Discovery of potential FGFR3 inhibitors via QSAR, pharmacophore modeling, virtual screening and molecular docking studies against bladder cancer

Abstract

Background: Fibroblast growth factor receptor 3 is known as a favorable aim in vast range of cancers, particularly in bladder cancer treatment. Pharmacophore and QSAR modeling approaches are broadly utilized for developing novel compounds for the determination of inhibitory activity versus the biological target. In this study, these methods employed to identify FGFR3 potential inhibitors.

Methods: To find the potential compounds for bladder cancer targeting, ZINC and NCI databases were screened. Pharmacophore and QSAR modeling of FGFR3 inhibitors were utilized for dataset screening. Then, with regard to several factors such as Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) properties and Lipinski's Rule of Five, the recognized compounds were filtered. In further step, utilizing the flexible docking technique, the obtained compounds interactions with FGFR3 were analyzed.

Results: The best five compounds, namely ZINC09045651, ZINC08433190, ZINC00702764, ZINC00710252 and ZINC00668789 were selected for Molecular Dynamics (MD) studies. Off-targeting of screened compounds was also investigated through CDD search and molecular docking. MD outcomes confirmed docking investigations and revealed that five selected compounds could make steady interactions with the FGFR3 and might have effective inhibitory potencies on FGFR3.

Conclusion: These compounds can be considered as candidates for bladder cancer therapy with improved therapeutic properties and less adverse effects.

Keywords: Bladder cancer; Docking; Drug discovery; Molecular dynamics; Pharmacophore; QSAR.

Fig. 1

The schematic diagram of screening steps used in this study

Fig. 2

Final model proposed for FGFR3 protein using RCSB server

Fig. 3

FGFR3 protein domains and structure in cell membrane

Fig. 4

Illustrations for modeling and screening. A Pharmacophore hypothesis model ARRR-1 for FGFR3 inhibitor. B Arrangement of individual features in a fixed distance of pharmacophore hypothesis model ARRR-1 for FGFR3 inhibitor

Fig. 5

Developed QSAR model. R² = 0.8538

Fig. 6

Three-dimensional (3D) structure related to the interaction between FGFR3 protein and approved drugs. A Pemigatinib. B Infigratinib. C Nintedanib. D Ponatinib. Yellow color indicates molecular bonds (hydrogen and hydrophobic), blue color indicates approved drugs and green color was used to improve differentiate of molecular bonds

Fig. 7

Two-dimensional (2D) structure related to the interaction between FGFR3 protein and approved drugs. A Pemigatinib. B Infigratinib. C Nintedanib. D Ponatinib, Red dotted lines indicate hydrophobic bonds and hydrogen bonding shown by green dotted lines

Fig. 8

Chemical structure of candidate ligands

Fig. 9

Three-dimensional (3D) structure related to the interactions of FGFR3 protein and candidate ligands at a distance of 3 Å. A1 ZINC00668789. B1 ZINC00702764. C1 ZINC00710252. D1 ZINC08433190. E1 ZINC09045651. The two-dimensional structures are each candidate shown via A2, B2, C2, D2 and E2, respectively

Fig. 10

Three-dimensional structure (3D) related to the interaction between FGFR3 protein and approved drugs and candidate ligands after conducting molecular dynamics. Yellow color indicates molecular bonds (hydrogen and hydrophobic), blue color indicates approved drugs and green color was used to improve differentiate of molecular bonds

Fig. 11

Two-dimensional (2D) structure related to the interaction between FGFR3 protein and approved drugs after conducting molecular dynamics. Red dotted lines indicate hydrophobic bonds and hydrogen bonding shown by green dotted lines

Fig. 12

Root mean square deviation (RMSD) of free FGFR3 and FGFR3-ligand complexes during the 100 ns simulation

Fig. 13

Root mean square fluctuation (RMSF) value per residue of free FGFR3 and FGFR3-ligands complexes during the 100 ns simulation

Fig. 14

Radius of gyration (Rg) value of free FGFR3 and FGFR3-ligands complexes during the 100 ns simulation

Fig. 15

Numbers of intermolecular hydrogen bonds in FGFR3-ligands complexes during the 100 ns simulation

Fig. 16.

2D projections of the nine complexes on eigenvector 1 and eigenvector 2