Repositioning of Etravirine as a Potential CK1ε Inhibitor by Virtual Screening

Abstract

CK1ε is a key regulator of WNT/β-catenin and other pathways that are linked to tumor progression; thus, CK1ε is considered a target for the development of antineoplastic therapies. In this study, we performed a virtual screening to search for potential CK1ε inhibitors. First, we characterized the dynamic noncovalent interactions profiles for a set of reported CK1ε inhibitors to generate a pharmacophore model, which was used to identify new potential inhibitors among FDA-approved drugs. We found that etravirine and abacavir, two drugs that are approved for HIV infections, can be repurposed as CK1ε inhibitors. The interaction of these drugs with CK1ε was further examined by molecular docking and molecular dynamics. Etravirine and abacavir formed stable complexes with the target, emulating the binding behavior of known inhibitors. However, only etravirine showed high theoretical binding affinity to CK1ε. Our findings provide a new pharmacophore for targeting CK1ε and implicate etravirine as a CK1ε inhibitor and antineoplastic agent.

Keywords: CK1ε; abacavir; cancer; drug repurposing; etravirine; pharmacophore model.

Figure 3

Molecular dynamics analysis identified key inhibitor-CK1ε interactions. (A) Protein backbone and ligand RMSDs for apoCK1ε and CK1ε-inhibitor complexes. (B) Protein backbone RMSF for all systems. Secondary structure of CK1ε is shown at the bottom with regions comprising the catalytic domain in red squares. (C) Heatmap of the non-covalent interaction profile occurrence. For clarity, only residues with occurrence higher than 20% are shown.

Figure 4

Pharmacophore modeling. (A) CK1ε catalytic site representation with selected amino acids showed on licorice model. Residues forming hydrophobic interactions, hydrogen bonds, or stacking interactions are colored green, yellow and white, or purple, respectively. (B) The generated pharmacophore model included eight elements. Hydrogen bond donors (HBD) are represented as sphere grid colored white, hydrogen bond acceptor (HBA) on yellow, aromatic (Aro) on purple, and hydrophobic (Hyd) on green.

Figure 5

Docking poses of etravirine and abacavir. Overlap of pharmacophore features with the calculated binding modes of etravirine (magenta) and abacavir (cyan). Hydrogen bonds are shown as blue lines. Note that the pharmacophoric submodel used for abacavir lacks the dual element Aro1/Hyd1 (see text for details).

Figure 6

MD analyses for CK1ε-etravirine and CK1ε-abacavir complexes. (A) Protein backbone (black) and ligands (colored) RMSDs. (B) Heatmap of the non-covalent interactions occurrence for selected residues. (C) Protein backbone RMSF. The apo enzyme and the system CK1ε-IN1 are shown for comparison. Secondary structure of CK1ε is shown at the bottom with the catalytic domain in red squares.

Figure 7

Binding energy-analysis. (A) Energy decomposition from systems with etravirine or abacavir were calculated from MD simulations. The system with IN1 is included for comparison. (B) Per-residue energy decomposition. Highlighted residues are part of the catalytic domain.

Figure 1

Docking protocol reproduces the active conformer of IN1. (A) Hierarchical clustering of 96 poses obtained by molecular docking. Inset shows the largest subcluster. Poses are labeled using letters that indicate the combination of search algorithm/scoring function employed (see “Methods”) and a number indicating the ranked position. The crystal pose of IN1 (REF) and the best scored pose calculated using PLANTS scoring function and Optimizer search algorithm (pO-1) are highlighted in gray and pink, respectively. Color scale shows the RMSD between poses. (B) Superimposition of co-crystallized (gray) and pO-1 docked (pink) poses.

Figure 2

Binding modes calculated by molecular docking for the five inhibitors analyzed. Cartoon representation of CK1ε with residues Glu52, Glu83, and Leu85, and Ser88 on licorice model. Hydrogen bonds are shown as blue lines.