Filtration:Vendor Notes: The Effect of Flux Decay on a 20-nm Nanofilter for Virus Retention

ABSTRACT

Viral clearance steps are essential for maintaining the safety and integrity of biopharmaceutical products. ICH Q5A mandates that the manufacturing process remove or inactivate contaminants based on a process-specific virus clearance strategy. A 20-nm retentive virus removal filter can clear both large and small viruses, but a virus spiking trial is needed to validate the effectiveness of such a step. Virus retention studies were run with three lots of Virosart CPV, a 20-nm polyethersulfone virus filter, over a flow decay range of 0 to 90%. The model virus used was bacteriophage PP7 using Pseudomonas aeruginosa as the target and indicator cell. Four different protein solutions were spiked with PP7 and tested in triplicate runs. The retention goal of 4 log₁₀ was met and exceeded over the entire flow decay profile.

Innovative technologies in biopharmaceutical processing have resulted in greater production capabilities, and at the same time, growing concern over the viral safety of the products. Virus contamination of products derived from human or animal cell lines can have disastrous clinical consequences. Although there have been no reported infections or transmissions of Chinese hamster ovary (CHO) cell-related type A and C virus particles to date, viral clearance steps are vital for ensuring the safety and integrity of biopharmaceutical products.

The ICH Q5A regulatory guideline mandates that therapeutic biological product manufacturers implement technologies into their manufacturing process that remove or inactivate known or unknown contaminants based on a process-specific virus clearance strategy.¹ Such viral clearance steps must be validated and must also demonstrate that the implemented technologies effectively remove a range of known and unpredictable viruses.² Due to these stringent regulatory demands, bioprocess companies have struggled to find innovative methods to effectively eliminate and inactivate viruses in the bio-feedstream.

Virus filtration is an established and robust method for effectively reducing a range of viruses within a single stage of the downstream purification process.² As a part of the purification process of a biopharmaceutical, 20-nm retentive virus removal filters can clear both large and small viruses. A virus spiking trial must be performed to validate the efficiency of these 20-nm filters.³ This article describes trials with a particular 20-nm filter and four different protein solutions.

PROPERTIES OF A VIRUS FILTER

The ideal virus filter should retain all viruses and allow high protein transmission while maintaining a high flow rate without significant virus breakthrough. However, virus breakthrough seems to be a general phenomenon among current virus filters and it has been suggested that pore plugging causes a decline in the flow rate: "With virus breakthrough a general phenomenon for virus removal filters, one could argue that fouling by contaminants in virus spikes should be minimized so that conditions in filter validation most closely represent those in the manufacturing process."⁴ Others have shown that the virus reduction capability of some virus removal filters decreases with increasing flux decay.^5,6

Although contaminants and other various parameters may be main causes of filter breakdown, some nanofilters still efficiently remove viruses at high Log Reduction Value (LRV) even when experiencing high flow decay. A recent paper outlined the various titer reduction capabilities of virus retentive nanofilters.⁷ The data showed that not all filters tested for their LRV versus flow decay profile experienced a significant loss of titer reduction with increasing flow decay.

The existing virus filtration technologies available on the market have similarities and differences with respect to the physical parameters that affect small virus retention:

• pleated versus hollow fiber configuration

• pore size and pore size distribution

• pore density

• membrane symmetry

• membrane thickness

• number of membrane layers

• membrane material used.

Care should be taken with general statements about virus filter performance and LRV predictions. The filter configuration is not the only factor that influences the overall virus reduction capability. A potential decrease in virus reduction may depend on several factors, including protein concentration, buffer composition, product purity, and plugging mechanism (adsorptive versus pore plugging). The influence of these parameters on the overall flow decay profile of each virus filter has to be examined case by case.

Flow Decay Study Protocol

This article presents virus retention data for Virosart CPV, a 20-nm polyethersulfone (PESU) nanofilter from Sartorius AG (Gottingen, Germany). Different protein types and concentrations have been used to determine LRV versus flux decay profiles. Spiking studies were performed using a laboratory scale model of Virosart CPV (Virosart CPV Minisarts with an effective filter area of 5cm²).

Model Virus Used for the Nanofiltration Study

The model virus used for the assay was bacteriophage PP7, a small 25-nm, nonenveloped ss-RNA Pseudomonas phage from the Leviviridae family. Leviviridae have a simple construction and are not tailed. They consist of a capsid and are not enveloped. The capsid appears round, is 25 nm in diameter, and has an icosahedral symmetry. The genome consists of linear, nonsegmented, positive sense ssRNA of about 3.4-4.3 kb. Nucleic acid takes up 30% of the total phage weight. PP7 is a model virus used by several filter manufacturers for the evaluation of retention characteristics of 20-nm virus filters. The principal feasibility of PP7 as a model virus for mammalian viruses has been demonstrated in a previous study.⁸

The PP7 Assay

A plaque assay used Pseudomonas aeruginosa as the target and indicator bacterial cells to be infected by bacteriophage PP7. P. aeruginosa and the bacteriophage PP7 used were from the American Type Culture Collection, USA (ATCC-No. 15692 and ATCC-No. 15692-B2 respectively).

P. aeruginosa was grown as an overnight culture at 37 °C in nutrient broth supplemented with 0.5% NaCl. PP7 was harvested from suspensions of infected P. aeruginosa by low speed centrifugation (2000 ± 100 rev/min, 10 ± 1 min). The supernatant was pooled and filtered through a 0.2-µm PESU membrane filter, aliquoted, shock-frozen in liquid nitrogen, and stored at –66 ± 5 °C until further use.

To determine the PP7 titer, experimental samples (150-µL aliquots) were incubated with bacteria as a 50-µL aliquot of an overnight culture, diluted 1:100 in nutrient broth (Nutrient Broth Difco 23400, Becton Dickinson) for 10–20 min at room temperature. Subsequently, 3 mL of 0.5% agar (Difco Nutrient Agar, Becton Dickinson) were added. The entire volume was spread on a Petri dish with 1.5% solid nutrient agar. After an incubation period of 20–24 hr at 37 °C the plaques induced by bacteriophages were counted.

Calculation and Statistical Analysis

All experimental samples are assayed for bacteriophages in fourfold replicate (150-µL each) at undiluted (100) and three different dilutions (10¹, 10², and 10³). After an incubation period of 20–24 hr, the number of plaques on a confluent layer of P. aeruginosa were counted. Three dilutions with countable numbers of plaques were used for the calculation of the PP7 titer.

The Log₁₀ reduction factor in the nanofiltrate was calculated using the following equation:

in which R is the final log₁₀ reduction factor, A₀ is the phage titer/mL upstream of test item, and A_n is the phage titer/mL downstream of the test item.

Nanofiltration Study

Flow decay is defined as the decrease of flow during the nanofiltration when compared to the initial buffer flow during flushing of the filter. At 100% buffer flow, flow decay is 0%. As an example, a nanofiltration flow of 50% compared to the initial buffer flow is described as 50% flow decay. During the nanofiltration of the protein solution, instantaneous samples were taken at various levels of flow decays from 5% to around 90% and the LRF was calculated.

Figure 1

Three commercially available lots of Virosart CPV Minisarts were used for each nanofiltration study. Each lot was run in triplicate. Figures 1–4 display all nine data sets per protein solution filtered. For each test, protein type, protein concentration, pH, buffer type, and composition, as well as the spiking concentration, are outlined.

Four proteins were used for this nanofiltration study:

• clinical grade of human-derived intra-venous immune globlulin (IVIG) at 2 mg/mL

• monoclonal antibody at 8 mg/mL

• sheep-derived Fab₍₂₎ at 5 mg/mL

• bovine serum albumin (BSA) at 1 mg/mL.

Figure 2

Results

Figures 1–4 show LRV over a wide flow decay using PP7 as the model virus for small non-enveloped viruses. The Virosart CPV provided reliable virus retention of small viruses.

The goal was >4 log₁₀, which was achieved with little variation across the range. This shows that the capability of Virosart CPV to retain more than 4 log₁₀ of small non-enveloped viruses is not necessarily correlated to the flux decay. Reliable retention of more than 4 log₁₀ is shown for flux decay profiles of up to 90%.

Figure 3

The data shown are typical examples for different protein types, buffers, and concentrations. They cannot, however, be used to predict the nanofilter performance in general. Specific virus clearance studies under good laboratory practices conditions must be carried out for each biopharmaceutical product going into Phase 1 or 3 clinical trials. Finally, it is up to the end user to determine what levels of flow decay and virus retention are acceptable in order to achieve the targeted virus retention.

Figure 4

Upcoming studies

Upcoming studies will feature porcine parvo virus (PPV) and minute virus of mice (MVM) studies, each with co-spikes of bacteriophage PP7 to provide evidence of the capability of bacteriophage PP7 to serve as a model virus for small nonenveloped viruses. Different spike preparation scenarios will be outlined, and technologies are being screened for their capability to purify virus spikes for easier and more standardized nanofiltration spiking studies.

Klaus Tarrach is the senior product manager of purification technologies at Sartorius AG Germany in Gottingen, +49 551. 308. 3959, Klaus.Tarrach@sartorius.com

Anika Meyer is an application specialist in purification technologies for Central Europe at Sartorius AG Germany;

Julia Elena Dathe is a research assistant in the Sartorius virology department; and

Hanni Sun is a biologist in the Sartorius virology department.

REFERENCES

1. International Conference on Harmonization. Q5A. Viral safety evaluation of biotechnology products derived from cell lines of human or animal origin. Geneva, Switzerland; 1998.

2. Schmidt S, Mora, J, Dolan S, Kauling, J. An integrated concept for robust and efficient virus clearance and contaminant removal in biotech process. BioProcess Int. 2005 Sept.; 3(9):26-31.

3. Immelmann A, Kellings K, Stamm O, Tarrach K. Validation and quality procedures for virus and prion removal in biopharmaceuticals. BioProcess Int. 2005 Sept;3(9):38-44.

4. Cabatingan, M. Impact of virus stock quality on virus filter validation. BioProcess Int. 2005 July;3(7):39-43.

5. Bailey M. Normal flow virus filtration—Detection and assessment of endpoint in bioprocessing. PDA viral safety conference; 2005 May; Tampa, FL.

6. Bolton G, et al. Normal flow virus filtration—Detection and assessment of endpoint in bioprocessing. Biotechnol. Appl. Biochem. 2005; 42:133-142.

7. Brorson K. Virus filter validation and performance. Recovery conference; 2006 April; Litchfield, AZ.

8. Tarrach K. Integrative Strategies for viral clearance. 4th annual biological production forum. 2005 April 7; Edinburgh, Scotland.