Morphology and synaptic connections of myelinated primary axons in the ventrolateral region of rat trigeminal nucleus oralis

Neurons in the ventrolateral (VL) subdivision of rat trigeminal nucleus oralis (Vo) have most of their dendritic arbors confined within this region. This study examines the morphology and synaptic connections of a population of myelinated primary trigeminal axons that arborize within VL and are in a position to provide input directly to VL neurons. Primary axons were visualized for light and electron microscopic analysis by injecting 30% horseradish peroxidase (HRP) in 2% dimethylsulfoxide (DMSO) into the sensory root of the trigeminal nerve and allowing 24-36 hours for the anterograde transport of HRP into the terminal axonal arbors. This population is characterized by its cone-shaped terminal arbors, which generate many axonal endings (2-8 micron in diameter) along unmyelinated terminal strands. These arbors arise from collaterals emanating from thinly myelinated (2-5 micron in diameter) parent branches descending in the spinal V tract, which, on the basis of their size, are considered to be small myelinated (A sigma) primary trigeminal axons. HRP-labeled P endings belonging to this population of primary axons are scalloped, filled with spherical to ovoid (40-70 nm in diameter) synaptic vesicles, and lie centrally in glomeruli where they make asymmetrical axodendritic synapses on dendritic shafts and spine heads. It is at these synapses that this population of primary trigeminal axons is probably transferring its input directly to the dendritic arbors of VL neurons. The dendritic shafts and spine heads also receive symmetrical to intermediate axodendritic synapses from endings containing flattened (70 X 29 nm) synaptic vesicles. These terminals also establish axo-axonic synapses on the P ending. Other synaptic components found less often in the glomeruli include small terminals containing oval (14-23 nm) synaptic vesicles that establish symmetrical to intermediate synapses on the P ending, boutons containing pleomorphic (35-80 nm) synaptic vesicles that form symmetrical to intermediate synapses on the P ending as well as on dendritic shafts, and small peripheral endings containing round (20-40 nm) synaptic vesicles that establish asymmetrical synapses on dendritic shafts.     

Adaptations of synaptic form in an aberrant projection to the avian cochlear nucleus

Surgical removal of the otocyst in chick embryos induces axons from the contralateral cochlear nucleus (nucleus magnocellularis, NM) to form, in addition to their normal endings in nucleus laminaris (NL), anomalous and persistent functional contacts in the ipsilateral NM (Jackson and Parks, 1988). We have examined how interaction between the abnormal synaptic partners during development influences the form of the axon terminal and its relation to the target neuron. In the light microscope, aberrant axon terminals labeled in vitro with HRP appear to form boutons quite unlike the large calycine endbulbs made by the normal cochlear nerve (CN) endings in NM. In the electron microscope, however, the anomalous endings appear embedded in the NM cells, something never seen normally in NM or NL. Morphometric analyses were performed on electron micrographs from NM and NL in animals aged embryonic day (E) 19 to posthatching day (P) 2 from which the right otocyst had been removed on E3 and in normal control animals. Aberrant endings appose 18% of the circumference of operated NM cells, versus 45% for CN axons in the normal NM at this age. The mean length of membrane apposition for the anomalous NM-to-NM endings was 215% greater than for normal NM-to-NL endings but 54% smaller than that in normal CN endings. These results support the idea that developmental interactions between synaptic partners can influence the form of the contact between the 2 neurons. The results also demonstrate, however, that formation of persistent and functional synapses with NM neurons throughout development is not sufficient to induce any axon to assume the calycine form of a cochlear nerve endbulb.(ABSTRACT TRUNCATED AT 250 WORDS)     

Interaural timing difference circuits in the auditory brainstem of the emu (Dromaius novaehollandiae)

In the auditory system, precise encoding of temporal information is critical for sound localization, a task with direct behavioral relevance. Interaural timing differences (ITDs) are computed using axonal delay lines and cellular coincidence detectors in nucleus laminaris (NL). We present morphological and physiological data on the timing circuits in the emu, Dromaius novaehollandiae, and compare these results with those from the barn owl (Tyto alba) and the domestic chick (Gallus gallus). Emu NL was composed of a compact monolayer of bitufted neurons whose two thick primary dendrites were oriented dorsoventrally. They showed a gradient in dendritic length along the presumed tonotopic axis. The NL and nucleus magnocellularis (NM) neurons were strongly immunoreactive for parvalbumin, a calcium-binding protein. Antibodies against synaptic vesicle protein 2 and glutamic acid decarboxlyase revealed that excitatory synapses terminated heavily on the dendritic tufts, while inhibitory terminals were distributed more uniformly. Physiological recordings from brainstem slices demonstrated contralateral delay lines from NM to NL. During whole-cell patch-clamp recordings, NM and NL neurons fired single spikes and were doubly rectifying. NL and NM neurons had input resistances of 30.0 +/- 19.9 Momega and 49.0 +/- 25.6 Momega, respectively, and membrane time constants of 12.8 +/- 3.8 ms and 3.9 +/- 0.2 ms. These results provide further support for the Jeffress model for sound localization in birds. The emu timing circuits showed the ancestral (plesiomorphic) pattern in their anatomy and physiology, while differences in dendritic structure compared to chick and owl may indicate specialization for encoding ITDs at low best frequencies.     

Immunoelectron microscopic study of glutamate inputs from the retrosplenial granular cortex to identified thalamocortical projection neurons in the anterior thalamus of the rat

We have carried out an ultrastructural study to determine the characteristics and distribution of glutamate-containing constituents of the anterodorsal (AD) and anteroventral (AV) thalamic nuclei in adult rats. We used a polyclonal antibody to glutamate and a postembedding immunogold detection method in animals in which the neurons of AD/AV projecting to the cortex had been retrogradely labelled and the terminals of corticothalamic afferents anterogradely labelled by injection of cholera toxin-horseradish peroxidase (HRP) into the retrosplenial granular cortex. The heaviest immunogold labelling was over axon terminals 0.42 to 2.2 microm in diameter containing round synaptic vesicles and establishing Gray type 1 (asymmetric) synaptic contact (type 1 terminals) on HRP-labelled or non-labelled dendrites. Mean gold particle densities over such terminals were 3-4 times higher than the densities over the dendrites to which they were presynaptic and 5-6 times higher than over terminals establishing Gray type 2 (symmetric) synaptic contacts (type 2 terminals). Gold particle densities over neuronal cell bodies and dendrites and over a subpopulation of myelinated axons were intermediate between the densities over type 1 and type 2 terminals. In adjacent serial sections immunoreacted for gamma aminobutyric acid, type 2 terminals were heavily immunolabelled whereas type 1 terminals and other profiles with moderate gold particle densities after glutamate immunoreaction displayed very low labelling. A subpopulation of small type 1 axon terminals (up to 1 microm diameter) contained HRP reaction product identifying them as cortical in origin; they contacted small dendritic profiles (most <1 microm diameter) many of which also contained HRP reaction product. We conclude that terminals of the corticothalamic projection from retrosplenial granular cortex to AD/AV are glutamatergic and innervate predominantly distal dendrites of thalamocortical projection neurons.     

Changes in the length and organization of nucleus laminaris dendrites after unilateral otocyst ablation in chick embryos

To evaluate the contributions of the ear and acoustic stimulation to the structural development of central auditory neurons, the right otocyst was surgically destroyed in chick embryos on the third day of incubation. This pro-cedure prevents the formation of the inner ear and acousticovestibular nerve and thus removes the dominant afferent synaptic input to nucleus magnocellularis (NM). Subsequently, nucleus laminaris (NL), which receives afferent synaptic input to its dorsal dendrites from the ipsilateral NM and afferents to its ventral dendrites from the contralateral NM, was studied on both sides of the brain in Golgi preparations. By embryonic day 17, the total lengths of the individual NL dendritic fields connected to the right NM (i.e., the manipulated dendrites) were decreased by an average of 44% as compared to those NL dendrites connected to the left NM (i.e., the unmanipulated dendrites). The mean length of the unman-ipulated dendrites in experimental animals, however, did not differ from average dendritic length in normal control embryos. The amount of dendritic length lost by a NL neuron was strongly correlated with the length of the unmanipulated dendrites on the opposite side of the same neuron and with that neuron's position within NL. The lengths of dorsal and ventral dendrites on individual neurons were at least as highly correlated in experimental as in normal control animals. Correlation of dendrite length with the position of measured neurons within NL indicated that the large rostromedial-to-caudolateral gradient of increasing den-drite length present in the normal NL is also found in the manipulated dendrites in experimental animals. Regression and correlation analyses relating the length of dendrites to their longitudinal cross-sectional area revealed that there was no difference in mean dendritic diameter between the manipulated and unmani-pulated dendrites in experimental animals. The findings of a high dorsal-ventral length correlation in experimental animals and a normal spatial gradient of dendritic length among the manipulated dendrites suggests two explanations. Either (1) acoustically evoked synaptic activity is not essential for the develop-ment of these two aspects of dendritic organization, or (2) the normal NM afferents to the unmanipulated dendrites of each NL neuron in an animal with one ablated otocyst can, under the influence of acoustically driven activity, control develop-ment of the manipulated dendrites. These alternate hypotheses can be tested experimentally.     

Presumptive GABAergic centrifugal input to the lamprey retina: a double-labeling study with axonal tracing and GABA immunocytochemistry

The distribution of GABA-like immunoreactivity (GABA-LI) was performed in the lamprey retinopetal system which was previously identified by either anterograde or retrograde axonal tracing methods. This study was carried out at the ultrastructural level for the retina and under both the light and electron microscope for the mesencephalic retinopetal centers (M5 and RMA). The GABA-LI was distributed in about 40% of anterogradely HRP-labeled axon terminals in the inner retina. These made synaptic contacts upon either HRP-labeled ganglion cell dendrites or mostly on GABA-LI or on immunonegative amacrine cell dendrites and somata. The other immunonegative HRP-labeled axon terminals also established synaptic contacts on amacrine cell dendrites and somata. The mesencephalic retinopetal neurons, retrogradely labeled with HRP or [³H]proline, were GABA-LI in 65% of M5 somata and only in 15% of RMA neurons. M5 and RMA retinopetal neurons and dendrites, either GABA-LI or immunonegative, were contacted: (1) asymmetrically by HRP-labeled or unlabeled axon terminals containing rounded synaptic vesicles, always immunonegative and (2) symmetrically by HRP-unlabeled axon terminals containing pleiomorphic synaptic vesicles, which were either GABA-LI or immunonegative. The role of GABA as a putative neurotransmitter in the centrifugal visual system is discussed.     

An electron microscopic study of synaptic organization in the medial superior olive of normal and experimental chinchillas

The medial superior olive (MSO) was studied in normal animals to determine the types of synaptic endings and their distribution over the surface of MSO neurons. Unilateral lesions were made in the anteroventral cochlear nucleus (AVCN) of experimental animals to determine the source of at least one synaptic type in the MSO. The surfaces of MSO neurons in normal animals were studded with three distinct types of synaptic endings distinguished mainly by the size of their synaptic vesicles. There were endings with large vesicles, 510 Å in mean diameter; endings with small vesicles, 380 Å; and endings with vesicles intermediate in size. 435 Å. The large vesicle ending typically was greater than 2 μm in maximum diameter. It appeared as the termination of a myelinated axon or as a swollen portion of a node and made multiple asymmetrical synapses. Large vesicle endings occurred exclusively on dendrites where they formed 85% of the synaptic endings. Small vesicle endings typically were less than 2 μm in diameter. They appeared as the termination of a fine unmyelinated axon and made only one symmetrical synapse. Small vesicle boutons occurred infrequently over the entire neuronal surface. Intermediate vesicle synaptic endings were similar to large vesicle endings except that they were present only on the cell body, axon hillock, and proximal portions of the dendrites where they formed most of the synapses. In AVCN lesioned animals degenerating myelinated axons and large vesicle synaptic endings were distributed to the lateral dendrites of the ipsilateral MSO and medial dendrites of the contralateral one. In addition, a few degenerating axons and large vesicle endings were found among the ipsilateral medial dendrites. The changes in the degenerating endings were characterized by an early proliferation of neurofilaments and swelling of the endings followed by collapse of the endings and increase in electron density, disappearance of filaments and synaptic vesicles, and phagocytosis of the degenerated endings by reactive glial cells. No degenerative changes were observed in the small and intermediate vesicle endings. The results of this study indicate that the more numerous large vesicle endings presynaptic to the MSO dendrites are the axon terminals of neurons in the AVCN. The persistence after lesions of the small and intermediate vesicle endings suggests that they arise from as yet unidentified sources.     

Intracellular labeling of neurons in the medial accessory olive of the cat: III. Ultrastructure of axon hillock and initial segment and their GABAergic innervation

The gamma-aminobutyric acid (GABA) synaptic input of identified axons in the cat inferior olive was studied by use of combination of intracellular labeling with horseradish peroxidase and postembedding gold-immunocytochemistry. With this technique olivary cells were physiologically identified and light microscopically reconstructed, and the horseradish peroxidase reaction product and the immunogold labeling were subsequently simultaneously visualized for electron microscopic investigation with the use of serial ultrathin sections. The axons of cell type I (characterized by dendrites which radiate away from the cell body) originated from the soma, whereas those of type II neurons (characterized by dendritic trees which curve back towards the soma) were derived from a primary dendrite. The axons of olivary neurons stand out by the length of their axon hillock (up to 21 microns) and initial segment (up to 40 microns). The hillock forms various spiny appendages which were located within glomeruli together with dendritic spines of other olivary neurons. Axonal spines of type II neurons were more numerous and complex looking than those of type I. The axonal spines, the shaft of the axon hillock, and the transition between the hillock and initial segment were primarily innervated by GABAergic terminals (65%) but non-GABAergic terminals (35%) were present as well. The terminals apposed to the axons of type I neurons contacted mainly the axonal shafts, whereas most of the terminals adjacent to the axons of type II neurons established synaptic contacts with the axonal spines. The initial segments were largely devoid of synaptic input. Distally, the initial segment acquired a myelin sheath.     

Ultrastructural characteristics of callosal neurons in deep layers of the primary auditory cortex (AI) in cat]

An electron microscope study of retrogradely labelled nonpyramidal neurons has been carried out in layers V-VI of the primary auditory cortex (AI) after HRP injections into the contralateral AI of cats. From 2 to 9 synapses were usually revealed on somatic profiles of these callosal neurons. Synapses occupied 15.8 +/- 1.7% (on the average) of the somatic surface of these neurons. All of the revealed synapses on the somata of these callosal neurons had symmetric contacts and were formed by axon terminals with small elongated synaptic vesicles. An average length of these synaptic contacts in sections was 1.6 +/- 0.1 mm. HRP-labelled axon terminals of callosal fibres in layers V-VI contained round synaptic vesicles and formed asymmetric synapses on spines and dendrites. Possible functional significance of axo-somatic synapses in formation of impulsation patterns of the callosal neurons is discussed.     

Spiny neurons lacking choline acetyltransferase immunoreactivity are major targets of cholinergic and catecholaminergic terminals in rat striatum

The ultrastructural substrate for functional interactions between intrinsic cholinergic neurons and catecholaminergic afferents to the caudate-putamen nucleus and nucleus accumbens septi (NAS) was investigated immunocytochemically. Single sections of glutaraldehyde-fixed rat brain were processed 1) for the immunoperoxidase labeling of a rat monoclonal antibody against the acetylcholine-synthesizing enzyme choline acetyltransferase (CAT) and 2) for the immunoautoradiographic localization of a rabbit polyclonal antiserum against the catecholamine-synthesizing enzyme tyrosine hydroxylase (TH). The ultrastructural morphology and cellular associations did not significantly differ in the caudate-putamen versus NAS. Immunoperoxidase reaction for CAT versus NAS. Immunoperoxidase reaction for CAT was seen in perikarya, dendrites, and terminals, whereas immunoautoradiography for TH was in terminals. The perikarya and dendrites immunolabeled for CAT were large, sparsely spiny, and postsynaptic mainly to unlabeled axon terminals. Only 2-3% of the CAT-labeled terminals (n = 136) and less than 1% of the TH-labeled terminals (n = 86) were apposed to, or formed synapses with, perikarya or dendrites immunoreactive for CAT. Most unlabeled and all labeled terminals formed symmetric synapses. In the same sample, 18% of the CAT and 16% of the TH-labeled terminals were directly apposed to each other. Unlabeled dendritic shafts received the major (40% for CAT versus 23% for TH) synaptic input from cholinergic terminals, while unlabeled spines received the major (47% for TH versus 23% for CAT) synaptic input from catecholaminergic terminals. Neither the unlabeled dendrites or spines received detectable convergent input from CAT and TH-labeled terminals. Thirteen percent of the CAT-labeled and 14% of TH-labeled terminals were in apposition to unlabeled terminals forming asymmetric, presumably excitatory, synapses with unlabeled dendritic spines. We conclude that in both the caudate-putamen and NAS cholinergic and catecholaminergic terminals 1) form symmetric, most likely inhibitory, synapses primarily with non-cholinergic neurons, 2) differentially synapse on shafts or spines of separate dendrites, and 3) have axonal appositions suggesting the possibility of presynaptic physiological interactions. These results support the hypothesis that the cholinergic-dopaminergic balance in striatal function may be mediated through inhibition of separate sets of spiny projection neurons with opposing excitatory and inhibitory functions.     

Ultrastructural analysis of the innervation of TRH-immunoreactive neuronal elements located in the periventricular subdivision of the paraventricular nucleus of the rat hypothalamus

A combination of electron microscopic immunocytochemistry and autoradiography was employed to examine the synaptic organization of thyrotropin-releasing hormone (TRH) neurons in the periventricular subdivision of the paraventricular nucleus of the rat hypothalamus. TRH neurons were identified by immunocytochemistry. Selective uptake of tritiated serotonin (5-HT) was used to identify serotoninergic elements. TRH-immunoreactive axon terminals were found to be in synaptic contact with TRH-immunoreactive dendrites and with unlabeled dendritic branchlets. There were direct appositions between radiolabeled 5-HT terminals and TRH-immunoreactive dendrites, but differential synaptic contacts between 5-HT axonal elements and TRH neurons were not seen. TRH-immunopositive cell bodies and dendrites received a very intense innervation by unlabeled axon terminals or axonal varicosities showing morphologically defined synaptic junctions. These were mostly of the asymmetrical variety and different types could be distinguished. The findings substantiate the view that TRH neurons of the periventricular subvision of the paraventricular nucleus may be influenced by TRH axons, serotoninergic fibers and a large number of unidentified nerve terminals.     

Axonal endings terminating on dendrites of identified large trigeminospinal projection neurons in rat trigeminal nucleus oralis

The retrograde horseradish peroxidase technique was used to: (1) identify and assess the overall morphology of large neurons in the ventrolateral portion (VL) of rat trigeminal nucleus oralis projecting to cervical, thoracic and lumbosacral levels of the spinal cord; and (2) characterize the synaptic endings terminating on their dendrites. The morphology of large VL neurons projecting to all spinal levels is similar. They have 25–50 μm pyramidal-shaped somata which emit 3–6 primary dendrites. These primary dendrites give rise to spherical to elliptical-shaped dendritic arbors measuring up to 700 μm in diameter. Labeled axons enter either a deep axon bundle or the medial portion of the spinal V tract. Dendrites of labeled neurons are contacted by axonal endings of 3 types. The most numerous endings are filled with clear, spherical synaptic vesicles and usually form a single asymmetrical contacts along the entire length of dendritic shafts. Synapsing less frequently on dendritic shafts are endings containing pleomorphic synaptic vesicles and forming single symmetrical synaptic contacts. The least frequently encountered synaptic terminal contains flattened synaptic vesicles and makes a single symmetrical synaptic contact with a dendritic shaft.     

Compartment‐specific regulation of plasma membrane calcium ATPase type 2 in the chick auditory brainstem

Calcium signaling plays a role in synaptic regulation of dendritic structure, usually on the time scale of hours or days. Here we use immunocytochemistry to examine changes in expression of plasma membrane calcium ATPase type 2 (PMCA2), a high‐affinity calcium efflux protein, in the chick nucleus laminaris (NL) following manipulations of synaptic inputs. Dendrites of NL neurons segregate into dorsal and ventral domains, receiving excitatory input from the ipsilateral and contralateral ears, respectively, via nucleus magnocellularis (NM). Deprivation of the contralateral projection from NM to NL leads to rapid retraction of ventral, but not the dorsal, dendrites of NL neurons. Immunocytochemistry revealed symmetric distribution of PMCA2 in two neuropil regions of normally innervated NL. Electron microscopy confirmed that PMCA2 localizes in both NM terminals and NL dendrites. As early as 30 minutes after transection of the contralateral projection from NM to NL or unilateral cochlea removal, significant decreases in PMCA2 immunoreactivity were seen in the deprived neuropil of NL compared with the other neuropil that continued to receive normal input. The rapid decrease correlated with reductions in the immunoreactivity for microtubule‐associated protein 2, which affects cytoskeleton stabilization. These results suggest that PMCA2 is regulated independently in ventral and dorsal NL dendrites and/or their inputs from NM in a way that is correlated with presynaptic activity. This provides a potential mechanism by which deprivation can change calcium transport that, in turn, may be important for rapid, compartment‐specific dendritic remodeling. J. Comp. Neurol. 514:624–640, 2009. © 2009 Wiley‐Liss, Inc.     

Development of GABA immunoreactivity in brainstem auditory nuclei of the chick: ontogeny of gradients in terminal staining

The development of gamma-aminobutyric acid-immunoreactivity (GABA-I) in nucleus magnocellularis (NM) and nucleus laminaris (NL) of the chick was studied by using an antiserum to GABA. In posthatch chicks, GABA-I is localized to small, round punctate structures in the neuropil and surrounding nerve cell bodies. Electron microscopic immunocytochemistry demonstrates that these puncta make synaptic contact with neuronal cell bodies in NM; thus, they are believed to be axon terminals. GABAergic terminals are distributed in a gradient of increasing density from the rostromedial to the caudolateral regions of NM. The distribution of GABA-I was studied during embryonic development. At embryonic days (E) 9-11, there is little GABA-I staining in either NM or NL. Around E12-14, a few fibers are immunopositive but no gradient is seen. More GABA-I structures are present at E14-15. They are reminiscent of axons with varicosities along their length, preterminal axonal thickenings and fiber plexuses. At E15, terminals become apparent circumscribing neuronal somata and are also discernible in the neuropil of both nuclei. In E16-17 embryos, terminals are the predominantly labeled GABA-I structures and they are uniformly distributed throughout NM. The density of GABAergic terminals increases in caudolateral regions of NM such that by E17-19, there is a gradient of increasing density of GABA-I terminals from the rostromedial to caudolateral regions of NM. The steepness of this gradient increases during development and is the greatest in posthatch (P) chicks. Cell bodies labeled with the GABA antiserum are located around the borders of both NM and NL and in the neuropil between these two nuclei. Occasionally, GABA-I neurons can be found within these auditory brainstem nuclei in both embryonic and posthatch chicks. Nucleus angularis (NA) contains some GABAergic cells. The appearance of GABA-I terminals around E15 is correlated in time with the formation of end-bulbs of Held on NM neurons. Thus, the ontogeny of presumed inhibitory inputs to chick auditory brainstem nuclei temporally correlates with, and could modulate the development of, excitatory auditory afferent structure and function.     

Electron microscopic evidence that axon terminals from the mediodorsal thalamic nucleus make direct synaptic contacts with callosal cells in the prelimbic cortex of the rat

A combined study of anterograde axonal degeneration and HRP retrograde labeling has shown that there exist monosynaptic connections between afferent fibers from the mediodorsal thalamic nucleus (MD) and callosal cells in the prelimbic cortex of the rat. Degenerating axon terminals from MD made asymmetrical synaptic contacts with dendritic spines from apical dendrites of layer III pyramidal cells that were retrogradely labeled with HRP after its injection into the prelimbic cortex contralateral to MD lesions.     

Intrinsic circuitry: synapses involving the local axon collaterals of corticocortical projection neurons in the mouse primary somatosensory cortex

Pyramidal neurons in the mouse SmI cortex were labeled by the retrograde transport of horseradish peroxidase (HRP) injected into the ipsilateral MsI cortex. Terminals of the local axon collaterals of these neurons (CC terminals) were identified in SmI, and their distribution and synaptic connectivity were examined. To avoid confusion, terminals in SmI cortex labeled by the anterograde transport of HRP injected into MsI were eliminated by lesion-induced degeneration. Lesions of MsI were made 24 hours after the injection of HRP; postlesion survival time was 4 days. Most CC axon terminals occurred in layers III and V where they formed asymmetrical synapses. Of 139 CC synapses in layer III and 104 in layer V, approximately 13% were formed with dendritic shafts. Reconstruction of 19 of these dendrites from serial thin sections showed them to originate from both spiny and nonspiny neurons. Most synapses of CC terminals (about 87%) were onto dendritic spines. In contrast, White and Keller (1987) demonstrated that terminals belonging to the local axon collaterals of corticothalamic (CT) projection cells synapse mainly with dendritic shafts of nonspiny neurons: 92% onto shafts, the remainder onto spines. The distribution of asymmetical synapses onto spines and dendritic shafts was analyzed for neuropil in layers III, IV, and V. Depending on the layer, from 34 to 46% of the asymmetrical synapses in the neuropil were onto dendritic shafts. Results showing that CC and CT terminals form proportions of axodendritic vs. axospinous synapses that differ from each other, and from the neuropil, indicate that local axon collaterals are highly selective with regard to their postsynaptic elements.     

Dendroaxonic synapses in the substantia gelatinosa glomeruli of the spinal trigeminal nucleus of the cat

The glomeruli in the substantia gelatinosa layer of the spinal trigeminal nucleus of the cat contain three kinds of dendritic processes. One of these, the type 2 dendrite, contains large synaptic vesicles in its spine heads and in its shaft. The type 2 dendrite receives axodendritic synapses from primary trigeminal afferent (C) axons and an occasional axodendritic synapse from small axonal (P) endings with small synaptic vesicles. The type 2 dendrites in turn form dendroaxonic synapses on the C endings. The dendroaxonic synapse and the axodendritic synapse of the C ending typically occur in reciprocal pairs. The axodendritic synapse usually lies in the depths of scalloped depressions in the surface of the C ending while the dendroaxonic synapse is found on the rim of the depression. Type 1 spines, i.e., dendritic spines receiving axodendritic synapses from the primary ending and lacking synaptic vesicles, also receive dendrodendritic synapses from type 2 dendrites. The types 2 dendrite with its large, rounded synaptic vesicles is considered to be excitatory at its dendroaxonic and dendrodendritic synapses. The type 2 dendrites course from glomerulus to glomerulus receiving their excitatory input through the axodendritic synapses of C axons. A type 2 dendrite, in response to C axon excitation would activate type 1 spines directly through their dendrodendritic synapses (C→2→1) and indirectly by increasing transmitter release at the axodendritic synapses of the C axonal endings through their dendroaxonic synapses (2→C→1). The type 2 dendrites could serve two functions. First, they may prolong transmitter release from the axodendritic synapses of C axonal endings beyond the time of arrival of incoming action potentials because of the reciprocal pairing of dendroaxonic and axodendritic synapses (C?2). Second, they may extend the spatial range of the excitatory output of active primary afferent axons to type 1 spines of glomeruli whose primary afferent axons may be inactive (C→2→1).     

Synapses made by axons of callosal projection neurons in mouse somatosensory cortex: emphasis on intrinsic connections

This is one of a series of papers aimed at identifying the synaptic output patterns of the local and distant projections of subgroups of pyramidal neurons. The subgroups are defined by the target site to which their main axon projects. Pyramidal neurons in areas 1 and 40 of mouse cerebral cortex were labeled by the retrograde transport of horseradish peroxidase (HRP) transported from severed callosal axons in the contralateral hemisphere. Terminals of the local axon collaterals of these neurons ("intrinsic" terminals) were identified in somatosensory areas 1 and 40, and their distribution and synaptic connectivity were examined. Also examined were the synaptic connections of "extrinsic" callosal axon terminals labeled by lesion induced degeneration consequent to the severing of callosal fibers. A post-lesion survival time of 3 days was chosen because by this time the extrinsic terminals were all degenerating, whereas the intrinsic terminals were labeled by HRP. Both intrinsic and extrinsic callosal axon terminals occurred in all layers of the cortex where they formed only asymmetrical synapses. Layers II and III contained the highest concentrations of both types of callosal axon terminal. Analyses of serial thin sections through layers II and III in both areas 1 and 40 yielded similar results: 97% of the extrinsic (277 total sample) and of the intrinsic (1215 total sample) callosal axon terminals synapsed onto dendritic spines, likely those of pyramidal neurons; the remainder synapsed onto dendritic shafts of both spiny and nonspiny neurons. Thus the synaptic output patterns of intrinsic vs. extrinsic callosal axon terminals are strikingly similar. Moreover, the high proportion of axospinous synapses formed by both types of terminal contrasts with the proportion of asymmetrical, axospinous synapses that occur in the surrounding neuropil where only about 80% of the asymmetrical synapses are onto spines. This result is in accord with previous quantitative studies of the synaptic connectivities of both extrinsic and intrinsic axonal pathways in the cortex (White and Keller, 1989: Cortical Circuits; Boston: Birkhauser): in all instances, axonal pathways are highly selective for the types of elements with which they synapse.     

Pre- and postsynaptic sites for serotonin modulation of GABA-containing neurons in the shell region of the rat nucleus accumbens

The shell of the nucleus accumbens receives a dense serotonergic innervation and contains abundant gamma-aminobutyric acid (GABA)-immunoreactive neurons. Moreover, serotonin (5-hydroxytryptamine: 5-HT) and GABA have been implicated in a variety of common motivational and motor-related functions partially ascribed to this brain area. We used immunoelectron microscopy of antisera directed against 5-HT and GABA in the same section of tissue to examine whether there were cellular substrates that might indicate more specific sites for functional interactions involving these transmitters in the shell region of the rat nucleus accumbens. Immunogold-silver labeling for GABA was localized to perikarya, dendrites, axons and axon terminals, whereas immunoperoxidase labeling for 5-HT was restricted to axons and axon terminals. Approximately half (187/366) of the 5-HT-immunoreactive axon terminals apposed or formed synaptic junctions with postsynaptic neurons. These junctions were mainly of the symmetric-type (83/187) characteristic of inhibitory transmitters, and were equally prevalent on dendrites with and without detectable gold-silver labeling for GABA. Of the 187 5-HT-labeled axon terminals with recognized synaptic contacts, 36% also showed convergence on a common dendrite with a GABA-labeled axon terminal. In addition, 5-HT- and GABA-immunoreactive axon terminals were commonly (83/366) identified in direct apposition to one another. Within a single plane of section, 41% of the apposed GABA-immunoreactive axon terminals formed symmetric-type junctions with dendrites or somata, whereas, the apposed 5-HT-labeled axon terminals rarely showed postsynaptic contacts. These results indicate that 5-HT-containing axon terminals may postsynaptically inhibit GABAergic neurons and their targets within the shell of the rat nucleus accumbens. Additionally, our results strongly suggest that, in this brain region, appositions between 5-HT and GABA axons and axon terminals may facilitate presynaptic interactions between these transmitter systems. © 1996 Wiley-Liss, Inc.     

Ultrastructural analyses of afferent terminals in the subthalamic nucleus of the cat with a combined degeneration and horseradish peroxidase tracing method

The synaptic organization of the feline subthalamic nucleus (STN) was studied electron microscopically. Following horseradish peroxidase (HRP) injections into the globus pallidus (GP) and electrolytic lesions of the nucleus tegmenti pedunculopontinus pars compacta (TCP) in the same cat, both degenerating and HRP-labeled terminals were found in the STN with abundant retrogradely HRP-labeled neurons. Degenerating terminals of TPC origin were medium-sized and characterized by asymmetric synaptic contacts. They synapsed widely on the STN neuronal surface, including the somata, dendrites of varying dimensions, dendritic spines and vesicle-containing processes. They formed 25.1%, 65.1%, 4.7%, and 4.7%, respectively, of all TPC efferent terminals. Some of the postsynaptic components were labeled with HRP. Occasionally both degenerating terminals and HRP-labeled terminals were in synaptic contact with the same HRP-labeled neuron: therefore, afferents of TPC and GP converge on the same STN projection neuron. In order to discover the origin of these HRP-labeled terminals, a mixed solution containing HRP and kainic acid was injected into the GP. Numerous degenerating terminals were observed to synapse with HRP-labeled STN neurons, but no HRP-labeled terminal was observed. These degenerating terminals were similar in appearance to the above-mentioned HRP-labeled terminals. They were characterized by their relatively large size, predominantly symmetric synapses, and preferential distribution on the somata and large or medium-sized dendrites. They formed 39.6%, 20.1%, and 31.1%, respectively, of all GP efferent terminals. Therefore, it became clear that both the HRP-labeled terminals of the first experiment and the degenerating terminals of the second experiment originated from the GP. Following surgical ablations of the primary sensorimotor cortex (Cx), some axon terminals in the STN showed degeneration. These degenerating terminals were small and formed asymmetric synapses mainly with dendritic spines, small dendrites and vesicle-containing processes. They formed 48.0%, 28.0%, and 12.0%, respectively, of all Cx efferent terminals. These electron microscopic investigations reveal the convergence of TPC and GP afferents and that STN projection neurons relay the TPC and pallidal inputs directly to the GP.