombin and collagen averaged 47 J: 5%,
`53 J: 4% and 51 J: 5%, respectively in human PRP, and
`50 J: 5%, 56 J: 5% and 60 J: 1%, respectively, in rabbit
`PRP. These platelet responses were not significantly changed
`by apixaban at 1, 3 and 10 ]JM, but were almost completely
`inhibited by the GPIIb/IIIa antagonist DMP802 at 3 [JM (data
`not shown).
`
`In vivo studies
`
`AVST model Mean thrombus weights in the different
`vehicle-treated AVST rabbits were similar, and ranged from
`290 J: 11 to 327 J: 15 mg (n
`6 per group). As shown in
`Fig. 2, apixaban, fondaparinux and warfarin were efficacious
`in
`the AVST rabbits
`and produced
`dose-dependent
`antithrombotic effects;
`their ED50 values are reported in
`Table 2. At their top doses studied in this model, apixaban at
`3 mg kg] h_1 i.v., fondaparinux at
`1 mg kg] h_1 i.v. and
`warfarin at 3 mg kg] day_l p. 0. reduced thrombus weight by
`98%,
`86% and 77%,
`respectively,
`relative
`to
`their
`corresponding vehicle group. We observed that apixaban at
`0.03, 0.1, 0.3, 1 and 3 mg kg” h"' i.v. produced linear dose-
`proportional increases in plasma levels of 34 J: 2, 121 J: 9,
`490 J: 104, 1155 J: 153 and 3705 J: 525 nM,
`respectively
`(11
`3-7 per group). The EC50 for apixaban was estimated
`to be 357 J: 90 nM.
`
`VT model Mean thrombus weights in the different vehicle-
`treated VT rabbits were similar, and ranged from 64 J: 2 to
`79 J: 7 mg (n = 6 per group). In this model, apixaban,
`fondaparinux
`and warfarin
`produced
`dose-dependent
`antithrombotic effects (Fig. 2); their ED50 values are given in
`Table 2. At their top doses studied in this model, apixaban at
`1 mg kg] h_1, fondaparinux at 0.3 mg kg_1 h"l and warfarin
`at 3 mg kg_1day"1 reduced thrombus formation by 83%,
`74% and 84%, respectively, relative to their corresponding
`vehicle group.
`
`350
`
`300
`
`Rabbit AVST
`
`
`
`Thrombusweight(mg)
`
`250
`
`200
`
`50
`
`150
`
`100
`
`>(ov-_:v:_»~—(o
`0.00O
`
`>v—_Cq-CO
`O0
`
`Apixaban
`
`Fondaparinux
`
`Warfarin
`
`Dose (mg kg‘1 h"‘ IV or mg kg‘1 d“ P0)
`
`80
`
`O)O
`
`
`
`Thrombusweight(mg) 4>O
`
`20
`
`>
`
`0.03
`
`".03.
`O0
`
`Apixaban
`
`“.03.
`CO
`
`0.01
`
`0.03
`
`0.003
`Fondaparinux
`
`Warfarin
`
`Dose (mg kg‘1 h“ IV or mg kg‘1 d‘1 PO)
`
`Figure 3 (top panel) shows the
`Arterial thrombosis model
`effects of vehicle and apixaban on carotid blood flow after
`electrical stimulation. Basal carotid blood flow in the vehicle-
`
`the
`averaged 21 J: 4 mL min—1. After
`treated animals
`initiation of thrombosis, blood flow was gradually decreased,
`and the artery was totally occluded in about 35 min in vehicle-
`treated animals. Apixaban at 0.01—1 mg kg"1 h_1 i.v. produced
`
`Fig. 2. Antithrombotic effects in the arteriovenous-shunt thrombosis
`(AVST) and venous thrombosis (VT) rabbit models. Top panel: Effects of
`apixaban, fondaparinux and warfarin on thrombus formation in AVST.
`Means :: SE, and n = }7 per group; *P < 0.05 vs. the corresponding
`vehicle. Preliminary AVST data for apixaban at 0.0}1 mg kg'1 h‘1 were
`reported in Pinto et al. [7]. Bottom panel: Effects of apixaban, fonda-
`parinux and warfarin on thrombus formation in VT. Mean :: SE, and
`n = 5-6 per group; *P < 0.05 vs. the corresponding vehicle.
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`

`
`Apixaban in vitro/in vivo studies 825
`
`Table 2 In vivo potency of apixaban, fondaparinux and warfarin in rabbit models of arteriovenous-shunt thrombosis (AVST), venous thrombosis (VT),
`electrically mediated carotid arterial thrombosis (ECAT) and bleeding time (BT)
`
`Compound
`
`Apixaban
`Fondaparinux
`Warfarin
`
`AVST* EDSJ
`
`0.27 : 0.03
`0.05 : 0.01
`0.53 : 0.04
`
`VT EDSJ
`
`: 0.03
`0.11
`0.05 : 0.01
`0.27 : 0.01
`
`ECAT EDSJ
`
`0.02
`0.07
`0.27 :: 0.08
`0.08
`0.01
`
`BT ED3XT
`
`> 3
`> 3
`0.70
`
`0.07
`
`Results are expressed as mean :: SE. *Preliminary AVST data for apixaban were reported in Pinto et al. [7]. ‘Expressed in mg kg” h" i.v. for
`apixaban and fondaparinux and in mg kg"1 day'1 p.o. for warfarin.
`
`100
`
`80
`
`60
`
`’‘§
`'5
`3
`2,
`
`E,
`
`§
`3 40
`3-3
`
`Apixaban
`(mg kg"‘ h“ IV)
`
`-9- Vehicle
`
`-0- 0.01
`
`—E|- 0.03
`
`—I— 0.1
`
`fi 0.3
`
`—A— 1
`NO
`
`20
`
`Time after injury (min)
`
`9 8
`
`100
`
`CO0
`
`O)O
`
`-P0
`
`
`
`
`
`Integratedbloodflow(%control)
`
`V 0.010.030.1 0.3
`
`1
`
`V 0.1 0.3 1
`
`3
`
`V 0.03 0.1 0.3
`
`1
`
`3
`
`Apixaban
`
`Fondaparinux
`
`Warfarin
`
`Dose (mg kg“ h"‘ IV or mg kg“ d“ PO)
`
`Fig. 3. Antithrombotic effects in the rabbit model of electrically mediated carotid arterial thrombosis. Top panel: Effects of vehicle and apixaban
`on carotid blood flow (expressed as % of control carotid blood flow) after artery injury. Means :: SE, and n = 6 per group. Bottom panel: Dose-
`dependent effects of apixaban, fondaparinux and warfarin on integrated blood flow. Mean :: SE, and n = 6 per group. *P < 0.05 vs. the corresponding
`vehicle.
`
`a dose-dependent increase in duration of the patency of the
`injured artery. At 0.03—l mg kg_1 h"l
`i.v.,
`there was no
`occlusion in any of the animals up to 90 min.
`
`Figure 3 (bottom panel) also shows the effects of apixaban,
`fondaparinux and warfarin on integrated blood flow (ED50
`values are given in Table 2). We observed that apixaban at
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`

`
`826 P. C. Wong et al
`
`607 E] aP'|‘|'
`50*’ I PT
`
`*
`
`aPTTorPT(3)
`
`(A!C.’
`
`20‘
`
`10‘
`
`O
`
`>
`
`001
`
`003
`
`0.1
`
`03
`
`Apixaban
`
`Fondapari nux
`
`Dose (mg kg“ h“ IV or mg kg“ d" PO)
`
`Fig. 4. Ex vivo activated partial thromboplastin time (APTT) and prothrombin time (PT) effects of apixaban, fondaparinux and warfarin in rabbits.
`Means :: SE, and n = %19 per group. *P < 0.05 vs. the oorresponding vehicle.
`
`0.01, 0.03, 0.1, 0.3 and 1 mg kg_1 h_l i.v. produced linear dose-
`proportional increases in plasma levels of 13 J: 1, 45 :: 4,
`146 :: 8, 435 J: 25 and 1495 J: 50 nM, respectively (n = 6
`per group). The integrated blood flow EC50 for apixaban was
`estimated to be 106 J: 31 nM.
`
`Figure 4 shows the summary
`Ex vivo coagulation markers
`of the ex vivo APTT and PT responses
`to apixaban,
`fondaparinux and warfarin obtained from the AVST, VT
`and ECAT studies. Apixaban significantly prolonged ex vivo
`APTT at 3 mg kg] h_1, and PT at doses of 0.3 mg kg_1 h"l
`and higher (Fig. 4). Fondaparinux at the doses studied did not
`have significant effects on ex vivo APTT and PT. Warfarin
`significantly
`prolonged
`ex vivo APTT at
`doses
`of
`0.3 mg kg"1 day_1
`and higher,
`and PT at
`doses of
`0.1 mg kg"1 day_1 and higher (Fig. 4).
`Figure 5 (top panel) shows the ex vivo effect of apixaban on
`FXa and thrombin activity. Apixaban produced a dose-
`dependent inhibition of ex vivo FXa activity and did not
`change ex vivo thrombin activity. A good correlation was
`observed between the antithrombotic effect of apixaban and its
`ex vivo anti-FXa activity (r2 = 0.94), and also between the
`plasma
`concentrations
`and
`ex vivo
`anti-FXa
`activity
`(r2 = 0.90).
`
`Cuticle bleeding time model Mean cuticle BT in the vehicle-
`treated groups of apixaban, fondaparinux and warfarin were
`172 J: 2, 181 J: 7 and 183 J: 7 s, respectively (n = 6 per
`group). Warfarin at 0.1, 0.3,
`1 and 3 mg kg‘1day"1 p.o.
`produced dose-dependent
`increases
`in BT (228 J: 14,
`371 i 24, 929 J: 70 and 1129 J: 43 s, respectively), with an
`ED3X of 0.70 mg kg_1 day_1 (Table 2). As compared to
`vehicle, warfarin at 1 and 3 mg kg” day"' p.o. increased BT
`significantly by 5.1-fold and 6.2-fold, respectively (P < 0.05).
`
`In contrast, fondaparinux and apixaban at antithrombotic
`doses increased BT slightly (fondaparinux, 166 J: 4, 210 J: 12
`and 213 J: 11 s at 0.3, 1 and 3 mg kg] h_1 i.v., respectively;
`apixaban, 191 J: 8 and 228 J: 14 s at 1 and 3 mg kg” h"1
`i.v., respectively). At 3 mg kg‘1h"'
`i.v., fondaparinux and
`apixaban increased BT by 1.3-fold and 1.2-fold relative to their
`vehicle, respectively with ED3x of > 3 mg kg] h_1 for both
`compounds (Table 2).
`
`Discussion
`
`This study shows that apixaban is a potent, highly selective and
`direct inhibitor of FXa. It is also very efficacious for the
`prevention of arterial and venous thrombosis at doses that
`preserve hemostasis in rabbits.
`Analysis of enzyme kinetics shows that apixaban is a direct
`and competitive inhibitor of free hmnan FXa vs. a synthetic
`tripeptide substrate and does not need antithrombin III for its
`activity. It should be noted that apixaban is likely to behave as
`a mixed-type inhibitor of FXa activation of prothrombin in
`blood [18]. In addition, apixaban inhibited the prothrombin-
`ase-bound FXa activity in vitro, which resulted in the reduction
`of the conversion of prothrombin to thrombin with potency in
`the nanomolar range [18,19]. Together, these in vitro studies
`suggest that apixaban is a potent, selective, direct and effective
`inhibitor of both free and prothrombinase-bound FXa.
`Previously, we, as well as others, reported a species difference
`in FXa inhibition in hmnans, rabbits, rats and dogs with small-
`molecule, direct FXa inhibitors [6,20,21]. This study extends
`these findings, and shows that apixaban was more potent at
`prolonging APTT and PT in hmnan and rabbit plasma than in
`rat and dog plasma, which parallels its inhibitory potencies
`against human,
`rabbit,
`rat and dog FXa.
`Interestingly,
`apixaban was more potent in the prolongation of PT than
`
`© 2008 International Society on Thrombosis and Haemostasis
`
`

`
`FXa
`
`Thrombin
`
`Apixaban
`(mg kg"‘ h“ IV)
`
`100-
`
`- *
`
`Apixaban in vitro/in vivo studies 827
`
`sol
`
`c
`.9
`5 60*
`c
`E
`
`o\° 40
`
`20
`
`O_
`
`100
`
`so
`
`-5O
`
`N‘?
`
`0
`
`O
`
`
`
`
`
`
`
`Integratedbloodflow(O/0control)
`
`
`
`
`
`
`Apixabanplasmaconcentration(nM)
`100 1 0'
`
`E
`E
`E
`E
`E
`‘%'
`
`E
`
`E
`
`m
`
`01OOO
`
`1 000
`
`6*
`
`20
`
`40
`
`60
`
`80
`
`100
`
`EX vivo anti-FXa activity (%)
`
`0
`
`I
`l
`80
`60
`40
`20
`EX vivo anti-FXa activity (%)
`
`100
`
`Fig. 5. Ex vivo anti-FXa and antithrombin activities. Top panel: ex vivo anti-FXa and antithrombin effects of apixaban in arterial thrombosis rabbits from
`Fig. 3. Bottom panel: Correlation of ex vivo anti-FXa with antithrombotic effects and plasma concentrations of apixaban in arterial thrombosis rabbits.
`*P < 0.05, compared with the vehicle. Mean :: SE, and n = 6 per group.
`
`APTT in vitro and ex vivo. However, this effect may not extend
`to other direct FXa inhibitors, which may show varying effects
`on APTT and PT [3,4,2l]. Among the clotting assays studied,
`the mPT and HepTest have the highest sensitivity to apixaban
`in human plasma. Although the HepTest and mPT appear to
`be promising clotting assays to monitor the anticoagulant
`activity of apixaban, further clinical validation is needed.
`Because apixaban exhibited similar in vitro anti-FXa po-
`tency in humans and rabbits, we investigated its in vivo activity
`in well-characterized rabbit models of thrombosis and hemo-
`
`stasis. Although apixaban has hig