Paleovirology-- HIV-1s ancestors just got older

Lentiviruses, we think, are evolutionarily young.

For example-- you have retroviruses in your DNA (ERVs) from other genuses (genii?) genera of retrovirus that are millions and millions of years old. So we know those other genuses (genii?) genera are millions and millions of years old.

Lentiviruses, on the other hand, havent left us many of these kinds of fossils for us to figure out their age. We can still put out a good estimate, though. For instance-- SIV is all over the place in non-human primates in Africa. And while its ubiquitious in Africa (implying 'old'), its not found in non-human primates in Asia, or New World monkeys. It almost certainly wasnt around when the common ancestors of all these organisms were cohabitating. So, SIV is old, but not that old, evolutionarily. Previous estimates put SIV into chimpanzees in between 1266-1685.

Yeahhh... Looks like that number has been pushed back, for chimpanzees, about *squints at phylogenetic tree* 23,000 years. O.o

Island Biogeography Reveals the Deep History of SIV

This group of scientists looked for SIV in primates that had been separated from the African mainland for ~10,000 years. When they used those new sequences to recalibrate SIVs molecular clock, things got pushed back. Waaaaay back. SIV could be 32,821-132,780 (best estimate 76,794) years old.

SIV is, for the most part, non-pathogenic in African primates. Animals that can survive infection until reproductive age have been selected for, as well as variants of the virus that are less pathogenic (they reproduce and spread just fine, but they dont kill you ASAP). Countless primates had to die for this to happen. Darwinian medicine.

So it annoys me when, in a Science paper, they say something like this:

Although evidence of codivergence between SIV and its natural hosts is still lacking, our results are suggestive that generally low pathogenicity of SIV is likely a consequence of long-term host- virus coevolution. A similar accommodation between HIV and humans should therefore not be expected to arise soon.

Um, no, it should not be expected to arise ever, in humans. I mean, freak chance it could happen, but it wont happen for the same reasons it happened in non-human primates. Modern humans find Darwinian medicine unethical.

We do not say "Oh? You have HIV-1? Lets see what happens!" We give people antiretrovirals to prevent horizontal transmission and to 'artificially' extend the lives of HIV+ individuals. We give pregnant mothers antiretrovirals to prevent vertical transmission, thus even if Mom had The Good Genetics and was infected with The Good Virus, she would not pass it on to her offspring (who presumably would also have aspects of The Good Genetics).

Im not anticipating the evolutionary truce that African non-human primates have with their SIVs in humans at any point in the future. But this is still cool paleovirology :)

More like this

New Scientist has a short piece on this paper and seem to come to the opposite conclusion than the authors did saying:

"Because SIV does not cause AIDS in its primate hosts, some have speculated that HIV too might stop being lethal within a few centuries â but the discovery that SIV has had millennia to evolve into peaceful coexistence dashes these hopes."

I wonder why the last paragraph of that paper wasn't taken out during the editing process.

By Poodle Stomper (not verified) on 17 Sep 2010 #permalink

Poodle Stomper -- the quote you provide from New Scientist seems to be exactly the same as the what the authors of the Science paper were saying.

ERV -- In light of the comment quoted above from New Scientist, suggesting that others have been suggesting that HIV might stop being lethal within a few centuries, it doesnt seem patently unreasonable for the authors of this paper to point out that their findings should count against that suggestion.

"Um, no, it should not be expected to arise ever, in humans."

Um, no. Aggressively treating infections strongly selects for variants that cause asymptomatic infections, hastening attenuation over time.

This might be a stupid question for virology folks, but maybe someone can explain it anyway : When an exogenous retrovirus becomes endogenous, does it still cause disease(as in, the disease it causes while it's exogenous), or, at what point does it stop causing disease in the host ?

It's not a stupid question. They can and sometimes do. Sometimes they don't. The active ERV in AKR/J mice, for example leads to lymphomas, despite being endogenous. Also see ERV's post on the Koala ERV. However, other endogenized retroviruses are not necessarily active.

Our cells can chemically alter retroviral genomes by adding methyl groups (a carbon with three hydrogens) to the DNA sequences, silencing them. Sequences that are silenced in this way are no longer under evolutionary pressure to maintain a functional sequence and thus acquire mutations that can disrupt it. In this way many ERVs become nonfunctional simply because over time they've mutated to the point where they no longer "make sense" to the cell's machinery. As far as I've read, the vast majority of our ERVs are nonfunctional, although "fixing" mutations that rendered them silent can yield an active retrovirus (as has been done in a lab with the HERV-K sequence)

By Poodle Stomper (not verified) on 17 Sep 2010 #permalink


"the quote you provide from New Scientist seems to be exactly the same as the what the authors of the Science paper were saying."

You are correct. I misread ERV's quote to say "A similar accommodation between HIV and humans should therefore be expected to arise soon." and missed the "not". Thanks for the correction!

By Poodle Stomper (not verified) on 17 Sep 2010 #permalink

John makes an interesting point. If indeed we treat those infections only at later stages in the progression (for example at CD4s of under 350) with HAARTs, it would in fact lead to less therapy to individuals with attenuated strains that are incapable/less efficient at lowering CD4 counts. These people would be more likely to transmit the virus (assuming that the attenuation does not affect transmissibility) as compared to those treated with HAARTs whose VL have been suppressed. Either way, I don't think we have enough information, really, to say for sure whether or not attenuation will "definitely" arise.

By Poodle Stomper (not verified) on 17 Sep 2010 #permalink

Along with what John and Poodle Stomper said, just because treatments exist for viral infections doesn't mean there is no selection for changes in humans that result in reduced or eliminated symptoms, or in the virus for lower harm to the host.

In fact, some treated individuals might benefit more from treatment by also having some genetic advantage. The fact that they were not killed outright allows their descendants to fine tune a slight advantage into complete attenuation.

Your understanding of evolutionary analyses is suprisingly unsophisticated. You might want to consider taking and introductory class in evolution.

John-- Treating people with antiretrovirals selects for variants that are 1) drug resistant, 2) high replicative fitness. Fitness is associated with both transmission and disease progression. Transmission of drug resistant variants is increasing world-wide, implying that what youre suggesting isnt happening. I totally get what youre saying though-- that could happen, but I dont think it is happening :-/

Beebee-- Thats nice. Your comment was entirely devoid of content and contributed nothing to the discussion. You might want to consider lurking more.


On point 2), I was under the impression that "high replicative fitness" is what is *always* selected for.

Also, substitutions increasing replicative fitness in the presence of ARVs can impair replicative fitness in their absence, like M184V in RT for 3TC. Now if these strains are indeed getting transmitted, and this reduced fitness causes slower disease progression, wouldnt this be an instance of treatment selecting for slow progression? The mechanism is different to Johns though.

And whats the deal with "SIV does not cause AIDS in its primate hosts"? What happened to this result:

Just saw that, besides the New Scientist quote, the "SIV does not cause AIDS" statements were tempered with "generally" or "for the most part". My bad. :)

When we are talking about viruses in general, a virus doesnt 'want' to kill/sicken its host before it can spread to a new host. So too much replication is bad. And then the ERVs in our genome dont replicate at all, but every human on Planet Earth is 'infected' :P

When we are talking about HIV-1 specifically-- disease progression is associated with viral fitness, and transmission is associated with viral fitness (in that the variants that get transmitted are more fit than the variants within the transmitters quasispecies that did not transmit). This might make sense if you are thinking of 'fitness' as a numbers game-- the most fit makes the most babby viruses, so has a better chance of being transmitted. Nope! With a quasispecies, the 'most fit' variants are a minority population. Fitness in a quasispecies is like a bell-curve. If transmission is a stochastic event-- just a numbers game, then moderately fit variants would be transmitted... but its the fit ones!

Drug resistance-- Yes. The initial mutations for antiretroviral escape pretty much universally come at a fitness cost. But secondary compensatory mutations eventually make up for it. People infected with (or develop) drug-resistant HIV-1 progress to AIDS like everyone else if their treatment isnt modified, and resistance is maintained and can be transmitted.

... Did that make any sense? lol! Ive been running gels all day...

Hi again. Sense was made. :)

>People infected with (or develop) drug-resistant HIV-1 progress to AIDS like everyone else if their treatment isnt modified

Yip, but they might progress slower. :) I stumbled upon this:…
Short version: in patients with transmitted resistance, those with M184V indeed have lower pre-therapy viral loads.
This might translate to slower progression, although I'm not sure how long it takes before reversion to wildtype occurs.

Also, I get the quasispecies thing. Mostly. How robust is the 'only the fittest get transmitted' result? Is it *always*? Or just mostly? Also, is that fitness measured in the transmitter or the transmitee? Do we expect the difference (maybe from host HLA type or something) to be important? Can you point me to a decent paper?

>a virus doesnt 'want' to kill/sicken its host before it can spread to a new host
True, but selection within a single host should inevitably be for increased replication rates? Do you know of any work addressing the relation between the apparently contradictory within-host and across-host-population selection effects?

Goodnight for now.