The Na/K-ATPase α1 subunit fine-tunes platelet P2Y₁₂ function and mediates sex dimorphism–associated thrombosis

Key Points

• •NKA α1 attaches to specific platelet receptors and helps regulate platelet activation and clot formation.
• •Male hormones decrease NKA α1 in platelets, contributing to differences in blood clot risk between men and women.

Visual Abstract

Abstract

Sex differences are well recognized in thrombotic diseases, but the underlying mechanisms remain unclear. The sodium/potassium ATPase (NKA), composed of α and β subunits, regulates ion homeostasis and plays a key role in cardiovascular function. We investigated whether the NKA α1 subunit influences platelet activation and thrombosis. Using the ferric chloride (FeCl₃)–induced carotid artery injury thrombosis model in wild-type (WT; α1^+/+) and NKA α1 heterozygous (α1^+/−) mice, we found that NKA α1 haploinsufficiency significantly inhibited thrombosis in males but not in females, without affecting hemostasis. Platelet NKA α1 expression was halved in α1^+/− mice, but sodium homeostasis remained unchanged. Transfusion of α1^+/− platelets into thrombocytopenic WT mice prolonged the time to occlusive thrombus formation. Low-dose ouabain or marinobufagenin, which bind NKA α1, suppressed thrombosis. Mechanistically, α1 interacted with P2Y₁₂, and this interaction was disrupted by a leucine-glycine-leucine (LGL)→serine-phenylalanine-threonine mutation in either partner or by the LGL peptide. NKA α1 haploinsufficiency, ouabain, and LGL peptide treatment all reduced ADP-induced platelet aggregation. Female mice exhibited higher platelet α1 expression and shorter thrombosis times than males. Gonadectomy had no effect in females but abolished the antithrombotic phenotype in α1^+/– males, whereas orchiectomy increased platelet α1 expression. Although α1 haploinsufficiency did not affect thrombosis in the 10% FeCl₃ model, it prolonged thrombosis time in mice treated with low-dose clopidogrel or prasugrel, which alone had no effect. These findings identify NKA α1 as a key regulator of sex-specific platelet activation and thrombosis, suggesting its potential as a biomarker for thrombotic risk and a therapeutic target for antiplatelet and antithrombotic therapy.

Introduction

Thrombotic events represent a substantial global health concern, significantly contributing to morbidity and mortality worldwide.^1,2 Clinical studies have well-documented significant sex differences in the risk of thrombosis-related diseases, including myocardial infarction, stroke, and venous thromboembolism.^3,4 However, the underlying mechanisms remain unclear. Platelet activation and hyperaggregation at vascular injury sites constitute the primary pathogenic processes of thrombosis and are mediated by platelet surface glycoproteins and G protein–coupled receptors (GPCR).5, 6, 7, 8, 9 Current clinical strategies for preventing platelet-mediated thrombotic events rely on drugs against P2Y₁₂, protease-activated receptor (PAR), or thromboxane A2 receptor (TP); However, these therapies often have major side effects, including systemic hemorrhage.10, 11, 12 Additionally, some patients remain unresponsive to these treatments and experience recurrent thrombosis.² Certain medical conditions, such as heart failure (HF), further elevate thrombosis risk, but the underlying mechanisms remain poorly understood.

The sodium/potassium ATPase (NKA), or sodium pump, is a transmembrane protein composed of α and β subunits.¹³ Four α and 3 β isoforms have been identified, forming tissue- and cell-specific isoenzymes.14, 15, 16 NKA plays a crucial role in maintaining intracellular Na⁺ and K⁺ homeostasis. Additionally, the NKA α1 subunit, encoded by Atp1a1, functions as a signal transducer and is essential for cell growth and differentiation.^14,17, 18, 19 Proteomic studies suggest that α1β3 is the predominant NKA isoform in platelets.20, 21, 22 However, the role of NKA in platelet function and thrombosis remains unclear, with existing data presenting conflicting findings.23, 24, 25 A comprehensive investigation into NKA’s role in platelet biology and thrombosis is lacking.

NKA is highly sensitive to inhibition by cardiotonic steroids (CTS), particularly those of the digitalis class, which have been widely used in treating HF and arrhythmias.^17,26 Digoxin, a CTS, has been linked to increased all-cause mortality in women with HF but not in men.²⁷ Patients with HF also exhibit adverse platelet characteristics, including reduced survival time, increased mean platelet volume, and heightened activation.^28,29 Inhibition of NKA pump function by ouabain enhances Na⁺/Ca²⁺ exchange, leading to elevated intracellular Ca²⁺ levels, phosphatidylserine exposure, and increased platelet aggregation.^24,25,30 Several studies have also suggested that digoxin use increases the risk of thrombotic events.^31,32 However, the mechanisms underlying these observations remain unclear. In this study, we address these gaps by identifying a novel function for the platelet NKA α1 subunit and demonstrating that its expression plays a key role in thrombosis, including sex dimorphism–associated thrombosis.

Materials and methods

Additional details are provided in the supplemental Materials and Methods.

Animals

The α1^+/− mouse strain was maintained through heterozygous wild-type (WT) breeding.³³ Mice aged 8 to 16 weeks were used, with WT littermates (α1^+/+) as controls. All animal procedures were approved by the Institutional Animal Care and Use Committee of Marshall University (protocol number 1033528).

Murine FeCl₃−induced carotid artery injury thrombosis model and tail bleeding assay

The ferric chloride (FeCl₃)–induced carotid artery injury thrombosis model and the tail bleeding assay were performed as previously described.34, 35, 36, 37, 38

Zebrafish estrogen–induced venous thrombosis model

The Tg(fabp10a:fgb-eGFP) line producing green fluorescent protein−tagged fibrinogen was used to evaluate whether knockdown of atp1a1 paralogs affected venous thrombosis.^39,40 Single-guide RNAs targeting early exons of atp1a1 paralogs (Synthego) were mixed with Cas9 and injected into single-cell–stage embryos. Treatment with the estrogen analog mestranol was performed at 5 days after fertilization, followed by thrombosis evaluation at 6 days after fertilization.⁴¹ All data were collected by an observer blinded to the condition.

Platelet function and coagulation tests

Platelet-rich plasma (PRP) and washed platelets from α1^+/− and α1^+/+ male mice were used for aggregation assays.^36,38,42 Whole blood was subjected to a Cellix flow chamber–mediated adhesion assay on a collagen-coated surface.^38,43,44

PRP (2.5 × 10⁸ platelets per mL) was stimulated with 2.5 μM ADP for various periods, and platelets were pelleted and lysed for immunoblotting AKT (protein kinase B) phosphorylation.^36,45 Platelet-poor plasma was used for activated partial thromboplastin time (aPTT) assays. Healthy human PRP was also treated with different concentrations of ouabain and used for aggregation assays.

Platelet sodium pump function assays

Intraplatelet sodium concentration⁴⁶ and resting platelet membrane potential⁴⁷ were measured as indicators of NKA pump function.

Platelet transfusion assay

Thrombocytopenia was induced in α1^+/+ mice.⁴⁸ Donor platelets from age-matched male α1^+/− and α1^+/+ mice were stained with rhodamine 6G and transfused via jugular vein injection in a 1-donor–to–1-recipient ratio. The thrombosis assay was conducted 10 minutes after transfusion.³⁶

Identification of NKA α1–associated platelet GPCR

Human and mouse platelet lysates were used for coimmunoprecipitation (co-IP) or blue-native polyacrylamide gel electrophoresis to identify NKA α1–associated platelet GPCR. COS-7 cells were transfected with plasmids encoding human P2Y₁₂ or A₂B receptors, and co-IP was performed on cell lysates to assess the interaction between endogenous NKA α1 and these GPCRs. For α1 western blotting, samples were mixed with 2× Laemmli sample buffer and incubated at 37°C for 30 minutes.

The interaction between PAR4 and P2Y₁₂ in human platelets is mediated by the conserved leucine-glycine-leucine (LGL) motif,⁴⁹ corresponding to amino acids 121 to 123 in mouse P2Y₁₂. Human NKA α1 also contains an LGL motif, conserved as leucine-glycine-isoleucine (LGI) in mice. We generated mutant variants of mouse P2Y₁₂ and NKA α1 by replacing their LGL(I) sequence with serine-phenylalanine-threonine (SFT). WT or LGL→SFT-mutated P2Y₁₂ and NKA α1 constructs were cotransfected into COS-7 cells in various combinations, followed by co-IP assays to assess whether the LGL(I) motif affects their interaction.

NKA α1 serves as a dual antiplatelet target

To assess the translational relevance, WT males were treated with ouabain (100 μg/kg) or marinobufagenin (MBG, 100 μg/kg)⁵⁰ and subjected to in vivo thrombosis study 18 hours later. To determine whether α1 haplodeficiency enhances dual antiplatelet therapy, α1^+/– and α1^+/+ male mice were treated with clopidogrel (0.5 or 1 mg/kg)³⁸ or prasugrel (0.33 mg/kg)⁵¹ via gavage for 1 week, followed by thrombosis assessment using the 10% FeCl₃–induced thrombosis model.

Statistics

Data are expressed as mean ± standard error of the mean. Statistical analyses were performed using a 2-tailed Student t test, Mann-Whitney test, or 1-way analysis of variance with Bonferroni post hoc correction for multiple comparisons using GraphPad Prism (version 10.2.2). Some data were analyzed with log-rank testing using the Kaplan-Meier survival curve. A P value <.05 was considered statistically significant.

Consent was obtained as per the Declaration of Helsinki.

Results

α1 gene haploinsufficiency confers an antithrombotic effect in male mice without affecting bleeding

We investigated the role of NKA α1 in thrombosis using the 7.5% FeCl₃–induced carotid artery injury thrombosis model in male α1^+/− and α1^+/+ mice. As shown in Figure 1A-B, as well as supplemental Video 1, thrombus formation in α1^+/+ mice was comparable to that observed in previously reported C57BL6/J WT males,^44,52 with an average occlusion time of ∼11 minutes. In α1^+/− mice, initial platelet adhesion and aggregation appeared unaffected (Figure 1A; comparing images of 1 minute after injury); however, the second phase of platelet activation was significantly impaired, resulting in a prolonged time to occlusive thrombosis (Figure 1A-B; supplemental Video 2). Although large thrombi still formed in most α1^+/− mice, they failed to fully occlude the vessel within the 30-minute observation period. Histological analysis revealed the presence of gaps within thrombi from α1^+/− mice, which were absent in those from the α1^+/+ controls (supplemental Figure 1).

Figure 1.

a1 haploinsufficiency attenuates arterial thrombosis in male mice without affecting hemostasis. (A) Representative video images of thrombus formation in the carotid artery of male mice after 7.5% FeCl₃ treatment. Platelets were labeled via direct IV injection of rhodamine 6G. (B) Quantitative data showing the time to occlusive thrombus formation after injury. (C) Tail bleeding assay conducted by truncating the tail 1 cm from the tip. (D) aPTT assay using platelet-poor plasma. (E) Thrombus formation in the carotid artery of male mice induced by 10% FeCl₃. (F) Platelet α1 expression assessed by western blot. (G) Intracellular sodium concentration in platelets measured as an indicator of NKA pump function. (H) Male WT mice were lethally irradiated with X-RAD320 at a dose of 11 Gy. Five days later, mice received an infusion of α1^+/− or α1^+/+ platelets via jugular vein injection, followed by thrombosis induction using 7.5% FeCl₃ injury 10 minutes later. The log-rank (Mantel-Cox) test was used for statistical analyses in panels B-C,E,H. MW, molecular weight; plts, platelets.

α1 haploinsufficiency did not affect thrombosis in female mice (P = .96; Figure 1B; supplemental Figure 2). Consequently, in the following studies, only male mice were used except for investigating sex-associated thrombosis. Additionally, in a zebrafish model of estrogen-induced venous thrombosis, global knockdown of atp1a1 paralogs did not affect venous thrombosis (supplemental Figure 3). Therefore, this study focused solely on arterial thrombosis.

Despite the prolonged thrombosis time, the tail bleeding time and aPTT remained comparable between male α1^+/+ and α1^+/− mice (Figure 1C-D). Under more severe injury with 10% FeCl₃, the time to carotid artery occlusion was similar between the 2 genotypes (Figure 1E). These findings suggest that NKA α1 may play a mechanistic role in platelet activation and thrombosis, potentially influenced by sex hormones or sex-associated factors.

Atp1a1 haploinsufficiency significantly reduced platelet α1 expression (Figure 1F). However, this reduction did not affect intracellular sodium homeostasis or resting platelet membrane potential (Figure 1G; supplemental Figure 4), suggesting that the antithrombotic effect is independent of NKA’s ion-pumping function. To determine whether this phenotype was attributable specifically to reduced platelet α1 expression rather than α1 deficiency in other cell types, we performed a platelet transfusion assay.^36,43 As shown in Figure 1H, thrombocytopenic WT mice receiving α1^+/− platelets exhibited a significantly prolonged time to occlusive thrombosis compared to those receiving α1^+/+ platelets, indicating that platelet α1 is responsible for the antithrombotic phenotype observed in α1^+/− mice.

NKA α1 haploinsufficiency reduces low-dose ADP-induced platelet aggregation

α1^+/− platelets exhibited normal adhesion to collagen-coated surfaces (Figure 2A). Platelet aggregation in response to collagen was also comparable between α1^+/− and α1^+/+ platelets (Figure 2B-C; supplemental Figure 5A). Although 10μM ADP induced similar aggregation in both groups, 2.5μM ADP-induced aggregation was significantly reduced in α1^+/− platelets (Figure 2D-E), suggesting a potential interaction between α1 and the ADP receptors P2Y₁₂ or P2Y₁. In contrast, thrombin-, U-46619–, and AY-NH2–induced platelet aggregation remained unchanged (supplemental Figure 5B-D), indicating that α1 haploinsufficiency has only a moderate impact on platelet function.

Figure 2.

α1 haploinsufficiency inhibits ADP-induced platelet aggregation in vitro. (A) Whole blood from male α1^+/− or α1^+/+ mice was labeled with rhodamine 6G (50 mg/L), supplemented with CaCl₂/MgCl₂ to a final concentration of 1mM, and perfused through a collagen-coated (100 mg/L) Cellix flow chamber at a shear rate of 60 dyn/cm² for 3 minutes. The chambers were then washed with phosphate-buffered saline (PBS), and images were captured at predesignated marker positions. Fluorescence areas, representing platelet coverage, were quantified using ImageJ and analyzed by Student t test. (B-E) Whole blood from male α1^+/− or α1^+/+ mice was collected using 0.109M sodium citrate as an anticoagulant, and PRP was prepared by centrifugation. PRP was supplemented with CaCl₂/MgCl₂ to a final concentration of 1mM, and aggregation was initiated by adding collagen (panels B-C) or ADP (panels D-E). Statistical analysis was performed using Student t test.

NKA α1 forms a complex with P2Y₁₂, and α1 haploinsufficiency reduces ADP-stimulated AKT activation

To investigate the mechanism underlying α1-P2Y cross talk, we examined whether α1 forms a complex with P2Y in mouse platelets. Co-IP assay revealed that α1 binds P2Y₁₂ but not P2Y₁ (Figure 3A). Blue-native polyacrylamide gel electrophoresis assay confirmed this finding (Figure 3B). To determine whether α1-P2Y₁₂ binding is a general feature of these proteins, we transfected human P2Y₁₂ or adenosine A₂B receptor (A₂BR) into COS-7 cells, a green monkey kidney fibroblast-like cell line. Notably, green monkey α1 (XP_007975629.1) is fully conserved with human α1 (NP_000692.2). Both P2Y₁₂ and A₂BR are purinergic GPCR.⁵³ We found that P2Y₁₂ appeared in multiple forms, including monomers, dimers, and multimers. α1 preferentially bound to the P2Y₁₂ dimer over the monomer (Figure 3C). In contrast, α1 did not bind to A₂BR (Figure 3C), suggesting that α1-P2Y₁₂ binding is either P2Y₁₂ specific or dependent on a unique structural feature of these 2 proteins.

Figure 3.

α1 binds to P2Y₁₂, and α1 deficiency attenuates ADP-induced AKT activation in platelets. (A) WT platelets pooled from 6 male WT mice were lysed and subjected to co-IP to assess the interaction between NKA α1 and P2Y₁₂ or P2Y₁. (B) Platelet lysates from α1^+/− and α1^+/+ mice were analyzed using blue-native polyacrylamide gel electrophoresis (PAGE). The membrane was first probed for α1, then stripped and reprobed for P2Y₁₂. The image represents 2 independent experiments. (C) COS-7 cells were transfected with plasmids encoding human P2Y₁₂ (p-hP2Y₁₂; Addgene number 66471) or the A2B receptor (p-hA2BR; Addgene number 37202) for 36 hours. Cell lysates were then used for co-IP assays to examine α1 binding to P2Y₁₂ and A2B receptors. (D) PRP from α1^+/− and α1^+/+ mice was pooled from 5 males, adjusted to a final concentration of 2.5 × 10⁸ cells per mL using platelet-poor plasma, and divided into 4 aliquots. PRP was supplemented with MgCl₂/CaCl₂ (1mM final concentration) and stimulated with ADP (2.5μM final concentration) for different time points. The reaction was stopped using an EDTA/PGE1 cocktail, and platelets were lysed for western blot analysis. The blot represents 2 independent experiments. GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IB, immunoblotting; IgG, immunoglobulin G; IP, immunoprecipitation.

Both P2Y₁₂ and α1-mediated signaling activate the PI3K/AKT pathway.^38,54 To assess whether α1 affects P2Y₁₂ signaling, we treated α1^+/− and α1^+/+ platelets with ADP for varying durations and examined AKT phosphorylation. As shown in Figure 3D, although ADP activated AKT in both groups, the magnitude of AKT phosphorylation was significantly reduced in α1^+/− platelets. These data further suggest that, through regulating P2Y₁₂ signaling, NKA α1 modulates platelet function.

NKA α1 binds to LGL-containing platelet GPCRs

Sequence analysis revealed LGL motifs in several platelet GPCR, including P2Y₁₂ (115-117), PAR4 (194-196), PTGER3 (237-239), TP (76-78 and 163-165), and PI₂R (IP; 82-84 and 156-158; supplemental Figure 6). Co-IP assays confirmed that α1 binds to PAR4 and TP in human platelets (supplemental Figure 7), suggesting that α1 interacts with LGL-containing platelet GPCRs across species and may modulate GPCR-mediated platelet activation. Given the important role of P2Y₁₂ in platelet function and its clinical significance, we used it as a representative LGL-containing GPCR to investigate the pathophysiological relevance of this interaction.

To determine whether the LGL(I) motif in NKA α1 or P2Y₁₂ mediates their interaction, we mutated LGL(I) to SFT in both mouse α1 and P2Y₁₂. As shown in Figure 4A-D, mutation of the LGL motif in either α1 or P2Y₁₂ significantly reduced their interaction, with a stronger effect observed upon P2Y₁₂ mutation. Treatment with ouabain or digoxin also reduced α1-P2Y₁₂ binding, but it appeared to promote oligomer formation at specific molecular sizes (Figure 4D).

Figure 4.

LGL mediates the interaction between NKA α1 and P2Y₁₂. (A) COS-7 cells were transfected with plasmids encoding mouse WT P2Y₁₂ along with either WT or LGL→SFT mutant α1. (B-C) COS-7 cells were transfected with plasmids encoding mouse WT α1 along with either WT or LGL→SFT mutant P2Y₁₂, and cell lysates were used for co-IP assays (B); α1 band intensity was quantified and expressed as the ratio relative to the average of WT P2Y₁₂ cotransfected with WT α1 (C). (D) COS-7 cells were transfected with different plasmid combinations and treated with ouabain or digoxin for 36 hours before being lysed for co-IP assays. α1 band intensity was quantified with ImageJ and normalized to lane 1. (E) COS-7 cells were transfected with plasmids encoding human P2Y₁₂ for 24 hours, then treated with LGL peptide at the indicated dosages for 16 hours. Cells were harvested for co-IP analysis of α1 and P2Y₁₂. α1 band intensity was quantified with ImageJ and normalized to the non-LGL condition. GAPDH was blotted as a loading control for input. (F) Washed platelets from male α1^+/− or α1^+/+ mice were resuspended in PBS at 2.5 × 10⁸/mL, and 400 μL was used per aggregation assay. LGL peptide or a leucine-glycine mixture (control) was added to a final concentration of 50μM and incubated for 3 minutes before initiating aggregation. Data were expressed as (LGL – control)/control. An unpaired t test was used. IgG, immunoglobulin G; SB, 2× Laemmli sample buffer.

To determine whether an LGL peptide could act as a decoy, we tested its ability to disrupt α1-LGL GPCR complex formation. COS-7 cells were transfected with human P2Y₁₂ for 24 hours, then treated with LGL peptide (0, 10, and 100 μg/mL) for 16 hours in a complete culture medium. Co-IP assays showed that LGL treatment dramatically reduced α1-P2Y₁₂ binding (Figure 4E). Notably, LGL peptide treatment did not lead to cytotoxicity and also did not affect NKA α1 expression (supplemental Figure 8).

We then examined the effect of the LGL peptide on ADP-induced platelet aggregation using washed platelets. The minimal effective dose of LGL for inhibiting 2.5μM ADP-induced aggregation was 50μM (data not shown). As shown in Figure 4F, LGL reduced aggregation in α1^+/+ platelets by ∼15% compared to controls, in which platelets from the same mouse were treated with an equimolar mixture of leucine (33.3μM) and glycine (16.7μM). In contrast, the inhibitory effect was more pronounced in α1^+/− platelets, with an average reduction of 34%.

Therapeutic digoxin enhances α1-P2Y₁₂ dimer and oligomer formation

To investigate the effects of CTS on platelet function, we compared the activity of platelets from patients with HF on digoxin with those from healthy donors. ADP-induced platelet aggregation was not enhanced in patients with HF receiving digoxin; instead, there was a trend toward reduced aggregation (Figure 5A).

Figure 5.

Therapeutic digoxin enhances α1-P2Y12 dimerization while inhibiting SFK activation. Whole blood was drawn from healthy donors (control), patients with HF, and patients with HF on digoxin (Di.; HF + Di.). Platelets were isolated by centrifugation. (A) PRP was assessed for platelet aggregation in response to 10μM ADP. (B-C) Washed platelets were lysed in RIPA buffer, protein concentration was adjusted to 2 μg/μL, and samples were mixed with an equal volume of 2× Laemmli sample buffer in the presence (B) or absence of 2-mercaptoethanol (C) and incubated at 37°C for 30 minutes. Forty micrograms of total protein were analyzed by western blot. (D) Samples from panel B were used for western blot analysis of SFK activation, detected using a phospho-Src antibody. The membrane was stripped and reblotted with a total Src antibody as a loading control. Cont., control; D, dimer; M, monomer; MW, molecular weight; O, oligomer.

We then examined α1 and P2Y₁₂ expression in platelets from patients with HF, with and without digoxin treatment. Under reducing conditions, both α1 and dimeric P2Y₁₂ were decreased in HF platelets. However, in patients with HF receiving digoxin, P2Y₁₂ dimer and oligomer formation were markedly increased under both reducing and nonreducing conditions (Figure 5B-C). Additionally, in nonreducing conditions, a band >250 kDa was detected in the α1 blot, which likely represents tetrameric α1 or α1β3 diprotomer,⁵⁵ or other α1-associated complexes. This high–molecular weight complex was increased in HF platelets and was further enhanced by digoxin treatment.

To further assess platelet activity, we measured the activity of Src-family kinases (SFK), which are critical regulators of platelet activation.⁵⁶ SFK phosphorylation was significantly increased in HF platelets but not in controls (Figure 5D). Notably, digoxin treatment markedly reduced SFK activation, suggesting a potential modulatory role of digoxin in platelet signaling.

Platelet α1 expression correlates inversely with plasma androgen levels

To clarify the mechanism underlying the sex difference in thrombosis in α1^+/− mice, we analyzed our thrombosis data pool of WT C57BL/6J mice and randomly extracted 10 males and 10 females. Female mice exhibited a shorter time to occlusive thrombus formation (Figure 6A) but a higher platelet α1 expression than males (Figure 6B; supplemental Figure 9A). α1 expression was the same in other organs examined, including hearts, white blood cells, and red blood cells (supplemental Figure 9B-D). We then performed gonadectomy in α1^+/+ and α1^+/− mice. Although orchiectomy (ORX) nearly depleted plasma androgens in males, ovariectomy only reduced plasma 17-β estradiol levels to ∼70% at 2 weeks after surgery (Figure 6C-D).

Figure 6.

Androgens suppress platelet α1 expression and exert antithrombotic effects. (A) WT male and female mice were subjected to the 7.5% FeCl₃–induced thrombosis model. (B) Platelet α1 expression was assessed in WT male and female mice. (C-D) WT female and male mice underwent ovariectomy (OVX; C) or ORX (D), and plasma 17-β estradiol and testosterone levels were measured 2 weeks after surgery. (E-F) Female and male α1^+/− and α1^+/+ mice underwent OVX (E) or ORX (F), respectively. Eight weeks later, they were subjected to the 7.5% FeCl₃–induced thrombosis model, and thrombosis time was compared with nongonadectomized controls. (G-H) Whole blood was collected from male α1^+/+ and α1^+/− mice with or without ORX to measure plasma testosterone (G) and platelet α1 expression (H). F, female; M, male; MW, molecular weight.

To allow for a more complete hormonal adjustment, we extended the postsurgery period to 8 weeks before conducting thrombosis assays. Ovariectomy did not affect thrombosis in either genotype (Figure 6E). Unexpectedly, ORX significantly enhanced thrombosis in α1^+/− mice, effectively abolishing their antithrombotic phenotype (Figure 6F). Notably, plasma 17-β estradiol and testosterone levels remained stable between 2 and 8 weeks after gonadectomy (data not shown). Interestingly, α1^+/− mice had significantly higher plasma testosterone levels than α1^+/+ mice (Figure 6G). ORX eliminated this difference, along with the disparity in platelet α1 expression (Figure 6H), suggesting that androgen regulates the platelet α1 expression.

α1 haploinsufficiency enhances the antithrombotic efficacy of P2Y₁₂ antagonists, and low-dose CTS suppresses both platelet aggregation and thrombosis

Treatment of WT mice with 1 mg/kg clopidogrel has previously shown inconsistent effects on thrombosis inhibition.³⁸ To investigate whether α1 haploinsufficiency affects the efficacy of P2Y₁₂ antagonists, we administered 0.5 or 1 mg/kg clopidogrel via gavage for 1 week to α1^+/+ and α1^+/− mice, followed by thrombosis assessment using the 10% FeCl₃–induced carotid artery injury model. As shown in Figure 7A, 0.5 mg/kg clopidogrel did not significantly prolong the time to thrombosis in α1^+/− mice. However, 1 mg/kg clopidogrel significantly delayed occlusive thrombus formation. Similarly, 0.33 mg/kg prasugrel showed an antithrombotic effect in α1^+/− mice but had no effect in α1^+/+ mice (Figure 7B). Along with the observation that the LGL peptide more effectively inhibits ADP-induced aggregation of α1^+/− platelets (Figure 4F), these data suggest that α1 haploinsufficiency lowers the effective dose of P2Y₁₂ antagonists required to inhibit thrombosis.

Figure 7.

α1 haploinsufficiency reduces the effective clopidogrel dose, and a low-dose CTS inhibits thrombosis and ADP-induced platelet aggregation. (A) Male α1^+/− and α1^+/+ mice were gavage-fed clopidogrel (0.5 or 1 mg/kg) for 1 week, followed by the 10% FeCl₃–induced thrombosis model. (B) Male α1^+/− and α1^+/+ mice were gavage-fed prasugrel (0.33 mg/kg) for 1 week, followed by the 10% FeCl₃–induced thrombosis model. (C) Male WT mice received a single bolus intraperitoneal injection of ouabain (100 μg/kg) or MBG (100 μg/kg) and were subjected to the 7.5% FeCl₃–induced thrombosis model 18 hours later. (D) Human PRP was pretreated with ouabain at the indicated concentrations for 5 minutes, and aggregation was induced with 2.5 μM ADP. The log-rank test was used for statistical analyses in panels A-B. Clop., clopidogrel; Cont., control.

To determine whether CTS treatment mimics α1^+/− mice, we administered ouabain (100 μg/kg) via intraperitoneal injection in WT mice and assessed thrombosis 18 hours later. As shown in Figure 7C, a single dose of ouabain significantly inhibited thrombosis (P = .005 vs control), aligning with findings in α1^+/− mice.

MBG, a bufadienolide CTS, also inhibits NKA function and shares a conserved structural core with ouabain and digoxin.^57,58 A previous study demonstrated that intraperitoneal MBG (560 μg/kg) significantly attenuated zymosan-induced inflammation.⁵⁰ To assess its antithrombotic effects, we administered a low dose of MBG (100 μg/kg) overnight. As shown in Figure 7C (blue line), MBG also significantly inhibited thrombosis (P = .003 vs control), further supporting that low-dose CTS suppresses thrombosis.

Because rodent α1 is ∼100- to 1000-fold less sensitive to ouabain than human α1,¹⁵ we examined whether this species difference influences our findings. Human PRP from healthy donors (aged 18-50 years) was pretreated with ouabain for 5 minutes before aggregation was initiated with 2.5μM ADP. As shown in Figure 7D, ouabain dose dependently inhibited ADP-induced platelet aggregation in 2 of 6 donors (33.3%), whereas no effect was observed in the remaining 4 (66.7%; 1 man and 3 women). Notably, the 2 ouabain-sensitive individuals were men who exercised regularly and did not use anabolic steroids. This variability suggests that individual androgen levels may influence platelet responsiveness to CTS. Together, these data indicate that NKA α1 contributes to platelet activation and that its function may be modulated by plasma androgen levels, providing a mechanistic link between α1 and the antithrombotic effects of low-dose CTS.

Discussion

This study represents, to our knowledge, the first systematic investigation of NKA’s role in thrombosis, using genetically modified mice, zebrafish, and multiple CTSs in both in vivo and in vitro models. We identify novel mechanisms regulating platelet activation and thrombosis. Key findings include the following: (1) α1 haploinsufficiency exerts a significant antithrombotic effect in male mice without disturbing hemostasis; (2) α1 forms complexes with P2Y₁₂, PAR4, and TP, all receptors containing LGL motifs, and its haploinsufficiency reduces ADP-stimulated AKT activation and platelet aggregation; (3) α1 haploinsufficiency enhances the efficacy of P2Y₁₂ antagonist in vivo; (4) low-dose CTS inhibit thrombosis in mice and attenuate ADP-induced platelet aggregation in some human donors; and (5) androgen levels inversely correlate with platelet α1 expression.

Human platelets contain ∼2000 copies of NKA α1 subunit, 500 copies of α4, and 2500 copies of β3.²⁰ The α4 subunit, exclusive to sperm and 78% identical to α1,⁵⁹ was undetected in previous human and mouse platelet proteomics studies.^21,22 The ouabain-sensitive α2 and α3 isoforms were also absent in platelets.20, 21, 22 Therefore, the predominant platelet NKA isoform is presumed to be α1β3. Additionally, attempts to generate PF4-Cre⁺/α1^f/f mice were unsuccessful (data not shown), suggesting that complete α1 deletion in megakaryocytes or platelets may be embryonically lethal. These findings support the conclusion that α1 is the sole NKA α isoform in platelets.

The role of CTS on human platelet activation is under debate.²³ Studies have suggested that ouabain potentiates platelet aggregation by increasing intracellular Ca²⁺ levels, which promotes phosphatidylserine exposure.^24,25,30 However, these studies used supraphysiological ouabain concentrations (4-200 μM),^24,25,32 far exceeding its nanomolar 50% inhibitory concentration in human cells.⁶⁰ In contrast, our findings demonstrate that low-dose ouabain treatment (100 μg/kg; intraperitoneal) significantly inhibits thrombosis in vivo, emphasizing the need to consider dose-dependent effects. Given that ∼60% of a mouse’s body weight is water, a dose of 100 μg/kg ouabain is estimated to yield a plasma concentration of ∼285nM (416nM for MBG), assuming uniform distribution after injection. With a plasma half-life of 18 to 22 hours for ouabain,⁶¹ mice may have blood ouabain levels of ∼100 to 150nM at the time of the in vivo thrombosis assay.

Additionally, we observed that ouabain inhibits ADP-induced aggregation in only a subset of human donors, suggesting variability in platelet sensitivity to CTS, possibly due to differences in α1 expression, GPCR interactions, or genetic polymorphisms. Further studies should explore whether patient-specific factors influence CTS responsiveness, particularly in individuals with cardiovascular disease.

Sex differences in α1 expression may contribute to thrombosis risk, because female mice exhibited higher platelet α1 levels and a more prothrombotic phenotype than males (Figure 6A-B; supplemental Figure 9A). This aligns with prior findings that the expression of renal transport proteins, including NKA α1, is higher in animals with a female gonadal phenotype or sex chromosome complement.⁶² The loss of the antithrombotic effect in α1^+/− males after ORX, which unexpectedly have higher testosterone levels, suggests that androgens regulate α1 expression. This novel finding may explain clinical observations linking male hypogonadism to cardiovascular diseases.⁶³ Because testosterone and estrogen contain a steroidal structure similar to CTS, they may act as “endogenous CTS.” Further studies should investigate whether testosterone and estrogen differentially modulate α1 expression and thrombosis susceptibility, particularly in the context of HF, postmenopausal thrombosis, and hormone replacement therapy.

Global knockdown of atp1a1 paralogs in zebrafish did not affect venous thrombosis, and aPTT was not affected by α1 haplodeficiency, suggesting that the intrinsic and common coagulation pathways remain normal in α1^+/− mice. Transfusion of α1^+/− platelets prolonged the time to occlusive thrombus formation, suggesting a predominant role of platelet α1 in arterial thrombosis. Accumulating evidence indicates that GPCRs form homodimers or heterodimers, as well as higher-order oligomers, with dimerization/oligomerization being essential for signaling initiation.64, 65, 66, 67 Our study identifies a novel mechanism of platelet activation via NKA α1–GPCR interaction, mediated by the LGL motif, which facilitates the complex formation with platelet receptors, such as P2Y₁₂. These interactions may involve a 2-domain mechanism, because the LGL motif is present in both binding partners. LGL(I)→SFT mutation in α1 and P2Y₁₂ significantly attenuates their binding, and treatment of platelets with LGL peptide attenuated platelet aggregation, with a pronounced effect in α1^+/− platelets, suggesting that targeting α1-GPCR interaction could represent a novel antithrombotic strategy.

Clinically, ouabain and digoxin have been used to treat HF and atrial fibrillation. Digoxin has been linked to increased platelet and endothelial cell activation in patients with atrial fibrillation.³¹ In our study, we found that therapeutic doses of digoxin enhanced the formation of dimers and oligomers involving α1 and P2Y₁₂ (may be PAR4 and TP too), possibly promoting platelet activation. In HF platelets, a 75- to 80-kDa band was detected by an anti-α1 antibody (Figure 5B), which was absent in healthy donors and those receiving digoxin. A recent study indicated that ouabain binding to α1 enhances α1 trypsinolysis.⁶⁸ Increased endogenous digitalis-like compounds have been reported in patients with HF, with circulating levels ranging from 1nM to 9nM.^69,70 However, some groups have challenged this concept, asserting that “no endogenous ouabain is detectable in human plasma.”⁷¹ Nonetheless, the presence of endogenous ouabain or other factors may partially explain our findings, because low-dose ouabain can activate c-Src,⁷² and digoxin-ouabain antagonism can diminish c-Src activation.⁷³ Whether sex hormones could serve as “endogenous CTS” and whether the 75- to 80-kDa peptide found in HF platelets results from α1 trypsinolysis and affects platelet function warrant further investigation. Additionally, the lack of clarification regarding the paradoxical discrepancy between the increased aggregation of P2Y₁₂ receptors in platelets from patients with HF treated with digoxin and their reduced platelet aggregation, although not statistically significant, represents a limitation of this study that warrants further investigation. These results may shed light on the mechanism underlying digoxin-associated thrombosis in HF.

Current antiplatelet therapies require chronic administration to maintain efficacy and, targeting major platelet activators carries significant bleeding risks. Because α1 haploinsufficiency does not impair hemostasis and has no effect on thrombosis under more severe injury (Figure 1E), and because low-dose CTS significantly inhibit thrombosis in vivo, α1 could serve as a target for a new generation of dual antiplatelet therapy.⁷⁴ This is further supported by our observation that the effectiveness of P2Y₁₂ antagonists was enhanced in α1^+/− mice. Because cells require only 20% to 30% of maximal NKA expression to maintain ion homeostasis,^75,76 partial α1 inhibition could offer a safe and effective antithrombotic strategy.

In summary, this study provides, to our knowledge, the first comprehensive analysis of NKA α1 in platelet activation and thrombosis, identifying a novel α1-GPCR interaction mediated by the LGL motif. We demonstrate that α1 deficiency enhances antithrombotic efficacy without impairing hemostasis and that low-dose CTS exerts antithrombotic effects in both mice and certain human donors. These findings lay the groundwork for α1-targeted therapies as a novel antiplatelet and antithrombotic strategy, with potential translational applications in cardiovascular disease.

Conflict-of-interest disclosure: The authors declare no competing financial interests.