Mystery Rays from Outer Space has a good essay on the evolution of spumaviruses ("foamy viruses") which are cytologically fatal in the lab, but which are latent in most body cell types in the nonhuman species they inhabit. It turns out that they have evolved over a long time to be nonvirulent by only infecting epithelial cells that are somatic dead-ends anyway, as they are about to be shed, and so they don't measurably affect the fitness of their hosts.
This leads directly to the question: why do some pathogens (viruses, bacteria, fungal infections etc.) evolve to low virulence ("virulence" is a term that basically means "does nasty stuff to the host"). at all? Wouldn't it be better to rip the guts out of the resources in a kind of biological strip mining and then get the hell out of there before the immunological EPA comes and fines your arse? That, it turns out, is an interesting question.
The traditional view was that given enough time, pathogens would evolve towards commensuality by natural selection and coadaptation between host and infection. This was the view presented by Macfarlane Burnet and his collaborator Peter Medawar, who won the Nobel prize for medicine in 1960 for their work on the immune system. Mac, as he was known, developed the current theory of immunology in vertebrates, known as the "clonal selection theory". I know he did - I have seen his notebook pages, two of them, where he first came up with it.
But Mac and Medawar were not infallible, Mac particularly in his social views in Credo and Comment, about which another time. In this case, it seems that some diseases remain virulent over evolutionary time, while others are inexplicably benign like these spumaviruses when they ought, physiologically, be highly virulent. Why do they only infect the epithelials?
An answer, unjustly ignored in my view, was given by Paul Ewald, in a highly significant book for evolutionary theory as well as disease epidemiology, Evolution of Infectious Disease. Ewald makes the following point: in interactions over evolutionary time, the genetic interests of two species will coincide only when their reproduction is linked. That is, for commensuality to evolve, it has to be in the interests of the pathogen to minimse the damage done to the host, which effectively means that for the pathogen to infect another host they have to be passed on through the host's reproductive process.
Otherwise, competition between variants of the pathogen will select for those who can get themselves replicated and passed on rapidly. Such pathogens are considered highly virulent. But if the host prevents transmission except during reproduction - as happens with gastric commensuals, or gut flora - then it is in the interests of the pathogen to do as little harm so as to survive until it gets passed on.
Effectively, this is the observation that the epidemiology of the host is the genetics of the pathogen. The models of epidemiology translate directly to the population genetics of the various pathogens. Ewald made some interesting observations about the nature of other biological functional categories as well: if a thousand vampire bats will kill a cow, is that parasitism or predation? If a lion ate part of a gazelle but the gazelle survived, would that be parasitism? Ewald shows this way that some of our favourite categories in biology are in fact relative and subjective. Likewise terms like "commensual" and "pathogen". What is actually going on is the interplay of the interests of organisms from different species, and selection will tend to shape the populations according to what is the local optimum.
This has major implications for the treatment of disease. First of all, hygiene is the single best way to reduce the virulence of disease. For example, with HIV, practising safe sex, using condoms to make it harder for the disease to spread, means that eventually, the variant of the virus that spreads will be less virulent, since variants that kill their hosts will be selected out by the death of their hosts. It's not fair for those who have the virulent forms (and we should continue to find ways to prevent their deaths), but making the transmission much harder this way, no matter how socially unpopular it may be, will bear great fruit in the long run. Free needles for IV drug users, and free condoms is mandated by public health. Right now we should be sending these through the developing world to ensure that eventually HIV will become another commensual virus, of no more account than infectious mononucleosis.
- Log in to post comments
um. "low virulence" means "does nasty stuff to the host"?
Thanks for spotting that...
Didn't May & Anderson make the same point in about 1980? They described some neat results showing that the myxoma virus in Australia had evolved to reduced virulence, after which they threw some differential equations around.
I can't stand differential equations. Don't understand 'em.
An answer, unjustly ignored in my view, was given by Paul Ewald, in a highly significant book for evolutionary theory as well as disease epidemiology
I don't think it's been ignored, John. I think it's fairly well accepted, as conventional theory, now. I talked about that (with some other examples) in an earlier post on Mystery Rays here.
@Bob (#3) - That post also discusses the evolution of myxoma viruses in some detail A critical point about myxoma evolution to "reduced" virulence is that the "reduced" virulence is still pretty damn virulent -- 50 to 70% mortality rate, in the same ballpark as Ebola and smallpox (variola major), which few people would consider apathogenic.
The increased rabbit survival in AUstralia also has a lot to do with the rabbits evolving resistance, of course; I talked about that in a post here.
Thanks for noting the post, John. The whole question of virulence evolution is fascinating, and of course much more complicated than the general simple understanding most people have.
I'm not really sure what you mean by this:
From your later statements about HIV it seems that you mean "dependent on the host's reproductive behavior to be passed on". Precious few parasites do this.
I think the solution to "why low virulence?" is simply a balance of benefits. Let's imagine two alternative strategies: (1) strip-mining the host to convert as much of its biomass as possible into baby parasites (here' i'm actually thinking 'virus', rather than 'fluke') before killing the host or being killed by it. (2) producing a long series of parasite babies at minimal cost to the host. (r- and K-selected strategies, obviously.)
From the perspective of the parasite, the 'best' strategy is the one that leaves the largest number of viable offspring (say, grandchildren). If killing the host results in more grandkids than "letting" it live, then this is the best strategy, and vice versa. This doesn't necessarily reduce to connection with the reproduction of the host (as you seem to imply--correct me if i have misread): it's related to the likelihood that a host will encounter another uninfected host and pass on the infection. Bob O'H mentioned the myxoma virus and the Australian rabbits. Dead bunnies don't infect other bunnies. Virulent virus variants (HAD to structure my sentence to let me say that!) are at a disadvantage. Nothing here about host reproductive success.
p.s. I have a better species for your analogy of lion half-eating a gazelle: cookie-cutter sharks.
http://www.amonline.net.au/fishes/fishfacts/fish/ibrasil.htm
Is this a predator or a parasite?
@ djlactin :
Dead bunnies don't infect other bunnies.
This is the common belief, and it's not really true; that's the point John (and I) want to make here. I've talked about this in much more detail on my blog (originally at http://www.iayork.com/MysteryRays/2007/08/26/rabbits-1-virus-1-evolutio… and also with a number of further posts).
The key is not that the virus benefits from letting its hosts survive. As John says (and see references on my blog) the key is optimizing viral transmission, and this doesn't necessarily mean host survival. If a virus will transmit itself to more new hosts by letting its host survive, then (absent host evolution) that's what it will move toward. If killing the host causes better transmission, then it will happily kill the host.
Myxoma is an example of this. People have heard that myxoma evolved toward reduced virulence, but they rarely realize what this actually means. Myxomavirus evolved to the same "low" virulence as such cheery little companions as smallpox and ebola virus. Even the "reduced virulence" myxomavirus kills 50-80% of its hosts.
And the reason is because of transmission. Myxoma spreads (in Australia) via fleas and mosquitos. Rabbits that are almost dead can't shoo away mosquitos or scratch fleas. The original virus killed rabbits too fast; they were dead before the mosquitos had a chance to bite them. The new, improved myxoma didn't "want" them to live. It "wanted" its hosts to be nearly dead for a long time. And that's the big change in the virus. It just let its hosts live a little longer, in a moribund state, before killing them.
Myxoma is not unique in this. There are plenty of other pathogens that optimize their transmission by being virulence. There's a fair bit of literature on this -- I have pointers to some, including reviews.
By the way, John, I had an earlier comment that seems to have been swallowed up.Did it get tagged as spam, or did I do something wrong?
My spam filter is clean, and I have published all but a couple of personal messaqes, so I think something went wrong before it got to the blog host.
Thanks for the long response. And your blog.
@liayork
We seem to be arguing the same point:
You said:
I said:
The information you provided about the Myxoma virus is very interesting and certainly educated me.
But your statement addresses only a minor of my comment:
I continued:
I agree that for parasites (read, microbes) that depend on their host's reproduction to spread (e.g. herpes virus) the best strategy is probably to adapt their virulence to allow the host to survive to reproduce (or at least, to indulge in reproductive behavior). My point though, is that this lifestyle is is not universal in viruses that infect multicellular organisms. Respiratory viruses that depend for transmission on their host coughing and sneezing form a class of counterexample. My bottom line remains as stated in the earlier post: the best strategy for a parasite is the one that produces the most "grandchildren". This strategy does not necessarily reduce to connection with the host's reproductive process.
And of course we can't assume that evolution of the parasite's reproductive strategy occurs
.
But that's a separate discussion.
@ djlactin - I'm sorry, I think I misinterpreted your "grandchildren" as host's rather than pathogen's grandchildren.
A minor point: parasites (read, microbes) that depend on their host's reproduction to spread (e.g. herpes virus) ... -- you probably mean "herpes simplex" rather than "herpes" here, and it's not really a good illustration of the allow-reproduction concept - only herpes simplex type II is sexually transmitted among the 8 human herpesviruses, yet all of them (and probably all of the millions of herpesviridae) establish life-long, latent/persistent infection. So the link between life-long infection, and spread via reproduction, isn't a strong one.
Finally -- And of course we can't assume that evolution of the parasite's reproductive strategy occurs (absent host evolution).
But that's a separate discussion.
If you're interested, I talked about the co-evolution story more at http://www.iayork.com/MysteryRays/2008/03/02/hostvirus-co-evolution/ .
@iayork
reading my post, i can see that i was not clear on this point.
and about my herpes error: thanks for the info. i'm not a virologist; only an entomologist.
And I will dip into your blog asap.
This point was made by Dawkins in The Extended Phenotype, which, it seems, came out before Ewald's book. I don't know if this was Dawkins' original insight.