Pruning an axon piece by piece: a new mode of synapse elimination

The process by which excess axons are pruned during development has remained unclear. In this issue of Neuron, Bishop et al. use time-lapse imaging and serial electron microscopy of developing neuromuscular junctions to describe a novel cellular mechanism in which retracting axon branches shed fragments rich in normal synaptic organelles. These "axosomes" are engulfed by adjacent Schwann cells and may be assimilated into the glial cytoplasm. Shedding of axosomes and glial engulfment may represent a widespread mechanism of synapse elimination.     

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.     

Essential role of the apoptotic cell engulfment genes draper and ced-6 in programmed axon pruning during Drosophila metamorphosis

Axon pruning is a common phenomenon in neural circuit development. Previous studies demonstrate that the engulfing action of glial cells is essential in this process. The underlying molecular mechanisms, however, remain unknown. We show that draper (drpr) and ced-6, which are essential for the clearance of apoptotic cells in C. elegans, function in the glial engulfment of larval axons during Drosophila metamorphosis. The drpr mutation and glia-specific knockdown of drpr and ced-6 by RNA interference suppress glial engulfment, resulting in the inhibition of axon pruning. drpr and ced-6 interact genetically in the glial action. Disruption of the microtubule cytoskeleton in the axons to be pruned occurs via ecdysone signaling, independent of glial engulfment. These findings suggest that glial cells engulf degenerating axons through drpr and ced-6. We propose that apoptotic cells and degenerating axons of living neurons are removed by a similar molecular mechanism.     

The Drosophila cell corpse engulfment receptor Draper mediates glial clearance of severed axons

Neuron-glia communication is central to all nervous system responses to trauma, yet neural injury signaling pathways remain poorly understood. Here we explore cellular and molecular aspects of neural injury signaling in Drosophila. We show that transected Drosophila axons undergo injury-induced degeneration that is morphologically similar to Wallerian degeneration in mammals and can be suppressed by the neuroprotective mouse Wlds protein. Axonal injury elicits potent morphological and molecular responses from Drosophila glia: glia upregulate expression of the engulfment receptor Draper, undergo dramatic changes in morphology, and rapidly recruit cellular processes toward severed axons. In draper mutants, glia fail to respond morphologically to axon injury, and severed axons are not cleared from the CNS. Thus Draper appears to act as a glial receptor for severed axon-derived molecular cues that drive recruitment of glial processes to injured axons for engulfment.     

Asynchronous synapse elimination in neonatal motor units: studies using GFP transgenic mice

In developing muscle, synapse elimination reduces the number of motor axons that innervate each postsynaptic cell. This loss of connections is thought to be a consequence of axon branch trimming. However, branch retraction has not been observed directly, and many questions remain, such as: do all motor axons retract branches, are eliminated branches withdrawn synchronously, and are withdrawing branches localized to particular regions? To address these questions, we used transgenic mice that express fluorescent proteins in small subsets of motor axons, providing a unique opportunity to reconstruct complete axonal arbors and identify all the postsynaptic targets. We found that, during early postnatal development, each motor axon loses terminal branches, but retracting branches withdraw asynchronously and without obvious spatial bias, suggesting that local interactions at each neuromuscular junction regulate synapse elimination.     

THREE-DIMENSIONAL ULTRASTRUCTURE OF THE CRAYFISH NEUROMUSCULAR APPARATUS

The synapse-bearing nerve terminals of the opener muscle of the crayfish Procambarus were reconstructed using electron micrographs of regions which had been serially sectioned. The branching patterns of the terminals of excitatory and inhibitory axons and the locations and sizes of neuromuscular and axo-axonal synapses were studied. Excitatory and inhibitory synapses could be distinguished not only on the basis of differences in synaptic vesicles, but also by a difference in density of pre- and postsynaptic membranes. Synapses of both axons usually had one or more sharply localized presynaptic "dense bodies" around which synaptic vesicles appeared to cluster. Some synapses did not have the dense bodies. These structures may be involved in the physiological activity of the synapse. Excitatory axon terminals had more synapses, and a larger percentage of terminal surface area devoted to synaptic contacts, than inhibitory axon terminals. However, the largest synapses of the inhibitory axon exceeded in surface area those of the excitatory axon. Both axons had many side branches coming from the main terminal; often, the side branches were joined to the main terminal by narrow necks. A greater percentage of surface area was devoted to synapses in side branches than in the main terminal. Only a small fraction of total surface area was devoted to axo-axonal synapses, but these were often located at narrow necks or constrictions of the excitatory axon. This arrangement would result in effective blockage of spike invasion of regions of the terminal distal to the synapse, and would allow relatively few synapses to exert a powerful effect on transmitter release from the excitatory axon. A hypothesis to account for the development of the neuromuscular apparatus is presented, in which it is suggested that production of new synapses is more important than enlargement of old ones as a mechanism for allowing the axon to adjust transmitter output to the functional needs of the muscle.     

Differences in synaptic output between excitatory and inhibitory motoneurons in a crayfish muscle

A pair of antagonistic motoneurons, one excitatory and one inhibitory, innervates the distal accessory flexor muscle in the walking limb of the crayfish Procambarus clarkii. The number and size of synapses formed by these two axons on the muscle fibers (neuromuscular synapses) and on each other (axo-axonal synapses) were estimated using thin-section electron microscopy. Although profiles of nerve terminals of the two axons occur in roughly equal proportions, the frequency of occurrence of neuromuscular synapses differed markedly: 73% were excitatory and 27% were inhibitory. However, inhibitory synapses were 4–5 times larger than excitatory ones, and consequently, the total contact areas devoted to neuromuscular synapses were similar for both axons. Axo-axonal synapses were predominantly from the inhibitory axon to the excitatory axon (86%), and a few were from the excitatory axon to the inhibitory axon (14%). The role of the inhibitory axo-axonal synapse is presynaptic inhibition, but that of the excitatory axo-axonal synapse is not known. The differences in size of neuromuscular synapses between the two axons may reflect intrinsic determinants of the neuron, while the similarity in total synaptic area may reflect retrograde influences from the muscle for regulating synapse number.     

Semaphorin-1a acts in concert with the cell adhesion molecules fasciclin II and connectin to regulate axon fasciculation in Drosophila

Semaphorins comprise a large family of phylogenetically conserved secreted and transmembrane glycoproteins, many of which have been implicated in repulsive axon guidance events. The transmembrane semaphorin Sema-1a in Drosophila is expressed on motor axons and is required for the generation of neuromuscular connectivity. Sema-1a can function as an axonal repellent and mediates motor axon defasciculation. Here, by manipulating the levels of Sema-1a and the cell adhesion molecules fasciclin II (Fas II) and connectin (Conn) on motor axons, we provide further evidence that Sema-1a mediates axonal defasciculation events by acting as an axonally localized repellent and that correct motor axon guidance results from a balance between attractive and repulsive guidance cues expressed on motor neurons.     

In vivo time-lapse imaging of synaptic takeover associated with naturally occurring synapse elimination

During development, competition between axons causes permanent removal of synaptic connections, but the dynamics have not been directly observed. Using transgenic mice that express two spectral variants of fluorescent proteins in motor axons, we imaged competing axons at developing neuromuscular junctions in vivo. Typically, one axon withdrew progressively from postsynaptic sites and the competing axon extended axonal processes to occupy those sites. In rare instances when the remaining axon did not reoccupy a site, the postsynaptic receptors rapidly disappeared. Interestingly, the progress and outcome of competition was unpredictable. Moreover, the relative areas occupied by the competitors shifted in favor of one axon and then the other. These results show synaptic competition is not always monotonic and that one axon's contraction in synaptic area is associated with another axon's expansion.     

Innervation of chromatophore muscle fibres in the octopus Eledone cirrhosa

Summary Cephalopod chromatophores are made of a central pigment cell surrounded by 10 to 20 radially arranged muscle fibres under direct nervous control. Innervation of these muscle fibres was studied with anterograde cobalt fills of peripheral nerve bundles and light and electron microscopy. Individual axons branch repeatedly to innervate the muscles of chromatophores scattered over several millimeters. Axons contained in several dermal nerves converge to innervate the same chromatophores. Among the chromaophores, axons were found running either singly or in small bundles, often accompanied by sheath cells. Single chromatophore muscles were innervated by at least one axon running across or along its length. Since nerves terminating on chromatophore muscles are very rare, neuromuscular contact seems to be made en passant. Varicosities of the axons apposed to the muscles are thought to be presynaptic sites. However, morphological differentiations of the pre-or post-synaptic membranes were not visible. Two types of innervating processes were found containing either electron-clear or a mixture of electron-clear and dark-core synaptic vesicles.Supported by a postgraduate award from the University of Aberdeen (GB)     

Coordinated motor neuron axon growth and neuromuscular synaptogenesis are promoted by CPG15 in vivo

We have used in vivo time-lapse two-photon imaging of single motor neuron axons labeled with GFP combined with labeling of presynaptic vesicle clusters and postsynaptic acetylcholine receptors in Xenopus laevis tadpoles to determine the dynamic rearrangement of individual axon branches and synaptogenesis during motor axon arbor development. Control GFP-labeled axons are highly dynamic during the period when axon arbors are elaborating. Axon branches emerge from sites of synaptic vesicle clusters. These data indicate that motor neuron axon elaboration and synaptogenesis are concurrent and iterative. We tested the role of Candidate Plasticity Gene 15 (CPG15, also known as Neuritin), an activity-regulated gene that is expressed in the developing motor neurons in this process. CPG15 expression enhances the development of motor neuron axon arbors by promoting neuromuscular synaptogenesis and by increasing the addition of new axon branches.     

Crustacean Neuromuscular Mechanisms

Properties of crustacean muscle fibers and neuromuscular synapsesare discussed, with particular reference to the problems offast and slow contraction, synaptic diversity, and peripheralinhibition. Electrical and mechanical responses of crustacean muscle fibersare variable, and govern to a large extent the muscle's performance.Fast and slow contractions are often mediated by distinct "phasic"and "tonic" muscle fibers, as in abdominal muscles, in whichsuch fibers are segregated into two parallel sets of muscles.In leg muscles the fibers are often heterogeneous in propertiesand innervation. In doubly-motor-innervated muscles of crabsthe axons producing fast and slow contractions preferentiallyinnervate rapidly and slowly contracting fibers, respectively. Crustacean neuromuscular synapses vary greatly in electricalbehavior and in ultrastructural characteristics. Some motoraxons possess both facilitating and nonfacilitating synapses.The proportion of the different types of synapse associatedwith a motor axon probably determines in large measure the propertiesof the postsynaptic potentials evoked by that axon. Pre-synaptic and post-synaptic inhibition both occur, sometimesin the same muscle. The latter type is more common. Pre-synapticinhibition is thought to be mediated by the action of an inhibitorytransmitter-substance on receptors of the motor nerve terminals.     

The finestructure of nerve branches and neuromuscular junctions in the coxal adductor of the cockroach, Gromphadorhina portentosa

The neuromuscular junctions of a fast coxal adductor of Gromphadorhina portentosa show great variability in both axon terminal diameter and extent of post-junctional sarcoplasmic specializaton. Finestructural equivalents of both cone and brush type nerve endings are present. The large motor axons innervating this muscle are surrounded by a pervasive lemnoblast sheath, leaving the axon surface exposed only in the area of synaptic contact. Connective tissue covers the nerve and fills the spaces between sheath cell processes in the nerve trunk, but is lost after it enters the muscle. The role of sheath cells in nerve function is discussed in the light of these findings.     

In vivo observations of terminal Schwann cells at normal, denervated, and reinnervated mouse neuromuscular junctions

We found a low-molecular-mass, fluorescent dye, Calcein blue am ester (CB), that labels terminal Schwann cells at neuromuscular junctions in vivo without damaging them. This dye was used to follow terminal Schwann cells at neuromuscular junctions in the mouse sternomastoid muscle over periods of days to months. Terminal Schwann cell bodies and processes were stable in their spatial distribution over these intervals, with processes that in most junctions were precisely aligned with motor nerve terminal branches. Three days after nerve cut, the extensive processes elaborated by terminal Schwann cells in denervated muscle were labeled by CB. The number and length of CB-labeled terminal Schwann cell processes decreased between 3 days and 1 month after denervation, suggesting that terminal Schwann cell processes are only transiently maintained in the absence of innervation. During reinnervation after nerve crush, however, terminal Schwann cell processes were extended in advance of axon sprouts, and these processes persisted until reinnervation was completed. By viewing the same junctions twice during reinnervation, we directly observed that axon sprouts used existing Schwann cell processes and chains of cell bodies as substrates for outgrowth. Thus, CB can be used to monitor the dynamic behavior of terminal Schwann cells, whose interactions with motor axons and their terminals are important for junction homeostasis and repair.     

On the role of the 200-kDa neurofilament protein at the developing neuromuscular junction

To examine whether the 200-kDa neurofilament protein (200K NFP) is involved in mechanically stabilizing axons, we studied the developmental appearance of immunoreactivity to nonphosphorylated and phosphorylated 200K NFP at the neuromuscular junction. Polyinnervated rat muscle fibers become singly innervated during the first 3 weeks of postnatal life through the process of synapse elimination. If production or post-translational modification of the 200K NFP is actively involved in imparting mechanical stability on neuromuscular synapses, then the selective presence of this protein in only one of several axons at each developing end plate region might make that one axon selectively resistant to elimination. The remaining axons would then be eliminated. Immunoreactivity to the 200K NFP is present on Gestational Day 14 and can be seen in more than one preterminal axon in the end plate region of a muscle fiber during the period of synapse elimination. These results suggest that the 200K NFP is present and phosphorylated early in development and, although the 200K NFP may increase the mechanical stability of axons, this increased stability does not determine the final outcome of synapse elimination.     

Neuromuscular synapses can form in vivo by incorporation of initially aneural postsynaptic specializations

Synapse formation requires the coordination of pre- and postsynaptic differentiation. An unresolved question is which steps in the process require interactions between pre- and postsynaptic cells, and which proceed cell-autonomously. One current model is that factors released from presynaptic axons organize postsynaptic differentiation directly beneath the nerve terminal. Here, we used neuromuscular junctions (NMJs) of the zebrafish primary motor system to test this model. Clusters of neurotransmitter (acetylcholine) receptors (AChRs) formed in the central region of the myotome, destined to be synapse-rich, before axons extended and even when axon extension was prevented. Time-lapse imaging revealed that pre-existing clusters on early-born slow (adaxial) muscle fibers were incorporated into NMJs as axons advanced. Axons were, however, required for the subsequent remodeling and selective stabilization of synaptic clusters that precisely appose post- to presynaptic elements. Thus, motor axons are dispensable for the initial stages of postsynaptic differentiation but are required for later stages. Moreover, many AChR clusters on later-born fast muscle fibers formed at sites that had already been contacted by axons, suggesting heterogeneity in the signaling mechanisms leading to synapse formation by a single axon.     

Single-cell analysis of Drosophila larval neuromuscular synapses

The neuromuscular system of Drosophila has been widely used in studies on synaptic development. In the embryo, the cellular components of this model system are well established, with uniquely identified motoneurons displaying specific connectivity with distinct muscles. Such knowledge is essential to analyzing axon guidance and synaptic matching mechanisms with single-cell resolution. In contrast, to date the cellular identities of the larval neuromuscular synapses are hardly established. It is not known whether synaptic connections seen in the embryo persist, nor is it known how individual motor endings may differentiate through the larval stages. In this study, we combine single-cell dye labeling of individual synaptic boutons and counterstaining of the entire nervous system to characterize the synaptic partners and bouton differentiation of the 30 motoneuron axons from four nerve branches (ISN, SNa, SNb, and SNd). We also show the cell body locations of 4 larval motoneurons (RP3, RP5, V, and MN13-Ib) and the types of innervation they develop. Our observations support the following: (1) Only 1 motoneuron axon of a given bouton type innervates a single muscle, while up to 4 motoneuron axons of different bouton types can innervate the same muscle. (2) The type of boutons which each motoneuron axon forms is likely influenced by cell-autonomous factors. The data offer a basis for studying the properties of synaptic differentiation, maintenance, and plasticity with a high cellular resolution.     

Molecular and cellular characterization of the glial roof plate of the spinal cord and optic tectum: a possible role for a proteoglycan in the development of an axon barrier.

Certain types of glial structures, located at strategic positions along axon pathways, may provide the mechanical and/or chemical elements for the construction of barriers which can grossly direct the elongation of axons during development. The roof plate, a putative axon barrier, is located along the dorsal midline of the developing spinal cord and may be important for the guidance of the commissural and dorsal column axons. We examined the roof plate to determine the developmental morphology of the region and to determine which molecules were correlated with the barrier function when axons were growing nearby. Light and electron microscopic observations of the roof plate revealed that this glial domain undergoes a dramatic change in shape from a "wedge" with large extracellular spaces between the cell apices at E12.5 to a thin, dense septum with reduced extracellular space at E15.5. Immunocytochemical techniques demonstrated that highly sialylated neural cell adhesion molecule (N-CAM), the carbohydrate recognized by L2 monoclonal antibody, cholinesterase, stage-specific embryonic antigen 1, and a ligand that binds tetragonolobus purpureas agglutinin are expressed by the roof plate. These molecules, however, were also found in other regions of the spinal cord which are permissive or attractive to axon growth. A molecule which is unique to the roof plate when axons grow close to, but do not cross, the dorsal midline is a glycosaminoglycan (GAG), keratan sulfate. Keratan sulfate is also present in the tectal midline and in other noninnervated regions such as the outer epidermis and developing cartilage. Our data suggest that keratan sulfate, alone or in combination with other molecules expressed by the roof plate, may be responsible, in part, for the inhibition of axon elongation through the roof plate in the embryonic spinal cord.     

The transport properties of axonal microtubules establish their polarity orientation

It is well established that axonal microtubules (MTs) are uniformly oriented with their plus ends distal to the neuronal cell body (Heidemann, S. R., J. M. Landers, and M. A. Hamborg. 1981. J. Cell Biol. 91:661-665). However, the mechanisms by which these MTs achieve their uniform polarity orientation are unknown. Current models for axon growth differ with regard to the contributions of MT assembly and transport to the organization and elaboration of the axonal MT array. Do the transport properties or assembly properties of axonal MTs determine their polarity orientation? To distinguish between these possibilities, we wished to study the initiation and outgrowth of axons under conditions that would arrest MT assembly while maintaining substantial levels of preexisting polymer in the cell body that could still be transported into the axon. We found that we could accomplish this by culturing rat sympathetic neurons in the presence of nanomolar levels of vinblastine. In concentrations of the drug up to and including 100 nM, the neurons actively extend axons. The vinblastine- axons are shorter than control axons, but clearly contain MTs. To quantify the effects of the drug on MT mass, we compared the levels of polymer throughout the cell bodies and axons of neurons cultured overnight in the presence of 0, 16, and 50 nM vinblastine with the levels of MT polymer in freshly plated neurons before axon outgrowth. Without drug, the total levels of polymer increase by roughly twofold. At 16 nM vinblastine, the levels of polymer are roughly equal to the levels in freshly plated neurons, while at 50 nM, the levels of polymer are reduced by about half this amount. Thus, 16 nM vinblastine acts as a "kinetic stabilizer" of MTs, while 50 nM results in some net MT disassembly. At both drug concentrations, there is a progressive increase in the levels of MT polymer in the axons as they grow, and a corresponding depletion of polymer from the cell body. These results indicate that highly efficient mechanisms exist in the neuron to transport preassembled MTs from the cell body into the axon. These mechanisms are active even at the expense of the cell body, and even under conditions that promote some MT disassembly in the neuron. MT polarity analyses indicate that the MTs within the vinblastine-axons, like those in control axons, are uniformly plus-end-distal.(ABSTRACT TRUNCATED AT 400 WORDS)     

Chemical degeneration of indolamine axons in rat brain by 5,6-dihydroxytryptamine

Summary Evidence has been obtained by electron microscopy of a direct cytotoxic effect of intraventricularly administered 5,6-dihydroxytryptamine (5,6-DHT) on unmyelinated axons in the rat brain. Ultrastructural signs of axonal damage were observed in areas rich in indolamine nerve terminals as early as 2 hrs after injection. By 6–24 hrs, characteristic and more dramatic signs of degeneration developed, involving coalescence of all axonal constituents—often in combination with a uniform osmiophilic impregnation of the axoplasm—accompanied by engulfment of the dystrophic structures by glial processes. During the next five days, the degenerating axons and axon terminals appeared to be removed by glial cell phagocytosis, whose equivalents were the inclusion of axonal residues into membrane-bound lysosome-like bodies. Concomitantly, there was a progressively increasing number of extremely large and dilated axons in all regions analysed. These axonal swellings, which have an ultramorphology similar to that of dilated stumps of mechanically severed monoamine axons, correspond most probably to proximal, dilated portions of drug-damaged axons.The present results, in combination with biochemical and fluorescence microscopical data, indicate that within a proper dose range the 5,6-DHT-induced degeneration is largely restricted to indolamine axons and axon terminals. However, unselective effects on other unmyelinated axons, on myelin, and on glial cells were observed in narrow subependymal zones close to the lateral ventricles, i.e. close to the injection cannula.Supported by grants from the Deutsche Forschungsgemeinschaft.Supported by grants from the National Institutes of Health, USPHS (NS-06701-06) and from the Swedish Medical Research Council (grants No. B72-14X-712-07B and B72-14X-56-08B).