Abstract
Narcotic analgesics cause addiction by poorly understood mechanisms, involving μ opioid receptor (MOR). Previous cell culture studies have demonstrated significant basal, spontaneous MOR signaling activity, but its relevance to narcotic addiction remained unclear. In this study, we tested basal MOR-signaling activity in brain tissue from untreated and morphine-pretreated mice, in comparison to antagonist-induced withdrawal in morphine-dependent mice. Using guanosine 5′-O-(3-[35S]thio)triphosphate ([35S]GTPγS) binding and adenylyl cyclase activity assay in brain homogenates, we demonstrated that morphine pretreatment of mice enhanced basal MOR signaling in mouse brain homogenates and, moreover, caused persistent changes in the effects of naloxone and naltrexone, antagonists that elicit severe withdrawal in dependent subjects. Naloxone and naltrexone suppressed basal [35S]GTPγS binding (acting as “inverse agonists”) only after morphine pretreatment, but not in drug-naive animals. Moreover, naloxone and naltrexone stimulated adenylyl cyclase activity in striatum homogenates only after morphine pretreatment, by reversing the inhibitory effects of basal MOR activity. After cessation of morphine treatment, the time course of inverse naloxone effects on basal MOR signaling was similar to the time course of naltrexone-stimulated narcotic withdrawal over several days. The neutral antagonist 6β-naltrexol blocked MOR activation without affecting basal signaling (G protein coupling and adenylyl cyclase regulation) and also elicited substantially less severe withdrawal. These results demonstrate long-lasting regulation of basal MOR signaling as a potential factor in narcotic dependence.
G protein-coupled receptors tend to display basal signaling activity in the absence of agonists (also referred to as constitutive or spontaneous activity) (Leurs et al., 1998). Only a few studies, for example on β2-adrenergic (Maack et al., 2000) and histamine receptors (Smit et al., 1996; Morisset et al., 2000; Gbahou et al., 2003), have addressed the physiological and pharmacological relevance of basal receptor activity in vivo. Although antagonists by definition block agonist-mediated receptor activation, they can have varying effects on basally active receptors. Those suppressing basal signaling activity are referred to as inverse agonists, or antagonists with negative intrinsic activity, whereas neutral antagonists or antagonists with no intrinsic activity do not affect basal signaling (Kenakin, 2001). Inverse agonists and neutral antagonists are essential tools for studying basal receptor activity and have been identified for several G protein-coupled receptors, including opioid receptors (Costa and Herz, 1989; Wang et al., 2001b).
The μ opioid receptor (MOR), considered the main mediator of opioid analgesia and addiction (Matthes et al., 1996), exhibits basal signaling activity in tissue culture (Wang et al., 1994; Burford et al., 2000). Moreover, in vitro experiments using cells transfected with recombinant MOR indicate that basal activity of the receptor is regulated by morphine pretreatment, resulting in at least two significant changes in receptor and ligand properties. First, morphine paradoxically enhances basal MOR signaling even though maximal agonist stimulation is blunted (Wang et al., 1994, 2000, 2001a,b; Liu and Prather, 2001; Liu et al., 2001). Second, morphine pretreatment affects the antagonist properties of naloxone and naltrexone. Naloxone and naltrexone produce little effect per se in untreated cells, acting as neutral antagonists, whereas in morphine-pretreated cells, they exhibit inverse agonist effects that suppress basal MOR activity (Wang et al., 1994, 2001b; Burford et al., 2000; Liu and Prather, 2001; Liu et al., 2001). Similarly, chronic agonist treatment converts neutral antagonists into inverse agonists at δ opioid receptors in cell culture (Liu and Prather, 2002). Changes in ligand properties from neutral antagonist to inverse agonist effects also occur at other G protein-coupled receptors. Prior stimulation at the β2-adrenergic receptor, for example, has been shown to enhance the inverse agonist activity of certain antagonists (Chidiac et al., 1994). Furthermore, regulation of basal signaling of the bradykinin B2 receptor determines the efficacy of a ligand as either partial agonist or inverse agonist (Fathy et al., 1999). Similarly, protean agonism at histamine H3 receptors has been reported in vitro and in vivo, indicating that a ligand can be an agonist, neutral antagonist, or inverse agonist at the same receptor type, depending on the cell context (Gbahou et al., 2003). Therefore, the intrinsic efficacy of a ligand, previously thought of as an independent property, appears to depend upon the context of tissue and cellular environment (Fathy et al., 1999; Kenakin, 2002). The relevance of these results to opioid pharmacology in intact animals and human subjects remained unclear.
In morphine-dependent subjects, naloxone and naltrexone potently elicit severe withdrawal, whereas the effects on drug-naive subjects are minimal (Martin and Eades, 1977; Heishman et al., 1990; Baldwin and Koob, 1993; June et al., 1995; Parker and Joshi, 1998). Antagonist effects are not readily accounted for by simply blocking morphine at the MOR (Wang et al., 1994). This suggested the hypothesis that naloxone and naltrexone precipitate withdrawal by suppressing basal MOR signaling in the opioid-dependent state, in addition to blocking agonists at the MOR (Wang et al., 1994). Moreover, it implies that basal MOR signaling plays a role in narcotic dependence. However, known mechanisms of receptor regulation at the protein level commonly occur over relatively short time periods (seconds to hours), whereas opioid dependence and withdrawal extends over several days or even longer (Martin and Eades, 1977; Heishman et al., 1990; Baldwin and Koob, 1993; June et al., 1995; Parker and Joshi, 1998). To account for processes underlying narcotic dependence and withdrawal, we postulated that regulation of basal MOR signaling must parallel the time course of narcotic withdrawal. In this study, we tested the presence of basal MOR activity in mouse brain tissues and its regulation by morphine pretreatment. This was compared with naltrexone-induced withdrawal symptoms over several days after a morphine pretreatment period.
Materials and Methods
Materials. Morphine, naltrexone, naloxone, BNTX, and 6β-naltrexol were obtained from the National Institute on Drug Abuse (NIDA) drug supply program (NIDA, National Institutes of Health, Bethesda, MD). DAMGO, β-CNA, and ICI 174,864 were obtained from Sigma/RBI (Natick, MA). [35S]GTPγS was obtained from PerkinElmer Life and Analytical Sciences (Boston, MA). The [3H]cAMP assay kit was obtrained from Amersham Biosciences Inc. (Piscataway, NJ). Other reagents were obtained from Sigma/RBI.
Animal Treatment. Male ICR mice (20–25 g; Harlan, Indianapolis, IN) were administered either as a single dose of morphine (100 mg/kg s.c.; acute dependence) or as multiple doses for 3 days (every 8 h with increasing doses each day, 30, 60, and 100 mg/kg per injection, respectively; chronic model of dependence). Mice were sacrificed at different times after the last morphine dose. Brains were removed and immediately frozen in liquid nitrogen, then stored at –80°C until use for GTPγS binding assay. Alternatively, different brain regions were dissected and used freshly for the adenylyl cyclase activity assay as described below. MOR knockout (MOR–/–) and control mice (MOR+/+) (Matthes et al., 1996) were sacrificed without any treatment. The brains were removed and immediately frozen in liquid nitrogen. Brains were shipped in dry ice and stored at –80°C until used for the [35S]GTPγS binding assay.
Brain Membrane Preparation. Brains were homogenized in buffer containing 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, 1 mM DTT, 1 mM MgCl2, and 0.25 M sucrose at 4°C and then centrifuged at 27,000g for 10 min (4°C). Pellets were washed twice with homogenizing buffer and stored at –80°C.
[35S]GTPγS Binding. [35S]GTPγS binding assays were carried out as described (Wang et al., 2001b) with minor modifications. To test agonist effects, 10 μg of membrane proteins was incubated with 0.1 nM [35S]GTPγS and different concentrations of agonists (10 nM to 100 μM) in buffer containing 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 10 mM MgCl2, 1 mM DTT, 100 μM GDP, 1 mM EDTA, 0.1% bovine serum albumin, at 30°C for 1 h. To test inverse agonist effects, we used 4 mM MgCl2 and 10 μM GDP in the assay buffer, and reactions were incubated at 30°C for 30 min.
Adenylyl Cyclase Activity. Brain tissues were homogenized in ice-cold buffer (40 ml for 100 mg of tissue) containing 20 mM Tris-HCl, pH 7.4, 1 mM EDTA, 0.5 mM DTT, 1 mM MgCl2, 250 mM sucrose, and 0.5 mM phenylmethylsulfonyl fluoride, then centrifuged at 27,000g for 15 min. The pellets were washed once with the same buffer. Pellets from the second centrifugation were resuspended in the same buffer at a protein concentration of ∼10 mg/ml. Adenylyl cyclase activity was analyzed in assay buffer containing 50 mM Tris-HCl, pH 7.6, 5 mM MgCl2, 250 μM ATP, 5 mM creatine phosphate, 250 μg/ml creatine kinase, 100 μg of membrane protein, and 1 μM test drugs in a total volume of 60 μl. The reaction mixture was incubated at 37°C for 10 min and then stopped by boiling for 2 min. The levels of cAMP in the supernatant were determined by using a [3H]cAMP assay kit (Amersham Biosciences Inc.).
Naltrexone-Induced Withdrawal Jumping. Mice were pretreated with morphine over 3 days as described above (Bilsky et al., 1996; Wang et al., 2001b). Naltrexone (0.1 mg/kg) and 6β-naltrexol (1 mg/kg) [equipotent doses with respect to antagonism of morphine analgesia (Wang et al., 2001b)] were injected i.p. at different times (4, 8, 12, 24, 48, and 72 h) after the last morphine dose. Several experiments were also performed with 1 mg/kg naltrexone and 10 mg/kg 6β-naltrexol at 8 and 48 h after the last morphine dose. Mice were videotaped for 20 min after antagonist injection. A trained observer who was blinded to experimental treatments reviewed the videotapes.
Antagonism of Morphine-Induced Locomotor Activity. Locomotor activity was measured using a Coulbourn Instruments Activity Monitoring System (Coulbourn Instruments, Allentown, PA) and Truscan 99-005-00 software. The chambers are composed of Plexiglas walls with dimensions of 10 inches wide × 10 inches long × 16 inches high and removable base plates. Each chamber contains a sensor ring and 16 infrared beams on each of the four sides, which transect the length and width of the chamber (beam spacing = 0.6 inch, resolution of 32 × 32 squares). Mice were individually placed into the chambers and allowed to habituate for 30 min. After 30 min, morphine was injected at a dose of 30 mg/kg (s.c.), and the mice were immediately placed back into the chambers for an additional 30 min of monitoring. Naltrexone (1.0 mg/kg), 6β-naltrexol (10 mg/kg), or saline injections were then administered i.p., and mice were monitored for an additional 60 min. Measurement of x-y position was taken every 100 ms, and total distance traveled in 1-min intervals (bins) was calculated.
Mass Spectrometric Analysis of Naltrexone and 6β-Naltrexol in Mouse Brain Homogenates. Mice were injected i.p. with either 1.0 mg/kg naltrexone, or 1.0 or 10 mg/kg 6β-naltrexol, and sacrificed after 10 min. Blood samples and brains were collected. Whole brain was excised and homogenized in a 1:3 volume (grams per milliliter) of phosphate-buffered saline solution (pH 7.3), and the homogenate was stored at –70°C. Fifty microliters of mouse brain homogenate were extracted with methyl-t-butyl ether. Each sample was analyzed for naltrexone and 6β-naltrexol using mass spectrometry, with a liquid chromatography/tandem mass spectrometry PE-Sciex API 3000 system equipped with a BDS C18 column [250 × 4.6 mm, 5 μm; mobile phase, acetonitrile/water/acetic acid (65:35:0.05) (v/v) with 5 mM ammonium acetate, with a flow of 1.42 ml/min]. Mass-spectrometric detection employed a heated nebulizer sample-inlet, positive ionization atmospheric pressure chemical ionization, and mass scanning by multiple reaction monitoring. Naltrexone was measured at m/z 342 to 324 and 6β-naltrexol at m/z 344 to 326. The retention time was 3.3 min for naltrexone and 3.0 min for 6β-naltrexol. Nalbuphine (m/z 358–340) served as the internal standard (3.3-min retention time). The method was sensitive to ∼1 ng/g tissue (∼3 nM), with a linear range of 1 to 200 ng/g.
Results
[35S]GTPγS Binding in Brain Homogenates. G protein coupling of MOR was assessed in mouse brain homogenates by measuring the incorporation of [35S]GTPγS into membrane fractions (Wang et al., 2001b), a measure of G protein activation that detects exchange of GDP for GTP as the first step in G protein-coupled receptor signaling. Addition of morphine or the MOR-selective peptide DAMGO to the homogenates accelerated [35S]GTPγS binding by 30 to 40% and 50 to 60%, respectively, of the control (Fig. 1, A and B), with EC50 values of 1.1 ± 0.2 μM and 211 ± 32 nM, respectively. No increase was seen in homogenates from MOR-deficient mice (Fig. 1C), confirming that the increase in [35S]GTPγS binding is due to MOR activation. Upon pretreatment with a single dose of morphine, the dose-response curve (measured 4.5 h after morphine injection) shifted to the right, with EC50 values increasing to 2.9 ± 0.5 μM (morphine) and 400 ± 27 nM (DAMGO), respectively (n = 3, p < 0.05, t test). This finding indicates partial desensitization of MOR to subsequent agonist stimulation. This commonly observed phenomenon is thought to contribute to narcotic tolerance. However, the shift after a single dose of morphine was rather modest and did not further increase substantially with 3-day multiple dosing (Fig. 1, A and B). The EC50 values for morphine and DAMGO in brain homogenates from mice after 3-day morphine pretreatment were 3.2 ± 0.7 μM and 523 ± 77 nM, respectively. Since desensitization of MOR coupling to G proteins was relatively limited, this mechanism may account only for a small portion of the pronounced antinociceptive tolerance to morphine in vivo.
To test for basal MOR activity in brain homogenates, we took advantage of the full inverse agonists β-CNA and BNTX, shown to suppress basal MOR activity in tissue culture (Wang et al., 2000, 2001b). Indeed, both agents significantly suppressed [35S]GTPγS binding in brain homogenates, suggesting the presence of basal MOR activity in vivo (Fig. 2A). Whereas these effects are relatively small compared with background [35S]GTPγS binding, which should reflect a summation of all basal and nonspecific binding, it was nevertheless surprising that basal signaling of a single receptor was significantly detectable above background in brain homogenates. It represented ∼7% versus 35% of maximal [35S]GTPγS binding elicited by morphine. We also performed [35S]GTPγS binding assays in homogenates from brain stem rather than whole brain tissue, but the results were similar, even though the brain stem is expected to express slightly more MOR protein per gram of tissue (data not shown). Dose-response curves of β-CNA and BNTX yielded EC50 values of 80 ± 20 nM and 187 ± 55 nM, respectively (Fig. 2B, β-CNA dose-response curve). However, high concentrations (>10–5 M) of β-CNA nonspecifically decreased [35S]GTPγS binding even in brain homogenates from MOR–/– mice. Therefore, the maximum concentration that we used for β-CNA is 1 μM, which has negligible nonspecific effect (Fig. 3, MOR–/– mice). β-CNA is an irreversible MOR ligand, so that assumptions concerning binding equilibrium and the law of mass action may not be valid. However, over the relatively short time period of incubation, β-CNA may have served mainly as a reversible ligand. Moreover, for the purpose of the present study, the mode of binding to MOR does not substantively affect the interpretation of inverse effects.
Pretreatment of mice with 3-day multiple doses of morphine further increased the inverse agonist effects of β-CNA and BNTX (Fig. 2, p < 0.05 and p < 0.01 comparing saline with 3-day morphine pretreatment), as previously reported in tissue culture (Wang et al., 2000, 2001b). Inverse agonist effects of β-CNA and BNTX in the single-dose morphine pretreatment group did not differ significantly from that in the saline-treated group (p > 0.05) or from that in the 3-day multiple-dose morphine pretreatment group (p > 0.05). Thus, in morphine-pretreated animals, the inverse effect approximated 30 to 40% of the maximal morphine agonist effect. Dose-response curves for β-CNA did not show any shift after either single or 3-day multiple doses of morphine pretreatment (Fig. 2B; see figure legend for EC50 values). In brain homogenates from MOR-deficient mice, β-CNA (1 μM) and BNTX (10 μM) had no significant effects on basal [35S]GTPγS binding (Fig. 3), supporting the hypothesis that the effects of β-CNA and BNTX (which also interact with δ receptors) are largely due to suppression of basal MOR activity. A δ opioid receptor-inverse agonist, ICI 174,864, had no detectable effects on basal [35S]GTPγS binding (Fig. 3) in brain homogenates from both MOR+/+ and MOR–/– mice, further supporting the notion that δ opioid receptors do not play a major role in basal signaling activity of opioid receptors in brain homogenates, or are undetectable under the experimental conditions used. For the sensitive detection of δ opioid receptor basal signaling, different experimental conditions may be needed.
We next tested the effects of naloxone and naltrexone on basal [35S]GTPγS binding (Fig. 4, A and B). Both agents showed a small degree of agonist-like activity in brain homogenates from untreated animals, as observed in tissue culture homogenates (Wang et al., 2001b). In contrast, after morphine pretreatment (3-day multiple doses), these two agents served as inverse agonists, decreasing [35S]GTPγS binding (Fig. 4, A and B). In contrast to the results with naloxone and naltrexone, the neutral antagonist 6β-naltrexol (Wang et al., 2000, 2001b) failed to show any inverse agonist activity in treated and untreated brain homogenates over the entire dose range (Fig. 4C). Therefore, the naloxone-induced decrease in basal [35S]GTPγS binding (naloxone-induced inverse effects) can serve as an indicator of basal signaling activity of MOR in the dependent state, as a result of morphine pretreatment.
To test how long naloxone-sensitive basal MOR signaling lasts after single or 3-day multiple-dose morphine pretreatment, we measured naloxone-induced inverse effects in brain homogenates from mice at different times after the last morphine dose. After a single dose of morphine, naloxone-induced inverse effects returned only gradually toward baseline values over a period of 48 h (Fig. 5). Importantly, after 3-day multiple-dose morphine pretreatment in the chronic model of narcotic dependence, inverse naloxone effects on basal MOR activity lasted for at least 72 h (Fig. 5). Five days after the last morphine dose, naloxone-induced [35S]GTPγS binding leveled out at ∼+1 to 2%, which was no longer significantly different from the control value (+4%). The precision of the assay was insufficient to determine whether chronic morphine pretreatment caused a small but even longer-lasting change in inverse naloxone effects.
Adenylyl Cyclase Activity Assay in Brain Homogenates. G protein coupling is the first step in MOR activation, followed by effector regulation. Specifically, MOR inhibits adenylyl cyclase via inhibitory Gαi/o subunits. Therefore, inverse agonists are expected to increase adenylyl cyclase activity by removing the inhibition generated from basal MOR activity, as shown in cell culture (Wang et al., 2001b). As shown in Fig. 6A, in striatum membrane homogenates from control mice, the inverse agonists BNTX and β-CNA significantly increased adenylyl cyclase activity, whereas naloxone and naltrexone had no effect. After chronic morphine pretreatment (3-day injections), naloxone and naltrexone increased adenylyl cyclase activity, indicating that they now act as inverse agonists. In contrast, the effects of BNTX and β-CNA did not change, indicating that BNTX and β-CNA are full inverse agonists before morphine treatment in this assay. The effects of naloxone and naltrexone in pretreated samples are comparable with that of BNTX and β-CNA. The inverse agonist effects of naloxone and naltrexone were sustained for at least 24 h after the last morphine dose. In contrast to naloxone and naltrexone, the neutral antagonist 6β-naltrexol had no effect on adenylyl cyclase activity regardless of morphine pretreatment. These results are consistent with results from [35S]GTPγS binding assays. In other brain regions, i.e., brain stem, midbrain, cortex, and hippocampus, similar results were obtained for naloxone and 6β-naltrexol (Fig. 6B), although the inverse naloxone effects did not reach statistical significance in some tissues at either 4 or 24 h after the last morphine dose.
Antagonist-Induced Withdrawal Jumping. To test whether suppression of basal MOR activity by naloxone (naloxone-induced inverse effects) correlates with the time course of antagonist-induced withdrawal, we measured withdrawal jumping in morphine-dependent mice (single dose and 3-day morphine pretreatment) using naltrexone and 6β-naltrexol. In mice, vertical jumping is widely considered to be the most sensitive and reliable index of withdrawal intensity and is the most commonly used measure (Way et al., 1969; Saelens et al., 1971; Smits, 1975; Bilsky et al., 1996; Kest et al., 2001). After a single i.p. dose of 100 mg/kg morphine, naltrexone induced withdrawal jumping, which was maximal at 4 h and then ceased between 6 and 10 h (data not shown). However, after 3-day multiple doses of morphine pretreatment, naltrexone (0.1 mg/kg)-induced withdrawal jumping was pronounced at 4 to 8 h, remained detectable at 48 h after the last morphine dose, and subsided at 72 h (Fig. 7A). In contrast, the neutral antagonist 6β-naltrexol, given at equipotent doses (1 mg/kg) for blocking morphine analgesia (Wang et al., 2001b), induced significantly fewer jumps than did naltrexone at all time points tested (Fig. 7A). When naltrexone or 6β-naltrexol was injected at higher doses (1.0 and 10 mg/kg, respectively, at 8 and 48 h after morphine dose), 6β-naltrexol precipitated 25 jumps, which was approximately 38% of the withdrawal jumping elicited by naltrexone (66 jumps) at 8 h (Fig. 7B). However, no significant withdrawal jumping was observed with 6β-naltrexol at 48 h after the last morphine dose (compared to saline injection, p > 0.05), whereas significant jumps were still observed with naltrexone injection (p < 0.05) (Fig. 7B). These results clearly distinguish the pharmacological actions of the inverse agonist naltrexone and the neutral antagonist 6β-naltrexol given at equipotent doses [antagonism of morphine-induced antinociception: ID50 values (95% confidence interval), 0.22 mg/kg (0.11–0.43) for naltrexone and 1.0 mg/kg (0.58–1.7) for 6β-naltrexol (Wang et al., 2001b)]. Moreover, the time course of naltrexone-induced withdrawal jumping was similar to the time course of naloxone-sensitive basal MOR signaling in brain tissue (Fig. 5).
Morphine-Induced Locomotor Activity. To rule out the possibility that the difference between naltrexone and 6β-naltrexol in causing withdrawal jumping has arisen from different rates by which these agents enter the central nervous system, we observed the reversal of morphine-induced locomotor activity by naltrexone (1 mg/kg) and 6β-naltrexol (10 mg/kg), given at equipotent doses, for antagonism of morphine antinociception (Wang et al., 2001b). As shown in Fig. 8, A and B, both antagonists rapidly reversed morphine effects at the same rate. This suggests that the influx rate of these two drugs into the brain is comparable at the doses used for precipitating withdrawal jumping.
Plasma and Brain Levels of Naltrexone and 6β-Naltrexol. Naltrexone has been reported to metabolize to 6β-naltrexol (Chatterjie et al., 1974; Ferrari et al., 1998), a process that could be reversible in vivo. To ascertain the level of administered drugs and metabolites in the blood and brain after administration of naltrexone and 6β-naltrexol in mice, the compounds were extracted from blood and brain tissue and measured by liquid chromatography/tandem mass spectrometry (W. Shi and E. T. Gee, Protocol for UCSF Drug Studies Units, DSU # 94-A, 2000). As shown in Table 1, at 10 min after a dose of 1.0 mg/kg naltrexone and an equipotent dose of 10.0 mg/kg 6β-naltrexol, brain levels were 69 ± 20 ng/g naltrexone and 89 ± 6 ng/g 6β-naltrexol, respectively (S.D., n = 3). Therefore, the concentration of 6β-naltrexol in the brain is comparable with that of naltrexone at 10 min after injection. Neither 6β-naltrexol nor naltrexone was detectable (<1 ng/g; <3 nM) as a metabolite of each other. Thus, both naltrexone and 6β-naltrexol exerted their effect in the brain of mice without substantial metabolic interconversion.
Discussion
In this study, we have shown that basal MOR activity is detectable in mouse brain tissue and that it is up-regulated by morphine treatment. A role for basal MOR signaling in narcotic dependence is buttressed by the finding that changes in basal MOR activity persist for prolonged time periods, consistent with prolonged signs of antagonist-induced withdrawal after exposure to morphine.
The basal signaling activity of MOR was detected with the use of inverse agonists, β-CNA and BNTX (Wang et al., 2001b). Since basal MOR signaling spontaneously increases receptor/G protein coupling, inverse agonists are expected to decreased basal receptor/G protein coupling. We used two assays to test basal MOR activity, [35S]GTPγS binding and adenylyl cyclase activity assays. [35S]GTPγS binding measures direct receptor/G protein coupling, whereas adenylyl cyclase activity is an immediate downstream event, capable of amplifying the signal, and therefore, a potentially more sensitive assay (Wang et al., 2001b). Both assays support the notion that the basal signaling activity of MOR is detectable in brain homogenates and is regulated by morphine pretreatment of mice. After 3-day multiple doses of morphine pretreatment, basal MOR signaling was significantly enhanced, as judged by a small but detectable increase in the inverse effect of β-CNA and BNTX in the GTPγS binding assay (Fig. 2A). A similar finding was previously reported in transfected human embryonic kidney cells (Wang et al., 2001b). Using an adenylyl cyclase activity assay, we did not observe increased effects of β-CNA and BNTX after morphine pretreatment, either because the effect was too small to be detectable or because downstream signaling to adenylyl cyclase had been altered. The results suggest that β-CNA and BNTX behave as full inverse agonists even before morphine pretreatment in this assay. On the other hand, a striking difference after morphine pretreatment are the inverse properties of naloxone and naltrexone, two antagonists capable of precipitating immediate and severe morphine withdrawal in dependent subjects. Whereas naloxone and naltrexone slightly increased [35S]GTPγS binding in brain homogenates from drug-naive mice, morphine pretreatment reversed those effects such that naloxone and naltrexone decreased binding in the dependent state. Conversely, naloxone and naltrexone enhanced adenylyl cyclase activity in brain homogenates only after morphine pretreatment. Recently, Gbahou et al. (2003) have reported that proxyfan, a histamine H3 receptor ligand, can behave as a partial agonist, full agonist, neutral antagonist, partial inverse agonist, or full inverse agonist within a given cell or a given animal species, as well as at a given response within different animal species. This phenomenon is termed protean agonism. Protean agonism indicates that a given drug may produce different physiological responses that do not simply reflect its intrinsic property but depend on the constitutive activity of the system or other factors. Our results suggest that naloxone and naltrexone are protean ligands under certain circumstances. For example, morphine pretreatment of cells or tissues may alter constitutive G protein coupling of MOR, thereby allowing for the expression of the inverse agonist effects of these ligands. Both partial desensitization by receptor phosphorylation and release from calmodulin-induced constraints on basal MOR activity (Wang et al., 1999, 2000, 2001b) could have contributed to the observed effects. Because the inverse effects produced by naloxone and naltrexone in morphine-pretreated brain tissue did not differ statistically from that of BNTX and β-CNA, we assume that naloxone and naltrexone did convert into full inverse agonists with respect to adenylyl cyclase activity.
Our results confirm earlier reports on a change in the antagonist properties of naloxone and naltrexone after morphine pretreatment in cell culture (Wang et al., 1994, 2001b; Liu and Prather, 2001). Therefore, whether a ligand serves as a partial agonist, neutral antagonist, or inverse agonist appears to depend upon the cellular context, the in vitro incubation conditions, and the state of narcotic dependence. We had previously identified different assay conditions for [35S]GTPγS binding for detecting maximal agonist and inverse agonist response (Wang et al., 2001b). Optimal conditions, specifically concentrations of GDP and Mg2+, differed between MOR-transfected GH3 and HEK293 cells (Liu and Prather, 2001; Wang et al., 2001b). When observing the switch of naloxone and naltrexone from partial agonist to inverse agonist conditions in the morphine-dependent state in the present study, we used identical assay condition, e.g., 4 mM Mg2+ and 10 μM GDP, for both experiments. Therefore, the changes in the properties of naloxone and naltrexone are due to increased or altered basal MOR activity, not from different assay conditions. This change could be caused by post-translational modification of MOR, such as phosphorylation and/or changes in receptor aggregate formation in the dependent state. The fact that we observed significant naloxone-induced increases in adenylyl cyclase activity in the dependent state indicates that the inverse naloxone activity detected with [35S]GTPγS binding has potential pharmacological and biological consequences.
A single dose of morphine has a tendency to enhance the inverse effects of BNTX and β-CNA, whereas a 3-day pretreatment period caused only a marginal further increase (Fig. 2B). This is supported by the results obtained with naloxone and naltrexone. A single morphine dose is sufficient to change naloxone and naltrexone into inverse agonists (data not shown). This was unexpected if one assumes that the inverse agonist effects of naloxone and naltrexone are directly related to the severity of antagonist-stimulated withdrawal, which increases with increasing dependence. However, the cAMP system (a key component of dependence) is also markedly up-regulated in the dependent state, such that naloxone's effects are amplified, yielding a larger cAMP overshoot with increasing dependence. Other signaling pathways may also contribute, in the dependent state, to accentuating the response to an inverse agonist. Thus, the inverse effects of naloxone and naltrexone provide a sensitive measure of basal MOR activity in the dependent state. Dose-response curves for naloxone and naltrexone in morphine-dependent brain homogenates show that 10 to 100 nM concentrations are required to detect the inverse effect using [35S]GTPγS binding. Since the binding Kd values for naloxone and naltrexone range from 1 to 10 nM, it is possible that the inverse effects occur in vitro at a receptor conformation with relatively low affinity. This requires further study. Similarly, the EC50 values for morphine and DAMGO are relatively high (above 1 μM and 200 nM, respectively, in untreated homogenates), in agreement with previous reports (Gonzalez-Maeso et al., 1999). Therefore, it is difficult to infer functional states from affinity-potency data in tissue homogenates.
Antagonist-induced withdrawal is also observable after a single dose of morphine, and it subsides only gradually, long after morphine has been eliminated from the body (Martin and Eades, 1977; Heishman et al., 1990; Baldwin and Koob, 1993; June et al., 1995; Parker and Joshi, 1998). Therefore, we tested whether the inverse naloxone effects on basal MOR signaling lasted for similar time periods. Indeed, after a single morphine dose, naloxone-induced inverse effects lasted for at least 24 h (Fig. 5), consistent with prolonged signs of dependence and aversive naloxone effects after a single opiate exposure (lasting 24 h) (Gianutsos et al., 1975; Heishman et al., 1990; Baldwin and Koob, 1993; June et al., 1995; Parker and Joshi, 1998). More strikingly, after 3-day morphine pretreatment, naloxone-induced inverse effects lasted at least 72 h (Fig. 5). This result indicates that whereas a single morphine dose is sufficient to change basal MOR signaling, multiple doses are needed to stabilize this change over a prolonged time period (Fig. 5). This parallels the time course of naltrexone-induced withdrawal jumping, which also lasted considerably longer after multiple doses (Fig. 7A) than a single morphine dose (Gianutsos et al., 1975; Heishman et al., 1990; Baldwin and Koob, 1993; June et al., 1995; Parker and Joshi, 1998). We have used, here, naltrexone rather than naloxone (which yield indistinguishable results in all assays) for better comparison with the chemically similar 6β-naltrexol. Parallel effects of morphine pretreatment on basal MOR activity and withdrawal response support the view that basal signaling is intricately involved in narcotic dependence.
If inverse agonists cause withdrawal jumping by suppressing basal MOR signaling, then we expect antagonists, judged to be neutral in biochemical tests, not to cause withdrawal, or less withdrawal, if given in equipotent doses. We had determined previously that 0.2 mg/kg naltrexone and 1.0 mg/kg 6β-naltrexol are equipotent in antagonizing antinociceptive effect of morphine 10–20 min after the dose (Wang et al., 2001b). In agreement with this hypothesis, equipotent doses of naltrexone (0.1 and 1.0 mg/kg) and the neutral antagonist 6β-naltrexol (1.0 and 10 mg/kg) differed markedly in their ability to elicit withdrawal jumping. At 1.0 mg/kg, 6β-naltrexol failed to cause withdrawal jumping over the entire observation period (Fig. 7A), while 10 mg/kg caused an attenuated withdrawal effect (40% of that elicited by 1.0 mg/kg naltrexone) only at 8 h after morphine pretreatment. We had observed similar results with 6β-naltrexol previously at a single time point (4 h) after the last morphine dose (Wang et al., 2001b). This result argues against the possibility that the naltrexone effects could have been caused to a large extent by displacement of residual morphine or endogenous opioids, which would have been displaced by 6β-naltrexol as well. Therefore, naltrexone's effect appears to be at least in part caused by suppression of basal MOR activity in vivo.
Long-lasting changes in neuronal activity have been associated with altered gene expression during narcotic addiction (Nestler and Aghajanian, 1997), whereas direct regulation of protein activity usually occurs over short time periods. However, the changes of MOR signaling reported here occur over an unprecedented time scale for biochemical regulation at the level of receptor protein. A molecular basis for the lasting regulation of basal MOR activity was suggested by our observation that calmodulin binds to the intracellular i3 loop of MOR in competition with G proteins, thereby suppressing inherent basal G protein coupling by MOR in untreated tissue (Wang et al., 1999, 2000, 2001b). We propose that basal MOR activity is constrained in drug-naive tissue by the action of calmodulin (Wang et al., 1999, 2001b) or other cellular factors. Upon calmodulin release by receptor stimulation, basal G protein coupling can occur at an accelerated pace. A single dose of morphine is sufficient to elevate basal MOR activity; however, multiple doses appear to induce a vicious circle, probably involving gene expression, which maintains unconstrained basal signaling over prolonged time periods. This provides a potential link between biochemical changes at the receptor, changes in gene expression, and the long time scales involved in narcotic addiction.
In conclusion, we have demonstrated basal MOR signaling in mouse brain tissues and provided evidence in support of the hypothesis that basal MOR activity is regulated in the morphine-dependent state. This could play a role in narcotic dependence. The mechanisms underlying these observations require further study. Ability of neutral antagonists to block MOR activation by narcotic agonists while causing less withdrawal promises new avenues in the treatment of drug addiction, overdose, and pain management.
Acknowledgments
Morphine, naltrexone, and 6β-naltrexol were obtained from the NIDA drug supply program. We thank Jennifer Blair for help with the data collection.
Footnotes
-
This study was supported by National Institute on Drug Abuse Grants DA 04166 and DA 06284, a grant from the Wheeler Center for Neurobiology of Addiction, University of California San Francisco, and funds from the College of Medicine and Public Health, Ohio State University. J.J.L. received support from a Summer Undergraduate Research Fellowship from the American Society for Pharmacology and Experimental Therapeutics and a Maine Space Grant Consortium Fellowship.
-
DOI: 10.1124/jpet.103.054049.
-
ABBREVIATIONS: MOR, μ opioid receptor; BNTX, 7-benzylidenenaltrexone; NIDA, National Institute on Drug Abuse; DAMGO, [d-Ala2,N-Me-Phe4,Gly-ol5]-enkephalin; β-CNA, β-chlornaltrexamine; ICI 174,864, N,N-diallyl-Tyr-Aib-Aib-Phe-Leu-OH; [35S]GTPγS, guanosine 5′-O-(3-[35S]thio)triphosphate; DTT, dithiothreitol.
- Received May 6, 2003.
- Accepted October 31, 2003.
- The American Society for Pharmacology and Experimental Therapeutics