In Silico Molecular Modeling of Four New Afatinib Derived Molecules Targeting the Inhibition of the Mutated Form of BCR-ABL T315I

Abstract

Four afatinib derivatives were designed and modeled. These derivatives were compared to the known tyrosine-kinase inhibitors in treating Chronic Myeloid Leukemia, i.e., imatinib and ponatinib. The molecules were evaluated through computational methods, including docking studies, the non-covalent interaction index, Electron Localization and Fukui Functions, in silico ADMET analysis, QTAIM, and Heat Map analysis. The AFA(IV) candidate significantly increases the score value compared to afatinib. Furthermore, AFA(IV) was shown to be relatively similar to the ponatinib profile when evaluating a range of molecular descriptors. The addition of a methylpiperazine ring seems to be well distributed in the structure of afatinib when targeting the BCR-ABL enzyme, providing an important hydrogen bond interaction with the Asp381 residue of the DFG-switch of BCR-ABL active site residue and the AFA(IV) new chemical entities. Finally, in silico toxicity predictions show a favorable index, with some molecules presenting the loss of the irritant properties associated with afatinib in theoretical predictions.

Keywords: BCR-ABL; CML; NCI; TKI; molecular modeling.

Figure 1

The BCR-ABL enzyme in its (a) DFG-in (PDB: 2GQG) and (b) DFG-out (PDB: 3QRJ) conformations. Color code: P-loop orange, c-alpha helix cyan, hinge red, catalytic loop green, and activation loop yellow. The Phe382 residue is shown in the CPK representation.

Figure 2

Representation of the 2D molecular structures of afatinib, imatinib, and ponatinib, and the afatinib derivatives, AFA(I), AFA(II), AFA(III), and AFA(IV).

Figure 3

Isosurfaces of the frontier orbitals of (a) afatinib, (b) imatinib, (c) ponatinib, (d) AFA(I), (e) AFA(II), (f) AFA(III), and (g) AFA(IV) at B3LYP/6-311+G(d,p) level. The HOMO isosurfaces are depicted at the top, and the LUMO isosurfaces at the bottom. The images were rendered with an isovalue of 0.05 e-.Å⁻³.

Figure 4

Fukui functions (f⁻ at the left, f⁺ at the right) of (a) afatinib, (b) imatinib, (c) ponatinib, (d) AFA(I), (e) AFA(II), (f) AFA(III), and (g) AFA(IV) at B3LYP/6-311+G(d,p) level. Positive regions are represented in green, while negative regions are presented in cyan. Rendering was obtained with an isovalue of 0.005.

Figure 5

Plotting of the 2D-Electron Localization Function around the central amide moiety at B3LYP/6-311+G(d,p) level. The scaffold represents the afatinib derivatives. (a) afatinib, (b) imatinib, (c) ponatinib, (d) AFA(I), (e) AFA(II), (f) AFA(III), and (g) AFA(IV).

Figure 6

2D representation of the Vina docking conformations in the 1OPJ protein showing the distances (Å) of calculated structures. (a) AFA(I), (b) AFA(II), (c) AFA(III), and (d) AFA(IV).

Figure 7

2D representation of the Vina docking conformations in the 3QRJ protein. Distances are measured in Å. (a) AFA(I), (b) AFA(II), (c) AFA(III), and (d) AFA(IV).

Figure 8

NCI analysis based on promolecular densities of the docked conformations of the ligands in the 1OPJ protein. (a) afatinib, (b) imatinib, (c) ponatinib, (d) AFA(I), (e) AFA(II), (f) AFA(III), and (g) AFA(IV).

Figure 9

NCI analysis based on promolecular densities of the docked conformations of the ligands in the 3QRJ protein. (a) afatinib, (b) imatinib, (c) ponatinib, (d) AFA(I), (e) AFA(II), (f) AFA(III), and (g) AFA(IV).

Figure 10

Estimated distances (Å) between key residues and each ligand in the chosen docking conformations for the wild-type enzyme (PDB 1OPJ). The volume of each bar is normalized.

Figure 11

Estimated distances (Å) between key residues and each ligand in the chosen docking conformations for the mutated enzyme (PDB 3QRJ). The volume of each bar is normalized.

Figure 12

Heat maps showing the overall similarity between the derived molecules and the established inhibitors: (a) General similarity considering molecular descriptors; (b) General similarity considering molecular descriptors associated with the binding energy performance regarding the wild-type enzyme; (c) General similarity considering molecular descriptors associated with the binding energy performance regarding the mutated enzyme.