Thrombosis Research 131 (2013) 59–63

Contents lists available at SciVerse ScienceDirect

Thrombosis Research

j ourna l homepage: www.e lsev ie r .com/ locate / thromres

Regular Article

The effects of warfarin and edoxaban, an oral direct factor Xa inhibitor, ongammacarboxylated (Gla-osteocalcin) and undercarboxylated osteocalcin(uc-osteocalcin) in rats

Yoshiyuki Morishima, Chikako Kamisato, Yuko Honda, Taketoshi Furugohri ⁎, Toshiro ShibanoBiological Research Laboratories, R&D Division, Daiichi Sankyo Co., Ltd., 1-2-58 Hiromachi, Shinagawa-ku, Tokyo 140–8710, Japan

⁎ Corresponding author. Tel.: +81 3 3492 3131; fax:E-mail address: furugori.taketoshi.v5@daiichisankyo

0049-3848/$ – see front matter © 2012 Elsevier Ltd. Allhttp://dx.doi.org/10.1016/j.thromres.2012.08.304

a b s t r a c t

a r t i c l e i n f o

Article history:

Received 27 March 2012Received in revised form 7 August 2012Accepted 21 August 2012Available online 19 September 2012

Keywords:EdoxabanOsteocalcinOsteoporosisCarboxylationWarfarin

Introduction: Osteocalcin plays a role in bone homeostasis. The vitamin K cycle is essential for thegamma-carboxylation of glutamic acid residues in osteocalcin. Some evidence suggests that long-term war-farin therapy, which inhibits the vitamin K cycle and prevents gamma-carboxylation, is associated with in-creased bone-fracture risk. The aim of this study was to determine the effects of warfarin and edoxaban, adirect factor Xa inhibitor, on the serum concentration of total, gamma-carboxylated (Gla-osteocalcin) andundercarboxylated osteocalcin (uc-osteocalcin) in rats.Materials and Methods: Rats received orally administered warfarin or edoxaban, and 24 h later serum andplasma were prepared. Osteocalcin level in serum was measured with ELISA. A Gla-osteocalcin was precipi-tated by the addition of hydroxyapatite, and the resulting supernatant was used for measuringuc-osteocalcin. Prothrombin time (PT) of plasma was also measured.Results: Warfarin at 1 mg/kg (a dose which prolonged PT 2.62-fold) markedly increased the serum level of

uc-osteocalcin and slightly increased the total osteocalcin level compared with control in rats. SerumGla-osteocalcin significantly decreased by warfarin. Edoxaban at 1 mg/kg (an antithrombotic dose) and54 mg/kg (a dose which prolonged PT 2.25-fold) had no effects on total, uc-, and Gla-osteocalcin levels.Conclusions: This study demonstrates that warfarin impaired the carboxylation of osteocalcin in rats. In con-trast, edoxaban at or higher doses than needed for an antithrombotic effect sustained the circulatingGla-osteocalcin level. These findings suggest that edoxaban has no effects on the production ofGla-osteocalcin and thus, may have a lower risk of adverse effects on bone health.

© 2012 Elsevier Ltd. All rights reserved.

Introduction

Osteocalcin is a bone-specific protein produced by osteoblasts andcontains 3 gamma-carboxyl glutamic acid residues that facilitate thebinding of calcium and hydroxyapatite in bone. The serum osteocalcinconcentration is commonly used as a marker of bone formation/turnover[1,2]. The subfraction of osteocalcin that is an incomplete gamma-carboxylated form (undercarboxylated osteocalcin: uc-osteocalcin) isalso a marker of bone turnover. Increased serum uc-osteocalcin concen-tration is associated with an increased risk of hip fracture and lowbone-mineral density [3]. Thesemarkers of bone turnover are considereduseful diagnostic tools for the evaluation of bone formation and resorp-tion and prediction of the risk of osteoporosis.

Gamma-carboxylation of glutamic acid residues ismediated by vita-min K dependent gamma-glutamyl carboxylase. Therefore, serumuc-osteocalcin concentration is also a marker of vitamin K status inbone [4], i.e. low food intake of vitamin K resulted in elevated serumuc-osteocalcin level. Warfarin is an oral anticoagulant with vitamin K

+81 3 5436 8587..co.jp (T. Furugohri).

rights reserved.

antagonism. This drug inhibits gamma-carboxylation of coagulationfactors II, VII, IX and X, which in turn exerts an anticoagulant effect.Based on its mechanism of action, warfarin treatment increases circu-lating undercarboxylated osteocalcin, thus affecting uc-osteocalcin toosteocalcin ratio [5]. Furthermore, some reports demonstrate thatlong-termwarfarin therapy is associatedwith lowbonemineral density[6], increased osteoporosis and fracture risks [7,8], and warfarinembryopathy (severe bone defects during pregnancy) [9,10].

Edoxaban (the free form of edoxaban tosilate hydrate: JapaneseAccepted Name) is an oral direct factor Xa (FXa) inhibitor with highpotency and selectivity [11]. This compound is marketed for the pro-phylaxis of venous thromboembolism after orthopedic surgeries inJapan [12–14] and under evaluation in phase III clinical studies forthe prevention of stroke in patients with atrial fibrillation (ENGAGEAF-TIMI 48, NCT00781391) [15] and treatment and prevention of re-current venous thromboembolism (Hokusai VTE, NCT00986154).

Since the risks of osteoporosis and fracture are concerns of thelong-term anticoagulant therapy with vitamin K antagonists like war-farin, we evaluated the comparative effects of warfarin and edoxabanon the serum concentrations of total, gamma-carboxylated andundercarboxylated osteocalcin in rats.

Fig. 1. Effects of edoxaban and warfarin on serum total osteocalcin, undercarboxylatedosteocalcin (uc-osteocalcin), and gamma-carboxylated osteocalcin (Gla-osteocalcin)levels in rats. Edoxaban and warfarin were orally administered to rats 24 h beforeblood sampling. Data represent mean±SD (n=5). *** Pb0.001 compared with thecontrol group (Student t-test or Welch's test).

60 Y. Morishima et al. / Thrombosis Research 131 (2013) 59–63

Materials and Methods

Reagents and Drugs

Edoxaban tosilate hydrate (JAN) was synthesized at DaiichiSankyo Co., Ltd. (Tokyo, Japan). Warfarin sodium was purchasedfromWako Pure Chemical Industries, Ltd. (Osaka, Japan). Hydroxyap-atite was purchased from Sigma Aldrich, Inc. (St. Louis, MO) andOsteocalcin Rat ELISA system was purchased from GE HealthcareJapan (Hino, Japan). Thromboplastin C Plus and Platerin LS II wereobtained from Sysmex Corporation (Kobe, Japan) and Kyowa MedexCo., Ltd. (Tokyo, Japan), respectively. Human FXa and S-2222 wereobtained from Enzyme Research Laboratories, Inc. (South Bend, IN)and Chromogenix Instrumentation Laboratory S.p.A. (Milano, Italy),respectively. Thiopental sodium was obtained from MitsubishiTanabe Pharma Corporation (Osaka, Japan).

Animals

Animal facilities, animal care, and experimental procedures wereperformed in accordance with the in-house guidelines of the Institu-tional Animal Care and Use Committee of Daiichi Sankyo. Male Slc:Wistar rats (9 or 12-week-old) were purchased from Japan SLC, Inc.(Hamamatsu, Japan) and maintained on 7:00 am/7:00 pm light/dark schedule. Rats were housed 4 – 6 per cage and food and waterwere available ad libitum. They were acclimated for more than1 week.

Drug Administration and Preparation of Plasma and Serum

Edoxaban (1 or 54 mg/kg), warfarin (1 mg/kg), or solvent (0.5%methyl cellulose) were orally administered to rats (12 weeks old,n=5 per group). The animals were anesthetized with thiopental so-dium (100 mg/kg, i.p.) 24 h after the administration. Blood (0.54 ml)was collected from the jugular vein using a syringe containing 1/10volume (0.06 ml) of 3.13% sodium citrate tribasic dihydrate solutionfor anticoagulant and anti-FXa activity assay. Plasma was preparedfrom blood samples containing sodium citrate by centrifugation at1,500 ×g for 10 min at 4 °C. In addition, 3.5 ml of blood was collectedfrom the abdominal aorta using another syringe without the citratesolution for measurement of serum osteocalcin levels. Serum wasprepared from uncitrated blood samples containing serum separator(BD Vacutainer Blood Collection Tube 365920) by centrifugation at1,400 ×g for 10 min at room temperature.

Measurement of Serum Osteocalcin Levels

Osteocalcin levels in serum were measured with Osteocalcin RatELISA system. Uc-osteocalcin in serum were differentiated by absorp-tion of gamma-carboxylated osteocalcin (Gla-osteocalcin) by hy-droxyapatite (100 mg/ml) for 1 h at 4 °C, and then centrifuged at6,400 ×g for 5 min at 4 °C. The supernatant was used for measuringuc-osteocalcin. Gla-osteocalcin level in the serum sample was calcu-lated by subtracting uc-osteocalcin from total osteocalcin.

Anticoagulant and Anti-FXa Activity Assay

To evaluate the anticoagulant and anti-FXa activities of edoxaban,edoxaban (1 and 54 mg/kg) was orally administered to rats (9 weeksold, n=4 per group) and plasma was prepared 0.5, 1, 2, 4 and 24 hafter dosing.

Prothrombin time (PT) and activated partial thromboplastin time(APTT) were measured with a microcoagulometer, Amelung KC-10Amicro (Trinity Biotech Plc.). PT was measured by incubating plasma(50 μl) and saline (50 μl) for 1 min at 37 °C, followed by the additionof 100 μl Thromboplastin C Plus. APTT was measured by incubating

plasma (50 μl), saline (25 μl) and Platerin LS II (50 μl) for 5 min at37 °C, followed by the addition of 25 μl of 50 mM calcium chloridesolution. Anti-FXa activity in plasma was measured with a chromo-genic substrate S-2222. A plasma sample (5 μl) was added to the re-action mixture (95 μL) of human FXa (final conc. 0.016 IU/ml) andS-2222 (final conc. 300 mM). Amidolysis of S-2222 was measuredfor 10 min at room temperature.

Statistical Analysis

All data represent the mean±standard deviation (SD). The distri-butions of the data were determined by Shaprio-Wilk test using JMPrelease 9.0.2 (SAS Institute Inc., Cary, NC, USA). After confirmationof the normal distribution of the data, statistical significance betweenthe control and treatment groups was analyzed by F test for the ho-moscedasticity followed by Student t-test or Welch's test using SASrelease 8.2 (SAS Institute Japan, Tokyo, Japan). The statistical signifi-cance level was Pb0.05.

Results

Effects of Edoxaban and Warfarin on Serum Osteocalcin Levels in Rats

The serum level of total osteocalcin, uc-osteocalcin, and Gla-osteocalcin in control group was 52.4±6.3, 0.7±0.2, and 51.6±6.2 ng/ml, respectively (Fig. 1). In the warfarin group, both serumosteocalcin (70.1±3.0 ng/ml) and uc-osteocalcin (39.7±1.9 ng/ml)significantly increased as compared to the control group. On the otherhand, serum Gla-osteocalcin significantly decreased (30.5±3.0 ng/ml).These results indicate the inhibition of gamma-carboxylation ofosteocalcin (Fig. 1). In contrast, edoxaban (1 mg/kg and 54 mg/kg) hadno effects on total (57.6±7.4 and 57.7±4.5 ng/ml), uc- (0.8±0.2 and0.6±0.1 ng/ml) and Gla-osteocalcin (56.7±7.2 and 57.1±4.5 ng/ml)levels (Fig. 1).

Plasma Coagulation Times in Edoxaban- and Warfarin-treated Rats, andPlasma Anti-FXa Activity in Edoxaban-treated Rats

Warfarin at 1 mg/kg prolonged PT 2.62-fold and APTT 1.58-fold(Fig. 2A,B) 24 h after administration. This PT prolongation was within

61Y. Morishima et al. / Thrombosis Research 131 (2013) 59–63

the therapeutic range of warfarin. Edoxaban (1 mg/kg and 54 mg/kg)did not prolong PT or APTT 24 h after administration (Fig. 2A,B), butshowed 78% anti-FXa activity at 54 mg/kg (Fig. 2C).

Time Course of PT and Anti-FXa Activity of the Plasma from Rats OrallyAdministered Edoxaban

Since edoxaban is a reversible FXa inhibitor and its activity clearlydepends on plasma concentration [16], 24 h after administration isnot a suitable time to evaluate the activity of edoxaban. Therefore,we examined the effect of edoxaban on PT at earlier times (0.5 to4 h after the dosing). Edoxaban at 1 mg/kg did not prolong PT butdemonstrated approximately 60% anti-FXa activity for the 4 h afteradministration (Fig. 3A,B). Edoxaban at 54 mg/kg prolonged PT2.25-fold with ≥95% anti-FXa activity 1 h after administration andPT prolongation was sustained for at least 4 h, indicating that the an-ticoagulant effect at this dose was comparable to 1 mg/kg warfarin(Fig. 3A,B).

Discussion

Anticoagulant therapy with vitamin K antagonists like warfarinraises the concern about an increased risk of osteoporosis and warfa-rin embryopathy (severe bone defects during pregnancy), probablyrelated to its mechanism of action, vitamin K epoxide reductase inhi-bition. Vitamin K antagonists suppress the gamma-carboxylation ofglutamic acid residues not only in coagulation factors but also inosteocalcin, a bone-specific protein. Edoxaban is a new oral anticoag-ulant with direct FXa inhibition. We hypothesized that edoxaban hasno effects on the production of Gla-osteocalcin because the mecha-nism of action of edoxaban is unrelated to inhibition of vitamin Kcycle/gamma-carboxylation. In this study, we evaluated the compar-ative effects of edoxaban and warfarin on the serum concentrationsof total, gamma-carboxylated and undercarboxylated osteocalcin inrats.

Osteocalcin is synthesized and subsequently gamma-carboxylatedby vitamin K-dependent gamma-glutamyl carboxylase in osteoblasts.The resulting Gla-osteocalcin binds to hydroxyapatite (a kind of calci-um phosphate and the main component of bone and teeth) of boneand accumulates in the bone matrix. A portion of Gla-osteocalcin issecreted into blood, thus it is used as an indicator of osteogenesis[1,2]. In contrast, uc-osteocalcin, which is incomplete gamma-carboxylated form, has a low affinity to bone matrix; released intoblood, it serves as an index of bone resorption. In warfarin treated pa-tients, the uc/Gla-osteocalcin ratio was significantly increased [5].Long-term use of warfarin increased osteoporotic fracture risk in riband vertebra [7,8] by impairing cortical bone material quality as theosteocalcin content decreased [17], while some reports have shownthat warfarin use is not associated with osteoporotic fractures[19,20]. These conflicting results might be dependent on the durationof warfarin use and the skeletal sites [21]. The occurrence of warfarinembryopathy clearly demonstrates that warfarin has an adverse ef-fect on developing bone [21].

Long-term unfractionated heparin treatment is also associated withan elevated risk of osteoporosis, so called heparin-inducedosteoporosis.Some reports indicate that unfractionated heparin and low molecularweight heparins significantly inhibit proliferation of osteoblasts andprotein synthesis of bone metabolism related factors [22,23]. A neworal anticoagulant rivaroxaban (another FXa inhibitor) also significantlyreduced the osteoblast function (alkaline phosphatase activity) and the

Fig. 2. Plasma coagulation times in edoxaban- and warfarin-treated rats (A and B), andplasma anti-FXa activity in edoxaban-treated rats (C). Edoxaban and warfarin wereorally administered to rats. Blood samples were collected 24 h after dosing. PT (A),APTT (B) and FXa activity (C) of plasma were measured. Data represent mean±SD(n=5). *** Pb0.001 vs. control (Student t-test or Welch's test).

Fig. 3. Time course of PT and anti-FXa activity of the plasma from rats orally adminis-tered edoxaban. Blood samples were collected 0.5 to 4 h after dosing. PT (A) and FXaactivity (B) of plasma were measured. The data represent the mean±SD (n=4).

62 Y. Morishima et al. / Thrombosis Research 131 (2013) 59–63

mRNA expression of the bone marker like osteocalcin in osteoblasts[24]. However, fondaparinux, an antithrombin dependent FXa inhibitor,had no effects on the proliferation or protein synthesis of osteoblasts invitro [22,23]. Therefore, it is unlikely that the effects of unfractionatedheparin, low molecular weight heparins and rivaroxaban on the osteo-blast functions depend on FXa inhibition. Determination of the effect ofedoxaban on these osteoblast functions would clear the mechanisms ofthe anticoagulants in osteoblasts.

In this study, edoxaban or warfarin were orally administered torats. Serum total osteocalcin, uc-osteocalcin, and Gla-osteocalcinlevels were measured 24 h after administration. Twenty-four-hourwas a sufficiently long interval to evaluate the effect of warfarin ongamma-carboxylation of osteocalcin because warfarin prolonged PTmore than 2-fold at this time point and the half-life of osteocalcin inblood is about 5 min [18]. Warfarin impaired the carboxylation ofosteocalcin in rats at the dose which prolonged PT 2.62 fold. This PTprolongation is within a therapeutic range of warfarin, suggestingthat anticoagulant therapy with warfarin is associated with impair-ment of gamma-carboxylation of osteocalcin. In contrast, edoxabanleft unchanged the circulating active osteocalcin at an antithromboticdose (1 mg/kg) [11] or a 50-fold higher dose (54 mg/kg) whichprolonged PT comparable to warfarin (1 mg/kg). These findings

indicate that the anticoagulant effect of edoxaban is not associatedwith suppression of Gla-osteocalcin production and thus may beexpected to have lower risks of osteoporosis or embryopathy.

In conclusion, we demonstrated that warfarin impaired the car-boxylation of osteocalcin in rats, whereas edoxaban at antithromboticor higher doses did not reduce circulating active osteocalcin. Thesefindings suggest that edoxaban has no effects on the production ofGla-osteocalcin and thus may provide an alternative to vitamin-K an-tagonist anticoagulant therapy, which does not pose an increased riskof osteoporosis.

Conflict of Interest Statement

All authors are employees of Daiichi Sankyo Co., Ltd.

References

[1] Kruse K, Kracht U. Evaluation of serum osteocalcin as an index of altered bonemetabolism. Eur J Pediatr 1986;145:27-33.

[2] Slovik DM, Gundberg CM, Neer RM, Lian JB. Clinical evaluation of bone turnoverby serum osteocalcin measurements in a hospital setting. J Clin EndocrinolMetab 1984;59:228-30.

[3] Vergnaud P, Garnero P, Meunier PJ, Bréart G, Kamihagi K, Delmas PD.Undercarboxylated osteocalcin measured with a specific immunoassay predictship fracture in elderly women: the EPIDOS study. J Clin Endocrinol Metab1997;82:719-24.

[4] Sokoll LJ, O'Brien ME, Camilo ME, Sadowski JA. Undercarboxylated osteocalcin anddevelopment of a method to determine vitamin K status. Clin Chem 1995;41:1121-8.

[5] Menon RK, Gill DS, Thomas M, Kernoff PB, Dandona P. Impaired carboxylation ofosteocalcin in warfarin-treated patients. J Clin Endocrinol Metab 1987;64:59-61.

[6] Philip WJ, Martin JC, Richardson JM, Reid DM, Webster J, Douglas AS. Decreasedaxial and peripheral bone density in patients taking long-term warfarin. Q JMed 1995;88:635-40.

[7] Caraballo PJ, Heit JA, Atkinson EJ, Silverstein MD, O'Fallon WM, Castro MR, et al.Long-term use of oral anticoagulants and the risk of fracture. Arch Intern Med1999;59:1750-6.

[8] Gage BF, Birman-Deych E, Radford MJ, Nilasena DS, Binder EF. Risk of osteoporoticfracture in elderly patients taking warfarin. Results from the National Registry ofAtrial Fibrillation 2. Arch Intern Med 2006;166:241-6.

[9] Pettifor JM, Benson R. Congenital malformations associated with the administra-tion of oral anticoagulants during pregnancy. J Pediatr 1975;86:459-62.

[10] Pauli RM, Lian JB, Mosher DF, Suttie JW. Association of congenital deficiency ofmultiple vitamin K-dependent coagulation factors and the phenotype of the war-farin embryopathy: clues to the mechanism of teratogenicity of coumarin deriva-tives. Am J Hum Genet 1987;41:566-83.

[11] Furugohri T, Isobe K, Honda Y, Kamisato-Matsumoto C, Sugiyama N, Nagahara T,et al. Shibano T.DU-176b, a potent and orally active factor Xa inhibitor: in vitroand in vivo pharmacological profiles. J Thromb Haemost 2008;6:1542-9.

[12] Fujita S, Fuji T, Tachibana S, Nakamura M, Kawai M. Safety and efficacy ofedoxaban in patients undergoing hip fracture surgery. Pathophysiol HaemostThromb 2010;37:A95 [P366].

[13] Fuji T, Fujita S, Tachibana S, Kawai Y, Koretsune Y, Yamashita T, et al. Efficacy andsafety of edoxaban versus enoxaparin for the prevention of venous thromboem-bolism following total hip arthroplasty: STARS J-V trial. Blood 2010;116:3320[ASH Annual Meeting Abstracts].

[14] Fuji T, Fujita S, Tachibana S, Kawai Y. A dose-ranging study evaluating the oral fac-tor Xa inhibitor edoxaban for the prevention of venous thromboembolism in pa-tients undergoing total knee arthroplasty. J Thromb Haemost 2010;8:2458-68.

[15] Ruff CT, Giugliano RP, Antman EM, Crugnale SE, Bocanegra T, Mercuri M, et al.Evaluation of the novel factor Xa inhibitor edoxaban compared with warfarin inpatients with atrial fibrillation: design and rationale for the EffectiveaNticoaGulation with factor xA next GEneration in Atrial Fibrillation-Thrombolysis In Myocardial Infarction study 48 (ENGAGE AF-TIMI 48). AmHeart J 2010;160:635-41.

[16] Ogata K, Mendell-Harary J, Tachibana M, Masumoto H, Oguma T, Kojima M, et al.Clinical safety, tolerability, pharmacokinetics, and pharmacodynamics of thenovel factor Xa inhibitor edoxaban in healthy volunteers. J Clin Pharmacol2010;50:743-53.

[17] Sugiyama T, Takaki T, Sakanaka K, Sadamaru H, Mori K, Kato Y, et al.Warfarin-induced impairment of cortical bone material quality and compensato-ry adaptation of cortical bone structure to mechanical stimuli. J Endocrinol2007;194:213-22.

[18] Broulík P, Hrubý M. Acute effects of hemodialysis on circulating parathormone,osteocalcin and alkaline phosphatase in patients with chronic renal failure. CasLek Cesk 1993;132:681-3.

[19] Jamal SA, Browner WS, Bauer DC, Cummings SR. Warfarin use and risk for osteo-porosis in elderly women. Study of Osteoporotic Fractures Research Group. AnnIntern Med 1998;128:829-32.

63Y. Morishima et al. / Thrombosis Research 131 (2013) 59–63

[20] Pilon D, Castilloux AM, Dorais M, LeLorier J. Oral anticoagulants and the risk of os-teoporotic fractures among elderly. Pharmacoepidemiol Drug Saf 2004;13:289-94.

[21] Pearson DA. Bone health and osteoporosis: the role of vitamin K and potential an-tagonism by anticoagulants. Nutr Clin Pract 2007;22:517-44.

[22] Matziolis G, Perka C, Disch A, Zippel H. Effects of fondaparinux compared withdalteparin, enoxaparin and unfractionated heparin on human osteoblasts. CalcifTissue Int 2003;73:370-9.

[23] Handschin AE, Trentz OA, Hoerstrup SP, Kock HJ, Wanner GA, Trentz O. Effect oflow molecular weight heparin (dalteparin) and fondaparinux (Arixtra) onhuman osteoblasts in vitro. Br J Surg 2005;92:177-83.

[24] Solayar GN, Walsh PM, Mulhall KJ. The effect of a new direct Factor Xa inhibitor onhuman osteoblasts: an in-vitro study comparing the effect of rivaroxaban withenoxaparin. BMC Musculoskelet Disord 2011;12:247-54.