Abstract
A C6 glioma cell line stably transfected with the humankappa opioid receptor (κOR) was used to characterize receptor binding and G protein activation via the κOR by a comprehensive series of opioid ligands. The ligand-binding affinity for [3H]5α,7α,8β(-)-N-methyl-N-(7-Cl-pyrrolidinyl)-1-oxaspiro(4,5)dec-8-yl)benzene acetamide (U69593) was similar to that observed in monkey brain membranes and was 10-fold lower in the presence of sodium and GDP. Both peptide and nonpeptide agonists maximally stimulated [35S]GTPγS binding. The stimulation of [35S]GTPγS binding was blocked by pretreatment of cells with pertussis toxin. Partial stimulation of [35S]GTPγS binding via the κOR was observed for several ligands that are antagonists at the mu opioid receptor, suggesting an additional mechanism of drug action. The ability of isomers of tifluadom and levallorphan to stimulate [35S]GTPγS binding indicates that the chiral carbon of levallorphan, a benzomorphan derivative, imparts a greater degree of stereoselectivity than does the chiral carbon in the benzodiazepine derivative tifluadom. In addition, (−)tifluadom, the less potent isomer of tifluadom, which is also a γ-aminobutyric acidA receptor agonist, stimulated [35S]GTPγS binding. In contrast,d-pentazocine, (+)SKF10047, (+)cyclazocine, andd-ethylketocyclazocine displayed no agonist activity. κOR-selective antagonist norbinaltorphimine competitively inhibited the stimulation of [35S]GTPγS binding by the active isomers of ethylketocyclazocine, cyclazocine, and nalorphine to the same degree, indicating that all three ligands are eliciting an effect via the κOR. The results suggest that these cells express a homogeneous population of κOR, and that their [35S]GTPγS-binding properties make them an excellent means to assess κOR efficacy.
Opioids can differ in their ability to provoke a response once they have bound to their receptor, and these differences in efficacy have been characterized by a number of procedures. Opioid-sensitive in vivo assays, in which different intensities of a particular stimulus are used, have proven very helpful in ordering opioids with respect to their efficacies. In thermal analgesia assays, for example, low-efficacy mu agonists such as nalbuphine or buprenorphine are less able to produce analgesia at warmer temperatures than are high-efficacy agonists such as fentanyl. At warmer temperatures, where the low-efficacy agonists are ineffective, these low-efficacymu agonists are able to antagonize the analgesia produced by higher efficacy agonists (Walker et al., 1993, 1995). Fewer experiments of this type have been undertaken with kappa agonists.McClane and Martin (1967) showed that the ability of the putative high-efficacy kappa agonist, cyclazocine, or the putative low-efficacy kappa agonist, nalorphine, to depress the extensor reflex in the spinal dog depended in part on the strength of the stimulus that elicited the reflex. A sufficiently large dose of cyclazocine depressed the reflex elicited by the strongest stimulus, but the most effective dose of nalorphine was less able to depress this most intense reaction.
In vitro assay procedures have utilized measures of the maximum membrane binding of [35S]GTPγS, a nonhydrolyzable form of GTP, as an indicator of opioid efficacy at their G protein-coupled receptors. In parallel with in vivoassay systems, most in vitro measures have measured efficacy differences with the mu opioid system. In SH-SY5Y cells, in which mu receptors predominate, differences in maximum [35S]GTPγS binding produced by a series ofmu and kappa agonists were proportional to their differences in presumed efficacy (Traynor and Nahorski, 1995). Subsequently, this finding has been replicated and extended in more selective mu and delta clones (Emmerson et al., 1996; Clark et al., 1997). Liu-Chen and colleagues (Zhu et al., 1997) recently reported kappa efficacy-related differences in maximum [35S]GTPγS binding by a number ofkappa opioids in a human kappa receptor clone expressed in Chinese hamster ovary (CHO) cells. Dynorphin A 1–17, (±) ethylketocyclazocine, U50488H and β-funaltrexamine were among the drugs that exhibited high-efficacy profiles, whereas nalorphine and pentazocine produced decreasing maximum binding levels indicative of reduced efficacy. In the present study, these observations have been extended to the human kappa receptor expressed in the C6 glioma cell line. A large series of opioids that varied in affinity at the kappa receptor as well as having potential differences in efficacy at this site were evaluated. These included a set of stereoisomers with distinct positions of chirality. The aim of this study was to support and extend in the C6 cells the data provided byZhu et al. (1997) in CHO cells as well as to determine whether different positions of chirality produced distinct changes in efficacy.
Materials and Methods
Materials.
[35S]GTPγS (1300 Ci/mmol) and [3H]naloxone (53 Ci/mmol) were purchased from DuPont-NEN (Boston, MA), and [3H]5α,7α,8β(-)-N-methyl-N-(7-Cl-pyrrolidinyl)-1-oxaspiro(4,5)dec-8-yl)benzene acetamide (U69593; 60 Ci/mmol) was obtained from Amersham (Arlington Heights, IL). The cyclic AMP assay kit was purchased from Diagnostic Products Corp. (Los Angeles, CA). β-chlornaltrexamine (β-CNA) was obtained from Research Biochemicals International (Natick, MA). All other opioids and their isomers used in this study were obtained through the Opioid Basic Research Center at the University of Michigan (Ann Arbor, MI). Geneticin was purchased from Mediatech, Inc. (Herndon, VA). Pertussis toxin was purchased from List Biological Laboratories, Inc. (Campbell, CA). Dulbecco’s modified Eagle’s medium (DMEM), fetal bovine serum, Trizma, and other biochemicals were purchased from Sigma (St. Louis, MO).
Cell Culture.
The cDNA encoding the human kappaopioid receptor (κOR) in a pCDNA3 expression vector (Mansson et al., 1994) was used to stably express the receptors in C6 glioma cells. Twenty micrograms of plasmid DNA was transfected into a 100-mm dish of cells using the method of Chen and Okayama (1987). Two days after transfection, cells were maintained in tissue culture medium (DMEM and 10% fetal bovine serum) with 1 mg/ml geneticin for 14 days. After this selection period, individual cells were removed and plated in 24-well plates, maintaining antibiotic selection pressure. Stably transfected cells were able to grow in the presence of geneticin. Individual colonies were screened for opioid receptor binding and opioid-stimulated [35S]GTPγS binding. A single clone (C6κ2a) was used for this study.
Cells were grown to confluence under 5% CO2 in DMEM containing 10% fetal bovine serum and 1 mg/ml geneticin. The cells were typically subcultured at a ratio of 1:20 with partial replacement of the media on days 3 and 6, and harvested on day 7. Pertussis toxin treatment was carried out by the addition of pertussis toxin (100 ng/ml) at the time of media refreshment 18 h before harvesting.
Membrane Preparation.
Plasma membranes were prepared by lysis of cells in isotonic sucrose (Emmerson et al., 1996). Cells were washed twice with ice-cold phosphate-buffered saline (0.9% NaCl, 0.61 mM Na2HPO4, and 0.38 mM KH2PO4, pH 7.4). Cells were detached from flasks by incubation in lifting buffer (5.6 mM glucose, 5 mM KCl, 5 mM HEPES, 137 mM NaCl, and 1 mM EGTA, pH 7.4) at 37°C and pelleted by centrifugation at 200g for 3 min. The cell pellet was resuspended in 10 volumes of ice-cold 0.32 M sucrose and 1 mM Tris-HCl (pH 7.4) using a Teflon-glass dounce mounted to a Tri-R Stir-R motor at 1000 rpm. The suspension was then centrifuged for 10 min at 1000g at 4°C, and the supernatant was removed and kept on ice. The resuspension and centrifugation were repeated with the remaining pellet an additional three times, saving the supernatant from each spin in tubes kept on ice, to further break up the membranes and increase the yield. The combined supernatants were then centrifuged at 15,000g for 20 min at 4°C. After the centrifugation, the upper pellet was removed from the lower pellet by gently washing with ice-cold 0.32 M sucrose. The upper pellet was resuspended in 50 mM Tris-HCl buffer (pH 7.4) and centrifuged 20 min at 15,000g and 4°C. The final pellet was resuspended in 50 mM Tris buffer and frozen at −80°C in 0.5-ml aliquots (0.6–1.0 mg/ml).
A crude membrane preparation was prepared when clones were being screened for receptor density and agonist-stimulated [35S]GTPγS binding. A crude membrane preparation was also used for toxin-treated cells, with some loss in receptor density. Cells were collected as described above and resuspended in 10 volumes of hypotonic phosphate buffer (0.61 mM Na2HPO4, 0.38 mM KH2PO4, and 0.2 mM MgSO4, pH 7.4) by glass-glass dounce homogenization and centrifugation for 20 min at 20,000g at 4°C. The pellet was then resuspended in 50 mM Tris buffer and aliquots of 0.6 to 1 mg/ml were frozen at −80°C.
Protein Determination.
Protein concentration was determined by the method of Lowry et al. (1951) using bovine serum albumin standard. Samples were dissolved with 1 N NaOH for 30 min at room temperature before protein determination.
Receptor-Binding Assay.
Ligand binding was carried out as described previously (Fischel and Medzihradsky, 1981). In brief, the assay medium for determination of [3H]U69593 binding contained membrane protein (20 μg) diluted in Tris-Mg buffer (50 mM Tris-HCl and 5 mM MgCl2, pH 7.4), 50 μl of water or unlabeled ligand (1 μM naloxone final concentration for maximum specific displacement), and 25 μl of [3H]U69593 (0.09–7 nM) in a final volume of 525 μl. After the membranes were preincubated for 15 min at 25oC in the assay buffer, the binding was initiated by addition of unlabeled and radiolabeled ligands. After incubation for 90 min at 25°C to reach equilibrium, the samples were quickly filtered through glass fiber filters (No. 32, Schleicher & Schuell, Keene, NH) and mounted in a Brandel cell harvester (Biomedical Research and Development Laboratories, Gaithersburg, MD). Each filter was removed and placed in a 5-ml polypropylene scintillation vial with 0.4 ml of ethanol and 4 ml of Ultima Gold (Packard Instrument Co., Meriden, CT) scintillation cocktail and subjected to liquid scintillation counting.
Ligand-binding affinity in Tris-Mg buffer was determined by displacement of 0.6 nM [3H]U69593. Six concentrations of competing ligand in duplicate were included in the binding assay. Ki values were calculated from the EC50 for inhibition of the specific binding of tritiated ligand obtained from two to three experiments and analyzed using the one-site competition curve (where the top was held constant) using GraphPad Prism (San Diego, CA).
The assay medium for [3H]naloxone contained membrane protein (20–38 μg), water or unlabeled ligand (1 μM naloxone final concentration for maximum specific displacement), and [3H]naloxone (final concentration of 7.5 nM for competition experiment or 0.09–26 nM for saturation curve) in 50 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM dithiothreitol, and 50 μM GDP in a total volume of 525 μl for competition assay or 100 μl for equilibrium binding. For the determination of Ki values (7.5 nM [3H]naloxone), binding was evaluated in the presence of competing ligand as described above.
[35S]GTPγS-Binding Assay.
Agonist stimulation of [35S]GTPγS binding was measured as described by Tian et al. (1994). Membranes (5–20 μg/tube) were mixed with ligand and preincubated for 10 min at 25°C. The experiment was initiated by the addition of [35S]GTPγS to yield a final concentration in 100 μl of 50 mM Tris-HCl, 100 mM NaCl, 5 mM MgCl2, 1 mM EDTA, 1 mM dithiothreitol (added fresh), 50 μM GDP, and 50 pM [35S]GTPγS (pH 7.4). Tubes were incubated for 30 min at 25°C and the reaction was terminated by diluting the sample with 2 ml of ice-cold 50 mM Tris-HCl buffer containing 5 mM MgCl2 and 100 mM NaCl and rapidly filtering the tube contents through glass fiber filters (Schleicher & Schuell 32). The filters were then washed an additional three times with 2 ml of buffer. Filters were placed in vials containing 400 μl of ethanol and 4 ml of Econo-Safe scintillation cocktail for liquid scintillation counting. Basal activity was defined by the difference between the [35S]GTPγS binding in the absence or presence of 50 μM unlabeled GTPγS. To determine the percentage of increase in [35S]GTPγS binding over basal, binding in the presence of 50 μM unlabeled GTPγS and the basal binding was subtracted from each point, and each value was divided by the basal value and then multiplied by 100%. The experiment was performed three to four times in duplicate.
Inhibition of agonist-simulated [35S]GTPγS binding by nor-binaltorphimine (norBNI) was evaluated by addition of 1 nM antagonist at the time of agonist addition to the membranes.Ke values for norBNI inhibition was calculated by the following equation:Ke = [1 nM norBNI]/(EC50′/EC50 − 1) where EC50′ and EC50 are the concentrations of agonist in the presence and absence of norBNI that half maximally stimulated [35S]GTPγS binding.
Whole-Cell Adenylyl Cyclase Assay.
Inhibition of forskolin-stimulated adenylyl cyclase was performed as described previously (Clark et al., 1997).
Data Analysis.
Saturation binding data for [3H]U69593 and [3H]naloxone were fit to a one-site binding hyperbola using GraphPad Prism. [35S]GTPγS-binding data from three to five experiments were combined and fit to a sigmoidal curve with a variable slope using GraphPad Prism and radioligand-binding displacement curves were best fit to one-site competition curves. Kivalues were calculated as IC50/(1 + [3HL]/Kd) (Cheng and Prusoff, 1973) using 13.7 nM for the naloxone Kdvalue and 0.6 nM for the U69593 Kd value. Efficacy was calculated as the fraction of the maximum stimulation of [35S]GTPγS binding by [N-methyl-Tyr1,N-α-methyl-Arg7-d-Leu8]dynorphin A-(1–8)ethylamide (E2078).
Results
κOR Expression in C6 Glioma Cells.
The level of κOR expression was significantly lower in six different C6 glioma cells stably expressing the human κOR (C6κ) clones as compared with clones expressing either the mu opioid receptor (Emmerson et al., 1996) or the delta opioid receptor (Clark et al., 1997). Membrane preparations from these cells expressed approximately 0.1 to 1 pmol of receptor/mg membrane protein (Fig.1). The agonist-stimulated [35S]GTPγS binding was dependent on receptor density. The binding data in Fig. 2 as well as all subsequent experiments were performed in the clone expressing the greatest amount of receptor. However, the expression level in a single clone was dependent on the lot of fetal bovine serum in the cell culture media. Because opioid-stimulated GTPγS binding is dependent on receptor expression levels, all [35S]GTPγS-binding assays examining agonist efficacy were performed in membranes with approximately 2.8 pmol receptor/mg membrane protein. In control experiments, we found that the EC50 values for E2078 stimulation of [35S]GTPγS binding are similar in membranes expressing 101 to 4433 fmol receptor/mg membrane protein (data not shown), suggesting that agonist potency is independent of receptor number under the conditions of the experiments performed here.
Evaluation of Ligand-Binding Affinities.
Equilibrium binding of agonist [3H]U69593 and antagonist [3H]naloxone revealed a single population of saturable binding sites on membranes prepared from C6κ (Fig. 2).Ki values for several agonists and antagonists were determined by displacement of [3H]naloxone or [3H]U69593 (Table1). The [3H]naloxone-binding assay was performed under the same conditions as the measurements of [35S]GTPγS binding in buffer containing 100 mM NaCl and 50 μM GDP. The [3H]U69593 displacement assay was performed in the presence of 5 mM MgCl2. Seven ligands ([(trans)-3,4-dichloro-N-methyl-N-[2-(2-pyrrolidinyl)-cyclohexyl]benzeneacetamide (U50488H), dynorphin 1–17, ethylketocyclazocine (EKC), nalorphine, nalbuphine, naloxone, and norBNI) had very similar affinities for the human κOR expressed in C6 glioma cells as compared with human κOR expressed in CHO cells (Table 1; Zhu et al., 1997). We extended the evaluation of opioid affinity and efficacy at the κOR by examining several other pharmacologically interesting opioids. Inhibition of agonist binding by 100 mM sodium and 50 μM GDP was evident by an approximate 10-fold shift in binding affinity (Table 1). Several agonists (oxilorphan, nalmefene, levallorphan, cyclazocine, and β-CNA) displayed subnanomolar affinities for the human κOR, whereas agonist nalorphine had nanomolar affinity.
Determination of Potencies and Efficacies of Opioid Ligands to Stimulate [35S]GTPγS Binding.
Identical conditions were used here to evaluate κOR pharmacology as were used to examinemu (Emmerson et al., 1996) and delta (Clark et al., 1997) efficacy. The kappa agonist-stimulated [35S]GTPγS-binding dependence on the GDP concentration was identical with that observed for mu- (Emmerson et al., 1996) and delta- (Clark et al., 1997) stimulated binding (data not shown). The presence of 50 μM GDP markedly reduced basal [35S]GTPγS binding, which yielded the highest possible sensitivity to agonist stimulation. Maximal stimulation of [35S]GTPγS binding was significantly lower here (188 ± 6%) (Fig. 3 and Table 1) compared with the mu and delta opioid receptors expressed in C6 glioma which increased [35S]GTPγS binding by 300 and 600%, respectively (Emmerson et al., 1996; Clark et al., 1997). As observed in C6μ and C6δ cell membranes, agonist potency, as measured by EC50to stimulate [35S]GTPγS binding, did not correlate with efficacy. Efficacy was measured by the ratio of maximal stimulation of [35S]GTPγS binding for the ligand in question to the maximal stimulation produced by E2078. E2078, bremazocine, EKC, U50488H, and U69593 are all full agonists at the κOR with E2078, a hydrolysis-resistant dynorphin derivative, slightly more efficacious than all of the other compounds (Fig. 3, A and B). Both dynorphin 1–17 and 1–13 and both isomers of tifluadom stimulated [35S]GTPγS binding by approximately 150%, although the pairs had greatly differing potencies (Fig. 3B). Cyclazocine, β-CNA, levallorphan, oxilorphan, nalorphine, and nalmefene were all partial agonists whereas norBNI and naloxone possessed no agonist properties (Fig. 3C; Table 1). In control experiments, relative agonist (E2078, EKC, and cyclazocine) efficacy as well as EC50 values were independent of receptor number (femtomoles receptor per milligram of membrane protein). In preparations containing 1540, 2850, and 4433 fmol receptor/mg membrane protein, EC50 values for these three agonists did not differ and the rank order efficacy was identical where E2078 ≥ EKC > cyclazocine (data not shown).
As shown by Zhu et al. (1997), we found that agonist-stimulated [35S]GTPγS binding was eliminated by pretreatment with pertussis toxin (data not shown). In addition,kappa agonists were able to inhibit forskolin-stimulated adenylyl cyclase activity but the maximal inhibition observed by 1 μM EKC was 15 ± 4%. The inhibition was opioid-specific in that it was reversed by norBNI (data not shown); however, the low degree of inhibition made evaluation of ligand efficacy infeasible.
The Relationship between Stereoselectivity and Efficacy.
The (−) isomer of tifluadom displayed agonist properties, albeit at 32-fold lower potency than (+)-tifluadom (Fig. 3B; Table 1). The inactive isomers of EKC (10 μM d-EKC), cyclazocine (10 μM (+)-cyclazocine), and pentazocine (10 μMd-pentazocine) did not stimulate [35S]GTPγS binding (Fig.4). In addition, neither 10 nor 100 μM (+)SKF10047 was able to stimulate [35S]GTPγS binding. Although 10 μM dextrallorphan stimulated [35S]GTPγS binding by 15 ± 7%, the increase was not blocked by 1 μM norBNI.
Antagonism of the Stimulation of [35S]GTPγS Binding by NorBNI.
NorBNI is a κOR-selective potent antagonist (Portoghese et al., 1987). When 1 nM norBNI was added at the same time as agonist, it inhibited the ability of EKC, nalorphine, and cyclazocine to stimulate [35S]GTPγS binding by 16- to 18-fold (Fig. 5). Very similarKe values (0.06–0.07 nM) for inhibition of the full and two partial agonists response indicate that all three agonists are activating the same receptor.
Discussion
The purpose of this study was to examine the efficacy of several ligands at the human κOR by evaluating the magnitude of κOR-mediated stimulation of [35S]GTPγS binding. After the cloning of the human κOR (Mansson et al., 1994;Simonin et al., 1995), receptor pharmacology and function has been examined after stable (Zhu et al., 1997; Blake et al., 1997) and transient (Simonin et al., 1995) expression in a variety of cell lines and compared with the pharmacology of the mouse and rat clones (Simonin et al., 1995). Given the widespread distribution of the κOR in the human brain and spinal cord (Simonin et al., 1995), the efficacy of ligands at the human κOR is of great interest in the evaluation of compounds for potential clinical use. We have extended an initial characterization (Zhu et al., 1997) and have evaluated the efficacy of several compounds at the human κOR by examination of ligand-stimulated [35S]GTPγS binding. There appear to be no significant spare receptors in this assay in that the EC50 values for ligand-stimulated [35S]GTPγS binding are higher than the ligand affinities determined under identical assay conditions. The relatively high κOR expression level in these cells compared with brain tissue, and the lack of spare receptors, provide a robust agonist stimulation of [35S]GTPγS binding for full agonists. The observation that receptor expression roughly correlates with agonist-stimulated GTPγS binding also suggests that there are no spare receptors.
The affinity of agonist [3H]U69593 for the C6κ receptor (0.6 nM) is similar to that observed in monkey membranes (0.95 nM; Emmerson et al., 1994) as well as in membranes prepared from PC12 cells stably expressing the cloned mouse κOR (2.8 nM; Raynor et al., 1994). The naloxone affinity for the κOR is approximately 7-fold weaker than that observed in monkey brain membranes (Emmerson et al., 1994), yet it is comparable with the affinity observed for naloxone in CHO cells stably expressing the human κOR (4.5 ± 1.1 nM; Zhu et al., 1997). Similar values were obtained in cell membrane preparations from COS-1 cells transiently expressing the rat and human κOR (Meng et al., 1995; Simonin et al., 1995).
In a general comparison of our and Liu-Chen’s data (Zhu et al., 1997), ligand efficacy at the human kappa receptor is apparently independent of the cell line in which the receptor is expressed. E2078, a potentially clinically useful stable derivative of dynorphin that is able to penetrate the blood-brain barrier (Terasaki et al., 1989,1991), was most efficacious in increasing [35S]GTPγS binding by 188 ± 6%, whereas several peptide and nonpeptide kappa agonists (dynorphin 1–17, dynorphin 1–13, bremazocine, EKC, U50488H, U69593, and tifluadom isomers) displayed high efficacy (∼ 150% stimulation of [35S]GTPγS binding). Of these agonists, bremazocine, EKC, U69593, and U50488H were full agonists when tested for analgesic activity by measuring the latency for monkeys to remove their tails from a thermos containing 55° C water (France et al., 1994). In addition, the binding Kivalues for bremazocine, EKC, U50488, and nalorphine in the C6κ cells correlate well with the dose required to produce a half-maximal EKC-discriminative stimulus effects in monkeys (France et al., 1994). Surprisingly, both dynorphin 1–17 and 1–13 had similar efficacies in the stimulation of [35S]GTPγS binding although dynorphin 1–13 was approximately 30-fold less potent.
Based on the limited ligands selective for putative κOR subtypes, the human clone appears to be similar to thekappa1 subtype. The C6κ cells are U69593 sensitive; U69593 binds selectively, and with high affinity, to thekappa1 but notkappa2 site (Zukin et al., 1988). Nalorphine has been characterized as akappa3 analgesic. Animals tolerant to U50488H were not tolerant to nalorphine (Paul et al., 1991). In the C6κ cells, we found that norBNI was equally efficacious to block EKC, cyclazocine, and nalorphine-stimulated [35S]GTPγS binding, suggesting that although nalorphine may interact with the kappa3 opioid receptor, in this cell line, it is functional at thekappa1 opioid receptor subtype.
By suppression of basal [35S]GTPγS binding by excess GDP, we were able to evaluate partial agonist efficacy. Efficacy differences at the mu opioid receptor were magnified by increasing GDP concentrations, indicating that the activity state of G proteins can affect agonist efficacy (Selley et al., 1997). Several of the partial agonists have been previously characterized asmu opioid receptor antagonists. For example, nalmefene is marketed as a long-acting mu-selective antagonist with no efficacy. However, monkey discriminative effects at the κOR have been observed (Woods et al., 1986) in addition to high affinity for both the monkey mu and kappa opioid receptor (0.13 and 0.28 nM, respectively (Emmerson et al., 1994)). The data presented here suggest that nalmefene is a potent yet weak partial agonist and may exhibit its discriminative effects via the κOR.
Natural opioids are levorotatory, whereas several of the synthetic opioids are racemic. The benzomorphan dextrorotatory (+)-enantiomers do not possess opioid activity (for review, see Musacchio, 1990). The ability of isomers of tifluadom and levallorphan to stimulate [35S]GTPγS binding indicates that the chiral carbon of levallorphan, a benzomorphan derivative, imparts a greater degree of stereoselectivity than does the chiral carbon in the benzodiazepine derivative tifluadom. In addition (−)-tifluadom, the less potent isomer of tifluadom, which is also a GABAA agonist, stimulated [35S]GTPγS binding. In contrast,d-pentazocine, (+)SKF10047, (+)cyclazocine, and (d)EKC displayed no agonist activity. The (+) isomer of SKF10047 is a putative sigma receptor ligand and anN-methyl-d-aspartate receptor noncompetitive antagonist.
One striking difference between the results obtained here and those in the CHO cells is the sensitivity to sodium and GDP observed here. Despite the observation that several ligands had similar efficacies and affinities (in the presence of GDP and sodium) for the human κOR expressed in C6 glioma cells as compared to human κOR expressed in CHO cells, we found the C6κ receptor to be sensitive to sodium and GDP whereas the CHO cell-expressed human κOR was not (Zhu et al., 1997). Relative lack of κOR sensitivity to sodium compared to themu opioid receptor was also described in monkey brain membrane preparations (Emmerson et al., 1994); however, ligand binding in guinea pig cerebellar membranes was strongly inhibited by NaCl and a nonhydrolyzable guanine nucleotide analog (Gairin et al., 1989). In addition, agonist binding to the κOR (kappa1 subtype) expressed in the mouse R1.1 lymphoma cell line was reduced to 30% of control binding by 30 mM NaCl and 100 μM GTP (Lawrence and Bidlack, 1992).
We observed minimal κOR-mediated inhibition of adenylyl cyclase in contrast to the results of Liu-Chen and colleagues (Zhu et al., 1997) and in contrast to stably expressed mu and deltaopioid receptors in C6 glioma cells. Differences in adenylyl cyclase inhibition in the two clonal cell lines (CHO and C6 glioma cells) may be due to a different complement of G proteins in CHO and C6 glioma cells or to the presence of an yet undescribed factor which modulates opioid receptor sodium and/or guanine nucleotide sensitivity. The isoform of adenylyl cyclase as well as the G protein subunit (alpha or beta-gamma) that mediates inhibition is unknown in the C6 glioma cells and the mechanisms of inhibition may vary between receptor types (mu, delta, andkappa). In these same clones, Gutstein et al. (1997) found that mu and delta receptor stimulation activated extracellular signal-related kinase but kappa stimulation did not.
Prather et al. (1995) found that rat κOR stably transfected in CHO cells was able to interact with multiple G proteins (Go, Gi2, and Gi3) as measured by agonist-stimulated incorporation of azidoanilido-GTP. This pattern was not unlike that observed for both delta andmu opioid receptors, indicating that receptors did not selectively couple to a single isoform of G protein. The coupling of κOR to G protein was apparently weaker, as measured by an agonist-stimulated increase in GTPγS binding over basal binding, than that observed in C6 cells expressing the mu ordelta opioid receptor (Yabaluri and Medzihradsky, 1996,Clark et al., 1997). These differences may be due to the relatively lower expression levels observed in the κOR clones. Indeed, when additional experiments were performed to evaluate agonist EC50 values over a range of receptor expression levels, the extent of kappa agonist-stimulated GTPγS-binding via the κOR was similar to that of deltaagonists acting via the δOR with similar receptor expression levels (data not shown). Lower expression levels of the κOR were also observed when all three opioid receptors were individually expressed in baculovirus-infected insect cells (Massotte et al., 1997).
By the characterization of the binding affinity and efficacy at the κOR of a wide range of opioids, the results of this study contribute to the assessment of opioid efficacy in stimulating G protein, a first step in the signal transduction cascade. An assessment of ligands that vary in efficacy at the κOR at the cellular level will promote the use of kappa opioid ligands as pharmacological tools and perhaps as potential therapeutic agents.
Footnotes
-
Send reprint requests to: Ann E. Remmers. Ph.D., Department of Pharmacology, 1303 MSRB III, 1150 W. Medical Center Drive, University of Michigan, Ann Arbor, Michigan 48109-0632. Email:aremmers{at}umich.edu
-
1 This work was supported by grants from the U.S. Public Health Service to F.M (DA04087), J.H.W. (DA00254), and H.A. (National Institute on Drug Abuse DA02265 and DA08920).
- Abbreviations:
- κOR
- kappa opioid receptor
- C6κ
- C6 glioma cells stably expressing the human kappaopioid receptor
- E2078
- [N-methyl-Tyr1,N-α-methyl-Arg7-d-Leu8]dynorphin A-(1–8) ethylamide
- EKC
- ethylketocyclazocine
- U69593
- 5α,7α,8β(−)-N-methyl-N-(7-Cl-pyrrolidinyl)-1-oxaspiro(4,5)dec-8-yl)benzene acetamide
- U50488H
- (trans)-3,4-dichloro-N-methyl-N-[2-(1-pyrrolidinyl)-cyclohexyl]benzeneacetamide
- β-CNA
- β-chlornaltrexamine
- norBNI
- nor-binaltorphimine
- GTPγS
- guanosine-5′-O-(3-thio)triphosphate
- DMEM
- Dulbecco’s modified Eagle’s medium
- CHO
- Chinese hamster ovary
- Received April 14, 1998.
- Accepted September 9, 1998.
- The American Society for Pharmacology and Experimental Therapeutics