The oxylipin analog CS585 prevents platelet activation and thrombosis through activation of the prostacyclin receptor

Abstract

Cardiovascular disease remains the primary cause of morbidity and mortality globally. Platelet activation is critical for maintaining hemostasis and preventing the leakage of blood cells from the vessel. There has been a paucity in the development of new drugs to target platelet reactivity. Recently, the oxylipin 12(S)-hydroxy-eicosatrienoic acid (12-HETrE), which is produced in platelets, was shown to limit platelet reactivity by activating the prostacyclin receptor. Here, we demonstrated the synthesis of a novel analog of 12-HETrE, known as CS585. Human blood and mouse models of hemostasis and thrombosis were assessed for the ability of CS585 to attenuate platelet activation and thrombosis without increasing the risk of bleeding. Human platelet activation was assessed using aggregometry, flow cytometry, western blot analysis, total thrombus formation analysis system, microfluidic perfusion chamber, and thromboelastography. Hemostasis, thrombosis, and bleeding assays were performed in mice. CS585 was shown to potently target the prostacyclin receptor on the human platelet, resulting in a highly selective and effective mechanism for the prevention of platelet activation. Furthermore, CS585 was shown to inhibit platelet function in human whole blood ex vivo, prevent thrombosis in both small and large vessels in mouse models, and exhibit long-lasting prevention of clot formation. Finally, CS585 was not observed to perturb coagulation or increase the risk of bleeding in the mouse model. Hence, CS585 represents a new validated target for the treatment of thrombotic diseases without the risk of bleeding or off-target activation observed with other prostaglandin receptor agonists.

Graphical abstract

Figure 1.

Structure and kinetics of CS585. Potency of CS585 and 12-HETrE. (A) Structure of CS585. (B) Synthesis of CS585 involved a 3-step synthesis starting with a pyrazine derivative. The pyrazine derivative reacted with an alcohol to form 4-((5-diphenylpyrazin-2-yl)(isopropyl)amino)butan-1-ol. This intermediate was converted to a second intermediate, 4-((5-diphenylpyrazin-2-yl)(amino)butyl 4-methanebenzenesulfonate. Finally, the prostacyclin receptor agonist CS585 (5-((4-((5,6-diphenylpyrazin-yl)(isopropyl)amino)butyl)thio)-2,4-dihydro-3H-,12,4-triazol-3-one) is formed. (C) PK were assessed in mouse plasma over a 24-hour period. (D) Aggregation dose-response comparison of washed human platelets treated with vehicle, CS585 (3.125-100 nM), or 12-HETrE (500-15 000 nM), stimulated with collagen (0.5 μg/mL) (n = 4). (E) Integrin αIIbβ3 activation, α-granule, and dense granule secretion dose-response comparison of washed human platelets treated with vehicle, CS585 (3.125-100 nM) or 12-HETrE (500-15 000 nM) stimulated with convulxin (25 ng/mL) (n = 4). Data represent mean ± SEM. A one-way analysis of variance (ANOVA) was performed using an uncorrected Fisher least significant difference post hoc test. Asterisks denote statistical differences between the vehicle and treated groups: ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. MFI, mean fluorescence intensity.

Figure 2.

CS585 potently inhibits human platelet activity. (A) Aggregation of washed human platelets treated with vehicle or CS585 (0.2-100 nM), stimulated with thrombin (0.5 nM), collagen (0.5 μg/mL), or ADP (20 μM). Data represent mean ± SEM. One-way ANOVA with Dunnett comparison (n = 4-6). (B) Representative aggregation curves for panel A. (C) Aggregation of human PRP treated with vehicle or CS585 (0.25-5 μM), stimulated with collagen (1 and 2 μg/mL, closed and open boxes, respectively) (n = 4). Data represent mean ± SEM. Mixed effects analysis with Dunnett comparison. Collagen aggregation of 1 μg/mL response compared with that of the vehicle; ∗∗P < .01; ∗∗∗∗P < .0001. Collagen aggregation of 2 μg/mL response compared with that of the vehicle; ^##P < .01; ^####P < .0001. (D) Representative aggregation curves for panel C. (E) Aggregation of washed human platelets treated with CS585 (1 μM), stimulated with thrombin (0.5-5 nM), collagen (0.5-10 μg/mL), or ADP (10-80 μM) (n = 4). (F) Expression of VASP phosphorylation (Ser157 VASP, 50 kilodalton [kDa]; total VASP, 46 and 50 kDa), and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (37 kDa) on washed human platelets treated with vehicle or CS585 (10 pM-100 nM) for 10 minutes (n = 4). Representative blots are included above the graph. Data represent mean ± SEM. One-way ANOVA and Dunnett comparison. Asterisks denote statistical differences between vehicle and treated groups. (G) Activation of integrin αIIbβ3, α-granule, and dense granule secretion of washed human platelets treated vehicle or CS585 (6.2-100 nM), stimulated with convulxin (25 ng/mL) (n = 6). Data represent mean ± SEM. One-way ANOVA with Dunnett comparison. Asterisks denote statistical differences between vehicle and treated groups. (H) cAMP and cGMP levels measured in washed human platelets treated with vehicle, CS585 (0.1-100 nM), or positive controls forskolin (10 μM) and PAPA-NONOate (PAPA-NO, 5 μM) (n = 4). Data represent mean ± SEM. One-way ANOVA with Dunnett multiple comparison test. (I) Expression of Ser157 VASP (50 kDa) and GAPDH (37 kDa) in washed human platelets treated with vehicle or PKA inhibitor H89 (100 μM) before incubation with vehicle or CS585 (100 nM) (n = 6). Data represent mean ± SEM. One-way ANOVA with Tukey multiple comparison test. (J) Representative blot of panel I. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 3.

CS585 is selective to activation of the IP receptor. (A) Aggregation of washed human platelets treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM), DP1 receptor inhibitor (MK-0542, 4 nM), or EP₂ and EP₄ receptor inhibitors (TG4-155, 2 μM and CJ-42794, 80 nM, respectively) before treatment with vehicle or CS585 (0.05-100 nM) stimulated with collagen (0.5 μg/mL; n = 4-5). Data represent mean ± SEM. Two-factor mixed-effects analysis with Tukey multiple comparisons test. Each condition is compared with its corresponding vehicle. (B) Representative aggregation curves for panel A. (C) Activation of integrin αIIbβ3, α-granule, and dense granule secretion of washed human platelets treated with the vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM), DP1 receptor inhibitor (MK-0542, 4 nM), or EP₂ and EP₄ receptor inhibitors (TG4-155, 2 μM and CJ-42794, 80 nM) before treatment with the vehicle or CS585 (12.5-100 nM), stimulated with convulxin (25 ng/mL) (n = 4-5). The results are expressed as the relative mean (percentage of vehicle MFI) ± SEM. Two-factor mixed-effects analysis with Sidak multiple comparisons. Each condition is compared with the corresponding vehicle. (D) Expression of Ser157 pVASP (50 kDa) and GAPDH (37 kDa) in washed human platelets treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM), DP1 receptor inhibitor (MK-0542, 4 nM), or EP₂ and EP₄ receptor inhibitors (TG4-155, 2 μM and CJ-42794, 80 nM, respectively) before treatment with vehicle or CS585 (10 pM-100 nM; n = 4). Data represent mean ± SEM. Two-way ANOVA with Dunnett correction. Asterisks denote statistical differences between vehicle and treated groups. (E) Expression of Ser157 pVASP (50 kDa) and GAPDH (37 kDa) in washed platelets from WT and IP receptor-deficient (IP^−/−) mice treated with vehicle, CS585 (1, 5, or 10 μM), 12-HETrE (10 or 25 μM), or forskolin (5 μM; n = 3; 2 mice pooled per n). ∗P < .05; ∗∗P < .01; ∗∗∗∗P < .0001.

Figure 4.

CS585 attenuates ex vivo platelet adhesion and thrombus formation under flow. (A) Human whole blood treated with vehicle or CS585 (0.062-5 μM) was perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (B) Quantification and representative images of platelet adhesion and accumulation in human whole blood stained with DiOC₆ and treated with vehicle or CS585 (250 nM-1 μM) perfused through a collagen-coated chamber at arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (C) Human whole blood treated with vehicle or iloprost (0.125-2.5 nM) and then perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (D) Quantification and representative images of platelet adhesion and accumulation of human whole blood stained with DiOC₆ and treated with vehicle or iloprost (0.25-1 nM) perfused through a collagen-coated chamber at arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (E) Human whole blood treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM) before treatment with vehicle or CS585 (1 μM) perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Tukey multiple comparison test. (F) Quantification and representative images of platelet adhesion and accumulation in human whole blood stained with DiOC₆ and treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM) before treatment with vehicle or CS585 (1 μM) perfused through a collagen-coated chamber under arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (G) Quantification and representative contour plots of phosphatidylserine exposure in washed human platelets treated with vehicle, CS585 (100 nM), or iloprost (20 nM), incubated with annexin V and an antibody for CD62P, then stimulated with 0.5 nM thrombin or 50 ng/mL (n = 5). Data represent mean percentage of dual-positive platelets ± SEM. One-way ANOVA with Tukey multiple comparison test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 4.

CS585 attenuates ex vivo platelet adhesion and thrombus formation under flow. (A) Human whole blood treated with vehicle or CS585 (0.062-5 μM) was perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (B) Quantification and representative images of platelet adhesion and accumulation in human whole blood stained with DiOC₆ and treated with vehicle or CS585 (250 nM-1 μM) perfused through a collagen-coated chamber at arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (C) Human whole blood treated with vehicle or iloprost (0.125-2.5 nM) and then perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (D) Quantification and representative images of platelet adhesion and accumulation of human whole blood stained with DiOC₆ and treated with vehicle or iloprost (0.25-1 nM) perfused through a collagen-coated chamber at arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (E) Human whole blood treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM) before treatment with vehicle or CS585 (1 μM) perfused on a collagen-coated PL chip at arterial shear using T-TAS (n = 5). Data represent mean ± SEM. One-way ANOVA with Tukey multiple comparison test. (F) Quantification and representative images of platelet adhesion and accumulation in human whole blood stained with DiOC₆ and treated with vehicle or an IP receptor inhibitor (Ro 1138542, 5 μM) before treatment with vehicle or CS585 (1 μM) perfused through a collagen-coated chamber under arterial shear (n = 6). Data represent mean ± SEM. Two-way ANOVA. Scale bars represent 100 μm. (G) Quantification and representative contour plots of phosphatidylserine exposure in washed human platelets treated with vehicle, CS585 (100 nM), or iloprost (20 nM), incubated with annexin V and an antibody for CD62P, then stimulated with 0.5 nM thrombin or 50 ng/mL (n = 5). Data represent mean percentage of dual-positive platelets ± SEM. One-way ANOVA with Tukey multiple comparison test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 5.

CS585 prevents thrombosis in small and large vessels. (A) WT mice were administered vehicle or CS585 (0.5-6 mg/kg), followed by the induction of cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (B) Representative images of panel A showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (C) WT mice were administered vehicle or CS585 (0.5-6 mg/kg) and the carotid artery FeCl₃-induced thrombosis model was assessed (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (D) Representative images of panel C. Scale bars represent 500 μm. (E) WT mice were administered vehicle or CS585 (6 mg/kg) and the stability of CS585 in the blood was determined by induction of the cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury at 4 hours and 18 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (F) Representative images of panel E, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (G) WT mice were administered a single dose of vehicle or CS585 (1.5-6 mg/kg) per oral (PO) and the functional efficacy of CS585 in the blood was determined by cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury 4 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (H) Representative images of panel G, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (I) WT mice were administered vehicle or CS585 (3 mg/kg) PO and the stability of CS585 in the blood was determined by cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury 24, 48, and 72 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (J) Representative images of panel I, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (K) Quantification of phosphatidylserine exposure in washed platelets from WT mice IV administered with vehicle or CS585 (6 mg/kg; CS585 [IV]) and stained with annexin V, stimulated with convulxin (50 ng/mL). Washed platelets from untreated WT mice treated ex vivo with CS585 (10 μM) before annexin V staining and stimulation with convulxin (50 ng/mL) (n = 5). Quantification of the percentage of annexin V–positive events represent the levels of PS exposure. Data represent mean ± SEM. One-way ANOVA with Dunnett multiple comparisons. (L) WT mice treated with vehicle or CS585 (6 mg/kg) daily starting 24 hours before thrombus initiation. The venous thrombus mass was measured 2 days after inferior vena cava ligation (n = 5). Data represent mean ± SEM. Welch t-test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 5.

CS585 prevents thrombosis in small and large vessels. (A) WT mice were administered vehicle or CS585 (0.5-6 mg/kg), followed by the induction of cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (B) Representative images of panel A showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (C) WT mice were administered vehicle or CS585 (0.5-6 mg/kg) and the carotid artery FeCl₃-induced thrombosis model was assessed (n = 5). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (D) Representative images of panel C. Scale bars represent 500 μm. (E) WT mice were administered vehicle or CS585 (6 mg/kg) and the stability of CS585 in the blood was determined by induction of the cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury at 4 hours and 18 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (F) Representative images of panel E, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (G) WT mice were administered a single dose of vehicle or CS585 (1.5-6 mg/kg) per oral (PO) and the functional efficacy of CS585 in the blood was determined by cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury 4 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (H) Representative images of panel G, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (I) WT mice were administered vehicle or CS585 (3 mg/kg) PO and the stability of CS585 in the blood was determined by cremaster arteriole laser-induced injury to assess platelet and fibrin accumulation at the site of injury 24, 48, and 72 hours after administration (n = 3 mice per condition/8 injuries per mouse). Data represent mean ± SEM. Two-way ANOVA. (J) Representative images of panel I, showing platelet plug formation (green) and fibrin accumulation (red). Scale bars represent 50 μm. (K) Quantification of phosphatidylserine exposure in washed platelets from WT mice IV administered with vehicle or CS585 (6 mg/kg; CS585 [IV]) and stained with annexin V, stimulated with convulxin (50 ng/mL). Washed platelets from untreated WT mice treated ex vivo with CS585 (10 μM) before annexin V staining and stimulation with convulxin (50 ng/mL) (n = 5). Quantification of the percentage of annexin V–positive events represent the levels of PS exposure. Data represent mean ± SEM. One-way ANOVA with Dunnett multiple comparisons. (L) WT mice treated with vehicle or CS585 (6 mg/kg) daily starting 24 hours before thrombus initiation. The venous thrombus mass was measured 2 days after inferior vena cava ligation (n = 5). Data represent mean ± SEM. Welch t-test. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001.

Figure 6.

CS585 does not alter hemostasis. (A) WT mice were treated with vehicle or CS585 (0.5 or 6 mg/kg), and bleeding time of the tail was assessed (n = 10). Data represent mean ± SEM. One-way ANOVA with Dunnett correction. (B-E) Human whole blood was treated with the vehicle, CS585 (100 nM or 1 μM), rivaroxaban (250 ng/mL), or CS585 (100 nM or 1 μM) + rivaroxaban (250 ng/mL). Coagulation parameters were assessed using TEG (n = 4). (B) Reaction time is the time to initial fibrin thread formation (minutes). (C) Maximum amplitude and the maximum strength of the clot (mm). (D) K time represents the time until the clot reaches a strength of 20 mm (minutes). (E) α angle, rate of clot formation (degree). One-way ANOVA with Dunnett correction. (F) Representative tracing of the coagulation parameters measured in panels B-E. (G) The maximum rate of thrombin generation was measured by TEG using a velocity curve that graphs the clot strength over time (n = 4). One-way ANOVA with Dunnett correction. (H) Representative tracing of the velocity curve used to calculate the rate of thrombin generation. ∗P < .05; ∗∗P < .01; ∗∗∗P < .001; ∗∗∗∗P < .0001. ns, nonsignificant.