Are in vitro tests sufficient screening tools for viral safety and product purity?

It’s not hard to spot bacteria in cell culture. The samples look cloudy, often smell badly, and a decent low-power microscope can actually pick up the granule-shaped microbes in the culture medium. Both red flags are enough to trigger an investigation at a Good Manufacturing Practices (GMP) facility.

Viruses, on the other hand, are largely invisible to the naked eye. They don’t light up like a neon sign or emit strange odors. But they sure do cause a lot of mischief when they pop up where they shouldn’t.

So pharmaceutical companies routinely screen cell banks and biologics for the presence of these so-called adventitious agents, industry-speak for microorganisms (viruses as well as bacteria and other pathogens) that unintentionally get introduced into the manufacturing process. By all accounts, the strategy works well; so well, in fact, that a 40-year-old standard using both in vivo and in vitro assays to screen for known and unknown contaminants may be superfluous.

A systematic evaluation led by Charles River and published last month in the journal Vaccine1 found that in vitro assays did as good a job — if not better — than in vivo tests at detecting a panel of 16 lab-adapted viruses ranging from measles, mumps and influenza A to adenoviruses and Coxsackie viruses.

The findings aren’t just academic. Given the call to reduce, refine and replace (3Rs) the use of animals in product safety testing, and the proliferation of new, more sensitive technologies such as deep sequencing and current generation mass spectrometers, the study’s authors suggest it may be time to re-examine a system that has been with us almost since the birth of the polio vaccine.

The impetus for the study emerged from US meetings in 2004 and 2009 that dealt with product safety issues and the emerging field of biologics. The more recent meeting, a workshop on Microbial Agents in Animal Cell Substrates, concluded that despite the many in vivo tests conducted over the years, no adventitious agents were detected in this way that were not also detected using in vitro methods.

The rationale behind this two-tiered approach dates back to the 1950s and early 1960s, when doctors began immunizing children routinely for polio, mumps, measles and rubella. For public safety reasons, the US Food and Drug Administration (FDA) — followed later by European regulatory authorities and eventually the World Health Organization — required vaccine manufacturers to inoculate animals and cell cultures with undiluted virus stocks to detect adventitious agents that could possibly contaminate vaccines. Laboratories commonly used suckling mice, guinea pigs, and embryonated chicken eggs for the in vivo tests, and three cell lines, including a human diploid cell line and a monkey kidney cell line, for the cell culture tests. These assays continue to be used today, with little evidence that contaminated vaccine products are making it to market.

(A notable exception: Common pig viruses that turned up in batches of Merck’s Rota Teq and GlaxoSmithKline’s Rotarix vaccines. Though not considered to be a risk to the public, the FDA temporarily suspended the use of the rotavirus vaccines until an investigation was completed.)

As the largest public funder of vaccine research and development in the world, the US National Institute of Allergy and Infectious Diseases (NIAID) thought it time to fund a study to compare and validate the current menu of assays, and Charles River, with its long track record of screening for adventitious viruses, was awarded the grant to conduct the analysis. The study screened for two strains of adenovirus, two strains of bovine virus, two strains of Coxsackie virus, echovirus 11, herpes simplex type 1, influenza A, measles, mumps, rhinovirus 2, rubella, simian cytomegalovirus, SV40 and vesicular stomatitis virus (VSV).

In vivo assays were performed using suckling mice, post-weaning mice and embryonated hens’ eggs. For the in vitro tests, 10-fold dilutions of each virus were inoculated onto MRC-5, HeLa and Vero cells, with a fourth cell line added for six of the 16 viruses.

All viruses were detected in vitro — though one of the bovine viruses required special conditions to detect and rubella was detected at low sensitivity. In contrast, in vivo, only six of the 11 viruses were detected and only two of them — influenza and VSV — were found at lower amounts in vivocompared to in vitro.

James Gombold, Senior Director of Technical Services at Charles River’s Malvern facility and the lead author of the study, said the data counter some longstanding assumptions about which assays work best. He used Coxsackie viruses (enteroviruses that live in the digestive tract) as an example. “It had always been said that Coxsackie viruses did not replicate well in cell culture and were more exquisitely detected in vivo,” said Gombold. “That dogma came from old virology literature when labs first started isolating this virus. The only way they were able to do it was in suckling mice, and so it was thought that if you wanted to isolate, detect and grow it, it had to be done in animals.”

The study includes some caveats. For practical reasons, scientists had to use a limited number of viruses, and left out entirely viral families that are unlikely contaminants of biologicals. Moreover, the 16 viruses had to be lab-adapted strains, as opposed to naturally-occurring strains that circulate in the environment. This latter limitation, says Gombold, could admittedly account for the efficient detection of viruses in the in vitro assays, especially in the cases of viruses expected to be detected better in vivo, such as Coxsackie.

What impact this study will have in re-shaping industry standards — or at least the discussion on adventitious agents — remains to be seen. But the study couldn’t come at a more productive time for vaccine manufacturing. In 2015, pharmaceutical manufacturers are expected to produce an estimated 35 billion doses of vaccine for diseases as common as the seasonal flu and as rare as the monkeypox virus, which circulates in central and West Africa’s rainforest region. Most of the two dozen or so vaccines will be viral vaccines, made with either inactivated (killed) viruses or live attenuated viruses (those whose genetic machinery has been intentionally weakened so they no longer cause disease.)

“We are at a turning point, if you will, in adventitious agents testing in the biopharmaceutical industry. There are many new methods coming available with tremendous promise but also a lot of unknowns,” noted Rebecca Sheets, a Vaccine Scientific and Regulatory Specialist at NIAID and a co-author of the study. “Regulators are looking at those new test methods and trying to make a decision; where do they fit and where might they be best utilized.”


  1. Vaccine. 32 (24) 2916 2014