Antithrombotic phosphoinositide 3-kinase β inhibitors in humans: a ‘shear’ delight!


Shaun P. Jackson, Australian Centre for Blood Diseases, 6th Level, Burnet Building, 89 Commercial Road, Melbourne, Victoria 3004, Australia
Tel.: +61 3 9903 0131; fax: +61 3 9903 0228.


Jackson SP, Schoenwaelder SM. Antithrombotic phosphoinositide 3-kinase β inhibitors in humans: a ‘shear’ delight!J Thromb Haemost 2012; 10: 2123–6.See also Nylander S, Kull B, Bjorkman JA, Ulvinge J-C, Oakes N, Emanuelsson BM, Andersson M, Skarby T, Inghardt T, Fjellstrom O, Gustafsson D. Human target validation of phosphoinositide 3-kinase (PI3K)β:effects on platelets and insulin sensitivity, using AZD6482 a novel PI3Kβ inhibitor. This issue, pp 2127–36.

Antiplatelet agents have become the cornerstone pharmacologic therapy for acute coronary syndrome (ACS). Used in combination with percutaneous coronary intervention, antithrombotic therapy has had a major positive impact on outcomes in this patient group [1–3]. The relatively high frequency of recurrent thrombosis with single antiplatelet therapy has necessitated the need for more intensive approaches, with dual antiplatelet therapy (aspirin and P2Y12 antagonists) becoming the standard of care for ACS. Although patient outcomes continue to improve, major cardiovascular events can still occur in up to 10% of patients within 3 years of commencing dual antiplatelet therapy. These statistics are based on findings from closely managed phase III clinical trials, and the ‘real-world’ incidence of second coronary events is likely to be considerably higher.

Although more potent regimens of antithrombotic agents are reducing ischemic events, this is coming at the cost of increased bleeding. Dual antiplatelet therapy is associated with a 4% annual incidence of major bleeding, and triple therapy, in which an anticoagulant is combined with dual antiplatelet therapy, is associated with major bleeding in up to 12% of patients [4]. Bleeding is associated with adverse cardiovascular outcomes, and carries a similar risk of death as recurrent ischemia [5].

Bleeding is also the paramount consideration for ischemic stroke, the second commonest manifestation of cardiovascular disease. Unlike in ACS, the benefits of dual antiplatelet therapy are partially outweighed by negative impact of intracerebral hemorrhage. Furthermore, unlike in ACS, antiplatelet agents are not routinely coadministered with thrombolytic agents in stroke, owing to an increased risk of intracerebral bleeding [6]. Increasing awareness of the adverse impact of bleeding may necessitate a re-evaluation of the antithrombotic regimens required to maximize efficacy without increasing the risk of bleeding.

In this context, the report by Nylander et al. [7], in this edition of the journal, is timely. They have performed a preclinical and phase I human study investigating the effects of a phosphoinositide 3-kinase (PI3K) inhibitor (AZD6482), selective for the p110β isoform (PI3Kβ), on platelet function, thrombosis, and bleeding. PI3Kβ has previously been proposed as an antithrombotic target, on the basis of the ability of isoform-selective PI3Kβ inhibitors to inhibit shear activation of platelets, stable platelet aggregation and thrombus formation without causing a significant increase in bleeding [8]. The comprehensive studies reported by Nylander et al. confirm this, and demonstrate an impressive separation of antiplatelet and antithrombotic effects from bleeding, in both humans and dogs. These findings provide important validation of PI3Kβ as a potentially important antithrombotic target in humans, and provide further support for the concept that PI3Kβ inhibitors may have a wide therapeutic window and represent a potentially safe antithrombotic approach [8].

PI3Ks and platelet function

The PI3Ks are an intensely investigated family of signaling enzymes that regulate a broad range of cellular responses, including cell survival and proliferation, cell adhesion, cytoskeletal remodeling, and membrane trafficking [9,10]. These enzymes are divided into three distinct classes, on the basis of their structure, substrate preferences, and regulatory mechanisms. The class I subfamily is by far the most extensively characterized, and consists of four distinct isoforms: PI3Kα, PI3Kβ, PI3Kδ, and PI3Kγ. All four isoforms preferentially phosphorylate phosphatidylinositol 4,5-bisphosphate to generate the lipid second messenger phosphatidylinositol 3,4,5-trisphosphate. PI3Kα and PI3Kβ are ubiquitously expressed, whereas the expression of PI3Kδ and PI3Kγ is more restricted, with high levels of expression in hemopoietic cells. Although platelets express each of the class I isoforms, PI3Kβ appears to play the dominant role in regulating platelet adhesive function.

Numerous studies with the isoform-selective inhibitor TGX-221 [8,11–17], and more recent studies with genetically engineered mouse models lacking PI3Kβ catalytic function [18,19], have revealed a major role for PI3Kβ in regulating signaling downstream of tyrosine kinase and G-protein-coupled signaling (Fig. 1). In platelets, PI3Kβ appears to play a major role in regulating signals downstream of the Gi-coupled P2Y12 receptor necessary for Rap1b activation and sustained integrin αIIbβ3 (glycoprotein [GP]IIb–IIIa) activation [15]. PI3Kβ also appears to play an important role in adhesion receptor signaling, including that involving GPIbα, integrins αIIbβ3 and α2β1, and the major ITAM receptor, GPVI [8,11,13,14,18,19]. It is likely that the major effect of PI3Kβ inhibition on platelet function occurs through the modulation of cytosolic calcium flux and activation of Rap1b (and possibly AKT), which are key signaling events modulating sustained integrin αIIbβ3 activation. As a consequence, PI3Kβ inhibitors undermine stable platelet aggregation in vitro and platelet thrombus formation in vivo, affording protection from thrombotic occlusion in arteries [8,19].

Figure 1.

 Phosphoinositide 3-kinase p110β (PI3Kβ) signaling in platelets. Platelet activation pathways are under the influence of multiple input signals, induced by adhesive proteins and soluble agonists. Ligation of glycoprotein (GP)Ib–V–IX and GPVI by von Willebrand factor (VWF) and collagen, respectively, initiates intracellular signaling via non-receptor tyrosine kinases, leading to the activation of phosphoinositide 3-kinase (PI3K), calcium mobilization, and Rap1b activation. A similar pathway operates downstream of integrin αIIbβ3 to promote platelet activation. The local release and/or generation of soluble agonists at sites of vascular injury potentiates integrin αIIbβ3 activation and thrombus growth. Receptors for thrombin (protease-activated receptor-1 [PAR1]), thromboxane A2 (TXA2) (TP) and ADP (P2Y1) are G-protein-coupled receptors linked to Gq, leading to mobilization of intracellular calcium and activation of Rap1b. ADP also signals through a second purinergic receptor, P2Y12, which is a Gi-coupled receptor that potentiates the activation of PI3K and Rap1b. Several of these pathways have been targeted therapeutically, with P2Y12 antagonists (clopidogrel, prasugrel, and ticagrelor), integrin αIIbβ3 antagonists (abciximab, tirofiban, and eptifibatide), PAR1 antagonists (SCH530348 and E5555), inhibitors of platelet signaling enzymes (cyclooxygenase-I: aspirin), and inhibitors of thrombin generation or function (including the heparins, direct thrombin inhibitors, and vitamin K antagonists). PI3Kβ is an important signaling molecule regulating integrin αIIbβ3 activation in platelets, and is the major PI3K isoform signaling downstream of GPIb–V–IX, integrins αIIbβ3 and α2β1, and GPVI. It also plays a major role in sustaining Rap1b activation downstream of P2Y12. Inhibition of PI3Kβ dysregulates cytosolic calcium flux and Rap1b activation, thereby undermining sustained integrin αIIbβ3 activation and stable thrombus growth. LMWH, low molecular weight heparin; UFH, unfractionated heparin.

PI3Kβ inhibitor AZD6482

AZD6482 is the active enantiomer of a racemic mixture, originally named KN-309, a closely related structural analog of TGX-221 with improved pharmaceutical properties [20]. It has been extensively characterized in the study by Nylander et al. [7], in terms of its PI3K isoform selectivity, and its effects on platelet aggregation and thrombus formation, bleeding, and insulin signaling. It is a potent PI3Kβ inhibitor (IC50 10 nm) with high selectivity for PI3Ks relative to a broad range of protein kinases. It has a rapid onset of action and a short half-life in rats, dogs, and humans, so sustained platelet inhibition requires continuous infusion of the compound. This is potentially advantageous, as the antiplatelet effect of AZD6482 can be rapidly reversed with cessation of the infusion.

AZD6482 was demonstrated to be an effective inhibitor of ADP and collagen-induced platelet aggregation, with a similar potency to that reported with TGX-221 and PI3Kβ kinase-dead mouse platelets. AZD6482 is also a highly effective inhibitor of shear-induced platelet aggregation. The dose-dependent effects of AZD6482 on ADP-induced platelet aggregation correlate closely with the effects on thrombus formation in the dog, with complete protection from arterial thrombotic occlusion at a plasma concentration (1 μm) that inhibited aggregation by > 80%. Notably, at these plasma levels, AZD6482 had no effect on bleeding time or blood loss [7].

AZD6482 appears to be well tolerated in humans following a 3-h intravenous infusion, with no significant drug-related adverse events [7]. At ‘therapeutic’ plasma concentrations (∼ 1 μm) that produce marked inhibition of ADP-induced and shear-induced platelet aggregation, there was no increase in skin bleeding time. Only a slight increase in bleeding time (1.6-fold) was noted at five-fold to six-fold higher concentrations, which is broadly consistent with findings in rats and mice. The ex vivo antiplatelet effect and bleeding time data in the dog paralleled the human findings, and it appears that the impressive antithrombotic therapeutic window of PI3Kβ inhibitors observed in preclinical animal studies may be translatable to humans.

A concern with PI3Kβ inhibitors is their potential impact on insulin signaling. Whereas PI3Kβ was initially thought to play a major role in promoting GLUT4 transport downstream of the insulin receptor, more recent studies have demonstrated a dominant role for the PI3Kα isoform in this process. The findings from Nylander’s study support this, demonstrating that a short infusion of AZD6482 at ‘therapeutically relevant’ concentrations is unlikely to have a clinically significant impact on insulin signaling and plasma glucose levels. Whether chronic PI3Kβ inhibition will lead to insulin resistance is an important issue for future investigation, as mice lacking catalytically active PI3Kβ develop mild insulin resistance with aging.

PI3Kβ inhibitors in the clinic

So what are the potential indications for a short-acting antiplatelet agent that does not have a marked effect on bleeding? The authors suggest the possible use of AZD6482 in patients undergoing cardiopulmonary bypass surgery involving extracorporeal circulation. Platelet activation is a common feature of extracorporeal circulation, owing to platelet contact with artificial surfaces and exposure to shear stress, leading to platelet dysfunction, which contributes to bleeding and thromboembolic complications [21]. Blockade of GPIIb–IIIa is an effective means of reducing platelet interaction with artificial surfaces; however, the half-lives of GPIIb–IIIa inhibitors used in the clinic are sufficiently long to potentially exacerbate postoperative bleeding [22]. The PI3Kβ inhibitor TGX-221 has previously been demonstrated to reduce platelet interactions in extracorporeal circulation, reducing platelet aggregation and platelet–leukocyte interactions [23,24]. It remains to be demonstrated whether a PI3Kβ inhibitor can be used safely in combination with standard-dose unfractionated heparin, or may require a reduction in the level of anticoagulation. Notably, in preclinical animal models, PI3Kβ inhibitors in combination with anticoagulants cause significantly less bleeding than aspirin or P2Y12 antagonists [8], indicating that this approach may be well tolerated.

A rapid-acting PI3Kβ inhibitor could also be beneficial in the acute management of ischemic stroke. Combining antiplatelet agents with thrombolytic therapy has proven to be beneficial in the setting of ACS, improving vessel reperfusion and reducing reocclusion. In stroke, this strategy has proven less effective, owing to the increased incidence of intracerebral hemorrhage, particularly with GPIIb–IIIa inhibitors [25]. Aspirin and clopidogrel have a slow onset of action, and have the potential to exacerbate bleeding in combination with tissue-type plasminogen activator [26]. Whether a rapid-acting PI3Kβ inhibitor will enhance vessel reperfusion and reduce vessel reocclusion without increasing bleeding during thrombolysis remains to be established. A theoretical concern with PI3Kβ inhibitors is their propensity to promote distal embolism of thrombotic material and exacerbate downstream vessel occlusion. In our own experience, P2Y12 antagonists cause a similar defect in platelet thrombus stability and distal embolization, so it remains to be seen whether these features of PI3Kβ inhibitors are problematic.

If PI3Kβ inhibitors are ultimately found to be safe and effective in the clinic, particularly when used in combination with other antithrombotic and thrombolytic agents, there is enormous scope for their use in the management of a broad range of cardiovascular diseases. The studies presented here on AZD6482 represent a promising next step in the validation of a potentially exciting new antithrombotic target.

Disclosure of Conflict of Interests

The authors were inventors on the patent originally describing the racemic mixture KN-309, from which AZD6482 was derived.