Isotonic concentrations of excipients control the dimerization rate of a therapeutic immunoglobulin G1 antibody during refrigerated storage based on their rank order of native‐state interaction

Abstract

Inert co-solutes, or excipients, are often included in protein biologic formulations to adjust the tonicity of liquid dosage forms intended for subcutaneous delivery. Despite the low concentration of their use, many of these excipients alter protein-protein interactions such as dimerization and aggregation rates of high concentration monoclonal antibody (mAb) therapeutics to varying extents during long-term refrigerated clinical storage, challenging the formulation scientist to make informed excipient selections at the earliest stages of development when protein supply and time are often limited. The objectives of this study were to better understand how isotonic concentrations of excipients influence the dimerization rates of a model mAb stored at refrigerated and room temperatures and explore protein sparing biophysical methods capable of predicting this dependence. Despite their prevalence of use in the biopharmaceutical industry, methods for assessing conformational stability such differential scanning calorimetry and isothermal equilibrium unfolding showed little predictive power and we highlight some of the assumptions and technical challenges of their use with mAbs. Conversely, measures of colloidal stability of the native-state such as preferential interaction coefficients measured by vapor pressure osmometry and solubility assessed by polyethylene-glycol induced precipitation correlated reasonably well with the mAb dimerization data and are most consistent with the excipients tested minimizing dimerization by interacting favorably with the residues comprising the protein-protein association interface.

Suppression of Electrostatic Mediated Antibody Liquid–Liquid Phase Separation by Charged and Noncharged Preferentially Excluded Excipients

“…These values and the corresponding interaction parameters and transfer free energies (Figure S5 of the Supporting Information) indicate thermodynamically unfavorable protein–excipient interactions (negative Γ μ3 and positive Δμ 2,tr ) by amounts that show no apparent relation to the excipient rank order of LLPS attenuation (Figure ). Similar findings have been reported for other mAbs where preferentially excluded excipients have been shown to enhance solubility. ,− We believe these results, at least those for mAb-B, suggest that the excipient–protein interactions cannot be assumed to be delocalized or evenly distributed over the surface of the protein. In this case, excipient preferential interaction of only the specific residues involved in protein–protein interaction would determine the excipient dependence of LLPS and not excipient net accumulation/depletion relative to its concentration in bulk solution.…”

“…32,33 Unfortunately, testing this nonnative LLPS mechanism for mAb-B is not straightforward given the difficulty in measuring mAb conformational stability due to their lack of unfolding reversibility. 11 We therefore measured the thermal stability of mAb-B formulated in each excipient by DSC because this method is one of the most widely used within the industry as a metric of conformational stability.…”

“…One potential mechanism for how these excipients may be lowering the mAb-B C T is if the PPIs responsible for mAb-B LLPS are mediated through exposed hydrophobic side chains and/or the peptide backbone typically found in the native state interior. A number of the excipients studied in this report are known to be preferentially excluded from these residues , and their additions have been used successfully in past reports to shift the equilibrium away from unfolded/partially unfolded species (i.e., increase conformational stability) to limit the non-native aggregation rates of other proteins, particularly when stored under stressful conditions such as at elevated temperatures, extreme pH, or in the presence of low concentrations of chemical denaturants. , Unfortunately, testing this non-native LLPS mechanism for mAb-B is not straightforward given the difficulty in measuring mAb conformational stability due to their lack of unfolding reversibility . We therefore measured the thermal stability of mAb-B formulated in each excipient by DSC because this method is one of the most widely used within the industry as a metric of conformational stability.…”

“…A better understanding of this mechanism is of significant practical importance to the biopharmaceutical industry in saving the time and resources needed to conduct large empirical formulation screening studies at the earliest stages of development when protein supply and time are often limited. Prior research in this area has examined both the effects excipients have on mAb conformational stability and how this may alter non-native PPIs − as well as how excipients influence the PPIs of the native state. − This investigation explores both mechanisms with an emphasis on the latter using an immunoglobulin 4PE (IgG4PE) named mAb-B as a model protein that undergoes liquid–liquid phase separation (LLPS) or the partitioning into protein-poor and protein-rich liquid phases at room temperature when formulated at low ionic strength and buffered at neutral pH near its isoelectric point.…”

Analytical Workflows to Unlock Predictive Power in Biotherapeutic Developability

“…A significant observation from this analysis is that the majority of samples fall on the two quadrants where there is either a colloidal or a conformational liability, while the other two quadrants are sparsely populated. Several mechanisms driving the effect of different salts and excipients on protein stability have been proposed, including changes in the water structure, preferential binding or exclusion from the protein surface or competition for hydration between the protein and ions or excipients [ 40 , 42 , 46 , 47 ]. What all the various theories have in common is that they observe a trade-off between conformational and colloidal stability.…”