glial cells missing
gcm is initially expressed in a patch 16 cells wide by 12 cells in length in the anterior ventral region of the cellular blastoderm (Jones, 1995). gcm is expressed in progeny of specific neuroblasts in the CNS and in the brain during stages 11 and 12. These cells include glioblast precursors of longitudinal glia, exit glia, glia asssociated with peripheral axons, and glia associated with sensory organs (Jones, 1995). There is no gcm expression in midline glial cells. gcm is also expressed in the most anterior region of the presumptive mesoderm (Hosoya, 1995).
Hemocytes are cells of mesodermal origin that disperse along migratory pathways in the embryo. The first stage at which they can be identified is late stage 10 in a subpopulation of cells located in the head of the embryo. During the first wave of apoptosis, normally occurring during midembryogenesis, some of them convert to macrophages that engulf and degrade cells undergoing programmed cell death. Glial cell differentiation in Drosophila melanogaster requires the activity of glide (glial cell
deficient), also known as glial cells missing (gcm). The role of this gene is to direct the cell fate switch between neurons and
glial cells by activating the glial developmental program in multipotent precursor cells of the nervous
system. glide/gcm is also expressed and required in the lineage of
hemocytes/macrophages, scavenger cells that phagocytose cells undergoing programmed cell death. The earliest glide/gcm expression in the hemocyte lineage can be detected in the head region at the end of the blastoderm stage. By stage 11, glide/gcm expression in the hemocyte lineage decreases, while its expression in glial cells of the peripheral and central nervous systems becomes evident. For glial cells, glide/gcm plays an instructive role in hemocyte differentiation.
Interestingly, it has been shown that in the development of the fly adult nervous system the role of
scavenger cells is played by glial cells. glide/gcm expression in the hemocyte lineage requires serpent. These data and and evidence regarding on the dual role of glide/gcm
indicate that glial cells and hemocytes/macrophages are functionally and molecularly related (Bernardoni, 1997).
Neurons and glia are produced in stereotyped patterns after neuroblast cell division during development of the Drosophila
central nervous system. The first cell division of thoracic neuroblast 6-4 (NB6-4T) is asymmetric, giving rise to a glial
precursor cell and a neuronal precursor cell. In contrast, abdominal NB6-4 (NB6-4A) divides symmetrically to produce two
glial cells. To understand the relationship between cell division and glia-neuron cell fate determination, the effects of known cell division mutations on the NB6-4T and NB6-4A lineages were examined and
compared. Based on observation of
expression of glial fate determination and early glial differentiation genes, the onset of glial differentiation occurs in
NB6-4A but not in NB6-4T when both cell cycle progression and cytokinesis are genetically arrested. In contrast,
glial differentiation starts in both lineages when cytokinesis is blocked with intact cell cycle progression. These results
show that NB6-4T, but not NB6-4A, requires cell cycle progression for acquisition of glial fate, suggesting that distinct
mechanisms trigger glial differentiation in the different lineages (Akiyama-Oda, 2000).
Cell division
mutants stg, cycA, and pbl were used to
investigate the relationship between cell division and glia-
neuron cell fate determination in NB6-4.
To determine the effects of cell cycle arrest on cell fate of
NB6-4, expression of a glial fate determination protein, Gcm, and an early glial marker protein, Repo, were examined in stg mutant embryos. In normal development of
NB6-4T, these proteins are detectable in the medial daughter
cell after the first cell division and in its progeny cells, while they are expressed in both
daughter cells of NB6-4A. In stg mutant embryos, neither
Gcm nor Repo is detectable in NB6-4T, whereas both proteins are expressed in NB6-4A
at levels comparable to those of wild-type embryos. This indicates that stg activity is required
for the onset of glial differentiation in neuroglioblast NB6-
4T, but not in glioblast NB6-4A. GCM mRNA was examined to determine whether
the lack of Gcm expression in stg mutant NB6-4T resulted
from loss of transcription or failure of posttranscriptional
regulation. In contrast to wild-type embryos,
in which GCM mRNA is detected in both the NB6-4T and
the NB6-4A lineages, GCM mRNA is detected in NB6-4A, but not in
NB6-4T of stg mutant embryos. This
indicates that transcription of gcm in NB6-4T does not
occur or occurs at only a very low level in the mutant
embryos (Akiyama-Oda, 2000).
There are at least two
possibilities that explain how stg mutation affects the onset
of glial differentiation in NB6-4T: (1) that phosphatase
activity of Stg protein is directly required for Gcm expression
and (2) that stg-mediated cell cycle progression
is needed for function and/or distribution of some
factors that are necessary for Gcm expression. The latter possibility is favored, since not all the stg-induced cells
expressed the glial proteins in the rescue experiments. The normal expression of glial markers in
cycA mutant embryos, in which the first cell
division is normal, is consistent with the notion that
the first cell division is a critical point for the onset of glial
differentiation in the NB6-4T lineage (Akiyama-Oda, 2000).
In pbl mutant embryos, which lack cytokinesis, all the
nuclei of NB6-4T express the glial proteins,
suggesting that cytokinesis is not required for the onset of
glial differentiation in the NB6-4T lineage. It has been
suggested that cytokinesis may be required for negative
regulation of glial differentiation, since more than three
nuclei in the pbl mutant NB6-4T, in contrast to three glial cells
in the wild-type, express Repo.
In wild-type embryos, expression of Gcm protein becomes
prominent in one of the daughter cells shortly after
the first cell division of NB6-4T.
Gcm protein is not detected in NB6-4T of the stg mutant,
while beta-gal is detected, although rather weakly, in the
cell of stg mutants bearing gcmp-lacZ (gcm promoter regulating lacZ expression). This
indicates that the gcm promoter may be initially activated
even when cell cycle progression is arrested by stg mutation (Akiyama-Oda, 2000).
Cell cycle progression of NB6-4T appears to be more
closely related to up-regulation of Gcm expression. This
regulatory mechanism may lead to a sufficient level of Gcm
expression, which enables transcription of downstream
glia-specific genes including repo. GCM
mRNA is present from before the first cell division of
NB6-4T in wild-type embryos. The level of Gcm protein in
the NB6-4T lineage is possibly controlled by several steps of
regulation, including transcription, stability of mRNA, and
translation. The findings of this study suggest that such
regulatory mechanisms involve stg-mediated cell cycle progression (Akiyama-Oda, 2000).
NB6-4T and NB6-4A are the corresponding cells in thoracic
and abdominal segments that share expression of
some marker genes, including en and eg. However, these cells show distinct patterns of proliferation
and cell fate. NB6-4T produces three glial cells
and four to six neuronal cells, while NB6-4A produces only
two glial cells. Despite this difference,
glial fate arises from both lineages. The analyses using stg
mutant embryos reveals that the effects of the mutation
on glial differentiation in the NB6-4T and NB6-4A lineages
are distinct. In the mutant, expression of Gcm and Repo
is detected in NB6-4A but not in NB6-4T. This indicates
that the start of glial differentiation in NB6-4T is dependent
on stg-mediated cell cycle progression but that in NB6-4A this is not the case (Akiyama-Oda, 2000).
This raises the question of whether the different regulatory
mechanisms for glial differentiation in these cells are reflected
by their distinct cell types: neuroglioblast and glioblast. In the
other glioblast GP, Gcm and Repo expression are detected in
the stg mutant, indicating that glial differentiation
in the glioblasts is independent of cell cycle progression.
In addition to these glioblasts, a few cells expressing
the glial proteins are observed in the mutant. These cells
might have been neuroglioblasts, suggesting that there may be
another mechanism to control the onset of glial differentiation
in neuroglioblast lineages (Akiyama-Oda, 2000).
In the NB6-4T lineage, the first cell division is a critical
step for triggering glial differentiation. Coincident with the
onset of glial differentiation is the occurence of cell fate bifurcation.
In the cell division rescue experiments using eg-GAL4, Gcm-positive and Gcm-negative cells appear after the
first cell division, although surrounding cells are still stg
mutant. The cell fate bifurcation is probably regulated cell
intrinsically and coupled to cell division. In contrast, all the
nuclei of NB6-4T in the pbl mutant express Gcm and
Repo. This may be because these proteins contain
the nuclear localization signal that enables them to enter
the nuclei within the single cell after translation, even if
asymmetry might initially appear within the cell (Akiyama-Oda, 2000).
During cell division of NBs, the transcription factor
Prospero is asymmetrically segregated to ganglion mother
cells, in which this protein has a role in specification
of cell identity. The first cell division of
NB6-4T shows some similarity to such NB division, since
the transcription factor Gcm is expressed preferentially in
one daughter cell after cell division to start sublineage-specific
differentiation. There may be a repressor and/or an
activator of Gcm expression, which should be segregated to
or expressed only in the neuronal daughter cell and the glial
daughter cell (Akiyama-Oda, 2000).
Asymmetric cell divisions and segregation of fate determinants are crucial events in the generation of cell diversity. Fly neuroblasts, the precursors that self-reproduce and generate neurons, represent a clear example of asymmetrically dividing cells. Less is known about how neurons and glial cells are generated by multipotent precursors. Flies provide the ideal model system to study this process. Indeed, neuroglioblasts (NGBs) can be specifically identified and have been shown to require the gcm fate determinant to produce glial cells, which otherwise would become neurons. The division of a specific NGB (NGB6-4T), which produces a neuroblast (NB) and a glioblast (GB), has been followed. To generate the glioblast, gcm RNA becomes progressively unequally distributed during NGB division and preferentially segregates. Subsequently, a GB-specific factor is required to maintain gcm expression. Both processes are necessary for gliogenesis,
showing that the glial vs. neuronal fate choice is a two-step process. This feature, together with gcm subcellular RNA distribution and the behavior of the NGB mitotic apparatus identify a novel type of division generating cell diversity (Ragone, 2001).
The study of NGB6-4T has allowed a novel type of asymmetric division to be identified with respect to several features. Neuroblasts produce cells with different size and fate. NGB daughter cells show minor differences in size. Thus, the two features can be
uncoupled. In neuroblasts, the mitotic spindle shows signs
of asymmetry by anaphase, while in the NGB6-4T it is already
asymmetric by prophase and stays asymmetric until the
end of division, as shown by mitotic spindle and centrosome
labeling. More importantly, asymmetry of the mitotic
spindle is always associated with different fates, irrespective
of the size of daughter cells. In the future, it will be important
to determine whether differences in mitotic spindle
size are the cause rather than the consequence of the asymmetry (Ragone, 2001).
Fate choices autonomously regulated require the acquisition
of an appropriate cell polarity. It is generally thought
that delamination from the epithelial layer provides neuroblasts
with an apico-basal polarity that ensures the generation
of different fates. However, NGB6-4T, which also delaminates from the epithelium and
produces different cell types, does not divide along the
apico-basal axis. It will be interesting to determine whether
the apically localized Bazooka protein, which
controls epithelial polarity, directs the orientation of neuroglioblast
6-4T, possibly in combination with specific positional cues, or whether different mechanisms control orientation of division according to the differentiative potentials.
NGB division also differs from other asymmetric divisions
described in the fly nervous system. In the sensory
organ lineage, the PI precursor divides along a planar
polarity and shows no asymmetry in the mitotic spindle; in contrast to this,
the PIIb precursor divides like a classic neuroblast. Interestingly, the PIIb
division, which produces the glial precursor of peripheral
nervous system, is at an oblique angle from the apicobasal
axis, much like that of the NGB. It will be important to
clarify whether the oblique angle is an important distinction (Ragone, 2001).
It is interesting to note that some features make the NGB multipotent precursor an intermediate between pure neuroblasts and epidermoblasts. Not only is the orientation of its division oblique, but also the newly duplicated centrosomes are always apical, whereas epidermoblast centrosomes are initially located basally and those of neuroblasts are apical or basal (Ragone, 2001).
The simultaneous utilization of different types of markers
labeling the mitotic apparatus, the cell fate determinants,
and the NGB6-4T lineage has revealed a new mode
of division in the nervous system. This type of division was
overlooked in the past due to the lack of adequate tools. The
NGB6-4T is one of the simplest neuroglioblasts to analyze,
because its lineage tree is already known and because it
produces neurons and glial cells just after delamination.
Other lineages, such as the 5-6, initially produce neurons
and only after some divisions give rise to glial cells (Ragone, 2001).
The establishment of different fates upon asymmetric
division relies on the unequal distribution of fate determinant(s). Prospero is expressed in the neuroblast and inherited in the GMC, where there is no de novo pros transcription. Thus, all the information is provided by the mother, i.e., the neuroblast. The initial asymmetry established in the mother must be sustained and amplified in the progeny. gcm RNA asymmetric distribution does result in the
preferential segregation to the GB. gcm expression, however,
is maintained and amplified in this cell through direct
autoregulation and the expression of cell-specific cofactors. Gcm accumulation eventually
activates the glial fate and represses the neuronal fate. The progressive activation of the glial fate is also confirmed by the finding that Repo, which contains
several binding sites for Gcm, is expressed in the GB
progeny but is almost undetectable in the GB itself (Ragone, 2001).
The present study shows that Pros transcription factor is
necessary to maintain gcm expression and thereby
activate the glial program in the glioblast. Indeed, in the
absence of Pros, gcm RNA progressively disappears
from the GB. The gain-of-function phenotype also demonstrates
that Pros is not sufficient to initiate gcm
expression nor to induce the glial fate on its own. Pros
protein and RNA most likely form a complex with Staufen
and Miranda. In the absence of Miranda, which is necessary
to localize them, both daughter cells inherit the RNA and the protein.
In addition, they both inherit gcm RNA. As a consequence,
the two daughter cells adopt the GB fate. It is speculated that Pros is required for both gcm-independent and gcm-dependent maintenance, since
in its absence the levels of gcm expression are even
lower than in the absence of autoregulation. The gcm
promoter contains, indeed, several binding sites for the Pros
protein. Finally, Pros does not affect all glial cells, therefore it is likely that
specific factors will be required in other lineages (Ragone, 2001).
gcm displays several differences with respect to
pros with regard to RNA localization. (1) asymmetric distribution is not evident before metaphase; (2) asymmetry occurs progressively during cell division rather than being sharply apical at interphase and basal at metaphase; (3) gcm
transcripts are present at the cortex and in the cytoplasm. These differences suggest the existence of different RNA localization pathways in asymmetrically dividing cells.
That Stau and Mira may participate to the process is
suggested by the mislocalization of gcm RNA in stau
and mira mutants. In addition, the gcm 3'UTR displays a
stem-loop secondary structure, a conformation that is necessary for the interaction
of Staufen with Bicoid 3'UTR. However, this mechanism is
not sufficient to ensure a correct fate choice. Moreover,
mira and stau are not fully penetrant with respect to
gcm RNA distribution. Finally, and more importantly,
the cytoplasmic localization of some gcm
transcripts as well as the kinetics of asymmetry calls for a
cortical microfilament independent mechanism. Thus, the
same RNA may be the target of two localization pathways: this complements the observation that the same RNA binding protein may localize transcripts using pathways with different cytoskeletal requirements (Ragone, 2001).
It is speculated that due to its features, gcm RNA
localization arises by differential degradation/stabilization,
as has been shown for Hsp83 mRNA. In the future, it will be important to analyze RNA asymmetric distribution in vivo by using transgenic flies that allow the NGB6-4T lineage to be followed simultaneously with the gcm RNA. It will also be interesting to analyze other lineages in order to determine whether pure neural precursors require
one RNA localizing pathway, whereas all mixed precursors require two. The presence of two pathways may act as a backup system, as suggested by the absence of a glial
phenotype in stau embryos, or may allow a fine-tuning of
RNA distribution. It is known that gcm dosage is
important to trigger the glial fate. The identification of the molecules acting in gcm RNA localization will be necessary to better understand the
glia versus neuron fate choice. Finally, it will be important
to define how the activity of these molecules is connected
with cell cycle and positional cues in the developing nervous system (Ragone, 2001).
In Drosophila, cell-fate determination of all neuroectoderm-derived glial cells depends on the transcription factor Glial cells missing (Gcm), which serves as a binary switch between the neuronal and glial cell fates. Because the expression of Gcm is restricted to the early phase of glial development, other factors must be responsible for the terminal differentiation of glial cells. Expression of three transcription factors, Reversed polarity, Tramtrack p69 (Ttk69) and PointedP1 (PntP1), is induced by Gcm in glial cells. Repo is a paired-like homeodomain protein, expressed exclusively in glial cells, and is required for the migration and differentiation of embryonic glial cells. To understand how Repo functions in glial terminal differentiation, the mechanism of gene regulation by Repo was analyzed. Repo is shown to act as a transcriptional activator through the CAATTA motif in glial cells, and three genes are defined whose expression in vivo depends on Repo function. In different types of glial cells, Repo can act alone, or cooperate with either Ttk69 or PntP1 to regulate different target genes. Coordination of target gene expression by these three transcription factors may contribute to the diversity of glial cell types. In addition to promoting glial differentiation, it was found that Repo is also necessary to suppress neuronal development, cooperating with Ttk69. It is proposed that Repo plays a key role in both glial development and diversification (Yuasa, 2003).
Although ectopic Repo induces the appearance of many non-glial
lacZ-expressing cells in the dorsal epidermis, cells within the CNS do not respond to ectopic Repo. In fact, even in the wild-type background, not all Repo-positive glia in the CNS expressed the ftz HDS reporter. This suggests that the mechanism by which Repo regulates transcription may be different in the CNS from the one for peripheral glia. One possible scenario is that the functions of Repo in the CNS require cooperation with one or more other factors, and that these interactions preclude Repo from acting through
the CAATTA motif. Ttk69 and PntP1 are good candidates for such co-factors,
because ttk and pointed are both required for the
development of CNS glial cells. Although repo, ttk and pointed are expressed in overlapping subsets of CNS glial cells, their expression is mutually independent; Repo continues to be expressed in the ttk or pointed mutant background, and lacZ expression levels in enhancer-trap lines of ttk or pointed are unaffected in repo mutant embryos. Moreover, ectopic expression of Repo in the entire neuroectoderm does not increase the expression of pointed P1 mRNA or
Ttk69, nor does ectopic expression of either Ttk69 or PntP1 affect Repo
expression. All three genes are most probably regulated
independently, downstream of the glial determinant Gcm (Yuasa, 2003).
Although glial specification by Gcm is well established, how the
characteristics of individual glial cells are determined is poorly understood. Gcm expression is confined to the early stage of glial development, suggesting that Gcm itself does not participate in the terminal differentiation of glia.
Moreover, Gcm also directs blood cell development; Gcm is expressed in
macrophage precursors and ectopic expression of Gcm in crystal cell precursors causes the transformation of crystal cells to macrophages. These results clearly show that the expression of Gcm does not always lead to the determination and terminal differentiation of glia. In glial cells, Gcm induces the expression of three transcription factors, Repo, Ttk69, and PntP1, and the
loss of these proteins causes abnormal glial development, although Gcm
expression remains normal. Although gcm can direct
repo expression in various contexts, repo is not expressed endogenously in blood cells, but is confined to
Gcm-positive glial cells, lasting even after gcm expression has
ceased. In repo mutant embryos, the migration, survival and terminal differentiation of glial cells are abnormal. This study shows that Repo activates gene expression in glia, and also demonstrates that Repo mediates the suppression of neuronal differentiation. These
results suggest that Repo is the major factor that is necessary for glial development (Yuasa, 2003).
The conclusion that the expression of M84 and loco are achieved by a cooperation of Repo and Ttk69/PntP1 does not rule out the possibility that these genes are also direct targets of Gcm. In fact, reporter genes driven by loco enhancer elements are expressed normally in stage 14 repo mutant embryos, indicating that other factor(s) activate their transcription at the onset of gliogenesis.
Because the loco enhancer element contains Gcm-binding sites, Gcm can directly regulate loco. However, since the expression of Gcm in glia is transient, transcription
initiated by Gcm must be sustained by other factors. Repo and PntP1 are the best candidates for factors that maintain loco expression throughout glial development and functioning (Yuasa, 2003).
Glial fate determination involves not only the promotion of glial
differentiation but also the suppression of neuronal properties. Because
ectopic Gcm can induce neurogenesis in certain contexts, it
is unlikely that Gcm directly represses neuronal differentiation. Ttk69 has been proposed to inhibit neuronal differentiation, mainly because of its loss-of-function phenotype in the sensory organ. Here,
it has been shown that the co-expression of Repo and Ttk69 has a potent
neuron-suppressing activity, and further demonstrated that the repo mutant permits neuronal differentiation even when Gcm is overexpressed. This strongly suggests that Repo functions not only to activate the transcription of glial genes, but also to prevent the neuronal differentiation of presumptive glial cells (Yuasa, 2003).
If glia and neuron represent two mutually exclusive cell states that must be chosen between early in development, it is somewhat strange that suppression of neuronal development should be carried out by proteins that are expressed throughout glial differentiation. The existence of continuous suppression of neuronal properties in glia suggests that cells within the
nervous system may retain the potential to become neurons or glia throughout their cellular history. This idea is supported by the observation that Gcm is able to transform post-mitotic neurons into glia. Conversely, in the vertebrate nervous system, glial cells (astrocytes and oligodendrocyte-precursor) can respond to environmental signals and function as neural stem cells, generating neurons. The
role of Repo and Ttk69 may be to suppress the ability of glia to respond to cues that would cause them to change into neurons or neural precursors (Yuasa, 2003).
The bristle mechanosensory organs of the adult fly are
composed of four different cells that originate from a single
precursor cell, pI, via two rounds of asymmetric cell
division. The pattern of cell
divisions in this lineage have been examined by time-lapse confocal microscopy
using GFP imaging and by immunostaining analysis. pI
divides within the plane of the epithelium and along the
anteroposterior axis to give rise to an anterior cell, pIIb,
and a posterior cell, pIIa. pIIb divides prior to pIIa (it has been previously reported that pIIa divides prior to pIIb) to
generate a small subepithelial cell (not previously described) and a larger daughter
cell, named pIIIb. This unequal division, oriented
perpendicularly to the epithelium plane, has not been
described previously. pIIa divides after pIIb, within the
plane of the epithelium and along the AP axis, to produce
a posterior socket cell and an anterior shaft cell. Then pIIIb
divides perpendicular to the epithelium plane to generate
a basal neuron and an apical sheath (glial) cell.
The small
subepithelial pIIb daughter cell (not previously described) has been identified as a sense
organ glial cell: it expresses glial cell missing, a selector
gene for the glial fate and migrates away from the
sensory cluster along extending axons. It is proposed that
mechanosensory organ glial cells, the origin of which has been
until now unknown, are generated by the asymmetric
division of pIIb cells. Both Numb and Prospero segregate
specifically into the basal glial and neuronal cells during
the pIIb and pIIIb divisions, respectively. This revised
description of the sense organ lineage provides the basis for
future studies on how polarity and fate are regulated in
asymmetrically dividing cells (Gho, 1999).
Some neurons and glial cells originate from neuroblasts and glioblasts, stem cells that delaminate from the ectoderm of developing fly embryos. A second class of glial
cells and neurons differentiates from multipotent precursors, the neuroglioblasts. The differentiation of both glial cell types depends on glial cell
missing. Although it has been shown that this transcription factor promotes gliogenesis at the expense of neurogenesis, the cellular mechanisms underlying
this fate choice are poorly understood. Using loss and gain of function gcm mutations it has been shown that the cell fate choice takes place in the neuroglioblast,
which divides and produces a glioblast and a neuroblast. Such choice requires the asymmetric distribution of GCM mRNA, which accumulates preferentially on
one side of the neuroglioblast and is inherited by one cell, the presumptive glioblast. Interestingly, glial cells can differentiate from cells that delaminate as
neuroglioblasts or they can arise from cells that start expressing gcm several hours after delamination of a neuroblast. Altogether, these findings identify a novel
type of asymmetric cell division and disclose the lineage relationships between glia and neurons. They also reveal the mode of action of the gcm
factor (Bernardoni, 1999).
The transformation of glioblasts into neuroblasts in
gcm embryos has prompted a study of the cellular
mechanism underlying this cell fate choice and an analysis of
the lineage that gives rise to glial descendants. Although
the loss and gain of function phenotypes suggest that
gcm plays a role in the neuroglioblasts, the enhancer trap line that carries a P element in the gcm
locus is not expressed in these cells. To solve this
conundrum, the profile of expression of
gcm was examined at early stages of development and a double
immunolabeling was performed with anti-Gcm and anti-Eagle to recognize
the 6-4 thoracic stem cell. Interestingly, this cell expresses gcm before any cell division. Thus, the neural stem cell already contains the
Gcm, while the enhancer trap line only
reflects the late pattern of expression, which is restricted to
the glial lineage (glioblasts and glial cells).
Surprisingly, the Gcm product is
observed to be localized in the cytoplasm of the NGB. Upon NGB
division, Gcm is mostly detected in the glioblast,
although in some cases it is also present at low levels in the
NB. Since no asymmetric
distribution of Gcm is observed in the NGB, the preferential
accumulation of the product in one of the daughter cells
most likely depends on de novo synthesis. This is in
agreement with the observation that Gcm contains a
PEST sequence, typical of proteins characterized by a rapid
turnover. Just after the division of the
NGB, Gcm is localized in the cytoplasm and in the
nucleus of the GB. Later on, by the time when the NB
issued from the first division has already divided, Gcm is detectable in the GB, which has not yet
divided, and is mostly localized in the nucleus.
The expression of Gcm in stem cells
has been confirmed by using anti-Gcm and anti-Dpn.
Finally, the protein data have been corroborated using, simultaneously, an
anti-Dpn and a Gcm riboprobe. These
results show that Dpn is expressed in mixed
precursors (NGBs) before cell division and in the precursors
of pure lineages (NBs and GBs), such as the longitudinal
glioblast (LGB). In the progeny
of the mixed precursors (NGBs), Dpn expression is limited
to one cell, the presumptive NB, most likely
because its expression is repressed in the presumptive GB
by the Gcm product.
These data also suggest that transport of the Gcm
protein from the cytoplasm to the nucleus takes place in
the glial lineage (Bernardoni, 1999).
To elucidate the process that controls the glioblast vs
neuroblast choice, it had to be determined whether GCM RNA
is asymmetrically distributed in the stem cell. To this aim,
a Gcm-positive cell was examined that is located at the position
of neuroglioblast 5-6, which gives rise to three to five
neurons and to two to five glial cells.
The identity of this cell was confirmed using the sevenup-lacZ
transgenic line as a lineage-specific marker. Indeed, the
Gcm transcript is unevenly distributed (Bernardoni, 1999).
It is worth noting that the 6-4 and the 5-6
lineages display rather different behaviors. Indeed, while
the 6-4 stem cell delaminates at stage 10 and expresses
Gcm before any cell division, neuroglioblast 5-6 is one
of the earliest delaminating stem cells (first wave, stage 8) but only starts expressing
gcm almost 4 h later, at the end of stage 11, after the
production of neuronal progeny. Strikingly, no asymmetric
distribution was seen in two pure glioblasts:
the longitudinal glioblast, which produces seven to nine
glial cells, and the abdominal 6-4, which gives rise to two
glial cells. This indicates that this
mechanism is restricted to lineages in which a choice
between neurons and glial cell must take place (Bernardoni, 1999).
The Prospero protein also localizes to the basal side of the
neuroblast and is inherited by the ganglion mother cell. It has been suggested that PROS mRNA and protein may
act redundantly to establish ganglion mother cell-specific
patterns of gene expression. Strikingly,
while GCM mRNA is asymmetrically distributed,
the protein is present in the NGB but is not asymmetrically
distributed. It is most likely that the protein is inherited by
both cells, glioblast and neuroblast, but is rapidly degraded
due to the presence of the PEST motif.
In the glioblast, RNA accumulation allows gcm-positive
autoregulation to take place.
This reinforces the initial asymmetry and specifically promotes
the establishment of the glial fate in this cell. Thus
it is speculated that, in the case of Gcm, rapid protein
degradation in both daughter cells and positive autoregulation
in one of the two cells functionally replace the protein
asymmetric distribution observed for Pros (Bernardoni, 1999).
In order to gain insight into the neuronal and glial lineage specificity of neural progenitor cells during postembryonic Drosophila brain development, an extensive mosaic analysis was carried out. In contrast to embryonic CNS development, it was found that most postembryonic neurons and glial cells of the optic lobe and central brain originate from segregated progenitors. This analysis provides relevant information about the origin and proliferation patterns of several postembryonic lineages such as the superficial glia and the medial-anterior Medulla neuropile glia. Additionally, the spatio-temporal relationship was studied between gcm expression and gliogenesis. It was found that gcm expression is restricted to the post-mitotic cells of a few neuronal and glial lineages and it is mostly absent from postembryonic progenitors. Thus, in contrast to its major gliogenic role in the embryo, the function of gcm during postembryonic brain development seems to have evolved to the specification and differentiation of certain neuronal and glial lineages (Colonques, 2007).
These data indicate that neurogenesis and gliogenesis follow rather diverse developmental patterns in different regions of the larval brain. One main conclusion is that the great majority of postembryonic neurons and glia from the central brain (CB), outer proliferation center (OPC), and larval brain surface originate from segregated progenitors (either NBs or GBs). Thus, only mixed lineages (neurons and glia) were found in radial CB-ME clones and in a minority of the CB clones. Part of these lineages may correspond to the mushroom bodies, which have been previously shown to be generated by mixed lineages. The segregation of neuronal and glial progenitors in the larval brain is in accordance with a clonal analysis carried out in the ventral ganglion, which found only mixed lineages in 5% of the clones. Also, a mosaic analysis focusing on the first optic ganglion concluded that lamina neurons and glia are derived from distinct lineages in the lamina progenitor cell (LPC) and glial progenitor cells (GPCs), respectively. Nevertheless, a further study found that some GPC progenitors in second instar larvae had the potential to generate both neurons and glia. However, only glia are generated from the GCM expressing progenitors of the GPCs and only neurons are produced from the LPC. It was also found that in addition to LA neurons, some ME neurons also originate from the LPC (Colonques, 2007).
Strikingly, a recent study has concluded that glial proliferation in the larval CB occurs mainly from NGBs (Pereanu, 2005). How can one explain such contrasting conclusions between this studies? Part of the difference may result from the different techniques (FLP/FRT vs. MARCM) and Gal4 drivers (elav-Gal4 vs. tubP-GAL4) that were used to generate the clones. Since lamelliform processes of cortical glia form a very tight scaffold around neurons and neuroblasts, an important factor that might have also contributed to the different conclusions is the use of membrane bound GFP. This makes rather difficult to distinguish whether the neurons embedded inside a labeled glial scaffold are part of a clone or simply caged by the labeled glial processes. To avoid this problem, UAS-nls-LacZ, which labels cell nuclei, was used in this study. Also, given the very short cell cycle of postembryonic neural progenitors, 10 min pulses for BrdU in vitro labeling were applied instead of the long (12 h) BrdU in vivo pulses used by Pereanu. Finally, a fourth factor might be the timing of clone generation. Since most of the clonal analyses were carried out from second instar larvae, the possibility cannot be ruled out that glial and neuronal progenitors had a common lineage in earlier larval stages. However, the few CB clones that were produced in early larvae were entirely neuronal. Thus, the current results favor the idea that CB NBs and GBs diverged during embryonic or very early larval development. This also seems to be the case for most ME lineages. In this case, most neurons are born from the OPC while neuropile glia are generated from two sources: the lateral-posterior ME glia from the GPCs and the medial-anterior ME glia from progenitors located at the CB-OPC border (Colonques, 2007).
Mosaic analysis also sheds some light on the origin of certain cell types and their patterns of proliferation. For instance, although both surface glia and cortical glia processes contribute to the glial layer surrounding the larval brain (Pereanu, 2005), superficial clones exclusively contained surface glia. This indicates a complete lineage separation between cortical and surface glia. Interestingly, superficial glial clones often do not respect the morphologic borders of proliferative primordia (i.e., OPC, LPC, CB, etc.). This suggests that the development of these superficial glia may be subject to independent morphogenic cues. In addition, the data clearly show that superficial glia are proliferative. Their low mitotic index probably indicates that either they have slow cell cycles with a long interphase or that only a subset of the superficial glia are progenitor cells (Colonques, 2007).
The pattern of immunostaining of the CB neuronal clones (one large progenitor and several small ELAV positive cells) fits well with the idea that CB lineages are generated by scatteredly distributed individual NBs through repetitive asymmetric divisions similar to those of embryonic NBs. With regards to the origin of CB glia, the fact that pure glial CB clones contain a single progenitor and a few glial cells, favors a repetitive asymmetric (stem cell like) pattern of division for CB GBs. Nevertheless, given the few clones of this type obtained, other possibilities should not be ruled out (Colonques, 2007).
The pronounced changes in the geometry of OPC clones during development reflect time dependent changes in the pattern of OPC NB proliferation. Thus, the change in the shape from a broad band when generated in early stages to a thin stream of cells when generated in late stages is consistent with a switch in the division pattern from tangentially oriented symmetric divisions to radially oriented asymmetric divisions that OPC NBs experience during the third instar stage. Interestingly, LPC-OPC clones expand tangentially in the LPC, strongly suggesting that LPC progenitors divide symmetrically a few times before yielding progeny orientated towards the LA and/or ME through asymmetric divisions (Colonques, 2007).
In the embryonic CNS, the identity of individual cells generated in small lineages is achieved by the sequential expression of a repertoire of transcription factors. Among them, gcm regulates the neuronal vs. glial cell fate decision in mixed lineages. In the larval brain gcm is expressed early in the lamina glial precursors of the GPCs and it is required for their differentiation. gcm probably plays a differential role in the specification of these two ME glial lineages. The lateral-posterior ME glia originate from gcm expressing precursors located at the GPCs but they lose GCM expression during their migration to the ME neuropile. In contrast, it has been shown in this study that the medial-anterior ME neuropile glia originate from progenitors located in the CB-OPC border which lack GCM. In these glia, GCM appears to be expressed post-mitotically all along the migratory pathway and after glia reach the neuropile. Interestingly, both groups of glia seem to migrate from their respective sites of origin to either side of the ME neuropile along axon scaffolds formed by neurons generated earlier in the respective lineages. Thus, this appears to be a strategy to direct glial migration to the correct location (Colonques, 2007).
Interestingly, gcm is also expressed in the precursors of lamina neurons in the LPC. These results point to a dual role of gcm in gliogenesis and neurogenesis during postembryonic CNS development. Nevertheless, in the CB gcm is only expressed in post-mitotic neurons of a single neuronal lineage. Together with the apparent absence of GCM expression in larval CB progenitor cells, this result strongly suggests that GCM plays a role in the specification and differentiation of this neuronal lineage rather than in neurogenesis or gliogenesis. This is also supported by the limited transformation of neurons into glial cells that was found after the ectopic expression of gcm in the larval brain. Thus, the gliogenic capacity of gcm seems to be restricted to a small subset of larval brain progenitor cells. This may explain why mammalian Gcm genes do not seem to play a gliogenic role in spite of being able to functionally substitute to Drosophila GCM. Interestingly, a chicken Gcm ortholog is expressed in early neuronal lineages of the spinal cord and is required for neuronal differentiation. Concomitant with a possible role in neurogenesis, it has been also found that the precocious expression of gcm in the peripheral nervous system induces neurogenesis (Colonques, 2007).
In summary, the current results favor a role of gcm in the specification of a few postembryonic brain lineages. Thus, a different genetic strategy appears to have been selected during evolution for the generation of larger lineages of similar cells mostly arising from segregated neuronal and glial progenitors. It is hypothesized that the regulated expression of a repertoire of transcription factors, including gcm, will presumably be used to orchestrate lineage specificity in the adult brain of Drosophila (Colonques, 2007).
The effects of gcm
ectopic expression on the fate of non-neural cells have been tested. When gcm expression is continuously induced in
epidermal cells from around stage 9, these cells started to exhibit mesenchymal cell morphology at
stage 13, which is preceded by the onset of expression of Repo, a glial marker. The morphological
change is coincident with loss of expression of epidermal cell-adhesion molecule FasIII. Dorsal closure is defective in embryos in which gcm is expressed ectopically in the ectoderm. Elav is induced in the epidermis with a one hour lag time between the onset of ectopic Gcm expression and that of Elav. There are four sites to which Gcm can bind in a 3.5 kb segment upstream of the elav gene. In addition to
the epidermis, the fate of mesodermal cells is also affected by gcm ectopic expression. These findings
suggest that gcm can convert gene expression and cell morphology even outside the neuroectoderm (Akiyama-Oda, 1998).
Glial cells constitute the second component of the nervous system and are important during
neuronal development. The gene described in this paper, glial cell deficient (glide) also known as glial cells missing, is
necessary for glial cell fate commitment in Drosophila. Mutations at the glide locus
prevent glial cell determination in the embryonic central and peripheral nervous systems. Moreover,
the absence of glial cells is the consequence of a cell fate switch from glia to
neurons. This suggests the existence of a multipotent precursor cells in the nervous system. glide
mutants also display defects in axonal navigation, which confirms and extends previous results
indicating a role for glial cells in these processes (Vincent, 1996).
In gcm mutants, glia precursors become neural. A glial phenotype is evident in gcm expressing neural precursors, engineered to express gcm ectopically. In gcm mutants, longitudinal CNS tracts are abnormal (Jones, 1995). "Pioneer" neurons in gcm mutants find correct pathways, indicating that GCM is not required for initial axon pathfinding (Hosoya, 1995), but it is required later in the axonogenesis process.
Whereas considerable progress has been made in
understanding the molecular mechanisms of axon guidance
across the midline, it is still unclear how the axonal
trajectories are establised for longitudinal pioneer neurons, which never
cross the midline. Longitudinal glia of the embryonic Drosophila CNS direct
formation of pioneer axon pathways. By ablation and
analysis of glial cells missing mutants, it has been demonstrated that
glia are required for two kinds of processes. (1) Glia
are required for growth cone guidance, although this
requirement is not absolute. The route of
extending growth cones is rich in neuronal cell bodies and
glia, and also in long processes from both these cell types.
Interactions between neurons, glia and their long processes serve to
orient extending growth cones. (2) Glia direct the
fasciculation and defasciculation of axons, which pattern
the pioneer pathways. Together these events are essential
for the selective fasciculation of follower axons along the
longitudinal pathways (Hidalgo, 2000 and references therein).
Longitudinal pathways are pioneered by pCC, MP1, dMP2
and vMP2, which extend in pairs in opposite directions.
pCC and vMP2 extend together anteriorly, whereas MP1 and
dMP2 extend together posteriorly. All four contact half way to
establish the first, single longitudinal fascicle. Subsequently,
pioneer axons undergo a series of defasciculation and
refasciculation events to establish, by the end of
embryogenesis, two primary fascicles at each side of the
midline; pCC fasciculates with dMP2 along the first fascicle,
closest to the midline, vMP2 runs also along this pathway, but
in a more ventral plane defasciculated from pCC/dMP2, and
MP1 runs along the second, central fascicle. A third, outer fascicle is visualised by fasII, but
its pioneer neurons remain undiscovered. It is not known what
governs the defasciculation and refasciculation events that
build the final trajectories of pioneer axons (Hidalgo, 2000).
At the beginning of axonogenesis (stage 12.3), the growth
cones of pioneer neurons pCC and vMP2 extend together
anteriorly, as the growth cones of MP1 and dMP2 extend
posteriorly. Antibody 22c10
recognizes the axons of dMP2 and vMP2, whereas fasII
recognizes the axon of pCC.
Glia migrate ahead of extending growth cones. When the
progeny of the glioblast reach the cell bodies of the pioneer
neurons, they stop migrating in the dorsoventral direction and
cluster around the neurons. At stage 12.3, the growth cones of
dMP2 and vMP2 extend toward the glia. Glia do not
form a prepattern for the axonal trajectories, as most of the
segment at this stage is free of glia. Glia migrate to the
posterior, concomitently with pioneer growth cone extension;
in particular, the axon of dMP2 follows the glia,
and sends projections toward the glia. The growth cone
of vMP2/pCC appears to extend further than the most anterior
glia, as visualized with the nuclear marker Repo. However Heartless-positive glial cytoplasm
abuts the pCC axon and glial
projections extend ahead of the pCC/vMP2 growth cone. dMP2/MP1 and pCC/vMP2 meet over a glial cell
to form the first longitudinal pathway, which by stage 13 is
covered in glia.
The distance travelled by the vMP2 and dMP2 growth cones
is rich in cell membranes, asrevealed with rhodamin-phalloidin. Glial projections contribute to this membrane mesh (Hidalgo, 2000).
Longitudinal glia coexpress the nuclear protein Repo and the
membrane protein Heartless. At the time of pioneer
growth cone extension and prior to their fasciculation,
Heartless-positive glial projections reach across adjacent
segments and make glia-to-glia contact.
Neuronal filopodia also cover the distance to be travelled by
the pioneer growth cones. The pCC growth
cone extends long filopodia that contact glia
from the adjacent anterior segment prior to
fasciculation with the dMP2/MP1 growth
cone. While the vMP2 and dMP2
growth cones are still far apart, the
pCC/vMP2 growth cone also makes contact
with intermediate neurons, which can be
visualized with fasII and 22c10. One of these intermediate neurons
is SP1. Glia migrate along
the cell bodies of fasII-positive cells located
ahead of the glia, along their trajectory. It
appears that the same intermediate neurons
are contacted by both growth cones and
migrating glia (Hidalgo, 2000).
The fact that glia precede extending
growth cones during axonogenesis and that
growth cones send filopodia toward glia
suggests that glia attract pioneer growth
cones. However, along the trajectory of
migrating glia and extending growth cones,
there are intermediate neurons that are
contacted by both cell types. This suggests
that bidirectional contact between neurons
and glia occurs during the establishment of
the first longitudinal trajectory (Hidalgo, 2000).
The role of glia in guidance was studied by
ablating longitudinal glia with
expression of the toxin Ricin, driven by the GAL4 system, and by studying
glial cells missing (gcm) mutations, in which
glia are transformed toward a neuronal fate. The s-gcmGAL4 line
drives Ricin expression in the longitudinal glioblast in a mosaic
fashion, so that in most embryos only one or two glioblasts are
killed. This line is also expressed in other
glial types; however, expression is also mosaic in these glia so
that secondary damage to the ventral nerve cord is minimized.
Only the catalytic subunit (A) of Ricin is expressed, such that
the toxin cannot exit the expressing cells. Targeted ablation with Ricin-A is cell
autonomous, since cells adjacent to ablated cells are unaffected.
At the beginning of axonogenesis (stage 12.3), dMP2
extends posteriorly, fasciculating with MP1 and aCC and
surrounded by glia for a short distance. The growth
cone branches at the location of a glial cell (choice point 1): one branch carries on toward the muscle (the aCC
axon), and the other branch (dMP2) heads posteriorly to meet
the vMP2/pCC growth cone (choice point 2).
When the glioblast is ablated, the dMP2 growth cone extends
more slowly, either because the axon is shorter than that in a normal
adjacent segment or because it does not meet the pCC/vMP2 fascicle.
It appears thickened, and has a large growth
cone. However, dMP2 does eventually
extend; slightly later, the growth cone may not
branch out: the dMP2/MP1 axon heads toward the
muscle fasciculating with aCC.
At stage 13, normally the ascending growth
cone of vMP2 and descending dMP2 meet and
fasciculate over a glial cell to form
the first longitudinal single fascicle.
Ablation of longitudinal glia can lead to failure of
contact and fasciculation between pCC/vMP2 and
dMP2/MP1 fascicles, or to loss of the
dMP2/MP1 pathway. In many
cases, formation of the first fascicle is not affected
by the absence of glia.
These data suggest that glia attract the
dMP2/MP1 growth cone, provoke axonal
defasciculation at the branching choice point, and
aid fasciculation of all growth cones into a single
fascicle (Hidalgo, 2000).
Patterns of pioneer axons change with time. These changes in axonal
fasciculation are controlled by glia. In the wild-type
CNS, glia are found at the choice points where
fasciculation and defasciculation of axons take place.
In the absence of glia from these positions, either through
ablation or gcm mutations, pioneer neurons do not
defasciculate from sister axons. This leads to axonal
phenotypes that recapitulate the fasciculation patterns typical
of earlier developmental stages. For instance, at choice point 1
where dMP2 normally separates from aCC to head posteriorly,
ablation of glia leads to the misrouting of dMP2 toward the
muscle, since it does not defasciculate from aCC. Similarly,
whereas normally at choice point 2 the two dMP2/MP1 and
vMP2/pCC growth cones meet and fasciculate together,
absence of glia can prevent contact between the two growth
cones. Hence, when glial function is compromised, more axons
take the route of the intersegmental nerve, toward the muscle.
Glia are also located at two other choice points, 3 and 4, where
pCC/vMP2 defasciculates from dMP2/MP1 and subsequently
along the separating dMP2 and MP1 axons. By the end of
embryogenesis, instead of the normal three fascicles, absence
of glia often leads to fusion of the fascicles into a single
fascicle. The formation of a single fascicle is consistent with the fact
that formation of the first longitudinal fascicle is not severely
affected by loss of glia (Hidalgo, 2000).
Several pieces of evidence indicate that these recapitulation
phenotypes are due to failed defasciculation. The data demonstrate that glia
drive the defasciculation and refasciculation events that shape
pioneer axon patterns. The most remarkable phenotype
from these experiments is the general disruption of the scaffold
formed by the three axonal facicles separated by glia. Both
upon ablation and mutation this scaffold is damaged. Follower
axons may still extend along the longitudinal pathways;
however, their selective fasciculation routes are altered. It
would be interesting to know if these defects lead to altered
connections of follower axons with their targets (Hidalgo, 2000).
While survival of CNS neurons appears to depend on multiple neuronal and non-neuronal factors, it remains
largely unknown how neuronal survival is controlled during development. Glia regulate neuronal survival during formation of the Drosophila embryonic CNS. When glial function is impaired either by
mutation of the glial cells missing gene, which transforms
glia toward a neuronal fate, or by targeted genetic glial
ablation, neuronal death is induced non-autonomously.
Pioneer neurons, which establish the first longitudinal axon
fascicles, are insensitive to glial depletion, whereas the
later extending follower neurons die. To investigate whether the missing longitudinal tracts reflect
neuronal loss, it was asked if pioneer neurons disappear upon glial
ablation. The pioneer neurons MP1 and dMP2 express the
nuclear marker Odd and are both present despite being isolated from neighbouring glia, and are occasionally duplicated. The pioneer neurons vMP2, MP1 and dMP2 were monitored with the membrane
marker FasII at stage 14 and were present in hemisegments with ablated interface
glia. (Glia were ablated by expressing the catalytic subunit of Ricin
toxin with the Gal4 system: this subunit is unable to leave the cell, hence leading to cell
autonomous ablation). The cell bodies of the longitudinal pioneer
neuron pCC and the motorneurons aCC and RP2, were
visualized with anti-Eve antibodies. When longitudinal glia are ablated, RP2 is present and is occasionally duplicated. Ablation of the glioblast alone leaves other interface glia that contact aCC
and pCC. However, in severe embryos where more interface
glia are ablated, pCC and aCC as well as RP2 also remain in
the absence of neighboring glia. The possibility that progenitors of pioneer neurons may undergo
further cell divisions to maintain a constant number of pioneer
progeny in the absence of glia cannot be ruled out. Nevertheless, no net loss of pioneer neurons is seen when interface and other glia are ablated. This suggests that pioneer neurons do not depend on glia for survival. Once the first longitudinal fascicles are formed, as follower neurons cross the midline and fasciculate with the pioneer axons,
longitudinal glia regulate their survival as the mature CNS is built (Booth, 2000).
In flies, the choice between neuronal and glial fates depends
on the asymmetric division of multipotent precursors, the
neuroglioblast of the central nervous system and the IIb
precursor of the sensory organ lineage. In the central
nervous system, the choice between the two fates requires
asymmetric distribution of the Glial cell missing RNA in the neuroglioblast. Preferential accumulation of the transcript in one of the daughter cells
results in the activation of the glial fate in that cell, which
becomes a glial precursor. gcm is
necessary to induce glial differentiation in the peripheral
nervous system. Evidence suggests that GCM
RNA is not necessary to induce the fate choice in the
peripheral multipotent precursor. Indeed, GCM RNA
and protein are first detected in one daughter of IIb but not
in IIb itself. Thus, gcm is required in both central and
peripheral glial cells, but its regulation is context
dependent. Strikingly, it has been found that only subsets of
sensory organs are gliogenic and express gcm. The
ability to produce glial cells depends on fixed, lineage
related, cues and not on stochastic decisions. After gcm expression has ceased, the IIb
daughter migrates and divides symmetrically to produce
several mature glial cells. Thus, the gcm-expressing
cell, also called the fifth cell of the sensory organ, is indeed
a glial precursor. This is the first reported case of
symmetric division in the sensory organ lineage. These data
indicate that the organization of the fly peripheral nervous
system is more complex than previously thought (Van De Bor, 2000).
To understand how gliogenesis takes place in the peripheral
nervous system (PNS), the wing, where the array
of glial cells and sensory organs has already been described, was analyzed.
Mechano- and chemo-sensory organs differentiate from ventral
and dorsal epithelia, most of them being located along the
anterior margin (also called L1 vein) and the L3 vein. Axons from the sensory neurons fasciculate into two
nerves that navigate along the L1 and L3 veins to reach the
thoracic ganglia. Like sensory organs, most glial
cells differentiate along L1 and L3 and require the
activity of proneural genes in order to differentiate. Subsequently, glial cell nuclei
are found along the whole nerves, due to distal to proximal
migration along the axonal fibers.
In an attempt to identify the molecule(s) involved in
peripheral glial differentiation, the role of
gcm, which is required for the differentiation of all lateral
glial cells in the embryonic CNS, was analyzed. Since gcm is an
embryonic lethal mutation, gcm mitotic
clones were induced using the FLPase recombinase and a strain carrying myc
FRT that allows the recognition of mutant territories. The glial phenotypes were then analyzed using
the antibody against the reverse polarity (Repo) protein, which
is expressed in all but midline glial cells during embryonic and
post-embryonic development (Van De Bor, 2000).
In a gcm null clone
straddling the dorsoventral boundary, almost two thirds of the
anterior margin is affected by the mutation. The rest of the
margin is wild-type on one or both epithelia. Strikingly, the
number of glial cells along L1 is drastically decreased. The few
glial cells detected in the mutant territory are not themselves
mutant, since they are all Myc positive. These wild-type
glial cells are likely to have differentiated distally to the
clone and subsequently migrated into it by following the axon
bundle. Similar results have been obtained in clones affecting L3. In all cases, no example of gcm mutant glial
cells was ever observed, indicating that gliogenesis strictly
depends on gcm (Van De Bor, 2000).
In the embryonic CNS, lack of gcm leads to the
generation of supernumerary neurons. It was therefore asked
whether lack of wing glial cells is accompanied by an increased
number of neurons. As a marker for neuronal nuclei
the embryonic lethal abnormal vision (elav) gene was used. Mitotic clones were generated and
simultaneously labelled with anti-Myc, anti-Repo and
anti-Elav, in order to recognize simultaneously the glial and
neuronal nuclei as well as the mutant territories. Indeed, more
Elav labelling is observed in mutant clones than in wild-type
tissues (Van De Bor, 2000).
It has been recently shown that sensory organ precursor
(SOP) cells of notum microchaetes generate glial cells. These
SOP cells divide asymmetrically and produce the PIIa and the
PIIb cells, two second order precursor cells that also divide
asymmetrically. PIIa generates the tormogen and the
trichogen. PIIb produces a glial cell and a PIIIb precursor,
which in turn divides and generates a neuron and a sheath
cell. If this type of lineage were common to all sensory organs, the
number of supernumerary neurons in gcm clones should
correspond to the number of sensory organs affected by the
mutation. However, clones located along the L3 vein and
affecting four sensory organs (L3-v, ACV, L3-1, L3-
2) only display two supernumerary neurons (4 clones were
analyzed). Two alternative mechanisms might explain this
result: either the fate transformation is not complete in the
mutant cells or two types of sensory organs exist, gliogenic and
non-gliogenic (Van De Bor, 2000).
To determine whether all sensory organs produce glial
cells, two approaches were used. First, by analyzing gcm
clones affecting individual sensory organs one
supernumerary neuron was detected when L3-1, L3-3
or L3-v were mutated, whereas none was detected
in clones affecting L3-2. No clones affecting only ACV
were found (Van De Bor, 2000).
Second, the profile of gcm expression was analyzed in
different wing sensory organ lineages. Along the anterior
margin, gcm is detected in the inner-most cell of chemo-and
mechanosensory organs, when the precursors have already
divided more than once. This indicates that
both types of sensory organs can generate glial cells. gcm
expression is also observed in large campaniform sensilla of
the L3 vein. Surprisingly, however, not all of them are
gcm positive. L3 campaniform sensilla have been divided
into two classes, early and late, according to their birth time,
central projections and electrophysiological behavior. Early sensilla include ACV and L3-2; late
sensilla include L-3v, L3-1 and L3-3. gcm
is only expressed in late sensilla.
Expression is first detectable in the sub-epithelial cell produced
by the PIIb division. In this cell, it persists until the five cell
stage, but becomes undetectable later on. Interestingly, notum macrochaetes behave like the
early sensilla in that they do not express gcm nor repo,
although the mutant phenotype has not been assessed (Van De Bor, 2000).
These results indicate that at least two types of sensory organs
exist -- a gliogenic one and a non-gliogenic one -- and that lack of
Gcm leads to the complete transformation of glial
cells into neurons in the gliogenic ones. Interestingly,
supernumerary neurons are not adjacent to the sensory organs,
due to migration (Van De Bor, 2000).
The finding that some sensory organs are not gliogenic
prompted studies to follow their lineage using triple labelling with anti-Elav,
anti-Cut and anti-Prospero (anti-Pros). cut is initially
expressed in the SOP and in the two daughter cells, PIIa and
PIIb. Upon divisions of these
two cells, it is present in all the progeny, however it is
expressed at higher levels in the tormogen and the trichogen
cells than in the inner, neuronal and sheath, cells. Anti-Pros
identifies PIIb and its progeny.
The lineage of late campaniform sensilla is the same as that
of notum microchaete. PIIb divides and produces PIIIb, and a small sub-epithelial
cell. The small cell accumulates the highest
levels of Pros, expresses gcm and, later on, repo. Pros is asymmetrically localized in the cortex of the
dividing PIIIb. As in microchaete, the daughter that
accumulates Pros at high levels is the neuron, however, at later
stages, the sheath cell shows higher Pros levels than the neuron,
most likely due to de novo synthesis (Van De Bor, 2000).
At early stages, the lineage of non-gliogenic sensilla is
similar to that of gliogenic ones. Starting from the division of
PIIb, however, there are a number of differences. (1) The PIIb
cell produces two daughters of equal size: PIIIb and a cell
located in the sub-epithelial layer that preferentially inherits
Pros. This cell, which continuously expresses Pros at
high levels and does not express glial-specific genes,
corresponds to the sheath cell. (2) When PIIIb
divides, Pros is present but is not asymmetrically localized. The two PIIIb daughters initially accumulate Pros at
similar levels but at later stages express it at different levels. Interestingly, both cells are labelled by the Elav-neuronal
marker: one of them belongs to the early
campaniform sensillum, the other, which at late stages
accumulates higher levels of Pros and migrates, constitutes one
of the 'extra' neurons (Van De Bor, 2000).
In the CNS, asymmetric GCM mRNA distribution dictates
the fate choice in the NGB. Surprisingly, in the PNS, gcm is
not detectable in the multipotent precursor, PIIb, suggesting
that it does not dictate fate choices. The analysis of
mutant lineages also confirms this hypothesis. If indeed
gcm were expressed and required to induce a fate choice
in PIIb, lack of its product would induce the transformation of
the glial cell into a PIIIb. This would entail the presence of two
neurons and two sheath cells in a sensory organ consisting of
six cells. However, mutant sensory organs never contain two
sheath cells and are always composed of five cells, as in the
wild-type. All these data
strongly suggest that gcm is only expressed and required
in the PIIb daughter and that its role in the PNS is different
from that observed in the CNS (Van De Bor, 2000).
To better understand the mode of action of gcm, gain of function phenotypes were produced by ubiquitously
expressing gcm at different stages of L3-3 development.
Heat shock before SOP birth, at around 0 hour APF, produces
no effects. Heat shock at 2 hours APF results
in the production of a sensory organ consisting of four cells:
two cells express the Repo glial-specific marker, while the Elav
labelling is completely absent. Heat shock at 5 hours
APF results in an L3-3 consisting of five cells, three of which
express Repo. One of the Repo-positive cells also
expresses Elav at very low levels. In all cases, the cells that are
not Repo-positive express high levels of cut, which identifies
them as tormogen and trichogen. These data suggest that the
2-hour APF shock leads to both PIIb daughters adopting the
glial fate, while the 5-hour APF shock leads to both PIIIb
daughters adopting the glial fate. Surprisingly, in no case was a complete transformation of the sensory organ
cells into glial cells observed. This result is even more
striking when compared to the effects obtained in the
embryonic CNS, in which the same UAS-gcm transgene is
able to transform all neural stem cells into glioblasts. A different competence to acquire
the glial fate among the cells of the sensory organ lineage is
also confirmed by the observation that the fate transformation
remains partial even when higher doses of ectopic Gcm
are provided (Van De Bor, 2000).
gcm is expressed in a cell of the sensory organ and ceases
to be expressed after that cell starts migrating, whereas repo is
expressed in the glial lineage until terminal differentiation . By comparing the profile of expression of the two genes,
it was found that while the number of gcm-positive cells
corresponds to that of gliogenic sensory organs, the number of
Repo-positive cells is much higher. For example, on L3 it goes
from 15 to 22 (vs. three gliogenic sensory organs). A possible
explanation for this result is that the sensory organ lineage
contains more than five cells, due to the division of the
gcm-expressing cell. In this way, a determination was made of the precise
number of glial cells arising from each sensory organ as well
as the profile of glial proliferation in an individual sensory
organ (Van De Bor, 2000).
Analysis of gcm clones affecting L3 gliogenic sensory
organs individually or in combination allows the relative contribution of each sensory organ to the total
number of glial cells to be established. The three L3 gliogenic
sensory organs produce on average six cells. This analysis
confirms that the gcm-expressing cell is a glial precursor
cell and shows that the number of divisions producing glial
cells is comparable among the sensory organs.
The profile of proliferating glial cells during development
was obtained by double labelling with anti-Repo and anti-PH3,
an antibody that recognizes the phosphorylated form of histone
H3 present in the chromatin of dividing cells. Since glial cells migrate from distal to proximal during
development, focus was placed on the gliogenic
lineage most-distally located and nuclei between L3-
3 and L3-1 were counted. While at 9 hours APF only one Repo-positive cell
is present in that region, five to eight Repo-positive
cells can be detected at 36 hours APF.
Between 17 hours and 30 hours APF several mitotic figures
are observed, concomitant with the progressive increase in
Repo-positive cells. Thus, the fifth cell of the sensory organ
proliferates and produces several glial cells. While this cell
expresses both gcm and repo, its daughters express
only repo. This cell will be referred to as GP, for glial precursor.
Given the final number of glial cells,
it is clear that more than one division occurs. At least two
modes of proliferation can account for the results observed: (1)
the GP divides asymmetrically and produce another glial
precursor and a differentiated glial cell; (2) the GP divides
symmetrically and produces two proliferating daughters. Two
Repo-positive dividing cells were indeed observed in the L3-3
to L3-1 region, strongly supporting the model in which
both GP daughters have proliferative potential (Van De Bor, 2000).
It is important to note that the time of division is not fixed,
since at any given stage the number of proliferating glial cells
varies from wing to wing.
Similarly, the number of cell divisions is not tightly fixed
within a given lineage, since the total number of glial cells is
not constant among the different wings analyzed.
The average number of glial cells associated with each sensory
organ as well as the average number of glial cells on L3 suggests
that on average, more than two rounds of cell division take place
for each GP.
It is also worth noting that clones lacking gcm display
sensory organs composed of five cells carrying only one
additional neuron indicating that this neuron does not
proliferate. This explains why the number of additional
neurons in mutant territories is always lower than the number
of Repo-positive cells found in corresponding regions in the
wild-type background (Van De Bor, 2000).
By using gain-of-function mutations it has been proposed
that vertebrate Notch promotes the glial fate. In
vivo glial cells are produced at the expense of neurons
in the peripheral nervous system of flies lacking Notch and
that constitutively activated Notch produces the opposite
phenotype. Notch acts as a genetic switch between neuronal
and glial fates by negatively regulating
glial cells missing, the gene required in the glial precursor
to induce gliogenesis. Moreover, Notch represses
neurogenesis or gliogenesis, depending on the sensory
organ type. Numb, which is asymmetrically localized in the
multipotent cell that activates the glial cell fate, inducing
glial cells at the expense of neurons. Thus, a cell-autonomous
mechanism inhibits Notch signaling (Van De Bor, 2001).
Strikingly, N seems to act in opposite directions
in fly and some vertebrate peripheral glial cells. Indeed,
gain-of-function N mutations promote differentiation of
Müller, radial and Schwann glial cells. Two possible
explanations can account for these results: (1) the genetic
switch between neuronal and glial fates has different
requirements in flies and vertebrates; (2) the role of N
depends on the subtype of glial cell. The analysis of other
classes of fly glial cells will help elucidate this point.
Preliminary analyses on the embryonic CNS
suggest that the response of central glial cells to N depends
on the subtype. The observation that
oligodendrocyte differentiation, like fly peripheral glial cells,
is repressed by N, also argues in favor of the second
hypothesis (Van De Bor, 2001).
One of the most striking results is that repression of the N
pathway throughout the development of the sensory organ
(obtained by N loss-of-function mutations or by Numb
overexpression) leads to sensory organs composed of six glial
cells. The competence to adopt
the glial fate is restricted to some cells of the sensory organ
lineage; the strongest phenotype observed upon overexpression
of gcm throughout the lineage is the differentiation of a
sensory organ composed of five cells, three of which are Repo-positive. Thus, gcm is not sufficient
to induce a IIa into IIb transformation. This indicates that the
pathway mediated by N and Numb affects the competence of
sensory organ cells to adopt the glial fate. In molecular terms,
this implies the control of expression of gcm regulators,
positive co-factors and/or repressors (Van De Bor, 2001).
Neurons and glial cells depend on similar developmental pathways and often originate from common precursors; however, the
differentiation of one or the other cell type depends on the activation of cell-specific pathways. In Drosophila, the differentiation of
glial cells depends on a transcription factor, Gcm. This glial-promoting factor is both necessary and sufficient to induce the
central and peripheral glial fates at the expense of the neuronal fate. In a screen for mutations affecting the adult peripheral nervous
system, a dominant mutation inducing supernumerary sensory organs was found. Surprisingly, this mutation is allelic to
gcm and induces precocious gcm expression, which, in turn, activates the proneural genes. As a consequence, sensory organs are induced. Thus,
temporal misregulation of the Gcm glial-promoting factor reveals a novel potential for this cell fate determinant. At the molecular level, this implies unpredicted
features of the gcm pathway. These findings also emphasize the requirement for both spatial and temporal gcm regulation to achieve proper cell
specification within the nervous system (Van De Bor, 2002).
These results show that gcm is able to induce PNS formation. Interestingly, supernumerary sensory organs are induced only when gcm is expressed precociously, due to the activation of the AS-C genes. Indeed, the stage at which gcm induces PNS formation is the same stage at which ectopic expression of the AS-C induces PNS formation. Together with the analysis of the epistatic relationship these results suggest that all the effects observed in gcmPyx (Polythryx, the dominant mutation of gcm) and in hs-gcm flies are mediated by ectopic AS-C expression. It is possible that Gcm acts on the AS-C directly, since several Gcm-binding sites are present in the promoter sequences that induce ac and sc expression. For example GBSs are found in the regulatory elements that promote AS-C expression in the dorso-central and scutellar clusters. A search for GBSs in the ASC promoter from Drosophila virilis has revealed that several sites do exist. Some of them are located in the region that corresponds to the dorso-central enhancer, even though the precise position of GBSs is not conserved between D. virilis and D. melanogaster. It will be interesting to determine whether the sites present in the ASC promoter are functional and important during development. This might provide insight into a previously unknown regulation of the AS-C genes by Gcm or related proteins. It is worth mentioning that precocious gcm induces the AS-C, but not other proneural genes such as ato (Van De Bor, 2002).
It is tempting to speculate that the supernumerary bristles induced by precocious gcm are due to defects in positional information. A number of genes controlling the AS-C and sensory organ differentiation are required to define specific territories in a given tissue. This class of genes, to which pannier and the iroquois complex belong, has also been called prepatterning genes. Gcm shares some features with such genes in the sense that it affects the expression of proneural genes and thereby PNS development. Interestingly, however, while pannier mutations also affect notum differentiation, precocious gcm does not seem to alter the general structure of the tissue in which it is expressed. It is unlikely that gcm acts by regulating prepattern genes, since its precocious expression does not modify the expression of pannier. Thus, even though gcmPyx and pannier affect the same sensory organs (dorso-central bristles), it is likely that they act independently (Van De Bor, 2002).
Glial cells function as guide post cells for axonal pathfinding. Due to the difficulty in completely eliminating glial
cells during development, their functions in axonal pathfinding
have not been critically evaluated. In Drosophila
gcm mutant embryos, glial cells are eliminated,
providing a unique opportunity to investigate
glial functions in nervous system formation. Even in the absence of glial cells the initial
axonal extension of pioneer neurons is essentially normal.
However, at later stages, axon bundle formation and
pathfinding are disturbed in the absence of glial cells,
and abnormal migration of glial cells leads to misrouting
of axons. This indicates that glial cells are required for
correct pathfinding at later stages. It is proposed that glial
cells function in a stage-specific manner; they are not required
for the initial extension of pioneers but essential
for the subsequent extension of pioneers and followers as
well as axon bundle formation (Takizawa, 2001).
Although initial behaviors of pioneer neurons were normal
in the glia-less environment, distinct
pathfinding errors in subsequent stages have been noted. The dMP2 axon
in gcm mutants was traced using the 15J2 Gal4 line,
which expresses Gal4 in this neuron. In some hemisegments of gcm mutant embryos, the dMP2 neurons send out axons in an abnormal orientation,
occasionally joining with a motor pathway. This indicates that without functional
glial cells, the descending dMP2 axon can correctly fasciculate with an ascending vMP2 axon, but can not perform subsequent pathfinding tasks. Anti-Fas ll (mAb1D4) staining also reveals that the pCC pathway (medial most axon pathway) is broken or becomes thinner at stage 15 (Takizawa, 2001).
The effect of mislocated lateral glia (LG) and axonal
tracts of follower axons wes tested using a hypomorphic allele, gcm5. Since LG are dislocated in this allele, the relationship between location of LG and axon pathfinding could be studied. The embryos were doubly stained with anti-Repo and mAb1D4 to simultaneously visualize
both migration of LG and axonal extension.
In the wild-type embryo, LGB, a single precursor
of LG, divides and migrates medially, and associates
with the MP1/dMP2 fascicle at stage 13. Thereafter,
their mediolateral position does not change. In gcm5 mutants, although the number of Repo-positive LG is reduced, their location is relatively
normal. This means that glial migration and the initial orientation of pioneers are not affected in this gcm hypomorphic allele until stage 13.
At stage 14, some glial cells migrate to abnormal positions; some are located laterally and associated with motor nerves where segmental (SNG) or intersegmental
nerve root glia (ISNG) reside at stage 15. Such abnormalities occur in 40% of abdominal hemisegments. These mislocated glial cells are likely to be LG that have migrated abnormally because of their round nuclei characteristic of LG and
expression of LG-specific enhancer trap line 3.267. Longitudinal connectives are frequently disrupted in the segments where LG migrate abnormally. These findings suggest that LG play a crucial role in the guidance of the follower axons. Pathfinding of dMP2 is followed by visualizing axon
patterning and glial cells simultaneously. In wild-type
embryos, dMP2 fascicle is formed and LG align regularly
along the longitudinal tract. Defects are detected in axon bundle formation in the hemisegments where glial cells are missing and pathfinding errors are also found where glial cells are mislocated. The frequency
of these phenotypes is higher in the gcm5 mutant
than in the null gcm1, where no glia are present. Thus, abnormal
location of LG in the gcm5 embryo is the cause, and
not the result, of the pathfinding errors. Therefore, it is
concluded that glial cells are required for correct pathfinding
of the dMP2 axon after fasciculation with the vMP2
axon. In gcm mutants, longitudinal connectives are missing
in some segments even though pioneers must have
connected with the longitudinal tracts. This indicates that
the formation of longitudinal connectives requires not
only correct pathfinding of pioneers but also glial assistance
at later stages. Hidalgo and Booth (2000) reported
that glial cells are required for defasciculation and refasciculation
of pioneer axons. Consistent with this earlier work are results from the current study, observed in various gcm alleles.
Previous studies have also shown that, in the absence of
pioneer neurons, follower axons initially have defects
but eventually correct their trajectory (Lin, 1995).
Observations from the present study demonstrate that association of glial
cells with pioneer axons is a prerequisite for the correct
pathfinding by the follower axons (Takizawa, 2001).
In the embryonic neuromuscular system, about 40 motoneurons
per hemisegment innervate 30 muscles. All the
cells involved are identifiable so they are ideal for the
analysis of neuronal pathfinding. The motor pathway in each
hemisegment consists of two nerve roots, intersegmental
nerve root (ISN) and segmental nerve root (SN). Several glial cells are located
along the motor pathways at positions that coincide with
choice points of pathfinding. For instance, EG glial cells
are shown to be located at the branch point where ISNb
axons depart from the ISN and SBC glia; ISNG and SNG are located on the motor path
between the longitudinal tract and the exit junction. All
of these glial cells are born in the CNS and migrate laterally
to assume their final position in the periphery. During
stage 16 of the gcm null embryo, no Repo-positive cells
exist along the motoneuron pathway. This is because
these presumptive glial cells have failed to migrate to
their normal position. The pathfinding
of motor axons was studied by staining stage 16 wild-type
and gcm null embryos with anti-Repo antibody and
mAb1D4 that stains motor axons. In the gcm null embryo, ISN and SN are separated and do not meet at the exit junction at stage 15 as in
wild-type embryos. Moreover, the
dorsoventral position where the two nerve roots exit the
CNS is abnormal; ISN and SN exit from neuromeres of
about the same dorsoventral level in wild-type embryos,
whereas they exit at different dorsoventral levels in gcm null
embryos. After the motoneurons
exit the CNS, axon bundle formation is also affected
in the peripheral nervous system (PNS); axon
bundles are detached from each other in gcm null mutant
embryos. This indicates that glial cells
are required for the correct pathfinding of motoneurons. Whether these abnormally guided motor
axons are still able to find their correct target muscles was examined. In
gcm null embryos, ISNb is formed and the axons apparently
innervate target muscles in the absence of glial cells; ISNb motor neurons connect with ventral muscles 12 and 13 (Takizawa, 2001).
glial cells missing is the primary regulator of glial cell fate in Drosophila. In addition, gcm has a role in the
differentiation of the plasmatocyte/macrophage lineage of hemocytes. Since mutation of gcm causes only a decrease in
plasmatocyte numbers without changing their ability to convert into macrophages, gcm cannot be the sole determinant of
plasmatocyte/macrophage differentiation. A gcm homolog, gcm2, has been characterized. gcm2 is expressed at low levels in
glial cells and hemocyte precursors. gcm2 has redundant functions with gcm and has a minor role promoting
glial cell differentiation. More significant, like gcm, mutation of gcm2 leads to reduced plasmatocyte numbers. A deletion
removing both genes has allowed the role of these redundant genes in plasmatocyte development to be characterized. Animals
deficient for both gcm and gcm2 fail to express the macrophage receptor Croquemort. Plasmatocytes are reduced in number,
but still express the early marker Peroxidasin. These Peroxidasin-expressing hemocytes fail to migrate to their normal locations and do not complete their conversion into macrophages. These results suggest that both gcm and gcm2 are required together for the proliferation of plasmatocyte precursors, the expression of Croquemort protein, and the ability of
plasmatocytes to convert into macrophages (Alphonso, 2002).
gcm2 expression was followed in embryos using in situ
hybridization to mRNA. Similar results were obtained
whether a full-length probe was used for gcm2 or a probe
made from the first 930 bp of the gcm2 cDNA that ensured
no cross-hybridization with gcm mRNA. Cross-hybridization
with gcm was of concern because the expression
profiles of the two genes overlap, though gcm2 is expressed
at much lower levels. With gcm2 and gcm probes of similar
lengths under identical hybridization conditions, gcm2
transcripts were first detected after 3 h of reaction, while
gcm transcripts were detected after only 5 min of reaction. Like gcm, gcm2 is first detected in an anterior ventral region in stage 5 embryos. During gastrulation, these cells invaginate at the end of the ventral furrow just anterior to the cephalic furrow, in the primordium of presumptive hemocyte precursors. Expression in hemocyte precursors persists
through stage 11, after which it rapidly fades. At late
stage 9, gcm2 is first detected in the neuroectoderm in each
hemisegment in a single cell at the lateral edge of the CNS. At stage 11, gcm2 continues to be expressed in a
single cell per CNS hemisegment, which is now in the
position of neuroblast NB1-3 or its progeny. NB1-3
gives rise to several CNS and peripheral glial. By stage 12, gcm2 is detected in longitudinal glia precursors, as well as other CNS glia, and in a stripe of ectodermal cells of the lateral body wall in each hemisegment. Through stage 15, gcm2 continues to be detected in the longitudinal glia, in other CNS glia at very low levels, and in the lateral ectoderm.
After stage 15, gcm2 expression rapidly fades.
In summary, the expression pattern of gcm2 in part
mirrors that of gcm, but expression is at very low levels
compared with that of gcm. gcm2 expression is highest in the hemocyte primordia and in the longitudinal and nerve root glia (Alphonso, 2002).
The presence of a second factor promoting gliogenesis
explains why a small number of glial cells still develop in
gcm null mutants. gcm2 is expressed at very low levels in
lateral glial cells. The complete deletion of both gcm and
gcm2 results in the elimination of all lateral glial cell development. Ectopic expression of gcm2 induces gliogenesis, and its effectiveness is indistinguishable from gcm. These results suggest that Gcm and Gcm2 proteins have redundant biochemical capabilities, which are likely to be mediated through the similar gcm-motif DNA-binding domains (Alphonso, 2002).
Mutation of gcm2 alone is viable and has little effect on
glial cell differentiation, confirming that the presence of
normal gcm expression is sufficient to carry out gliogenesis
in the absence of gcm2. However, when gcm2 mutation is
crossed with a deletion removing both gcm and gcm2, the
resulting transheterozygote is lethal, and glial cell deficiencies
are visible. Thus, a twofold reduction of gcm and the
complete removal of gcm2 reveal a small contribution of
gcm2 to glial cell differentiation, especially in the LG
lineage, where gcm2 expression is highest. This phenotype
shows that glial cell differentiation is sensitive to the
dosage of gcm and gcm2 gene products. It has been shown that in gcm mutant embryos there is a reduction in the glial expression of gcm2 transcripts; in addition, ectopic gcm expression induces
gcm2 expression, and vice versa. These results suggest that
cross-regulation between the two genes may contribute to
the phenotypes observed (Alphonso, 2002).
Dosage sensitivity and the regulation of gcm2 by gcm
may explain why gcm2 has a weak effect on glial cell
differentiation in gcm mutants. gcm2 is expressed at such
low levels that Gcm2 protein may be at concentrations
below a threshold that triggers glial cell differentiation.
Sporadically it surpasses this threshold in some neural
progenitors, triggering occasional longitudinal glia and
nerve root glia differentiation in the absence of gcm. It has
been proposed that a fairly high threshold for Gcm protein
is required to trigger glial cell differentiation in neuroglioblasts
that give rise to both neurons and glia. In these
neuroglioblasts, low levels of Gcm expression are not
sufficient to trigger glial cell fate, but may be necessary to
confer glial potential when upregulated in daughter cells (Alphonso, 2002).
It is clear that both gcm and gcm2 are required for the
proper differentiation of the plasmatocyte lineage. Mutation
of either gcm or gcm2 results in deficits in plasmatocyte
numbers, and the
ectopic expression of gcm leads to extra plasmatocytes. Previous reports had suggested that gcm specifies the plasmatocyte lineage, and that perhaps the existence of a second gcm-motif gene in Drosophila
accounts for the presence of only reduced number of plasmatocytes in gcm mutants rather than their elimination (Alphonso, 2002).
In light of this, it was surprising to find, in embryos
deficient for both gcm and gcm2, plasmatocyte-like cells
still developing and expressing the early plasmatocyte
marker Peroxidasin. However the number of Peroxidasin-labeled
hemocytes in gcm/gcm2-deficient embryos is reduced
by 60% compared with wild type. This reduction
roughly corresponds to the combined reduction in plasmatocytes
in gcm and gcm2 mutant embryos together (40%
for gcm mutants plus 25% for gcm2 mutants). It has
previously been observed that the procephalic mesoderm,
from which plasmatocytes develop, forms a mitotic domain
that undergoes four divisions during embryonic stages 8-11. After the final division, most procephalic mesoderm cells are recognizable as plasmatocytes
and undergo no further divisions. These
cell divisions are coincident with the highest levels of gcm
and gcm2 expression. These results suggest that gcm and
gcm2 promote the proliferation of plasmatocyte precursors,
rather than their initial specification (Alphonso, 2002).
It is believed that the increased number of plasmatocytes
observed when embryos receive an overexpression of gcm
under control of the heat shock promoter is due to an over-proliferation of plasmatocytes in the procephalic mesoderm, which subsequently migrate
throughout the embryo. Ectopic expression of gcm in the nervous system leads to an increase in Peroxidasin-labeled hemocytes clustered
around the CNS. This phenotype has been ascribed to a transformation
of CNS to hemocyte cell fate. The number of Peroxidasin-labeled hemocytes were counted in sca-Gal4;UAS-gcm embryos, where gcm was ectopically expressed in all neuroblasts, and it was found that the number of plasmatocytes in stage 15 embryos is not significantly increased
(average of 320 Peroxidasin-labeled hemocytes per half embryo), even though their distribution had changed, with more macrophages observed around the CNS. It is thought that this phenotype arises from the recruitment of macrophages to the CNS due to increased apoptosis, rather than transformation of neurons into macrophages. These results show that ectopic expression of gcm transforms presumptive neurons into glial cells but not into plasmatocytes, consistent with combinatorial models of transcription factor action (Alphonso, 2002).
Hemocyte precursors in the embryo give rise to two
populations of blood cells, crystal cells and plasmatocytes, promoted by the
GATA transcription factor Serpent.
Lack of Serpent results in the complete absence of hemocytes. Crystal cell development is promoted by the AML-1 transcription factor homolog Lozenge. In the absence of both gcm and gcm2, the number and location of crystal cells remains the same. It is concluded that gcm and gcm2 do not act as genetic switches between plasmatocyte and crystal cell fate.
However, ectopic expression of Gcm in the crystal cell lineage causes them to express Croquemort and assume plasmatocyte-like morphology showing that there is some plasticity between hemocyte lineages (Alphonso, 2002).
The persistent expression of Serpent suggests that it may
continue to have a role in plasmatocyte development after
the initial specification of hemocyte precursors. In addition, a second factor, U-shaped, is expressed in plasmatocytes. U-shaped
acts to limit the proliferation of crystal cells. In the absence
of U-shaped, there is an increase in the crystal cell population,
and forced expression of U-shaped reduces the crystal
cell population. It is suggested that a
combination of Serpent, U-shaped, lack of Lozenge expression,
and perhaps unidentified factors promotes the initial
specification of plasmatocytes and Peroxidasin expression (Alphonso, 2002).
The results are consistent with a requirement of gcm and
gcm2 for the conversion of plasmatocytes into macro
phages. Deletion of both gcm and gcm2 results in the
complete absence of Croquemort expression in hemocytes.
These mutant hemocytes retain Peroxidasin expression and
have some migratory properties characteristic of plasmatocytes;
some move out of the procephalon and are ultimately distributed in the hemolymph. However, many of these mutant hemocytes remain in the head; those that do migrate, migrate dorsally, and do not follow stereotypic
ventral paths along the surface of the CNS. Normal plasmatocytes
migrate to prominent sites of programmed cell death. The distribution of plasmatocytes in embryos lacking gcm and gcm2 functions suggests that they are not attracted to cells undergoing programmed cell death or that their ability to migrate is impaired (Alphonso, 2002).
In addition, the majority of mutant plasmatocytes fail to
enlarge, nor do they contain large vacuoles filled with dark
inclusions. This phenotype suggests that most mutant
plasmatocytes lack phagocytic activity. The possibility cannot be ruled out
that they are incabable of phagocytosis; the fact that a small
number of mutant plasmatocytes enlarge slightly may
indicate that they are competent for phagocytosis. Plasmatocytes
mutant for Croquemort protein are impaired in
their ability to engulf apoptotic cells, but will still recognize
and engulf bacteria. Whether gcm/gcm2-deficient plasmatocytes will convert into macrophages in response to bacterial invasion or other insults
remains to be explored. The morphology of these cells and
high level of Peroxidasin expression suggest characteristics
of plasmatocytes just before their conversion to
macrophages. A model is suggested whereby gcm and gcm2 promote the expansion and differentiation of a Peroxidasin-expressing population of
hemocytes, or 'proplasmatocytes'. Expression of gcm and
gcm2 is required for the differentiation of these proplasmatocytes
into plasmatocytes that are competent for
macrophage conversion, and they do so by initiating the
expression of the macrophage receptor Croquemort and
other genes that promote macrophage morphogenesis and function (Alphonso, 2002).
Locomotion relies on stable attachment of muscle fibres to their target sites, a process that allows for muscle contraction to generate movement. gcm and gcm2, the fly glial cell determinants, are expressed in a subpopulation of embryonic tendon cells and required for their terminal differentiation. By using loss-of-function approaches, it has been shown that in the absence of both genes, muscle attachment to tendon cells is altered, even though the molecular cascade induced by stripe, the tendon cell determinant, is normal. Moreover, gcm activates a new tendon cell gene independently of stripe. Finally, segment polarity genes control the epidermal expression of gcm and determine, within the segment, whether it induces glial or tendon cell-specific markers. Thus, under the control of positional cues, gcm triggers a new molecular pathway involved in terminal tendon cell differentiation, which allows the establishment of functional muscle attachment sites and locomotion (Soustelle, 2004).
gcm and gcm2 are required in segment border tendon
cells. By contrast to the nervous system, where these genes are transiently
expressed and play a role in fate choice, their transcripts are present in
tendon cells until the end of embryogenesis. Moreover, the gcm complex is
not controlled by the stripe tendon cell fate determinant and is not
necessary for the expression of stripe targets. Finally, the gcm
complex does not trigger tendon cell differentiation. Thus, gcm and
gcm2 do not play a role in fate choice in the epidermis (Soustelle, 2004).
Several observations indicate that the gcm complex controls terminal
tendon cell differentiation and thereby affects attachment site integrity.
(1) gcm-gcm2 embryos do express genes (delilah and
ß1tubulin) that are activated by the establishment of
muscle-tendon cell contact and only show muscle defects after stage 16, once these
contacts have been established. (2) gcm-gcm2 tendon cells and
muscles are unable to form a functional attachment site, as shown by muscle
midline crossing and defective hemiadherens junctions (HAJs). (3) Tendon cell-specific inactivation of gcm complex activity results in massive muscle disorganisation and defective locomotion. Interestingly, locomotion defects have also been observed in larvae mutant for flapwing, a gene encoding a phosphatase 1ß, known for its role in muscle attachment maintenance. It
will be interesting to determine whether gcm complex targets represent
Flapwing substrates, even though it is clear that Flapwing also affects a
gcm-independent pathway since it is expressed and required both in
tendon cells and muscles (Soustelle, 2004).
Although Stripe represents the tendon cell fate determinant, it is
expressed until the end of embryogenesis, suggesting an additional, late,
role. Indeed, while early inactivation of the stripe pathway (stage
11) by the dominant negative approach and null mutations induce the same
phenotype of detached muscles, late inactivation (stage 15) induces a muscle
midline crossing phenotype similar to the one observed in
gcm-gcm2 embryos. Thus, the gcm complex and stripe are
necessary for terminal tendon cell differentiation and probably act on common
targets such as alien. Interestingly, the gcm complex also controls
its own pathway, independently of stripe. Understanding the
relative contribution of the two pathways will require the identification of
the stripe and gcm targets. Finally, the observation that
tendon cell-specific mutations alter muscle organisation highlights the
importance of cell-cell interactions throughout muscle and tendon cell
development. Identifying mutations affecting either cell type will help elucidate the bases of such interactions (Soustelle, 2004).
gcm expression is controlled along the antero-posterior axis via
wingless (wg) and hedgehog (hh) signalling
pathways. Hh- and Wg-secreted proteins are expressed anterior to segment
border cells; therefore, a non-autonomous process controls gcm
expression in the epidermis. Hh and Wg act respectively via Cubitus
interruptus (Ci) and Pangolin (Pan) transcription factors,
which bind to the promoter of stripe and activate its expression.
Since the first 6 kb of the gcm promoter that are sufficient to
induce epidermal expression contain
no predicted sites for Pan and Ci, Hh and/or Wg signalling pathways probably
act on gcm indirectly (Soustelle, 2004).
While dorso/ventral (D/V) patterning contribution to stripe
expression and tendon cell differentiation is not yet elucidated, a microarray
study has identified gcm as a target of Dorsal, the maternally provided D/V patterning factor.
However, no Dorsal binding sites are found in the gcm promoter,
suggesting that this regulation is also indirect. Altogether, the results show
that stripe and gcm expression is mutually independent and
associated with segment border cell identity. Therefore, positional cues
trigger the expression of Gcm and Stripe, which in turn control several
aspects of tendon cell differentiation. Detailed analyses of the
gcm promoter will allow identification of the transcription factors
upstream from gcm in tendon cells (Soustelle, 2004).