Meperidine Exerts Agonist Activity at the α^2B-Adrenoceptor Subtype

All reagents were obtained from Sigma (St. Louis, MO) except bovine calf serum (Gemini Bio-Products, Calabasas, CA), gentamicin (Boehringer Mannheim, Indianapolis, IN), pCDNA-3 (Invitrogen, Carlsbad. CA), [³H]RX821002 (Amersham, Piscataway, NJ), RX821002 (RBI, Natick, MA), and meperidine (AstraZeneca, Wilmington, DE).

Genes encoding human α^2A(or C10, 7α^2B[or C2]) 8and α^2C(or C4) 9adrenoceptors were cloned into the multiple cloning site of pCDNA-3 vectors. Each adrenoceptor was expressed transiently in COS-7 cells using diethylaminoethyl-dextran–mediated transfection. 10COS-7 cells were maintained in Dulbecco modified Eagle medium supplemented with 10% bovine calf serum and 25 mg/l gentamicin.

Cells were collected 3 days after transfection and rinsed twice with cold phosphate-buffered saline, and then 7 ml ice-cold lysis buffer (10 mm Tris-HCl, 1 mm EDTA, pH 7.4) was added. Cells were scraped off the flask and collected. The flask was rinsed with an additional 3 ml lysis buffer, and the washings were pooled. Cells were homogenized by four 8-s bursts at full speed using a Polytron homogenizer. The nuclei were pelleted by centrifugation at 220 g  for 5 min at 4°C, and the supernatant was centrifuged at 30,000 g  for 30 min at 4°C. The pelleted membranes were resuspended in an appropriate volume of binding buffer (75 mm Tris-HCl, 12.5 mm MgCl², 1 mm EDTA, 0.01% CHAPS, pH 7.4). Binding experiments were performed in 500-μl volumes of buffer for 90 min at 25°C, using a nonselective α²ligand, [³H]RX821002 (1813 GBq). The bound radiolabeled ligand was separated from the free ligand by filtration through GF/C filters using a vacuum filtration manifold (Brandel Cell Harvester; Brandel, Gaithersburg, MD). Saturation binding isotherms were performed by incubating membranes with varying concentrations of the radioligand [³H]RX821002 (2.0, 13.0, and 3.2 nm for α^2A, α^2B, and α^2Cadrenoceptors, respectively), and nonspecific binding was determined by adding 10 μm rauwolscine to radiolabeled binding studies. Competition experiments were conducted by incubating membranes with varying concentrations of competing ligand (meperidine and morphine) and a constant concentration of [³H]RX821002, and nonspecific binding was determined in the presence of 10 μm rauwolscine. Equilibrium dissociation constants (K  ^dand K  ⁱvalues) and EC⁵⁰values were determined from saturation isotherms and competition curves. Saturation isotherm data were analyzed by a nonlinear least squares curve-fitting technique, and the competition data were analyzed according to the Cheng-Prusoff equation using GraphPAD Prism (GraphPAD Software Inc., San Diego, CA).

Cells were collected as for radiolabeled ligand binding of transfected cells. Membranes from cells (at least 1.5 × 10⁶for each 24-well assay) were prepared fresh just before use by homogenizing in 600 μl of ice-cold assay buffer (75 mm Tris, 25 mm MgCl², 2 mm EDTA, pH 7.4) with 15 up–down strokes in a 2-ml glass homogenizer. Membrane preparation (20 μl) was added to 20 μl of “cyclase mixture” (2.7 mm mono{cyclohexylammonium} phosphoenolpyruvate, 53 μm GTP3Na.3 H²O, 0.2 IU pyruvate kinase, 1.0 IU myokinase, 125 μm adenosine triphosphate) and 10 μl of “reaction mixture.” To determine the inhibition of forskolin-stimulated adenylyl cyclase activity by meperidine and dexmedetomidine, the reaction mixture was composed of 100 μm forskolin and varying concentrations of meperidine or dexmedetomidine. Furthermore, to investigate the effect of rauwolscine on the inhibition of forskolin-stimulated adenylyl cyclase activity by meperidine, the reaction mixture was composed of 100 μm forskolin, 0.5 μm meperidine, and 20 μm rauwolscine. After addition of the membrane preparation, the cells were shaken (220 rpm) at 37°C for 10 min. The reaction was terminated by adding 10 μl ice-cold 0.8 m perchloric acid. The samples were left at room temperature for 20 min, after which 15 μl of 2N KOH in 25 mm HEPES was added. Tubes were centrifuged for 5 min in a microcentrifuge, and the supernatant was removed and assayed for cyclic adenosine monophosphate (cAMP) by the scintillation proximity assay kit (Amersham Pharmacia Biotech, Piscataway, NJ). Reaction mixture–inhibited cAMP accumulation was expressed as a percent of forskolin-alone stimulation; these data were analyzed by a nonlinear least squares curve-fitting technique, and EC⁵⁰values were determined using GraphPAD Prism (Graph-Pad Software Inc.).

All experimental protocols were reviewed and approved by the Subcommittee for Animal Studies (Veterans Affairs Palo Alto Health Care System, Palo Alto, CA) before the initiation of work. All protocols conform to the guidelines for the study of pain in awake animals as established by the International Association for the Study of Pain. Every effort was made to minimize animal suffering and to use only the number of animals necessary to produce reliable scientific data. In all studies presented, male C57BL/6 mice aged 12–16 weeks were used (Jackson Labs, Bar Harbor, ME). These mice were kept in cages (4–6 mice per cage) with a light–dark cycle of 12 h–12 h. Food and water were provided ad libitum . Animals were weighed on the day of experimentation for the calculation of drug dosages.

Drugs were injected alone or in combination subcutaneously in a total volume of approximately 50 μl. The animals were returned to their cages for 25 min after the injection of drugs and before hot-plate assay.

The hot-plate assay was performed as described by O'Callaghan and Holtzman, 11with equipment from IITC (Woodland Hills, CA). The hot plate was set thermostatically at 54 ± 0.1°C. Mice were placed in a rectangular clear enclosure on the hot plate, and the time to licking of a hind paw, the endpoint of the assay, was measured with a stopwatch. Animals were left on the hot plate no longer than 60 s to prevent injury.

Initial transmembrane topology and secondary structure predictions were made by submitting the sequence of the human α^2Badrenoceptor to the Predict Protein Server at Columbia University (New York, NY, on August 19, 2000). The PHDhtm algorithm was used to make the topology predictions. 12The secondary structure analysis of the rhodopsin template (chain A of protein ID 1F88 in the protein database of the Research Collaboratory for Structural Bioinformatics **), and the production of an annotated template file was performed using the Kabsch and Sander 13algorithm. The Seqfold algorithm, 14available through the InsightII 2000 Modeling Environment (Molecular Simulations, Inc., San Diego, CA), was then used to align the sequence of the α^2Aadrenoceptor to that of bovine rhodopsin. This algorithm allows the alignment of a sequence of unknown 3-dimensional structure to a template of known 3-dimensional structure based on sequence similarity, as scored by the Gonnet matrix, 15as well as fold similarity. Further refinement of the sequence alignment was made by importing it into the SwissPDB Viewer 16and calculating the mean force potential of Sippl 17for several additional manual sequence threadings. Final alignment of the sequence of the α^2Breceptor to the rhodopsin template was refined based on the experimental data derived in this laboratory. The coordinates of the atoms within the α^2Breceptor were then assigned. Side chains were initially optimized using the rotamer refinement function of the SwissPDB Viewer. 16Energy optimization of the structure was performed using the Hyperchem 5.1 software package (Hypercube Inc., Gainesville, FL) with the Amber 3.0 all-atom forcefield. This was performed initially only on the side chains and then on the entire structure without restraints. Optimization was initially performed via  the steepest descents method followed by the method of Polak-Ribiere.

The initial meperidine molecule was constructed and optimized using the Builder module within InsightII 2000. A model of the α^2Badrenoceptor was constructed based on homology modeling techniques. In an attempt to understand the selective binding of meperidine to the human α^2B-adrenoceptor subtype, simulated docking experiments were performed using the Affinity module within the InsightII 2000 Modeling Environment (Molecular Simulations, Inc.).

Statistical significance was determined by using one-way analysis of variance followed by Bonferroni multiple comparison tests. The EC⁵⁰of the cAMP inhibition curves were calculated by nonlinear regression fitting to a sigmoidal dose–response model with variable slope.

Analysis of the equilibrium dissociation constants (K  ^d) of [³H]RX821002 for human α²-adrenoceptor subtypes showed that the ligand has lowest affinity for the human α^2B-adrenoceptor subtype (α^2B, 13.1 ± 1.5 nm; α^2A, 2.2 ± 0.3 nm; α^2C, 6.2 ± 0.3 nm; mean ± SD, P < 0.01, n = 4.). Meperidine competes for the binding of [³H]RX821002 for each of the human α²-adrenoceptor subtypes (fig. 1); morphine, another opioid ligand, does not displace the radiolabeled ligand off any of the human α²-adrenoceptor subtypes (fig. 2). Analysis of the inhibitory constant of meperidine showed that it has highest affinity for the human α^2B-adrenoceptor subtype (α^2B, 8.6 ± 0.3 μm; α^2C, 13.6 ± 1.5 μm, P < 0.05; α^2A, 38.6 ± 0.7 μm, P < 0.001 from the α^2B-receptor value; mean ± SD, n = 4).

Meperidine potently inhibits forskolin-stimulated adenylyl cyclase activities in cells transfected with the human α^2B-adrenoceptor subtype over the same concentration range as its analgesic activity (0.63 μm). The inhibitory action of meperidine on forskolin-stimulated adenylyl cyclase activity at the α^2Breceptor, as defined as the EC⁵⁰that achieved half maximal inhibition, is approximately 1,000-fold less than is required to reach the same level of inhibition with the other receptor subtypes (α^2A, 1.3 mm; α^2C, 0.3 mm;fig. 3). In fact, meperidine is as efficacious as dexmedetomidine, the highly selective and potent α²agonist, at inhibiting forskolin-stimulated adenylyl cyclase activities in cells transfected with the α^2B-adrenoceptor subtype (fig. 4). That meperidine exerts its inhibitory action through α²adrenoceptors itself is evident from the fact that the inhibitory action of meperidine (50 μm) on α^2B-mediated inhibition of forskolin-stimulated cAMP production is blocked by the α²-receptor antagonist rauwolscine (20 μm) (meperidine, 73.8 ± 5.1%; meperidine plus rauwolscine, 101 ± 2.5% of control level; mean ± SD, P < 0.01, n = 4).

The data in figure 5demonstrate that mice treated with 50 mg/kg meperidine had prolonged latencies for hind-paw licking when compared with saline-injected animals. Preliminary experiments established that 50 mg/kg meperidine was an approximate ED⁵⁰for meperidine in this assay. This antinociceptive effect was completely blocked by the coadministration of 1 mg/kg naloxone, an opioid receptor antagonist. However, neither of two α²-adrenoceptor antagonists, namely, the nonimidazoline yohimbine (1 mg/kg) or the imidazoline RX 821002 (2 mg/kg), significantly reduced the meperidine-mediated analgesia. These doses of α²-adrenoceptor antagonists were previously shown to antagonize the antinociceptive effects of clonidine in mice when using the hot-plate assay. 18 

Assuming that the amino acid residues that could be affecting binding are within 0.4 nm of the ligand, there is a potential binding pocket made up of transmembrane domains 3, 4, 5, 6, and 7 (fig. 6). Further examination within this 4-Å radius reveals the amino acids in the α^2Badrenoceptor, which differ from their counterparts in the α^2Aadrenoceptor. They are LEU 145, ILE 173, SER 177, and SER 390. The corresponding residues in the α^2Aadrenoceptor are PHE 166, VAL 197, CYS 201, and THR 393, respectively, which will impose different constraints on the ability of the ligand to dock (fig. 6). The corresponding residues in the α^2Cadrenoceptor are PHE 184, ILE 211, CYS 215, and ILE 400. Another possible residue that is in close proximity to the meperidine (but just outside of the arbitrary 4-Å limit) is GLY 394 on the α^2Badrenoceptor; the corresponding residue on the other receptor subtypes is THR 397 in the α^2Aadrenoceptor and TYR 404 in the α^2Cadrenoceptor. Thus, there exist structural reasons why meperidine should have activity at one receptor subtype and not another.

The opiate receptor agonist meperidine competes directly with α²-adrenoceptor ligands for the human α²-adrenoceptor subtypes (fig. 1). Morphine, another μ-opiate receptor agonist, has no affinity for these sites at clinically relevant concentrations (fig. 2). Meperidine has highest affinity for the α^2B-adrenoceptor subtype, where it exerts comparable agonist efficacy to that seen with the highly selective α²agonist dexmedetomidine (fig. 4). This agonist action is blocked by an α²antagonist. The analgesic action of meperidine is not blocked by α²antagonists but is blocked by an opiate antagonist.

The α²-adrenoceptor agonists are either imidazoline, guanidinium, or phenylethanolamine compounds and are structurally dissimilar from the phenylpiperidine meperidine. However, there are examples of other piperidine compounds that exert their pharmacologic activity at the α²adrenoceptor. 19Interestingly, another piperidine compound causes a diuretic action via  renal α²adrenoceptors 20and is considered to be mostly, if not exclusively, of the α^2Bvariety. Furthermore, as is seen in the results of the molecular modeling, there appear to be structural differences that may grossly account for the differential, subtype-specific binding of meperidine. In particular, there are probably differences in the binding pocket size and relative affinity for meperidine in the α^2Aadrenoceptor and the α^2Cadrenoceptor because of the proximity of a phenylalanine in the binding pocket, where there is a much smaller ILE in the more favorable binding pocket of the α^2Badrenoceptor. Likewise, there is a glycine in the α^2Badrenoceptor, whereas there are the much larger and more polar THR in the α^2Aadrenoceptor and the TYR in the α^2Cadrenoceptor.

The fact that meperidine is a potent agonist at the α²adrenoceptors raises the possibility that some of its actions are transduced by this mechanism. In reviewing the pharmacologic actions of α²-adrenoceptor agonists and those of meperidine, there is correspondence for analgesia and antishivering actions. The free plasma concentration of meperidine corresponding to antishivering effect is approximately 1–3 μm. 21To reduce cAMP concentrations to 50% of maximal, a concentration of 1–3 μm meperidine is certainly sufficient. It is unknown what proportion of α^2Badrenoceptors must be occupied by meperidine to achieve an antishivering effect, but because 1 μm meperidine has significant cyclase inhibitory power at the α^2Breceptor, it is feasible that 1–3 μm would be sufficient. These clinically achieved meperidine concentrations would also be analgesic. 22We have effectively ruled out the participation of α²adrenoceptors in the analgesic action of meperidine because neither the imidazoline (RX821002) nor the nonimidazoline (yohimbine) α²-adrenoceptor antagonists attenuated its antinociceptive action. As expected, the analgesic action of meperidine was reversed by a nonselective opiate receptor antagonist.

Of the many responses transduced by α²adrenoceptors, few are transduced by the α^2B-adrenoceptor subtype. 23Of the α²-adrenoceptor–mediated actions attributed to the α^2Bsubtype, the most well substantiated is the direct vasoconstrictive response in the peripheral vasculature. 5Concerning the analgesic action of the α²agonists, earlier pharmacologic evidence suggested that only the α^2Asubtype transduced the analgesic action 24,25; however, evidence is now accumulating for a role for the other two subtypes. Prazosin-sensitive α²-adrenoceptor subtypes (either the α^2Bor α^2C) were shown to inhibit release of substance P in a spinal cord preparation. 26In addition, ST-91, a non–α^2Asubtype preferring α²agonist, was shown to induce antinociception in rats when it was administered intrathecally, an effect that was blocked by prazosin. 27–30We recently showed that the antinociceptive action of nitrous oxide was reduced by 65% in the hot-plate assay and totally absent in the tail-flick assay in mice lacking the α^2B-adrenoceptor subtype. These data therefore suggested that the α^2B-adrenoceptor subtype has a role to play in nociception. However, there are no direct data implicating one or other α²-adrenoceptor subtype for the antishivering action of α²agonists.

Postanesthetic shivering occurs through a disorder in thermoregulation partly caused by differential recovery of excitatory and inhibitory centers. Thermoregulation is composed of multiple feedback systems and neurotransmitters together with the integration of information at each of several levels, including afferent impulses, hypothalamic centers, and skeletal muscles. 31Based largely on responsiveness to pharmacologic probes, adrenergic, 32serotonergic, 3,33and cholinergic 34pathways have each been causally implicated. Neuroanatomic studies have demonstrated that noradrenergic nuclei in the locus coeruleus are interconnected with hypothalamic thermoregulatory centers 35–37; furthermore, there is evidence that α adrenoceptors mediate the decrease in temperature after central administration of norepinephrine. 38The distribution of the α^2B-adrenoceptor subtype may shed light on whether meperidine transduces its antishivering action through the α^2B-adrenoceptor subtype. In rat spinal cord and dorsal root ganglion, mRNA for the α^2A-, α^2B-, and α^2C-adrenoceptor subtypes have been identified. Within the brain, the most prevalent sites for α^2BmRNA are in the thalamus and the cerebellum. 39 

The antishivering action of meperidine has previously been attributed to its action on κ-opioid receptors. After administration of naloxone, an opioid receptor antagonist that blocks both μ receptors and κ receptors at high doses and only μ receptors when administered at low doses, a differential effect of meperidine was revealed. Only a high dose of naloxone was able to prevent the effect of meperidine on shivering; when administered in low doses, naloxone did not influence the action of meperidine, thus indicating that κ receptors and not μ receptors may mediate its antishivering action. 2However, other more κ-receptor–selective drugs, such as pentazocine, failed to inhibit postanesthetic shivering, thus suggesting that κ-opioid receptors are not implicated. 40In addition, the effect of meperidine on shivering is more effective than that of pure μ-opioid receptor agonists at equianalgesic doses. 1,41Thus, an α²-adrenoceptor–mediated mechanism of action of meperidine on postanesthetic shivering as suggested by the data presented here is conceivable. Clinically, α²-adrenoceptor agonists such as clonidine have been successfully used both for prevention and treatment of postanesthetic shivering, 3,42–45indicating that α²adrenoceptors mediate this action. Our current findings suggest that it is likely to be caused by the activation of the α^2B-adrenoceptor subtype.

The authors thank the Medical Research Council (London, UK), National Institutes of Health (Bethesda, Maryland), and the Department of Veterans Affairs (Washington, DC).