Role of Endogenous Serotonin in Modulating Genioglossus Muscle Activity in Awake and Sleeping Rats
http://www.100md.com
《美国呼吸和危急护理医学》
Departments of Medicine and Physiology, University of Toronto, Toronto, Canada
ABSTRACT
Rationale: Exogenous serotonin at the hypoglossal motor nucleus (HMN) stimulates genioglossus (GG) muscle activity. However, whether endogenous serotonin contributes to GG activation across natural sleep–wake states has not been determined, but is relevant given that serotonergic neurons have decreased activity in sleep and project to pharyngeal motoneurons.
Objectives: To determine the role of endogenous serotonin at the HMN in modulating GG activity across natural sleep–wake states.
Methods: Ten rats were implanted with electroencephalogram and neck muscle electrodes to record sleep–wake states, and GG and diaphragm wires for respiratory muscle recordings. Microdialysis probes were implanted into the HMN for perfusion of artificial cerebrospinal fluid and the serotonin receptor antagonist mianserin (100 μM).
Measurements and Main Results: In room air, there was no effect of mianserin on respiratory-related or tonic GG activities across sleep–wake states (p > 0.300). In hypercapnia, however, the normal declines in GG activity from non-REM to REM sleep, and wakefulness to REM sleep, were reduced with mianserin (p < 0.005). These data demonstrate a normally low endogenous serotonergic drive modulating GG activity unless augmented by reflex inputs. We also demonstrated a significant serotonergic drive modulating GG activity in vagotomized rats, but not in vagi-intact rats, under anesthesia, suggesting that previous results in reduced preparations may have been influenced by vagotomy.
Conclusions: The results show a minimal endogenous serotonergic drive at the HMN modulating GG activity across sleep–wake states, unless augmented by reflex inputs. This result has implications for pharmacologic strategies aiming to increase GG activity by manipulating endogenous serotonin in patients with obstructive sleep apnea.
Key Words: hypoglossal motor nucleus obstructive sleep apnea serotonin sleep
Obstructive sleep apnea (OSA) is a common disorder affecting up to 4% of adults (1). There has been increasing interest in development of pharmacologic treatments for OSA, although initial studies using a variety of agents, such as ventilatory stimulants, have been largely unsuccessful (2–4). Given the sleep-state dependence of this disorder, however, newer approaches have attempted to modify those sleep state–dependent neural systems modulating pharyngeal motoneurons (5, 6). Nevertheless, the relative success or failure of such approaches ultimately depends on knowledge of the relevant neurotransmitters and, importantly, whether they provide a physiologically relevant drive to the pharyngeal muscles across sleep–wake states and are amenable to intervention.
For clinical studies, most attention has focused on the potential role of serotonin (5-hydroxtytryptamine [5-HT]) in the pharmacologic treatment of OSA because of its stimulatory effects on pharyngeal motoneurons. Accordingly, one approach has been to increase such stimulation via application of broad-spectrum or receptor-specific 5-HT agonists, or agents to increase 5-HT production (7, 8). The rationale for this approach comes from basic studies in animal preparations. In vitro studies of brainstem tissue slices show that 5-HT depolarizes and increases the excitability of hypoglossal motoneurons (9), the source of motor outflow to the genioglossus (GG) muscle of the tongue. Likewise, 5-HT at the hypoglossal motor nucleus (HMN) in decerebrate cats (10, 11) and anesthetized rats (12) increases motor output to GG. In freely behaving and naturally sleeping rats, tonic GG activation occurs when 5-HT is applied directly to the HMN, with the increased GG activity in non-REM sleep maintained for as long as 5-HT is applied (i.e., several hours), although the response is attenuated in REM sleep (13). Although it will be difficult to target the pharyngeal motoneurons specifically with systemic interventions in conscious animals, systemic administration of 5-HT precursors (14) or 5-HT2 receptor agonists (15) increase GG activity in acutely anesthetized rats; however, significant elevations in blood pressure have been reported (15).
A second major approach in the serotonergic modulation of pharyngeal motoneurons in patients with OSA has been to increase endogenous 5-HT via systemic administration of selective serotonin reuptake inhibitors (SSRIs) (16–19). However, it is again difficult to target the relevant pharyngeal motoneurons specifically with systemic interventions, with responses observed clinically being potentially confounded by influences on sleep or indirect effects on presynaptic and autoreceptor elements (that can, themselves, influence pharyngeal motor activity [5, 20–22]). Nevertheless, the conceptual rationale for this approach comes from basic studies in animal preparations. Medullary raphe neurons provide tonic 5-HT inputs to hypoglossal motoneurons, with these neurons showing declining discharge rates from wakefulness to non-REM and REM sleep (23) and activation by arousing stimuli, such as CO2 (24, 25), which contribute to the GG muscle responses to hypercapnia in anesthetized or decerebrate animals (12, 26). There is decreased discharge of medullary raphe neurons projecting to the HMN in a pharmacologic model of REM sleep in vagotomized and decerebrate cats (22). This change is accompanied by reduced 5-HT at the HMN in association with reduced hypoglossal nerve activity (27, 28). These observations in reduced animal preparations have led to the concept that increased raphe neuronal activity in wakefulness contributes to a physiologically relevant 5-HT drive to the HMN that is then withdrawn in sleep—especially REM sleep (29). However, it has not yet been tested whether 5-HT receptor antagonism at the HMN will decrease GG activity in wakefulness and sleep in a manner consistent with a significant role for endogenous 5-HT in the physiologic control of GG muscle activity, and this hypothesis is the focus of the present study. These results have relevance for pharmacologic strategies aiming to modulate GG activity by manipulating endogenous 5-HT in patients with OSA. Some of the results of this study have previously been reported in the form of an abstract (30).
METHODS
A more detailed account of the methods is given in the online supplement. Experiments were performed on 13 male Wistar rats (mean body weight = 261 ± [SEM] 7.3 g). All procedures conformed to the recommendations of the Canadian Council on Animal Care, and the University of Toronto Animal Care Committee approved the experimental protocols.
Anesthesia and Surgical Procedures
Sterile surgery was performed under general anesthesia for the chronic implantation of EEG and neck EMG electrodes to determine sleep–wake states, and GG and diaphragm electrodes for respiratory muscle recordings (13, 31, 32). Tests for the accurate placement of the GG electrodes and their function throughout the experiments (13, 31–33) are described in the online supplement. During surgery, microdialysis guides were targeted 3 mm above the HMN (13, 31, 32). The rats recovered for an average of 6.5 d (range, 5–8 d) before the studies.
Protocol and Data Analysis
On the day of the experiment, the microdialysis probe was inserted into the HMN (13, 31, 32). Measurements of sleep–wake states and respiratory muscle activities were made during both room-air and CO2-stimulated breathing using 7.5% inspired CO2 (31–33). Measurements were also made during hypercapnia, as 5-HT raphe neurons are activated by CO2 (24, 34) and contribute to the GG responses to respiratory stimulation in reduced preparations (12, 26). In 10 rats, the probes were flushed at a flow rate of 2.1 μl · min–1 with artificial cerebrospinal fluid (ACSF) or mianserin (5-HT receptor antagonist, 100 μM) dissolved in ACSF. Importantly, we have shown, using the exact same methodology, that 100 μM mianserin at the HMN in anesthetized rats produces robust decreases in GG activity during both room-air breathing and hypercapnia (12). Additional experiments in three rats were performed with the more specific 5-HT2 receptor antagonist MDL100907 at the HMN. The applied dose of 20 μM is 10 times higher than that causing major (62%) suppression of hypoglossal activity in anesthetized rats (35).
Sleep–wake states and respiratory muscle activities were analyzed as previously described (31–33) and are detailed in the online supplement. GG activity was quantified as tonic and respiratory-related activity. Data were analyzed at least 60 min after a switch between drugs. After the experiments, the rats were reanesthetized and tests for GG electrode function were performed (see online supplement). The sites of microdialysis were confirmed with histology (13, 31–33).
Experiments in Acutely Anesthetized Rats
Previous studies demonstrating a significant endogenous 5-HT drive to the HMN were performed in anesthetized or decerebrate vagotomized animals (10, 12, 22, 26, 27, 35). To characterize the potential role of the vagus nerves influencing GG responses to 5-HT receptor antagonism at the HMN, we performed additional experiments in 6 urethane-anesthetized rats with the vagus nerves intact and compared responses with those obtained in 10 vagotomized rats using the same methodology (12).
Statistical Analysis
Data were analyzed using ananlysis of variance with repeated measures, and a Student-Newman-Keuls test was used for post hoc comparisons. Analyses were performed using SigmaStat (SPSS, Inc., Chicago, IL). All data are expressed as mean ± SEM.
RESULTS
The full experimental protocol was completed in 11 of the 13 rats studied. In the two remaining rats, the mianserin protocol was also complete, except that REM sleep was not observed when ACSF was microdialized into the HMN in one rat, nor was REM sleep observed when mianserin was microdialized during hypercapnia in the other rat. For the group of 10 mianserin rats, a total of 44,053 5-s epochs (i.e., a total of 61.2 h of data) were included in the analysis, of which 29,111 epochs (66%) were from periods of wakefulness, 11,902 epochs (27%) were from non-REM sleep, and 3,040 epochs (7%) were from REM sleep. In the three rats with MDL100907 at the HMN, a total of 3,474 5-s epochs of wakefulness, 5,225 epochs of non-REM sleep, and 902 epochs of REM sleep were analyzed.
Function of GG Electrodes and Histology
Stimulation of the GG electrodes at the end of the experiment showed that the voltages required to cause tongue movements were not different from those at the time of surgery (0.80 ± 0.07 vs. 0.77 ± 0.05 V), demonstrating that the electrodes were in place and functional throughout the study.
Figure 1A shows an example of a lesion site made by the microdialysis probe in the HMN, with the dark stain produced by the potassium permanganate marking the probe site. Figure 1B shows the locations of all the microdialysis sites from each of the 10 mianserin and 3 MDL100907 rats. In all the experiments, the microdialysis sites were within the HMN or in the midline immediately adjacent to both the nuclei.
GG Activity across Sleep and Awake States
An example of the effects of sleep and awake states on respiratory muscle activity during room-air and CO2-stimulated breathing is shown in Figures 2A and 2B, respectively. Note the presence of prominent tonic GG activity during room-air breathing in wakefulness, as well as discernable respiratory-related GG activity (i.e., patterns typical of previous studies [31–33]). Also note that GG activity is markedly decreased from wakefulness to non-REM sleep, and is effectively abolished in those periods of REM sleep without the transient GG muscle twitches that also occur in REM sleep (see below). Figure 2 also shows that although GG activity is increased during CO2-stimulated breathing, similar changes in GG activity occur from wakefulness to non-REM sleep, including the periods of major suppression of GG activity in REM sleep.
Figure 2C shows an example of the change in GG activity at the transition from non-REM to REM sleep, as well as the variations in GG activity during REM sleep. This example illustrates that the periods of major suppression of GG activity in REM sleep typically occur at the onset of the REM sleep episodes, with the transient GG muscle twitches occurring later in the REM episodes. Such changes are typical. They have been analyzed in detail in previous articles (31, 32) and also occur in other animals, such as cats (36) and dogs (37). In addition, they are believed to be responsible for the sporadic restorations of airflow during obstructive apneas in REM sleep in humans (38).
Effects of 5-HT Receptor Antagonism with Mianserin on GG Activity
Wakefulness.
When the large quantities of data recorded in each individual rat were analyzed (i.e., those including both quiet wakefulness and active behaviors), there was a significant effect of 5-HT receptor antagonism at the HMN on respiratory-related and tonic GG activities (Figures 3A and 4A, respectively; both F1,9 > 5.33, p < 0.047) that did not depend on whether the animal was breathing room air or CO2 (both F1,9 < 1.31, p > 0.283). However, this overall effect of 5-HT receptor antagonism was in the opposite direction to that predicted by the hypothesis (i.e., mianserin increased rather than decreased respiratory-related and tonic GG activities; Figures 3A and 4A).
Further analysis was performed to determine if this result was indeed due to a specific effect of 5-HT receptor antagonism at the HMN or was likely an epiphenomenon secondary to a nonspecific generalized motor activation caused by the rat being engaged in more motor behaviors during wakefulness. If the latter were the case, there would also be an expected, concomitant activation of an unrelated control postural muscle as well as GG. This additional analysis showed that, during these same periods of wakefulness, there was also a significant increase in the activity of the control postural muscle in the neck (F1,9 = 15.24, p = 0.004) that also did not depend on whether the animal was breathing room or CO2-enriched air (F1,9 = 0.16, p = 0.701). This result suggested that the increased GG activity recorded at the time of 5-HT receptor antagonism at the HMN, when all of the wakefulness epochs were analyzed together (i.e., those including both quiet wakefulness and active behaviors), was most likely due to nonspecific effects related to active behaviors that would not otherwise have been detected without reference to the control muscle activity in the neck. Such a nonspecific general motor activation also likely explains the absence of a measurable increase in respiratory-related GG activity in response to CO2 when all the awake epochs were analyzed together (Figure 3A, F1,9 = 1.03, p = 0.337), as the active behaviors obscured any respiratory response. Nevertheless, the expected significant increases in indices of lung ventilation were observed in hypercapnia; hypercapnic respiratory stimulation increased diaphragm amplitude, respiratory rate, and diaphragm minute activity (all F1,9 > 52.41, p < 0.001). For further description, see CHANGES IN CONTROL VARIABLES ACROSS SLEEP–WAKE STATES.
To control for the overall potential confounding influence of active behaviors, obscuring the GG muscle responses to 5-HT receptor antagonism at the HMN and responses to CO2, further analyses were performed in those periods of relaxed wakefulness in which the rats were not physically active and engaged in behaviors.
Wakefulness without active behaviors.
In contrast to the main hypothesis of the study, there was no evidence of a significant suppressant effect of 5-HT receptor antagonism at the HMN on either respiratory-related or tonic GG activities in periods of quiet wakefulness (Figures 3B and 4B, respectively; both F1,9 < 1.10, p > 0.322). During these periods of wakefulness, however, the stimulating effects of CO2 on respiratory-related GG activities were readily apparent (Figure 3B; F1,9 = 9.86, p = 0.012), unlike during active behaviors (compare Figures 3A and 3B, and see above). There was no statistically significant effect of CO2 stimulation on tonic GG activity (Figure 4B; F1,9 = 1.16, p = 0.310) or neck muscle activity (F1,9 = 2.97, p = 0.119) as expected from previous results (33). The expected significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity in response to CO2 stimulation were all observed (all F1,9 > 24.69, p < 0.001).
Non-REM sleep.
As for wakefulness, there was no statistically significant suppressant effect of 5-HT receptor antagonism at the HMN on GG activity. There was clearly no effect of 5-HT receptor antagonism on respiratory-related GG activity (Figure 3C; F1,9 = 0.072, p = 0.795), although the effect on tonic GG activity was closer to statistical significance (Figure 4C; F1,9 = 3.41, p = 0.098). Nevertheless, despite being closer to statistical significance, this trend was also in a direction opposite to that predicted by the hypothesis (Figure 4C), providing further support for the concept of a minimal role of 5-HT mechanisms providing a physiologically relevant endogenous excitatory 5-HT drive to GG muscle in freely behaving rats in non-REM sleep.
During non-REM sleep, the stimulating effect of CO2 on respiratory-related GG activity was also readily apparent (Figure 3C; F1,9 = 14.25, p = 0.004), as was the case for quiet wakefulness (see above). As expected (33), there was no effect of CO2 stimulation on tonic GG activity (Figure 4C; F1,9 = 2.47, p = 0.151) or neck muscle activity (F1,9 = 1.41, p = 0.266). The expected significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity in response to CO2 stimulation were all observed (all F1,9 > 25.23, p < 0.001).
REM sleep.
In REM sleep, there was also no statistically significant effect of 5-HT receptor antagonism at the HMN on either respiratory-related or tonic GG activities (Figures 3D and 4D, respectively; both F1,9 < 2.21, p > 0.166). Although the data show a trend for increased GG activity after 5-HT antagonism at the HMN in the presence of hypercapnia in REM sleep (Figures 3D and 4D), there was no statistically significant interaction between the presence of mianserin at the HMN and inspired CO2 level (F1,7 < 3.93, p > 0.088; i.e., mianserin had no statistically significant effect on GG activity, whether the animal was breathing room air or CO2-enriched air).
Despite the robust increases in respiratory-related GG activity observed during non-REM sleep and wakefulness with CO2-stimulated breathing, such a CO2-mediated increase in GG activity was not apparent in REM sleep (Figure 3D; F1,9 = 0.277, p = 0.610). This result is in agreement with the previously observed major suppression of GG activity in REM sleep despite strong respiratory stimulation by hypercapnia (33). Also consistent with that previous study (33), significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity still occurred in response to CO2 stimulation in REM sleep (all F1,9 > 35.40, p < 0.001).
GG Activity across Sleep–Wake States with Mianserin at the HMN
The analysis described above provides evidence for a minimal effect of 5-HT receptor antagonism at the HMN on GG activity in freely behaving rats within individual states of wakefulness and sleep. However, that analysis precluded observations of the changes in GG activity across sleep–wake states and the effects of 5-HT receptor antagonism. The results of such an analysis across sleep–wake states are described below.
Room Air.
During room-air breathing, there was no statistically significant effect of 5-HT receptor antagonism on the changes in respiratory-related (Figure 5A) or tonic (Figure 5B) GG activities recorded across sleep–wake states (F1,9 < 1.21, p > 0.301). This result confirmed the separate analyses performed above for the individual states of wakefulness and sleep. Importantly, although there was a statistically significant modulating effect of sleep–wake states per se on both respiratory-related and tonic GG activities (F2,18 > 4.65, p < 0.024), there was no statistically significant interaction with the presence of 5-HT receptor antagonism at the HMN (F2,17 < 1.82, p > 0.192; i.e., the changes in GG activities from wakefulness to sleep were statistically indistinguishable between ACSF controls and mianserin, implying a minimal role for 5-HT receptor mechanisms modulating GG activity across sleep–wake states during basal breathing). Post hoc analysis confirmed a significant decline in respiratory-related and tonic GG activities from wakefulness to non-REM sleep (Figures 5A and 5B; p < 0.035) and from wakefulness to REM sleep for tonic GG activity (Figure 5B; p = 0.021). Although post hoc tests also showed that overall GG activities were similar between non-REM and REM sleep (p > 0.451), this result reflected that the GG values were obtained from REM sleep episodes that contained both the periods of major suppression of GG activity and the transient GG muscle twitches that typify this state, as shown in Figure 2C and documented in previous studies (13, 31, 32).
As controls for the changes in physiologic parameters across sleep–wake states we observed the expected significant alterations in neck muscle activity (Figure 6; F2,18 = 22.03, p < 0.001), with progressive declines from wakefulness to non-REM sleep (p = 0.008, post hoc test) and from non-REM to REM sleep (p = 0.002). The changes in other sleep and breathing variables are analyzed in the last section of RESULTS and are summarized in Figure 6.
CO2-stimulated breathing.
There was evidence for a statistically significant effect of 5-HT receptor antagonism at the HMN on respiratory-related GG activity across sleep–wake states, indicative of a serotonergic contribution to GG muscle tone, but this effect was observed only under conditions of increased drive to the HMN caused by hypercapnia. Although there was no overall difference in the levels of respiratory-related or tonic GG activities between ACSF and 5-HT receptor antagonism at the HMN during hypercapnia (Figures 5C and 5D; F1,9 < 2.25, p > 0.167), in confirmation of the analyses performed separately for the individual states of wakefulness and sleep described above (see Figures 3B–3D and 4B–4D), there was the expected significant modulation of GG activities across sleep–wake states (both F2,18 > 4.04, p < 0.036). Importantly, however, the change in respiratory-related GG activity across these sleep–wake states was different for mianserin compared with ACSF (F2,17 = 10.35, p = 0.001). Post hoc analyses showed that, with ACSF at the HMN, there were significant declines in respiratory-related GG activity from non-REM to REM sleep (p = 0.004) and from wakefulness to REM sleep (p < 0.001), whereas these normal declines in respiratory-related GG activity were abolished with 5-HT receptor antagonism (Figure 5C; all p > 0.254). In contrast to GG activity, however, the normal changes in neck muscle activity across states of wakefulness and sleep were preserved (Figure 6; F2,18 = 21.88, p < 0.001, and described below), showing that the effects were specific to 5-HT receptor antagonism at the HMN.
Changes in Control Variables across Sleep–Wake States
Figure 6 shows the grouped mean data for the effects of mianserin versus ACSF at the HMN on respiratory and sleep-related parameters during both room-air and CO2-stimulated breathing. Importantly, in both room-air and CO2-stimulated breathing, there was no statistically significant independent effect of mianserin on respiratory rate, respiratory-related diaphragm activity, neck muscle activity, or the ratio of high to low frequencies in the EEG (i.e., % 2/% 1) (all F1,9 < 2.99, p > 0.116). This demonstrates that any effects at the HMN were specific to the GG muscle. The independent effects of sleep and awake states on respiratory variables (e.g., the changes in respiratory rates in REM sleep) and sleep variables (e.g., the changes in EEG frequencies and neck EMG activities), and their response to CO2 stimulation, were typical and are as previously described (13, 33).
Responses to MDL100907 at the HMN
Microdialysis perfusion of the more specific 5-HT2 receptor antagonist MDL100907 (35) into the HMN also failed to decrease GG activity across sleep and awake states. Data are shown in Figure 7 for respiratory-related GG activity, and a similar lack of effect of MDL100907 was observed for tonic GG activity. The overall results are similar to those obtained with mianserin (compare Figures 3 and 7). Statistical analysis confirmed that MDL100907 had no effect on respiratory-related or tonic GG activities compared with ACSF controls (all F1,2 < 13.72, p > 0.065). Interestingly, the only comparison that approached borderline statistical significance (p = 0.066; all others, p > 0.115) was the increased GG activity after introduction of MDL100907 during hypercapnia in REM sleep, which was also observed with mianserin (Figure 3D); as described above, mianserin reduced the normal decline in GG activity in REM sleep during hypercapnia. Therefore, these results with MDL100907 are consistent with those obtained with mianserin, and support the concept of a minimal endogenous 5-HT drive at the HMN in these intact animals unless augmented by reflex inputs.
GG Responses to 5-HT Receptor Antagonism at the HMN with the Vagus Nerves Intact or Cut
Figure 8 shows an example of the effects of 5-HT receptor antagonism at the HMN with mianserin on GG activity in an anesthetized rat with the vagus nerves intact. Note the minimal suppression of GG activity with 5-HT receptor antagonism in this rat compared with the same intervention in a rat with the vagus nerves cut. Figure 9 shows group data for the effects of 5-HT receptor antagonism at the HMN with mianserin in the vagotomized and vagus nerve–intact rats. These data show that the effect of 5-HT receptor antagonism at the HMN on GG activity depended on whether the vagus nerves were intact or cut (Figure 9; F1,14 = 7.17, p = 0.018). Post hoc analyses confirmed that GG activity decreased significantly with mianserin at the HMN in the vagotomized rats (p < 0.001), but the decrease was small and statistically insignificant in those with intact vagus nerves (p = 0.601). These data strongly support the concept of a normally low endogenous 5-HT drive to the HMN in intact animals.
DISCUSSION
The results show that during normal room-air breathing, 5-HT receptor antagonism at the HMN has minimal effects on respiratory-related and tonic GG activities in freely behaving rats, both within individual states of wakefulness and sleep as well as across states. In contrast, under conditions of increased drive to the HMN caused by hypercapnia, there was a statistically significant effect of 5-HT receptor antagonism at the HMN on respiratory-related GG activity across sleep–wake states, indicative of a serotonergic contribution to GG muscle tone; in hypercapnia, the normal significant declines in respiratory-related GG activity from non-REM to REM sleep, and from wakefulness to REM, with ACSF were reduced with mianserin at the HMN (Figure 5). Taken together, these results suggest a normally low endogenous 5-HT tone at the HMN having a minimal effect on GG activity unless this tone is augmented by reflex inputs. Hypercapnia would be expected to cause a reflex increase in 5-HT drive to the HMN, as some medullary raphe neurons show increased activity in response to CO2 stimulation, both in vitro (25) and in vivo (24). In addition, we have shown using the same techniques that 5-HT at the HMN plays a role in the full manifestation of the GG response to CO2 stimulation in acutely anesthetized animals (12).
5-HT and Modulation of GG Activity
Although application of exogenous 5-HT excites hypoglossal motoneurons in vitro (9), and in decerebrate (10, 11), anesthetized (12, 35), and behaving animals in vivo (13), no previous study has assessed the physiologic role of endogenous 5-HT acting directly at the HMN to modulate motor outflow in an intact, freely behaving, and naturally sleeping preparation. However, unlike the present results in intact animals that show a normally low endogenous 5-HT tone at the HMN, studies in reduced animal preparations have demonstrated a significant serotonergic modulation of pharyngeal muscle activity during basal breathing. For example, in decerebrate or anesthetized, paralyzed, vagotomized, and artificially ventilated cats or rats, 5-HT receptor antagonism at the HMN significantly reduced hypoglossal nerve activity (10, 35). In the pharmacologic model of REM sleep produced by microinjection of carbachol into the pontine reticular formation of decerebrate and vagotomized cats, the reduced activity of 5-HT raphe neurons (22) is also associated with reduced 5-HT at the HMN (27), also suggesting the presence of a tonic serotonergic drive. In agreement with this result, destruction or reversible blockade of raphe neurons in decerebrate (or anesthetized), vagotomized piglets also reduces hypoglossal nerve activity (26). Importantly, we also showed that 5-HT receptor antagonism at the HMN decreased GG activity during room-air and CO2-stimulated breathing in a previous study performed using exactly the same techniques and dose of mianserin as the present study, albeit in vagotomized, anesthetized rats, as opposed to intact, conscious rats (12). Overall, the common features of these studies, showing a robust tonic 5-HT drive to the HMN-modulating GG activity during basal breathing, are that they were performed in reduced (i.e., anesthetized or decerebrate) preparations, and that the animals were vagotomized.
Although there are some, albeit few, lines of evidence suggesting that anesthesia or decerebration could have augmented medullary raphe activity, leading to a larger 5-HT drive to the HMN compared with awake or naturally sleeping animals (39, 40), there is evidence suggesting that vagotomy can increase raphe activity. In vagotomized animals, at least 50% of medullary raphe neurons show respiratory-related activity (41), with about one-third of these also stimulated by chemoreceptor stimuli (42). Because most medullary raphe neurons that project to motoneurons are inhibited by vagal afferents (43, 44), the removal of this inhibitory influence by vagotomy could have increased raphe modulation of GG activity compared with vagus nerve–intact and behaving preparations. In support of this potential effect of vagotomy increasing serotonergic activity, the reduction in 5-HT at the HMN after introduction of pontine carbachol in the pharmacologic model of REM sleep was only 29% in vagus nerve–intact cats (28) versus 68–58% in vagotomized cats (27) when data obtained with the same microdialysis probes in each study were compared. Additional experiments in the present study showed that there was a significant reduction in GG activity after 5-HT receptor antagonism at the HMN in anesthetized, vagotomized rats, but not in vagus nerve–intact rats (Figures 8 and 9). These results support the concept of a normally low endogenous 5-HT drive to the HMN in intact preparations unless increased by vagotomy.
The concept derived from the present study of a normally low endogenous 5-HT tone modulating GG activity in intact animals, unless driven by reflex inputs, is also supported by previous studies using systemically administered serotonergic drugs in intact preparations. Human patients with OSA exhibit a neuromuscular compensatory mechanism for a narrow airway that increases pharyngeal muscle activity in wakefulness and promotes airway patency (45, 46). English bulldogs, a breed with a narrow airway and sleep-disordered breathing, also exhibit increased pharyngeal muscle activity (47), as do anesthetized obese Zucker rats (48), which have a narrower (49) and more collapsible upper airway compared with lean control animals. This increased GG activity in awake bulldogs and anesthetized obese Zucker rats is diminished by intravenous ritanserin, a 5-HT receptor antagonist, indicating a serotonergic component to the response (48, 50). However, the decrements in GG activity after ritanserin in anesthetized obese rats (48) are not observed in lean rats (51), suggesting that different endogenous 5-HT mechanisms may be operating in conditions associated with an altered upper airway, such as obesity. In addition, ritanserin increased work of breathing in obese but not lean Zucker rats, an effect that was attributed to effects on the upper airway (48). Together, these studies support the concept that 5-HT receptor antagonism at the HMN has minimal effects on GG activity unless there is augmented serotonergic drive to the HMN. In our study this augmented 5-HT drive was produced by hypercapnia, and in bulldogs and obese Zucker rats by compensation for a chronically narrowed upper airway. Whether the increased GG activity recorded in anesthetized obese Zucker rats (48) persists in awake animals and is mediated by 5-HT remains to be determined.
Specificity of Responses
During CO2-stimulated breathing, the changes in GG activity observed from quiet wakefulness to non-REM sleep and REM sleep with 5-HT receptor antagonism at the HMN occurred without any associated changes in neck muscle activity (as a control postural muscle), diaphragm activity (as a control respiratory muscle), or EEG activity (nonspecific marker of arousal; i.e., responses to interventions at the HMN were specific to GG muscle). Measurement of these control variables was also important, as they showed that the unexpected increase in GG activity observed with 5-HT receptor antagonism, when all of the awake periods were analyzed together (i.e., regardless of quiet wakefulness or active behaviors), was most likely explained by a nonspecific, generalized motor activation. This is because, during these same periods, there was also increased activity in neck muscle, whereas this was not observed in quiet wakefulness. The GG responses to 5-HT receptor antagonism also tended toward increased rather than the hypothesized decreased GG activity compared with ACSF (Figures 3 and 4), further suggesting that the absence of a statistically significant decrease under basal breathing was not due to borderline significant results; the large quantity of data analyzed per rat suggests that we would have detected such a decrease if one were present. Interestingly, increased GG activity has also been observed in anesthetized rats after systemic administration of mirtazapine, which has 5-HT receptor antagonist properties (52).
In the present study, we chose to apply mianserin, a broad spectrum 5-HT receptor antagonist (53), because the endogenous excitatory 5-HT drive to the HMN is effectively antagonized by mianserin in decerebrate (10) or anesthetized and vagotomized animals (12), with the latter study having been performed using the same methodology as the present experiments in behaving animals. Mianserin was also chosen because we had difficulty in dissolving other more specific 5-HT-receptor antagonists, such as ritanserin (50), in ACSF. It is a concern, however, that use of this broad-spectrum 5-HT antagonist may have masked a true suppressant effect on GG activity, as it has some 2 and 5-HT1B antagonist properties, receptors that exert inhibitory effects at the HMN (21, 54, 55). Therefore, it is possible, albeit unlikely given the results in reduced preparations (10, 12), that a more specific antagonist for those excitatory 5-HT2 receptors on hypoglossal motoneurons (10, 56, 57) might have yielded different results. Accordingly, in three additional rats, the more specific 5-HT2 receptor antagonist MDL100907 was perfused into the HMN at a dose 10 times higher than that which causes major (62%) suppression of hypoglossal motor activity in anesthetized and vagotomized rats (35). Similar to the effects observed with mianserin, MDL100907 caused no suppression of GG activity in these intact, behaving animals (Figure 7) a result that is consistent with a minimal endogenous excitatory 5-HT drive to the HMN across sleep–wake states.
Raphe Neurons and 5-HT Modulation of GG Activity
Although raphe neurons exhibit well-established changes in neuronal discharge across sleep–wake states (23), the magnitude of change is different for dorsal raphe neurons involved in ascending brain arousal compared with the caudal medullary groups modulating motor and autonomic functions (23). For example, dorsal raphe activity in cats is strongly modulated by changes in sleep–wake state, with large alterations in discharge frequency paralleling the changes in behavioral state (58), whereas caudal raphe neurons are only weakly modulated from wakefulness to non-REM sleep, although there remains a large suppression of activity in REM sleep (59). Serotonergic caudal raphe neurons are also believed to be more activated during motor tasks associated with rhythmic behaviors, such as chewing and licking (60, 61), and may have a lesser role in modulating the respiratory component of GG activity during basal breathing (21). Nevertheless, although the results of the present study show only a weak serotonergic modulation of the HMN during unstimulated breathing in this intact, freely behaving rodent preparation, this does not contradict the concept that raphe neurons per se can significantly influence GG activity in a sleep- and wakefulness-dependent manner, as nonserotonergic raphe neurons also project to the HMN (62–64). In this respect, raphe neurons have been implicated in the state-dependent modulation of ventilation and responses to CO2 (65), although the relative roles of serotonergic versus nonserotonergic mechanisms in mediating these effects are less clear. Because raphe neurons also contain the excitatory neurotransmitters thyrotropin-releasing hormone and substance P (62, 66), modulation of these transmitters may themselves contribute to changes in pharyngeal motor tone across sleep–wake states. In addition, in vitro evidence suggests that a major component of the raphe pallidus inputs to hypoglossal motoneurons are glutamatergic, with 5-HT inhibiting the release of the excitatory neurotransmitter glutamate from these projections via presynaptic effects (63).
Clinical Relevance and Conclusions
Overall, the absence of a physiologically relevant endogenous 5-HT drive acting at the HMN to modulate GG activity across states of wakefulness and sleep, unless augmented by reflex inputs, has clinical relevance to the potential efficacy of attempts to modulate endogenous 5-HT levels with SSRIs as a pharmacologic treatment of OSA. In general, however, although there are some improvements in indices of sleep-disordered breathing after SSRI administration in humans, such responses have been of limited clinical benefit (5, 6). Decreases in apnea–hypopnea index have been observed in non-REM sleep (16, 17), although not consistently (18), and there is some direct (19) and indirect (18) evidence for augmented pharyngeal muscle tone.
Nevertheless, it is possible that the potential role of 5-HT in modulating pharyngeal muscle activity in vivo has been overestimated from previous animal experiments in reduced preparations because of the effects of vagotomy. As such, SSRIs may exert only weak effects in some humans if similar mechanisms are operative. It is also possible that the magnitude of 5-HT–dependent inputs to pharyngeal motoneurons may differ between individuals, depending on whether they have a primary anatomical predisposition to OSA and whether 5-HT is involved in the compensatory reflex responses to maintain an open airspace. This compensation would be expected to be reduced in sleep as 5-HT neuronal activity is decreased. Indeed, many patients with OSA exhibit increased GG activity during wakefulness, indicating neuromuscular compensation for a narrow airspace (45, 46); compensatory reflexes are likely involved in increasing upper airway motor activity (45, 67). If such reflex compensations involve 5-HT, then certain patients with OSA may benefit from strategies to augment 5-HT activity (e.g., via SSRIs). It has also been demonstrated that repeated bouts of respiratory stimulation by hypoxia cause 5-HT–mediated augmentations of hypoglossal motor activity (68) that may contribute to the chronic augmentations of GG activity in patients with OSA. These data fit with the concept that, although 5-HT may normally contribute little to baseline GG activation of pharyngeal motoneurons, reflex responses may increase the role of 5-HT–mediated activation. Such effects may explain why in bulldogs, a breed with compromised upper airway anatomy, there is significant worsening of sleep-disordered breathing in after 5-HT receptor antagonism (50).
FOOTNOTES
Supported by the Canadian Institutes of Health Research (grant MT-15563).
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.200502-258OC on July 14, 2005
Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
REFERENCES
Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230–1235.
Sanders MH. Medical therapy for obstructive sleep apnea–hypopnea syndrome. In: Kryger MH, Roth T, and Dement WC, editors. Principles and Practice of Sleep Medicine, 3rd ed. Philadelphia: WB Saunders; 2000. pp. 879–893.
Hudgel DW, Thanakitcharu S. Pharmacologic treatment of sleep-disordered breathing. Am J Respir Crit Care Med 1998;158:691–699.
Smith IE, Quinnell TG. Pharmacotherapies for obstructive sleep apnoea: where are we now Drugs 2004;64:1385–1399.
Horner RL. The neuropharmacology of upper airway motor control in the awake and asleep states: implications for obstructive sleep apnoea. Respir Res 2001;2:286–294.
Veasey SC. Serotonin agonists and antagonists in obstructive sleep apnea: therapeutic potential. Am J Respir Med 2003;2:21–29.
Schmidt HS. L-tryptophan in the treatment of impaired respiration in sleep. Bull Eur Physiopathol Respir 1983;19:625–629.
Veasey SC, Fenik P, Panckeri K, Pack AI, Hendricks JC. The effects of trazodone with L-tryptophan on sleep-disordered breathing in the English bulldog. Am J Respir Crit Care Med 1999;160:1659–1667.
Berger AJ, Bayliss DA, Viana F. Modulation of neonatal rat hypoglossal motoneuron excitability by serotonin. Neurosci Lett 1992;143:164–168.
Kubin L, Tojima H, Davies RO, Pack AI. Serotonergic excitatory drive to hypoglossal motoneurons in the decerebrate cat. Neurosci Lett 1992;139:243–248.
Douse MA, White DP. Serotonergic effects on hypoglossal neural activity and reflex responses. Brain Res 1996;726:213–222.
Sood S, Liu X, Liu H, Nolan P, Horner RL. 5-HT at hypoglossal motor nucleus and respiratory control of genioglossus muscle in anesthetized rats. Respir Physiolo Neurobiol 2003;138:205–221.
Jelev A, Sood S, Liu H, Nolan P, Horner RL. Microdialysis perfusion of 5-HT into hypoglossal motor nucleus differentially modulates genioglossus activity across natural sleep–wake states in rats. J Physiol 2001;532:467–481.
Berry RB, Hayward LF. Selective augmentation of genioglossus electromyographic activity by L-5-hydroxytryptophan in the rat. Pharmacol Biochem Behav 2003;74:877–882.
Ogasa T, Ray AD, Michlin CP, Farkas GA, Grant BJ, Magalang UJ. Systemic administration of serotonin 2A/2C agonist improves upper airway stability in Zucker rats. Am J Respir Crit Care Med 2004;170:804–810.
Hanzel DA, Proia NG, Hudgel DW. Response of obstructive sleep apnea to fluoxetine and protriptyline. Chest 1991;100:416–421.
Kraiczi H, Hedner J, Dahlof P, Ejnell H, Carlson J. Effect of serotonin uptake inhibition on breathing during sleep and daytime symptoms in obstructive sleep apnea. Sleep 1999;22:61–67.
Berry RB, Yamaura EM, Gill K, Reist C. Acute effects of paroxetine on genioglossus activity in obstructive sleep apnea. Sleep 1999;22:1087–1092.
Sunderram J, Parisi RA, Strobel RJ. Serotonergic stimulation of the genioglossus and the response to nasal continuous positive airway pressure. Am J Respir Crit Care Med 2000;162:925–929.
Singer JH, Bellingham MC, Berger AJ. Presynaptic inhibition of glutamatergic synaptic transmission to rat motoneurons by serotonin. J Neurophysiol 1996;76:799–807.
Singer JH, Berger AJ. Presynaptic inhibition by serotonin: a possible mechanism for switching motor output of the hypoglossal nucleus. Sleep 1996;19:S146–S149.
Woch G, Davies RO, Pack AI, Kubin L. Behaviour of raphe cells projecting to the dorsomedial medulla during carbachol-induced atonia in the cat. J Physiol 1996;490:745–758.
Jacobs BL, Azmitia EC. Structure and function of the brain serotonin system. Physiol Rev 1992;72:165–229.
Veasey SC, Fornal CA, Metzler CW, Jacobs BL. Response of serotonergic caudal raphe neurons in relation to specific motor activities in freely moving cats. J Neurosci 1995;15:5346–5359.
Wang W, Pizzonia JH, Richerson GB. Chemosensitivity of rat medullary raphe neurones in primary tissue culture. J Physiol 1998;511:433–450.
Dreshaj IA, Haxhiu MA, Martin RJ. Role of the medullary raphe nuclei in the respiratory response to CO2. Respir Physiol 1998;111:15–23.
Kubin L, Reignier C, Tojima H, Taguchi O, Pack AI, Davies RO. Changes in serotonin level in the hypoglossal nucleus region during carbachol-induced atonia. Brain Res 1994;645:291–302.
Lai YY, Kodama T, Siegel JM. Changes in monoamine release in the ventral horn and hypoglossal nucleus linked to pontine inhibition of muscle tone: an in vivo microdialysis study. J Neurosci 2001;21:7384–7391.
Kubin L, Davies RO, Pack L. Control of upper airway motoneurons during REM sleep. News Physiol Sci 1998;13:637–656.
Sood S, Raddatz E, Liu X, Liu H, Horner RL. Modulation of respiratory motor output by serotonergic neurons of the medullary raphe obscurus . Sleep 2004;27(Suppl.):A28.
Morrison JL, Sood S, Liu H, Park E, Nolan P, Horner RL. Role of inhibitory amino acids in control of hypoglossal motor outflow to genioglossus muscle in naturally sleeping rats. J Physiol 2003;552:975–991.
Morrison JL, Sood S, Liu H, Park E, Nolan P, Horner RL. GABA-A receptor antagonism at the hypoglossal motor nucleus increases genioglossus muscle activity in NREM but not REM sleep. J Physiol 2003;548:569–583.
Horner RL, Liu X, Gill H, Nolan P, Liu H, Sood S. Effects of sleep–wake state on the genioglossus vs. diaphragm muscle response to CO2 in rats. J Appl Physiol 2002;92:878–887.
Wang W, Tiwari JK, Bradley SR, Zaykin RV, Richerson GB. Acidosis-stimulated neurons of the medullary raphe are serotonergic. J Neurophysiol 2001;85:2224–2235.
Fenik P, Veasey SC. Pharmacological characterization of serotonergic receptor activity in the hypoglossal nucleus. Am J Respir Crit Care Med 2003;167:563–569.
Richard CA, Harper RM. Respiratory-related activity in hypoglossal neurons across sleep-waking states in cats. Brain Res 1991;542:167–170.
Goh AS, Issa FG, Sullivan CE. Upper airway dilating forces during wakefulness and sleep in dogs. J Appl Physiol 1986;61:2148–2155.
Younes M. Contributions of upper airway mechanics and control mechanisms to severity of obstructive apnea. Am J Respir Crit Care Med 2003;168:645–658.
Mosko SS, Jacobs BL. Electrophysiological evidence against negative neuronal feedback from the forebrain controlling midbrain raphe unit activity. Brain Res 1977;119:291–303.
Roda F, Pio J, Bianchi AL, Gestreau C. Effects of anesthetics on hypoglossal nerve discharge and c-Fos expression in brainstem hypoglossal premotor neurons. J Comp Neurol 2004;468:571–586.
Lindsey BG, Hernandez YM, Morris KF, Shannon R. Functional connectivity between brain stem midline neurons with respiratory-modulated firing rates. J Neurophysiol 1992;67:890–904.
Li Z, Morris KF, Baekey DM, Shannon R, Lindsey BG. Responses of simultaneously recorded respiratory-related medullary neurons to stimulation of multiple sensory modalities. J Neurophysiol 1999;82:176–187.
Blair RW, Evans AR. Responses of medullary raphespinal neurons to electrical stimulation of thoracic sympathetic afferents, vagal afferents, and to other sensory inputs in cats. J Neurophysiol 1991;66:2084–2094.
Evans AR, Blair RW. Responses of medullary raphe neurons to electrical and chemical activation of vagal afferent nerve fibers. J Neurophysiol 1993;70:1950–1961.
Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal electromyogram in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism). J Clin Invest 1992;89:1571–1579.
Suratt PM, McTier RF, Wilhoit SC. Upper airway muscle activation is augmented in patients with obstructive sleep apnea compared with that in normal subjects. Am Rev Respir Dis 1988;137:889–894.
Hendricks JC, Petrof BJ, Panckeri K, Pack AI. Upper airway dilating muscle hyperactivity during non–rapid eye movement sleep in English bulldogs. Am Rev Respir Dis 1993;148:185–194.
Nakano H, Magalang UJ, Lee SD, Krasney JA, Farkas GA. Serotonergic modulation of ventilation and upper airway stability in obese Zucker rats. Am J Respir Crit Care Med 2001;163:1191–1197.
Magalang UJ, Farkas GA, Najdzionek JS, Nakano H. Obese Zucker rats have narrower upper airways compared to lean littermates. Am J Respir Crit Care Med 2000;161:A87.
Veasey SC, Panckeri KA, Hoffman EA, Pack AI, Hendricks JC. The effects of serotonin antagonists in an animal model of sleep-disordered breathing. Am J Respir Crit Care Med 1996;153:776–786.
Richmonds CR, Hudgel DW. Hypoglossal and phrenic motoneuron responses to serotonergic active agents in rats. Respir Physiol 1996;106:153–160.
Berry RB, Koch GL, Hayward LF. Low-dose mirtazapine increases genioglossus activity in the anesthetized rat. Sleep 2005;28:78–84.
Zifa E, Fillion G. 5-Hydroxytryptamine receptors. Pharmacol Rev 1992;44:401–458.
Adachi T, Robinson DM, Miles GB, Funk GD. Noradrenergic modulation of XII motoneuron inspiratory activity does not involve alpha2-receptor inhibition of the Ih current or presynaptic glutamate release. J Appl Physiol 2005;98:1297–1308.
Okabe S, Kubin L. Role of 5HT1 receptors in the control of hypoglossal motoneurons in vivo. Sleep 1996;19:S150–S153.
Okabe S, Mackiewicz M, Kubin L. Serotonin receptor mRNA expression in the hypoglossal motor nucleus. Respir Physiol 1997;110:151–160.
Zhan G, Shaheen F, Mackiewicz M, Fenik P, Veasey SC. Single cell laser dissection with molecular beacon polymerase chain reaction identifies 2A as the predominant serotonin receptor subtype in hypoglossal motoneurons. Neuroscience 2002;113:145–154.
Trulson ME, Jacobs BL. Raphe unit activity in freely moving cats: correlation with level of behavioral arousal. Brain Res 1979;163:135–150.
Heym J, Steinfels GF, Jacobs BL. Activity of serotonin-containing neurons in the nucleus raphe pallidus of freely moving cats. Brain Res 1982;251:259–276.
Ribeiro-do-Valle LE, Lucena RL. Behavioral correlates of the activity of serotonergic and non-serotonergic neurons in caudal raphe nuclei. Braz J Med Biol Res 2001;34:919–937.
Jacobs BL, Fornal CA. 5-HT and motor control: a hypothesis. Trends Neurosci 1993;16:346–352.
Henry JN, Manaker S. Colocalization of substance P or enkephalin in serotonergic neuronal afferents to the hypoglossal nucleus in the rat. J Comp Neurol 1998;391:491–505.
Bouryi VA, Lewis DI. The modulation by 5-HT of glutamatergic inputs from the raphe pallidus to rat hypoglossal motoneurones, in vitro. J Physiol 2003;553:1019–1031.
Bouryi VA, Lewis DI. Enkephalinergic inhibition of raphe pallidus inputs to rat hypoglossal motoneurones in vitro. Neuroscience 2004;129:55–64.
Nattie E. Multiple sites for central chemoreception: their roles in response sensitivity and in sleep and wakefulness. Respir Physiol 2000;122:223–235.
Johansson O, Hokfelt T, Pernow B, Jeffcoate SL, White N, Steinbusch HW, Verhofstad AA, Emson PC, Spindel E. Immunohistochemical support for three putative transmitters in one neuron: coexistence of 5-hydroxytryptamine, substance P- and thyrotropin releasing hormone-like immunoreactivity in medullary neurons projecting to the spinal cord. Neuroscience 1981;6:1857–1881.
Fogel RB, Malhotra A, Shea SA, Edwards JK, White DP. Reduced genioglossal activity with upper airway anesthesia in awake patients with OSA. J Appl Physiol 2000;88:1346–1354.
Bocchiaro CM, Feldman JL. Synaptic activity-independent persistent plasticity in endogenously active mammalian motoneurons. Proc Natl Acad Sci USA 2004;101:4292–4295.(Sandeep Sood, Janna L. Morrison, Hattie )
ABSTRACT
Rationale: Exogenous serotonin at the hypoglossal motor nucleus (HMN) stimulates genioglossus (GG) muscle activity. However, whether endogenous serotonin contributes to GG activation across natural sleep–wake states has not been determined, but is relevant given that serotonergic neurons have decreased activity in sleep and project to pharyngeal motoneurons.
Objectives: To determine the role of endogenous serotonin at the HMN in modulating GG activity across natural sleep–wake states.
Methods: Ten rats were implanted with electroencephalogram and neck muscle electrodes to record sleep–wake states, and GG and diaphragm wires for respiratory muscle recordings. Microdialysis probes were implanted into the HMN for perfusion of artificial cerebrospinal fluid and the serotonin receptor antagonist mianserin (100 μM).
Measurements and Main Results: In room air, there was no effect of mianserin on respiratory-related or tonic GG activities across sleep–wake states (p > 0.300). In hypercapnia, however, the normal declines in GG activity from non-REM to REM sleep, and wakefulness to REM sleep, were reduced with mianserin (p < 0.005). These data demonstrate a normally low endogenous serotonergic drive modulating GG activity unless augmented by reflex inputs. We also demonstrated a significant serotonergic drive modulating GG activity in vagotomized rats, but not in vagi-intact rats, under anesthesia, suggesting that previous results in reduced preparations may have been influenced by vagotomy.
Conclusions: The results show a minimal endogenous serotonergic drive at the HMN modulating GG activity across sleep–wake states, unless augmented by reflex inputs. This result has implications for pharmacologic strategies aiming to increase GG activity by manipulating endogenous serotonin in patients with obstructive sleep apnea.
Key Words: hypoglossal motor nucleus obstructive sleep apnea serotonin sleep
Obstructive sleep apnea (OSA) is a common disorder affecting up to 4% of adults (1). There has been increasing interest in development of pharmacologic treatments for OSA, although initial studies using a variety of agents, such as ventilatory stimulants, have been largely unsuccessful (2–4). Given the sleep-state dependence of this disorder, however, newer approaches have attempted to modify those sleep state–dependent neural systems modulating pharyngeal motoneurons (5, 6). Nevertheless, the relative success or failure of such approaches ultimately depends on knowledge of the relevant neurotransmitters and, importantly, whether they provide a physiologically relevant drive to the pharyngeal muscles across sleep–wake states and are amenable to intervention.
For clinical studies, most attention has focused on the potential role of serotonin (5-hydroxtytryptamine [5-HT]) in the pharmacologic treatment of OSA because of its stimulatory effects on pharyngeal motoneurons. Accordingly, one approach has been to increase such stimulation via application of broad-spectrum or receptor-specific 5-HT agonists, or agents to increase 5-HT production (7, 8). The rationale for this approach comes from basic studies in animal preparations. In vitro studies of brainstem tissue slices show that 5-HT depolarizes and increases the excitability of hypoglossal motoneurons (9), the source of motor outflow to the genioglossus (GG) muscle of the tongue. Likewise, 5-HT at the hypoglossal motor nucleus (HMN) in decerebrate cats (10, 11) and anesthetized rats (12) increases motor output to GG. In freely behaving and naturally sleeping rats, tonic GG activation occurs when 5-HT is applied directly to the HMN, with the increased GG activity in non-REM sleep maintained for as long as 5-HT is applied (i.e., several hours), although the response is attenuated in REM sleep (13). Although it will be difficult to target the pharyngeal motoneurons specifically with systemic interventions in conscious animals, systemic administration of 5-HT precursors (14) or 5-HT2 receptor agonists (15) increase GG activity in acutely anesthetized rats; however, significant elevations in blood pressure have been reported (15).
A second major approach in the serotonergic modulation of pharyngeal motoneurons in patients with OSA has been to increase endogenous 5-HT via systemic administration of selective serotonin reuptake inhibitors (SSRIs) (16–19). However, it is again difficult to target the relevant pharyngeal motoneurons specifically with systemic interventions, with responses observed clinically being potentially confounded by influences on sleep or indirect effects on presynaptic and autoreceptor elements (that can, themselves, influence pharyngeal motor activity [5, 20–22]). Nevertheless, the conceptual rationale for this approach comes from basic studies in animal preparations. Medullary raphe neurons provide tonic 5-HT inputs to hypoglossal motoneurons, with these neurons showing declining discharge rates from wakefulness to non-REM and REM sleep (23) and activation by arousing stimuli, such as CO2 (24, 25), which contribute to the GG muscle responses to hypercapnia in anesthetized or decerebrate animals (12, 26). There is decreased discharge of medullary raphe neurons projecting to the HMN in a pharmacologic model of REM sleep in vagotomized and decerebrate cats (22). This change is accompanied by reduced 5-HT at the HMN in association with reduced hypoglossal nerve activity (27, 28). These observations in reduced animal preparations have led to the concept that increased raphe neuronal activity in wakefulness contributes to a physiologically relevant 5-HT drive to the HMN that is then withdrawn in sleep—especially REM sleep (29). However, it has not yet been tested whether 5-HT receptor antagonism at the HMN will decrease GG activity in wakefulness and sleep in a manner consistent with a significant role for endogenous 5-HT in the physiologic control of GG muscle activity, and this hypothesis is the focus of the present study. These results have relevance for pharmacologic strategies aiming to modulate GG activity by manipulating endogenous 5-HT in patients with OSA. Some of the results of this study have previously been reported in the form of an abstract (30).
METHODS
A more detailed account of the methods is given in the online supplement. Experiments were performed on 13 male Wistar rats (mean body weight = 261 ± [SEM] 7.3 g). All procedures conformed to the recommendations of the Canadian Council on Animal Care, and the University of Toronto Animal Care Committee approved the experimental protocols.
Anesthesia and Surgical Procedures
Sterile surgery was performed under general anesthesia for the chronic implantation of EEG and neck EMG electrodes to determine sleep–wake states, and GG and diaphragm electrodes for respiratory muscle recordings (13, 31, 32). Tests for the accurate placement of the GG electrodes and their function throughout the experiments (13, 31–33) are described in the online supplement. During surgery, microdialysis guides were targeted 3 mm above the HMN (13, 31, 32). The rats recovered for an average of 6.5 d (range, 5–8 d) before the studies.
Protocol and Data Analysis
On the day of the experiment, the microdialysis probe was inserted into the HMN (13, 31, 32). Measurements of sleep–wake states and respiratory muscle activities were made during both room-air and CO2-stimulated breathing using 7.5% inspired CO2 (31–33). Measurements were also made during hypercapnia, as 5-HT raphe neurons are activated by CO2 (24, 34) and contribute to the GG responses to respiratory stimulation in reduced preparations (12, 26). In 10 rats, the probes were flushed at a flow rate of 2.1 μl · min–1 with artificial cerebrospinal fluid (ACSF) or mianserin (5-HT receptor antagonist, 100 μM) dissolved in ACSF. Importantly, we have shown, using the exact same methodology, that 100 μM mianserin at the HMN in anesthetized rats produces robust decreases in GG activity during both room-air breathing and hypercapnia (12). Additional experiments in three rats were performed with the more specific 5-HT2 receptor antagonist MDL100907 at the HMN. The applied dose of 20 μM is 10 times higher than that causing major (62%) suppression of hypoglossal activity in anesthetized rats (35).
Sleep–wake states and respiratory muscle activities were analyzed as previously described (31–33) and are detailed in the online supplement. GG activity was quantified as tonic and respiratory-related activity. Data were analyzed at least 60 min after a switch between drugs. After the experiments, the rats were reanesthetized and tests for GG electrode function were performed (see online supplement). The sites of microdialysis were confirmed with histology (13, 31–33).
Experiments in Acutely Anesthetized Rats
Previous studies demonstrating a significant endogenous 5-HT drive to the HMN were performed in anesthetized or decerebrate vagotomized animals (10, 12, 22, 26, 27, 35). To characterize the potential role of the vagus nerves influencing GG responses to 5-HT receptor antagonism at the HMN, we performed additional experiments in 6 urethane-anesthetized rats with the vagus nerves intact and compared responses with those obtained in 10 vagotomized rats using the same methodology (12).
Statistical Analysis
Data were analyzed using ananlysis of variance with repeated measures, and a Student-Newman-Keuls test was used for post hoc comparisons. Analyses were performed using SigmaStat (SPSS, Inc., Chicago, IL). All data are expressed as mean ± SEM.
RESULTS
The full experimental protocol was completed in 11 of the 13 rats studied. In the two remaining rats, the mianserin protocol was also complete, except that REM sleep was not observed when ACSF was microdialized into the HMN in one rat, nor was REM sleep observed when mianserin was microdialized during hypercapnia in the other rat. For the group of 10 mianserin rats, a total of 44,053 5-s epochs (i.e., a total of 61.2 h of data) were included in the analysis, of which 29,111 epochs (66%) were from periods of wakefulness, 11,902 epochs (27%) were from non-REM sleep, and 3,040 epochs (7%) were from REM sleep. In the three rats with MDL100907 at the HMN, a total of 3,474 5-s epochs of wakefulness, 5,225 epochs of non-REM sleep, and 902 epochs of REM sleep were analyzed.
Function of GG Electrodes and Histology
Stimulation of the GG electrodes at the end of the experiment showed that the voltages required to cause tongue movements were not different from those at the time of surgery (0.80 ± 0.07 vs. 0.77 ± 0.05 V), demonstrating that the electrodes were in place and functional throughout the study.
Figure 1A shows an example of a lesion site made by the microdialysis probe in the HMN, with the dark stain produced by the potassium permanganate marking the probe site. Figure 1B shows the locations of all the microdialysis sites from each of the 10 mianserin and 3 MDL100907 rats. In all the experiments, the microdialysis sites were within the HMN or in the midline immediately adjacent to both the nuclei.
GG Activity across Sleep and Awake States
An example of the effects of sleep and awake states on respiratory muscle activity during room-air and CO2-stimulated breathing is shown in Figures 2A and 2B, respectively. Note the presence of prominent tonic GG activity during room-air breathing in wakefulness, as well as discernable respiratory-related GG activity (i.e., patterns typical of previous studies [31–33]). Also note that GG activity is markedly decreased from wakefulness to non-REM sleep, and is effectively abolished in those periods of REM sleep without the transient GG muscle twitches that also occur in REM sleep (see below). Figure 2 also shows that although GG activity is increased during CO2-stimulated breathing, similar changes in GG activity occur from wakefulness to non-REM sleep, including the periods of major suppression of GG activity in REM sleep.
Figure 2C shows an example of the change in GG activity at the transition from non-REM to REM sleep, as well as the variations in GG activity during REM sleep. This example illustrates that the periods of major suppression of GG activity in REM sleep typically occur at the onset of the REM sleep episodes, with the transient GG muscle twitches occurring later in the REM episodes. Such changes are typical. They have been analyzed in detail in previous articles (31, 32) and also occur in other animals, such as cats (36) and dogs (37). In addition, they are believed to be responsible for the sporadic restorations of airflow during obstructive apneas in REM sleep in humans (38).
Effects of 5-HT Receptor Antagonism with Mianserin on GG Activity
Wakefulness.
When the large quantities of data recorded in each individual rat were analyzed (i.e., those including both quiet wakefulness and active behaviors), there was a significant effect of 5-HT receptor antagonism at the HMN on respiratory-related and tonic GG activities (Figures 3A and 4A, respectively; both F1,9 > 5.33, p < 0.047) that did not depend on whether the animal was breathing room air or CO2 (both F1,9 < 1.31, p > 0.283). However, this overall effect of 5-HT receptor antagonism was in the opposite direction to that predicted by the hypothesis (i.e., mianserin increased rather than decreased respiratory-related and tonic GG activities; Figures 3A and 4A).
Further analysis was performed to determine if this result was indeed due to a specific effect of 5-HT receptor antagonism at the HMN or was likely an epiphenomenon secondary to a nonspecific generalized motor activation caused by the rat being engaged in more motor behaviors during wakefulness. If the latter were the case, there would also be an expected, concomitant activation of an unrelated control postural muscle as well as GG. This additional analysis showed that, during these same periods of wakefulness, there was also a significant increase in the activity of the control postural muscle in the neck (F1,9 = 15.24, p = 0.004) that also did not depend on whether the animal was breathing room or CO2-enriched air (F1,9 = 0.16, p = 0.701). This result suggested that the increased GG activity recorded at the time of 5-HT receptor antagonism at the HMN, when all of the wakefulness epochs were analyzed together (i.e., those including both quiet wakefulness and active behaviors), was most likely due to nonspecific effects related to active behaviors that would not otherwise have been detected without reference to the control muscle activity in the neck. Such a nonspecific general motor activation also likely explains the absence of a measurable increase in respiratory-related GG activity in response to CO2 when all the awake epochs were analyzed together (Figure 3A, F1,9 = 1.03, p = 0.337), as the active behaviors obscured any respiratory response. Nevertheless, the expected significant increases in indices of lung ventilation were observed in hypercapnia; hypercapnic respiratory stimulation increased diaphragm amplitude, respiratory rate, and diaphragm minute activity (all F1,9 > 52.41, p < 0.001). For further description, see CHANGES IN CONTROL VARIABLES ACROSS SLEEP–WAKE STATES.
To control for the overall potential confounding influence of active behaviors, obscuring the GG muscle responses to 5-HT receptor antagonism at the HMN and responses to CO2, further analyses were performed in those periods of relaxed wakefulness in which the rats were not physically active and engaged in behaviors.
Wakefulness without active behaviors.
In contrast to the main hypothesis of the study, there was no evidence of a significant suppressant effect of 5-HT receptor antagonism at the HMN on either respiratory-related or tonic GG activities in periods of quiet wakefulness (Figures 3B and 4B, respectively; both F1,9 < 1.10, p > 0.322). During these periods of wakefulness, however, the stimulating effects of CO2 on respiratory-related GG activities were readily apparent (Figure 3B; F1,9 = 9.86, p = 0.012), unlike during active behaviors (compare Figures 3A and 3B, and see above). There was no statistically significant effect of CO2 stimulation on tonic GG activity (Figure 4B; F1,9 = 1.16, p = 0.310) or neck muscle activity (F1,9 = 2.97, p = 0.119) as expected from previous results (33). The expected significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity in response to CO2 stimulation were all observed (all F1,9 > 24.69, p < 0.001).
Non-REM sleep.
As for wakefulness, there was no statistically significant suppressant effect of 5-HT receptor antagonism at the HMN on GG activity. There was clearly no effect of 5-HT receptor antagonism on respiratory-related GG activity (Figure 3C; F1,9 = 0.072, p = 0.795), although the effect on tonic GG activity was closer to statistical significance (Figure 4C; F1,9 = 3.41, p = 0.098). Nevertheless, despite being closer to statistical significance, this trend was also in a direction opposite to that predicted by the hypothesis (Figure 4C), providing further support for the concept of a minimal role of 5-HT mechanisms providing a physiologically relevant endogenous excitatory 5-HT drive to GG muscle in freely behaving rats in non-REM sleep.
During non-REM sleep, the stimulating effect of CO2 on respiratory-related GG activity was also readily apparent (Figure 3C; F1,9 = 14.25, p = 0.004), as was the case for quiet wakefulness (see above). As expected (33), there was no effect of CO2 stimulation on tonic GG activity (Figure 4C; F1,9 = 2.47, p = 0.151) or neck muscle activity (F1,9 = 1.41, p = 0.266). The expected significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity in response to CO2 stimulation were all observed (all F1,9 > 25.23, p < 0.001).
REM sleep.
In REM sleep, there was also no statistically significant effect of 5-HT receptor antagonism at the HMN on either respiratory-related or tonic GG activities (Figures 3D and 4D, respectively; both F1,9 < 2.21, p > 0.166). Although the data show a trend for increased GG activity after 5-HT antagonism at the HMN in the presence of hypercapnia in REM sleep (Figures 3D and 4D), there was no statistically significant interaction between the presence of mianserin at the HMN and inspired CO2 level (F1,7 < 3.93, p > 0.088; i.e., mianserin had no statistically significant effect on GG activity, whether the animal was breathing room air or CO2-enriched air).
Despite the robust increases in respiratory-related GG activity observed during non-REM sleep and wakefulness with CO2-stimulated breathing, such a CO2-mediated increase in GG activity was not apparent in REM sleep (Figure 3D; F1,9 = 0.277, p = 0.610). This result is in agreement with the previously observed major suppression of GG activity in REM sleep despite strong respiratory stimulation by hypercapnia (33). Also consistent with that previous study (33), significant increases in diaphragm amplitude, respiratory rate, and diaphragm minute activity still occurred in response to CO2 stimulation in REM sleep (all F1,9 > 35.40, p < 0.001).
GG Activity across Sleep–Wake States with Mianserin at the HMN
The analysis described above provides evidence for a minimal effect of 5-HT receptor antagonism at the HMN on GG activity in freely behaving rats within individual states of wakefulness and sleep. However, that analysis precluded observations of the changes in GG activity across sleep–wake states and the effects of 5-HT receptor antagonism. The results of such an analysis across sleep–wake states are described below.
Room Air.
During room-air breathing, there was no statistically significant effect of 5-HT receptor antagonism on the changes in respiratory-related (Figure 5A) or tonic (Figure 5B) GG activities recorded across sleep–wake states (F1,9 < 1.21, p > 0.301). This result confirmed the separate analyses performed above for the individual states of wakefulness and sleep. Importantly, although there was a statistically significant modulating effect of sleep–wake states per se on both respiratory-related and tonic GG activities (F2,18 > 4.65, p < 0.024), there was no statistically significant interaction with the presence of 5-HT receptor antagonism at the HMN (F2,17 < 1.82, p > 0.192; i.e., the changes in GG activities from wakefulness to sleep were statistically indistinguishable between ACSF controls and mianserin, implying a minimal role for 5-HT receptor mechanisms modulating GG activity across sleep–wake states during basal breathing). Post hoc analysis confirmed a significant decline in respiratory-related and tonic GG activities from wakefulness to non-REM sleep (Figures 5A and 5B; p < 0.035) and from wakefulness to REM sleep for tonic GG activity (Figure 5B; p = 0.021). Although post hoc tests also showed that overall GG activities were similar between non-REM and REM sleep (p > 0.451), this result reflected that the GG values were obtained from REM sleep episodes that contained both the periods of major suppression of GG activity and the transient GG muscle twitches that typify this state, as shown in Figure 2C and documented in previous studies (13, 31, 32).
As controls for the changes in physiologic parameters across sleep–wake states we observed the expected significant alterations in neck muscle activity (Figure 6; F2,18 = 22.03, p < 0.001), with progressive declines from wakefulness to non-REM sleep (p = 0.008, post hoc test) and from non-REM to REM sleep (p = 0.002). The changes in other sleep and breathing variables are analyzed in the last section of RESULTS and are summarized in Figure 6.
CO2-stimulated breathing.
There was evidence for a statistically significant effect of 5-HT receptor antagonism at the HMN on respiratory-related GG activity across sleep–wake states, indicative of a serotonergic contribution to GG muscle tone, but this effect was observed only under conditions of increased drive to the HMN caused by hypercapnia. Although there was no overall difference in the levels of respiratory-related or tonic GG activities between ACSF and 5-HT receptor antagonism at the HMN during hypercapnia (Figures 5C and 5D; F1,9 < 2.25, p > 0.167), in confirmation of the analyses performed separately for the individual states of wakefulness and sleep described above (see Figures 3B–3D and 4B–4D), there was the expected significant modulation of GG activities across sleep–wake states (both F2,18 > 4.04, p < 0.036). Importantly, however, the change in respiratory-related GG activity across these sleep–wake states was different for mianserin compared with ACSF (F2,17 = 10.35, p = 0.001). Post hoc analyses showed that, with ACSF at the HMN, there were significant declines in respiratory-related GG activity from non-REM to REM sleep (p = 0.004) and from wakefulness to REM sleep (p < 0.001), whereas these normal declines in respiratory-related GG activity were abolished with 5-HT receptor antagonism (Figure 5C; all p > 0.254). In contrast to GG activity, however, the normal changes in neck muscle activity across states of wakefulness and sleep were preserved (Figure 6; F2,18 = 21.88, p < 0.001, and described below), showing that the effects were specific to 5-HT receptor antagonism at the HMN.
Changes in Control Variables across Sleep–Wake States
Figure 6 shows the grouped mean data for the effects of mianserin versus ACSF at the HMN on respiratory and sleep-related parameters during both room-air and CO2-stimulated breathing. Importantly, in both room-air and CO2-stimulated breathing, there was no statistically significant independent effect of mianserin on respiratory rate, respiratory-related diaphragm activity, neck muscle activity, or the ratio of high to low frequencies in the EEG (i.e., % 2/% 1) (all F1,9 < 2.99, p > 0.116). This demonstrates that any effects at the HMN were specific to the GG muscle. The independent effects of sleep and awake states on respiratory variables (e.g., the changes in respiratory rates in REM sleep) and sleep variables (e.g., the changes in EEG frequencies and neck EMG activities), and their response to CO2 stimulation, were typical and are as previously described (13, 33).
Responses to MDL100907 at the HMN
Microdialysis perfusion of the more specific 5-HT2 receptor antagonist MDL100907 (35) into the HMN also failed to decrease GG activity across sleep and awake states. Data are shown in Figure 7 for respiratory-related GG activity, and a similar lack of effect of MDL100907 was observed for tonic GG activity. The overall results are similar to those obtained with mianserin (compare Figures 3 and 7). Statistical analysis confirmed that MDL100907 had no effect on respiratory-related or tonic GG activities compared with ACSF controls (all F1,2 < 13.72, p > 0.065). Interestingly, the only comparison that approached borderline statistical significance (p = 0.066; all others, p > 0.115) was the increased GG activity after introduction of MDL100907 during hypercapnia in REM sleep, which was also observed with mianserin (Figure 3D); as described above, mianserin reduced the normal decline in GG activity in REM sleep during hypercapnia. Therefore, these results with MDL100907 are consistent with those obtained with mianserin, and support the concept of a minimal endogenous 5-HT drive at the HMN in these intact animals unless augmented by reflex inputs.
GG Responses to 5-HT Receptor Antagonism at the HMN with the Vagus Nerves Intact or Cut
Figure 8 shows an example of the effects of 5-HT receptor antagonism at the HMN with mianserin on GG activity in an anesthetized rat with the vagus nerves intact. Note the minimal suppression of GG activity with 5-HT receptor antagonism in this rat compared with the same intervention in a rat with the vagus nerves cut. Figure 9 shows group data for the effects of 5-HT receptor antagonism at the HMN with mianserin in the vagotomized and vagus nerve–intact rats. These data show that the effect of 5-HT receptor antagonism at the HMN on GG activity depended on whether the vagus nerves were intact or cut (Figure 9; F1,14 = 7.17, p = 0.018). Post hoc analyses confirmed that GG activity decreased significantly with mianserin at the HMN in the vagotomized rats (p < 0.001), but the decrease was small and statistically insignificant in those with intact vagus nerves (p = 0.601). These data strongly support the concept of a normally low endogenous 5-HT drive to the HMN in intact animals.
DISCUSSION
The results show that during normal room-air breathing, 5-HT receptor antagonism at the HMN has minimal effects on respiratory-related and tonic GG activities in freely behaving rats, both within individual states of wakefulness and sleep as well as across states. In contrast, under conditions of increased drive to the HMN caused by hypercapnia, there was a statistically significant effect of 5-HT receptor antagonism at the HMN on respiratory-related GG activity across sleep–wake states, indicative of a serotonergic contribution to GG muscle tone; in hypercapnia, the normal significant declines in respiratory-related GG activity from non-REM to REM sleep, and from wakefulness to REM, with ACSF were reduced with mianserin at the HMN (Figure 5). Taken together, these results suggest a normally low endogenous 5-HT tone at the HMN having a minimal effect on GG activity unless this tone is augmented by reflex inputs. Hypercapnia would be expected to cause a reflex increase in 5-HT drive to the HMN, as some medullary raphe neurons show increased activity in response to CO2 stimulation, both in vitro (25) and in vivo (24). In addition, we have shown using the same techniques that 5-HT at the HMN plays a role in the full manifestation of the GG response to CO2 stimulation in acutely anesthetized animals (12).
5-HT and Modulation of GG Activity
Although application of exogenous 5-HT excites hypoglossal motoneurons in vitro (9), and in decerebrate (10, 11), anesthetized (12, 35), and behaving animals in vivo (13), no previous study has assessed the physiologic role of endogenous 5-HT acting directly at the HMN to modulate motor outflow in an intact, freely behaving, and naturally sleeping preparation. However, unlike the present results in intact animals that show a normally low endogenous 5-HT tone at the HMN, studies in reduced animal preparations have demonstrated a significant serotonergic modulation of pharyngeal muscle activity during basal breathing. For example, in decerebrate or anesthetized, paralyzed, vagotomized, and artificially ventilated cats or rats, 5-HT receptor antagonism at the HMN significantly reduced hypoglossal nerve activity (10, 35). In the pharmacologic model of REM sleep produced by microinjection of carbachol into the pontine reticular formation of decerebrate and vagotomized cats, the reduced activity of 5-HT raphe neurons (22) is also associated with reduced 5-HT at the HMN (27), also suggesting the presence of a tonic serotonergic drive. In agreement with this result, destruction or reversible blockade of raphe neurons in decerebrate (or anesthetized), vagotomized piglets also reduces hypoglossal nerve activity (26). Importantly, we also showed that 5-HT receptor antagonism at the HMN decreased GG activity during room-air and CO2-stimulated breathing in a previous study performed using exactly the same techniques and dose of mianserin as the present study, albeit in vagotomized, anesthetized rats, as opposed to intact, conscious rats (12). Overall, the common features of these studies, showing a robust tonic 5-HT drive to the HMN-modulating GG activity during basal breathing, are that they were performed in reduced (i.e., anesthetized or decerebrate) preparations, and that the animals were vagotomized.
Although there are some, albeit few, lines of evidence suggesting that anesthesia or decerebration could have augmented medullary raphe activity, leading to a larger 5-HT drive to the HMN compared with awake or naturally sleeping animals (39, 40), there is evidence suggesting that vagotomy can increase raphe activity. In vagotomized animals, at least 50% of medullary raphe neurons show respiratory-related activity (41), with about one-third of these also stimulated by chemoreceptor stimuli (42). Because most medullary raphe neurons that project to motoneurons are inhibited by vagal afferents (43, 44), the removal of this inhibitory influence by vagotomy could have increased raphe modulation of GG activity compared with vagus nerve–intact and behaving preparations. In support of this potential effect of vagotomy increasing serotonergic activity, the reduction in 5-HT at the HMN after introduction of pontine carbachol in the pharmacologic model of REM sleep was only 29% in vagus nerve–intact cats (28) versus 68–58% in vagotomized cats (27) when data obtained with the same microdialysis probes in each study were compared. Additional experiments in the present study showed that there was a significant reduction in GG activity after 5-HT receptor antagonism at the HMN in anesthetized, vagotomized rats, but not in vagus nerve–intact rats (Figures 8 and 9). These results support the concept of a normally low endogenous 5-HT drive to the HMN in intact preparations unless increased by vagotomy.
The concept derived from the present study of a normally low endogenous 5-HT tone modulating GG activity in intact animals, unless driven by reflex inputs, is also supported by previous studies using systemically administered serotonergic drugs in intact preparations. Human patients with OSA exhibit a neuromuscular compensatory mechanism for a narrow airway that increases pharyngeal muscle activity in wakefulness and promotes airway patency (45, 46). English bulldogs, a breed with a narrow airway and sleep-disordered breathing, also exhibit increased pharyngeal muscle activity (47), as do anesthetized obese Zucker rats (48), which have a narrower (49) and more collapsible upper airway compared with lean control animals. This increased GG activity in awake bulldogs and anesthetized obese Zucker rats is diminished by intravenous ritanserin, a 5-HT receptor antagonist, indicating a serotonergic component to the response (48, 50). However, the decrements in GG activity after ritanserin in anesthetized obese rats (48) are not observed in lean rats (51), suggesting that different endogenous 5-HT mechanisms may be operating in conditions associated with an altered upper airway, such as obesity. In addition, ritanserin increased work of breathing in obese but not lean Zucker rats, an effect that was attributed to effects on the upper airway (48). Together, these studies support the concept that 5-HT receptor antagonism at the HMN has minimal effects on GG activity unless there is augmented serotonergic drive to the HMN. In our study this augmented 5-HT drive was produced by hypercapnia, and in bulldogs and obese Zucker rats by compensation for a chronically narrowed upper airway. Whether the increased GG activity recorded in anesthetized obese Zucker rats (48) persists in awake animals and is mediated by 5-HT remains to be determined.
Specificity of Responses
During CO2-stimulated breathing, the changes in GG activity observed from quiet wakefulness to non-REM sleep and REM sleep with 5-HT receptor antagonism at the HMN occurred without any associated changes in neck muscle activity (as a control postural muscle), diaphragm activity (as a control respiratory muscle), or EEG activity (nonspecific marker of arousal; i.e., responses to interventions at the HMN were specific to GG muscle). Measurement of these control variables was also important, as they showed that the unexpected increase in GG activity observed with 5-HT receptor antagonism, when all of the awake periods were analyzed together (i.e., regardless of quiet wakefulness or active behaviors), was most likely explained by a nonspecific, generalized motor activation. This is because, during these same periods, there was also increased activity in neck muscle, whereas this was not observed in quiet wakefulness. The GG responses to 5-HT receptor antagonism also tended toward increased rather than the hypothesized decreased GG activity compared with ACSF (Figures 3 and 4), further suggesting that the absence of a statistically significant decrease under basal breathing was not due to borderline significant results; the large quantity of data analyzed per rat suggests that we would have detected such a decrease if one were present. Interestingly, increased GG activity has also been observed in anesthetized rats after systemic administration of mirtazapine, which has 5-HT receptor antagonist properties (52).
In the present study, we chose to apply mianserin, a broad spectrum 5-HT receptor antagonist (53), because the endogenous excitatory 5-HT drive to the HMN is effectively antagonized by mianserin in decerebrate (10) or anesthetized and vagotomized animals (12), with the latter study having been performed using the same methodology as the present experiments in behaving animals. Mianserin was also chosen because we had difficulty in dissolving other more specific 5-HT-receptor antagonists, such as ritanserin (50), in ACSF. It is a concern, however, that use of this broad-spectrum 5-HT antagonist may have masked a true suppressant effect on GG activity, as it has some 2 and 5-HT1B antagonist properties, receptors that exert inhibitory effects at the HMN (21, 54, 55). Therefore, it is possible, albeit unlikely given the results in reduced preparations (10, 12), that a more specific antagonist for those excitatory 5-HT2 receptors on hypoglossal motoneurons (10, 56, 57) might have yielded different results. Accordingly, in three additional rats, the more specific 5-HT2 receptor antagonist MDL100907 was perfused into the HMN at a dose 10 times higher than that which causes major (62%) suppression of hypoglossal motor activity in anesthetized and vagotomized rats (35). Similar to the effects observed with mianserin, MDL100907 caused no suppression of GG activity in these intact, behaving animals (Figure 7) a result that is consistent with a minimal endogenous excitatory 5-HT drive to the HMN across sleep–wake states.
Raphe Neurons and 5-HT Modulation of GG Activity
Although raphe neurons exhibit well-established changes in neuronal discharge across sleep–wake states (23), the magnitude of change is different for dorsal raphe neurons involved in ascending brain arousal compared with the caudal medullary groups modulating motor and autonomic functions (23). For example, dorsal raphe activity in cats is strongly modulated by changes in sleep–wake state, with large alterations in discharge frequency paralleling the changes in behavioral state (58), whereas caudal raphe neurons are only weakly modulated from wakefulness to non-REM sleep, although there remains a large suppression of activity in REM sleep (59). Serotonergic caudal raphe neurons are also believed to be more activated during motor tasks associated with rhythmic behaviors, such as chewing and licking (60, 61), and may have a lesser role in modulating the respiratory component of GG activity during basal breathing (21). Nevertheless, although the results of the present study show only a weak serotonergic modulation of the HMN during unstimulated breathing in this intact, freely behaving rodent preparation, this does not contradict the concept that raphe neurons per se can significantly influence GG activity in a sleep- and wakefulness-dependent manner, as nonserotonergic raphe neurons also project to the HMN (62–64). In this respect, raphe neurons have been implicated in the state-dependent modulation of ventilation and responses to CO2 (65), although the relative roles of serotonergic versus nonserotonergic mechanisms in mediating these effects are less clear. Because raphe neurons also contain the excitatory neurotransmitters thyrotropin-releasing hormone and substance P (62, 66), modulation of these transmitters may themselves contribute to changes in pharyngeal motor tone across sleep–wake states. In addition, in vitro evidence suggests that a major component of the raphe pallidus inputs to hypoglossal motoneurons are glutamatergic, with 5-HT inhibiting the release of the excitatory neurotransmitter glutamate from these projections via presynaptic effects (63).
Clinical Relevance and Conclusions
Overall, the absence of a physiologically relevant endogenous 5-HT drive acting at the HMN to modulate GG activity across states of wakefulness and sleep, unless augmented by reflex inputs, has clinical relevance to the potential efficacy of attempts to modulate endogenous 5-HT levels with SSRIs as a pharmacologic treatment of OSA. In general, however, although there are some improvements in indices of sleep-disordered breathing after SSRI administration in humans, such responses have been of limited clinical benefit (5, 6). Decreases in apnea–hypopnea index have been observed in non-REM sleep (16, 17), although not consistently (18), and there is some direct (19) and indirect (18) evidence for augmented pharyngeal muscle tone.
Nevertheless, it is possible that the potential role of 5-HT in modulating pharyngeal muscle activity in vivo has been overestimated from previous animal experiments in reduced preparations because of the effects of vagotomy. As such, SSRIs may exert only weak effects in some humans if similar mechanisms are operative. It is also possible that the magnitude of 5-HT–dependent inputs to pharyngeal motoneurons may differ between individuals, depending on whether they have a primary anatomical predisposition to OSA and whether 5-HT is involved in the compensatory reflex responses to maintain an open airspace. This compensation would be expected to be reduced in sleep as 5-HT neuronal activity is decreased. Indeed, many patients with OSA exhibit increased GG activity during wakefulness, indicating neuromuscular compensation for a narrow airspace (45, 46); compensatory reflexes are likely involved in increasing upper airway motor activity (45, 67). If such reflex compensations involve 5-HT, then certain patients with OSA may benefit from strategies to augment 5-HT activity (e.g., via SSRIs). It has also been demonstrated that repeated bouts of respiratory stimulation by hypoxia cause 5-HT–mediated augmentations of hypoglossal motor activity (68) that may contribute to the chronic augmentations of GG activity in patients with OSA. These data fit with the concept that, although 5-HT may normally contribute little to baseline GG activation of pharyngeal motoneurons, reflex responses may increase the role of 5-HT–mediated activation. Such effects may explain why in bulldogs, a breed with compromised upper airway anatomy, there is significant worsening of sleep-disordered breathing in after 5-HT receptor antagonism (50).
FOOTNOTES
Supported by the Canadian Institutes of Health Research (grant MT-15563).
This article has an online supplement, which is accessible from this issue's table of contents at www.atsjournals.org
Originally Published in Press as DOI: 10.1164/rccm.200502-258OC on July 14, 2005
Conflict of Interest Statement: None of the authors have a financial relationship with a commercial entity that has an interest in the subject of this manuscript.
REFERENCES
Young T, Palta M, Dempsey J, Skatrud J, Weber S, Badr S. The occurrence of sleep-disordered breathing among middle-aged adults. N Engl J Med 1993;328:1230–1235.
Sanders MH. Medical therapy for obstructive sleep apnea–hypopnea syndrome. In: Kryger MH, Roth T, and Dement WC, editors. Principles and Practice of Sleep Medicine, 3rd ed. Philadelphia: WB Saunders; 2000. pp. 879–893.
Hudgel DW, Thanakitcharu S. Pharmacologic treatment of sleep-disordered breathing. Am J Respir Crit Care Med 1998;158:691–699.
Smith IE, Quinnell TG. Pharmacotherapies for obstructive sleep apnoea: where are we now Drugs 2004;64:1385–1399.
Horner RL. The neuropharmacology of upper airway motor control in the awake and asleep states: implications for obstructive sleep apnoea. Respir Res 2001;2:286–294.
Veasey SC. Serotonin agonists and antagonists in obstructive sleep apnea: therapeutic potential. Am J Respir Med 2003;2:21–29.
Schmidt HS. L-tryptophan in the treatment of impaired respiration in sleep. Bull Eur Physiopathol Respir 1983;19:625–629.
Veasey SC, Fenik P, Panckeri K, Pack AI, Hendricks JC. The effects of trazodone with L-tryptophan on sleep-disordered breathing in the English bulldog. Am J Respir Crit Care Med 1999;160:1659–1667.
Berger AJ, Bayliss DA, Viana F. Modulation of neonatal rat hypoglossal motoneuron excitability by serotonin. Neurosci Lett 1992;143:164–168.
Kubin L, Tojima H, Davies RO, Pack AI. Serotonergic excitatory drive to hypoglossal motoneurons in the decerebrate cat. Neurosci Lett 1992;139:243–248.
Douse MA, White DP. Serotonergic effects on hypoglossal neural activity and reflex responses. Brain Res 1996;726:213–222.
Sood S, Liu X, Liu H, Nolan P, Horner RL. 5-HT at hypoglossal motor nucleus and respiratory control of genioglossus muscle in anesthetized rats. Respir Physiolo Neurobiol 2003;138:205–221.
Jelev A, Sood S, Liu H, Nolan P, Horner RL. Microdialysis perfusion of 5-HT into hypoglossal motor nucleus differentially modulates genioglossus activity across natural sleep–wake states in rats. J Physiol 2001;532:467–481.
Berry RB, Hayward LF. Selective augmentation of genioglossus electromyographic activity by L-5-hydroxytryptophan in the rat. Pharmacol Biochem Behav 2003;74:877–882.
Ogasa T, Ray AD, Michlin CP, Farkas GA, Grant BJ, Magalang UJ. Systemic administration of serotonin 2A/2C agonist improves upper airway stability in Zucker rats. Am J Respir Crit Care Med 2004;170:804–810.
Hanzel DA, Proia NG, Hudgel DW. Response of obstructive sleep apnea to fluoxetine and protriptyline. Chest 1991;100:416–421.
Kraiczi H, Hedner J, Dahlof P, Ejnell H, Carlson J. Effect of serotonin uptake inhibition on breathing during sleep and daytime symptoms in obstructive sleep apnea. Sleep 1999;22:61–67.
Berry RB, Yamaura EM, Gill K, Reist C. Acute effects of paroxetine on genioglossus activity in obstructive sleep apnea. Sleep 1999;22:1087–1092.
Sunderram J, Parisi RA, Strobel RJ. Serotonergic stimulation of the genioglossus and the response to nasal continuous positive airway pressure. Am J Respir Crit Care Med 2000;162:925–929.
Singer JH, Bellingham MC, Berger AJ. Presynaptic inhibition of glutamatergic synaptic transmission to rat motoneurons by serotonin. J Neurophysiol 1996;76:799–807.
Singer JH, Berger AJ. Presynaptic inhibition by serotonin: a possible mechanism for switching motor output of the hypoglossal nucleus. Sleep 1996;19:S146–S149.
Woch G, Davies RO, Pack AI, Kubin L. Behaviour of raphe cells projecting to the dorsomedial medulla during carbachol-induced atonia in the cat. J Physiol 1996;490:745–758.
Jacobs BL, Azmitia EC. Structure and function of the brain serotonin system. Physiol Rev 1992;72:165–229.
Veasey SC, Fornal CA, Metzler CW, Jacobs BL. Response of serotonergic caudal raphe neurons in relation to specific motor activities in freely moving cats. J Neurosci 1995;15:5346–5359.
Wang W, Pizzonia JH, Richerson GB. Chemosensitivity of rat medullary raphe neurones in primary tissue culture. J Physiol 1998;511:433–450.
Dreshaj IA, Haxhiu MA, Martin RJ. Role of the medullary raphe nuclei in the respiratory response to CO2. Respir Physiol 1998;111:15–23.
Kubin L, Reignier C, Tojima H, Taguchi O, Pack AI, Davies RO. Changes in serotonin level in the hypoglossal nucleus region during carbachol-induced atonia. Brain Res 1994;645:291–302.
Lai YY, Kodama T, Siegel JM. Changes in monoamine release in the ventral horn and hypoglossal nucleus linked to pontine inhibition of muscle tone: an in vivo microdialysis study. J Neurosci 2001;21:7384–7391.
Kubin L, Davies RO, Pack L. Control of upper airway motoneurons during REM sleep. News Physiol Sci 1998;13:637–656.
Sood S, Raddatz E, Liu X, Liu H, Horner RL. Modulation of respiratory motor output by serotonergic neurons of the medullary raphe obscurus . Sleep 2004;27(Suppl.):A28.
Morrison JL, Sood S, Liu H, Park E, Nolan P, Horner RL. Role of inhibitory amino acids in control of hypoglossal motor outflow to genioglossus muscle in naturally sleeping rats. J Physiol 2003;552:975–991.
Morrison JL, Sood S, Liu H, Park E, Nolan P, Horner RL. GABA-A receptor antagonism at the hypoglossal motor nucleus increases genioglossus muscle activity in NREM but not REM sleep. J Physiol 2003;548:569–583.
Horner RL, Liu X, Gill H, Nolan P, Liu H, Sood S. Effects of sleep–wake state on the genioglossus vs. diaphragm muscle response to CO2 in rats. J Appl Physiol 2002;92:878–887.
Wang W, Tiwari JK, Bradley SR, Zaykin RV, Richerson GB. Acidosis-stimulated neurons of the medullary raphe are serotonergic. J Neurophysiol 2001;85:2224–2235.
Fenik P, Veasey SC. Pharmacological characterization of serotonergic receptor activity in the hypoglossal nucleus. Am J Respir Crit Care Med 2003;167:563–569.
Richard CA, Harper RM. Respiratory-related activity in hypoglossal neurons across sleep-waking states in cats. Brain Res 1991;542:167–170.
Goh AS, Issa FG, Sullivan CE. Upper airway dilating forces during wakefulness and sleep in dogs. J Appl Physiol 1986;61:2148–2155.
Younes M. Contributions of upper airway mechanics and control mechanisms to severity of obstructive apnea. Am J Respir Crit Care Med 2003;168:645–658.
Mosko SS, Jacobs BL. Electrophysiological evidence against negative neuronal feedback from the forebrain controlling midbrain raphe unit activity. Brain Res 1977;119:291–303.
Roda F, Pio J, Bianchi AL, Gestreau C. Effects of anesthetics on hypoglossal nerve discharge and c-Fos expression in brainstem hypoglossal premotor neurons. J Comp Neurol 2004;468:571–586.
Lindsey BG, Hernandez YM, Morris KF, Shannon R. Functional connectivity between brain stem midline neurons with respiratory-modulated firing rates. J Neurophysiol 1992;67:890–904.
Li Z, Morris KF, Baekey DM, Shannon R, Lindsey BG. Responses of simultaneously recorded respiratory-related medullary neurons to stimulation of multiple sensory modalities. J Neurophysiol 1999;82:176–187.
Blair RW, Evans AR. Responses of medullary raphespinal neurons to electrical stimulation of thoracic sympathetic afferents, vagal afferents, and to other sensory inputs in cats. J Neurophysiol 1991;66:2084–2094.
Evans AR, Blair RW. Responses of medullary raphe neurons to electrical and chemical activation of vagal afferent nerve fibers. J Neurophysiol 1993;70:1950–1961.
Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal electromyogram in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism). J Clin Invest 1992;89:1571–1579.
Suratt PM, McTier RF, Wilhoit SC. Upper airway muscle activation is augmented in patients with obstructive sleep apnea compared with that in normal subjects. Am Rev Respir Dis 1988;137:889–894.
Hendricks JC, Petrof BJ, Panckeri K, Pack AI. Upper airway dilating muscle hyperactivity during non–rapid eye movement sleep in English bulldogs. Am Rev Respir Dis 1993;148:185–194.
Nakano H, Magalang UJ, Lee SD, Krasney JA, Farkas GA. Serotonergic modulation of ventilation and upper airway stability in obese Zucker rats. Am J Respir Crit Care Med 2001;163:1191–1197.
Magalang UJ, Farkas GA, Najdzionek JS, Nakano H. Obese Zucker rats have narrower upper airways compared to lean littermates. Am J Respir Crit Care Med 2000;161:A87.
Veasey SC, Panckeri KA, Hoffman EA, Pack AI, Hendricks JC. The effects of serotonin antagonists in an animal model of sleep-disordered breathing. Am J Respir Crit Care Med 1996;153:776–786.
Richmonds CR, Hudgel DW. Hypoglossal and phrenic motoneuron responses to serotonergic active agents in rats. Respir Physiol 1996;106:153–160.
Berry RB, Koch GL, Hayward LF. Low-dose mirtazapine increases genioglossus activity in the anesthetized rat. Sleep 2005;28:78–84.
Zifa E, Fillion G. 5-Hydroxytryptamine receptors. Pharmacol Rev 1992;44:401–458.
Adachi T, Robinson DM, Miles GB, Funk GD. Noradrenergic modulation of XII motoneuron inspiratory activity does not involve alpha2-receptor inhibition of the Ih current or presynaptic glutamate release. J Appl Physiol 2005;98:1297–1308.
Okabe S, Kubin L. Role of 5HT1 receptors in the control of hypoglossal motoneurons in vivo. Sleep 1996;19:S150–S153.
Okabe S, Mackiewicz M, Kubin L. Serotonin receptor mRNA expression in the hypoglossal motor nucleus. Respir Physiol 1997;110:151–160.
Zhan G, Shaheen F, Mackiewicz M, Fenik P, Veasey SC. Single cell laser dissection with molecular beacon polymerase chain reaction identifies 2A as the predominant serotonin receptor subtype in hypoglossal motoneurons. Neuroscience 2002;113:145–154.
Trulson ME, Jacobs BL. Raphe unit activity in freely moving cats: correlation with level of behavioral arousal. Brain Res 1979;163:135–150.
Heym J, Steinfels GF, Jacobs BL. Activity of serotonin-containing neurons in the nucleus raphe pallidus of freely moving cats. Brain Res 1982;251:259–276.
Ribeiro-do-Valle LE, Lucena RL. Behavioral correlates of the activity of serotonergic and non-serotonergic neurons in caudal raphe nuclei. Braz J Med Biol Res 2001;34:919–937.
Jacobs BL, Fornal CA. 5-HT and motor control: a hypothesis. Trends Neurosci 1993;16:346–352.
Henry JN, Manaker S. Colocalization of substance P or enkephalin in serotonergic neuronal afferents to the hypoglossal nucleus in the rat. J Comp Neurol 1998;391:491–505.
Bouryi VA, Lewis DI. The modulation by 5-HT of glutamatergic inputs from the raphe pallidus to rat hypoglossal motoneurones, in vitro. J Physiol 2003;553:1019–1031.
Bouryi VA, Lewis DI. Enkephalinergic inhibition of raphe pallidus inputs to rat hypoglossal motoneurones in vitro. Neuroscience 2004;129:55–64.
Nattie E. Multiple sites for central chemoreception: their roles in response sensitivity and in sleep and wakefulness. Respir Physiol 2000;122:223–235.
Johansson O, Hokfelt T, Pernow B, Jeffcoate SL, White N, Steinbusch HW, Verhofstad AA, Emson PC, Spindel E. Immunohistochemical support for three putative transmitters in one neuron: coexistence of 5-hydroxytryptamine, substance P- and thyrotropin releasing hormone-like immunoreactivity in medullary neurons projecting to the spinal cord. Neuroscience 1981;6:1857–1881.
Fogel RB, Malhotra A, Shea SA, Edwards JK, White DP. Reduced genioglossal activity with upper airway anesthesia in awake patients with OSA. J Appl Physiol 2000;88:1346–1354.
Bocchiaro CM, Feldman JL. Synaptic activity-independent persistent plasticity in endogenously active mammalian motoneurons. Proc Natl Acad Sci USA 2004;101:4292–4295.(Sandeep Sood, Janna L. Morrison, Hattie )