"Assessing Impact of Hinge Flexibility on Predicted Second Osmotic Virial Coefficients".

Abstract

Monoclonal antibodies (MAbs) are a key modality for treating a range of diseases because of their unique biophysical properties, such as high binding affinity and high specificity. However, MAb solutions can have unpredictable behavior that is detrimental to the drug product including aggregation and self-association, and high viscosity at elevated protein concentrations. Coarse-grained (CG) molecular simulations provide an opportunity to probe antibody behavior and self-interactions early in development without large experimental or computational burden. Recent work used a 1-bead-per-charge with 1-bead-per-domain (1bC/D) to model a combination of screened electrostatic, steric, and short-ranged non-electrostatic interactions to accurately predict experimental protein self-interactions for MAbs but neglected the influence of MAb hinge flexibility. This work includes the effects of flexibility of the hinge region while maintaining the 1bC/D resolution and computational efficiency. The flexibility is modeled by intramolecular rotations and flexing of antibody fragments about the central hinge to capture literature results for the distribution of internal structures for a single MAb. The difference between flexible and rigid models are analyzed for two body interactions for a reasonably large data set (n = 63) of different MAbs at typical commercial solution conditions. The net behavior showed small differences for the flexible vs. rigid model for most MAbs, within the range of experimental results, with a small number of exceptions. While the overall MAb-MAb self-interactions were not largely dependent on intramolecular degrees of freedom of the hinge region, there were some predicted differences in particular amino acid pairwise interactions from flexible to rigid models, which may indicate the additional computational burden of including hinge flexibility would be useful for future work focused on protein design and extensions to high protein concentration drug development where there are multi-protein spatial correlations.