OR WAIT null SECS
A quality-by-design approach that implements PAT offers advantages in upstream cell-culture processing.
Much of the commercial production of biologic therapeutics is based on the use of recombinant cell lines, and product quality is verified through testing. This testing approach typically depends on fixed process conditions and extensive testing of the end product (1). However, using a quality-by-design (QbD) approach that employs predictive modeling can improve the quality of end product as well as the efficiency of the bioprocess.
Integrating process analytical technology (PAT) into the bioproduction process enables the biomanufacturer to move from a quality-by-testing approach to a QbD approach, which is more flexible (1). By applying advanced sensor systems combined with mathematical modeling techniques, one can gain an enhanced understanding of the processes that produce the biologic product.
In addition, this use of PAT would allow for on-line prediction of the critical quality attributes (CQAs) required to maintain product quality and offer the ability for real-time product quality control.
Mammalian cells are most frequently used as the hosts for manufacturing complex biopharmaceuticals, and cultivation of these host cells is a key factor in the production of biotherapeutics. The cultivation step impacts both product yield and product quality (1).
Because there is a lack of techniques for real-time measurement of product attributes, quality assurance in biopharmaceutical processes still mainly relies on the repetition of identical process settings and extensive end-product testing.
Even in a bioprocess with identical process settings, however, variability in end-product quality output is still likely to occur due to the variability of inputs, such as the raw materials and the living host cells themselves (1).
Using a QbD strategy that implements PAT offers the advantage of real-time process monitoring and, subsequently, quality control based on the application of multivariate data analysis and mathematical modeling techniques that can help predict product outcome. QbD starts by defining the relevant target product profile and CQAs.
Subsequently, development of the manufacturing process takes into account the likely impact that process parameters will have on process response, using knowledge gained through risk-assessment analysis (e.g., instant failure mode and effects analysis).
Once those steps have been developed, operators can determine the relationships between critical process parameters and product quality attributes by applying statistical design experiments, multivariate data analysis, and mathematical modeling techniques. This will, in turn, inform the design of the operation space and the sensor technologies needed to monitor the critical process variables in real-time (1).
Because the QbD approach aims to achieve closed-loop CQA control, model predictive control (MPC), a methodology used for multivariate control in many other process industries, would be best suited for bioprocessing. The goal of MPC is to meet various CQA specifications by manipulating process inputs while also taking into consideration process constraints.
1. W. Sommeregger et. al, “Quality by Control: Towards Model Predictive Control of Mammalian Cell Culture Bioprocesses,” Biotechnology Journal online, DOI: 10.1002/biot.201600546, March 31, 2017.
Volume 30, Number 12
When referring to this article, please cite as F. Mirasol, “QbD Improves Cell-Culture Process Control,” BioPharm International 30 (12) 2017.