Send to printer »

Tutorials : Sep 15, 2013 (Vol. 33, No. 16)

Selecting the Right Spike for Virus Clearance

Data Supports the Use of Highly Purified Viral Preparations
  • Damon Asher, Ph.D.
  • ,
  • Ushma Mehta
  • ,
  • Annie Leahy
  • ,
  • Patricia Greenhalgh, Ph.D.

Monoclonal antibody (mAb) manufacturing processes typically contain several steps that are evaluated for their ability to remove or inactivate viruses that may be present in the process material.

The availability of cleaner virus stocks for virus clearance studies has raised the question of what level of virus purity is best suited for virus clearance evaluations of various unit operations. One approach is to use minimally purified virus spikes for testing unit operations that come early in a drug manufacturing process and spikes of higher purity for testing of later steps.

An alternative technique would be to use highly purified stocks for all tests because less purified virus stocks have the potential to distort performance of any unit operation, thus making the scaled-down model less representative of the manufacturing process.

A study was conducted to investigate whether highly purified spikes are appropriate for use in evaluations of several different types of virus clearance operations.

Parvovirus and retrovirus stocks of two different purities were produced and used as virus spikes for low pH virus inactivation, anion exchange (AEX) flow-through chromatography, and virus filtration operations. The effects on virus reduction as well as unit operation performance were measured.

Results showed that the less purified virus spikes adversely affected the performance of the scaled-down models, but levels of virus reduction across the unit operations were not influenced by spike purity. This suggests that highly purified spikes are well suited for use in evaluation of various types of virus clearance operations.

Materials

Minute Virus of Mice (MVM; parvovirus, nonenveloped, ~20 nm diameter) and Xenotropic Murine Leukemia Virus (X-MuLV; retrovirus, enveloped, ~100 nm diameter) stocks were prepared at two levels of purity:

Crude: Virus grown in serum-containing media and clarified.

Ultrapure: Virus harvested under serum-free conditions, then purified by ultrafiltration, ultracentrifugation, and flow-through chromatography steps.

Results

Ultrapure virus stocks have increased purity as assessed by lower DNA and protein levels than the crude stocks. Crude stocks show high levels of stained protein in contrast to ultrapure preps at equivalent loading. The ultrapure purification process also concentrated the virus, yielding higher titer stocks.

Purified monoclonal antibody feed with a protein concentration of 6 g/L (in 20 mM acetic acid buffer; pH 5.8) was spiked with crude and ultrapure X-MuLV of two different strains. The pH was then lowered to 3.8 at 20°C and held for 0–30 minutes. Results showed that the kinetics of retrovirus inactivation were unaffected by the purity level of the virus spike, but use of higher titer ultrapure virus resulted in greater calculated log reduction value (LRV) (data not shown).

Virus clearance studies were performed using ChromaSorb™ anion exchange devices (EMD Millipore). Use of crude preparations of both MVM and X-MuLV resulted in increased breakthrough of host-cell protein (HCP) compared to unspiked material.

In contrast, HCP capacity was not affected by the ultrapure spikes; the HCP breakthrough curve was almost identical to that of unspiked material (Figures 1 and 2).

Despite the differences in HCP retention, the levels of virus retention were similar in tests with crude and ultrapure spikes at the loadings examined (Table).

The purity of the virus spike preparation had a notable impact on the performance of the filtration operation using Viresolve Pro Micro devices (Figure 3).

The use of a crude virus spike caused a notable decrease in flux, which was not observed when ultrapure virus spike was used. The use of ultrapure virus not only enabled higher throughput but also higher virus spike levels resulting in greater measured LRV across the virus filtration device. LRV measured for the crude virus spike was ≥5.2; LRV measured for the ultrapure virus was ≥ 6.5.

Conclusions

The LRVs obtained for low pH inactivation, flow-through anion exchange chromatography, and virus filtration unit operations were equivalent regardless of virus purity level. However, the loading capacities of the AEX and filtration models were reduced when the less purified virus spikes were used.

In the case of filtration, these results could lead to oversizing of manufacturing-scale devices to compensate for loss of capacity observed in the clearance study, even though the loss is caused by spike impurities that would not be present in the manufacturing process.

These data support the use of highly purified virus spikes for virus clearance studies of all unit operations. This single-stock strategy would facilitate comparison of results across studies of different clearance steps.

Furthermore, exclusive use of high-titer, ultrapure virus could provide more rigorous clearance testing by enabling spiking to higher levels without impacting performance of unit operations.