Molecular optimization of O-glycoside inhibitors of blood coagulation factors.

Dabigatran etexilate, a selective thrombin (fIIa) inhibitor, and factor Xa (fXa)-selective inhibitors, namely apixaban and rivaroxaban, are new oral active anticoagulants (NOACs), which overcome a number of drawbacks associated to traditional oral anticoagulants (e.g., warfarin) in the therapy of thrombotic disorders (1). Recently, we reported compound 1, an isonipecotamide-based inhibitor of the serine proteases of the blood coagulation cascade (2,3,4), and its β-D-glucose-containing analogue 2. The latter compound proved to be a picomolar inhibitor of fXa, with good anticoagulant and profibrinolytic activities. Interestingly, glucosilation resulted in a significant increase of fXa/fIIa inhibition (2,4) (2, fXa Ki = 0.090 nM; fIIa Ki = 100 nM). As shown previously, the chlorothiophene moiety is essential for binding of both compounds, whereas comparing the inhibition constant value of 2 with that of the parent compound 1 clearly showed that the removal of the glucose moiety reduces the affinity for fXa by less than ten-fold and for fIIa by more than two orders of magnitude. Moreover, removing the piperidine moiety does decrease affinity to fXa by several orders of magnitude. Experimental deconstruction of 2 into smaller fragments revealed a binding cooperativity of the piperidine and propylene-linked β-D-glucose fragments, stronger in fIIa (15.5 kJ∙mol-1) than in fXa (2.8 kJ∙mol-1). For a better understanding of the observed binding cooperativity, the crystal structure of the human α-thrombin in complex with the O-glucoside derivative 2 (pdb: 4N3L) has been determined at 1.94 Å resolution, which revealed critical hydrogen bond interactions between the glucose moiety and two basic residues of the Na+-binding site (R221a and K224), involved in allosteric activation of thrombin. Replacing the glucose moieties with other sugars (i.e. galactose, xylose, and glucuronic acid) revealed the importance of maintaining the β-glucose moiety to stabilize the ligand/enzymes complex. Surface plasmon resonance (SPR) studies and docking calculations provided helpful information for optimizing the design of novel fXa/fIIa inhibitors. SPR has been also used to preliminarily assess parameters related to bioavailability.