Optimizing antithrombotic therapy
Antiplatelet drugs are used for preventing and treating thrombotic diseases. However, they have serious adverse effects, including risk of excessive bleeding. A synthetic peptide (mP6) targeting the integrin signaling has been shown to inhibit thrombosis without affecting bleeding. Its application was hindered by inefficient and impractical intracellular delivery. Now, Pang et al. developed lipid-stabilized, high-loading peptide nanoparticles (HLPNs) for efficient in vivo delivery of the antithrombotic peptide. Injection of HLPNs loaded with engineered mP6 prevented thrombosis without affecting bleeding in mice and had therapeutic effects in a model of ischemia-reperfusion injury. This strategy might be successful in minimizing the side effects of current antiplatelet therapy.
Abstract
Inefficient delivery is a major obstacle to the development of peptide-based drugs targeting the intracellular compartment. We recently showed that selectively inhibiting integrin outside-in signaling using a peptide (mP6) derived from the Gα13-binding ExE motif within the integrin β3 cytoplasmic domain had antithrombotic effects. Here, we engineered lipid-stabilized, high-loading peptide nanoparticles (HLPN), in which a redesigned ExE peptide (M3mP6) constituted up to 70% of the total nanoparticle molarity, allowing efficient in vivo delivery. We observed that M3mP6 HLPN inhibited occlusive thrombosis more potently than a clopidogrel/aspirin combination without adverse effects on hemostasis in rodents. Furthermore, M3mP6 HLPN synergized with P2Y12 receptor inhibitors or the clopidogrel/aspirin combination in preventing thrombosis, without exacerbating hemorrhage. M3mP6 HLPN also inhibited intravascular coagulation more potently than the P2Y12 inhibitor cangrelor. Postischemia injection of M3mP6 HLPN protected the heart from myocardial ischemia–reperfusion injury in a mouse model. This study demonstrates an efficient in vivo peptide delivery strategy for a therapeutic that not only efficaciously prevented thrombosis with minimal bleeding risk but also protected from myocardial ischemia–reperfusion injury in mice.
INTRODUCTION
The inability to efficiently deliver peptides in vivo into cells is a major obstacle in developing peptide-based drugs targeting the intracellular compartments. Thus, despite the low toxicity and high specificity of natural peptides in general, successes in developing cell-penetrating drugs targeting intracellular compartments are scarce (1, 2). An efficient method for delivering peptides into cells in vivo would be an important advancement in peptide-based drug development.
Thrombotic cardiovascular disease causes more deaths than any other disease in the world (3). Blood platelets not only physiologically mediate hemostatic thrombus formation to prevent bleeding but are also critical in the development of occlusive thrombosis (4, 5). Antiplatelet therapy is therefore pivotal in the treatment of thrombotic diseases and in preventing thrombosis in patients undergoing invasive vascular procedures (6, 7). The important role of platelets in hemostasis and thrombosis requires the adhesion receptor integrin αIIbβ3 [also named glycoprotein (GP) IIb-IIIa] (8). Currently available antiplatelet drugs suppress thrombus formation either by inhibiting the activation of integrin αIIbβ3 [such as the cyclooxygenase (COX) inhibitor aspirin or inhibitors of the adenosine diphosphate (ADP) P2Y12 receptor, clopidogrel, ticagrelor, and cangrelor] or directly blocking the ligand binding function of αIIbβ3 (integrin antagonists abciximab, eptifibatide, and tirofiban) (9, 10). The P2Y12 inhibitors such as clopidogrel either with or without aspirin are the current standard of care (9, 10). However, these drugs have the serious adverse effect of excessive bleeding (11–14) because of the importance of integrin-mediated primary platelet adhesion and aggregation in hemostasis. Hemorrhage is strongly associated with poor outcomes and increased mortality (11, 15–17), thus a new generation of antiplatelet drugs that minimally affect hemostasis are needed (18). Recently, inhibitors of thrombin receptors, the protease-activated receptor 1 (PAR1) inhibitor vorapaxar (19) and PAR4 inhibitor BMS-986120 (20), have been shown to partially reduce hemorrhage as compared with P2Y12 inhibitors in animal studies. However, clinical trials revealed clear adverse hemorrhagic effects associated with vorapaxar (19). Although the mechanism for reduced hemorrhage with these inhibitors in animal models remains unclear, the hemorrhagic effects of both PAR1 and PAR4 inhibitors are consistent with the knowledge that these thrombin receptors are important for activation of the ligand binding function of integrin αIIbβ3.
Ligand binding to integrin αIIbβ3 not only mediates platelet adhesion but also transmits signals, leading to greatly expanded thrombus size, important for vascular occlusion (12, 21). We recently discovered a Gα13-dependent mechanism of integrin outside-in signaling and proposed the concept of selectively targeting this pathway without affecting the ligand binding function of integrin αIIbβ3 (21, 22). A synthetic peptide (mP6; Myr-FEEERA), derived from the Gα13-binding ExE motif of integrin β3’s cytoplasmic domain, inhibited thrombosis without affecting hemostasis (21). However, in vivo delivery of this peptide with liposomes or lipid micelles was inefficient and impractical for therapeutic use. To resolve this problem and address the challenge of intracellular peptide delivery in vivo, we engineered lipid-stabilized, high-loading peptide nanoparticles (HLPN) that incorporate high concentrations of the modified ExE peptide, M3mP6, and demonstrated superior characteristics of M3mP6 in inhibiting occlusive thrombosis without causing excessive bleeding. M3mP6 HLPN also synergistically enhanced the antithrombotic effects of current standard antiplatelet treatments while minimizing the adverse effect of hemorrhage. Furthermore, postischemia injection of M3mP6 HLPN protected the heart from ischemia/reperfusion injury in mice.
RESULTS
High-loading ExE peptide nanoparticles as a new antiplatelet drug
The Gα13-binding ExE motif of the β3 cytoplasmic domain (FEEERA) selectively mediates outside-in signaling and occlusive thrombosis (21). Although a synthetic peptide based on this sequence can disrupt Gα13-β3 interaction and outside-in signaling, it is a challenge to efficiently deliver such a peptide in vivo into cells for therapeutic use. Although liposomes or lipid micelles can be used (2), they only incorporate low concentrations of peptide (21), making it difficult to achieve efficacious doses for clinical use, except for a few extremely high-affinity drugs. We found that only 1 mM mP6 (4% of total molar content) could be achieved in the lipid micellar suspension (21). Although this concentration had effects in mice when injecting the maximum possible volume (21), it is not practical for clinical use. For this purpose, we developed lipid-stabilized HLPN for intracellular delivery of peptides in vivo. The main component of HLPN is an amphiphilic peptide, which is capable of self-assembling to form micellar nanoparticles. The peptide micellar nanoparticles are stabilized by a low percentage of phosphatidylcholine and protected by 1,2-distearoyl-sn-glycero-3-phosphoethanolamine (DSPE)–attached polyethylene glycol (PEG), forming the hydrophilic outer layer (Fig. 1A). We reengineered mP6 to enhance its incorporation into HLPN. The new peptide, M3mP6 (Myr-FEEERL), contains the critical ExE motif, and its N-terminal phenylalanine is myristoylated. The C-terminal alanine was changed to leucine, whose long hydrophobic side chain facilitates packing higher concentrations of the peptide, together with lipids into nanoparticles (Fig. 1B). This peptide retained the function of mP6 to inhibit Gα13-β3 interaction as indicated by coimmunoprecipitation (Fig. 1, C and D). The M3mP6 HLPN achieved high M3mP6 peptide loading, reaching 70% of the total nanoparticle (mol/mol) and a high peptide concentration of >10 mM in an injectable suspension [>10 times more concentrated than the original lipid micelle formulation of mP6 (21)]. As analyzed by dynamic light scattering (DLS), the vast majority (99.8 to 100%) of these lipid-stabilized, PEG-coated M3mP6 HLPN had an average size ranging from 6 to 25 nm (in different preparations), with occasional appearance of very small populations (0 to 0.2%) with larger diameters (~50 to 500 nm) (Fig. 1E). The lyophilized powder of M3mP6 was readily soluble in physiological saline for intravenous injection and was stable for >18 months, with a similar DLS profile and pharmacological effect when stored at −20°C, and for at least 2 weeks at room temperature (22°C) (fig. S1). In a pilot maximal tolerated dose (MTD) test in mice, we did not observe signs of toxicity after one bolus injection of up to 60 mg peptide/kg (~60 μmol/kg) (table S1A). Further pilot studies in rats revealed that the MTD exceeded 100 mg of peptide bolus (80× converted efficacy dose) (table S1B), and rats exhibited no observable toxic reaction to M3mP6 HLPN after a 5-day continuous infusion at 150 and 300 mg/kg per day (table S1, C and D). M3mP6 HLPN dose-dependently inhibited human platelet granule secretion and secretion-dependent secondary platelet aggregation induced by low-dose thrombin in vitro (Fig. 2, A to C), but had no effect on platelet aggregation induced by a high dose of thrombin (Fig. 2C), although platelet granule secretion was still partially inhibited by M3mP6 even at higher thrombin concentrations (Fig. 2D). M3mP6 HLPN also partially inhibited collagen-induced (Fig. 2E) and U46619-induced [thromboxane A2 (TXA2) analog] platelet aggregation (Fig. 2F), but did not affect ADP-induced platelet aggregation (Fig. 2G). M3mP6 HLPN (Fig. 2, H and I) or dimethyl sulfoxide (DMSO)–dissolved M3mP6 (fig. S2) did not affect JonA (Fig. 2H and fig. S2A) or fibrinogen (Fig. 2I and fig. S2, B and C) binding to platelets induced by PAR4 agonist peptide (PAR4AP) and, under identical conditions as the fibrinogen binding assay, M3mP6 HLPN had no effect on high-dose PAR4AP–induced platelet aggregation, but partially inhibited low-dose PAR4AP–induced platelet aggregation (fig. S2). These data suggest that M3mP6 does not directly affect inside-out signaling or the ligand binding function of αIIbβ3 but inhibits secondary platelet responses to integrin outside-in signaling. Fluorescence microscopy of fluorescently labeled M3mP6 demonstrated that M3mP6 HLPN entered platelets after incubation (fig. S3). Using flow cytometry, we found that entry of fluorescent M3mP6 HLPN into platelets was markedly greater than that of fluorescent M3mP6 in DMSO (Fig. 3A). Consistently, we found only one-fourth the concentration of M3mP6 in HLPN was required to achieve an inhibitory effect on platelet aggregation comparable to that of M3mP6 dissolved in DMSO (Fig. 3B), underscoring the improved intracellular delivery. The DMSO-dissolved M3mP6 peptide had no in vivo effect on arterial thrombosis (Fig. 3C). In contrast, M3mP6 HLPN dose-dependently inhibited occlusive thrombosis using a robust FeCl3-induced carotid artery thrombosis model in mice (Fig. 3, C and D; see fig. S4 for comparison with negative control peptides), suggesting the protective effect of the HLPN on peptide viability in vivo. Near-maximal antithrombotic effect of M3mP6 HLPN was observed at doses above 5 μmol peptide/kg when administered by retro-orbital injection (Fig. 3D), and 2.5 μmol/kg with tail vein injection (fig. S5). The HLPN containing 50% (mass/mass) M3mP6 peptide had a similar antithrombotic effect to HLPN containing 36% M3mP6 when identical amounts of peptide (5 μmol/kg) were injected, confirming that peptide concentrations but not changes in formulation determined the antithrombotic effects (fig. S6). Pharmacokinetic studies indicated that blood and plasma concentrations of M3mP6 reached a maximum within 10 min at 59.63 μg/ml in plasma (Fig. 3E) and 33.36 μg/ml in whole blood (Fig. 3F) after intravenous injection of M3mP6 (5 mg/kg; 5 μmol/kg) with t1/2−λz = 2.46 hours (table S2). Consistently, the antithrombotic effect of M3mP6 was demonstrated within 5 min after injection and lasted until about 45 min after injection (Fig. 3G). A rat 5-day infusion study showed a t1/2−λz (half-life after cessation of infusion) of 3.1 (female) and 3.9 (male) (see table S2 for pharmacokinetic characteristics). Thus, M3mP6 HLPN is a fast-acting and reversible antiplatelet drug likely suitable for intravenous injection; if needed, then its therapeutic effect might be prolonged by continuous infusion.
(A) A schematic of lipid-stabilized, high-loading peptide nanoparticles (HLPN). (B) Structure of M3mP6 peptide. (C) Dose-dependent inhibitory effects of M3mP6 on the coimmunoprecipitation of integrin β3 and Gα13 in α-thrombin (0.025 U/ml) stimulated human platelets compared with scrambled peptide (Scra). (D) Quantification of coimmunoprecipitation blots in (C) (n = 3). OD, optical density. (E) The size distribution of a single preparation of M3mP6 HLPN (6.6 ± 0.8 nm) as analyzed by DLS (43).
(A) Dose-dependent effects of M3mP6 HLPN on thrombin-induced human platelet secretion in vitro. Data from three to four experiments. (B) Dose-dependent effects of M3mP6 HLPN on thrombin (0.03 U/ml)–induced human platelet aggregation. (C) Effects of M3mP6 HLPN on thrombin-induced human platelet aggregation induced by low- and high-dose thrombin (n = 3; Scra, scrambled control peptide). (D) Effects of M3mP6 HLPN on low-dose thrombin– and higher-dose thrombin–induced secretion in human platelets (n = 3). (E) Effects of M3mP6 HLPN on collagen (1 μg/ml)–induced mouse platelet aggregation compared with scrambled peptide control (n = 3). (F) Effects of M3mP6 HLPN on U46619 (0.5 μM)–induced mouse platelet aggregation compared with scrambled peptide HLPN. (G) Effects of M3mP6 HLPN on ADP (5 μM)–induced human platelet aggregation (PRP). (H) Effects of M3mP6 HLPN on 100 μM PAR4AP–induced binding of phosphatidyl ethanolamine (PE)–conjugated antibody against activated αIIbβ3, JonA, to mouse platelets. (I) Effects of M3mP6 HLPN on 100 μM agonist (PAR4AP)–induced binding of Oregon Green–labeled fibrinogen to mouse platelets. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; ns, not significant; unpaired t test in (A), (C), (D), (E), and (F); data presented as means ± SEM. ATP, adenosine triphosphate.
(A) Flow cytometry comparison of the uptake of fluorescence (FL)–labeled M3mP6 dissolved in DMSO with that in HLPN into mouse platelets (n = 3). (B) Comparable effect of 40 μM M3mP6 in HLPN with 160 μM DMSO–solubilized M3mP6 on human platelet aggregation (n = 3). (C) Effects of M3mP6 HLPN (10 μmol/kg) on FeCl3-induced carotid artery thrombosis (n = 15), as compared with DMSO (10 μmol/kg)–solubilized M3mP6 (n = 6) and with saline control (n = 15). (D) Dose response of M3mP6 HLPN in inhibiting FeCl3-induced carotid artery occlusive thrombosis after retro-orbital injection 15 min before procedure (n = 3). (E) Pharmacokinetic study on plasma concentrations of M3mP6 HLPN after retro-orbital injection (5 μmol/kg). (F) Pharmacokinetic study on whole-blood level concentrations of M3mP6 HLPN after retro-orbital injection (5 μmol/kg). (G) Kinetics of antithrombotic effect of M3mP6 HLPN (5 μmol/kg; retro-orbital injection) on FeCl3-induced carotid artery occlusive thrombosis (n = 3 to 5). *P < 0.05, **P < 0.01, and ****P < 0.0001; one-way ANOVA with multiple comparisons in (A) and (B); Mann-Whitney test in (C); unpaired t test in (D) and (G); data presented as means ± SEM.
Comparison of antithrombotic efficacy between M3mP6 HLPN and aspirin
Next, we used a robust FeCl3-induced mouse carotid artery thrombosis model to compare M3mP6 HLPN with aspirin in inhibiting occlusive thrombosis. We used mouse models because the β3 cytoplasmic domain is identical between humans and mice. After injury induction, control C57BL/6 mice formed stable occlusive thrombosis, with a median time of 199 s, which was significantly delayed in mice injected with a one-time bolus of M3mP6 HLPN (5 or 10 μmol/kg, 15 min before procedure, retro-orbital; P < 0.0001) (Fig. 4A). Occlusion time in aspirin-treated mice (4.3 mg/kg, 2 hours before the procedure, oral) was similar to controls (Fig. 4A). To exclude the possibility of insufficient aspirin dosing, we also tested the effect of a very high dose of aspirin [36 mg/kg, intraperitoneally (ip), 1 hour before experiment, sufficient to severely prolong bleeding (see below)]. This dose of aspirin still had no obvious effect on occlusive thrombus formation (Fig. 4B), which contrasted with the marked antithrombotic effect of M3mP6 HLPN (5 μmol/kg). Combined use of M3mP6 HLPN (5 μmol/kg) and aspirin (36 mg/kg) was similar to M3mP6 HLPN alone in inhibiting thrombosis (Fig. 4B). The results suggest that M3mP6 HLPN might be superior to aspirin in inhibiting FeCl3-induced arterial thrombosis.
(A) Comparison of saline control (n = 17), M3mP6 HLPN (10 μmol/kg; 15 min before procedure, n = 18), aspirin (4.3 mg/kg; oral 2 hours before procedure, n = 11), and M3mP6 HLPN plus aspirin (n = 11) treatments. (B) Comparison of saline control (n = 8), M3mP6 HLPN (5 μmol/kg; 15 min before procedure, n = 8), high-dose aspirin (36 mg/kg, ip; 1 hour before procedure, n = 10), and M3mP6 HLPN plus high-dose aspirin (n = 6) treatments. (C) Comparison of saline control (n = 30), M3mP6 HLPN (10 μmol/kg; n = 18), clopidogrel (4 mg/kg; 2 hours before procedure, n = 8), M3mP6 HLPN plus clopidogrel (n = 8), ticagrelor (3 mg/kg; 2 hours before procedure, n = 13), and M3mP6 HLPN plus ticagrelor (n = 14) treatments. (D) Comparison of the saline control (n = 27), M3mP6 HLPN (n = 18), aspirin (n = 11), clopidogrel (n = 8), clopidogrel plus aspirin (n = 9), and M3mP6 HLPN plus clopidogrel plus aspirin (n = 10) treatments. (E) Comparison of the saline control (n = 29), M3mP6 HLPN (10 μmol/kg; n = 26), a high-dose cangrelor (30 μg/kg; n = 20), the high-dose cangrelor (30 μg/kg) plus M3mP6 HLPN (n = 15), a low-dose cangrelor (10 μg/kg; n = 17), and the low-dose cangrelor (10 μg/kg) plus M3mP6 HLPN (n = 14) treatments. Certain identical experimental data are used for different comparisons in different panels. *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; Mann-Whitney test.
Comparison of antithrombotic efficacy of M3mP6 HLPN and oral P2Y12 inhibitors
We further compared the antithrombotic effect of M3mP6 HLPN with the P2Y12 antagonist clopidogrel, which is a current standard-of-care antiplatelet drug more potent than aspirin. One-time oral administration of a loading dose of clopidogrel (4 mg/kg, 2 hours before procedure) delayed occlusive thrombus formation after FeCl3-induced injury but was less effective than M3mP6 HLPN (Fig. 4C). We also compared the antithrombotic effect of M3mP6 HLPN with the more potent oral P2Y12 antagonist, ticagrelor, which does not require hepatic conversion. One-time injection of M3mP6 HLPN had similar antithrombotic effects to a one-time ingestion of high-dose ticagrelor (3 mg/kg, 2 hours, oral) (Fig. 4C). We found that the median occlusion time in mice treated with M3mP6 HLPN together with clopidogrel (762 s) or ticagrelor (845 s), respectively, was longer than that of mice treated with either of the drugs alone (Fig. 4C) and was longer than the total median occlusion time calculated by adding that of the M3mP6 HLPN group with that of either the clopidogrel (660 s) or ticagrelor (692 s) groups (Fig. 4C), respectively. These data suggest that M3mP6 HLPN might have synergistic effects with clopidogrel and ticagrelor in inhibiting thrombosis.
Comparison of antithrombotic efficacy of M3mP6 HLPN and clopidogrel plus aspirin
Clopidogrel in combination with aspirin is the recommended antiplatelet treatment for patients with myocardial infarction. Thus, we investigated the effects of M3mP6 when combined with clopidogrel plus aspirin. One-time oral administration of a loading dose of clopidogrel combined with aspirin had a similar antithrombotic effect to clopidogrel alone and was significantly less effective in inhibiting occlusive thrombosis than one-time M3mP6 HLPN injection alone (P < 0.0001; Fig. 4D). M3mP6 HLPN in combination with aspirin and clopidogrel had a significantly enhanced antithrombotic effect compared with clopidogrel combined with aspirin (P < 0.0001; Fig. 4D), but was similar to the antithrombotic effect of M3mP6 HLPN combined with clopidogrel alone (Fig. 4D). These data suggest that M3mP6 HLPN might be superior to the combination of aspirin and clopidogrel in preventing arterial thrombosis and that M3mP6 HLPN might exert a synergistic antithrombotic effect when used in combination with clopidogrel and aspirin.
Comparison between M3mP6 HLPN and the intravenous P2Y12 inhibitor cangrelor
Recently, an intravenous direct P2Y12 inhibitor, cangrelor, was approved by the U.S. Food and Drug Administration for acute antithrombosis treatment, which not only exhibits a more potent antithrombotic effect than oral P2Y12 inhibitors but also causes more severe bleeding (23, 24). We compared the effects of a loading dose of cangrelor with M3mP6 HLPN in preventing occlusive carotid artery thrombosis induced by higher concentrations of FeCl3. Under these conditions, a one-time injection of a loading dose of cangrelor (30 μg/kg, retro-orbital) caused a variable inhibition of occlusive thrombosis in different individual mice or experiments. In the majority of mice tested, the effect on vessel occlusion was similar to that of M3mP6 HLPN (Fig. 4E). However, a small population of mice treated with cangrelor showed much longer occlusion time (Fig. 4E). Thus, the overall effect of cangrelor was moderately better than M3mP6 HLPN (Fig. 4E). M3mP6 HLPN combined with cangrelor yielded a longer median occlusion time than M3mP6 HLPN or cangrelor alone (Fig. 4E), which was longer than that calculated by adding the individual increases caused by M3mP6 HLPN and cangrelor alone (>900 s versus 857 s at 30 μg/kg cangrelor, and 791 s versus 688 s at 10 μg/kg cangrelor), suggesting a likely synergism between M3mP6 HLPN and cangrelor in antithrombotic efficacy.
Comparison between M3mP6 HLPN and current antiplatelet drugs in tail bleeding time analysis
A mouse tail bleeding time test was used to evaluate the effect of M3mP6 HLPN on hemostasis. M3mP6 HLPN, when injected at the same doses that potently inhibited occlusive thrombosis (Fig. 4A), showed no effects on mouse tail bleeding time (Fig. 5A). In contrast, the high-dose aspirin, which had a minimal antithrombotic effect in the FeCl3-induced thrombosis model (Fig. 4B), increased tail bleeding time (Fig. 5B). In addition, clopidogrel (Fig. 5C), clopidogrel/aspirin combination (Fig. 5D), ticagrelor (Fig. 5C), and cangrelor (Fig. 5E) increased tail bleeding time in mice. Furthermore, the bleeding time in mice treated with M3mP6 HLPN combined with aspirin, clopidogrel, ticagrelor, or cangrelor was similar to that in mice treated with each of these drugs alone (Fig. 5, A to E), again highlighting the benefit of M3mP6 HLPN in minimizing bleeding risk. Together, our data indicate that M3mP6 HLPN might be effective in potently inhibiting thrombosis without causing excessive bleeding.
(A) Comparison of saline control (n = 16), M3mP6 HLPN (10 μmol/kg; 15 min before procedure, n = 18), aspirin (4.3 mg/kg; oral 2 hours before procedure, n = 11), and M3mP6 HLPN plus aspirin (n = 11) treatments. (B) Comparison of saline control (n = 8), M3mP6 HLPN (5 μmol/kg; 15 min before procedure, n = 8), high-dose aspirin (36 mg/kg, ip; 1 hour before procedure, n = 11), and M3mP6 HLPN plus high-dose aspirin (n = 6) treatments. (C) Comparison of saline control (n = 30), M3mP6 HLPN (10 μmol/kg; n = 18), clopidogrel (4 mg/kg; 2 hours before procedure, n = 8), M3mP6 HLPN plus clopidogrel (n = 8), ticagrelor (3 mg/kg; 2 hours before procedure, n = 13), and M3mP6 HLPN plus ticagrelor (n = 14) treatments. (D) Comparison of saline control (n = 27), M3mP6 HLPN (n = 18), aspirin (n = 11), clopidogrel (n = 8), clopidogrel plus aspirin (n = 9), and M3mP6 HLPN plus clopidogrel plus aspirin (n = 10) treatments. (E) Comparison of saline control (n = 29), M3mP6 HLPN (10 μmol/kg; n = 25), a high-dose cangrelor (30 μg/kg; n = 17), the high-dose cangrelor (30 μg/kg) plus M3mP6 HLPN (n = 17), a low-dose cangrelor (10 μg/kg; n = 17), and the low-dose cangrelor (10 μg/kg) plus M3mP6 HLPN (n = 14) treatments. Certain identical data are used for different comparisons in different panels. ***P < 0.001 and ****P < 0.0001; Mann-Whitney test.
Comparison of M3mP6 HLPN with cangrelor in an artery perforation model of surgical bleeding
Clinically, the risk of hemorrhage during antiplatelet therapy is often associated with traumatic intervention, particularly intravascular intervention when perforation of a vascular wall is a necessity. To more closely mimic hemorrhage during intravascular/surgical procedures, we designed a carotid artery perforation model of surgical hemorrhage (Fig. 6A). In this model, the common carotid artery was surgically exposed and perforated with a needle. The site of perforation was immediately covered with a hemostatic pad (CERTI-GAUZE) to stop bleeding, replicating the procedure performed during vascular intervention. Blood absorbed on the pad was eluted and quantified to indicate amounts of hemorrhage. In normal controls, bleeding at the perforation site was quickly stopped by the hemostatic pad with the extravasation of a minimal amount of blood, but the loading dose of cangrelor caused excessive bleeding despite the use of the hemostatic pad (Fig. 6B). In contrast, M3mP6 HLPN–treated mice were not different from control mice (Fig. 6B), further demonstrating M3mP6 HLPN to be a safe antiplatelet drug that does not cause excessive bleeding under conditions mimicking an invasive vascular intervention. We used this model to verify that M3mP6 HLPN does not exacerbate the adverse bleeding effect of a P2Y12 antagonist. There was no difference in hemorrhage between groups administered cangrelor alone or cangrelor plus M3mP6 HLPN, either at high- or low-dose cangrelor (Fig. 6B). These data demonstrated that M3mP6 HLPN exerted a potent antiplatelet effect with minimal bleeding risk. Moreover, M3mP6 HLPN combined with cangrelor did not increase bleeding risk over that of cangrelor treatment alone.
(A) Illustration of a mouse surgical bleeding model. (B) Comparison of M3mP6 HLPN (10 μmol/kg; n = 20) with high-dose cangrelor (30 μg/kg; n = 16), high-dose cangrelor plus M3mP6 HLPN (n = 17), low-dose cangrelor (10 μg/kg; n = 11), low-dose cangrelor plus M3mP6 HLPN (n = 12), and normal control (n = 27) in the artery perforation model of surgical bleeding. (C) BMBT test of the effect of M3mP6 HLPN on hemostasis in dogs (n = 3). ****P < 0.0001; Mann-Whitney test in (B) and unpaired t test in (C); data presented as means ± SEM.
No effect of M3mP6 HLPN in a dog buccal mucosal bleeding time test
To determine whether M3mP6 may affect hemostasis in large animals, we used buccal mucosal bleeding time (BMBT), a routine bleeding time test, in three dogs before and during M3mP6 HLPN infusion. No differences in bleeding time were observed between pre- and postadministration of M3mP6 HLPN, both of which fell in the normal bleeding time range (<4 min in dogs) (Fig. 6C). These data showed that M3mP6 HLPN did not cause excessive bleeding not only in rodents but also in dogs.
Synergistic antithrombotic effect without excessive bleeding using a combination of M3mP6 HLPN with low-dose cangrelor
Our results showed that a loading dose of cangrelor not only had a potent antithrombotic efficacy but also markedly increased hemorrhage, which contrasted with M3mP6 HLPN’s selective inhibition of thrombosis without causing bleeding (Figs. 4E, 5E, and 6B). Our data also demonstrated that M3mP6 HLPN combined with this high dose of cangrelor enhanced antithrombotic efficacy compared with high-dose cangrelor alone (Fig. 4E), but appeared to have an adverse effect on bleeding similar to that of cangrelor alone (Figs. 5E and 6B). Thus, we hypothesized that the combination of M3mP6 HLPN with low-dose cangrelor may synergistically enhance antithrombotic efficacy but with reduced adverse effects on bleeding. M3mP6 HLPN in combination with low-dose cangrelor (10 μg/kg) had an antithrombotic effect that was greater than even the maximum clinical dose of cangrelor (Fig. 4E); however, tail bleeding time or bleeding in a carotid artery perforation model of surgical hemorrhage was not increased over control or either of the drugs individually (Figs. 5E and 6B). Thus, M3mP6 HLPN used together with low-dose cangrelor might be an optimal choice for acute antithrombotic treatment due to its powerful antithrombotic therapeutic effect and lack of increased bleeding risk.
Acute postischemic antithrombotic effects of M3mP6 HLPN in comparison with cangrelor
An advantage for intravenous antithrombotics is fast action, which makes them suitable for emergency treatment of acute thrombosis. We thus compared the acute antithrombotic effect of cangrelor with M3mP6 HLPN under experimental conditions mimicking a thrombotic emergency, where the drugs were retro-orbitally injected after the initiation of carotid arterial thrombosis. There was no difference between M3mP6 (10 μmol/kg) and a loading dose of cangrelor (30 μg/kg) in antithrombotic effect under these conditions (Fig. 7A). However, cangrelor but not M3mP6 HLPN showed a prolonged tail bleeding time (Fig. 7B), indicating the superior bleeding safety of M3mP6 HLPN.
(A) Effects of M3mP6 HLPN (10 μmol/kg) and cangrelor (30 μg/kg) on occlusive thrombosis when injected 5 min after FeCl3 (3.75%)–induced carotid artery injury. (B) Tail bleeding time of mice as described in (A). (A and B) n = 7 for normal control, n = 9 for M3mP6 HLPN, and n = 6 for cangrelor. (C) Comparison of the median integrated platelet fluorescence intensity (DyLight 649–labeled anti-GPIbβ) at the sites of laser-induced injury in cremaster arterioles of control (black), M3mP6 HLPN (10 μmol/kg)–treated (red), and cangrelor (30 μg/kg)–treated (blue) mice (27 injury sites each for control and M3mP6- and cangrelor-treated groups). (D) Comparison of the median integrated fibrin fluorescence intensity (Alexa Fluor 488–labeled anti-fibrin) at the sites of the laser-induced injury in cremaster arterioles of control (black), M3mP6 HLPN (10 μmol/kg)–treated (red), and cangrelor (30 μg/kg)–treated (blue) mice. (E and F) Quantification of total fluorescence as detected over time in (C) for platelet thrombus (E) and in (D) for fibrin (F) (27 injury sites each for control and M3mP6- and cangrelor-treated groups). *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001; Mann-Whitney test.
Effects of M3mP6 HLPN and cangrelor on platelet thrombus formation and intravascular coagulation using a laser-induced cremaster arteriolar thrombosis model in mice
Recent studies demonstrated that integrin outside-in signaling plays an important role not only in platelet thrombus formation but also in intravascular coagulation under flow shear (25). Current antiplatelet drugs are ineffective in inhibiting intravascular coagulation (26, 27). Thus, we further tested the effects of M3mP6 HLPN compared with cangrelor in inhibiting both platelet thrombus formation and intravascular coagulation using a laser-induced cremaster arterial thrombosis model. When pre-injected intravenously, M3mP6 HLPN (10 μmol/kg) was similar to a loading dose of cangrelor (30 μg/kg) in potently inhibiting platelet thrombus formation (Fig. 7, C and E). However, M3mP6 HLPN also almost completely inhibited intravascular fibrin clot formation at the site of vascular injury, whereas cangrelor had only a moderate partial effect (Fig. 7, D and F). Thus, M3mP6 HLPN was effective in inhibiting not only platelet thrombus formation but also intravascular coagulation in vivo, and this effect might be superior to that achieved by the most potent P2Y12 inhibitor cangrelor.
Treatment of myocardial ischemia–reperfusion injury with M3mP6 HLPN
The prevailing treatment for myocardial infarction/ischemia (MI) is to perform surgical or percutaneous coronary interventions to invasively reperfuse the occluded artery. Reperfusion of ischemic tissues, however, may cause myocardial ischemia–reperfusion (MI/R) injury, where an acute thrombo-inflammatory reaction of the ischemic tissue occurs upon reexposure to oxygenated blood, resulting in damage to cardiac function and death. To evaluate the potential therapeutic effect of M3mP6 HLPN in treating MI and MI/R injury, a severe MI was induced in mice by ligating the left anterior descending (LAD) branch coronary artery for 45 min followed by reperfusion. To mimic the clinical setting in which antithrombotic therapies are provided during reperfusion, M3mP6 or control HLPN were injected postischemically 35 min after the induction of MI (Fig. 8A). Compared with the control group, the M3mP6 HLPN treatment group showed lower infarct area/risk area ratio (Fig. 8, B and C) but similar risk area/total area ratio (Fig. 8D) as indicated by triphenyltetrazolium chloride (TTC)/Evans blue staining, suggesting reduced infarction under similar ischemia/reperfusion assault. M3mP6 HLPN also prevented damage of cardiac function as indicated by echocardiography performed at 24 hours after the procedure (Fig. 8, E and F). Immunohistochemistry indicated that MI/R-induced microvascular thrombosis (Fig. 8G) and neutrophil infiltration (Fig. 8H) in the reperfused cardiac tissue were both reduced by the M3mP6 HLPN treatment. M3mP6 also reduced plasma myeloperoxidase (MPO) concentrations (Fig. 8I), which is an indicator of neutrophil activation, in MI/R mice. In addition, M3mP6 greatly reduced mortality rate during the 7-day postprocedure monitoring (Fig. 8J). These data suggest that M3mP6 was an effective treatment of MI/R-induced thrombosis/inflammation and cardiac injury in the mouse model.
(A) Schematic protocol of MI/R study. iv, intravenous. (B) Representative images of heart sections from mice treated with M3mP6 or scrambled peptide HLPN (Evans blue/TTC stain). (C) Quantification of the infarct area (white) as percentage of the area at risk (white and red) as shown in (A) (n = 4 for each group, t test, ✽P < 0.05). (D) Quantification of risk area as percentage of the entire heart section (n = 4 for each group, t test, P > 0.05). (E) Representative M-mode long-axis echo images for (i) sham control, (ii) MI/R treated with scrambled peptide HLPN, and (iii) MI/R treated with M3mP6 HLPN. (F) Mouse left ventricle ejection fraction (n = 11 for the sham and M3mP6 groups, and n = 10 for the scrambled peptide group; mean ± SEM, one-way ANOVA). (G) Immunohistochemistry staining of platelets in mouse heart sections 24 hours after MI/R using rat anti-integrin αIIb antibody. The representative sections are from (i) sham control, (ii) MI/R treated with scrambled peptide HLPN, and (iii) MI/R treated with M3mP6 HLPN. (iv) αIIb positive staining was quantified using ImageJ and presented as percentage of total tissue area (n = 9 for sham and n = 10 for both M3mP6 and scrambled peptide groups; mean ± SEM, Mann-Whitney test). (H) Immunohistochemistry staining of neutrophils in mouse heart sections 24 hours after MI/R using rat anti-Ly6G antibody. The representative sections are from (i) sham control, (ii) MI/R treated with scrambled peptide HLPN, and (iii) MI/R treated with M3mP6 HLPN. (iv) Ly6G-positive cells/mm2 were counted using ImageJ software (n = 6 for sham, n = 22 for scrambled peptide, and n = 12 for M3mP6; mean ± SEM, Mann-Whitney test, **P < 0.01). (I) Mouse plasma MPO concentrations 24 hours after MI/R (sham, n = 6; scrambled HLPN treatment, n = 11; M3mP6 HLPN treatment, n = 9; mean ± SEM, t test, ✽P < 0.05). (J) Kaplan-Meier survival curve of mice 7 days after MI/R surgery treated with M3mP6 HLPN or scrambled peptide HLPN (n = 12 for M3mP6 treatment, n = 14 for scrambled peptide treatment). Sham surgery caused no death in six tested mice.
DISCUSSION
Three important advances are reported in this study: (i) We have developed lipid-stabilized HLPN for efficient delivery of an ExE motif peptide (M3mP6) in vivo into the intracellular compartment for therapeutic use; (ii) we showed that M3mP6 HLPN potently inhibited occlusive thrombosis without observable adverse effect on hemorrhage in mice and dogs; and (iii) postischemia injection of M3mP6 HLPN inhibited microvascular thrombosis/inflammation and improved cardiac function and survival in a mouse model of MI/R injury. This compound not only showed antithrombotic synergism with the current standard-of-care antiplatelet drugs (P2Y12 inhibitors) without increasing bleeding risk but also demonstrated a marked therapeutic effect in the treatment of ischemia/reperfusion injury after MI in animal models. Thus, M3mP6 HLPN has the potential to improve the current antiplatelet therapy in the treatment of myocardial infarction while concomitantly improving the safety profile.
Despite having advantages of high specificity due to unique amino acid sequences and low risk of toxic nonspecific effects, development of peptide-based drugs targeting intracellular compartments has lagged (1, 2). Among the major reasons is the lack of technology to efficiently deliver sufficient amounts of peptides into cells in vivo. Thus, facilitating intracellular delivery of peptides should accelerate the development and clinical use of peptide-based therapeutic agents. Although numerous tools have been developed to enable peptides to cross cell membranes, none have been completely satisfactory for therapeutic use (1, 2). Among them, lipidization is effective to render peptides membrane permeable. However, the myristoylated peptide M3mP6, although membrane permeable and able to inhibit platelet function in vitro, was ineffective in vivo. Thus, membrane permeability alone is not sufficient to render peptides effective in vivo. Liposomes and lipid micelles have been relatively more efficient and often used for drug delivery in vivo (2). However, these approaches usually require large amounts of lipid to encapsulate a limited amount of peptide, thus necessitating unusually high peptide affinity or avidity for its targets. For example, a lipid micellar formulation of the synthetic peptide mP6 contains a maximal 4% (mol/mol) peptide, yielding a maximal concentration of mP6 of <1 mM, which is insufficient for clinical use (21). The lipid-stabilized M3mP6 HLPN in the present study contains up to 70% (mol/mol) peptide, achieving a high peptide concentration of >10 mM, suitable for bolus intravenous injection in humans. The lipid-stabilized HLPN also facilitates peptide drug entry into cells, resulting in much less peptide being required for comparable antiplatelet efficacy compared with the peptide alone. These superior characteristics rendered M3mP6 HLPN a potent antithrombotic in vivo. Thus, our data demonstrate that the lipid-stabilized HLPN improves efficiency of intracellular delivery of a peptide-based drug in vivo, enabling effective in vivo use of peptide drugs having relatively moderate affinity for their intracellular targets. This approach may be useful for facilitating development of peptide-based drugs targeting the intracellular compartment for therapeutic use.
Circulating platelets are normally in a resting state and become activated only when exposed to platelet agonists. Platelet agonists elicit platelet activation via various receptor-mediated intracellular signaling pathways (28, 29). These intracellular signals converge as an “inside-out” signal to transform αIIbβ3 from a “resting” state to an “activated” state (8, 30). This enables integrin’s ligand binding function, a requirement for platelet adhesion and aggregation (12). Current antiplatelet drugs either inhibit platelet activation signaling pathways, leading to integrin activation, or directly inhibit the ligand binding function of the integrin αIIbβ3 (10, 12). Although these drugs have become cornerstones in the treatment of patients with cardiovascular diseases (6, 12), they also inhibit hemostasis, causing excessive bleeding, which can be life-threatening (11–14). Bleeding effects of these drugs also limit the use of optimal doses and, therefore, their antithrombotic efficacy. Bleeding in thrombotic patients receiving a percutaneous coronary intervention is associated with poor prognosis and increased mortality (11, 15–17). Hence, the balance of optimal platelet inhibition with minimal bleeding complications is an important current clinical challenge, particularly for patients requiring traumatic procedures or suffering from thrombotic stroke, in which hemorrhage is life-threatening. Recently, various thrombin receptor inhibitors were shown to have reduced adverse effects on hemorrhage in animal models (19, 20). However, this reduction in adverse effect on hemorrhage was measured relative to clopidogrel and was only partial (19, 20). Animals administered thrombin receptor inhibitors still had increased bleeding compared with normal controls (19, 20). In contrast, we observed no difference in bleeding between normal controls and animals administered M3mP6 HLPN. Clinical trials of the PAR1 inhibitor vorapaxar reported excessive bleeding in human patients (19), suggesting that this drug still poses a clinically relevant hemorrhagic risk. The adverse hemorrhagic effects of these drugs are consistent with them inhibiting the activation of the ligand binding function of integrin αIIbβ3 and are thus not fundamentally different from the current antiplatelet drugs (29). The compound described here is based on a different concept of selectively targeting integrin outside-in signaling without directly affecting the ligand binding function of the platelet integrin αIIbβ3. Thus, M3mP6 permits primary platelet adhesion and aggregation to occur but prevents outside-in signaling–mediated secondary thrombus expansion and consequent vascular occlusion (21, 22). As primary platelet aggregation in the collagen-rich tissues outside the vascular lumen is likely to be robust, the secondary outside-in amplification signals are unlikely to be required for hemostasis in most hemorrhagic situations but are critical during thrombosis intravascularly. Furthermore, outside-in signaling is important for activating platelet procoagulant activity, which is selectively required for intravascular coagulation and thrombus stability in the presence of blood flow shear (25), providing a second mechanism by which selectively blocking outside-in signaling does not affect hemostasis but inhibits occlusive intravascular thrombi. Thus, our combination of an innovative HLPN delivery system with a peptide-based outside-in signaling inhibitor allowed us to demonstrate the ability of M3mP6 HLPN to inhibit occlusive thrombosis without observable effects on hemostasis using three different bleeding assays in mice and dogs.
M3mP6 HLPN is not only itself a potent antithrombotic, but it also exerts a synergistic effect when used in combination with the current standard of care, P2Y12 inhibitors with or without aspirin. This synergistic effect occurred with clopidogrel and with the more potent new direct P2Y12 inhibitors ticagrelor and cangrelor, even at the highest clinical doses. These synergistic effects suggest that the integrin-dependent secondary amplification of thrombus formation requires release of known agonists such as ADP and TXA2 and also activation of additional signaling pathways that are not inhibited by ADP receptor antagonists or inhibitors of the TXA2 synthesis pathway. M3mP6 HLPN did not increase the bleeding risks when combined with aspirin and P2Y12 inhibitors, suggesting that this synergistic effect has the potential to increase antithrombotic efficacy to an extent that is not currently possible with antiplatelet therapy without bleeding risk. Consistent with this notion, we showed that the combination of M3mP6 with low-dose cangrelor not only enhanced antithrombotic efficacy beyond that achievable with the highest clinical dose of cangrelor but also abolished the bleeding risk associated with high-dose cangrelor in mice. Thus, combined use of M3mP6 HLPN with low doses of a P2Y12 inhibitor represents a potential antithrombotic strategy that might have important implications for improving clinical outcomes. Our data further suggest that this compound might be useful for treating acute MI/R injury. Postischemia injection of M3mP6 HLPN inhibited both microvascular thrombosis and inflammation in reperfused cardiac tissue, rescued cardiac function, and improved survival in a mouse model of MI/R. It is important to note that the therapeutic effect of M3mP6 HLPN on MI/R injury may not only necessarily be due specifically to its antiplatelet effect, but it is also possible that this drug may affect integrin signaling in other cells such as leukocytes. Both thrombosis and leukocyte-mediated inflammation (31) play important roles in MI/R injury.
A major limitation of our study is that our in vivo data are primarily obtained in mice, which show not only similarities with humans in integrin outside-in signaling and β3 cytoplasmic domain sequence but also differences in thrombosis and inflammation. In addition, M3mP6 HLPN has only been shown to be effective intravenously and has a relatively short half-life in blood, which limits its out-of-hospital use or prolonged use. Thus, M3mP6 in current formulation is not for routine prevention of thrombosis in place of current oral antiplatelet drugs. However, M3mP6 HLPN is fast acting, has a relatively longer in-plasma half-life (~3.5 hours in rats) than cangrelor (3 to 6 min), and thus might be suitable for bolus intravenous injection in emergency in patients suffering from acute myocardial infarction. Future preclinical safety studies and clinical trials should demonstrate whether the compound and the strategy herein are suitable for use in patients.
In summary, we have engineered lipid-stabilized HLPN to efficiently deliver an inhibitor peptide of integrin outside-in signaling in vivo, resulting in a fast-acting, reversible, and potent antithrombotic compound with minimal bleeding risk in mice. Postischemic injection of this compound effectively reduces MI/R injury and mortality in mice. Furthermore, we developed an antiplatelet strategy in which M3mP6 HLPN in combination with a low dose of current antiplatelet drugs enhances antithrombotic efficacy while minimizing bleeding risk in mice.
MATERIALS AND METHODS
Study design
This study aimed at developing HLPN for in vivo delivery of an inhibitor peptide, M3mP6, to its intracellular target for treating thrombosis and MI/R injury. We performed a cohort of experiments using multiple in vitro and in vivo approaches to characterize the properties and antiplatelet/antithrombotic/hemorrhagic effects of M3mP6 HLPN, the details of which are described in the Supplementary Materials and Methods and in previous publications (21, 32–35). We compared the antithrombotic effects of M3mP6 HLPN with current antiplatelet drugs using the FeCl3-induced mouse carotid artery thrombosis (34, 36) and laser-induced mouse cremaster arterial thrombosis models (37). We compared the bleeding effects of current antiplatelet drugs with M3mP6 HLPN using mouse tail bleeding time (21, 38), mouse carotid artery perforation bleeding analysis (we developed in this study), and dog BMBT test (39). In all these experiments, mice of either sexes (8 to 10 weeks of age) were randomly selected into control and different treatment groups. Per animal protection requirements and when scientifically valid, data collected in identical in vivo experiments were used for multiple purposes to avoid unnecessary mouse sacrifice. The thrombosis and bleeding tests were not blinded because of the obviousness of the differences in occlusion times between M3mP6 HLPN and control or aspirin-treated groups, and in bleeding between M3mP6 HLPN and P2Y12 antagonists during procedures. The mouse LAD artery ligation model (40, 41) was used to study the effect of M3mP6 HLPN in treating MI/R injury, and this study was blinded by concealing the identities of treatment groups to the operator. The confidence and sample sizes of key experiments were evaluated by Power analysis using the GPower 3.1 software (42) (see source files for details). For in vitro experiments, sample size prediction was performed on the basis of our previous experience with n = 3 to 6 (21, 25). To determine the suitability of controls used for our studies, we compared HLPN of scrambled peptide and AAA mutant peptide with physiological saline to show their similarity (fig. S4).
Statistics
Sample normality and distribution tests were performed using GraphPad PRISM 5.0. For parametric data, statistical significance was analyzed using Student’s t test (two-group comparison) or analysis of variance (ANOVA; multiple-group comparison) using GraphPad PRISM 5.0. For nonparametric data, statistical significance was determined using Mann-Whitney test. Power analysis was performed using G-Power 3.1 software. Survival analysis was performed using the Kaplan-Meier method. A P value <0.05 was considered as significant.
SUPPLEMENTARY MATERIALS
stm.sciencemag.org/cgi/content/full/12/552/eaaz7287/DC1
Supplementary Materials and Methods
Fig. S1. Stability of M3mP6 HLPN during storage.
Fig. S2. Effect of M3mP6 peptide and HLPN on JonA and fibrinogen binding to platelets and platelet aggregation induced by PAR4AP.
Fig. S3. Uptake of fluorescently labeled M3mP6 HLPN into mouse platelets.
Fig. S4. Comparison of the effects of M3mP6 HLPN, scrambled control peptide (Scra) HLPN, AAA mutant peptide (Myr-FAAARL) HLPN, and physiological saline solution on 7.5% FeCl3–induced carotid artery thrombosis.
Fig. S5. Dose-response analysis of tail vein–injected M3mP6 HLPN in occlusive thrombosis using the 7.5% FeCl3–induced carotid artery thrombosis model.
Fig. S6. Comparison of effects of 50% versus 36.8% M3mP6 HLPN on FeCl3-induced thrombosis.
Table S1. Summary of toxicological studies of M3mP6 HLPN.
Table S2. M3mP6 peptide pharmacokinetic parameters.
Movie S1. Laser-induced cremaster arteriolar thrombosis after saline injection as indicated by platelet (red) and fibrin (green) stains as described in Methods.
Movie S2. Laser-induced cremaster arteriolar thrombosis 15 min after bolus injection of M3mP6 HLPN as indicated by platelet (red) and fibrin (green) stains.
Movie S3. Laser-induced cremaster arteriolar thrombosis immediately after injection of cangrelor as indicated by platelet (red) and fibrin (green) stains.
Data file S1. Raw data.
This is an article distributed under the terms of the Science Journals Default License.
REFERENCES AND NOTES
- Copyright © 2020 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works
