Ticagrelor inverse agonist activity at the P2Y12 receptor is non-reversible versus its endogenous agonist adenosine 5´-diphosphate

Abstract

Background and purpose: Ticagrelor is labelled as a reversible, direct-acting platelet P2Y₁₂ receptor (P2Y₁₂ R) antagonist that is indicated clinically for the prevention of thrombotic events in patients with acute coronary syndrome (ACS). As with many antiplatelet drugs, ticagrelor therapy increases bleeding risk in patients, which may require platelet transfusion in emergency situations. The aim of this study was to further examine the reversibility of ticagrelor at the P2Y₁₂ R.

Experimental approach: Studies were performed in human platelets, with P2Y₁₂ R-stimulated GTPase activity and platelet aggregation assessed. Cell-based bioluminescence resonance energy transfer (BRET) assays were undertaken to assess G protein-subunit activation downstream of P2Y₁₂ R activation.

Key results: Initial studies revealed that a range of P2Y₁₂ R ligands, including ticagrelor, displayed inverse agonist activity at P2Y₁₂ R. Only ticagrelor was resistant to washout and, in human platelet and cell-based assays, washing failed to reverse ticagrelor-dependent inhibition of ADP-stimulated P2Y₁₂ R function. The P2Y₁₂ R agonist 2MeSADP, which was also resistant to washout, was able to effectively compete with ticagrelor. In silico docking revealed that ticagrelor and 2MeSADP penetrated more deeply into the orthosteric binding pocket of the P2Y₁₂ R than other P2Y₁₂ R ligands.

Conclusion and implications: Ticagrelor binding to P2Y₁₂ R is prolonged and more akin to that of an irreversible antagonist, especially versus the endogenous P2Y₁₂ R agonist ADP. This study highlights the potential clinical need for novel ticagrelor reversal strategies in patients with spontaneous major bleeding, and for bleeding associated with urgent invasive procedures.

Keywords: P2Y12 receptor; acute coronary syndrome; blood platelets; irreversibility; ticagrelor.

FIGURE 1

Ligand‐dependent regulation of P2Y₁₂R responsiveness as assessed by a BRET‐based assay. HEK 293T cells were co‐transfected with human FLAG‐P2Y₁₂R and heterotrimeric G proteins, RlucII‐Gαi, untagged Gβ and GFP‐Gγ. Forty‐eight hours after transfection, cells were treated with ligand for 5 min at 37°C and subsequent changes in BRET signal were measured with reduction in BRET signal, a consequence of G protein‐subunit disassociation. Results are expressed as delta BRET. (a) Ligand‐induced changes in receptor activation as assessed by BRET following treatment with ticagrelor, 2MeSADP, ADP and AR‐C66096 versus vehicle (0.1% DMSO) control. (b) Ligand‐induced changes in receptor activation as assessed by BRET following treatment with ticagrelor, cangrelor, elinogrel and AZD1283 versus vehicle (0.1% DMSO) control. Data shown are the means ± SEM of at least five independent experiments, each performed in triplicate.

FIGURE 2

Ticagrelor‐dependent activity at the P2Y₁₂R is resistant to washout. HEK 293T cells were co‐transfected with human FLAG‐P2Y₁₂R and heterotrimeric G proteins, RlucII‐Gαi, untagged Gβ and GFP‐Gγ. Forty‐eight hours after transfection, cells were treated with ticagrelor (10 μM) or vehicle control for 30 min at 37°C. Following ticagrelor treatment, cells were washed (Tic washed) for either 3 × 10‐min washes (a, b) or 3 × 30‐min washes (a), and receptor activity was compared with that following acute ticagrelor treatment (10 μM; 5 min; Tic). In (a), ADP‐stimulated (10 μM; 5 min) activity was assessed in non‐ticagrelor‐treated cells (ADP), in cells treated acutely with ticagrelor (Tic + ADP) or in cells following more prolonged ticagrelor treatment and subsequent washing (Tic washed + ADP). In (b), 2MeSADP‐stimulated (10 μM; 5 min) activity was assessed in non‐ticagrelor‐treated cells (2MeSADP), in cells treated acutely with ticagrelor (Tic + 2MeSADP) or in cells following more prolonged ticagrelor treatment and subsequent washing (Tic washed + 2MeSADP). Data shown are the means ± SEM of at least five independent experiments, each performed in triplicate. Statistical analysis was performed using one‐way ANOVA and followed by Bonferroni's multiple comparison test.

FIGURE 3

Prasugrel active metabolite but not cangrelor or AR‐C66096 activity at the P2Y₁₂R is resistant to washout. HEK 293T cells were co‐transfected with human FLAG‐P2Y₁₂R and heterotrimeric G proteins, RlucII‐Gαi, untagged Gβ and GFP‐Gγ. Forty‐eight hours after transfection, cells were treated with P2Y₁₂R antagonist (AR‐C66096 [10 μM; a]), cangrelor (10 μM; b) or the active metabolite of prasugrel (1‐μM R‐138727; c) for 30 min at 37°C. Following antagonist treatment, cells were washed for 3 × 10 min, and receptor activity was compared with that following acute antagonist treatment alone (5 min). ADP‐stimulated (10 μM; 5 min) activity was assessed in non‐ticagrelor‐treated cells (ADP), in cells treated acutely with ticagrelor (Tic + ADP) or in cells following more prolonged ticagrelor treatment and subsequent washing (Tic washed + ADP). In (b), 2MeSADP‐stimulated (10 μM; 5 min) activity was assessed in non‐ticagrelor‐treated cells (2MeSADP), in cells treated acutely with ticagrelor (Tic + 2MeSADP) or in cells following more prolonged ticagrelor treatment and subsequent washing (Tic washed + 2MeSADP). Data shown are the means ± SEM of at least five independent experiments, each performed in triplicate. Statistical analysis was performed using one‐way ANOVA and followed by Bonferroni's multiple comparison test ([a] *P < 0.05 AR‐C66096 + ADP vs. AR‐C66096 washed + ADP; [c] *P < 0.05 cangrelor vs. cangrelor washed and *P > 0.05 cangrelor + ADP vs. cangrelor washed + ADP).

FIGURE 4

Ticagrelor and 2MeSADP activity at the P2Y₁₂R is resistant to washout. HEK 293T cells were co‐transfected with human FLAG‐P2Y₁₂R and heterotrimeric G proteins, RlucII‐Gαi, untagged Gβ and GFP‐Gγ. Forty‐eight hours after transfection, receptor activity (inverse agonist: ticagrelor, 0.4/10 μM; AZD1283, 10 μM; cangrelor, 10 μM; and elinogrel, 10 μM; agonist: ADP, 10 μM; 2MeSADP, 10 μM) was compared in cells treated with P2Y₁₂R ligand (30 min) versus that in cells treated with ligand for 30 min followed by washout for 3 × 10 min. Data are expressed as % loss of P2Y₁₂R ligand‐induced activity following washing and represent means ± SEM of at least five independent experiments, each performed in triplicate.

FIGURE 5

Resistance to washout of ticagrelor‐dependent activity is maintained in a P2Y₁₂R mutant (C194A) displaying reduced ticagrelor activity. HEK 293T cells were transfected with either the human FLAG‐P2Y₁₂R or FLAG‐C194A‐P2Y₁₂R and heterotrimeric G proteins, RlucII‐Gαi, untagged Gβ and GFP‐Gγ. Forty‐eight hours after transfection, cells were treated with ticagrelor (10 μM) or vehicle control for 30 min at 37°C. Following ticagrelor treatment, cells were washed (Tic washed) for 3 × 10 min, and receptor activity was compared with that following acute ticagrelor treatment (10 μM; 5 min; Tic). ADP‐stimulated (10 μM; 5 min) activity was assessed in non‐ticagrelor‐treated cells (ADP), in cells treated acutely with ticagrelor (Tic + ADP) or in cells following more prolonged ticagrelor treatment and subsequent washing (Tic washed + ADP). Data shown are the means ± SEM of at least five independent experiments, each performed in triplicate. Statistical analysis was performed using one‐way ANOVA and followed by Bonferroni's multiple comparison test (*P < 0.05 ticagrelor response in FLAG‐P2Y₁₂R vs. FLAG‐C194A‐P2Y₁₂R).

FIGURE 6

Ticagrelor‐dependent effects on P2Y₁₂R activity are unaffected by washout in human platelets. Human washed platelets (1 × 10⁹/ml) were untreated or treated with (a) ticagrelor (10 μM) and (b) AR‐C66096 (10 μM) or vehicle (0.1% DMSO) for 30 min at 37°C. Platelets were then washed three times with intervals of 10 min before snap frozen in equal volume of fractionation buffer in liquid nitrogen. Membrane fractions were collected by ultracentrifugation. (c) The purity of our cell membrane preparation was assessed by western blot comparing the expression of platelet membrane (integrin β3) and cytosolic (Syk) expressed proteins in membrane versus cytosolic cell fractions. Platelet cell membranes were treated either with ADP (10 μM) or vehicle (0.1% DMSO) before incubation with recombinant GTP (2 μM) and DTT (1 mM) for 1 h at room temperature. GTPase‐Glo reagents were added to measure GTP hydrolysis. Data shown are the means ± SEM of at least five independent experiments. Statistical analysis was performed using one‐way ANOVA and followed by Bonferroni's multiple comparison test ([b] *P < 0.05 AR‐C66096 + ADP vs. AR‐C66096 washed + ADP).

FIGURE 7

Inhibition of ADP‐stimulated platelet aggregation by ticagrelor is not reversed by extensive washout. (a, b) PRPs were treated either with vehicle (0.1% DMSO), ticagrelor (1 or 10 μM) or AR‐C66096 (10 μM) for 30 min at 37°C. Treated samples were either unwashed (before washout) or underwent either one 10‐min wash termed Washout 1 or two 10‐min washes termed Washout 2. As outlined in the methods following centrifugation steps, platelets were resuspended in PPP. In all cases, aggregation responses were recorded following the addition of ADP (10 μM) or vehicle (0.1% DMSO). (a) Representative aggregatory traces. (b) Data shown are the means ± SEM of at least five independent experiments. Statistical analysis was performed using one‐way ANOVA and followed by Bonferroni's multiple comparison test (*P < 0.05 DMSO vs. Tic [1 μM] or DMSO vs. Tic [1 μM] comparing before washout, after Washout 1 or after Washout 2). (c) As above, PRP was treated with vehicle (0.1% DMSO), AR‐C66096 (10 μM) or ticagrelor (10 μM) for 30 min at 37°C. Samples underwent one 10‐min wash and resuspended in PPP for 1, 4 or 24 h. Aggregation responses were subsequently recorded following the addition of ADP (10 μM) or vehicle (0.1% DMSO). (d) Platelet viability was assessed by measuring annexin V levels by FACs in platelet samples washed for 4 and 24 h as described in (c). Data shown are the means ± SEM of at least five independent experiments.

FIGURE 8

Ticagrelor is resistant to washout in platelets treated in whole blood. DMSO vehicle, AR‐C66096 (10 μM) or ticagrelor (10 μM) was added to whole blood for 10 min before centrifugation, washing and reconstitution of treated platelets into fresh, untreated plasma/red blood cells and re‐treated with vehicle, AR‐C66096 or ticagrelor as indicated for a further 10 min before flowing over collagen at arterial shear, fixation and visualization in order to monitor platelet deposition. (a) In vitro thrombus maximum projected surface area (n.s., not significant Veh vs. ARC/Veh; *P < 0.05 Veh vs. Tic/Veh; *P < 0.05 Veh vs. ARC/ARC; *P < 0.05 Veh vs. Tic/Tic). (b) In vitro thrombus volume (n.s., not significant Veh vs. ARC/Veh; *P < 0.05 Veh vs. Tic/Veh; *P < 0.05 Veh vs. ARC/ARC; *P < 0.05 Veh vs. Tic/Tic). (c) Representative images from a single n. Bars represent mean ± SEM from six independent experiments. Data were analysed by matched ANOVA with Dunnett's test comparing each treatment condition versus vehicle.

FIGURE 9

In silico docking reveals that ticagrelor and 2MeSADP penetrate more deeply into the orthosteric binding pocket of the P2Y₁₂R than cangrelor. (a) Overlay of the agonist and antagonist models of the P2Y₁₂R with 2MeSADP (blue), ADP (purple), ticagrelor (orange) and cangrelor (green) docked within the receptor orthosteric site. Model also shows the outmost plane of both the extracellular and intracellular membranes. The 2MeSADP binding pose is based on reported crystal structure, whilst that for ADP, ticagrelor and cangrelor is based on simulated docking results. (b) Geometric centre (centroid) of each ligand shown. Centroid distance to the extracellular membrane for each ligand calculated and summarized in (e). (c) Deepest point for each ligand identified and distance to extracellular membrane plane calculated and displayed in (e). (d) Focus zoom in of residues of 2MeSADP, ticagrelor and cangrelor identified to show deepest penetration into orthosteric binding pocket. (e) Summary of distances for the deepest penetration points and ligand centroid from the extracellular membrane plane for each ligand.