Karin Westerberg

Modeling and robust pooling design of a preparative cation-exchange chromatography step for purification of monoclonal antibody monomer from aggregates.

This study has implemented and calibrated a model that describes the separation of the monomer of monoclonal antibodies from the dimer and larger oligomers on preparative-scale using cation-exchange chromatography. A general rate model with temperature dependent diffusion was coupled to a pH- and temperature-dependent steric mass action model. The model was shown to predict the retention of the monomer, dimer, and oligomer at low loadings for different pH levels and temperatures. Additionally, the model was shown to adequately predict the elution behavior of the monomer and soluble aggregates at high loadings within the same ranges with some limitations. The model was not able to accurately describe the shape of the product break-through curves or the slight levels of co-elution of the dimer and oligomer with the monomer at higher pH. The model was used to predict how 12 process variations impact the separation. The model is used to establish an elution end collection criterion such that the step can robustly provide the target purity of monomers.

Effects of uncertainties in experimental conditions on the estimation of adsorption model parameters in preparative chromatography

Model-based process design is increasingly popular when designing pharmaceutical purification processes. The effect of uncertainties in concentration measurements on the estimation of model parameters is analyzed for two cases of non-isocratic adsorption chromatography. A model, calibrated to experiments, is used to generate data by adding a Monte Carlo sampled error in the inlet concentrations. New model parameters are estimated by minimizing the deviation between the synthetic data and the model. The first case is a separation of rare earth elements by ion-exchange chromatography and the second case is a purification of insulin from a product-related impurity by reversed-phase chromatography. It is shown that normally distributed errors in the concentrations result in deviations in the UV-signal that are not normally distributed. With the applied method, known concentration distributions can be translated into probability distributions of the model parameters, which can be taken into account in the model-based process design

Pooling control in variable preparative chromatography processes.

Preparative chromatographic columns that run at high loads are highly sensitive to batch-to-batch disturbances of the process parameters, placing high demands on the strategy used for pooling of the product fractions. A new approach to pooling control is presented in a proof-of-concept study. A model-based sensitivity analysis was performed identifying the critical process parameters to product purity and optimal cut points. From this, the robust fixed cut points were found and pooling control strategies for variations in the critical parameters were designed. Direct measurements and indirect measurements based on the UV detector signal were used as control signals. The method is demonstrated for two case studies of preparative protein chromatography: hydrophobic interaction and reversed phase chromatography. The yield improved from 88.18 to 92.88% when changing from fixed to variable pooling in hydrophobic interaction chromatography, and from 35.15 to 76.27% in the highly sensitive reversed phase chromatography.

Model‐based risk analysis of coupled process steps

A section of a biopharmaceutical manufacturing process involving the enzymatic coupling of a polymer to a therapeutic protein was characterized with regards to the process parameter sensitivity and design space. To minimize the formation of unwanted by‐products in the enzymatic reaction, the substrate was added in small amounts and unreacted protein was separated using size‐exclusion chromatography (SEC) and recycled to the reactor. The quality of the final recovered product was thus a result of the conditions in both the reactor and the SEC, and a design space had to be established for both processes together. This was achieved by developing mechanistic models of the reaction and SEC steps, establishing the causal links between process conditions and product quality. Model analysis was used to complement the qualitative risk assessment, and design space and critical process parameters were identified. The simulation results gave an experimental plan focusing on the “worst‐case regions” in terms of product quality and yield. In this way, the experiments could be used to verify both the suggested process and the model results. This work demonstrates the necessary steps of model‐assisted process analysis, from model development through experimental verification. Biotechnol. Bioeng. 2013; 110:2462–2470.

Numerical Analysis of Model Parameter Uncertainties as a Result of Experimental Uncertainty — An Example from Preparative Chromatography

Abstract Model calibration, and in particular model parameter uncertainty caused by experimental errors, is the focus of this work. Computer simulations were used to design a purification step for insulin by reversed-phase chromatography. The effect of errors in the protein sample concentration and purity, and in the modifier concentration in the sample, equilibration, and elution buffers was studied on the calibration of the adsorption kinetic parameters by the inverse method. The overall error, including experimental errors, was not normally distributed and not uncorrelated. Monte Carlo simulations were performed where the calibrated model was used to generate new data sets and a random error was added on the experimental conditions. New model parameter sets were found by recalibrating the model to the data sets from Monte Carlo simulations and the model parameter covariances were estimated from these. A control strategy which was robust to uncertainty in both model and process was designed from the resulting model parameter distribution and the expected variations in the process variables.