Andréa Dumoulin

Target-Dependent Use of Coreleased Inhibitory Transmitters at Central Synapses

Corelease of GABA and glycine by mixed neurons is a prevalent mode of inhibitory transmission in the vertebrate hindbrain. However, little is known of the functional organization of mixed inhibitory networks. Golgi cells, the main inhibitory interneurons of the cerebellar granular layer, have been shown to contain GABA and glycine. We show here that, in the vestibulocerebellum, Golgi cells contact both granule cells and unipolar brush cells, which are excitatory relay interneurons for vestibular afferences. Whereas IPSCs in granule cells are mediated by GABAA receptors only, Golgi cell inhibition of unipolar brush cells is dominated by glycinergic currents. We further demonstrate that a single Golgi cell can perform pure GABAergic inhibition of granule cells and pure glycinergic inhibition of unipolar brush cells. This specialization results from the differential expression of GABAA and glycine receptors by target cells and not from a segregation of GABA and glycine in presynaptic terminals. Thus, postsynaptic selection of coreleased fast transmitters is used in the CNS to increase the diversity of individual neuronal outputs and achieve target-specific signaling in mixed inhibitory networks.

Formation of mixed glycine and GABAergic synapses in cultured spinal cord neurons

In the spinal cord, GABA and glycine mediate inhibition at separate or mixed synapses containing glycine and/or GABAA receptors (GlyR and GABAAR, respectively). We have analysed here the sequence of events leading to inhibitory synapse formation during synaptogenesis of embryonic spinal cord neurons between 1 and 11 days in vitro (DIV). We used immunocytochemical methods to detect simultaneously an antigen specific to inhibitory terminals, the vesicular inhibitory amino acid transporter (VIAAT), and one of the following postsynaptic elements: GlyR, GABAAR or gephyrin, the anchoring protein of GlyR, which is also associated with GABAAR. Quantitative analysis revealed that until 5 DIV most gephyrin clusters were not adjacent to VIAAT‐positive profiles, but became associated with them at later stages. In contrast, GlyR and GABAAR clustered predominantly in front of VIAAT‐containing terminals at all stages. However, about 10% of receptor aggregates were detected at nonsynaptic loci. The two receptors colocalized in 66.2 ± 2.5% of the inhibitory postsynaptic domains after 11 DIV, while 30.3 ± 2.6% and 3.4 ± 0.8% of them contained only GlyR and GABAAR, respectively. Interestingly, at 3 DIV GABAAR clustered at a postsynaptic location prior to gephyrin and GlyR; GABAAR could thus be the initiating element in the construction of mixed glycine and GABAergic synapses. The late colocalization of gephyrin with GABAAR, and the demonstration by other groups that, in the absence of gephyrin, postsynaptic GABAAR is not detected, suggest that gephyrin is involved in the stabilization of GABAAR rather than in its initial accumulation at synaptic sites.

Cellular transport and membrane dynamics of the glycine receptor

Regulation of synaptic transmission is essential to tune individual-to-network neuronal activity. One way to modulate synaptic strength is to regulate neurotransmitter receptor numbers at postsynaptic sites. This can be achieved either through plasma membrane insertion of receptors derived from intracellular vesicle pools, a process depending on active cytoskeleton transport, or through surface membrane removal via endocytosis. In parallel, lateral diffusion events along the plasma membrane allow the exchange of receptor molecules between synaptic and extrasynaptic compartments, contributing to synaptic strength regulation. In recent years, results obtained from several groups studying glycine receptor (GlyR) trafficking and dynamics shed light on the regulation of synaptic GlyR density. Here, we review (i) proteins and mechanisms involved in GlyR cytoskeletal transport, (ii) the diffusion dynamics of GlyR and of its scaffolding protein gephyrin that control receptor numbers, and its relationship with synaptic plasticity, and (iii) adaptative changes in GlyR diffusion in response to global activity modifications, as a homeostatic mechanism.

Immunolocalization of Bcl-xL/S in the central nervous system of neonatal and adult rats

Abstract.A polyclonal antibody raised against a peptide corresponding to the (2–19) amino-terminal sequence of the Bcl-xL/S protein was used to localize Bcl-x immunostaining in the central nervous system of rats at various postnatal ages. Whereas Bcl-x immunostaining was present in virtually all neurons of young animals (4 days postnatal), this staining became progressively restricted during the course of postnatal development. In adults, Bcl-x immunostaining was particularly strong in certain neurons present in a few hypothalamic nuclei, such as the supraoptic or the arcuate nuclei. Moderate staining was observed in some discrete brain regions, such as the olfactory bulb, the hippocampus, some catecholaminergic nuclei of the brainstem, and the cerebellum. Strong Bcl-x immunostaining was also exhibited in axon-like fibers located in the pyriform cortex, the median eminence, the dorsal medulla oblongata, and spinal cord. Bcl-x immunostaining was also present in astrocytes scattered throughout the white matter in the brain and the spinal cord, but was absent from those located in gray matter. Staining was particularly strongly expressed in reactive astrocytes densely packed along the borders of a central lesion or surrounding them, and in a large number of reactive astrocytes detected at a distance from the lesion. Our data suggest that, in addition to the possible stimulating effects on cell survival generally ascribed to Bcl-x, its maintained expression throughout adulthood or its re-expression following injury characterizes those neuronal or non-neuronal cells of the adult central nervous system that synthesize a range of molecules enabling them to adapt rapidly and successfully to a changing environment.

Mixed Inhibitory Synaptic Balance Correlates with Glutamatergic Synaptic Phenotype in Cerebellar Unipolar Brush Cells

Inhibitory synapses display a great diversity through varying combinations of presynaptic GABA and glycine release and postsynaptic expression of GABA and glycine receptor subtypes. We hypothesized that increased flexibility offered by this dual transmitter system might serve to tune the inhibitory phenotype to the properties of afferent excitatory synaptic inputs in individual cells. Vestibulocerebellar unipolar brush cells (UBC) receive a single glutamatergic synapse from a mossy fiber (MF), which makes them an ideal model to study excitatory–inhibitory interactions. We examined the functional phenotypes of mixed inhibitory synapses formed by Golgi interneurons onto UBCs in rat slices. We show that glycinergic IPSCs are present in all cells. An additional GABAergic component of large amplitude is only detected in a subpopulation of UBCs. This GABAergic phenotype is strictly anti-correlated with the expression of type II, but not type I, metabotropic glutamate receptors (mGluRs) at the MF synapse. Immunohistochemical stainings and agonist applications show that global UBC expression of glycine and GABAA receptors matches the pharmacological profile of IPSCs. Paired recordings of Golgi cells and UBCs confirm the postsynaptic origin of the inhibitory phenotype, including the slow kinetics of glycinergic components. These results strongly suggest the presence of a functional coregulation of excitatory and inhibitory phenotypes at the single-cell level. We propose that slow glycinergic IPSCs may provide an inhibitory tone, setting the gain of the MF to UBC relay, whereas large and fast GABAergic IPSCs may in addition control spike timing in mGluRII-negative UBCs.

Heat Shock Cognate Protein 70 Regulates Gephyrin Clustering

Formation and stabilization of postsynaptic glycine receptor (GlyR) clusters result from their association with the polymerized scaffold protein gephyrin. At the cell surface, lateral diffusion and local trapping of GlyR by synaptic gephyrin clusters is one of the main factors controlling their number. However, the mechanisms regulating gephyrin/GlyR cluster sizes are not fully understood. To identify molecular binding partners able to control gephyrin cluster stability, we performed pull-down assays with full-length or truncated gephyrin forms incubated in a rat spinal cord extract, combined with mass spectrometric analysis. We found that heat shock cognate protein 70 (Hsc70), a constitutive member of the heat shock protein 70 (Hsp70) family, selectively binds to the gephyrin G-domain. Immunoelectron microscopy of mouse spinal cord sections showed that Hsc70 could be colocalized with gephyrin at inhibitory synapses. Furthermore, ternary Hsc70-gephyrin-GlyR coclusters were formed following transfection of COS-7 cells. Upon overexpression of Hsc70 in mouse spinal cord neurons, synaptic accumulation of gephyrin was significantly decreased, but GlyR amounts were unaffected. In the same way, Hsc70 inhibition increased gephyrin accumulation at inhibitory synapses without modifying GlyR clustering. Single particle tracking experiments revealed that the increase of gephyrin molecules reduced GlyR diffusion rates without altering GlyR residency at synapses. Our findings demonstrate that Hsc70 regulates gephyrin polymerization independently of its interaction with GlyR. Therefore, gephyrin polymerization and synaptic clustering of GlyR are uncoupled events.

Biphasic response of spinal GABAergic neurons after a lumbar rhizotomy in the adult rat.

The expression of γ‐aminobutyric acid (GABA) and of the isoforms of the enzyme involved in its synthesis, glutamic acid decarboxylase (GAD), is modified in several rat brain structures in different injury models. The aim of the present work was to determine whether such plasticity of the GABAergic system also occurred in the deafferented adult rat spinal cord, a model where a major reorganization of neural circuits takes place. GABAergic expression following unilateral dorsal rhizotomy was studied by means of non‐radioactive in situ hybridization to detect GADs67 mRNA and by immunohistochemistry to detect GAD67 protein and GABA. Three days following rhizotomy the number of GAD67 mRNA‐expressing neurons was decreased in the superficial layers of the deafferented horn, while GABA immunostaining of axonal fibres located in this region was highly increased. Seven days after lesion, on the other hand, many GAD67 mRNA‐expressing neurons were bilaterally detected in deep dorsal and ventral layers, this expression being correlated with the increased detection of GADs67 immunostained somata and with the reduction of GABA immunostaining of axons. GABA immunostaining was frequently found to be associated with reactive astrocytes that exhibited intense immunostaining for glial fibrillary acidic protein (GFAP) but remained GADs67 negative. These results indicate that degeneration of afferent terminals induces a biphasic response of GABAergic spinal neurons located in the dorsal horn and show that many spinal neurons located in deeper regions re‐express GAD67, suggesting a possible participation of the local GABAergic system in the reorganization of disturbed spinal networks.

The Role of Synaptopodin in Membrane Protein Diffusion in the Dendritic Spine Neck

The dynamic exchange of neurotransmitter receptors at synapses relies on their lateral diffusion in the plasma membrane. At synapses located on dendritic spines this process is limited by the geometry of the spine neck that restricts the passage of membrane proteins. Biochemical compartmentalisation of the spine is believed to underlie the input-specificity of excitatory synapses and to set the scale on which functional changes can occur. Synaptopodin is located predominantly in the neck of dendritic spines, and is thus ideally placed to regulate the exchange of synaptic membrane proteins. The central aim of our study was to assess whether the presence of synaptopodin influences the mobility of membrane proteins in the spine neck and to characterise whether this was due to direct molecular interactions or to spatial constraints that are related to the structural organisation of the neck. Using single particle tracking we have identified a specific effect of synaptopodin on the diffusion of metabotropic mGluR5 receptors in the spine neck. However, super-resolution STORM/PALM imaging showed that this was not due to direct interactions between the two proteins, but that the presence of synaptopodin is associated with an altered local organisation of the F-actin cytoskeleton, that in turn could restrict the diffusion of membrane proteins with large intracellular domains through the spine neck. This study contributes new data on the way in which the spine neck compartmentalises excitatory synapses. Our data complement models that consider the impact of the spine neck as a function of its shape, by showing that the internal organisation of the neck imposes additional physical barriers to membrane protein diffusion.

The Susd2 protein regulates neurite growth and excitatory synaptic density in hippocampal cultures.

Complement control protein (CCP) domains have adhesion properties and are commonly found in proteins that control the complement immune system. However, an increasing number of proteins containing CCP domains have been reported to display neuronal functions. Susd2 is a transmembrane protein containing one CCP domain. It was previously identified as a tumor-reversing protein, but has no characterized function in the CNS. The present study investigates the expression and function of Susd2 in the rat hippocampus. Characterization of Susd2 during development showed a peak in mRNA expression two weeks after birth. In hippocampal neuronal cultures, the same expression profile was observed at 15days in vitro for both mRNA and protein, a time consistent with synaptogenesis in our model. At the subcellular level, Susd2 was located on the soma, axons and dendrites, and appeared to associate preferentially with excitatory synapses. Inhibition of Susd2 by shRNAs led to decreased numbers of excitatory synaptic profiles, exclusively. Also, morphological parameters were studied on young (5DIV) developing neurons. After Susd2 inhibition, an increase in dendritic tree length but a decrease in axon elongation were observed, suggesting changes in adhesion properties. Our results demonstrate a dual role for Susd2 at different developmental stages, and raise the question whether Susd2 and other CCP-containing proteins expressed in the CNS could be function-related.