Search

The genesis of the cuneothalamic neurons (CTNs) in the rat cuneate nucleus was determined by a double-labeling method using 5'-bromodeoxyuridine (BrdU), the thymidine analogue, and Fluoro-Gold (FG), a retrograde fluorescent tracer. BrdU-positive cells were observed in the cuneate nucleus in all rats receiving BrdU injection at embryonic days (E) E13--E16; none was detected in rats given BrdU injection at E12. At E13 and E14, BrdU-positive cells were randomly distributed. However, at E15, the number of BrdU-positive cells was clearly reduced and the majority of them was located at the dorsolateral or peripheral region of the nucleus. FG/BrdU double-labeling study showed the existence of BrdU-labeled CTNs when the mother rat received BrdU injection at E13 and E14, being more numerous at E13 in which the neurons were scattered throughout the nucleus. At E14, however, the majority of the BrdU-labeled CTNs were located superficially in the nucleus. Double-labeled cells were undetected in rats that had been exposed to BrdU at E15 and E16. Quantitative data showed that the majority (ca 70-80%) of the CTNs were generated at E13, and were markedly decreased at E14 (ca 4-6%). Using nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d) histochemistry coupled with BrdU immunohistochemistry, we have shown the NADPH-d/BrdU double-labeled neurons in the nucleus between E13 and E15, with the majority of them occurring at E14, but absent at E16. The present results suggest that the CTNs are generated prior to the NO-containing neurons in the cuneate nucleus.

Nicotinamide adenine dinucleotide phosphate-diaphorase (NADPH-d) histochemistry and nitric oxide synthase (NOS) immunohistochemistry have been used to characterize the nitric oxide (NO)-containing neurons in the rat cuneate nucleus. The present results showed that NADPH-d-positive/NOS-immunoreactive (-IR) neurons were distributed in the entire rostrocaudal extent of the nucleus. In the caudal region (approximately 1-2 mm caudal to the obex), NADPH-d/NOS-IR neurons were aggregated along the dorsal area of the nucleus notably in the lateral aspect. When traced rostrally, labeled neurons were progressively reduced and the cells were randomly distributed. The labeled neurons varied from round, ovoid to spindle-shaped with a mean profile area of about 140.1+/-1.7 microm(2) (n=720). They made up 7-10% of the neuronal population in the cuneate nucleus. By immunoelectron microscopy, the immunoreaction product was deposited throughout the cytoplasm extending from the soma to the proximal and distal dendrites. Results of NADPH-d staining paralleled that of NOS immunohistochemistry. Furthermore, NADPH-d reactivity and NOS-IR were colocalized in the same neurons following double labeling. Using NADPH-d histochemistry along with anti-gamma-aminobutyric acid (GABA) and -glycine postembedding immunolabeling for identification of GABA- and glycine-IR neurons, respectively, about 33% of the NADPH-d-positive neurons contained both GABA and glycine, 26% of them contained only glycine, while 41% of them showed neither GABA nor glycine labeling. Cuneothalamic neurons (CTNs) were identified by injecting the retrograde tracer Fluorogold (FG) into the ventrobasal complex of the thalamus. Numerous FG-labeled neurons were present in the contralateral cuneate nucleus, but none were reactive for NADPH-d. The present results suggest that approximately 60% of the NADPH-d/NOS-IR neurons in the cuneate nucleus are interneurons containing GABA and/or glycine.

This study was aimed to clarify whether the primary afferent terminals (PATs), GABAergic terminals, and glutamatergic terminals made direct synaptic contacts with glycine-IR neurons in the cuneate nucleus of rats. In this connection, injection of the anterograde tracer WGA-HRP into brachial plexus, antiglycine preembedding immunoperoxidase, and anti-GABA, along with antiglutamate postembedding immunogold labeling, were used to identify the PATs, glycine-IR neurons, GABA-IR terminals, and glutamate-IR terminals, respectively. The present results showed that HRP-labeled PATs, immunoperoxidase-labeled glycine-IR terminals, immunogold-labeled GABA-IR, and glutamate-IR terminals made axodendritic synaptic contacts with immunoperoxidase-labeled glycine-IR neurons. The latter three presynaptic elements also formed axosomatic synapses with glycine-IR neurons. Statistical analysis has shown that the minimum diameter of the glycine-IR dendrites postsynaptic to the above-mentioned four presynaptic elements did not differ significantly. In addition, the synaptic ratio of the glutamate-IR terminals on the glycine-IR dendrites was higher than that of GABA-IR terminals. The synaptic ratio of the GABA-IR terminals on glycine-IR dendrite was in turn higher than that of the PATs and glycine-IR terminals. It is suggested that the PATs and glutamate-IR terminals on the glycine-IR neurons may be involved in subsequent postsynaptic inhibition for spatial precision of lateral inhibition. On the other hand, the GABA-IR and glycine-IR terminals which make synaptic contacts with the dendrites of glycine-IR neurons may provide a putative means for disinhibition or facilitation to maintain the baseline neuronal activity in the rat cuneate nucleus. The results of quantitative analysis suggest that glutamate act as the primary excitatory neurotransmitter, while GABA, when compared with glycine, may serve as a more powerful inhibitory neurotransmitter on glycine-IR neurons in the rat cuneate nucleus.

This study was aimed to investigate the possible involvement of neurons in the cuneate nucleus (CN) in the processing of A beta afferent inputs evoked by electrical stimulation of constricted median nerve in rats with behavioral signs of neuropathic pain. Immunohistochemical localization of Fos protein was used to examine the neuronal activation, and the combination of Fos immunohistochemistry with the retrograde labeling of Fluoro-Gold (FG) injected into the ventrobasal complex of the thalamus was used to characterize the activated neurons. Two weeks after unilateral median nerve constriction injury, the rats exhibited behavioral signs of neuropathic pain in the affected forepaws. In rats after nerve injury but without electrical stimulation, some Fos-like immunoreactive (Fos-LI) neurons were detected in the dorsal horn of the seventh cervical segment (C7) but none was found in the CN. Similar features were also noted when the stimulation of the intact median nerve served as an additional control. After A beta-fiber intensity stimulation of the previously constricted median nerve, an increase in number of Fos-LI neurons occurred in the medial half of the ipsilateral C7 dorsal horn as well as in the ipsilateral CN. In the latter, the Fos-LI neurons were located in the median nerve projection territory throughout the nucleus. Most of the Fos-LI neurons were distributed in the middle region of the CN, with about 78% of them emitting FG fluorescence indicating that they were cuneothalamic projection neurons. The results of this study suggest that the dorsal column-medial lemniscal system may contribute to the transmission and modulation of A beta-fiber mediated neuropathic pain signals.

This study was aimed to localize and characterize the somatostatin-immunoreactive (SOM-IR) neurons in the rat cuneate nucleus (CN). By immuno-histochemistry, the SOM-IR neurons, which were widely distributed in the nucleus, were round, spindle or multiangular in shape (mean area = 226.1 +/ -3.1 microm(2), n = 1016). By electron microscopy, the neurons shared all the ultrastructural features of the cuneothalamic neurons (CTNs) which showed a slightly indented nucleus and a fairly rich cytoplasm containing well-developed Golgi apparatuses and rough endoplasmic reticulum (rER). The SOM immunoreaction product filled the cytoplasm of the neurons extending from the soma to the proximal and distal dendrites, which were postsynaptic to unlabeled boutons. In addition to soma and dendrites, SOM-IR boutons were also identified which made axodendritic synaptic contacts with SOM-IR dendrites. The SOM-IR neurons were characterized by using anti-SOM pre-embedding immunolabeling coupled with horseradish peroxidase (HRP) retrograde method, or SOM immunolabeling along with anti-glutamate, gamma-aminobutyric acid (GABA) or glycine post-embedding immunolabeling for identification of CTNs, glutamate-IR, GABA-IR and glycine-IR neurons, respectively. It was shown that more then 80% of the CTNs contained SOM and, furthermore, they contained glutamate but not GABA or glycine. On the basis of present findings, it is suggested the majority of the SOM-IR neurons in the rat CN are CTNs and that they may be involved in modulation of somatosensory synaptic transmission.

This study was aimed to clarify whether the cuneothalamic relay neurons (CTNs) in the rat cuneate nucleus contained glycine or whether the neurons were modulated directly by presynaptic glycine-IR terminals. For this purpose, retrograde transport of wheat germ agglutinin conjugated with horseradish peroxidase (WGA-HRP) and immunoperoxidase labelling for glycine have been used to ascertain if the CTNs in the rat are glycine-immunoreactive (glycine-IR). Our results have shown that the WGA-HRP-labelled CTNs (mean area = 318 +/- 6.5 microm(2)) were not reactive for glycine. Glycine immunoreactivity, however, was localized in some small-sized neurons (mean area = 210 +/- 6.2 microm(2)) and axon terminals associated with the CTNs. The synaptic organization between the glycine-IR terminals and CTNs was further analyzed using anti-glycine postembedding immunogold labelling. By electron microscopy, the immunogold-labelled glycine-IR terminals containing pleomorphic synaptic vesicles formed symmetrical synaptic contacts with the dendrites, dendritic spines, and somata of CTNs. Quantitative estimation showed that the mean ratios of glycine-IR terminals to total terminals associated with the soma, proximal dendrites and distal dendrites of the CTN were 49.5, 45.2, and 45.8%, respectively. The higher incidence of glycine-IR terminals on the soma, however, was not significantly different from that of the proximal and distal dendrites. Notwithstanding the above, this study has shown a large number of glycine-IR terminals making direct synaptic contacts with CTNs, suggesting that glycine is one of the important neurotransmitters involved in postsynaptic inhibition on the cuneothalamic relay neurons to modulate incoming somatosensory information from forelimb areas in the rat., (Copyright 2000 Wiley-Liss, Inc.)

The present study describes the ultrastructural synaptic relationships between corticocuneate terminals (CCTs) and glycine-immunoreactive (glycine-IR) neurons in the cuneate nucleus of rats using anterograde tract-tracing of wheatgerm agglutinin conjugated with horseradish peroxidase (WGA-HRP) and anti-glycine immunoperoxidase labeling methods. The HRP-labeled CCTs made axodendritic synapses preferentially in the ventral part of the cuneate nucleus near the obex. In a total of 182 CCTs surveyed, 14 of them made direct synaptic contacts with immunoperoxidase-labeled glycine-IR dendrites. The present results suggest that cortical modulation on the sensory transmission of cuneate nucleus may be mediated through glycine-IR neurons.

The distribution of glycine-immunoreactive (glycine-IR) neurons and their associated axon terminals in the rat cuneate nucleus was studied using antiglycine postembedding immunoperoxidase labelling and immunogold staining, respectively. The immunoperoxidase-labelled glycine-IR neurons were widely distributed in the entire rostrocaudal extent of the nucleus. They made up 30.8% (9671/31368) of the neurons surveyed. Quantitative evaluation showed that the percentage of glycine-IR neurons in the caudal level was significantly higher than that in the middle and rostral levels. The glycine-IR neurons were small cells (mean area = 198+/-1.9 microm2, n = 2862) with ovoid or spindle-shaped somata. Statistical analysis showed that the size of the glycine-IR neurons in the rostral level was significantly smaller than that in the middle and caudal levels. Immunogold labelled glycine-IR terminals which contained predominantly pleomorphic synaptic vesicles were mostly small (mean area = 1.24+/-0.03 microm2, n = 286) and they constituted 24.7% (286/1158) of the total terminals surveyed. They formed axodendritic, axosomatic and axoaxonic synapses with unlabelled elements. It is suggested from this study that glycine is one of the major neurotransmitters involved in the depression of synaptic transmission in the cuneate nucleus.

This study reports the reactivities of acetylcholinesterase (AChE) and choline acetyltransferase (ChAT) in some of the nonneuronal elements in the external cuneate nucleus (ECN) of gerbils. AChE reaction products were localized in some oligodendrocytes in their cisternae of rough endoplasmic reticulum, nuclear envelope and Golgi saccules. The basal lamina lining the capillary endothelia also displayed AChE reactivity. In ChAT immunocytochemistry, the reaction products were found to be associated with the vascular basal lamina as well as the endothelial plasma membrane facing the lumen. The most remarkable finding was the localization of ChAT immunoreactivity in some oligodendrocytes and occasional glioblasts (small glial precursor cells containing a thin rim of cytoplasm surrounding an irregular nucleus with homogeneous chromatin materials). The ChAT-positive oligodendrocytes consisted of two types, medium-dense and dark cells, either associated with blood vessels or ChAT-stained neuronal elements. It is suggested from these new findings that at least some of the oligodendrocytes and glioblasts in the ECN of gerbils may be involved in the synthesis, storage, release and degradation of acetylcholine.

The present study described an ultrastructural synaptic configurations between primary afferent terminals (PATs), cuneothalamic relay neurons (CTNs) and GABA-immunoreactive (GABA-IR) boutons in the cuneate nucleus of rats using cervicothoracic dorsal rhizotomies, retrograde transport of wheat germ agglutinin conjugated with horseradish peroxidase complex (WGA-HRP) and anti-GABA immunogold labelling methods. With this procedure, direct synaptic relationships between the PATs, CTNs and GABA-IR boutons have been demonstrated. The most remarkable feature was the observation whereby an immunogold-labelled GABA-IR bouton was presynaptic to a WGA-HRP labelled dendrite of CTN and a degenerating PAT; the same PAT was in turn presynaptic to the HRP-labelled dendrite. This was evident in ten out of a total of 133 synaptic configurations that were closely scrutinized. Results of this study support the concept that GABA-IR boutons are not only involved in presynaptic inhibition on the primary afferent input to the cuneothalamic relay neurons, but also exert a simultaneous postsynaptic inhibition on these cells.

Using acetylcholinesterase histochemical and choline acetyltransferase immunocytochemical localization methods, this study has provided conclusive evidence for the existence of cholinergic neurons in the external cuneate nucleus of gerbils. By light microscopy, both acetylcholinesterase and choline acetyltransferase labelling was confined to the rostral portion of the external cuneate nucleus. Ultrastructurally, acetylcholinesterase reaction products were found in the nuclear envelope, cisternae of rough endoplasmic reticulum and Golgi saccules of some somata and large dendrites as well as in the membranes of small dendrites, myelinated axons and axon terminals. These neuronal elements were also stained for choline acetyltransferase; immunoreactivity was associated with nuclear pores, nuclear envelope, perikaryal membrane and all the membranous structures within the cytoplasm. Of the total choline acetyltransferase-labelled neuronal profiles analysed, 79% were myelinated axons, 15% dendrites, 4% somata and 2% axon terminals. The immunostained axon terminals consisted of two types containing either round (Rd type; 62.5%) or pleomorphic (Pd type; 37.5%) vesicles. Both were associated directly with choline acetyltransferase-positive dendrites. In contrast to the paucity of choline acetyltransferase-labelled axon terminals, numerous choline acetyltransferase-positive myelinated axons were present. It may thus be hypothesized that most, if not all, of the external cuneate nucleus cholinergic neurons are projection cells; such cells may give rise to axonal collaterals which synapse onto their own dendrites for possible feedback control. Choline acetyltransferase-positive dendrites were contacted by numerous unlabelled presynaptic boutons, 60% of which contained round or spherical synaptic vesicles (Rd boutons) and 40% flattened vesicles (Fd boutons), suggesting that these neurons are under strong inhibitory control. The preferential concentration of cholinergic components in the rostral external cuneate nucleus may be significant in the light of the highly organized somatotopy in the external cuneate nucleus and its extensive efferent projections to medullary autonomic-related nuclei. Our results suggest that the cholinergic neurons may be involved in somatoautonomic integration.

The synaptic organisation of the primary afferents from the brachial and cervical plexuses to the external cuneate nucleus of gerbils was compared following an intraneural injection of horseradish peroxidase into the musculocutaneous, median, ulnar and radial nerves of the brachial plexus or the main branches of the cervical plexus; 407 labelled primary afferent terminals from the brachial and 459 from the cervical plexus were studied. These boutons made synaptic contacts with 586 and 633 dendritic profiles, respectively. 99.0% of the primary afferent boutons from the brachial plexus contained clear round synaptic vesicles (R boutons); the remaining 1% of boutons contained pleomorphic synaptic vesicles (P boutons). For boutons from the cervical plexus, 95% were R boutons and 5% were P boutons. The labelled R bouton profiles had a wide range of cross-sectional area from 0.4 to 13.1 microns 2, while the P boutonal profiles were of a small variety (range, 0.4-2.3 microns 2; mean, 1.5; S.D., 0.6 micron 2). The R boutons from the brachial plexus (mean, 3.9 microns; S.D., 2.1 microns 2) were generally larger than those from the cervical plexus (mean, 3.3 microns 2; S.D., 1.9 microns 2). On close analysis, 72.4% of R boutons from the brachial plexus were found to synapse on distal dendrites, 15.9% on secondary dendrites, 9.5% on dendritic spines and 2.2% on proximal dendrites. For R boutons from the cervical plexus, 81.1% synapsed on distal dendrites, 12.1% on dendritic spines and 6.8% on secondary dendrites; none was observed on proximal dendrites. Such a different synaptic organisation between the two nerve plexuses may be related to their different perceptuomotor executions.

The present study revealed the efferent projections from the external cuneate nucleus (ECN) to various medullary nuclei in the gerbil as demonstrated in fresh living brainstem slices by using in vitro anterogradely tracing with the dextran-tetramethyl-rhodamine-biotin. The tracer-labelled ECN axon terminals were observed (1) in most of the vital autonomic-related nuclei: the nucleus solitary tractus, nucleus ambiguus, rostroventrolateral reticular nucleus and C2 adrenergic area, (2) in the reticular formation: the medullary, parvocellular, intermediate, gigantocellular, dorsal paragigantocellular and lateral paragigantocellular reticular nuclei and medullary linear nucleus, and (3) in sensory nuclei: the cuneate nucleus, spinal trigeminal nuclei caudalis and interpolaris, paratrigeminal nucleus, medial and spinal vestibular nuclei, inferior olive and prepositus hypoglossal nucleus. These new findings are discussed in relation to possible roles of the ECN in cardiovascular, respiratory and sensorimotor controls.

The present study is concerned with the connections of the external cuneate nucleus (ECN) in the gerbil following an injection of horseradish peroxidase (HRP) into the ventralis posterior pars oralis (VPLo) or adjacent nuclei of the thalamus. The number, soma size and distribution of the retrograde-labelled ECN neurons were studied and quantified. The application of two retrograde fluorescent tracers was also used to determine whether the ECN neurons would project to the thalamus as well as to the cerebellum through their collaterals. The HRP-positive ECN neurons projecting to the thalamic VPLo were confined to the contralateral caudal half of the ECN, primarily within the intermediate portion represent the forearm and arm territories with a small part of the thoracic and shoulder areas. Labelled neurons were classified into small and medium-sized cells. The majority (96%) of the external cuneothalamic neurons were of the small variety. No double-labelled cells were detected in the ECN following injections of Rhodamine-labelled latex microspheres and Fast blue into the cerebellum and thalamus respectively, suggesting that the ECN neurons projecting to the thalamus form a separate cell group different from those projecting to the cerebellum. The injected HRP into the VPLo was also transported in an anterograde direction by the thalamocortical fibers. The HRP-labelled axonal terminals were distributed within motor area 4 and the dysgranular zones (DZs) of the primary somatosensory cortex (SmI), reaching the deep layers IV and VI as well as superficial layer I. The external cuneothalamocortical pathway shown in the present study may be related to the proprioceptive feedback control of the coordinating motor activity, especially during forelimb muscle movement.

The projection from the external cuneate nucleus (ECN) to the cerebellum was studied in the gerbil following the retrograde transport of minute injections (0.05-0.1 microliter, 30% solution) or implantations of horseradish peroxidase (HRP) in various folia of the cerebellar cortex and deep nuclei. The morphology of the labelled neurons, as well as their quantitative distribution, was also examined. The projections were in general bilateral but predominantly ipsilateral. A remarkable finding, however, was the predominant contralateral projection from the medial cell column of ECN representing the paw region. Topographically, the nuclear areas receiving the thoracic, shoulder and neck muscle afferents projected bilaterally to vermis lobules I-V and VIII-IX, lobulus simplex, crura Ia and Ib, paraflocculus and flocculus, and contralaterally to crus II and nucleus interpositus; the nuclear regions receiving the forelimb muscle afferents sent fibers bilaterally to vermis lobules V-VIII, paramedian lobule and copula pyramidis, and ipsilaterally to crus II and nucleus interpositus. Based on their somal size, three classes of neurons were distinguished: small, medium and large cells. The small and medium cells constituted 79 and 20% of the population of the labelled cells respectively, whereas the large cells were only occasionally identified. In the cell columns of the caudal ECN representing the forelimb (Lan et al. 1994), most labelled neurons were medium cells which made up 65% of the total labelled medium cells. In contrast, the majority (90%) of the labelled cells in the rostral ECN were of the small-sized variety. The functional significance of the segregation, both in the projection patterns and sizes of somata of the cuneocerebellar neurons is discussed.

The projections of muscle afferents from six regions (hand, forearm, arm, thorax, shoulder and neck) to the external cuneate nucleus (ECN) of gerbils were investigated using transganglionic transport of horseradish peroxidase (HRP). These were chiefly ipsilateral projections which terminated in six discrete longitudinal columns in the ECN. A medial cell column confined to the caudal ECN received projections from the hand muscles, while the ventromedial and ventrolateral cell columns in the posterior half of the ECN received projections from the forearm and arm muscles respectively. In the caudal ECN, the thoracic and shoulder muscles were represented by the dorsomedial and dorsolateral cell columns respectively. The last mentioned two columns extended anteriorly into the rostral ECN occupying a dorsal and intermediate portion leaving the ventral portion which represented the neck muscles. The latter also extended posteriorly into the middle 1/3 of ECN occupying a lateral portion. Serial sections and reconstruction studies showed that the ECN was located in the dorsal region of the upper medulla, approximately 0.3-1.8 mm rostral to the obex. The rostrocaudal extent of the six cell columns from hand proximally to neck were measured to be 0.5-0.8, 0.3-1.1, 0.3-1.3, 0.4-1.8, 0.4-1.8 and 0.8-1.8 mm rostral to the obex.

This study investigates the synaptic relation between gamma-aminobutyric acid-immunoreactive (GABA-IR) and cuneothalamic relay neurons (CTNs) in the rat cuneate nucleus. Retrograde transport of wheat germ agglutinin conjugated with horseradish peroxidase complex (WGA-HRP) was used to label CTNs while anti-GABA immunogold serum was used for the detection of GABA-IR boutons associated with CTNs. With these procedures, immunogold-labelled GABA-IR boutons were found to form axosomatic, axodendritic and axospinous synapses with the WGA-HRP-labelled but immunonegative CTNs. Quantitative estimation showed that the mean ratios of GABA-IR to GABA-immunonegative boutons making synaptic contacts with somata, proximal dendrites, and distal dendrites were 47.9%, 49.1% and 34.7%, respectively. Statistical analysis showed that the incidence of GABA-IR boutons on the somata and proximal dendrites of CTNs was significantly higher than on the distal dendrites. Our results indicate that GABA is the primary inhibitory neurotransmitter in the cuneate nucleus, thereby emphasizing the importance of postsynaptic inhibition on cuneothalamic relay neurons.

The present study examined the synaptic relation between the primary afferent terminals and intrinsic neuronal elements in the rat cuneate nucleus. For this purpose, experimental degeneration after multiple cervicothoracic dorsal rhizotomies or anterograde transport of wheatgerm agglutinin conjugated to horseradish peroxidase were used to identify the primary afferent terminals, while immunogold postembedding staining was employed to identify the GABA-immunoreactive boutons. The combined procedure allowed us to demonstrate a direct synaptic relationship between the primary afferent terminals and GABA-immunoreactive boutons. At least two types of synaptic relation were observed between the primary afferent terminals, identified by their degenerating features or labeled by wheatgerm agglutinin conjugated to horseradish peroxidase, and the immunogold-labeled GABA-immunoreactive boutons (i) a GABA-immunoreactive bouton making a simple presynaptic contact with the primary afferent terminal; and (ii) a synaptic glomerular complex in which the centrally located primary afferent terminal was postsynaptic to a GABA-immunoreactive bouton and presynaptic to dendrites closely associated with it; both terminals were sometimes presynaptic to a common dendrite. It is speculated from this study that the incoming impulses from the forelimb area are modulated by the GABA-immunoreactive boutons in the cuneate nucleus of the rat.

Immunogold labelling was used to identify GABA-immunoreactive (GABA-IR) terminals in the cuneate nucleus of the rat. About 30% of the terminals surveyed were GABA-IR. They were mostly small, although a few of medium size were encountered. The terminals contained either polymorphic or round synaptic vesicles; these occurred in an approximately 5:1 ratio. Most of these terminals formed symmetric synapses with dendrites of various sizes. The GABA-IR terminals also made synaptic contacts with somata of cuneate neurons and axon terminals of unknown origin. In addition, a few GABA-IR terminals forming asymmetric synapses were found in the cuneate neuropil. The possible functional significance of the synaptic organisation is discussed.