Involvement of AMPA receptors in posterior locomotor activity in the rabbit: An in vivo study

J Physiology 0 Elsevier,

(Par-is) Paris

(1998)

92,5-l

5

Involvement of AMPA receptors in posterior locomotor activity in the rabbit: An in vivo study Agnks Bonnot, Marc Corio, Anne-Marie Laboratoire

des Neurosciences

de la Motricit6,

(Received

UMR-CNRS

3 February

5807,

Avenue

Bout, Denise Viala des FacultPs,

1998; accepted 23 February

33405

Talence

cedex,

France

1998)

Abstract -Although AMPA receptors are known to be widely involved in excitatory synaptic neurotransmission at the spinal level, very little is known about their role in modulating motor activity in mammals. In curarized decerebrate or spinalized rabbit preparations, fictive locomotion was monitored on hindlimb nerves after either activation or blockade of AMPA receptors. In decerebrate preparations, the administration of the antagonist, NBQX (3.5 mg/kg i.p.) or the agonist, AMPA (0.5 mg/kg i.v.) produced, in both cases, a depression of locomotor activities induced by stimulation of cutaneous afferents (evoked locomotor activity). This potent effect was transient with AMPA (recovery after 20 min) and followed by the occurrence of spontaneous locomotor sequences, while no recovery was observed with NBQX treatment. In spinal preparations where a continuous ‘spontaneous’ locomotor activity resulted from the pharmacological activation of noradrenergic descending pathways (nialamide-DOPA pretreatment), the same drugs injected at higher doses (5 mg/kg NBQX i.p. and I mg/kg AMPA i.v.) only weakly affected the frequency of ‘spontaneous’ and evoked locomotor bursts while they exerted inhibitory and facilitatory effects on the burst amplitude respectively. The results suggest that AMPA receptors are involved at spinal level: I) in direct mediation of cutaneous afferent excitatory effects on the posterior locomotor generators (pLG); 2) in indirect mediation of a supraspinal descending inhibition controlling, likely presynaptically, the cutaneous afferent activation; and 3) in transmission to motoneurons of the output signals from the pLG. Finally, tight spinal interactions between potent descending noradrenergic pathways and spinal AMPA neurotransmission were disclosed. (0 Elsevier, Paris). non-NMDA

receptors

I locomotor

activity

/ cutaneous

afferents

1. Introduction It is well established that glutamate and aspartate are the main endogenous neurotransmitters responsible for excitatory synaptic transmission of sensory information in the mammalian spinal cord. These endogenous ligands act on N-methyl-D-aspartate (NMDA) and AMPA/kainate (non-NMDA) ionotropic receptors. AMPA (a-amino-3-hydroxy-5methyl-isoxaolepropionate) receptors are used extensively at glutamatergic synapses throughout the vertebrate central nervous system, where they are involved in fast excitatory neurotransmission. However, the lack of selective antagonists for nonNMDA receptors has long delayed studies on their involvement in physiological and pathological processes, compared to those concerning NMDA receptors. Selective antagonism has been described with the quinoxaline derivatives [22]; NBQX (2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinoxaline) was thought to specifically antagonize the effects of AMPA subtype of non-NMDA receptors since it is some 30-fold less potent at kainate receptors, and in contrast to the other quinoxalines, it does not act at the glycine site of NMDA receptors [29]. Several studies have indicated that AMPA antagonists can attenuate both noxious and non-noxious

I noradrenergic

activation

transmission in the spinal cord [6, 24, 381, disclosing their contribution to afferent transmission [27, 311. A substantial body of immunocytochemical data indicates the presence of moderate to high levels of AMPA receptors in motor nuclei [ 16, 32, 391, supporting electrophysiological observations that these motoneurons are depolarized via AMPA receptors [26]. Electrophysiological findings that non-NMDA antagonists evoke muscle relaxation [35] and depress monosynaptic spinal reflex in mice [34] also support the involvement of AMPA receptors in the mediation of excitation at motoneuronal level [9]. However, it has been observed from electrophysiological studies in vitro [3, 4, 18, 301 and in vivo [lo] that AMPA receptor agonists have a lower capability, as compared to other EAA agonists such as NMDA and kainate, to induce motor activity in higher vertebrates. Consequently, to date, no electrophysiological or behavioral studies have focused on motor control performed by AMPA receptors. The aim of the present electrophysiological study is to determine the role of AMPA receptors in modulating the activity of the posterior locomotor generators (pLG) in the rabbit. Decerebrate and spinal rabbit preparations were used to compare the effects of systemic administration of both the agonist, AMPA and the antagonist, NBQX on spontaneous and evoked locomotor activity.

A. Bonnot

et al.

2. Materials and methods

2.4. Drug trials

2.1. Surgery

The following drugs were used: 1) DL-a-AMPA (Sigma) was dissolved in saline (0.9% NaCl solution) for i.v. administration (0.5 or 1 mg/kg); 2) NBQX (Novo Nordisk) was dissolved in dilute aqueous 5% ethanol absolute for i.p. administration (3.5 or 5 mg/kg). In addition, when studying the drug effects on spinal preparations, the animals were pretreated with a monoaminoxidase inhibitor, nialamide (Niamide Pfizer, 120 mg/kg i.p.) just after the general anesthesia so that the injection of L-DOPA (from 90 to 120 mg/kg i.v.) a few minutes after spinal transection could be active. When, at the end of some experiments, the blood pressure dropped, it was maintained above 80 mmHg by a slow perfusion of adrenaline in saline (0.04 mg/kg i.v.). Control preparations were used to check that dilute aqueous ethanol administration alone was inefficacious and that frequency and amplitude of locomotor activities did not change over time in response to repeated peripheral stimulations.

The experiments were carried out using 21 adult rabbits (2.3-2.9 kg body weight). After anesthesia with sodium pentobarbitate (pentobarbital, 35 mg/kg, i.v.), the animal was tracheotomized. A jugular cannula was inserted for further injections and a carotid cannula was used for blood pressure recording. The second carotid artery was also ligated. Then the animals were curarized (gallamine triethyiodide, flaxedil 5 mgkg i.v.) and artificially ventilated. The rabbit’s head was placed in a stereotaxic contention frame and, after craniectomy, a decerebration was performed at a precollicular-premammilar level. The remaining exposed brain was protected with agar. Such decerebrate animals were no longer capable of perceiving pain and, therefore, did not require the subsequent use of general anesthetics [ 11. Right hindlimb muscle nerves to tibialis anterior (TA, flexor muscle) and to gastronecmius medialis (GM, extensor muscle) and the sural nerve were isolated and severed in their distal part. The dissected nerves were maintained in mineral oil. A restricted low thoracic laminectomy was performed on some of the animals for subsequent spinal transection during the experiment. The dura was removed just before spinal transection and lidocaine (xylocaine, 1% adrenaline) was infiltrated to avoid mechanical stimulation. 2.2. Recordings Since general anesthesia greatly depresses locomotor function, recordings were started only 6 h after the onset of surgery, when the main effects of anesthesia had disappeared. Such unanesthetized decerebrate animals were capable of a wide range of spontaneous and stimulated muscle nerve activities which were recorded from their proximal cut end with a pair of silver wires connected to short time-constant amplifiers (bandwidth 100 Hz-l 0 kHz), before and up to 120 min after drug administration. These recordings were fed into a multichannel tape recorder and also monitored during the experiment on a 4-channel digital oscilloscope (Yokogawa DL 1200 A) and on an g-channel ink recorder (Siemens), as raw or filtered signals. In order to determine the precise intensity required to stimulate either A fiber group alone or both A and C fiber groups, the afferent volley, evoked by stimulation of the sural nerve, was recorded in a proximal part of this nerve through one active electrode; another was placed in an adjacent muscle and both were connected to an amplifier. Heart rate, blood pressure and rectal temperature were constantly monitored. 2.3. Stimulations Stimulation of the sural nerve was achieved through a pair of silver wires placed on the central cut end. The intensity of the isolated shocks was adjusted in order to obtain selective activation ofA@ primary afferent fibers or to excite all group A fibers (duration 0.1 ms in both cases). A higher intensity of electrical stimulation was used to recruit both A and C fibers (duration 1 ms). Locomotion was evoked through repetitive electrical stimulation (at 10 Hz during 10 s) of these fibers.

2.5. Data processing The periodic bursts in hindlimb muscle nerves were digitized from the filtered activities and analyzed in their different parameters by the 4-channel digital oscilloscope. Unlike regular firing of spinal preparations, locomotor firing was not regular during a 10 s stimulation (progressive decrease) in decerebrate preparations. Consequently average frequency, defined as the number of bursts over a 10 s period of cutaneous stimulation or spontaneous activity, was considered as an index of the effects of the drugs. Prior to application of any drug, the average burst frequency (F) and the average burst amplitude (A) (considered separately for tibialis anterior (TA) and gastrocnemius medialis (GM) nerve activities) were defined on 10 s periods. After administration of the drug, the resulting changes in burst frequency and amplitude were expressed as a percentage of the control value. Only one administration was performed in each experiment, and the experiments were repeated on several preparations to determine the mean effect.

3. Results 3.1. Decerebrate

preparations

Fourteen experiments were performed to study NBQX (3.5 mg/kg i.p.) and AMPA (0.5 mg/ kg iv.) effects on decerebrate preparations. In all of them, locomotor sequences could be elicited at any moment through repetitive stimulation of hindlimb cutaneous afferents. When the sural nerve was stimulated, locomotor bursts occurred alternately on TA and GM nerves; the greater the number of fibers recruited, the higher the frequency. Different intensities of cutaneous stimulation were compared in order to distinguish between the effects of the drugs on the activation pathways of the locomotor network and the operation of the network itself.

7

AMPA receptors in posterior locomotor activity

3.2. NBQX administration In eight experiments, the injection of the AMPA receptor antagonist, NBQX, resulted within about 10 min in the disappearanceof all bursts when AC@ and sometimesAC@? fiber groups were stimulated but not when high threshold primary afferent fibers (A+C) were recruited. In 10 other experiments, a weaker NBQX effect enabled the observation of a progressive depressionof the elicited locomotor sequences as shown in figure IA. The frequency of

locomotor activity progressively dropped over time during the first 15 min after NBQX injection and then remained close to 27%, 24% and 56% of the control value respectively for AC@, Aa@ and A+C stimulation (figure IB); as both A and C fibers were recruited, the decreasein locomotor frequency after NBQX administration was less perceptible. It was also possible to observe, after NBQX administration, a progressive reduction of the amplitude of the locomotor bursts to the sameextent for both TA and GM nerves. In addition, the stronger the stimulation,

A+C

Figure 1. Effect of NBQX on decere4

1AaaS1

brate preparations. The effect of NBQX on locomotor activities elicited by stimulationof the sural nerve (AaD, Aa@ or A+C fibers) in the decerebrate preparation. A. Loco-

motor activities recordedfrom hind-

1 A+C 1 80

limb muscle nerves (TA and GM) just before (control, C) and at different times after injection of 3.5 mg/kg NBQX. The bar underneath the records represents the 10 s duration of the cutaneous stimulation. B. Temporal evolution of two parameters of rhythmic activity: burst frequency (F) and burst amplitude (A) for TA and GM. Each value plotted is expressed as a percentage of the control value established from two experiments (n = 2).

A. Bonnot et al.

the smaller the decrease in burst amplitude, whatever the time after NBQX injection. Four h after NBQX administration, no recovery of its inhibitory effect on frequency and amplitude of the evoked locomotor activities was observed.

NBQX, such rapid locomotor inhibition was always followed, about 25-30 min after administration, by a partial recovery (figure 2A, B). However, as with the antagonist NBQX, the stronger the stimulation, the weaker the effect. Thus, both the frequency and amplitude of bursts were depressed with a much larger effect for locomotor sequences evoked by AC@ fiber group stimulation (total disappearance of flexor bursts and the amplitude of the remaining tonic extensor activity reached 12% of the control value at t 15 min) than for those evoked by all A

3.3. AMPA administration In four experiments, AMPA (0.5 mg/kg) deeply depressed evoked locomotor activities within 15 min after administration. In contrast to the effect of

A AaP

A+C

AaBs

TA

3o

rr-1

nn

GM

115 Figure 2. Effect of AMPA on decerebrate preparation. The effect of AMPA on spontaneous or evoked activities in the decerebrate preparation. A. Motor activities from the hindlimb nerves either elicited by stimulation of the sural nerve (Asp, AC@ or A+C fibers) in control and for three times in succession after administration of 0.5 mg/kg i.v. AMPAor occurring spontaneously from 30 min. The bar underneath the records represents the 10 s duration of the cutaneous stimulation. B. Temporal evolution of the two parameters of rhythmic activity: frequency (F) and amplitudes (A) for TA and GM, expressed as a percentage of the control value elicited in the same experiment before the OCcurrence of spontaneous activities.

A

m n

AA

An 111

ilnh

n

I

I

L

A .

n n

AllI

B 1 A+C 1

OF(n)

+TA (%)&M

(%)

AMPA receptors in posterior locomotor activity

fiber group stimulation (9%, 67% and 41% of the control values for F, A (TA) and A (GM) respectively at t 15 min); it was still weaker when all sural nerve fibers (A+C) were recruited (73%, 81% and 72% of the control values for F, A (TA) and A (GM) respectively at t 15 min) (figure 2B). It is worth noting that AMPA was able, in both experiments, to induce ‘spontaneous’ bursts (i.e., without any peripheral stimulation) (figure 2A, SPONT). This unpredictable and irregular firing disturbed the usual response to peripheral stimulation. Therefore, after the onset of these bursts at time 30 min, we could no longer take into account the evoked locomotor activities. Such spontaneous bursts first appeared on the flexor nerve, grouped in sequences of 2-8 bursts of 300-500 ms duration at about 0.8 Hz in frequency. Extensor bursts, alternating with the flexor bursts appeared from 75 min. Such a spontaneous rhythm, which belongs to locomotor activity, continued for a further 30 min, always as short or longer sequences interrupted by silent motor periods. In the two other experiments, AMPA led, in the same delay, to a total disappearance of all flexor bursts when group A fibers were stimulated, an effect that did not occur when C fibers were also recruited. However, neither recovery of the evoked locomotor responses, nor spontaneous activity occurred in the later experiments. 3.4. Spinal preparations After acute spinal transection, locomotor activities could no longer be elicited by cutaneous nerve stimulation, even when all fiber groups were stimulated. In six out of seven animals pretreated with nialamide, L-DOPA administration allowed locomotor bursts to occur ‘spontaneously’, i.e., without any electrical stimulation, after a 20 min delay, on flexor and extensor muscle nerves (figure 3A). It has previously been shown that such regular and uninterrupted bursting at about 1 Hz could be maintained stable for more than 40 min [36]. As soon as such regular bursting was obtained, either NBQX (5 mg/kg i.p.) or AMPA (1 mg/kg i.v.) was injected and their effects on either ‘spontaneous’ activities or evoked activities were recorded. Whatever the injected drug and in spite of the larger doses, we did not obtain such large effects on the spinal preparation as on the decerebrate preparation.

9

plies for both spontaneous rhythmic activities and for those elicited by peripheral electrical stimulation. Indeed, whatever the experimental conditions, locomotor frequency remained close to the control value (figure 3B). On the other hand, from 10 min after its administration, NBQX globally reduced the amplitude of GM bursts, especially for locomotor sequences elicited by group A fiber stimulation (79% and 81% of the control value between 10 and 30 min, when stimulating respectively group AC@ and A@38 fibers) and for ‘spontaneous’ extensor activities (78% of the control value). This decrease of extensor amplitude did not occur when all afferent fibers, i.e., group A+C, were recruited (around 100% of the control value). As far as the flexor burst amplitude is concerned, it remained relatively unchanged when locomotor sequences were evoked whatever the kind of stimulation (92%, 98% and 105% for AC@, At$G and A+C fibers, respectively), whereas it clearly increased (115% of the control value) in the absence of any stimulation. 3.6. AMPA administration In three experiments in which AMPA administration followed low spinal transection and DOPA activation, the frequency of locomotor activities was not greatly altered 10 to 15 min after AMPA administration (10 to 15% around the control value) while a pronounced increase in the amplitude of spontaneous activities appeared in every case (figure 4). In two experiments where nialamide-DOPA pretreatment did not enable the occurrence of spontaneous locomotor activities, they appeared after AMPA administration and progressively increased in amplitude against time. In the third experiment, spontaneous rhythm was already present in control and the TA burst amplitude reached about 170% of the control value, 45 min after injection. For evoked locomotor activities, it appears clearly that the stronger the electrical stimulation, the weaker the effect on TA amplitude (135% for AaD fibers, almost no change for A@8 (90%) and A+C fibers (105%)). In return, the effects on the GM burst amplitude were not homogenous for the various kinds of considered activities: positive effect on spontaneous activities (255%) and A+C evoked activities (115%), slightly negative for A@ (80%) and Aa@ (60%) evoked activities.

3.5. NBQX administration In three experiments, administration of NBQX did not induce large changes in locomotor frequency and flexor burst amplitude while extensor burst amplitude was more clearly altered (figure 3A). This ap-

4. Discussion The systemic injection of either an agonist or an antagonist of AMPA receptors on spinal preparations

A. Bonnot et al.

10

induces only weak effects on the frequency and amplitude of the locomotor activities as compared to the strong decreaseof both parametersobserved on decerebrate preparations, The interpretation of the data must be guided by the following knowledge of the level of interactions on rhythmic patterns. The locomotor occurrence may be modified by a modulation which operates either directly on the pattern generator or on inputs to the generator. A modulation of the rhythmic burst amplitude is presumed

to reflect mainly the output drive from the rhythmically active interneurons to motoneurons. Finally, to compare the effects of a samedrug on decerebrate and spinal preparations could help to determine the spinal or supraspinal origin of the modulation. 4.1. Decerebrate preparations Considered separately, the depressant effect of NBQX on evoked locomotor activities in decere-

A

WONT

c-

A+C

B

1 AaS1

I 160,

,120

$120 g

60

6C

40

0

60

s

$40

h

Figure 3. Effect of NBQX on spinal

preparations. Effect of NBQX on spi&locomotor activities,either ‘spontaneously’ occurring (pretreatment with nialamideandDGPA)or elicited by electricalstimulationof the sural nerve. A. Motor activities recorded from TA and GM nerves. B. Graphs of temporal evolution for ‘spontaneous’ locomotor activities and for electrically evoked activities (n = 3).

) AapS ) 100

60

50

40

0

0 5 10 15 26 26 JO

60

100 50

40

n

0

0 0

5

Time (min)

0 A(TA)%

I A(GM)%+

F(%)

10

16

M

Time (min)

25

Jo

AMPA receptors in posterior locomotor activity

brate animals could be explained by the blockade of an AMPA-mediated excitation of the pLG performed either by cutaneous afferents or by descending pathways. However, the fact that administration of AMPA in decerebrate preparations shows a similar depressant effect as NBQX on evoked locomotor activities rules out the possibility of such a simple AMPA mediation. We suggest that our pharmacological data are the result of both spinal excitatory and descending inhibitory influences on the pLG mediated by AMPA receptors. The possibility that they may constitute two independent pathways acting directly on the pLG can not be excluded. However, the finding that sensory messages that reach the mammalian spinal cord are controlled by GABAergic presynaptic inhibition [28] strongly supports the possibility that descending inhibition from supraspinal levels would interact with the afferent excitatory inputs on the pLG (see figure 5). In such conditions, the NBQX relief of the AMPA-mediated supraspinal inhibition is without excitatory effect on the locomotor response to peripheral stimulation since in parallel it blocks the AMPA activation of the pLG by these cutaneous afferents or the direct AMPA activation on spinal neuron excitability. In contrast, the AMPA activation of the supraspinal inhibition would be responsible for the depression of the locomotor responses to cutaneous stimulations. However, it can not be excluded that AMPA could mimic the effects of an antagonist by desensitizing its postsynaptic receptors. The weakness of drug effects generally observed with A+C stimulation could be due to the fact that the intensity of stimulation used for group C fiber recruitment results in such an activation of the pLG that pharmacological treatment has less influence in modulating the induced locomotor pattern. In the same way, most of the changes in output amplitude could be simply the result of the synaptic weight on the pLG. The interpretation of our results is strongly supported by a study reporting that either NBQX or baclofen, a GABAn agonist, relieved the behavioral sensitivity to innocuous mechanical stimulation that followed spinal cord ischemia in rats [37]. Examination of the time course of AMPA effect in the decerebrate preparation shows a maximal inhibition of induced locomotor activities 15-20 min after injection, followed by a significant recovery at 30 min. The second period, thereafter, is characterized by a delayed occurrence of spontaneous and short-lasting locomotor sequences. Why are such spontaneous bursts delayed in time after AMPA application? The occurrence after a 30 min delay of spontaneous activities could result from a rebound of the pLG activity rather than from a direct exci-

11

tatory effect of AMPA. The recovery of peripheral stimulation efficiency observed with AMPA and the possible occurrence of spontaneous activities can be seen as a removal of AMPA-mediated inhibition,

lsPoNTl cn 2000 $i 1500

160 120

0" c

2

1000

80

2

500

40

g

e a

o

0 0

s

15

25

30

40

45

k

50

200

120

150

90

100

60

50

30

0

0 0

5

IO

15

25

30

40

45

I-aq

120,

,120

100

90

80

60

60 30

40

0

20 0

150

5

10

15

25

30

40

45

1200

I

125

150

100

100

75

50

50

0 0

5

10

15

25

30

40

45

Time (min)

1 -+F (%) OA(TA) Figure 4. Effect of AMPA on spinalpreparations. Temporal

evolutionof the different parametersof locomotoractivity after treatmentwith AMPA (1 mg/kgi.v.) in spinalpreparations (n = 2) (samepresentationasinfigure 38).

12

A. Bonnot et al.

Supraspinal descending pathways

I

NA

.

NA

.................... ;? .-.* :...v

1 ,B6\ l A

1

4 -

Excitatory Inhibitory

a @

Cutaneous afferents

AMPA receptors PLG

Figure 5. Simplified scheme indicating the different spinal levels at which neuromodulation on AMPA receptors could affect the resulting peripheral fictive locomotor activity. NA, noradrenaline; pLG, posterior locomotor generator; Mn, motoneuron.

likely

corresponding

to the metabolism

or uptake of

AMPA. Alternatively, such a phenomenon could simply reverse the desensitizing effect of AMPA on spinal neuron excitability. In return, the uptake phenomenon can not occur for a competitive antagonist such as NBQX whose effect was therefore not reversible. 4.2. Spinal preparations

The lack of effect of systemic injection of NBQX on the frequency of ‘spontaneous’and evoked spinal locomotor activities could suggestthe absenceof locomotor-related cells sensitive to AMPA at spinal

level on the afferent pathway to the generator. Such an interpretation implying an exclusive location at the supraspinal level of the AMPA receptors involved in pLG modulation can not account, as seen above, for the effects of the agonist and the antagonist on decerebrate preparations. In addition, a large body of neuroanatomical evidence has revealed the expression of AMPA receptors throughout the rostro-caudal extention of the mammalian spinal cord, with a prominent localization in the superficial laminae of the dorsal horn [20, 21, 23, 32, 391. Furthermore, there is immunohistological evidence for a close relationship between the glutamatergic nociceptive primary afferent system [2]

AMPA receptors in posterior locomotor activity

and the AMPA receptors located on dorsal horn neurons. Finally, the observation of AMPA effects on the frequency of ‘spontaneous’ or evoked locomotor activities in spinal preparations also discloses the presence of AMPA receptors on the inputs to the pLG. Thus, both the weakness (for NBQX) and the heterogeneity (for AMPA) of drug effects on the frequency of locomotor activities in spinal preparations could be partly explained if the supraspinal inhibition is AMPA-mediated at the spinal level. Indeed, a recent study suggested a contribution of glycinergic and GABAergic inhibitory interneurons to descending antinociceptive actions [25]. It has also been shown that about 30% of lumbar neurons producing fos protein in response to noxious stimuli were found to be GABA- and/or glycine-immunoreactive [33]. In addition, it has been reported that non-NMDA receptors are largely involved in the polysynaptic excitation of GABAergic interneurons controlling primary afferents [ 121. Finally, in the isolated brainstem-spinal cord preparations of newborn rats, the finding of a strong spinal GABAergic inhibition of the lumbar locomotion generator [5] suggests that the induction of locomotor activity results from a balance between excitatory and inhibitory inputs from supraspinal levels. In the same way, in the present study, the transmission from all cutaneous fiber groups could be affected with a different balance in each case, which could explain the occurrence of either a slight excitatory or a slight inhibitory effect of AMPA. Concurrently, midlumbar interneurons which project to hindlimb motoneurons have been electrophysiologically characterized in viva in the cat; these excitatory and inhibitory intemeurons received strong input from all muscular and cutaneous afferents [I l] and were specifically contacted by different descending motor pathways [g]. These interneurons can therefore be considered as nodal points where convergent descending and afferent information is integrated and modulated and a set of them could be involved in the AMPA-mediated effect described in this report. However, it is likely that nialamide-DOPA pretreatment of spinal preparations contributes to the much less pronounced depressive effects of AMPA and NBQX in spinal rather than in decerebrate preparations. If such an activation was expected to compensate the strong spinal depression following spinal cord transection, it results in a boosted pLG able to produce continuous locomotor activities. Indeed, such an activation of noradrenergic descending pathways is known to produce a powerful release of the lumbar locomotion generators [ 17, 361 and a strong inhibitory control of painful afferent transmission [ 141. Recently, the study offos protein

13

expression following nialamide-DOPA treatment able to induce locomotion, has revealed a strong immunostaining in all parts of the spinal cord (personal observation). Therefore, we assume that the activation of descending noradrenergic pathways could greatly mask the afferent glutamatergic control on the pLG. In accordance with this interpretation is the recent finding that presynaptic inhibition evoked by stimulation of the adjacent dorsal roots would be tonically suppressed by descending noradrenergic neurons [ 191. The neuroanatomical description of a massive projection of A5 and A7 noradrenergic groups in the vicinity of motoneurons [ 151 supports the possibility they could also contribute to a powerful supraspinal control of AMPA-mediated effect on the outputs from the pLG. This supraspinal noradrenergic control could account for the increase in amplitude of spontaneous bursts after AMPA administration in the spinal preparation while such an activation would be blocked after NBQX administration which is able to reduce by 20% the amplitude of extensor locomotor bursts. Furthermore, the effects of agonist and antagonist on the efferent pathways to the pLG are disclosed. Thus, our study strongly supports the idea that nialamide-DOPA pretreatment produces such an activation of the pLG that the influence of the various afferent controls is greatly reduced compared to those exerted in decerebrate preparations. Such an interpretation is consistent with a recent study where iontophoretic injection of AMPA antagonists no longer reduced the evoked response of spinal neurons when cell excitability was increased with NMDA [7]. They conclude that the antagonists only indirectly affected the responses to peripheral stimulations by decreasing cell excitability. Thus, some of the AMPA-mediated effects described in our study could be explained on the basis of an action on cell excitability rather than an effect on synaptic transmission. In addition, it is not excluded that the used drugs have produced secondary effects on locomotion by acting on neural circuitry involved in other central mechanisms. In conclusion, our study provides the first data suggesting a role for AMPA receptors in controlling hindlimb locomotion in in vivo mammalian preparations, prominently via an AMPA-mediated spinal presynaptic inhibition. Such a spinal AMPA-mediated control results, as for the NMDA-mediated control of locomotion [13], in inhibitory effects performed on afferents to the pLG and driven by supraspinal descending pathways. Finally, in addition to a functional cooperation in neuromodulation, it will be important to determine if? in spite of a weaker inductor effect, AMPA receptors could contribute with NMDA receptors to locomotor genesis.

14

A. Bonnot

et al.

1151 Fritschy

Acknowledgments

descending

The authors thank Novo Nordisk laboratory (Denmark) for providing NBQX and Pfizer (France) for the generous supply of nialamide. They are indebted to Drs. M.J. O’Donovan, D. Orsal and D. Morin for helpful comments on this manuscript and to Dr. T. Durkin for English revision. This work was supported by INSERM (CRE 901103).

J.M., Grzanna noradrenergic

R., Demonstration of two separate pathways to the rat spinal cord:

evidence for an intragriseal trajectory of locus coeruleus axons in the superficial layers of the dorsal horn, J. Comp. Neurol.

[IhI

291 (1990)

553-582.

Furuyama T., Kiyama H., Sato K., Park H.T., Maeno H., Takagi H., Tohyama M., Regio-specific expression of subunits of ionotropic glutamate receptors (AMPA-type, KA-type and NMDA receptors) in the rat spinal cord with special reference to nociception, Mol. Brain Res. 18 (1993) 141-151.

[I71 Grillner

S., Neurobiological bases of rhythmic in vertebrates, Science 228 (1985) 143-149.

References

motor

acts

1181 Hagevik Adams R.D., in: Isselbacher E., Petersdorf R.G., Wilson ciples of Internal Medicine, New York, 1980. 13-20.

PI

[31

141

[51

K.J., Adams R.D., Braunwald J.D. (Eds.), Harrison’s Prinninth edition, McGraw-Hill,

Al-Ghoul W.M., Li Volsi G., Weinberg R.J., Rustioni A., Glutamate immunochemistry in the dorsal horn after injury or stimulation of the sciatic nerve of rats, Brain Res. Bull. 30 (1993) 453-459. Barry M.J., O’Donovan M.J., The effects of excitatory amino acids and their antagonists on the generation of motor activity in the isolated chick spinal cord, Dev. Brain Res. 36 (1987) 271-276. Cazalets J.R., Sqalli-Houssaini Y., Clarac F., Activation of the central pattern generator for locomotion by serotonin and excitatory amino acids in neonatal rat, J. Physiol. 455 (1992) 187-204. Cazalets J.R., Sqalli-Houssaini Y., Clarac F., GABAergic inactivation of the central pattern generators for locomotion in isolated neonatal rat spinal cord, J. Physiol. 474 (1994) 173-181.

[61

Cumberbatch M.J., Chizh B.A., Headley PM., AMPA receptors have an equal role in spinal nociceptive and nonnociceptive transmission, Neuroreport 5 (1994) 877-880.

[71

Cumberbatch M.J., Herrero J.F., Headley P.M., Studies of synaptic transmission in viva: indirect versus direct effects of (RS)-a-amino-3-hydroxy-5-methyl-4-isoxazole-propionic acid/kainate antagonists on rat spinal sensory responses, Neurosci. Lett. 204 (1996) 33-36. Davies H.E., Edgley S.A., Inputs to group II-activated midlumbar interneurones from descending motor pathways in the cat, J. Physiol. 479 (1994) 463473. Davies J., Watkins J.C., Role of excitatory amino acid receptors in mono- and polysynaptic excitation in the cat spinal cord, Exp. Brain Res. 49 (1983) 280-290. Douglas J.R., Noga B.R., Dai X., Jordan L.M., The effects of intrathecal administration of excitatory amino acid agonists and antagonists on the initiation of locomotion in the adult cat, J. Neurosci. 13 (1993) 99&1000.

PI [91 1101

[III

Edgley S.A., Jankowska E., An intemeuronal relay for group I and II muscle afferents in the middle lumbar segments or the cat spinal cord, J. Physiol. 389 (1987) 647-674.

WI

Evans R.H., Long S.K., Primary afferent depolarization in the rat spinal cord is mediated by pathways utilising NMDA and non-NMDA receptors, Neurosci. Lett. 120 (1989) 231-236. Fenaux E, Corio M., Palisses R., Viala D., Effects of an NMDA-receptor antagonist, MK-801, on central locomotor programming in the rabbit, Exp. Brain Res. 86 (1991) 393401. Fields H.L., Heinricher M.H., Masson P, Neurotransmitters in nociceptive modulatory circuits, Annu. Rev. Neurosci. 14 (1991) 219-244.

[I31

[I41

A., McClellan A.D., Role of excitatory amino acids in brainstem activation of spinal locomotor networks in larval lamprey, Brain Res. 636 (1994) 147-152.

U91 Hasegawa

Y., Ono H., Descending noradrenergic tonically suppress spinal presynaptic inhibition Neuroreport 7 (1995) 262-266.

r201 Henley

J.M., Localisation rat CNS using antipeptide 334-336.

of AMPA antibodies,

neurones in rats,

receptor subunits in Neuroreport 4 (1993)

I211 Henley

J.M., Jenkins R., Hunt S.P., Localisation of glutamate receptor binding sites and mRNAS to the dorsal horn of the rat spinal cord, Neuropharmacology 32 (1993) 3741.

WI

Honor& T., Davies S.N., Drejer J., Fletcher E.J., Jacobsen I?, Lodge D., Nielsen F.E., Quinoxalinediones: potent competitive non-NMDA glutamate receptor antagonists, Science 2412 (1988) 701-703.

~231 Jansen K.L.R., Autoradiographic kainate receptors (1990) 53-57.

Faull R.L.M., Dragunow M., Waldvogel H., localisation of NMDA, quisqualate and in human spinal cord, Neurosci. Lett. 108

1241 King A.E., Lopez-Garcia

J.A., Excitatory amino acid receptor-mediated neurotransmission from cutaneous afferents in rat dorsal horn in vitro, J. Physiol. (Lond.) 472 (1994) 443-457.

~251 Lin Q., Peng Y., Willis W.D., Glycine

and GABAA gonists reduce the inhibition of primate spinothalamic neurons produced by stimulation in periaqueductal Brain Res. 654 (1994) 286-302.

WI

antatract gray,

Long S.K., Smith D.A.S., Siarey R.J., Evans R.H., Effect of 6-cyano-2,3,-dihydroxy-7-nitro-quinoxaline (CNQX) on dorsal root-, NMDA-, kainateand quisqualate-mediated depolarization of rat motoneurones in vitro, Brain J. Pharmacol. 100 (1990) 850-854.

V., Liicke T., Schaible 1271 Neugebauer of N-methyl-D-aspartate (NMDA) antagonists on the responses of rat input, Neurosci. Lett. 155 (1993)

H.G., Differential effects and non-NMDA receptor spinal neurons with joint 29-32.

[281 Rudomin

P., Presynaptic inhibition of muscle spindle and tendon organ afferents in the mammalian spinal cord, TINS 13 (1990) 499-505.

~291 Sheardown

M.G., Nielsen E.O., Hansen A.J., Jacobsen P., Honor& T., 2,3-dihydroxy-6-nitro-7-sulfamoyl-benzo(F)quinoxaline: a neuroprotectant for cerebral &hernia, Science 247 (1990) 571-573.

[301 Smith J.C., Feldman

J.L., In vitro brainstem-spinal cord preparation for study of motor systems for mammalian respiration and locomotion, J. Neurosci. Methods 21 (1987) 321-333.

r311 Song X.J., Zhao Z.Q., Differential

effects of NMDA and nonNMDA receptor antagonists on spinal cutaneous vs. muscular nociception in the cat, Neuroreport 4 (1993) 17-20.

AMPA receptors in posterior locomotor activity

15

[32]

Tachibana M., Wenthold R.J., Morioka H., Petralia R.S., Light and electron microscopic immunocytochemical localization of AMPA-selective glutamate receptors in the rat spinal cord, J. Comp. Neurol. 344 (1994) 431-454.

[36]

Viala D., Buser P., Modalitts d’obtention de rythmes locomoteurs chez le lapin spinal par traitements pharmacologiques (DOPA, 5-HTP, D-amphetamine), Brain Res. 35 (1971) 151-165.

[33]

Todd A.J.. Spike R.C., Brodbelt A.R., Price R.F., Shehab S.A.S., Some inhibitory neurons in the spinal cord develop c-fos-immunoreactivity after noxious stimulation. Neuroscience 63 (1994) 805-816.

[37]

[34]

Turski L., Bressler K., Klockgether T., Stephens D.N., Differential effects of the excitatory amino acid antagonists, 6-cyano-7-nitroquinoxaline-2.3-dione (CNQX) and 3-((f)2-carboxypiperazin-4-yl)-propyl-1-phosphonic acid (CPP), on spinal reflex activity in mice, Neurosci. Lett. 113 (1990) 6671.

Xu X.J., Hao J.X., Seiger A., Wiesenfeldhallin Z., Systemic excitatory amino acid receptor antagonists of the a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor and of the N-methyl-D-aspartate (NMDA) receptor relieve mechanical hypersensitivity after transient spinal cord ischemia in rats, J. Pharmacol. Exp. Ther. 267 (1993) 140-144.

[38]

Dougherty Willis W.D., amino acid lamic tract

[39]

Bonnot A., Corio M., Tramu., Viala D., Immunocytochemical distribution of ionotropic glutamate receptor subunits in the spinal cord of the rabbit, J. Chem. Neuroanat. 11 (1996) 267-278.

[35]

Turski L., Klockgether T., Turski W.A., Ikonomidou C., Schwarz M., Sontag K.H., Muscle relaxant action of excitatory amino acid antagonists: sites of action, in: Cavalheiro E.A., Lehmann J., Turski L. (Eds.), Frontiers in Excitatory Amino Acid Research, Alan R. Liss, New York, 1988, pp. 343-350.

P.M., Palecek J., Paleckova V., Sorkin L.S.. The role of NMDA and non-NMDA excitatory receptors in the excitation of primate spinothaneurons, Neuroscience 12 (1992) 3025-3041.