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 segments. It is a very interesting paper.
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 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. Kash et al. also looked at which genes of the mouse immune system might be activated by the different viruses. Other studies had shown that the increase in cytokines in mouse lung infected with the 2:6 1918 could be attributed to the increased number of inflammatory and scavenger cells, suggesting the increase in cytokines was an effect rather than a cause. The same phenomenon was seen in this study with r1918, i.e., an increase in cytokines from increased recruitment of white blood cells and macrophages (a kind of scavenger cell) in the lungs. The microarray method used to detect which genes were upregulated during infection could not distinguish what kind of lung cell was involved, but did show that that the severe disease produced by r1918 was correlated with genes involved in inflammatory response. Similar findings have been seen with H5N1, suggesting that H5N1 shares some ominous similarities in pathogenic mechanisms with the 1918 virus, features it doesn't share with contemporary H1N1 human viruses.
The interplay of functional components of the immune system appears very complicated, either because it is or because we have yet to unravel the simpler coupled relationships that produce its complicated dynamics. This study tried to tease out some of these inter-relationships. The overall picture is that the virus causes the body to turn on a pro-inflammatory response and also to initiate a programmed cell death response. Whether the latter is a result of the former, or vice versa or they are independent or partially independent is not known. Programmed cell death (apoptosis) is different than a cell being killed by some exogenous agent. Apoptosis is a cell suicide routine, and like the inflammatory response, is part of the protective and development repertoire of normal cells. Cells can be induced to kill themselves if their death is required by shaping an organ system during development or to protect the host from an infected or cancerous cell. But like inflammation, which is meant to recruit infection-fighting cells to an area of microbial invasion, apoptosis can do much damage if uncontrolled. It appears that something like this may be part of the devastating pathology of r1918 and probably H5N1.
In this study the FAS and caspase cascade are upregulated. They are part of the cell death program. Researchers are still trying to tease out the convoluted chain of events related to programmed cell death and it appears there may be several different parallel chains capable of initiating it. One of them involves the cell's energy generator, the sub-cellular organelle called the mitochondrion. Caspase-9 which is one of the upregulated proteins in this study, is activated in the cell death routine by one of the contents of the mitochondrion, cytochrome c. Recent work indicates that a protein coded by an alternate reading frame of the PB1 gene segment of the influenza virus affects the mitochondrion in ways that might increase release of cytochrome c. An alternate reading frame is a way to make a different protein from a gene sequence by shifting where one starts to code the protein's amino acids. Amino acids are coded by triples of gene bases but there is no spacing or punctuation to indicate where to start the triple. If you move over one base you will get a different sequence of triples and produce a different protein. That is what happens with the PB1-F2 segment which takes the PB1 gene segment and translates it again slightly differently. It is of interest that none of the viruses in this study, other than r1918, used the PB1 gene segment, i.e., none of tx91, 2:6 1918 or 5:3 1918 had the 1918 PB1 that was in r1918. It would be of interest to do the same experiment with an incomplete 1918 genome but one that included PB1.
This paper is an important contribution to understanding the role of host immune responses in virulent influenza infections. So far it appears the full 1918 genome is needed to produce the catastrophic response we associate with 1918 H1N1 and the current H5N1 subtype. We have yet to discover what parts of the 1918 virus are mirrored by current strains of H5N1 now circulating in many areas of the world in birds, with sporadic cases in humans, and incomplete knowledge of other possible reservoirs. It is possible that these virulence factors of 1918 are all present, given the clinical picture in humans. The question would then shift to those factors which govern the transmissibility of the virus from bird to human and human to human, since H5N1 currently is not easily transmitted to people as the 1918 virus was.
The basic science of influenza virology is moving relatively quickly, but the pace of science, no matter how rapid, is slow compared to the possible time frame of an emerging pandemic. It's a race whose outcome we don't as yet know.
- Log in to post comments
These are interesting findings regarding the human immune response to H5N1. But there is a more urgent task that needs to be addressed.
There are now four human bird flu clusters in Indonesia. they are at Tulungung, Bandung, Sumatra and Tasikmalaya.
There is obviously the probability of H2H, which can be concluded by observing the onset dates of disease.
In none of the clusters do the human sequences match the poultry isolates. This means the human cases are probably being infected from H5N1 evolving in a reservoir other than domestic poultry.
Unless there is an attempt to determine what that hidden reservoir is, it will be impossible to slow down the spread of human H5N1 infections. And if this attempt is not made, then there is a good chance you will be able to observe the effects on human H5N1 infection on the human immune system by observing those you know dying from H5N1 as they drown on their own body fluids if the pandemic hits the US.
For the public health community to ignore the scientific evidence that there is probably now a hidden reservoir of human H5N1 infection, and to make no attempt to find this reservoir, is an act of barbaric ingnorance.
"So far it appears the full 1918 genome is needed to produce the catastrophic response we associate with 1918 H1N1 and the current H5N1 subtype."
I seem to remember reading something at the beginning of this year about researchers combining segments of r1918 with ordinary human flu and infecting mice. I think it was reported that virulence was greatly increased. This was from some little 2 paragraph summary in something like New Scientist magazine, but I can't exactly remember.
William,
Don't worry. God has put Bush on this planet to save the world from H5N1. And he's doing a heckuva job I might add. I know you're outside of the US, but Bush's allmightiness transcends international borders.
concerning the race : science vs. pandemic,
I'd be interested, how much such a study might cost.
How much is being invested in this ?
Is the prevention of the pandemic just a question
of money ? How much ? I feel, that we should have solved
all these problems, should have done all these studies
much earlier.
william: Which human isolates don't match which bird isolates? In what way? There is never a complete match. Is the mismatch of the type that definitely signals the one couldn't be derived from the other (after all the are likely in the same clade)? Are you matching a human case in a cluster (remember cluster does not mean H2H, it means a grouping of cases in space and time; most clusters are common source clusters in epidemiology) with the sick poultry in the same area? I not saying you are not. I am just asking the questions since your statement is so definitive and I want to know how you made the deduction.
Edmund/anon: Yes, there has been some previous work with the 2:6 1918 and some others. This work is part of the process of parsing the existing evidence. Which goes to anon's question. This is difficult work that has to be done under BSL3 plus conditions. It takes time to do and is less a matter of money than available expertise and facilities. There are many questions to answer and not that many people able to answer them and of those that can, they have only a finite amount of time and resources. Yes, it would have been much better if we'd done a lot of this earlier. Some of this is because our non-leaders didn't lead us to do it earlier. Some of it has to do with scientific advances that had to come first and new techniques and facilities coming on line.
If it were just money we'd have cured cancer by now. It isn't. Nature doesn't give up her secrets easily. People are working hard at this, which is why I get annoyed when I see them being dinged for one thing and another. They are appropriate targets for some kinds of criticism (they work in a competitive field which both increases their productivity but also raises obstacles to sharing that could increase net productivity still more), but they are doing a lot more than the rest of us to solve this problem.
You can't do this work without adequate funding, but the rate limiting step isn't money.
Edmund. Webster and others did a split off back in December? of the H1N1 strain, inserted the non-lethal parts of another virus H5N1 into it and produced vaccine using egg technology. The idea was that the non lethal parts would keep the egg alive so as to produce vaccine. The vaccine worked.. sort of. It too was so hot that you could die from the vaccine. Yikes!
The levels of the titers I read were great but the body responses of the test subjects were almost as bad as getting full blown H5N1. You also had to have already taken this prophylactically or so I understood it. Its not get H5 and then take it post infection. It would surely kill you.
The mice thing came as an adjunct to this later. One side trying to find vaccine, another creating a situation that could be used as a bioweapon. Infect mice, then turn them loose. Only they didnt turn them loose. H1 is alive and well in pigs in Tennessee/Kentucky and Georgia. Hate to think what will happen if it gets into a swine population that already has seasonal pig flu. Humans with H5N1 infecting pigs with H1N1. New strain? Surfs up!
Revere,
Thank you for responding to my post. I have used the information posted by Henry Niman at www.recombinomics.com.
Therefore I will now quote some of his posts:
August 31:
"The release of recent H5N1 bird sequences in Indonesia have identified a duck that has the novel cleavage site,RESRRKKR. This sequence has been detected in almost all human H5N1 isolates in West Java. The failure to find the sequence in domestic poultry has raised concerns regarding the origin of the human infections."
September 8, this post relates to the 47th victim.
"Although there were dying poultry in the proximity of the confirmed case, the link to domestic poultry remains weak."
"In 2006, all human H5N1 isolates, other than the cluster from Karo, have had a novel HA cleavage site, RESRRKKR. This change is associated with a number of additional changes in HA as well as the other seven gene segments which are not present in poultry H5N1 isolates."
And Niman, in later posts, keeps repeating the same theme in regard to all of the recent clusters since September 8; and he requests that other mammals be tested, such as dogs, cats, and pigs.
In a post here, on your blog site, Niman said he thought pigs and humans were the best possibilites for the unknown reservoir.
To date the only sequence that has matched has been from a cat. Andrew Jereminko said on your site that the human H5N1 sequences have to be matched with the animal sequences in order to determine the reservoir.
Over and over again, in the new clusters in Indonesia, there is no match between the human and poultry sequences.
Therefore, how can poultry be the reservoir?
Please go to Niman's site and review the information there to bring the information you need up to date.
I would sincerely like to know what conclusions you reach regarding the information I have presented here; and the information Niman is presenting.
I realize Indonesia is a difficult field assignment for WHO employees; and they are risking their lives. Therefore I have no intention of criticizing them.
But could not field studies be done in Indonesia by veterinarians, by taking blood from dogs, cats, and pigs, in an attempt to find a match with the human sequences?
Do you believe the identification of the reservoirs is a critical issue; or do you believe, in the end, it will not make any difference, if a pandemic occurs?
I am concerned by the increasing frequency of human H5N1 infections in Indonesia, and I believe this may indicate we are in the beginning stages of a pandemic. WHO representatives have stated that in at least 6 instances there was probably human to human infection of H5N1. The 1918 pandemic started slowly, and then increased in speed over about a 4 month period. I sincerely hope that is not the case now; but all of these clusters in Indonesia may indicate otherwise.
Edmond,
Today is historic with the passage of a law that officially makes the US a fascist police state, and gives the president the power to throw any American citizen in jail, and throw away the key.
Under this new law defendants will lack rights to confront accusers, exclude heresay accusations, or bar evidence obtained through coercive interrogations. Do you think torture might be considered coercive interrogation? I do.
They will not be guaranteed a public or speedy trial, and will lack the right to choose a military counsel, who in turn would not be guaranteed equal access to evidence held by prosecutors.
And the Posse Comitatus Act, passed in 1879; which prohibits the National Guard and the US military from acting in a local law enforcement capacity, will soon be abolished.
Abolition of this law will allow federal control of the National Guard.
Under Executive Order 13010, Bush has the power to declare martial law in time of an emergency. Bush will decide what is or is not an emergency. It can be a natual disaster, a bird flu pandemic, a terrorist attack, etc.
Martial law means the military controls the civil population. Do you think Bush would be crazy enough to declare martial law, in order to stay in power? I do.
Recent news indicates the intelligence commuity considers any leftist or anti-globalization proponent to be a terrorist. We are talking about US citizens that are considered terrorists by our own intelligence community.
And the congressional award of a contract to a subsidiary of Halliburton, for $385 million, to build detention centers in the US, supposedly for illegal aliens,
will of course serve as nice concentration camps for all those who oppose the new fascist state called the United States. Do you think sadists in those detention centers might torture you, if you refuse to be quarantined during a bird flu pandemic? Of course not, don't be silly. One of the prisoners in a US Army prison in Iraq had his legs smashed to pulp, and he died of a heart attack. Do you think that could happen in a Hallburton constructed detention center in the US? Of course not, since we are protected by the Constitution.
Or we were protected by the Constitution until today. With the new law we are not.
I am glad I live in Medellin Colombia. I feel a lot safer here. I know that sounds crazy, but under the new fascist state that is the US, I do not want to be there.
The BBC's report on the study quotes lead author Dr John Kash, as saying: "What we think is happening is that the host's inflammatory response is being highly activated by the virus, and that response is making the virus much more damaging to the host. The host's immune system may be overreacting and killing off too many cells, and that may be a key contributor to what makes this virus more pathogenic."
It occured to me that this is also a very fitting metaphor for one of your other interests, Revere--the War of Terror.
...Sorry: the war ON terror.
Name; Good point and good metaphor. There is a different possibility than the one Kash suggests that appeals to me more. Apoptosis is important in the immune system, where autoimmune cells are culled. But feedback from certain classes of cells also are needed to stop inflammatory responses. It could be apoptosis of an immune subset happens first and fails to turn off the pro-inflammatory immune response. I'd bet on that before I'd bet on Kash's view, but there is no data to support either scenario at present and it's just speculating.
I do not think that the problem is an immune system out of control...over-reacting due to a trigger from H5N1.
I think the problem is a normal immune sytem reacting normally to a highly efficient, effective, and replicative virus.
It's just that in winning the battle, the immune system loses the war...the cytokines are just the 'smoke of the battle'.
Hi Revere. I have a question about something you said in the article above. "...there is no spacing or punctuation to indicate where to start the triple. If you move over one base you will get a different sequence of triples and produce a different protein."
I acknowledge my limited expertise in the area of genetics: as a secondary school science teacher, I am a generalist. But my general understanding is that there is a specific nucleotide sequence that acts as a "start" codon to which the initiation complex binds. This ensures that the translation into a protein of any RNA sequence will begin at the proper triplet code on the chain, otherwise the amount of frame shift mutations would be profound. This seems at odds with your comment about their being no "punctuation" to indicate where to start the triple. Is this something specific to viruses only, or am I misunderstanding something? Could you please explain? Thanks.
MiH: No you are correct, in general, but once started the process takes it three at a time and if it moves one base (a frame shift) it changes the coding output. When I said no punctuation (there are start and stop codons) I meant that the polymerase doesn't know if the triple it is reading is the right one or not if it started at a place not modulo three of the "usual" start. This can happen normally. The PB1-F2 is an alternative reading frame.
Tom: We are all speculating. I would bet that you are wrong, but no one has any good evidence right now to tell your hypothesis from mine from the Nature paper from yet another one. I do think, along with Kash et al., that immune dysregulation is the key here, but we'll see. There is a lot of work being done on this and I think we'll know much more within a year. If we are still around, that is (that's a joke -- sort of).
Kash's nature paper is a follow on to work he's been doing for several years. I've always said that while we might "assign" a given virulence mechanism to a particular protein from a given gene, it is the sum of these effects that makes the virus so bad. Studying gene products in isolation in recombination experiments is important, but doesn't tell the whole story.
The prematurely induced apoptosis via the caspase-9 mechanism is really bad news (and by the way, Revere, I agree with you on your speculation). This provides more evidence, in my opinion for cytokine dysregulation in which colloquially, the virus "steps on the gas" with the pro-inflammatory cytokines, and "hits the brakes" on the anti-inflammatory cytokines, by inhibiting the feedback loops. What I find so amazing, is that 8 small RNA segments can defeat all the cell's beautifully regulated loops bichemically. It does this very quickly--within 1-3 hours. Hit fast; hit hard. H5N1's motto.
"What I find so amazing, is that 8 small RNA segments can defeat all the cell's beautifully regulated loops bichemically. It does this very quickly--within 1-3 hours. Hit fast; hit hard. H5N1's motto."
Marissa. That is one basis of my argument.
Do not underestimate the effect of a massive number of cells infected and the debris created when those cells are removed by a normal immune system.
Deaths from primary viral pneumonia, without secondary issues, are not uncommon in the animal world.
Marissa: re your statement "the virus does this very quickly - in 1 to 3 hours." Correct me if I'm wrong, but you seem to be saying that it reproduces so rapidly within any given cell that it dysregulates the cell's feedback loops and thus causes the cytokine "storm". And are you saying that it completes an entire replication cycle, from insertion to release of new completed virions in only 1 to 3 hours? Wow!! But Tom seems to be saying that it is this rapid replication on a massive scale, and the consequent volume of dead cells, that causes the cytokine storm..the host's reaction to all those dead cells lying about. I know there have been studies done in vitro to see how fast this virus replicates: so is this where you got the 1 to 3 hours data? Is there also data on how many virions are made per cycle (and thus released to infect surrounding cells?) And would this high speed replication have anything to do with the mitochondrial release of cytochrome C, since the mitochondria are having to supply all the "power" in the form of ATP to drive this replication and assemble the viral proteins? It wasn't entirely clear from the dialogue above under what circumstances mitochondria release cytochrome C...that is, is this a normal product, a byproduct, or is it a breakdown component from mitochondria under stress? If it the latter, might there not be a stepwise causal chain something like this: virus infects cell and goes into overdrive of reproduction, causing stress to mitochondria, which breaks down and releases cytochrome C, which activates Caspase 9 which causes cell apoptosis (in addition to all the other factors above?) It would seem the culprit to look for that is at the root of all this is what enzyme the virus has or activates that promotes rapid replication and/or dysregulates controls on rate of replication. The rapidity of replication causes all the rest, separately and in concert.
Revere: The more I read about this (cytokine dysregulation), the more impressed I am by the extent of our lack of fundamental knowledge (and the concomitant lack of confidence that is inspired) in this particular area.
Is it your position that the dysregulation of the immune reaction is the product of discrete mechanisms; one which initiates cellular apoptosis within the infected cells, before any sort of chemical messengers could alert the immune system's natural apoptotic inducing mechanisms to begin initiating the process in these infected cells? As MIH suggests, something along the lines of mitochondrial stress inducing cell death, effectively "localizing" the apoptosis, until a significant "swarm" of virions (which were initially masked by this process) can be expressed; and which will subsequently evoke an extraordinary response from the immune system, including apoptosis inducing elements? The result being that two essentially independent processes conjoin to maximize both the apoptotic effect, and the toxic effect of the cytokine storm? Would this be a reasonable interpretation of the "cause-effect mechanism?"
And if this isn't an approximation of the process (and I am certainly in no position to judge), then how might we go about determining what the process does actually consist of? Will we just have to wait until we have a sufficient number of subjects, who expire at various stages of presentation, before we can arrive at any reasonable interpretation of the process? I'm not being facetious here, at all. I just wonder how we are going to go about this; how, exactly, do we begin to approach the problem? Are there experimental avenues available to us, here?
I should have mentioned that I understand the use of "surrogates," for experimental purposes, here; but will these be sufficiently close to approximating us, in structural terms, to allow for any reasonably reliable conclusions, in this remarkably complex area?
Mary, the appearance of cytochrome C is a biocehmical marker of cell death--and end result, not any cause. I have a Japanese paper around here somewhere on this and will try to dig it out some time--though I'm moving across town and it's difficult to find it.
Many papers support the timelines I indicated. My point is there is massive upregulation of pro-inflammatory cytokines and downregulation ofantiinflammatory cytokines that occurs very quickly. These cytokines then target other cell types, spreading the havoc--which is Tom's point, with which I agree. What this boils down is that H5N1 leverages our own immune system against itself. I don't see any specific involvement of mitochondria. H5N1 has to leave energy production intact or it cannot propagate itself or do much damage.
Revere(s), while I am not a virologist, I have followed the story of 1918 H1N1 and Taubenberger and colleagues' journey to reconstruct the virus and characterize its biology with great scientific and medical fascination.
With my limited background, is not the current Science report consistent with the unusually high frequency and rapidity of death observed in young and otherwise very healthy American servicemen, particularly at New Jersey's Ft. Dix in 1918? Today, we worry first about immunizing infants and the very elderly. With 1918 H1N1, one would think that those with the most healthy immune systems and intact apoptotic cascades would be most susceptible to this strain of influenza.
Apologies if this idea has been put forth before, but thank you for your coverage of this continuing development of investigations into this historic and horrific infectious disease outbreak. Simply fascinating....and frightening.
Dylan: You are right we are all waving our hands here. Having said that, let me try to wave mine a little more clearly than I did in the comments you are probably referring to.
The immune reaction -- particularly the generic, non-specific "innate" reaction -- appears to us as a very complicated back and forth of signaling between different immune cell types. It may just look complicated because we don't understand it very well, like describing somebody chewing food in minute detail without quite understanding eating. But from what we see so far, cells/tissues when under stress or threat from an invading parasite send out a general call for help, one result of which is the pro-inflammatory response, the influx into the area of inflammatory cells whose job it is to fight the invader, often with crude, non-specific tools. Like sending in the SWAT team. They in turn are capable of calling for still more help. The first ones are holding the line until more help comes. And the new ones can call for still more help. Etc.
Normally this doesn't keep going on in an escalating fashion like this, because there are still other cells whose job it is to send out the Stop signal. OK. Mission accomplished (excuse the bad associations with current events). My thought was that the virus might be inducing those "brake" cells to commit suicide, hence no Stop signal and a runaway reaction. Kash's thought is that the virus causes the accelerator to be stuck in the "on" position, and both could be happening. Tom think it is just that the situation is so bad the immune system keeps calling for more and more help because it needs it. In fact, all of these could be playing some part.
The only way I know to approach this is by taking apart the signalling systems and figuring out how they work, not quantitatively but qualitatively. That is what mathematical modelers do in the field of dynamical systems. There are people who have been working on this for a long time, and I think someday we will understand it. Meanwhile we have to get lucky. Earlier I posted on the experimental drug that went wrong and induced a cytokine storm. We'll have to follow up on things like that to see if we can find some handle or foothold, like just a part of the process, like a protein, that can be used as a target for intervention, even if we don't quite know how the whole thing works. Trial and error.
Abel: The "W" shaped mortality curve of 1918 has been one of the great puzzles. Indeed in every pandemic the hallmark has been a shift of the age distribution to the left. Thus 1951 Liverpool was still an H1N1 and had higher mortality than 1918 but a "normal" age distribution so wasn't a pandemic strain. The age distribution of victims of H5N1 is very mysterious and no one has been able to explain it (this is also true of other things, like flu seasonality, which we take for granted but we don't know why it occurs).
We're still trying to understand apoptosis (which seems to have multiple pathways and signalling modes), so trying to figure out its place in this is pretty difficult. The new report is indeed consistent with what happened in 1918, but it is more like another data point that shows what we would expect but not why we get it.
Marissa: You wrote "Mary, the appearance of cytochrome C is a biocehmical marker of cell death" and that "I don't see any specific involvement of mitochondria."
I got the idea that Mitochondria and cytochrome C were involved in the apoptosis based on the statement below, taken from the article that started this discussion.
"One of them involves the cell's energy generator, the sub-cellular organelle called the mitochondrion. Caspase-9 which is one of the upregulated proteins in this study, is activated in the cell death routine by one of the contents of the mitochondrion, cytochrome c."
In that article, it also states "Programmed cell death (apoptosis) is different than a cell being killed by some exogenous agent. Apoptosis is a cell suicide routine, and like the inflammatory response, is part of the protective and development repertoire of normal cells."
Note the use of the word in "routine" in the last two paragraphs. Putting this together, the author would appear to be saying "cytochrome c activates Caspase-9 which is involved in the cell suicide routine." You seem to be saying the opposite, that cytochrome C is a biproduct of ordinary cell death (as opposed to the suicide routine) rather than a causative agent. Am I misreading you?
Mary: okay, I'll try to be clearer. (Sounds like the author of the article could have been clearer, too.) When a cell undergoes apoptosis, (and the caspase-9 pathway is only one trigger--there are others), all the cell's contents, (after some digesting that goes on within the cell as it breaks up) are released into what we call the third space, that is the interstitial spaces between cells. Mitochondria are broken up and the cytochrome C, one of the many entities involved in oxidative phosphorylation--the aerobic mechanism of producing ATP in cells) is also released. Apoptosis, or programmed cell death occurs when the cell figures it has lost the battle with the invader. (I'm sure Randy could come up with a nice war metaphor.) In influenza, this occurs when the flu strain has caused too much damage in a cell. I finally found the Japanese reference I was looking for whose main conclusion is: "The present results indicate that cytochrome c is a useful marker to follow patients with influenza-associated encephalopathy and suggest that an apoptosis of cells in several organs including the cerebrum and liver under the influence of hypercytokinemia is a possible mechanism of the disease."
Here's a link to the full abstract.