Brain Research, 530 (1990) 215-222
215
Elsevier BRES 15930
Direct observations of synapses between GABA-immunoreactive boutons and muscle afferent terminals in lamina VI of the cat's spinal cord D.J. Maxwell, W.M. Christie, A.D. Short and A.G. Brown Department of Preclinical Veterinary Sciences, University of Edinburgh, Summerhall, Edinburgh (U.K.) (Accepted 10 April 1990)
Key words: Spinal cord ; Cat; Ultrastructure; Muscle spindle afferents, y Aminobutyric acid Single group Ia muscle afferent fibres in the lumbar spinal cord of the cat were impaled with microelectrodes and labelled with horseradish peroxidase. Two collateral axons were prepared for combined light and electron microscopy. Arbors selected from lamina VI were processed by the postembedding immunogold technique with antiserum which specifically recognizes GABA in glutaraldehyde-fixed tissue. Twelve Ia boutons were examined through series of thin sections with the electron microscope and all of them were associated with presynaptic axon terminals which were positively labelled for GABA. Some Ia boutons received synaptic contacts from several GABAergic terminals. The present study establishes that a GABA-like substance is present in axon terminals presynaptic to Ia afferent boutons in lamina VI of the spinal cord. This evidence provides a morphological basis for presynaptic inhibition of Ia afferent input into lamina VI.
INTRODUCTION
MATERIALS AND METHODS
Collateral axons of group Ia muscle afferent fibres form characteristic branching patterns in the L6-S1 segments of the cat's spinal cord 4'6'17. In these segments,
Experiments were performed on two adult cats anaesthetized with choralose (70 mg/kg) following induction of anaesthesia with halothane. The animals were paralyzed with gallamine triethiodide and artificially ventilated throughout the experiments. End-tidal CO 2, carotid arterial blood pressure and rectal temperature were monitored continuously. The level of anaesthesia was assessed from the continuous blood pressure record and by examining the degree of pupillary constriction. End-tidal CO2 was kept between 3.5-4.0% by adjusting the stroke volume of the respiratory pump and rectal temperature was maintained with an electric blanket under the animal. Glass microelectrodes containing 8% horseradish peroxidase (HRP) in Tris-HCI buffer with 0.2 M KC! added were used to impale single Ia muscle afferent fibres near their dorsal root entrance zone in the L6 and L7 segments3°. When a succesful impalement was made, putative Ia fibres were identified according to their characteristic resting discharges which altered in response to gentle manipulation of muscle4 and HRP was injected by passing depolarizing pulses through the electrode. A final classification was made on the basis of collateral morphology (see refs. 3 and 4). The two collateral axons examined in this study displayed characteristic morphological features of Ia fibres and each of them formed terminations in laminae VI, VII and IX (for further details on the identification of muscle afferent axons see ref. 3). At the conclusion of experiments animals were rapidly perfused, initially with a saline flushing solution consisting of 300 ml of 0.9% saline, heparin (100 units/ml) and sodium nitrite (0.1%) at 37 °C, and subsequently with 1 litre of fixative containing 2.5% glutaraldehyde and 1% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 37 °C and 2 litres at 4 °C. Blocks of cord containing identified Ia axons were stored in the same fixative for approximately 8 h and were washed several times in phosphate buffer (pH 7.4) before transverse sections (50 #m) were cut with a Vibratome. Sections
they terminate in 3 distinct regions: (1) lamina IX (the m o t o r nucleus); (2) lamina VII (the Ia inhibitory intern e u r o n region); and (3) lamina VI (where n u m e r o u s i n t e r n e u r o n s in reflex pathways are located9'11"lS'19'27'29). Ultrastructural observations on Ia boutons in laminae VI and IX demonstrate that they are postsynaptic to small terminals (the P terminals of Conradi 5) which contain irregulary-shaped agranular vesicles 6'14'2°. Primary afferent depolarization has been recorded from group I fibres 12'19'26'28 and ionophoretic application of G A B A into the ventral horn or intermediate nuclei mimics this effect 7'8"26. It seems likely therefore that the depolarizing action is mediated through P terminals which contain GABA. In the present study we combined intra-axonal labelling of group Ia fibres with horseradish peroxidase ( H R P ) and the postembedding immunogold technique 31'34 using antiserum which specifically recognizes the presence of glutaraldehyde-fixed G A B A in electron microscope sections 16,31-33. We examined Ia boutons in lamina VI in order to determine if G A B A was located within their associated P terminals.
Correspondence: D.J. Maxwell, Department of Preclinical Veterinary Sciences, University of Edinburgh, Summerhall, Edinburgh, EH9 1QH, U.K. 0006-8993/90/$03.50 t~) 1990 Elsevier Science Publishers B.V. (Biomedical Division)
216 were reacted to reveal the presence of HRP using diaminobenzidine (DAB) as a chromogen. Finally, they were immersed in a solution of osmium tetroxide, dehydrated through a series of ethanol solutions, cleared in propylene oxide and flat-embedded in Durcupan between foils of cellulose acetate. Labelled collateral axons were photographed, drawn and reconstructed and 2 collaterals (one from each animal) were selected for combined electron-microscopic and immunochemical analysis. Lamina VI was identified in osmium-treated sections according to the criteria formulated by Rexed 25, who defined it as a broad band occupying the base of the dorsal cell column with a medial zone containing small- or medium-sized cells and a lateral zone consisting of large star-shaped cells. Sections were attached to blocks and cut on an ultramicrotome. Series of thin sections were collected on Formvar-coated single slot gold grids. The postembedding immunogold procedure has been described elsewhere in detail 22"3~. Briefly, sections were etched and deosmificated using 1% solutions of periodic acid and sodium periodate. They were washed in distilled water between changes and were transferred to droplets of Tris-phosphate-buffered saline (TPBS). Sections were incubated in normal goat serum prior to placing them on droplets containing the primary antiserum. The primary antiserum (GABA-9) was kindly supplied to us by Dr. P. Somogyi (University of Oxford) and was used at dilutions which ranged from 1:2000-1:6000. This antiserum is known to recognize fixed GABA in tissue 16'31-33 and its labelling properties with spinal material has been characterized previously22. Following incubation in primary antiserum, grids were washed in Tris-HC1 buffer containing polyethylene glycol and placed in a similar solution containing anti-rabbit IgG coupled to 15-nm gold spheres. Finally the grids were contrasted with lead citrate. Criteria for the identification of synapses between P terminals and Ia boutons were similar to those of ConradiS: (1) cell membranes of boutons displayed immediate appositions which were often associated with contrast-rich material; (2) synaptic vesicles accumulated next to the membrane appositions in presynaptic boutons; (3) triangular electron-dense structures were present at presynaptic appositions. All synapses were examined through series of thin sections and, where necessary, a goniometer was used to tilt the section in order to resolve the active zone. Quantitative analysis of electron micrographs was performed using a Reichert Videoplan and a Magiscan Image Analysis System (Joyce Loebl). The Magiscan program (see ref. 23) enabled Ia and P terminal profiles to be delineated and numbers of gold particles contained within outlined profiles were counted manually. Particle density was calculated and expressed in terms of numbers of particles per am 2. Estimates of average tissue particle densities were made by dropping a card of unit area from a standard height onto the same micrographs from which Ia and P terminal data were gathered and measuring the numbers of particles contained within these randomly defined areas. Measurements were compared statistically using Student's t-test.
r a n g e d in l e n g t h f r o m 1.24/~m to 4.15 p m ( m e a n _+ S . D . = 2.97 _+ 0.9). B o u t o n s f o r m e d synaptic associations with b e t w e e n 1 and 3 p o s t s y n a p t i c structures. T h e s e structures consisted o f s o m a t a (2 contacts), spines (3 contacts) and d e n d r i t e s (13 contacts) (see Figs. 1-3). D e n d r i t e s r a n g e d in d i a m e t e r f r o m 1 . 0 p m to 1 4 . 6 p r o ( m e a n _+ S . D . = 3.99 + 3.71). E x a m i n a t i o n of serial sections r e v e a l e d that synapses w e r e f o r m e d by P t e r m i n a l s on Ia b o u t o n s which satisfied the criteria o u t l i n e d a b o v e in the m e t h o d s section (e.g. see Fig. 2C). A l l P b o u t o n s e x h i b i t e d a c c u m u l a t i o n s of vesicles and d e n s e t r i a n g u l a r structures at t h e i r appositions with Ia t e r m i n a l s , thus indicating that t h e y w e r e p r e s y n a p t i c to Ia b o u t o n s and c o n f i r m i n g p r e v i o u s claims that P b o u t o n s are p r e s y n a p t i c in a x o - a x o n i c a r r a n g e m e n t s with t e r m i n a l s of muscle a f f e r e n t axons 5"6'14"20. O c c a s i o n a l l y , a few vesicles in Ia b o u t o n s w e r e s e e n to be associated with r e g i o n s of m e m b r a n e s
apposed
to P
t e r m i n a l s (e.g. see Fig. 2 E ) . D e s p i t e careful analysis of serial sections, j u n c t i o n a l specializations w e r e n o t a p p a r ent at t h e s e sites of a p p o s i t i o n and t h e r e f o r e t h e y did not fully satisfy the criteria for the i d e n t i f i c a t i o n o f synapses. A t p r e s e n t , the significance of this o b s e r v a t i o n is unclear. A characteristic f e a t u r e o f all Ia b o u t o n s e x a m i n e d was that e a c h r e c e i v e d at least 1, and up to 3, synaptic contacts from P terminals;
8 Ia b o u t o n s
received 1
c o n t a c t f r o m a P t e r m i n a l , 2 r e c e i v e d 2 c o n t a c t s and 2 r e c e i v e d 3 contacts. P t e r m i n a l s r a n g e d in size f r o m 0.42 ktm to 1.26/~m ( m e a n + S . D . = 0.88 + 0.25; n = 18). As
expected
from
the
previous
study 2° t h e s e
small
b o u t o n s c o n t a i n e d i r r e g u l a r l y - s h a p e d a g r a n u l a r vesicles. S o m e b o u t o n s f o r m e d c o m p l e x synaptic a r r a n g e m e n t s , for e x a m p l e on 4 o c c a s i o n s P t e r m i n a l s a p p e a r e d to be p r e s y n a p t i c to Ia b o u t o n s and d e n d r i t e s which w e r e also c o n t a c t e d by the s a m e Ia b o u t o n s and thus f o r m e d triadic a r r a n g e m e n t s (Figs. 1 and 2). A l t h o u g h t h e m a j o r i t y of Ia b o u t o n s w e r e o r g a n i z e d relatively simply, a few w e r e c o m p l e x and w e r e f o u n d to be p r e s y n a p t i c to several structures and p o s t s y n a p t i c to a n u m b e r of P t e r m i n a l s (Figs. 1 and 2).
RESULTS
P o s t e m b e d d i n g i m m u n o l a b e l l i n g d e m o n s t r a t e d that all P t e r m i n a l s in c o n t a c t with Ia b o u t o n s w e r e e n r i c h e d with
In all, 12 Ia b o u t o n s f r o m l a m i n a VI (7 f r o m o n e a f f e r e n t fibre and 5 f r o m the o t h e r ) w e r e e x a m i n e d
G A B A - l i k e i m m u n o r e a c t i v i t y . T h r o u g h series of sections P t e r m i n a l s w e r e consistently strongly l a b e l l e d with g o l d
t h r o u g h series of thin sections. T h e s a m p l e c o n t a i n e d 7 en passant and 5 t e r m i n a l b o u t o n s and their long axes
particles w h e r e a s a s s o c i a t e d structures, such as Ia boutons, w e r e always w e a k l y l a b e l l e d (Fig. 2 A , B ) . Q u a n t i -
Fig. 1. A terminal swelling in a la axon (B~) is illustrated in the light micrograph (C). B l is associated with a neuron (N) of lamina VI. The same structures are illustrated in the electron micrograph (B) and B 1 is illustrated at a higher magnification in A. B t is apposed to the perikaryon (P) of N and is postsynaptic to a small bouton (3) which is heavily labelled with gold particles indicating the presence of GABA within it. D illustrates B~ further through the series of sections. B l is no longer apposed to the perikaryon (P) of N but is now presynaptic to a dendrite (Den) and to a small spine (Sp). Three boutons (1, 2 and 3) which display heavy immunoreactions are associated with B 1. Bouton 3 is the same bouton designated 3 in A through serial sections and bouton 1 is presynaptic to B t and the dendrite (Den) and so forms a triad. These structures are illustrated in more detail in Fig. 2.
~
~ ~
~
~
W~
~
~ lklg~k,2gl ~
"~
~ ~"~'JW ~
~
J~'~*
, , ~ ~
218
Fig. 2. Further details of immunoreactive boutons illustrated in Fig. 1. A and B show B1, bouton 1 and Den in 2 consecutive serial sections. Bouton 1 is consistently labelled heavily with gold particles in both sections whereas B1 and Den are weakly labelled. C: a magnified view of the apposition between bouton 1 and B 1 which has been tilted to reveal a synaptic junction (arrow). Note: (i) the immediate apposition between the two boutons and the slight increase in contrast along the region of apposition; (ii) the accumulation of vesicles in bouton 1 at the region of apposition (also seen in serial section Fig. 2A), and (iii) the triangular presynaptic density at the region of apposition. Bouton 1 was also found to be presynaptic to Den (synaptic specializations are not clearly resolved in the material shown here). D: bouton 2 is heavily labelled with gold particles and is presynaptic to B 1. Part of the synaptic junction formed by B 1 with D e n is also shown in this illustration (between the arrows). E: details of bouton 3. This is presynaptic to B 1 and is heavily labelled with gold particles.
219
Fig. 3. A: a light micrograph illustrating part of a Ia axon that was labelled with horseradish peroxidase. Several swellings are evident. O n e of t h e m (B2) is illustrated in the electron micrographs (B,C,D). BE was found to be an en passant bouton and is located at some distance from a n e u r o n (N) of lamina VI. B: an electron micrograph illustrating B z (boxed area: illustrated in greater detail in C). B z is associated with a large proximal dendrite which through serial sections was found to emerge from N. Ax = axon of N. C: a magnified view of the structures contained within the box in B. B e is associated with a small bouton (boxed area) which is shown at higher magnification in D. D: the small bouton is heavily labelled with gold particles which indicate that it contains G A B A and is presynaptic to B 2 (junction is shown between the arrows).
220 tative analysis confirmed specific differential labelling of P terminals. P terminals were associated with a mean (+ S.D.) particle density of 40.21 (+ 27.93) particles//~m2 (n = 18), Ia boutons with 5.01 (+ 3.69) paricles//~m 2 (n = 12) and the average tissue density was 6.52 (+ 7.64) particles//~m2 (n = 18). Student's t-test demonstrated that the particle density associated with P terminals was significantly different from the average tissue density and Ia densities (P < 0.001) but the two latter densities were not significantly different from each other. From these figures it follows that P terminals are associated with 8.03 times the particle density of Ia boutons and 6.16 times the particle density of tissue in general.
DISCUSSION
The immunogold procedure Labelling properties of the antiserum used in this study have been described in detail elsewhere 16'31-33 and it has been rigorously tested on spinal material 22. Preadsorption of the primary antiserum with GABA coupled to polyacrylamide beads with glutaraldehyde abolishes the immunoreaction whereas preadsorption with other common amino acids prepared in a similar manner has no effect 16. Therefore the antiserum specifically recognizes GABA fixed in tissue with glutaraldehyde and so it is likely that the immunoreaction observed in the present study is due to a GABA-Iike molecule. The postembedding immunogold procedure has a number of advantages over the pre-embedding method and is particulary suitable for studies involving combinations of intracellular staining and immunolabeUing (see ref. 33). The ultrastructural preservation of material treated in this manner is of superior quality to material prepared for pre-embedding immunolabelling and difficulties encountered as a consequence of penetration are largely overcome since the immunoreaction is performed on thin sections. Furthermore, the immunogold reaction does not obscure internal details of structures such as vesicles and junctional specializations; thus synaptic relationships between immunolabelled structures and processes labelled with HRP can be established with a greater degree of certainty than in similar studies using pre-embedding immunostaining methods (e.g. see ref. 21). Labelling in the neuropil of lamina VI was differential and heavy labelling of gold particles was confined to particular structures such as P terminals and a proportion of other boutons and somata in this region. This impression was confirmed by analysis of serial sections; heavy immunoreactions were usually confined to particular structures which were always heavily labelled through the series. Other structures, such as Ia boutons,
were weakly labelled. It has been suggested recently that the density of particle labelling in postembedding immunogold reactions reflects the concentration of antigen 24. It is likely therefore that the observed intense immunogold reaction associated with P terminals is indicative of high concentrations of G A B A within them.
The function of P terminals Some years ago Gray a5 postulated that axo-axonic synaptic arrangements provided the morphological basis of presynaptic inhibition. Barber and his coUeagues 1, in a study of the superficial dorsal horn, extended these observations and demonstrated immunoreactivity for glutamate decarboxylase (GAD) in boutons which were presynaptic to degenerating primary afferent terminals and concluded that the presynaptic boutons were GABAergic. Similar evidence has been presented for identified hair follicle-afferent axon terminals 21 which are postsynaptic to boutons which show a positive immunoreaction for GAD. However, until the present report, little information was available on the relationships between the GABAergic systems and identified primary afferent fibres and no information was available on the relationships formed by GABA-containing boutons with primary afferent fibres which terminate in regions of gray matter outside laminae 1-IV. Observations made in previous ultrastructural studies 6'14'2° of identified Ia boutons in laminae VI and IX demonstrated that they were postsynaptic to other axon terminals but the identity of the neurotransmitter associated with the P terminals was not known although it was suspected to be GABA. The present study demonstrates that boutons which are presynaptic to terminals of Group Ia afferent axons in lamina VI contain a GABA-Iike substance. Activity in segmental and descending axons can depolarize central terminals of Gp I fibres in lamina V111"27"28 and ionophoretic application of G A B A into lamina VI mimics this effect7'8'26. Evidence has occurred suggesting that this primary afferent depolarization (PAD) is a specific effect which is mediated by GABA through bicuculline-sensitive (probably GABAA) receptors 7'8'26. However, there is evidence for the existence of another type of G A B A receptor on Ia terminals which is insensitive to bicuculline but may be activated by the antispasticity drug, baclofen s'13. These G A B A a receptors are thought to reduce the calcium influx into nerve terminals and thus to inhibit synaptic transmission by reducing the amount of neurotransmitter released ~°a3. It follows therefore that GABA contained within P terminals is likely to inhibit transmission from Ia terminals in lamina VI by acting presynaptically through GABA A and GABA B receptors. Previous observations z° on Ia boutons in lamina VI
221 revealed that most (92%) of them were associated with one or more P terminals. In the present study similarly,
neurons in laminae V and VI generates P A D in Ia fibres 19'27 and, in the rat, G A D - i m m u n o r e a c t i v e n e u r o n s
all Ia boutons examined were associated with at least one P terminal. In this respect Ia boutons in lamina VI may differ from those in the motor nucleus where approxi-
have b e e n observed in these laminae 2. It is clear that the
mately a quarter of boutons are associated with P terminals 6 (but see also ref. 14) and in Clarke's column where even fewer Ia boutons are associated with presynaptic structures 35.
boutons rather than globally throughout their arbors.
P terminals containing G A B A were observed to be presynaptic to both terminal and en passant boutons thus suggesting that transmission may be inhibited at terminal b o u t o n s and/or at en passant boutons located at some distance from terminals. Ia afferent input to lamina VI
in an axonal triad. Such triads have been reported previously in ultrastructural studies of Ia fibres in the
therefore appears to be u n d e r considerable presynaptic control and the nature of this control may vary depending u p o n the organization of the G A B A e r g i c interneurones responsible for it. A proportion of single Ia boutons receive synaptic contacts from several G A B A - c o n t a i n i n g boutons. A t present the significance of this observation is unclear; these multiple contacts could originate from several different i n t e r n e u r o n s or, alternatively, they could originate from a single interneuron. The location of these i n t e r n e u r o n s is at present u n k n o w n but it is likely that they are local circuit neurons; microstimulation of
REFERENCES 1 Barber, R.P., Vaughn, J.E., Saito, K., McLaughlin, B.J. and Roberts, E., GABAergic terminals are presynaptic to primary afferent terminals in the substantia gelatinosa of the spinal cord, Brain Research, 141 (1978) 35-55. 2 Barber, R.P., Vaughn, J.E. and Roberts, E., The cytoarchitecture of GABAergic neurons in rat spinal cord, Brain Research, 238 (1982) 305-328. 3 Brown, A.G., Organization in the Spinal Cord, Springer, Berlin, 1982. 4 Brown, A.G. and Fyffe, R.E.W., The morphology of group Ia afferent fibre collaterals in the spinal cord of the cat, J. Physiol., 274 (1978) 111-127. 5 Conradi, S., Ultrastructure of dorsal root boutons on lumbosacral motoneurons of the adult cat as revealed by dorsal root section, Acta Physiol. Scand., Suppl. 332 (1969) 85-115. 6 Conradi, S., CuUheim, S. Gollvik, L. and Kellerth, J.-O., Electron microscopic observations on the synaptic contacts of group Ia muscle spindle afferents in the cat lumbosacral spinal cord, Brain Research, 265 (1983) 31-39. 7 Curtis, D.R. and Malik, R., The effects of GABA on lumbar terminations of rubrospinal neurons in the cat spinal cord, Proc. R. Soc. Lond. Ser. B Biol. Sci., 223 (1984) 25-33. 8 Curtis, D.R., Gynther, B.D. and Malik, R., A pharmacological study of group I muscle afferent terminals and synaptic excitation in lamina VI and Clarke's column of the cat spinal cord, Exp. Brain Res., 64 (1986) 105-113. 9 Czarkowska, J., Jankowska, E. and Sybriska, E., Common interneurones in reflex pathways from group Ia and Ib afferents of knee flexors and extensors, J. Physiol., 310 (1981) 367-380. 10 Dunlap, K., Two types of },-aminobutyric acid receptor on embryonic sensory neurones, Brit. J. Pharmacol., 74 (1981) 579-585. 11 Eccles, J.C., Eccles, R.M. and Lundberg, A., Types of neurone
inhibitory input to Ia fibres in lamina VI is precisely organized and probably acts locally at individual Ia Occasionally, a P terminal was observed to form synapses with a Ia b o u t o n and with a dendrite which was also postsynaptic to the Ia b o u t o n and thus participated
motor nucleus 14. Therefore some of the i n t e r n e u r o n s responsible for presynaptic inhibition of input from Ia axons may also directly inhibit n e u r o n s of lamina VI through postsynaptic mechanisms. In conclusion Ia axon terminals in lamina VI receive synaptic contacts from G A B A - c o n t a i n i n g P boutons which are organized in a precise m a n n e r . It is likely therefore that P b o u t o n s control Ia afferent input to lamina VI presynaptically, by releasing G A B A . Acknowledgements. GABA antiserum was generously donated by P. Somogyi (MRC Anatomical Neuropharmacoiogy Unit, Oxford). We wish to thank C.A. Ingham for discussions and comments on the manuscript and H. Anderson and D. Hall for excellent technical assistance. This work was financed by the MRC.
in and around the intermediate nucleus of the lumbosacral cord, J. Physiol., 154 (1960) 89-144. 12 Eccles, J.C., Kostyuk, P.G. and Schmidt, R.E, Central pathways responsible for depolarization of primary afferent fibres, J. Physiol., 161 (1962) 237-257. 13 Edwards, ER., Harrison, P.J., Jack, J.J.B. and KuUmann, D.M., Reduction by baclofen of monosyrtaptic EPSPs in lumbosacral motoneurones of the anaesthetized cat, J. Physiol., 416 (1989) 439-556. 14 Fyffe, R.E.W. and Light, A.R., The ultrastructure of group Ia afferent fibre synapses in the lumbosacral spinal cord of the cat, Brain Research, 300 (1984) 201-209. 15 Gray, E.G., A morphological basis for presynaptic inhibition?, Nature, 193 (1962) 82-83. 16 Hodgson, A.J., Penke, B., Erdei, A., Chubb, I.W. and Somogyi P., Antisera to ~-aminobutyric acid, I. Production and characterization using a new model system, J. Histochem. Cytochem., 33 (1985) 229-239. 17 Ishizuka, N., Mannen, H., Hongo, T. and Sasaki, S., Trajectory of group Ia afferent fibres stained with horseradish peroxidase in the lumbosacral spinal cord: three-dimensional reconstructions from serial sections, J. Comp. Neurol., 186 (1979) 189-212. 18 Jankowska, E., Johannisson, T. and Lipski, J., Common interneurones in reflex pathways from group Ia and Ib afferents of ankle extensors in the cat, J. Physiol., 310 (1981) 381-402. 19 Jankowska, E., McCrea, D., Rudomin, P. and Sykova, E., Observations on neuronal pathways subserving primary afferent depolarization, J. Neurophysiol., 46 (1981) 506-516. 20 Maxwell, D.J. and Bannatyne, B.A., Ultrastructure of muscle spindle afferent terminations in lamina VI of the cat spinal cord, Brain Research, 288 (1983) 297-301. 21 Maxwell, D.J. and Noble, R., Relationships between hairfollicle afferent terminations and glutamic acid decarboxylasecontaining boutons in the cat's spinal cord, Brain Research, 408 (1987) 308-312.
222 22 Maxwell, D.J., Christie, W.M. and Somogyi, P., Synaptic connections of GABA-containing boutons in the lateral cervical nucleus: an ultrastructural study employing pre-and post-embedding immunocytochemical methods, Neuroscience, 33 (1989) 169-184. 23 Maxwell, D.J., Christie, W.M., Short, A.D., Storm-Mathisen, J. and Ottersen O.P., Central boutons of glomeruli in the spinal cord of the cat are enriched with L-glutamate-like immunoreactivity, Neuroscience, 36 (1990) 83-104. 24 Ottersen, O.P., Postembedding immunogold labelling of fixed glutamate: an electron microscopic analysis of the relationship between gold particle density and antigen concentration, J. Chem. Neuroanat., 2 (1989) 57-66. 25 Rexed, B., The cytoarchitectonic organization of the spinal cord in the cat, J. Comp. Neurol., 96 (1952) 415-466. 26 Rudomin, P., Engberg, I. and Jim6nez, I., Mechanisms involved in presynaptic depolarization of group I and rubrospinal fibres in cat spinal cord, J. Neurophysiol., 46 (1981) 532-548. 27 Rudomfn, P., Jim6nez, I., Solodkin, M. and Duefias, S., Sites of action of segmental and descending control of transmission on pathways mediating PAD of Ia- and Ib-afferent fibres in cat spinal cord, J. Neurophysiol., 50 (1983) 743-769. 28 Rudomin, P., Solodkin, M. and Jim6nez, I., PAD and PAH response patterns of group Ia- and Ib-fibres to cutaneous and descending inputs in the cat spinal cord, J. Neurophysiol., 56 (1986) 987-1006.
29 Rudomfn, P., Solodkin, M. and Jim6nez, I., Synaptic potentials of primary afferent fibres and motoneurons evoked by single intermediate nucleus interneurones in the cat spinal cord, J. Neurophysiol., 57 (1987) 1288-1313. 30 Snow, P.J., Rose, P.K. and Brown, A.G., Tracing axons and axon collaterals using intracellular injection of horseradish peroxidase, Science, 143 (1976) 312-313. 31 Somogyi, P. and Hodgson, A.J., Antiserum to y-aminobutyric acid, III. Demonstration of GABA in Golgi-impregnated neurons and in conventional electron microscopic sections of cat striate cortex, J. Histochem. Cytochem., 33 (1985) 249-257. 32 Somogyi, P., Hodgson, A.J., Chubb, I.W., Penke, B. and Erdei, A., Antiserum to ~,-aminobutyric acid, II. Immunocytochemical application to the central nervous system, J. Histochem. Cytochem., 33 (1985) 240-248. 33 Somogyi, P. and Solt6sz, I., Immunogold demonstration of GABA in synaptic terminals of intracellularly recorded, horseradish peroxidase-filled basket cells and clutch cells in the cat's visual cortex, Neuroscience, 19 (1986) 1051-1065. 34 Van den Pol, A.N., Colloidal gold and biotin-avidin conjugates as ultrastructural markers for neural antigens, Q. J. Exp. Physiol., 69 (1984) 1-33. 35 Walmsey, B., Wieniawa-Narkiewicz, E. and Nicol, M.J., Ultrastructural evidence related to presynaptic inhibition of primary muscle afferents in Clarke's column of the cat, J. Neurosci., 7 (1987) 236-243.
215
Elsevier BRES 15930
Direct observations of synapses between GABA-immunoreactive boutons and muscle afferent terminals in lamina VI of the cat's spinal cord D.J. Maxwell, W.M. Christie, A.D. Short and A.G. Brown Department of Preclinical Veterinary Sciences, University of Edinburgh, Summerhall, Edinburgh (U.K.) (Accepted 10 April 1990)
Key words: Spinal cord ; Cat; Ultrastructure; Muscle spindle afferents, y Aminobutyric acid Single group Ia muscle afferent fibres in the lumbar spinal cord of the cat were impaled with microelectrodes and labelled with horseradish peroxidase. Two collateral axons were prepared for combined light and electron microscopy. Arbors selected from lamina VI were processed by the postembedding immunogold technique with antiserum which specifically recognizes GABA in glutaraldehyde-fixed tissue. Twelve Ia boutons were examined through series of thin sections with the electron microscope and all of them were associated with presynaptic axon terminals which were positively labelled for GABA. Some Ia boutons received synaptic contacts from several GABAergic terminals. The present study establishes that a GABA-like substance is present in axon terminals presynaptic to Ia afferent boutons in lamina VI of the spinal cord. This evidence provides a morphological basis for presynaptic inhibition of Ia afferent input into lamina VI.
INTRODUCTION
MATERIALS AND METHODS
Collateral axons of group Ia muscle afferent fibres form characteristic branching patterns in the L6-S1 segments of the cat's spinal cord 4'6'17. In these segments,
Experiments were performed on two adult cats anaesthetized with choralose (70 mg/kg) following induction of anaesthesia with halothane. The animals were paralyzed with gallamine triethiodide and artificially ventilated throughout the experiments. End-tidal CO 2, carotid arterial blood pressure and rectal temperature were monitored continuously. The level of anaesthesia was assessed from the continuous blood pressure record and by examining the degree of pupillary constriction. End-tidal CO2 was kept between 3.5-4.0% by adjusting the stroke volume of the respiratory pump and rectal temperature was maintained with an electric blanket under the animal. Glass microelectrodes containing 8% horseradish peroxidase (HRP) in Tris-HCI buffer with 0.2 M KC! added were used to impale single Ia muscle afferent fibres near their dorsal root entrance zone in the L6 and L7 segments3°. When a succesful impalement was made, putative Ia fibres were identified according to their characteristic resting discharges which altered in response to gentle manipulation of muscle4 and HRP was injected by passing depolarizing pulses through the electrode. A final classification was made on the basis of collateral morphology (see refs. 3 and 4). The two collateral axons examined in this study displayed characteristic morphological features of Ia fibres and each of them formed terminations in laminae VI, VII and IX (for further details on the identification of muscle afferent axons see ref. 3). At the conclusion of experiments animals were rapidly perfused, initially with a saline flushing solution consisting of 300 ml of 0.9% saline, heparin (100 units/ml) and sodium nitrite (0.1%) at 37 °C, and subsequently with 1 litre of fixative containing 2.5% glutaraldehyde and 1% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 37 °C and 2 litres at 4 °C. Blocks of cord containing identified Ia axons were stored in the same fixative for approximately 8 h and were washed several times in phosphate buffer (pH 7.4) before transverse sections (50 #m) were cut with a Vibratome. Sections
they terminate in 3 distinct regions: (1) lamina IX (the m o t o r nucleus); (2) lamina VII (the Ia inhibitory intern e u r o n region); and (3) lamina VI (where n u m e r o u s i n t e r n e u r o n s in reflex pathways are located9'11"lS'19'27'29). Ultrastructural observations on Ia boutons in laminae VI and IX demonstrate that they are postsynaptic to small terminals (the P terminals of Conradi 5) which contain irregulary-shaped agranular vesicles 6'14'2°. Primary afferent depolarization has been recorded from group I fibres 12'19'26'28 and ionophoretic application of G A B A into the ventral horn or intermediate nuclei mimics this effect 7'8"26. It seems likely therefore that the depolarizing action is mediated through P terminals which contain GABA. In the present study we combined intra-axonal labelling of group Ia fibres with horseradish peroxidase ( H R P ) and the postembedding immunogold technique 31'34 using antiserum which specifically recognizes the presence of glutaraldehyde-fixed G A B A in electron microscope sections 16,31-33. We examined Ia boutons in lamina VI in order to determine if G A B A was located within their associated P terminals.
Correspondence: D.J. Maxwell, Department of Preclinical Veterinary Sciences, University of Edinburgh, Summerhall, Edinburgh, EH9 1QH, U.K. 0006-8993/90/$03.50 t~) 1990 Elsevier Science Publishers B.V. (Biomedical Division)
216 were reacted to reveal the presence of HRP using diaminobenzidine (DAB) as a chromogen. Finally, they were immersed in a solution of osmium tetroxide, dehydrated through a series of ethanol solutions, cleared in propylene oxide and flat-embedded in Durcupan between foils of cellulose acetate. Labelled collateral axons were photographed, drawn and reconstructed and 2 collaterals (one from each animal) were selected for combined electron-microscopic and immunochemical analysis. Lamina VI was identified in osmium-treated sections according to the criteria formulated by Rexed 25, who defined it as a broad band occupying the base of the dorsal cell column with a medial zone containing small- or medium-sized cells and a lateral zone consisting of large star-shaped cells. Sections were attached to blocks and cut on an ultramicrotome. Series of thin sections were collected on Formvar-coated single slot gold grids. The postembedding immunogold procedure has been described elsewhere in detail 22"3~. Briefly, sections were etched and deosmificated using 1% solutions of periodic acid and sodium periodate. They were washed in distilled water between changes and were transferred to droplets of Tris-phosphate-buffered saline (TPBS). Sections were incubated in normal goat serum prior to placing them on droplets containing the primary antiserum. The primary antiserum (GABA-9) was kindly supplied to us by Dr. P. Somogyi (University of Oxford) and was used at dilutions which ranged from 1:2000-1:6000. This antiserum is known to recognize fixed GABA in tissue 16'31-33 and its labelling properties with spinal material has been characterized previously22. Following incubation in primary antiserum, grids were washed in Tris-HC1 buffer containing polyethylene glycol and placed in a similar solution containing anti-rabbit IgG coupled to 15-nm gold spheres. Finally the grids were contrasted with lead citrate. Criteria for the identification of synapses between P terminals and Ia boutons were similar to those of ConradiS: (1) cell membranes of boutons displayed immediate appositions which were often associated with contrast-rich material; (2) synaptic vesicles accumulated next to the membrane appositions in presynaptic boutons; (3) triangular electron-dense structures were present at presynaptic appositions. All synapses were examined through series of thin sections and, where necessary, a goniometer was used to tilt the section in order to resolve the active zone. Quantitative analysis of electron micrographs was performed using a Reichert Videoplan and a Magiscan Image Analysis System (Joyce Loebl). The Magiscan program (see ref. 23) enabled Ia and P terminal profiles to be delineated and numbers of gold particles contained within outlined profiles were counted manually. Particle density was calculated and expressed in terms of numbers of particles per am 2. Estimates of average tissue particle densities were made by dropping a card of unit area from a standard height onto the same micrographs from which Ia and P terminal data were gathered and measuring the numbers of particles contained within these randomly defined areas. Measurements were compared statistically using Student's t-test.
r a n g e d in l e n g t h f r o m 1.24/~m to 4.15 p m ( m e a n _+ S . D . = 2.97 _+ 0.9). B o u t o n s f o r m e d synaptic associations with b e t w e e n 1 and 3 p o s t s y n a p t i c structures. T h e s e structures consisted o f s o m a t a (2 contacts), spines (3 contacts) and d e n d r i t e s (13 contacts) (see Figs. 1-3). D e n d r i t e s r a n g e d in d i a m e t e r f r o m 1 . 0 p m to 1 4 . 6 p r o ( m e a n _+ S . D . = 3.99 + 3.71). E x a m i n a t i o n of serial sections r e v e a l e d that synapses w e r e f o r m e d by P t e r m i n a l s on Ia b o u t o n s which satisfied the criteria o u t l i n e d a b o v e in the m e t h o d s section (e.g. see Fig. 2C). A l l P b o u t o n s e x h i b i t e d a c c u m u l a t i o n s of vesicles and d e n s e t r i a n g u l a r structures at t h e i r appositions with Ia t e r m i n a l s , thus indicating that t h e y w e r e p r e s y n a p t i c to Ia b o u t o n s and c o n f i r m i n g p r e v i o u s claims that P b o u t o n s are p r e s y n a p t i c in a x o - a x o n i c a r r a n g e m e n t s with t e r m i n a l s of muscle a f f e r e n t axons 5"6'14"20. O c c a s i o n a l l y , a few vesicles in Ia b o u t o n s w e r e s e e n to be associated with r e g i o n s of m e m b r a n e s
apposed
to P
t e r m i n a l s (e.g. see Fig. 2 E ) . D e s p i t e careful analysis of serial sections, j u n c t i o n a l specializations w e r e n o t a p p a r ent at t h e s e sites of a p p o s i t i o n and t h e r e f o r e t h e y did not fully satisfy the criteria for the i d e n t i f i c a t i o n o f synapses. A t p r e s e n t , the significance of this o b s e r v a t i o n is unclear. A characteristic f e a t u r e o f all Ia b o u t o n s e x a m i n e d was that e a c h r e c e i v e d at least 1, and up to 3, synaptic contacts from P terminals;
8 Ia b o u t o n s
received 1
c o n t a c t f r o m a P t e r m i n a l , 2 r e c e i v e d 2 c o n t a c t s and 2 r e c e i v e d 3 contacts. P t e r m i n a l s r a n g e d in size f r o m 0.42 ktm to 1.26/~m ( m e a n + S . D . = 0.88 + 0.25; n = 18). As
expected
from
the
previous
study 2° t h e s e
small
b o u t o n s c o n t a i n e d i r r e g u l a r l y - s h a p e d a g r a n u l a r vesicles. S o m e b o u t o n s f o r m e d c o m p l e x synaptic a r r a n g e m e n t s , for e x a m p l e on 4 o c c a s i o n s P t e r m i n a l s a p p e a r e d to be p r e s y n a p t i c to Ia b o u t o n s and d e n d r i t e s which w e r e also c o n t a c t e d by the s a m e Ia b o u t o n s and thus f o r m e d triadic a r r a n g e m e n t s (Figs. 1 and 2). A l t h o u g h t h e m a j o r i t y of Ia b o u t o n s w e r e o r g a n i z e d relatively simply, a few w e r e c o m p l e x and w e r e f o u n d to be p r e s y n a p t i c to several structures and p o s t s y n a p t i c to a n u m b e r of P t e r m i n a l s (Figs. 1 and 2).
RESULTS
P o s t e m b e d d i n g i m m u n o l a b e l l i n g d e m o n s t r a t e d that all P t e r m i n a l s in c o n t a c t with Ia b o u t o n s w e r e e n r i c h e d with
In all, 12 Ia b o u t o n s f r o m l a m i n a VI (7 f r o m o n e a f f e r e n t fibre and 5 f r o m the o t h e r ) w e r e e x a m i n e d
G A B A - l i k e i m m u n o r e a c t i v i t y . T h r o u g h series of sections P t e r m i n a l s w e r e consistently strongly l a b e l l e d with g o l d
t h r o u g h series of thin sections. T h e s a m p l e c o n t a i n e d 7 en passant and 5 t e r m i n a l b o u t o n s and their long axes
particles w h e r e a s a s s o c i a t e d structures, such as Ia boutons, w e r e always w e a k l y l a b e l l e d (Fig. 2 A , B ) . Q u a n t i -
Fig. 1. A terminal swelling in a la axon (B~) is illustrated in the light micrograph (C). B l is associated with a neuron (N) of lamina VI. The same structures are illustrated in the electron micrograph (B) and B 1 is illustrated at a higher magnification in A. B t is apposed to the perikaryon (P) of N and is postsynaptic to a small bouton (3) which is heavily labelled with gold particles indicating the presence of GABA within it. D illustrates B~ further through the series of sections. B l is no longer apposed to the perikaryon (P) of N but is now presynaptic to a dendrite (Den) and to a small spine (Sp). Three boutons (1, 2 and 3) which display heavy immunoreactions are associated with B 1. Bouton 3 is the same bouton designated 3 in A through serial sections and bouton 1 is presynaptic to B t and the dendrite (Den) and so forms a triad. These structures are illustrated in more detail in Fig. 2.
~
~ ~
~
~
W~
~
~ lklg~k,2gl ~
"~
~ ~"~'JW ~
~
J~'~*
, , ~ ~
218
Fig. 2. Further details of immunoreactive boutons illustrated in Fig. 1. A and B show B1, bouton 1 and Den in 2 consecutive serial sections. Bouton 1 is consistently labelled heavily with gold particles in both sections whereas B1 and Den are weakly labelled. C: a magnified view of the apposition between bouton 1 and B 1 which has been tilted to reveal a synaptic junction (arrow). Note: (i) the immediate apposition between the two boutons and the slight increase in contrast along the region of apposition; (ii) the accumulation of vesicles in bouton 1 at the region of apposition (also seen in serial section Fig. 2A), and (iii) the triangular presynaptic density at the region of apposition. Bouton 1 was also found to be presynaptic to Den (synaptic specializations are not clearly resolved in the material shown here). D: bouton 2 is heavily labelled with gold particles and is presynaptic to B 1. Part of the synaptic junction formed by B 1 with D e n is also shown in this illustration (between the arrows). E: details of bouton 3. This is presynaptic to B 1 and is heavily labelled with gold particles.
219
Fig. 3. A: a light micrograph illustrating part of a Ia axon that was labelled with horseradish peroxidase. Several swellings are evident. O n e of t h e m (B2) is illustrated in the electron micrographs (B,C,D). BE was found to be an en passant bouton and is located at some distance from a n e u r o n (N) of lamina VI. B: an electron micrograph illustrating B z (boxed area: illustrated in greater detail in C). B z is associated with a large proximal dendrite which through serial sections was found to emerge from N. Ax = axon of N. C: a magnified view of the structures contained within the box in B. B e is associated with a small bouton (boxed area) which is shown at higher magnification in D. D: the small bouton is heavily labelled with gold particles which indicate that it contains G A B A and is presynaptic to B 2 (junction is shown between the arrows).
220 tative analysis confirmed specific differential labelling of P terminals. P terminals were associated with a mean (+ S.D.) particle density of 40.21 (+ 27.93) particles//~m2 (n = 18), Ia boutons with 5.01 (+ 3.69) paricles//~m 2 (n = 12) and the average tissue density was 6.52 (+ 7.64) particles//~m2 (n = 18). Student's t-test demonstrated that the particle density associated with P terminals was significantly different from the average tissue density and Ia densities (P < 0.001) but the two latter densities were not significantly different from each other. From these figures it follows that P terminals are associated with 8.03 times the particle density of Ia boutons and 6.16 times the particle density of tissue in general.
DISCUSSION
The immunogold procedure Labelling properties of the antiserum used in this study have been described in detail elsewhere 16'31-33 and it has been rigorously tested on spinal material 22. Preadsorption of the primary antiserum with GABA coupled to polyacrylamide beads with glutaraldehyde abolishes the immunoreaction whereas preadsorption with other common amino acids prepared in a similar manner has no effect 16. Therefore the antiserum specifically recognizes GABA fixed in tissue with glutaraldehyde and so it is likely that the immunoreaction observed in the present study is due to a GABA-Iike molecule. The postembedding immunogold procedure has a number of advantages over the pre-embedding method and is particulary suitable for studies involving combinations of intracellular staining and immunolabeUing (see ref. 33). The ultrastructural preservation of material treated in this manner is of superior quality to material prepared for pre-embedding immunolabelling and difficulties encountered as a consequence of penetration are largely overcome since the immunoreaction is performed on thin sections. Furthermore, the immunogold reaction does not obscure internal details of structures such as vesicles and junctional specializations; thus synaptic relationships between immunolabelled structures and processes labelled with HRP can be established with a greater degree of certainty than in similar studies using pre-embedding immunostaining methods (e.g. see ref. 21). Labelling in the neuropil of lamina VI was differential and heavy labelling of gold particles was confined to particular structures such as P terminals and a proportion of other boutons and somata in this region. This impression was confirmed by analysis of serial sections; heavy immunoreactions were usually confined to particular structures which were always heavily labelled through the series. Other structures, such as Ia boutons,
were weakly labelled. It has been suggested recently that the density of particle labelling in postembedding immunogold reactions reflects the concentration of antigen 24. It is likely therefore that the observed intense immunogold reaction associated with P terminals is indicative of high concentrations of G A B A within them.
The function of P terminals Some years ago Gray a5 postulated that axo-axonic synaptic arrangements provided the morphological basis of presynaptic inhibition. Barber and his coUeagues 1, in a study of the superficial dorsal horn, extended these observations and demonstrated immunoreactivity for glutamate decarboxylase (GAD) in boutons which were presynaptic to degenerating primary afferent terminals and concluded that the presynaptic boutons were GABAergic. Similar evidence has been presented for identified hair follicle-afferent axon terminals 21 which are postsynaptic to boutons which show a positive immunoreaction for GAD. However, until the present report, little information was available on the relationships between the GABAergic systems and identified primary afferent fibres and no information was available on the relationships formed by GABA-containing boutons with primary afferent fibres which terminate in regions of gray matter outside laminae 1-IV. Observations made in previous ultrastructural studies 6'14'2° of identified Ia boutons in laminae VI and IX demonstrated that they were postsynaptic to other axon terminals but the identity of the neurotransmitter associated with the P terminals was not known although it was suspected to be GABA. The present study demonstrates that boutons which are presynaptic to terminals of Group Ia afferent axons in lamina VI contain a GABA-Iike substance. Activity in segmental and descending axons can depolarize central terminals of Gp I fibres in lamina V111"27"28 and ionophoretic application of G A B A into lamina VI mimics this effect7'8'26. Evidence has occurred suggesting that this primary afferent depolarization (PAD) is a specific effect which is mediated by GABA through bicuculline-sensitive (probably GABAA) receptors 7'8'26. However, there is evidence for the existence of another type of G A B A receptor on Ia terminals which is insensitive to bicuculline but may be activated by the antispasticity drug, baclofen s'13. These G A B A a receptors are thought to reduce the calcium influx into nerve terminals and thus to inhibit synaptic transmission by reducing the amount of neurotransmitter released ~°a3. It follows therefore that GABA contained within P terminals is likely to inhibit transmission from Ia terminals in lamina VI by acting presynaptically through GABA A and GABA B receptors. Previous observations z° on Ia boutons in lamina VI
221 revealed that most (92%) of them were associated with one or more P terminals. In the present study similarly,
neurons in laminae V and VI generates P A D in Ia fibres 19'27 and, in the rat, G A D - i m m u n o r e a c t i v e n e u r o n s
all Ia boutons examined were associated with at least one P terminal. In this respect Ia boutons in lamina VI may differ from those in the motor nucleus where approxi-
have b e e n observed in these laminae 2. It is clear that the
mately a quarter of boutons are associated with P terminals 6 (but see also ref. 14) and in Clarke's column where even fewer Ia boutons are associated with presynaptic structures 35.
boutons rather than globally throughout their arbors.
P terminals containing G A B A were observed to be presynaptic to both terminal and en passant boutons thus suggesting that transmission may be inhibited at terminal b o u t o n s and/or at en passant boutons located at some distance from terminals. Ia afferent input to lamina VI
in an axonal triad. Such triads have been reported previously in ultrastructural studies of Ia fibres in the
therefore appears to be u n d e r considerable presynaptic control and the nature of this control may vary depending u p o n the organization of the G A B A e r g i c interneurones responsible for it. A proportion of single Ia boutons receive synaptic contacts from several G A B A - c o n t a i n i n g boutons. A t present the significance of this observation is unclear; these multiple contacts could originate from several different i n t e r n e u r o n s or, alternatively, they could originate from a single interneuron. The location of these i n t e r n e u r o n s is at present u n k n o w n but it is likely that they are local circuit neurons; microstimulation of
REFERENCES 1 Barber, R.P., Vaughn, J.E., Saito, K., McLaughlin, B.J. and Roberts, E., GABAergic terminals are presynaptic to primary afferent terminals in the substantia gelatinosa of the spinal cord, Brain Research, 141 (1978) 35-55. 2 Barber, R.P., Vaughn, J.E. and Roberts, E., The cytoarchitecture of GABAergic neurons in rat spinal cord, Brain Research, 238 (1982) 305-328. 3 Brown, A.G., Organization in the Spinal Cord, Springer, Berlin, 1982. 4 Brown, A.G. and Fyffe, R.E.W., The morphology of group Ia afferent fibre collaterals in the spinal cord of the cat, J. Physiol., 274 (1978) 111-127. 5 Conradi, S., Ultrastructure of dorsal root boutons on lumbosacral motoneurons of the adult cat as revealed by dorsal root section, Acta Physiol. Scand., Suppl. 332 (1969) 85-115. 6 Conradi, S., CuUheim, S. Gollvik, L. and Kellerth, J.-O., Electron microscopic observations on the synaptic contacts of group Ia muscle spindle afferents in the cat lumbosacral spinal cord, Brain Research, 265 (1983) 31-39. 7 Curtis, D.R. and Malik, R., The effects of GABA on lumbar terminations of rubrospinal neurons in the cat spinal cord, Proc. R. Soc. Lond. Ser. B Biol. Sci., 223 (1984) 25-33. 8 Curtis, D.R., Gynther, B.D. and Malik, R., A pharmacological study of group I muscle afferent terminals and synaptic excitation in lamina VI and Clarke's column of the cat spinal cord, Exp. Brain Res., 64 (1986) 105-113. 9 Czarkowska, J., Jankowska, E. and Sybriska, E., Common interneurones in reflex pathways from group Ia and Ib afferents of knee flexors and extensors, J. Physiol., 310 (1981) 367-380. 10 Dunlap, K., Two types of },-aminobutyric acid receptor on embryonic sensory neurones, Brit. J. Pharmacol., 74 (1981) 579-585. 11 Eccles, J.C., Eccles, R.M. and Lundberg, A., Types of neurone
inhibitory input to Ia fibres in lamina VI is precisely organized and probably acts locally at individual Ia Occasionally, a P terminal was observed to form synapses with a Ia b o u t o n and with a dendrite which was also postsynaptic to the Ia b o u t o n and thus participated
motor nucleus 14. Therefore some of the i n t e r n e u r o n s responsible for presynaptic inhibition of input from Ia axons may also directly inhibit n e u r o n s of lamina VI through postsynaptic mechanisms. In conclusion Ia axon terminals in lamina VI receive synaptic contacts from G A B A - c o n t a i n i n g P boutons which are organized in a precise m a n n e r . It is likely therefore that P b o u t o n s control Ia afferent input to lamina VI presynaptically, by releasing G A B A . Acknowledgements. GABA antiserum was generously donated by P. Somogyi (MRC Anatomical Neuropharmacoiogy Unit, Oxford). We wish to thank C.A. Ingham for discussions and comments on the manuscript and H. Anderson and D. Hall for excellent technical assistance. This work was financed by the MRC.
in and around the intermediate nucleus of the lumbosacral cord, J. Physiol., 154 (1960) 89-144. 12 Eccles, J.C., Kostyuk, P.G. and Schmidt, R.E, Central pathways responsible for depolarization of primary afferent fibres, J. Physiol., 161 (1962) 237-257. 13 Edwards, ER., Harrison, P.J., Jack, J.J.B. and KuUmann, D.M., Reduction by baclofen of monosyrtaptic EPSPs in lumbosacral motoneurones of the anaesthetized cat, J. Physiol., 416 (1989) 439-556. 14 Fyffe, R.E.W. and Light, A.R., The ultrastructure of group Ia afferent fibre synapses in the lumbosacral spinal cord of the cat, Brain Research, 300 (1984) 201-209. 15 Gray, E.G., A morphological basis for presynaptic inhibition?, Nature, 193 (1962) 82-83. 16 Hodgson, A.J., Penke, B., Erdei, A., Chubb, I.W. and Somogyi P., Antisera to ~-aminobutyric acid, I. Production and characterization using a new model system, J. Histochem. Cytochem., 33 (1985) 229-239. 17 Ishizuka, N., Mannen, H., Hongo, T. and Sasaki, S., Trajectory of group Ia afferent fibres stained with horseradish peroxidase in the lumbosacral spinal cord: three-dimensional reconstructions from serial sections, J. Comp. Neurol., 186 (1979) 189-212. 18 Jankowska, E., Johannisson, T. and Lipski, J., Common interneurones in reflex pathways from group Ia and Ib afferents of ankle extensors in the cat, J. Physiol., 310 (1981) 381-402. 19 Jankowska, E., McCrea, D., Rudomin, P. and Sykova, E., Observations on neuronal pathways subserving primary afferent depolarization, J. Neurophysiol., 46 (1981) 506-516. 20 Maxwell, D.J. and Bannatyne, B.A., Ultrastructure of muscle spindle afferent terminations in lamina VI of the cat spinal cord, Brain Research, 288 (1983) 297-301. 21 Maxwell, D.J. and Noble, R., Relationships between hairfollicle afferent terminations and glutamic acid decarboxylasecontaining boutons in the cat's spinal cord, Brain Research, 408 (1987) 308-312.
222 22 Maxwell, D.J., Christie, W.M. and Somogyi, P., Synaptic connections of GABA-containing boutons in the lateral cervical nucleus: an ultrastructural study employing pre-and post-embedding immunocytochemical methods, Neuroscience, 33 (1989) 169-184. 23 Maxwell, D.J., Christie, W.M., Short, A.D., Storm-Mathisen, J. and Ottersen O.P., Central boutons of glomeruli in the spinal cord of the cat are enriched with L-glutamate-like immunoreactivity, Neuroscience, 36 (1990) 83-104. 24 Ottersen, O.P., Postembedding immunogold labelling of fixed glutamate: an electron microscopic analysis of the relationship between gold particle density and antigen concentration, J. Chem. Neuroanat., 2 (1989) 57-66. 25 Rexed, B., The cytoarchitectonic organization of the spinal cord in the cat, J. Comp. Neurol., 96 (1952) 415-466. 26 Rudomin, P., Engberg, I. and Jim6nez, I., Mechanisms involved in presynaptic depolarization of group I and rubrospinal fibres in cat spinal cord, J. Neurophysiol., 46 (1981) 532-548. 27 Rudomfn, P., Jim6nez, I., Solodkin, M. and Duefias, S., Sites of action of segmental and descending control of transmission on pathways mediating PAD of Ia- and Ib-afferent fibres in cat spinal cord, J. Neurophysiol., 50 (1983) 743-769. 28 Rudomin, P., Solodkin, M. and Jim6nez, I., PAD and PAH response patterns of group Ia- and Ib-fibres to cutaneous and descending inputs in the cat spinal cord, J. Neurophysiol., 56 (1986) 987-1006.
29 Rudomfn, P., Solodkin, M. and Jim6nez, I., Synaptic potentials of primary afferent fibres and motoneurons evoked by single intermediate nucleus interneurones in the cat spinal cord, J. Neurophysiol., 57 (1987) 1288-1313. 30 Snow, P.J., Rose, P.K. and Brown, A.G., Tracing axons and axon collaterals using intracellular injection of horseradish peroxidase, Science, 143 (1976) 312-313. 31 Somogyi, P. and Hodgson, A.J., Antiserum to y-aminobutyric acid, III. Demonstration of GABA in Golgi-impregnated neurons and in conventional electron microscopic sections of cat striate cortex, J. Histochem. Cytochem., 33 (1985) 249-257. 32 Somogyi, P., Hodgson, A.J., Chubb, I.W., Penke, B. and Erdei, A., Antiserum to ~,-aminobutyric acid, II. Immunocytochemical application to the central nervous system, J. Histochem. Cytochem., 33 (1985) 240-248. 33 Somogyi, P. and Solt6sz, I., Immunogold demonstration of GABA in synaptic terminals of intracellularly recorded, horseradish peroxidase-filled basket cells and clutch cells in the cat's visual cortex, Neuroscience, 19 (1986) 1051-1065. 34 Van den Pol, A.N., Colloidal gold and biotin-avidin conjugates as ultrastructural markers for neural antigens, Q. J. Exp. Physiol., 69 (1984) 1-33. 35 Walmsey, B., Wieniawa-Narkiewicz, E. and Nicol, M.J., Ultrastructural evidence related to presynaptic inhibition of primary muscle afferents in Clarke's column of the cat, J. Neurosci., 7 (1987) 236-243.
