Virus spiking strategies
As discussed in ICH Q5A, to characterize the capacity of the unit
operation to remove and/or inactivate viruses, the model virus should
provide useful information about the performance of the unit operation.
Of the viruses suggested in ICH Q5A, the parvovirus MMV represents a
worst-case virus spike for small virus retentive filter validation
studies. Thus, it is often used as a relevant model virus for scale-down
viral filtration studies, along with other parvoviruses, such as PPV
which can also be grown to high titers. It has been demonstrated that
canine parvovirus (CPV) can pass through small virus retentive filters
and can be construed as a worst case model
(Nowak, Popp et al. 2019).However, the
relevance of CPV to the manufacturing process should be taken into
consideration, because, unlike MMV, CPV has not been a real-world
contaminant in biomanufacturing.
When designing viral spiking strategies for viral filtration studies,
several things should be considered including viral titer or purity of
viral spike material. Studies have shown that some virus filters have a
finite capacity to retain small viruses
(Lute, Bailey et al. 2007). In these
cases, an excessive virus spike e.g. 1014 viral
particles per/m2 would not be representative of
typical manufacturing conditions, may lead to lower LRVs
(Lute, Riordan et al. 2008), and should
be avoided (Chen and Chen 2015). Newer
generation filters may be more robust to higher spikes
(Mattila, Clark et al. 2016,
Strauss, Goldstein et al. 2017). High
purity and low aggregate virus stocks are desirable to reduce the risk
of filter fouling caused by virus spike
(Khan, Parrella et al. 2009,
Chen 2014).
Infectivity assays such as median tissue culture infective dose (TCID50)
or plaque assays are the preferred assays to quantify virus clearance
LRVs. Large volume testing is a common method to increase assay
sensitivity and claimable LRVs. qPCR assays are rarely used currently
due to the detection of residual free viral nucleic acids which may or
may not be active viral particles. Detection of free nucleic acid that
can pass through the virus filter resulting in artificially lower LRVs
unless the buffer pH values pose the potential for inactivation.
For further information about viral spikes, the PDA Technical Report
#47 may be consulted(PDA 2010).