In this blog post, we discuss how to develop validation studies for the virus removal step in drug manufacture.
Establishing virus validation process parameters
Validation of virus removal is essential to ensuring safe biologic drug products. To develop the most effective virus reduction protocols, including virus filtration, clearance studies should be performed under simulated process-specific conditions. This is done using a downscale model that accurately reflects the commercial process.
The viruses used in viral clearance studies are selected after conducting a risk assessment of potential contamination sources. If known or suspected viruses cannot be used in virus clearance studies, for instance because they cannot be grown to sufficiently high titers, then specific model viruses with similar physicochemical properties must be identified. To verify the robustness of the viral filtration process, non-specific model viruses are also used in order to characterize the capacity of the manufacturing process to remove/inactivate a range of viruses with different properties, including unknown viruses.
The virus filtration process design space is validated after assessing the impact of a range of possible process and product parameters on virus filtration performance. The optimum approach is to use bracketing studies to establish the window of process conditions where consistent virus removal can be achieved. As well as determining the maximum allowable volume throughput and flow decay, studies can also confirm the impact of process interruptions. Historical data collected in similar process studies are often considered in addition to newly collected results when using the bracketing approach, which is well-accepted by global regulatory authorities.
Connection between process and validation targets
Virus filtration validation studies involve passing process fluid intentionally spiked with potential adventitious viruses through the virus filter using a downscale model of the commercial process. This approach is necessary but challenging. Under process conditions, the viral load in a potentially contaminated batch is very low because virus filtration is one of the last downstream operations. Several other viral clearance steps have already lowered the initial contamination level by multiple orders of magnitude. The preceding purification processes also remove any impurities associated with the virus.
Spiked solutions in the lab, however, contain higher levels of the test viruses and therefore high concentrations of any contaminants present in the virus stock preparations. The addition of a virus spike consequently has the potential to cause premature filter fouling, which can result in a throughput lower than that observed during process development and can often be a risk factor for virus breakthrough.
To avoid these issues, the virus stock should be highly purified. In addition, the target virus removal level that must be demonstrated should be well understood so that the virus spike contains only as much virus as is needed to demonstrate the desired level of removal, keeping safety margins. With this approach, the likelihood of the downscale model operating as close as possible to the targets established during process development is much greater.
Factors to consider when developing virus filter validation studies
Purification of the virus stock is important because impurities such as DNA and protein aggregates could potentially block the filter or interact with the product solution. In most cases, standard ultracentrifugation can purify the spike sufficiently to prevent flux decay problems. Where additional purification is required, many virus validation laboratories offer ultra-pure versions of their virus stocks. Since most samples used in these studies have been frozen, the impact of freezing and thawing of the process solution must also be understood to further avoid premature filter fouling.
The decision tree shown in our application note on Validating Pegasus™ Prime Virus Membrane filters, can be used for effective spike selection. Following this chart will help to determine the minimum required spike and ensure that the filterability of the scaled down study is not impacted by this spike. The use of large-volume assays is highly recommended to allow for increased assay sensitivity and calculation of a higher log reduction value at a given spike ratio.
With this approach, a strong downscale model can be established that enables validation of the full process throughput and ensures a trouble-free regulatory submission.