THE JOURNAL OF COMPARATIW NEUROLOGY 301:604-617 (1990)

Anhtracelldar Study ofRespirato~ Neurons in the RostralVentrolateral Medda of the Rat and Their Relationship to Catecholamine-ContainingNeurons PAUL M.PILOWSKY, CHUN JIANG, AND JANUSZ LIPSKI Department of Physiology, School of Medicine, University of Auckland, Auckland, New Zealand

ABSTRACT Intracellular recording and labelling with Lucifer yellow of respiratory neurons in the rostral ventrolateral medulla were carried out in urethane-anaesthetised rats. A combined immunofluorescence and immunoperoxidase technique enabled an assessment of the tyrosine hydroxylase immunoreactivity, as well as an examination of the morphology of inspiratory and expiratory neurons in this part of the medulla oblongata. The results demonstrate: a) that respiratory neurons in the rostral ventrolateral medulla of the rat are intermingled with catecholamine-containingneurons of the C 1 cell group, but are not themselves catecholaminecontaining; b) that many non-spinally projecting respiratory neurons have axonal arborisations within the ventrolateral medulla in the same region as the C1 cell group, other respiratory neurons, and neurons reported to have a cardiovascular function; and c) that the dendrites of respiratory neurons in this region radiate throughout the ventrolateral medulla and frequently approach the ventral surface. Key words: in vivo, Lucifer yellow, immunohistochemistry,electrophysiology,tyrosine hydroxylase

The rostral ventrolateral medulla oblongata is widely considered to be an important site for the integration of many autonomic and somatic functions (e.g., Howard and Tabatabai, '75; Ciriello and Calaresu, '77; Brown and Guyenet, '84, '85; Barman and Gebber, '85; Lovick, '86; McAllen, '86; Pilowsky et al., '85; Terui, '86; Kamiya et al., '88). Electrical or chemical stimuli in this region elicit a wide range of effects, including marked changes in ventilation and blood pressure (Ross et al., '84a; Pilowsky et al., '85, '86a; Gatti et al., '86; Hodgson et al., '87; Chen et al., '88; Haxhiu et al., '88). Furthermore, a number of electrophysiological studies have identified neurons in the ventrolateral medulla that discharge in relation to the respiratory or cardiac rhythm (Howard and Tabatabai, '75; Brown and Guyenet, '84, '85; McAllen, '87; Saether et al., '87; Ellenberger and Feldman, '88; Ezure et al., '88; Morrison et al., '88). In parallel with these physiological experiments, numerous reports have documented the location and neurotransmitter content of cell bodies and terminals in this region. These include neurons containing catecholamines, serotonin, gamma-aminobutyricacid, substance P, and neuropeptide Y (Hokfelt et al., '74; Ross et al., '84a,b; Kalia et al., '85a,b; Pilowsky et al., '86b; Agnati et al., '88; Milner et al., Q

1990 WILEY-LISS, INC.

'88; Livingston and Berger, '89). Each of these groups of neurons is organised as a cell column that extends from the caudal border of the facial nucleus to the cenricomedullary junction. In addition, these functionally or histochemically identified cell groups have been shown to overlap with each other, either wholly or partially (e.g., Hodgson et al., '87; Ellenberger et al., '90). For example, electrophysiological studies (Howardand Tabatabai, '75; Suzue, '84; Brown and Guyenet, '85; Murakoshi et al., '85; Saether et al., '87; Ezure et al., '88; Onimaru et al., '88, '89) have demonstrated neurons with a respiratory-related discharge in the area which overlaps with the region of the Al/Cl catecholamine-synthesizing neurons. However, it is still not possible to say whether or not these electrophysiologicallyidentified neurons, often referred to as the Ventral Respiratory Group (VRG; Saether et al., '87; Onimaru et al., '88; Ezure et al., '88; Wilhelm et al., '901, are catecholamine-containing.

Accepted August 2,1990. Address correspondence to J. Lipski, Dept. Physiology, University of Auckland, Private Bag, Auckland I, New Zealand. Paul M. Pilowsky is now at Department of Medicine, Flinders Medical Centre, Flinders University, Adelaide, Australia.

RAT RESPIRATORY NEURONS AND CATECHOLAMINES

605

The present study was carried out with two main aims. border of the facial nucleus and the rostral part of the The first was to assess tyrosine hydroxylaseimmunoreactiv- nucleus ambiguus (pars compacta, c.f. Bieger and Hopkins, ity in respiratory neurons located in the rostral part of the '87) was mapped by recording the antidromic field potential VRG cell column. This was achieved by making intracellu- obtained by stimulation of the facial nerve and the superior lar recordings from these neurons, labelling them with laryngeal nerve, respectively. At the same time, it was Lucifer yellow, and then using immunofluorescence to usually possible to locate the region with a multiunit assess their content of tyrosine hydroxylase. Secondly, we respiratory activity. The location of these units generally wanted to provide a morphological description of the cell coincided with, and was ventral and medial to, the antidrobodies, dendrites, and axons of functionally identified respi- mic field potential elicited by stimulation of the superior ratory neurons. A preliminary account of this work has laryngeal nerve, and immediately caudal to the facial nucleus. An example of the mapping of the antidromic field been presented (Pilowskyet al., '89). potentials from the facial and superior laryngeal nerves is shown in Figure 1. MATEJ3JALSANDMETHODS Following the initial mapping, the microelectrodes were Animalpreparation replaced with those suitable for intracellular recording. Twenty-four male Wistar rats weighing 295-310 g were Glass microelectrodes containing 4% Lucifer yellow CH in anaesthetised with urethane (1.1&g in 10 ml of lactated 0.5 M LiCl were used (resistance, 25-35 Rohm). In all cases, Ringer solution, i.p.1. Additional smaller doses of anaes- impaled neurons were characterised according to the relathetic were administered intravenously as required. The tionship of their membrane potential with the discharge of trachea was cannulated and a femoral artery and vein the phrenic nerve. Neurons were classified as inspiratory if catheterised to enable measurement of arterial blood pres- depolarization of the membrane potential occurred during sure and administration of drugs respectively. The phrenic phrenic nerve discharge, and expiratory if hyperpolarizaand superior laryngeal nerves were dissected in the neck tion occurred during this period. Where possible, the antidroand prepared for standard bipolar recording or stimulation. mic response to stimulation of the spinal cord, facial nerve, Bilateral incisions were made on the cheeks to expose the and superior laryngeal nerve was determined. If there was marginal mandibular branch of the facial nerve for stimula- no clear relationship between the membrane potential and tion. The rat was mounted in a stereotaxic frame that the discharge from the phrenic nerve, the neuron was not enabled fixation of the skull and spinal column. Mean included in this study. Lucifer yellow was injected by using arterial blood pressure, expired CO,, tracheal pressure, and hyperpolarizing current pulses (1-5 nA) with a 50% duty rectal temperature were monitored. In order to obtain cycle for 2-5 minutes. In each rat, we attempted to fill no minimal movement of the brainstem during recording, a more than two neurons (at least 300400 p,m apart) on each portion of the parietal bones and the occipital bone were side. removed, and a laminectomy from C3 to C6 was performed. Histological analysis The caudal part of the cerebellum was removed with suction. The animals were paralysed with pancuronium At least 1 hour after the last intracellular injection of bromide (1.0 mgkg i.v.; Pavulon) and artificiallyventilated. Lucifer yellow was made, the rat was perfused transcardially with 200-300 ml of 0.9% NaCl containing 10 mM Stimulationandrecording sodium nitrite, followedby 1litre of 0.1 M phosphate buffer Phrenic nerve discharge was amplified (bandwidth, 100 containing 4% formaldehyde and 0.2% glutaraldehyde. The Hz-3 kHz), full wave rectified, and integrated (time con- brainstem was then removed and stored overnight in the stant, 50 ms). The tidal volume and frequency of the same fixative at 4°C. Transverse sections (50 pm) of the ventilator were adjusted to ensure that the frequency of the medulla oblongata from about 250 p,m caudal to obex up to bursts of phrenic nerve activity were not synchronised with the level of the facial nucleus were made by using a the ventilator. The intensity of stimulation of the superior Vibratome, and the sections were mounted serially with a laryngeal nerve was set at three times the threshold glycerol-carbonate buffer (1:1; pH 8.0). All incubations required to inhibit phrenic nerve activity during a train at involving immunohistochemical reagents were conducted 50 Hz (0.1 ms duration). The stimulus intensity (0.1 ms at room temperature with continuous agitation and were duration) for the marginal mandibular branch of the facial separated with three 10 minute washes in buffer containing nerve was initially set to be just above the threshold Triton X-100 (0.3%), and (in mM) NaCl 120, KC1 5, required to elicit a twitch in the surrounding muscle. Two NaH,PO, 1.5, N%HPO, 8.5, Tris base 10, sodium merthiomonopolar stimulating electrodes were placed bilaterally in late 1, pH 7.4. All incubations in primary antisera also the lateral funiculus of the C3 segment of the spinal cord contained 10%heat-inactivated porcine serum, while incujust rostral to the phrenic nucleus, and positioned at the bations in secondary antibody contained 1%heat-inactidepth required to obtain the largest evoked potential from vated porcine serum. the phrenic nerve when the current was passed between the The histological analysis was carried out in three stages. two electrodes with the cathode being on the side of the In the first step, Lucifer yellow-containing neurons were phrenic recording (up to 1 mA, 0.1 ms). If a neuron located with the aid of the fluorescence microscope, and responded antidromically to this stimulus (as determined then the sections were stained to reveal tyrosine hydroxyby the collision test; e.g., Lipski, '811, the current was lase immunoreactivity. In the next step the sections were decreased close to threshold and the polarity of the current reexamined with the fluorescence microscope to see if reversed in order to determine if the neuron had an Lucifer yellow+ontaining cells (yellow fluorescence) were also catecholamine-containing (red, rhodamine fluoresipsilateral or contralateral projection. In each experiment, recordings were made initially with cence),and to assess the relationship of the Lucifer yellowglass microelectrodes containing 3M NaCl (tip diameter, 2 filled neurons to the Al/Cl cell group. Finally, an immunoto 4 pm; resistance, 2 to 4 Ma). The location of the caudal peroxidase reaction was carried out with antibodies to

606

P.M. PILOWSKY ET AL.

t v 1.0m m

FN

t SLN

Fig. 1. A diagram of a coronal section through the rostral medulla of the rat. Antidromic field potentials recorded following stimulation of the facial nerve (FN) and the superior laryngeal nerve (SLN) at the sites indicated by the dots are shown on the right as a, b, c, and d. Following SLN stimulation, a field potential could be recorded when the electrode was located within or close to the nucleus ambiguus pars compacta

(NAc).When the electrode was located within the facial nucleus (VIIj, a larger field potential was recorded following FN stimulation. The field potential recorded after stimulation of the SLN provided a good index of the mediolateral position of the examined respiratory cell group, while the field recorded after stimulation of the FN was useful in determining the rostral boundary of this cell group. PYR, pyramidal tract.

Lucifer yellow, thus allowing a detailed reconstruction of Lucifer yellow-stained respiratory neurons. These steps are described in detail below. Firstly, the sections were examined by using a fluorescence microscope (Leitz, filter block D). Sections containing cell bodies with Lucifer yellow fluorescence were removed from the slide and incubated for 24 to 48 hours in a rabbit antibody t o tyrosine hydroxylase (1:lO; Eugene Tech). The sections were then incubated in a rhodamine-labelled sheep antirabbit antibody for 12-24 hours (1:300;Silenus, Australia). Sections rostral and caudal to these sections that contained fluorescent neuronal processes belonging to Lucifer yellow-labelled neurons were incubated in a polyclonal mouse antibody to Lucifer yellow for 24-48 hours (1:1,000; R.Prestidge, Auckland). Secondly, the sections that had been incubated in antibody to tyrosine hydroxylase and rhodamine-labelled second antibody were again mounted in glycerol-carbonate buffer and reexamined with the fluorescence microscope. Filter block D was again used to locate the Lucifer yellowcontaining cell body, while filter block N2 was used to locate rhodamine fluorescence (indicating the presence of tyrosine hydroxylase immunoreactivity). By changing between the two filters, it was possible to determine whether or not Lucifer yellow-containingneurons were also immunoreac-

tive for tyrosine hydroxylase, and to examine the relationship of Lucifer yellow-containing neurons to tyrosinehydroxylase-immunoreactiveneurons. With this system, it was generally possible to discriminate between the two fluorophores. In some cases the Lucifer yellow fluorescence was detectable as a faint red fluorescence by using the N2 filter block. In this situation it was still possible to differentiate tyrosine hydroxylase immunoreactivity from Lucifer yellow fluorescence, since Lucifer yellow always gave a strong nuclear stain while tyrosine hydroxylase immunoreactivity was not present in the nucleus. Furthermore, Lucifer yellow staining became weaker at sites distal to the cell body whereas tyrosine hydroxylase immunoreactivity was of equal intensity in all parts of the neuron. Thirdly, the sections containing Lucifer yellow cell bodies were again removed from the slides and incubated in the mouse antibody to Lucifer yellow for 24-48 hours. The sections, both those containing cell bodies and those rostral and caudal containing dendrites or axons, were then incubated in a biotinylated sheep antimouse antibody for 12-24 hours (1:300;Sigma), followed by an avidin-horseradish peroxidase conjugate for 3-12 hours (1:1,000; ExtrAvidin peroxidase; Sigma). Lucifer yellow immunoreactivity was revealed by incubation in a solution containing diaminobenzidine (0.05%),nickel ammonium sulphate (0.6%), and

RAT RESPIRATORY NEURONS AND CATECHOLAMINES hydrogen peroxide (0.003%)for 15-20 minutes. Finally, the sections were mounted serially, and reconstructions of the stained neurons were made by using a camera lucida technique.

RESULTS Electrophysiologidcharacteristics Stable intracellular recordings were made from a total of 48 neurons in the rostral ventrolateral medulla. Forty-five of these neurons lay within 600 pm of the caudal pole of the facial nucleus. This region corresponds to the location of the rostral VRG, the C1 neurons, and the bulbospinal sympathoexcitatory neurons (Howard and Tabatabai, '75; Brown and Guyenet, '84, '85; Saether et al., '87; Ezure et al., '88; Morrison et al., '88; Ellenberger et al., '90; Saji and Miura, '90). The remaining three neurons were located between 600 pm and 1000 km from the facial nucleus and were not used for histological analysis. The relationship of these neurons to the nucleus ambiguus compact formation (Bieger and Hopkins, '87) and the ventral surface of the medulla at a point approximately 300 pm caudal to the caudal pole of the facial nucleus is shown in Figure 2. The average membrane potential during the initial period of recording was -29 mV (range -15 to -50 mV). Thirtythree neurons were classified as expiratory, while 15 were classified as inspiratory. Five out of 29 (17%) expiratory neurons were spinally projecting (three ipsilateral and two contralateral). One out of 14 (7%) inspiratory neurons tested was spinally projecting (ipsilateral). Out of ten nonspinally projecting neurons tested (seven expiratory, three inspiratory) none was antidromically activated from the superior laryngeal nerve.

_ -_

/I

0

o

During the iontophoresis of Lucifer yellow, the hyperpolarisation of membrane potential that was seen during phrenic nerve discharge in expiratory neurons was often decreased in amplitude (in 15 of 28 neurons). In most of these neurons (n = 131, after prolonged iontophoresis of Lucifer yellow, these hyperpolarisations reversed in polarity; i.e., the membrane potential became more depolarised during phrenic nerve discharge. An example is shown in Figure 3.

Lucifer yellow fluorescence Lucifer yellow was iontophoresed into 37 respiratory neurons (28 expiratory and nine inspiratory). Of these, 16 neurons containing Lucifer yellow fluorescencewere found on examination with the fluorescence microscope (11expiratory and five inspiratory). The usual reasons for failure to locate an injected neuron appeared to be either a small total charge transfer (less than 5 nA . minute), or a poor membrane potential at the end of the iontophoresis (less negative than -10 mV) indicating cell damage and leakage of Lucifer yellow from the cytoplasm. Well-stained neurons showed extensive filling of the dendritic tree and axons (Fig. 4A,C).

Tyrosine hydmxyhse immunofluorescence Tyrosine hydroxylase immunoreactivity was found within the cell bodies and fibres of numerous neurons in the rostral ventrolateral medulla. The distribution of tyrosine hydroxylase immunoreactivity within cells was the same as in previous studies (Kalia et al., '85a; Sun et al., '88b): strong immunoreactivity was seen within the cytoplasm of the soma and fibres, but the nucleus was unstained (Fig.

A

0

!

607

*.

A h.'

f l . 0

Fig. 2. A composite reconstruction of a coronal section through the rostral medulla of a rat at a site approximately 300 pn caudal to the facial nucleus. The right-hand side shows the location of the 48 respiratory-related units recorded. Their locations were mapped on the basis of measurements taken during the recording sessions. The filled triangles indicate inspiratory units, the open triangles expiratory units.

The left-hand side shows the location of tyrosine hydroxylase immunoreactivity in cell bodies mapped from two consecutive 50 Fm sections. It can be seen that the two cell groups form largely separate columns, with considerable overlap. XII, hypoglossal nucleus; NTS, nucleus tractus solitarius; NAc, nucleus ambiguus pars compacta; 10, inferior olive; PYR, pyramidal tract.

I

m

i n

n 5

n

c

RAT RESPIRATORY NEURONS AND CATECHOLAMINES 4B,D,F). Immunoreactive cells extended in a broad arc, from ventral to the nucleus ambiguus (compact formation), towards the medial border of the pyramidal tract (Fig. 2). This distribution corresponded to that described by previous authors (e.g., Kalia et al., '85a; Ellenberger et al., '90). Tyrosine hydroxylase immunoreactivity was not detected in any of the 16 Lucifer yellow-labelled neurons described above. Although none of the respiratory neurons that we examined contained tyrosine hydroxylase immunoreactivity, all of the labelled neurons were in close proximity to cell bodies and dendrites of immunoreactive neurons, The general relationship of the respiratory neurons to the AlIC1 cell group is shown in Figure 2. Examples of individual neurons are shown in Figure 4.

Luciferydowbunoreactivity The use of antibodies to Lucifer yellow combined with immunoperoxidase histochemistry allowed us to stain injected neurons in such a way as to enable detailed reconstructions by using the camera lucida technique. Twelve of the 16 Lucifer yellow-labelled neurons that were found with the fluorescence microscope were sufficiently well stained so as to allow a detailed reconstruction. In these cases, most dendrites appeared to be filled to their terminations, and in many cases axons and axon terminations were well seen, even though this was not possible when looking at Lucifer yellow fluorescence. A low-power photomicrograph of a well-stained neuron is shown in Figure 5A.

609

not significantly different from those of the tyrosinehydroxylase-immunoreactive neurons. Only three inspiratory neurons could be examined in this study, and the diameters of these ranged from 9.1 to 16.2 pm (short axis) and from 16.4 to 23.3 pm (long axis).

Dendritic morphology Cranial motoneurons. The dendrites of the two neurons with cell bodies in the nucleus ambiguus (compact formation) were oriented at about 20-45" from the vertical in the coronal plane. In both neurons, dendrites travelling in a ventral direction came to within 300 pm of the ventral surface of the medulla (Fig. 7). Ventral respiratory group neurons. The dendrites of the other neurons arborised throughout the ventrolateral medulla, and in many cases terminated within 100-200 pm of the ventral surface of the medulla. The dendrites were usually present in the same coronal plane as the cell body, or at least within 100-200 pm rostrocaudally. In some cases, however, two to three dendrites could be traced that travelled more than 500 pm rostrocaudally. The shafts of the dendrites were usually smooth in their proximal regions, but often became highly specialized at their terminations, with multiple appendages of varying shape (Fig. 5). These dendritic specialisations were sometimes seen close to the ventral surface of the medulla, or close to blood vessels.

Cell bodies ofrespiratoryneurons

AXOnS

Cranial motoneurons. Two of the Lucifer yellowfilled cell bodies were located within the borders of the rostral nucleus ambiguus (cornpact formation), and were probably laryngealor pharyngeal motoneurons. This conclusion was also supported by the analysis of the axonal trajectory in one of the two cells (see below); no axon could be located in the second. Both of these neurons were expiratory, in that the membrane potential became hyperpolarised during phrenic nerve discharge. Ventral respiratory group neurons. The cell bodies of the remaining neurons (seven expiratory, three inspiratory) were located medial or ventral to the nucleus ambiguus compact formation (Figs. 3, 6). There was no obvious difference in the shape or appearance of inspiratory, compared with expiratory, neurons. Both types of neuron were multipolar with the majority of primary dendrites present in the coronal plane. In comparison with tyrosine-hydroxylase-immunoreactiveneurons, respiratory neurons tended to be similar in size (e.g., Fig. 4). An analysis of the long and short axis of 30 randomly selected tyrosine-hydroxylaseimmunoreactive neurons revealed that the short axis was 15.6 pm (SD, 6.7 pm), and the long axis was 29.6 pm (SD, 8.3 pm). The diameters of the seven expiratory neurons in this study were 17.3 pm (SD, 7.5 pm) and 35.1 hm (SD, 9.1 pm). A Student's t test showed that these diameters were

Cranial motoneurons. In only one of the two neurons located within the nucleus ambiguus (compact formation) was an axon found. This neuron is shown in Figure 7. The axon coursed rostrally, dorsally and medially, until it reached the region of the nucleus tractus solitarius, where it turned sharply laterally. The Lucifer yellow immunoreactivity became very pale at this point, and the axon could not be traced to its exit from the medulla. Ventral respiratory group neurons. None of the remaining neurons where the Lucifer yellow immunoreactivity was adequate to allow reconstruction had a spinal projection on electrophysiological criteria. On the other hand these neurons frequently had very fine axons with numerous branches and terminations in the region of the rostral nucleus ambiguus, and ventral and medial to this nucleus. This includes the region of the C 1 catecholamine neurons, although no evidencewas obtained in this study to demonstrate that catecholamine-containingneurons in this region receive a direct synaptic input from respiratory neurons. In two cases, axon terminations were seen within the nucleus ambiguus (compact formation), while in another case axons with numerous boutons-both terminal and en passant-were seen projecting dorsally and medially in the lateral tegmental field, although never as far as the nucleus tractus solitarii (Figs. 3,6).

Fig. 3. Camera lucida reconstructions of two expiratory neurons. The insets show the general location of each neuron in the medulla. Neither neuron had a spinally projecting axon. In both cases, ventrally projecting dendrites can be seen that travel to within 300-400 pm from the ventral surface of the medulla oblongata. In the neuron on the right, an axon is present with arborizations in the nucleus ambiguus pars compacta (NAc) and in the area ventral and medial to this nucleus. The chart recordings show waves of the membrane potential (MP)

synchronized with discharges of the phrenic nerve (Phr) and not with fluctuations of tracheal pressure (TP) due to ventilation. The first part of each record was made immediately after impalement. The gap in records indicates the period of Lucifer yellow injection. In both cases it can be seen that there is a reversal of the polarity of the waves in the membrane potential after injection of Lucifer yellow. The Calibration bar is 100 pm for the reconstructions and 1.0 mm for the insets. The calibration bar for the chart record is 1.2 seconds.

P.M. PILOWSKY ET AL.

610

Fig. 4. Photomicrographs illustrating the lack of tyrosine hydroxylase immunoreactivity within Lucifer yellow-filled respiratory neurons. A,C,E are examples of Lucifer yellow-filled neurons (filter block D). B,D,F are the same sites photographed with filter block N2 showing rhodamine fluorescence within tyrosine-hydroxylase-immunoreactive neurons. Although none of the respiratory neurons was immunoreac-

tive for tyrosine hydroxylase, dl were located in close proximity to other tyrosine-hydroxylase-immunoreactive neurons. Lucifer yellow fhorescence was stronger in the nucleus than the cytoplasm, whereas tyrosine hydroxylase immunoreactivity was only present in the cytoplasm. This effect is best seen in E and F. Calibration bars, 50 pm.

DISCUSSION

contain phenylethanolamine-N-methyltransferaseimmunoreactivity (e.g., Howe et al., '80; Kalia et al., '85b; Ellenberger et al., 'go), and this region is therefore the same as the C1 area originally described by Hokfelt et al. ('74) None of the respiratory neurons that were adequately filled1 with Lucifer yellow in this study were found to contain

Relationship Of

respiratory@up neuronstocatecholaminenneurons

Previous studies have shown that the vast majority of catecholamine neurons in the rostral ventrolateral medulla

RAT RESPIRATORY NEURONS AND CATECHOLAMINES

611

Fig. 5. Photomicrographs of Lucifer yellow immunoreactivity. A well-filled neuron is shown in A. An axon with terminal boutons is shown in B. Examples of dendritic morphology are shown in C-F. Note the highly specialized nature of these terminal dendrites, with many

appendages (C,D,F). An example of two dendrites that terminate close to the ventral surface of the medulla is shown in E. Calibration bars for A, D, F are 10 hm; the calibration for E is 100 pm.

tyrosine hydroxylase immunoreactivity (a marker for all catecholamine neurons), although all were found to have cell bodies and dendrites that were located in the same region as tyrosine-hydroxylase-immunoreactive neurons and dendrites. This latter finding is supported by the recent reports by Ellenberger et al. ('87, 'go), who have examined the relationship between neurons in the ventrolateral medulla that can be retrogradely labelled from the phrenic

nucleus, neurons that are labelled from injections in the ventral respiratory group, and neurons that are immunoreactive for catecholamine-synthesizingenzymes (Al/Cl neurons). In their study, it was found that there was some intermingling of the respiratory and catecholamine-containing populations of neurons (cf. Arata et al., '901, especially in the rostral ventrolateral medulla, although they formed largely separate cell columns. Furthermore, it was noted by

Fig. 6. Camera lucida reconstructions of two inspiratory neurons. In both cases fine wit.h extensive arborizations were present in the region ventral and medial to the nucleus ambiguus pars compacta (NAc). i n the rieuioii 011 t h e right, a dorsomedially projecting axon is present that gave rise til numerous boutons The ventral branch of this axon arhorizrs in the region that contains a high density of tyrosine-hydroxyiasc-

&~IIS

immunoreactive neurons. The axon on the left shows a limited arborization within the NAc. In both cases, ventrally projecting dendrites are present that terminate within 100 to 200 prn from the ventral surface of the medulla oblongata. In the neuron on the right it can be seen that the find portiozs of these dendrites become highly specialised. c~!alibratisribar is 100 urn for the reconstructions and 1.0 mm for the insets.

Ventral Surface

A /

F

c N

Q)

RAT RESPIRATORY NEURONS AND CATECHOLAMINES Ellenberger et al. ('90) that in spite of this intermingling, no double labelling was observed. The conclusion that propriobulbar respiratory neurons do not contain tyrosine hydroxylaseimmunoreactivity must be made with some caution. Firstly, it is possible that Lucifer yellow or chloride ions interfered with tyrosine hydroxylase immunoreactivity resulting in a false-negative finding. This possibility seems unlikely since other workers using the same primary antibody in "in vitro" studies were able to demonstrate immunoreactivity within Lucifer yellow-filled neurons in the midbrain, locus coeruleus, and ventrolateral medulla (Sun et al., '88b; Grace and Onn, '89). Secondly, it is possible that the cells we recorded from had become damaged by the impalement, and this damage resulted in a loss of tyrosine hydroxylase immunoreactivity. We have attempted to record from, and label, nonrespiratory neurons within the ventrolateral medulla in order to demonstrate tyrosine hydroxylase immunoreactivity, but this has so far not been possible in vivo (unpublished observations). It remains to be determined to what extent neuronal depolarisation resulting from impalement and neuronal damage affects the immunoreactivity of intracellularly labelled neurons. Finally, a morphometric analysis of the soma size of tyrosine-hydroxylase-immunoreactive neurons suggests that these neurons are generally similar in size to the respiratory neurons that we examined. It therefore seems unlikely that a size difference between these two groups of neurons could account for the findings reported here. The question of the functional role of catecholaminergic and other neurons in the ventrolateral medulla has also been addressed by Sun et al. ('88a,b) using an "in vitro" approach. These workers have studied the electrophysiological properties of neurons in this medullary area and found that tyrosine hydroxylase immunoreactivity was contained within silent cells (Sun et al., '88b). On the basis of this result, and of their hypothesis that sympathoexcitatory neurons have endogenous pacemaker activity "in vivo" (Sun et al., '88a), these authors concluded that C1 neurons were unlikely to be responsible for the generation of basal sympathetic tone. Instead, it was suggested that they may be involved in relaying "respiratory-related information to the spinal autonomic nucleus" (Sun et al., '88a). This possibility is supported by a recent extracellular study from the same laboratory that suggests that C1 neurons may receive an input from central respiratory neurons (Haselton and Guyenet, '89). The finding in the present study that respiratory neurons have axonal arborisations within the rostral ventrolateral medulla in the region of the C1 cell group provides indirect support for this hypothesis. Unfortunately, the functional role of these catecholaminecontaining cells still remains to be determined.

613

berger and Feldman, '88; Wilhelm et al., '90). Further support for this notion has been provided by a number of workers using a neonatal rat brainstem-spinal cord preparation (Suzue, '84; Murakoshi et al., '85; Onimaru et al., '88, '89; McCrimmon et al., '89). The location of the cell bodies of respiratory neurons recorded from in this study corresponds with that found in previous studies using extracellular recording techniques (Howard and Tabatabai, '75; Saether et al., '87; Ellenberger and Feldman, '88; Ezure et al., '88). In determining the location of the ventral respiratory group, antidromic mapping of the field potentials that can be recorded following stimulation of the facial nerve and the superior laryngeal nerve was invaluable. Recent electrophysiological (Yajima and Hayashi, '89) and anatomical (Bieger and Hopkins, '87) studies have shown that motoneurons with axons in the superior laryngeal nerve have their cell bodies within the rostral nucleus ambiguus (compactformation). This finding is confirmed in the present study. The antidromic field elicited from the facial nerve was maximal immediately rostral to, and deep to, the field from the superior laryngeal nerve. Electrophysiologically, neurons with a respiratoryrelated discharge pattern were found caudal to the antidromic field potential from the facial nerve, and at the level of, and deep to, the antidromic field potential from the superior laryngeal nerve. Anatomically,the ventral respiratory group overlaps with the nucleus ambiguus (compact and external formations) and the rostral third of the nucleus paragigantocellularis as defined by retrograde tracing and cresyl violet staining studies (Andrezik et al., '81; Bieger and Hopkins, '87; Portillo and Pasaro, '88). Other cell groups that have also been localised to this region and that overlap with the ventral respiratory group include the C1 cell group and the ventral presympathetic group, raising the possibility that some of these cells may interact or share common inputs. Although anatomical tracing studies have documented a pathway from the ventrolateral medulla to the phrenic motor nucleus and thoracic expiratory motor neurons (Onai et al., '87; Guyenet and Young, '87; Ellenberger and Feldman, '88; Yamada et al., '88; Saji and Miura, '901, only a relatively small proportion of medullary respiratory neurons (13%)have been shown to project to the spinal cord by using electrophysiological techniques (Saether et al., '87). In this study we found that 14% of neurons tested had a spinal axon (5/29; Expiratory, 1/14;Inspiratory), which is in close agreement with the findings of Saether et al. 1'87). It therefore seems unlikely that this small percentage of spinally projecting neurons found in our study is due to a sampling bias with the intracellular technique. This is also supported by the finding that most bulbospinal neurons of the ventral respiratory group in the rat arise from the area of the caudal part of the rostral ventrolateral medulla (Onai organisation of respiratmyneurons et al., '87). It is at present still unclear whether or not the organisaThe importance of the ventrolateral medulla in the control of respiration in the adult rat has been addressed in tion of the ventral respiratory group in the rat is similar to a small number of recent studies. Extracellular electrophys- that seen in the cat. Some differences are likely since, for iological studies have demonstrated the existence of units example, most of the expiratory neurons in the rostral that have a firing pattern synchronised with the discharge ventrolateral medulla of the rat are not spinally projecting, of the phrenic nerve or with diaphragmatic EMG or with in contrast to the Botzinger neurons of the cat that often spontaneous ventilation, indicating that they may be in- have a spinal axon (Fedorko and Merrill, '84). It may volved in the generation or transmission of central inspira- therefore be inappropriate at this stage to use terminology tory activity (Howard and Tabatabai, '75; Brown and that has developed to describe respiratory neurons in the Guyenet, '85; Saether et al., '87; Ezure et al., '88; Ellen- cat, until the role of these ventral medullary respiratory

P.M. PILOWSKY ET AL.

614

. MP

Phr

TP

Ventral Surface

/------

Figure 7

RAT RESPIRATORY NEURONS AND CATECHOLAMINES neurons has been more accurately determined. In this report we have used the term "ventral respiratory group" to refer to all neurons in the ventrolateral medulla that have a respiratory-related modulation of their membrane potential. During iontophoresis of Lucifer yellow, the hyperpolarisation of the membrane potential seen in expiratory neurons during phrenic nerve discharge often became reversed to a depolarisation probably due to a concomitant iontophoresis of chloride ions. This reversal suggests an active inhibition of expiratory neurons during inspiration rather than a disfacilitation, a finding similar to that in the cat (e.g., Ballantyne and Richter, '86; Arita et al., '87). However, the present result must be treated cautiously, since the membrane potentials at the end of iontophoresis were generally poor, and no tests with current injection during the initial period of recording were carried out.

615

study of Botzinger complex neurons in the cat has demonstrated the presence of axonal branches with boutons in this region (Otake et al., '87). The finding in the present study that some respiratory neurons in the rat have extensive axonal terminations within the same region suggests that similar interactions may occur in the rat. For example, it seems likely that axon terminations of expiratory neurons would provide an inhibitory input to inspiratory neurons. In addition, these axonal terminations are located in the same region as neurons known to be responsible for the generation of basal sympathetic tone. This association raises the possibility that these propriobulbar respiratory neurons may be responsible for the well-known respiratory modulation of sympathetic preganglionic neurons (e.g., Gilbey et al., '86) through an action on sympathetic premotor neurons (McAllen, '87; Haselton and Guyenet, '89). Another possible target for these inputs are the C 1 neurons that are located in the same region.

Dendrites andaxonsof intracellularly labelledneurons

Neurotransmittercontent of respiratoryneurons

The wide arborisation of the dendritic trees of respiratory neurons illustrated in this study is noteworthy, and has At the present time, none of the functionally identified also been shown in two recent studies of respiratory neurons known to be present in the rostral ventrolateral neurons in the rostral medulla of the cat (Otake et al., '87; medulla have been characterized in terms of their neuroGrblot et al., '88). Although the location of the cell bodies in transmitter content. This problem is made all the more any individual rat tends to be confined to the nucleus difficult since neurons in the rostral ventrolateral medulla ambiguus and the region medial and ventral to it, the are known to subserve a number of different physiological dendrites of these neurons project over a much wider area. functions, and to contain a number of neurotransmitters. This finding suggests that physiological and anatomical With regard to the regulation of central respiratory neustudies that only focus on the region of the cell bodies may rons, pharmacological and immunocytochemical studies significantly underestimate the functional input to respira- have implicated a number of putative neurotransmitters tory neurons in the ventrolateral medulla. including excitatory amino acids (McCrimmon et al., '89), Of particular interest in this regard are those dendrites GABA (Holtman and King, '88; Lipski et al., '901, acetylchothat approach the ventral surface. Since the report by line (Bradley and Lucy, '831, serotonin (Holtman et al., '86; Feldberg ('76) that the application of various pharmacolog- Voss et al., 'go), adrenaline (Bolme et al., '741, substance P ical agents to the ventral surface of the medulla induces a (Chen et al., '88), angiotensin-I1 (Aguirre et al., '89a), range of cardiovascular and respiratory effects, consider- neurotensin (Haxhiu et al., "9,neuropeptide Y (Fuxe et able attention has been focussed on the neural substrates al., '83; Aguirre et al., '89b), and somatostatin (Yamamoto that may mediate these responses. From experiments con- et al., '88). ducted in the cat it was found that these topically applied Future studies of the type described here may enable the agents exerted their effects only upon structures located identification of individual neurotransmitters within respiwithin about 350 +m of the ventral surface (Rohlicek and ratory and other functionally identified neurons in the Polosa, '83; Prabhakar et al., '89). The neuronalreconstruc- medulla. tions presented in this study indicate that these effects could be mediated through the ventrally projecting dendrites of respiratory neurons whose cell bodies are located A C I ( " T S deeper in the medulla. The presence of ventrally projecting We would like to thank Ms. Karen Bark and Ms. Vinese dendrites was noted both within neurons located in the nucleus ambiguus and in other neurons located ventrally Yardley for their excellent technical assistance. This work was supported by grants from the Medical Research Counand medially to this nucleus. Previous electrophysiological studies in the cat have cil and the National Heart Foundation of New Zealand and shown that there is a synaptic interaction between respira- the Wellcome trust (UK). P.M.P. is an Overseas Research tory neurons in the ventrolateral medulla (e.g., Segers et Fellow of the National Heart Foundation and the National al., '87; Jiang and Lipski, '90). In addition, an intracellular Health and Medical Research Council of Australia.

Fig. 7. A camera lucida reconstruction of an expiratory motoneuron located within the nucleus ambiguus pars compacta (NAc). The axon of this neuron can be seen projecting dorsomedially, before turning sharply lateral. Prominent ventrally projecting dendrites are present that terminate within 100 p,m of the ventral surface of the medulla oblongata. The chart record shows waves of hyperpolarization of

~~

membrane potential that were synchronised with the discharges of the phrenic nerve. Calibration bars are 100 p m for the reconstruction, and 1.0 mm for the inset. The time calibration bar for the chart record is 1.2 seconds. MP, membrane potential; Phr, integrated discharge of the phrenic nerve; TP, tracheal pressure.

P.M. PILOWSKY ET AL.

616

LrrERAmcrrED Agnati, L.F., K. Fuxe, M. Zoli, I. Zini, A. Harfstrand, G. Toffano, and M. Goldstein (1988) Morphometrical and microdensitometrical studies on phenylethanolamine-N-methyltransferaseand neuropeptide Y-immunoreactive neurons in the rostral medulla oblongata of the adult and old male rat. Neuroscience 26t461-478. Aguirre, J.A., R. Covenas, D. Croix, J.R. Alonso, J.A. Narvaez, G. Tramu, and S. Gonzdez-Baron (1989a) Immunocytochemical study of angiotensin-I1 fibres and cell bodies in the brainstem respiratory areas of the cat. Brain Res. 489:311-317. Aguirre, J.A., R. Covenas, M.S. David-Milner,J.R. Alonso, G. Garcia-Herdugo, and S. Gonzdez-Baron (1989b) Neuropeptide Y-like immunoreactivity in brainstem respiratory nuclei of the cat. Brain Res. Bull. 23~201-207. Andrezik, J.A., V. Chan-Palay, and S.L. Palay (1981) The nucleus paragigantocellularis lateralis in the rat. Anat. Embryol. (Berl.) 161:355-371. Arata, A., H. Onimaru, and I. Homma (1990) Respiration related neurons in the ventral medulla of newborn rats in vitro. Brain Res. Bull. 24:599604. Arita, H., N. Kogo, and N. Koshiya (1987) Morphological and physiological properties of caudal medullary expiratoly neurons of the cat. Brain Res. 401,258-266. Ballantyne, D., and D.W. Richter (1986) The non-uniform character of expiratory synaptic activity in expiratory bulbospinal neurones of the cat. J. Physiol. (Lond.) 370~433456. Barman, S.M., and G.L. Gebber (1985) Axonal projection patterns of ventrolateral medullospinal sympathoexcitatoryneurons. J. Neurophysiol. 53t1551-1566. Bieger, D., and D.A. Hopkins (1987) Viscerotopic representation of the upper alimentary tract in the medulla oblongata of the rat: The nucleus ambiguus. J. Comp. Neurol. 262546-562. Bolme, P., H. Corrodi, K. Fuxe, T. Hokfelt, P. Lidbrink, and M. Goldstein (1974) Possible involvement of central adrenaline neurons in vasomotor and respiratory control. Studies with clonidine and its interactions with piperoxane and yohimbine. Eur. J. Pharmacol. 2839-94. Bradley, P.B., and A.P. Lucy (1983) Cholinoceptive properties of respiratory neurones in the rat medulla. Neuropharmacology 22t853-858. Brown, D.L., and P.G. Guyenet (1984) Cardiovascular neurons of brain stem with projections to spinal cord. Am. J. Physiol. 247:R1009-R1016. Brown, D.L., and P.G. Guyenet (1985) Electrophysiological study of cardiovascular neurons in the rostral ventrolateral medulla in rats. Circ. Res. 56: 359-369. Chen, Z., T. Hedner, and J. Hedner (1988) Hypoventilation and apnoea induced by the substance P antagonist [D-Pro', d-Trp7W3P in the ventrolateral rat medulla. Acta Physiol. Scand. 134:153,154. Ciriello, J., and F.R. Calaresu (1977) Lateral reticular nucleus: A site of somatic and cardiovascular integration in the cat. Am. J. Physiol. 233~R100-Rl09. Ellenberger, H.H., and J.L. Feldmsn (1988) Monosynaptic transmission of respiratory drive to phrenic motoneurons from brainstem bulhospinal neurons in rats. J. Comp. Neurol. 2 6 9 ~ 4 7 5 7 . Ellenberger, H.H., W.Z. Zhan, and J.L. Feldrnan (1987) Anatomical relationship between respiratory and catecholamine neurons in ventrolateral medulla of rat SOC.Neurosci. Abs. 13,809. Ellenberger, H.H., J.L. Feldman, and W.-Z. Zhan (1990) Subnuclear organization ofthe lateral tegmental field of the rat. 11: Catecholamine neurons and the ventral respiratory group. J. Comp. Neurol. 294.212-222. Ezure, K., M. Manabe, and H. Yamada (1988) Distribution of medullary respiratory neurons in the rat. Brain Res. 455262-270. Fedorko, L., and E.G. Merrill (1984) Axonal projections from the rostral expiratory neurons of the Botzinger complex to medulla and spinal cord in the cat. J. Physiol. (Lond.)350:487496. Feldberg, W. (1976) The ventral surface of the brain stem: A scarcely explored region of pharmacological sensitivity. Neuroscience lr427-441. Fuxe, K., L.F. Agnati, A. Harfstrand, I. Zini, K. Tatemoto, E.M. Pich, T. Hokfelt, V. Mutt, and L. Terenius (1983) Central administration of neuropeptide Y induces hypotension bradypnea and EEG synchronization in the rat. Acta Physiol. Scand. 118t189-192. Gatti, P.J., W.P. Norman, A.M.T. Dasilva, and R.A. Gillis (1986) Cardiorespiratory effects produced by microinjecting L-glutamic acid into medullary nuclei associated with the ventral surface of the feline medulla. Brain Res. 381:281-288. Gilbey, M.P., Y. Numao, and K.M. Spyer (1986) Discharge patterns of cervical sympathetic preganglionic neurones related to central respiratory drive in the rat. J. Physiol. (Lond.) 378,253365.

Grace, A.A., and S.-P. Onn (1989) Morphology and electrophysiologid properties of immunocytochemically identified rat dopamine neurons recorded in vitro. J. Neurosci. 9t3463-3481. Grelot, L., A.L. Bianchi, S. Iscoe, and J.E. Remmers (1988) Expiratory neurones of the rostral medulla: Anatomical and functional correlates. Neurosci. Lett. 89~140-145. Guyenet, P.G., and B.S. Young (1987) Projections of nucleus paragigantocellularis lateralis to locus coeruleus and other structures in the rat. Brain Res. 406:171-184. Haselton, J.R., and P.G. Guyenet (1989) Electrophysiological characterization of putative C1 adrenergic neurons in the rat. Neuroscience 30:199214. Haxhiu, M.A., E.C. Deal, R.D. Trivedi, E.V. Lunteren, and N.S. Cherniak (1988) Tracheal and phrenic responses to neurotensin applied to ventral medulla. Am. J. Physiol. 255:R780-R786. Hodgson, A., J. Minson, J. Chalmers, and P. Pilowsky (1987) The use of microinjected colloidal gold and immunocytochemistry to localize pressor sites in the rostral medulla oblongata of the rat. Neurosci. Lett. 77:125-130. Hokfelt, T., K. Fuxe, M. Goldstein, and 0. Johansson (1974) Immunohistochemical evidence for the existence of adrenaline neurons in the rat brain. Brain Res. 66:235-251. Holtman, J.R., and K.A. King (1988) Regulation of respiratory motor outflow to the larynx and diaphragm by GABA receptors. Eur. J. Pharmacol. 156:181-187. Holtman, J.R., T.E. Dick, andA.J. Berger (1986)Involvement of serotonin in the excitation of phrenic motoneurons evoked by stimulation of the raphe obscurus. J. Neurosci. 6~1185-1193. Howard, B.R., and M. Tabatabai (1975) Localization of the medullary respiratory neurons in rats by microelectrode recording. J. Appl. Physiol. 39:8 12-8 17. Howe, P.R.C., M. Costa, J.B. Furness, and J.P. Chalmers (1980) Simultaneous demonstration of phenylethanolamine N-methyltransferaseimmunofluorescent and catecholamine fluorescent nerve cell bodies in the rat medulla oblongata. Neuroscience 5,2229-2238. Jiang, C., and J. Lipski (1990) Monosynaptic inhibition of respiratory neurons in the ventral respiratory group from augmenting expiratory neurons of the ipsilateral Botzinger complex. Exp. Brain Res. in press. Kalia, M., K. Fuxe, and M. Goldstein (1985a) Rat medulla oblongata. 11. Dopaminergic, noradrenergic (A1 and A2) and adrenergic neurons, nerve fibres and presumptive terminal processes. J. Comp. Neurol. 233:308332. Kalia, M., K. Fuxe, and M. Goldstein (1985b) Rat medulla oblongata. 111. Adrenergic (C1 and C2) neurons, nerve fibres and presumptive terminal processes. J. Comp. Neurol. 233:333-349. Kamiya, H., K. Itoh, Y. Yasui, T. Ino, and N. Mizuno (1988) Somatosensory and auditory relay nucleus in the rostral part of the ventrolateral medulla: A morphological study in the cat. J. Comp. Neurol. 273421435. Lipski, J. (1981) Antidromic activation of neurones as an analytic tool in the study of the central nervous system. J. Neurosci. Methods 4: 1-32. Lipski, J., H.J. Waldvogel, P.M. Pilowsky, and C. Jiang (1990) GABAimmunoreactive boutons make synapses with inspiratory neurons of the dorsal respiratory group in the cat. Brain Res. in press. Livingaton, C.A., and A.J. Berger (1989) Immunocytochemical localization of GABA neurons projecting to the ventrolateral nucleus of the solitary tract. Brain Res. 494,143-150. Lovick, T.A. (1986) Analgesia and the cardiovascular changes evoked by stimulating neurones in the ventrolateral medulla in rats. Pain 25~259268. McAllen, R.M. (1986) Location of neurones with cardiovascular and respiratory actions, at the ventral surface of the cat's medulla. Neuroscience 18r43-49. McAllen, R.R. (1987) Central respiratory modulation of subretrofacial bulbospinal neurones in the cat. J. Physiol. (Lond.)388:533-545. McCrimmon, D.R., J.C. Smith, and J.L. Feldman (1989) Involvement of excitatory amino acids in neurotransmission of inspiratory drive to spinal respiratory motoneurons. J. Neurosci. 9:1910-1921. Milner, T.A., V.M. Pickel, C. Abate, T.H. Joh, and D.J. Reis (1988) Ultrastructural characterization of substance P-like immunoieactive neurons in the rostral ventrolateral medulla in relation to neurons containing catecholamine-synthesizingenzymes. J. Comp. Neurol. 270; 427-445. Morrison, S.F., T.A. Milner, and D.J. Reis (1988) Reticulospinal vasomotor neurons of the rat rostral ventrolateral medulla: Relationship to sympa-

RAT RESPIRATORY NEURONS AND CATECHOLAMINES thetic nerve activity and the C1 adrenergic cell group. J. Neurosci. 8:1286-1301. Murakoshi, T., T. Suzue, and S. Tamai (1985) A pharmacological study in the isolated brainstem-spinal cord preparation of the newborn rat. Br. J. Pharmacol. 86:95-104. Onai, T., M. Saji, and M. Miura (1987) Projections of supraspinal structures to the phrenic motor nucleus in rats studied by a horseradish peroxidase microinjection method. J. Auton. Nerv. Syst. 21:233-239. Onimaru, H., A. Arata, and I. Homma (1988) Primary respiratory rhythm generator in the medulla of hrainstem-spinal cord preparation from newborn rats. Brain Res. 445:314-324. Onimaru, H., A. Arata, and I. Homma (1989) Firing properties of respiratory rhythm generating neurons in the absence of synaptic transmission in rat medulla in vitro. Exp. Brain Res. 76:530-536. Otake, K., H. Sasaki, H. Mannen, and K. Ezure (1987) Morphology of expiratory neurons of the Botzinger complex: An HRP study in the cat. J. Comp. Neurol. 258:565-579. Pilowsky, P.M., M.J. West, and J.P. Chalmers (1985) Renal sympathetic nerve responses to stimulation, inhibition and destruction of the ventrolateral medulla in the rabbit. Neurosci. Lett. 605-55. Pilowsky, P.M., V. Kapoor, J.B. Minson, M.J. West, and J.P. Chalmers (1986a) Spinal cord serotonin release and raised blood pressure after brainstem kainic acid injection. Brain Res. 366:354-357. Pilowsky, P.M., J.B. Minson, A.J. Hodgson, P.R.C. Howe, and J.P. Chalmers (198613) Does substance P coexist with adrenaline in neurones of the rostral ventrolateral medulla in the rat? Neurosci. Lett. 71:293-298. Pilowsky, P.M., Jiang C., and Lipski J. (1989) Cardiovascular and respiratory neurones in the rostral ventrolateral medulla oblongata of the rat. Neurosci. Lett. S34:135. Portillo, F., and R. Pasaro (1988) Location of bulbospinal neurons and of laryngeal motoneurons within the nucleus ambiguus of the rat and cat by means of retrograde fluorescent labelling. J. Anat. 159:ll-18. Prabhakar, N.R., J. Mitra, E.M. Adams, and N.S. Cherniack (1989) Involvement of ventral medullary surface in respiratory responses induced by 2,4-dinitrophenol. J. Appl. Physiol. 665984505. Rohlicek, C.V., and C. Polosa (1983) Mediation of pressor responses to cerebral ischemia by superficial ventral medullary areas. Am. J. Physiol. 245:H962-H968. Ross, C.A., D.A. Ruggiero, D.H. Park, T.H. Joh, A.F. Sved, J. FernandezPardal, J.M. Saavedra, and D.J. Reis (1984a) Tonic vasomotor control by the rostral ventrolateral medulla: Effect of electrical or chemical stimulation of the area containing C1 adrenaline neurons on arterial pressure,

617

heart rate and plasma catecholamines and vasopressin. J. Neurosci. 4:474494. Ross, C.A., D.A. Ruggiero, T.H. Joh, D.H. Park, and D.J. Reis (1984b) Rostral ventrolateral medulla: Selective projections to the thoracic autonomic cell column from the region containing C1 adrenaline neurons. J. Comp. Neurol. 226168-185. Saether, K., G. Hilaire, and R. Monteau (1987) Dorsal and ventral respiratory groups of neurons in the medulla of the rat. Brain Res. 419:87-96. Saji, M., and M. Miura (1990) Thoracic expiratory motor neurons of the rat: localization and sites of origin of their premotor neurons. Brain Res. 507:247-253. Segers, L.S., R. Shannon, S. Saporta, and B.G. Lindsey (1987) Functional associations among lateral medullary respiratory neurons in the cat. I. Evidence for excitatory and inhibitory actions of inspiratory neurons. J. Neurophysiol. 57: 1078-1100. Sun, M.-K., B.S. Young, J.T. Hackett, and P.G. Guyenet (1988a) Reticulospinal pacemaker neurons of the rat rostral ventrolateral medulla with putative sympathoexcitatory function: an intracellular study in vitro. Brain Res. 442:229-239. Sun, M.-K., B.S. Young, J.T. Hackett, and P.G. Guyenet (1988b) Rostral ventrolateral medullary neurons with intrinsic pacemaker properties are not catecholaminergic. Brain Res. 451:345-349. Suzue, T. (1984) Respiratory rhythm generation in the in vitro brain stem spinal cord preparation of the neonatal rat. J. Physiol. (Lond.) 354:173183. Terui, N., Y. Saeki, and M. Kumada (1986) Barosensory neurons in the ventrolateral medulla in rabbits and their responses to various afferent inputs from peripheral and central sources. Jpn. J. Physiol. 36:11411164. Voss, M.D., D. de Castro, J. Lipski, P.M. Pilowsky, and C. Jiang (1990) Serotonin immunoreactive boutons form close appositions with respiratory neurons of the dorsal respiratory group in the cat. J. Comp. Neurol. 295:208-218. Wilhelm, Z., S.W. Schwarzacher, K. Anders, and D.W. Richter (1990) Medullary respiratory neurones in rat. Pflugers Arch. [Suppl.] 415353. Yajima, Y., and Y.Hayashi (1989) Electrophysiologicalevidence for axonal branching of ambiguus laryngeal motoneurons. Brain Res. 478t309-314. Yamada, H., K. Ezure, and M. Manabe (1988) Efferent projections of inspiratory neurons of the ventral respiratory group. A dual labeling study in the rat. Brain Res. 455t283-294. Yamamoto, Y., M. Runold, N. Prabhakar, T. Pantaleo, and H. Lagercrantz (1988) Somatostatin in the control of respiration. Acta Physiol. Scand. 134:529-533.