Avaliação computacional de um conjunto de fragmentos de anticorpos anti-DLL1 como base para proposição de um novo biofármaco

Abstract

Cancer is the name that characterizes a set of disorders marked by the presence of cells that do not respond to normal cell division controls. It is estimated that in 2018, such diseases caused 9,6 million deaths worldwide. Conventional therapies are not targeted and commonly are followed by the manifestation of various adverse effects. In this context, new treatments, such as immunotherapy using immune checkpoints inhibitors, have gained prominence. The humanized monoclonal antibody pidilizumab was initially selected for its ability to stimulate lymphocyte proliferation and its anti-tumor activity. A large number of promising clinical studies indicated that such antibodyhad a typical pharmacological profile of a PD-1 checkpoint protein antagonist. However, it has recently been elucidated that its genuine molecular target is the DLL1 protein, a Notch signaling pathway ligand whose blockade has been associated with inhibition of cell proliferation and induction of tumor cell apoptosis to glioma, demonstrating its potential as a therapeutic target for cancer treatment. Given the good results achieved with the use of pidilizumab in several clinical trials, this study aimed to indicate changes to scFv antibody fragments produced from pidilizumab in order to optimize interactions with their possible molecular targets, PD-1 and DLL1 proteins.For this, comparative modeling of scFvs light and heavy chain sequences obtained from Integrity databases was performed. The addition of three flexible linker peptides of different length between chains resulted in the production of three scFvs. These molecules were docked to PD-1 and DLL1 proteins through a previously validated methodology and the resulting complexes were submitted to molecular dynamics simulations. The simulations showed the complete absence of interactions between scFvs and PD-1. Furthermore, by performing MM/PBSA calculations, it was possible to elucidate that the most energetically favorable mode of binding between scFv and DLL1 is found through positioning the antibody fragment in the antigen C2 domain, and that the major free energy components contributing to the establishment of the interaction are the non-polar, Coulomb and van der Waals terms. Finally, the main interface residues involved in the interaction were identified and energetically favorable mutations were proposed on the scFv sequence. The application and validation of such mutations by molecular dynamics resulted in the proposition of a new optimized biopharmaceutical.