Cannibalism & evolution

Cannibalism is a controversial topic. It is routine for particular societies to accuse "barbarians", enemies, or evil mythological figures, of cannibalism. When it comes to the archaeological record some skeptics have claimed that like "sacred objects" too often human remains found in peculiar circumstances are ascribed to human sacrifice or cannibalism. In Did Adam and Eve Have Navels? Martin Gardner lays out the skeptical case for why cannibalism is rare to non-existent, and rather something which emerges from the imaginations of ethnographers and archaeologists, or is rooted in scurrilous insults hurled between ethnic groups which have antagonistic relationships. Since the arguments Gardner lays out were presented it seems that the skeptical case is looking weaker, though controversy remains in specific instances. In the domain of genetics, there is some evidence of natural selection on genetic loci which imply widespread prion diseases in the past. Diseases which are often the outcomes of cannibalism. These sorts of molecular genetic data should perhaps change our perspective as to the imaginative color which archaeologists and paleoanthropologists might add to their inferences from ancient or prehistoric human remains.

But the case of cannibalism among the Fore people of Papua New Guinea it is not speculation or a matter of historical or archaeological inference. They were engaging in the practice as late as the 1960s. So it is of interest that a new paper has come out reinforcing the finding that the kuru epidemics might have left a genetic imprint, A Novel Protective Prion Protein Variant that Colocalizes with Kuru Exposure:

Results Persons who were exposed to kuru and survived the epidemic in Papua New Guinea are predominantly heterozygotes at the known resistance factor at codon 129 of the prion protein gene (PRNP). We now report a novel PRNP variant -- G127V -- that was found exclusively in people who lived in the region in which kuru was prevalent and that was present in half of the otherwise susceptible women from the region of highest exposure who were homozygous for methionine at PRNP codon 129. Although this allele is common in the area with the highest incidence of kuru, it is not found in patients with kuru and in unexposed population groups worldwide. Genealogic analysis reveals a significantly lower incidence of kuru in pedigrees that harbor the protective allele than in geographically matched control families.

Conclusions The 127V polymorphism is an acquired prion disease resistance factor selected during the kuru epidemic, rather than a pathogenic mutation that could have triggered the kuru epidemic. Variants at codons 127 and 129 of PRNP demonstrate the population genetic response to an epidemic of prion disease and represent a powerful episode of recent selection in humans.

Kuru is a prion caused illness, related to Creutzfeldt-Jakob disease (a.k.a. "mad cow disease", though that is somewhat of a misnomer). It is a neurodegenerative ailment which can take up to five decades to incubate. The oral history of the Fore suggests that the first cases occurred around 1900. The disease spread mostly to women and young children because of the practice of mortuary feasts where the maternal kin of the deceased would dismember and consume the remains. The Fore only eat healthy people who die. Additionally, males above the age of seven were not participants in the feasts. Because the feasts ceased after ~1960, the kuru epidemic abated (though incubation means that individuals succumbed long after the ending of the feasts).

These researchers had previously found a region of the prion protein gene (PRNP) which conferred protection against kuru at codon 129. Those who were homozygotes were susceptible, while those who were heterozygotes were not. The molecular genetic rationale seems to be that homozygous state allows for homologous
protein-protein interactions which foster the progression of the disease. When the locus is a heterozygote the facilitative interactions do not occur, and that dampens manifestation of kuru.

In this study the researchers found a second region of PRNP, at codon 127, which also seems to confer resistance. The map below shows the the concentration of the 127V ("V" for valine) allele in areas where kuru was endemic and at high frequency:

Even in regions where kuru was common the frequency of the 127V allele is extant at only 8%, and no homyozygotes were found in their sample. Previous work had shown that the allele at 129 was out of Hardy-Weinberg equilibrium. HWE is defined by:

1 = p² + 2pq + q²

In other words, the proportion of homozygotes and heterozygotes can be predicted by the formalism above. 129 is out of HWE because heterozygotes are in excess, suggesting positive (balancing) selection for that state. Since the frequency of the 127V allele was so low they didn't have the statistical power to explore this particular issue (127VV did not exist, but the N was too small to judge whether it was significant since its expectation should be less than 1% anyhow). It is important to note that they found that deviation was significant in particular for older women, precisely the group which would have been subject to selection against homozygosity. They did not find it for men who, likely avoided the peak years of the epidemic because they were not party to the mortuary feasts.

It was also found that codons 129 and 127 were not at equilibrium with each other. In particular, heterozygosity at 129 was negatively correlated with heterozygosity with 127 within individuals. The hypothesis which naturally presents itself then is that the mutation on 127 rose in frequency in the background of the population segment which was vulnerable to kuru. A population vulnerable to kuru will always exist in the previously known gentic architecture because selection for heterozygotes (at 129) is going to result in the production of homozygotes naturally out of the process of Mendelian segregation (heterozygote parents are likely to produce some homozygote children). The fitness increase conferred upon heterozygotes in kuru afflicted areas was very high, vulnerable homozygotes had relative fitness in the range 0.2-0.75 depending on how pervasive the disease was in their region (areas hard hit by kuru were in the low range, areas less afflicted in the high range). This is a huge selective pressure, so it would not be surprising if the 127 codon allele arose in the genetic background due to great reduced population mean fitness. The presence of 129 homozygotes is naturally guaranteed to open up such an opportunity. As noted above, in the sample those on codon 129 who were homozygote were particularly likely to carry 127V.

127V is extremely efficacious against fatality due to kuru. Looking through the pedigrees in a region of very high kuru exposure the researchers found that of the individuals who carried 127V, only 1 out of 36 in the parent generation died of kuru. By contrast, 33 of 218 parents from those carrying 127G only (the modal allele) had died of kuru (some of these presumably would also have carried the protective variant of 129). When the researchers looked at the 127V haplotype, the nature of the variation around this mutation implied that a common ancestor existed ~10 generations ago, with a 95% confidence interval 7 to 15 generations. That means that all of the copies of 127V extant today in the Fore descend from one particular copy present on the order of 250 years in the past within the population.

The authors conclude:

Our new data thus provide evidence of a complex selection event in the Fore population at PRNP during the kuru epidemic, with balancing selection acting to maximize heterozygosity at codon 129, coincident with positive selection acting to increase 127V alleles on a 129M background. Whether putative 127VV homozygotes would have had higher susceptibility to kuru than 127GV heterozygotes, in an analogous way to the situation at codon 129, remains a matter of speculation. The relative viability of combined codon 127-129 genotypes of PRNP in elderly persons from the kuru region suggests that there was stronger selection, in this specific situation, than that in the classically quoted examples of endemic malaria and hemoglobin S or C alleles.

Malaria is a classic example of selection in human beings. In the 1970s Richard Lewontin was apt to assert that it was the example of heterozygote advantage in human beings. There are some theoretical reasons to be skeptical of pervasive heterozygote advantage. The intuitive reason is clear, imagine a population which is subject to heterozygote advantage at locus after locus; the likelihood that any given individual is going to be all heterozygote at locus after locus is very low. The predicted range of fitness is simply not plausible. If powerful heterozygote advantage occurs, it can't be spread out across too many genes, because the likelihood of homozygosity emerging out of Mendelian reproduction is not trivial, and homozygosity on some loci is then inevitable.

One can illustrate this rather simply. Imagine there are three genes which are the targets of selection so that heterozyogsity is favored, A, B and C. Differentiate the variants by case, so:

AA = homozygote
Aa = heterozyogte
aa = homozyogote

Assume that homozygotes have equal distance from heterozygotes in fitness, so at equilibrium the heterozygote genotypes should be extant at frequencies of 0.50. The allelic frequencies should be balanced. So HWE:

(0.50)² + 2(0.50)(0.50) + (0.50)² = 1

The maximal fitness for an individual would be:

AbBbCc

The minimal would be:

aabbcc

If for all the genes 50% of the genotypes are heterozygotes at birth, then the chance that any random individual is going to have the optimal gentoype is 12.5% (0.50³). If all of the homozygotes are extremely detrimental to fitness, as is the case with malaria or kuru, then we're in a world of hurt.

Heterozygote advantage on one locus seems likely to be a short-term fix. Over time one might imagine a host of genetic modifiers emerging so as to slowly dampen the drag on homozygotes; 127V was likely favored in part because of the drag on homozygotes of 129. If kuru emerged only recently we saw the first sweep of selection, but not its later equilibrium. But I wonder as to what kuru itself, and the practice of cannibalism, tells us about the power of selection. Over the past few weeks I've mentioned between vs. within group dynamics in natural selection, and it seems that from a functional perspective cannibalism among the Fore was extremely detrimental, dragging down the population mean fitness. Apparently like most people the Fore had little understanding of the causes of disease, and attributed kuru to some sort of curse. Neighboring populations were not subject to kuru. Why exactly were not the Fore conquered and their deleterious cultural practices eliminated? Perhaps if "nature" had been allowed to work they would have been. On the other hand, perhaps we're seeing here the long time frame which selection upon groups has to work, and the power of norms within groups to generate "irrational herds." What would have happened to women who refused to engage in cannibalism? On the one hand they would not have suffered from kuru, but perhaps they would have been ostracized from the community and their fitness reduced. The Fore were developing genetic adaptations which mitigated the fitness drag from from kuru, and they were not being conquered by their neighbors. The answer for why this process was allowed to occur, that is, allowing slow genetic solutions to emerge when a quick cultural fix was available, might be explained by the geography of the highlands of Papua New Guinea, which encourage and foster isolation and fragmentation. It makes me think about the Shifting Balance. Though in this case the Fore wandered into a fitness valley of cultural maladaptation, instead of a genetic one.

Citation: Mead, Simon, Whitfield, Jerome, Poulter, Mark, Shah, Paresh, Uphill, James, Campbell, Tracy, Al-Dujaily, Huda, Hummerich, Holger, Beck, Jon, Mein, Charles A., Verzilli, Claudio, Whittaker, John, Alpers, Michael P., Collinge, John. A Novel Protective Prion Protein Variant that Colocalizes with Kuru Exposure. New England Journal of Medicine, 2009; 361 (21): 2056 DOI: 10.1056/NEJMoa0809716