LED3Score in De Novo Design of Synthetically Accessible Inhibitors of Monoglyceride Lipase
Martin Šícho,2, Alan Kai Hassen3, Anthe Janssen4, Yorick van Aalst1, Sohvi Luukkonen1, Gerard van Westen1 and Mike Preuss3
1Leiden Academic Centre of Drug Research, Leiden University, The Netherlands
2CZ-OPENSCREEN: National Infrastructure for Chemical Biology, Department of
Informatics and Chemistry, Faculty of Chemical Technology, University of Chemistry and
Technology Prague, Technick ́a 5, 166 28, Prague, Czech Republic
3Leiden Institute of Advanced Computer Science, Leiden University, The Netherlands
4Leiden Institute of Chemistry, Leiden University, The Netherlands
Monoglyceride lipase (MAGL) has recently became the focus of many drug dis-covery efforts due to its important role in lipid metabolism and its involvement in various neurological and other diseases . MAGL regulates the levels of bioactive lipids such as endocannabinoids. Inhibiting MAGL activity has been shown to reduce endocannabinoid levels, which in turn can affect behaviour and other functions of the nervous system. Therefore, MAGL is a prospective target in virtual screening and de novo drug design.
However, in de novo drug design the transfer from virtual molecules to the wet lab has always been a challenge. Most commonly encountered issues usually stem from Synthetic Accessibility (SA) of the generated compounds. One of the tools for comprehensive SA modeling is Computer Aided Synthesis Planning (CASP), facilitated by tools such as AiZynthFinder . However, the computational complexity of this method often limits its practicality in high throughput applications. In the current study, we utilize the DrugEx molecular generator [3, 4] for de novo design of novel MAGL inhibitors in combination with a novel SA score, the LED3Score, as an inherent component of the objective space. Moreover, we show that even with a limited set of building blocks that are readily available in house a synthetic route can be identified for the majority of the generated molecules. Therefore, our results not only demonstrate the feasibility of finding a large and diverse set of novel potentially active ligands for a specific target, but also directly provides routes for their synthesis.
 Zanfirescu, A. et al. Targeting Monoacylglycerol Lipase in Pursuit of Therapies for Neurological
and Neurodegenerative Diseases. Molecules 2021, 26, DOI: 10.3390/molecules26185668.
 Genheden, S. et al. AiZynthFinder: a fast, robust and flexible open-source software for retrosyn-
thetic planning. Journal of Cheminformatics 2020, 12, 70.
 Sicho, M. et al. DrugEx: Deep Learning Models and Tools for Exploration of Drug-like Chemical
Space. 2023, in prepration.
 Liu, X. et al. DrugEx v3: Scaffold-Constrained Drug Design with Graph Transformer-based
Reinforcement Learning. 2021, DOI: 10.26434/chemrxiv-2021-px6kz.