| Kate Smith |
The potential for vaccines and other biological pharmaceutical products to be contaminated with viruses has been known for decades. Awareness goes right back to the early days of the polio vaccine, more than half a century ago, when early versions of the life-saving vaccine were contaminated with simian virus-40 (SV-40) from infected rhesus monkey kidney cells. There have been numerous instances since then, some more serious than others, and even some that have led to manufacturing facilities being closed down for long periods of time for decontamination.
Any manufacturing process that relies on starting materials from an animal source can be contaminated with viruses. Many biological drugs are made via cell culture, typically using Chinese Hamster Ovary (CHO) cells or murine cells, and these cells invariably (and unavoidably) contain endogenous viral contaminants. While these viruses may have no health effects in humans, their presence in final drug products is far from ideal. Similarly, the use of media that contain components from animal sources, such as serum, can cause viral contamination of a product.
Human plasma is another potential source of viral contamination. Blood donations are screened, but there is rarely time to carry out full viral analyses to test for the presence of viral contaminants before they are used. Instead, companies who use blood-derived materials in their processes rely on donor screening programmes to minimise the chance of dangerous viruses such as HIV or Hepatitis B and C inadvertently being passed on to patients via drug products.
Starting materials are not the only source of viral contamination, however. Any raw material that is used in a process, regardless of its source, could potentially be contaminated with a virus. There is also the possibility that viral contaminants could enter the manufacturing process from a human operator within the facility.
The impact of viral contamination can be enormous. The SV-40 present in early polio vaccine led to a widespread distrust of the vaccine, and it remains unclear today whether the inadvertent infection of recipients has had long-term health effects. The headline-hitting vesivirus contamination at Genzyme’s manufacturing facility in Boston in 2009 led to shortages of two drugs to treat rare conditions, Cerezyme for Gaucher’s disease and Fabrazyme for Fabry’s disease, while the facility was closed for months as it was being decontaminated. In addition, the detection of porcine circovirus in both GSK and Merck’s rotavirus vaccines led to brief product suspensions, before FDA declared that there is no risk to human health from either virus and allowed the vaccines to be used again.
These incidents highlight the importance of continual testing for viruses during a biological product’s manufacture process and throughout its lifecycle, as it becomes possible to detect new viruses. Even if there is no health risk, viruses should be inactivated or removed from final drug products, if at all possible.
Managing the risk
Screening is the initial defence against viral contamination. The regulators have clear rules about the required testing procedures necessary for all cell lines and other raw materials that are used in biomanufacturing processes. While this is a time-consuming process – the full characterisation of both master cell banks and working cell banks may take up to three months – it is essential. These tests will highlight the presence of any adventitious biological contaminants, including viruses.
Cell-based assays are used to test for any viral contaminants. Usually, multiple cell lines – maybe three or four – will be employed in the test runs, preferred because of their susceptibility to a range of viruses. One of these will be the cell line that is being used in the manufacturing process, such as the CHO cell line, and another will be known to be susceptible to human viruses. The choice is dependent on the potential contaminants; either as endogenous contaminants or from the way the cells are handled.
Some viruses will not grow in in vitro assays, and therefore in vivo tests must be used to check for these. Again, the species chosen will depend on the nature and source of the cells being used in the biomanufacturing process. As with any procedures using animals, these tests must be carefully controlled.
Other tests can also be employed. Notably, retrovirus can be detected using electron microscopy or reverse transcriptase assays. Those viruses that are typically present in rodent cell lines can also be detected using antibody production tests. Massively parallel sequencing is a relatively new technique that is now becoming established in the field. While it is not intended for routine quality control use, it does allow detailed studies to be made on raw materials. Once potential viral contaminants have been identified, the results can be used to develop specific Polymerase Chain Reaction (PCR) tests for routine QC use.
Viral inactivation and clearance
While the testing of raw materials and cell lines provides important information about the presence of viral contamination, it does not guarantee there is none. Fundamentally, it will only detect viruses that are actively tested for. Unknown viruses may still be present. Therefore, for safety reasons, regulators have instituted specific requirements to incorporate steps capable of inactivation or removal of potential viral contamination, from those known to be present, to those that are currently unknown.
Generic platform-type processes are typically used to manufacture biopharmaceutical products like monoclonal antibodies. The purification processes are, therefore, similar, so it is relatively straightforward to incorporate viral inactivation steps to prevent viral contamination of the final product. Regulators demand the inclusion of at least two robust techniques to ensure minimal amounts of the virus remain; these must operate via different mechanisms. The two most common techniques are incubation at low pH, which inactivates enveloped viruses, and nanofiltration, which is able to remove both enveloped and non-enveloped viruses.
It is usually fairly easy to include a low pH inactivation treatment in the purification process, especially if the first capture step of the purification is a Protein A-based affinity step. In this process, the pH of the buffer is reduced to elute the product, therefore giving an obvious point for the introduction of a low-pH inactivation step. To insure viral inactivation, the pH is adjusted to a defined range, usually between 3.5 and 3.9, and held there for one to four hours.
This technique is very effective at inactivating many enveloped viruses, particularly those with low, or low-to-medium, resistance to physicochemical influences. The efficiency of the ‘removal process’ is checked in the lab via samples spiked with model viruses, typically murine leukaemia retrovirus and pseudorabies herpesvirus.
There is a risk that the product may not be sufficiently stable for long periods at this pH, and thus its stability towards acid cleavage must also be checked to ensure it will not decompose on purification. If this is the case, alternative inactivation techniques must be investigated.
Solvent detergent treatment, which was developed in the blood products industry, can also inactivate viruses. It does this by interfering with the virus’s lipid coat, thus disrupting the viral membrane. The product is incubated in a defined concentration of solvent, which is typically 0.15 to 0.30 per cent TnBP, to accelerate the interaction between the lipid coat of the virus and the detergent. This is usually Triton X-100 or Tween 80, and a typical concentration of detergent is 0.1 to 1 per cent. As with low-pH treatment, the technique is most effective on enveloped viruses.
Both of these methods work well for enveloped viruses, but rarely for non-enveloped viruses, as they are normally resistant to low pH and not affected by detergent treatment. A number of other techniques are also occasionally used to inactivate viruses. Gamma irradiation is frequently employed in the sterilisation of finished medical devices. UV irradiation is now starting to gain ground in the purification of monoclonal antibodies. Beta-propiolactone treatment is routine for vaccine products, while formaldehyde treatment is also used in some cases.
Removing viruses
The most robust – and most routinely used – method of virus removal, nanofiltration, relies on the relatively large size of the virus compared to the product. Filters designed to remove viruses from monoclonal antibodies and other protein products have pore sizes in the range of 15–20nm, and the smallest pore size that still allows the product to pass through successfully should be selected, as this will retain the maximum amount of virus. Nanofiltration can be particularly effective for the removal of common viral contaminants that low pH treatment fails to inactivate, including minute virus of mice, which is often present in murine cell lines.
The starting point for a biomanufacturer will invariably be the brand of filter they already use within their platform process; all the major filter manufacturers produce nanopore filters for virus removal, including Asahi, Millipore, Pall and Sartorius. However, this may not be the best choice for an individual product, and test runs using different filters should be carried out to identify which gives optimal results while minimising the filter area required, for cost reasons. Regulators are looking for at least a 4-log reduction of model viruses in test runs.
During the process development Vmax studies are conducted to define the operational capacity for the filter. This helps determine the defined membrane area required to process the production scale batch. Careful design of the spiking study is critical to ensure that this defined capacity can successfully be validated in the study. Of course, the virus spikes must be of sufficient quality to ensure they do not adversely impact the validation study.
All filters are not the same, and this is even more the case in real-world process use. They are all able to remove viruses, but filter capacity and flow decay is different from filter to filter, process to process and product to product. As a result, the brand of filter a manufacturer is most familiar with may not be the one that gives the best results. This is why careful testing of more than one brand of filter is essential for each different product.
Alternative removal strategies
While nanofiltration is by far the most robust technique for removing viruses, there are other, albeit less reliable, techniques. Perhaps the most important of these is chromatography, which is commonly included in the platform process’s purification steps. Its lack of robustness results from the number of variable operational parameters that make it difficult to replicate conditions precisely from one run to the next. These include operating capacities, flow rates, buffer pH and conductivities. Even minute variations can have an impact. Proving that changes in these parameters had no negative impact on the effectiveness of viral clearance would be both challenging and expensive.
However, chromatography does have its place in the viral removal arsenal, as it can provide an additional ‘belt and braces’ level of confidence that a virus is being removed effectively. The precise nature of the viral clearance via chromatography varies. Commonly, the virus binds to the column as the product flows through, and is then removed from the column before it is reused via a regeneration procedure. In other chromatographic processes, the difference in binding affinity for the column enables the virus and the product to be separated.
In some cases, particularly where a Protein A affinity step is being used, a combination of inactivation and removal by chromatographic separation is taking place. If this is the case, tests should be run to identify how much of the virus is being inactivated and how much removed. A PCR-based assay that detects the viral genome will be able to provide this information, since it does not discriminate between inactivated and infectious virus.
Improving safety
It remains impossible to ensure that no viral contamination whatsoever enters a biomanufacturing process. Some viruses are inherently present in the cells that are used in cell culture; others commonly occur in other starting materials. However, a comprehensive programme of efficient testing strategies and effective removal and inactivation processes will ensure the safety of biological products. Regulatory guidelines have been established for a reason: mainly, patient safety; only by implementing them carefully and fully validating their effectiveness, can this be achieved.