C.D. Barnes and 0. Pompeiano (Ed$.)
Progress in Brain Research, Vol. 88
197
0 1991 Elsevier Science Publishers B.V.
CHAPTER 14
Mechanisms of opioid actions on neurons of the locus coeruleus M.J. Christie Department of Pharmacology, University of Sydney, N.S. W., Australia
The locus coeruleus (LC) has provided a useful model for pioneering studies of the mechanisms underlying the acute and chronic actions of opioid drugs. Acutely, opioids inhibit the electrical activity of single neurons in the rat and guinea pig LC. Inhibition is due to a membrane hyperpolarisation. In these cells, opioids act on y-receptors to increase the opening of inwardly rectifying potassium channels, thus leading to hyperpolarisation. The y-receptors are coupled to potassium channels via G-proteins which are sensitive to inactivation by
pertussis toxin. This coupling process is quite direct, in that it does not involve freely diffusible intracellular second messengers. Agonists specific for other receptors, such as aZ-and somatostatin-receptors, are capable of opening the same population of potassium channels on LC neurons. Following chronic treatment of animals with morphine, a specific deficit develops in the ability of p-receptors to open potassium channels, producing reduced sensitivity of LC neurons to inhibition by opioids.
Key words: locus coeruleus, receptor opioid, potassium channel, tolerance, G-protein, chronic drug
Introduction Opioids directly inhibit the electrical activity of neurons in many regions of the nervous system, including the locus coeruleus (LC) (see Duggan and North, 1983, for review). This inhibition arises from changes in ionic conductances in the somatic membrane or nerve terminal, leading ultimately to a reduced firing of action potentials and depressed release of neurotransmitter. The ionic mechanisms underlying opioid actions were first elucidated in the LC (Williams et al., 19821, and the molecular mechanisms involved have been more thoroughly studied in these neurons than in other regions of the CNS. It is now clear that
opioids inhibit LC neurons by activating p-receptors, which couple to the opening of potassium channels via G-proteins (North et aL, 1987). Within the LC neuron, the coupling process between receptor and potassium channel is localised to a membrane area of less than a few square microns, and does not involve freely diffusible second messengers (Miyake et al., 1989). No action of endogenously released opioids has yet been clearly identified in the LC, so the physiological role in the whole animal of this action on the LC soma is still unclear. Much of the impetus to understand the acute actions of opioids on CNS neurons has arisen from the problems of tolerance and physical de-
198
pendence. The LC has been a focus of much of this work, partly because of the similarity between the behavioural effects of electrical stimulation of the LC and the opiate withdrawal syndrome. While the mechanisms underlying dependence remain elusive, toIerance involves a specific deficit in the ability of opioid receptors to open potassium channels (Christie et al., 1987). Acute actions of opioids Opioids directly inhibit LC neurons Opioids inhibit the firing of action potentials when applied to neurons of many regions of the CNS (see Duggan and North, 1983, for review). In most cases where opioids excite neurons, it has been shown to occur as a result of the disinhibition of inhibitory interneurons (e.g., Madison and Nicoll, 1987). The first in vivo studies demonstrated inhibition of sensory-evoked action potential firing following intravenous administration of morphine (Korf et at., 1974). Inhibition was specific in that it was reversed by the opioid antagonist, naloxone. Similar results were reported in cat LC (Strahlendorf et al., 1980). Inhibition of spontaneous firing was also shown in rat LC in viuo following intravenous or localised (iontophoretic) application of opioids (Bird and Kuhar, 1977; Guyenet and Aghajanian, 1977). This observation has been substantiated in uiuo and in ilitro for species in which LC neurons can be identified as a fairly uniform population of noradrenaline-containing cells (see below). Opioid receptor type on LC neurons Opioids are known to affect cells via activation of at least three distinct membrane receptors; p-, 6- and K-receptors (see Leslie, 1987, for review). With the advent of brain slice techniques to study CNS neurons in vitro, the opportunity arose to classify the opioid receptor(s1 responsible for inhibition in LC neurons. Quantitative pharmacological techniques have demonstrated that inhibition is due to activation of p-receptors in the rat
(Williams and North, 1984). In addition to the actions of non-selective agonists such as methionine-enkephalin (met-enkephalin) and D-Ala-DLeu-enkephalin (DADLE), LC neurons were directly inhibited by p-agonists such as morphine and D-Ala-methionine-enkephalin-glyol(DAGO). However, the 6-agonist, D-Pen-D-Pen-enkephalin (DPDPE) (North et al., 1987) and the K-agonist, U50488H (Williams and North, 1984) were ineffective in rat LC, even at very high concentrations, suggesting that 6- and K-receptors were not responsible for inhibition. Quantification of the equilibrium dissociation constant (K,) of an antagonist is the most reliable method to define receptor type. This is because, unlike agonist actions, it does not depend on efficiency of coupling between receptor and effector. When determined by Schild analysis, the K , of naloxone to antagonize the action of met-enkephalin was 2 nM (Williams and North, 1984). This result was very similar to the value obtained for p-receptors in biochemical assays and differed by 10 to 30-fold from the value expected €or 6-or K-receptors. Moreover, the &receptor antagonist, ICI174864, blocked the effects of met-enkephalin very poorly ( K , > 5 pM).Thus, it is clear that inhibition of LC neurons by opioids is due solely to the activation of p-receptors. p-Receptors act cia G-proteins Although direct information on the primary aminoacid structure of the p-receptor is still lacking, there is little doubt that it belongs to the family of receptors which exert their cellular effects via GTP-binding proteins (G-proteins, see Birnbaumer, 1990, for review). Evidence that p receptors act through G-proteins includes inhibition of adenylate-cyclase in various preparations including LC (Duman et al., 1988), co-purification of p-binding sites with pertussis toxin-sensitive G-proteins (Wong et al., 1989) and the ability of purified p-receptors to reconstitute with purified G, and Go (Ueda et al., 1988). In LC, treatment of rats with pertussis toxin, which inactivates &-subunits of G, and Go, prevents inhibi-
199
tion of LC neurons by opioids (Aghajanian and Wang, 1986).
p-Receptor distribution on LC neurons Opioid receptors are distributed over the plasma membrane and nerve terminal regions of LC neurons. Physiological studies with patchclamp electrodes (see below) suggest a fairly uniform distribution of receptors over the perikaryal membrane of isolated LC neurons, although this result could have been influenced by enzymatic and mechanical procedures used to dissociate the LC. Some studies have also suggested that pagonists decrease the excitability of LC axon terminals. p-agonists inhibit noradrenaline release from brain slices of LC projection areas (Werling et al., 1987) and increase the threshold for antidromic activation of LC fibers when infused into cerebral cortex in civo (Nakamura et al., 1982). However, because these studies could not be conducted in the absence of synaptic transmission, they did not rule out indirect effects of p-agonists on the LC terminal membrane. p-Receptors activate potassium channels Techniques for intracellular recording from brain slices have also facilitated the elucidation of ionic mechanisms underlying opioid inhibition. The first intracellular recordings from LC neurons demonstrated inhibition to be due to a membrane hyperpolarization in guinea-pig (Pepper and Henderson, 1980). The hyperpolarization was associated with a fall in input resistance, and persisted under conditions in which synaptic transmission was blocked, indicating that opioids act directly on the impaled neuron. Williams et al. (1982) demonstrated that the hyperpolarization was due to an increased conductance to potassium ions in rat LC neurons. This was due to opening of membrane potassium channels, shifting the membrane potential from rest (about - 55 mV) towards the potassium equilibrium potential (about - 110 mV>. The macroscopic properties of the potassium
conductance were similar to those of a class of potassium channels known as inward rectifiers, first characterised in invertebrate egg cells (for review see Hille, 1984). These channels are so named because they do not open until the membrane potential approaches the potassium equilibrium potential, thus allowing potassium currents to flow more readily in the inward than outward direction. The mid-point of the conductance activated by opioids in LC neurons (about - 60 mV) is close to action potential threshold (about -55 mV), so these channels are capable of hyperpolarizing the membrane (North and Williams, 1985; Williams et al., 1988). Opioid-activated potassium channels have been demonstrated directly in acutely dissociated rat LC neurons (Miyake et at., 1989). When patchclamp pipettes were sealed onto the cell surface, single potassium channels were observed when opioids were included in the pipette solution, but not when opioids were applied to the rest of the cell. The probability of channel opening was dependent on agonist concentration and was antagonized when naloxone was included in the pipette. The channels had a unit conductance of about 45 pS in isotonic potassium and displayed bursts of activity lasting several seconds to minutes, separated by similar closed periods. A thorough analysis of channel kinetics was precluded by these long, closed periods, but within the periods of activity there may be two closed states and one open state of the channel. The open state is prolonged when higher concentrations of opioids are tested. All agonist-filled pipettes of about the same size contained similar numbers of channels, suggesting a fairly uniform distribution over the perikaryal membrane. The number of opioidactivated channels per cell was estimated to be about 1,000, given the diameter of recording pipettes and the total surface area of the LC neuron. This result suggests that p-receptors must also be fairly uniformly distributed over the soma, because they are within the few square microns of membrane contained within the pipette.
200
Mechanism of coupling between opioid receptors and potassium channels The main inference drawn from the experiments described above was that the coupling between opioid receptor activation and potassium channel opening is localized to the patch of membrane enclosed by the patch pipette, therefore ruling out the involvement of freely diffusible second messengers. Although second messenger systems such as the adenylate cyclase to protein kinase-A cascade are inhibited by opioids in LC (see below), these are not involved in the activation of potassium channels. One report, which suggested that agents which activate protein kinase-A reversed the hyperpolarization produced by morphine (Andrade and Aghajanian, 1985) was subsequently shown to result from activation of a small inward current by these agents (Wang and Aghajanian, 1987). Activators of protein kinase-A failed to prevent opioid hyperpolarizations, or potassium currents in other experiments (North and Williams, 1985). Beyond the lack of involvement of freely diffusible second messengers, and an involvement of G-proteins sensitive to pertussis toxin, the nature of the coupling between preceptor and potassium channel is unknown. It is possible that activated G-proteins bind directly to the channel, or that biochemical events confined to a region in close proximity ( < 1 pm) to the receptor and channel mediate the effects. These issues also need to be resolved for a variety of receptors which couple to potassium channels in LC and other cells. Multiple receptors open a single population of potassium channels Agonists selective for several receptors have been demonstrated to hyperpolarize, increase a potassium conductance and/or open potassium channels on LC neurons. These include agonists selective for adrenergic az- (North and Williams, 1983, somatostatin- (Inoue et al., 1988) and GABA,- (Osmanovic and Shefner, 1988) receptors. Adenosine receptor agonists also hyperpo-
larize LC neurons, possibly by activating a potassium conductance (Shefner and Chiu, 1986). This list is not exhaustive, agonists selective for some other receptors inhibit LC (see Olpe and Steinmann, this volume), and others might not have been tested. For the only combination tested, different receptors have been shown to activate the same population of potassium channels as do p-receptors. This conclusion has been inferred from simultaneous application of different agonists to LC neurons, e.g., if sufficient p-agonist is applied to an LC neuron to fully activate the potassium conductance, a,-agonists cannot activate the conductance any further, even though an a,-agonist applied alone can activate the conductance to the same extent as the p-agonist (North and Williams, 1985). This observation held for any combinations of p - and a,-agonists. The location in the sequence of biochemical events where different receptors converge to open one population of potassium channels is not known. It might be that each receptor can activate a different population of G-proteins which can all open the same population of channels, or in the extreme alternative, convergence to a single population of G-proteins may occur. In either case, it is not surprising that different receptors can open the same population of potassium channels on LC neurons. The same is true for p- (Christie and North, 1988), and other receptors in a variety of cells (see North, 1989, for review). These receptors generally belong to a family of G-protein-coupled receptors which have also been shown to inhibit adenylate cyclase (see Birnbaumer, 1990, for review). Like in the LC, where studied in detail, this coupling has been shown to involve pertussis toxin-sensitive G-proteins, but not freely diffusable second messengers such as CAMP.
Do opioid receptors influence other ion channels on LC neurons? Not only do multiple G-protein-coupled receptors converge onto single effectors, single receptors are able to influence multiple effectors in
20 1
single cells via G-proteins (see Birnbaumer, 1990, for review). In addition to opening potassium channels, opioid receptors have been shown to influence the activity of voltage-dependent calcium channels in several cell types. Both 6- and K-receptors inhibit calcium currents in cultured cells (Macdonald and Werz, 1986; Hescheler et al., 1987), and p-receptors inhibit calcium currents in a human neuroblastoma cell-line (E. Seward and G. Henderson, personal communication). Although p-agonists inhibited the calcium component of the action potential in LC neurons, this was shown to be secondary to potassium current activation (North and Williams, 1983). When studied directly, voltage-dependent calcium currents were unaffected in LC neurons by high concentrations of the non-selective opioid, met-enkephalin (Williams, Christie and North, unpublished observations). In the same experiments, muscarinic receptor agonists inhibited the current. To date, the electrophysiological actions of opioids on the soma of LC neurons seem to be confined to the opening of potassium channels.
Presynaptic actions of opioids Opioids have been shown to inhibit synaptic potentials impinging on neurons throughout the nervous system. In the LC, studies of opioid actions on synaptic potentials are complicated by the post-synaptic actions of p-opioids, i.e. , the large post-synaptic increase in potassium conductance affects both the amplitude and time course of synaptic potentials (see Williams et at., this volume). In contrast, although K-agonists have no post-synaptic effects, they have been reported to produce a 38% inhibition of the amplitude of fast synaptic potentials which were evoked by electrical stimulation in rat brain slices (McFadzean et al., 1987). The fast synaptic potential is due to the release of at least three neurotransmitters: an excitatory aminoacid, GABA and glycine (see Williams et al., this volume). It is unknown which of these are inhibited by K-opioids, and which projections to LC are responsible.
Physiological role of opioids in LC The physiological role of p-receptor-mediated inhibition of LC neurons has yet to be established. Fibers and terminals containing products of pro-opiomelanocortin, pro-dynorphin and pro-enkephalin are present in the vicinity of the LC (see Sutin and Jacobowitz, this volume) and may make synaptic contacts with LC neurons (Pickel et al., 1979). However, attempts to detect a post-synaptic response to endogenous opioids released by electrical or high potassium stimulation within slices of LC met with no success, even in the presence of peptidase inhibitors (Williams et al., 1987). This result is not surprising, an endogenous opioid-mediated synaptic response has yet to be reported on any neuron. A possible explanation is that unusual stimulus parameters are needed to evoke release of endogenous opioids and other peptides, such as somatostatin, within the LC. It is also possible that endogenous opioids act on LC neurons in a diffuse, non-synaptic, manner. One report of a naloxone-reversible inhibition of LC firing following stimulation of the pro-opiomelanocortin containing arcuate nucleus (Strahlendorf et al., 1980) has yet to be shown to be a direct effect on LC using intracellular recording. Opioid tolerance and physical dependence in LC The notion that tolerance to opioids, i.e., reduced responsiveness to an agonist following chronic exposure to it, involves cellular adaptive processes within the neuron that bears opioid receptors is generally accepted (see Koob and Bloom, 1988, for review). The mechanisms underlying physical dependence are less clearly understood than those underlying tolerance. Physical dependence is characterised in whole animals by withdrawal signs upon removal of agonist, or administration of antagonist. Withdrawal signs are often qualitatively opposite to the acute agonist actions, e.g., hyperalgesia is a withdrawal sign opposite to the acute analgesic action of opioids. At the level of single neurons, signs of dependence would be
202
considered as the occurrence of enhanced spontaneous activity or heightened excitability. Tolerance and physical dependence following chronic exposure to opioids have been extensively investigated in the LC. This interest arose partly from the simiIarity between the behavioral effects of electrical stimulation of the LC and the opiate withdrawal syndrome. Indeed, this similarity provided the rationale for the use of clonidine in the treatment of opiate withdrawal in man (Gold et al., 1978). Clonidine, and other a,-adrenoceptor agonists, by acting through the same cellular pathway as opioids inhibits LC neurons and thereby reduces withdrawal excitation. Tolerance The electrophysiological mechanisms underlying tolerance have been studied in any detail only in LC neurons (Andrade et al., 1983; Christie et al., 1987). Tolerance was observed as a reduction in the sensitivity of LC neurons to agonists such as met-enkephalin and normorphine. The maximum potassium current induced by normorphine was also reduced (Christie et al., 1987). This reduced sensitivity to opioids was observed for many hours after removal of tissue from animals and hence withdrawal of morphine. These effects were indistinguishable from inactivation of a fraction of the p-receptors on the normal cell surface with the irreversible antagonist, p-chlornaltrexamine. It was, therefore, concluded that tolerance was equivalent to inactivation of p-receptors, This interpretation is functionally equivalent to a decrease in the ability of individual p-receptors to open channels. This conclusion was similar to those drawn from isolated organ experiments (Chavkin and Goldstein, 1984). No effects were found on the K , of naloxone, the potassium conductance itself, or the ability of aZadrenoceptor agonists to open the same population of potassium channels (Christie el al., 1987). No signs indicative of physical dependence (withdrawal, or antagonist-induced excitation) were observed (Andrade et al., 1983; Christie et al., 1987).
These studies localized tolerance specifically to coupling between p-receptors and potassium channels, because the channels themselves and the ability of other receptors to couple to them were unaffected. Tolerance must therefore be due to a reduced preceptor density on the cell surface, or reduced ability of receptors to couple to G-proteins. These two possibilities could not be distinguished in physiological studies. Examples of both possibilities have been observed in biochemical studies of p- and &receptors in cultured cells (Law et al., 1983; Werling et al., 1989) and brain membrane preparations (Rogers and El-Fakahany, 1986). Beyond this level of analysis, little is known of the mechanisms underlying opioid tolerance. Changes occur with chronic opioids in the adenylate-cyclase to protein kinase-A cascade (Duman et al., 1988; Nestler and Tallman, 1988) and Gprotein ribosylation in LC (Nestler et al., 1989); however, the physiological relationship of these adaptations to tolerance is unknown. An observation which could be related to tolerance development is that p-receptors were desensitised for 15-20 min after occupation by high concentrations of agonists (G.C. Harris and J.T. Williams, personal communication). This acute desensitisation is analogous to homologous desensitisation of other G-protein coupled receptors, e.g., padrenoceptors are phosphorylated and inactivated by p-adrenergic receptor-kinase only when occupied by an agonist (Benovic et al., 1989). Whether similar mechanisms play a role in the development of the long-term tolerance described above is unknown. Physical dependence Dependence (increase in extracellular action potential frequency) was first reported in LC neurons in vivo after iontophoretic application of the opioid antagonist, naloxone (Aghajanian, 1978). These results were replicated using intracerebroventricular application of an antagonist, naltrexone (Valentino and Wehby, 1989). However, these observations were not replicated using
203
extracellular (Andrade et af., 1983) or intracellular (Christie et af., 1987) recording in vitro, upon withdrawal of opioids, or upon application of naloxone. There are several possible explanations for this discrepancy. First, increased firing frequency in ciuo might simply be due to reversal of the effects of the agonist (morphine) which was still present in the tissue. However, the naloxoneinduced frequency was much greater than that observed in naive animals (Aghajanian, 1978; Valentino and Wehby, 1989), ruling out this possibility. Second, if dependence in LC dissipates very rapidly, it may have declined during the preparation of brain slices and impalement of neurons (usually longer than 30 min in the absence of an agonist). However, maintenance of tissue in morphine (1 FM) during brain slice preparation did not lead to the observation of withdrawal excitation upon superfusion of naloxone (J.T. Williams, personal communication), ruling out this possibility. Finally, increased frequency could have been due to altered synaptic activity, which would bc disrupted in brain slices. This would suggest that dependence develops in nerve terminals and/or cells projecting to the LC. This possibility has not been explored in vitro, but lesions to nucleus paragigantocellularis, which provides the main excitatory afferents to the LC, attenuated withdrawal excitation in uiuo (Rasmussen and Aghajanian, 1989). These studies imply that tolerance and dependence can be dissociated at the neuronal level and are, therefore, consequences of different cellular mechanisms. This interpretation was consistent with similar observations reported for isolated organ preparations (Wuster et al., 1982), contrary to earlier theories of opioid tolerance and dependence, which posited a unitary mechanism for both phenomena (Collier, 1980). This interpretation might imply that tolerance occurs in many types of opioid sensitive neuron, but dependence develops only in a subset. Some studies support this possibility. Cells in mouse dorsal root ganglion explants develop hyper-excitability following chronic treatment with morphine (Crain
et a/., 1988). Another possibility is that dependence requires intact neural networks or requires some form of synaptic interactions. This might explain the discrepancy between the observation of dependence in LC neurons in uivo and its absence in uitro. Conclusions The LC has contributed greatly to our understanding of the acute and chronic actions of opioids. Within the LC opioids have direct inhibitory effects on the soma, as well as presynaptic inhibitory effects. It is also likely that opioids have direct inhibitory effects on the terminals of LC neurons. The somatic effects are due to opening of inwardly rectifying potassium channels. p-receptors, as well as other membrane receptors, couple to these channels via the activation of pertussis toxin-sensitive G-proteids). Diffusible second messengers, such as CAMP, are not involved in the coupling process between preceptor and potassium channel. The physiological role of inhibition of adenylate cyclase by preceptors in LC is still unknown, but might involve long-term processes such as gene regulation. This understanding of the mechanisms of acute actions has facilitated analysis of the chronic effects of opioids in LC. Tolerance to opioids in LC in uitro is due either to a reduced number of preceptors on the soma, or to a reduced efficiency of coupling between preceptors and potassium channels. This deficit is specific in that a,-adrenoceptor agonists can still fully activate the same population of potassium channels. Signs of opioid withdrawal have been observed in LC neurons in uiuo, but not in brain slices in citro. The reasons for this discrepancy are unclear. The observation of tolerance, but not physical dependence, in LC neurons in uitro suggests the mechanisms underlying the two phenomena differ. The presynaptic actions of opioids are less clearly understood than the post-synaptic effects. K-receptor agonists depress fast synaptic potentials evoked in the LC by electrical stimulation.
204
The neurotransmitters involved and anatomical source of these synaptic potentials have yet to be described. A clear demonstration of synaptic responses to endogenously released opioids is lacking in the LC, as it is elsewhere in the nervous system. Acknowledgements Drs. J.T. Williams and S.M. Johnson are gratefully acknowledged for their comments and suggestions. References Aghajanian, G.K. (1978) Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature (London), 276: 86-188. Aghajanian, G.K. and Wang, Y.Y. (1986) Pertussis toxin blocks the outward currents evoked by opiate and a, agonists in locus coeruleus neurons. Brain Res., 371: 390-394. Andrade, R. and Aghajanian, G.K. (1985) Opiate- and alpha,-adrenoceptor-induced hyperpolarizations of locus coeruleus neurons in brain slices: Reversal by cyclic adenosine 3’,5’-monophosphate analogues. J. Neurosci., 5: 2359-2364. Andrade, R., VanderMaelen, C.P. and Aghajanian, G.K. (1983) Morphine tolerance and dependence in locus coeruleus: Single cell studies in brain slices. Eur J. Pharmacol., 91: 161-169. Benovic, J.L., DeBlasi, A., Stone, W.C., Caron, M.G. and Lefkowitz, R.J. (1989) P-adrenergic receptor kinase: Primary structure delineates a multigene family. Science, 246: 235-240. Bird, S.J. and Kuhar, M.J. (1977) Iontophoretic application of opiates to the locus coeruleus. Brain Res., 122: 523-533. Birnbaumer, L. (1990) G proteins in signal transduction. Ann. Rec. Pharmacol. Toxicol., 30: 675-705. Chavkin, C. and Goldstein, A. (1984) Opioid receptor reserve in normal and morphine-tolerant guinea pig ileum myenteric plexus. Proc. Natl. Acad. Sci. USA, 81: 7253-7257. Christie, M.J. and North, R.A. (1988) Agonists at p opioid, M, muscarinic and GABA, receptors increase the same potassium conductance in rat lateral parabrachial neurons. Br. J. Pharmacol., 95: 896-902. Christie, M.J., Williams, J.T. and North, R.A. (1987) Cellular mechanisms of opioid tolerance: Studies in single brain neurons. Mol. Pharmacol., 32: 633-638. Collier, H.O.J. (1980) Cellular site of opiate dependence. Nature (London), 283: 625-629. Crain, S.M., Shen, K.-F. and Chalazonitis, A. (1988) Opioids excite rather than inhibit sensory neurons after chronic opioid exposure of spinal cord-ganglion cultures. Brain Res., 455: 99-109.
Duggan, A.W. and North, R.A. (1983) Electrophysiology of opioids. Pharmacol. Reu., 35: 219-281. Duman, R.S., Tallman, J.F. and Nestler, E.J. (1988) Acute and chronic opiate-regulation of adenylate cyclase in brain: Specific effects in locus coeruleus. J. Pharrnacol. Exp. Ther., 246: 1033-1039. Gold, M.S., Redmond, D.E. and Kleber, H.D. (1978) Clonidine in opiate withdrawal. Lancet, i: 929-930. Guyenet, P.G. and Aghajanian, G.K. (1977) Excitation of neurons in the nucleus locus coeruleus by substance-P and related peptides. Brain Rex, 136: 178-184. Hescheler, J., Rosenthal, W., Trautwein, W. and Schultz, G. (1987) The GTP-binding protein, Go, regulates neuronal calcium channels. Nature (London), 325: 445-447. Hille, B. (1984) Ionic Channels in Excitable Membranes, Sinauer, Sunderland, MA, 426 pp. Inoue, M., Nakajima, S. and Nakajima Y. (1988) Somatostatin induces an inward rectification in rat locus coeruleus neurons through a pertussis toxin-sensitive mechanism. J. Physiol. (London), 407: 177-198. Koob, G.F. and Bloom, F.E. (1988) Cellular and molecular mechanisms of drug dependence. Science, 242: 715-723. Korf, J., Bunney, B.S. and Aghajanian, G.K. (1974) Noradrenergic neurons: Morphine inhibition of spontaneous activity. Eur. J. Pharmacol., 25: 165-167. Law, P.Y., Hom, D.S. and Loh, H.H. (1983) Opiate receptor down-regulation and desensitization in neuroblastoma X glioma NG108-15 hybrid cells are two separate cellular adaptation processes. Mol. Pharmacol., 24: 413-424. Leslie, F.M. (1987) Methods used for the study of opioid receptors. Pharmacol. Reu., 39: 197-249. Macdonald, R.L. and Werz, M.A. (1986) Dynorphin A decreases voltage-dependent calcium conductance of mouse dorsal root ganglion neurons. J. Physiol. (London), 377: 237-249. Madison, D.V. and Nicoll, R.A. (1987) Enkephalin hyperpolarizes interneurons in rat hippocampus. J. Physiol. (London), 398: 123-130. McFadzean, I., Lacey, M.G., Hill, R.G. and Henderson, G. (1987) Kappa opioid receptor activation depresses synaptic input to rat locus coeruleus neurons in vitro. Neuroscience, 20: 231-239. Miyake, M., Christie, M.J. and North, R.A. (1989) Single potassium channels opened by opioids in rat locus coeruleus neurons. Proc. Natl. Acad. Sci. USA, 86: 34193422. Nakamura, S., Tepper, J.M., Joung, S.J., Ling, N. and Groves, P.M. (1982) Noradrenergic terminal excitability: Effects of opioids. Neurosci. Lett., 30: 57-62. Nestler, E.J. and Tallman, J.F. (1988) Chronic morphine treatment increases cyclic AMP-dependent protein kinase activity in the rat locus coeruleus. Mol. Pharmacol., 33: 127-132. Nestler, E.J., Erdos, J.J., Terwilliger, R., Durnan, R.S. and Tallman, J.F. (1989) Regulation of G proteins by chronic morphine in rat locus coeruleus. Brain Res., 476: 230-239. North, R.A. (1989) Drug receptors and the inhibition of nerve cells. Br. J. Pharmacol., 98: 13-28.
205 North, R.A. and Williams, J.T. (1983) Opiate activation of potassium conductance inhibits calcium action potentials in rat locus coeruleus neurons. Br. J. Pharmacol., 80: 225-228. North, R.A. and Williams, J.T. (1985) On the potassium conductance increased by opioids in locus coeruleus neurons. J. Physiol. (London), 364: 265-280. North, R.A., Williams, J.T., Surprenant, A. and Christie, M.J. (1987) p and S opioid receptors both belong to a family of receptors which couple to a potassium conductance. Proc. Natl. Acad. Sci. USA, 84: 5487-5491. Osmanovic, S.S. and Shefner, S.A. (1988) Baclofen increases the potassium conductance of rat locus coeruleus neurons recorded in brain slices. Brain Res., 438: 124-136. Pepper, C.M. and Henderson, G. (1980) Opiates and opioid peptides hyperpolarize locus coeruleus neurons in vitro. Science, 209: 394-396. Pickel, V.M., Joh, T.H., Reis, D.J., Leeman, S.E. and Miller, R.J. (1979) Electron microscopic localization of substance P and enkephalin in axon terminals related to dendrites of catecholaminergic neurons. Brain Res., 160: 387-400. Rasmussen, K. and Aghajanian, G.K. (1989) Withdrawal-induced activation of locus coeruleus neurons in opiate-dependent rats: Attenuation by lesions of the nucleus paragigantocellularis. Brain Res., 505: 346-350. Rogers, N.F. and El-Fakahany, E. (1986) Morphine-induced opioid receptor down-regulation detected in intact adult rat brain cells. Eur. J. Pharmacol., 124: 221-230. Shefner, S.A. and Chiu, T.H. (1986) Adenosine inhibits locus coeruleus neurons: A n intracellular study in a rat brain slice preparation. Brain Res., 366: 364-368. Strahlendorf, H.K., Strahlendorf, J.C. and Barnes, C.D. (1980) Endorphin mediated inhibition of locus coeruleus neurons. Brain Rex, 191: 284-288. Ueda, H., Harada, H., Nozaki, M., Katada, T., Ui, M., Satoh, M. and Takagi, H. (1988) Reconstitution of rat brain 1 opioid receptors with purified guanine nucleotide binding regulatory proteins, Gi and Go. Proc. Natl. Acad. Sci. USA, 85: 7013-7017.
Valentino, R.J. and Wehby, R.G. (1989) Locus ceruleus discharge characteristics of morphine-dependent rats: Effects of naltrexone. Brain Res., 488: 126-134. Wang, Y.Y. and Aghajanian, G.K. (1987) Excitation of locus coeruleus neurons by adenosine 3’,5’-cyclic monophosphate-activated inward current: Extracellular and intracellular studies in rat brain slices. Synapse, 1: 481-487. Werling, L.L., Brown, S.L. and Cox, B.M. (1987) Opioid receptor regulation of the release of norepinephrine in brain. Neuropharmacology, 26: 987-996. Werling, L.L., Mcmahon, P.N. and Cox, B.M. (1989) Selective changes in p opioid receptor properties induced by chronic morphine exposure. Proc. Natl. Acad. Sci. USA, 86: 63936397. Williams, J.T. and North, R.A. (1984) Opiate-receptor interactions on single locus coeruleus neurons. Mol. PharmaC O ~,. 26: 67-70. Williams, J.T., Egan, T.M. and North, R.A. (1982) Enkephalin opens potassium channels in mammalian central neurons. Nature (London), 299: 74-76. Williams, J.T., Christie, M.J., North, R.A. and Roques, B.P. (1987) Potentiation of enkephalin action by peptidase inhibitors in rat locus coeruleus in vitro. J. Pharmacol. Exp. Ther., 243: 397-401. Williams, J.T., North, R.A. and Tokimasa, T. (1988) Inward rectification of resting and receptor-linked potassium currents in rat locus coeruleus neurons. J. Neurosci., 8: 42994306. Wong, Y.H., Demoliou-Mason, C.D., Barnard, E.A. (1989) Opioid receptors in magnesium-digitonin-solubilized rat brain membranes are tightly coupled to a pertussis toxinsensitive guanine nucleotide-binding protein. J. Neurochem., 52: 999-1009. Wuster, M., Schulz, R. and Herz, A. (1982) The development of opiate tolerance may dissociate from dependence. Life Sci., 31: 1695-1698.
Progress in Brain Research, Vol. 88
197
0 1991 Elsevier Science Publishers B.V.
CHAPTER 14
Mechanisms of opioid actions on neurons of the locus coeruleus M.J. Christie Department of Pharmacology, University of Sydney, N.S. W., Australia
The locus coeruleus (LC) has provided a useful model for pioneering studies of the mechanisms underlying the acute and chronic actions of opioid drugs. Acutely, opioids inhibit the electrical activity of single neurons in the rat and guinea pig LC. Inhibition is due to a membrane hyperpolarisation. In these cells, opioids act on y-receptors to increase the opening of inwardly rectifying potassium channels, thus leading to hyperpolarisation. The y-receptors are coupled to potassium channels via G-proteins which are sensitive to inactivation by
pertussis toxin. This coupling process is quite direct, in that it does not involve freely diffusible intracellular second messengers. Agonists specific for other receptors, such as aZ-and somatostatin-receptors, are capable of opening the same population of potassium channels on LC neurons. Following chronic treatment of animals with morphine, a specific deficit develops in the ability of p-receptors to open potassium channels, producing reduced sensitivity of LC neurons to inhibition by opioids.
Key words: locus coeruleus, receptor opioid, potassium channel, tolerance, G-protein, chronic drug
Introduction Opioids directly inhibit the electrical activity of neurons in many regions of the nervous system, including the locus coeruleus (LC) (see Duggan and North, 1983, for review). This inhibition arises from changes in ionic conductances in the somatic membrane or nerve terminal, leading ultimately to a reduced firing of action potentials and depressed release of neurotransmitter. The ionic mechanisms underlying opioid actions were first elucidated in the LC (Williams et al., 19821, and the molecular mechanisms involved have been more thoroughly studied in these neurons than in other regions of the CNS. It is now clear that
opioids inhibit LC neurons by activating p-receptors, which couple to the opening of potassium channels via G-proteins (North et aL, 1987). Within the LC neuron, the coupling process between receptor and potassium channel is localised to a membrane area of less than a few square microns, and does not involve freely diffusible second messengers (Miyake et al., 1989). No action of endogenously released opioids has yet been clearly identified in the LC, so the physiological role in the whole animal of this action on the LC soma is still unclear. Much of the impetus to understand the acute actions of opioids on CNS neurons has arisen from the problems of tolerance and physical de-
198
pendence. The LC has been a focus of much of this work, partly because of the similarity between the behavioural effects of electrical stimulation of the LC and the opiate withdrawal syndrome. While the mechanisms underlying dependence remain elusive, toIerance involves a specific deficit in the ability of opioid receptors to open potassium channels (Christie et al., 1987). Acute actions of opioids Opioids directly inhibit LC neurons Opioids inhibit the firing of action potentials when applied to neurons of many regions of the CNS (see Duggan and North, 1983, for review). In most cases where opioids excite neurons, it has been shown to occur as a result of the disinhibition of inhibitory interneurons (e.g., Madison and Nicoll, 1987). The first in vivo studies demonstrated inhibition of sensory-evoked action potential firing following intravenous administration of morphine (Korf et at., 1974). Inhibition was specific in that it was reversed by the opioid antagonist, naloxone. Similar results were reported in cat LC (Strahlendorf et al., 1980). Inhibition of spontaneous firing was also shown in rat LC in viuo following intravenous or localised (iontophoretic) application of opioids (Bird and Kuhar, 1977; Guyenet and Aghajanian, 1977). This observation has been substantiated in uiuo and in ilitro for species in which LC neurons can be identified as a fairly uniform population of noradrenaline-containing cells (see below). Opioid receptor type on LC neurons Opioids are known to affect cells via activation of at least three distinct membrane receptors; p-, 6- and K-receptors (see Leslie, 1987, for review). With the advent of brain slice techniques to study CNS neurons in vitro, the opportunity arose to classify the opioid receptor(s1 responsible for inhibition in LC neurons. Quantitative pharmacological techniques have demonstrated that inhibition is due to activation of p-receptors in the rat
(Williams and North, 1984). In addition to the actions of non-selective agonists such as methionine-enkephalin (met-enkephalin) and D-Ala-DLeu-enkephalin (DADLE), LC neurons were directly inhibited by p-agonists such as morphine and D-Ala-methionine-enkephalin-glyol(DAGO). However, the 6-agonist, D-Pen-D-Pen-enkephalin (DPDPE) (North et al., 1987) and the K-agonist, U50488H (Williams and North, 1984) were ineffective in rat LC, even at very high concentrations, suggesting that 6- and K-receptors were not responsible for inhibition. Quantification of the equilibrium dissociation constant (K,) of an antagonist is the most reliable method to define receptor type. This is because, unlike agonist actions, it does not depend on efficiency of coupling between receptor and effector. When determined by Schild analysis, the K , of naloxone to antagonize the action of met-enkephalin was 2 nM (Williams and North, 1984). This result was very similar to the value obtained for p-receptors in biochemical assays and differed by 10 to 30-fold from the value expected €or 6-or K-receptors. Moreover, the &receptor antagonist, ICI174864, blocked the effects of met-enkephalin very poorly ( K , > 5 pM).Thus, it is clear that inhibition of LC neurons by opioids is due solely to the activation of p-receptors. p-Receptors act cia G-proteins Although direct information on the primary aminoacid structure of the p-receptor is still lacking, there is little doubt that it belongs to the family of receptors which exert their cellular effects via GTP-binding proteins (G-proteins, see Birnbaumer, 1990, for review). Evidence that p receptors act through G-proteins includes inhibition of adenylate-cyclase in various preparations including LC (Duman et al., 1988), co-purification of p-binding sites with pertussis toxin-sensitive G-proteins (Wong et al., 1989) and the ability of purified p-receptors to reconstitute with purified G, and Go (Ueda et al., 1988). In LC, treatment of rats with pertussis toxin, which inactivates &-subunits of G, and Go, prevents inhibi-
199
tion of LC neurons by opioids (Aghajanian and Wang, 1986).
p-Receptor distribution on LC neurons Opioid receptors are distributed over the plasma membrane and nerve terminal regions of LC neurons. Physiological studies with patchclamp electrodes (see below) suggest a fairly uniform distribution of receptors over the perikaryal membrane of isolated LC neurons, although this result could have been influenced by enzymatic and mechanical procedures used to dissociate the LC. Some studies have also suggested that pagonists decrease the excitability of LC axon terminals. p-agonists inhibit noradrenaline release from brain slices of LC projection areas (Werling et al., 1987) and increase the threshold for antidromic activation of LC fibers when infused into cerebral cortex in civo (Nakamura et al., 1982). However, because these studies could not be conducted in the absence of synaptic transmission, they did not rule out indirect effects of p-agonists on the LC terminal membrane. p-Receptors activate potassium channels Techniques for intracellular recording from brain slices have also facilitated the elucidation of ionic mechanisms underlying opioid inhibition. The first intracellular recordings from LC neurons demonstrated inhibition to be due to a membrane hyperpolarization in guinea-pig (Pepper and Henderson, 1980). The hyperpolarization was associated with a fall in input resistance, and persisted under conditions in which synaptic transmission was blocked, indicating that opioids act directly on the impaled neuron. Williams et al. (1982) demonstrated that the hyperpolarization was due to an increased conductance to potassium ions in rat LC neurons. This was due to opening of membrane potassium channels, shifting the membrane potential from rest (about - 55 mV) towards the potassium equilibrium potential (about - 110 mV>. The macroscopic properties of the potassium
conductance were similar to those of a class of potassium channels known as inward rectifiers, first characterised in invertebrate egg cells (for review see Hille, 1984). These channels are so named because they do not open until the membrane potential approaches the potassium equilibrium potential, thus allowing potassium currents to flow more readily in the inward than outward direction. The mid-point of the conductance activated by opioids in LC neurons (about - 60 mV) is close to action potential threshold (about -55 mV), so these channels are capable of hyperpolarizing the membrane (North and Williams, 1985; Williams et al., 1988). Opioid-activated potassium channels have been demonstrated directly in acutely dissociated rat LC neurons (Miyake et at., 1989). When patchclamp pipettes were sealed onto the cell surface, single potassium channels were observed when opioids were included in the pipette solution, but not when opioids were applied to the rest of the cell. The probability of channel opening was dependent on agonist concentration and was antagonized when naloxone was included in the pipette. The channels had a unit conductance of about 45 pS in isotonic potassium and displayed bursts of activity lasting several seconds to minutes, separated by similar closed periods. A thorough analysis of channel kinetics was precluded by these long, closed periods, but within the periods of activity there may be two closed states and one open state of the channel. The open state is prolonged when higher concentrations of opioids are tested. All agonist-filled pipettes of about the same size contained similar numbers of channels, suggesting a fairly uniform distribution over the perikaryal membrane. The number of opioidactivated channels per cell was estimated to be about 1,000, given the diameter of recording pipettes and the total surface area of the LC neuron. This result suggests that p-receptors must also be fairly uniformly distributed over the soma, because they are within the few square microns of membrane contained within the pipette.
200
Mechanism of coupling between opioid receptors and potassium channels The main inference drawn from the experiments described above was that the coupling between opioid receptor activation and potassium channel opening is localized to the patch of membrane enclosed by the patch pipette, therefore ruling out the involvement of freely diffusible second messengers. Although second messenger systems such as the adenylate cyclase to protein kinase-A cascade are inhibited by opioids in LC (see below), these are not involved in the activation of potassium channels. One report, which suggested that agents which activate protein kinase-A reversed the hyperpolarization produced by morphine (Andrade and Aghajanian, 1985) was subsequently shown to result from activation of a small inward current by these agents (Wang and Aghajanian, 1987). Activators of protein kinase-A failed to prevent opioid hyperpolarizations, or potassium currents in other experiments (North and Williams, 1985). Beyond the lack of involvement of freely diffusible second messengers, and an involvement of G-proteins sensitive to pertussis toxin, the nature of the coupling between preceptor and potassium channel is unknown. It is possible that activated G-proteins bind directly to the channel, or that biochemical events confined to a region in close proximity ( < 1 pm) to the receptor and channel mediate the effects. These issues also need to be resolved for a variety of receptors which couple to potassium channels in LC and other cells. Multiple receptors open a single population of potassium channels Agonists selective for several receptors have been demonstrated to hyperpolarize, increase a potassium conductance and/or open potassium channels on LC neurons. These include agonists selective for adrenergic az- (North and Williams, 1983, somatostatin- (Inoue et al., 1988) and GABA,- (Osmanovic and Shefner, 1988) receptors. Adenosine receptor agonists also hyperpo-
larize LC neurons, possibly by activating a potassium conductance (Shefner and Chiu, 1986). This list is not exhaustive, agonists selective for some other receptors inhibit LC (see Olpe and Steinmann, this volume), and others might not have been tested. For the only combination tested, different receptors have been shown to activate the same population of potassium channels as do p-receptors. This conclusion has been inferred from simultaneous application of different agonists to LC neurons, e.g., if sufficient p-agonist is applied to an LC neuron to fully activate the potassium conductance, a,-agonists cannot activate the conductance any further, even though an a,-agonist applied alone can activate the conductance to the same extent as the p-agonist (North and Williams, 1985). This observation held for any combinations of p - and a,-agonists. The location in the sequence of biochemical events where different receptors converge to open one population of potassium channels is not known. It might be that each receptor can activate a different population of G-proteins which can all open the same population of channels, or in the extreme alternative, convergence to a single population of G-proteins may occur. In either case, it is not surprising that different receptors can open the same population of potassium channels on LC neurons. The same is true for p- (Christie and North, 1988), and other receptors in a variety of cells (see North, 1989, for review). These receptors generally belong to a family of G-protein-coupled receptors which have also been shown to inhibit adenylate cyclase (see Birnbaumer, 1990, for review). Like in the LC, where studied in detail, this coupling has been shown to involve pertussis toxin-sensitive G-proteins, but not freely diffusable second messengers such as CAMP.
Do opioid receptors influence other ion channels on LC neurons? Not only do multiple G-protein-coupled receptors converge onto single effectors, single receptors are able to influence multiple effectors in
20 1
single cells via G-proteins (see Birnbaumer, 1990, for review). In addition to opening potassium channels, opioid receptors have been shown to influence the activity of voltage-dependent calcium channels in several cell types. Both 6- and K-receptors inhibit calcium currents in cultured cells (Macdonald and Werz, 1986; Hescheler et al., 1987), and p-receptors inhibit calcium currents in a human neuroblastoma cell-line (E. Seward and G. Henderson, personal communication). Although p-agonists inhibited the calcium component of the action potential in LC neurons, this was shown to be secondary to potassium current activation (North and Williams, 1983). When studied directly, voltage-dependent calcium currents were unaffected in LC neurons by high concentrations of the non-selective opioid, met-enkephalin (Williams, Christie and North, unpublished observations). In the same experiments, muscarinic receptor agonists inhibited the current. To date, the electrophysiological actions of opioids on the soma of LC neurons seem to be confined to the opening of potassium channels.
Presynaptic actions of opioids Opioids have been shown to inhibit synaptic potentials impinging on neurons throughout the nervous system. In the LC, studies of opioid actions on synaptic potentials are complicated by the post-synaptic actions of p-opioids, i.e. , the large post-synaptic increase in potassium conductance affects both the amplitude and time course of synaptic potentials (see Williams et at., this volume). In contrast, although K-agonists have no post-synaptic effects, they have been reported to produce a 38% inhibition of the amplitude of fast synaptic potentials which were evoked by electrical stimulation in rat brain slices (McFadzean et al., 1987). The fast synaptic potential is due to the release of at least three neurotransmitters: an excitatory aminoacid, GABA and glycine (see Williams et al., this volume). It is unknown which of these are inhibited by K-opioids, and which projections to LC are responsible.
Physiological role of opioids in LC The physiological role of p-receptor-mediated inhibition of LC neurons has yet to be established. Fibers and terminals containing products of pro-opiomelanocortin, pro-dynorphin and pro-enkephalin are present in the vicinity of the LC (see Sutin and Jacobowitz, this volume) and may make synaptic contacts with LC neurons (Pickel et al., 1979). However, attempts to detect a post-synaptic response to endogenous opioids released by electrical or high potassium stimulation within slices of LC met with no success, even in the presence of peptidase inhibitors (Williams et al., 1987). This result is not surprising, an endogenous opioid-mediated synaptic response has yet to be reported on any neuron. A possible explanation is that unusual stimulus parameters are needed to evoke release of endogenous opioids and other peptides, such as somatostatin, within the LC. It is also possible that endogenous opioids act on LC neurons in a diffuse, non-synaptic, manner. One report of a naloxone-reversible inhibition of LC firing following stimulation of the pro-opiomelanocortin containing arcuate nucleus (Strahlendorf et al., 1980) has yet to be shown to be a direct effect on LC using intracellular recording. Opioid tolerance and physical dependence in LC The notion that tolerance to opioids, i.e., reduced responsiveness to an agonist following chronic exposure to it, involves cellular adaptive processes within the neuron that bears opioid receptors is generally accepted (see Koob and Bloom, 1988, for review). The mechanisms underlying physical dependence are less clearly understood than those underlying tolerance. Physical dependence is characterised in whole animals by withdrawal signs upon removal of agonist, or administration of antagonist. Withdrawal signs are often qualitatively opposite to the acute agonist actions, e.g., hyperalgesia is a withdrawal sign opposite to the acute analgesic action of opioids. At the level of single neurons, signs of dependence would be
202
considered as the occurrence of enhanced spontaneous activity or heightened excitability. Tolerance and physical dependence following chronic exposure to opioids have been extensively investigated in the LC. This interest arose partly from the simiIarity between the behavioral effects of electrical stimulation of the LC and the opiate withdrawal syndrome. Indeed, this similarity provided the rationale for the use of clonidine in the treatment of opiate withdrawal in man (Gold et al., 1978). Clonidine, and other a,-adrenoceptor agonists, by acting through the same cellular pathway as opioids inhibits LC neurons and thereby reduces withdrawal excitation. Tolerance The electrophysiological mechanisms underlying tolerance have been studied in any detail only in LC neurons (Andrade et al., 1983; Christie et al., 1987). Tolerance was observed as a reduction in the sensitivity of LC neurons to agonists such as met-enkephalin and normorphine. The maximum potassium current induced by normorphine was also reduced (Christie et al., 1987). This reduced sensitivity to opioids was observed for many hours after removal of tissue from animals and hence withdrawal of morphine. These effects were indistinguishable from inactivation of a fraction of the p-receptors on the normal cell surface with the irreversible antagonist, p-chlornaltrexamine. It was, therefore, concluded that tolerance was equivalent to inactivation of p-receptors, This interpretation is functionally equivalent to a decrease in the ability of individual p-receptors to open channels. This conclusion was similar to those drawn from isolated organ experiments (Chavkin and Goldstein, 1984). No effects were found on the K , of naloxone, the potassium conductance itself, or the ability of aZadrenoceptor agonists to open the same population of potassium channels (Christie el al., 1987). No signs indicative of physical dependence (withdrawal, or antagonist-induced excitation) were observed (Andrade et al., 1983; Christie et al., 1987).
These studies localized tolerance specifically to coupling between p-receptors and potassium channels, because the channels themselves and the ability of other receptors to couple to them were unaffected. Tolerance must therefore be due to a reduced preceptor density on the cell surface, or reduced ability of receptors to couple to G-proteins. These two possibilities could not be distinguished in physiological studies. Examples of both possibilities have been observed in biochemical studies of p- and &receptors in cultured cells (Law et al., 1983; Werling et al., 1989) and brain membrane preparations (Rogers and El-Fakahany, 1986). Beyond this level of analysis, little is known of the mechanisms underlying opioid tolerance. Changes occur with chronic opioids in the adenylate-cyclase to protein kinase-A cascade (Duman et al., 1988; Nestler and Tallman, 1988) and Gprotein ribosylation in LC (Nestler et al., 1989); however, the physiological relationship of these adaptations to tolerance is unknown. An observation which could be related to tolerance development is that p-receptors were desensitised for 15-20 min after occupation by high concentrations of agonists (G.C. Harris and J.T. Williams, personal communication). This acute desensitisation is analogous to homologous desensitisation of other G-protein coupled receptors, e.g., padrenoceptors are phosphorylated and inactivated by p-adrenergic receptor-kinase only when occupied by an agonist (Benovic et al., 1989). Whether similar mechanisms play a role in the development of the long-term tolerance described above is unknown. Physical dependence Dependence (increase in extracellular action potential frequency) was first reported in LC neurons in vivo after iontophoretic application of the opioid antagonist, naloxone (Aghajanian, 1978). These results were replicated using intracerebroventricular application of an antagonist, naltrexone (Valentino and Wehby, 1989). However, these observations were not replicated using
203
extracellular (Andrade et af., 1983) or intracellular (Christie et af., 1987) recording in vitro, upon withdrawal of opioids, or upon application of naloxone. There are several possible explanations for this discrepancy. First, increased firing frequency in ciuo might simply be due to reversal of the effects of the agonist (morphine) which was still present in the tissue. However, the naloxoneinduced frequency was much greater than that observed in naive animals (Aghajanian, 1978; Valentino and Wehby, 1989), ruling out this possibility. Second, if dependence in LC dissipates very rapidly, it may have declined during the preparation of brain slices and impalement of neurons (usually longer than 30 min in the absence of an agonist). However, maintenance of tissue in morphine (1 FM) during brain slice preparation did not lead to the observation of withdrawal excitation upon superfusion of naloxone (J.T. Williams, personal communication), ruling out this possibility. Finally, increased frequency could have been due to altered synaptic activity, which would bc disrupted in brain slices. This would suggest that dependence develops in nerve terminals and/or cells projecting to the LC. This possibility has not been explored in vitro, but lesions to nucleus paragigantocellularis, which provides the main excitatory afferents to the LC, attenuated withdrawal excitation in uiuo (Rasmussen and Aghajanian, 1989). These studies imply that tolerance and dependence can be dissociated at the neuronal level and are, therefore, consequences of different cellular mechanisms. This interpretation was consistent with similar observations reported for isolated organ preparations (Wuster et al., 1982), contrary to earlier theories of opioid tolerance and dependence, which posited a unitary mechanism for both phenomena (Collier, 1980). This interpretation might imply that tolerance occurs in many types of opioid sensitive neuron, but dependence develops only in a subset. Some studies support this possibility. Cells in mouse dorsal root ganglion explants develop hyper-excitability following chronic treatment with morphine (Crain
et a/., 1988). Another possibility is that dependence requires intact neural networks or requires some form of synaptic interactions. This might explain the discrepancy between the observation of dependence in LC neurons in uivo and its absence in uitro. Conclusions The LC has contributed greatly to our understanding of the acute and chronic actions of opioids. Within the LC opioids have direct inhibitory effects on the soma, as well as presynaptic inhibitory effects. It is also likely that opioids have direct inhibitory effects on the terminals of LC neurons. The somatic effects are due to opening of inwardly rectifying potassium channels. p-receptors, as well as other membrane receptors, couple to these channels via the activation of pertussis toxin-sensitive G-proteids). Diffusible second messengers, such as CAMP, are not involved in the coupling process between preceptor and potassium channel. The physiological role of inhibition of adenylate cyclase by preceptors in LC is still unknown, but might involve long-term processes such as gene regulation. This understanding of the mechanisms of acute actions has facilitated analysis of the chronic effects of opioids in LC. Tolerance to opioids in LC in uitro is due either to a reduced number of preceptors on the soma, or to a reduced efficiency of coupling between preceptors and potassium channels. This deficit is specific in that a,-adrenoceptor agonists can still fully activate the same population of potassium channels. Signs of opioid withdrawal have been observed in LC neurons in uiuo, but not in brain slices in citro. The reasons for this discrepancy are unclear. The observation of tolerance, but not physical dependence, in LC neurons in uitro suggests the mechanisms underlying the two phenomena differ. The presynaptic actions of opioids are less clearly understood than the post-synaptic effects. K-receptor agonists depress fast synaptic potentials evoked in the LC by electrical stimulation.
204
The neurotransmitters involved and anatomical source of these synaptic potentials have yet to be described. A clear demonstration of synaptic responses to endogenously released opioids is lacking in the LC, as it is elsewhere in the nervous system. Acknowledgements Drs. J.T. Williams and S.M. Johnson are gratefully acknowledged for their comments and suggestions. References Aghajanian, G.K. (1978) Tolerance of locus coeruleus neurons to morphine and suppression of withdrawal response by clonidine. Nature (London), 276: 86-188. Aghajanian, G.K. and Wang, Y.Y. (1986) Pertussis toxin blocks the outward currents evoked by opiate and a, agonists in locus coeruleus neurons. Brain Res., 371: 390-394. Andrade, R. and Aghajanian, G.K. (1985) Opiate- and alpha,-adrenoceptor-induced hyperpolarizations of locus coeruleus neurons in brain slices: Reversal by cyclic adenosine 3’,5’-monophosphate analogues. J. Neurosci., 5: 2359-2364. Andrade, R., VanderMaelen, C.P. and Aghajanian, G.K. (1983) Morphine tolerance and dependence in locus coeruleus: Single cell studies in brain slices. Eur J. Pharmacol., 91: 161-169. Benovic, J.L., DeBlasi, A., Stone, W.C., Caron, M.G. and Lefkowitz, R.J. (1989) P-adrenergic receptor kinase: Primary structure delineates a multigene family. Science, 246: 235-240. Bird, S.J. and Kuhar, M.J. (1977) Iontophoretic application of opiates to the locus coeruleus. Brain Res., 122: 523-533. Birnbaumer, L. (1990) G proteins in signal transduction. Ann. Rec. Pharmacol. Toxicol., 30: 675-705. Chavkin, C. and Goldstein, A. (1984) Opioid receptor reserve in normal and morphine-tolerant guinea pig ileum myenteric plexus. Proc. Natl. Acad. Sci. USA, 81: 7253-7257. Christie, M.J. and North, R.A. (1988) Agonists at p opioid, M, muscarinic and GABA, receptors increase the same potassium conductance in rat lateral parabrachial neurons. Br. J. Pharmacol., 95: 896-902. Christie, M.J., Williams, J.T. and North, R.A. (1987) Cellular mechanisms of opioid tolerance: Studies in single brain neurons. Mol. Pharmacol., 32: 633-638. Collier, H.O.J. (1980) Cellular site of opiate dependence. Nature (London), 283: 625-629. Crain, S.M., Shen, K.-F. and Chalazonitis, A. (1988) Opioids excite rather than inhibit sensory neurons after chronic opioid exposure of spinal cord-ganglion cultures. Brain Res., 455: 99-109.
Duggan, A.W. and North, R.A. (1983) Electrophysiology of opioids. Pharmacol. Reu., 35: 219-281. Duman, R.S., Tallman, J.F. and Nestler, E.J. (1988) Acute and chronic opiate-regulation of adenylate cyclase in brain: Specific effects in locus coeruleus. J. Pharrnacol. Exp. Ther., 246: 1033-1039. Gold, M.S., Redmond, D.E. and Kleber, H.D. (1978) Clonidine in opiate withdrawal. Lancet, i: 929-930. Guyenet, P.G. and Aghajanian, G.K. (1977) Excitation of neurons in the nucleus locus coeruleus by substance-P and related peptides. Brain Rex, 136: 178-184. Hescheler, J., Rosenthal, W., Trautwein, W. and Schultz, G. (1987) The GTP-binding protein, Go, regulates neuronal calcium channels. Nature (London), 325: 445-447. Hille, B. (1984) Ionic Channels in Excitable Membranes, Sinauer, Sunderland, MA, 426 pp. Inoue, M., Nakajima, S. and Nakajima Y. (1988) Somatostatin induces an inward rectification in rat locus coeruleus neurons through a pertussis toxin-sensitive mechanism. J. Physiol. (London), 407: 177-198. Koob, G.F. and Bloom, F.E. (1988) Cellular and molecular mechanisms of drug dependence. Science, 242: 715-723. Korf, J., Bunney, B.S. and Aghajanian, G.K. (1974) Noradrenergic neurons: Morphine inhibition of spontaneous activity. Eur. J. Pharmacol., 25: 165-167. Law, P.Y., Hom, D.S. and Loh, H.H. (1983) Opiate receptor down-regulation and desensitization in neuroblastoma X glioma NG108-15 hybrid cells are two separate cellular adaptation processes. Mol. Pharmacol., 24: 413-424. Leslie, F.M. (1987) Methods used for the study of opioid receptors. Pharmacol. Reu., 39: 197-249. Macdonald, R.L. and Werz, M.A. (1986) Dynorphin A decreases voltage-dependent calcium conductance of mouse dorsal root ganglion neurons. J. Physiol. (London), 377: 237-249. Madison, D.V. and Nicoll, R.A. (1987) Enkephalin hyperpolarizes interneurons in rat hippocampus. J. Physiol. (London), 398: 123-130. McFadzean, I., Lacey, M.G., Hill, R.G. and Henderson, G. (1987) Kappa opioid receptor activation depresses synaptic input to rat locus coeruleus neurons in vitro. Neuroscience, 20: 231-239. Miyake, M., Christie, M.J. and North, R.A. (1989) Single potassium channels opened by opioids in rat locus coeruleus neurons. Proc. Natl. Acad. Sci. USA, 86: 34193422. Nakamura, S., Tepper, J.M., Joung, S.J., Ling, N. and Groves, P.M. (1982) Noradrenergic terminal excitability: Effects of opioids. Neurosci. Lett., 30: 57-62. Nestler, E.J. and Tallman, J.F. (1988) Chronic morphine treatment increases cyclic AMP-dependent protein kinase activity in the rat locus coeruleus. Mol. Pharmacol., 33: 127-132. Nestler, E.J., Erdos, J.J., Terwilliger, R., Durnan, R.S. and Tallman, J.F. (1989) Regulation of G proteins by chronic morphine in rat locus coeruleus. Brain Res., 476: 230-239. North, R.A. (1989) Drug receptors and the inhibition of nerve cells. Br. J. Pharmacol., 98: 13-28.
205 North, R.A. and Williams, J.T. (1983) Opiate activation of potassium conductance inhibits calcium action potentials in rat locus coeruleus neurons. Br. J. Pharmacol., 80: 225-228. North, R.A. and Williams, J.T. (1985) On the potassium conductance increased by opioids in locus coeruleus neurons. J. Physiol. (London), 364: 265-280. North, R.A., Williams, J.T., Surprenant, A. and Christie, M.J. (1987) p and S opioid receptors both belong to a family of receptors which couple to a potassium conductance. Proc. Natl. Acad. Sci. USA, 84: 5487-5491. Osmanovic, S.S. and Shefner, S.A. (1988) Baclofen increases the potassium conductance of rat locus coeruleus neurons recorded in brain slices. Brain Res., 438: 124-136. Pepper, C.M. and Henderson, G. (1980) Opiates and opioid peptides hyperpolarize locus coeruleus neurons in vitro. Science, 209: 394-396. Pickel, V.M., Joh, T.H., Reis, D.J., Leeman, S.E. and Miller, R.J. (1979) Electron microscopic localization of substance P and enkephalin in axon terminals related to dendrites of catecholaminergic neurons. Brain Res., 160: 387-400. Rasmussen, K. and Aghajanian, G.K. (1989) Withdrawal-induced activation of locus coeruleus neurons in opiate-dependent rats: Attenuation by lesions of the nucleus paragigantocellularis. Brain Res., 505: 346-350. Rogers, N.F. and El-Fakahany, E. (1986) Morphine-induced opioid receptor down-regulation detected in intact adult rat brain cells. Eur. J. Pharmacol., 124: 221-230. Shefner, S.A. and Chiu, T.H. (1986) Adenosine inhibits locus coeruleus neurons: A n intracellular study in a rat brain slice preparation. Brain Res., 366: 364-368. Strahlendorf, H.K., Strahlendorf, J.C. and Barnes, C.D. (1980) Endorphin mediated inhibition of locus coeruleus neurons. Brain Rex, 191: 284-288. Ueda, H., Harada, H., Nozaki, M., Katada, T., Ui, M., Satoh, M. and Takagi, H. (1988) Reconstitution of rat brain 1 opioid receptors with purified guanine nucleotide binding regulatory proteins, Gi and Go. Proc. Natl. Acad. Sci. USA, 85: 7013-7017.
Valentino, R.J. and Wehby, R.G. (1989) Locus ceruleus discharge characteristics of morphine-dependent rats: Effects of naltrexone. Brain Res., 488: 126-134. Wang, Y.Y. and Aghajanian, G.K. (1987) Excitation of locus coeruleus neurons by adenosine 3’,5’-cyclic monophosphate-activated inward current: Extracellular and intracellular studies in rat brain slices. Synapse, 1: 481-487. Werling, L.L., Brown, S.L. and Cox, B.M. (1987) Opioid receptor regulation of the release of norepinephrine in brain. Neuropharmacology, 26: 987-996. Werling, L.L., Mcmahon, P.N. and Cox, B.M. (1989) Selective changes in p opioid receptor properties induced by chronic morphine exposure. Proc. Natl. Acad. Sci. USA, 86: 63936397. Williams, J.T. and North, R.A. (1984) Opiate-receptor interactions on single locus coeruleus neurons. Mol. PharmaC O ~,. 26: 67-70. Williams, J.T., Egan, T.M. and North, R.A. (1982) Enkephalin opens potassium channels in mammalian central neurons. Nature (London), 299: 74-76. Williams, J.T., Christie, M.J., North, R.A. and Roques, B.P. (1987) Potentiation of enkephalin action by peptidase inhibitors in rat locus coeruleus in vitro. J. Pharmacol. Exp. Ther., 243: 397-401. Williams, J.T., North, R.A. and Tokimasa, T. (1988) Inward rectification of resting and receptor-linked potassium currents in rat locus coeruleus neurons. J. Neurosci., 8: 42994306. Wong, Y.H., Demoliou-Mason, C.D., Barnard, E.A. (1989) Opioid receptors in magnesium-digitonin-solubilized rat brain membranes are tightly coupled to a pertussis toxinsensitive guanine nucleotide-binding protein. J. Neurochem., 52: 999-1009. Wuster, M., Schulz, R. and Herz, A. (1982) The development of opiate tolerance may dissociate from dependence. Life Sci., 31: 1695-1698.
