Iron lung
Polio patient in iron lung (1949)

It’s well known that HIV mutates rapidly and continuously in infected people. An individual is infected with a handful of HIV viruses, but quickly becomes the host of a vast cloud of virus genomes, with the dominant strain of HIV evolving over time.

There are several factors selecting which HIV sequences are dominant at any one time. The most interesting force selecting mutations, at least from my view, is the immune system — this has been shown in many papers; one of the most recent comes from Bruce Walker’s empire1 in a paper I may talk more about some time — but other factors, often lumped together as “viral fitness”, are also important. A successful immune response to HIV is probably one that forces the virus (in its attempts to avoid the immune system) to lose overall fitness.

HIV isn’t unique in this, of course, although it’s probably the clearest example. Other persistent viruses very likely do the same thing, with mutations and selection driving the virus toward a more “fit” phenotype. An unusual example was just described in the Journal of Virology.2

Poliovirus and vaccination

Wild polio vs vaccine polio, JAMA 2004The oral poliovirus vaccine is remarkably safe; even though the vaccine is live virus, the vaccine itself only causes paralysis in 1 out of 3 million vaccinees. 3 (See the chart to the right4 for rates of wild polio vs. rates of vaccine polio — note that the inset, showing total cases of wild polio, has a log scale.  Click for a larger version.) The reason that’s remarkable is that the virus — both the vaccine virus, and the wild-type virus — has very high mutation rates, even higher than HIV’s. So you might wonder why the vaccine strain doesn’t rapidly mutate back to a virulent form as soon as it gets into the host and starts replicating — that is, why doesn’t the virus mutate toward a more fit form, as happens with HIV.

A big part of the answer, in the case of polio, is the immune system. Unlike HIV, the immune system normally rapidly controls poliovirus replication; that means that the virus only has a narrow window between infection and clearance in which to find a virulent sequence.

In that case, what if the immune system doesn’t work properly? If the virus isn’t eliminated by the immune system, it should have more chances to find a virulent variant. In fact, as you’d expect, immune deficient people are at much higher risk of vaccine-associated polio disease — thousands of times higher; the risk is something like 3000 cases of disease per million vaccinees.

Mutation of the vaccine strain

Polio quarantine (NLM)This has all been known for quite a while. The new paper2 looked in detail at the evolution of vaccine-strain polio in an immune deficient (hypogammaglobulinemic) patient over a couple of years. (This was actually a re-analysis of an old case; the patient was infected in 1963, as part of a study that I’m pretty sure would never pass an ethics committee today. Normally, known immune deficient patients are not, of course,  given oral polio vaccine, because of the obvious risk. Although it does happen that people are vaccinated and later found to be immune deficient, in this study, an immune deficient patient was deliberately infected, along with a treatment that was hypothesized to protect him. That meant stool samples, with excreted poliovirus, were saved from the initial stages, and we are now at the stage where it’s technically feasible to sequence large chunks, and in some cases whole genomes, of hundreds of the excreted viruses.)

There were lots of interesting findings in the evolution that I’m going to ignore here. The bottom line is that the vaccine virus mutated fast (six changes within the first 21 days) and continuously, and before long reverted to wild-type characteristics; ultimately the patient was excreting polioviruses that were “essentially wild-type viruses in their measurable properties“.

That means that it’s possible for vaccinated immune-deficient patients to act as functional reservoirs of wild-type poliovirus, for a long time. The child in this study died in 1966, but there’s a case of an immune deficient person who has excreted vaccine-derived virus for over 20 years.5 That has obvious implications for eliminating polio from a country. I think there are also implications for HIV vaccination — a vaccine that forces the virus into an unfit state, which is a goal of some vaccine strategies, will produce a vast reservoir of virus that can rapidly switch back into a wild type, virulent form in unvaccinated people. That would probably be better than we have today, but it’s something to think about.


  1. Marked Epitope- and Allele-Specific Differences in Rates of Mutation in Human Immunodeficiency Type 1 (HIV-1) Gag, Pol, and Nef Cytotoxic T-Lymphocyte Epitopes in Acute/Early HIV-1 Infection. Brumme et al., Journal of Virology, September 2008, p. 9216-9227, Vol. 82, No. 18 doi:10.1128/JVI.01041-08. []
  2. J. K. Odoom, Z. Yunus, G. Dunn, P. D. Minor, J. Martin (2008). Changes in Population Dynamics during Long-Term Evolution of Sabin Type 1 Poliovirus in an Immunodeficient Patient Journal of Virology, 82 (18), 9179-9190 DOI: 10.1128/JVI.00468-08[][]
  3. Of course, even that rate is higher than anyone would like; as the rates of authentic polio have dropped to nearly zero, the relative risk of the vaccine have increased; and so the current recommendation is for the killed polio vaccine, which is even safer.[]
  4. Vaccine Policy Changes and Epidemiology of Poliomyelitis in the United States. Alexander et al., JAMA. 2004;292:1696-1701.[]
  5. MacLennan, C., G. Dunn, A. P. Huissoon, D. S. Kumararatne, J. Martin, P. O’Leary, R. A. Thompson, H. Osman, P. Wood, P. Minor, D. J. Wood, and D. Pillay. 2004. Failure to clear persistent vaccine-derived neurovirulent poliovirus infection in an immunodeficient man. Lancet 363:1509-1513[]