Pharmacological inhibition of Septins with Forchlorfenuron attenuates thrombus formation in experimental thrombotic mice models with modulating multiple signaling pathways in platelets

Graphical abstract

Highlights

• •The widespread use of forchlorfenuron in agricultural production harbours a potential risk to human health.
• •Forchlorfenuron is a specific inhibitor for studies on the biological function of the septin family of the cytoskeleton.
• •Administration of forchlorfenuron in mice has no effect on physiological haemostasis and peripheral platelet properties.
• •Pharmacological inhibition of septins associated platelet inactivation may be helpful in the treatment of thrombotic diseases.
• •Intravenous administration of forchlorfenuron attenuates the formation of pathological thromboses.

Abstract

Introduction

The Septin family of cytoskeletal proteins is abundant in platelets. When these proteins are functionally blocked using the compound forchlorfenuron (FCF), it hampers the normal activation processes of purified human platelets.

Objectives

To evaluate the in vivo effects of FCF on physiological haemostasis and pathological thrombosis in mice and to investigate possible molecular mechanisms.

Methods

The impact of FCF on haemorrhage risk in the brain, liver, and tail of mice was investigated. Using several experimental models, thrombus development in the lung, mesenteric arteries, and postcava was studied. Functional assays were performed on mice and human platelets, both with and without FCF pretreatment. These tests included aggregation, granule release, ROS production, integrin αIIbβ3 activation, cytoskeletal remodeling imaging, and clot retraction.

Results

Neither oral nor intravenous administration of FCF showed any apparent impairment of haemostasis in the tissues studied, but only later administration resulted in a significant reduction in thrombus formation in different mice vessel types. FCF generally inhibited agonist-induced platelet aggregation, degranulation, ROS burst, morphological expansion on the fibrinogen matrix with completely disordered dynamic organizations of the cytoskeleton for septin, tubulin and actin. In addition, FCF was found to antagonise agonist-induced dephosphorylation of VASP (Ser239) and PI3K/AKT and ERK1/2 phosphorylation.

Conclusion

FCF showed preferences in attenuating pathological thrombus formation, apart from physiological haemostasis, with possible mechanisms to prevent cytoskeletal remodelling and signal transduction of AKT, ERK1/2 and VASP signalling pathways, suggesting that Septin may serve as a promising target for the prevention and treatment of thrombotic diseases.

Introduction

Thrombocytes are small, non-nucleated cells with a diameter of about 1.5–3 µm, which mainly originate from megakaryocytes in the bone marrow and circulate in the blood. They play an important role in maintaining homeostasis through physiological haemostasis and in the formation of thrombi when a vascular injury occurs [1]. However, platelet over-activity is a critical factor in certain cardiovascular diseases (CVD). It contributes to thrombosis, which restricts or blocks blood flow in the vessels and leads to ischaemic tissue damage such as stroke and myocardial infarction [2]. Given the central role of platelets in thrombosis, antiplatelet agents such as aspirin (an irreversible COX inhibitor), ticagrelor (an antagonist for the ADP receptor) and the glycoprotein IIb-IIIa (GPIIb/IIIa) inhibitor tirofiban are widely used for the clinical treatment of cardiovascular thrombotic diseases. However, the adverse side effects of these drugs, such as bleeding, skin rashes, neutropenia and thrombocytopenia severely limit their widespread clinical use [3]. Therefore, it is worthy to identify novel anti-platelet or related pharmacological agents.

Septins (septins) are evolutionarily conserved small GTP-binding proteins found in organisms from yeast to mammals. They are crucial for cell motility, cytokinesis, and morphological changes [4], [5], [6]. Humans possess 13 septin subunits (SEPT1-12 and SEPT14) [7], [8]. These proteins form complex structures like filaments and rings, typically as heterotrimers or octamers. Such structures are important to organize the actin-microtubule cytoskeleton [4], [5], [6] and act as scaffolds for protein interactions, facilitating various signalling pathways [7], [9]. It is commonly known that platelets contain large amounts of septins, such as SEPT2, 4, 5, 6, 7, 8, 9, and 11 [8], [10], and their different molecular interactions have also been determined. Indeed, septins in platelets exhibit different modes of distribution in different states. In resting platelets, SEPT5 and SEPT6 accumulate at the periphery, forming a ring-like structure that colocalizes with α-tubulin [11]. During platelet morphological expansion, SEPT5 formed a centralized ring without any interaction with actin [12] and SEPT8 showed both ring-like structures in the periphery and distributed across cytoplasm [7]. The functions of septins in platelet biology are poorly understood. Only Sept5 and Sept8 have been found to impact platelet functioning in studies using knockout mice, however, there are discrepancies in their apparent phenotypes. Sept5 knockout platelets respond strongly to substantial amounts of collagen, which triggers aggregation and release of serotonin and ATP [13]. Depletion of Sept8 decreased platelets sensitivity to convulxin- or thrombin-induced aggregation, integrin αIIbβ3 activation and α-granule exocytosis [7].

A recent study by Kim et al. [14], demonstrated that forchlorfenuron (FCF), a specific small molecule inhibitor of septin organization, impairs both structural and functional aspects of purified human platelets. This finding underscore septins’ importance in platelet biology, though the precise molecular mechanisms have not been described in detail. It should be noted that FCF was originally introduced as a plant growth regulator or herbicide/pesticide and is still widely used in today’s agriculture to promote fruit size. However, the “watermelon explosion” event in 2011 has led to ongoing public concern about the potential risks of FCF residues in agricultural products or contaminated water or soil. The potential hazard of FCF to the cardiovascular system has been demonstrated in recent studies by the toxicities and functionalities of cultured H9c2 cardiomyocytes [15], human umbilical vein endothelial cells (HUVEC) [16], and by cardiac deformation and systolic dysfunction in zebrafish [15]. The main purpose of this study was to evaluate the risk potential of FCF on physiological haemostasis and its pharmacological effects on pathological thrombi development using experimental mouse models and to investigate possible molecular mechanisms.

Materials and methods

Reagents and antibodies

Forchlorfenuron (FCF, Catalog no. 32974) was purchased from Sigma (Shanghai, China) with purity over 98 % and dissolved in DMSO. SB216763 (SB, Catalog no. S1075) was bought from Cell Signaling Technology (Shanghai, China). Thrombin (Catalog no. P/N 386), Collagen (Catalog no. P/N 385) and ADP (Catalog no. P/N 384) were from Chrono-log (Havertown, PA, United States). Platelet-activating factor (PAF, Catalog no. P4904) were purchased from Sigma (Shanghai, China). Fibrinogen (Catalog no. F8050) was purchased from Solarbio (Beijing, China). PGE1 (Catalog no. 900100P) and Methyl cellulose (M0512) were purchased from Sigma (Shanghai, China). TRITC Phalloidin (Catalog no. 40734ES75) was purchased from Yeasen Biotechnology (Shanghai, China). H2DCFDA (Catalog no. D399) reagent was bought from Thermo Fisher Scientific (Shanghai, China). PE anti-mouse/rat CD62P (Catalog no.148306) antibody was purchased from Biolegend (Beijing,China). Rabbit anti-JNK (Catalog no. 24164-1-AP), anti-GSK3β (Catalog no. 22104-1-AP), anti-Sept8 (Catalog no. 11769-1-AP), anti-Sept5 (Catalog no.11631-1-AP) and anti-β-actin (Catalog no. 20536-1-AP) antibodies were purchased from Proteintech (Wuhan, China). Rabbit anti-Phospho-p38 (Thr180/Tyr182, Catalog no. 4511P), anti-p38 (Catalog no. 8690P). anti-Phospho-JNK (Thr183/Tyr185, Catalog no. 4668P), anti-Phospho-AKT (Thr308, Catalog no. 13038P), anti-Phospho-GSK3β (Ser9, Catalog no. 5558P), anti-Phospho-ERK1/2 (Thr202/Tyr204, Catalog no. 4370S), anti-ERK1/2 (Catalog no.4695S), anti-Acetyl-α-tubulin (Lys40, Catalog no. 5335S) and anti-Phospho-PI3K p85 (Tyr458)/p55 (Tyr199, Catalog no. 17366S) antibodies were purchased from Cell Signaling Technology (Shanghai, China). Rabbit anti-Phospho-VASP (Ser157, Catalog no. AP0763) antibody was purchased from ABclonal (Wuhan, China). Rabbit anti-α-tubulin (Catalog no. AF0001) and anti-Phospho-VASP (Ser239, Catalog no.AF55962) antibodies were purchased from Beyotime (Shanghai, China). Rabbit anti-Sept7 (Catalog no. ab186021) antibody was purchased from Abcam (Shanghai, China). Mouse anti-AKT (Catalog no. 60203-2-Ig), anti-Sept7 (Catalog no. 66542-1-Ig), anti-Sept2 (Catalog no. 60075-1-Ig) and anti- PI3K p85 (Catalog no. 60225-1-Ig) antibodies were purchased from Proteintech (Wuhan, China). Mouse anti-β-tubulin (Catalog no. A01030) antibody was purchased from Abbkine (Wuhan, China). Mouse anti-GAPDH (Catalog no. 60004-1-Ig) antibody was purchased from Proteintech (Wuhan, China). Mouse anti-VASP (Catalog no. BF8131) antibody was purchased from Affinity Biosciences (Jiangsu, China). Goat anti-human Fibrinogen (Catalog no. SF108) antibody was purchased from Solarbio (Beijing, China).

Platelets preparation

Healthy blood donors were informed consent of the experiments and agreed to participant in the research and 20 mL blood were drawn from cubitus vein. Blood from anesthetized mice was taken from inferior vena cava. Platelets were prepared from human and mouse blood as described previously [17]. Briefly, blood was drawn into anti-coagulated tubes with 3.8 % sodium citrate (the volume ratio to blood is 1:9) and mixed with equal volume of pre-warmed Tyrode’s buffer containing 1 μM PGE1, supernatant platelet-rich plasma (PRP) was collected after centrifugation for 10 min at 1000 rpm, which was further centrifuged at 2100 rpm for 15 min and platelets pellets were washed three times by Tyrode’s buffer for further functional assays.

Platelet aggregation and ATP releases

Aggregation analysis was carried out on mice and human-washed platelets, at a concentration of 3 × 10⁸ platelets/mL. The process utilized a 4-channel Aggregometer (Tailixin, Shandong, China) and lasted for 5 min, employing specified agonists. Supernatants were collected by centrifugation at 3000 rpm at 4 °C, and ATP contents were assessed with luciferase activity-based ATP determination kit (Beyotime Biotechnology, Shanghai, China) by microplate reader (Thermo Fisher Scientific, USA).

Flow cytometry

Platelets suspension (1 × 10⁷/mL) was pretreated with or without FCF (100 µM) for 10 min at 37 °C, and further stimulated with thrombin (0.01 U/mL) and CaCl₂ (1 mM) in the presences of primary antibodies of PE-CD62P or FITC-fibrinogen. Fluorescence signals were analyzed using Monisight421 flow cytometry (Gaugene, China).

Clot retraction assay

Washed platelets (3 × 10⁸/mL, 200 µL) from mice and humans were mixed with equal volume of mouse or human plasma, and clot retraction was initiated by adding 2 mM CaCl₂ and 0.01 U/mL thrombin at 37 °C. Clot retractions were photographed at 15 min intervals. The clot areas at each time point were measured using Image J software, and the ratios were calculated to quantify the progression of clot retraction.

Platelets viability examination

Platelets (3 × 10⁷/mL, 200 µL) were incubated with Alamar Blue reagent in dark at 37 °C for 4 h and platelets viability were examined with fluorescent intensities (Ex550nm /Em590nm) by microplate reader (Thermo Fisher Scientific, USA).

Platelet adhesion, spreading and imaging

Platelets (3 × 10⁷/mL in Tyrode’s buffer) were seed on cover-slides pre-coated with fibrinogen (50 µg/mL) and further stimulated with combination of CaCl₂ (1 mM) and agonists (thrombin 0.01 U/mL; collagen 2 µg/mL; ADP 20 µM) for indicated time intervals at 37 °C. After fixation with 4 % PFA and permeation with 0.5 % TritonX-100/PBS, platelets were separately incubated with TRITC-Phalloidin, primary antibodies against Sept7, β-tubulin or CD62P and subsequently with corresponding fluorescent labeled second antibodies. An immunofluorescence (IF) microscope was used to visualize and document platelet morphology and target proteins. The adherent or spreading platelets number, area, and mean fluorescence intensity (MFI) were further calculated with Image J software.

ROS measurement

Platelets were pre-loaded with H2DCFDA (10 µM) for 30 min and stimulated with thrombin (0.25 U/mL), and reactive oxygen species (ROS) signals were determined by florescent microplate reader (Thermo Scientific, USA).

Cell cultures

MEG01 cells were maintained in RPMI-1640 supplemented with 10 % fetal bovine serum, along with penicillin (100 mg/mL) and streptomycin (100 mg/mL). Cells were kept in a 5 % CO₂ and 95 % air humidified incubator at 37 °C.

Western blotting

Platelets lysates after aggregation or adhesion assays were subjected into SDS-PAGE, transferred to nitrocellulose membrane, and incubated with specific primary antibodies overnight at 4 °C, followed by a secondary antibody (goat anti-rabbit IRDye 800CW or goat anti-mouse IRDye 680RD). The membranes were then scanned using the Odyssey CLx Imaging System (LI-COR, USA).

Animals

C57BL/6 male mice with an age of 8–10 weeks were used in this study and were maintained in Laboratory Animal Center of Dalian Medical University under the specific pathogen-free condition. For oral administration, randomly grouped mice received with solvent (Sol., 0.1 % DMSO), FCF (50, 100, and/or 250 mg/kg) and asprin (ASP, 10 mg/kg) for once treatment or daily for 9 days. For intravenous administration, mice were injected with solvent (0.1 % DMSO), FCF (2.5, 5, and/or 10 mg/kg) and asprin (1 mg/kg) via tail vein. The maximal oral dose of 250 mg/kg FCF in mice was consistent with the normalized interspecies allometric scaling factor established by the Food and Drug Administration (FDA) to be the equivalent of the maximum toxic in humans as described in previous publications [18]. All the following animal experimental models were conducted in mice under anesthesia, with 2 % isoflurane inhalation showing no response of the toes to being clipped by forceps.

Bleeding assessments

The haemorrhage tests followed the protocols described by Shen et al. [19] with some modifications. The tail haemorrhage tests were performed by immersing the tails of the mice in 37 °C PBS immediately after a 3 mm separation of the distal tip. Bleeding effects were assessed by accumulated bleeding time and frequency of bleeding/haemostasis within 20 min. For the liver haemorrhage study, the peritoneal lavage of the anaesthetised mice was opened along the median abdominal line and the liver tissue was punctured with a calibrated piece (∼5 mg) from the lower edge of the right lobe of the liver. The blood collected in the peritoneal lavage was counted with a haemocytometer 20 min after the liver injuries. For the cerebral haemorrhage study, a 2 mm diameter hole was punctured in the bregma of the anaesthetised mice and a 25 G needle was inserted 4 mm deep into the brain to create a brain lesion. Brain tissue from the euthanised mice was harvested to create 2 mm sections. The haemorrhage status was recorded with a digital camera and then quantified using Image J software.

FeCl₃-induced mesenteric thrombosis

The mesentery of anaesthetised mice was exposed through a 2 cm incision in the ventral midline and removed sequentially on a surgical drape soaked in warm saline. A filter paper (1 × 1 mm²) saturated with 10 % FeCl₃ solution was applied to the mesenteric vessel surface for 5 min. Subsequently, the dynamic blood flow in the injured vessel was monitored for 30 min using an RFLSI III/III-SE Laser Doppler (RWD, China).

Model of inferior vena cava stenosis

Lower vena cava of anesthetized mice was exposed and subjected into a partial ligation leaving 26 G needle space. The venous thrombus was removed from the ligation site in 2 days later, and further measured the length and mass.

Mice pulmonary embolism model

Mice were intravenously injected with one dose of ADP (300 mg/kg) through the tail vein. Two drops of blood were collected from the angular vein 5 min before and after ADP administration to measure changes in peripheral platelet counts. 10 min after ADP application, the lungs were dissected for subsequent immunohistochemical analysis.

Statistics

Data are represented as mean ± standard deviation (S.D.) or standard error of mean (S.E.M.) where indicated. Unpaired student’s t-test and One-way ANOVA followed by Bonferroni post hoc test statistical methods were separately used to compare data from two or more than two groups using GraphPad Prism 8 software, and p < 0.05 showed that the difference was statistically significant.

Results

FCF administration did not affect mice physiological hemostasis

To investigate the potential effects of FCF on physiological haemostasis capabilities, mice were orally administered a dose of FCF (50, 100, 250 mg/kg) or solvent (0.1 % DMSO) 30 min prior to haemorrhage assay, which was performed in various tissues, including tail, liver and brain (Fig. 1A). The tail haemorrhage model in mice showed no obvious effects of FCF on total bleeding time or haemostasis/bleeding frequency, even at the maximum oral dose of 250 mg/kg (Fig. 1B&C), while asprin (ASP, 10 mg/kg) showed significant prolongation of bleeding times (Fig. 1B). Furthermore, in a tissue excision-induced liver haemorrhage test, FCF (250 mg/kg) treatment resulted in similar red blood cell (RBC) accumulation in the peritoneal cavity compared to solvent-treated mice. However, mice administered ASP exhibited significantly more pronounced effects (Fig. 1D). Trauma-induced cerebral haemorrhage in brain slices was also unaffected by FCF, but it increased significantly after ASP treatment (Fig. 1E&F). Platelets isolated from mice given an oral FCF application (250 mg/kg) showed aggregative responses to collagen comparable to those from solvent-treated mice (Fig. 1G&H). In addition, we also examined the effects of prolonged FCF ingestion (9 days, spanning one platelet renewal cycle in mice) on platelet generation and haemorrhage risk (Fig. 1I). It was shown that FCF at the tested doses did not neither affect mice platelet properties (including peripheral counts, MPV and other parameter, Supplementary Fig. 1A-E), nor the duration and frequency of tail haemorrhages (Fig. 1J&K). These data indicate that oral administration of FCF does not interfere mouse platelet biogenesis and physiological haemostasis.

Fig. 1.

The haemorrhage evaluations of FCF oral administration in mice. Mice were i.g. administered solvent (Sol., 0.1 % DMSO), FCF (50, 100, and/or 250 mg/kg), aspirin (ASP, 10 mg/kg), and further subjected to the following bleeding and aggregation assessments. (A-F) Mice were dosed orally once, and bleeding was assessed 30 min later (n = 8 each group); (A) Schematic of experimental design, Mice tail bleeding assay was determined by (B) accumulated bleeding duration and (C) bleeding/re-bleeding times, (D) Mice tissue excision-induced liver haemorrhage tests were determined with RBC counts in the peritoneal cavity (FCF, 250 mg/kg; aspirin, 10 mg/kg), (E) Representative images of trauma-induced cerebral haemorrhage in mice brain slices and (F) quantifications of brain bleeding area (FCF, 250 mg/kg; aspirin, 10 mg/kg). (G) The representative showing aggregative curves of isolated mice platelets in response to collagen (2 µg/mL) stimulation and (H) quantified aggregation rates (n = 3 each group). (I-K) Mice were administered orally daily for 9 days and evaluated in a hemorrhage assay (n = 8 each group); (I) Schematic of experimental design, Tail bleeding assay was determined by (J) accumulated bleeding duration and (K) bleeding/re-bleeding times. Data are shown as the mean ± S.D., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.) or aspirin.

Next, we investigated the effects of i.v. administration of FCF on physiological haemostasis in mice following acute (Fig. 2A) or long-term (Fig. 2I) protocols. We found that FCF (2.5, 5, 10 mg/kg for acute treatment and 10 mg/kg for long-term treatment) did not affect the duration (Fig. 2B&J) and frequency (Fig. 2C&K) of tail haemorrhage in mice. Besides, in both liver (Fig. 2D) and brain haemorrhage models (Fig. 2E&F), the interference of i.v. administration of FCF (2.5 mg/kg) in tissue haemostasis was also excluded. Moreover, mice receiving chronic i.v FCF treatment did not exhibit any obvious changes in peripheral platelet numbers as well as other platelets parameters (Supplementary Fig. 1F-J). All these observations strongly suggest that FCF does not significantly impair the physiological haemostasis capabilities of mice regardless of the route and duration of administration. Interestingly, platelets isolated 30 min after i.v. administration of FCF (2.5 mg/kg) exhibited reduced responses to collagen-induced aggregation (Fig. 2G&H), suggesting FCF may play a role in modulating pathological thrombosis.

Fig. 2.

Bleeding risk assessments of mice after intravenous FCF administration. Mice were i.v. administered with solvent (Sol., 0.1 % DMSO), FCF (2.5, 5, and/or 10 mg/kg), aspirin (ASP, 1 mg/kg), and further subjected to the following bleeding and aggregation assessments. (A-F) Mice were injected once into the tail vein and assessed for hemorrhage 30 min later (n = 8 each group); (A) Schematic of experimental design, Mice tail haemorrhage assay was determined within 20 min by assessing (B) accumulated bleeding duration and (C) bleeding/re-bleeding times, (D) Mice tissue excision-induced liver haemorrhage tests evaluated with RBC counts in the peritoneal cavity within 20 min (FCF, 10 mg/kg; aspirin, 1 mg/kg), (E) Schematic diagram of trauma-induced cerebral haemorrhage in mice brain slices and (F) percentages of changed brain bleeding area within 20 min (FCF, 10 mg/kg; aspirin, 1 mg/kg). (G) Mice platelets were isolated and subjected to collagen (2 µg/mL) initiated aggregation assays and (H) bars graph showing the quantified inhibition rate (n = 3 each group). (I-K) Mice were injected daily with the drug in the tail vein and evaluated for hemorrhage after 9 days (n = 8 each group). (I) Schematic of experimental design, Mice tail bleeding assay was determined within 20 min by assessing (J) accumulated bleeding duration and (K) bleeding/re-bleeding times. Data are shown as the mean ± S.D., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.) or aspirin.

FCF prevents mice thrombus formation in multiple types of vessels

To comprehensively clarify whether FCF would prevent platelet function and thrombus formation in vivo, mice pretreated with FCF (i.v. 2.5 mg/kg) have been studied in different thrombosis models. Initially, deep vein thrombosis (DVT) was induced by inferior vena cava stenosis (Fig. 3A). FCF significantly reduced DVT formation in both length and mass, showing effects comparable to ASP (Fig. 3B&C). Additionally, blood flow was dynamically monitored using laser Doppler in a ferric chloride-induced mice model of mesenteric vascular thrombosis (Fig. 3D&E). FCF prolonged vascular occlusion time similarly effective as ASP (Fig. 3F, solvent 11.7 ± 1.8 min, FCF 17.7 ± 3.5 min, ASP 17.0 ± 1.9 min). Next, we investigated the potential of FCF to modulate microvascular thrombus formation in an ADP-induced pulmonary embolism model. ADP induced significant microvascular occlusion in the lung, evidenced by the number of thrombotic vessels (Fig. 3G&H) and decreased peripheral platelet counts (Fig. 3I). These effects were notably reduced in mice following i.v. administration of FCF (Fig. 3G-I). These results indicate that FCF has beneficial effects on the prevention of pathological thrombosis in veins, arteries and microvessels without affecting physiological haemostasis. However, those anti-thrombotic effects were unable to be reproduced in mice after orally treated with high doses of FCF (Supplementary Fig. 2).

Fig. 3.

FCF prevents thrombosis formation in mice vasculature. Mice were i.v. administered with solvent (Sol., 0.1 % DMSO), FCF (2.5 mg/kg), aspirin (ASP, 1 mg/kg) for once treatment 20 min in prior of the following thrombosis models. (A-C) Deep vein thrombus formation was initiated by postcava constriction, and (A) representative images showing the thrombus dissected two days later and evaluations of thrombus (B) weight and (C) length (n = 6 each group). (D-F) 10 % FeCl₃ was used to cause injuries in the mesenteric arteries, and blood flow was dynamically monitored by Laser Doppler for 30 min (n = 8 each group); (D) Representative images showing the changes of blood flow in the injured vessel within the indicated time points, (E) arbitrary reperfusion curve plots, and (F) calculated vessel occlusion times. (G-I) ADP (300 mg/kg) was used to induce pulmonary embolism via tail vein injection (n = 5 each group); (G) H&E staining images showing thrombus occurrence in mouse lungs, and (H) quantifications of thrombotic vessels in lung tissue sections and (I) percentage of peripheral platelet drops after ADP treatment. Data are shown as the mean ± S.D., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.) or aspirin.

FCF antagonizes agonist-induced platelet aggregation, degranulation, integrin αIIbβ3 activation and ROS generation

To verify direct pharmacological links of FCF with platelets functions, mice platelets were purified and subjected to functional in vitro assays in which a dose range of FCF (25, 50 and 100 µM) was pre-incubated. Initially, FCF inhibited agonist-induced platelet aggregation in a concentration-dependent manner (Fig. 4A, Supplementary Fig. 3A) using thrombin (0.01 U/mL), collagen (2 µg/mL), ADP (20 µM). An Alamar Blue assay confirmed FCF’s safety, showing no effect on platelet viability even at the highest tested dose of 100 µM (Supplementary Fig. 3B). Meanwhile, ATP secretion, a typical indicator of dense granule exocytosis associated with platelets activation, was strongly stimulated upon thrombin stimulation, whereas the effects were completely abolished by FCF (Fig. 4B). In parallel, in dose-dependent manner, FCF inhibited thrombin-enhanced α-degranulation, as shown by flow cytometric analysis of CD62P surface translocation (Fig. 4C&D). Immunofluorescence (IF) staining further supported these findings, demonstrating FCF-preserved intracellular CD62P accumulation in thrombin-stimulated platelets (Supplementary Fig. 3C&D). Besides, FCF was also able to suppress platelet-activating factor (PAF, 1 µM)-elicited aggregations and α-degranulation (Supplementary Fig. 3E-H). Transmission electron microscopy (TEM) analysis confirmed FCF’s inhibitory effects on degranulation, revealing substantial granule retention in FCF-pretreated human platelets after thrombin stimulation (Supplementary Fig. 4A). To assess FCF’s impact on platelet activation, a FITC-conjugated anti-fibrinogen antibody was used to detect thrombin-activated integrin αIIbβ3. The effects were significantly reduced by FCF in dose-dependent manner (Fig. 4E&F). Furthermore, FCF suppressed thrombin-induced reactive oxygen species (ROS) (Fig. 4G), suggesting that ROS-linked platelet signal transduction, activation, and thrombosis may be involved in FCF’s pharmacological mechanism.

Fig. 4.

The in vitro responses of FCF on mice platelet to agonists stimulation. Washed mouse platelets were pre-incubated with different doses of FCF (25, 50, and/or 100 μM) at 37 °C for 10 min, subsequently subjected into platelets aggregation, degranulation and integrin αIIbβ3 activation reactions stimulated with thrombin (0.01 U/mL), collagen (2 µg/mL) and/or ADP (20 µM). (A) Bars graph showing the quantified inhibition rates of platelet aggregation (n = 3 each group). (B) Platelets supernatants from aggregation assay (FCF, 100 µM) were harvested for ATP releases by luciferin/luciferase reagents and the data are presented as the relative percentage to solvent treated platelets in resting stage (n = 3 each group). (C) Flow cytometry showing CD62P expression on the platelet surface determined with primary antibody against CD62P-PE, and (D) quantified by the MFI (n = 3 each group). (E) Flow cytometry showing activation status of integrin αIIbβ3 determined with fibrinogen-FITC, and (F) quantified by the MFI (n = 3 each group). (G) After pretreatment with FCF (100 µM) for 10 min, thrombin (0.25 U/mL) induced ROS generation was dynamically monitored in platelets loaded with H2DCFDA probe (n = 10 each group). After treatment with indicated concentrations of FCF, washed mouse platelets were allowed to adhere and spread on FN coated coverslips for 60 min in presence of 0.01 U/mL thrombin; (H) Representative images for platelet morphology were visualized by TRITC-phalloidin staining (Red) and bar graphs (I) summarized adherent platelets number and (J) averaged areas (n = 3 each group). (K) Representative images showing the effects of FCF on thrombin (0.01 U/mL) induced mouse platelets clots progression and (L) quantified by the time-dependent curve plots showing the percentages of clots area to initial area (n = 3 each group). The graphs summarize the data from at least three times independent experiments, data are shown as the mean ± S.E.M., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Activated integrin αIIbβ3 initiates outside-in signalling, leading to adhesion, spreading on the extracellular matrix, and clot retraction for stable aggregate formation. To examine FCF’s effect on these processes, adhesion and spreading of thrombin-stimulated mice platelets were analyzed over an observation period of 60 min. It was found that FCF dose-dependently reduced the number of adherent platelets and the area of platelet spreading (Fig. 4H-J). Furthermore, platelet clot retraction was observed to progress over time, and FCF significantly slowed this process in a dose-dependent manner (Fig. 4K&L). Overall, these results indicated that FCF would directly improve agonist-induced platelet functions in mice.

FCF modulates agonists-mediated AKT, ERK and VASP phosphorylation

To investigate the mechanism behind FCF’s pharmacological suppression of platelet functions stimulated by various agonists, we examined several common signalling pathways involved in agonist signal transduction. For this purpose, the phosphorylation levels of AKT and extracellular signal-regulated kinase (ERK) were analyzed separately. FCF showed dose-dependent inhibitory effects on agonist-induced phosphorylation of AKT and ERK (Fig. 5A, thrombin; B, collagen; C, ADP; D&E, quantification). However, FCF did not affect phosphorylation levels of other key mitogen-activated protein kinases (MAPKs), p38 or c-Jun N-terminal kinase (JNK) (Supplementary Fig. 5). Additionally, FCF reduced phosphorylation levels of AKT, Glycogen synthase kinase3β (GSK3β), and ERK during platelet spreading (Fig. 5F&G). Moreover, FCF appeared to selectively maintain platelet quiescent signalling, as phosphorylation of vasodilator-stimulated phosphoprotein (VASP) at Ser239, but not at Ser157, was immensely preserved during platelets activation with various agonists (Supplementary Fig. 6), suggesting that FCF may also contribute to cyclic nucleotide GMP (cGMP)/PKG signal transduction.

Fig. 5.

The impacts of FCF pretreatment on platelets signalling transduction in mice. Mouse platelets were pretreated with FCF at the indicated doses, and further stimulated with (A) thrombin (0.01 U/mL), (B) collagen (2 µg/mL) and (C) ADP (20 µM) for 5 min. Representative western blotting showing phosphorylation levels of AKT and ERK1/2 as well as (D&E) quantifications (n = 3 each group). After treatment with indicated concentrations of FCF, washed mouse platelets were allowed to adhere and spread on FN coated coverslips for 60 min in presence of 0.01 U/mL thrombin; (F) Western blotting showing the phosphorylation levels of AKT, GSK3β and ERK1/2 as well as (G) quantifications (n = 3 each group). After treatment with FCF (100 µM) in combination with or without SB216763 (10 µM), washed human platelets were allowed to adhere and spread on FN coated coverslips for 60 min in presence of 0.01 U/mL thrombin; (H) Representative images showing platelet morphology changes visualized by TRITC-phalloidin staining (Red) and (I) quantifications for averaged areas of platelets (n = 3 each group). (J) Representative images showing platelet clot retracts formation within one hour and (K) the bar graph summarizing the ratio of clots area to initial area (n = 3 each group). The graphs summarize the data from at least three times independent experiments, data are shown as the mean ± S.E.M., and analyzed using a One-way ANOVA, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

To explore the link between FCF-disrupted signalling and platelet dysfunction, we examined whether GSK3β inhibition could mitigate FCF-induced platelet impairment. Human platelets were pretreated with FCF, with or without the GSK3 inhibitor SB216763 (SB). Results showed that GSK3 inhibition partially reversed FCF-induced delays in platelet expansion (Fig. 5H&I) and reductions in platelet clot progression (Fig. 5J&K).

FCF inhibits functional performances and signaling transduction in human platelets

Although FCF shows a significant antiplatelet effect in mice, these results need to be confirmed in human platelets. Consistently, FCF attenuates agonist-induced platelet aggregation in human platelets (Fig. 6A, Supplementary Fig. 7A). FCF’s effects are associated with decreased phosphorylation of PI3K/AKT/GSK3β and ERK pathways, and increased VASP phosphorylation at Ser239 (Fig. 6B, Supplementary Fig. 7B-E, Supplementary Fig. 7I). Other key pathways for platelet aggregation, including JNK, p38, and VASP (Ser157), remain unaffected (Fig. 6B, Supplementary Fig. 7F-H). FCF also inhibits “outside-in” signaling in human platelets, as shown by spreading on fibrinogen (FN) (Fig. 6C-E) and delayed clot retraction (Fig. 6G&H). In addition, FCF maintained intracellular CD62P accumulation in human platelets stimulated with thrombin (Fig. 6F). Furthermore, FCF significantly inhibited thrombin-stimulated ATP exocytosis in human platelets (Fig. 6I). Therefore, the pharmacological effects and signal modulation effects of FCF are identical in mouse and human platelets.

Fig. 6.

The effect of FCF on human platelet to agonists stimulation. Washed human platelets were pre-incubated with different doses of FCF (25, 50, and/or 100 µM) at 37 °C for 10 min, subsequently subjected into platelets aggregation, western blotting, adhesion and clot retraction assays in response to stimulation with thrombin (0.01 U/mL), collagen (2 µg/mL) and/or ADP (20 µM). (A) Bars graph showing the quantified inhibition rates of platelet aggregation (n = 3 each group). (B) Representative western blotting showing phosphorylation levels of PI3K, AKT, GSK3β, ERK1/2, p-p38, JNK, VASP (S157) and VASP (S239) (n = 3 each group). (C) Human platelets were subjected on FN matrix for one hour in presence of 0.01 U/mL thrombin, and IF images showing platelet morphological alternation and intracellular CD62P expression (Red, F-actin; Green, CD62P). Bar graphs quantified (D) adherent platelet number, (E) averaged platelet area and (F) CD62P MFI values (n = 3 each group). (G) The representative images showing human platelets colt retract progression and (H) the curve plots summarized relative colt retraction (n = 3 each group). (I) human platelets supernatants from aggregation assay (FCF, 100 µM) were harvested for ATP releases by luciferin/luciferase reagents and the data are presented as the relative percentage to solvent (n = 3 each group). The graphs summarize the data from at least three times independent experiments, data are shown as the mean ± S.E.M., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

FCF disrupts platelet cytoskeleton rearrangement

Experimental evidence from genetically modified animals suggests that individual Septin knockout or overexpression would alter platelets biological functions, such that a specific Septin inhibitor FCF could theoretically be associated with Septin dysfunction. To test this hypothesis, we first showed that the major Septin members (including Sept7, 2, 5 and 8) are ubiquitously expressed in human, rat and mouse platelets as well as in human MEG01 cell lines (Fig. 7A). To characterize the dynamic changes of septin complexes during platelet activation, IF staining against SEPT7 (an indispensable subunit to form the core element of the septin hetero-complex) was visualised separately with actin or microtubules during platelets spreading on the FN matrix. As shown in resting platelets (Fig. 7B), SEPT7 polymerized predominantly as a ring-like structure in the periphery and was strongly colocalised with the microtubule border band. Along with platelet spreading, SEPT7 underwent dramatic remodelling. In the early spreading phase, SEPT7 filaments were found together with actin filaments and β-tubulin in filopodia and at the same time SEPT7 dots were also observed in the cytoplasm (Fig. 7C). In platelets that spread further, the filaments disappeared and fewer SEPT7 dots were occasionally seen in the lamellipodial region. Strikingly, the polymerized SEPT7 was not only punctate in the cytoplasm, but also formed a discontinuous ring-like structure that exhibited intense co-staining with centralized microtubules (Fig. 7C). In fully expanded platelets, microtubules were further segmented and depolymerized, forming network structures. We observed strong overlaps of SEPT7 with these reorganized microtubules (Fig. 7C).

Fig. 7.

The distribution of SEPT7 during human platelets spreading. (A) Representative western blotting showing the protein expression of SEPT7, 2, 5 and 8 in mouse brain (positive control), washed platelets (mouse, rat, and human) and MEG01 cell line. IF images showing different localizations of SEPT7 with F-actin or microtubule in (B) resting human platelets or (C) platelets spreading on FN matrix in the presence of 0.01 U/mL thrombin for 15 min. Green, SEPT7; Red, F-actin or β-tubulin. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Remarkably, FCF resulted in an abnormal platelets population with hyper-polymerized Septin filaments aligned along the direction of spreading (indicated by asterisks in Fig. 8A), leading to a disrupted organization of the actin and microtubule cytoskeletal system. Notably, FCF increased the platelet population in the filopodial stage from 15.4 ± 2.1 % to 39 ± 3.1 %, while decreasing lamellipodial platelets from 77.6 ± 1.6 % to 51.2 ± 4.0 % compared to solvent treatment (Fig. 8B). Surprisingly, most FCF-treated platelets displayed unusual high-order microtubule structures with condensed microtubule filaments perfectly co-localized with hyper-polymerized septin filaments (Fig. 8C, indicated by arrows). These structures were observed in 52.8 ± 1.8 % of FCF-treated platelets but were undetectable in solvent-treated platelets (Fig. 8D). Consequently, FCF-treated platelets exhibited a lower percentage (14.4 ± 1.7 %) of centralized microtubule spirals and segmented microtubules compared to 31.4 ± 2.3 % in solvent-treated platelets (Fig. 8D).

Fig. 8.

The influences of FCF on platelet cytoskeletal remodeling. (A&C) The representative IF images showing influences of FCF (100 µM) on SEPT7, actin and microtubule dynamics during thrombin-stimulated human platelet extension on FN matrix for 15 min (Red, F-actin or β-tubulin; Green, SEPT7). Quantifications for (B) platelets percentage in bound, filopodial and lamellipodial stages as well as (D) platelets with indicated microtubule dynamic organization phase, M.B: marginal band, C.C: centralized coil, S/D: segmentation/depolymerization, I.F: irregular filaments; n.d: none detectable. (E) Resting and spreading human platelets (with or without FCF) were subjected to immunoprecipitation with anti-SEPT7 antibody and anti-IgG antibody, followed by western blot analysis with anti-β-actin, β-tubulin, and α-tubulin antibodies. (F) Western blotting showing altered acetylation levels of α-tubulin as well as (G) quantifications (n = 3 each group). The graphs summarize the data from at least three times independent experiments, data are shown as the mean ± S.E.M., and analyzed using a One-way ANOVA or Unpaired student’s t-test, and *p < 0.05, **p < 0.01, and ***p < 0.001 versus solvent (Sol.) or indicated treatment group. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Platelet spreading requires the synergistic involvement of various skeletal proteins. Co-immunoprecipitation (Co-IP) experiments demonstrated increased interactions between SEPT7 and β-actin following platelet activation, which could be further strengthened by FCF treatment (Fig. 8E). However, the interactions of SEPT7 with α- and β-tubulin remained consistent before and after activating spreading, but slightly decreased by FCF (Fig. 8E). Microtubule deacetylation is a hallmark of dynamic remodelling of platelet microtubules and subsequent activation. Strikingly, we found that FCF strongly inhibits thrombin-induced deacetylation of α-tubulin Lys40 (Fig. 8F&G).

Discussion

Our study shows FCF doesn’t significantly impair haemostasis but effectively prevents thrombosis in the mesenteric arteries, deep veins and pulmonary vasculature in mice regardless of route and frequency of administration. In vitro functional studies demonstrated that FCF antagonizes platelet aggregation, inhibits degranulation, ROS formation, and integrin αIIbβ3 activation. It also suppresses platelet spreading and impairs clot retraction. The antiplatelet effects of FCF may be due to its inhibition of AKT and ERK phosphorylation, VASP dephosphorylation, and cytoskeleton remodeling.

FCF, a commonly used agricultural reagent to increase plant size and production, has drawn attention due to its potential toxicity to humans [20]. Previous in vitro studies showed that FCF inhibited thrombin-stimulated integrin α2bβ3 activation and blood clot retraction in human platelets [14], raising concerns about potential physiological haemostatic disorders. Our findings partially align with these observations but provide a more comprehensive understanding of FCF’s effects in vivo. In a long-term study, no obvious adverse effects were found in rats receiving daily oral FCF administration of 100 mg/kg for 28 days [21]. Based on the rules for the conversion of the effective drug dose between the body surfaces of the different species, a dose of 100 mg/kg in rats is equivalent to a dose of 140 mg/kg in mice [18]. We established a dose range of 50–250 mg/kg in our study. Surprisingly, we found that both single and continuous oral administration of FCF up to 250 mg/kg showed no obvious effects on promoting or inhibiting haemorrhage in mice tail, liver, and brain. Moreover, platelets isolated from these mice maintained normal aggregation ability, indicating that oral administration of FCF does not significantly alter platelet function in vivo.

FCF was administered i.v. at dosages of up to 10 mg/kg to determine its effect on tissue haemostasis. It was found that FCF at 2.5 mg/kg inhibited platelet responses to collagen, resulting in fewer aggregation. These results demonstrate that FCF does not impair normal physiological haemostasis, regardless of administration route, and may even reduce agonist-induced pathological thrombosis. Surprisingly, only intravenous, not oral, administration showed effective anti-thrombotic activity in several vessel types examined using thrombotic mouse models. This difference in pharmacological effects could result from changes in FCF’s pharmacokinetic features between the two forms, including maximum serum concentration, half-life, and other undiscovered aspects due to our technical limitations, necessitating future investigation.

To clarify potential mechanisms, it is important first to confirm the association between FCF and its known targets, septins. In platelets and other mammalian cells, this family of cytoskeletal proteins is essential for cell shape formation and ongoing cytoskeletal remodeling [5], [22]. Consistent with previous studies [7], [10], [11], [23], [24], we found expression of major septin family members i.e., Sept7, 2, 5, and 8 in platelets and their consistency in dynamic remodelling into typical centralized ring-like structures, scattered dots or filaments together with active spreading on fibrinogen, where we examined that septins serve as an upstream modulator for the dynamic reorganization of actin and microtubulin in platelets. Upon actin remodelling in spreading platelets, septins enhanced molecular interactions with actin (detected by Co-IP and IF co-staining) and facilitated the development of filopodium and lamellipodium (detected by IF co-staining). We also observed that FCF altered septin skeleton assembly, resulting in the creation of thick, centrally developed septin filaments that expanded outward, replacing normally centered septin rings. This disruption also resulted in the development of irregular actin filaments and altered tubulin structures, which significantly reduced the formation of actin-activated filopodia and lamellipodia.

Our findings highlight the dynamic organization of tubulins in platelets is of great importance for the correct assembly of septins. Despite the molecular associations between tubulins (α- and β-chains) and Sept7 remaining largely unchanged before and after platelet expansion on fibrinogen. FCF treatment was found to transform the central tubulin rings into abnormal filament structures. These altered tubulin structures showed extensive co-staining with the disordered septin filaments induced by FCF. Furthermore, microtubulin skeletal dynamics were initiated by HDAC6-catalysed deacetylation of Lys40 of α-tubulin [25], [26], and we found that higher Ac-Lys40 levels were maintained in thrombin-activated platelets with FCF pretreatment. This aligns with an earlier study by Ageta-Ishihara et al. showing that Sept7 knockout resulted in hyperacetylated microtubules that impaired cerebrospinal axon projections and dendritogenesis in perinatal mice. A mechanistic study suggests that Septins may serve as a scaffold to link HDAC6 to microtubules, contributing to the successful deacetylation of α-tubulin [27]. Our findings suggest that disruption of dynamic Septin assembly may also impair α-tubulin deacetylation, indicating that Septins are necessary for regulating platelet tubulin deacetylation. However, this hypothesis requires further careful investigation using FCF-treated platelets or those with individual Septin gene deletions. In agreement with previous studies in karyocytes [4], [5], our data further confirmed that precise dynamic organization of Septins is indispensable for regulating actin and microtubule rearrangement in platelets. We found that FCF not only directly inhibits Septin assembly but also secondarily interferes with the dynamic remodelling of actin and tubulin rearrangement, leading to inadequate platelet spreading and clot retraction.

Our findings regarding FCF’s effects on platelet morphology and exocytosis differed from some previous reports. While earlier studies by Kim et al. [14] showed that FCF induces resting platelets to change from a disc-like to an elongated shape, we were unable to reproduce this observation. Our TEM imaging and mean platelet volume (MPV) data showed no morphological changes caused by FCF application, even at doses up to 100 μM. These discrepancies could be due to different experimental settings or the indicative septins used for our and their studies, perhaps reflecting different roles of septin members in maintaining platelet morphology.

Regarding platelet exocytosis, our results also diverged from study by Kim et al. reported that FCF promotes surface expression of CD62P in resting platelets and phosphatidylserine exposure in activated platelets, which are normally associated with increased platelet aggregability and activations [14]. However, our study revealed a slight increase of CD62P expression in FCF treated platelet (data not show), but we also demonstrated that FCF plays a significant role in inhibiting agonist-provoked CD62P membrane expression. A similar change was also observed in the detection of P/S exposure (data not show). Interestingly, Our TEM analysis demonstrated substantial granule retention in FCF-treated, thrombin-stimulated platelets, contrasting with the rare granules observed in thrombin-activated controls.

In addition, we observed an inhibitory effect of FCF in thrombin-induced platelet exocytosis. The effect might be due to dysregulation of septin organization by FCF, as Sept8-depleted platelets show reduced α-degranulation and ATP release [7]. It is obvious that septin members are essential for platelet biological functions and play both positive or negative role in platelet activation, exocytosis and thrombosis.

To date, platelet septin functions have only been experimentally demonstrated in Sept5 and Sept8 knockout mice. Sept5-depleted platelets exhibited increased in α- and δ-exocytosis via interaction with α-syntaxin4 and enhanced aggregative responses to collagen [13], [28]. Although, Sept8 depletion attenuated convulxin-triggered CD62P exocytosis and reduced thrombin-induced aggregation, integrin αIIbβIII activation, and spreading to fibronectin [7]. It is hypothesized that the thrombin actions interrupted by FCF are partly due to the inhibition of Sept8. Contradictorily, FCF also inhibited ADP-triggered aggregation and thrombin-induced α- and δ-degranulation, but these were not affected by Sept8 knockout. On the contrary, FCF suppressed collagen-induced platelet aggregation, which is in contrast to Sept5 knockout platelets. Therefore, we hypothesized that in addition to Sept8 and 5, other functional, as yet unidentified septin subunits are also targeted by FCF and should contribute to these impaired platelet functions.

We also observed that FCF suppresses thrombin-induced ROS formation, an important mediator for enhancing agonist-induced platelet activation and thrombosis [29]. This observation suggests novel pathways of septins in the regulation of platelet activation. In our study, we explored the influence of septins on ROS formation in platelets, though we couldn’t definitively pinpoint the exact mechanism. However, previous study provided some clues. For instance, Rac1, a member of the Rho GTPase family, plays multiple roles in platelet function. It not only modulates platelet skeletal remodeling but also contributes to ROS generation triggered by agonists. This occurs through its interaction with p67phox and its influence on the organization of the NADPH oxidase (Nox) complex. In addition, Rac1 was found to induce septin filament formation in focal adhesions (FA) and facilitate microtubule growth in FA of endothelial cells [30]. From this, we deduced that septins may also contribute to platelet ROS generation in a Rac1-dependent manner. This should be investigated in further studies, focussing on testing their associations and the effects of FCF on Rac1/Nox-mediated ROS generation.

Our study provides novel insights into the role of septins in platelet signalling regulation. Platelets respond to various stimuli, including exposed extracellular matrix proteins, adhesive proteins, and agonists, through surface receptors. These receptors initiate signaling cascades that lead to a range of platelet responses, such as aggregation, degranulation, β3-integrin activation, and morphological changes. The universal inhibition of these processes by FCF suggests that septins may be crucial for one or more common downstream signaling pathways.

PI3K/AKT/GSK3β pathway plays critical roles in directing platelet biological activities as targeting this pathway by genetic knockout or pharmacological inhibition would remarkably attenuate agonists stimulated platelets aggregation and thrombus formation. FCF was also found to inhibit PI3K/AKT/GSK3β phosphorylation in platelets undergoing inside-out and outside-in processes. The involvement of septins in AKT signaling pathways has been further confirmed by GSK3β inhibition in which FCF impaired platelet aggregation, spreading and clot retraction would be partially reversed, indicating other pathways might also be affected. Mitogen-activated protein kinases (MAPKs) are known essential for platelets activation, exocytosis and aggregation. We demonstrated FCF preferentially inhibited ERK1/2 phosphorylation, while the phosphorylation statuses of JNK and p38 (the other two MAPKs) were hardly influenced. A similar effect has also been observed in our previous study in which FCF treatment or silencing expression of SEPT2 or SEPT7 would suppress breast cancer cell migration and invasion via ERK1/2 inactivation. Besides, cAMP/PKA and cGMP/PKG pathways are known to play negative roles in regulating platelet aggregation, degranulation and actin assembly, and their activities are commonly evaluated by the phosphorylation levels of VASP at Ser157 and Ser239 respectively. Intriguingly, FCF can selectively antagonize agonists-induced dephosphorylation of VASP at Ser239 but not Ser157, suggesting septins might participate cGMP/PKG/VASP activation, and the details of molecular interplays require further insight studies. Therefore, we demonstrated the novel roles of septins in regulating platelet ERK1/2, AKT and PKG/VASP signaling pathways, which may account for possible mechanisms for FCF induced anti-platelet actions.

Furthermore, our data show for the first time that septins are involved in the regulation of platelet signalling. By sensing extracellular matrix proteins exposed in damaged vessels, adhesive proteins or agonists, platelet surface receptors transduce the signals through their activation pathways and induce a range of platelet responses, including aggregation, degranulation, β3-integrin activation and adhesion/spreading (morphological changes). These processes are intricately modulated by various intracellular signalling pathways. Since FCF showed universal inhibition of these processes, this suggests that septins may be required for one or more common downstream signalling pathways. The PI3K/AKT/GSK3β pathway plays a critical role in controlling platelet biological activities, as targeting this pathway by genetic knockout or pharmacological inhibition would significantly attenuate agonist-stimulated platelet aggregation and thrombus formation [31], [32], [33]. FCF has also been found to inhibit PI3K/AKT/GSK3β phosphorylation in platelets undergoing inside-out and outside-in processes. The involvement of septins in AKT signalling pathways was further confirmed by GSK3β inhibition, in which platelet aggregation, spreading and clot retraction impaired by FCF was partially reversed, suggesting that other signalling pathways may also be affected. MAPKs are known to be essential for platelet activation, exocytosis and aggregation [34]. We could show that FCF mainly inhibited ERK1/2 phosphorylation, whereas the phosphorylation states of JNK and p38 were hardly affected. A similar effect was also observed in our previous study, in which FCF treatment or silencing expression of SEPT2 or SEPT7 suppressed breast cancer cell migration and invasion by inactivating ERK1/2 [35]. Furthermore, the cAMP/PKA and cGMP/PKG signalling pathways are known to play a negative role in regulating platelet aggregation, degranulation and actin assembly, and their activities are usually assessed by the phosphorylation levels of VASP at Ser157 and Ser239, respectively [36], [37], [38], [39]. Interestingly, FCF can selectively antagonise agonist-induced dephosphorylation of VASP at Ser239, but not at Ser157. This suggests that septins may be involved in cGMP/PKG/VASP activation, and the details of the molecular interplay require further investigation. Therefore, our study reveals novel roles for septins in regulating platelet ERK1/2, AKT, and PKG/VASP signalling pathways, explaining FCF’s inhibitory effects. Further studies needed to explore the molecular interplay between septins and these pathways to understand septin function in platelet biology.

Consider potential septin-independent mechanisms of FCF action. Research has shown FCF’s off-target effects in human epithelial cells, inhibiting cell motility and MAPK signalling even without SEPT7 (achieved through siRNA knockdown and CRISPR/Cas9 knockout) [40]. This finding suggests that FCF may interact with additional cellular targets beyond septins. To fully elucidate FCF’s pharmacological mechanisms in platelets, future studies should employ advanced chemical-biological approaches to identify all potential molecular targets. Such comprehensive analysis will provide a more complete understanding of FCF’s effects on platelet function and thrombosis prevention.

FCF demonstrates unexpected potential in preventing thrombotic diseases. Our main objective was only to investigate whether FCF would interfere with physiological haemostasis. However, our detailed experiments showed that intravenous application of FCF in mice was effective against pathological thrombosis, but preserved the normal physiological ability to stop bleeding associated with tissue injury. FCF’s safety profile has been well-established over the past two decades. The US government authorized its use in agricultural products, and extensive research has consistently shown FCF to have low toxicity in humans [41] and other mammals [20], [42], [21]. It is a well-accepted strategy for targeting cytoskeletal proteins in clinical usages against tumor or cardiovascular diseases. A best example was targeting microtubule in tumor therapy by Taxol as well as in cardiac ischemic reperfusion injuries by Colchicine. However, whether or not Septin may serve as a therapeutic target is rarely known, especially in cardiovascular diseases. Our study provided strong evidences to support Septin inhibition with FCF would efficiently antagonize pathological developments of thrombotic diseases with less hemorrhage risks. Therefore, we proposed targeting Septins would be a novel therapeutic direction for thrombotic diseases. Moreover, thrombotic inflammation is acknowledged as important pathological mechanisms in multiple CVDs, which were greatly linked with higher platelets activities. Therefore, further studies would consider investigating the effects of FCF or platelet Septin inhibitions on preventions of cardiac or cerebral tissues after ischemic reperfusion.

There are several unsolved issues or apperceptions derived from our study worthy of being mentioned. First, safety and anti-thrombotic evaluation of FCF were only examined on adult animals. Inspired by metabolism disorders in children [43] and advancement of puberty of rats receiving FCF [20], we realized the populations of children and adolescence have been overlooked, and the future study have to be specifically confirmed in younger animals. Second, options of proper aspirin dose in thrombotic animal models. According to the Guidance of “Equivalent dose calculation based on body surface area between human and animals” [18], the prophylactic doses of aspirin (1–4 mg/kg) in human would be converted into 12–49 mg/kg in mice. Of the note, it seems that the dosages of aspirin used in animal experiments are quite variable from 1 mg/kg to 100 mg/kg by different research teams. For example, Dütting et al, showing 1 mg/kg aspirin is efficient to increase mice tail bleeding [44], while the other studies treated mice with aspirin at 100 mg/kg or 20 mg/kg [45], [46]. Therefore, a standardized dosage range of aspirin as positive anti-platelets reagent is required in case overdose associated unexpected phenotypes. Third, our findings have to be further supported by Sept7 knockout platelets, which raises a requirement of generating platelet specific Sept7 knockout mice as conventional Sept7 knockout mice are embryonic lethal [47]. In fact, the required mice strain is under breeding in our group and we wish to have a chance to report the importance of Sept7 in platelets biological functions and thrombosis in the future.

Conclusion

FCF has desirable pharmacological effects to prevent platelet activation and thrombus formation without obvious interference with physiological haemostasis. Our study shows that actin and microtubule remodelling coordinated by septin are essential for the morphological changes associated with platelet activation. The antagonistic effects of FCF may result from impaired morphological remodeling and insufficient signal transduction of the AKT, ERK1/2, and VASP signaling pathways in platelets during agonist stimulation. The targeted influencing of septin activities or dynamic organization could open up innovative research directions for the treatment of platelets or thrombotic diseases.

Compliance with ethics requirements

Ethics statement: All experiments involving animals were conducted according to the ethical policies and procedures approved by the ethics committee of Dalian Medical University, China (Approval no. AEE22131). Platelets were from healthy volunteers (no anti-platelets medicines intakes) who have been informed consent of the experiments and signed the agreement of participation in this research. Human blood donation and platelet experiments were approved by ethic committee of Dalian Medical University, China (Approval no. DMU2022-010) and conformed to the directives of Declaration of Helsinki.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article: