The variable domains in Fab regions are well-established contributors to high concentration viscosity of IgG, primarily through charge and hydrophobic interactions. In contrast, the roles of the Fc and the number of self-interacting sites (valency) are less well understood. Here, we investigate the relative contributions of Fab–Fab and Fab–Fc interactions to high concentration viscosity for a diverse panel of 20 IgG 1 antibodies, by rheometry, coarse-grained simulations, and molecular surface property analysis. Strikingly, fragmentation of IgG 1 into F(ab′) 2 plus Fc reduced viscosity (−11 to −93%) for all antibodies tested, demonstrating prevalent contributions of Fc to viscosity. Coarse-grained simulations with one site per Fab and two sites per Fc qualitatively tracked trends in experimental rheometry data for 20 parental antibodies and their fragments. In these simulations Fab–Fab and Fab–Fc interaction strengths were independently varied to capture possible interaction differences arising from parental sequences or any mutations. These coarse-grained simulations suggest that Fab–Fc attractions generate branched IgG 1 networks and disproportionately larger clusters relative to Fab–Fab interactions of comparable strength. This study suggests that a four-site self-interaction model, previously proposed for a single antibody (omalizumab), is broadly applicable to diverse IgG 1 . Beyond well-established variable domain engineering, this self-interaction model predicts that Fc engineering may reduce IgG 1 viscosity, a much sought after goal to enable subcutaneous delivery. Clinically validated Fc mutations are demonstrated here to substantially reduce the viscosity for multiple IgG 1 (−33 to −91% reduction, n = 6), supporting this emerging antibody design concept.
Heisler et al. (Tue,) studied this question.