Process Analytical Technology (PAT) for Bioprocessing

Presented by Cyrus Agarabi

Cyrus Agarabi is senior regulatory research review officer at the Division of Biotechnology Review and Research-II OBP/OPQ/CDER, US Food and Drug Administration.

Disclaimer: This article reflects the views of the author and should not be construed to represent official views or policies of the US FDA.

The office of pharmaceutical quality (OPQ) oversees eight offices, of which the Division of Biotechnology is one. The Division of Biotechnology is responsible for overseeing monoclonal antibodies (MAbs) and therapeutic proteins. We perform regulatory research as an outreach activity with internal and external stakeholders to determine where the pinch-points in production of biological products occur and where they are not being addressed by industry or academia. That information helps our office to support FDA guidance documents on areas such as process analytical technology (PAT) and quality by design (QbD). It also helps us to provide internal and external training on bioprocessing equipment and analytics to our reviewers to support biotech regulatory and review decisions and develop policy.

In this presentation, I discussed the types of regulatory research we perform in the context of PAT. The essence of PAT is controlling and monitoring product quality during manufacturing processes. I demonstrated where PAT can be useful in bioprocess development for the production of MAbs in two different case studies. In the first case study, I examined the nutrient content in 5-L bioreactors to determine its effects on critical quality attributes (CQAs). I compared our standard approach, which is to use Bioprofile® flex analyzer (Nova Biomedical) for on line auto-sampling, with a BioPAT® Trace (Sartorius Stedim Biotech) that performs real-time nutrient analysis.

The benefit of using BioPat Trace is that it is a direct measurement. There is no modeling because it uses an internal calibration. But it does require an orthogonal offset, so it is not a standalone piece of equipment. Additionally, the BioPat Trace can run in dialysis mode, in which it does not require a large sample volume and can be run at high cell densities. For our bioprocess runs we used a murine suspension hybridoma (MAb IgG3k), cultured in either batch mode using serum-free media or fed-batch mode using media with glutamine (60 mL, 1.5 mmol/L) added at 48 and 72 hours or media with glutamine (60 mL, 1.5 mmol/L) and a nonessential amino acid solution (1× concentration) added at 48 and 72 hours. We monitored glucose, cell density, productivity, and glycan content. The data showed that fed-batch cultures performed better with increased viable cell density and antibody production. Product quality also was improved with changes in glycan patterns with fedbatch media. On-line glucose analysis results from the BioPat Trace, and the Bioprofile® Flex analyzer results were comparable. Therefore, we concluded that using in-line glucose analysis would provide increased amounts of data and could be a good potential tool for feeding decisions.

In the second case study, we examined perfusion culture in 5-L bioreactors to determine its effects on CQAs. We compared the Bioprofile Flex analyzer for on-line autosampling with a Biomass probe (ABER), which performs real-time biomass analysis using dielectric spectroscopy technology. In collaboration with the MIT Engineering Practice School, we ran batch, fed-batch, and perfusion cultures in 5-L bioreactors for 14 days. For our bioprocess runs, we used MAb IgG3k and Chinese hamster ovary (CHO) cell line DG-44 cultured in chemically defined medium (NaHCO3 buffered, 5 g/L glucose) and supplemented with 10 mM glutamine and 3% antifoam C added as required. We monitored viable cell density, glucose, and titer. Data showed that viable cell density and yield were higher in the perfusion cultures. The data also showed our feeding strategy was insufficient for all cultures, and the automated feeding strategy developed from the ABER probe data and feedback loops is needed. Real-time viable cell density results from the ABER probe and the Bioprofile Flex analyzer were comparable. The probe also was able to track trends and fluctuations in cellular growth and viability throughout the culture but does require development of an offset based on robust off-line data.

In conclusion, PAT can be used in bioprocessing for real-time decision making and has a promising future in applications such as continuous manufacturing and advanced technologies to support CQAs. PAT may allow scientists to make decisions based on feed timing, quantity, and composition, as well as when to harvest a product. It also can help global culture health trajectory and support QbD and multivariate data analysis (MVDA).