Antibodies constitute the most rapidly growing class of human therapeutics for the treatment of numerous cancers, chronic inflammatory diseases, and infectious diseases. These antibodies are stored for long term under high concentration conditions as required for administration. These antibodies are, however, thermodynamically unstable under these conditions and degrade due to aggregation. Currently, there is very little mechanistic understanding of how these molecules aggregate. We employ molecular simulation tools in collaboration with experimental techniques to understand the mechanism behind antibody aggregation.

Here we present the insights gained into aggregation through the molecular modeling of therapeutic antibodies. We have performed detailed atomistic simulations for the glycosylated Fc and Fab fragments of the antibody. In order to achieve this, we developed the necessary model parameters for sugar groups involved in glycosylation. We have also built a full antibody from the Fab and Fc fragments using another full antibody as a model template. From these molecular simulations, we studied the dynamic fluctuations and solvent exposure of various regions in the antibody, and how these depend on the different glycosylation patterns. These characteristics are used to develop a measure that indicates the regions of the antibody most prone to aggregation. These aggregation prone regions predicted from molecular simulations are currently being validated through experiments.