Single-Use Devices for Alternating Tangential-Flow Filtration: Highly Efficient Perfusion Processes Over 50 Days (Webcast Recap)

Perfusion is being explored as a viable option for upstream-process intensification. Research on such intensification concerns increased productivity and decreased production costs and time; therefore, perfusion, as a continuous addition of fresh media and removal of used media, could reduce the burden on upstream processes. However, scalability remains challenging. Geometric similarity between laboratory-scale and production-scale bioreactors often is considered. In fully turbulent conditions, bioreactor shape, volumetric power input (P/V), fluid profile, and power-input distribution should be consistent at all scales. In an April 2024 Ask the Expert webinar, Jan Ott (research associate at the Zurich University of Applied Sciences (ZHAW) School of Life Sciences and Facility Management) presented with support from Repligen about scaling perfusion processes based on novel bioreactor technologies, such as single-use vessels.

Ott’s Presentation

The heart of each perfusion process is the cell-retention device. Such devices use methods such as tangential-flow filtration (TFF) or alternating tangential-flow (ATF) filtration to separate the cells from the cultivation medium. ATF is the most commonly used method in biomanufacturing because of its low residence time, less membrane fouling, and ease of use compared with TFF.

ZHAW explored scalability of a 50-day perfusion process from 2-L glass reactors to 50-L single-use reactors. The process involved culture of Chinese hamster ovary (CHO) cells engineered to express trastuzumab. The ZHAW team installed XCell ATF 2 and XCell

ATF 6 single-use devices for cell retention in the 2-L and 50-L bioreactors, respectively.

First, the team had to determine scaling criteria between a cylindrical 2-L bioreactor and a cube-shaped 50-L bioreactor. A fixed P/V of 40 W/m³ was chosen for the pilot-scale cultivation. A volumetric oxygen transfer coefficient (k_La) was used as a scaling criterion to accommodate targeted cell densities of >100 million cells/mL. ATF flow also was chosen as a scaling criterion to reach the same shear rate in both the ATF2 and ATF6 devices. Flow rates were set at 17 L/min for the ATF6 cultivation and 0.9 L/min for the ATF2 cultivation. A perfusion rate of 1 vessel volume per day (VVD) was run at both scales.

The team achieved a viable cell volume (VCV) of >100 mm³/mL after seven days for both cultivations, as expected. Trastuzumab production reached 1.5–2 g/L throughout most of the experiment’s 50-day duration. Product retention through ATF increased until the end of cultivation, resulting in ~40% retention. Viability remained >90% throughout testing. The team was able to harvest >1.2 g/L of trastuzumab at nearly any point of the experiment. That harvest resulted in an average volumetric productivity of >1 g/L/day over the production phase. Productivity potentially could have increased to >1.2 g/L/day by recycling bleed to obtain antibodies that otherwise would not be used for downstream processing. All cultivations had average production rates of 10–12.5 μg/mm³/day.

The team measured product quality attributes of the harvested cultivations. For all processes, antibodies harvested between days 20–50 had stable glycosylation profiles — e.g., 60–70% of antibodies bearing a G0F glycan and ~20% bearing a G1F glycan. The two cultivations had slightly different charge-variant profiles: The 2-L process had a slightly higher acidic peak ratio, whereas the 50-L cultivations had a slightly higher neutral peak ratio.

Questions and Answers

How did you choose between ATF and TFF? ATF backflush enables longer cultivation. We would not have been able to reach the same retention after 50 days with TFF.

Did you use custom or off-the-shelf culture media? We used Gibco’s commercially available High-Intensity Perfusion CHO media.

Did you have issues using a mass balance to control the harvest pump due to fluctuating culture levels related to ATF backflow? No, because we used the TruBio software (Thermo Fisher Scientific), which enables us to place a time-based filter over the process value so that the bioreactor’s weight value is averaged over one minute, for example. That average value is used to control the pump.

How did you decide on the perfusion rate for your experiment, and how did that affect cell health and yield? We have a cell line that can exceed a VCV of 100 mm³/mL at a rate of only 1 VVD. It might be possible to increase that perfusion rate and overall yields, but oxygen input rates should be considered before establishing higher values. Cell health was not affected by the rates that we used. The 1-VVD yield was higher compared with yields from other published cultivations.

For more information, view ZHAW’s research article published in Processes at https://doi.org/10.3390/pr12040806.

Find the full webinar online at www.bioprocessintl.com/category/webinars.