Titolo della tesi: Computational Methods For the Development of Biotherapeutic Products
The development process of biotherapeutic products faces remarkable challenges in identifying novel, effective and safe candidates, with chemical stability emerging as a critical obstacle that hinders the discovery process. This thesis provides a comprehensive evaluation of the thermodynamics and kinetics of two of the most common spontaneous Post-Translational Modifications (PTMs), namely asparagine deamidation and methionine oxidation, whose occurrence can compromise the efficacy and immunogenicity of the biotherapeutics under development. By employing a rigorous theoretical-computational framework, that integrates Quantum Mechanics, hybrid Quantum Mechanics/Molecular Mechanics and Molecular Dynamics approaches, we sought to decipher the molecular mechanism underlying these PTMs reactions.
A bottom-up computational strategy was adopted, which comprises investigations on simple molecular models, extension to real systems and quantitative validation against tailored experiments and literature data. This strategy has yielded novel insights into the behavior of such reactions in proteins, unveiling the critical role of hydrogen peroxide diffusion in modulating methionine oxidation. Conversely, features such as amide acidity, conformational degrees of freedom, and environmental electrostatics were identified as key determinants of asparagine reactivity towards deamidation. These findings enhanced our understanding of these chemical reactions and facilitated more accurate predictions in several proteins, paving the way for future studies on a more extended scale.
Moreover, this thesis demonstrates the efficacy of an integrative modelling approach to rationalize mutagenesis effects, which is essential to capture the inherent flexibility of monoclonal antibodies (mAb), the dominant class of biotherapeutics. Indeed, by coupling computational and experimental assays, we elucidated the structural-dynamics features behind the novel identified framework positions to tune mAb-antigen binding, which could provide valuable insights in the sequence optimization stage of development.