Another "big" science story on the mutations in H5N1 that will a make it a pandemic strain. Same ending as the other stories. Not exactly.
Some of the blame for this rests with the scientists who can't resist going beyond what's in the paper when talking to reporters. I understand. I've done it myself, probably, although I try not to. On the one hand there are scientific conventions that suppress over interpretation in the published report, even when there are plausible speculations about larger meanings. On the other hand, there is the natural tendency to please the reporter, who is not looking for a "maybe this, maybe that" quote.
The paper du jour is by a large international team, published yesterday in Nature. It addresses what is different about the H5 protein of those H5N1 strains that recognize avian cell surface receptors compared to those that recognize human cell surface receptors. If you are interested in the scientific background to this, you can read about it in a four part series we did recently (part I, part II, part III, part IV). The same question was approached recently in a paper by Chen et al., using computational techniques that looked at all eight gene segments rather than just one. This paper is not computational, so while more restricted in its question it may be more reliable in its results (which differ from Chen's). The results are interesting and may even be important, but we will only know that with time. At the moment they should be considered quite limited.
The researchers constructed artificial cell surfaces by making the putative receptor molecule for bird and human receptors anchored to an artificial surface. They tested to see which viruses bound to the avian and which to the human receptors. The tested viruses were constructed from bird and human isolates from Vietnam and also artificially constructed viruses that had the same HA sequences as viruses isolated in Thailand, Vietnam and Cambodia. The constructed viruses had either a bird H5 or an H5 from a virus isolated from a human together with the N1 from one of the H5N1s and six internal genes from PR8/H1N1, a common laboratory strain. Thus they were not testing H5N1 viral mutants isolated from birds and humans against human and bird cells, but laboratory constructed H5N1s unlike any found in nature. The thing that differed between the tested strains was in the H5 glycoprotein. By doing this they were able to answer what, if anything, was different about the HAs that bound to bird versus human receptors, although the receptors were not "natural" either. This is good scientific method. You isolate one thing at a time. However it makes generalization more difficult, since there are now many differences between the experimental set-up and the real world of natural infections. In the real world, the receptors are on cell surfaces, not on artificial surfaces and there may be other factors on the glycoprotein besides the single linking difference on the terminal sialic acid that was tested here. In addition, since we are fairly sure there is some kind of cooperative relationship between what is going on with the surface proteins HA and NA and internal genes like PB2 or NS1, the binding assay reported here may or not be critical to the ability to infect humans readily.
Under these conditions, the paper reports that changes at positions 182 and 192 worked independently to affect a change in binding affinity from avian to human cell receptors. Comparison with a putative crystal structure suggests this conclusion is plausible because these positions might affect the binding site. Combinations of changes at other positions also affected binding. Binding is clearly a complex process and it is not unexpected it might be altered by a variety of changes. The changes at positions 182 and 192 assumed more importance when they were shown to be present in two individuals from Azerbaijan and one from Iraq, but not in any of more than 600 avian isolates. The 192 change also enhanced binding to human receptor in the H5 of a completely different sublineage (clade 2) from Indonesia, but the 182 change abolished binding to both bird and human receptors. Clearly the story isn't simple.
This means to us that this is another brick in the slow accretion of facts about H5N1. So far, it isn't a breakthrough or a magic key signalling when H5N1 might become a pandemic strain. Since these changes have already occurred in some isolates, they are clearly insufficient to make the virus readily transmissible. It isn't even clear they are of any clinical or public health significance. The concluding sentence of the paper is appropriately careful:
Thus, such residues might provide useful molecular markers in assessments of H5N1 field isolates for their capacity to replicate in humans -- an essential indicator of pandemic potential. (Yamada et al. in Nature, vol. 444, 16 November 2006).
They might indeed provide useful markers. Then again, they might not. I don't think that's the kind of tag line a reporter is interested in.
- Log in to post comments
The patients in Azerbaijan and Iraq are almost certainly from clusters (can't be verified because WHO is still hoarding those sequences), but they certainly are Qinghai abd have PB2 E627K.
Moreover, the latest fatality in Egypt has M230I, which is in human influnza A (H1N1 and H3N2) and influenza B and is adjacent to the receptor binding domain - creating an exact match with infleunza B at positions 226-230 (QSGRI). The patient died a couple of weeks ago, but the HA sequence was released before her death.
Now there are four hospitalized patients in Egypt (3 siblings and neighbor). Not confirmed, but they are suspect cases and two are unconscious.
Plot is thickening
http://www.recombinomics.com/News/11160602/H5N1_Suhaj_4.html
typo killed the link in OP
Revere - You commented on the State Departments message to US Consulates a few days ago in your blog 'Bloggers catch US State Dept. on bird flu preps' Many 'flubies' are speculating on what could have prompted the message to prep for 12 weeks.
Do you see any connection between that announcement and the research in this Nature article?
Chuck: No, I don't see a connection. I think that the advice was specific to the consulate personnel as they could easily be caught in some very aggressive HK measures in the opening weeks of a pandemic. I don't see any specific elements suggesting we are a step closer (or further) from a pandemic. I'm not sure we will see anything. If it happens, it'll just happen without our seeing it coming. I know Henry feels differently. That's my opinioin, however.
Greenhammer cites the U.S. Embassy in Tunesia as recommending 12 weeks preparation as well.
Chuck, The story is in the sequence, and it was a triple wammy that caused the concern. WHO's worst fear is a change in the receptor binding domain without a decrease in the fatality rate, and the news on three fronts indicated that was going to happen, and it was likley to happen soon.
The PNAS paper showed that the Fujian strain was spreading, but it also shjowed many chnages in the receptor binding domain. The latest case in Egypt also has a change that matched all three seasonal flu's (Influenza A H1N1 and H3N2 as well as influenza B). The Nature paper showed that there were more changes that could affect the receptor binding domain, but these chnages were also in the Midle East where Qinghai was circulationg with PB2 E627K.
That raised the concern level worldwide, and Hong Kong would be concerned because the dead wild birds at its doorstep were all the Fujian strain, which had caused all of the human infections in China.
In the past WHO has forecasted infections in 1/4 to 1/3 of teh world's population, and if the case fatality rate remains in the 50% range (or anything double digits) the pandemic will be much worse than 1918.
revere,
FWIW, I read that paper EXACTLY the way you read it. Almost every sentence of what you wrote. Seeing that you are the expert, it gives me some comfort that I wasn't quite barking up the wrong tree. Thank you, it makes my day.
anon_22: "FWIW" -- since I know how smart you are, it's worth a lot. Thanks.
Revere and Dr. Niman, your responses are very much appreciated. It is so helpful to have people with expertise present different interpretations of what is a very complex and confusing situation.
As an added note on the sequences, WHO and consultants have problems understanding how the sequences change (they are still clinging to "random mutations"), but even though they have the wrong mechanism, they can read the sequences (although they have problems with influenza B).
They do see the chnages in the recptor binding domain and do know what the chnages mean, so they know there is a major problem, and they know that the situation is rapidly deteriorating.
This is about 100% off-topic. :)
But while digging through the web doing research on an related topic (LPAI detections in continental NA), I ran across a freely downloadable PDF on avian influenza here. The people who put it together appear to be medical professionals.
I wonder if you have looked at this, and if so, what your opinions are.
Charles: Yes, I have. It is excellent.
It being sunday, my day to clean house (which I avoid at all costs) I've been doing a little research and thinking I would like to share regarding the idea that recombination is the cause of the changes in the virus leading to increased infectivity in humans.
Hypothesis: we have two viruses circulating in the same host, one avian with serine in position 223, making it able to form a 2,3 linkage to a cell (which in birds is usually more commonly in the digestive tract and in humans is only common far down in the respiratory tract)and the other with asparagine in that position, making it able to form a 2,6 linkage, which in humans is common in the upper respiratory tract (and I'm not sure where it is in birds, if at all). According to recombinomics theory, the avian flu virus and the human flu virus could both infect a single cell and then switch sections of their genes, so that the avian flu virus acquired the genetic code for asparagine in the 223 position, and the human virus acquired the code for serine in the 223 position. This would give the avian flu virus the ability to infect the upper respiratory tract of humans without losing any of the other parts of its genetic make-up that cause it to be so lethal. Right so far?
This presupposes that the host cell of the human has receptors for both 2,3 and 2,6 linkages on a single cell.
Is this true? Are there simply more 2,3 sites in the lower respiratory tract and more 2,6 sites in the upper tract?
And since this would seemingly require the switch of the avian flu gene with a mammalian one, is there any evidence that any common mammalian viruses in the area of an outbreak have picked up the 223 mutation for serine in place of their asparagine? That would seem to support recombination theory.
On the other hand, mutation by substitution at the 223 site seems to me also a distinct possibility, since the genetic code for Asparagine AAU or AAC is only one nucleotide different from codes for Serine AGU or AGC.
A simple error in transcription as the virus begins copying its genetic material inside the host cell could easily cause the changes.
Most likely both recombination and substitution play a part, IMHO.