A molecular switch for the neuron-glia developmental decision

The genetic analysis of nervous system development has revealed a’wealth of information about the molecular mechanisms that generate neural precursors and neuronal cell type diversity (Shankland and Macagno, 1992). Until recently, however, relatively little has been known about the genetic control of gliogenesis, in part because of the comparative lack of early markers specific for glial cellsor their immediate precursorsin organisms amenable to genetic analysis, such as Drosophila. One general rule that has emerged from lineage analysis of neurogenesis is that neurons and glia share a common precursor. This has been demonstrated in both the CNS and the PNS of both vertebrates and invertebrates (for reviews, see Anderson, 1969; McConnell, 1991; Doe and Technau, 1993; Jan and Jan, 1994). Since neurons and glia share a common precursor, there may be genes that control the decision between these alternative fates. Now, two papers from the Hotta and Goodman laboratories (Hosoya et al., 1995; Jones et al., 1995) report the identification of an important component of this mechanism. Specifically, they have isolated a gene, glial cells missing (gem), whose presence or absence determines the choice between neuronal and glial fates. The idea that the product of a single gene can determine a choice between neuronal and glial lineages is not unprecedented. For example, in the grasshopper CNS, the development of precursors of midline glia and their neuronal siblings can be controlled in a reciprocal manner by the activity of the homeoprotein engrailed (Condron et al., 1994) as well as by protein kinase A (Condron and Zinn, 1995). Furthermore, in the vertebrate PNS, the choice between neurbnal and glial fates by multipotent neural crest cells can be controlled by glial growth factor (GGF)/neuregulin, a growth factor in the epidermal growth factor/ transforming growth factor a superfamily. In the presence of GGFlneuregulin, cells that would otherwise have become neurons instead choose a glial fate (Shah et al., 1994). The mutation characterized by the Hotta and Goodman laboratories was first identified not by its glial phenotype but rather in the course of screens for mutations that affect axon pathfinding. gem homozygotes show a disruption of axons within the longitudinal connectives (Hosoya et al., 1995; Joneset al., 1995).Thesamemutation wasindependently identified in a P element screen for mutations affecting PNS development (Kania et al., 1995). Subsequent analysis of the gem expression pattern revealed a striking specific and transient expression in essentially all glial precursors in the embryonic CNS and PNS, with the exception of the CNS midline glia, a special population of Minireview