High-Throughput Chemical Screening and Structure-Based Models to Predict hERG Inhibition

Area of Science:

• Toxicology
• Cardiovascular Pharmacology
• Computational Chemistry

Background:

• Inhibition of the human ether-a-go-go-related gene (hERG) potassium channel prolongs the QT interval, potentially causing severe cardiotoxicity.
• hERG channel inhibition is a major reason for drug attrition during development.
• Environmental chemicals and toxicants may also inhibit hERG, contributing to cardiovascular disease.

Purpose of the Study:

• To screen the Tox21 library for chemicals that inhibit hERG channel activity.
• To identify chemical features associated with hERG inhibition.
• To develop quantitative structure-activity relationship (QSAR) models for predicting hERG liabilities.

Main Methods:

• Quantitative high-throughput screening (qHTS) of ~7871 unique chemicals using a U2OS cell line thallium flux assay.
• Chemical structure-based clustering and chemotype enrichment analysis.
• Machine learning approaches, including random forest, to build QSAR prediction models.

Main Results:

• A robust dataset of chemicals perturbing hERG activity was generated.
• Molecular features responsible for hERG activity were identified.
• A random forest model achieved ~92.6% balanced accuracy in predicting hERG liabilities.

Conclusions:

• The study successfully identified hERG inhibitors and associated chemical features.
• Developed QSAR models provide valuable in vitro and in silico tools for drug development and environmental chemical screening.
• Open-access data and scripts facilitate translational toxicology research.