Computational Methods to Investigate Dual-Target-Directed Ligands

Molecules designed to modulate single targets typically lack efficacy when dealing with multifactorial diseases. Even combining single-target agents through drug cocktails represents a less-than-ideal strategy, as it can lead to less predictable pharmacokinetics and an increased risk of drug-drug interactions. To overcome these limitations, researchers shift their attention to the rational design of compounds capable of simultaneously modulating multiple targets, thus influencing interconnected biological pathways within a single molecular framework. These compounds, commonly referred to as MultiTarget-Directed Ligands (MTDLs), hold the potential to elicit an additive, and ideally synergistic, disease-modifying effect, while avoiding the issues arising from drug drug interactions. However, their rational design remains challenging, as it depends on both the biological and structural similarity of targets of interest and on the existence of common chemical space between their ligands. In this thesis, we aimed to expand the current understanding of this growing field, focusing primarily on the simplest example of MTDLs, which is represented by compounds rationally designed to modulate two distinct proteins, commonly known as Dual-Target-Directed Ligands (DTDLs). Several computational approaches were employed to systematically analyze the state-of-the-art of the field, revealing common methodologies adopted in the target selection and rational design of dual-active compounds. Building on the obtained insights, we also developed a data-driven pipeline to identify and leverage candidate target pairs that are “hidden in plain sight”, demonstrating that, before venturing into more complex polypharmacological scenarios, many readily available target combinations remain unexplored. One of these identified candidate target pairs, consisting of the Serotonin 3a (5-HT3A) receptor and soluble epoxide hydrolase (sEH), is presented and discussed in detail as a case study. Subsequently, we shifted our attention from data-driven chemoinformatic approaches to the actual discovery of dual dopamine D3 receptor (D3R) / glycogen synthase kinase-3 beta (GSK-3β) modulators from ultra-large libraries-derived chemical space, with potential applications in neuropsychiatric disorders. By combining two complementary computational approaches, we successfully prioritized and identified compound ARN27663, which exhibited the desired dual pharmacological profile.