@misc{oai:ir.soken.ac.jp:00001192,
 author = {関川, 明生 and セキガワ, アキオ and SEKIGAWA, Akio},
 month = {2016-02-17},
 note = {The neocortex is composed of excitatory (pyramidal) and inhibitory (GABAergic nonpyramidal) neurons. Pyramidal neurons receive excitatory synaptic input from their own recurrent collaterals as well as thalamic fibers. Pyramidal neurons also receive inhibitory synaptic input from local GABAergic nonpyramidal cells. This mixture of synaptic input maintains the excitatory and inhibitory balance in the cortex. Neocortical GABAergic cells are morphologically and physiologically heterogeneous, but specific subtypes can be identified based on differential expression of specific peptides and proteins. Individual GABAergic cell subtypes tend to innervate specific surface domains of other cortical cells. Somatostatin-expressing Martinotti cells mostly innervate thin dendritic shafts and spines, whereas parvalbumin fast-spiking (FS) basket cells also make synapses on somata. Thus, cortical inhibition is differentially exerted onto specific cellular domains based on the innervation patterns of different interneuron subtypes. Therefore, to understand the mechanisms that maintain the excitatory and inhibitory activity balance in the cortex, it is necessary to reveal the specific excitatory and inhibitory input patterns onto individual GABAergic cell subtypes.<br />　　GABAergic neuron subtypes show differential dendritic spatial extension, branching patterns, and spine densities. The local input impedance influences local postsynaptic potentials induced by active synaptic conductances, and is in turn dependent on the postsynaptic dendritic geometry. Local synaptic current amplitudes are related to the postsynaptic synapse density and junctional area related to the receptor number. The total excitatory depolarization would be determined by interaction between the activated excitatory and inhibitory synapses. However, it remains to be investigated how local postsynaptic morphologies, important for the local synaptic integration and current transfer to the soma, are related to synaptic density. Furthermore, it is not known if these relationships are different between excitatory and inhibitory terminals onto the various GABAergic neuron subtypes. The cell body integrates all excitatory currents from the dendrite and generates depolarization for spike induction. The differences in excitatory and inhibitory balances would affect the firing regulation a lot.<br />　　The majority of GABAergic neurons can be identified by chemical expression of parvalbumin, calretinin and somatostatin. These chemical classes are further divided into subtypes, such as a somatostatin subpopulation expressing nitric oxide synthase (NOS). Here they have investigated the relationships between postsynaptic density of GABA-positive and GABA-negative terminals onto different GABAergic neuron subtypes.<br />　　First they confirmed that substance P receptors (SPR) were selectively expressed in NOS cells, a subpopulation of somatostatin cells (13% of somatostatin cells in layer II/III, 20% in layer V and 25% in layer VI) by double immunofluorescence. Parvalbumin and calretinin cells were not positive for SPR. Next they labeled the somata and dendrites of 4 chemically defined nonpyramidal neuron subtypes positive for somatostatin, SPR, parvalbumin, or calretinin by pre-embedding immunohistochemistry using Ni-DAB reaction. These sections were embedded in Epon for electron microscopic observations. Some immunostained somata and dendrites were reconstructed 3-dimensionally at the light microscopic level using the Neurolucida system. Immunopositive tissues were serially sectioned in 90 nm thicknesses. To identify GABAergic terminals, they applied GABA postembedding immunohistochemistry to ultrathin sections, detected by colloidal gold particles. Synaptic boutons were quantitatively divided into two classes on the basis of gold particle densities for GABA immunohistochemistry. The particle density differences between GABA-negative and -positive terminals were similar among the materials immunostained for the above 4 chemical markers.<br />　　The labeled somata and dendrites and associated structures were reconstructed from serial electron microscopic images by a 3D reconstruction system using the software package ‘Reconstruct’. From the reconstructed dendrites, they measured the length and surface area, followed by a calculation of the averaged cross-sectional area. In individual reconstructed dendritic segments, they counted GABA-positive and -negative synapses, followed by evaluation of their density per surface area. <br />　　Cell bodies of 4 chemical types were partially reconstructed, and somatic synaptic input patterns were compared between them. GABA-positive synapse densities on the soma were similar between the subtypes, but GABA-negative densities were significantly different. Parvalbumin cells had higher densities of GABA-negative synapses than did calretinin and somatostatin cells. Therefore, the proportion of GABA-positive synapses on the soma was significantly different between the 4 classes. Somatostatin somata had a higher proportion of GABA-positive synapses than did SPR and parvalbumin somata. Calretinin-positive somata had a higher proportion of GABA-positive synapses than those of parvalbumin cells. These indicate that nonpyramidal neuron subtype influences the ratio of inhibitory to excitatory somatic input.<br />　　Dendritic spines were found in somatostatin cells, but not in those of parvalbumin and calretinin cells. Although SPR cells were a subpopulation of somatostatin cells, spines were not identified in SPR dendritic segments.<br />　　The dendritic synaptic densities and cross-sectional areas were well correlated in GABA-negative synapses. Larger dendrites were lower in GABA-negative synapse density, and smaller dendrites had higher synaptic densities. The density dependency on the postsynaptic dendritic dimension was most prominent in SPR cells and least in calretinin cells. On the other hand the correlation between GABA-positive synapse densities and dendritic dimensions was weaker than that of GABA-negative synapses. These data show that GABA synapse density is relatively constant between dendritic locations, but excitatory input density changes according to the postsynaptic dendritic dimension and location.<br />　　They next compared dendritic synaptic densities as a whole. GABA-positive synapse densities on dendrites were similar between the neuronal subtypes, but GABA-negative synaptic densities were significantly different. Calretinin dendrites had lower GABA-negative densities than did parvalbumin and SPR cells. Somatostatin dendrites were lower in GABA-negative densities than were parvalbumin-positive neurons.<br />　　These observations revealed that the GABAergic inhibitory synaptic density is similar between the subtypes, the somata and dendrites, the dendritic surface locations, or the dendritic dimensions. On the other hand, the excitatory density varies between the subtypes. It is higher in dendrites than in somata, and also higher in distal thinner dendrites.<br /><br />, 総研大甲第1171号},
 title = {Quantification of excitatory and inhibitory synapseson GABAergic nonpyramidal cell subtypes　in the rat cerebral cortex},
 year = {}
}