INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

WE RECENTLY REPORTED that neurons in the region of the ventrolateral nucleus of the solitary tract (vlNTS) of the rat have an important role in the control of respiratory function (50). Our hypothesis was that these neurons act as part of the "inspiratory off-switch" (IOS) and terminate the inspiratory phase of the respiratory cycle. This was tested using drugs that hyperpolarize (GABA), disfacilitate (kynurenic acid), and block both axonal conduction and action potential generation (tetrodotoxin). These drugs were used as tools to ascertain the function of the vlNTS and were not used to delineate the neurotransmitter systems involved. Bilateral microinjection of each of these agents prolonged inspiratory duration (TI) and evoked an apneustic pattern of breathing. These results supported our hypothesis that neurons in the region of the vlNTS promote the transition from inspiration to expiration and function as part of the IOS.

As vlNTS neurons appear to function as an integral part of the IOS in the rat, we next focused our attention on the question of which central nervous system neurotransmitter systems control the level of activity of these neurons. Recent data of Ellenberger (11) suggest a widespread role for GABAergic regulation of respiratory function within respiratory-related portions of the NTS (e.g., ventrolateral, interstitial, and commissural subnuclei). This was demonstrated for glutamic acid decarboxylase (GAD) immunohistochemistry. Relatively high GAD immunoreactivity was found associated with each of these subnuclei. Terai and colleagues (45) found positive staining of presumptive terminals in the vlNTS using a polyclonal antibody to GABA. These investigators also examined the distribution of GABA_A receptors in the NTS and reported that this subunit was moderately to highly expressed within the vlNTS. They used a monoclonal antibody to the 2/3 subunit of the GABA_A receptor. Finally, we and others reported that TI increases after intravenous administration of baclofen to selectively activate GABA_B receptors (20, 44).

Thus the primary aim of the present study was to evaluate the role of GABAergic neurotransmission in the vlNTS of the rat in the control of respiratory function with a specific focus on control of the inspiratory phase of the respiratory cycle.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

General Experimental Methods

Stereotaxic Surgery and Microinjection Procedure

Assessment of Apneusis, Apnea, Analysis of Data

An apneustic breath was defined as a prolongation of inspiration in which the depth of breathing (i.e., iEMG amplitude) was maintained at a level equal or nearly equal to the maximum iEMG amplitude of control breaths for a duration >650 ms and having a TI of >900 ms. Onset of apneusis was judged to occur at the second apneustic breath. Duration of apneustic breathing was determined as the time from the second apneustic breath (i.e., onset) to the occurrence of the second-to-last apneustic breath. Apnea was defined as a cessation of inspiratory activity of >5 s duration occurring after drug administration. It was observed (see RESULTS) that unilateral microinjection of vehicle occasionally produced transient respiratory effects, principally a prolongation of TE but sometimes included cessation of breathing that reversed with the termination of the microinjection. For this reason, we defined drug-induced apnea as occurring only after administration of the drug had been completed.

If a drug either prolonged TI, evoked apneusis, or both, then cardiorespitory variables were measured at the time of maximal TI and 5 min after the drug administration. The means of five successive respiratory cycles were used to calculate the values reported in this study. In instances where other than obvious reduction of TI, apnea, or both was produced, postdrug cardiorespiratory values were taken within the first 5 s after microinjection had been completed. This was done to minimize the effects of altered respiration on cardiovascular function. The duration of apnea was determined as the time between cessation of breathing after completion of drug administration and the last point in the experimental record at which respiration could not be maintained unsupported.

Experimental Protocols

Studies were designed to maximize the amount of information obtained from each animal. Thus a number of different microinjection protocols were used depending on the drugs used and the expected response observed from preliminary studies. All experiments using GABA receptor agonists or reuptake inhibitors involved bilateral microinjection as it was noted previously (50) that apneusis was only produced when the region of the vlNTS was inhibited bilaterally. All antagonist studies involved unilateral microinjections as it was observed in preliminary experiments that this was sufficient to produce significant respiratory alterations. Additionally, GABAergic antagonists were not microinjected before GABAergic agonists/reuptake inhibitors because they produced clear respiratory effects on their own. Thus, unless the antagonist was given to assess antagonist effects per se, antagonists were given unilaterally after bilateral agonist-induced apneusis had stabilized; generally, this occurred 20-60 s after the end of agonist administration.

Cardiorespiratory effects of either dose of nipecotic acid (i.e., 5 nmol or 25 nmol) were brief (<4 min in all cases) and demonstrated complete recovery. Because of this, both doses of drug could be tested in the same animal. Typically this involved either bilateral microinjection of one dose followed by a 10- to 20-min waiting period before bilateral microinjection of the other dose (i.e., dose response) into the same site or the bilateral microinjection of the same dose into a different site (i.e., into the mNTS or DMRN for site response). Experiments were conducted such that dose or site order was randomized. The time interval allowed to elapse between injections was chosen to be more than double the longest duration of cardiorespiratory effects observed with the highest dose of nipecotic acid tested in the vlNTS to avoid the effects of a prior intervention. No animal received more than two sets of microinjections.

As mentioned previously, unilateral microinjection of GABA receptor antagonists could elicit robust respiratory effects, depending on the site of microinjection. Respiratory rhythm frequently returned to approximately baseline values, which allowed us to test the opposite side of the brain after a 15- to 20-min period of recovery. No animal received more than one microinjection on either side of the brain stem. Cardiorespiratory effects of L-glutamate microinjection were of even shorter duration and were made unilaterally (as this was shown to produce cardiorespiratory effects). Thus, because of the short duration of the L-glutamate effect, multiple unilateral microinjections could be made in each animal. In these experiments, the dose was always the same (300 pmol) but the target site was chosen randomly as was site order. No animal received more than four unilateral microinjections.

In experiments in which drugs produced apnea, animals were ventilated only if mean arterial pressure fell markedly, typically 30 s after the onset of apnea. Once placed on the ventilator, the animal was tested once or twice per minute to determine if spontaneous respiration would occur with removal from artificial ventilation. The animal was placed back on the respirator until testing indicated the animal could maintain respiratory rhythm unsupported.

Histology

View larger version (51K): [in this window] [in a new window]   Fig. 1.   Response to bilateral microinjection of nipecotic acid into the ventrolateral nucleus of the solitary tract (NTS) in the anesthetized, vagotomized, and spontaneously breathing rat. A: photomicrograph of ventrolateral NTS (vlNTS) microinjection sites, denoted by arrows, which are observed at 0.1 mm caudal to obex. Bar = 0.25 mm. B: baseline cardiorespiratory activity. Time bar = 1 s. C: microinjection of 5 nmol/45 nl nipecotic acid into the left vlNTS (right vlNTS microinjection accomplished within the preceding 45 s). Black arrow denotes the end of microinjection. D: 2 min postbilateral 5 nmol nipecotic acid microinjection, note return to baseline cardiorespiratory activity. E: 400 breath time-course depicting the effect of nipecotic acid on inspiratory duration (TI). Bilateral nipecotic acid administration occurs during the inset bar and denoted "nip." F: baseline cardiorespiratory activity 21 min after D. Gray arrow denotes the start of microinjection of 25 nmol/45 nl nipecotic acid into the left vlNTS region (right vlNTS microinjection accomplished within the preceding 30 s) and ends at black arrow in G. H: 2 min postbilateral microinjection of 25 nmol nipecotic acid. Note apneusis is still evident. I: 400 breath time course depicting the effect of 25 nmol nipecotic acid on TI. BP, blood pressure; dEMG, raw diaphragmatic electromyogram signal; iEMG, integrated dEMG signal; HR, heart rate; XII, hypoglossal nucleus; DMV, dorsal motor nucleus of the vagus; TS, solitary tract.

Immunohistochemistry Protocol

Tissue preparation. Five rats that received no prior surgical intervention were deeply anesthetized with pentobarbital sodium (50 mg/kg) and perfused through the aorta with 250 ml of a 0.1 M PBS (pH 7.4) solution. This was followed by 500 ml of 2% paraformaldehyde solution in 0.1 M PBS solution (pH 7.4). The brains were removed and postfixed in this fixative solution for at least 24 h at 4°C. Brains were then transferred to a 30% sucrose solution in PBS, pH 7.3, and allowed to equilibrate for at least 48 h at 4°C. Brains were removed, cut in 30 µm sections and mounted on gelatin-coated Superfrost+ glass slides (Fisher Scientific; Pittsburgh, PA). Brain slices from the level of the caudal brain stem to 1 mm rostral to the opening of the 4th ventricle were stored at 80°C until processed for immunohistochemistry. These slices usually included cerebellar tissue in addition to medullary tissue. As the distribution of GABA receptor subtypes in the cerebellum has been previously described (14, 27), the cerebellar tissue was used as a positive control for our antisera.

Antisera. To label GABA_A receptors, we used a commercially available (Upstate Biotechnology, Lake Placid, NY), affinity purified polyclonal rabbit antibody formed against the synthetic peptide QPSQDELUDNTTUFT, which corresponds to amino acids 1-15 of the subunit of this receptor. Similarly, to label GABA_B receptors, we used a polyclonal guinea pig antibody (Chemicon, Temecula, CA) against the synthetic peptide PSEPPDRLSCDGSRVHLLYK that is common to both R1 and R1 isoforms.

Protocol. Each coronal section was circled with a PAP pen (Research Products International, Mount Prospect, IL) to form incubation wells. The sections were then rehydrated for 20 min with PBS solution, blocked for 20 min with a 5% reconstituted milk solution containing 0.3% Triton X-100 along with 0.2% sodium azide in PBS, and incubated with one or more primary antibodies in this blocker solution. Primary antibody incubation with either the rabbit anti-GABA_A receptor (1:300 dilution), guinea pig anti-GABA_B receptor (1:500), or a combination of these primary antibodies was allowed to proceed for 1 h at room temperature followed by 18-24 h of incubation at 4°C. Primary solution was washed off and sections were then washed in PBS and incubated for 2 h in either FITC-conjugated donkey anti-rabbit IgG, TRITC-conjugated donkey anti-guinea pig IgG diluted 1:200 in PBS (Jackson ImmunoResearch Laboratory, West Grove, PA), or both solutions. Sections were then placed under a coverslip with anti-fading and fluorescence mounting medium (Vectashield, Vector, Burlingame, CA).

Immunofluorescence assessment. Immunofluorescence was examined using a Nikon fluorescence microscope (model Eclipse E600, Tokyo, Japan). Images were captured using a charge-coupled device camera (Optronics Engineering model DEI-750, Golete, CA) and Bioquant TCW95 v.250.4 software (Bioquant-R&M Biometrics, Nashville, TN). Images were then processed using Adobe Photoshop 4.0 (Adobe, Mountain View, CA). Image processing consisted of reduction of brightness, increase of contrast, and sharpening to enhance fluorescence. Control sections stained with either one or none of the primary antibodies in the presence of both secondary antibodies produced staining selective only for the primary antibody (i.e., no "cross-over" or "bleed-through" staining was observed upon switching from FITC to Texas red visualization or vice versa; data not shown).

Data Analysis

Research described in this paper fully conforms with the American Physiological Society's "Guiding Principles for Research Involving Animals and Human Beings."

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effects of Bilateral Microinjection of Nipecotic Acid Into the vlNTS Region

Bilateral microinjection of 25 nmol of nipecotic acid in eight rats produced even greater changes in respiratory function than the 5 nmol dose (Table 1, Fig. 1, F-I). TI was more markedly affected (285 ± 46% of baseline; Table 1) and apneusis was produced in all eight animals studied. TE was reduced (to 65 ± 12% of baseline) as with the 5 nmol dose. Unlike the low dose group, however, respiratory rate was reduced in animals receiving the 25 nmol dose. No significant changes were seen in the amplitude of the iEMG after bilateral microinjection of 25 nmol nipecotic acid. Prolongation of TI and shortening of TE could be observed with unilateral administration of this dose of nipecotic acid although apneusis was only seen after microinjection into the vlNTS on both sides. Cardiovascular effects consisted of an elevation of mean arterial pressure without alteration of heart rate. Onset of respiratory effects and, indeed apneusis, was rapid (within seconds, Table 1; Fig. 1G). The duration of these effects was longer lasting than after the 5 nmol dose but was still relatively short lived (Table 1; Fig. 1I). The time course of the increase in blood pressure mirrored the duration of respiratory effects with the exception that there appeared to be a progressive increase in blood pressure during microinjection on both sides. However, 2 of 8 animals receiving 25 nmol nipecotic acid did not exhibit a rise in blood pressure (e.g., Fig. 1, F-H). Cardiorespiratory changes were not significantly different from baseline 5 min after bilateral microinjection of drug (Table 1).

Microinjection sites for the experiment depicting the effects of 5 and 25 nmol of nipecotic acid in a single animal appear as Fig. 1A and indicate that the micropipette tips were located in the region of the vlNTS. The sites of nipecotic acid microinjection into the vlNTS region as well as those targeting the mNTS (described below) are shown in Fig. 2. All microinjection sites were found located within a region 0.5 mm caudal to 0.1 mm rostral to the opening of the central canal into the 4th ventricle (obex). Microinjections into the region of the vlNTS were distributed from an area lateral to the solitary tract to a region lateral and ventral to the solitary tract. These microinjection sites are located within the general borders of the ventrolateral and interstitial subnuclei of the solitary tract as defined by Kalia and Sullivan (18). Focusing on microinjection sites depicting the effects of 5 nmol of nipecotic acid, as mentioned above, an apneustic pattern of breathing occurred in 3 of 12 animals. These sites were always in the region of the vlNTS. Apneusis appeared to be an all-or-near nothing event. That is, in the nine animals where apneusis was not observed, TI was only increased slightly, if at all, even though the micropipette tips were located in the region of the vlNTS. Thus our impression was that 5 nmol of nipecotic acid was close to a threshold dose for inducing apneusis.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Summary of microinjection sites. Camera lucida reconstruction detailing microinjection sites of nipecotic acid of two different doses: 5 (A) and 25 nmol (B). Microinjection sites that were associated with apneusis are indicated by black circles; those that were not associated with apneusis are indicated with gray circles. Note that microinjections of 5 nmol nipecotic acid into the medial NTS described in RESULTS are also included in this figure. Coordinates listed describe location relative to obex; see METHODS for description. AP, area postrema.

Effects of Bilateral Microinjections of Nipecotic Acid Into the mNTS Region

View larger version (62K): [in this window] [in a new window]   Fig. 3.   Response to bilateral microinjection of nipecotic acid into the medial NTS. A: baseline cardiorespiratory activity. Time bar = 1 s. B: microinjection of 5 nmol nipecotic acid (black arrow) into the left medial NTS (right mNTS microinjection accomplished within the preceding 30 s). Note the lack of effect on respiration but an increase in blood pressure. C: 1 min after microinjection demonstrates blood pressure remians elevated with no alterations in inspiratory duration. D: cardiorespiratory traces obtained 5 min after medial NTS (mNTS) microinjection. Blood pressure returned to baseline. E: 400 breath time course depicting the lack of an effect of 5 nmol nipecotic acid on inspiratory duration. F: photomicrograph of mNTS microinjection sites denoted by arrows which are observed at 0.2 mm caudal to obex. Scale bar = 0.5 mm.

Effects of Bilateral Microinjection of Muscimol and Baclofen Into the vlNTS Region

Bilateral microinjection of the GABA_B receptor agonist baclofen (75 pmol) produced a very similar profile of cardiorespiratory effects as observed with muscimol. TI significantly increased, TE significantly decreased, and apneusis was produced in 10 of 11 animals. Respiratory rate was decreased due to the increase in TI, and the iEMG amplitude was elevated significantly at the time of the peak respiratory effects (Table 2). Although TI was often prolonged and TE reduced after unilateral microinjection, apneusis was never observed after unilateral microinjection of baclofen. However, apneusis developed nearly immediately (14 ± 4 s) after completion of bilateral microinjection of baclofen. Peak respiratory effects (i.e., apneusis) were observed 56 ± 12 s after bilateral microinjection of baclofen and generally appeared to have the same time course of action as muscimol. As in the case of muscimol, if iEMG amplitude was assessed 10-20 s after developing baclofen-induced apneusis (rather than during the peak effect approximately a minute after drug administration) it was not significantly different from baseline (+18 ± 17%). Cardiovascular alterations also were similar to that of muscimol: mean arterial pressure was elevated and heart rate remained unchanged with the cardiovascular time course mirroring that of the respiratory effects (Table 2).

To assure that the doses of drugs tested were effective and selective in stimulating each type of GABA receptor, studies were performed wherein apneustic breathing was produced by bilateral microinjection of each GABA receptor agonist into the vlNTS region followed by a unilateral microinjection of the respective GABA receptor antagonist into the same site. As can be seen in Figs. 4A and 5A, apneustic breathing produced by 100 pmol muscimol was readily reversed by subsequent administration of the GABA_A receptor antagonist bicuculline methiodide, 25 pmol (n = 5). Additionally, pressure increases observed with bilateral muscimol administration into the vlNTS also were reversed by bicuculline (Fig. 4A). The specific GABA_B receptor antagonist CGP-35348 was similarly able to reverse apneusis produced by 75 pmol baclofen (Figs. 4B and 5B) and also reduced the baclofen-induced pressor response (Fig. 4B). Although apneusis was produced only after bilateral microinjection of the agonist, reversal of apneusis required only unilateral administration of the antagonist. Additionally, reversal of apneustic breathing always preceded the reversal of GABA receptor agonist-induced pressor response, usually by 15 s to 1 min.

View larger version (28K): [in this window] [in a new window]   Fig. 4.   Reversal of muscimol and baclofen-induced apneusis with bicuculline and CGP-35348, respectively. A: left depicts apneusis produced by bilateral microinjection of muscimol 100 pmol into the vlNTS. Middle indicates the effect of unilateral microinjection of bicuculline 25 pmol into the vlNTS (black arrow denotes end of microinjection). Reversal of apneusis occurs during administration. Right obtained 1 min after bicuculline microinjection. B: left depicts apneusis produced by 75 pmol baclofen. Gray arrow at the end of the left denotes beginning of microinjection of CGP-35348. Middle indicates completion of CGP-35348 microinjection 40 s later (black arrow) with clear reversal of apneusis during drug administration. Right obtained 1 min after CGP 35348 injection. Time bars = 1 s.

View larger version (38K): [in this window] [in a new window]   Fig. 5.   GABA receptor subtype antagonists act specifically to reverse only corresponding agonist-induced apneusis. Agonists were microinjected bilaterally into the vlNTS (muscimol 100 pmol, baclofen 75 pmol) while antagonists were microinjected unilaterally (bicuculline 25 pmol, CGP-35348 5 nmol). Bar groups denoted A and B depict appropriate agonist/antagonist competition while bar groups C and D depict the results of cross-blocking experiments. Agonist data obtained at maximal response. Postantagonist data obtained 1 min after antagonist administration. * P < 0.05 vs. baseline.

Cross-blocking studies in which the GABA_B receptor antagonist was given after GABA_A receptor agonist administration and the GABA_A receptor antagonist given after GABA_B receptor agonist demonstrated the specificity of these drugs for the receptor subtypes. CGP-35348 was unable to reduce the apneustic breathing produced by muscimol (Fig. 5C), and bicuculline was ineffective in reversing baclofen-induced apneustic breathing (Fig. 5D). Furthermore, CGP-35348 was unable to reverse the increased blood pressure produced by muscimol and bicuculline was unable to counteract the increase in pressure after microinjection of baclofen (data not shown).

Effects of Bilateral Microinjection of Drug Vehicle Into the vlNTS Region

Effects of Unilateral Microinjection of GABA Receptor Antagonists Into the vlNTS Region

View larger version (48K): [in this window] [in a new window]   Fig. 6.   Response to unilateral microinjection of GABA receptor subtype antagonists into the vlNTS. A, B, and C depict unilateral microinjection of bicuculline 25 pmol, CGP-35348 5 nmol, and a cocktail of bicuculline and CGP-35348, respectively, into 3 different animals. Left in each case are baseline recordings of cardiorespiratory activity. Time bars = 1 s. Middle, effects produced by microinjection of antagonist(s). Gray and black arrows denote the start and end of microinjection of antagonist, respectively. Asterisk in A denotes noise generated by a ventilator switched on but not applied to the animal. Right panels are taken 1 min after microinjection of antagonist. Note that the animal depicted in C returned to a normal pattern of breathing over the succeeding 10 min. Photomicrographs on the far right are of the identified microinjection sites in each experiment that were observed at 0.05 mm (A), 0.3 mm (B), and 0.2 mm (C) relative to obex. Abbreviations and arrows in the photomicrographs are as in Fig. 1.

In contrast to the intense respiratory effect observed with bicuculline methiodide, unilateral microinjection of CGP-35348 (5 nmol/45 nl) only rarely produced apnea. In the 10 experiments performed in seven animals, only one animal exhibited apnea, and this event lasted just 13 s. The primary effects of CGP-35348 were to decrease TI and to increase TE. Thus, at 30 s after unilateral microinjection of CGP-35348, TI was reduced from 451 ± 19 to 423 ± 13 ms (P < 0.05) and TE was increased from 604 ± 48 to 742 ± 53 ms (P < 0.05) (see Fig. 6B). Recovery from these effects of CGP-35348 occurred within 1 min. Blood pressure was observed to be significantly decreased 30 s after unilateral microinjection of CGP-35348 (79 ± 4 to 67 ± 5 mmHg); however, heart rate remained unchanged.

Three animals received a combination of the two antagonists (bicuculline methiodide, 25 pmol, and CGP-35348, 5 pmol, in a total volume of 45 nl), and this cocktail was microinjected unilaterally into the vlNTS in each animal with the procedure repeated in two animals on the contralateral side for a total of five microinjections. In all instances, apnea occurred immediately, and in three cases required artificial ventilation of the animal (average duration of apnea in each animal tested was 154 ± 91 s; range 17-329 s). Measurements of TI, TE, iEMG amplitude, mean blood pressure, and heart rate were not obtained because of the quick onset of apnea and the influence of respiratory changes on cardiovascular function. A representative experiment depicting the effects of the cocktail appears as Fig. 6C.

Quaternary salts of bicuculline are also capable of blocking calcium-dependent potassium channels (37). This is not the case for the GABA_A receptor antagonist SR-95531 (Gabazine) (37). Hence, we performed three experiments in three rats wherein unilateral microinjections of 25 pmol of SR-95531 were made into the vlNTS region. In each instance, an immediate apnea occurred, and this response was indistinguishable from that produced by bicuculline methiodide. A representative experiment appears as Fig. 7.

View larger version (12K): [in this window] [in a new window]   Fig. 7.   Response to unilateral microinjection of GABAA receptor antagonist SR-95531 (Gabazine) into the vlNTS. A: baseline cardiorespiratory activity, bar = 1 s. B: unilateral microinjection of SR-95531 (25 pmol) into the right vlNTS produces an immediate apnea that necessitated placing the animal on a respirator (not shown) within seconds after the end of the panel. The animal was incapable of supporting its own breathing and required artificial ventilation for 10.5 min after this unilateral microinjection. C: 18 min after unilateral microinjection, the animal is taken off the respirator but respiratory rhythm was still erratic.

Effects of GABA Receptor Antagonists vs. Nipecotic Acid-Induced Changes in Cardiorespiratory Function

View larger version (30K): [in this window] [in a new window]   Fig. 8.   GABA receptor antagonist reversal of nipecotic acid-induced apneusis. A, C, and E, 400 breath time course depicting a representative experiment in which apneusis produced by bilateral microinjection of nipecotic acid (Nip) 25 nmol was followed by unilateral microinjection of bicuculline 25 pmol (Bic; A), CGP-35348 5 nmol (CGP; C), cocktail of bicuculline and CGP-35348 (Combo; 25 pmol and 5 nmol, respectively; E). Note that bicuculline fails to reverse apneustic breathing produced by nipecotic acid and CGP-35348 produces a brief reduction in TI during administration which immediately returns to apneustic breathing at the end of injection. Additionally, minutes after unilateral bicuculline injection, inspiratory duration becomes significantly shortened and transient episodes of apnea develop. Unilateral microinjection of nipecotic acid (N) to the same side as bicuculline immediately restores respiration to eupnea. Unilateral microinjection of the cocktail of bicuculline and CGP-35348, however, immediately reverses nipecotic-induced apneusis (E) and produces a late developing apnea that is effectively reversed by unilateral microinjection nipecotic acid to the same side as the cocktail. B, D, and F, group means for animals receiving bicuculline (n = 5), CGP-35348 (n = 4), or the combination cocktail (n = 8) after apneusis evoked by nipecotic acid. Only microinjection of the antagonist combination was able to reverse nipecotic acid-induced apneusis. * P < 0.05 vs. baseline, # P < 0.05 vs. peak drug effect.

Effects of Unilateral Microinjection of Kynurenic Acid Into the vlNTS Region on Bicuculline-Induced Apnea

View larger version (26K): [in this window] [in a new window]   Fig. 9.   Blockade of bicuculline-induced respiratory alterations with kynurenic acid. Note this is a representative experiment conducted in a single animal. A: left is a baseline recording of cardiorespiratory activity. Right indicates the effects of bicuculline methiodide (25 pmol) that was microinjected unilaterally into the vlNTS (gray arrow denotes start, black arrow denotes end of microinjection). Apnea developed during microinjection and the animal was placed on a ventilator 30 s later (not shown). Total period of apnea lasted slightly over 6 min (367 s) before returning to normal 25 min postbicuculline. B: left is a cardiorespiratory trace obtained 1 min postkynurenic acid (1 nmol) microinjection into the vlNTS and just before a second bicuculline microinjection (right, arrows). Note apnea does not develop after bicuculline and alterations in respiratory activity are slight and disappear by the end of the tracing. C: left obtained 20 min later and respiratory activity appears similar to baseline. Right panel indicates the effect of a third unilateral microinjection of bicuculline (arrows) that produces mild to moderate prolongation of expiratory duration (TE) and shortening of TI, which lasts for 7 min before returning to normal (not shown).

Effects of Microinjection of Glutamate Into the Dorsomedial Medulla

View larger version (28K): [in this window] [in a new window]   Fig. 10.   Microinjection of glutamate into the dorsal medulla. A representative experiment conducted in 1 animal in which glutamate 300 pmol was microinjected unilaterally into 3 regions of the medulla: A: vlNTS; B, medial NTS 7.7 min after previous microinjection; and C, dorsomedial reticular nucleus (~0.5 mm lateral to vlNTS) 6.2 min after the previous injection. Both ventrolateral and medial NTS microinjections of glutamate produce apnea but demonstrate different effects on blood pressure. Microinjection of glutamate into the dorsomedial reticular nucleus produced no effect on cardiorespiratory function. Gray and black arrows denote start and end of glutamate microinjection, respectively. Time bars = 1 s. Group means are tabulated in Table 3.

Distribution of GABA_A and GABA_B Receptors in the vlNTS Region

View larger version (36K): [in this window] [in a new window]   Fig. 11.   Distribution of GABAA and GABAB receptors in the vlNTS and cerebellum of the rat. Coronal sections of the medulla at the level of the AP and cerebellum were double-labeled using antibodies against the 1-subunit of the GABAA receptor and the GABAB receptor. A-C are taken from the vlNTS region. A depicts GABAA receptor immunoreactivity (GABAA-IR) within the vlNTS while B depicts GABAB-IR. Colocalization of receptors is shown in C. D-F are taken from a double-labeling experiment in the cerebellum demonstrating the specificity of the antibodies used. As previously described (14, 28) the granular layer is found to be GABAA-IR positive (D) while GABAB-IR is found primarily on Purkinje neurons (E). Little colocalization of receptors is found in this region (F). Scale bar in A is 50 µm and indicates dorsal and medial orientation within the section.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Our data suggest that GABA plays a role at the vlNTS in the central control of breathing. This was established using three pharmacologic approaches consisting of microinjection of 1) specific antagonists of GABA_A and GABA_B receptors, 2) a GABA reuptake inhibitor, and 3) specific agonists of GABA_A and GABA_B receptors into the vlNTS of the rat. Data obtained with antagonists of GABA_A or GABA_B receptors, bicuculline methiodide, or CGP-35348, respectively, indicate a reduced duration of TI and an increased duration of TE. Bicuculline, in contrast to CGP-35348, nearly always produced apnea. The production of apnea with GABA_A receptor blockade suggests the strongest synaptic GABAergic input to the vlNTS is mediated by GABA_A receptors. However, endogenous GABA also activates GABA_B receptors at the vlNTS as reflected by the changes in cycle duration after CGP-35348 microinjection. Activation of GABA receptors with muscimol and baclofen and, indirectly, with nipecotic acid increased the duration of TI and decreased the duration of TE. Whereas the responses obtained with muscimol and baclofen were through activation of GABA_A and GABA_B receptors, respectively, the responses obtained with nipecotic acid were mediated through both GABA_A and GABA_B receptors. Finally, both the GABA receptor agonists and the GABA receptor uptake inhibitor affected TI so strongly that apneustic breathing nearly always occurred. Alterations in GABAergic transmission using the above drugs had little effect on the intensity of inspiratory drive as reflected by the lack of significant effects on the amplitude of the iEMG.

Other investigators have studied GABAergic neurotransmission in the NTS in the rat but not specifically in the vlNTS (e.g., Ref. 41). We are aware of only one other study relating to drugs affecting GABAergic neurotransmission in the vlNTS to respiratory function and that is by Miyazaki and colleagues (29). These investigators, using extracellular recordings from respiratory neurons in the vlNTS combined with iontophoretic application of pharmacologic agents, describe a type of respiratory neuron that responds to both bicuculline and the GABA_B weak antagonist/partial agonist saclofen. This respiratory neuron received monosynaptic input from slowly adapting pulmonary stretch receptors (SARs; Refs. 28, 29) as well as rhythmic inputs from the central respiratory system. One of these inputs resulted in facilitation of firing at around the period from late inspiration to early expiration (IE facilitation; Fig. 12). IE facilitation was reduced by iontophoretic application of the NMDA receptor channel blocker MK-801, indicating that a glutamatergic synapse on the respiratory neuron was responsible for this excitation during the time of transition from inspiration to expiration (29). Thus blockade of this phasic drive (Fig. 12) would be expected to prolong TI and could potentially cause apneustic breathing.

View larger version (56K): [in this window] [in a new window]   Fig. 12.   Model of probable inputs to vlNTS respiratory neurons. Schematic representation of synaptic inputs to a type of respiratory neuron documented by Miyazaki et al. (29) to be present in the vlNTS of the rat. IE input is active at the transition period from inspiration to expiration. (This figure is adapted with permission from Fig. 12 of Ref. 29). SAR, slowly adapting stretch receptor. Receptors on respiratory neuron: A/K, non-NMDA; N, NMDA; A, GABAA; B, GABAB.

These investigators found that iontophoretic application of bicuculline markedly enhances IE facilitation. This, according to their model adapted in Fig. 12, could theoretically shorten TI, or, if intense enough, produce apnea. Indeed, both responses were observed in our study. A presynaptic location on glutamatergic terminals for at least a portion of GABA_A receptors was suggested based on the demonstration that after bicuculline, the firing of the neuron during the IE period (IE facilitation) was greater relative to baseline firing than that seen before bicuculline iontophoresis. According to the model in Fig. 12, excessive glutamate stimulation of these neurons would be expected to shorten TI or, if the stimulation were strong enough, produce apnea. In support of this idea and schematically depicted in Fig. 12 are our findings that 1) microinjection of glutamate into the vlNTS produces apnea (Fig. 11A; also see Ref. 25) and 2) preliminary studies suggesting that pretreatment with the nonselective glutamate antagonist kynurenic acid, which blocks both NMDA and non-NMDA receptors, prevents bicuculline-induced apnea (Fig. 10B). Although kynurenic acid will not distinguish between glutamate receptor subtypes, taken together, these data suggest that one role of GABA acting at GABA_A receptors at the vlNTS is to oppose or balance glutamatergic drive at this site and allow an adequate period for completion of the inspiratory phase of respiration.

Miyazaki and colleagues (29) reported that bicuculline increased baseline firing of the respiratory neuron and this is the rationale for our model depicting tonic GABA input synapsing directly onto the respiratory neuron (Fig. 12). Saclofen also increased the baseline firing of the respiratory neuron but had no effect on IE facilitation. We assume therefore that GABA exerts a tonic inhibitory effect on both GABA_A and GABA_B receptors located postsynaptically on the respiratory neuron (Fig. 12) and speculate that blockade of these receptors is largely responsible for shortening TI. We also noted that blockade of GABA_A and GABA_B receptors at the vlNTS also prolonged TE. The reason for this is unclear but would not be unexpected because the respiratory neuron depicted in Fig. 12 receives monosynaptic input from SAR afferents (28, 29) and may function as a neuron mediating the Breuer-Hering reflex-induced prolongation of expiration (15).

Most importantly, the schema shown in Fig. 12 predicts that local administration of GABA, GABA_A, and GABA_B receptor agonists or the GABA reuptake inhibitor nipecotic acid should all produce apneustic breathing. GABA_A receptor agonists would act both directly (postsynaptic) and indirectly (presynaptic) to silence the respiratory neuron, whereas the GABA_B receptor agonist would act directly to evoke the same response. Nipecotic acid would lead to an increase in the concentration of GABA at the synapse and could result in activation of both GABA_A and GABA_B receptors. In fact, in our study we found that nipecotic acid-induced apneusis could only be reversed by a combination of bicuculline and CGP-35348.

We used an immunohistochemical approach to assess whether GABA receptors are present in the vlNTS. Our data demonstrate that both GABA_A and GABA_B receptors are localized primarily to neuronal processes and cell bodies of neurons located in the vlNTS, respectively. Terai and colleagues (45) used a monoclonal antibody directed against a subunit of the GABA_A receptor (2/3), which is commonly expressed along with the ₁-subunit in native GABA_A receptors (26). They reported that GABA_A receptors were moderately to highly expressed within the vlNTS and ventral NTS, with moderate immunoreactivity to the 2/3 subunit in the interstitial subnucleus. Immunoreactivity to the 2/3 subunit decreased rapidly medially, and was almost absent within the mNTS itself (45). Additionally, Terai and colleagues (45) used a polyclonal antibody to GABA and found GABA-immunoreactive puncta (which they state are probably nerve terminals) in the vlNTS. According to these investigators, immunoelectron microscopic studies of others have indicated that such puncta are GABA-containing presynaptic terminals, and this supports the tonic GABA input to IE neurons depicted in Fig. 12. To our knowledge, no studies have been published in which the distribution of GABA_B receptors within the subnuclei of the rat NTS has been assessed.

We do not directly record from neurons in the vlNTS in our experiments and demonstrate that the neurons exhibit a respiratory pattern of discharge. Nevertheless, this seems likely as respiratory neurons can be found in this region (7, 28, 42, 46) and the neurons we are affecting appear to have similar inputs to those described by Miyazaki and colleagues (28, 29). Thus our data strongly suggest that GABA plays an important role in controlling the activity of respiratory neurons located in the dorsal respiratory group of rats and that this takes place specifically in the vlNTS. Our evidence for site specificity is based primarily on our nipecotic acid data and on our recently published data on GABA (50). Data obtained with microinjection of 5 nmol of nipecotic acid into the vlNTS or the adjacent mNTS revealed significant respiratory changes elicited from the vlNTS but no significant respiratory changes evoked from the mNTS. Evidence that nipecotic acid microinjections into the mNTS were accurately placed was demonstrated not only anatomically (Figs. 2 and 3) but functionally wherein a pressor effect occurred that was at least as great if not greater than the pressor effect obtained with both 5 and 25 nmol of nipecotic acid microinjected into the vlNTS (Table 1). Furthermore, our previous study (50), which compared the respiratory effects of microinjection of GABA into the vlNTS and the mNTS, revealed identical respiratory changes could be elicited from the vlNTS to those noted with nipecotic acid and muscimol from the vlNTS in the present study, but no comparable respiratory changes were evoked from the mNTS. On the other hand, GABA elicited robust pressor effects from the mNTS that were at least equal in magnitude (if not greater) than that produced from the vlNTS (50).

vlNTS neurons in other species such as the cat represent a major monosynaptic drive to the phrenic motor nucleus and central pattern generation (1, 8, 32). Functional evidence for this drive was not observed in our studies. Perturbation of GABAergic transmission had relatively little or no effect on iEMG amplitude but, in contrast, affected rhythm generation. This might be expected because the rat showed only a minor monosynaptic connection to the phrenic motor nucleus (7, 9, 12, 36, 51). However, the vlNTS has extensive projections to other brain stem regions (16, 28, 51). Data obtained from both electrophysiologic studies (28) and neuroanatomical studies (28) in the rat indicated neural connectivity to the dorsolateral pons, primarily to the Kölliker-Fuse nucleus with some projections to adjacent parabrachial nuclei (28). Since apneustic breathing as well as apnea can be evoked by changing the activity of neurons in the region of the Kölliker-Fuse nucleus (6, 30), we speculate that the connectivity demonstrated between the vlNTS and the dorsolateral pons might explain a significant portion of the respiratory changes evoked from the vlNTS.

Studies performed in the cat have provided evidence of GABAergic inputs to respiratory neurons of the vlNTS from the Bötzinger complex, presumably from expiratory neurons (22, 23). This connectivity from the Bötzinger complex to the vlNTS, however, seems to be absent in the rat (3). Thus the source of afferent GABAergic tone into the vlNTS of the rat (Fig. 12) is not known. Apneustic breathing required that a bilateral microinjection of direct and indirect GABA receptor agonists be administered, whereas apnea could be elicited by unilateral microinjection of a GABA receptor antagonist. These results suggest that the vlNTS receives bilateral inhibitory (GABA) drive to maintain the integrity of the IOS and that excitatory input to each vlNTS region is capable or sufficient to affect inspiratory termination. Apneusis, therefore, would not be expected to occur unless the vlNTS was inhibited bilaterally. Unilateral stimulation of these neurons within the vlNTS, either directly with glutamate microinjection or indirectly with bicuculline, would be expected to produce high rates of firing in these neurons. As one intact vlNTS region is capable of effecting the IOS, hyperstimulation of these neurons would be expected to produce a prolonged activation of the IOS resulting in apnea.

Focusing on the possibility that GABAergic transmission within the vlNTS participates in the control of the IOS, there are published reports from our laboratory and others indicating that systemic administration of baclofen produces apneustic breathing without consistent changes in TE (19, 35, 44). Since microinjection of baclofen into the vlNTS evokes the same pattern of apneustic breathing without an increase in TE (Table 2), it is tempting to suggest that the site of action of systemically administered baclofen is the vlNTS.

Our data obtained with L-glutamate (Table 3) and data of Marchenko and Sapru (24) suggest that vlNTS neurons are dedicated for controlling respiratory function. Evidence for this is that microinjection of an excitatory amino acid into the vlNTS in a small volume (present study and see Ref. 24) produces only a change in respiratory function and no change in cardiovascular function. The mNTS, considered a primary site for cardiovascular regulation, has also been implicated in the control of respiration and contains second-order neurons in the Breuer-Hering reflex pathway (2). Indeed, these investigators reported that with microinjection of small volumes of an excitatory amino acid (not exceeding 3 nl), only respiratory effects could be observed. In our study and in the study of Marchenko and Sapru (24) we could not separate respiratory effects from cardiovascular effects upon microinjection of glutamate into the mNTS. However, we did note only cardiovascular effects when 5 nmol of nipecotic acid were microinjected into the mNTS (Table 1). Regarding cardiovascular effects elicited from the mNTS with nipecotic acid, Sved (40) reported that GABA under the conditions of reuptake inhibition with nipecotic acid acts primarily on GABA_B receptors to elicit a cardiovascular response (i.e., a rise in arterial pressure). On the other hand, both bicuculline and CGP-35348 microinjected into the mNTS exerted effects per se. Both agents given separately decreased arterial pressure, approximately by the same magnitude (41). Thus tonic GABAergic drive to the mNTS acts to raise arterial pressure, presumably by acting equally on GABA_A and GABA_B receptors, whereas the indirect GABAergic effect of nipecotic acid results in a rise in arterial pressure due to excitation of GABA_B receptors.

In summary, evidence from pharmacologic and immunohistochemical approaches performed in the vlNTS of the anesthetized rat reveal that GABAergic transmission is crucially important in controlling the activity of these neurons and therefore in modulating respiratory function. Furthermore, the vlNTS appears to contain a pool of neurons dedicated to respiration and forms a major component of the IOS.