Discovering Potent and Diverse Agonists for the β₂‑Adrenergic Receptor via Machine Learning

Area of Science:

• Drug discovery and development
• Computational chemistry and cheminformatics
• Pharmacology and medicinal chemistry

Background:

• Diverse compound screening is crucial for novel drug discovery, offering advantages like patent navigation and toxicity avoidance.
• G protein-coupled receptors (GPCRs) are key drug targets, with the beta-2 adrenergic receptor (β2AR) being a significant target for respiratory disease treatments.
• Existing β2AR agonists require optimization for enhanced safety, efficacy, and selectivity.

Purpose of the Study:

• To develop and apply machine learning (ML) and computational methods for efficient screening of a large compound library.
• To identify novel, potent, and structurally diverse agonists targeting the β2AR.
• To address the need for improved therapeutic agents for respiratory conditions.

Main Methods:

• Utilized machine learning (ML) integrated with computational approaches for high-throughput screening.
• Screened a comprehensive library of 19 million chemical compounds.
• Validated identified agonists through cellular functional assays.

Main Results:

• Identified several highly potent β2AR agonists with EC50 values ranging from 0.2 to 20 nM.
• Discovered agonists possessing novel chemical structures, distinct from previously reported molecules.
• Confirmed the efficacy of identified compounds through experimental validation.

Conclusions:

• Machine learning and computational screening offer a powerful strategy for discovering novel GPCR drug candidates.
• The identified potent and structurally diverse agonists represent promising leads for developing next-generation β2AR therapeutics.
• This approach accelerates the design of new drug candidates for GPCR targets, addressing unmet medical needs.