In Silico Prediction of Drug-Drug Interactions Mediated by Cytochrome P450 Isoforms

Area of Science:

• Pharmacology
• Computational Chemistry
• Drug Discovery

Background:

• Drug-drug interactions (DDIs) can lead to toxicity, reduced efficacy, and adverse reactions.
• Assessing DDIs mediated by cytochrome P450 (CYP) enzymes is crucial in drug development.
• Seven key CYP isoforms (CYP1A2, CYP2B6, CYP2C19, CYP2C8, CYP2C9, CYP2D6, CYP3A4) are central to drug metabolism.

Purpose of the Study:

• To create a computational model for predicting DDIs mediated by major CYP enzymes.
• To enable prediction of DDIs for both known drugs and new chemical entities.
• To provide a publicly accessible web resource for DDI predictions.

Main Methods:

• Development of structure-activity relationship (SAR) models.
• Utilized Prediction of Activity Spectra for Substances (PASS) software.
• Employed Pairs of Substances Multilevel Neighborhoods of Atoms (PoSMNA) descriptors.
• Trained models on approximately 2500 DDI records mediated by target CYP isoforms.

Main Results:

• Achieved an average prediction accuracy of 0.92 for DDIs mediated by various CYP isoforms.
• The developed SAR models can predict DDIs for known and novel substances.
• The models demonstrated high predictive performance using leave-one-out cross-validation (LOO CV).

Conclusions:

• The study successfully created a predictive computer model for CYP-mediated DDIs.
• The model offers a valuable tool for drug discovery and development, aiding in the early identification of potential DDIs.
• The publicly available web resource facilitates DDI assessment for a wide range of compounds.