Speaker: Christopher Roberts, University of Delaware, USA

The Antibodies in Solution: a LINXS – NIST Webinar Series provides background information related to the currently ongoing LINXS antibody research program. This is a concerted experimental and theoretical effort that aims to investigate the properties of monoclonal antibodies in solution, which comprise a major platform for potential drug candidates and are of high academic and pharmaceutical interest. An international consortium of researchers at academic institutions, research centers, NIST and Novartis has teamed up for this. Didactical lectures given by members of the consortium on different experimental and theoretical topics that are highly relevant for state-of-the-art antibody research as well as insights from pharmaceutical industry will be broadcasted. A central aspect of the webinar series will be to illustrate the full power of neutron and X-ray scattering science that can be achieved in combination with complementary experimental methods and different unifying simulation techniques.

Abstract:

Protein-protein self-interactions are involved in a number of important aspects of controlling the solution properties and stability of therapeutic proteins that affect how challenging it can be to develop a viable drug product for biotherapeutics such as monoclonal antibodies (MAbs) and novel engineered antibody formats with antibodies or antibody fragments with payload proteins / domains (e.g., Fc-fusion constructs).   This presentation focuses on a combination of experimental approaches to characterize protein-protein interactions such as small-angle static and dynamic scattering methods, along with coarse-grained molecular models to help address issues such as: predicting high-concentration behaviour; identifying which domains or regions of a protein are most promising to “redevelop” to improve net protein-protein interactions and solution behaviour; and identifying key amino acids for targeted “local protein engineering” to improve the drug product physical characteristics while minimizing risk to clinical efficacy.  Examples include a series of MAbs at low and high concentration conditions and a set of Fc-fusion protein constructs (monovalent and bivalent).  The results highlight that coarse-grained models with sufficient granularity at the level of charge distributions can offer a useful computationally accessible balance for predicting key experimental behaviour with computational burden that should be reasonable for scaling to the larger needs of the industry.