NIS 2463, pp. 265-273
Journal of Physiology (1994), 478.2
265
Sulphorhodamine-labelled cells in the neonatal rat spinal cord following chemically induced locomotor activity in vitro Ole Kjaerulff, Isabella Barajon and Ole Kiehn * Division of Neurophysiology, Department of Medical Physiology, The Panum Institute, University of Copenhagen, Blegdamsvej 3, DK-2200 Copenhagen, Denmark 1.
2.
3.
4.
Sulphorhodamine 101, a fluorescent dye and newly identified activity marker, was used to localize potential spinal locomotor networks in the neonatal rat spinal cord. Preparations of the spinal cord with one entire hindlimb attached or the spinal cord in isolation were kept in vitro. Spinal locomotor activity was maintained chemically with NMDA (5-7 5 uM), in combination with 5-HT (7-5-20 /M), for 4-4 5 h in the presence of 0 0001-0 0005 % sulphorhodamine 101. Matched non-locomoting controls were exposed to the dye in the absence of transmitters for a comparable time. Transverse sections of the lumbar spinal cord (L1-L6) were screened for rhodamine emission using an epifluorescence microscope. In hindlimb-attached locomoting preparations with intact dorsal roots, labelled cells were found on the leg side in the dorsal horn (mainly laminae II-IV), in the intermediate grey (lamina VI-VII) and around the central canal (lamina X). Dorsal rhizotomy was performed on the leg side, to prevent synaptic activity due to afferent inflow. This largely reduced the number of labelled cells in the dorsal horn and in the lateral part of the intermediate grey matter. A further reduction of labelling in these areas was seen after complete isolation of the cord or when compared to the legless side, with the majority of labelled cells persisting in a bilateral cluster close to the central canal and in the medial intermediate grey. Few labelled cells were observed in non-locomoting preparations. The intensity of motoneuronal labelling was variable. We suggest that the sulphorhodamine-labelled cells located in the intermediate grey and around the central canal are part of the neuronal network which generates spinal locomotor activity in the neonatal rat.
Rhythmic movements, such as walking, swimming and breathing, are basic features of all animals. In almost all species studied so far these movements can be generated in the absence of afferent input (Delcomyn, 1980), and the neuronal networks which generate the rhythmic motor patterns have been termed central pattern generators (CPGs). Considerable detail of how cellular and network mechanisms function together in generating motor programs has been outlined in invertebrates and also in some simple vertebrates (Roberts, Soffe & Dale, 1986; Harris-Warrick & Marder, 1991; Grillner & Matsushima, 1991). Much less is known about these mechanisms for both spinal and supraspinal motor CPGs in mammals. In the present study we have used the in vitro neonatal rat spinal cord preparation in order to explore the spinal CPG engaged in locomotion. Spinal locomotor activity can readily be induced in this preparation by external drug
application (Kudo & Yamada, 1987) and it can be monitored in great detail when the limb muscles are preserved. The neonatal rat preparation has also been used successfully to unravel some of the basic features of the respiratory network (Feldman, Smith & Liu, 1991). As a first step towards identifying cells that belong to spinal locomotor networks, we have used sulphorhodamine 101 in combination with chemically induced locomotion. Sulphorhodamine is one among several fluorescent probes which have been used to selectively mark electrically stimulated snake motor nerve terminals (Lichtman, Wilkinson & Rich, 1985). Recent experiments have shown that sulphorhodamine 101 also appears to be taken up in an activity-dependent manner in vertebrate central neurones (Keifer, Vyas & Houk, 1992), and it has been used to label active neuronal circuits of the turtle brain in vitro (Kriegstein, Avilla & Blanton, 1988; Keifer et al. 1992), as
*To whom correspondence should be addressed.
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
266
0. Kjarulff, I. Barajon and 0. Kiehn
well as spinal frog motoneurones engaged in executing central motor programmes (Giszter, Mussa-Ivaldi & Bizzi, 1993). In our experiments we have used sulphorhodamine in an attempt to label active spinal locomotor circuits in the in vitro neonatal rat, focusing on the distribution of labelled cells in the lumbar spinal cord during different experimental conditions. We suggest that our findings may provide a first insight into the localization of interneurones involved in the generation of spinal locomotor activity in the neonatal rat. Preliminary results from these experiments have been presented in an abstract (Kiehn, Kjaerulff & Barajon, 1993).
METHODS Dissection Preparations of the spinal cord with one entire hindlimb attached (n =9) or the spinal cord in isolation (n =3) were made in neonatal rats aged 0-4 days. The animals were anaesthetized with ether, decapitated and eviscerated before being transferred to a dissection chamber filled with oxygenated, ice-cold, low-calcium Krebs solution (the reduction in calcium was 90 % and obtained by replacing CaCl2 with equimolar concentrations of MgSO4; see below). The neuroaxis was cut at the medullospinal transition and the entire spinal cord was exposed by a ventral laminectomy. In the spinal cord-hindlimb preparations all dorsal and ventral roots, except those from Li-L6 on the leg side, were cut. In some experiments (n= 4) the remaining dorsal roots were also cut. All preparations were pinned down ventral side upwards in an experimental chamber perfused with normal Krebs solution (pH 7 4) containing (mM): 118 NaCl, 4-69 KCl, 25 NaHCO3, 1418 KH2PO4, 1P25 MgSO4, 2-52 CaCl2 and 11 glucose, which was oxygenated with 95 % 02 and 5 % CO2. The temperature of the chamber was kept constant at 25 + 1 'C.
Recording In the spinal cord-hind limb preparations muscle activity was recorded from selected muscles with pairs of Teflon-coated platinum-iridium wires (25-50 ,usm diameter; Clark Electromedical Instruments, Reading, UK). In the isolated spinal cord preparations ventral roots were placed in suction electrodes for recording. All recordings were bandpass filtered (100-10000 Hz), digitized, stored on a digital tape recorder (Biologic DTR 1800) and printed on thermosensitive paper (Gould 4000, Ilford, UK).
Induction of locomotor activity, and staining Long-lasting locomotor activity was achieved by adding a combination of N-methyl-D-aspartic acid (NMDA, 5-7-5 #LM; Sigma, UK) and 5-hydroxytryptamine (5-HT, 7-5-20 #M; Sigma) to the superfusion medium. Sulphorhodamine 101 (Molecular Probes, Eugene, OR, USA, or Sigma) was added after initiation of locomotion, which was then maintained for 4-4 5 h. Following staining the spinal cord was washed in tetrodotoxin-containing medium (10 /M) without transmitter for 30-45 min. Initially we used sulphorhodamine at the concentration (0'01 %) used by Lichtman et al. (1985) and Keifer et al. (1992). We discovered that at this concentration the dye itself produced frequent and centrally generated
J. Physiol 478.2
synchronized motor activity, which caused the transmitterinduced locomotor pattern to become very irregular. However, by reducing the dye concentration to 0-0001-0-0005 % (0-1 x 100-l-0 5 x 100- mg ml-) the locomotor pattern was undisturbed in most experiments, while in a few some disturbances of the pattern were observed. The latter experiments were excluded. As a consequence of the reduced dye concentration we implemented 4-4 5 h staining sessions instead of the 45 min sessions used in previous studies (Keifer et al. 1992). The EMG activity on the leg side in the hindlimbattached preparations or the ventral root discharges on the left and right side of the cord in the isolated preparations were monitored on the chart recorder throughout the staining session. Motor activity was not monitored on the legless side in the hindlimb-attached preparations. However, in other experiments where the ventral root discharges were monitored on the side opposite to the preserved leg, locomotor activity, when induced, was always present bilaterally (0. Kjaerulff & 0. Kiehn, unpublished observations). In conclusion, only preparations where it was possible to maintain long-lasting regular locomotor activity in the presence of sulphorhodamine were included. As controls, spinal cords were exposed to sulphorhodamine without transmitters in the superfusion medium. Preparations where EMG activity was observed were discarded.
Anatomy After the experiments the lumbar spinal cord segments were fixed by overnight immersion in 4 % paraformaldehyde phosphate buffer (PFA, pH 7 4), cryoprotected for 24-48 h in 30 % sucrose, and cut in 20 ,um thick transverse sections on a cryostat. The sections were dehydrated at 4 °C in a desiccator for 48 h and successively cleared in xylene prior to mounting in Entellan (Merck). Mounted sections were then stored at -20 0C. Light microscopic observations and photographic documentation were performed on a Zeiss microscope equipped with rhodamine epifluorescence. Reconstructions of the cellular distribution in lumbar double segments (L1-L2, L3-L4 and L5-L8) were made from drawings of projections (Zeiss projector) of micrograph negatives. Three sections (one rostral, one mid and one caudal) from each double segment were superimposed to construct the maps in Figs 3 and 5. Figure 4 is based on cell counts from these maps. The intensity of cell labelling was variable, with brightly and faintly stained cells intermingled (e.g. Fig. 2). In Figs 3-5 we have only plotted the most intensively stained cells. This method is likely to underestimate the number of active cells (see Results). For evaluation of normal cytoarchitecture, two animals were anaesthetized and perfused transcardially with ice-cold heparinized (15000 i.u. ml') Krebs solution, followed by 4 % PFA (pH 7 4). The lumbar spinal cord was removed from these animals and the tissue was post-fixed in 4 % PFA for 24 h before being embedded in paraffin and cut in 20 ,sm thick transverse sections. The sections were stained according to the Nissl method. Since there is no available reference for the outline of the laminar structure of the spinal cord grey matter in neonatal rats we have adapted Rexed's scheme obtained in adult rats (Molander, Xu & Grant, 1984). We use this scheme in combination with the the terminology of Altman & Bayer (1984) when referring to localization of cells.
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
Intact dorsal roots
RESULTS Sulphorhodamine labelling in hindlimb-attached
preparations The initial experiments were aimed at examining the spatial pattern of sulphorhodamine labelling in the lumbar spinal cord as a consequence of chemically induced locomotor activity in hindlimb-attached preparations. Figure IA shows the experimental set-up for the hindlimbattached preparation (note that only one leg is preserved). EMG activity of selected hindlimb flexors and extensors served to monitor spinal locomotor activity before (B; NMDA plus 5-HT) and after (C; transmitters plus sulphorhodamine) sulphorhodamine was added to the mixture. The quality of the labelling as it appears in epifluorescence viewing of 20,m sections is depicted in Fig. 2A and B. A Nissl-stained section with the corresponding spinal cord areas marked is shown below (Fig. 2C). Figure 2A shows a high-power micrograph of intensively stained cells located dorsolateral to the central canal of the spinal cord in Ll-L2. The dye was mainly found in cell bodies, although in some cells labelled processes extended from the soma. When nuclei were present in the plane of section they could be seen as unstained circular profiles in which nucleoli were sometimes detected (not shown). The low-power micrograph (Fig. 2B) illustrates the topography of labelled cells around the central canal (lamina X) and in the medial intermediate grey (lamina VII) in Ll-L2.
A
267
Labelling of locomotor-related cells
J. Physiol. 478.2
The distribution of sulphorhodamine-labelled cells in one of the hindlimb-attached preparations with intact dorsal roots is illustrated in Fig. 3A. Maps of labelled cells were derived from superimposed drawings, as described in Methods. Although some variations in the spatial pattern were observed among animals, a consistent distribution of cellular elements appeared from surveying the sections. In all animals and segments a clear asymmetry was observed, with more cells labelled on the ipsilateral 'leg side' than on the contralateral 'legless' side. On the leg side numerous cells were found in the superficial layers of the dorsal horn (mainly laminae II-III) and in the neck region (lamina IV-V). These small densely packed cells had no visible dendritic staining, but showed intensively labelled somas. Cells were also found in the nucleus of the dorsolateral funiculus ('DN'). Intensively labelled cells were located dorsal and lateral to the central canal, in lamina X and the medial part of lamina VII. These groups of cells were connected with clusters of cells in the intermediate grey (lamina VII in L1-L2 and laminae VI-VII in L3-L6). In these laminae large cells with labelled processes were scattered among smaller cells with mainly somatic labelling (see Fig. 2A). The laminar distribution with the actual number of labelled cells on the leg side in three different preparations is shown in Fig. 4A (filled bars). In general, very few cells were found in lamina VIII, and this lamina is shown, together vith laminae VI and VII, in Fig. 4.
C
B
NMDA + 5-HT + sulphorhodamine
NMDA + 5-HT
AM
14J1J4i4J
TA+N4 IL
RF
I
AiLli
10
7-
ilku-
Wfl. 6-4
, *.. ~.U4H_M44 4 ,,gL
jIII
I LAl I I'
1 s
Figure 1. Transmitter-induced locomotor activity in the in vitro neonatal rat spinal cord-hindlimb preparation A, experimental set-up. B, rhythmic EMG activity in hindlimb muscles following bath application of 75 AuM NMDA in combination with 15 AM 5-HT. The activity is seen to alternate between flexors, iliopsoas (IL) and tibialis anterior (TA), and extensors, adductor magnus (AM) and rectus femoris (RF). C, recordings from the same muscles after 3 h of continuous locomotor activity in the presence of 00002 % sulphorhodamine (w/v).
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
268
0. Kjaerulff, I. Barajon and 0. Kiehn
On the legless side consistently fewer cells were found in laminae II-V in the dorsal horn, especially in the lateral part (Figs 3A and 4A and B). Also in the lateral part of the intermediate grey matter the number of labelled cells was considerably lower when compared to the leg side (Fig. 4A-B), while in the medial part of the intermediate grey (laminae VI-VII) the reduction was less pronounced (see later). In two animals, faintly stained motoneurones (recognized by their size) could be detected in the lateral part of the ventral horn bilaterally (not plotted in Fig. 3A; see Methods). Very intensively stained motoneurones (plotted in Fig. 3A) were seen on the legless side in one
J. Physiol. 478.2
animal, associated with the presence of stained cells located laterally in the intermediate grey. Motoneurones are not plotted in Fig. 4.
Dorsal rhizotomy To identify the source of the asymmetrical cell labelling found in hindlimb-attached preparations with intact dorsal roots on the leg side, we cut all dorsal roots on this side. The results from such an experiment are shown in Fig. 3B. In the three preparations with cut dorsal roots the total number of cells was reduced compared with the total cell count in the three preparations with intact dorsal roots. The difference was not significant (two-tailed t test,
C
Figure 2. Photomicrographs of sulphorhodamine-labelled cells following spinal locomotion A, labelled neurones located dorsolateral to the central canal. Note processes extending from some of the labelled cells. B, low-power photomicrograph of labelled cells situated around the central canal and in the medial part of the intermediate grey. C, photomicrograph of a Nissl-stained transverse section of the lumbar spinal cord. The small and innermost frame indicates the position of the section in A, while the large outermost frame indicates the position of the section in B. Scale bars: 25 ,um in A, 100 um in B and 200 ,sm in C.
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
J. Physi'ol. 478.2
Labelling of locomotor-related cells
P > 0 05), possibly due to the small sample size and the high variability in cell numbers in individual preparations
(e.g. Fig. 4). On the leg side the number of labelled cells in the dorsal horn as a whole was decreased compared with preparations with intact dorsal roots. The decrease was mainly confined to the lateral part of lamina I-II (cut vs. intact; P < 0 10) and the lateral part of lamina III-IV (P < 0-05). The
A
B
269
number of labelled cells in the lateral part of the intermediate grey (V-VII) was also reduced (P < 0410). However, there was still pronounced labelling in the medial part of laminae JJ-IV and in lamina V (P > 0.10 for both areas). Furthermore, the number of labelled cells remained high in the medial part of laminae VI-VII and in lamina X (P > 0-10 for both areas). These changes in distribution can also be seen in Fig. 4A. C
Ll -L2
L3-L4
'-II L5-L6 I.
V1
y.. Vill
Figure 3. Distribution of sulphorhodamine-labelled cells in spinal segments LI-L6 Results from hindlimb-attached preparations. Data from the leg which was preserved is given in the left side of panels A-C. All ventral and dorsal roots were cut on the side opposing the leg. In A, ventral and dorsal roots of L1-L6 were left intact on the leg side while in B dorsal roots were cut. Locomotion was maintained with NMDA and 5-HT in the presence of sulphorhodamine. The staining period was 4 h in A and 4 h 35 min in B. Each diagram includes all labelled cells in three representative sections from each of the double segments L1-L2, L3-L4 and L5-L6. Each dot represents one cell. Large dots indicate motoneurones. C is a control matching B (cut dorsal roots), stained for 4 h in the absence of transmitters. Controls matching A are described in the text. DN, nucleus of the dorsolateral funiculus. C, Clarkes column. Numerals I-X indicate the laminar structure of the lumbar spinal cord, while the dashed line shows the division between the medial and lateral parts of the laminae.
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
0. Kjyrulff, I. Barajon and 0. Kiehn
270
J. Physiol. 478.2
A 300
-
250
-
n 2000
o 1506
Z 10050
Figure 4. Laminar distribution of sulphorhodaminelabelled cells in the hindlimb-attached preparations with the dorsal roots cut (open bars) and with intact dorsal roots (filled bars) A shows the distribution on the leg side, while B shows data from the legless side. The laminae have been divided into medial and lateral parts (see Fig. 3C). Each bar indicates one animal and the sequence of animals is the same in A and B. The number of labelled cells in a given area was obtained by summing the cell numbers from that area in segments Ll-L6. Lamina IX is not shown (see text).
-
0-
B 300 -
250
-
200
-
o 150
-
co
Legless side
6
Z 10050 0- I _n
I FL
I-
-
-
-
X c--'a
E
'
E
>
D I
u Cl'ZiX (a >
'
E
x
E
> >I I_> ~~~~~~~~~>
-
I.
Li -L2
L3-L4
Figure 5. Distribution of sulphorhodamine-labelled cells in spinal segments Ll-L6 in an isolated spinal cord preparation Locomotor activity was maintained by a combination of NMDA and 5-HT for 4 h in the presence of sulphorhodamine. Each diagram includes all labelled cells in three representative sections from each double segment, and each dot represents one cell. Motoneurones are indicated by large dots.
L5-L6
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
J. Physiol. 478.2
Labelling of locomotor-related cells
On the legless side dorsal rhizotomy did not cause any major change in the distribution of labelled cells compared with the legless side with intact dorsal roots (Fig. 4B; P > 0X10 for all areas). Motoneuronal staining was similar to that seen in preparations with dorsal horns intact, with motoneurones intensively stained on the legless side in two preparations. As in the preparation with intact dorsal roots this staining pattern was associated with cell labelling laterally in the intermediate grey (Fig. 3B). Faintly stained motoneurones were observed bilaterally in three preparations. It appears from these experiments that the dorsal rhizotomy on the leg side reduced the density of labelling in the dorsal horn, especially laterally. In contrast, little change was observed in the medial part of the dorsal horn, in the cluster of cells situated around the central canal (lamina X) and in the medial intermediate grey (laminae VI-VII). Moreover, the labelling pattern on the legless side was similar in all the preparations (with or without dorsal roots).
Labelling asymmetry in the hindlimbattached preparations The difference in cell distribution between the leg side and the legless side was evaluated for all hindlimb-attached preparations using a variance component model (Miller, 1986). Three areas of the grey matter were considered: (i) laminae I-V, (ii) the lateral part of laminae VI-VIII and (iii) the medial part of laminae VI-VIII together with lamina X. It was found for all three areas that the difference between the number of cells on the leg side and that on the legless side was not changed by cutting the dorsal roots. The cell number was significantly higher on the leg side than on the legless side in laminae I-V (P=0 001) and in the lateral part of laminae VI-VIII (P < 005). On the other hand, no significant difference in cell number was found between the two sides in the medial part of laminae VI-VIII and lamina X (P > 0410).
Non-locomoting controls Three matched non-locomoting control animals (two with intact and one with cut dorsal roots) were superfused with sulphorhodamine-containing solution without transmitter for 4-4 5 h. In two of these controls a few labelled cells were scattered in laminae VII-VIII in L3-L4 (n = 1; intact dorsal roots) or in the dorsal horn (n = 1; cut dorsal roots; Fig. 3C). In the third control no labelled cells were found.
Sulphorhodamine labelling in isolated spinal cord preparations It was somewhat surprising that the dorsal rhizotomy did not eliminate dorsal horn staining on the leg side, especially since on the legless side very few labelled cells were present in this region (Fig. 3A and B). However, since
271
it is known that afferents are present in the ventral roots in neonatal rats (Jiang, Shen, Wang & Dun, 1991), we completed the deafferentation in three experiments by isolating the spinal cord. As a result, most of the staining in the dorsal horn on both sides was eliminated in two out of three animals (Fig. 5). In the third animal a few cells were still scattered in the dorsal horn (not shown). Labelled cells now appeared concentrated in two columns lateral to the central canal in the medialmost part of the intermediate grey throughout the lumbar spinal cord. In L1-L4 these columns were bridged by cells dorsal to the central canal. Unilateral bright motoneuronal labelling accompanied by a varying number of stained cells located dorsolaterally in lamina VII (compare Fig. 3) was still observed in one of these animals, while in two rats motoneurones were visible bilaterally (Fig. 5). At present we have no explanation for the remaining asymmetrical staining in the isolated spinal cord (e.g. the unilateral motoneuronal labelling). In particular, the asymmetry was not easily explained by the locomotor performance, since alternating motor activity was continuously present on both sides of the cord throughout the experiments
Labelling in the thoraco-lumbar transition Labelling in the rostral and caudal extension of the lumbar spinal cord was not systematically investigated. However, the caudal thoracic segments (Thll-Th13) were inspected in four preparations (two hindlimb-attached preparations with intact dorsal roots and two isolated preparations). Characteristically, the large majority of labelled cells were found around the central canal and in the medial intermediate grey. The density of labelled cells in these areas was lower than in the corresponding areas in L1-L6, whereas the labelling intensity of individual cells was similar.
DISCUSSION These experiments provide a first attempt to localize en bloc neurones which may be engaged in spinal locomotor generation in the neonatal rat. A survey of the spatial patterns indicate that distinctive distributions of labelled cells were obtained in the different experimental situations. Characteristically, the number of labelled cells in the dorsal horn (laminae I-V) and in the lateral intermediate grey (lateral laminae VI-VII) was reduced with progressive deafferentation, possibly due to a reduction of afferent inflow through both dorsal and ventral roots. In contrast, labelled cells in the dorsomedial, central and medial intermediate grey (medial laminae VI-VII and lamina X) remained after total isolation of the cord. We suggest that cells located in these areas are engaged in the generation of spinal locomotor activity in
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
272
0. Kjerulff, I. Barajon and 0. Kiehn
the neonatal rat. Before discussing this topic, we will briefly comment on the evidence that sulphorhodamine uptake is activity dependent.
Sulphorhodamine as an activity marker Keifer et al. (1992) have provided strong evidence that sulphorhodamine uptake is dependent on synaptic activity. The specific labelling thus disappeared when synaptic transmission was blocked in low-calcium, highmagnesium solution or with the excitatory amino acid transmitter blockers 6-cyano-7-nitroquinoxaline-2,3,-dione (CNQX) and D-aminophosphonovalerate (APV). These experiments complemented previous reports which showed that sulphorhodamine was taken up selectively in electrically activated motor terminals (Lichtman et al. 1985) and in nerve terminals in the cortex at epileptic sites (Kriegstein et al. 1988). Together these experiments support the idea of an activity-dependent neuronal uptake, which is related to synaptic recycling (see also Betz & Bewick, 1992). In our experiments we have not provided evidence for a direct link between synaptic activity and sulphorhodamine labelling, but the decrease in labelling following deafferentation certainly does suggest such a link. Dorsal rhizotomy would thus be expected to reduce afferent inflow, since it is known that locomotor-related afferent input is conveyed onto the spinal cord via the dorsal roots in the cat (Arshavsky, Berkinblit, Fukson, Gelfand & Orlovsky, 1972). Electrophysiological experiments in the neonatal rat also indicate that a strong afferent input can reach the spinal cord via ventral roots (Jiang et al. 1991), although at the moment there is no physiological evidence which suggests that locomotor-related sensory information can reach the dorsal horn from ventral roots. The lack of staining in non-locomoting animals and the fact that less intensive or no labelling was obtained when the staining period was decreased in locomoting animals (0. Kjaerulff, I. Barajon & 0. Kiehn, unpublished results) also indicate that sulphorhodamine is taken up in an activity-dependent manner. Since we used neuroactive drugs (5-HT and NMDA) to induce the locomotor activity it is conceivable that some neurones not related to locomotion are labelled simply because they are strongly activated by the drugs. However, as prolonged locomotor activity is the dominant motor output from the cord after drug application, we suggest that many of the sulphorhodamine-labelled cells are indeed related to locomotion. Moreover, the fact that rhizotomies change the staining pattern seems to support the idea of a locomotor-related staining pattern, rather than an unspecific transmitter activation of neurones. At present we have no explanation for the variability of motoneuronal staining. The faint staining might be due to the large size of these neurones or a less effective reuptake mechanism in the cord compared to other neurones. This would explain why the faint labelling was mainly found on the leg side, where staining might be hampered by the
J. Physiol. 478.2
distance between the main synaptic uptake site for the dye (presumably the neuromuscular endplate). In contrast, the bright labelling, which was pronounced on the legless side and in the isolated preparations, might be explained by the generally easier access of the dye to the motoneuronal somata via the cut ventral roots. It appears unlikely that this staining is due to an activity-independent bulk uptake, since no retrograde labelling of motoneurones was observed in the three control experiments. It is of interest that Giszter et al. (1993) have shown that frog motoneurones in vivo are strongly labelled with sulphorhodamine after electrical activation. This difference in results might be explained by the fact that these authors used a much higher sulphorhodamine concentration (0 5 %) than used by us. Due to the side-effects of sulphorhodamine described in Methods, we have not tried similar high doses of sulphorhodamine in the neonatal rat.
Labelling pattern The most consistent labelling in all experimental conditions was found throughout the lumbar spinal cord in the dorsomedial, the central and the intermediate grey areas (laminae VI-VII; lamina X). Thus the number of sulphorhodamine-labelled cells varied little in these areas in the different experimental conditions. We suggest that cells located in these areas are engaged in the generation of locomotion in the neonatal rat. This indicates that the spinal locomotor network is distributed, although not necessarily evenly, to all segments. This is in agreement with physiological experiments, which have shown that a single segment of the cord is capable of generating a rhythmic motor output (Harder & Schmidt, 1992). Interestingly, previous studies in the rabbit and cat using various activity markers have pointed to the intermediate grey and the area around the central canal as important for generating spinal locomotion. In the rabbit the uptake of 2-deoxyglucose during L-f-3,4-dihydroxyphenylalanine (L-DOPA)-induced fictive locomotion was found to be localized in cells in the intermediate grey (Viala, Buisseret-Delmas & Portal, 1988), while in locomoting deafferented cats c-fos was detected in cells in the intermediate grey and area X (Dai, Douglas, Nagy, Noga & Jordan, 1990). Isopotential field studies in the cat have also identified these areas as potential network areas (Jordan, 1991). In addition, studies using extracellular recordings have revealed rhythmically active (unidentified) cells in the intermediate grey during fictive locomotion (Orlovsky & Feldman, 1972; Jordan, 1991). Recent experiments using optical imaging and ablation have also pointed towards ventral and intermediate regions as potential locomotor generators in the chick embryo spinal cord (O'Donovan, Sernagor, Sholomenko, Ho, Antal & Yee, 1992). In this context it is of interest that motoneurones have proved difficult to label both with the c-fos and 2-deoxyglucose methods (Viala et al. 1988; Dai et al. 1990;
Barajon, Gossard & Hultborn, 1992).
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
J. Phy8iol.478.2
Labelling of locomotor-related cells
It should be noted that the observation that labelled cells were found in the intermediate zone and the central grey of the thoracic cord (where all afferent input was abolished) is consistent with the idea of these areas being engaged in generating rhythmic motor output. Thus, recent experiments have shown that rhythmic motor output emerges from thoracic ventral roots when the entire spinal cord is superfused with locomotor-inducing drugs (0. Kjaerulff & 0. Kiehn, unpublished observations, and B. Schmidt & N. Kudo, personal communications). At present no electrophysiological data are available on the localization of rhythmically active interneurones in the neonatal rat spinal cord. However, our findings might provide a framework for conducting such studies. The recent observation that groups of sensory signals have direct access to the rhythm generator even at birth (Kiehn, Iizuka & Kudo, 1992) has added a further tool for identifying potential locomotor network neurones in the neonatal rat spinal cord.
REFERENCES ALTMAN, J. & BAYER, S. A. (1984) The development of the rat spinal cord. Advances in Anatomy Embryology and Cell Biology 85,1-166. ARSHAVSKY, Y. I., BERKINBLIT, M. B., FUKSON, 0. I., GELFAND, I. M. & ORL,OVSKY, G. N. (1972). Recordings of neurons of the dorsal spinocerebellar tract during evoked locomotion. Brain Research 43,272-275. BARAJON, I., GOSSARD, J.-P. & HULTBORN, H. (1992). Induction of c-fos expression by activity in the spinal rhythm generator for scratching. Brain Research 588, 168-172. BETZ, W. J. & BEWICK, G. S. (1992). Optical analysis of synaptic vesicle recycling at the frog neuromuscular junction. Science 255, 200-203. DAI, X., DOUGLAS, J. R., NAGY, J. I., NOGA, B. R. & JORDAN, L. M. (1990). Localization of spinal neurons activated during treadmill locomotion using the c-fos immunohistochemical method. Society for Neuroscience Abstracts 16, 889. DELCOMYN, F. (1980). Neural basis of rhythmic behavior in animals. Science 210, 492-498. FELDMAN, J. L., SMITH, J. C. & Liu, G. (1991). Respiratory pattern in generation in mammals: in vitro en bloc analyses. Current Opinion in Neurobiology 1, 590-594. GISZTER, S. F., MUSSA-IVALDI, F. A. & Bizzi, E. (1993). Convergent force fields organized in the frog's spinal cord. Journal of Neuroscience 13, 467-491. GRILLNER, S. & MATSUSHIMA, T. (1991). The neural network underlying locomotion in lamprey. Neuron 7, 1-15. HARDER, K. C. & SCHMIDT, B. J. (1992). Effects of inhibitory amino acid agonists and antagonists on locomotor rhythmicity in the rat spinal cord in vitro. Society for Neuroscience Abstracts 18, 315. HARRIS-WARRICK, R. M. & MARDER, E. (1991). Modulation of neural networks for behavior. Annual Review of Neuroscience 14, 39-57. JIANG, Z. G., SHEN, E., WANG, M. Y. & DUN, N. J. (1991). Excitatory postsynaptic potentials evoked by ventral root stimulation in neonate rat motoneurons in vitro. Journal of Neurophysiology 65, 57-66. JORDAN, L. M. (1991). Brainstem and spinal cord mechanisms for initiation of locomotion. In Neurobiological Basis of Human Locomotion, ed. SHIMAMURA, M., GRILLNER, S. & EDGERTON, V. R., pp. 3-20. Japan Scientific Societies Press, Tokyo.
273
KEIFER, J., VYAS, D. & HOUK, J. C. (1992). Sulphorhodamine labeling of neural circuits engaged in motor pattern generation in the in vitro turtle brainstem-cerebellum. Journal of Neuroscience 12, 3187-3199. KIEHN, O., IIZUKA, M. & KuDo, N. (1992). Resetting from low threshold afferents of N-methyl-D-aspartate-induced locomotor rhythm in the isolated spinal cord-hindlimb preparation from newborn rats. Neuroscience Letters 148, 43-46. KIEHN, O., KJA:RULFF, 0. & BARAJON, I. (1993). Spinal locomotor network in the neonatal rat revealed by neuronal uptake of sulphorhodamine. Society for Neuroscience Abstracts 19, 557. KRIEGSTEIN, A. R., AVILLA, J. G. & BLANTON, M. G. (1988). Distribution of increased synaptic activity during focal and generalized epileptiform activity revealed by presynaptic uptake of fluorescent dyes. Society for Neuroscience A bstracts 14, 471. KuDo, N. & YAMADA, T. (1987). N-Methyl-D,L-aspartate-induced locomotor activity in a spinal cord-hindlimb muscles preparation of the newborn rat studied in vitro. Neuroscience Letters 75, 43-48. LICHTMAN, J. W., WILKINSON, R. S. & RICH, M. M. (1985). Multiple innervation of tonic endplates revealed by activity-dependent uptake of fluorescent probes. Nature 314, 357-359. MILLER, R. G. (1986). Beyond ANOVA: Basics of Applied Statistics. Wiley, New York. MOLANDER, C., Xu, Q. & GRANT, G. (1984). The cytoarchitectonic organization of the spinal cord in the rat. I. The lower thoracic and lumbosacral cord. Journal of Comparative Neurology 230, 133-141. O'DONOVAN, M., SERNAGOR, E., SHOLOMENKO, G., Ho, S., ANTAL, M. & YEE, W. (1992). Development of spinal motor networks in the chick embryo. Journal of Experimental Zoology 261, 216-273. ORLOVSKY, G. N. & FELDMAN, A. G. (1972). Classification of lumbosacral neurones according to their discharge patterns during evoked locomotion. Neurophysiology (Kiev) 4, 410-417. ROBERTS, A., SOFFE, S. R. & DALE, N. (1986) Spinal interneurones and swimming in frog embryos. In Neurobiology of Vertebrate Locomotion, ed. GRILLNER, S., STEIN, P. S. G., STUART, D. G., FORSSBERG, H. & HERMAN, R. M., pp. 279-306. Macmillan, London. VIALA, D., BUISSERET-DELMAS, C. & PORTAL, J. J. (1988). An attempt to localize the lumbar locomotor generator in the rabbit using 2-deoxy-[14C]glucose autoradiography. Neuroscience Letters 86, 139-143.
Acknowledgements We would like to express our gratitude to Dr Morten M0ller for providing anatomical help throughout the study and to Per Kragh Andersen for advice on statistical problems. The research was supported by Carlsberg Fondet, lb Henriksens Fond, LIege Sofus Carl Emil Friis og Hustru Olga Doris Friis' Legat, Novos Fond, Fonden til Fremme af Eksperimentel Neurologisk Forskning and The Human Frontier Science Program. Ole Kiehn is a senior Research Fellow supported by the Weimann Foundation.
Author's present address Dr Isabella Barajon: Instituto di Anatomia Umana, Via Mangiagalli 31, 20133 Milano, Italy. Received 11 June 1993; accepted 26 November 1993.
Downloaded from J Physiol (jp.physoc.org) at Københavns Universitetsbibliotek on April 14, 2009
