New αIIbβ3 antagonists: Existing receptor antagonists are ligand-mimetics, which may cause a conformation change in αIIbβ3 to a high-affinity state, leading to either paradoxical platelet activation or exposure of ligand-induced binding sites that trigger antibody-mediated platelet clearance and thrombocytopenia in some patients (32, 33). They also are highly potent and associated with a significant increase in bleeding risk, and all require intravenous administration, limiting their utility for long-term therapy. Zalunfiban (RUC-4) is a small molecule inhibitor designed to alleviate this, as it binds to the metal ion-binding site on GPIIIa, maintaining the receptor in the low-affinity state incapable of fibrinogen binding, and therefore, does not induce a conformational change (34, 35). It can also be subcutaneously administered, which would favor its use in urgent settings (34). It showed efficacy in a Phase I study in healthy volunteers and stable CAD patients on aspirin, as it produced high-grade inhibition of ADP-induced platelet aggregation within 15 min of administration with rapid return to normal platelet function within the next 2 h (36). It is currently in a Phase IIb trial. Intracellular inhibitors of αIIbβ3 have also been developed which disrupt integrin activation, preventing the switch to the high-affinity state and subsequent outside-in signaling (37).

New PAR1 antagonists: G protein-coupled receptors (GPCR) are cell surface receptors which upon ligand binding undergo conformational change and activate the associated cytosolic G protein, which further activates an intracellular signaling process. Pepducins are cell-penetrating lapidated peptides that are designed to selectively target the intracellular receptor-effector interface of a GPCR, by conjugating the intracellular loop portion of the receptor, including PAR1 (38). Existing PAR1 inhibitors interfere with both prothrombotic and cytoprotective downstream pathways, while the pepducin technology allows for selective control of downstream signaling pathways (39). PZ-128 is a pepducin inhibitor of PAR1 proposed for CAD treatment that has completed a Phase II clinical trial, showing that it was well tolerated in ACS patients undergoing PCI and did not cause bleeding even when administered on top of DAPT and heparin (40).

P2Y1 targeting: Platelets express P2Y1 and P2Y12 ADP-receptors (15, 46). MRS2500, an adenosine analog developed as a specific P2Y1 antagonist, has been shown to provide strong protection against systemic thromboembolism upon intravenous injection of collagen and adrenalin in mice, while only moderately prolonging bleeding time (47), and exhibited antithrombotic effects in cynomolgus monkeys (48). GLS-409, an analog of the naturally occurring compound adenosine tetraphosphate, inhibits ADP-induced platelet aggregation, and significantly inhibits thrombosis in animal models, with minimal increase in bleeding time (49). To date, there is no clinical trial assessing the P2Y1 antagonists.

Glycoprotein VI (GPVI) targeting: GPVI is the major collagen receptor on platelets (50). Glenzocimab (ACT017) is a humanized monoclonal fragment antigen-binding (Fab) domain against platelet GPVI, which has been shown to inhibit aggregation and procoagulant activity on collagen-stimulated platelets, as well as adhesion and thrombus formation on collagen surfaces in vitro (51, 52). A Phase II/III trial is currently being conducted to evaluate its efficacy and safety in acute ischemic stroke and a Phase IIb in treating myocardial infarction is planned for the near future. SAR264565 is another humanized Fab directed against GPVI, which potently inhibits collagen-induced platelet aggregation (53). Revacept, on the other hand, is a soluble dimeric GPVI-Fc fusion protein containing the fragment crystallizable (Fc) portion of human IgG1, and was developed to specifically bind fibrillar collagen (54). Revacept binds collagen, preventing interaction with circulating platelets and von Willebrand factor (VWF), and resulting in antithrombotic effects. As it does not directly bind platelets, it does not interfere with platelet activity or cause thrombocytopenia (54, 55). It has recently completed a Phase II clinical trial (56, 57). Recently, new selective GPVI antagonists have been identified from large chemical database screening as reported by Olğaç et al. (58), which showed promising antithrombotic properties in vitro. Finally, nanobodies raised against the extracellular domain of GPVI have emerged as potential therapeutics (59).

GPIb-VWF interaction targeting: GPIbα is a central component of the GPIb-IX-V complex expressed on platelets, which binds to VWF (60). VWF binding to collagen and its unfolding under higher shear rates exposes binding sites to platelet GPIb and mediates platelet adhesion and thrombus formation (61). Several GPIb-blocking antibodies have been developed and tested for their antiplatelet effects, such as the monoclonal antibody h6B4-Fab, which inhibits ristocetin-induced platelet aggregation in non-human primates, with only a mild prolongation of skin bleeding time (62, 63). Anfibatide is a C-type lectin (CLEC) purified from snake venom, which inhibits the binding of VWF and α-thrombin to GPIbα (64). It showed strong inhibition of murine and human platelet thrombus formation at low and high shears in ex vivo experiments (65), with a Phase I study in healthy volunteers showing good inhibitory effects without significant prolongation in bleeding time. Serine protease inhibitor (SERPIN)-based fusion proteins have also been proposed as antiplatelet therapy (66). A fusion protein called TaSER (targeted SERPIN), consisting of a variable heavy domain of heavy chain (VHH) with function-blocking activity against GPIbα and a thrombin-specific SERPIN, has been shown to block VWF binding and limit thrombus formation (67). Various monoclonal antibodies directed against VWF are also being pursued as antiplatelet therapy. AJW200 is a humanized IgG4 monoclonal antibody against the A1 domain of VWF, whereas 82D6A3 is directed against the A3 domain (68, 69). They both show promise in animal models and the former has completed a Phase I study, showing therapeutic promise without prolonging the skin bleeding time. ALX-0081 is a first-in-class, bivalent humanized nanobody directed against the A1 domain of VWF (70). The bivalency allows for high-affinity interaction with VWF-A1, leading to potent inhibition. Preclinical studies in cynomolgus monkeys showed potency in inhibiting ristocetin-induced platelet aggregation with 1.6- and 6-fold less prolongation of bleeding time compared to clopidogrel and abciximab, respectively (70), while studies in a guinea pigs thrombotic stroke model demonstrated reduction in brain infarct size (71). Caplacizumab, derived from ALX-0081, was approved in Europe in 2018 and in the USA in 2019 for use in acquired thrombotic thrombocytopenia purpura (TTP) (72). However, after an inconclusive Phase II trial, drug development for atherothrombotic indications was discontinued (73). Finally, nucleic acid aptamer technology has also been pursued. Aptamers are single-strand DNA or RNA molecules that form 3D structures which specifically bind to their target protein. They are potentially superior to antibodies due to their manufacturing cost, specificity, small size, lack of immunogenicity, and ease of reversal. ARC1779 is a nuclease-resistant aptamer designed to bind to VWF-A1, inhibiting VWF-dependent platelet aggregation (74). ARC15105 is a second generation VWF-A1 aptamer with improved potency and pharmacokinetics (75). BT200, derived from ARC15105, has been successfully tested in healthy volunteers and in a Phase IIa trial (76, 77). More recently, a novel DNA aptamer TAGX-0004, which targets VWF-A1 with very high affinity and specificity was generated (78). TAGX-0004 contains a unique mini-hairpin DNA structure which confers further resistance to nuclease degradation thus extending its half-life in vivo. It showed superior thrombus inhibition to ARC1779 and comparable to caplacizumab (78).

Adenosine A2A and A2B receptor targeting: Adenosine is a purine metabolite in plasma resulting from ecto-50-nucleotidase activity (7). It has a very short (∼1 s) half-life due to enzymatic conversion and has been used widely as an antiarrhythmic agent (7). Platelets express adenosine GPCR receptors, and their activation leads to inhibition of platelet activation and aggregation by increasing the cellular cAMP level (79). Several adenosine receptor agonists are being evaluated as antiplatelet agents. Previously evaluated existing adenosine receptor agonists include 5′-N-ethylcarboxamidoadenosine (NECA), HE-NECA, CGS 21680, 2-chloroadenosine, and PSB-15826, which were shown to possess antiplatelet effects in vitro, and antithrombotic effects in vivo (80). However, relatively high doses were required for efficacy, leading to a strong vasodilatory activity and various off target effects.

C-type lectin-like type II (CLEC-2) targeting: Platelets express CLEC-2, which activates various pathways including platelet activation and thromboinflammation via various endogenous ligands such as podoplanin and rhodocytin (81). A recombinant rhodocytin, derived from a snake venom protein, was developed and shown to inhibit CLEC-2 in in vitro and animal models (82). A small molecule, 2CP, was also reported to act as a chemical inhibitor of CLEC-2 (83). Anti-CLEC-2 mAb 2A2B10 has also been shown to suppress thrombus formation without significant bleeding effects in animal cancer model (84).

Serotonin/ 5-hydroxytryptamine (5-HT)-receptor interaction targeting: Platelets express 5-HT subtype 2A receptor (85). Serotonin, packaged in dense granules, is a mild platelet activator, and selective serotonin reuptake inhibitors (SSRIs) inhibit 5-HT reuptake by blocking the serotonin transporter (SERT), thereby acting as antiplatelet agents (86). Direct 5-HT2A receptor antagonists are also being investigated for antiplatelet activity. Currently tested SSRIs are MCI-9042 (sarpogrelate) is an SSRI currently being evaluated (87, 88), while several selective small molecule 5-HT2A receptor antagonists are being investigated, including APD791 (temanogrel), SL65.0472-00, and two existing antidepressant cyproheptadine and pizotifen (89–91). Sarpogrelate significantly reduced serotonin and collagen-induced acute pulmonary thromboembolic death in mice (87). In a trial with stroke patients, however, sarpogrelate failed to demonstrate noninferiority to aspirin, although it was associated with a reduced rate of bleeding complications (88).

Targeting platelet signaling components (Figure 4). (1)

Targeting signaling downstream of PAR1: Parmodulins are novel small molecule allosteric inhibitors of PAR1 which act on cytosolic Gαq subunit signaling downstream of the receptor, inhibiting integrin activation and platelet aggregation. Parmodulin 2, the most selective compound developed, further exhibits anti-inflammatory activity through PAR1 signaling inhibition in endothelial cells, in mouse models (92, 93). No parmodulin clinical trials have started yet.

P-selectin targeting: Platelets store the adhesion molecule P-selectin in α granules and express it on the cell surface during activation (94). P-selectin plays a critical role in mediating platelet-leukocyte adhesion, and is therefore implicated in both thrombosis and inflammation (95). Antiplatelet effects of P-selectin inhibition have been recognized by investigations using small molecules or monoclonal antibodies. PSI697, a small molecule P-selectin antagonist, reduced both arterial and venous thrombosis in animals (96), while THCMA, a nanomolecule, was shown to reduce venous thrombosis without inducing bleeding in animal models (97). Crizanlizumab, a humanized IgG2 monoclonal antibody, was approved in 2019 for patients with sickle cell disease with vaso-occlusive symptoms (98). Inclacumab, a monoclonal antibody that reduces myocardial damage in NSTEMI patients treated with PCI without bleeding complications, is currently being investigated in a Phase III trial (99).

Phosphoinositide 3-kinase β (PI3Kβ) targeting: PI3K signaling in platelets plays a major role in the formation of stable shear-dependent αIIbβ3 platelet adhesion (100). Platelets express class I and class II PI3Ks and among these, the class I PI3Kβ plays an important role in platelet activation responses triggered by the P2Y12 activation, GPVI ligation, and by αIIbβ3 outside-in signaling (100). Several PI3Kβ antagonists have been studied and shown to exhibit strong antithrombotic effects without significantly affecting hemostasis (101), including a peptide inhibitor TGX-221, which are consistent with the phenotype of PI3Kβ knockout mice (102). AZD6482, a selective inhibitor of PI3Kβ that blocks its interaction with ATP, is currently in a Phase IIa study (103). MIPS-9922, another selective inhibitor of PI3Kβ, showed promise in mouse studies, while berberine and 2v, its derivative, have been investigated and show antiplatelet aggregation effects (104, 105).

Spleen tyrosine kinase (Syk)-targeting: Syk is a non-receptor tyrosine kinase important for immunoreceptor tyrosine-based activation motif (ITAM)-dependent platelet activation and in GPVI-induced platelet activation (106). Inhibition of Syk in a mouse model protects against arterial thrombosis without altering bleeding time (106). Fostamatinib is a first-in-class product and the only Syk inhibitor approved by the FDA, and is indicated for treating immune thrombocytopenia (ITP) (107). It provides additional, mild inhibition of platelet aggregation when combined with aspirin and/or ticagrelor (108).

Tyrosine kinase targeting: Tyrosine kinase inhibitors (TKIs) are widely used as targeted strategies in cancer treatment, with many tyrosine kinases being highly expressed in platelets. Indeed, many TKIs are associated with suppression of platelet activation and mild bleeding. Ibrutinib, a Bruton tyrosine kinase irreversible inhibitor, was shown to reduce platelet adhesion and thrombus formation in vitro (109, 110). However, due to the ubiquitous expression of tyrosine kinases throughout the body, off-target effects might be inevitable.

Protein disulfide isomerase (PDI)-targeting: PDI catalyzes disulfide bond formation and cleavage and acts as chaperone of protein folding (111). PDI is expressed on the surface of resting platelets and secreted upon activation (111). Platelet PDI was shown to regulate thrombus growth without affecting platelet adhesion or fibrin generation in an arterial injury mouse model. Anti-PDI antibodies and bacitracin, a nonselective PDI inhibitor, reduced thrombus formation and fibrin generation (112). As PDI is highly ubiquitous in tissue distribution, and furthermore, available PDI inhibitors are non-selective and often cytotoxic, off target effects remain a challenge in drug development. Recently, however, isoquercetin and myricetin, members of the flavonoid family, have gained interest and are being investigated as antiplatelet agents, with promising results in a Phase II clinical trial for the former (113, 114).

12-Lipoxygenase (12-LOX) targeting: 12-LOX is an oxygenase predominantly expressed in platelets, participates in arachidonic acid metabolism, producing downstream eicosanoids which are involved in platelet activation and proinflammatory response. In platelets, 12-LOX plays a role in αIIbβ3 activation and FcγRIIa-, PAR4-, and GPVI-mediated platelet activation (115, 116). ML355 is a potent and selective 12-LOX inhibitor, which has been investigated as an antiplatelet and shown promising results. ML355 inhibited platelet adhesion and thrombus formation under flow in vitro, and similarly reduced thrombus formation in mice without significantly affecting normal hemostasis function (117).

FXa mimetics: Andexanet alfa is a recombinant version of FXa in which the membrane-binding Gla domain has been removed and the active site serine residue mutated. Thus, andexanet alfa is catalytically inactive but still able to bind FXa inhibitors, either pharmacologic or endogenous (134). Unlike idarucizumab, andexanet alfa does appear to have some risk of overcorrection, as a recent safety study reported that ∼10% of patients experienced at least one thrombotic event within 30 days of receiving andexanet alfa, unless anticoagulation therapy was restarted (135). This is consistent with in vitro work, which indicated that andexanet alfa can increase plasma thrombin generation in the absence of FXa inhibition (136, 137). This effect may be due to the interaction of andexanet alfa with plasma FXa inhibitors, such as antithrombin and tissue factor pathway inhibitor (TFPI), and their neutralization.

As an alternative approach, Thalji et al. described FXa^I16l (138). This single amino acid substitution destabilizes the FXa active site enough that it is resistant to inactivation by plasma inhibitors, such as antithrombin and TFPI, and effectively reverses both rivaroxaban and dabigatran in murine models. FXa^I16l has been safe in Phase I and Phase Ib clinical trials (139, 140). Similarly, Verhoef et al. described a FXa homolog found in the venom of Pseudonaja textilis (the eastern brown snake), which is resistant to direct FXa inhibitors (141). They utilized this homolog as the basis to design human FXa mutants, with alterations in the substrate binding pocket, which exhibit similar resistance properties. One of these compounds, termed VMX-C001, recently completed a Phase I clinical trial.

Small molecules: Ciraparantag is a small peptide mimetic, which was originally developed to reverse heparins, but found to also reverse DOACs (both thrombin and FXa inhibitors) (142). Ciraparantag functions by directly binding the thrombin and FXa inhibitors and blocking interactions with their respective target proteases. It safely reversed the anticoagulant activities of edoxaban, apixaban, and rivaroxaban in healthy subjects (143, 144).

Monoclonal antibodies: Another approach involves parenteral administration of monoclonal antibodies that block clotting factor activation and/or activity. Several antibodies targeting FXI are in development, with differing modes of action: (a) Abelacimab binds the active site of FXI/XIa, locking it in an inactive zymogen-like state. As such, it both prevents the activation of FXI by thrombin or FXII and blocks the activity of FXIa (158, 159). (b) Osocimab binds an allosteric site on FXIa, blocking its activation of FIX (160, 161). Xisomab, inhibits the activation of FXI by FXIIa (162). Garadacimab (formerly CSL 312) is one of the few agents targeting FXII as a monoclonal antibody (163, 164). Unlike ASOs, monoclonal antibodies can be used both in acute and chronic settings, as they have a faster response. ASOs require uptake in the liver and clearance of the existing plasma protein, before their effect may be realized (151).

Small molecule inhibitors: There are both orally and parenterally available small-molecule inhibitors of FXI, such as milvexian (BMS-986177/JNJ-70033093) and asundexian (BAY 2433334), in development. In a Phase II trial of 1242 knee arthroplasty patients, milvexian (25–200 mg) twice daily effectively prevented venous thromboembolism with lower bleeding risk than enoxaparin. Once daily milvexian also showed efficacy in VTE prevention (165). Milvexian is now being assessed in a Phase III trial. Similarly, asundexian at doses of 20 mg and 50 mg once daily demonstrated near-complete in vivo FXIa inhibition. These dosages resulted in reduced rates of bleeding, and similar rates of thrombosis, compared to the standard dosing of apixaban (166), suggesting that asundexian is a safer alternative. However, there are concerns about the efficacy of FXIa inhibition, based on these results (152).

Targeting activators of the contact system: In addition to directly targeting the contact pathway, it may also be possible to target components that activate this pathway, such as NETs and polyphosphates. As recently reviewed elsewhere (167), multiple therapeutic strategies are in pre-clinical or clinical stages of development to target NETs, or the generation of NETs. These include approaches to directly or indirectly inhibit the release of nuclear content from neutrophils, the degradation of DNA by DNAse, and the targeting of NET-associated proteins, such as myeloperoxidase. In addition, Baron et al. recently reported that Selinexor, a first-in-class inhibitor of nuclear export approved for use in multiple myeloma patients, effectively prevents NET release in vitro (168). Selinexor may not be an effective antithrombotic, however, as it is associated with reduced platelet count, due to an off-target effect on megakaryocytes (169).

Similarly, polyphosphates may be targeted to prevent thrombosis. La et al. developed a polycationic compounds (termed macromolecular polyanion inhibitors, MPIs) that bind and neutralizes polyphosphates (170). MPIs bind to platelet-released polyphosphates, and reduce thrombus formation in a cremaster muscle laser injury model in mice. In contrast, they did not impair hemostasis in a tail bleeding model.