Abstract
While viral safety is a major concern for biologics manufactured using mammalian cells, the numbers of viral contamination events reported in the literature are quite low. During a 10-year period during which biologics derived from a variety of mammalian cell culture processes were evaluated using the in vitro virus screening assay, only reovirus type 2 and Cache Valley virus were detected, and these only in biologics manufactured in Chinese hamster ovary (CHO) cells. From the literature, we know that the murine parvovirus mouse minute virus has also been detected in biologics manufactured using CHO cells. The manifestations of these viral contaminants within the biotech manufacturing processes are discussed, as are the likely sources of the contaminants and some possible approaches for mitigating the risk of future occurrences of these types of contamination.
Introduction
While a constant concern, the contamination of a biotechnology manufacturing process with a virus happens very rarely (1). When it does occur, as happened recently at Genzyme (2), it can adversely impact the biopharmaceutical company as well as patients relying on its products. A widely publicized contamination event such as this causes ripples throughout the biotech industry. One company's misfortune becomes many other companies' concern as each scrambles to assess its own risk of experiencing the virus. This is especially true in cases, such as this, where the same viral entity is believed to have contaminated more than one product manufacturing process and/or biopharmaceutical entity (3).
Key to the handling of a viral contamination event in a manufacturing campaign is the rapid identification of the virus and subsequent steps taken to mitigate the risk of further occurrences. By examining past viral contamination events within the biotech industry, we can better predict the types of future occurrences, mitigate the risk of such events, develop rapid analytical methods for detection of contaminants, and therefore more rapidly identify future contaminants.
Adventitious virus testing is performed using the in vitro virus screening assay on a lot-by-lot basis for biopharmaceuticals. The testing is done at the bulk harvest stage, prior to any purification steps, so that the screening assay has the best chance of detecting any adventitious viral contaminants infecting the upstream manufacturing process (4, 5). The in vitro virus screening assay detects viral contaminants through a classical virology approach of inoculating bulk harvest material onto various host (detector) cells and observing the inoculated cells for 14 or 28 days for the presence of viral cytopathic effects. Additional endpoints used in the assay include evaluation of the cultures for hemadsorption and hemagglutination of erythrocytes derived from multiple animal species.
The viral screening assay can only determine whether a virus capable of causing an endpoint response is present in the sample. It cannot, on its own, identify the virus present. In order to accomplish the latter, the virus must be isolated and subjected to methodologies appropriate for viral identification (e.g., immunofluorescent antibody staining, electron microscopy, nucleic acid detection and sequencing, etc.).
In this paper, we examine the viral contaminants detected in biopharmaceutical bulk harvest samples tested at BioReliance over the past decade. Some of the manifestations of these contaminants within the biotech manufacturing processes are discussed, as are the likely sources of the contaminants. Finally, possible approaches for mitigating the risk of future occurrences of these types of contamination are discussed.
Detection Experience at BioReliance
During 10 years of adventitious virus screening of proteins, antibodies, vectors, vaccines, and oncolytics at BioReliance, adventitious viral contaminants were detected only a handful of times, and only in processes involving Chinese hamster ovary (CHO) cells. CHO cell processes have been employed extensively in this industry for manufacture of antibodies, peptibodies, and other recombinant proteins. The cell substrate is well characterized, is free of infectious retrovirus (6), and has proven to be relatively resistant to infection by adventitious viral contaminants (7, 8).
Mouse Minute Virus (MMV)
One of the few viruses that has been isolated from CHO-cell manufacturing processes on more than one occasion is the mouse parvovirus, mouse minute virus (MMV) (9). Sources of the past infections have been difficult to ascertain (9), but non-homogeneous contamination of a raw material has been considered likely. Ultimately, some form of rodent contamination must have been involved, as mice are the natural hosts for the virus.
Indication of an MMV contamination of a biologics manufacturing process may take the form of in-process results suggestive of poor cell viability or growth. In some cases, however, the infection may not become apparent until in vitro virus screening is performed at the bulk harvest stage. In the latter case, spread of the virus throughout the facility may have occurred by the time the infection is detected (9).
Parvoviruses are non-enveloped and relatively small, and are one of the few types of viruses that can survive on environmental surfaces. Due to the fact that these viruses are difficult to inactivate using physical and chemical means, parvoviruses such as MMV or bovine parvovirus are often employed as model viruses for cleaning and viral clearance validation studies. The U.S. Food and Drug Administration (FDA) Points to Consider guidance for the Manufacture and Testing of Monoclonal Antibody Products for Human Use (10) specifically discusses MMV as a virus of concern for biological production processes using hamster-cell bioreactors, and it indicates that tests that can detect this virus should be employed as part of routine testing of the unprocessed bulk.
MMV contamination at a manufacturing site can represent a pervasive problem due to the ability of the virus to survive on surfaces and ventilation ducts, and to the relative resistance of the virus to commonly employed cleaning agents. It is typical for manufacturers to resort to fumigation of facilities with agents such as vaporized hydrogen peroxide in order to eradicate the virus and prevent reinfection of subsequent upstream production processes.
Mitigation of the risk of MMV contamination has typically involved one or more of the following approaches: (1) early detection of infection in upstream processes through the use of nucleic acid–based in-process testing in order to prevent spread of the infection through the facility (9) and (2) application of raw material treatments such as UVC (11, 12) or high-temperature short-time (HTST) (12, 13).
MMV, despite being a known contaminant of CHO cell processes, was never detected using cell- or nucleic acid-based assays specifically designed, optimized, and validated at BioReliance for its detection, during the 10 year period addressed by this paper. Early detection of MMV infections and/or mitigation strategies to prevent infection of upstream processes at biologics manufacturing facilities may have contributed to this relative rarity of detection in lot release viral screening assays. The viruses that were detected at BioReliance during this period included the reovirus REO type 2 (14) and the bunyavirus Cache Valley virus (14, 15).
Reovirus (REO)
Reovirus (REO) contamination of biologics manufactured using CHO cell substrates was detected at least twice during the period of time covered by this paper. In each case, fetal bovine serum (FBS) was used at some point in the manufacturing process, and this animal-derived material was implicated as the source of the infection. This, unfortunately, was never proven through direct detection of the virus in the FBS lots used, either through prospective screening using the 9 CFR 113.53-based cell infectivity assays (16) in which REO viruses are specifically probed for, or through retrospective examination of the FBS lots using specific nucleic acid–based methods. REO viruses are known to infect cattle; however, they also infect other animal species, including man. Adding to the difficulty of assigning a specific source is the fact that the typing of REO viruses (e.g., types 1, 2, or 3) does not lead to an indication of its animal source. Thus, in the contamination cases occurring during this period, typing of the viruses as REO type 2 did not assist in the assignment of source. In each case, the contamination was eventually attributed to non-homogeneous contamination of the FBS with REO virus. This assignment was due in part to the fact that multiple bulk harvest lots manufactured using the same FBS lot became contaminated with the virus, while bulk harvest lots utilizing different FBS lots were not affected.
A REO virus infection of a biologics manufacturing process can be relatively stealthy and typically does not manifest itself through adverse changes in upstream in-process parameters. Little or no change in cell viability may occur. In monolayer processes, cell sheets may remain intact. In bioreactor processes, oxygen and base utilization, and metabolite production, may appear normal. Upstream production processes typically go to completion, with product yields and quality appearing to be normal. Viral contamination is unexpected, and is typically not discovered until indicated through cytopathic effects (Figure 1) observed in certain detector cells (especially CHO-K1, 324K) used in adventitious viral screening assays performed on the bulk harvest samples.
Human 324K cells: control culture (A), culture infected with REO type 2 virus (B) on day 20 post-inoculation.
REO viruses are relatively large (60–80 nm), non-enveloped RNA viruses. They are therefore somewhat resistant to inactivation by chemical means (11), although UV treatment and gamma-irradiation can be effective (11, 17, 18). Elimination of bovine serum from the upstream manufacturing process, or gamma-irradiation of the serum, should mitigate the risk of experiencing this virus.
Cache Valley Virus
As previously reported (15), the bunyavirus Cache Valley virus has been detected on at least three occasions over the period of time covered by this paper. The authors are aware of at least one additional incident, involving Cache Valley virus contamination of a CHO cell manufacturing process, that has not been reported in the literature. In each case, bovine serum was used at some point in the upstream manufacturing process, and it is this animal-derived material that has been implicated as the cause of the infections. As was the case with the REO virus contaminations described above, Cache Valley virus contamination of the lots of bovine serum used has never been proven through prospective screening using cell infectivity assays or through retrospective examination of the serum lots using specific nucleic acid–based methods. Cache Valley virus is known to infect livestock and is transmitted through a mosquito vector. In each case, the contamination was attributed to non-homogeneous contamination of the bovine serum with Cache Valley virus, again in part due to concordance of the bulk harvest lot infection patterns with specific serum lot usage.
In contrast to the case for REO virus contamination of a biologics manufacturing process, which was characterized above as being “stealthy,” Cache Valley virus contamination is manifested by striking adverse impacts on the upstream process parameters. These typically include obvious changes in cell viability. In monolayer processes, cell sheets may be lost. In bioreactors, oxygen and base utilization and metabolite production appear abnormal, indicative of cell death. These changes typically result in the upstream production processes being aborted prematurely, triggering an investigation. Due to the rapidity of cell death caused by exposure of fresh cells to the harvested bulk harvest fluids, cytotoxicity may even be suspected. Viral contamination is, however, confirmed through investigation. The rapid destruction of detector cells by this virus (Figure 2) is somewhat indicative of a Cache Valley virus infection.
Vero cells: 24 h post-infection (A) or 48 h post-infection (B) with Cache Valley virus.
Cache Valley virus is a relatively large (70–100 nm), enveloped RNA virus. It is susceptible, therefore, to inactivation by both physical and chemical means. Elimination of bovine serum from the manufacturing process, gamma-irradiation of the serum (18), and/or UVC or HTST treatment of culture media containing the serum (11, 12) should mitigate the risk of experiencing this virus.
Summary
The detection of adventitious viral contaminants in biopharmaceutical bulk harvests over the 10 year period addressed in this paper has been a rare event indeed, considering the great numbers of samples tested. The few viruses that have been detected at BioReliance over the past decade were found to have contaminated CHO cell processes. This is not believed to reflect a heightened susceptibility of this cell substrate to adventitious viral contamination, which appears on the contrary to be relatively resistant to infection. Rather, it likely reflects the high frequency with which this cell substrate is employed in this industry and the relatively great volumes of animal-derived materials associated with the affected CHO cell processes. The viruses detected (REO virus and Cache Valley virus) were likely introduced via contaminated bovine serum. Elimination or treatment of the serum or the serum-containing culture media through gamma-irradiation, UVC treatment, or HTST treatment should help to mitigate the risk of experiencing these viruses.
- © PDA, Inc. 2010
