JOURNALOFNEUROPHYSIOL~GY
Vol. 63, No. 2, February
1990. Printed
in U.S.A.
Location and Morphology of Dorsal Spinocerebellar Tract Neurons That Receive Monosynaptic Afferent Input From Ankle Extensor Muscles in Cat Hindlimb B. WALMSLEY AND M. J. NICOL Neural Research Laboratory, School of Anatomy, University of New South Wales, Kensington, New South Wales 2033, and Experimental Neurology Unit, John Curtin School of iMedical Research, Australian National University, Australian Capital Territory 2601, Australia
SUMMARY
AND
CONCLUSIONS
1. The present experiments were carried out to investigate the morphology and somatotopic location of dorsal spinocerebellar tract (DSCT) neurons that receive monosynaptic group I afferent input from hindlimb ankle extensor muscles in the cat. 2. Intracellular recordings were obtained from DSCT neurons throughout the rostrocaudal extent of the L3 dorsal root entry zone of the spinal cord. DSCT neurons, physiologically identified as receiving monosynaptic group I input from the ankle extensor muscles, were injected with horseradish peroxidase (HRP) and subsequently reconstructed under the light microscope. 3. In contrast to previous HRP studies of DSCT neurons, these cells were found to have extremely extensive and complex dendritic trees, that often extend beyond the region of Clarke’s column. Dendrites were found to extend into the white matter of the dorsal columns, and/or into the spinal gray matter in a ventrolatera1 direction. The large dendritic spread of DSCT neurons was found to occupy up to 60% or more of the cross-sectional area of Clarke’s column. 4. DSCT neurons receiving monosynaptic group I input from the single functional group of ankle extensor muscles were not found to be confined within a specific transverse region of Clarke’s column, in contrast to a previous proposal. Instead, these cells could be found throughout Clarke’s column. 5. The present results demonstrate that DSCT neurons, physiologically identified as receiving group I muscle afferent input, exhibit dendritic trees that are considerably more extensive and morphologically complex than indicated by previous studies. In addition, the present results do not support a previous proposal of a strict somatotopic arrangement for DSCT neurons and their dendritic envelopes within Clarke’s column in the transverse plane. INTRODUCTION
Three major types of cells have been found within Clarke’s column, the largest of which (the Type C cells) are the primary cells of origin of the dorsal spinocerebellar tract (DSCT) (Loewy 1970; see review by Mann 1973). Golgi and Nissl staining have shown these DSCT cells to be large bodied, with a complex dendritic tree oriented primarily in a rostrocaudal direction (Boehme 1968; Loewy 1970). The dendrites of these cells have been described as either being totally confined to the column (Rethelyi 1968) or to have a few branches that extend beyond the column boundary (Boehme 1968; Clarke 1859; Loewy 1970). Kuno et al. ( 1973) proposed that cells in Clarke’s column 286
0022-3077/90
$1 SO Copyright
might obey some form of somatotopy after finding DSCT neurons that received input from posterior biceps/semitendinosus in more lateral locations than those that received input from triceps surae muscles. Hongo and co-workers (Hong0 et al. 1982, 1985) have proposed that DSCT neurons in Clarke’s column are somatotopically organized with respect to the muscle providing their primary afferent input. Hongo (1985) has further suggested that both the cell body and the dendritic tree of a DSCT neuron are confined to the “appropriate” somatotopic location. This proposal has since been correlated with observations that identified group I muscle afferents terminate in Clarke’s column in a strictly musculotopic arrangement (Hong0 et al. 1987; see also Mense and Craig 1988). Two studies have utilized intracellular horseradish peroxidase (HRP) labeling techniques to visualize physiologically identified DSCT neurons, although the results of these studies are largely conflicting. Randic et al. (198 1) labeled cells that received identified afferent input from the hindlimb. The reconstructions presented by Randic et al. (198 1) showed the cells to be far less morphologically complex than suggested by previous Golgi studies. In contrast, Houchin et al. (1983) showed DSCT neurons labeled with HRP to be morphologically complex, with extensively branched dendritic trees confined almost entirely to the boundaries of Clarke’s column. The cells reconstructed by Houchin et al. (1983) were, however, identified only on the basis of sciatic nerve stimulation. In view of these limited and conflicting results, there is a need for further information on the morphology of DSCT neurons in Clarke’s column that have been positively identified physiologically. The experiments described in this paper were carried out to examine in detail DSCT neurons that receive monosynaptic group I input from a specific functional muscle group, the hindlimb ankle extensor muscles in the cat. The cells were first physiologically identified and subsequently labeled intracellularly with HRP to determine the morphology, location, and possible somatotopic arrangement of identified DSCT neurons in Clarke’s column. METHODS
Experiments were performed on 23 young adult cats weighing 1S-3.5 kg. Cats were anesthetized with pentobarbital sodium (35 mg/kg ip) and maintained on supplementary doses (5 mg/kg iv)
0 1990 The American
Physiological
Society
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND MORPHOLOGY
OF DSCT NEURONS
281
FIG. 1. A: an HRP-labeled DSCT neuron that received input from plantaris muscle shown in transverse section. B: higher magnification photomicrograph of the neuron in A. C: full reconstruction in the transverse plane of the HRP-labeled DSCT neuron in Clarke’s column shown in A and B.
DORSAL COLUMN
CENTRAL CANAL
0.1 \
’
mm ’
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
288
B. WALMSLEY
as required. Mean arterial pressure and end-tidal CO2 were monitored. The ankle extensor muscles medial gastrocnemius (MG), lateral gastrocnemius (LG), soleus (Sol), and plantaris (PI) were exposed in the left hindlimb and their nerves cleared in continuity for part of their length. The tendon of each muscle was cut at its insertion to allow the muscles to be individually stretched. The cat was fixed in a rigid animal frame, and a conventional laminectomy was performed from L2 to L,. The exposed hindlimb muscles and spinal cord were covered with pools of mineral oil maintained at 35-37OC by infrared heating. Bipolar stimulating electrodes were placed on the sciatic nerve and its branches to the exposed muscles. Intracellular recording electrodes filled with 10% HRP in 2 M K-methyl sulphate were driven into Clarke’s column at the L3-L4 spinal levels. Antidromic identification of the DSCT neurons was achieved by stimulation of the dissected dorsolateral fasciculus (DLF) at the C2 level (Houchin et al. 1983; Tracey and Walmsley 1984). Synaptic input to the cells was identified as monosynaptic on the basis of group I latency and high-frequency following of the synaptic potential in response to nerve stimulation (Tracey and Walmsley 1984). HRP was delivered through the electrodes by pressure injection. This method afforded two major advantages over ionophoresis. First, the time required to fill a neuron adequately was much less: m-20-60 s for pressure injection, compared with lo-20 min
AND
M.
J. NICOL
for ionophoresis. Second, a high concentration of electrolyte could be used in the electrode, because the amount of HRP delivered was not dependent on this concentration. A simple method for pressure injection was used in this study: a 5-ml or lo-ml disposable syringe was used to deliver pressure to the electrode holder via a flexible plastic tube. A small hole (-3 mm diam) was drilled into the syringe, which equilibrated the electrode to atmospheric pressure in the release position. A spring inserted into the syringe ensured that the plunger returned to the extended position with the release of external pressure. Pressure was applied by hand in pulses, with a return to atmospheric pressure between each pulse. The amount of pressure applied was gauged by the volume of compression. During application of pressure, it was useful to observe the resistance of the electrode ( 1 ms/O. 1 nA current pulse through the electrode), because this resistance generally decreased during extrusion of electrolyte from the electrode tip. After HRP injection into a cell, the animal was perfused with 2.5% glutaraldehyde in phosphate buffer (pH 7.2). Spinal cord segments L3 and L4 were removed and postfixed for - 10 h in the same fixative. Either transverse or parasagittal sections ( 100 pm) of these segments were cut with the use of a vibratome and reacted for HRP with the use of a cobalt-enhanced diaminobenzidine technique (Adams 1977, 198 1). The sections were then examined and photographed, and the DSCT neurons were traced out by the use of a drawing tube attachment. In each case, the boundary of Clarke’s column was determined from 20 sequential counter-
MG,LG, SOL,PL
FIG. 2. Reconstructions in the transverse plane of 5 DSCT neurons in Clarke’s column. Each of these cells received input from one or more of the ankle extensor muscles, as indicated. Dotted outline indicates the boundary of Clarke’s column in each cast.
DORSAL
LATERAL
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND MORPHOLOGY
stainedsections centeredaround the cell body of the HRP-labeled DSCT neuron. RESULTS
Intracellular recordings were obtained in 40 DSCT neurons that were positively identified as receiving monosynaptic group I afferent input from one or more of the ankle extensor muscles. Figure 1 shows a DSCT neuron that received group I afferent input from PI muscle only. Figure 1A shows a photomicrograph of the cell as it appeared in a single loo-pm transverse section of the spinal cord. The cell is
OF DSCT NEURONS
289
situated in the ventromedial region of Clarke’s column, close to the central canal. The same cell is shown at higher magnification in Fig. 1B, which also reveals other large counterstained cell bodies in Clarke’s column. A full reconstruction of this cell is shown in Fig. 1C. The soma and dendrites of the cell were heavily labeled with HRP reaction product, as was the axon, which could be followed over 1.5 mm as it travelled ventrally and then laterally, finally to ascend in the DLF. No axon collaterals were observed. The cell soma measured 48 X 53 pm in this plane, and the dendrites were confined almost entirely to the column. The dendrites appear complex, spreading laterally and dorsally over one-half the column in cross-section. A few of the dendrites are shown to extend dorsally
A
FIG. 3. Reconstruction in the transverse plane of the axonal trajectory and dendritic envelope (w) of 4 DSCT neurons in Clarke’s column. These neurons all received monosynaptic group I afferent input from ankle extensor muscles. Dotted outlines indicate the boundary of Clarke’s column in each case. (Cells in Band D are those shown in Figs. 1 and 2C, respectively.)
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
290
B. WALMSLEY
AND
out of the column, one reaching almost 200 pm into the white matter of the dorsal columns. Each of the cells shown reconstructed in Fig. 2, A-E, received monosynaptic group I afferent input from one or more of the ankle extensor muscles (as indicated). As with the DSCT neuron illustrated in Fig. 1, each cell exhibits corn plex dendrit ic bran ches and a.n extensive dendritic tree, both within, and, in most cases, outside, the boundary of Clarke’s column. Although all of these cells received input from the same functional muscle group, they are located throughout the column in the transverse plane. Figure 2A shows a DSCT neuron that received afferent input from PI only. This cell, one of the smallest examined, has a soma size 35 X 45 pm and is quite compact, its dendritic spread confined almost exclusively to the column. The DSCT neuron shown in Fig. 2B received input from both LG and Sol. In this plane, the cell body measures 65 X 53 pm and is situated centrally in the column. The most striking feature of this cell is the extent to which the dendrites leave the column, up to 620 pm in the dorsal direction and 450 pm ventrolaterally. The dendrites of the cell shown in Fig. 2C also extend well beyond the column,
/ FIG. 4.
extensor
Sagittal reconstructions muscle input.
of 5 DSCT
neurons
in Clarke’s
M.
J. NICOL
up to 340 pm laterally and ventrolaterally. This cell, which received input from PI only, exhibits a soma size of 55 X 50 pm in this plane and occupies a lateral position within Clarke’s column. Both neurons reconstructed in Fig. 2, D and E, received afferent input from all four of the ankle extensor muscles. The cell shown in Fig. 20 is dorsomedial, and the cell shown in Fig. 2E is in the ventrolateral region of the column. The dendrites of the cell represented in Fig. 20 are quite restricted, with only a few dendrites extending beyond the column boundary in the dorsal direction. In contrast, the cell shown in Fig. 2E has an extensive dendritic spread, reaching well beyond the column boundary ventrolaterally and covering ~60% of the column in crosssection. The extent of the dendritic spread of the DSCT neurons receiving group I afferent input from ankle extensor muscles is again illustrated in Fig. 3 (w). Each reconstruction, A-D, shows the dendritic envelope and axonal trajectory of an individual DSCT neuron. The dendritic envelope can be seen to be either confined to the column (Fig. 3, A and C) or to extend beyond it in a limited (Fig. 3B) or extensive
100 ym column
that
received
CAUDA monosynaptic
group
I ankle
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND
MORPHOLOGY
(Fig. 30) manner. In all cases, the dendrites cover a substantial proportion of Clarke’s column itself. The axonal trajectory, illustrated in Fig. 3, A-D, is generally ventral or ventolateral on leaving the soma and then lateral to the DLF, at which point the axon changes direction abruptly and ascends towards the cerebellum. No axon collaterals were observed along any of the axons examined. In the sagittal plane, DSCT neurons were again shown to be large cells, with extensively branched, complex dendritic trees oriented primarily in the rostrocaudal direction. Figure 4 shows sagittal reconstructions of five DSCT neurons, all of which received monosynaptic afferent input from one or more of the ankle extensor muscles. In this plane, the cell bodies of the labeled neurons ranged from 32 X 67 pm to 39 X 119 pm (average, 43 X 94 pm). The dendrites extended up to 3 mm rostrocaudally, compared with a much more restricted range in the transverse plane (see Fig. 2). The number of trunk dendrites ranged from six to nine
OF
DSCT
Clarke’s
NEURONS
column
_-----------------------0 .--------------------------------------
291
1.0
\ 9
------------=. ..-
,r;“-;ri
---------
8 -----------------
mm I--w -we
B
Clarke’s
Column
1. Summarvw ofthe monosynuptic muscle inputs received . by 40 DSCT neurons
TABLE
Cell Number
PL
SOL
MG
1 2 3
0 0 0
0 0 l
0
0
4
0
0
0
l
5 6
0 0
0 0
0
0
0
0
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
0
0
0 0 0
0
0 0
lateral
LG FIG. 5.
A: schematic reconstruction in the sagittal plane of the L3 dorsal root entry zone, with the recorded position of 28 DSCT neurons receiving ankle extensor muscle input marked as individual filled circles. B: schematic reconstruction in the transverse plane showing the position of 12 DSCT neurons in Clarke’s column. These neurons all received group I afferent input from the ankle extensor muscles.
0 0
l 0
0 0 0 0 0 0
0
0 0 0 0 0 0 0 l
DSCT, dorsal spinocerebellar dial gastrocnemius; LG, lateral
0 0 0 0 0 l
0 0 0 0 0 0 0 0 0 0 0 tract; PI, plantaris; gastrocnemius.
Sol, soleus;
MG,
me-
(average, 6.4). [As noted by Houchin et al. (1983), it was often difficult to accurately determine the number of trunk dendrites, as many branched almost immediately after leaving the soma. The number of dendrites counted was, therefore, the most conservative estimate.] The diameters of the trunk dendrites were measured just beyond the initial tapered region and found to have an average diameter of 8.8_pm. Many small branchlets were seen arising from the dendrites of these cells. Beading along some of the dendrites was observed [in accordance with Loewy (1970) and Houchin et al. ( 1983)], although dendritic spines were very rare, and the dendrites were relatively smooth in appearance. DSCT neurons recorded during these experiments often received group I monosynaptic afferent input from more than one of the ankle extensor muscles. A summary of the muscle inputs received by 40 DSCT neurons is presented in Table 1. Five of these cells were shown to receive input from all ankle extensor muscles (LG, MG, Sol, and Pl). Two neurons received input from three of the muscles, 8 received input from two of the muscles, and the remaining 25 cells received input from only one of the muscles. Of these, two received input from LG, eight from MG, seven from Sol, and eight from Pl. Figure 5A shows a schematic reconstruction of the L3 dorsal root entry zone viewed in sagittal section. The location of 28 neurons receiving group I ankle extensor input is plotted, showing that these cells could be located along the
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
292
B. WALMSLEY
entire length of the L3 root, from rostra1 L3 past the caudal end of L3. (The apparent clustering of the cells toward the caudal end of L3 is not thought to be significant, because electrode tracking was often concentrated in this region.) Figure 5B shows the location in cross-section of the cell bodies of 12 DSCT neurons that received monosynaptic group I input from one or more of the ankle extensor muscles. It should be noted that this schematic diagram illustrates the relative location of cells within Clarke’s column, because the exact size and shape of Clarke’s column varies among animals (see Fig. 2). As clearly shown in Fig. 2, DSCT cells receiving ankle extensor input can be found throughout Clarke’s column and are not strictly confined to a particular region.
Golgi studies originally suggested that the cells in Clarke’s column giving rise to the DSCT are large, with profuse dendritic trees oriented in a primarily rostrocaudal direction (Boehme 1968; Loewy 1970). Subsequent attempts to label DSCT neurons intracellularly with HRP have revealed conflicting results (Houchin et al. 1983; Randic et al. 198 1). The present results have demonstrated that DSCT neurons receiving group I muscle afferent input have large cell bodies and extremely complex dendritic trees that may extend 3 mm in the rostrocaudal direction. In cross-section, the profusely tangled appearance of the dendrites is quite striking, as is the extent of their vast spread within Clarke’s column. In addition, the dendrites of these cells are often found to extend well beyond Clarke’s column in the lateral, ventral, and dorsal directions, in contrast to the results of Houchin et al. (1983) and Hongo ( 1985). A particularly interesting observation is that dendrites are found to reach some distance into the white matter of the dorsal columns. White-matter dendrites have also been observed arising from cervical motoneurons by Rose and Richmond ( 198 1). Electron microscopy showed these dendrites to receive synaptic contacts within the white matter (Rose and Richmond 198 1). Kuno et al. (1973) proposed that the somata of DSCT neurons in Clarke’s column were somatotopically organized, showing that cells that received input from gastrocnemius/soleus were more medially located than cells that receive biceps femoris input. Hongo and co-workers (Hong0 et al. 1982, 1985) returned to the question of somatotopic location of cell bodies of DSCT neurons, concluding that, in the transverse plane, the cells are arranged with respect to the proximity of the muscle supplying their afferent input. In this scheme, DSCT neurons that receive toe-muscle input are located dorsomedially in the column, those receiving shank input in the intermediate region, and those receiving input from thigh muscles in the ventrolatera1 zone. Hongo (1985) has proposed that these regions are quite definite (although they exhibit some overlap), with both the cell body and the dendrites confined to their respective functional regions. DSCT neurons labeled in the present study correspond
AND
M.
J. NICOL
with those receiving input from the shank muscles as defined by Hongo (1985). However, the present results indicate that these cells can be located throughout Clarke’s column in the transverse plane. This result does not, of course, rule out a general trend for these cells to be located in a particular region of Clarke’s column. Nevertheless, DSCT neurons receiving ankle extensor-muscle input may be located throughout Clarke’s column, including the central, dorsomedial, and ventrolateral regions. Furthermore, the dendritic trees of these neurons, rather than being confined to a particular functional zone, may occupy up to 60% or more of the transverse area of Clarke’s column and often extend well beyond the column boundary. The results presented here do not, therefore, provide convincing evidence for a strict transverse musculotopic arrangement of DSCT neurons and their dendrites within Clarke’s column. DSCT neurons have commonly been regarded in the role of relay neurons, conveying signals from restricted hindlimb receptor groups to the cerebellum. However, Osborn and Poppele (1988) have recently indicated that although DSCT neurons may receive a limited monosynaptic input, they receive significant polysynaptic influences from receptors throughout a widespread region of the hindlimb. They propose that DSCT neurons should be regarded in the role of integration of afferent information, rather than as simple relay neurons (Osborn and Poppele 1988). The present demonstration of complex dendritic trees that extend over a wide area both within and outside Clarke’s column supports the suggestion that these cells might play a much more complex role than previously imagined. We arc grateful to Dr. R. E. W. Fyffc for his comments on the manuscript. We also thank K. Howarth for participation in several of the experiments. This work was supported by a grant from the National Health and Medical Research Council of Australia. Address for reprint requests: B. Walmsley, Neural Research Laboratory, School of Anatomy, University of New South Wales, P. 0. Box I, Kensington, N. S. W. 2033, Australia. Received
12 June
1989; accepted
in final form
25 September
1989.
REFERENCES J. C. Technical considerations on the use of horseradish peroxidase as a marker. Ncz4rosc1icr7(‘I’2: 14 I- 145, 1977. ADAMS, J. C. Heavy metal intensification of DAB based HRP reaction product. J. lJistochcm. Cytochon. 29: 775, 198 1. BOEHME, C. C. The neural structure of Clarke’s nucleus of the spinal cord. J. Camp. Nvzrro~. 132: 445-462, 1968. CLARKE, J. C. Further researches on the grey substance of the spinal cord. Phi/ox Trans. R. Sot. Lond. 149: 437-467, 1859. HONGO, T. Functional and morphological organization of DSCT cells in Clarke’s column. In Development, Orgunization and Processing in Sensory Puthwuys, edited by M. Rowe and W. D. Willis. New York: Liss, 1985, p. 149-156. HONGO, T., KUDO, N., SASAKI, S., YAMASHITA, M., YOSHIDA, K., ISHIZUKA, N., AND MANNEN, H. Somatotopic organization of Clarke’s column in the cat (Abstract). N~wr-oscicvx~~~ Lctt. 9, Suppl.: S107, 1982. HONGO, T., Kur>o, N., SASAKI, S., YAMASHITA, M., YOSHIDA, K., ISHIADAMS,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND
MORPHOLOGY
N., AND MANNEN, H. Trajectory of group Ia and Ib fibers from the hind-limb muscles at the L3 and L4 segments of the spinal cord of the cat. J. Camp. Neurol. 262: 159-194, 1987. HOIJCHIN, J., MAXWELI,, D. J., FYFFE, R. E. W., AND BROWN, A. G. Light and electron microscopy of dorsal spinocerebellar tract neurons in the cat: an intracellular horseradish peroxidase study. Q. J. Exp. Physid. 68: 7 19-732, 1983. KUNO, M., MUNOZ-MARTINEZ, E. J., AND RANDIC, M. Sensory inputs to neurons of Clarke’s column from muscle, cutaneous and joint receptors. J. Physiol. Lord. 228: 327-342, 1973. LOEWY, A. D. A study of neuronal types in Clarke’s column of the adult 1970. cat. J. Comp. N~wrol. 139: 53-80, MANN, M. D. Clarke’s column and the dorsal spinocerebellar tract: a review. Brain Bchav. Evol. 7: 34-83, 1973. MENSE, S. AND CRAIG, A. D. Spinal and supraspinal terminations of ZUKA,
OF primary
DSCT afferent
NEURONS
293
fibres from
the gastrocnemius-soleus muscle in the cat. 1988. OSBORN, C. E. AND POPPELE, R. E. The extent of polysynaptic responses in the dorsal spinocerebellar tract to stimulation of group I afferent fibers in gastrocnemius-soleus. J. Neuroxi. 8: 3 16-3 19, 1988. RANDIC, M., MILETIC, V., AND LOEWY, A. D. A morphological study of cat dorsal spinocerebellar tract neurons after intracellular injection of horseradish peroxidase. J. Camp. Ncwrd. 198: 453-466, 198 1. RETHELYI, M. The Golgi architecture of Clarke’s column. Acta Morphol. Hung. 16: 31 l-330, 1968. ROSE, P. K. AND RICHMOND, F. J. White matter dendrites in the upper cervical spinal cord of the adult cat: a light and electron microscopic study. J. Camp. Neurd. 199: 19 l-203, 198 1. TRACEY, D. J. AND WALMSLEY, B. Synaptic input from identified muscle afferents to neurons of the dorsal spinocerebellar tract in the cat. J. Physiol. Lord. 350: 599-6 14, 1984.
Neuroscience 26: 1023-1035,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
Vol. 63, No. 2, February
1990. Printed
in U.S.A.
Location and Morphology of Dorsal Spinocerebellar Tract Neurons That Receive Monosynaptic Afferent Input From Ankle Extensor Muscles in Cat Hindlimb B. WALMSLEY AND M. J. NICOL Neural Research Laboratory, School of Anatomy, University of New South Wales, Kensington, New South Wales 2033, and Experimental Neurology Unit, John Curtin School of iMedical Research, Australian National University, Australian Capital Territory 2601, Australia
SUMMARY
AND
CONCLUSIONS
1. The present experiments were carried out to investigate the morphology and somatotopic location of dorsal spinocerebellar tract (DSCT) neurons that receive monosynaptic group I afferent input from hindlimb ankle extensor muscles in the cat. 2. Intracellular recordings were obtained from DSCT neurons throughout the rostrocaudal extent of the L3 dorsal root entry zone of the spinal cord. DSCT neurons, physiologically identified as receiving monosynaptic group I input from the ankle extensor muscles, were injected with horseradish peroxidase (HRP) and subsequently reconstructed under the light microscope. 3. In contrast to previous HRP studies of DSCT neurons, these cells were found to have extremely extensive and complex dendritic trees, that often extend beyond the region of Clarke’s column. Dendrites were found to extend into the white matter of the dorsal columns, and/or into the spinal gray matter in a ventrolatera1 direction. The large dendritic spread of DSCT neurons was found to occupy up to 60% or more of the cross-sectional area of Clarke’s column. 4. DSCT neurons receiving monosynaptic group I input from the single functional group of ankle extensor muscles were not found to be confined within a specific transverse region of Clarke’s column, in contrast to a previous proposal. Instead, these cells could be found throughout Clarke’s column. 5. The present results demonstrate that DSCT neurons, physiologically identified as receiving group I muscle afferent input, exhibit dendritic trees that are considerably more extensive and morphologically complex than indicated by previous studies. In addition, the present results do not support a previous proposal of a strict somatotopic arrangement for DSCT neurons and their dendritic envelopes within Clarke’s column in the transverse plane. INTRODUCTION
Three major types of cells have been found within Clarke’s column, the largest of which (the Type C cells) are the primary cells of origin of the dorsal spinocerebellar tract (DSCT) (Loewy 1970; see review by Mann 1973). Golgi and Nissl staining have shown these DSCT cells to be large bodied, with a complex dendritic tree oriented primarily in a rostrocaudal direction (Boehme 1968; Loewy 1970). The dendrites of these cells have been described as either being totally confined to the column (Rethelyi 1968) or to have a few branches that extend beyond the column boundary (Boehme 1968; Clarke 1859; Loewy 1970). Kuno et al. ( 1973) proposed that cells in Clarke’s column 286
0022-3077/90
$1 SO Copyright
might obey some form of somatotopy after finding DSCT neurons that received input from posterior biceps/semitendinosus in more lateral locations than those that received input from triceps surae muscles. Hongo and co-workers (Hong0 et al. 1982, 1985) have proposed that DSCT neurons in Clarke’s column are somatotopically organized with respect to the muscle providing their primary afferent input. Hongo (1985) has further suggested that both the cell body and the dendritic tree of a DSCT neuron are confined to the “appropriate” somatotopic location. This proposal has since been correlated with observations that identified group I muscle afferents terminate in Clarke’s column in a strictly musculotopic arrangement (Hong0 et al. 1987; see also Mense and Craig 1988). Two studies have utilized intracellular horseradish peroxidase (HRP) labeling techniques to visualize physiologically identified DSCT neurons, although the results of these studies are largely conflicting. Randic et al. (198 1) labeled cells that received identified afferent input from the hindlimb. The reconstructions presented by Randic et al. (198 1) showed the cells to be far less morphologically complex than suggested by previous Golgi studies. In contrast, Houchin et al. (1983) showed DSCT neurons labeled with HRP to be morphologically complex, with extensively branched dendritic trees confined almost entirely to the boundaries of Clarke’s column. The cells reconstructed by Houchin et al. (1983) were, however, identified only on the basis of sciatic nerve stimulation. In view of these limited and conflicting results, there is a need for further information on the morphology of DSCT neurons in Clarke’s column that have been positively identified physiologically. The experiments described in this paper were carried out to examine in detail DSCT neurons that receive monosynaptic group I input from a specific functional muscle group, the hindlimb ankle extensor muscles in the cat. The cells were first physiologically identified and subsequently labeled intracellularly with HRP to determine the morphology, location, and possible somatotopic arrangement of identified DSCT neurons in Clarke’s column. METHODS
Experiments were performed on 23 young adult cats weighing 1S-3.5 kg. Cats were anesthetized with pentobarbital sodium (35 mg/kg ip) and maintained on supplementary doses (5 mg/kg iv)
0 1990 The American
Physiological
Society
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND MORPHOLOGY
OF DSCT NEURONS
281
FIG. 1. A: an HRP-labeled DSCT neuron that received input from plantaris muscle shown in transverse section. B: higher magnification photomicrograph of the neuron in A. C: full reconstruction in the transverse plane of the HRP-labeled DSCT neuron in Clarke’s column shown in A and B.
DORSAL COLUMN
CENTRAL CANAL
0.1 \
’
mm ’
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
288
B. WALMSLEY
as required. Mean arterial pressure and end-tidal CO2 were monitored. The ankle extensor muscles medial gastrocnemius (MG), lateral gastrocnemius (LG), soleus (Sol), and plantaris (PI) were exposed in the left hindlimb and their nerves cleared in continuity for part of their length. The tendon of each muscle was cut at its insertion to allow the muscles to be individually stretched. The cat was fixed in a rigid animal frame, and a conventional laminectomy was performed from L2 to L,. The exposed hindlimb muscles and spinal cord were covered with pools of mineral oil maintained at 35-37OC by infrared heating. Bipolar stimulating electrodes were placed on the sciatic nerve and its branches to the exposed muscles. Intracellular recording electrodes filled with 10% HRP in 2 M K-methyl sulphate were driven into Clarke’s column at the L3-L4 spinal levels. Antidromic identification of the DSCT neurons was achieved by stimulation of the dissected dorsolateral fasciculus (DLF) at the C2 level (Houchin et al. 1983; Tracey and Walmsley 1984). Synaptic input to the cells was identified as monosynaptic on the basis of group I latency and high-frequency following of the synaptic potential in response to nerve stimulation (Tracey and Walmsley 1984). HRP was delivered through the electrodes by pressure injection. This method afforded two major advantages over ionophoresis. First, the time required to fill a neuron adequately was much less: m-20-60 s for pressure injection, compared with lo-20 min
AND
M.
J. NICOL
for ionophoresis. Second, a high concentration of electrolyte could be used in the electrode, because the amount of HRP delivered was not dependent on this concentration. A simple method for pressure injection was used in this study: a 5-ml or lo-ml disposable syringe was used to deliver pressure to the electrode holder via a flexible plastic tube. A small hole (-3 mm diam) was drilled into the syringe, which equilibrated the electrode to atmospheric pressure in the release position. A spring inserted into the syringe ensured that the plunger returned to the extended position with the release of external pressure. Pressure was applied by hand in pulses, with a return to atmospheric pressure between each pulse. The amount of pressure applied was gauged by the volume of compression. During application of pressure, it was useful to observe the resistance of the electrode ( 1 ms/O. 1 nA current pulse through the electrode), because this resistance generally decreased during extrusion of electrolyte from the electrode tip. After HRP injection into a cell, the animal was perfused with 2.5% glutaraldehyde in phosphate buffer (pH 7.2). Spinal cord segments L3 and L4 were removed and postfixed for - 10 h in the same fixative. Either transverse or parasagittal sections ( 100 pm) of these segments were cut with the use of a vibratome and reacted for HRP with the use of a cobalt-enhanced diaminobenzidine technique (Adams 1977, 198 1). The sections were then examined and photographed, and the DSCT neurons were traced out by the use of a drawing tube attachment. In each case, the boundary of Clarke’s column was determined from 20 sequential counter-
MG,LG, SOL,PL
FIG. 2. Reconstructions in the transverse plane of 5 DSCT neurons in Clarke’s column. Each of these cells received input from one or more of the ankle extensor muscles, as indicated. Dotted outline indicates the boundary of Clarke’s column in each cast.
DORSAL
LATERAL
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND MORPHOLOGY
stainedsections centeredaround the cell body of the HRP-labeled DSCT neuron. RESULTS
Intracellular recordings were obtained in 40 DSCT neurons that were positively identified as receiving monosynaptic group I afferent input from one or more of the ankle extensor muscles. Figure 1 shows a DSCT neuron that received group I afferent input from PI muscle only. Figure 1A shows a photomicrograph of the cell as it appeared in a single loo-pm transverse section of the spinal cord. The cell is
OF DSCT NEURONS
289
situated in the ventromedial region of Clarke’s column, close to the central canal. The same cell is shown at higher magnification in Fig. 1B, which also reveals other large counterstained cell bodies in Clarke’s column. A full reconstruction of this cell is shown in Fig. 1C. The soma and dendrites of the cell were heavily labeled with HRP reaction product, as was the axon, which could be followed over 1.5 mm as it travelled ventrally and then laterally, finally to ascend in the DLF. No axon collaterals were observed. The cell soma measured 48 X 53 pm in this plane, and the dendrites were confined almost entirely to the column. The dendrites appear complex, spreading laterally and dorsally over one-half the column in cross-section. A few of the dendrites are shown to extend dorsally
A
FIG. 3. Reconstruction in the transverse plane of the axonal trajectory and dendritic envelope (w) of 4 DSCT neurons in Clarke’s column. These neurons all received monosynaptic group I afferent input from ankle extensor muscles. Dotted outlines indicate the boundary of Clarke’s column in each case. (Cells in Band D are those shown in Figs. 1 and 2C, respectively.)
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
290
B. WALMSLEY
AND
out of the column, one reaching almost 200 pm into the white matter of the dorsal columns. Each of the cells shown reconstructed in Fig. 2, A-E, received monosynaptic group I afferent input from one or more of the ankle extensor muscles (as indicated). As with the DSCT neuron illustrated in Fig. 1, each cell exhibits corn plex dendrit ic bran ches and a.n extensive dendritic tree, both within, and, in most cases, outside, the boundary of Clarke’s column. Although all of these cells received input from the same functional muscle group, they are located throughout the column in the transverse plane. Figure 2A shows a DSCT neuron that received afferent input from PI only. This cell, one of the smallest examined, has a soma size 35 X 45 pm and is quite compact, its dendritic spread confined almost exclusively to the column. The DSCT neuron shown in Fig. 2B received input from both LG and Sol. In this plane, the cell body measures 65 X 53 pm and is situated centrally in the column. The most striking feature of this cell is the extent to which the dendrites leave the column, up to 620 pm in the dorsal direction and 450 pm ventrolaterally. The dendrites of the cell shown in Fig. 2C also extend well beyond the column,
/ FIG. 4.
extensor
Sagittal reconstructions muscle input.
of 5 DSCT
neurons
in Clarke’s
M.
J. NICOL
up to 340 pm laterally and ventrolaterally. This cell, which received input from PI only, exhibits a soma size of 55 X 50 pm in this plane and occupies a lateral position within Clarke’s column. Both neurons reconstructed in Fig. 2, D and E, received afferent input from all four of the ankle extensor muscles. The cell shown in Fig. 20 is dorsomedial, and the cell shown in Fig. 2E is in the ventrolateral region of the column. The dendrites of the cell represented in Fig. 20 are quite restricted, with only a few dendrites extending beyond the column boundary in the dorsal direction. In contrast, the cell shown in Fig. 2E has an extensive dendritic spread, reaching well beyond the column boundary ventrolaterally and covering ~60% of the column in crosssection. The extent of the dendritic spread of the DSCT neurons receiving group I afferent input from ankle extensor muscles is again illustrated in Fig. 3 (w). Each reconstruction, A-D, shows the dendritic envelope and axonal trajectory of an individual DSCT neuron. The dendritic envelope can be seen to be either confined to the column (Fig. 3, A and C) or to extend beyond it in a limited (Fig. 3B) or extensive
100 ym column
that
received
CAUDA monosynaptic
group
I ankle
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND
MORPHOLOGY
(Fig. 30) manner. In all cases, the dendrites cover a substantial proportion of Clarke’s column itself. The axonal trajectory, illustrated in Fig. 3, A-D, is generally ventral or ventolateral on leaving the soma and then lateral to the DLF, at which point the axon changes direction abruptly and ascends towards the cerebellum. No axon collaterals were observed along any of the axons examined. In the sagittal plane, DSCT neurons were again shown to be large cells, with extensively branched, complex dendritic trees oriented primarily in the rostrocaudal direction. Figure 4 shows sagittal reconstructions of five DSCT neurons, all of which received monosynaptic afferent input from one or more of the ankle extensor muscles. In this plane, the cell bodies of the labeled neurons ranged from 32 X 67 pm to 39 X 119 pm (average, 43 X 94 pm). The dendrites extended up to 3 mm rostrocaudally, compared with a much more restricted range in the transverse plane (see Fig. 2). The number of trunk dendrites ranged from six to nine
OF
DSCT
Clarke’s
NEURONS
column
_-----------------------0 .--------------------------------------
291
1.0
\ 9
------------=. ..-
,r;“-;ri
---------
8 -----------------
mm I--w -we
B
Clarke’s
Column
1. Summarvw ofthe monosynuptic muscle inputs received . by 40 DSCT neurons
TABLE
Cell Number
PL
SOL
MG
1 2 3
0 0 0
0 0 l
0
0
4
0
0
0
l
5 6
0 0
0 0
0
0
0
0
7 8
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
0
0
0 0 0
0
0 0
lateral
LG FIG. 5.
A: schematic reconstruction in the sagittal plane of the L3 dorsal root entry zone, with the recorded position of 28 DSCT neurons receiving ankle extensor muscle input marked as individual filled circles. B: schematic reconstruction in the transverse plane showing the position of 12 DSCT neurons in Clarke’s column. These neurons all received group I afferent input from the ankle extensor muscles.
0 0
l 0
0 0 0 0 0 0
0
0 0 0 0 0 0 0 l
DSCT, dorsal spinocerebellar dial gastrocnemius; LG, lateral
0 0 0 0 0 l
0 0 0 0 0 0 0 0 0 0 0 tract; PI, plantaris; gastrocnemius.
Sol, soleus;
MG,
me-
(average, 6.4). [As noted by Houchin et al. (1983), it was often difficult to accurately determine the number of trunk dendrites, as many branched almost immediately after leaving the soma. The number of dendrites counted was, therefore, the most conservative estimate.] The diameters of the trunk dendrites were measured just beyond the initial tapered region and found to have an average diameter of 8.8_pm. Many small branchlets were seen arising from the dendrites of these cells. Beading along some of the dendrites was observed [in accordance with Loewy (1970) and Houchin et al. ( 1983)], although dendritic spines were very rare, and the dendrites were relatively smooth in appearance. DSCT neurons recorded during these experiments often received group I monosynaptic afferent input from more than one of the ankle extensor muscles. A summary of the muscle inputs received by 40 DSCT neurons is presented in Table 1. Five of these cells were shown to receive input from all ankle extensor muscles (LG, MG, Sol, and Pl). Two neurons received input from three of the muscles, 8 received input from two of the muscles, and the remaining 25 cells received input from only one of the muscles. Of these, two received input from LG, eight from MG, seven from Sol, and eight from Pl. Figure 5A shows a schematic reconstruction of the L3 dorsal root entry zone viewed in sagittal section. The location of 28 neurons receiving group I ankle extensor input is plotted, showing that these cells could be located along the
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
292
B. WALMSLEY
entire length of the L3 root, from rostra1 L3 past the caudal end of L3. (The apparent clustering of the cells toward the caudal end of L3 is not thought to be significant, because electrode tracking was often concentrated in this region.) Figure 5B shows the location in cross-section of the cell bodies of 12 DSCT neurons that received monosynaptic group I input from one or more of the ankle extensor muscles. It should be noted that this schematic diagram illustrates the relative location of cells within Clarke’s column, because the exact size and shape of Clarke’s column varies among animals (see Fig. 2). As clearly shown in Fig. 2, DSCT cells receiving ankle extensor input can be found throughout Clarke’s column and are not strictly confined to a particular region.
Golgi studies originally suggested that the cells in Clarke’s column giving rise to the DSCT are large, with profuse dendritic trees oriented in a primarily rostrocaudal direction (Boehme 1968; Loewy 1970). Subsequent attempts to label DSCT neurons intracellularly with HRP have revealed conflicting results (Houchin et al. 1983; Randic et al. 198 1). The present results have demonstrated that DSCT neurons receiving group I muscle afferent input have large cell bodies and extremely complex dendritic trees that may extend 3 mm in the rostrocaudal direction. In cross-section, the profusely tangled appearance of the dendrites is quite striking, as is the extent of their vast spread within Clarke’s column. In addition, the dendrites of these cells are often found to extend well beyond Clarke’s column in the lateral, ventral, and dorsal directions, in contrast to the results of Houchin et al. (1983) and Hongo ( 1985). A particularly interesting observation is that dendrites are found to reach some distance into the white matter of the dorsal columns. White-matter dendrites have also been observed arising from cervical motoneurons by Rose and Richmond ( 198 1). Electron microscopy showed these dendrites to receive synaptic contacts within the white matter (Rose and Richmond 198 1). Kuno et al. (1973) proposed that the somata of DSCT neurons in Clarke’s column were somatotopically organized, showing that cells that received input from gastrocnemius/soleus were more medially located than cells that receive biceps femoris input. Hongo and co-workers (Hong0 et al. 1982, 1985) returned to the question of somatotopic location of cell bodies of DSCT neurons, concluding that, in the transverse plane, the cells are arranged with respect to the proximity of the muscle supplying their afferent input. In this scheme, DSCT neurons that receive toe-muscle input are located dorsomedially in the column, those receiving shank input in the intermediate region, and those receiving input from thigh muscles in the ventrolatera1 zone. Hongo (1985) has proposed that these regions are quite definite (although they exhibit some overlap), with both the cell body and the dendrites confined to their respective functional regions. DSCT neurons labeled in the present study correspond
AND
M.
J. NICOL
with those receiving input from the shank muscles as defined by Hongo (1985). However, the present results indicate that these cells can be located throughout Clarke’s column in the transverse plane. This result does not, of course, rule out a general trend for these cells to be located in a particular region of Clarke’s column. Nevertheless, DSCT neurons receiving ankle extensor-muscle input may be located throughout Clarke’s column, including the central, dorsomedial, and ventrolateral regions. Furthermore, the dendritic trees of these neurons, rather than being confined to a particular functional zone, may occupy up to 60% or more of the transverse area of Clarke’s column and often extend well beyond the column boundary. The results presented here do not, therefore, provide convincing evidence for a strict transverse musculotopic arrangement of DSCT neurons and their dendrites within Clarke’s column. DSCT neurons have commonly been regarded in the role of relay neurons, conveying signals from restricted hindlimb receptor groups to the cerebellum. However, Osborn and Poppele (1988) have recently indicated that although DSCT neurons may receive a limited monosynaptic input, they receive significant polysynaptic influences from receptors throughout a widespread region of the hindlimb. They propose that DSCT neurons should be regarded in the role of integration of afferent information, rather than as simple relay neurons (Osborn and Poppele 1988). The present demonstration of complex dendritic trees that extend over a wide area both within and outside Clarke’s column supports the suggestion that these cells might play a much more complex role than previously imagined. We arc grateful to Dr. R. E. W. Fyffc for his comments on the manuscript. We also thank K. Howarth for participation in several of the experiments. This work was supported by a grant from the National Health and Medical Research Council of Australia. Address for reprint requests: B. Walmsley, Neural Research Laboratory, School of Anatomy, University of New South Wales, P. 0. Box I, Kensington, N. S. W. 2033, Australia. Received
12 June
1989; accepted
in final form
25 September
1989.
REFERENCES J. C. Technical considerations on the use of horseradish peroxidase as a marker. Ncz4rosc1icr7(‘I’2: 14 I- 145, 1977. ADAMS, J. C. Heavy metal intensification of DAB based HRP reaction product. J. lJistochcm. Cytochon. 29: 775, 198 1. BOEHME, C. C. The neural structure of Clarke’s nucleus of the spinal cord. J. Camp. Nvzrro~. 132: 445-462, 1968. CLARKE, J. C. Further researches on the grey substance of the spinal cord. Phi/ox Trans. R. Sot. Lond. 149: 437-467, 1859. HONGO, T. Functional and morphological organization of DSCT cells in Clarke’s column. In Development, Orgunization and Processing in Sensory Puthwuys, edited by M. Rowe and W. D. Willis. New York: Liss, 1985, p. 149-156. HONGO, T., KUDO, N., SASAKI, S., YAMASHITA, M., YOSHIDA, K., ISHIZUKA, N., AND MANNEN, H. Somatotopic organization of Clarke’s column in the cat (Abstract). N~wr-oscicvx~~~ Lctt. 9, Suppl.: S107, 1982. HONGO, T., Kur>o, N., SASAKI, S., YAMASHITA, M., YOSHIDA, K., ISHIADAMS,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
LOCATION
AND
MORPHOLOGY
N., AND MANNEN, H. Trajectory of group Ia and Ib fibers from the hind-limb muscles at the L3 and L4 segments of the spinal cord of the cat. J. Camp. Neurol. 262: 159-194, 1987. HOIJCHIN, J., MAXWELI,, D. J., FYFFE, R. E. W., AND BROWN, A. G. Light and electron microscopy of dorsal spinocerebellar tract neurons in the cat: an intracellular horseradish peroxidase study. Q. J. Exp. Physid. 68: 7 19-732, 1983. KUNO, M., MUNOZ-MARTINEZ, E. J., AND RANDIC, M. Sensory inputs to neurons of Clarke’s column from muscle, cutaneous and joint receptors. J. Physiol. Lord. 228: 327-342, 1973. LOEWY, A. D. A study of neuronal types in Clarke’s column of the adult 1970. cat. J. Comp. N~wrol. 139: 53-80, MANN, M. D. Clarke’s column and the dorsal spinocerebellar tract: a review. Brain Bchav. Evol. 7: 34-83, 1973. MENSE, S. AND CRAIG, A. D. Spinal and supraspinal terminations of ZUKA,
OF primary
DSCT afferent
NEURONS
293
fibres from
the gastrocnemius-soleus muscle in the cat. 1988. OSBORN, C. E. AND POPPELE, R. E. The extent of polysynaptic responses in the dorsal spinocerebellar tract to stimulation of group I afferent fibers in gastrocnemius-soleus. J. Neuroxi. 8: 3 16-3 19, 1988. RANDIC, M., MILETIC, V., AND LOEWY, A. D. A morphological study of cat dorsal spinocerebellar tract neurons after intracellular injection of horseradish peroxidase. J. Camp. Ncwrd. 198: 453-466, 198 1. RETHELYI, M. The Golgi architecture of Clarke’s column. Acta Morphol. Hung. 16: 31 l-330, 1968. ROSE, P. K. AND RICHMOND, F. J. White matter dendrites in the upper cervical spinal cord of the adult cat: a light and electron microscopic study. J. Camp. Neurd. 199: 19 l-203, 198 1. TRACEY, D. J. AND WALMSLEY, B. Synaptic input from identified muscle afferents to neurons of the dorsal spinocerebellar tract in the cat. J. Physiol. Lord. 350: 599-6 14, 1984.
Neuroscience 26: 1023-1035,
Downloaded from www.physiology.org/journal/jn by ${individualUser.givenNames} ${individualUser.surname} (139.184.014.150) on January 7, 2019.
