The neuronal precursor cells in the ventricular zone of the embryonic brain look and act more or less the same. Yet these precursors ultimately give rise to postmitotic cells that are enormously diverse in form and function. The spinal cord, cerebellum, cerebral cortex, and subcortical nuclei (including the basal ganglia and thalamus) each contain several dozen neuronal cell types distinguished by morphology, neurotransmitter content, cell surface molecules, and the types of synapses they make and receive. On an even more basic level, the stem cells of the ventricular zone produce both neurons and glia, cells with markedly different properties and functions. How and when are these different cell types determined?

One possibility is that the precursors of different populations of neurons (or neurons and glia) are established very early in development, perhaps at the formation of the neural plate (Figure 22.8A). Separate types of precursor cells would then exist in the ventricular zone, each giving rise to a particular type of cell in the adult. According to this school of thought, a cell's fate is a function of its lineage: Neurons of different types would have distinct “ancestors,” as would glial cells. At the other extreme, the ancestral precursor cells might provide essentially no information about eventual phenotype; in this scenario, all subsequent differentiation depends on interactions with other cells in particular brain microenvironments (Figure 22.8B).

There is little evidence that precursor cells are committed irrevocably early on to produce particular types of daughter cells. In fact, in several brain regions precursor cells generate postmitotic daughter cells throughout the course of development that assume a number of different phenotypes (Figure 22.9). In the retina, for example, experiments using lineage-marking techniques have shown that a precursor cell can generate any combination of cell types found in the retina, including photoreceptors, bipolar cells, amacrine cells, ganglion cells, and even glial cells. The developing spinal cord, superior colliculus, cerebellum, and cerebral cortex also appear to lack clearly specified precursor cells. Apparently, cell lineage plays only a minor role in specifying cell fate for at least these components of the developing brain.

The bulk of the evidence favors the view that neuronal differentiation is based primarily on cell-cell interactions (Figure 22.8B). Historically, most experimental approaches to this issue have relied on transplantation strategies, such as moving bits of a particular brain region to a different location in the brain of a host animal to determine whether the transplanted cells acquire the host phenotype or retain their original fate during subsequent development. When very young precursor cells are transplanted, they tend to acquire the host phenotype. Transplants at increasingly older ages, however, usually retain the original phenotype. The progressive restriction of possible phenotypes that a given cell can assume almost certainly results from local cellular and molecular cues that progressively limit the complement of genes expressed in that cell.

One of the best-understood examples of how such cellular and molecular interactions generate diverse neuronal phenotypes is the development of the eye in Drosophila. The fly eye is a stereotyped structure consisting of hundreds of repeating cellular units called ommatidia, each comprising eight distinct receptor cells (R1–R8; Figure 22.10). The assignment of phenotypes to individual cells is independent of the lineage of precursor cells, depending instead on a series of cell-cell interactions involving cell-specific ligands and receptors. These interactions have been analyzed in a strain of mutant flies called sevenless (because they are missing one receptor, R7). A similar phenotype has been observed in another mutation involving R8, called boss (for bride of sevenless). These studies have shown that the boss gene encodes a tyrosine kinase-linked membrane receptor, while sevenless encodes a membrane-bound ligand for this receptor. Each of these genes is expressed in a separate population of cells in the developing eye. The interaction between these two proteins therefore leads to a series of intercellular events resulting in the differentiation of the R7 neuroblast. Based on this and much other work, signaling via receptor kinases and their ligands is thought to influence a range of cellular differentiations in the nervous system.

These observations indicate that neuronal differentiation relies on signaling between distinct cell classes (see Figure 22.8B). Recent evidence indicates that a broad range of signaling molecules, including FGFs and TGFs as well as members of the vertebrate notch family of cell surface ligands and receptors, contribute to this process. Accordingly, neuroblasts must be in appropriate positions to participate in the signaling events that lead to their ultimate differentiation.