Influenza H5N1 x H1N1 reassortants: ignore the headlines, it’s good science

7 May 2013

Those of you with an interest in virology, or perhaps simply sensationalism, have probably seen the recent headlines proclaiming another laboratory-made killer influenza virus. From The Independent: ‘Appalling irresponsibility: Senior scientists attack Chinese researchers for creating new strains of influenza virus'; and from InSing.com: ‘Made-in-China killer flu virus’. It’s unfortunate that the comments of several scientists have tainted what is a very well done set of experiments. Let’s deconstruct the situation with an analysis of the science that was done.

It is known that avian influenza H5N1 viruses can occasionally infect but not transmit among humans, while the 2009 pandemic H1N1 virus (which continues to circulate) readily transmits from person to person. The investigators asked whether reassortants of the two viruses – which could arise in nature – might confer transmissibility to H5N1 virus. To answer this question they produced 127 different reassortants of the two viruses, and tested their ability to transmit by aerosol among guinea pigs. The latter have been used for transmission studies on influenza, notably to understand the seasonality of infection. Ferrets have been more famously used for influenza virus transmission studies.

Rather than describe the results, I’ve made an illustration that shows what I believe to be the most important conclusions of the study (click for a larger version):

h1n1 h5n1 reassortants

The H5N1 virus (red RNAs) is not transmissible among guinea pigs, while the H1N1 virus (green RNAs) has highly efficient transmission. Exchange of the H5N1 RNA coding for PA or NS from H1N1 produces a highly transmissible virus. Exchange of the H5N1 RNA coding for NA or M produces a less efficiently transmitted virus. These are interesting and novel findings. It will be of great interest to determine how the PA, NS, NA, or M genes mechanistically enhance aerosol transmission. This is important information because our understanding of the determinants of transmission is very poor.

All the reassortant viruses shown in the figure have the H5 HA; when only the H1 of the H1N1 virus was substituted with the H5 HA, the reassortant virus transmitted efficiently among guinea pigs. In ferrets the H5 HA is not compatible with aerosol transmission. Therefore guinea pigs are clearly different from ferrets with respect to the determinants of transmissibility.

I cannot understand why some scientists have called these experiments ‘appallingly irresponsible’ and of no scientific use. I can only assume that they are not familiar with the literature on viral transmission and do not appreciate how the results advance our understanding of the field. It also seems irresponsible to predict that these viruses, should they escape from the laboratory, could kill millions of people. If you accept guinea pigs as a predictor of human pathogenicity – which I do not – then there is no reason for fear because none of the reassortants were lethal. I do not believe that any animal model predicts what will occur in humans, and so I am even less concerned about the safety of these experiments. I firmly believe that laboratory-constructed viruses do not have what it takes to be a human pathogen: only viral evolution in nature can produce the right combination of RNA segments and mutations. I also believe that scientists are quite responsible when it comes to safe handling of pathogens. If we worry about every type of transmission experiment involving influenza H5N1 virus, we will never make progress in understanding why this virus does not transmit among humans. The moratorium on H5N1 transmission research is over; let’s move beyond the sensational headlines and get back to the science.

In summary, I believe that these are well designed experiments which show that single RNA exchanges with H1N1 virus can produce an H5N1 virus that transmits via aerosol among guinea pigs. The relevance of these findings to humans is not known; nevertheless understanding how the individual viral proteins identified in this study enhance transmission may be mechanistically informative. I believe that the news headlines depicting these experiments as irresponsible and dangerous are based on uninformed statements made by scientists who are not familiar with the literature on influenza virus transmission. I wonder if they even read the paper in its entirety before making their comments.

  • gsgs

    > I firmly believe that laboratory-constructed viruses do not have what it
    > takes to be a human pathogen: only viral evolution in nature can
    > produce the right combination of RNA segments and mutations.
    you will hardly find an influenza scientist who joins you in that belief.
    (ask them !)

  • Pingback: Influenza H5N1 x H1N1 reassortants: ignore the headlines, it’s good science | The Great Plagues

  • Kenzibit

    You are right but was thinking, What if truly they were specifically constructed to be human pathogens?

  • DavidS

    It may very well be good and important science. The following statements are VERY troubling, however.

    “It also seems irresponsible to predict that these viruses, should they escape from the laboratory, could kill millions of people. If you accept guinea pigs as a predictor of human pathogenicity – which I do not – then there is no reason for fear because none of the reassortants were lethal. I do not believe that any animal model predicts what will occur in humans, and so I am even less concerned about the safety of these experiments.”

    I’m sorry, but these sentences are ridiculous. If it is true that it is wrong to believe that “any animal model predicts what will occur in humans,” why are we conducting animal experiments? It is silly also to suggest that humans cannot accidentally create threatening pathogens.

    “I firmly believe that laboratory-constructed viruses do not have what it takes to be a human pathogen: only viral evolution in nature can produce the right combination of RNA segments and mutations. I also believe that scientists are quite responsible when it comes to safe handling of pathogens.”

    Your beliefs are not relevant. What is your EVIDENCE? It may be true that viral evolution in nature is more efficient, but it may be possible for human technology to create a virus that is given a head-start on that evolutionary pathway. Your beliefs about scientists are belied by the anthrax case and others. It could well be that most scientists are quite responsible. Some may not be.

  • Megan

    Well said! As a scientist I am often appalled at the science headlines and even more so with the uninformed comments left on the article forums. Scientists as a whole are responsible and doing great work to advance the field. The conspiracy theorists need to see the whole picture and stop making assumptions that there are rogue intentions!

  • Josh P.

    Kudos…I agree with your Vincent, this is quality science and much more relevant than the gain of function experiments done last year in ferrets. I will tell you this, the PA segment of H1N1 is a nasty bugger. I find this segment increases the replication fitness within seven segments of H3N2 swine over the parent strain in both human and swine cells when I make reassortments. The more alarming scenario that isn’t mentioned is that multiple genome segments of pH1N1 are circulating in the swine population right now..including pH1N1 PA. However there is a little surveillance data on internal influenza genes as mostly just sequencing is done with NA, HA and sometimes M. I hypothesize based on my co-infection reassortment studies in swine (Hopefully some upcoming animal data) and data in the influenza online database (although limited) that the PA gene segment of pH1N1 will be an issue for years to come. I bet if you took H7N9 and replaced the PA gene with pH1N1 you would get a very nasty virus…this could very well happen in pigs. 50 million pigs in the H7N9-infected region, knowing how reassortments behave is VERY important.

  • Magpie

    And not just pigs! A quite enormous number of humans, too. Given the length of time the H7N9 infections are taking people to shake off, and the likelihood that there are a large number of asymptomatic carriers of H7N9 running about, I reckon we’ve got a real possibility of seeing some reassortment events come the northern-hemisphere’s next flu season.

    My money is still on “won’t happen this time, and if it does happen, it probably won’t be too bad” – but not ALL my money. I do think it’s fair to say that the next year or two has a much higher than normal chance for a pandemic to emerge. Understanding the specific changes required to form a pandemic represents our only decent shot at shutting it down before it gets away from us – or at least implementing early and effective delaying measures to buy time for vaccine development and production.

  • josh

    The only curve ball there is that humans usually don’t harbor avian flu….pigs readily do. That means, what’s more likely…a human co-infected with two flu strains that allow for robust infection and reassortment or pigs? In the case of H7N9…I put my money on pigs being the thing “we missed” if this thing takes off. I don’t think it will, but never underestimate the 50 million livestock reservoir in the region.

  • gsgs

    I’m not aware of pigs harboring avian flu.
    There are some events …i.e. with H9N2 …
    but still no evolution of avian flu in pigs.
    Maybe they get it more often than humans, OK.

  • gsgs

    also my first thoughts. It sounds ridiculous, self contradicting, (maybe misunderstandings ?).
    He will be attacked for it, he should have got it reviewed,discussed
    before posting (IMO)

  • EllenHunt

    Agreed. Not only that, but to the extent there is an issue, those yelling about irresponsibility are the ones creating a problem.

    http://www.omicsonline.org/2157-2526/2157-2526-S3-013.php?aid=11953
    Security in a Goldfish Bowl

    Abstract

    If our potential adversaries are understood, publicity drawing the attention of bioterrorists to plausible bioweapons is the primary contribution the NSABB makes by censorship recommendations. Attempting censorship by attracting the entire world’s attention to information of concern is obviously counterproductive.
    It is equally problematic to curb research in an attempt to prevent those outside our borders from understanding how to create bioweapons. That horse is, “Out of the barn and in the next county”. Locking the barn door is not going to change this. Proliferation of molecular biology, vaccine and other knowledge across the world has already taken place. Thus, the best course is to make sure that we understand what human bioterrorists could do as well as we understand what nature can do. …

  • Magpie

    There’s also the likelihood that a human-native reassortment that survives will (almost necessarily) be fitter in a human population – even if hopefully less harmful.
    One asymptomatic human case of H7N9 means, to me, a LOT of asymptomatic cases, given the likelihood of finding any given case (from my reading of it, they’re still relying mostly on clinical follow-ups, with PCR for only the closest of contacts). So if it’s still infecting people come flu season, then there’s a chance of coinfection (though I’m profoundly ignorant about how often that actually happens – flu’s weird), and then… we’ll see.
    ‘Course, “more likely for a pandemic than usual” may mean (hand-wavey numbers) that we go from a 3% chance to a 6% chance this year. I ‘aint stocking up on baked beans just yet!

  • Pingback: Credenze « Oggi Scienza

  • Pingback: Credenze – Ocasapiens - Blog - Repubblica.it

  • http://www.virology.ws profvrr

    Animal models are not predictors of what will happen in humans. We study infections in animals to obtain mechanistic clues; conclusions about what happens in people require further testing. There are dozens of examples of this in the literature. Drugs and vaccines are tested in several animal models, but are the drugs and vaccines then released for humans? Of course not; clinical trials are done.

    We’ve had a great deal of experience in this laboratory constructing viruses with various mutations, and none have ever been more virulent; most are attenuated. The same applies for research done in other laboratories. We have little clue how to make a more virulent virus. Part of the problem is that we focus on amino acid changes in isolation; in nature these are accompanied by hundreds of other changes which are eventually selected in various hosts to make the final pathogen. We don’t have a chance at duplicating this and I’ve never seen any laboratory come close. Just take a look at the scientific literature on viral pathogenesis.

    My beliefs are indeed relevant. They are based on 30+ years of doing research in the laboratory on viral pathogenesis, and keeping up with the literature.

  • http://www.virology.ws profvrr

    I can find plenty of virologists who would have the same view. We are talking about humans, not animal models.

  • gsgs

    profvrr wrote:
    > I firmly believe that laboratory-constructed
    > viruses do not have what it
    > takes to be a human pathogen: only
    > viral evolution in nature can
    > produce the right combination of RNA
    > segments and mutations.
    > … I can find plenty of virologists who
    > would have the same view.
    > We are talking about humans, not animal models.

    laboratory-constructed viruses do have what it takes
    to be a ferret pathogen ?
    Humans are animals. We are not so special from
    the virus’ POV. Why might it work in different animals
    but not in humans ?
    Presumably only because we can’t test it in humans
    for ethical reasons. There is no magical difference
    that separates humans from ferrets but not ferrets from
    mice or guinea pigs. We cannot be 100% sure
    before we tested it in humans – but maybe 80%,
    and that’s still useful. So lab-viruses probably often
    do have “what it takes to be an animal or human pathogen”
    but tests to confirm this for humans are rarely available.

  • gsgs

    [ I also put it here :
    http://gsgs2.blogspot.de/2013/05/man-made-pandemics.html
    and
    here:
    http://www.flutrackers.com/forum/showpost.php?p=497438

    for better formatting and possible later editing and additions and
    graphics
    and responses and discussion ]

    ——————————————————————————————————–

    profvrr wrote on 2013.May.13 in response to DavidS

    > Animal models are not predictors of what will happen in humans.

    somehow they are.That’s why we are using them, why you say “it’s good
    science”
    You probably meant: are not so good predictors (as most
    think)

    > We study infections in animals to obtain mechanistic clues;

    you say mechanistic clues, I’d say probability estimates

    > conclusions about what happens in people require further testing.
    There are dozens
    > of examples of this in the literature. Drugs and
    vaccines are tested in several animal models,
    > but are the drugs and
    vaccines then released for humans? Of course not; clinical
    > trials are
    done.

    which often however confirm the animal tests. I’d guess in ~70% of
    cases

    > We’ve had a great deal of experience in this laboratory constructing
    viruses with
    > various mutations, and none have ever been more virulent;
    most are attenuated.
    > The same applies for research done in other
    laboratories. We have little clue how
    > to make a more virulent virus.

    but there are many papers demonstrating just this increase of
    virulence.
    I don’t know much outside influenza, but especially with
    reassorting influenza viruses :
    we have 256 possible children, some are
    usually
    more virulent than the parents, some are less virulent. Only ~10% do
    survive and replicate
    and most transmit worse than in nature. But now I feel
    that we are getting close to successfully
    create/detect and filter the
    potential candidates. We are clearly better than nature here,
    which requires
    a (rare in humans) double-infection and competition (through immunity) with
    other viruses,
    while we can learn and design and select to specifically
    target and optimize the outcome.
    Since above study is not public and already
    starts with a HP-virus,
    let me select this as an example instead, to
    demonstrate what I mean Sun et.al., 2011:

    [url]http://www.pnas.org/content/early/2011/02/23/1019109108[/url]
    [ For
    an overview of other reassortment studies
    see
    [url]http://www.flutrackers.com/forum/showthread.php?t=203696[/url]
    ]
    While most (46 out of 127) [H9N2+pH1N1)-reassortant viable children had
    lower pathogenicity
    than the parents, there were still 8 out of 127 that had
    higher pathogenicity in mice than both
    parents.Researchers and terrorists
    will learn how to create higher pathogenicity by reassortment
    or passaging or
    combinations of both. And how to enhance it to ferrets,pigs,…,humans.

    This was for reassortment experiments, but also for single mutations and
    passaging
    there are many papers how these may increase virulence.
    Some
    keywords that come to mind:: N66S in PB1-F2, E627K in PB2, H5N1 in mouse-brain

    after passaging, chicken-adapted,quail-adapted H9N2, D225G in HA of pH1N1
    and 1918-H1N1,
    search “virulence mutations”, “influenza”

    > Part of the problem is that we focus on amino acid
    > changes in
    isolation; in nature these are accompanied by hundreds of other changes
    >
    which are eventually selected in various hosts to make the final pathogen.

    in flu it’s often just one mutation. And the mutations usually
    accumulate one by one.
    Each of the viruses in that chain must be viable. We
    have many influenza sequences
    meanwhile, so we can study the chains. And in
    the labs they can create these mutations
    by passaging.

    > We don’t
    have a chance at duplicating this and I’ve never seen any laboratory come close.

    Palese,Fouchier,Kawaka,…I feel they are pretty close

    > Just take a look at the scientific literature on viral
    pathogenesis.
    > My beliefs are indeed relevant. They are based on 30+
    years of doing research in the
    > laboratory on viral pathogenesis, and
    keeping up with the literature.

    but things have changed a lot recently. Now we have reverse genetics,
    better organized
    labs with rooms full of ferret-cages, 200000 flu sequences
    at genbank, 70000 flu-papers at pubmed.
    Next Generation Sequencing

  • Magpie

    I think it’s fair to say, though, that any particular non-human virus has an extremely tiny chance of being harmful to humans. Why would a lab-derived animal virus be so much more likely, virus-by-virus, to be a human pandemic potential? We’re exposed to, god, billions of viruses, types and subtypes, as a species, every day – and even though those viruses are perfectly viable in their animal hosts they have a vanishingly small chance of knocking us off. What, then, are the chances of any particular assortment being immediately viable in a human host?

    Sure we’re experimenting with known human-capable pathogens, but those same experiments happen in nature a gob-smackingly vast number of times every second. For a lab-derived reassortment – whose *only* selective pressure has been the ability to exist in a non-human mammal – for that to turn out to ALSO be well adapted to humans, especially given the extremely small number of test subjects, just seems… minute, to say the least.

    Throw in even the most basic of infection control and surely we’re about as safe from that chain of god-awful luck as we are from any wandering black hole or alien war fleet.

    (That said, our lack of orbital defence platforms is a bit embarrassing).

  • gsgs

    Magpie, replace humans with ferrets or mice or even cell-cultures in your

    post above and tell me, at what point your argument differs.

    Take a random avian virus out of the 5000 avian flu-genomes at
    genbank.

    Reassort it with seasonal flu, say pH1N1(2009), 254 new possibilities.

    Test them in mice. How many % of these 254 do you estimate

    are viable in mice, replicate well and cause disease ?

    How many % would infect and transmit well in ferrets by aerosoles ?

    Many previous experiments tell us that it’s not that tiny amount

    that you suggest. It’s a managable method to create new flu-viruses.

    Indistinguishable from real flu-viruses.

  • Pingback: Further defense of the Chinese H1N1 – H5N1 study

  • Pingback: TWiV 233: We’re surrounded

  • Rohan

    i think H7N9*H5N1 and H7N9*H1N1 reassortant studies can give more insights into whether or not the current H7N9 virus is capable of causing a pandemic. Of course you can test it only with animal models but that ll provide a good head start.

  • Elial Rebollo

    Greetings, not enough to take preventive actions and vaccines our immune system must be 100% in tune …! and that only the transfer factors can. The 4Life Transfer Factor has the property that combat all types of viruses and many more!.

    I leave my site to investigate a little more of this product that can be your salvation when a health problem.

    elialr.my4life.com
    @elialr4life
    elialr@my4life.com
    +00-58-412-8853656