Role of the medial medullary reticular formation in relaying vestibular signals to the diaphragm and abdominal muscles

Role of the medial medullary reticular formation in relaying vestibular signals to the diaphragm and abdominal muscles

Brain Research 902 (2001) 82–91 www.elsevier.com / locate / bres Research report Role of the medial medullary reticular formation in relaying vestib...

872KB Sizes 0 Downloads 31 Views

Brain Research 902 (2001) 82–91 www.elsevier.com / locate / bres

Research report

Role of the medial medullary reticular formation in relaying vestibular signals to the diaphragm and abdominal muscles R.L. Mori, A.E. Bergsman, M.J. Holmes, B.J. Yates* Departments of Otolaryngology and Neuroscience, University of Pittsburgh, Pittsburgh, PA 15213, USA Accepted 27 February 2001

Abstract Changes in posture can affect the resting length of respiratory muscles, requiring alterations in the activity of these muscles if ventilation is to be unaffected. Recent studies have shown that the vestibular system contributes to altering respiratory muscle activity during movement and changes in posture. Furthermore, anatomical studies have demonstrated that many bulbospinal neurons in the medial medullary reticular formation (MRF) provide inputs to phrenic and abdominal motoneurons; because this region of the reticular formation receives substantial vestibular and other movement-related input, it seems likely that medial medullary reticulospinal neurons could adjust the activity of respiratory motoneurons during postural alterations. The objective of the present study was to determine whether functional lesions of the MRF affect inspiratory and expiratory muscle responses to activation of the vestibular system. Lidocaine or muscimol injections into the MRF produced a large increase in diaphragm and abdominal muscle responses to vestibular stimulation. These vestibulo–respiratory responses were eliminated following subsequent chemical blockade of descending pathways in the lateral medulla. However, inactivation of pathways coursing through the lateral medulla eliminated excitatory, but not inhibitory, components of vestibulo–respiratory responses. The simplest explanation for these data is that MRF neurons that receive input from the vestibular nuclei make inhibitory connections with diaphragm and abdominal motoneurons, whereas a pathway that courses laterally in the caudal medulla provides excitatory vestibular inputs to these motoneurons.  2001 Elsevier Science B.V. All rights reserved. Theme: Endocrine and autonomic regulation Topic: Respiratory regulation Keywords: Reticular formation; Motoneuron; Respiration; Vestibular; Diaphragm; Abdominal muscle

1. Introduction Changes in posture, such as standing from a supine position in humans or nose-up tilt from a supine position in quadrupeds, can alter the resting length of the diaphragm [12,13,19,23,31], requiring rapid compensation if ventilation is to be unaffected. This compensation involves an increase in activity of both the abdominal muscles and the diaphragm [6,7,10,11,23,31]. Although vagal afferents are known to play a role in eliciting compensatory changes in the activity of the diaphragm and abdominal muscles during movement [9,11], considerable evidence suggests *Corresponding author. Department of Otolaryngology, University of Pittsburgh, Room 106 Eye and Ear Institute, Pittsburgh, PA 15213, USA. Tel.: 11-412-647-9614; fax: 11-412-647-0108. E-mail address: [email protected] (B.J. Yates).

that vestibular inputs also participate in triggering these respiratory alterations [21,29,32–34]. Nonetheless, the neural pathways that relay vestibular signals to diaphragm and abdominal motoneurons have not been defined. Neurons in the dorsal and ventral respiratory groups do not play a critical role in producing vestibulo–respiratory responses [29,32,34], indicating that the respiratory rhythm and labyrinthine signals are conveyed to respiratory motoneurons through different pathways. However, recent transneuronal tracing studies have shown that in addition to respiratory group neurons, cells in the medial medullary reticular formation (MRF) make synaptic connections with both diaphragm and abdominal motoneurons [1,2,35]. Because MRF neurons receive substantial vestibular inputs [3,25], it is possible that reticulospinal projections relay vestibular signals to respiratory motoneurons. The goal of the present study was to determine whether

0006-8993 / 01 / $ – see front matter  2001 Elsevier Science B.V. All rights reserved. PII: S0006-8993( 01 )02370-8

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

the MRF participates in mediating vestibular influences on the diaphragm and abdominal muscles. For this purpose, vestibulo–respiratory reflexes were compared before and after inactivation of the MRF using either lidocaine or muscimol. Vestibulo–respiratory responses were elicited by electrical stimulation of the vestibular nerve, because previous studies have shown that such electrical stimulation elicits highly reproducible responses whose magnitude remains constant from trial to trial [29,33]. For comparison, in some experiments pathways coursing in the lateral brainstem were blocked using injections of lidocaine. The results raise the possibility that inhibitory MRF neurons convey vestibular signals to respiratory motoneurons, whereas excitatory components of vestibulo–respiratory reflexes are elicited through unidentified projections that course in the lateral medulla.

2. Materials and methods All of the procedures used in this study were approved by the University of Pittsburgh’s Institutional Animal Care and Use Committee (IACUC) and conformed to National Institutes of Health guidelines.

83

to provide access to the costal diaphragm on one side. Following implantation of diaphragm electrodes, the abdominal musculature was closed with the use of sutures, and EMG electrodes were attached to rectus abdominis and the external oblique. Animals remained unventilated during most of the course of the experiment. However, following lateral brainstem lesions it was common for spontaneous respiration to disappear, and in these cases artificial respiration using a positive pressure ventilator (rate of |30 cycles / min) was employed. During artificial ventilation, end tidal CO 2 was monitored and maintained near 4%. In the other three animals, vestibular-elicited activity was recorded from lumbar nerves innervating abdominal muscles. These nerves were dissected using procedures described previously [29,33,34] and mounted on silver bipolar recording electrodes. During nerve recordings, paralysis was induced by an initial intravenous (i.v.) injection of 10 mg / kg gallamine triethiodide (Sigma), and maintained using hourly injections of 5 mg / kg. Paralyzed animals were artificially ventilated as described above. At the end of each experiment the animal was euthanized with an i.v. injection of pentobarbital sodium (120 mg / kg). Subsequently, the brainstem was removed for histological analysis.

2.1. Surgical procedures 2.2. Stimulation of the vestibular nerve Data were collected from 15 adult cats of either sex. Animals were initially anesthetized with halothane vaporized in nitrous oxide and oxygen. Catheters were placed in each femoral vein for administration of drugs, and blood pressure was monitored from a femoral artery using a Millar Instruments Mikrotip  transducer. The carotid arteries were ligated, a pre-collicular decerebration was performed, a craniotomy was conducted to expose the posterior half of the cerebellum, and the posterior cerebellar cortex was aspirated to visualize the dorsal brainstem. Following the decerebration, the administration of halothane was discontinued. Rectal temperature was maintained between 36 and 388C using an infrared lamp and heating pad. When necessary, an intravenous infusion of Aramine (metaraminol bitartrate, Merck, Sharp & Dohme, 80 mg / ml) in a lactated Ringers solution was used to maintain blood pressure above 100 mmHg. In 12 of the 15 animals, electromyographic (EMG) activity was recorded from the diaphragm, external oblique, and rectus abdominis during vestibular stimulation. Muscle activity was recorded using a pair of Tefloninsulated stainless steel wires (Cooner Wire, Chatsworth, CA, USA); each wire was stripped of insulation for approximately 5 mm and either sutured to the muscle epimysium together with an insulating patch of Silastic sheeting or inserted directly into a muscle and secured using sutures. To place EMG electrodes for recording diaphragm activity, a small incision was made through linea alba, and the liver and adjacent viscera were retracted

The vestibular nerve on each side was stimulated using procedures described in detail in previous manuscripts [18,29,33,36]. A pair of ball electrodes (1–2 mm separation) was implanted into each labyrinth using a ventral approach. The vestibular nerve was stimulated using square-wave current pulses that were 0.15 ms in duration; trains of five or 50 pulses (interpulse interval of 3 ms) were used to produce responses. The stimulus repetition rate was 1 train per 3 s. Thresholds for vestibular stimulation were expressed as multiples of the threshold (T ) required to produce eye movements, which was determined prior to decerebration. Typically, intensities 1–3 times this threshold were used to elicit vestibulo–respiratory responses. The current intensity required to produce stimulus spread to the facial nerve (the closest non-target nerve to the stimulation site) was also determined, as in prior studies [18,29,33,36]. In every animal, stimulus strengths used to elicit vestibulo–respiratory responses included those less than this intensity. It was previously demonstrated that stimulus strengths applied to the vestibular nerve that are subthreshold for producing current spread to the facial nerve selectively activate vestibular afferents [18,29,33].

2.3. Recording of nerve and EMG activity EMG or abdominal nerve activity was amplified (factor of 10 000–50 000), filtered with a bandpass of 1–10 000

84

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

Hz, full-wave rectified with a time constant of 10 ms, and led into a Cambridge Electronic Design 1401-plus data interface that was connected to a Macintosh G4 computer for averaging and storage; the sampling rate was 1000 Hz. Typically, 50 waveforms were averaged to cancel-out spontaneous muscle or nerve discharges. Runs were repeated several times and at several stimulus intensities. Stability of EMG or nerve recordings was established by verifying that response patterns and amplitudes remained consistent for at least 1 h before the injection of lidocaine or muscimol.

2.4. Procedures for inactivating brainstem regions In most animals, chemical inactivation of the MRF was produced by making 0.1 ml injections of one of the following agents into the brainstem: 2% lidocaine (Xylocaine; Astra), a local anesthetic that affects both cell bodies and fibers of passage, or muscimol (Sigma, 0.5 nmoles per injection, dissolved in phosphate-buffered saline at pH 7.4), a GABA receptor agonist that predominantly acts on cell bodies and dendrites. Injection solutions were saturated with Fast Green or Eosin dye so that injection sites could be reconstructed histologically. Injection procedures were similar to those described in previous publications [22,33,34]. Because MRF neurons that make connections with respiratory motoneurons are diffusely organized in the brainstem [1,2,35], it was necessary to inactivate a large region of the medulla through multiple injections. Typically, injections were spaced apart by 1 mm, and were made from 1 to 4 mm rostral to the obex and up to 2 mm lateral to the midline at depths of 1, 2, 3 and 4 mm from the floor of the fourth ventricle. Although multiple injections were employed in these experiments, the entire time required for the procedure was less than 10 min. If initial injections did not eliminate vestibulo–respiratory responses, a second series of larger injections (0.3 ml) spaced apart by 1 mm were made into the MRF to assure that this region was completely inactivated. In addition, injections confined to discrete regions of the MRF were sometimes made prior to the large injections described above. In two control experiments, multiple 0.1-ml injections of phosphate-buffered saline at pH 7.4 were made into the MRF, as described previously. These control experiments served to determine whether the process of making multiple injections of vehicle into the brainstem in itself could alter respiratory muscle activity. In some experiments, including those in which the MRF was functionally lesioned using lidocaine or muscimol, pathways coursing in the lateral brainstem were inactivated by making 0.1-ml injections of lidocaine near the level of the obex. These injections were made at 1-mm intervals, in tracks at approximately 4 and 5 mm lateral to the midline. It was essential to confine the lateral chemical lesions to the vicinity of the obex, so that regions of the vestibular

nuclei that mediate vestibulo–respiratory responses (which are located in the lateral medulla at approximately 4.5–6 mm rostral to the obex [33]) were not inadvertently inactivated. Following the assessment of the effects of lateral lidocaine injections on respiratory muscle activity, the lateral brainstem near the obex was sectioned using a blunt spatula to assure that all pathways coursing in this region were interrupted.

2.5. Recordings from vestibular nuclei following brainstem functional lesions In cases where drug injections abolished vestibulo– respiratory reflexes, field potential recordings were subsequently made from the vestibular nuclei to confirm that neurons located there still responded to stimulation of the labyrinth. This control experiment served to demonstrate that the effects of injections were not due to the spread of drugs to regions of the vestibular nuclei that mediate vestibulo–respiratory responses. A low-impedance (0.5 MV) glass-insulated tungsten electrode was used to record vestibular nucleus field potentials. Signals were led into an AC microelectrode amplifier (A-M Systems Model 1800) with a bandpass of 10–10 000 Hz and sampled at 1000 or 10 000 Hz. During field potential recordings, the vestibular nerve was stimulated using single shocks of 0.15 ms duration and trains of up to five shocks with a 3 ms interpulse interval.

2.6. Histology Following the experiment, the brainstem was removed and fixed in 10% formalin. The tissue was cut into 100-mm transverse sections, and every other section was counterstained with thionine. Locations of dye marks designating injection sites or mechanical brainstem lesions were reconstructed on standard sections.

3. Results In all animals, electrical stimulation of the vestibular nerve on either side evoked EMG responses in the diaphragm and abdominal muscles or alterations in activity recorded from respiratory nerves. Typically stimulus intensities 1–3-times the threshold required to elicit vestibulo– ocular reflexes were used to evoke vestibulo–respiratory responses. The minimal stimulus intensity required to produce eye movements, which was determined prior to decerebration, was 144615 (S.E.M.) mA. In contrast, the mean stimulus intensity required to produce stimulus spread to the closest non-target nerve to the stimulation site in the labyrinth, the facial nerve [18,29,33], was significantly (P,0.0001, Wilcoxon signed rank test) larger: 309632 mA. Although the maximal stimulus intensities delivered to the vestibular nerve were sometimes slightly

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

higher than those required to activate motor axons in the facial nerve, similar responses could also be elicited by current intensities that did not stimulate facial efferents. In addition, brainstem functional lesions had comparable effects on vestibulo–respiratory responses elicited by all stimulus intensities employed. Thus, we are confident that the data considered in this study reflect the effects of selective activation of vestibular afferents on respiratory activity.

3.1. Effects of inactivation of the MRF on vestibulo– respiratory responses In nine animals, respiratory responses elicited by electrical stimulation of the vestibular nerve were recorded before and after inactivation of the MRF using multiple 0.1-ml injections containing 2% lidocaine (n55) or 0.5 nmoles muscimol (n54). Typically, EMG responses were recorded from the diaphragm and abdominal muscles in unparalyzed animals, but in two experiments (one in which lidocaine was injected and another in which muscimol was employed) vestibular-evoked activity was recorded from lumbar abdominal nerves in paralyzed animals. The muscimol or lidocaine injections produced a large increase in the amplitude of EMG or abdominal nerve responses to vestibular nerve stimulation, as well as an increase in spontaneous respiratory activity. Results from three experiments are illustrated in Fig. 1. Fig. 2A compares the amplitude of EMG responses recorded from rectus abdominis, external oblique, and the diaphragm before and after inactivation of the MRF; for this analysis, the effects of injections on responses evoked by all stimulus intensities employed were pooled for each animal, and subsequently the pooled means for all animals were averaged. Inactivation of the MRF produced a sixfold increase in the amplitude of vestibular-elicited EMG responses recorded from rectus abdominis, and over a threefold increase in the amplitudes of responses recorded from external oblique and the diaphragm. The increases in response amplitude were statistically significant for each muscle (P,0.05; Wilcoxon Signed Rank Test), and occurred after the injection of either lidocaine or muscimol. For example, injection of lidocaine produced a 410% increase in the amplitude of vestibular-evoked EMG responses recorded from rectus abdominis, whereas injection of muscimol resulted in a 860% increase in response amplitude. Although transient changes in blood pressure occurred in some animals during the injections, blood pressure stabilized before maximal respiratory effects were observed. Thus, alterations in respiratory activity produced by the functional lesions were likely not directly related to these brief cardiovascular disturbances. Inactivation of the MRF also resulted in a large increase in baseline activity of abdominal muscles, as shown in Fig. 2B. Examples of spontaneous EMG activity recorded from

85

the external oblique and rectus abdominis before and immediately after functional lesions of the MRF are illustrated in Fig. 3. Following injection of either muscimol or lidocaine into the MRF, activity of both rectus abdominis and external oblique increased approximately eightfold; these effects were statistically significant (P, 0.05; Wilcoxon Signed Rank Test). However, the change in diaphragm activity following inactivation of the MRF was less pronounced and not statistically significant (see Fig. 2B). Following the initial chemical lesions, a second series of larger (0.3 ml) injections was typically made into the MRF to insure that this region was completely inactivated. The augmentation of responses produced by the initial injections was not diminished following the subsequent larger injections. These observations suggest that the effects of chemical lesions of the MRF were due to inactivation of this entire region, and not just a subset of cells located near the midline. In two additional experiments in paralyzed animals, vestibular-elicited responses were recorded from lumbar nerves innervating abdominal muscles before and after the injection of muscimol or lidocaine into the MRF. In both cases, inactivation of the MRF produced a large increase in response amplitude, as illustrated in Fig. 1D. In this particular animal, vestibulo–respiratory responses were too small for their amplitude to be accurately measured until after MRF inactivation. One complication in these studies is that injections of drugs into the medial medulla affect both the MRF and the raphe nuclei, which are located near the midline. In order to determine whether inactivation of the medullary midline alone was sufficient to produce maximal increases in baseline respiratory activity or vestibulo–respiratory responses, in five experiments the effects of lidocaine (n54) or muscimol (n51) injections confined to the midline were compared to the effects of subsequent larger injections that also affected more lateral regions of the MRF. When injections were made in tracks through the midline, only a relatively small increase in spontaneous respiratory discharges or in the magnitude of vestibulo–respiratory responses occurred, as illustrated in Figs. 3B and 4B, respectively. However, pronounced increases in spontaneous activity and evoked responses did occur following additional injections in multiple tracks at 1 and 2 mm lateral to the midline, as shown in Figs. 3B and 4C. These data indicate that inactivation of neurons dispersed throughout the medial medullary reticular formation, and not just those located near the midline, is required to produce large increases in baseline respiratory activity and in the magnitude of vestibulo–respiratory responses. In two additional control experiments, multiple 0.1-ml injections of phosphate-buffered saline at pH 7.4 were made into the MRF. In neither case did these injections result in the pronounced increases in the magnitude of vestibulo–respiratory responses that occurred following the

86

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

Fig. 1. Effects of inactivation of the MRF using multiple 0.1-ml injections of muscimol or lidocaine on vestibulo–respiratory responses. In all examples, the vestibular nerve was stimulated at intensities that were subthreshold for current spread to the facial nerve, the nearest non-target nerve to the stimulation site. All traces are the average of approximately 50 sweeps. An arrow indicates the onset of vestibular stimuli. (A) Effects of injections of muscimol into the MRF on rectus abdominis responses to stimulation of the vestibular nerve (50 shocks at 250 mA intensity). (B) Reconstructions of brainstem regions inactivated in this experiment; the region affected by muscimol was estimated from the location of Fast Green dye, which was dissolved in the injection solution. The area containing dye is indicated by shading, and is plotted on standard transverse sections of the brainstem. The values next to each section indicate the relative distance rostral to the obex. (C) Effects of muscimol injections into the MRF of a different animal on vestibular-elicited activity recorded from the diaphragm. A train of 50 shocks at 160 mA intensity elicited the responses. (D) Effects of lidocaine injections into the MRF on abdominal nerve responses to vestibular nerve stimulation (five shocks at 340 mA intensity). At this stimulus intensity, only shock artifacts were present prior to MRF inactivation, but afterwards a prominent short-latency response appeared.

injection of muscimol or lidocaine. These control experiments indicated that the effects of lidocaine or muscimol injections into the MRF were due to inactivation of medial reticular formation neurons, and not nonspecific conditions related to making multiple injections into the brainstem.

3.2. Effects on vestibulo–respiratory responses of inactivation of descending pathways in the lateral medulla following MRF functional lesions In five of the experiments in which inactivation of the

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

Fig. 2. Effects of inactivation of the MRF on the amplitude of vestibulo– respiratory responses (A) and on spontaneous respiratory muscle activity (B). Error bars indicate one S.E.M. An asterisk above a bar indicates that MRF inactivation produced a statistically significant effect (P,0.05; Wilcoxon signed rank test). Abbreviations: DIA, diaphragm; EO, external oblique; RA, rectus abdominis.

MRF enhanced vestibulo–respiratory reflexes (one in which EMG responses were increased following the injection of muscimol, three in which EMG responses were augmented following the injection of lidocaine, and one in which the magnitude of abdominal nerve responses was augmented following the injection of lidocaine), additional chemical lesions were made in the brainstem to determine the location of excitatory pathways mediating these effects. In all of these cases, responses were abolished by making multiple 0.1-ml injections of lidocaine spaced 1 mm apart in bilateral tracks at 4–5 mm lateral to the obex; an example is illustrated in Fig. 4D. These injections presumably inactivated descending pathways coursing in the lateral medulla. Following the lateral injections of lidocaine, vestibular-elicited field potentials and single unit activity could be recorded from the region of the vestibular nuclei known to mediate vestibulo–respiratory responses [33], indicating that this area had not been inactivated. However, one caveat in these experiments is

87

Fig. 3. Effects of inactivation of the MRF on spontaneous activity of abdominal muscles. Each panel shows EMG recordings made immediately before and after injection of lidocaine (A) or muscimol (B) into the brainstem. Within each panel, the top trace (A1 and B1) illustrates external oblique activity, whereas the bottom trace (A2 and B2) illustrates rectus abdominis activity. The approximate time required to produce each series of injections is also indicated. In panel B, two groups of injections were made: the first only affected the midline, whereas the second included more lateral regions of the MRF.

that the lateral lidocaine injections abolished spontaneous breathing, and animals were artificially ventilated. Nonetheless, spontaneous activity persisted in respiratory muscles and nerves, and it seems likely that diaphragm and abdominal motoneurons retained sufficient excitability to respond to vestibular stimulation.

3.3. Effects of selective inactivation of descending pathways in the lateral medulla on vestibulo–respiratory responses Because initial experiments suggested that excitatory vestibulo–respiratory reflexes are mediated through pathways coursing laterally in the medulla, in four additional experiments we examined the effects of selective inactivation of these lateral pathways on either vestibular-elicited EMG responses recorded from abdominal muscles and the diaphragm (three cases) or vestibular-evoked activity recorded from lumbar abdominal nerves (one case). Descending projections in the lateral medulla were inactivated

88

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

Fig. 4. Relationship between area of the brainstem inactivated and the magnitude of vestibular-elicited responses recorded from rectus abdominis. All traces are the average of approximately 50 sweeps; the vestibular nerve was stimulated using a train of 50 shocks at 200 mA intensity (which was subthreshold for current spread to the facial nerve). An arrow indicates the onset of the stimulus train. (A) EMG response prior to injections. (B) Response following muscimol injections in tracks through the midline, at 1, 2, 3 and 4 mm rostral to the obex. Within each track, injections were made at depths of 1, 2, 3 and 4 mm from the floor of the fourth ventricle. (C) Response following the subsequent additional injections of muscimol in tracks at 1 and 2 mm lateral to the midline on each side. Because the response duration was very long following the additional injections, the trace is shown using a longer time base than in other panels. (D) Response following the injection of lidocaine in tracks at 4 and 5 mm lateral to the midline, at the level of the obex. Inactivation of lateral portions of the caudal medulla abolished the prominent responses that resulted from injection of muscimol into the MRF.

by making multiple bilateral 0.1-ml injections of 2% lidocaine at the level of the obex, in tracks at 4 and 5 mm lateral to the midline. Artificial ventilation was required following these injections, as animals no longer breathed spontaneously. In all experiments, inactivation of lateral pathways resulted in a decrease in the amplitude of excitatory components of vestibulo–respiratory reflexes, whereas inhibitory components persisted. Subsequently, a blunt spatula was used to transect the region previously injected with lidocaine to insure that all axons coursing in the lateral medulla were severed. These mechanical brainstem lesions always resulted in a complete disappearance of excitatory responses elicited by vestibular stimulation, which were typically replaced with inhibitory responses. Examples from two experiments are shown in Fig. 5.

4. Discussion The present results demonstrate that although the MRF is not the only brainstem area that conveys vestibular signals to respiratory motoneurons, this region appears to mediate inhibitory components of vestibulo–respiratory reflexes. This conclusion is based on the following observations: (1) inactivation of the MRF using either lidocaine or muscimol resulted in an increase in the amplitude of vestibulo–respiratory responses, and (2)

inactivation of projections in the lateral medulla eliminated excitatory, but not inhibitory, components of vestibulo– respiratory reflexes. The extensive projections from the MRF to respiratory motoneurons [1,2,35] are well-suited to adjust respiratory muscle activity according to body position in space, as this region integrates signals from the labyrinth, skin, and muscle [5,8,20,24–26]. Furthermore, the MRF receives inputs from motor cerebral cortex [4,24,25], and thus could participate in regulating respiratory muscle activity during particular voluntary movements. Evidence that MRF neurons may regulate respiratory muscle activity as part of a motor program lies in the observation that lesions of this area block some respiratory adjustments during exercise [28]. Although the simplest explanation for the present data is that MRF neurons exclusively inhibit spinal respiratory motoneurons, other possibilities also exist. For example, it is feasible that both excitatory and inhibitory MRF neurons affect respiratory activity, but that the excitatory neurons are silent in the decerebrate cat preparation. The techniques used in this study could not distinguish this possibility from the existence of a selective inhibitory pathway from the MRF to diaphragm and abdominal motoneurons. Further experiments will be required to establish the neurotransmitters produced by MRF neurons involved in respiratory control. The origin of descending projections coursing in the

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

89

Fig. 5. Effects of selective inactivation of pathways coursing through lateral portions of the caudal medulla on vestibulo–respiratory responses. EMG responses were elicited by stimulation of the vestibular nerve using a train of 50 shocks at either 345 mA (A) or 120 mA (C) intensity; these current strengths were subthreshold for activating facial nerve efferents. An arrow indicates the onset of the stimulus train. All waveforms are the average of approximately 50 sweeps. Post-functional lesion traces were recorded after injection of lidocaine into the lateral medulla on each side and the subsequent transection of the same areas. (A) Effects of inactivation of lateral medullary pathways on vestibular-elicited EMG responses recorded from external oblique and the diaphragm. Post-functional lesion responses are shown at twice the gain of those recorded prior to removal of lateral projections. (B) Regions of the brainstem containing Fast Green dye (which was dissolved in the injection solution) or exhibiting mechanical damage following the functional lesions in this experiment. The combined chemical and mechanical lesions affected portions of the lateral medulla from approximately 1 mm caudal to the obex to about 3 mm rostral to the obex, but did not apparently extend rostrally to regions of the vestibular nuclei that mediate vestibulo–respiratory responses (which are located 4.5–6 mm rostral to the obex [33]). (C) Effects of lateral lesions on vestibular-elicited EMG responses recorded from rectus abdominis in another experiment.

lateral medulla that mediate excitatory components of vestibulo–respiratory responses is yet to be determined. The lateral medullary injections of lidocaine that eliminated increases in respiratory muscle activity during vestibular stimulation affected respiratory group neurons, as evidenced by the fact that spontaneous ventilation ceased following these injections. Nonetheless, respiratory group neurons are not essential for the production of excitatory components of vestibulo–respiratory responses [29,32,34], indicating that other excitatory neurons must also relay labyrinthine signals to diaphragm and abdominal motoneurons. One possibility is vestibulospinal neurons. Vestibular nucleus neurons were not transneuronally labeled within short survival periods following injection of viral tracers into respiratory muscles [1,2,35], indicating

that vestibulospinal projections do not make direct synaptic contacts with respiratory motoneurons. Nonetheless, these projections could provide inputs to premotor respiratory interneurons in the spinal cord, which in turn relay labyrinthine signals to diaphragm and abdominal motoneurons. This possibility, as well as the prospect that other descending pathways mediate excitatory components of vestibulo–respiratory reflexes, waits to be tested experimentally. Inactivation of MRF neurons also produced an increase in baseline respiratory muscle activity. This finding is not surprising, as previous studies have suggested that MRF neurons can inhibit respiratory-related discharges of the diaphragm [17,27,30]. In addition, anatomical studies have shown that many MRF neurons synthesize the inhibitory

90

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91

neurotransmitters GABA and glycine [14–16], although there is no direct evidence that inhibitory MRF neurons make connections with respiratory motoneurons. Together, these data support our conclusion that MRF neurons, including those receiving vestibular inputs, inhibit respiratory activity. In summary, the present data suggest that the MRF provides inhibitory influences on diaphragm and abdominal muscle activity, and that the vestibular system contributes to regulating the level of this inhibition. Thus, the MRF may participate in adjusting respiratory muscle activity during movements and changes in posture that activate labyrinthine receptors. Because the MRF also receives a large number of additional sensory inputs [5,8,20,24–26], as well as signals from cerebral cortex [4,24,25], this region may also be involved in adjusting the background activity of respiratory motoneurons during a variety of behaviors. The definition of the full role of the MRF in regulation of contractions of respiratory muscles awaits further experiments.

Acknowledgements Drs. Joseph Furman and Robert Schor provided helpful comments on an earlier version of this manuscript. The authors thank Ms. Lucy Cotter for valuable technical assistance. This work was supported by grants DC00693, DC03732, and DC03417 from the National Institutes of Health. Electronics support was provided through core grant EY08098 from the National Institutes of Health. M.J. Holmes was supported by the NASA’s Graduate Student Researcher Program, fellowship GSRP 99-075.

References [1] I. Billig, J.M. Foris, J.P. Card, B.J. Yates, Transneuronal tracing of neural pathways controlling an abdominal muscle, rectus abdominis, in the ferret, Brain Res. 820 (1999) 31–44. [2] I. Billig, J.M. Foris, L.W. Enquist, J.P. Card, B.J. Yates, Definition of neuronal circuitry controlling the activity of phrenic and abdominal motoneurons in the ferret using recombinant strains of pseudorabies virus, J. Neurosci. 20 (2000) 7446–7454. [3] P.S. Bolton, T. Goto, R.H. Schor, V.J. Wilson, Y. Yamagata, B.J. Yates, Response of pontomedullary reticulospinal neurons to vestibular stimuli in vertical planes. Role in vertical vestibulospinal reflexes of the decerebrate cat, J. Neurophysiol. 67 (1992) 639–647. [4] A. Canedo, J.A. Lamas, Pyramidal and corticospinal synaptic effects over reticulospinal neurones in the cat, J. Physiol. – London 463 (1993) 475–489. [5] K.L. Casey, Somatic stimuli, spinal pathways, and size of cutaneous fibers influencing unit activity in the medial medullary reticular formation, Exp. Neurol. 25 (1969) 35–56. [6] A. De Troyer, Mechanical role of the abdominal muscles in relation to posture, Respir. Physiol. 53 (1983) 341–353. [7] A. De Troyer, M. Estenne, V. Ninane, D. Van Gansbeke, M. Gorini, Transversus abdominis muscle function in humans, J. Appl. Physiol. 68 (1990) 1010–1016.

[8] T. Drew, T. Cabana, S. Rossignol, Responses of medullary reticulospinal neurones to stimulation of cutaneous limb nerves during locomotion in intact cats, Exp. Brain Res. 111 (1996) 153–168. [9] G.A. Farkas, R.E. Baer, M. Estenne, A. De Troyer, Mechanical role of expiratory muscles during breathing in upright dogs, J. Appl. Physiol. 64 (1988) 1060–1067. [10] G.A. Farkas, M. Estenne, A. De Troyer, Expiratory muscle contribution to tidal volume in head-up dogs, J. Appl. Physiol. 67 (1989) 1438–1442. [11] G.A. Farkas, M.A. Schroeder, Mechanical role of expiratory muscles during breathing in prone anesthetized dogs, J. Appl. Physiol. 69 (1990) 2137–2142. [12] J. Finkler, S. Iscoe, Control of breathing at elevated lung volumes in anesthetized cats, J. Appl. Physiol. 56 (1984) 839–844. [13] M. Green, J. Mead, T.A. Sears, Muscle activity during chest wall restriction and positive pressure breathing in man, Respir. Physiol. 35 (1978) 283–300. [14] C.J. Holmes, L.S. Mainville, B.E. Jones, Distribution of cholinergic, GABAergic and serotonergic neurons in the medial medullary reticular formation and their projections studied by cytotoxic lesions in the cat, Neuroscience 62 (1994) 1155–1178. [15] J.C. Holstege, Ultrastructural evidence for GABAergic brain stem projections to spinal motoneurons in the rat, J. Neurosci. 11 (1991) 159–167. [16] J.C. Holstege, C.M. Bongers, A glycinergic projection from the ventromedial lower brainstem to spinal motoneurons. An ultrastructural double labeling study in rat, Brain Res. 566 (1991) 308–315. [17] K. Kawahara, M. Suzuki, Descending inhibitory pathway responsible for simultaneous suppression of postural tone and respiration in decerebrate cats, Brain Res. 538 (1991) 303–309. [18] I.A. Kerman, B.J. Yates, Regional and functional differences in the distribution of vestibulosympathetic reflexes, Am. J. Physiol. 275 (1998) R824–R835. [19] A.M. Leevers, J.D. Road, Effect of lung inflation and upright posture on diaphragmatic shortening in dogs, Respir. Physiol. 85 (1991) 29–40. [20] R.A. Maunz, N.G. Pitts, B.W. Peterson, Cat spinoreticular neurons: locations, responses and changes in responses during repetitive stimulation, Brain Res. 148 (1978) 365–379. [21] A.D. Miller, T. Yamaguchi, M.S. Siniaia, B.J. Yates, Ventral respiratory group bulbospinal inspiratory neurons participate in vestibular– respiratory reflexes, J. Neurophysiol. 73 (1995) 1303–1307. [22] A.D. Miller, B.J. Yates, Evaluation of role of upper cervical inspiratory neurons in respiration, emesis and cough, Brain Res. 606 (1993) 143–147. [23] S.L. Newman, J.D. Road, A. Grassino, In vivo length and shortening of canine diaphragm with body postural change, J. Appl. Physiol. 60 (1986) 661–669. [24] B.W. Peterson, M.E. Anderson, M. Filion, Responses of pontomedullary reticular neurons to cortical, tectal and cutaneous stimuli, Exp. Brain Res. 21 (1974) 19–44. [25] B.W. Peterson, M. Filion, L.P. Felpel, C. Abzug, Responses of medial reticular neurons to stimulation of the vestibular nerve, Exp. Brain Res. 22 (1975) 335–350. [26] O. Pompeiano, D. Manzoni, U.C. Srivastava, G. Stampacchia, Convergence and interaction of neck and macular vestibular inputs on reticulospinal neurons, Neuroscience 12 (1984) 111–128. [27] C.A. Richard, R.W. Stremel, Involvement of the raphe in the respiratory effects of gigantocellular area activation, Brain Res. Bull. 25 (1990) 19–23. [28] C.A. Richard, T.G. Waldrop, R.M. Bauer, J.H. Mitchell, R.W. Stremel, The nucleus reticularis gigantocellularis modulates the cardiopulmonary responses to central and peripheral drives related to exercise, Brain Res. 482 (1989) 49–56. [29] K. Shiba, M.S. Siniaia, A.D. Miller, Role of ventral respiratory group bulbospinal expiratory neurons in vestibular–respiratory reflexes, J. Neurophysiol. 76 (1996) 2271–2279.

R.L. Mori et al. / Brain Research 902 (2001) 82 – 91 [30] R.W. Stremel, T.G. Waldrop, C.A. Richard, G.A. Iwamoto, Cardiorespiratory responses to stimulation of the nucleus reticularis gigantocellularis, Brain Res. Bull. 24 (1990) 1–6. [31] E. van Lunteren, M.A. Haxhiu, N.S. Cherniack, M.D. Goldman, Differential costal and crural diaphragm compensation for posture changes, J. Appl. Physiol. 58 (1985) 1895–1900. [32] S.F. Woodring, B.J. Yates, Responses of ventral respiratory group neurons of the cat to natural vestibular stimulation, Am. J. Physiol. 273 (1997) R1946–R1956. [33] B.J. Yates, J. Jakus, A.D. Miller, Vestibular effects on respiratory outflow in the decerebrate cat, Brain Res. 629 (1993) 209–217.

91

[34] B.J. Yates, M.S. Siniaia, A.D. Miller, Descending pathways necessary for vestibular influences on sympathetic and inspiratory outflow, Am. J. Physiol. 268 (1995) R1381–R1385. [35] B.J. Yates, J.A. Smail, S.D. Stocker, J.P. Card, Transneuronal tracing of neural pathways controlling activity of diaphragm motoneurons in the ferret, Neuroscience 90 (1999) 1501–1513. [36] B.J. Yates, Y. Yamagata, P.S. Bolton, The ventrolateral medulla of the cat mediates vestibulosympathetic reflexes, Brain Res. 552 (1991) 265–272.