Killing 2 proteinases with 1 (dual-acting) stone

In this issue of Blood, Olson et al establish the biochemical basis for the specificity and selectivity of a dual-action anticoagulant with the designed feature of rapid neutralization, making it extremely attractive for use in anticoagulated patients if surgery is warranted.¹ 

The endogenous inhibitory and anticoagulant reactions of antithrombin III, tissue factor pathway inhibitor, and activated protein C highlight the effectiveness of targeting multiple steps in the blood coagulation cascade to limit the clotting response.²  Olson et al apply an analogous concept to inhibit both thrombin function and its formation with a bifunctional inhibitor targeting thrombin and factor Xa, the proteinase required for prothrombin activation. Although their approach mirrors the multireactant targeting strategy of endogenous regulation, the twist here is that the bifunctional EP217609 molecule inhibits its 2 targets by different mechanisms.

EP217609 contains a fondaparinux-like component fused through a spacer bearing a biotin moiety to a derivative of N-α-(2-naphthylsulfonylglycyl)-4-amidinophenylalanine piperidide (NAPAP). The fondaparinux-like functionality is intended to bind endogenous antithrombin III with high affinity and enhance the suicide inactivation of factor Xa.³  On the other hand, the NAPAP analog is intended to reversibly bind to the active site of thrombin and inhibit function.⁴  The biotin moiety provides a potential handle for the neutralization of the inhibitory properties of EP217609 by abstraction through high-affinity ligation with avidin. Olson et al's design draws on the established efficacy of heparin pentasaccharide derivatives as anticoagulants without many of the drawbacks of unfractionated heparin but also seeks to resolve the problem of “thrombin-rebound” associated with the cessation of heparin and pentasaccharide therapy.⁵ 

The cleverest of intentions in designing novel inhibitors can be laid to waste by the complexities of coagulation proteinase enzymology. For example, there is no reason to imagine that specificity, selectivity, or other beneficial features of the individual functional groups will be preserved in the fusion construct. There is also the question of whether the NAPAP-like moiety can effectively inhibit thrombin when EP217609 circulates bound to antithrombin III. Despite these qualifications, the approach seemingly works! A prior version of such a fusion-lacking biotin has shown to be a superior anticoagulant in comparison to its components in animal models of arterial and venous thrombosis and also to prevent “thrombin rebound.”⁶  A phase 1 study has reported EP217609 to be well tolerated in healthy subjects with neutralization of its anticoagulant effects after the administration of avidin.⁷  The drug is now in phase 2 clinical trials for cardiopulmonary bypass.

In a second twist that runs counter to the typical course of anticoagulant development, it is only now that proper expertise has been applied to establish the biochemical basis for the apparent efficacy of EP217609. Olson et al, leading investigators in proteinase, serpin, and heparin enzymology, present a thorough study of EP217609 and related compounds to provide the quantitative basis for its selectivity and the function of the components within the fusion construct. Unexpectedly, they find that incorporation of the NAPAP analog into the fusion increases its affinity for thrombin into the 30pM range, yielding an inhibitor that ranks among the highest affinity inhibitors known for this proteinase. High selectivity for thrombin is evident from its 1000-fold weaker binding constant for the next ranked proteinase of relevance to hemostasis. The fondaparinux-like functionality is also preserved, allowing EP217609 to bind antithrombin with nM affinity and act with similar selectivity to accelerate the inhibition of Xa over other proteinase targets. Importantly, they establish only modest deleterious linkage effects, within a 5-fold range, of ligation at one site in affecting the functionality of the second and vice versa. Finally, they also establish that the basis for inhibitor neutralization by avidin results from an approximately 100-fold decrease in thrombin inhibition and an approximately 30-fold decrease in its binding to antithrombin III. This study highlights the power of enzymology, done expertly, in establishing the basis for the selectivity of this novel class of dual-action anticoagulants and provides the mechanistic foundation to show the way forward for their development as therapeutics.