Application of QSPR models to predict CYP450 enzymes activity

Abstract

Cytochromes P450 are a large group of superfamily enzymes mainly concentrated in the liver which play a key role in phase I drug metabolism. The important isforms of these types of enzymes are: CYP3A4, CYP2D6, CYP2C9, CYP2C19 and CYP1A2. Due to their significant contribution in drug metabolism, inhibition or induction of these enzymes can trigger drug-drug interactions, drug toxicity, as well as adverse drug reactions when CYP450-metabolized therapeutic agents are co-administered simultaneously. Therefore, prediction of activity of such enzymes is of great importance in pharmaceutical sciences especially in early stages of drug design and discovery process. Aim: The purpose of this investigation was to develop QSPR models capable of predicting the activity of CYP450 enzymes at the level of mRNA or protein expression. Materials and methods: The compounds showed increasing/decreasing effect on CYP450 mRNA or protein expression were extracted from CTD database. The molecular structures of the extracted compounds were generated and energy minimized by HyperChem program. The molecular descriptors of the chemical compounds were calculated by Dragon software. Then, DTC-QSAR software was utilized for data pretreatment and QSPR model generation followed by model validation process. Results: Analyses of the QSPR models developed for four CYP450 isoenzymes including CYP1A2, CYP3A4, CYP2C9, and CYP2C19, revealed that the generated models were highly predictive with accuracy, precision, and sensitivity of greater than 80%, 80%, and 70%, respectively. Conclusion: The results of the current work can be used for prediction of CYP450 enzymes activity in the field of drug toxicity and drug-drug interactions studies.