Light and electron microscopic study of corticorubral synapses in adult cat: evidence for extensive synaptic remodeling during postnatal development.

Spine-like dendritic protrusions (SLDPs) emanating from developing dendrites have been proposed to play an important role in early synaptogenesis. We previously analyzed synaptic termination sites on soma-dendritic membrane of newborn cats and found that corticorubral (CR) axons form synapses preferentially on SLDPs (Saito et al., 1997). In the present study, we examined CR synapses in adult cats to elucidate the maturation process of CR synapses in relation to SLDPs. Electron microscopic observation of serial thin sections of Phaseolus vulgaris-leucoagglutinin-labeled axons revealed that approximately 60% of CR terminals in adult cats formed synapses on dendritic spines. We also found that CR axons terminate on dendritic spines originating from the intermediate or distal dendrites of rubrospinal cells (more than 200 microm apart from the soma), in contrast to kittens in which CR fibers terminate on SLDPs originating from the proximal dendrites (less than 100 microm apart from the soma) of rubrospinal cells (Saito et al. [1997] J. Neurosci. 17:8792-8803). These results suggest that CR synapses undergo remarkable remodeling after initial termination on SLDP during postnatal development.

Morphology of individual axons in crossed corticorubral projections in developing cats and effects of partial denervation.

Ordered neuronal connections in mature brains are thought to be sculpted from initially diffuse projections by elimination of inappropriate projections and strengthening of appropriate ones. Although evidence suggests that neuronal activity plays a role in these processes, the mechanism behind the modification of neuronal connections remains obscure. To gain insight into the mechanisms of axonal elimination and projection strengthening, we examined the morphology of individual axons that were to be eliminated as well as the consequences of partial denervation. While corticorubral projections in adult cats are thought to be uncrossed, early in postnatal development and after early unilateral lesions to the sensorimotor cortex, however, a significant amount of crossed corticorubral projections occurs. We examined the morphology of individual corticorubral axons in fetal cats and kittens from embryonic day 59 to postnatal day 48 and those that had received early unilateral lesions to the cortex, by serial reconstruction of Phaseolus-vulgaris-leucoagglutinin- or biocytin-labeled axons. For about 2 weeks during pre- and postnatal development, crossed axons remained simple in morphology, with few branches. Thereafter, they showed an increase in branch number, but then began to show fewer branches again. Axons and their collaterals were found in nonrestricted areas of the red nucleus (RN) throughout the period of observation, indicating that axons can sit at an inappropriate target for weeks but fail to ramify. In contrast, crossed corticorubral axons in kittens with cortical lesions showed terminal-arbor-like structures in the RN region that are in mirror symmetry to topographically appropriate areas in the ipsilateral RN, although some showed simple morphology without arbors. These complicated forms of morphology of individual axons during development and after partial denervation may not be explained by a simple activity-dependent mechanism.

Axonal-growth-associated intracellular molecule in the rat central nervous system recognized by the monoclonal antibody 5H.

In order to discover molecules involved in axonal outgrowth during development of the central nervous system (CNS), monoclonal antibodies (MAbs) were raised against homogenates of the early postnatal rat cerebral cortex. A novel MAb, 5H, was obtained which recognizes a developmentally regulated antigen in the CNS. In the cerebral cortex of postnatal (PN) day 0-6 rats, 5H-immunoreactive punctate or fiber-like processes were observed. In the PN day 12 cortex, 5H immunostaining was mostly present as puncta. The expression of the 5H antigens decreased in the second PN week and completely disappeared from the cortex of rats older than PN day 18. In the hippocampal region, 5H immunoreactivity exhibited similar developmental changes, but with a more prolonged period of expression of the antigen (up to 5 weeks postnatally). They were also absent from the hippocampus of the adult rat. Similar developmental changes were also observed in the other CNS regions, but the only region in the adult CNS that showed 5H immunoreactivity was the olfactory bulb, in which synaptic turnover normally occurs even in the adult. Immunoelectron microscopy revealed that 5H immunoreactivity was localized to the cytoplasm of neuronal processes, including synaptic boutons. The expression of the 5H antigen in neurons cultured from rat cerebral cortex obtained from embryonic day 16 to PN day 6 was mostly restricted to neurite tips, including growth cones. These results suggest that the MAb 5H antigen is associated with axonal outgrowth during early development in the CNS.

Electron microscopic localization and expression on cultured cells of a novel surface-associated molecule in the rat central nervous system.

Subcellular localization and in vitro expression of a newly raised monoclonal antibody 374 antigen, which outlines a subset of neurons in various regions of the rat CNS, were studied. Electron microscopy revealed that the immunoreactivity is associated with the outer surfaces of neuronal subsets, the surfaces of glial cell processes facing such neurons and less frequently, with the Golgi apparatus and inner cell surfaces. The monoclonal antibody also stained part of the cytoplasm of a population of dendrites. In order to test whether the antigen originates from neurons, we examined the expression of the antigen in cultured cortical cells. Double immunofluorescence staining with neuron-specific enolase and glial fibrillary acidic protein revealed that the antigen is expressed on a subset of neuronal cells and to a lesser extent on glial cells. The antigen is considered to be expressed on the outer surface because the cells were stained by the monoclonal antibody irrespective of whether they were alive or fixed. The monoclonal antibody 374 antigen may be a novel neuronal surface molecule which plays a role in neuron-neuron or glial-neuron interactions in the CNS.

Plasticity of neuronal connections in developing brains of mammals.

Although mature nervous systems show substantial malleability following various surgical or environmental manipulations, developing brains show far more prominent plasticity, particularly in terms of morphological features. Neuronal circuits, for example, can be dramatically rewired following neonatal but not adult brain lesions. It remains unknown why neuronal circuits in developing brains show such remarkable plasticity. A number of anatomical and physiological studies suggest that there are transient projections in developing brains and they are eliminated by cell death and/or collateral elimination as development proceeds. This raises a possibility that aberrant projections observed following various surgical or environmental manipulations such as partial denervation, results from retention or stabilization of transient projections. However, evidence suggests that cell death does not play an important role in developmental fine-tuning of neuronal projections. Furthermore, although the elimination of axon collaterals takes place, individual neurons appear to elaborate axonal arbors in appropriate target areas, resulting in a net increase in the size of axonal arbor emerging from individual neurons. In accord with these observations, the number of synapses appear to increase during the period when axonal elimination proceeds. Taken together, reinforcement of appropriate projections rather than elimination of excessive connections plays a major role in developmental specification of neuronal connections. Appearance of aberrant projections after partial denervation may not be a consequence of disordered axonal growth, since they form topographic maps which precisely mirrors those for normal projections. They may be induced due to reinforcement of pre-existing neuronal connections rather than to construction of novel pathways. Observations of axonal morphology in denervated areas indicate that lesion-induced enlargement of projections is due to transformation of axonal morphology, from simple and poorly branched to multiply branched. Perhaps such simple and poorly branched axons in inappropriate target areas may represent ones in the course of elimination but they may serve as a source of sprouting when denervated. In other words, after total elimination of axons any surgical or environmental manipulation cannot induce enlargement of projections. The mechanisms underlying such modifiability of neuronal connections remains unclarified but possible participation of an activity-dependent competitive mechanism is discussed.

Developing corticorubral axons of the cat form synapses on filopodial dendritic protrusions.

Developing neurons transiently grow numerous spine- or filopodium-like dendritic protrusions (SLDPs). Electron microscopy on identified input and intracellular staining of postsynaptic cells were performed to gain insight into their significance. Newborn kitten-corticorubral axons, labelled with biocytin, commonly made synapses on SLDP, often multiply invaginated by the SLDPs. Correspondingly, intracellularly labelled kitten rubrospinal cells had numerous SLDPs. Taking into account that corticorubral synapses are largely formed on dendritic shafts in adult cats, it is likely that the SLDPs play some important role in the development of corticorubral synapses. We hypothesize that rubrospinal cells elongate SLDPs searching for corticorubral axons to form synapses.

Usefulness of thyroglobulin antibody detected by ultrasensitive enzyme immunoassay: a good parameter for immune surveillance in healthy subjects and for prediction of post-partum thyroid dysfunction.

OBJECTIVE: Using newly developed ultrasensitive enzyme immunoassay (EIA) for thyroglobulin antibody (TgAb), we have evaluated physiological and pathological implications of the antibody in healthy subjects as well as in autoimmune thyroid diseases. MEASUREMENTS: This EIA was based on the immune complex transfer method, and was 10(4)-fold more sensitive compared with the conventional haemagglutination assay (HA); the detection limit was 0.1 micrograms IgG/I, and the specificity of the assay was confirmed from the unequivocal decrease in the fluorescence intensity by the preincubation of test serum with Tg and/or inactive beta-D-galactosidase which blocks antibodies to the enzyme. RESULTS: TgAb was detectable in 159 (91%) of 175 healthy subjects aged 3rd to 7th decade (96 men and 79 women), and did not exhibit age or sex-associated change. In nine healthy women, the TgAb level significantly decreased as pregnancy progressed but increased transiently after delivery. TgAb was detectable in 52 (98%) of 53 patients with Graves' disease and all (100%) of 107 patients with chronic thyroiditis. Abnormal high TgAb values (> 40 micrograms/I), determined from the 95th percentile in healthy subjects, were shown in 40 (75%) with the former disease and 94 (88%) with the latter disease. Moreover, in 14 goitrous patients with biopsy-proved chronic thyroiditis with negative HA results, 12 (86%) showed abnormal high TgAb levels. In 69 patients with post-partum thyrotoxicosis in Graves' disease, 15 (79%) of 19 patients with the TgAb level of more than 2 x 10(3) micrograms/I in early pregnancy showed destructive thyrotoxicosis and 46 (92%) of 50 with less than this level showed stimulative thyrotoxicosis. This TgAb test could discriminate the two types of thyrotoxicosis more clearly than could the conventional TGHA test. In chronic thyroiditis, the mean TgAb value in early pregnancy was significantly higher in patients with postpartum hypothyroidism than in those without thyroid dysfunction. Hypothyroidism developed in 80% of the patients with a TgAb value of more than 10(3) micrograms/I. CONCLUSIONS: The ultrasensitive TgAb EIA was useful for detecting the physiological changes in autoantibody formation in healthy subjects and the TgAb value was useful for predicting post-partum thyroid dysfunction in autoimmune thyroid diseases. This EIA is useful for the evaluation of the immune surveillance in patients with autoimmune thyroid diseases as well as in healthy subjects.

Ultrastructural localization of telencephalin, a telencephalon-specific membrane glycoprotein, in rabbit olfactory bulb.

Localization of a telencephalon-specific glycoprotein, telencephalin (TCLN), in the olfactory bulb of the rabbit was studied with an electron microscope. Anti-TCLN antisera appeared to stain plasma membrane, Golgi apparatus and multivesicular bodies of granule cells which are local circuit interneurons in the bulb. Principal neurons, mitral and tufted cells, were not immunoreactive. No glial cells showed immunoreactivity. Thus, expression of telencephalin is specific not only to the telencephalic segment of the brain, but also to the neuronal types.

Use of inactive beta-D-galactosidase for elimination of interference by anti-beta-D-galactosidase antibodies in immune complex transfer enzyme immunoassay for anti-thyroglobulin IgG in serum using beta-D-galactosidase from Escherichia coli as label.

A novel enzyme immunoassay (immune complex transfer enzyme immunoassay) for anti-thyroglobulin IgG using beta-D-galactosidase from Escherichia coli as label was reported previously. This immunoassay was highly sensitive in demonstrating anti-thyroglobulin IgG not only in all patients with Graves' disease and chronic thyroiditis but also in a large proportion of healthy subjects. However, the detection of anti-thyroglobulin IgG at low levels in some serum samples was difficult, probably due to the presence of anti-beta-D-galactosidase antibodies. In the present study, the use of inactive beta-D-galactosidase was tested for elimination of interference by anti-beta-D-galactosidase antibodies. Preincubation of serum samples with excess of inactive beta-D-galactosidase resulted in sufficiently low backgrounds to detect low levels of anti-thyroglobulin IgG with little effect on the dose-response of anti-thyroglobulin IgG. As a result, it was revealed that anti-thyroglobulin IgG was present in almost all healthy subjects as well as all patients with Graves' disease and chronic thyroiditis.

Immunocytochemical study of GABAergic neurons containing the calcium-binding protein parvalbumin in the rat hippocampus.

The structural features of PV-immunoreactive (PV-I) neurons, a particular subpopulation of GABAergic neurons, in the hippocampus were studied by immunocytochemistry. The PV-I cell bodies were concentrated within the stratum pyramidale (SP) and stratum oriens (SO) in the hippocampus. PV-I puncta were frequent in SP, while they were rarely seen in other layers. The dendritic arborization of PV-I neurons resembled that of some of the nonpyramidal cells observed after Golgi-impregnation. The most commonly observed PV-I neurons had their perikarya located in SP with dendrites extending into SO and the stratum radiatum (SR). Most of the dendrites in SR had typical beaded or varicose segments. The dendrites extending into SO had few beaded parts. There were many bipolar and multipolar neurons with smooth dendrites in SO, but only a small number of such multipolar neurons in SR. An electron microscopic analysis revealed that PV-I products were located to perikarya, dendrites, myelinated axons and synaptic boutons. The perikarya of PV-I neurons exhibited several ultrastructural features of nonpyramidal cells, e.g., abundant cisternae of endoplasmic reticulum, mitochondria and other perikaryal organelles, an infolded nuclear envelope and intranuclear inclusions. They received many asymmetric synapses with round presynaptic vesicles. There were numerous PV-I boutons, presumably axonal endings, covering the pyramidal cell bodies. The PV-I boutons also contacted the axon initial segments and proximal dendrites of the pyramidal cells. In addition PV-I terminals were found on somata and dendrites of both PV-I or PV-negative nonpyramidal cells. The results suggest that PV-containing neurons include basket and axo-axonic cells.

Gap junctions on GABAergic neurons containing the calcium-binding protein parvalbumin in the rat hippocampus (CA1 region).

Gap junctions were identified for the first time on chemically defined neurons in the central nervous system. Gap junctions were thus demonstrated on GABAergic neurons containing the calcium-binding on GABAergic neurons containing the calcium-binding protein parvalbumin (PV) in the rat hippocampus. Thin and semithin (0.5 micron thick) sections were cut alternately and consecutively from osmium-fixed tissue which was embedded in epoxy resin and usable for conventional electron microscopic studies. The semithin sections were processed for postembedding immunocytochemistry using an anti-PV serum. Structures corresponding to the PV-immunoreactive (PV-I) profiles on the semithin sections were easily identified on electron micrographs from the adjacent thin sections. Using this technique gap junctions were found (1) between PV-I dendrites, (2) between PV-I dendrites and PV-I somata and (3) between PV-I dendrites and small processes whose origin could not be identified. Despite a systematic search, we did not find gap junction between PV-negative processes.

Formation of new corticorubral synapses as a mechanism for classical conditioning in the cat.

An electron microscopic quantitative study of corticorubral synapses was performed in the cat which acquired classical conditioning. Conditioned stimulus was applied to the cerebral peduncle and the unconditioned stimulus was an electrical shock to the forelimb skin. The proportion of corticorubral synapses contacting with somata and proximal dendrite was increased after conditioning. It was suggested that collateral sprouting and the formation of new synapses underlie classical conditioning.

GABAergic neurons containing the Ca2+-binding protein parvalbumin in the rat hippocampus and dentate gyrus.

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.

Dendritic and somatic appendages of identified rubrospinal neurons of the cat.

Giant neurons of the red nucleus of the cat were stained intracellularly with horseradish peroxidase and examined using light microscopy, electron microscopy of thin sections, and high voltage electron microscopy of thick sections (2-5 microns). Special attention was paid to the arrangement of dendritic spines and other appendages relative to the distribution of synaptic contacts from known sources. In the region of the neuron known to receive synaptic contacts from the nucleus interpositus of the cerebellum (soma and proximal 200-300 microns of dendrites), the dendrites were relatively unbranched, and free of long spines or complex appendages. The surface of the neurons in this region was covered with a dense layer of short thin appendages that invaginated or penetrated between the synaptic terminals that cover this part of the cells. The small spines received synapses of the types associated both with the cerebellar afferent fibers and with the local inhibitory interneurons. These same terminals made synaptic contacts directly onto the surface of the neurons and onto the lateral surfaces of the spines, suggesting that the spines may serve primarily to increase the available synaptic surface area. The more distal portion of the dendritic field, where cerebellar afferents do not make synaptic contacts, exhibited a dramatically different appearance. The dendrites were much more branched, and exhibited many and varied dendritic appendages. The appendages were of three general types. One was a large protrusion with a cup-shaped head that formed the principal postsynaptic component of a glomerular arrangement also involving an axon terminal and usually a presynaptic dendrite. A second was a long thin filiform process that usually occurred around the glomeruli. This appendage was occasionally postsynaptic. The third was a spherical appendage containing many lysosomal organelles resembling residual bodies. The glomerular dendritic protrusions were very common in the distal portion of the dendritic field, numbering at least 1000 per cell. At least some of the glomeruli are specialized for receipt of synaptic input from the corticorubral pathway, since lesions of sensorimotor cortex resulted in degeneration of the central synaptic terminal in some glomeruli on horseradish peroxidase-injected rubrospinal neurons. These specializations of dendritic structure may contribute to the differences in excitatory postsynaptic potential wave shape between cortical and cerebellar inputs, and they may play a role in the changes in the cortical excitatory postsynaptic potential that develop after lesions of cerebellar inputs.(ABSTRACT TRUNCATED AT 250 WORDS)

Fast spiking cells in rat hippocampus (CA1 region) contain the calcium-binding protein parvalbumin.

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.

Sprouting of GABAergic synapses in the red nucleus after lesions of the nucleus interpositus in the cat.

An immunocytochemical study using anti-GAD serum was performed to examine the plastic changes of GABAergic inhibitory synapses in the red nucleus (RN) after lesions of the nucleus interpositus (IP) of the cat. Light-microscopic analyses revealed that 20-175 d after the unilateral lesion of the IP, somatic profiles of large neurons in the magnocellular RN contralateral to the lesion were more densely covered with GAD-immunoreactive puncta than those in the ipsilateral RN. Electron-microscopic analyses demonstrated that the GAD-immunoreactive puncta observed with the light microscope were synaptic terminals and that the number of GAD-immunoreactive synaptic terminals per unit length of somatic membrane of RN neurons was increased on the deafferented side. The GAD-immunoreactive terminals on somata of RN neurons made symmetric synaptic contacts with somatic membranes on both the deafferented and control sides. The number of immunoreactive synapses on somata of RN neurons was markedly increased on the deafferented side following IP lesion, whereas that of the unlabeled asymmetric synapses was decreased. These observations indicate that new GABAergic synapses were formed on somata of RN neurons after deafferentation from the IP.

Reorganization of corticorubral synapses following cross-innervation of flexor and extensor nerves of adult cat: a quantitative electron microscopic study.

A quantitative electron microscopic study of corticorubral synapses was performed in the red nucleus (RN) of adult cats to determine the morphological correlates for the changes in time course of corticorubral excitatory post-synaptic potentials, which occur following cross-innervation of forelimb extensor and flexor nerves. Corticorubral synaptic endings were identified by anterograde degeneration after lesions of the ipsilateral sensorimotor cortex. Rubrospinal neurons innervating upper spinal segments were electrophysiologically identified and filled with horseradish peroxidase (HRP). These cells were mainly situated in the dorsomedial part of RN. Electron micrographs of the degenerating corticorubral synaptic endings were taken in the region surrounding HRP-filled neurons and the diameter of the dendrites contacted by such terminals was measured. In the cross-innervated animals many degenerating terminals were found to synapse on dendrites with large diameter and the somata of neurons in RN. This is in contrast to the previous observations in normal cats, in which very few corticorubral synapses were found to synapse on proximal dendrites and somata of RN neurons. The diameter of HRP-filled neurons in cats which were cross-innervated was slightly smaller than those observed in normal animals. These results indicate that new corticorubral synapses were formed on proximal dendrites and somata of RN neurons as a consequence of cross-innervation.

GABAergic intrinsic interneurons in the red nucleus of the cat demonstrated with combined immunocytochemistry and anterograde degeneration methods.

The presence of glutamic acid decarboxylase (GAD), the enzyme synthesizing gamma-aminobutyric acid (GABA), was investigated in the red nucleus by an immunocytochemical method. The ipsilateral sensorimotor cortex was ablated prior to the immunocytochemical procedures to examine whether cortical neurons make synaptic contacts with GAD-immunoreactive neurons. Small GAD-immunoreactive neurons with a major diameter of 16.1 +/- 3.2 micron (mean +/- S.D.) were observed in the red nucleus under both light and electron microscopy. They were uniformly distributed throughout the nucleus. Degenerating axon terminals were found making synaptic contact with GAD-immunoreactive neurons in the red nucleus, which suggests that there is an input from the ipsilateral sensorimotor cortex to these neurons. This observation, along with our previous findings that GABAergic axon terminals make synaptic contact with the rubrospinal neurons, provides anatomical evidence for the presence of intrinsic GABAergic interneurons which mediate cortical inhibition in cat rubrospinal neurons.

Immunocytochemical demonstration of GABAergic synapses on identified rubrospinal neurons.

GABAergic synapses on rubrospinal neurons were demonstrated with immunocytochemistry combined with intracellular injection of horseradish peroxidase. Sections containing red nucleus neurons were processed for glutamic acid decarboxylase (GAD) immunohistochemistry. GAD-immunoreactive synaptic endings formed synaptic contacts with somata and dendrites of red nucleus neurons and identified rubrospinal neurons. Our observation provides further evidence that GABA acts as an inhibitory transmitter mediating cortically evoked inhibitory postsynaptic potentials in red nucleus neurons.

A quantitative study of synaptic reorganization in red nucleus neurons after lesion of the nucleus interpositus of the cat: an electron microscopic study involving intracellular injection of horseradish peroxidase.

A quantitative electron microscopic analysis of the corticorubral projection was performed in the red nucleus (RN) of adult cats to determine morphological correlates of synaptic reorganization that occur following a lesion of the interpositus nucleus (IP). Corticorubral synaptic endings were identified by lesioning the sensorimotor cortex 2-6 days before electrophysiological experiments. Horseradish peroxidase (HRP) was injected into electrophysiologically identified RN neurons. Sagittal sections 100 micrometers thick were cut and reacted by diaminobenzidine. Sections containing HRP-positive neurons were selected and embedded in Epon. In normal cats, degenerating corticorubral terminals in the RN region frequently made contact with dendritic profiles, having small cross-sections, while a few made contact with somatic profiles. Similar results were obtained when degenerating terminals making contact with HRP-filled dendrites were analyzed. In the experimental animals, the cortical lesion was performed more than 8 weeks after lesion of the IP. In these animals, degenerating corticorubral terminals were frequently found on proximal dendrites and somata in RN region and HRP-positive neurons in contrast to the findings in normal cats. The results indicate that new corticorubral synapses were formed on proximal dendrites and somata of RN neurons as a consequence of IP lesions.