Reduction of NF- κ B Signals in Platelets and Prolongation of Platelet Plug Formation against High Shear Flow in Whole Blood on Human Subject by Columbianadin

: Myocardial infarction and cerebral ischemic stroke during the process of arterial thrombosis are prominently causes of death worldwide. Platelets are anucleated cells and play a critical factor in these diseases. Columbianadin (CBN), a coumarin derivative from plants, inhibits e ﬀ ective platelet activation. In this study, platelet function analysis revealed that the closure time of the platelet plug in human whole blood signiﬁcantly prolonged by CBN, whereas CBN did not pointedly prolong the bleeding time in mice. BAY11-7082 (an inhibitor of I κ B kinase) and MG-132 (an inhibitor of proteasome) inhibited collagen-stimulated platelet aggregation and ATP-release in human platelets, BAY11-7082 exhibited a higher potency than MG-132. Moreover, CBN markedly reduced NF- κ B activation (e.g., I κ B α and p65 phosphorylation) and reversed I κ B α degradation in activated platelets. We investigated intercellular signaling events between mitogen-activated protein kinases and NF- κ B, and found that BAY11-7082 abolished JNK1 / 2 and ERK1 / 2 phosphorylation. Interestingly, SP600125 (an inhibitor of JNK) but not PD98059 (an inhibitor of ERK) had no e ﬀ ect in NF- κ B activation in activated platelets. Moreover, CBN but not BAY11-7082 signiﬁcantly reduced hydroxyl radical (HO • ) formation in platelets. Therefore, we propose that CBN inhibits NF- κ B activation in human platelets and could present a potent clinical treatment for thromboembolic diseases.


Platelet Function Analysis in Human Whole Blood
The Dade Behring PFA-100 System (Marburg, Germany) was used to analyze platelet function [13]. Cartridges containing collagen/ADP (CADP)-coated membranes were preincubated with CBN (45 and 90 µM) or the solvent control (0.1% DMSO) for 2 min. Whole blood aliquots (0.8 mL/cartridge) were applied to the cartridges before contents were exposed to high shear flow conditions (5000-6000/s). The closure time (CT) was defined as time required for a platelet plug to occlude the aperture in the collagen membrane [13].

Tail Bleeding Time
BioLasco, Taipei, Taiwan supplied male ICR mice (20-25 g, 5-6 weeks). Affidavit of Animal Use Protocol, Taipei Medical University (LAC-2018-0360) approved all procedures and protocols. The bleeding time was measured after 30 min of intraperitoneal administration of CBN (10 and 20 mg/kg), 0.1% DMSO or aspirin (50 mg/kg). The mice tail was cut in 3 mm, immersed in normal saline directly and the time of bleeding was recorded until no sign of further bleeding for at least 10 s.

Platelet Aggregation
The directives of the Helsinki Declaration were conformed this study and further approved by the Institutional Review Board of Taipei Medical University (TMU-N201812024). According to our earlier study, the platelet suspensions were prepared from 36 healthy human bloods [14] and mixed with an acid-citrate-dextrose solution (9:1, v/v). The platelet-rich plasma separated by centrifugation was supplemented with PGE 1 (0.5 µM) and heparin (6.4 IU/mL). A 3.5 mg/mL of BSA and 1 mM of Ca 2+ added Tyrode's solution was used to make the final suspensions of washed human platelets.
Platelet aggregation was measured by using a Lumi-Aggregometer (Payton, Scarborough, ON, Canada) through turbidimetric method [14]. Different concentrations of CBN, 0.1% DMSO, or other substances were preincubated with platelet suspensions (3.6 × 10 8 cells/mL) for 3 min before the addition of collagen (1 µg/mL). The incubation was continued further 6 min, and the degree of aggregation was measured as a percentage of the control (Tyrode's solution-treated group) in light-transmission units. A 20 µL of a luciferin/luciferase mixture was added 1 min before the addition of collagen, and the final suspension was used to measure ATP release using Hitachi Spectrometer F-7000 (Tokyo, Japan).

Data Analysis
The results are represented as the mean ± standard error of the mean (SEM) and are convoyed by the number of observations (n). Differences between mice groups were analyzed using unpaired Student's t tests. Differences among the experimental setups were designed using a one-way analysis of variance (ANOVA). If ANOVA revealed significant differences in the group means, then each group was compared using the Student-Newman-Keuls method. p < 0.05 was reflected statistically significant.

Effect of CBN in Platelet Plug Formation in Human Whole Blood and Tail Bleeding in Experimental Mice
Hou et al. [12] reported that 10 mg/kg CBN treatment markedly prolonged occlusion time in the mesenteric microvessels of mice. In the current study, we further confirmed the effective antithrombotic activity of CBN by using the platelet function analyzer (PFA-100) ( Figure 1A), which consists of an instrument and test cartridges in which the process of platelet adhesion and aggregation following a vascular injury is stimulated in vitro. The single-use cartridges consists of a number of integrated parts including a capillary, a sample reservoir and a biochemically active membrane with a central circular aperture ( Figure 1A). Citrated whole blood is aspirated from the sample reservoir through the capillary and aperture, which expose platelets to high shear flow conditions (5000-6000/s). The membrane is coated with collagen, a subendothelial protein generally believed to be the initial matrix for platelet attachment. In addition, the membrane is coated with ADP, which is other physiological agonist, along with collagen, are extensively used to trigger platelets in aggregometry testing. The time needed to occlude the aperture is described as closure time (CT). In the present study, the CT of the CADP membrane in whole blood treated with solvent control (0.1% DMSO) was 81.5 ± 5.1 s (n = 8); CBN (90 µM) significantly prolonged CT (45 µM, 99.8 ± 5.6 s, n = 8; p > 0.05; 90 µM, 114.0 ± 9.8 s, n = 8; p < 0.05) compared with the solvent control ( Figure 1A).
We evaluated the effects of CBN on bleeding time, since it is a general side effect of antiplatelet drugs when they are using in clinical trials. This effect was also compared with antiplatelet drug aspirin. The bleeding time was 165.4 ± 15.9 s (n = 8) in the solvent control group ( Figure 1B). After 30 min of intraperitoneal treatment of CBN at 10 and 20 mg/kg, the bleeding time was 180.4 ± 20.1 and 191.0 ± 26.1 s (n = 8), respectively; therefore, the bleeding time was not significantly affected by those doses. In addition, aspirin was administered at 50 mg/kg, and the bleeding time was markedly prolonged after 30 min from 165.4 ± 15.9 s (0.1% DMSO-treated group) to 225.7 ± 6.3 s (n = 8, p < 0.05) ( Figure 1B). Appl. Sci. 2020, 10, x FOR PEER REVIEW 5 of 15 We evaluated the effects of CBN on bleeding time, since it is a general side effect of antiplatelet drugs when they are using in clinical trials. This effect was also compared with antiplatelet drug aspirin. The bleeding time was 165.4 ± 15.9 s (n = 8) in the solvent control group ( Figure 1B). After 30 min of intraperitoneal treatment of CBN at 10 and 20 mg/kg, the bleeding time was 180.4 ± 20.1 and

Inhibition of Platelet Aggregation and ATP-Release by BAY11-7082, MG-132, and CBN in Collagen-Stimulated Platelets
BAY11-7082 is an IKK inhibitor that possesses several biological activities, including anticancer, neuroprotective, and anti-inflammatory effects [16]. MG-132 belongs to a proteasome inhibitor that blocks the activation of NF-κB by inhibiting IκB degradation [17]. In the present study, we compared the relative activities of BAY11-7082 and MG-132 in the inhibition of platelet aggregation and determined that in washed human platelets, both BAY11-7082 (4 and 8 µM) and MG-132 (60 and 120 µM) inhibited collagen (1 µg/mL)-stimulated platelet aggregation in a concentration-dependent manner, similar to CBN (45 and 90 µM) (Figure 2A). BAY11-7082 exhibited a higher potency than MG-132 in on molar basis ( Figure 2A). Besides, the release of granular contents (e.g., ATP) is associated with platelet activation, which causes platelet aggregation. As illustrated in Figure 2B, BAY11-7082 also showed higher activity than MG-132 at inhibiting ATP-release stimulated by collagen. Therefore, BAY11-7082 was employed to clarify the role of NF-κB in CBN-mediated antiplatelet activity.

Reduction of NF-κB Signals by CBN in Platelets
NF-κB exists as an inactive cytoplasmic complex, composed of heterodimer p50 and p65 subunits that are tightly bound to IκB inhibitory proteins [18]. Figure 3 shows that IκBα and p65 phosphorylation and IκBα protein degradation were significantly increased after stimulation with collagen (1 µg/mL) in human platelets. CBN (45 and 90 µM) reduced IκBα and p65 phosphorylation ( Figure 3A,B) and reversed IκBα degradation ( Figure 3C) caused by collagen stimulation. Assembled data of Figure 3 are displayed in right panels. These results suggest that the inhibition of NF-κB signals may play a critical role in CBN-mediated antiplatelet activity.

Reduction of NF-κB Signals by CBN in Platelets
NF-κB exists as an inactive cytoplasmic complex, composed of heterodimer p50 and p65 subunits that are tightly bound to IκB inhibitory proteins [18]. Figure 3 shows that IκBα and p65 phosphorylation and IκBα protein degradation were significantly increased after stimulation with collagen (1 μg/mL) in human platelets. CBN (45 and 90 μM) reduced IκBα and p65 phosphorylation ( Figure 3A,B) and reversed IκBα degradation ( Figure 3C) caused by collagen stimulation. Assembled data of Figure 3 are displayed in right panels. These results suggest that the inhibition of NF-κB signals may play a critical role in CBN-mediated antiplatelet activity.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 10 of 15 phosphorylation were detected in the subcellular extracts of the collected platelets. The matching statistical data of (B,C) are displayed on the right panel of each figure. All phosphorylated proteins were normalized to the total proteins. * p < 0.05 and *** p <0.001, compared with resting platelets; # p < 0.05, ## p < 0.01, and ### p < 0.001, compared with the 0.1% DMSO-treated group (n = 4).

Effect of CBN and BAY11-7082 on HO • Formation
Reactive oxygen species (ROS) generated from platelet activation may further increase platelet reactivity during thrombus formation. However, the controlling pathways of ROS, particularly HO • , in platelet activation remain unclear. As illustrated in Figure 6, a classic ESR signal of HO • was triggered by collagen (1 μg/mL), which was less in resting platelets ( Figure 6A,B); 90 μM but not 45 μM of CBN obviously reduced HO • formation ( Figure 6C,D), whereas BAY11-7082 (8 μM) had no noteworthy effect under the same condition ( Figure 6E).

Effect of CBN and BAY11-7082 on HO • Formation
Reactive oxygen species (ROS) generated from platelet activation may further increase platelet reactivity during thrombus formation. However, the controlling pathways of ROS, particularly HO • , in platelet activation remain unclear. As illustrated in Figure 6, a classic ESR signal of HO • was triggered by collagen (1 µg/mL), which was less in resting platelets ( Figure 6A,B); 90 µM but not 45 µM of CBN obviously reduced HO • formation ( Figure 6C,D), whereas BAY11-7082 (8 µM) had no noteworthy effect under the same condition ( Figure 6E).

Discussion
Columbianadin is a natural coumarin derivative from Angelica decursiva (Umbelliferae) have various biological activities, including anti-inflammatory and anticancer [10]. Collagens from the subendothelial matrix activates platelet adhesion and aggregation at the site of vascular endothelial cell injury, which is then, induced arterial thrombus formation. Previous study [12] demonstrated that CBN significantly reduced the mortality of mice those were subjected to ADP-induced acute pulmonary thromboembolism. Platelet adhere to collagen depended on flow conditions, and inactive platelets were unable to adhere to the CADP membrane under flow conditions. In the present study,

Discussion
Columbianadin is a natural coumarin derivative from Angelica decursiva (Umbelliferae) have various biological activities, including anti-inflammatory and anticancer [10]. Collagens from the subendothelial matrix activates platelet adhesion and aggregation at the site of vascular endothelial cell injury, which is then, induced arterial thrombus formation. Previous study [12] demonstrated that CBN significantly reduced the mortality of mice those were subjected to ADP-induced acute pulmonary thromboembolism. Platelet adhere to collagen depended on flow conditions, and inactive platelets were unable to adhere to the CADP membrane under flow conditions. In the present study, we confirmed that CBN markedly prolonged closure time (CT) in whole blood by using the PFA-100 instrument. Therefore, CBN may act as potent agent in the inhibition of human platelet activation and suggesting that it can have therapeutic or prophylactic applications. Moreover, the effect of CBN on bleeding time was examined by using the tail transection model in mice. Aspirin is the highly recommended antiplatelet drug for the prevention or treatment of cerebro-cardiovascular diseases. CBN-treated mice were not significantly prolongation of bleeding time, whereas aspirin-treated mice caused the prolongation under the same condition. Therefore, CBN may signify an active coumarin derivative for treating thromboembolic disorders without causing the side effect of bleeding.
In this study, CBN inhibited collagen-stimulated NF-κB activation in human platelets, indicating that NF-κB signals play a crucial role in CBN-mediated antiplatelet activity. The function of NF-κB in nucleated cells has been studied extensively. Diverse stimuli, including bacterial infection, cytokines, and free radicals can induce NF-κB activation. NF-κB regulates genes involved in cell survival, cell proliferation and inflammation responses among others [18]. Therefore, NF-κB is an ideal targeting transcription factor for therapeutic interventions against inflammatory diseases and cancer. Several studies have demonstrated that platelets express several transcription factors [20,21], suggesting that these transcription factors have a nongenomic function in platelets. However, whether NF-κB is practically present in a novel manner in platelets, unrelated to transcriptional regulation, remains unknown. NF-κB was reported to be involved in thromboxane A 2 (TxA 2 ) formation, P-selectin expression, fibrinogen adhesion in platelets [22]. In the current study, BAY11-7082 inhibited platelet aggregation, ATP-release reaction and phosphorylation of IκBα, which definitively demonstrated that NF-κB is involved in platelet activation through various signal pathways. So far, we are not clear whether CBN directly regulates NF-κB, or through inhibition of its upstream regulator, hence it needs to be investigated in the future. Liu et al. [23] were the first to demonstrate the expression of NF-κB in platelets in 2002, revealing that thrombin-induced platelet activation triggers the degradation of IκBα following its serine 32 residue phosphorylation. [23].
As described previously [11], CBN inhibits platelet activation through interfering in integrin α IIb β 3 inside-out signals. Salanova et al. [24] reported that platelet integrin α IIb β 3 colocalized with integrin β 2 and cooperated in NF-κB activation, our results accord with their findings. MAPKs are serine/threonine protein kinases consist of of three major subgroups, p38 MAPK, ERK1/2, and JNK1/2, which occurred in platelets and activated by various agonists [25]. Our findings demonstrated that BAY11-7082 attenuates collagen induced JNK1/2 and ERK1/2 phosphorylation in human platelets. Similarly, Kauskot et al. [26] found the involvement of JNK1 in ADP-dependent collagen-induced platelet aggregation. They also demonstrated that glycoprotein (GP)Ib-von Willebrand factor (vWF) interaction triggers integrin α IIb β 3 activation in a JNK1-dependent manner during rolling and adhesion of platelets to vWF. Several studies with different experimental setups have indicated that ERK2 is also involved in platelet aggregation. Roger et al. [27] and Falker et al. [28] have independently reported the involvement of ERK2 in platelet aggregation. They demonstrated that collagen-or thrombin-induced ERK2 activation depends on TxA 2 formation and ADP/ATP release reaction. Furthermore, Toth-Zsamboki et al. [29] have indicated that ERK2 affects platelet secretion.
A partial amount of hydrogen peroxide formed by platelets is converted into HO • , which are involved in the initial phase of platelet activation [15]. The ESR study revealed that BAY11-7082 did not scavenge HO • in activated platelets, indicating that platelet HO • formation does not seem to be regulated by NF-κB signal, and which may be played an upstream regulator for NF-κB signal. However, we do not rule out the possibility that other unknown signals are involved in the NF-κB-mediated inhibition of platelet activation by CBN.

Conclusions
This study presents a distinctive inhibitory pathway of NF-κB-mediated MAPK activation by CBN in human platelets and corroborates findings that CBN has potent activity and thus should be investigated as a prophylactic or clinical therapy for cardiovascular diseases (Figure 7). Platelet activation is involved in inflammation; thus, blocking platelet function by inhibiting NF-κB signal could also be considered for treating various inflammatory diseases.
Appl. Sci. 2020, 10, x FOR PEER REVIEW 13 of 15 activation is involved in inflammation; thus, blocking platelet function by inhibiting NF-κB signal could also be considered for treating various inflammatory diseases.