• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of jphysiolThe Journal of Physiology SiteMembershipSubmissionJ Physiol
J Physiol. Apr 15, 1999; 516(Pt 2): 571–582.
PMCID: PMC2269272

Inspiration-promoting vagal reflex under NMDA receptor blockade in anaesthetized rabbits


  1. This study describes a novel vagal respiratory reflex in anaesthetized rabbits. In contrast to the well-known inspiratory (I) off-switching by vagal afferent excitation, this vagal reflex initiates and maintains the central I activity of phrenic nerve discharges in rabbits pre-treated with antagonists of N-methyl-D-aspartate-type excitatory amino acid receptors (NMDA-Rs).
  2. Under NMDA-R blockade with either dizocilpine (0·025-0·3 mg kg−1), D-2-amino-5-phosphonopentanoic acid (AP5, 0·5-1 mg, i.c.v.) or ketamine (10 mg kg−1), vagal stimulation at low frequencies (5-40 Hz) during the I phase prevented or markedly delayed the spontaneous I termination. In contrast, stimulation of the same vagal afferent at the same intensity but at a higher frequency (100-160 Hz) during the I phase immediately terminated the I phase.
  3. In non-vagotomized rabbits, maintaining the tidal volume at end-expiratory levels during the I phase prevented spontaneous I termination and maintained apneusis after NMDA-R blockade with dizocilpine.
  4. Brief stimulation of vagal afferents at low frequency (5-40 Hz) during the expiratory (E) phase constantly initiated phrenic I discharge after NMDA-R block.
  5. We conclude that low-frequency discharge of vagal pulmonary stretch receptor afferents, as when lung volume is near functional residual capacity, promotes central I activity under NMDA-R blockade.

The respiratory rhythm in vertebrates can be automatically generated in the central nervous system without afferent information (Bianchi et al. 1995). Afferent signals from various peripheral receptors, such as mechanoreceptors and chemoreceptors, modify this central respiratory pattern for the optimum control of ongoing respiratory movements (Feldman et al. 1990). Of these afferent signals, the mechanosensory information reflecting transpulmonary pressure in the bronchi is most important for determining the depth and rate of breathing (Clark & von Euler, 1972). This afferent information is detected by slowly adapting pulmonary stretch receptors (SA-PSRs) and sent to the brainstem via the vagus nerves. As first described by Hering & Breuer, this afferent information arising from SA-PSRs induces bi-modal responses of the central respiratory network to the lung volume: (1) inspiratory promotion with expiratory inhibition and (2) inspiratory suppression with expiratory promotion (reviewed in Widdicombe, 1961; Paintal, 1973; Comroe, 1974).

Electrophysiological studies have revealed that a single shock stimulation of the vagus nerve during the central inspiratory phase induces early termination of the inspiratory discharge of the phrenic nerve in anaesthetized cats (Cohen, 1975). The same response is also elicited by artificial inflation of the lung (Cohen, 1975). These results indicate that excitation of SA-PSR afferents induces the Hering-Breuer inspiratory suppression/expiratory promotion reflex (‘inflation reflex’). Recent studies with selective antagonists have clearly shown that N-methyl-D-aspartate-type ionotropic glutamate receptors (NMDA-Rs) appear not to be involved in this reflex pathway (Bonham, 1995; Denavit-Saubié & Foutz, 1996).

In contrast to this ‘inflation reflex’, the neural mechanisms underlying the Hering-Breuer deflation reflex, that is, the promotion of inspiratory activity at low transpulmonary pressures near functional residual capacity (FRC) during eupnoea (Widdicombe, 1961; Koller & Ferrer, 1970; Sant'Ambrogio, 1982) are still poorly understood. There has been no evidence that the electrical stimulation of vagal afferents could promote inspiration.

Here we describe for the first time an inspiration (I)-promoting effect, in contrast to the well studied I-terminating effect, of vagal afferent stimulation. This rather paradoxical effect was revealed by blocking NMDA-Rs with low-frequency afferent stimulation of the vagus nerve in vagotomized animals or when the lung was inflated to FRC in animals with intact vagi. Artificially ventilated rabbits were used because PSR afferents are more important for determining central respiratory rhythm in this species than in other mammals (Widdicombe, 1961).

We propose that the inspiration-promoting vagal reflex described in this study is a neural component of the Hering-Breuer deflation reflex. The present study also provides clear evidence that the same peripheral information arising from a single set of receptor groups can produce completely inverse responses from the central pattern generator depending on the frequency of the discharge and the activation patterns of neurotransmitter/receptor systems.


Surgical procedures

All experiments were performed in accordance with the Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences of the Physiological Society of Japan (1998). Adult Japanese white rabbits of either sex weighing 2·5-3·6 kg were anaesthetized with urethane (0·5 g kg−1, i.p., then 0·5 g kg−1, i.v.). After tracheal cannulation, the animals were immobilized with gallamine triethiodide (6 mg kg−1, i.v.; supplemental dose, 3 mg kg−1 h−1) and artificially ventilated with room air with a constant-volume animal ventilator (Shinano, SN-480-5, Tokyo, Japan). The tidal volume and ventilation rate were manually adjusted so that the end-tidal CO2 concentration as monitored with a gas analyser (NEC-San-ei: 1H31, Tokyo, Japan) was within 4·0-4·5 %. The vagus, depressor and sympathetic nerves were severed at the neck for experiments in which the vagus nerve was stimulated. These nerves were kept intact for experiments in which the lung no-inflation test was performed. The femoral vein and artery were cannulated for intravenous injection of drugs and for measurement of arterial blood pressure, respectively. For experiments in which the drugs were administered intracerebroventricularly (i.c.v.), the rabbit's head was fixed to a stereotaxic apparatus and the drug solution was injected into the lateral or third ventricle with a microsyringe (Kloehn, Whittier, CA, USA) fixed to a micromanipulator (Narishige, Tokyo, Japan). The injection site and drug distribution were confirmed by injecting the same volume of dye at the end of the experiment. The injection into the lateral or third ventricle ensured effective distribution of the NMDA-R antagonist to the possible sites of action from the rostral pons to the medulla, which are located downstream of the injection site. The rectal temperature was maintained at about 37°C with a thermostatically controlled heating pad (Nihon-Koden, AD-1100, Tokyo, Japan). Blood pressure, heart rate (in rabbits with intact vagi) and electroencephalograms (EEG) were recorded simultaneously to monitor the general state of the animal, especially the level of anaesthesia. Additional injections of urethane were never required as judged from these parameters over the course of the present experiments, which usually lasted 6-10 h. At the end of the experiments, the animals were killed by intravenous injection of an excessive dose (> 100 mg kg−1) of pentobarbitone.

Recording of neural activities and vagus nerve stimulation

The bilateral phrenic nerves were sectioned in the neck and the central cut end of a nerve was hooked on a bipolar platinum electrode for recording. The compound phrenic nerve discharge, EEG, arterial blood pressure, concentration of the expired CO2 and normalized pulses of the stimulation were recorded on digital audio tapes (Sony, PC108M, Tokyo, Japan).

Due to their lower threshold of excitation, A-α and -β afferent fibre groups in the vagus nerve, which ascend from PSRs, can be selectively activated by controlling the voltage of stimulation (Trenchard, 1977). Our preliminary assessment of conduction velocity by double-electrode stimulation/recording showed that a stimulus of ~0·5 V amplitude and 100 μs duration (Nihon-Kohden, SEN-7203) was suprathreshold in activating those fibres involved in the vagal I off-switch. With these stimulation conditions, no changes in blood pressure or EEG were detected (e.g. Fig. 4C), even after drug injection (see Results).

Figure 4
Effect of no-inflation test and low-frequency vagal stimulation on the phrenic nerve discharge

For experiments in which the effects of lung-volume holding were examined, after the ventilator circuit was closed at the end-expiration level, a known volume of room air was injected directly into the lung via the tracheal catheter with a 10 or 50 ml syringe. Passive ventilatory movements of the chest were recorded with a tension transducer attached to the chest wall of the rabbit in the supine position.


Dizocilpine (0·01-0·3 mg kg−1), ketamine (5-10 mg kg−1, Research Biochemicals International) and pentobarbitone (4-8 mg kg−1, Nembutal, Abbott) were injected intravenously. 2-Amino-5-phosphonopentanoic acid (AP5, Research Biochemicals International) and 6-nitro-7-sulphamoylbenzo-(f)-quinoxaline-2,3-dione (NBQX, Tocris) were injected into the lateral or third ventricle of the animals whose skulls had been fixed in a stereotaxic apparatus in the prone position. Before injection, AP5 and NBQX were dissolved in saline to concentrations of 10 g l−1 and 5 g l−1, respectively. The pH of the NBQX solution was adjusted to 8·5-9 with NaHCO3. A final volume of 10-100 μl was injected slowly with a microsyringe to obtain the final quantity (0·25-0·5 mg for NBQX and 0·5-1 mg for AP5).

Data analysis

The phrenic nerve discharge was sampled at 1000 Hz after low-cut filtration at 200 or 500 Hz for further analyses, including the measurement of inspiratory time (TI), expiratory time (TE) and total respiratory duration (TTOT) using the MacLab interface (AD Instruments, Castle Hill, Australia). The TI/TTOT ratio was estimated as follows. First, the phrenic nerve discharge was sampled for 15 s before stimulation. Then the total duration of all I phases appearing within this 15 s time window was divided by 15 s and expressed as a percentage of 15 s. Then, to evaluate the effects of vagal stimulation, another 15 s sampling was made from the beginning of the stimulation which always lasted more than 15 s. Similarly, the total duration of all I phases within this window was divided by 15 s. These values are extremely close to the TI/TTOT calculated with the conventional method. We used this method for the following reasons: (1) when inspiration was extremely prolonged and no E phase appeared during the stimulation period (Fig. 1A and B), the TTOT could not be measured; and (2) when the I termination was completely suppressed (Fig. 1A and B), TI was dependent on the duration of the sustained stimulation period, which was selected arbitrarily with a minimal value always exceeding 15 s. The fixed window was thus given a value of 15 s, a time frame in which all monitored biological activities were stable. With this calculation, TI/TTOT is 100 % when apneusis lasts 15 s (or more) and TI/TTOT is 0 % when there is no I activity.

Figure 1
Effect of dizocilpine on changes in phrenic nerve discharge induced by vagal stimulation

The values are expressed as mean values ±s.d. Differences were compared using Student's two-tailed paired t tests for values with Gaussian distribution and using the Mann-Whitney U test for non-parametric values. Differences with a probability (P) less than 0·05 were considered significant.


Frequency dependence of the respiratory response to vagal stimulation before and after dizocilpine injection

Before dizocilpine

The central end of a vagus nerve was stimulated at frequencies of 0·5-160 Hz to observe effects on spontaneous rhythmic discharge of the phrenic nerve in five vagotomized animals (Fig. 1A, left). The response of the central rhythm to vagal stimulation was frequency dependent and was classified into three ranges. (1) Stimulation at extremely low frequencies (0·5-1 Hz) induced no detectable changes in the phrenic nerve discharge (data not shown). (2) Stimulation at low frequencies (2·5-40 Hz) consistently accelerated the central respiratory rhythm (Fig. 1A, left). Stimulation at 40 Hz significantly shortened both TI and TE, leading to a significant decrease in TTOT (Table 1). (3) Stimulation at higher frequencies (100-160 Hz) significantly shortened or suppressed inspiratory bursts (Fig. 1A, left) and markedly prolonged the E phase to several seconds (Table 1).

Table 1
Effect of vagal stimulation on respiratory phase duration

After dizocilpine

The same protocol of vagus stimulation was used after injection of dizocilpine (0·025-0·3 mg kg−1, i.v.) into the femoral vein. The effects of dizocilpine were confirmed by the appearance of a longer I phase (i.e. apneustic breathing), similar to that observed in other mammalian species in vivo (Denavit-Saubié & Foutz, 1996). Ten minutes after injection of dizocilpine (0·1-0·3 mg kg−1), the TI increased significantly to 3094 ± 1263 % of control (n= 6). However, TE was unchanged after dizocilpine injection (at 10 min after injection, 90·5 ± 16·7 %; not significantly different from before injection, n= 6).

In contrast to that before dizocilpine injection, low-frequency (5-40 Hz) stimulation of the vagus nerve after injection further prolonged the I phase (Fig. 1A, right). With large doses (≥ 0·1 mg kg−1) of dizocilpine, sustained low-frequency stimulation completely suppressed I termination (Fig. 1A right, 10-40 Hz; Fig. 1B 2).2). During the first 30 min, when the drug effect was greatest, the central spontaneous I termination was completely prevented for more than 30 s if low-frequency stimulation was continued (Fig. 1B). The recovery of this effect paralleled that of the apneustic breathing effect (Fig. 1B, bottom). After injection of a small dose of dizocilpine (0·025-0·05 mg kg−1), low-frequency stimulation significantly prolonged TI but a few short pauses of I burst appeared during sustained stimulation of longer than 15 s.

Figure 2
Time course of effect of dizocilpine on spontaneous TI (above) and TI/TTOT (below) during sustained vagal stimulation at 20 Hz

To evaluate changes in I duration with vagal stimulation, we calculated TI/TTOT (Fig. 1C). Before dizocilpine administration, low-frequency vagal stimulation increased the TI/TTOT only slightly. This small increase in TI/TTOT reflects the increase in respiratory frequency due to shortening of the I and E phases to a similar extent during stimulation (upper traces in Fig. 1A, left). In contrast, vagal stimulation markedly reduced TI/TTOT at frequencies higher than 100 Hz, indicating TTOT lengthening due to E prolongation.

These differences in response to different stimulation frequencies were much greater after dizocilpine administration. In addition to a significant rise in TI/TTOT resulting from apneustic breathing after dizocilpine without stimulation, low-frequency stimulation further increased the TI/TTOT to near 100 % (i.e. complete prevention of I termination). In contrast, the effect of high-frequency stimulation was largely unchanged regardless of whether dizocilpine was injected. As this insensitivity of the I off-switching effects of high-frequency vagal stimulation to dizocilpine has also been described in other mammalian species (see Denavit-Saubié & Foutz, 1996), we studied the effects of low-frequency stimulation in further detail.

To compare the sensitivity to NMDA-R blockade of the spontaneous I prolongation (apneustic breathing) and vagally induced I lengthening, we checked the difference in time course of the effects of dizocilpine on TI and TI/TTOT. Figure 2 shows the time course of the recovery of these variables. Even at 120 min after injection of dizocilpine (0·3 mg kg−1), low-frequency vagal stimulation still effectively lengthened I phase (Fig. 2, lower panel) despite apparent recovery of TI from extreme prolongation observed until 60 min after administration (Fig. 2, upper panel). It is concluded that expression of I lengthening with low-frequency vagal stimulation under NMDA-R blockade does not necessarily depend on the expression of extremely prolonged apneustic I discharge.

To confirm that the threshold of PSR afferent fibres was not affected by dizocilpine, we tested different stimulation intensities (0·1-5·0 V) before and after dizocilpine administration. Stimulation intensities of less than 0·1 V were ineffective at any frequency and in any situation. Responses as in Fig. 1A were consistently elicited by intensities of 0·2-5·0 V. Stimulation intensities higher than 3 V induced transient and slight changes in blood pressure. Even with stimulation at this intensity, however, the effect on phrenic nerve discharges was unchanged. These intensity ranges and responses were unchanged even after 0·1-0·3 mg kg−1 of dizocilpine had been injected. Therefore, we concluded that the stimulation intensity of 0·5 V used in the present study is appropriate for eliciting consistent responses before and after administration of NMDA-R antagonists. We concluded that excitation threshold of these fibre groups was relatively insensitive to NMDA-R blockade.

The effect of low-frequency stimulation and excitatory amino acid receptors

Other NMDA-R antagonists

To confirm whether the I-lengthening effect of low-frequency vagal stimulation can be consistently observed under NMDA-R blockade or is simply an unknown secondary effect of dizocilpine, we tested other NMDA-R blockers. Intravenous injection (10 mg kg−1) of ketamine, another non-competitive NMDA-gated channel blocker, had very similar apneustic effects on the spontaneous respiratory rhythm to those observed after dizocilpine administration. The I lengthening was also similarly elicited by low-frequency (10-40 Hz) vagal stimulation after ketamine (data not shown). AP5, a competitive NMDA-R antagonist at glutamate-binding sites, injected i.c.v. induced apneustic I discharges (Fig. 3A right, control) similarly to dizocilpine. This effect lasted for 30-60 min. When this effect was observed, sustained stimulation of the vagal afferent at low frequencies of 5-40 Hz delayed (5 Hz in Fig. 3A, right) or prevented (10-40 Hz) the rhythmic I termination. This extreme prolongation of the I phase resulted in a clear increase in TI/TTOT (Fig. 3B) similar to that observed after dizocilpine (Fig. 1C). These results indicate that the delayed I termination with low-frequency stimulation is observed under NMDA-R blockade with any pharmacological agent.

Figure 3
Effect of AP5 on changes in phrenic nerve discharges induced by vagal stimulation

Low-frequency vagal stimulation under pharmacological respiratory depression

The NMDA-R blockade was always accompanied by prolonged apneustic I discharges. To examine whether this I-lengthening vagal reflex is related to pharmacological prolongation of spontaneous I phases, the vagus nerve was stimulated at various frequencies during pharmacological slowing of the respiratory rhythm with a non-NMDA receptor blockade and with a barbiturate. Administration of NBQX (0·5 mg, i.c.v.), a competitive antagonist of non-NMDA-type excitatory amino acid receptors, significantly increased TI (before, 0·48 ± 0·09 s; after, 0·79 ± 0·21 s, n= 4). Vagal stimulation at 20 Hz significantly shortened, but never prolonged the I phase before (to 0·37 ± 0·07 s) as well as after (to 0·67 ± 0·13 s) NBQX injection. The effect of 160 Hz stimulation on TI was not significant either before or after NBQX administration (before, 0·38 ± 0·08 s during stimulation; after, 0·74 ± 0·16 s during stimulation; Mann-Whitney U test).

After i.v. injection of pentobarbitone (4-8 mg kg−1), despite manifest prolongation of the I phase (TI before pentobarbitone, 0·59 ± 0·08 s; after, 0·96 ± 0·23 s; n= 5), vagal stimulation failed to lengthen the I phase at any frequency from 5 to 160 Hz. Vagal stimulation at 20 Hz significantly shortened the I phase before (to 0·39 ± 0·07 s) as well as after (to 0·59 ± 0·10 s) pentobarbitone administration. Interestingly, the shortening of TI by high-frequency stimulation was greater after pentobarbitone (TI during 160 Hz stimulation; before pentobarbitone, 0·41 ± 0·05 s; after, 0·18 ± 0·24 s). Thus, we conclude that sustained I activity during low-frequency stimulation is closely related to the elimination of NMDA-R-mediated synaptic transmission in the central respiratory networks.

Effect of physiological reduction of lung volume on the central I phase under dizocilpine

As the PSR encodes the transpulmonary pressure to the discharge frequency of vagal afferents, low-frequency stimulation of PSR afferents may well convey physiologically relevant information. To determine whether the central I responses to vagal stimulation after NMDA-R blockade can be reproduced by a physiologically induced change in PSR excitation levels, we performed experiments in which lung volume was held isobarically. Figure 4 shows typical results of lung-volume holding and electrical vagal stimulation performed sequentially in a single rabbit. Figure 4A shows the effect of ‘no-inflation’ of the lung (holding lung volume at FRC) after dizocilpine administration (0·3 mg kg−1) in a rabbit with intact vagi. During the no-inflation test, phrenic I activity was sustained in a manner extremely similar to that observed with low-frequency vagal stimulation (inflation (-) in Fig. 4A). This effect lasted until holding was terminated. Soon after this test in the same animal, we severed the vagal nerves bilaterally in the neck. Vagotomy successfully eliminated PSR afferents because (1) the phrenic nerve respiratory burst was no longer synchronized with the ventilatory cycle (Fig. 4B 11--3),3), and ((2)2) the no-inflation test had no effect on phrenic discharges (Fig. 4B, right). The latter finding indicates that the I-lengthening effect of lung-volume holding (Fig. 4A) is mediated by vagal afferents. Ten minutes after vagotomy, vagus nerve stimulation at 20 Hz prevented spontaneous I termination (Fig. 4C) in a manner very similar to that observed with the no-inflation test before vagotomy (Fig. 4A).

To study the effects of lung volume changes on central I activity, a constant volume was injected into the lung immediately following the no-inflation test in the animals with intact vagi after dizocilpine administration (Fig. 5). The no-inflation test consistently prolonged the I phase (arrows in Fig. 5a-g); however, injection of various volumes into the lung produced spontaneous cyclic bursts whose rhythm and pattern were dependent on the volume injected (Fig. 5a-g). Isobaric holding of the lung at a volume at least 20 ml greater than FRC prolonged TE as with stimulation of the vagus nerve at frequencies of 100 Hz or higher in vagotomized animals (compare the bottom traces in Fig. 1A right and the right part of the traces in Fig. 5f and g). These responses were consistently observed in all four animals to which the protocol was applied. These observations indicate that sustained electrical stimulation of the vagus nerve at various frequencies gives relevant physiological information to the central respiratory rhythm generator that the lung volume is set at various levels.

Figure 5
Effect of lung-volume holding on the phrenic nerve discharge in a rabbit with bilateral intact vagi after dizocilpine administration (0·3 mg kg−1, i.v.)

I on-switch by low-frequency vagal stimulation during the E phase

As low-frequency vagal stimulation produced a central response similar to that observed when the lung volume is held near FRC, we expected that low-frequency stimulation would induce an E-to-I phase transition as in the Hering-Breuer deflation reflex. We applied a very short train of vagal stimulation during the E phase in dizocilpine-treated rabbits with the same stimulation parameters. A train of 10-12 pulses delivered at an interval of 25 ms (i.e. 40 Hz) during the E phase immediately induced I activity in the phrenic nerve discharge regardless of the timing within the E phase (Fig. 6A6A22--4).4). Surprisingly, a single pulse or a brief train of two to three pulses at intervals of 10-50 ms was sufficient to induce E-to-I phase switching (Fig. 6B). Figure 7 demonstrates that vagal stimulation at 20 Hz can initiate and sustain I activity when delivered during the E and I phases, respectively. These effects were reproduced consistently in the five animals tested. This vagally induced effect is hereafter referred to as the I on-switching vagal reflex.

Figure 6
Effect of low-frequency vagal stimulation delivered at the E phase
Figure 7
Effect of low-frequency vagal stimulation delivered at the E and I phases after injection of dizocilpine (0·2 mg kg−1)

To examine the contribution of NMDA-R blockade to this I on-switching vagal reflex, we compared the responses of the phrenic nerve discharge to low-frequency vagal stimulation delivered at the mid-E phase before and after dizocilpine administration in three animals. Even before dizocilpine administration, a train of two pulses with an interval of 50 ms (i.e. 20 Hz) could initiate I phase with a latency of about 150-200 ms (Fig. 8A, above). About 20 ms after each pulse, a short temporal burst consistently appeared in the phrenic nerve discharge (two arrows in Fig. 8A), but the I burst started after a pause of about 100 ms in length (asterisk in Fig. 8A) which followed the short phrenic burst. The duration of this 100 ms pause substantially varied among trials in each animal, making it difficult to ascertain the advance of I onset with vagal stimulation in about 25 % of trials. In contrast, dizocilpine abolished this 100 ms phrenic pause and advanced the onset of stimulation-induced I discharge by about 100 ms (Fig. 8A, bottom). In all three animals, the same stimulation with longer trains delivered during the I phase prolonged the I discharge.

Figure 8
Effect of dizocilpine on brief low-frequency stimulation of the vagus nerve at the mid-expiratory phase

The latency from the first pulse to the I onset was measured from digitally integrated phrenic nerve discharge (horizontal lines with bi-directional arrows in the left panel of Fig. 8B). In all three animals with which the same stimulation procedure was tested, dizocilpine markedly shortened the latency of stimulation-induced I activity; the I discharge started within about 50 ms of either of two pulses in all respiratory cycles after dizocilpine, whereas the I discharge started never earlier than 100 ms after the first stimulus before dizocilpine injection (Fig. 8B, right). These results indicate that the I on-switch by low-frequency vagal stimulation at mid-E phase is facilitated by blockade of NMDA-Rs.


A novel type of reflex mediated by slowly adapting pulmonary stretch receptor afferents

Here we describe a novel vagally evoked respiratory reflex. The characteristics of this reflex are: (1) the response is the promotion of I activity, i.e. on-switching and maintenance of the I phase; (2) it is elicited by the same vagal afferents involved in the I-terminating reflex by lung inflation arising from PSRs; (3) it is activated only with low-frequency afferent discharges corresponding to the transpulmonary pressure near FRC; and (4) it is apparent only when the NMDA-R is inactivated. Mechanisms underlying these characteristics and their functional significance are discussed below.

Functional implications of the I-lengthening vagal reflex

In the present study, under NMDA-R blockade I termination was prevented only with low-frequency stimulation of the vagus nerve in vagotomized rabbits or when tidal volume was held at FRC in non-vagotomized rabbits. The I discharges in phrenic nerve activity during sustained low-frequency stimulation cannot be stimulation artefacts or recurring evoked potentials because similar prolonged phrenic discharges were produced during the no-inflation test (Fig. 4). We believe that this reflex is mediated by identical afferent fibre groups, possibly those from SA-PSRs, involved in the I termination by high-frequency stimulation and lung inflation to the end-I level, because a simple change in stimulation frequency (Figs 1, ,22 and and3)3) or in tidal volume (Fig. 5) was sufficient to evoke these bi-directional responses. As lowering the stimulation frequency is unlikely to newly activate another group of fibres of higher threshold, this I-lengthening reflex may be mediated by the same fibres involved in the vagal I off-switch or by a subset of SA-PSR afferent fibres with a lower threshold. The sustained I discharge during lung-volume holding at FRC argues that this I-lengthening effect is mediated mainly by SA-PSRs. The I on-switch, elicited by a brief vagal stimulation of the same intensity during E phase, may involve excitation of afferent fibres from both or either of SA-PSRs and rapidly adapting PSRs (RA-PSRs) both of which are activated by stimulus intensities similar to those used in the present study (Berger & Dick, 1987).

As the discharge frequency of SA-PSR afferents changes linearly in response to changes in transpulmonary pressure (Davenport et al. 1981), low-frequency stimulation of this afferent can resemble a message from PSRs to the brainstem respiratory network that the lung has deflated. Interestingly, the I-lengthening effect was greatest at a frequency of 20 Hz (Fig. 2), at which the TI was greater than with no stimulation. Trenchard (1977) has electrically stimulated the vagus nerve of rabbits while continuously varying its frequency throughout the respiratory cycle and estimated that the discharge frequency at end-expiration was 5-40 Hz in rabbits. Davenport et al. (1981) measured the PSR unit discharge frequency in anaesthetized cats and found two types of units: a phasic type that does not discharge during expiratory phase, and a tonic type that shows a frequency-modulated variation throughout the respiratory cycle and continues to discharge at 10-40 Hz when the lung volume is equal to FRC. This type of ‘tonic’ PSR unit comprises 44 % of all PSR fibres in rabbits (Sant'Ambrogio, 1982). Therefore, electrical stimulation of vagus afferents at 10-40 Hz may send a signal falsely informing the brainstem respiratory network that lung volume is held at end-expiratory FRC. Indeed, an effect similar to that of the 20 Hz vagal stimulation preventing spontaneous I termination was observed in animals with intact vagi during the no-inflation test (Figs 4 and and5).5). In this context, we argue that the extremely prolonged apneustic I activity during the no-inflation test in animals with NMDA-Rs inactivated and intact vagi described in previous studies (for example, see Fig. 2 in Foutz et al. 1989; Figs 1 and and22 in Foutz & Denavit-Saubié, 1989; Figs 3 and and66 in Feldman et al. 1992; and Fig. 1 in Ling et al. 1994) was not due simply to the effect of NMDA-R blockade but was further intensified by the I-lengthening vagal reflex described here. In early inspiration, when the PSR discharge frequency is still low (Paintal, 1973), this reflex, by suppressing central I off-switch mechanisms, might activate the inspiratory pattern generator and help it to maintain its activity until the lung is sufficiently inflated (Comroe, 1974).

Mechanism of frequency dependency of the I-lengthening vagal reflex

The frequency specificity of the responses observed in the present study is not surprising as the effect of vagal stimulation on the central respiratory rhythm has already been shown to depend both quantitatively and qualitatively on the frequency (Wyss, 1946). However, in most subsequent studies, the vagus nerve has been stimulated at relatively high frequencies, such as 100 Hz (Iscoe et al. 1979; Karius et al. 1994) to study its I-terminating effects. The effect of low-frequency stimulation has been paid little attention. This is, we believe, the main reason that this I-lengthening reflex under NMDA-R blockade has not been described previously.

The frequency response to high-frequency stimulation (100-160 Hz) was not affected by any of the NMDA-R blockers (Figs 1--3),3), a result consistent with previous reports of the insensitivity of I-terminating and E-lengthening vagal reflexes to NMDA-R blockade in rats or cats (Iscoe et al. 1979; Feldman et al. 1992; Bonham et al. 1993; Pierrefiche et al. 1994; however, see the effect of AP5 in Karius et al. 1994).

Although the present results cannot elucidate the mechanism of this frequency-dependent switching of respiratory responses, one of possible sites where this ‘frequency filtering’ probably occurs is the neuronal network within the nucleus tractus solitarii (NTS), where primary afferent fibres from PSRs terminate (Donoghue et al. 1982; Berger & Averill, 1983; Pantaleo & Corda, 1986; Bonham et al. 1993). It is also the nucleus in which the information from PSRs is integrated and sent to other nuclei (Champagnat et al. 1986; Ezure & Tanaka, 1996). Miles (1986) stimulated the solitary tract fibres at various frequencies and demonstrated a frequency-dependent depression of the postsynaptic potential (PSP) of second-order neurones in guinea-pig NTS slice (Miles, 1986). However, despite lack of direct evidence, other respiratory networks downstream to NTS, such as the pontine respiratory group where the pneumotaxic centre is located (Bianchi et al. 1995), can also underlie the frequency-dependent discrimination of afferent information, as this type of frequency-dependent change in central network responses has also been reported in many other systems (see Discussion of Miles, 1986).

Stimulation of the vagus afferent provokes a biphasic response, excitation followed by depression, of the phrenic nerve discharge in anaesthetized and decerebrate cats (Iscoe et al. 1979). Haji et al. (1996) described a short-latency excitatory PSP followed by a transient inhibitory PSP in I neurones evoked by vagus nerve stimulation during the I phase in decerebrate cats. A recent report from the same group demonstrated that a long-latency brief phrenic burst appeared following the transient inhibitory PSP of I neurones after vagal stimulation at mid-E phase (Pierrefiche et al. 1998), which is consistent with the present observation. They speculated that this short phase of brief phrenic burst, which was little affected by NMDA-R blockade, represents a higher-order network-generated event, not a simple direct response to the peripheral input. This idea supports the view that low-frequency vagal stimulation during the I phase raises and maintains the activity of central I generator through these higher-order events and blocks inhibitory perturbation from central intrinsic off-switching mechanisms that has been weakened or suppressed by NMDA-R blockade (Pierrefiche et al. 1992). As a whole, these multiphasic responses to vagal afferents of the I generators may serve as a frequency discriminator by accumulating either excitatory or inhibitory responses depending on the afferent discharge intervals and the latency of each response.

Mechanism of excitatory amino acid receptor-type dependence of the I-lengthening vagal reflex

An important aspect of the I-promoting vagal reflex in the present study is that it was apparent only when the NMDA-R was blocked pharmacologically. The principle excitatory amino acid released from primary afferent terminals of SA-PSRs is glutamate (Meeley et al. 1989), and the receptor primarily involved in the inflation reflex of the second-order NTS neurone is a non-NMDA-type receptor (Bonham et al. 1993). Recently, a slowly activating and inactivating NMDA-R current together with a faster AMPA-type receptor (AMPA-R) current have been shown in response to the solitary tract stimulation in NTS neurones in brainstem slices (Aylwin et al. 1997). This finding suggests that both non-NMDA-Rs and NMDA-Rs expressed in the same cell are involved in the transmission from PSR afferents to second-order NTS cells (Miller & Felder, 1988).

An important question is why this reflex is manifested only under NMDA-R blockade. We propose two different, not necessarily mutually exclusive, interpretations.

(1) The NMDA-R blockade might have unmasked and intensified AMPA-R-mediated responses of the respiratory network which had been obscured in anaesthetized and paralysed animals as in the present study. Among possible mechanisms, the depression of AMPA-R current by inactivation of metabotropic glutamate receptors due to reduced glutamate release (Glaum & Miller, 1993) or by long-term depression through calcium-dependent desensitization of AMPA-Rs in NTS (Zhou et al. 1997) deserves to be examined in vivo. The priority or efficacy balance between the NMDA-R and AMPA-R systems may change according to on-going situations, such as incoming synaptic inputs. For example, the effect of NMDA-R activation is weaker when it occurs during the silent phase of respiratory neurones (Pierrefiche et al. 1991, 1994).

(2) Another possibility is that the mechanism of the I off-switch is completely different in NMDA-R-intact and NMDA-R-blocked animals (Morin-Surun et al. 1995), leading to a different response to the low-frequency PSR inputs at the I termination. In vagotomized animals, the I termination is facilitated mainly by the pontine respiratory neurone group (PRG) located in the nucleus parabrachialis medialis and Kölliker-Fuse nucleus. Systemic blockade of NMDA-Rs (Foutz et al. 1988a, b, 1989) or microinjection of NMDA antagonists to the PRG (Ling et al. 1994) in vagotomized animals prolongs the I phase (apneustic breathing). The mechanism underlying the spontaneous termination of this apneustic ‘reversible’ I phase (Baker & Remmers, 1980) after NMDA-R blockade in animals without vagal PSR input is still unclear. In an in vitro preparation of isolated brainstem of adult guinea-pigs, intra-arterial perfusion of dizocilpine did not prolong the I phase (Morin-Surun et al. 1995). Also, in a similar isolated brainstem preparation of kittens, TI remained unchanged after complete ablation of the pons (Kato et al. 1996). These findings indicate that another mechanism in the medulla can terminate the I phase in the absence of both NMDA-R systems and PRG. Our present hypothesis is that PSR afferents discharging at low frequency suppressed this mechanism under the present experimental conditions. Whether this I-promoting reflex functions in animals with intact NMDA-R systems should be examined in future studies.

I on-switching vagal reflex facilitated by NMDA-R blockade

The present study indicates that afferent signals arising from the same fibre groups involved in the I-lengthening vagal reflex can also initiate I activity when activated during the E phase. The decrease in TE by low-frequency vagal stimulation and by lung-volume holding to a low level before and after dizocilpine administration may be partly due to this I on-switching vagal reflex. The new finding is that this vagally induced I on-switching is facilitated by NMDA-R blockade in anaesthetized rabbits. The I on-switch by afferent fibre stimulation during the E phase has been described in decerebrate cats. Pierrefiche et al. (1998) reported that stimulation of the vagus nerve during mid-E phase initiated I activity that occurred after the three-phased response in PSPs and phrenic discharge, each of which was dizocilpine resistant. They speculated that the third phase response characterized by a brief phrenic burst is an ‘abortive I on-switch’: however, they did not measure latency from this third phase (or from the stimulus) to the resumption of the I burst comparing before and after dizocilpine injection. In the present study, the latency from stimulation to I onset became markedly shorter and more stable after dizocilpine administration; the stimulation-evoked brief phrenic burst seemed to develop directly to the I burst without any interruption after dizocilpine administration (Fig. 8). One of the possible interpretations of the present result is that the neural process from this ‘abortive on-switch’ to establishment of the central I phase is facilitated by NMDA-R blockade. This neural process bridging over peripherally evoked short-term responses and the central global I on-switch should be operating within the higher-order network of the respiratory centre which integrates information mediated by single or combined activation of various afferent fibres including afferents other than the vagus nerve (Oku & Dick, 1992; Oku et al. 1993; Pierrefiche et al. 1998) and generates various short-term motor responses, such as in the cranial efferent (Hayashi & McCrimmon, 1996).


The I-promoting vagal reflex described in the present study is the first evidence that peripheral information arising from a single group of receptor cells can produce completely inverse responses to the central pattern generator depending on the frequency of the afferent fibre discharge under certain conditions. This type of reflex may contribute to the precise and complex real-time adaptation of the central activity to the incoming peripheral information. On the basis of the present findings, we propose that results of the conventional ‘no-inflation’ test in studies of the central respiratory network in animals with NMDA-R blockade should be carefully interpreted as the response may be simply due to absent vagal afferents or to active information from PSRs that the lung volume is at FRC.


The authors are grateful to Takako Matsuo for her technical assistance during experiments and preparation of the manuscript. Continuous encouragement by Professor Masahiro Kawamura was invaluable. This work was initially inspired by the late Professor Takehiko Hukuhara who died in 1993 after many years of devotion to the development of respiration physiology in Japan. The authors wish to dedicate this article to his memory.


  • Aylwin ML, Horowitz JM, Bonham AC. NMDA receptors contribute to primary visceral afferent transmission in the nucleus of the solitary tract. Journal of Neurophysiology. 1997;77:2539–2548. [PubMed]
  • Baker JP, Remmers JE., Jr Temporal correlation of graded reversible inspiratory inhibition with discharge patterns of late inspiratory neurons located in the dorsal respiratory group in cats. Brain Research. 1980;200:331–340. [PubMed]
  • Berger AJ, Averill DB. Projection of single pulmonary stretch receptors to solitary tract region. Journal of Neurophysiology. 1983;49:819–830. [PubMed]
  • Berger AJ, Dick TE. Connectivity of slowly adapting pulmonary stretch receptors with dorsal medullary respiratory neurons. Journal of Neurophysiology. 1987;58:1259–1274. [PubMed]
  • Bianchi AL, Denavit-Saubié M, Champagnat J. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters. Physiological Reviews. 1995;75:1–45. [PubMed]
  • Bonham AC. Neurotransmitters in the CNS control of breathing. Respiration Physiology. 1995;101:219–230. [PubMed]
  • Bonham AC, Coles SK, McCrimmon DR. Pulmonary stretch receptor afferents activate excitatory amino acid receptors in the nucleus tractus solitarii in rats. The Journal of Physiology. 1993;464:725–745. [PMC free article] [PubMed]
  • Champagnat J, Denavit-Saubié M, Grant K, Shen KF. Organization of synaptic transmission in the mammalian solitary complex, studied in vitro. The Journal of Physiology. 1986;381:551–573. [PMC free article] [PubMed]
  • Clark FJ, Euler C, von On the regulation of depth and rate of breathing. The Journal of Physiology. 1972;222:267–295. [PMC free article] [PubMed]
  • Cohen MI. Phrenic and recurrent laryngeal discharge patterns and the Hering-Breuer reflex. American Journal of Physiology. 1975;228:1489–1496. [PubMed]
  • Comroe JH. Physiology of Respiration. Chicago: Year Book Medical Publishers; 1974. Reflexes from the lungs; pp. 72–82.
  • Davenport PW, Frazier DT, Zechman FW., Jr The effect of the resistive loading of inspiration and expiration on pulmonary stretch receptor discharge. Respiration Physiology. 1981;43:299–314. 10.1016/0034-5687(81)90111-0. [PubMed]
  • Denavit-Saubié M, Foutz AS. Neuropharmacology of respiration. In: Miller AD, Bianchi AL, Bishop BP, editors. Neural Control of the Respiratory Muscles. New York: CRC Press, Boca Raton; 1996. pp. 143–157.
  • Donoghue S, Garcia M, Jordan D, Spyer KM. The brain-stem projections of pulmonary stretch afferent neurones in cats and rabbits. The Journal of Physiology. 1982;322:353–363. [PMC free article] [PubMed]
  • Ezure K, Tanaka I. Pump neurons of the nucleus of the solitary tract project widely to the medulla. Neuroscience Letters. 1996;215:123–126. 10.1016/S0304-3940(96)12968-2. [PubMed]
  • Feldman JL, Smith JC, Ellenberger HH, Connelly CA, Liu G, Greer JJ, Lindsay AD, Otto MR. Neurogenesis of respiratory rhythm and pattern: emerging concepts. American Journal of Physiology. 1990;259:R879–886. [PubMed]
  • Feldman JL, Windhorst U, Anders K, Richter DW. Synaptic interaction between medullary respiratory neurones during apneusis induced by NMDA-receptor blockade in cat. The Journal of Physiology. 1992;450:303–323. [PMC free article] [PubMed]
  • Foutz AS, Champagnat J, Denavit-Saubié M. N-Methyl-D-aspartate (NMDA) receptors control respiratory off-switch in cat. Neuroscience Letters. 1988a;87:221–226. 10.1016/0304-3940(88)90452-1. [PubMed]
  • Foutz AS, Champagnat J, Denavit-Saubié M. Respiratory effects of the N-methyl-D-aspartate (NMDA) antagonist, MK-801, in intact and vagotomized chronic cats. European Journal of Pharmacology. 1988b;154:179–184. 10.1016/0014-2999(88)90095-7. [PubMed]
  • Foutz AS, Champagnat J, Denavit-Saubié M. Involvement of N-methyl-D-aspartate (NMDA) receptors in respiratory rhythmogenesis. Brain Research. 1989;500:199–208. 10.1016/0006-8993(89)90314-4. [PubMed]
  • Foutz AS, Denavit-Saubié M. Differentiation of phencyclidine and sigma receptor types affecting the central inspiratory termination mechanism in cat. Life Sciences. 1989;45:1285–1292. 10.1016/0024-3205(89)90131-8. [PubMed]
  • Glaum SR, Miller RJ. Activation of metabotropic glutamate receptors produces reciprocal regulation of ionotropic glutamate and GABA responses in the nucleus of the tractus solitarius of the rat. Journal of Neuroscience. 1993;13:1636–1641. [PubMed]
  • Haji A, Pierrefiche O, Foutz AS, Champagnat J, Denavit-Saubié M, Takeda R. Pharmacological properties of peripherally induced postsynaptic potentials in bulbar respiratory neurons of decerebrate cats. Neuroscience Letters. 1996;211:17–20. 10.1016/0304-3940(96)12708-7. [PubMed]
  • Hayashi F, McCrimmon DR. Respiratory motor responses to cranial nerve afferent stimulation in rats. American Journal of Physiology. 1996;271:R1054–1062. [PubMed]
  • Iscoe S, Feldman JL, Cohen MI. Properties of inspiratory termination by superior laryngeal and vagal stimulation. Respiration Physiology. 1979;36:353–366. 10.1016/0034-5687(79)90047-1. [PubMed]
  • Karius DR, Ling L, Speck DF. Nucleus tractus solitarius and excitatory amino acids in afferent-evoked inspiratory termination. Journal of Applied Physiology. 1994;76:1293–1301. [PubMed]
  • Kato F, Morin-Surun M-P, Denavit-Saubié M. Coherent inspiratory oscillation of cranial nerve discharges in perfused neonatal cat brainstem in vitro. The Journal of Physiology. 1996;497:539–549. [PMC free article] [PubMed]
  • Koller EA, Ferrer P. Studies on the role of the lung deflation reflex. Respiration Physiology. 1970;10:172–183. 10.1016/0034-5687(70)90081-2. [PubMed]
  • Ling L, Karius DR, Speck DF. Role of N-methyl-D-aspartate receptors in the pontine pneumotaxic mechanism in the cat. Journal of Applied Physiology. 1994;76:1138–1143. [PubMed]
  • Meeley MP, Underwood MD, Talman WT, Reis DJ. Content and in vitro release of endogenous amino acids in the area of the nucleus of the solitary tract of the rat. Journal of Neurochemistry. 1989;53:1807–1817. [PubMed]
  • Miles R. Frequency dependence of synaptic transmission in nucleus of the solitary tract in vitro. Journal of Neurophysiology. 1986;55:1076–1090. [PubMed]
  • Miller BD, Felder RB. Excitatory amino acid receptors intrinsic to synaptic transmission in nucleus tractus solitarii. Brain Research. 1988;456:333–343. 10.1016/0006-8993(88)90236-3. [PubMed]
  • Morin-Surun M-P, Boudinot E, Kato F, Foutz AS, Denavit-Saubié M. Involvement of NMDA receptors in the respiratory phase transition is different in the adult guinea pig in vivo and in the isolated brainstem preparation. Journal of Neurophysiology. 1995;74:770–778. [PubMed]
  • Oku Y, Dick TE. Phase resetting of the respiratory cycle before and after unilateral pontine lesion in cat. Journal of Applied Physiology. 1992;72:721–730. [PubMed]
  • Oku Y, Dick TE, Cherniack NS. Phase-dependent dynamic responses of respiratory motor activities following perturbation of the cycle in the cat. The Journal of Physiology. 1993;461:321–337. [PMC free article] [PubMed]
  • Paintal AS. Vagal sensory receptors and their reflex effects. Physiological Reviews. 1973;53:159–227. [PubMed]
  • Pantaleo T, Corda M. Respiration-related neurons in the medial nuclear complex of the solitary tract of the cat. Respiration Physiology. 1986;64:135–148. 10.1016/0034-5687(86)90037-X. [PubMed]
  • The Physiological Society of Japan. Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences. Japanese Journal of Physiology. 1998;48:219.
  • Pierrefiche O, Foutz AS, Champagnat J, Denavit-Saubié M. The bulbar network of respiratory neurons during apneusis induced by a blockade of NMDA receptors. Experimental Brain Research. 1992;89:623–639. [PubMed]
  • Pierrefiche O, Foutz AS, Champagnat J, Denavit-Saubié M. NMDA and non-NMDA receptors may play distinct roles in timing mechanisms and transmission in the feline respiratory network. The Journal of Physiology. 1994;474:509–523. [PMC free article] [PubMed]
  • Pierrefiche O, Haji A, Foutz AS, Takeda R, Champagnat J, Denavit-Saubié M. Synaptic potentials in respiratory neurones during evoked phase switching after NMDA receptor blockade in the cat. The Journal of Physiology. 1998;508:549–559. [PMC free article] [PubMed]
  • Pierrefiche O, Schmid K, Foutz AS, Denavit-Saubié M. Endogenous activation of NMDA and non-NMDA glutamate receptors on respiratory neurons in cat medulla. Neuropharmacology. 1991;30:429–440. 10.1016/0028-3908(91)90003-T. [PubMed]
  • Sant'Ambrogio G. Information arising from the tracheobronchial tree of mammals. Physiological Reviews. 1982;62:531–569. [PubMed]
  • Trenchard D. Role of pulmonary stretch receptors during breathing in rabbits, cats and dogs. Respiration Physiology. 1977;29:231–246. 10.1016/0034-5687(77)90096-2. [PubMed]
  • Widdicombe JG. Respiratory reflexes in man and other mammalian species. Clinical Sciences. 1961;21:163–170. [PubMed]
  • Wyss OAM. Excitatory and inhibitory pathways involved in lower reflex integration. Experientia. 1946;2:381–385. [PubMed]
  • Zhou Z, Champagnat J, Poon C-S. Phasic and long-term depression in brainstem nucleus tractus solitarius neurons: differing roles of AMPA receptor desensitization. Journal of Neuroscience. 1997;17:5349–5356. [PubMed]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


  • Compound
    PubChem Compound links
  • MedGen
    Related information in MedGen
  • PubMed
    PubMed citations for these articles
  • Substance
    PubChem Substance links

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...