The authors examined the interaction of ketamine with recombinant mu, kappa, and delta opioid receptors and recombinant orphan opioid receptors expressed in Chinese hamster ovary cells (CHO-mu, CHO-kappa, CHO-delta, and CHO(ORL1), respectively).
CHO-mu, CHO-kappa, and CHO-delta membranes were incubated with the opioid receptor radioligand [3H]diprenorphine at room temperature. Ketamine (racemic, R(-) and S(+)) was included at concentrations covering the clinical range. CHO(ORL1) membranes were incubated with [125I]Tyr(14)nociceptin and racemic ketamine at room temperature. The effects of racemic ketamine and selective opioid receptor agonists (mu: [D-Ala2, MePhe4, Gly(ol)5] enkephalin (DAMGO); kappa: spiradoline or delta: [D-pen2, D-pen5] enkephalin (DPDPE)) on forskolin-stimulated cyclic adenosine monophosphate formation also were examined. Data are mean +/- SEM.
Racemic ketamine increased the radioligand equilibrium dissociation constant for [3H]diprenorphine from 85+/-5 to 273+/-11, 91+/-6 to 154+/-16, and 372+/-15 to 855+/-42 pM in CHO-mu, CHO-kappa, and CHO-delta, respectively. The concentration of radioligand bound at saturation was unaffected. In CHO-mu and CHO-kappa cells, racemic ketamine did not slow the rate of naloxone-induced [3H]diprenorphine dissociation. Ketamine and its isomers also displaced [3H]diprenorphine binding to mu, kappa, and delta receptors in a dose-dependent manner, with pKi values for racemic ketamine of 4.38+/-0.02, 4.55+/-0.04, and 3.57+/-0.02, respectively. S(+)-ketamine was two to three times more potent than R(-)-ketamine at mu and kappa receptors. Racemic ketamine displaced [125I]Tyr(14)nociceptin with an estimated affinity constant of 0.5 mM. Racemic ketamine inhibited the formation of cyclic adenosine monophosphate (naloxone insensitive) in a dose-dependent manner (concentration producing 50% inhibition approximately 2 mM) in all cell lines, including untransfected CHO cells. Ketamine (100 microM) reversed DAMGO (mu) and spiradoline (kappa) inhibition of formation of cyclic adenosine monophosphate.
Ketamine interacts stereoselectively with recombinant mu and kappa opioid receptors.
OPIOID receptors are classified as [micro sign][small delta, Greek], [small kappa, Greek] based on the pharmacology of a range of selective agonists and antagonists. All have now been cloned and sequenced, and their activation is capable of producing analgesia. [1-3]The International Union of Pharmacology, however, has reclassified the receptors as OP1 ([small delta, Greek]), OP2 ([small kappa, Greek]), and OP3 ([micro sign]), because these receptors bind OPioids. Shortly after the cloning of the classic opioid receptors, a further “orphan” receptor was identified by several groups in areas of the brain involved in perception of pain. The “orphan” receptor did not bind classic opioids. An endogenous 17 amino acid ligand, nociceptin or orphanin FQ, was soon identified that produced analgesic, hyperalgesic, and antiopioid actions depending on the site of administration. [5-7]In this article, the “orphan” receptor is referred to as opioid receptor-like 1 (ORL1) and the endogenous peptide nociceptin.
Ketamine interacts with opioid receptors. [8-12]In vitro radioligand binding and bioassay studies have shown that ketamine interacts stereoselectively with [micro sign] and [small kappa, Greek] opioid receptors with S(+)-ketamine, being two to three times more potent than R(-)-ketamine. [9,12]In agreement with these findings, clinical observations also indicate that S(+)-ketamine is two- to threefold more potent as an analgesic than R(-)-ketamine, although this is likely to result, in part, from N-methyl-D-aspartate (NMDA) receptor antagonism. For example, Smith et al. reported that the analgesic action of ketamine could not be antagonized by naloxone microinjected into the periaqueductal gray region of the rat brain, an area rich in [micro sign] but not [small kappa, Greek] opioid receptors. In addition, microinjection of ketamine into the periaqueductal gray region did not produce analgesia but antagonized morphine analgesia. These findings suggested that ketamine analgesia is unlikely to be mediated through [micro sign] opioid receptors in the central nervous system.
We have examined in detail the interaction of ketamine and its optical isomers with opioid receptors using Chinese hamster ovary (CHO) cells expressing a homogenous population of recombinant [micro sign] opioid, [small kappa, Greek] opioid, [small delta, Greek] opioid, and ORL1 [14-18]receptors (designated CHO-[micro sign], CHO-[small kappa, Greek], CHO-[Greek small letter delta, and CHOORL1, respectively). Opioid and ORL1 receptors are negatively coupled to adenylyl cyclase and hence inhibit the formation of cyclic adenosine monophosphate (cAMP). We used the inhibition of formation of cAMP as an index of functional opioid receptor activation, and using this system, we have examined the functional consequences of interaction of ketamine with [micro sign], [small kappa, Greek], and [small delta, Greek] opioid receptors.
Materials and Methods
Racemic ketamine was purchased from Sigma (Poole, Dorset, UK). S(+)- Lot Q and R(-)- Lot U ketamine (> 90% pure) were donated from Parke Davis (Ann Arbor, MI). [(3) H]Diprenophine (specific activity, 41 Ci/mmol) and [(125) I]Tyr14nociceptin (specific activity, 2,000 Ci/mmol) were purchased from Amersham International (Bucks, UK). All tissue culture media and supplements were purchased from Life Technologies (Paisley, UK). All other drugs were purchased from Sigma or Calbiochem (Notts, UK). All other reagents were of the highest purity available.
Membrane Preparation and Cell Culture
Chinese hamster ovary [micro sign], [small kappa, Greek], and [small delta, Greek] cells, and in some studies untransfected wild-type CHO cells, were maintained in Hams F12 medium supplemented with 100 IU/ml penicillin, 100 [micro sign]g/ml streptomycin, 2.5 [micro sign]g/ml fungizone, and 10% fetal calf serum. Chinese hamster ovary cells expressing the recombinant human ORL1 receptor (CHOORL1) were maintained in Dulbecco's modified Eagle's medium:F12 (50:50) containing 5% fetal calf serum, 2 mM glutamine, 200 [micro sign]g/ml hygromycin B, and 200 [micro sign]g/ml G418. Cultures were maintained at 37[degree sign]C in 5% CO2/humidifiedair, fed every 2-3 days, and passaged every 7 days. Experiments were performed on days 5-7 after subculture. All cells were harvested for use by the addition of 0.9% saline containing HEPES (10 mM)/EDTA (0.02%). Cells were homogenized at 4[degree sign]C using a tissue Tearor (setting 5,5x30-s bursts) in 50 mM Tris HCl buffer (pH 7.4). The homogenate was centrifuged at 18,000g for 10 min, and the pellet was resuspended in Tris HCl buffer. This homogenization and centrifugation procedure was repeated twice more. Membranes of each cell line (CHO-[micro sign], passage number 13-16; CHO-[small kappa, Greek], passage number 5-8; CHO-[small delta, Greek], passage number 7-10; CHOORL1, passage number 5-10) were prepared and used fresh daily.
Theoretical Considerations in Radioligand Binding
Three main types of radioligand binding studies are commonly used to characterize drug-receptor interaction. These are saturation, displacement, and kinetic. ,Figure 1A shows a typical saturation experiment in which receptors-specific binding is calculated as the difference between total and nonspecific binding (i.e., nonreceptor binding). As the concentration of radiolabel increases, the amount of specific binding increases until saturation occurs. The concentration of radioligand bound at saturation (Bmax) is a measure of receptor density. The concentration at which half Bmaxis obtained is the radioligand equilibrium dissociation constant (Kd). Specific binding data can be linearized using a Scatchard transformation (Figure 1B) of the data, in which the x-intercept defines Bmaxand -1/slope defines Kd. The binding characteristics of an unlabeled compound can be estimated by measuring the displacement of a fixed concentration of a receptor-specific radioligand by increasing concentrations of the displacer (Figure 1C). The concentration producing 50% inhibition (IC50) is related to the affinity. The position of the displacement curve is determined by the concentration of radioligand used (i.e., more radioligand requires more displacer to produce the same degree of displacement). Values of IC50are corrected for this competing mass of radioligand using the Cheng and Prusoff Equation toyield the affinity constant (Ki). An additional estimate of drug Kdcan be made kinetically in which the rate of association and dissociation are estimated. If a drug interacts with a receptor at a site other than the radioligand binding site, then the binding of the radioligand may be influenced in an allosteric fashion. This may slow the rate of radioligand dissociation from the receptor. This type of experiment is illustrated in Figure 2B and Figure 2D.
[(3) H]Diprenorphine Binding
The binding of [(3) H]diprenorphine was performed in 1-ml volumes of Tris HCl buffer containing [tilde operator] 200 [micro sign]g of membranes at 20[degree sign]C for 90 min. Nonspecific binding was defined in the presence of 10-5M naloxone. Typically, this was < 10% in all cells around the radioligand Kd. After incubation, each sample was filtered (and washed) under vacuum through Whatman GF/B filters using a Brandel cell harvester. Filter retained radioactivity was extracted for >or= to 8 h in 4 ml scintillation fluid. In displacement studies, the interaction of racemic (3 x 10-7- 10-2M), S(+) and R(-)(3 x 10-6- 10-3M) ketamine with [micro sign], [small kappa, Greek], or [small delta, Greek] opioid receptors was determined by displacement of 0.5 nM [(3) H]diprenorphine (n = 6). Cyprodime (3 x 10-10- 10-5M), norbinaltorphimine (10-12- 10-5M), and naltrindole (10-11- 10-5M) were included as [micro sign], [small kappa, Greek], and [small delta, Greek] selective reference compounds, respectively. Saturation analysis to determine Kdand Bmaxof CHO-[micro sign], -[small kappa, Greek], or -[small delta, Greek](n = 5) in the presence and absence of racemic ketamine (100, 50, or 350 [micro sign]M, respectively) was performed using increasing concentrations of [(3) H]diprenorphine (0.03-3.00 nM). The ketamine concentrations used are about 1.5-fold the Kivalue obtained for each cell line in [(3) H]diprenorphine displacement experiments.
Effects of Ketamine on Naloxone-induced [(3) H]Diprenorphine Dissociation
To determine whether any interaction of ketamine with opioid receptors was allosteric, membranes were labeled with 0.5 nM [(3) H]diprenorphine to equilibrium (90 min); then radioligand dissociation was initiated by addition of 10 [micro sign]M naloxone. The time course for dissociation was followed by filtration at various times up to 90 min as described earlier. Naloxone-induced dissociation was determined in the presence and absence of ketamine (200 [micro sign]M for CHO-[micro sign] and 100 [micro sign]M for CHO-[small kappa, Greek]). These concentrations are higher than those used in saturation studies to maximize the chances of detecting any differences between control and ketamine-treated preparations).
[(125) I]Tyr14nociceptin Binding Assay
The effects of racemic ketamine on the binding of [(125) I]Tyr (14) nociceptin to CHOORL1cells were determined in 1-ml volumes of Tris-HCl 50 mM, MgSO45 mM, and bovine serum albumin 0.5% buffer containing 30 [micro sign]M of peptidase inhibitors; captopril, amastatin, bestatin, and phosphomaridon (to prevent nociceptin breakdown) at pH 7.4. Unlabeled nociceptin was included as a high-affinity reference compound. Membranes were incubated for 30 min at room temperature. In all studies, [tilde operator]1 pM [(125) I]Tyr14nociceptin was used in the absence and presence of increasing concentrations of racemic ketamine.
Measurement of Formation of Cyclic Adenosine Monophosphate
Whole cells (CHO-[micro sign], -[small kappa, Greek], -[small delta, Greek] and wild-type cells) were suspended in 0.3 ml Krebs/HEPES buffer, pH 7.4, and incubated in the presence of isobutylmethylxanthine (1 mM) with or without (for the basal) forskolin (1 [micro sign]M) at 37[degree sign]C for 15 min. To obtain ketamine dose-response curves for inhibition of formation of cAMP, the cells were incubated additionally with or without racemic ketamine (3 x 10-6- 10 (-2) M). Naloxone (10 [micro sign]M) was included in some experiments. To study the nature of any interaction of ketamine with [micro sign], [small kappa, Greek], [small delta, Greek], receptors, racemic ketamine was coincubated with [D-Ala2, MePhe4, Gly(ol)5] enkephalin (DAMGO; 100 nM [micro sign]-selective agonist), spiradoline (2 nM [small kappa, Greek]-selective agonist and [D-pen2, D-pen5] enkephalin [DPDPE; 5 nM [small delta, Greek]-selective in various combinations. The concentration of each selective opioid receptor agonist was approximately the IC50value for inhibition of forskolin-stimulated formation of cAMP obtained in preliminary experiments. All reactions were terminated by the addition of 20 [micro sign]l HCl (10 M), 20 [micro sign]l NaOH (10 M), and 180 [micro sign]l Tris buffer (1 M, pH 7.4). The concentration of cAMP was measured in the supernatant using a specific radioreceptor mass assay. 
Data and Statistical Analysis
All data are expressed as mean +/− SEM. Bmaxand Kdwere derived from Scatchard plots of the specific binding data. In radioligand displacement and cAMP studies, the concentration of displacer producing 50% displacement of specific binding or inhibition of cAMP formation (IC50) was obtained by computer-assisted curve fitting (GRAPHPAD-PRISM, GraphPad Software, San Diego, CA). In radioligand displacement studies, the IC50was corrected for the competing mass of [(3) H]diprenorphine/[(125) I]Tyr14Nociceptin according to Cheng and Prusoff to yield Ki. The effects of ketamine on naloxone-induced [(3) H]diprenorphrine dissociation were analyzed using a one-phase exponential decay (GRAPHPAD-PRISM). Statistical analysis was performed by analysis of variance and unpaired t test as appropriate. A probability value < 0.05 was considered statistically significant.
[(3) H]Diprenorphine and [(125) I]Tyr14nociceptin Binding
The specific binding of [(3) H]diprenorphine to membranes from each cell line was dose-dependent and saturable. Scatchard analysis revealed that ketamine significantly increased the Kdin CHO-[micro sign], -[small kappa, Greek], and -[small delta, Greek] cells, with B (max) remaining unchanged (Figure 2A and Figure 2C and Table 1). Two hundred micromoles per liter of ketamine (and 100 [micro sign]M, data not shown) in CHO-[micro sign] and 100 [micro sign]M (and 50 [micro sign]M, data not shown) in CHO-[small kappa, Greek] did not affect the dissociation of [(3) H]diprenorphine (Figure 2B and Figure 2D). Cyprodine, norbinaltorphimine, and naltrindole displaced [(3) H]diprenorphine binding with high affinity in CHO-[micro sign], -[small kappa, Greek], or -[small delta, Greek] cells with pKi(mean in nM) values of 7.89 +/− 0.04 (12.9), 9.74 +/− 0.05 (0.18), and 9.61 +/− 0.03 (0.24), respectively (Figure 3A, Figure 3B, and Figure 3C). Ketamine and its isomers also displaced [(3) H]diprenorphine binding in a dose-dependent manner at [micro sign], [small kappa, Greek], or [small delta, Greek] receptors, with S(+)-ketamine being significantly more potent than R(-)-ketamine (Figure 3A, Figure 3B, Figure 3C, and Table 2) and with the rank order potency [small kappa, Greek] > [micro sign] > [small delta, Greek](Figure 3D and Table 2). Unlabeled nociceptin and racemic ketamine displaced [(125) I]Tyr14nociceptin in CHOORL1cells with estimated pKivalues of 9.80 +/− 0.01 (0.16 nM) and 3.30 +/− 0.05 (0.5 mM) respectively (Figure 4).
Inhibition of Forskolin-stimulated Formation of Cyclic Adenosine Monophosphate
Ketamine (in a dose-dependent manner)(Figure 5A) and naloxone (irreversibly)(Figure 5B) inhibited forskolin-stimulated formation of cAMP in all cell lines, including wild-type CHO cells. The IC50values were not significantly different among all cell lines tested (Table 3). Ketamine (1 mM) in conjunction with DAMGO (100 nM), spiradoline (2 nM), or DPDPE (5 nM) in CHO-[micro sign], -[small kappa, Greek], and -[small delta, Greek] was neither additive nor synergistic (Figure 6). Ketamine (100 [micro sign]M, a concentration that produced > 75% diprenorphine displacement in CHO-[micro sign] and CHO-[small kappa, Greek] but failed to inhibit formation of cAMP directly) reversed the inhibition of formation of cAMP by DAMGO and spiradoline (Figure 7). These studies were not performed in CHO-[small delta, Greek] cells because ketamine concentrations that produced significant diprenorphine displacement (> 300 [micro sign]M) also produced direct inhibition of cAMP.
The dissociative anesthetic agent ketamine is a noncompetitive blocker of glutamate NMDA receptors. Ketamine is administered clinically as a racemate but is composed of two isomers, S(+) and R(-), with the S(+) isomer being more potent at the NMDA receptor, and hence a more potent anesthetic agent. It is noteworthy that the stereoselectivity observed at NMDA receptors and opioid receptors is not observed for the interaction with Ca (2+) channels, and less than twofold selectivity is observed at muscarinic receptors. A wide variation in peak concentrations of ketamine in serum associated with anesthesia have been reported, ranging from 9.3-14.5 [micro sign]M [26,27]to 94.1 [micro sign]M, although these values are reduced by protein binding.
Ketamine is known to produce analgesia, and the effect is stereoselective with the S(+) isomer of ketamine being two- to threefold more potent than the R(-) isomer. Similarly, in vitro opioid receptor binding studies in isolated tissue homogenates [9,12]have shown that S(+)-ketamine interacts with [micro sign] and [small kappa, Greek] opioid receptors two- to threefold more potently than R(-)-ketamine. Stereoselectivity was not observed at the [small delta, Greek] opioid receptor. In the current study, high but nevertheless clinically achievable concentrations of ketamine [26-29]produced a stereoselective displacement of [(3) H]diprenorphine binding to recombinant [micro sign] and [small kappa, Greek] but not [small delta, Greek] opioid receptors. Supraclinical concentrations of ketamine produced a small but statistically significant stereoselective displacement at [small delta, Greek] opioid receptors. The rate of radioligand dissociation in CHO-[micro sign] and CHO-[small kappa, Greek] was not affected by coincubation with ketamine. If an allosteric interaction was present, ketamine would have slowed the rate of dissociation (K-1). Collectively, these data are consistent with simple competitive interaction.
The use of recombinant receptors in this study offers several advantages over the use of tissue homogenates, the most important being the expression of a homogenous receptor population considerably simplifying interpretation. Receptor density should be controlled carefully, however, as second messenger coupling is often expression dependent. In this study, receptor density was not vastly different from to that seen physiologically. Finck et al. suggested that ketamine analgesia may be mediated by [micro sign] or [small delta, Greek] opioid receptors, as morphine-tolerant animals, showing upregulation of [micro sign] and [small delta, Greek] receptors, are cross-tolerant to the analgesic effect of ketamine. Smith et al., however, suggested that ketamine may be a [micro sign] antagonist and a [small kappa, Greek] agonist, as microinjection of ketamine into the rat periaqueductal gray region containing [micro sign](but not [small kappa, Greek]) opioid receptors antagonized morphine morphine analgesia as effectively as naloxone. The effect of ketamine was not antagonized by naloxone. In addition, it was suggested that ketamine analgesia might result from local anesthetic action as microinjection of lignocaine into the periaqueductal gray region also antagonized morphine analgesia. Consistently, we recently reported that lignocaine displaced [(3) H]diprenorphine binding to opioid receptors. 
Opioid receptors are negatively coupled to adenylyl cyclase via a pertussis toxin-sensitive G-protein and inhibit formation of cAMP. [1-4,34]This inhibition of formation of cAMP would be expected to reduce neuronal excitability by inhibiting the hyperpolarization-activated (Ih) current. In addition to inhibiting the formation of cAMP, opioid receptors close voltage-sensitive Ca2+channels and enhance an outward K (+) conductance leading to hyperpolarization. [3,34]Opioids also exert excitatory effects in many systems, and we have reviewed this subject recently. To examine the nature of the interaction of ketamine with opioid receptors, we measured concentrations of cAMP in CHO cells exposed to ketamine and a range subtype selective agonists in various combinations. In the current study, we found that ketamine inhibited forskolin-stimulated formation of cAMP in a dose-dependent manner not only in CHO-[micro sign], -[small kappa, Greek], and [small delta, Greek] also in untransfected wild-type CHO cells with similar IC50values ([tilde operator] mM). This IC50value lay significantly to the right of diprenorphine displacement curves and was beyond the clinical range [26,29]and, therefore, is likely to be a nonspecific effect. Ketamine in conjunction with selective opioid receptor agonists produced neither and additive nor synergistic interaction. As the interaction was less than additive, we suspected that perhaps ketamine might be antagonizing the opioid-mediated inhibition of forskolin-stimulated formation of cAMP. To further test this hypothesis, we used 100 [micro sign]M ketamine in CHO-[micro sign] and CHO-[small kappa, Greek] cells (which produced > 75% displacement of [(3) H]diprenorphine but failed to inhibit formation of cAMP alone) and coincubated this with DAMGO or spiradoline. In this experimental paradigm, 100 [micro sign]M ketamine was as effective as 10 [micro sign]M naloxone in revising [micro sign] and [small kappa, Greek] receptor-mediated adenylyl cyclase inhibition. As there was significant direct adenylyl cyclase inhibition in CHO-[small delta, Greek] cells at concentrations of ketamine producing a significant displacement of [(3) H]diprenorphine, these experiments could not be performed. Any in vitro antagonism at the [small delta, Greek] receptor is likely to be of little clinical significance, as the displacement curve for [(3) H]diprenorphine binding lies outside the range of clinically relevant concentrations. [26-29]
The current study suggests that clinically achievable concentrations of ketamine interact with [micro sign] and [small kappa, Greek] but not [small delta, Greek] opioid receptors in a competitive fashion. In addition, ketamine failed to interact with the ORL1 receptor. This produces functional antagonism of opioid receptor-mediated cellular signalling. The clinical significance of these data remains to be explored fully but suggests that ketamine may be a [micro sign] and [small kappa, Greek] antagonist, which implicates a nonopioid mechanism of analgesic action for this intriguing compound.
The authors thank Drs. F. Marshall and N. Bevan (Glaxo-Welcome Research and Development, Stevenage, Herts, UK) for providing CHOORL1cells and Parke-Davis (Ann Arbor, MI) for providing R(-)- and S(+)-ketamine.