Comparative In Vitro Study of Various α2-Adrenoreceptor Agonist Drugs for Ticagrelor Reversal

Ticagrelor, an antiplatelet adenosine diphosphate (ADP)-P2Y12 receptor antagonist, increases the risk of bleeding. Its management is challenging because platelet transfusion is ineffective and no specific antidote is currently available. Epinephrine, a vasopressor catecholamine prescribed during shock, restores platelet functions inhibited by ticagrelor through stimulation of α2A-adrenoreceptors. It subsequently inhibits cyclic adenosine monophosphate (cAMP) pathway and PI3K signaling. However, since epinephrine may expose a patient to deleterious hemodynamic effects, we hypothesized that other α2-adrenoreceptor agonist drugs used in clinical practice with fewer side effects could reverse the antiplatelet effects of ticagrelor. We compared in vitro the efficacy of clonidine, dexmedetomidine, brimonidine, and norepinephrine with epinephrine to restore ADP- and PAR-1-AP-induced washed platelet aggregation inhibited by ticagrelor, as well as resulting platelet cAMP levels. In ticagrelor-free samples, none of the α2-adrenoreceptor agonists induced aggregation by itself but all of them potentiated ADP-induced aggregation. Compared with epinephrine, norepinephrine, and brimonidine partially restored ADP- and fully restored PAR-1-AP-induced aggregation inhibited by ticagrelor while clonidine and dexmedetomidine were ineffective. Indeed, this lack of effect resulted from a lower decrease in cAMP concentration elicited by these partial α2-adrenoreceptor agonists, clonidine, and dexmedetomidine, compared with full α2-agonists. Our results support the development of specific full and systemic α2-adrenoreceptor agonists for ticagrelor reversal.


Introduction
Ticagrelor is a direct-acting, selective, and reversible antagonist of the platelet adenosine diphosphate (ADP) P2Y 12 receptor that is recommended for the treatment of acute coronary syndromes [1]. Like any antithrombotic drug, ticagrelor exposes to bleeding complications, management of which is challenging. Indeed, ticagrelor is not dialysable and the circulating concentrations of ticagrelor and its active metabolite make platelet transfusion totally inefficient within at least 24 h after intake. Moreover, hemostatic agents including recombinant activated factor VII, tranexamic

Materials and Methods
Blood samples were obtained from the French Blood Bank Institute (Etablissement Français du Sang, Paris, France, convention ref. C CPSL UNT n • 12/EFS/038) from healthy volunteers after obtaining written informed consent. Blood samples were collected in acid citrate dextrose (ACD) tubes (BD Vacutainer, citric acid 5.7 mM, trisodium citrate 11.2 mM, and dextrose 20 mM, final concentrations).

Isolation and Preparation of Washed Platelet Suspension
Experiments were performed on washed platelets in order to investigate the direct effects of α 2 -adrenoreceptor agonists on platelets, independently of the plasma proteins, and the anticoagulant used for blood sampling. The washed platelet suspension was prepared as previously described [11]: briefly, blood samples collected in ACD tubes were mixed with wash buffer (citric acid 36 mM, glucose 5 mM, KCl 5 mM, CaCl 2 2 mM, MgCl 2 1 mM, and NaCl 103 mM; pH = 6.5) containing 0.2 µM prostaglandin E1 (PGE1, Sigma-Aldrich, St. Louis, MO, USA) and 0.03 IU/mL apyrase (Agro-Bio, La Ferté-Saint-Aubin, France) to avoid platelet activation during preparation. Platelet-rich plasma was obtained after the first spin (216 g for 11 min at room temperature). Platelets were washed twice with wash buffer, containing PGE1 and apyrase, and centrifuged (1200× g for 11 min). The last pellet was resuspended in suspension buffer (Hepes 10 mM, NaCl 140 mM, KCl 3 mM, NaHCO 3 5 mM, MgCl 2 0.5 mM, and glucose 10 mM; pH = 7.35) to a final concentration of 3 × 10 8 platelets/mL (=300 G/L, within the physiological range for platelet count). Apyrase (0.03 IU/mL) was added to prevent ADP receptor desensitization; then, CaCl 2 2 mM was added [13].

Agonists and Inhibitors
Ticagrelor was extracted from tablets (Brilique ® 90 mg, AstraZeneca, Mölndal, Sweden) as previously described [11]. Briefly, tablets were dispersed in 10 mM HCl and the water-soluble excipient extracted by centrifugation (5 min, 5000× g, 20 • C). The pellet was washed 3 times with distilled water followed by centrifugation (5 min, 5000× g, 20 • C). The pellet was resuspended in distilled water, snap-frozen in liquid nitrogen, and then lyophilized. Ticagrelor was extracted from the lyophilized material by incubation in pure DMSO. Insoluble material in the ticagrelor-saturated DMSO was removed by two centrifugations (5 min, 5000× g, 20 • C). The concentration of ticagrelor-saturated DMSO solution was approximately 100 mg/mL, and was diluted 1/10 in pure DMSO prior storage at −80 • C. Further dilutions were performed in suspension buffer containing 1% DMSO. The final DMSO concentration in all experiments was 0.001%. The final concentration of ticagrelor was assessed by reference to ticagrelor raw material provided by AstraZeneca (Mölndal, Sweden).
Washed platelets were incubated with ticagrelor for 10 min at 37 • C prior to use. A final concentration of 100 nM ticagrelor was selected from a dose-ranging study since it inhibited at least 80% of platelet aggregation [11]. ADP (Roche Molecular Biochemicals, Meylan, France) diluted in suspension buffer was used at either 5 µM final concentration, selected to induce less than 50% platelet aggregation in all samples, in order to capture potential α 2 -adrenoreceptor agonist synergism, or 10 µM final concentration, to maximize potential restoration in the presence of inhibitor. Since ADP acts as a second agonist released from the dense granules in response to strong platelet stimulation, we repeated the experiments using proteinase-activated receptor-1 activating peptide at 2 µM final concentration (PAR-1-AP, Bachem, Basel, Switzerland) as a primary agonist. Here, the aim was to use PAR-1-AP at a concentration sufficient to induce ADP release from dense granules and to assess subsequent ADP-dependent aggregation, not to induce maximal aggregation, independent of ADP, as observed at higher PAR-1-AP concentrations.
Epinephrine (Aguettant, Lyon Gerland, France) was diluted in suspension buffer and used at a final concentration of 10 µM which was selected from a dose-ranging study as the lowest concentration increasing ADP-induced platelet aggregation to a similar level to 10 µM PAR-1-AP, a surrogate for maximal activation and aggregation. Norepinephrine (Sigma-Aldrich, St. Louis, MO, USA), clonidine (Boehringer Ingelheim, Paris, France), dexmedetomidine (Baxter SA, Guyancourt, France), and brimonidine (Sigma-Aldrich, St. Louis, MO, USA) were diluted in suspension buffer and used at a final concentration of 10 µM (except dexmedetomidine, which cannot be dissolved at a higher concentration than 8 µM). For clonidine, dexmedetomidine, and brimonidine, final concentrations were selected according to the available data in vitro and dose-ranging studies.

Measurement of Intracellular cAMP
Intracellular cAMP concentration was measured using the cAMP kit (Cisbio Bioassays, Codolet, France) based on an immuno-enzymatic method, according to the manufacturer's protocol. Briefly, 40 µL washed platelets (5 × 10 8 platelets/mL) pre-treated with ticagrelor or vehicle (resuspension buffer/DMSO) were incubated with 5 µL PGE1 (final concentration 1 µM) allowing cAMP production in experimental conditions. After 30 s, 5 µL of agonists (ADP or a combination of ADP and α 2 -adrenoreceptor agonists) were added, and the reaction was stopped 1 min and 3 min later using 50 µL lysis buffer containing the phosphodiesterase pan-inhibitor 3-isobutyl-1-methylxanthine (final concentration 2 mM). The samples were then centrifuged at 4 • C for 1 min at 13,200 rpm and the supernatant was stored at −20 • C, and subsequently thawed immediately prior testing. The results were expressed in picomoles adjusted to 10 9 platelets (pmol/10 9 platelets).

Statistical Analysis
Quantitative data were expressed as medians (minimum-maximum) or box-and-whisker plots showing median, interquartile range, minimum, and maximum values. Statistical analysis was performed using Prism v6.0 (GraphPad software, San Diego, CA, USA). Nonparametric tests were used for comparisons. Variables were compared using the Friedman test, followed by the Wilcoxon test (paired samples) or the Mann-Whitney (independent samples). All statistical tests were two-sided and p < 0.05 was considered statistically significant.

Statistical Analysis
Quantitative data were expressed as medians (minimum-maximum) or box-and-whisker plots showing median, interquartile range, minimum, and maximum values. Statistical analysis was performed using Prism v6.0 (GraphPad software, San Diego, CA, USA). Nonparametric tests were used for comparisons. Variables were compared using the Friedman test, followed by the Wilcoxon test (paired samples) or the Mann-Whitney (independent samples). All statistical tests were twosided and p < 0.05 was considered statistically significant.

Norepinephrine and Brimonidine Fully Inhibit PGE1-Stimulated cAMP Production Whereas Clonidine and Dexmedetomidine Induce Only Weak Inhibition
To further explore the differences between various α2-adrenoreceptor agonists in restoring platelet aggregation inhibited by ticagrelor, we assessed the level of cAMP in platelets sensitized with

Discussion
Previous studies investigating the in vitro effects of α 2 -adrenoreceptor agonists on platelet functions mainly focused on epinephrine. In the present study, we aimed to compare the effects on platelet aggregation of four marketed α 2 -adrenoreceptor agonists compared with those of epinephrine. We confirmed that all α 2 -adrenoreceptor agonists potentiated ADP-induced platelet aggregation. More importantly, we provided evidence that norepinephrine or brimonidine in synergism with ADP restored platelet aggregation inhibited by ticagrelor with the same efficacy as epinephrine, whereas dexmedetomidine and clonidine failed to overcome the inhibitory effect of ticagrelor.
Several mechanistic hypotheses may be advanced to explain these discrepancies among α 2 -adrenoreceptor agonists. First, it is well known that cAMP is the main intracellular regulator of platelet activation and consecutively of platelet aggregation. Dexmedetomidine and clonidine may induce insufficient suppression of cAMP production to elicit platelet aggregation. Indeed, we observed that dexmedetomidine and clonidine had a weak effect on the intracellular cAMP level unlike epinephrine, norepinephrine, and brimonidine. This is in line with the pharmacological characteristics identifying full α 2 -adrenoreceptor agonists, including norepinephrine and brimonidine, and partial α 2 -adrenoreceptor agonists, including dexmedetomidine and clonidine. Kafka et al. demonstrated that clonidine was as potent as epinephrine and norepinephrine in binding to the receptor but was only 1/10 as potent in inhibiting PGE1-stimulated cAMP production [14]. Second, we previously showed that restoration of platelet aggregation by epinephrine is due to cAMP pathway inhibition and PI3K pathway activation, but simultaneously requires ADP-induced P2Y 1 -receptor-related signaling and Ca 2+ mobilization [11]. We cannot exclude that dexmedetomidine and clonidine exert only partial activation of the PI3K pathway, thereby limiting subsequent signaling involved in sustained activation, granule secretion, and stable thrombus formation.
Finally, dexmedetomidine and clonidine are able to bind I 1 -imidazoline receptors, distinct from α 2 -adrenoreceptors on the platelet membrane surface [15,16]. Since I 1 -receptor stimulation induces guanylate cyclase activation and the subsequent increase in the intracellular cGMP level, dexmedetomidine and clonidine can also participate in platelet aggregation inhibition [17]. Consequently, dexmedetomidine and clonidine have both enhancing and suppressive effects on platelet functions through their opposite action on α 2 -adrenoceptors and on I 1 -imidazoline receptors, respectively [18]. In our experimental conditions, these two antagonist mechanisms resulted in potentiation of ADP-induced platelet aggregation insufficient to induce platelet aggregation in the presence of ticagrelor. We nevertheless noted that dexmedetomidine was more efficient than clonidine.
These in vitro findings translated into conflicting in vivo observations. Adam et al. showed that clonidine added ex vivo to samples from patients treated with dual antiplatelet therapy (aspirin and clopidogrel) did not affect ADP-induced platelet aggregation compared with control [19]. However, Janatmakan et al. recently reported in a randomized study conducted in 120 patients undergoing spinal surgery that clonidine or dexmedetomidine reduced intraoperative blood loss compared with the control group. However, none of the patients received antiplatelet agents [20]. In contrast, Mizrak et al. reported that dexmedetomidine increased bleeding in pediatric patients undergoing adenotonsillectomy [21].
Our study aimed to identify potential α 2 -adrenoreceptor agonists as efficient as epinephrine in restoring ticagrelor-inhibited platelet aggregation in vitro, but with fewer deleterious hemodynamic effects. So far studies performed with epinephrine are encouraging. Using a model of artery stenosis, Yao et al. and Samama et al. have reported in canines and pigs, respectively, that epinephrine infusion fully restored platelet thrombus formation abolished by clopidogrel [22,23] Singh et al. recently confirmed that even low clinically-relevant doses of epinephrine infusion (0.15 µg/kg/min, leading to epinephrine concentration 20 nM) slightly improved platelet aggregation and clot formation in healthy volunteers receiving ticagrelor [24]. However, it is likely that higher epinephrine doses would be required to translate biological surrogates into clinically-relevant effects on bleeding control, exposing patients to hypertension and bleeding enhancement, and serious cerebral and cardiovascular adverse effects [25]. Norepinephrine has not been assessed with the aim to restore hemostasis. However, norepinephrine is a catecholamine cumulating strong α 1 -, α 2 -, and βeffects, and 10 times less potent than epinephrine suggesting a supratherapeutic dose to exert potential platelet effects, exposing patients to unacceptable hemodynamic effects. Finally, the more attractive candidate was brimonidine, a pure full α 2 -adrenoreceptor agonist. However, its restricted cutaneous or intraocular administration limits the scope of its use to restore platelet function inhibited by ticagrelor.

Study Limitations and Conclusions
Regarding limitations, we acknowledge that we assessed a way to improve platelet functions inhibited by ticagrelor rather than a specific antidote, which is in development but not marketed yet. The main weakness of our study lies in the in vitro design, which makes it difficult to transpose our findings to the clinical management of ticagrelor-induced bleeding. To go further, it would be interesting to assess the effects of α 2 -adrenoreceptor agonists on various assays using platelets from ticagrelor-treated patients.
Norepinephrine, recommended for hemorrhagic shock, may expose patients to side effects, especially in the absence of shock, and brimonidine is limited by its topical administration. Further research is required to assess the reversal effects of these drugs in the clinical setting and to support the development of pure full and systemic α 2 -adrenoreceptor agonists for ticagrelor reversal.