How the Chordates Got a Head

The vertebrate head is derived largely from the neural crest cells. Thus, Hall (2000) considers vertebrates to be not merely triploblastic animals but quadroblastic, with the neural crest constituting a fourth germ layer. Holland and Chen (2001) have even proposed calling vertebrates and their fossilized precursors “cristozoa,” the “crest-animals.” But how did these unique neural crest cells arise?

This is a critically important question in evolutionary developmental biology because it goes to the heart of evolutionary novelty. A turtle can form bones that no other organism has ever made—but they are still bones. A crustacean can turn a leg into a mouthpart, but the mouthpart is made of the same materials as the leg. But how does a completely new type of cell emerge? How did bone cells come into being for the first time? For this to happen, new combinations of transcription factors have to form a new and stable network (see the Introduction to Part IV). Scientists today are busy figuring out how combinations of gene expression patterns can become stabilized into a new cell type.

Only vertebrates have a neural crest, and the cranial neural crest is responsible for forming the bones and cartilage of the face, the migratory pigment cells, the peripheral nervous system, and much of the skull (see Chapter 10). Although we do not know how neural crest cells arose, there are cells at the neural plate/epidermal boundaries of cephalochordate and urochordate embryos that might be “latent homologues” of the neural crest. Both the urochordate tunicates and the cephalochordate Amphioxus express in their dorsal midline ectoderm many of the same genes expressed in vertebrate neural crest cells; however, they haven’t evolved the pathways that integrate these genes (Holland and Holland 2001; Wada 2001; Stone and Hall 2004; Sauka-Spengler et al. 2007).

Amphioxus is an invertebrate chordate that has a notochord, somites, and a hollow neural tube. It lacks a brain and facial structures, and most importantly, it lacks neural crest cells. In Amphioxus, the neural plate/epidermal border contains cells expressing several of the same genes expressed in vertebrate neural crest cells—Bmp2, Pax3/7, Msx, Dll, and Snail. However, the protochordate cells expressing these genes do not migrate, nor do they differentiate into a wide range of tissues (Figure 1A; Holland and Holland 2001).

Figure 1

Figure 1   Neural crest evolution. (A) One possible phylogeny of Chordata, showing a possible origin of the neural crest genetic regulatory network. The boxes across the top of the tree represent the elements of the neural crest genetic regulatory network defined by Sauka-Spengler and colleagues (2007). The patterning signals and many of the neural crest specifiers are found throughout the animal kingdom; however, the complete range of neural crest specifiers and effectors appear only in the vertebrates. Neural crest involvement is seen in novel structures appearing at each node (gold dots). (B) Expression in the mouse embryo of lacZ transgene on an Amphioxus Hox enhancer causes expression of the transgene in the neural crest cells of the first and second pharyngeal arches (those arches that will form jaws). (C) HNK1-expressing migratory pigment cells (stained brown) surrounding siphon of the tunicate Ecteinascidia turbinata. (A after Donoghue et al. 2008; B from Manzanares et al. 2000; C from Jeffery et al. 2004.)

In addition, many of the cis-regulatory elements needed for the positioning of neural crest cells in the vertebrate body are also present in Amphioxus. Manzanares and colleagues (2000) added regulatory regions of Amphioxus Hox genes onto a reporter lacZ gene and observed the expression of these chimeric genes in mouse and chick embryos. Remarkably, some of the Amphioxus Hox regulatory sequences caused the expression of the reporter gene in the neural crest cells and neural placodes of mice and chicks (Figure 1B).

Two transcription factors, AP2 and Distal-less, may have been critical in the formation of neural crest cells. Vertebrate AP2 is expressed in non-neural ectoderm, cranial neural crest, and neural tube cells, even in jawless fish. It appears to be essential for Hoxa2 expression in the neural crest cells. However, AP2 expression is confined to the non-neural ectoderm of Amphioxus. Meulemans and Bronner-Fraser (2002) speculate that altering the expression pattern of AP2 was critical in establishing the neural crest cell lineage.

Holland and colleagues (1996) have suggested that the origin of the neural crest involves the duplication and divergence of the Distal-less genes, which are found throughout the animal kingdom and are expressed in those tissues that stick out from the body axis, notably limbs and antennae (Panganiban et al. 1997). But in vertebrates, Distal-less has acquired new functions. Like Drosophila, Amphioxus has only one copy of the Distal-less gene per haploid genome, and as in Drosophila, this gene is expressed in the epidermis and central nervous system.

Vertebrates, however, have five or six closely related copies of Distal-less, all of which probably originated from a single ancestral gene that resembles the one in Amphioxus (Price 1993; Boncinelli 1994; Neidert et al. 2001). These Distal-less homologues have found new functions. Some are expressed in the mesoderm, a place where Distal-less is not expressed in Amphioxus. At least three of the vertebrate Distal-less genes function in the patterning of neural crest cells, and deleting them results in the absence or malformation of the pharyngeal arches, face, jaws, teeth, and vestibular apparatus (Qiu et al. 1997; DePew et al. 1999). Although it remains to be proved, it is possible that a new type of Distal-less gene could have played a major role in allowing migratory ectodermal cells of Amphioxus to evolve into neural crest cells.

Another hypothesis is based on finding neural crest-like cells in tunicates. In working with the tunicate Ecteinascidia turbinata, Jeffery and colleagues (2004) discovered a migratory cell population that emerges from the neural tube and that synthesizes two of the markers for neural crest cells—HNK1 (Figure 1C) and Zic. These tunicate cells do not make skeletons, however, but pigment. So it is possible that the neural crest arose not for skeleton formation, but to make pigments that could protect an ascidian-like ancestor from solar radiation. Indeed, new genomic and structural data suggest that the tunicates, and not Amphioxus, are the closest relatives of the vertebrates (Delsuc et al. 2006).

The development of the vertebrate skull was one of the greatest unsolved problems of evolutionary morphology. Alexander Kowalevski (1871) had demonstrated that the tunicates and Amphioxus were indeed chordates, having notochords and pharyngeal pouches, but he had no way of showing how the head forms in the vertebrates. Others therefore criticized his work and proposed numerous other progenitors for vertebrates (including worms and spiders; see Bowler 1996; Gee 1996). Now that the specification of the neural crest is known, we should soon be able to determine the mechanisms by which the neural crest evolved and became “the fourth germ layer” of the vertebrates (Donaghue et al. 2008; Bronner-Fraser and Sauka-Spengler 2010).