| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Centre de Recherche en Sciences Neurologiques, Department of Physiology, Pavillon Paul-G.-Desmarais, Université de Montréal, Montreal, Quebec, Canada
Submitted 1 September 2004; accepted in final form 2 January 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
), a behavior that is essentially generated by a spinal "central pattern generator" (CPG) (Grillner 1981). The question studied here concerns the segmental distribution of this locomotor network and whether it is regionalized or distributed throughout the spinal cord of the cat. In the in vitro neonatal rat, the thoraco-lumbar cord constitutes a key area for locomotor pattern generation (Cazalets 2000) although the sacral cord also has autonomous rhythmogenic capabilities (Cazalets and Bertrand 2000a; Lev-Tov et al. 2000). Other works also suggest a regionalized organization of rhythmogenicity based on the segmental distribution of various serotonin (5-HT) receptors (Jordan and Schmidt 2002). Chronic lesions of the gray matter at the level of L2 by kainic acid results in paraplegia in adult rats (Magnuson et al. 1999). Other studies in which spinal lesions and neurotransmitters were also used on various spinal segments suggest that the rhythmogenic potential is distributed throughout the spinal cord, with rostral lumbar segments being more excitable (Kjaerulff and Kiehn 1996).
It was suggested that, in cats, the rostral lumbar segments L3–L5 play a leading role during scratching but that rhythmogenicity was present in lower L6–S1 segments (Deliagina et al. 1983), a conclusion also reached in studies of "fictive locomotion" induced by L-3,4-dihydroxyphenylalanine (L-DOPA) in acute spinal cats (Grillner and Zangger 1979). We have shown potent effects on locomotion of small 100-µl injections of 
2-noradrenergic agonist (clonidine) and antagonist (yohimbine) through a chronically implanted intrathecal cannula that delivered the drugs in a limited region around L3–L4 (Giroux et al. 2001). We also demonstrated that in decerebrated cats spinalized at T13, intraspinal microinjections of clonidine at L3–L4 segments could evoke a locomotor pattern that could be abolished by a further microinjection of yohimbine. Similarly, after an intravenous injection of clonidine, locomotion was blocked by microinjections of yohimbine in the L3 or the L4 segment (Marcoux and Rossignol 2000). Finally, we also showed that locomotion, evoked by intravenous or intrathecal injections of clonidine, disappeared after a second lesion at L4, suggesting that the integrity of the segments above L4 was necessary to trigger and to maintain spinal locomotion. However, we could not determine whether the abolition of locomotion was due to depressive effect on the cord of this second acute spinal lesion or to the disconnection of the mid-lumbar segments from the lower lumbar segments. The present work presents a study of this problem using a double chronic spinal lesion. Cats were first spinalized at T13 and trained to regain spontaneous spinal locomotion on the treadmill. Then a single second spinal lesion was performed at various lower lumbar levels. Our results show that walking capabilities remained after lesions at L2 or rostral L3 but was permanently abolished after spinal transections that disconnected the midlumbar segments (caudal L3 or L4) from the L5–S1 segments. However, other rhythmic patterns such as fast paw shakes could be elicited even though locomotion could no longer be evoked. Because the main motoneuron pools of the hindlimbs are located in the L5–S1 segments (Vanderhorst and Holstege 1997) and because they can be recruited in rhythmic patterns other than locomotion after lesions at L3–L4, it is concluded that the midlumbar segments provide essential inputs to organize the locomotor pattern in hindlimb motoneurons. Part of this work has been published in abstract form (Langlet and Rossignol 2002).
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
GENERAL PROTOCOL. First, cooperative adult cats (n = 7) were trained to walk on a motor-driven treadmill at different speeds (0.3–0.6 m/s) every day during 1 mo. Once the cats were able to walk at a constant speed for 15–20 min, intramuscular electrodes were implanted chronically in various flexor and extensor muscles of the hindlimbs. After implantation, locomotion was recorded to establish baseline values in the intact state. The cats were then spinalized at the last thoracic vertebra (T13, labeled A in Fig. 1) and trained to walk on the treadmill every day for 3–4 wk without any drugs. When they regained a stable locomotor pattern, a second spinalization was performed at L2 (n = 2; cats B and C), L3 (n = 2, cats D and E), L4 (n = 2; cats F and G), and L5 (n = 1, cat H) and, again, the ability to walk was documented for several weeks. In Table 1, each cat is identified by a letter and the level of the second lesion is shown for each cat as well as the number of days between the first spinal section at T13 and the second spinal section plus the number of days the animals were kept after the second lesion. At the end of the experimental series, an acute experiment was carried out in the cats that were spinalized at L2 and at rostral L3 to perform other acute lesions at more caudal levels. Two other cats, spinalized only at T13 and trained to walk on the treadmill for 1 mo, were also studied in acute experiments. All procedures were approved by the Ethics Committee of Université de Montréal.
|
|
A Fentanyl (2.5 mg, 25 µg/h) transdermal patch was applied on the lower back of the cat for continuous analgesia for the first three postoperative days and an antibiotic (Cephatab, 100 mg po) was given prophylactically for 10 days after surgery.
SPINAL CORD LESIONS.
After obtaining the baseline values of kinematics and EMG activity during locomotion in the intact state, the cats were anesthetized, and a laminectomy was performed at the T13 vertebra. The dura was removed carefully and Lidocain (2% xylocaine) was applied on the surface of the cord and 
100 µl injected bilaterally into the spinal cord. The spinal cord was then completely transected using a pair of surgical scissors. The space between the rostral and caudal ends was packed solidly with a sterile absorbable hemostat (Surgicel, oxidized regenerated cellulose). An antibiotic (Ayercilline, 40,000 IU/kg im) and an analgesic (Anafen, 2 mg/kg sc) were administered at the end of the surgery.
The same method was applied for the second spinal cord lesion at lower lumbar levels. The different levels of spinal transections are illustrated in Fig. 1 (vertical lines, B–H). The spinal cord of the cat is composed of 13 thoracic and 7 lumbar segments. We used bony landmarks to guide the sites of laminectomies and identified the roots before making the second spinal lesion. The level of these lesions in relation to the entry level of dorsal roots was always checked postmortem. We always attempted to make spinal lesions as straight as possible so that there would be little variation in the level of the section from the dorsal to the ventral side of the cord.
POSTOPERATIVE CARE.
Immediately after surgery, the cats were placed in a temperature-controlled incubator. Once the animal regained consciousness, it was returned to its individual cage with full access to food and water. The cage floor was covered with a foam mattress to avoid skin ulceration. Cats were attended daily for general inspection, cleaning of the hindquarters and manual bladder expression (Bélanger et al. 1996).
RECORDING AND ANALYSIS PROCEDURE. After the EMG implantation and before spinalization, locomotor performance (EMG activity synchronized to the video images) in the intact state was recorded. Cats were placed on a treadmill belt equipped with a Plexiglas enclosure and were trained to walk freely at the various imposed speeds.
Spinal cats were trained, without drugs, to walk every day until the animals recovered a well-coordinated locomotor pattern. For spinal locomotion, the forelimbs were placed on a platform located above the treadmill while the hindlimbs walked on the belt. At the beginning of training, perineal stimulation was used, and the experimenter had to support the weight of the hindquarters of the cat. After 2–3 wk, the cats could walk without stimulation and weight support of the hindquarters while the experimenter held the tail merely to provide lateral stability.
The EMG signal was amplified differentially (bandwidth: 100 Hz to 3 kHz) and digitized at 2 kHz; this is adequate to document the amplitude and to establish proper timing between EMG bursts. EMGs were synchronized to the video images of the hindlimbs using a digital time code, recorded in the computer file and the video tape. Reflective markers were placed on the skin overlying the iliac crest, the femoral head, the knee joint, the lateral malleolus, the metatarsophalangeal joint (mtp), and the tip of the fourth toe. Video images of the locomotor movement (left side) were captured by a digital camera and recorded on a videocassette recorder at 30 frames/s.
The bursts of activity of all EMGs were rectified and integrated to quantify their amplitude. Their onset and offset were identified automatically by a homemade software and corrected manually by the experimenter to determine temporal parameters of the EMG for each cycle. The video images were digitized and the x-y coordinates of various joint markers were obtained at the frequency of 60 fields/s, by de-interlacing the video frames taken at 60 Hz. These coordinates were used to calculate angular joint movements and could be displayed as continuous angular displacements or stick diagrams of one-step cycle (see Fig. 2A). The step length was calculated as the distance between two consecutive contacts of the same foot. The time interval between the onset and the offset of each muscle was measured and synchronized to the onset of the St muscle.
|
PHARMACOLOGICAL STIMULATION.
Clonidine was injected ip (150 µg/kg) once, 2–3 days after the first spinal transection (T13) because clonidine has been shown to induce locomotion in spinal cats (Barbeau et al. 1987; Chau et al. 1998b; Forssberg and Grillner 1973). The reason for testing the effect of clonidine at this stage was to serve as a comparison for ulterior lesions after which clonidine was also injected to try and induce locomotion. In cases where cats were unable to walk several days after the second spinal lesion, clonidine was injected once again. It should be made clear, however, that clonidine was not used otherwise to train the cats to walk after spinalization. It should also be stated that the effect of clonidine lasts for a period of only 4–6 h (Chau et al. 1998a), and it is thus unlikely that the injection of the drug only once after spinalization exerts a long-time effect.
Acute experiments
Three chronic cats with the second transection located at L2 (cats B and C) and rostral L3 (cat D) as well as two other cats with only one transaction at T13 were used for acute experiments. These acute experiments aimed at avoiding a third chronic lesion in cats that recovered locomotion after the most rostral second lesion (L2 and rostral L3) and at confirming results obtained in chronic cats with lesions at lower levels.
Under general anesthesia (see Spinalization for details) and endotracheal intubation through a tracheotomy, one carotid artery was cannulated for monitoring blood pressure and the other one was ligated. One jugular vein was cannulated for the administration of fluid and medication. The temperature was measured with a rectal thermometer and maintained 38°C by a feedback-controlled heating element using DC and with heating lamps. The end-expiratory pCO2 was monitored using a Datex Monitor during normal or assisted ventilation and maintained between 3.5 and 4.5%. The blood pressure was maintained within physiological limits. A laminectomy was performed, removing T13–L6 vertebrae, and the cats were then mounted to a stereotaxic frame attached to a motor-driven treadmill. The spine was fixed with three pairs of lateral pins: one at the iliac crest, one at L4–5 and the other at L1. A precollicular, postmamillary decerebration was performed with a spatula and the rostral nervous tissues removed. Anesthesia was then discontinued. The dura was opened and the spinal cord covered with warm mineral oil. The spinal segments were determined by identifying the most rostral and the most caudal dorsal rootlets at each segment.
One hour after the cessation of anesthesia, locomotor capacities of the cat were evaluated on a treadmill at a speed of 0.2–0.3 m/s while using perineal and/or abdominal manual stimulation. All cats received an intravenous injection of clonidine (500 µg/kg). To study the importance of each segment, dorsal funiculi lesion, dorsal hemisection and complete cord transections were made sequentially at different levels between L3 and L5. The same recording and analysis procedures as in the chronic experiments were used.
Histology
At the end of the experiments, the cats were killed with an overdose of pentobarbital sodium and lesions were verified postmortem. Each spinal segment was first identified by the location of the dorsal roots, and after visual examination of the lesion, relevant portions of the cord were removed and fixed by immersion in formalin (10%) for routine histology. Inspection of serial Nissl-stained cross sections (30 µm thick) of the spinal cord was then made for all cats to ascertain the completeness and the rostrocaudal extent of the lesions. Signs of obvious cellular damage due to the second spinal transaction was evaluated under the microscope. The rostrocaudal extent of damage tothe cord resulting from the second spinal transection is given for each cat, except cat B, in Table 1. Overall, in these chronic experiments, an average total of 7.1 ± 1.2 (SD) mm of nervous tissue was damaged by the lesion. In contrast, the sections adjacent to the second lesion in all acute experiments did not reveal any evidence of significant damage.
Statistical analysis
Results are presented as means ± SD for cycle periods and amplitude, and the difference between groups was tested by a one-way ANOVA. Bonferroni's t-test was used as a post hoc test to determine the statistical significance (probability level of <0.05).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
EFFECTS OF THE FIRST THORACIC LESION (T13) ON LOCOMOTION (CATS WITH LESIONS AT LEVEL A ON FIG. 1). Figure 2A represents the typical locomotor pattern of a cat (cat B) in the intact condition before the first spinal lesion at T13. The characteristics of locomotion are reflected in the stick diagrams representing one-step cycles (Fig. 2A, left) and normalized angular plots of the hip, knee, ankle and mtp joints (Fig. 2A, middle). EMG activity of the corresponding sequence is shown to the right of Fig. 2A. EMG activity illustrates a regular alternation of activity on the left (L) and right (R) sides, as well as between flexors (St, Srt) and extensors (VL and GL) on the same sides, organized reciprocally, with respect to muscle activity on the other side.
Eight days after T13 spinalization, this cat was unable to walk when placed on the moving treadmill, even after attempts at locomotor training with perineal stimulation (Fig. 2B). The EMG activity was not rhythmically organized. Now and then, only slight flexion movements were observed and, of course, there was no weight support.
It was important to document the locomotor performance with clonidine at this early stage after a T13 lesion because, in the subsequent phases of the experiment, we intended to use pharmacological stimulation (clonidine) to try and induce locomotion after lesions at lower levels. Within 20 min of an intraperitoneal injection of clonidine (150 µg/kg), a robust locomotor pattern (rhythmic alternation of flexor and extensors muscles) was obtained with perineal stimulation. There was some foot drag at the onset of the swing (Fig. 2C, left) and only partial weight support because the foot was often placed behind the hip. But the cat could walk at imposed speed, ranging from 0.2 to 0.6 m/s, with placement of the foot. However, the duration of the flexor and the extensor bursts was comparable to the intact values obtained before spinalization (St: 283.8 ± 6.9 vs. 295 ± 48 ms; Srt: 349 ± 50 vs. 355.88 ± 100 ms, VL : 787.5 ± 112.5 vs. 604.33 ± 46.7 ms, GL : 786.1 ± 114.0 vs. 544.5 ± 62.7 ms, NS for all previous values at P < 0.05).
Such drug injection was performed only once between the first and second spinal lesions and no training was performed under clonidine. However, this cat was trained on the treadmill for a period of 31 days (3–4 training sessions per day). Figure 2D shows that this cat had recovered spontaneous locomotion by day 21 with proper weight support but that the pattern was somewhat more irregular than in the intact condition. The step length was decreased compared with intact conditions (intact: 453 ± 16 mm vs. T13: 204 ± 37 mm, P < 0.05) or compared with values obtained after the first injection of clonidine, when the cat had not been trained (clonidine: 332 ± 32 mm vs. T13: 204 ± 37 mm, P < 0.05). The cycle duration was correspondingly significantly decreased (intact: 1,253.8 ± 173.9 ms, clonidine: 1,045.3 ± 116.2, T13: 666.9 ± 79.9 ms, P < 0.05). The foot was often placed slightly behind the hip during stepping. This locomotor behavior is consistent with our previous reports (Barbeau and Rossignol 1987; Bélanger et al. 1996; Chau et al. 1998a, b; Rossignol et al. 2000).
EFFECTS OF A SECOND LESION AT THE L2 LUMBAR LEVEL (CATS B AND C).
Several weeks after the first transection at T13 (see Table 1), after cats had regained the ability to walk on the treadmill with adequate weight support, a second lesion was performed at L2 and recovery of locomotion was studied. In one cat (cat B in Fig. 1), the transection was made at the rostral L2 level and, in the other cat (cat C in Fig. 1), at the caudal L2 level. Postmortem histological verification confirmed that the spinal lesions were indeed complete and that the primary and secondary damage that had taken place after spinal injury affected the nervous tissue in the proximity of the aforementioned regions (not shown). For example, necrosis and scar tissues could be identified at the site of the lesion of cat C for a total length of 7.5 mm. Assuming an equal distribution of damage on each side of the lesion, it is likely that 3.75 mm of tissue is affected caudally to the lesion. Given the length of the spinal segment, it is possible that the lesion reached a small portion of rostral L3. Because these cats regained locomotion, their recovery is described in two stages : early and late periods.
Early period.
Two days after the second lesion, both cats B and C were able to walk on the treadmill at speeds ranging from 0.2 to 0.6 m/s with perineal stimulation. No important changes were observed in hip excursion before or after the second lesion, but some decreases were seen in knee excursion as well as in ankle and mtp extension. For cat B, the St burst duration was decreased (–20.5%, P < 0.05) in comparison to the value obtained 31 days after the T13 lesion, i.e., the day before the second lesion. No changes in cycle duration, burst duration of the other muscles, or amplitude were observed after this second lesion when comparing values obtained before (Fig. 2) and after the second lesion (Table 2). There was no change in amplitude right after the lesion, except for the VL (+54.4%, P < 0.05).
|
|
|
Late period.
Cats B and C were trained to walk on the treadmill and, by the end of the 22nd day, these two cats were able to walk at various speeds on the treadmill, without perineal or pharmacological stimulation. The cats showed the same pattern of locomotion as that obtained at T13 at 21 days, except for the knee flexion (Fig. 3C). Nevertheless, we have noticed some tendency to abduct the hindlimbs after the second transection (not shown).
For cat B, when comparing the values obtained after 31 days of recovery after the T13 lesion to those after 27 days of training after the second lesion at L2, we found that the cycle duration remained unchanged (see Table 2), whereas there was a decrease of the Srt burst duration and amplitude (–48 and –27%, P < 0.05). However, a significant increase was observed for the burst duration and amplitude of the VL (+14 and +83%, P < 0.05).
Comparisons of "before and after training" values showed a tendency for the cycle duration (10.7%, NS) and for the flexors and GL burst duration (3–31%, NS) to increase. Only the VL burst duration was significantly higher after training than it was before (+30%, P < 0.05). There was no change of the step length.
In cat C, spinalized at caudal L2, there was a significant decrease of the cycle and burst durations (22–26%, P < 0.05) when comparing values gathered 27 days after the T13 lesion with those obtained 22 days after the second lesion. No significant changes were observed for the amplitude. The St burst duration was higher (+35%, P < 0.05) after training than before, whereas the VL burst duration was shorter (–15%, P < 0.05). However, we noted no change in the step length and the time interval (Fig. 4) between the onset and the offset of each muscle (Srt, VL, and GL) between the values obtained 27 days after the T13 lesion with those obtained 22 days after the second lesion.
In conclusion, the animals that had a previous lesion at T13 and had recovered locomotion could still walk early after a second spinal lesion at L2. However, there were some changes in the locomotor characteristics (cycle and EMG burst duration) after 2 wk of training on the treadmill, which suggests that these higher lumbar segments do exert some role in the expression of the spinal locomotor pattern. Injection of clonidine also modulated the spinal locomotor pattern, by increasing flexion movement and yielding more robust EMG activity.
EFFECTS OF A SECOND LESION AT THE L3 LUMBAR LEVEL (CATS D AND E). In these cats, the second transection was made in the rostral part (cat D in Fig. 1) or the caudal part of L3 (cat E in Fig. 1). There was a major difference in the locomotor recovery between these two levels of transection.
With the rostral L3 lesion (cat D), occasional locomotor activity was observed, but it was much less well organized and less sustained than after a transection at L2. Three days after this second lesion, the injection of clonidine did not initiate locomotor activity (not shown). On the contrary, clonidine induced a state of high excitability of the hindlimbs. For instance, perineal stimulation evoked a response of bilateral hyperextension 
70 min postinjection. Attempts to train this animal to walk on the treadmill in the following week were unsuccessful. Seven days after the transection, perineal stimulation induced bilateral tonic hyperflexion. With strong perineal stimulation, only slight alternating movements were observed. After 11 days of locomotor training and, after another injection of clonidine, a good stepping pattern was established with alternation between flexors and extensors and between the left and right hindlimbs. The cat could perform 15 well-organized consecutive cycles. However, the cat could not follow treadmill speeds >0.3 m/s. Twelve days after this lesion and without clonidine, the cat could walk for several cycles, but the cycles were not well organized and not constant. Furthermore, hyperflexions were often observed. After this experiment (from days 13 to 18), the cat was unable to walk but continued to hyperflex the hindlimbs in response to perineal stimulation.
In cat E, which had a second lesion at the level of caudal L3, all rhythmic locomotor activity that had been regained after the T13 lesion (Fig. 5A) was lost. It is important to specify that the histology revealed that exactly 6.12 mm of tissue was affected by the spinal injury. It is thus possible that some portion of rostral L4 was also damaged by the lesion. Six days after this second transaction, strong perineal stimulation elicited bilateral tonic flexion of the hindlimbs and some small movements but no rhythmic movements (Fig. 5B). Neither could locomotion be triggered with clonidine; in fact, clonidine evoked a strong and maintained bilateral hyperextension that was exacerbated by perineal stimulation. At day 20 after the second spinal transection, another injection of clonidine also produced only bilateral hindlimb extensions. These hyperextensions are illustrated by the tonic EMG activity obtained in extensor muscles and by the stick figures (Fig. 5C). We were unable to induce locomotion despite 3 wk of intensive training. However, vigorous muscle activity was observed during reflex movements by pinching the skin at various locations or during fast paw shake responses (FPS in Fig. 5D and see later).
|
EFFECTS OF A SECOND LESION AT THE L4 LUMBAR LEVEL (CATS F AND G). The second spinal transection at middle L4 (cat F) and caudal L4 (cat G; Fig. 1) also abolished locomotion completely and permanently. When compared with cats lesioned at caudal L3 level transection, there was even less tonic flexion with perineal stimulation or faint locomotor movements of the hindlimbs.
Again, we were interested in finding out if clonidine could induce locomotion as it did after the first lesion at T13. Figure 6 illustrates the effect of clonidine in cat F 6 days after the lesion at T13 (Fig. 6A), then 16 days (Fig. 6B) after the second transection at L4. Whereas it was clear that stepping movements could be enhanced by clonidine after the T13 lesion (Fig. 6A), clonidine only induced a bilateral and sustained hyperextension of the hindlimbs (Fig. 6B), much as what had been seen in cat E after a lesion at caudal L3 (Fig. 5C). Perineal stimulation indeed induced tonic hip extension in both legs. The amplitude of the knee extensor VL was as high (LVL) or higher (RVL) during these hyperextensions after the second lesion as during locomotion evoked by clonidine after the first lesion at T13. This pattern of clonidine-induced bilateral hyperextension was consistent in cats F and G for injections made at 2 and 16 days after the second spinalization. The abolition of locomotor activity persisted despite intensive attempts at locomotor training on the treadmill for several days (see Table 1). Occasionally, we observed some rapid and clonic movements of the hindlimbs with perineal stimulation and weak alternation of the hindlimbs but no locomotor movement.
|
|
FAST PAW SHAKE. We wanted to know if the fact that the spinal cat could not express locomotion after the second transection at L3, L4, or L5 meant that motoneurons were not functional or that the segments below the second lesion had lost all their potential for rhythmogenesis. To that end, we tested another rhythmic activity, the fast paw shake (FPS) response that was elicited in each cat in the intact condition, after the first thoracic lesion and after the second transection by dipping the paw in water (Fig. 8B).
|
), and this synergy was still present after the transection at T13 and after lumbar lesions (see Fig. 8, A–C, top). The frequency of the FPS was 9.92 ± 1.07 Hz in the intact cats. Thirty-five days after the first spinal transection at T13, the frequency was similar to that obtained in the intact condition (9.07 ± 1.41 Hz, mean on 10 cycles/ recording session). Whatever the level of the second transection at the lumbar level, a strong sequence of FPS was observed along with a tendency to increase its frequency compared with the intact condition or to the period after the first T13 lesion. The frequency of the FPS obtained for the cat spinalized at L3 or L4 level tended to be higher (11.39 ± 1.23 Hz, mean on 20 cycles/ test session) than in cats spinalized at L2 (10.98 ± 1.73 Hz, mean on 20 cycles/ recording session), but this difference was not significant. The total duration of FPS episodes also tended to be somewhat longer (but not statistically significant) after a second lesion at L3–L4 (1,124 ± 628 ms) compared with after the first lesion at T13 (963.74 ± 240 ms) or the intact condition (926.58 ± 307 ms). Administration of clonidine abolished the FPS responses, as reported before (Barbeau and Rossignol 1991). In conclusion, after a second transection at the caudal L3, L4, or L5 level, all locomotor activity was abolished, but the cat was able to produce other rhythmic activities such as FPS, which even became somewhat more vigorous.
Serial spinal transections in acute experiments
Because it was not deemed justifiable to perform a third chronic spinal lesion, acute experiments were carried out in three cats that continued to walk after the second lesion, i.e., the two cats spinalized at L2 (cats B and C) and rostral L3 (cat D). These acute experiments also allowed us to corroborate the results obtained in the chronic state. Furthermore, it was possible to perform progressive dorsoventral lesions at different lumbar levels; this gave us an indication of the importance of dorsal and ventral pathways. In two other cats spinalized only at T13, we also investigated the effect of serial and progressive spinal transections.
After the decerebration and removal of the anesthesia, none of the cats displayed spontaneous locomotor activity, but reflexes could be elicited by pinching the skin of the hindlimbs. It should be remembered that, in this protocol, the cats are pinned firmly to a stereotaxic frame (see METHODS), which can prevent spontaneous locomotion. Administration of a higher dose of clonidine (500 µg/kg iv) was thus required, in these acute experiments, to generate locomotion in these cats.
After the establishment of the locomotor rhythm with clonidine, successive cuts were made to the spinal cord caudally to the last chronic lesion. In cats that had been chronically spinalized at L2, successive lesions were made at rostral L3, mid L3, caudal L3, rostral L4, mid-L4, etc. At each level, the following sequence of lesion was strictly observed: removal of the dorsal funiculi, dorsal hemisection, and complete transection of the cord. After each lesion, the locomotor capacity of the animal was assessed for 20–30 min before performing the subsequent lesion and until locomotion was totally abolished. Further injections of clonidine were always made when needed to make sure that the absence of locomotion was not due to a lack of central stimulation. For the cat spinalized at rostral L3, the same paradigm was observed but the lesions were initiated at caudal L3 and at rostral L2 for the two cats spinalized at T13.
SPINAL T13–L2/L3 CATS. Figure 9 illustrates the effect of progressive dorsoventral lesions at different segmental levels of cat C that was previously chronically spinalized at caudal L2. Figure 9A illustrates the locomotor activity obtained 30 min after injection of clonidine and before the additional lesions. Normal amplitude of flexion of the hip, knee, and ankle, plantar foot placement (Fig. 9A, right) as well as a robust EMG activity with a left-right alternation (Fig. 9A, left) was obtained. This locomotion can be compared with the one obtained in this cat during the chronic experiments before decerebration (compare with Fig. 3D)
|
In the cat that was chronically spinalized at rostral L2 (cat B), locomotion was obtained until a complete transection at rostral L3 was performed. After this transection, all rhythmic activity was abolished (not shown). For the cat chronically spinalized at L3 (cat D), abolition of locomotion was obtained after a complete lesion at mid-L4.
SPINAL T13 CATS. The two cats spinalized at T13 were trained for 5 wk and recovered locomotion in the hindlimbs before the acute experiment. They were not submitted to a second chronic spinal lesion but only additional transections in the acute experiment. One hour after decerebration, locomotion over the treadmill was assessed before the second transection was performed. After intravenous injection of clonidine, a rhythmic activity of the hindlimbs was obtained with perineal stimulation at a speed of 0.3 m/s with alternation of the left and right side and with flexion and extension of the knee and the ankle. The amplitude of the swing and of the stance phases was decreased compared with that obtained before the injection. When an adequate locomotor pattern was obtained, a second transection was made. For the two cats, the first second lesion was done at rostral L2. This transection did not affect the locomotor performance. The cats were still able to walk on the treadmill with a proper alternation of the hindlimbs and a robust EMG activity. Locomotor activity persisted after transections at the rostral L3 level and at mid-L3 level, but there was no foot placement and the foot still landed behind the hip. But in one cat, we removed the vertebral pins used for fixation, and a well-organized locomotor pattern was obtained with amplitude of the swing and the stance phase comparable to those observed in the chronic state (flexion and extension of all joints).
All rhythmic activity was abolished when the transection was performed at the rostral L4 level in one cat and at the mid-L4 level for the other cat. At this time, only strong movements of the hindlimbs and hyperflexion, but no locomotor movement was observed. A last injection of clonidine was applied to ensure that the absence of locomotion was not due to a lack of central stimulation. In these cats, FPS could not be obtained because it is abolished by the injection of clonidine as mentioned earlier. Therefore the acute experiments confirm the results obtained with the chronic lesions. Spinal locomotion can be obtained in cats previously spinalized at T13 until the L3–L4 segments are disconnected from the lower lumbar levels.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
TREADMILL LOCOMOTION.
The main conclusion of the present study is that spinal locomotion of the cat on a treadmill relies critically on the integrity of the L3–L4 spinal segments of the spinal cord. It is well known that a few weeks after spinalization at T13, kittens and adult cats can walk on a treadmill with plantar foot contact during stance and good weight support of the hindquarters (Barbeau and Rossignol 1987; Bélanger et al. 1996; Chau et al. 1998a; de Leon et al. 1998; Forssberg et al. 1980; Lovely et al. 1990; Rossignol 2000). Lesions at other spinal levels have rarely been reported. In kittens, one study (Shurrager and Dykman 1951) mentions a cat spinalized at L3 that could walk, but unfortunately no details were given. A study using chronic double lesions of the cord (Afelt 1970) concluded that there is no locomotion with a spinalization below L4, although other simpler reflex behaviors were seen (rhythmic patterns such as FPS were not reported). The present work shows that a second spinal transection that disconnects the segments above L4 from the more caudal lumbar segments abolishes spinal locomotion, even after 2–4 wk of sustained attempts at locomotor training on the treadmill. Our previous results had reported that a spinal transection at L4 abolished all locomotion in acute preparations (Marcoux and Rossignol 2000). This condition cannot be interpreted as being due to a secondary spinal shock after the second spinal lesion (which should have subsided after a week or so) but is due instead to the disconnection of the mid-lumbar segments with the more caudally located motoneurons. Because in the present report, spinal locomotion was still possible after the second lesions were performed at L2 or rostral L3, it can be concluded that merely performing a second lesion caudal to T13 is not sufficient to abolish locomotion but that the specific inactivation of caudal L3 and L4 is necessary.
FICTIVE RHYTHMIC PATTERNS.
A study in paralyzed cats, using a combination of lesions and a cooling probe to inactivate specific spinal segments, has demonstrated that "fictive scratching" was strongly dependent on mid-lumbar segments L3–L5 (Deliagina et al. 1983), although rhythmic oscillations could be obtained in the isolated L5 segment. In acute curarized spinal cats (T12–13) injected with nialamide and L-DOPA, rhythmic alternating activity in the TA and GL nerves remained, even after a lesion at rostral L5, when only segments L6, L7, and S1 were intact (Grillner and Zangger 1979). This shows clearly that the L5–S1 segments have the necessary interneuronal circuitry to produce spinal rhythmicity. In this acute experimental condition, it is probable that the chemical stimulation produced by DOPA through the still-functional monoamine terminals may indeed mimic the inputs normally received by supraspinal pathways.
PHARMACOLOGICAL STIMULATION.
We have not studied DOPA in our chronically spinalized preparations because, in principle, the noradrenergic terminals have degenerated, and, therefore DOPA, the precursor of NA that needs to be taken up by the terminals to release NA, would be inefficient. Accordingly, we have studied only the effect of an alpha-2 receptor agonist, clonidine, which has been shown to induce locomotion in spinal cats in many previous studies (Barbeau et al. 1987; Chau et al. 1998b; Forssberg and Grillner 1973; Giroux et al. 2001). In the present study, clonidine injected in cats with a single chronic lesion at T13 also induced robust locomotion with marked effects on the overall step cycle but especially on flexor burst duration.
It was striking and unexpected to observe that clonidine injected after a second lesion at L3, L4, or L5 induced a tonic bilateral hyperextension on the treadmill only when combined with perineal stimulation. These puzzling effects suggest that segmental distribution of receptors may account for this finding. Indeed, it can be postulated that, if receptors for various transmitters are unevenly distributed along the length of the spinal cord (Rossignol et al. 2002), transmitter agonists may induce different pharmacological effects depending on the level of the transection. Such a regional difference in the activation of locomotion by 5-HT was shown in the neonatal rat (Cowley and Schmidt 1997). It is possible, therefore that the net effect of clonidine on the low lumbar spinal cord may favor extensor hyperexcitability, while the effect on the upper lumbar cord may have a net effect that is biased toward the flexors. Because various types of spinal interneurons are sensitive to adrenergic stimulation (Hammar and Jankowska 2003), it is indeed tempting to speculate that the effects of pharmacological stimulation may depend on the segmental distribution of receptors for a given neurotransmitter. Clearly, however, the effect of clonidine on segments L5–S1 seems to be predominant in the extensor muscles and are nonlocomotor in the chronic spinal cat.
Mechanisms of locomotor abolition after L4 lesions
DAMAGE TO AFFERENT INPUTS.
One mechanism that could be implicated with the L3–L4 transection is a partial deafferentation from proximal joints, muscles, or skin. After a unilateral hindlimb deafferentation (caudal to L2 or L3 in decerebrate but not spinal cats), it was reported that hindlimb locomotion could be abolished (Orlovsky and Feldman 1972), although lumbar interneurons still discharged rhythmically in the absence of forelimb movements. In decerebrate cats with mesencephalic stimulation (MLR), locomotion could be induced after unilateral deafferentation even though the pattern was more labile (Grillner and Zangger 1984). Here again, inputs from the forelimbs could have played a role in providing timing cues to the deafferented hindlimb. However, deafferentation alone, involving L3–L4, does not seem to be sufficient to abolish locomotion in decerebrate cats.
Some studies have looked at deafferentation in spinal cats. In one study, unilateral lumbar deafferentation was done first in otherwise intact cats, and they recovered locomotion. After performing a hemisection on the deafferented side, the cats stopped walking (Goldberger 1977). On the other hand, complete spinal cats that have recovered locomotion will still be able to perform some stepping movements on a the treadmill after a unilateral deafferentation (Giuliani and Smith 1987).
Although we cannot discard categorically that our lesions at caudal L3 and L4 abolish critical sensory inputs (for instance from hip muscle, joint, or skin afferents), the above deafferentation experiments involving L3–L4 dorsal roots lead us to believe that this cannot be the main mechanism, suggesting that deafferentation alone of the L3–L4 segments might not be sufficient to explain the total absence of locomotion reported here, after a chronic lesion at caudal L3 or L4.
Damage to motoneurons
One obvious concern in such experiments is whether the second lesion could induce damage to motoneurons, thus preventing any motor pattern. First our histological assessment of the second lesion clearly indicated damage of 7 mm, including the lesion itself and damage above and below the lesion. This could hardly explain how a lesion at L4 could abolish activity in motoneurons of L7–S1. However, interestingly, it might explain that lesions close to the L3–L4 border may have variable effects that may evolve with time and the progressive extent of the damage that could gradually affect L4. Furthermore, when comparing the levels of transection necessary to abolish locomotion in acute and chronic spinal cats, it was observed that a slightly higher number of caudal lesions in the acute experiments was needed to abolish locomotion. This may suggest that, in the chronic cats, more rostral lesions induce damage lower down in the L4 segments.
However, the following results are even more compelling. Although there was no spinal locomotion after the second spinal lesions below caudal L3, vigorous fast paw shakes could still be elicited and tend to be even more vigorous. The characteristic synergy of TA and VL co-activation (Pearson and Rossignol 1991; Smith et al. 1985) was preserved after the first and the second spinal transections. This shows that most motoneurons controlling hindlimbs muscles of the cat, located in L5–S1 segments (Sherrington 1892; Vanderhorst and Holstege 1997) were not significantly damaged by the second lesion, except in cat H, lesioned at L5. Although stimulation of L4 leads to abdominal contractions and hip flexion (Sherrington 1892), hip flexor muscles such as Ip or Srt also receive an innervation through the L5 ventral root, so that a transection at caudal L3 or L4 should not denervate these proximal muscles. The abolition of spinal locomotion reported here, therefore is not a consequence of damage to motoneurons but to a lack of their recruitment into a locomotor pattern. This finding also establishes that, despite the absence of locomotion, the spinal cord still has a strong rhythmogenic potential in segments below L4 and that the circuitry responsible for FPS is largely distinct from that of the locomotor CPG. Because we are looking at actual behaviors, we can be certain that the rhythms observed correspond to distinct behaviors; this might not always be the case when looking at recordings in paralyzed preparations.
Damage to interneurons
The most likely cause of the abolition of locomotion after our secondary lesion is a disconnection of neural circuits in the midlumbar segments that are important for spinal locomotion. The spinal cord is endowed with a rich network of propriospinal interneurons (Kostyuk and Vasilenko 1979), which make excitatory and inhibitory connections with motoneurons through the DLF (Kostyuk et al. 1971). In the cat, some interneurons, in the intermediate zone and in the ventral horn of mid-lumbar segments (mainly L4) of the spinal cord, receive dominant inputs from group II muscle spindle afferents (especially quadriceps and Srt) and project to motoneurons in more caudal segments through the lateral funiculi (Edgley and Jankowska 1987). It was proposed that these mid-lumbar interneurons (especially the ventral group) might play a role in locomotion (Edgley et al. 1988) and were shown directly to be activated during MLR-evoked fictive locomotion in decerebrate cats (Shefchyk et al. 1990). Other commissural interneurons located at mid-lumbar levels and that normally receive supraspinal inputs (such as reticular formation) and project to contralateral motoneurons (Jankowska et al. 2003) could also be implicated. It is thus possible that all these groups of interneurons located rostrally in the lumbar cord may play a role in locomotion and that after spinalization some may actually become even more critical for the expression of locomotion. This is in keeping with suggestions made by others (Jordan and Schmidt 2002; Shik 1983) that chains of spinal propriospinal interneurons may play a key role in organizing locomotion. Our present work would suggest that the L3–L4 interneurons are critical in that chain and may actually be the last representative of that chain after chronic spinal transection at T13. We do not infer here that the CPG for locomotion is located at L3–L4 but rather that these segments provide critical inputs for the function of the locomotor CPG. Obviously, it is hard at this stage and with this approach to determine what constitutes a CPG and what characteristics should identify unequivocally the critical role of certain neurons in the generation of locomotion. The rhythmogenic capabilities of the whole cord is clear. However, the recruitment of motoneurons in various different rhythmic patterns must depend on specific inputs even in spinal cats that can display several different rhythmic patterns. We suggest here that, in chronic spinal cats, this critical organizing input for locomotion originates from the mid-lumbar segments.
Locomotor-like activity can also be elicited in in vitro neonatal preparations (Kudo and Yamada 1987). It was suggested (Bertrand and Cazalets 2003; Cazalets 2000; Cazalets and Bertrand 2000a, b; Cazalets et al. 1995) that the thoraco-lumbar cord is critical for the generation of locomotor rhythms, although rhythmogenic capabilities were found in lower lumbar segments (Kudo and Yamada 1987) and sacral segments (Bonnot and Morin 1998; Bonnot et al. 1998; Cazalets and Bertrand 2000a; Lev-Tov et al. 2000). A modified version of this hypothesis was formulated on the basis of the distribution of receptor subtypes capable of inducing different rhythmic patterns (Jordan and Schmidt 2002). Other studies in neonatal rats (Bracci et al. 1996; Ho and O'Donovan 1993; Kjaerulff and Kiehn 1996) have suggested that the whole spinal cord has rhythmogenic potential but that the upper lumbar segments appear to have a leading role. An excellent review on this subject has been published recently (Clarac et al. 2004). Furthermore, intraspinal injections of kainic acid into the T9 and L2 segments to destroy the gray matter only, while sparing the white matter of the spinal cord of rats (Magnuson et al. 1999), induced paraplegia. Finally, adult spinal rats receiving grafts of 5-HT mesencephalic embryonic cells below the spinal lesion at T8 recovered spinal locomotion only if the 5-HT terminals re-innervated the rostral L1–L2 segments (Gimenez y Ribotta et al. 2000).
It can thus be concluded that the upper lumbar segments in rats (L1–L2) and mid-lumbar segments in cats (L3–L4) play a crucial role in the organization of the spinal locomotor pattern. An appealing consequence of this work is the possibility to target circumscribed critical spinal segments for pharmacological stimulation (Marcoux and Rossignol 2000), electrical stimulation (Barthélemy et al. 2002), or cell grafts (Gimenez y Ribotta et al. 2000), to induce locomotion after spinal cord injury.
|
GRANTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
FOOTNOTES |
|---|
Address for reprint requests and other correspondence: S. Rossignol, Centre de Recherche en Sciences Neurologiques, Pavillon Paul-G.-Desmarais, C.P. 6128, Succ. Centre-ville, Montreal, QC H3C 3J7, Canada (E-mail: Serge.Rossignol{at}umontreal.ca)
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Barbeau H, Julien C, and Rossignol S. The effects of clonidine and yohimbine on locomotion and cutaneous reflexes in the adult chronic spinal cat. Brain Res 437: 83–96, 1987.[CrossRef][ISI][Medline]
, paw contact; 
, paw lift; horizontal line, stance phase. Flexion always corresponds to downward deflections of the angular traces. Right: raw electromyographic (EMG) activity of the flexor and extensor muscles of the left (L) and right (R) hindlimb during locomotion. LSt, left Semitendinosus; LSrt, left Sartorius; LVL, left Vastus Lateralis; LGL, left Gastrocnemius Lateralis; RSt, right Semitendinosus; RSrt, right Sartorius; RVL, right Vastus Lateralis; RGL, right Gastrocnemius Lateralis. The scale bar equals 1 s and applies for all EMG traces.
