Abstract
CONFERENCE PROCEEDING Proceedings of the PDA/FDA Adventitious Viruses in Biologics: Detection and Mitigation Strategies Workshop in Bethesda, MD, USA; December 1–3, 2010
Guest Editors: Arifa Khan (Bethesda, MD), Patricia Hughes (Bethesda, MD) and Michael Wiebe (San Francisco, CA)
The production of biologic drugs using mammalian cell production systems offers the benefits of high yield, proper protein folding, and faithful post-translational modifications. However, mammalian cell culture is vulnerable to contamination with adventitious agents, including mouse minute virus (MMV). The case study presented here demonstrates that MMV is a ubiquitous threat to CHO (Chinese hamster ovary) cell-based production of biologic drugs and that animal-free media components can be a contamination source. Compounding the risk posed by MMV, the contamination may be “silent,” with no impact on cell viability and product titers. Furthermore, contamination may not be detected using in vitro virus assays, and assays based on PCR (polymerase chain reaction) are required for reliable detection. The development of effective corrective and preventative action (CAPA) was greatly aided by the identification of the source of the contamination as an animal-free recombinant media additive. The execution of a CAPA that included disposal of contaminated materials, decontamination of the facility, and replacement of the contaminated raw material allowed the resumption of MMV-free production.
Introduction
Biologic drug manufacturing using mammalian cell culture is vulnerable to contamination with adventitious agents, including viruses such as mouse minute virus (MMV), also known as minute virus of mice (1, 2), Cache Valley virus (3), epizootic hemorrhagic disease virus (EHDV) (4, 5), and vesivirus (6). This contamination poses risks to patients, both a direct risk from the contaminating virus and an indirect risk from interruptions in drug supply. Despite industry efforts, viral contamination events continue to occur, and anecdotal data suggest that these occurrences are more common than have been reported in the literature (7).
Here we describe the contamination of a Chinese hamster ovary (CHO) cell production system with MMV that occurred at Merrimack Pharmaceuticals, Inc. (Merrimack; Cambridge, MA), during the production of a biologic drug for clinical use. The determination of the contamination's root cause was invaluable in the development of corrective and preventative action (CAPA) that included disposal of contaminated materials, decontamination of the facility, and replacement of the contaminated raw material. The detection of viruses in raw materials is often difficult, which poses major challenges in determining root causes of viral contaminations. We demonstrate that an approach where suspect materials were used in scaled-down CHO cell cultures that mimicked the conditions used in the good manufacturing practice (GMP) cell culture process was successful in identifying the source of the contamination.
MMV Background
MMV belongs to the Parvoviridae family of viruses and is a common infection of mice that is excreted in urine and feces (8). Because MMV is robust, infectious virus can persist in the environment, and this provides many potential opportunities for the virus to be introduced into the manufacturing process. Both U.S. Food and Drug Administration (FDA) and European Medicines Evaluation Agency (EMEA) guidances reflect the risk posed by MMV and require testing specifically for MMV (9, 10).
Typically, three types of assays are used to detect MMV: in vitro screening assays using CHO cells, an in vitro assay using a 324K cell assay, or polymerase chain reaction (PCR) assays designed to detect viral DNA.
In vitro screening assays using CHO cells are typically run for either a 14 or 28 day incubation period and utilize cytopathic effect (CPE) and/or hemagglutination as detection endpoints (11). In vitro assays using the 324K cell (an SV-40 transformed human newborn kidney cell) have been specifically developed and validated for the detection of MMV. The cultures are observed for CPE, sub-cultured, and split after 7 days. The endpoints of the assay are CPE throughout 21 days and hemagglutination using guinea pig and mouse erythrocytes (1, 2).
A variety of PCR assays have been developed for the detection of MMV (12, 13). In most of these assays the primers and probes are designed to hybridize to the NS1 gene sequence of MMV. The NS1 gene is conserved in parvoviruses (14, 15), and primers and probes against the NS1 gene detect common strains and related viruses (13). The assays are highly sensitive with a typical limit of detection (LOD) of 10 copies per assay. In quantitative PCR assays, this LOD typically correlates to approximately 3000 copies of MMV target DNA per milliliter of unprocessed bulk cell culture. It should be noted that the PCR assays do not discriminate between infectious virions and inactive copies of viral DNA, so it is possible to obtain a positive PCR result without the presence of infectious virus.
Transmission electron microscopy (TEM) analysis of unprocessed bulk cell culture has the potential to reveal MMV virions, but the assays typically used for the quantitation of endogenous retroviral particles have not been validated for the identification and quantification of MMV, and MMV particles might not be observed.
The Discovery of MMV Contamination in a CHO Cell Production Campaign
Phase 2 clinical supplies were being produced in a five-run campaign of batch-fed CHO cell cultures using 1000 L single-use bioreactors when routine testing of unprocessed bulk cell culture material from the third bioreactor run tested positive in a PCR assay for MMV contamination. The contract testing laboratory that performed the test suggested that because this result was in contrast to the “none detected” result reported for the routine in vitro virus assay of the same material, and that because PCR assays are inherently highly susceptible to false positives, the initial “out of specification” (OOS) positive PCR result required confirmation. An investigation initiated at the contract testing laboratory included retesting of the sample, and this retesting confirmed that the sample was positive for MMV DNA. As part of Merrimack's investigation, retained samples of the unprocessed bulk cell culture material were tested in a 324K cell assay that had been validated for the detection of infectious MMV, examined by thin section TEM, and also retested at a second independent laboratory using a quantitative MMV PCR assay. The 324K cell assay results were negative, and no particles resembling MMV were observed in the TEM results. However, the PCR assay was again positive for the presence of MMV DNA.
While the retesting of the lot 3 material was underway, the bulk cell culture material from the fourth bioreactor run and the seed train cell culture material from the fifth production run also tested positive for MMV in PCR assays. The viral load was estimated by quantitative PCR to be 4.9 × 104 and 2.8 × 106 copies/mL of unprocessed bulk in production lots 3 and lot 4, respectively.
It is important to note that these MMV contamination events were cryptic; no impact on cell viability or product titers was observed. This is different from what has been reported for a Cache Valley virus contamination of CHO cell production that caused an absence of cell sheets in a monolayer process and a perturbation in bioreactor parameters (O2 and base addition) (3) as well as an EHDV (a member of the reovirus family) contamination of CHO cell production where a decrease in product titer, cell death, and a generalized CPE was observed (4, 5).
The negative in vitro assays and the lack of any observable effect on bioreactor cell viability is in contrast to the observations of Genentech (1) during an MMV contamination in which positive hemadsorption at day 5 was observed in the in vitro virus test. This raised the possibility that virions detected were non-infectious. However, the high copy numbers observed suggested that viral replication had occurred during the cell culture process, and subsequent experiments (described below) confirmed that the MMV was indeed infectious.
The Response to the MMV Contamination
Upon confirmation of the initial positive MMV test result, Merrimack immediately halted all production of GMP material for clinical use and initiated an investigation to determine the risk to the patients, manufacturing staff, environment, manufacturing facility, and drug supply. As the MMV-contaminated upstream harvest material from runs 3 and 4 of the campaign had been processed through downstream purification, there was an immediate need to segregate and quarantine these bulk drug substance batches as well as all remaining materials and equipment used in their production. As none of the contaminated bulk drug substance from production runs 3 and 4 had been used to produce drug product vials, there was no direct potential risk to patients. At the request of regulatory authorities, the negative test results of the production runs 1 and 2 were confirmed with additional testing of unprocessed bulk retains prior to clinical use of drug product produced from these materials.
Merrimack next began to develop a CAPA plan that was based upon a risk assessment of the facility, equipment, raw materials, products, and procedures. The determination of the root cause became a priority of the investigation, as it assisted in establishing the extent of contamination and was helpful in development of CAPAs. These investigation and CAPA activities required significant time and resources and involved many functional departments at Merrimack (manufacturing, quality control, quality assurance, regulatory, and clinical) as well as the efforts of the FDA and raw material suppliers.
The Determination of the Root Cause of MMV Contamination
There are major challenges in determining the root cause of viral contaminations, including the fact that there are many potential entry points for viruses in the production system, and the detection of viruses in raw materials or facilities is difficult. The potential source of contaminating viruses is greatly influenced by the type of virus detected. Certain viruses are known to be associated with animal-derived raw materials such as serum or trypsin through infection of the donor animals—Cache Valley virus (3), EHDV (5), bovine viral diarrhea virus (BVDV) (16, 17), porcine parvovirus (16), porcine circovirus virus (18)—while other viruses have the potential to be introduced via environmental contamination of materials or equipment (e.g., MMV) (1, 2, 17).
Merrimack's investigation began with a thorough review and evaluation of the manufacturing process and facilities to understand where the MMV contamination originated, how to control the contamination, and how to prevent future contaminations. As MMV is spread by mice, there were many opportunities for contamination of raw materials or equipment. However, the upstream manufacturing process utilized disposable cell culture vessels (from seed train through 1000 L single-use bioreactors) and is an essentially closed process. Therefore, the introduction of the virus into the cell culture through a containment breach was considered unlikely.
The evaluation pointed to raw material as the source of the contamination, as the most likely source of viral contamination of CHO cell cultures is typically raw materials, particularly media and media components (16, 19). Contamination events with Cache Valley virus (3) and EHDV (5) have been attributed to serum. However, the media formulations used in the bioreactors did not contain serum, which eliminated this possible source and focused the investigation on other raw materials used to formulate the media.
A review and evaluation of Merrimack's facility was conducted to determine if the source of the contamination could be internal to Merrimack. The Merrimack facility, including storage areas, animal care facility, and pest control program, was reviewed for its potential to have caused contamination. A review of pest control service reports indicated no evidence of pest mice observed anywhere in Merrimack's facilities, and there was no evidence of MMV infection of mice in the animal care facility. Furthermore, there was no other evidence that indicated that the contamination was caused by exposure to MMV during internal storage or handling practices. Based upon these findings, it was concluded that contamination of materials within Merrimack's facility was unlikely.
The root cause investigation next focused directly on the raw materials used in the affected manufacturing campaigns. A comprehensive review of all raw materials used for the production campaign from seeding through harvest was performed. Consideration included the raw material used in the manufacturing process, the storage of the source materials, and packaging of the material. An elimination scheme was developed using an algorithm (Figure 1) that eliminated raw materials that had production lots/batches used in the non-contaminated lots 1 and 2, and that eliminated raw materials that had production lots/batches not used in the MMV contaminated lots 3, 4, and 5.
Raw material algorithm.
A number of assumptions were used for the purpose of narrowing the focus of the investigation. The assumptions included that lots 1 and 2 tested negative for MMV and that lots 3, 4, and 5 tested positive for MMV, and the materials used to manufacture those lots were considered in accordance with the MMV test results. Media and cell culture supplements became the prime focus. As the MMV positive results were seen in lot 5 during the seed train (approximately 15 generations from seed thaw), it was possible to narrow the investigation further and focus on the complements of media used from the cell bank thaw media and media used in the seed train culture. Figure 2 outlines the upstream process manufacturing steps where MMV was detected. Two components passed the elimination algorithm (Figure 1) and were identified as suspect raw materials: a single lot of a recombinant media additive and a single lot of CHO cell media. It is important to note that both of these materials were comprised of animal-free components and had been chosen for use, in part, because of the presumed low risk of contamination with adventitious agents.
Detection of MMV in the upstream manufacturing process.
MMV PCR testing of samples of the two suspect raw materials was performed, and the results were negative for both materials. However, as the detection of virus in raw materials is unlikely due to low titers and non-homogenous distribution (2), the negative test results did not eliminate the two suspect raw materials as potential sources of the MMV contamination.
Concurrent with this MMV PCR testing, the suppliers of the raw materials were contacted and their facilities and manufacturing practices were audited to evaluate the risk of contamination. These suppliers initiated internal investigations, which included direct testing of materials, and the results of these tests were also negative.
As a result of the non-informative results from the direct testing of the raw materials, it was decided to test the suspect raw materials in an experiment where the suspect materials were used in scaled-down CHO cell cultures that mimicked the conditions used in the GMP cell culture process. This approach included 15 generations of CHO cell culture doublings, which would provide an opportunity for low-input titers of MMV to replicate to levels that could be detected in MMV PCR assays.
The experiment was initiated with a master cell bank vial from the same master cell bank lot used in the GMP production campaign. These investigational research cultures utilized the suspected raw materials as well as new lots of the raw materials as negative controls, and each condition was run in triplicate. The media components used in the preparation of the test media are described in Figure 3, and the combinations of cell media components were chosen so that the possible experimental outcomes could support the following conclusions:
If condition 1 was positive (using suspect commercial media and suspect media additive), and condition 2 (using suspect commercial media and new non-suspect media additive) and condition 3 (control, using new non-suspect commercial media and new non-suspect media additive) were negative, then this would confirm the media additive as the source MMV.
If condition 1 (using suspect commercial media and suspect media additive) and condition 2 (using suspect commercial media and new non-suspect media additive) were both positive, and condition 3 (control, using new non-suspect commercial media and new non-suspect media additive) was negative, then this would confirm that the commercial media as the source of MMV.
Results of root cause investigation experiment.
Samples of the CHO cell cultures representing the various experimental conditions were taken from the CHO cell cultures approximately 15 generations from seed thaw. The samples were sent to a contract research laboratory for analysis by quantitative PCR (QPCR). The results of the QPCR testing exonerated the suspect lot of commercial media as sources of the MMV contamination (Figure 3). In contrast, the results of the QPCR testing strongly suggested that the suspect lot of the media additive was the single source of contamination. Merrimack concluded that this lot of media additive was contaminated with MMV and that the use of this lot was the root cause of the contamination of lot 3, lot 4, and lot 5 cell cultures. While the source of contamination was established to be the media additive, Merrimack could not determine how or where the material may have become contaminated. However, based on our internal investigation, we think it is unlikely that contamination of the media additive occurred in the Merrimack facility.
Corrective and Preventive Actions (CAPAs)
Corrective Actions
Merrimack developed and executed a CAPA plan that was based upon a risk assessment of the facility, equipment, raw materials, products, and procedures used during the production and processing of the MMV-contaminated materials.
The corrective actions began with the goal to eliminate the threat posed by the presence of the MMV-contaminated materials in the facility. A policy that included disposal of materials that could have been contaminated was initiated (if required, these materials were decontaminated prior to disposal). This included not only the drug substance produced from production runs 3 and 4 and cell culture material from production run 5 (decontaminated with 20% bleach prior to disposal), but also a number of unused materials (tubing, disposable product bags, etc.) that had been located in the production suites during the contamination events.
Because the contaminated unprocessed cell culture material from production runs 3 and 4 underwent recovery, clarification, and purification, each unit operation of the process was evaluated for potential contamination, and a rigorous decontamination of the facility and equipment was initiated that utilized a number of different approaches that have been demonstrated to be effective inactivating methods for parvoviruses (Table I).
MMV Decontamination Methods
In the upstream bioreactor, cell culture material recovery, and clarification steps, the product contact components are disposable. As these processes are run as a closed system that utilizes single-use disposable equipment, the risk of facility contamination from these steps was deemed to be low. Furthermore, the disposal of these product contact materials as biohazardous waste minimized any potential contamination of the facility or the environment.
The manufacturing suite (including both the cell culture and downstream processing cleanrooms) was decontaminated using vapor phase hydrogen peroxide (VHP). Non-disposable equipment such as tissue culture incubators, bioreactor components and control systems, membrane filtration cassette holders, pumps, and the chromatography skid were present in the manufacturing suite during the VHP process. The tissue culture incubators were then further decontaminated with a 9 h, 90 °C heating cycle (moist heat).
The downstream processing of the contaminated material posed a more significant risk. The downstream process incorporates disposable components where possible (disposable bags and tubing for buffers, intermediates, product, and waste stream). However, the process uses a chromatography skid, chromatography columns, and an ultrafiltration membrane filter holder that are not disposable. The columns and chromatography resins were decontaminated with 1 N NaOH, and the external surfaces of the columns were cleaned with 70% ethanol. The resins were then removed from the columns and discarded. The empty columns were decontaminated using 1 N NaOH, and all elastomers were replaced. The columns were repacked and sanitized before the next production campaign.
A similar process was used to decontaminate the ultrafiltration membrane filter holder and chromatography skid. These devices were first decontaminated with NaOH and then disassembled to remove disposable components. After the disposable components were replaced and the equipment was reassembled, the flow path was again sanitized with 1 N NaOH.
As part of a preplanned facility expansion, new heating, ventilation, and air conditioning (HVAC) ductwork and High-Efficiency Particulate Air (HEPA) filters were installed in the cleanrooms. Following these procedures, the manufacturing suite was again decontaminated using VHP prior to restart of the manufacturing facility.
Preventative Actions
In addition to the corrective actions described above, a number of preventive actions were instituted. The most valuable preventive action was the elimination of the media additive that had been identified as the root cause from the media formulation. A like-for-like change of this material from a new vendor was instituted. This new vendor employed different production methods that were considered to present a much lower risk of MMV contamination.
A second preventive action was initiated for the early and rapid detection of any future MMV contamination. In-process testing of seed train and bioreactor samples by MMV PCR assays prior to harvest of material is now performed to provide an opportunity to contain any MMV contamination to the upstream manufacturing process, which due to its single-use disposable product contact materials and closed systems would greatly simplify decontamination and control of the contamination.
Finally, Merrimack is in the process of developing a high-temperature, short-time (HTST) pasteurization process for cell culture media. HTST technology inactivates a wide range of adventitious agents and has been shown to provide an effective barrier against contamination (29).
Conclusions
The case study presented here demonstrates that MMV is a ubiquitous threat to CHO cell–based production of biologic drugs, and that animal-free media components can be a contamination source. Compounding the risk posed by MMV, the contamination may be “silent,” with no impact on cell viability and product titers. Furthermore, contamination may not be detected using in vitro virus assays, and PCR-based assays are required for reliable detection. The development of effective CAPAs was greatly aided by the identification of the source of the contamination as an animal-free recombinant media additive. The root cause determination required testing of suspect raw materials in an experiment that mimicked upstream production culture conditions to allow the contaminating virus to replicate to detectable levels. The execution of a CAPA that included disposal of contaminated materials, decontamination of the facility, and replacement of the contaminated raw material with a new, lower-risk source allowed Merrimack to resume MMV-free production.
Acknowledgments
We acknowledge the help of our colleagues who assisted in the investigation and in developing and executing the CAPA plan: Tony Awad, Ann Farnsworth, Kevin Murphy, Drew O'Brien, and Melissa Whitten.
- © PDA, Inc. 2011
References
Jump to section
Related Articles
Cited By...
- k-mer-Based Metagenomics Tools Provide a Fast and Sensitive Approach for the Detection of Viral Contaminants in Biopharmaceutical and Vaccine Manufacturing Applications Using Next-Generation Sequencing
- CRISPR/Cas9 gene editing for the creation of an MGAT1 deficient CHO cell line to control HIV-1 vaccine glycosylation
- Detection of Adventitious Agents Using Next-Generation Sequencing
- Meeting Report: 2013 PDA Virus & TSE Safety Forum
- Use of MMV as a Single Worst-Case Model Virus in Viral Filter Validation Studies