Virus reduction

Hygiene and Regulation Almost unique to MF is the influence of regulatory concerns in selection and implementation of a suitable microfilter. Since MF is heavily involved with industries regulated by the Food and Drug Administration, concerns about process stability consistency of manufacture, virus reduction, pathogen control, and material safety loom far larger than is usually found in other membrane separations. 

The virus reduction factor of an individual purification or removal—inactivation step is defined as the log10 of the ratio of the virus load in the pre-purification material divided by the virus load in the post-purification material. A clearance factor for each stage can be calculated and the overall clearance capacity of the production process assessed. Total virus reduction is calculated as the sum of individual log reduction factors. Individual manufacturing steps must possess fundamentally different mechanisms of virus removal or inactivation in order for values to be considered cumulative. Additionally, because viruses vary greatly with regard to inactivation or removal profiles, only data for the same model virus can be cumulative. 

Plasma-derived therapeutic proteins are parenteral biologies that are purified on an industrial scale. All biologies derived from human sources, such as plasma, carry the risk of viral contamination. Thus, in order to market a medicinal product derived from human plasma, manufacturers must assure the absence of specific viral contamination. Virus validation studies are performed to evaluate the capacity of a manufacturing process to remove viral contaminants. Virus clearance across three different terminal inactivation steps, low pH incubation of immunoglobulins (IgG), pasteurization of albumin, and freeze dry/dry heat treatment of plasma-derived products (Factor VIII and Protein G), is discussed in this article. The data show that, like all other upstream virus reduction steps, the methods used for terminal inactivation are process and product dependent, and that the reduction factors for an individual step may be overestimated or underestimated due to inherent limitations or inadequate designs of viral validation studies. 

The overall virus reduction capacity of a manufacturing process is the sum of the individual reduction factors, and should be greater than the potential virus load in the starting material. At least one step in the process should clear significant levels of infectious virus so that the overall clearance is not made up of individual small, and possibly negligible, reductions. 

Virus reduction steps may be placed upstream or at the very end of the manufacturing process. 

The virus reduction studies of the three process steps discussed here were performed with HFV-l, Bovine viral diarrhea virus (BVDV), Pseudorabies virus (PRV), Reovirus type 3 (Reo), Hepatitis A virus (HAV), and Porcine parvovirus (PPV). HIV-1 was included as a relevant enveloped virus, while BVDV and PRV were tested as specific model viruses for HCV and HBV, respectively (Table 1). Reo was chosen as a non-specific model non-enveloped virus, HAV was included as a relevant virus and PPV was used as a surrogate for human parvovirus B19. All viruses were propagated using standard cell culture conditions. " The appropriate cell lines were infected, at a low multiplicity of infection, and incubated until 4-1- cytopathic effects were observed. The infected cells were frozen and thawed three times to release virus, centrifuged at low speed to remove cell debris and the clarified supernatants were removed for use as virus spikes. 

To control these experimental differences, BVDV inactivation was monitored, in parallel, during low pH incubation, in TgG solutions 1 and 3, using the same stock of virus (Fig. 3 and Table 2). The data show minor changes in pH and temperature resulted in major differences in virus reduction. For example, BVDV reduction in TgG solution 1, 20° C, was... 

Table 2 Summary of virus reduction during incubation at low pH...

Enveloped virus inactivation was biphasic, as HIV-1, PRV, and BVDV were quickly inactivated within the first 2 hr of pasteurization. Although inactivation was much slower after 2 hr, all enveloped virus infectivity was below detection after 5 hr (Fig. 4). Although not completely inactivated, the levels of non-enveloped viruses, Reo and HAV (strain HM175/18f), decreased significantly, resulting in 5.6 and 4.4 logio reduction, respectively (Fig. 4 and Table 3). PPV was the most resistant to pasteurization, as less than 2 logio virus reduction was observed (Fig. 4 and Table 3). 

In summary, pasteurization is effective for inactivation of many enveloped and non-enveloped viruses but their kinetics of inactivation are different. In the experiments with 25 /o albumin, enveloped viruses were below detection after heating for 5 hr, leaving a wide margin of safety. In contrast, non-enveloped virus reduction was slower and required the entire 10 hr of heating. Temperature, time, protein concentration, and possibly stabilizers were critical parameters for virus inactivation. 

Table 3 Summary of virus reduction during pasteurization of 25% albumin...

The impact of formulation on virus reduction and product recovery (e.g., potency) was evaluated during terminal freeze dry/dry heat treatment studies with an unlicensed Fraction I derived product, that will be designated Protein G. As shown in Fig. 10, the optimum formulation, that achieved >4 logio PPV inactivation and 80% product recovery, was one that contained 2% albumin, no NaCl, and <0.3%o moisture by the Karl Fischer coulometric method. In contrast, freeze dry/dry heat treatment of product formulated with no albumin, 150mM NaCl and low moisture resulted in approximately 3 logio PPV reduction and 35% product recovery. Thus, minor changes in formulations such as the addition of 2% albumin may impact virus reduction and product recovery during a freeze dry/dry heat treatment. 

Table 5 Summary of virus reduction during 80° C heat treatment of freeze dried AHF...

Logio virus reduction Moisture Logio virus reduction Moisture Logio virus reduction Moisture... 

Logio virus reduction = Logio total viruspre-pD-Logio total viruSpost-so"c heat-... 

Fig. 10 Effect of formulation on virus reduction and product recovery. Protein G was formulated in buffer containing albumin (0% or 2%) and NaCI (OmM or 150mM). Virus was added and a sample was immediately removed for virus titration (Pre-FD). The virus-spiked material was aseptically filled into vials, freeze dried, and then heated at 80°C, 72 hr. Mock-spiked Protein G was processed like the virus-spiked samples but was used to measure product recovery (potency).

Screening and selection of the source plasma will only avoid contamination by known pathogens. The protein purification steps and specific virus reduction methods used in production processes, however, will inactivate and/or remove both known and unknown viruses. Terminal virus inactivation treatments are applied to product in final container and must balance virus inactivation with any modifications to protein immunogenicity, activity, and yield. While many upstream virus inactivation steps rely on chemical methods that involve the addition and subsequent removal of toxic agents (e.g., solvent/detergent), physical methods for virus inactivation, such as pH and heat, are used for terminal steps. 

The results from the albumin studies demonstrated that virus selection (e.g., HM175/18f and HM175/24a) and preparation (e.g., chloroform extracted and untreated virus) were just as important as the process conditions (e.g., protein concentration and temperature) in determining the outcome of virus reduction during pasteurization. Since laboratory adapted viruses and naturally occurring viruses may differ in their sensitivity to physico-chemical treatments, the results observed with model viruses must be carefully interpreted before they can be extrapolated to relevant viruses of concern. 

Certain process steps may be easier to model than others and duplicating the freeze dry/dry heat step at small scale is very difficult. The conclusions drawn from virus clearance studies are reliable only when the appropriateness of the small-scale models can be demonstrated. During the freeze dry/dry heat step of Koate -DVI, virus reduction was dependent on moisture levels so even the formulation of stoppers, which could impact the absorption of water during autoclaving and its release to freeze dried material, must be considered and tested. 

Manufacturing processes have evolved dramatically over the last few years. In the late 1980s, 76 /o of hemophiliacs were HCV positive, ° and between 1979 and 1985, approximately 50 /o of hemophiliacs had acquired HIV from plasma-derived FVIII. Since then, however, most U.S.-licensed plasma derivatives have not transmitted HBV, HCV, or HIV as a result of improvements in donor screening and test methods, and the inclusion of effective upstream virus-reduction and terminal virus-inactivation steps in manufacturing processes. Residual risks of virus transmission from plasma-derived products are now largely associated with non-enveloped viruses. " Thus, the need for additional terminal or upstream virus inactivation/removal steps still exists, but the current challenge is to develop cost effective methods against physico-chemically resistant non-enveloped viruses, such as human parvovirus B19. 

The models can be scaled to various sizes to fit the needs of the experiment. For example, a very small-scale nanofiltration system, such as a Planova P-15 hollow-fiber cartridge with O.OOl-m surface area (Asahi Kasei Corporation, Japan), can be used to study virus retention capabilities of a virus reduction step in a biological manufacturing process, whereas a scaled-up version of the same system with a surface area of 0.01 m provides an excellent way to study the nanofiltration process variables. In a nanofiltration validation study, a feed sample is typically spiked with a known quantity of a model virus. The mixture is filtered under the expected process... 

Table 3.4 Model virus reduction for rAHF-PFM processing [8]...

Specific virus reduction values achieved with each individual viral removal/inactiva-tion step are cumulative, so that the final viral reduction value attained reflects multiple removal and/or inactivation procedures occurring throughout the process. In summary, the processing of rAHF-PFM is associated with a significant viral reduction capacity for both lipid-enveloped and non-lipid-enveloped viruses and provides additional assurances of reliability with respect to pathogen safety. 

Steps to be Validated for Viral Clearance. Process validation for viral clearance is conducted only on robust steps that can (1) be scaled down accurately and (2) reproducibly and effectively remove and/or inactivate a wide variety of potential viral contaminants under a wide variety of process conditions [5,41]. The number of steps selected for validation depends on estimated viral clearance effectiveness based on historical data and target clearance values [5]. The FDA demands at least two different steps for virus reduction to guarantee safety and efficacy [7]. Potentially only two steps are required for antibody processes that use serum-free medium, but additional steps might be required if viral contamination risk is increased by using serum-containing medium [7]. Due to the use of live viruses to perform clearance studies, this work usually is outsourced to reduce cross-contamination issues [14]. 

---