Work

The neural crest cell hypothesis: no unified explanation for domestication

Year: 2021
Type: article
Abstract: A few years ago, Wilkins et al. (2014) advanced the neural crest cell hypothesis of domestication as an explanation for the “domestication syndrome” in animals, which refers to a collection of characteristic traits associated with domestication. Since then, the neural crest cell hypothesis has risen to become arguably the most popular explanation for the genetics of the initial phase of the domestication process (Zeder 2017; Thomas and Kirby 2018; Kikusui et al. 2019; Zanella et al. 2019). Neural crest cells contribute to many different parts of the vertebrate body. They are the origin of most of the peripheral nervous system including parasympathetic and sympathetic ganglia, the enteric nervous system and sensory ganglia; sensory receptor cells in the skin; pigment cells, except in the retina; endocrine cells in the adrenal medulla and thyroid; the frontal part of the head, including bone, connective tissues, and muscle; and parts of the cardiovascular system (reviewed by Dupin et al. 2006). The neural crest cell hypothesis was motivated by a search for common factors between tameness and other domestication syndrome traits, and an apparent similarity between domestic animals and humans with neurocristopathies, posited by Wilkins et al. (2014). They hypothesized that if tameness is mainly caused by changes in the adrenal glands, selection on tameness might change neural crest development to cause correlated changes such as brain size reduction, floppy ears, altered facial shape, and altered pigmentation. The hypothesis entails that a reduction in neural crest cell proliferation and migration is a core genetic mechanism of early domestication. This implies that the genetic architecture of domestication is by necessity pleiotropic, with variants in genes affecting neural crest development and distribution causing multiple domestication traits. The authors also suggest that neural crest cell function is polygenic, affected by many genetic variants with individually small effects. Furthermore, they have suggested that it involves epigenetic inheritance, genetic assimilation, or epistatic interaction between genes where changes to the neural crest may potentiate later evolution of other domestication traits (Wilkins 2020). Because pleiotropy is at the core of the neural crest cell hypothesis of domestication, we need to clarify what we mean by this term. Paaby and Rockman (2013) distinguish between when genes have multiple functions, for example in different pathways, reactions or protein–protein interactions, and when alleles have effects on multiple traits, for example, by causing a heritable syndrome. As this distinction pertains to the neural crest cell hypothesis, many genes involved in the neural crest are pleiotropic in the sense that they participate in multiple developmental processes. However, for the neural crest cell hypothesis to work, there also have to be pleiotropic alleles, selected during domestication, that affect multiple domestication traits through their effect on the neural crest. The development of neural crest cells is regulated by a complex gene regulatory network that spans multiple cell types and temporal phases (reviewed by Betancur et al. 2010; Rogers et al. 2012). Neural crest cells originate at the border of the neural plate. After induction, they undergo an epithelial-to-mesenchymal transition (where they lose their epithelial properties), delaminate (split off) from the neural tube, and migrate to their target locations where they differentiate. Neural crest induction happens in response to secreted signaling molecules [including Bone morphogenetic proteins (BMPs), Fibroblast growth factors (FGFs), and Wnt ligands] secreted by surrounding tissues, and Notch signaling, which is based on cell–cell contact. These signals lead to the expression of neural plate border specifiers (homeobox genes including Msx1 and Pax3 and zinc finger Zic genes), a group of transcription factors that trigger neural crest specifiers. Neural crest specifiers are the key transcription factors expressed in neural crest cells themselves before and during migration (including Snai1/2, Id, cMyc, Sox9, and Sox10). Furthermore, there is a set of genes (including FoxD3, Snai1/2, and Sox5) that regulate how neural crest cells delaminate and go through epithelial-to-mesenchymal transition, needed for them to migrate. This process involves changes to the extracellular matrix, and regulation of cell adhesion molecules including cadherins and integrins. When arriving at their destination, neural crest cells typically stop expressing neural crest specifiers, but some of them keep being expressed and serve to regulate differentiation (e.g., Sox10 in the peripheral nervous system and in melanocytes). Pleiotropic genetic variants acting on neural crest cell development, which is what the neural crest cell hypothesis predicts, could affect any of these processes, up until but likely not including the regulation of differentiation at the target site. Since the publication of Wilkins et al. (2014), several selective sweep scans in domestic animals have reported enrichments of neural crest-related genes and individual candidate genes related to the neural crest (reviewed by Wilkins 2017). There are several weaknesses in the evidence for the neural crest cell hypothesis that, in our view, make it implausible and premature as an explanation for domestication traits and as a unifying hypothesis for the domestication syndrome. In this letter, we present three arguments against the neural crest cell hypothesis as a unified explanation for traits under domestication. These are: Several of the key traits potentially explained by neural crest cell mechanisms are not parts of a universal domestication syndrome. To the extent that there is a domestication syndrome, that does not imply a universal genetic mechanism. Even if we postulate a universal genetic mechanism, the evidence that this mechanism is neural crest cell proliferation and migration is indirect and weak. We then examine in detail some of the evidence from genome scans that is purported to support the neural crest cell hypothesis, which is weaker than it first might seem. Finally, we suggest some future research directions for the genomics of domestication in the light of these arguments. The idea of a domestication syndrome is some 40 years older than the neural crest cell hypothesis, and not dependent on any particular genetic mechanism. However, for the neural crest cell hypothesis to have anything to explain, there must be an at least somewhat universal domestication syndrome. The term “domestication syndrome” was popularized for animals by Price (1984, 2002), though the original idea came from plant biology (Faegri 1981; attributed by Faegri to Swiss botanist Albert Thellung). At its core, the domestication syndrome consists of alterations in behavior (increased tameness), body size and composition, brain size and composition, and color [though even this was not considered initially (Price 2002; Wright et al. 2020)]. These core domestication traits have been consistently altered in almost every domestic animal, but the type and direction of changes are not always consistent in all domestic animals. If we take brain size and composition as an example, the telencephalon is most consistently increased in relative size to the rest of the brain in pigs, llama, and alpaca (Kruska and Röhrs 1974; Kruska 1982, 1988), but not in the case of sheep (Ebinger 1974). In the dog, brain composition is highly variable between breeds (Hecht et al. 2019). Similarly, in domestic birds the cerebellum and cerebral hemisphere have consistently increased relative to the rest of the brain in domestic chickens, geese, pigeons, and turkeys, but not in domestic ducks (Ebinger and Löhmer 1984, 1987 ; Ebinger 1995; Ebinger and Röhrs 1995; Gille and Salomon 2000; Rehkämper et al. 2008; Henriksen et al. 2016). There is a wide degree of variation in the extent of allometric brain size reduction between domestic species, with pigeons only showing a 7% reduction, whereas turkeys exhibit a 35% reduction. Body composition and skull changes also show a similar pattern. Sánchez-Villagra et al. (2017) looked in detail on skull morphology in pairs of wild and domestic animals and Wilson (2018) at developmental trajectories of skull morphology within domestic species. Neither found a single universal pattern. That is, skull morphology has changed in many domestic animals, but it has not changed in the same way. The domestication syndrome is often expanded to include a multitude of other traits, including hair and feather morphology, dentition, and many others characteristics (Sánchez-Villagra et al. 2016; Lord et al. 2020; Wright et al. 2020). Several of the key traits that inspired the neural crest cell hypothesis—facial morphology, floppy ears, tail shape, and so on—should not be considered part of the domestication syndrome proper, because there is no good evidence that they are universal in domestic animals. We also need to keep in mind that differences observed between breeds of the same species may have happened during breed formation and selective breeding, and are not necessarily reflective of changes during the initial phase of domestication. For example, the floppy ears of certain dog breeds and experimentally domesticated foxes may or may not be due to changes in the regulation of neural crest cells; this is currently not known. However, even if they were, the occurrence of floppy ears is not a universal feature of domestic animals. The same goes for most other proposed domestication traits, as reviewed in detail by Sánchez-Villagra et al. (2016) and Lord et al. (2020). While there are knowledgeable scientists who argue otherwise (Lord et al. 2020), we believe that the domestication syndrome is a meaningful observation that warrants explanation, when restricted to core traits that are universal. One can certainly imagine neural crest-driven mechanisms for these core features: tameness might be caused by endocrine effects in the adrenal glands; brain size and composition might be indirectly affected by the size and shape of the cranium. However, the explanatory power of the neural crest cell hypothesis shrinks when we keep in mind that many of the features it aims to explain are not universal features of domestication, but specific to the history and biology of particular species. Even where traits are consistently altered in domestic species, or change together in experimental domestication studies, the mechanism may not be pleiotropic alleles. In support of the neural crest cell hypothesis, Wilkins (2020) argues that “The simplest way to view the situation is that a GRN [gene regulatory network], affecting all these traits, is involved and that mutations in relatively upstream elements in this GRN account for many of the traits ….” However, given recent developments in quantitative genetics, explanations based on a single gene regulatory network appear to be the exception rather than the rule. Recent evidence from quantitative genetics (Boyle et al. 2017) in humans suggests that polygenic traits tend to be omnigenic. That is, heritable variation in traits is not mainly due to a core biological process or pathway, but causative variants may act on almost any gene expressed in a relevant cell type, mediated by trans-regulatory variation (Liu et al. 2019). Experimental domestication studies provide a way to isolate the effect of selection on behavior, and reveal genuine correlated responses. However, a correlated response does not imply pleiotropy. The most famous domestication experiment is the farm fox experiment of Belyaev and co-workers (reviewed by Belyaev 1979; Trut et al. 2009; Lord et al. 2020). The experiment has been widely claimed (for example by Wilkins et al. 2014) to support a universal domestication syndrome caused by selection on behavior, and it is one of the main pieces of evidence that strong pleiotropy underpins domestication. However, there are several issues that have been raised that cast doubt as to how applicable this experiment is to a domestication model, and also how the traits themselves are correlated and documented. Importantly, the farmed fox experiment was actually founded on already domesticated animals (Statham et al. 2011; Lord et al. 2020). The co-occurrence of domestication syndrome traits in natural domesticates may be due to selection on multiple traits, and recent evidence suggests that this might be the case even for the farm fox domestication experiment. As such, the different domestication features that were present in the selected lines have already been present in the population and not developed as a result of selection. Linked selection can cause apparent pleiotropy between unrelated genes, when causative variants for different traits are located close to each other in the genome, even if the variants themselves share no This mechanism can potentially also cause changes between different species, to the extent that gene is between species. Recent of experimental evolution from suggest that selection is a of change in selection and 2020). This effect is likely to be even in studies with population and correlated selection in the farm fox experiment as evidence of pleiotropy is Furthermore, there are other explanations for between domestication traits than a origin in the neural crest. For example, selection on behavior certain universal feature of may affect through changes in the nervous upstream in the A effect of selection for tameness on the endocrine system was initially hypothesis before it was as a effect of changes to the neural crest by Wilkins et al. (2014). evidence of pleiotropic alleles the domestication syndrome is studies of domestic species can the extent of pleiotropy and close between different domestication traits, with the and power to small effects in from that domestication traits initially with one when an population is but when are and between these are found to actually be and rather than pleiotropic et al. Similarly, of lines selected for and initially by Belyaev et al. found only between domestication traits. of tameness in the farm foxes themselves not the but one associated with different of tameness behavior et al. genetic of this might or pleiotropic effects on other traits. When expression quantitative are considered the and a of are no evidence of pleiotropy et al. et al. 2016; et al. 2020). In even if we that there is a core domestication syndrome to selection on behavior, as the farm fox experiment that response may be due to or by pleiotropy that is mediated through and mechanisms that may or may not the neural crest. The neural crest cell hypothesis may be a explanation than potentially many biological processes, but based on quantitative genetics and we have to complex traits to be by a single biological Recent studies et al. et al. 2018; et al. have found evidence of selection at a few of genes that have associated with the neural but that in is not good evidence that neural crest cell proliferation and migration are the domestication syndrome. For that, we need evidence that these alleles were selected for their function in the neural crest and that they cause pleiotropic effects on other traits. There are three main the selection provide only evidence for the neural crest cell there are to selection that provide an view of sweep scans are in their to polygenic of complex traits that may be due to small in at many and sweep scans also with from variation and which be to contribute to domestication in natural domesticates and in experimental population where there is for mutations to selective likely traits or for complex traits. This is consistent with selective sweep signals at genes for or traits in domestic animals, the in et al. genes in et al. and the gene associated with in et al. 2012). recent that in natural domesticated animals are likely to recent such as breed formation and For example, an how strong in the domestic at and were selected later than early domestication et al. et al. 2017). to that the same may be for other “domestication even with genuine signals of selection. when we a of genes with a certain function in selective sweep how is We have no for how many related genes be observed in a genome term are for their and of when in a genome et al. gene set enrichments and to be of gene set enrichments from selective sweep scans be with Furthermore, the of enrichments is because developmental genes are highly The same genes and are in many developmental For example, the same signaling and have different during different phases of neural crest development and temporal The same transcription neural in the nervous neural crest cell and in cells et al. and multiple can to the same pathways, and our to process from pattern. Furthermore, even gene expression is not an of cell type or cell the pleiotropy of many of these transcription because some of the factors that act as neural crest specifiers may be expressed in a particular this is certainly no that the cells involved are neural crest cells or from the neural crest. there is no evidence for the core of the neural crest cell hypothesis, that these selected alleles have pleiotropic effects on domestication traits. sweep scans provide any such evidence in as they not any When genome scans reveal selection at genes that have multiple functions, that is not necessarily good evidence that the selected pleiotropic effects on multiple traits. A selective sweep in or close to a key developmental regulatory gene might affect its expression in any of many different cell which need not be We on this in the which with some of the candidate domestication genes potentially related to the neural crest. In of neural crest-related genes in selective are not good evidence for the neural crest cell hypothesis of domestication. This is because selection is to complex traits in early domestication, because are to selective sweep scans and developmental processes, and because evidence of pleiotropic alleles, as by the neural crest cell hypothesis, is We examine in detail the evidence for the neural crest cell hypothesis found in studies of domesticated animals et al. 2014) and et al. and then in a selection of the farm fox experiment et al. In the case of et al. (2014) the authors themselves not the neural crest cell this is due to Wilkins One keep in mind that genetic and selection are and have and the of evidence is not evidence of We are not the population genetic in these studies, or that these are strong evidence in of any other what we argue is that they provide only evidence for the neural crest cell and are transcription factors with developmental associated with selection signals in the domestic et al. Wilkins (2017) them as support for the neural crest cell in the original they were in of neural genes are involved in the neural crest gene regulatory network et al. but this is by no their is a that cell et al. to the neural is expressed in the neural before neural crest induction and neural crest formation (reviewed by and 2017). is in neural crest cells induction, and during but in a of neural crest cells that are to become peripheral or cells et al. of the neural neural by differentiation is expressed in neural of the nervous where it differentiation et al. is expressed during development of the nervous including in the neural and the and et al. in and and in with suggest that to of the nervous system et al. However, a recent of the neural crest gene regulatory network by and found evidence that is involved in neural crest cell development, being associated with et al. 2019). A selective sweep in et al. selection on and Wnt signaling as support for the neural crest cell The evidence consists of three related Wnt signaling Wnt signaling pathway, cell and regulation of Wnt signaling of of and a of other genes and of of Wnt signaling is a developmental many other processes, a Wnt is involved in of the neural tube, and of the and of the nervous system (reviewed by and signaling, on the other was not but by the of and in selected are involved in almost all tissues, development and and The farm fox selection of et al. (2018) support for the neural crest cell hypothesis in its and in its but to the main the support to the of and genes with Wnt signaling to the development of every part of the body and a in neural crest cells, but they are also involved in nervous system development, neural and et al. and In the genes that have been crest cell is but They are involved in a multitude of developmental in a wide of cell and any evidence that their in domestication is mediated by the neural crest is This type of be anything as because is the of is not Several of the candidate genes and have in the development of the nervous system and of the neural crest. selection on the nervous system and on the neural crest cell might to similar of the is with its it to the that the evidence consists of a few genes or a few At the same et al. (2018) and et al. (2018) the neural crest cell hypothesis, because it is the one domestication that is popular at the The is that the with the that the neural crest cell hypothesis of domestication is by evidence from multiple species, when the evidence is at and We that a of the neural crest cell hypothesis lead to research than to for evidence of neural crest cell While there may be many different of research and correlated traits under domestication, we suggest some directions for genomics of population The of and the of population genetic for selection have the selection studies of domestication. are at from variation and at with complex it is to the population genetic that is already there with population genetic and of That make it to between species and which genes or are universal. For example, there is evidence of of selection during domestication between sheep and et al. While the of selection this is to be than gene from studies that them in different genomics of The development of has the to be a for domestication is to of gene expression and in of (e.g., et al. et al. and at with If to wild and domestic animals, such studies could what transcription and between wild and domestic animals. If to the neural crest cells, for example is by et al. in a we if the and differentiation of neural crest cells are different between wild and domestic animals. of the neural crest cell hypothesis of domestication is that it has increased the in development domestication There have been to differences in the neural crest function in domestication for example in (Wilkins et al. Since these studies have not we that this of studies is than There is also a in genetics from a set of genes and pathways, and scans a that with the genome, and may genes that are has the to make relevant on a A change of from to a we to that a universal mechanism at the core of domestication might not or might be in what traits and species it we to research particular traits, species, and their rather than the universal. There have been many of research on domestic animals that in on specific traits that are relevant to different phases of domestication, in et al. and to in domestic et al. 2020). The neural crest cell hypothesis has an domestication, including on the on Sánchez-Villagra and for a of the history and of the If there is a between evolution and domestication, it might be to evolution by the genetic of domestication. However, we argue that any universal genetic neural crest-related or proposed for domestication are currently not for to other In the neural crest cell hypothesis is an explanation for a and it is implausible in the light of quantitative The evidence suggests that some variants genes involved in the neural crest have been but not that they were selected for their function in the neural that they have pleiotropic effects that domestication traits. the evidence signaling such as Wnt and that contribute to many developmental the neural crest. the of the genetics of early domestication, we suggest that it is premature to any unifying mechanism.
Source: Genetics
Language: English
Subfield: Genetics
Open Access status: hybrid
Estimated APC paid: $4,000
OpenAlex ID: W3177596996
FWCI: 4.653
Cites: 75
Cited by: 59
Related to: 9