Computational Procedure for Predicting Excipient Effects on Protein-Protein Affinities

Area of Science:

• Biochemistry
• Computational Biology
• Pharmaceutical Sciences

Background:

• Protein-protein interactions are central to biological processes.
• These interactions pose challenges in formulating biologic drugs like antibodies.
• Excipients are small molecules used to weaken protein-protein interactions and mitigate formulation issues.

Purpose of the Study:

• To develop a computationally efficient model for predicting the viscosity-reducing effects of excipients.
• To assess the model's ability to rank excipients based on their impact on protein solution viscosity.
• To explore the utility of the model for predicting excipient effects on protein association and phase separation.

Main Methods:

• Atomic-resolution molecular dynamics (MD) simulations.
• Binding polynomial analysis.
• Thermodynamic perturbation theory.

Main Results:

• A computationally efficient model was developed by integrating MD simulations, binding polynomials, and thermodynamic perturbation theory.
• The model successfully ranked the viscosity-reducing efficacy of four common excipient types for a monoclonal antibody solution.
• The findings demonstrate the model's potential for predicting excipient performance.

Conclusions:

• The developed computational approach effectively predicts the viscosity-reducing effects of excipients on protein solutions.
• This model can aid in the rational design of stable biologic drug formulations.
• The approach is valuable for understanding excipient impact on protein association, phase separation, and buffer effects.