Both Methanol and Water Mixture Extracts of Panax Notoginseng Flower Affect Platelet Function and Thrombus Growth by Down-Regulating PI3K/AKT and MAPKs Signaling Pathways

Xiao Zuo School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Nan Qin Department of Clinical laboratory, The Second A liated Hospital of Guangzhou Medical University, Guangzhou, Guangdong Province 510260 Yu-Heng Mao School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Lijuan Ma State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Taipa, Macao, 999078 Qing Li Department of Food and Nutrition Sciences, Life Science, MMW505, The Chinese University of Hong Kong, Shatin, N.T. Zezhong Tian School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Mingzhu Zhao School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Qiuhua Ji School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Yiting Chen School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080 Jian-Bo Wan State Key Laboratory of Quality Research in Chinese Medicine, Institute of Chinese Medical Sciences, University of Macau, Taipa, Macao, 999078 Yan Yang (  yangyan3@mail.sysu.edu.cn ) School of Public Health (Shenzhen), Sun Yat-sen University, Guangzhou, Guangdong Province 510080


Introduction
Due to the critical role of platelet in thrombus formation, antiplatelet therapy has become a useful strategy to prevent acute thromboembolic artery occlusions in cardiovascular diseases (Schror, 1995;Sharma & Berger, 2011). However, most of the antiplatelet agents may lead to severe adverse effects, especially under a high dosage or a long course of treatment. The increasing bleeding risk limited the clinical applications of many antiplatelet medicines (Michelson, 2010). Therefore, the development of novel natural antiplatelet agents without bleeding risk have been attracted the increasing attention.
Panax notoginseng, a well-known and valuable Chinese medicine, has been widely used for the treatment of CVD more than 400 years(L. Wang et al., 2013). Among the different parts of the plant,P. notoginseng owers (PNF) contain the highest abundance of saponins. A numerous studies indicated that PNF exert the bene cial effects on hypertension, insomnia, and stomatitis. Our previous study reported that two monomers (G-Rb2 and G-Rd2) from PNF signi cantly inhibited human platelet aggregation and activation induced by adenosine diphosphate (ADP)in vitro. However, the effects of PNF mixture extracts on platelet function were unclear yet.

Platelet ATP release assay
The secretion of ATP was determined in a Chrono-log lumiaggregometer according to the manufacturer's instructions. Brie y, PRP at 2.5×10 8 platelets/mL were incubated with different concentrations of PNFM, PNFW and the control buffer at 37℃ for 20 min as described above. Luciferin-luciferase reagent was added directly to platelet suspensions, which were continually stirred at 1,000 rpm at 37 °C. 5 μM or 10 μM ADP was added to activate platelets and real-time ATP secretion was monitored.

Assay of soluble β-thromboglobulin (β-TG)
To detect platelet β-TG in vitro, PRP (2.5×10 8 platelets/mL) was pre-incubated with different concentrations of PNFM, PNFW or control buffer for 20 min at 37℃ as described, and then stimulated with 5 μM ADP, followed by centrifugation at 10000 × g for 5 min at 4°C. The cell-free supernatant was collected by a new clean tube and stored at -80°C until use. β-TG levels in supernatant were determined using a β-TG ELISA kit (BlueGene Biotech, Shanghai, China) according to the manufacturer's instruction.

Assay of platelet Ca 2+ mobilization
The intracellular calcium ion concentration was measured using Fluo-3AM as a calcium ion uorescence probe. Briefly, PRP (2.5 × 10 8 platelets/mL) was incubated with 10 μM Fluo-3AM for 30 min at 37°C. The Fluo-3-loaded platelets were pre-incubated with different concentrations of PNFM or PNFW for 20 min at 37°C in the presence of 1mM CaCl 2 , and then stimulated with 200 μM ADP or 0.5 U thrombin. Fura-3 uorescence in the cytosol was measured by a spectro uorometer as the following formula: where 224 is the dissociation constant of the Fura-3-Ca 2+ complex, and F min and F max are the uorescence intensities at very low and very high Ca 2+ concentrations, respectively.

Platelet spreading on immobilized Fibrinogen
Chamber slides with microtiter wells were coated with 100 μg/mL brinogen overnight at 4℃. PRP were incubated with different concentrations of PNFM, PNFW and the control buffer for 20 min at 37℃ as described above. The platelets were allowed to adhere and spread on brinogen-coated wells at 37℃ for 1 h. After washing, the cells were xed, permeabilized, and stained with Alexa Fluor 488-conjugated phalloidin before observing with an inverted uorescence microscope. The spreading areas of single platelet were measured using ImageJ software. Ten randomly selected elds from at least three parallel tests were used for statistical analysis.

Western blot analysis
After incubation with PNFM, PNFW and the control buffer for 20 min,platelets were activated with ADP (5 μM) or Thrombin (0.5 U) in the presence of 1 mM Ca 2+ for 5 min. Platelet were harvested and lysed with RIPA buffer supplemented with protease and phosphatase inhibitors for 30 min on ice. After centrifugation at 12000 g for 15 min. the supernatants were collected as platelet total protein for western blot analysis. The protein concentrations were determined using a commercial BCA kit (Thermo Fisher Scienti c, Rockford, IL). Equal amounts of protein (20 -50 μg) were fractionized on 10% SDS-PAGE gels and transferred onto a polyvinylidene di uoride (PVDF) membrane. The membranes were blocked with 5% nonfat milk in TBST and then incubated with each primary antibodies including β-actin, GAPDH, phospho-PI3K, phospho-Akt, phospho-Erk1/2, phospho-JNK, phospho-p38 and corresponding secondary antibodies. The target proteins were detected with an enhanced chemiluminescence (ECL) reagent (Thermo Scienti c, Waltham, MA, USA) in an automatic chemiluminescence image analysis system. 2017-0080).Brie y, after injection of PNFM, PNFW (30μg/g BW), or control buffer via the tail vein followed by inject calcein-labeled (4 μg/mL) mice platelets.The dose of PNFM and PNFW in animal experiments was calculated according to the dose of in vitroexperiments, approximately 500 μg/mL as the nal concentration in the murine blood. The injury was induced by topical application of FeCl 3 . Images of thrombus formation and dissolution were visualized by a uorescence microscope (Leica Microsystems, Wetzlar, Hesse, Germany). Based on the time to complete the vessel occlusion, the images from each group are compared with thosefrom the other groups.

Assay of bleeding time in mice
Male C57BL/6J mice (6-8 weeks old) were injected with PNFM, PNFW (30μg/g BW) or the control buffer via the tail vein 20 min before the bleeding time assay. Mice were then anesthetized with sodium pentobarbital and maintained at 37℃ on a heating pad during the experiment. The 5 mm tip of tail was cut off and placed into 37℃ saline solutions immediately. The bleeding time was calculated from the moment blood began emerging to the moment bleeding ceased.

Statistical analysis
GraphPad Prism version 5.01 software (GraphPad Inc., San Diego, CA, USA) was used for statistical analyses. Data are expressed as the means ± standard deviation (SD)of at least three independent experiments. The statistical signi cance among different groups was determined using one-way ANOVA. Differences were considered signi cant at P< 0.05.

PNFM and PNFW inhibit human platelet aggregation in vitro
Both PNFM and PNFW signi cantly inhibited ADP-and thrombin-induced human platelet aggregation in a dose-dependent manner ( Fig. 2A-C). Additionally, PNFW at 300 μg/mL exhibited a more potent effect on the suppression platelet aggregation induced by ADP at both 5 μM and 10 μM, compared with PNFM at the same dose ( Fig.2A-B). The dosages of PNFM and PNFW used in this study had no cytotoxicity on human platelets as demonstrated by the LDH leakage assay (Supplementary Figure 1).
3.3 PNFM and PNFW attenuate human platelet surface CD62P expression, PAC-1 binding, and platelet binding to brinogen in vitro CD62P expressed on platelet surface is a marker of platelet activation. Upon agonist stimulation, the transduction of intraplatelet signals leads PAC-1 to switch from a low-a nity to high-a nity state for brinogen, which initiates and ampli es platelet aggregation and thrombus consolidation (Ya et al., 2018). As shown in Fig. 3, both PNFM and PNFW dose-dependently inhibited ADP-and thrombin-induced CD62P expression, PAC-1 activation, and brinogen binding. We found the inhibitory effects of PNFW on human platelet CD62P expression were stronger comparing with PNFM at low concentration (100 μg/mL). On the contrary, the PNFM showed more potent inhibition on PAC-1 activation at all concentrations compared with PNFW. Additionally, the inhibitory effects of PNFM on thrombin-induced PAC-1 activation and brinogen binding were signi cantly stronger compared with PNFW at relatively higher concentrations (300 or 500 μg/mL).

PNFM and PNFW inhibit human platelet ATP release, β-TG expression, and Ca 2+ mobilization in vitro
Granule secretion is an important marker of platelet activation prior to aggregation and thrombus formation. As shown in Fig. 4A-C, both PNFM and PNFW decreased intraplatelet ATP release induced by ADP (5 μM and 10 μM) and thrombin (0.1 U) in a dose-dependent manner. Both PNFM and PNFW also dose-dependently reduced the intraplatelet Ca 2+ mobilization induce by ADP and thrombin (Fig. 4D-E). Additionally, the release of intraplatelet β-TG were attenuated by PNFM and PNFW at the dose of 500 μg/mL (Fig. 4 F).

PNFM and PNFW decrease human platelet spreading on immobilized brinogen in vitro
After brinogen binding to platelet PAC-1, it transduces outside-in signals into the cell and triggers platelet spreading. We next investigated whether PNFM and PNFW in uence platelet spreading on immobilized brinogen. As shown in Fig. 4G, PNFM and PNFW substantially decreased the surface area of the spreading platelets on immobilized brinogen. Additionally, the inhibitory effects of PNFM on platelet spreading were more potent than those of PNFW at relatively lower dose (100 μg/mL) rather than other two higher doses.

Effects of PNFM and PNFW on phosphorylation of PI3k/Akt and MAPKs in platelets
PI3k/Akt and MAPKs (ERK, JNK, and p38 MAPK) pathways play a crucial role in platelet activation. Our results demonstrated that both PNFM and PNFW markedly down-regulated the phosphorylation of PI3k, Akt, JNK, ERK, and p38 MAPK in ADP-and thrombin-treated human platelets in a dose-dependent manner (Fig. 5). These results suggested that the ameliorative effects of PNFM and PNFW on platelet hyperactivity were possibly mediated by concomitant inhibition of PI3k/Akt and MAPK signaling pathways.

PNFM and PNFW inhibit FeCl 3 -induced mesenteric thrombus formation in mice
The results of in vivo thrombus model experiment showed that 500 μg/mL of PNFM and PNFW effectively suppressed FeCl 3 -induced mesenteric thrombus formation as demonstrated by the prolonged vessel occlusion time (Fig. 6A-6B).We also found that PNFW exerted more potent inhibitory effects on thrombus formation than PNFM treatment. Additionally we found that compared with the control, the treatment of either PNFM or PNFW at 500 μg/mL had no signi cant in uence on the tail bleeding time, which is an indicator of normal hemostasis and coagulation functions (Fig.6 C).

Discussion
Resting platelets circulate in the form of small discs without interacting with each other or the vascular endothelium (Blache, 1992). Upon vascular injury, platelets adhere to the subendothelial matrix becoming activated and aggregated (Z. Li et al., 2010). Platelets could be activated by various extracellular stimuli, including ADP and thrombin.ADP activates platelets mainly via the P2Y 1 and P2Y 12 receptors through positive feedback loops that amplify platelet activation (Hollopeter et al., 2001). Thrombin is regarded as the most powerful physiological platelet agonist. Thrombin could activate human platelets via a dual system of G-protein coupled protease-activated receptors (PAR), namely PAR1 and PAR4, which promote platelet shape change, aggregation, mobilization of P-selectin to platelet surface, integrin αIIbβ3 activation and granule secretion (Angiolillo, Capodanno, & Goto, 2010). The present study showed that PNFM and PNFW effectively inhibited human platelet aggregation induced by both ADP and thrombin.
Since ADP and thrombin have different receptors, we speculated that PNFM and PNFW have more comprehensive effects onplatelets via various pathways, instead of working as a single speci c antagonist to inhibit the binding to platelet membrane receptors.
Although the antithrombotic bene ts of P.notoginseng roots have been well documented before (Fu et al., 2021), few studies have evaluated the medicinal use of PNF for the prevention and treatment of cardiovascular diseases. Our previous studies showed that two saponin monomers (Rb2 and Rd2) in PNF signi cantly attenuated platelet function (Zuo et al., 2021). In the present study, we demonstrated that both the methanol (PNFM) or water aqueous (PNFW) mixture extracts from PNF can also effectively inhibited human ADP-and thrombin-induced platelet activation, granule secretion, aggregation and platelet spreading on brinogen in vitro. Compared with the saponin monomers that are suitable for pharmaceuticalapplication, PNF mixture extracts are more available and convenient to be utilized as a nutritional ingredient in functional food.
The pharmacological activities may differ among different ginsenosides. Our previous study has identi ed different ginsenoside pro les between PNFM and PNFW. The chromatographic analysis revealed that the compositional ratio of original types saponins in PNFM were obviously higher than PNFW, while the transformed types saponins were higher in PNFW than PNFM. Speci cally, during the water extraction process, the β -(1→2) -glucosidic linkage at the C-3 site of ginsenosidesis selectively cleaved while otherglycosidic linkage keeps intact. The major saponins in PNF methanol mixture extract (Fig. 1A), ginsenoside Rb1 (G-Rb1), G-Rc, G-Rb2, G-Rb3, and G-Rd were partially converted into other saponins, including gypenoside XVII (GY-XVII), notoginsenoside Fe (NG-Fe), G-Rd2 and NG-Fd and G-F2 (Fig. 1B) (Ma et al., 2017). In this study, although both PNFM and PNFW exerted signi cant effects on human platelet functions with similar dose-dependent trends, minor differences were found among different doses. The inhibitory effects of PNFW on human platelet aggregation, CD62p expression, platelet spreading and ATP release and thrombus formation were more potent than PNFM, while the effects of PNFM on αIIbβ3 activation, brinogen binding, and Ca 2+ mobilization were stronger than PNFW. The difference in the antithrombotic effects of PNFM and PNFW could be partially attributed to the different saponin pro les.Another studyalso reported that the water and methanol extracts of a traditional medicine (Ocimumamericanum) displayed different healthy effect (Zengin et al., 2019). But the speci c mechanism remained to be further studied.
Notoginsenoside Fc (NG-Fc), a main composition of both PNFM and PNFW, has been shown to inhibit platelet aggregation induced by ADP, thrombin, and collagen at submaximal concentrations (Liu et al., 2018).Except NG-Fc, among the 10 saponins isolated from PNF extracts in our previous studies have indicated that, only G-Rb2, and G-Rd2 could inhibit ADP-induced platelet aggregation at the concentration of 100 μg/mL (Zuo et al., 2021).Additionally, in the present study, our results showed that PNFM and PNFW may have stronger antiplatelet effect than saponin monomers, as they were still effective with higher doses of agonists (10 μM ADP and 0.5 U thrombin). Notably, according to the quantitative determination of the saponins in the PNFM and PNFW (Figure 1), the effective dosage of the NG-Fc, G-Rb2, and G-Rd2 in our previous study (Zuo et al., 2021) are much higher than those saponin monomers in PNFM and PNFW. This may due to the synergistic effects of various saponin monomers in two PNF extracts, which is also consistent to a reported study that ginsenosideRp3 synergistically inhibits platelet aggregation with other ginsenosides (Irfan et al., 2018). Additionally, some other saponins with lower concentrations in PNFM and PNFW should be taken into consideration in future studies.It indicated that PNF mixture extracts at a lower dosage could confer the similar therapy compared to the saponin monomers.
Bioavailabilityand pharmacokinetics are two major issues in uencing the health effects of natural compounds used as dietary supplements.A previous study indicated that NG-Fc have a slow elimination from plasma with a long t 1/2 (approximately 22-30 h) and its oral bioavailability was 0.10-0.14% (He et al., 2015). Moreover, it was reported that the oral bioavailability of Rb1 was 4.35% after administering PNS to rats and the maximal concentration of plasma G-Rb1 could be reached about 50 μM after 5 min of intravenous dosing 5 μmol/kg body weight of G-Rb1 in rats (Q. F. Xu, Fang, & Chen, 2003). Additionally, it is estimated that the serum concentration of PNS after administrating Xueshuantong® Injection (a clinical medicine in which PNS is the main component) is compatible to the concentration of 100μg/mL used in thisstudy(R. L. Li et al., 2020). Although the systemic bioavailability of PNFM and PNFW are not clearly, health effects were observed in animal studies. Orally administration of 120 mg/kg/day PNF extractions for 9 weeks signi cantly improved the ventricular hypertrophy state in mice(Y. Wang et al., 2012). In a myocardial infarction rats model, oral 1000 mg/kgP. notoginseng ower decoction could relieve symptoms of MI (Zhou, Li, Chen, & Xie, 2019). Although whether orally administration of PNFM and PNFW can modify platelet function and attenuate thrombus formation is unclear yet, we demonstrated platelet inhibitory effects of PNFM and PNFW in vitroin the present study, and the bioavailability and pharmacokinetics of PNFM and PNFW and their potential health bene ts in vivo will be clari ed in the further studies based on the present results.
PI3K/Akt and the MAPK pathways play an important role in human platelet activation and are reported in many studies (Ma et al., 2017). The inhibition of PI3K/Akt could prevent integrin αIIbβ3-mediated platelet adhesion and thrombus formation (Morello, Perino, & Hirsch, 2009). MAPKs, including JNK1, ERK2, and p38, have been identi ed to be involved in platelet activation induced by various agonists (Bugaud, Nadal-Wollbold, Levy-Toledano, Rosa, & Bryckaert, 1999). In the present study, we found that PNFM and PNFW signi cantly suppressed the activation of PI3K, Akt, and MAPKs induced by both ADP and thrombin in a dose-dependent manner. Therefore, we concluded that the inhibitory effects of PNFM and PNFW on platelet hyperactivity may partly result from concomitant inhibition of PI3K/Akt and MAPK signaling. Our results were consistent with previous study of PNF saponin monomers. G-Rb2 and G-Rd2 work on platelets by modulating P2Y12-mediated signaling in a way of up-regulating cAMP/PKA signaling and down-regulating PI3K/Akt/Erk1/2 signaling pathways (Zuo et al., 2021).

Conclusion
In conclusion, both PNF water and methanal mixture extracts inhibited ADP-and thrombin-induced platelet aggregation, activation, granule secretion and spreading without causing bleeding risk. Their inhibitory effects on platelet function might be associated with suppressing phosphorylation of PI3K/Akt and MAPKs. Compared with saponin monomers, the PNF mixture extracts are more accessible and easily produced. As natural plant extracts, they would be more acceptable by patients compared with traditional medicines.Therefore, our ndings suggest that the PNF mixture extractsare valuable ingredients for application in pharmacal industry in the future and could be potentiallyapplied in preventing thrombotic and cardiovascular diseases in the future. -Consent for publication Not applicable.
-Availability of supporting data Not applicable.

-Competing interests
The authors declare that they have no known competing nancial interests or personal relationships that could have appeared to in uence the work reported in this paper. -Acknowledgements We thank all the participants in this study. 31. Zengin, G., Ferrante, C., Gnapi, D. E., Sinan, K. I., Orlando, G., Recinella, L., . . . Menghini, L. (2019).
Comprehensive approaches on the chemical constituents and pharmacological properties of owers    Effects of PNFM and PNFW on human platelet activation. Human platelet-rich plasma (PRP) or gelltered platelets were pre-incubated with control buffer or indicated doses (100, 300, and 500 μg/mL) of PNFM or PNFW for 20 min at 37℃. Effects of PNFM and PNFW on CD62P expression, αIIbβ3 activation and brinogen binding were measured by ow cytometry analysis. Values are mean ± SD, n = 4. *p< 0.05, **p< 0.01 and ***p< 0.001, as compared to control buffer.

Figure 4
Effects of PNFM and PNFW on human platelet ATP release, β-TG expression, Ca2+ mobilization and platelet spreading on immobilized brinogen. Human PRP or gel-ltered platelets were pre-incubated with control buffer or the indicated doses (100, 300, and 500 μg/mL) of PNFM and PNFW for 40 min at 37℃.  Effects of PNFM and PNFW on platelet PI3k Akt and P38 MAPK phosphorylation. PRP or gel-ltered platelets were initially treated with indicated doses of PNFM, PNFW or vehicle, followed by activation with 5 μM ADP or 0.5 U thrombin. Platelet were collected and the cell lysates were analyzed for PI3k Akt and P38 MAPK, Erk and JNK activation. Data are presented as mean± SD. **p< 0.01 and ***p< 0.001, as compared to the ADP or thrombin control group.

Figure 6
Effects of PNFM, PNFW on thrombosis formation and bleeding time. (A-B) Thrombus formation was initiated by topical application of FeCl 3 on mesenteric arterioles in C57BL/6 male mice, which were injected with uorescent-labeled platelets and different extracts or control buffer. Thrombus formation was compared between groups based on the time to complete vessel occlusion. Values are mean ± SD, n = 10. ***p< 0.001, as compared to control. (C) Tail-vein bleeding times were examined in C57BL/6 mice.
Either control buffer or PNFM/PNFW were administered via the tail vein 20 min before the bleeding time was determined. Values are mean ± SD, n = 8.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Supplementarydata.docx