Anesthetic Potency and Influence of Morphine and Sevoflurane on Respiration in μ-Opioid Receptor Knockout Mice

Mice with a disruption of exon 2 of the μOR gene and wild-type animals of the same mouse strain have been reported previously. 18All animals are 1:1 hybrids from 129/SV and C57BL/6 mouse strains, and hence both genotypes have an identical genetic background. The mice were genotyped using Southern blot analysis on mouse tail DNA using specific external probes to distinguish deletion of the μOR gene from the intact gene (see Matthes et al.  18for the details and results of the genotyping). Experiments were performed in mice of either sex (weight, 21–41 g) while the animals were 2–6 months of age. All experiments were conducted after obtaining approval for the protocol from the local Animal Ethics Committee (Dier Ethische Commissie, Leiden University Medical Center). Animal care was conducted in accordance with institutional guidelines.

Respiratory activity was measured using whole body plethysmography, originally described by Drorbaugh and Fenn, 25with an important adaptation allowing the continuous flow of dry gas through the measurement and referential chambers. A barometric measurement of ventilation is based on the principle that when an animal in a closed chamber inhales a tidal volume, this volume is warmed to body temperature and saturated with water vapor. On expiration, the gas is cooled down to ambient temperature and desaturated to the conditions in the chamber. The temperature and water vapor changes are the cause of small pressure changes in the chamber, which are measured with a pressure transducer. The pressure changes are proportional to the inspired and expired tidal volumes. By using a differential pressure measurement between the chamber that contains the animal (measurement chamber) and the reference chamber, pressure changes in the room where the experiment is taking place are canceled out. Both chambers are transparent and are approximately 600 ml each. The flow and composition of the gases was set by three mass flow controllers (Bronkhorst High Tec, Veenendaal, The Netherlands). The chambers were kept at room temperature (24–26°C). Data acquisition was started after an animal was placed in the chamber and had ample time for habituation.

Two nociceptive tests were performed: the tail-immersion test and the hot-plate test. In the tail-immersion test, the tails of the mice were immersed approximately 2 cm in water that was 54°C, and the latency time to a rapid tail flick was recorded. The cutoff time for this test was 15 s. In the hot-plate test, mice were placed on a plate heated to 52°C. The latency time to paw licking was determined, with a cutoff of 30 s.

Anesthetic potency was determined by assessing the minimum alveolar concentration (MAC) of sevoflurane according to the following method. 26,27Mice were anesthetized in an acrylic cylinder. The temperature of the mice was monitored and kept between 36 and 38°C with a heating lamp. Initially, mice randomly received one of two sevoflurane concentrations (1.3 and 4.0% in oxygen) for 20 min. A tail clamp (alligator clip) was applied to the tail for 30 s, and the mice were observed for movement in response to the stimulation. In case of a positive–negative motor response, the anesthetic was increased–decreased in steps of 0.3% until the positive–negative response reversed. Consecutive tail clamps were applied proximal to the previous test site, and only the middle third part of the tail was used for MAC assessment. After reversal of response had occurred, the study ended. The MAC was defined as the concentration midway between the concentration that permitted movement and the concentration that prevented movement. Twenty-minute inhalation periods were allowed for equilibration after each change in anesthetic concentration. 28,29Body temperature was measured via  a probe inserted approximately 1 cm rectally.

Breathing was measured while the animals breathed a gas mixture of 21% oxygen in nitrogen. Subsequently, to obtain a measure of the hypercapnic ventilatory response (HCVR), two to three elevations in inspired carbon dioxide were applied (inspired carbon dioxide concentrations: 3, 5, and 7%). When online analysis showed that a ventilatory steady state had not been reached, the duration of hypercapnia was extended. Minute ventilation, tidal volume, and respiratory frequency were calculated per breath. These data were averaged over 50 breaths and stored for further analysis. The HCVR was determined by the slope of the relation between ventilation and inspired carbon dioxide using least-squares regression analysis. In each animal (table 1), measurements were performed on five separate occasions, with at least 3 weeks between measurements.

Morphine (intraperitoneal administration of 6, 20, and 100 mg/kg; volume for each dose, 0.12 ml) and placebo (0.9% NaCl, 0.12 ml) were administered in random order to each animal (table 1), with at least 3 weeks between studies. Twenty minutes after the drug was administered, respiratory measurements as described above were performed.

After naloxone (intraperitoneal administration of 100 μg/kg, 0.1 ml) respiratory measurements as described above were performed. After a resting period, morphine (intraperitoneal administration of 100 mg/kg, 0.1 ml) plus naloxone (intraperitoneal administration of 100 μg/kg, 0.1 ml) were given and respiratory measurements repeated.

After assessment of control respiratory measurements as described above, 1% sevoflurane was added to the gas mixture flowing through the chambers. After an equilibration period of at least 20 min, 28,29and the observation that ventilation had reached a new steady state, another set of respiratory measurements was obtained.

After intraperitoneal administration of saline, 6, 20, or 100 mg/kg morphine (volume, 0.12 ml), and an equilibration period of 20 min, the nociceptive tests were performed by a researcher blinded to the doses, with at least 3 weeks between studies (table 1).

In one set of animals (table 1), baseline anesthetic potency was assessed, followed by MAC determination after the intraperitoneal administration of 20 mg/kg morphine. In another set of animals (table 1), baseline anesthetic potency was assessed, followed by MAC determination after intraperitoneal administration of 100 μg/kg naloxone.

The control respiratory data obtained during air breathing and the slope of the HCVR were compared between genotypes using analysis of variance. Because morphine doses of 6 and 20 mg/kg had little effect on the slope of the hypercapnic ventilatory response curve and respiratory depression (in terms of a depression of the slope of ventilation vs.  carbon dioxide) was observed only after a dose of 100 mg/kg (see Results), we performed a statistical analysis on the saline and 100-mg/kg morphine data using two-way analysis of variance (factors: genotype and treatment) and least-significant differences test. 30Data from naloxone-treated mice were compared with data from placebo-treated mice of the same genotype using Student t  tests. In wild-type animals, the significance of morphine-induced changes in naloxone pretreated mice were tested using paired t  tests. Sevoflurane-induced changes in respiratory parameters were compared between genotypes with Student t  tests. Statistical analysis of the analgesic and MAC data was conducted using two-way analysis of variance and least significant differences test. P  values < 0.05 were considered significant. All values reported are mean ± SD.

There were no obvious morphologic or behavioral abnormalities in the animals, apart from a tendency for exaggerated aggressive behavior in the male mutant mice. A total of 20 μOR^−/−and 22 μOR^+/+mice were used in the experiments. The two groups were matched with respect to weight and sex. During resting, morphine, and naloxone respiratory studies, the mice were relatively quiescent with their eyes open.

A total of 90 experiments were performed (50 and 40 in wild-type and mutant mice, respectively). Representative sample tracings of respiration during air breathing in μOR^+/+and μOR^−/−mice are shown in figure 1. There were subtle differences in breathing pattern between genotypes (values are the mean of animal means obtained at five sessions): tidal volume, 48 ± 4 μl in μOR^+/+and 44 ± 3 μl in μOR^−/−mice (nonsignificant [NS]); respiratory frequency, 188 ± 14 breaths/min in μOR^+/+and 220 ± 16 breaths/min in μOR^−/−mice (P < 0.0001); ventilation, 8.4 ± 0.9 ml/min in μOR^+/+and 9.5 ± 0.9 ml/min in μOR^−/−mice (P = 0.03). The slope of the ventilatory response to inspired carbon dioxide did not differ between genotypes (μOR^+/+1.7 ± 0.4 ml · min⁻¹·%⁻¹vs. μOR^-/-1.6 ± 0.8 ml · min⁻¹·%⁻¹; NS).

Representative sample tracings of the influence of 100 mg/kg morphine on respiration at two levels of inspired carbon dioxide in a μOR^+/+mouse and a μOR^−/−mouse are given in figure 2. Note the reduced breathing frequency in the μOR^+/+mouse relative to the μOR^-/-mouse at both carbon dioxide concentrations. Examples of the influence of 100 mg/kg morphine on the HCVR in a mutant mouse and a wild-type mouse are shown in figure 3A. In the mutant mouse, the response to inspired carbon dioxide was not affected by morphine. In contrast, in the mouse with an intact μOR gene, morphine caused the decrease of resting ventilation and the slope of the HCVR. The influences of the three tested morphine doses (6, 20, and 100 mg/kg) and saline on the slope of ventilation versus  carbon dioxide are shown in figure 4A. In μOR^−/−mice, 6, 20, and 100 mg/kg morphine caused no change in slope of the carbon dioxide response curve relative to saline. In μOR^+/+animals, morphine at 6 and 20 mg/kg morphine had no effect on the slope of the carbon dioxide response curve relative to saline, whereas an approximately 50% depression was observed at 100 mg/kg.

Table 2and figure 4Bgive the averaged effects of 100 mg/kg morphine on respiration relative to saline. Morphine, at a dose at which respiratory depression was obvious in μOR^+/+mice, had no effect on any of the measured variables in μOR^−/−knockout mice. In wild-type animals, ventilatory depression (as observed by a decrease in ventilation during air breathing) was caused by a reduction of breathing frequency but not tidal volume. The reduction in slope of the HCVR was caused by reductions of the tidal volume and respiratory frequency responses to carbon dioxide.

Ventilatory parameters after naloxone were as follows: ventilation, 11.6 ± 2.0 ml/min (P = 0.002 vs.  saline-treated animals of the same genotype); respiratory frequency, 278 ± 40 breaths/min (P = 0.001); tidal volume, 43 ± 11 ml (NS); slope of the HCVR, 2.3 ± 0.5 ml/min per percent carbon dioxide (P = 0.02).

Morphine in naloxone pretreated animals did not affect any of the measured respiratory variables: ventilation, 11.2 ± 5.5 ml/min (NS vs.  naloxone-treated animals); respiratory frequency, 256 ± 47 breaths/min (NS); tidal volume, 43 ± 15 ml (NS); slope of the HCVR, 2.1 ± 0.6 ml/min per percent carbon dioxide (NS) (figure 4B).

Ventilatory parameters after naloxone were as follows: ventilation, 15.1 ± 1.6 ml/min (P < 0.0001 vs.  saline-treated animals of the same genotype); respiratory frequency, 294 ± 32 breaths/min (P = 0.002); tidal volume, 52 ± 6 ml (P = 0.01); slope of the HCVR, 2.1 ± 0.3 ml/min per percent carbon dioxide (P = 0.04) (figure 4B).

Examples of the influence of 1% sevoflurane on the HCVR in a mutant mouse and a wild-type mouse are shown in figure 3B. In both mice, sevoflurane caused a decrease in resting ventilation and slope of the HCVR.

The changes (Δ) in tidal volume, timing parameters, and ventilation caused by inhalation of 1% sevoflurane did not differ between genotypes:Δ tidal volume, −1 ± 3 and −6 ± 11 μl in μOR^+/+and μOR^−/−mice, respectively (NS); Δ respiratory frequency, −36 ± 25 and −26 ± 27 breaths/min in μOR^+/+and μOR^−/−mice, respectively (NS); Δ ventilation, −1.5 ± 1 and −1.3 ± 2 ml/min in μOR^+/+and μOR^−/−mice, respectively (NS). Relative to control, sevoflurane reduced the slope of the HCVR by approximately 40% in both genotypes (Δ slope =−1.1 ± 0.3 and −1.0 ± 0.3 ml · min⁻¹·%⁻¹in μOR^+/+and μOR^−/−mice, respectively; NS).

In both nociceptive tests (hot plate and tail immersion), μOR^−/−mice displayed the absence of morphine-induced analgesia. At 6, 20, and 100 mg/kg morphine, latencies for hind paw licking and tail withdrawal were not different from those after saline (fig. 5). In contrast, in μOR^+/+animals, morphine showed a dose-dependent increase in latencies, with cutoff values reached for both tests at 100 mg/kg morphine (fig. 5). At all morphine doses tested, latencies were greater in μOR^+/+mice compared with μOR^−/−mice, whereas saline latencies did not differ between genotypes.

In the first set of mice (table 1), baseline sevoflurane MAC values differed by approximately 20% between genotypes: MAC in μOR^+/+mice = 2.7 ± 0.2 vol%versus μOR^−/−mice = 3.3 ± 0.5 vol% (P = 0.02). Morphine reduced the sevoflurane MAC by approximately 50% in wild-type animals (MAC after morphine = 1.4 ± 0.4%;P < 0.001 vs.  untreated animals of the same genotype) but did not affect the MAC of sevoflurane in mutant mice (MAC after morphine = 3.6 ± 0.6%; NS vs.  untreated animals of the same genotype) (figure 6).

In the second set of mice, baseline sevoflurane MAC values differed by approximately 20% between genotypes: MAC in μOR^+/+mice = 2.9 ± 0.2 vol%versus μOR^−/−mice = 3.5 ± 0.2 vol% (P = 0.02). Naloxone increased the sevoflurane MAC by 18% in wild-type animals (MAC after naloxone = 3.4 ± 0.2%;P < 0.01 vs.  untreated animals of the same genotype) but had no effect in mutant mice (MAC after naloxone = 3.6 ± 0.2%; NS vs.  untreated animals of the same genotype;fig. 6).

Using a knockout mouse model, we examined the influence of the μOR gene product on the control of breathing in the absence and presence of morphine, naloxone, and sevoflurane, and on the anesthetic potency of sevoflurane in the absence and presence of morphine and naloxone. Our main findings are as follows. (1) Relative to mice that do posses the μOR, mice that lack the μ-receptor displayed an approximately 10% greater level of ventilation (because of greater breathing frequencies). The slope of the ventilatory response to inspired carbon dioxide did not differ between genotypes. (2) Morphine, at a dose that caused analgesia and overt respiratory depression in wild-type animals, had no effect on ventilation and the ventilatory response to inspired carbon dioxide in mice with a deficient μOR. (3) Naloxone, at a dose that blocks μ-, κ- and δ-receptors, 31,32caused an increase in ventilation (by 50–90%) and slope of the ventilatory response to inspired carbon dioxide (by 25–35%) in mice with and without intact μ-receptors. (4) Morphine had no respiratory effect in naloxone-pretreated wild-type mice. (5) The volatile anesthetic sevoflurane caused respiratory depression independent of the μOR. (6) Morphine analgesia was absent in μOR-deficient mice. (7) Anesthetic potency was reduced in μOR^−/−mice relative to mice that do possess the μ-receptor. (8) While morphine increased and naloxone decreased anesthetic potency in wild-type mice, these agents had no effect in μ-receptor–deficient mice. Our observations indicate an important function for the endogenous opioid system in the physiology of the control of breathing with, however, only a modest role for the μOR gene product. Our findings implicate the μOR as sole mediator of the respiratory and analgesic actions of morphine. They further indicate the modulation of anesthetic potency but not anesthesia-induced respiratory depression by the endogenous μ-opioid system. Our studies confirm and extend our earlier studies in the same mouse strain, which showed the involvement of the μOR in morphine spinal and supraspinal analgesia, reward, physical dependence, respiratory depression, and immunosuppression. 18,33,34 

We remained uninformed on the arterial or end-tidal carbon dioxide and oxygen tensions in the animals. As a consequence, we calculated the slope of the HCVR from the inspired carbon dioxide tension. The major drawback of this approach is that a reduction of the slope of the relation between ventilation and inspired carbon dioxide (as was observed after 100 mg/kg morphine in wild-type animals only and during sevoflurane inhalation in both genotypes) may be related not only to opioid- and anesthetic-induced depression of respiratory neurons, but also to a reduction in metabolic rate. The differences in effect of morphine on slope between genotypes should therefore be viewed as the action of the μOR either to cause respiratory depression or to reduce metabolism or, most probably, to both. Note, however, that although at this stage we are unable to discriminate between these two mechanisms, there is ample evidence for a direct respiratory-depressant effect of opioids on brainstem neurons independent of metabolism. 5,8Hence, it is not unreasonable to attribute the reduction of slope predominantly to an effect of morphine on respiratory neurons in the brainstem. The same reasoning applies to the influences of sevoflurane on slope, although non–μ-opioid mechanisms are responsible for the respiratory effects of sevoflurane. Furthermore, during sevoflurane inhalation, because of overt hypoventilation or atelectasis, the measured carbon dioxide drive may reflect the combination of a hypoxic plus hypercapnic drive. Because this may have occurred in both genotypes, we believe that this did not influence our conclusions.

Our findings of a difference in resting ventilation between the two tested genotypes (table 2) and increase in ventilation after naloxone in both genotypes (fig. 4B) suggest the involvement of the opioid system in the physiology of the control of breathing. The genotype differences and the naloxone effect were primarily in breathing frequency. This is not surprising considering that endogenous opioid peptides and ORs are found in areas of the central nervous system involved in respiratory rhythmogenesis and frequency control, such as the preBötzinger complex (see below). 5Taking into account the finding that naloxone increased breathing frequency in mice with and without functional μ-receptors, and the small differences in resting values between μOR mutant mice (table 2), our data suggest only a minor role for the μ-opioid system. This may be related to the fact that the knockout mice in our study lacked the μOR gene throughout development. Functional compensation may have (partly) masked the phenotype (i.e. , the frequency generated by the respiratory rhythm generator) resulting from the long-term absence of the targeted gene. Furthermore, it may well be that differences in breathing frequencies between genotypes and naloxone effect are accentuated during stressful stimulation of the ventilatory control system, which may occur during exercise, chronic hypoxia, chronic elevations of arterial carbon dioxide concentration, pain, and suffering. For now, our in vivo  data do indicate the existence of an inhibitory tonic drive on respiratory rhythmogenic neurons from opioid peptides acting predominantly through non-μORs. A plausible candidate receptor is the δOR. Further studies in δOR knockout mice and studies using specific δ-opioid antagonists are needed to elucidate this matter.

Morphine, even at 100 mg/kg, had no effect on ventilatory control in mice lacking the μOR. This stands in contrast with the data from μOR-intact mice, showing reduction in ventilation, breathing frequency, and slopes of the ventilation, tidal volume, and respiratory frequency responses to carbon dioxide at 100 mg/kg morphine (table 2and figs. 2, 3A, and 4). These data show that the μ-opioid gene and receptor are essential in generating the respiratory effects of morphine. Finding no respiratory depression from 100 mg/kg morphine in μOR^−/−mice suggests further the absence of involvement of δ- and κ-receptors in morphine respiratory depression or, in analogy to the finding that δ-analgesia is dependent on functional μ-receptors, 33the need for functional μ-receptors in the production of respiratory depression from μ-receptors.

The site at which morphine induces bradypnea needs to contain neurons that express μORs and that are involved in the control of breathing frequency. We propose the preBötzinger complex as a possible site of the effect of morphine on breathing frequency. Recently, Gray et al.  5studied the in vitro  effects of injections of DAMGO, a μOR agonist, into the preBötzinger complex of rodents. DAMGO decreased endogenous respiratory-related rhythm. The preBötzinger complex is part of a narrow column of respiratory neurons in the ventromedial medulla extending from the facial nucleus to the spinal cord (the ventral respiratory group) and involved in respiratory rhythmogenesis and frequency control. 5,35We are aware that this may be just one of many sites in the peripheral and central nervous system via  which exogenous administered opioids may exert an effect on the control of breathing (other sites include the locus coeruleus and the tractus nucleus solitarius). 3,4Note also that the carotid bodies contain μORs. 2These small organs, part of the peripheral nervous system, hold the peripheral chemoreceptors and are located at the bifurcations of the carotid arteries and are responsible for 20–30% of ventilatory drive at rest and for more than 80% during hypoxia. 36Further studies are needed to identify sites involved in exogenous opioid-induced reduction of breathing frequency and tidal volume.

Between genotypes, we observed no differences in sevoflurane-induced changes in resting ventilation and slope of the HCVR (fig. 3B). The respiratory depression caused by 1% sevoflurane is in agreement with observations in other species. 37Anesthetics and opioids affect ventilatory control via  distinct neuronal pathways. Recent studies indicate that inhalational anesthetics, such as sevoflurane, cause respiratory depression via  reduction of glutamatergic excitation in the brainstem 38and possibly also via  stimulation of an oxygen-sensitive background K⁺channel in type-I cells of the carotid bodies. 39Our data indicate the absence of functional interaction between the endogenous μ-opioid system and neuronal networks involved in anesthesia-induced respiratory depression. On the other hand, we did observe the modulation—i.e.,  potentiation—of anesthetic potency (defined as the concentration sevoflurane necessary to suppress a motor response to noxious stimulation of the tail) by μ-opioid-receptors. Mice with active μ-receptors had a 20% lower MAC relative to the μOR^−/−mice, and naloxone increased the MAC in wild-type animals exclusively (fig. 6). This further indicates the absence of involvement of δ- and κ-ORs in the modulation of anesthetic potency. Our data are in agreement with clinical observations of synergistic μ-opioid–anesthetic interaction on anesthetic potency. 23,24We hypothesize that the differences in μ-receptor involvement in the respiratory system and in anesthetic potency are a result of different neuronal networks and possibly also different receptor systems involved in anesthesia-induced respiratory depression, and anesthesia-induced suppression of somatic responses (MAC is predominantly mediated by spinal pathways), 40with absent and active interactions with the endogenous μ-opioid system, respectively.

In conclusion, our data indicate that the endogenous opioid system plays an important role in the control of breathing frequency. The finding that knockout mice, without μ-receptors, were unaffected by morphine, showing neither respiratory depression nor analgesia, confirms the concept that μORs are the essential targets of morphine on both respiratory control and pain response.