- Email: [email protected]
Respiratory activity in neonatal rats
Autonomic Neuroscience: Basic and Clinical 84 (2000) 19–29 www.elsevier.com / locate / autneu
Respiratory activity in neonatal rats a b a, Mathias Dutschmann , Richard J.A. Wilson , Julian F.R. Paton * a
b
Department of Physiology, School of Medical Sciences, University of Bristol, Bristol BS8 1 TD, UK Heritage Medical Research Building, Department of Medical Physiology and Biophysics, University of Calgary, 3330 Hospital Drive N.W., Calgary, Alberta, Canada T2 N4 N1 Received 12 April 2000; accepted 17 May 2000
Abstract In neonatal animals in vitro preparations have been employed widely to study the central control of respiration. These preparations have limitations in that reflex afferent inputs and kinesiological studies cannot be performed. Here, we describe an alternative in situ experimental model for studying both peripheral and central control of the respiratory system in neonatal rats. Using technology based on adult mammals, we introduce an intra-arterially perfused working heart–brainstem preparation (WHBP) that permits studies on eupnoeic respiration in neonatal rats from within a few hours of birth. Using this preparation we demonstrate a three-phase respiratory rhythm as revealed by the activity in phrenic and recurrent laryngeal motor nerves, the respiratory modulation of laryngeal resistance and the firing patterns of respiratory neurones recorded from the ventrolateral medulla. We conclude that the neonatal rat WHBP is an in situ preparation because it produces a respiratory rhythm similar to that of adult in vivo mammal preparations but distinct from in vitro preparations. 2000 Elsevier Science B.V. All rights reserved. Keywords: Respiration; Pre-Botzinger complex; Rhythm generation; In vitro; Neonatal rat
1. Introduction A novel in situ intra-arterially perfused working heart– brainstem preparation (WHBP) of adult mouse for studying cardiorespiratory function at the systemic and cellular levels has been described (Paton, 1996a). Like other arterially perfused preparations (e.g., Hayashi et al., 1991), this model was shown to generate a ramp-inspiratory motor pattern comparable to that seen in vivo and defined by Cohen as eupnoea 1 (see Fig. 1 of Cohen, 1979). A
*Corresponding author. Tel.: 144-117-928-7818; fax: 144-117-9288923. E-mail address: [email protected] (J.F.R. Paton). 1 An internationally accepted definition for eupnoeic respiration is required. We raise the question of whether this can be defined in in vitro neonatal preparations when a single motor outflow such as either C1 discharge or hypoglossal motor activity is recorded. The problem lies in that a single respiratory motor outflow can be rhythmic under a number of behavioural situations such as coughing, hic-cupping, sneezing, swallowing, retching and vomiting. Here we wish to define eupnoeic respiration based on a number of criteria including motor pattern, frequency, inspiratory time to total respiratory cycle time, presence of early expiratory activity in cranial motor outflows and respiratory modulation of glottal muscles and the heart.
subsequent study of the in situ WHBP of adult mouse (Paton, 1996b) purported respiratory neurone types with similar membrane potential trajectories and firing patterns to those found in a variety of adult mammals in vivo (e.g., Lawson et al., 1989; Richter et al., 1975; Zheng et al., 1991). This provided validation for using the preparation to study eupnoeic respiration. We have now extended the WHBP to the neonatal rat from within a few hours of birth. Our aim was to describe the kinesiology of the respiratory system in the neonatal rat and critically compare the respiratory activity generated with that described previously in vivo and in vitro. We report that the in situ WHBP of the neonatal rat is an appropriate preparation for studying neural mechanisms controlling breathing.
2. The in situ WHBP of neonatal rat The procedures used were based on those detailed previously in adult mice (Fig. 1; Paton, 1996a). Rat pups from 1 h to 4 days old were studied as this age range is commonly used for in vitro studies. Animals were decerebrated at the precollicular level, transected below the diaphragm and perfused via the descending aorta at a flow
1566-0702 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S1566-0702( 00 )00177-6
20
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 1. Schematic of the in situ WHBP of neonatal rat. Neonatal rats between 1 h and 4 days old were decerebrated at the precollicular level, transected below the diaphragm, cerebellectomised to expose the IVth ventricle and perfused arterially via the aorta with Ringers containing ficoll at 318C. With the open chest the lungs did not inflate but the vagi remained intact. In paralysed preparations, respiratory neurones were recorded intracellularly using sharp microelectrodes (80–120 MV) placed into the ventrolateral medulla by passing through the dorsal medulla. A kinesiological approach was adopted to functionally monitor respiratory motor activity on laryngeal resistance. This was achieved by perfusing the glottis with a constant flow of warm, moistened oxygen in the expiratory direction and subglottal pressure measured. Increases and decreases in pressure indicated adduction (i.e. constriction) and abduction (dilatation) of the glottis, respectively. The oxygen exhausted through a hole cut in the pharynx. Centrally generated respiratory activity was recorded from mixed motor nerves via suction electrodes: XII, hypoglossal nerve activity; RLNA, recurrent laryngeal nerve activity; PNA, phrenic nerve activity. The ECG was recorded and from this heart rate derived. Peripheral chemoreceptors were stimulated by aortic injection of 25–75 ml of 0.03% sodium cyanide (NaCN) solution.
rate of 18–25 ml / min at 318C with Ringers containing 1.25% ficoll. Due to size restrictions we were unable to monitor perfusion pressure within the aorta. Rhythmic contractions of respiratory muscles returned with the onset of re-perfusion. Rhythmic phrenic discharges persisted in paralysed preparations for up to 5 h. The preparation set up time was 20–25 min.
2.1. Oxygenation of the medulla To utilise in situ (or in vitro) preparations for studies on the neural control of respiration, the medulla must be oxygenated adequately. To assess medullary oxygenation in the in situ WHBP, we made PO 2 depth profiles using a Clark-style microelectrode connected to a polargraphic amplifier (Fig. 2A). Tracks began from the dorsolateral surface of the medulla and extended to the ventral respiratory group. Depth profiles immediately after re-perfusion revealed the core and ventral parts of the medulla to be anoxic (Fig. 2B, filled symbols). However, after |40 min of perfusion tissue PO 2 had increased significantly (Fig.
2B, open symbols; ANOVA: P , 0.05, n 5 4) such that the core PO 2 reached 117.5629.5 (S.E.M.) Torr. Thus, the in situ WHBP of the neonatal rat is supplied with adequate oxygen for aerobic metabolism. In contrast, the in vitro brainstem–spinal cord preparation (or in vitro en bloc) has an anoxic core which may extend from 450 to 500 mm below the surface despite the imposed reduction in metabolic demand caused by the hypothermic experimental conditions (278C; e.g., Okada et al., 1993). While the cell bodies of respiratory neurons in the ventrolateral medulla may be outside the anoxic core, they are likely to have dendrites and axons within it. They may receive synaptic inputs from neurones that reside within the anoxic region. Further, anoxia and / or hypothermia have profound effects on membrane potential (see Ballanyi et al., 1996) and firing properties, both of which may alter the ability of neurones to respond appropriately to synaptic inputs and / or regulate synaptic release (Ramirez et al., 1998). Moreover, the lack of oxygen in the core of these preparations results in metabolic acidosis (Okada et al., 1993) making surrounding areas highly
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
21
Fig. 2. PO 2 profiles in an in situ WHBP of neonatal rat. Profiles were made using an O 2 -sensitive Clark-style microelectrode with a tip diameter of |20 mm. Profiles began close to the dorsolateral surface of the medulla and microelectrodes directed to intercept with the ventral respiratory group. (A) Raw data from a 4 day old rat in situ. (B) Immediately after re-perfusing the ventral medulla was anoxic (open symbols; n 5 4). However, following 40 min of perfusion, oxygen tension within the ventrolateral medulla was .100 Torr (filled symbols). During this 40 min, we presume there is a further recruitment of blood vessels which permits respiratory frequency to be maintained at approximately 46 phrenic nerve discharges per minute.
acidic. This changes ionic homeostasis which will effect synaptic transmission and possibly saturate central CO 2 / H 1 respiratory chemosensitive mechanisms. In summary, an anoxic core in the medulla could alter the function and ultimately the configuration of the respiratory network. Consequently, the in situ preparation has a distinct advantage over the in vitro preparations in studies on neural regulation of respiration.
and a rapid onset-decrementing burst which started at the cessation of inspiration and continued to decrement for at least 50% of the expiratory interval (Fig. 3). Typically, the peak amplitude of the post-inspiratory activity was greater than the peak inspiratory discharge, indicating that postinspiratory activity in rat pups is pronounced. Both the hypoglossal and recurrent laryngeal nerve activity patterns were similar to that found in WHBP of adult rats (Paton et al., 1999; Paton and Nolan, 2000).
2.2. Motor patterns recorded from phrenic, hypoglossal and recurrent laryngeal nerves (n 5 6)
2.3. Respiratory modulation of the upper airway (n 5 6)
In the in situ WHBP of neonatal rats, phrenic nerve outflow occurred at a frequency of 4664 per minute at 318C (Fig. 3), close to that found in vivo (Smith et al., 1990; Wang et al., 1996). The motor pattern included incrementing discharges of 150–200 ms duration. The inspiratory time to overall cycle length was |15% and comparable to the adult rat (Paton et al., 1999). Inspiratory related activity was also recorded from the hypoglossal nerve which could exhibit a ramping discharge (Fig. 3). The recurrent laryngeal nerve showed both an inspiratory discharge (i.e. coincident with phrenic nerve discharge)
In the in situ WHBP we employed a kinesiological approach to study the respiratory modulation of glottal resistance. We perfused the larynx with a constant stream of warmed, humidified oxygen in the expiratory direction while monitoring pressure changes sub-glottaly (Fig. 1). Increases and decreases in sub-glottal pressure were indicative of constriction and dilatation respectively, thereby giving a direct index of the dynamic changes in glottal resistance during the respiratory cycle. In neonatal rats in situ we found decreases in sub-glottal pressure during inspiration (i.e. dilatation) followed by an
22
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 3. Simultaneous recording of multiple respiratory motor outflows in a P3 rat in situ. Respiratory activity was recorded from the phrenic (PNA), recurrent laryngeal (RLNA) and hypoglossal nerves simultaneously. Each motor outflow exhibited an augmenting inspiratory discharge pattern. The RLNA displayed the three respiratory phases: inspiratory (Insp), since it coincided with PNA, post-inspiratory (Post-Insp) and expiratory (Exp). The expiratory phase is seen as quiescence in the neurogram. The post-inspiratory activity observed in the RLNA is comparable to that seen in adult rats (see Paton and Nolan, 2000), but is absent in the in vitro preparations (see text for discussion). Respiratory frequency of this preparation averaged 47 per min. The time constant of the integrator was 0.15 s.
abrupt and potent increase in pressure coinciding with the end of inspiration and start of the post-inspiratory phase (Fig. 4). The latter indicated a powerful constriction of the glottis. During the expiratory interval, sub-glottal pressure declined and reached a steady state before decreasing during the next neural inspiratory discharge. From these pressure recordings we could clearly define three phases of respiration (see Fig. 4). This ongoing pattern of laryngeal muscle activation as reflected by sub-glottal pressure changes is similar to that recently published for the adult, perfused rat (Paton and Nolan, 2000). Fine control of the upper airway musculature is important at birth since swallowing begins immediately after parturition. Since swallowing involves activation of postinspiratory mechanisms centrally (Paton et al., 1999; Shiba et al., 1999) it is not surprising that post-inspiratory activity is prevalent in the newborn rat. It was of further interest that respiratory sinus arrhythmia was present in the neonate (Figs. 4 and 6). Sinus arrhythmia may result from excitatory drive from a subclass of post-inspiratory neurones impinging on cardio-inhibitory vagal motoneurones (Richter and Spyer, 1990). This further emphasises that the basic medullary neuronal connections important for controlling the cardiorespiratory system are present at birth in the rat. This is supported by our previous finding that the baroreceptor reflex sensitivity in newborn rats is comparable to adults (Kasparov and Paton, 1997). Indeed, postinspiratory activity in neonatal rats may be higher relative
to adults. First, a rapid onset post-inspiratory constriction of the glottis is important for slowing expiratory airflow out of the lungs. In the neonatal rat, where respiratory frequency is relatively high, slowing expiration, at least transiently, would prolong the time for adequate gas exchange. Second, post-inspiratory activity is important for maintaining functional residual capacity (FRC) and preventing lung collapse. The compliance of the neonatal lung and chest wall is higher than that in the adult, which means that lung volume must be regulated very precisely. The Hering-Breuer reflex which regulates lung volume is fully functional at birth, at least in kittens (Kalia, 1976). Since the Hering-Breuer reflex inputs excite post-inspiratory neurones (Haji et al., 1999; Rybak et al., 1997) a fine balance is needed between mechanisms activating postinspiratory activity (i.e. swallowing; see Paton et al., 1999) and lung inflation to maintain FRC. In the absence of pulmonary feedback, (e.g., WHBP and the in vitro preparations), post-inspiratory activity may be prevalent.
2.4. Respiratory neurones in the in situ WHBP of neonatal rats We recorded over 264 respiratory neurones from the ventrolateral medulla of neonatal rats (Figs. 5 and 6). Post-inspiratory neurones were abundant in the ventrolateral medulla of neonatal rats. They represented by far the highest number of expiratory neurones (i.e. 82 of 130
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
23
Fig. 4. Three-phase respiratory modulation of laryngeal resistance in a WHBP of neonatal rat. Changes in pressure recorded below the larynx were monitored during constant perfusion of the upper airway with humidified oxygen in the expiratory direction (see Fig. 1 and text). The gas exited the airway through a small hole cut into the pharynx (Fig. 1). Increases and decreases in sub-glottal pressure indicated constriction and dilatation respectively. This permitted a kinesiological approach for determining the function of respiratory modulated activity in nerves innervating the glottis. As seen, the glottis dilated during neural inspiration (i.e. coincident with phrenic nerve activity; PNA), but showed an abrupt constriction during early expiration which slowly relaxed; the latter formed the post-inspiratory phase (Post-Insp; compare with Fig. 3). The expiratory phase (Exp) was determined by the steady-state pressure prior to inspiration. Note also that there was an ongoing respiratory sinus arrhythmia with a mild bradycardia coinciding with the onset of the post-inspiratory phase.
expiratory cells) and constituted almost one-third of all respiratory neurone types encountered (Fig. 6). The high numbers of post-inspiratory neurones is consistent with both previous in vivo rat studies (Schwarzacher et al., 1991), our present finding of a pronounced post-inspiratory discharge in the recurrent laryngeal nerve and the dramatic glottic constriction following inspiration (see Fig. 4). Other neurones which we encountered were pre-inspiratory, inspiratory, late inspiratory (Fig. 5), tonic expiratory, biphasic expiratory (also described as pre-inspiratory by Onimaru and Homma, 1992) and augmenting expiratory (Fig. 5) neurones; their relative proportions are shown in Fig. 6. The respiratory neurone types we found in the neonatal rat in situ are comparable to those described in a range of adult mammals (e.g., Lawson et al., 1989; Paton, 1996b; Richter et al., 1975; Zheng et al., 1991).
3. Modulating respiratory activity in the neonatal rat
3.1. Peripheral chemoreceptor activation (n 5 6) An advantage of the arterially perfused in situ neonatal rat preparation is that numerous peripheral receptors and afferents can be stimulated. Fig. 7 shows the effect of
stimulating peripheral chemoreceptors with low doses of sodium cyanide injected into the aorta. An increase in the amplitude and frequency of phrenic nerve activity concomitant with a reflex bradycardia was evoked. These data indicate that, first, both respiratory and cardiac components of the peripheral chemoreceptor reflex are functional in the newborn rat pup; second, synaptic transmission through the nucleus of the solitary tract in the WHBP of neonatal rats is preserved; third, the preparation is not generating ceiling-limit respiratory activity since both the amplitude and frequency of inspiratory activity were elevated, which is consistent with our data purporting a well perfused brainstem (see Fig. 2).
3.2. Systemic hypoxia (n 5 5) Gassing the perfusate with 5% oxygen and 5% carbon dioxide produced a biphasic pattern of response consisting of an initial augmentation in phrenic motor outflow (amplitude and frequency) followed by a depression (Fig. 8). During the depression, the motor pattern decremented consistently and remained stable at a frequency of 7–8 per min; apnoea occurred after prolonged exposure to hypoxia (i.e. .10 min). PO 2 was measured during exposure to hypoxia and fell to zero mmHg.
24
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 5. Membrane potential trajectories and firing patterns of respiratory neurones in neonatal rats in situ. Representative intracellular recordings are shown of four different respiratory neurones. Inspiratory neurones were most frequently encountered. Although more rare, late inspiratory cells were found and exhibited a characteristic hyperpolarisation prior to neural inspiration and a discharge late in inspiration. The most common expiratory neurone was the post-inspiratory type depicting a rapid membrane depolarisation and burst of action potentials coincident with the end of neural inspiration followed by a decrementing discharge pattern during the expiratory period. Note the similarity in the firing pattern and membrane potential trajectory of this neurone with the recurrent laryngeal nerve activity in Fig. 3 and the sub-glottal pressure recording in Fig. 4. Augmenting expiratory neurones were also observed. These cell types are all found in adult mammals and, with the exception of inspiratory neurones, rarely described in in vitro preparations. Solid bars represent 1 s.
3.3. Transecting at the ponto-medullary level (n 5 5) In the in situ WHBP of neonatal rats, sectioning the neuraxis at the ponto-medullary junction had a profound effect on respiratory activity. Both the frequency of rhythmic activity declined to |2 per min and the pattern was altered either exhibiting a prolonged inspiratory discharge or a decrementing pattern (Fig. 9); this is consistent with that found in vivo (see St-John, 1998, for a review). These data purport the importance of pontine structures for production of eupnoea as originally described (Lumsden, 1923). In contrast, in the en bloc or brainstem spinal cord preparations pontine transection resulted in an increase in respiratory frequency (Hilaire et al., 1989; Reckling and Feldman, 1998; Smith et al., 1990).
4. Some advantages of the WHBP In vitro preparations of neonatal rats were introduced for studying respiratory rhythm and pattern generation in the 1980s (Suzue, 1984). These now include a ‘rhythmic’ transverse slice (Smith et al., 1991), tilted-sagittal slice (Paton et al., 1994), and an en bloc and brainstem–spinal cord preparation (Onimaru and Homma, 1992; Smith et al.,
1990). All benefit from retaining at least one respiratory motor nucleus which allows an analysis of cellular and synaptic mechanisms in the context of the ‘system’. These motor outflows include hypoglossal and vagal rootlets which are monitored as an index of in vitro inspiratory activity. However, these are mixed nerves with opposing functions: in addition to the stiffening muscles in the tongue, the hypoglossal motor fibres innervate both protruder and retractor genioglossus muscles which can be activated during inspiration and expiration respectively. Similarly, vagal efferents contain fibres discharging during inspiration and expiration: laryngeal abductors, which fire with inspiration, and laryngeal adductors, which are active in the post-inspiratory phase. The in situ WHBP permits a kinesiological approach to assess the function of respiratory motor activity in mixed nerves such as the central vagus or recurrent laryngeal nerve through measurements of changes in glottal resistance, for example. This clearly is not possible in vitro and may prevent a clear distinction between inspiratory and expiratory discharges recorded from hypoglossal and vagal rootlets. A notable advantage is that peripheral reflexes such as the arterial chemoreceptor reflex are preserved. This will now permit in situ studies on the development of respiratory reflexes at the systemic and cellular levels. Since
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
25
Fig. 6. Respiratory cell types and their frequency of occurrence in the ventrolateral medulla of the WHBP of neonatal rat. From a total of 264 neurones, inspiratory neurones were most abundant (115); these were followed by: post-inspiratory neurones (82), tonic expiratory (22), augmenting expiratory (i.e. E2, 21), late-inspiratory (9), early-inspiratory (7), biphasic expiratory (5) and pre-inspiratory (3). The high number of post-inspiratory neurones is consistent with the robust post-inspiratory discharge in the recurrent laryngeal nerve and the post-inspiratory laryngeal constriction. The in situ WHBP appears distinct from in vitro preparations where post-inspiratory neurones are scarce (see text).
preparations demonstrated a respiratory sinus arrhythmia, the opportunity for studying the development of central cardiorespiratory coupling is possible. Further, unlike in vitro preparations, removal of pontine structures in the
WHBP has a dramatic effect on neural respiration which is consistent with that seen in in vivo neonatal rats (Fung and St-John, 1995). Thus, in the in situ WHBP, pontine structures appear to function and play an important role in
Fig. 7. Stimulation of the peripheral chemoreceptors in a neonatal rat in situ. Aortic injections of sodium cyanide (0.03% solution) were used to stimulate peripheral chemoreceptors. This evoked a reflex bradycardia and increases in the frequency and amplitude of phrenic nerve activity (PNA, integrated; 100 ms time constant). This suggests that homeostatic regulatory reflex pathways through the medulla are functional in the WHBP of neonatal rats.
26
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 8. Respiratory response to systemic hypoxia in a neonatal rat in situ. A continuous record of integrated phrenic nerve activity (ePNA) during exposure to 5% O 2 and 5% CO 2 (balance nitrogen) demonstrates a biphasic response: initially there was an increase in respiratory amplitude and frequency ((a) to (b): from 45 to 79 burst per minute; bpm) followed by a depression (13 discharges per minute). Note, PNA discharges persisted during the hypoxic insult but with a decrementing pattern reminiscent of a gasp. Following termination of the hypoxia, phrenic nerve activity slowly recovered its frequency and augmenting pattern similar to that seen in control (compare (d) to (a)).
Fig. 9. Ponto-medullary transection affects both respiratory pattern and frequency in neonatal rats in situ. The response in phrenic nerve activity (PNA; raw and integrated) to removal of the pons is shown in a WHBP of a P2 rat. Removal of pontine structures seriously disrupted the rate and pattern of phrenic motor outflow. The absence of the pons produced a slow, decrementing discharge. Integrator time constant 100 ms.
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
determining the frequency and pattern of respiratory motor activity during normoxia. Although the WHBP permits good pharmacological access to the brain and appears mechanically stable, the in vitro preparations permit an advantaged access to respiratory neurones for intracellular recording. In the in situ WHBP we used sharp microelectrodes and approached respiratory cells from the dorsal surface. This is not ideal as the target area cannot be visualised. Further, the sharp microelectrodes we used had high resistance (.100 MV) which limited the electrical access to the cell but prevented cell dialysis which occurs when using patch pipettes. However, we found that immature respiratory neurones did not recover consistently after cell penetration with sharp microelectrodes. Hence we are establishing a ventral approach to the medulla to permit utilisation of patch pipettes as performed in vitro. This will improve recording quality and time and allow a better electrical access to the cell for both current / voltage clamp experiments and intracellular labeling.
5. Differences between the WHBP and in vitro preparations Based on the evidence presented we propose that the respiratory activity generated by the in situ WHBP of neonatal rat is eupnoea. Since some investigators have raised the possibility that the rhythm generated in vitro is different to eupnoea (Alsop et al., 1997; Ballanyi et al., 1999; St-John, 1998; Wang et al., 1996), it seems reasonable to compare the respiratory activity generated in vitro with that in the WHBP of equivalently aged rats. A striking difference between the in situ and in vitro preparations was the relatively fast respiratory frequency and short inspiratory phase; the frequency in the WHBP was comparable to that of neonatal rats in vivo (Fung and St-John, 1995; Wang et al., 1996). Unlike that proposed by Ballanyi et al. (1999), we emphasise that this difference cannot be accounted for by a difference in pulmonary stretch receptor feedback because both preparations lack such feedback. In in vitro preparations, approximately nine discharges occur per min at 278C, but this increases to 15 per min at 378C (Smith et al., 1990). In comparison, the WHBP averaged 45 per min at 318C. Although an augmenting inspiratory discharge pattern was seen in the WHBP, this was not easily detectable every cycle due to the short burst duration. Thus, inspiratory motor pattern might not be a reliable index of eupnoea. Therefore, the most important criteria for defining eupnoea in neonatal rats is: (i) the frequency and short duration of phrenic nerve discharges (i.e. |46 per min and |200 ms duration); (ii) an inspiratory to total respiratory cycle time ratio of approximately 15%; (iii) presence of post-inspiratory discharge in the recurrent laryngeal or central vagal
27
outflow; and (iv) laryngeal dilatation and constriction during inspiration and post-inspiration respectively. Another major difference between the WHBP and in vitro is the paucity of post-inspiratory activity found in recordings from motor nerves and respiratory neurones in vitro. In the en bloc or brainstem–spinal cord preparation, both hypoglossal and vagal rootlets rarely show activity following the inspiratory burst. In cases where it is present (see Smith et al., 1990) it is unlike that seen in the in situ preparation. It is of low amplitude, highly staccato, does not decrement and appears to be inconsistent within and across studies. In central recording studies using in vitro preparations most authors fail to find post-inspiratory neurones (e.g., Brockhaus and Ballanyi, 1998; Herlenius and Lagercrantz, 1999; Johnson et al., 1994; Smith et al., 1990, 1991). When described they are either rare (i.e ,7% of all cells; Kawai et al., 1996) or very different (see Fig. 3 of Ramirez et al., 1998) to both those we described here and those seen in adult mammals in vivo. The predominant cell type found in vitro is inspiratory and the most frequently encountered expiratory neurones types were termed tonic and biphasic expiratory cells (Herlenius and Lagercrantz, 1999; Kawai et al., 1996; Onimaru and Homma, 1992; Smith et al., 1990). The dearth of post-inspiratory activity may have a major bearing on the type of respiratory rhythm being generated in vitro since post-inspiratory neurones constitute an essential element of the basic rhythm generator in vivo by providing both an irreversible inspiratory off-switch (Richter et al., 1986, 1992; Rybak et al., 1997) and inhibition of expiratory neurones (i.e. augmenting expiratory or E2 type — see Fig. 5 ‘Expiratory’; Klages et al., 1993). Since post-inspiratory neurones inhibit augmenting expiratory neurones, an absence of post-inspiratory neurones in vitro explains why internal intercostal nerves fire in the postinspiratory phase and not late expiration (see Iizuka, 1999). The in vitro finding of expiratory neurones that show no pattern (i.e. tonically fire throughout the expiratory interval) can also be accounted for by an absence of postinspiratory activity. Post-inspiratory neurones are also essential for control of glottal musculature, which is functional at birth (see Harding, 1984, for a review) and causes inspiratory opening and early expiratory constriction, as well as for basic behaviours such as vocalisation and swallowing. This raises controversial questions of whether activity generated in vitro is ‘breathing’ (for recent viewpoints, see Cohen, 1999; Koshiya and Smith, 1999; Reckling and Feldman, 1998; St-John, 1998) and whether basic mechanisms of eupnoeic rhythm generation can be studied in the in vitro preparations. Clearly a form of rhythm generation is maintained in vitro, but whether it would permit ventilation is unclear. The pacemaker or pacemaker-hybrid system may form a component of the rhythm generator which when embedded in the intact perfused brain is superceded by network properties which alone are sufficient to drive eupnoea (e.g., Rybak et al.,
28
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
1997). A possibility suggested previously is that pacemaker activity produces a central apnoeic rhythm which provides a protective function to ‘kick-start’ inspiration following an extreme insult of hypoxia or ischaemia (see Paton, 1997).
6. Conclusions The in situ WHBP has permitted a thorough description of the respiratory activity generated by a well oxygenated medulla in the neonatal rat. We suggest that the WHBP offers an alternative preparation to study the neurogenesis of eupnoea in an immature mammalian respiratory network. This in situ approach also provides opportunities to study the ontogenesis of reflexes regulating respiration, reflexes altering cardiac function and central cardiorespiratory coupling. The pattern, frequency and central organisation of respiratory activity generated in the arterially perfused in situ WHBP of 1 h to 4 day old rats is distinct to in vitro rhythmic preparations of animals of the same age. The respiratory system of the WHBP of rat pups is more akin to the in vivo adult rat; this conclusion is similar to that of Wang et al. (1996) who showed that eupnoea in vagotomised pups resembles closely that of the vagotomised adult preparations. We must now define whether in vitro preparations are generating eupnoea and interpret data from these preparations accordingly (see Remmers, 1998). A wrong interpretation concerning the central mechanisms governing breathing is not only scientifically inaccurate but potentially dangerous clinically.
Acknowledgements We would like to thank Dr. Thomas E. Dick and Professor John E. Remmers for their helpful suggestions concerning the manuscript. MD was supported by a travel fellowship from the Deutsche Forschungsgemeinschaft. RJAW is supported by a Parker B. Francis Fellowship in Pulmonary Research and was in receipt of a Wellcome Trust travel fellowship. JFRP is supported by the British Heart Foundation (BS / 93003).
References Alsop, S., Bhatia, A., Paton, J.F.R., 1997. Characteristics of the respiratory network in an arterially perfused preparation of the neonatal rat. J. Physiol. 504P, 200P. Ballanyi, K., Doutheil, J., Brockhaus, J., 1996. Membrane potentials and microenvironment of rat dorsal vagal cells in vitro during energy depletion. J. Physiol. 495, 769–784. Ballanyi, K., Onimaru, H., Homma, I., 1999. Respiratory network in the isolated brainstem–spinal cord of newborn rats. Prog. Neurobiol. 59, 583–634.
Brockhaus, J., Ballanyi, K., 1998. Synaptic inhibition in the isolated respiratory network of neonatal rats. Eur. J. Neurosci. 10. Cohen, M.I., 1979. Neurogenesis of respiratory rhythm in the mammal. Physiol. Rev. 59, 1105–1173. Cohen, P., 1999. Cells with rhythm keep us breathing. New Scientist 286, 1566. Fung, M.-L., St-John, W.M., 1995. The functional expression of a pontine pneumotaxic centre in neonatal rats. J. Physiol. 489 (2), 579–591. Haji, A., Okazaki, M., Takeda, R., 1999. GABA(A) receptor-mediated inspiratory termination evoked by vagal stimulation in decerebrate cats. Neuropharmacol. 38, 1261–1272. Harding, R., 1984. Function of the larynx in the fetus and newborn. Annu. Rev. Physiol. 46, 645–659. Hayashi, F., Jiang, C., Lipski, J., 1991. Intracellular recording from respiratory neurones in the perfused ‘in situ’ rat brain. J. Neurosci. Methods 36, 63–70. Herlenius, E., Lagercrantz, H., 1999. Adenosinergic modulation of respiratory neurones in the neonatal rat brainstem in vitro. J. Physiol. 518 (1), 159–172. Hilaire, G., Monteau, R., Errchidi, S., 1989. Possible modulation of the medullary respiratory rhythm generator by noradrenergic A5 area: an in vitro study in the newborn rat. Brain Res. 485, 325–332. Iizuka, M., 1999. Intercostal expiratory activity in an in vitro brainstem– spinal cord–rib preparation from the neonatal rat. J. Physiol. 520 (1), 293–302. Johnson, S.M., Smith, J.C., Funk, G.D., Feldman, J.L., 1994. Pacemaker properties of respiratory neurons in medullar slices from neonatal rats. J. Neurophysiol. 72, 2598–2608. Kalia, M., 1976. Visceral and somatic reflexes produced by J pulmonary receptors in new born kittens. J. Appl. Physiol. 41, 1–6. Kasparov, S., Paton, J.F.R., 1997. Changes in baroreceptor vagal reflex ¨ Arch. 434, 438–444. performance in the developing rat. Pflug. Kawai, A., Ballantyne, D., Muckenhoff, K., Scheid, P., 1996. Chemosensitive medullary neurones in the brainstem–spinal cord preparation of the neonatal rat. J. Physiol. 492 (1), 277–292. Klages, S., Bellingham, M.C., Richter, D.W., 1993. Late expiratory inhibition of stage 2 expiratory neurons in the cat — a correlate of expiratory termination. J. Neurophysiol. 70, 1307–1315. Koshiya, N., Smith, J.C., 1999. Neuronal pacemaker for breathing visualized in vitro. Nature 400, 360–363. Lawson, E.E., Richter, D.W., Bischoff, A., 1989. Intracellular recordings of medullary respiratory neurons in the lateral medulla of piglets. J. Appl. Physiol. 66, 983–988. Lumsden, T., 1923. Observations on the respiratory centres in the cat. J. Physiol. 57, 153–160. Okada, Y., Muckenhoff, K., Holtermann, G., Acker, H., Scheid, P., 1993. Depth profiles of pH and PO 2 in the isolated brain stem–spinal cord of the neonatal rat. Respir. Physiol. 93, 315–326. Onimaru, H., Homma, I., 1992. Whole cell recodings from respiratory neurons in the medulla of brainstem–spinal cord preparations isolated ¨ Arch. 420, 399–406. from newborn rats. Pflug. Paton, J.F.R., 1996a. A working heart–brainstem preparation of the mouse. J. Neurosci. Methods 65, 63–68. Paton, J.F.R., 1996b. The ventral respiratory network of the mature mouse studied in a working heart brainstem preparation. J. Physiol. 493 (3), 819–831. Paton, J.F.R., 1997. Rhythmic bursting of pre- and post-inspiratory neurones during central apnoea in mature mice. J. Physiol. 502 (3), 623–639. Paton, J.F.R., Nolan, P.J., 2000. Parallelism in reflex control of laryngeal and cardiac vagal motor neurones. Respir. Physiol. (in press). Paton, J.F.R., Li, Y.-W., Kasparov, S., 1999. Reflex response and convergence of pharyngoesophageal and peripheral chemo-receptors in the nucleus of the solitary tract. Neuroscience 93, 143–154. Paton, J.F.R., Ramirez, J.-M., Richter, D.W., 1994. Functionally intact in vitro preparation generating respiratory activity in neonatal and ¨ Arch. 428, 250–260. mature mammals. Pflug.
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29 Ramirez, J.M., Quellmalz, U.J.A., Wilken, B., Richter, D.W., 1998. The hypoxic response of neurones within the in vitro mammalian respiratory network. J. Physiol. 507 (2), 571–582. Reckling, J.C., Feldman, J.L., 1998. Prebotzinger complex and pacemaker neurons: hypothesized site and kernel for respiratory rhythm generation. Annu. Rev. Physiol. 60, 385–405. Remmers, J.E., 1998. Central neural control of breathing. In: Altose, M., Kawakami, Y. (Eds.), Control of Breathing in Health and Disease. Marcel Dekker, New York. Richter, D.W., Ballantyne, D., Remmers, J.E., 1986. How is the respiratory rhythm generated? A model. NIPS 1, 109–112. Richter, D.W., Ballanyi, K., Schwarzacher, S.W., 1992. Mechanisms of respiratory rhythm generation. Curr. Opin. Neurobiol. 2, 788–793. Richter, D.W., Heyde, F., Gabriel, M., 1975. Intracellular recordings from different types of medullary respiratory neurons of the cat. J. Neurophysiol. 38, 1162–1171. Richter, D.W., Spyer, K.M., 1990. Cardiorespiratory control. In: Lowey, A.D., Spyer, K.M. (Eds.), Central Regulation of Autonomic Functions. Oxford University Press, New York, pp. 189–207. Rybak, I.A., Paton, J.F.R., Schwaber, J.S., 1997. Modelling neural mechanisms for genesis of respiratory rhythm and pattern: II. Network models of the central respiratory pattern generator. J. Neurophysiol. 77, 2007–2026.
29
Schwarzacher, S.W., Wilhelm, Z., Anders, K., Richter, D.W., 1991. The medullary respiratory network in the rat. J. Physiol. 435, 631–644. Shiba, K., Satoh, I., Kobayashi, N., Hatashi, F., 1999. Multifunctional laryngeal motoneurons: an intracellular study in the cat. J. Neurosci. 19, 2717–2727. Smith, J.C., Ellenberger, H.H., Ballanyi, K., Richter, D.W., Feldman, J.L., ¨ 1991. Pre-Botzinger complex: a brainstem region that may generate respiratory rhythm in mammals. Science 254, 726–729. Smith, J.C., Greer, J.L., Liu, G., Feldman, J.L., 1990. Neural mechanisms generating respiratory pattern in mammalian brainstem–spinal cord in vitro. I. Spatiotemporal patterns of motor and medullar neuron activity. J. Neurophysiol. 64, 1149–1169. St-John, W.M., 1998. Neurogenesis of patterns of automatic ventilatory activity. Prog. Neurobiol. 56, 97–117. Suzue, T., 1984. Respiratory rhythm generation in the in vitro brainstem– spinal cord preparation of neonatal rat. J. Physiol. 354, 173–183. Wang, W., Fung, M.-L., Darnall, R.A., St-John, W.M., 1996. Characterizations and comparisons of eupnoea and gasping in neonatal rats. J. Physiol. 490 (1), 272–292. Zheng, Y., Barillot, J.C., Bianchi, A.L., 1991. Patterns of membrane potentials and distributions of the medullary respiratory neurons in the decerebrate rat. Brain Res. 546, 261–270.
Respiratory activity in neonatal rats a b a, Mathias Dutschmann , Richard J.A. Wilson , Julian F.R. Paton * a
b
Department of Physiology, School of Medical Sciences, University of Bristol, Bristol BS8 1 TD, UK Heritage Medical Research Building, Department of Medical Physiology and Biophysics, University of Calgary, 3330 Hospital Drive N.W., Calgary, Alberta, Canada T2 N4 N1 Received 12 April 2000; accepted 17 May 2000
Abstract In neonatal animals in vitro preparations have been employed widely to study the central control of respiration. These preparations have limitations in that reflex afferent inputs and kinesiological studies cannot be performed. Here, we describe an alternative in situ experimental model for studying both peripheral and central control of the respiratory system in neonatal rats. Using technology based on adult mammals, we introduce an intra-arterially perfused working heart–brainstem preparation (WHBP) that permits studies on eupnoeic respiration in neonatal rats from within a few hours of birth. Using this preparation we demonstrate a three-phase respiratory rhythm as revealed by the activity in phrenic and recurrent laryngeal motor nerves, the respiratory modulation of laryngeal resistance and the firing patterns of respiratory neurones recorded from the ventrolateral medulla. We conclude that the neonatal rat WHBP is an in situ preparation because it produces a respiratory rhythm similar to that of adult in vivo mammal preparations but distinct from in vitro preparations. 2000 Elsevier Science B.V. All rights reserved. Keywords: Respiration; Pre-Botzinger complex; Rhythm generation; In vitro; Neonatal rat
1. Introduction A novel in situ intra-arterially perfused working heart– brainstem preparation (WHBP) of adult mouse for studying cardiorespiratory function at the systemic and cellular levels has been described (Paton, 1996a). Like other arterially perfused preparations (e.g., Hayashi et al., 1991), this model was shown to generate a ramp-inspiratory motor pattern comparable to that seen in vivo and defined by Cohen as eupnoea 1 (see Fig. 1 of Cohen, 1979). A
*Corresponding author. Tel.: 144-117-928-7818; fax: 144-117-9288923. E-mail address: [email protected] (J.F.R. Paton). 1 An internationally accepted definition for eupnoeic respiration is required. We raise the question of whether this can be defined in in vitro neonatal preparations when a single motor outflow such as either C1 discharge or hypoglossal motor activity is recorded. The problem lies in that a single respiratory motor outflow can be rhythmic under a number of behavioural situations such as coughing, hic-cupping, sneezing, swallowing, retching and vomiting. Here we wish to define eupnoeic respiration based on a number of criteria including motor pattern, frequency, inspiratory time to total respiratory cycle time, presence of early expiratory activity in cranial motor outflows and respiratory modulation of glottal muscles and the heart.
subsequent study of the in situ WHBP of adult mouse (Paton, 1996b) purported respiratory neurone types with similar membrane potential trajectories and firing patterns to those found in a variety of adult mammals in vivo (e.g., Lawson et al., 1989; Richter et al., 1975; Zheng et al., 1991). This provided validation for using the preparation to study eupnoeic respiration. We have now extended the WHBP to the neonatal rat from within a few hours of birth. Our aim was to describe the kinesiology of the respiratory system in the neonatal rat and critically compare the respiratory activity generated with that described previously in vivo and in vitro. We report that the in situ WHBP of the neonatal rat is an appropriate preparation for studying neural mechanisms controlling breathing.
2. The in situ WHBP of neonatal rat The procedures used were based on those detailed previously in adult mice (Fig. 1; Paton, 1996a). Rat pups from 1 h to 4 days old were studied as this age range is commonly used for in vitro studies. Animals were decerebrated at the precollicular level, transected below the diaphragm and perfused via the descending aorta at a flow
1566-0702 / 00 / $ – see front matter 2000 Elsevier Science B.V. All rights reserved. PII: S1566-0702( 00 )00177-6
20
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 1. Schematic of the in situ WHBP of neonatal rat. Neonatal rats between 1 h and 4 days old were decerebrated at the precollicular level, transected below the diaphragm, cerebellectomised to expose the IVth ventricle and perfused arterially via the aorta with Ringers containing ficoll at 318C. With the open chest the lungs did not inflate but the vagi remained intact. In paralysed preparations, respiratory neurones were recorded intracellularly using sharp microelectrodes (80–120 MV) placed into the ventrolateral medulla by passing through the dorsal medulla. A kinesiological approach was adopted to functionally monitor respiratory motor activity on laryngeal resistance. This was achieved by perfusing the glottis with a constant flow of warm, moistened oxygen in the expiratory direction and subglottal pressure measured. Increases and decreases in pressure indicated adduction (i.e. constriction) and abduction (dilatation) of the glottis, respectively. The oxygen exhausted through a hole cut in the pharynx. Centrally generated respiratory activity was recorded from mixed motor nerves via suction electrodes: XII, hypoglossal nerve activity; RLNA, recurrent laryngeal nerve activity; PNA, phrenic nerve activity. The ECG was recorded and from this heart rate derived. Peripheral chemoreceptors were stimulated by aortic injection of 25–75 ml of 0.03% sodium cyanide (NaCN) solution.
rate of 18–25 ml / min at 318C with Ringers containing 1.25% ficoll. Due to size restrictions we were unable to monitor perfusion pressure within the aorta. Rhythmic contractions of respiratory muscles returned with the onset of re-perfusion. Rhythmic phrenic discharges persisted in paralysed preparations for up to 5 h. The preparation set up time was 20–25 min.
2.1. Oxygenation of the medulla To utilise in situ (or in vitro) preparations for studies on the neural control of respiration, the medulla must be oxygenated adequately. To assess medullary oxygenation in the in situ WHBP, we made PO 2 depth profiles using a Clark-style microelectrode connected to a polargraphic amplifier (Fig. 2A). Tracks began from the dorsolateral surface of the medulla and extended to the ventral respiratory group. Depth profiles immediately after re-perfusion revealed the core and ventral parts of the medulla to be anoxic (Fig. 2B, filled symbols). However, after |40 min of perfusion tissue PO 2 had increased significantly (Fig.
2B, open symbols; ANOVA: P , 0.05, n 5 4) such that the core PO 2 reached 117.5629.5 (S.E.M.) Torr. Thus, the in situ WHBP of the neonatal rat is supplied with adequate oxygen for aerobic metabolism. In contrast, the in vitro brainstem–spinal cord preparation (or in vitro en bloc) has an anoxic core which may extend from 450 to 500 mm below the surface despite the imposed reduction in metabolic demand caused by the hypothermic experimental conditions (278C; e.g., Okada et al., 1993). While the cell bodies of respiratory neurons in the ventrolateral medulla may be outside the anoxic core, they are likely to have dendrites and axons within it. They may receive synaptic inputs from neurones that reside within the anoxic region. Further, anoxia and / or hypothermia have profound effects on membrane potential (see Ballanyi et al., 1996) and firing properties, both of which may alter the ability of neurones to respond appropriately to synaptic inputs and / or regulate synaptic release (Ramirez et al., 1998). Moreover, the lack of oxygen in the core of these preparations results in metabolic acidosis (Okada et al., 1993) making surrounding areas highly
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
21
Fig. 2. PO 2 profiles in an in situ WHBP of neonatal rat. Profiles were made using an O 2 -sensitive Clark-style microelectrode with a tip diameter of |20 mm. Profiles began close to the dorsolateral surface of the medulla and microelectrodes directed to intercept with the ventral respiratory group. (A) Raw data from a 4 day old rat in situ. (B) Immediately after re-perfusing the ventral medulla was anoxic (open symbols; n 5 4). However, following 40 min of perfusion, oxygen tension within the ventrolateral medulla was .100 Torr (filled symbols). During this 40 min, we presume there is a further recruitment of blood vessels which permits respiratory frequency to be maintained at approximately 46 phrenic nerve discharges per minute.
acidic. This changes ionic homeostasis which will effect synaptic transmission and possibly saturate central CO 2 / H 1 respiratory chemosensitive mechanisms. In summary, an anoxic core in the medulla could alter the function and ultimately the configuration of the respiratory network. Consequently, the in situ preparation has a distinct advantage over the in vitro preparations in studies on neural regulation of respiration.
and a rapid onset-decrementing burst which started at the cessation of inspiration and continued to decrement for at least 50% of the expiratory interval (Fig. 3). Typically, the peak amplitude of the post-inspiratory activity was greater than the peak inspiratory discharge, indicating that postinspiratory activity in rat pups is pronounced. Both the hypoglossal and recurrent laryngeal nerve activity patterns were similar to that found in WHBP of adult rats (Paton et al., 1999; Paton and Nolan, 2000).
2.2. Motor patterns recorded from phrenic, hypoglossal and recurrent laryngeal nerves (n 5 6)
2.3. Respiratory modulation of the upper airway (n 5 6)
In the in situ WHBP of neonatal rats, phrenic nerve outflow occurred at a frequency of 4664 per minute at 318C (Fig. 3), close to that found in vivo (Smith et al., 1990; Wang et al., 1996). The motor pattern included incrementing discharges of 150–200 ms duration. The inspiratory time to overall cycle length was |15% and comparable to the adult rat (Paton et al., 1999). Inspiratory related activity was also recorded from the hypoglossal nerve which could exhibit a ramping discharge (Fig. 3). The recurrent laryngeal nerve showed both an inspiratory discharge (i.e. coincident with phrenic nerve discharge)
In the in situ WHBP we employed a kinesiological approach to study the respiratory modulation of glottal resistance. We perfused the larynx with a constant stream of warmed, humidified oxygen in the expiratory direction while monitoring pressure changes sub-glottaly (Fig. 1). Increases and decreases in sub-glottal pressure were indicative of constriction and dilatation respectively, thereby giving a direct index of the dynamic changes in glottal resistance during the respiratory cycle. In neonatal rats in situ we found decreases in sub-glottal pressure during inspiration (i.e. dilatation) followed by an
22
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 3. Simultaneous recording of multiple respiratory motor outflows in a P3 rat in situ. Respiratory activity was recorded from the phrenic (PNA), recurrent laryngeal (RLNA) and hypoglossal nerves simultaneously. Each motor outflow exhibited an augmenting inspiratory discharge pattern. The RLNA displayed the three respiratory phases: inspiratory (Insp), since it coincided with PNA, post-inspiratory (Post-Insp) and expiratory (Exp). The expiratory phase is seen as quiescence in the neurogram. The post-inspiratory activity observed in the RLNA is comparable to that seen in adult rats (see Paton and Nolan, 2000), but is absent in the in vitro preparations (see text for discussion). Respiratory frequency of this preparation averaged 47 per min. The time constant of the integrator was 0.15 s.
abrupt and potent increase in pressure coinciding with the end of inspiration and start of the post-inspiratory phase (Fig. 4). The latter indicated a powerful constriction of the glottis. During the expiratory interval, sub-glottal pressure declined and reached a steady state before decreasing during the next neural inspiratory discharge. From these pressure recordings we could clearly define three phases of respiration (see Fig. 4). This ongoing pattern of laryngeal muscle activation as reflected by sub-glottal pressure changes is similar to that recently published for the adult, perfused rat (Paton and Nolan, 2000). Fine control of the upper airway musculature is important at birth since swallowing begins immediately after parturition. Since swallowing involves activation of postinspiratory mechanisms centrally (Paton et al., 1999; Shiba et al., 1999) it is not surprising that post-inspiratory activity is prevalent in the newborn rat. It was of further interest that respiratory sinus arrhythmia was present in the neonate (Figs. 4 and 6). Sinus arrhythmia may result from excitatory drive from a subclass of post-inspiratory neurones impinging on cardio-inhibitory vagal motoneurones (Richter and Spyer, 1990). This further emphasises that the basic medullary neuronal connections important for controlling the cardiorespiratory system are present at birth in the rat. This is supported by our previous finding that the baroreceptor reflex sensitivity in newborn rats is comparable to adults (Kasparov and Paton, 1997). Indeed, postinspiratory activity in neonatal rats may be higher relative
to adults. First, a rapid onset post-inspiratory constriction of the glottis is important for slowing expiratory airflow out of the lungs. In the neonatal rat, where respiratory frequency is relatively high, slowing expiration, at least transiently, would prolong the time for adequate gas exchange. Second, post-inspiratory activity is important for maintaining functional residual capacity (FRC) and preventing lung collapse. The compliance of the neonatal lung and chest wall is higher than that in the adult, which means that lung volume must be regulated very precisely. The Hering-Breuer reflex which regulates lung volume is fully functional at birth, at least in kittens (Kalia, 1976). Since the Hering-Breuer reflex inputs excite post-inspiratory neurones (Haji et al., 1999; Rybak et al., 1997) a fine balance is needed between mechanisms activating postinspiratory activity (i.e. swallowing; see Paton et al., 1999) and lung inflation to maintain FRC. In the absence of pulmonary feedback, (e.g., WHBP and the in vitro preparations), post-inspiratory activity may be prevalent.
2.4. Respiratory neurones in the in situ WHBP of neonatal rats We recorded over 264 respiratory neurones from the ventrolateral medulla of neonatal rats (Figs. 5 and 6). Post-inspiratory neurones were abundant in the ventrolateral medulla of neonatal rats. They represented by far the highest number of expiratory neurones (i.e. 82 of 130
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
23
Fig. 4. Three-phase respiratory modulation of laryngeal resistance in a WHBP of neonatal rat. Changes in pressure recorded below the larynx were monitored during constant perfusion of the upper airway with humidified oxygen in the expiratory direction (see Fig. 1 and text). The gas exited the airway through a small hole cut into the pharynx (Fig. 1). Increases and decreases in sub-glottal pressure indicated constriction and dilatation respectively. This permitted a kinesiological approach for determining the function of respiratory modulated activity in nerves innervating the glottis. As seen, the glottis dilated during neural inspiration (i.e. coincident with phrenic nerve activity; PNA), but showed an abrupt constriction during early expiration which slowly relaxed; the latter formed the post-inspiratory phase (Post-Insp; compare with Fig. 3). The expiratory phase (Exp) was determined by the steady-state pressure prior to inspiration. Note also that there was an ongoing respiratory sinus arrhythmia with a mild bradycardia coinciding with the onset of the post-inspiratory phase.
expiratory cells) and constituted almost one-third of all respiratory neurone types encountered (Fig. 6). The high numbers of post-inspiratory neurones is consistent with both previous in vivo rat studies (Schwarzacher et al., 1991), our present finding of a pronounced post-inspiratory discharge in the recurrent laryngeal nerve and the dramatic glottic constriction following inspiration (see Fig. 4). Other neurones which we encountered were pre-inspiratory, inspiratory, late inspiratory (Fig. 5), tonic expiratory, biphasic expiratory (also described as pre-inspiratory by Onimaru and Homma, 1992) and augmenting expiratory (Fig. 5) neurones; their relative proportions are shown in Fig. 6. The respiratory neurone types we found in the neonatal rat in situ are comparable to those described in a range of adult mammals (e.g., Lawson et al., 1989; Paton, 1996b; Richter et al., 1975; Zheng et al., 1991).
3. Modulating respiratory activity in the neonatal rat
3.1. Peripheral chemoreceptor activation (n 5 6) An advantage of the arterially perfused in situ neonatal rat preparation is that numerous peripheral receptors and afferents can be stimulated. Fig. 7 shows the effect of
stimulating peripheral chemoreceptors with low doses of sodium cyanide injected into the aorta. An increase in the amplitude and frequency of phrenic nerve activity concomitant with a reflex bradycardia was evoked. These data indicate that, first, both respiratory and cardiac components of the peripheral chemoreceptor reflex are functional in the newborn rat pup; second, synaptic transmission through the nucleus of the solitary tract in the WHBP of neonatal rats is preserved; third, the preparation is not generating ceiling-limit respiratory activity since both the amplitude and frequency of inspiratory activity were elevated, which is consistent with our data purporting a well perfused brainstem (see Fig. 2).
3.2. Systemic hypoxia (n 5 5) Gassing the perfusate with 5% oxygen and 5% carbon dioxide produced a biphasic pattern of response consisting of an initial augmentation in phrenic motor outflow (amplitude and frequency) followed by a depression (Fig. 8). During the depression, the motor pattern decremented consistently and remained stable at a frequency of 7–8 per min; apnoea occurred after prolonged exposure to hypoxia (i.e. .10 min). PO 2 was measured during exposure to hypoxia and fell to zero mmHg.
24
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 5. Membrane potential trajectories and firing patterns of respiratory neurones in neonatal rats in situ. Representative intracellular recordings are shown of four different respiratory neurones. Inspiratory neurones were most frequently encountered. Although more rare, late inspiratory cells were found and exhibited a characteristic hyperpolarisation prior to neural inspiration and a discharge late in inspiration. The most common expiratory neurone was the post-inspiratory type depicting a rapid membrane depolarisation and burst of action potentials coincident with the end of neural inspiration followed by a decrementing discharge pattern during the expiratory period. Note the similarity in the firing pattern and membrane potential trajectory of this neurone with the recurrent laryngeal nerve activity in Fig. 3 and the sub-glottal pressure recording in Fig. 4. Augmenting expiratory neurones were also observed. These cell types are all found in adult mammals and, with the exception of inspiratory neurones, rarely described in in vitro preparations. Solid bars represent 1 s.
3.3. Transecting at the ponto-medullary level (n 5 5) In the in situ WHBP of neonatal rats, sectioning the neuraxis at the ponto-medullary junction had a profound effect on respiratory activity. Both the frequency of rhythmic activity declined to |2 per min and the pattern was altered either exhibiting a prolonged inspiratory discharge or a decrementing pattern (Fig. 9); this is consistent with that found in vivo (see St-John, 1998, for a review). These data purport the importance of pontine structures for production of eupnoea as originally described (Lumsden, 1923). In contrast, in the en bloc or brainstem spinal cord preparations pontine transection resulted in an increase in respiratory frequency (Hilaire et al., 1989; Reckling and Feldman, 1998; Smith et al., 1990).
4. Some advantages of the WHBP In vitro preparations of neonatal rats were introduced for studying respiratory rhythm and pattern generation in the 1980s (Suzue, 1984). These now include a ‘rhythmic’ transverse slice (Smith et al., 1991), tilted-sagittal slice (Paton et al., 1994), and an en bloc and brainstem–spinal cord preparation (Onimaru and Homma, 1992; Smith et al.,
1990). All benefit from retaining at least one respiratory motor nucleus which allows an analysis of cellular and synaptic mechanisms in the context of the ‘system’. These motor outflows include hypoglossal and vagal rootlets which are monitored as an index of in vitro inspiratory activity. However, these are mixed nerves with opposing functions: in addition to the stiffening muscles in the tongue, the hypoglossal motor fibres innervate both protruder and retractor genioglossus muscles which can be activated during inspiration and expiration respectively. Similarly, vagal efferents contain fibres discharging during inspiration and expiration: laryngeal abductors, which fire with inspiration, and laryngeal adductors, which are active in the post-inspiratory phase. The in situ WHBP permits a kinesiological approach to assess the function of respiratory motor activity in mixed nerves such as the central vagus or recurrent laryngeal nerve through measurements of changes in glottal resistance, for example. This clearly is not possible in vitro and may prevent a clear distinction between inspiratory and expiratory discharges recorded from hypoglossal and vagal rootlets. A notable advantage is that peripheral reflexes such as the arterial chemoreceptor reflex are preserved. This will now permit in situ studies on the development of respiratory reflexes at the systemic and cellular levels. Since
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
25
Fig. 6. Respiratory cell types and their frequency of occurrence in the ventrolateral medulla of the WHBP of neonatal rat. From a total of 264 neurones, inspiratory neurones were most abundant (115); these were followed by: post-inspiratory neurones (82), tonic expiratory (22), augmenting expiratory (i.e. E2, 21), late-inspiratory (9), early-inspiratory (7), biphasic expiratory (5) and pre-inspiratory (3). The high number of post-inspiratory neurones is consistent with the robust post-inspiratory discharge in the recurrent laryngeal nerve and the post-inspiratory laryngeal constriction. The in situ WHBP appears distinct from in vitro preparations where post-inspiratory neurones are scarce (see text).
preparations demonstrated a respiratory sinus arrhythmia, the opportunity for studying the development of central cardiorespiratory coupling is possible. Further, unlike in vitro preparations, removal of pontine structures in the
WHBP has a dramatic effect on neural respiration which is consistent with that seen in in vivo neonatal rats (Fung and St-John, 1995). Thus, in the in situ WHBP, pontine structures appear to function and play an important role in
Fig. 7. Stimulation of the peripheral chemoreceptors in a neonatal rat in situ. Aortic injections of sodium cyanide (0.03% solution) were used to stimulate peripheral chemoreceptors. This evoked a reflex bradycardia and increases in the frequency and amplitude of phrenic nerve activity (PNA, integrated; 100 ms time constant). This suggests that homeostatic regulatory reflex pathways through the medulla are functional in the WHBP of neonatal rats.
26
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
Fig. 8. Respiratory response to systemic hypoxia in a neonatal rat in situ. A continuous record of integrated phrenic nerve activity (ePNA) during exposure to 5% O 2 and 5% CO 2 (balance nitrogen) demonstrates a biphasic response: initially there was an increase in respiratory amplitude and frequency ((a) to (b): from 45 to 79 burst per minute; bpm) followed by a depression (13 discharges per minute). Note, PNA discharges persisted during the hypoxic insult but with a decrementing pattern reminiscent of a gasp. Following termination of the hypoxia, phrenic nerve activity slowly recovered its frequency and augmenting pattern similar to that seen in control (compare (d) to (a)).
Fig. 9. Ponto-medullary transection affects both respiratory pattern and frequency in neonatal rats in situ. The response in phrenic nerve activity (PNA; raw and integrated) to removal of the pons is shown in a WHBP of a P2 rat. Removal of pontine structures seriously disrupted the rate and pattern of phrenic motor outflow. The absence of the pons produced a slow, decrementing discharge. Integrator time constant 100 ms.
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
determining the frequency and pattern of respiratory motor activity during normoxia. Although the WHBP permits good pharmacological access to the brain and appears mechanically stable, the in vitro preparations permit an advantaged access to respiratory neurones for intracellular recording. In the in situ WHBP we used sharp microelectrodes and approached respiratory cells from the dorsal surface. This is not ideal as the target area cannot be visualised. Further, the sharp microelectrodes we used had high resistance (.100 MV) which limited the electrical access to the cell but prevented cell dialysis which occurs when using patch pipettes. However, we found that immature respiratory neurones did not recover consistently after cell penetration with sharp microelectrodes. Hence we are establishing a ventral approach to the medulla to permit utilisation of patch pipettes as performed in vitro. This will improve recording quality and time and allow a better electrical access to the cell for both current / voltage clamp experiments and intracellular labeling.
5. Differences between the WHBP and in vitro preparations Based on the evidence presented we propose that the respiratory activity generated by the in situ WHBP of neonatal rat is eupnoea. Since some investigators have raised the possibility that the rhythm generated in vitro is different to eupnoea (Alsop et al., 1997; Ballanyi et al., 1999; St-John, 1998; Wang et al., 1996), it seems reasonable to compare the respiratory activity generated in vitro with that in the WHBP of equivalently aged rats. A striking difference between the in situ and in vitro preparations was the relatively fast respiratory frequency and short inspiratory phase; the frequency in the WHBP was comparable to that of neonatal rats in vivo (Fung and St-John, 1995; Wang et al., 1996). Unlike that proposed by Ballanyi et al. (1999), we emphasise that this difference cannot be accounted for by a difference in pulmonary stretch receptor feedback because both preparations lack such feedback. In in vitro preparations, approximately nine discharges occur per min at 278C, but this increases to 15 per min at 378C (Smith et al., 1990). In comparison, the WHBP averaged 45 per min at 318C. Although an augmenting inspiratory discharge pattern was seen in the WHBP, this was not easily detectable every cycle due to the short burst duration. Thus, inspiratory motor pattern might not be a reliable index of eupnoea. Therefore, the most important criteria for defining eupnoea in neonatal rats is: (i) the frequency and short duration of phrenic nerve discharges (i.e. |46 per min and |200 ms duration); (ii) an inspiratory to total respiratory cycle time ratio of approximately 15%; (iii) presence of post-inspiratory discharge in the recurrent laryngeal or central vagal
27
outflow; and (iv) laryngeal dilatation and constriction during inspiration and post-inspiration respectively. Another major difference between the WHBP and in vitro is the paucity of post-inspiratory activity found in recordings from motor nerves and respiratory neurones in vitro. In the en bloc or brainstem–spinal cord preparation, both hypoglossal and vagal rootlets rarely show activity following the inspiratory burst. In cases where it is present (see Smith et al., 1990) it is unlike that seen in the in situ preparation. It is of low amplitude, highly staccato, does not decrement and appears to be inconsistent within and across studies. In central recording studies using in vitro preparations most authors fail to find post-inspiratory neurones (e.g., Brockhaus and Ballanyi, 1998; Herlenius and Lagercrantz, 1999; Johnson et al., 1994; Smith et al., 1990, 1991). When described they are either rare (i.e ,7% of all cells; Kawai et al., 1996) or very different (see Fig. 3 of Ramirez et al., 1998) to both those we described here and those seen in adult mammals in vivo. The predominant cell type found in vitro is inspiratory and the most frequently encountered expiratory neurones types were termed tonic and biphasic expiratory cells (Herlenius and Lagercrantz, 1999; Kawai et al., 1996; Onimaru and Homma, 1992; Smith et al., 1990). The dearth of post-inspiratory activity may have a major bearing on the type of respiratory rhythm being generated in vitro since post-inspiratory neurones constitute an essential element of the basic rhythm generator in vivo by providing both an irreversible inspiratory off-switch (Richter et al., 1986, 1992; Rybak et al., 1997) and inhibition of expiratory neurones (i.e. augmenting expiratory or E2 type — see Fig. 5 ‘Expiratory’; Klages et al., 1993). Since post-inspiratory neurones inhibit augmenting expiratory neurones, an absence of post-inspiratory neurones in vitro explains why internal intercostal nerves fire in the postinspiratory phase and not late expiration (see Iizuka, 1999). The in vitro finding of expiratory neurones that show no pattern (i.e. tonically fire throughout the expiratory interval) can also be accounted for by an absence of postinspiratory activity. Post-inspiratory neurones are also essential for control of glottal musculature, which is functional at birth (see Harding, 1984, for a review) and causes inspiratory opening and early expiratory constriction, as well as for basic behaviours such as vocalisation and swallowing. This raises controversial questions of whether activity generated in vitro is ‘breathing’ (for recent viewpoints, see Cohen, 1999; Koshiya and Smith, 1999; Reckling and Feldman, 1998; St-John, 1998) and whether basic mechanisms of eupnoeic rhythm generation can be studied in the in vitro preparations. Clearly a form of rhythm generation is maintained in vitro, but whether it would permit ventilation is unclear. The pacemaker or pacemaker-hybrid system may form a component of the rhythm generator which when embedded in the intact perfused brain is superceded by network properties which alone are sufficient to drive eupnoea (e.g., Rybak et al.,
28
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29
1997). A possibility suggested previously is that pacemaker activity produces a central apnoeic rhythm which provides a protective function to ‘kick-start’ inspiration following an extreme insult of hypoxia or ischaemia (see Paton, 1997).
6. Conclusions The in situ WHBP has permitted a thorough description of the respiratory activity generated by a well oxygenated medulla in the neonatal rat. We suggest that the WHBP offers an alternative preparation to study the neurogenesis of eupnoea in an immature mammalian respiratory network. This in situ approach also provides opportunities to study the ontogenesis of reflexes regulating respiration, reflexes altering cardiac function and central cardiorespiratory coupling. The pattern, frequency and central organisation of respiratory activity generated in the arterially perfused in situ WHBP of 1 h to 4 day old rats is distinct to in vitro rhythmic preparations of animals of the same age. The respiratory system of the WHBP of rat pups is more akin to the in vivo adult rat; this conclusion is similar to that of Wang et al. (1996) who showed that eupnoea in vagotomised pups resembles closely that of the vagotomised adult preparations. We must now define whether in vitro preparations are generating eupnoea and interpret data from these preparations accordingly (see Remmers, 1998). A wrong interpretation concerning the central mechanisms governing breathing is not only scientifically inaccurate but potentially dangerous clinically.
Acknowledgements We would like to thank Dr. Thomas E. Dick and Professor John E. Remmers for their helpful suggestions concerning the manuscript. MD was supported by a travel fellowship from the Deutsche Forschungsgemeinschaft. RJAW is supported by a Parker B. Francis Fellowship in Pulmonary Research and was in receipt of a Wellcome Trust travel fellowship. JFRP is supported by the British Heart Foundation (BS / 93003).
References Alsop, S., Bhatia, A., Paton, J.F.R., 1997. Characteristics of the respiratory network in an arterially perfused preparation of the neonatal rat. J. Physiol. 504P, 200P. Ballanyi, K., Doutheil, J., Brockhaus, J., 1996. Membrane potentials and microenvironment of rat dorsal vagal cells in vitro during energy depletion. J. Physiol. 495, 769–784. Ballanyi, K., Onimaru, H., Homma, I., 1999. Respiratory network in the isolated brainstem–spinal cord of newborn rats. Prog. Neurobiol. 59, 583–634.
Brockhaus, J., Ballanyi, K., 1998. Synaptic inhibition in the isolated respiratory network of neonatal rats. Eur. J. Neurosci. 10. Cohen, M.I., 1979. Neurogenesis of respiratory rhythm in the mammal. Physiol. Rev. 59, 1105–1173. Cohen, P., 1999. Cells with rhythm keep us breathing. New Scientist 286, 1566. Fung, M.-L., St-John, W.M., 1995. The functional expression of a pontine pneumotaxic centre in neonatal rats. J. Physiol. 489 (2), 579–591. Haji, A., Okazaki, M., Takeda, R., 1999. GABA(A) receptor-mediated inspiratory termination evoked by vagal stimulation in decerebrate cats. Neuropharmacol. 38, 1261–1272. Harding, R., 1984. Function of the larynx in the fetus and newborn. Annu. Rev. Physiol. 46, 645–659. Hayashi, F., Jiang, C., Lipski, J., 1991. Intracellular recording from respiratory neurones in the perfused ‘in situ’ rat brain. J. Neurosci. Methods 36, 63–70. Herlenius, E., Lagercrantz, H., 1999. Adenosinergic modulation of respiratory neurones in the neonatal rat brainstem in vitro. J. Physiol. 518 (1), 159–172. Hilaire, G., Monteau, R., Errchidi, S., 1989. Possible modulation of the medullary respiratory rhythm generator by noradrenergic A5 area: an in vitro study in the newborn rat. Brain Res. 485, 325–332. Iizuka, M., 1999. Intercostal expiratory activity in an in vitro brainstem– spinal cord–rib preparation from the neonatal rat. J. Physiol. 520 (1), 293–302. Johnson, S.M., Smith, J.C., Funk, G.D., Feldman, J.L., 1994. Pacemaker properties of respiratory neurons in medullar slices from neonatal rats. J. Neurophysiol. 72, 2598–2608. Kalia, M., 1976. Visceral and somatic reflexes produced by J pulmonary receptors in new born kittens. J. Appl. Physiol. 41, 1–6. Kasparov, S., Paton, J.F.R., 1997. Changes in baroreceptor vagal reflex ¨ Arch. 434, 438–444. performance in the developing rat. Pflug. Kawai, A., Ballantyne, D., Muckenhoff, K., Scheid, P., 1996. Chemosensitive medullary neurones in the brainstem–spinal cord preparation of the neonatal rat. J. Physiol. 492 (1), 277–292. Klages, S., Bellingham, M.C., Richter, D.W., 1993. Late expiratory inhibition of stage 2 expiratory neurons in the cat — a correlate of expiratory termination. J. Neurophysiol. 70, 1307–1315. Koshiya, N., Smith, J.C., 1999. Neuronal pacemaker for breathing visualized in vitro. Nature 400, 360–363. Lawson, E.E., Richter, D.W., Bischoff, A., 1989. Intracellular recordings of medullary respiratory neurons in the lateral medulla of piglets. J. Appl. Physiol. 66, 983–988. Lumsden, T., 1923. Observations on the respiratory centres in the cat. J. Physiol. 57, 153–160. Okada, Y., Muckenhoff, K., Holtermann, G., Acker, H., Scheid, P., 1993. Depth profiles of pH and PO 2 in the isolated brain stem–spinal cord of the neonatal rat. Respir. Physiol. 93, 315–326. Onimaru, H., Homma, I., 1992. Whole cell recodings from respiratory neurons in the medulla of brainstem–spinal cord preparations isolated ¨ Arch. 420, 399–406. from newborn rats. Pflug. Paton, J.F.R., 1996a. A working heart–brainstem preparation of the mouse. J. Neurosci. Methods 65, 63–68. Paton, J.F.R., 1996b. The ventral respiratory network of the mature mouse studied in a working heart brainstem preparation. J. Physiol. 493 (3), 819–831. Paton, J.F.R., 1997. Rhythmic bursting of pre- and post-inspiratory neurones during central apnoea in mature mice. J. Physiol. 502 (3), 623–639. Paton, J.F.R., Nolan, P.J., 2000. Parallelism in reflex control of laryngeal and cardiac vagal motor neurones. Respir. Physiol. (in press). Paton, J.F.R., Li, Y.-W., Kasparov, S., 1999. Reflex response and convergence of pharyngoesophageal and peripheral chemo-receptors in the nucleus of the solitary tract. Neuroscience 93, 143–154. Paton, J.F.R., Ramirez, J.-M., Richter, D.W., 1994. Functionally intact in vitro preparation generating respiratory activity in neonatal and ¨ Arch. 428, 250–260. mature mammals. Pflug.
M. Dutschmann et al. / Autonomic Neuroscience: Basic and Clinical 84 (2000) 19 – 29 Ramirez, J.M., Quellmalz, U.J.A., Wilken, B., Richter, D.W., 1998. The hypoxic response of neurones within the in vitro mammalian respiratory network. J. Physiol. 507 (2), 571–582. Reckling, J.C., Feldman, J.L., 1998. Prebotzinger complex and pacemaker neurons: hypothesized site and kernel for respiratory rhythm generation. Annu. Rev. Physiol. 60, 385–405. Remmers, J.E., 1998. Central neural control of breathing. In: Altose, M., Kawakami, Y. (Eds.), Control of Breathing in Health and Disease. Marcel Dekker, New York. Richter, D.W., Ballantyne, D., Remmers, J.E., 1986. How is the respiratory rhythm generated? A model. NIPS 1, 109–112. Richter, D.W., Ballanyi, K., Schwarzacher, S.W., 1992. Mechanisms of respiratory rhythm generation. Curr. Opin. Neurobiol. 2, 788–793. Richter, D.W., Heyde, F., Gabriel, M., 1975. Intracellular recordings from different types of medullary respiratory neurons of the cat. J. Neurophysiol. 38, 1162–1171. Richter, D.W., Spyer, K.M., 1990. Cardiorespiratory control. In: Lowey, A.D., Spyer, K.M. (Eds.), Central Regulation of Autonomic Functions. Oxford University Press, New York, pp. 189–207. Rybak, I.A., Paton, J.F.R., Schwaber, J.S., 1997. Modelling neural mechanisms for genesis of respiratory rhythm and pattern: II. Network models of the central respiratory pattern generator. J. Neurophysiol. 77, 2007–2026.
29
Schwarzacher, S.W., Wilhelm, Z., Anders, K., Richter, D.W., 1991. The medullary respiratory network in the rat. J. Physiol. 435, 631–644. Shiba, K., Satoh, I., Kobayashi, N., Hatashi, F., 1999. Multifunctional laryngeal motoneurons: an intracellular study in the cat. J. Neurosci. 19, 2717–2727. Smith, J.C., Ellenberger, H.H., Ballanyi, K., Richter, D.W., Feldman, J.L., ¨ 1991. Pre-Botzinger complex: a brainstem region that may generate respiratory rhythm in mammals. Science 254, 726–729. Smith, J.C., Greer, J.L., Liu, G., Feldman, J.L., 1990. Neural mechanisms generating respiratory pattern in mammalian brainstem–spinal cord in vitro. I. Spatiotemporal patterns of motor and medullar neuron activity. J. Neurophysiol. 64, 1149–1169. St-John, W.M., 1998. Neurogenesis of patterns of automatic ventilatory activity. Prog. Neurobiol. 56, 97–117. Suzue, T., 1984. Respiratory rhythm generation in the in vitro brainstem– spinal cord preparation of neonatal rat. J. Physiol. 354, 173–183. Wang, W., Fung, M.-L., Darnall, R.A., St-John, W.M., 1996. Characterizations and comparisons of eupnoea and gasping in neonatal rats. J. Physiol. 490 (1), 272–292. Zheng, Y., Barillot, J.C., Bianchi, A.L., 1991. Patterns of membrane potentials and distributions of the medullary respiratory neurons in the decerebrate rat. Brain Res. 546, 261–270.