Let’s start from the very beginning, as this can be simply retold step by step as a story. It’s useful to keep the picture above open in a tab. The very first step is the induction of the neural plate, which is simply a sheet of neuron progenitor cells derived from the ectoderm – those ectodermal cells that don’t go into the neural plate will later become the epidermis. The specification of the neural plate cells happens pretty early during gastrulation, under the action of several signalling molecules (they’re different in every vertebrate lineage).
Under orders from surrounding border cells, it invaginates. Neurulation then happens, during which the two halves are joined to form the neural tube (last step in the picture). The neural tube gets specified anteriorly to make the brain and posteriorly to make the spinal cord; dorsoventrally, the neural crest cells get specified, as well as the epidermis. The brain gets further subdivided into fore- (prosencephalon), mid- (mesencephalon) and hindbrain (rhombencephalon).
First we’ll look at the neural crest. The migrating neural crest is often (rightly) touted as the innovation that really allowed vertebrates to become as successful as they have (in fact, it’s a common thing among invertebrate zoologists to say that “the only interesting thing about vertebrates is the neural crest”; I don’t know who came up with it) – even though its evolution was not a single key event, but rather a series of progressive stages leading up to the vertebrate condition. The neural crest is a population of multipotent neural stem cells found in the embryo that are derived from the neural tube.
What separates the vertebrates from the rest of the chordates is that in vertebrates, the neural crest cells can migrate. Mechanically, the cell-cell adhesion that usually holds the cells in groups gets lost, allowing them to move freely. They do this along specific pathways, differentiating along the way to become the target cell type – this differentiation involves cytoskeletal changes, along with new proteins in the cell membrane that allow all the signalling interactions to coordinate the whole process.
There are two categories (see picture below): the ectomesenchymal group are the ‘skeletal’ derivatives (bone, dentine, cartilage); the non-ectomesenchymal type are neurons, glia and pigment cells.
It is this migration that allows the facial skeleton and teeth to be built, as well as the anterior brain and the peripheral nervous system. The reason why it’s viewed as the major evolutionary novelty of the vertebrates is because it allowed the active predatory lifestyle that characterises them (for comparison, amphioxus adults dig themselves in the ground, leaving only their mouth open to the rest of the ocean); all the additional complexity that the vertebrates have in their head and body is directly due to the neural crest. The picture below summarises the former case.
The diagram shows the skulls of a chick and a mouse. The area outlined by the black dots is all derived from the neural crest and is not present in non-vertebrates.
The skull itself is not a single homogeneous structure. In addition to putting it together, all the facial muscles, the dermis, the connective tissues and the nervous system of the head are of primary importance for vertebrates – if anything goes wrong during development, the best one can hope for is a deformed face. The neural crest is almost single-handedly responsible for forming the whole craniofacial complex; problems with the neural crest can lead to First Arch syndromes and other such disorders.
All this time, I mention control of the cells; this is achieved mostly by gene expression. In the case of the neural crest, the surrounding tissues release specific signalling molecules that belong to the Wnt, FGF and BMP families. They basically turn transcription factors on and off (in this case, members of the Pax, Zic, Snail, Sox and Msx families), which in turn regulates what genes are active or not, which in turn specifies what the cell will become. This is all part of the “developmental cascade” – which actually begins from the 16-cell stage of the embryo and even earlier – and elucidating these pathways is what molecular developmental biologists do all day.
The Wnt family is one of the most important in developmental biology – whenever talking about vertebrates, you will definitely encounter it. It’s a family of glycoproteins (biochemically, they’re characterised by the palmitoylation of one of the cysteines in its structure). The first of these proteins was discovered in 1982 in cancerous tumours and only later was it found to play a key role in morphogenesis. Their most important role is in cell-cell communication, but they also regulate individual cellular movements, cell fates, cell death and cell polarity. They can either have a very local effect, only messing around with a small cluster of cells, or they can act tissue-wide. They are not only active in the embryo but also in the adult, where they play a role in neurogenesis and in regulating stem cells.
As signalling molecules, they act together with transmembrane proteins called Frizzled receptors (FZ). When the proteins meet, another protein, Dishevelled (DVL) gets activated inside the cell, and it enacts the message that the Wnt was conveying. This may be one of three possible pathways: the planar cell polarity pathway (PCP), the WNT/ß-catenin pathway or the WNT/calcium pathway. These are summarised in the picture below. It’s no use rephrasing my main reference here; I just wanted to show that no matter how complex a process might be, there are always logical pathways to explain it – this is especially true in cell biology.
With all that in mind, let’s look at how the neural crest develops in vertebrates (using the model example of a mouse). Neural crest cells get induced at the neural plate border, undergoing a transformation from epithelial to mesenchymal cells. This involves the activation of Snail transcription factors, which suppress various cadherins (cell adhesion molecules), and this is the first step that allows the neural crest cells to migrate (note that I am not reviewing the roles of all signalling pathways and their interaction, as it would be overkill – I can go into more detail if wished though!). The migration starts from the hindbrain rhombomeres (see later) and proceeds both forward to the anteriormost part of the embryo and backwards towards the tail. There are four pathways in total, corresponding to the four types of neural crest cells: cranial, cardiac, vagal and trunk.
The cranial neural crest cells are again divided into fore-, mid- and hindbrain. They migrate along the cranial mesoderm along very specific (and highly conserved) paths. The forebrain cells go towards the frontonasal and periocular regions, i.e. at the anteriormost part of the embryo (nose, eyes). The midbrain cells go into the first branchial arch. The hindbrain cells emigrate to colonise the first three branchial arches. The picture below shows the hindbrain neural crest cells’ migratory paths and what they end up forming at their destination. It’s important to note that these pathways are not stored in the cell but are guided by the environment, i.e. taking a cell from rhombomere 1 and placing it in rhombomere 3 will not screw anything up; the cell will just take the clues secreted from the surrounding cells.
In all, the cranial neural crest cells allow the formation of the whole tooth and jaw system (from tooth stem cells to the pulp, cement, dentine, etc. to the mandible itself), the meninges and all associated nerve ganglia and bones. They also give rise to various pigment cells, smooth muscle, connective tissue and cartilage – basically, your whole face, skull and everything in between comes from the cranial neural crest cells. I mentioned jaws in the parentheses there. It should be noted that the upper jaw (maxillary) and lower jaw (mandibular) are significantly different both in their development and morphology.
Given that the focus here is on the head, I’ll just mention that the neural crest also contributes to the heart (cardiac neural crest cells); vagal and trunk cells can also produce pigment cells, smooth muscle et al., but they are involved in the postcranial part of the animal, i.e. the endocrine system, the peripheral nervous system and the skin. But I will not go into this further.
As we’ve seen already, the skull is unique to the vertebrates – however, saying that it is a ‘singular’ structure is false, since it’s made up of three fused parts of separate origins. The most ancestral part is the viscerocranium (also called the splanchnocranium), which evolved from the pharyngeal arches (the structures that support the gills), meaning it’s derived from the neural crest; it makes up a large part of the jaw (in the agnathans, it’s a gill attachment site). The neurocranium is the part of the skull that protects the brain and sensory organs; its evolution is a matter of debate, with the tentative concensus being that it formed by the fusion of several vertebrae, in which case it’s of mesodermal origin. The dermatocranium is made up of dermal bones and is simply added to the other two parts; it can either be mesodermally or neural crest-derived.
The brain itself is much more developed than in tunicates and lancelets, which only have a neural tube. They all share one ancestral property, that they are divided into the fore-, mid- and hindbrain. However, the gross enlargement and structural differentiation of the vertebrate forebrain and midbrain are unique, as is the presence of rhombomeres. Another difference is in the mid-hindbrain boundary (the isthmic junction), which in the vertebrates has a fundamental role in controlling the brain’s patterning.
In the vertebrate forebrain, you find the telencephalon anteriorly and the diencephalon posteriorly. In mammals, it is where the cerebral cortex is found – the center of mammalian intelligence and by far the most elaborate brain structure in mammals. The telencephalon develops into the cerebral hemispheres, while the diencephalon develops into the thalamus and hypothalamus. The figure below shows how all this happens in the embryo.
On the left, you see a fate map: a diagram showing what each region will eventually become in the adult. Blue and green show the eventual telencephalon, pink the diencephalon. The red line is the notochord, and the dark blue-coloured area at the front is the anterior neural ridge (ANR). This is at a very early stage in the embryo. The middle picture shows a more advanced stage, with the telecencephalon (tel) and diencephalon (di) clearly differentiated. The arrows and their respective labels are signalling chemicals that are released and control the growth and continued differentiation of the area, all of which (alone!) lead to the picture on the right: the basal brain (this is generalised for all vertebrates). For orientation, the pallium is the ancestral structure from which the mammalian coretx is derived (see the picture below to see the pallium in different vertebrates). Abbreviations: thalamus (thal); hypothalamus (hyp); mesencephalon (mes); isthmus (is); cerebellum (cb); metencephalon (met).
The pallium is the shaded part. There are way too many abbreviations to list here. I can copy-paste them from the paper, but it’s kind of useless.
The telencephalon (i.e. cerebral hemispheres) in mice embryos divides into two regions: the dorsal pallium and the ventral subpallium – again, this is all controlled by differential gene expression. In the subpallium, Dlx and Gsh genes are expressed and eventually, it forms (among others) the basal ganglia, the central and medial parts of the amygdala and a large part of the septum. In the pallium, other genes are expressed: Pax6, Emx1/2, Neurogenin2, T-brain genes and Lhx9 (a homeodomain-containing gene), giving rise to (among others) the rest of the amygdala and septum, the cortex and the claustrum. These genes are not just coincidentally there: knocking them out (not allowing them to be expressed) results in serious defects and malformations.
And this principle applies to all the subregions within the pallium and subpallium. These gene expression profiles are so specific that a lot of the imaging in molecular developmental biology depends on staining genes to be able to tell the limits of each segment/section/compartment/etc.
I will not discuss the brain of any specific vertebrate group here, since they can be very different (see picture below). A comparative look at them may be the topic of a later post, if requested.
The midbrain does not have any subdivisions; it gives rise to the cerebral aqueduct.
Now let’s look at the hindbrain, an absolutely essential part of the CNS since it controls all the ‘automatic’ actions like breathing and heartbeat, motor control, as well as the facial musculature (including the eyes). In the adult, it’s divided into three main parts: the cerebellum, the pons and the medulla. In the embryo, it has two subdivisions: the anterior metencephalon and posterior myelencephalon. The former gives rise to the cerebellum (movement control), while the latter gives rise to the medulla omblongata (respiratory, cardiac, etc. control). Before that stage, however, it is divided into seven rhombomeres. The picture below shows what the hindbrain looks like at this stage.
Rhombomeres are extremely significant – it is from them that the neural crest cells get their migratory pathways. In a way, they are ultimately responsible for the development of the head; it is therefore of no surprise that they are only found in the vertebrates and none of the other chordate groups. So that you’re not confused, the picture below shows how early in the embryo this happens (the pharyngula stage of PZ fame).
And that does it for this series. As always, I did not go into the details, since this is just an introduction. And anyway, the level above this one simply involves looking at all the genetic and signalling pathways – not the job of a blog, that’s textbook or presentaion stuff.
References and Further Reading
Melton, K. R., Iulianella, A. & Trainor, P. 2004. Gene expression and regulation of hindbrain and spinal cord development. Frontiers in Bioscience 9, 117-138.