Generation of respiratory activity by the lamprey brain exposed to picrotoxin and strychnine, and weak synaptic inhibition in motoneurons

0306-4522/83 33.00 + 0.00

Neuroscience Vol. 10, No. 3, pp. 875-882, 1983 Printed in Great Britain

Pergamon Press Ltd 0 1983IBRO

ACTIVITY BY THE GENERATION OF RESPIRATORY LAMPREY BRAIN EXPOSED TO PICROTOXIN AND STRYCHNINE, AND WEAK SYNAPTIC INHIBITION IN MOTONEURONS C. M. ROVAINEN Department of Physiology and Biophysics, Washington University School of Medicine, 660 South Euclid Avenue, St. Louis, MO 63110, U.S.A. Ahstmct-The roles of Cl-dependent synaptic inhibition in the generation of fictive breathing were tested in isolated brains of adult lampreys, Ichthyomyzon unicuspis. Only a few inhibitory synaptic potentials were recorded in respiratory motoneurons between excitatory bursts. This was also true after Cl- injections inverted them to depolarizing potentials. A weak and variable phase of Cl-sensitive synaptic inhibition occurred at the ends of excitatory bursts. Respiratory motoneurons had a pronounced post-spike hyperpolarization, which was distinct from synaptic inhibition and appeared to be a more important mechanism for termination of firing. The production of the basic rhythm for respiration was tested in strychnine, picrotoxin, bicuculline and Cl-free fluid. Low concentrations of the blocking drugs prevented the inhibitory effects of bath-applied glycine and y-aminobutyric acid, but essentially normal respiratory bursts still occurred. Equilibration of isolated brains in high concentrations of strychnine and picrotoxin did not prevent periodic activities, but burst durations were increased and inter-burst intervals were longer and less regular than normal. Similar bursts could also occur transiently in Cl-free fluid. Recordings from the IX and X motor nuclei indicated that respiratory neurons produced the periodic bursts in the presence of strychnine and picrotoxin. Hemisections of the brain behind the V motor nuclei eliminated the bursts ipsilaterally. This indicated that descending excitation was necessary during pattern generation both in normal fluid and in the presence of antagonists of synaptic inhibition. Conventional synaptic inhibition does not appear to be essential for respiratory pattern generation in the adult lamprey but may contribute to its modulation. The hypothetical neural oscillator may consist of excitatory bursting intemeurons.

Synaptic inhibition can contribute to pattern generation in the nervous system in two ways. First, it can be part of the mechanism for oscillation. The contributions of inhibitory synaptic connections to the generation of basic rhythms have been well studied in the leech6~1*~25 and in the stomatogastric ganglion of the lobster.” Second, even if the basic rhythm is produced by some other mechanism, synaptic inhibition can be added to adjust the durations of phases, to introduce inhibitory phases and to provide reciprocal inhibition between antagonistic groups of neurons. The purpose of the present experiments has been to test the contributions of synaptic inhibition to the generation of respiratory activity in the isolated brain of the lamprey. In the adult lamprey, breathing is produced by synchronous excitation of expiratory motoneurons in the medulla with no known antagonist or antiphasic activity.‘3*20*2’ Antidromic stimulation,*O curarization and recordings in tetrodotoxin (C. Rovainen, unpublished observations) indicate that the motoneurons themselves are not the pattern generators. Inhibition has not been obvious during intracellular recordings from respiratory motoneurons in lampreys, except for a transient hyperGABA, y-aminobutyrate; potential.

Abbreviations:

post-synaptic

IPSP, inhibitory 875

polarization following the excitatory burst? One aim of the present experiments has been to test for the occurrence of inhibitory postsynaptic potentials (IPSPs) in motoneurons before and after Cl- injections. Similar injections of Cl- into motoneurons and interneurons in the lamprey spinal cord have shown clear bursts of IPSPs between the normal periods of excitation during fictive swimming.‘2T23Likewise, in the respiratory system of the cat, Cl- injections into intercostal motoneurons” and into medullary inspiratory and expiratory neurons” have revealed phases of active synaptic inhibition in addition to the phases of excitation. The contribution of synaptic inhibition to the generation of the basic rhythm for breathing in the lamprey has been tested in the present work by bathing the brain in Cl-free fluid or in solutions of drugs which block inhibitory receptors. Both spinal interneurons and Miiller cells in the midbrain and medulla of the lamprey are inhibited by glycine and y-aminobutyrate (GABA). These amino acids produce increases in Cl- conductance, and the effects are antagonized by strychnine or by picrotoxin and bicuculline, respectively. “,I6 Moreover, evoked IPSPs in Miiller cells are also sensitive to changes in Clconcentrations and are blocked by strychnine.i6 In the lamprey spinal cord the oscillator mechanism for the generation of fictive swimming appears to be

876

C M. Rovainen

sensitive to synaptic inhibition. Glycme decreases the frequency of fictive swimming, while low concentrations of strychnine increase frequency, and high strychnine concentrations or low Cl- fluid disrupt it.’

In contrast, the present respiratory activity can lamprey brain has been centrations of strychnine

experiments indicate that continue after the isolated equilibrated with high conand picrotoxin.

EXPERIMENTAL

PROCEDURES

The

present experiments were performed on isolated brain-skull preparations of adult silver lampreys, Ichthyunicuspis, 13-32cm, from Iowa. Animals were anesthetized by immersion in 1mM Tricaine, tissues were trimmed from the head, the brain was exposed dorsally, and the choroid plexus was removed to expose the ventricles. The ventral skull and lateral meninges were left to support the brain in a dish with Sylgard bottom. In some experiments the brain was dissected from the skull and cleaned of meningeal tissue to promote penetration of drugs. The preparations were immersed in l-5 ml physiological fluid (7-14”) of the following composition: 115 mM NaCl, 2 mM KCl, 2.6 mM CaCl,, 2 mM M.&l,, 3 mM NaHCO, and 6mM glucose. In some experiments the standard fluid was diluted half and half with l2OmM NaCI. Sodium isethionate and propionate salts of the minor cations were substituted in Cl-free fluid. Low Ca2+ fluid contained 1mM MgCI, and no added CaCI,. As in other preparations’ longer times were required to wash CaZ+ from the brain than to restore activities after Ca’+ was added back to the bathing fluid (see Results section). Breathing activities can continue in the isolated lamprey brain in solutions with 0.2-0.3mM Ca2+.” Drugs were obtained from Sigma, St Louis (except tetrodotoxin from Calbiochem), and were mixed from stock solutions directly into the bath. Glass micropipettes were filled with 4 M K acetate for routine intracellular recordings. For inversion of IPSPs, micropipettes were filled with 3 M KCl, and steady negative currents of several nA were passed into motoneurons until all spontaneous synaptic potentials were depolarizing.‘* Discharges of motor axons were recorded extracellularly by placing 50-150 ,um suction electrodes on Intracranial roots of the glossopharyngeus (IX) or vagus (X) cranial nerve. Respiratory motoneurons in the lamprey he in the IX and X motor nuclei under the ependyma of the fourth ventricle, and similar respiratory bursts could be recorded with the suction electrodes placed against the ventricular surface over these nuclei. Extracellular recordings were displayed AC and intracellular recordings DC. omyzon

RESULTS

Weak synaptic inhibition in motoneurons during fictive breathing Motoneurons were identified in the vagal (X) and glossopharyngeal (IX) motor nuclei by intracellular stimulation to produce one-for-one evoked spikes in an ipsilateral root of the X or IX nerve. During fictive breathing, motoneurons were driven synchronously by brief bursts of EPSPs to produce periodic spike bursts in the nerves.20,2’ Figure l(A) illustrates a spontaneous action potential, an excitatory burst and subsequent hyperpolarizations in a respiratory motoneuron. The mechanisms for the hyperpolarizations, post-spike KC condnetance ami synaptically mediated Cl- conductance were tested separately in the following two sections. During the

Fig. 1(A). Intracellular recording from a respiratory rnotoneuron in the X motor nucleus with a K acetate electrode. The first action potential occurred spontaneot&y and was followed by a post-spike. hypagoLriza$ior~ (PSH) (arrow). During the respiratory burst in the X nerve (upper trace) the motoneuron fired three times. The following hyperpolarization (dot) could have been produced by a combination of PSH and synaptic inhibition. Fig. l(B). Recording from a respiratory motoneuron after intracellular Cl- injection. The spontaneous inverted IPSPs occurred at an unusually high frequency in this cell compared to other motoneurons.

intervals between excitatory bursts, few spontaneous synaptic potentials were recorded in most of the motoneurones tested previously2o or in the present experiments. Eighteen respiratory c&s were injected with Cl- by iontophoresis from KC1 electrodes to produce inverted IPSPs. The highest resting frequency of IPSPs which was recorded is shown in Fig. 1B. More often, zero, one, or a few (Fig. 2B) inverted IPSPs were recorded during each cycle. Another way in which IPSPs could be observed was just after impalement with a K acetate electrode when the cells were depolarized by damage (Fig. 2A, r.p. about -40 to 45 mV). Presumably, the equilibrium potential for synaptic inhibition remained near the original resting potential (about - 65 mV) to produce a large driving force for inhibitory synaptic current. In 18 damaged motoneurons, the consistent observation was that few identifiable IPSPs were present in respiratory motoneurons during the intervats between bursts. A variable period of Cl-sen&ve .synaPGc inhibition appeared to occur at the end of the excitatory burst in some respiratory motoueurous. In damaged cells, one or a few II%% c&d W be distinguished at the end of the Elrsp, burst, These were recognized by abrupt deemaees in pore&&I to interrupt or termmate excitiacion (&g, ZC). A&&$h interruptions between ex&Wsry paatt+ynatptie potentials (EPSPs) during bursts mi& produce ‘t&r&r events, the putative IPSPs disappeared during recov-

817

Respiratory pattern generation

B-

recordings from the nucleus, 2 motor cells showed weak periodic EPSPs and 23 exhibited periodic IPSPs in phase with respiratory bursts. The IPSPs had the same average timing as in the IX and X motor cells, 25-40 ms after the onset of spikes in the respiratory nerves. The occurrence of the IPSPs alone in some trigeminal cells provided additional evidence for periodic respiratory inhibition. Post-spike hyperpolarization

Fig. 2. Unitary inhibitory postsynaptic potentials in “damaged” motoneurons. (A). Extracellular AC recording of periodic spike bursts in a root of the X nerve (upper trace), and intracellular DC recording from a motoneuron shortly after impalement with a micropipette filled with K acetate (lower trace). The EPSP bursts no longer elicited action potentials due to membrane depolarization and inactivation. Spontaneous negative IPSPs are labeled with dots. (B). Recording from a motoneuron which was injected with Cl- from a microoinette filled with KCl. Same brain as in A. (C). Recording from a motoneuron with a K acetate electrode at a faster sweep speed to illustrate possible IPSPs (dots) occurring after the excitatory phase.

ery of the cell, whereas the EPSPs became more

prominent, Likewise, the putative IPSPs were not present during all cycles and were not observed during the early part of the EPSP burst but had average occurrences 25-60 ms after the onset of the spike burst in the IX or X nerve. In contrast, the average onset of the EPSP bursts was lo-23 ms prior to the spike burst. In occasional damaged motoneurons the IPSPs were more prominent than EPSPs (Fig. 3Al). Hyperpolarization of one such cell changed the compound synaptic response to a depolarization continuing for longer than 100 ms (Fig. 3B3). A similar persistent depolarization was observed during hyperpolarization in other respiratory neurons, particularly after Cl- injections (Fig. 3B). The slow transient was unlikely to have been produced by passive electrical properties of the cell because it persisted after an action potential was elicited by intracellular stimulation (Fig. 3B3). The presence of an undershoot after the action potential (arrow) also indicated that the depolarization was not produced by an elevation of external K+ concentration. Thus, recordings from damaged motoneurons and from Cl-injected cells in the IX-X motor nuclei indicated that weak synaptic inhibition could occur near the ends of the excitatory respiratory bursts. Trigeminal motoneurons in the adult lamprey innervate the muscles of the sucker disc, pharynx and piston apparatus,9 none of which is involved in respiratory movements. Surprisingly, in a total of 92

The other mechanism which could help terminate firing in respiratory motoneurons is post-spike hyperpolarization (PSH). This was recorded in a variety of lamprey nerve cells, including respiratory motoneurons (Fig. 4), trigeminal motoneurons, and the large Miiller cells in the lamprey brain. Post-spike hyperpolarization was probably produced by a Cadependent K+ conductance, for instance, as described in frog motoneurons.’ Post-spike hyperpolarization could be distinguished from synaptic inhibition in several ways. Its apparent reversal potential was the same as for the rapid undershoot of the action potential and probably corresponded to the equilibrium potential for K+. The hyperpolarization and its reversal potential remained below resting potential after Cl- injections (Fig. 4B). Post-spike hyperpolarization was still present in nerve cells after conventional synaptic inhibition had been blocked with high concentrations of picrotoxin and strychnine (Fig. 4C). Finally, bathing the brain in lo-’ M tetrodotoxin eliminated the periodic bursts

3

A

3

ii

Fig. 3. Hyperpolarization and Cl- injection reveal phasic inhibition in respiratory neurons. (Al). Intracellular recordings of predominantly inhibitory responses in a respiratory motoneuron. (A2). Firing of the motoneuron after recovery. (A3). A combined excitatory-inhibitory potential during steady hyperpolarization of the motoneuron. All traces are aligned to the onset of spike bursts in the X nerve (top trace). (Bl). Intracellular recording of a compound depolarizing potential in a motoneuron injected with Cl-. The first phase is produced by EPSPs, and the second (dot) is probably an inverted IPSP. (B2). A prominent inverted phase (dot) in another respiratory neuron which was injected with Cl- and held steadily hyperpolarized. (B3). Action potentials and undershoots (arrow) were produced by repetitive current pulses in the same cell, but they did not eliminate the late, Cl-sensitive phase (dot).

878

C. M Rovamen

effect of 0.5 mM GABA. The only consistent change in activity produced by picrotoxin was an increased frequency of episodes of arousal, indicated by steady discharges in the spinal cord and the IX and X nerves (Fig. 5A). Similar episodes also occurred in untreated preparations. A stronger effect was produced by 3-5 PM strychnine in the bathing fluid. This concentration was sufficient to block the inhibitory effect of 0.6- 1.2 mM glycine. After 2-8 min of strychnine exposure, preparations exhibited seizure-like activities conjointly in the IX and X nerves and in the spinal cord. The discharges in the respiratory nerves consisted of high frequency bursts superimposed on steady activity, while discharges in the spinal cord were more continuous. Respiratory motoneurons showed a remarkable degree of synchronous firing on the two sides during seizures (Fig. SB). After episodes of seizurelike activity, respiratory bursts reappeared in 5 out of 6 preparations and were synchronous on the two sides, as in normal fluid. On the average the durations of bursts were longer in strychnine; this appeared to be due to a mixture of periods with bursts of normal Fig. 4. Post-spike and post-tetanic hyperpolarizations in respiratory neurons. (A). A train of 8 action potentials was elicited by direct intracellular stimulation with current pulses. The subsequent hyperpolarization (arrow) prevented firing from the periodic RPSP burst (E). (B). Intracellular Cl- injection inverted IPSPs (dots) in a X respiratory motoneuron, but repetitive intracellular stimulation still produced a hyperpolarixing after-potential (arrow). Same cell as in Fig. l(B). (C). Repetitive intracellular stimulation of a nerve cell in the X motor nucleus still produced post-spike hyperpolarixations after exposure of the isolated brain to 20 PM strychnine for 99 min and to 40 ~1M picrotoxin for 111 min. BPSPs (E) occurred in phase with periodic bursts in the ipsilateral (i) and contralateral (c) IX and X motor nuclei.

A 20 pkl picrotoxin

of EPSPs and IPSPs within 5-9 min and later blocked action potentials recorded in the cell body. Depolarizing current pulses in cells exposed to tetrodotoxin could still elicit after-hyperpolarizations similar to those following spikes in normal fluid. Efsects of glycine, y-aminobutyrate, and Cl-free fluid

antagonist drugs,

Addition of 0.2-0.5 mM glycine or y-aminobutyrate (GABA) to the bathing fluid had an inhibitory effect on respiratory activity. At the higher concentration, respiratory discharges disappeared in 5 preparations treated with glycine and 3 with GABA. Lower concentrations reduced the intensities and sometimes the durations of spike bursts in the IX and X nerves. These. effects were probably due in part to direct inhibition of motoneurons. In two out of four tests the frequency of bursts was also reduced, indicating that the pattern generator was also sensitive to the amino acids. Picrotoxin at 20 PM in the bathing &rid d@ not interfere with the periodic respiratory activity. This concentration was sufficient to block the inhibitory

Fig. 5. Effects of blockers of synaptic inhibition WI respiratory activity. (A) Normal respiratqy bursts (IV) stfB occurred in the X nerve after the b&n -had 14phi picrotoxin for 40 mitt and to 20 11 mm. A period of sustained activity also occurred for a few seconds in the X nerve and in spinal cord 2 mnt behind the ~obex. (B) bursts (N) still occurred in right and left nerves 11 min after 5 PM strychnine was bathing &rid. Seizure-like activities are present& ti last haif of the trace and are remark&Q sin&r in the eachsi&ofthebrain.(C)Synehrquabur%sof duration in the left and tight X nerves 9 mm aft& 4 the brain in Cl-free fluid.

Respiratory pattern generation 30-50 ms, and of longer ones, 100-200 ms. Strychnine had a long-term deleterious effect on respiratory activity, perhaps due to the fatigue of preparations during seizures. Respiratory bursts became intermittent or disappeared after lo-20 min. It was possible that the respiratory pattern generator utilized a combination of inhibitory mechanisms and that blocking of either GABA or glycine sensitivity still allowed some to remain. Addition of both 20 yM picrotoxin and 3 PM strychnine to two preparations had the same effects as just described for strychnine alone: the appearance of seizures, the progressive weakening of activity, and importantly the continuation of essentially normal periodic bursts in the respiratory nerves. Bicuculline at 7-10 PM in the bathing fluid blocked the inhibitory effects of both 0.5-l mM GABA and 0.8 mM glycine, evidently due to cross reaction between the two receptor types. Essentially normal respiratory bursts still occurred in one preparation treated with bicuculline alone and in one preparation with both bicuculline and strychnine. Another possibility was that some of the inhibitory synapses required longer diffusion times and higher concentrations of the drugs to be blocked. In fact, Matthews and Wickelgren16 reported that IPSPs in the midbrain Miiller cells required 20 PM strychnine to be eliminated and that the IPSP in the Miiller cell I, was not totally blocked until 30 min after 1 PM strychnine was added to the bathing fluid over the dorsal surface of the brain. Therefore, the present experiments were modified to promote equilibration of the drugs at higher concentrations. Brains of three lampreys (plus four others in the next section) were removed from the skull, and the meninges were dissected from the lateral and ventral surfaces. Each brain was suspended in the dish with both the ventricular and pial surfaces exposed to bathing fluid, and the roots of the IX and X nerves were drawn into suction electrodes for recording respiratory activities. In order to prevent excessive seizures, the brains were washed 2-5 times in low Ca*+ fluid. The frequency of respiratory bursts increased transiently, probably due to a lower threshold for generator activity,*’ but after 5-20min the bursts ceased and were replaced by steady spontaneous firing. Picrotoxin (30 PM) was added at 1I-30min and strychnine (20pM) at 40-60 min after the initial washing in Ca-free fluid. A few spontaneous seizures occurred after mixing and were probably due to some residual Ca*+ in the brain and the low threshold for activity. Finally, after 30 min of equilibration in strychnine, 1 mM Ca*+ was added to the bath to restore chemical transmission. Seizures began in 30-60 s, but after 2-4 min during the intervals between them, synchronous bursts occurred in the IX and X nerves (Fig. 6). These bursts were abnormal with respect to their long duration (140-500 ms) and their low and irregular frequency as compared to control respiratory bursts in normal fluid. Nevertheless, on a slower time scale the bursts

879

duration,

Fig. 6. Activities in the left and right X nerves after prolonged exposure to a combination of 20 p M strychnine and 30 PM picrotoxin. Bursts occurred at irregular intervals and were longer in duration (300ms) than normal (about 50 ms) but were still synchronous on the two sides. Seizurelike activities are present in both nerves near the end of the trace. See text for the sequence of changes in the bathing

fluid; these traces were recorded 75min after addition of picrotoxin and 43 min after addition of strychnine. in high concentrations of strychnine and picrotoxin were basically similar to those during normal respiration. In an intact lamprey such activity would be expected to produce intense periodic contractions of the branchial basket and movements of water in and out of the gill chambers, not unlike normal tidal ventilation. The alternative method for disrupting synaptic inhibition is removal of external Cl-. This was done in 6 preparations with the brain and meninges still attached to the ventral skull. The preparations were washed two or more times in Cl-free fluid, and after l-2min, episodes of increased activity occurred in the IX and X nerves and sometimes separately in the spinal cord. Respiratory bursts became longer in duration (200-400 ms) and less regular in frequency. Seizure-like activities commenced at 2-8 min, and ultimately breathing ceased. The record in Fig. 5(C), taken after 9min in Cl-free fluid, illustrates the activities which most nearly resembled normal respiration in the preparations tested. Note that the bursts were still coordinated bilaterally even though they were longer and less regular than in normal fluid. Did the periodic bursts occur in respiratory neurons?

In the preceding experiments it was unlikely that another population of vagal motoneurons became active and replaced the activity of respiratory motoneurons. The dominant efferent population in the X nerves of the lamprey is to the branchial muscles, and the vagal motor innervation of the gut and the heart is quite sparse” in comparison to higher vertebrates. Furthermore, the IX nerve, which has no abdominal projections, exhibited identical activities to those in the X nerve roots before and after equilibration with picrotoxin and strychnine. An alternative site for extracellular recording from respiratory neurons is the ependymal surface over the

880

C. M. Rovamen

Fig. 7. Periodic bursts in the IX and X motor nuclei and the effects of sections of the medulla after equilibration for over one hour in 20nM strychnine and 4OpM picrotoxin. (A) Recordings from the ventricular surfaces of cranial motor nuclei of the intact brain. Periodic bursts occur in IX and X motor nuclei but not in the V motor nucleus. Seizures occurred synchronously at all recording sites. (B) Hemi-

section of the medulla behind the left V motor nucleus eliminated periodic bursts but not seizures in the X motor nucleus on the same side. (C) Complete transection of the medulla eliminated all periodic bursts, but seizures still occurred independently in the rostra1 and caudal parts of the brain. Periodic respiratory bursts in normal fluid are affected in the same way by sections.

IX and X motor nuclei. Stimulation of the ipsilateral IX and X nerve roots elicits antidromic spikes, and periodic spike bursts are similar to those recorded in the roots. Four additional brains were cleaned of meninges and equilibrated in 40 pM picrotoxin plus 20 PM strychnine in low Ca*+ fluid, as in the preceding section. The brains were mounted dorsal side up, and electrodes were placed against the ventricular surfaces of the V, IX, and X motor nuclei. Addition of 1 mM Ca*+ to the bathing fluid produced synchronous seizures in all three motor nuclei and in the spinal cord when its activity was recorded. Additional short, periodic bursts occurred in the IX and X motor nuclei (Figs 4C, 7A), but not in the nearby V motor nucleus or in the spinal cord. Intracellular recordings were made from six nerve cells in the IX and X motor nuclei after equilibration in picrotoxin and strychnine. All of the nerve cells depolarized during seizures, and five of t&m aIao exhibited periodic depolarizations during t&e short periodic bursts recorded extracellularly (Fig. 4C). Thus, all three recording conditions for respiratory

bursts in normal fluid also showed periodic activity in picrotoxin and strychnine. This congruence indicates that respiratory motoneurons were active under both conditions. Components of the respiratory pattern generator also appeared to be necessary for periodic bursts in picrotoxin and strychnine. Hemisection of the medulla behind the V motor nucleus in the isolated brain of the adult lamprey eliminates bursts on that side (C. Rovainen, unpublished observations). The same hemisections of the four brains which were equilibrated in picrotoxin and strychnine likewise eliminated the short periodic bursts on that side only but did not prevent generalized seizures (Fig. 7B). Complete transection of the isolated lamprey brain behind the V motor nuclei in normal fluid eliminates respiratory bursts in the IX and X nerves on both sides. Two of the brains in picrotoxin and strychnine were completely transected, and the short periodic bursts disappeared on both sides. Seizures still occurred synchronously in the IX and X motor nuclei and independently in the V motor nucleus (Fig. 7C). Therefore, in both normal fluid and in the presence of blockers of synaptic inhibition, an ipsilateral descending excitatory pathway appears to be necessary for the expression of periodic bursts. DISCUSSION

One conclusion of this paper is that synaptic inhibition is not very important for the modulation of activity in motoneurons during fictive breathing activity in the brain of the adult lamprey. The IPSPs which were recorded between EPSP bursts in respiratory motoneurons were few and variable. This contrasts with the strong bursts of IPSPs in motoneurons in the lamprey spinal cord during fictive swimming.‘2,23 Similarly, during breathing activity in the cat, bursts of Cl-sensitive IPSPs occur between periods of excitation in intercostal motoneurones2’ and in medullary respiratory neuronsI Different phases of inhibition in the medulla are sensitive to strychnine and to bicuculline or picrotoxin4 Some synaptic inhibition occurred near the end of EPSP bursts in respiratory motoneurons in the lamprey and was sensitive to Cl- injection. Although this inhibition could help terminate firing, post-spike hyperpolarization appeared to exert a stronger effect (Fig. 4A). The second and major conclusion of this paper is that conventional synaptic inhibition was not required for the generation of the basic rhythms for breathing in the lamprey. In the vertebrate central nervous system in general, all fast synaptic inhibition appears to be mediated by receptors activated by either glycine or GABA, antagonized by either strychnine or picrotoxin and bicucuIline, and metiiated by increased Cl- conductanee5 In the &&prey, a simpler but typical vertebrate, all known FastIPSPs in the caudal brain stem’” and in the spinal cord2,‘2,21

Respiratory pattern generation

are sensitive to strychnine and to altered Cl-. If alternative mechanisms of inhibition such as increased K+ conductance are present, they are not sufhciently strong to prevent seizures in the lamprey brain stem and spinal cord in strychnine and Cl-free fluid. In particular, no negative IPSPs were observed in respiratory motoneurons after Cl- injections or after prolonged exposure to strychnine and picrotoxin. Because the pattern-generating interneurons for respiration in the lamprey have not yet been identified, it was not possible to test directly for synaptic inhibition in them, However, it is likely that synaptic inhibition in respiratory interneurons is similar to that in the larger nerve cells which have been tested in greater detail in the lamprey brain and spinal cord. The concentrations of strychnine and picrotoxin used in the present experiments were higher than those required to block the inhibitory effects of glycine and GABA on Miiller cells,16 spinal interneurons,‘O and fictive breathing. Likewise, the time for equilibration was longer than that required for the block of an IPSP by a low concentration of strychnine.16 The distances for diffusion are not great, since the walls of the lamprey medulla are approximately 0.5 mm thick. Anionic dyes easily penetrate the lamprey brain through the pial surface.** Removal of the meninges to expose the pial surface of the brain to the bathing fluid should have supplemented the entry of drugs through the ventricular surface. However, the present experiments do not rule out the possibility that the generator interneurons utilize unconventional synaptic inhibition, such as slow changes in K+ conductance produced by other putative transmitters. Overlapping populations of respiratory motoneurons and interneurons appeared to be active both in normal fluid and after equilibration with picrotoxin and strychnine. Although intracellular recordings were not made from the same identified respiratory motoneurons under both conditions, representative surveys of the IX and X motor nuclei showed the same periodic bursts as were recorded in the nerve roots in the presence of the synaptic blockers. These bursts were not observed in the V motor nuclei or in the spinal cord, neither of which is involved in fictive respiration in the adult lamprey. In the larval lamprey, the major respiratory muscles are in the velum. These contract periodically prior to branchial compressions and are innervated by motoneurons in the V motor nucleus.’ Hemisections of the larval medulla behind the V motor nucleus do not interfere with periodic velar activity but do eliminate periodic branchial bursts from the ipsilateral IX and X motor nuclei. This indicates that the oscillator for the respiratory rhythm is in or rostra1 to the trigeminal region, and that an ipsilateral descending pathway drives the branchial motoneurons.’ The same hemisections block ipsilateral branchial bursts in the isolated brain of the adult lamprey. This suggests that the trigeminal region still contains essential com-

881

ponents for pattern generation, even though V motor cells and muscles are no longer used for respiration after metamorphosis. The loss of ipsilateral periodic bursts in the IX and X motor nuclei after hemisection of the medulla in picrotoxin and strychnine is additional evidence that normal respiratory pathways are used to produce the rhythm. The two basic types of mechanisms which can produce rhythmic activity in a nervous system are intrinsic cellular pacemaker properties and emergent periodic firing patterns of neurons in synaptic circuits, particularly of interneurons with inhibitory connections. Both of these mechanisms are combined to produce patterned activities in two well-studied invertebrate preparations.“,” The pyloric system of the lobster stomatogastric ganglion includes neurons with endogenous bursting properties and also extensive reciprocal inhibitory synapses between cells. Either the network interactions or the endogenous bursting can produce periodic activities, but both are used together to produce the overall pattern.” The heartbeat of the leech is generated and coordinated by a network of inhibitory interneurons which also have endogenous bursting properties.3*‘8,25Low Clfluid severely reduces synaptic inhibition by these interneurons and eliminates both their modulation of heart motoneurons and their own reciprocal inhibitory interactions. Under these conditions, endogenous bursting is released from control by the network and produces periodic bursts in the interneurons at frequencies at least twice the normal rate.3 Although pacemaker properties are sufficient to produce periodic activity in this system, the network of inhibitory synapses among the intemeurons is essential for the overall pattern of activity and its transmission to motoneurons and the heart.3.‘8 The present experiments on the lamprey brain indicate that mechanisms other than conventional synaptic inhibition are sufficient to produce bursting activity. Under conditions of inhibitory synaptic blockade, the pattern generator for respiration in the adult lamprey may consist of groups of interneurons which are excitatorily coupled and also excite motoneurons during their peaks of activity. A mechanism which might terminate their excitatory phase could be a Ca-dependent K+ conductance, as appears to be present in other nerve cells in the lamprey. The recovery from the hyperpolarization produced by this conductance might also be combined with inward currents to produce slow rhythmic firing, as described in mammalian inferior olivary neuronsI and in the thalamus.14 (Also see discussion in ref. 13.) The present experiments do not exclude the contribution of synaptic inhibition to the development of the final motor program. Respiratory bursts in the present experiments became longer in duration and occurred at longer intervals in Cl-free fluid and in high concentrations of strychnine and picrotoxin. This suggests that the generator intemeurons for respiration in the adult lamprey may be modulated by

C. M. Rovainen

887

conventional synaptic inhibition. The modulation could be either tonic. with indirect effects on burst duration and interval, or phasic, with direct effects on the emerging pattern. It seems clear that the oscillatory interneurons and their cellular

themselves properties

need to be identified and synaptic inputs

determined in order to test these possibilities directly. Acknovledgemenls-This work was supported by USPHS grant NS 09367. Helpful comments on the manuscript were orovided bv Dr P. Stein. Dr P. Gettine. Dr A. Cohen and L. Rovainen.

REFERENCES

1. Barrett E. F. and Barrett J. N. (1976) Separation of two voltage-sensitive potassium currents, and the demonstration of a tetrodotoxin-resistant calcium current in frog motoneurones. J. Physiol., Land. 255, 737-774. 2. Buchanan J. T. (1982) Identification of interneurons with contralateral, caudal axons in the lamprey spinal cord: synaptic interactions and morphology. J. Neurophysiol. 47, 961-975. 3. Calabrese R. L. (1979) The roles of endogenous membrane properties and synaptic interaction in generating the heartbeat rhythm of the leech, Hirudo medicinalis. J. Expl Biol. 82, 163-176. 4. Champagnat J., Denavit-Saubie M., Moyanova S. and Rondouin G. (1982) Involvement of amino acids in periodic inhibitions of bulbar respiratory neurones. Brain Res. 237, 351-365. 5. Curtis D. R. and Johnston G. A. R. (1974) Amino acid transmitters in the mammalian central nervous system. Ergebn. Physiol. 69, 97- 188. 6. Friesen W. O., Poon M. and Stent G. S. (1978) Neuronal control of swimming in the medicinal leech--IV. Identification of a network of oscillatory interneurones. J. expl Biol. 75, 25-43. 7. Grillner S. and Wallin P. (1980) Does the central pattern generation for locomotion in the lamprey depend on giycine inhibition? Acra physiol. stand. 110, 103-105. 8. Homma S. (1975) Velar motoneurons of lamprey larvae. J. camp. Physiol. 104, 175-183. 9. Homma S. (1978) Organization of the trigeminal motor nucleus before and after metamorphosis in lampreys. Brain Res. 140, 33-42. 10. Homma S. and Rovainen C. M. (1978) Conductance increases produced by glycme and y-aminobutyric acid in lamprey intemeurones. J. Physiol., Lond. 279, 231-252. 11. Johnels A. G. (1956) Gn the peripheral autonomic nervous system of the trunk region of Lamperra planeri. Acta zool. Stockh. 37, 25 l-286. 12. Kahn J. A. (1982) Patterns of synaptic inhibition in motoneurons and interneurons during fictive swimming in the lamprey, as revealed by Cl- injections. J. camp. Physiol. 147, 189-194. 13. Kawasaki R. (1979) Breathing rhythm-generation in the adult lamprey, Entosphenus japonicus. Jap. J. Physiol. 29, 327-338. 14. Llinls R. and Jahnsen H. (1982) Electrophysiology of mammalian thalamic neurones in vitro. Nature, Land. 297,

406-408. 15. Llinas R. and Yarom Y. (1981) Properties and distribution of ionic conductances generating electraresponsiveness of mammalian inferior olivary neurones in vitro. J. Physiol., L.ond. 315, 569-584. 16. Matthews G. and Wickelgren W. 0. (1979) Glycine, GABA and synaptic inhibition of reticulospinal neurones of lamprey. J. Physiol., Lond. 293, 393-415. 17. Miller J. P. and Selverston A. I. (1982) Mechanisms underlying pattern generation in lobster stamatopstric ganglion as determined by selective inactivation of identified neurons-IV. Network properties of pyloric system. J. Neurophysiol. 48, 1416-1432. 18. Peterson E. L. and Calabrese R. L. (1982) Dynamic analysis of a rhythmic neural circuit in the leech Hirudo medicinalis. J. Neurophysiol. 47, 256-271. 19. Richter D. W., Camerer H., Meesmami M., Rijhrig N. and Bischoff A. (1979) Studies on the synaptic interconnection between bulbar respiratory neurones of cats. r’fugers Arch. ges. Physiol. 380, 245-251. 20. Rovainen C. M. (1974) Respiratory motoneurons in lampreys. J. camp. Physiol. !J4, 57-68. 21. Rovainen C. M. (1977) Neural control of ventilation in the lamprey. Fedn Proc. Fedn Am. Sots exp. Biol. 36,2386-2389. 22. Rovainen C. M., Lemcoe G. E. and Peterson A. (1971) Structure and chemistry of glucose-producing cells in meningeal tissue of the lamprey. Bruin Res. 30, 99-118. 23. Russell D. F. and Walltn P. (1980) On the pattern generator for fictive swimming in the lamprey, Ichthyomyzon unicuspu. Acta physiol. scand. 108, 9A.

23a. Russell, D. F. and Wallen, P. (1983) On the control of myotomal motoneurones during “fictive swimming” in the lamprey spinal cord in vitro. Acta physiol. stand. 117, 161-170. 24. Sears T. A. (1964) The slow potentials of thoracic respiratory motoneurones and their relation to breathing. J. Physiol., Land. 175, 404-424. 25. Stent G. S., Thompson W. J. and Calabrese R. L. (1979) Neural control of heartbeat in the leech and in some other invertebrates. Physiol. Rev. 59, 101 - 136.

(Accepted

10 March 1983)