1521-0081/70/3/526–548$35.00 https://doi.org/10.1124/pr.117.014530PHARMACOLOGICAL REVIEWS Pharmacol Rev 70:526–548, July 2018Copyright © 2018 by The American Society for Pharmacology and Experimental Therapeutics

ASSOCIATE EDITOR: LORI L. ISOM

Platelet Signaling and Disease: Targeted Therapy forThrombosis and Other Related Diseases

Jennifer Yeung, Wenjie Li, and Michael Holinstat

Departments of Pharmacology (J.Y., W.L., M.H.) and Internal Medicine, Division of Cardiovascular Medicine (M.H.), University ofMichigan, Ann Arbor, Michigan

Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 526I. Introduction. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527II. General Function of Platelets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 527III. Platelet Role in Disease . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 528

A. The Role of Platelets in Arterial Thrombosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 529B. The Role of Platelets in Venous Thrombosis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530C. The Role of Platelets in Immune Responses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 530D. Platelets in Cancer Metastasis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531

IV. Regulators of Platelet Activation and Signaling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531A. G Protein–Coupled Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 531

1. Protease-Activated Receptors 1 and 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5322. Purinergic Receptors P2Y1 and P2Y12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5333. Prostaglandin Receptors TPa, IP, EP, and DP . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 533

B. Integrin Receptors aIIbb3, a2b1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 534C. Immunoreceptor Tyrosine-Based Activation Motif Receptors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535D. Enzymes Targeted for Regulation of Platelet Function. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 535

1. Cyclooxygenase-1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5352. 12-Lipoxygenases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5353. Cyclic Nucleotide Phosphodiesterases. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536

V. Platelet Pharmacological Targets and Interventions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536A. Oxygenase Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536B. ADP Receptor Inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 537C. Integrin aIIbb3 Receptor Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 540D. Collagen Receptor Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 541E. Thrombin Receptor Inhibitors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 541F. Glycoprotein Ib-IX-V Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 542G. IP Receptor Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543H. Phosphodiesterase Inhibitors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 543

VI. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544

Abstract——Platelets are essential for clotting in theblood and maintenance of normal hemostasis. Underpathologic conditions such as atherosclerosis, vascularinjury often results in hyperactive platelet activation,resulting in occlusive thrombus formation, myocardialinfarction, and stroke. Recent work in the field haselucidated a number of platelet functions unique from

that of maintaining hemostasis, including regulation oftumor growth and metastasis, inflammation, infection,and immune response. Traditional therapeutic targetsfor inhibiting platelet activation have primarily beenlimited to cyclooxygenase-1, integrin aIIbb3, and theP2Y12 receptor. Recently identified signaling pathwaysregulating platelet function have made it possible to

This work was supported by National Institutes of Health [Grants HL114405 (to M.H.), GM105671 (to M.H.), and HL129481 (to J.Y.)]. Thecontent is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Address correspondence to: Dr. Michael Holinstat, Department of Pharmacology, University of Michigan, 1150 West Medical CenterDrive, Room 2220D MSRB III, Ann Arbor, MI 48109. E-mail: mholinst@umich.edu

https://doi.org/10.1124/pr.117.014530.

526

by guest on June 23, 2021D

ownloaded from

https://doi.org/10.1124/pr.117.014530mailto:mholinst@umich.eduhttps://doi.org/10.1124/pr.117.014530

develop novel approaches for pharmacologicalintervention in the blood to limit platelet reactivity.In this review, we cover the newly discovered roles forplatelets as well as their role in hemostasis andthrombosis. These new roles for platelets lendimportance to the development of new therapies

targeted to the platelet. Additionally, we highlight thepromising receptor and enzymatic targets that mayfurther decrease platelet activation andhelp to addressthe myriad of pathologic conditions now known toinvolve platelets without significant effects onhemostasis.

I. Introduction

Platelets are anucleate cells produced by megakaryo-cytes in the bone marrow and lungs (Weyrich andZimmerman, 2013; Lefrançais et al., 2017) existing inthe vessel for 5–10 days before they are removed fromcirculation by the spleen (Kaplan and Saba, 1978;Kuter, 1996). Their function in the body has predomi-nantly been linked to maintaining normal flow in theblood vessel through a process known as hemostasis.When the vessel integrity is challenged, however, eitherthrough vessel injury, atherosclerotic plaque rupture,or chronic inflammatory conditions, platelets respondby clotting to form a thrombus at the site of injury. Inaddition to normal hemostasis, platelet activation oftenresults in the formation of an occlusive thrombusleading to myocardial infarction and stroke. Due tothe high turnover of the platelet in the body and itsessential role in hemostasis and thrombosis, the plate-let has long been a primary target for therapeuticintervention for the prevention of occlusive thromboticevents. This review is focused on delineating ourcurrent understanding of the roles that platelets playin both physiologic and pathophysiological conditions,the various potential drug targets expressed in theplatelet or on its surface, and how classic and newlydeveloped therapeutics have taken advantage of thesetargets to limit platelet activation in a number ofpathophysiological conditions.

II. General Function of Platelets

Platelets are thought to be the primary mediators ofhemostasis and thrombosis (Semple and Freedman,2010). Under physiologic conditions, platelets circulatein the blood to maintain the blood constituents withinthe vessel (Jackson, 2011; Holinstat, 2017; Tomaiuoloet al., 2017) (Fig. 1). Due to the biologic and physicalproperties of the blood, including the numerous blood

cells and plasma (white blood cells, red blood cells,platelets, plasma constituents) and the shear forceinside the blood vessel, platelets are physically excludedfrom the central flow of the vessel and as a result areprimarily found near the vessel wall. The location ofplatelets in the vessel due to this physical constraintenables the platelets to play a principal role in thequick hemostatic response following a vascular injury(Holinstat, 2017). When a vascular insult or injuryoccurs, platelets initially tether to the subendothelialextracellular matrix (ECM) through multiple receptors,including the collagen receptors a2b1 and glycoprotein(GP)VI and von Willebrand factor (vWF) receptorglycoprotein receptor Ib-V-IX (Holinstat, 2017). Follow-ing firm adhesion to the subendothelial ECM, plateletsundergo spreading, activation, and eventual aggrega-tion to form a thrombus. Activated platelets also releasegranules or signals to aid in the recruitment andactivation of nearby platelets to the localized thrombus.

Secondary activation via granule secretion and oxy-genase catalysis from cyclooxygenase (COX)-1 and12-lipoxygenase (LOX) is mediated through autocrinepositive feedback on the platelet and paracrine signal-ing that stimulates circulating or loosely bound plate-lets to integrate into the existing clot to form anirreversible platelet plug at the site of injury. Althoughthe platelet contains several types of granules that playunique roles in regulating platelet activity [dense (d), a,and lysosomal], it is the dense granule that releasessmall molecules such as ATP, ADP, epinephrine, andserotonin, which play a predominant role in granule-dependent paracrine activation of the surroundingplatelets in the blood. The most highly investigatedsmall molecule released from the dense granule is ADP,which further signals through the platelet purinergicreceptors P2Y1 and P2Y12, and it is this pathway that istargeted clinically for prevention of occlusive thrombo-sis. Similar to granule secretion, following plateletactivation, the bioactive lipid products of free fatty

ABBREVIATIONS: 12(S)-HETE, 12(S)-hydroxyeicosatetraenoic acid; AA, arachidonic acid; ACS, acute coronary syndrome; ASA,acetylsalicylic acid; BPS, beraprost sodium; CLEC-2, C-type lectin-like receptor 2; COX, cyclooxygenase; cPLA2, cytocolic phosholipase A2;CYP450, cytochrome P450; DAG, diacylglycerol; DGLA, dihomo-g-linolenic acid; ECM, extracellular matrix; EHEC, enterohemorrhagicEscherichia coli; EPA, eicospentaenoic acid; FcgR, Fc receptor g-chain; FDA, Food and Drug Administration; GP, glycoprotein; GPCR, Gprotein–coupled receptor; GPI, glycosylphosphatidylinositol; HCMV, human cytomegalovirus; HIT, heparin-induced thrombocytopenia; HUS,hemolytic-uremic syndrome; IP3, inositol triphosphate; ITAM, immunoreceptor tyrosine-based activation motif; LAT, linker for activatedT cells; LOX, lipoxygenase; LPS, lipolysaccharide; MI, myocardial infarction; MP, microparticle; NET, neutrophil extracellular trap; PAD,peripheral arterial disease; PAH, pulmonary arterial hypertension; PAR, protease-activated receptor; PCI, percutaneous coronaryintervention; PDE, phosphodiesterase; PDPN, podoplanin; PF4, platelet factor 4; PG, prostaglandin; PI3K, phosphoinositide 3-kinase;PLC, phospholipase C; PUFA, polyunsaturated fatty acid; TIA, transient ischemic attack; TLR, Toll-like receptor; TTP, thromboticthrombocytopenic purpura; TxA2, thromboxane A2; UA/NSTEMI, unstable angina/non-ST elevation myocardial infarction; VEGF, vascularendothelial growth factor; VTE, venous thromboembolism; vWF, von Willebrand factor.

Platelet Therapeutics 527

acids from arachidonic acid (AA), including prostaglan-dins (PG)E2 and thromboxane A2 (TxA2), formed viaCOX-1 and eicosanoids [12(S)-hydroxyeicosatetraenoicacid (12(S)-HETE)], produced via 12-LOX, interact withtheir respective G protein–coupled receptors (GPCRs)to reinforce platelet activation (Holinstat, 2017). Thiscomplex process of platelet adhesion followed by aggre-gation and recruitment of platelets to the site of injury isbest exemplified by recent studies showing a thrombuscomposed of an outer loose shell overlaying the denselypacked inner core. P-selectin–positive platelets arelocalized to the inner core area of a thrombus at thesite of injury and are tightly packed together in anirreversible clot, whereas the platelets located in theshell of the thrombus are more sensitive to the inhibi-tion of positive feedback signaling through the throm-boxane receptor (TPa) and the ADP receptor (P2Y12)(Stalker et al., 2013; Welsh et al., 2014, 2016).Beyond the well-established role for platelet regula-

tion of hemostasis and thrombosis, it has been proposedthat platelets have distinctive roles in regulating andassisting immune responses and inflammatory reac-tions (von Hundelshausen and Weber, 2007). Theadhesive molecule P-selectin that is highly expressedon the platelet surface following activation regulatesthe extent of interaction between activated plateletsand P-selectin glycoprotein ligand-1–expressing im-mune cells (lymphocytes, neutrophils, and monocytes),leading to micro-aggregate formation and leukocyterolling and arrest (Diacovo et al., 1996a,b). Otherplatelet surface receptors, including aIIbb3 (Oki et al.,2006), CD40 ligand (Prasad et al., 2003), intercellularadhesion molecule 2 (Diacovo et al., 1994), junctionaladhesion molecules (Prota et al., 2003), and chemokinereceptors (Murphy et al., 2000), have been proposed tolink the immune system to platelets. Additionally, allnine Toll-like receptors (TLRs) were found to beexpressed on platelets (Holinstat and Tourdot, 2015;Koupenova et al., 2015). These receptors are known to

be involved in innate immunity against viral, bacterialinfection, and even tumors, providing evidence thatplatelets are an integral component of the immunereporting system. To this end, platelets have previouslybeen reported to exhibit autophagy in activation(Ouseph et al., 2015). Additionally, platelets are knownto sample the blood environment and assist otherimmune cells by presenting foreign pathogens (Assinger,2014; Holinstat, 2017; Koupenova et al., 2018). Hence,platelet whose function was once thought to be limitedto maintaining the vessel integrity through regulationof hemostasis and thrombosis is now known to partic-ipate in a number of additional functions, includingimmunity and inflammation. As we continue to inves-tigate the ever-expanding role of the platelet in circu-lation, it is likely that its role in the body will likewiseexpand to include regulation of tumor growth andmetastasis and signaling distal tissue beds in thevascular tree through microparticle (MP) communica-tion (Gastpar, 1977; Jurasz et al., 2004; Boilard et al.,2010; Italiano et al., 2010).

III. Platelet Role in Disease

The pathophysiological mechanisms underlying theformation of arterial and venous thrombi are distinct.Arterial clots are formed under high shear stress,typically after rupture of an atherosclerotic plaque orother damage to the blood vessel wall. They are deemedplatelet-rich or white clots and are generally treatedwith antiplatelet drugs (see Platelet PharmacologicalTargets and Interventions). In contrast, venous throm-boses are largely fibrin-rich or red clots formed underlow shear stress on the surface of largely intactendothelium, and anticoagulants are clinically used totreat patients with venous thromboembolism (VTE).There is accumulating evidence that whereas thevenous thrombotic clot is a condition rich in red bloodcells, platelets represent a major component in the

Fig. 1. The physiologic and pathophysiological roles of platelets. Platelets participate in hemostasis to prevent blood loss by forming a hemostatic plugfollowing a vascular insult. In contrast, platelets can also partake in arterial and venous thrombosis, increasing the likelihood of vessel occlusion. Thepathophysiological mechanisms of arterial and venous thrombus formation are distinct by which arterial thrombosis normally occurs following anatherosclerotic plaque rupture, leading to damaged endothelial cells, whereas, in venous thrombosis, the endothelial cells remained intact.

528 Yeung et al.

development of VTE (Fig. 1) as well as their activeparticipation in immune response to foreign substances,including drug, bacteria, and viruses (Fig. 2).There is also a growing body of evidence that activated

platelets contribute to other pathophysiological diseasestates, including arterial and venous thrombosis, heparin-induced thrombocytopenia (HIT), cancer, and sickle celldisease (Italiano et al., 2010), through shedding of MPs orvesicular membrane fragments. Platelet MPs are highlyheterogeneous in size, ranging from 0.05 to 1 mm, organ-elle content, and phosphatidylserine surface expression(Italiano et al., 2010; Boilard et al., 2015). Interestingly,several studies have shown platelet MPs can engulf aswell as transfer their contents (proteins or nucleic acids) tonearby cells, most notably tumor cells (Best et al., 2015;Michael et al., 2017). Depending on the contents, plateletMPs can either suppress or enhance the metastaticcapabilities of tumor cells (Dashevsky et al., 2009;Varon et al., 2012; Michael et al., 2017). Despite theirapparent participation in these physiologic or patho-physiological processes, the fundamental aspects un-derlying their mechanisms remain largely unexplored.Currently, there are ongoing efforts to improve themethodologies in characterizing the diversity of MPs toaid in platelet-related disease biopsies or diagnosis.This section will detail the various pathophysiological

conditions in which platelets play an important role andfor which antiplatelet drugs are likely to represent asignificant treatment option in the patient.

A. The Role of Platelets in Arterial Thrombosis

The pathogenesis of arterial thrombosis is complexand dynamic. Arterial thrombosis is initiated followingendothelial damage from vascular injury or pathologicatherosclerotic plaque rupture under conditions of highshear rates (Fig. 1). Circulating platelets are rapidlydecelerated and transiently interact with the damagedand exposed subendothelial connective tissue containingimmobilized vWF bound to collagen (types I, III, and VI).vWF-collagen interacts with the platelet glycoproteinreceptor Ib-V-IX complex, allowingplatelets to translocatealong the vessel wall and engage their receptors, GPVIand integrin a2b1, with the subendothelial fibrillar colla-gen. This firm interaction also facilitates platelets toassociate with fibronectin through its engagement withintegrin a5b1.

The engagement of platelet receptors with collagenalso induces an inside-out cellular signaling cascadethat leads to integrin aIIbb3 activation (Coller andShattil, 2008; Shattil et al., 2010). This process involvesintermediary proteins, talin and kindlin, which bind tothe cytoplasmic domain of b3 integrin to shift aIIbb3

Fig. 2. The role of platelets in immune response. Platelets can directly or indirectly interact with components of the virus, bacteria, or even drugs suchas heparin to induce platelet activation and interaction with neutrophils or other phagocytes, resulting in neutrophil–platelet aggregate formation orthrombocytopenia.

Platelet Therapeutics 529

from an inactive to active state (Tadokoro et al., 2003;Moser et al., 2008). Active aIIbb3 conformation increasesits affinity for adhesive proteins, vWF, fibrinogen,fibrin, and fibronectin. These interactions are essentialfor the platelets to form stable aggregates with otheractivated platelets to promote thrombus growth.Following firm adhesion and aggregate formation,

platelets also release or locally generate soluble ago-nists, such as ADP, TxA2, and thrombin, to mediatefeed-forward autocrine and paracrine platelet activa-tion via their respective GPCRs. The activation ofGPCRs initiates a series of intracellular signalingevents, including generation of second messengers[diacylglycerol (DAG) and inositol triphosphate (IP3)].Eventually, the cascade of downstream signaling eventsculminates in the secretion of soluble factors, plateletspreading, and integrin activation. The secreted solubleagonists act on circulating platelets to be recruited andincorporated into a growing thrombus.

B. The Role of Platelets in Venous Thrombosis

VTE comprises deep vein thrombosis and pulmonaryembolism. The underlying pathophysiology of venousthrombosis has predominantly been attributed to Virch-ow’s Triad: hypercoagulability, alterations in blood flow(stasis and turbulence), and endothelial dysfunction(Mackman, 2012). Platelets have traditionally not beenthought to be a major player of venous thrombosis;however, increasing experimental evidence has demon-strated an important platelet contribution to the path-ophysiology of venous thrombosis (Montoro-Garcíaet al., 2016). Platelet membranes are thought to con-tribute approximately 95% of the MPs found in circula-tion. Elevated MPs have been extensively described inpatients with VTE, especially in cancer patients(Chirinos et al., 2005; Ay et al., 2009; Tesselaar et al.,2009; Zwicker et al., 2009; Garcia Rodriguez et al., 2010;Manly et al., 2010). In an ectopic cancer model, micetreated with clopidogrel, a P2Y12 inhibitor, decreasedtumor size and restored hemostasis by preventing theaccumulation of cancer cell–derived MPs at the site ofthrombosis (Mezouar et al., 2015). The in vivo use ofantiplatelet drugs, clopidogrel and acetylsalicylic acid(ASA), in dogs, rat, rabbits, as well as the in vitromodels, by which they reduced venous thrombus for-mation, has supported the role of platelets in thepathophysiology of venous thrombosis (Herbert et al.,1993; Bernat and Herbert, 1994; Savi et al., 1994, 2000;Imbault et al., 1996; Arroyo et al., 2001; Moore andDeschler, 2007; Wang et al., 2007). Deletion of the P2Y1receptor, a purinergic receptor expressed in the platelet,also showed reduced venous thrombosis formation (Birdet al., 2012). Furthermore, the ASPIRE andWARFASAclinical trials have provided strong evidence that ASAafter initial anticoagulation is ceased reduces the rate ofrecurrence of VTE in patients with a prior unprovokedVTE (Becattini et al., 2012; Brighton et al., 2012). In

fact, aspirin was shown to reduce recurrent events bymore than one-third without significantly increasingthe risk of bleeding.

C. The Role of Platelets in Immune Responses

Extending beyond the classic roles of hemostasis andthrombosis, platelets are increasingly being recognized fortheir functions in immune-pathologic disorders in theblood, such as inflammation and non- or proinfectiousimmunologic functions (Sullam et al., 1988; Pampolinaand McNicol, 2005; Fitzgerald et al., 2006b; Arman et al.,2014; Boilard et al., 2014). Although there is developingunderstanding on the mechanisms by which foreignsubstances initiate host immune response via plateletinteractions, the role of platelets in immunity is complex.Depending on the type of foreign agent, platelet activationcan be enhanced, leading to either prothrombotic eventsor dampened platelet response, resulting in bleedingcomplications. In this section, an overview of plateletoverlap with the immune response following bacterial,viral, or drug exposure will be discussed briefly.

Human and murine platelets express all nine TLRs(Koupenova et al., 2018) in addition to Fcg receptors thatare used by the innate and adaptive immune cells(macrophages, neutrophils, and dendritic cells) (Shirakiet al., 2004) to induce platelet response that indirectlybridges communication between platelets and other my-eloid progenitor cells. For instance, Gram-negativebacteria-derived endotoxin, lipolysaccharide (LPS), playsa fundamental role in sepsis through the activation of theTLR4onneutrophils. Similarly, LPS induction ofTLR4onplatelets can lead to platelet–neutrophil aggregate forma-tion and subsequent neutrophil extracellular trap (NET)activation knownasNETosis (Clark et al., 2007). NETosisis a dynamic process that may either have a beneficialeffect for the host in isolating and preventing the spread ofinvading bacteria or detrimental outcome by whichplatelet-induced activation of neutrophils promotes injuryto the host. Platelet TLR4 activation by LPS has also beenshown to induce in vivomicrovascular thrombosis inmice,resulting in thrombocytopenia (Zhang et al., 2009), as wellas enhanced ex vivo platelet secretion and aggregation.

TLR4 also plays a role in hemolytic-uremic syndrome(HUS), characterized by nonimmune microangiopathichemolytic anima, thrombocytopenia, and renal failure(Prohászka, 2008). Human platelets have been demon-strated to bind to the O157:H7 LPS serotype derived fromenterohemorrhagic Escherichia coli (EHEC) throughTLR4 (Ståhl et al., 2006). Such binding of LPS is shownin platelets from childrenwithHUSafter EHEC infection,but not in children who did not develop HUS after EHECinfection, suggesting that platelet–LPS interaction maycontribute to thrombocytopenia during HUS. In addition,platelets from HUS patients show elevated plateletmarkers [P-selectin, platelet factor 4 (PF4)], MPs, andb-thromboglobulin (Appiani et al., 1982; Katayama et al.,1993; Galli et al., 1996).

530 Yeung et al.

Besides TLR4, platelets can interact with variousbacteria (Staphylococci family, Neisseria gonorrheae,Porphyromonas gingivalis, and Helicobacter pylori)(Fitzgerald et al., 2006a; Yeaman, 2010; Hamzeh-Cognasse et al., 2015) by using their FcgRIIa (Fcfragment of IgG receptor IIa), complement receptors,or glycoprotein receptors. Following bacterial interaction,activated platelets enhance P-selectin expression to me-diate its association with the P-selectin receptor onneutrophils as well as secrete the antimicrobial peptide,b-defensin, to induce NET formation (Kraemer et al.,2011). Similar to HIT, a life-threatening disorder that ischaracterized by low platelet count and thromboemboliccomplications, circulating PF4 can also recognize Gram-negative bacteria, leading to the formation of PF4/heparin-like epitopes. Exposed PF4/bacterial epitopesare recognized by autoantibodies, which in turn can bindto the platelet FcgRIIa and induce platelet activation ordestruction by opsonization mediated by neutrophils(Krauel et al., 2011, 2012).Viruses also rely on the same family of receptors as the

bacteria to mediate platelet–virus interaction. Humancytomegalovirus (HCMV) bound to platelet TLR2 resultsin the release of proinflammatory CD40L, interleukin-1b,and vascular endothelial-derived growth factor (VEGF).Although HCMV does not induce platelet adhesion oraggregation, enhanced platelet–neutrophil heterotypicaggregates and neutrophil activation are observed in thepresence of HCMV-treated platelets (Assinger et al.,2014). In addition, antibodies developed against influenzaH1N1 virus have been shown to activate human plateletsindependently through both FcgRIIa signaling andthrombin generation (Boilard et al., 2014). Finally, denguevirus infection is characterized by profound hemorrhagicfever and thrombocytopenia in humans. In a rhesusmacaquemodel infectedwith dengue virus, platelets wereshown to be engulfed by monocytes with observed in-creased permeability of the endothelium (Onlamoon et al.,2010). In general, these studies provide evidence that byrecognizing bacterial, drug, or viral components, plateletscanmediate prothrombotic or proinflammatory pathwaysthrough varying host defense mechanisms (Fig. 2).

D. Platelets in Cancer Metastasis

Although not originally appreciated as being associatedwith platelet function, it is now well recognized that theprogression of tumors in cancer patients is accompanied byincreased risk of thrombotic episodes (Khorana and Fine,2004; Khorana et al., 2008; Khorana and Connolly, 2009;Lyman andKhorana, 2009). Current evidence supports animportant role for platelet regulation of tumor growth andmetastasis with platelet–tumor cross-talk contributing toincreased tumor metastasis, angiogenesis, growth, pro-liferation, and even enhanced platelet activation. Plateletsare now recognized to be a major source of metastatic andproangiogenic or survival growth factors (Assoian et al.,1983), such as transforming growth factor-b1 and VEGF.

Mechanistically, platelet-derived transforming growthfactor-b1 has been observed to act through the p-Smadsignaling pathway to induce phenotypic conversion incancer cells, such as the transition from epithelial tomesenchymal-like cells. Mice exposed to human coloncarcinoma cells (HT29) cocultured or primed with humanplatelets exhibited a higher incidence of lung metastasescompared with untreated HT29 cells. Interestingly, ASAadministration to mice has been shown to prevent theincreased rate of lung metastasis in vivo as well asdownregulating E-cadherin and upregulating transcrip-tion factor Twist1 that is associated with metastaticcancer (Guillem-Llobat et al., 2016). In addition to theprometastatic role of platelets, platelets also promoteangiogenesis, a fundamental process for tumor growthand survival, through the secretion of VEGF released froma-granules, which acts on the VEGF receptor, VEGF-R2,on endothelial cells (Möhle et al., 1997; Italiano et al.,2008). Successful metastasis is now known to rely oncancer cell adhesion to platelets, which can bemediated bysurface proteins, including P-selectin and integrin aIIbb3.

In addition to regulation of tumor function and metas-tasis by platelets, tumor cells can also directly triggerplatelet activation, through either the induction of ago-nists (TxA2 and ADP) or direct physical contact withplatelets (Grignani et al., 1989; Zucchella et al., 1989;Kato et al., 2005; Mitrugno et al., 2014). Podoplanin(PDPN), a transmembrane sialoglycoprotein, is highlyexpressed on metastatic tumor cells and can interactwith C-type lectin-like receptor 2 (CLEC-2), a hemi-immunoreceptor tyrosine-based activation motif (ITAM)on the platelet surface, to induce platelet activation andin turn further promote tumor growth and metastasis.Mice administered with anti-PDPN antibody, MS-1,which prevented PDPN/CLEC-2 interaction, exhibited asignificant reduction in tumor metastasis and growth(Takagi et al., 2013). In summary, the mechanistic andcellular contributions of platelets to tumor survival andmetastasis suggest the validity of targeting platelets incancer as a new avenue for therapy.

IV. Regulators of Platelet Activationand Signaling

A. G Protein–Coupled Receptors

GPCRs are seven-transmembrane receptors with anintracellular C terminus and an extracellular N termi-nus (Dohlman et al., 1987; Kroeze et al., 2003; Woulfe,2005). GPCRs signal through physical interaction withheterotrimeric G proteins (Moers et al., 2003), which arelocated on the surface of the internal membrane (Wonget al., 1990). G proteins associated with GPCRs can becategorized into four families, Gs, Gi, Gq, and G12/13,accordingly to the a subunit identities and functions(Wettschureck et al., 2004). G protein selectivity andresultant function in the platelet are not always obviousbased on receptor activation, because many GPCRs can

Platelet Therapeutics 531

be stimulated by more than one platelet agonist andcoupled to single or multiple G protein families (Woulfe,2005; Li et al., 2010). A number of GPCRs are expressedon the surface of the platelet, and activation of thesereceptors by their respective ligands dictates the extentof activation (Fig. 3) or inhibition of platelets in thevessel.1. Protease-Activated Receptors 1 and 4.

Protease-activated receptors (PARs) are widely expressedon platelets and are primarily activated by the potentserine protease thrombin (Ossovskaya andBunnett, 2004;Arachiche and Nieman, 2017). In addition to thrombin,other serine proteases, such as calpain, granzyme, andfactor Xa, are able to activate PAR1 and PAR4 (Zhaoet al., 2014). The PAR family of GPCRs is comprised offour members: PAR1, PAR2 (not expressed on plate-lets), PAR3 (not expressed on human platelets), andPAR4; and the general mechanism of PAR activation issummarized as follows: 1) the N terminus is proteolyt-ically cleaved to expose the tethered ligand, and 2) thenewly formed N-terminal exodomain bends over andbinds to extracellular loop 2 of the receptor, leading toreceptor activation, G protein activation, and down-stream signal transduction (Seeley et al., 2003). Part ofthe challenge in studying platelet activation is thathuman platelets express both PAR1 and PAR4 on itssurface, whereas most vertebrates below primate expressPAR3 and PAR4 on their platelet surface (Ossovskayaand Bunnett, 2004).

Whereas PAR1 and PAR4 are activated by the sameligand and share similar binding to Gaq and Ga13,they only share 27% amino acid sequence identity,suggesting that PAR1 in the human cannot be directlycompared with PAR4 in either the human or mousemodels of platelet activation and signaling (Xu et al.,1998). Compared with PAR1, PAR4 cleavage by throm-bin is not as efficient if expressed by itself and needs10 times more thrombin to be activated (Jacques andKuliopulos, 2003; Nieman, 2008). However, the inher-ent coexpression of PAR1 and PAR4 on human plateletsis postulated to potentially enhance the PAR4 cleavagerate by approximately 6- to 10-fold, suggesting thatdimerization or oligomerization of the PARs on thesurface of the platelet may play an important role inregulation of platelet activation (Jacques and Kuliopulos,2003; Nieman, 2008; Arachiche et al., 2013).

PAR1 and PAR4 in human platelets signal throughthe a subunit of heterotrimeric Gq andG13 proteins (Gaqand Ga13); however, the coupling of PARs to Gai inplatelets has not been confirmed to date (Holinstatet al., 2006, 2009; Kim et al., 2006; McCoy et al., 2012;Arachiche and Nieman, 2017). Gaq transmits its signalprimarily through the activation of phospholipase C(PLC)b, which subsequently induces second messengersignaling through formation of IP3 and DAG (Hunget al., 1992; Offermanns et al., 1997; Holinstat et al.,2009; Stalker et al., 2012; Edelstein et al., 2014). IP3stimulates the intracellular calcium mobilization, and

Fig. 3. GPCR signaling in platelet function. Platelets express thrombin (PAR1/PAR4) and purinergic (P2Y1 and P2Y12) receptors. Each receptor iscoupled to either Gi, Gq, or G13, which is involved in platelet activation (granule release, integrin activation).

532 Yeung et al.

DAG activates protein kinase C, respectively (Hunget al., 1992; Offermanns et al., 1997). As a result, adownstream integrin activation pathway is generated,including a Ca2+-dependent guanine nucleotide ex-change factor for Rap1, an adaptor Rap1-GTP–interacting adaptor molecule, and the proteins kindlinand talin that directly bind to the cytosolic domain of theintegrin aIIbb3 (Shattil et al., 2010). In addition toactivation of aIIbb3, a number of other platelet re-sponses are regulated by the Gaq-mediated pathway,including granule secretion and platelet aggrega-tion (Ossovskaya and Bunnett, 2004; Arachiche andNieman, 2017). The signal transduced through Ga13activates the Rho guanine nucleotide exchange factorp111RhoGEF, resulting in the activation of RhoA, andits effectors, including Rho-activated kinase and LIM-kinase, to stimulate platelet shape change (Moerset al., 2003; Huang et al., 2007). Although shapechange is thought to be predominantly regulated byGa13, activation of Gaq has also been shown to beinvolved in platelet shape change downstream ofmyosin light chain kinase in a Ca2+-dependent man-ner (Offermanns, 2001).Whereas PAR1 and PAR4 share some overlapping

signaling events, each receptor appears to signal plateletactivity through a unique set of pathways with differingkinetics. The intracellular Ca2+ signaling duration differsdramatically betweenPAR1andPAR4 (Covic et al., 2000).Compared with PAR1, PAR4 has sustained Ca2+ signal-ing, which is beneficial for stable clot formation andfibrinogen fully spreading (Mazharian et al., 2007).Membrane lipid signaling identified on platelets is regu-lated via PAR1 and PAR4 differently. Membrane-boundneutral sphingomyelinase, which regulates platelets byinducing themitogen activated protein kinase pathway, isincreasingly associated with PAR4, but not PAR1, inhuman platelets when responding to thrombin (Chenet al., 2013). Compared with PAR1, stimulation of PAR4induces significantly higher TxA2 formation, which is ametabolite of lipid membrane AA and activates plateletsvia TPa (Holinstat et al., 2011). Integrin aIIbb3 isactivated by PAR1, instead of PAR4, through phosphoi-nositide 3-kinase (PI3K) to induce platelet aggregation(Holinstat et al., 2007; Voss et al., 2007). All of the aboveleads to the rejection of the original hypothesis that PAR4is just a redundant receptor for PAR1, reinforcing theconcept that pharmacological perturbation of each recep-tor is warranted to develop therapeutic target (to bediscussed in section Platelet Pharmacological Targetsand Interventions of this review).2. Purinergic Receptors P2Y1 and P2Y12. The

purinergic receptors can be classified as either aden-osine receptors (P1) or nucleotide receptors (P2).Ligand-gated ion channels (P2X) and GPCRs (P2Y) arethe twomajormembers of the P2 family (Murugappa andKunapuli, 2006). Human platelets express the two P2Yreceptor subtypes, P2Y1 and P2Y12, that are activated by

ADP (Fig. 3) and are prime targets for antiplatelettherapy (Murugappa and Kunapuli, 2006; Yeung andHolinstat, 2012). Optimal platelet activation by ADPrequires both receptors, by which each receptor subtypecontributes uniquely through its associated G proteins.P2Y1 is a GPCR coupled to Gaq that mediates activationof PLCb and subsequent production of IP3 and DAG,causing Ca2+ release and protein kinase C activation(Ayyanathan et al., 1996; Murugappa and Kunapuli,2006). Interestingly, shape change is also observed inP2Y1 activation, suggesting Gaq is also involved in shapechange as observed in Ga13. This is most likely due tocalcium or calmodulin, or Rac-dependent contractilesignaling in the Gaq pathway (Offermanns et al., 1997;Soulet et al., 2005).

In contrast to the wide tissue distribution of P2Y1,being expressed in platelets, central nervous system,and peripheral tissues, P2Y12 is only found in plateletsand brain, which has made it an appealing antithrom-botic drug target (Murugappa and Kunapuli, 2006).P2Y12 is coupled to both Gaq and Gai, which results innot only PLCb activation, but also inhibition of adenylylcyclase (Hollopeter et al., 2001; Kauskot and Hoylaerts,2012). The activation of P2Y12 has been demonstratedto induce Gbg activation that mediates dense granulesecretion through the PI3K/Akt pathway (Kauffensteinet al., 2001; Murugappa and Kunapuli, 2006). Althoughmany GPCR targets exist on the surface of the platelet,at least four Food and Drug Administration (FDA)–approved drugs have been developed to target P2Y12, asa standard-of-care treatment of a number of thromboticconditions, used either as monotherapy or polytherapyin conjunctionwith ASA (referred to as dual antiplatelettherapy).

3. Prostaglandin Receptors TPa, IP, EP, and DP.Prostaglandins (PGs) are lipid-derivedautacoids formedbya 20-carbon unsaturated fatty acid, AA, sequential metab-olism catalyzed by COX, and prostaglandin synthaseenzymes (Hata and Breyer, 2004). TxA2, PGD2, PGE2,and prostacyclin (PGI2) are four predominant bioactiveprostaglandins formed in vivo that are known to regulateplatelet function through their respective receptors, TPa,prostaglandin D2 receptor 1 (DP1), prostaglandin E2receptors (EP1–4), and prostacyclin receptor (IP) (Fig. 4).TxA2, derived from COX-1 oxidation of AA followingphospholipase A2 activation, is a potent agonist for plateletshape change and aggregation (Woulfe, 2005). Althoughthere are two variants of the TPs, TPa and TPb (Habibet al., 1999), TxA2 interacts with the TPa receptor coupledtoGaq andGa13 onplatelets (Raychowdhury et al., 1994) tomediate the PLCb and RhoA pathways (Woulfe, 2005;Gong et al., 2010). In contrast, PGD2 activatesDP1 coupledto Gas that stimulates adenylyl cyclase to enhance cAMPproduction to inhibit platelet activation.

Interestingly, PGE2 exhibits a biphasic, dose-dependenteffect on platelet function through its receptors, EP1, EP2,EP3, andEP4. Each of the EP receptors is unique such that

Platelet Therapeutics 533

it has different binding affinity for PGE2 as well as coupledto different G proteins, allowing PGE2 to either activate orinhibit platelet function at low or high concentrations,respectively (Kauskot and Hoylaerts, 2012). Whereas EP1and EP2 are low-affinity receptors (Kd . 10 nM), EP3 andEP4 bind to PGE2 with higher affinity (Kd , 10 nM)(Abramovitz et al., 2000). Furthermore, EP2 and EP4 arecoupled to Gas, which activates adenylyl cyclase to evokecAMP generation, resulting in inhibition of platelet func-tion. EP1 is reported to couple to Gaq based on theobservation of increased calcium mobilization (Sugimotoand Narumiya, 2007). Although EP3 has been predomi-nantly demonstrated to associate with either Gai or Gaq,there are reports ofGas orGa13 coupling. Thus, the exactGproteins involved in EP3 signaling are still unclear. Over-all, the dual effect regulated by at least four GPCRs givesthe platelet a significant level of control over the overallPGE2 effect on platelet function and activation.PGI2 or prostacyclin is a known inhibitor of platelet

activation derived from AA (Kauskot and Hoylaerts,2012). PGI2 maintains platelets in a quiescent state inthe absence of vascular injury or agonist activation(Cheng et al., 2002; Woulfe, 2005; Tourdot et al., 2017)by activating its receptor, IP, coupled to Gas tostimulate adenylyl cyclase and subsequent cAMPformation and protein kinase A activation (Kauskotand Hoylaerts, 2012).

B. Integrin Receptors aIIbb3, a2b1In addition to GPCRs, platelets express multiple non-

GPCRs, including several integrin receptors, such asfibrinogen receptor, aIIbb3; collagen receptor, a2b1; andlaminin receptor, a6b1, which share similar signalingmechanisms (Li et al., 2010). The fibrinogen receptor

aIIbb3 is the most abundant surface integrin (40,000–80,000 copies per cell) with a 148-kDa aIIb and a 95-kDab3 subunit (Kauskot and Hoylaerts, 2012). Normally,aIIbb3 is at low affinity or resting state; however, uponplatelet response to activating agonists, aIIbb3 un-dergoes an inside-out process by which it switches toan activated state with high affinity (Li et al., 2010) forfibrinogen. The shift in integrin affinity or conforma-tional change is facilitated by the binding of talin andkindlin, to the intracellular b3 domain (Shattil et al.,2010). The underlying mechanism accommodating thetalin and kindlin binding is attributed to Rap1 activa-tion and Rap1-GTP–interacting adaptor molecule me-diated by Ca2+-dependent guanine nucleotide exchangefactor for Rap1 (Lafuente et al., 2004). In addition, theplatelet can be secondarily activated by direct aIIbb3interaction with fibrinogen, vWF, and other matrixproteins, to transmit signals to the cytoplasmic andcytoskeletal domains. This process is termed outside-insignaling, by which matrix proteins directly activateaIIbb3.

Furthermore, integrin a2b1 is one of two receptorsexpressed on the platelet that binds with high affinity tocollagen (Cosemans et al., 2008). a2b1 has primarilybeen recognized in providing platelets firm adhesion tothe subendothelial wall following platelet translocationalong the vessel wall (see section General Function ofPlatelets). The role played by a2b1 in the collagen-induced platelet activation still remains unclear. Manyof the contradicting results on whether collagen bindingto a2b1 is critical for platelet activation were due todiffering experimental conditions. In contrast, onegeneral consensus was that genetic ablation or phar-macological inhibition of a2b1 integrin in platelets

Fig. 4. Prostanoid receptors on human platelets. Each prostanoid receptor, IP, DP, EP1–4, and TPa, is uniquely defined by its lipid ligands as well asits associated G proteins. Depending on the type of oxylipins or doses, platelet function can either be inhibited or activated. This is dictated by theircognate receptor activation, which is coupled to either inhibitory Gas or activating Ga13, Gaq, and Gai.

534 Yeung et al.

delayed platelet response to collagen without affectingthe final extent of activation compared with controls. Ithas been demonstrated that, analogous to aIIbb3 bind-ing to fibrinogen, a2b1 integrin undergoes a shift fromlow- to high-affinity state for collagen following stimu-lation (Jung and Moroi, 1998), supporting an inside-outsignaling mechanism. In contrast, an outside-in modelhas also been demonstrated by which platelets spreadon GFOGER motifs. Platelet spreading on a2b1 recog-nition motif, GFOGER, induced activation of the Srckinases and subsequent Syk recruitment and PLCg2activation (Inoue et al., 2003). This observation was alsosupported by PLCg2-deficient mouse platelets, whichexhibited limited spreading on GFOGER-coated sur-face. Interestingly, this study also provided insights tothe underlying intracellular signaling events involvedin a2b1-mediated platelet adhesion and spreading.a2b1-mediated spreading was demonstrated to sharesimilar signaling pathways as the outside-in models ofaIIbb3 and FcgRIIa.

C. Immunoreceptor Tyrosine-Based ActivationMotif Receptors

Platelets express three (hemi)ITAM receptors, GPVI,FcgRIIa, and CLEC-2 (Fig. 5). Although both GPVI andFcgRIIa belong to the Ig family of receptors and signalthrough the dual ITAM consensus sequence (L/I/V/S)-X-Y-X-X-(L/V) (Kauskot and Hoylaerts, 2012) in platelets,they are structurally distinct in their relation to ITAMlocalization and activation. GPVI noncovalently asso-ciates with the Fc receptor g-chain (FcgR) containingthe ITAM, whereas FcgRIIa possesses the ITAM con-sensus sequence in its cytoplasmic tail (Clemetson andClemetson, 2001; Bergmeier and Stefanini, 2013). UponGPVI and FcgRIIa cross-linking or clustering, theITAM is phosphorylated by Src family kinases, Lynand Fyn, followed by recruitment and activation ofproximal effector tyrosine kinase Syk. Syk initiates adownstream signaling cascade that includes the phos-phorylation of transmembrane adapter linker for acti-vated T cells (LAT) and assembly of a signalosome. Thecore of the signalosome consists of phosphorylatedtransmembrane adaptor LAT (Pasquet et al., 1999)and cytosolic adaptors, Src homology 2 domain-containing leukocyte phosphoprotein of 76 kDa boundto Gads or Grb2 (Judd et al., 2002; Hughes et al., 2008).These three proteins associate with a number ofsignaling molecules, including Bruton tyrosine kinase(Quek et al., 1998), GTP exchange factors (Vav1 andVav3) (Pearce et al., 2002, 2004), small GTPase Rac, andthe a and b isoforms of PI3K, which are critical for therecruitment and activation of PLCg2. For instance, thepleckstrin homology domain of PLCg2 facilitates itsrecruitment to the plasma membrane through itsbinding of the PI3K product, phosphatidylinositol (3,4,5)-triphosphate. This binding plays an important role in themaximal activation of PLCg2, which liberates second

messengers DAG and IP3 from phosphatidylinositol-4,5-bisphosphate.

CLEC-2 is a novel member of the ITAM plateletreceptor family that has a carbohydrate-like extracel-lular domain. Currently, CLEC-2 is found to be impor-tant in lymphatic development and thrombosis. Micelacking CLEC-2 in platelets exhibit impaired plateletaggregate formation and lower susceptibility to arterialthrombosis (May et al., 2009; Suzuki-Inoue et al., 2010;Herzog et al., 2013). Unlike the ITAM receptors, CLEC-2has a single YxxL motif in its cytoplasmic tail that alsoemploys the same signaling effectors as the ITAMs;however, differences of ITAM and hemi(ITAM) signalinghave been reported. Syk has been shown to phosphorylatethe hemi(ITAM), followed by Src family kinase recruit-ment and activation of LAT (Stegner et al., 2014). TxA2has been shown to play an important role in CLEC-2–induced Syk and PLCg2 phosphorylation. In addition,activation of ADP and PAR receptors potentiate CLEC-2signaling (Badolia et al., 2017).

D. Enzymes Targeted for Regulation ofPlatelet Function

Polyunsaturated fatty acids (PUFAs), including AA,docosahexaenoic acid, docosapentaenoic acid, dihomo-g-linolenic acid (DGLA), eicospentaenoic acid (EPA), andlinoleic acid, are hydrolyzed by cytocolic phosholipase A2(cPLA2) inplatelets, and the subsequent free fatty acids aremetabolized by COXs, LOXs, and cytochrome P450(CYP450) to generate structurally distinct oxylipins thatexhibit either pro- or antiplatelet functions (Yeung et al.,2017).

1. Cyclooxygenase-1. COX-1 is highly expressed inplatelets and generates a number of PGs, includingTxA2, through PUFAmetabolism (Rouzer andMarnett,2009). Through COX-1, AA is transformed to series2 PGs (including TxA2, PGD2, PGE2, and PGI2), whichare either prothrombotic or antithrombotic (Ricciottiand FitzGerald, 2011; Chandrasekharan et al., 2016;Yeung et al., 2017). In addition, COX-1 converts the v-6PUFA, DGLA, to series 1 PGs (including TxA1, PGD1,and PGE1) that are known to inhibit platelet activityboth in vivo and in vitro (Lagarde et al., 2013; Sergeantet al., 2016). Similarly, the v-3 PUFA, EPA, reacts withCOX-1 to form series 3 PGs (including TxA3, PGD3,PGE3, and PGI3) to inhibit platelet functions (Fischerand Weber, 1985; Krämer et al., 1996).

2. 12-Lipoxygenases. 12-LOX can be categorized intoone of three forms (platelet, leukocyte, and epithelial)based on its cellular expression, and, among which, theplatelet-type 12-LOX is found in all mammalian species(Funk et al., 1990). 12(S)-LOX, rather than 12(R)-LOX, isthe predominant 12-LOX protein reported to regulateplatelet function (Yeung et al., 2017). 12-LOX converts AAto 12(S)-hydroperoxyeicosatetraenoicacid, which is thenquickly reduced to 12(S)-HETE (Yeung et al., 2017). 12(S)-HETE has been shown to have contradicting roles in

Platelet Therapeutics 535

platelet function. 12(S)-HETE has been reported to exertantiplatelet effects by inhibiting cPLA2 (Chang et al.,1985; Yamamoto, 1992) and preventing TxA2 binding toits TPa receptor (Fonlupt et al., 1991). Conversely, 12(S)-HETE has been reported to be proaggregatory or pro-thrombotic by inducingTxA2 and dense granule secretion,as well as inhibiting PGE1-induced cAMP formation(Calzada et al., 1997; Yeung et al., 2012, 2016; Adiliet al., 2017). Moreover, other 12-LOX–derived metabo-lites, including 12(S)-hydroxyeicosapentanoic acid fromEPA, 11/14-hydroxydocosahexaenoic acid from docosa-hexaenoic acid, 11/14-hydroxydocosapentaenoic acid fromdocosapentaenoic acid, and 12-hydroxyeicosatrienoic acidfrom DGLA, are shown to have antithrombotic effectin vivo and in vitro (Takenaga et al., 1986; Sun et al., 2015;Yeung et al., 2016).3. Cyclic Nucleotide Phosphodiesterases. There are

at least seven isoenzymes of phosphodiesterases (PDEs)distinguished by their structural and enzymatic prop-erties; however, platelets express PDE2, PDE3, andPDE5. PDEs regulate intracellular levels of cyclicnucleotides, such as cGMP and cAMP, in the plateletsthrough hydrolysis (Cheung et al., 1996; Degermanet al., 1997). As a result of decreased cGMP or cAMP,platelets are prevented from being inhibited. Thus,PDEs are major targets of antiplatelet therapy toregulate platelet function throughmodulation of endog-enous cAMP or cGMP (Conti et al., 1995; Manganielloet al., 1995).

V. Platelet Pharmacological Targetsand Interventions

Antiplatelet agents, whether administered as amono- or polytherapy, are the cornerstone of clinicaltreatment of arterial thrombotic events and ischemic

stroke. Although antiplatelet targets have been limitedto primarily COX-1, P2Y12 receptor, and integrin aIIbb3,recent advances have revealed a number of newertargets that have led to novel antiplatelet drugs, eithercurrently in use or in preclinical or early clinical stagesof development. These agents target surface receptors(glycoproteins and GPCRs), oxygenases, and PDEs (Fig.6; Table 1). With the availability of several antiplateletagents with different safety and efficacy profiles, find-ing the best antithrombotic drug to rapidly and potentlycurtail thrombotic-associated events without increasingserious bleeding is becoming an attainable goal.

A. Oxygenase Inhibitors

Platelets express two classes of oxygenases, COX-1and 12-LOX, which generate an assortment of uniquelipid mediators (oxylipins) from PUFAs that exhibitpro- or antithrombotic activity. The most abundantPUFA in the lipid bilayer is AA, which is acted uponby both oxygenases to form the prothrombotic oxylipins,TxA2 and 12(S)-HETE. Significant efforts have resultedin a number of therapeutic drugs being developed totarget the formation of these prothrombotic oxylipins toprevent their potentiation of platelet activation.

Aspirin or ASA, one of the first drugs developed forprevention of thrombosis, irreversibly targets COX-1 inthe platelets to block the prostanoid production fromAA(Capone et al., 2010). ASA is rapidly absorbed in theupper gastrointestinal tract following oral administra-tion, leading to measurable platelet inhibition within60 minutes. The plasma half-life of ASA is approxi-mately 15minutes, and peak plasma levels are achieved30–40 minutes after ingestion (Patrono et al., 2005;Capodanno and Angiolillo, 2016). The ISIS-2 clinicaltrial showed ASA therapy was associated with asignificant reduction in vascular mortality in patients

Fig. 5. ITAM and integrin receptors on platelets. GPVI, FcgIIa, and CLEC-2 belong to a class of hemi(ITAM) receptors that are involved in plateletactivation. Whereas GPVI and a2b1 are activated by collagen, FcgIIa recognizes IgG immune complexes to induce integrin aIIbb3 activation. CLEC-2contains a single YxxL motif that is activated by podoplanin. Ligand binding to GPVI or FcgIIa results in Syk and subsequent PLCg2 activation,leading to platelet aggregation, mediated by the active conformation of integrin aIIbb3.

536 Yeung et al.

with suspected acute MI, either as a stand-alonetherapy or in a combination with streptokinase. Thisestablished ASA as the first-line therapy in patientswith ST-elevation myocardial infarction, and recom-mended to be takenwhen presentedwith symptoms andindefinitely irrespective of treatment strategy (Levineet al., 2011; O’Gara et al., 2013). More recently, ASA hasbeen used as a dual antiplatelet therapy with P2Y12receptor inhibitors (clopidogrel, prasugrel, or ticagrelor)to prevent blood clots after percutaneous coronaryinterventions (PCIs). Although ASA has effectivelyreduced morbidity and mortality, an increase in bleed-ing has also been associated with its use, especiallygastrointestinal hemorrhage (Huang et al., 2011).Another approach to antiplatelet therapy is targeting

12-LOX from forming 12(S)-HETE. Initial 12-LOX in-hibitors (baicaelin, nordihydroguaiaretic acid, 5,8,11,14-eicosatetraynoic acid, OPC-29030, L-655,238, andBW755C) were found to be nonselective, as they werefound to target cPLA2, COX-1, COX-2, and other LOXs(15-LOX-1, 15-LOX-2, 5-LOX) in addition to 12-LOX.Recently, a more selective 12-LOX inhibitor, ML355, hasbeen developed with no inhibitory activity identified for15-LOX, 5-LOX, or COX, which exhibits potent inhibitionof platelet activation in vivo (Luci et al., 2010; Adili et al.,2017).Moreover,ML355, whichwas given orally twice perday for 2 days to wild-type mice dose dependently (1.88,3.75, 7.5, 15, and 30 mg/kg), inhibited thrombus growthand vessel growth following FeCl3-induced mesentericand laser-induced cremaster arteriole injury. ML355 hasa favorable pharmacokinetic profile in which themaximalconcentration of ML355 in plasma was approximately57 mmol/l within 30 minutes, and approximately 5 mmol/lwas detected in plasma at 12 hours following oraladministration of 30 mg/kg for 2 days. ML355 was alsoable to inhibit human platelet adhesion and thrombus

growth over collagen-coated ex vivo flow chambers, con-firming the in vivo observations (Adili et al., 2017).Importantly, bleeding diathesis based on in vivo hemo-static plug formation and tail bleeding in mice orallygavaged with 30 mg/kg ML355 was not altered. Addition-ally, the utility of ML355 for inhibition of a variety ofthrombotic diseases was established through inhibitionof immune-mediated platelet activation and prevention ofHIT in an ex vivomousemodel (Yeung et al., 2014). Takentogether, these preclinical studies demonstrated thepotential therapeutic benefits of targeting 12-LOXwith ML355 for effectively preventing thrombotic eventswith minimal impact on bleeding.

B. ADP Receptor Inhibitors

P2Y12 receptor inhibitors are divided into prodrugand active antagonists. The thienopyridines (ticlopi-dine, clopidogrel, and prasugrel) are FDA-approvedprodrugs, whereby the active metabolites of the thieno-pyridine prodrugs covalently and irreversibly bind tothe receptor during the entire life span of the platelets(7–10 days). Although the thienopyridines requireCYP450 isoenzyme metabolism for the generation ofactive metabolites, the pathways leading to their activemetabolites differ between the prodrugs. In contrast,the nucleoside analogs (ticagrelor and cangrelor) areactive P2Y12 receptor antagonists that do not necessar-ily require metabolic conversion. The nucleoside ana-logs are reversible inhibitors due to their distinctbinding site from ADP binding domain.

One of the first thienopyridines, ticlopidine (tradename Ticlid), which has an onset of action of 1–2 hoursafter a single oral dose (250 mg), was initially shown tobe useful for preventing coronary stent occlusions andstrokes. Due to the serious side effects [bone marrowsuppression, thrombotic thrombocytopenia purpura

Fig. 6. The major antiplatelet targets. Drugs inhibit surface receptors, GPCRs (PAR1/4, P2Y12) and glycoproteins (integrin aIIbb3, GPVI, and GIb-IX-V), oxygenases (COX-1, 12-LOX), and phosphodiesterases to regulate platelet function. Agents coded in black are currently FDA approved, whereasdrugs labeled as blue are under investigation in preclinical or clinical phases.

Platelet Therapeutics 537

TABLE

1FDA

appr

oved

andinve

stigationa

lan

tiplatelet

drug

s

Targe

t(s):

Oxy

gena

seIn

tegrin

aIIbb3

Collage

nreceptor

Thrombinreceptors

Dru

gAsp

irin

ML35

5Abcixim

abTirofiban

Eptifab

tide

GPVI-Fc

Atopa

xar

Vorap

axar

BMS-986

120

Trade

nam

eBay

erAsp

rin

(Inv

estiga

tion

al)

ReoPro

Agg

rastat

Integrilin

Rev

acep

t(Inv

estiga

tion

al)

Zontivity

(Inv

estiga

tion

al)

Molecule

type

Acetylsalicylic

acid

Smallmolecule

Human

ized

mou

semon

oclona

lan

tibo

dy

Non

peptide

RGD

mim

etic

KGD-con

taining

heptap

eptide

Solub

ledimeric

glycop

rotein

VI-Fc

fusion

protein

Bicyclicam

idine

Tricyclic

3-ph

enylpy

ridine

Smallmolecule

Bindingtype

Irreve

rsible

N.D

.Irreve

rsible

Rev

ersible

Rev

ersible

N.D

.Rev

ersible

Rev

ersible

Rev

ersible

Formation

Oral

Oral

IVIV

IVIV

Oral

Oral

Oral

Half-life

15–20

min

(Altman

etal.,20

04)

2.5h(m

urine)

(Adiliet

al.,

2017

)

10–30

min

(De

Luc

a,20

12)

1.5–

2h(D

eLuca,

2012

)2–

2.5h(D

eLuc

a,20

12)

67–13

7hde

pendingon

thedo

ses(U

ngerer

etal.,20

11)

22–26

h(Tello-

Mon

toliuet

al.,20

11)

7–13

days

(Tello-

Mon

toliuet

al.,

2011

)

4h(W

ilson

etal.,20

17)

Prodr

ug

No

N.D

.No

No

No

No

No

No

No

Onsetof

action

1h(E

ikelbo

omet

al.,20

12)

N.D

.Within

20min

(Kinget

al.,

2016

)

Within

20min

(Kinget

al.,

2016

)

Within

20min

(King

etal.,20

16)

2h(U

ngerer

etal.,20

11)

3.5h(Tello-

Mon

toliuet

al.,20

11)

1–2h(Tello-

Mon

toliuet

al.,

2011

)

2h(W

ilson

etal.,20

17)

Offsetof

action

7–10

days

N.D

..4h(D

eLuca,

2012

).4h(D

eLuca,

2012

)4–

8h(Top

ol,19

99;

Vorch

heim

eret

al.,19

99)

7da

ys(U

ngerer

etal.,

2013

)3–

5da

ysde

pendingon

thedo

ses

4 –8wee

ks(K

osog

louet

al.,20

12)

24h(W

ilsonet

al.,20

17)

Indication

(s)

Non

fatal

thrombo

tic

even

ts;

patien

tswithSTEMI

(Franchi

etal.,

2017)

Preclinical

data

show

redu

ction

inthrombo

tic

card

iova

scular

even

tsan

dpr

even

tion

ofIT

T(A

diliet

al.,

2017

;Yeu

nget

al.,20

14)

Preve

ntionof

isch

emic

complications

dueto

PCI

and

patien

tswith

UA/N

STEMI

pre-PCI

(Amsterda

met

al.,20

14a;

Jneidet

al.,

2012

)

Red

uction

ofcard

iova

scular

even

tsin

ACS

patien

tswith

UA/N

STEMI

(Amsterda

met

al.,20

14a;

Jneidet

al.,

2012

)

Red

uction

ofacute

card

iacisch

emic

even

ts(dea

than

d/or

MI)

forPCI,

corona

rysten

ts,pa

tien

tswithUA/STEMI

(Amsterda

met

al.,

2014

a;Jn

eidet

al.,

2012

)

Und

erinve

stigation

fortrea

tmen

tof

patien

tswithACS

andstroke

Unde

rinve

stigation

fortrea

tmen

tof

patien

tswithACS

andstroke

Red

uction

ofthrombo

tic

card

iova

scular

even

tsin

patien

tswith

prev

ious

MIor

periph

eral

arterial

diseas

e

Unde

rinve

stigation

forredu

ction

ofthrombo

tic

even

tsin

high

risk

patien

ts

Targe

t(s):

PDE

ADP

Recep

tors

GPIb-IX-V

Dru

gDipyridam

ole

Cilostazol

Ticlopidine

Clopido

grel

Prasu

grel

Ticag

relor

Can

grelor

Anfibatide

Trade

nam

ePersa

ntine

Pletal

Ticlid

Plavix

Effient

Brilinta,

Briliqu

e,Possia

Ken

grea

lN.D

.

Molecule

type

Pyrim

ido-py

rimidine

deriva

tive

Quinolinede

riva

tive

Thieno

pyridine

Thien

opyridine

Thien

opyridine

CPTP

ATPan

alog

C-typ

electin-

like

protein

derive

dfrom

veno

mof Agk

istrod

onac

utus

Bindingtype

Rev

ersible

Rev

ersible

Irreve

rsible

Irreve

rsible

Irreve

rsible

Rev

ersible

Rev

ersible

Rev

ersible

Formation

Oral

Oral

Oral

Oral

Oral

Oral

IVIP

(mur

ine);IV

(human

s)(Lei

etal.,20

14;Li

etal.,

2015

;Zhen

get

al.,

2016

) (con

tinued

)

538 Yeung et al.

TABLE

1—Con

tinued

Targe

t(s):

PDE

ADP

Recep

tors

GPIb-IX-V

Dru

gDipyridam

ole

Cilostazol

Ticlopidine

Clopido

grel

Prasu

grel

Ticag

relor

Can

grelor

Anfibatide

Half-life

10h(G

rego

vet

al.,

1987

)10

–11

h(E

ikelbo

omet

al.,20

12)

24–36

h(E

ikelbo

omet

al.,20

12)

20min

(Umem

ura

andIw

aki,20

16)

30min

to7.5h

(Umem

uraet

al.,

2016)

7–9h(Ten

gan

dButler,

2013

)3–

5min

(Sible

and

Naw

arsk

as,20

17)

5–7h(m

urine)

(Zhe

nget

al.,

2016

)Prodr

ug

No

Yes

(CYP3A

5an

dCYP2C

19)(Y

ooet

al.,20

09,20

10)

Yes

Yes

Yes

No

No

No

Onsetof

action

1–2h(G

regovet

al.,

1987

)N.D

.1–

3h(E

ikelbo

omet

al.,20

12)

2–8h(C

apod

annoet

al.,

2013

)30

min

to4h

(Cap

odan

noet

al.,

2013

)

30min

to4h

(Cap

odan

noet

al.,20

13)

Secon

ds(C

apod

anno

etal.,20

13)

N.D

.

Offsetof

action

N.D

.12

–16

h(Y

amam

oto

etal.,20

08)

7–10

days

7–10

days

(Cap

odan

noet

al.,20

13)

7–10

days

(Cap

odan

noet

al.,

2013

)

3–5da

ys(C

apod

anno

etal.,

2013

)

1h(C

apod

annoet

al.,20

13)

N.D

.

Indicatio

n(s)

Adjunct

therap

ywithoral

anticoag

ulantin

thepr

even

tion

ofpo

stop

erative

thrombo

embo

lic

complications

Claudication

inpa

tien

tswith

periph

eral

arterial

diseas

e(Fax

onet

al.,20

04)

Red

ucerisk

ofthrombo

tic

stroke

andpr

even

tion

ofcorona

ryartery

sten

tthrombo

sis

inpa

tien

tsintolerant

toas

pirin

Com

mon

lyused

withas

pirinfor

dual

antiplatelet

therap

yto

redu

ceMIan

dstroke

inpa

tien

tswith

NSTEMIor

acute

STEMI(Franch

iet

al.,20

17)

Red

uce

rate

ofthrombo

tic

card

iova

scular

even

ts,

including

sten

tthrombo

sis,

inACS

patien

tswhoare

unde

rgoing

PCI,

withUA/

NSTEMI,

STEMI

(Franchi

etal.,

2017

)

Red

ucerate

ofrate

ofsten

tthrombo

sis,

card

iova

scular

death,M

I,an

dstroke

inpa

tien

tswithACSor

ahistory

ofMI

(Franchi

etal.,

2017

)

Usedas

anad

junc

tto

PCIfor

redu

cing

therisk

ofpe

riproced

ural

MI,

repe

atcorona

ryreva

scularization,

andsten

tthrombo

sisin

patien

tswitho

utP2Y

12inhibitor

(Sible

and

Naw

arskas,2

017)

Preclinical

data

supp

ort

trea

tmen

tof

isch

emic

even

tsan

dTTP

(Zhen

get

al.,20

16);

stillunde

rinvestigation

inpa

tien

tswith

NST

EMI

andST

EMI

CPTP,c

yclope

ntyl-triazolo-pyrim

idine;

NSTEMI,

non

-STelev

ationmyo

cardialinfarction

;STEMI,

STelev

ationmyo

cardialinfarction

.

Platelet Therapeutics 539

(TTP), neutropenia], use of ticlopidine was limited topatients for whom ASA was not tolerated. Because ofticlopidine’s limited tolerability and significant sideeffects, an alternative thienopyridine analog was devel-oped [clopidogrel (Plavix)]. Clopidogrel is currently themost widely used oral antithrombotic agent. Its irre-versible effect on the P2Y12 receptor is due to itscovalent binding to the cysteine sulphydryl residueswithin the receptor. Although it is well established thatclopidogrel is effective in providing significant protec-tion against thrombotic episodes when administered incombination with ASA (Fox et al., 2004), clopidogreldoes have some drawbacks. Its onset of action is rela-tively slow following initiation of standard dosing, andinhibitory effects on platelet function may be highlyvariable. A major contributing factor to clopidogrel’svariable response is the polymorphism of the requisiteCYP450 isoenzyme, CYP2C19, that is required for itsconversion from prodrug to active drug. Furthermore,gender, ethnicity, body mass index, polymorphismin paraoxnase-1, and comorbidities, such as liverdisease and insulin resistance, can significantly affectclopidogrel efficacy.High on-clopidogrel platelet reactivity or poor re-

sponder to clopidogrel, especially due to the describedfactors above, has been shown to be an importantpredictor of adverse thrombotic outcomes. Thus, prasu-grel (trade name Effient), a newer irreversible agentand third generation thienopyridine, has been devel-oped to accommodate patients with high on-clopidogrelplatelet reactivity. Prasugrel has a higher bioavailabil-ity profile compared with clopidogrel, resulting in fasteronset of action, enhanced platelet inhibition, and lowerinterindividual variability of platelet response. Basedon the TRITON-TIMI 38 clinical trial, prasugrel wasshown to significantly reduce the primary efficacy endpoint (composite cardiovascular deaths, nonfatal myo-cardial infarction (MI), or nonfatal stroke) in patientswith moderate- to high-risk acute coronary syndrome(ACS) undergoing PCI by 19% compared with clopidog-rel over a medium follow-up of 14.5 months (Wiviottet al., 2007). Overall, due to significant bleeding risk inprasugrel compared with clopidogrel, its use is re-stricted to patients who are less than 75 years old andexceed a body weight of 60 kg.Ticagrelor (trade names Brilinta, Brilique, Possia),

which belongs to the cyclopentyl-triazolo-pyrimidineclass, was the first oral allosteric and reversible inhib-itor of the P2Y12 receptor and was approved for use toprevent thrombotic events in patientswithACSorMIwithST elevation. Unlike the thienopyridines, ticagrelor doesnot require metabolic conversion; however, it can be me-tabolized toAR-C124910XXbyCYP3A4. Its rapid on-rate ofplatelet inhibition is due to its quick absorption followingoral administration. Ticagrelor shows a higher consistentand more predictable level of platelet inhibition com-pared with clopidogrel. Although ticagrelor is reversible,

its off-rate is quite slow, bywhich 20% of platelet inhibitionstill remains 3 days after administration. Ticagralor’s slowoff-rate is thought to be due to its binding to circulatingplasma proteins, which lowers its rate of clearance.According to thePLATOclinical trial, ticagrelorwas shownto be more efficacious than clopidogrel and reducedprimary endpoint (composite of cardiovascular death,nonfatal MI, and stroke) in ACS patients (James et al.,2009; Wallentin et al., 2009).

Whereas ticagrelor is deemed to bemore preferable thanclopidogrel, absorption of oral P2Y12 inhibitors prior to PCIis still an issue. Thus, an i.v. and reversible formulation,cangrelor (trade name Kengreal in the United States andKangrexal in Europe) was developed to expedite plateletinhibition. Cangrelor is a reversible P2Y12 platelet inhib-itor that blocks ADP-mediated platelet activation and iscurrently FDA approved as an adjunct to PCI for reducingthe risk of periproceduralMI, stent thrombosis, and repeatcoronary revascularization (Franchi et al., 2017) in pa-tients who have not been treated with other P2Y12inhibitors or glycosylphosphatidylinositols (GPIs). Liketicagrelor, it has a chemical structure that resemblesATP and does not require metabolic conversion byCYP450 to become active. The half-life of cangrelor is veryshort (3–5 minutes), and its inhibitory effects on plateletfunction disappear within minutes following infusion(Storey et al., 2001; Sible and Nawarskas, 2017).

C. Integrin aIIbb3 Receptor Inhibitors

Receptor aIIbb3 is targeted by a class of inhibi-tors categorized into small (tirofiban, eptifibatide) andnonsmall (abciximab) molecules. These molecules exerttheir antithrombotic effects by preventing the binding offibrinogen to the most abundant glycoprotein on theplatelet surface (60,000–80,000 copies per platelet)(Wagner et al., 1996), aIIbb3, thereby blocking plateletaggregation. According to the guidelines of AmericanCollege of Cardiology/American Heart Association,aIIbb3 inhibitors are recommended for patients withunstable angina/non-ST elevation myocardial infarc-tion (UA/NSTEMI) who cannot tolerate clopidogrel andare undergoing PCI (Jneid et al., 2012; Amsterdamet al., 2014b). The most serious complications of theGPIs are bleeding and thrombocytopenia. Thrombocy-topenia typically occurs within 24 hours of i.v. admin-istration; thus, platelet counts are frequentlymonitoredfollowing infusion.

Abciximab (trade name ReoPro), a recombinantmonoclonal antibody recognizing the epitope of the b3motif, was the first aIIbb3 inhibitor to be approved by theFDA for use in the clinic for PCI. Clinical trials (EPIC,EPILOG, EPISTENT) have shown that abciximabtreatment reduced all-cause mortality by approxi-mately 20% during long-term follow-up after PCI. TheFab fragment of abciximab was also found to cross-reactwith aVb3 and aMb2 (Mac-1, CD11b/CD18), which areexpressed by endothelial, smooth muscle cells, and

540 Yeung et al.

leukocytes, respectively. Due to its cross-reactivity,which may be responsible for its unique anti-inflammatoryproperties, abciximab may provide additional benefitscompared with the other GPIs. Abciximab has a slow off-rate because it remains bound to the platelets for anextended time (12–24 hours) and may dissociate andreassociatewith new platelets for as long as 21 days aftercessation of the agent (Kereiakes et al., 2000).Unlike abciximab, eptifibatide (trade name Integrilin)

and tirofiban (trade name Aggrostat) are smaller mole-cules with a lesser degree of affinity for the receptor andshorter duration of action. Tirofiban is a synthetic non-peptide tyrosine derivative that acts as a RGD (arginine-glycine-aspartic acid) mimetic that has a plasma half-lifeof 1.5–2 hours and a short biologic half-life, which reflectsits reversible and relatively low-binding affinity. Aftercessation of tirofiban, partial platelet function is recoveredwithin 4 hours. Comparative studies have shown thatabciximab and eptifibatide were more effective thantirofiban in inhibiting platelet aggregation (Topol et al.,2001). Interestingly, whereas the TARGET clinical trialalso showed abciximab was more effective than thestandard dose of tirofiban in preventing ischemic events(Herrmann et al., 2002), a subsequent clinical trial(ADVANCE) suggested that tirofiban was not inferior toabciximab when given at a higher dose (Schneider et al.,2003; Valgimigli et al., 2008).Eptifibatide, a cyclic heptapeptide, has been shown to

rapidly inhibit platelet aggregation, attaining maximaleffect within 15 minutes after its bolus injection andmaintaining its potent inhibition during infusion.Whereas epitfibatide has a long plasma half-life of;2.5 hours, the platelet-bound half-life is short (sec-onds). Eptifibatide is a reversible inhibitor, as observedby platelet function returning to normal function within4–8 hours following cessation of drug. Both PURSUITand ESPRIT clinical trials have shown the efficacy ofeptifibatide at higher doses in reducing mortality by25%–35% with a modest increase in bleeding comparedwith the placebo-treated group (Harrington, 1997;ESPRIT Investigators: Enhanced Suppression of thePlatelet IIb/IIIa Receptor with Integrilin Therapy,2000; Lincoff et al., 2000; O’Shea et al., 2001; O’Sheaand Tcheng, 2001). Importantly, eptifibatide is cur-rently the most widely used of the aIIbb3 inhibitors forACS patients in PCI, coronary stents, and UA/NSTEMIpre-PCI, and for medical stabilization of UA/NSTEMI.

D. Collagen Receptor Inhibitors

GPVI-Fc (tradename Revacept) is a recombinant GPVImimetic bywhich a fusion proteinwas formed between theextracellular collagen-binding domain of GPVI and theC-terminal of human Ig Fc domain to forma solubleGPVI.The GPVI-Fc predominantly targets the collagen found onthe exposed subendothelial ECM at the injured vascularsite and prevents the ability of platelet GPVI to bind to theexposed subendothelial ECM. Preclinical studies have

successfully proven that injection of GPVI-Fc improvedendothelial dysfunction and vascular in atheroscleroticrabbits (Ungerer et al., 2013) as well as reduced arterialthrombus formation, cerebral infarct size, and edema afterstroke (Goebel et al., 2013), with improved functional andprognostic outcome with intracranial bleeding. A phase Iclinical trial shows infused GPVI-Fc (10, 20, 40, 80, or160 mg) doses efficiently inhibit collagen-induced plateletaggregation with no alteration in primary hemostasiscompared with predosing. Both tumor necrosis factor–related activation protein– and ADP-induced plateletaggregation were not affected by GPVI-Fc treatment.Revacept could still be detected 3 days after administra-tion of the 10 mg and higher doses, indicating drugelimination is slow. The terminal half-life of the lowerdoses (10 and 20 mg) was lower than in the groupsreceiving higher doses (40, 80, and 160 mg) (Ungereret al., 2011). Platelet inhibition occurred at least 2 hoursafter infusion of Revacept, and the inhibitory effectsremained up to 24 hours after infusion of higher doses(40, 80, and 160mg). Ten and twentymilligrams Revaceptweakly inhibited collagen-induced platelet aggregation7 days after drug administration, whereas higher dosesof Revacept prolonged inhibition of collagen-mediatedplatelet aggregation at later time points. These datasupport the safety and efficacy of Revacept in people, buthave yet to be verified in patients with ACS. A phase IIclinical trial is currently investigating Revacept in stablecoronary artery disease patients scheduled for PCI. Acomposite of death and/orMI injurywill be assessed as theprimary endpoint in this ongoing trial.

E. Thrombin Receptor Inhibitors

Thrombin is a potent agonist that not only activatesplatelets through the proteolytic cleavage of thrombinreceptors PAR1 and PAR4, but also plays a key role inthe coagulation cascade, which is essential for hemo-stasis. An appeal to targeting PARs is that thrombin isthe most potent agonist, and therefore blocking PARactivation would also block the release of secondarymediators (ADP and TxA2) from activated platelets. Inaddition, targeting PARs would avoid directly interfer-ing with thrombin-induced fibrin production involved inhemostatic plug formation. Thus, inhibiting PAR1- orPAR4-induced platelet activation while leaving theother functions of thrombin intact may mitigate therisk of bleeding and provide a more selective effect forprevention of arterial thrombosis.

One of the first PAR antagonists, atopaxar, a small-molecule and reversible inhibitor of PAR1, derived frombicyclic amidine, was shown to competitively bind at ornear the tethered ligand binding site within the secondextracellular loop of the receptor. In a guinea pig modelof photochemically induced thrombosis, orally adminis-tered atopaxar (30 mg/kg) prolonged the time to vesselocclusion by approximately 2 hours compared withcontrol. The pharmacokinetic assessment showed that

Platelet Therapeutics 541

atopaxar was metabolized by CYP3A4, and the onset ofaction occurred after 3.5 hours. However, the agentexhibited a slow rate of elimination with a half-life of23 hours. A phase I study demonstrated that a single doseof 50, 100, and200mgdrug significantly reduced thrombin-induced platelet aggregationwith no severe adverse eventsreported. Two clinical trials (LANCELOT-ACS andLANCELOT-CAD) were conducted to assess the safety ofatopaxar in ACS and chronic coronary artery disease.These trials showed relevant bleeding increased numeri-cally but not significantly in the ACS trial; however, thebleeding rates increased in the CAD trial. Adverse events,such as liver function abnormalities and QT intervalprolongation, were observed with higher doses ($200 mg)of atopaxar administration. Currently, there are no phase3 trials registered to further investigate atopaxar.Vorapaxpar (trade name Zontivity), a synthetic tricyclic

3-phenylpyridine derived from himbacine, has been FDAand EMA approved for the reduction of thromboticcardiovascular events in patients with previous MI orperipheral arterial disease (Déry et al., 2016). Vorapaxar isa reversible, competitive antagonist of PAR1 that binds ator near the tethered ligand binding site within the secondextracellular loop of PAR1 (Kosoglou et al., 2012) with anestimated Kd of 1.2 6 0.3 nM (Hawes et al., 2015).Preclinical studies show vorapaxar can successfully blockthrombin- and PAR1-induced human platelet aggregationwith an IC50 of 47 and 25 nM. Additionally, vorapaxar isshown to have a high oral bioavailability pharmacokineticprofile by which its onset of action occurs 1–2 hours afteroral ingestion (5–40 mg) and has a terminal plasma half-life of 125–269hours. Platelets fromhealthy subjects orallyadministered with a single loading dose of vorapaxarranging from 5 to 40 mg showed PAR1-induced inhibitiongreater than 90% for more than 72 hours followingadministration of the drug (Becker et al., 2009).Vorapaxar was assessed for efficacy and safety in two

large phase 3 clinical trials that includedmore than 40,000patients enrolled with ACS (TRACER clinical trial) andstable atherosclerosis (TRA2°P-TIMI 50 clinical trial). TheTRACER trial concluded that vorapaxar in combinationwith the standard antiplatelet therapy provided no signif-icant net clinical benefit in patients with ACS, butsignificantly increased the risk of major bleeding, mostnotably intracranial hemorrhage in a subset of patients.This was most apparent in patients with prior history ofstroke or transient ischemic attack (TIA), by whichbleedingwas observed to increase incrementally over time,leading to early termination of the TRACER trial. TRA 2°P-TIMI50clinical trialwas thenmodified to enroll patientswith stable atherosclerosis and a history of MI or ischemicstroke (between 2and12months before enrollment) and/orperipheral arterial disease (PAD) (either after revascular-ization or with an ankle–brachial index of ,0.085), butexcluded subset of patients with bleeding risk (history ofstroke or TIA) described in TRACER. Unlike TRACER,TRA 2°P-TIMI 50 excluded the 40 mg single loading dose,

but maintained the daily dose of 2.5 mg. Both the primary(composite of cardiovascular death, MI, or stroke) andsecondary (composite of cardiovascular death, MI, stroke,or urgent revascularization) endpoints TRA 2°P-TIMI50 were reduced compared with the placebo group;however, as observed in TRACER, bleeding complicationswere increased in patients receiving vorapaxar. As a resultof these trials, the FDA approved vorapaxar for theprevention of thrombotic cardiovascular events in patientswith a history of MI or PAD.

In light of the bleeding complications due to PAR1inhibition, PAR4 antagonismhas also been developed as apotential antiplatelet therapeutic avenue. There areseveral lines of evidence that suggest targeting PAR4may also protect against occlusive thrombus formationwithout interfering hemostasis. For instance, PAR1 andPAR4 signal with different kinetics in human plateletactivation, and these differences are thought to contributeto different phases of platelet activation. PAR1-inducedactivation by low thrombin is thought to contribute to theinitial phase of platelet activation involved in hemostasis,whereas later stages of platelet activationmediated by thelow-affinity thrombin receptor, PAR4, are thought to beimportant for occlusive thrombosis. PAR4 activation alsoinduces platelet spreading, contributing to the irrevers-ible formation of a stable thrombus. Therefore, targetingPAR4 while preserving PAR1 signaling may selectivelyinhibit thrombosis without altering hemostasis.

Anti-PAR4 antibodies were used to validate the efficacyand safety of PAR4 inhibition and were shown to restorecarotid blood flow following iron (III) chloride (FeCl3)-induced thrombotic injury compared with the control IgGantibodies. Importantly, anti-PAR4 antibodies did notsignificantly prolong bleed times in the guinea pig cuticle-and renal-provoked bleeding time models (Wong et al.,2017). These data demonstrated the efficacy and lowbleeding liability of PAR4 antagonism, prompting thedevelopment of the first, orally bioavailable, small-moleculeinhibitor, BMS-986120. BMS-986120 was assessed in arandomized, blinded, placebo-controlled study in healthycynomolgus monkeys, demonstrating its efficacy in re-ducing thrombus weight by 83% at 1mg/kg with a twofoldincrease in bleeding time. In contrast, 0.3 mg/kgclopidogrel-administered monkeys showed a 48% reduc-tion in thrombus weight, but with a seven- to eightfoldincrease in bleeding times (Wong et al., 2017). Recently,the completed phase I PROBE clinical trial showed BMS-986120 to exhibit high oral bioavailability in healthyhuman participants (Wilson et al., 2018). The preclinicalandphase I trials have successfully shown the efficacy andlimited bleeding risk in reducing thrombosis in healthyanimals and volunteers, but these findings have yet to beconfirmed in patients at risk for thrombosis.

F. Glycoprotein Ib-IX-V Inhibitors

Anfibatide (also known as agkicetin) is a C-type lectin-like protein derived from the venom of Agkistrodon

542 Yeung et al.

acutus (Chen and Tsai, 1995), which competitively bindsto the region of GPIba that recognizes the A1 domain ofvWF and prevents GPIba-dependent platelet adhesion andactivation under high arterial shear rates (Lei et al., 2014).Preclinical studieswere able to demonstrate anfibatide as apotent and safe antithrombotic agent in inhibiting plateletadhesion and thrombus formation in vitro as well as in vivomodels of thrombosis.Additionally, anfibatidewasshowntoimprove ischemic stroke model induced by middle cerebralartery occlusion in mice. Importantly, anfibatide-treatedmice showed lower incidence of intracerebral hemorrhageand shorter tail bleeding time compared with the tirofiban-treated