A flu virus protein that promotes secondary bacterial infection

Interesting paper from McAuley et al. (St. Jude's) on the PB1-F2 protein produced by an alternative reading frame on the PB1 gene of the influenza A virus. Most of you know that genes encode proteins via a three letter code. If you read the sequence of three letters by starting one letter earlier or later you will get a different sequence (e.g., ABCDEF is two three letter sequences ABC and DEF, but if you start a letter later you get BCD EF+whatever would have started the next three letter sequence originally). You have shifted the three letter frame for reading, hence the designation as an alternative reading frame. This way you can code more than one protein from the same set of (overlapping) letters. The F2 protein is produced in this way on the PB1 segment (one of the eight gene segments of the flu virus). PB1 is part of a complex of genes that produces the RNA replicating machinery of the virus (the RNA-dependent RNA polymerase). Whatever this protein is doing, it isn't needed for the virus to replicate. You can knock out the ability of the virus to make it and it will still replicate in eggs and cell culture. But it's doing something, and what it seems able to do is not very nice.

One thing the F2 fragment seems able to do is promote cell apoptosis, often called programmed cell death. This is a normal process needed for growth, development and protection, although it can also can go wrong and create pathology. The mechanism of apoptosis is very complicated, but at least one way to set it in motion is to do something to a small cell organ (an organelle), the mitochondrion, often called the cell's powerhouse, although we now know it has many other functions. Like initiating apoptosis. The F2 protein appears to target the mitochondrion and cause it to malfunction, initiating apoptosis and inducing inflammation. You don't need the virus for this to happen. You can do it with a synthetic full length F2. We know also that PB1-F2 will enhance viral pathogenicity in mice. So the St. Jude team took a closer look to see if the protein enhanced secondary bacterial infections, a major cause of morbidity and mortality in seasonal and pandemic flu.

Using H1N1 flu viruses that either do or don't produce PB1-F2, they established that its presence increases the incidence of bacterial pneumonia in a mouse model and affects the way the mouse reacts to the resulting infection with bacteria, increasing the response of cell-mediated and innate immune response but not the response of antibody producing cells. Having shown that the virally produced F2 enhanced seconadary bacterial infections, they tried to determine if what they were seeing was from F2 alone or connected with other viral proteins. They used fragments of the F2 produced by a common laboratory flu virus, H1N1/PR8, and found that a fragment from one end (the C-terminal end) was able to cause inflammation and promote severe bacterial pneumonia in the mouse model independently of the virus. The C-terminal end of the F2 protein was responsible from this lab flu virus.

Is the lab virus (PR8) F2 different than other F2s, in particular, the 1918 H1N1? By comparing the gene sequence of PR8 and the reconstructed 1918 virus on PB1, McAuley et al. engineered a 1918 PB1-F2 inside a PR8 virus. Interestingly none of the genetic changes in this reading frame affected the proteins made in the other reading frame (there is redundancy in the code so different triplets can code for the same amino acid). Thus the only thing that changed in the newly engineered virus was the F2 alternative reading frame protein. The 1918 F2 increased the virulence, accelerated the production of viral load (but didn't increase it) and changed the cellular response to inflammation compared to the usual PR8. The 1918 F2 infected mice had a severe pneumonia reminiscent of the pathology in mice infected with the full 1918 virus. There was no apparent effect on early cytokine storm, so F2 seems to be most important for the secondary bacterial infections that are a prominent part of influenza illness.

Those are the results. This work confirms the suspicion that the F2 protein produced by the PB1 gene from an alternative reading frame is a "virulence factor" for the influenza virus, i.e., a factor associated with virulence that may vary from virus to virus. Could it be responsible for the variation in severity of flu epidemics? In the Discussion section of this paper the authors note that after 1956, H1N1 viruses encoded for a PB1-F2 that is shorter than its predecessors. The truncated version lacks the part of F2 most important for inducing apoptosis and producing inflammation. The differences between the full length PR8 F2 and the 1918 F2 are also of obvious interest. The 1918 F2 seems to be affecting replication, even though the protein itself isn't necessary. This was a surprise and suggests that something about the changes in the 1918 sequence may have had effects on other viral proteins or genes.

As I said, pretty interesting. But there is an awful lot to learn about this. We don't know how PB1-F2 is enhancing inflammation and secondary bacterial infection, and when it does, what the consequences are. Does it set up a positive cytokine feedback produces a runaway loop with disastrous results? How is apoptosis and the inflammatory effects related (if they are)? Are there other F2s like the 1918 F2 or is it unique? There are many contradictory and confusing elements to this story at this point:

Clearly, much remains to be learned about PB1-F2 and its contribution to viral virulence. We have demonstrated here that the protein is proinflammatory, can contribute to virulence, and facilitates secondary bacterial infections. The ability of 1918 PB1-F2 to enhance the pathogenicity of PR8 may only represent the tip of the iceberg of the full pathogenic potential of this protein when expressed in its natural context of the complete 1918 strain genome. (McAuley et al., Cell Host & Microbe)

So this is the beginning of the story, not the end. But at least it's a story. And a fascinating one.

More like this

We've now had a chance to take a look at a new paper in Nature (advance online publication 27 September 2006 | doi:10.1038/nature05181) on increased host immune and cell death responses in mice infected with the reconstructed 1918 virus compared to other viruses with only some of the 1918 gene…
Every couple of months a major flu paper appears purporting to reveal why the 1918 H1N1 virus was so horrifically virulent in comparison to the other pandemic viruses of the last century, H2N2 (1957 pandemic) and H3N2 (1968 pandemic). It's not just the H1N1 subtype of the influenza A virus that…
Nature has just published another new paper on the basic biology of influenza virus. Unlike other recent papers it doesn't purport to reveal the secret of why some flu (e.g., H5N1, 1918 H1N1) is so virulent and "normal" seasonal influenza much less so. Instead it involves a process and structures…
I know I said I was going to discuss a bit more about pandemic preparedness today, but I think I'll hold off on that to discuss this story: It sounds like a sci-fi thriller. For the first time, scientists have made from scratch the Spanish flu virus that killed millions of people in 1918. Why? To…

This is an interesting piece of work, particularly in regard to bacterial infection. You might recall that during the 1918 pandemic researchers got fixated on B. influenzae as the causative agent, especially after Avery? invented the blood agar culture medium. I wonder what the proportion of secondary bacterial infections was in 1918 vs. the 1957 and 1968 pandemics?

Congratulations to St. Jude on a real step forward.

It seems to me that we must also concentrate how to make them better...after we make them sick.

What is vaccines affect on this developement in mice. What is Tamiflu's effect on this development in mice.

What is the effect of prednisolone and electrolytes and broad spectrum antibiotics after the condition is more or less fully developed...

...and most importantly to me...what is the effect of moderate relative doses of prednisolone, given for a short period of time in concert with broad spectrum antibiotics and electrolytes.,,given proactively early in the infection which would mimic readily achievable conditions in a pandemic...

...if the Governments and Regulators provide home treatment kits including oral prednisolone, oral electrolyte powders to be mixed with tap water, broad spectrum antibiotics and over the counter antifever medications...

...This whole package would probably cost less than twenty dollars per person and certainly no more than thirty dollars...seems like a small investment in our children...

...or are we not only too dumb to know the truth but also too dumb to treat our loved ones...in the absence of a functioning healthcare system.


Tom: I know you have a strong interest in prednisolone, but on the basis of the evidence we have it is a very dangerous drug to distribute and have people take without supervision or instruction. One lesson of the science here is that there are many interacting components and we don't know how this all works. Jumping to a remedy like this is a recipe for disaster, IMO.

Revere. Prednisolone at 50 mg orally for three to five days is dispensed to bee sting allergies in emergency rooms without any requirement for revisits or monitoring of any kind...how does this square with the dangerous drug analogy.

It has been commonly used in a variety of intensive care situations for more than fifty years...

...I don't see doctors using protective equipment when they inject it.

It is not immunosuppressive when moderate dosages (50 mg per day) are used for less than 7 days and certainly when used for 3-5 days when the patient should be weaned onto over the counter anti-fever medications like acetominophen (tylenol) which by the way, is way more dangerous than prednisolone and is a lethal weapon.

Prednisolone is only dangerous when it is used in massive dosages for more than 5-7 days...which is exactly what happens in SARS, H5N1 and ARDS.

It's just like Tamiflu...it didn't work so they doubled the doses...now they are talking about using it in combinations.

If a drug doesn't work as designed and in the first 36 hours then it isn't going to work...no matter how much you give and for how long.

Like I said...it seems that the general public is too dumb to know the scientific truth...are they also too dumb to have a chance to save their children.

I am tired of doctors restricting a potentially lifesaving drug under pandemic conditions...because they can.

If it is dangerous prove it...if it is immunosuppressive under the treatment I suggest prove it.

I don't need to know the explantation of how it works to know that it works.

Thanks as always.

ps. there is no other solution antivirals and vaccinations do not work with this influenza viruses which are too unstable and shift genetically faster than you can make a vaccine etc.

Thanks as always.

Tom: People will often use drugs incorrectly. I know, because I do it all the time. Tylenol is the leading cause of liver failure in the US. Prednisolone is not a benign OTC drug. It is pharmacologically active (or it wouldn't work for flu). Your only evidence that it is a good thing to do is anecdotal. You might be right, but I doubt it very much. It is your hobby horse and you are entitled to it, but IMO it is a dangerous piece of advice.

Revere It is not my hobby horse. If I have one, it is that environmental conditions including but not only sunspots and global warming are driving variabilities in parasites, bacteria, viruses and probably fungi as well.

This is one of the factors that brought me some time ago to the conclusion that an H5N1 influenced pandemic was inevitable and imminent (although imminent in nature's time is an unknown)

The reason that we are discussing prednisolone openly is because, based on my 15 years of anecdotal experience, it works...and since H5N1 and SARS are animal diseases, I believe I have a right to say my peace.

This study once again indicates to me that there is a reason that prednisolone works and nothing else does for what is referred to as the cytokine storm (the cytokines to me are secondary incidental to the massive replication of virus in some patients).

Under pandemic conditions (at the moment with all our eggs in the antiviral and vaccine basket, neither of which have a hope in hell of working) with no functioning healthcare system...

...we are going to leave young parents of young children with what...illusions...vials of water called vaccine on one day and antibiotics on the next and morphine when they don't work.

We must have treatments avaliable for people to treat H5N1 at home with professional advice if necessary...

...and these packages must include what I have already said plus morphine for the third of the CFR that is going to die no matter what treatment is given from the direct effects of the virus (this is also based on my anecdotal experience)...

...the thing is that these kinds of outbreaks may occur once every fifty years in human medicine but occur once a week in veterinary medicine.

...and by the way, if prednisolone is so dangerous...then why is it that it is a standard treatment dispensed orally from emergency departments.

This drug, in my opinion, is still wearing the stigma of the name 'steroid' and the unfortunate incidents when it was introduced in the nineteen fifties when it was thought of as a panacea.

In my opinion, prednisolone in responsible dosages for responsible time periods is an immune modulator and not an immune suppressor...

...and this study once again explains why it would work to prevent the cytokine storm.

The issue for this drug to be effective is timing...and under pandemic conditions, early enough treatment should not be a problem.



Revere: This research seems to extrapolate, and capably modify -- by filling in revealed gaps, and oversights -- the previously (somewhat confused) material generated by the Nature paper (headed "The Nature paper on the 1918 virus and immune reaction") that you posted on this site on 9/29/2006. The current paper does much to address the structural inadequacies of the research (which probably could not have been effectively addressed, at that time, in any case) that the earlier paper was devoted to. It also suggests, to me, that those inherent inadequacies were essentially unavoidable. That paper, as a consequence, seemed to generate far more questions than it could possibly resolve. The current paper seems to do an admirable job of tidying up some of the messes. That paper did promote some very interesting commentary, though, just the same.

The H1N1 viruses used for that earlier research would almost certainly have been post-1947 viruses, which would have ensured the introduction of truncated (essentially non-functional) PB1-S2's, as you pointed out in your presentation of this current paper.

From your explication of the modification of the viruses employed in that earlier research:

"The authors (Kash et al.) infected mice intranasally (through the nose) with four viruses, one a currently circulating human H1N1 virus, A/Texas/36/91, dubbed Tx91 for short, and three others, two produced by replacing first two of the eight gene segments of tx91 with the corresponding ones from the 1918 virus (HA and NA, thus designated 2:6 1918) and then five of the eight gene segments of Tx91 (HA, NA, M, NP and NS, 5:3 1918). tx91 was also an H1N1 virus so no new sub-type was produced. Kash et al. also infected the mice with the fully reconstructed 1918 H1N1 virus (r1918), that is, an H1N1 that had all of the original eight of the 1918 gene segments."

The 1918 H1N1 virus was unmodified, retaining its original PB1 segment intact. The experiment did not select for PB1 as the focus of any anomalous virulence contributions (it couldn't), so it apparently played no part at all in assessing the outcome of the research. Unfortunately.

From your further comments on the earlier research:

"The most lethal combination was r1918. Moreover the pathology of the lung lesions looked distinctly different with r1918 compared to the other viruses, suggesting that the pathology and virulence of the virus is not due to specific genes but rather the cooperation or combinations of proteins from genes on different segments."

Some very important, fundamental questions have just been resolved, with this latest research, as I see it.

Dylan: I think your comments are very astute, although I am not quite as confident this work as answered open questions as much as opened up new questions. There seems to be interaction with other viral proteins or processes with F2 and we don't know what they are at the moment. But it is another piece of the jigsaw puzzle, with most pieces still missing and nothing fitting together nicely yet (eventhough we often try to mke them fit). So I think we need to just stay tuned.

but what is the virus' benefit ?
Why should virus evolution favour
secondary bacterial infection ?

anon: The virus doesn't think or want. Evolution tends to maximize the chance of finding another host, which could be another cell in the same host. If the host lives longer it also ha a chance to find a host and if it is coughing up sputum to infect one. Maybe the F2 virulence is just a hitchhiker on another more advantageous trait. Many, many possibilities.

An analysis of the causes of death during the 1918 Spanish Flu pandemic showed a link with tuberculosis. Many of the Spanish Flu victims were also infected with tuberculosis; just as today in Africa, many Aids victims become infected with tuberculosis.
How many bird flu victims are also infected with tuberculosis? What happens if extreme drug resistant tuberculosis produces an epidemic, at the same time an influenza pandemic develops?
Would the recent increase in those infected with tuberculosis create a situation in which an influenza pandemic might more easily spread?

Viral epidemics/pandemics have many potential factors.

Ecological is another. See David Quammen's Deadly Contact.

Tuberculosis would make sense, especially during that time period. Might be interesting to see how viruses and Mycobacterium tuberculosis interrelate today.

Add in vitamin D deficiency and low amount of defensins. Put them altogether and an interesting hypothesis is born.

Now. Where's the grant?

By phytosleuth (not verified) on 12 Oct 2007 #permalink

An immunologist, Dr.Warwick Britton, at the University of Sydney; is studying the interaction of influenza and tuberculosis infections in the lung. These are the two most common lung infections.
Tuberculosis is a major cause of death globally. The interaction between different types of infection in the lung is poorly understood.
Many victims of bird flu may have also been infected with tuberculosis at the time of death. If that is true, which disease killed the victim? The same issue existed in 1918 during the Spanish Flu pandemic.
Now that H5N1 has adapted to the lung of humans, how could tuberculosis facilitate the process of H5N1 human to human contagion?

Revere, I do not expect this comment to appear but I am releasing a book and have been trying to speak to for some time. You appear to be very well informed and have many intuitive insights and genetic knowledge relating to the current Avian situation. We are currently doing research in several medical areas and what we do and so forth is virtually unknown on the net. After 30 years in the medical field, but outside the mainstream doctor group, much as Gates was outside the mainstream mainframe thinking, we have data you might find interesting, but is not exactly well received in terms of the current mindset of using Tamiflu when it is not effective, ignoring significant information in the search to correctly avoid a Pandemic (new one). Currently I wanted to speak to you concerning the possibility of interviral, or bacterial chemical communication where a "pack theory" - no opportunistic but diseases working together and occurring literally at the same time. Their are huge inaccuracies in current medical theory and practice. Shamman's dancing around a fire. MC

By medclinician (not verified) on 14 Oct 2007 #permalink

Um, medclinician, I don't even know exactly what you are advocating for, but you just demonstrated just about every "quack" tendency short invoking Hitler, including:

-claiming suppression
-claiming to be a visionary like Bill Gates (just because an idea is mocked doesn't make it right; wrong ideas make up the bulk of mockery targets)
-comparing modern medicine to shamans dancing around a fire

Of course modern medicine will look primitive in 25 years. We keep studying and moving it forward. Here's an idea bud, get a grant, do some real research, and publish it in a real journal. Or at the very least state your point clearly without all the cries of discrimination. Like I said, I haven't even heard your point yet, and I'm already skeptical because of how you presented it.

By Leukocyte (not verified) on 15 Oct 2007 #permalink