Development of glial cells in the cerebral wall of ferrets: direct tracing of their transformation from radial glia into astrocytes.

Coronal sections of the cerebral wall from developing ferrets (newborn to adult) were double-stained with antibodies to vimentin and glial fibrillary acidic protein (GFAP). At birth, the dominant glial population was radial glia and these cells labeled only for vimentin. A small population of immature astrocytes in the cortical plate was double labeled for GFAP and vimentin. In successive days, the number of vimentin-positive radial glia gradually decreased and they disappeared entirely at about 21 days. During this same period, the double-stained astrocytes increased in number and were distributed throughout the cortical plate and intermediate zone. After 6 weeks of age the astrocytes were mostly confined to the developing white matter. Around this time they gradually lost their vimentin staining, and in the adult no vimentin-positive elements were seen except at the ependymal surface. In newborn ferrets single radial glial cells were also visualized by applying the carbocyanine dye DiI onto the pial surface of fixed brains. While most radial glia extended from the ventricular zone to the pial surface, a substantial fraction of them had lost their contact to the ventricular zone. Their somata were displaced into the subventricular zone and lower portion of the intermediate zone. The possibility that radial glia transform into astrocytes was directly tested by injecting fluorescent dyes under the pial surface of newborn ferrets at a time when virtually no GFAP-positive astrocytes are present. The tracer, which was taken up in the upper portion of the cortical plate, stained the radial glial cell somata in the ventricular zone in a similar way as the dye DiI did in the fixed brains. As the radial glial cells disappeared at successively longer survival times, the tracer was ultimately found within newly formed GFAP-positive astrocytes. These results provide strong support for the hypothesis that radial glia cells are the immature form of astrocytes (Choi and Lapham: Brain Res. 148:295-311, '78; Schmechel and Rakic: Anat. Embryol. (Berl.) 156:115-152, '79), and also show that, at least in the ferret cortex, the transformation is accompanied by a change in the expression of intermediate filament protein.