Ask the Expert: Predicting Viral Clearance During Downstream Development

Until recently, downstream process development (PD) teams have lacked methods for easy, effective, and economical estimation of a process’s viral clearance capability. David Cetlin (senior director of R&D at Cygnus Technologies) delivered an “Ask the Expert” presentation on 13 October 2020 describing how MockV kits containing noninfectious mock-virus particles (MVPs) could fill that gap.

Cetlin’s Presentation
Viral clearance studies tend to be outsourced to contract research organizations (CROs) because they require biosafety level (BSL)-2 and -3 facilities for working with infectious agents. But high costs and challenging logistics associated with CROs can push clearance assessment late into process characterization or just before validation.
MockV kits containing noninfectious MVPs enable benchtop estimation of viral clearance performance. Kits are available for model viruses such as minute virus of mice (MVM). A spiking study using MVPs is performed similarly to that for live virus.

Analysts spike a sample with a known quantity of MVPs, perform downstream processing, then quantify MVPs in the eluate to derive a log reduction value (LRV). To provide the requisite level of assay sensitivity, the MockV MVM kit provides components to perform immuno-quantitative polymerase chain reaction (immuno-qPCR) analysis.

Effective Mimicry: Cetlin explained that MVM surrogates are designed to express the major MVM capsid protein (VP2), which facilitates imitation of the virus’s physicochemical properties. In collaboration with the US Food and Drug Administration (FDA), a MockV team compared the surface charge and hydrophobicity of MVM, MVM-MVPs, and the similarly sized bacteriophage PP7. The particles registered isoelectric points (pI) of 5.99, 5.81, and 4.74, respectively; hydrophobicity values came to 0.28, 0.35, and 0.61. Thus, whereas phage PP7 resembled MVM in size only, MVM-MVPs also approximated its charge and surface chemistry.

Predictive Value: Studies performed with RegenXBio, Texcell, and Thermo Fisher Scientific evaluated the MVM kit’s ability to estimate the viral clearance of an adenoassociated virus (AAV) process purification step. Teams ran 5-mL columns prepacked with AAVX affinity resin (Thermo Fisher Scientific) under center-point (two runs) and worst-case (one run) conditions. The former required spiking ~150 mL of AAVX load to starting concentrations of 10 log10 MVPs/mL or ~8.0 TCID50/mL of live virus. Worst-case conditions used a higher load volume (200 mL) and a longer residence time.

Eluate from MVP runs at center-point conditions showed LRVs of 4.91 and 5.16, with most surrogates clearing during flow-through (52.6% and 67.0%) and a benzonase wash (10.4% during both runs). Worst-case conditions yielded 4.07 LRV with 79.1% of MVPs clearing during flow-through. Evaluation of live MVM clearance yielded similar results: 4.35 ± 0.38 LRV at center-point and 3.58 ± 0.46 LRV in worst-case conditions. In both cases, ~54% of MVM particles cleared during flow-through.

MockV MVM kits are equally effective for predicting clearance in anion-exchange chromatography (AEX) steps. The MockV program partnered with GlaxoSmithKline to develop a design of experiments (DoE) evaluation of how pH and conductivity influence AEX retention of MVM and MVPs. Cetlin highlighted that columns run at pH 7.0 and 3.0 mS/cm conductivity exhibited ~4-log clearance of MVM and MVPs. Increasing conductivity to 13.0 mS/cm resulted in ~2-log decreases in clearance for both types of particles. These findings not only corroborated the predictive value of MockV MVM kits, but also enabled GSK scientists to map optimal conditions for MVP application.

High Throughput: Cetlin reported that Cygnus has worked with the US National Institutes of Health Vaccine Research Center (NIH-VRC) to develop a high-throughput method for measuring >4-log clearance of MVM-MVPs. The team evaluated MVP retention in eight AEX resins at three pH levels. Higher levels of conductivity buffer were added incrementally. The team later performed those same tests on eight cation exchangers. Results are forthcoming, but Cetlin said that the study has enabled the NIH-VRC to improve its selection of resins for PD, optimization, and scale-up.

Cygnus MockV MVM kits enable PD teams to perform their own spiking studies. A standard kit provides enough material for 10 experiments (150-mL loads spiked to concentrations of 10⁹ particles/mL). With this kit, a user can assess resins and techniques easily and cost-effectively before performing more extensive work with a CRO.

Questions and Answers
How closely do MVP LRVs trend with MVM LRVs? MVP LRVs come within
1 log of MVM LRVs in ~90% of cases. MVP values come within 0.5 log roughly half of the time.

How do regulatory agencies regard data from MVP studies? Currently, studies using MVPs cannot substitute for live-virus studies required for applications to regulators. However, Cygnus hopes that agencies soon will consider MVP data as important supplements to results from live-viral clearance studies.

You can watch the full presentation in our online webcast archive.