Self-­�Interaction Chromatography (SIC) of mAbs: New Methods for Estimating the Dead Volume in SIC and using SIC to Predict mAb Stability

S. Hedberg*, A. Quigley*, J.Y.Y Heng*, D.R. Williams* and J. Liddell**

*Surface and Particle Engineering Laboratory **Process and Analytical Sciences Department of Chemical Engineering Fujifilm Diosynth Biotechnologies Imperial College London, SW7 2BY, Billingham, TS23 1LH, United Kingdom United Kingdom

ABSTRACT

Protein-­�protein molecular interactions are known to be involved in protein solution aggregation behaviour; however the mechanisms leading to protein aggregation are still not fully understood. The osmotic second virial coefficient (B22) is a fundamental physiochemical property that describes protein-­�protein interactions solution, which can be a useful tool to predict aggregation propensity of proteins. This work includes two experimental SIC studies on both model proteins and therapeutic mAbs of different sizes. The first study is an evaluation of two different experimental techniques used to determine SIC dead volumes and the second study uses SIC results for mAb to predict stability.
Accurate dead retention volumes are essential for the accurate determinations of B22. The traditional method of estimating dead volume for SIC includes the use of a dead column (without protein immobilised) where the retention volume for proteins can be established. For this technique the dead volume was established for the proteins over a wide range of solution conditions (pH and salt concentrations), and then compared with a new method, where a number of non-­â?interacting dextrans of different molecular weights (MW) (including the MWâ??s of the protein) were employed to find the dead retention volume. The results for the traditional technique with a dead column changed depending on the protein used; only certain model proteins kept a constant dead retention volume when the pH was changing under a constant high salt concentration to minimise protein-­â?surface interactions. Several proteins, including the mAb, exhibited an increased dead retention volume especially when exposed to lower pH. From this it can be concluded that there is no absolute dead volume that can be determined by this technique which are independent of solution conditions. The new technique involving dextrans gives a better overall result for the dead volume for proteins such as mAbs.
The second study shows that the SIC data gives a strong direct correlation between
B22 values and aggregation behaviour of a range of proteins in solution. SIC data for a mAb was not strongly affected by increasing salt concentration; with a B2 (dimensionless osmotic second virial coefficient) of around -­�1 to -­�2. This data confirms the previous established stability for this mAb which made it as a suitable therapeutic candidate.