Mystery Rays from Outer Space

The recent papers on TRIM5-cyclophilin fusion proteins as retroviral restriction systems,¹ and my nattering about myxomavirus the other day, reminded me of another paper I recently read.

TRIM5 is one of a family of proteins that prevent replication of certain retroviruses; the TRIM5-CYP fusion protein protects against infection so that, for example, feline immunodeficiency virus doesn’t effectively infect rhesus macaque cells. This is a specific instance of a general problem: Why are are some viruses only able to infect certain species, while others leap around between species like the chamois of the alps? This isn’t just relevant to retroviruses. Emerging virus infections are an ongoing problem, and lots of the emerging viruses are jumping into humans from some other species — HIV, Ebola, and SARS being particularly obvious examples out of many.

In many cases, of course, viruses are specifically adapted to their receptors, and another species may not have the precise receptor, or the precise shape or sequence the virus needs to get in a cell in the first place. Similarly, the virus may depend on some species-specific structure for some step in replication. But there are viruses that can cheerfully replicate in insects, birds, and mammals; fundamental replication equipment is pretty conserved.

In other cases — like the TRIM5 story — hosts may have evolved systems to restrict viruses, and only those viruses that were resistant, or that developed resistance, to the restriction systems were able to persist. In this case, host restriction is the flip side of viral immune evasion;² viruses that are unable to counteract the host’s immune adaptations go extinct in that host. Most such extinctions are invisible; species-specific viruses are (at least in some cases) the fossil evidence for ancient viruses.

Beta-herpesviruses — cytomegaloviruses, for example — are among the most species-specific virus families. (I’m approaching my point here.) Broadly speaking, the way you get a new beta-herpesvirus is when its host speciates. (The same is generally true for other herpesvirus families, but there’s more evidence for cross-species jumping of some of the other families.) How come? What is it that’s restricting mouse cytomegalovirus (MCMV) to mice, and prevents them from infecting humans? The answer was a bit of a surprise to me, actually: Apoptosis is a major factor.

If you’d asked me a couple of years ago, I’d have guess that MCMV would simply never get in to human cells — wrong receptor, I’d have suggested. I would have been wrong.³ Jurak and Brune (and here I have reached my point) showed a couple of years ago that MCMV gets into human cells just fine, ⁴ but in most cases it’s not able to do anything afterward; it doesn’t spread to neighboring cells, and eventually it just fades away from the culture.

The exceptions, as is often the case, were interesting. In two of the cell lines they tested, MCMV not only entered, but replicated and spread (though a thousand times less well than in permissive mouse cells). Those cell lines are 293 and 911 cells, and right now the virologists⁵ in my audience⁵ are nodding and saying, “Aha!” — because 293 and 911 cells have something in common: They both contain chunks of a human adenovirus genome.⁶ There are a bunch of functions included in those chunks of genomes, but to cut to the chase, it was the anti-apoptosis function that proved to be important here.

Human cells infected with MCMV undergo apoptosis, dying before the virus can replicate. Swapping various anti-apoptotic genes into human cells renders them sensitive to MCMV infection.

Blocking apoptosis is a very common, if not universal, function of viruses, so clearly apoptosis is a hugely important antiviral process. Paradoxically, that’s why I was surprised to find out that apoptosis limited MCMV replication in human cells. The apoptosis pathway is pretty conserved, and most of the viral anti-apoptotic molecules I could name off the top of my head^7,8 act on many species. (The baculovirus protein IAP (“Inhibitor of APoptosis”) — evolved to function in cells of infected insects — works just fine in mammalian cells,⁹ for example.) It’s not at all clear to me why MCMV, which can presumably inhibit apoptosis in mouse cells, can’t do so in human cells. Human CMV does block apoptosis in human cells, through more than one pathway, and swapping the HCMV gene into MCMV makes the mouse virus able to replicate in human cells. Presumably the MCMV gene(s) that have the similar function are tuned to the mouse version of the apoptosis pathway.

Incidentally, since apoptosis is an anti-tumor mechanisms as well as an antiviral, many tumors have down-regulated components of their apoptosis pathways. This, at least potentially, makes tumors targets for viruses that can only grow where apoptosis is inhibited, and this is an active area of research for the various oncolytic viruses. As well MCMV (which probably isn’t a good candidate for oncolysis, because it replicates poorly in human cells even when apoptosis is blocked) our old friend myxomavirus is at least partly restricted by apoptosis; knocking out the myxoma M-T5 protein, an anti-apoptosis gene, eliminates its pathogenicity altogether in rabbits, “making MV an excellent oncolytic candidate”. ¹⁰

I don’t want to leave the impression that apoptosis is the usual reason for host restriction. There are all kinds of other explanations. TRIM family proteins, as I said, limit retroviruses, as do APOBEC family proteins; SARS adapted to humans by (among other things) mutating its receptor to bind better to human angiotensin-converting enzyme 2; ¹¹ myxomavirus are apparently restricted to some species because their immune response modifiers are species-specific¹² (and I will probably talk more about that some time); and so on.

1. Virgen CA, Kratovac Z, Bieniasz PD, Hatziioannou T (March 4, 2008) Independent genesis of chimeric TRIM5-cyclophilin proteins in two primate species. Proceedings of the National Academy of Sciences 105.:3563-3568 10.1073/pnas.0709258105
   Wilson SJ, Webb BLJ, Ylinen LMJ, Verschoor E, Heeney JL, et al. (March 4, 2008) Independent evolution of an antiviral TRIMCyp in rhesus macaques. Proceedings of the National Academy of Sciences 105.:3557-3562 10.1073/pnas.0709003105
   Brennan G, Kozyrev Y, Hu S (March 4, 2008) From the Cover: TRIMCyp expression in Old World primates Macaca nemestrina and Macaca fascicularis. Proceedings of the National Academy of Sciences 105.:3569-3574 10.1073/pnas.0709511105[↩]
2. For rather generous uses of “immune evasion”[↩]
3. And should have known better, because it was shown 35 years ago that MCMV can enter humans cells: Kim KS , Carp RI (1972) Abortive infection of human diploid cells by murine cytomegalovirus. Infect Immun 6: 793-797 [↩]
4. Jurak, I., Brune, W. (2006). Induction of apoptosis limits cytomegalovirus cross-species infection. The EMBO Journal, 25(11), 2634-2642. DOI: 10.1038/sj.emboj.7601133[↩]
5. If any[↩][↩]
6. For 193 cells: Graham FL , Smiley J , Russell WC , Nairn R (1977) Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36: 59-74
   For 911 cells:
   Fallaux FJ , Kranenburg O , Cramer SJ , Houweling A , van Ormondt H , Hoeben RC , van der Eb AJ (1996) Characterization of 911: a new helper cell line for the titration and propagation of early region 1-deleted adenoviral vectors. Hum Gene Ther 7: 215-222 [↩]
7. A fun ice-breaker for your next cocktail party — name that viral apoptosis inhibitor![↩]
8. And this is why I don’t get invited to many cocktail parties[↩]
9. Duckett CS, Nava VE, Gedrich RW, Clem RJ, Van Dongen JL, et al. (June 3, 1996) A conserved family of cellular genes related to the baculovirus iap gene and encoding apoptosis inhibitors. EMBO J 15.:2685-94.[↩]
10. Werden SJ, McFadden G (January 2008) The role of cell signaling in poxvirus tropism: The case of the M-T5 host range protein of myxoma virus. Biochimica et Biophysica Acta (BBA) – Proteins & Proteomics 1784.:228-237 doi:10.1016/j.bbapap.2007.08.001[↩]
11. Li W, Zhang C, Sui J, Kuhn JH, Moore MJ, et al. (April 20, 2005) Receptor and viral determinants of SARS-coronavirus adaptation to human ACE2. EMBO J 24.:1634-43.
    Sheahan T, Rockx B, Donaldson E, Sims A, Pickles R, et al. (March 2008) Mechanisms of zoonotic severe acute respiratory syndrome coronavirus host range expansion in human airway epithelium. J Virol 82.:2274-85.[↩]
12. Wang F , Ma Y , Barrett JW , Gao X , Loh J , Barton E , Virgin HW , McFadden G (2004) Disruption of Erk-dependent type I interferon induction breaks the myxoma virus species barrier. Nat Immunol 5: 1266–1274[↩]

I’ve previously talked about the common misconception that viruses evolve toward benignity. This is usually phrased something like, “Natural selection favours viruses with low pathogenicity/virulence (so they don’t eradicate their hosts)“, or “Viral pathogenesis is an abnormal situation of no value to the virus“. This claim is clearly wrong — “clearly” both through common sense, and through observation.

Common sense tells us that pathogenicity does benefit at least some viruses. If a virus is spread through fecal/oral contamination (like noroviruses, say), then increasing fecal production will increase viral spread; we perceive that as diarrhea, fluid loss, illness and perhaps death — pathogenicity. In this case, the virus obviously benefits from virulence; there’s no pressure on the virus to evolve toward reduced pathogenicity, because that would reduce its transmission.

I talked about this with myxomavirus — one of the few well-studied cases where the concept can actually be tested. Myxoma was introduced to Australia in the early 1950s as a biological control agent for the rabbit plague. At first, the virus killed virtually every rabbit it infected (99.8% lethality), reducing the rabbit population by 85%, to a mere 100,000,000; but after some years of adaptation, most rabbits survived infection, and the rabbit population rebounded. However, the virus did not evolve to “low pathogenicity”; rather it evolved to a point where it killed about 70% of rabbits it infected. (To put that in context, it’s the same ballpark as smallpox, or Ebola — viruses which are not usually considered to have low virulence.) In this case, the reduced (but still very high) pathogenicity correlated with a longer period of transmission; the rabbits were sick for weeks, and less able to fend off the mosquitos and fleas that spread the virus. The most virulent virus killed the rabbits so fast that not many mosquitos got a chance to bite them. Hence, evolution to reduced, but still very severe, virulence.

But there’s another side of the equation. The host would, very obviously, benefit from reduced viral pathogenicity. This is something that’s hard to separate in natural experiments. When we see, for example, deer mice surviving with minimal disease while infected with Sin Nombre virus, how can we tell if the virus has adapted to the host, the host to the virus, or both?

In this natural myxomavirus experiment, we can in fact address that question, because Australian rabbits had to deal with the virus while their European cousins didn’t. And as you’d expect, Australian rabbits are actually relatively resistant to myxomavirus.

Below are data from Fenner¹ showing survival of wild Australian rabbits when exposed to the same strain of myxomavirus (click for a larger version). (Because the virus was deliberately introduced, it was known what the original strain was, and new strains with defined virulence in European rabbits could be isolated and saved.) The Y axis is the percent of rabbits responding with a certain level of disease; the X axis shows the number of epidemics that had spread through the country before sampling the rabbits.

In the first couple epidemics rabbits were almost all severely infected; but after selection by two or three epidemics, many of the rabbits are only moderately affected by the same strain of virus that killed almost all of their ancestors, and after six or seven epizootics, most rabbits survived and many were only mildly sick:

The belief that evolution is a slow process, undetectable by our mayfly eyes, is long obsolete. Here’s one example of how fast changes can become fixed in a population. As Fenner said:

it was then thought that it might be decades before resistance became evident, and that challenge with something less than fully virulent virus would make it easier to observe the first changes. To our surprise, within a few years, in a population of’ rabbits that had been exposed annually to myxomatosis, the case fatality rate fell from about 90% to about 50%, and eventually even lower than that.

But it’s not so simple. For example, why did it take a few epidemics before resistance arose in the rabbits?

Probably the reduced viral virulence played an important part in that. With a 99.8% mortality, there’s not much room for resistance to arise. After the virus settled down and spared 30% of its hosts, there was room for survivors to pass on their resistance genes. But there are limits:

… the rate of acquisition of resistance had reached a plateau after about six generations of selection (W. R. Sobey, personal communication, 1982). Perhaps this represents a limit to what can be achieved by selection of the pre-existing genotypes; any further change may require the occurrence and selection of mutations for resistance.

(That was written in 1983, and I don’t actually know if resistance has progressed further.)

We don’t know (or at least I don’t know; and I haven’t seen anything published on it) exactly what the resistance factors are in the resistant rabbits. However, it’s interesting that apparently a major cause of myxomavirus disease is “an acute and overwhelming immunopathological response to the virus“,² because this brings me back to the deer mouse/hantavirus story I mentioned earlier. It seems that mice respond to Sin Nombre virus with a regulatory T cell response — dampening immune responses rather than enhancing them — and this may be one reason they don’t show disease (i.e. they’ve removed the immunopathological side of the equation). I wonder if Australian rabbits have also evolved a tendency to develop TReg responses to myxoma virus.

In any case I can’t offer a better summary than Fenner’s:

Observations of its effects after some 30 years show that a fantastically virulent virus did not eradicate the rabbit, but that selection for transmissibility, especially during the off-season (winter in Australia) allowed the early emergence and rapid dominance of strains of lowered virulence. We now see a fascinating interplay between genetic changes in host and virus.

1. Fenner F. Biological control, as exemplified by smallpox eradication and myxomatosis. Proceeding of the Royal Society of London, Series B, The Florey Lecture, 1983: Biological Control, as Exemplified by Smallpox Eradication and Myxomatosis 1983;218:259-285.[↩]
2. Silvers L, Inglis B, Labudovic A, Janssens P, van Leeuwen B, et al. (April 25, 2006) Virulence and pathogenesis of the MSW and MSD strains of Californian myxoma virus in European rabbits with genetic resistance to myxomatosis compared to rabbits with no genetic resistance. Virology 348.:72-83 doi:10.1016/j.virol.2005.12.007[↩]