Modulation of human platelet activation and in vivo vascular thrombosis by columbianadin: regulation by integrin αIIbβ3 inside-out but not outside-in signals

Background Columbianadin (CBN) is one of the main coumarin constituents isolated from Angelica pubescens. The pharmacological value of CBN is well demonstrated, especially in the prevention of several cancers and analgesic activity. A striking therapeutic target for arterial thrombosis is inhibition of platelet activation because platelet activation significantly contributes to these diseases. The current study examined the influence of CBN on human platelet activation in vitro and vascular thrombotic formation in vivo. Methods Aggregometry, immunoblotting, immunoprecipitation, confocal microscopic analysis, fibrin clot retraction, and thrombogenic animals were used in this study. Results CBN markedly inhibited platelet aggregation in washed human platelets stimulated only by collagen, but was not effective in platelets stimulated by other agonists such as thrombin, arachidonic acid, and U46619. CBN evidently inhibited ATP release, intracellular ([Ca2+]i) mobilization, and P-selectin expression. It also inhibited the phosphorylation of phospholipase C (PLC)γ2, protein kinase C (PKC), Akt (protein kinase B), and mitogen-activated protein kinases (MAPKs; extracellular signal-regulated kinase [ERK] 1/2 and c-Jun N-terminal kinase [JNK] 1/2, but not p38 MAPK) in collagen-activated platelets. Neither SQ22536, an adenylate cyclase inhibitor, nor ODQ, a guanylate cyclase inhibitor, reversed the CBN-mediated inhibition of platelet aggregation. CBN had no significant effect in triggering vasodilator-stimulated phosphoprotein phosphorylation. Moreover, it markedly hindered integrin αIIbβ3 activation by interfering with the binding of PAC-1; nevertheless, it had no influences on integrin αIIbβ3-mediated outside-in signaling such as adhesion number and spreading area of platelets on immobilized fibrinogen as well as thrombin-stimulated fibrin clot retraction. Additionally, CBN did not attenuate FITC-triflavin binding or phosphorylation of proteins, such as integrin β3, Src, and focal adhesion kinase, in platelets spreading on immobilized fibrinogen. In experimental mice, CBN increased the occlusion time of thrombotic platelet plug formation. Conclusion This study demonstrated that CBN exhibits an exceptional activity against platelet activation through inhibition of the PLCγ2-PKC cascade, subsequently suppressing the activation of Akt and ERKs/JNKs and influencing platelet aggregation. Consequently, this work provides solid evidence and considers that CBN has the potential to serve as a therapeutic agent for the treatment of thromboembolic disorders.


Introduction
Arterial thrombosis can lead to the development of cardiovascular diseases (CVDs) such as myocardial infarction, atherosclerosis, and even ischemic stroke. When vascular subendothelial connective tissues are exposed due to injury, platelets move, adhere at the site of injury, and subsequently initiate the vascular thrombosis. Collagen contained in the basement membrane induces a change in shape from discoid to spheroid with pseudopodic projections of platelets. The combination of platelet secretion from the granules contain ADP/ATP, Ca 2+ , and fibrinogen, allows engagement of platelet receptors initiates intra-platelet signaling pathways, which activates platelet integrin α IIb β 3 and enables platelet aggregation [1]. In resting platelets, integrin α IIb β 3 exists in a low activation state and is unable to interact with its specific ligands such as fibrinogen, fibronectin, and von Willebrand factor. Platelet activation stimulated by various agonists induces a conformational change in integrin α IIb β 3 , enabling it to bind to its ligands, resulting in the onset of platelet aggregation; this process is known as inside-out signal transduction [1]. Moreover, the binding of fibrinogen to activated integrin α IIb β 3 initiates a series of intracellular signaling events, such as tyrosine phosphorylation of numerous proteins and cytoskeleton reorganization; this process is referred to as outside-in signaling [1]. These outside-in reactions, originating in the integrin α IIb β 3 bound to fibrinogen, are required for maximal secretion, procoagulation, and clot retraction [1].
Columbianadin (CBN; Fig. 1a) is a natural coumarintype compound isolated from the root of Angelica pubescens Maxim. f. biserrata Shan et Yuan, which is mainly used to treat rheumatism, spasm, and headache in clinics, according to Chinese Pharmacopoeia [2]. Many coumarin derivatives are isolated from A. pubescens, of which CBN is one of the main bioactive constituents. CBN has attracted considerable attention due to its pharmacological properties, such as prevention of several types of cancers (i.e., human leukemia, bladder carcinoma, and colon cancer). It can effectively suppress the growth of colon cancer cells by inducing apoptosis at low concentrations (~25 μM) and necroptosis at high concentrations (50 μM). The induction of apoptosis by CBN is correlated with the modulation of caspase-9, caspase-3, Bax, Bcl-2, Bim, and Bid, and the induction of necroptosis is related to receptor-interacting protein kinase-3 and caspase-8 [2]. CBN also possesses analgesic properties. Moreover, it causes the inhibition of inflammatory responses, which markedly inhibited edema and the vascular permeability in mice and reduced the inflammatory response in LPSinduced lung injury through the downregulation of inducible nitric oxide synthase in mice [3,4].
CBN was also preliminary reported to exhibit antiplatelet activity stimulated by ADP in rat platelets [5]; however, the effects and mechanisms of this compound on human platelets have not been investigated. Our initial screening exhibited that CBN significantly inhibits aggregation in human platelets. This result inspired us to conduct a thorough investigation on the influence of CBN on human platelets to support the scientific rationale for its clinical use (i.e., to treat CVDs).

Platelet preparation, aggregation, and ATP release
This study complied with the directives of the Helsinki Declaration and was approved by the Institutional Review Board of Taipei Medical University. Informed consent was obtained from all human volunteers who participated in this study. Washed human platelets (3.6 × 10 8 cells/mL) were prepared as described previously [6], and CBN (10-100 μM), aspirin (20-100 μM) or solvent control (0.1% DMSO) was incubated with the platelets for 3 min before stimulation. ATP release was measured using Hitachi Spectrometer F-7000 (Tokyo, Japan) according to the manufacturer's protocol.

Intracellular [Ca 2+ ]i mobilization and lactate dehydrogenase assays
To measure the intracellular calcium [Ca 2+ ]i, citrated whole blood was centrifuged, and the supernatant was incubated with 5 μM Fura 2-AM, which was then Fig. 1 Inhibitory activities of columbianadin (CBN) on platelet aggregation and cytotoxicity stimulated by agonists. Washed human platelets (3.6 × 10 8 cells/mL) were preincubated with a solvent control (0.1% DMSO), CBN (40-100 μM) (a, chemical structure) or aspirin (20-100 μM) and subsequently treated with 1 μg/mL of collagen, 0.01 U/mL of thrombin, 1 μM U46619, or 60 μM of arachidonic acid (AA) to stimulate platelet aggregation (b). Concentration-response histograms of CBN in inhibition of platelet aggregation (%) (c). To assess the cytotoxicity (d), platelets were preincubated with the solvent control (0.1% DMSO) or CBN (60, 80, and 160 μM) for 20 min, and a 10-μL aliquot of the supernatant was deposited on a Fuji Dri-Chem slide LDH-PIII. Data (c and d) are presented as the means ± SEM (n = 4) measured using a Hitachi Spectrometer F-7000 (Tokyo, Japan). [Ca 2+ ]i was measured at excitation wavelengths of 340 and 380 nm and an emission wavelength of 500 nm [7]. Furthermore, cytotoxic effect was examined by determining the level of lactate dehydrogenase (LDH). Washed platelets were preincubated with CBN (60, 80, and 160 μM) or 0.1% DMSO for 20 min at 37°C. An aliquot of the supernatant (10 μL) was deposited on a Fuji Dri-Chem slide LDH-PIII (Fuji, Tokyo, Japan), and the absorbance was read using a spectrophotometer (UV-160; Shimadzu, Japan). The maximal value of LDH was observed in triton-treated platelets.
Surface P-selectin expression and integrin α IIb β 3 activation Briefly, washed platelets were preincubated with CBN (60 and 80 μM) and the FITC-conjugated anti-P-selectin mAb (2 μg/mL) or PAC-1 mAb (2 μg/mL) for 3 min and then stimulated by collagen (1 μg/mL). For other experiments, fluorescence-conjugated triflavin, a specific integrin α IIb β 3 antagonist, was prepared as described previously [8]. The final concentration of FITC-triflavin was adjusted to 1 mg/mL. Washed platelets were preincubated with EDTA (2 mM), CBN (60 and 80 μM) or solvent control (0.1% DMSO), followed by the addition of FITCtriflavin (2 μg/mL) for 3 min. The suspensions were then assayed for fluorescein-labeled platelets on a flow cytometer (FAC Scan system, Becton Dickinson, San Jose, CA, USA). Data were collected from 50,000 platelets per experimental group, and the platelets were identified based on their characteristic forward and orthogonal light-scattering profiles. All experiments were repeated at least four times to ensure reproducibility.

Confocal microscopic analysis of platelet adhesion and spreading on immobilized fibrinogen
Platelet spreading on immobilized fibrinogen was analyzed as described previously [9]. In brief, platelets were stained with FITC-labeled phalloidin and visualized with a Leica TCS SP5 microscope equipped with a 100×, 1.40 NA oil immersion objective (Leica, Wetzlar, Germany). The number of platelet adhesion events and the platelet spreading surface area were determined using the NIH ImageJ software (NIH, Bethesda, MD; http://rsbweb.nih.gov/ij/).

Platelet-mediated fibrin clot retraction
Washed platelets were suspended in Tyrode's solution containing 2 mg/mL fibrinogen and 1 mM CaCl 2 in tubes designed for aggregation [10]. The platelet suspensions were preincubated in CBN (60 and 80 μM) or 0.1% DMSO for 3 min prior to thrombin (0.01 U/mL)-induced clot retraction without stirring. The reaction was photographed at 15 and 30 min, respectively.

Immunoblotting
Washed platelets (1.2 × 10 9 cells/mL) were preincubated with CBN (60 and 80 μM) or 0.1% DMSO, and collagen was subsequently added to trigger activation. The platelet suspensions were lysed and separated on a 12% SDS-PAGE. Several proteins were detected by specific primary antibodies. Respective quantitative results were obtained by quantifying the optical density of protein bands using a video densitometer and Bio-profil Biolight software, Version V2000.01 (Vilber-Lourmat, Marne-la-Vallée, France).

Immunoprecipitation
Dishes (6-cm diameter) were precoated with fibrinogen (100 μg/mL) overnight and then blocked with 1% BSA. Washed platelets were preincubated with CBN (80 μM) or the solvent control (0.1% DMSO) for 3 min and then allowed to spread on dishes for 60 min. The platelets were lysed and centrifuged; subsequently, Protein G Mag Sepharose Xtra beads (10 μL) was added and the platelets were incubated with the anti-integrin β 3 mAb (1 μg/mL) for immunoblotting as described previously.

Vascular thrombus formation in mouse mesenteric microvessels irradiated by sodium fluorescein
The method applied to the thrombogenic animal model in this experiment conformed to the Guide for the Care and Use of Laboratory Animals (8th edition, 2011), and we received an affidavit of approval for the animal use protocol from Taipei Medical University. In brief, external jugular veins of mice (6 weeks old) were cannulated with a polyethylene (PE)-10 tube for administration of the sodium fluorescein (15 μg/kg) and CBN (5 and 10 mg/kg) intravenously as described previously [11]. Venules (30-40 μm) were irradiated with wavelengths of < 520 nm to produce a microthrombus, and the time required for the thrombus to occlude the microvessel (occlusion time) was recorded.

ADP-induced acute pulmonary thromboembolism in mice
Acute pulmonary thromboembolism was induced according to a previously described method [12]. Various doses of CBN (5 and 10 mg/kg), aspirin (20 mg/kg) or solvent control (0.1% DMSO) (all in 50 μL) were administered through intraperitoneal injection to mice. After 5 min, ADP (0.7 mg/g) was injected into the tail vein. The mortality of mice in each group within 10 min after injection was determined.

Statistical analysis
The results are expressed as the means ± SEM and are accompanied by the number of observations (n). n refers to the number of experiments, and each experiment was conducted using different blood donors. The unpaired Student's t test or analysis of variance was used to determine the significant differences among the groups. When this analysis indicated significant differences, the groups were compared using the Student-Newman-Keuls method. Statistical significance was set at p < 0.05.

Results
Inhibitory activities of CBN in platelet aggregation stimulated by various agonists (Fig. 1a). Li et al. [5] reported that CBN significantly inhibits rat platelet aggregation stimulated by ADP. However, no other study has reported this effect of CBN. In the current study, CBN (40-80 μM) more selectively inhibited human platelet aggregation stimulated by collagen (1 μg/mL) than AA, thrombin, or U46619 (a thromboxane A 2 receptor agonist). Although CBN slightly but no significantly inhibited platelet aggregations even at concentrations up to 100 μM (Fig.  1b-c). The 50% inhibitory concentration (IC 50 ) of CBN for collagen-induced platelet aggregation was approximated at 60 μM (Fig. 1c). Moreover, aspirin (20, 50, and 100 μM) concentration-dependently inhibited platelet aggregation stimulated by collagen (1 μg/mL), and its IC 50 value was approximated at 70 μM ( Fig. 1b and c). The solvent control (0.1% DMSO) did not exert any significant effects on platelet aggregation (Fig. 1b). The LDH study revealed that CBN (60, 80, and 160 μM) did not alter LDH release or have any cytotoxic effects on platelets (Fig. 1d). This result revealed that CBN did not affect platelet permeability or induce platelet cytolysis.
The antithrombotic activity of CBN was observed in experimental mice. The occlusion time in the mesenteric microvessels of mice pretreated with 15 μg/kg fluorescein sodium was approximately 120 s. The resulting occlusion times were significantly extended after 5 and 10 mg/kg CBN treatments compared with those after 0.1% DMSO treatment (control vs. 5 mg/kg CBN, 118.0 ± 14.1 s vs. 113.1 ± 11.2 s, n = 8, p > 0.05; control vs. 10 mg/kg CBN, 102.3 ± 21.5 s vs. 377.7 ± 41.2 s, n = 8, p < 0.001; Fig. 7d). After irradiation, a thrombotic platelet plug was observed in the mesenteric microvessels at 5 and 150 s, in either 5 mg/kg CBN-or 0.1% DMSOtreated group (Fig. 7d; left panel, arrows). On administration of 10 mg/kg CBN, platelet plug formation was only observed at 5 s, but not at 150 s after irradiation (Fig. 7d). Furthermore, we also investigated and compared the therapeutic effects of CBN with aspirin in preventing acute pulmonary embolism death in mice as shown in Table 1. The results indicated that treatment with CBN at 5 and 10 mg/kg significantly lowered the ADP (0.7 mg/g)-induced mortality rate in mice from 100% (8 dead, n = 8) to 50% (4 dead, n = 8), and 0% (0 dead, n = 8), respectively. In addition, aspirin (20 mg/kg) also reduced the mortality to 25% (6 dead, n = 8) in this experiment (Table 1).

Discussion
This study reveals that in addition to the well-known properties of CBN, it also possesses antiplatelet activity in humans. It can be satisfactorily absorbed from gastrointestinal tract into bloodstream and distributed into organs [19]. Thus, the intake of CBN or natural of nontoxic prophylactic agents, such as food products and nutritional supplements, is ideal to prevent atherothrombotic events.
In the current study, CBN more potently inhibited collagen-induced platelet aggregation, but only slightly (not statistically significant) inhibited other platelet agonists; this implied that CBN was effective in inhibiting platelet aggregation through a prominent PLC-dependent mechanism. The platelet stimulation by agonists, for example collagen, noticeably modified phospholipase activation. The PLC activation resulted in IP 3 and DAG formation, which activated PKC, inducing p47 protein phosphorylation [20]. PLC enzyme is composed of several subtypes in which PLCγ family can be further divided into two isozymes, namely PLCγ1 and PLCγ2. PLCγ2 participates in collagendependent signaling in platelets [21]. In our present study, CBN reduced the collagen-activated PLCγ2/ PKC phosphorylation but without inhibition of PDBuinduced platelet aggregation; this suggested that CBN had no direct effects on PKC. Akt (downstream regulator of PI3K)-knockout mice have defective platelet activation [22]. Hence, Akt activation may be an attractive target for the development of antithrombotic therapeutics. Although effectors through which Akt contributes to platelet activation are not definitively known, several candidates have been discussed, including glycogen synthase kinase 3β, phosphodiesterase 3A, and the integrin β 3 [22]. Additionally, it has been observed that both PI3K/Akt and MAPKs are mutually activated and PKC is the upstream regulator in platelets [23].
MAPKs constitute a family of serine/threonine kinases that convert extracellular stimuli into cellular responses. Conventional MAPKs can be divided into the ERK1/2, p38 MAPK (α, β, γ, and δ), JNK1/2, and big MAPK (ERK5) [24]. The ERK1/2, JNK1/2, and p38 MAPK have been found to participate in platelet activation [24]. All of these kinases are activated by specific MAPK kinases (MEKs). The intracellular roles of JNK1/2 and ERK1/2 in platelets remain unclear, but evidence shows that the suppression of integrin α IIb β 3 activation may be involved [25]. Moreover, ERK activation is essential for collageninduced platelet aggregation [26]. Cytosolic phospholipase A 2 (cPLA 2 ), which catalyzes AA release to produce thromboxane A 2 , which is an important substrate of p38 MAPK activation induced by various platelet agonists such as thrombin [27]. The present study revealed that CBN-mediated inhibition of collagen-stimulated platelet activation involved ERK1/2 and JNK1/2 activation, but not p38 MAPK activation, which may explain why CBN presents higher effectiveness for collagen stimulation than that for AA, U46619, and thrombin. Moreover, Fan et al. [24] reported that ERK1/2 and JNK1/2, but not p38 MAPK, are the major mitogen-activated protein 3 kinase (MEKK3) downstream signaling molecules in platelet activation. Therefore, we speculated that CBN may act on the MEKK3, resulting in inhibition of ERK1/ 2 and JNK1/2 phosphorylation. However, further studies are required for clarification. Elevation of cyclic nucleotides, such as cyclic AMP and cyclic GMP, in platelets activates their respective protein kinase A and protein kinase G. This modulates platelet activation by phosphorylating intracellular protein substrates, such as VASP, which are involved in the inhibition of platelet aggregation and platelet adhesion [28]. Increased levels of cyclic nucleotides prevent most of the platelet responses and decrease the intracellular [Ca 2+ ]i through Ca 2+ uptake into the dense tubular system, which suppresses the activation of PLC/PKC signaling. In this study, neither SQ22536 nor ODQ significantly reversed the CBN-mediated inhibitory response, and CBN had no effects on VASP phosphorylation. Therefore, the CBNmediated inhibition of platelet activation is independent of the intracellular cyclic nucleotides/VASP pathway.
The fibrinogen-integrin α IIb β 3 binding belongs to a major component of activated platelets. Integrin α IIb β 3 undergoes conformational changes upon activation, generating a unique and specific ligand-binding site for the fibrinogen, von Willebrand factor, and fibronectin [1]. Platelet adheres to immobilized fibrinogen and mediates clot retraction; these processes are involved in integrin α IIb β 3 outside-in signaling [1]. PAC-1 reacts with the activation-induced conformational epitope of integrin α IIb β 3 [29], and the PAC-1 binding was observed to be markedly reduced by CBN treatment. In addition, platelet-mediated fibrin clot retraction is also mediated by integrin α IIb β 3 . Integrin α IIb β 3 -mediated signaling begins immediately after a fibrinogen molecule binds to the integrin; this outside-in signaling results in tyrosine phosphorylation of numerous proteins, such as the Src family kinases (SFK; e.g., Src, Lyn, and Fyn), FAK, and the cytoplasmic tail of integrin β 3 at Tyr759, a process dependent on outside-in signaling and cytoskeleton reorganization [1]. The critical role of integrin β 3 at Tyr759 in platelets was demonstrated in vivo, and its mutation led to bleeding disorder and strongly affected clot retraction responses in vitro [30]. FAK, a cytoplasmic tyrosine kinase located at focal adhesion points, plays a vital role in cytoskeleton regulation and integrin α IIb β 3 activity [31]. Platelet adhesion to immobilized fibrinogen requires FAK activation through integrin α IIb β 3 , and in turn activation of FAK requires autophosphorylation [31]. In this study, CBN had no effects on platelet adhesion and spreading and fibrin clot retraction, as well as phosphorylation of integrin β 3 , Src, and FAK on immobilized fibrinogen, indicating that CBN influences integrin α IIb β 3 inside-out but not outside-in signaling.
After vascular endothelial cell injury, exposure to subendothelial collagen is the major trigger that initiates platelet adhesion and aggregation at the injury site, followed by vascular thrombosis. Animal models of vascular thrombosis are necessary in order to understand the effectiveness of test compounds for this disease. An ideal mouse model should be technically simple, quick in operation, and easily reproducible. In a vascular thrombotic mice model [11], mesenteric venules were continuously irradiated by fluorescein sodium throughout the experimental period, leading to strong damage to the endothelium, treatment with 10 mg/kg CBN significantly extended the occlusion times; in studies on acute pulmonary thromboembolism, platelet aggregation is intimately involved in experimental thrombosis, and CBN effectively prevented ADP-induced thromboembolic death. We also found that CBN is more effectiveness than aspirin at lowering mortality in acute pulmonary thromboembolism. These data are consistent with the fact that platelet aggregation is a more crucial factor causing vascular thrombosis. Therefore, CBN may represent a potential natural compound for treating thromboembolic disorders.

Conclusion
This study revealed a novel role of CBN in the inhibition of platelet activation in humans, suggesting that it can be used for potential therapeutic or prophylactic applications. The outcome of this study may provide a new insight into the role of CBN in human platelet activation because it significantly inhibited platelet activation by hindering the PLCγ2-PKC cascade and subsequently, suppressed the activation of Akt and ERKs/JNKs. These changes decrease the release, such as [Ca 2+ ]i, followed by integrin α IIb β 3 inside-out signaling and inhibition of platelet aggregation.