Activation of Opioid μ Receptors in Caudal Medullary Raphe Region Inhibits the Ventilatory Response to Hypercapnia in Anesthetized Rats

The experimental protocols were approved by the Institutional Animal Care and Use Committee in Lovelace Respiratory Research Institute, Albuquerque, New Mexico, accredited by the Association for Assessment and Accreditation of Laboratory Animal Care International, USA. The experiments were performed in tracheotomized and spontaneously breathing Sprague-Dawley adult male rats (350–450 g).

Animals (n = 32) were initially anesthetized with sodium pentobarbital (30 mg/kg) and urethane (600 mg/kg, intraperitoneal). Supplemental urethane was administered, if needed, to reach an adequate level of anesthesia where rats exhibited neither an eye-blink nor limb-withdrawal reflex throughout the experiment. The right femoral vein and artery were cannulated, the former for drug administration and the latter for monitoring mean arterial blood pressure and heart rate (HR). The trachea below the larynx was exposed through a midline incision, tracheotomized by blunt dissection, and cannulated. The tracheal cannula was connected with a pneumotachograph to record airflow. The pneumotachograph had a linear flow–pressure relation in the range of 2–20 ml/s, a flow resistance of 0.046 cm H²O ml/s, and a dead space of 0.2 ml. Another end of the pneumotachograph was placed (approximately 5 mm deep) in a plastic tube with a diameter fivefold greater than that of the pneumotachograph. A three-way stopcock was attached to the other side of the plastic tube and connected to a supplemental gases device through a gas-mixing flowmeter (GF-3MP; Cameron Instrument Co., Port Aransas, TX) with the flow volume at 1.0 l/min. By turning the switch, the gases to be inhaled from the different tanks were controlled. End-tidal pressure of carbon dioxide (Petco²) was measured via  a carbon dioxide analyzer (MicroCapStar end-tidal carbon dioxide analyzer, model 15-10000; CWE, Inc., Ardmore, PA) connected to a side port of the tracheal cannula. Animals were placed into a rigid metal frame with their heads fixed and centered in a stereotaxic apparatus (model 1404; Kopf, Tujunga, CA). A hole (approximately 10 mm in diameter) was drilled at the midline of the skull in some rats for microinjection into the MRR. The animals' core temperature was monitored with a rectal probe and maintained at 36.5°–37.5°C by a heat pad and radiant heat lamp.

The rats were exposed to four progressive, steady state concentrations of carbon dioxide, i.e. , 3, 5, 7, and 9% CO²(balanced with 40% O²and varied percentages of nitrogen). Each concentration lasted continuously for 2 min before switching to the next higher carbon dioxide concentration. It should be noted that the V^Eand V^Tresponses to hypercapnia did not always reach the plateau during the 2-min exposure in the current study.

A 25-gauge, 0.5-μl needle (Hamilton, Reno, NV), prefilled with saline or agents, was inserted into the given region of the MRR. According to the rat stereotaxic atlas of Paxinos and Watson,35the MRR, extended from 9–12 mm caudal to the bregma, was divided into three subregions, i.e. , the rostral, middle, and caudal MRRs (rMRR, mMRR, and cMRR, respectively), located at 9.0, 10.5, and 12.0 mm caudal to the bregma, respectively. The rMRR contained the magnus nucleus; the mMRR contained the magnus nucleus and its neighbor pallidus nucleus (RPa). The central sites for the mMRR and rMRR were localized 9 mm ventral to the cerebellar surface, whereas the central sites of the obscurus nucleus (ROb) and RPa that compose the cMRR were localized 8.3 and 9.3 mm ventral to the cerebellar surface. Each site of the rMRR and mMRR received 100 nl microinjection. Because the two nuclei in the cMMR were located separately and distantly, two injections (each 100 nl) were given. That is, the first was given when the needle was placed into the center of the ROb. After completion of the first injection, the needle was advanced 1 mm deeper, equivalent to the RPa, followed by the second injection. To test the unique role of the cMRR, DAMGO was purposely administered into the regions outside the cMRR in few rats, in which DAMGO was microinjected into the sites lateral (0.5 mm) to the cMRR.

Hyperoxia (40% O², 1 l/min) was applied throughout the experiment to serve as a baseline. After stabilization of the cardiorespiratory baseline values for at least 10 min, effects of intravenous injection of DAMGO (0.3 ml, 100 μg/kg, intravenous; similar to the dose used in previous studies36) on baseline ventilation were tested in 16 rats. Fourteen of these rats received the progressive hypercapnic exposure twice before systemic DAMGO injection and once 5 and 30 min after injection. The same protocols were repeated 2 h later in 2 of 14 rats to test the reproducibility of the DAMGO effect on the response to carbon dioxide over time. To determine the role of cMRR μ receptors in the ventilatory responses to systemic DAMGO, the effects of systemic DAMGO on ventilation and the responses to carbon dioxide were tested before and 2–3 min after local administration of CTAP (100 ng/100 nl) into cMRR in 5 of 14 rats.

The effects of stimulating the MRR μ receptors on cardiorespiratory activity were estimated by microinjection of DAMGO (35 ng/100 nl) into the cMRR, mMRR, and rMRR in three groups of rats (n = 7, 6, and 6, respectively), and the dose was chosen according to the previous studies.37,38Among these animals, three rats in each group came from those previously receiving an intravenous administration of DAMGO, and the remaining 10 rats underwent no previous systemic DAMGO. In those animals receiving previous intravenous DAMGO, a 2-h interval was allowed for recovery from systemic administration of DAMGO before microinjection into the MRR subregions, because this intravenously injected agent has a brief elimination half-life around 15 min in mammals.39Our preliminary studies showed that microinjection of DAMGO into the cMRR rather than mMRR or rMRR evoked remarkable respiratory changes, and the recovery time was longer as compared with the responses induced by intravenous injection. Therefore, progressive hypercapnia was performed twice before microinjection and once 5, 30, 60, and 90 min after microinjection.

Six other rats were also used to test whether systemic and local injection of vehicle (n = 3) or whether local injection of DAMGO into the regions outside the cMRR (n = 3) could alter the respiratory responses.

After completion of the experiment, all animals were killed by an overdose of anesthetic and cervical dislocation. The brainstem was removed and fixed by soaking in 4% paraformaldehyde (pH 7.4) for at least 36 h at 4°C, and subsequently sectioned at a 40-μm thickness by a slicing machine (Leica, CM 1850; Microsystems GMbH, Nussioch, Germany). The area marked by Chicago sky blue was identified under a microscope, and the center of the stained area was used as the location of microinjection.

DAMGO and CTAP (Sigma-Aldrich, St. Louis, MO) were dissolved in 0.9% saline containing 1% Chicago sky blue for central injection. DAMGO for systemic injection was prepared in 0.9% saline.

Raw data of the airflow signal, blood pressure, HR, Petco², and rectal temperature were digitized, monitored, and recorded by using a PowerLab/8sp (model ML 785; AD Instruments Inc., Colorado Springs, CO) connected to a computer using the PowerLab Chart 5 software. The airflow signals were integrated to generate tidal volume (V^T), respiratory frequency (f), and minute ventilatory volume (V^E). After stabilization, the cardiorespiratory baseline was determined by averaging the variables for 1 min immediately before administration of DAMGO. Because DAMGO injection into the femoral vein or the cMRR inhibited V^Eand increased Petco², this V^Ewas defined as hypercapnic baseline V^E. Three steps (isocapnic calculation) were taken to correct the DAMGO-induced hypercapnic effects on the baseline V^E. We calculated (1) ΔPetco²(the differences of Petco²before and after DAMGO administration), (2) ΔV^E(slope of V^E− Petco²×ΔPetco²), and (3) isocapnic baseline V^Eafter DAMGO administration (the differences between the hypercapnic baseline V^Eand ΔV^E). The adjustment of baseline V^Tand f after DAMGO administration was conducted by using the same approach. The responses to progressive hypercapnia were defined by averaging the variables for the last 10 s of the 2 min of given carbon dioxide exposure. Our preliminary studies showed linear correlations of respiratory (V^E, V^T, and f; P < 0.05) instead of cardiovascular variables, with Petco²as reported previously.15,40The linear correlation between V^E(V^Tor f) and Petco²was tested by Pearson correlation with Microsoft Excel (Microsoft Corporation, Redmond, WA). Subsequently, if significant correlations existed, least squares analyses of linear regression were applied to obtain the values of the slopes. These slopes were used as the indices of ventilatory responsiveness to hypercapnia (chemosensitivity). The blood pressure and HR responses to carbon dioxide were presented as percent change from control (without hypercapnia). The control ventilatory responses to carbon dioxide were determined by averaging the relevant variables from the two tests before administration of DAMGO.

All data are presented as mean ± SE. A paired t  test was used to test the significant differences of cardiorespiratory baseline variables obtained immediately before and after administration (systemic or local) of DAMGO. One-way analysis of variance for repeated measures was used to compare the differences of the cardiorespiratory responses to progressive hypercapnia: (1) before and 5 and 30 min after systemic or local administration of DAMGO into the rMRR and mMRR; (2) before and 5, 30, 60, and 90 min after microinjection of DAMGO into the cMRR; and (3) before and after systemic DAMGO with and without pretreatment of CTAP. The Fisher least significant difference posttest was used if the overall analysis of variance (an omnibus test) had a P  value less than 0.05. The software Statistica 6.0 (StatSoft, Inc., Tulsa, OK) was used for statistical analysis. Differences are considered significant at a P  value less than 0.05.

In the current study, baseline Petco², V^E, f, V^T, blood pressure, and HR in the 32 anesthetized rats were 43.2 ± 2.6 mmHg, 198.6 ± 8.4 ml/min, 103 ± 4 breaths/min, 1.9 ± 0.1 ml, 122 ± 4 mmHg, and 367 ± 11 beats/min, respectively, which are similar to those reported previously.41Systemic DAMGO significantly increased baseline Petco²(from 41.3 ± 2.4 to 48.4 ± 2.6 mmHg; n = 16; P < 0.01) and decreased baseline V^Eby 20% due to the depression of f, and lowered blood pressure by 18% with no change in HR (fig. 1). The baseline V^E, f, and V^Tbefore and after systemic DAMGO were 201 ± 10 versus  161 ± 6 ml/min (P < 0.01), 105 ± 5 versus  76 ± 4 breaths/min (P < 0.01), and 2.1 ± 0.2 versus  2.2 ± 0.2 ml (P > 0.05), respectively. After isocapnic calculation, the adjusted baseline V^E, f, and V^Tafter systemic DAMGO were significantly reduced to 56 ± 8 ml/min, 72 ± 6 breaths/min, and 1.2 ± 0.3 ml, respectively (P  values < 0.01 as compared with the data obtained before DAMGO). As compared with baseline, 9% CO²increased the responses of V^E, f, and V^Tby 234 ± 16, 18 ± 3, and 182 ± 14%, respectively. The representative recordings of the cardiorespiratory responses to progressive hypercapnia before and after intravenous DAMGO are shown in figure 2, whereas the group data of respiratory responsiveness to hypercapnia are presented in figure 3. Intravenous administration of DAMGO significantly depressed the V^Eand V^Tresponsiveness to carbon dioxide by 31% and 33%, respectively, with little effect on f responsiveness (figs. 2 and 3). The ventilatory changes disappeared approximately 30 min after DAMGO administration. With respect to cardiovascular activity, no significant blood pressure (−0.3 ± 1.1%) or HR responses (−2.9 ± 1.9%) to hypercapnia were found, as they were not affected by intravenous DAMGO (0.2 ± 0.9% and 0.3 ± 0.8%; n = 14; P > 0.05). It should be noted that these cardiovascular data were averaged from the responses to four degrees of hypercapnia because DAMGO did not alter the cardiovascular response to each hypercapnic level. This was the same for the subsequent experiments. Intravenous injection of the vehicle, 0.9% saline alone, produced no effect on the cardiorespiratory variables.

To evaluate the role of cMRR μ receptors in the ventilatory responses to systemic DAMGO, the effects of systemic DAMGO on ventilation and the responses to carbon dioxide were tested before and 2–3 min after local blockade of cMRR μ receptors in five rats. As illustrated in figure 4, the systemic DAMGO-induced inhibition of baseline ventilation (−20%) and the response to carbon dioxide (−31%) was substantially diminished by pretreatment of the cMRR with CTAP (100 ng/100 nl). After the pretreatment, the systemic DAMGO only inhibited the baseline ventilation by 11% via  acting on f and the response to carbon dioxide by 14% via  changing V^Tresponse. That is, local CTAP pretreatment in the cMRR diminished the systemic DAMGO-induced inhibitions for baseline ventilation and the response to carbon dioxide by 45% and 55%, respectively. Because the interval between the first and second systemic DAMGO (after the local injection of CTAP) was approximately 2 h, we tested whether the respiratory responses to the systemic DAMGO were changed over this period in two rats. We found no differences between the first and second systemic DAMGO-induced reduction of ventilation (−19 ± 2 vs. −20 ± 3%) and the response to carbon dioxide (−30 ± 3 vs. −31 ± 2%).

As shown in figure 5, microinjection of DAMGO into the cMRR in seven rats significantly increased Petco ²(from 43.3 ± 2.5 to 47.6 ± 3.1; n = 7; P < 0.05) and reduced baseline V^Eby 22%, mainly due to depressing baseline V^T(by 18%), whereas the baseline values for f, blood pressure, and HR were not significantly affected. As compared with control, microinjection of DAMGO into the cMRR significantly diminished the baseline V^E(212 ± 12 vs.  166 ± 11 ml/min; P < 0.01) and V^T(2.1 ± 0.2 vs.  1.7 ± 0.2 ml; P < 0.05), with little effect on f (98 ± 8 vs.  98 ± 10 breaths/min; P > 0.05). After isocapnic calculation, a greater decrease in V^E(105 ± 8 ml/min) and V^T(1.3 ± 0.2 ml) was observed as compared with those without isocapnic calculation. It should be noted that the adjusted baseline f (97 ± 8 breaths/min) was still not significantly different from the control (P > 0.05). Microinjection of DAMGO into the cMRR significantly depressed V^Echemosensitivity to carbon dioxide by 24% and 28% 5 and 30 min after microinjection, respectively. This inhibitory effect was due to the depression of both V^Tand f responsiveness. The animals' ventilatory responses to hypercapnia were recovered 60 min after DAMGO microinjection. The typical experimental recordings and the group data showing the cardiovascular and/or respiratory responses to progressive hypercapnia before and after microinjection are displayed in figures 6 and 7, respectively. The absence of cardiovascular response to hypercapnia was not altered by microinjection of DAMGO into the cMRR (blood pressure, 3.75 ± 1.98% and HR, −2.21 ± 1.88%; n = 7; P > 0.05). Microinjection of vehicle into the cMRR produced no remarkable effect on baseline respiration and the response to carbon dioxide in three rats (table 1). In addition, microinjections were given in the regions outside the cMRR in three rats as marked in figure 7D. These injections did not significantly change the baseline ventilation and the response to carbon dioxide (table 1). In these cases, absence of significant changes of baseline blood pressure (129 ± 12 vs.  128 ± 13 mmHg) and HR (345 ± 12 vs.  347 ± 13 beats/min) in response to DAMGO were also observed.

As shown in figures 8 and 9, microinjection of DAMGO into the mMRR produced no significant effect on baseline cardiorespiratory variables or the respiratory response to progressive hypercapnia. The same results were also observed when DAMGO was microinjected into the rMRR (figs. 10 and 11). Similar to microinjection into the cMRR, the absence of cardiovascular response to hypercapnia was not affected by DAMGO microinjected into the mMRR (blood pressure, 2.47 ± 1.81% and HR, 0.86 ± 0.57%; n = 6; P > 0.05) or the rMRR (blood pressure, 2.97 ± 1.12% and HR, 1.62 ± 0.71%; n = 6; P > 0.05).

In the current study, we found that intravenous administration of DAMGO significantly inhibited baseline ventilation by 20% mainly due to depressing f, associated with an increase in Petco ². After isocapnic adjustment, a much greater inhibition was uncovered. A large number of previous studies has demonstrated that opioids administered systemically decrease breathing mainly through slowing f in humans and animals, especially in the anesthetized condition3,8,12,42–44(see the review8). In the current study, we found that as compared with control, hypercapnia increased the V^Eresponses by 234% mainly via  increasing V^T(182%), which is similar to the results reported previously.8Systemic administration of DAMGO attenuated the ventilatory responsiveness to progressive hypercapnia by 31% owing to a reduction of V^Trather than f responsiveness. In agreement, it has been indicated that the slopes of the hypercapnic ventilatory responses were decreased and shifted to the right after systemic delivery of opioid agonists in humans, monkeys, rats, and mice.9–13Among these studies, activation of μ receptor–induced attenuation of V^T(up to 29%) is dominant as compared with the slowing of f.11,13,45,46 

The MRR is an area responsible for the ventilatory responsiveness to hypercapnia24–27and nociception.23Although in rats the MRR has a rich distribution of μ receptors,34it is unknown whether MRR μ receptors are involved in opioid-induced depression of ventilation and the ventilatory responses to hypercapnia. Our major finding in the current study is that microinjection of DAMGO into the cMRR, but not the mMRR or rMRR, decreased baseline ventilation by 22%, with an associated increase in Petco ². After isocapnic calculation, this inhibition became 50%, predominantly via  reducing V^T. More important, microinjection of DAMGO into the cMRR attenuated the hypercapnic ventilatory responsiveness by 28% due to a decrease in both responses of V^Tand f. This is the first evidence to show that among the MRR, the cMRR μ receptors are uniquely inhibitory to baseline ventilation and the ventilatory response to hypercapnia. To estimate the contribution of cMRR μ receptors to the systemic DAMGO-induced ventilatory responses, we compared the baseline ventilatory responses and hypercapnic ventilatory responses to systemic DAMGO before and after blockade of cMRR μ receptors by local injection of CTAP. We found that the inhibition of baseline (−20%) and carbon dioxide ventilatory responses (−31%) to systemic DAMGO was diminished to −11% and −14%, respectively. That is, blockade of cMRR μ receptors diminished the systemic DAMGO-induced inhibitions for baseline ventilation and the response to carbon dioxide by 45% and 55%, respectively, suggesting that these μ receptors play a major role in inhibiting ventilation and attenuating the ventilatory responses to carbon dioxide. Although opioids inhibit respiration mainly via  stimulating central μ receptors,12,15–20some early studies indicated involvement of both μ and Δ receptors21,22and even showed that Δ receptors could modulate or counteract the respiratory depression induced by μ receptors.16Our data cannot rule out the possible involvement of MRR Δ receptors in the respiratory modulations. Directly answering whether the hypercapnia-recruited cMRR neurons containing μ receptors are chemosensitive and/or respiratory modulated is beyond the scope of the current study. The presence of chemosensitive neurons in the cMRR has been well established.24,25,28On the other hand, investigators have also shown that the cMRR contains respiratory-modulated neurons in cats,47,48and the firing patterns of these neurons are altered during fictive cough.47Ultimately, further studies are required to clarify which types of cMRR neurons contain μ receptors (chemosensitive and/or respiratory-modulated neurons), and how stimulation of these receptors alters the ventilatory responsiveness to hypercapnia.

The significance of our study is evident. The most adverse effect of opioids, even at therapeutic doses, is the substantial depression of breathing that could be lethal. In fact, the rate of opioid-induced deaths in 1999 was 122 per million in Australia.49Opioids markedly depress the ventilatory response to hypercapnia that is essential for maintaining a normal eupneic breathing and critical for survival in awake and/or anesthetized humans and animals.2–6,12,16–18To date, however, how morphine depresses ventilation, leading to respiratory failure in the clinic, remains unclear. Our finding that cMRR μ receptors play an important role in the systemic DAMGO-induced inhibition of ventilation and the response to carbon dioxide benefits our understanding of central morphine depression of ventilation, especially the ventilatory response to carbon dioxide. In fact, there is evidence to suggest some functional and anatomical similarities in the MRR between humans and rats. MRR serotonergic neurons have been believed to be involved in chemosensitivity to carbon dioxide in vivo  24,25and in vitro .28,29A lesion or inhibition of these neurons limited in the MRR profoundly attenuates the ventilatory response to hypercapnia in animals.26,27In infants, serotonergic neurons are heavily distributed in the MRR that is homologous in position to chemosensitive neurons in rats.50The abnormality of medullary serotonergic system (the MRR) has been linked to some respiratory disorders such as sudden infant death syndrome.51,52The important role of cMRR μ receptors in morphine-induced attenuation of the ventilatory response to carbon dioxide also builds a foundation for further defining whether cMRR μ receptors modulate the ventilatory chemosensitivity via  local serotonergic mechanism.

Mu receptors in the MRR are not critical for cardiovascular regulation. In this study, intravenous DAMGO can significantly depress blood pressure by 18% (fig. 1), similar to the previous studies reported in anesthetized rats.53Unexpectedly, this cardiovascular modulation was not observed when DAMGO was microinjected into the MRR subregions we tested. There are three reasons that may account for the absence of cardiovascular effect in our study. First, the depressed cardiovascular effect by systemic administration of opioids was blocked by naloxone methiodide, an opioid antagonist restricted to the periphery,53suggesting peripheral μ receptors' cardiovascular modulation. Second, instead of the MRR, μ receptors are also located within other brainstem regions responsible for cardiorespiratory modulation.54Therefore, their influence on cardiovascular modulation cannot be ruled out. Third, it is possible that activation of the whole MRR rather than only one of its subnuclei is required for involvement of MRR μ receptors in cardiovascular regulation.

There are several limitations in this study. As mentioned in the Materials and Methods section, the V^Eand V^Tresponses to hypercapnia did not always reach the plateau during the 2-min exposure in the current study. This suggests that the ventilatory sensitivity reported here seems not to be obtained in a full hypercapnic steady state condition. However, it would not invalidate our finding that DAMGO inhibits ventilation and the responsiveness to hypercapnia because the rats treated without and with DAMGO had undergone the same hypercapnic exposure. Moreover, the ventilatory responses to progressive hypercapnia by rebreathing carbon dioxide have been widely used to reflect the central ventilatory chemosensitivity in animals and humans.13,55–57It was reported that the systemic administration of opioids had greater potency and efficacy in male than in female rodents58–60although debatable in humans.61No attempt was made in the current study to delineate whether the sex difference also exists in the mice. Urethane was chosen in the current study because of its minimal cardiorespiratory depression compared with other anesthetics. We cannot rule out the possible interaction between the anesthetic and DAMGO in changing ventilation and the response to carbon dioxide. We did not directly measure the diffused size of microinjection of DAMGO. However, because injecting the same volume of DAMGO (100 nl) into the regions outside the cMRR (0.5 mm lateral to the cMRR) or into the mMRR (1.5 mm rostral to the cMRR) did not alter ventilation and the response to hypercapnia, the spread should be considerably limited. In addition, because of the cylinder-shaped ROb in the cMRR, the microinjection used in this study was not expected to block all μ receptors in this region. Unlike the systemic administration of DAMGO, the effects of microinjection into the cMRR on the ventilatory response to carbon dioxide lasted much longer (5 min vs.  30 min), similar to application of morphine on the medullary ventral surface in anesthetized cats.7It is unknown why the effects elicited by local microinjection are prolonged. However, the prolonged respiratory depression after injection into the cMRR may be related to the diffusion of the injection from each other of the ROb and RPa or a longer downstream effect of DAMGO.

In summary, it is well documented that opioids centrally inhibit respiration and depress the ventilatory responsiveness to hypercapnia in anesthetized humans and animals. However, the mechanisms related to this depressed response remain unknown. Our results show that microinjection of DAMGO into the cMRR rather than mMRR or rMRR significantly depressed ventilation and the response to hypercapnia in anesthetized rats. Moreover, blockade of cMRR μ receptors diminished the systemic DAMGO-induced inhibitions for ventilation and the response to carbon dioxide by 45% and 55%, respectively. We conclude that cMRR μ receptors are at least one of the major contributors to opiate-induced depression of the ventilatory response to hypercapnia.

The authors thank Jianguo Zhuang, M.D., Ph.D. (Associate Research Scientist, Pathophysiology Program, Lovelace Respiratory Research Institute, Albuquerque, New Mexico), for his assistance with the experiment setup and statistics.