Synapse selection based on differences in synapse turnover

Rat retinal neurons formed transient synapses with rat muscle cells in culture only during a discrete period in development, from the 20th day of embryonic development to the 7th neonatal day. In contrast, chick embryo spinal cord neurons formed synapses at all developmental stages tested, from the 2nd to the 18th day of embryonic development. The percentage of cells from the spinal cord that formed synapses with muscle cells was maximum at 4 days of embryonic development and decreased thereafter. However, the number of synapses with muscle formed by cells from 8-day embryonic spinal cord did not decrease during 14 days of culture. Under identical conditions, all synapses formed between rat retinal neurons and muscle cells were terminated during 7 days of culture. These results show that differences in the rates of turnover of two populations of cholinergic synapses can result in the selective retention of one population of synapses and the loss of the other, and thereby alter the specificity of synaptic connections.     

Early appearance of and neuronal contribution to agrin-like molecules at embryonic frog nerve-muscle synapses formed in culture.

Antibodies against chicken and Torpedo agrin were used for immunofluorescent staining in order to assess the spatial distribution and temporal appearance of agrin-like molecules at newly formed synaptic contacts in cultures of embryonic Xenopus nerve and muscle cells. The antibodies stained Xenopus neuromuscular junctions and removed ACh receptor (AChR)-aggregating activity from extracts of Xenopus brain. Immunofluorescence was observed at almost all nerve-induced AChR aggregates, even at microaggregates in cocultures as young as 7.5 hr and at nerve-muscle contacts less than 2 hr old. Microdeposits of immunofluorescence extended as far distally as, or farther than, the microaggregates of AChRs along young nerve-muscle contacts. They also occurred along portions of growing neurites that were not in contact with muscle. By contrast, immunofluorescence was rarely observed at the nonsynaptic aggregates of AChRs that form on noninnervated muscle cells. These results raise the possibility that neuronally derived microaggregates of agrin-like molecules may be primary sites of nerve-induced clustering of AChRs, and they indicate that these molecules are present at embryonic nerve-muscle synapses from the very onset of AChR aggregation. The cellular origin of the agrin-like molecules at synapses was examined in cross-species cocultures in which the neurons and muscle cells were obtained from embryos of Xenopus laevis and Rana pipiens. Immunofluorescent staining with anti-agrin antibodies reactive at both Rana and Xenopus neuromuscular junctions revealed immunofluorescence at AChR aggregates along nerve-muscle contacts involving both cross-species combinations. Immunofluorescent staining with an anti-agrin antibody reactive at Rana but not at Xenopus neuromuscular junctions was positive only at cross-species nerve-muscle contacts involving Rana neurons. These results provide the first demonstration that embryonic neurons supply agrin-like molecules to the synapses they form with embryonic muscle cells.     

Human muscle cultured in monolayer and cocultured with fetal rat spinal cord: importance of dorsal root ganglia for achieving successful functional innervation

Adult human muscle cultured in monolayer was cocultured with explants of 13-14-d-old rat embryo using (a) ventral spinal cord (VSC), (b) transverse section of whole spinal cord (WSC), and (c) WSC with dorsal root ganglia (DRG) attached (WSC + DRG). AChR clusters and AChE-positive patches, both at the nerve-muscle contacts, were studied at 5, 12, and 21 d of coculture with each of the 3 spinal cord preparations. In addition, AChE-positive patches were studied after 31-64 d of coculture with WSC + DRG to evaluate further organization of those patches. Compared to VSC and WSC cocultures, WSC + DRG induced significantly more AChR clusters per muscle fiber at the nerve-muscle contacts at 5 d of coculture, and the percentage of muscle fibers containing AChR clusters was higher at all 3 time points quantitated. The number of AChE-positive sites was the same with all 3 spinal cord preparations in early (day 5) cocultures. Between 12 and 21 d of coculture, the number of muscle fibers containing AChE patches increased significantly only with WSC + DRG, correlating with the increased number of contracting muscle fibers in that coculture system. Only in human muscle cocultured with WSC + DRG was successful innervation of the cultured muscle fibers achieved, as manifested by (1) contractions in a continuous rhythm of large groups of muscle fibers that were reversibly blocked by 1 mM d-tubocurarine (aneurally cultured human muscle does not spontaneously contract); (2) well-developed cross-striations throughout the fiber; (3) well-organized AChE-positive sites; and (4) a trend from multifocal toward unifocal innervation of those muscle fibers. Our studies demonstrate that adult human muscle cultured in monolayer can be innervated by fetal rat spinal cord and that, in our system, DRG are essential for achieving functional innervation.     

Possible pathogenic role of muscle cell dysfunction in motor neuron death in spinal muscular atrophy

We have previously shown that myofibers formed by fusion of muscle satellite cells from spinal muscular atrophy (SMA) I or II undergo degeneration 1 to 3 weeks after innervation by rat embryonic spinal cord explants, whereas normal myofibers survive for several months. In the "muscle component" of the coculture, the only cells responsible for the degeneration are the SMA muscle satellite cells. Moreover, SMA muscle satellite cells do not fuse as rapidly as do normal muscle satellite cells. To determine whether death of muscle cells precedes that of motor neurons, we studied the origin and kinetics of release of apoptotic microparticles. In SMA cocultures, motor neuron apoptosis occurred before myofiber degeneration becomes visible, indicating that SMA myofibers were unable to sustain survival of motor neurons. In normal cocultures, motor neuron apoptosis occurred 4 days after innervation. However, it did not continue beyond 2 days. These results strengthen the hypothesis that SMA is due to a defect in neurotrophic muscle cell function.     

Differentiation of glial cells and motor neurons during the formation of neuromuscular junctions in cocultures of rat spinal cord explant and human muscle

Motor axons extending from embryonic rat spinal cord explants form fully mature neuromuscular junctions with cocultured human muscle. This degree of maturation is not observed in muscle innervated by dissociated motor neurons. Glial cells present in the spinal cord explants seem to be, besides remaining interneurons, the major difference between the two culture systems. In light of this observation and the well documented role of glia in neuronal development, it can be hypothesized that differentiated and long-lived neuromuscular junctions form in vitro only if their formation is accompanied by codifferentiation of neuronal and glial cells and if this codifferentiation follows the spatial and temporal pattern observed in vivo. Investigation of this hypothesis necessitates the characterization of neuronal and glial cell development in spinal cord explant-muscle cocultures. No such study has been reported, although these cocultures have been used in numerous studies of neuromuscular junction formation. The aim of this work was therefore to investigate the temporal relationship between neuromuscular junction formation and the differentiation of neuronal and glial cells during the first 3 weeks of coculture, when formation and development of the neuromuscular junction occurs in vitro. The expression of stage-specific markers of neuronal and glial differentiation in these cocultures was characterized by immunocytochemical and biochemical analyses. Differentiation of astrocytes, Schwann cells, and oligodendrocytes proceeded in concert with the differentiation of motor neurons and neuromuscular junction formation. The temporal coincidence between maturation of the neuromuscular junction and lineage progression of neurons and glial cells was similar to that observed in vivo. These findings support the hypothesis that glial cells are a major contributor to maturity of the neuromuscular junction formed in vitro in spinal cord explant-muscle cocultures.     

Accumulation of acetylcholine receptors is a necessary condition for normal accumulation of acetylcholinesterase during in vitro neuromuscular synaptogenesis

To study a step of the very complex processes of the formation of the neuromuscular junction (NMJ), we have analysed the clustering of acetylcholine receptors (AChR) and acetylcholinesterase (AChE) in myotubes cultured in various conditions. On the surface of rat myotubes cultured in the presence of spinal cord cells from embryonic rat, numerous AChE clusters appeared. Such clusters are always co-localized with AChR clusters, but the reverse is not true: the number of AChR clusters largely exceeds that of AChE clusters. Very few AChE clusters formed when such co-cultures were treated with monoclonal antibodies (mAbs) against the main immunogenic region (MIR) of the AChR, which provoke internalization and degradation of the AChRs of the muscular membrane. The total levels of AChE and proportions of molecular forms were unaffected. We also used non-innervated myotubes in which addition of agrin, a protein normally synthesized by motoneurons, transported to nerve terminals and inserted into the synaptic basal lamina, induces the formation of small clusters of AChE. When added to rat myotubes devoid of membrane AChR, agrin-induced AChE clusters did not form. Finally, we analysed the capacity of the variant of the C2 mouse muscle cell line deficient in AChR (1R-) to form clusters of AChE in co-cultures with spinal cord cells from rat: no formation of AChE clusters could be observed. In all these different systems of cultures, the conditions which prevented clustering of AChR (anti-AChR antibodies, deficiency of the variant C2 cell line) also suppressed AChE clustering. We concluded that clustering of AChR is a prerequisite for clustering of AChE, so that NMJ formation implies the sequential accumulation of these two components.     

Influence of spinal cord stimulation on the innervation pattern of muscle fibers in vivo.

In chick embryo, chronic stimulation of the brachial spinal cord at a fast rhythm from days 7 to 18 of development induced an increase in AChE activity sites and ACh receptor (AChR) clusters in slow anterior latissimus dorsi (ALD) muscle. Most AChR clusters and AChE spots were contacted by nerve endings. A previous study showed that such spinal cord stimulation causes changes in ALD muscle properties, especially the appearance of a high proportion of fast type II fibers (Fournier Le Ray et al., 1989). Analysis of the synaptic pattern in different fiber types of experimental ALD muscle indicated a decrease in the distance between successive AChE spots in slow type III fibers compared to controls, whereas the intersynaptic distance in fast type II fibers was very similar to that in the rare fast fibers developing in control ALD. Fast fibers of experimental muscles exhibited less AChR than did slow fibers. The increased number of neuromuscular junctions in ALD muscle after spinal cord stimulation appeared to be preferentially located in slow fibers. Electron microscopy showed no change in the number of axons in ALD nerve after spinal cord stimulation. The activity imposed on brachial motoneurons apparently caused terminal sprouting of ALD nerve in target muscle, thus accounting for the increase in neuromuscular contacts in ALD muscle fibers. Differences in the distribution of nerve contacts indicate that the type of muscle fiber innervated may play a critical role in the synaptic pattern during chick embryogenesis.     

Acetylcholine receptors and acetylcholinesterase accumulate at the nerve-muscle contacts of de novo grown human monolayer muscle cocultured with fetal rat spinal cord

Adult human biopsied muscle grown de novo from myoblasts in monolayer was cocultured with spinal cord explants of 13- to 14-day-old rat embryos. Cocultures were established 15 days after myoblast fusion. Autoradiography of 125I-alpha-bungarotoxin and acetylcholinesterase staining were carried out on 27- to 56-day-old muscle-cord cocultures. Large clusters of 125I-alpha-bungarotoxin, indicating clusters of acetylcholine receptors, were present at nerve-muscle contacts but not elsewhere on the muscle fibers. Accumulation of acetylcholinesterase was also present at nerve-muscle contacts. This study demonstrates that monolayer cultured adult human muscle can be innervated by embryonic rat spinal cord neurons and thus there is no requirement for an original basal lamina tube as it is present when organ-cultured human muscle is innervated.     

Early effects in vitro of the muscular dysgenesis mutation on nervous tissue in the mouse

Muscular dysgenesis (mdg), a disease expressed in embryonic mice, severely affects the formation and differentiation of skeletal musculature. The focus of this investigation was an analysis of dysgenic nervous tissue with special attention centered on interactions between muscle and nerve cells in vitro. Results indicate that mdg/mdg spinal cord cells can form functional neuromuscular junctions in nervehyphen;muscle cocultures and induce contractions in dysgenic muscle. However, dysgenic spinal cord cells induce fewer myotubes to contract and result in a delayed induction of dysgenic myotube contractile activity. Furthermore, mdg/mdg nervous tissue, or its conditioned medium, is associated with a higher incidence of morphologically abnormal myotube contractures. The results from this investgation demonstrate that there are functional abnormalities in both dysgenic muscle and nervous tissues which are stable and expressed for up to 3 weeks in vitro.     

Isoform pattern and AChR aggregation activity of agrin expressed by embryonic chick retinal ganglion neurons

Several isoforms of chick agrin, which differ in their activity to aggregate AChRs at the neuromuscular junction, are generated by alternative splicing at splice site B. We analyzed the isoform pattern and the functional properties of agrin in a defined population of CNS neurons. At all developmental stages retinal ganglion cells purified by immunopanning expressed the agrin B0, B11, and B19 isoforms. Single-cell RT-PCR of individual retinal ganglion cells revealed simultaneous expression of B0 and B11 isoforms in about half of the neurons analyzed. Despite the expression of agrin isoforms active in AChR aggregation, ganglion cells did not aggregate AChRs when cocultured with myotubes. Addition of exogenous agrin to myotube-ganglion cell cocultures indicated that AChR aggregation is inhibited. These results demonstrate that a defined population of CNS neurons can simultaneously express several agrin isoforms and that the AChR aggregation activity of agrin might be regulated not only by alternative splicing but also on the protein level.     

Cell-type specific organization of glycine receptor clusters in the mammalian spinal cord

Glycinergic synapses play a major role in shaping the activity of spinal cord neurons. The spatial organization of postsynaptic receptors is likely to determine many functional parameters at these synapses and is probably related to the integrative capabilities of different neurons. In the present study, we have investigated the organization of gephyrin expression along the dendritic membranes of α- and γ-motoneurons, Ia inhibitory interneurons, and Renshaw cells. Gephyrin is a protein responsible for the postsynaptic clustering of glycine receptors, and the features of gephyrin and glycine receptor α₁-subunit immunofluorescent clusters displayed similar characteristics on ventral horn spinal neurons. However, the density of clusters and their topographical organization and architecture varied widely in different neurons and in different dendritic regions. For motoneurons and Ia inhibitory interneurons, cluster size and complexity increased with distance from the soma, perhaps as a mechanism to enhance the influence of distal synapses. Renshaw cells were special in that they displayed an abundant complement of large and morphologically complex clusters concentrated in their somas and proximal dendrites. Serial electron microscopy confirmed that the various immunoreactivity patterns observed with immunofluorescence accurately parallel the variable organization of pre- and postsynaptic active zones of glycinergic synapses. Finally, synaptic boutons from single-labeled axons of glycinergic neurons (Ia inhibitory interneurons) were also associated with postsynaptic receptor clusters of variable shapes and configurations. Our results indicate that mechanisms regulating receptor clustering do so primarily in the context of the postsynaptic neuron identity and localization in the dendritic arbor. J. Comp. Neurol. 379:150-170, 1997. © 1997 Wiley-Liss, Inc.     

Schwann cells and astrocytes induce synapse formation by spinal motor neurons in culture

Glia constitute 90% of cells in the human nervous system, but relatively little is known about their functions. We have been focusing on the potential synaptic roles of glia in the CNS. We recently found that astrocytes increase the number of mature, functional synapses on retinal ganglion cells (RGCs) by sevenfold and are required for synaptic maintenance in vitro. These observations raised the question of whether glia similarly enhance synapse formation by other neuron types. Here we have investigated whether highly purified motor neurons isolated from developing rat spinal cords are able to form synapses in the absence of glia or whether glia similarly enhance synapse number. We show that spinal motor neurons (SMNs) form few synapses unless Schwann cells or astrocytes are present. Schwann cells increase the number of functional synapses by ninefold as measured by immunostaining, and increase spontaneous synaptic activity by several hundredfold. Surprisingly, the synapses formed between spinal motor neurons were primarily glutamatergic, as they could be blocked by CNQX. This synapse-promoting activity is not mediated by direct glial-neuronal cell contact but rather is mediated by secreted molecule(s) from the Schwann cells, as we previously found for astrocytes. Interestingly, the synapse-promoting activity from astrocytes and Schwann cells was functionally similar: Schwann cells also promoted synapse formation between retinal ganglion cells, and astrocytes promoted synapse formation between spinal motor neurons. These studies show that both astrocytes and Schwann cells strongly promote synapse formation between spinal motor neurons and demonstrate that glial regulation of synaptogenesis extends to other neuron types.     

Interferon-γ Alters Nerve-Induced Redistribution of Acetylcholine Receptors in Cultured Rat Skeletal Muscle Cells

The influence of recombinant interferon-γ (rIFN-γ) on the development of acetylcholine receptor (AChR) aggregates in cocultures of rat embryonic muscle cells and spinal cord neurons was studied by counting the number of AChR aggregates in relation to cholinergic nerve fibers coming to the muscle fibers. rIFN-γ caused no decrease in the number of cholinergic nerve fibers, but inhibited the increase in the number of AChR aggregates that occurs early during cocultivation and is an early sign in the development of neuromuscular junctions. rIFN-γ stimulated release of nitric oxide, but no effects on aggregation of AChRs occurred after exposure to a nitric oxide synthase inhibitor, l-N^G-monomethylarginine, or by the addition of nitroprusside, a generator of nitric oxide. No effect was seen on the number of AChR aggregates when the cultures were exposed to rIFN-γ at later time points of cocultivation, when the increase in number of AChRs had already occurred. These studies indicate that the key immunoregulatory cytokine IFN-γ can cause alterations in the early process of synapse formation and that these effects are independent of the nitric oxide release caused by the cytokine.     

Different degradation rates of junctional and extrajunctional acetylcholine receptors of human muscle cultured in monolayer and innervated by fetal rat spinal cord neurons.

It is well demonstrated that in intact animals the degradation rate of the junctional acetylcholine receptor (AChR) is significantly slower than that of the extrajunctional receptor. Such data, however, are not available for human AChRs because the required experimentation cannot be performed in humans. We have now studied the degradation rate of the junctional and extrajunctional AChRs, utilizing our tissue culture model, in which well-differentiated neuromuscular junctions (NMJs) form on human muscle cultured in monolayer and innervated long-term by fetal rat spinal cord neurons. Half-life of AChRs was studied by a method utilizing the autoradiography of 125I-alpha bungarotoxin and computerized video image analysis. Extrajunctional AChRs degraded with a half-life of 1.3 days whereas junctional AChRs degraded with a half-life of 3.5 days. Our studies demonstrate for the first time that in innervated cultured human muscle: (a) the life span of human junctional AChR, is approximately 3 times longer than that of the extrajunctional AChR and (b) the stability of human AChR is neuronally regulated. This system can now be applied to evaluate the influence of pharmacologic agents on the stability of human junctional AChR, which is of potential importance in the treatment of myasthenia gravis and other diseases of the NMJ.     

Inhibition by transforming growth factor beta of choline acetyltransferase stimulation in a co-culture of spinal cord and muscle cells from mice

Choline acetyltransferase (CAT) activity increased 11-fold in co-cultures of spinal cord and muscle cells from fetal mice relative to cultures of spinal cord cells alone. The addition of transforming growth factor-beta (TGF-beta) to the medium at 30 pM throughout the culture period inhibited the increase of CAT activity in the co-cultures, but did not affect the activity in cultures of spinal cord cells alone. TGF-beta did not inhibit glutamic acid decarboxylase activity in the co-cultures. Other growth factors such as epidermal growth factor, fibroblast growth factor and beta-NGF had little or no effect on CAT activity. TGF-beta markedly inhibited the fusion of myoblasts to myotubes and the expression of marker enzymes for muscle differentiation. When TGF-beta was included during muscle culture and removed before inoculation with spinal cord cells, myoblasts did not subsequently form myotubes. CAT activity in the spinal cord cells, however, markedly increased in co-cultures with the undifferentiated myoblasts. When TGF-beta was added to the co-cultures after myotube formation was complete, the increase in CAT activity was inhibited according to the length of TGF-beta treatment. These results suggest that TGF-beta inhibits the muscle-induced stimulation of CAT activity by inhibiting the production, secretion and/or action of trophic factors from muscle.     

Synaptic connections formed by grafts of different types of cholinergic neurons in the host hippocampus

The present experiment was performed to determine whether different types of grafted central cholinergic neurons are able to form synaptic contacts with host hippocampal neurons. Grafts from the septal-diagonal band area, which contain the neurons that normally innervate the hippocampal formation, were compared to those from the nucleus basalis magnocellularis region (NBM), the striatum, the pontomesencephalic tegmentum of the brain stem, and the spinal cord. The regions were dissected from 14- to 16-day-old rat fetuses, and the same number of viable cells (35 x 10(4] from each of the different regions was stereotaxically injected as a cell suspension into the hippocampus of rats subjected to a complete fimbria-fornix lesion, transecting the intrinsic septohippocampal pathways. At 14 to 17 weeks after transplantation, the brains were processed for choline acetyltransferase (ChAT) immunocytochemistry at the light and electron microscopic levels and acetylcholinesterase (AChE) histochemistry at the light microscopic level. There was a great variation in the number of surviving ChAT-positive cells among the different graft types. The septal grafts contained the highest number of ChAT-positive cells, and the striatal grafts showed the lowest numbers. The NBM, brain stem, and spinal cord grafts were in between. The differences in the number of ChAT-positive neurons between the groups matched, in general, the differences found in the magnitude of graft-derived AChE-positive fiber growth into the host hippocampal formation. At the electron microscopical level, all types of grafts were capable of forming synaptic contacts with host elements, however, with vast differences in the number of synapses found. The septal grafts produced the highest number of contacts, whereas the striatal and spinal cord grafts produced very few contacts. The ultrastructure of the cholinergic fibers from grafts obtained from the forebrain areas, i.e., septum, NBM, and striatum all appeared normal, whereas brain stem and spinal cord grafts produced different types of anomalies. The results show that grafted cholinergic neurons, that normally do not innervate the hippocampus, can send axons and form synaptic contacts in the host hippocampus. The ability to reinnervate the denervated hippocampal target appears to be shared by the embryologically closely related forebrain cholinergic neuron types, i.e., the septal, NBM, and striatal neurons. The marked differences in overall fiber ingrowth and number of synapses observed between these different types of grafts could be explained largely on the basis of differences in survivability of each grafted neuron type. By contrast, the reinnervation obtained from the grafted brain stem and spinal cord neurons were both quantitatively and qualitatively abnormal.(ABSTRACT TRUNCATED AT 400 WORDS)     

Cultured smooth muscle targets lack survival activity for ciliary ganglion neurons

Cultured neurons require specific trophic agents in order to survive. This dependence is thought to resemble the neuron-target interdependence that develops in vivo during synaptogenesis and neuronal cell death. The notion that neurons in general derive trophic support from their synaptic targets is based primarily on studies of peripheral neurons and motor neurons. To assess the general applicability of this nerve-target relationship, we tested the ability of vascular smooth muscle (VSM) to support dissociated neurons from the chick ciliary ganglion. The ciliary ganglion contains 2 distinct neuronal populations, one of which innervates striated muscle, the other VSM. Striated muscle cocultures are known to support all of the neurons in the ganglion for extended periods. Dissociated neurons were therefore cocultured in microwells containing VSM derived from the rat or chick aorta and from the choroid coat of the chick eye. Surviving neurons were counted after 1, 2, 5, and 7 d. Striated muscle is able to support full neuronal survival in the same assay. However, in no case was VSM capable of contributing to neuronal survival in vitro. The neurons in the VSM cocultures were able to form neurites and make contacts with their putative targets, as confirmed by scanning electron and light microscopy. The presence of viable and differentiated smooth muscle cells was demonstrated in the cultures by transmission electron microscopy and analysis of smooth muscle alpha-actin. The failure of VSM and even the choroid target tissue to support the survival of their innervating neurons suggests that novel mechanisms may operate to provide trophic support for neurons innervating VSM targets.     

Trophic support by neural explants of cultured muscle fibers

The influence of various neural explants on the morphology and the survival of chick muscle fibers was studied. A method was developed to evaluate the condition of the muscle fibers using the following four morphological parameters: cross striation, thickness, number of fibers, and absence of vacuoles. Chick as well as mouse spinal cord explants appeared to have a distinct favorable influence on the muscle fibers. Chick ciliary ganglia and mouse cortex explants had less effect and chick sympathetic ganglia and mouse dorsal root ganglia had no effect. Innervation by spinal cord neurons did not lead to a change in resting membrane potential of the muscle fibers. The amount of cross striation in muscle fibers in the vicinity of mouse spinal cord explants was positively correlated with the frequency of spontaneous end-plate potentials in these muscle fibers. d-Tubocurarine did not interfere with the trophic support of muscle cells by nerve cells, although it reversibly blocked neuromuscular transmission throughout the experiment. This demonstrates that neither the acetylcholine receptor nor the activity induced in the muscle fiber mediate the trophic action. The data suggest that a humoral trophic factor, released at the neuromuscular junction or at a region of close cell-cell contact, is needed for normal development and maintenance of muscle fiber morphology.     

Target-dependent synaptogenesis: Inductive effect of pituitary melanotrophs on their central innervation

The glandular activity of the vertebrate pituitary intermediate lobe (IL) is regulated by direct cellular innervation, in contrast with the purely humoral regulation of adjacent pituitary anterior lobe (AL). Thus in the rat IL, melanotrophs receive a dopaminergic and GABAergic innervation from the basal hypothalamus, which tonically inhibit their glandular activity. We studied this model of neuron–target interactions in cocultures in defined medium of fetal hypothalamic neurons with neonate pituitary glandular cells. In the cocultures with IL cells, neuroglandular contacts occurred after 4 days in vitro (DIV) but required another 8 DIV to exhibit ultrastructural and immunocytochemical features of fully differentiated functional synapses; by contrast, neuroneuronal synapses developed much faster and could already be detected after 4 DIV. In the cocultures with AL cells, neuroglandular contacts never mature in differentiated synapses. Confocal microscope observation revealed that dopaminergic neurons, which represented less than 1% of total neurons in the cocultures, established 50% of the synapses detected on the melanotrophs. These cells are thus able, contrary to the AL cells, to promote the establishment of functional synapses and, to some extent, to select their specific innervation. Synapse 27:267–277, 1997. © 1997 Wiley-Liss, Inc.     

Possible trophic role on the neuromuscular junction of a neuropeptide co-existing with acetylcholine in motor neurons of the spinal cord

The Calcitonin-Gene Related Peptide (CGRP), a neuropeptide present in chick spinal cord motoneurons, increases the levels of surface acetylcholine receptor (AChR) and of the AChR alpha-subunit mRNA in cultured chick myotubes. Cholera toxin (CT), an activator of adenylate cyclase, produces a similar effect which does not add up with that of CGRP. Consistent with this observation, CGRP increases the content of cyclic AMP in chick muscle cells in culture. Tetrodotoxin (TTX), a blocker of voltage-sensitive Na+ channels, elevates the levels of AChR and of AChR alpha-subunit mRNA. This effect is additive with that of CGRP or CT. TPA (12-O-tetradecanoyl phorbol-13-acetate), an activator of protein kinase C, decreases the level of AChR but has no effect on the level of AChR alpha-subunit mRNA. Interestingly, TPA inhibits the increase of AChR alpha-subunit mRNA caused by TTX without affecting that produced by CGRP or CT. These data suggest that CGRP, which coexists with acetylcholine in spinal cord motoneurons, could be one of the anterograde factors (or a model of such factor) responsible for the enhanced expression of the genes coding for AChR subunits in subneural nuclei, via the activation of adenylate cyclase. Muscle electrical activity would then inhibit the expression of the same genes in extrajunctional nuclei, via another intracellular pathway.