Synergistic inhibitory effects of clopidogrel and rivaroxaban on platelet function and platelet‐dependent thrombin generation in cats

Abstract Background Dual antithrombotic treatment (DAT) with clopidogrel and rivaroxaban sometimes is prescribed to cats with hypertrophic cardiomyopathy at risk of thromboembolism. To date, no studies have evaluated their combined effects on platelet function. Objectives/Hypothesis Evaluate the safety of DAT in healthy cats and compare, ex vivo, platelet‐dependent thrombin generation and agonist‐induced platelet activation and aggregation in cats treated with clopidogrel, rivaroxaban, or DAT. We hypothesized that DAT would safely modulate agonist‐induced platelet activation and aggregation more effectively than single agent treatment. Animals Nine apparently healthy 1‐year‐old cats selected from a research colony. Methods Unblinded, nonrandomized ex vivo cross‐over study. All cats received 7 days of rivaroxaban (0.6 ± 0.1 mg/kg PO), clopidogrel (4.7 ± 0.8 mg/kg PO), or DAT with defined washout periods between treatments. Before and after each treatment, adenosine diphosphate (ADP)‐ and thrombin‐induced platelet P‐selectin expression was evaluated using flow cytometry to assess platelet activation. Platelet‐dependent thrombin generation was measured by fluorescence assay. Platelet aggregation was assessed using whole blood impedance platelet aggregometry. Results No cats exhibited adverse effects. Of the 3 treatments, only DAT significantly decreased the number of activated platelets (P = .002), modulated platelet activation in response to thrombin (P = .01), dampened thrombin generation potential (P = .01), and delayed maximum reaction velocity (P = .004) in thrombin generation. Like clopidogrel, DAT inhibited ADP‐mediated platelet aggregation. However, rivaroxaban alone resulted in increased aggregation and activation in response to ADP. Conclusion and Clinical Importance Treatment combining clopidogrel and rivaroxaban (DAT) safely decreases platelet activation, platelet response to agonists, and thrombin generation in feline platelets more effectively than monotherapy with either clopidogrel or rivaroxaban.

thrombin generation was measured by fluorescence assay. Platelet aggregation was assessed using whole blood impedance platelet aggregometry.
Results: No cats exhibited adverse effects. Of the 3 treatments, only DAT significantly decreased the number of activated platelets (P = .002), modulated platelet activation in response to thrombin (P = .01), dampened thrombin generation potential (P = .01), and delayed maximum reaction velocity (P = .004) in thrombin generation. Like clopidogrel, DAT inhibited ADP-mediated platelet aggregation. However, rivaroxaban alone resulted in increased aggregation and activation in response to ADP.
Conclusion and Clinical Importance: Treatment combining clopidogrel and rivaroxaban (DAT) safely decreases platelet activation, platelet response to agonists, and thrombin generation in feline platelets more effectively than monotherapy with either clopidogrel or rivaroxaban.

| INTRODUCTION
Hypertrophic cardiomyopathy (HCM), characterized by concentric left ventricular hypertrophy and diastolic dysfunction, is the most common cardiac disease in cats and affects approximately 15% of the general feline population. 1,2 Cardiogenic arterial thromboembolism (CATE), a devastating complication that occurs in 6%-17% of cats with underlying cardiomyopathies, often carries a poor prognosis with a high mortality rate of up to 67%. 3 It occurs when an intracardiac thrombus, most commonly formed in the left auricle, embolizes to an artery and impedes blood flow, leading to tissue ischemia and ischemic reperfusion injury. 4 In 90% of cases, CATE results in acute distal aortic occlusion, also known as a saddle thrombus. 4 For that reason, CATE is a very distressing emergency for owners and veterinarians because of the sudden occurrence of extreme pain, hindlimb or forelimb paralysis and often congestive heart failure without any forewarning. Despite a high mortality rate, thromboprophylaxis to prevent intracardiac thrombosis and recurrence of CATE is limited, and further research in this area is crucial. 4,5 Cats with HCM are hypercoagulable because of a pathologically enlarged left atrium promoting blood stasis and endothelial injury. 6,7 This situation places cats with HCM at increased risk of CATE, with up to 11.3% of cats developing CATE within 10 years of HCM diagnosis. 8,9 Other outcomes of HCM in cats include a persistent subclinical state, left-sided congestive heart failure (CHF), and sudden cardiac death. 8 Although HCM affects domestic cats of any breed, sex, and age >3 months, purebred Maine Coon and Ragdoll cats homozygous for breed-specific myosin binding protein-C gene (MYBPC3) mutations develop severe HCM and are at higher risk for CATE. 10 Although there is no universal standard of care exists for CATE, current guidelines recommend the use of clopidogrel, an antiplatelet drug that irreversibly inhibits the platelet adenosine diphosphate (ADP) receptor, P2Y 12 , to prevent CATE. 11,12 Previous studies showed that clopidogrel is well-tolerated in cats and is superior to aspirin at decreasing recurrence rate and prolonging the time to recurrence of CATE. 13 However, on-treatment recurrence rate remained high at 36%, indicating that clopidogrel alone is not effective at preventing CATE. 13 A plausible explanation for this finding is resistance to clopidogrel, reported in up to 15.4% of cats with HCM because of genetic polymorphism in the ADP receptor gene P2RY1. 11,14 Because of the variable response to clopidogrel and high mortality rate of CATE, a multimodal approach in thromboprophylaxis currently is employed in cats at risk of thrombosis at our institution. Administration of anticoagulant treatment including unfractionated heparin and low molecular weight heparin has been described in cats, but the required frequent SC administration of up to q6h compromises owner and patient compliance. 15 Rivaroxaban is a direct PO anticoagulant drug that prevents clot formation by inhibiting activated factor X (factor Xa). 16 In healthy adult cats, once daily administration of rivaroxaban is safe, well-tolerated and effective at prolonging clotting times. 17 In humans, dual treatment consisting of rivaroxaban and clopidogrel decreases the incidence of thrombosis in patients with coronary artery disease and is superior to single-agent treatment in preventing acute cardiovascular events. 18,19 However, their synergistic effects on platelet function are poorly understood. In an in vitro study in humans, ticagrelor, a P2Y 12 inhibitor similar to clopidogrel, enhanced the inhibitory effect of rivaroxaban on plateletdependent thrombin generation. 20 A retrospective study in cats found favorable outcomes in those receiving dual antithrombotic treatment (DAT). The study found no first-time occurrence of CATE after commencement of DAT and clinically unimportant bleeding adverse events in 15% of cats. 20 However, little is known about the synergistic inhibitory effects of clopidogrel and rivaroxaban on platelets in cats.
Because rivaroxaban may not only improve owner compliance but also enhance the antiplatelet effects of clopidogrel, we sought to evaluate the safety and efficacy of DAT in healthy cats. We hypothesized that DAT with clopidogrel and rivaroxaban would safely modulate agonist-induced platelet activation and aggregation more effectively than single agent treatment. To test our hypothesis, we aimed to (a) examine the safety of combining standard dose PO rivaroxaban with clopidogrel in cats and (b) compare, ex vivo, plateletdependent thrombin generation and agonist-induced platelet activation and aggregation in cats treated with clopidogrel alone, rivaroxaban alone or DAT (clopidogrel and rivaroxaban).

| Subject selection
The study protocol was approved by the Institutional Animal Care and Use Committee of the University of California, Davis (IACUC protocol number 20565). Nine healthy cats (6 intact females and 3 intact males, 1 year of age) were selected from a colony of Maine Coon/outbred mixed domestic cats bred and raised at the UC Davis Feline HCM Research Laboratory. Of the 9 cats, 6 heterozygous and 3 wild-type cats for the A31P variant in the myosin binding protein C gene (MYBPC3) gene were included. Cats were housed in group housing kennels and cared for by dedicated animal care technicians. Environmental enrichment was provided in the form of toys, boxes, shelves, and social interaction with cats and humans. Physical examination, CBC, serum biochemistry panel, and echocardiography performed by a board-certified cardiologist (JAS) were performed to ensure all cats were clinically healthy before enrollment.
Cats were monitored twice daily for adverse drug reactions including gastrointestinal signs and bleeding diathesis.

| Drug administration
We utilized an unblinded, nonrandomized ex vivo cross-over design. All cats received 7 days of rivaroxaban (2.5 mg PO q24h), clopidogrel (18.75 mg PO q24h), or DAT ( Figure 1). The 3-day and 15-day washout periods after rivaroxaban and clopidogrel treatments, respectively, have proven to be sufficient for coagulation and platelet function to return to baseline. 17,21 Data measured during the study confirmed rivaroxaban assay and platelet aggregometry results had returned to baseline before advancing to the next treatment period. A sole researcher (STL) administered all medications to all cats to minimize variation. Citrated plasma samples, collected 3 hours after administration of rivaroxaban, were sent to the Cornell University Animal Health Diagnostic Center for measurements of plasma rivaroxaban concentration based on factor Xa inhibition.
Cats did not receive any other medications for the duration of the study. day. Blood samples on day 7 of the rivaroxaban and DAT trials were collected 3 hours after rivaroxaban administration to capture peak anti-Xa activity. 17 Blood samples on day 7 of clopidogrel treatment were collected 1.5-2 hours after clopidogrel administration to capture maximum clopidogrel active metabolite (CAM) plasma concentration. 22 Complete blood counts performed using citrated blood and an automatic analyzer (HM5 Hematology Analyzer, Abaxis, Abbott Group, Union City, CA). Heparinized whole blood was utilized once after the first blood collection for a biochemistry panel (VetScan V2 Chemistry Analyzer, Abaxis, Abbott Group, Union City, CA).

| Whole blood impedance platelet aggregometry
Platelet aggregation was assessed using an automated whole blood impedance platelet aggregometer (Multiplate, Roche, Mannheim, Germany) according to manufacturer's instructions. 23,24 In brief, after F I G U R E 1 Flow diagram illustrating drug administration and blood sampling timeline. Clop, Clopidogrel; DAT, dual antithrombotic treatment; Riva, Rivaroxaban. 30 minutes of resting at room temperature, 300 μL of heparinized whole blood was added to pre-heated test cells containing prewarmed 0.9% sodium chloride solution. After 3 minutes of incubation at 37 C under physiologic shear rate created by an 800-rpm-spinning Teflon-coated magnetic stir bar, ADP (final concentration, 6.25 μM; MilliporeSigma, Burlington, MA) was added and electrical impedance was recorded for 6 minutes. Platelet aggregation, measured as electrical impedance over time on a pair of silver-coated electrodes, was reported as velocity (AU/min), area under the curve (AUC; AU Â min), and maximum aggregation (Ag max). Response to treatment was calculated based on the formula: MFI fold change log 10 ð Þ¼ log 10 MFI Activated ð Þ À log 10 MFI Resting À Á 2.6 | Platelet-dependent thrombin generation Citrated PRP was further analyzed to determine thrombin generation over time using a commercially available fluorescence assay (Technothrombin TGA, Diapharma, Westchester, OH) according to the manufacturer's instructions. Before analysis, all reagents were allowed to reach room temperature. Platelet rich plasma first was standardized to 1.5 Â 10 8 cells/mL with autologous platelet poor plasma, generated as described above, within 2 hours after collection. Thrombin generation was measured as time (seconds) to maximum thrombin generation (T max ), peak thrombin generation (relative fluorescence unit; RFU), maximum velocity (V max , FI/s) and thrombin generation potential (AUC) over 1 hour.

| Rivaroxaban assay
Plasma rivaroxaban concentrations were measured before and after 7 days of rivaroxaban and DAT. Platelet-poor plasma was generated by centrifuging the remaining citrated blood after extracting PRP at 5000 Â g for 10 minutes, aliquoted and stored at À80 C. Batched samples were analyzed using the Comparative Coagulation Laboratory at the Cornell University Animal Health Diagnostic Center for measurement of anti-Xa activity chromogenic assay with a rivaroxaban calibration standard as described. 26

| Statistical analysis
Preliminary data based on a previous study was used to estimate the number of cats needed to achieve an approximately 33% decrease in platelet aggregation in response to clopidogrel treatment. 14 With a 2-tailed design, 9 cats were needed to detect significant changes with 80% power and a priori alpha of 0.05. Each cat served as its own control in order to determine the response to each of the 3 treatments.

| Platelet activation by flow cytometry
Platelet activation in the presence or absence of ex vivo treatment of ADP or thrombin was characterized using platelet surface P-selectin expression by flow cytometry (Figure 3). Baseline P-selectin expressions measured on D0 were not significantly different (P = .3) before each treatment to ensure that all measurements returned to baseline before the next treatment. We found that in unstimulated platelets (resting), DAT resulted in a significantly lower level of P-selectinpositive platelets (D0 = 44.0% ± 19.6 vs D7 = 16.7% ± 12.8; F I G U R E 2 Scatter dot plots demonstrating platelet inhibition, measured as percent (%) inhibition, in 9 cats after 7 days of rivaroxaban (RIVA), clopidogrel, or dual antithrombotic treatment (DAT) with rivaroxaban and clopidogrel. ADP-induced whole blood impedance aggregometry was measured as maximum aggregation (Ag max), velocity and area under the cure (AUC). Percent inhibition >0% indicates platelet inhibition (up arrows) and inhibition (%) < 0% indicates platelet activation (down arrows) after 7 days of treatment. Clopidogrel alone and DAT resulted in significant inhibition compared to Riva alone. Riva alone resulted in platelet activation in 7 of the 9 cats. Line represents median and error bars represent interquartile range. *P < .05; **P < .005; ***P < .001.
T A B L E 1 Summary of ADP-mediated whole blood platelet aggregometry in 9 cats receiving 7 days of rivaroxaban, clopidogrel or dual agent therapy. When platelets were treated with 20 μM ADP, CLOP or DAT significantly modulated response to ADP, while 7 days of RIVA treatment resulted in an increase response to ADP, which were significantly higher on D7 compared to CLOP or DAT. (C) Only DAT modulated platelet response to 0.005 U/mL thrombin, as shown by decreased percentage of P-selectin-positive platelets, and MFI FC from D0 to D7. Response to thrombin was significantly higher on D7 in cats treated with RIVA alone. *P < .05; **P < .005; ***P < .001; ****P < .0005. P = .002; Figure 3A). When platelets were activated with ADP, rivaroxaban treatment alone led to an increased response to ADP ( Figure 3B; MFI fold change log 10 D0 = 0.11 ± 0.096 vs D7 = 0.33 ± 0.20; P = .003). This increased response to ADP after 7 days of rivaroxaban was significantly higher compared to the clopidogrel and DAT treated cats on D7 (P < .01; Figure 3B)  Figure 3C).

| Platelet-dependent thrombin generation
Representative thrombograms of PRP in a cat receiving rivaroxaban and DAT are shown in Figure 4A,

| Correlations between plasma anti-Xa levels and platelet-dependent thrombin generation
A strong and significant negative correlation was found between plasma rivaroxaban concentrations and change (%) in V max in thrombin generation before and after rivaroxaban treatment. After 7 days of DAT, both peak thrombin generation and V max were strongly and negatively correlated with plasma rivaroxaban concentrations. Data are summarized in Table 2.

| DISCUSSION
We found a synergistic inhibitory effect of clopidogrel and rivaroxaban on platelet activation and thrombin generation potential in feline platelets. Specifically, DAT with clopidogrel and rivaroxaban safely Platelet activation, in the presence or absence of physiological agonists, was measured by upregulation of platelet surface P-selectin, a marker of outside-in signaling that leads to alpha granule secretion.
In addition, tissue factor-induced thrombin generation was measured in PRP to evaluate the pharmacodynamic effects of DAT on the link between primary and secondary hemostasis. Our data indicate that DAT in cats not only decreased the number of activated platelets but also modulated their thrombin generation potential. A plausible mechanism of these findings is that rivaroxaban, a direct factor Xa inhibitor, modulates the activation of thrombin during the initial phase of coagulation when tissue factor and activated factor VII complex activate factors X and V. Evidence in human platelets suggests that upon activation, platelets undergo de novo synthesis of tissue factor leading to its subsequent externalization and interaction with extracellular factor VIIa. 27  Although increased platelet activation previously has been observed in cats with HCM and other cardiomyopathies, its clinical and prognostic importance is unclear. In human beings, a persistent increase in activated platelets is associated with acute myocardial infarction and F I G U R E 5 Scatter dot plots demonstrating platelet dependent thrombin generation over time in 9 cats, measured as (A) time (seconds) to maximum thrombin generation (T max ), (B) peak thrombin generation (Relative Fluorescence Unit [RFU]), (C) thrombin generation potential (Area Under the Curve [AUC]) over 1 hour and (D) maximum velocity (V max , unit/s). After 7 days of treatment with rivaroxaban (RIVA), T max and peak thrombin generation were significantly modulated from D0 to D7. In addition to T max and peak thrombin generation, dual agent treatment (DAT) modulated thrombin generation potential and maximum velocity of thrombin generation from D0 to D7. Compared to RIVA, DAT also significantly prolonged T max compared to clopidogrel (CLOP) alone on D7. *P < .05; **P < .005; ***P < .001.
T A B L E 2 Correlation between plasma rivaroxaban concentrations and changes in platelet dependent thrombin generation. Upon treatment of PRP with supraphysiologic amounts of tissue factor, DAT was more effective at suppressing thrombin generation than rivaroxaban treatment alone. These findings are in contrast to studies in humans that showed that rivaroxaban administration modulates thrombin generation potential in tissue factor-treated plasma. 29 This may be a result of assay and species differences. Thrombin propagation by activated platelets is partially dependent on their reactivity to physiologic agonists. 32 Consistent with our thrombin generation data, platelet reactivity to ADP and thrombin was significantly decreased after DAT. As expected and based on its mechanism of action, clopidogrel decreased reactivity to ADP but did not elicit a consistent modulatory effect on thrombin-mediated platelet activation in all cats. This finding is important because thrombin is the most potent physiologic agonist that elicits strong outside-in signaling, leading to marked platelet activation, calcium mobilization, integrin activation and platelet aggregation via the protease activating receptor (PAR) and co-receptor of GP1bα. 32 36 A plausible explanation is that in vitro platelet activation, which commonly occurs in feline platelets, could dampen their reactivity to ADP before treatment. 37 Rivaroxaban treatment may therefore lower pre-activation of platelets at rest, resulting in potentiated activation and aggregation in the presence of physiologic agonists in vitro. However, because no significant differences in P-selectin positive platelets were found before and after rivaroxaban treatment, an underlying platelet priming mechanism mediated by rivaroxaban causing unrestrained platelet aggregation and activation also is possible. The mechanism by which rivaroxaban primes feline platelets to ADP is unclear and further research is warranted.
Our study had several limitations. First, the small sample size and 7-day duration of each treatment may limit the clinical recognition of adverse events that are less frequent or take longer to be manifested.
Second, ADP was the only agonist used for the platelet aggregation assay. Utilization of different agonists such as arachidonic acid, tissue factor or collagen may have identified differences among treatments on platelet aggregation. Third, predetermined doses were given without dose adjustments for weight. Although this design accurately represents clinical usage of these drugs, the variable dosages may have resulted in large SDs in drug effect on platelet function. However, this concern may not be an important limitation because many cats were of similar body weight. Finally, the cats included in ours study were all healthy and selected from a research colony of related cats. This feature may affect our ability to extrapolate the physiological response to DAT to the general population of cats, and in hypercoagulable cats with HCM. Future prospective clinical trials examining dual clopidogrel and rivaroxaban treatment in a large heterogeneous population of cats with CATE or HCM are needed.
Overall, we found that DAT with clopidogrel and rivaroxaban lowered the number of activated circulating platelets, dampened platelet response to thrombin, and suppressed tissue-factor-induced thrombin generation in healthy cats more effectively than monotherapy with clopidogrel or rivaroxaban alone. The potentiated response to ADP in platelets secondary to rivaroxaban requires further investigation.
Because cats with severe HCM have increased platelet activation and increased platelet priming, randomized clinical trials are needed to determine if DAT is superior to monotherapy in preventing intracardiac thrombosis and CATE in cats with HCM. 11 Additionally, we did not characterize the ADP receptor genes P2RY1 and P2RY12 in our small population of cats to determine which cats could have been genetically resistant to clopidogrel's platelet inhibition effects. 14 Our study suggests that DAT with clopidogrel and rivaroxaban holds promise for thromboprophylaxis in cats with cardiomyopathy and that clinical trials in such patients are warranted.
Sara T. Lo was funded by the Students Training in Advanced Research program, NIH T35.