Opioid Action on Respiratory Neuron Activity of the Isolated Respiratory Network in Newborn Rats

The study was performed in accordance with the protocol approved by the institutional Animal Care and Use Committee of the Karolinska Institute, Stockholm, Sweden. Brainstem–spinal cord preparations from 0- to 4-day-old Sprague-Dawley rats (163 preparations) were used. 8,9,11Brainstem and spinal cord were isolated with deep ether anesthesia. 3The brainstem was rostrally decerebrated between the VIth cranial nerve roots and the lower border of the trapezoid body, so that most of the pons was removed. The preparation with the ventral surface up was then placed in a perfusion chamber (2 ml) continuously superfused at a rate of 3.0–3.5 ml/min with artificial cerebrospinal fluid: 124 mm NaCl, 5.0 mm KCl, 1.2 mm KH²PO⁴, 2.4 mm CaCl², 1.3 mm MgSO⁴, 26 mm NaHCO³, and 30 mm glucose, equilibrated with 95% O²and 5% CO², and kept at 26–27°C and a pH of 7.4.

Intracellular whole cell recordings were made from three major classes of respiratory neurons using the blind patch clamp technique. 10,13–15Electrodes were pulled in a single stage from thin-wall borosilicate glass (GC100TF-10, OD = 1.0 mm, with a filament, Clark Electromedical, Reading, United Kingdom) on a vertical puller. The electrodes had an inner tip diameter of 1.2–2.0 μm, a DC resistance of 3–8 MΩ, and the electrode solution consisted of 130 mm potassium gluconate, 10 mm EGTA, 10 mm HEPES, 2 mm Na²–adenosine triphosphate, 1 mm CaCl², and 1 mm MgCl², with a pH of 7.2–7.3 adjusted by KOH. The electrodes were placed in the RVLM, including the ventral respiratory group, by insertion through a small region of the ventral surface of the RVLM where the pia mater was removed with a thin glass needle. The E^m(millivolts) was recorded with a single-electrode voltage clamp amplifier (CEZ-300, Nihon Kohden, Tokyo, Japan) after compensation of the series resistance (20–60 MΩ) and capacitance. For comparison with previous studies, 13,14the membrane potential values were not corrected for junction potentials (< −10 mV). Input membrane resistance (R^m, megaohms) was determined from the voltage change in response to intracellular injection of DC current pulses (500 ms, −20 and −40 pA) during the silent phase of the respiratory cycle or during negative holding potentials (−50 or −55 mV) of expiratory neurons. Discharges of respiratory motor activity were recorded extracellularly with suction electrodes applied to the proximal ends of cut ventral roots of spinal (C4 or C5) nerves and through a high-pass filter with a 0.3-s time constant. During the experiments, neuronal activity was displayed on a chart recorder, monitored on oscilloscope, digitized (Digidata 1200B, Axon Instruments, Foster, CA), and stored on a DAT tape (RD-120TE, TEAC, Tokyo, Japan) or a hard disk for offline analysis with a personal computer and data acquisition software (Axoscope, Axon Instruments).

Membrane potential was recorded from three major classes of respiratory neurons in the RVLM (fig. 1). Preinspiratory neuron action potential discharges are observed during the preinspiratory and postinspiratory phases. 13,16The expiratory neurons are characterized by tonic action potential discharges during the expiratory phase and by inhibitory postsynaptic potentials (IPSPs) that cause hyperpolarization and inhibited spontaneous firing during the inspiratory phase. 14,17Preinspiratory and expiratory neurons receive inhibitory postsynaptic inputs from inspiratory neurons. 10,11Inspiratory neurons were classified into three subtypes. 13,14,18Type I inspiratory neurons show excitatory postsynaptic potentials (EPSPs) before onset and after termination of inspiratory nerve bursts. In type II neurons, EPSPs are only revealed during the inspiratory phase, whereas type III neurons are hyperpolarized by synchronized IPSPs during the preinspiratory and postinspiratory phases. 8,13Types I and III inspiratory neurons receive excitatory and inhibitory postsynaptic inputs, respectively, from preinspiratory neurons. 10,11 

Neuronal projections and evoked EPSPs were studied using whole cell recordings of inspiratory neurons (n = 17). Contralateral stimulation of the RVLM (0.1 ms duration single square wave impulses of 15 V) was delivered by tungsten electrodes with 2 MΩ resistance and 30-μm tip diameters. Simultaneous recordings of intracellular and extracellular E^mof inspiratory–preinspiratory neuronal pairs (in the contralateral medulla) were made in 10 preparations to elucidate whether opioid receptor agonists interfered with the correlation between these neuronal bursts. Extracellular unit activity of respiratory neurons in the RVLM was recorded by using glass microelectrodes filled with 0.5 m sodium acetate (5–20 MΩ).

Actions of four opioid receptor agonists were examined:μ-opioid receptor agonists by the use of morphine hydrochloride (Pharmacia, Stockholm, Sweden) and DAGO (Try-D-Ala- Gly-[N MePhe]-Gly-ol; Sigma, St. Louis, MO), κ-opioid receptor agonist by U50488 (trans-(±)-3,4-Dichloro-N -methyl-N -[2-(1-pyrrolidiny)-cyclohexyl]-nzeneacetamide methanesulfonate; Research Biochemicals International, Natick, MA), and the δ-opioid receptor agonist by DPDPE ([D-pen^2,5]-enkephalin; Sigma). Drug concentration and application time were determined according to results from previous pharmacologic studies. 4,6Artificial cerebrospinal fluid containing 1 μm DAGO and 10 μm morphine were applied for 10 min, while artificial cerebrospinal fluid with 10 μm DPDPE and 10 μm U50488 were applied for 20 min. This was followed by the application of opioid receptor antagonist naloxone hydrochloride (Sigma; 1 μm) for 10 min. In 68 preparations (47 inspiratory, 11 preinspiratory, and 10 expiratory neurons), the opioid receptor agonists were applied after blockade of ongoing synaptic activity with 0.5 μm tetrodotoxin (Sigma). The whole brainstem–spinal cord preparation was exposed via  bath application of pharmacologic agents. All solutions were adjusted to a pH of 7.4 and were applied to the recording chamber through the perfusing system.

Repeated-measurement analysis of variance and Bonferroni tests were performed to distinguish within-group differences over time. The one-way analysis of variance and Bonferroni tests were performed to evaluate differences among the respiratory neurons and the test drugs. All statistical analyses were conducted using the SPSS version 8.0J (SPSS Inc., Chicago, IL). All values are reported as the mean ± SD, and all P  values < 0.05 were considered statistically significant.

Control membrane properties of 163 respiratory neurons are presented in table 1. The values were comparable to those of previous studies. 13,14 Table 2summarizes effects of opioids on the various neuron groups in standard artificial cerebrospinal fluid and tetrodotoxin-containing solution.

Application of morphine, DAGO, and U50488 resulted in a synchronous decrease in the burst rates of all classes of inspiratory neuron and of the C4 activity (table 3, figs. 2 and 3). The depressant effects by these three opioid receptor agonists were antagonized by 1 μm naloxone. However, inspiratory neuron intraburst firing frequency did not significantly change during opioid-induced respiratory depression (17.0 ± 6.8 to 16.8 ± 6.8 Hz).

In four (27%) of 15 inspiratory neurons, the decrease in burst rate was associated with a hyperpolarizing shift in resting E^mby, on average, 5.5 ± 2.9 mV (range, 3–8 mV) and a reduction in R^mby, on average, 77.5 ± 39.7 MΩ (i.e. , 24%; range, 36–121 MΩ) after application of morphine (fig. 2). Resting E^mand R^min the remaining 11 inspiratory neurons did not change despite a decrease in C4 and inspiratory neuron burst rates. To further study the membrane properties of the hyperpolarized neurons, we analyzed the current–voltage relations of steady state E^mresponses to injection of DC pulses. In the example shown in fig. 2C, the control current–voltage relation and that obtained during morphine application intersected at −70 mV. This E^mvalue represents the reversal potential (E^rev) of the opioid-induced hyperpolarization, which averaged −75.1 ± 18.6 mV (n = 4).

DAGO resulted in a hyperpolarization by, on average, 5.0 ± 2.5 mV (range, 2–9 mV) and a decrease in R^mby, on average, 78.3 ± 33.2 MΩ (i.e. , 20%; range, 25–130 MΩ), whereas E^revaveraged −76.2 ± 12.1 mV in 11 (50%) of 22 inspiratory neurons (fig. 3A). The remaining 11 neurons were unchanged for resting E^mand R^m.

Within 15 min after application of U50488, steady state hyperpolarization of down to 5.7 ± 1.7 mV (range, 2–7 mV) occurred, and R^mdecreased by, on average, 65.2 ± 32.2 MΩ (i.e. , 19%; range, 21–125 MΩ) in nine (45%) of 20 inspiratory neurons (fig. 3B). E^revaveraged −83.0 ± 15.5 mV (n = 9). The remaining 11 inspiratory neurons did not change resting E^mand R^m.

In all three classes of inspiratory neurons, the δ-opioid receptor agonist DPDPE did not cause any effects on burst rate, E^m, and R^m(n = 9).

As shown in fig. 3, the type I and III inspiratory neurons showed a high probability of EPSPs and IPSPs, respectively, during the preinspiratory and postinspiratory phases. EPSPs and IPSPs, as indicated by arrows in figure 3, were still observed periodically despite a decrease in the burst rates of C4 and inspiratory neuron during opioid-induced respiratory depression. Summarizing the effects of all three agonists (morphine, DAGO, and U50488), approximately 42% inspiratory neurons responded with membrane hyperpolarization and R^mdecrease during application of these agonists (table 2).

Opioid receptor agonists did not affect E^mand R^min preinspiratory neurons (n = 15: 5 DAGO, 4 morphine, 4 U50488, 2 DPDPE). As shown in figure 4, the preinspiratory neurons were discharging at unchanged rate while the C4 respiratory rate decreased markedly after application of DAGO (table 4). Similar effects were also caused by morphine or U50488 (table 4). Intraburst firing frequency of preinspiratory neurons did not significantly change during opioid-induced respiratory depression (12.4 ± 3.3 to 12.1 ± 2.9 Hz).

In the 12 expiratory neurons (4 DAGO, 3 morphine, 3 U50488, 2 DPDPE), opioid receptor agonists did not affect action potentials, E^m, and R^m(fig. 5A). Intraburst firing frequency of these expiratory neurons did not significantly change during opioid-induced respiratory depression (8.3 ± 4.0 to 8.5 ± 3.4 Hz). However, we observed that U50488 hyperpolarized the membrane by 3–5 mV and decreased intraburst firing frequency from 5.3 ± 1.1 to 1.3 ± 1.1 Hz in a naloxone-reversible manner in two (40%) of five expiratory neurons (fig. 5Band table 2).

The stimulation induced orthodromic responses in the 15 of 17 neurons, whereas two neurons were antidromically activated (6 DAGO, 4 morphine, 4 U50488, 3 DPDPE). The latencies of orthodromic EPSPs in the individual neurons were determined from responses to three trials, and the SD indicates variance of the latencies less than 0.6 ms. The average latencies of the first EPSPs (9.2 ± 0.9 ms, n = 15) closely fitted with antidromic latencies (9.4 ± 0.2 ms, n = 2).

After application of opioid receptor agonists, latencies of EPSP responses did not significantly change (table 5), whether neurons examined were hyperpolarized or not. In contrast, the slope in the rising phase of the first EPSPs, as indicated by arrowheads in figure 6, was reduced during opioid-induced respiratory depression (n = 12: 4 DAGO, 4 morphine, 4 U50488). Therefore, latencies of first action potential responses were significantly prolonged after application of μ- and κ-opioid receptors (table 5). Seven (58%) of these 12 inspiratory neurons did not change the resting E^mand R^m(fig. 6A). The other five neurons were hyperpolarized by 4–7 mV, and R^mdecreased by 36–130 MΩ during opioid-induced respiratory depression (fig. 6B).

After treatment with DAGO (n = 4), morphine (n = 3), and U50488 (n = 3), the inspiratory neuron burst rate and the C4 respiratory rate decreased, while the preinspiratory neuron was still discharging at the same rate as in the control state (fig. 7). Furthermore, we confirmed that EPSPs and IPSPs, which corresponded to the period of extracellular preinspiratory neuron burst, were still observed in type I and III inspiratory neurons during opioid-induced respiratory depression (fig. 7).

After the application of DAGO, 7 (39%) of 18 inspiratory neurons hyperpolarized by, on average, 3.4 ± 1.6 mV (range, 2–6 mV) and decreased R^mby, on average, 68.0 ± 46.9 MΩ (i.e. , 27%; range, 24–151 MΩ). E^revaveraged −65.5 ± 4.6 mV (n = 7;fig. 8). Application of U50488 resulted in a hyperpolarization by, on average, 5.6 ± 2.8 mV (range, 2–11 mV), and R^mdecreased by, on average, 46.6 ± 27.5 MΩ (i.e. , 20%; range, 22–109 MΩ), while E^revaveraged −83.2 ± 10.8 mV in 9 (56%) of 16 inspiratory neurons. Morphine also hyperpolarized by, on average, 5.3 ± 3.9 mV (range, 3–11 mV) and decreased R^mby, on average, 51.0 ± 34.9 MΩ (i.e. , 18%; range, 17–84 MΩ), while E^revaveraged −82.6 ± 21.1 mV in four (40%) of 10 inspiratory neurons. This means that a membrane hyperpolarization and reduction of R^mwere recorded in 20 (45%) of 44 inspiratory neurons during application of these three agonists. In inspiratory neurons, DPDPE did not cause any effects on E^mand R^m. The opioid receptor agonists with tetrodotoxin did not affect E^mand R^min preinspiratory and expiratory neurons.

The major findings in this investigation were that both μ- and κ-opioid receptor agonists inhibited inspiratory neuron bursts, whereas these agonists were less effective on the preinspiratory and expiratory neurons, as well as on the synaptic transmission from preinspiratory neurons to inspiratory neurons. Furthermore, this inhibition of inspiratory neurons seems to be caused by both a presynaptic and postsynaptic inhibition. Thus, opioids in used concentration caused reduction of final motor outputs by mainly inhibiting medullary inspiratory neuron network, exerting only moderate effects on the central respiratory rhythm produced by preinspiratory neurons. However, it should be carefully considered which neuron activity is a good indicator of the central rhythm generator. The current results are consistent with a hypothesis that the preinspiratory neuron network is a primary rhythm generator for respiration in the neonatal in vitro  preparation. Therefore, measurement of central burst frequency could be practically performed by measuring preinspiratory burst rate in the current study.

In the current study, central mechanisms of respiratory depression induced by opioid receptor agonists were investigated using a brainstem–spinal cord preparation from newborn rats. Although this reduced preparation is advantageous for the purpose, extrapolation of the current results to the mature and intact animal should be made with a high degree of caution.

The current results showing an inhibitory effect of opioid receptor agonists on the C4 respiratory frequency were consistent with those of previous studies. 6,19In detail, μ- and κ-opioid receptors were involved in the reduction of C4 and inspiratory neuron burst frequency, whereas δ-opioid receptor did not participate in respiratory depression. Although drugs were applied to the isolated brainstem preparation in a perfusion chamber, concomitant inhibition of inspiratory neuron discharge in the medulla and C4 inspiratory activity indicates that such inhibitory effects on respiratory outputs originated from the effects of drugs in the medulla. This is further supported by the lack of effects on preinspiratory burst rate. In rats, μ- and κ-opioid receptors are present at birth and have a similar distribution in the brainstem, whereas δ-opioid receptors appear during the postnatal period. 20–22These results from previous histologic studies may explain pharmacologic findings that activation of δ-opioid receptor in the neonatal preparations did not always produce consistent inhibitory effects. 4–6It is possible that δ-opioid receptors in the medulla oblongata may participate in respiratory depression produced by opioids in older rats. 23,24 

Using the neonatal rat brainstem–spinal cord preparation, Greer et al.  4reported that DAGO (0.2–1.0 μm) decreased in the C4 respiratory frequency, whereas the respiratory frequency was unaffected by U50488 (0.1–4.0 μm) and DPDPE (0.1–2.0 μm). However, they measured respiratory frequency only 5 min after drug application. It was also shown that inhibitory effects via κ-opioid receptor could be observed with a slower time course than those mediated via μ-opioid receptor. 6In the current study, we showed that the effects of U50488 usually appeared within 15 min after the application. Several in vivo  studies have shown that κ-opioid receptors do not participate in the opioid-induced respiratory depression. 4,24,25This discrepancy between the results of the current in vitro  study and these in vivo  studies suggests that the suprabulbar structures may modify the medullary κ-opioid receptor-mediated respiratory depression.

Concerning action of endogenous opioids, we did not test the sole administration of naloxone in the current study. Previous studies have shown that naloxone has no effect on medullary respiratory control during similar experimental conditions as in the current study, and that endogenous opioids do not participate in basal respiratory control in the medulla oblongata of newborn rats. 4,6 

Forty-two percent of the inspiratory neurons were hyperpolarized and decreased in R^mduring opioid-induced respiratory depression. In addition, a similar percentage of inspiratory neurons were also hyperpolarized and decreased in R^mduring application of μ- and κ-opioid receptor agonists after blockade of ongoing synaptic activity by tetrodotoxin. This type of neuron is thought to be inhibited by postsynaptic action of opioids. Results from the current–voltage relation of the neuron and the value of the E^revare consistent with a suggestion that opioid-induced hyperpolarization is induced by an increase in the K⁺conductance. These results are consistent with those of rhythmically active slices from mice in which a subpopulation of inspiratory neurons was hyperpolarized via  direct postsynaptic action accompanying an increase in K⁺conductance during application of μ-opioid receptor agonist. 12Interestingly, half of inspiratory neurons inhibited with μ- and κ-opioid receptor agonists remained unchanged for resting E^mand R^m. These inhibitory actions on neuron burst discharges may be induced by a decrease in excitatory drives from presynaptic neurons. Opioids in presynaptic sites inhibit Ca²⁺-dependent neurotransmitter release along with K⁺channel activation and inhibitory actions on voltage-dependent Ca²⁺channels. 26,27It has been suggested that the main action of opioids is a presynaptic inhibition of the excitatory synaptic input. 28Intermediate- and high-voltage–activated Ca²⁺channels may contribute to the potentiation of burst activity in respiratory neurons. 10In addition to K⁺conductance increase, therefore, we presume that the insufficient activation of Ca²⁺channels at the presynaptic and postsynaptic membrane may be involved in the inhibition of inspiratory neuron bursts by opioids.

The slope of the first EPSPs, evoked by stimulation to the contralateral medulla, was reduced in inspiratory neurons during opioid-induced respiratory depression. Fifty-eight percent (7 of 12) of these inspiratory neurons did not show any significant change in the resting E^mand R^m. We hypothesized that the first EPSPs were induced monosynaptically because of the steady latency, which corresponds well to antidromic latency and significant small variance of latencies of EPSPs. Based on this assumption, the results indicate that opioids exert inhibitory effects at presynaptic sites on these inspiratory neurons, as suggested in other neuronal systems. 26,27Even in hyperpolarized neurons (42%), the slope of the EPSPs was reduced, which suggests the presence of neurons that simultaneously show both presynaptic and postsynaptic inhibitory effects. The reduction of slope of EPSPs causes delay of initiation of action potentials and, consequently, may elicit failure of synaptic transmission at polysynaptic pathways in the inspiratory neuron network. In addition to the postsynaptic inhibition, the presynaptic inhibition caused by opioids may effectively depress inspiratory burst generation by mutual excitatory couplings between inspiratory neurons. 29 

Ballanyi et al.  5reported that opioid receptor agonists blocked respiration-related E^mfluctuations in a cyclic adenosine monophosphate–dependent manner. Their data also indicated a possibility that the neuronal responses to opioid receptor agonists differ depending on subtypes of respiratory neurons. Type I and III inspiratory neurons receive excitatory and inhibitory synaptic inputs, respectively, from preinspiratory neurons. 10,11These synaptic potentials, which correlated to preinspiratory neuron burst, remained during opioid-induced respiratory depression, as confirmed by simultaneous recordings from inspiratory and preinspiratory neurons. Thus, we conclude that μ- and κ-opioid receptor agonists inhibited inspiratory neuron burst via  direct or indirect synaptic mechanism. Synaptic inputs from preinspiratory neurons in the neonatal brainstem–spinal cord preparation were, however, unchanged, indicating a lack of effect of these agonists on central respiratory rhythm generation.

We observed that the burst activities of preinspiratory neurons were still present even with the disappearance of inspiratory neuron bursts and C4 respiratory discharges. In addition, E^mand R^mdid not change during administration of opioid receptor agonists with or without tetrodotoxin perfusion. Such resistance of preinspiratory neurons to opioids does not mean that inspiratory neuron burst and C4 inspiratory activity occur out of phase with the preinspiratory neuron burst during opioid-induced respiratory depression. Rather, we noted that during opioid-induced respiratory depression, the firing of a preinspiratory neuron always preceded the initiation even of slow inspiratory neuron bursts.

In the previous studies using the brainstem–spinal cord preparation from neonatal rats, it was suggested that bath-applied neuromodulators induced change in respiratory rhythm through primarily altering preinspiratory neuron activity. 8,11,30In contrast, inhibition of respiratory rhythm (i.e. , reduction of frequency of the C4 motoneuron outputs) by opioids was caused by inhibition of the inspiratory neuron network but not attributable to effects on the preinspiratory neuron network in the medulla. Hence, inspiratory neurons (i.e. , output neurons) are affected either directly or via  drive reduction without any significant effects on the central rhythm generated by preinspiratory neurons.

Most expiratory neurons showed unchanged discharging action potentials, E^m, and R^mduring application of opioid receptor agonists. However, the κ-opioid receptor agonist U50488 with standard artificial cerebrospinal fluid induced a naloxone-reversible hyperpolarization and decrease of spike firing frequency in two expiratory neurons. This response may be a result of an activation of K⁺conductance via  opioid receptors present on expiratory neurons. Expiratory neurons in the RVLM of neonatal rats express three different K⁺channels. One of these, which may be expressed specifically in the expiratory neuron, is voltage-dependent and determines the neuronal level of depolarization. 31Although most expiratory neurons are thought to be basically insensitive to opioids, some expiratory neurons have opioid receptors and may be influenced by opioids. 5 

In conclusion, respiratory depression caused by μ- and κ-opioid receptor agonists in the brainstem–spinal cord preparation is caused by a combined and reversible presynaptic and postsynaptic inhibition of inspiratory neurons, while membrane properties of preinspiratory and most expiratory neurons are left unaffected. Thus, opioids exerted the inhibitory effects mainly on the inspiratory neuron network in the medulla and only minor effects on the central respiratory rhythm by preinspiratory neurons.