Heparin, a glycosaminoglycan (GAG), is a complex biopolymer of varying chain length and consisting of uronic acid and glucosamine residues, which are sulfated at various positions. The interaction of heparin with antithrombin is the basis for anticoagulation therapy. Heparin accelerates the antithrombin mediated inhibition of factor Xa and thrombin by a conformational activation mechansism and bridging mechanism, respectively. The sequence specific pentasaccharide DEFGH in full length heparin is the most important fragment for high affinity and activation of antithrombin, without which the heparin is incapable of binding to antithrombin. Although heparin is a commonly used anticoagulant, it suffers from serious side effects including bleeding complications, heparin-induced thrombocytopenia, and intra- and inter-patient dose response variability. Desai and co-workers have shown that it is possible to replace the GAG skeleton by small, non-saccharide sulfated molecules as antithrombin activators. However, the designed molecules were found to be weak activators of antithrombin due to their binding to the extended heparin-binding site (EHBS), instead of the pentasaccharide-binding site (PBS), of antithrombin. To design better non-saccharide antithrombin activators, a virtual screening-based approach was employed. Combinatorial virtual screening of 24576 molecules based on tetrahydroisoquinoline core scaffold resulted in 92 hits that were predicted to bind preferentially in the PBS of activated antithrombin with good affinity. The work resulted in a predicted pharmacophore consisting of a 5,6-disulfated bicyclic tetrahydroisoquinoline and a 2′,5′-disulfated unicyclic phenyl ring connected by a 4- to 5-carbon linker. The work has led to several hypotheses, which are being tested in the laboratory through synthesis and biochemical evaluation. To understand the mechanism of heparin binding to thrombin in greater detail, structural biology and molecular modeling approaches were used. More specifically, the nature of the heparin binding to thrombin was studied with a special focus on understanding the specificity of recognition. Comparative analysis was performed with heparin–antithrombin interaction to assess similarities and differences between the two heparin binding systems. In antithrombin, three important amino acids are involved in heparin pentasaccharide binding, while in thrombin, at least seven basic amino acids are predicted to be involved. For biological systems, one would expect greater specificity with more interacting points. However, the heparin–thrombin system interestingly displays a lack of specificity. The molecular basis for this lack of specificity is not clear. A study of antithrombin and thrombin crystal structures with regard to surface exposure, flexibility, and geometry of basic amino acids present in the respective heparin binding site provides the basis for the specificity of recognition (or lack thereof) in the two systems. Interestingly, analysis of thrombin exosite-II showed that Arg101, Arg165 and Arg233 are spatially conserved and form a local asymmetric center. Using in-silico docking techniques, selected tetrasaccharide sequences were found to specifically recognize this triad of amino acids indicating the possibility of specific recognition of thrombin. This hypothesis led to the design of a putative lead sequence that is 50% smaller in size and contains 62.5% fewer charges in comparison to the literature reported known exosite II sequence. The design of novel putative ‘specific’ exosite II sequence challenges the idea that the thrombin–heparin interaction is completely non-specific and gives rise to novel opportunities of designing specific thrombin exosite-II ligands.

Anticoagulants are the basis for treatment and prevention of thrombotic diseases. The currently available medicines are associated with a wide range of adverse reactions that mandates developing new anticoagulants. Several lines of evidence support the superiority of factor Xa (FXa) as a promising target to develop novel anticoagulants. This work focuses on the design of direct and indirect FXa inhibitors using an interdisciplinary approach. As indirect FXa inhibitors, a focused library of tetrasulfated N–arylacyl tetrahydroisoquinoline (THIQ) nonsaccharide allosteric antithrombin activators was designed, synthesized, and biochemically evaluated to establish their structure–activity relationship (SAR). An N–arylacyl THIQ analog having carboxylate at position–3, two sulfate groups at positions–5 and –8 of THIQ moiety, butanoyl linker, and two sulfate groups at positions–2 and –5 of the phenolic monocyclic moiety was identified as the most promising nonsaccharide antithrombin activator with KD of 1322 ± 237 μM and acceleration potential of 80–fold. Its biochemical profile indicates a strong possibility that it activates antithrombin by the pre–equilibrium pathway rather than the induced–fit mechanism utilized by heparin analogs. A similar interdisciplinary approach was exploited to design direct FXa inhibitors that possess high selectivity and are potentially orally bioavailable. Structurally, the designed direct FXa inhibitors are neutral THIQ dicarboxamides. THIQ dicarboxamide is a privileged structure with a semi–rigid character, a structural feature that potentially offers high selectivity for targeting FXa over other coagulation and digestive proteases. It can also be thought of as an amino acid–like structure, which affords accessibility to a large number of compounds using well established peptide chemistry. Mechanistically, the designed inhibitors were expected to bind to FXa in the active site and function as orthosteric inhibitors. These direct FXa active site inhibitors are also likely to inhibit clot–bound enzyme. Nearly 60 THIQ dicarboxamides were synthesized and biochemically evaluated. Through detailed SAR analysis, the most potent analog was designed and found to exhibit an IC50 of 270 nM (Ki = 135 nM), an improvement of more than 207–fold over the first inhibitor synthesized in the study. The most potent inhibitor displayed at least 1887–fold selectivity for FXa over other coagulation enzymes and a selectivity index of at least 279–fold over the digestive serine proteases. This analog doubled plasma clotting times at 17–20 μM, which are comparable to those of agents being currently studied in clinical trials. Overall, allosteric and orthosteric approaches led to the design of indirect and direct small molecule inhibitors of FXa based on the THIQ scaffold. This work introduces two promising molecules, a tetrasulfated N–arylacyl THIQ analog as a heparin mimetic and a neutral THIQ dicarboxamide as a potent, selective, and potentially bioavailable peptidomimetic, for further advanced medicinal chemistry studies.