Light and Electron Microscopical Studies of the GABA Innervation of the Dorsal Column Nuclei and the Lateral Cervical Nucleus in the Primate Species Macaca Fascicularis and Papio Anu bis

In 4 monkeys of the species Macaca fascicularis (2 animals) and Papio anubis ( 2 animals) the three dorsal column nuclei and the lateral cervical nucleus have been investigated immunocytochemically with antiserum against gamma-aminobutyric acid (GABA). Light microscopic studies demonstrated the presence of GABA-positive cells in the gracile nucleus, the internal cuneate nucleus and the lateral cervical nucleus but not in the external cuneate nucleus.Although labeled cells seemed fairly evenly spread in the nuclei there was an increased amount between the clusters of the internal cuneate nucleus and in the border zone between the gracile and the cuneate nucleus. Electron microscopical investigation showed .GABA labeling in fairly small neurons with relatively large cell nuclei and low somatic bouton covering. GABA-positive terminals with rounded synaptic vesicles were present in all the investigated nuclei also the external cuneate one. No apparent difference in number of such boutons in the different nuclei or parts of the nuclei was found. GABA-positive boutons mostly synapsed with dendrites but in the dorsal column nuclei also with other larger boutons. Axosomatic contacts between labeled terminals and neuronal perikarya were more common in the lateral cervical nucleus than in the dorsal column nuclei. The results from the different nuclei in the monkey were compared with the results of similar investigations in the cat. It is concluded that there are important species differences especially on the light microscopical level in the lateral cervical nucleus. Thus the GABA labeled cells is rather evenly spread over the nucleus in the monkey whereas in the cat they are concentrated to the ventromedial region.

neurons and terminals containing GABA or glutamic acid decarboxylase (GAD), the biosynthetic enzyme of GABA, have been found in the dorsal column nuclei (DCN) (9,11,22,25) in the lateral cervical nucleus (LCN) (43) and in the nucleus of the spinal tract of the trigeminal nerve (3,14).Similar observations have been made in the feline somatosensory thalamic regions, where cell bodies and terminals containing GAD (17,18,21,23) or GABA (13,19,26) have been demonstrated.On the basis of studies with the Golgi technique and retrograde tracing experiments it has been concluded, that the GABA-immunoreactive neurons in the DCN, LCN and somatosensory thalamus are local circuit neurons (5,8,17,18,21,22,23).
So far, however, no studies in primates have been performed in any of the lower somatosensory relay nuclei to confirm that the conditions referred to above also are valid in the highest animals and thus possibly also man.In the present study two different species of monkey, the Crab-eating monkey, Macaca fascicularis , and the Olive baboon, Papio anubis, have been investigated for the presence of GABA in the DCN and LCN.Preliminary results have been published (10).

MATERIALS AND METHODS
Four adult monkeys, two of the species Macaca fascicularis and two of the species Papio anubis were used for the study of gamma-aminobutyric acid (GABA).Under anaesthesia induced by sodium pentobarbital (40-60 mg/kg), the animals intended for the GABA study were perfused through the heart with 2.5 % glutaraldehyde and 2.0 % formaldehyde in 0.1 M phosphate buffer (38"C, pH 7.4), preceded by a rinse with 0.1 M phosphate buffer.After termination of the perfbsion the brain and spinal cord were removed and kept in fresh fixative at 4°C.
Thereafter the lower brain stem and upper cervical spinal cord containing the dorsal column nuclei (DCN) and the lateral cervical nucleus (LCN) were cut transversely on a Vibratome (Oxford Instruments) into sections 100 and 60 pm thick.The thicker sections which were regularely spaced with a t least 3 for each mm were used for orientation purposes and were not subjected t o immunocytochemistry.The 60 pm sections were processed further for GABA immunohistochemistry, by a procedure previously described in detail (5).After incubation for 30 minutes in 1% NaBH4 in phosphate buffered saline (PBS) at room temperature to reduce unspecific activity due to binding of antibodies to free aldehyde groups (12), the sections were transferred to the primary antibody solution, (rabbit anti-GABA antiserum, Immunonuclear Corp, Stillwater, USA) diluted 1500 and normal swine serum 1:30 in PBS and incubated for 36 hours.After rinsing in PBS, the sections were transferred to the secondary antibody solution (swine anti-rabbit 1:30 in PBS) and incubated for 60 minutes at room temperature under agitation; thereafter they were incubated in the PAP complex solution (1:120 in PBS) under the same conditions.The immunocomplexes were localized by incubating the sections for 3-6 minutes in a solution containing 75 mg of diaminobenzidine and 30 pl of 30% H202/100 ml of Tris-HCI buffer.The reaction was terminated by transferring the sections to a bath of Tris-HC1 buffer.They were then postfixed for 20 minutes in 2% Os04 dissolved in cacodylate buffer, dehydrated in a graded series of ethanol and embedded in Epon between acetate foils.The 100pm sections were taken directly to osmification and subsequent dehydration and embedding in Epon.After light microscopic investigation suitable 60 pm thick Vibratome sections were mounted onto Epon blocks and resectioned for electron microscopy.Some of the 60 pm sections, which were exclusively intended for light microscopy were preincubated with 0,2 % Triton X-100 in PBS for one hour a t 20OC before GABA immunocytochemistry.
Control experiments: From every series consecutive sections were incubated in parallel with those intended for GABA immunohistochemistry, the only difference in treatment being that the primary antiserum was omitted in the first incubation of the control sections.In two of the animals, one from each of the used species two additional types of controls (6,16) were made.The first one was a specificity test where the used GABA-serum was tested against test filters containing different amino acids (0-alanine, glutamate, GABA, glycine, glutamine, aspartate and taurine) coupled to rat brain extracts by means of glutaraldehyde.The other one was a n inhibition test where the amino acids were mixed with glutaraldehyde and used to block the primary antiserum.

Control experiments:
The specificity tests demonstrated that the used antiserum labelled the GABA test conjugate but also to a slighter extent 0-alanine, gIycine and glutamine.In the immunocytochemical studies of the Vibratome sections three kinds of intensively labelled structures were found in the light microscope: small cell somata, minute granular structures in the size of boutons and thin fibers sometimes connected to these granular structures.These labelled structures were absent from the controls where the primary antiserum had been omitted or blocked with GABA.When the other amino acids were used for blocking, however, the three kinds of labelled structures were present.Below these structures will be referred to as GABA-positive.GABA immunoreactivity in the DCN Light microscopic studies demonstrated the presence of GABA-positive cells in the gracile (Fig 1A ) and the internal cuneate nucleus (Fig. 1B) but not in the external cuneate nucleus.The labeled cells seemed fairly evenly spread over the total longitudinal extent of the nuclei.There was however, an increased amount of labeled cells between the clusters of the internal cuneate nucleus and in the border zone between the gracile and the cuneate nucleus.The cells were small to mediumsized and rounded or elongated in cross sections from the brain stem (Fig. lA,B).
Electron microscopical investigations demonstrated that the reaction products were only present close to the surface of the Vibratome section.Labeled neurons (Fig. 2A) had a mean diameter of about 15pm, a relatively large cell nucleus and a low somatic bouton covering.Most terminals in contact with labeled cells were unlabeled.
GABA-positive terminals were present in all the investigated DCN (Figs lB, C and Figs 2A,,B, C) also the external cuneate one (Figs 3A, B).NO apparent difference in number of such boutons in the different nuclei or parts of the nuclei was found.GABA-positive boutons had a profile mean diameter of about lpm, contained rounded synaptic vesicles and mostly synapsed with dendrites (Figs 2B, C and 3A) but also with other boutons (Figs 2B, 3B)).These postsynaptic terminals were often conspicuously large up to 5pm in diameter.Axosomatic contacts were rather sparse.GABA immunoreactivity in the LCN: Light microscopy demonstrated a few small rounded or elongated GABA-positive cells (Fig. 4A, B and D).The labeled cells had no preferential localization within the LCN, whose overwhelming majority of neurones were unlabeled and larger than the labeled ones.In the electron microscope (Fig. 5A) the labeled cells demonstrated a fairly large nucleus and rather low somatic bouton covering compared to unlabeled cells .Most of the boutons in contact with labeled cells were unlabeled.Both in the light (Fig. 4C) and electron microscope (Fig. 5A-C) numerous labeled terminals with a mean diameter of about 1 pm were seen without any larger regional differences in number or size.These boutons contained rounded vesicles and synapsed with other dendrites (Fig. 5B) or perikarya (Fig. 5 C).Most terminals were, however, unlabeled and contained rounded or elongated synaptic vesicles.No apparent difference between the two monkeys species Macaca fascicularis and Papio anubis was detected.

DlscuSsION
The present study has demonstrated that the GABA distribution in the two medial dorsal column nuclei (DCN) of the monkey is the same as has been reported for cat (9,111 and could be assumed from studies of glutamic acid decarboxylase, the biosynthetic ezyme of GABA in the cat (22,251 and rat (2).Thus the DCN contain numerous GABA-positive neurons all over the two medial nuclei especially between the cell clusters.These cells are fairly small with a relatively large cell nucleus and few somatic synapses.The GABA-positive terminals have synaptic contact with dendrites and other large boutons probably emerging from the primary afferentdn this way the GABA-positive cells, which most probably are local circuit neurones can exert their presynaptic inhibition through axo-axonal contacts (1,24).It seems likely that the same basic organization is valid also for man although this has to be demonstrated by further studies.
The present study of the external cuneate nucleus appears to be the first electron microscopical one in the primate.It confirms earlier reports of axodendritic and axoaxonal synapses in the cat (15) and rat (20).In the cat the large terminals postsynaptic to smaller knobs have been demonstrated to be derived from primary afferents (15).The present study has shown that the small presynaptic knobs are GABA-positive and thus presumably GABAergic.Further studies are necessary to elucidate the origin of these terminals, which seems to be extrinsic, as the present study failed to detect any GABA-positive perikarya.In the LCN there is a difference in distribution of GABA between the cat ( 5 ) and monkey (present study).Thus the GABA-positive cells although few could be found in all parts of the monkey LCN while they seemed to be confined to to the ventromedial part of the cat LCN.In both species, however, the question remains whether the few GABA-positive cells could give origin to all the terminals or if an additional external GABAergic inhibitory input (7) has to be assumed.The present study has confirmed that also in the primate there is important differences between the synaptic contacts of the GABA-positive terminals in the DCN and the LCN.Thus in the DCN the GABA-positive and presumably inhibitory terminals are presynaptic to other larger terminals at least partially primary afferent ones.In the LCN such axo-axonal contacts are absent.The physiological differences between the DCN and LCN which are closely related to these anatomical differences are to find e.g. in the type of inhibition exerted in the two areas.In the DCN this is of the "surround" type enhancing the contrast of the system, in the LCN it is of a more wide-spread type affecting larger areas.On the other hand there seem to be at least morphologically some similarities especially in the appearence of the GABA-positive and supposedly inhibitory local circuit neurones.Both in the DCN and LCN they are fairly small, have a proportionally large cell nucleus and a low somatic bouton covering.Although further studies of more quantitative nature need to be done, this might be an indication of a general morphologic similarity between inhibitory internuncial neurones in somatosensory systems and perhaps the whole subcortical central nervous system.3.

A.
5500X.GABA positive cell from the internal cuneate nucleus.The cell is B. 19 OOOX Three labeled terminals (arrows) from the gracile nucleus.The lower C. 15 OOOX Two GABA positive gracile terminals, both in contact with dendrites

Fig. 3 .
Fig.3.Electron micrographs of GABA in the external cuneate nucleus.A. 15000X Numerous labeled structures (arrows) of which most are boutons.B. 21 OOOX Three GABApositive terminals in synaptic contact (small arrows) with three larger unlabeled terminals (B1, B2, and B3, respectively).Both labeled and unlabeled terminals contain rounded synaptic vesicles.Note the GABApositive axon (large arrow) connected to the labeled terminal (above).

(
arrows) close to the cell.The right one is in contact with the perikaryon.A. 7000X Labeled neurone from the central part.Two GABApositive terminals B. 20 OOOX Two labeled axodendritic terminals (arrows) from the lateral part.C. 21 OOOX Array of labeled axosomatic terminals with rounded vesicles from the lateral part of LCN in synaptic contact (arrows) with a cell soma to the left.
Dr. Ole Petter Ottersen and Anna T. Bore, Department of Anatomy, University of Oslo, generously provided me with brain stem and spinal cord of perfused monkeys of the species Papio anubis filters for specificity tests and instructions for inhibition tests.I also wish to thank Mrs. Karstin Flink, Mrs. Ingmarie Olsson and Mrs. Kerstin Rystedt for skillful technical assistance.This work was supported by grants from the Swedish Medical Research Council (project No. 2710) and from the Torsten and Ragnar Soderberg Foundation.R E F X R E N m1.Andersen, P., Eccles, JC., Oshima, T. & Schmidt, R.F.: Mechanisms of synaptic transmission in the cuneate nucleus.J Neurophysiol27:1096-1116, 1964.Barbaresi, P. , Spreafico, R., Frassoni, C. & Rustioni, A.: GABAergic neurons are present in the dorsal column nuclei but not in the ventroposterior complex of rats.Brain Res 382:305-326,1986.Basbaum, A I ., Glazer, E J. & Oertel, W.: Immunoreactive glutamic acid decarboxylase in the trigeminal nucleus caudalis of the cat: a light-and electron-microscopic analysis.Somatosens Res 4:77-94, 1986.