Probing Small-Molecule Binding to the Liver-X Receptor: A Mixed-Model QSAR Study

Area of Science:

• Toxicology
• Computational Chemistry
• Drug Discovery

Background:

• Endocrine-disrupting chemicals pose risks to human health.
• Accurate prediction of chemical-protein interactions is crucial for safety assessment.
• VirtualToxLab offers automated screening for toxicological potential.

Purpose of the Study:

• To integrate and validate a Liver X Receptor (LXR) model within the VirtualToxLab platform.
• To screen natural compounds for their potential to bind to LXRs.
• To establish the predictive capability of the LXR model for compound screening.

Main Methods:

• Utilized molecular docking to predict ligand binding to LXR protein structures.
• Employed multidimensional Quantitative Structure-Activity Relationship (mQSAR) for binding affinity quantification.
• Validated the model using an external set of 17 oxysterols and performed scramble tests.

Main Results:

• Successfully screened 161 natural compounds for LXR binding.
• The LXR model demonstrated predictive accuracy when compared with available experimental data.
• Model robustness was confirmed through consensus scoring and elimination of chance correlations.

Conclusions:

• The validated LXR model in VirtualToxLab is effective for predicting compound binding affinity.
• This automated approach facilitates efficient screening of drugs, chemicals, and natural products for endocrine-disrupting potential.
• The model serves as a valuable tool for early-stage safety assessment in drug discovery and chemical evaluation.