Continuous Processing. Performance Enhancements for Perfusion Applications in 50L to 500L Single-Use Bioreactors: A Technical Comparison of Performance Characterization, Cell Culture & Scale-Up Modeling.

Improvements in Single-Use systems have allowed implementation of high-density cultures in emerging bioprocess workflows while progressive advances in media optimization and improved clone genetic selection have underscored the perceived performance limitations of Single-Use bioreactors (S.U.B.s).

Strategic enhancements to the sparge and agitation systems of Thermo Scientific™ HyPerforma™ S.U.B.s have revealed the potential for a three- to four-fold improvement of mixing and mass transfer performance compared to legacy designs.
New data reveal the benefits of design improvements by demonstrating:

  • Bioreactor characterization, TruBio™ DeltaV™ controller optimization, online process analytics, and scalability analysis of the S.U.B. when targeting perfusion applications from 50L pilot scale to 500L production scale working volumes.
  • High-density culture results (>260E06 cells/mL) while maintaining proper operating parameters. New data reveal how a 50L S.U.B. – equipped with a specialized precision drilled-hole sparger (DHS), Single-Use foam probe, and oversized impeller – is able to improve overall S.U.B. operating efficiency. Results also include specific suggestions on how to maintain a nearly ideal dissolved carbon dioxide environment, reduce headspace foam generation, and produce lower overall shear levels, thus yielding excellent cell viability.

The work also demonstrates best practices and the desirable process benefits that can be achieved through reduced technical risk, lower labor, and simplified technical transfer of a completely disposable processing assembly. Further evidence is presented on the advantages of continuous processing when used in high-density seed train intensification or as a compact production-scale bioreactor system operating at reasonable media exchange rates of one to two vessel volumes per day (VVD).


1. S.U.B. Design Enhancements for Continuous Processing
Table 1 presents the enhancements made to legacy HyPerforma S.U.B.s from 50-500L to specifically convert them to perfusion systems where higher mass transfer and more automation are required.




2. Materials and methods
2.1 Mass transfer testing
Mass transfer studies were performed in all vessels using the standard dynamic method(2) in a test solution at 37oC composed of 1 g/L poloxamer 188 and 3.5 g/L HEPES. Oxygen and CO2 mass transfer was investigated at various gas flow rates and calculated power inputs
(nP = 2.1).


2.2 Cell culture testing
Cell culture was performed in a 50L HyperformaS.U.B. at 40 L working volume integrated with an XCell ATF 6 system. CHO cells were seeded at 0.4 x 106 cells/ in Gibco™ CD OptiCHO™ AGT™ medium with supplemented with necessary growth factors. Operating conditions were set as 30% dissolved oxygen (DO), 7.0 ± 0.2 pH, and 37oC. DO was controlled via oxygen through the DHS and agitation varied as needed to increase mass transfer as oxygen demand increased. Cell counts, nutrients, metabolites, electrolytes, and protein yields were measured offline. Additionally, automation and control of most parameters were desired throughout the culture to ensure a high probability of success. To that end, the following components were integrated into the controller:

  • Standard Single-Use dissolved oxygen sensor, pH sensor, and temperature sensor
  • Foam sensor to automate antifoam delivery and control foam levels. The built-in foam control was linked to the antifoam pump and antifoam was delivered, as needed, to minimize total antifoam usage.
  • Cell density sensor (AberTM). The Aber transmitter signal was integrated into the controller and a bleed pump was linked to the measured cell density to cycle on and off, as needed, to maintain target cell density.
  • Bioreactorload cells to maintain target volume. A calculation block was built using a simple Heaviside equation and linked to the media addition pump to avoid over/under-filling the S.U.B. with new media.

Media for the culture was formulated in a 1,000L Single-Use Mixer and sterilely filtered into a 1,000L S.U.B. maintained at <8oC with a TCU. Spent media from the process was transferred to a 1,000L container.


3. Results
3.1 Improved mass transfer due to design improvements
Figure 2 displays O2 and CO2 mass transfer results for the enhanced Single-Use Bioreactor (S.U.B.) and O2 mass transfer for legacy S.U.B.s using only the DHS at maximum recommended impeller speeds (30 W/m3 for legacy systems, 100 W/m3 for enhanced S.U.B.s). Whereas legacy systems tend to limit near 10-13/hr for kLa, results for the enhanced S.U.B.s show three- to four-fold increases in O2 kLa using only the DHS.



Oxygen mass transfer with the DHS is shown to be higher in larger S.U.B.s, thereby allowing oxygen and air flow rates to be balanced against agitation rates to achieve desired O2 and CO2 mass transfer across all vessel sizes.
Importantly, the CO2:O2 mass transfer ratio of the DHS across vessel sizes, gas flow rates, and mixing speeds remains between 0.3-0.5. This ratio is well suited to balance oxygen demand while maintaining a sufficient level of CO2 stripping to keep dCO2 levels in physiological ranges (30-80 mmHg)(3-4).


3.2 Ultra-high cell density achievable with room to spare
Cell culture was performed in a 50L system . to test system capabilities including the S.U.B., the controller, and the XCell ATF6 Single-Use system to achieve ultra-high cell densities. The reactor was seeded at 0.4E06 cells/mL and grown to target cell densities of 40, 100, 150, and 200E06 cells/mL. Following two days of culture at 200E06 cells/mL, the final objective was to stress test the perfusion system. Thus, the cells were purposely allowed to grow without constraint, reaching >260E06 cells/mL. Perfusion rates were increased incrementally to support cell growth. Glucose was supplemented as needed to maintain 1-3 g/L in culture. The run was finally terminated after 26 days when the ATF unit fouled completely. Viable cell density (VCD) and viability for the culture are shown in Figure 3.



Culture controller setpoints were maintained throughout the run including DO at 30 ± 4%, temperature at 37 ± 0.05oC, and pH at 7 ± 0.20. Other culture parameters were measured offline and used to determine proper media exchange and glucose feed rates to maintain glucose levels between 1-3 g/L while maintaining minimal buildup of lactate and ammonia.


3.3 Robust mixing and gassing offers targeted mass transfer performance
Bench-scale testing was performed separately, showing robustness of this cell line to perform in extreme mixing conditions with tip speeds in excess of 1.5 m/s and power input greater than 400 W/m3. Therefore, agitation was occasionally adjusted between 20-100 W/m3 while observing oxygen flow rates and effects on dCO2 (Figure 4).



The data suggest a correlation among operating parameters such as gas flow rate, mixing speed, and concentration of sparged gasses, dCO2, pH, and achievable cell density. The literature has established empirical and theoretical models for cell culture in stirred tank reactors using oxygen transfer rates (OTRs), cell density, mixing power, gas flow rate, and partial pressures in sparged gasses(2-5). Similarly, dCO2 and pH can be modeled from the culture data to demonstrate the effectiveness of balancing process parameters. These models are plotted in Figures 6 and 7. Both models show excellent correlation among the parameters suggesting culture conditions (gas flow rate, mixing speed, and concentration of sparged gasses) can be dialed specifically to process requirements to achieve desired cell density. This establishes the S.U.B. as an ideal platform for obtaining predictive culture conditions for ideal growth and productivity.


4. Design improvements lead to best-in-class S.U.B.s for high-density perfusion cultures
With recent improvements in media selection, cell screening, and the increased adoption of continuous processing in the biotechnology industry stresses on the S.U.B. have increased and demanded that performance must improve to accommodate these new processes. Enhancements to the legacy S.U.B. have been implemented to achieve peak performance targeting high-demand cultures such as those seen in continuous processing applications. These enhancements, focusing on impeller and DHS configurations and designs, show improved mixing and mass transfer performance across vessel sizes from 50-500 L. Additionally, the improvements allow for easily scaling processes across these vessel sizes while maintaining scale parameters such as PIV and gas flow rate, especially when utilizing the enhanced DHS.

Importantly, the large increases in mass transfer were achieved using the DHS only without the frit as a mass transfer aide. Previous testing has shown the frit is prone to fouling in high density and/or long duration cultures leading to highly variable performance, even significant loss in performance due to changes in bubble formation on the frit surface. However, the DHS, with its benefits of low cell shear stress and high kLa, has shown robust performance under high-demand conditions due to high resistance to cell debris fouling.
The oxygen and CO2 mass transfer data generated through extensive testing show predictability and scalability of these parameters within specific vessel sizes as well as across vessels. These enhancements have shown increased performance three- to four-fold above previous legacy systems while maintaining mixing and sparging within the bounds of recommended reactor settings.
A confirmation cell run was performed to highlight the effect of these mixing and mass transfer enhancements, achieving stable cell densities at various setpoints up to 200E06 cells/mL, while achieving final growth to 260E06 cells/mL. While most processes would never require such high cell densities due to filter limitations and clone stability, the data demonstrate that the S.U.B. is capable of supporting very high demand cultures.

Importantly, near full automation of many key reactor settings was achieved in the culture including automated media delivery based on load cells, antifoam delivery as needed based on the Single-Use foam sensor, cell bleed for targeted cell density based on the integrated cell density probe into the DeltaV controller, and gas control up to 0.2 vessel volume per minute to maintain DO levels at setpoint. This automation allowed for minimal required user input even under demanding conditions of ultra-high cell density, leading to a successful cell run.

Mass transfer models from the cell culture data were generated showing the effect of mixing power and gas flow rates on predicting viable cell density and dCO2 levels during the culture. While each culture, media, cell line, and process will show inherent variability, the data suggest that modeling can help achieve ideal growth conditions and allow the end user to dial in parameters as needed. Further automation could be achieved by integrating other sensors (dCO2, Raman) to monitor and control levels of dCO2, glucose, and lactate by controlling mixing speeds, gas flow rates, inlet gas concentrations, and perfusion rates.

Partager l’article


Ben is Senior engineer in Thermo Fisher Scientific R&D team since more than 6 years. He is working on the integration and evaluation of external technologies into Single-Use bioreactor systems (Perfusion/continuous processing, research into sparging technologies for improved and reliable product) and new product development and Single-Use improvements for product expansion. Before this position, Ben has worked as research assistant in the Utah State University. Ben owns a BS and a PhD in Biological Engineering from the Utah State University.




1. Zhu, Y., Cuenca, J. V., Zhou, W. & Varma, A. NS0 cell damage by high gas velocity sparging in protein-free and cholesterol-free cultures. Biotechnology and Bioengineering 101, 751–760 (2008).
2. Doran, P. M. Mass Transfer. in Bioprocess Engineering Principles 190-217 (Elsevier, 1995).
3. Zhu, M. M. et al. Effects of elevated pCO2 and osmolality on growth of CHO cells and production of antibody-fusion protein B1: a case study. Biotechnology Progress 21, 70–77 (2005).
4. Mostafa, S. S. & Gu, X. Strategies for improved dCO2 removal in large-scale fed-batch cultures. Biotechnology Progress 19, 45–51 (2003).
5. Garcia-Ochoa, F. & Gomez, E. Bioreactor scale-up and oxygen transfer rate in microbial processes: An overview. Biotechnology Advances 27, 153-176 (2009).


BPC: BioProcessing Container or Single-Use assembly DHS: drilled-hole sparger
DO: Dissolved Oxygen
P/V: power input-to-volume
SUB: Single-Use Bioreactor
VVD: Vessel Volumes per Day