Abstract
Agonist efficacy may influence the magnitude of neuroadaptation in response to chronic drug exposure. Chronic administration of either Δ9-tetrahydrocannabinol (THC), a partial agonist, orR-(+)-[2,3-dihydro-5-methyl-3-[(morpholinyl)methyl]pyrrolo-[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanone mesylate (WIN55,212-2), a full agonist, for G protein activation produces tolerance to cannabinoid-mediated behaviors. The present study examined whether chronic administration of maximally tolerated doses of Δ9-THC and WIN55,212-2 produces similar cannabinoid receptor desensitization and down-regulation. Mice were treated with escalating doses of agonist for 15 days, with final doses of 160 mg/kg Δ9-THC and 48 mg/kg WIN55,212-2. Tolerance to cannabinoid-mediated hypoactivity, hypothermia, and antinociception was found after treatment with Δ9-THC or WIN55,212-2. In autoradiographic studies, cannabinoid-stimulated guanosine 5′-O-(3-[35S]thio)triphosphate ([35S]GTPγS) binding was significantly decreased in all regions of Δ9-THC- and WIN55,212-2-treated brains. In addition, Δ9-THC-treated brains showed greater desensitization in some regions than WIN55,212-2-treated brains. Concentration-effect curves for cannabinoid-stimulated [35S]GTPγS binding confirmed that decreases in the hippocampus resulted from loss of maximal effect in both WIN55,212-2- and Δ9-THC-treated mice. In the substantia nigra, theEmax decreased and the EC50value increased for agonist stimulation of [35S]GTPγS binding in Δ9-THC-treated mice. [3H]N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide (SR141716A) binding was decreased in all brain regions in Δ9-THC- and WIN55,212-2-treated mice, with no difference between treatment groups. These results demonstrate that chronic treatment with either the partial agonist Δ9-THC or the full agonist WIN55,212-2 produces tolerance to cannabinoid-mediated behaviors, as well as cannabinoid receptor desensitization and down-regulation. Furthermore, Δ9-THC produced greater desensitization than WIN55,212-2 in some regions, indicating that agonist efficacy is one determinant of cannabinoid receptor desensitization in brain.
Chronic Δ9-THC administration produces tolerance (Carlini, 1968; McMillan et al., 1971), but the cellular mechanisms underlying cannabinoid tolerance have only recently been investigated, and the relationship between in vivo and in vitro adaptation has not been elucidated. CB1 receptors activate G proteins of the Gi/Go subfamily (Howlett et al., 1986; Glass and Northup, 1999) and modulate adenylyl cyclase, potassium, and calcium channel conductance (Howlett et al., 1988;Mackie et al., 1995). Chronic cannabinoid administration alters signaling at each level of the CB1 signal transduction cascade. Down-regulation (loss of binding sites) has been reported after chronic administration of Δ9-THC, as well as the synthetic cannabinoid agonist CP55,940 (Oviedo et al., 1993; Fan et al., 1996; Romero et al., 1998; Breivogel et al., 1999). Chronic Δ9-THC administration also produces desensitization (defined herein as a decrease in G protein activation) throughout the brain (Sim et al., 1996; Breivogel et al., 1999). The effect of chronic cannabinoid treatment on effectors is not well understood. Studies have reported increased basal levels of cAMP and cAMP-dependent protein kinase after Δ9-THC administration (Rubino et al., 2000b), but no alteration in basal or cannabinoid-mediated inhibition of adenylyl cyclase after chronic CP55,940 treatment (Fan et al., 1996; Rubino et al., 2000a).
Many studies have examined the effects of chronic administration of Δ9-THC, the primary psychoactive compound in marijuana. However, the development of synthetic cannabinoid ligands with increased potency and/or efficacy allows more extensive investigation of CB1 receptor adaptation. For CB1 receptors, WIN55,212-2 is a full agonist for activation of G proteins, as measured using agonist-stimulated [35S]GTPγS binding (Sim et al., 1996; Breivogel et al., 1998). Methanandamide and CP55,940 are moderate- and high-efficacy partial agonists, respectively, whereas Δ9-THC is a low-efficacy partial agonist for G protein activation. The effect of treatment with cannabinoid agonists of different efficacies on in vivo and in vitro measures of cannabinoid tolerance has not been directly examined in animals. However, studies have shown that low- (Δ9-THC), moderate- (methanandamide), and high (CP55,940)-efficacy agonists produce CB1 receptor desensitization and down-regulation, as well as tolerance to cannabinoid-mediated behaviors (Oviedo et al., 1993; Fan et al., 1996;Sim et al., 1996; Breivogel et al., 1999; Romero et al., 1999; Rubino et al., 2000a).
Treatment with cannabinoid agonists of different efficacies has been directly compared in a cell culture model that heterologously expresses CB1 receptors (Hsieh et al., 1999). Brief exposure to high-efficacy agonists (WIN55,212-2, CP55,940, and HU210) promoted rapid internalization of CB1 receptors, whereas low- and moderate-efficacy agonists (Δ9-THC and methanandamide, respectively) produced less internalization. Similarly, studies in N18TG2 cells showed that treatment with desacetyllevonantradol, a high-efficacy agonist, produced greater desensitization of cannabinoid-inhibited cAMP accumulation than treatment with Δ9-THC (Dill and Howlett, 1988). However, it is not clear whether the same results will be obtained in brain because the stoichiometric relationship between CB1 receptors and G proteins, as well as the cellular complement of receptors and signaling proteins, differs between brain and cell models. Furthermore, different effects may be seen in different brain regions based on these factors.
Previous studies have shown that alterations in CB1 receptor signaling exhibit regional differences in parameters such as the rate and magnitude of development of desensitization and down-regulation (Breivogel et al., 1999). Differences in the rate of recovery from tolerance to cannabinoid-mediated antinociception and hypomotility have also been reported (Bass and Martin, 2000), suggesting that regional differences in CB1 signaling alterations may correlate with differences in behavioral measures of cannabinoid tolerance. CB1 receptor levels vary among brain regions (Herkenham et al., 1991), providing one possible explanation for these findings. Regions of high CB1 receptor density exhibit a small degree of receptor reserve for G protein activation, whereas low-density regions show no receptor reserve (Breivogel and Childers, 2000; Selley et al., 2001). CB1 receptor reserve may explain why Δ9-THC produces comparable behavioral effects to higher efficacy agonists (Compton et al., 1992a,b) even though it is a partial agonist for G protein activation. However, Δ9-THC may require greater cannabinoid receptor occupancy to produce in vivo effects. Thus, chronic administration of equally effective behavioral doses of Δ9-THC and WIN55,212-2 might result in greater desensitization after Δ9-THC treatment because it must occupy more receptors to produce a comparable response. Therefore, these studies were conducted to measure CB1 receptor desensitization and down-regulation, as well as tolerance to cannabinoid-mediated behaviors, after chronic treatment with the full agonist WIN55,212-2 or the partial agonist Δ9-THC.
Materials and Methods
Materials.
Mice (ICR male, 24–30 g) were obtained from Harlan (Indianapolis, IN) and maintained on 14:10 light/dark cycle with ad libitum food and water. [35S]GTPγS (1250 Ci/mmol) was purchased from PerkinElmer Life Sciences (Boston, MA). [3H]SR141716A (49 Ci/mmol) was purchased from Amersham Biosciences (Piscataway, NJ) Δ9-THC was provided by the Drug Supply Program of the National Institute on Drug Abuse. Methanandamide, WIN55,212-2, and GDP were purchased from Sigma/RBI (Natick, MA). Kodak X-O-mat and Biomax MS film were purchased from PerkinElmer Life Sciences. All other reagent grade chemicals were obtained from Sigma-Aldrich (St. Louis, MO) or Fisher Scientific (Pittsburgh, PA).
Chronic Drug Administration.
Δ9-THC and WIN55,212-2 were dissolved in a 1:1:18 solution of ethanol/Emulphor/saline. Mice were injected s.c. twice daily (7:00 AM and 3:00 PM) with Δ9-THC, WIN55,212-2, or vehicle for 15 days. Initial doses of drugs were 10 mg/kg Δ9-THC and 3 mg/kg WIN55,212-2, which were determined to be equally active in behavioral tests after acute administration. The doses of drugs were doubled every 3 days to final doses of 160 mg/kg Δ9-THC and 48 mg/kg WIN55,212-2. Twenty-four hours after the final injection, mice were either sacrificed for autoradiographic assays or tested in behavioral assays to assess tolerance.
Behavioral Evaluation.
Mice were tested using paradigms previously established in the laboratory to assess tolerance (Fan et al., 1994). Mice were acclimated to the observation room overnight before evaluation for hypomotility, rectal temperature, and antinociception. Baseline measures of temperature and antinociception were obtained before challenge. Rectal temperature was measured using a telethermometer (Yellow Springs Instrument Co., Yellow Springs, OH) and a thermistor probe. Antinociception was assessed using tail-flick reaction time to a heat stimulus. Mice from Δ9-THC-, WIN55,212-2-, and vehicle-treated groups were then administered a single intravenous injection of appropriate drug or vehicle in the tail vein. Spontaneous activity was assessed by placing mice in individual photocell chambers (11 inches × 6.5 inches) 5 min after injection. Activity was measured for 10 min in a Digiscan animal activity monitor (Omnitech Electronics Inc., Columbus, OH) as the number of interruptions of 16 photocell beams per chamber. Mice were tested in the tail-flick assay at 20 min postinjection. A 10-s maximum latency was used to avoid tail injury. Rectal temperature was measured as described above at 60 min after injection.
Agonist-Stimulated [35S]GTPγS Binding.
Mice were sacrificed by rapid decapitation 24 h after the final injection. Brains were removed and immediately frozen in isopentane at −30°C. Twenty-micrometer coronal sections were cut on a cryostat maintained at −20°C and thaw-mounted onto gelatin-coated slides. Slides were collected in a humidified chamber and stored desiccated at 4°C overnight. Slides were then stored desiccated at −80°C until the day of the assay. Agonist-stimulated [35S]GTPγS binding was assessed as described previously (Sim et al., 1996), with minor modification. Slides were brought to room temperature under cool air then equilibrated in 50 mM Tris-HCl, pH 7.4, with 3 mM MgCl2, 0.2 mM EGTA, 100 mM NaCl, and 0.5% bovine serum albumin (assay buffer) for 10 min at 25°C. Slides were transferred to assay buffer containing 2 mM GDP and 9.5 mU/ml adenosine deaminase and incubated for 15 min at 25°C. The addition of adenosine deaminase to the assay has been shown to decrease basal [35S]GTPγS binding by degrading endogenous adenosine that is tightly bound to its receptors (Moore et al., 2000). Slides were then incubated in the presence or absence (basal) of 10 μM WIN55,212-2 or 10 μM methanandamide in assay buffer containing 2 mM GDP, 9.5 mU/ml adenosine deaminase, and 0.04 nM [35S]GTPγS for 2 h at 25°C. Concentration-effect curves were generated using 0.03 to 10 μM WIN55,212-2. Slides were rinsed twice for 2 min each in 50 mM Tris buffer, pH 7.4, at 4°C and then rinsed for 30 s in deionized water. Slides were dried thoroughly and placed in cassettes with14C microscales and Kodak X-O-mat film. Films were developed after exposures of 4 to 7 days.
[3H]SR141716A Binding.
Slides were brought to room temperature then equilibrated in assay buffer for 20 min at 30°C. Slides were then incubated in assay buffer containing 0.8 nM [3H]SR141716A at 30°C for 90 min. Nonspecific binding was assessed in the presence of 5 μM unlabeled SR141716A. Slides were rinsed four times for 10 min each in assay buffer at 25°C then for 30 s in deionized water on ice. Slides were dried thoroughly, transferred to cassettes containing3H microscales and exposed to Kodak Biomax MS film for 12 weeks.
Data Analysis: Behavioral Evaluation.
Antinociception was calculated as percentage of maximum possible effect ([(test latency − control latency)/(10 s − control latency)] × 100). Spontaneous activity was expressed as the percentage of the activity of vehicle-treated mice challenged with vehicle. Change in rectal temperature was calculated by (control temperature − test temperature). Dose-response curves were generated by administering increasing doses of Δ9-THC and WIN55,212-2 to groups of six to 12 mice. ED50 values were calculated based upon least-squares linear regression followed by calculation of 95% confidence limits (Bliss, 1967). Significance was determined by calculating the potency ratio between the groups (Colquhoun, 1971) and considered significant when the lower 95% confidence limit was >1.
Data Analysis: Autoradiography.
Films were digitized with an XC-77 videocamera (Sony, Tokyo, Japan) and analyzed using the NIH Image program for Macintosh computers. Data are reported as mean values ± S.E.M. of triplicate sections of brains from eight mice per group. Net [35S]GTPγS binding is defined as (agonist-stimulated [35S]GTPγS binding − basal [35S]GTPγS binding). Specific [3H]SR141716A binding is calculated as (total [3H]SR141716A binding − nonspecific binding). Nonlinear iterative regression analyses of agonist concentration-effect curves were performed with JMP, version 2.0.5 (SAS Institute, Inc., Cary, NC). Binding values were converted to femtomoles or picomoles per milligram of protein based on specific activity of the isotopes and the ratio of milligrams of protein per milligram of tissue. 14C values were corrected for35S based upon incorporation of35S into sections of frozen brain paste. Statistical comparison of brains from vehicle-, Δ9-THC-, and WIN55,212-2-treated mice was conducted by analysis of variance followed by post hoc analysis with the Tukey-Kramer test in each region using JMP. Statistical significance of differences in each specific brain region between Δ9-THC- and WIN55,212-2-treated mice was determined by the Tukey-Kramer test.
Results
Behavioral Evaluation.
Preliminary studies were first conducted in naı̈ve mice to determine the doses of Δ9-THC and WIN55,212-2 that produced approximately the same magnitude of effects. It was found that an acute injection of 10 mg/kg Δ9-THC and 3 mg/kg WIN55,212-2 produced a decrease of 4.6 and 4.3°C in rectal temperature, respectively, and both drugs at these doses produced the same degree of inhibition of spontaneous activity and tail-flick response (data not shown). After 15 days of treatment (s.c.) with escalating doses of either Δ9-THC or WIN55,212-2, mice exhibited tolerance to the effects of Δ9-THC and WIN55,212-2 produced by acute i.v. challenge in all paradigms tested (Fig.1; Table1). A high degree of tolerance is evident by the dramatic shifts to the right of all dose-response curves. Additionally, challenge with a very high dose of Δ9-THC (60 mg/kg) in the chronic Δ9-THC-treated mice produced less than 50% effect in the tail-flick assay and in lowering rectal temperature. Chronic treatment with Δ9-THC or WIN55,212-2 produced 50- to 100-fold shifts in the ED50values for antinociception and hypothermia, indicating that similar levels of tolerance were produced by treatment with both drugs. The degree of tolerance for Δ9-THC in the hypoactivity measure was only 6-fold, which was considerably less than the corresponding finding for WIN55,212-2 (72-fold).
Agonist-stimulated [35S]GTPγS Binding.
[35S]GTPγS binding was assessed using 10 μM WIN-55,212-2, which has previously been shown to be maximally effective in this assay (Sim et al., 1996; Breivogel et al., 1998). Brain sections from mice chronically treated with WIN-55,2212-2 or Δ9-THC had visibly reduced levels of WIN55,212-2-stimulated [35S]GTPγS binding in all regions compared with the vehicle control group (Fig.2). Basal binding did not seem to differ between the three groups of mice, and this observation was confirmed by densitometric analysis (data not shown). Statistical analysis (two-way ANOVA) of net WIN55,212-2-stimulated [35S]GTPγS binding showed a significant interaction between drug and region (F15,284 = 112, p < 0.0001), and post hoc analysis confirmed that net-stimulated binding was significantly reduced in virtually all regions of brains from Δ9-THC- and WIN55,212-2-treated mice compared with vehicle-treated controls (Table 2). Statistical analysis (two-way ANOVA) of net agonist-stimulated [35S]GTPγS binding in brains from Δ9-THC- compared with WIN55,212-2-treated mice also revealed a significant interaction between drug and region (F14,187 = 178, p < 0.0001), indicating that regional differences exist in desensitization produced by treatment with Δ9-THC versus WIN55,212-2. Because significant effects of drug and region were found, net-stimulated [35S]GTPγS binding in Δ9-THC- and WIN55,212-2-treated brains was compared by post hoc analysis in each region for effect of drug to determine whether treatment with the two agonists produced differences in cannabinoid-mediated G protein activation. This analysis revealed significant differences in net-stimulated [35S]GTPγS binding in prefrontal cortex, nucleus accumbens, cingulate cortex, caudate-putamen, hippocampus, amygdala, entopeduncular nucleus, and cerebellum (p < 0.05) between WIN55,212-2- and Δ9-THC-treated mice. No significant differences in net WIN55,212-2-stimulated [35S]GTPγS binding between Δ9-THC- and WIN55,212-2-treated mice were found in the remaining regions: globus pallidus, hypothalamus, thalamus, entorhinal cortex, substantia nigra, and periaqueductal gray (PAG) (p > 0.05). Data from cannabinoid-treated mice were also expressed as a percentage of net WIN55,212-2-stimulated [35S]GTPγS binding in vehicle-treated mice to reveal the relative levels of desensitization produced by the two drugs in different brain regions (Fig. 3). These data illustrate region-specific differences in cannabinoid-mediated desensitization produced by treatment with WIN55,212-2 versus Δ9-THC.
Brain sections were also processed for agonist-stimulated [35S]GTPγS binding using a maximally effective concentration (10 μM) of methanandamide, a stable analog of the endocannabinoid anandamide and a partial agonist for cannabinoid receptor activation of G proteins. The level of overall stimulation produced by methanandamide in vehicle-treated brains was approximately 30 to 40% of that produced by WIN55,212-2, reflecting the partial agonist activity of methanandamide for G protein activation. Statistical analysis (two-way ANOVA) confirmed an interaction of drug and region (F8,146 = 57,p < 0.0001) on desensitization and post hoc analysis showed significant decreases in [35S]GTPγS binding in all regions of brains from Δ9-THC- and WIN55,212-2-treated mice. However, in contrast to data obtained using WIN55,212-2 as the in vitro agonist, there did not seem to be as much difference in the degree of desensitization between regions. There was also no significant difference in the level of net-stimulated [35S]GTPγS binding in brains from Δ9-THC- versus WIN55,212-2-treated mice except in cerebellum (p > 0.05).
The preceding experiments were conducted using maximally effective concentrations of agonists to stimulate [35S]GTPγS binding. However, it cannot be determined from these data whether decreased stimulation of [35S]GTPγS binding resulted from a change in the maximal effect or potency of the agonist for G protein activation. To answer this question, concentration-effect curves of net WIN55,212-2-stimulated [35S]GTPγS binding were generated in brain sections adjacent to those shown in Fig. 2C that included the substantia nigra and hippocampus (Fig.4; Table3). Statistical analysis (two-way ANOVA of Emax values) showed an interaction between drug and region (F3,34 = 112,p < 0.0001). TheEmax values calculated in the hippocampus from Δ9-THC- and WIN55,212-2-treated mice were significantly lower than those obtained in vehicle-control mice (p < 0.01). In addition, theEmax values calculated in Δ9-THC- and WIN55–212-2-treated hippocampi were significantly different from each other (p < 0.01). In the substantia nigra, Δ9-THC-treated brains had significantly decreasedEmax values compared with vehicle-treated mice (p < 0.05), but there was no difference between WIN55,212-2-treated mice and mice treated with vehicle or Δ9-THC. There was a significant interaction between drug and EC50 values only in the substantia nigra (F2,16 = 6,p < 0.01), where the EC50 value for WIN55,212-2-stimulated [35S]GTPγS binding in Δ9-THC-treated mice was significantly different from vehicle control. These data show that in the hippocampus desensitization resulted from a decrease in maximal effect, whereas in the substantia nigra, both the Emaxand EC50 values for WIN55,212-2-stimulated [35S]GTPγS binding were significantly different between Δ9-THC- and vehicle-treated mice.
[3H]SR141716A Binding.
Decreased [3H]SR141716A binding was also seen in the brains of WIN55,212-2- and Δ9-THC-treated mice and confirmed by computer-assisted densitometry (Table4). Statistical analysis (two-way ANOVA) revealed an interaction between drug and region (F15,308 = 17, p < 0.0001) on [3H]SR141716A binding. Post hoc analysis showed that [3H]SR141716A binding was reduced in virtually all regions of brains from Δ9-THC- and WIN55,212-2-treated mice. The exception to this result was the substantia nigra, in which a ≤5% decrease was seen in Δ9-THC-treated mice compared with vehicle control. Data from drug-treated mice were expressed as percentage of [3H]SR141716A binding in vehicle control brains in each region (Fig. 5). As seen with desensitization, the loss in [3H]SR141716A binding varied by brain region. In contrast to desensitization, there was very little difference in the magnitude of down-regulation in brains from Δ9-THC- compared with WIN55,212-2-treated mice. The results of receptor binding studies showed a similar regional distribution for down-regulation as seen for desensitization, and a correlation was found between the levels of down-regulation and desensitization in both Δ9-THC- and WIN55,212-2-treated brains (r = 0.800, p = 0.0006 for Δ9-THC; r = 0.805,p = 0.005 for WIN55,212-2).
Discussion
This study investigated the effect of agonist efficacy on cannabinoid receptor adaptation by treating mice with Δ9-THC or WIN55,212-2, low- and high-efficacy agonists for G protein activation, respectively. Both WIN55,212-2 and Δ9-THC treatment produced desensitization and down-regulation throughout the brain, but the magnitude of effect varied among regions. In addition, regional differences were found in the magnitude of desensitization in brains from Δ9-THC compared with WIN55,212-2-treated mice. In all cases, Δ9-THC treatment produced greater desensitization than treatment with WIN55,212-2. In contrast, no difference in the magnitude of down-regulation was found in brains from Δ9-THC- versus WIN55,212-2-treated mice. These results indicate that region-specific properties of cannabinoid receptors and/or signaling proteins affect the magnitude of desensitization and down-regulation.
An important consideration is the dose of each drug used in the chronic administration paradigm. Our goal was to select dosing regimens that would be approximately equivalent pharmacologically. Obviously, development of chronic treatment regimens that are pharmacologically equivalent is problematic because of the likely pharmacodynamic and pharmacokinetic differences between Δ9-THC and WIN55,212-2. Therefore, we initiated dosing with quantities of Δ9-THC and WIN55,212-2 that were equally effective acutely and doubled the doses every 3 days during treatment. Although there is no certainty that longer duration or higher dose regimens would have produced greater tolerance, the degree of tolerance produced in these regimens is far greater than that reported previously. It is unlikely that greater tolerance can be achieved without the danger of toxic side effects. This treatment paradigm produced a high degree of tolerance to cannabinoid-mediated hypothermia and antinociception. The lower amount of tolerance that developed to the effect of Δ9-THC on spontaneous activity should be interpreted cautiously because the high doses of Δ9-THC required to develop a dose-response relationship in the Δ9-THC-tolerant mice may have produced some nonreceptor-mediated motor effects.
An experimental concern in this study was the possibility that residual drug would remain in the brain. Previous studies have demonstrated that a single injection of Δ9-THC does not alter WIN55,212-2-stimulated [35S]GTPγS binding or [3H]WIN55,212-2 binding (Sim et al., 1996;Romero et al., 1998). In the present study, no difference in basal [35S]GTPγS binding was found in any region between vehicle and treated brains. In addition, concentration-effect curves showed no change in the EC50 value of WIN55,212-2 in stimulating [35S]GTPγS binding, except in the substantia nigra of Δ9-THC-treated mice. Finally, no difference in [3H]SR141716A binding was found between Δ9-THC- and WIN55,212-2-treated brains. If the differences in desensitization between Δ9-THC- and WIN55,212-2-treated brains were due to a higher level of residual drug in the Δ9-THC-treated brains, residual Δ9-THC should inhibit [3H]SR141716A binding. Moreover, previous studies showed no change in the KD of SR141716A in brains from animals treated for up to 21 days with Δ9-THC (Breivogel et al., 1999). Taken together, these data indicate that changes in [35S]GTPγS and [3H]SR141716A binding in Δ9-THC- and WIN55,212-2-treated brains do not result from residual drug.
Chronic treatment with Δ9-THC produced greater desensitization of cannabinoid-mediated G protein activity than treatment with WIN55,212-2 in most regions. However, significant differences were confined mainly to regions of moderate (2–6 fmol/mg) levels of cannabinoid-stimulated [35S]GTPγS binding, including prefrontal cortex, nucleus accumbens, cingulate cortex, caudate-putamen, hippocampus, amygdala, and cerebellum. Regions with very high levels of cannabinoid-stimulated [35S]GTPγS binding (>6 fmol/mg), such as globus pallidus and substantia nigra, showed no difference between the two treatment groups. Similarly, no difference was found between Δ9-THC- and WIN55,212-2-treated mice in regions expressing low (<2 fmol/mg) levels of cannabinoid-stimulated [35S]GTPγS binding, including thalamus, hypothalamus, and PAG. The levels of CB1 receptors in these regions (Table 4; Herkenham et al., 1991) correspond to the level of cannabinoid-stimulated [35S]GTPγS binding (Table 2), indicating that receptor density may affect desensitization. Interestingly, globus pallidus and substantia nigra, regions containing very high levels of CB1 receptors, also showed a smaller degree of down-regulation. These results suggest that receptor recycling may also vary by region due in part to receptor density.
Concentration-effect curves of WIN55,212-2-stimulated [35S]GTPγS binding were generated to determine whether loss of activity resulted from changes inEmax and/or EC50values of WIN55,212-2. In the hippocampus, both WIN55,212-2- and Δ9-THC-treated mice exhibited decreased maximal effect of WIN55,212-2 for G protein activation, with no difference in EC50 value. Moreover, the decrease in maximal effect was nearly twice as great for Δ9-THC-treated than WIN55,212-2-treated mice. In the substantia nigra, only Δ9-THC-treated brains showed a loss in maximal effect, which was accompanied by a nearly 2-fold increase in the EC50 value of WIN55,212-2 for G protein activation. These findings indicate that there is receptor reserve for cannabinoid-mediated G protein activation in the substantia nigra that is lost after treatment with Δ9-THC. Furthermore, these results support the hypothesis that receptor density may be one factor that influences the magnitude of desensitization.
Differences in the chemical structure of WIN55,212-2 and Δ9-THC could also contribute to differences in desensitization due to differential molecular interactions with the CB1 receptor. Mutagenesis studies have shown that specific residues in transmembrane 3 of the CB1 receptor distinguish binding of WIN55,212-2 from other cannabinoid ligands, including Δ9-THC (Song and Bonner, 1996; Chin et al., 1999). However, it is not clear whether differences in ligand binding at this site affect CB1 receptor-G protein coupling. Moreover, although differences in desensitization were produced in some regions by treatment with WIN55,212-2 versus Δ9-THC, this finding was not true in all brain regions, and no differences were found in receptor down-regulation. These results suggest that subtle differences in the molecular interaction between WIN55,212-2 and Δ9-THC and the CB1 receptor do not affect adaptation to chronic drug treatment. This finding is in agreement with previous studies showing that WIN55,212-2 and Δ9-THC produce similar acute behavioral effects (Compton et al., 1993) and that cross-tolerance can be demonstrated between these two cannabinoid agonists (Pertwee et al., 1993; Fan et al., 1994). It is also possible that binding of WIN55,212-2 to a non-CB1 G protein-coupled receptor (Breivogel et al., 2001) produces differences between WIN55,212-2- and Δ9-THC-treated brains. However, the anatomical distribution of this novel receptor (high in cortex, midbrain, and hippocampus; absent in basal ganglia and cerebellum) does not correlate with the regional profile of areas with different levels of desensitization in WIN55,212-2- and Δ9-THC-treated brains (Table 2). Moreover, no difference in down-regulation was found after treatment with WIN55,212-2 compared with Δ9-THC.
Another factor that could contribute to differences in desensitization between Δ9-THC- and WIN55,212-2-treated mice is the regional distribution of signaling proteins, particularly those involved in agonist-mediated desensitization. CB1 receptors primarily activate Gi/Go proteins (Glass and Northup, 1999; Prather et al., 2000). Studies in purified reconstituted cell models have shown that Δ9-THC is partial agonist for activation of both Gαi and Gαo, whereas WIN55,212-2 is a full agonist for activation of Gαi and a partial agonist for Gαo (Glass and Northup, 1999). However, studies in brain have shown that WIN55,212-2 is a full agonist in all regions (Breivogel and Childers, 2000) despite the fact that Gαo is the predominant Gα subtype in brain. Perhaps the localization of specific isoforms of G protein-coupled receptor kinase or arrestin, which mediate CB1 receptor desensitization (Jin et al., 1999), may be a relevant factor. Another possibility is that Gβγ subtype localization affects desensitization because 1) each exhibits a unique anatomical distribution (Betty et al., 1998) and 2) Gβγ recruits G protein-coupled receptor kinase to the receptor (Pitcher et al., 1998).
The relationship between agonist efficacy and tolerance has perhaps been most extensively investigated for the μ-opioid receptor, which is also coupled to Gαi/Gαo. A number of studies have shown that the magnitude of tolerance to opioid-mediated analgesia is inversely correlated with the efficacy of the treatment drug (Stevens and Yaksh, 1989; Duttaroy and Yoburn, 1995; Walker and Young, 2001). Although the present study did not demonstrate obvious differences in tolerance after treatment with Δ9-THC compared with WIN55,212-2, it is likely that the very high doses administered produced maximal tolerance as assessed in this paradigm. The parameter that did differ between Δ9-THC- and WIN55,212-2-treated brains was desensitization, which showed region-specific differences in the magnitude of desensitization produced by the two drugs. This suggests that differential receptor-G protein desensitization produced by treatment with agonists of different efficacies could produce differences in tolerance at the in vivo level. Recent studies have shown that mice lacking β-arrestin do not exhibit tolerance to analgesia after chronic opioid treatment (Bohn et al., 2000), providing evidence that desensitization is an underlying cellular mechanism of tolerance for G protein-coupled receptors. In summary, these results show that treatment with either Δ9-THC or WIN55,212-2 produces cannabinoid receptor desensitization and down-regulation, as well as tolerance to cannabinoid-mediated hypoactivity, antinociception, and hypothermia. The regional profile of desensitization and down-regulation, as well as the results of concentration-effect studies, indicates that the level of cannabinoid receptors and receptor reserve for cannabinoid-mediated G protein activation are factors that influence the development of desensitization and down-regulation. Moreover, the specific regional distributions of cannabinoid receptors and signaling proteins may affect the magnitude of these adaptations. Finally, region-dependent differences in desensitization were detected after chronic treatment with Δ9-THC versus WIN55,212-2, indicating that agonist efficacy influences desensitization of cannabinoid receptors in brain.
Acknowledgments
Leah Brunk and Renee Jefferson provided excellent technical assistance in these studies. We thank Drs. Aron Lichtman and Dana E. Selley for helpful discussions.
Footnotes
-
These studies were supported by U.S. Public Health Service Grants DA-00287 (to L.J.S.) and DA-03672 (to B.R.M.) from the National Institute on Drug Abuse.
-
DOI: 10.1124/jpet.102.035618
- Abbreviations:
- Δ9-THC
- Δ9-tetrahydrocannabinol
- CB
- cannabinoid
- CP55,940
- (1α,2β)-R-5-(1,1-dimethylheptyl)-2-[5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]phenyl
- WIN55,212-2
- R-(+)-[2,3-dihydro-5-methyl-3-[(morpholinyl)methyl]pyrrolo[1,2,3-de]-1,4-benzoxazinyl]-(1-naphthalenyl)methanone mesylate
- [35S]GTPγS
- guanosine 5′-O-(3-[35 S]thio)triphosphate
- SR141716A
- [N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamidehydrochloride]
- HU210
- (−)-11-OH-Δ8-dimethylheptyl-tetrahydrocannabinol
- ANOVA
- analysis of variance
- PAG
- periaqueductal gray
- Received March 1, 2002.
- Accepted April 25, 2002.
- The American Society for Pharmacology and Experimental Therapeutics