Riding the right wave: would the real neural crest please stand up?

Abstract

The neural crest has been subjected to the close attention of both embryologists and evolutionary biologists. It has been accepted that the neural crest was a vertebrate innovation that allowed the vertebrate head to appear since the fundamental and influential article by Gans and Northcutt was published 25 years ago (Gans and Northcutt 1983). However, recent data have once again stimulated interest in this issue, forcing us to reconsider this belief. In particular, cells that resemble the neural crest have been described in tunicates (Jeffery et al. 2004) and recent evidence based on molecular phylogenetics indicates that tunicates replace cephalochordates as the sistergroup to vertebrates (Delsuc et al. 2006). These data led Jeffery (2007) to propose that primordial neural crest cells were present in the ancestor of tunicates and vertebrates, although it was during vertebrate evolution that the crest began to give rise to derivatives other than pigment cells. Two lines of discussion immediately emerged from this proposal: (a) rather than a vertebrate novelty, the neural crest now appears to have been a novelty in invertebrates and (b) in addition to concentrating on the origin of the neural crest, perhaps attention should be devoted to studying the evolution of the neural crest itself. These two points were considered and discussed in a recent review by Donoghue et al. (2008), and similarly, Hall (2008) also addressed these issues, using the example of the Snail1 gene to illustrate that the new phylogeny is forcing us to reconsider what we believe to be the origins of the neural crest. Here, we would like to present our thoughts on these two issues and to clarify some aspects of Snail gene expression that might have been misunderstood in this latter review. First of all, and in complete accordance with Donoghue et al. (2008), the main characteristics of the neural crest can be used to produce a more accurate definition of this cell population, specifically, (a) cells that emerge from the neural tube and migrate to the periphery and (b) cells that differentiate into one or more cell types. Accordingly, ascidian cells that emigrate from the neural tube, express HNK-1 (a prominent neural crest marker), and differentiate into pigment cells (a prominent neural crest derivative) can be considered bona fide neural crest cells. This is perhaps the first indication that we need to reconsider the concept of the crest as a vertebrate novelty, although this would not undermine its importance in the generation of the vertebrate head. Having mentioned one neural crest marker, the efforts made by many groups to look for the genes involved in neural crest development should be acknowledged. In general, genes that have been described as ‘‘markers’’ for neural crest show broader expression patterns. For instance, the transcription factors Zic or Pax3/7 are involved in dorso-ventral patterning of the neural tube, and other proteins, such as cadherins or RhoGTPases, execute cell adhesion and/or motility programs in many developmental processes. In many cases, genes defined as neural crest markers were already present in invertebrates but expressed in other tissues. Hence, it appears that these genes, or even complete regulatory networks, could have been co-opted into neural cells from adjacent tissues in order to generate the neural crest cells (Meulemans and BronnerFraser 2004). This may well be the case for transcription factors such as Id, FoxD3, and AP2, which are particularly relevant for the consolidation of the crest population, the changes in cell adhesion to allow them to emigrate, and the survival of the migratory cells, respectively. Indeed, Id and FoxD3 are expressed in the mesoderm whereas AP2 is expressed in the ectoderm of invertebrates, including nonvertebrate chordates. Thus, it appears that they have subsequently been co-opted to fulfill a novel function in the vertebrate neural tube. We believe that perhaps it is not that important to define a specific molecular signature to describe the neural crest, but rather, we should continue to gain insight into the molecular networks required for the development of these cells (Meulemans and Bronner-Fraser 2004). The co-option hypothesis is even more appealing if we consider that both the mesoderm and neural crest have common characteristics: they migrate away from their place of origin (dorsal neural tube, primitive streak or blastopore) and they differentiate into a variety of cell types depending on their origin and the local interactions they experience as they migrate. Snail1 (or Snail2 depending on the species) is another transcription factor that should be considered when discussing co-option by the neural tube. Hall (2008) has used the example of the differential expression of Snail in ascidians and cephalochordates to support the argument that ascidians and EVOLUTION & DEVELOPMENT 10:5, 509 –510 (2008)