Development of Classification Models for Identifying “True” P-glycoprotein (P-gp) Inhibitors Through Inhibition, ATPase Activation and Monolayer Efflux Assays

• Main Authors: Anna Maria Bianucci, Simona Rapposelli, Alessio Coi, Marcello Imbriani
• Format: Article
• Language: English
• Published: MDPI AG 2012-06-01
• Series: International Journal of Molecular Sciences
• Subjects: P-glicoprotein; decision trees; classification model; consensus model; P-gp inhibitors; MDR1 ligands
• Online Access: http://www.mdpi.com/1422-0067/13/6/6924

P-glycoprotein (P-gp) is an efflux pump involved in the protection of tissues of several organs by influencing xenobiotic disposition. P-gp plays a key role in multidrug resistance and in the progression of many neurodegenerative diseases. The development of new and more effective therapeutics targeting P-gp thus represents an intriguing challenge in drug discovery. P-gp inhibition may be considered as a valid approach to improve drug bioavailability as well as to overcome drug resistance to many kinds of tumours characterized by the over-expression of this protein. This study aims to develop classification models from a unique dataset of 59 compounds for which there were homogeneous experimental data on P-gp inhibition, ATPase activation and monolayer efflux. For each experiment, the dataset was split into a training and a test set comprising 39 and 20 molecules, respectively. Rational splitting was accomplished using a sphere-exclusion type algorithm. After a two-step (internal/external) validation, the best-performing classification models were used in a consensus predicting task for the identification of compounds named as “true” P-gp inhibitors, <em>i.e.</em>, molecules able to inhibit P-gp without being effluxed by P-gp itself and simultaneously unable to activate the ATPase function.