If H5N1 is going to "go pandemic" it has to be transmissible from person to person. This occurs, but rarely. Why? The 1918 virus was not only lethal but easily transmissible. What's the difference between the 1918 pandemic strain and all the H5N1 strains we've seen so far?
One of the theories was that H5N1 didn't transmit readily because it didn't infect readily, and the explanation for that was that the avian versions hooked on to cellular receptors that are readily accessible in birds but not in humans. We've discussed the underlying science frequently (here [first in five part series, with other links available there], here, here). The story was that the only place where there were suitable avian type receptors in humans was deep in the lung and it was difficult for the H5N1 to get there. But it turns out that humans also have avian-style receptors on cells in the upper respiratory tract. The explanation didn't quite fit. A new paper in the journal Science provides new evidence but doesn't completely solve the problem. It is a remarkable and elegant paper.
A team of scientists from CDC, Mt. Sinai and USDA used a reconstructed version of the pandemic H1N1 virus obtained from the autopsy of a soldier who died in South Carolina in 1918 (SC18). The virus was used to infect ferrets, small carnivorous mammals that looks like (but aren't) rodents and are considered to be a good animal model for human influenza. Again it is the avian (designated α 2,3 sialic acid) and human (designated α 2,6 sialic acid) receptor types that are involved. Ferrets have receptor types distributed more like humans than other common laboratory animals. Uninfected ferrets were caged next to the infected animals and the virus was passed between them, although there was no direct or indirect contact. Sick ferrets sneeze and the virus apparently went airborne. The animals died with the typical severe lung lesions also seen in H5N1 infections. So far this confirms what we already knew: the 1918 virus was very transmissible, was very virulent and looked a lot like H5N1 pathologically.
Next ferrets were infected with a human seasonal H1N1 virus first isolated in Texas in 1991 (Tx91) and an avian (bird) virus isolated from a duck in Alberta, Canada in 1976 (Dk/Alb). These viruses also infected the ferrets but only the human virus (Tx91) transmitted to the uninfected ferrets caged next to them. The ferrets originally infected (by squirting the virus into their nose) had much less severe illness and lesions in their lungs. The fact that the human virus could infect was not a surprise, but so did the duck virus, which infected the upper respiratory tract which shed copious virus into nasal washes. But it didn't transmit and the ferrets didn't get very sick.
The punchline of this story comes next, however. The research team then introduced two amino acid changes into the hemagglutinin (HA) protein of the 1918 virus, one at a time. The changes were at positions known to affect the ability of the virus to bind to either avian or human receptors. The first change converted the virus into another virus seen in 1918, this one isolated during the same time period as the South Carolina virus from a dead soldier in New York. It is designated NY18 and the change made the virus ambidextrous for avian and human receptors, i.e., it bound to either one, although not as tightly as viruses that bind to only one or the other separately. They then changed a second amino acid which fully converted the propensity to bind to human receptors to the propensity to bind to avian receptors (now labeled AV18). Thus we have two steps from the fully human SC18 to the hybrid NY18 to the fully avian AV18, the latter two accomplished by genetic manipulation of the original SC18 virus.
All three viruses infected the intranasally inoculated ferrets. The NY18 virus transmitted to the uninfected ferrets, although less efficiently than either SC18 or
Tx18Tx91, both of which had full human receptor affinities. AV18, like Dk/Alb, did not transmit at all. In other words, the viruses with the full avian receptors didn't transmit from ferret to ferret and the one with a combined avian/human receptor did so, but not as efficiently. While not transmitting well, NY18 proved extremely virulent, like its cousin SC18. Virulence, remember, refers to the severity of the disease the virus produces, not its infectiousness. Just as significantly, AV18 also proved extremely virulent, although it didn't transmit at all. All three were able to infect the upper respiratory tract of ferrets.
Thus the receptor binding isn't involved in the ability to infect the upper tract nor is it necessary for the extreme virulence of the 1918 virus in the lower tract. It is involved in the ability to transmit. These are obviously very important findings, although how to relate them to evaluating the pandemic risk from H5N1 isn't immediately clear. For one thing, we know that the changes in H5N1 involved in receptor binding are different than the ones in the 1918 virus. These changes seem to be subtype specific. Nor do we know all the changes that might shift binding specificity. Nor do we know if a change in binding is the only way an influenza virus can become more transmissible. finally, we still don't know how receptor binding is involved in transmissibility.
It is important to point out that we needed to do these experiments to learn this. This required reconstructing the 1918 virus. It is a risky experiment, but in our view, worthwhile and this result shows one reason why.
It is a first small step, but a very important one.
Early in the third from the last paragraph, you reference a Tx18 virus. Should this actually be Tx91? Apologies if I'm wrong -- it's too early in the day for my brain to work right.
Oops! That should be "early in the fourth from the last paragraph, ..."
Stan: Thank you. Late night blogging is hazardous. Will correct.
Thank you for the concise explanation.
Beyond all the ignorance highlighted in the last paragraph, are there reasonable surmises and suspicions?
Would it not be of value to know what can reasonably be projected and hypothesized, based on these scientific observations?
Is that too speculative that we need wait one or a few years for the unveiling of the next facts?
For the past 2 years, I repeatedly do not see how we have the time to wait around for the next "doubt" to clear up. We lost a year to "it's gotta be reassortment". We lost another year to the barren "random and point mutations".
Are we now about to lose another year to "I don't know what's happening"? [This virus has no interest in engaging in these "scientific" investigations; it's evolving rapidly and more and more geographically broadly every single day. With curious and concerned scientists underfunded and quite randomly directed (from an outsider's point of view), I am concerned that the endgame show of Phase 5 will appear without more than a Tamiflu Maginot line. Will there ever be a Manhattan Project before Manhattan is cleared of its citizenry by this virus?]
Relevant to the questions I want to ask:
1) "Uninfected ferrets were caged next to the infected animals and the virus was passed between them, although there was no direct or indirect contact. Sick ferrets sneeze and the virus apparently went airborne."
2) "The ferrets originally infected (by squirting the virus into their nose) had much less severe illness and lesions in their lungs."
When you say "The ferrets originally infected ... had much less severe illness and lesions in their lungs," are you referring comparatively to the ferrets that were infected by the nasally-infected ferrets?
If so, that would seem to me to be more consistent (a la Tellier's review on aerosol transmission) with aerosol transmission and less compatible with large-droplet transmission.
Did the authors say anything about their thoughts on the primary mechanism of the apparent airborne transmission between the caged ferrets? (I'm guessing they would have set the experiment up so they could gain information along that line.)
Stan: All ferrets were intranasally inoculated with one of the viruses. The ones with the 1918 virus (SC18, NY18, AV18) all had severe lesions in their lower respiratory tract. The ones with the contemporary virus or avian virus (Tx91, D/Alb) had much less severe pathology or no pathology, although the cells of the upper tract were still infected and they shed virus. But neither AV18 nor Dk/Alb transmitted to nearby ferrets. Thus the aability to transmit was reelated to the receptor affinity change but not to the virulence of the infection. I don't think this sheds light on the large versus small droplet question.
"Thus the aability to transmit was reelated to the receptor affinity change but not to the virulence of the infection."
If virulence and transmissibility are not both inversly porportional and inter-related, we have a major problem down the road.
Tom: Not necessarily. To the extent they are independent, they are just that. If they are related, we don't know how tightly and in what way. They are clearly not inversely related n a deterministic way as this experiment shows. And knowing it is valuable.
Sorry to have to come back on this, but I still don't have a clear picture and it is an important point to me. I'll try to ask the question a different way.
In the case of each of the five different viruses, ferrets were infected intranasally by the virus. Dk/Alb and AV/18 did not transmit to other ferrets at all, but the other three viruses did.
In the fourth paragraph, you write, "The ferrets originally infected (by squirting the virus into their nose) had much less severe illness and lesions in their lungs."
1) Did the ferrets that became infected with SC18 via airborne transmission have more severe illness and lesions in their lungs than the ferrets that were infected with an intranasal squirt of SC18?
2) Did the ferrets that became infected with Tx91 via airborne transmission have more severe illness and lesions in their lungs than the ferrets that were infected with an intranasal squirt of Tx91?
3) Did the ferrets that became infected with NY18 via airborne transmission have more severe illness and lesions in their lungs than the ferrets that were infected with an intranasal squirt of NY18?
"Not necessarily. To the extent they are independent, they are just that."
Where does that leave the argument, repeated over and over and over again, that virulence would have to decrease for H5N1 to achieve pandemic transmissibility; a step that several studies have now estimated at two mutations (the quotes were as high as twenty mutations in Jan 2006).
H5N1 has increased in virulence as it has adapted to mammals...unprecedented in history...
...and transmissibility is now unrelated and "independent" of virulence...
...I think we have a major problem down the road...and we won't have to wait long to see and feel it...of course I hope in the end to be wrong.
Not to butt in to your discussion, but hopefully just to add to it.
Osterholm, I think in his City Pages interview last Spring (May?), has said the argument that it has to decrease its virulence in order to "go pandemic" is bullcrap (or something like that). I'm nowhere near an expert, but I had basically the same feeling as Osterholm quite a while before he said what he said. It just doesn't make sense unless H5N1 started killing people before they became contagious. I think it's just one more falsehood that makes it harder for "us" to get the message across about how serious this problem is.
I wish there were a Web site that specialized in identifying such stuff.
Stan: Severe disease among the contact ferrets of SC18 but not from NY18 or Tx91.
Stom, Stan: Regarding the "mandatory" decrease in virulence, we have discussed this several times here. For some diseases this has been true and not for others (e.g., HIV/AIDS, smallpox). Theoretical arguments show that there are ways that virulence can increase with transmissibility. See .