Artificial Intelligent Deep Learning Molecular Generative Modeling of Scaffold-Focused and Cannabinoid CB2 Target-Specific Small-Molecule Sublibraries

Abstract

Design and generation of high-quality target- and scaffold-specific small molecules is an important strategy for the discovery of unique and potent bioactive drug molecules. To achieve this goal, authors have developed the deep-learning molecule generation model (DeepMGM) and applied it for the de novo molecular generation of scaffold-focused small-molecule libraries. In this study, a recurrent neural network (RNN) using long short-term memory (LSTM) units was trained with drug-like molecules to result in a general model (g-DeepMGM). Sampling practices on indole and purine scaffolds illustrate the feasibility of creating scaffold-focused chemical libraries based on machine intelligence. Subsequently, a target-specific model (t-DeepMGM) for cannabinoid receptor 2 (CB2) was constructed following the transfer learning process of known CB2 ligands. Sampling outcomes can present similar properties to the reported active molecules. Finally, a discriminator was trained and attached to the DeepMGM to result in an in silico molecular design-test circle. Medicinal chemistry synthesis and biological validation was performed to further investigate the generation outcome, showing that XIE9137 was identified as a potential allosteric modulator of CB2. This study demonstrates how recent progress in deep learning intelligence can benefit drug discovery, especially in de novo molecular design and chemical library generation.

The application of artificial intelligence to accelerate G protein‐coupled receptor drug discovery

“…After biological validation of the generated molecules, a potential allosteric modulator, XIE9137, of the CB 2 receptor was identified (Bian & Xie, 2022).…”

“…The potential for the generation of molecules conditional on the protein pockets has also been widely explored (Li et al, 2021; Luo et al, 2021; Peng et al, 2022). The potential of the deep generative recurrent neural network model was showcased by successfully building scaffold‐focused chemical libraries for the CB 2 receptor (Bian & Xie, 2022). After biological validation of the generated molecules, a potential allosteric modulator, XIE9137, of the CB 2 receptor was identified (Bian & Xie, 2022).…”

A Structure-Based Allosteric Modulator Design Paradigm

“…Such models strive to extract representative features automatically and learn the distribution of the ligand set in the training dataset, allowing the generation of ligands that bear a resemblance to those in the training set yet also display a degree of novelty. To this end, Bian and Xie [ 84 ] trained one of the generative models, long short-term memory recurrent neural networks, on half a million drug-like molecules. Next, they constructed a CB2 target-specific model, named t-DeepMGM, by fine-tuning the generative model on a dataset of reported CB2 modulators.…”

Cannabinoid type-2 receptors: An emerging target for regulating schizophrenia-relevant brain circuits

“…Collectively, these pharmacological and genetic tools have facilitated detailed studies examining how the endocannabinoid system functions under physiological conditions. Moreover, this toolbox is constantly expanding and is aided by recent insights into the structure of the CB2 receptor ( Hua et al, 2020 ; Xing et al, 2020 ) that have allowed in silico approaches to discover novel compounds ( Bian and Xie, 2022 ).…”