Innovative strategies for the drug lead identification and optimization

This PhD thesis was focused on the exploration of several computational strategies that are employed in computer-aided drug design and in the study of small molecules interacting with different targets. The aim of the thesis was to validate the reliability and the effectiveness of these procedures and to combine them in order to develop efficient virtual screening (VS) protocols or target prediction platforms and to structurally optimize promising lead compounds. In the first part, the capability of a pharmacophore-enhanced consensus docking approach to improve VS results and reduce the normally required computing time was tested and validated. Chapter 1 reports the development of a VS protocol combining the pharmacophore filtering of a commercial database with a consensus docking approach that allowed the identification of the first STARD3 inhibitor. Herein, consensus docking was employed for both a qualitative and quantitative analysis. In Chapter 2, the same methodology was applied to a different target and led to the discovery of new potent MbtI inhibitors. Differently, Chapter 3 reports the extensive evaluation of docking-based reverse screening approaches in identifing the proper target of a query ligand. This analysis involved the use of several docking procedures in target prediction studies performed on a benchmark dataset including different targets and a set of known-active compounds. Finally, Chapter 4 describes the structural optimization of a class of potent MAGL inhibitors. Several in vitro analysis and preliminary in vivo studies were performed on the best representatives of this class of compounds.