Improving Product Quality by Understanding Bioreactor Conditions

By: Dr. Nick Hutchinson

Conditions within the bioreactor can heavily influence the quality of biopharmaceutical products. Often significant concerns at this stage in the process are those quality attributes affected by post-translational modifications such as glycosylation patterns. Decisions that can influence product quality from the very start of projects include both the choice of cell line and appropriate clone selection.

During cell culture development, companies are increasingly adopting a quality-by-design approach that is support by the prevailing regulatory environment. This approach requires scientists to identify those bioreactor conditions that generate product with the necessary critical quality attributes during process development. They achieve this by manipulating process parameters and measuring product quality under the different process conditions generated.  There are a variety of different methods for doing this, however, a DOE approach is highly efficient and allows scientists to identify interactions between process variables. It is often useful to use two-stage approach to DOE. In the first stage, an initial screen of process parameters identifies those that exert an effect on product quality. During the second stage, firms elucidated the nature of the relationships between this subset of parameters and quality attributes using a response surface method. Determining these relationships allows companies to define a design space within which they can operate their process whilst achieving the desired product quality. They achieve a greater understanding of the way in which process parameters influence the quality of their products.

Miniature Bioreactors Allow Rapid Design Space Characterization

Characterizing a cell culture process in this way is rigorous and allows biomanufacturers to demonstrate to regulators that they understand the impact of process changes on patient safety. Performing the large number of necessary experiments in parallel using automated miniaturised bioreactors avoids excessive media costs and ensure adherence to short project timelines. Companies adopting this approach must consider how they will manage the analysis of the large number of samples that the experiments will generate and the volume of data that they will accumulate.

Companies must be careful that the design space they define in the laboratory is one in which they can operate within at the intended scale of production. Scientists working for biopharmaceutical firms, have in the past, defined design spaces in which production scale equipment could not operate. To avoid this scenario, scientist should confirm the performance characteristics of large-scale bioreactors before running extensive multivariate experiments in the laboratory. This, along with careful scale-up using appropriate models, will allow companies to have a high degree of confidence in their ability to produce product with the required quality attributes at the commercial-scale.

New Sensors Provide Unprecedented Amount of Data on Cultures

Sensors and other process analytical technologies allow cell culture engineers to monitor conditions within the bioreactor. This is a rapidly developing field with new sensors and formats becoming available to the industry on a regular basis. For example, the industry has started using impedance spectroscopy to monitor viable biomass in real-time within single-use bioreactors. It is likely that we can expect this trend to continue giving engineers unprecedented amounts of data from their cell cultures. To convert this data into knowledge, however, requires its effective analysis. Multivariate methods such as PCA and PLS can be used to make sense of the myriad of data generated by the bioreactors used for cell culture.

Upstream processing engineers have used PCA to visualize the variation observed between comparable cell cultures. This can help them identify outlying batches that warrant further investigation. They can then put in place actions that prevent batches from outlying in the same way.

Real-time Multivariate Process Monitoring

While this is useful, multivariate tools that allow companies to analyse data from their cell cultures in real-time can allow interventions that save batches that would otherwise be lost. All the data from the batch is summarised in a single curve following a “golden batch” trajectory with upper and lower limits. Engineers use a statistical analysis of previous runs that generated product with the necessary quality attributes to derive the “golden batch” trajectory.

Merck is an example of a company that has previously reported on the successful use of real-time multivariate process monitoring. The company’s shift supervisors say, “the biggest benefit is early detection of faults and deviations”. The company has communicated that real-time batch trajectories have informed actions in half of all batches where they used the technique and saved three batch failure over an 18 month period.

The industry can continue to improve its approach to ensuring product quality through an ever-greater understanding of process conditions. However, it has already made significant improvements in this field over the last 20 years. Processes are increasing well characterized during the development phases and the tools available for cell culture process monitoring continue to improve. A more widespread adoption of real-time multivariate data analysis will help the industry gain even greater control of biopharmaceutical product quality.


Dr Nick Hutchinson


About the author: Nick Hutchinson is a Technical Content Marketing Manager at Sartorius Stedim Biotech.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s