Evolutionary Anthropology 12:247–251 (2003)

CROTCHETS & QUIDDITIES

Dinner at Baby’s: Werewolves, Jaws, Hen’s Teeth, and Horse Toes

KENNETH WEISS AND SAMUEL SHOLTIS

Occasionally traits arise that appear to be atavistic throwbacks to the remote vism.” Our primate ancestors were past. How can this make evolutionary sense? fully furred, and when a variant allele or new arises that causes a person to be very hairy, he or she may As we get older we have a tendency tionary earlier states. In Descent of understandably be seen as ape-like. to become nostalgic and think back on Man, he argued that these reversions Such human atavisms, because we old times. We’ve recently seen a surge were “well worthy of attention.” Such tend to think of ourselves as advanced of nostalgia for the 1950s; among the traits are easier to explain as reflecting and gentile, are often popularly por- remarkable comebacks are the new historical connections than by cre- trayed as haunts of a brutish past, as old diners, like Baby’s here in State ationist arguments (for a popular in the “abominably hairy” atavism College (Figure 1). These new-old treatment, see1). Many such traits oc- Red Eye of Jack London’s book Before wonders proffer burgers and shakes cur naturally, but recently some sur- Adam who “was a monster in all like they used to be in the good old prising examples have arisen out of ways.” In our attention-hungry society days. The decades of change in the experiments in developmental biol- this has even been likened to were- competitive fast-food industry seem ogy. wolves.2 not to matter at all. The old taste is Another common reversion dis- back! Even Patsy Cline and Elvis are cussed by Darwin is supernumerary still singing the same songs in the NATURAL “ATAVISMS” nipples (polythelia). The “milk line” is background. The occasional presentation of ex- well known in mammals. It runs along There is a similar phenomenon in treme hairyness in humans is an ex- the thorax and abdomen on both sides . Nobody accepts Ernst Haeck- ample of a naturally occurring “ata- of the midline. Different species have el’s famous recapitulation argument that, as , we literally go through the adult stages of our ances- tors. Nonetheless, many seem to think the evolutionary past can rise again. Can it?

“A CIRCUMSTANCE WELL WORTHY OF ATTENTION” One of the key facts in Darwin’s “long argument” for were atavisms—“throwbacks”—to evolu-

Ken Weiss is a biological anthropologist, and Sam Sholtis is a Weiss graduate fel- low (no relation to the first author), both at Penn State University. E-mail: [email protected]

© 2003 Wiley-Liss, Inc. DOI 10.1002/evan.10125 Published online in Wiley InterScience (www.interscience.wiley.com). Figure 1. Dinner at Baby’s (State College, PA). 248 Weiss and Sholtis CROTCHETS & QUIDDITIES

Figure 2. Atavisms. (A) Werewolves, and other “atavisms.” Source: http://www.e- pub.org.br/cm/n09/fastfacts/atavis- mo_i.htm from “Evolution: Zufall oder Sinn?” Bild der Wissenschaft, 4: 114-126, 1979 (B) Horses’s toes (which are the horse, which Caesar?). Source: Modi- fied from Bettman Archive.

differing numbers of nipples located early neural tube.6 The epithelium dif- showed molecularly that the mesen- along this line, and this, as Darwin ferentiates into ameloblasts that pro- chyme in fact expresses chick, not noted, can vary naturally and be her- duce enamel, while the mesenchyme mouse, . There is still much to be itable. becomes odontoblasts that produce learned about the crosstalk between The evolution of digits has long dentine. When these two tissues come these two tissues required for tooth de- been of interest to anthropologists.3 in contact early in embryonic develop- velopment as suggested by the seem- Julius Caesar’s horse is said to have ment, the information for initiating ingly contradictory results of two recent had toes (as have the horses of Napo- dental patterning resides in the epi- papers. Chen et al.10 show that path- leon and Alexander the Great). thelial layer, which activates the mes- ways involved in tooth development are Whether seeing these thundering toes enchyme, and the two then interact conserved and can be turned on in in battle terrorized the barbarians we during tooth morphogenesis. chick mesenchyme by the addition of a can only speculate, but the fossil Kollar and Fisher7 wondered single signaling factor (Bmp4) to the record shows that horses haven’t had whether teeth could be induced in chick epithelium. However, the chick toes since Merychippus during the up- . They grafted chick embryonic epithelium is perfectly capable of initi- per Miocene (and they were already jaw epithelium to mouse dental mes- ating tooth-like formation when chick reduced and probably bore little or no enchyme, and found the development neural crest is replaced by neural crest weight).4 of what appeared to be teeth (Figure from a mouse.11 Another classic is the occasional ap- 4). This famous “hen’s tooth” experi- pearance of rudimentary hindlimbs in ment was questioned because it whales.5 Whale legs and horse toes are seemed implausible that chick epithe- Dinosaur Jaws related to Darwin’s notion of vestigial lium could be re-awakened to make It is exciting to perform an experi- traits, because vestiges of the horse’s teeth tens of millions of years after ment for one reason, and discover original toes are present in the form of chick ancestors had any choppers to something totally unexpected. You splint bones in the legs of all modern chop with. The tooth seems to contain scurry to the nearest textbook to find horses, and many whales retain bones enamel, but chicks appear not to have out what has occurred. If you were or cartilage related to hindlimbs hid- the required (amelogenin), and it working on a detailed biomedical den beneath their incredible mass. was molariform, but chicks don’t have problem, but discovered something of molars in their family tree (We as- evolutionary interest—like an atavis- EXPERIMENTAL ATAVISMS sume as an artifact that it looked more tic trait—it would make a quote-wor- like a human than a mouse molar.) thy thing to jazz up a paper. Hen’s Teeth The standard explanation was that This kind of surprise has resulted Teeth develop through a process of this was really a mouse tooth because from transgenic mouse experiments, tissue induction between two embry- the grafted mesenchyme was contami- and a particularly interesting case in- onic tissues in the developing jaw nated with mouse epithelium. But the volves engineered to inacti- (Figure 3B). The overlying dental lam- experiments have been replicated8,9 in vate a gene believed to be involved ina is an epithelial layer. Underneath many ways generating tooth-like struc- with jaw and tooth development. Two it is a mesenchyme largely composed tures, but without enamel. Wang et al.9 components made up the jaws of of neural crest cells that have physi- reversed the combination (mouse epi- mammalian ancestors, Meckel’s carti- cally migrated to the jaws from the thelium and chick mesenchyme) and lage in the lower jaw, and the palato- CROTCHETS & QUIDDITIES Dinner at Baby’s 249

even partially, to a previous stage al- ready realized in the ranks of its an- cestors.” Marshall et al.15 addressed this from a genetic point of view, showing by plausibility calculations that accumulating mutations are likely to destroy genes not maintained by selection by around 10 million years. Subtle coordination and inter- action among genes are probably the first to go, but eventually unused genes simply mutate into oblivion. Why can’t such a gene be restored by reverse mutation? With few excep- tions (olfactory receptors being one possible example), the probability that reverse mutation could restore the complex function of a gene is Figure 3. Hen’s tooth (A) original and (B) a confirmation showing the morphology and somewhere between impossible and corresponding different dental tissues. Mouse dental mesenchyme grown in culture in unbelievable. A gene long silent (as combination with chick “dental” epithelium. am, ameloblasts; en, enamel; de, dentine; od, between a chicken and it’s former odontoblasts. G and H are from the original panel notation. In G, e indicates area of teeth) cannot be resuscitated. Avian invaginating mouse epithelium; arrows indicate chick mesenchyme. (A: Reprinted with permission from Kollar and Fisher7) B. Wang et al.9 (Reprinted with permission, copyright 1998 amelogenin appears to be but a mem- Developmental Dynamics, Wiley.) ory. Sometimes a given complex trait does recur independently in related quadrate cartilage in the upper (Fig- structures to substantiate claims of lineages, as for example, complex eyes, or immature stages in various ure 4A). The palatoquadrate has been homology, much less the atavistic rec- amphibian or sea urchin larvae.16,17 severely reduced during mammalian reation of an ancestral organized trait. Traits of similar general form may be evolution with an anterior surviving The assignment of homology often rather simple to initiate but these re- remnant, the alisphenoid, that forms implies similar developmental origin currences are probably not identical part of the braincase, and a posterior of structures, but gene phylogeny and at the gene level. And because they surviving remnant that is the incus expression by David Stock, working in occur among evolutionary lineages, ossicle of our middle-ear (Figure 4B). our lab, showed that Dlx-2 existed and rather than within a lineage over time, The Dlx-2 DNA regulatory gene is ex- was expressed in embryonic jaws long we would not normally call them ata- pressed early in jaw development. before reptiles came on the scene. The vistic throwbacks to a former state. When Dlx-2 was experimentally inac- ancestral jaw developed with, not tivated in a mouse, the abnormally de- without that gene, so its experimental veloping jaws possessed anomalous deletion could hardly duplicate ances- What Probably Is cartilage12 (Figure 4C). The investiga- tral processes in dinosaur jaws. Smith At any given time, traits can come tors suggested that this cartilage was and Schneider offer the more plausi- and go across generations. Recessive homologous to the ancestral palato- ble explanation that cartilage forma- traits are familiar examples. But blue quadrate, as if ancient dinosaur jaw tion relies on a threshold level of cell eyes are not an interesting atavistic development had been recreated. condensation in development. Exper- throwback even if both of one’s par- Similar atavistic explanations have imental disruption of early jaw devel- ents had brown eyes but at least one been offered for anomalous cartilage opment could lead to the piling up of grandparent had blue ones. This is in experimental results inactivating migrating cranial neural crest cells in variation still circulating in the popu- the Hoxa2, MHox, Otx2, and retinoic uncharacteristic locations, leading to lation even if rare or masked in some acid receptor genes. anomalous local cartilage formation. individuals. Similarly with polydacty- In fact, this probably reflects misun- Consistent with this is that the exper- ly: it is only a kind of statistical typol- derstandings of the path of evolution imental had numerous other ogy to say that horses have hooves and a simplistic notion of atavisms.13 craniofacial developmental anoma- rather than toes. The cartilage formed in these mutants lies. Sewall Wright18 and others made bears no real resemblance to the an- formal breeding studies of existing cestral condition, and thus fails a key GENETIC EXPLANATIONS digit variation in guinea pigs early in criterion for identifying atavisms.5 the 20th century. Wright showed evi- What Probably Isn’t Comparing the results of the Dlx-2 dence for genetic causation, as well as knockout with the appearance of the A general principle of evolution is random effects among and within in- actual palatoquadrate cartilage shows expressed by Dollo’s14 famous rule bred strains. Interestingly, because too vague a resemblance of these that “an organism is unable to return, other South American Caviidae have 250 Weiss and Sholtis CROTCHETS & QUIDDITIES

Similar arguments apply to super- numerary nipples, variation in dental formulas including transitory dental rudiments in the diastema of develop- ing mouse jaws which have been de- scribed as representing “missing” teeth, and the occasional presence of “tails” in humans. Given their rather common natural occurrence, if these are atavistic throwbacks, they are not throwing very far back.

Then How Do Horses Get Toes and Hens Get Teeth? A lot has been written about these same traits, often in terms of develop- mental genetic “thresholds” and con- served developmental “programs,” but this amounts to little more than hand waving. However, we can make ge- netic sense of it. Most examples of atavisms are in traits produced by pe- riodic patterning processes, a point famously stressed by William Bate- son.19 These serially homologous structures are produced by generic de- velopmental processes that involve signaling-factor molecules that dif- fuse between cells in a particular de- veloping tissue. The relative local con- centration of these molecules affects gene expression in the cells, resulting in the production of epithelial pla- codes, tissue-layer invagination, branching, and periodically spaced growth and inhibition zones where members of a series (e.g., a tooth) de- velop.3,6 The same genes are used to pattern many structures, including the classic “atavism-prone” traits like limbs, mammary glands, teeth, vertebrae, and hair. The genes that make teeth in primates are still around biting and kicking in hens. A tooth is develop- Figure 4. Dinosaur jaw “atavisms” that aren’t really. (A) Primitive tetropod jaw, (B) normal mentally also a hair, feather, scale, or mouse and (C) Dlx-2 knockout “atavism.” (Black, upper first arch elements or palatoquad- nipple. For example, the close rela- rate cartilage; dark grey, lower first arch element, or Meckel’s cartilage; light grey, (ectopic tionship of teeth and hair is shown by cartilage.) (Source: Smith and Schneider13 with permission). the interesting effect of mis-express- ing the regulatory gene Lef-1 in mice, which leads to tooth development in thumbs, big and little toes, Wright re- polygenic causation, the same trait the lip furrow or hair development ferred to their occurrence in guinea arises in many different genotypes, among developing teeth. (Is that pigs as atavistic, despite the degree of implying that the same genetic basis woolly morning-after mouth a re- natural variation he observed. He of these toes in guinea pigs and other membrance of times past?) Natural as noted that the genetic basis of these Cavids past or present is unlikely. It’s well as experimental alteration of recurrences was probably different in similarly unlikely that the genetic ba- these developmental processes can different strains, and attributed this to sis of Caesar’s horse was the same as produce variation in the number, size, polygenic threshold causation. With the toes on ancestral horses. and other characteristics of the ele- CROTCHETS & QUIDDITIES Dinner at Baby’s 251 ments in the structures they pattern— independent evolution has occurred 1 Gould SJ. 1980. Hen’s teeth and horses toes. sometimes resulting in variation that in each genome for one to be a part of Natural History 89:24–28. resembles traits from distant earlier or function with the other. In this 2 Figuera LE, Pandolfo M, Dunne PW, Cantu JM, Patel PI. 1995. Mapping of the congenital gener- 20 ages. The genetic pathway that sense, unicorns might just be conceiv- alized hypertrichosis locus to chromosome could turn the splint bones in horses able, but Centaurs and mermaids, Xq24-q27.1. Nat Genet 10:202–207. into actual toes still exists in the sadly not. By contrast, at least in the 3 Chiu C-H, Hamrick M. 2002. Evolution and horse’s remaining digit. Many genes past, hens had teeth, and horses had development of the primate limb skeleton. Evol Anthropol 11:94–107. are involved in these processes, but toes, so hen-like and horse-like ge- 4 Carroll RL. 1988. Vertebrate Paleontology and they may actually be simpler than the nomes once were compatible with Evolution. New York: WH Freeman. polygenic control envisioned by teeth and toes. 5 Hall BK. 1984. Developmental mechanisms un- Wright. Still, from a gene-regulation, devel- derlying the formation of atavisms. Biological Hen’s teeth aren’t really teeth, be- opmental point of view, the verisimil- Reviews 59:89–124. cause some of the patterning instruc- itude of atavism does indicate a con- 6 Jernvall J, Jung HS. 2000. Genotype, pheno- type, and developmental biology of molar tooth tions and the enhancers needed to in- nection to the past, though it’s not an characters. Am J Phys Anthropol Suppl 31:171– tegrate the pathway in one of its awakening of long dormant develop- 190. contexts are gone. The chick epithelial mental programs. Recent develop- 7 Kollar EJ, Fisher C. 1980. Tooth induction in layer has lost these capabilities, and mental has been showing chick epithelium: expression of quiescent genes for enamel synthesis. Science 207:993–995. mouse mesenchyme can’t induce how deeply conserved developmental 8 Lemus D. 1995. Contributions of heterospecific them, but mouse epithelium has the processes really are. Developmentally, tissue recombinations to odontogenesis. Int J Start instructions and can invoke the hen’s teeth are truly tooth-like. But Dev Biol 39:291–297. responses in chick mesenchyme. It is they are not truly teeth. Evolution 9 Wang YH, Upholt WB, Sharpe PT, Kollar EJ, instructive that the missing ameloge- doesn’t generally reverse itself, but the Mina M. 1998. Odontogenic epithelium induces similar molecular responses in chick and mouse nin gene that is used only in the pro- past does seem to weave in and out of mandibular mesenchyme. Dev Dyn 213:386–397. duction of enamel was not protected complex traits, in that elements of the 10 Chen Y, Zhang Y, Jiang TX, Barlow AJ, St from obliteration by pleiotropy. mechanisms producing those traits do Amand TR, Hu Y, Heaney S, Francis-West P, In most cases, only some aspects of seem to be conserved in interesting Chuong CM, Maas R. 2000. Conservation of early odontogenic signaling pathways in Aves. Proc the process have been anomalously ways. The atavism story is subtle and Natl Acad SciUSA97:10044–10049. expressed, and atavistic changes are we’re now getting a genetic under- 11 Mitsiadis TA, Cheraud Y, Sharpe P, Fontaine- usually not completely normal. Thus standing of why we see what we see. Perus J. 2003. Development of teeth in chick em- the hen’s tooth has no enamel, and Thus, from what we know Dollo will bryos after mouse neural crest transplantations. Proc Natl Acad SciUSA100:6541–6545. supernumerary nipples usually are rule. Genetic engineers might produce 12 Qiu M, Bulfone A, Martinez S, Meneses JJ, only imperfectly formed and do not or chickens with teeth, but they won’t be Shimamura K, Pedersen RA, Rubenstein JL. cannot function normally in humans. real atavisms. It’s like that with retro 1995. Null mutation of Dlx-2 results in abnormal diners, too. They have a nostalgic ap- morphogenesis of proximal first and second branchial arch derivatives and abnormal differ- BUT THERE AREN’T ANY pearance and Patsy Cline may be sing- entiation in the forebrain. Genes Dev 9:2523– CENTAURS ing the same songs in the background, 2538. but most everything else in these din- 13 Smith KK, Schneider RA. 1998. Have gene We can imagine all sorts of assem- ers has changed, and we’re “So knockouts caused evolutionary reversals in the mammalian first arch? Bioessays 20:245–255. blages of characters. One example wrong,” if we think we can really be 14 Dollo L. 1893. Les lois de l’evolution. Bulletin might be the Centaurs of classical my- “Back in Baby’s arms.” de la Societe Belge de Geologie, de Paleontologie thology. These half-horse half-human et d’Hydrologie 7:164–166. creatures have not actually been seen 15 Marshall CR, Raff EC, Raff RA. 1994. Dollo’s recently, but we have evolutionary NOTES law and the death and resurrection of genes. Proc reasons to assert that they, like uni- Natl Acad SciUSA91:12283–12287. We would welcome comments on 16 Wray GA, Hahn MW, Abouheif E, Balhoff JP, corns and mermaids, aren’t real. It this column: [email protected]. Crotch- Pizer M, Rockman MV, Romano L. 2003. The would be impossible, based on every- etyComments are maintained on: Evolution of Transcriptional Regulation in Eu- karyotes. Mol Biol Evol. In press. thing we know, for the genetic and http://www.anthro.psu.edu/weiss_lab/ developmental organization required 17 Weiss KM. 2002. How the eye got its brain. index.html. We thank Anne Buchanan, Evol Anthropol 11:215–219. to make a human torso also to make a Kathleen Smith, and John Fleagle for 18 Wright S. 1968. Evolution and the genetics of horse’s aft. The forelimb and hindlimb critically reading this manuscript. populations: a treatise. Chicago: University of of mammals develop using at least Chicago Press. some of the same genes, and there is a 19 Weiss KM. 2002. Good vibrations: the silent genetically based correlation in the REFERENCES symphony of life. Evol Anthropol 11:176–182. form of the two limbs. The divergence 20 Salazar-Ciudad I, Jernvall J. 2002. A gene net- work model accounting for development and between primate and horse genomes Many things discussed here can be evolution of mammalian teeth. Proc Natl Acad happened so long ago that too much profitably explored by web searching. SciUSA99:8116–8120.