Equipment
The virus removal filtration equipment consisted of an electronic
recorder, two pairs of redundant pressure sensors (placed up- and
down-stream of the pre-filter) and temperature sensors as generally
described elsewhere (Wieser, (2015)).
A pressure vessel made of electropolished 316L steel with a capacity of
5 Liters that could be pressurized to 7 bar (Millipore Cat. No.
XX6700P05) was used as a feed tank.
Pressure was applied to the tank via a manual controller for nitrogen
gas pressure. The nitrogen gas was passed over a 0.2µm air filter in
order to avoid introduction of bacterial contaminants.
The filtrate was collected in vessels placed on an electronic balance.
The data collected by electronic recorder was used to calculate process
feed loads, feed flow rates, process duration and duration of pressure
interruptions.
The virus removal filters used were:
- 1st generation filters:
- Asahi Planova (Asahi Kasei Medical Co., Ltd., Tokyo, Japan) 15N
filters; 10 cm², rating: mean pore size 15 ± 2 nm (Asahi KASEI,
(2018))
- Millipore (Merck KGaA, Darmstadt, Germany) NFP; 3.5 cm², rating:
>4 log removal of small viruses (Merck Millipore,
(2012))
- Pall DV20 (Pall Corp., Port Washington, USA); 10 cm², rating:
> 3 log for 20 nm viruses (Pall Life Sciences, (2010))
- Pall Pegasus SV4; 9.6 cm², rating: 20 nm (nominal) (Pall Life
Sciences, (2012))
- 2nd generation filters
- Asahi Planova BioEX filters; 3 cm², rating: designed to deliver
> 4.0 logs of parvovirus removal (Asahi KASEI, (2018))
- Millipore Barrier; 3.1 cm², rating: ≥ 3.0 log removal of parvovirus
(Merck Millipore, (2021))
- Pall Pegasus Prime; 2.8 cm², rating: 20 nm (nominal) (Pall Life
Sciences, (2016))
- Sartorius CPV; 5 cm², rating: 20 nm (nominal) (Sartorius Stedim
Biotech GmbH, (2011))
- Sartorius HF; 1.7 cm², rating: 20 nm (nominal) (Sartorius Stedim
Biotech GmbH, (2014))
Each set-up included a Millipore Durapore pre-filter (rating: 0.1 µm
(nominal), 13 cm²) set-up in-line in a dead-end filtration
configuration. The 0.1 µm pre-filter was flushed and conditioned with
water for injection before it was autoclaved at 100°C for 30 minutes.
Then it was re-introduced into the filtration equipment and conditioned
with WFI again (together with the virus removal filter).
0.1µm Pall Acrodisc syringe filters were used to remove virus aggregates
from the thawed virus stocks.
The virus-spiked fermenter feed media was passed over a 0.2 µm Pall
Vacucap 90PF filter as needed in order to avoid the introduction of
bacterial contaminants.
The pre-use suitability of each viral filter used was verified according
to the vendors instructions:
Asahi filters were flushed with WFI, subjected to a pre-use air
leakage test and then filled with WFI again. Only filters passing this
pre-use test were used for individual experiments – none of the
filters tested failed.
All other filters were subjected to normalized water flow tests during
which water was passed through the filter at specified pressures and
the flow rate measured. Only filters with water flow rates within the
ranges specified by the manufacturer were used for individual
experiments – none of the filters failed.
After each experiment the integrity of each viral filter used was
verified according to the vendors instructions. For the Asahi filters
this was identical to the pre-use leakage test. For all other filters
the test was a pressure hold test. Only filters for which no air was
seen exiting the filter at the specified testing pressure within a given
time period were seen to be integral.