Formulation and Quality – Biomolecular
2019 PharmSci 360
Monoclonal antibody stability can be modified with the appropriate selection of excipient. We have used molecular dynamics simulations to study how various excipients (including salts, carbohydrates, and amino acids) affect aggregation and viscosity. By connecting molecular-level interactions of antibody residues with these excipients to experimental data, we have identified trends in how excipients interact with the antibody. For example, excipient carbohydrate molecules tend to interact with hydrophobic patches on the antibody surface, leading to a reduction in aggregation. Additionally, we have used machine learning algorithms to develop models for these antibody-excipient interactions, allowing these interaction patterns to be modeled very quickly, without having to perform lengthy molecular dynamics simulations. These algorithms have identified the relative importance of different features, such as surface area, charge, and hydrophobicity, in determining the local interactions with each excipient molecule, and their speed makes this information accessible in early-stage candidate discovery and development.