Brain Research, 509 (1990) 137-140 Elsevier
137
BRES 23957
Effects of bicuculline and strychnine on synaptic inputs to edge cells during fictive locomotion Simon Alford, Thelma L. Williams and Karen A. Sigvardt Department of Physiology, St. George's Hospital Medical School, London (U. K.)
(Accepted 24 October 1989) Key words: ~-Aminobutyric acid; Fictive locomotion; Strychnine; Lamprey; Edge cell; Mechanoreceptor; Spinal cord
Edge cells are mechanoreceptive neurones located in the lateral tracts of the lamprey spinal cord. Phasic activation of these cells by lateral bending can entrain the activity of the locomotor central pattern generator. During fictive locomotion induced by bath-applied NMDLA (N-methyl-D,L-aspartate) or sensory stimulation, edge cells receive synaptic input. In this paper we provide evidence that the tonic inhibition received during sensory-induced fictive locomotion is glycinergic. We also show that during fictive locomotion induced by application of NMDLA to the spinal cord, bicuculline unveils phasic synaptic activity in edge cells. During sensory-evoked fictive locomotion, by contrast, only tonic synaptic activity is apparent in the presence of bicuculline. Fictive swimming in the in vitro preparation of the lamprey spinal cord may be activated either by the bath application of certain excitatory amino acids 4 or by various means of sensory stimulation 1~-13. Such activity requires no sensory feedback to occur but may be entrained by the phasic activation of a class of sensory cells, the edge cells 6'7. These cells are identified by their lateral location in the spinal cord and their characteristic dendritic ramifications w,15. The membrane potential of the edge cell depolarises in response to stretch of the lateral margin of the spinal cord and hyperpolarises with release 1°. Edge cells receive both excitatory and inhibitory input from other cells of the spinal cord 14'15 and are tonically inhibited during episodes of fictive locomotion induced by sensory stimulation 3. In this study we have sought to characterise further the input to edge cells by application of the glycine and v-aminobutyric acid (GABA) antagonists strychnine and bicuculline, respectively, during fictive locomotion induced either by sensory stimulation or by application of the excitatory amino acid agonist N-methyl-D,L-aspartate (NMDLA). Experiments were performed on newly transformed Petromyzon marinus. Two preparations were used. For sensory stimulation the spinal cord and notochord were isolated with a portion of the tailfin left intact 12. Both the region of stimulation and the region of ventral root recording were left attached to the notochord for support, and mounted dorsal side up in a sylgard-lined dish. In the intermediate region the notochord and meningeal
tissue were removed and this portion of the spinal cord was turned ventral side up on an additional layer of sylgard to ease edge cell impalement (Fig. 1A). The tailfin was stimulated mechanically with a pair of forceps and released after an episode of fictive locomotion was over. The second preparation included a section of spinal cord isolated entirely, desheathed and mounted ventral side up. Fictive locomotion was activated by bath application of 100 ktM N M D L A . Since the membrane potential of the edge cells (--60 to - - 7 0 mV) is near the chloride equilibrium potential, IPSPs are not easily observed. Therefore, in many experiments the edge cell was injected with chloride ions to reverse and magnify the IPSPs which occur during an episode of fictive locomotion (cf. ref. 3). Strychnine has been shown to abolish left-right alternation during sensory-induced fictive locomotion 2, or during excitatory amino-acid induced fictive locomotion immediately following strychnine pretreatment 5. Either method results in rhythmic coactivation of ventral roots of the same segment, with retention of the rostro-caudal phase lag characteristic of fictive locomotion 2, Fig. 1 shows the response recorded in edge cells during an episode of sensory-evoked fictive locomotion and during rhythmic coactivation following addition of 2-5 /~M strychnine. During sensory-evoked fictive locomotion the edge cell receives a tonic barrage of IPSPs coincident with the episode but not phasically related to the motor pattern recorded in the ventral roots (Fig. 1B), as reported previously 3. During the bath application of
Correspondence: T.L. Williams, Physiology Department, St. George's Hospital Medical School, Tooting, London SW17 ORE, U.K.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
138
A
Edge (ell l\'t, I]
Stimulus
/,
~pinal Cord
~4
lailfin
Control evr A f t e r Cl" I n j e c t i o n ee i
'1
~ll~i
, Ii~q
i
d~111lllil l
err" C Strychninei A f t e r CI- / Injection
[
. i
. [
. i
.
t
/
.
2s
oe I k i l l
l
l
1
tt
ivr
,~,
~,
~,
,~
~..
,.
. it
a,
CVF
~
.,~ ,~, w
J. l ~. at . .J
.
.
,,,,,
,
5mY] ls
Fig. 1. Effects of strychnine on synaptic inputs to edge cells during rhythmic activity. A: experimental preparation (see text). B: tailfin stimulation evokes an episode of fictive locomotion. Impaled edge cell has been injected with CI-- to reverse and enlarge IPSPs (cf. ref 3). C: after the application of 5/~M strychnine, tailfin stimulation evokes an episode of rhythmic left-right coactivation. Phasic depolarisations now occur in phase with the ventral root burst and a sustained depolarisation is seen during the episode of activity, ec, edge cell membrane potential; ivr, ipsilateral ventral root; cvr, contralateral ventral root.
strychnine (2/~M) edge cells exhibit a phasic depolarisation of the m e m b r a n e potential coincident with the ventral r o o t burst of the same segment (Fig. 1C). These phasic d e p o l a r i s a t i o n s are similar to, but much smaller than, the rhythmic depolarisations seen in ventral horn n e u r o n e s u n d e r the same conditions ~'2. This activity, in edge cells, occurs s u p e r i m p o s e d over a tonic depolarisation sustained t h r o u g h o u t the episode of activity (Fig. 1C). The activity seen in edge cells i m p a l e d with m i c r o e l e c t r o d e s containing KCI to load the neurones with
C I in o r d e r to depolarise the CI reversal potential, ~s the same as that recorded with microelectrodes containing potassium acetate. Wash-in of the G A B A A r e c e p t o r antagonist bicuculline m e t h o b r o m i d e (10-20 /~M) in addition to the strychnine had no effect on the response recorded in the edge cells. This indicates that the depolarisation does not result from G A B A A r e c e p t o r activation. It seems most likely, therefore, that this depolarisation is due to excitatory synaptic input. During fictive locomotion activated by the application of N M D L A there is an a p p a r e n t increase in edge cell synaptic activity (Fig. 2A) c o m p a r e d to that seen in the quiescent spinal cord after N M D L A washout (Fig. 2B). As during sensory-evoked fictive l o c o m o t i o n , this activity appears to be tonic rather than phasic. W h e t h e r there are EPSPs as well as IPSPs is not clear, since positive-going changes in m e m b r a n e potential m a y r e p r e s e n t the decay of IPSPs (see Fig. 2A). There were no large changes in the average m e m b r a n e potential before and after N M D L A application and small changes would be unreliable in these circumstances, so it is not known whether
A
Control Bicuculline
[3 A f t e r CI- I n j e c t i o n & B i c u c u l l i n e
ec
i~r ~ cvr
r" ~
"~
"*'
. . . ~ I , . , , , - ~ . - . * L . . . . - ~ . . . , , , ...............
I Stimulus C
NMDLA & B i c u e u l l i n e
A f t e r Ul- lnjeeLion
e(
]t1~ .I,L l~,.i_, IL . i iil~ l~lllih~ il~:,~ii i A NMDLA ivr
ivr 10mV]
_.
is
Washout ivr ......
e
c
, 4
~
,,I
2s
[ 10mV
Fig. 2. A: synaptic input to edge cells during fictive locomotion induced by bath application of 100 /~M NMDLA. B: synaptic activity in the same edge cell following NMDLA washout, ivr, ipsilateral ventral root; ec, edge cell membrane potential. Voltage calibration applies to edge cell records.
'
Fig. 3. Input to edge cells during fictive locomotion in the presence of bicuculline. A: fictive locomotion induced by tailfin stimulation in the presence of bicuculline, no chloride injection. B: tonic inhibition reversed by injection with C1--. Tailfin stimuli applied at arrow for A and B. C: fictive locomotion induced by hath application of 100/~M NMDLA in the presence of bicuculline in a different preparation. Edge cell (injected with cl'floride ions) receives bursts of large inverted IPSPs in phase with ipsilateral ventral root activity (out of phase with contralateral ventral root activity), ec, edge cell membrane potential; ivr, ipsilateral ventral root; cvr, contralateral ventral root. Voltage calibration applies to edge cell membrane potentials. 10/~M bicuculline present in all 3 experiments.
139 tonic inhibition or excitation also occurred. Strychnine application to the spinal cord during fictive locomotion induced by excitatory amino acid leads to tonic efferent activity recorded from the ventral roots 8. It has, therefore, not been possible to study the effects of strychnine on the increased synaptic activity illustrated in Fig. 2. Addition of bicuculline methobromide has no apparent effect on the pattern of ventral root activity recorded in NMDLA, though the application of GABA does slow the frequency of rhythmic activity8. Correspondingly, the presence of bicuculline during induction of fictive locomotion by tailfin stimulation prolongs the episode duration, but does not disrupt the pattern of ventral root activity significantly2. In the present study, we have recorded from edge cells in the presence of bicuculline. During fictive locomotion induced by tailfin stimulation, the edge cells undergo tonic inhibition in the presence of bicuculline, as shown in Fig. 3A, and previous injection of chloride ions makes this clearer (Fig. 3B). The inhibition is more marked than in the absence of bicuculline, as can be seen by comparing Fig. 3B with Fig. lB. In the absence of bicuculline, the IPSPs appear discrete, with little or no summation apparent (Fig. 1B), whereas in the presence of bicuculline, the IPSPs appear to occur upon a plateau. This was seen in all 15 of the edge cells tested. The particular cell exemplified in Fig. 3B was chosen to illustrate the association of this plateau with the fictive locomotor activity. During the first 4 bursts of ventral root activity, the input to the edge cell appears not to be associated with one ventral root more than the other. Following this rhythmic sequence, there is pause in the rhythmic ventral root activity, which is accompanied by a trailing off of the plateau of reversed inhibition in the edge cell. With the final burst of activity in the left (contralateral) ventral root, the plateau rises again briefly. Hence it would appear that fictive locomotion induced by sensory stimulation, gives rise to a tonic inhibitory input to edge cells which is masked by GABAergic transmission. It is not clear whether this represents the masking of a separate input to the edge cells or simply a gain control of the inhibitory input already apparent during sensory-induced fictive locomotion in the absence of blocking agents 3 (Fig. 1B). Intracellular recordings from edge cells during NMDLA-induced fictive locomotion following bath application of bicuculline also reveals inhibitory synaptic input. In contrast with the sensory-induced case, however, this inhibition is rhythmic, occurring in phase with the ipsilateral ventral root burst, out of phase with the
contralateral ventral root (3 edge cells in 3 preparations). This can best be seen in an edge cell that has been injected with chloride ions to accentuate the IPSPs (Fig. 3C). No large changes in resting potential were seen with bicuculline application, so it is not possible to determine whether the conversion of tonic inhibition to phasic is brought about by enhancement of the inhibition during the ipsilateral root burst, or repression during the contralateral root burst. Nonetheless, this result, compared with that obtained with sensory-induced fictive locomotion (Fig. 3B), indicates that bath application of NMDLA gives rise to cyclic activity in phase with fictive locomotion, which is not present during fictive locomotion induced by endogenous transmitter release. In summary, the following inputs to edge cells are apparent during fictive locomotion: (1) During an episode of sensory-evoked fictive locomotion, edge cells receive a volley of tonic IPSPs which are enhanced by bicuculline and blocked by strychnine. In addition, strychnine unmasks excitatory synaptic input to edge cells in phase with the ventral root activity. (2) During NMDLA induced fictive locomotion, edge cells receive tonic IPSPs. By contrast with the sensory-induced case, however, bicuculline converts this tonic activity to phasic activity. The effects of strychnine cannot be tested. The rhythmic synaptic input to edge cells seen during fictive locomotion induced by bath-applied NMDLA but not with sensory stimulation is of some concern. The bath application of excitatory amino acids and their analogues to the lamprey spinal cord has provided a fruitful means of study of this motor system, and it is clear that endogenous excitatory amino acid transmission plays an essential role in the spinal central pattern generator 9. However, it is clear from this study that, in at least one class of cells, bath application of NMDA-receptor agonists may produce rhythmic activity in phase with the ventral root bursts which does not occur during fictive locomotion induced by sensory stimulation. During tictive locomotion following sensory stimulation only endogenously released transmitter is available. The site of release, and therefore action, of endogenously released transmitter is, of course, much more specific than that of bath applied agonists. Thus data obtained during bath application of excitatory amino acids must be interpreted with care.
1 Alford, S. and Sigvardt, K.A., Excitatory neurotransmission activates voltage dependent properties in the spinal motor system of the lamprey, J. NeurophysioL, 62 (1989) 334-341.
2 Alford, S. and Williams, T., Endogenous activation of glycineand NMDA-receptors in the lamprey spinal cord during fictive locomotion, J. Neurosci., 9 (1989) 2792-28(10.
We thank Dr. Avis Cohen for supplying lampreys. This work was supported by the SERC (U.K.) and by NIH Grant NINCDSNS22360.
41) 3 Alford, S. and Williams, T.L., Inhibitory synaptic input to edge cells during fictive locomotion, Brain Research, 409 (1987) 139-142. 4 Cohen, A.H. and Wall6n, P., The neuronal correlate of locomotion in fish. 'Fictive swimming' induced in an in vitro preparation of the lamprey, Exp. Brain Res., 41 (1980) 11-18. 5 Cohen, A.H. and Harris-Warwick, R.M. Strychnine eliminates alternating motor output during fictive locomotion in the lamprey, Brain Research, 293 (1984) 164-167. 6 Grillner, S., McClellan, A. and Sigvardt, K., Mechanosensitive neurons in the spinal cord of the lamprey, Brain Research, 235 (1982) 169-173. 7 Grillner, S., McClellan, A.D. and Perret, C., Entrainment of the spinal central pattern generators for swimming by mechanosensitive elements in the lamprey spinal cord in vitro, Brain Research, 217 (1981) 380-386. 8 Grillner, S. and WaUEn, P., Does the central pattern generation for locomotion in lamprey depend on glycine inhibition?, Acta Physiol. Scand., 110 (1980) 103-105. 9 Grillner, S., Wall6n, R, Dale, N., Brodin, L., Buchanan, J. and Hill, R., Transmitters, membrane properties and network circuitry in the control of locomotion in lamprey, Trends
Neurosci., 10 (1987) 34-41. 10 Grillner, S., Williams, T. and Lagerb~ick, P.A., The edge cell, a possible intraspinal mechanoreceptor, Science, 223 (19841 500503. 11 McClellan, A.D., Descending control and sensory gating ol 'fictive' swimming and turning responses elicited in an in vitro preparation of the lamprey brainstem/spinal cord, Brain Research, 302 (1984) 151-162. 12 McClellan, A.D. and Grillner, S., Initiation and sensory gating of 'fictive' swimming and withdrawal responses in an in vitro preparation of the lamprey spinal cord. Brain Research, 269 (19831 237-250. 13 McClellan, A.D. and Grillner, S., Activation of 'fictive swimming' by electrical microstimulation of brainstem locomotor regions in an in vitro preparation of the lamprey central nervous system, Brain Research, 300 (1984) 357-361. 14 Rovainen, C.M., Synaptic interactions between retieulospinal neurons and nerve cells in the spinal cord of the sea lamprey, J. Comp. Neurol., 154 (1974) 207-224. 15 Rovainen, C.M., Synaptic interactions of identified nerve cells in the spinal cord of the sea lamprey, J. Comp. Neurol., 154 (1974) 189-206.
137
BRES 23957
Effects of bicuculline and strychnine on synaptic inputs to edge cells during fictive locomotion Simon Alford, Thelma L. Williams and Karen A. Sigvardt Department of Physiology, St. George's Hospital Medical School, London (U. K.)
(Accepted 24 October 1989) Key words: ~-Aminobutyric acid; Fictive locomotion; Strychnine; Lamprey; Edge cell; Mechanoreceptor; Spinal cord
Edge cells are mechanoreceptive neurones located in the lateral tracts of the lamprey spinal cord. Phasic activation of these cells by lateral bending can entrain the activity of the locomotor central pattern generator. During fictive locomotion induced by bath-applied NMDLA (N-methyl-D,L-aspartate) or sensory stimulation, edge cells receive synaptic input. In this paper we provide evidence that the tonic inhibition received during sensory-induced fictive locomotion is glycinergic. We also show that during fictive locomotion induced by application of NMDLA to the spinal cord, bicuculline unveils phasic synaptic activity in edge cells. During sensory-evoked fictive locomotion, by contrast, only tonic synaptic activity is apparent in the presence of bicuculline. Fictive swimming in the in vitro preparation of the lamprey spinal cord may be activated either by the bath application of certain excitatory amino acids 4 or by various means of sensory stimulation 1~-13. Such activity requires no sensory feedback to occur but may be entrained by the phasic activation of a class of sensory cells, the edge cells 6'7. These cells are identified by their lateral location in the spinal cord and their characteristic dendritic ramifications w,15. The membrane potential of the edge cell depolarises in response to stretch of the lateral margin of the spinal cord and hyperpolarises with release 1°. Edge cells receive both excitatory and inhibitory input from other cells of the spinal cord 14'15 and are tonically inhibited during episodes of fictive locomotion induced by sensory stimulation 3. In this study we have sought to characterise further the input to edge cells by application of the glycine and v-aminobutyric acid (GABA) antagonists strychnine and bicuculline, respectively, during fictive locomotion induced either by sensory stimulation or by application of the excitatory amino acid agonist N-methyl-D,L-aspartate (NMDLA). Experiments were performed on newly transformed Petromyzon marinus. Two preparations were used. For sensory stimulation the spinal cord and notochord were isolated with a portion of the tailfin left intact 12. Both the region of stimulation and the region of ventral root recording were left attached to the notochord for support, and mounted dorsal side up in a sylgard-lined dish. In the intermediate region the notochord and meningeal
tissue were removed and this portion of the spinal cord was turned ventral side up on an additional layer of sylgard to ease edge cell impalement (Fig. 1A). The tailfin was stimulated mechanically with a pair of forceps and released after an episode of fictive locomotion was over. The second preparation included a section of spinal cord isolated entirely, desheathed and mounted ventral side up. Fictive locomotion was activated by bath application of 100 ktM N M D L A . Since the membrane potential of the edge cells (--60 to - - 7 0 mV) is near the chloride equilibrium potential, IPSPs are not easily observed. Therefore, in many experiments the edge cell was injected with chloride ions to reverse and magnify the IPSPs which occur during an episode of fictive locomotion (cf. ref. 3). Strychnine has been shown to abolish left-right alternation during sensory-induced fictive locomotion 2, or during excitatory amino-acid induced fictive locomotion immediately following strychnine pretreatment 5. Either method results in rhythmic coactivation of ventral roots of the same segment, with retention of the rostro-caudal phase lag characteristic of fictive locomotion 2, Fig. 1 shows the response recorded in edge cells during an episode of sensory-evoked fictive locomotion and during rhythmic coactivation following addition of 2-5 /~M strychnine. During sensory-evoked fictive locomotion the edge cell receives a tonic barrage of IPSPs coincident with the episode but not phasically related to the motor pattern recorded in the ventral roots (Fig. 1B), as reported previously 3. During the bath application of
Correspondence: T.L. Williams, Physiology Department, St. George's Hospital Medical School, Tooting, London SW17 ORE, U.K.
0006-8993/90/$03.50 © 1990 Elsevier Science Publishers B.V. (Biomedical Division)
138
A
Edge (ell l\'t, I]
Stimulus
/,
~pinal Cord
~4
lailfin
Control evr A f t e r Cl" I n j e c t i o n ee i
'1
~ll~i
, Ii~q
i
d~111lllil l
err" C Strychninei A f t e r CI- / Injection
[
. i
. [
. i
.
t
/
.
2s
oe I k i l l
l
l
1
tt
ivr
,~,
~,
~,
,~
~..
,.
. it
a,
CVF
~
.,~ ,~, w
J. l ~. at . .J
.
.
,,,,,
,
5mY] ls
Fig. 1. Effects of strychnine on synaptic inputs to edge cells during rhythmic activity. A: experimental preparation (see text). B: tailfin stimulation evokes an episode of fictive locomotion. Impaled edge cell has been injected with CI-- to reverse and enlarge IPSPs (cf. ref 3). C: after the application of 5/~M strychnine, tailfin stimulation evokes an episode of rhythmic left-right coactivation. Phasic depolarisations now occur in phase with the ventral root burst and a sustained depolarisation is seen during the episode of activity, ec, edge cell membrane potential; ivr, ipsilateral ventral root; cvr, contralateral ventral root.
strychnine (2/~M) edge cells exhibit a phasic depolarisation of the m e m b r a n e potential coincident with the ventral r o o t burst of the same segment (Fig. 1C). These phasic d e p o l a r i s a t i o n s are similar to, but much smaller than, the rhythmic depolarisations seen in ventral horn n e u r o n e s u n d e r the same conditions ~'2. This activity, in edge cells, occurs s u p e r i m p o s e d over a tonic depolarisation sustained t h r o u g h o u t the episode of activity (Fig. 1C). The activity seen in edge cells i m p a l e d with m i c r o e l e c t r o d e s containing KCI to load the neurones with
C I in o r d e r to depolarise the CI reversal potential, ~s the same as that recorded with microelectrodes containing potassium acetate. Wash-in of the G A B A A r e c e p t o r antagonist bicuculline m e t h o b r o m i d e (10-20 /~M) in addition to the strychnine had no effect on the response recorded in the edge cells. This indicates that the depolarisation does not result from G A B A A r e c e p t o r activation. It seems most likely, therefore, that this depolarisation is due to excitatory synaptic input. During fictive locomotion activated by the application of N M D L A there is an a p p a r e n t increase in edge cell synaptic activity (Fig. 2A) c o m p a r e d to that seen in the quiescent spinal cord after N M D L A washout (Fig. 2B). As during sensory-evoked fictive l o c o m o t i o n , this activity appears to be tonic rather than phasic. W h e t h e r there are EPSPs as well as IPSPs is not clear, since positive-going changes in m e m b r a n e potential m a y r e p r e s e n t the decay of IPSPs (see Fig. 2A). There were no large changes in the average m e m b r a n e potential before and after N M D L A application and small changes would be unreliable in these circumstances, so it is not known whether
A
Control Bicuculline
[3 A f t e r CI- I n j e c t i o n & B i c u c u l l i n e
ec
i~r ~ cvr
r" ~
"~
"*'
. . . ~ I , . , , , - ~ . - . * L . . . . - ~ . . . , , , ...............
I Stimulus C
NMDLA & B i c u e u l l i n e
A f t e r Ul- lnjeeLion
e(
]t1~ .I,L l~,.i_, IL . i iil~ l~lllih~ il~:,~ii i A NMDLA ivr
ivr 10mV]
_.
is
Washout ivr ......
e
c
, 4
~
,,I
2s
[ 10mV
Fig. 2. A: synaptic input to edge cells during fictive locomotion induced by bath application of 100 /~M NMDLA. B: synaptic activity in the same edge cell following NMDLA washout, ivr, ipsilateral ventral root; ec, edge cell membrane potential. Voltage calibration applies to edge cell records.
'
Fig. 3. Input to edge cells during fictive locomotion in the presence of bicuculline. A: fictive locomotion induced by tailfin stimulation in the presence of bicuculline, no chloride injection. B: tonic inhibition reversed by injection with C1--. Tailfin stimuli applied at arrow for A and B. C: fictive locomotion induced by hath application of 100/~M NMDLA in the presence of bicuculline in a different preparation. Edge cell (injected with cl'floride ions) receives bursts of large inverted IPSPs in phase with ipsilateral ventral root activity (out of phase with contralateral ventral root activity), ec, edge cell membrane potential; ivr, ipsilateral ventral root; cvr, contralateral ventral root. Voltage calibration applies to edge cell membrane potentials. 10/~M bicuculline present in all 3 experiments.
139 tonic inhibition or excitation also occurred. Strychnine application to the spinal cord during fictive locomotion induced by excitatory amino acid leads to tonic efferent activity recorded from the ventral roots 8. It has, therefore, not been possible to study the effects of strychnine on the increased synaptic activity illustrated in Fig. 2. Addition of bicuculline methobromide has no apparent effect on the pattern of ventral root activity recorded in NMDLA, though the application of GABA does slow the frequency of rhythmic activity8. Correspondingly, the presence of bicuculline during induction of fictive locomotion by tailfin stimulation prolongs the episode duration, but does not disrupt the pattern of ventral root activity significantly2. In the present study, we have recorded from edge cells in the presence of bicuculline. During fictive locomotion induced by tailfin stimulation, the edge cells undergo tonic inhibition in the presence of bicuculline, as shown in Fig. 3A, and previous injection of chloride ions makes this clearer (Fig. 3B). The inhibition is more marked than in the absence of bicuculline, as can be seen by comparing Fig. 3B with Fig. lB. In the absence of bicuculline, the IPSPs appear discrete, with little or no summation apparent (Fig. 1B), whereas in the presence of bicuculline, the IPSPs appear to occur upon a plateau. This was seen in all 15 of the edge cells tested. The particular cell exemplified in Fig. 3B was chosen to illustrate the association of this plateau with the fictive locomotor activity. During the first 4 bursts of ventral root activity, the input to the edge cell appears not to be associated with one ventral root more than the other. Following this rhythmic sequence, there is pause in the rhythmic ventral root activity, which is accompanied by a trailing off of the plateau of reversed inhibition in the edge cell. With the final burst of activity in the left (contralateral) ventral root, the plateau rises again briefly. Hence it would appear that fictive locomotion induced by sensory stimulation, gives rise to a tonic inhibitory input to edge cells which is masked by GABAergic transmission. It is not clear whether this represents the masking of a separate input to the edge cells or simply a gain control of the inhibitory input already apparent during sensory-induced fictive locomotion in the absence of blocking agents 3 (Fig. 1B). Intracellular recordings from edge cells during NMDLA-induced fictive locomotion following bath application of bicuculline also reveals inhibitory synaptic input. In contrast with the sensory-induced case, however, this inhibition is rhythmic, occurring in phase with the ipsilateral ventral root burst, out of phase with the
contralateral ventral root (3 edge cells in 3 preparations). This can best be seen in an edge cell that has been injected with chloride ions to accentuate the IPSPs (Fig. 3C). No large changes in resting potential were seen with bicuculline application, so it is not possible to determine whether the conversion of tonic inhibition to phasic is brought about by enhancement of the inhibition during the ipsilateral root burst, or repression during the contralateral root burst. Nonetheless, this result, compared with that obtained with sensory-induced fictive locomotion (Fig. 3B), indicates that bath application of NMDLA gives rise to cyclic activity in phase with fictive locomotion, which is not present during fictive locomotion induced by endogenous transmitter release. In summary, the following inputs to edge cells are apparent during fictive locomotion: (1) During an episode of sensory-evoked fictive locomotion, edge cells receive a volley of tonic IPSPs which are enhanced by bicuculline and blocked by strychnine. In addition, strychnine unmasks excitatory synaptic input to edge cells in phase with the ventral root activity. (2) During NMDLA induced fictive locomotion, edge cells receive tonic IPSPs. By contrast with the sensory-induced case, however, bicuculline converts this tonic activity to phasic activity. The effects of strychnine cannot be tested. The rhythmic synaptic input to edge cells seen during fictive locomotion induced by bath-applied NMDLA but not with sensory stimulation is of some concern. The bath application of excitatory amino acids and their analogues to the lamprey spinal cord has provided a fruitful means of study of this motor system, and it is clear that endogenous excitatory amino acid transmission plays an essential role in the spinal central pattern generator 9. However, it is clear from this study that, in at least one class of cells, bath application of NMDA-receptor agonists may produce rhythmic activity in phase with the ventral root bursts which does not occur during fictive locomotion induced by sensory stimulation. During tictive locomotion following sensory stimulation only endogenously released transmitter is available. The site of release, and therefore action, of endogenously released transmitter is, of course, much more specific than that of bath applied agonists. Thus data obtained during bath application of excitatory amino acids must be interpreted with care.
1 Alford, S. and Sigvardt, K.A., Excitatory neurotransmission activates voltage dependent properties in the spinal motor system of the lamprey, J. NeurophysioL, 62 (1989) 334-341.
2 Alford, S. and Williams, T., Endogenous activation of glycineand NMDA-receptors in the lamprey spinal cord during fictive locomotion, J. Neurosci., 9 (1989) 2792-28(10.
We thank Dr. Avis Cohen for supplying lampreys. This work was supported by the SERC (U.K.) and by NIH Grant NINCDSNS22360.
41) 3 Alford, S. and Williams, T.L., Inhibitory synaptic input to edge cells during fictive locomotion, Brain Research, 409 (1987) 139-142. 4 Cohen, A.H. and Wall6n, P., The neuronal correlate of locomotion in fish. 'Fictive swimming' induced in an in vitro preparation of the lamprey, Exp. Brain Res., 41 (1980) 11-18. 5 Cohen, A.H. and Harris-Warwick, R.M. Strychnine eliminates alternating motor output during fictive locomotion in the lamprey, Brain Research, 293 (1984) 164-167. 6 Grillner, S., McClellan, A. and Sigvardt, K., Mechanosensitive neurons in the spinal cord of the lamprey, Brain Research, 235 (1982) 169-173. 7 Grillner, S., McClellan, A.D. and Perret, C., Entrainment of the spinal central pattern generators for swimming by mechanosensitive elements in the lamprey spinal cord in vitro, Brain Research, 217 (1981) 380-386. 8 Grillner, S. and WaUEn, P., Does the central pattern generation for locomotion in lamprey depend on glycine inhibition?, Acta Physiol. Scand., 110 (1980) 103-105. 9 Grillner, S., Wall6n, R, Dale, N., Brodin, L., Buchanan, J. and Hill, R., Transmitters, membrane properties and network circuitry in the control of locomotion in lamprey, Trends
Neurosci., 10 (1987) 34-41. 10 Grillner, S., Williams, T. and Lagerb~ick, P.A., The edge cell, a possible intraspinal mechanoreceptor, Science, 223 (19841 500503. 11 McClellan, A.D., Descending control and sensory gating ol 'fictive' swimming and turning responses elicited in an in vitro preparation of the lamprey brainstem/spinal cord, Brain Research, 302 (1984) 151-162. 12 McClellan, A.D. and Grillner, S., Initiation and sensory gating of 'fictive' swimming and withdrawal responses in an in vitro preparation of the lamprey spinal cord. Brain Research, 269 (19831 237-250. 13 McClellan, A.D. and Grillner, S., Activation of 'fictive swimming' by electrical microstimulation of brainstem locomotor regions in an in vitro preparation of the lamprey central nervous system, Brain Research, 300 (1984) 357-361. 14 Rovainen, C.M., Synaptic interactions between retieulospinal neurons and nerve cells in the spinal cord of the sea lamprey, J. Comp. Neurol., 154 (1974) 207-224. 15 Rovainen, C.M., Synaptic interactions of identified nerve cells in the spinal cord of the sea lamprey, J. Comp. Neurol., 154 (1974) 189-206.
