Non-native protein aggregation is a ubiquitous concern when bringing a new protein-based drug or formulation to market. Aggregation can proceed through a variety of mechanisms and is influenced by a number of factors. The resulting aggregates adversely affect product potency, must be closely controlled to meet regulatory requirements, and potentially pose an immunogenicity concern. Rapidly creating a stable formulation that minimizes aggregation is an on-going challenge within the pharmaceutical industry, and would benefit from more predictive computational and experimental tools. In the longer term, this work seeks to utilize a combination of experimental and computational techniques to determine biophysical properties and aggregations behavior across various formulation conditions. The present work first focuses on characterizing a series of monoclonal antibodies (mAbs) that display qualitatively different interactions and aggregation behavior at pH and ionic strength ([NaCl]) conditions that are commonly employed for product formulation. The short and medium time-scale aggregation rates/behaviors were measured and compared against the thermal conformational stability, cloud point, viscosity measurements, and protein-protein weak interaction parameters (kD, B22, and G22) obtained from light scattering measurements. This was done in order to assess if useful correlations between protein aggregation behavior and biophysical properties were viable. Additionally, computer simulations of protein-protein interactions (e.g. B22 and G22) were performed via a coarse-grained molecular model coupled with Monte Carlo algorithms and compared against experimental B22 and G22 values, as well as experimental kD values. B22 values from a priori calculations matched experimental B22 values semi-quantitatively for some situations, whereas others required additional parameter tuning. Contrary to prior literature results, thermal scanning thermodynamic properties did not correlate with accelerated aggregation behavior for any of the mAbs studied. Weak protein-protein interactions correlated with aggregation rates for some but not all antibodies, and in some cases were also indicative of electrostatically-mediated phase separation.

Towards predictive models of antibody biophysical properties and aggregation

253rd National Meeting (2017)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.