Journal of Atherosclerosis and Thrombosis Vol.18, No.8 659 Original Article

Antiplatelet Activity of Obovatol, a Biphenolic Component of Magnolia Obovata, in Rat Arterial Thrombosis and Rabbit Platelet Aggregation

Eun-Seok Park1, Yong Lim2, Seung-Ho Lee1, Byoung-Mog Kwon3, Hwan-Soo Yoo1, Jin-Tae Hong1, 4, and Yeo-Pyo Yun1

1College of Pharmacy, Chungbuk National University, Cheongju, Republic of Korea 2Department of Clinical laboratory Science, Dong-eui University, Busan, Republic of Korea 3 Laboratory of Chemical Biology and Genomics, Korea Research Institute of Bioscience and Biotechnology, Daejeon, Republic of Korea 4College of Pharmacy, Medical Research Center, Chungbuk National University, Cheongju, Republic of Korea

Aim: Thrombosis occurs in the coronary arteries via the activation of platelets, and leads to acute myocardial infarction and sudden death. Obovatol, a major biphenolic component of Magnolia obo- vata leaves, displays anti-inflammatory and acyl Co-A acyltrasferase inhibitory effects. The purpose of this study was to determine the effects of obovatol on thrombus formation in vivo and platelet activation in vitro and ex vivo. Methods: We investigated the antiplatelet and antithrombotic activities of obovatol in rat carotid ar- terial thrombosis in vivo along with platelet aggregation in vitro and ex vivo. Its possible cellular mechanism of antiplatelet activity was investigated by testing PLC-γ2 activation, arachidonic acid cascade, calcium mobilization and granule secretion. Results: Oral administration of obovatol prevented carotid thrombosis, but also significantly inhibit- ed collagen-induced platelet aggregation. Obovatol did not change coagulation times, such as activat- ed partial thromboplastin time and prothrombin time, indicating that the antithrombotic effect of obovatol might be due to antiplatelet activity rather than anticoagulation activity. Obovatol inhibited in vitro collagen- and arachidonic acid-induced rabbit platelet aggregation in a concentration-depen- dent manner (1-10 μM), with IC50 values of 2.4±0.8 and 4.8±0.9 μM, respectively. Obovatol blocked collagen-mediated phospholipase C-γ2 phosphorylation, cytoplasmic calcium mobilization, arachidonic acid liberation and serotonin secretion. Conclusion: Obovatol has a potent antithrombotic effect, which may be due to antiplatelet activity. The antiplatelet activity of obovatol is mediated by inhibition of PLC-γ2 phosphorylation. Thus, obovatol may be a potential candidate to treat cardiovascular disease.

J Atheroscler Thromb, 2011; 18:659-669.

Key words; Obovatol, Antiplatelet activity, PLC-γ2, Thrombosis, FeCl3-induced thrombosis model

array of ligands, such as collagen, von Willebrand fac- Introduction tor (vWF), ADP and thrombin, are generated at the Once blood vessels become damaged, a diverse injury site and induce platelet adhesion, followed by activation and aggregation1). Collagen enhances the Address for correspondence: Yeo-Pyo Yun, College of Pharmacy, adhesion of platelets to the injury site and induces Chungbuk National University, 48 Gaeshin-Dong, Heungduk- Gu, Cheongju, 361-763, Korea platelet activation through a tyrosine kinase-based sig- E-mail: [email protected] naling pathway. Collagen activates the kinase Syk and Received: October 24, 2010 phospholipase C-γ2 (PLC-γ2), and leads to increased Accepted for publication: February 15, 2011 cytoplasmic calcium, shape change and granule re- 660 Park et al. Antiplatelet Activity of Obovatol 661

lease2-5). Collagen-induced platelet activation results in tional University, Korea. platelet shape change and the release of substances that recruit activated platelets to the developing Arterial thrombus formation in vivo 6, 7) thrombus . The rat FeCl3 thrombosis model used for these When various inducers activate platelets, arachi- studies was modified from that described19). Obovatol donic acid is increased markedly via PLA2-mediated was orally administered to male Sprague-Dawley rats phospholipid hydrolysis8, 9). There are two main path- (240-260 g) daily at doses of 50 and 100 mg/kg for 3 ways of arachidonic acid metabolism in platelets. One days or carboxymethylcellulose (CMC) as a vehicle. is the cyclooxygenase (COX) pathway, which produc- The rats were anaesthetized with sodi- es TXA2; the other is the 12-lipoxygenase (12-LOX) um salt (60 mg/kg, intraperitoneally). A segment of pathway, which produces 12-hydroxy-eicosatetraenoic the right carotid artery was exposed and dissected free acid (12-HETE). 12-HETE plays a positive feedback of the vagus nerve and surrounding tissues and then role in thrombus formation in vivo10) and gives rise to fitted on the Doppler flow probe (1 mm diameter). platelet aggregation and aortic smooth muscle cell mi- Blood flow was measured with a Doppler velocimeter 11, 12) gration in vitro . 12-HETE and TXA2 appear to (Crystal Biotech, Hopkinton, MA, USA). After be involved in the initiation and propagation of 60-min administration of obovatol, thrombus was in- thrombotic and atherosclerotic disorders. Although duced by placing a 2-mm2 Whatman no. 1 filter pa- various arachidonic acid metabolites are produced per saturated with 50% ferric chloride (FeCl3) on the from COX and LOX, TXA2 and 12-HETE activities carotid artery near the probe for 10 min. The time play a major role in thrombosis13). needed for occlusion to occur was measured for up to Magnolia (Hou Po) was used medicinally as early 60 min, and the occlusion time was assigned as a val- as the first century and is a commonly prescribed Chi- ue of 60 min for vessels that did not occlude within nese traditional medicine. It possesses antioxidant ac- 60 min. tivity approximately 1000 times greater than α-to- copherol14). The biphenolic components from Magno- Platelet aggregation and coagulation times assay ex lia obovata, including , and obova- vivo tol, exhibit antiplatelet15), antitumor16) and antiin- Male Sprague-Dawley rats (240-260 g) were oral- flammatory activities17, 18). In our ongoing search for ly administered obovatol (50 and 100 mg/kg) for 3 antiplatelet agents from natural resources, we found days or CMC as a vehicle, and then blood was collect- that obovatol exhibited potent antiplatelet activity. In ed and platelet-rich plasma (PRP) was prepared by the study, we investigated the antiplatelet activity of centrifuging the blood sample at 180 ×g for 10 min obovatol by measuring in vitro and ex vivo agonist-in- and platelet-poor plasma (PPP) was obtained by cen- duced platelet aggregation. Its possible cellular mecha- trifuging the PRP at 2100 ×g for 10 min. PRP was nism of antiplatelet activity was investigated by testing adjusted to a density of 3×108 platelets/ml with PPP. PLC-γ2 activation, arachidonic acid cascade, calcium Platelet aggregation was measured as previously de- mobilization and granule secretion. We also evaluated scribed20) using an aggregometer (Chrono-Log Co., the antithrombotic activity of obovatol in the ferric Havertown, PA, USA). ADP (2 μM) and collagen (10 chloride-induced arterial thrombosis model in rats. μg/ml) were used as agonists for platelet aggregation. Anticoagulation activity was evaluated by measuring plasma-clotting times. The plasma activated partial Materials and Methods thromboplastin times (APTT) and prothrombin times Experimental animals (PT) were automatically measured using an Automat- Male Sprague-Dawley rats were purchased from ed Coagulation Laboratory 100 Instrument (Instru- Dae-Han Biolink Co. (Eum sung, Korea) and accli- mentation Laboratory Co., Milano, Italy). mated for 1 week at 24±1 ℃ and 55±5% humidity. Male New Zealand white rabbits (n =8) were pur- Measurement of washed rabbit platelet preparation chased from SamTako Bio Korea Inc. (Osan, Korea) and aggregation and acclimated for at least 1 week at 24±1 ℃ and 55 Rabbits were fixed in a cage. Blood was drawn ±5% humidity, with free access to drinking water and from the ear artery of conscious rabbits and collected a commercial pellet diet obtained from SamTako Bio directly into Vacutainer tubes containing anticoagu- Korea Inc. before experiments. Animal experiments lant solution (composed of 0.8% citric acid, 2.2% tri- were carried out in accordance with the Guide for the sodium citrate, and 2% dextrose (w/v)). PRP was pre- Care and Use of Laboratory Animals, Chungbuk Na- pared by centrifugation at 230 ×g for 10 min at room 660 Park et al. Antiplatelet Activity of Obovatol 661

temperature. Platelets were sedimented by centrifuga- samples were then centrifuged at 2100 ×g at 4℃ for tion of PRP at 2100 ×g for 10 min and then washed 10 min, and the upper phase was removed and evapo- twice with HEPES buffer (137 mM NaCl, 2.7 mM rated to dryness under . Residues were dis- KCl, 1 mM MgCl2, 5.6 mM glucose, 0.35% bovine solved in / (2:1, v/v) and applied serum albumin, and 3.8 mM HEPES, pH 6.5) con- to thin layer chromatography plates. The plates were taining 0.4 mM EGTA. After centrifugation, the pel- developed in petroleum ether//acetic acid lets were resuspended in HEPES buffer (pH 7.4). The (40:40:1, v/v/v). The area corresponding to arachi- platelet concentration was counted using a Coulter donic acid or diacylglycerol was scraped off, and ra- Counter (Coulter Electronics, Hialeah, FL, USA) and dioactivity was determined. adjusted to 4×108 platelets per milliliter. Platelet ag- gregation was measured as previously described20). Measurement of serotonin secretion Briefly, washed platelet suspensions were incubated at Serotonin concentration was determined by the 37℃ in the aggregometer with stirring at 1000 rpm. fluorimetric method of Jin et al.20). Washed platelets After incubation with various concentration of obova- were then pretreated with imipramine, serotonin reup- tol (1-10 μM), DMSO (0.5%) as a control for 3 min- take inhibitor (5 μM) and obovatol at 37 ℃ for 5 min, utes, platelet aggregation was induced by the addition and collagen (10 μg/mL) or arachidonic acid (100 μM) of collagen (10 μg/mL), arachidonic acid (100 μM), was subsequently added. The reaction was stopped by U46619 (1 μM), thrombin (0.05 U/mL), respectively. the addition of 5 mM EDTA on ice, and then the su- The resulting aggregation measured as the change in pernatant was centrifuged at 12000 ×g for 2 min. light transmission was recorded for 10 min. The supernatant was mixed with 6 M trichloroacetic acid (TCA) and centrifuged at 12000 ×g for 2 min. A Measurement of thromboxane B2, prostaglandin 0.6 mL aliquot of TCA supernatant was mixed with D2 and 12-HETE 2.4 mL of the solution (0.5% ο-phthalaldehyde in Thromboxane B2, prostaglandin D2 and 12-HETE ethanol diluted 1:10 with 8 N HCl), placed in a boil- generations were measured as previously described20). ing water bath for 10 min, and then cooled on ice. Ex- Briefly, washed rabbit platelets (4×108 platelets/mL) cess lipids were extracted with chloroform, and the were preincubated with various concentrations of ob- flurophore was measured at the excitation (360 nm) ovatol at 37℃ for 5 min and then further incubated and emission (475 nm) wavelengths. Serotonin creati- with a mixture of [3H]-AA (1 μCi/mL) and unlabeled nine sulfate was used as a standard solution to calcu- AA (2 μM) for 5 min. The reaction was terminated by late the extent of serotonin release. the addition of stop solution (2.6 mM EGTA, 130 μM BW755C, a dual COX and LOX inhibitor). Lipids Measurement of intracellular Ca2＋ mobilization were extracted and separated by silica gel G thin-layer Cytoplasmic Ca2＋ was measured as previously chromatography (TLC) on silica gel G plates (An- described20). In brief, rabbit platelets were incubated altech, Newark, DE). The plates were developed in eth- with 2 μM Fura-2/AM at room temperature for 60 yl acetate/isooctane/acetic acid/H2O (9:5:2:10, v/v/ min in loading buffer (137 mM NaCl, 27 mM KCl, v/v). The area corresponding to each lipid standard 0.4 mM NaH2PO4, 10 mM HEPES, 12 mM NaH- was scraped off, and radioactivity was determined by a CO3, 5.5 mM dextrose, and 0.35% BSA, pH 7.4). liquid scintillation counter (Beckman, Fullerton, CA, Excess Fura-2/AM was removed by centrifugation and USA). the platelets were suspended in fresh buffer without EGTA. Aliquots of platelet suspension (2.5 ml) were Measurement of arachidonic acid and diacylglycerol added to 4-ml cuvettes containing a Teflon-coated stir 2＋ Arachidonic acid liberation was measured as pre- bar (Chrono-Log Co.). Just before [Ca ]i was mea- viously described9). In brief, PRP was preincubated sured, Ca2＋ was added back to the buffer to a final with [3H]-AA (1 μCi/ml) at 37℃ for 1.5 hours and concentration of 1 mM, and then samples and ago- 3 8 2＋ then washed. The [ H]-AA-labeled platelets (4 ×10 nists were added. The measurement of [Ca ]i was platelets/mL) were pretreated with 100 μM BW755C performed at room temperature in an MSIII fluorom- and various concentrations of obovatol at 37℃ for 5 eter (Photon Technology International, Monmouth min in the presence of 1 mM CaCl2, and then stimu- Junction, NJ, USA) using excitation wavelengths of lated with collagen (10 μg/ml). The reaction was ter- 340 and 380 nm and an emission wavelength of 505 2＋ minated by the addition of chloroform/methanol/HCl nm. [Ca ]i was calculated using the SPEX dM3000 (200:200:1, v/v/v). After the mixture was vortexed, 5 software package. mM EGTA (containing 0.1 M KCl) was added. The 662 Park et al. Antiplatelet Activity of Obovatol 663

Western blotting assay tive compound was the same as reported previously23). For analysis of total platelet proteins, the reaction The dried and milled sample (1 kg) was soaked in was terminated by the addition of Laemmli sample chloroform/ (1:1, v/v, 5 L) at room tempera- buffer, and the mixture was then boiled for 5 min and ture for 5 days. The extract was concentrated under resolved by 7.5% SDS-PAGE. Western blotting assays reduced pressure and purified by silica gel, C18 col- were performed as previously described21). For West- umn chromatography and preparative thin layer chro- ern blotting assays, proteins were electrically trans- matography (TLC). Finally, we purified the active ferred to a polyvinylidene difluoride membrane for 80 compound by high-performance liquid chromatogra- min at 120 mA. Blots were incubated for 4 hours with phy (HPLC) (Phenomenex, Ultracarb 10ODS; 250× 5% (w/v) BSA in TBS to block residual protein bind- 21.2 mm at 285 nm, flow rate 3.5 mL/min) using a ing sites. Phospho-PLC-γ2 was detected by rabbit an- linear gradient rising from 80 to 90% methanol for 50 ti-phospho-PLC-γ2 antibody (Q-20, 1 μg/mL) in min. The tR in HPLC of active compound (120 mg) TBS containing 5% BSA for 4 hours. The primary was 35 min. The mass spectroscopic, infra-red spec- antibody was removed, and blots were washed three troscopic and nuclear magnetic resonance data for the times in TBS with 0.05% Tween-20. Blots were incu- purified compound were in good agreement with the bated with alkaline phosphatase-conjugated anti-rab- spectral data for obovatol (C18H18O3, mol. Wt 282 bit antibody (New England Biolabs, MA, USA) dilut- kDa) (Fig.1). ed to 1:5000 in TBS containing 5% BSA for 5 hours, and then washed five times in TBS with 0.05% Statistical analysis Tween-20. They were exposed to hyper-film-enhanced The experimental results are expressed as the chemiluminescence (Amersham Biosciences, Ltd., means±S.E.M. One-way analysis of variance (ANO- Buckinghamshire) for 5 min. The intensities of phos- VA) was used for multiple comparisons (GraphPad pho-PLC-γ2 and total PLC-γ2 bands were quantified Prism version 4.00 for Windows; San Diego, CA, by Scion-Image for Windows (Scion Corporation, USA). If there was a significant variation between MA, USA). treated groups, the Dunnett test was applied. The data were considered significant when P＜0.05. Materials The leaves of Magnolia obovata were harvested in Results Daejon, Korea, and identified by Dr. Byoung-Mog Kwon, Molecular Cancer Research Center, Korea Re- Extension of time to arterial thrombus formation search Institute of Bioscience and Biotechnology. A To investigate the effects of obovatol on patho- voucher specimen was deposited in the same laborato- logic and occlusive thrombus formation in vivo, rat ry. Obovatol (Fig.1) was isolated from Magnolia obo- carotid artery thrombus was induced by FeCl3 3 days vata leaves as described previously22, 23). Indomethacin, after oral administration of obovatol or carboxymeth- imipramine, serotonin creatinine sulfate and fura-2 ylcellulose as a vehicle control. The injured vessels acetoxymethyl ester (fura-2/AM) were purchased from were occluded within 12.8±1.3 min in the control Sigma Chemical Co. (St. Louis, MO, USA). [3H]-Ar- group. After oral administration of obovatol to rats, achidonic acid (220 μCi/mmol) was obtained from the occlusion times were significantly prolonged to Perkin-Elmer Life and Analytical Sciences (Boston, 22.4±5.4 and 31.6±6.4 min at doses of 50 and 100 MA, USA). Collagen, arachidonic acid and thrombin mg/kg body weight (P＜0.01, n =10), respectively were purchased from Chrono-Log Co. U46619 (Fig.2). Obovatol-treated groups at 50 and 100 mg/ (9,11-dideoxy-9α,11α-methanoepoxyprostaglandin kg showed prolongation of the thrombus time by 1.8- F2α), TXB2, PGD2 and 12-HETE were purchased and 2.5- fold, respectively, compared to the control. from Cayman Chemical Co. (Ann Arbor, MI, USA). Anti-PLC-γ2 and anti-phospho-PLC-γ2 (Q-20) anti- Inhibition of ex vivo platelet aggregation bodies were obtained from Upstate Biotechnology Inc. To distinguish whether the antithrombotic effect (Lake Placid, NY, USA). All other chemicals were of of obovatol was due to the inhibition of either platelet analytical grade. aggregation or coagulation activity, they were exam- ined after oral administration of obovatol at 50 and Isolation and purification of obovatol from 100 mg/kg to rats. Obovatol significantly inhibited Magnolia obovata collagen-induced platelet aggregation by 39.5±4.5 The leaves of Magnolia obovata were harvested in and 94.9±1.6% (P＜0.01, n =7) at doses of 50 and Daejeon, Korea and the isolation procedure of the ac- 100 mg/kg body weight, respectively; however, obova- 662 Park et al. Antiplatelet Activity of Obovatol 663

Fig.1. Chemical structure of obovatol Fig.2. Extension of arterial thrombus formation time in vivo Rat carotid artery was subjected to chemical injury by placing a 2-mm 2 Whatman no.1 filter paper saturated with 50 % FeCl3 on the carotid artery for 10 min, and then blood flow was measured with a Doppler velocimeter. Obovatol (50 and 100 mg/kg) was orally administered for 3 days, and the results are expressed as the oc- clusion time (n=10). Significantly different from control at ＊＊P＜0.01.

A B

Fig.3. Effect of obovatol on platelet aggregation and coagulation times ex vivo Obovatol (50 and 100 mg/kg) was orally administered for 3 days, and blood was collected 60 min after the last treatment. In the ex vivo plate- let aggregation assay, PRP was obtained by centrifugation blood samples at 180 ×g for 10 min. Platelet aggregation was induced by collagen (10 μg/mL) or ADP (2 μM) (A). In the coagulation assay, PPP was obtained by centrifuged at 2100 ×g for 10 min, and then APTT and PT were measured as described in Materials and Methods. Data (mean±S.E.M.) are from seven animals in each group. Significantly different from control at ＊＊p＜0.01.

tol failed to inhibit ADP-induced platelet aggregation tol does not change coagulation times (Fig.3B). (Fig.3A). The APTT and PT in the control group were 23.2±0.9 and 18.8±0.3 sec, respectively; how- Washed rabbit platelet aggregation in vitro ever, APTT and PT in obovatol-treated groups were Obovatol inhibited collagen- (10 μg/mL) and ar- 22.6±0.7 and 19.4±0.2 sec at a dose of 50 mg/kg, achidonic acid (100 μM) -challenged washed rabbit and 22.8±1.6 and 18.7±0.3 sec at a dose of 100 mg/ platelet aggregation in a concentration-dependent kg, respectively. Thus, these results suggest that obova- manner, with IC50 values of 2.4±0.8 and 4.8±0.9 μM, 664 Park et al. Antiplatelet Activity of Obovatol 665

respectively (Fig.4). However, obovatol had no effect on U46619- (1 μM) and thrombin (0.05 U/mL) -in- duced platelet aggregation, which indicates that ob- ovatol may be involved in the inhibition of collagen- mediated platelet aggregation. In addition, platelet vi- ability was determined by a MTT assay in order to ex- clude the possibility that the antiplatelet activity of obovatol was related to cytotoxity. Obovatol did not affect platelet viability compared with the control at concentrations of up to 50 μM (date not shown).

Inhibition of arachidonic acid metabolism In order to examine the effect of obovatol on the arachidonic acid metabolism pathways, washed rabbit platelets were preincubated with obovatol (5, 10 µM) and then further incubated with a mixture of [3H]-ar- achidonic acid and unlabeled arachidonic acid. Ob- Fig.4. Inhibition of washed rabbit platelet aggregation in vitro ovatol suppressed TXB2, PGD2 and 12-HETE genera- by obovatol tion induced by the addition of [3H]-arachidonic acid In the aggregation assay, the platelet suspension was incubated at to rabbit platelets in a concentration-dependent man- 37℃ in an aggregometer with stirring at 1000 rpm. After preincu- bation with obovatol for 3 min, platelet aggregation was induced ner (Fig.5). TXB2 production of obovatol at concen- by collagen (10 μg/mL), arachidonic acid (100 μM), U46619 (1 trations of 5 and 10 μM was inhibited by 80.8 and μM) and thrombin (0.05 U/mL). The degree of aggregation was 86.9% (Fig.5A), PGD2 generation by 49.3 and expressed as the percentage of maximum aggregation stimulated by 61.2% (Fig.5B) and 12-HETE production by 72.3 the respective inducers. Data are expressed as the mean±S.E.M. and 74.4% (Fig.5C), respectively. The inhibition of (n=3). TXB2 formation by obovatol occurred at a much low- er concentration, 5 μM, than by aspirin, a positive control, 50 μM, suggesting that obovatol may be a potent inhibition of TXB2 formation.

A B C

Fig.5. Inhibition of obovatol to arachidonic acid metabolism Washed rabbit platelets were preincubated with various concentrations of obovatol, DMSO (0.5%) as a control or aspirin (ASA, 50 μM) as a positive control for 5 min, and then further incubated with a mixture of [3H]-AA (1 μCi/mL) and unlabeled AA (2 μM) for 5 min. (A) 3 3 [ H]TXB2, (B) [ H]PGD2 and (C) 12-HETE were extracted and separated by TLC. The area corresponding to each lipid was scraped off and the radioactivity was determined by liquid scintillation counting. Data are expressed as the means±S.E.M. (n=3). Significantly different from control at ＊＊P＜0.01 664 Park et al. Antiplatelet Activity of Obovatol 665

A B

C D

Fig.6. Inhibitory effect of obovatol on the liberation of arachidonic acid and diacylglycerol, and and the secretion of serotonin (A) Arachidonic acid and (B) DAG liberation assays were performed, using [3H]-arachidonic acid-labeled platelets. After platelet suspensions were incubated with obovatol (1-10 μM) or with U73122 (50 μM), a PLC inhibitor, in the presence of 100 μM BW755C, the platelets were stimulated with 10 μg/mL collagen for 2 min. Washed rabbit platelet suspension was incubated with imipramine (5 μM), a serotonin reup- take inhibitor, and obovatol (1-10 μM) at 37℃ for 5 min prior to the addition of (C) collagen (10 μg/mL) or (D) arachidonic acid (100 μM). The serotonin concentration was determined by a fluorimetric method. Data are expressed as the means±S.E.M. (n=3). Significantly differ- ent from the control at ＊P＜0.05, ＊＊P＜0.01.

Decreased liberation of arachidonic acid and duced secretion of serotonin by 11.9, 22.1 and 70.9%, diacylglycerol and reduced secretion of serotonin respectively (Fig.6D). Pretreatment of rabbit platelets with obovatol at concentrations of 1, 3 and 10 μM significantly de- Inhibitory effect of obovatol on cytoplasmic creased collagen-induced arachidonic acid release from calcium mobilization membrane phospholipids by 14.3, 39.9 and 61.2% PLC catalyzes the simultaneously liberation of (Fig.6A). Collagen-induced diacylglycerol formation diacylglycerol and IP3, which induces calcium release of rabbit platelets was inhibited by obovatol (Fig.6B). from the dense tubular system leading to platelet ag- Inhibition by 1, 3 and 10μM obovatol of the diacylg- gregation. Cytosolic calcium mobilization is very im- lycerol liberation component to the concurrent con- portant in collage-mediated platelet aggregation and trols was 16.7, 60.3 and 71.1%, respectively. U73122, depends on the cytosol levels of IP3. We hypothesized a phospholipase C inhibitor, was used at 50 μM as a that obovatol may affect cytosolic calcium mobiliza- positive control and blocked the liberation of arachi- tion. Collagen- and arachidonic acid-induced cytosol- donic acid and diacylglycerol. Obovatol at 1, 3 and 10 ic calcium mobilization was completely blocked by μM reduced the collagen-induced secretion of sero- obovatol at 10 μM. These observations well correlated tonin in a concentration-dependent manner by 30.4, with the results of platelet aggregation (Fig.7). 56.4 and 86.6%, respectively (Fig.6C). Obovatol at 1, 5 and 10 μM also decreased the arachidonic acid-in- 666 Park et al. Antiplatelet Activity of Obovatol 667

A B

2＋ Fig.7. Inhibitory effect of obovatol on [Ca ]i mobilization in rabbit platelets caused by collagen and arachidonic acid

CaCl2 was added at a concentration of 1 mM, and fura 2/AM-loaded platelets were preincubated with obovatol (1-10 μM) for 5 min. (A) Ei- ther collagen (10 μg/mL) or (B) arachidonic acid (100 μM) was added. The traces shown are representative of results obtained from separate experiments conducted in triplicate, and the panels show the average data from separate experiments. Significantly different from the control at ＊＊P＜0.01 vs. stimulus control. Inhibitory effect of obovatol on PLC-γ2 phosphorylation Obovatol inhibited collagen-induced diacylglyc- erol formation and cytoplasmic calcium mobilization in rabbit platelets. Thus, we investigated the expres- sion of PLC-γ2, an upstream component of collagen- induced platelet aggregation, to determine whether PLC-γ2 may be a target for obovatol inhibition of platelet aggregation. Washed rabbit platelets were stimulated with 10 μg/mL collagen in the presence or absence of obovatol, and PLC-γ2 phosphorylation levels were examined. Pretreatment with obovatol at concentrations of 1, 3 and 10 μM significantly inhib- ited collagen-induced PLC-γ2 phosphorylation with inhibition percentages of 28.2, 47.2 and 96.6%, re- spectively (Fig.8).

Discussion Thrombosis occurs in the coronary arteries via the activation of platelets, and leads to acute myocar- 19) Fig.8. Inhibitory effect of obovatol on PLC-γ2 phosphoryla- dial infarction and sudden death . In the present tion study, we demonstrated that obovatol had antithrom- Washed rabbit platelets in the presence of EGTA (1 mM) and in- botic and antiplatelet activities. In the rat artery domethacin (20 μM) were incubated for 3 min with obovatol (1, thrombosis model, ferric chloride caused oxidative ar- 3 and 10 μM) or DMSO, and then stimulated with collagen (10 terial injury and exposed the subendothelial matrix. μg/mL) for 5 min. Cells were lysed and 20 μg/mL protein was an- Platelets interact with collagen and von Willebrand alyzed by SDS-PAGE. Relative densities were quantified by scan- ning densitometry and the density levels are shown as the relative factor (vWF) in the matrix via platelet membrane value of total PLC-γ2. Data are expressed as mean±S.E.M. (n=3). GPIb-V-IX and αIIbβ3. In platelets, glycoprotein VI Significantly different from the control at ＊＊p＜0.01. binding to collagen leads to platelet activation, fol- 666 Park et al. Antiplatelet Activity of Obovatol 667

lowed by calcium mobilization, the release of ADP tion2-4). Calcium mobilization is a major cellular pro- and TXA2, platelet recruitment and activation and cess in collagen-mediated platelet activation and is 24) 33) thrombus formation . Thus, platelet aggregation mainly IP3-dependent . Collagen- and arachidonic contributes to thrombus formation. Oral administra- acid-induced cytoplasmic calcium mobilization was tion of obovatol significantly increased the occlusion completely blocked by obovatol (Fig.7). Therefore, time (Fig.2), which indirectly indicated that obovatol we hypothesized that obovatol may also affect cyto- was able to inhibit thrombus formation in vivo via the plasmic calcium mobilization based on the inhibition inhibition of platelet aggregation, because obovatol of collagen-induced diacylglycerol formation and pre- did not prolong APTT or PT (Fig.3B). The exact sumably the elevation of IP3. Inhibition of calcium mechanism by which thrombus formation is triggered mobilization by obovatol in platelets was well correlat- in this model is not clear, but it has been shown that ed with the inhibition of platelet aggregation, and se- the morphology of the thrombi is similar to in hu- rotonin secretion from the dense tubular system in a mans25). Thus, the antithrombotic effect of obovatol concentration-dependent manner (Fig.6C and D). on the ferric chloride-mediated rat artery thrombosis Collagen-mediated cytoplasmic calcium mobili- model may be due to its antiplatelet activity, rather zation results from the activation of PLC-γ2 to gener- 26) 7) than anticoagulation activity . ate IP3 . Obovatol inhibited the collagen-induced vWF plays an important role in platelet adhesion phosphorylation of PLC-γ2 in a similar concentration and thrombus formation. vWF binds to the platelet range with the inhibition of platelet aggregation membrane GPIb-IX-V and initiates signals leading to (Fig.8). These results suggest that inhibition of PLC- platelet activation27). These events were promoted by γ2 phosphorylation by obovatol may be responsible the action of TXA2 produced from arachidonic acid. for the inhibition of collagen-mediated platelet aggre- TXA2 has proaggregatory and vasoconstrictor proper- gation. The intracellular target of obovatol in the col- ties that amplify the initiating stimulus and contribute lagen-mediated signaling transduction pathway re- to platelet plug formation28). Arachidonic acid released mains unknown. Obovatol had no effect on throm- from platelets is metabolized to TXA2 on the COX bin-induced platelet aggregation, while U73122, a pathway followed by PLC activation, an increase in PLC inhibitor, inhibited both collagen and thrombin- 2＋ 29, 30) [Ca ]i and protein kinase C activation . For this induced platelet aggregation (data not shown). Thus, reason, the arachidonic acid-mediated pathway plays a it is reasonable to speculate that the potential target central role in platelet activation by vWF. Collagen for obovatol may be involved in prior inhibition of the adheres to platelets, leading to platelet shape change, upstream signal(s) of PLC-γ2 rather than direct inhi- and the release of serotonin and TXA2 from dense bition of PLC, as obovatol had no effect on thrombin- granules. Obovatol potently inhibited collagen- and or U46619-induced platelet aggregation. The up- arachidonic acid-induced washed rabbit platelet aggre- stream component of PLC includes collagen receptor gation (Fig.4). In platelets, arachidonic acid is further and tyrosine kinase, which may be a potential target metabolized either to TXA2 and PGs through the of obovatol. Future studies of obovatol-induced inhi- COX pathway or to 12-HETE through the LOX bition of platelet aggregation need to be pursued. The pathway31). Obovatol inhibited the biosynthesis of inhibition of tyrosine kinase by obovatol may also both TXA2 and PGD2, and suppressed 12-HETE gen- contribute to the inhibition of collagen-induced plate- eration induced by arachidonic acid in platelets let aggregation, as polyphenols have been reported to (Fig.5), indicating that the release of arachidonic acid exhibit such inhibitions on tyrosine kinase34-36). Fur- is suppressed by obovatol, thereby inhibiting the pro- ther study of this issue is still needed. duction of TXA2, PDG2 and 12-HETE. Taken together, obovatol significantly protection Diacylglycerol and IP3 are produced from mem- against artery thrombosis in vivo, which may be due brane phospholipids by PLC-γ2 after stimulation to antiplatelet activity. In addition, obovatol inhibited with collagen. Diacylglycerol activates platelets via platelet aggregation induced by collagen and arachi- protein kinase C with the activation of integrin αIIbβ donic acid, and the cellular mechanism of the anti- 332). Obovatol inhibited collagen-stimulated arachi- platelet activity of obovatol appears to be mainly me- donic acid liberation and diacylglycerol formation diated by the inhibition of PLC-γ2 phosphorylation. (Fig.6). It has been reported that collagen induces This beneficial property of obovatol may be important platelet activation through a tyrosine kinase-based sig- in improving thrombosis because platelet aggregation naling pathway that includes Syk and PLC-γ27), lead- is an important determinant of thrombus initiation 2＋ ing to [Ca ]i increase, platelet shape change and and development, and obovatol may contribute to the dense granule release, adhesion and platelet aggrega- treatment of cardiovascular diseases. 668 Park et al. Antiplatelet Activity of Obovatol 669

Conflicts of interest cer Metastasis Rev, 1992; 11: 389-409 The authors state no conflicts of interest. 14) Lo YC, Teng CM, Chen CF, Chen CC, Hong CY: Magn- olol and honokiol isolated from Magnolia officinalis pro- tect rat heart mitochondria against lipid peroxidation. Acknowledgements Biochem Pharmacol, 1994; 47: 549-553 15) Pyo MK, Lee Y, Yun-Choi HS: Anti-platelet effect of the This work was supported by the Priority Re- constituents isolated from the barks and fruits of Magno- search Centers Program through the National Re- lia obovata. Arch Pharm Res, 2002; 25: 325-328 search Foundation of Korea (NRF) funded by the 16) Ikeda K, Nagase H: Magnolol has the ability to induce Ministry of Education, Science and Technology apoptosis in tumor cells. Biol Pharm Bull, 2002; 25: (2009-0094035), and by a Korea Research Founda- 1546-1549 tion Grant (MRC, R13-2008-001-00000-00). 17) Matsuda H, Kageura T, Oda M, Morikawa T, Sakamoto Y, Yoshikawa M: Effects of constituents from the bark of Magnolia obovata on production in lipopoly- saccharide-activated macrophages. Chem Pharm Bull (To- References kyo), 2001; 49: 716-720 1) Jackson SP, Schoenwaelder SM: Antiplatelet therapy: in 18) Park J, Lee J, Jung E, Park Y, Kim K, Park B, Jung K, Park search of the ‘magic bullet’. Nat Rev Drug Discov, 2003; 2: E, Kim J, Park D: In vitro antibacterial and anti-inflam- 775-789 matory effects of honokiol and magnolol against Propi- 2) Cowan DH, Robertson AL, Shook P, Giroski P: Platelet onibacterium sp. Eur J Pharmacol, 2004; 496: 189-195 adherence to collagen: role of plasma, ADP, and divalent 19) Kurz KD, Main BW, Sandusky GE: Rat model of arterial cations. Br J Haematol, 1981; 47: 257-267 thrombosis induced by ferric chloride. Thromb Res, 3) Cowan DH: Platelet adherence to collagen: role of prosta- 1990; 60: 269-280 glandin-thromboxane synthesis. Br J Haematol, 1981; 49: 20) Jin YR, Im JH, Park ES, Cho MR, Han XH, Lee JJ, Lim Y, 425-434 Kim TJ, Yun YP: Antiplatelet Activity of Epigallocatechin 4) Sixma JJ, van Zanten GH, Huizinga EG, van der Plas Gallate Is Mediated by the Inhibition of PLCgamma2 RM, Verkley M, Wu YP, Gros P, de Groot PG: Platelet ad- Phosphorylation, Elevation of PGD2 Production, and hesion to collagen: an update. Thromb Haemost, 1997; Maintaining Calcium-ATPase Activity. J Cardiovasc Phar- 78: 434-438 macol, 2008; 51: 45-54 5) Nieswandt B, Watson SP: Platelet-collagen interaction: is 21) Kim TJ, Zhang YH, Kim Y, Lee CK, Lee MK, Hong JT, GPVI the central receptor? Blood, 2003; 102: 449-461 Yun YP: Effects of on the serum- and platelet de- 6) Gibbins JM: Platelet adhesion signalling and the regula- rived growth factor-BB-induced proliferation of rat aortic tion of thrombus formation. J Cell Sci, 2004; 117: 3415- vascular smooth muscle cells. Planta Med, 2002; 68: 605- 3425 609 7) Farndale RW, Sixma JJ, Barnes MJ, de Groot PG: The 22) Ito K, Iida T, Ichino K, Tsunezuka M, Hattori M, Namba role of collagen in thrombosis and hemostasis. J Thromb T: Obovatol and obovatal, novel biphenyl ether Haemost, 2004; 2: 561-573 from the leaves of Magnolia obovata Thunb. Chem 8) Pollock WK, Rink TJ, Irvine RF: Liberation of [3H] ara- Pharm Bull (Tokyo), 1982; 30: 3347-3353 chidonic acid and changes in cytosolic free calcium in fu- 23) Kwon BM, Kim MK, Lee SH, Kim JA, Lee IR, Kim YK, ra-2-loaded human platelets stimulated by ionomycin and Bok SH: Acyl-CoA : cholesterol acyltransferase inhibitors collagen. Biochem J, 1986; 235: 869-877 from Magnolia obovata. Planta Med, 1997; 63: 550-551 9) Akiba S, Nagatomo R, Ishimoto T, Sato T: Effect of ber- 24) Furie B, Furie BC: Thrombus formation in vivo. J Clin bamine on cytosolic phospholipase A2 activation in rabbit Invest, 2005; 115: 3355-3362 platelets. Eur J Pharmacol, 1995; 291: 343-350 25) Farrehi PM, Ozaki CK, Carmeliet P, Fay WP: Regulation 10) Eynard AR, Tremoli E, Caruso D, Magni F, Sirtori CR, of arterial thrombolysis by plasminogen activator inhibi- Galli G: Platelet formation of 12-hydroxyeicosatetraenoic tor-1 in mice. Circulation, 1998; 97: 1002-1008 acid and thromboxane B2 is increased in type IIA hyper- 26) Yun YP, Do JH, Ko SR, Ryu SY, Kim JH, Song HC, Park cholesterolemic subjects. Atherosclerosis, 1986; 60: 61-66 YD, Ahn KS, Kim SH: Effects of Korean red ginseng and 11) Nakao J, Ooyama T, Chang WC, Murota S, Orimo H: its mixed prescription on the high molecular weight dex- Platelets stimulate aortic smooth muscle cell migration in tran-induced blood stasis in rats and human platelet ag- vitro. Involvement of 12-L-hydroxy-5,8,10,14-eicosatet- gregation. J Ethnopharmacol, 2001; 77: 259-264 raenoic acid. Atherosclerosis, 1982; 43: 143-150 27) Meyer D, Girma JP: von Willebrand factor: structure and 12) Nyby MD, Sasaki M, Ideguchi Y, Wynne HE, Hori MT, function. Thromb Haemost, 1993; 70: 99-104 Berger ME, Golub MS, Brickman AS, Tuck ML: Platelet 28) Kroll MH, Harris TS, Moake JL, Handin RI, Schafer AI: lipoxygenase inhibitors attenuate thrombin- and throm- von Willebrand factor binding to platelet GpIb initiates boxane mimetic-induced intracellular calcium mobiliza- signals for platelet activation. J Clin Invest, 1991; 88: tion and platelet aggregation. J Pharmacol Exp Ther, 1568-1573 1996; 278: 503-509 29) Siess W, Siegel FL, Lapetina EG: Arachidonic acid stimu- 13) Chen YQ, Liu B, Tang DG, Honn KV: Fatty acid modu- lates the formation of 1,2-diacylglycerol and phosphatidic lation of tumor cell-platelet-vessel wall interaction. Can- acid in human platelets. Degree of phospholipase C acti- 668 Park et al. Antiplatelet Activity of Obovatol 669

vation correlates with protein phosphorylation, platelet A: Evidence for a role for phospholipase C, but not phos- shape change, serotonin release, and aggregation. J Biol pholipase A2, in platelet activation in response to low Chem, 1983; 258: 11236-11242 concentrations of collagen. Thromb Haemost, 2001; 85: 30) Brass LF, Shaller CC, Belmonte EJ: Inositol 1,4,5-triphos- 882-889 phate-induced granule secretion in platelets. Evidence 34) Tachibana H, Koga K, Fujimura Y, Yamada K: A receptor that the activation of phospholipase C mediated by plate- for green tea polyphenol EGCG. Nat Struct Mol Biol, let thromboxane receptors involves a guanine nucleotide 2004; 11: 380-381 binding protein-dependent mechanism distinct from that 35) Stoclet JC, Chataigneau T, Ndiaye M, Oak MH, El of thrombin. J Clin Invest, 1987; 79: 1269-1275 Bedoui J, Chataigneau M, Schini-Kerth VB: Vascular pro- 31) Coleman RA, Smith WL, Narumiya S: International tection by dietary polyphenols. Eur J Pharmacol, 2004; Union of Pharmacology classification of prostanoid recep- 500: 299-313 tors: properties, distribution, and structure of the recep- 36) Wright B, Moraes LA, Kemp CF, Mullen W, Crozier A, tors and their subtypes. Pharmacol Rev, 1994; 46: 205-229 Lovegrove JA, Gibbins JM: A structural basis for the inhi- 32) Blockmans D, Deckmyn H, Vermylen J: Platelet activa- bition of collagen-stimulated platelet function by querce- tion. Blood Rev, 1995; 9: 143-156 tin and structurally related . Br J Pharmacol, 33) Lockhart LK, Pampolina C, Nickolaychuk BR, McNicol 2010; 159: 1312-1325