Oroxylin A reverses SHP-2 oxidative inactivation in GPVI signaling to suppress platelet activation and thrombus formation

Abstract

Background

Arterial thrombotic events are the leading causes of death worldwide, and the therapeutic effects of current antiplatelet drugs are not fully satisfactory. Oroxylin A (OroA), a flavone compound extracted from Scutellaria radix, possesses cardioprotective and many other pharmacological effects. While platelets play a crucial role in the development of myocardial infarction, the direct effects of OroA on platelet activation and thrombosis have yet to be investigated.

Methods

FeCl₃-induced arteriole thrombosis and whole-blood perfusion were used to assess the inhibitory effect of OroA on thrombus formation. A myocardial ischemia model was employed to evaluate the protective effect of OroA on myocardial injury. Multiple platelet function studies including platelet aggregation, platelet spreading, clot retraction were performed. Network pharmacology, flow cytometry, enzyme-linked immunosorbent assay, co-immunoprecipitation and western blot were utilized to explore the mechanism of OroA on platelet activation.

Results

OroA inhibited thrombus formation with less bleeding risk compared with aspirin. OroA protected against myocardial injury by suppressing microvascular thrombosis and platelet infiltration. OroA suppressed different agonist-induced platelet activation in a concentration-dependent manner, showing greater antiplatelet activity against collagen-induced platelet aggregation compared to ADP or thrombin-induced aggregation. OroA decreased granule release, integrin αIIbβ3 activation, platelet spreading and clot retraction. As a flavone, OroA boosted superoxide dismutase (SOD) and glutathione (GSH) activities and decreased malondialdehyde (MDA), oxidized glutathione (GSSG) and ROS levels in platelets during oxidative stress. OroA binds to SHP-2 and prevents its oxidative inactivation, leading to the tyrosine dephosphorylation of Src, Syk and PLCγ2, as well as the reduction of Ca²⁺ influx and PKC phosphorylation in GPVI signaling.

Conclusions

OroA inhibits platelet activation, thrombus formation and myocardial injury via reversing SHP-2 oxidative inactivation thereby attenuating collagen-induced GPVI signaling. With minor bleeding risk and no obvious pharmacological toxicity, OroA holds promising therapeutic potential as an antithrombotic drug.

Introduction

Arterial thrombotic events, such as myocardial infarction (MI) and stroke, are the leading causes of morbidity and mortality worldwide [1,2,3]. Platelet activation and its subsequent arterial thrombogenesis are key pathological processes in arterial thrombosis [4]. Unfortunately, despite antiplatelet drugs are proven to be beneficial in coronary artery disease (CAD), morbidity and mortality rates remain high [5,6,7]. This underscores the urgent need to explore new antiplatelet agents to reduce thrombotic events in CAD patients [8].

Oroxylin A (OroA), an O-methylated flavone, is primarily found in Scutellaria baicalensis, S. lateriflora and Oroxylum indicum [9,10,11], which are widely used in Traditional Chinese Medicine and Ayurveda [12]. OroA has demonstrated a variety of pharmacological properties, including anti-oxidant [13, 14], anti-inflammatory [15], anti-liver fibrosis [16], neuroprotective [17] and anti-cancer [18, 19] properties. Previous studies also revealed that OroA exerts beneficial effects on myocardial infarction and limb ischemia by reducing cardiomyocytes ferroptosis and increasing angiogenesis of endothelial cell [20, 21]. Moreover, OroA has been reported to protect against cardiac structural and functional damage caused by diabetes [22]. However, the antithrombotic effects of OroA, especially its direct effect on platelet activation during myocardial infarction, are not yet fully understood.

The accumulation of oxidative stress is a key factor in cardiovascular diseases (CVD), which contributes to atherosclerosis, MI and stroke [23, 24]. Current evidence suggests that reactive oxygen species (ROS) activates inflammatory response, exacerbates endothelial cell permeability and enhances smooth muscle cell proliferation involving in cardiovascular pathogenesis [25, 26]. Antioxidants, such as methotrexate and doxycycline, exhibit multiple benefits in atherosclerosis and myocardial infarction [27]. Mechanistic studies revealed that ROS oxidative modify and manipulate AMPK [28], PKG1α [29] and SERCA [30] activity, which respectively contributes to macrophage efferocytosis, vasoconstriction and cardiac remodeling. Intracellular ROS are also accumulated during platelet activation, which further enhances platelet function and thrombus formation [31, 32]. Therefore, the suppression of ROS production exhibits a potent antiplatelet and antithrombotic effect [33, 34]. Given that flavones possess strong antioxidant properties [35, 36], whether OroA can directly decrease ROS and manipulate oxidative modification to inhibit platelet activation remains unclear.

In this study, we demonstrated that OroA attenuates thrombus formation and ameliorates myocardial injury post-MI in vivo and ex vivo. We elucidated the underlying mechanism by which OroA decreases oxidative stress and reverses SHP-2 oxidative inactivation, thereby suppressing the Src/Syk/PLCγ2/(Ca²⁺ + PKC) signaling pathway downstream of GPVI activation by collagen. We also confirmed that OroA exhibits a minor bleeding risk without obvious pharmacological toxicity, which holds a promising therapeutic potential as an antiplatelet drug.

Materials and methods

Reagents

Oroxylin A (Baicalein 6-methyl ether, high-performance liquid chromatography grade with ≥ 98%) was provided by Sigma-Aldrich (Germany). The OroA was dissolved in DMSO and stored at -20 ℃ [37]. The final concentration of DMSO did not exceed 0.1% throughout the experiment. In the animal experiment, OroA was administered via oral gavage.

Adenosine diphosphate (P/N 384), collagen (type I, P/N 385), thrombin (P/N 386) and luciferin (P/N 395) were purchased from Chrono-Log (Havertown, PA, USA). Apyrase grade VII, aspirin, human fibrinogen and quinacrine dihydrochloride were purchased from Sigma-Aldrich (Germany). Fura-2 AM, calcein AM, Lipid Peroxidation MDA Assay Kit and total superoxide dismutase assay kit were purchased from Beyotime Biotechnology (Shanghai, China). Rhodamine phalloidin was purchased from Thermo Fisher Scientific (Waltham, MA USA). The antibody of FITC-labeled PAC-1 and PE-labeled CD62P used for human platelets in flow cytometry were purchased from Biolegend (San Diego, CA, USA). The antibody of PE-labeled JON/A and FITC-labeled CD62P were purchased from Emfret. Iron trichloride was purchased from Sinopharm Chemical Reagent Co., Ltd. Acetylcysteine, 6-carboxy-2’,7’-dichlorodihydrofluorescein diacetate (DCFDA) and Digitonin were purchased from MedChemExpress (USA). PF4 ELISA kit was purchased from Abcam (Cambridge, MA, USA). lactate dehydrogenase (LDH) cytotoxicity assay kit was purchased from Biosharp. Total glutathione/Oxidized glutathione assay kit was purchased from Nanjing Jiancheng Bioengineering Institute (Nanjing, China). Glutamic-pyruvic Transaminase (GPT/ALT) Activity Assay Kit, Micro Glutamic-oxalacetic Transaminase (GOT/AST) Assay Kit, Micro Uric Acid (UA) Content Assay Kit and Urea Nitrogen (UN) Content Assay Kit were purchased from Solarbio (Beijing, China). Antibodies against phospho-Src Family (Tyr416), phospho-Syk (Tyr525/526), phoshpo-PLCγ2 (Tyr1217), phospho-(Ser) PKC Substrate, total-Src, total-Syk, total-PLCγ2, Integrin beta 3, Akt1 (C73H10) Rabbit mAb and SHP-2 Polyclonal antibody were purchased from Cell Signaling Technology (Beverly, MA, USA). Anti-phosphotyrosine Mouse mAb was purchased from PTM BIO. Anti-Integrin beta 3 (phospho Y773), anti-AKT1 monoclonal antibody (ab235958), Goat Anti-rabbit IgG (H + L) Antibody (HRP) and Goat Anti-mouse IgG (H + L) Antibody (HRP) were purchased from Abcam (Cambridge, MA, USA). PDGFR alpha Monoclonal antibody and GAPDH Polyclonal antibody were purchased from Proteintech (USA). Oxidized PTP Active Site Antibody and IgG antibody were purchased from R&D Systems (Minneapolis, MN, USA). All other reagents were reagent grade, and deionized water was used throughout. All primary antibodies were diluted at a 1:1000 ratio and second antibodies were diluted at a 1:5000 ratio in western blotting. The dilution for SHP-2 Polyclonal antibody was at a 1:50 ratio. The dilution for Oxidized PTP Active Site Antibody and IgG antibody were at 1:1000 ratios.

Subjects

Human blood samples were obtained from healthy volunteers who did not take antiplatelet or nonsteroidal anti-inflammatory drugs for at least 14 days before blood collection. All experiments using human subjects were performed in accordance with the Declaration of Helsinki and approved by the Institutional Review Board of Yueyang Hospital of Integrated Traditional Chinese and Western Medicine, Shanghai University of Traditional Chinese Medicine (2022-051).

Animals

All animal procedures were approved by the Ethical Committee of Shanghai University of Traditional Chinese Medicine (Approval No. PZSHUTCM2304240008) and conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals (NIH Publication No. 85 − 23, revised 1996). Wild-type C57BL/6 male mice were purchased from Bikai Keyi and housed in specific pathogen-free (SPF) facilities. all mice were weighed and euthanized using sodium pentobarbital for blood sample collection. All in vivo experiments were performed in a blinded manner, with the data analysts remaining unaware of the experimental groups throughout the study.

Intravital microscopy of FeCl3-induced thrombosis in mouse mesenteric arteriole

Intravital microscopy of FeCl₃-injured thrombus formation in mouse mesenteric arterioles was performed as described previously with minor modification [38]. Mice were administered with OroA (20 mg/kg or 40 mg/kg) [16, 39], aspirin (20 mg/kg) or 0.1% DMSO via oral gavage (o.g.) daily over a three-day period. Washed platelets from mice were labeled with calcein AM and subsequently injected into the mice through the jugular vein. The mesenteric arterioles were injured by applying a filter (1 mm × 2 mm) saturated with 10% FeCl₃ solution for 1 min. Thrombosis was observed for 20 min using an inverted fluorescence microscope (Zeiss, Axio Observer 7). Images were acquired in time series and the diameters and areas of the platelet-covered areas were measured using Image-Pro software (Media Cybernetics Inc., Rockville, MD, USA). The time to first thrombus (> 20 μm and stable for > 2 min) [40] and the final areas in 20 min are analyzed.

Thrombus formation under flow conditions

Mice were administered with OroA (20 mg/kg or 40 mg/kg, o.g.) [16, 39], aspirin (20 mg/kg, o.g.) or 0.1% DMSO daily over a three-day period. Bioflux plates were coated with collagen (40 µg/mL) overnight and blocked with 5% BSA. Anticoagulated whole blood samples were collected from abdominal aortas using syringes prefilled with an 0.1 ml ACD solution (85 mM sodium citrate, 71.38 mM citric acid and 27.78 mM glucose) and incubated with quinacrine dihydrochloride [41,42,43] for a duration of 30 min. Blood was perfused over the plates at a shear force of 1000 s^− 1 using a Bioflux-200 system (Fluxion, South San Francisco, CA). Platelets were permitted to adhere to the collagen surface for a duration of 3 min, and thrombus formation was visualized in real-time using an Olympus IX73 inverted fluorescence microscope. Images were acquired and the platelet-covered areas were measured using Image-Pro software.

Pulmonary embolism model

Mice were randomly assigned to receive either 0.1% DMSO or OroA (40 mg/kg, o.g.) for 3 days prior to the experiment. A pulmonary embolism model was then induced by intravenous administration of collagen (100 µg/kg) and epinephrine (600 µg/kg). After a 5 h observation period, the mice were euthanized under anesthesia, and lung tissues were harvested for histological analysis using H&E and CD61⁺ (platelet marker) staining.

Bleeding assay

Mice were randomly assigned to receive 0.1% DMSO, OroA, or aspirin for 3 days. The bleeding assay was performed using the tail amputation method as previously described [44]. Mice were placed prone on a warming pad with the tail extending outward. The tail was transected 3 mm from the tip and immediately immersed in 10 ml of saline, maintained at 37 °C. Bleeding time was recorded from the moment of tail transection until blood flow ceased for more than 2 min. After the addition of red blood cell lysis buffer, red blood cells were lysed. Blood loss was calculated using a standard curve produced from known volumes of mouse blood.

LDH assay

Following the incubation with OroA (50 or 100 µM) or Digitonin (50 µg/mL) for varying durations (2 h, 4 h, 8 h, 24 h), human washed platelets were centrifugated and the supernatant was subjected to LDH assay. Wild-type mice were randomly assigned to one of three treatment groups: 0.1% DMSO, OroA (40 mg/kg, o.g.), or OroA (360 mg/kg, o.g.). Following treatment, the mice were anesthetized, and blood samples were collected. Plasma was then separated from the blood for subsequent analysis. The activity of released LDH was measured using a commercial ELISA kit (Biosharp). The absorbance of the solution was measured at 450 nm in the Spark reader (TECAN). The LDH release was expressed as % of total LDH level, and the total LDH level was detected in platelet supernatant lysed by 0.5% Triton X-100.

Determination of AST, ALT, UA and BUN

Wild-type mice were divided into three groups and administered either 0.1% DMSO, OroA (40 mg/kg, o.g.), or OroA (360 mg/kg, o.g.). Following anesthesia, blood was collected, and PPP was separated. The level of alanine aminotransferase (ALT), aspartate transaminase (AST), blood urea nitrogen (BUN) and blood uric acid (UA) in plasma were measured using the respective assay kits. Optical density (OD) values were determined with Spar Multimode Microplate Reader (Tecan).

Myocardial ischemia model

Mice were anesthetized with 1.0% isoflurane and intubated endotracheally for mechanical ventilation using a rodent respirator, set to an inspiratory tidal volume of 250 µL at a rate of 130 breaths per minute. A left thoracotomy was performed at the fourth intercostal space, and the left anterior descending (LAD) coronary artery was visualized and ligated with an 6 − 0 silk suture. In the sham-operated group, the same procedure was performed, except the suture was passed under the LAD artery without tying.

The infarct size following MI injury was assessed as previously described [45]. Mouse hearts were flushed with PBS to remove blood and then excised. The hearts were fixed in 4% paraformaldehyde for at least 48 h, embedded in paraffin, and cut transversely into 5 μm thick sections. To determine the infarct size, heart sections were stained with Masson’s trichrome, and the infarct size was calculated as the ratio of the infarct circumference to the total left ventricular (LV) circumference× 100%. The collagen volume fraction was evaluated in 5 randomly selected sections. All measurements and calculations were performed using Image-Pro software.

Echocardiography was performed using a Vevo 2100 system (Visual Sonics, Toronto, Ontario, Canada) equipped with an MS-400 imaging transducer. Mice were anesthetized with isoflurane and positioned supine. After shaving the chest, a left parasternal short-axis view was recorded at the level of the papillary muscles. Simultaneous transverse M-mode tracings were captured at the mid-left ventricular (LV) cavity. Left ventricular end-diastolic and end-systolic diameter (LVDd and LVDs, respectively) were measured via M-mode and left ventricular end-diastolic and end-systolic volume (LVEDV and LVESV) were calculated. LV ejection fraction (LVEF) was determined using Vevo 2100 software.

Immunohistochemistry

Following MI injury, mouse hearts were harvested after perfusion, embedded in paraffin, and cut transversely into 5 μm thick sections. Platelets were stained overnight at 4 °C using rabbit anti-mouse CD61 antibody (1:100, Invitrogen) and rabbit anti-mouse 4-HNE antibody (1:100, LSBio). The slides were then rinsed and subjected to biotin-conjugated goat anti-rabbit Ig, followed by avidin-linked enzyme peroxidase complex, with 3,3′-diaminobenzidine serving as the substrate. After counterstaining with hematoxylin and dehydration, images were captured using an Olympus microscope, and the positive staining in each section was quantified using Image-Pro software.

Platelet isolation and platelet functional studies

Human blood was drawn from the antecubital vein and mixed with an ACD solution (85 mM sodium citrate, 71.38 mM citric acid and 27.78 mM glucose) in a 9:1 volume ratio. Platelet-rich plasma (PRP) and platelet-poor plasma (PPP) were prepared by centrifugation, as described previously [45]. To prepare washed platelets, PRP was centrifuged at 1900 rpm, and the resulting pellet was resuspended in Tyrode’s buffer (134 mM NaCl, 2.9 mM KCl, 340 µM Na₂HPO₄•12H₂O, 12 mM NaHCO₃, 20 mM HEPES, 1 mM MgCl₂, 5 mM glucose, pH 7.35, 37 °C) at a concentration of 3 × 10⁸ platelets/mL. Mouse blood was drawn from the portal vein, then PRP, PPP and platelet suspensions in Tyrode’s buffer were prepared similarly by centrifuge as previously described [46]. Before experiment, washed platelets were incubated with different concentrations of OroA for 5 min. For controls, washed platelets were treated without OroA in a comparable amount of DMSO. The final concentration of DMSO is 0.1% (v/v).

Platelet aggregation in response to agonists (collagen, thrombin, or ADP) was measured using the aggregometers (Model 700, Chrono-Log) under stirring conditions (1200 rpm) at 37 °C as described previously [47].

Platelet dense granule secretion was simultaneously monitored by measuring ATP release using CHRONO-LUME reagent (Chrono-Log).Platelets were pre-incubated with different concentrations of OroA or DMSO under stirring before platelet aggregation was initiated. PF4 secretion from α granules in platelets from healthy volunteers was detected with ELISA kits.

For clot retraction, washed platelets were separated as previously described [48]. WP (0.4 mL) from healthy subjects and wild-type mice were incubated with different concentrations of OroA or DMSO (0.1%, v/v) for 5 min. Fibrinogen (2 mg/mL) was added, and clot retraction was initiated by thrombin (1.0 U/mL) at 37 °C. Photographs were captured at specified time intervals using a digital camera. Clot surface area was quantified using ImageJ software (version 1.54 g).

For platelet spreading, washed platelets were pre-incubated with/without OroA and incubated onto 100 µg/mL fibrinogen-coated glass coverslips at 37 ℃ for varying durations. After fixed, permeabilized, and blocked with 0.5% BSA, platelets then were incubated with rhodamine phalloidin for 1 h. Inverted fluorescent microscope (Zeiss, Axio Observer 7) was used to acquire images and platelet areas were calculated by ImageJ software (version 1.54 g).

Platelet P-selectin and activated integrin ΑIIbβ3 expression

Fluorophore-labelled antibodies were utilized for the detection of active form of integrin αIIbβ3 (PAC-1-FITC for human platelets, JON/A-PE for mouse platelets, respectively) and P-selectin expression (CD62P-PE for human platelets, CD62P-FITC for mouse platelets, respectively) and. Washed platelets (3 × 10⁷/mL) were activated by different agonists for 5 min [45], and then were incubated with fluorophore-labelled antibodies in the dark at room temperature for 30 min without stirring. At least ten thousand events were recorded for every sample. P-selectin expression, PAC-1 binding, and JON/A binding were subsequently analyzed by flow cytometry (FACSCalibur, BD Biosciences).

Determination of SOD, MDA, GSH and GSSG

PPP and platelets from healthy volunteers and mice were separated from blood by centrifuge. SOD, MDA, GSH and GSSG were determined by ELISA kits according to the manufacturer’s instructions. The Total Superoxide Dismutase Assay Kit, Lipid Peroxidation MDA Assay Kit and Total Glutathione/Oxidized Glutathione Assay Kit are listed above. Optical density (OD) values were measured using a multifunctional microplate reader (Biotek Instruments, Inc.).

ROS measurement

Washed platelets were incubated with 10 µM DCFDA at 37 °C for 30 min and subsequently washed twice with Tyrode’s buffer. Platelets were preincubated with DMSO (0.1%, v/v), OroA or NAC for 5 min, followed by stimulation with collagen (1 µg/ml). ROS generation in platelets was then detected by flow cytometry.

Screening of common targets for oroxylin A and MI

To predict the targets of OroA, four public databases were utilized: the Encyclopedia of Traditional Chinese Medicine (ETCM, http://www.tcmip.cn/ETCM/), Super-Enhancer Archive (SEA, http://sea.edbc.org), Swiss Target Prediction (STP, http://www.swisstargetprediction.ch/), and Super-PRED (https://prediction.charite.de/). The identified protein targets were standardized and converted to their corresponding official gene symbols using the UniProt database (https://www.uniprot.org/).

Data related to myocardial infarction (MI) was gathered from GeneCards (https://www.genecards.org/), Online Mendelian Inheritance in Man (OMIM, https://www.omim.org/), and the Comparative Toxicogenomics Database (CTD, https://ctdbase.org/). The Venn diagram was created to identify the overlapping genes shared by OroA and MI related. The overlapping genes intersect with those associated with platelet activation from GeneCards. Based on the relative scores of platelet activation-related genes, the top 10 genes and their corresponding proteins are presented. The relative input terms and output results are in the supplementary excel.

Molecular docking of core targets

To further identify the potential target of OroA in platelets, molecular docking was performed to evaluate the affinity of OroA to above potential proteins. The crystal structures of PDGFRA, Akt1 and SHP-2 were retrieved from the Protein Data Bank (PDB). The AutoDock Vina 1.1.2 was used to generate the docking input files, and the results were analyzed with the protein–ligand interaction profiler (PLIP). The visualization of the docking results was done using PyMOL.

Cellular thermal shift assay (CETSA) - western blot (WB)

The Cellular Thermal Shift Assay (CETSA) capitalizes on this principle of thermal denaturation to study protein stability and interactions [49,50,51]. The protein lysate of platelets was aliquoted into PCR tubes and treated with OroA (100 µM) or DMSO (0.1%, v/v) for 1 h at room temperature prior to CETSA heat pulse. The solutions were heated at the indicated temperatures for 3 min, followed by cooling at 4 °C for 3 min. After centrifugation for 15 min (12000 rpm, 4 °C), the supernatant was subject to immunoblotting.

Drug affinity-responsive target stability assay (DARTS)

Platelets were lysed in lysis buffer on ice for 15 min. The lysate was then divided into two aliquots: one serving as the control and the other treated with OroA (100 µM) at room temperature for 1 h. The aliquoted lysates were subsequently digested with pronase (0.5 µg/µl) at ratios of 1:500 and 1:1000 for 30 min. Following digestion, the samples were centrifuged at 12,000 rpm for 15 min at 4 °C, and the supernatant was subjected to immunoblotting.

Co-immunoprecipitation

Human washed platelets were pre-incubated with different concentrations of OroA or DMSO (0.1%, v/v) under stirring for 5 min before platelet aggregation was initiated. Platelets were stimulated by collagen (1 µg/mL) for the appropriate time at 37 °C with stirring in Chrono-Log aggregometer and reactions were terminated by addition of equal volumes of chilled 2 × NP-40 lysis buffer (100 mM Tris-HCL pH 7.4, 300 mM NaCl, 2 mM NaF, 2% NP-40, 2 mM EDTA, and protease and phosphatase inhibitor solution) on ice for 15 min. Immunoprecipitation was carried out using anti-SHP-2 antibody and corresponding IgG for 2 h and then incubated with protein A/G-agarose beads overnight on rocker at 4 °C. The beads were then harvested and rinsed 3 times with 1 × NP-40 lysis buffer. Bead-captured proteins were detected by immunoblotting using anti-oxPTP antibody. The primary antibodies and dilutions used in these assays are listed above.

Immunoblotting

Human washed platelets were pre-incubated with different concentrations of OroA or 0.1% DMSO under stirring for 5 min before platelet aggregation was initiated. Mice were randomly assigned to receive either 0.1% DMSO or OroA (40 mg/kg, o.g.) daily for 3 days prior to the experiment. We collected the blood sample from mice and prepared the washed mouse platelets. Washed platelets were stimulated by collagen (1 µg/mL) for the appropriate time at 37 °C with stirring in Chrono-Log aggregometer. Reactions were terminated by addition of RIPA buffer containing protease inhibitor and phosphatase inhibitor. The reducing agents were included in the protein loading buffer (5X). Proteins were separated by 10% SDS-polyacrylamide gel electrophoresis (PAGE), transferred onto PVDF membranes, and then subjected to immunoblotting. The immunoreactive bands were visualized by enhanced chemiluminescence and imaged with Tanon 5200muti (Tanon Science, Shanghai, China) after incubation with the corresponding secondary antibodies. The primary antibodies and dilutions used in these assays are listed above. After the visualization, the immunoreactive bands were subjected to Image J for grayscale intensity measurement.

Calcium mobilization

Agonist-induced Ca²⁺ influx in platelets was performed as described previously [34]. Human washed platelets were prepared by centrifuge as previously described [45]. Briefly, platelet-rich plasma was separated by centrifuging whole blood at 300 g for 10 min. After centrifuging platelet-rich plasma at 900 g for 10 min, the platelet poor plasma was collected and the platelet pellet was then resuspended in Tyrode’s buffer and loaded with 5 µM Fura-2 AM at 37 °C for 30 min. After washing, the platelets were incubated with OroA for 5 min, and then stimulated with collagen (1 µg/ml) at 37 °C under continuous stirring for 5 min. Fura-2 AM was alternately excited at 340 nm and 380 nm, with fluorescence emission detected at 510 nm. The fluorescence signal was recorded, and the maximum 340/380 ratio values were calculated using a Duetta fluorescence spectrophotometer (HORIBA Scientific) according to manufacturer’s instructions.

Statistical analysis

Data are expressed as mean ± SD. For normally distributed data with one variable, differences between the two groups were analyzed using unpaired Student’s t-test, and differences among more than two groups were analyzed using one-way analysis of variance (ANOVA) followed by Tukey multiple comparison test. For normally distributed data with two variables, differences were evaluated by two-way ANOVA with Tukey multiple comparison test. All analyses were conducted using Prism 9.0 software (GraphPad Inc., San Diego, CA, USA), with statistical significance defined as P < 0.05.

Results

OroA inhibits thrombus formation

We first evaluated the effects of OroA on thrombus formation using different animal models. Wild-type mice received intragastric administration of OroA (20 or 40 mg/kg, once per day) for three days according to the in vivo cardioprotective experiments as previously described [39], and FeCl₃-induced mesenteric arterioles injury was initiated and recorded 20 min after the last administration. As shown in Fig. 1A, Aspirin (20 mg/kg), as a positive antithrombotic drug, remarkedly prolonged the time to first thrombus (706.6 ± 80.6 versus 265.8 ± 40.5 s) and decreased the area of thrombus (4.7 ± 1.8 versus 32.6 ± 3.3 × 10³ µm²) at 20 min after FeCl₃ injury. Intriguingly, OroA dose-dependently suppressed thrombus formation, and OroA (20 mg/kg) prolonged the time to first thrombus (438.3 ± 81.5 versus 265.8 ± 40.5 s), and reduced the area of thrombus (18.7 ± 2.1 versus 32.6 ± 3.3 × 10³ µm²) at 20 min after FeCl₃ injury. The high-dose OroA (40 mg/kg) presented a similar antithrombotic effect to aspirin, which dramatically prolonged the time to first thrombus (669.3 ± 169.0 s of OroA versus 706.6 ± 80.6 s of aspirin), and markedly reduced the area of thrombus at 20 min (6.6 ± 2.2 × 10³ µm² of OroA versus 4.7 ± 1.8 × 10³ µm² of aspirin). We then assessed the role of OroA on thrombus formation under arterial shear stress using a microfluidic whole-blood perfusion assay. Whole blood was drawn from wild-type mice which received intragastric administration of OroA (20 or 40 mg/kg, once per day) and aspirin (20 mg/kg, once per day) for 3 days, and perfused the channel precoated with collagen at the shear rate of 1000 s^− 1 for 3 min. Similarly, OroA dose-dependently decreased platelet adhesion on the channel throughout the perfusion period (Fig. 1B). Since the 40 mg/kg OroA treatment demonstrated a better anti-thrombotic effect compared to the 20 mg/kg dose, we chose 40 mg/kg as the treatment concentration for the subsequent in vivo animal studies. We further determined the antithrombotic effect of OroA using a pulmonary embolism model induced by epinephrine and collagen. Consistently, H&E staining revealed that 40 mg/kg OroA treatment improved the lung structure after epinephrine and collagen injection, which showed less intrapulmonary thrombi in the pulmonary artery (Fig. 1C), and further analysis revealed that OroA diminished platelet infiltration (CD61⁺ area) in lung. We also evaluated the bleeding risk of OroA by measuring bleeding time and blood loss after tail snip. As shown in Fig. 1D, at the same dose used for in vivo antithrombotic study, OroA (40 mg/kg) slightly increased bleeding time from 484 ± 94 s to 660 ± 132 s, much less than 1287 ± 294 s (P < 0.001) induced by aspirin (20 mg/kg); blood loss was dramatically less after OroA treatment compared with that after aspirin treatment (319 ± 78 versus 571 ± 121 µl, P < 0.001), suggesting that OroA poses a lower bleeding risk compared to aspirin.

Oroxylin A exhibits no obvious pharmacological toxicity

Evaluating the potential toxicity of a pharmaceutical compound is essential for its development and application. To further assess the in vivo toxicity of OroA, whole blood samples from wild-type mice were taken after intragastric administration of OroA for 7 days. Previous study has proven that LDH is released during cell death and tissue damage [52, 53]. No significant difference (p < 0.05) in plasma LDH activity was observed among the groups of mice (vehicle, 1.467 ± 0.2313 U/ml; OroA 40 mg/kg, 1.472 ± 0.1542 U/ml; OroA 360 mg/kg, 1.517 ± 0.08944 U/ml), suggesting that OroA administration in vivo did not cause any obvious tissue damage (Fig. 2A). Similarly, we found that both the experimental dose (40 mg/kg, once per day) and a higher dose (360 mg/kg, once per day) of OroA did not result in alterations for the counts of white blood cells, red blood cells or platelets (Fig. 2B), and showed no obvious liver or kidney damage (Fig. 2C). We also examined the LDH activity in the supernatant of washed platelets treated with OroA for 2–24 h. The result (Fig. 2D) revealed that OroA, at the concentration of 50 and 100 µM, exhibited no obvious cytotoxicity to platelets compared with the cytotoxicity-positive control Digitonin, ruling out the possibility that OroA could be developed as a safe antiplatelet drug.

Fig. 1

Oroxylin A inhibits thrombus formation. A, OroA inhibited thrombus formation in mesenteric arterioles of wild-type mice. Wild-type mice randomly received 0.1% DMSO (once per day, o.g.) or OroA (20 or 40 mg/kg, once per day, o.g.), or aspirin (20 mg/kg, once per day, o.g.) for 3 days before the experiment. FeCl₃-induced thrombosis was performed and the thrombus area was recorded. Representative images at different time points are shown. Statistical analysis of thrombosis assessed using area of thrombus at indicated times and time to first thrombus (> 20 μm) are shown (n = 8). Scale bar = 100 μm. B, OroA inhibited thrombus formation over an immobilized collagen surface at a shear rate of 1000 s^− 1 in whole blood from wild-type mice. After treatment with 0.1% DMSO (once per day, o.g.), OroA (20 or 40 mg/kg, once per day, o.g.), or aspirin (20 mg/kg, once per day, o.g.), whole blood was perfused through collagen-coated bioflux plates. Representative images of thrombus formation at the indicated time points are presented (n = 6). Scale bar = 100 μm. C, OroA reduced the number of pulmonary microthrombus and platelet infiltration in mice with pulmonary embolism. Wild-type mice randomly received 0.1% DMSO (once per day, o.g.) or OroA (40 mg/kg, once per day, o.g.) for 3 days before the experiment. The mixture of collagen (100 µg/kg) and epinephrine (600 µg/kg) was used to induce pulmonary embolism. After 5 h, the mice were anesthetized and the lungs were taken for H&E and CD61 (platelet marker) staining. Representative images and quantification of number of thrombi observed in a random 10× field or CD61 positive areas expressed as a percentage of the field, respectively (n = 6). Col, collagen; Epi, epinephrine. Scale bar = 100 μm (high magnification view) and 2 mm (low magnification view). D, OroA slightly prolonged the tail bleeding time in mice. Wild-type mice randomly received 0.1% DMSO (once per day, o.g.), OroA (40 mg/kg once per day, o.g.), or aspirin (20 mg/kg once per day, o.g.) for 3 days before the experiment. Bleeding time and tail blood loss were measured, and each dot represents a single mouse (n = 12).Statistical analyses were performed using one-way ANOVA followed by Tukey’s multiple comparison test in (A and D), and two-way ANOVA followed by Sidak’s multiple comparison test in (B), and unpaired two-tailed Student’s t-test in (C). NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

Fig. 2

Oroxylin A exhibits no significant pharmacological toxicity. A, OroA administration did not affect the LDH activity in plasma from dosed mice. LDH release was then quantified using an LDH assay kit (n = 8). Summary data of 8 experiments are shown. B, OroA did not affect blood cell count in mice. Summary data of 8 experiments are presented. WBC, white blood cells; RBC, red blood cells; PLT, platelets. C, OroA administration caused no obvious liver or kidney damage in mice. Analysis of hepatic and renal damage by the level of ALT, AST, BUN and UA in plasma. ALT, alanine aminotransferase; AST, aspartate transaminase; BUN, blood urea nitrogen; UA, blood uric acid. Summary data of 8 experiments are presented. D, OroA did not induce LDH release in platelets. Washed human platelets were incubated with OroA (50 or 100 µM), or Digitonin (50 µg/mL) for 2, 4, 8, and 24 h. LDH release was then quantified using an LDH assay kit (n = 3). Summary data of 3 experiments are shown. Male C57BL/6 mice were randomly received OroA (40 mg/kg, or 360 mg/kg, once per day, o.g.) for 7 days, and whole blood was collected for analysis (A, B and C). All statistical analyses were performed using one-way ANOVA followed by Tukey’s multiple comparison test. NS, no significance; ***P < 0.001

OroA alleviates microvascular obstruction and platelet activation in MI

When the major vessel is occluded, activated platelets infiltrate heart tissues, leading to infarct expansion and a progressive decline in heart function during the progression of myocardial infarction [45, 54]. After highlighting the in vivo antithrombotic role of OroA, we further determined the cardioprotective effects of OroA on MI expansion and microthrombosis caused by ischemia. The left anterior descending coronary artery in mice was permanently ligated to induce MI. As shown in Fig. 3A, compared with vehicle treatment, OroA (40 mg/kg) markedly decreased the infarct size on day 7 post-MI. Cardiac function was also evaluated by echocardiography. In contrast to vehicle-treated mice, those treated with OroA showed improved cardiac function, including left ventricular ejection fraction, left ventricular fractional shortening and left ventricular end-diastolic volume (Fig. 3B). Immunostaining with platelet-specific marker CD61 demonstrated that OroA dramatically decreased platelet infiltration in heart tissue (Fig. 3C). Moreover, treatment with OroA (40 mg/kg) significantly reduced P-selectin (CD62P) exposure and integrin αIIbβ3 (JON/A) activation in platelets from mice after MI surgery induced by collagen, with a more pronounced effect compared to thrombin-induced activation (Fig. 3D). Taken together, these data reveal that OroA may be a potential antiplatelet drug in inhibiting microvascular thrombus formation and preventing MI aggravation.

Fig. 3

Oroxylin A alleviates microvascular obstruction and platelet activation to prevent MI progression. A, Masson’s trichrome staining of heart tissues from wild-type mice on day 7 post-MI. OroA (40 mg/kg once per day, o.g.) was administered for 7 days following MI surgery. The infarct size was quantified as a percentage of LV area (n = 8). Scale bar = 2 mm. B, Representative M-mode echocardiograms of the three study groups on day 7 post-MI. Echocardiography quantification of LVEF, LVFS, LVESV, LVEDV, LVDs and LVDd in mice of three study groups (n = 8). C, CD61 (platelet marker) staining of heart tissues from three study groups. Representative images are shown along with quantification of CD61-positive areas, expressed as a percentage of the field (n = 8). Scale bar = 50 μm (high magnification view) and 1 mm (low magnification view). D, OroA inhibited platelet activation after MI. On day 7 post-MI, washed platelets were separated, and activated by collagen (1 µg/mL) and thrombin (0.05 U/mL). Representative histograms obtained by flow cytometry show P-selectin (CD62P) exposure and activated integrin αIIbβ3 (JON/A) on the platelet surface (n = 8). The dotted lines highlight the mean values of the fluorescence intensities of P-selectin (CD62P) exposure and integrin αIIbβ3 (JON/A) activation on platelets from the vehicle-treated MI mice. Statistical analyses were performed using unpaired two-tailed Student’s t-test in (A), and one-way ANOVA followed by Tukey’s multiple comparison test in (B - D). NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001. MI, myocardial infarction; LV, left ventricular; LVEF, LV ejection fraction; LVFS, LV fractional shortening; LVESV, LV end-systolic volume; LVEDV, LV end-diastolic volume; LVDs, LV dimension-systole; LVDd, LV dimension-diastole

Oroxylin A directly inhibits human platelet function

After confirming the anti-thrombotic effect of OroA both in vivo and in vitro, we sought to establish its direct impact on platelet activation. We first used blood samples from healthy volunteers to detect platelet aggregation and granule release. Washed platelets were treated with vehicle or different concentrations of OroA at 37 °C for 5 min and then stimulated by different agonists. We demonstrated that OroA inhibited platelet aggregation induced by collagen, ADP and thrombin in a concentration-dependent manner. Especially, OroA demonstrated a superior efficacy in inhibiting collagen-induced aggregation (Fig. 4A1). We also detected ATP release from dense granules using a luciferin-luciferase luminescence assay simultaneously with aggregation, and calculated PF4 release from α granules using commercial ELISA kits. Similarly, OroA inhibited ATP release of dense granules in a concentration-dependent manner (Fig. 4A2) and PF4 release of α granules (Fig. 4A3) induced by different agonists. Among these, OroA exhibited the most potent inhibition of granule release induced by collagen. Since both 100 µM and 200 µM OroA significantly inhibited collagen-induced platelet aggregation, we selected 100 µM as the highest concentration for the subsequent in vitro studies instead of 200 µM. Upon agonist stimulation, the transduction of inside-out signaling leads to integrin αIIbβ3 switching into a high-affinity state for ligands [55] and P-selectin (CD62P) exposure on the surface [56], thus we examined the effect of OroA on αIIbβ3 activation and CD62P expression induced by collagen, ADP and thrombin by flow cytometry. Consistent with platelet aggregation, OroA (100 µM) inhibited αIIbβ3 activation and CD62P expression after different agonist stimulation, suggesting that OroA could reduce inside-out signaling to modulate platelet activation (Fig. 4B). Inside-out signaling subsequently trigger outside-in signaling, which induces spreading and clot retraction contributing to thrombus formation [55]. As expected, OroA exhibited a concentration-dependent inhibition of platelet spreading (Fig. 4C) and clot retraction (Fig. 4D), with a more pronounced inhibitory effect at a concentration of 100 µM.

Fig. 4

Oroxylin A directly attenuates human platelet function. A, OroA inhibited platelet aggregation (A1), ATP release (A2) and PF4 release in a concentration-dependent manner (A3). Washed human platelets were pre-incubated with OroA (50 or 100 or 200 µM) or 0.1% DMSO for 5 min before activation with different agonists. Representative curves and summary data of 3 to 6 experiments are presented. B, OroA inhibited integrin αIIbβ3 activation and P-selectin (CD62P) exposure in a concentration-dependent manner. Representative histogram obtained by flow cytometry for activated αIIbβ3 (PAC-1) or P-selectin expression (CD62P) on platelet surface, respectively. Results and summary data of 3 experiments are presented. The dotted lines highlight the mean values of the fluorescence intensities of P-selectin (CD62P) exposure and integrin αIIbβ3 activation on platelets treated with the agonist-vehicle. C, Representative fluorescence images (phalloidin) showing that OroA inhibited platelet spreading on immobilized fibrinogen. After preincubation with OroA (50 or 100 µM) or 0.1% DMSO for 5 min, washed human platelets were allowed to spread for the indicated times. Representative results and summary data of 6 experiments are presented. Scale bar = 10 μm. D, OroA inhibited clot retraction induced by thrombin. WP was preincubated with OroA (50 or 100 µM) or 0.1% DMSO for 5 min and then stimulated with thrombin (1 U/mL). Representative results and summary data of 6 experiments are presented. Statistical analyses were performed using one-way ANOVA followed by Tukey’s multiple comparison test was performed in (A-C), and two-way ANOVA followed by Sidak’s multiple comparison in (D). NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

Intragastric administration of oroxylin A inhibits platelet function in mice

Following the administration of OroA at a dosage of 40 mg/kg (o.g.) per day for three consecutive days, blood samples were collected from the mice to evaluate platelet function. OroA treatment markedly reduced platelet activation in response to different agonists. Consistent with its inhibitory effect on human platelets, OroA shows the stronger inhibiting effect on collagen-induced platelet aggregation (Fig. 5A). Further experiments revealed that OroA suppressed platelet spreading and attenuated clot retraction (Fig. 5B and C). Flow cytometry assay showed that OroA also diminished integrin αIIbβ3 activation and inhibited CD62P expression induced by different agonists (Fig. 5D). Both in vitro human platelet function studies in Fig. 4 and ex vivo mice platelet function studies in Fig. 5 demonstrated that OroA shows an antiplatelet effect, particularly by inhibiting platelet activation induced by collagen.

Fig. 5

Intragastric administration of Oroxylin A attenuated platelet function in mice. A, OroA inhibited platelet aggregation induced by collagen, ADP and thrombin using washed mouse platelets. Wild-type mice randomly received vehicle or OroA (40 mg/kg, once per day, o.g.) for 3 days before the experiment. Representative aggregation tracings and summary data of 6 mice are presented. B, OroA inhibited platelet spreading on immobilized fibrinogen. Wild-type mice randomly received vehicle or OroA (40 mg/kg, once per day, o.g.) for 3 days before the experiment. Washed platelets were allowed to spread for the indicted times. Representative results and summary data of 6 mice are presented. Scale bar = 10 μm. C, OroA inhibited clot retraction induced by thrombin. Wild-type mice randomly received vehicle or OroA (40 mg/kg, once per day, o.g.) for 3 days before the experiment. WP was separated and stimulated with thrombin (1 U/mL). Representative results and summary data of 6 mice are presented. D, OroA inhibited integrin αIIbβ3 activation and P-selectin (CD62P) exposure in vitro in a concentration-dependent manner. Washed mouse platelets were pre-incubated with different concentrations of OroA or vehicle for 5 min before activation with different agonists. Representative histogram obtained by flow cytometry for activated αIIbβ3 (JON/A) or P-selectin (CD62P) expression on platelet surface, respectively. The dotted lines highlight the mean values of the fluorescence intensities of P-selectin (CD62P) exposure and integrin αIIbβ3 activation on platelets treated with the agonist-vehicle. Results and summary data of 3 experiments are presented. Statistical analyses were performed using unpaired two-tailed Student’s t-test in (A), two-way ANOVA followed by Sidak’s multiple comparison in (B-C), and one-way ANOVA followed by Tukey’s multiple comparison test was performed in (D). NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

Oroxylin A reduces reactive oxygen species production to inhibit platelet activation

It has been known that ROS is accumulated and initiates oxidative damage during myocardial infarction, and the ligation of the left anterior descending coronary artery in mice inevitably causes myocardial damage, leading to an elevation in oxidative stress levels in heart [57]. After MI surgery in mice, 4-HNE immunostaining, a marker of oxidative stress, revealed that OroA mitigated the high oxidative stress status in heart tissue (Fig. 6A). Similarly, OroA treatment increased superoxide dismutase (SOD) activity, elevated concentrations of reduced glutathione (GSH), and attenuated both malondialdehyde (MDA) and oxidized glutathione (GSSG) levels in plasma after MI surgery, which indicated that the cardioprotective role of OroA may related to its antioxidation effect (Fig. 6B). And we further demonstrated that OroA markedly attenuated oxidative stress in resting platelets, as evidenced by an increase in SOD activity, elevated levels of GSH, and a decrease in both MDA and GSSG concentrations (Fig. 6C). Additionally, OroA also increased SOD activity and GSH content while reducing MDA and GSSG level in activated human platelets stimulated by collagen in vitro (Fig. 6D). It has been demonstrated that SHP-2 oxidation can be reversed by antioxidant NAC [58, 59]. Similar to NAC, OroA reduced collagen-induced ROS generation in human platelets detected by flow cytometry (Fig. 6E). Interestingly, we found that the suppressing effect on platelet aggregation of 50 µM OroA was similar to 1 mM NAC, and no statistically difference was observed in platelet aggregation between OroA plus NAC treatment and NAC treatment alone (Fig. 6F), showing no synergistic effect of OroA plus NAC, and again highlighting the antioxidant effect of OroA contributes to its antiplatelet potential.

Fig. 6

OroA attenuates oxidative stress in platelets. A, 4-HNE staining of heart tissues from wild-type mice on day 7 post-MI. OroA (40 mg/kg, once per day, o.g.) was administered for 7 days following MI surgery. Representative images are shown along with quantification of 4-HNE positive areas, expressed as a percentage of the field (n = 8). Scale bar = 50 μm (high magnification view) and 1 mm (low magnification view). B, OroA increased SOD activity and GSH levels, while decreasing MDA and GSSG levels in plasma from MI mouse. SOD activity, MDA, GSH and GSSG concentrations were quantified using respective biochemical assay kits (n = 8). C, OroA increased SOD activity and GSH levels, while decreasing MDA and GSSG levels in washed platelets from MI mouse. SOD activity, MDA, GSH and GSSG concentrations were measured using corresponding assay kits (n = 8). D, OroA increased SOD activity and GSH levels, while decreasing MDA and GSSG levels in washed human platelets induced by collagen in a concentration-dependent manner. Platelets were pre-incubated with OroA (50 or 100 µM) or 0.1% DMSO for 5 min and stimulated with collagen (1 µg/mL). SOD activity, MDA, GSH and GSSG concentrations were measured using corresponding assay kits. Results and summary data of 3 experiments are presented. E, OroA inhibited ROS production in washed human platelets induced by collagen in a concentration-dependent manner. Platelets loaded with DCFDA were pretreated with OroA (50 or 100 µM) or NAC (1 mM) for 5 min and then induced by collagen (1 µg/mL) for 5 min. The 2’,7’-Dichlorofluorescein (DCF) fluorescence intensity was measured using flow cytometry. The dotted lines highlight the mean values of the fluorescence intensities of ROS produced in platelets treated with the agonist-vehicle. Representative histograms and summary data of 3 experiments are presented. F, OroA attenuated platelet aggregation similar to antioxidant NAC. Washed human platelets were pretreated OroA (50 µM) and/or NAC (1 mM) for 5 min and then stimulated by collagen (1 µg/mL). Representative aggregation tracings and summary data of 3 experiments are presented. All statistical analyses were performed using one-way ANOVA followed by Tukey’s multiple comparison test. NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

Oroxylin A targets SHP-2 in platelets

The therapeutic mechanisms of natural drugs are complex, often involving multiple pathways and targets. This part aims to assess the target of OroA in treating myocardial infarction, guiding further research into its antiplatelet mechanisms. Initially, we identified 5967 MI-related target genes from three disease databases (GeneCards, OMIM, and CTD) and 183 OroA-related target genes from four natural drug and compound databases (ETCM, SEA, Swiss Target Prediction, and Super-PRED) (Fig. 7A1). A Venn diagram was created to identify 126 overlapping target genes between MI and OroA (Fig. 7A2). We further crossed these 126 overlapping genes with platelet activation-related genes from GeneCards database and found 118 potential genes (Fig. 7B1). Of which, top 10 potential target proteins include PDGFRA, AKT1, SHP-2, tPA, PI3K-γ, MAPK8, STAT1, TERT, c-KIT and ESR1, displayed according to their relative scores in platelet activation (Fig. 7B2).

We then conducted molecular docking experiments between the top 3 potential target proteins and OroA. The binding energy scores showed that OroA has the strongest affinity for SHP-2 rather than PDGFRA and AKT1 (Fig. 7C1). The interaction between OroA and the SHP-2 was visualized in Fig. 7C2 via PyMOL software. The PLIP results suggest that SHP-2 forms hydrophobic interactions with OroA at the protein residues 357 threonine (THR), 465 arginine (ARG) and 507 threonine (THR), while establishing a hydrogen bond at residue 510 glutamine (GLN). Next, Cellular Thermal Shift Assay (CETSA), and Drug Affinity-responsive target stability assay (DARTS) were introduced to further identify target of OroA. CETSA showed that OroA (100 µM) increases the thermal stability of SHP-2 at 49 °C and 53 °C in platelets, while it did not increase the thermal stability of PDGFRA, and AKT1 (Fig. 7D). Immunoblot analysis of DARTS samples also showed an increased stabilization of SHP-2 during proteolysis when treated with OroA (100 µM) in platelets, while PDGFRA and AKT1 were proteolysed in the process (Fig. 7E). This data indicated that OroA binds to SHP-2 involving in its anti-platelet activity.

Fig. 7

Oroxylin A targets SHP-2 in platelets. A, Venn diagram showing the intersection of OroA target and MI-related genes. To explore the target of OroA in AMI treatment, OroA-related targets were identified from four databases including ETCM, SEA, STP and Super-PRED, while MI-related targets were obtained from three separate databases as OMIM, Genecards and CTD (A1). The overlapping genes in Venn diagram (A2) show the potential target of OroA in MI treatment. B, Venn diagram showing the intersection of potential target genes and platelet activation-related genes obtained from Genecards database (B1). Top 10 target proteins ranked by the relevance scores were displayed in the table (B2). mC, Docking scores of top 3 potential target proteins were listed in the table (C1) and the docking diagram of OroA with SHP-2 was displayed in the C2. D, Immunoblot analysis of CETSA (cellular thermal shift assay) samples. OroA (100 µM) increased the thermal stability of SHP-2 at 49 °C and 53 °C. Representative results and summary data of 3 experiments are presented. E, Immunoblot analysis of DARTS (drug affinity responsive target stability) samples. OroA (100 µM) enhanced the stability of SHP-2 during proteolysis. Representative results and summary data of 3 experiments are presented. All statistical analyses were performed using two-way ANOVA followed by Sidak’s multiple comparison in. NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

Oroxylin A inhibits SHP-2 oxidative inactivation to suppress GPVI signaling in platelets

Previous studies revealed that ROS upregulated oxidative modification of SHP-2 to decrease its tyrosine phosphatase activity [60]. Considering our results that OroA binds to SHP-2, we sought to assess the potential of OroA on SHP-2 oxidation and its tyrosine phosphatase activity. As shown in Fig. 8A, collagen increased SHP-2 oxidation, which was reversed by OroA (50 and 100 µM) in a concentration-dependent manner. Consistently, collagen induced an increase in the protein tyrosine phosphorylation of numerus proteins compared with resting platelets, in contrast, pretreatment of human platelets with OroA resulted in a reduction in tyrosine phosphorylation (Fig. 8B1), and administration of OroA (40 mg/kg, o.g.) led to a similar decrease in protein tyrosine phosphorylation in wild-type platelets (Fig. 8B2).

Fig. 8

Oroxylin inhibits SHP-2 oxidative inactivation to suppress GPVI signaling in platelets. A, OroA inhibited SHP-2 oxidation in washed human platelets induced by collagen in a concentration-dependent manner. Washed human platelets were pre-incubated with different concentrations of OroA or NAC for 5 min, and then stimulated with collagen (1 µg/mL). Platelets were lysed and immunoprecipitated with SHP-2 antibody. The oxidation of SHP-2 was examined by immunoblotting with anti-oxPTP antibody. Representative results and summary data of 3 experiments are presented. B, OroA inhibited tyrosine phosphorylation (pTyr) in washed human platelets (B1) and washed mouse platelets (B2) induced by collagen (1 µg/mL). Representative results and summary data of 3–6 experiments are presented. C, OroA inhibits the phosphorylation of Src, Syk, PLCγ2 and Integrin β3 in washed human platelets (C1) and washed mouse platelets induced by collagen (1 µg/mL) (C2). Representative results and summary data of 3–6 experiments are shown. D, OroA inhibited Ca²⁺ influx induced by collagen (1 µg/mL). Washed human platelets loaded with Fura-2AM were incubated with OroA (50 or 100 µM) for 5 min and measured for 300 s with collagen (1 µg/mL). Representative results and summary data of 3 experiments are presented. E, OroA inhibits the phosphorylation of PKC in washed human platelets (E1) and washed mouse platelets (E2). Representative results and summary data of 3–5 experiments are shown. All statistical analyses were performed using one-way ANOVA followed by Tukey’s multiple comparison test. NS, no significance; *P < 0.05; **P < 0.01; ***P < 0.001

In response to collagen, SHP-2 oxidative inactivation induces the tyrosine phosphorylation of Syk, the crucial signaling downstream of GPVI receptor, which subsequently leads to calcium influx, PKC phosphorylation and integrin αIIbβ₃ activation in platelets [61, 62]. As expected, immunoblotting demonstrated that collagen increased the tyrosine-phosphorylation of Src (Y416), Syk (Y525/526), PLCγ2 (Y1217) and αIIbβ3 (Y773), which was dramatically reduced by OroA (100 µM) treatment (Fig. 8C1) in human platelets. Similar reduction of tyrosine-phosphorylation of Src, Syk, PLCγ2 and αIIbβ3 was observed in platelets from wild-type mice after oral administration of OroA (Fig. 8C2). Consistently, collagen-induced Ca²⁺ influx (Fig. 8D) and protein kinase C (PKC) phosphorylation (Fig. 8E) were downregulated by OroA in platelets.

Discussion

Cardiovascular diseases including thrombotic disorders are the leading causes of mortality and morbidity worldwide. Growing evidence from randomized controlled trials (RCTs) highlights the beneficial effects of traditional Chinese medicine on cardiovascular diseases [63]. For instance, Tongxinluo has shown significant improvements in major adverse cardiac and cerebrovascular events (MACCEs) in patients with MI [64]. Oroxylin A (OroA), a flavone primarily found in Scutellaria baicalensis, S. lateriflora and Oroxylum indicum, has shown a variety of pharmacological properties, including anti-cancer, anti-liver fibrosis and neuroprotective properties [36]. Previous studies have highlighted that OroA increases coronary blood flow and accelerates perfusion recovery during ischemia [20, 65], which has led to further exploration of its potential antithrombotic effects in MI. In this study, we demonstrated that (1) OroA directly attenuates in vivo thrombosis with a lower bleeding risk compared to aspirin; (2) OroA decreases platelet infiltration and ameliorates myocardial injury post-MI in mice; (3) OroA suppresses platelet activation in a concentration-dependent manner, and shows the better antiplatelet activity for collagen-induced platelet aggregation; (4) OroA exhibits its inhibitory effects by reversing oxidative stress and binding to SHP-2 to manipulate SHP-2 oxidative inactivation in GPVI signaling; (5) OroA exhibits no obvious pharmacological toxicity. Collectively, our results indicated that OroA may have therapeutic potential as an antithrombotic drug in MI.

Platelet activation plays a significant role in the increased prevalence of coronary artery disease by promoting arterial thrombus formation. As a result, antiplatelet therapy is crucial in the prevention and treatment of MI with well-established benefits. However, all current antiplatelet drugs, including the widely used aspirin, carry an inherent risk of bleeding [66, 67]. This risk not only limits the ability to increase doses to enhance antithrombotic efficacy but also contributes to mortality rates [68, 69]. Recent studies have revealed a delicate difference between physiological hemostasis and pathological thrombosis, which may offer new opportunities for developing antiplatelet therapy without bleeding risk [70, 71]. Collagen-induced GPVI signaling has been identified as a promising target [72, 73] for the development of an efficient and safe antiplatelet agent. OroA exhibited a more prominent inhibitory effect on collagen-induced platelet activation, strongly suggesting its impact on GPVI. Previous studies revealed that ROS contributes to ADP or thrombin-induced platelet activation [74,75,76,77]. As an antioxidant flavonoid, OroA decreases ROS level in platelets which may explain the antiplatelet effect of OroA on ADP- and thrombin-induced platelet activation. Consistent with two antiplatelet agents targeting GPVI signaling under clinical trial [78, 79], our study further expands these findings and demonstrated that OroA inhibited GPVI signaling to reduce thrombus formation in vivo with minor bleeding risk, demonstrating its efficacy and safety as a candidate for antiplatelet therapy. But the pharmacokinetics data of OroA need to be further investigated to evaluate the potential of OroA as an antiplatelet drug.

Accumulated research has identified oxidative stress as an important pathophysiological pathway in the development and progression of CVD. With regards to the antioxidant effects of the flavone, OroA has been shown to reduce oxidative stress during pulmonary injury [80], alcoholic liver disease [81] and skin aging [82]. In this study, we further demonstrated that OroA dramatically alleviated oxidative stress during MI as evidenced by decreasing the end product of oxidative stress 4-HNE staining in heart tissue, as well as increasing SOD activity and GSH concentration and attenuating MDA and GSSG levels in plasma. Consistent with the result of decreasing platelet infiltration in heart, we found that the administration of OroA suppressed SOD activity, elevates GSH concentration and attenuates MDA and GSSG levels and decreases ROS accumulation in platelets. Further aggregation results suggest that NAC and OroA both act as antioxidants, with no additional inhibition in combination.

Reactive oxygen species accumulation in response to collagen has been found to be responsible for the propagation of GPVI signaling in platelets [83,84,85]. Superoxide undergoes dismutation to H₂O₂ which targets protein-tyrosine phosphatases (PTPs) and inhibits their activity in GPVI signaling [86, 87]. Network pharmacology and molecular docking technology were employed to confirm the exact PTP of OroA, and our results suggest that OroA binds to SHP-2, one of PTPs in GPVI signaling, thereby inhibiting its oxidative modification. The CETSA and DARTS experiments confirmed the binding of OroA and SHP-2, and co-IP experiment further showed that OroA reversed ROS-induced SHP-2 oxidative modification during collagen-induced platelet activation. OroA probably binds onto SHP-2 and inhibits the cysteine residue oxidation in its phosphatase domain thereby regulating its enzymatic activity. The binding between OroA and SHP-2 needs to be further investigated in a pull-down assay using the OroA-specific column in the future.

ROS accumulation upon GPVI stimulation leads to the oxidative modification of SHP-2 and thereby suppresses its tyrosine phosphatase activity, which promotes phosphorylation of tyrosine substrates Src, Syk and PLCγ2 [45, 61, 88]. After showing the inhibiting effect of OroA on SHP-2 oxidative modification, we further found that collagen dramatically increased tyrosine phosphorylation events and pretreatment of platelets with OroA results in an inhibition of protein tyrosine phosphorylation, especially suppresses tyrosine-phosphorylation of Src, Syk and PLCγ2. Previous studies demonstrated that Ca²⁺ influx and PKC phosphorylation are synergistical to activate platelets downstream of Syk signaling, thereby resulting in platelet activation [34, 62]. We consistently found that Ca²⁺ influx and PKC phosphorylation were reversed by OroA, suggesting that OroA inhibits Syk signaling downstream of GPVI to suppress collagen-induced platelet activation.

Conclusion

According to the above analysis, our data provide brand new evidence that Oroxylin A attenuates platelet activation and thrombus formation and therefore alleviates cardiac function deterioration and MI expansion. The markedly antiplatelet effect of OroA depends on the antioxidant property, which binds to SHP-2 and decreases its oxidative inactivation, thereby causing tyrosine-dephosphorylation of Src/Syk/PLCγ2 and decrease Ca²⁺ influx and PKC activation. Our findings suggest that supplementation of OroA in MI prevents platelet activation and thrombotic complications with minor bleeding. This exciting new opportunity awaits further deep research and subsequent clinical translation to assess the safety and the therapeutic efficacy of OroA as a potent antiplatelet agent.