Recovery of locomotion in the cat following spinal cord lesions

Brain Research Reviews 40 (2002) 257–266www.elsevier.com/ locate/brainresrev

Review

R ecovery of locomotion in the cat following spinal cord lesions* ´S. Rossignol , L. Bouyer, D. Barthelemy, C. Langlet, H. Leblond

´ ´ ´ ´Centre de recherche en sciences neurologiques, Pavillon Paul-G.-Desmarais, 2960 Chemin de la Tour, Universite de Montreal, Montreal (Quebec),Canada H3T 1J4

Abstract

In most species, locomotor function beneath the level of a spinal cord lesion can be restored even if the cord is completely transected.This suggests that there is, within the spinal cord, an autonomous network of neurons capable of generating a locomotor patternindependently of supraspinal inputs. Recent studies suggest that several physiological and neurochemical changes have to occur in theneuronal networks located caudally to the lesion to allow the expression of spinal locomotion. Some evidence of this plasticity will beaddressed in this review. In addition, original data on the functional organisation of the lumbar spinal cord will also be presented. Recentworks in our lab show that segmental responsiveness of the spinal cord of the cat to locally micro-injected drugs in different lumbarsegments, in combination with complete lesions at various level of the spinal cord, suggest a rostro-caudal organisation of spinallocomotor control. Moreover, the integrity of midlumbar segments seems to be crucial for the expression of spinal locomotion. These datasuggest that the regions of critical importance for locomotion can be confined to a restricted portion of the spinal cord. Later, thesemidlumbar segments could be targeted by electrical stimulation or grafts to improve recovery of function. Understanding the changes inspinal cord neurophysiology and neurochemistry after a lesion is of critical importance to the improvement of treatments for locomotorrehabilitation in spinal-cord-injured patients. 2002 Elsevier Science B.V. All rights reserved.

Theme: Motor systems and sensorimotor integration

Topic: Spinal cord and brainstem

Keywords: Central pattern generator; Plasticity; Midlumbar segments

Contents

1 . Introduction ............................................................................................................................................................................................ 2572 . Locomotor training and recovery in complete or partial spinal cats.............................................................................................................. 2593 . Plasticity of spinal locomotor controls....................................................................................................................................................... 259

3 .1. Tibialis anterior and extensor digitorum longus lesions ...................................................................................................................... 2603 .2. Lateral gastrocnemius-soleus lesions................................................................................................................................................. 2603 .3. Cutaneous denervation of the hindfeet .............................................................................................................................................. 261

4 . Functional organisation of the spinal cord with respect to locomotion.......................................................................................................... 2615 . General conclusions................................................................................................................................................................................. 264Acknowledgements ...................................................................................................................................................................................... 264References................................................................................................................................................................................................... 264

1 . Introduction

Major efforts have been made to try to understand therespective roles of the spinal cord, supraspinal structures*Corresponding author. Tel.:11-514-343-6366; fax:11-514-343-and peripheral afferents in the control of normal locomo-6113.

E-mail address: [email protected](S. Rossignol). tion and these studies have been summarized quite exten-

0165-0173/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0165-0173( 02 )00208-4

mailto:[email protected]

258 S. Rossignol et al. / Brain Research Reviews 40 (2002) 257–266

sively [1,34,48,57]. Fig. 1 outlines the general framework It is thought that this CPG can be activated throughof our current understanding on how animal locomotion is different descending pathways coursing through the ven-generated and controlled. The locomotor pattern (see insert trolateral (VLF) or dorsolateral (DLF) pathways throughof EMGs) is rather complex and involves the sequential propriospinal interneurones (PS). Fig. 1 also illustrates thebilateral activation of muscles acting at different joints, spinal cord in a pseudo 3D configuration to indicate that, ineach having more or less its own characteristic signature of the cat and probably in most limbed animals, there is also aamplitude and timing. Within the spinal cord, networks of segmental organisation of the locomotor generating mecha-interconnected interneurones constitute a central pattern nisms whereby rostral and caudal segments may playgenerator (CPG) capable of activating motoneurones in an different roles. Agonists and antagonists of neurotrans-appropriate sequence as well as setting the excitability of mitters can also activate, block or modulate the activity ofother types of interneurones involved in transmitting this CPG. Finally, peripheral afferent inputs do not appearinformation from descending pathways and sensory affer- to be essential for the generation of the basic rhythments so that corrections through these pathways are com- although they are crucial for the appropriate and adaptedmensurate with the various phases of the locomotor cycle. expression of the locomotor pattern.

Fig. 1. General framework of locomotor control in the cat. The EMGs of the inset at the bottom center are taken from Semitendinosus (St, a kneeflexor /hip extensor), Sartorius (Srt, a hip flexor), Vastus Lateralis (VL, a knee extensor), Gastrocnemius Medialis (GM, an ankle extensor) and Tricepsbrachialis (Tri, an elbow extensor); L and R refer respectively to left and right. DRG, dorsal root ganglion; CPG, central pattern generator; E, extensormotoneurones; F, flexor motoneurones; VLF, ventrolateral funiculus; DLF, dorsolateral funiculi, PS, propriospinal interneurones; L3–S1 refer loosely tovarious lumbar and sacral spinal segments; NE, norepinephrine; 5-HT, serotonine, DA, dopamine, Glu, glutamate. A more complete description of thevarious control pathways are found in Ref. [48].

S. Rossignol et al. / Brain Research Reviews 40 (2002) 257–266 259

Although the normal controls of locomotion represent an sible since the cats only have weak movements withoutexquisite and dynamic balance between different levels of weight support during that period. In four cats, sponta-the nervous system, after a large or after a complete lesion neous locomotion (i.e. not drug-induced) with plantar footof the spinal cord, many animal species can re-express placement and weight support of the hindquarters could belocomotion of the hindlimbs below the spinal lesion obtained after as little as a week of training with the drug;[48,49,51,52]. These studies show a remarkable ability of the performance of these cats remained thereafter good forthe spinal cord to optimize the locomotor function within several weeks.the remnant structures of the CNS and this, in turn, What is the effect of training? Edgerton and his col-suggests plasticity of locomotor controlling mechanisms. leagues [18–20,24,26,36] have concluded that trainingWe will review here some of the direct and indirect exerts its effects mainly through central mechanisms andevidence suggesting that, after a lesion, the spinal cord is not peripheral muscular mechanisms and that cats trainedgradually modified and eventually regains the ability to specifically to walk will walk better than stand whereasexpress locomotion, among other motor functions. cats trained to stand still will stand better than walk. This

We will first review some of the evidence of recovery of suggests some degree of plasticity within the spinal cordfunction after spinal lesion in the cat, then address more which eventually leads to the ability to walk after a spinaldirectly the question of spinal locomotor plasticity and lesion.finally discuss some new work which suggests a spinal Is spinal locomotion learned? Spinal locomotion can berostro-caudal organisation of spinal locomotor control. expressed very early after spinalisation with noradrenergic

drugs, either on a treadmill or in ‘fictive’ conditions i.e.after paralysis [28,35]. The spinal cord, when acutely

2 . Locomotor training and recovery in complete or separated from supraspinal structures, can therefore, all bypartial spinal cats itself, generate a locomotor pattern without having to

‘relearn’ walking. Kittens spinalised before having learntSpinal animals can step with the hindlimbs after a few to walk, can walk with their hindlimbs on a treadmill

weeks without any special training. However, regular [29,30]. This shows that the basic locomotor pattern istreadmill training placing the hindfeet on the belt while the innate and does not require to be ‘learnt’ for its basicanimal maintains its forequarters on a stationary platform, expression.will improve significantly the quality of walking and the However, training can alter the expression of thisability to support weight of the hindquarters during pattern; it can accelerate the rate of locomotor recovery orwalking [4–6,22,25,41,58]. Initially, the hindlimbs are favour standing over walking as already mentioned. Younglargely extended and the feet drag on the dorsum while spinal rabbits can be trained to either express an alternatesome weak alternating movements can be observed at the or an in-phase coupling of the hindlimbs [65]. Since bothhip. These weak movements are incapable of bringing the forms of coupling modes exist in the rabbit, the patternsfoot forward past the vertical projection of the hip joint to are not, as such, learned but one or the other mode ofthe ground. Cats can regain a secure plantar foot contact coupling becomes predominant after training. The spinalafter a few weeks of training on the treadmill. cord is probably incapable of learning completely new

Cats with large but incomplete lesions of the spinal cord motor patterns which are not, to some degree, alreadycan also eventually regain voluntary quadrupedal locomo- extant. In that sense the spinal cat does ‘relearn’ to walk.tion. With lesions of the dorsal column and dorsolateral Nevertheless, it seems that the spinal cord can, withquadrants [37], cats can regain quadrupedal locomotion training, by itself and in isolation from descending path-with relatively minor deficits mainly manifested as a foot ways, modify the expression of the locomotor patterndrag at the onset of swing as well as an inability to when having to deal with changes in sensory inputs oradequately clear an obstacle placed in its path of walk on even changes in the peripheral neuro-muscular apparatus.the treadmill (see also Drew, this issue). With ventral and This adaptation of spinal locomotion, induced by repetitiveventrolateral lesions [12,52] cats behave as complete spinal training, may be seen as a form of ‘learning’ by the spinalcats for several weeks but then they gradually regain the cord. In the next section, we will summarize some of theability to stand and walk voluntarily with all four limbs, work which indeed suggests rather directly that adaptivethe forelimbs becoming more propulsive than the hin- changes occur in the cord after spinalisation and, further-dlimbs, contrary to the normal cat. more, that even in the spinal state, cats can up to a certain

Since the alpha-2 adrenergic agonist clonidine can limit adapt to supplementary lesions made, for instance, toinduce locomotion with plantar foot contact and ample peripheral nerves.movements for several hours [3,17,28], we have used thatwindow of opportunity to train cats every day after aninjection of clonidine (either i.p. or i.t.) during the first 3 . Plasticity of spinal locomotor controlspost-spinalisation week [2,16]. This allowed for an inten-sive early locomotor training which is otherwise impos- Many early studies on locomotor plasticity were made


using tendon transfer to see if the spinal cord could change3 .1. Tibialis anterior and extensor digitorum longusthe timing of activity within the step cycle of muscles lesionswhose biomechanical actions were changed by the tendontransfer [62]. Sperry pointed out that animals could We first sectioned the innervation of major ankle flexorsactually develop compensatory strategies using intact un- Tibialis Anterior (TA) and Extensor Digitorum Longustransposed muscles unless they were forced to use almost (EDL) on one side of the animal (spinal cord intact) [13].exclusively the transposed muscles, in which case there A detailed kinematic analysis revealed a significant reduc-was little or no adaptation of the muscle usage [59–61]. tion of ankle flexion albeit not complete. EMG analysesThis hardwiring of the locomotor network was also seen showed some increase in amplitude of the knee flexorby others [31]after transposing the proximal portion of the Semitendinosus (St) and/or the hip flexor Sartorius (Srt)lower tendon of the medial gastrocnemius (MG) muscle which were probably enough to lift the foot during swingwith the distal part of the cut tibialis anterior (TA) tendon. despite the lack of ankle flexion. These changes wereEven after 3 years, one cat still showed locomotor deficits altogether quite discrete so that cats walking on theas seen from EMG recordings and kinematic analysis and treadmill appeared more or less normal. In two cats, anever changed the timing of activation of MG from spinalisation was then performed at T13 and the cats wereextension to flexion. trained to walk for several weeks. Contrary to the usual

On the other hand, there is some evidence of spinal locomotor performance of spinal cats [4,6,16], theseplastic changes after tendon transfers. After transposing animals had a quite abnormal asymmetrical walk withwrist muscles, cats showed gradual locomotor compensa- hyperflexions of the denervated limb at the hip and kneetion [70]. After crossing Soleus and EDL muscle nerves and foot contact on the dorsum without weight support. As[42] it was shown that the phase of activity of these a control, one other cat was first spinalised and thenmuscles during locomotion became variable after crossing, neurectomized. Contrary to cats first neurectomized andwhich was attributed to the combination of a lack of then spinalised, this cat continued to walk symmetricallyproprioceptive information or some degree of central after the procedure albeit with an absence of ankle flexionremodelling. leading to foot drag and an increase in proximal flexors.

Some changes in the simplest connections between There was no hyperflexion nor disorganized asymmetricalafferents and motoneurones (the monosynaptic EPSPs) hindlimb locomotor movements as observed in the previ-were also found long-term after antagonist nerve crossing ous cats.[23]. After crossing the proximal end of the Peroneus (Per) It can be concluded from these experiments that, inmuscle nerve with the distal end of the MG muscle in normal cats, there are changes at different levels of the1–25-day-old kittens, intracellular recordings were made CNS during adaptation to a denervation. That there arein Per and MG motoneurones 21–38 weeks later. Some changes occurring in the spinal cord itself is revealed bynew connections were made between previously antagonist the quite abnormal expression of the locomotor patternmuscles, but the change was incomplete. This work clearly after spinalisation since normally the spinal locomotorshowed that muscle transposition would induce some pattern is quite regular. By removing supraspinal inputsdegree of plastic changes in the spinal connectivity, even one can thus reveal the changes that have occurred withinthough the authors tended to support Sperry’s view. the cord itself and show that, in absence of the supraspinal

In conclusion, locomotor adaptation to muscle or nerve inputs, the changes could not compensate for the lesion.transpositions across a joint represents a very demandingtask and, in most cases, animals seem to develop compen-satory strategies by recruiting other muscles to produce 3 .2. Lateral gastrocnemius-soleus lesionsmore or less the same biomechanical function rather thanchanging activity in the transposed muscles, thus sug- Pearson and collaborators [67,69] cut the muscle nervegesting a rather limited locomotor plasticity, in the classi- to the Lateral Gastrocnemius Soleus (LGS) and Soleuscal sense. One major question in this context of spinal (SOL) and studied the locomotor recovery of these other-learning is then: are spinal animals capable of developing wise intact cats with kinematic and EMG recordings. Theysuch compensatory tricks? If so, this compensation could saw a rapid kinematic compensation, as previously re-certainly be considered as a form of spinal plasticity. Our ported [66]. They also showed increased activity in at leastgeneral approach to study directly the locomotor plasticity one of the remaining ankle extensors MG, and alsoof spinal animals has been to observe the changes occur- changes in polysynaptic reflex pathways (as tested indirect-ring in the locomotor pattern as a function of time after ly by measuring effectiveness of increases in step cycleperipheral denervation of flexor, extensor or cutaneous duration [68]). Furthermore, in two out of five animals,nerves of the hindpaws. In all cases, we were able to show these changes were retained after acute spinalisation [69].some spinal component to the compensation after peripher- Together with Whelan and Pearson [11], we have studiedal nerve section. similar denervations in chronic spinal cats. The nerve


lesion caused a yield of the ankle during stance. Re- to some degree to compensate for supplementary lesions ofmarkably, the kinematic locomotor recovery observed after peripheral nerves [9,10].the LGS lesion was similar, both in time course and inextent of recovery to that seen in intact cats [50]. There-fore, the spinal locomotor circuitry can detect a musclenerve lesion and induce a locomotor compensation to 4 . Functional organisation of the spinal cord withreinstate the normal locomotor kinematics. respect to locomotion

The studies reported in the two previous sections3 .3. Cutaneous denervation of the hindfeet establish quite clearly that spinal animals can express a

locomotor pattern and that probably several types of plasticThe experimental series so far described involved changes occur in the cord resulting in the capability to

damage to muscle nerves and thus direct damages to the express locomotion. We have reviewed elsewhere some ofeffector. In the present series, we restricted our lesions to the neurochemical changes occurring in the cord aftercutaneous nerves of the hindpaws [7–9]. Combined spinalisation [33,53] and will not comment on these in thekinematics and EMG recordings were obtained from cats present chapter.trained to walk regularly on a treadmill, before and after a However, several lines of evidence now suggest thatcomplete, bilateral removal of cutaneous inputs from the some regions of the spinal cord may come to occupy ahindlimb paws. The following nerves were cut: superficial central role in the expression of locomotion after spinalisa-peroneal (SP), tibial (Tib), caudal cutaneous sural (CCS), tion. Firstly, a series of experiments in spinal rats showedsaphenous (Sap), and the cutaneous branch of the deep that grafts of E14 mesencephalic cells below the lesion siteperoneal nerve (DPc). Two days following this major could facilitate the expression of locomotion if, and onlychange in sensory feedback from the hindpaws, these if, the 5-HT re-innervation reached the L2 level [32]. Thisotherwise intact cats walked normally on the treadmill with is in agreement with the work of others [14,15]. We thenonly small kinematic changes accompanied by modest but sought to see if locomotion could be triggered in decere-consistent increase in EMG activity of knee and ankle brated cats spinalised 1 week earlier by microinjectingflexors. These cats could no longer walk across an clonidine in restricted rostral lumbar segments [43]. In-horizontal ladder for 3–6 weeks but eventually even deed, we found that these localized injections could triggerregained that capability, at which point they were spinal- locomotion of the hindlimb on the treadmill especiallyised. when applying perineal stimulation and that this locomotor

Contrary to the non-denervated spinal cats, spinal cats pattern was not improved by a further i.v. injection ofwith cutaneous denervation of both hindpaws could no clonidine, suggesting that the full locomotor capabilitylonger place the foot properly on the treadmill belt. They could be triggered by the restricted injections. Further-dragged their legs across the treadmill, keeping the dorsum more, after i.v. injections of clonidine, an injection of theof the foot in contact with the belt at all times (forward alpha-2 noradrenergic receptor blocker yohimbine in themovement during swing and backward movement during same restricted L3–L4 segments, could block locomotion.stance) and performed rapid movements of small am- More recent experiments in curarized cats have shownplitude. that microinjections of clonidine can induce ‘fictive’

In another chronic cat, we left one cutaneous nerve with locomotion with restricted injections to the rostral lumbara small receptive field intact, the DPc. After spinalisation segments (Leblond and Rossignol, unpublished). However,however, the cat was trained to walk and eventually this is much more difficult when there are no afferentregained plantar foot contact, thus showing a remarkable inputs since, in many experiments, there was no activityability,in the spinal state, to adapt its locomotion to a new under curare but the animal could walk later on theconstraint imposed by a major reduction of sensory input. treadmill when the curare wore off. On the other hand, weAfter cutting the remaining DPc nerve, however, the cat could also show that fictive locomotion induced by l-permanently lost the ability to place the foot on the plantar DOPA could specifically be blocked by applying yohim-surface, similarly to the cats which had had the complete bine in the rostral segments.cutaneous denervation in a single surgical procedure. In During real or fictive locomotion evoked in cats spinal-some cats, we did a progressive cutaneous denervation of ised already 1 week before at the T13 level, we have alsoone paw after the cats had regained treadmill walking. evaluated the effects of a second acute spinal section at L3After each nerve cut, there was a recovery of locomotion or L4 (Fig. 2). Fig. 2A–C illustrate firstly the EMGafter a few days until all nerves were cut, in which case the recordings of a bout of treadmill walking in the spinal catspinal cat could no longer place the foot properly as did (T13) after an i.v. injection of clonidine. A second lesioncats neurectomized before spinalisation. This suggests was then performed at L3 and this markedly deterioratedtherefore that even in the spinal state, cats have the ability the walking pattern which slowed down and became


Fig. 2. Effects of supplementary spinal lesions on locomotion. This figure compares hindlimb EMGs in a walking cat (A–C) and hindlimb ENGs in aparalyzed cat. (A) Treadmill locomotion was triggered by intravenous administration of clonidine (500mg/kg). (B) 150 min after a section at caudal L3,190 min after the first injection of clonidine and 60 min after a second injection of clonidine (same dosage). (C) 80 min after a section at caudal L4 andafter three injections of clonidine (5 h, 3 h and 17 min before). (D) fictive locomotion after i.v. injection of clonidine. (E) 7 min after section at caudal L3,30 min after the injection of clonidine. (F) 15 min after the section at caudal L4, 60 min after the first injection of clonidine and 20 min after a second one.Note that in both preparations, the rhythm slows down after the L3 lesion and that only tonic activity is recorded after the lesion at caudal L4. Note also thedifferent time scale for the real and fictive locomotion panels. (Leblond and Rossignol, unpublished). Gastrocnemius lateralis (GL, an ankle extensor),Semimenbranosus-Anterior Biceps nerve (SmAB, hip extensor), Posterior Biceps-Semitendinosus nerve (PBSt), a knee flexor).

erratic. A further lesion at L4 abolished all rhythmic spinal cord (see also [46]). Fig. 4 illustrates one experi-activity even though tonic activity could still be recorded ment in which the cord was stimulated at caudal L5 on thein the muscles. This inability to walk lasted for more than left side with repetitive trains of stimuli that evoked a1 h after the last lesion. A similar sequence of lesions was regular rhythmic locomotor activity on the right side. Noteperformed during fictive locomotion with similar results in Fig. 4A that the stimulus more or less comes at around(see Fig. 2D–F). the time of discharge of the right St which triggers a

In one experiment, after establishing a locomotor pattern prolonged St burst followed by an extensor burst of(Fig. 3A), the spinal cord was cut at the L4–L5 junction at activity in both the VL and GL. After sectioning the cord atprogressively larger depths. In Fig. 3B, only the dorsal L3–L4, it was still possible to trigger generalized andcolumns were cut and the locomotor pattern continued synchronous activity on both sides of the cord but withoutunaltered. A further lesion involving the dorsal half of the any rhythmic organization. This again suggests that thecord had similarly little effect on the walking pattern (Fig. effects of the spinal electrical stimulation was relayed3C), which, however, was completely abolished when the somehow through the rostral lumbar segments or at leastventral half of the cord was sectioned as well. This again depended on the integrity of these segments. The underly-illustrates the importance of these rostral segments and ing spinal circuitry at the basis of this pattern generationespecially the ventral pathways interconnecting the rostral must still be worked out but most certainly should involveand caudal segments in the generation of the locomotor the important intrinsic propriospinal circuitry [38–rhythm. 40,54,64], the midlumbar interneurones shown to be active

On-going experiments in decerebrate cats spinalised 1 during fictive locomotion [27,55] or cholinergic inter-week before also suggest that in some cases a locomotor neurones shown, by c-fos and CHAT-immunolabeling, topattern can be evoked by electrical microstimulation of the be active during locomotion [56].


Fig. 3. Effects of progressive dorso–ventral lesions at the L4–L5 junction on treadmill locomotion in an acutely decerebrated cat spinalised 1 week beforeat T13. (A) raw EMG traces of spinal locomotion after clonidine injection i.v. (500mg/kg) while the segments of L4 and L5 are still intact; (B) after thesection of the dorsal funiculi at the junction of L4–L5 segments; (C) after completing the section to include the dorsal half of the cord; (D) after the

´complete spinalisation. (Barthelemy, Leblond and Rossignol, unpublished).

Can animals re-express locomotion if these upper lum- enlargement [45]. These authors concluded that the patternbar segments are lesioned or inactivated chronically? generating capacity of the anterior half of the hindlimbUsing a cooling probe, it was shown that inactivation of enlargement is greater than the posterior half.these upper lumbar segments in cats abolished scratching In recent experiments we have studied the effect of a[21]. In the acute experiments reported above [43], a lesion second lesion at around L3–L4 in cats previously spinal-of the cord at the L4–L5 junction abolished all spinal ised at T13 and which had recovered a good locomotorlocomotion. After complete transection of the spinal cord pattern. In all cases, locomotion was abolished and did notjust posterior to the forelimb enlargement in the turtle, the reappear despite intensive daily training over severalhindlimbs exhibit a motor task called scratch reflex [44]. weeks. However, vigorous muscle activity was observedThis scratch reflex is elicited by tactile stimulation to a site during fast paw shake or when pinching the feet. Consider-innervated by spinal segments caudal to the level of ing that the main motoneurones of the hindlimbs are belowtransection. Three forms of scratch reflex were produced in the L4 level [63], are still functional (fast paw shake) andresponse to tactile stimulation [47] and key elements of the available to produce locomotion, one must conclude thatcentral pattern generator of these three scratch forms were the integrity of the midlumbar spinal segments are essen-found in the anterior three segments of the hindlimb tial for the expression of spinal locomotion.


Fig. 4. Effects of a second spinal lesion on locomotion evoked by electrical stimulation of the cord. Effects of electrical stimulations applied at Caudal L5level before and after a second spinalisation at the junction of L3–L4, in an acutely decerebrated cat spinalised 7 days before (T13). (A) EMG averagesafter phasic electrical stimulations at caudal L5, 1 mm left to the midline and 0.5 mm under the dorsal surface after clonidine injection i.v. (500mg/kg),while the segments of L3 and L4 are still intact. The stimulation consists of 200 ms trains (250ms pulses, 20mA, 300 Hz) repeated every second. The

´averages are triggered on the onset of the train. (B), EMG averages with similar trains after a complete transection at the junction of L3–L4. (Barthelemy,Leblond and Rossignol, unpublished).

5 . General conclusions Many thanks to Dr. Trevor Drew for his helpful commentson the manuscript.

The above accounts make three major points. Firstly, thespinal cord below a complete T13 spinal transection iscapable of generating an elaborate locomotor pattern with R eferencesplantar foot placement, weight support and adaptability totreadmill speed. Secondly, this re-expression of locomo-

[1] D.M. Armstrong, Supraspinal contributions to the initiation andtion, suggests plastic changes within the spinal cord and control of locomotion in the cat, Prog. Neurobiol. 26 (1986) 273–this plasticity can be directly illustrated by the ability of 361.

[2] H. Barbeau, C. Chau, S. Rossignol, Noradrenergic agonists andspinal animals to adapt to peripheral deficits. The expres-locomotor training affect locomotor recovery after cord transectionsion of locomotion after spinalisation depends on neuro-in adult cats, Brain Res. Bull. 30 (1993) 387–393.chemical changes as well as on afferent inputs, namely

[3] H. Barbeau, C. Julien, S. Rossignol, The effects of clonidine andcutaneous. Thirdly, the midlumbar region of the spinal yohimbine on locomotion and cutaneous reflexes in the adultcord appears to play a critical role in the expression of chronic spinal cat, Brain Res. 437 (1987) 83–96.

[4] H. Barbeau, S. Rossignol, Recovery of locomotion after chronicspinal locomotion since all locomotion is abolished whenspinalization in the adult cat, Brain Res. 412 (1987) 84–95.these segments are non-functional.

[5] H. Barbeau, S. Rossignol, Enhancement of locomotor recoveryThis has several potential implications for further re-following spinal cord injury, Curr. Opin. Neurol. 7 (1994) 517–524.

search using electrical microstimulation or grafts since it ´[6] M. Belanger, T. Drew, J. Provencher, S. Rossignol, A comparison ofcircumscribes regions of critical importance for locomo- treadmill locomotion in adult cats before and after spinal transection,

J. Neurophysiol. 76 (1996) 471–491.tion. It also has implications for work in spinal cord[7] L. Bouyer, S. Rossignol, Hindfoot cutaneous input contribution toinjured patients.

locomotion in intact and spinal cats, Soc. Neurosci. Abstr. 23 (1997)759.

[8] L. Bouyer, S. Rossignol, The contribution of cutaneous inputs toA cknowledgements locomotion in the intact and the spinal cat, in: O. Kiehn, R.M.

Harris-Warrick, L.M. Jordan, H. Hultborn, N. Kudo (Eds.), Neuronalmechanisms for generating locomotor activity, Annals of the NewWe wish to acknowledge the invaluable technical sup-York Acad. of Sci., Vol. 860, New York, 1998, pp. 508–512.port of J. Provencher, F. Lebel, C. Gagner, P. Drapeau, D.

[9] L. Bouyer, S. Rossignol, Locomotor plasticity in the spinal cat afterCyr. The support of various granting agencies is also a progressive cutaneous denervation, Soc. Neurosci. Abstr. 25acknowledged: the Canadian Institute of Health Research (1999) 1153.and the Spinal Cord Research Foundation and the FCAR.[10] L. Bouyer, S. Rossignol, Spinal cord plasticity associated with


locomotor compensation to peripheral nerve lesions in the cat, in: [31] H. Forssberg, G. Svartengren, Hardwired locomotor network in catM.M. Patterson, J.W. Grau (Eds.), Spinal Cord Plasticity: Alterations revealed by a retained motor pattern to gastrocnemius after musclein Reflex Function, Kluwer Academic Publishers, Boston, Massa- transposition, Neurosci. Lett. 41 (1983) 283–288.chusetts, 2001, pp. 207–224. [32] M. Gimenez y Ribotta, J. Provencher, D. Feraboli-Lohnherr, S.

[11] L.J.G. Bouyer, P. Whelan, K.G. Pearson, S. Rossignol, Adaptive Rossignol, A. Privat, D. Orsal, Activation of locomotion in adultlocomotor plasticity in chronic spinal cats after ankle extensors chronic spinal rats is achieved by transplantation of embryonic rapheneurectomy, J. Neurosci. 21 (2001) 3531–3541. cells reinnervating a precise lumbar level, J. Neurosci. 20 (2000)

[12] E. Brustein, S. Rossignol, Recovery of locomotion after ventral and 5144–5152.ventrolateral spinal lesions in the cat. II. Effects of noradrenergic [33] N. Giroux, S. Rossignol, T.A. Reader, Autoradiographic study ofand serotoninergic drugs, J. Neurophysiol. 81 (1999) 1513–1530. a -, a -Noradrenergic and Serotonin receptors in the spinal cord1 2 1A

[13] L. Carrier, L. Brustein, S. Rossignol, Locomotion of the hindlimbs of normal and chronically transected cats, J. Comp. Neurol. 406after neurectomy of ankle flexors in intact and spinal cats: model for (1999) 402–414.the study of locomotor plasticity, J. Neurophysiol. 77 (1997) 1979– [34] S. Grillner, Control of locomotion in bipeds, tetrapods, and fish, in:1993. J.M. Brookhart, V.B. Mountcastle (Eds.), Handbook of Physiology,

[14] J.-R. Cazalets, Organization of the spinal locomotor network in The Nervous System II, Am. Physiol. Soc, Bethesda, Maryland,neonatal rat, in: R.G. Kalb, S.M. Strittmater (Eds.), Neurobiology of 1981, pp. 1179–1236.Spinal Cord Injury, Humana Press, Totowa NJ, 2000, pp. 89–111. [35] S. Grillner, P. Zangger, On the central generation of locomotion in

[15] J.-R. Cazalets, M. Borde, F. Clarac, Localization and organization of the low spinal cat, Exp. Brain Res. 34 (1979) 241–261.the central pattern generator for hindlimb locomotion in newborn [36] J.A. Hodgson, R.R. Roy, R.D. de Leon, B. Dobkin, V.R. Edgerton,rat, J. Neurosci. 15 (1995) 4943–4951. Can the mammalian lumbar spinal cord learn a motor task?, Med.

[16] C. Chau, H. Barbeau, S. Rossignol, Early locomotor training with Sci. Sports Exer. 26 (1994) 1491–1497.clonidine in spinal cats, J. Neurophysiol. 79 (1998) 392–409. [37] W. Jiang, T. Drew, Effects of bilateral lesions of the dorsolateral

[17] C. Chau, H. Barbeau, S. Rossignol, Effects of intrathecal a - and funiculi and dorsal columns at the level of the low thoracic spinal1

a -noradrenergic agonists and norepinephrine on locomotion in cord on the control of locomotion in the adult cat: I. Treadmill2

chronic spinal cats, J. Neurophysiol. 79 (1998) 2941–2963. walking, J. Neurophysiol. 76 (1996) 849–866.[18] R.D. de Leon, J.A. Hodgson, R.R. Roy, V.R. Edgerton, Full weight- [38] P.G. Kostyuk, D.A. Vasilenko, Spinal interneurons, Annu. Rev.

bearing hindlimb standing following stand training in the adult Physiol. 41 (1979) 115–126.spinal cat, J. Neurophysiol. 80 (1998) 83–91. [39] P.G. Kostyuk, D.A. Vasilenko, E. Lang, Propriospinal pathways in

[19] R.D. de Leon, J.A. Hodgson, R.R. Roy, V.R. Edgerton, Locomotor the dorsolateral funiculus and their effects on lumbosacralcapacity attributable to step training versus spontaneous recovery motoneuronal pools, Brain Res. 28 (1971) 233–249.after spinalization in adult cats, J. Neurophysiol. 79 (1998) 1329– [40] V.M. Kozhanov, The propriospinal monosynaptic effects of the1340. ventral descending pathways on the cat lumbar motoneurones,

[20] R.D. de Leon, J.A. Hodgson, R.R. Roy, V.R. Edgerton, Retention of Sechenov Physiol. J. USSR 60 (1974) 171–178.hindlimb stepping ability in adult spinal cats after the cessation of [41] R.G. Lovely, R.J. Gregor, R.R. Roy, V.R. Edgerton, Effects ofstep training, J. Neurophysiol. 81 (1999) 85–94. training on the recovery of full-weight-bearing stepping in the adult

[21] T.G. Deliagina, G.N. Orlovsky, G.A. Pavlova, The capacity for spinal cat, Exp. Neurol. 92 (1986) 421–435.generation of rhythmic oscillations is distributed in the lumbosacral [42] A.R. Luff, S.N. Webb, Electromyographic activity in the cross-spinal cord of the cat, Exp. Brain Res. 53 (1983) 81–90. reinnervated soleus muscle of unrestrained cats, J. Physiol. (Lond.)

[22] R.A. Dykman, P.S. Shurrager, Successive and maintained con- 365 (1985) 13–28.ditioning in spinal carnivore, J. Comp. Physiol. Psychol. 49 (1956) [43] J. Marcoux, S. Rossignol, Initiating or blocking locomotion in spinal27–35. cats by applying noradrenergic drugs to restricted lumbar spinal

[23] J.C. Eccles, R.M. Eccles, F. Magni, Monosynaptic excitatory action segments, J. Neurosci. 20 (2000) 8577–8585.on motoneurones regenerated to antagonistic muscles, J. Physiol. [44] L.I. Mortin, J. Keifer, P.S. Stein, Three forms of the scratch reflex in154 (1960) 68–88. the spinal turtle: movement analyses, J. Neurophysiol. 53 (1985)

[24] V.R. Edgerton, C.P. de Guzman, R.J. Gregor, R.R. Roy, J.A. 1501–1516.Hodgson, R.G. Lovely, Trainability of the spinal cord to generate [45] L.I. Mortin, P.S. Stein, Spinal cord segments containing keyhindlimb stepping patterns in adult spinalized cats, in: M. Shima- elements of the central pattern generators for three forms of scratchmura, S. Grillner, V.R. Edgerton (Eds.), Neurobiological Basis of reflex in the turtle, J. Neurosci. 9 (1989) 2285–2296.Human Locomotion, Japan Scientific Societies Press, Tokyo, 1991, [46] V.K. Mushahwar, D.F. Collins, A. Prochazka, Spinal cord mi-pp. 411–423. crostimulation generates functional limb movements in chronically

[25] V.R. Edgerton, D.J. Johnson, L.A. Smith, K. Murphy, A. Eldred, J.L. implanted cats, Exp. Neurol. 163 (2000) 422–429.Smith, Effects of treadmill exercises on hindlimb muscles of the [47] G.A. Robertson, L.I. Mortin, J. Keifer, P.S. Stein, Three forms of thespinal cat, in: C.C. Kao, R.P. Bunge, P.J. Reier (Eds.), Spinal Cord scratch reflex in the spinal turtle: central generation of motorReconstruction, Raven Press, New York, 1983, pp. 435–444. patterns, J. Neurophysiol. 53 (1985) 1517–1534.

[26] V.R. Edgerton, R.R. Roy, R.D. de Leon, N. Tillakaratne, J.A. [48] S. Rossignol, Neural control of stereotypic limb movements, in:Hodgson, Does motor learning occur in the spinal cord?, Neurosci- L.B. Rowell, J.T. Sheperd (Eds.), Exercise: Regulation and Integra-entist 3 (1997) 287–294. tion of Multiple Systems, Handbook of Physiology, Vol. Section 12,

[27] S.A. Edgley, E. Jankowska, S. Shefchyk, Evidence that mid-lumbar Oxford University Press, New York, 1996, pp. 173–216.´neurones in reflex pathways from group II afferents are involved in [49] S. Rossignol, M. Belanger, C. Chau, N. Giroux, E. Brustein, L.

locomotion in the cat, J. Physiol. 403 (1988) 57–71. Bouyer, C.-A. Grenier, T. Drew, H. Barbeau, T. Reader, The spinal[28] H. Forssberg, S. Grillner, The locomotion of the acute spinal cat cat, in: R.G. Kalb, S.M. Strittmatter (Eds.), Neurobiology of Spinal

injected with clonidine i.v, Brain Res. 50 (1973) 184–186. Cord Injury, Humana Press, Totowa, 2000, pp. 57–87.[29] H. Forssberg, S. Grillner, J. Halbertsma, The locomotion of the low [50] S. Rossignol, L. Bouyer, P.J. Whelan, K.G. Pearson, Chronic spinal

spinal cat. I. Coordination within a hindlimb, Acta Physiol. Scand. cats can recover locomotor function following transection of an108 (1980) 269–281. extensor nerve, Soc. Neurosci. Abstr. 23 (1997) 761.

´[30] H. Forssberg, S. Grillner, J. Halbertsma, S. Rossignol, The locomo- [51] S. Rossignol, C. Chau, E. Brustein, M. Belanger, H. Barbeau, T.tion of the low spinal cat: II. Interlimb coordination, Acta Physiol. Drew, Locomotor capacities after complete and partial lesions of theScand. 108 (1980) 283–295. spinal cord, Acta Neurobiol. Exp. 56 (1996) 449–463.


[52] S. Rossignol, T. Drew, E. Brustein, W. Jiang, Locomotor per- nerve regeneration and muscle transposition, Quat. Rev. Biol. 20formance and adaptation after partial or complete spinal cord lesions (1945) 311–369.in the cat, Prog. Brain Res. 123 (1999) 349–365. [63] V.G.J.M. Vanderhorst, G. Holstege, Organization of lumbosacral

[53] S. Rossignol, N. Giroux, C. Chau, J. Marcoux, E. Brustein, T. motoneuronal cell groups innervating hindlimb, pelvic floor, andReader, Pharmacological aids to locomotor training after spinal axial muscles in the cat, J. Comp. Neurol. 382 (1997) 46–76.injury in the cat, J. Physiol. (Lond.) 533.1 (2001) 65–74. [64] D.A. Vasilenko, Principles of neuronal organization of the motor

[54] A.I. Shapovalov, Neurones and Synapses of Supraspinal Motor control systems, Proc. XII all Union Physiol. Soc. Congr. Tbisilis,Systems, in: Nauka, Leningrad, 1975, pp. 58–64, in Russian. Nauka, Leningrad, 2001, pp. 121–122 (in Russian).

[55] S. Shefchyk, D.A. McCrea, D. Kriellaars, P. Fortier, L. Jordan, [65] D.Viala, G.Viala, N. Fayein, Plasticity of locomotor organization inActivity of interneurons within the L4 spinal segment of the cat infant rabbits spinalized shortly after birth, in: M.E. Goldberger, A.during brainstem-evoked fictive locomotion, Exp. Brain Res. 80 Gorio, M. Murray (Eds.), Development and Plasticity of the(1990) 290–295. Mammalian Spinal Cord, Liviana Press, Padova, 1986, pp. 301–310.

[56] F.E. Sherriff, Z. Henderson, A cholinergic propriospinal innervation [66] M.C. Wetzel, R.L. Gerlach, L.Z. Stern, L.K. Hannapel, Behavior andof the rat spinal cord, Brain Res. 634 (1994) 150–154. histochemistry of functionally isolated cat ankle extensors, Exp.

[57] M.L. Shik, G.N. Orlovsky, Neurophysiology of locomotor automat- Neurol. 39 (1973) 223–233.ism, Physiol. Rev. 56 (1976) 465–500. [67] P.J. Whelan, G.W. Hiebert, K.G. Pearson, Plasticity of the extensor

[58] J.L. Smith, L.A. Smith, R.F. Zernicke, M. Hoy, Locomotion in group I pathway controlling the stance to swing transition in the cat,exercised and non-exercised cats cordotomized at two or twelve J. Neurophysiol. 74 (1995) 2782–2787.weeks of age, Exp. Neurol. 76 (1982) 393–413. [68] P.J. Whelan, G.W. Hiebert, K.G. Pearson, Stimulation of the group I

[59] R.W. Sperry, The functional results of muscle transposition in the extensor afferents prolongs the stance phase in walking cats, Exp.hind limb of the rat, J. Comp. Neurol. 73 (1940) 379–404. Brain Res. 103 (1995) 20–30.

[60] R.W. Sperry, The effect of crossing nerves to antagonistic muscles in [69] P.J. Whelan, K.G. Pearson, Plasticity in reflex pathways controllingthe hind limb of the rat, J. Comp. Neurol. 75 (1941) 1–19. stepping in the cat, J. Neurophysiol. 78 (1997) 1643–1650.

[61] R.W. Sperry, Transplantation of motor nerves and muscles in the [70] H. Yumiya, K.D. Larsen, H. Asanuma, Motor readjustment andforelimb of the rat, J. Comp. Neurol. 76 (1942) 283–321. input–output relationship of motor cortex following cross-connect-

[62] R.W. Sperry, The problem of central nervous reorganization after ion of forearm muscles in cats, Brain Res. 177 (1979) 566–570.

Recovery of locomotion in the cat following spinal cord lesions

Documents

Transcript of Recovery of locomotion in the cat following spinal cord lesions