Opioid Receptors

Opioid Receptors

Alistair Corbett, Sandy McKnight and Graeme Henderson

Dr Alistair Corbett is Lecturer in the School of Biological and Biomedical Sciences, Glasgow Caledonian University, Cowcaddens Road, Glasgow G4 0BA, UK.

Dr Sandy McKnight is Associate Director, Parke-Davis Neuroscience Research Centre, Cambridge University Forvie Site, Robinson Way, Cambridge CB2 2QB, UK.

Professor Graeme Henderson is Professor of Pharmacology and Head of Department, Department of Pharmacology, School of Medical Sciences, University of Bristol, University Walk, Bristol BS8 1TD, UK.

Introduction

Preparations of the opium poppy papaver somniferum have been used for many hundreds of years to relieve pain. In 1803, Sertürner isolated a crystalline sample of the main constituent alkaloid, morphine, which was later shown to be almost entirely responsible for the analgesic activity of crude opium. The rigid structural and stereochemical requirements essential for the analgesic actions of morphine and related opioids led to the theory that they produce their effects by interacting with a specific receptor.¹ The concept that there is more than one type of opioid receptor arose to explain the dual actions of the synthetic opioid nalorphine, which antagonises the analgesic effect of morphine in man but also acts as an analgesic in its own right. Martin (1967) concluded that the analgesic action of nalorphine is mediated by a receptor, later called the k-opioid receptor, that is different from the morphine receptor.² Evidence for multiple receptors, m, k and s, came from the demonstration of different profiles of pharmacological activity in the chronic spinal dog with the prototype agonists morphine, ketazocine and N-allylnormetazocine (SKF 10047).³ The existence of the d-receptor was subsequently proposed to explain the profile of activity in vitro of the enkephalins (the first endogenous opioid peptides), and on the basis of the relative potency of the non-selective opioid antagonist naloxone to reverse endogenous opioid peptide inhibition of the nerve-evoked contractions of the mouse vas deferens.⁴ Its existence was further confirmed by radioligand binding studies using rat brain homogenates.

It is now clear from work carried out in many laboratories over the last 20 years that there are 3 well-defined or "classical" types of opioid receptor µ, d and k . Genes encoding for these receptors have been cloned.^{5, 6, 7,
8} More recently, cDNA encoding an "orphan" receptor was identified which has a high degree of homology to the "classical" opioid receptors; on structural grounds this receptor is an opioid receptor and has been named ORL₁ (opioid receptor-like).⁹ As would be predicted from their known abilities to couple through pertussis toxin-sensitive G-proteins, all of the cloned opioid receptors possess the same general structure of an extracellular N-terminal region, seven transmembrane domains and intracellular C-terminal tail structure. There is pharmacological evidence for subtypes of each receptor and other types of novel, less well-characterised opioid receptors, e, l, i, z, have also been postulated. The s-receptor, however, is no longer regarded as an opioid receptor.

Receptor Subtypes

m-Receptor subtypes
The MOR-1 gene, encoding for one form of the m-receptor, shows approximately 50-70% homology to the genes encoding for the d-(DOR-1), k-(KOR-1) and orphan (ORL1) receptors. Two splice variants of the MOR-1 gene have been cloned, differing only in the presence or absence of 8 amino acids in the C-terminal tail. The splice variants exhibit differences in their rate of onset and recovery from agonist-induced internalization but their pharmacology does not appear to differ in ligand binding assays.¹⁰ Furthermore, in the MOR-1 knockout mouse, morphine does not induce antinociception demonstrating that at least in this species morphine’s analgesia is not mediated through d- or k-receptors.¹¹ Similarly morphine did not exhibit positive reinforcing properties or an ability to induce physical dependence in the absence of the MOR-1 gene.

m₁ and m₂: The m₁/m₂ subdivision was proposed by Pasternak and colleagues to explain their observations, made in radioligand binding studies, that [³H]-labelled-m, -d and -k ligands displayed biphasic binding characteristics.¹² Each radioligand appeared to bind to the same very high affinity site (m₁) as well as to the appropriate high affinity site (m, d or k) depending on the radioligand used. Naloxazone and naloxonazine were reported to abolish the binding of each radioligand to the m₁-site. Furthermore, in in vivo studies it was observed that naloxazone selectively blocked morphine-induced antinociception but did not block morphine-induced respiratory depression or the induction of morphine dependence.^{13,
14} Subsequent work in other laboratories has failed to confirm this classification.

Is there another, novel form of the m-opioid receptor?
Several related observations suggest the existence of a novel form of m-receptor at which analogues of morphine with substitutions at the 6 position (e.g. morphine-6b-glucuronide, heroin and 6-acetyl morphine) are agonists, but with which morphine itself does not interact.¹⁵ In antinociception tests on mice it has been reported that morphine does not exhibit cross tolerance with morphine-6b-glucuronide, heroin or 6-acetyl morphine. Furthermore, in mice of the CXBX strain morphine is a poor antinociceptive agent whereas morphine-6b-glucuronide, heroin and 6-acetyl morphine are all potently antinociceptive. The 6-substituted morphine analogues do not appear to be acting through d- or k-receptors because the antinociception they induce is not blocked by selective d- or k-receptor antagonists, whereas 3-methoxynaltrexone has been reported to antagonise morphine-6b-glucuronide- and heroin-induced antinociception without affecting that induced by morphine, [D-Pen², D-Pen⁵]enkephalin (DPDPE , d-selective) or U50488 (k-selective).¹⁶

Recently it has been reported that heroin and morphine-6-glucuronide, but not morphine, still produce antinociception in MOR-1 knockout mice in which the disruption in the MOR-1 gene was engineered in exon-1.¹⁷ The same authors observed that in other MOR-1 knockout mice in which exon-2, not exon-1, had been disrupted, all three agonists were ineffective as antinociceptive agents. They conclude that the antinociceptive actions of heroin and morphine-6-glucuronide in the exon-1 MOR-1 mutant mice are mediated through a receptor produced from an alternative transcript of the MOR-1 gene differing from the MOR-1 gene product, the m-opioid receptor, in the exon-1 region. To substantiate this conclusion they report that in RT-PCR experiments using primers spanning exons 2 and 3, a MOR-1 gene product was still detected in MOR-1 knockout mice.

d-Receptor subtypes
The DOR-1 gene is the only d-receptor gene cloned to date. However, two, overlapping subdivisions of d-receptor have been proposed (d₁/d₂ and d_cx/d_ncx) on the basis of in vivo and in vitro pharmacological experiments.

d₁ and d₂: The subdivision of the d-receptor into d₁ and d₂ subtypes was proposed primarily on the basis of in vivo pharmacological studies (Table 1). In rodents in vivo, the supraspinal antinociceptive activity of DPDPE can be selectively antagonised by 7-benzylidenenaltrexone (BNTX) or [D-Ala², D-Leu⁵]enkephalyl-Cys (DALCE)^{18, 19} whereas the antinociceptive activity of [D-Ala²]-deltorphin II (deltorphin II) and [D-Ser², Leu⁵]enkephalyl–Thr (DSLET) can be reversed by naltriben or naltrindole 5¢-isothiocyanate (5¢-NTII).^{18, 19,
20} Furthermore, while mice develop tolerance to the antinociceptive effects of repeated injections of either DPDPE or deltorphin II, this tolerance appears to be homologous in that there is no cross tolerance between these ligands.²¹ In vivo, d₁- and d₂-receptor-induced antinociception can be differentially antagonised by blockers of different types of potassium channels.²²

Table 1 . Putative ligands for d-receptor subtypes

Receptor subtype	Antagonists
Receptor subtype	Competitive	Nonequilibrium
d ₁	DPDPE / DADLE	BNTX	DALCE
d₂	Deltorphin II / DSLET	Naltriben	5¢-NTII

N.B. DPDPE may not in fact be a selective d₁ agonist but may also be a partial agonist at d₂-sites.²³

The best evidence from in vitro experiments to support the d₁ and d₂ subdivision of d-receptors comes from inhibition of adenylyl cyclase activity in membranes from rat brain^{24, 25} and from the d-receptor-mediated elevations of intracellular Ca²⁺ in the ND8-47 cell line²⁶ where BNTX selectively antagonised DPDPE, and naltriben selectively antagonised deltorphin II. Surprisingly, little selectivity was seen in radioligand displacement studies.²⁴ The converse has been observed in studies on neuronal cell lines. Two distinct d-receptor binding sites were observed in radioligand binding experiments on SK-N-BE cells.²⁷ Studies on NG108-15 cells²⁸ or the human neuroblastoma cell line, SH-SY5Y,^{29, 30} have failed to find any functional evidence for d-receptor subtypes.

The pharmacological properties of the cloned DOR-1 receptor are somewhere between those predicted for either the d₁ or d₂ subtypes. DPDPE and deltorphin II are both potent displacers of [³H]-diprenorphine binding to mouse and human recombinant receptors, which is not consistent with either the d₁ or d₂ classifications.³¹ In contrast, [³H]-diprenorphine binding to the mouse recombinant receptor is more potently displaced by naltriben than BNTX, suggesting that the cloned receptor is of the d₂ subtype. It will be of importance to determine in the DOR-1 knockout mouse if analgesia can still be induced by either d₁- or d₂-receptor selective agonists.

d_cx and d_ncx: The d_cx and d_ncx subdivision of d-receptors was based on the hypothesis that one type of d-receptor (d_cx) was complexed with m-receptors (and perhaps also k-receptors) whereas the other type of d-receptor (d_ncx) was not associated with an opioid receptor complex.³² It was originally observed that sub-antinociceptive doses of agonists at the d_cx receptor (e.g. low doses of DPDPE), potentiated m-receptor-mediated analgesia, an effect which could be antagonised by 5¢-NTII. On the other hand, at higher doses, DPDPE then acted as an agonist at the d_ncx-receptor and itself induced analgesia which was reversed by DALCE. Data obtained from subsequent radioligand binding studies have been interpreted as demonstrating the existence of further subtypes of the d_ncx receptor i.e. d(_ncx-1) and d(_ncx-2). More recently it has been suggested that the d(_ncx-1) receptor is in fact synonymous with the d₁-receptor and the d_cx-receptor synonymous with the d₂-receptor of the previous classification.³³

k-Receptor subtypes
The situation regarding the proposals for subtypes of the k-receptor is rather more complex than for the m- and d-receptors, perhaps because of the continuing use of non-selective ligands to define the putative sites. The evidence for the need for sub-division of the k-receptor comes almost entirely from radioligand binding assays.

The first characterisation of a k-receptor binding site in brain came from work using [³H]-ethylketocyclazocine (EKC).³⁴ Crucial to this success was the use of the guinea-pig brain where k-sites are present in relative abundance, and of "suppression", or quenching of the binding of this non-selective ligand to m- and d-sites, by incubation with non-radioactive ligands that bound selectively at these other sites.

Studies of [³H]-EKC binding in guinea-pig spinal cord pointed to the existence of a non-homogeneous population of high-affinity binding sites, and led to the first proposal for k₁- and k₂-sites distinguished by their sensitivity to DADLE.³⁵ The DADLE-sensitive k₂ site bound b-endorphin with high affinity, and was later identified with the recognition site of the e-receptor in brain.³⁶ Another study using [³H]-EKC identified a k-site in bovine adrenal medulla, with a pharmacology similar to that of the k₁-site in guinea-pig cord³⁷ but labelling with [³H]-etorphine revealed two additional sites, one resembling k₂ that bound [Met]enkephalyl-Arg-Gly-Leu with high affinity and another termed "k₃" or "MRF" that bound [Met⁵]enkephalyl-Arg-Phe with high affinity.

The k₁/k₂ terminology has more recently been applied by other groups to the putative subtypes defined in other tissues in their hands, but it is not always clear how closely the common nomenclature reflects a common pharmacology. The introduction of the first selective k-agonist U-50,488 and its congeners (U-69,593, PD 117302, CI 977, ICI 197067) led to a refinement of the definition of the putative subtypes, but pointed to the need for careful considerations of the effect of technical differences in assays and of species as a possible explanation for discrepancies. Thus a direct comparison of the binding of [³H]-EKC in guinea-pig and rat (with suppression of binding to m- and d-sites) pointed to the existence of a high affinity k₁-site that predominated in guinea-pig brain and was selectively sensitive to U-69,593, and a low affinity, U-69,593-insensitive k₂-site that predominated in rat brain.³⁸ Others resorted to the binding of [³H]-bremazocine to reveal U-69,593-insensitive k₂-binding sites; in contrast to the k₂-site originally defined in guinea-pig spinal cord, the k₂-site in brain after suppression of k₁ was insensitive to DADLE.³⁹

Subdivision of the k₁-site in guinea-pig brain into k_1a and k_1b, was proposed to resolve the complex displacement of either [³H]-EKC or [³H]-U-69,593 with dynorphin B and a-neo-endorphin which both preferentially bound to the proposed k_1b sub-subtype.⁴⁰ The same study proposed the existence of a k₃ subtype, insensitive to U-50,488, that was identified from the binding of [³H]-naloxone benzoylhydrazone. The pharmacology of this later "k₃-site" is rather different from the k₃/MRF site of bovine adrenal medulla, and has been proposed to be the receptor mediating the antinociceptive effect of nalorphine, Martin’s "N"-receptor.⁴¹

Nomenclature differences appear to have arisen in the context of subtyping of the k₁-subtype. Using binding surface analyses to allow highly accurate estimation of binding parameters, the binding of [³H]-U-69,593 resolved two binding sites termed k_1a and k_1b. The ligand demonstrating the highest affinity, and around 30-fold preference, for the "k_1a binding site" was a-neo-endorphin.⁴² More recently putative k_1a- and k_1b-sites in mouse brain were identified from complex displacement curves against the binding of [³H]-U-69,593, in an attempt to compare the pharmacology of the mouse k₁-sites, with that at the cloned rat KOR stably expressed in a host neuroblastoma cell line.⁴³ Based on the high affinity of bremazocine and a-neo-endorphin, it was deemed "consistent to term the cloned KOR a k_1b subtype".

Rothman (1990) also reported subdivision of the k₂-binding of [³H]-bremazocine into 2a- and 2b-sub-subtypes.⁴² The k_2b-site had high affinity for b-endorphin and DADLE, reminiscent of the original k₂-binding site of guinea-pig spinal cord. The k_2a- and k_2b-sites in guinea-pig brain have undergone a further subdivision (sub-sub-subtypes?) on the basis of investigations using a combination of depletion (of m- and d-sites) and suppression, against the binding of 6b-[¹²⁵I]-3,14-dihydroxy-17-cyclopropylmethyl-4,5a-epoxymorphinan ([¹²⁵I]OXY).⁴⁴ So were defined the k_2a-1 and k_2a-2 sites, having relatively high and low affinities respectively for nor-BNI and enadoline (CI-977), and k_2b-1 and k_2b-2 sites with high and low affinities for DAMGO and a-neo-endorphin.

Definitive functional pharmacological evidence supporting the existence of this confusing number of putative subtypes of the k-receptor is lacking, because of the absence of subtype-specific antagonists. It has been reported however, that pretreatment with the isothiocyanate analogue of U-50,488 called (-)-UPHIT, was able to produce a long-lasting block of the antinociceptive effect of U-69,593 in the mouse without affecting the action of bremazocine, while treatment with the non-selective antagonist WIN 44,441 (quadazocine) blocked selectively the antinociception with bremazocine.^45,46 These findings provide obvious support for the k₁-k₂ subdivision; the pharmacological corollary is that (-)-UPHIT and WIN 44,441 are antagonists with selectivity for the k₁-and k₂-subtypes respectively, at least in the mouse.

Correlating genes with m-, d- and k-receptor subtypes
Although there is as yet little evidence for different genes encoding the different subtypes of m-, d- and k-receptor these subtypes may result from different post-translational modifications of the gene product (glycosylation, palmytoylation, phosphorylation, etc), from receptor dimerization to form homomeric⁴⁷ and heteromeric complexes,^{48, 49, 50} or from interaction of the gene product with associated proteins such as RAMPs.⁵¹

The Orphan Receptor

Extending the screening of genomic and cDNA libraries, perhaps in an effort to identify putative subtypes of the classical opioid receptors, resulted in the identification of a novel receptor that bore as high a degree of homology towards the classical opioid receptor types, as they shared among each other. The receptor was identified in three species: rat, mouse and man, with the degree of homology among the species variants more than 90%. Although the putative receptor has had as many names as the number of groups who reported its identification,⁵² there is some consensus for the use of the original designation for the human form, "ORL₁". Workers in the field are, however, divided in their preferred terminology for the endogenous peptide agonist for ORL₁ with both "nociceptin"⁵³ or "orphanin FQ"⁵⁴ being used with roughly equal frequency.

Although the ORL₁ receptor was accepted as a member of the "family" of opioid receptors on the basis of its structural homology towards the classical types, there is no corresponding pharmacological homology. Even non-selective ligands that exhibit uniformly high affinity towards m-, k- and d-receptors, have very low affinity for the ORL₁ receptor, and for this reason as much as for the initial absence of an endogenous ligand, the receptor was called an "orphan opioid receptor". Close comparison of the deduced amino-acid sequences of the four receptors highlights structural differences that may explain the pharmacological anomaly. Thus there are sites near the top of each of the trans-membrane regions, that are conserved in the m-, k- and d-receptors, but are altered in ORL₁. Work with site-directed mutants of ORL₁ (rat) has shown that it is possible to confer appreciable affinity on the non-selective benzomorphan bremazocine by changing Ala²¹³ in TM5 to the conserved Lys of m, k and d, or by changing the Val-Gln-Val^276-278 sequence of TM6 to the conserved Ile-His-Ile motif.⁵⁵

A splice variant of the ORL₁ receptor from rat has been reported ("XOR")⁵⁶ with a long form (XOR1L) containing an additional 28 amino acids in the third extracellular loop. In the homologous receptor from mouse (also sometimes referred to as "KOR-3") five splice variants have been reported to date.⁵⁷

ORL₁-Receptor subtypes
Selective high affinity ligands with which to attempt pharmacological definitions of the ORL₁ receptor are few in number (Table 2). Besides the natural heptadecapeptide agonist nociceptin/orphanin FQ and some closely related peptides, the only other ligands offering high affinity and selectivity belong to a class of peptides obtained by a positional scanning approach to combinatorial libraries of hexapeptides.⁵⁸ Being basic peptides highly susceptible to degradation, all of those agents are chancy tools in the hands of the unwary. So the paucity of safe and sure pharmacological tools may partly explain some of the confusion in the literature regarding the effect of nociceptin in tests of response latency to noxious stimulation; antinociception, pro-nociception/hyperalgesia, allodynia, or no overt effect, have all been reported.

Table 2. Selective opioid ligands

Receptor type	µ-Receptor	d -Receptor	k -Receptor	ORL₁
Selective agonists	endomorphin-1 endomorphin-2 DAMGO	[D-Ala²]-deltorphin I [D-Ala²]-deltorphin II DPDPE SNC 80	enadoline U-50488 U-69593	nociceptin / OFQ Ac-RYYRWK-NH₂*
Selective antagonists	CTAP	naltrindole TIPP-y ICI 174864	nor-binaltorphimine	None as yet**
Radioligands	[³H]-DAMGO	[³H]-naltrindole [³H]-pCI-DPDPE [³H]-SNC 121	[³H]-enadoline [³H]-U69593	[³H]-nociceptin

*Related combinatorial library hits are also selective agonists.⁵⁸ **Ac-RYYRIK-NH₂ has been proposed to be an ORL₁ antagonist⁶¹ whereas the putative antagonist [Phe¹y(CH₂-NH)Gly²]nociceptin(1-13)NH₂⁵⁹ appears to be a partial agonist.

Although the results of some studies have been interpreted as pointing to the existence of subtypes of ORL₁, this conclusion is so far premature in most cases. The most reliable pharmacological definition of receptors is based on differences in antagonist affinity, and in this context the absence of useful antagonists for ORL₁ is particularly galling to pharmacologists. Although the synthetic analogue of the N-terminal tridecapeptide of nociceptin, [Phe¹y(CH₂-NH)Gly²]nociceptin(1-13)NH₂ was first reported to be a selective antagonist,⁵⁹ increased use of this peptide points to it having agonist actions. There are no grounds for saying that this peptide is an antagonist at ORL₁ receptors in the periphery, but an agonist in the brain (not least because agonist actions in the periphery, and antagonist actions in the brain have been reported) and that these differences in efficacy point to differences in the receptors. Although differences in the affinity for [Phe¹y(CH₂-NH)Gly²]nociceptin(1-13)NH₂ may be found between central and peripheral sites,⁶⁰ and there may indeed be different "subtypes" of ORL₁ in the brain and periphery, the safest conclusion for the moment is just that [Phe¹y(CH₂-NH)Gly²]nociceptin(1-13)NH₂ is a partial agonist, and that the observed differences in efficacy are consistent with differences in receptor reserve.

Very recently a peptide related to the combinatorial hexapeptide library hit acetyl-Arg-Tyr-Tyr-Arg-Trp-Lys-NH₂ (Ac-RYYRWK-NH₂; Table 2), but with isoleucine substituting for tryptophan, was reported to block the effects of nociceptin/orphanin FQ in rat cortex (stimulation of GTPg³⁵S binding) or heart (positive chronotropic effect in isolated myocytes). Although this peptide, like all of its structural homologues, was originally reported to be a potent agonist, but with somewhat less than full efficacy,⁵⁸ it will be important to see if the antagonist activity of Ac-Arg-Tyr-Tyr-Arg-Ile-Lys-NH₂ (Ac-RYYRIK-NH₂ ) at the ORL₁ receptor⁶¹ is confirmed.

Less Well-Characterised Opioid Receptors

In addition to the µ-, d-, k- and ORL₁-receptors, several other types of opioid receptor have been postulated. Since the contractions of the isolated vas deferens of the rat are much more sensitive to inhibition by b-endorphin than by other opioid peptides, it was suggested that this tissue contains a novel type of opioid receptor, the e-receptor, that is specific for b-endorphin.⁶² The rabbit ileum has been proposed to possess i-receptors, for which the enkephalins have high affinity but which are distinct from d-receptors.⁶³ A very labile l-binding site with high affinity for 4,5 epoxymorphinans has been found in freshly-prepared rat membrane fragments64 and there is evidence that opioids inhibit growth in S20Y murine blastoma cells by an action at yet another receptor type called the z-receptor.⁶⁵ The e-, l-, i-and z-receptors are poorly characterised and wider acceptance of their existence awaits further experimental evidence, in particular isolation of their cDNAs.

Although originally classified as such, the s-receptor appears not to be an opioid receptor but rather the target for another class of abused drugs, phencyclidine (PCP) and its analogues.⁶⁶ Phencyclidine is an effective blocker of the ion channel associated with the N-methyl-D-aspartate (NMDA) receptor where it binds to the same site as MK 801.⁶⁷

Endogenous Ligands

In mammals the endogenous opioid peptides are mainly derived from four precursors: pro-opiomelanocortin, pro-enkephalin, pro-dynorphin and pro-nociceptin/orphanin FQ^{.68, 69, 70, 53} Nociceptin/orphanin FQ is processed from pro-nociceptin/orphanin FQ and is the endogenous ligand for the ORL₁-receptor; it has little affinity for the µ-, d- and k-receptors.^{53, 54} The amino acid sequence of nociceptin/orphanin FQ has homology with other opioid peptides especially the prodynorphin fragment dynorphin A (Table 3), suggesting a close evolutionary relationship between the precursors. Nociceptin/orphanin FQ, however, has a C-terminal phenylalanine (F) whereas peptides derived from the other precursors all have the pentapeptide sequence TyrGlyGlyPheMet/Leu (YGGFM/L) at their N-termini. These peptides vary in their affinity for µ, d- and k-receptors, and have negligible affinity for ORL₁-receptors, but none binds exclusively to one opioid receptor type.^{71, 53} b-endorphin is equiactive at µ-and d-receptors with much lower affinity for k-receptors; the post-translational product, N-acetyl-b-endorphin, has very low affinity for any of the opioid receptors.^{72, 73} [Met]- and [Leu]enkephalin have high affinities for d-receptors, ten-fold lower affinities for µ-receptors and negligible affinity for k-receptors.⁷¹ Other products of processing of pro-enkephalin, which are N-terminal extensions of [Met]enkephalin, have a decreased preference for the d-receptor with some products, e.g. metorphamide displaying highest affinity for the µ-receptor.⁷¹ The opioid fragments of pro-dynorphin, particularly dynorphin A and dynorphin B, have high affinity for k-receptors but also have significant affinity for µ- and d-receptors.⁷¹

Endomorphin-1 and endomorphin-2 are putative products of an as yet unidentified precursor, that have been proposed to be the endogenous ligands for the µ-receptor where they are highly selective.⁷⁴ The endomorphins are amidated tetrapeptides and are structurally unrelated to the other endogenous opioid peptides (Table 3). Although the study of the cellular localisation of these peptides is at an early stage, endomorphin-2 is found in discrete regions of rat brain, some of which are known to contain high concentrations of m-receptors.⁷⁵ Endomorphin–2 is also present in primary sensory neurones and the dorsal horn of the spinal cord where it could function to modulate nociceptive input.⁷⁶

In comparison to the mainly non-selective mammalian opioid peptides (the exceptions being the endomorphins), amphibian skin contains two families of D-amino acid-containing peptides that are selective for µ- or d-receptors. Dermorphin is a µ-selective heptapeptide Tyr-D-Ala-Phe-Gly-Tyr-Pro-Ser-NH₂ without significant affinity at d- and k-receptors.⁷⁷ In contrast, the deltorphins - deltorphin (dermenkephalin; Tyr-D-Met-Phe-His-Leu-Met-Asp-NH₂), [D-Ala²]-deltorphin I and [D-Ala²]-deltorphin II (Tyr-D-Ala-Phe-Xaa-Val-Val-Gly-NH₂, where Xaa is Asp or Glu respectively) - are highly selective for d-opioid receptors.⁷⁸

Table 3. Mammalian endogenous opioid ligands

Precursor	Endogenous peptide	Amino acid sequence
Pro-opiomelanocortin	b-Endorphin	YGGFMTSEKSQTPLVTL- FKNAIIKNAYKKGE
Pro-enkephalin	[Met]enkephalin [Leu]enkephalin Metorphamide	YGGFM YGGFL YGGFMRF YGGFMRGL YGGFMRRV-NH₂
Pro-dynorphin	Dynorphin A Dynorphin A(1-8) Dynorphin B a-neoendorphin b-neoendorphin	YGGFLRRIRPKLKWDNQ YGGFLRRI YGGFLRRQFKVVT YGGFLRKYPK YGGFLRKYP
Pro-nociceptin / OFQ	Nociceptin	FGGFTGARKSARKLANQ
Pro-endomorphin*	Endomorphin-1 Endomorphin-2	YPWF-NH₂ YPFF-NH₂

*Presumed to exist, awaiting discovery

Effector Mechanisms

The opioid receptor family, in common with the somatostatin receptor family, is somewhat unusual in that all of the cloned opioid receptor types belong to the G_i/G_o-coupled superfamily of receptors. Opioid receptors do not couple directly with G_s or G_q and none of the cloned receptors forms a ligand-gated ion channel. It was originally thought that m- and d-receptors coupled through G_i/G_o proteins to activate an inwardly rectifying potassium conductance and to inhibit voltage-operated calcium conductances whereas k-receptors only inhibit voltage-operated calcium conductances. However it is now known that the k-receptor is, in some cell types, also coupled to activation of an inwardly rectifying potassium conductance.⁷⁹ It seems highly likely, therefore, that all of the opioid receptors will share common effector mechanisms. Indeed, many papers have recently appeared demonstrating that the ORL₁-receptor couples to the same effector systems as the other more extensively studied opioid receptors. It should be borne in mind that, given the heterogeneity of a_i, a_o, b and g subunits which may combine to form a trimeric G protein, there may well be some subtle differences in the downstream effector mechanisms to which opioid receptors are coupled if one type of opioid receptor is unable to interact with a certain form of G_i/G_o heterotrimer. However, different responses evoked in different cell types in response to activation of different opioid receptors or even in response to activation of the same receptor are likely to reflect changes in the expression of G proteins and effector systems between cell types rather than any inherent differences in the properties of the receptors themselves.

Opioid receptor activation produces a wide array of cellular responses (Table 4). Although the pertussis toxin sensitivity has not been assessed in all instances it is highly likely that in each the first step is activation of G_i or G_o . The functional significance of many of these opioid receptor-mediated effects is still unclear, but two recent observations on changes in neurotransmitter release following acute and chronic exposure to opioids are worthy of special mention because they provide potential solutions to long-asked questions.

Table 4. Opioid receptor-evoked cellular responses

Direct G-protein bg or a subunit-mediated effects
• activation of an inwardly rectifying potassium channel
• inhibition of voltage operated calcium channels (N, P, Q and R type)
• inhibition of adenylyl cyclase

Responses of unknown intermediate mechanism
• activation of PLA₂
• activation of PLC b (possibly direct G protein bg subunit activation)
• activation of MAPKinase
• activation of large conductance calcium-activated potassium channels
• activation of L type voltage operated calcium channels
• inhibition of T type voltage operated calcium channels
• direct inhibition of transmitter exocytosis

Responses which are a consequence of opioid-evoked changes in other effector pathways
• activation of voltage-sensitive potassium channels (activation of PLA₂)
• inhibition of M channels (activation of PLA₂)
• inhibition of the hyperpolarisation-activated cation channel (Ih) (reduction in cAMP levels following inhibition of adenylyl cyclase)
• elevation of intracellular free calcium levels (activation of PLCb, activation of L type voltage operated calcium conductance)
• potentiation of NMDA currents (activation of protein kinase C)
• inhibition of transmitter release (inhibition of adenylyl cyclase, activation of potassium channels and inhibition of voltage operated calcium channels)
• decreases in neuronal excitability (activation of potassium channels)
• increases in neuronal firing rate (inhibition of inhibitory transmitter release - disinhibition)
• changes in gene expression (long-term changes in adenylyl cyclase activity, elevation of intracellular calcium levels, activation of cAMP response element binding protein (CREB))

The periaqueductal grey region (PAG) is a major anatomical locus for opioid activation of descending inhibitory pathways to the spinal cord and is thus an important site for m-receptor-induced analgesia. Opioids do not excite descending fibres directly but disinhibit them by inhibiting spontaneous GABA release from local GABAergic interneurones.⁸⁰ This inhibition of transmitter release results from activation of a dendrotoxin-sensitive, voltage-sensitive potassium conductance. The mechanism by which the voltage-sensitive potassium conductance is activated appears to be through activation of phospholipase A₂ (PLA₂) with subsequent metabolism of arachidonic acid along the 12¢-lipoxygenase pathway because the inhibition of GABA release can be inhibited by quinacrine and 4-bromo-phenacylbromide, inhibitors of PLA₂, and by baicalein, an inhibitor of 12¢-lipoxygenase. This proposed mechanism of opioid action also explains the synergy between opioids and non-steroidal analgesic drugs (NSAIDs) in producing analgesia because in the presence of a NSAID, with the cyclo-oxygenase enzymes inhibited, more of the arachidonic acid produced by opioid activation of PLA₂ can be diverted down the 12¢-lipoxygenase pathway.

The cellular locus of opiate withdrawal has long been the Holy Grail of opioid biologists. Over 20 years ago, it was shown that following chronic exposure of NG108-15 neuroblastoma x glioma hybrid cells to opiates, withdrawal resulted in a rebound increase in adenylyl cyclase;⁸¹ the functional significance of this observation for opiate withdrawal in brain neurones has remained obscure. Recently, Williams and colleagues⁸² have observed an increase in the release of the inhibitory neurotransmitter GABA, in the nucleus accumbens during opiate withdrawal. This effect could be mimicked by the adenylyl cyclase activator, forskolin, and inhibited by protein kinase A inhibitors. Therefore, as proposed over 25 years ago by the late Harry Collier, rebound adenylyl cyclase activity in withdrawal may be the fundamental step in eliciting the withdrawal behaviour.⁸³

Development and Clinical Applications of Opioid Ligands

Among the receptors for the many neuropeptides that exist in the nervous system, the opioid receptors are unique in that there existed before the discovery of the natural agonists, an abundance of non-peptide ligands with which the pharmacology of the receptors was already defined. In current terms relating to the drug-discovery process, we would consider the 4,5-epoxy-methylmorphinan opioid alkaloids morphine, codeine and thebaine as "natural-product hits" on which were based chemical programmes to design analogues with improved pharmacology (Figure 1). The effects of morphine to reduce sensitivity to pain or to inhibit intestinal motility and secretion, have continued to be exploited clinically, however the presence of other undesirable effects (e.g. depression of respiration, tolerance/dependence, effects on mood) provided the stimulus to seek analogues that were selective in producing analgesia. Thus a semi-synthetic di-acetylated analogue of morphine was introduced in the 19th century in the mistaken belief that this compound (heroin) had those desired properties. More radical changes to the morphinan nucleus were subsequently explored in various synthetic programmes, in many early cases resulting in the development of low efficacy partial agonists.

With the benefit of hindsight, it is possible to conceive an evolution of those opioid analogues, with a progressive simplification of chemical structure from the epoxymorphinans (nalorphine, nalbuphine) through the morphinans such as levorphanol, and the benzomorphans such as pentazocine, to the phenyl-piperidines including pethidine and the 4-anilino-piperidines as exemplified by fentanyl (Figure 1). The ultimate simplification of the morphine structure was in the methadone class, with methadone itself and d-propoxyphene (Darvon). Although thebaine is virtually inactive, the compound itself was an important chemical precursor in the synthesis of 14-hydroxy derivatives of morphine, most particularly the antagonists naloxone and naltrexone. Also derived from thebaine were the oripavine derivatives, and here the trend of chemical "simplification" was reversed with the introduction of an additional six-membered ring that appeared to enhance biological potency. For example, etorphine is about one thousand times more potent than morphine as an analgesic, but its use is limited to veterinary medicine as a sedative for large animals.

For the most part, such compounds have highest affinity for the m-receptor, and to a greater or lesser extent produce the full panoply of effects, good and bad, obtained with morphine. Depending on the level of affinity and efficacy, such compounds have been used acutely or chronically, to provide analgesia in cases of mild, through moderate to severe pain, alone or with adjuncts. The piperidines related to fentanyl include the most potent non-peptide m-agonists known, and are generally used peri-operatively, often for the induction and maintenance of anaesthesia. The use of many of the benzomorphans (as had been found with the first of the "duallists" nalorphine) has been associated with dysphoric and psychotomimetic effects in man, a property originally thought to be attributable to affinity at the non-opioid s-site.

The attractiveness of the prospect for development of selective k-agonists as analgesics was based on the preclinical pharmacology in animals of the 6,7-benzomorphans such as ketazocine and its derivatives (Figure 1). Although those agents are not selective in terms of affinity, their utility as pharmacological tools is based on their functional selectivity for the k-receptor, where their efficacy is high. Such agents produced a powerful antinociceptive effect, but did not substitute for morphine in dependent animals. A full biochemical and pharmacological characterisation of the k-receptor was not possible until the discovery of highly selective agonists in the aryl-acetamides that appear unrelated structurally to any of the morphine derivatives. The first compound of this class was U-50,488, but its importance was also as a chemical lead for the attempted design of related compounds of greater selectivity and potency. At least two such compounds have entered clinical trials as centrally acting analgesics, Spiradoline (U-62,066) and enadoline (CI-977). Although CNS-mediated, mechanism-related side effects of sedation and dysphoria may limit the potential for development of such compounds, the prospects for analogues with limited brain penetration to produce a peripherally mediated analgesic effect in inflammatory conditions is under exploration, with at least one compound (asimadoline, EMD-61753) in clinical trials for osteoarthritis. The observation of neuroprotective properties of k-agonists in pre-clinical models of cerebral ischaemia has lead to consideration of the possible clinical development of selective k-agonists for stroke or traumatic head injury. In this context the sedative properties of k-agonists, and even perhaps their characteristic diuretic action, may be advantageous.

The discovery of the enkephalins and of the d-receptor, led to the idea that the peptides themselves might be taken as "leads" for the synthesis of a new class of opioid agonist that lacked the addictive properties of morphine. Although such synthetic activities produced many useful experimental tools, no direct benefit in the form of a drug appeared, in spite of the attempted development of several enkephalin analogues. It did become clear from the work of a number of laboratories that activation of the d-receptor is associated with antinociception in animals, and the development of a selective non-peptide agonist is under consideration by a number of commercial drug houses. In some cases the synthetic strategy is based directly on structural considerations of the first non-peptide with significant selectivity, the 6,7-indole analogue of naltrexone, naltrindole.⁸⁴ Applying the "message-address" concept that produced the antagonist naltrindole to a novel series of octahydroisoquinoline derivatives has been successful in producing non-peptide d-selective agonists TAN-67⁸⁵ or SB 213698.⁸⁶ Similar considerations do not serve to explain the existence of another series of novel piperazine derivatives d-agonists, BW 373U86⁸⁷ or SNC 80.⁸⁸ Preclinical studies suggest that d-agonists may have a superior profile as analgesics, but this will only be established when such an agent is successfully introduced into clinical investigation; other possible applications of selective ligands for this receptor may emerge from clinical experience.

The prospects for clinical utilities of agonists or antagonists for the ORL₁ receptor can only be the subject of speculation. Elucidation of the role of the nociceptin/ORL-receptor system in pain control (and in other areas, for the peptide and its receptor have a dense and wide investment in the nervous system) must await the initial results of the drug-discovery process. Only with the availability of non-peptide selective agonists, and perhaps more particularly antagonists, will it be possible to undertake the definitive pre-clinical studies that will serve for the identification of possible clinical targets. There is some agreement that activation of the ORL₁ receptor in the brain leads to a motor impairment, so it may be that the development of ORL₁ agonists would be difficult.

References

1. Beckett and Casey (1954) J.Pharm.Pharmacol. 6 986.
2. Martin (1967) Pharmacol.Rev. 19 463.
3. Martin et al (1976) J.Pharmacol.Exp.Ther. 197 517.
4. Lord et al (1977) Nature 267 495.
5. Evans et al (1992) Science 258 1952.
6. Kieffer et al (1992) Proc.Natl.Acad.Sci.USA 89 12048.
7. Chen et al (1993) Mol.Pharmacol. 44 8.
8. Minami et al (1993) FEBS Lett. 329 291.
9. Mollereau et al (1994) FEBS Lett. 341 33.
10. Koch et al (1998) N.S. Archives of Pharmacology 357 SS 44.
11. Matthes et al (1996) Nature 383 818.
12. Wolozin and Pasternak (1981) Proc.Natl.Acad.Sci.USA 78 6181.
13. Ling et al (1984) Science 226 462.
14. Ling et al (1985) J.Pharmacol.Exp.Ther. 232 149.
15. Rossi et al (1996) Neuroscience Letters 216 1.
16. Brown et al (1997) J.Pharmacol.Exp.Ther. 282 1291.
17. Schuller et al (1999) Nature Neuroscience 2 151.
18. Jiang et al (1991) J.Pharmacol.Exp.Ther. 257 1069.
19. Sofuoglu et al (1993) Life Sci. 52 769.
20. Sofuoglu et al (1991) J.Pharmacol.Exp.Ther. 257 676.
21. Mattia et al (1991) J.Pharmacol.Exp.Ther. 258 583.
22. Wild et al (1991) Eur.J.Pharmacol. 193 135.
23. Vanderah et al (1994) Eur.J.Pharmacol. 252 133.
24. Buzas et al (1994) Life Sci. 54 PL101.
25. Noble and Cox (1995) J.Neurochem. 65 125.
26. Tang et al (1994) J.Pharmacol.Exp.Ther. 270 40.
27. Polastron et al (1994) J.Neurochem. 62 898.
28. Ho et al (1997) Eur.J.Pharmacol. 319 109.
29. Connor et al (1997) Neuropharmacology 36 125.
30. Toll et al (1994) Regul.Peptides 54 303.
31. Law et al (1994) J.Pharmacol.Exp.Ther. 271 1686.
32. Rothman et al (1993) In: Handbook Exp.Pharmacol. Ed. A. Herz 104/1 p217.
33. Xu et al (1993) Peptides 14 893.
34. Kosterlitz et al (1981) Br.J.Pharmacol. 73 939.
35. Attali et al (1982) Neuropeptides 3 53.
36. Chang et al (1984) Mol.Pharmacol. 26 484.
37. Castanas et al (1985) J.Neurochem. 45 688.
38. Zukin et al (1988) Proc.Natl.Acad.Sci.USA 85 4061.
39. Tiberi and Magnan (1989) Can.J.Physiol.Pharmacol. 67 1336.
40. Clark et al (1989) J.Pharmacol.Exp.Ther. 251 461.
41. Paul et al (1991) J.Pharmacol.Exp.Ther. 257 1.
42. Rothman et al. (1990) Peptides 11 311.
43. Lai et al (1994) Neuroreport 5 2161.
44. Ni et al (1995) Peptides 16 1083.
45. Horan et al (1991) J.Pharmacol.Exp.Ther. 257 1154
46. Horan et al (1993) J.Pharmacol.Exp.Ther. 266 926.
47. Cvejic and Devi (1997) J.Biol.Chem. 272 26959.
48. Jones et al (1998) Nature 396 674.
49. White et al (1998) Nature 396 679.
50. Kaupman et al (1998) 396 683.
51. McLatchie et al (1998) Nature 393 333-339.
52. Henderson and McKnight (1997) TiPS 18 293.
53. Meunier et al (1995) Nature 377 532.
54. Reinscheid et al (1995) Science 270 792.
55. Meng et al (1996) J.Biol.Chem. 271 32016.
56. Wang et al (1994) FEBS Lett. 348 75.
57. Pan et al (1998) FEBS Lett. 435 65.
58. Dooley et al (1997) J.Pharmacol.Exp.Ther. 283 735.
59. Guerrini et al (1998) Br.J.Pharmacol. 123 163.
60. Mason et al (1998) Soc. for Neurosci. 24 536.20.
61. Berger et al (1999) Br.J.Pharmacol. 126 555.
62. Wuster et al (1978) Neurosci.Lett. 15 193.
63. Oka (1980) Br.J.Pharmacol. 68 198.
64. Grevel and Sadee (1983) Science 221 1198.
65. Zagon et al (1989) Brain Res. 482 297.
66. Vincent et al (1979) Proc.Natl.Acad.Sci.USA 76 4578.
67. Hollmann and Heinemann (1994) Ann.Rev.Neurosci. 17 31.
68. Nakanishi et al (1979) Nature 278 423.
69. Kakidani et al (1982) Nature 298 245.
70. Noda et al (1982) Nature 295 202.
71. Corbett et al (1993) In: Handbook Exp.Pharmacol. Ed. A. Herz 104/1 p645.
72. Kosterlitz and Paterson (1985) Philos.Trans.R.Soc.Lond. 308 291.
73. Akil et al (1981) Peptides 2 289.
74. Zadina et al (1997) Nature 386 499.
75. Schreff et al (1998) Neuroreport 9 1031.
76. Martin-Schild et al (1997) Peptides 18 1641.
77. Amiche et al (1988) Int.J.Pept.Protein Res. 32 50.
78. Erspamer et al (1989) Proc.Natl.Acad.Sci.USA 86 5188.
79. Grudt and Williams (1993) Proc.Natl.Acad.Sci.USA 90 11429.
80. Vaughan et al (1997) Nature 390 611.
81. Sharma et al (1975) ) Proc.Natl.Acad.Sci.USA 72 3092.
82. Chieng and Williams (1998) J.Neurosci. 18 7033.
83. Collier et al (1974) Nature 249 471.
84. Portoghese et al (1990) J.Med.Chem. 43 1714.
85. Nagase et al (1994) Jap.J.Pharamacol. 64 (suppl. 1) 35.
86. Dondio et al (1995) Analgesia 1 394.
87. Chang et al (1993) J.Pharmacol.Exp.Ther. 267 852.
88. Calderon et al (1994) J.Med.Chem. 37 2125.