Purpose
To identify the molecular features of well characterized anti-VEGFA (bevacizumab, ranibizumab and aflibercept) as a reference guide to design new effective anti-angiogenic molecules.
Methods
We carried out the molecular modeling of aflibercept binding domain (R1D2_R2D3), characterized by the fusion of domain D2 of VEGFR1 and domain D3 of VEGFR2, by means of SwissModel web server using as template the crystal structure of VEGFR2 (PDB: 2X1W). All-atom molecular dynamics of R1D2_R2D3 in explicit water was carried out for 10 ns. Three protein-protein docking software were used (ZDock, GRAMMX and Pydock) in order to carry out docking of VEGFA with bevacizumab, ranibizumab and modeled aflibercept binding domains.
Results
The protein-protein docking gave consistent results for ranibizumab and bevacizumab in complex with VEGF-A in comparison to the correspondent crystal structures, respectively 1CZ8 and 1BJ1. Since the docking protocol was successful for ranibizumab and bevacizumab, modeling of VEGF-A interaction with modeled R1D2_R2D3 was carried out. Interactions of R1D2_R2D3 in complex with VEGF-A resembled the binding of VEGF-C with VEGFR2 (Fig. 1). Pydock had a prediction rate comparable to the other software hereby used, in addition pydock output was informative of binding energy terms: electrostatic (ELEC), desolvation (DES), Van Der Waals (VDW), total docking score (TOT). Interestingly the binding of VEGF-A to ranibizumab and bevacizumab was characterized by different stabilizing energy terms, in comparison to aflibercept binding domain (table1).
Conclusions
Molecular modeling approaches are feasible to evaluate binding features of VEGF in complex with anti-angiogenic factors, and might help to design new anti-VEGF drugs with better inhibitory effects.
Keywords: 473 computational modeling •
543 growth factors/growth factor receptors