Discussion

Glial Cells in the Optic Lobe - Discussion

Help SiteNavigator Index Where Am I? Search Page Status

E. Eule, S. Tix (1), and K.-F. Fischbach
Institut für Biologie III, Schänzlestrasse 1, 79104 Freiburg i. Br., Germany. (1) New address: California Institute of Technology, Pasadena, California 91125, USA

Discussion

Glial Cells in the Optic Lobe fall into genetically diverse classes Previously, glial cells in the adult optic lobe of insects have been classified according to their shape and position. The enhancer trap technique yields additional information: it allows in principle the detection of genetic heterogeneity in a morphologically homogeneous cell population, and it provides an estimate of the degree of genetic equivalence of positionally and morphologically diverse cell populations. Our initial pool of 87 enhancer trap lines was preselected because of lacZ expression in the embryonic nervous system (Christian Klämbt, personal communication). Of these, 6 lines with reporter gene expression in glial cells of the adult CNS were used for detailed analysis (Table III). With one exception (3-159), the lines studied label several neuronal and glial cell classes, and most cell types were detected by more than one enhancer trap line. Interestingly, the giant glial cells of the inner and outer chiasm were always labelled together indicating similarity in their state of differentiation. Some glial cell types were detected in only one enhancer trap line: the small outer chiasm glial cells showed up in line T1184, and the lobula plate satellite glia was uncovered in line 3-74. In the latter, many other cell types were labelled as well. In contrast, the reporter gene expression in line 3-159 solely and specifically marks the medulla neuropile glial cells in the adult optic system.

After Trujillo Cenoz (1972) the medulla neuropile glia of insects is situated at the interface between the medulla cortex and the neuropile. The cell bodies marked in our lines lie in the most distal region of the neuropile. Interestingly, reporter gene expression of line 3-159 starts late in pupal development at 8085 h apf, i.e. well after the development of the basic structural design of the adult optic ganglia. This late onset of lacZ expression indicates a late maturation of these glial cells and a major function in the adult medulla neuropile. In this context, the dependence of their number on neuropil volume (see below) may suggest a mainly throphic function, and their genetic uniqueness may reflect special needs of medullar neurons.

Three glial cell types, namely the lobula neuropile glia (LoNGl), the lobula plate neuropile glia (LPNGl) and lobula plate satellite glia (LPSaGl) did not show up in any of our enhancer trap lines and were identified in osmium fixed specimen only. Some glial cell types of the optic lobe may still be undetected.

The perineurial cells ensheath the entire central nervous system and constitute the bloodbrain barrier. In our study, no enhancer trap line could be identified that selectively detects perineurial glial cells of the optic lobe. Therefore, it is likely that these cells are part of a homogeneous population of total CNS perineurial cells. The enhancer trap lines also did not establish any distinction between subperineurial glial cells in the optic lobe or elsewhere in the CNS.

Tracheal cells in the optic lobe were not only always labelled together with tracheal cells elsewhere in the CNS, but even always together with tracheal cells outside the nervous system. This indicates that these cells constitute a homogeneous cell population irrespective of location. For this reason it might be better not to categorize tracheal cells as "glia". Similarly, in vertebrates, components of blood vessels are not considered to be glial cells.

Taken together, our observations show that glial cells fall into genetically distinct classes. Our genetic definition of glial cell populations agrees well with previous morphological classification schemes, e.g. all the glial cell types in the lamina detected in our study, can easily be homologized with the classes described in Musca (SaintMarie & Carlson1983).

Number and position of certain glial cells may vary

It was shown by Giangrande et al. (1993) that number and position of glial cells in the wing imaginal disc is variable. We present evidence that certain glial cells in the optic lobe as well may not display a stereotyped arrangement. We counted the cell number of medulla neuropile glia and of inner chiasm giant glia in wild type and mutant whole mounts. Even in wild type we found significant variability (Table II). The following arguments are in favour of the assumption that this variability is not due to experimental artefacts: First, counting errors can be excluded as the stained nuclei are large and easy to detect. Second, there is no large variability of staining between individual cells; all cells of a population are similarly labelled. Therefore, it is unlikely that some members of a population were not detected.

The position of glial cells also does not seem to be very stereotyped. In Fig.9 it is shown that the medulla neuropile glial cells of one focal plane are not arranged in a very stereotyped way. This seems to contrast somewhat with the situation in the lamina, where the arrangement of the epithelial glial cells seems to be fairly fixed. One possible explanation for variability may be given by the ability of glial cells to migrate to their final destination (Giangrande et al. 1993, Choi and Benzer 1994). With regard to this, the apparently randomized arrangement of inner chiasm giant glia in irre C mutants (Fig. 11), deserves, however, a second look. It reflects the disruption of the inner chiasm and the existence of ectopic fibre bundles. The glial cells are still closely associated with those. It will be intriguing to analysis the factors that regulate number and position of glial cells. As a first step we analysed the glial constitution of several structural brain mutants.

Factors regulating the size of different glial cell populations

Winberg et al. (1992) showed that glial cells in the developing lamina need retinal input to complete differentiation. Only glial cells in the neighborhood of ingrowing axons fully differentiate. It is, therefore, not surprising that disco mutants of the unconnected phenotype without a connection of retinal fibres with the optic lobe lack besides all neuronal also all glial constituents of the lamina.

We have selected medulla neuropile glia and inner chiasm giant glia for a closer investigation of factors regulating the number of glial cells. We took advantage of the fact that several structural brain mutants are known that reduce the volume of the medulla, but not the number of visual columns (Fischbach and Heisenberg, 1981, 1985). The known mutants reduce the medulla volume differently (Fischbach et al. 1989). If the number of cells in both glial populations are plotted as a function of medulla neuropile volume it is apparent that they respond quite differently. The number of medulla neuropile glia is strongly dependent on the medulla volume (Fig.10a) and clearly not correlated with the number of visual columns, as this number in wild type and solEE111 flies is the same. In disco flies with the unconnected phenotype, medulla volume and the number of visual columns are severely, or respectively, completely reduced. Nevertheless, a few medulla neuropile glial cells could be detected (Fig. 12b). It is therefore concluded that the number of medulla neuropile glial cells is regulated by the size of the medulla neuropile. It may well reflect the number of neurons that have to be supplied.

On the other hand, the number of inner chiasm giant glial cells does not display a strong dependence on the medulla volume as long as the number of visual columns remains normal (Fig.10b). Although the fibre bundles connecting medulla and lobula complex are much thinner than normal, their number and that of associated inner chiasm giant glial cells is retained. This is true even for the irreC mutants with ectopic fibre bundles (Boschert et al. 1990). This, however, is reflected in a displacement of the glial cells (Fig. 11a) rather than in a reduction of number.

Giant glial cells separate axonal bundles of different horizontal planes in the outer and inner chiasms

Three types of chiasm glial cells have been characterized: the inner and outer chiasm giant glial cells and the small outer chiasm glial cells (the small inner chiasm glial cells may have remained undetected). All chiasm glial cells are arranged from dorsal to ventral in a crescent, stuck inbetween the crossings of axonal fiber bundles. The giant glial cells of the outer and inner chiasm bear homologous characteristics with respect to their nuclear shape and position as well as to their genetic identity. They are always labelled together. These findings seem to indicate a homologous function of both giant glial cell types. In mutants without a lamina, e.g. in unconnected disco lobes, the giant glial cells of the outer chiasma are absent, indicating that their develop ment or maintenance is dependent on the fibre bundles in the first optic chiasm. On the other hand, these giant cells may well be involved in the establishment of normal optic chiasms. Their strategic location inbetween horizontal fibre bundles makes them good starting points for a closer inspection of chiasm formation (Tix et al., in preparation).

Add comment to poster

Flybrain Poster Sessions Glial Cells in the Optic Lobe Discussion