An in silico approach to peptide-based dual-receptor targeting for IL13RA2 and VEGFR-2 extracellular domain

Vascular endothelial growth factor receptor-2 (VEGFR-2) and interleukin 13 receptor subunit-2 (IL13Rα2) are major drug targets due to their overexpression in several cancers that lead to tumor cell proliferation, metastasis, and poor prognosis. In this work, we have designed novel peptides that have the potential for dual targeting of both VEGFR-2 and IL13Rα2. Using the tumor homing peptide ACGEMGWVRCGGGS (pep1IL) as a starting point, we designed seven new peptides through single point variations within the sequences. The binding affinities and stabilities of the peptides were evaluated through molecular docking and molecular dynamics (MD) simulations. Our results indicated that among the peptide variants, ACGHMGWVRCGGGS and ACGEMGWVRCGGGT formed the most stable complexes with VEGFR-2 and IL13Rα2 respectively, while ACGEMGWVSCGGGS displayed strong binding with both receptors. In general, binding occurred with residues encompassing the D2 and D3 domains of VEGFR-2 and D1, D2, and D3 domains of IL13Rα2. Of particular note is the importance of MET143 and THR229 residues of IL13Rα2 that were critical in binding of most of the peptides. For VEGFR-2, several of the designed peptides formed hydrogen bonds with residues including TYR 165, SER 193, TYR 194, and ASN253 which have been implicated in mediating binding with VEGF-A/C/D. Thus, some of the peptides may act as competitive inhibitors, that may potentially play a role in impeding angiogenic signaling. We also designed the corresponding disulfide-bridged peptides that were docked with both receptors. MD simulations were run for two of the most optimal of those peptides and the disulfide-bridged serine variant sequence showed stable binding with both receptors. MM-GBSA results demonstrated that Van der Waals and electrostatic interactions played a key role in binding. Additionally, ACGEMGWVRCGGGT showed specificity, toward IL13Rα2 receptor and stronger binding than the original tumor homing peptide. Likewise, ACGDMGWVRCGGGS and ACGHMGWVRCGGGS showed significantly stronger binding with VEGFR-2 compared to IL13Rα2. Thus, this study shows the potential of our approach to design selective peptide variants that can be utilized for tumor targeting. Moreover, the designed peptides may be further explored in conjugation with anti-cancer drugs for future synthesis and therapeutic applications.

To determine the complete 3D structure of the extracellular domains of VEGFR-2 and IL13RA2, the AlphaFold 3 web server was used. Unbound (apo) receptor simulations were run using DESMOND to ensure the stability of the receptors. AntiCP web server was used to predict the anticancer potential of the peptides, while ADMETlab3.0 web server was used to determine the drug likeliness of the peptides. PEP-FOLD3 and MolProbity web servers were utilized to determine structural information about the peptides. Monte Carlo simulations were done using the MCPep server to predict membrane interactions of the peptides with model cell membranes. The free energy difference between the peptide sequences in the presence of the membrane and in aqueous phase (ΔG_tot) was determined as a sum of differences of the following parameters: electrostatic (ΔG_elc) and non-polar (ΔG_np) contributions to the solvation free energy (ΔG_sol = ΔG_elc + ΔG_np), peptide immobilization (ΔG_imm), lipid structural changes (ΔG_lip), peptide conformation changes (ΔG_con), and membrane deformation (ΔG_def). Atomistic molecular dynamics simulations were carried out for the peptide-receptor complexes using DESMOND, from Schrödinger Suite. Each of the simulations was carried out for 400 ns in an NPT ensemble at 310 K employing the OPLS4 force field. Binding energy was evaluated by molecular mechanics generalized born surface area (MM/GBSA) studies using the prime module of Schrodinger. PCA analysis was also conducted to determine the most prominent components that contribute to the dynamic motion of the complexes and the unbound receptor using Schrodinger. Molecular docking studies were carried out using Autodock Vina 1.1.2 as well as DockThor and HPEPDOCK webservers.