Proceedings of the 2019 Viral Clearance Symposium, Session 8. Conference Summary: Key Discussion and Outcomes, Pending Questions, and Proposed Experiments and Actions

Key Items and Outcomes

1. Although there has been acceptance of generic claims for low pH inactivation, detergent inactivation, and viral clearance by used protein A resins, there has been so far non-acceptance in Europe of generic claims for anion-exchange (AEX) chromatography and viral filtration.

2. Generic claims for detergent inactivation (Triton CG-110) and low pH inactivation were proposed by Genentech for Biologics License Application (BLA) procedures.

3. A consensus outcome from the session was that at least a parvovirus-only validation for virus filtration should be applicable for marketing applications.

4. As part of the ASTM International Standard initiative, Janssen presented efforts that were made to standardize the small virus filtration step for retrovirus removal. A 5 Log10 retrovirus reduction was proposed for filtration steps performed under predefined conditions for a variety of different molecules.

5. Data were presented showing that the difference between duplicate log reduction value (LRV) measurements was based more on process step performance than assay variability, and therefore use of a worst case between the replicates may not always be justified.

6. For rAAV vectors, viral clearance is desirable but the guidance is not clear. Virus filtration, heat inactivation, detergent inactivation, ion-exchange chromatography, or affinity chromatography may be considered for viral clearance, though the effectiveness of these steps may vary by rAAV serotype.

7. Platform/modular claims may be accepted for more steps in the future, including AEX and viral filtration steps, but should be based on a very strong scientific justification including a detailed risk assessment and scientific results. This is also being discussed as part of the update to ICH Q5A (1).

Pending Questions and Proposed Actions

1. As part of the ASTM initiative, Janssen presented efforts that were made to standardize also the small virus filtration step for retrovirus removal. According to the standard, it is proposed that a 5 Log10 reduction can be claimed if filtration is performed under predefined conditions for a variety of different molecules. However, it was found during discussion that the currently defined conditions are not sufficient to make this claim, and updating and refining of conditions is therefore warranted.

2. Additional discussion on the use of worst case or average LRV values between replicates will be carried out within the industry and as part of the ICH Q5A revision.

3. Additional clarity on expectations and methods of viral clearance for viral vectors would be beneficial.

Key Items Discussed and Outcomes

1. Flow-through AEX or mixed-mode AEX (MMAEX) steps were shown to be potent tools for removal of retroviruses and under defined conditions also for parvoviruses.

2. Depth filtration successfully removes retroviruses; parvovirus removal rather by co-precipitation.

3. Cation-exchange (CEX) bind/elute, hydrophobic interaction chromatography (HIC), and depth filtration can also be potent for virus removal but less robust for parvoviruses

4. MMAEX can provide robust and effective virus removal, but this may depend on the process conditions and impurity levels.

5. Virus removal by AEX flow-through may be product dependent.

6. For CEX bind-elute steps, there is no clear worst case for load density. Loadings that are representative of manufacturing conditions can provide representative measurements of viral clearance.

7. “No salt” HIC shows good virus removal for xenotropic murine leukemia virus (X-MuLV), reovirus-3 (Reo-3), and pseudorabies virus (PRV).

Pending Questions and Proposed Actions

1. Further investigation is still required for a clear understanding of the molecular mechanisms of virus removal and the relevance of potential interactions with resins or viruses.

2. For bind-elute steps like protein A and CEX, there is no consistent guidance or practice of product loading during demonstration of viral clearance. However, use of product loading at the upper end of the acceptable range offers benefits. The viral loading is higher when using high product loading; typically, higher loading is used in manufacturing to improve economics, and typically a loading range or limit is filed with agencies.

Key Items Discussed and Outcomes

1. Further scientific understanding of novel continuous processing technologies and potential alternative strategies for continuous processing viral validation were presented and discussed.

Pending Questions and Proposed Actions

1. The data presented at this session demonstrate that viral validation in a continuous process is achievable.

2. Future work is needed to solidify the robust scientific evidence seen in past batch mode viral validation.

Key Items Discussed and Outcomes

1. Three significant areas were identified for prioritization in an ICH Q5A update: Next-generation sequencing (NGS), resin reuse, and consistency runs.

2. It would be beneficial to provide the flexibility to utilize NGS in lieu of in vivo assays or in vitro co-cultivation.

3. The combination of key outcomes of the EMA Workshop on prior knowledge and prior published scientific knowledge indicates that resin reuse studies are not necessary for protein A especially when performance attributes (e.g., antibody step yield and breakthrough of impurities) are consistent over time/cycles (2).

4. For resin reuse evaluations, the major source of variability is due to the lack of concurrent runs (new and used resins) based on a Paul-Ehrlich-Institut data summary and a similar conclusion from BioPhorum Operations Group assessment for protein A and AEX.

5. There appears to be no significant impact of resin reuse on viral clearance for protein A or AEX, recapitulating and building on some of the outcomes of previous discussions on this topic, indicating that clearance studies with used protein A resins are not necessary.

6. Detergent treatment is generally applicable to a broad variety of molecules.

7. Low pH inactivation is consistent at the usual protein concentrations and consistent across formats, with acetate buffers observed to be somewhat more robust than citrate.

Pending Questions and Proposed Actions

1. A similar approach to the one used for protein A could be employed to obviate the need to perform resin reuse studies for non-product binding AEX chromatography (AEX Flowthrough), though it is challenging to demonstrate a failure scenario within typical operating conditions. Surrogate indicators (height equivalent to the theoretical plate, impurity breakthrough, etc.) could be used to ensure column performance.

2. Viral surrogates (e.g., noninfectious minute virus of mice-mock virus particle [MMV-MVP]) matched the general trends in LRV, though some differences were observed in quantification. Additional investments in the detection assay (quantitative polymerase chain reaction) and assessing means of stabilization of the prep (sensitivity to freeze/thaw was observed) to increase sensitivity and robustness are required to overcome the current limitations of the approach.

3. The potential for support of modular claim for AEX flow-through leveraging data sets was presented by several researchers at the symposium. Potential parameters that would be conserved (aka generic operating conditions) are conductivity, pH, and loading. An understanding of the effects of buffer compositions, impurity levels, and protein–virus interactions would be beneficial to advance these efforts.

4. The use of Bayesian statistics to develop a modular claim is a novel approach to assessing viral clearance and warrants additional investigation.

5. One novel approach discussed during the session was based on the concept of combining CEX and detergent inactivation. It would be important to demonstrate a uniform distribution of detergent on the column and to leverage existing tools for column assessments.

6. The outcome of the discussion on the development of a potential ASTM International Standard covering retrovirus clearance by viral filtration or flow-through AEX was the need to focus on sharing of data sets across organizations to support the development of modular claims.

7. The potential impact of the molecule format on viral filtration is complicated by varying freeze/thaw impacts on filter fouling.

Key Items Discussed and Outcomes

1. Alternative eco-friendly and cost-effective detergents, including sugar-based detergents and Triton X-100 analogs, were demonstrated to provide robust viral inactivation in the biologics manufacturing processes. These will help meet the requirements of European REACH (registration, evaluation, and authorization of chemicals) regulations.

2. Consistent pH measurement and an understanding of pH meter/probe differences between lab-scale, contract research organizations (CROs) for virus spike study, and manufacturing is critical to ensure scale-down models are truly representative of manufacturing.

3. A bracketed approach may not be applicable at pH > 3.70. Instead, a pH target of 3.65 can balance effective viral inactivation with desired product stability, suitable for acid labile biotherapeutics.

4. For a solvent/detergent (S/D) viral inactivation step used in a recombinant enzyme process, pre-mixed S/D stock addition reduced cloudiness. In addition, poloxamer, as a shear protectant used in the cell culture, did not impact the effectiveness of the S/D viral inactivation.

5. On-column viral inactivation using a detergent-containing wash buffer can improve product stability and eliminate a unit operation. Three approaches to viral clearance validation were suggested: delta determination, simplified batch mode, and loaded resin batch mode. The loaded resin batch mode approach is generally recommended as it allows for claiming the full removal/inactivation capacity of both the chromatography step and the detergent-mediated activation, and a broader acceptance by the authorities could be expected.

Pending Questions and Proposed Actions

1. Further exploration of novel detergents, on-column inactivation, and development of standardization of pH measurement techniques would be beneficial.

Key Items Discussed and Outcomes

1. A comparison between two different CROs revealed differences in viral clearance provided by a filter, possibly due to differences in the quality of the viral preparations.

2. The Viresolve Pro (Merck) was studied using MMV and a range of pressures and pauses reflective of potential manufacturing conditions. Complete or nearly complete virus removal was achieved for all conditions, indicating that the step robustly provides viral clearance.

3. Product-related characteristics and storage method can impact filtration during virus clearance studies. This can be improved using an adsorptive prefilter. If prefiltration is done before virus spiking, the improvement in throughput may be reduced. If virus is loaded on the prefilter, any removal should be quantitated and assessed for mechanism versus other viral removal steps before asserting a clearance claim.

4. The capacity of a modified polyether sulfone (PES) viral filtration step was reduced by the addition of a hydrophobic interaction chromatography column despite low aggregate levels. The process was revised with a cellulose viral filter, transferred to the contract manufacturing organization, and completed successfully.

5. High volumetric loading, as opposed to protein mass loading, is considered a worst-case challenge for demonstrating the viral clearance of the filter step. High volumetric loading will also provide a high viral challenge.

6. Previous-generation viral filters suffered a decline in viral retention at lower pressures or with flow pauses. Next-generation viral filters appear to be robust to variations in pressure and extended flow pauses.

7. Despite the robustness of newer viral filters, it is considered best practice to include high pressure, low pressure, a flow pause, and a buffer chase during each scale-down run demonstrating viral clearance of the filter. These data can be used to support the assertion of viral retention for the ranges of pressure and flow pauses that might occur during routine manufacturing.

8. To support process deviations, reduced-scale studies with Planova 20N filters were tested with extended pauses and depressurization and found to robustly provide MMV retention. The data can be used to support deviations and reduce reprocessing requirements.

Pending Questions and Proposed Actions

1. The impact of virus spike quality (size and percentage of noninfectious virus) should be evaluated before selecting a CRO.

2. A comprehensive review of viral filter retention data across companies may be used in the future to support a generic viral clearance claim as has been done for the low pH viral inactivation step.

Key Items Discussed and Outcomes

Proof of concept using NGS has shown that it can be as effective at detecting MMV as traditional cell culture methods.

1. Upstream media filters are very effective at removing parvoviruses from media and feeds and can be a good alternative to high-temperature short time.

2. Testing of Mycobacterium species has been requested by the Chinese health authority, but the risk of mycobacterial contamination is minimal for well-characterized cell banks, and robust downstream processes are in place to remove any potential adventitious contaminants such as viruses and bacteria; therefore, it is not needed.

Pending Questions and Proposed Actions

1. Viral metagenomics can serve as an alternative method to replace in vivo or in vitro methods although more data and understanding of the technology should be developed.