Integrating qHTS and QSAR Models to Identify Safe GPCR-Targeted Compounds: A Focus on hERG-Dependent Cardiotoxicity

Area of Science:

• Pharmacology and Toxicology
• Computational Chemistry
• Drug Discovery

Background:

• G-protein-coupled receptors (GPCRs) are crucial drug targets for numerous diseases.
• hERG channel inhibition by drugs can cause life-threatening cardiac arrhythmias.
• Assessing GPCR-hERG interactions is vital for safe drug development.

Purpose of the Study:

• To identify novel GPCR modulators with reduced hERG liability.
• To develop and validate machine learning models for predicting GPCR activity and hERG safety.
• To provide efficient strategies for safer drug lead discovery.

Main Methods:

• Quantitative high-throughput screening (qHTS) of the Tox21 10K compound library for GPCR activity.
• Development and application of machine learning (ML)-based quantitative structure-activity relationship (QSAR) models.
• Virtual screening of ~360K compounds and experimental validation of top predictions.

Main Results:

• Identified selective GPCR agonists and inhibitors with minimal hERG liability.
• ML-QSAR models accurately predicted GPCR-targeting compounds with reduced hERG risk.
• Validated models using LOPAC and identified novel compounds with desired profiles.

Conclusions:

• Machine learning-based QSAR models offer an efficient approach for discovering GPCR modulators.
• This strategy effectively minimizes cardiac risks associated with hERG channel inhibition.
• The findings support the development of safer therapeutics targeting GPCRs.