Impact of Glycosylation on Protein-Protein Self-Interactions of Monoclonal Antibodies

Protein self-interactions measured via second osmotic virial coefficients (B-22) and dynamic light scattering interaction parameter values (k(D)) are often used as metrics for assessing the favorability of protein candidates and different formulations during monoclonal antibody (MAb) product development. Model predictions of B-22 or k(D) typically do not account for glycans, though glycosylation can potentially impact experimental MAb self-interactions. To the best of our knowledge, the impact of MAb glycosylation on the experimentally measured B-22 and k(D) values has not yet been reported. B-22 and k(D) values of two fully deglycosylated MAbs and their native (i.e., fully glycosylated) counterparts were measured by light scattering over a range of pH and ionic strength conditions. Significant differences between B-22 and k(D) of the native and deglycosylated forms were observed at a range of low to high ionic strengths used to modulate the effect of electrostatic contributions. Differences were most pronounced at low ionic strength, indicating that electrostatic interactions are a contributing factor. Though B-22 and k(D) values were statistically equivalent at high ionic strengths where electrostatics were fully screened, we observed protein-dependent qualitative differences, which indicate that steric interactions may also play a role in the observed B-22 and k(D) differences. A domain-level coarse-grained molecular model accounting for charge differences was considered to potentially provide additional insight but was not fully predictive of the behavior across all of the solution conditions investigated. This highlights that both the level of modeling and lack of inclusion of glycans may limit existing models in making quantitatively accurate predictions of self-interactions.