Predictive chromatographic simulations for the optimization of recovery and aggregate clearance during the capture of monoclonal antibodies

Abstract

Predictive chromatographic simulations were used to assess whether significant aggregate clearance, in addition to high step recovery and limited eluate pool volumes, can be achieved during protein A affinity chromatography capture steps. Such aggregates of the antibody monomer are commonly found in manufacturing processes. A lumped desorption-kinetic limiting model was used to describe the elution from the chromatography column, as batch isotherm measurements indicated no adsorption under elution conditions. In order to quantify the trade-off between step recovery and aggregate clearance, independent experiments were first performed to obtain the key kinetic parameters. These parameters were used in simulations to predict the behavior of bench-scale protein A column runs and identify robust operating windows within which good yields and significant aggregate clearance can be achieved. Two examples are described. For antibody A, a robust window of operation was identified. In this case, the optimal conditions were transferred to pilot-plant scale, and the resulting experimental data were shown to be in good agreement with model predictions. For antibody B, it was found that conditions resulting in high recovery and good aggregate clearance were not robust: at the optimal elution conditions, changes of ±0.1 units in pH or ±1 mS/cm in conductivity affected the results substantially.