Premium
Optimizing QSAR Models for Predicting Ligand Binding to the Drug‐Metabolizing Cytochrome P450 Isoenzyme CYP2D6
Author(s) -
Saraceno Marilena,
Massarelli Ilaria,
Imbriani Marcello,
James Thomas L.,
Bianucci Anna M.
Publication year - 2011
Publication title -
chemical biology and drug design
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.59
H-Index - 77
eISSN - 1747-0285
pISSN - 1747-0277
DOI - 10.1111/j.1747-0285.2011.01137.x
Subject(s) - quantitative structure–activity relationship , cyp2d6 , cytochrome p450 , applicability domain , isozyme , molecular descriptor , computational biology , chemistry , biological system , computer science , stereochemistry , biology , biochemistry , enzyme
The cytochrome P450 isozyme CYP2D6 binds a large variety of drugs, oxidizing many of them, and plays a crucial role in establishing in vivo drug levels, especially in multidrug regimens. The current study aimed to develop reliable predictive models for estimating the CYP2D6 inhibition properties of drug candidates. Quantitative structure–activity relationship (QSAR) studies utilizing 51 known CYP2D6 inhibitors were carried out. Performance achieved using models based on two‐dimensional (2D) molecular descriptors was compared with performance using models entailing additional molecular descriptors that depend upon the three‐dimensional (3D) structure of ligands. To properly compute the descriptors, all the 3D inhibitor structures were optimized such that induced‐fit binding of the ligand to the active site was accommodated. CODESSA software was used to obtain equations for correlating the structural features of the ligands to their pharmacological effects on CYP2D6 (inhibition). The predictive power of all the QSAR models obtained was estimated by applying rigorous statistical criteria. To assess the robustness and predictability of the models, predictions were carried out on an additional set of known molecules ( prediction set ). The results showed that only models incorporating 3D descriptors in addition to 2D molecular descriptors possessed the requisite high predictive power for CYP2D6 inhibition.