Evidence that glycine and GABA mediate postsynaptic inhibition of bulbar respiratory neurons in the cat.

Evidence that glycine and GABA mediate postsynaptic inhibition of bulbar respiratory neurons in the cat AKIRA

HAJI,

RYUJI

TAKEDA,

AND

The Respiratory Research Group, Faculty Calgary, Albertu T2N 4N1, Canada

JOHN

E. REMMERS

of Medicine, University

of Calgary

Health Science Centre,

nally, tonic postsynaptic inhibition, i.e., inhibition extending throughout the respiratory cycle, has been documented in inspiratory neurons (7,20). The sites of origin 2333-2342, 1992.-Experiments were carried out on decere- of these inhibitory inputs are known in some cases; spebrate cats to identify transsynaptic mediators of spontaneous cifically, expiratory neurons of the retrofacial area appostsynaptic inhibition of bulbar inspiratory and postinspirainhibit augmenting inspiratory tory neurons. Somatic membrane potentials were recorded pear to postsynaptically neurons of the solitary tract and para-ambigual nuclei (5, through the central micropipette of a coaxial multibarreled HAJI, AKIRA, RYUJI TAKEDA, AND JOHN E. REMMERS. Evidence that glycine and GABA mediate postsynaptic inhibition of bulbar respiratory neurons in the cat. J. Appl. Physiol. 73(6):

Blockers of type A y-aminobutyric acid (GABA-A) and glycine receptors were iontophoresed extracellularly from peripheral micropipettes surrounding the central pipette. Effective antagonismwas demonstratedby iontophoresis of agonists with antagonists; application of strychnine antagonized the action of glycine but not GABA, and application of bicuculline antagonized the action of GABA but not glycine. In both types of neurons, iontophoresis of either antagonist depolarized the somatic membrane and increased input resistance throughout the respiratory cycle. Bicuculline preferentially depolarized the somatic membranein both types of neuronsduring inactive phases.Strychnine increasedthe firing rate of inspiratory neurons during inspiration despite maintenance of somaticmembranepotential at preiontophoresislevels. Tetrodotoxin reduced the effects of iontophoresed bicuculline and strychnine, suggesting that the action of the antagonists required presynaptic axonal conduction. The present results suggest that presynaptic releaseof both GABA and glycine contributes to tonic postsynaptic inhibition of bulbar respiratory neurons. GABA-A receptors appear to contribute to inhibition during inactive phasesin inspiratory and postinspiratory neurons, whereasglycinergic mechanismsappear to contribute to inspiratory inhibition in inspiratory neurons. electrode.

control of breathing; bicuculline; strychnine FLUCTUATION OF MEMBRANE POTENTIAL in synchrony

with the respiratory motor output is a hallmark of bulbar respiratory neurons in mammals (18-20) and reptiles (24). This periodic undulation results, in part, from rhythmic variation in presynaptic neuronal activity, which induces periodic waves of inhibitory and excitatory postsynaptic potentials (1, 18, 19). Such postsynaptic inhibition occurs at precise times in the respiratory cycle. For example, in most respiratory neurons, a prominent barrage of inhibitory postsynaptic potentials (IPSPs) occurs during the nonspiking or inactive phase of the respiratory cycle, suggesting that postsynaptic inhibition plays a role in gating the spike activity of respiratory neurons (1, 2, 17). In addition, IPSPs Occur during the active or spiking periods in some respiratory neurons, and this process is thought to control the rate of firing of action potentials during active phases (1). Fi-

6, 13).

The neurotransmitters

mediating

postsynaptic

inhibi-

tion in respiratory neurons have not been unequivocally

identified. y-Aminobutyric acid (GABA) and glycine are obvious candidates, inasmuch as these amino acids appear to mediate postsynaptic inhibition in the spinal cord, cortex, cerebellum, and brain stem (4, 10, 11). Haji et al. (8) recently demonstrated that iontophoresed GABA and glycine decrease input resistance and increase the chloride conductance of bulbar respiratory neurons, evidence consistent with these amino acids being transsynaptic mediators of postsynaptic inhibition in the bulbar respiratory neuronal network. Champagnat et al. (3) examined the response of inspiratory neuron spiking to iontophoresis of the type A GABA (GABA-A) receptor antagonist bicuculline and the glycine receptor antagonist strychnine. They observed that bicuculline caused spiking during late expiration, whereas strychnine induced action potentials during early expiration. Because this study employed extracellular recording, the nature of the postsynaptic response to the antagonist is uncertain. To clarify the roles of GABA-A and glycine receptors in mediating postsynaptic inhibition in the respiratory neuronal network, we have examined the response of membrane potential and input resistance of bulbar respiratory neurons to iontophoretic application of bicuculline and strychnine close to the impaled neuron. Tetrodotoxin was iontophoresed to block pre- and postsynaptic events dependent on “fast” sodium channels (12). Because the postsynaptic actions of endogenously released GABA and glycine do not involve postsynaptic sodium channels, the action of tetrodotoxin on the effects of strychnine and bicuculline reveals the role of presynaptic action potentials in mediating release of inhibitory amino acids. METHODS

General Surgery

Experiments were carried out on 36 adult cats of either gender, weighing 2.7-4.3 kg. The animals were anesthetized with halothane (L&-2.2%). Catheters were inserted

0161-7567/92 $2.00 Cupyright 0 1992 the AmericanPhysiological Society

2333

Downloaded from www.physiology.org/journal/jappl by ${individualUser.givenNames} ${individualUser.surname} (132.210.236.020) on January 15, 2019.

2334

GA3A

AND GLYCINE

TABLE 1. Numbers of nonantidrumically activated neurons, vugul motoneurons, and bulbospinul neurons Neuron

Type

Inspiratory

Postinspiratory

ES

14 3 4

16 4 0

Total

21

20

NAA Vagal

NAA, nonantidromically

activated; BS, bulbospinal.

into the trachea, femoral artery, femoral vein, and urethra. The external carotid arteries were ligated, and the head was mounted in a stereotaxic frame. A temporal craniotomy was performed, and the brain was transected at the midcollicular level. After the brain was aspirated rostra1 to the transection and hemostasis was secured, the concentration of halothane was reduced (0.5~1.0%). Muscular paralysis was instituted by an intravenous loading dose of gallamine triethiodide (5 mglkg), followed by a continuous intravenous infusion (1-2 mg +kg-l. h-l). The lungs were mechanically ventilated with O,-enriched air. End-tidal CO, concentration was monitored and held between 0.04 and 0.05 by adjusting pulmonary ventilation. Systemic blood pressure was continuously monitored, and the mean value was main-

IPSPs

tained at >80 mmHg by intravenous infusion of a lactate-Ringer solution. Rectal temperature was held at 373VC by a thermostatically controlled heating pad. The cervical vagus nerves were sectioned bilaterally, and the central cut end of each was mounted on a pair of silver electrodes for electrical stimulation. A C&-C!, laminectomy was performed, and a linear array of five coaxial stimulating electrodes was positioned transversely in the ventrolateral spinal cord. The most rostra1 root of the phrenic nerve was exposed bilaterally and cut distally. The central cut end of each was desheathed and placed on a bipolar recording electrode. The dorsal and lateral aspects of the medulla oblongata were exposed through an occipital craniotomy. Movement of the brain stem associated with ventilation was minimized by a bilateral pneumothorax. To maintain end-expiratory lung volume above residual volume, end-expiratory tracheal pressure was maintained at 2 cmH,O by an expiratory threshold load. After this surgery was completed, halothane was discontinued, and 23 h elapsed before respiratory neuronal activities were recorded. Recording Procedures

The efferent discharge of the phrenic nerve was ampliband pass), and averaged

fied, filtered (lo- to 3,000-Hz

A

MP mv -60

FIG. 1. Effects of glycine iontophoresis (horizontal bar) on membrane potential (MP) of an inspiratory neuron (nonantidromically activated) before (A) and during (3) iontophoresis of strychnine. PN, efferent activity of phrenic nerve. A: iontophoresis of glycine (100 nA) caused hyperpolarization of membrane and blocked action potentials during inspiration. B: effect of glycine was antagonized by iontophoresis - of strychnine (40 nA, started 20 s before glycine ejection). This dose of strychnine had a negligible effect on membrane potential trajectory.

PN ~mvll-ulyb----*ryI

B

mV

-60

2sec


GABA

AND

GLYCINE

IPSPs

2335

mV -60

\

\

a

l-Jv[ m

-1,

B -40 mV

\ \

-60

.5mV[

m

FIG. 2. Effects of iontophoresis of GABA (horizontal bar) before (A) and during (B) iontophoresis of bicuculline onto a postinspiratory vagal motoneuron. A: iontophoresis of GABA (100 nA) produced striking hyperpolarization during stages I and II of expiration. B: tracing taken 45 s after onset of bicuculline iontophoresis (100 nA) showing a decrease in membrane potential fluctuations and synaptic noise. Iontophoresis of GABA had no effect. Bottom traces in A and B represent higher-speed sweeps of segments a and b of top

-

2 SEC

using a Paynter filter with a 100-ms rise time (BAK PF1). Neurons were impaled by a compound coaxial microelectrode consisting of a central recording micropipette encircled by six iontophoresis micropipettes with tips recessed 20-40 pm from the tip of the recording pipette (22). The latter was filled with 2 M potassium citrate and connected to a high-impedance electrometer (Dagan 8100-l) by a silver-silver chloride electrode. The directcurrent resistance of the recording micropipette ranged from IO to 40 MQ. The compound microelectrode was positioned by a micromanipulator driven by a stepping motor (SPI Nano-Stepper, type B). Respiratory neurons included in the present study were located 3.0-4.5 mm lateral to the midline, O-3.0 mm rostra1 to the obex, and 2.7-4.7 mm below the dorsal surface. These neurons were classified by the temporal relationship of membrane potential trajectory to the phrenic discharge, as described by Richter (18). Two types of respiratory neurons, inspiratory and postinspiratory, were encountered. The former displayed progressive membrane depolarization during inspiration and hyperpolarization during expiration; the latter exhibited hyperpolarization during inspiration followed by rapid depolariza-

tion coincident with the abrupt downstroke in the phrenic neurogram. According to their response to electrical stimulation of the ipsilateral vagus nerve (3 V, 0.1 ms) and the cervical spinal cord (10 V, 0.2 ms), neurons were classified as vagal motoneurons, bulbospinal neurons, or nonantidromically activated neurons (18). Table 1 provides the numbers of inspiratory and postinspiratory neurons in each category. Approximately one-third of inspiratory neurons were demonstrated to have vagal or spinal cord axons. The fraction of total inspiratory neurons classified as nonantidromically activated (66%) is somewhat larger than previously reported (1, 20). These differences may be related to spinal cord electrode position or variation in sampling of the rather small number of inspiratory neurons in the present study. Input resistance of each neuron was measured by applying constant-current pulses intracellularly (-0.5 to -2.5 nA, loo-250 ms) through the recording pipette by use of a balanced bridge method. The pulses were delivered at 2 Hz by a pulse generator triggered by the onset or the rapid downstroke of the phrenic burst. The interval between peak phrenic discharge and the onset of the next inspiration was divided into stage I and


GABA AND GLYCINE

2336

lmV[

4sec 3. Effects of saline and bicuculline iontophoresis on membrane potential of an inspiratory neuron (nonantidromically activated). A: membrane potential (top trace) and phrenic neurogram (bottom trace) recorded before iontophoresis. B: similar recordings made during a 30-s control test when lOO-nA iontophoretic current was applied to saline barrel. C: similar recordings obtained 20 s after beginning of a bicuculline iontophoresis test using 100 nA. Saline iontophoresis produced no change in membrane potential tracing, and bicuculline iontophoresis depolarized membrane and induced action potentials during stage II of expiration. FIG.

stage II of expiration, as described previously (17, 19). The trajectory of membrane potential of postinspiratory neurons displayed depolarization during stage I and repolarization during stage II. This distinction was apparent in some inspiratory neurons as hyperpolarization during stage I followed by a plateau in stage II. Data Analysis

The phrenic neurogram and membrane potential were recorded on a polygraph and on magnetic tape for subsequent computer analysis. Respiratory cycle averages of membrane potential, input resistance, and action potentials were obtained using a computer (IBM-AT) and software (Computer Scope, RC Electronics) triggered by a pulse coincident with the onset or rapid decline of the phrenic burst. To systematically quantify the responses of the neuronal population recorded, membrane potential was measured at two points: the most depolarized point (Dmax) and the most hyperpolarized point (Hmax) of the respiratory cycle. Mean values of these variables observed before and during iontophoresis were compared using a paired two-tailed t test. An analysis of variance for repeated measures was used to evaluate differences in effects of each antagonist or differences in the responses of inspiratory and postinspiratory neurons. This analysis yielded an F value, which was used to indicate significant differences between means. Iontophoresis Procedures

The peripheral barrels of the compound microelectrode were filled with the following solutions: physiological saline (165 mM NaCl), glycine hydrochloride (1 M in

IPSPs

distilled water, pH 4.0), GABA (1 M in distilled water, pH 4.5), strychnine sulfate (5 mM in saline solution, pH 7.5), bicuculline methiodide (5 mM in saline solution, pH 3.0), and tetrodotoxin (0.5 mM in saline solution, pH 6.5). The resistances of the iontophoresis pipettes ranged from 2 to 50 MQ. During iontophoresis trials, cationic iontophoretie currents (20400 nA) were applied to these solutions through silver-silver chloride electrodes by use of a programmable iontophoresis current source (WPI S7100A). Anionic retaining currents (-2 to -5 nA) were applied to all solutions between trials. The saline pipette was used as a current sink during drug iontophoresis trials and served as an iontophoretic source during control tests. Control tests were performed by application of 100 nA of iontophoretic current to the saline barrel before drug iontophoresis. Such control tests produced changes in membrane potential 0.05, Student’s t test).

Tonic Changes Caused by the Antagonists Effects on Hmux and Dmax. A total of 50 iontophoresis trials (30 with bicuculline and 20 with strychnine) were performed during stable recordings of membrane potential in 41 neurons: 21 inspiratory and 20 postinspiratory neurons (Table 1). Control iontophoretic currents applied to the saline micropipette caused no change in mean value or respiratory fluctuation of membrane potential, as illustrated in Fig. 3, A and B. Mean values of Dmax and Hmax, recorded before and during iontophoresis of antagonist, are provided in Table 2. The value of both variables increased significantly (i.e., the membrane depolarized) during iontophoresis of bicuculline or strychnine onto either type of neuron. Furthermore, mean Hmax and Dmax increased comparably during iontophoresis, indicating that the membrane was depolarized throughout the respiratory cycle (Figs. 2-5 and 7). Such tonic depolarization throughout the respiratory cycle was the common, but not exclusive, pattern of response; an increase in Hmax, i.e., hyperpolarization of the most hyperpolarized segment of the cycle (range -1 to -5 mV), was observed in 9 (6 postinspiratory neurons and 3 inspiratory neurons) of 50 trials. By contrast, Dmax always shifted to more depolarized values. As shown in Table 2, the changes in Dmax and Hmax caused by bicuculline iontophoresis significantly exceeded those caused by strychnine. However, changes in these variables did not differ significantly between neuron type. Effects on input resistance. Iontophoresis of bicuculline or strychnine increased input resistance in both types of neurons. Figure 4B illustrates such an action caused by bicuculline in a postinspiratory neuron; the change in membrane potential, caused by negative current pulses (-1 nA, 150 ms) delivered to the intracellular electrode, increased during bicucufline iontophoresis. Similar changes are demonstrated in the cycle-triggered average of membrane potential of an inspiratory neuron (Fig. 5C). Values for input resistance during the preiontophoresis period (left) averaged 4.2 MQ for expiration and 2.0 MQ for inspiration. Bicuculline iontophoresis (right) produced a twofold increase in input resistance; average

IPSPs

values of 8.6 and 4.0 Ma were observed during expiration and inspiration, respectively. Group mean values for input resistance are presented in Table 3. Input resistance, measured during iontophoresis, significantly exceeded preiontophoresis values for both antagonists with both types of neurons. Phasic Changes Caused by the Antagonists Effects on the shape of the membrane potential trajectory. To evaluate the effects of the antagonists on the shape of the respiratory wave in membrane potential, cycle-triggered averages of membrane potential were recorded before and during iontophoresis of the antagonist. Such comparisons were achieved in 10 inspiratory and 13 postinspiratory neurons with use of both bicuculline and strychnine. Figure 5B shows cycle-triggered averages of membrane potential of an inspiratory neuron recorded before (left) and during (right) bicuculline iontophoresis. Comparison of the two averages reveals that

A -66 . mV

-75

I

I

I

I

I

I

I

I

I

I

I

I

I

J

B -78

mV -84 f

1

I

1

I

I

I

I

I

Scale : 400 msec FIG. 6, Effects of bicuculline on average membrane potential trajectory of a postinspiratory neuron (nonantidromically activated) before (A) and after (B) iontophoresis of tetrodotoxin (50 nA, 1 min). Each tracing represents cycle-triggered average of membrane potential for 5 consecutive respiratory cycles. A: tracings obtained before (Control) and during (BIG) bicuculline iontophoresis. During iontophoresis of bicuculline (50 nA), membrane potential was adjusted by manual current clamping at the most depolarized point of respiratory cycle (Dmax) to preejection level. In B, tracings obtained before and during iontophoresis of bicuculline are superimposed.


GABA AND GLYCINE STRYCtiN1NE

A 1 mv

1 OONA

CURRENT

1 MIN

2339

IPSPs CLAMP

’

[ L

CYCLE

TRlGGERED

1

5 SEC

HISTOGRAM

FIG.

II

11

11

‘ONSET

OF

11

PHRENIC

7. Effects of strychnine

on firing of action

potentials in an inspiratory neuron (nonantidromi1 ’ 1 tally activated). Top: tracings obtained before (A) and after 1 (B) and 1.5 min (C) of iontophoresis of strychnine (100 nA). A-C: higher-speed sweeps of inspiratory activity from segments presented in tup; cycle-triggered histograms of action potentials were obtained in 3 consecutive respiratory cycles before and during each period of iontophoresis (histogram scale: spikes/s). A: before iontophoresis; B: 1 min after onset of iontophoresis; C: 1.5 min after onset of iontophoresis where somatic membrane potential has been returned to control levels by intracellular injection of negative current (current clamp).

DISCHARGE

Y-7 20 mV [ L

1 mV [

II

I +

I1

II

I

I]

SCALE:25Omsec

I

L 1 SEC

the GABA antagonist altered the shape of the membrane potential trajectory such that the membrane depolarized more rapidly during the first portion of expiration. This more rapid early-expiratory depolarization was observed in all 10 inspiratory neurons in which cycle-triggered averages of membrane potential were compared before and during bicuculline iontophoresis. Strychnine iontophoresis had no effect on the shape of the membrane potential trajectory. The overall depolarization caused by bicuculline cannot explain the preferential depolarization during expiration; if the antagonist had not influenced postsynaptic inhibition during expiration, the hyperpolarizing effects of expiratory IPSPs would have been augmented compared with control, because the average value of membrane potential was less negative during iontophoresis. Comparable shifts in the trajectory of membrane potential were noted during inspiration and stage II of expiration for postinspiratory neurons, as illustrated in Fig. 6A, where cycle-triggered averages of membrane poten-

tial recorded before and during bicuculline iontophoresis are superimposed. In this example, the bicuculline-induced depolarization was reversed by intracellular current injection, so that Dmax was the same before and during iontophoresis. In addition to reducing membrane hyperpolarization during inspiration, bicuculline retarded repolarization of the membrane during stage II of expiration, suggesting that the drug blocked IPSPs during both inactive phases of postinspiratory neurons. Figure 4A shows that bicuculline reduces the magnitude of membrane hyperpolarization of a postinspiratory neuron during inspiration. Before iontophoresis (Fig. 4A, left), a 5-mV hyperpolarizing shift in membrane potential occurred at the transition from stage II of expiration to inspiration, whereas during iontophoresis (Fig. 4A, right) this hyperpolarizing step was virtually eliminated. Of 13 postinspiratory neurons examined with bicuculline iontophoresis, 9 displayed a reduction in the inspiratory hyperpolarizing step similar to that shown in Fig. 4A and the other 4 displayed the depolarizing shift in membrane


2340

GABA AND GLYCINE

IPSPs

3

A 0 MP W

**

-70

PN

0’

C

mV

-70

FIG. 8. Effects of iontophoresed tetrodotoxin on membrane and action potentials of an inspiratory vagal motoneuron. A-C were obtained from the same neuron. A: tracings were taken before (1) and 10 s (2) and 10 min after (3) beginning of iontophoresis of tetrodotoxin (50 nA). 23:records of antidromic action potential induced by ipsilateral vagal nerve stimulation (1.5 V, 0.2 ms) and excitatory postsynaptic potentials (EPSPs) evoked by spinal cord stimulation (SCst) before (1) and 10 min after (2) iontophoresis of tetrodotoxin. C: cycle-triggered averages of membrane potential for 5 consecutive respiratory cycles taken before (1) and 10 (2) and 25 min after (3) beginning of iontophoresis of tetrodotoxin. Note progressive reduction of membrane potential fluctuations and hyperpolarizing shift of membrane potential caused by tetrodotoxin iontophoresis.

L

2 0 mV

-70

potential during expiration resembling that shown in Fig. 6A. By contrast, no such preferential depolarization of the membrane during inactive phases was observed when strychnine was iontophoresed onto 13 postinspiratory neurons. Effects on spike activity. Both antagonists increased the rate of firing of action potentials during active phases in both types of neurons. In addition, bicuculline, but not strychnine, initiated repetitive action potentials during inactive phases. The bicuculline-induced augmentation in spike frequency during the active and inactive phases was eliminated when the somatic membrane potential was returned to preiontophoresis levels by intracellular injection of negative current. Accordingly, this spikeaugmenting action of bicuculline appeared to result from the generalized membrane depolarization caused by the antagonist rather than a phase-related blockade of IPSPS. Strychnine iontophoresis, by contrast, augmented spike frequency of inspiratory neurons during inspiration, despite maintenance of somatic membrane potential at preiontophoresis levels by negative intracellular current. Figure 7 provides a detailed analysis of such an effect of strychnine on the discharge activity of an inspiratory neuron. The tracings of membrane potential (top) were obtained before and during strychnine iontophoresis. Figure 7, A-C, presents expanded views of membrane potential tracings (left) together with corresponding cy-

cle-t8riggered histograms of action potentials (right). Before iontophoresis (A), the membrane potential and firing pattern displayed an augmenting pattern, but membrane depolarization and firing frequency were attenuated shortly before the end of the phrenic burst. During iontophoresis (B), the membrane depolarized 5 mV, and spike frequency in the inspiratory phase was increased greatly. That depolarization of the somatic membrane did not contribute to this strychnine-induced augmented firing is demonstrated by the similarity of the histograms shown in B and C, the latter having been recorded when somatic membrane potential was returned to the preiontophoresis value by intracellular injection of negative current. Similar results were not obtained with postinspiratory neurons; strychnine-induced increase of action potential frequency during the active phase (stage I of expiration) was reversed by returning the somatic membrane potential to preiontophoresis levels. Effects of tetrodotoxin. Tetrodotoxin was applied to block pre- and postsynaptic processes dependent on fast sodium channels (21). As shown in Fig. 8A (1 and 2), tetrodotoxin consistently suppressed spontaneous action potentials and depolarized the membrane during the initial lo-30 s of iontophoresis. Thereafter, membrane potential returned to preiontophoresis values (Fig. 8A, 3). In 9 of 14 cells, the respiratory fluctuations of membrane potential were reduced by 250% after tetrodotoxin application, as illustrated by the cycle-triggered averages


GABA AND GLYCINE

of membrane potential shown in Figs. 6B and 8. Tetrodotoxin blocked antidromic spike generation by vagal simulation and greatly reduced the excitatory postsynaptic potentials evoked by spinal cord stimulation (Fig. 8B). The cycle-triggered averages of membrane potential shown in Fig. 6 demonstrate that tetrodotoxin blocked the effect of iontophoresed bicuculline on a postinspiratory neuron Before iontophoresis of tetrodotoxin (A), bicuculline depolarized the neuron during inspiration and stage II of expiration, whereas after tetrodotoxin (B), bicuculline iontophoresis exerted little effect on the respiratory waveform of membrane potential. Such elimination of the action of the iontophoresed antagonist by tetrodotoxin was observed in all neurons tested, i.e., in five neurons using bicuculline and in four neurons using strychnine. DISCWSSION

The present results demonstrate that antagonists of inhibitory amino acids, iontophoresed close to the soma of inspiratory or postinspiratory neurons, exert tonic and phasic actions consistent with blockade of postsynaptic inhibition. Iontophoresis of each amino acid antagonist depolarized the somatic membrane and increased input resistance in all phases of the respiratory cycle. This indicates that endogenous release of transmitter acts on GABA-A and glycine receptors to tonically maintain membrane potential at a relatively hyperpolarized value. Richter et al. (20) showed that injection of chloride into bulbar respiratory neurons resulted in a depolarizing shift of membrane potential throughout the respiratory cycle. They suggested that this was due to tonic inhibitory inputs, possibly from pontine structures. Tonic inhibition may control the rate and duration of spike activity, as recently pointed out by Feldman and Smith (7). Blockade of GABA-ergic tonic inhibition may explain the prolongation of the inspiratory burst observed by these workers in the in vitro neonatal rat brain preparation when bicuculline was added to the superfusate. Iontophoresed bicuculline exerted phase-specific actions, as manifested by preferential depolarization of the somatic membrane of inspiratory and postinspiratory neurons during inactive phases. In agreement with the results of Champagnat et al. (3), we observed that bicuculline induced action potentials during stage II of expiration in inspiratory neurons. Similarly, the drug induced action potentials during inspiration and stage II of expiration in postinspiratory neurons. A recent report (23) showed that iontophoresis of flurazepam, a specific enhancer of GABA-mediated chloride conductance (16), potentiates the periodic IPSP waves during inactive phases of postinspiratory and inspiratory neurons. This action is blocked by coiontophoresis of bicuculline. Together, these results suggest that phasic release of GABA acting on GABA-A receptors contributes to inactive phase inhibition in both types of neurons. A possible correlative finding is that GABA-ergic neurons of the retrofacial nucleus project caudally to respiration-related bulbar areas (12). Strychnine increased the frequency of action poten-

IPSPs

2341

tials during the active phase of inspiratory and postinspiratory neurons. In the latter case, this action was reversed when the somatic membrane potential was returned to control levels. Therefore, this effect of strychnine appears to be the consequence of generalized somatic membrane depolarization caused by the antagonist. By contrast, in inspiratory neurons, the strychnineinduced augmented firing rate persisted despite return of somatic membrane potential to control levels. This result suggests that glycinergic mechanisms contribute to active phase postsynaptic inhibition of inspiratory, but not postinspiratory, neurons. This apparent disparity between the two types of neurons corresponds to differences in identified postsynaptic inhibition during the active phase in each type of neuron; active phase postsynaptic inhibition has been demonstrated by chloride reversal in inspiratory (l), but not postinspiratory, neurons (17, 18). These results are at variance with the results of Champagnat et al. (3), who observed that strychnine iontophoresis on inspiratory neurons elicited action potentials during stage I of expiration. Such negative evidence is not convincing. For instance, our negative results may reflect failure of the drug to reach synapses located remotely on the dendritic tree. Another explanation may be that strychnine iontophoresis might have caused more pronounced depolarization in the study of Champagnat et al. than in our study. The more depolarized inspiratory neuron might be expected to discharge action potentials during stage I of expiration. The results obtained with coiontophoresis of agonist and antagonist provide evidence regarding specificity of the agonist, because strychnine blocked the action of giytine but not GABA and bicuculline blocked the action of GABA but not glycine. Bicuculline appears to be a rather specific antagonist of GABA-A receptors (ll), but at concentrations >lO PM the drug depolarizes the membrane and increases input resistance by nonsynaptic actions (9). Strychnine, on the other hand, has been shown to attenuate the actions of @alanine and taurine when iontophoresed with large currents (4, lo), so that the observed actions of iontophoresed strychnine might have been due to antagonisms of either of these potential transmitters. Because our results were obtained with low iontophoretic currents, the effects can probably be attributed to the specific action of the antagonist, i.e., antagonism of GABA by bicuculline and antagonism of glycine by strychnine. The question of site of action of the iontophoresed antagonist, i.e., whether on the impaled neuron or at a presynaptic site, cannot be answered definitively from the present results. Because the iontophoretic source was positioned close to the neuronal soma, the diffusion distances to somatic synapses on the impaled neuron were, presumably, relatively short (20-40 pm). By contrast, distances to remote denditric synapses on the neuron were undoubtedly larger. The persistence of respiratory fluctuation in membrane potential after iontophoresis of tetrodotoxin suggests that presynaptic axons remote from the recording site were not blocked by the drug because of long diffusion pathways. The observed reduction of antagonist effect by tetrodotoxin suggests that


2342

GABA AND GLYCINE

presynaptic axonal conduction in the vicinity of the impaled neuron is required for antagonist action. This conclusion assumes that the primary effects of GABA and glycine are not dependent on the sodium channels of the impaled postsynaptic neuron. These considerations suggest that iontophoresed GABA and glycine acted on postsynaptic receptors of the impaled neuron, or of neighboring neurons, to block the actions of presynaptically released amino acid. In summary, the results demonstrate tonic postsynaptic inhibition of bulbar inspiratory and postinspiratory neurons of the cat owing to transmitter release activation of GABA-A and glycine receptors. Phasic activation of GABA-A receptors inhibits both types of neurons during their inactive phases. Glycinergic mechanisms play a role in controlling spike frequency of inspiratory neurons during inspiration. The authors thank Marguerite Schultz and Ashley Micheals for expert assistance in the preparation and typing of the manuscript. This research was supported by grants from the Alberta Heritage Foundation for Medical Research and the Medical Research Council of Canada. Address for reprint requests: J. E. Remmers, Dept. of Medicine, Faculty of Medicine, University of Calgary, 3330 Hospital Dr. NW, Calgary, Alberta T2N 4Nl, Canada. Received 19 December 1991; accepted in final form 12 June 1992. REFERENCES 1. BALLANTYNE, D., AND D. W. RICHTER. Postsynaptic inhibition of bulbar inspiratory neurones. J. Physiol. Lo&. 348: 67-87, 1984. 2. BALLANTYNE, D., AND D. W. RICHTER. The non-uniform character of expiratory synaptic activity in expiratory bulbospinal neurones of the cat. J. Physiol. Land. 370: 433-456, 1986. 3. CHAMPAGNAT, J., M. DENAVIT-SAIJBIE, S. MOYANOVA, AND G. R~ND~UIN. Involvement of amino acids in periodic inhibitions of bulbar respiratory neurons. Brat’n Res. 327: 351-365, 1982. 4. CURTIS, D. R., L. HOSLI, ANI) G, A. R. JOHNSTON. A pharmacological study of the depression of spinal neurons by glycine and related amino acids. &p. Bruin Res. 6: l-18, 1968. 5. EZURE, K., AND M. MANABE. Decrementing expiratory neurons of the Botzinger complex. II. Direct inhibitory synaptic linkage with ventral respiratory group neurons. Ex~. Brain Res. 72: 159-166, 1988. 6. FEDORKO, L., AND E. G. MERRILL. Axonal projections from the rostra1 expiratory neurones of the Botzinger complex to medulla and spinal curd in the cat. J. Physiol. Land. 350: 487-496, 1984.

IPSPs

7, FELDMAN, J., AND J. C. SMITH. Cellular mechanisms underlying modulation of breathing pattern in mammals. Ann. NY Acud. Sci. 563: 114-130,1989. 8. HAJI, A., J. E. REMMERS, C. A. CONNELLY, AND R. TAKEDA. Effects of glycine and GABA on bulbar respiratory neurons of the cat. J. Neurophysiot. 63: 955-965, 1990. 9. HEYER, E.J.,L.M. NOWAK,AND P.L. MACDONALD. Bicuculline:A convulsant with synaptic and non-synaptic actions. Neurology 31: 1381-1390,1981. 10. KRNJEVIC, K. Chemical nature of synaptic transmission in vertebrates. Physdol. Reu. 54: 418-450, 1974. 11. KRNJEVIC, K., E. PUIL, AND R. WERMAN. Bicuculline, benzyl penicillin and inhibitory amino acids in the spinal cord of cat. Can. J. Physiol. Pharmucol. 55: 670-680, 1977. C. A., AND A. J. BERGER. Tmmunocytochemical local12. LIVINGSTON, ization of GABA in neurons projecting to the ventrolateral nucleus of the solitary tract. Bruin Res. 494: 143-150, 1989. 13. MERRILL, E. G., J. LIPSKI, L, KUBIN, AND L. FEDORKO. Origin of the expiratory inhibition of nucleus tractus solitarius inspiratory neurones. Brain Res. 263: 43-50, 1983. 14. Moss, I., M. DENAVIT-SAUBIE, F. L. ELDRIDGE, R. A. GILLIS, M. HERKENHAM, AND S. LAHIRI. Neuromodulators and transmitters in respiratory control. Federation Proc. 45: 2133-2147, 1986. 15. MUELLER, R. A., D. B. A. LUNDBERG, G. R. BREESE, J. HEDNER, T. HEDNER, AND J. JONASON. The neuropharmacology of respiratory control. Phurmucol. Reu. 34: 255-285, 1982. 16. POLC, P. Electrophysiology of benzodiazepine receptor ligands: multiple mechanisms and sites of action. Prog. Neurobiol. 31: 349423, 1988. 17. REMMERS, J. E., D. W. RICHTER, D. BALLANTYNE, C. R. BAINTON, AND J. P. KLEIN. Reflex prolongation of stage I of expiration. Pfluegers Arch. 407: 190-198, 1986. 18. RICHTER, D. W. Generation and maintenance of the respiratory rhythm. J. Exp. Biol. 100: 93-107, 1982. 19, RICHTER, D. W., D. BALLANTYNE, AND J. E. REMMERS. How is the respiratory rhythm generated? A model. News Physiol. Sci. 1: 109112,1986. 20. RICHTER, D. W., H. CAMERER, M. MEESMANN, AND N. ROHRIG. Studies on the synaptic interconnection between bulbar respiratory neurones of cats. Pfluegers Arch. 380: 245-257, 1979. 21. RITCHIE, J. M. Tetrodotoxin and saxitoxin and sodium channels of excitable tissues. Trends Pharmucol. Sci. 1: 275-279, 1980, 22, SONNHOF, U. A multi-barrelled coaxial electrode for iontophoresis and intracellular recording with a gold shield of the central pipette for capacitance neutralization. Pflwgers Arch. 341: 351-358, 1973. 23. TAKEDA, R., AND A. HAJI. Micro-iontophoresis of flurazepam on inspiratory and postinspiratory neurons in the ventrolateral medulla of cats: an intracellular study in vivo. Neurosci. Lett. 102: 261-267,1989. 24. TAKEDA, R., J. E. REMMERS, J. P. BAKER, K. P. MADDEN, AND J. P. FARBER. Postsynaptic potentials of bulbar respiratory neurons of the turtle. Respir. Physiol. 64: 149-160, 1986.


Not GABA but glycine mediates segmental, propriospinal, and bulbospinal postsynaptic inhibition in adult mouse spinal forelimb motor neurons.

Valproic acid augments GABA-mediated postsynaptic inhibition in cultured mammalian neurons.

Glycine, GABA and synaptic inhibition of reticulospinal neurones of lamprey.

Modification of orientation sensitivity of cat visual cortex neurons by removal of GABA-mediated inhibition.

GABA, glycine, aspartate, glutamate and taurine in the vestibular nuclei: an immunocytochemical investigation in the cat.

The scotopic threshold response of the cat ERG is suppressed selectively by GABA and glycine.

Enhancement of GABA-mediated postsynaptic inhibition in cultured mammalian spinal cord neurons: a common mode of anticonvulsant action.

Guanidino compounds that are increased in hyperargininemia inhibit GABA and glycine responses on mouse neurons in cell culture.

In vivo evidence for an inhibitory glutamatergic control of serotonin release in the cat caudate nucleus: involvement of GABA neurons.

Baroreceptor inhibition of subretrofacial neurons: evidence from intracellular recordings in the cat.

Bulbar reticular neurons relaying somatosensory information to the mesencephalic parabrachial area of the cat.

The effects of intranigral GABA and dynorphin A injections on striatal dopamine and GABA release: evidence that dopamine provides inhibitory regulation of striatal GABA neurons via D2 receptors.

Accommodative reactions of medullary respiratory neurons of the cat.

Benzodiazepines: GABA and glycine receptors on single neurons in the rat medulla.

Spatial distribution of various types of bulbar respiratory neurons in the rabbit.

Activation-Dependent Rapid Postsynaptic Clustering of Glycine Receptors in Mature Spinal Cord Neurons.

The role of GABA in inhibitory synaptic inputs on inhibitory burst neurons in the cat.

Immunocytochemical evidence that GABA and neurotensin exist in different neurons in laminae II and III of rat spinal dorsal horn.

Cl(-)-mediated interaction between GABA and glycine currents in cultured rat hippocampal neurons.

Morphology and electrophysiology of superior laryngeal nerve afferents and postsynaptic neurons in the medulla oblongata of the cat.

Postembedding immunocytochemistry for GABA and glycine reveals the synaptic relationships of the dopaminergic amacrine cell of the cat retina.

Involvement of GABA and glycine in recurrent inhibition of spinal motoneurons.

Synaptic response of bulbar respiratory neurons to hypercapnic stimulation in peripherally chemodenervated cats.

Neostriatal GABAergic Interneurons Mediate Cholinergic Inhibition of Spiny Projection Neurons.