This is my first attempt at a blog post. I've avoided the genre up until now because my personal life is exceptionally dull. ( I could tell you a lot about bedroom ceilings.) Besides reading and sleeping, I do a good bit of thinking. Some of these thoughts may be of interest to others.
While I'm not in a position to say if science (or Science) has gained from the debate over XMRV, I can say what I have learned. While others seem fascinated with who wins and loses, my focus has been on stray facts that bubble to the surface of this stew. Some of them fit together.
The Knox/Levy paper introduced an apparent side issue, that the virus they were using was inactivated via a complement reaction in blood from both healthy controls and patients. This was taken to mean the virus was very unlikely to persist as a human infection. For me this illuminated the problem from a different angle.
Following their citation, you find this is triggered by a viral epitope resembling a particular molecule found on the outer coat of some bacteria. This is an innate rather than acquired immune response. Those bacteria are most likely found in the gut. If they appear in the bloodstream that is a sign of septicemia, for which a quick and decisive immune response is necessary. In this case, it came via the component known as complement.
This fits other pieces of evidence. The virus was originally found in the prostate. A later report put it in pulmonary secretions. Infection of rhesus macaques showed a tropism for reproductive and lymphoid organs. Virus was also found in their waste. What do these sites have in common?
My common factor would be mucous membranes. This points to a feature of immune response which still presents theoretical difficulties. While a great deal of immune activity can be simply described as distinguishing 'self' from 'foreign', these membranes present more complex challenges. Any sign of bacteria invading the blood should be dealt with sharply, but harmless bacteria just a few layers away should be left to perform a natural function, like aid digestion. We simply don't know the whole story for how these distinctions are made.
So, my inference was that virions, or infected cells displaying viral epitopes on their surfaces, would be tolerated in an environment like the gut, but suppressed if detected in blood. The key word here is 'detected'. Undetected latent provirus would be undisturbed.
I am not some oracle who can tell you who is right and wrong. (In a new field of science it is sometimes safe to bet that everyone is wrong.) I am not some superscientist who can follow details to some level far removed from the rest of the human race. What I can do is identify assumptions.
The scientific debate has covered a great deal of territory, with many curious twists and turns. The key difference between those still claiming positive results from those that are solidly negative appears to revolve around the question of detecting virus in blood. This carries with it an unstated assumption.
For most, "virus in the blood" = viremia = "free virions in blood". This is the easiest case to detect. There is a second possibility, cellular viremia. In this case, virions may not be found outside of cells, but the cells themselves may have active infections. In this case, some pieces of virus floating around loose inside the cell are likely to be carried to the cell surface, where they can be detected as viral epitopes. Again, if there is an active viral infection in blood, this should be (relatively) easy to detect.
However, if innate immune response inactivates virions, and removes cells with active infections, there is a third possibility. Virus present as latent provirus can still transit the blood, to be activated by conditions at another site more favorable for the virus.
It is possible people who have been sure they could detect latent provirus in blood have been unconsciously relying on the plausible assumptions that there must be some free virions and cells with cellular viremia present as well. We may not have seen a human retroviral infection with such a high percentage of latent provirus before. This doesn't mean it is impossible. Some attempts to treat AIDS resulted in reduced active infection, without changing total viral load. Active virus simply became latent until conditions became more favorable. This also happens with HTLV-1. Measuring proviral load is a very tricky problem.
We have animal models in related viruses. Significant percentages of latent provirus have been seen in mice infected with MLV. The problem for the virus is that the two-month life cycle of mice means it must replicate quickly, if it is to infect the next generation. What happens in larger animals with longer life cycles?
Bovine leukemia virus (a delta retrovirus) is an example. Here it may only be possible to detect one infected cell in 5,000 to 50,000 in blood, if you can't detect latent provirus. This is the kind of ratio which could yield the sharp contrast we see between researchers using different methods and assumptions. In the case of BLV, this problem lies behind the continuing difficulty in removing the infection from existing herds.
Culturing virus from blood plasma would certainly seem to imply there were active virions in cell-free blood. The problem with culture is that it can amplify a single virion into any number. It is certainly possible for viral titers to fall below detection thresholds for any except perfect assays. At this point there are no such assays.
Even the fragments of information in abstracts fit an idea concerning a common missing piece of all negative studies. I've been thinking all these groups made a standard assumption that a viral infection which can be detected in blood must propagate there. If this were true, the culturing step would not be essential, and would have the potential to introduce contamination.
If all virions and cells expressing viral epitopes in the blood are likely to be inactivated by complement reactions, only latent provirus in infected cells would be likely to evade this response. Hypermutation prior to insertion, and DNA damage from reactive oxygen species associated with mitochondrial damage, (common in this illness, but not most others,) would compound the problem of detection by PCR. Considerable DNA repair takes place during mitosis. This is also a time when production of hypermutation enzymes stops, to avoid interference with the cell's own reproduction.
My suspicion has been that the activation and culturing steps are critical to detection. Claims that this should not be necessary to detect virus need to be separated from claims that this was actually done in cases where no virus was detected. A further claim that the necessary step was not mentioned in the original research can be directly refuted. I had thought people were calling foul play over an explanation coming after the original paper. I had read an open letter to a patient group which spelled things out clearly enough for some laymen. I was not thinking about official response letters published in Science, but those also spelled out the importance of activation and culturing.
It turns out two important features of the process were in that original research publication. They were relegated to a supplement and encrypted in jargon, quite likely because of page restrictions imposed by editors, but they were present.
From the detailed supplementary information supplied with the original 2009 paper in Science:
> The light density fraction (buffy coat) was collected and
> washed twice with PBS. PBMC were activated by 1 __g/mL PHA (Abbott
> Diagnostics, Abbott Park, IL) and after 72 hours the cells were cultured with 20
> units/mL of IL-2 (Zeptometrix, Buffalo, NY) and subcultured every 3-5 days.
Activated PBMCs undergo mitosis as part of the natural amplification of tiny signals from detected epitopes into a strong immune response. This is crucial to detection of sequences only present as latent provirus in an environment which is hostile to virions and cells expressing viral sequences on their surfaces.
The distinction between culture and co-culture also seems to have been missed. Latest research, which I have only seen in abstracts, indicates the virus can sometimes infect cells without the XPR1 receptor previously identified. It is, so far, unclear if this takes place via a different class of receptor, or if it takes place by the formation of a viral synapse, which bypasses the need for specific receptors.
If the mechanism uses a viral synapse, culturing blood cells directly in contact with cells of the reported LNCaP line would allow infection to propagate by a different mechanism from the classic one used by free virions. This is one of those important details which can get lost in attempts at culturing virus. Culturing HIV-1 from blood is still not particularly reliable.
A virus replicating on mucous membranes need not propagate in blood. In the case of the "sin nome" strain of hantavirus, contamination of humans from viral particles in dust stirred up while sweeping did lead to lethal infections in the lungs before it passed through blood. (Hantavirus is a virus found in mouse droppings, though not a retrovirus.)
From a site like the surface of the lungs, this virus could spread through several mechanisms. It could infect immune cells there, which would then pass the infection to other cells in lymphoid organs. It could infect cells which could still pass through blood before they showed signs of infection. Infected cells on the surface might shed virions into the blood fast enough to overcome defenses, at least for a short time, as in the tests on rhesus macaques. This would be sufficient to spread the infection to other parts of the body.
The pulmonary mucosa are contiguous with the nasopharyngeal mucosa. Mobile infected cells could travel along this path without entering the bloodstream. From here, they could travel down the esophagus to stomach and gut. Bacteria have a hard time traveling between colon and prostate, despite the short distance. Do virions have better chances? Quite possibly. Based on present evidence, I would hesitate to say there are any mucous membranes the infection absolutely cannot reach.
We might be getting to the point of understanding what is happening, if anyone is still capable of thinking rationally.
I'm afraid some disputes come down to arguments that this virus refuses to play by rules set by virologists. It's hard to see how anything could. The original discovery reported 94% homologous sequences to X-MLV, which supplied the name, and 95% homologous sequences to a HERV. This leaves precious little room for anything distinct. If it does stay within that range, you can argue the lack of variation marks a contaminant. If other sequences turn up, you can argue they are unrelated.
Such possibilities have been explored in this debate. It seems the only safe way out is to avoid discovering anything. A number of participants might prefer that.
We now have several groups reporting evidence of immune responses to gamma retroviruses in patients, which contaminated PCR would not cause. One possible explanation, cross reactivity with sequences of HTLV-1, is not especially reassuring as this is known to cause HAM/TSP and ATLL, which are commonly lethal.
We are still learning.
While I'm not in a position to say if science (or Science) has gained from the debate over XMRV, I can say what I have learned. While others seem fascinated with who wins and loses, my focus has been on stray facts that bubble to the surface of this stew. Some of them fit together.
The Knox/Levy paper introduced an apparent side issue, that the virus they were using was inactivated via a complement reaction in blood from both healthy controls and patients. This was taken to mean the virus was very unlikely to persist as a human infection. For me this illuminated the problem from a different angle.
Following their citation, you find this is triggered by a viral epitope resembling a particular molecule found on the outer coat of some bacteria. This is an innate rather than acquired immune response. Those bacteria are most likely found in the gut. If they appear in the bloodstream that is a sign of septicemia, for which a quick and decisive immune response is necessary. In this case, it came via the component known as complement.
This fits other pieces of evidence. The virus was originally found in the prostate. A later report put it in pulmonary secretions. Infection of rhesus macaques showed a tropism for reproductive and lymphoid organs. Virus was also found in their waste. What do these sites have in common?
My common factor would be mucous membranes. This points to a feature of immune response which still presents theoretical difficulties. While a great deal of immune activity can be simply described as distinguishing 'self' from 'foreign', these membranes present more complex challenges. Any sign of bacteria invading the blood should be dealt with sharply, but harmless bacteria just a few layers away should be left to perform a natural function, like aid digestion. We simply don't know the whole story for how these distinctions are made.
So, my inference was that virions, or infected cells displaying viral epitopes on their surfaces, would be tolerated in an environment like the gut, but suppressed if detected in blood. The key word here is 'detected'. Undetected latent provirus would be undisturbed.
I am not some oracle who can tell you who is right and wrong. (In a new field of science it is sometimes safe to bet that everyone is wrong.) I am not some superscientist who can follow details to some level far removed from the rest of the human race. What I can do is identify assumptions.
The scientific debate has covered a great deal of territory, with many curious twists and turns. The key difference between those still claiming positive results from those that are solidly negative appears to revolve around the question of detecting virus in blood. This carries with it an unstated assumption.
For most, "virus in the blood" = viremia = "free virions in blood". This is the easiest case to detect. There is a second possibility, cellular viremia. In this case, virions may not be found outside of cells, but the cells themselves may have active infections. In this case, some pieces of virus floating around loose inside the cell are likely to be carried to the cell surface, where they can be detected as viral epitopes. Again, if there is an active viral infection in blood, this should be (relatively) easy to detect.
However, if innate immune response inactivates virions, and removes cells with active infections, there is a third possibility. Virus present as latent provirus can still transit the blood, to be activated by conditions at another site more favorable for the virus.
It is possible people who have been sure they could detect latent provirus in blood have been unconsciously relying on the plausible assumptions that there must be some free virions and cells with cellular viremia present as well. We may not have seen a human retroviral infection with such a high percentage of latent provirus before. This doesn't mean it is impossible. Some attempts to treat AIDS resulted in reduced active infection, without changing total viral load. Active virus simply became latent until conditions became more favorable. This also happens with HTLV-1. Measuring proviral load is a very tricky problem.
We have animal models in related viruses. Significant percentages of latent provirus have been seen in mice infected with MLV. The problem for the virus is that the two-month life cycle of mice means it must replicate quickly, if it is to infect the next generation. What happens in larger animals with longer life cycles?
Bovine leukemia virus (a delta retrovirus) is an example. Here it may only be possible to detect one infected cell in 5,000 to 50,000 in blood, if you can't detect latent provirus. This is the kind of ratio which could yield the sharp contrast we see between researchers using different methods and assumptions. In the case of BLV, this problem lies behind the continuing difficulty in removing the infection from existing herds.
Culturing virus from blood plasma would certainly seem to imply there were active virions in cell-free blood. The problem with culture is that it can amplify a single virion into any number. It is certainly possible for viral titers to fall below detection thresholds for any except perfect assays. At this point there are no such assays.
Even the fragments of information in abstracts fit an idea concerning a common missing piece of all negative studies. I've been thinking all these groups made a standard assumption that a viral infection which can be detected in blood must propagate there. If this were true, the culturing step would not be essential, and would have the potential to introduce contamination.
If all virions and cells expressing viral epitopes in the blood are likely to be inactivated by complement reactions, only latent provirus in infected cells would be likely to evade this response. Hypermutation prior to insertion, and DNA damage from reactive oxygen species associated with mitochondrial damage, (common in this illness, but not most others,) would compound the problem of detection by PCR. Considerable DNA repair takes place during mitosis. This is also a time when production of hypermutation enzymes stops, to avoid interference with the cell's own reproduction.
My suspicion has been that the activation and culturing steps are critical to detection. Claims that this should not be necessary to detect virus need to be separated from claims that this was actually done in cases where no virus was detected. A further claim that the necessary step was not mentioned in the original research can be directly refuted. I had thought people were calling foul play over an explanation coming after the original paper. I had read an open letter to a patient group which spelled things out clearly enough for some laymen. I was not thinking about official response letters published in Science, but those also spelled out the importance of activation and culturing.
It turns out two important features of the process were in that original research publication. They were relegated to a supplement and encrypted in jargon, quite likely because of page restrictions imposed by editors, but they were present.
From the detailed supplementary information supplied with the original 2009 paper in Science:
> The light density fraction (buffy coat) was collected and
> washed twice with PBS. PBMC were activated by 1 __g/mL PHA (Abbott
> Diagnostics, Abbott Park, IL) and after 72 hours the cells were cultured with 20
> units/mL of IL-2 (Zeptometrix, Buffalo, NY) and subcultured every 3-5 days.
Activated PBMCs undergo mitosis as part of the natural amplification of tiny signals from detected epitopes into a strong immune response. This is crucial to detection of sequences only present as latent provirus in an environment which is hostile to virions and cells expressing viral sequences on their surfaces.
The distinction between culture and co-culture also seems to have been missed. Latest research, which I have only seen in abstracts, indicates the virus can sometimes infect cells without the XPR1 receptor previously identified. It is, so far, unclear if this takes place via a different class of receptor, or if it takes place by the formation of a viral synapse, which bypasses the need for specific receptors.
If the mechanism uses a viral synapse, culturing blood cells directly in contact with cells of the reported LNCaP line would allow infection to propagate by a different mechanism from the classic one used by free virions. This is one of those important details which can get lost in attempts at culturing virus. Culturing HIV-1 from blood is still not particularly reliable.
A virus replicating on mucous membranes need not propagate in blood. In the case of the "sin nome" strain of hantavirus, contamination of humans from viral particles in dust stirred up while sweeping did lead to lethal infections in the lungs before it passed through blood. (Hantavirus is a virus found in mouse droppings, though not a retrovirus.)
From a site like the surface of the lungs, this virus could spread through several mechanisms. It could infect immune cells there, which would then pass the infection to other cells in lymphoid organs. It could infect cells which could still pass through blood before they showed signs of infection. Infected cells on the surface might shed virions into the blood fast enough to overcome defenses, at least for a short time, as in the tests on rhesus macaques. This would be sufficient to spread the infection to other parts of the body.
The pulmonary mucosa are contiguous with the nasopharyngeal mucosa. Mobile infected cells could travel along this path without entering the bloodstream. From here, they could travel down the esophagus to stomach and gut. Bacteria have a hard time traveling between colon and prostate, despite the short distance. Do virions have better chances? Quite possibly. Based on present evidence, I would hesitate to say there are any mucous membranes the infection absolutely cannot reach.
We might be getting to the point of understanding what is happening, if anyone is still capable of thinking rationally.
I'm afraid some disputes come down to arguments that this virus refuses to play by rules set by virologists. It's hard to see how anything could. The original discovery reported 94% homologous sequences to X-MLV, which supplied the name, and 95% homologous sequences to a HERV. This leaves precious little room for anything distinct. If it does stay within that range, you can argue the lack of variation marks a contaminant. If other sequences turn up, you can argue they are unrelated.
Such possibilities have been explored in this debate. It seems the only safe way out is to avoid discovering anything. A number of participants might prefer that.
We now have several groups reporting evidence of immune responses to gamma retroviruses in patients, which contaminated PCR would not cause. One possible explanation, cross reactivity with sequences of HTLV-1, is not especially reassuring as this is known to cause HAM/TSP and ATLL, which are commonly lethal.
We are still learning.
