Combined Ligand- and Structure-Based Strategies to Design Potential Inhibitors of Multidrug Resistance (MDR) Protein ABCB1

Abstract

Intrinsic or acquired resistance to chemotherapy, known as multidrug resistance (MDR) is one of the major obstacles in the treatment of cancer. Several MDR-causing factors have been elucidated, however, accelerated drug efflux mediated by the over-expression of the ATP-binding cassette (ABC) transporters are commonly recognized as clinically crucial. Among the 49 distinct human ABC transporters, ABCB1 has been recognized as the leading cause of MDR which prevents the intracellular accumulation of drugs, including front-line chemotherapeutic agents such as doxorubicin, paclitaxel, and vincristine. Over the past few decades, three generations of ABCB1-inhibitors have been developed to overcome the innate or up-regulated ABCB1 expression. Yet, to date, no single drug has been approved due to a combination of poor pharmacokinetic properties, lack of selectivity, inter-patient variability, and poor toxicity profiles. However, the advent of structural data for human ABCB1 in the past few years allowed to design optimized ABCB1 inhibitors which are more potent, less toxic, and clinically efficacious. The study presents a virtual screening pipeline using a pharmacophore of the structurally diverse and highly selective dataset of ABCB1 inhibitors. Finally, a selected model delineating three hydrophobic (Hyd), one hydrogen bond acceptor (HBA), and one aromatic (Aro) features at a certain mutual distance, was used for the screening of the Chembridge library. The presented virtual screening pipeline along with various toxicity profiling filters were used for the identification and short-listing of the potential hits against ABCB1. Subsequently, the selected six hit compounds were evaluated experimentally for efficacy and potency using fluorescent drug transport assay in Flp-In 293 and the Flp-In-ABCB1 cell lines. It was observed that four (‘A’, ‘D’, ‘E’, and ‘F) out of six compounds showed half-maximal inhibitory concentrations (IC50) in the low nanomolar range (from 1.35 to 26.4 nM). Moreover, the optimized concentration of these leads was tested against taxol and it was observed that the two most promising compounds ‘A’, and ‘D’ were also able to re-sensitize ABCB1-expressing cells to taxol at 100 nM as well.

XVIII

Further, to probe the binding hypothesis and physiochemical profiles, the identified leads (‘A’, ‘D’, ‘E’, and ‘F’) were docked into a model structure of the human ABCB1 (6QEX) followed by molecular dynamics (MD) simulations. All four lead compounds mostly showed hydrophobic interactions with Phe-303, Tyr-307, Phe-343, Met-986, and Gln-990, π-π interactions with Trp-232, and hydrogen bonding with Gln-990. The binding hypotheses were also compared to the wider curated dataset of highly selective and potent compounds, previously reported in the literature. Moreover, the binding stability of all four lead compounds, within the binding pocket of ABCB1 was indicated by very low positional changes of the Root Mean Square Deviation (RMSD) values. Similarly, in the case of RMSF, small peaks were observed in the transmembrane regions however, relatively large fluctuations were observed in the loop regions of ABCB1 due to their flexible nature. Hence, Protein-ligand interactions, molecular docking, MD simulations, LipE profiling, and statistical along with the pharmacokinetic analyses, were indicative of potent and selective inhibition of ABCB1. Remarkably, four out of the six new compounds, identified using our pharmacophore model were found to inhibit ABCB1 with high potency, providing proof-of-principle of our computational approach and the utility of the cryo-EM structure data for this purpose. Moreover, these lead compounds also showed stable interactions with minimum fluctuations in RMSD and RMSF during MD simulations and ideal pharmacokinetic properties, LipE and clogP values, which will help to realize their potential as candidate inhibitors of ABCB1.