Design, synthesis, and characterisation of small molecule inhibitors of the coagulation factor XIIa (FXIIa)

Thromboembolic disorders are the major cause of death and disability worldwide. Anticoagulants are the mainstay in the prevention and treatment of thromboembolic diseases. However, almost all the currently available anticoagulants cause therapy-related haemorrhages as a side effect. In the recent years, FXIIa was highlighted as an attractive target for the development of new anticoagulant drugs with low rates of therapy-related haemorrhages. In this work, the development of a new class of chemical inhibitors as potent and selective nonpeptidic inhibitors of FXIIa has been described. The structural information of FXIIa is not as prevalent as that of FXa and thrombin. Therefore, the 3D-structure of FXIIa in the active conformation was elaborated by homology studies. Given the structural and functional similarities between FXIIa and FXa, rivaroxaban, a known FXa inhibitor, was used as a starting point for this project. Using a combination of medicinal chemistry and computational chemistry strategies, rivaroxaban analogues with acceptable inhibitory activity against FXIIa were readily identified. However, these early analogues showed poor selectivity profiles. Two important structural features of FXIIa inhibitors have been extracted from iterative make-test cycles, an amine derivative at the P1-position and piperazine derivative at the P4-position. Compound 4-(aminomethyl)-N-({(5S)-2-oxo-3-[4-(piperazin-1-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)benzamide (v8) was synthesised with piperazine at the P4-position and 4-(aminomethyl)benzoyl at the P1-position. It was found that compound v8 inhibited FXIIa activity in a chromogenic biochemical assay with an IC50 value of 0.18 ± 0.1 μM. Interestingly, this compound was found to be 72-fold more potent against FXIIa than FXa. In another set of compounds, compound 4-carbamimidoyl-N-({(5S)-2-oxo-3-[4-(piperazin-1-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)benzamide (z8) was synthesised with a 4-carbamimidoylbenzoyl group at the p1-position and piperazine at the P4-position. This compound was 14-fold more potent against FXIIa than FXa. Also, this compound was 1.5-fold more potent against FXIIa than compound v8, but it was less selective to FXIIa. In a further structural refinement, 4-(aminomethyl)benzoyl was replaced with a 4-carbamoylbenzoyl at the P1-position. This afforded compound N1-({(5S)-2-oxo-3-[4-(piperazin-1-yl)phenyl]-1,3-oxazolidin-5-yl}methyl)benzene-1,4-dicarboxamide (w8) with an IC50 against FXIIa of 0.16 ± 0.05 µM, but with significant improvement in the selectivity factor to FXIIa (selectivity factor = 206) compared to compounds z8 and v8. Docking studies showed that FXa inhibitory activity depends mainly on hydrophobic interactions at the S1- and S4-subpockets. Compounds deprived of the Cl-π interactions at the S1-pocket and/or π-π stacking at the S4-pocket were almost always less potent against FXa than compounds that can make such interactions. However, the inhibitory potency against FXIIa depends mainly on electrostatic interaction in the S1- and S4-pockets especially with Asp189 residue at the bottom of the S1-pocket