Fears about mutant H5N1 hinge on ferrets as flu model

Dec 23, 2011 (CIDRAP News) – In the influenza literature, it's a given that ferrets are the best animal model for influenza in humans. They show similar clinical signs of disease, such as fever, coughing, and sneezing, and flu viruses that spread among humans usually spread in ferrets as well.

But if a particular strain of flu, such as H5N1, causes severe illness and spreads easily in ferrets, does that mean it will do the same things in humans? That's a key question raised by two controversial unpublished studies in which ferrets were infected with mutant H5N1 avian flu viruses—and virologists say there's room for doubt about the answer.

This week the Department of Health and Human Services (HHS), following advice from the independent National Science Advisory Board for Biosecurity (NSABB), recommended that the journals Science and Nature withhold key details of two studies in which mutant H5N1 flu strains gained the ability to spread between ferrets by the airborne route. HHS promised to work on a mechanism for providing the full details to scientists who need them.

One of the studies was done by a team led by Dr. Ron Fouchier of Erasmus University in the Netherlands and was submitted to Science. The other, led by Dr. Yoshihiro Kawaoka of the University of Wisconsin and the University of Tokyo, was submitted to Nature.

The editors of the journals have said they are considering the government recommendation and waiting to see what sort of plan officials propose for sharing the details with experts.

The concern is that publication of the reports could lead to the unleashing, through a laboratory accident or criminal activity, of a highly dangerous virus: an H5N1 strain that could spread quickly from person to person across the globe. H5N1 disease is often lethal in humans, with a fatality rate of about 59% in confirmed cases, but the virus has not gained the ability to spread easily in people. The World Health Organization (WHO) has confirmed 574 cases of H5N1 since 2003, including 337 deaths.

The NSABB in its deliberations spent considerable time discussing the ferret model of influenza, according to James A. Roth, DVM, PhD, a member of the board and director of the Center for Food Security and Public Health at Iowa State University's College of Veterinary Medicine in Ames.

"The consensus is that the ferret is the most reliable animal model for human influenza," Roth told CIDRAP News. "However, it is not perfect. The results in ferrets are likely to predict the transmissibility and severity of a given influenza virus in humans. Therefore there is a need for caution when an influenza virus with an H5 hemagglutinin is transmissible in ferrets and has the potential to produce morbidity and mortality in humans. Especially considering that the human population has no immunity to H5 hemagglutinin."

H5N1 transmission experiment
Naturally occurring H5N1 viruses act much the same in ferrets as in humans: They cause severe disease but they do not spread efficiently, either by respiratory droplets or direct contact. Fouchier and his colleagues wanted to identify mutations that would make H5N1 spread more easily in ferrets, and by implication in humans as well, so that scientists could be alert for those mutations in wild strains.

Fouchier's team introduced various mutations into H5N1 viruses and found that as few as five single mutations enabled them to bind to human nasal and tracheal cells in lab cultures, according to a Scientific American report of Fouchier's comments at a European meeting in September. But this mutant strain still didn't spread easily in ferrets.

The scientists then allowed the mutant strain to evolve naturally in ferrets, according to the report. They infected one ferret and, after it got sick, used a specimen from it to infect a second ferret. Ten repetitions of this "passaging" process led to an H5N1 strain that could spread among ferrets by air, without direct contact. This mutant strain was said to be just as lethal in the ferrets as its predecessor.

No details about the Kawaoka experiment have been released, but it is said to have yielded similar results to Fouchier's.

Ferret-human parallels
Influenza researchers and published research reports agree that ferrets are the best animal model for studying human flu.

"The ferret is an excellent animal model for human influenza. The virulence and transmissibility of a wide range of influenza viruses are found to be similar between ferrets and humans," said Anice Lowen, PhD, assistant professor of microbiology and immunology at the Emory University School of Medicine in Atlanta, whose research focuses on flu transmission.

"For example," she commented by e-mail, "human seasonal strains cause mild disease in ferrets and transmit very well from ferret to ferret; whereas wild-type H5N1 strains cause severe (fatal) disease and do not transmit readily. Similarly, low pathogenic avian influenza viruses do not, in general, transmit among either ferrets or humans. Thus, to date, ferrets have been a highly reliable model for human influenza."

Experts also point out that the clinical signs of flu in ferrets are generally similar to those in humans. The typical ones include sneezing, fever, and nasal discharge, says a report by Centers for Disease Control and Prevention (CDC) authors in the 2009 annual publication Advances in Viral Research.

Other parallels, less obvious, have to do with pathophysiology. In both species, the receptors on epithelial cells of the upper airway are marked predominantly by alpha2,6-linked sialic acids, unlike the epithelial cells in birds, which feature alpha2,3-linked sialic acids, according to the Advances article. In addition, avian H5N1 and human H3N2 viruses attach to human and ferret tissues in similar patterns.

Researchers also have found parallels between ferrets and humans in the early response to flu virus infection. Recently a team from the CDC and the University of Quebec studied the local innate immune responses to flu in the upper respiratory tract of ferrets, including evidence of cytokines, chemical messengers that trigger inflammation.

They found a strong temporal correlation between clinical signs, virus shedding, transmission dynamics, and the expression of particular cytokines that are associated with similar clinical signs in humans, according to their report in the Journal of Infectious Diseases.

The similarities between ferret and human flu are such that ferrets have been useful in evaluating vaccines for human use and in showing the effects of mutations that confer resistance to antiviral drugs on viral fitness and transmissibility, according to a review published by CDC scientists in September in Disease Models and Mechanisms.

Flu manifestations not all the same
Despite all of these parallels, there are some differences between flu in ferrets and humans, researchers say.

For example, Lowen commented, "Ferrets sneeze frequently when infected with seasonal influenza viruses, whereas humans tend to cough and not to sneeze." Also, she said, "Seasonal influenza in ferrets is perhaps milder than in humans, given that the ferrets in question are naive hosts (analogous to children infected for the first time)."

Another difference is that flu causes neurologic signs more often in ferrets, according to Peter Palese, PhD, professor and chair of the Department of Microbiology at Mount Sinai School of Medicine in New York City, who has been critical of the NSABB move to suppress details of the Fouchier and Kawaoka studies.

"Ferrets have a lot of neurologic symptoms with flu, and that certainly doesn't happen in humans," he said. "There's plenty of evidence that ferrets are a much too sensitive system."

He pointed to a 2007 study in the Journal of Virology, in which ferrets that were inoculated with high doses of H5N1 strains from Vietnam showed neurologic signs before dying. Also, in an earlier study, published in Avian Diseases in 2003, scientists found that ferrets infected with the 1997 strain of H5N1 from the first human cases ever confirmed, in Hong Kong, had neurologic manifestations of the infection.

Another researcher who has studied flu in animals, Daniel R. Perez, PhD, of the University of Maryland, agreed that neurologic manifestations are more common in ferrets than what the literature says about human cases. Perez is an associate professor of virology in the Department of Veterinary Medicine in the university's College of Agriculture and Natural Resources and program director for the Avian Influenza Cooperative Agricultural Project there.

Perez said there's a further difference: "In general, I'd say ferrets are more susceptible than humans to H5N1. The manifestations of the disease seem to be a lot more dramatic and noticeable in ferrets than humans. So the human immune system tends to handle the infection better than ferrets."

But he cautioned that not all strains of H5N1 have been tested in ferrets, and the animals are exposed to the virus in controlled laboratory conditions, unlike humans, making comparisons tricky.

He also observed that scientists face some technical limitations in studying the immune response of ferrets: "We are limited in making parallels between what happens in humans and ferrets in terms of the immune responses. We don't have reagents that are ferret-specific that allow us to study immune responses in ferrets the way we've studied them in mice."

Implications unclear
But what, if anything, these differences in clinical manifestations of flu say about the risk represented by the reported mutant H5N1 strains is unclear.

Some virologists say that since ferrets aren't humans, one can't assume that a virus that's highly pathogenic and transmissible in ferrets will have the same qualities in human hosts. Further, they assert that a virus that has been passaged multiple times in an animal, as was reportedly done in Fouchier's experiment, may become attenuated, or less virulent, in humans.

Other experts, however, say that if a flu virus spreads efficiently in ferrets, it will probably spread readily in humans as well. And while passaging a virus in some animal hosts may lead to attenuation in humans, they say, it's not clear if this is necessarily true in the specific case of ferrets.

Questioning assumptions
One virologist who questions the assumption that the mutant H5N1 strains identified by Fouchier and Kawaoka would be highly dangerous is Vincent Racaniello, PhD, Higgins professor of microbiology and immunology at Columbia University and author of Virology Blog. He has been critical of the move to suppress details of the H5N1 studies.

Racaniello acknowledges that ferrets are considered the best flu model. However, he said in a recent interview, "Ferrets aren't people. . . . You can get some good information out of it [ferret research], but to say it's predicting how flu will be in people is wrong. We just can't say that."

For example, he said that when the 2009 H1N1 virus emerged, a number of research teams studied it in ferrets, and some found it could be lethal, with one study showing a 30% to 50% mortality rate. One scientist concluded that it was a very dangerous virus, he said.

"As we found out subsequently, it was really quite mild in people," Racaniello said. "I just don't see how you can say this [Fouchier's H5N1 virus] is the most dangerous virus ever made when you don't know what happens in people." (To be fair, some public health experts have objected to the description of the 2009 pandemic as mild, because, unlike seasonal flu, the virus disproportionately affected children and nonelderly adults.)

Palese voiced similar views. Asserting that it would be "ridiculous" not to publish the Fouchier and Kawaoka data in full, he said, "I don't think the evidence indicating this is really dangerous in humans is there."

Racaniello also observed that passing a virus through an animal host multiple times is a standard way to potentially reduce its virulence in humans. "That can often result in increased virulence in the animal, but when it goes into humans, it has reduced virulence; it's adapted to the animals. The Sabin polio vaccine was handled this way. The same with the yellow fever vaccine.

"That doesn't predict what H5N1 in ferrets is going to do, but one possibility is that it will become less virulent in people," he said. "What they [Fouchier's team] did was select for transmissibility. You're better adapting it to ferrets, essentially, and that could be accompanied by reduced virulence or transmissibility in humans."

A good chance of transmissibility
Richard Webby, PhD, offered a more nuanced view of the issues in an interview this week. He is director of the WHO Collaborating Center for Studies on the Ecology of Influenza in Animals and Birds at St. Jude Children's Research Hospital in Memphis. Like Racaniello, he said he'd prefer to see the full versions of the Fouchier and Kawaoka studies published, though he acknowledged the concerns about them.

On the transmissibility question, Webby said, "Personally, my feeling is that if a virus transmits readily in a ferret model, there's a pretty good chance it will do the same in naive humans, keeping in mind that the only thing we can't mimic particularly well in ferrets is the preexisting immunity" that humans have by virtue of previous exposures to flu viruses and vaccines. The effects of that population immunity are unpredictable, he said.

"On the other side of the coin, pathogenicity is a little less clear to me, whether something that causes severe disease in ferrets would necessarily cause severe disease in humans," Webby said. "As far as I'm concerned, the jury is still out on that."

As for the possible effects of passaging an H5N1 virus in ferrets, he said it's "certainly a possibility" that the virus would become less virulent. "Viruses that have been passaged a lot in mice do tend to lose some of their pathogenicity for some of the [other] models, but I don't think we know enough about what passaging in a ferret would do to human infectivity or virulence."

Lowen voiced a similar view. She concurred that serial passaging of a virus in an animal host is a "classical" way of making a strain that's attenuated for humans. "I would agree that often and probably usually serial passage in an animal host leads to attenuation in humans," she said.

But she added, "Whether this is true of ferrets is not clear, and there may (possibly) be a difference since ferrets are naturally very susceptible to human influenza viruses; in other words, an influenza virus that is highly adapted to growth in humans will generally also show a high fitness in ferrets. Again, whether the reverse is true (ie, a ferret-adapted virus is well adapted to humans) is not clear."

In response to a question, Webby acknowledged another possibility: Since the case-fatality rate for H5N1 illness in humans is astronomical at about 59%, even an attenuated strain could be dangerous. "That would be fair to say. Even an attenuated H5N1 virus would certainly be equivalent to at least the 1957 and 1968 pandemics," he said.

Lowen observed that uncertainty is inescapable when it comes to interpreting the results of animal experiments. "There is always a caveat with any animal model: Animals are not humans and for a given situation we can't know for sure that what is seen in an animal model is predictive of what would be seen in humans until the same experiment is done (eg, by nature) in humans."

See also:

Dec 20 CIDRAP News story "US government urges journals to omit details of H5N1 studies"

Nov 17 CIDRAP News story "H5N1 transmission experiment stirs concern"

September review in Disease Models and Mechanisms on the ferret as a model for human flu

Dec 9 Journal of Infectious Diseases abstract by CDC authors on local immune responses to flu in ferrets

2007 Journal of Virology report noting neurologic signs of H5N1 in ferrets

2003 Avian Diseases report of neurologic signs of H5N1 in mammals

Dec 6 Vincent Racaniello blog post

Nov 28 CIDRAP News item on NSABB review of the two studies

This week's top reads