Genes Involved in Drosophila Brain Development

Two recent publications provide further information about the specific roles of developmental genes in Drosophila brain development. In addition, both may shed light on evolutionary changes in specific molecular components of the regulatory cascade which forms the anterior posterior body axis. First, orthodenticle was shown to have distinct roles in the development of the brain and ventral nerve cord, which could be replaced by overexpression of homologous human genes (Leuzinger et. al, 1998). The second demonstrated that the homeotic selector genes labial and Deformed seem to be required for the step in nervous system development which specifies the neuronal identity of cells in two specific regions of the Drosophila brain (Hirth et. al, 1998)

Orthodenticle-dependent development of the anterior nervous system

Drosophila null mutants for orthodenticle lack the anterior region of the brain. In the mouse, knockout of either of the orthodenticle homologues, hox genes OTX1 and Otx2, results in similarly abnormal brain structure. Specifically, OTX1 null mutation leads to both epilepsy and the loss of posterior brain structures, including the telencephalon, mesencephalon, and cerebellum. The Otx2 null mutation results in loss of the forebrain and midbrain. Based on these similarities, Leuzinger et. al proposed the hypothesis that orthodenticle and the Otx genes have function which was conserved by evolution. (Leuzinger et. al, 1998)

Leuzinger's experimental strategy

The laboratories collaborating in these experiments prepared strains of transgenic flies with a heat shock promoter-controlled gene for either Drosophila otd or human OTX1, or OTX2. These genes were introduced into either wild type flies or flies with the orthodenticle null mutant background. Heat shock was applied to the mutant flies to determine whether ubiquitous overexpression of either orthodenticle, OTX1, or OTX2 could replace the function of the missing orthodenticle gene in either the Drosophila brain or the ventral nerve cord. Brain rescue was evaluated on the basis of the presence of the anterior protocerebral cells which normally express the brain specific homeobox gene, and observation of total neural structure present. Ventral nerve cord rescue was evaluated on the basis of separation of commissures and connectives, and on the basis of midline position of engrailed-expressing cells. (Leuzinger et. al, 1998)

The experimental design seems appealing in its apparent cleanness and straightforward organizational structure. However, ubiquitous overexpression of a gene is not necessarily a perfect proxy for the regulated expression of the genes in its normal developmental context. One possible future experiment to avoid this possible difficulty of interpretation might involve preparing transgenic flies with the orthodenticle null mutant background, and with OTX1 or OTX2 linked to the normal Drosophila orthodenticle promoter rather than to a heat shock promoter.

Rescue of orthodenticle mutants by otd, OTX1, and OTX2

The data suggest that otd, OTX1, and OTX2 can all rescue the orthodenticle null mutant phenotype. This said, the data command attention to the differential efficacy of these genes in rescue and the distinct roles of these genes in development of the ventral nerve cord and the brain. First, orthodenticle linked to the heat shock promoter was imperfectly successful at restoring normal brain morphology. In the laser confocal microscopy images of the otd-rescued Drosophila brain Leuzinger presented, the structure differs from the wild type control, but, on the other hand, the massive deletion which results from the orthodenticle mutation was successfully ameliorated. Overall, otd restoring the deleted regions, the protocerebral anlage and preoral commissure, in 57.6% of the mutants tested. OTX1 and OTX2 were comparatively less effective, acheiving rescue in 22.6% and 45.5% of the mutants tested, respectively. In the ventral nerve cord, orthodenticle achieved restoration of the malformed structure in 100% of the mutants tested, while OTX1 and OTX2 were 94.0% and 100.0% successful, respectively. Additionally, the timing requirements differed between the two brain areas; the deleted brain structure could be restored only if orthodenticle or homologue was induced by stage 7-8 of development, while the ventral nerve cord defects could be remedied as late as stage 10-11. (Leuzinger et. al, 1998)

Analysis of the results

These data provide additional information about the role of orthodenticle in brain development by indicating that the gene has distinct roles in the development of the ventral nerve cord and the brain. Ventral nerve cord malformation and brain malformation differed qualitatively, the former consistent with improper cell patterning or location and the latter with deletion of the affected structures. Also, the two areas were distinct with respect to both the timing requirements for orthodenticle expression and the efficiency of rescue by orthodenticle. This leads to the interpretation that the distinct milieu of regulatory processes operative within these two areas of the Drosophila embryo affects orthodenticle's behavior. While this interpretation is necessarily vague, it does provide a basis for several further questions about orthodenticle's action within the nervous system. For example, the possibility that the specific time phase of orthodenticle expression has a critical role in its regulatory function represents a particularly interesting question, one which probably would benefit from an experimental methodology which permits more specific control of the timing and location of gene expression.

In addition, Leuzinger et. al (1998) present the surprising result that orthodenticle's activity seems to be replaceable by ubiquitous overexpression of the human genes OTX1 or OTX2. While the sixty amino acid homeodomain sequences of orthodenticle, OTX1, and OTX2 are homologous, the human and Drosophila proteins are dissimilar elsewhere. Consequently, Leuzinger's data (1998) apparently argue that it is the homeodomain region of the protein which has the critical role in determining the large-scale morphological characteristics which were altered by the orthodenticle mutation. However, these data must also be interpreted within the limits of their method; the possibility remains open that the non-homeodomain region of orthodenticle has a role in its regulatory functions. The method of ubiquitous, heat shock controlled overexpression renders it difficult to make conclusive interpretations about the comparative success of orthodenticle and OTX1 or OTX2 in restoring nervous system structure. In particular, the incomplete efficiency and accuracy of the orthodenticle-mediated rescue of the orthodenticle mutant caution against considering these data as presenting a complete picture of orthodenticle's developmental functions in the nervous system.

Homeotic genes in Drosophila brain development

In the Drosophila trunk, homeotic genes are activated subsequent to the pair rule genes, gap genes, such as otd, and segmentation polarity genes. The homeotic genes, which encode homeodomain DNA-binding proteins, seem to have the function of specifying cell identity. This may be seen in observations that the homeotic genes regulate the expression both of other regulatory genes, and of genes which are unique to specific cell types in the adult organism. (Gilbert, 1997)

While the role of homeotic genes in the trunk is understood to some extent, little direct information was available about the function of homeotic genes in the development of the Drosophila brain. In order to approach this topic, Hirth and colleagues systematically examined the expression and mutant phenotypes of these genes in the Drosophila brain. (Hirth et. al, 1998)

Experimental Design

Hirth and colleagues (1998) first immunohistochemically identified the locus of expression of each of the homeotic genes in the Drosophila brain. Next, examined the phenotype of the Drosophila brain in knockouts of each of the genes in order to identify possible gene functions. When mutants which had defects visible at the scale of tissue organization were identified, Hirth proceeded to characterize the specific abnormality resulting from the mutation. (Hirth et. al, 1998)


Immunohistochemical analysis of homeotic gene expression demonstrated that each of the genes was expressed in a localized region of the Drosophila brain which corresponded to their linear order of position on the Drosophila chromosome except for the transposition of labial and proboscipedia. (Hirth et. al, 1998)

On this basis, mutant analysis was implemented. It was determined that only labial and Deformed had phenotypic effects on the scale of abnormal tissue structure. Both mutant phenotypes were similar in terms of the observed defect morphology. The defect seemed to consist of abnormal axonal structure, at the level of deleted commissures which connect the hemispheres of the Drosophila brain, and of abnormal axonal structure. In addition, both mutant phenotypes were localized to the site of the gene's expression; the labial null mutation affected the tritocerebral neuromere, while the Deformed null mutation affected the mandibular neuromere. (Hirth et. al, 1998)

Figure 1
Figure 1   Co-staining of lab-lacZ transgenic Drosophila brain with B-gal (green) and anti-HRP (red) indicates the tritocerebral neurons in the area of the commissure in the wild-type, left. In the labial mutant, right, the red-stained region which overlaps the area of the B-galactosidase-expressing cells is absent, while the B-galactosidase expressing cells are present.

After identifying the lab and Dfd defects as problems of axonal patterning, Hirth implemented a series of experiments to define the specific defects which resulted from the mutation. One possibile explanation for the defective axonal structure was the loss of the cells that normally project these axons. This was tested by analyzing the CNS phenotype of transgenic Drosophila with the labial null mutation and with a lac Z reporter construct controlled by the labial promoter. This experiment demonstrated that the cells which normally express labial were not deleted in the null mutant, as B-galactosidase expressing cells were observed in the position normally occupied by labial-expressing cells. The alternative possibility, that the null mutant led to cell death, would have resulted in the loss of B-galactosidase expression concomitant with absent labial expression. (Hirth et. al, 1998)

Figure 2
Figure 2   Co-staining of the Drosophila tritocerebral area with B-gal (red) and anti-tubulin (green). In the wild type, pictured above left, the staining indicates that the labial-expressing cells of the tritocerebral domain give rise to the axons which compose the tritocerebral commissure. In the labial mutant, above right, the red cells which normally express labial are present, while the anti-tubulin staining structure (indicated by the arrows) is absent, suggesting the loss of projecting axons.

Given that the cells were still present, the next experiment characterized the axonal patterning defect which resulted from the labial mutation. Staining the lab-lac Z reporter flies with anti-tubulin allowed analysis of the relationship between labial expression and axon projection. Hirth showed that in the wild type, the labial expressing cells, which also express B-galactosidase, extend the axons which normally compose the tritocerebral commissure by immunohistochemistry for tubulin, a cytoskeletal protein expressed along the length of axons. However, in the labial null mutant, the lab-lac Z cells give rise neither to the commissural axons nor to any other axonal projections. In addition to this cell-autonomous defect, Hirth determined that the axons of cells from other areas which normally project to the area of labial expression acquire abnormal structure in the labial null mutant, either stopping at the border of the region or projecting around it. (Hirth et. al, 1998)

Finally, Hirth tested the hypothesis that the cells of the brain affected by the labial mutation do not differentiate into neurons using immunohistochemistry for a variety of proteins known to be neuron-specific. Cells were found to express the proteins encoded by the head gene hunchback and the ganglion mother cell gene prospero. However, labial mutant flies lost expression of the RNA-binding protein ELAV, and the cell adhesion molecule Fasciclin II. This data suggested that the step which specifies neuronal cell fate was dysfunctional in the labial mutant strain. (Hirth et. al, 1998)

In addition to immunohistochemically analyzing cell morphology, Hirth characterized the effect of the labial mutation on the expression of other genes. It was determined that in the lab mutant domain, empty spiracles was activated while proboscipedia was inactivated, indicating that labial has a role in regulating the expression of other regulatory genes. In addition, heat-shock induced overexpression of labial resulted in ectopic Dfd expression. These data are consistent with posteriorizing gene expression in the lab-overexpressing condition and anteriorizing gene expression in the lab-mutant condition. However, the cell-structure data indicate that this change in gene expression is not adequate to transform cell identity, likely implying that interaction between the anteriorly expressed genes affected by lab and other genes unaffected by the lab mutation is involved in specification of cell identity. (Hirth et. al, 1998)

Summary of Hirth's results

  • Homeotic genes are expressed in the Drosophila head.
  • Mutation of the homeotic gene labial results in abnormal organization of the tritocerebral lobe, while mutation of the homeotic gene Dfd results in abnormal mandibular lobe organization.
  • Mutation and overexpression of labial result in altered homeotic gene expression in the tritocerebrum.
  • Mutation of lab does not lead to the death of cells in the tritocerebral lobe.
  • Cells in the tritocerebrum of lab null mutant Drosophila fail to project axons, while axons from other areas do not project normally into or through the tritocerebrum.
  • Cells in the tritocerebrum of lab mutant flies are not specified as neurons.
  • The Dfd mutation results in failure of neronal specification localized to the mandibular lobe.

Analysis: Questions raised by Hirth's findings

From the point of view of homeotic gene expression and function, Hirth et. al (1998) have demonstrated that the homeotic genes are active in locally defined regions of the head, but that only mutation of labial and Deformed has a phenotypic effect which could be characterized on the basis of brain morphology. It would be interesting to learn whether the individual mutations of the other homeotic genes alter expression of other regulatory genes in a manner analogous to the labial mutation, since it might be possible that this type of mutation has a cellular effect not observable on the scale of this data. In the same line, it would be useful to identify the targets of regulation by the homeotic genes expressed within each of the areas of the Drosophila brain. This might permit discrimination of developmental differences within different categories of nerve cell within the brain, as the homeotic gene expression pattern seems to provide distinct groups of cells along the anterior-posterior axis with distinct regulatory environments. In addition, information about the regulatory targets of the homeotic genes might be useful for beginning to compare the mechanisms which regulate neuronal cell fate in distinct areas of the Drosophila brain. Clearly, they differ to the extent that cell fate does not universally depend on homeotic gene expression patterns. The details of the process which intervenes between regulated homeotic gene expression and neuronal identity, and the results in the differences between each area of the brain at the level of this process remain to be studied.

Evolutionary Possibilities: How much is a mouse like a fly?

The results presented by both Hirth et. al (1998) and Leuzinger et. al (1998) include information which indicates certain similarities in the regulatory mechanisms which occur during the development of Drosophila and vertebrates. Leuzinger (1998) showed that human OTX1 or 2 could effect rescue of the orthodenticle null mutant phenotype, while Hirth (1998) demonstrated that the linear order of expression of homeotic genes in the Drosophila brain matches the order of expression of the homeotic homologues, hox genes, in the mouse brain.

The similarity that Leuzinger (1998) identified seemed to be located primarily at the level of the conservation of the structure of a regulatory protein. The events which lead to regulated expression of orthodenticle differ from those which lead to OTX expression. This follows from the observation, among others, that Drosophila axis determination takes place in a syncytium, while vertebrate embryonic organization relies on intercellular regulatory events. However, Leuzinger's (1998) observation that inserting a foreign, but similar, gene into the regulatory environment established during Drosophila development indicates that Drosophila and vertebrates are probably similar at the level the of the molecular events involved in the regulation of gene transcription. In addition, Leuzinger (1998) has shown that the regulatory targets of this process are common to the extent that orthodenticle has a role in brain development which was at least partially conserved in evolution. This data provides compelling evidence for the utility of Drosophila as model for investigating the genetic regulation of brain development in a way which might be informative for understanding, or for posing hypotheses about, the genetic regulation of mouse or human brain development. In addition, this initial identification between the similarities between the two genes provides a position which might be useful for beginning to investigate the downstream targets of these regulatory genes. For example, learning about the downstream targets of orthodenticle in the Drosophila brain might present a valuable starting point for identifying genes which might provide interesting knockout phenotypes in mice. While the targets of regulation are likely divergent in humans and flies, it is possible that learning about the points of similarity could lead to experimental analysis and understanding of the distinctions between the events which regulate the formation of the vertebrate brain and the chordate brain.

The finding in Hirth et. al's (1998) article that the order of homeotic gene expression in the Drosophila brain matches the order of hox gene expression in the mammalian brain represents a different sort of similarity. This evidence does not speak directly to the detailed molecular interactions involved in brain development, but indicates that the genes which identify position along the anterior-posterior axis seem to be the same. It is also interesting that this close match exists at the homeotic/hox level in particular, given their status as selectors of cell identity. Upstream genes, which lead to formation of localized areas of gene expression along the axis would seem to need to differ due to the differences inherent in development according to position in a single cytoplasm versus development according to interaction with neighboring cells. Downstream genes, which lead to cell specification and the detailed acquisition of cellular identity and adult tissue organization, probably differ in a way which reflects the distinct structures and computational tasks of the vertebrate and invertebrate brains. However, the developmental point at which cells are identified according to position along the axis,but have not yet acquired identity, would seem to be one amenable to retaining similarity between distinct organisms. Despite this similarity in the order of expression, Hirth's data also demonstrate a clear difference in that single homeotic genes are expressed in areas of the developing Drosophila brain, whereas multiple hox genes are expressed in the mammalian brain. This divergence seems consistent, however, with the greater degree of complexity of the regulatory tasks which are involved in the organization of the human or mouse brain by comparison to the brain of Drosophila.

Thus, considered from an evolutionary perspective, this data regarding brain development may provide examples of conserved processes at the level of body plan organization. Orthodenticle and its homologs act to regulate brain formation by acting on its targets using the same DNA-binding domain, but apparently with different phenotypic outcomes in the distinct regulatory and cellular environments of the fly brain and the mouse or human brain. The homeotic genes seem to provide information about cell position along the anterior-posterior axis in Drosophila and mice, but the effect of this information seems to differ between organisms. Together, this evidence suggests that both the targets of regulation of genes involved in brain development and the mechanisms which lead to formation of regulated gene expression along the body axis would both be interesting areas for further research.

Literature Cited

Gilbert, S. F. (1997) Developmental Biology (5th ed.) Sinauer Associates, Inc., Sunderland.

Hirth, F., Hartmann, B., and Reichert, H. (1998) Homeotic gene action in embryonic brain development in Drosophila. Development 125: 1579-1589.

Leuzinger, S., Hirth, F., Gerlich, D., Acampora, D., Simeone, A., Gehring, W. J., Finkelstein, R., Furukubo-Tokunaga, K., and Reichert, H. (1998) Equivalence of the fly orthodenticle gene and the human OTX genes in embryonic brain development of Drosophila. Development 125: 1703-1710.

© All the material on this website is protected by copyright. It may not be reproduced in any form without permission from the copyright holder.