Opioid receptor mechanisms at the hypoglossal motor pool and effects on tongue muscle activity in vivo

Abstract

Opioids can modulate breathing and predispose to respiratory depression by actions at various central nervous system sites, but the mechanisms operating at respiratory motor nuclei have not been studied. This study tests the hypotheses that (i) local delivery of the μ-opioid receptor agonist fentanyl into the hypoglossal motor nucleus (HMN) will suppress genioglossus activity in vivo, (ii) a component of this suppression is mediated by opioid-induced acetylcholine release acting at muscarinic receptors, and (iii) δ- and κ-opioid receptors also modulate genioglossus activity. Seventy-two isoflurane-anaesthetised, tracheotomised, spontaneously breathing rats were studied during microdialysis perfusion into the HMN of (i) fentanyl and naloxone (μ-opioid receptor antagonist), (ii) fentanyl with and without co-application of muscarinic receptor antagonists, and (iii) δ- and κ-opioid receptor agonists and antagonists. The results showed (i) that fentanyl at the HMN caused a suppression of genioglossus activity (P < 0.001) that reversed with naloxone (P < 0.001), (ii) that neither atropine nor scopolamine affected the fentanyl-induced suppression of genioglossus activity, and (iii) that δ-, but not κ-, opioid receptor stimulation also suppressed genioglossus activity (P= 0.036 and P= 0.402 respectively). We conclude that μ-opioid receptor stimulation suppresses motor output from a central respiratory motoneuronal pool that activates genioglossus muscle, and this suppression does not involve muscarinic receptor-mediated inhibition. This μ-opioid receptor-induced suppression of tongue muscle activity by effects at the hypoglossal motor pool may underlie the clinical concern regarding adverse upper airway function with μ-opioid analgesics. The inhibitory effects of μ- and δ-opioid receptors at the HMN also indicate an influence of endogenous enkephalins and endorphins in respiratory motor control.

The analgesic properties of the juice extracted from the opium poppy have been recognized for centuries, and opioid drugs remain the mainstay of pain management in clinical medicine today (Gutstein & Akil, 2001; Dahan, 2007b). Morphine and other analgesic drugs derived from opium exert their effects by mimicking the action of endogenous opioid peptides, the natural ligands for opioid receptors. Most of the exogenously administered opioid drugs used clinically preferentially activate μ-opioid receptors (Gutstein & Akil, 2001; Dahan, 2007b). However, the most widely recognized adverse side effect of μ-opioid analgesics is respiratory depression, which can be lethal (Santiago & Edelman, 1985; McCrimmon & Alheid, 2003; Wang & Teichtahl, 2007; Pattinson, 2008), especially with overdose (Gutstein & Akil, 2001; Teichtahl & Wang, 2007). Medical concern for opioid-induced respiratory depression is such that provision of effective analgesia is often suboptimal (Pattinson, 2008).

Opioid-induced respiratory depression may be caused by effects at multiple central nervous system sites. These include effects at medullary respiratory neurons leading to suppression of respiratory rate and inhibition of premotor respiratory drive (Greer et al. 1995; Ballanyi et al. 1997; Gray et al. 1999; Takeda et al. 2001; Lalley, 2003; Manzke et al. 2003; McCrimmon & Alheid, 2003; Mellen et al. 2003), effects at sites of chemoreception leading to suppression of ventilatory responses to hypoxia and hypercapnia (Bailey et al. 2000; Zhang et al. 2007; Pattinson, 2008), suppression of brain arousal systems leading to sedation (Teichtahl & Wang, 2007; Wang & Teichtahl, 2007), and/or suppression of activity at central respiratory motoneurons innervating the respiratory muscles. Importantly for the latter possibility, despite the fact that respiratory motoneurons are the final common output pathway for the influence of the brain on the muscles of breathing (Sherrington, 1906; Burke, 2007), we know of no studies determining whether opioids suppress respiratory muscle activity by direct influences at central respiratory motor nuclei, and there is also no such information in major reviews of either motor function or opioids and breathing (Santiago & Edelman, 1985; Yeadon & Kitchen, 1989; Rekling et al. 2000; McCrimmon & Alheid, 2003; Dahan, 2007b; Pattinson, 2008).

Upper airway obstruction and hypoventilation are serious clinical concerns associated with opioid administration (Kryger, 2000; Brown et al. 2006; Lerman, 2006). The hypoglossal motor nucleus innervates the genioglossus (GG) and other muscles of the tongue, relaxation of which is a key factor in the pathogenesis of upper airway obstruction (Remmers et al. 1978). Clinically, adult patients with obstructive sleep apnoea (OSA) are at particular risk for respiratory complications post-operatively, i.e. at a time when opioids are routinely administered (Kryger, 2000). Children undergoing adenotonsillectomy for OSA have a higher sensitivity to μ-opioid receptor agonists, and higher post-operative difficulty in breathing and oxygen desaturation compared to children without OSA (Brown et al. 2006; Lerman, 2006). Caution and close respiratory monitoring is considered ‘mandatory’ in OSA patients post-operatively, and it is recommended that opioid drugs ‘only be administered if the airway is under control’ (Kryger, 2000). In practice, however, there is a large population of patients with undiagnosed OSA (Young et al. 2002), and many individuals with patient-controlled analgesia are at risk for impaired breathing post-operatively (Dahan, 2007a; Teichtahl & Wang, 2007; Wang & Teichtahl, 2007). Based on the above considerations, this study aims to characterize for the first time the effects on respiratory muscle activity of opioid receptor agonists and antagonists applied directly to a respiratory motor pool in vivo, to test the hypothesis that application of the μ-receptor agonist fentanyl to the hypoglossal motor nucleus will suppress GG muscle activity.

A recent study has also shown that application of morphine to the hypoglossal motor nucleus of isoflurane-anaesthetized rats leads to increased acetylcholine at that motor pool, a change that was postulated to have an inhibitory effect on motor activity (Skulsky et al. 2007). However, the physiological consequence of that opioid-induced acetylcholine release on GG activity was not investigated (Skulsky et al. 2007). This is particularly relevant as we have shown previously that increased endogenous acetylcholine at the hypoglossal motor nucleus suppresses GG activity via a muscarinic receptor mechanism (Liu et al. 2005). Accordingly, a second aim of the present study was to test the hypothesis that a component of the fentanyl-induced suppression of GG activity is reduced by muscarinic receptor antagonism at the hypoglossal motor nucleus. Finally, given that this study is the first to characterize the effects on respiratory muscle activity of opioid receptor stimulation at a respiratory motor pool, further studies were performed to determine the effects on GG muscle activity of δ- and κ-opioid receptor stimulation at the hypoglossal motor pool.

Methods

Studies were performed on a total of 72 adult male Wistar rats (mean body weight = 270 ± 2 g (s.e.m.), range = 230–330 g, Charles River). The number of animals studied for each experiment is listed where appropriate. All procedures conformed to the recommendations of the Canadian Council on Animal Care, and the University of Toronto Animal Care Committee approved the experimental protocol.

Surgical preparation

All experiments were performed in isoflurane-anaesthetized rats. Following induction of surgical levels of anaesthesia, as judged by the abolition of the hindlimb withdrawal and corneal blink reflexes, the rats were tracheotomized and the femoral artery and vein were cannulated. Throughout the surgery, and the experiments, the rats spontaneously breathed a 50: 50 mixture of room air and oxygen. In initial experiments in 33 rats (Study 1, Protocols 1a–d, see below), anaesthesia was maintained with isoflurane (typically 0.2–0.9%) after initial initiation with urethane (1 g kg⁻¹) via intraperitoneal injection. We have routinely used this anaesthetic regime in previous studies because it provides highly reliable preparations (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008), compared to, for example, periodic bolus doses of supplemental urethane, which can transiently suppress GG activity and alter the electroencephalogram (EEG). Once isoflurane was initiated within an animal, typically no further adjustment was necessary across the experiment to maintain stable EEG activity. Additional experiments (Studies 2–4) were performed in 39 rats that were anaesthetized with isoflurane only (2–3.2%). These experiments were performed with isoflurane because of a recent study showing that μ-opioid receptor stimulation at the hypoglossal motor nucleus can lead to increased levels of acetylcholine at that motor pool in isoflurane-anaesthetized rats (Skulsky et al. 2007), and we wished to study the physiological consequence of that effect on GG activity in a similar preparation, i.e. without the potential influence of urethane.

In all experiments, core body temperature was maintained between 36 and 38°C with a water pump and heating pad (T/Pump-Heat Therapy System, Gaymar, NY, USA). The rats received continuous intravenous fluid (0.4 ml h⁻¹) containing 7.6 ml saline, 2 ml 5% dextrose and 0.4 ml of 1 m NaHCO₃. For electromyogram (EMG) recordings of diaphragm activity, two insulated, multi-stranded stainless steel wires (AS636; Coorner Wire, Chatsworth, CA, USA) were sutured into the costal diaphragm via an abdominal approach. The rats were placed in a stereotaxic apparatus (Kopf Model 962, Tujunga, CA, USA) and the head was secured with blunt ear bars. To ensure consistent positioning between rats, the flat skull position was achieved with an alignment tool (Kopf Model 944). To record the cortical EEG two stainless steel screws attached to insulated wire were implanted onto the skull over the frontal-parietal cortex (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008). Two stainless steel wire electrodes were also inserted bilaterally under direct observation into the body of the tongue, via an oral approach, to record GG activity. We have shown that such electrode placements record activity from muscles innervated by the medial branch of the hypoglossal nerve (Morrison et al. 2002).

Microdialysis

A microdialysis probe (CMA/11 14/01, CSC, St Laurent, QC, Canada) was placed through a small hole drilled at the junction of the interparietal and occipital bones. The probes were lowered into the hypoglossal motor nucleus at the following final coordinates: 13.6 ± 0.05 mm posterior to bregma, 0.1 ± 0.02 mm lateral to the midline and 9.9 ± 0.09 mm ventral to bregma. In each rat, a brief burst of GG activity was observed when the probe initially penetrated the hypoglossal motor nucleus, and then the probe was advanced another 0.5 mm before being left at the final coordinates. This burst of GG activity during probe insertion was transient (typically < 5 min) and did not affect diaphragm activity, blood pressure or respiratory rate, and was useful as a preliminary indication of probe placement (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008). The rats stabilized for an average of 30 ± 5 min before any interventions.

The microdialysis probes were 240 μm in diameter with a 1 mm cuprophane membrane and a 6000 Da cut-off. The probes were connected to FEP Teflon tubing (inside diameter = 0.12 mm) and connected to 1.0 ml syringes via a zero dead space switch (Uniswitch, B.A.S. West Lafayette, IN, USA). The probes were continually flushed with artificial cerebrospinal fluid (ACSF) at a flow rate of 2.0 μl min⁻¹ using a syringe infusion pump (Pump 22; Harvard Apparatus; Holliston, MA, USA). The lag time for fluid to travel to the tip of the probe at this flow rate was approximately 5 min. The composition (in mm) of the ACSF was NaCl (125), KCl (3), KH₂PO₄ (1), CaCl₂ (2), MgSO₄ (1), NaHCO₃ (25) and glucose (30). The ACSF was made fresh each day, warmed to 37°C and bubbled with CO₂ to a pH of 7.42 ± 0.003.

Recording

The electrical signals were amplified and filtered (Super-Z head-stage amplifiers and BMA-400 amplifiers/filters, CWE Inc., Ardmore, PA, USA). The EEG was amplified and filtered between 1 and 100 Hz, whereas the EMG signals were amplified and filtered between 100 and 1000 Hz. The GG and diaphragm signals were recorded at the same amplification across all experiments. It was not necessary to alter the gain of the recording apparatus between experiments and baseline EMG activity was similar across rats. The electrocardiogram was removed from the diaphragm signal using an oscilloscope and electronic blanker (Model SB-1, CWE Inc.). In addition, the moving time averages (time constant = 100 ms) of the GG and diaphragm signals were obtained (Coulbourn S76-01, Lehigh Valley, PA, USA). Each signal, along with blood pressure (DT-XX transducer, Ohmeda, Madison, WI, USA and PM-1000 Amplifier, CWE Inc.) were digitized and recorded on computer (Spike 2 software, 1401 interface, CED Ltd., Cambridge, UK).

Protocol

Multiple studies were performed in separate groups of rats to determine the mechanisms underlying suppression of GG activity by opioids applied to the hypoglossal motor nucleus. Drugs were purchased from Sigma (St. Louis, MO, USA).

Study 1: Modulation of GG activity by μ-opioid receptor mechanisms at the hypoglossal motor pool

Signals were recorded continuously during microdialysis perfusion of ACSF into the hypoglossal motor nucleus (control condition), and then during perfusion of the μ-opioid receptor agonist fentanyl at 1, 10 and 100 μm. For these experiments, following application of the last dose of fentanyl the perfusion medium was then switched to one of (a) the μ-opioid receptor antagonist naloxone (100 μm, n= 9 rats, Protocol 1a); (b) the muscarinic receptor antagonist atropine (10 μm, n= 8 rats, Protocol 1b), with previous studies (using the same techniques) showing that this dose of atropine significantly reduces the suppression of GG activity produced by increases in endogenous acetylcholine at the hypoglossal motor nucleus (Liu et al. 2005); or (c) back to ACSF, i.e. washout (n= 8 rats, Protocol 1c). In each of these protocols, each intervention lasted 30 min. Importantly, to next examine if any of the observed changes in GG activity were caused by effects of time per se, i.e. independent of the drug interventions, further experiments were performed in which repeated switches to perfusion of ACSF into the hypoglossal motor nucleus were performed (i.e. ‘sham interventions’) over the same time course as the drug interventions (n= 8 rats, Protocol 1d).

Study 2: Continuous μ-opioid receptor stimulation at the hypoglossal motor pool with and without muscarinic receptor antagonism

An additional study was performed to further characterize μ-opioid induced suppression of GG muscle activity and the potential role of fentanyl-induced acetylcholine release at the hypoglossal motor nucleus in contributing to the GG suppression. Given that the experiments performed in Study 1 were to characterize GG activity in response to increasing doses of fentanyl, further experiments were performed in which ACSF was applied to the hypoglossal motor nucleus for 60 min, followed by a switch to continuous delivery of one dose of 100 μm fentanyl for 90 min (in three 30 min periods). In these experiments, fentanyl was delivered to the hypoglossal motor nucleus either alone (n= 6 rats, Protocol 2a) or in the combined presence of 10 μm atropine (n= 6 rats, Protocol 2b). Again, previous studies using the same techniques have shown that this dose of atropine significantly reduces the suppression of GG activity by increases in endogenous acetylcholine at the hypoglossal motor nucleus (Liu et al. 2005). In a third set of experiments, 100 μm fentanyl was also continuously applied to the hypoglossal motor nucleus in the presence of a different muscarinic receptor antagonist, (–)-scopolamine methyl bromide (100 nm, n= 4 rats, Protocol 2c). This dose of scopolamine delivered by microdialysis perfusion into the spinal cord significantly inhibits the release of acetylcholine in isoflurane-anaesthetized rats (Hoglund et al. 2000). Finally, to again examine if any of the observed changes in GG activity were caused by effects of time per se, i.e. independent of the drug interventions, further experiments were performed in which repeated switches to perfusion of ACSF into the hypoglossal motor nucleus were performed (i.e. ‘sham interventions’) over the same time course as the drug interventions (n= 8 rats, Protocol 2d).

Study 3: δ-opioid receptor stimulation at the hypoglossal motor pool

δ-Receptors have also been localized to the hypoglossal motor pool, and as such are also positioned to modulate GG activity (Sales et al. 1985; Richardson & Gatti, 2005). Accordingly, signals were monitored continuously during microdialysis perfusion of ACSF into the hypoglossal motor nucleus (control), and then during perfusion of the δ-opioid receptor agonist [d-Pen_2,5]-enkephalin hydrate (DPDPE) at 1, 10 and 100 μm (n= 9 rats, Protocol 3). Following application of the last dose of DPDPE, the perfusion medium was then switched to the δ-opioid receptor antagonist naltrindole (100 μm). Any changes in GG muscle activity were compared to the equivalent time control experiments performed in Protocol 2d (see above).

Study 4: κ-opioid receptor stimulation at the hypoglossal motor pool

In a final set of experiments, signals were monitored during microdialysis perfusion of ACSF into the hypoglossal motor nucleus (control), and then during perfusion of the κ-opioid receptor agonist U-50488 at 1, 10 and 100 μm (n= 6 rats, Protocol 4). Following application of the last dose of U-50488, the perfusion medium was switched to the κ-opioid receptor antagonist nor-binaltorphimine dihydrochloride (nor-BNI) (100 μm). Any changes in GG muscle activity were compared to the equivalent time-control experiments performed in Protocol 2d (see above).

Analyses

For each agent delivered to the hypoglossal motor nucleus, measurements were taken over 1 min periods at the end of each 30 min drug or sham intervention (i.e. Studies 1, 3 and 4), and at the end of each successive 30 min time period for the more prolonged interventions (Study 2). Breath-by-breath measurements of GG and diaphragm activities were calculated and averaged in consecutive 5 s time bins (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008). All values were written to a spreadsheet and matched to the corresponding intervention at the hypoglossal motor nucleus to provide a grand mean for each variable, for each intervention, in each rat. The EMG signals were analysed from the moving time average signals (above electrical zero) and quantified in arbitrary units and expressed as a percentage (%) of the respective ACSF controls. Electrical zero was the voltage recorded with the amplifier inputs grounded. GG activity was quantified as mean tonic activity (i.e. basal activity during expiration), peak activity and respiratory-related activity (i.e. peak inspiratory minus tonic activity). In practice there was no tonic GG activity in these experiments performed under anaesthesia, so only respiratory-related GG activity is presented. Mean diaphragm amplitudes (i.e. respiratory-related diaphragm activity), respiratory rate and mean arterial blood pressure were also calculated for each 5 s period. The EEG was sampled by computer at 500 Hz and subjected to a fast-Fourier transform for each 5 s time bin, and the power within frequency bands spanning the 0.5–30 Hz range was calculated (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008). The ratio of high (β₂, 20–30 Hz) to low (δ₁, 2–4 Hz) frequency activity was calculated and used as a relative marker of EEG activation (Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008).

Tests of function of hypoglossal motor nucleus and histology

At the end of each experiment for each protocol, 10 mm serotonin (creatinine sulphate complex) was applied to the hypoglossal motor nucleus as a positive control to confirm that it was still functional and able to respond to manipulation of neurotransmission as judged by the expected increase in GG activity (Jelev et al. 2001; Liu et al. 2005; Sood et al. 2005; Steenland et al. 2008). At the end of each study the rats were killed under anaesthesia by intravenous injection of 3–5 ml of saturated KCl and a high dose of anaesthetic. The rats then were perfused intracardially with 0.9% saline and 10% formalin, following which the brain was removed and fixed in 10% formalin. Medullary regions containing the hypoglossal motor nucleus were blocked and transferred to a 30% sucrose solution for cryoprotection. The tissue was then sectioned at 50 μm using a cryostat (CM1850, Leica, Nussloch, Germany). Sections were mounted and stained with neutral red, and the lesion sites left by the microdialysis probes were recorded on a corresponding standard cross-section using a stereotaxic atlas of the rat brain (Paxinos & Watson, 1998).

Statistics

The analyses performed for each statistical test are included in the text where appropriate. For all comparisons, differences were considered significant if the null hypothesis was rejected at P < 0.05. Where post hoc comparisons were performed after analysis of variance with repeated measures (ANOVA-RM), the Bonferroni-corrected P value was used to infer statistical significance. Analyses were performed using SigmaStat (Systat Software Inc., San Jose, CA, USA). Data are presented as means ±s.e.m. unless otherwise indicated.

Results

Sites of microdialysis

Figure 1 shows an example of a lesion site left by a microdialysis probe in the hypoglossal motor nucleus. The distribution of microdialysis sites from all the experiments in Study 1 are also shown in Fig. 1, with the sites located within or immediately adjacent to the hypoglossal motor nuclei in all animals.

Figure 1. Example and group data showing the location of the microdialysis probes from all the experiments in Protocols 1a–d from Study 1: modulation of genioglossus activity by μ-opioid receptor mechanisms at the hypoglossal motor pool.

The top images show histological sections with the location of microdialysis site indicated by the arrow, with adjacent regions of the hypoglossal motor pool (rostral and caudal) also shown. Abbreviation: HMN, hypoglossal motor nucleus.

Study 1: Modulation of GG activity by μ-opioid receptor mechanisms at the hypoglossal motor pool

Opioid-induced suppression of GG activity

Figure 2A shows an example of the progressive decrease in GG activity with increasing concentration of the μ-opioid receptor agonist fentanyl applied locally to the hypoglossal motor pool. The group data in Fig. 2B show that there was a statistically significant effect of fentanyl on GG activity (P < 0.001, 2-way ANOVA-RM). For each of Protocols 1a–c, significant GG suppression first occurred at 10 μm fentanyl compared to the preceding baseline ACSF controls (P < 0.001, post hoc paired t-tests, asterisks in Fig. 2B). Nevertheless, for each of Protocols 1a–c, it was at 100 μm fentanyl that the levels of GG activity were all significantly reduced compared to the same point in the corresponding ACSF time-control experiments (Fig. 2B, each P < 0.001 post hoc paired t-tests, symbol ‘+’ in Fig. 2B). Finally, although there was a decline in GG activity in the time-control experiments with continuous application of ACSF to the hypoglossal motor nucleus (Fig. 2B), this only became statistically significant at the last intervention (P= 0.015, post hoc paired t-test, asterisks for Protocol 1d in Fig. 2B, P > 0.178 for each of the other comparisons to baseline ACSF in this protocol).

Figure 2. Modulation of GG activity by μ-opioid receptor mechanisms at the hypoglossal motor pool.

A, example in one rat showing selective depression of respiratory-related genioglossus activity with microdialysis perfusion of the μ-opioid receptor agonist fentanyl into the hypoglossal motor pool, and the subsequent reversal of this suppression with application of the μ-opioid receptor antagonist naloxone. B, group data showing significant suppression of respiratory-related genioglossus activity with fentanyl at the hypoglossal motor pool in Protocols 1a–c. Subsequent application of naloxone led to a significant increase in genioglossus activity (Protocol 1a) that did not occur with a switch to either atropine or artificial cerebrospinal fluid (ACSF) (Protocols 1b and 1c respectively). See text for further details. *P < 0.05 compared to baseline ACSF controls; +P < 0.05 compared to the corresponding ACSF time-control experiment (i.e. Protocol 1d); #P < 0.05 from fentanyl to naloxone. Abbreviations: EEG, electroencephalogram; EMG, electromyogram; MTA, moving-time average.

Effects of applied antagonists

The effect on GG activity of a switch from 100 μm fentanyl at the hypoglossal motor nucleus to a particular antagonist depended on the experimental protocol (P < 0.001, 2-way ANOVA-RM). A switch from fentanyl to naloxone at the hypoglossal motor nucleus led to a significant increase in GG muscle activity (P < 0.001, post hoc paired t-test, symbol ‘#’ in Fig. 2B) to levels that were still reduced, but not significantly different, from the levels of GG activity recorded in the time-control experiments with ACSF (label ‘ns’ in Fig. 2B). This result confirmed that the suppression of GG activity with fentanyl at the hypoglossal motor nucleus was indeed mediated by a μ-opioid receptor mechanism per se and not an effect of time. In contrast to this result with naloxone, a switch from fentanyl to atropine at the hypoglossal motor nucleus in Protocol 1b had no effect on GG activity (P= 0.252, post hoc paired t-test, Fig. 2B), such that GG activity remained significantly lower than the levels recorded in the time-control experiments with ACSF (symbol ‘+’ in Fig. 2B). This lack of effect with atropine, despite being an effective dose to block inhibitory cholinergic effects at the hypoglossal motor nucleus (Liu et al. 2005), showed that any potential effect of increased acetylcholine at the hypoglossal motor pool (Skulsky et al. 2007) did not play a significant role in the fentanyl-induced suppression of GG activity. Finally, a switch back to ACSF after fentanyl in Protocol 1c also had no effect on GG activity (P= 0.342, post hoc paired t-test, Fig. 2B), such that again GG activity remained significantly lower than the levels recorded in the ACSF time-control experiments (symbol ‘+’ in Fig. 2B). This result also showed that fentanyl had persistent effects on GG activity despite its washout, and that the increase previously observed with naloxone was not due to washout of fentanyl per se but was due to the antagonist.

Specificity of responses to fentanyl

In these anaesthetized rats baseline respiratory rate averaged 73.1 ± 1.6 min⁻¹, blood pressure averaged 72.4 ± 3.4 mmHg, and respiratory-related GG and diaphragm muscle activities averaged 177.3 ± 19.7 and 385.1 ± 40.4 arbitrary units respectively. Unlike the clear suppression of GG activity with fentanyl applied locally to the hypoglossal motor nucleus in Protocols 1a–c (Fig. 2), there were no significant effects of these fentanyl applications on the amplitude of diaphragm activity, respiratory rate or blood pressure compared to the changes over time observed in the ACSF time-control experiments (Fig. 3, P= 0.092, 0.065 and 0.376, respectively, 2-way ANOVA-RM), i.e. the effects on GG activity were specific to the interventions at the hypoglossal motor pool and did not significantly affect other variables. Likewise, there was no effect of the fentanyl applications on the ratio of high (β₂, 20–30 Hz) to low (δ₁, 2–4 Hz) frequency activity in the EEG, indicating no relative change in EEG activation compared to the ACSF time-control experiments (P= 0.451, 2-way ANOVA-RM).

Figure 3. Responses to the μ-opioid receptor agonist fentanyl at the hypoglossal motor pool were specific to the genioglossus muscle as there were no significant effects of these fentanyl applications on the amplitude of diaphragm activity, respiratory rate or blood pressure in Protocols 1a–c compared to the changes over time observed in the ACSF time-control experiments (Protocol 1d).

See text for further details.

Study 2: Continuous μ-opioid receptor stimulation at the hypoglossal motor pool with and without muscarinic receptor antagonism

The distribution of microdialysis sites from all the experiments in Study 2 are shown in Fig. 4, which illustrates that all sites were located within or immediately adjacent to the hypoglossal motor nuclei in all animals. The group data in Fig. 5 show that there was a statistically significant effect of experimental protocol on GG activity (P= 0.005, 2-way ANOVA-RM), with the responses to the μ receptor agonist fentanyl (Protocols 2a–c) being significantly different from the ACSF time-control experiment (Protocol 2d, all P < 0.032, post hoc paired t-tests). For each of Protocols 2a–c, there was significant suppression of GG activity compared to the preceding baseline ACSF controls (all P < 0.001, post hoc paired t-tests, asterisks in Fig. 5). These decreases in GG activity with fentanyl at the hypoglossal motor nucleus were due to the effects of fentanyl per se and not due to the effects of time, as shown by the significant difference in GG activity with fentanyl compared to the same point in the corresponding ACSF time-control experiments (all P < 0.041, post hoc paired t-tests, symbols ‘+’ in Fig. 5). Importantly, however, there was no difference in the magnitude of GG suppression with fentanyl alone at the hypoglossal motor nucleus (Protocol 2a) compared to when fentanyl was co-applied with either atropine or scopolamine (Protocols 2b and 2c, P= 1.00, post hoc paired t-tests, Fig. 5). This lack of effect of atropine and scopolamine showed that any potential effect of fentanyl on acetylcholine at the hypoglossal motor pool (Skulsky et al. 2007) did not play a significant role in the fentanyl-induced suppression of GG activity via a cholinergic-induced muscarinic receptor-mediated inhibition.

Figure 4. Location of the microdialysis probes from all the experiments in Protocols 2a–d from Study 2: continuous μ-opioid receptor stimulation at the hypoglossal motor pool with and without muscarinic receptor antagonism.

Abbreviations are as for Fig. 1.

Figure 5. Group data showing significant suppression of respiratory-related genioglossus activity with microdialysis perfusion of the μ-opioid receptor agonist fentanyl into the hypoglossal motor pool in Protocols 2a–c, and the lack of effect of co-applied atropine or scopolamine on these responses.

See text for further details. *P < 0.05 compared to baseline ACSF controls; +P < 0.05 compared to the corresponding ACSF time-control experiment (i.e. Protocol 2d).

Finally, although there was a decline in GG activity in the time-control experiments with continuous application of ACSF to the hypoglossal motor nucleus, this only became statistically significant at the last intervention (P= 0.008, post hoc paired t-test, asterisks for Protocol 2d in Fig. 5). Moreover, at this time in the ACSF time-control experiments, GG activity was still significantly increased compared to the same time point in the fentanyl experiments (i.e. compared to Protocols 2a–c, each P < 0.016, post hoc paired t-tests, symbols ‘+’ in Fig. 5).

Specificity of responses

In these anaesthetized rats, baseline respiratory rate averaged 50.2 ± 1.9 min⁻¹, blood pressure averaged 70.4 ± 3.2 mmHg, and respiratory-related GG and diaphragm muscle activities averaged 200.7 ± 45.7 and 331.3 ± 63.9 arbitrary units, respectively. Unlike the clear suppression of GG activity with fentanyl applied locally to the hypoglossal motor nucleus in Protocols 2a–c (Fig. 5), there were no significant effects of experimental protocol on the amplitude of diaphragm activity, respiratory rate or blood pressure during the course of the experiments (P= 0.252, 0.069 and 0.411 respectively, 2-way ANOVA-RM), i.e. the effects on GG activity observed in Fig. 5 were specific to the interventions at the hypoglossal motor pool and did not significantly affect other variables. Likewise, there was no effect of the fentanyl applications on the ratio of high (β₂, 20–30 Hz) to low (δ₁, 2–4 Hz) frequency activity in the EEG, indicating no relative change in EEG activation compared to the ACSF time-control group (P= 0.617, 2-way ANOVA-RM).

Study 3: Responses to δ-opioid receptor stimulation at the hypoglossal motor pool

The distribution of microdialysis sites from Protocol 3 are shown in Fig. 6A. The group data in Fig. 6B show there was a significant effect of the experimental protocol on GG activity (P= 0.043, 2-way ANOVA-RM), with GG activity in the presence of DPDPE being significantly reduced compared to the ACSF time-control interventions (P= 0.043, post hoc paired t-test). Significant GG suppression occurred at each dose of DPDPE compared to the preceding baseline ACSF controls (all P < 0.013, post hoc paired t-tests, asterisks in Fig. 6B). These decreases in GG activity with DPDPE at the hypoglossal motor pool were due to the effects of the δ-opioid receptor agonist per se and not the effects of time, as shown by the significant difference in DPDPE response compared to the ACSF time controls (P= 0.036 2-way ANOVA-RM). Further analyses showed that the levels of GG activity with 10 and 100 μm DPDPE were significantly reduced compared to the same point in the corresponding ACSF time-control experiments (Fig. 6B, each P < 0.013 post hoc paired t-tests, symbols ‘+’ in Fig. 6B).

Figure 6. Responses to δ-opioid receptor stimulation at the hypoglossal motor pool.

A, location of the microdialysis probes from all the experiments in Protocol 3 from Study 3: δ-opioid receptor stimulation at the hypoglossal motor pool. B, group data illustrating significant suppression of respiratory-related genioglossus activity with microdialysis perfusion of the δ-opioid receptor agonist DPDPE into the hypoglossal motor pool. *P < 0.05 compared to baseline ACSF controls; +P < 0.05 compared to the corresponding ACSF time-control experiment. See text for further details.

Effects of the antagonist

GG activity did not significantly increase following a switch from 100 μm DPDPE at the hypoglossal motor nucleus to the antagonist naltrindole, such that any observed change in GG activity was statistically indistinguishable from the change observed in the ACSF time-control experiment (P= 0.743, 2-way ANOVA-RM, see Fig. 6B). The relatively large error bar for GG activity in the presence of naltrindole in Fig. 6B is because GG activity changed inconsistently following the switch from DPDPE (range of −27.2% to +173%), with increases occurring in only three of nine rats.

Specificity of responses

Baseline respiratory rate averaged 54.6 ± 2.3 min⁻¹, blood pressure averaged 68.9 ± 2.9 mmHg, and respiratory-related GG and diaphragm muscle activities averaged 95.7 ± 18.7 and 576.7 ± 110.5 arbitrary units respectively. Unlike the suppression of GG activity with DPDPE applied to the hypoglossal motor pool (Fig. 6B), there were no effects of experimental protocol on the amplitude of diaphragm activity, respiratory rate or blood pressure during the course of the experiments (P= 0.910, 0.059 and 0.587 respectively, 2-way ANOVA-RM), i.e. the effects on GG activity observed in Fig. 6B were specific to the interventions at the hypoglossal motor pool and did not significantly affect other variables. Likewise, there was no effect of the DPDPE applications on the ratio of high (β₂, 20–30 Hz) to low (δ₁, 2–4 Hz) frequency activity in the EEG, indicating no relative change in EEG activation compared to the ACSF time-control experiments (P= 0.872, 2-way ANOVA-RM).

Study 4: Responses to κ-opioid receptor stimulation at the hypoglossal motor pool

The distribution of microdialysis sites from Protocol 4 are shown in Fig. 7A, and the data for effects on GG activity are shown in Fig. 7B. Statistical analyses showed that there was a decline in GG activity over the course of the experiment (P < 0.001, 2-way ANOVA-RM), which first became significant at a time point corresponding to 10 μm U-50488 for Protocol 4 and the same for the time control (P= 0.016, post hoc paired t-test, asterisk in Fig. 7B). Importantly, however, there was no effect of experimental protocol per se on GG activity (P= 0.402). This latter result indicated that any effect on GG activity observed in Protocol 4 with κ-receptor stimulation at the hypoglossal motor pool was statistically indistinguishable from the corresponding sham experiment with ACSF, i.e. the time control. Moreover, there was no significant effect on GG activity following a switch from 100 μm of the κ-receptor agonist U-50488 to the antagonist nor-BNI (P= 0.123, 2-way ANOVA-RM), and again no difference between experimental protocols (P= 0.182, 2-way ANOVA-RM). Overall, these data indicate that compared to the ACSF time-control group, there was no significant difference in effect of κ-receptor agonists or antagonists at the hypoglossal motor pool.

Figure 7. Responses to κ-opioid receptor stimulation at the hypoglossal motor pool.

A, location of the microdialysis probes from all the experiments in Protocol 4 from Study 4: κ-opioid receptor stimulation at the hypoglossal motor pool. B, group data illustrating that there was a decline in genioglossus activity over the course of the experiment with microdialysis perfusion of the κ-opioid receptor agonist U-50488 into the hypoglossal motor pool, which first became statistically significant (P < 0.05) at 10 μm U-50488 (indicated by *). However, there was no significant difference in the effect of κ-receptor agonists or antagonists at the hypoglossal motor pool compared to the ACSF time controls. See text for further details.

Nevertheless, because of the apparent visual impression of a larger suppression of GG activity with U-50488 compared to the ACSF time-control experiments in Fig. 7B (especially at 100 μm), we further inspected the data and observed that three animals showed a decline of GG activity to less than 50% of the initial baseline value with 100 μm U-50488 (range, 41.5 to 44.4%) whereas the others showed a lesser decline (range, 60.6 to 94.3%) that was in the range of the ACSF controls at the same time point (mean = 78.1 ± 8.2%). Importantly, none of those three animals that showed the larger decline in GG activity with U-50488 responded to subsequent application of the antagonist, providing further evidence for a lack of a specific inhibitory effect of the κ-receptor agonist at the hypoglossal motor pool.

Specificity of responses

Baseline respiratory rate averaged 60.9 ± 2.7 min⁻¹, blood pressure averaged 71.8 ± 3.6 mmHg, and respiratory-related GG and diaphragm muscle activities averaged 69.3 ± 11.1 and 404.3 ± 81.9 arbitrary units respectively. There were no effects of experimental intervention (i.e. a switch from ACSF to the κ-opioid receptor drugs) or protocol (i.e. Protocol 4 vs. the ACSF time-control experiments) on the amplitude of diaphragm activity, blood pressure, or the ratio of high (β₂, 20–30 Hz) to low (δ₁, 2–4 Hz) frequency activity in the EEG (all P > 0.077, 2-way ANOVA-RM). However, there was a significant difference in the change in respiratory rate response between the ACSF time-control group and the κ-receptor agonist applied to the hypoglossal motor pool (P= 0.009, 2-way ANOVA-RM), with respiratory rate decreasing less with U-50488 at all time points compared to the ACSF time controls (all P < 0.012, post hoc paired t-test).

Discussion

Major findings and their interpretation

This paper contributes to the field of physiology by being the first to report a suppression of respiratory muscle activity caused by the presence of opioids at a distinct central nervous system motor pool, and the first characterization of the receptor mechanisms underlying this suppression. The results indicate that the presence of the μ-opioid receptor agonist fentanyl at the hypoglossal motor pool causes the suppression of tongue respiratory muscle activity, with this suppression reversed by application of the μ-opioid receptor antagonist naloxone. Further experiments involving different protocols for delivery of fentanyl to the hypoglossal motor pool (i.e. incrementing doses or continuous steady-state infusion), with and without co-application of different muscarinic receptor antagonists (i.e. atropine or scopolamine), also indicated that a potential μ-opioid-induced release of acetylcholine at the hypoglossal motor nucleus (Skulsky et al. 2007) did not contribute to the observed suppression of GG activity via cholinergic-induced muscarinic receptor-mediated inhibition (Liu et al. 2005). With respect to the possible involvement of acetylcholine, we performed the additional experiments with muscarinic receptor antagonism because a previous study using the same anaesthetized rat preparation had shown that microdialysis perfusion of the μ-opioid receptor agonist morphine into the hypoglossal motor pool led to an increase in acetylcholine levels by ∼25%, as measured by high performance liquid chromatography in the recovered perfusate, a level of change that was postulated to have an inhibitory effect on motor activity (Skulsky et al. 2007). However, although this percentage increase may at first seem large, the levels of acetylcholine were normalized to basal levels before morphine, i.e. a time when absolute levels of acetylcholine would be expected to be low in the anaesthetized preparation. Consequently, the physiological consequence of that increase in acetylcholine on tongue muscle activity would be uncertain, and was not measured in that study (Skulsky et al. 2007).

This discussion of a potential μ-opioid receptor-induced release of acetylcholine at the hypoglossal motor pool is relevant because we have previously identified that increases in endogenous acetylcholine at the hypoglossal motor nucleus cause robust suppression of GG activity, with this suppression being significantly attenuated by muscarinic receptor antagonism (Liu et al. 2005). Such an effect of increased endogenous acetylcholine is also of potential relevance to upper airway function clinically (Eikermann et al. 2007, 2008). Of importance with respect to the interpretation of results and validity of the conclusions from the present study, it is notable that we used a dose of atropine delivered to the hypoglossal motor pool that we have shown previously to be effective in significantly attenuating the inhibitory effects on GG activity of increased endogenous acetylcholine (Liu et al. 2005). However, in the present study using the same techniques, there was clearly no effect of the applied atropine on the magnitude of the fentanyl-induced suppression of GG activity, with similar results observed with scopolamine (Figs 2B and 5). This dose of scopolamine delivered by microdialysis perfusion is also effective in inhibiting the release of acetylcholine in the spinal cord of isoflurane-anaesthetized rats (Hoglund et al. 2000). Overall, these considerations do not dispute that an increase in endogenous acetylcholine can occur at the hypoglossal motor pool in the presence of μ-opioid receptor stimulation, and we consider it highly unlikely that such an effect would be specific only to morphine (Skulsky et al. 2007). Rather, we conclude that any increase in endogenous acetylcholine that may have occurred with fentanyl at the hypoglossal motor pool was simply insufficient to significantly modulate GG activity.

Finally, in addition to this characterization of a μ-opioid receptor-induced suppression of hypoglossal motor activity and the lack of influence of associated muscarinic receptor-mediated inhibition, we also characterized the effects on GG activity of δ- and κ-opioid-receptor stimulation at the hypoglossal motor pool. These data identified a significant suppression of GG muscle activity with δ-opioid receptor stimulation at the hypoglossal motor pool.

Physiological and clinical relevance

The results of the present study identify a mechanism whereby the presence of opioids in the mammalian brainstem could directly predispose to hypoventilation and respiratory depression by an inhibitory effect at central respiratory motor pools. In the case of the hypoglossal motor pool, this inhibition would exert an effect on the tongue musculature, and in humans so contribute to hypoventilation via an increase in upper airway resistance, which is a principal contributor to hypoventilation (Henke et al. 1990, 1992) and upper airway obstruction (Remmers et al. 1978). However, if this opioid-induced suppression of motor activity is typical of other respiratory motor pools, it may contribute to a more generalized respiratory depression via reduced tidal volume and functional residual capacity. Such an effect would therefore contribute to respiratory depression independently of, and likely in addition to, the more commonly discussed mechanisms of inhibition of respiratory proprio-bulbar and pre-motor neurons, so leading to suppression of respiratory rate and pre-motor respiratory drive (Greer et al. 1995; Ballanyi et al. 1997; Gray et al. 1999; Takeda et al. 2001; Lalley, 2003; Manzke et al. 2003; McCrimmon & Alheid, 2003; Mellen et al. 2003), suppression of chemoreflex responses (Bailey et al. 2000; Zhang et al. 2007; Pattinson, 2008), and/or sedation (Teichtahl & Wang, 2007; Wang & Teichtahl, 2007). Indeed, the observed suppression of tongue muscle activity by fentanyl at the hypoglossal motor pool lends credence to the clinical caution regarding the upper airway in adult and pediatric OSA patients peri- and post-operatively (Kryger, 2000; Brown et al. 2006; Lerman, 2006), and may provide the neurophysiological basis for this clinical concern.

Finally, the suppressant effect on GG activity of μ- and δ-opioid receptor agonists at the hypoglossal motor pool may also have physiological relevance to endogenous control of respiratory motor activity, and a functional role of these receptors in the control of tongue muscle activity. For example, nerve terminals containing enkephalins, the endogenous ligands for opioid receptors, are present in the hypoglossal motor nucleus (Connaughton et al. 1986; Aldes, 1998; Richardson & Gatti, 2005). Medullary enkephalinergic pathways have been hypothesized to function in the normal modulation of oro-facial motor tone (Aldes, 1998), even in a state-dependent fashion (Fort et al. 1998). However, these conjectures remain to be formally tested.

Critique of preparation and importance of adequate controls

The present study was performed in intact anaesthetized rats to allow measurements of GG muscle activity during controlled delivery of a variety of opioid receptor agonists and antagonists into the hypoglossal motor pool in different protocols, without the influence of unpredictable changes in ongoing behavioural state and spontaneous motor behaviours that typify the conscious preparation (Steenland et al. 2008). The levels of respiratory-related GG activity are also consistent, robust and of larger magnitude with isoflurane anaesthesia than without anaesthesia (Roda et al. 2004; Sood et al. 2006; Steenland et al. 2008), so making any motor suppression more easy to detect and quantify (e.g. as compared to quantifying suppression of a lower amplitude signal), more easy to compare across different protocols during stable conditions, and to characterize better the potential μ-, δ- and κ-opioid receptor influences. Importantly, effort was also made for each experimental protocol to determine if the decreases in GG activity that occurred in the presence of the different opioid receptor agonists at the hypoglossal motor pool were due to the effects of the drugs per se or the effects of time. This was achieved by comparing the responses to the respective agonists with appropriate ACSF time controls, and it is from such comparisons that evidence for a μ- and δ-opioid receptor-mediated suppression of hypoglossal motor output to GG muscle was obtained.

The main focus of this paper was identification of μ-opioid receptor-induced suppression of hypoglossal motor activity and the potential role of cholinergic-induced muscarinic receptor-mediated inhibition in that response. However, this paper also provides some evidence for δ-opioid receptor-induced suppression of hypoglossal motor activity. This latter evidence was based largely on the significant suppression of GG activity following application of the selective δ-opioid receptor agonist DPDPE to the hypoglossal motor pool compared to the preceding baseline ACSF values and, importantly, the corresponding ACSF time controls. These data are consistent with anatomical evidence for δ-opioid receptors at the hypoglossal motor pool (Richardson & Gatti, 2005). Nevertheless, application of the δ-opioid receptor antagonist naltrindole caused inconsistent and non-significant changes in GG activity (Fig. 6B). This lack of response to naltrindole may relate to the insufficient time for the response to manifest after 30 min application. Longer periods of observation and application of other agonists and antagonists were not performed in this study.

With respect to the experimental protocols, it was also noted that there was no effect on GG activity of a switch from fentanyl to ACSF in Study 1 (Protocol 1c in Fig. 2B), i.e. an increase toward the values observed in the corresponding ACSF time-control experiment did not occur. The most likely explanations for this persistent suppression of GG activity after 30 min washout of fentanyl include lingering inhibition due to downstream signalling mechanisms activated by the μ-opioid receptor activation, relatively tight binding of fentanyl to the μ-opioid receptor such that it is not so readily uncoupled by washout with ACSF compared to a switch to naloxone, and insufficient time for full washout and recovery with ACSF. Changes in GG activity with longer periods of fentanyl washout were not investigated in this study.

In addition to the identification of a μ- and δ-opioid receptor-mediated suppression of GG activity following perfusion of the respective opioid agonists into the hypoglossal motor pool, the specificity of this motor suppression in response to these interventions was also indicated by the lack of changes in control variables compared to the ACSF time controls. Such control variables included the amplitude of diaphragm activity (a control respiratory motor signal that should be unaffected by the interventions at the hypoglossal motor pool), respiratory rate (an indicator of the rhythmic activity generated by nearby medullary respiratory neurons that should also be unaffected by interventions at the hypoglossal motor pool), blood pressure (to rule out reflex hemodynamic effects on GG activity), and EEG activity (used as a marker of arousal state which may also indirectly influence GG activity). Importantly, unlike the significant suppressant effects on GG activity with μ- and δ-opioid receptor stimulation at the hypoglossal motor pool, there were no significant effects of those interventions on any of the aforementioned control variables compared to the ACSF time controls. The only statistically significant change observed in any control variable across each of Studies 1 to 4 was a lesser decline in respiratory rate compared to the ACSF time controls during κ-opioid receptor stimulation at the hypoglossal motor pool (i.e. Study 4). Overall, therefore, the most parsimonious explanation of the results reported in this paper is that the observed changes in GG activity in response to application of the opioid receptor agonists and antagonists were the result of specific effects occurring at the hypoglossal motor pool and not via indirect influences at other brainstem sites. Nevertheless, it is relevant to mention that in Studies 1–3 the decreases in respiratory rate approached statistical significance in each case (P= 0.065, 0.069 and 0.059), perhaps indicative that the application of fentanyl to the hypoglossal motor pool was capable of affecting distant structures, although not so readily and consistently as the robust local effects on hypoglossal motor output.

Previous studies and relevance to current findings

Although the isolated changes in GG activity in response to the interventions at the hypoglossal motor pool strongly suggest that opioid-induced responses were confined to influences at the motor nucleus, a justifiable criticism of the microdialysis method to locally apply these agents is that it is not possible to determine whether the observed changes in motor activity are exerted via pre- and/or post-synaptic receptor elements. Effects could even be mediated via influences on hypoglossal interneurons, although these constitute a small population in the hypoglossal motor nucleus (Travers, 1995). The concern regarding an inability to separate pre- and/or post-synaptic effects is fully acknowledged. Nevertheless, it is important to mention that this concern is more relevant to studies investigating the control of motor function by endogenously released neurochemicals, and is less of a concern with studies aiming to characterize the effects on motor activity with agents that are commonly used clinically, and that will access brainstem motor nuclei indiscriminately when so administered.

Nevertheless, with this caveat in mind, the simplest explanation of the present results in light of previous data regarding the distribution of opioid receptors in the hypoglossal motor nucleus is that the observed suppression of GG activity with δ-opioid receptor stimulation was mediated via presynaptic inhibition of excitatory inputs to hypoglossal motoneurons (Richardson & Gatti, 2005). In Richardson & Gatti (2005), the presynaptic location of δ-opioid receptors was hypothesized to inhibit excitatory glutamatergic inputs, inputs which mediate the transmission of respiratory drive to hypoglossal motoneurons in vitro (Greer et al. 1991; Funk et al. 1993) and in vivo (Steenland et al. 2006, 2008). In contrast, μ- but not δ-opioid receptors are located on the presynaptic terminals of raphé inputs to hypoglossal motoneurons, where they also act to decrease glutamate release from these projections (Bouryi & Lewis, 2004), an effect that would also contribute significantly to the observed suppression of hypoglossal motor activity (Sood et al. 2006). The clear suppression of GG activity by fentanyl at the hypoglossal motor pool is consistent with the presence of μ-opioid receptor binding sites at this motor pool, these being present in amounts over twice that of δ-opioid receptor binding sites (Sales et al. 1985; Xia & Haddad, 1991).

One previous study performed using slices of neonatal rodent medulla in vitro showed suppression of hypoglossal nerve activity with bath application of μ-opioid receptor agonists (Greer et al. 1995). However, whether that suppression occurred via effects at the motor nuclei per se or via suppression of premotor inputs such as from the medullary raphé (Bouryi & Lewis, 2004; Zhang et al. 2007) or central respiratory premotor neurons (Greer et al. 1995; Ballanyi et al. 1997; Gray et al. 1999; Takeda et al. 2001; Lalley, 2003; Manzke et al. 2003; McCrimmon & Alheid, 2003; Mellen et al. 2003) was not determined as the drugs were bath applied (Greer et al. 1995). Likewise, the mechanisms underlying the major suppression of hypoglossal nerve activity following intravenous administration of morphine was not determined in a previous study in decerebrate cats (Bartlett & St John, 1986). The results of the present study with local application of these agents to the hypoglossal motor pool suggests that the suppression of hypoglossal nerve activity in these previous studies may have occurred via effects occurring at the motor pool per se.

Summary

The results of the present study show that μ-opioid receptor mechanisms operating at the hypoglossal motor pool cause a suppression of the drive to the GG muscle of the tongue, and that any effect of muscarinic receptor-mediated inhibition does not play a significant role in this suppression. This μ-opioid receptor-induced suppression of tongue muscle activity by effects at the hypoglossal motor pool may underlie the clinical concern regarding adverse upper airway function in adult and pediatric OSA patients peri- and post-operatively in the presence of μ-opioid analgesics (Kryger, 2000; Brown et al. 2006; Lerman, 2006). The inhibitory effects of μ- and δ-opioid receptors at the hypoglossal motor pool also indicate an influence of the endogenous enkephalinergic and endorphinergic systems in respiratory motor control.

Glossary

Abbreviations

DPDPE

[d-Pen_2,5]-enkephalin hydrate

GG

genioglossus

HMN

hypoglossal motor nucleus

nor-BNI

nor-binaltorphimine dihydrochloride

OSA

obstructive sleep apnoea

Author contributions

M.H.: (1) conception, design, analysis and interpretation of data, (2) drafting the article and revising it critically for important intellectual content, (3) final approval of the version to be published. M.-A.D.: (1) conception, design, analysis and interpretation of data, (2) drafting the article and revising it critically for important intellectual content, (3) final approval of the version to be published. H.L.: (1) conception, design, analysis and interpretation of data, (2) revising the article critically for important intellectual content, (3) final approval of the version to be published. R.L.H.: (1) conception, design, analysis and interpretation of data, (2) drafting the article and revising it critically for important intellectual content, (3) final approval of the version to be published. The experiments were performed in the Medical Sciences Building at the University of Toronto.