Human Evolution: A Legacy ofCannibalism in our Genes?

Dispatch

John F.Y. Brookfield

A new study of genetic variation in the human prionprotein gene suggests that balancing selection hasoperated on an amino acid sequence polymorphismin the gene during the last five hundred thousandyears. Is this a legacy of widespread cannibalism byour ancestors?

The patterns of DNA sequence variation in humanpopulations are the result of migration, selection andrandom genetic drift operating, as a function of theeffective population size, on human populations overthe last million years. Given this, the possibility existsof using the structure of the human variability to infersomething about human migration, selection andeffective population sizes in the past. A new study ofDNA sequence variation at the prion protein locus [1]has been interpreted as suggesting a sustainedheterozygote advantage created by a lifestyle of habit-ual cannibalism, implying a new vision of the lifestylesof our ancestors.

One of the most challenging questions about humansis to know the extent to which natural selection operat-ing in our prehistoric ancestry has shaped our presentphenotypes and behaviours. Evolutionary psychologytries to explain human behaviour in terms of the selec-tive advantage conferred by individual behaviouraltraits. This approach falls into the biological tradition oftrying to explain aspects of an organism’s phenotypethrough the fitness advantage that the phenotype gen-erates. With humans, however, the explanation ofbehaviour in adaptive terms is more difficult than inother species, as the societies in which we live todayare immeasurably different from those in which ourgene pool evolved. While we may be adept at hang-gliding, the genes that make us so were unquestionablynot shaped by selection against individuals who hang-glided unsuccessfully in our evolutionary past.

For this reason, selective explanations of humanbehaviour cannot be judged on the basis of the con-sequences of the behaviour for modern humans, butonly in terms of whether the genes responsible wouldhave been favoured by selection during the prehistoriceras when our genome evolved. For this reason, manyhave speculated about the ‘environment of evolution-ary adaptedness’ [2], the hypothetical pleistoceneenvironment to which our genes were adapted byselection. While some insights into our lifestyle duringthe time of the environment of evolutionary adapted-ness can be gleaned from studies of hunter-gatherersliving today, and from the archaeological record, thiscan never be known with certainty. This is obviously a

problem for the adaptive interpretation of humanbehaviour, as it is very easy to hypothesize selectionhaving operated on any arbitrary behaviour by makingappropriate guesses about the nature of the environ-ment of evolutionary adaptedness.

There is, however, a completely different, and com-plementary, approach to discovering the lifestyle of ourancestors — one which starts with the identification ofselection, and moves from this to inference aboutlifestyle. The idea is to identify genetic variability withknown fitness consequences in present-day pop-ulations, and to use the patterns of variability to infersomething about selection in the past. Then inferencescan be drawn about the lifestyle that would have beennecessary for such selection to have operated.

A remarkable new study provides an example ofthis approach [1]. The tragic occurrence of new variantCreutzfeld-Jakob Disease (vCJD) in the UK has led tostudies of the genotype of the affected individuals atthe prion protein locus. There is a polymorphism foralleles of this locus that encode either methionine orvaline at codon 129 (M129V) of the polypeptideproduct. Remarkably, all vCJD-affected individuals todate have been homozygotes for the methionine-encoding allele. There is also evidence that sporadicand iatrogenic cases of CJD have stronger and morefrequent effects on each of the homozygotes for thispolymorphism. The most powerful data set concernsthe Fore tribe of the Eastern highlands of Papua NewGuinea. Here the systematic consumption of diseasedrelatives at mortuary feasts spread kuru, the spongi-form encephalopathy associated with the conversionof the cellular form of the prion protein, PrPc, into itsscrapie form PrPSc. Mead et al. [1] screened Forewomen over fifty years of age, many of whose cohortwould have died of kuru. Among these women, thereis very significant heterozygote excess for the M129Vpolymorphism, implying a heterozygote resistance tothe disease.

The inference is therefore that the kuru disease iscreating heterozygote advantage at this locus.Heterozygote advantage, of course, is one of the fewways in which a genetic polymorphism can be stablymaintained. A process balancing a polymorphism inthis way that operated for a long period — tens ofthousands of generations or more — would beexpected to start to create a ‘signature’ of its effects inthe structure of the genetic variability. Specifically, onewould start to see more and more DNA sequence vari-ants with intermediate frequencies in the population —more than would be expected from a neutral processinvolving only mutation and genetic drift. This structurewould be detected by positive values of the statisticknown as Tajima’s D [3]. Furthermore, one would seevariants at this and flanking loci being associated intodifferent haplotypes, showing many base polymor-phisms in linkage disequilibrium with the selected poly-morphic site. In the new study of the structure of the

Current Biology, Vol. 13, R592–R593, August 5, 2003, ©2003 Elsevier Science Ltd. All rights reserved. DOI 10.1016/S0960-9822(03)00522-0

Institute of Genetics, University of Nottingham, QueensMedical Centre, Nottingham NG7 2UH, UK.

prion protein variation [1], in both European andAfrican populations in addition to the Fore themselves,both of these predictions were confirmed.

The implication is that the heterozygote advantageidentified in individuals exposed to kuru has been aconstant feature of our evolutionary past, which itselfimplies that acquired prion diseases were themselvessufficiently common for such selection to be effective.Mead et al. [1] suggest that, while the regularacquisition of prion diseases through eating animals isa possibility, it is also possible that this polymorphismis a legacy of human cannibalism, as indicated bysome fossil evidence [4,5].

Clearly, the situation is not entirely clear-cut.Heterozygote advantage is maintained by selectionagainst both homozygotes, and the equilibriumfrequencies of two alleles depend on the relativestrength of selection against the two homozygotes —the stronger the selection, the lower an allele’sfrequency. Thus, it is surprising that it is homozygotesfor the commoner methionine 129 allele who appear tobe more sensitive to vCJD in the UK data.

More generally, this study [1] shares with the recentdemonstration of an apparent selective sweep at theFOXP2 locus [6] the property that part of the evidencefor selection comes from the frequency of variantalleles in the population. In studies like this [7], theidentification of selection comes from a comparison offrequencies of variant alleles with predictions of thestandard neutral model — which makes manyassumptions, a critical one being that there is nopopulation subdivision. With population subdivision,predictions should be modified, but it is hard to knowhow to do this without independent evidence of themigrations and gene flow in our ancestry.

Furthermore, with population substructure, samplingstrategy becomes very important. Imagine a populationthat consists of twenty fairly reproductively isolatedsubpopulations created simultaneously at a time morethan N generations ago, where N is the effective popu-lation size. Now, suppose that we sampled 20 alleles ofour favourite gene from this population. Our inferencesabout selection will almost certainly depend on oursampling strategy. If our sample is of one allele fromeach of the twenty populations, almost all variants willbe seen only once in the sample, and we will concludethat we are seeing the ‘signature’ of a selective sweepcreated by an adaptive substitution. If, however, wechoose to sample ten alleles from each of two popula-tions, most variants will be at intermediate frequencies,and we will identify a history of balancing selection.

References1. Mead, S., Stumpf, M. P. H., Whitfield, J., Beck, J.A., Poulter, M.,

Campbell, T., Uphill, J.B., Goldstein, D., Alpers, M., Fisher, E.M.C.,and Collinge, J. (2003). Balancing selection at the prion protein geneconsistent with prehistoric kurulike epidemics. Science 300, 640-643.

2. Irons, W. (2003). Adaptively relevant environments versus the envi-ronment of evolutionary adaptedness. Evolutionary Anthropology421, 63-66.

3. Tajima, F. (1989). Statistical method of testing the neutral mutationhypothesis by DNA polymorphism. Genetics 123, 585-595.

4. Defleur, A., White, T., Valensi, P., Slimak, L., and Crégut-Bonnoure,E. (1999). Neanderthal cannibalism at Moula-Guercy, France.Science 286, 128-131.

5. Marlar, R.A., Banks, L.L., Billman, B.R., Lambert, P.M., and Marlar,J.E. (2000). Biochemical evidence of cannibalism at a prehistoricPuebloan site in southwestern Colorado. Nature 407, 74-80.

6. Enard, W., Przeworski, M., Fisher, S.E., Lai, C.S.L., Wiebe, V.,Kitano, T., Monaco, A.P., and Pääbo, S. (2002). Molecular evolutionof FOXP2, a gene involved in speech and language. Nature 418,869-872

7. Bamshad, M., and Wooding, S.P. (2003). Signatures of selection inthe human genome. Nature Rev. Genet. 4, 99-111.

Current BiologyR593