We have examined the consequences of surface interactions with glial and nonglial cells on the in vitro growth of CNS neurons. When cerebellar or spinal cord cells were plated onto monolayers highly enriched in cortical astrocytes or sciatic nerve Schwann cells, neurons generally grew as single cells and showed relatively little tendency to aggregate. Similarly, neurites showed little tendency to fasciculate. In contrast, when plated onto fibroblast, heart muscle-fibroblast, or astrocyte-free meningeal monolayers, neurons rapidly aggregated, and neurite outgrowth was primarily in large fascicles. There were no glia detectable in the majority of aggregates or fascicles, suggesting that aggregation and fasciculation were due to interactions between neurons. Neurite outgrowth over 24 hr was also greater on astrocytes than on nonglia. Whether or not aggregation and fasciculation occurred was due to surface properties of the glial and nonglial cells. When neurons were added to astrocyte and nonglial monolayers growing in medium conditioned by a large excess of co-cultured nonglia or astrocytes, respectively, the pattern of neuronal growth was determined by the type of monolayer with which the neurons were in contact. Moreover, the initial growth of neurons on heat-killed astrocytes was indistinguishable from growth on living astrocytes. The pattern of neuronal growth on these different monolayers suggests that neurons are more adherent to glia than to other neurons but are more adherent to other neurons than to nonglia. Such an adherence hierarchy could explain the consistent finding of an apposition of neurons to glial surfaces during neuronal migration and axon outgrowth. Our findings also suggest that the interaction of axons with the non-neuronal milieu through which they grow may play an important role in regulating fasciculation, a process which has generally been treated as due primarily to axon-axon interactions.

The neuroepithelial cells of the mammalian neural tube are thought to give rise to all classes of differentiated neurons and macroglial cells in the adult CNS. In most cases, the regulation and timing of commitment of neuroepithelial cells to specific differentiative pathways are unknown. It has been proposed that in developing spinal cord, the macroglial cells--astrocytes and oligodendrocytes--arise either by the direct transformation of radial glial cells in the developing cord or, alternatively, by the differentiation of distinct precursor cells which migrate to presumptive white matter from the region of the central canal during development. In this study, the timing of oligodendrocyte differentiation in different levels of the spinal cord and the capacity of specific regions of the spinal cord to give rise to oligodendrocytes at various ages was tested in vitro. At embryonic day 14, all complete segments, as well as all ventral regions along the rostral-caudal axis of the spinal cord, have the capacity for oligodendrogenesis. By contrast, dorsal regions of the thoracic and lumbar spinal cord do not develop the capacity for oligodendrogenesis until later in development. The capacity of dorsal rat spinal cord to give rise to oligodendrocytes appears to be associated with the ventral-to-dorsal migration of oligodendrocyte precursors. These observations suggest that commitment to an oligodendrocyte differentiative pathway appears to occur in a distinct population of ventrally located glial precursors in the embryonic rat spinal cord.

No abstract