Analysis of Drosophila photoreceptor axon guidance in eye-specific mosaics

During development of the adult Drosophila visual system, axons of the eight photoreceptors in each ommatidium fasciculate together and project as a single bundle towards the optic lobes of the brain. Within the brain, individual photoreceptor axons from each bundle then seek specific targets in distinct layers of the optic lobes. The axons of photoreceptors R1-R6 terminate in the lamina, while R7 and R8 axons pass through the lamina to terminate in separate layers of the medulla. To identify genes required for photoreceptor axon guidance, including those with essential functions during early development, we have devised a strategy for the simple and efficient generation of genetic mosaics in which mutant photoreceptor axons innervate a predominantly wild-type brain. In a large-scale saturation mutagenesis performed using this system, we recovered new alleles of the gene encoding the receptor tyrosine phosphatase PTP69D. PTP69D has previously been shown to function in the correct targeting of motor axons in the embryo and R1-R6 axons in the visual system. Here, we show that PTP69D is also required for correct targeting of R7 axons. Whereas mutant R1-R6 axons occasionally extend beyond their normal targets in the lamina, mutant R7 axons often fail to reach their targets in the medulla, stopping instead at the same level as the R8 axon. These targeting errors are difficult to reconcile with models in which PTP69D plays an instructive role in photoreceptor axon targeting, as previously proposed. Rather, we suggest that PTP69D plays a permissive role, perhaps reducing the adhesion of R1-R6 and R7 growth cones to the pioneer R8 axon so that they can respond independently to their specific targeting cues.     

Stromal cell‐derived factor‐1 and hepatocyte growth factor guide axon projections to the extraocular muscles

Vertebrate eye movements depend on the co‐ordinated function of six extraocular muscles that are innervated by the oculomotor, trochlear, and abducens nerves. Here, we show that the diffusible factors, stromal cell‐derived factor‐1 (SDF‐1) and hepatocyte growth factor (HGF), guide the development of these axon projections. SDF‐1 is expressed in the mesenchyme around the oculomotor nerve exit point, and oculomotor axons fail to exit the neuroepithelium in mice mutant for the SDF‐1 receptor CXCR4. Both SDF‐1 and HGF are expressed in or around the ventral and dorsal oblique muscles, which are distal targets for the oculomotor and trochlear nerves, respectively, as well as in the muscles which are later targets for oculomotor axon branches. We find that in vitro SDF‐1 and HGF promote the growth of oculomotor and trochlear axons, whereas SDF‐1 also chemoattracts oculomotor axons. Oculomotor neurons show increased branching in the presence of SDF‐1 and HGF singly or together. HGF promotes the growth of trochlear axons more than that of oculomotor axons. Taken together, these data point to a role for both SDF‐1 and HGF in extraocular nerve projections and indicate that SDF‐1 functions specifically in the development of the oculomotor nerve, including oculomotor axon branch formation to secondary muscle targets. HGF shows some specificity in preferentially enhancing development of the trochlear nerve. © 2010 Wiley Periodicals, Inc. Develop Neurobiol 70: 549–564, 2010     

Muscle founder cells regulate defasciculation and targeting of motor axons in the Drosophila embryo.

During Drosophila embryogenesis, motor axons leave the central nervous system (CNS) as two separate bundles, the segmental nerve (SN) and intersegmental nerve (ISN). From these, axons separate (defasciculate) progressively in a characteristic pattern, initially as nerve branches and then as individual axons, to innervate target muscles [1] [2]. This pattern of branching resembles the outgrowth and defasciculation of motor axons from the neural tube of vertebrate embryos. The factors that trigger nerve branching are unknown. In vertebrate limbs, the branched innervation may depend on mesodermal cues, in particular on the connective tissues that organise the muscle pattern [3]. In Drosophila, the muscle pattern is organised by specific mesodermal cells, the founder myoblasts, which initiate the development of individual muscles [4][5][6]. Founder myoblasts fuse with neighbouring non-founder myoblasts and entrain these to a specific muscle programme, which also determines their innervation [4] [7]. In the absence of mesoderm, ISN and SN can form, but motor axons fail to defasciculate from these bundles [7]. The cue(s) for nerve branching therefore lie within the mesoderm, most likely in the muscles and/or in the precursor cells of the adult musculature [8]. Here, we show that founder myoblasts are the source of the cue(s) that are required to trigger defasciculation and targeted growth of motor axons. Moreover, we found that a single founder myoblast can trigger the defasciculation of an entire nerve branch. This suggests that the muscle field is structured into sets of muscles, each expressing a common defasciculation cue for a particular nerve branch.     

Common mechanisms underlying growth cone guidance and axon branching

During development, growth cones direct growing axons into appropriate targets. However, in some cortical pathways target innervation occurs through the development of collateral branches that extend interstitially from the axon shaft. How do such branches form? Direct observations of living cortical brain slices revealed that growth cones of callosal axons pause for many hours beneath their cortical targets prior to the development of interstitial branches. High resolution imaging of dissociated living cortical neurons for many hours revealed that the growth cone demarcates sites of future axon branching by lengthy pausing behaviors and enlargement of the growth cone. After a new growth cone forms and resumes forward advance, filopodial and lamellipodial remnants of the large paused growth cone are left behind on the axon shaft from which interstitial branches later emerge. To investigate how the cytoskeleton reorganizes at axon branch points, we fluorescently labeled microtubules in living cortical neurons and imaged the behaviors of microtubules during new growth from the axon shaft and the growth cone. In both regions microtubules reorganize into a more plastic form by splaying apart and fragmenting. These shorter microtubules then invade newly developing branches with anterograde and retrograde movements. Although axon branching of dissociated cortical neurons occurs in the absence of targets, application of a target-derived growth factor, FGF-2, greatly enhances branching. Taken together, these results demonstrate that growth cone pausing is closely related to axon branching and suggest that common mechanisms underlie directed axon growth from the terminal growth cone and the axon shaft.     

Nerve growth factor-induced growth of sympathetic axons into the optic tract of mature mice is enhanced by an absence of p75NTR expression

Postganglionic sympathetic axons display a remarkable ability for new collateral growth in response to local increases in nerve growth factor (NGF). Elevating NGF levels within the brain also induces the directional growth of sympathetic axons, but not within myelinated pathways of adult mammals. In this investigation, we provide in vivo evidence that sympathetic axons are capable of NGF-induced collateral growth through the microenvironment of mature myelinated pathways, especially in the absence of the p75 neurotrophin receptor (NTR). In transgenic mice overexpressing NGF centrally and expressing p75NTR, only a few varicose sympathetic axons invade the optic tract after the first month of postnatal life. In other transgenic mice overexpressing NGF centrally but lacking p75NTR expression, the incidence of sympathetic axons within this myelinated tract substantially increases. Moreover, numerous unmyelinated sympathetic axons cluster together to form large processes extending through the optic tract; such structures are first seen 8 weeks after birth. Only these large axon bundles display prominent immunostaining for GAP-43, which is preferentially localized to the sympathetic fibers, since nonmyelinating Schwann cells are not associated with these axon bundles. These data provide the first direct evidence that sympathetic axons are indeed capable of NGF-induced collateral growth into myelinated tracts of mature mammals, and that their continued growth through this microenvironment is markedly enhanced by the absence of p75NTR expression. We propose that p75NTR among sympathetic axons may either directly or indirectly limit collateral branching of these fibers in response to increased levels of NGF.     

In vivo imaging of axonal degeneration and regeneration in the injured spinal cord

The poor response of central axons to transection underlies the bleak prognosis following spinal cord injury. Here, we monitor individual fluorescent axons in the spinal cords of living transgenic mice over several days after spinal injury. We find that within 30 min after trauma, axons die back hundreds of micrometers. This acute form of axonal degeneration is similar in mechanism to the more delayed Wallerian degeneration of the disconnected distal axon, but acute degeneration affects the proximal and distal axon ends equally. In vivo imaging further shows that many axons attempt regeneration within 6-24 h after lesion. This growth response, although robust, seems to fail as a result of the inability of axons to navigate in the proper direction. These results suggest that time-lapse imaging of spinal cord injury may provide a powerful analytical tool for assessing the pathogenesis of spinal cord injury and for evaluating therapies that enhance regeneration.     

Axon degeneration mechanisms: commonality amid diversity

A wide range of insults can trigger axon degeneration, and axons respond with diverse morphology, topology and speed. However, recent genetic, immunochemical, morphological and pharmacological investigations point to convergent degeneration mechanisms. The principal convergence points - poor axonal transport, mitochondrial dysfunction and an increase in intra-axonal calcium - have been identified by rescuing axons with the slow Wallerian degeneration gene (Wld(S)) and studies with blockers of sodium or calcium influx. By understanding how the pathways fit together, we can combine our knowledge of mechanisms, and potentially also treatment strategies, from different axonal disorders.     

Distinct mechanisms triggering glial differentiation in Drosophila thoracic and abdominal neuroblasts 6-4

Neurons and glia are produced in stereotyped patterns after neuroblast cell division during development of the Drosophila central nervous system. The first cell division of thoracic neuroblast 6-4 (NB6-4T) is asymmetric, giving rise to a glial precursor cell and a neuronal precursor cell. In contrast, abdominal NB6-4 (NB6-4A) divides symmetrically to produce two glial cells. To understand the relationship between cell division and glia-neuron cell fate determination, we examined and compared the effects of known cell division mutations on the NB6-4T and NB6-4A lineages. Based on observation of expression of glial fate determination and early glial differentiation genes, the onset of glial differentiation occurred in NB6-4A but not in NB6-4T when both cell cycle progression and cytokinesis were genetically arrested. On the other hand, glial differentiation started in both lineages when cytokinesis was blocked with intact cell cycle progression. These results showed that NB6-4T, but not NB6-4A, requires cell cycle progression for acquisition of glial fate, suggesting that distinct mechanisms trigger glial differentiation in the different lineages.     

Genetic Study of Axon Regeneration with Cultured Adult Dorsal Root Ganglion Neurons

It is well known that mature neurons in the central nervous system (CNS) cannot regenerate their axons after injuries due to diminished intrinsic ability to support axon growth and a hostile environment in the mature CNS^1,2. In contrast, mature neurons in the peripheral nervous system (PNS) regenerate readily after injuries³. Adult dorsal root ganglion (DRG) neurons are well known to regenerate robustly after peripheral nerve injuries. Each DRG neuron grows one axon from the cell soma, which branches into two axonal branches: a peripheral branch innervating peripheral targets and a central branch extending into the spinal cord. Injury of the DRG peripheral axons results in substantial axon regeneration, whereas central axons in the spinal cord regenerate poorly after the injury. However, if the peripheral axonal injury occurs prior to the spinal cord injury (a process called the conditioning lesion), regeneration of central axons is greatly improved⁴. Moreover, the central axons of DRG neurons share the same hostile environment as descending corticospinal axons in the spinal cord. Together, it is hypothesized that the molecular mechanisms controlling axon regeneration of adult DRG neurons can be harnessed to enhance CNS axon regeneration. As a result, adult DRG neurons are now widely used as a model system to study regenerative axon growth^5-7.Here we describe a method of adult DRG neuron culture that can be used for genetic study of axon regeneration in vitro. In this model adult DRG neurons are genetically manipulated via electroporation-mediated gene transfection^6,8. By transfecting neurons with DNA plasmid or si/shRNA, this approach enables both gain- and loss-of-function experiments to investigate the role of any gene-of-interest in axon growth from adult DRG neurons. When neurons are transfected with si/shRNA, the targeted endogenous protein is usually depleted after 3-4 days in culture, during which time robust axon growth has already occurred, making the loss-of-function studies less effective. To solve this problem, the method described here includes a re-suspension and re-plating step after transfection, which allows axons to re-grow from neurons in the absence of the targeted protein. Finally, we provide an example of using this in vitro model to study the role of an axon regeneration-associated gene, c-Jun, in mediating axon growth from adult DRG neurons⁹.     

Drosophila Dscam is required for divergent segregation of sister branches and suppresses ectopic bifurcation of axons

Axon bifurcation results in the formation of sister branches, and divergent segregation of the sister branches is essential for efficient innervation of multiple targets. From a genetic mosaic screen, we find that a lethal mutation in the Drosophila Down syndrome cell adhesion molecule (Dscam) specifically perturbs segregation of axonal branches in the mushroom bodies. Single axon analysis further reveals that Dscam mutant axons generate additional branches, which randomly segregate among the available targets. Moreover, when only one target remains, branching is suppressed in wild-type axons while Dscam mutant axons still form multiple branches at the original bifurcation point. Taken together, we conclude that Dscam controls axon branching and guidance such that a neuron can innervate multiple targets with minimal branching.     

Gogo Receptor Contributes to Retinotopic Map Formation and Prevents R1-6 Photoreceptor Axon Bundling

Background

Topographic maps form the basis of neural processing in sensory systems of both vertebrate and invertebrate species. In the Drosophila visual system, neighboring R1–R6 photoreceptor axons innervate adjacent positions in the first optic ganglion, the lamina, and thereby represent visual space as a continuous map in the brain. The mechanisms responsible for the establishment of retinotopic maps remain incompletely understood.

Results

Here, we show that the receptor Golden goal (Gogo) is required for R axon lamina targeting and cartridge elongation in a partially redundant fashion with local guidance cues provided by neighboring axons. Loss of function of Gogo in large clones of R axons results in aberrant R1–R6 fascicle spacing. Gogo affects target cartridge selection only indirectly as a consequence of the disordered lamina map. Interestingly, small clones of gogo deficient R axons perfectly integrate into a proper retinotopic map suggesting that surrounding R axons of the same or neighboring fascicles provide complementary spatial guidance. Using single photoreceptor type rescue, we show that Gogo expression exclusively in R8 cells is sufficient to mediate targeting of all photoreceptor types in the lamina. Upon lamina targeting and cartridge selection, R axons elongate within their individual cartridges. Interestingly, here Gogo prevents bundling of extending R1-6 axons.

Conclusion

Taken together, we propose that Gogo contributes to retinotopic map formation in the Drosophila lamina by controlling the distribution of R1–R6 axon fascicles. In a later developmental step, the regular position of R1–R6 axons along the lamina plexus is crucial for target cartridge selection. During cartridge elongation, Gogo allows R1–R6 axons to extend centrally in the lamina cartridge.     

Defining the Ligand Specificity of the Deleted in Colorectal Cancer (DCC) Receptor

The growth and guidance of many axons in the developing nervous system require Netrin-mediated activation of Deleted in Colorectal Cancer (DCC) and other still unknown signaling cues. Commissural axon guidance defects are more severe in DCC mutant mice than Netrin-1 mutant mice, suggesting additional DCC activating signals besides Netrin-1 are involved in proper axon growth. Here we report that interaction screens on extracellular protein microarrays representing over 1,000 proteins uniquely identified Cerebellin 4 (CBLN4), a member of the C1q-tumor necrosis factor (TNF) family, and Netrin-1 as extracellular DCC-binding partners. Immunofluorescence and radio-ligand binding studies demonstrate that Netrin-1 competes with CBLN4 binding at an overlapping site within the membrane-proximal fibronectin domains (FN) 4–6 of DCC and binds with ∼5-fold higher affinity. CBLN4 also binds to the DCC homolog, Neogenin-1 (NEO1), but with a lower affinity compared to DCC. CBLN4-null mice did not show a defect in commissural axons of the developing spinal cord but did display a transient increase in the number of wandering axons in the brachial plexus, consistent with a role in axon guidance. Overall, the data solidifies CBLN4 as a bona fide DCC ligand and strengthens its implication in axon guidance.     

Messages from forgotten friends: classic cell adhesion molecules inhibit regeneration too

A necessary step toward complete functional recovery after spinal cord injury is the regeneration of axons. Axon regrowth after injury is prevented by a myriad of intrinsic and extrinsic factors. In this issue of The EMBO Journal, Huang et al ( 2016 ) demonstrate that the cell adhesion molecule NB‐3 (CNTN6) functions as a major brake on axon regrowth when it is activated by NB‐3 from scar‐forming cells at the injury site. Disruption of this NB‐3 trans‐cellular signaling led to impressive axon regrowth after spinal cord transection.     

UNC-6 expression by the vulval precursor cells of Caenorhabditis elegans is required for the complex axon guidance of the HSN neurons

Netrin is an evolutionarily conserved axon guidance molecule that has both axonal attraction and repulsion activities. In Caenorhabditis elegans, Netrin/UNC-6 is secreted by ventral cells, attracting some axons ventrally and repelling some axons, which extend dorsally. One axon guided by UNC-6 is that of the HSN neuron. The axon guidance process for HSN neurons is complex, consisting of ventral growth, dorsal growth, branching, second ventral growth, fasciculation with ventral nerve cords, and then anterior growth. The vulval precursor cells (VPC) and the PVP and PVQ neurons are required for the HSN axon guidance; however, the molecular mechanisms involved are completely unknown. In this study, we found that the VPC strongly expressed UNC-6 during HSN axon growth. Silencing of UNC-6 expression in only the VPC, using a novel tissue-specific RNAi technique, resulted in abnormal HSN axon guidance. The expression of Netrin/UNC-6 by only the VPC in unc-6 null mutants partially rescued the HSN ventral axon guidance. Furthermore, the expression of Netrin/UNC-6 by the VPC and the ventral nerve cord (VNC) in unc-6 null mutants restored the complex HSN axon guidance. These results suggest that UNC-6 expressed by the VPC and the VNC cooperatively regulates the complex HSN axon guidance.     

Role of netrin UNC-6 in patterning the longitudinal nerves of Caenorhabditis elegans

The nervous system of Caenorhabditis elegans comprises circumferential and longitudinal axon tracts. Netrin UNC-6 is required for the guidance of circumferential axon migrations and is expressed by ventral neuroglia and neurons in temporally and spatially regulated patterns. Migrating axons mediate the UNC-6 signal through the UNC-5 and UNC-40 receptors. It is thought that UNC-6 is secreted and becomes associated with basement membranes and cell surfaces to form gradients that direct circumferentially migrating axons toward or away from the ventral UNC-6 sources. Little is known about the effects of UNC-6 on longitudinally migrating axons. In unc-6, unc-5, and unc-40 null mutants, some longitudinal nerves are dorsally or ventrally misdirected. Furthermore, the organization of axons are disrupted within nerves. We show that cells ectopically expressing UNC-6 can redirect the migrations of some neighboring longitudinal axons, suggesting that the gradients postulated to direct circumferential migration also help specify the dorsoventral positions of these longitudinal nerves. We also manipulated the temporal and spatial expression pattern of UNC-6 by two different means. First, we removed the PVT midline neuron which expresses UNC-6 for a short time during axon outgrowths. Second, we expressed UNC-6 uniformly in the nervous system throughout development. The results suggest that changing UNC-6 expression patterns modify the distribution of the cue by providing new localized sources. This new guidance information is critical for organizing the axons of longitudinal nerves.     

Roundabout gene family functions during sensory axon guidance in the drosophila embryo are mediated by both Slit-dependent and Slit-independent mechanisms

roundabout (robo) family genes play key roles in axon guidance in a wide variety of animals. We have investigated the roles of the robo family members, robo, robo2, and robo3, in the guidance of sensory axons in the Drosophila embryo. In robo(-/-), slit(-/-), and robo(-/+) slit(-/+) mutants, lateral cluster sensory neurons misproject to cells and axons in the nearby ventral' (v') cluster. These phenotypes, together with the normal expression pattern of Slit and Robo, suggest that Slit ligand secreted from the epidermis interacts with Robo receptors on lateral cluster sensory growth cones to limit their exploration of nearby attractive substrates. The most common sensory axon phenotype seen in robo2(-/-) mutants was misprojection of dorsal cluster sensory axons away from their normal growth substrate, the transverse connective of the trachea. slit appears to play no role in this aspect of sensory axon growth. Robo2 is expressed, not on the dorsal sensory axons, but on the transverse connective. These results suggest a novel, non-cell-autonomous mechanism for axon guidance by robo family genes: Robo2 expressed on the trachea acts as an attractant for the dorsal sensory growth cones.     

Reversing the outcome of synapse elimination at developing neuromuscular junctions in vivo: evidence for synaptic competition and its mechanism

During mammalian development, neuromuscular junctions and some other postsynaptic cells transition from multiple- to single-innervation as synaptic sites are exchanged between different axons. It is unclear whether one axon invades synaptic sites to drive off other inputs or alternatively axons expand their territory in response to sites vacated by other axons. Here we show that soon-to-be-eliminated axons rapidly reverse fate and grow to occupy vacant sites at a neuromuscular junction after laser removal of a stronger input. This reversal supports the idea that axons take over sites that were previously vacated. Indeed, during normal development we observed withdrawal followed by takeover. The stimulus for axon growth is not postsynaptic cell inactivity because axons grow into unoccupied sites even when target cells are functionally innervated. These results demonstrate competition at the synaptic level and enable us to provide a conceptual framework for understanding this form of synaptic plasticity.     

Ephrin-A5 inhibits growth of embryonic sensory neurons

EphA-ephrin signaling has recently been implicated in the establishment of motor innervation patterns, in particular in determining whether motor axons project into dorsal versus ventral nerve trunks in the limb. We investigated whether sensory axons, which grow out together with and can be guided by motor axons, are also influenced by Eph-ephrin signaling. We show that multiple EphA receptors are expressed in DRGs when limb innervation is being established, and EphA receptors are present on growth cones of both NGF-dependent (predominantly cutaneous) and NT3-dependent (predominantly proprioceptive) afferents. Both soluble and membrane-attached ephrin-A5 inhibited growth of approximately half of each population of sensory axons in vitro. On average, growth cones that collapsed in response to soluble ephrin-A5 extended more slowly than those that did not, and ephrin-A5 significantly slowed the extension of NGF-dependent growth cones that did not collapse. Finally, we show that ectopic expression of ephrin-A5 in ovo reduced arborization of cutaneous axons in skin on the limb. Together these results suggest that sensory neurons respond directly to A-class ephrins in the limb. Thus, ephrins appear to pattern sensory axon growth in two ways-both directly, and indirectly via their inhibitory effects on neighboring motor axons.     

Sequential activation of butyrylcholinesterase in rostral half somites and acetylcholinesterase in motoneurones and myotomes preceding growth of motor axons

By applying double-staining procedures that combine cholinesterase histochemistry (acetyl- and butyrylcholinesterase, respectively) as indicators of neuronal and myotomal tissue differentiation on longitudinal sections, together with detection of motor axons with antibodies to G4 antigen, we here describe the spatiotemporal expression of all components of the segmental motor units along the trunk of chicken embryos between stages 16-20. In particular, BChE expression is spatially elevated on the rostral part of the differentiating somite. About 2-3 somites more rostrally (and thus developmentally later), AChE is expressed almost simultaneously in a nonsegmented fashion in neuronal cell bodies of the ventral horn and in the corresponding dermomyotomes. There it is first detectable in a rostromedial sector. With a delay (4-6 somites compared with AChE in motoneurones), motor axons begin to grow exclusively through the BChE-rich sclerotomal space towards the AChE-activated myotome anlage. On motor axons, AChE detection is significantly retarded. We conclude that the rostrocaudal segmental asymmetry is not restricted to the sclerotomes (which other authors have described before by using different markers), but it extends into the dermomyotome, in which cholinesterases introduce an early subdivision. Hence, the entire process of first myotome differentiation, motor axon growth and establishment of first target contacts are taking place within the rostral half somite. We suggest that both cholinesterases might be involved in processes of motor unit differentiation and fibre guidance.