In Silico Prediction of Inhibition of Promiscuous Breast Cancer Resistance Protein (BCRP/ABCG2)

Background: Breast cancer resistant protein has an essential role in active transport of endogenous substances and xenobiotics across extracellular and intracellular membranes along with P-glycoprotein. It also plays a major role in multiple drug resistance and permeation of blood-brain barrier. Therefore, it is of great importance to derive theoretical models to predict the inhibition of both transporters in the process of drug discovery and development. Hitherto, very limited BCRP inhibition predictive models have been proposed as compared with its P-gp counterpart. Methodology/Principal Findings: An in silico BCRP inhibition model was developed in this study using the pharmacophore ensemble/support vector machine scheme to take into account the promiscuous nature of BCRP. The predictions by the PhE/SVM model were found to be in good agreement with the observed values for those molecules in the training set (n = 22, r2 = 0.82, = 0.73, RMSE = 0.40, s = 0.24), test set (n = 97, q2 = 0.75–0.89, RMSE = 0.31, s = 0.21), and outlier set (n = 16, q2 = 0.72–0.91, RMSE = 0.29, s = 0.17). When subjected to a variety of statistical validations, the developed PhE/SVM model consistently met the most stringent criteria. A mock test by HIV protease inhibitors also asserted its predictivity. Conclusions/Significance: It was found that this accurate, fast, and robust PhE/SVM model can be employed to predict the BCRP inhibition of structurally diverse molecules that otherwise cannot be carried out by any other methods in a high-throughput fashion to design therapeutic agents with insignificant drug toxicity and unfavorable drug–drug interactions mediated by BCRP to enhance clinical efficacy and/or circumvent drug resistance.