The new paper on transmissibility and reassortment in PNAS

The long awaited reassortment experiment where the 1997 H5N1 virus was combined with a seasonal strain of human H3N2 virus is being reported today in the Proceedings of the National Academy of Sciences (PNAS, known colloquially amongst scientists as "penis"). As of this writing I was not able to get a .pdf so all I know is what is being reported on the newswires as a result of a press conference held by CDC Director Gerberding and a co-author of the study, Dr. Jacquiline Katz of CDC. It's hard to know the real significance of this work until we see the actual paper but here is the synopsis as provided by reporters.

One of the big concerns about the evolution of avian flu is the possibility it will reassort with a human strain when it co-infects a cell in a human or some other animal like a pig capable of being infected by both human and bird viruses. The influenza A virus's genetic material is not all in one piece but divided into eight segments, which encode of 11 proteins. Each of the segments can suffer a genetic change (heritable alterations of genetic material, called mutations); or the segments from one virus can be shuffled so that a new virus is produced that has some segments of one virus and some of another. This is called reassortment. Reassortment is thought to be the way the pandemics of 1957 and 1968 got started, whereas the 1918 virus is now thought to have arisen by changes within some of the segments which allowed a virus that formerly infected birds to infect humans.

We have only a hazy idea of the genetic changes needed to make a bird virus a pandemic virus. One of the possibilities is a reassortment with a circulating human strain. Using a technique called reverse genetics, now being used for rapid production of vaccine seed strains, the CDC researchers recombined the segments in the H5N1 and H3N2 virions to produce new viruses with various combinations. I figure there are 62 possible ways to produce a new H5N1 this way (i.e., keeping the H5 and the N1 segments and swapping out various combinations of the other six segments but not counting the two originals). It was not clear from the news reports how many of those combinations were tried or if all of them were. The paper should be interesting in revealing what were the permitted combinations, at least for these two strains of H5N1 and H3N2. Once a replicable virus segment combination was obtained it was tested in ferrets(thought to be a good model for human influenza) to see if it became more easily transmissible. One report also said the ability to cause severe disease was also investigated and it implied none of the combinations was able to do so. Since H5N1 does cause severe disease, this also implies the virulence of the virus was moderated in the reassortants, but until we see the paper we don't know what was meant by this.

It sounds as if the press conference announcing these results was relatively free of spin, although there was plenty of headline induced spin in the reporting. An example of a more or less accurate headline comes from Randolph Schmid's AP report: Flu Viruses Don't Transmit Easily in Test. An example of a poor headline can be found atop Amanda Gardern's HealthDay report: Bird Flu Pandemic May Not Develop: Virus doesn't show signs of mutating easily, U.S. government researchers now say. Despite the atrociously misleading headline, Gardner's article is among the best and most complete. She's not responsible for the headlines and was badly served by this one.

I still have my doubts the paper can sustain the claims being made for it, particularly that this experiment shows it isn't easy to change a virus from one not easily transmissible to one that is and that the process is not simple but complicated. Marc Siegel (who sounds pretty sane on the subject of bird flu these days) is again quoted as saying it will take multiple mutations for this to happen so a pandemic is not imminent. This may be true. But it may not be and this experiment does not shed much light on that question. It can plausibly be said to show that such an outcome is not a frequent result of a reassortment with a human flu subtype, and that's information worth knowing. And while it doesn't say that some so far untested reassortant might not do it, at least it says it doesn't happen a lot and in many different ways. Since we now know reassortment is much more common than previously thought, this is an important point.

On the other hand, without knowing the factors that make a virus easily transmissible it is unwarranted to say this experiment shows it is either difficult or complicated. This virus is capable of amazing parlor tricks and parlor tricks often look impossible until you see how they are done. Then they look simple. Until we have a better picture of how the virus does its various tricks we won't know if it is simple or complicated, easy or difficult, likely or unlikely. So this is a step forward but a relatively small one in terms of the distance we need to travel.

Addendum: Someone just sent me a .pdf of the paper which I'll try to read tomorrow and update if there is any more of interest to report.

More like this

I mentioned just last month the dearth of research into what really makes an influenza virus easily transmissible from host to host, noting: If we had a better handle on the factors that caused an avian strain of influenza virus to be more efficiently transmitted among humans, then we could better…
I have now had a chance to read the PNAS paper by Maines et al. and it is surprising in two respects. The first is it isn't that interesting. The second is related to the first. Why did they bother to hold a press conference about it? Even more, why did the press conference focus on the…
Last week another expedited-review paper appeared in the high profile journal Science, this one summarizing the genetics of the novel H1N1 influenza A virus causing the current outbreak cum pandemic. This time there is quite a bit of interesting material in this paper for non-virologist scientists…
When influenza viruses with different genetic make-up co-infect a cell there is the possibility that they will mix their genetic endowments. The influenza virus is designed to do only one thing: make a copy of itself. It does this by tricking the host's protein manufacturing machinery to use the…

After reading your discussion, I went to Recombinomics. This is the last sentence of Niman's comments on the PNAS paper:

"These data [meaning not only the difficulties of reassortment but also his analysis of the four extant H5N1 clades] should focus attention on the changing, not swapping of genes. The changing of genes is driven by recombination."

Can you offer a critique of Niman's recombination theory? Or, if you've already done so, provide a link?

slovenia: I will try to read the PNAS paper carefully today. As far as I can see they only tested 4 combinations by reverse genetics, so it doesn't say anything about how likely reassortment is in the wild or what kind of combination would be produced. We know that within subtypes there is a lot of reassortment, so I'm not sure why this shows anything much about the propensity of reassortment in H5N1, but I'll think about it some more.

Of note in the paper (which has data that was discussed in 2004) is the fact that although there are no human genes or changes in the receptor binding domain, the human isolate from Indonesia (5/05) grew to HIGH titers in the UPPER respiratory tract of infected ferrets.

So I understand the titers level but for the layman most will think this is a good thing. Viruses for Dummies-Only if its response to vaccine. What are you saying Henry? Is the titer level high because of the infection they induced without human sequences?

I read the PDF and it was pretty weak as best I could tell and a lot of holes in the results. Doesnt look to me that it could be used as a basis for making the statements that are being kicked out as the Gospel according to Gerberding.

By M. Randolph Kruger (not verified) on 01 Aug 2006 #permalink

The paper showed that combinations of whole genes from humans and birds wasn't very effective. 8 gene segments from one species or the otehr either grew or transmitted more efficiently. Most of the data was on 1997 and this data was generated years ago (Nancy Cox indicated the results were not that interesting and had not been published when asked - I think in 2004).

The new paper is a slight upgrade of the old data because some initial studies with one 2003 and two 2005 isoaltes were described. One of the 2005 isolates was from Indonesia. The H5N1 wih no human genes killed the ferrets, but also grew to high levels in the UPPER resiratory tract. However, the ferrets with the H5N1 locked and loaded in their noses didn't sneeze, so there was no transmission.

Thus, a pandemic strain was one sneeze away. The changes required to cause sneezing are not known, but there is no data indicating that human genes are required for sneezing. The same is true for receptor bindng domain changes. The H5N1 from Indonesia has an avian receptor binding domain sequence, yet it grew to high levels in the UPPER respiratory tract of ferrets.

Thus, H5N1 in Indonesia has evolved WITHOUT human genes or changes in the receptor binding domain to a pint where it is one sneeze away from a pandemeic.

Thus, the "significant changes" (reassortment with human or swine genes or changes in positions 226 and 288 in the receptor binding domain), do not appear to be all that important for H5N1.

Maybe all that is required is H5N1 in a person with an allergy for efficient transmission.

WHO is holding the huamn H5N1 sequences and looking for non-events, which is NOT reassuring.

The following quote is in an earlier commentary

http://www.recombinomics.com/News/03230502/CDC_H5N1_Mix_Again.html

"It may be that we find out that it's really not as scary a situation as one would think," Dr. Cox says. "Or we may find some gene combinations work well together." Similar CDC experiments using an H5N1 virus from an outbreak in Hong Kong in 1997 didn't produce "sensational" results, she says. Details of that research are still being written up for publication.

Understandably, most of the interested public's initial discussion of this paper seems to be about the transmissibility of the reassortants, and how to interpret that with regard to the likelihood of H5N1 "going pandemic."

But is it -- or is it not -- an important paper with regard to the "caged ferret" transmission model? Is this of more interest to scientists than to the public? Or is this model either old news, or not especially useful?

Revere, I hope in your reading of the paper, you will address this.

By Path Forward (not verified) on 01 Aug 2006 #permalink

Maines et al state that their "use of the term respiratory droplet transmission ... does not imply an understanding of the droplet size involved in virus spread between ferrets.'

Apparently not.

In support of "droplet" (as opposed to "airborne") transmission of influenza, they cite a review paper that actually provides more evidence for the latter.

In terms of the experimental set-up, they housed their ferrets within a Duo-Flo BioClean unit; this includes a HEPA filtration system and "is usually operated in the positive mode to protect experimental animals" [Am Ind Hyg Assoc J. 1992 Feb;53(2):138-45] i.e. to limit airborne exposure.

And, having having made the point that H5N1 has an affinity for binding sites in the lower respiratory tract of humans and ferrets, decided to use nasal [i.e. about as "upper" as you can get] washes / titres as their index of exposure / transmission. Why not broncho-alveolar lavage? (Two of their 3 HK486s contact ferrets sero-converted despite the absence of detectable virus in their nasal wash.)

So what have they assessed?

Is a bird version of the 1918 virus available?

Has anybody run experiments essentially identical to the above except using the 1918 bird virus instead of today's bird virus? And if so, did the experiments produce a readily transmissable virus?

Like Slovenia, I am not clear on the difference between reassortment and recombination. Hopefully, Dr. Niman or Revere will spoon-feed me...

Although reassortment and recombination have serveral aspects in common, there are some major differences. Flu has its genes segmented onto 8 pieces of RNA. As a result those genes can get shuffled when one cell is infected with two different viruses. Recombination also requires a dual infection, but the recombination involves only portions of genes. That happens when a gene is used as a template for new RNA and the new RNA hops off template 1 and then falls onto template 2. For this to happen, the two viruses have to have regions of identity and the hopping can happen where the two sequences are the same. This process is called homologous recombination, because a region of homology is required.

Recombination creates a new gene, with some genetic information from one virus and some from another. Thus, for recombination all eight gene segments could get changed in one dual infection. In reassortment, the virus may have a new set of genes, but each gene will match one of the two parents.

Reassortment is pretty easy to observe becaise the whole gene is swapped. For recombination, the amount of genetic information acquired can be as small as a single nucleotide.

As far as H5n1 is concerned. everyone is in agreement that to date there has never been a reported isoalte with a human gene. WHO and consultants are waiting for such a change. In teh PNAS paper, sucg chnages were introduced artificially. These reassortants however were not as effective as either parental strain (one with all 8 gene segments from either and avian source or a human source.

However, H5N1 has evolved by chnaging its genes. Some think these chnages are just random mutations that can't be predicted. I think they are just more examples of recombination, where most of the genetic information exchnaged is the same and only single nucleotides are different.

This is supported by the fact that the new infromation usually can be found in isoaltes that would be expected to be involved in dual infections. Frequently the new sequences in H5N1 in wild birds are the same changes in other wild birds, but the combinations change. Other times they are from H5N1 isoaltes from a year or two earlier in areas where wild birds frequent.

In other cases the H5N1 picks up mammalian sequences, but again they can be found in mammals (usually pigs) and in many instances from pigs are in specific regions.

In any event, WHO and consultants are good at identifying reassortment, which has never happened with H5N1 and human genes, but they are not good at detecting recombiantion, and in fact maintain that recombiantion rarely happens. In instances where recombiantion is obvious, they either ignore it or in some cases withhold parts of sequences when recombination is expected. In the more common tyes of recombination, where the two strains are closely related, the acquired information is said to be due to random mutations.

In any event, WHO is looking for reassortment, and the CDC experiments provide data indicating that such changes are highly unlikely because they do not offer an advantage in replication or transmission.

I took the liberty of translating your post into Brazilian - Portuguese and posting it in my Blog, for the benefit of non-Englhish readers.

I hope you don't mind.

Dr. Niman: You seem to be confirming what I have feared all along, that WHO has predetermined a certain set of circumstances - Reassortment of H5N1 avian genes with human genes - as critical to determining a human pandemic is imminent and raising the alert level. They apparently used this criteria with the Karo outbreak: since they found no reassortment, they decided it was not a significant change even though the virus had definitely achieved H2H transmission in a chain of at least 3. This is just plain bad science; predetermining what the expected outcome is and closing itself to all other possibilities, especially in view of what you and others have just enlightened us on regarding the fact that the most devastating pandemic in history, the one H5N1 has already shown itself to be most similar to in many ways, became one through recombination rather than reassortment. What is most worrisome is that WHO will completely fail to warn the world until the pandemic is already well past the point of containment because of this blinders-on mindset.

By mary in hawaii (not verified) on 01 Aug 2006 #permalink

juan carlos: I am delighted. We have a Creative Commons license on the site so all is free for the taking (with attribution). There is another post tomorrow on the same subject.

mary: Henry has his take on WHO's motives on this that are his own and not shared by us. There is genuine disagreement about the extent of recombination and interpreting the data is not straightforward (he acknowledges some of the difficulties in his explanation). Anyway, it is not my view that WHO is deliberately covering up the existence of recombination and I wanted to say that clearly. Henry is entitled to his views, as are we.

Revere: I didn't mean to intimate in any way that WHO was "covering up the existence of recombination" and I'm sorry if my wording made it sound that way. I was saying that they are downplaying its importance as a factor in increasing transmissability of the virus among humans, and instead are focussing on evidence of reassortment as the key parameter for deciding the virus has mutated into a H2H form that warrants raising the alert level. It seems that, short of a sudden rapid H2H spread that goes beyond a familial cluster, they will focus on this genetic clue. It may be valid, or it may not. We have discussed at length the wisdom of authorities testing humans for H5N1 based on the criteria that there must be dead birds involved to warrant such tests. The consensus seemed to be that placing such limits was dangerously shortsighted considering we have a rapidly evolving virus and very little knowledge about it. I am saying similarly to presume that reassortment must be a factor in H5N1's evolution to a human pandemic might be also dangerously shortsighted.

By mary in hawaii (not verified) on 01 Aug 2006 #permalink

mary: WHO also is concerned with point mutations, so they are not just looking for reassortment. Whether they are right to downplay recbomination only time will tell. Henry is very definite but most virologists don't believe there is much evidence of recombination in negative sense RNA viruses like influenza. I'm not going to get into this argument as it doesn't interest me much. I wasn't scolding you for saying WHO was covering up, just making clear my own view of the matter which is different than Henry's.

Ferrets? I'm going to need some help translating this for my readers. The reassortment/recombination part I don't care about so much. What is the bottom line for Jane Average?

Melanie. The bottom line is that they should not have released an incomplete study unless there was a pressing need to do so (ex. the longterm hormone replacement study)...there was no pressing need.

I have no doubt that H5N1 and Nature are chuckling at the conclusions.

Revere, Please refrain from posting what you think I think. I have not said that WHO is covering up recombination. I have said that WHO consultants have published partial sequences, and the missing data contains areas where recombination is likely, based on more complete sequences. There are many reasons for the missing data. The primers may not have reacted well in the recombined region. The sequencers may have not thought the additional sequence data was important. They may have not wanted to interpret the "anamolous" result.

Similarly, WHO consultants have published papers with obvious recombination which was not mentions. They may not have thought it was important, may not have wanted to run experiments to look at its importance, or may not recognize obvious recombination because they have not looked for it.

WHO consultants have also indicated that they looked for recombination but could not find it. However, it is unclear on how hard they looked, if they would recognize it, or had some preconceived expectation of some frequency of recombination that was not met.

I realize this is not your area of expertise or interest, which is why I fail to understand you have a need to constantly tell others what I think about recombination. Similarly, you have commented that neither you nor I am a virologist - you are half right - my training and background are in virology and I have published in Nature, Science, PNAS, Cell, Journal of Virology, and Virology on various viruses, including influenza.

Tom DVM, Much of the data in the PNAS report was generated several years ago. It contained data of combinations that the author's thought were most likely to increase the growth or transmission of H5N1 reassortants.

The data indicated that the combinations used decreased replication or transmissions. It is not necessarily the final word, but is consistant with the lack of any reported reassortment between H5N1 and mammalian genes.

The data presented can be interpreted by readers of the article, and the data for the examples reported were quite clear.

WHO still reports failures to find reassortment. The PNAS paper helps put those pronouncements into proper perspective.

Henry: Thank you for clarifying your views on WHO. I didn't mean to misrepresent you, although I hope you can see why the way you phrased it made such a conclusion reasonable.

revere, I usually draw distinctions between WHO and its consultants and also usually lay out the facts and let the reader interpret the data.

On recombination, WHO is pretty much dependent on what their consultants tell them, and those consultants are much more interested in reassortment than recombination.

Similarly, for the term "cover-up", I generally only apply that to a situation that is blatant.

I would use cover-up on the withholding of disease onset dates and the relationship between family members in the Turkey outbreak at the beginning of this year. WHO personel issued comments initially indicating several members of two large families were involved. However, when those family members were confirmed to be H5N1 positive, disease onset dates were withheld, and the facts that the two families were cousins and had been at a family gathering were withheld.

The tactics worked quite well, because most still don't realize how large the cluster was in Turkey, or how extended the transmission chain was. Thus, the withholding of that information could be characterized as a cover-up.

Niman: "~ lay out the facts and let the reader interpret the data"

Unfortunately, your performance at this is not good. Much improvement could be had by simply rereading your posts, correcting keyboarding errors and improving the clarity; let's call it respect for the data, the analyst (often yourself) and the reader.

Come on Tympanachus, that was a tad meeeooowww... Surely you must be aware of the massive hourly influx of data folk like Dr Niman are required to multitask digest, interpret and then publicly post within many forums. Keyboarding errors show him as just a human whizzing along keeping up with the daily H5N1 nitemare -- deal with it, honey!

By Jon Singleton (not verified) on 02 Aug 2006 #permalink

I'm a Niman fan; been reading him for a couple years now. I hope he's right and is finally well compensated for his insight and diligent prosecution of his case.

His writing sucks. It really has little to do with keyboarding errors.

By tympanachus (not verified) on 02 Aug 2006 #permalink

In an earlier post Henry Niman states "Similarly, you have commented that neither you nor I am a virologist - you are half right - my training and background are in virology and I have published in Nature, Science, PNAS, Cell, Journal of Virology, and Virology on various viruses, including influenza."

However, a search of ISI Web of Science from 1970 to the present using the author Niman does not come up with a single research article on virology or one that was published in ANY of the journals mentioned. In fact the most recent publication by an H Niman was in 1996 and all the papers were cancer related.

So perhaps Niman could post the citations to this paper to verify his claims?

It is possible that when Niman said "published in" he did not mean original research but was referring to correspondence, like a letter to the editor. Well that is not being "published in".

Another thing is that Niman has been talking about the recombination/reassortment issue for years, mainly with the argument that once the WHO releases sequence data he can prove everything.

Well, 3 years down the line and there is plenty of data available, including human sequences from China amd Vietnam, but still nothing has been published. As Niman is always telling us about specific recombination events he must have enough data to PROVE his theory, so why not shut up about it and get it published? There are many journals that would jump at a chance to publish such research.

In science publishing is the way to gain acceptance, otherwise it is just a bunch of unproven hotair. And when someone has not published for years, keeps on going with the air, and fails to publish years after the start of the talk, then likely as not they have nothing but a strong desire for publicity.

Anon2. Choosing not to publish in scientific journals does not make Dr. Niman's hypothesis wrong however.

Dr. Niman is in a public forum and his data is clearly avaliable for peer review...whether his peers choose to do so is another matter however...they appear remarkably silent on the issues raised.

C'mon folks, ease up on Henry. What's more important? How he spells/writes, or the speediness of his reports and the information he provides us? He's been ahead of the game in most emerging stories of H5N1 activities, and while I realize that subsequent confirmation is necessary before jumping to conclusions, I still appreciate his "hot-off-the-press" machine translations and follow-up interpretations (thanks, Henry!) to let me know what's brewing. A few misspelled words are insignificant.

I don't understand anon2's attack on Dr. Niman. His website gives plenty of links and data that one can easily check to ascertain his professional credentials and expertise. He has 9 patents in his name for research technologies he's developed related to cancer research, as well as one pending that was filed in June 2005 for his current research into viral recombination. The fact that he shifted his research emphasis is not any cause for question: Francis Crick shifted from DNA research to the neurochemistry involved in learning in later years. So what? If he has a patent pending on his more recent line of research, he may want to avoid publishing papers until the patent is through that process in order to protect his IP interests. I also don't understand why anything he has said would cause even a dispute. He has only published data about verified changes in genetic sequences of H5N1 as it travels and mutates, most often not doing more than interpreting that data for us with minimal opinion on its meaning. I think that since WHO and other scientists seem to be focussing more on reassortment as an indicator of potential H2H transmissability, it is of great benefit to have this other side presented.

By mary in hawaii (not verified) on 04 Aug 2006 #permalink

Yes Mary, but the reason he supposedly wants the nasty WHO to release the sequences is so that he can save the world. So holding back because of a patent or whatever is just as bad as withholding information while waiting publication.

And to TomDVM and others, it is impossible to ascertain what Niman is talking about as he has never presented a methodology. Detecting recombination on a small scale is problematic and would probably require a new approach/methodology. As far as I can tell from his website, Niman does not mention any methodology, just listing nucleotide differences. But just the presence of a nucleotide change does not provide any information regarding whether it arose through mutation or recombination.

So yes, Niman makes his contentions in public but there is no way to assess his theory as there is no data available.

Influenza is a negative stranded RNA virus with no proof reading mechanism when copying of genetic material during virus replication. This means that influenza has a high mutation rate, even among viruses.

Also, the segmented nature of the influenza genome makes it less likely that recombination would occur. Not that it is impossible, just that it has not been frequently identified. This is mostly due to the problems with the methodology for detecting recombination.

Then there is the issure of quasi-species that Niman has never addressed. Imagine that a drop of influenza virus contains 1 million virus particles. Because of the high mutation rate, theoretically, each vurus particle could be genetically distinct, even if by just a single nucleotide. Therefore, what we see as a point mutation could in fact be a change in the proportion of virus particles with that particular nucleotide as it adapts to its host.

So as I said you would need a new method. If Niman has a new method, then great, I would love to be proved wrong as it would be a great help to my work, and truly revolutionary. But there is absolutely nothing about a methodology on his website, the best you get is an statement that "influenza evolves through recombination" which has never been shown in the literature.

This leads me to conclude that he is full of hot air and using this issue purely for publicity.