Biomedical Engineering Reference
In-Depth Information
8.4.2 Unknown Starting Virus Load
Once an acceptable virus clearance level CQA is targeted, the manufacturing process is
evaluated for its capacity to clear viruses. The targeted virus clearance level is compared
to the virus clearance capability to determine if the process provides an adequate safety
margin. Such final determination requires a reasonable estimate of the starting virus
contaminant load. For endogenous viruses, it is possible tomake a reasonable estimate of
starting virus load, as these viruses should be routinely expressed in each fermentation
run. The handling of adventitious viruses is much more subjective: these viruses, by
definition, do not routinely and predictably contaminate the product. Only through
unpredictable circumstances, virus might contaminate a givenmanufacturing run. In this
case, the virus number and identity are not knowable a priori.
8.4.3 Differences Between Model or Surrogate Virus
and Wild-Type Virus
Quantifying process capacity for virus clearance requires scale-down virus clearance
studies. Inasmuch as the process-specific wild-type virus contaminant may be either
unidentified, not able to be cultured, or otherwise unable to be used, common practice is
to test with a panel of surrogate
1
viruses (“relevant” or “model” viruses) that represent
ranges of physical and biochemical attributes. Suggestions on the identity of these
viruses are found in ICH Q5A and are summarized in Table 8.2. Relevant and model
viruses may differ in properties that affect virus clearance. Because there has not been a
systematic comparison of the common relevant andmodel viruses, the differences are not
always well understood. It is well known, however, that virus shape, aggregation state,
and exterior surface charge influence virus removal, in particular for virus filtration [4].
Significant differences in susceptibility to virus clearance may exist even between
genetically related viruses. For example, it has been shown that B19 virus, a relevant
parvovirus for plasma products, is more susceptible to some inactivation methods than
the commonly used model parvovirus MMV (mouse minute virus) [5].
8.4.4 Virus Spike Preparation Quality
Commonly, high-titer, laboratory virus preparations are inoculated into a process
intermediate test article and the spiked fluid is subjected to the virus clearance step.
Issues related to virus spike quality can confound one's ability to reproduce test results
with the accuracy needed to derive reliable data for statistical analysis. Specific virus
spike quality issues include variability invirus stock titers, existence and variability in the
amount of host cell impurities, presence or absence of stabilizing agents such as BSA,
1
ICHQ5Adescribes relevant viruses as those that are either identified as, or are the same species as, viruses that
are known or likely to contaminate the cell substrate or any other reagents or materials used in the production
process. Amodel virus is one that is closely related to the known or suspected virus (same genus or family). For
biotechnology processes, model viruses with differing properties are used to characterize manufacturing
process capacity.
