InteractiveFly: GeneBrief

Vsx1 and Vsx2: Biological Overview | References

BIOLOGICAL OVERVIEW

The central neuroendocrine system in the Drosophila brain includes two centers, the pars intercerebralis (PI) and pars lateralis (PL). The PI and PL contain neurosecretory cells (NSCs) which project their axons to the ring gland, a complex of peripheral endocrine glands flanking the aorta. This paper presents a developmental and genetic study of the PI and PL. The PI and PL are derived from adjacent neurectodermal placodes in the dorso-medial head. The placodes invaginate during late embryogenesis and become attached to the brain primordium. The PI placode and its derivatives express the homeobox gene Dchx1 (Vsx1) and can be followed until the late pupal stage. NSCs labeled by the expression of Drosophila insulin-like peptide (Dilp), FMRF, and myomodulin form part of the Dchx1 expressing PI domain. NSCs of the PL can be followed throughout development by their expression of the adhesion molecule FasII. Decapentaplegic (Dpp), secreted along the dorsal midline of the early embryo, inhibits the formation of the PI and PL placodes; loss of the signal results in an unpaired, enlarged placodeal ectoderm. The other early activated signaling pathway, EGFR, is positively required for the maintenance of the PI placode. Of the dorso-medially expressed head gap genes, only tailless (tll) is required for the specification of the PI. Absence of the corpora cardiaca, the endocrine gland innervated by neurosecretory cells of the PI and PL, does not affect the formation of the PI/PL, indicating that inductive stimuli from their target tissue are not essential for early PI/PL development (de Velasco, 2007).

The insect neuroendocrine system consists of several populations of neurosecretory cells (NSCs) with peripheral axons terminating in contact with specialized neurohemal glands where the neurohormones are released. The majority of NSCs are found in the dorso-medial protocerebrum, the so-called pars intercerebralis (PI) and pars lateralis (PL). These NSCs project their axons towards a set of small glands, the corpora cardiaca (CC), and corpora allata (CA). In Drosophila, the CC and CA, along with a third neuroendocrine gland, the prothoracic gland (PTG), are fused into a single complex, the ring gland, which surrounds the anterior tip of the aorta. The PI-PL/ring gland complex of insects has been repeatedly compared to the hypothalamus-pituitary axis in vertebrates, based on clear similarities between the two, anatomically and functionally (i.e., their shared role in energy metabolism, growth, water retention, and reproduction; reviewed in Nassel, 2002). Previous studies of the insect neuroendocrine system have focused on the neurosecretory cells of the PI and PL; however, not much is known about the different types of non-secretory PI neurons, and even less information exists about the development of this important part of the insect brain. This paper focused on the formation of the PI during Drosophila embryonic and larval development (de Velasco, 2007).

The PI is histologically recognized as the unpaired antero-medial domain of the protocerebral cortex that is located anterior of the calyces of the mushroom bodies and dorsal of the central complex. Beside innervating the neuroendocrine glands and thereby acting as the uppermost center of endocrine release, the neurites of PI neurons are structurally and functionally integrated into the medial compartments of the protocerebrum. Some of the cells of the PI were shown to play a role in various behaviors, including locomotor activity and flight behavior. Moreover, an early role of certain PI cells as pioneer neurons during the formation of protocerebral axon tracts has also been reported (de Velasco, 2007).

The extremely limited information about the development of the PI stems mainly from studies on grasshopper and Drosophila. For grasshopper, the dorso-medial subpopulation of brain neuroblasts, measuring approximately 20 in number, was tentatively assigned to the formation of the PI, although detailed lineage studies have not yet been carried out. Three groups of cells that form part of the adult PI have been followed in more detail from embryonic stages onward (Ludwig, 2002). These include (part of) the NSCs, a group of unpaired cells derived from a cell called the 'dorsal midline progenitor', and a small set of early differentiating neurons expressing the antigen TERM-1 that act as pioneers of the brain commissure (de Velasco, 2007).

Cells situated along the antero-dorso-medial edge of the population of brain neuroblasts can be assumed to produce neuronal offspring that gets incorporated into the PI, although specific lineages have not yet been followed. The neurectoderm of the so-called 'head midline', which gives rise to at least part of (if not the entire) PI was described as a specialized region with morphological and molecular similarities to the ventral (trunk) midline (mesectoderm). Both trunk and head midline domains contain neuronal progenitors that do not delaminate as individual cells like the neuroblasts of the lateral neurectoderm, but that invaginate as an elongated furrow (ventral midline) or as several separate placodes (head midline). Furthermore, both trunk and head midline require the activity of EGFR signaling for survival and fate specification (Dumstrei, 1998). Another molecular characteristic shared by the midline of the trunk and head is the extended expression of the neurogenic genes of the E(spl) complex. These genes are activated in the neurectoderm (in general) by Notch signaling, and are responsible for inhibiting proneural gene expression, thereby mediating the lateral inhibition process that delimits the number of neuroblasts delaminating at any given position from the neurectoderm. At the stage when no more neuroblasts are born, Notch signaling and the expression of E(spl) transcripts cease. This happens around stage 11 in the lateral neurectoderm of the trunk and the head. However, in the midline domains, expression continues far beyond that stage, up until stage 14 (Dumstrei, 1998). This finding was interpreted to indicate a temporally extended neurogenic potential of the midline cells. In other words, neurectodermal cells along the head midline are still neurogenic at a stage when other ectodermal cells are already on their way to become epidermal cells (de Velasco, 2007).

A comprehensive developmental study of the Drosophila pars intercerebralis and pars lateralis has been undertaken with the aim of understanding their embryonic origin from the dorsal head midline, developmental morphogenesis, and architecture. The PI, marked by its continued expression of the homeobox gene Dchx1, is derived from a placode located anteriorly in the neurectoderm of the dorso-medial head. Three additional placodes are situated posterior to the PI placode; one, marked by the expression of FasII, gives rise to a cluster of neurons that form the pars lateralis (PL), which represents the second domain within the brain that contains neurosecretory cells (Siegmund, 2001). The remaining two placodes express the homeobox gene drx (Davis, 2003) and become part of the protocerebum surrounding the PI and PL. This study further investigated the role in PI development of early acting signaling pathways (Dpp, EGFR) and transcription factors expressed in the dorso-medial head neurectoderm [tailless (tll), ventral nerve cord defective (vnd), single minded (sim); orthodenticle (otd)] and in the corpora cardiaca [sine oculis (so), glass (gl), forkhead (fkh)] in the PI placode. The findings provide a developmental genetic framework for the study of the Drosophila central neuroendocrine system (de Velasco, 2007).

Pars intercerebralis: structure and development during the postembryonic phase: To analyze the PI in Drosophila larvae, pupae, and adult, brains were used in which GFP was driven by Chat-Gal4, which is expressed in most differentiating neurons, including their axons. This staining visualized most, if not all neuronal cell bodies and their proximal axons that comprise the PI. The PI of the adult and 3 day pupa appears as an unpaired cluster of cell bodies filling the cleft between the brain hemispheres. At most positions along the antero-posterior axis, it is well demarcated from the lateral cortex by a glial lamella. The following will distinguish the anterior PI (PIa), located around the medial lobes of the mushroom body, from the central PI (PIc; dorsal to the central complex) and posterior PI (in between central complex and protocerebral bridge). Most PIa neurons are small and weakly Chat-GFP-positive; in addition, one can distinguish a subset of large, strongly Chat-GFP-positive somata with axons that pass through the superior medial protocerebrum and form the NccI nerve that projects to the ring gland. Based on their trajectory, these large neurons comprise the neurosecretory cells of the PI. The central PI (PIc) is comprised of neuronal somata that form a rather homogenous population in regard to size and Chat-GFP expression levels. Axons of PIc neurons turn laterally and appear to branch in the posterior part of the superior-medial protocerebrum (called DP compartment in the larva). Neurons of the posterior PI (PIp) are small and densely packed; many PIp neurons form axons that fasciculate in a fiber tract projecting straight ventrally and then anteriorly into the fan-shaped body. One cannot define a clear boundary between the PIp and the laterally/posteriorly adjacent protocerebral cortex (de Velasco, 2007).

It is clear from the above said that the PI is formed by neurons that contribute to many different brain circuits, and that much work is needed to elucidate the exact structure and function of these cells. Do all of these cells share a common origin? In other words, can one define one or more specific neural lineages whose cells exclusively contribute to the PI? And is there an ontogenetic relationship between the PI and the PL, the second cluster of neurosecretory cells that, without specific markers, cannot be recognized within the brain of the larva, pupa or adult? To approach these questions, the development of the PI was followed backward in time. In the early pupa and larva, all components of the PI defined in the previous section for the late pupa and adult can be recognized; however, they do not form an unpaired cluster in the brain midline, but are split into bilaterally symmetric clusters. Thus, as an anatomical entity, the pars intercerebralis does not exist in the larva, but evolves during the pupal period (de Velasco, 2007).

In the one day-old pupa, cells with the characteristics of the anterior PI (strongly Chat-GFP-positive neurons located dorsal of the medial lobes) form axons that branch in the DA and CA compartments and continue peripherally as the NccI nerve (Ncc stands for nervus corporis cardiaci). In the central PI, large, strongly Chat-GFP-positive cells cover the developing ellipsoid body and fan-shaped body. The same picture presents itself in the wandering larva. Besides their large size and high expression level of Chat-Gal4, neurons of the PIa and PIc are distinguished from the lateral cortex by lacking any Neurotactin expression. Neurotactin expression defines a distinct, transient stage in neural differentiation. Thus, all primary neuroblasts and neurons of the late embryonic brain express neurotactin, to then lose it in the early larva. When neuroblasts become reactivated and produce secondary neurons, these also express Neurotactin until mid-pupal stages. The fact that the larval PIa/c has no Neurotactin-positive cells says that this domain consists only of primary neurons, or is characterized by a distinct mode of Neurotactin expression that sets it apart from other neurons. In either case, it serves as a suitable global marker of the larval PIa/c. Using the neuron-specific marker Elav supports the view that the PIa/c consists mainly of primary neurons which can be distinguished from undifferentiated secondary neurons by their large cell and nuclear size: anti-Elav reveals large neurons in the PIa/c which contrast sharply from the small secondary neurons of the laterally adjacent cortex (de Velasco, 2007).

As expected from its adult and late pupal morphology, the posterior PI (PIp) is difficult to define with any confidence in the larval brain. Neurons in the posterior-medial cortex, around the emerging commissure of the lateral horn which served as a topological landmark for the PIp at later stages, have axons directed anteriorly towards the central complex primordium. These neurons, which should include the cells comprising the later PI, belong to several lineages of secondary neurons of the DPM group. Without more specific markers, it is not possible to ascertain exactly which DPM lineages produce the posterior PI, and whether these lineages contribute only to the PI, or to laterally adjacent parts of the brain as well (de Velasco, 2007).

Definition of the PIa/c as a expression domain of the homeobox gene Dchx: The vertebrate gene Chx10 (also called Vsx; Levine, 1994) appears in the anlage of the forebrain during neurulation and is later expressed and required for neurons of the retina. The Drosophila genome contains two closely associated Chx10 homologs, Dchx1 and Dchx2 (Vsx2). Their expression profile shows similarities to the expression of their vertebrate counterparts. Dchx1 appears in the early anlage of the visual system (the optic lobe placode/optic anlage of the late embryo and larva), as well as in differentiating neurons of the optic lobe. In addition, Dchx1 is expressed in a dorso-medial domain of the brain that overlaps with the PIa/c, and at least in part, the PIc as defined above. Thus, large Dchx-positive primary neurons are located anterior and dorsal of the medial lobe of the mushroom body and are flanked by the secondary lineages of the DAM and DPMl groups, respectively. Based on these criteria, the Dchx-positive neurons are recognized as the larval PIa/c as described above. Dchx-positive primary neurons of the PIa/c can be followed backward in time through early larval stages. Posterior to the group of large PIa/c neurons, a cluster of small Dchx-positive cells are found that overlap with the group of DPMm lineages. It is speculated that these neurons represent the primordium of the posterior PI. As expected, the Dchx-positive secondary neurons of the PIp are born in the late larva, when DPM lineages (along with most other lineages of the central brain) proliferate. In early larval and embryonic brains, only few scattered Dchx-positive cells appear in the postero-medial brain (de Velasco, 2007).

The expression of GFP driven by a Dchx promoter construct (Landgraf, 2003) visualizes the projection pattern of the Dchx-positive neurons populating the PI. Most, if not all of these cells conform to a relatively simple commissural pattern, whereby axons cross in the anterior part of the supraesophageal commissure. Axonal and dendritic branches form a dense plexus that fills out the DA and CA compartments. A subset of PI neurons, notably those that co-express Drosophila insulin-like peptide (Dilp; Rulifson, 2002) form axons that leave the dorso-medial brain and project to the ring gland, as well as the subesophageal ganglion/tritocerebrum (de Velasco, 2007).

Neurosecretory cells targeting the ring gland lie within the PIa/c: The larval PI includes most, if not all, of the neurosecretory cells that project towards the corpora cardiaca/corpora allata through the NccI. Larval brains were labeled with antisera against Dilp, FMRFamide, and myomodulin. In all of these experiments, peptidergic neurons with axons to the ring gland were located in the Dchx-positive/neurotactin-negative PIa/c domain. Peptidergic axons branch in the CA/DA compartments, cross the midline, and then extend posterior, passing underneath the supraesophageal commissure. This nerve connection constitutes the NccI nerve. The axonal marker FasII reveals the NccI has already developed in the late embryo (Nassif, 2003). A second connection between the dorso-lateral protocerebrum and the ring gland, called NccII, also expresses FasII from embryonic stages onward. The FasII-positive neurons that give rise to the NccII are not part of the PI, but instead turn out to represent the Pars lateralis (PL), which occupies a position in the dorsal brain cortex, anteriorly adjacent to the calyx of the mushroom body. The NccII root emitted by the PL forms a conspicuous tract of the larval brain that passes the calyx and peduncle medially before reaching the medial edge of the protocerebrum where it joins the NccI on its way to the ring gland (de Velasco, 2007).

The PI and PL are derived from a series of neuroepithelial placodes in the embryonic head: The Dchx-expressing cells of the PIa/c can be followed into the early embryonic period when they form a narrow cluster of approximately 40-50 cells in the antero-medial procephalon, right behind the furrow that separates the procephalon from the clypeolabrum. Two similar-sized domains are labeled by FasII and the Drosophila Rx homolog, Drx. Together, these three markers define linearly arranged, non-overlapping domains along the dorsal midline of the procephalon. The Dchx domain will be called 'embryonic PIa/c'. During later embryogenesis, a global morphogenetic movement shifts all of these domains posteriorly; at the same time, the FasII and Drx-positive domains also move laterally. Dchx remains strongly expressed in a cohesive cluster that develops into the larval PIa/c. It is currently unclear whether the secondary neuroblast(s) that later give rise to the neurons of the PIp are derivatives of the PIa/c, or whether they are recruited from outside this domain. In late embryos, faintly Dchx-positive neurons and neuroblasts can be recognized outside the PIa/c (not shown) (de Velasco, 2007).

Cells of the FasII-positive domain give rise to the NccII and therefore constitute the embryonic primordium of the PL. In the late embryo, shortly before the somata of these neurons lose FasII expression, one can follow a FasII-positive axon tract from the FasII cluster towards the primordium of the ring gland. This tract represents the embryonic NccII, given that it exhibits the same topographical characteristics as the NccII of the larva (point of origin in postero-lateral protocerebrum; passing the mushroom body towards medially; joining nascent NccI towards ring gland) (de Velasco, 2007).

The PI, PL and Drx domains visible in the mid-stage embryo (stages 11-13) form neuroepithelial placodes that split by invagination from the procephalic ectoderm. This mode of neurogenesis differs from the mechanism of delamination that produces all other brain neuroblasts within the surrounding neurectoderm. The use of anti-Crumbs (anti-Crb) as a marker for apical membrane domains of epithelial cells reveals a stereotypic pattern of small invaginations in the dorso-medial procephalon of stage 12/13 embryos. The anterior invagination ('1') is formed/surrounded by Dchx-positive cells and corresponds to the PIa/c; the second, intermediate invagination ('2') lies in the FasII-positive PL, and two posterior invaginations, one medially ('3'), one further laterally ('4'), flank the Drx domain. During stage 13/14, the invaginations pinch off the surface ectoderm and form small vesicles in between the outer epithelium (the nascent head epidermis) and the brain surface (de Velasco, 2007).

Two additional markers, a reporter construct of the E(spl)m5 gene and the neuronal differentiation marker Elav, were used to further characterize the placodeally derived PI/PL and Drx domain. The 'dorsal head midline', a domain that, along with the midline of the trunk, is set apart from the lateral neurectoderm by the prolonged expression of the Notch target E(spl), as well as the expression of/dependence on EGFR signaling (Dumstrei, 1998). Triggered by Notch activation, E(spl) is expressed in the neurectoderm as long as this layer is 'active', i.e., produces neuroblasts. The phenomenon of prolonged E(spl) expression was interpreted as an indication for an extended period of neurogenic potential (de Velasco, 2007).

The previously defined dorsal head midline coincides with the PIa/c and PL, as evident from the double labeling experiments using the E(spl)-m5 reporter, anti-FasII and anti-Crb. E(spl)m5 is expressed in these domains throughout late embryogenesis into the early larval period. Furthermore, double-labeling with anti-Elav demonstrates that neuronal differentiation begins relatively late in these domains. Throughout stages 12, 13 and much of 14, the E(spl)m5-positive domains remain Elav-negative; Elav signal comes up faintly during stage 15, and only in the stage 16 embryo do most neurons of these domains express Elav (de Velasco, 2007).

The placodeally derived, Crb-positive vesicles remain visible at the brain surface until late embryogenesis (stage 16). Subsequently, the cells lose Crb expression; it is assumedd that the intermediate and posterior vesicles ('2-4') convert into neurons that become incorporated in the brain, although this needs to be definitively shown. The anterior pair of vesicles ('1') appears to become incorporated into the corpora allata, the dorsal most part of the ring gland. The early development of this endocrine gland which secretes juvenile hormone during larval stages has not been clearly documented for Drosophila. In other insects, ectodermal placodes that invaginate from the ectoderm of gnathal segments were observed to give rise to both prothoracic glands (source of ecdysone) and corpora allata. The present data, which need to be substantiated by additional markers, indicate that at least part of the CA is derived from the PIa/c. Thus, in the late embryo, the anterior vesicles, still expressing Crb and E(spl)m5, approach each other and eventually fuse in the midline, forming a cluster of cells that moves posteriorly, behind the level of the brain commissure, and becomes incorporated into the dorsal part of the ring gland (de Velasco, 2007).

Genetic specification of the PI: The dorsal procephalic ectoderm that gives rise to the PI is patterned by several signaling pathways, in particular the Dpp and DER pathway. Furthermore, the head gap genes tailless (tll) and orthodenticle (otd) are expressed in the dorsal procephalon. The Tll expression domain includes the PI/PL placodes; these placodes, on the other hand, are not part of the Otd domain. The midline determinant Single minded (Sim) is expressed faintly from stage 13 onward in a subset of cells in the PI primordium. Ventral nerve cord defective (Vnd) is expressed in a longitudinal medial domain of the ventral neurectoderm flanking the mesectoderm. The paramedian stripe of Vnd expression continues in the head, but ends laterally/ventrally of the PI/PL (de Velasco, 2007).

Loss of Dpp results in the absence of the dorso-medial head epidermis that in wild-type separates the bilateral PI anlagen from each other. The PI anlagen are enlarged and fused in the dorsal midline. In DER mutants, Dchx expression is virtually absent, supporting the previously reported finding (Dumstrei, 1998) that the dorso-medial procephalic ectoderm ('dorsal head midline') requires DER signaling. Dchx is also eliminated in tll mutants, but is present, if possibly reduced, in embryos mutant for otd and other head gap genes (de Velasco, 2007).

Aside from these genes, which are expressed in the PI anlage, the possibility was tested of inductive interactions between the PI and neighboring tissues, in particular the foregut and stomatogastric nervous system (which transiently contacts the dorso-medial procephalic ectoderm) and the corpora cardiaca, which receive axons from the PI derived neurons. To remove the foregut, a null mutation was used in the fkh gene; the stomatogastric nervous system was eliminated by a mutation in so, and the corpora cardiaca by a mutation in gl. The results indicate that none of these genetic manipulations grossly affects the formation of the PI. The gl null mutation survives until the late larval period, which gave an the opportunity to analyze the structure and innervation of the ring gland that lacks the corpora cardiaca. Both Dilp and FasII-positive axons reach the ring gland and send axons towards the corpora allata, the dorsal part of the ring gland, which is unaffected in gl mutants. The only abnormality in ring gland innervation was the defasciculation and aberrant projection of FasII and FMRFamide-positive axons. In particular, FasII-positive axons frequently followed Dilp fibers onto the dorsal vessel, a behavior not observed in wild-type. The number of FMRFamide-positive axons reaching the corpora allata was reduced (de Velasco, 2007).

Definition of the pars intercerebralis and pars lateralis in Drosophila: According to the classical anatomical definition that applies to adult insect brains, one can recognize the PI as an unpaired cluster of neuronal cell bodies located along the dorsal midline of the brain. In its anterior and intermediate part (PIa/c), the pars intercerebralis is clearly set apart from the adjacent lateral cortex by a glial lamella; at posterior levels, the boundary between the PI and neighboring cortex domains is fluid. The PI is comprised of several hundreds of neuronal cell bodies that include as a relatively small minority large NSCs with axonal projections to the ring gland. The PI as an unpaired midline structure appears first during the late pupal phase. Before that stage, cells of the PI form bilateral clusters in the dorso-medial cortex of both brain hemispheres. Two criteria allowed recognition the PI at these stages and follow it backward into the embryonic period. One was the expression of the homeobox gene Dchx1, the other the idiosyncratic proliferatory properties of the PI primordium. Dchx1 is expressed in bilateral placodes in the antero-medial neurectoderm of the early embryonic head. These placodes form the early primordium of the PI. In the late embryo, they move interiorly and become part of the dorso-medial brain cortex. During later larval stages, the PI primordium, aside from the continued expression of Dchx1, sets itself apart from the lateral cortex by the absence of stem cell-like neuroblasts producing secondary lineages. Instead, the PI primordium appears to grow, at a rather slow rate, by symmetric cell division. During metamorphosis, the cortex and neuropile of the brain hemispheres fuse, giving rise to the unpaired median PI and central complex (de Velasco, 2007).

As often in development, it is difficult to state with any certainty whether the boundaries of the PI as defined in the adult coincide precisely with those visible earlier. In other words, it cannot be presently stated with any certainty that the sharp PI boundary defined, in the adult and late pupal brain, by glial septa coincides with the Dchx-positive cluster of the embryo and larva. It can be said, however, that many, if not all, of neurosecretory cells of the classically defined PI fall within the Dchx-expressing cell cluster. This was demonstrated in this study for Dilp, FMRFamide, and myomodulin; it can be said confidently that the NSCs of the PI expressing other peptides (Siegmund, 2001) will also be included within the Dchx-positive domain (de Velasco, 2007).

The pars lateralis (PL) has been defined as a cluster of NSCs that lie outside the PI, and whose peripherally projecting axons form the NccII. The NccII can be recognized from early embryonic stages onward by its expression of the adhesion molecule FasII. The FasII-positive cells that give rise to the NccII, and that therefore should be considered as the primordium of the PL, are derived from a neurectodermal placode located posteriorly adjacent to the placode that forms the PI. It may be significant that both PI and PL, the centers including neurosecretory cells, are derived from placodeal neurectoderm (de Velasco, 2007).

A third set of placodes appear in the dorso-medial protocerebral neurectoderm, posterior to the FasII-positive PL placode. The posterior placodes, at least partially, overlap with the main expression domain of the homeobox gene Drx. Drx-positive cells forming within this region can be followed into the larval stage. They spread out over a relatively large area of the dorso-posterior cortex. The expression of Drx may be significant given the fact that the vertebrate homolog of this gene is expressed and required in the primordium of the hypothalamus. However, in Drosophila, the relationship of the posterior placodes and their Drx-positive derivatives to the PI/PL is not clear. It is possible that non-ring gland-associated NSCs are derived from it; in addition, the drx-positive neurons may be functionally and anatomically closely connected to the PI/PL (de Velasco, 2007).

Origin of PI/PL from neurectodermal placodes: This paper has shown that the PI/PL originate as placodes from the dorso-medial neurectoderm of the head, in a way that is similar to the formation of the optic lobe. In all of these cases, small domains of the neurectoderm are seen which, during stages 10 or early 11 of development, adopt the shape of placodes, with cells elongating in the apico-basal axis and expressing a higher level of apical markers such as Crb at their apical surface. Eventually, all of these placodes invaginate and sever their connection to the ectoderm several hours after their initial appearance (stages 12-13). Subsequently, cells of the placodes lose their epithelial phenotype and directly turn into neural cells (as in the case of the PI placode, or the SNS placodes), or give rise to neuroblasts (as in the case of the optic lobe). In addition, during the interval between their first appearance and invagination, the placodes give rise to 'early neural progenitors' which delaminate from the surface and move inside. For example, the optic lobe placode gives rise to at least four neuroblasts that delaminate during stage 11 and then proliferate, like all other neuroblasts, in a stem cell-like manner. Similarly, individual neurons delaminate from the SNS placodes before these structures invaginate. It is considered likely that the dorso-medial placodes described in this study also give rise to several neuroblasts (de Velasco, 2007).

Most neurons of the insect brain are formed as part of fixed lineages, each lineage being produced by a stem cell-like neuroblast. Neuroblasts delaminate as individual cells, or small clusters of cells, from the neurectoderm, leaving behind other cells that then become specified and differentiate as epidermal cells. This peculiar mode of neural cell birth and proliferation is a derived feature found in insects and many crustaceans; it is not present in taxa considered basal in the arthropods, and taxa outside the arthropods. Early neurogenesis in chelicerates, myriapods and chilopods have recently been analyzed, and led to the interesting discovery that in these animals, the neurectoderm produces a large array of small placodes which subsequently invaginate and (after some additional rounds of mitosis) turn into the neurons and glial cells of the ventral nerve cord and brain. In terms of number and pattern, the placodes are comparable to the array of neuroblasts in insects, leading to the speculation that one might be able to define homologies between individual placodes and neuroblasts. To go a step further, one could speculate that at the root of arthropods, the neurectoderm was subdivided into a mosaic of small domains, each of which invaginated as a placode to then give rise to a specific part of the CNS. In time, this mode of neurogenesis was supplanted by the 'invention' of stem cell-like neuroblasts: instead of the entire placode invaginating, a single (or a few) cell(s) was selected from the placode at an early stage which then delaminated and continued to proliferate in an asymmetrical, stem cell-like manner. If this interpretation of neuroblasts vs. placodes among arthropods is correct, one would have to conclude that the occurrence of placodes along the ventral midline and head midline (as well the stomatogastric nervous system and optic lobe) of insects represents the phylogenetically older mode of neurogenesis. Likewise (and this notion is of course even more speculative), one could argue that molecular mechanisms at work in these placodes or the function of brain parts derived from them is phylogenetically more ancient compared to structures developing from neuroblasts. The same rationale has traditionally been put forward to argue that the 'fringe domains' of the cerebral cortex, including the archicortex (hippocampus) and paleocortex (entorhinal cortex) constitute the phylogenetically older regions of the mammalian brain (de Velasco, 2007).

Pars intercerebralis and hypothalamus: Similarities between the neuroendocrine system of vertebrates and arthropods on the structural, functional and developmental level have been emphasized in many previous studies. In both vertebrates and arthropods, the highest command center of the neuroendocrine system is comprised of groups of NSCs located in the brain; these cells, besides innervating brain centers and thereby influencing neural circuits as 'neuromodulators', send their axons to peripheral neurohemal glands in which the hormones produced by the NSCs are stored and released. In vertebrates, neurosecretory cells are located in the hypothalamus. The endocrine gland they act upon is the pituitary. The corresponding structures in arthropods would be the PI/PL and their peripheral targets, the CC/CA, respectively. The main hormone produced by the CC is adipokinetic hormone (AKH) that mobilizes lipids and carbohydrates from the fat body (see Van der Horst, 2001). AKH shares common functions with the vertebrate hormone glucagon that is produced in endocrine cells of the pancreas, as well as peptidergic neurons in the brain. AKH also shows some sequence similarity with the N-terminus of glucagons. Similar and possibly homologous to the relationship between Drosophila insulin-like peptides and AKH (Kim, 2004), the function of glucagon in vertebrates is antagonized by insulin. Insulin itself is expressed like glucagon in the endocrine pancreas, but a whole family of insulin-like growth factors is found in hypothalamic NSCs. Other neuropeptides found in NSCs of hypothalamus and PI alike are FMRFamides and tachykinins. Also, the sequence similarity between vertebrate CRF and insect CRF-like diuretic hormone deserves attention in this context. Here a scenario might be considered in which ancestrally a peptide directly exerted a diuretic effect, a condition maintained in the arthropod line of evolution; in the line of evolution leading up to chordates, other hormones (ACTH, aldosterone) were 'interpolated' between the original peptide (CRF) and the action on excretory cells (de Velasco, 2007).

Developmental similarities between vertebrate hypothalamus and arthropod PI are also strong. The anlagen of the pituitary and hypothalamus are neighboring structures within the anterior neural plate. Cells that will give rise to the anterior lobe of the pituitary (adenohypophysis) are anteriorly adjacent to the cells which will become the hypothalamus and the posterior pituitary. Numerous signals were found to be involved in delimiting the anlage of the neuroendocrine system within the anterior neural plate; they include Sonic hedgehog (Shh), members of the bone morphogenetic protein family (BMP7), and fibroblast growth factor family (FGF4/8). Among the molecular determinants that are switched on by these signaling pathways are the homeobox genes six3/6 and Rx, the paired-box genes pax6 and Nkx2.1/2, the PAS-bHLH gene sim1, and the orphan nuclear receptor Tlx. Genes acting further downstream in determining specific hypothalamic cell fates are POU III-related homeobox genes Brn-1, 2, and 4. It has been shown that loss-of-function mutant mice lacking Brn-2 do not develop part of the hypothalamus; Sim1 knock-out mutations in mice cause a similar phenotype to that one for Brn-2 (de Velasco, 2007 and references therein).

This study has shown that homologs of three of the transcription factors expressed in the anlage of the vertebrate hypothalamus also appear in or adjacent to the anlage of the Drosophila PI/PL: the Nkx2.1/2 homolog vnd, the Sim1 homolog sim, and the Rx homolog Drx. In addition, the Six3/6 homolog optix also appears to overlap with the PI/PL of the st.11 embryo. The role of these Drosophila genes in PI development awaits further study. A previous paper (DeVelasco, 2004; see Embryonic development of the corpus cardiacum, a component of the ring gland) has shown that the Six1 homolog sine oculis (so) is expressed and required for the formation of the CC and the ontogenetically closely related stomatogastric nervous system. In contrast, counterpart(s) of Dchx1, the gene that is expressed strongly and continuously throughout Drosophila PI development, apparently do not play a role in neuroendocrine development in vertebrates. Detailed expression studies of Chx10/Vsx1 documented that this gene is expressed in numerous tissues outside of the retina, notably the ventral hindbrain, the diencephalic-mesencephalic boundary, and the epithalamus (Liu, 1994). However, neither expression nor function of Chx10/Vsx1 in the ventral diencephalon has been reported (de Velasco, 2007).

In conclusion, this study presents evidence for a number of conserved properties in the way the progenitors of the central neuroendocrine system in vertebrate and Drosophila embryos are spatially laid out, and employ cassettes of signaling pathways and fate determinants. One may speculate that there existed in the common bilaterian ancestor a simple anterior brain with which sensory afferents and groups of neurosecretory cells were associated. These cells might have played pivotal roles in feeding behavior (olfactory/gustatory perception of food sources; feed back information from the intestinal tract and body cavity regarding the degree of urgency of feeding) and reproductive behavior, and could have evolved into the much more complex neuroendocrine systems that is found in today's highly derived bilateria, such as insects and vertebrates (de Velasco, 2007).

The origin of islet-like cells in Drosophila identifies parallels to the vertebrate endocrine axis

Single-cell resolution lineage information is a critical key to understanding how the states of gene regulatory networks respond to cell interactions and thereby establish distinct cell fates. This study identified a single pair of neural stem cells (neuroblasts) as progenitors of the brain insulin-producing neurosecretory cells of Drosophila, which are homologous to islet β cells. Likewise, a second pair of neuroblasts was identified as progenitors of the neurosecretory Corpora cardiaca cells, which are homologous to the glucagon-secreting islet α cells. Both progenitors originate as neighboring cells from anterior neuroectoderm, which expresses genes orthologous to those expressed in the vertebrate adenohypophyseal placode, the source of endocrine anterior pituitary and neurosecretory hypothalamic cells. This ontogenic-molecular concordance suggests that a rudimentary brain endocrine axis was present in the common ancestor of humans and flies, where it orchestrated the islet-like endocrine functions of insulin and glucagon biology (Wang, 2007).

The principal insulin producing-cells (IPCs) in higher metazoans, such as flies and mammals, direct organismal growth, metabolism, aging, and reproduction via a conserved signal transduction pathway. Gut- or pancreas-based IPCs, with endodermal origin, emerged as the principal IPC locus with the evolution of lower vertebrates such as the jawless fish. In contrast, the principal IPCs of invertebrates are found in the nervous system and are likely of ectodermal origin. Despite this difference, the possibility that gene regulatory modules may be conserved for cell fate programming the principal IPCs of all higher animals, irrespective of germ layer origin, has led the development of islet-like cells to be addressed in Drosophila (Wang, 2007).

Brain IPCs in Drosophila were first recognized by their expression of insulin (Drosophila insulin-like peptide, Dilp2) at the end of embryonic development. The goal of this work was to understand the developmental origin of these cells. The absence of morphological and vital markers for identifying brain neuroblasts for dye-labeled lineage tracing necessitated the combined use of mosaic analysis to demonstrate lineage relationships and immunohistology to follow cell identities. In this study, 16 molecular lineage markers corresponding to conserved genes were used to follow cells in fixed embryos. To identify genes involved in early IPC lineage development, before the differentiation of IPCs, 650 transposable GAL4-transgene insertions, obtained from public collections, that reported gene enhancer activity (GAL4 enhancer traps) in the CNS, were screened. Enhancer-driven GAL4 activity was used to trigger heritable and irreversible lineage labeling, which was assayed for coexpression with Dilp2 in late larval brains, thereby identifying lineage markers and potential developmental determinants. It was found that enhancers near the genes dachshund (dac), eyeless (ey), optix, and tiptop (tio) each triggered IPC lineage labeling by the time of Dilp2 expression onset just before hatching (late-stage 17). tio enhancer-triggered labeling was highly specific to the IPCs within the pars intercerebrallis (PI), the dorsomedial brain region harboring the IPCs and other neurosecretory cells. Antibody staining of Dac, Ey, and Optix proteins recapitulated enhancer reporter labeling and revealed expression in the tio⁺ cell cluster in late-stage embryos just after IPC differentiation, and before IPC differentiation at early-stage 17. Thus, a bilateral cluster of 10-12 Dac⁺ Ey⁺ cells were identified, 6-8 of which expressed tio before continuing on to express insulin (Dilp2) slightly later in development (Wang, 2007).

The hypothesis was tested that the Dac⁺ Ey⁺ cluster is generated by the proliferation of a single neuroblast. The pre-Dilp2 Dac⁺ Ey⁺ cluster comprised 10-12 cells at stage 17, but only a single Dac⁺ cell at stage 12, suggesting that a lineage expanded from a single progenitor beginning at stage 12. The Dac⁺ cluster maintains a posterior and lateral position within the anterior PI, identified by dChx1 expression, which allows following it during the morphogenetic changes in the developing brain. To mark progenitors and their lineage descendants, stage 11-12 embryos harboring both a heat-shock promoter-flip recombinase (hsp70-flp) transgene and an FRT-mediated flip-out Actin promoter-LacZ reporter were heat-shocked to induce random clone marking events in cell lineages. After aging embryos for 6 h at 25°C to reach stage 16-17, marked clusters of clonally related cells were occasionally recovered that comprised the 10-12 cell Dac⁺ Ey⁺ cluster. Clones that partly labeled the Dac⁺ Ey⁺ cluster, which were posterior in the cluster, were interpreted as being labeled by a lineage marking event induced after the neuroblast had divided one or more times. It was unlikely that multiple marking events accounted for the apparent clonal labeling of IPCs because the frequency of marked clone induction was extremely low (tens per brain). Clones were also found that labeled neighboring cells, but do not label Dac⁺ Ey⁺ cells, suggesting there is a lineage restriction that defined the Dac⁺ Ey⁺ cluster. Thus, all data are consistent with a lineage model whereby one neuroblast produced 10-12 Dac⁺ Ey⁺ cells, 6-8 of which were IPCs (Wang, 2007).

Whether the single Dac⁺ cell progenitor of IPCs seen at stage 12 was indeed a neuroblast was further tested by using markers of neuroblast lineage development. Asymmetrically dividing neuroblasts can be identified by nuclear expression of the pan-neuroblast marker Deadpan (Dpn) and Prospero (Pros) localization to the plasma membrane. It was found that the single Dac⁺ cell expressed Dpn and also showed Pros localization at the plasma membrane, which indicated that it was a neuroblast. As the Dac⁺ cluster increased in cell number with age, it was found that Pros was present in the nucleus of Dac⁺ cells anterior to the Dac⁺ neuroblast, which indicated that these were the neuroblast daughter cells, or ganglion mother cells (GMCs) generated by asymmetric neuroblast divisions. By stage 14, the most anterior Dac⁺ cells in the cluster lacked Dpn and Pros, suggesting that they were early, undifferentiated neurons or neurosecretory cells generated by GMC cell divisions. It was also found that tio expression occurs in the most anterior Dac⁺ cells of the lineage group, furthest from the posterior-located Dac⁺ neuroblast, suggesting that the six to eight IPCs are the products of the first three to four GMCs to be generated by asymmetric neuroblast division. This observation confirmed the interpretation of the marked clone data that showed partial labeling by a clone occupies the posterior, more recently formed region of the Dac⁺ Ey⁺ cluster, near the IPC neuroblast. Thus, a histological pattern of cell identities and divisions within the Dac⁺ IPC lineage group was observed that was consistent with the generic lineage development of a single neuroblast, with the IPCs being produced from the first three to four GMCs formed (Wang, 2007).

Further attempts were made to identify the precise origin of the IPC neuroblast within the neuroectoderm epithelium and the blastoderm embryo to place this lineage in the context of early axial patterning. The IPC neuroblast was first recognized by Dac expression only after neuroblast formation, but before its first division. However, preceding the formation of the IPC neuroblast, the markers Castor (Cas) and dChx1 and the proneural factor Lethal of Scute (L'Sc) showed coexpression in eight nearby cells of the neuroectoderm epithelium. Cas and dChx1 were maintained in all neuroblast lineages that delaminated from this group, as indicated by coexpression of Dpn. The IPC neuroblast was the only neuroblast from this group to express Dac, and it was always the first Dpn⁺ neuroblast to delaminate, becoming the most posterior in a chain of delaminating Cas⁺ dChx1⁺ neuroblasts. The Cas⁺ dChx1⁺ L'Sc⁺ proneural group lies within a 'gap gene' head stripe corresponding to the Bicoid responsive giant head stripe 1 (gt1), which suggested that the IPC neuroblast, or its earliest progenitor, arose from this pattern element of the precellular blastoderm (Wang, 2007).

β Cell and α cell development in mammals shares a largely common pathway. Thus attempts were made to study the origin of the α-like cells in Drosophila and their development relative to the IPC lineage. Corpora cardiaca (CC) cells are analogous in function to islet α cells (Kim, 2004). These neuroendocrine cells reside in the endocrine ring gland, just dorsal to the brain. CC cells produce and secrete a glucagon-like peptide, adipokinetic hormone, in response to circulating glucose levels, via a conserved K_atp sensor. The gene glass (gl) is a marker of CC cells and their precursors that specifically labels the CC lineage beginning at stage 10. The Gl⁺ group of cells expands in number to form a bilateral pair of six to eight cell clusters, aligned at the border of the brain and the developing foregut (stage 13). The Gl⁺ clusters then migrated out of the protocerebrum (stage 14), and posterior along the roof of the pharynx, to ultimately coalesce at the midline within the prospective ring gland (stage 16). Remarkably, the first Gl⁺ cells appeared a single cell diameter apart from the dChx1⁺ cluster containing the IPC neuroblast, also within the gt1 stripe (Wang, 2007).

These results suggested that the CC cell lineage, like the IPC lineage, is also generated from a progenitor within the gt1⁺ dorsal neuroectoderm. Indeed, a neuroblast progenitor for CC cells was suggested by expression of a Kruppel reporter (Kr-GFP) found to specifically label the Gl⁺ cells and an adjacent cell that both was Dpn⁺ and showed membrane localized Pros, indicating that it was a neuroblast. As for IPCs, tests were made to see if CC cells are derived from a single progenitor, perhaps the Kr-GFP⁺ neuroblast. Gl⁺ β-gal⁺-marked clones were recovered that comprised all or part of a CC cell cluster, after their migration to the prospective ring gland at stage 16. Because labeled CC cells had moved from their point of origin in the developing PI, it could not be determine whether a progenitor also produced other cells besides the CC cells, which did not similarly migrate. Together, these observations suggest that the CC cells are related by lineage to a neuroblast progenitor (Wang, 2007).

Typically, neuroblasts inherit the expression of cell specification factors from their point of origin in the patterned neuroectoderm before the neuroblast forms. It was found that this was the case with the IPC neuroblast, which retains dChx1 and Cas expression from the neuroectoderm. It was therefore hypothesized that this may also be the case for the CC cell neuroblast. CC cell specification was shown to require the function of gt, sine oculis (so), twist (twi), and snail (sna). Indeed, it was found that all of these factors are expressed in the Gl⁺ CC cell lineage. Moreover, the Kr-GFP⁺ cell group, containing the neuroblast and CC cell precursors, also expressed Eyes absent (Eya), the cognate protein tyrosine phosphatase of So. It was subsequently found that at stage 10, the time that Gl⁺ cells are first detected, a region of gt1⁺ neurectoderm shows expression of So. It was also found that one to two So⁺ gt1⁺ neuroblasts can be detected by labeling with Dpn at this stage. Thus, it is proposed that the So⁺ Eya⁺ gt1⁺ neuroectoderm gives rise to the Kr-GFP⁺ So⁺ Eya⁺ gt1⁺ neuroblast, which is the single progenitor of the CC cells (Wang, 2007).

The model of a dorsal neurectoderm origin for CC cells is in disagreement with another extant model. The anterior ventral furrow (AVF) epithelium was suggested to be the CC cell origin based on gene expression and function studies implicating So, Gt, Twi, and Sna in CC cell formation. To distinguish between the AVF and dorsal neuroectoderm as possible origins of CC cells, two newly available gt promoter fragment reporters were used whose expression persists late enough in development, beyond endogenous protein and transcript expression, to serve as a coarse-grain lineage marker of CC cells. The AVF is marked by the gt23 reporter, whose expression is limited to the two gt head stripes posterior to gt1 at the blastoderm stage. This reporter does not label the Gl⁺ cells. However, as has been shown, the Gl⁺ cells arise in the context of the most anterior gt head stripe, gt1, which reaffirms the proposed origin from the gt1⁺ neuroectoderm (Wang, 2007).

The organization of this gt1⁺ segment-derived proendocrine neuroectoderm was investigated with respect to the conserved factors Optix, So, Eya, and dChx1. Optix and Eya expression aligned with the gt1 reporter expression domain. The D-six4 gene also shows expression specific to this domain. Labeling studies showed that this domain is subdivided into several small compartments of 2-12 cells with discrete gene expression profiles. The data indicate that the IPC neuroblast was derived from compartment B (Optix⁺, dChx1⁺, Cas⁺, So^-, low-level Eya) and the CC cell neuroblast arose from the adjacent compartment C (Optix⁺, So⁺, Eya⁺, dChx1^-). This somewhat surprising finding suggests that the largely common developmental pathway of β and α cells may be partly conserved in Drosophila, perhaps with respect to a domain of Sine oculis/Six family and Eya gene expression (Wang, 2007).

The early expression of the mouse ortholog of the Drosophila homeodomain gene optix, Six6, demarcates the hypophyseal placode and infundibular region, which give rise to the anterior pituitary and neurosecretory hypothalamus, respectively. Mutation of the Six6 gene leads to reduction of the pituitary in mice and humans. The hypophyseal placode and adjacent ectoderm also expresses the other so-called 'placode genes,' Six1, Six4, and Eya, and this coexpression pattern is conserved in amphibians, fish, and lower chordates such as ascidians. In mice, the anterior pituitary is reduced in size in the double mutant of Eya1 and Six1, and in zebrafish, Eya1 is essential for differentiation of all pituitary cell types except for prolactin-expressing cells. In Drosophila, So and Eya are essential for CC cell formation. Thus, there is a striking conservation of the molecular signature of tissues that give rise to elements of the brain endocrine axis in flies, mammals, lower vertebrates, and lower chordates (Wang, 2007).

There are also parallels between vertebrate and fly with respect to tissue morphogenesis within the developing brain endocrine system and adjacent oral ectoderm, although there appears to be considerable variation on a general theme. For example, in mouse, the progenitors of the anterior pituitary and neurosecretory hypothalamus appear to arise respectively from Rathke's pouch, an invagination of the oral ectoderm, and the neurectoderm, which do not start as neighboring regions, but come into direct contact only after neurulation. However, in the zebrafish, which does not form a Rathke's pouch, the progenitors of the anterior pituitary and neurosecretory hypothalamic cells (GnRH1⁺) arise from neighboring regions of the hypohyseal placode, which is situated directly dorsal to the stomodeal ectoderm. In Drosophila, the ventral cells of the gt1⁺ Optix⁺ Eya⁺ ectoderm invaginate to form the roof of the pharynx, the fly's oral ectoderm, whereas the dorsal cells contribute to the endocrine axis. Therefore, there is considerable evidence for evolutionarily conservation of the close relationship between the oral ectoderm and the developing compartments of the endocrine axis, all of which express the hypophyseal placode genes. The gene expression profile and specification of endocrine cell functions from the anterior ectoderm appears to be more 'fixed' across the bilateria, whereas the pattern of accompanying tissue morphogenesis and diversity of cell types is more variable, just as has been demonstrated for the specification of the bilaterian CNS, eye, gut, and heart (Wang, 2007).

The model proposed in this study contrasts with the prior suggestion, based on the proximity of developing CC cells to the posterior foregut in the moth, Manduca, that CC cells originate from neurogenic placodes of the foregut that engender the stomatogastric nervous system. Because CC cell progenitors were not identified in those studies, and subsequent mutational analysis in Drosophila demonstrated that the CC cells develop independently of the stomatogastric nervous system and posterior foregut, it is suggested that the current model of CC cell origin is the most strongly supported (Wang, 2007).

It is proposed that the brain endocrine systems of invertebrates and vertebrates are derived from a common ancestry because they both develop from a domain of Eya and sine oculis/Six family gene expression that comprises the anterior neuroectoderm and adjacent oral ectoderm. Indeed, these results extend prior observations that the neurosecretory cells of the PI and ring gland show other aspects of homology to the hypothalamic-pituitary axis. The specification of islet-like cells within a conserved brain endocrine axis raises the intriguing possibility that islet organogenesis, which is a derived feature of vertebrates, may have coopted brain endocrine cis-regulatory modules for specification of islet fates in endoderm. Indeed, the ectopic expression of the nominal rat insulin promoter reporter in anterior pituitary and hypothalamus underscores the similar gene regulatory state of these endocrine tissues. It is expected that further genetic analysis of endocrine cell fate specification within the gt1 domain of Drosophila will lead to insights into the patterning and organogenesis of endocrine compartments and provide the basis for identifying conserved pan-IPC regulatory modules with relevance to mammalian systems (Wang, 2007).

Conserved role of the Vsx genes supports a monophyletic origin for bilaterian visual systems

Components of the genetic network specifying eye development are conserved from flies to humans, but homologies between individual neuronal cell types have been difficult to identify. In the vertebrate retina, the homeodomain-containing transcription factor Chx10 is required for both progenitor cell proliferation and the development of the bipolar interneurons, which transmit visual signals from photoreceptors to ganglion cells. This study shows that dVsx1 and dVsx2, the two Drosophila homologs of Chx10, play a conserved role in visual-system development. DVSX1 is expressed in optic-lobe progenitor cells, and, in dVsx1 mutants, progenitor cell proliferation is defective, leading to hypocellularity. Subsequently, DVSX1 and DVSX2 are coexpressed in a subset of neurons in the medulla, including the transmedullary neurons that transmit visual information from photoreceptors to deeper layers of the visual system. In dVsx mutant adults, the optic lobe is reduced in size, and the medulla is small or absent. These results suggest that the progenitor cells and photoreceptor target neurons of the vertebrate retina and fly optic lobe are ancestrally related. Genetic and functional homology may extend to the neurons directly downstream of the bipolar and transmedullary neurons, the vertebrate ganglion cells and fly lobula projection neurons. Both cell types project to visual-processing centers in the brain, and both sequentially express the Math5/ATO and Brn3b/ACJ6 transcription factors during their development. These findings support a monophyletic origin for the bilaterian visual system in which the last common ancestor of flies and vertebrates already contained a primordial visual system with photoreceptors, interneurons, and projection neurons (Erclik, 2008).

The eyes of vertebrates and flies, despite their very different structures, rely on a shared regulatory network of genes (such as Pax6/eyeless) for their specification. This led to the proposal that invertebrate and vertebrate eyes are derived from a common ancestor. It is not known, however, whether homologies exist between the cell types that make up the neural circuit required for vision, although it has been suggested that the vertebrate eye is a composite structure made up of distinct types of photoreceptive cells of independent evolutionary origins (Erclik, 2008).

In the vertebrate retina, transmission of the signal from photoreceptors to ganglion cells is mediated by bipolar interneurons. The neurons of the retina develop from a pool of proliferating and multipotent neuroepithelial cells known as the retinal progenitor cells (RPCs). A suite of transcription factors regulates retinal proliferation and neurogenesis. One of these (the homeodomain- and CVC-domain-containing transcription factor Chx10) is required for both RPC proliferation and bipolar cell specification (Burmeister, 1996; Erclik, 2008 and referencestherein).

In the mouse retina, Chx10 is expressed in the RPCs as well as in developing and mature bipolar cells (Liu, 1994). Chx10 mutant retinas are hypocellular because of a marked decrease in the rate of RPC proliferation, and bipolar cells are completely absent (Burmeister, 1996). Additionally, Chx10 mutant eyes are micropthalmic, suggesting that Chx10 nonautonomously controls the growth of the entire eye. The role of Chx10 in eye development is conserved in zebrafish and in humans; null mutations in Chx10 have been identified in families with congenital micropthalmia (Ferda Percin, 2000). Vertebrates contain a second Vsx-family gene, Vsx1, and in the mouse, retina expression of Vsx1 is restricted to a subset of differentiating and mature cone bipolar cells, where it is required for their late differentiation and function (Ohtoshi, 2004; Chow, 2004). In C. elegans, the Vsx homolog ceh-10 is required for the development of the AIY thermoregulatory interneuron (Svendsen, 1995). Because the AIY neuron may be directly targeted by a photoreceptive cell in some nematodes, it has been suggested that the Vsx genes play a conserved role in photoreceptor target-interneuron development that extends from humans to worms (Erclik, 2008 and references therein).

This study characterized the role of the two Drosophila homologs of Chx10 in visual-system development. The Drosophila visual system is made up of two major structures: the compound eye and the underlying optic lobe, which is located in the brain. The photoreceptors in the compound eye project to either the first or second ganglion of the optic lobe, respectively the lamina and the medulla. Thus, in contrast to vertebrates, in which the photoreceptors and their target interneurons are located together in the retina, in the fly, the target neurons of the photoreceptors are found in the brain. Transmedullary neurons then relay visual information to the third optic ganglion, the lobula complex. In turn, neurons in the lobula complex project from the optic lobe to the higher-order visual-processing centers of the brain (Erclik, 2008).

The compound eye and optic lobe develop from distinct epithelial structures in the larva. The photoreceptor cells are derived from the eye imaginal disc, and the optic lobe develops from two groups of neuroepithelial progenitor cells, termed the inner and outer optic anlagen (the IOA and OOA, respectively). Whereas the genetic control of neurogenesis has been intensively studied in the eye disc, comparatively little is known about the genes required for progenitor proliferation and neuronal cell-type specification in the developing optic lobes (Erclik, 2008).

This study demonstrates that the two Drosophila homologs of Chx10, dVsx1 and dVsx2, are required for optic-lobe development. DVSX1 is expressed in a population of multipotent and proliferating OOA progenitor cells, and both DVSX1 and DVSX2 are expressed in the developing and mature medulla, which derives from the OOA. dVsx-positive neurons in the medulla include both local neurons and transmedullary neurons that directly connect the inner photoreceptors, R7 and R8, to the lobula complex. In dVsx mutants, the optic lobes are very small, the medulla being most strongly affected. The hypocellularity observed in the adult optic lobes results from an early defect in OOA progenitor proliferation. The expression pattern and mutant phenotypes observed for the dVsx genes in the fly visual system are similar to those previously described for Chx10 in the vertebrate retina. On the basis of several functional and molecular similarities, it is proposed that (1) the progenitors of the OOA are evolutionarily related to the retinal progenitors of vertebrates, and (2) the photoreceptor target interneurons and visual-projection neurons of flies and vertebrates share several homologous features (Erclik, 2008).

Full-length cDNAs encoding two Drosophila Chx10 homologs, dVsx1 and dVsx2, were cloned from embryonic and adult head cDNA libraries, respectively. Their predicted proteins possess a domain structure characteristic of all Vsx homologs (Chow, 2001) : a paired-like homeodomain and an adjacent CVC domain (a putative DNA-binding domain, which defines the family). DVSX1 and DVSX2 each share 81% amino acid identity with Chx10 across the homeodomains and CVC domains. Unlike DVSX1, DVSX2 also possesses a C-terminal OAR domain, which is believed to function in transcriptional repression and which is present in a subset of Vsx homologs (Bernier, 2001; Erclik, 2008).

Sequence alignment and phylogenetic analyses suggest that DVSX1, DVSX2, and the C.elegans Vsx protein, CEH-10, are ancestrally more closely related to Chx10 than the second vertebrate Vsx protein, Vsx1. This finding suggests that the dVsx genes are the product of a duplication event from an ancestral Vsx gene, which occurred after the split of chordates, nematodes, and arthropods. Consistent with the theory of a duplication event, dVsx1 and dVsx2 are located at the same locus on the X chromosome (polytene bands 5A3-7), separated by a 33 kb intergenic region (Erclik, 2008).

DVSX1, but not DVSX2, is embryonically expressed in the pars intercerebralis, ventral nerve cord, pharynx, optic anlage, and isolated cells throughout the brain. Because Chx10 in vertebrates is expressed in, and required for, the proliferation of retinal progenitor cells, focus was placed on the expression pattern of DVSX1 in the optic-lobe progenitors. In the embryo, DVSX1 expression in the visual system was first observed in the preinvagination optic placode. After optic-placode invagination, DVSX1 expression was detected in approximately 12 cells in the center of each optic anlage. The progenitors of the optic anlagen remain quiescent until midway through the first larval instar, when they segregate into two morphologically distinct columnar epithelia, the IOA and OOA, and commence symmetric divisions. Shortly after the resumption of proliferation, DVSX1 expression was detected in approximately nine cells in a central region of the OOA, and colabeling with a phosphohistone H3 (PH3) antibody indicated that these cells are mitotically active. In the third-instar larva, when the majority of optic progenitors convert to asymmetrically dividing neuroblasts that generate the neurons of the adult optic lobe, DVSX1 expression remained restricted to a small central population of OOA epithelial progenitors. A second domain of DVSX1 expression also became evident in a subset of cells in the nascent medulla cortex (Erclik, 2008).

To establish the DVSX1-positive progenitor cells' contribution to the optic lobe, a cell-lineage-tracing experiment was performed with the MzVUM-Gal4 driver, which faithfully reports for DVSX1 expression MzVUM-Gal4 is a P element inserted 886 bp upstream of dVsx2 that drives Gal4 in dVsx-expressing cells. In this experiment, DVSX1-expressing cells and their progeny are permanently marked with lacZ expression. In the third-instar larval brain, β-Galactosidase (β-Gal) expression was detected throughout the outer optic lobe, including in cells of the OOA, outer neuroblasts, lamina, and medulla, but not in the lobula complex or eye disc. Thus, DVSX1-positive cells in the OOA are competent to give rise to cells throughout the outer optic lobe (Erclik, 2008).

The above data demonstrate that DVSX1 is (1) expressed in proliferating and uncommitted OOA progenitors, (2) downregulated as OOA progenitors become neuroblasts, and (3) subsequently upregulated in the medulla cortex. This expression pattern strongly resembles that of Chx10 in vertebrates; Chx10 is expressed in multipotent and proliferating neuroretinal progenitors, downregulated before neurogenesis, and then upregulated in bipolar interneurons (Liu, 1994; Passini, 1997). Of note, one significant difference between the visual-system expression patterns for the two genes is that whereas DVSX1 expression is restricted to a central subset of OOA progenitors, Chx10 is expressed in all RPCs (Erclik, 2008).

DVSX1 is expressed in a central region of the OOA, where it is required for progenitor cell proliferation. The spatial restriction of DVSX1 to the center of the OOA is unexpected because it has been assumed that the cells of the OOA constitute a homogeneous population of symmetrically proliferating progenitors. The data indicate that molecular and, perhaps, functional differences are present among OOA progenitor cells as early as the stage 16 embryo. It is possible that the DVSX1-positive central OOA cells and the DVSX1-negative peripheral OOA cells differ with respect to their cell-cycle kinetics and/or the neuronal populations that they generate. In vertebrate RPCs, Chx10 regulates the expression of the cyclin-dependent kinase inhibitor, p27Kip1, and CyclinD1 (Green, 2003). It will be important to determine whether dVsx1 regulates the expression of the Drosophila homologs dacapo and Cyclin D in the progenitors of the OOA (Erclik, 2008).

dVsx expression was observed in the transmedullary neurons that arborize in the M3 and M6 layers, as well as those that arborize in layers targeted by the lamina monopolar neurons, suggesting that the dVsx genes may function in both the color-vision and motion-detection pathways. A complete list of dVsx-positive transmedullary neurons will be required to determine whether they act as a bridge for all visual signals, as Chx10-positive bipolar cells do in the vertebrate retina. It is notable that not all dVsx-positive neurons project out of the medulla. Some are amacrine or intrinsic neurons that make only local connections within the medulla. These neurons are functionally analogous to the horizontal and amacrine neurons of the vertebrate retina, cell types in which Chx10 is not expressed. It is proposed that the dVsx-positive amacrine and intrinsic cells may have evolved from an ancestral population of dVsx-positive transmedullary neurons. The observation that sets of transmedullary and local neurons exhibit highly similar arborization patterns supports the hypothesis that these neurons may be ancestrally related (Erclik, 2008).

Expression of the dVsx genes is absent in the lamina, which houses the target neurons of the outer photoreceptors R1-R6. The lack of dVsx expression in the lamina may be a product of the evolutionary history of the Drosophila photoreceptors. Phylogenetic and ontogenic considerations suggest that R8 represents the ancestral photoreceptor (Meinertzhagen, 1991). Because R8 terminates in the medulla, the fly neurons that are most closely related to the ancestral photoreceptor target neurons may be found in the medulla, not the lamina (Erclik, 2008).

Debate regarding the phylogenetic origins of the bilaterian visual system is ongoing. The traditional view has held that eyes have evolved multiple times, an assertion that is supported by the existence of eyes with different morphologies, photoreceptor types, and embryonic origins. This view has been challenged by the surprising observation that morphologically disparate eyes rely on a conserved network of genes for their formation. Recently, the polyphyletic view has been argued with data from Platynereis, a model for the last common ancestor of bilateria. It has been proposed that the insect eye evolved from the rhabdomeric photoreceptors found in the eye of the last common ancestor, whereas the vertebrate eye evolved from the population of ciliary photoreceptors found in its brain. The rhabdomeric photoreceptors were then assimilated into the evolving retina, where they became ganglion cells (Erclik, 2008).

The identification of cell-type homologies between the vertebrate and fly visual systems supports a monophyletic origin for bilaterian visual systems. It is proposed that the last common ancestor of flies and vertebrates, Urbilateria, contained Vsx-positive interneurons that relayed information from a mixture of ciliary and rhabdomeric photoreceptors to projection neurons that targeted the brain. These neuronal populations developed from progenitor cells dependent on the Vsx genes for their proliferation. In the evolutionary line leading to vertebrates, the rhabdomeric photoreceptors were lost, and the remaining ciliary photoreceptors and their neuronal targets developed together in the retina. In contrast, in the line leading to arthropods, the ciliary photoreceptors were lost, and the progenitors for the rhabdomeric photoreceptors and the deeper neuronal layers were separated during development. Consequently, the photoreceptors in arthropods develop in the ectoderm, whereas the progenitors for their target neurons invaginate and develop in the brain. The progenitors for the visual system interneurons remained Vsx dependent, whereas those for the photoreceptors and the projection neurons became Vsx independent (Erclik, 2008).

The eyes of several extant organisms contain a mixture of ciliary and rhabdomeric photoreceptors, including the Brachiostoma cerebral vesicle and the larval eye of Pseudoceros. Thus, it is possible that the eyes of Urbilateria contained a mixture of both photoreceptor types and that the situation in Platynereis is derived rather than ancestral. The proposal that genetic and functional homology extends to the lobula projection neurons and ganglion cells is at odds with the suggestion, based partly on the shared expression of melanopsin, that vertebrate ganglion cells and fly rhabdomeric photoreceptors are ancestrally related cell types. Because only a very small subset of ganglion cells express melanopsin, which is also expressed in several other retinal cell types, expression of this photopigment in rhabdomeric photoreceptors and ganglion cells may be coincidental. It will be important to determine where the homologs of the Vsx, Lhx, Math5, and Brn3b genes are expressed in extant models of Urbilateria, such as Platynereis or the flatworm Macrostomum sp (Erclik, 2008).

Optix defines a neuroepithelial compartment in the optic lobe of the Drosophila brain

During early brain development, the organisation of neural progenitors into a neuroepithelial sheet maintains tissue integrity during growth. Neuroepithelial cohesion and patterning is essential for orderly proliferation and neural fate specification. Neuroepithelia are regionalised by the expression of transcription factors and signalling molecules, resulting in the formation of distinct developmental, and ultimately functional, domains. This study discovered that the Six3/6 family orthologue Optix is an essential regulator of neuroepithelial maintenance and patterning in the Drosophila brain. Six3 and Six6 are required for mammalian eye and forebrain development, and mutations in humans are associated with severe eye and brain malformation. In Drosophila, Optix is expressed in a sharply defined region of the larval optic lobe, and its expression is reciprocal to that of the transcription factor Vsx1. Optix gain- and loss-of-function affects neuroepithelial adhesion, integrity and polarity. Restricted cell lineage boundaries were found that correspond to transcription factor expression domains. It is proposed that the optic lobe is compartmentalised by expression of Optix and Vsx1. These findings provide insight into the spatial patterning of a complex region of the brain, and suggest an evolutionarily conserved principle of visual system development (Gold, 2014).

Integration of temporal and spatial patterning generates neural diversity

In the Drosophila optic lobes, 800 retinotopically organized columns in the medulla act as functional units for processing visual information. The medulla contains over 80 types of neuron, which belong to two classes: uni-columnar neurons have a stoichiometry of one per column, while multi-columnar neurons contact multiple columns. This study shows that combinatorial inputs from temporal and spatial axes generate this neuronal diversity: all neuroblasts switch fates over time to produce different neurons; the neuroepithelium that generates neuroblasts is also subdivided into six compartments by the expression of specific factors (see The OPC neuroepithelium is patterned along its dorsal-ventral axis). Uni-columnar neurons are produced in all spatial compartments independently of spatial input; they innervate the neuropil where they are generated. Multi-columnar neurons are generated in smaller numbers in restricted compartments and require spatial input; the majority of their cell bodies subsequently move to cover the entire medulla. The selective integration of spatial inputs by a fixed temporal neuroblast cascade thus acts as a powerful mechanism for generating neural diversity, regulating stoichiometry and the formation of retinotopy (Erclik, 2017).

The optic lobes, composed of the lamina, medulla and the lobula complex, are the visual processing centres of the Drosophila brain. The lamina and medulla receive input from photoreceptors in the compound eye, process information and relay it to the lobula complex and central brain. The medulla, composed of ~40,000 cells, is the largest compartment in the optic lobe and is responsible for processing both motion and colour information. It receives direct synaptic input from the two colour-detecting photoreceptors, R7 and R8. It also receives input from five types of lamina neuron that are contacted directly or indirectly by the outer photoreceptors involved in motion detection (Erclik, 2017).

Associated with each of the ~800 sets of R7/R8 and lamina neuron projections are 800 medulla columns defined as fixed cassettes of cells that process information from one point in space. Columns represent the functional units in the medulla and propagate the retinotopic map established in the compound eye. Each column is contributed to by more than 80 neuronal types, which can be categorized into two broad classes. Uni-columnar neurons have arborizations principally limited to one medulla column and there are thus 800 cells of each uni-columnar type. Multi-columnar neurons possess wider arborizations, spreading over multiple columns. They compare information covering larger receptor fields. Although they are fewer in number, their arborizations cover the entire visual field (Erclik, 2017).

The medulla develops from a crescent-shaped neuroepithelium, the outer proliferation centre (OPC). During the third larval instar, the OPC neuroepithelium is converted into lamina on its lateral side and into medulla neuroblasts on its medial side. A wave of neurogenesis moves through the neuroepithelial cells, transforming them into neuroblasts; the youngest neuroblasts are closest to the neuroepithelium while the oldest are adjacent to the central brain. Neuroblasts divide asymmetrically multiple times to regenerate themselves and produce a ganglion mother cell that divides once more to generate medulla neurons. Recent studies have shown that six transcription factors are expressed sequentially in neuroblasts as they age: neuroblasts first express Homothorax (Hth), then Klumpfuss (Klu), Eyeless (Ey), Sloppy-paired 1 (Slp1), Dichaete (D) and Tailless (Tll). This temporal series is reminiscent of the Hb --> Kr --> Pdm --> Cas --> Grh series observed in Drosophila ventral nerve cord neuroblasts that generates neuronal diversity in the embryo. Indeed, distinct neurons are generated by medulla neuroblasts in each temporal window. Further neuronal diversification occurs through Notch-based asymmetric division of ganglion mother cells. In total, over 20 neuronal types can theoretically be generated using combinations of temporal factors and Notch patterning mechanisms. However, little is known about how the OPC specifies the additional ~60 neuronal cell types that constitute the medulla (Erclik, 2017).

To understand the logic underlying medulla development, late larval brains were stained with 215 antibodies generated against transcription factors and 35 genes were identifiied that are expressed in subsets of medulla progenitors and neurons. The OPC neuroepithelial crescent can be subdivided along the dorsal-ventral axis by the mutually exclusive expression of three homeodomain-containing transcription factors: Vsx1 is expressed in the central OPC (cOPC), Rx in the dorsal and ventral posterior arms of the crescent (pOPC), and Optix in the two intervening 'main arms' (mOPC). These three proteins are regionally expressed as early as the embryonic optic anlage and together mark the entire OPC neuroepithelium with sharp, non-overlapping boundaries. Indeed, these three regions grow as classic compartments: lineage trace experiments show that cells permanently marked in the early larva in one OPC region do not intermingle at later stages with cells from adjacent compartments. Of note, Vsx1 is expressed in cOPC progenitor cells and is maintained in a subset of their neuronal progeny whereas Optix and Rx are not expressed in post-mitotic medulla neurons. The OPC can be further subdivided into dorsal (D) and ventral (V) halves: a lineage trace with hedgehog-Gal4 (hh-Gal4) marks only the ventral half of the OPC, bisecting the cOPC compartment. As hedgehog is not expressed in the larval OPC, this dorsal-ventral boundary is set up in the embryo. Thus, six compartments (ventral cOPC, mOPC and pOPC and their dorsal counterparts) exist in the OPC. The pOPC compartment can be further subdivided by the expression of the wingless and dpp signalling genes. Cells in the wingless domain behave in a very distinct manner from the rest of the OPC, and have been described elsewhere (Erclik, 2017).

The Hth --> Klu --> Ey --> Slp1 --> D --> Tll temporal progression is not affected by the compartmentalization of the OPC epithelium; the same neuroblast progression throughout the entire OPC. Thus, in the developing medulla, neuroblasts expressing the same temporal factors are generated by developmentally distinct epithelial compartments (Erclik, 2017).

To test whether the intersection of the dorsal-ventral and temporal neuroblast axes leads to the production of distinct neural cell types, focus was placed on the progeny of Hth neuroblasts, which maintain Hth expression. In the late third instar, Hth neurons are found in a crescent that mirrors the OPC (see Distinct neuronal cell types are generated along the dorsal-ventral axis of the OPC). The Notch^ON (N^ON) progeny of Hth⁺ ganglion mother cells express Bsh and Ap, and they are distributed throughout the entire medulla crescent. In contrast, the Notch^OFF (N^OFF) progeny, which are Bsh⁻Hth⁺ neurons, express different combinations of transcription factors, and can be subdivided into three domains along the dorsal-ventral axis: (1) in the cOPC, N^OFFHth⁺ neurons express Vsx1, Seven-Up (Svp) and Lim3; (2) in the pOPC, these neurons also express Svp and Lim3, but not Vsx1; (3) in the ventral pOPC exclusively, these neurons additionally express Teashirt (Tsh). N^OFFHth⁺ cells are not observed in the mOPC. Rather, Cleaved-Caspase-3⁺ cells are intermingled with Bsh⁺ neurons. When cell death is prevented, Bsh⁺Hth⁺ cells become intermingled with neurons that express the N^OFF marker Lim3, confirming that the N^OFFHth⁺ progeny undergo apoptosis in the mOPC (Erclik, 2017).

It was therefore possible to distinguish three regional populations of Hth neurons (plus one that is eliminated by apoptosis) and a fourth population that is generated throughout the OPC. The neuronal identity of each of these populations was identified, as follows. (1) Bsh is a specific marker of Mi1 uni-columnar interneurons that are generated in all regions of the OPC. (2) To determine the identity of Hth⁺N^OFF cOPC-derived neurons, Hth⁺ single cell flip-out clones were generated (using hth-Gal4) in the adult medulla. The only Hth⁺ neurons that are also Vsx1⁺Svp⁺ are Pm3 multi-columnar local neurons. (3) For Hth⁺N^OFF pOPC-derived neurons, 27b-Gal4 was used; it drives expression in larval pOPC Hth⁺N^OFF neurons and is maintained to adulthood. Flip-out clones with 27b-Gal4 mark Pm1 and Pm2 neurons, as well as Hth^- Tm1 uni-columnar neurons that come from a different temporal window. Both Pm1 and Pm2 neurons (but not Tm1) express Hth and Svp. Pm1 neurons also express Tsh, which only labels larval ventral pOPC neurons (Erclik, 2017).

Thus, in addition to uni-columnar Mi1 neurons generated throughout the OPC, Hth neuroblasts generate three region-specific neuronal types: multi-columnar Pm3 neurons in the cOPC; multi-columnar Pm1 neurons in the ventral pOPC; and multi-columnar Pm2 neurons in the dorsal pOPC (Erclik, 2017).

To determine the contribution of the temporal and spatial factors to the generation of the different neuronal fates, the factors were mutated them and whether neuronal identity was lost was examined. To test the temporal axis, hth was mutated. As previously reported, Bsh expression is lost in hth mutant clones. Loss of hth in clones also leads to the loss of the Pm3 marker Svp without affecting expression of Vsx1, indicating that Vsx1 is not sufficient to activate Svp and can only do so in the context of an Hth⁺ neuroblast. Hth is also required for the specification of Pm1 and Pm2 in the pOPC as Svp and Tsh expression is lost in hth mutant larval clones. Ectopic expression of Hth in older neuroblasts is not able to expand Pm1, 2 or 3 fates (on the basis of the expression of Svp) into later born neurons, although it is able to expand Bsh expression. Thus, temporal input is necessary for the specification of all Hth⁺ neuronal fates but only sufficient for the generation of Mi1 neurons (see Temporal and spatial inputs are required for neuronal specification in the medulla. ) (Erclik, 2017).

Next, whether regional inputs are necessary and/or sufficient to specify neuronal fates in the progeny of Hth⁺ neuroblasts was determined. In Vsx1 RNA interference (RNAi) clones, Svp expression is lost in the cOPC but Bsh is unaffected. Additionally, Hth⁺Lim3⁺ cells are absent, suggesting that N^OFF cells undergo apoptosis in these clones. Conversely, ectopic expression of Vsx1 leads to the expression of Svp in mOPC Hth⁺ neurons but does not affect Bsh expression. Therefore, Vsx1 is both necessary and sufficient for the specification of Pm3 fates in the larva. However, unlike the temporal factor Hth, Vsx1 does not affect the generation of Mi1 neurons (Erclik, 2017).

In Rx whole mutant larvae and in mutant clones, Svp⁺Lim3⁺Hth⁺ larval neurons (that is, Pm1 and Pm2 neurons) in the pOPC are lost. Additionally, the Pm1 marker Tsh is lost in ventral pOPC Hth⁺ cells. Consistent with the Vsx1 mutant data, larval Bsh expression is not affected by the loss of Rx. In adults, the Pm1/Pm2 markers (Svp, Tsh and 27b-Gal4) are lost in the medulla (Erclik, 2017).

Ectopic expression of Rx leads to the activation of Svp in mOPC Hth⁺ neurons, but does not affect the expression of Bsh. It also leads to the activation of Tsh, but only in the ventral half of mOPC Hth⁺ neurons, suggesting that a ventral factor acts together with Rx to specify ventral fates. Taken together, the above data show that Rx is both necessary and sufficient for the specification of Pm1/2 neurons but (like Vsx1) does not affect the generation of Mi1 neurons (Erclik, 2017).

Finally, the role of the mOPC marker Optix in neuronal specification was examined. In Optix mutant clones, Svp is ectopically expressed in the mOPC, but Bsh expression is not affected. Of note, these ectopic Svp⁺ neurons fail to express the region-specific Pm markers Vsx1 or Tsh (in ventral clones), which suggests that they assume a generic Pm fate. Conversely, ectopic expression of Optix leads to the loss of Svp expressing neurons in both the cOPC and pOPC but does not affect Bsh. These N^OFF neurons die by apoptosis as no Lim3⁺ neurons are found intermingled with Bsh⁺Hth⁺N^ON neurons. When apoptosis is prevented in mOPC-derived neurons, Svp is not derepressed in the persisting Hth⁺N^OFF neurons, which suggests that Optix both represses Svp expression and promotes cell death in Hth⁺N^OFF neurons (Erclik, 2017).

The above data demonstrate that input from both the temporal and regional axes is required to specify neuronal fates. The temporal factor Hth is required for both Mi1 and Pm1/2/3 specification. The spatial genes are not required for the specification of N^ON Mi1 neurons, consistent with the observation that Mi1 is generated in all OPC compartments. The spatial genes, however, are both necessary and sufficient for the activation (Vsx1 and Rx) or repression (Optix) of the N^OFF Pm1/2/3 neurons. Thus, Hth⁺ neuroblasts generate two types of progeny: N^OFF neurons that are sensitive to spatial input (Pm1/2/3) and N^ON neurons that are refractory to spatial input (Mi1). Vsx1 expression in the cOPC is only maintained in Hth⁺N^OFF neurons, suggesting that spatial information may be 'erased' in Mi1, thus allowing the same neural type to be produced throughout the OPC (Erclik, 2017).

Do spatial genes regulate each other in the neuroepithelium? In Vsx1 mutant clones, Optix (but not Rx) is derepressed in the cOPC epithelium. Conversely, ectopic Vsx1 is sufficient to repress Optix in the mOPC and Rx in the pOPC. Similarly, Optix, but not Vsx1, is derepressed in Rx mutant clones in the pOPC epithelium and ectopic Rx is sufficient to repress Optix in the mOPC (but not Vsx1 in the cOPC). In Optix mutant clones, neither Vsx1 nor Rx are derepressed in the mOPC epithelium, but ectopic Optix is sufficient to repress both Vsx1 in the cOPC and Rx in the pOPC. The observation that Optix is not necessary to suppress Vsx1 or Rx in the mOPC neuroepithelium is surprising because Svp is activated in a subset of Hth⁺ neurons in the mOPC in Optix mutant clones. Nevertheless, when cell death in the mOPC is abolished, the ectopic undead N^OFF neurons express Lim3 but not Svp, which confirms that Optix represses Svp expression in mOPC neurons. Taken together, these results support a model in which Optix is sufficient to repress Vsx1 and Rx, to promote the death of Hth⁺N^OFF neurons and to repress Pm1/2/3 fates (see Spatial genes cross-regulate each other in the OPC neuroepithelium). Vsx1 and Rx act to promote Pm3 (Vsx1) or Pm1/2 (Rx) fates but can only do so in the absence of Optix (Erclik, 2017).

These results suggest that multi-columnar neurons are generated at specific locations in the medulla crescent. However, since these neurons are required to process visual information from the entire retina in the adult medulla, how does the doral-ventral position of neuronal birth in the larval crescent correlate with their final position in the adult? Lineage-tracing experiments were performed with Vsx1-Gal4 to permanently mark neurons generated in the cOPC and with Optix-Gal4 for mOPC neurons, and the position of the cell bodies of these neurons was analyzed. In larvae, neurons from the cOPC or from the mOPC remain located in the same dorsal-ventral position where they were born. However, in adults, both populations have moved to populate the entire medulla cortex along the dorsal-ventral axis (see Neuronal movement during medulla development is restricted to multi-columnar cell types). The kinetics of cell movement during development was analyzed by following cOPC neurons. Neurons born in the cOPC remain tightly clustered until 20 h after puparium formation (P20), after which point the cell bodies spread throughout the medulla cortex. By P30 the neurons are distributed over the entire dorsal-ventral axis of the medulla cortex. In the adult, most neurons derived from the cOPC neuroepithelium are located throughout the cortex although there is an enrichment of neurons in the central region of the cortex (Erclik, 2017).

To determine whether these observed movements involve the entire neuron or just the cell body, the initial targeting of cOPC or mOPC-derived neurons in larvae was examined before the onset of cell movement. In larvae, both populations send processes that target the entire dorsal-ventral axis of the medulla neuropi. Therefore, medulla neurons first send projections to reach their target columns throughout the entire medulla. Later, remodelling of the medulla results in extensive movement of cell bodies along the dorsal-ventral axis, leading to their even distribution in the cortex (Erclik, 2017).

What is the underlying logic behind why some neurons move while others do not? Markers were studied for the Mi1 (Bsh), Pm2 (Hth⁺Svp⁺), Pm1 (Hth⁺Svp⁺Tsh⁺), and Pm3 (Vsx1⁺Svp⁺Hth⁺) populations of neurons through pupal stages and up to the adult. Mi1 neurons are generated evenly throughout the larval OPC and remain regularly distributed across the dorsal-ventral axis at all stages. The lineage-tracing experiment was repeated with Vsx1-Gal4 to follow Mi1 neurons produced by the cOPC. These Mi1 neurons remain exclusively in the centre of the adult medulla cortex, demonstrating that they do not move. In contrast, Pm3 neurons remain tightly clustered in the central region until P20, at which point they move to occupy the entire cortex (Erclik, 2017).

However, not all multi-columnar neurons have cell bodies that move to occupy the entire medulla cortex. Unlike Mi1 and Pm3, adult Pm1 and Pm2 cell bodies are not located in the adult medulla cortex but instead in the medulla rim, at the edges of the cortex. Pm1 and Pm2 markers remain clustered at the ventral (Pm1) or dorsal (Pm2) posterior edges of the medulla cortex throughout all pupal stages. In the adult, both populations occupy the medulla rim from where they send long horizontal projections that reach the entire dorsal-ventral axis of the medulla neuropil. The pOPC may be a specialized region where many of the medulla rim cell types are generated. Even though most of cOPC-derived neurons move during development, a cOPC-derived multi-columnar neuron (TmY₁₄) was identified that sends processes targeting the entire dorsal-ventral length of the medulla neuropil but whose cell bodies remain in the central medulla cortex in the adult (Erclik, 2017).

Thus, the four populations of Hth neurons follow different kinetics: Mi1 neurons are born throughout the OPC and do not move; Pm3 neurons are born centrally and then move to distribute throughout the entire cortex; and Pm1/Pm2 neurons are born at the ventral or dorsal posterior edges of the OPC and occupy the medulla rim in adults (Erclik, 2017).

It is noted that uni-columnar Mi1 neurons, whose cell bodies do not move, reside in the distal cortex whereas multi-columnar Pm3 neurons, which move, reside in the proximal cortex. The hypothesis was thus tested that neurons whose cell bodies are located distally in the medulla cortex represent uni-columnar neurons generated homogeneously throughout the OPC that do not move. In contrast, proximal neurons, which are fewer in number and are generated in specific subregions of the medulla OPC, would be multi-columnar and move to their final position (Erclik, 2017).

It was first confirmed that neurons that move have their cell bodies predominantly in the proximal medulla cortex. The cell body position of neurons born ventrally that have moved dorsally was analyzed using the hh-Gal4 lineage trace: in the adult, the cell bodies found dorsally are mostly in the proximal medulla cortex, whereas the cell bodies in the ventral region are evenly distributed throughout the distal-proximal axis of the ventral cortex. They probably represent both distal uni-columnar neurons that did not move as well as proximal multi-columnar neurons that remained in the ventral region (Erclik, 2017).

Next the pattern of movement of Tm2 uni-columnar neurons from the ventral and dorsal halves of the OPC was analyzed using the hh-based lineage-trace. The cell bodies of Tm2 neurons are located throughout the dorsal-ventral axis in the adult medulla cortex but are co-labelled with the hh lineage marker only in the ventral half. Thus, like Mi1, Tm2 uni-columnar neurons do not move. Furthermore, uni-columnar Tm1 neurons, labelled by 27b-Gal4, are born throughout the dorsal-ventral axis of the OPC crescent with distal cell bodies in the adult, suggesting that they also remain where they were born (Erclik, 2017).

Conversely, it was asked whether neurons that are specified in only one region, such as the Vsx⁺ neurons of the cOPC, are multi-columnar in morphology. By sparsely labelling cOPC-derived neurons using the Vsx1-Gal4 driver, 13 distinct cell types were characterized that retain Vsx1 expression in the adult medulla. Strikingly, all are multi-columnar in morphology, further supporting the model that it is the multi-columnar neurons that move during pupal development (Erclik, 2017).

Finally, MARCM clones were generated in the OPC neuroepithelium and visualized using cell-type-specific Gal4 drivers in the adult medulla. Two classes of adult clone distribution were observed: clones in which neurons are tightly clustered, and clones in which neurons are dispersed. Consistent with the model, the clustered clones are those labelled with uni-columnar neuronal drivers, whereas the dispersed clones are those labelled with a multi-columnar driver (Erclik, 2017).

Taken together, these data demonstrate that neurons that do not move are uni-columnar (with cell bodies in the distal cortex), whereas most multi-columnar neurons (with cell bodies in the proximal cortex) move (Erclik, 2017).

This study shows that combinatorial inputs from the temporal and spatial axes act together to promote neural diversity in the medulla. Previous work has shown that a temporal series of transcription factors expressed in medulla neuroblasts allows for a diversification of the cell types generated by the neuroblasts as they age. This study now shows that input from the dorsal-ventral axis leads to further diversification of the neurons made by neuroblasts; at a given temporal stage, neuroblasts produce the same uni-columnar neuronal type globally as well as smaller numbers of multi-columnar cell types regionally. This situation is reminiscent of the mode of neurogenesis in the Drosophila ventral nerve cord, in which each neuroblast also expresses a (different) temporal series of transcription factors that specifies multiple neuronal types in the lineage. Spatial cues from segment polarity, dorsal-ventral and Hox genes then intersect to impart unique identities to each of the lineages. However, neuroblasts from the different segments give rise to distinct lineages to accommodate the specific function of each segment. In contrast, in the medulla, the entire OPC contributes to framing the repeating units that form the retinotopic map. It is therefore likely that each neuroblast produces a common set of neurons that connect to each pair of incoming R7 and R8 cells, or L1-L5 lamina neurons. This serves to produce 800 medulla columns with a 1:1 stoichiometry of medulla neurons to photoreceptors. The medulla neurons that are produced by neuroblasts throughout the dorsal-ventral axis of the OPC are thus uni-columnar The production of the same neuronal type along the entire OPC could be achieved by selectively 'erasing' spatial information in uni-columnar neurons, as observed in Mi1 neurons (Erclik, 2017).

Regional differences in the OPC confer further spatial identities to neuroblasts with the same temporal identity, and lead to specific differences in the lineages produced in the compartments along the dorsal-ventral axis of the medulla. These differences produce smaller numbers of multi-columnar neurons whose stoichiometry is much lower than 1:1. The majority of these neurons move during development to be uniformly distributed in the adult medulla cortex. This combination of regional and global neuronal specification in the medulla presents a powerful mechanism to produce the proper diversity and stoichiometry of neuronal types and generate the retinotopic map (Erclik, 2017).

A core paired-type and POU homeodomain-containing transcription factor program drives retinal bipolar cell gene expression

The diversity of cell types found within the vertebrate CNS arises in part from action of complex transcriptional programs. In the retina, the programs driving diversification of various cell types have not been completely elucidated. To investigate gene regulatory networks that underlie formation and function of one retinal circuit component, the bipolar cell, transcriptional regulation of three bipolar cell-enriched genes was analyzed. Using in vivo retinal DNA transfection and reporter gene constructs, a 200 bp metabotropic glutamate receptor 6 (Grm6) enhancer sequence, a 445 bp calcium-binding protein 5 (Cabp5) promoter sequence, and a 164 bp Chx10 enhancer sequence, were defined, each driving reporter expression specifically in distinct but overlapping bipolar cell subtypes. Bioinformatic analysis of sequences revealed the presence of potential paired-type and POU homeodomain-containing transcription factor binding sites, which were shown to be critical for reporter expression through deletion studies. The paired-type homeodomain transcription factors (TFs) Crx and Otx2 and the POU homeodomain factor Brn2 are expressed in bipolar cells and interacted with the predicted binding sequences as assessed by electrophoretic mobility shift assay. Grm6, Cabp5, and Chx10 reporter activity was reduced in Otx2 loss-of-function retinas. Endogenous gene expression of bipolar cell molecular markers was also dependent on paired-type homeodomain-containing TFs, as assessed by RNA in situ hybridization and reverse transcription-PCR in mutant retinas. Cabp5 and Chx10 reporter expression was reduced in dominant-negative Brn2-transfected retinas. The paired-type and POU homeodomain-containing TFs Otx2 and Brn2 together appear to play a common role in regulating gene expression in retinal bipolar cells (Kim, 2008).

Paired and LIM class homeodomain proteins coordinate differentiation of the C. elegans ALA neuron

The ancient origin of sleep is evidenced by deeply conserved signaling pathways regulating sleep-like behavior, such as signaling through the Epidermal growth factor receptor (EGFR). In Caenorhabditis elegans, a sleep-like state can be induced at any time during development or adulthood through conditional expression of LIN-3/EGF. The behavioral response to EGF is mediated by EGFR activity within a single cell, the ALA neuron, and mutations that impair ALA differentiation are expected to confer EGF-resistance. This study describes three such EGF-resistant mutants. One of these corresponds to the LIM class homeodomain (HD) protein CEH-14/Lhx3, and the other two correspond to Paired-like HD proteins CEH-10/Chx10 and CEH-17/Phox2. Whereas CEH-14 is required for ALA-specific gene expression throughout development, the Prd-like proteins display complementary temporal contributions to gene expression, with the requirement for CEH-10 decreasing as that of CEH-17 increases. Evidence is presented that CEH-17 participates in a positive autoregulatory loop with CEH-14 in ALA, and that CEH-10, in addition to its role in ALA differentiation, functions in the generation of the ALA neuron. Similarly to CEH-17, CEH-10 is required for the posterior migration of the ALA axons, but CEH-14 appears to regulate an aspect of ALA axon outgrowth that is distinct from that of the Prd-like proteins. These findings reveal partial modularity among the features of a neuronal differentiation program and their coordination by Prd and LIM class HD proteins (Van Buskirk, 2010).

Chx10 consolidates V2a interneuron identity through two distinct gene repression modes

During development, two cell types born from closely related progenitor pools often express identical transcriptional regulators despite their completely distinct characteristics. This phenomenon implies the need for a mechanism that operates to segregate the identities of the two cell types throughout differentiation after initial fate commitment. To understand this mechanism, this study investigated the fate specification of spinal V2a interneurons, which share important developmental genes with motor neurons (MNs). The paired homeodomain factor Chx10 (see Drosophila Vsx1 and Vsx2) was shown to function as a critical determinant for V2a fate and is required to consolidate V2a identity in postmitotic neurons. Chx10 actively promotes V2a fate, downstream of the LIM-homeodomain factor Lhx3, while concomitantly suppressing the MN developmental program by preventing the MN-specific transcription complex from binding and activating MN genes. This dual activity enables Chx10 to effectively separate the V2a and MN pathways. This study uncovers a widely applicable gene regulatory principle for segregating related cell fates (Clovis, 2016).

Bernier, G., et al. (2001). Isolation and characterization of a downstream target of Pax6 in the mammalian retinal primordium. Development 128 3987-3994. PubMed ID: 11641222

Burmeister, M., et al. (1996). Ocular retardation mouse caused by Chx10 homeobox null allele: Impaired retinal progenitor proliferation and bipolar cell differentiation. Nat. Genet. 12: 376-384. PubMed ID: 8630490

Chow, R. L. et al. (2001). Vsx1, a rapidly evolving paired-like homeobox gene expressed in cone bipolar cells. Mech. Dev. 109: 315-322. PubMed ID: 11731243

Chow, R. L., et al. (2004). Control of late off-center cone bipolar cell differentiation and visual signaling by the homeobox gene Vsx1. Proc. Natl. Acad. Sci. 101: 1754-1759. PubMed ID: 14745032

Clovis, Y. M., Seo, S. Y., Kwon, J. S., Rhee, J. C., Yeo, S., Lee, J. W., Lee, S. and Lee, S. K. (2016). Chx10 consolidates V2a interneuron identity through two distinct gene repression modes. Cell Rep [Epub ahead of print]. PubMed ID: 27477290

Davis, R. J. et al. (2003). Drosophila retinal homeobox (drx) is not required for establishment of the visual system, but is required for brain and clypeus development, Dev. Biol. 259: 272-287. PubMed ID: 12871701

De Velasco, B., Shen, J., Go, S. and Hartenstein, V. (2004). Embryonic development of the Drosophila corpus cardiacum, a neuroendocrine gland with similarity to the vertebrate pituitary, is controlled by sine oculis and glass. Dev. Biol. 274(2): 280-94. PubMed ID: 15385159

de Velasco, B., Erclik, T., Shy, D., Sclafani, J., Lipshitz, H., McInnes, R. and Hartenstein, V. (2007). Specification and development of the pars intercerebralis and pars lateralis, neuroendocrine command centers in the Drosophila brain. Dev. Biol. 302(1): 309-23. PubMed ID: 17070515

Dumstrei, K., et al. (1998). EGFR signaling is required for the differentation and maintenance of neural progenitors along the dorsal midline of the Drosophila embryonic head. Development 125: 3417-3426. PubMed ID: 9693145

Erclik, T., Hartenstein, V., Lipshitz, H. D. and McInnes, R. R. (2008). Conserved role of the Vsx genes supports a monophyletic origin for bilaterian visual systems. Curr. Biol. 18(17): 1278-87. PubMed ID: 18723351

Erclik, T., Li, X., Courgeon, M., Bertet, C., Chen, Z., Baumert, R., Ng, J., Koo, C., Arain, U., Behnia, R., Rodriguez, A. D., Senderowicz, L., Negre, N., White, K. P. and Desplan, C. (2017). Integration of temporal and spatial patterning generates neural diversity. Nature [Epub ahead of print]. PubMed ID: 28077877

Ferda Percin, E., et al. (2000). Human microphthalmia associated with mutations in the retinal homeobox gene CHX10. Nat. Genet. 25: 397-401. PubMed ID: 10932181

Gold, K. S. and Brand, A. H. (2014). Optix defines a neuroepithelial compartment in the optic lobe of the Drosophila brain. Neural Dev 9: 18. PubMed ID: 25074684

Green, E. S., Stubbs, J. L. and Levine, E. M. (2003). Genetic rescue of cell number in a mouse model of microphthalmia: Interactions between Chx10 and G1-phase cell cycle regulators. Development 130: 539-552. PubMed ID: 12490560

Kim, D. S., Matsuda, T. and Cepko, C. L. (2008). A core paired-type and POU homeodomain-containing transcription factor program drives retinal bipolar cell gene expression. J. Neurosci. 28(31): 7748-64. PubMed ID: 18667607

Liu, I. S., et al. (1994). Developmental expression of a novel murine homeobox gene (Chx10): evidence for roles in determination of the neuroretina and inner nuclear layer. Neuron 13: 377-393. PubMed ID: 7914735

Meinertzhagen, I. (1991). Evolution of the cellular organization of the arthropod compound eye and optic lobe. In: J.R. Cronly-Dillon and R. Gregory, Editors, Vision and Visual Dysfunction Volume 2, Macmillan, London, pp. 341-363.

Ohtoshi, A., et al. (2004). Regulation of retinal cone bipolar cell differentiation and photopic vision by the CVC homeobox gene Vsx1. Curr. Biol. 14: 530-536. PubMed ID: 15043821

Passini, M. A., et al. (1997). Vsx-1 and Vsx-2: Differential expression of two paired-like homeobox genes during zebrafish and goldfish retinogenesis. J. Comp. Neurol. 388: 495-505. PubMed ID: 9368856

Svendsen, P. C. and McGhee, J. D. (1995). The C. elegans neuronally expressed homeobox gene ceh-10 is closely related to genes expressed in the vertebrate eye. Development 121: 1253-1262. PubMed ID: 7789259

Kim, S. K. and Rulifson, E. J. (2004). Conserved mechanisms of glucose sensing and regulation by Drosophila corpora cardiaca cells. Nature 431: 316-320. PubMed ID: 15372035

Landgraf, M., et al. (2003). Charting the Drosophila neuropile: a strategy for the standardised characterisation of genetically amenable neurites, Dev. Biol. 260: 207-225. PubMed ID: 12885565

Levine, E. M. et al. (1994). Restricted expression of a new paired-class homeobox gene in normal and regenerating adult goldfish retina, J. Comp. Neurol. 348: 596-606. PubMed ID: 7836564

Liu, I. S., et al. (1994). Developmental expression of a novel murine homeobox gene (Chx10): evidence for roles in determination of the neuroretina and inner nuclear layer. Neuron 13: 377-393. PubMed ID: 7914735

Ludwig, L., Williams, J. L. D. and Boyan, G. (2002). The pars intercerebralis of the locust brain: a developmental and comparative study. Microsc. Res. Tech. 56: 174-188. PubMed ID: 11810721

Meinertzhagen, I. (1991). Evolution of the cellular organization of the arthropod compound eye and optic lobe. In: J.R. Cronly-Dillon and R. Gregory, Editors, Vision and Visual Dysfunction Volume 2, Macmillan, London, pp. 341-363.

Nassel, D. R. (2002) Neuropeptides in the nervous system of Drosophila and other insects: multiple roles as neuromodulators and neurohormones. Prog. Neurobiol. 68: 1-84. PubMed ID: 12427481

Nassif, C., Noveen, A. and Hartenstein, V. (2003). Early development of the Drosophila brain: III. The pattern of neuropile founder tracts during the larval period. J. Comp. Neurol. 455: 417-434. PubMed ID: 12508317

Ohtoshi, A., et al. (2004). Regulation of retinal cone bipolar cell differentiation and photopic vision by the CVC homeobox gene Vsx1. Curr. Biol. 14: 530-536. PubMed ID: 15043821

Passini, M. A., et al. (1997). Vsx-1 and Vsx-2: Differential expression of two paired-like homeobox genes during zebrafish and goldfish retinogenesis. J. Comp. Neurol. 388: 495-505. PubMed ID: 9368856

Rulifson, E. J. Kim, S. K. and Nusse, R. (2002). Ablation of insulin-producing neurons in flies: growth and diabetic phenotypes. Science 296: 1118-1120. PubMed ID: 12004130

Siegmund T. and Korge, G. (2001). Innervation of the ring gland of Drosophila melanogaster, J. Comp. Neurol. 431: 481-491. PubMed ID: 11223816

Svendsen, P. C. and McGhee, J. D. (1995). The C. elegans neuronally expressed homeobox gene ceh-10 is closely related to genes expressed in the vertebrate eye. Development 121: 1253-1262. PubMed ID: 7789259

Van Buskirk, C. and Sternberg, P. W. (2010). Paired and LIM class homeodomain proteins coordinate differentiation of the C. elegans ALA neuron. Development 137(12): 2065-74. PubMed ID: 20501595

van der Horst, D. J., van Marrewijk, W. J. and Diederen, J. H. (2001). Adipokinetic hormones of insect: release, signal transduction, and responses. Int. Rev. Cytol. 211: 179-240. PubMed ID: 11597004

Wang, S., Tulina, N., Carlin, D. L. and Rulifson, E. J. (2007). The origin of islet-like cells in Drosophila identifies parallels to the vertebrate endocrine axis. Proc. Natl. Acad. Sci. 104(50): 19873-8. PubMed ID: 18056636

date revised: 30 September 2016

Home page: The Interactive Fly © 2008 Thomas Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.