jp911706q_si_001.pdf (31.42 kB)

Prediction of Aggregation Prone Regions of Therapeutic Proteins

Download (31.42 kB)
journal contribution
posted on 20.05.2010, 00:00 by Naresh Chennamsetty, Vladimir Voynov, Veysel Kayser, Bernhard Helk, Bernhardt L. Trout
Therapeutic proteins such as antibodies are playing an increasingly prominent role in the treatment of numerous diseases including cancer and rheumatoid arthritis. However, these proteins tend to degrade due to aggregation during manufacture and storage. Aggregation decreases protein activity and raises concerns about an immunological response. We have recently developed a method based on full antibody atomistic simulations to predict antibody aggregation prone regions [Proc. Natl. Acac. Sci. 2009, 106, 11937]. This method is based on “spatial-aggregation-propensity (SAP)”, a measure of the dynamic exposure of hydrophobic patches. In the present paper, we expand on this method to analyze the aggregation prone regions over a wide parameter range. We also explore the effect of different hydrophilic mutations on these predicted aggregation prone regions to engineer antibodies with enhanced stability. The mutation to lysine is more effective than serine but less effective than glutamic acid in enhancing antibody stability. Furthermore, we show that multiple simultaneous mutations on different SAP peaks can have a cumulative effect on enhancing protein stability. We also investigate the accuracy of various cheaper alternatives for SAP evaluation because the full antibody atomistic simulations are highly computationally expensive. These cheaper alternatives include antibody fragment (Fab, Fc) simulations, implicit solvent models, or direct computations from a static structure (i.e., a structure from X-ray or homology modeling). The SAP evaluation from the static structure is 200 000 times faster but less accurate compared to the SAP from explicit atom simulations. Nevertheless, the SAP from a static structure still predicts most of the major aggregation prone regions, making it a potential approach for use in high-throughput applications. Thus, the SAP technology described here could be employed either in high-throughput developability screening of therapeutic protein candidates or to improve their stability at later stages of manufacturing.