Predicting Formulation Conditions During Ultrafiltration and Dilution to Drug Substance Using a Donnan Model with...

Area of Science:

• Biopharmaceutical Manufacturing
• Computational Chemistry
• Protein Chemistry

Background:

• Manufacturing therapeutic monoclonal antibodies (mAbs) involves purification steps like ultrafiltration/diafiltration (UF/DF).
• Accurate formulation requires understanding Gibbs-Donnan and volume-exclusion effects on pH and buffer concentration during UF/DF.
• Current methods often rely on extensive experimental characterization.

Purpose of the Study:

• To develop and validate an in silico model for predicting drug substance (DS) pH and buffer concentration.
• To account for Gibbs-Donnan and volume-exclusion effects in UF/DF, dilution, and conditioning steps.
• To compare structure-based vs. sequence-based protein charge calculations.

Main Methods:

• Development of a computational model to predict DS pH and buffer concentration.
• Validation of the model against experimental data for nine different mAbs.
• Utilizing statistical analysis for model performance assessment.
• Employing a structure-based approach for calculating protein charge.

Main Results:

• The in silico model accurately predicted DS pH and buffer concentration.
• A structure-based protein charge calculation proved more accurate than a sequence-based method.
• The model provided insights into the Gibbs-Donnan effect, emphasizing protein charge concentration's role.

Conclusions:

• An in silico model can effectively predict Gibbs-Donnan and volume-exclusion effects in mAb purification.
• This computational approach may reduce the necessity for extensive experimental work.
• Structure-based charge calculation enhances model accuracy for biopharmaceutical development.