January 1, 2013 (Vol. 33, No. 1)
Scale-Down Reactor Automates Parameter Control and Facilitates Development
The selection and optimization of new recombinant clones is a key step in early bioprocess development. Traditionally, approaches such as shake flasks and benchtop bioreactors have been used to select the most productive clones, as well as define the optimum media and bioprocessing conditions to culture them.
Due to the scale and limited throughput of these strategies, researchers are faced with a number of limitations. The capacity for using benchtop bioreactors is especially limited as it is resource-intensive and has high capital equipment and infrastructure costs.
Consequently, scientists frequently cannot perform full DOE (design-of-experiments) and are generally only able to take one or two of their most promising clones forward for partial DOE runs in benchtop bioreactors. These clones often perform as expected in scale-up, but scientists are never certain if they would have identified critical bioprocessing parameters or an even better performing clone if they could have performed more runs in benchtop bioreactors.
This need to conduct large numbers of experiments under bioreactor conditions has resulted in the development and widespread use of miniaturized high-throughput technologies including bubble columns and microplate-based bioreactors.
The drawback with many current miniaturized bioreactor systems is that they do not mimic the sparged, stirring action of a bioreactor since they do not have internal impellers but use instead a gas bubble, shaking, or another method of stirring. They also tend to have a working volume of less than 1 mL, which can make running analytical testing on cell cultures problematic because there are often insufficient amounts of material for testing, particularly where multiple samples are required from a single reactor during the run.
Novel systems have been introduced with much more bioreactor-like characteristics and volumes in the 10–15 mL scale, which have been shown to be an accurate model system for use in processes such as cell-line selection. However, these tools do not have the extended feed and control capabilities that make them suitable for microbial fermentation or for use as accurate scale-down models for production-scale systems.
Therefore, there remains a need for enhanced bioreactor models that have the capability to automatically control temperatures, sparge, stir, and add feeds with suitable characteristics to support the demands of microbial fermentation, serve as a more precise scale-down model for manufacturing, and provide enough working volume to collect larger sample sizes and allow more frequent sampling for analysis.
TAP Biosystems recently announced the expansion of its bioprocessing portfolio with a new system offering increased volumes and compatibility with microbial fermentation. The ambr250 automated bioreactor has three components: ambr250 single-use fully disposable 250 mL bioreactors, automated ambr250 workstation, and ambr250 software (Figure 1).
The platform, which builds on the ambr micro bioreactor technology, was developed with the following improvements in mind: increased fermentation volumes, pumped liquid delivery for continuous feeds and automated individual bioreactor temperature, and impeller control. These features allow a more realistic, frequent feeding regime, as well as larger volumes to be sampled to perform a wider range of analytical tests.
This, combined with the parallel control of culture conditions and feeds, provides a true scale-down bioreactor model which will support Quality by Design (QbD).
The workstation can be configured for independent parallel control of 12 or 24 single-use 250 mL bioreactors and includes liquid-handling automation to provide culture set-up and inoculation, automated addition of feeds and alkali, and culture sampling.
Additionally, the workstation controls the stir speed, gas supply, and temperature and because it integrates laminar flow enables automated aseptic feed additions and sampling.
The ambr250 includes single-use, fully disposable bioreactors (100–250 mL working volume), each with independent temperature and impeller control. The bioreactors are supplied sterile and individually wrapped. They have been designed to be installed with only three quick steps so that setting up 24 bioreactors in the workstation manifold takes less than 30 minutes.
The bioreactors have a cap for automated sampling and incorporate integrated sensors for DO and pH, which provide individual closed loop control of these parameters. The microbial bioreactors have a dual 20 mm Rushton impeller, automated pH regulation through control of CO2, and liquid alkali addition.
A cascade approach is possible for DO control for fermentation and through control of O2 and N2/air in the mammalian configuration. Since there is no need to clean or sterilize the bioreactor between runs, there is a fast turnaround between experiments, thus saving time and providing a high-throughput system.
The ambr250 system is supported by proprietary software, which includes modules for defining and running experiments and for viewing the resulting data. In the definition module, users can program the workstation to perform actions such as add feeds and induce their cells at specific time points. The definition module is intuitive to use for bioprocess scientists as it lists all the steps that can occur in any type of fermentation so researchers simply pick the parameters they need to use.
In the run module, scientists can start their fermentation runs with one button press and because the software is very flexible, they can edit their process parameters if they need to, while the runs are in progress. For example, with E. coli fermentation, users can change an exponential feed to a linear one without interrupting the run. In the view module, researchers can view their data and also easily export it into Excel for further analysis and production of graphical or tabular results.
The application of the ambr250 for determining fermentation condition is demonstrated in experiments using Pichia pastoris and E. coli. With P. pastoris, parallel fermentations were performed in 2 x 30 L bioreactors, 11 x 3 L bioreactors with the ambr250 demo system using the same fermentation conditions, and P. pastoris clones. The results (Figure 2) show similar cell titers and biomass throughout the 140 hour fermentation run between all three bioreactor types.
Using E. coli, parallel fermentations were performed in three 30 L bioreactors, with the ambr250 demo system using the same fermentation conditions and E. coli clones. Data from two x 30 L historical bioreactor runs was also included in the experimental comparison. The results (Figure 3) show similar oxygen uptake rates (OUR) throughout the 30-hour fermentation run between both bioreactor types, demonstrating that ambr250 can support high-density fermentation.
The application of the ambr250 for determining bioprocess parameters is demonstrated in experiments using CHO clones expressing a recombinant antibody. Fermentations were performed in four 3 L bioreactors, with the ambr250 demo system using the same bioprocessing conditions and CHO clones.
The results (figure not shown) demonstrate comparable cell viability and antibody concentration between both bioreactor types throughout the 14 day bioprocessing runs.
The new ambr250 system replicates the physical characteristics of 3 L and 30 L bioreactors. As a result of being able to rapidly evaluate small volume multiple bioreactors in parallel, the study of many fermentation and bioprocess parameters is no longer limited by availability of benchtop bioreactors, operator time, and facility infrastructure.
The application of single-use bioreactors could reduce the need for aseptic process training on the use of benchtop bioreactors, as well as reduce the number of cleaning and sterilization processes, and the amount of sanitization equipment required.
The automation and on-line monitoring of fermentation and bioprocess parameters could eliminate unsociable working hours for scientists and also provide a more accurate process understanding with less manual effort. The implementation of this new platform could improve data quality and may contribute to allowing more complex statistically designed experiments in both fermentation and bioprocess development.