Corresponding author W. M. St.-John: Department of Physiology, Dartmouth Medical School, Dartmouth-Hitchcock Medical Center, Lebanon, NH 03756, USA. Email: email@example.com
During eupnoea, rhythmic motor activities of the hypoglossal, vagal and phrenic nerves are linked temporally. The inspiratory discharges of the hypoglossal and vagus motor neurones commence before the onset of the phrenic burst. The vagus nerve also discharges in expiration. Upon exposure to hypocapnia or hypothermia, the hypoglossal discharge became uncoupled from that of the phrenic nerve. This uncoupling was evidenced by variable times of onset of hypoglossal discharge before or after the onset of phrenic discharge, extra bursts of hypoglossal activity in neural expiration, or complete absence of any hypoglossal discharge during a respiratory cycle. No such changes were found for vagal discharge, which remained linked to the phrenic bursts. Intracellular recordings in the hypoglossal nucleus revealed that all changes in hypoglossal discharge were due to neuronal depolarization. These results add support to the conclusion that the brainstem control of respiratory-modulated hypoglossal activity differs from control of phrenic and vagal activity. These findings have implications for any studies in which activity of the hypoglossal nerve is used as the sole index of neural inspiration. Indeed, our results establish that hypoglossal discharge alone is an equivocal index of the pattern of overall ventilatory activity and that this is accentuated by hypercapnia and hypothermia.
To maintain patency of the upper airway during breathing, the discharges of cranial and spinal nerves must be coordinated (Remmers et al. 1978; Dutschmann & Paton, 2002). Prominent among these discharges is the hypoglossal nerve, which innervates the muscles of the tongue. The onset of discharge of the hypoglossal nerve occurs before the onset of activity of the phrenic nerve. This early onset of hypoglossal activity, termed the ‘pre-inspiratory’ discharge, stabilizes the upper airway prior to the commencement of airflow to the lungs (Remmers et al. 1978; Hwang et al. 1983; Remmers, 1998; St.-John, 1998; Bailey et al. 2001).
Brainstem mechanisms controlling the pre-inspiratory portion of the hypoglossal discharge may differ from those controlling hypoglossal activity during ‘neural inspiration’, which is concomitant with phrenic discharge. This possibility is supported by the finding that the pre-inspiratory portion of the hypoglossal discharge is separable from the ‘inspiratory’ discharge. In the in situ preparation of the juvenile rat, such a separation was evidenced by the occurrence of a transient reduction or total cessation of hypoglossal discharge at the start of the phrenic burst and/or by complete separation of pre-inspiratory and inspiratory hypoglossal bursts (Leiter & St.-John, 2004). Alterations in lung volume of in vivo preparations can also induce a separation of the pre-inspiratory and inspiratory components (Saito et al. 2002a,b). Additionally, high frequency oscillations in neural activities, assessed by power spectral analysis, differ between the components of the hypoglossal discharge. Peak frequency in the power spectra of the pre-inspiratory discharge was lower than that during the inspiratory component. This inspiratory component of hypoglossal discharge had the same peak frequency as that of the phrenic discharge and also both inspiratory and expiratory discharges of the vagus nerve (St.-John & Leiter, 2003; Leiter & St.-John, 2004). These differences in peak frequencies in the power spectra suggest a different origin of drive to the pre-inspiratory and inspiratory components of the hypoglossal discharge.
The present study was undertaken to assess further the mechanisms controlling the pre-inspiratory and inspiratory components of the hypoglossal discharge. If these components are indeed differentially regulated, we hypothesized that uncoupling between the two components of hypoglossal activity would be induced by perturbations that caused a reduction in eupnoeic ventilatory drive, including hypocapnia and hypothermia. In addition, we evaluated whether the uncoupling of the pre-inspiratory from the inspiratory hypoglossal discharges represented a transient, active inhibition that was superimposed on a continuous hypoglossal discharge. Finally, from recordings of efferent activity of the vagus nerve, we evaluated whether ‘uncoupling’ was a common characteristic of respiratory-modulated activities of cranial nerves other than the hypoglossal nerve. The results of these studies support the conclusion that the control of pre-inspiratory and also inspiratory hypoglossal activity differs from that of the vagal and bulbospinal–phrenic system.
We used 33 perfused preparations of juvenile rat. The preparation was identical to that previously described (St.-John & Paton, 2000). Preparations were deeply anaesthetized with halothane, as evidenced by a failure to respond to a noxious pinch of the paw. Under this halothane anaesthesia, the portion of the body caudal to the diaphragm was removed. The upper portion of the preparation was immersed in ice-cold artificial cerebrospinal fluid and decerebrated at a precollicular level. Halothane anaesthesia was then discontinued. The descending aorta was cannulated and perfusion was commenced. The perfusate contained the following in distilled water (mm): MgSO4 1.25, KH2PO4 1.25, KCl 5.0, NaHCO3 25, NaCl 125, CaCl2 2.5, dextrose 10 and Ficoll 70 0.1785. The perfusate was equilibrated with 95% O2–5% CO2. In some studies, a smaller catheter was inserted within the catheter in the aorta and 25–75 μl saline, containing sodium cyanide (0.03%), was injected to activate the peripheral chemoreceptors.
Because oxygenation of the preparation was via the perfusate, the lungs were not inflated. The vagi remained intact in all but five preparations in which one vagus was sectioned for recording of efferent activity.
The temperature of the perfusate was maintained at 31°C in 28 preparations, and at 27°C in the other five preparations. The temperature was measured as the perfusate entered the aorta. The temperature of the perfusate was regulated by a heat exchanger. The gas mixture with which the perfusate was equilibrated, and its temperature, could be altered as required for particular experiments.
Recording of neural activities
In all preparations, efferent activity of the phrenic nerve was recorded. Activities were also recorded from the central cut ends of the hypoglossal nerve in 23 preparations and from the vagus nerve in five preparations. Activities were recorded with glass ‘suction’ electrodes or bipolar metal electrodes, amplified, filtered (0.6–6.0 kHz) and integrated (50 ms time constant). Recordings were by Spike 2 (Cambridge Instruments) or DATAPAC (Run Technologies).
Recording of neuronal activities
In seven preparations, intracellular recordings were obtained from neurones in the region of the hypoglossal nucleus. Fine-tipped glass microelectrodes (borosilicate glass; GC 150F; Harvard) were pulled on a Sutter P-87 puller (Brown Flaming, USA) and filled with 3 m KCl. The impedance was 40–80 MΩ. Neuronal impalement was made by passing a transient, high amplitude positive current pulse. In some trials, inhibitory postsynaptic potentials were reversed by injecting constant hyperpolarizing current. Signals were amplified using a NPI bridge amplifier (Tamm, Germany), sampled at 5 kHz, digitized and recorded onto the hard drive of a computer (Spike 2 software).
Perfusion began with the perfusate equilibrated with 95% O2–5% CO2 and the temperature of the perfusate maintained at 31°C or 27°C, as noted above. After the start of perfusion, recordings were obtained as soon as rhythmic activities began. Approximately 15 min thereafter, ‘baseline activity’ was recorded in eupnoea. One or more of three experimental perturbations was performed: (1) the temperature of the perfusate was gradually reduced from 31°C to 27°C or elevated from 27°C to 31°C by altering the temperature in the heat exchanger; (2) the perfusate was made hypocapnic by equilibrating it with 97% O2–3% CO2; (3) the peripheral chemoreceptors were stimulated by injecting saline containing sodium cyanide through the inner catheter in the aortic cannula.
Analyses of data
For activities of the hypoglossal, vagus and phrenic nerves, the variable of interest was the difference in time of onset of activities. This variable was measured in a minimum of 25 ventilatory cycles commencing a minimum of 5 min after any experimental perturbation had been completed. These perturbations included changes from normocapnia to hypocapnia or a change in temperature of the perfusate delivered to the preparation. Under each experimental condition, mean values, standard deviations and coefficients of variation were determined for each preparation. ‘Return plots’ were also constructed in which the difference in the time of onset of hypoglossal or vagal activity compared to phrenic activity in one cycle was compared to the difference in onset in the subsequent cycle. Linear regression analyses and correlations were performed on these data. Statistical evaluations were by a Wilcoxon test; P < 0.05 was considered as significant.
For intracellular recordings of hypoglossal motoneurones, we noted changes in membrane potential following the change from normocapnia to hypocapnia and also following stimulation of the carotid chemoreceptors by sodium cyanide.
Changes in patterns of phrenic, vagal and hypoglossal activities after the onset of rhythmic activities
With the perfusate equilibrated with 95% O2–5% CO2 at a temperature of 31°C, a eupnoeic pattern of phrenic activity was recorded. This pattern was characterized by a sudden onset of discharge and a ramp-like rise to a peak integrated level in the mid to late portion of the burst (Fig. 1). Hypoglossal discharge commenced before the onset of the phrenic burst and terminated during the inspiratory phase of the phrenic burst. Vagal discharge also commenced before the onset of the phrenic burst, but lasted through phrenic inspiration and ended during the subsequent neural expiration (Fig. 1). No obvious differences in respiratory activity were observed for preparations having intact or a sectioned vagus nerve.
At 31°C, the pattern of phrenic discharge remained stable from the time at which this discharge commenced in the preparation. However, in every preparation, the pattern of hypoglossal discharge changed with time. When rhythmic phrenic activity first commenced, the hypoglossal discharge was characterized by a pause at or soon after the commencement of the phrenic burst (Fig. 1). This pause in hypoglossal discharge became less pronounced with time (Fig. 2). As was the case for phrenic activity, inspiratory vagal activity changed little with time (Fig. 1).
When rhythmic activity first commenced in preparations maintained at 27°C, phrenic discharge had two patterns (Fig. 1). In the first, phrenic discharge increased abruptly and was then maintained in a ramp-like pattern. In the second, the abrupt onset was absent and phrenic discharge gradually increased from the onset to the termination of inspiratory activity. For cycles in which the phrenic discharge had the abrupt onset, the hypoglossal discharge commenced in mid to late neural expiration. The pause between pre-inspiratory and inspiratory hypoglossal discharges was variable, from barely discernible to pronounced (Fig. 1). For cycles in which only an increasing phrenic pattern was noted, hypoglossal discharge commenced late in neural inspiration or was completely absent. These different patterns of hypoglossal discharge were also noted when the temperature of the preparation was reduced from 31 to 27°C (see below).
Changes in patterns of phrenic, vagal and hypoglossal activities with alterations in temperature
With the temperature of the preparation maintained at 31°C, hypoglossal discharge commenced before the onset of phrenic activity. In most cycles, the onset of discharge was in the mid-stages of neural expiration, and this onset was at very similar times in most cycles (Figs 2–4). An example of such similarity is shown in Fig. 3 for one preparation in which the differences in time of onset of hypoglossal and phrenic activities are presented for 50 respiratory cycles at 31°C (circles). In Fig. 4, mean values and standard deviations for this difference in time of onset of activities are presented for each of five preparations. Note in this figure that the coefficient of variation (open circles) for the difference in onset of hypoglossal compared to phrenic discharge was less than 100%.
When the temperature of the preparation was reduced to 27°C, the onset of hypoglossal discharge became extremely variable. Hence, as shown in Figs 2 and 3 (Fig. 2; triangles), while hypoglossal discharge still commenced before the phrenic discharge in most cycles, an onset after the onset of phrenic discharge was also common. Moreover, extra bursts of hypoglossal discharge in neural expiration were also seen. This increase in variability in time of onset of hypoglossal discharge is evidenced by the increase in the coefficient of variation in each of the five preparations with the change from 31°C to 27°C; this increase was statistically significant (Fig. 4; P < 0.05, Wilcoxon test).
In Fig. 5, ‘return plots’ are presented in which data from 20 respiratory cycles in each of the five preparations have been pooled. In these plots, the difference in the time of onset of hypoglossal and phrenic activity in one cycle is compared to this difference in the subsequent cycle. Linear regression analysis revealed a significant correlation for data obtained at 31°C but not at 27°C. Stated differently, the tight clustering of data points along the line of identity at 31°C indicates that the relationship between onset times of hypoglossal and phrenic activity was consistent across sequential respiratory cycles. But this relationship was not maintained following a reduction in the temperature of the preparation. Thus following this reduction, the onset of hypoglossal compared to phrenic activity became random, with there being no consistent relationship of onset times in serial respiratory cycles. Two additional preparations were maintained continuously at 27°C, and the onset of hypoglossal activity varied continuously in these preparations from the time of onset of rhythmic discharges. In these preparations, no consistent relationship between hypoglossal and phrenic onset was ever established.
As opposed to hypoglossal discharge, efferent activity of the vagus nerve showed little difference in variability in its time of onset of activity for examinations at 31°C or 27°C (Fig. 2). For the five preparations examined at both temperatures, coefficients of variation were very similar at each temperature (Fig. 4). Note that three preparations were initially examined at 27°C and then at 31°C; the others were examined at the higher temperature initially. Return plots of the onset of vagal activity relative to phrenic activity from five preparations examined at 27°C and seven preparations examined at 31°C revealed that onset times of sequential breaths were significantly correlated at both temperatures (Fig. 6). However, regardless of the sequence of temperature changes, the onset of vagal activity had a relatively fixed relationship to the onset of phrenic activity.
Changes in patterns of phrenic and hypoglossal activities in hypocapnia
Alterations in the relationship between hypoglossal and phrenic activities in hypocapnia were similar to those in hypothermia. In each of three preparations examined, the onset of hypoglossal discharge became exceedingly variable, following a reduction in the level of carbon dioxide in the perfusate from 5% to 3%. In some cases, the hypoglossal discharge actually commenced after the beginning of phrenic activity. Extra bursts of hypoglossal discharge during neural expiration were also observed.
The increase in variability in the three preparations was evidenced by an increase in the coefficient of variation in each (24, 34 and 19% with 5% CO2 in the perfusate and 168, 139 and 49% at 3% CO2). Return plots also showed a significant correlation between serial respiratory cycles obtained at normocapnia but not hypocapnia (Fig. 7).
Changes in activities of neurones in hypoglossal nucleus during uncoupling of phrenic and hypoglossal discharge
Intracellular recordings of 56 neurones were obtained in the region of the hypoglossal nucleus. With holding currents removed, the resting membrane potential averaged 57 ± 2.4 mV. These intracellular recordings were obtained while the activities of the phrenic and hypoglossal nerves were also recorded (Fig. 8). Inspiratory activity was demonstrated in 34 of the neurones. The neurone depolarized and the frequency of action potentials increased preceding the onset of phrenic discharge and continued throughout the phrenic discharge. There was a brief transient hyperpolarization at the termination of the burst, but no evidence of inhibitory postsynaptic potentials in neural expiration. Inhibitory potentials were seen during neural inspiration in 12 expiratory neurones that depolarized and exhibited action potentials in the period between phrenic bursts. These inhibitory potentials could be reversed by current injection, as shown in Fig. 9. For the neuronal activity in this Figure, a hyperpolarization to –80 mv caused a reversal of the inspiration-related inhibitory potential, indicating a fast chloride-mediated inhibition.
Four neurones had a phase-spanning inspiratory–expiratory pattern. These neurones commenced discharge before the onset of the phrenic burst and continued to discharge until the mid portion of neural expiration. Finally, six neurones discharged tonically throughout the entire respiratory cycle.
During uncoupling of hypoglossal and phrenic discharge, evidenced by extra bursts of the former, neurones in the hypoglossal nucleus showed a stereotypical depolarization and also action potentials in neural expiration (Fig. 8). Especially interesting was the observation that expiratory neurones showed an augmented burst of activity during the extra bursts of activity of the hypoglossal nerve in neural expiration. Thus, the separation of hypoglossal from phrenic discharges appears to reflect additional excitatory synaptic inputs in hypoglossal neurones, but not phrenic motoneurones.
The major conclusion of this study is that the control of respiration-modulated activity of hypoglossal motoneurones differs markedly from the control of vagal motoneurones and the bulbospinal–phrenic motor system. Stated differently, the linkage between the brainstem respiratory mechanisms generating eupnoea, as defined by activity of the phrenic nerve, and hypoglossal discharge is variable, whereas the relationship between phrenic and vagal activity is fixed and less susceptible to modification. Changes in temperature of the preparation, exposure to hypocapnia and stimulation of the peripheral chemoreceptors were used to induce modification. The variable nature of the relationship between hypoglossal and phrenic discharges is especially apparent for the component of hypoglossal discharge which commences in mid to late neural expiration.
Results of the present study further establish that the control of hypoglossal and phrenic activities differs. A variable link between the brainstem ventilatory control system and hypoglossal discharge was expected as the hypoglossal system is involved in many other rhythmic activities besides respiration, such as mastication and deglutition (Dick et al. 1993; Paton et al. 1999; Jean, 2001; Roda et al. 2002; Saito et al. 2002a,b). Hence, under conditions in which ventilatory drive is lowered, the link between hypoglossal discharge and the overall respiratory rhythm may become more tenuous. In the present study, lowered respiratory drive in hypocapnia and hypothermia resulted in uncoupling of hypoglossal activity from phrenic activity. Somewhat paradoxically, phrenic and hypoglossal activities also became uncoupled during stimulation of the carotid chemoreceptors. However, this uncoupling may not reflect the overall increase in respiratory activity as stimulation of the carotid bodies activates the brainstem reticular system (see references in Hafström et al. 2000). A greater dependency of the hypoglossal than phrenic systems upon reticular activation has long been proposed (for review see St.-John, 1986). Alternatively, the peripheral chemoreceptor stimulation may have elicited centrally generated swallowing, and hence extra bursts of hypoglossal activity (Paton et al. 1999).
Not only does the control of hypoglossal and phrenic activities per se differ, but various portions of the hypoglossal discharge may be differentially regulated. As noted in the Introduction, the pre-inspiratory and inspiratory components of the hypoglossal discharge are separable (Saito et al. 2002a,b; Leiter & St.-John, 2004). Furthermore, differences in high frequency oscillations of nerve activities between pre-inspiratory and inspiratory components of hypoglossal activity also imply that separate brainstem mechanisms control these two components of the hypoglossal discharge. Power spectral analysis of these oscillations during eupnoea revealed similar peak frequencies for phrenic discharge, vagal discharges in neural inspiration and expiration and hypoglossal discharge in neural inspiration. However, a significantly lower frequency was found for the pre-inspiratory portion of the hypoglossal discharge (Leiter & St.-John, 2004). High frequency oscillations in neural activities are considered signatures of the central pattern generator for eupnoea (St.-John & Leiter, 2003; Leiter & St.-John, 2004). Hence, the difference in peak frequency of the pre-inspiratory hypoglossal discharge implies that this component receives an input from a separate brainstem source than that to the other cranial and spinal nerves. The similarity of high frequency oscillations among inspiratory activity of the hypoglossal, vagal and phrenic activities implies that the inspiratory activation of these motor neurone pools originates from a common source.
Results from intracellular recordings also demonstrate a different control of hypoglossal compared to phrenic activities. Hence the depolarization of neurones in the hypoglossal neurones commenced before the phrenic discharge and/or during extra bursts of activity of the hypoglossal nerve in neural expiration with no concomitant changes in phrenic discharge. The excitatory events underlying these depolarizations appear to be differentially distributed among the neurones in the hypoglossal nucleus. Hence, despite an overall increase in activity of the hypoglossal nerve, some neurones showed little change in depolarization and no spiking. Similarly, neurones maintained depolarizations and even spiking during the pause in activity of the hypoglossal nerve at the commencement of phrenic discharge. Such pauses were minimal during recordings in the present study, as these were obtained long after the commencement of rhythmic respiratory action when such pauses were maximal. Yet in the study of Saito et al. (2002b), in which the pauses of pre-inspiratory and inspiratory discharges were marked, again hypoglossal motoneuronal activities showed variable changes, but largely a reduction in the degree of depolarization. Thus disfacilitation would appear responsible for these pauses, which is in agreement with the concept that the discharge of hypoglossal motoneurones is largely defined by excitatory synaptic events (Peever et al. 2002; Saito et al. 2002b).
Results from intracellular recordings establish that the difference between pre-inspiratory and inspiratory components of the discharge of the hypoglossal nerve cannot reflect neuronal discharges to different groups of muscles. Hence, while the hypoglossal nerve innervates a number of muscles of the tongue, a separation of pre-inspiratory and inspiratory discharges was uniformly observed for intracellular recordings of hypoglossal motoneuronal activities (Saito et al. 2002b). Moreover, such a separation was also observed for electromyographic activities of both the genioglossus and hyoglossal muscles (Bailey et al. 2001).
As to the brainstem source of respiration-modulated activity, neuroanatomical and neurophysiological observations demonstrated that the vast majority of ‘premotor’ pontile and medullary neurones to hypoglossal motor neurones differed from those to the bulbospinal–phrenic system (Ono et al. 1994; Dobbins & Feldman, 1995; Fay & Norgren, 1997; Peever et al. 2001, 2002; Travers & Rinaman, 2002; Li et al. 2003). Only a few premotor neurones which project to both the hypoglossal and phrenic motor nuclei have been identified in cats (Ono et al. 1994). Concerning medullary premotor neurones, there is no doubt that intrinsic medullary circuits can convey respiration-modulated activities to the hypoglossal and phrenic systems, as phrenic and hypoglossal discharges are synchronized during gasping in vivo and in situ and during rhythmic in vitro discharges (e.g. Smith et al. 1990; St.-John, 1990; Kato et al. 1996; St.-John & Leiter, 2003). Yet, these intrinsic medullary discharges differ from those during eupnoea in that phrenic and hypoglossal discharges commence at approximately the same time. In vitro preparations are of neonatal animals, whereas the majority of in vivo and in situ studies have involved juvenile and adult rats. The age of preparations would not appear to account for the difference in relative times of onset of hypoglossal and phrenic activities as hypoglossal discharge commences before phrenic discharge during eupnoea in neonatal in vivo rats (Fukuda, 2000). These times of onset become the same in gasping (Fukuda, 2000). Similarly, the onset of the efferent vagal activity precedes that of the phrenic during eupnoea in neonatal rats in vivo or in situ but again is the same in gasping (Wang et al. 1996; Dutschmann & Paton, 2002). This difference in the control of hypoglossal activities in vitro as compared to that during eupnoea in vivo or in situ might account for the findings of Kato et al. (1996) that, as opposed to the results herein, the synchronization between phrenic and hypoglossal discharges of an in vitro neonatal cat preparation was unaltered by changes in temperature. This synchronization was unaltered by removal of pons by a brainstem transection at the pontomedullary junction which supports our conclusion that these in vitro rhythms differ from those of eupnoea in vivo (St.-John, 1990, 1996, 1998; St.-John & Paton, 2003). Hence, medullary mechanisms alone may not be responsible for the regulation of hypoglossal and phrenic activities in eupnoea.
A prominent group of ‘premotor’ neurones to the hypoglossal nucleus are within the rostral pons, close to the pneumotaxic centre of the nucleus parabrachial medialis and Kolliker-Fuse nucleus (Dobbins & Feldman, 1995; Fay & Norgren, 1997; Li et al. 2003; Travers & Rinaman, 2002). While not evaluated in detail, there is evidence that pneumotaxic mechanisms can powerfully influence hypoglossal activity. Hence, electrical stimulation of the pneumotaxic centre, in regions that terminate phrenic discharge, causes augmentations in hypoglossal activity (St.-John, 1987). In apneusis in vivo, the difference in time of onset of phrenic and hypoglossal and phrenic activities is reduced or eliminated (St.-John et al. 1984). Thus, we submit that the complex interrelationships between hypoglossal and phrenic activities, described herein, are compatible with a significant role for pontile mechanisms in establishing this relationship.
Finally, results of the present study demonstrate that, during eupnoea, rhythmic activity of the hypoglossal nerve may deviate markedly from the overall respiratory rhythm as assessed by activities of the phrenic nerve and vagus nerve. Hence, from recordings of hypoglossal activity alone, rhythmic bursts cannot be concluded to be ‘inspiratory’ or even related to respiration. Thus, this raises concern over any experiment in which hypoglossal activity alone is used an index of neural inspiration. Typically, hypoglossal activity alone is recorded in vitro slice preparations and so caution should be exercised regarding the relevance to neural inspiration.
These studies were supported by grants from the National Heart, Lung and Blood Institute (26091), National Institutes of Health (USA) and from the British Heart Foundation.