The Marek Effect

Marek’s disease is caused by an alphaherpesvirus that infects poultry. Birds shed the virus (MDV) from the skin cells of their feather follicles, and other birds contract it by inhaling these cells. MDV can remain viable in poultry-farm dust for very long periods of time. It attacks the chicken’s memory T-cells, which go on to form visceral tumours and infiltrate nerves in the wings and legs, causing paralysis.

MDV has been around for a long time, but it did not become a serious problem until the 1950s, with the rise of industrialised agriculture, as chickens were raised increasingly in crowded conditions favourable to virus transmission. MDV was first identified in 1968, and since the 1970s the poultry industry has protected its chickens with attenuated virus vaccines. As everyone already knows, these vaccines leak, which is to say that they protect chickens from the most serious symptoms of Marek’s disease but do not stop MDV replication or transmission. Vaccinated chickens carry and shed MDV for life. Since the introduction of these vaccines, MDV has become vastly more pathogenic for unvaccinated chickens. Today Marek's disease has a nearly 100% infection fatality rate for the unvaccinated, killing all infected chicks within a few weeks of infection. Almost all newly hatched chicks are susceptible after their maternal antibodies wane. The older, milder MDV strains described in the early 20th century have totally disappeared.

The thesis that vaccination has driven MDV towards greater pathogenicity has been around for a while. The most recent argument to this effect is from Andrew F. Read. You can read his 2015 paper in PLOS Biology yourself, or you can just study this figure for a few moments.

These graphs illustrate the course of infection caused by five different MDV strains in 8-day old chickens. Solid lines and darker shading mark immunised chickens; dotted lines and lighter shading denote immunologically naive chickens. The three graphs in the left column track the least pathogenic strain studied, HPRS-B14. Towards the right the strains increase in pathogenicity, with Md5 and 675A representing the state-of-the-art in chicken virus apocalypse.

The bottom row, “Cumulative viral genomes shed from an infection”, tells the whole story. At the left, you can see that the MDV vaxx disadvantages the least virulent strain. Vaxxed chickens (black) shed less virus, unvaxxed chickens (grey) shed more. Dial the virulence up a notch, to strain 571, and suddenly this situation is reversed and the vaxxed chickens shed more. This is because their life expectancy extends beyond seven days and they can host the virus for a much longer period of time. The rare unvaxxed chickens who make it through their strain-571 infections in fact shed somewhat more virus (second row) but there’s fewer of them so it doesn’t matter. As you move to the right, this effect is only more pronounced. In a world where almost every last chicken is vaxxed, the worst strains have the greatest advantage.

Now, leaky vaxx clearly favours pathogenic strains of MDV in chickens, but we don’t know what drove the evolution of these strains in the first place. Marek’s disease was getting worse for chickens long before anybody invented a vaccine for MDV. It’s thought that the sharply reduced lifespans of factory-farmed chickens removed pressure for the virus to keep its host alive very long. Modern chickens are also highly uniform genetically, and they live in extremely unnatural, dense environments.

And this is true as well: Viruses only want to replicate. They might be fine with chickens dying paralysed and riddled with visceral tumours, but that’s not their goal. They aim only to make more copies of themselves. Pathogenicity is caused by many things. A greater rate of virus replication – the only thing the virus really wants – is just a part of it. The manner in which the virus interacts with host cells is another part of it. A leaky vaccine might therefore drive adaptation in more than one direction, encouraging higher rates of virus replication that before would have been selected against; while otherwise reducing selective pressure against strains that do more cellular damage.

Whatever the mechanisms, Marek’s Effect isn’t a universal law of immunology, or anything like that. There are a few other cases where it looks like leaky animal vaxx have caused greater virus pathogenicity in animals, but there are plenty of others where nothing like this has been observed. Marek’s Effect also took a long, long time to make MDV the extreme virus it has become – over ten years of widespread chicken vaxx, to be specific. There are probably three times more chickens than humans, and MDV has near universal prevalence across the entire poultry population. There must be many, many more instances of MDV replication than SARS-2 can hope for in the foreseeable future, and humans will never have the vaxx uptake rates that chickens do.

All of this is to say, that it’s not obvious to what degree Marek’s Effect will punish us for our vaccination follies. At the same time, it seems increasingly plausible that mass vaccination against SARS-2 and the rise of the Delta variant are related. As I wrote in my Corona Primer, Corona front-loads its attack, achieving high transmissibility fairly early, around the time of symptom onset. Delta represents a slightly more extreme version of this same strategy. It’s a balance, though: Too much virus replication too early, the infected person gets sick, stays home, and the virus gets nowhere. But the vaccines reduce symptoms, and they reduce Corona anxiety as well. The more vaccinated people there are, the less Delta is disadvantaged for pushing too hard too early.

Early, aggressive replication would also be a way getting the jump on immune response. Together, then, these would give us the twofold evolutionary mechanism that Read (at the link above) thought had selected ever more pathogenic MDV strains: “Within-host selection favouring virulent variants for their ability to evade immunity and vaccine-induced relaxation of between-host selection against virulence”.

If the Corona vaccines do end up causing more pathogenic (and not just more transmissible) strains, you will never hear about it. Marek’s Effect is limited mostly to poultry, but the whole problem has nevertheless inspired nervousness in Camp Vaccine, where everybody is worried about providing ammunition to their nemeses, the antivaxxers. But there’s another problem too: Despite everything you’ve read about how the variants cause worse disease, the evidence for this is pretty terrible, in no small part because nobody will induce experimental infections in human subjects to study the relative pathogenicity of different lineages. Future Marek’s-Effect strains would be concealed from us by the same ethical concerns. It would seem to us only that the virus was getting worse.