In 1955, over 200,000 children in the United States received polio vaccine contaminated with live polio virus. Popularly known as the Cutter incident, this was a defining moment in biomanufacturing that led to the creation of regulatory systems in the development of biological products.
Any biotherapeutic needs to maintain a high standard of purity. Recombinant protein therapeutics, vaccines, plasma products, and cell- and gene-based therapeutic products that use cell culture to produce biologics are susceptible to contamination with viruses. Viral safety, then, is critical. It is often said to benefit from a three-pronged approach: 1) select virus-free or low-risk source materials; 2) test processes and products for viral contaminants at selected steps in the manufacturing pipeline; 3) perform downstream viral clearance tasks that encompass the removal or inactivation of potential viral contaminants.
Since viruses are the most abundant biological entities on the planet, is it feasible to expect complete removal of contaminating viral particles from biologics in bioprocesses? “But of course,” says Thomas R. Kreil, PhD, associate professor of virology at the Medical University of Vienna and vice president of global pathogen safety at Takeda. “Nobody would tolerate the presence of viruses in any medical intervention. The only exception would be where you are using the viral vector as part of the therapy, such as in gene therapy.”
Kreil co-authored a paper that was submitted to Nature Biotechnology by members of the Consortium on Adventitious Agent Contamination in Biomanufacturing (CAACB), a body organized by the Massachusetts Institute of Technology (MIT). The paper, which was published last April, summarized all contamination events known to have occurred in biomanufacturing. The information in this paper, Kreil maintains, may lead to “safety measures that will prevent a repeat of these mistakes with gene therapy vectors.”
In 1995, the International Council for Harmonization of Technical Requirements for Pharmaceuticals for Human Use (ICH) first issued the Q5A Guidance on the Viral Safety of Biotechnology Products Derived from Cell Lines of Human or Animal Origin. The guidance, which was updated in 1999, has been the basis for the regulatory evaluation of a broad range of biologics. These include products derived from in vitro cell culture (for example, interferons, monoclonal antibodies, and recombinant proteins) as well as cell and gene therapy products.
A revision of the Q5A is currently underway, and it is expected to reflect biotechnological advances in manufacturing, emerging product types, analytical technologies, and viral clearance validation and risk mitigation strategies for advanced manufacturing, such as continuous processing. Demonstration of viral clearance is an integral part of all biomanufacturing developments and investigational new drug (IND) applications.
“Zero risk is impossible to achieve,” observes Kreil. “What manufactures need to demonstrate is adequate safety margins, and that is applicable for all biological medicinal products.”
Several methods are being used to test for harmful viruses in biologics. Detecting viral contamination in cultures is more complicated than detecting other contaminating microbes. Some viruses cause a visible morphological effect in cells they infect and can therefore be detected microscopically, but other viruses integrate into the infected cell’s genome as provirus, leaving no visible trace.
Traditionally, in vitro methods, such as cell density testing and polymerase chain reaction testing, and animal models have been used to test for the presence of contaminating viruses in the product stream. However, these methods for viral detection are labor and time intensive and yield limited data.
The advent of next-generation sequencing (NGS) holds great promise in the quick and accurate detection of known and unknown contaminating viruses. Performed upstream on ingredients used in production pipelines, NGS allows specific and sensitive testing of a wide range of viral contaminants, instead of the detection of only specific viruses allowed by current methods.
Detection of newly synthesized RNAs, a sign of the transcriptionally active virus, may be used to differentiate between inert and active viral contaminants, which allows for greater accuracy in risk assessments. Aligned to GLP (Good Laboratory Practices) and GMP (Good Manufacturing Practices) requirements for biomanufacturing, NGS is applicable to cell banks, cell therapy drug products, vaccines, raw materials, viral inactivation testing, and testing for the lack of replication-competent viruses.
“The pricing of NGS is not high compared to in vivo testing,” emphasizes Marc Eloit, PhD, professor of virology at the Veterinary School of Maisons-Alfort, head of the Pathogen Discovery Laboratory at the Institut Pasteur Paris, and a funder of and scientific adviser to PathoQuest. “You need a few days to carry out an NGS test, but for in vivo testing, you may need months. Operationally, it is very cost effective. Vaccine companies such as Sanofi and GlaxoSmithKline are pushing to replace in vivo testing with NGS because of its efficacy and price.”
“If you look back at stories of previous viral contaminations of biological products, you will see that if NGS had been in place, it would have easily detected the contaminating virus,” notes Eloit. “If you have a good bioprocess and NGS testing, the risk for viral contamination is very, very remote.”
“NGS won’t replace viral clearance,” clarifies Sean O’Donnell, PhD, research advisor in the purification development and virology group at Eli Lilly & Company. “What it will do is tell us whether there is any contaminating virus in unprocessed bulk harvest or cell banks that you’re starting with, that you cannot detect using cell-based or PCR assays. NGS may become a very useful tool in the future and replace in vitro viral testing where we use indicator cell lines, as well as in vivo animal testing.”
“Examples of low-hanging fruit for NGS in bioprocesses are in cell line characterization and in investigational tools,” adds Paul Barone, PhD, co-director of biomanufacturing at the MIT Center for Biomedical Innovation and the director of the CAACB. “Adoption of NGS will likely not impact viral clearance studies. The assays for viral clearance validation using model viruses and traditional assays are well established. They work. NGS would be overkill.”
Traditionally, biomanufacturing accomplishes viral clearance through heat, chromatographic separations (for example, through protein A/ion exchange columns), low-pH treatments, solvent/detergent (S/D) viral inactivation, and nanofiltration. A 2010 meta-analytical study on regulatory submissions showed that nearly half of all viral clearance claims were based on chromatography techniques. In addition, it showed that nearly a third were based on filtration techniques.
Viral contamination, while rare, does occur in cell culture-based bioprocesses adhering to GMP standards, curbing the supply of life-saving drugs and imposing substantial financial losses. One approach adopted in the biotech industry is to implement preventative measures such as virus-retentive filtration (an upstream viral barrier).
Viral clearance becomes tricky when the product itself is a viral vector. But even in such cases, strategic process design can achieve adequate viral clearance. “For example,” says Kreil, “when you produce adeno-associated virus (AAV), the workhorse vector of gene therapy, you can still include an S/D step that effectively inactivates lipid-enveloped viruses. AAV does not carry a lipid envelop, and so, its potency is fully preserved in the treatment. The same is true for nanofilters of larger pore size. You can choose the filter such that the very small AAV passes through, but larger contaminating viruses are retained.”
Despite the judicious use of cell lines that are not susceptible to viral infection, and despite advances in viral testing, multiple orthologous steps in viral clearance methods are not going to be eliminated from bioproduction pipelines.
“[With] the implementation of redundant measures … you can depend on multiple mechanisms [to reduce] the likelihood that mistakes [will] happen,” maintains Kreil. “We do that in all aspects of life. Think about cars—antilock braking systems, safety belts, air bags. You don’t want to take a risk. It’s the same idea in bioprocesses.”
Regulatory guidelines require biomanufacturers to incorporate separate individual mechanisms of viral clearance rather than repeat the same mechanism to achieve greater removal of virus. “A combination of low-pH treatment and S/D treatment, followed by column chromatography, such as that involving protein A or ion exchange, and viral filtration ensures viral safety,” says O’Donnell.
A nearly universal method adopted to inactivate lipid-coated viruses in biotherapeutics has been to use detergents such as Triton X-100. However, this popular detergent degrades into hormone-like compounds with estrogen-mimetic activity that may negatively affect wildlife. These environmental concerns have caused the European Union to prohibit the use of Triton X-100.
Environmentally friendly alternatives to Triton X-100 include Nereid, a proprietary new compound synthesized by Takeda’s R&D group, and reduced Triton X-100. “At Takeda, we have developed a new detergent that we believe is going to be a fully competent replacement of Triton X-100,” reports Kreil. “It has identical viral inactivation potency as Triton X-100 and is expected to avoid harming the environment.
“Although this new molecule is structurally almost identical to Triton X-100, it is not expected to metabolize into a hormone-like compound. We are doing the final testing of the molecule and looking at scaling up production. If successful, we want the community to have access to this innovation ASAP.”
“Polysorbate 80 (PS80) was one of the first Triton X-100 replacements that we looked at for viral inactivation,” recalls O’Donnell. “Only at low concentrations of PS80 in the presence of bioreactor harvest material did we see effective viral inactivation, but not in phosphate-buffered saline. We boiled that down to the need for CHO cellular enzymes to hydrolyze PS80, specifically phospholipase A2. When we used phospholipase A2 with PS80, we saw viral inactivation in matrices that did not contain CHO cells.”
PS80 is known to be hydrolyzed to produce fatty acids (oleic and lauric acids) by enzymes from production cells. It was found that oleic acid alone was capable of viral inactivation. “This makes it tricky for biomanufacturers to use PS80,” says O’Donnell. “It adds a level of complexity since you now have to demonstrate that you have a certain level of PS80 hydrolysis into x amount of oleic acid to demonstrate viral inactivation. That is how we came to investigate Simulsol™ SL 11W, which is made by a company called Seppic.”
Although purified oleic acid causes viral inactivation, it cannot be directly used for viral inactivation because when added directly to the bioreactor, it is insoluble in aqueous solution. This fouls filters, sticks to stainless steel, and becomes prohibitive for the manufacturing process. “Oleic acid works great for viral inactivation,” O’Donnell remarks, “but it’s just not feasible from a manufacturing standpoint.”
Simulsol SL 11W works optimally with different starting matrices and has fast kinetics very similar to Triton X-100. It works over a large temperature range, from 4 to 30°C, and it doesn’t foul any filters going onto the protein A column where it is removed from the product stream. The efficient removal of Simulsol SL 11W, confirmed by mass spectrometry, makes it very “palatable for manufacturing processes,” suggests O’Donnell. “We’re still investigating Simulsol SL 11W. We have not implemented it yet in our manufacturing processes, but we are very close.”
Viral filtration, a highly effective size-based removal of viral particles from the product stream, has its own set of challenges. Parvovirus filters have a pore size of around 20 nm. This is smaller than most common contaminating viruses. “When we test these filters, we use parvoviruses—either MMV (mouse minute virus) or PPV (porcine parvovirus)—that are 18 to 20 nm in size,” details O’Donnell.
Fouling of filters is the irreversible decline of flow through the filter that occurs when particles cake the membrane surface. Therefore, filters need periodical replacement or cleaning.
“We have seen that operating conditions can affect virus removal by these filters,” says O’Donnell. Some filters are sensitive to the start and stop of flow. “If you have a feed stream coming into the filter, stop the flow for whatever reason, and then repressurize the filter, you can see force breakthroughs of virus particles,” he continues. “You really have to understand the operating parameters of your viral filter and your feed streams.”
Viral clearance in batch versus continuous operations
Large-scale biomanufacturing is currently evolving from batch mode processing (BP) to continuous processing (CP). In contrast to modular BP unit operations separated by holding periods, CP accomplishes the entire pipeline in a seamlessly integrated manner and has several advantages, such as product stability, efficient use of resources, and reduction in operation size, thereby reducing the manufacturing footprint and rendering the process more cost effective and the medical product more affordable. The evolution of BP to CP is a welcome upgrade in productivity so long as there is sufficient and strong scientific evidence to assure viral clearance and safety.
“We’ve recently published two papers to show how S/D and low-pH treatment can be applied in CP,” Kreil mentions. “It is technically a bit more demanding, and the validation of the viral inactivation process is more complex. But it can be done. We have also shown that continuous nanofiltration can be run for weeks and that it is completely compatible with CP.” You could take two or more viral clearance methods and implement them redundantly in CP for enhanced safety.
Filtration-based viral removal methods are easily adaptable to continuous processing. “No matter whether it is a batch mode or a continuous single-unit operation, if you are operating within the manufacturing specifications of the viral filters, they are just as effective,” maintains O’Donnell. “As you increase volumetric throughput across the filters, you see virus breakthroughs when you get to critical thresholds. So, appropriate sizing and scaling is important in CP. Changing filters and increasing the surface area of the filters to accommodate the increased load is going to be part of it.”
When choosing viral clearance strategies to analyze products of CP pipelines, the design and qualification of scale-down models and integration of linked unit operations will need modification in viral clearance validation.
Key issues in the transition from BP to CP include determining how widely used viral clearance operations can be incorporated into CP pipelines in accordance with the Q5A guidelines, and how a representative scale model can be designed and implemented for viral clearance studies given the increased productivity and product load in CP.
After viral clearance, regulatory authorities require a demonstration of how much virus has been removed. At the beginning of the process, right after harvesting cells from the bioreactor, many manufacturers test for MMV since it is a common contaminant of CHO cells. After downstream viral removal and inactivation steps, biomanufacturers must demonstrate a reduction in viral load from the initial harvest step.
This is done through calculating the “retrovirus safety margin” in the case of CHO cells. This calculation considers bioreactor titers, purification yields, and PCR or transmission electron microscopy test results.
“Typically, for something that will go commercial, you want to show that you have six logs of retroviral safety margin after the downstream viral clearance steps are totaled,” states O’Donnell. “The more robust viral clearance steps accomplish greater than four logs of viral inactivation or removal, and you want multiple steps of viral removal in the process.”
“There is ample data from recombinant protein products to show that the current methods used in viral clearance work, even in removing viruses that you don’t know are there,” says Barone. The current viral clearance challenge, he continues, “is in emerging products such as viral vaccines, viral vectors, and especially cell therapies.”
Understanding the viral clearance parameters with the new CP pipelines is going to be challenging for biomanufacturers, particularly at the outset. “But many manufacturers are already beginning to understand how the operating conditions around CP parameters will work,” says O’Donnell.
The other big challenge is coming up with an alternative to Triton X-100, which has been prohibited by the European Union. “We’re one of many manufacturers that has come up with alternatives such as adding passive anion exchange columns to remove viruses, but these are costly and time consuming, O’Donnell complains. “Using something like Simulsol, which you can drop into any manufacturing process, is an attractive option.”
Viral safety a pandemic context
The potential impact of SARS-CoV-2 on biomanufacturing operations has raised several questions. “We have seen an increasing interest and an acceleration of the discussion on NGS,” says Eloit. “It provides COVID-19 vaccine developers the capacity to be faster in this race against time. Viral safety timelines are reduced four- to fivefold with the NGS approach compared to animal testing.”
“With any pandemic,” adds Barone, “a biotech company has to assess the risk to the process and the product that the emerging virus poses, as well as the risk to business continuity independent of whether the novel virus can be a product contaminant.”
The key questions regarding a newly emerging, pandemic-causing virus are: Is it a product contaminant? What are its sources? Is it airborne? Could it get into the raw materials and replicate in host cell lines? Would you detect the novel virus using existing assays? Will downstream viral clearance eliminate this novel virus?
Questions such as these were included in a recent survey of CAACB members. The survey’s results, which appeared in April 2020, suggest how manufacturers may respond to the COVID-19 pandemic.
A key lesson from the survey, says Barone, is as follows: “If you are using a production cell line that does not replicate the virus, and if you have robust downstream viral clearance, and if you are using a test that detects SARS-CoV-2, then there is very little risk to the process or the product. But with biotech products where these conditions do not hold—for example, a product of a cell line that can replicate SARS-CoV-2, where there is no downstream viral clearance, and where there is no time for in vitro viral testing, such as in some cell therapy processes—the risk is much higher.”
“In June 2020,” Kreil observes, “the FDA asked for research into whether cell lines used in the production of biologics are infectable by SARS-CoV-2, and whether our viral testing parameters can detect and remove SARS-CoV-2. We had already tested our platform cell lines, and we had determined that they are not infectible. Our testing assays for adventitious viruses can very comfortably detect the coronavirus, and our viral clearance strategies are capable of effectively removing and inactivating SARS-CoV-2.”
Since SARS-CoV-2 is a relatively large lipid-coated virus, the virus can be effectively cleared by S/D treatment, low-pH incubation, caprylate treatment, pasteurization, and dry heat, as well as by nanofiltration and fractionation.
“In the case of cell therapies, there is no viral clearance,” says Barone. “Right now, I can’t even imagine a technology that can remove viral infected cells from noninfected cells. That entire prong of the safety tripod is gone for cell therapy products. Also, for autologous cell therapy, because of the time constraint, where the patient needs the therapy immediately, you don’t have time to wait for viral assays.”
Although viral clearance is undoubtedly effective and works very well in the production of recombinant proteins, it is not an option for cell therapy products. “For cell therapy products,” Barone continues, “you really need to depend on ways to avoid introducing the virus in the first place, such as rigorous process controls, aseptic processing, and closed systems. Ideally, even if a very good viral clearance strategy is available, you don’t want to need it.”
“I’ve been in this line of work for the past 20 years, and I’ve spent a lot of time thinking through contamination issues,” declares Kreil. “We have seen that things can go sour. Contaminations can occur. So why put the lives of the patients we serve at risk? If we forget what our job is—to keep our patients safe and see that they have a steady supply of their medicines—then we do not deserve to be in this industry. We have witnessed that virology is an issue in every bioprocess. We need an adequate level of virological expertise to understand these concerns, to investigate and implement remediations, and to validate the remedial measures. Then we have done our job.”