Neurosteroids: Structure-Uptake Relationships and Computational Modeling of Organic Anion Transporting Polypeptides (OATP)1A2

Abstract

In this study, we investigated the delivery of synthetic neurosteroids into MCF-7 human breast adenocarcinoma cells via Organic Anionic Transporting Polypeptides (OATPs) (pH 7.4 and 5.5) to identify the structural components required for OATP-mediated cellular uptake and to get insight into brain drug delivery. Then, we identified structure-uptake relationships using in-house developed OATP1A2 homology model to predict binding sites and modes for the ligands. These binding modes were studied by molecular dynamics simulations to rationalize the experimental results. Our results show that carboxylic acid needs to be at least at 3 carbon-carbon bonds distance from amide bond at the C-3 position of the androstane skeleton and have an amino group to avoid efflux transport. Replacement of hydroxyl group at C-3 with any of the 3, 4, and 5-carbon chained terminal carboxylic groups improved the affinity. We attribute this to polar interactions between carboxylic acid and side-chains of Lys33 and Arg556. The additional amine group showed interactions with Glu172 and Glu200. Based on transporter capacities and efficacies, it could be speculated that the functionalization of acetyl group at the C-17 position of the steroidal skeleton might be explored further to enable OAT1A2-mediated delivery of neurosteroids into the cells and also across the blood-brain barrier.

Keywords: Organic Anion Transporting Polypeptides (OATPs); cellular uptake; docking; molecular modeling; neurosteroid.

Figure 1

Structures of studied compounds.

Figure 2

Cellular uptake of compounds 1–11 into the MCF-7 cells over a concentration range of 5–200 µM and the Eadie-Hofstee plots for transporter-mediated uptake (insets). The data is presented as mean ±SD (n = 3).

Figure 3

Cellular uptake of 25 µM compounds 1–11 into the MCF-7 cells at pH 7.4 (black bars) and at pH 5.5 (white bars). The data is presented as mean ± SD, n = 3 (*** p < 0.001, one-way ANOVA, followed by Tukey’s test).

Figure 4

Cellular uptake of 25 µM compounds 1–11 into the MCF-7 cells at pH 7.4 in the absence (black bars), and presence of efflux inhibitors (50 µM), elacridar (P-gp and BCRP inhibitor, white bars), and MK-571 (unselective MRP inhibitor, dotted bars). The data is presented as mean ± SD, n = 3 (* p < 0.05, ** p < 0.01, *** p < 0.001, one-way ANOVA, followed by Tukey’s test).

Figure 5

Human OATP1A2 homology model and its conserved potential binding pocket. Overview of the OATP1A2 model highlighting the volume of the two predicted binding pockets (Site 1, in pink, and Site 2, in blue) (A) followed by an insight of Site 2 (B). Predictions were performed using SiteMap and pockets were selected according to their druggability score, as described in the methods section. The compounds E3S and 1 were docked and simulated within the Site 2 and representative frames are depicted in (C) E3S and (D) compound 1, where one can observe the main residues with stable interactions in the simulations lying in the transmembrane helices 1 (highlighted in light purple), 4 (in green), and 11 (in yellow). (E) Individual residues and their interaction frequencies (% of simulation time) for each compound are shown as bar plots, where each interaction type is colored differently.

Figure 6

Stable binding mode of compounds 6–8 relies on polar contacts with Lys33, Arg168, and Arg556. Compounds 6 (A), 7 (B), and 8 (C) were docked and simulated within Site 2 of OATP1A2, and representative frames are depicted in (A–C), where one can observe the main residues with stable interactions in the simulations lying in the transmembrane helices 1 (highlighted in light purple), 4 (in green), and 11 (in yellow). (D) Individual residues and hydrogen bond interactions, and water interactions (represented as W), for each compound, are shown as heatmaps, where each interaction is colored according to the frequency.

Figure 7

Stable binding mode of compounds 9–11 relies on polar contacts with Lys33 and Arg556. Compounds 9 (A), 10 (B), and 11 (C) were docked and simulated within the Site 2 of OATP1A2, and representative frames are depicted in (A–C), where one can observe the main residues with stable interactions in the simulations lying in the transmembrane helices 1 (highlighted in light purple), 4 (in green), and 11 (in yellow). (D) Individual residues and hydrogen bond interactions, and water interactions (represented as W), for each compound, are shown as heatmaps, where each interaction is colored according to the frequency.

Figure 8

Summary of structure-uptake relationships of neurosteroids.

Figure 9

Principal Component Analysis (PCA) of OATP1A2-ligand bound complexes: Principal components analysis revealed two states: open (light gray) and closed (dark gray) from PC1. (A) Intracellular view of transmembrane helices displayed closed and open states, and the red arrows show the direction of movement of helices. (B) Lateral view of transmembrane helices showing the two states. (C),(D),(E),(F) Centroid distance between two helices measured along the simulations of each protein-ligand complex. (C) TM11 residues 544–555 to TM4 residues 173–185. (D) TM10 residues 528–538 to TM4 residues 173–185. (E) TM10 residues 528–538 to TM8 residues 373–383. (F) TM10 residues 528–538 to TM5 residues 192–201.