THE JOURNAL OF COMPARATIVE NEUROLOGY 32522-37 (1992)

The Termination Pattern and Postsynaptic Targets of Rubrospinal Fibers in the Rat Spinal Cord: A Light and Electron Microscopic Study M. ANTAL, G.N. SHOLOMENKO, A.K. MOSCHOVAKIS, J. STORM-MATHISEN, C.W. HEIZMANN, AND W. HUNZIKER Laboratory of Neural Control, NINDS, NIH, Bethesda, Maryland 20892 (M.A., G.N.S., A.K.M.), Department of Anatomy, University Medical School, Debrecen H-4012, Hungary (M.A.), Department of Basic Sciences, Division of Medicine, School of Health Sciences, University of Crete, Iraklion 71409, Crete, Greece (A.K.M.),Anatomical Institute, University of Oslo, Oslo, N-0317; Norway (J.S.-M.),and Department of Pediatrics, University of Zurich, Zurich CH-8032 (C.W.H.) and Central Research Units, F. Hoffman-La Roche Co., Base1 CH-4002 (W.H.),Switzerland

ABSTRACT The spinal course, termination pattern, and postsynaptic targets of the rubrospinal tract, which is known to contribute to the initiation and execution of movements, were studied in the rat at the light and electron microscopic levels by using the anterograde tracer Phaseolus uulgaris-leucoagglutinin(PHA-L) in combination with calbindin-D28k (CaBP),y-aminobutyric acid (GABA),and glycine immunocytochemistry. After injections of PHA-L unilaterally into the red nucleus, labelled fibers and terminals were detected at cervical, thoracic, and lumbar segments of the spinal cord. Most of the descending fibers were located in the dorsolateral funiculus contralateral to the injection site, but axons descending ipsilaterally were also revealed. Rubrospinal axon terminals were predominantly found in laminae V-VI and in the dorsal part of lamina VII at all levels and on both sides of the spinal cord, but stained collaterals were also seen in the ventrolateral aspect of Clark's column and in the ventral regions of lamina VII on both sides. The proportion of axonal varicosities revealed on the ipsilateral side varied at different segments and represented 10-28% of the total number of labelled boutons. Most of the labelled boutons were engaged in synaptic contacts with dendrites. Of the 137 rubrospinal boutons investigated, only 2 were found to establish axosomatic synaptic junctions in the lumbar spinal cord contralateral to the PHA-L injection. With the postembedding immunogold method, 80.8% of dendrites establishing synaptic contacts with rubrospinal terminals did not show immunoreactivity for either GABA or glycine, whereas 19.2%of them were immunoreactive for both amino acids. Rubrospinal axons made multiple contacts with CaBP-immunoreactive neurons in laminae V-VI. Synaptic contacts between rubrospinal terminals and CaBPimmunoreactive dendrites were identified at the electron microscopic level, and all CaBPcontaining postsynaptic dendrites investigated were negative for both GABA and glycine. The results suggest that rubrospinal terminals establish synaptic contacts with both excitatory and inhibitory interneurons in the rat spinal cord, and a population of excitatory interneurons P 1992 Wiley-Liss, Inc. receiving monosynaptic rubrospinal input is located in laminae V-VI. Key words: descending fibers, spinal interneurons, GABA, glycine, Phaseolus uulgaris-leucoagglutinin, calbindin-D28k

Although the rubrospinal tract is one of the most extensively investigated descending pathways in the spinal cord, there are still some contradictory and unresolved issues concerning the course, termination pattern, and postsynaptic targets of rubrospinal fibers. On the basis of findings obtained by using anterograde fiber degeneration and autoradiographic techniques, the rubrospinal tract has been O

1992 WILEY-LISS, INC.

described as a contralateral pathway descending in the dorsolateral funiculus (Waldron and Gwyn, '69; Edwards, Accepted M~~ 15, 1992. Address reprint requests to Miklds Antal, Dept. Anat., Univ. Med. School, Debrecen, H-4012 Hungary.

RUBROSPINAL TERMINALS IN THE RAT

Fig. 1, Camera lucida drawing of a transverse section of the mesencephalon showing the injection site of PHA-L (cross-hatched area). CS, superior colliculus; PVG, periventricular gray substance; GM, medial geniculate body; NR, red nucleus; SN, substantia nigra.

23 '72; Brown, '74). Most authors also agree that rubrospinal terminals are distributed in Rexed's laminae V-VII and appear only occasionally in more ventral regions of the spinal gray matter (Waldron and Gwyn, '69; Edwards, '72; Brown, '74; Holstege and Kuypers, '82; Holstege and Tan, '88). More recent studies utilizing horseradish peroxidase as an anterograde and retrograde tracer, however, indicate that there is an ipsilateral component of this descending system (Holstege and Kuypers, '82; Shieh et al., '83; Holstege, '87; Holstege and Tan, '88). In addition, Brown ('74) found contralaterally descending rubrospinal fibers that recross the midline at the spinal level and terminate in the ipsilateral spinal cord. Moreover, besides the termination in laminae V-VII, a focus of rubrospinal terminals in lamina IX has also been reported at the level of the cervical cord in the cat (Holstege, '87; McCurdy et al., '87; Robinson et al., '87) and at the level of the cervical and lumbosacral enlargements in the monkey (Holstege et al., '88). Despite many anatomical investigations, the synaptic relations of rubrospinal terminals have received very little attention, and no morphological attempt has been made to

Fig. 2. Photomicrographs showing PHA-L-labelled rubrospinal terminals in the cervical (a)and lumbar

(b-d)spinal cord. Arrows in d point to rubrospinal axons crossing the midline at the level of the lumbar spinal cord. CC, central canal. Bars = 50 km.

M. ANTAL ET AL.

24

locate and neurochemically characterize spinal neurons that receive monosynaptic rubrospinal input. Although electrophysiological evidence from the cat suggests that interneurons receiving monosynaptic excitatory postsynaptic potentials (EPSPs)from rubrospinal fibers are located in laminae V-VII and both excitatory and inhibitory interneurons are contacted by rubrospinal terminals (Baldissera et al., '71; Hongo et al., '72, '89a,b; Hultborn et al., '76; Illert et al., '77, '78; Alstermark et al., '84a,b; Harrison and Jankowska, '851, these findings have never been verified by morphological and neurochemical studies. In the present experiment, we studied the spinal course and termination pattern of rubrospinal fibers in the rat spinal cord employing the highly sensitive anterograde tracing substance Phaseolus vulgaris-leucoagglutinin (PHA-L) (Gerfen and Sawchenko, '84). Combining the axonal tracing with the immunocytochemical detection of calbindin-D28k (CaBP),a calcium-binding protein that has been reported to be a marker of certain subsets of spinal interneurons (Antal et al., '90, '91a), we intended to identify neurons contacted by rubrospinal terminals. The synaptic relations of rubrospinal terminals with CaBPimmunoreactive and unstained neurons were also investigated a t the electron microscopic level with special emphasis on the neurochemical characterization of the postsynaptic targets. Preliminary observations from this experiment have been reported in abstract form (Antal et al., '91b).

MATERIALS AND METHODS Animals, injection of PHA-L, and preparation of tissue sections Experiments were carried out on eight Sprague-Dawley rats. The skull was opened with a dental drill under deep sodium pentobarbital anaesthesia (35 mg/kg. i.p.1, while the animal was held in a stereotaxic frame. Glass micropipettes with a tip diameter of 20-30 p,m were filled with a 2.5% solution of PHA-L (Vector Labs.) dissolved in 0.05 M Tris-buffered saline (TBS, pH 7.4). The tracer was injected unilaterally into the red nucleus by iontophoresis, using positive direct current of 5 p,A with a pulse duration of 7 seconds followed by 3-second intervals for a period of 20 minutes. The coordinates for the injection were 1.8 mm rostra1 from the interauricular line, 0.9 mm from the midline, and 7.8 mm from the upper surface of the brain. After a 4 week survival period, the animals were reanesthetized with an overdose of sodium pentobarbital (70 mg/kg), and perfused transcardially with Tyrode's solution (oxygenated with a mixture of 95% 02,5% COz),followed by a fixative containing 2.5% glutaraldehyde, 0.5%paraformaldehyde, and 0.2% picric acid in 0.1 M phosphate buffer (PB, pH 7.4). The brain and spinal cord were removed and postfixed in the same fixative for 1-2 hours. Blocks of the cervical, thoracic, and lumbar segments of the spinal cord as well as the mesencephalon were dissected and immersed in 10% and 20% sucrose dissolved in 0.1 M PB until they sank. Tissue blocks were freeze-thawed in liquid nitrogen, sectioned at 50 p,m on a vibratome, and extensively washed in 0.1 M PB.

Preembedding immunocytochemistry For immunocytochemical detection of PHA-L, freefloating sections of the cervical, thoracic, and lumbar segments of the spinal cord as well as the mesencephalon

Fig. 3. Schematic representation of the distribution of descending rubrospinal fibers and terminals at the cervical, thoracic, and lumbar segments of the spinal cord. Dots represent descending fibers in the white matter and houtons in the gray matter. The borders of the gray and white matters as well as the Clark's column at the level of the ThlZ-L1 segments of the spinal cord are drawn with dashed lines.

were first incubated with biotinylated goat anti-PHA-L (Vector Labs., diluted 1:2,000) for 2 days at 4°C. Then the sections were transferred into a solution of avidinbiotinylated peroxidase complex (ABC; Vector Labs., diluted 1 : l O O ) for 4 hours at room temperature. The immunoreaction was completed with a nickel-intensified diaminobenzidine (DAB) chromogen reaction (Hancock, '84). For simultaneous visualization of PHA-L-labeled rubrospinal terminals and CaBP-containing spinal interneuions in the same material, sections of the cervical and lumbar segments of the spinal cord were processed by a double-immunostaining procedure. In order to enhance the immunoreactivity for CaBP, free-floating sections were treated with 1%sodium borohydride for 30 minutes, and then extensively washed in 0.1 M PB. First the sections were incubated in a mixture of biotinylated goat antiPHA-L (Vector Labs., diluted 1:2,000) and mouse antiCaBP (code: 13D3A10,diluted 1:5,000-1:10,000) for 2 days at 4°C. The immunological and immunocytochemical characteristics of the anti-CaBP antibody have been published earlier (Pinol et al., '90). Subsequently the sections were transferred into a mixture of avidin-biotinylated peroxidase

RUBROSPINAL TERMINALS IN THE RAT

Fig. 4. Electron micrographs showing PHA-L-labelled rubrospinal terminals that establish synaptic contacts with dendrites (a-d) and somata ( e )in the lumbar spinal cord. White asterisks label rubrospinal boutons. Arrows point to synaptic contacts established by labelled

25

rubrospinal terminals. Note that dendrites postsynaptic to rubrospinal terminals also receive synapses from unlabelled axons (arrowheads) in close proximity to the rubrospinal boutons. Bars = 1 pm.

26

M. ANTAL ET AL.

N= 135

Fig. 6. Distribution of the diameter of dendrites establishing synaptic contacts with rubrospinal terminals in the lumbar spinal cord.

Fig. 5. Electron micrographs showing a PHA-L-labelled rubrospinal terminal impinging upon a dendrite in close proximity to the cell body. b shows the area outlined in a at higher magnification. The rubrosuinal bouton is labelled with a white asterisk in b.-Bars = 10 km (a); 1km (b).

complex (ABC; Vector Labs., diluted 1:lOO) and goat antimouse IgG (Boehringer Mannheim, diluted 1 5 0 ) and left overnight at 4°C. The PHA-L-labeled axons were visualized with a nickel-enhanced DAB chromogen reaction (Hancock, '84). Sections were then treated with a mouse peroxidase anti-peroxidase complex (Boehringer Mannheim, diluted 1 : l O O ) and the immunostaining for CaBP was completed with a chromogen reaction using DAB alone. All incubations were performed under continuous gentle agitation. Before the antibody treatments, sections were kept in 20% normal goat serum (Vector Labs.) for 50 minutes. All of the antibodies were diluted in 0.01 M phosphate-buffered saline (PBS, pH 7.4) to which 0.1% Triton X-100 and 1%normal goat serum (Vector Labs.) were added. Between incubations in the antibody solutions, sections were rinsed three times for 30 minutes in the same buffer. Sections for light microscopy were mounted on gelatin-coated slides and covered with Permount neutral medium. Alternate sections from both incubation procedures were processed for electron microscopy in the same way except that Triton X-100 was omitted from all the incubation solutions. After the completion of the immunostaining, 4 in 0.1 M PB sections were treated with 0.5%0 ~ 0 dissolved for 30-60 minutes, and then dehydrated and flat-embedded into Durcupan ACM resin (Fluka).

27

RUBROSPINAL TERMINALS IN THE RAT

Fig. 7. Electron micrographs showing a PHA-L-labelled rubrospinal terminal in consecutive sections immunostained by the immunogold procedure for glycine (a)and GABA ( c ) , or counterstained with lead citrate (b).The dendrite (D) contacted by the rubrospinal terminal is negative for both GABA and glycine (GLY). Note that the dendrite

postsynaptic to the rubrospinal terminal also receives synapses from axons immunoreactive for GABA in close proximity to the rubrospinal bouton. White asterisks label the rubrospinal bouton. Black asterisks identify the same axonal profiles in consecutive sections. Arrows point to synaptic contacts. Bars = 1 pm.

Postembedding immunocytochemistry

immunostaining, many cells incorporated the tracer within the nucleus. The appearance of these cells, as well as the injection site, were similar to those reported in previous studies (Gerfen and Sawchenko, '85; Wouterlood and Groenewegen, '85). Although the tracer clearly spread into the surrounding white matter, labelled cell bodies were not found outside the red nucleus.

Selected sections processed for electron microscopy were reembedded, and serial ultrathin sections were cut and mounted on formvar-coated single-slot nickel grids. Sections on consecutive grids were processed for GABA and glycine immunocytochemistry, while sections on every third grid were counterstained with lead citrate. The immunostaining was performed according to the GABA-gold I procedure described by Somogyi and Hodgson ('85) with the modifications of Somogyi and Soltesz ('86). Sections were treated with 1%periodic acid for 8 minutes, and then transferred onto 2% sodium periodate for 10 minutes. Following treatment with 1%ovalbumin (Sigma) for 30 minutes, sections were incubated with rabbit antiGABA (code: 9, diluted 1:2,000) or rabbit anti-glycine (code: 290, diluted 1:2,000) for 90 minutes. The immunological and immunocytochemical characteristics of anti-GABA (Hodgson et al., '85; Somogyi et al., '85) and anti-glycine (Storm-Mathisen and Ottersen, '90; Walberg and Ottersen, '92) antibodies have been extensively tested and published earlier. Subsequently the grids were transferred onto goat anti-rabbit IgG-coated colloidal gold (15 nm, BioCell, diluted 1:20) for 2 hours. Sections were counterstained with uranyl acetate and lead citrate. As controls for the specificity of the immunostaining procedure, some sections were incubated in normal rabbit serum (diluted 1:lOO) instead of the primary antiserum. No specific staining was observed in these sections.

RESULTS Injection site of PHA-L

Distribution of rubrospinal fibers in the spinal cord Immunostained rubrospinal axons and collaterals with varicose axon terminals were found at cervical, thoracic, and lumbar segments of the spinal cord, although the labelling was more intense at rostral than at caudal spinal segments (Fig. 2). In agreement with previous studies (Waidron and Gwyn, '69; Brown, '74), the majority of descending fibers was found in a wedge-shaped region of the dorsolateral funiculus contralateral to the site of PHA-L application (Fig. 3). In addition to the contralateral descending tract, a few ipsilaterally descending axons were also observed. These axons were located in the dorsolateral funiculus and were present at all levels of the spinal cord (Fig. 3). Stained terminals were primarily distributed in laminae V-VI and the dorsal part of lamina VII at all levels and on both sides of the spinal cord (Fig. 3). Varicose axons occasionally extended into the ventral regions of lamina VII, but labelled fibers were never observed within the motor nucleus. Terminals were also found in the ventrolatera1 and ventromedial aspects of Clark's column on both sides (Fig. 3, Thl2-Ll). A few stained fibers crossed the midline and extended to the opposite side at all levels of the spinal cord (Fig. 2d). Varicosities along the stained axons, regarded as potential synaptic sites of the terminals, were counted at the level of the C6-7, Th5-9, Thl2-L1, L2-3, and L4-5 segments of ~~

The tracer was delivered into the rostral half of the mesencephalon. The injection site extended 300-400 bm along the rostrocaudal axis and involved the entire crosssectional area of the red nucleus (Fig. 1).As detected by the

28

M. ANTAL ET AL.

Figure 8

the spinal cord in five randomly selected sections at each level. On average, in a 50 pm thick transverse section 1,034, 339, 366, 526, and 336 boutons were counted, respectively. The proportion of varicosities revealed on the ipsilateral side varied at different levels and represented 10-28% of the total number of labelled boutons.

with spheroid vesicles, while others contained flattened or pleomorphic synaptic vesicles. A proportion of these axons formed asymmetric, while others established symmetric, synaptic contacts with the postsynaptic dendrites (Fig. 4a,b,d).

GABA and glycine immunoreactivity of rubrospinal terminals and postsynaptic dendrites

Postsynaptic targets of rubrospinal terminals Immunostained terminals were investigated at the ultrastructural level in the lumbar spinal cord contralateral to the injection site of PHA-L. The outline of the labelled profiles was dome shaped or slightly oval with a diameter of 0.5-3.0 pm (Figs. 4, 5). They were engaged in synaptic contacts mostly with dendrites (Figs. 4a-d). The diameter of the postsynaptic dendrites varied in a range of 0.4-4.0 pm with a peak between 1.0 and 2.0 pm (Fig. 6). Of the 137 rubrospinal boutons investigated only 2 (1.46%)were found to establish axosomatic synaptic contact (Fig. 4e). The ultrastructural appearance of the majority of postsynaptic dendrites was dominated by mitochondria and a prominent microtubular compartment (Figs. 4a,b,d, 5). In a proportion of dendrites impinged upon by rubrospinal terminals, however, microtubules were less frequently encountered, although they also contained numerous mitochondria (Fig. 4c). Synaptic contacts between labelled rubrospinal boutons and unstained spinal axonal profiles were never observed. Dendrites postsynaptic to rubrospinal terminals, however, frequently established synaptic contacts with unlabelled axons in close proximity to the rubrospinal bouton (Fig. 4a,b,d). The ultrastructural composition of these unstained boutons was variable. Some of them were densely packed

Rubrospinal terminals were never found immunoreactive for either GABA or glycine. However, unlabelled boutons impinging upon the same postsynaptic dendrites as rubrospinal boutons were frequently stained by the immunogold procedure. Some of these axonal varicosities showed GABA (Fig. 7) or glycine (Fig. 14) immunoreactivity without displaying positive staining for the other amino acid, and a proportion of them were found immunoreactive for both GABA and glycine (Figs. 8 , 9 ) .

Figs. 8 and 9. Electron micrographs showing a rubrospinal terminal in consecutive sections immunostained by the immunogold procedure for GABA (Fig. 9a)and glycine (Fig. 9b), or counterstained with lead citrate (Fig. 8 ) .The dendrite (D) contacted by the rubrospinal terminal is immunoreactive for both GABA and glycine (GLY). Note that the dendrite postsynaptic to the rubrospinal terminal also receives synapses from axons immunoreactive for both GABA and glycine in close proximity to the rubrospinal bouton. White asterisks label the rubrospinal bouton. Black asterisks identify the same axonal profiles in consecutive sections. Open stars mark dendrites that are negative for both GAl3A and glycine. Arrows point to synaptic contacts. Bars = 1

w.

Figure 9

M. ANTAL ET AL.

30

Fig. 10. Photomicrographs of sections double-immunostainedfor simultaneous visualization of PHA-Llabelled rubrospinal terminals (stained black by nickel-intensified DAB reaction; see Materials and Methods) and CaBP-containing neurons (stained brown by DAB without intensification) in the cervical (b,d,e,Dand lumbar (a,c) spinal cord. Arrows point to rubrospinal boutons that are in close apposition to dendrites or somata of CaBP-immunoreactive neurons. Bars = 10 rrm.

Of the 137 rubrospinal boutons studied at the ultrastruct u r d level, 103 were processed for the postembedding immunogold procedure t o reveal GABA and glycine immunoreactivity of the postsynaptic dendrites. The immunostaining for both GABA and glycine could be unambiguously evaluated in the case of 7 3 dendrites. Of the

73 postsynaptic dendrites, 59 (80.8%)were negative for both GABA and glycine (Fig. 71, whereas 14 (19.2%) were found immunoreactive for both amino acids (Figs. 8, 9). Postsynaptic dendrites displaying immunoreactivity for one of the amino acids and negative for the other were not observed.

RUBROSPINAL TERMINALS IN THE RAT

31

Fig. 11. Camera lucida drawings from double-immunostained sections of the cervical spinal cord showing close appositions between rubrospinal terminals and CaBP-immunoreactive neurons. Cells labelled with capital letters (A-D) were drawn from horizontal sections, whereas neurons marked with arabic numerals (14)were found in transverse sections. Dots on the inserted drawing of the transected

spinal cord indicate the location of CaBP-immunoreactiveneurons that receive direct rubrospinal contacts. Arabic numerals on the insert identify neurons shown on the camera lucida drawings. The borders of Rexed laminae I-VI are drawn with dotted lines and refer to Molander et al. ('89). Bar = 10 pm.

Location and morphology of CaBP-immunoreactive neurons contacted by rubrospinal terminals

of the perikarya endowed these cells with a bipolar appearance. The diameter of stem dendrites varied in a range of 1.5-5.0 km, with a mean value of 2.2 Fm (n = 24). The dendritic arbors were oriented mediolaterally, and the dendritic tree spanned across a few hundred micrometers in this direction (Fig. 12).

The overall distribution and morphology of spinal neurons that were found immunoreactive for CaBP in the present experiment were identical to those described previously in the rat (Yamamoto et al., '89; Antal et al., '90). The evaluation of sections double-stained for PHA-L and CaBP revealed close appositions between rubrospinal terminals and a population of CaBP-immunoreactive neurons both at the cervical and lumbar segments of the spinal cord (Figs. 10-12). Rubrospinal axons formed multiple contacts with the stained neurons, and the terminals impinged upon distal and proximal dendrites as well as perikarya of the CaBP-containing cells (Figs. 10-12). In the cervical cord, most of the neurons contacted by rubrospinal axons were located in lamina V and presented multipolar perikarya with three to six stem dendrites, but fusiform neurons were also contacted by stained boutons. The diameter of cell bodies varied in a range of 10-30 km, and the proximal dendrites that were revealed by the immunostaining were generally oriented in the rostrocaudal plane (Fig. 11). The morphology of neurons receiving rubrospinal contacts in the lumbar cord was strikingly different. Most of them were located in lamina VI and presented a smaller cell body with a diameter of 10-20 km. Two or three slender, poorly arborizing dendrites originating from opposite poles

GABA and glycine immunoreactivity of CaBP-containing postsynaptic dendrites Double-immunostained sections were also studied at the electron microscopic level. The peroxidase reaction endproduct in CaBP-immunoreactive dendritic profiles was principally associated with microtubules, but electron-dense deposits were also frequently precipitated on the outer membrane of mitochondria (Figs. 13, 14). This immunostaining pattern was clearly distinguishable from the PHA-L labelling of presynaptic boutons, which produced a homogeneous dense staining in the labelled profiles (Figs. 4 , 5 , 7 , 8, 13). Similar to the light microscopic findings, close appositions between rubrospinal terminals and CaBPimmunoreactive neurons were also revealed in sections processed for electron microscopy, and these appositions were identified as synaptic contacts at the ultrastructural level (Figs. 13, 14): Sixteen rubrospinal terminals establishing synaptic contacts with dendrites immunostained for CaBP were processed for the immunogold procedure to determine whether

M. ANTAL ET AL.

32

l \

Fig. 12. Camera lucida drawings from double-immunostained sections of the lumbar spinal cord showing close appositions between rubrospinal terminals and CaBP-immunoreactive neurons. All cells were drawn from transverse sections of the spinal cord. Dots on the insert (drawing of the transected spinal cord) indicate the location of

CaBP-immunoreactive neurons that receive direct rubrospinal contacts. Arabic numerals on the insert identify neurons shown on the camera lucida drawings. The borders of Rexed laminae I-Vl are drawn with dotted lines and refer to Molander et al. ('84).Bar = 10 wm.

GABA or glycine were present in the postsynaptic dendrites. The immunostaining for both GABA and glycine could be unambiguously evaluated in 12 dendrites, all of which were negative for both GABA and glycine (Fig. 14).

lian central nervous system (Gerfen and Sawchenko, '84, '85; Keller et al.,'85; Wouterlood and Groenewegen, '85; Grove et al., '86; Wouterlood et al., '87). Moreover, in unsuccessful injections in which PHA-L was delivered lateral or ventral to the red nucleus, stained fibers occasionally appeared in the anterior white matter, but they were never seen in the dorsolateral funiculus and we have never observed labelled axons more caudal than the cervical spinal cord. On the other hand, when the injection site was centered in the red nucleus, labelled descending axons were seen exclusively in the dorsolateral funiculus down to the lumbosacral segments of the spinal cord, Therefore, it seems likely that fibers of passage with unknown origin were not significantly labelled in our experiment. The PHA-L injection sites involved only a 300-400 km long segment of the red nucleus, and the lectin was delivered only into the rostral half of the nucleus. Thus we labelled only a certain proportion of the rubrospinal fibers. It is therefore necessary to raise the question of whether the labelled axons represent the spinal course and termination pattern of the entire rubrospinal tract. Neurons projecting to cervical and lumbar segments of the spinal cord are evenly distributed throughout the whole rostrocaudal extent of the red nucleus in the rat (Huisman et al., '81; Shieh et al., '83).Injecting 3H-leucine and horseradish peroxidase (HRP) into various regions of the red nucleus, Holstege and Tan ('88) did not find any clear differences between the spinal projections from the rostral and those from the

DISCUSSION Labelling of rubrospinal fibers Cells of origin of the rubrospinal tract are arranged in a somatotopic fashion. Neurons projecting to the cervical cord are located in the dorsomedial regions of the red nucleus and those projecting to the lumbosacral segments are distributed in the ventrolateral part of the nucleus (Hayes and Rustioni, '81; Huisman et al., '81; Shieh et al., '83; Holstege and Tan, '88). Since we intended to label both populations of the projecting neurons, we produced injections that covered the whole cross-sectional area of the red nucleus. In addition to the red nucleus, however, the PHA-L clearly spread into the surrounding white matter. Therefore in principle, fibers originating outside the red nucleus and traversing the white matter in close vicinity to the nucleus could also take up and transport the tracer, causing unspecific labelling in the spinal cord. It is highly probable, however, that this mechanism did not play a significant role in our experiment. Although fibers of passage are reported to take up and transport the lectin anterogradely (Cliffer and Giesler, '881, the efficacy of this labelling has always been found to be poor in the mamma-

RUBROSPINAL TERMINALS IN THE RAT

33

Fig. 13. a-c: Electron micrographs showing PHA-L-labelled rubrospinal terminals that establish synaptic contacts with dendrites

immunoreactive for CaBP. White asterisks label rubrospinal boutons. Arrows point to synaptic contacts. Bars = 1km.

caudal red nucleus in the cat. Furthermore, lesions of the rostra1 or caudal portions of the red nucleus produced almost identical patterns of degeneration in the rat spinal cord (Waldron and Gwyn, '69; Brown, '74). In the light of these results, we assume that descending fibers labelled in our experiments can be regarded as a representative sample of the entire rubrospinal tract.

the rat (Waldron and Gwyn, '69; Brown, '741, rubrospinal terminals have been revealed in laminae V-VII at all levels and on both sides of the spinal cord. Although rubrospinal axon terminals in lamina IX have recently been reported in the cat (Holstege, '87; McCurdy et al., '87; Holstege and Tan, '88) and in the monkey (Holstege et al., '88; Ralston et al., '881, stained fibers have not been observed among rat motoneurons in the present experiment.

Spinal course and termination pattern of rubrospinal fibers Based on the results of lesion-degeneration and autoradiographic studies, the mammalian rubrospinal tract has been described as a purely contralateral pathway descending in the dorsolateral funiculus and terminating in laminae V-VII of the spinal gray matter (for review, see Massion, '67). More recent studies, however, questioned the validity of this classical description. Utilizing 3H-leucine as an anterograde tracer, ipsilateral rubrospinal fibers have been demonstrated at cervical and upper thoracic segments of the cat spinal cord (Holstege, '87; Holstege and Tan, '88) and at the level of the cervical and lumbosacral enlargements in the monkey (Holstege et al., '88).In addition, after unilateral injection of HRP into the cervical and lumbar spinal cord of the rat, labelled neurons were revealed bilaterally in the red nucleus (Shieh et al., '83). In the present study, we have also demonstrated the presence of some ipsilaterally descending rubrospinal axons that extend as caudal as the lumbar spinal cord in the rat. Our results also show that the ipsilateral component represents a small but significant proportion (10-28%) of rubrospinal terminals. Axon terminals in the ipsilateral spinal cord may arise from two sources. A proportion of them may represent collaterals of ipsilaterally descending axons, but fibers descending in the contralateral dorsolateral funiculus and recrossing the midline at the level of spinal termination (e.g., Fig. 2d) may also add to the number of terminals in the ipsilateral gray matter. Confirming previous findings in

Synaptic relations of rubrospinal terminals Postsynaptic targets. In the only available ultrastructural study concerning the postsynaptic targets of rubrospinal axons in the rat, degenerating axons were found to make synaptic contacts exclusively with small and medium caliber dendrites following lesion of the red nucleus (Brown, '74). Terminals that impinged upon perikarya or dendrites with a diameter larger than 2.0 Fm were not seen to undergo degeneration. Our present results, however, demonstrate that terminals of rubrospinal axons do establish some synaptic contacts with somata in the rat lumbar spinal cord. In addition to the axosomatic contacts, a significant proportion of rubrospinal terminals impinged upon dendrites with a diameter larger than 2.0 pm in our material. Moreover, first order dendrites contacted by labelled boutons were also seen both at the light and electron microscopic levels. These observations suggest that rubrospinal terminals may be widely distributed on the dendritic tree of postsynaptic neurons in the rat spinal cord. They are located on both distal and proximal dendrites, and occasionally make synaptic contacts even with cell bodies. Neurochemical character ofpostsynapticdendrites. The present paper is the first morphological account of the neurochemical characterization of dendrites receiving synaptic contacts from rubrospinal terminals. We have demonstrated that GABA and glycine, which are known to be inhibitory neurotransmitters in the central nervous system, were both present within 19.2%of dendrites contacted by rubrospinal terminals, whereas 80.8% of the postsynap-

Figure 14

RUBROSPINAL TERMINALS IN THE RAT tic dendrites were negative for both amino acids. Postsynaptic dendrites displaying immunoreactivity for one of the amino acids and negative for the other have not been revealed. This perfect colocalization between GABA and glycine immunoreactivity in dendrites postsynaptic to rubrospinal terminals may raise questions about the specificity of the immunoreactions, with the possibility that both antisera recognize the same antigen. This is, however, unlikely since the immunocytochemical properties of both antisera have been extensively characterized and found to show a high degree of specificity in their reactions with fixed amino acids (Hodgson et al., '85; Somogyi et al., '85; Storm-Mathisen and Ottersen, '90; Walberg and Ottersen, '92). Furthermore, numerous dendrites without rubrospinal contact as well as unlabelled axon terminals showed positive staining for either GABA or glycine and were negative for the other amino acid in our preparations. On the other hand, it is conceivable that a certain amount of GABA and glycine was lost during the fixation procedure. Since the intensity of immunostaining in dendrites was always weaker than in axon terminals, this potential loss in the quantity of the amino acids may have caused falsenegative staining in a population of dendrites that contain GABA and/or glycine in a relatively low concentration. Dendrites containing only one of the amino acids could be among those that were not immunostained for these reasons, suggesting that the proportion of postsynaptic dendrites containing GABA andlor glycine may be somewhat underestimated in our study. Neurochemical character of rubrospinal terminals. Several lines of physiological evidence suggest that rubrospinal axons monosynaptically excite various spinal interneurons (Hongo et al., '69a,b, '72, '89a,b; Burke et al., '70; Baldissera et al., '71; Hultborn et al., '76; Illert et al., '77, '78; Fleshman et al., '88; Harrison and Jankowska, '85; Fujito et al., '91), and inhibitory monosynaptic effects evoked by the stimulation of the rubrospinal tract have never been reported. Following injections of HRP or wheat germ agglutinin (WGA)-HRP into the rat spinal cord, retrogradely labelled neurons in the red nucleus have also been found negative for glutamic acid decarboxylase (Beitz and Ecklund, '88). Confirming these previous results, we have detected neither GABA nor glycine in rubrospinal terminals, suggesting that the rubrospinal tract is an exclusively excitatory pathway in the rat. Synaptic surroundings o f synapses established by rubrospinal terminals. Rubrospinal terminals have not been found to be engaged in axoaxonic synaptic contacts. However, the dendrites postsynaptic to rubrospinal boutons have frequently been seen to receive synaptic contacts from axons of unknown origin in close proximity to rubrospinal axon terminals. Some of these axons contain GABA and/or glycine, while others do not show immunostaining for either inhibitory amino acid. These findings suggest that

Fig. 14. Electron micrographs showing a PHA-L-labelled rubrospinal terminal in consecutive sections immunostained by the immunogold procedure for GABA (b)and glycine (GLY) ( c ) , or counterstained with lead citrate (a). The dendrite (D) contacted by the rubrospinal terminal (white asterisk) was immunoreactive for CaBP and negative for both GABA and glycine. Note that this dendrite also received a synapse from an axon immunoreactive for glycine in close proximity to the rubrospinal bouton. Black asterisks identify the same axonal profiles, and open stars mark another dendrite in consecutive sections. Arrows point to synaptic contacts. Bars = 1 pm.

35

the rubrospinal synaptic input is under a presumably well-balanced inhibitory and excitatory postsynaptic spinal control, but presynaptic mechanisms do not play a significant role in the spinal modulation of rubrospinal signals in the rat.

Spinal neurons receiving monosynaptic rubrospinal input Volleys in the rubrospinal tract evoke short latency EPSPs in the spinal cord, suggesting monosynaptic connections between rubrospinal s o n s and various subgroups of spinal interneurons (for review, see Jankowska, '88). Most, if not all, of these interneurons are interposed in reflex pathways between various primary afferents and motoneurons, and represent both excitatory and inhibitory neurons located in laminae V-VII in the cat spinal cord (Baldissera et al., '71; Hongo et al., '72, '83, '89a,b; Brink et al., '83; Fleshman et al., '88). With CaBP immunocytochemistry, we have visualized a proportion of the excitatory population of these interneurons both in the cervical and lumbar segments of the rat spinal cord, providing direct morphological evidence that they indeed receive monosynaptic rubrospinal input. From the results presented in this paper, it can not be deduced how these neurons participate in spinal neuronal circuits. Physiological findings suggest that they may be first order interneurons of reflex pathways from group Ib or group I1 muscle afferents (Brink et al., '83; Hongo et al., '72, '83; Edgley and Jankowska, '87). They may also represent last order interneurons projecting directly to motoneurons and receiving mono- or polysynaptic connections from cutaneous afferents (Baldissera et al., '71; Fleshman et al., '88; Hongo et al., '89a,b). Although the functional identity of CaBP-immunoreactive neurons that receive synaptic contacts from rubrospinal terminals has yet to be determined, this study may initiate further efforts for a more detailed morphological characterization of spinal neuronal circuits involved in sensorimotor coordination.

CONCLUSIONS The rubrospinal tract in the rat is not a solely contralatera1 descending pathway as described in classical anatomical studies. Although most of the rubrospinal fibers are located in the contralateral dorsolateral funiculus, some of them descend in the ipsilateral white matter as caudal as the lumbar spinal cord. Rubrospinal terminals are distributed in laminae V-VII on both sides of the spinal cord. Terminals revealed on the ipsilateral side represent a significant proportion (10-28%) of the total number of boutons. Rubrospinal terminals are primarily engaged in synaptic contacts with dendrites, but occasionally they also make axosomatic synapses. Immunostaining for inhibitory amino acid neurotransmitters (GABA and glycine) suggests that the rubrospinal tract is an exclusively excitatory pathway establishing synaptic contacts with both excitatory and inhibitory spinal interneurons. The synaptic surroundings of rubrospinal terminals indicate that the rubrospinal synaptic input is under a strong postsynaptic spinal control, but presynaptic mechanisms presumably do not play a significant role in the spinal modulation of rubrospinal signals in the rat. The simultaneous visualization of rubrospinal terminals and CaBP-immunoreactive spinal neurons revealed that a

M. ANTAL ET AL.

36

population of excitatory spinal interneurons receiving monosynaptic input from rubrospinal axons is located in laminae V-VI at both the cervical and lumbar segments of the rat spinal cord.

ACKNOWLEDGMENTS The authors thank Dr. R.E. Burke for helpful discussions, Mrs. R. Azzam for excellent technical assistance, and N. Zhang and Ms. B. Riber for the preparation of the antisera to glycine. We are grateful to Dr. P. Somogyi, MRC Anatomical Neuropharmacology Unit, Department of Pharmacology, Oxford, United Kingdom, for the gift of antisera to GABA.

LITERATURE CITED Alstermark, B., A. Lundherg, and S. Sasaki i1984a) Integration in descending motor pathways controlling the forelimb in the cat. 11. Inhibitory pathways from higher motor centres and forelimb afferents to C3-C4 propriospinal neurones. Exp. Brain Res. 56:293-307. Alstermark, B., A. Lundherg, and S. Sasaki 1198413) Integration in descending motor pathways controlling the forelimb in the cat. 12. Interneurones which may mediate descending feed-forward inhibition and feedhack inhibition from the forelimb to C3-C4 propriospinal neurones. Exp. Brain Res. 56t308-322. Antal, M., T.F. Freund, and E. Polgar 11990) Calcium-binding proteins, parvalbumin- and calhindin-D28K-immunoreactive neurons in the rat spinal cord and dorsal root ganglia: A light and electron microscopic study. J. Comp. Neurol. 2 9 5 4 6 7 4 8 4 , Antal, M., E. Polgk, J. Chalmers, J.B. Minson, I. Llewellyn-Smith, C.W. Heizmann, and P. Somogyi (1991a) Different populations of parvalhumin- and calhindin-D28k-immunoreactive neurons contain GABA and accumulate 3H-D-aspartate in the dorsal horn of the rat spinal cord. J. Comp. Neurol. 314:114-124. Antal, M., G.N. Sholomenko, A.K. Moschovakis, J. Storm-Mathisen, C.W. Heizmann, and W. Hunziker (1991h) The termination pattern and postsynaptic targets of ruhro-spinal fibers in the rat spinal cord. Soc. Neurosci. Ahst. 17:470. Baldissera, F., G. Ten Bruggencate, and A. Lundherg (1971) Ruhrospinal monosynaptic connexion with last-order interneurones of polysynaptic reflex paths. Brain Res.27:390-392. Beitz, A.J., and L.J. Ecklund (1988) Colocalization of fixative-modified glutamate and glutaminase hut not GAD in rubrospinal neurons. J. Comp. Neurol. 274:265-279. Brink, E., E. Jankowska, D. McCrea, and B. Skoog (1983) Inhibitory interactions between interneurones in reflex pathways from group Ia and group Ih afferents in the cat. J. Physiol. (Lond.)343:361-373. Brown, L.T. (1974) Ruhrospinal projections in the rat. J. Comp. Neurol. 154:169-188. Burke, R.E., E. Jankowska, and G. Ten Bruggencate (1970) A comparison of peripheral and ruhrospinal synaptic input to slow and fast twitch motor units of triceps surae. J. Physiol. (Lond.) 207:709-732. Cliffer, K.D., and G.J. Giesler Jr. (1988) PHA-L can he transported anterogradely through fibers of passage. Brain Res. 458:185-191. Edgley, S.A., and E. Jankowska (1987) An interneuronal relay for group I and I1 muscle afferents in the midlumhar segments of the cat spinal cord. J. Physiol. (Lond.)389:647-674. Edwards, S.B. (1972) The ascending and descending projections of the red nucleus in the cat: An experimental study using an autoradiographic tracing method. Brain Res. 48:45-63. Fleshman, J.W., P. Rudomin, and R.E. Burke (1988) Supraspinal control ofa short-latency cutaneous pathway to hindlimb motoneurons. Exp. Brain Res. 69t449-459. Fujito, Y., T. Imai, and M. Aoki (1991) Monosynaptic excitation of motoneurons innervating forelimb muscles following stimulation of the red nucleus in cats. Neurosci. Lett. 127:137-140. Gerfen, C.R., and P.E. Sawchenko (1984) An anterograde neuroanatomical tracing method that shows the detailed morphology of neurons, their axons and terminals: Immunohistochemical localization of an axonally transported plant lectin, Phaseolus uulgaris leucoagglutinin (PHA-L). Brain Res. 290:219-238. Gerfen, C.R., and P.E. Sawchenko (1985) A method for anterograde axonal

tracing of chemically specific circuits in the central nervous system: Combined Phaseolus uulgaris-leucoagglutinin(PHA-L)tract tracing and immunohistochemistry Brain Res. 343: 144-150. Grove, E.A., V.B. Domesick, and W.H.H. Nauta (1986) Light microscopic evidence of striatal input to intrapallidal neurons of cholinergic cell group Ch4 in the rat: A study employing the anterograde tracer Phaseolus uulgaris leucoagglutinin (PHA-L).Brain Res. 367:379-384. Hancock, M.B. (1984)Visualization of peptide-immunoreactive processes on serotonin-immunoreactive cells using two-color immunoperoxidase staining. J. Histochem. Cytochem. 32:311-314. Harrison, P.J., and E. Jankowska (1985) Sources of input to interneurons mediating group I non-reciprocal inhibition of motoneurons in the cat. J. Physiol. (Lond.)361:379401. and A. Rustioni (1981) Descending projections from brainstem Hayes, N.L., and sensorimotor cortex to spinal enlargements in the rat. Single and double retrograde tracer study. Exp. Brain Res. 41:89-107. Hodgson, A.J., B. Penke, A. Erdei, I.W. Chuhb, and P . Somogyi (1985) Antiserum to y-aminobutyric acid. 1. Production and characterisation using a new model system. J. Histochem. Cytochem. 33929-239. Holstege, G. (1987) Anatomical evidence for an ipsilateral ruhrospinal pathway and for direct ruhrospinal projections to motoneurons in the cat. Neurosci. Lett. 74:269-274. Holstege, G., and H.G.J.M. Kuypers (1982) The anatomy of brain stem pathways to the spinal cord in cat. A labeled amino acid tracing study. In H.G.J.M. Kuypers and G.F. Martin (eds): Progress in Brain Research, Vol. 57, Anatomy of Descending Pathways to the Spinal Cord. Amsterdam: Elsevier, pp. 145-175. Holstege, G., and J. Tan (1988) Projections from the red nucleus and surrounding areas to the brainstem and spinal cord in the cat. An HRP and autoradiographical tracing study. Behav. Brain Res. 28:33-57. Holstege, G., B.F. Blok, and D.D. Ralston (1988)Anatomical evidence for red nucleus projections to motoneuronal cell groups in the spinal cord of the monkey. Neurosci. Lett. 9597-101. Hongo, T., E. Jankowska, and A. Lundherg (1969ai The rubrospinal tract. I. Effects on alpha-motoneurones innervating hindlimb muscles in cats. Exp. Brain Res. 7:344-364. Hongo, T., E. Jankowska, and A. Lundherg i1969b) The ruhrospinal tract. 11. Facilitation of interneuronal transmission in reflex paths to motoneurons. Exp. Brain Res. 7:365-391. Hongo, T., E. Jankowska, and A. Lundherg (1972) The ruhrospinal tract. IV. Effects on interneurones. Exp. Brain Res. 15:54-78. Hongo, T., E. Jankowska, T. Ohno, S. Sasaki, M. Yamashita, and K. Yoshida (1983)The same interneurons mediate inhibition of dorsal spinocerebellar tract cells and lumbar motoneurons in the cat. J. Physiol. (Lond.) 342:161-180. Hongo, T., S. Kitazawa, Y. Ohki, M. Sasaki, and M.-C. Xi i1989a) A physiological and morphological study of premotor interneurones in the cutaneous reflex pathways in cats. Brain Res. 505t163-166. Hongo, T., S. Kitazawa, Y. Ohki, and M.-C. Xi (1989h) Functional identification of last-order interneurones of skin reflex pathways in the cat forelimb segments. Brain Res. 505167-170. Huisman, A.M., H.G.J.M. Kuypers, and C.A. Verburgh (1981) Quantitative differences in collateralization of the descending spinal pathways from the red nucleus and other brain stem cell groups in rat as demonstrated with the multiple fluorescent retrograde tracer technique. Brain Res. 209:271-286. Hulthorn, H., M. Illert, and M. Santiani (1976) Convergence on interneurons mediating the reciprocal Ia inhibition of motoneurons. 111. Effects from supraspinal pathways. Acta Physiol. Scand. 96368-391. Illert, M., A. Lundherg, and R. Tanaka (1977) Integration in descending motor pathways controlling the forelimb in the cat. 3. Convergence on propriospinal neurones transmitting disynaptic excitation from the corticospinal tract and other descending tracts. Exp. Brain Res. 29:323346. Illert, M., A. Lundberg, Y. Padel, and R. Tanaka (1978) Integration in descending motor pathways controlling the forelimb in the cat. 5. Properties of and monosynaptic excitatory convergence on C3-C4 propriospinal neurons. Exp. Brain Res. 33:lOl-130. Jankowska, E. (1988) Target cells of ruhrospinal tract fibres within the lumbar spinal cord. Behav. Brain Res. 2 8 9 - 9 6 . Keller, A., E.L. White, and P.B. Cipolloni (1985) The identification of the thalamocortical axon terminal in the mouse Sml cortex using immunocytochemistry of anterogradely transported lectin (Phmeolus uulgaris leucoagglutinin). Brain Res. 343: 159-165. Massion, J. (1967) The mammalian red nucleus. Physiol Rev. 47:383-436. Molander, C., Q. Xu, and G. Grant (1984) The cytoarchitectonic organization

RUBROSPINAL TERMINALS IN THE RAT of the spinal cord in the rat. I . The lower thoracic and lumbosacral cord. J. Comp. Neurol. 230:133-141. Molander, C., Q. Xu, C. Rivero-Melian, and G. Grant (1989) Cytoarchitectonic organization of the spinal cord in the rat. 11. The cervical and upper thoracic cord. J. Comp. Neurol. 289:375-385. McCurdy, M.L., D.I. Hansma, J.C. Houk, and A.R. Gibson (1987) Selective projections from the cat red nucleus to digit motor neurons. J. Comp. Neurol. 265367-379. Pinol, M.R., U. Kapi, C.W. Heizmann, B. Vogel, J.-M. Sequier, W. Haas, and W. Hunziker (1990) Poly- and monoclonal antibodies against recombinant rat brain calbindin D-28k were produced to map its selective distribution in the central nervous system. J. Neurochem. 54:18271833. Ralston, D.D., A.M. Milroy, and G. Holstege (1988) Ultrastructural evidence for direct monosynaptic rubrospinal connections to motoneurons in Macaca rnulatta. Neurosci. Lett. 95r102-106. Robinson, F.R., J.C. Houk, and A.R. Gibson (1987) Limb specific connections of the cat magnocellular red nucleus. J. Comp. Neurol. 257:553-577. Shieh, J.Y., S.K. Leong, and W.C. Wong (1983) Origin of the rubrospinal tract in neonatal, developing, and mature rats. J. Comp. Neurol. 214.79-86. Somogyi, P., and A.J. Hodgson (1985) Antisera to y-aminobutyric acid. 111. Demonstration of GABA in Golgi-impregnated neurons and in conventional electron microscope sections of cat striate cortex. J. Histochem. Cytochem. 33:249-257. Sornogyi, P., and I. Soltesz (1986) Immunogold demonstration of GABA in synaptic terminals of intracellularly recorded, horseradish peroxidase-

37 filled basket cells and clutch cells in the cat’s visual cortex. Neuroscience 19:1051-1065. Somogyi, P., A.J. Hodgson, I.W. Chubb, B. Penke, and A. Erdei (1985) Antiserum to y-aminobutyric acid. 11. Immunocytocbemical application to the central nervous system. J. Histochem. Cytochem. 33240-248. Storm-Mathisen, J., and O.P. Ottersen (1990) Antibodies and fixatives for the immunocytochemical localization of glycine. in O.P. Ottersen and J. Storm-Mathisen (eds): Glycine Neurotransmission. New York: John Wiley & Sons, pp. 281-301. Walberg, F., and O.P. Ottersen (1992) Neuroactive amino acids in the area postrema: An immunocytochemical investigation in rat with some observations in cat and monkey (Macaca fasczcularis).Anat. Embryol. 185529-546. Waldron, H.A., and D.G. Gwyn (1969) Descending nerve tracts in the spinal cord of the rat. I . Fibers from the midbrain. J. Comp. Neurol. 137:143154. Wouterlood, F.G., J.G.J.M. Bol, and H.W.M. Steinbusch (1987) Double-label immunocytochemistry: Combination o f anterograde tracing with Phaseolus uulgaris-leucoagglutinin and enzyme immunocytochemistry o f the target neurons. J. Histochem. Cytochem. 35817-823, Wouterlood, F.G., and H.J. Groenewegen (1985) Neuroanatomical tracing by use o f Phaseolus uulgarzs-leucoagglutinin (PHA-L): Electron microscopy of PHA-L-filled neuronal somata, dendrites, axons and axon terminals. Brain Res. 326.188-191. Yamamoto, T., P.A. Carr, K.G. Braimbridge, and J.I. Nagy (1989) Parvalbumin- and calbindin D28k-imniunoreactive neurons in the superficial layers of the spinal cord dorsal horn of the rat. Brain Res. Bull. 23:493-508.

The termination pattern and postsynaptic targets of rubrospinal fibers in the rat spinal cord: a light and electron microscopic study.

The spinal course, termination pattern, and postsynaptic targets of the rubrospinal tract, which is known to contribute to the initiation and executio...
2MB Sizes 0 Downloads 0 Views