Integration of an Inhibitor-like Rule and Structure-based Virtual Screening for the Discovery of Novel Myeloperoxidase Inhibitors

Myeloperoxidase (MPO) is an attractive therapeutic target against inflammation. Herein, we developed an inhibitor-like rule, based on known MPO inhibitors, and generated a target database containing 6546 molecules with privileged inhibitory properties. Using a structure-based approach validated by decoys, robust statistical metrics, redocking, and cross-docking, we selected 10 putative MPO inhibitors with high chemical diversity. At 20 mu M, six of these 10 compounds (i.e., 60% success rate) inhibited more than 20% of the chlorinating activity of the enzyme. Additionally, we found that compound ZINC9089086 forms hydrogen bonds with Arg233 and with the hemic carboxylate. It makes a pi-stacking interaction with the heme group and displays a high affinity for the enzyme active site. When incubated with purified MPO, ZINC9089086 inhibited the chlorinating activity of the enzyme with an IC50 of 2.2 +/- 0.1 mu M in a reversible manner. Subsequent experiments revealed that ZINC9089086 inhibited hypochlorous acid production in dHL-60 cells and human neutrophils. Furthermore, the theoretical ADME/Tox profile indicated that this compound exhibits low toxicity risks and adequate pharmacokinetic parameters, thus making ZINC9089086 a very promising candidate for preclinical anti-inflammatory studies. Overall, our study shows that integrating an inhibitor-like rule with a validated structure-based methodology is an excellent approach for improving the success rate and molecular diversity of novel MPO inhibitors with good pharmacokinetics and toxicological profiles. By combining these tools, it was possible to increase the assurance rate, which ultimately diminishes the costs and time needed for the acquisition, synthesis, and evaluation of new compounds.