In The Reefs of Taprobane, Arthur C. Clarke recounted some underwater adventures he and Mike Wilson had exploring reefs around Sri Lanka. (He was ahead of his time in calling his books on diving the Blue Planet trilogy. That name was not used for a television series until much, much later.) In The Treasure of the Great Reef, he recounts how they discovered a wreck on Great Basses Reef, and tried to do a little archaeology in addition to treasure hunting. Clarke commented that this was about the only time he felt it was justified to investigate an archaeological site without a grid and careful measurements. He could have simply said it was impossible, because divers and equipment were being carried back and forth by the surge and backwash of waves even in calm weather. After storms even heavy objects found before would be in different places, so you could hardly call the site undisturbed. They did try to record the approximate location and circumstances where they found things, but this would never meet professional standards. Nonetheless, they were able to show that at least one ship had come to grief on that particular reef centuries ago, and they could tell a good bit about it.
The reefs and shipwrecks of most science are metaphorical, though nonetheless real disasters. The hazard to navigation I wish to describe today is called reification, from the Latin word for thing, res. It consists of creating a thing from an idea that isn't really a concrete thing at all. Since every example I can come up with is embroiled in some controversy, many quite heated, I may as well plunge headlong into the immediate topic, a virus.
We all have some idea what we mean by a virus, and few today would connect it with the Latin term for a poison, still less the original connotation of human semen. We think a virus is something really, really small which makes us sick -- not entirely wrong. We tend to equate the virus itself with what specialists call a virion, and sometimes, with retroviruses, also an inserted sequence in a chromosome called a provirus. (For DNA viruses we might also refer to an episome as the virus itself.) These are real things, but do they fit the meanings we attribute to them?
When we say we have a virus, we mean a viral infection -- a pathological process. Calling a process a thing raises a warning flag for anyone concerned with semantics. The problem is that virions are not alive, they have no metabolism; you could crystallize them and hide them in salt shakers (not that I recommend this.) Crystallizing the provirus might be harder, you would most likely end up with amorphous glop. What you would not find is significant activity of any kind. Virtually every action the virus takes is actually carried out by either the host cell or some other environment. The virus itself doesn't directly do much of anything except provide misleading information, rather like turning in a false fire alarm, or mailing yourself to some location by climbing into a package on the doorstep quickly after ringing the bell. Human beings did not invent deception.
The problem is that arguments about what a virus can or cannot do tend to be based on a perception that the virus is a fairly autonomous thing. Strictly speaking this is never true, but in a more general sense of behavior of hosts with and without the virus we can call some activity entirely viral. Some processes can be entirely ascribed to a viral infection. Other host processes may be present with or without the virus, but are coopted to serve viral replication and transmission in pathological cases, as with AIDS. This is at the root of huge controversies which have run for decades.
As a case in point, we have the latest installment on rumor viruses from eminent virological authorities. (You can find essentially the same party line in publications about 30 years old.) Part of this covers the story of the decline and fall of XMRV, though significant elements of the story are omitted. The second half covers the possible role of HERVs, and HERV-K/HML-2 in particular, in cancer. You will not find any pleas of mea culpa concerning research failures though there have been substantial blunders by virological prognosticators. Earlier publications by the same authorities listed 62 long HERVs in the HERV-K/HML-2 family. Work published last year showed more than 100 additional full-length copies of the provirus for HERV-K111 had been overlooked. Other work revealed that HERV-K insertions continue to alter human genes. You can't simply assume different individual humans have the same provirus in the same locations.
(One implication of recent research on HERVs is so specific to problems known to be associated with ME/CFS that I can't avoid mentioning it here: Endogenous retrovirus-K promoter: a landing strip for inflammatory transcription factors. )
If you read that sermon on rumor viruses above, you may find some nuggets hidden in the underbrush. There is no longer any question that complete functional copies of each of the original exogenous retroviral genes are present somewhere in the human genome. There is even a full-length provirus with complete open reading frames, HERV-K113, though it has not been possible to culture this in vitro in isolation. (Before you write this off, you might consider that nothing in vivo takes place in isolation, and culturing viruses has frequently stumped researchers for years. Culturing HIV-1 was a major breakthrough, and that virus replicates quite a bit more rapidly than some other known retroviruses.) A second hidden nugget is the admission that many HERV genes are really expressed, not only as RNA sequences, but also as proteins which provoke antibody responses.
I'll quote Robert Gallo from a paper on the discovery of the first human retrovirus, which other groups thought impossible at the time:
This bias has scarcely disappeared. Researchers continue to act as if small animal models directly translated to humans, even in diseases where perinatal transmission is important, and the time scale between nursing and sexual maturity is two orders of magnitude longer in humans. While HIV infection manifests itself as AIDS in about 5 years, HTLV often takes 20 or more years to produce pathological consequences. We know of two very different outcomes for HTLV infection, and we may be missing still more. (Look at all the manifestations of HIV infection.) If HTLV had been endemic in humans I would expect pathological consequences to be delayed until after reproductive age, which was about age 35 prior to modern medicine. Ancestors of those who would die more quickly would have died long ago without issue.
Researchers were able to discover HTLV and HIV in part because humans have no corresponding HERVs to confuse matters, and thus no experts to tell them these clues should be ignored. There were also distinct diseases appearing in new populations to sound the alarm. Had diseases caused by HIV been long accepted as part of "the human condition" neither factor would have favored recognition, and the delay in recognizing the cause of the resulting diseases would have been much greater.
Now let us consider what has been tacitly admitted, (with very little emphasis,) about the search for tumor viruses. Although it isn't directly stated, the ERVs designated pre-XMRV1 and pre-XMRV2 are not just slightly defective, they are mere fragments. Probability of recombination of random sequences in chromosomes is so low it may as well be neglected in most laboratory experiments. Start the process with a replication-competent retrovirus similar enough to allow RNA strands from ERVs to be packaged as dimers in virions, and the probability of a cross-over event during transcription of a complete viral genome is about an even-money bet. This is not likely to create a new, replication-competent virus on the first try, but a hundred or so events could easily be enough to succeed. In organisms with a trillion cells carrying ERVs this is quite likely, even if the original infection remains undetectable. This would explain the appearance of replication-competent recombinants only 8 days after the start of an experiment. We don't know what started the process, but a slow or latent retroviral infection seems most likely.
If you start with some kind of replication-competent retrovirus, the whole process makes sense. Without that it is a kind of Deus Ex Machina invoked to get the drama out of a wholly implausible impasse.
So, the first tacit admission is that undetected slow or latent retroviral infections are not uncommon, even in the most carefully studied laboratory animals. The alternative would be the assumption that a retrovirus scattered across many ERVs can reconstitute itself quickly without such a precursor. I can't think of any other alternative worth mentioning.
A second admission is that intact functional retroviral genes for complete retroviral genomes appear to persist in human hosts, though not necessarily in a single functional provirus.
A third admission is that HERV genes are in fact expressed, both as RNA sequences and proteins, in many human diseases.
Now, let's consider what elements of the process of retroviral infection are present. Transmission of retroviral genes to new hosts? Check. Transcription of those genes? Check. Expression of proteins coded for by those genes? Check.
(I wish to point out that immunosuppressive domains known to be in viral envelopes need not be part of functioning virions. Such proteins have been implicated in human diseases. Local immunosuppression could even be more important in etiology of cancer than random shuffling of genes. It is quite possible that individual cancerous cells arise frequently, and are promptly destroyed, without producing clinical disease.)
Even actual virions are present in a number of cases, though these are generally not replication-competent, or do not replicate rapidly enough to sit up and do tricks in laboratories. It looks like quite a number of retroviruses have simply become obligate parasites which have dispensed with inessential parts of their replication cycle. This evolutionary behavior is well known in the study of more conventional parasites.
A chicken is just an egg's way of making another egg. A virion is just a viral genome's way of making another genome. Inside an individual host cell such a genome can make copies of itself without creating virions. In this case it is called a (retro)transposon. This is confined to an individual cell, unless that cell exchanges cytoplasm with other cells. Such exchange happens in the case of exosomes. It can also happen when a cell forms viral or immunological synapses. These show up in cell-to-cell transmission of viral diseases and cancer. Individual host cells, especially immune cells, may even be transmitted to other hosts during sex or nursing. A disease which infects cells, and can be transmitted via exchange of host cells, would not be as rapidly contagious as your typical epidemic disease, but it would fit the requirements for an endemic human infection.
Such endemic retroviral diseases are known in other primate species: Gibbon Ape Leukemia Virus, Mason-Pfizer Monkey Virus. We can't tell if there are such viruses in humans because replication would be slower, and there are many similar HERVs. Humans are subject to breast cancer and chronic myeloid leukemia (CML), two diseases of unknown etiology corresponding to those found in the named species. (Two friends just happen to be afflicted with the named diseases this year.) The evidence that something very similar is going on in humans has become overwhelming. Researchers can exhibit sequences, virions, inserted provirus, proteins, antibodies and God knows what else. They cannot demonstrate the rapid replication in a laboratory required by the referenced paper, and if they did it would almost certainly be a contaminant. To reiterate: if a human retroviral disease passed near birth involved rapid replication it would kill hosts before being passed on at sexual maturity. This implies it could not be an endemic human infection resembling endemic retroviral infections in other species.
As said before, if the host has no ERV corresponding to an exogenous retrovirus the situation is much simpler. This specifically applies to HIV and HTLV. HIV-1 also replicates too rapidly to allow about 99% of hosts infected near birth to reproduce. This makes it plausible to assume the criteria for tumor viruses advocated in that paper were chosen to make sure that research on HIV maintained total control of funding in retrovirology research.
Consider what evidence those studying cancer have already produced. If retroviral sequences and resulting expressed proteins might come from HERVs we can essentially cross off all that work. Any sequence which rises above the background noise far enough to meet requirements will violate the rule I've stated about slow replication. You might consider variation in proviral insertion sites between individuals to be clear evidence of a slow infection still in progress, but that is being rejected in that paper, even though it is mentioned. What is now required is observation of an "insertion event". I'm not sure how you would observe such an event in the environment of human cells carrying hundreds of HERVs, with some actively transcribed. This takes us back to the requirement of replication-competent virions and reification of the process of retroviral infection. What had seemed at first to be discussion of new criteria simply ends by reiterating those criteria previously demanded.
I begin to think the message is really that virologists have little positive to contribute to cancer research, an hypothesis for which I can produce abundant documentation.
HIV has produced a particularly nasty epidemic disease which would kill about 99% of those infected without medical treatment. I am not advocating we stop research on the subject. It would, however, make sense to evaluate current research directions and accomplishments.
With the disclosure that the two Boston patients have gone back on HAART we can say that the only possible cure would be the "Berlin patient" who got a new immune system from an "elite controller" who happened to match tissue types. This kind of exceptional circumstance prevents this treatment from scaling. Without some means of eradicating the disease in infected patients we might need to start planning for a world in which the entire population is maintained on ARVs for life.
Just how do you eradicate an infection by a virus which has inserted about a trillion copies of itself in the chromosomes of host cells? You don't, unless you also have a way of eradicating an infection in which provirus is inherited in a trillion ways. The problem of curing AIDS is closely related to the problem of diseases caused by HERVs. This makes current funding disputes especially stupid.
The last 30 years of research on retroviruses has not exactly been an overwhelming success even w.r.t. AIDS. Most ARV drugs were found by sheer trial and error, little different from research which produced Salvarsan and Prontosil. The important contribution of virological research came in three parts: isolation of the virus, culturing the virus, detection of viral infection via antibodies and PCR. All three took place quite some time ago. The role of expert opinion in guiding research since has not been exactly prescient, except in the sense that the best predictor for future funding is past funding. Nowhere will you find people saying "forget what we told you 5 years ago" (or 20 or 30 years ago). In the toxic environment in which battles over research funding take place any admission of error is fatal. The result is near stasis when it comes to translating research into treatments for human diseases, or even explaining etiology of known human diseases.
(This is not to say there have been no serious errors in biomedical virology research. I've already discussed one horrible example with implications for wild bird species we still don't understand.)
If I haven't stirred up enough trouble already, here is one final blast. Anyone remember the flap over Bovine Spongiform Encephalopathy (BSE or "mad cow disease")? I could point to a corresponding spongiform encephalopathy in mice caused by a murine retrovirus. I can also reference recent work on activation of ERVs in prion disease. Now, consider what I said about eggs using chickens to make more eggs. Does this suggest that the "nude virus" once hypothesized for BSE was present all along as an ERV? Proteins with sequences derived from ERVs do show up in lesions of prion diseases. Ruling out the presence of active virions would have no effect on transmission using ERVs. The triggers for ERV expression could be much smaller and harder to find. Nobody knew this at the time BSE had its 15 minutes of fame. Experiments on which official reassurances were based wouldn't come close to meeting current standards for avoiding contamination by viruses and nucleic acid sequences, and badly underestimated the importance of ERVs.
The reefs and shipwrecks of most science are metaphorical, though nonetheless real disasters. The hazard to navigation I wish to describe today is called reification, from the Latin word for thing, res. It consists of creating a thing from an idea that isn't really a concrete thing at all. Since every example I can come up with is embroiled in some controversy, many quite heated, I may as well plunge headlong into the immediate topic, a virus.
We all have some idea what we mean by a virus, and few today would connect it with the Latin term for a poison, still less the original connotation of human semen. We think a virus is something really, really small which makes us sick -- not entirely wrong. We tend to equate the virus itself with what specialists call a virion, and sometimes, with retroviruses, also an inserted sequence in a chromosome called a provirus. (For DNA viruses we might also refer to an episome as the virus itself.) These are real things, but do they fit the meanings we attribute to them?
When we say we have a virus, we mean a viral infection -- a pathological process. Calling a process a thing raises a warning flag for anyone concerned with semantics. The problem is that virions are not alive, they have no metabolism; you could crystallize them and hide them in salt shakers (not that I recommend this.) Crystallizing the provirus might be harder, you would most likely end up with amorphous glop. What you would not find is significant activity of any kind. Virtually every action the virus takes is actually carried out by either the host cell or some other environment. The virus itself doesn't directly do much of anything except provide misleading information, rather like turning in a false fire alarm, or mailing yourself to some location by climbing into a package on the doorstep quickly after ringing the bell. Human beings did not invent deception.
The problem is that arguments about what a virus can or cannot do tend to be based on a perception that the virus is a fairly autonomous thing. Strictly speaking this is never true, but in a more general sense of behavior of hosts with and without the virus we can call some activity entirely viral. Some processes can be entirely ascribed to a viral infection. Other host processes may be present with or without the virus, but are coopted to serve viral replication and transmission in pathological cases, as with AIDS. This is at the root of huge controversies which have run for decades.
As a case in point, we have the latest installment on rumor viruses from eminent virological authorities. (You can find essentially the same party line in publications about 30 years old.) Part of this covers the story of the decline and fall of XMRV, though significant elements of the story are omitted. The second half covers the possible role of HERVs, and HERV-K/HML-2 in particular, in cancer. You will not find any pleas of mea culpa concerning research failures though there have been substantial blunders by virological prognosticators. Earlier publications by the same authorities listed 62 long HERVs in the HERV-K/HML-2 family. Work published last year showed more than 100 additional full-length copies of the provirus for HERV-K111 had been overlooked. Other work revealed that HERV-K insertions continue to alter human genes. You can't simply assume different individual humans have the same provirus in the same locations.
(One implication of recent research on HERVs is so specific to problems known to be associated with ME/CFS that I can't avoid mentioning it here: Endogenous retrovirus-K promoter: a landing strip for inflammatory transcription factors. )
If you read that sermon on rumor viruses above, you may find some nuggets hidden in the underbrush. There is no longer any question that complete functional copies of each of the original exogenous retroviral genes are present somewhere in the human genome. There is even a full-length provirus with complete open reading frames, HERV-K113, though it has not been possible to culture this in vitro in isolation. (Before you write this off, you might consider that nothing in vivo takes place in isolation, and culturing viruses has frequently stumped researchers for years. Culturing HIV-1 was a major breakthrough, and that virus replicates quite a bit more rapidly than some other known retroviruses.) A second hidden nugget is the admission that many HERV genes are really expressed, not only as RNA sequences, but also as proteins which provoke antibody responses.
I'll quote Robert Gallo from a paper on the discovery of the first human retrovirus, which other groups thought impossible at the time:
This bias has scarcely disappeared. Researchers continue to act as if small animal models directly translated to humans, even in diseases where perinatal transmission is important, and the time scale between nursing and sexual maturity is two orders of magnitude longer in humans. While HIV infection manifests itself as AIDS in about 5 years, HTLV often takes 20 or more years to produce pathological consequences. We know of two very different outcomes for HTLV infection, and we may be missing still more. (Look at all the manifestations of HIV infection.) If HTLV had been endemic in humans I would expect pathological consequences to be delayed until after reproductive age, which was about age 35 prior to modern medicine. Ancestors of those who would die more quickly would have died long ago without issue.
Researchers were able to discover HTLV and HIV in part because humans have no corresponding HERVs to confuse matters, and thus no experts to tell them these clues should be ignored. There were also distinct diseases appearing in new populations to sound the alarm. Had diseases caused by HIV been long accepted as part of "the human condition" neither factor would have favored recognition, and the delay in recognizing the cause of the resulting diseases would have been much greater.
Now let us consider what has been tacitly admitted, (with very little emphasis,) about the search for tumor viruses. Although it isn't directly stated, the ERVs designated pre-XMRV1 and pre-XMRV2 are not just slightly defective, they are mere fragments. Probability of recombination of random sequences in chromosomes is so low it may as well be neglected in most laboratory experiments. Start the process with a replication-competent retrovirus similar enough to allow RNA strands from ERVs to be packaged as dimers in virions, and the probability of a cross-over event during transcription of a complete viral genome is about an even-money bet. This is not likely to create a new, replication-competent virus on the first try, but a hundred or so events could easily be enough to succeed. In organisms with a trillion cells carrying ERVs this is quite likely, even if the original infection remains undetectable. This would explain the appearance of replication-competent recombinants only 8 days after the start of an experiment. We don't know what started the process, but a slow or latent retroviral infection seems most likely.
If you start with some kind of replication-competent retrovirus, the whole process makes sense. Without that it is a kind of Deus Ex Machina invoked to get the drama out of a wholly implausible impasse.
So, the first tacit admission is that undetected slow or latent retroviral infections are not uncommon, even in the most carefully studied laboratory animals. The alternative would be the assumption that a retrovirus scattered across many ERVs can reconstitute itself quickly without such a precursor. I can't think of any other alternative worth mentioning.
A second admission is that intact functional retroviral genes for complete retroviral genomes appear to persist in human hosts, though not necessarily in a single functional provirus.
A third admission is that HERV genes are in fact expressed, both as RNA sequences and proteins, in many human diseases.
Now, let's consider what elements of the process of retroviral infection are present. Transmission of retroviral genes to new hosts? Check. Transcription of those genes? Check. Expression of proteins coded for by those genes? Check.
(I wish to point out that immunosuppressive domains known to be in viral envelopes need not be part of functioning virions. Such proteins have been implicated in human diseases. Local immunosuppression could even be more important in etiology of cancer than random shuffling of genes. It is quite possible that individual cancerous cells arise frequently, and are promptly destroyed, without producing clinical disease.)
Even actual virions are present in a number of cases, though these are generally not replication-competent, or do not replicate rapidly enough to sit up and do tricks in laboratories. It looks like quite a number of retroviruses have simply become obligate parasites which have dispensed with inessential parts of their replication cycle. This evolutionary behavior is well known in the study of more conventional parasites.
A chicken is just an egg's way of making another egg. A virion is just a viral genome's way of making another genome. Inside an individual host cell such a genome can make copies of itself without creating virions. In this case it is called a (retro)transposon. This is confined to an individual cell, unless that cell exchanges cytoplasm with other cells. Such exchange happens in the case of exosomes. It can also happen when a cell forms viral or immunological synapses. These show up in cell-to-cell transmission of viral diseases and cancer. Individual host cells, especially immune cells, may even be transmitted to other hosts during sex or nursing. A disease which infects cells, and can be transmitted via exchange of host cells, would not be as rapidly contagious as your typical epidemic disease, but it would fit the requirements for an endemic human infection.
Such endemic retroviral diseases are known in other primate species: Gibbon Ape Leukemia Virus, Mason-Pfizer Monkey Virus. We can't tell if there are such viruses in humans because replication would be slower, and there are many similar HERVs. Humans are subject to breast cancer and chronic myeloid leukemia (CML), two diseases of unknown etiology corresponding to those found in the named species. (Two friends just happen to be afflicted with the named diseases this year.) The evidence that something very similar is going on in humans has become overwhelming. Researchers can exhibit sequences, virions, inserted provirus, proteins, antibodies and God knows what else. They cannot demonstrate the rapid replication in a laboratory required by the referenced paper, and if they did it would almost certainly be a contaminant. To reiterate: if a human retroviral disease passed near birth involved rapid replication it would kill hosts before being passed on at sexual maturity. This implies it could not be an endemic human infection resembling endemic retroviral infections in other species.
As said before, if the host has no ERV corresponding to an exogenous retrovirus the situation is much simpler. This specifically applies to HIV and HTLV. HIV-1 also replicates too rapidly to allow about 99% of hosts infected near birth to reproduce. This makes it plausible to assume the criteria for tumor viruses advocated in that paper were chosen to make sure that research on HIV maintained total control of funding in retrovirology research.
Consider what evidence those studying cancer have already produced. If retroviral sequences and resulting expressed proteins might come from HERVs we can essentially cross off all that work. Any sequence which rises above the background noise far enough to meet requirements will violate the rule I've stated about slow replication. You might consider variation in proviral insertion sites between individuals to be clear evidence of a slow infection still in progress, but that is being rejected in that paper, even though it is mentioned. What is now required is observation of an "insertion event". I'm not sure how you would observe such an event in the environment of human cells carrying hundreds of HERVs, with some actively transcribed. This takes us back to the requirement of replication-competent virions and reification of the process of retroviral infection. What had seemed at first to be discussion of new criteria simply ends by reiterating those criteria previously demanded.
I begin to think the message is really that virologists have little positive to contribute to cancer research, an hypothesis for which I can produce abundant documentation.
HIV has produced a particularly nasty epidemic disease which would kill about 99% of those infected without medical treatment. I am not advocating we stop research on the subject. It would, however, make sense to evaluate current research directions and accomplishments.
With the disclosure that the two Boston patients have gone back on HAART we can say that the only possible cure would be the "Berlin patient" who got a new immune system from an "elite controller" who happened to match tissue types. This kind of exceptional circumstance prevents this treatment from scaling. Without some means of eradicating the disease in infected patients we might need to start planning for a world in which the entire population is maintained on ARVs for life.
Just how do you eradicate an infection by a virus which has inserted about a trillion copies of itself in the chromosomes of host cells? You don't, unless you also have a way of eradicating an infection in which provirus is inherited in a trillion ways. The problem of curing AIDS is closely related to the problem of diseases caused by HERVs. This makes current funding disputes especially stupid.
The last 30 years of research on retroviruses has not exactly been an overwhelming success even w.r.t. AIDS. Most ARV drugs were found by sheer trial and error, little different from research which produced Salvarsan and Prontosil. The important contribution of virological research came in three parts: isolation of the virus, culturing the virus, detection of viral infection via antibodies and PCR. All three took place quite some time ago. The role of expert opinion in guiding research since has not been exactly prescient, except in the sense that the best predictor for future funding is past funding. Nowhere will you find people saying "forget what we told you 5 years ago" (or 20 or 30 years ago). In the toxic environment in which battles over research funding take place any admission of error is fatal. The result is near stasis when it comes to translating research into treatments for human diseases, or even explaining etiology of known human diseases.
(This is not to say there have been no serious errors in biomedical virology research. I've already discussed one horrible example with implications for wild bird species we still don't understand.)
If I haven't stirred up enough trouble already, here is one final blast. Anyone remember the flap over Bovine Spongiform Encephalopathy (BSE or "mad cow disease")? I could point to a corresponding spongiform encephalopathy in mice caused by a murine retrovirus. I can also reference recent work on activation of ERVs in prion disease. Now, consider what I said about eggs using chickens to make more eggs. Does this suggest that the "nude virus" once hypothesized for BSE was present all along as an ERV? Proteins with sequences derived from ERVs do show up in lesions of prion diseases. Ruling out the presence of active virions would have no effect on transmission using ERVs. The triggers for ERV expression could be much smaller and harder to find. Nobody knew this at the time BSE had its 15 minutes of fame. Experiments on which official reassurances were based wouldn't come close to meeting current standards for avoiding contamination by viruses and nucleic acid sequences, and badly underestimated the importance of ERVs.
