Intrinsic and extrinsic factors regulating vertebrate neurogenesis.

Recent studies of the factors regulating neurogenesis in vertebrates reveal three emerging themes. First, the number of cellular stages involved in this process may be greater than has previously been appreciated. Second, homologues of genes that regulate neurogenesis in invertebrates appear to play analogous roles in development of vertebrate nervous systems. Third, extrinsic factors can act to regulate neuron number during neurogenesis by controlling survival and differentiation, and not simply proliferation, of neural progenitor cells. Neurobiology


Introduction
In this review, I synthesize recent information on the molecular factors that regulate neurogenesis in vertebrates. I will focus on a few classes of molecules, both intrinsic and extrinsic to neuronal precursor cells, that regulate their proliferation, differentiation, and survival. The studies that will be discussed make use of several different vertebrate systems, including rodents, birds, and Xenopus, and employ varied techniques, such as tissue culture, genetic analysis, and experimental embryology. One area of vertebrate neurogenl:sis that, for reasons of space, is not dealt with here is the interesting story that is emerging about the relationship between neurogenesis and song learning in songbirds.
The reader is referred to a recent review by Doupe [l] for an introduction to this topic.

General features of vertebrate neurogenesis
The cells that make up vertebrate nervous systems ultimately descend from the ectodermally derived neuroepithelial cells of the neural tube and the neurogenie placodes [2]. In the peripheral nervous system, mesenchymal neural precursors (neural crest cells) emerge from these epithelial structures, and migrate to often distant targets, continuing to proliferate as they migrate [3]. In the central nervous system, most cell proliferation remains confined to an expanding neuroepithelium, which later comes to be called a germinal or ventricular zone as a growing mantle of postmitotic cells forms around it. Some exceptions to this rule occur in locations such as the cerebellum, where the precursors of granule neurons leave the ventricular zone and migrate to the pial surface of the developing brain to form a highly proliferative external granule layer [4].
Before the control of vertebrate neurogenesis can be fully understood, certain basic questions must be answered about how neurogenesis proceeds, such as what sorts of lineages give rise to the cells of the nervous system, and how many phenotypically distinct cellular stages precursors must pass through on their way to becoming tillly differentiated neurons.
The first question has been approached largely through lineage mapping and cell transplantation studies that involve the marking ofprecursors with non-transmissable viruses, dyes, or enzymcs. The results of these studies indicate that, in many cases, a rather large degree of multipotentiality exists among neural precursors, sometimes even very late in their development [5-131, although

Are there neuronal stem cells?
The stem cell question is a particularly thorny one, because, unlike regenerating tissues such as liver, muscle or blood, most vertebrate nervous systems do not replace lost neurons.
Thus, it is possible that most neuronal lineages operate without stem cells, simply producing

Cellular stages in vertebrate neurogenesis
Probably the least well understood aspect of vertebrate neurogenesis concerns the identification of the phenotypic stages that precursor cells pass through on the way to becoming neurons or glia. In other tissues, the cxistence of such stages has been deduced from progressive changes in morphology or growth factor requirements (e.g. the hernatopoietic system), changes in the expression of cell-surface antigens (e.g. lymphocytes), or changes in relative cell positions (e.g. epidermis and gastrointestinal epithelium) -reviewed in ( In addition to regulating the determination and proliferation of neural precursors, polypeptide growth factors apparently can also regulate the survival of neuronal precursor cells at particular stages of precursor differentiation. The clearest examples of this phenomenon, so far, come from irr l~itro studies of the actions of neurotrophins on the precursors that give rise to sympathetic neurons (sympathetic neuroblasts). Nerve growth factor (NGF), the prototypical neurotrophin, has long been known to be a survival factor for post-mitotic sympathetic neurons (for a recent review, see [83]). However, during the genesis of sympathetic neurons, sympathetic neuroblasts express trkC, the receptor for the neurotrophin NT-3, before they express t&A, the receptor for NGF [84*,85*]. Although NT-3 does stimulate proliferation of sympathetic neuroblasts to some extent, it also promotes survival of these cells when they are cultured as an isolated cell fraction, whereas FGF2, a stronger proliferationstimulating f&tor, does not enhance survival above control levels [&+*I. This separation of NT-~'S effects on proliferation and survival has been observed independently by DiCicco-Bloom and colleagues [85*], who foulid that NT-3 increased the number of sympathetic neuroblasts present at 24 hours in culture even when cells were blocked from entering S phase by treatment with aphidicolin, an inhibitor of DNA polymerase.
What could be the purpose for regulation by neurotrophins of the survival of proliferating neuronal precursor cells? Certainly this mechanism could provide a second level of control in regulating the size of neuronal precursor pools. Specifically, extrinsic factors could act to alter the number of precursors by controlling not only the number of precursor cell divisions, but also whether the progeny of these divisions survive to reach the next stage of their development.
Interestingly, however, in the case of sympathetic neuroblasts, neurotrophins can also act in another way to drive precursors toward the next stage of their development: culturing sympathetic neuroblasts in NT-3 not only promotes the survival of sympathetic neuroblasts, it also (at higher concentrations) induces expression of trkA and NGF responsiveness [86"]. The induction of trkA appears to be a consequence of the mitotic arrest of sympathetic neuroblasts that is brought about by high concentrations of NT-3. Indeed, both effects could be produced by treating cultures either with antimitotic agents (e.g. aphidicolin and mitomycin C) or with ciliary neurotrophic factor, a cytokine previously shown to inhibit proliferation of chick sympathetic neuroblasts [87]. NT-3 may also play a role in stimulating terminal neuronal differentiation by precursors in other areas of the nervous system. For example, a recent report indicates that NT-3 promotes differentiation into motoneurons of early (neural tube) progenitors cultured from quail [88].

Conclusions
Recent studies of precursors in several different neuronal lineages suggest that vertebrate neurogenesis is regulated by both intrinsic and extrinsic factors at every stage of neuronal precursor development, from initial commitment of multipotent progenitors to induction of the specific gene expression that is characteristic of terminal neuronal differentiation.
The experiments that have led to this view have, in the process, also begun to reveal distinct cellular stages through which neurogenesis in several systems proceeds. Based on analogies (and homologies) with Drosophila neurogenesis, and on the diversity of responses of vertebrate neural precursors to extrinsic signals, it seems likely that developing vertebrate nervous systems are complex mosaics of distinct precursor populations, distinguishable in terms of patterns of gene expression, as well as in terms of developmental potency.

Acknowledgements
The author thanks Arthur Lander for critical comments on the manuscript, and David Anderson and Jay Rothstein for sharing results prior to publication.
Work from the author's lab is supported by NIH grans NS32174 and 1X02180.

References and recommended reading
Papers of particular interest, published within the annual period oi review, have been highlighted as: .

13.
14. 18. Davis AA, Temple S: A self-renewing multipotential stem cell . in embryonic rat cerebral cortex. Nafure 1994, 372:263-266. Development of single cells isolated from embryonic day 12 and 14 rat cerebral cortex was followed in tissue culture. A small percentage (-7%) of cells could divide for long periods and give rise to neurons, astrocytes, and oligodendrocytes. Upon re-cloning, these putative stem-cell clones could again generate all three cell types, suggesting they had the property of self-renewal.

achaete-scute homolog (CASH-I)
is expressed in a temporally and spatially discrete manner in the developing nervous system. Development 1994, 120:769-783. Complete cDNA sequence for chicken achaere-scute homologue (CASH)-1 is reported. In situ hybridization analysis in retina suggests that CASH-l expression may define a stage in retinal precursor cell development.