Mechanisms of pattern generation in co-cultures of embryonic spinal cord and skeletal muscle

~

Pergamon

Int. J. Devl Neuroscience, Vol. 14, No. 2, pp. 137-148, 1996 Elsevier ScienceLtd Copyright © 1996 ISDN 0736-5748(95)00093-3 Printed in Great Britain. All rights reserved 0736-5748/96 $15.00+0.00

MECHANISMS OF PATTERN GENERATION IN CO-CULTURES OF EMBRYONIC SPINAL CORD A N D SKELETAL MUSCLE

JIJRG STREIT Institute of Physiology, University of Bern, B~hlplatz 5, 3012 Bern, Switzerland (Received 13 March 1995; revised 17 August 1995; accepted 29 September 1995) A~traet--Spontaneous output patterns of embryonic spinal cord slices in vitro were investigated in order to study the formation of pattern-generating networks. Patterns of spontaneous contractions of skeletal muscle fibers were recorded in co-cultures of embryonic rat spinal cord, dorsal root ganglia and skeletal muscle. A part of these contractions was shown to be driven by spinal circuits. These neuron-driven activity patterns changed from random to rhythmic when the inhibitory synapses in the spinal cord were blocked by strychnine, bicuculline or both. Rhythmic patterns consisted of bursts of activity (tetanic contractions) followed by periods of relaxation. The transition from random to rhythmic patterns occurred during a period of heavily increased rate of activity. Presynaptic inhibition was not involved critically in the generation of rhythmic patterns. Such patterns were, however, modulated through muscarinic and ~adrenergic receptors. Neither NMDA nor glutamate nor its uptake blocker dihydrokainate induced rhythmic patterns of contraction, although NMDA in the presence of low magnesium increased moderately the rate of random activity. In order to study the size of pattern-generating networks, parts of the spinal cord slices were sectioned during rhythmic activity. Tangential cuts at the lateral or dorsal side of the slices reduced either the rate or the duration of the bursts or both. Sagittal cuts suppressed the activity almost totally. These findings suggest that the pattern generators in the slices consist of excitatory networks covering the entire slice, and that these networks reverberate following spontaneous activity of some distributed elements. Key words: Spinal cord, development, pattern generator, reverberating circuit, spontaneous activity, slice culture, skeletal muscle.

Rhythmic activity has been shown to arise in a variety of fetal or neonatal neural networks in vitro, either spontaneouslyy ~'32"33'34'47for review see Ref. 8 or induced by block of inhibitory pathways,1'15 or activation of excitatory pathways.4'5'H'~7'24'48 In the spinal cord, rhythmic activity was shown to underlie simple patterns of locomotion and therefore was called fictive locomotion) 6"~7 These patterns can be activated either by afferent synaptic input to the spinal networks, TM or by excitatory neurotransmitters. 5'24'41'sl They are believed to be produced by local spinal networks, the spinal pattern generators. 25 Although the spinal pattern generators of some lower vertebrates have been analyzed in great detail, ~2'19'3° less is known about the structure, function and development of mammalian pattern generators. From studies in the chick, it can be assumed that these networks are organized in the form of distributed coupled local oscillators.:3 It has been shown previously that functional synaptic contacts develop in a culture system of slices of embryonic rat spinal cord and dorsal root ganglia,2'~°'45and that spontaneous rhythmic activity appears in the network of spinal interneurons in these cultures when strychnine or bicuculline is applied) 2 These findings suggest that spinal oscillators, which are revealed by disinhibition, are formed in embryonic slices of the cord. In the present paper, the circumstances under which rhythmic activity appears in the spinal cord slices is described in more detail, in order to understand the structure of the spinal networks underlying this activity. It is shown that spontaneous contractions of skeletal muscle fibers in these cultures appear with the same rhythmic patterns reported previously for the synaptic input to individual motoneurons.42 These patterns of muscle activity are used as an indicator of the output of the spinal networks in order to answer the following three questions: first, does rhythmic activity critically depend on presynaptic inhibition not blocked by strychnine? Second, can rhythmic activity be induced by an increase in excitation instead of disinhibition? Third, what are the components and the minimum size of the spinal networks producing rhythmic activity? Some of the results have been published previously in abstract form.43 137

138

J. Streit E X P E R I M E N T A L P R O C E D U RES

Co-cultures of spinal cord slices, dorsal root ganglia and skeletal muscle were obtained from 13 14-day-old rat embryos as described previously. 4° The spinal cord slices of the embryos were cut in such a way that a majority of the cultures contained perispinal myotomes from which skeletal muscle fibers developed. The cultures were maintained in incubators with stabilized atmospheres and constant movement in rotating drums for up to 4 weeks. 3-~4 For the experiments, a coverslip containing the culture was placed into a Perspex chamber mounted on an inverted microscope and superfused with a bath solution containing (mM): NaCI 150, KCl 4, MgC12 1, CaC12 2-5, HEPES 5, Na pyruvate 2, glucose 5, at pH 7.4. Experiments were performed at room temperature (23 ___2 C ) after 2 4 weeks in vitro. The light scatter produced by contraction of bundles of skeletal muscle was recorded using a Nikon inverted microscope and a custom made device consisting of a photodiode and a preamplifier. The microscope was equipped with a stabilized power supply. The photodiode could be exposed to light coming from restricted parts of the muscle by use of a pinhole. The data were stored on a digital audio tape recorder (Sony, modified according to 27 and later digitized at 0.2-1 kHz and analyzed using a MacLab interface and software running on a Macintosh Quadra 700 Microcomputer. Further analysis was performed using the IGOR (Lake Oswego) software package. Activity rates were measured as the number of contractions times the average length of individual contractions per minute. The rate and duration of the contractions were measured within 4 min and compared to the average of the values obtained under control conditions before and after the exposure to receptor antagonists. Effects were expressed in percent change from these average control values. Significance was tested by comparing the mean effects from three to six experiments to the mean control value in these experiments using the Student's t-test. The mean control value was the mean of all control sequences in these experiments, expressed in percent change from the initial control period in each experiment. Significance was defined on the 5% level (two-tailed). For the extracellular focal stimulation, glass pipettes with tip sizes between 50 and 100 #m, filled with normal bath solution, were used. Short voltage pulses (0.1 msec) of various amplitudes (0.1 10 V) were delivered by a Master-8 stimulator (A.M.P.I., Israel). The NMDA, glutamate (Sigma, Switzerland) and all receptor and re-uptake antagonists (RBI, Natick, MA) were applied by exchanging the bath solution for a solution containing the drug. The bath solution was exchanged every 10--15 rain. All recordings were made in absence of solution flow. Transections of the cultures were made using a fine scalpel. Contraction patterns of the muscle bundles were compared before and 5-15 min after the transections, when a steady state was reached.

RESULTS

Spontaneous muscle activity About 50% of the slice cultures of embryonic spinal cord contained skeletal muscle fibers originating from myoblasts of the circumspinal non-neuronal tissue. 4° In most cultures, these muscle fibers formed one or several bundles which contracted spontaneously. These contractions were recorded optically using their light-scattering effect. They usually lasted for 0.2-1 see. They were driven either by the spontaneous activity of the motoneurons in the spinal slice or by spontaneous electrical activity of the muscle itself. Neuron-driven activity (see Fig. 1) was recognized by its typical patterns and its spectrum of pharmacological sensitivity. The typical patterns were irregular synchronized contractions of all muscle fibers of at least one bundle and often of several bundles. This pattern of activity resembled the irregular, 'random' synaptic activity observed in most motoneu r o n s . 42'45 In agreement with the findings in motoneurons, these random patterns of muscle activity could be turned into episodic, 'rhythmic' patterns by the application of strychnine (1-10 ktM, n = 45), bicuculline (10 #M, n = 7) or both (n = 1). Rhythmic patterns consisted of synchronous tetanic contractions of the muscle fiber bundles lasting for several seconds, followed by relatively long periods of relaxation (see Fig. 1). Additional evidence for the neuron-driven origin of the spontaneous muscle contractions comes from the observation that these contractions were blocked

Pattern generation in embryonic spinal cord and muscle

139

1 m:m

control

+ strychnine

+ CNQX

_

_ _ "

-

. .

.

.

.

.

.

.

.

.

.

.

.

.

.

---7

. . . .

~ : - - -

~ _ 2 " d _

. . . .

5s Fig. 1. Overview of a slice culture of spinal cord (SC), dorsal root ganglia (SG) and bundles of skeletal muscle (M), 18 days in vitro. Below: optical recordings from the bundle of muscle fibers indicated by the circle, measured in control conditions, in presence of strychnine (10 pM) and in presence of CNQX (10 pM).

by C N Q X , a blocker o f glutamatergic neurotransmission within spinal cord circuits (Fig. 1) and by the nicotinic antagonist curare (experiments not shown, see4S). Neuron-driven muscle contractions also could be evoked by electrical stimulation at various sites o f the spinal cord slice. Besides the neuron-driven muscle activity, muscle contractions originating f r o m electrical activity o f the muscle fibers themselves were observed. 45 Typical patterns were frequent, weak and asyn-

140

J. Streit

IIIIIII

I

IIIIIIIII

lOs Fig. 2. Three subsequent segments of continuous optical recordings of muscle contractions induced by electrical stimulation at one discrete site in the ventral horn of the spinal slice. Vertical markers indicate the times of the electrical stimuli (series of stimuli are indicated by horizontal lines connecting the markers). Note that a regular pattern of spontaneous contractions appears under stimulation which persists after offset of stimulation. Following the cessation of the spontaneous activity, it cannot be triggered again by the same stimuli for several minutes. The experiment was performed in presence of CNQX to suppress the activity of the spinal network in the culture.

chronous contractions of single muscle fibers which could not be influenced by electrical stimulation of the spinal cord or by CNQX, strychnine or bicuculline. In addition, very regular patterns of contractions at 1 Hz were seen occasionally (see Fig. 2), which also were insensitive to CNQX, strychnine or bicucuUine. However, these patterns could be triggered by electrical stimulation at one defined site in the ventral part of the spinal cord slices. Electrical stimuli at these sites evoked stimulus-locked contractions which triggered long trains of regular spontaneous contractions at about 1 Hz. This type of muscle activity was independent of the stimulation and ceased spontaneously after 1-2 min. After that, the rhythmic contractions could not be triggered again by the same stimuli for several minutes. These patterns were probably produced by pacemaker potentials in muscle fibers innervated by silent motoneurons? 5 The three described patterns of activity were mutually exclusive, so that neuron-driven patterns were not contaminated by spontaneous activity of the muscle itself. Therefore, the contraction patterns of neuron-driven activity could be taken as an indicator of the output of the spinal networks. In this way, spinal cord activity could be observed with a minimum of experimental interference. When strychnine was applied to the cultures, the overall rate of activity first increased dramatically (see Fig. 3). During this period of high activity, the change from random to rhythmic patterns of activity occurred. Rhythmic patterns initially consisted of long burst (tetani) with short interburst intervals. During the first 5 min of rhythmic activity, the burst duration decreased gradually to a steady mean value of 5.1 + 0.7 sec ( + S.D., n = 9, range: 3.8~.3 sec) and the interval duration increased to a mean value of 8.6+4.5 sec (±S.D., n = 9, range: 3.7-14.4 sec). The burst rate was 4.7 +_ 1.3 per min ( + S.D., n = 9, range: 3 6.3 sec). The burst duration was correlated with the preceding, but not with the following, interval duration. After removal of strychnine, steadystate rhythmic activity was maintained for the first 10-15 min, followed by a slow decrease in burst and interval duration for the next 20 min. The burst duration was reduced gradually to single twitches, resulting in random activity after about 40 min (see Fig. 3).

The role of presynaptic inhibition in rhythmogenesis Various receptors are described or could potentially be involved in presynaptic inhibition in the spinal cord besides GABA A receptors: GABA B, muscarinic, ~-adrenergic, adenosine, 5-HT (serotonine), opioid and metabotropic glutamate receptors} 3'35"36'3s'39'46'52 In order to determine whether rhythmic activity is a function of purely excitatory networks, it had to be ruled out that any of these receptors was involved critically in rhythmic activity. To do this, rhythmic activity was first induced by strychnine (10-20 pM). Once steady-state bursting activity had developed, the

Pattern generation in embryonic spinal cord and muscle

a

141

+ strychnine 1.0-

;a~ 0.8-ov..~ ~ 0.6 ..~

•

•

~

•

~

""

OII

~..,.

•

0 0.4=

~0~

•

°:0°:

0 0.2=

o

o

0

!

I

2

4

7

ol

~

E

°o ~ °°° °~ ° ~

0.0 --

15

I

I

I

6

8

lO

~

~_o ~

12

~"

time (min)

D

"-"

+ strychnine

00.6- ' "~. 0.5-

7

-

1

0

~

f 50 0

q'~ 0,3-0 0

2O

0.10,0-

I

0

10

I

I

I

20

30

40

0 "~

time (min) Fig. 3. Timecourseof strychnine-inducedeffectsand recoveryfrom strychnineapplication.Rate of activity (solid line), burst length (filled circles) and interburst interval length (open circles) before, during and following the application of strychnine (10 mM) in one experiment. The rate of activity was determined during random and rhythmic activity (see experimental procedures). The burst and interval lengths could only be determined during rhythmic activity. The transition from random to rhythmic activity is therefore indicated by the beginning of the circles. (a) Time course of strychnine effects. (b) Recovery from strychnine application.

antagonists of the described receptors were applied (usually 10 #M) by a solution change in the bath. During the solution changes, the burst duration decreased transiently while the burst rate increased. Following the solution change, a new steady state developed within 2-3 min. All measurements were made in the steady state. The antagonists were applied in the absence or presence of strychnine. No difference between these two modes of application was seen since changes in activity patterns after removal of strychnine appeared only slowly (see Fig. 3). The GABA B antagonists phaclofen (n = 3) and saclofen (n = 3), the adenosine antagonist l-allyl-3,7-dimethyl-8-sulfophenylxanthine (n = 3), the 5-HT antagonist mianserine (n = 3), the opioid antagonist naloxone (n = 2) and the metabotropic glutamate antagonist L-2-amino-3-phosphonopropionic acid (L-AP3, n = 3) had no effect on the patterns of activity (see Figs 4 and 5). The muscarinic antagonist atropine increased the burst rate to 183_ 88% and decreased the burst duration to 55_ 2% (n = 5) of control values. The ct-adrenergic antagonist phentolamine increased the burst rate to 156_ 81% without a significant decrease of the burst duration (n = 4). None of these substances was able to interrupt the rhythmic patterns of activity. The only substance which clearly and reversibly interrupted the rhythmic activity patterns was caffeine (20-40 #M, n -- 6).

142

J. Streit

control ~~~~i saclofen~ washout~ caffeine*~~~ washout~'J~~~ lOs

Fig. 4. Effect of saclofen (10 #M) and caffeine (20 #M) on strychnine-induced rhythmic patterns of muscle contractions. Control and washouts were recorded in presence of strychnine (10/~M).

burst rate 11111

II

0

burst

length 50

i

Fig. 5. Mean effects of various antagonists to receptors involved in presynaptic inhibition on burst rate and burst duration of strychnine-induced rhythmic activity of muscle fibers. All results are expressed as a percentage of the control values. Light bars stand for rhythmic patterns, dark bars for random patterns. Stars indicate statistical significance; n = 2~5. Xanthine stands for the derivative 1-allyl-3,7-dimethyl-8sulfophenylxanthine. For details, see text.

The role of excitatory amino acids in rhythmogenesis It was investigated, w h e t h e r r h y t h m i c c o n t r a c t i o n p a t t e r n s also can be e v o k e d by a c t i v a t i o n o r strengthening o f e x c i t a t o r y s y n a p t i c p a t h w a y s . It was f o u n d t h a t N M D A ( 1 0 / a M ) increased the rate o f c o n t r a c t i o n s w i t h o u t i n d u c i n g r h y t h m i c activity. I n the presence o f n o r m a l extracellular m a g n e s i u m , only a t r a n s i e n t increase in the rate o f activity was seen (n = 5, see F i g s 6 a n d 7). W h e n m a g n e s i u m was r e m o v e d f r o m the extracellular solution, N M D A i n d u c e d a persistent increase o f 157 + 2 8 % (n = 5) in the rate o f c o n t r a c t i o n s ; however, n o r h y t h m i c activity was observed. T h e

Pattern generation in embryonic spinal cord and muscle

143

control

NMDA

0Mg.. 0 Mg++ + NMDA wash 0 Mg++ 0 Mg÷++ NMDA + APV 5s Fig. 6. Effects of N M D A (10/~M) with and without magnesium and the N M D A antagonist APV (10/~M) on the patterns of spontaneous muscle contractions.

event frequency

m'-,_

I

Fig. 7. Mean effects of excitatory amino acids on the rate of contractions during random activity of muscle fibers. All results are expressed as a percentage of the control values. Stars indicate statistical significance. Dark bars stand for random patterns; n = 3-7. For details, see text.

effect of N M D A in low magnesium was reversible and prevented by APV (10-20/~M, n = 4). Low magnesium alone had no effect on the rate of contractions (see Fig. 6). Glutamate at low doses (10 /~M) gave inconsistent results which appeared as an increase or decrease in the rate of contractions. On average no effect was seen (n = 7, see Fig. 7). At higher doses (0.1-1 mM), glutamate clearly decreased the rate of contractions by up to 97% (n = 6). The glutamate uptake inhibitor dihydrokainate (DHK, 10 #M) had no effect on the rate of contractions in three experiments either alone or in the presence of additional glutamate. In no experiment did glutamate o r D H K induce rhythmic contraction patterns. That muscle activity was mediated by glutamatergic neurotransmission was confirmed by the complete disappearance of spontaneous contractions by the non-NMDA receptor antagonist DNQX (10/~M, see Fig. 7).

144

J. Streit

Location of the networks involved in rhythmogenes& The minimal size of the spinal networks involved in rhythmic activity was investigated by cutting away parts of the spinal cord in culture using small scalpels and observing the patterns of contractions before and after these cuts. All these experiments were made in the presence of strychnine after the development of a stable bursting pattern. Cuts were made at two locations in the spinal cord (see Fig. 8): mid-sagittally along the central fissure, leaving two independent halves of the slices, or tangentially. The tangential cuts were made at the rim, around 100-200/~m inside the lateral or dorsal spinal cord. Mid-sagittal cuts terminated all spontaneous bursts of contractions in five out of six experiments for the rest of the observation period (up to 1 hr). In the remaining experiment, the mean burst rate was reduced from 7.5 bursts/min to 1.9 bursts/min and the mean burst duration was reduced from 3.2 to 1.5 sec. Tangential cuts produced four possible sorts of effects: in four experiments the mean burst rate and duration was reduced (to 52 and 63% of control values, respectively). In three experiments, the burst rate was reduced (48 %) while the burst duration remained the same or was even increased (111%). In one experiment, the burst rate was unchanged while the burst duration was reduced, and in one experiment nothing happened. Synchronous reduction of the mean burst rate and burst duration may have two possible reasons: either bursts of long duration are eliminated exclusively or the overall burst rate and the duration of all bursts

/,

OlOt ! \.

s.

ov-burst length (e)

b

5s

% Fig. 8. Effects of disconnection of pa~s of the spinal slices on st~chnine-induc~ rhyth~c actifity of muscle fi~rs. (a) A tangential cut (indicated by the fine in the sketch on the ri~t) reduces the rate and length of bursts in one experiment. Upper trace: ~fore the cut; lower trace: after the cut. Below: histo~ams of the length of the bursts ~fore (dotted line) and after (sofid line) the cut. Note the shift in the peak of the histogram, suggesting that all bursts became shorter. (b) A sagittal cut (hemisection) terminates all activity in the same experimentas shown in (a). (c) Two sequential tangential cuts with different effects in another experiment: the first cut decreases the length of the bursts with no change in the rate, the second decreases the burst rate with no additional change in the length.

Pattern generation in embryonic spinal cord and muscle

145

are reduced simultaneously. In order to distinguish between these two possibilities, the histograms of the burst lengths before and after the cuts were compared (see Fig. 8a). The shift of the histogram clearly favored the second possibility.

DISCUSSION In this paper, previous findings of rhythmic activity in spinal cord slice cultures were confirmed and extended.42The presented results lead to three new conclusions: first, rhythmic activity appears not only in the form of synaptic input to single motoneurons but also in the form of the output from the whole spinal cord slice to bundles of muscle fibers. Second, rhythmic activity critically involves a neural network covering the whole spinal cord slice. It is based on independent mechanisms controlling burst rate and burst duration. Third, presynaptic inhibition is not involved critically in rhythm generation, although the latter may be modulated via muscarinc or ct-adrenergic receptors. These findings confirm that rhythmic activity in the slice cultures is produced by predominantly excitatory networks. On the other hand, it cannot be induced by excitatory amino acids alone, even under conditions where the rate of activity is increased.

Rhythmic contractions of muscle fiber bundles It has been shown previously that spontaneous synaptic potentials recorded in motoneurons of spinal cord slice cultures are grouped in rhythmic patterns when the cultures are treated with strychnine or bicuculline. 42 However, since the recordings were made in single cells, it was not possible to ascertain whether the rhythmic activity in a pool of motoneurons in one culture was synchronous, as it is described for 'randomized' spinal networks. 2° In many cultures, however, bundles of cross-striated muscle fibers were present which were innervated by motoneurons and contracted synchronously.45Assuming that these bundles were innervated by a pool ofmotoneurons in the slice, their pattern of contractions reflect the output pattern of the motoneuronal pool. Bicuculline or strychnine induced rhythmic patterns of muscle contractions, suggesting a synchronous rhythmic activity pattern of the motoneuronal pool. However, it has to be taken into account that the pool ofmotoneurons innervating the muscle fiber bundles may be much smaller in individual cultures than the number of morphologically identified motoneurons, since many of the latter did not have contact with muscle fibers. 29It may even consist of a single motoneuron. Nevertheless, the experiments show that the output of spinal cord slices to muscle fiber bundles showed the same rhythmic patterns as the synaptic input to individual motoneurons. Single-twitch contractions of muscle fibers in culture were slow (see Figs 1, 2 and 4), certainly in part due to the low temperature (23°C). Therefore, contractions became tetanic at neuronal stimulation frequencies of above 4-6 Hz, which was the frequency range of rhythmic synaptic input to motoneurons. Size and components of spinal pattern generators in culture Spinal pattern generators usually produce alternating rhythmic patterns of activity on the two sides of the body and in agonist/antagonist muscle groups. ~6 Recent findings suggest that the rhythmic activity is produced by independent local networks (spinal oscillators) and that locomotion patterns are generated by the coupling of these spinal oscillators.23'49In the slice cultures, alternating activity was never observed. This observation may be explained by a synchronization of two independent oscillators by strychnine,zz However, in this case, one would expect that one half of the spinal slice is sufficient to produce the rhythmic activity. Clearly this was not the case, since the pattern generation was critically dependent on the integrity of the whole spinal cord slice. Therefore, it must be assumed that the spinal pattern generators in culture lack the symmetric structure necessary for alternating activity. At least the experiments show that one spinal half alone is not able to generate rhythms once it is separated acutely from the other half. The reason for this may be that the synaptic coupling between the two spinal hemispheres in culture is better than in vivo due to the much broader contact area between them (see Fig. 1). It has been suggested previously that the origin of the rhythmic activity is an oscillator network and not a cellular oscillator. 4z The finding that separation of marginal parts of the spinal network could influence the rate and the duration of the bursts independently of each other now definitely

146

J. Streit

rules out the possibility of bursting pacemaker cells as the origin of the activity patterns. It shows clearly that these two parameters are controlled by two different mechanisms and not by one bursting pacemaker cell. The only possible interpretation of these findings in terms of a bursting pacemaker model would be a selective disconnection of some of several variable and distributed pacemakers. However, the histogram of the burst duration clearly shows that this is not the case (see Fig. 8). An alternative model is that the burst duration is controlled by the reverberation of activity within the whole network, terminated by a gradual decrease in synaptic efficacy and/or in cellular excitability. ~ This model implies that bursts are initiated by one or a few spontaneous action potentials in several spinal cells, since the interburst intervals are much too long to be explained by 'silent' reverberation of activity within the spinal network not seen by the motoneurons. Several of these cells have to be distributed over the spinal network. Their role, however, is not that of a pacemaker but of a source of input to the network. The model was described already and favored by Corner and Crain to explain the origin of bioelectrical activity in a similar preparation? It best explains the experimental findings after disconnection of parts of the slices. Disconnection of one or several of the spontaneously active cells would lead to a decrease in burst rate, while disconnection of parts of the reverberating circuits would decrease the burst duration. In the case where the reverberating circuits stay intact, the decrease in burst rate leads to an increase in burst duration since the burst duration is correlated positively to the preceding interburst interval, due to the time requirement for recovery from synaptic depression. 42 Spontaneous oscillations based on reverberation of activity in a random excitatory network in the presence of strong synaptic depression have been found in an analytical computer model. 2°'37

Disinhibition but not strengthening of excitation induces rhythmic patterns In agreement with previous findings, rhythmic activity was most reliably induced by strychnine, bicuculline or a combination of the two. This suggests that rhythmic activity appears in cultures where inhibitory synaptic contacts are drastically reduced or blocked (disinhibited cultures). In the present study, this hypothesis was confirmed by excluding the possibility that presynaptic inhibition may be a prerequisite for rhythmic activity. However, these findings only show that within isolated spinal slice cultures these receptors are not involved critically in pattern generation. They do not exclude the possibility that these receptors are present in our cultures, nor do they exclude an involvement of these receptors in the modulation of spinal patterns; 5-HT, for example, has been shown to be involved in pattern generation in many spinal preparations. 4~'5° The main source of 5HTergic neurons is the brainstem, so probably they were not present in our cultures. The experiments suggest a modulation of pattern generation via muscarinic and ~-adrenergic receptors. Within the proposed model this effect can be interpreted in terms of an increase (direct or indirect) in the excitability of the spontaneously active neurons produced by the antagonists of these receptors, for example due to a decrease in a potassium conductance. Cholinergic as well as tyrosine hydroxylase positive neurons have been detected by histochemistry in the slice cultures (results not shown). The effect of caffeine probably is not related to its action as adenosine receptor antagonist since the more potent antagonist l-allyl-3,7-dimethyl-8-sulfophenylxanthine had no effect. More likely it is explained by the mobilization of calcium from intracellular calcium stores by caffeine, suggesting that intracellular calcium is involved critically in the mechanism of pattern generation. This is not surprising, since excitability, propagation of action potentials and synaptic efficacy all are controlled by intracellular calcium. 28 In many isolated spinal cord preparations, rhythmic activity can be induced by increased concentrations of N M D A , glutamate or glutamate uptake blockers. 4'5'~7'24'z6In a previous study, it was shown that in embryonic spinal cord slice cultures, N M D A antagonists did not interrupt rhythmic activity and that N M D A did not induce rhythmic activity measured as synaptic input to motoneurons. 43 In confirmation of these findings, neither N M D A nor glutamate nor the glutamate uptake blocker dihydrokainate induced rhythmic contractions of muscle fiber bundles. Glutamate at high doses drastically decreased the rate of spontaneous activity, probably caused by a desensitization of AMPA receptors, v and after prolonged exposure probably also due to cell death. 6 The N M D A , in the presence of low magnesium, increased the rate of random spontaneous activity, suggesting that N M D A receptors were expressed and played an important role provided their magnesium block was removed. 3~'53 However, even under these favorable conditions, N M D A did not induce

P a t t e r n generation in embryonic spinal cord a n d muscle

147

rhythmic activity. Looking at the time course of the appearance of rhythmic patterns under strychnine (Fig. 3), it is noticed that the rate of activity is initially heavily increased and that the transition from random to rhythmic activity occurs during this period of high activity. In the light of these findings it is possible that N M D A does not increase the overall rate of activity enough to provoke the transition to rhythmic patterns, since it potentiates excitatory and inhibitory transmission in parallel. In summary, the results presented here suggest that in embryonic spinal networks, in vitro rhythmic activity appears when the net excitatory transmission exceeds a certain level, and that the pattern generation is based on spontaneous activity of some distributed spinal neurons and reverberation of this activity within the entire spinal network. Acknowledyements--I wish to thank K. Rufener for excellent preparation and support of the cultures, D. DeLimoges for

the construction of critical parts of the equipment and H.-R. L~lscher and H. P. Clamann for helpful comments on the manuscript. This work was supported by the Swiss National Fund No. 31-27553.89 and 31-39419.93.

REFERENCES 1. Alford S., Sigvardt K. A. and Williams T. L. (1990) GABAergic control of rhythmic activity in the presence of strychnine in the lamprey spinal cord. Brain Res. 506, 303-306. 2. Baker R. E., Habets A. M. M. C., Brenner E. and Corner M. A. (1982) Influence of growth medium, age in vitro and spontaneous bioelectric activity on the distribution of sensory ganglion-evoked activity in spinal cord explants. Devl Brain Res. 5, 329-341. 3. Braschler U. F., Iannone A,, Spenger C., Streit J. and Ltischer H.-R. (1989) A modified roller tube technique for organotypic cocultures of embryonic rat spinal cord, sensory ganglia and skeletal muscle. J. Neurosci. Meth. 29, 121 129. 4. Brodin L., Tossman U., Ohta Y., Ungerstedt U. and Grillner S. (1988) The effect of an uptake inhibitor (dihydrokainate) on endogenous excitatory amino acids in the lamprey spinal cord as revealed by microdialysis. Brain Res. 458, 166-169. 5. Cazalets J. R., Grillner P., Menard I., Cremieux J. and Clarac F. (1990) Two types of motor rhythm induced by NMDA and amines in an in vitro spinal cord preparation of neonatal rat. Neurosci. Lett. 111, 116-121. 6. Choi D. W. (1987) Ionic dependence of glutamate neurotoxicity. J. Neurosci. 7, 369-379. 7. Colquhoun D., Jonas P. and Sakmann B. (1992) Action of brief pulses of glutamate on AMPA/kainate receptors in patches from different neurones of rat hippocampal slices. J. Physiol. 458, 261 287. 8. Comer M. A. (1994) Reciprocity of structure-function relations in developing neural networks: the odyssey of a selforganizing brain through research fads, fallacies and prospects. In Self-organizing Brain:from Growth Cones to Functional Networks (eds Van Pelt J., Comer M. A., Uylings H. B. M. and Dasilva F. H. L.), pp. 3-31. Elsevier, Amsterdam. 9. Corner M. A. and Crain S. M. (1972) Patterns of spontaneous bioelectric activity during maturation in culture of fetal rodent medulla and spinal cord tissues. J. Neurobiol. 3, 25-45. 10. Crain S. M. and Peterson E. R. (1967) Onset and development of functional interneuronal connections in explants of rat spinal cord-ganglia during maturation in culture. Brain Res. 6, 750-762. 11. Dale N. (1986) Excitatory synaptic drive for swimming mediated by amino acid receptors in the lamprey. J. Neurosci. 6, 2662-2675. 12. Dale N. and Roberts A. (1984) Excitatory amino acid receptors in Xenopus embryo spinal cord and their role in the activation of swimming. J. Physiol. 348, 527-543. 13. Duggan A. W., Hall J. G., Foong F. W. and Zhao Z. Q. (1985) A differential effect of naloxone on transmission of impulses in primary afferents to ventral roots and ascending spinal tracts. Brain Res. 344, 316-321. 14. G/ihwiler B. H. (1981) Organotypic monolayer cultures of nervous tissue. J. Neurosci. Meth. 4, 329-342. 15. G~ihwiler B. H. and Dreifuss J. J. (1980) Transition from random to phasic firing induced in neurons cultured from the hypothalamic supraoptic area. Brain Res. 193, 415-425. 16. Grillner S. (1975) Locomotion in vertebrates: central mechanisms and reflex interaction. Physiol. Rev. 55, 247-304. 17. Grillner S. (1985) Neurobiological bases of rhythmic motor acts in vertebrates. Science 228, 143-149. 18. Grillner S. and Wallrn P. (1985) Central pattern generators for locomotion, with special reference to vertebrates. Ann. Rev. Neurosci. g, 233-261. 19. Grillner S., Wallrn P., Brodin L. and Lansner A. (1991) Neuronal network generating locomotor behavior in lamprey: circuitry, transmitters, membrane properties, and simulation. Ann. Rev. Neurosci. 14, 169-199. 20. Gross G. W., Rhoades B. K. and Kowalski J. M. (1993) Dynamics of burst patterns generated by monolayer networks in culture. In Neurobionics (eds Bothe H.-W., Samii M. and Eckmiller R.), pp. 89-121. Elsevier, Amsterdam. 21. Habets A. M. M. C., Van Dongen A. M. J., Van Huizen F. and Comer M. A. (1987) Spontaneous neuronal firing patterns in fetal rat cortical networks during development in vitro: a quantitative analysis. Exp. Brain Res. 69, 43-52. 22. Hagevik A. and McClellan A. D. (1994) Coupling of spinal locomotor networks in larval lamprey revealed by receptor blockers for inhibitory amino acids: neurophysiology and computer modeling. J. Neurophysiol. 72, 1810-1829, 23. Ho S. and O'Donovan M. J. (1993) Regionalization and intersegrnental coordination of rhythm-generating networks in the spinal cord of the chick embryo. J, Neurosci. 13, 1354~1371. 24. Kudo N. and Yamada T. (1987) N-methyl-D,L-aspartate-induced locomotor activity in a spinal cord-hindlimb muscles preparation of the newborn rat studied in vitro. Neurosci. Lett. 75, 43-48. 25. Landmesser L. T. and O'Donovan M. J. (1984) Activation patterns of embryonic chick hindlimb muscles recorded in ovo and in an isolated spinal cord preparation, J. Physiol. 347, 189-204. 26. Lev-Tov A., Smith J. C. and Schmidt B. J. (1992) Neurochemical mechanisms for generation and transmission of locomotor rhythm in vitro (abstract). Soc. Neurosci. Abstr. 18, 317.

148

J. Streit

27. LOscher C., Frei P. and de Limoges D. (1992) Modification of a consumer digital audio tape (DAT) for analog data recording. J. Neurosci, Meth. 45, 155--158. 28. Liischer C., Streit J., Lipp P. and L~scher H.-R. (1994) Action potential propagation through embryonic dorsal root ganglion cells in culture. II. Decrease of conduction reliability during repetitive stimulation. J. Neurophysiol. 72, 634643. 29. Li~scher H.-R. and Streit J. (1992) A novel in vitro approach for studying the segmental motor system. N I P S 7, 249 253. 30. MacDermott A. B. and Dale N. (1987) Receptors, ion channels and synaptic potentials underlying the integrative actions of excitatory amino acids. T I N S 10, 280-288. 31. Mayer M. L. and Westbrook G. L. (1985) The action of N-methyl-D-aspartic acid on mouse spinal neurones in culture. J. Physiol. 361, 65-90. 32. O'Brien R. J. and Fischbach G. D. (1986) Isolation of embryonic chick motoneurons and their survival in vitro. J. Neurosci. 6, 3265-3274. 33. Provine R. R. (1972) Ontogeny of bioelectric activity in the spinal cord of the chick embryo and its behavioral implications. Brain Res. 41, 365 378. 34. Ransom B. R., Christian C. N., Bullock P. N. and Nelson P. G. (1977) Mouse spinal cord in cell culture. II. Synaptic activity and circuit behavior. J. Neurophysiol. 40, 1151-1161. 35. Salter M. W., Dekoninck Y. and Henry J. L. (1993) Physiological roles for adenosine and ATP in synaptic transmission in the spinal dorsal horn. ProRr. Neurobiol. 41, 125-156. 36. Scanziani M., G~ihwiler B. H. and Thompson S. M. (1993) Presynaptic inhibition of excitatory synaptic transmission mediated by alpha-adrenergic receptors in Area-CA3 of the rat hippocampus in vitro. J. Neurosci. 13, 5393-5401. 37. Senn W., Wyler K., Streit J., Larkum M., Liischer H.°R., Mey H., Milller L., Stainhauser D., Vogt K. and Wannier T. Dynamics of a random neural network with synaptic depression. Neural Network (in press). 38. Sillar K. T. and Simmers A. J. (1994) Presynaptic inhibition of primary afferent transmitter release by 5-hydroxytryptamine at a mechanosensory synapse in the vertebrate spinal cord. J. Neurosci. 14, 2636-2647. 39. Sladeczek F., Momiyama A. and Takahashi T. (1993) Presynaptic inhibitory action of a metabotropic glutamate receptor agonist on excitatory transmission in visual cortical neurons. Proc. R. Soc. Lond. B 253, 297-303. 40. Spenger C., Braschler U. F., Streit J. and Ltischer H.-R. (1991) An organotypic spinal cord-dorsal root ganglionskeletal muscle coculture of embryonic rat. 1. The morphological correlates of the spinal reflex arc. Eur. J. Neurosci. 3, 1037 1053. 41. Sqallihoussaini Y., Cazalets J. R. and Clarac F. (1993) Oscillatory properties of the central pattern generator for locomotion in neonatal rats. J. Neurophysiol. 70, 803-813. 42. Streit J. (1993) Regular oscillations of synaptic activity in spinal networks in vitro. J. Neurophysiol. 70, 871-878. 43. Streit J. (1994) Rhythmic output ofembryonic spinal networks in culture is induced bydisinhibition but not by increased excitation. Soc. Neurosci. Abstr. 20, 1594. 44. Streit J., Liischer C. and Liischer H.-R. (1992) Depression of postsynaptic potentials by high frequency stimulation in embryonic motoneurons grown in spinal cord slice cultures. J. Neurophysiol. 68, 1793-1803. 45. Streit J., Spenger C. and Liischer H.-R. (1991) An organotypic spinal cord~lorsal root ganglia-skeletal muscle coculture of embryonic rat. II. Functional evidence for the formation of spinal reflex arcs in vitro. Eur. J. Neurosci. 3, 1054-1068. 46. Stuart G. J. and Redman S. J. (1992) The role of GABA A and GABA B receptors in presynaptic inhibition of Ia EPSPs in cat spinal motoneurons. J. Physiol. 447, 675 692. 47. Swandulla D. and Misgeld U. (1990) Development and properties of synaptic mechanisms in a network of rat hypothalamic neurons grown in culture. J. Neurophysiol. 64, 715-726. 48. Tell F. and Jean A. (1990) Rhythmic bursting patterns induced in neurons of the rat nucleus tractus solitarii, in vitro, in response to N-methyl-D-aspartate. Brain Res. 533, 152-156. 49. Tunstall M. J. and Roberts A. (1991) Longitudinal coordination of motor output during swimming in Xenopus embryos. Proc. R. Soc. Lond. B 244, 27-32. 50. Wall6n P., Christenson J., Brodin L., Hill R., Lansner A. and Grillner S. (1989) Mechanisms underlying the serotonergic modulation of the spinal circuitry for locomotion in lamprey. Proyr. Brain Res. 80, 321 327. 51. Wall6n P. and Grillner S. (1987) N-methyl-D-aspartate receptor-induced, inherent oscillatory activity in neurons active during fictive locomotion in the lamprey. J. Neurosci. 7, 2745-2755. 52. Yoshioka K., Sakuma M. and Otsuka M. (1990) Cutaneous nerve-evoked cholinergic inhibition of monosynaptic reflex in the neonatal rat spinal cord: involvement on M2 receptors and tachykininergic primary afferents. Neuroscience 38, 195 203. 53. Ziskind Conhaim L. (1990) NMDA receptors mediate poly- and monosynaptic potentials in motoneurons of rat embryos. J. Neurosci. 10, 125-135.