Search

Corpora amylacea (C.A.) also named polyglucosan bodies (P.B.) are one of the hallmarks of normal brain aging. Although their functions are not yet clear, C.A. increase in number in patients suffering from neurodegenerative diseases. C.A. contain 88% of hexoses and 4% of proteins. Most of the proteins in C.A. are aging or stress proteins such as heat shock proteins, ubiquitinated proteins and advanced glycation end products which are also proinflammatory products. Stimulated by the potential role played by some S100 proteins in the inflammatory process which may be triggered in C.A., we investigated, by immunohistochemistry, the presence of different S100 proteins (S100A1, S100A2, S100A3, S100A4, S100A5, S100A6, S100A8, S100A9, S100A12 and S100B) in C.A. from normal human brain. Among the ten S100 proteins analyzed, nine (S100A) were detected in C.A. Three S100 proteins (S100A8, S100A9, S100A12) which are highly expressed in activated macrophages and used as inflammatory markers were detected in C.A. S100A8 was, in addition, found in thick neuronal processes from the pons. One (S100B) could not be found in C.A. although it was highly expressed in astrocytes. In C.A., the staining intensity was estimated by computer-assisted microscopy and gave the following order: S100A1 congruent withS100A8 congruent with S100A9>S100A5> or =S100A4>S100A12>S100A6> S100A2=S100A3. The potential inflammatory role played by S100 proteins in C.A. is discussed.

A single cerebroventricular injection of ethacrynic acid (EA), a Cl(-)-ATPase inhibitor, induces generalized tonic-clonic convulsions in mice. To clarify whether such convulsive stimulus triggers a long-lasting rearrangement of the neural circuitry culminating in seizure susceptibility, we examined molecular, cellular and behavioral changes following the EA-induced seizure. The expression of immediate early gene c-fos mRNA as an index for cellular activation increased biphasically, with an early transient increase at 60 min and a late prolonged increase on the 10th to 14th day post-EA administration, most remarkably in the hippocampus and pyriform cortex. On the 14th day post-EA seizure, subconvulsive dose of kainic acid (5-17.5 mg/kg) caused severe (stage 5) seizure in 77% of the mice, with 70% mortality. In addition, the expression of nerve growth factor (NGF) also showed biphasic increases with close spatiotemporal correlation with c-fos expression. Moreover, the number of cell somata and the density of axon fibers of parvalbumin (PARV)-positive cells, a subpopulation of GABAergic interneurons, decreased in area dentata, CA1 and CA3 on the 7th and 14th day post-EA seizure. In area dentata and CA1, the density of glutamic acid decarboxylase (GAD)-positive cells also decreased on the 14th day. Thus, the transient EA-induced seizures appear to develop seizure susceptibility by causing damage of a subpopulation of inhibitory interneurons along with increases in the expression of c-fos and NGF in limbic structures., (Copyright 1999 Elsevier Science B.V.)

In adult rodents, proliferating cells in the subventricular zone of lateral ventricle tangentially migrate into the olfactory bulb, where they become the interneurons. The present immunocytochemical analysis revealed that S100A6 (calcyclin), a specific calcium-binding protein of the S100 family, is restrictedly distributed in some astrocytes in the tangential migration pathway of the rat. These results suggest that a particular type of astrocytes containing S100A6 is associated with the tangential migration pathway.

In the present study we examined the laminar distributions of four types of chemically defined subpopulations of non-principal neurons, that is, those immunoreactive for parvalbumin (PV), calretinin (CR), nitric oxide synthase (NOS) and somatostatin (SS), in the rat hippocampus, by estimating their approximate numerical densities (NDs) and percentages in specific layers according to the 'disector' principle. CR-immunoreactive (CR-IR) neurons and NOS-IR neurons were scattered throughout layers, but among layers in each subdivision their NDs were largest in the principal cell layers, where 30-45% of CR-IR and NOS-IR somata in each subdivision were located. In addition, CR-IR and NOS-IR somata were also concentrated at the border between the stratum radiatum (SR) and stratum lacunosum moleculare (SLM) in the CA1 region, where the NDs of these neurons were far larger than those in the SR/SLM as a whole and close to those in the stratum pyramidale (SP) (CR-IR somata at the ventral level and NOS-IR somata at the dorsal level) or larger (NOS-IR neurons at the ventral level). The NDs of CR-IR somata were dorsoventrally different in all layers of the CA3 region, the SR/SLM in the CA1 region and the hilus and the granule cell layer (GCL) of the dentate gyrus (DG), whereas the NDs of NOS-IR somata were dorsoventrally different in all layers of the CA3 region and the SP in the CA1 region. In contrast, approx. 90% of somatostatin-like immunoreactive (SS-LIR) neurons were located in the stratum oriens/alveus (SO/SA) in the CA1 region and in the hilus of the DG, where they were the most predominant cell type among the four types of non-principal cells. In contrast, in the CA3 region, SS-LIR somata were scattered in various layers. The majority (50-70%) of PV-IR neurons were located in the principal cell layers, whereas one-fourth to one-third of them were located in the SO/SA and hilus. The NDs in the SP of the CA1 and CA3 regions showed a significant dorsoventral difference. Although PV-IR somata were most numerous among the four non-principal cell groups in the SP of the dorsal CA1 region, they were not necessarily predominant in the principal layers in other regions, that is, in the ventral CA1 region, CA3 region and DG, where the NDs of CR-IR and/or NOS-IR somata were nearly equal to or larger than that of PV-IR somata. The present study not only reveals the laminar distribution patterns of four types of non-principal neurons in each subdivision quantitatively, but also illustrates the prominent differences in the compositions of four types of non-principal cells in each layer of each subdivision.

Intensities of the immunoreactivities against two GAD isoforms, GAD65 and GAD67, were examined quantitatively in somata of GABAergic neurons in the mouse hippocampus proper. All labeled somata contained both isoforms but showed diverse immunoreactivities against them. The somata showing weak immunoreactivity for GAD65 but moderate to intense immunoreactivity for GAD67 were frequently located in the stratum pyramidale of the CA1 and CA3 regions and were mainly composed of parvalbumin-containing neurons, a particular subpopulation of hippocampal GABAergic neurons. These data revealed additional distinctive properties of parvalbumin-containing neurons and suggest their specialized roles in the hippocampal GABAergic system.

In the present study we examined the distribution of chemically identified subpopulations of nonprincipal neurons in the rat hippocampus, focusing on the dorsoventral differences in their distributions. The subpopulations analyzed were those immunoreactive for parvalbumin, calretinin, nitric oxide synthase, somatostatin, calbindin D28K, vasoactive intestinal polypeptide and cholecystokinin. Using a confocal laser scanning light microscope, we could confirm that the penetration of each immunostaining, except that of calbindin D28K, was complete throughout 50 microns thick sections under our immunostaining conditions. We counted numbers of immunoreactive somata according to the 'dissector' principle, measured areas of hippocampal subdivisions and the thickness of sections, and estimated the approximate numerical densities of these subpopulations, especially for those neurons immunoreactive for nitric oxide synthase, calretinin, somatostatin and parvalbumin. Generally speaking, neurons immunoreactive for parvalbumin showed no significant dorsoventral differences in the numerical densities in any of the subdivisions of the hippocampus, whereas the numerical densities of somata immunoreactive for calretinin, nitric oxide synthase and somatostatin were significantly larger in ventral levels than at dorsal levels of the hippocampus. The numerical density of somatostatin neurons was significantly larger in ventral levels than in dorsal levels of the denate gyrus, and, although not prominent, of the CA1 region. That of nitric oxide synthase positive neurons was significantly larger in ventral levels than in dorsal levels of the CA3 region as well as of the DG but not of the CA1 region. The numerical density of calretinin positive neurons was larger in ventral levels than in dorsal levels of all hippocampal subdivisions. The present study also revealed that dorsal and ventral levels of the hippocampus differ from each other in the composition of their nonprincipal neurons.

It has been believed so far that in the mammalian olfactory bulb, only granule cells form reciprocal synapses with mitral/tufted cells in the external plexiform layer (EPL). However, the present electron microscopic study demonstrated for the first time that another group of immunocytochemically defined interneurons different from granule cells, viz., neurons containing a specific calcium-binding protein parvalbumin, also made reciprocal synapses with mitral/tufted cells in the EPL.

Injections of fluorescent dyes were made in the prefrontal and motor cortex of owl monkeys and retrogradely labeled neurons in the mesencephalon were analyzed for tyrosine hydroxylase and calbindin-D28K immunostaining. Numbers of mesocortical dopaminergic neurons in the dorsal substantia nigra compacta and in the ventral tegmental area also contain calbindin-D28K. This cortically projecting calbindin-D28K containing subpopulation of the dopaminergic mesencephalic cells may be characterized by different electrophysiological properties and a lesser vulnerability to cell death.

The distribution and morphology of neurons containing the Ca-binding proteins parvalbumin (PV), calbindin-D28k (CaBP) and calretinin (CaR) are described in a rostral forebrain region (MNH) of the chick, known to be involved in auditory filial imprinting. PV immunoreactivity is chiefly a marker for numerous large to medium-sized neurons in the neostriatal part of MNH. They show patchy staining of their dendrites, but PV-positive spines are not visible. CaBP is represented in a different neuron population with on the average slightly smaller-sized somata, which carry long, spiny, CaBP-positive dendrites. In contrast to PV and CaBP, CaR immunoreactivity is a marker chiefly for neuropil in MNH but only for few stained neurons. They may be spiny and show the largest size variations. The density of CaR-immunoreactive neuropil is highest in the hyperstriatal part of MNH. Double immunostaining for PV and CaBP reveals that these proteins are expressed mostly in different neuron populations, with only few neurons containing both proteins. These neuron populations appear to form an interconnected network within MNH. A possible relationship between the expression of either Ca-binding protein and the presence of the inhibitory transmitter GABA is also examined. The GABA-antibody labels scattered, very small to medium-sized neurons and dense punctate neuropil. The comparison of the area histograms of somata reveals an overlap with all 3 Ca-binding protein containing cell populations, except for a large proportion of small GABA-positive neurons. The characteristics of immunostained neuron populations are compared to the previously described 3 Golgi-types of neurons in MNH, and possibilities of a functional implication of the proteins in MNH plasticity are examined.

Parvalbumin (PV)-immunoreactive neurons in rat neostriatum were studied under light and electron microscopes. A small number of neurons in the striatum were immunoreactive for PV (a Ca-binding protein). Most of them were also strongly immunoreactive for glutamate decarboxylase but were negative for NADPH-diaphorase activity. Light microscopic analysis revealed that PV-containing neurons have somata with fusiform or polygonal shape and are medium to large in size. The dendrites were smooth and cylindrical at the proximal portion but were varicose at the distal portion. Thin PV-immunoreactive fibers with large boutons were unevenly distributed in the striatum. Electron microscopy revealed that the somata of PV-immunoreactive neurons had a deeply indented nucleus with a nucleolus and often an intranuclear rod. These are the morphological features reported for interneurons of the striatum. Gap junctions formed between two neighboring PV-immunoreactive dendrites. A total of 175 boutons forming synapses with somata and dendrites of PV-immunoreactive neurons were examined. Of these, 115 were small in diameter (less than 1 micron), contained densely packed round vesicles and formed asymmetrical synapses mainly with dendrites. The other 60 boutons formed symmetrical synapses with somata and dendrites of PV-immunoreactive neurons. Both myelinated and unmyelinated axons with boutons were observed. PV-immunoreactive boutons had a diameter of 0.3-2 microns and contained round or elongated vesicles which were about 35 nm in diameter. The boutons formed symmetrical synapses with postsynaptic targets. Of the 100 PV-immunoreactive boutons, 51 were found on somata and proximal dendrites of medium-sized neurons containing a large, round, centrally located nucleus. The others formed synapses with dendrites of various sizes. It was occasionally observed that varicose dendrites free of spines were contacted by a large number of PV-immunoreactive boutons. The study indicates that, in the striatum, immunocytochemistry for PV selectively stains GABAergic interneurons and that the GABAergic interneurons are incorporated in a feed-forward inhibitory circuit of the striatum.

The coexistence of glycine- and PV-immunoreactivities was studied immunocytochemically in the nuclei of the superior olive, trapezoid body, cochlea and lateral lemniscus. All of the PV-immunoreactive neurons in the nuclei of the superior olive and trapezoid body were immunoreactive to glycine but not to GABA. In the dorsal cochlear nucleus, PV-positive neurons were sometimes immunoreactive to glycine. In the ventral nucleus of the lateral lemniscus, PV-positive cells were immunoreactive neither to glycine nor to GABA. Consequently, it was concluded that PV-immunoreactivity was distributed not only in the GABAergic neurons, but also in the glycinergic neurons and possibly in wider neuronal populations.

The organization of the GABAergic system in the rat main olfactory bulb was investigated immunohistochemically using antisera against glutamic acid decarboxylase (GAD), tyrosine hydroxylase (TH), parvalbumin (PV) and methionin-enkephalin-Arg6-Gly7-Leu8 (ENK). Some GABAergic neurons were shown to contain TH immunoreactivity in the glomerular layer, PV immunoreactivity in the external plexiform layer and ENK-like immunoreactivity in the granule cell layer, indicating the stratified organization of the GABAergic subsystems.

The possibility of the colocalization of a Ca2+-binding protein, parvalbumin (PV), with peptides, cholecystokinin (CCK) and somatostatin (SS), all of which are known to be contained gamma-aminobutyric acidergic neurons in the cerebral cortex, was tested in the rat visual cortex by means of immunohistochemistry using the adjacent section technique. No PV immunoreactive (PV-I) neurons showed CCK-like immunoreactivity, and almost none (932 out of 934 PV-I neurons examined) showed SS-like immunoreactivity. These observations suggest that the GABAergic system of the cerebral cortex may consist of some distinct subsystems with different chemical and presumably physiological characteristics.

Fast spiking cells in the CA1 region of the rat hippocampus were revealed as gamma-aminobutyric acid (GABA)ergic non-pyramidal cells containing the calcium-binding protein parvalbumin by intracellular injection of Lucifer yellow in vitro in combination with postembedding parvalbumin immunohistochemistry.

The distribution of Ca2+-binding protein, parvalbumin (PV), containing neurons and their colocalization with glutamic acid decarboxylase (GAD) were studied in the rat hippocampus and dentate gyrus using immunohistochemistry. PV immunoreactive (PV-I) perikarya were concentrated in the granule cell layer and hilus in the dentate gyrus and in the stratum pyramidale and stratum oriens in the CA3 and CA1 regions of the hippocampus. They were rare in the molecular layer of the dentate gyrus, in the stratum radiatum and in the stratum lacunosum-moleculare of the hippocampus. PV-I axon terminals were restricted to the granule cell layer, the stratum pyramidale and the immediately adjoining zones of these layers. Almost all PV-I neurons were also GAD immunoreactive (GAD-I), whereas only about 20% of GAD-I neurons also contained PV. The percentages of GAD-I neurons which were also immunoreactive for PV were dependent on the layer in which they were found; i.e. 40-50% in the stratum pyramidale, 20-30% in the dentate granule cell layer and in the stratum oriens of the CA3 and CA1 regions, 15-20% in the hilus and in the stratum lucidum of CA3 region and only 1-4% in the dentate molecular layer and in the stratum radiatum and the stratum lacunosum-moleculare of the CA3 and CA1 regions. PV-I neurons are a particular subpopulation of GABAergic neurons in the hippocampal formation. Based on their morphology and laminar distribution, they probably include basket cells and axo-axonic cells.

An antibody raised against chick intestinal calbindin D28K was used to study the number and size of calbindin immunoreactive neurones in postmortem human brains from neurologically normal controls and from patients with neuropathologically diagnosed Alzheimer-type dementia (ATD). In the controls, calbindin immunoreactive neurones were observed in all cerebral cortex areas examined including the frontal, temporal and parietal cortices. When compared with the controls, the number and size of calbindin immunoreactive neurones were significantly reduced in the cortices of patients with ATD. These findings suggest that calbindin containing neurones are affected in ATD.

The immunoreactivity of parvalbumin (PV), a Ca2+-binding protein present in a subpopulation of interneurons, was studied in the hippocampal CA1 region during kindling epileptogenesis, induced by tetanic stimulation of the Schaffer collateral/commissural fibers. PV-immunoreactivity was increased in comparison to controls after 13 afterdischarges and after the induction of generalized seizures. A quantification of the number of PV-immunoreactive somata showed an increase of 20% in both stages of kindling. This level had returned to baseline level 31 days after the last seizure. These results imply that changes in PV-immunoreactivity are related to seizure activity rather than to the long-term increase in seizure sensitivity in kindled animals. Co-localization study in controls showed that 32% of PV-immunoreactive somata were also immunopositive for GABA. A colocalization study in stratum oriens and pyramidale on the stimulated side of kindled animals 31 days after the last generalized seizure showed neither a reduction in the number of PV-immunoreactive somata nor in the number of GABA-immunopositive cell bodies that co-localized with PV. In contrast, the number of GABA-immunoreactive somata that did not co-localize with PV was reduced by 50%. It has been shown that a large influx of Ca2+ plays a crucial role in epileptogenesis. Here we demonstrate that the presence of the calcium-binding protein parvalbumin seems to exert a protective effect against the process that leads to a decrease in GABA content.

The characterization of neuron populations by their immunoreactivity against parvalbumin- and calbindin (28-kDa)-antisera has been used to study the postnatal development of the visual diencephalic nucleus rotundus and the mesencephalic nucleus isthmi complex in zebra finches. In nucleus rotundus, parvalbumin-immunoreactivity was restricted to the neuropil during the first 10 days and appears additionally in somata around day 12 where it remains until adulthood. Calbindin-immunoreactivity of the very scarce neuropil and the few somata, which can be observed during the first two weeks, disappears until adulthood. Thus, the adult nucleus rotundus shows an almost complementary distribution of calbindin- and parvalbumin-immunoreactive structures: the numerous, heavily parvalbumin-positive somata, which are surrounded by dense immunoreactive neuropil are in sharp contrast to the complete absence of calbindin-immunoreactive somata. Only a thin rim surrounding this nucleus contains punctate calbindin-positive neuropil. In the nucleus isthmi complex, parvalbumin and calbindin staining patterns show markedly different developmental profiles. While the density of parvalbumin-immunoreactive neuropil in the parvocellular part of the nucleus isthmi continuously increases and the somata remain unstained, the initially heavily calbindin-positive somata gradually lose their immunoreactivity during the first two weeks. In the adult nucleus isthmi complex, parvalbumin- and calbindin show nearly identical staining patterns. A comparison between the two calcium-binding proteins and GABA-immunoreactivity in adult brains revealed different relationships in the two nuclei: while in nucleus rotundus GABA-staining pattern neither resembles that of parvalbumin nor of calbindin, in the nucleus isthmi complex all three staining patterns coincide.

The type and cell size of parvalbumin-immunoreactive (PV-Ir) neurones were examined in 14 postmortem brains from elderly control and Alzheimer-type dementia (ATD) patients with the aid of an image analyser. Morphological features of PV-Ir neurones suggested the existence of PV in the non-pyramidal interneurones in the cerebral cortex. A significant loss of PV-Ir cells was found in the frontal and temporal cortex in ATD. A significant reduction in the size of PV-Ir cells was also noted in the temporal cortex in ATD. These findings suggested that PV-Ir neurones in the cortex are affected in ATD.

Lectins with specific affinity for terminal N-acetylgalactosamine, Vicia villosa agglutinin (VVA) and Glycine max agglutinin (soybean agglutinin; SBA), are shown to stain selectively a subpopulation of GABAergic neurons in the rat cerebral cortex. About 90% of VAA- and/or SBA-stained cells are also parvalbumin-immunoreactive, but no VVA-stained cells showed somatostatin-28-like or cholecystokinin-8-like immunoreactivities.