Does Tamiflu resistant swine flu virus really spread less well?

There is no way to keep up with all the flu news, so we pick and choose, usually based on some kind of point we want to make. That's both the good and the bad of this blog: the news comes with a point of view. But so does most news, and we try to make ours both explicit and scientifically as accurate as we can with the information at hand. Today is a typical example. Bloomberg is reporting that any swine flu virus resistant to oseltamivir (trade name Tamiflu), the only antiviral pill effective at all for the infection, transmits less well than swine flu that's sensitive to Tamiflu. The source of this information is Dr. David Mercer, identified as acting head of WHO's European region communicable disease unit. The article makes it sound pretty definite. I don't know if that's the reporter's interpretation or Dr. Mercer's, but, frankly, I don't have a lot of confidence in this report at the moment. Here's some background on this for those of you new to the issue.

Tamiflu (and Relenza, an inhaled antiviral also used to treat swine flu) works by blocking the action of an enzyme that is on the surface of the flu virus. The enzyme is neuriminidase (NA), a protein that comes in 9 different flavors on the influenza A virus. Those nine forms of neuriminidase are unimaginatively given numbers, 1 to 9, and the numbering gives the flu A subtypes the N-part of of the designation. The pandemic swine flu is H1N1, meaning that its neuriminidase surface protein is the one that bears the number 1. The other main surface protein, hemagglutinin (HA) is responsible for finding the doorway into the cell. It's called a receptor protein and it is responsible for getting the virus into the host cell where it has access to the replication machinery of the host that it hijacks to do the only thing in the world it does: make copies of itself. The HA proteins are also numbered, 1 to 16, and again it's number 1 that is in the H1N1 swine flu subtype. The HA grabs onto a matching receptor on the host cell which sets in motion a mechanism which gets the virus into the cell. There has to be a natural affinity or stickiness, sort of lock-and-key like, between the HA and the host cell. And that turns out to be a problem the virus has to solve on the other end, after it has gone through its replication inside the host cell and its millions or billions of copies are trying to leave the surface and find more host cells to infect. It is the neuriminidase on the virus that unsticks it and let's it go free. Perhaps the reason this doesn't prevent the cell from getting in is that the kinetics are such that infected host cells don't take up virus as easily uninfected ones or some other reason. In any event, if you block the action of neuriminidase the replicated viruses from the infected cell just stay stuck to the surface. You haven't prevented the cell from being infected but you've prevented most or many of the progeny viral particles from going on to infect other cells.

The story is more complicated because we still don't understand completely the receptor dynamics and what makes the virus infect one cell and not another or the cells of one tissue and not another or the cells of one species and not another. We've had a lot to say about this story (and the different kinds of receptors) over the years and you can find the posts we've done in the last three or four years scattered among the biology posts here (typical examples, here and here and an example of some of the complications, try this one). Most of those receptor posts are about what's on the cell, but the neuriminidase also differs on the virus and there are at least two groups. Tamiflu and Relenza were designed to work with group 2 NA (like the seasonal flu virus H3N2) and they also seem to work with group 1 (of which N1 is an example; the identity between the group number and the N-number is a coincidence, one of the many confusing things about this business). Yes, Tamiflu works on group 1 NA, but not exactly the same way, and given that we are still unsure about how the receptor-HA systems works (which is the system that the NA clips), you can see that, well, there's still a lot we don't know.

Which brings us back to the WHO claim that the resistant swine flu doesn't spread as well as the sensitive swineflu virus. First, in one sense this is an old story. Lab experiments done by the manufacturer of the drug, Roche, did seem to show that flu virus resistant to Tamiflu was less "genetically fit" than virus that wasn't resistant. This was highly reassuring because we know that when we treat people with Tamiflu usually a tiny fraction of them start producing virus that is resistant. In other words, they got infected with sensitive virus but the Tamiflu selects out some rare mutation that allows resistant virus to flourish inside the patient. But if the patient didn't transmit the virus (because it was "less fit"), then the resistance couldn't spread. We'd have only the rare sporadic case.

For a while that seemed to hold true. The biology seemed sound and all was very reassuring. Then, suddenly, in the 2007-2008 flu season the H1N1 seasonal influenza (not the swine flu H1N1) developed Tamiflu resistance and in a remarkably short time almost all of the H1N1 seasonal flu was resistant to Tamiflu. This happened in the US even though Tamiflu was infrequently used to treat flu. It seemed as if the Tamiflu resistance was "hitch hiking" on another trait that either cancelled out or conferred even more genetic fitness that was lost by becoming Tamiflu resistant. We still don't know how that happened, but while there are a few mutations that can cause Tamiflu resistance in the lab, in the real world they have almost all been one, H274Y in the NA protein. The Bloomberg news article doesn't say why Dr. Mercer believes the resistant virus doesn't spread as well as the sensitive one, but the reason is almost certainly because of the two dozen or so cases of resistance they have seen in Europe, all have been sporadic, that is, not epidemiologically linked. This suggests that the mutation involved, although not stated, is almost surely H274Y, the same mutation that confers resistance in seasonal H1N1, a virus which spreads very well indeed. Maybe swine flu virus is different, but recently CDC reported resistant isolates in two epidemiologically linked cases.

So here's the bottom line for us regarding this news report. Unless Dr. Mercer has a crystal ball, any report that Tamiflu resistant virus doesn't spread as well as Tamiflu sensitive swine flu virus is not only premature, but grotesquely rash. It's not just that we don't understand the whole system here (the HA-receptor-neuriminidase clipping system). It's that we have the same subtype H1N1, the seasonal variety, 100% resistant to Tamiflu via the same mutation and spreading with ease. There's clearly many important components to the ability to transmit from person to person, very little of which we have any understanding of, not to mention all the complications related to receptors and drug resistance.

In fairness, maybe Dr. Mercer's statements lost nuance in the reporting. But we have seen a lot of confused rosy thinking cheek by jowl with dire speculation coming from WHO, when neither is warranted in terms of the science. WHO is not the World Reassurance Organization but some officials can't seem to help themselves when talking to the press. Or maybe they are themselves engaged in whistling past the graveyard. I don't know.

But at this point, with a couple of dozen cases of resistance detected and the little we understand of how this virus works, I'm taking this optimistic report with a mountain of salt.

There. You got our point of view. Finally. It sure took us long enough.