Isolated ectodermal cells become neural. However, entire ectodermal explants do not. This is because BMP4 is produced in the ectoderm. As in the mesoderm, BMP is a ventralizing signal. Ventral ectoderm becomes epidermis. Hence, BMP produced by the ectodermal explants induces the entire explant into epidermis.

In the embryo, BMP antagonists secreted by the organizer (dorsal mesoderm) to overlying ectoderm inhibit the ventral fate. Furthermore, Isolated ectodermal cells produce BMP4, but it diffuses away and the ventral fate is thus not induced. Unless overridden by BMP4, ectodermal cells by default differentiate as neural even in the absence of chordin and noggin.

Injecting morpholinos against BMPs into frog embryos will cause excess neural tissue to form. This reinforces the theory that BMPs induce ectodermal cells to become epidermal, and that BMP antagonists prevent this to allow cells to default to a neural state.

Ectoderm appears to undergo a neural path by default, but in fact they have already received neural signals. For example, treatment of embryos with FGFs (fibroblast growth factors) stimulates neural development, and FGFs are produced throughout the embryo.

In invertebrates, neural precursor cells segregate from a specialized region within the ectoderm, called neurectoderm. The neurectoderm contains cells that will give rise to epidermis (the outer covering of the insect) and cells that will give rise to the nervous system. The neural precursors delaminate individually from the neurectoderm into the interior of the embryo, where they begin to proliferate and are referred to as neuroblasts.

This delamination involves a number of the cytoskeletal and cell adhesion mechanisms that we discussed previously. The delaminating cells lose their connections, via tight junctions and adherens junctions, to cells of the ectodermal epithelium; this is an epithelial to mesenchymal transition. In addition, the delaminating cells constrict at their apical surfaces via microfilaments, and elongate via microtubules.

Neuroblasts in insects are determined in a two-step mechanism.

Proneural

Proneural clusters are groups of neuroectoderm that become competent to give rise to neuroblasts. This is due to their activation of proneural gene expression. Proneural genes a group of bHLH (basic helix-loop-helix) transcription factors encoded by the Achaete-scute complex. The requirement for the proneural genes is shown by the fact that many neuroblasts do not appear in Achaete-scute-/- embryos.

Inhibition

One of the cells that has delaminated from the proneural cluster inhibits other cells of the cluster from becoming neuroblasts. This process is called lateral inhibition, and it is important for assuring that the correct number of cells become neuroblasts.

Lateral inhibition is mediated by Notch signal transduction. Notch is a transmembrane receptor that interacts specifically with Delta, a transmembrane protein. Therefore, Notch-Delta interaction requires direct cell-cell contact. Initially, all cells in the proneural cluster express low levels of Notch and Delta. One cell in each proneural cluster gains a slight advantage over the others by expressing slightly more Delta. This is likely stochastic (random). This cell thus engages more Notch receptors on neighboring cells, causing them to downregulate proneural gene and Delta expression whilst retaining Notch expression. These cells thus lose the ability to differentiate as neurons (due to decreased proneural gene expression) and also lose the ability to inhibit proneural expression in other cells (due to decreased expression of Delta). As a result the single cell that expressed slightly more Delta is the only cell that maintains proneural gene expression.

Genes involved in this inhibitory process are called neurogenic genes; their required role in this repression is demonstrated by the fact that, in an embryo homozygous for a null mutation in a neurogenic gene, the entire neurectoderm develops as neuroblasts (since all of the cells in the proneural clusters, which constitute the neuroectoderm, now develop as neuroblasts).

Notch and Delta are extremely important neurogenic genes in Drosophila. The Notch receptor and its various ligands have been found to play multiple roles in cell signaling and cell fate determination not only in Drosophila, but also in vertebrates. Moreover, their role in lateral inhibition is largely conserved in vertebrates.

Recall that in amphibian gastrulation, the mesodermal and endodermal cells must enter the interior of the embryo in order to correctly arrange the three germ layers. This is initiated through an invagination, followed by involution. The dorsal lip of the blastopore is the site where this process is initiated, and it forms opposite the site of sperm entry on the dorsal side of the embryo. Spemann and Mangold wondered if all cells are undetermined at the beginning of gastrulation. They transplanted the dorsal lip of the blastopore of an early gastrula from a light-colored newt into an ectopic region of a dark-colored newt early gastrula.

Spemann and Mangold observed that the donor tissue (dorsal blastopore lip) invaginated into the interior and a second embryonic axis developed. The notochord of this second embryonic axis was composed entirely of graft donor tissue, while the neural tube, and somites (a mesodermal tissue that we will discuss in a later lecture) were composed only partly of graft (donor) tissue. The notochord is the most dorsal mesodermal derivative. All other tissues of the new axis (eg, kidney tubules and gut) were entirely of host tissue. Spemann and Mangold concluded that the graft tissue induced a new embryonic axis.

The organizer (induced by the Nieuwkoop Center) dorsalizes the adjacent mesoderm by secreting Noggin, Chordin, Follistatin and Frzb. These secreted proteins induce the neural plate by antagonizing the ventralizing proteins BMP-4 (noggin, follistatin, and chordin) and Wnt-8 (frzb). Recall that the Organizer organizes the mesoderm around it into an axis by dorsalizing the mesoderm so that dorsal derivative like somites (future trunk muscle and bones) and intermediate derivatives (kidney and gonad) can form. In vertebrates, the most dorsal mesoderm forms the notochord; this requires the dorsalizing Organizer secretions.

Genes encoding Noggin, Chordin, Follistatin and Frzb are able to induce a second axis. Similarly, the major component of the Nieuwkoop center (β-catenin) can also induce a second axis. Thus, the major role of the Nieuwkoop center is to induce the organizer, which in turn produces BMP and Wnt antagonists. For example, injection of mRNA of noggin or chordin into an early embryo can induce a second axis. This second axis is due to the resultant protein’s ability to dorsalize ventral mesoderm (resulting in the notochord and somites), and to dosalize the overlying ectoderm (resulting in the neural plate).

Starting at the ventral end, the following structures comprise the spinal cord: floor plate; cell bodies of somatic motor neurons; cell bodies of commisural neurons; roof plate. The floor plate guides axons that grow into it, and the motor neurons send out axons that innervate muscles.

The notochord patterns the neural tube’s dorso-ventral axis. An additional notochord transplanted underneath a closing neural tube in the chick embryo induces an additional floorplate with adjacent motor neurons. Removing the notochord from the neural tube results in absence of a floor plate and motor neurons. Thus, the notochord induces at least the most dorsal neural tube structures. Furthermore, both the notochord and floorplate induce adjacent somite tissue into sclerotome, which gives rise to the vertebrae and ribs.

Vertebrate Sonic hedgehog protein (Shh, homolog of Drosophila Hedgehog) is localized to the notochord and floorplate of the spinal cord. Perhaps Shh is the notochord signal that patterns the neural tube in dorso-ventral axis, and that patterns the somites. Addition of cells producing Shh has a similar effect to adding an additional notochord: an ectopic floorplate and motor neurons. The floorplate differentiates where the level of ectopic Shh is highest and the motor neurons where the level is lower. This indicates that a gradient of Shh released from the notochord patterns the neural tube.

The key role of Sonic hedgehog was revealed by the phenotype of Sonic hedgehog knockout mice. Shh^-/- mice die early and are extremely defective. The defects observed are consistent with a role of Shh in inductions:

Floorplate	Shh form the notochord induces the floorplate. Shh-/- fetuses have severe ventral midline defects, i. e., the ventral midline seems to degenerate. This results in a fusion of the eye and nose primordia in the ventral midline (cyclopia). This defect is consistent with the known role of Shh in inducing the formation of the floorplate and other ventral cell types in the neural tube.
Sclerotomes	Shh from the notochord and floor induces sclerotomes. Another defect is the virtual absence of the vertebrae and the ribs. Again, this is consistent with the role of Shh in giving a more ventral identity (sclerotome) to a portion of each somite. We will discuss this aspect of Shh signaling in a future lecture.

The posterior neural tube gives rise to the spinal cord. The anterior neural tube gives rise to the brain. The neural tube begins subdividing into a complex series of bulges (later forming the brain) before it has even completely formed at the posterior. First, three primary vesicles are formed. Starting at the anterior: the forebrain (prosencephalon), the midbrain (mesencephalon) and the hindbrain (rhombencephalon). The prosencephalon then becomes subdivided into the telencephalon at the anterior and the diencephalon at the posterior. The telencephalon will give rise to the cerebral hemispheres. The diencephalon evaginates to form the optic cup (which will form much of the eye), and also will give rise to the thalamic and hypothalamic regions of the brain.

The rhombencephalon is the only part of the brain that appears to be segmented into separate bulges, referred to as rhombomeres. Each rhombomere is a morphologically distinct compartment. A compartment is a region of cell lineage restriction. Once boundaries between individual compartments form, cells are confined to their respective compartments and cannot cross from one side of a boundary to another. This implies that the cells within each compartment have differential affinities for each other, but not for cells from neighboring compartments. This is similar to the segments of the Drosophila embryo. Different types of appendages arise from different Drosophila segments. Similarly, distinct cranial nerves arise from different rhombomeres.

The strongest evidence that rhombomeres represent metameric units analogous to the segments of the Drosophila embryo is that their identities are controlled by Hox genes. Anterior boundaries of expression of individual Hox genes mark boundaries between rhombomeres. Knockouts of certain Hox genes in mice (for example Hoxa-1 and Hoxa-3) result in abnormal development of the anteriormost rhombomere in which the gene is normally expressed. As with Drosophila segments, loss of Hox genes leads to transformations to a more anterior identity. Hox genes thus appear to play similar roles in rhombomere identity, vertebral identity and Drosophila segment identity.

Hox genes are not expressed in the midbrain and forebrain. These regions are probably also divided up into subunits by the spatially localized expression of various other transcription factors. There are at least 25 homeobox-containing genes that are expressed in regionally restricted patterns in the forebrain. Many of the genes expressed in these overlapping patterns (Emx, Otx, Dlx) are homologous to genes (ems, otd, Dll) expressed in segmental patterns in the embryonic fly head. The few mutations observed so far have been instructive. For example, Drosophila otd (and its murine homolog Otx) mutants lack the anterior portion of the brain.

A key feature of cells located at the boundaries of compartments is that they often serve as signaling centers. We saw this when we considered the segment polarity genes of Drosophila. We saw that the cells on either side of the compartment boundary expressed either Hedgehog or Wingless proteins.

These proteins then diffused away from their sources and influenced the development of neighboring cells in the compartments. Similarly, boundaries of lineage restriction in the developing midbrain and forebrain act as signaling centers. For example, a signaling center forms at the boundary between the midbrain and hindbrain.

The cells at this boundary not only act as a barrier of cell lineage restriction, but also secrete extrinsic signals, such as FGF8 and Wnt1 (a homolog of Wingless). These molecules are thought to form gradients that pattern adjacent regions. Roles for FGF8 and Wnt1 are demonstrated by the fact that tissue-specific loss of these genes in neural tissue leads to loss of the midbrain. Other signaling centers have been found in the forebrain, but their functions are less well understood.

Injecting Sonic hedgehog protein (Shh) → extra floor plate. Sonic hedgehog is expressed in notochord and floorplate. Shh is a homolog of Drosophila hedgehog. High levels of Shh → floorplate. Low levels of Shh → motor neurons.

As the primitive streak regresses posteriorly, cells furthest from the node are no longer under the influence of FGF. Once outside the influence of FGF, paraxial mesoderm cells begin to compartmentalize into somites. Cells within an individual somite compact by increasing cadherin expression. These changes in cell adhesion cause the newly forming somite to separate from the rest of the paraxial mesoderm.

The region separating a newly formed epithelial somite from neighboring unsegmented paraxial mesoderm is a boundary. Boundaries arise in vertebrate brain rhombomere compartments and Drosophila embryo segments. Compartments are kept separate due to differences in cell adhesion. The same is true for cells in somites and in the unsegmented paraxial mesoderm.

The advantage of dividing the paraxial mesoderm into separate somites, each of which is an individual compartment, is that this allows the formation of repetitive structures such as vertebrae to use a common genetic pathway in each somite. The Notch pathway is critical for formation of a boundary between the newly formed somite and unsegmented mesoderm, and between rhombomeres.

The most apical (facing the lumen of the neural tube) zone is the ventricular zone; this is where neuronal progenitors undergo cell division. This can be seen as the region where 3H-thymidine is incorporated. Once cells are postmitotic (undergone their last division) they migrate away from the ventricular zone to differentiate. The identity of the newly born neuronal or glial cell is specified before it starts to migrate.

Asymmetric cell division may be involved in the formation of neuroblasts (cells committed to terminal differentiation) from neural progenitor cells in the ventricular zone. In Drosophila, asymmetric cell divisions lead to the segregation of cytoplasmic determinants that affect the Notch pathway. There is some evidence that similar mechanisms operate in parts of the vertebrate nervous system.

In order to generate a functional nervous system, the correct numbers of neurons must be generated, and they must differentiate into the correct cell types. How is the number of neurons regulated? Due to lateral inhibition, only one cell in each proneural cluster becomes a neuroblast. Similarly, in vertebrates, lateral inhibition controls the rate at which neuroblasts are specified from the cells in the ventricular zone.

Lateral inhibition is mediated by the Notch signal transduction system. Notch and Delta are transmembrane proteins. Notch is a receptor that interacts with specific ligands including Delta. Therefore, the Notch-Delta interaction requires direct cell-cell contact. Initially, all cells in the proneural cluster express low levels of Notch and Delta. Activating the Notch receptor inhibits expresion of proneural genes, and Notch-Delta interaction downregulates Delta expression. Through an unknown mechanism, one cell in each proneural cluster gains a slight advantage over the others by expressing slightly more Delta.

In a likely schochastic (random) process, one cells expresses slightly more Delta than its neighbors. This cell thus engages more Notch receptors on neighboring cells, downregulating their expression of proneural genes and Delta whilst retaining their Notch expression. Thus, these neighboring cells lose the ability to differentiate as neurons (due to decreased proneural gene expression) and also the ability to inhibit proneural expression in other cells (due to decreased expression of Delta). As a result the single cell that expressed slightly more Delta is the only cell in which proneural gene expression is maintained.

The importance of Notch-Delta signaling can be seen from analysis of mutants in this pathway. A clear example is provided by the sensory receptor cells of the inner ear. These cells are called hair cells for the characteristic cilia on their apical surfaces. These cilia are deflected by mechanical forces created by sound waves, and the signal is transmitted along axons to the brain. Hair cells are surrounded by various types of support cells.

pattern of hair cells in the inner ear is highly invariant, and this regular pattern is required for proper sound reception. In Delta^-/- mutants, lateral inhibition is defective, and all cells, including the support cells, develop as hair cells. Thus, lateral inhibition normally regulates the number of hair cells that develop, ensuring that there is a correct balance between hair cells and supporting cells.