Many sports fans will recogize the term "homefield advantage". Even I, who don't always know which teams are which, am aware that some consistently do better at home than away. This may depend on actual physical characteristics of that home field, which are less familiar to opposing teams, like the Green Monster in Fenway Park. It may also come from the lack of a need to travel to home games, or from enthusiastic fans who will threaten violence against opposition players or unpopular decisions. What is less widely known is that some viruses also experience homefield advantage.
Retroviruses are named for their ability to reverse the normal transcription process, turning RNA sequences into corresponding DNA sequences, which they insert in host chromosomes. If you take an example which is new in humans, like HIV-1, it is clear the inserted provirus is an advantage for the virus and not the host. Part of the reason is that the variant strains which are most successful in a particular host will come to dominate the population of viral sequences in that host. This is a standard evolutionary story of differential reproductive advantage. In the case of the virus you can't actually say that it is living, so you have to count replication of viral sequences as reproduction. Without a living host the virus can do nothing.
A second feature of retroviruses is that they package two strands of RNA in each of their virions. During reverse transcription it is fairly probable that a crossover event will occur, causing transcription to shift from one strand to another. If the two strands are identical this will not matter. If the two strands differ, there will be a recombination event. The result will be something like a hybrid of the two sequences, part of one and part of the other. The probability that a crossover will occur sometime during the transcription of an 8,000 base pair genome is fairly high. Whether this will result in recombination depends on the nature of those two sequences packaged in a single virion.
While many viruses evolve at breathtaking rates, retroviruses have an extraordinary ability to evolve within an individual host, and to preserve any evolutionary advantage in the host's own chromosomes. The result in the case of HIV-1 is that the strains which differ the most have sequences with only about 50% homology to each other. If we didn't have all the intervening sequences to show the connection, we would call these distinct species.
Rarely, a retrovirus will insert its genes in a germ-line cell rather than an ordinary somatic cell. Even more rarely, these genes will be passed on to offspring of the host without killing them before they reproduce. When both these events take place the provirus has become endogenous.
To avoid killing the new host it is more common than not for the endogenous retrovirus to contain defects. This is regularly said to mean the virus is dead, (as if it were alive in the first place.) Unfortunately, for the argument that these relics of past evolution can be ignored, they continue to show surprising signs, if not of life, then at least activity. It turns out that damn near anything is currently claimed to activate these relics when a researcher needs to invoke some cause. They are also said to recombine to form replication-competent viruses out of defective sequences.
This argument has a few problems, especially when ERVs involved are seriously defective fragments. It is as if a Frankenstein monster were to stitch itself together, and zap itself into life, without the assistance of Victor Frankenstein and Igor, plus lightning bolts. Recombination between random gene fragments in chromosomes is exceedingly slow. Forming a complete functional pathogen this way is extremely improbable in a short enough time to allow it to be observed in a laboratory. On the other hand we see many reports stating that this has apparently happened, typically within days or weeks of the start of some experiment. The conventional answer is that the result must be a contaminant.
In research publications, the contamination argument is actually older than retroviruses. Prior to 1970 there were strong reasons to deny that retroviruses, as currently defined, could exist. It took the independent discovery, in two laboratories, of an enzyme called reverse transcriptase to show that it was possible for a virus to reverse the flow of information from DNA to RNA to proteins. At this point one could start to look for viruses using this strategy, and call them retroviruses, without being shouted down.
Now consider what researchers saw before this was known. They could examine a culture very carefully without detecting any evidence of viral replication. At a later time they could find clear evidence of RNA viruses which had been absent. The obvious answer was that the virus was a contaminant. Without the understanding that provirus could lie latent in chromosomes, either by its own mechanisms or through host epigenetic control, there was scarcely any alternative. Generally speaking RNA sequences are less stable than DNA sequences. Without constant replication it is hard for RNA viruses to persist. Retroviruses found a loophole in this argument.
We also find many reports showing an association between activated endogenous retroviruses (ERV) and diseases of unknown etiology. (In humans ERVs are called HERVs.) In the case of breast cancer you can build a a tall stack of papers from the controversy over the existence of a human mammary tumor virus corresponding to MMTV, which causes 95% of mammary tumors in mice. This would be a beta retrovirus, and it just happens that the progenitor of HERV-K is believed to have been a beta retrovirus.
If this connection is valid it might help explain some success with monoclonal antibodies (trastuzumab, cetuximab, denosumab) and ARVs in treatment. The suspected retroviruses infect immune cells which play an ambiguous role in breast cancer pathology.
Recent research has revealed that one HERV designated HERV-K111 has more than 100 full-length copies in human genomes. All appear to be defective, but the fact that these were overlooked for so long is disturbing. (The hidden copies were in centromeres.) We also have paleontological evidence in the form of preliminary genomes for Neanderthals and Denisovans. These now appear to have been human subspecies which could interbreed with modern humans. Both vanished about 30,000 years ago.
(I refuse to speculate on the reason both vanished because I am related to the main suspects.)
Basically, it looks like HERV-K111 left one copy in chimpanzee genomes, which separated from human lineages millions of years ago. In those ancient human subspecies we find a mere handful of copies. (I'm not going to say the count is exact.) Both samples from which those genomes were derived were less than 50,000 years old. It appears that HERV-K111 has been very active in anatomically modern humans at some time after that.
Looking deeper, it also appears that you don't find the same number of copies in different living people, or copies with exactly the same defects. Trace the ancestry of those with differences to a recent common ancestor, and you have an upper bound on time in which that HERV has been active. With current fragmentary evidence, it looks like that bound is measured in centuries or less, not thousands or millions of years. There is good reason to believe this HERV is still active.
When you find an ERV in an individual you know two things: 1) an ancestor of that individual was infected by a retrovirus; 2) that ancestor survived long enough to reproduce, as did (some of) its offspring. Any evolutionary effect on fitness is limited to survival through reproductive age. If you are able to raise offspring, that is all that matters in this brutal logic. For most humans throughout history, not even considering prehistory, this meant that diseases which struck after about age 35 were irrelevant to differential reproductive success.
A quick look at on-line pictorial resources will reveal that human mammary hyperplasia contributes to reproductive success. A virus which induces local immune tolerance for the proteins in milk, and the cells that produce them, may well be of evolutionary value here, even if that same immune dysfunction makes cancer more likely after reproduction.
Local immune tolerance for gametes is likely to result in increased fertility, in both males and females. These cells, with their foreign DNA, are constantly at risk from adult immune response. So are zygotes resulting from fertilization. It takes no stretch of imagination to believe that limited retroviral infection, with local immune suppression, results in both increased fertility and later cancer in the affected organs.
From the standpoint of the virus, tying itself to host reproduction was a stroke of genius.
This brings us back to an important point: the virus can only survive as long as the host, unless it is passed from one host to a new one. While most viruses must replicate rapidly to do this, another strategy is open to retroviruses -- they can become ERVs. Once this happens, they have an interest in keeping the carrier of their genes alive and reproductive, (even if they don't know this.) Those that violate this rule tend to disappear along with the host. Those that act accordingly are now in a comfortable position where they will be passed on to offspring without taking any independent action. This is where an invading exogenous retrovirus becomes a threat to the virus as well as the host, and where ERVs can exploit a home-field advantage.
The key to this may be said to reside in the repeated sequences at either end of a retroviral genome, LTRs. These contain receptor elements which affect the probability the viral genes will be transcribed. It is significant that these sequences are the last parts of an ERV to disappear as mutations degrade the ERV.
The result is that a similar invading retrovirus will trigger transcription of the resident ERVs at the same time its own inserted provirus is transcribed. The resulting mRNA will bounce around in the cell until it is turned into polypeptide sequences and proteins, or is packaged in a new virion. Similar strands of RNA will pair up when they are put in a virion.
If the RNA from two ERVs is packaged in a virion, and both are defective, the virion will be useless to the virus. If one is defective, the probability it will be transcribed depends on the toss of a coin at the start of transcription, plus the frequency of crossover after this. By flooding a cell with defective sequences at the time an invading virus provirus is transcribed ERVs drastically reduce the chance a replication-competent virion will be produced. Furthermore, those virions which are produced may still insert defective provirus due to recombination. This will increase the number of defective retroviral sequences in somatic cells, and may eventually produce new ERVs in their own right, if inserted in germ-line cells. The ERV not only maintains its home-field advantage, it increases it.
What about the possibility of recreating the ancestral retrovirus responsible for the ERVs? The one advantage here is that we know the first host survived that infection long enough to reproduce. The invading retrovirus may be more lethal. Converting an invading retrovirus into the progenitor of an ERV is a reasonable strategy to deal with the problem, albeit less than perfect.
When the invading retrovirus is highly similar to the ERV this works fairly well. When the invading virus is very different, but triggers transcription of provirus under the same conditions, we get the widespread activation of HERVs seen with HIV infection, but without the benefits.
Other recent research is said to show that xenografting experiments are uniquely dangerous sources of recombinant retroviruses. Reports indicate that reverse transcription was observed 8 days after the start of the experiment. This is an upper bound on the time to form a replication-competent recombinant, because every detection method has a threshold. In the scenario described, where seriously defective fragments of retroviruses combine to produce a replication-competent virus, it is unlikely that the first recombination event would result in a virus which was more than marginally competent. Such a virus would be hard to detect. This results in a lower bound on the time remarkably close to zero.
Suppose the time really is zero? Suppose a replication-competent virus was already present prior to the recombination event. Can retroviruses lie latent and undetected for 8 days? They can lie latent for over 20 years, though you won't be able to demonstrate this in laboratory animals which don't live that long. A latent provirus produces no reverse transcription activity.
If you knew the sequence to look for you might be able to detect such a latent provirus, assuming you could distinguish it from ERVs. If we are talking about finding one latent replication-competent provirus with an unspecified sequence in a living mouse I think most rational people would say "don't be silly, of course you can't be sure." Those who claim certainty should be suspect. Certainty does, however, serve polemical purposes.
A second possibility is that introduced sequences resemble a "patch" for a computer program. If both the previous host, from which a sequence is derived, and the new host contain common ERVs with modest defects, it might be possible for much less than a complete virion to resurrect a "dead" ERV. A misplaced stop codon is only three base pairs long, a frameshift is caused by a sequence which is not a multiple of three base pairs. Both are considered "lethal" mutations for the virus, yet both could be corrected by a short sequence. It may be much easier to transmit such a sequence than an intact virion. The preconditions here are almost certain to be true if the new host is an offspring of the current host. This would be characteristic of infections passed perinatally via nursing, as happens in other mammals with beta and gamma retroviruses, and in humans with delta retroviruses. (There are problems with this idea, and I'll let others cope with them.)
After years of being told that HERVs are "dead", and that recombination is both random and slow, it is time to reassess the idea that a single replication-competent provirus in an environment dominated by ERVs will breed true. If that is wrong, the whole interpretation placed on a wide range of research results should be questioned.
Here I've stated an alternative interpretation, that recombination in such an environment is more likely than "breeding true". (This may not be original; there is a lot of literature I have not read.) This also suggests that the production of replication-competent retroviruses resembling progenitors may be a defensive action by the "home team" in response to infection by an exogenous retrovirus. What could be a sensitive test for the existence of previously undetected retroviruses has been used to discredit the idea that such viruses exist. This may well have set back research into the etiology of a variety of "real" diseases for the last 40 years.
Retroviruses are named for their ability to reverse the normal transcription process, turning RNA sequences into corresponding DNA sequences, which they insert in host chromosomes. If you take an example which is new in humans, like HIV-1, it is clear the inserted provirus is an advantage for the virus and not the host. Part of the reason is that the variant strains which are most successful in a particular host will come to dominate the population of viral sequences in that host. This is a standard evolutionary story of differential reproductive advantage. In the case of the virus you can't actually say that it is living, so you have to count replication of viral sequences as reproduction. Without a living host the virus can do nothing.
A second feature of retroviruses is that they package two strands of RNA in each of their virions. During reverse transcription it is fairly probable that a crossover event will occur, causing transcription to shift from one strand to another. If the two strands are identical this will not matter. If the two strands differ, there will be a recombination event. The result will be something like a hybrid of the two sequences, part of one and part of the other. The probability that a crossover will occur sometime during the transcription of an 8,000 base pair genome is fairly high. Whether this will result in recombination depends on the nature of those two sequences packaged in a single virion.
While many viruses evolve at breathtaking rates, retroviruses have an extraordinary ability to evolve within an individual host, and to preserve any evolutionary advantage in the host's own chromosomes. The result in the case of HIV-1 is that the strains which differ the most have sequences with only about 50% homology to each other. If we didn't have all the intervening sequences to show the connection, we would call these distinct species.
Rarely, a retrovirus will insert its genes in a germ-line cell rather than an ordinary somatic cell. Even more rarely, these genes will be passed on to offspring of the host without killing them before they reproduce. When both these events take place the provirus has become endogenous.
To avoid killing the new host it is more common than not for the endogenous retrovirus to contain defects. This is regularly said to mean the virus is dead, (as if it were alive in the first place.) Unfortunately, for the argument that these relics of past evolution can be ignored, they continue to show surprising signs, if not of life, then at least activity. It turns out that damn near anything is currently claimed to activate these relics when a researcher needs to invoke some cause. They are also said to recombine to form replication-competent viruses out of defective sequences.
This argument has a few problems, especially when ERVs involved are seriously defective fragments. It is as if a Frankenstein monster were to stitch itself together, and zap itself into life, without the assistance of Victor Frankenstein and Igor, plus lightning bolts. Recombination between random gene fragments in chromosomes is exceedingly slow. Forming a complete functional pathogen this way is extremely improbable in a short enough time to allow it to be observed in a laboratory. On the other hand we see many reports stating that this has apparently happened, typically within days or weeks of the start of some experiment. The conventional answer is that the result must be a contaminant.
In research publications, the contamination argument is actually older than retroviruses. Prior to 1970 there were strong reasons to deny that retroviruses, as currently defined, could exist. It took the independent discovery, in two laboratories, of an enzyme called reverse transcriptase to show that it was possible for a virus to reverse the flow of information from DNA to RNA to proteins. At this point one could start to look for viruses using this strategy, and call them retroviruses, without being shouted down.
Now consider what researchers saw before this was known. They could examine a culture very carefully without detecting any evidence of viral replication. At a later time they could find clear evidence of RNA viruses which had been absent. The obvious answer was that the virus was a contaminant. Without the understanding that provirus could lie latent in chromosomes, either by its own mechanisms or through host epigenetic control, there was scarcely any alternative. Generally speaking RNA sequences are less stable than DNA sequences. Without constant replication it is hard for RNA viruses to persist. Retroviruses found a loophole in this argument.
We also find many reports showing an association between activated endogenous retroviruses (ERV) and diseases of unknown etiology. (In humans ERVs are called HERVs.) In the case of breast cancer you can build a a tall stack of papers from the controversy over the existence of a human mammary tumor virus corresponding to MMTV, which causes 95% of mammary tumors in mice. This would be a beta retrovirus, and it just happens that the progenitor of HERV-K is believed to have been a beta retrovirus.
If this connection is valid it might help explain some success with monoclonal antibodies (trastuzumab, cetuximab, denosumab) and ARVs in treatment. The suspected retroviruses infect immune cells which play an ambiguous role in breast cancer pathology.
Recent research has revealed that one HERV designated HERV-K111 has more than 100 full-length copies in human genomes. All appear to be defective, but the fact that these were overlooked for so long is disturbing. (The hidden copies were in centromeres.) We also have paleontological evidence in the form of preliminary genomes for Neanderthals and Denisovans. These now appear to have been human subspecies which could interbreed with modern humans. Both vanished about 30,000 years ago.
(I refuse to speculate on the reason both vanished because I am related to the main suspects.)
Basically, it looks like HERV-K111 left one copy in chimpanzee genomes, which separated from human lineages millions of years ago. In those ancient human subspecies we find a mere handful of copies. (I'm not going to say the count is exact.) Both samples from which those genomes were derived were less than 50,000 years old. It appears that HERV-K111 has been very active in anatomically modern humans at some time after that.
Looking deeper, it also appears that you don't find the same number of copies in different living people, or copies with exactly the same defects. Trace the ancestry of those with differences to a recent common ancestor, and you have an upper bound on time in which that HERV has been active. With current fragmentary evidence, it looks like that bound is measured in centuries or less, not thousands or millions of years. There is good reason to believe this HERV is still active.
When you find an ERV in an individual you know two things: 1) an ancestor of that individual was infected by a retrovirus; 2) that ancestor survived long enough to reproduce, as did (some of) its offspring. Any evolutionary effect on fitness is limited to survival through reproductive age. If you are able to raise offspring, that is all that matters in this brutal logic. For most humans throughout history, not even considering prehistory, this meant that diseases which struck after about age 35 were irrelevant to differential reproductive success.
A quick look at on-line pictorial resources will reveal that human mammary hyperplasia contributes to reproductive success. A virus which induces local immune tolerance for the proteins in milk, and the cells that produce them, may well be of evolutionary value here, even if that same immune dysfunction makes cancer more likely after reproduction.
Local immune tolerance for gametes is likely to result in increased fertility, in both males and females. These cells, with their foreign DNA, are constantly at risk from adult immune response. So are zygotes resulting from fertilization. It takes no stretch of imagination to believe that limited retroviral infection, with local immune suppression, results in both increased fertility and later cancer in the affected organs.
From the standpoint of the virus, tying itself to host reproduction was a stroke of genius.
This brings us back to an important point: the virus can only survive as long as the host, unless it is passed from one host to a new one. While most viruses must replicate rapidly to do this, another strategy is open to retroviruses -- they can become ERVs. Once this happens, they have an interest in keeping the carrier of their genes alive and reproductive, (even if they don't know this.) Those that violate this rule tend to disappear along with the host. Those that act accordingly are now in a comfortable position where they will be passed on to offspring without taking any independent action. This is where an invading exogenous retrovirus becomes a threat to the virus as well as the host, and where ERVs can exploit a home-field advantage.
The key to this may be said to reside in the repeated sequences at either end of a retroviral genome, LTRs. These contain receptor elements which affect the probability the viral genes will be transcribed. It is significant that these sequences are the last parts of an ERV to disappear as mutations degrade the ERV.
The result is that a similar invading retrovirus will trigger transcription of the resident ERVs at the same time its own inserted provirus is transcribed. The resulting mRNA will bounce around in the cell until it is turned into polypeptide sequences and proteins, or is packaged in a new virion. Similar strands of RNA will pair up when they are put in a virion.
If the RNA from two ERVs is packaged in a virion, and both are defective, the virion will be useless to the virus. If one is defective, the probability it will be transcribed depends on the toss of a coin at the start of transcription, plus the frequency of crossover after this. By flooding a cell with defective sequences at the time an invading virus provirus is transcribed ERVs drastically reduce the chance a replication-competent virion will be produced. Furthermore, those virions which are produced may still insert defective provirus due to recombination. This will increase the number of defective retroviral sequences in somatic cells, and may eventually produce new ERVs in their own right, if inserted in germ-line cells. The ERV not only maintains its home-field advantage, it increases it.
What about the possibility of recreating the ancestral retrovirus responsible for the ERVs? The one advantage here is that we know the first host survived that infection long enough to reproduce. The invading retrovirus may be more lethal. Converting an invading retrovirus into the progenitor of an ERV is a reasonable strategy to deal with the problem, albeit less than perfect.
When the invading retrovirus is highly similar to the ERV this works fairly well. When the invading virus is very different, but triggers transcription of provirus under the same conditions, we get the widespread activation of HERVs seen with HIV infection, but without the benefits.
Other recent research is said to show that xenografting experiments are uniquely dangerous sources of recombinant retroviruses. Reports indicate that reverse transcription was observed 8 days after the start of the experiment. This is an upper bound on the time to form a replication-competent recombinant, because every detection method has a threshold. In the scenario described, where seriously defective fragments of retroviruses combine to produce a replication-competent virus, it is unlikely that the first recombination event would result in a virus which was more than marginally competent. Such a virus would be hard to detect. This results in a lower bound on the time remarkably close to zero.
Suppose the time really is zero? Suppose a replication-competent virus was already present prior to the recombination event. Can retroviruses lie latent and undetected for 8 days? They can lie latent for over 20 years, though you won't be able to demonstrate this in laboratory animals which don't live that long. A latent provirus produces no reverse transcription activity.
If you knew the sequence to look for you might be able to detect such a latent provirus, assuming you could distinguish it from ERVs. If we are talking about finding one latent replication-competent provirus with an unspecified sequence in a living mouse I think most rational people would say "don't be silly, of course you can't be sure." Those who claim certainty should be suspect. Certainty does, however, serve polemical purposes.
A second possibility is that introduced sequences resemble a "patch" for a computer program. If both the previous host, from which a sequence is derived, and the new host contain common ERVs with modest defects, it might be possible for much less than a complete virion to resurrect a "dead" ERV. A misplaced stop codon is only three base pairs long, a frameshift is caused by a sequence which is not a multiple of three base pairs. Both are considered "lethal" mutations for the virus, yet both could be corrected by a short sequence. It may be much easier to transmit such a sequence than an intact virion. The preconditions here are almost certain to be true if the new host is an offspring of the current host. This would be characteristic of infections passed perinatally via nursing, as happens in other mammals with beta and gamma retroviruses, and in humans with delta retroviruses. (There are problems with this idea, and I'll let others cope with them.)
After years of being told that HERVs are "dead", and that recombination is both random and slow, it is time to reassess the idea that a single replication-competent provirus in an environment dominated by ERVs will breed true. If that is wrong, the whole interpretation placed on a wide range of research results should be questioned.
Here I've stated an alternative interpretation, that recombination in such an environment is more likely than "breeding true". (This may not be original; there is a lot of literature I have not read.) This also suggests that the production of replication-competent retroviruses resembling progenitors may be a defensive action by the "home team" in response to infection by an exogenous retrovirus. What could be a sensitive test for the existence of previously undetected retroviruses has been used to discredit the idea that such viruses exist. This may well have set back research into the etiology of a variety of "real" diseases for the last 40 years.
