Performance of an acoustic settler versus a hollow fiber-based ATF technology for influenza virus production in perfusion

Abstract

Process intensification and integration is crucial regarding an ever increasing pressure on manufacturing costs and capacities in biologics manufacturing. For virus production in perfusion mode, membrane-based alternating tangential flow filtration (ATF) and acoustic settler are the commonly described cell retention technologies. While acoustic settlers allow for continuous influenza virus harvesting, the use of commercially available membranes for ATF systems typically results in the accumulation of virus particles in the bioreactor vessel. Accordingly, with one single harvest at the end of a cultivation, this increases the risk of lowering the product quality. To assess which cell retention device would be most suitable for influenza A virus production, we compared various key performance figures using AGE1.CR.pIX cells at concentrations between 25 and 50 × 10⁶ cells/mL at similar infection conditions using either an ATF system or an acoustic settler. Production yields, process-related impurities, and aggregation of viruses and other large molecules were evaluated. Taking into account the total number of virions from both the bioreactor and the harvest vessel, a 1.5-3.0-fold higher volumetric virus yield was obtained for the acoustic settler. In addition, fewer large-sized aggregates (virus particles and other molecules) were observed in the harvest taken directly from the bioreactor. In contrast, similar levels of process-related impurities (host cell dsDNA, total protein) were obtained in the harvest for both retention systems. Overall, a clear advantage was observed for continuous virus harvesting after the acoustic settler operation mode was optimized. This development may also allow direct integration of subsequent downstream processing steps. KEY POINTS: • High suspension cell density, immortalized avian cell line, influenza vaccine.

Keywords: Cell culture–based; Influenza; Perfusion; Virus.

Fig. 1

Perfusion cell culture setups with AGE1.CR.pIX cells during the influenza virus production phase. The cells were retained in the bioreactor using an alternating tangential flow filtration (A), or an acoustic settler with either a pump-based (B) or a valve-based recirculation strategy (C). The acoustic settler allowed continuous virus harvesting, which was not feasible here with the ATF system due to membrane clogging. Fresh medium was added continuously (green arrow) to feed the cells, while cell-free medium was removed (red arrow) to keep a constant bioreactor working volume

Fig. 2

Growth of AGE1.CR.pIX cells cultivated in perfusion mode using different cell retention technologies and recirculation strategies. a Viable cell concentration (filled symbols) and cell viability (empty symbols) of one representative ATF run (run 1) (black circle), one representative run for the acoustic settler with valve-based recirculation (AcSE valve, run 4) (blue circle), and two representative runs for the acoustic settler with pump-based recirculation (AcSE pump, run 5 (red circle), and run 10 (red triangle)). b Cell population doubling time (t_d) calculated during the cell growth phase in perfusion mode (average between each sampling time point for each run ± standard deviation). The values correspond to run 1 for ATF (black), runs 3 and 4 for the acoustic settler with valve-based recirculation (blue), and runs 5 and 10 for the acoustic settler with pump-based recirculation (red). A CSPR of 0.06 nL/cell/day was applied for every perfusion run. Detailed operation conditions in Table 1

Fig. 3

Influence of acoustic settler operation on viable cell concentration, viability, and lactate metabolism during the influenza A/PR/8/34 virus production phase with AGE1.CR.pIX cells. (a) Viable cell concentration, (b) cell viability, (c) lactate concentration in the bioreactor supernatant. (d) Y_Lac/Glc yield. Run 1 (black circle): performed with the ATF system. Run 3 (blue circle) and run 4 (blue triangle): performed with the valve-based recirculation mode of the acoustic settler. Run 5 (red circle), run 6 (red triangle), and run 7 (red square): performed with the pump-based recirculation mode of the acoustic settler. Detailed operation conditions in Table 1

Fig. 4

Host cell dsDNA and total protein impurity levels during influenza A/PR/8/34 virus production phase in AGE1.CR.pIX cells in perfusion mode using an ATF system (run 1, black) or an acoustic settler with pump-based recirculation (one representative optimized run, run 6, red). a Host cell dsDNA concentration (black circle) and total protein concentration (white circle). Dashed lines represent additional data from the ATF permeate line. b Accumulated dsDNA (black circle), accumulated total protein (white circle), and total number of produced virions over time (black triangle). c Host cell dsDNA per virion (striped columns) and total protein per virion (filled columns) at optimum harvest time point (average ± standard deviation of technical duplicates). For the ATF cultivation, the bioreactor content was harvested at 36 hpi. When using the acoustic settler, virions from the bioreactor were also collected at the optimum harvest time point which corresponded to 45 hpi. Detailed operation conditions in Table 1

Fig. 5

Infectious titer of influenza A/PR/8/34 virus and size distributions during influenza A/PR/8/34 virus production phase in AGE1.CR.pIX cells in perfusion mode using either an ATF system or an acoustic settler with pump-based recirculation. (a) Total number of infectious influenza virus particles produced using an ATF system (run 1, black circle) or an acoustic settler (one representative optimized run, run 6, red circle). Size distributions of run 1 (ATF, b) and run 6 (acoustic settler, c). All samples were measured from the crude bioreactor supernatant. For graph b, black, blue, and red lines correspond to 24, 36, and 47 hpi, respectively. For graph c, black, blue, and red lines correspond to 25, 33, and 45 hpi, respectively. Detailed operation conditions in Table 1