Sex-specific mechanisms of tolerance for the cannabinoid agonists CP55,940 and delta-9-tetrahydrocannabinol (Δ⁹-THC)

Abstract

Rationale

Tolerance to cannabinoids could limit their therapeutic potential. Male mice expressing a desensitization-resistant form (S426A/S430A) of the type-1 cannabinoid receptor (CB₁R) show delayed tolerance to delta-9-tetrahydrocannabinol (Δ⁹-THC) but not CP55,940. With more women than men using medical cannabis for pain relief, it is essential to understand sex differences in cannabinoid antinociception, hypothermia, and resultant tolerance.

Objective

Our objective was to determine whether female mice rely on the same molecular mechanisms for tolerance to the antinociceptive and/or hypothermic effects of cannabinoids that we have previously reported in males. We determined whether the S426A/S430A mutation differentially disrupts antinociceptive and/or hypothermic tolerance to CP55,940 and/or Δ⁹-THC in male and female S426A/S430A mutant and wild-type littermates.

Results

The S426A/S430A mutation conferred an enhanced antinociceptive response for Δ⁹-THC and CP55,940 in both male and female mice. While the S426A/S430A mutation conferred partial resistance to Δ⁹-THC tolerance in male mice, disruption of CB₁R desensitization had no effect on tolerance to Δ⁹-THC in female mice. The mutation did not alter tolerance to the hypothermic effects of Δ⁹-THC or CP55,940 in either sex. Interestingly, female mice were markedly less sensitive to the antinociceptive effects of 30 mg/kg Δ⁹-THC and 0.3 mg/kg CP55,940 compared with male mice.

Conclusions

Our results suggest that disruption of the GRK/βarrestin2 pathway of desensitization alters tolerance to Δ⁹-THC but not CP55,940 in male but not female mice. As tolerance to Δ⁹-THC appears to develop differently in males and females, sex should be considered when assessing the therapeutic potential and dependence liability of cannabinoids.

Introduction

Chronic pain (defined as pain exceeding three months in duration) currently affects over 25.3 million Americans (Nahin 2015), resulting in an annual socioeconomic burden in excess of $650 billion dollars (Gaskin and Richard 2012). While opioids have been commonly used for the management of chronic non-cancer pain (Boudreau et al. 2009; Campbell et al. 2010) their prolonged use and over prescription has contributed to the current “opioid crisis” (Kolodny et al. 2015; Vowles et al. 2015). Cannabinoid-based therapies offer a non-opioid alternative for the management of long-term and chronic neuropathic pain (for a review, see Mücke et al. 2018) and elicit analgesia through activation of two cannabinoid (CB) receptors, the cannabinoid type-1 receptor (CB₁R; Matsuda et al. 1990) and the cannabinoid type-2 receptor (CB₂R; (Munro et al. 1993). While CB₁R is expressed throughout the central nervous system, CB₂R tends to be expressed in the peripheral nervous system and in immune cells (Pertwee 1997).

Despite their therapeutic potential, tolerance to cannabinoids, including Δ⁹-tetrahydrocannabinol (Δ⁹-THC), represents a barrier to their therapeutic utility. Tolerance to Δ⁹-THC has been demonstrated clinically in both heavy recreational (Jones et al. 1981; Haney et al. 1999; D’Souza et al. 2008; Gorelick et al. 2012) and medicinal (Cuttler et al. 2020) users of cannabis, as well as pre-clinically in rodent models (Bass and Martin 2000; Henderson-Redmond et al. 2020; Nealon et al. 2019; Wakley et al. 2014b; Wiley et al. 2020). Desensitization of CB₁R represents one potential neuroadaptation that can contribute to tolerance to cannabinoid-induced antinociception and hypothermia and involves phosphorylation of the receptor by a G protein-coupled receptor kinase (GRK) and subsequent recruitment of (β-arrestin protein (Sim et al. 1996; Nguyen et al. 2012). Manipulation of this pathway through point mutations at serine residues S426 and S430 prevents CB₁R desensitization in vitro (Jin et al. 1999; Daigle et al. 2008a) and in vivo (Morgan et al. 2014; Nealon et al. 2019). The effects of this mutation have been shown to be agonist-specific, delaying tolerance to Δ⁹-THC (Morgan et al. 2014) and WIN55,212–2 (Nealon et al. 2019) but not CP55,940 (Nealon et al. 2019). However, it remains unknown whether these same mechanisms that mediate tolerance to Δ⁹-THC (and/or other cannabinoid agonist) in male mice are conserved in females or occur via a different mechanism.

Clinically, women report a greater incidence and/or tactile sensitivity for chronic (Nahin 2015), experimentally induced (Riley et al. 1998) and postoperative (Aubrun et al. 2005) pain compared with men. Preclinical research implicates sex as an important modulator of the antinociceptive response and tolerance to cannabinoids (Craft et al. 2012; Lafleur et al. 2018; Wakley et al. 2014a, b; Wiley et al. 2020), as females are more sensitive and develop tolerance faster to Δ⁹-THC and CP55,940 (for a review, see Cooper and Craft 2018). With more women than men utilizing Δ⁹-THC for medicinal purposes across different medical conditions (Cuttler et al. 2016), it is critical to assess the antinociceptive properties of cannabinoids, as well as tolerance, in females. Having previously shown that disruption of GRK-βarrestin2-induced desensitization of CB₁R delays tolerance to cannabinoids in male mice, the goal of the current study was to utilize this mouse model of “delayed tolerance” and determine whether GRK/βarrestin2-induced desensitization of CB₁R delayed tolerance to Δ⁹-THC and CP55,940 is sex-specific. We also sought to determine the extent to which the acute effects of Δ⁹-THC and CP55,940 (mixed CB₁R/CB₂R agonist) are mediated by CB₁R and CB₂R in male and female mutant and wild-type mice.

Methods

Subjects

Subjects included 566 experimentally naïve, age-matched (10‒16 weeks; 20‒35 g), adult male and female S426A/S430A mutant (KI; N = 277) and wild-type (WT; N = 289) mice backcrossed for 10 +generations onto a C57BL6/J background. Desensitization-resistant S426A/S430A mice were created as previously described by replacing serines 426 and 430 with alanines in the carboxy terminal of the CB₁R (Morgan et al. 2014). Mice were group housed (3‒5/cage) during all studies on a 12:12 h light/dark cycle (lights out at 18:00) with ad libitum access to food and water. Female mice, while group housed, were not monitored for estrus cycle. Mice were weighed daily prior to any administration of drug to ensure proper dosing. Animal care procedures were conducted in accordance with NIH guidelines for the Care and Use of Laboratory Animals (2011) and with approval from Marshall University’s and Pennsylvania State University’s Institutional Animal Care and Use Committee (IACUC).

Drugs/materials

Delta-9-tetrahydrocannabinol (Δ⁹-THC) was obtained from the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD) and 5-l(l,l-dimethylheptyl)-2-[(lR,2R,5R)-5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol [(−)-CP55,940], the selective CB₁ receptor inverse agonist Rimonabant (SR141716), and the CB₂ receptor inverse agonist, SR144528 were obtained from the Cayman Chemical Company (Ann Arbor, MI). For all experiments, Δ⁹-THC, CP55,940, Rimonabant, and SR144528 were dissolved in 0.9% saline, 5% Cremaphor EL, and 5% ethanol (18:1:1 v/v/v) and administered intraperitoneally (IP) in an injection volume of 10 ml/kg, either 30 or 60 min (see below) prior to testing. Doses of Δ⁹-THC and CP55,940 were selected based on previous data obtained in our lab to produce an approximate 70% maximum possible effect (%MPE) in the tail-flick assay in male mice (Henderson-Redmond et al. 2020), with the hypothesis that females would be more sensitive to the antinociceptive effects based on the pre-clinical literature (see Cooper and Craft 2018). Doses of Rimonabant and SR144528 were based on prior work with these two compounds (Yuill et al. 2017).

Tolerance to once-daily dosing

Male and female S426A/S430A mutant (N = 59) and wild-type (N = 57) mice were assessed for tolerance to the antinociceptive and hypothermic effects of seven days of once-daily IP injection of 30 mg/kg Δ⁹-THC or 0.3 mg/kg CP55,940. Antinociception was measured using a Columbus Instruments TF-1 tail-flick analgesia meter (Columbus, OH). The heat source on the apparatus was calibrated to an intensity of 5 to elicit a tail-flick latency of 3–4 s in untreated wild-type mice. A cutoff time of 10 s was used for tail-flick testing to avoid tissue damage. Tail-flick latencies were measured before and 60 min after Δ⁹-THC or CP55,940 injection to calculate the antinociceptive responses as the percentage of maximal possible effect (%MPE) using the following calculation: %MPE=[(post-drug latency)-(pre-drug latency)]/[Predetermined cut-off time (10 s)-(pre-drug latency)] × 100. Hypothermia was assessed by taking body temperatures using a mouse rectal thermometer (Physitemp Instruments, Clifton, NJ) immediately before and 60 min after drug injections. Values were converted to the percent change in body temperature (%A). Following our finding that doses of 30 mg/kg Δ⁹-THC and 0.3 mg/kg of CP55,940 induce little-to-no antinociceptive effect in female wild-type mice, a second group of mice was assessed for daily tolerance using either 50 mg/kg Δ⁹-THC or 0.6 mg/kg CP55,940. Unfortunately, day 1 of 50 mg/kg Δ⁹-THC likewise resulted in too low of an antinocieptive response to continue the 7-day experiment; however, 0.6 mg/kg of CP55,940 was sufficient to produce a similar (85‒100%MPE) response in male (N=19) and female (N=22) wild-type and mutant mice. Therefore, tail-flick latencies and body temperature were assessed immediately prior to and 60 min following injection (IP) of 0.6 mg/kg CP55,940 once-daily for 12 consecutive days.

Cumulative dose-response tolerance testing

Naïve male and female S426A/S430A mutant (N=137) and wild-type (N=146) littermate mice were tested using a range of escalating cumulative doses. Mice were cumulatively dosed to generate dose‒response curves of 0 (vehicle only), 1,3,10, 30 and 100mg/kg Δ⁹-THC or 0(vehicle only), 0.01, 0.03, 0.1, 0.3 and 1.0 mg/kg CP55,940. Briefly, tail-flick and body temperature measurements were taken immediately prior to and 60 min following vehicle injection and 60 min after injection with vehicle or each dose of cumulative drug. As an example, to achieve cumulative dosing, 1 h after injection with either 1 mg/kg Δ⁹-THC or 0.01 mg/kg CP55,940, mice were dosed with either 2 mg/kg Δ⁹-THC or 0.02 mg/kg CP55,940 to generate cumulative doses of 3 mg/kg Δ⁹-THC and 0.03 mg/kg CP55,940 and so on. Tail-flick antinociception was calculated as %MPE and body temperature was calculated as %Δ. To determine whether the dose response curve was shifted following repeated dosing, additional, drug-naive groups of male or female S426A/S430A mutant and wild-type mice were injected once-daily with either 30 mg/kg of Δ⁹-THC or 0.3 mg/kg of CP55,940 for 3 (Δ⁹-THC only), 7 (Δ⁹-THC and CP55,940) or 18 days (CP55,940 only). Following these repeated dosing regimens, full dose‒response curves to Δ⁹-THC or CP55,940 were performed. Different time points for Δ⁹-THC and CP55,940 were used in generating post dose‒response curve to reflect previous data that have shown mice demonstrate antinociceptive tolerance to 30 mg/kg Δ⁹-THC after 7 days (Morgan et al. 2014) and to 0.3 mg/kg CP55,940 after 15 days (Nealon et al. 2019) of once-daily treatment.

Mediation of antinociception and hypothermia by CB₁ and/or CB₂ receptors

Since Δ⁹-THC and CP55,940 are both mixed cannabinoid agonists, the goal of this experiment was to determine the extent to which CB₁ and CB₂ receptors mediated Δ⁹-THC and/or CP55,940-induced antinociception and hypothermia. To accomplish this, a separate group (N=41) of male and female wild-type (N=20) and S426A/S430A mutant (N=21) mice were assessed using a within subjects design to determine the effects of Vehicle (18:1:1; Veh), 10 mg/kg Rimonabant (SR141716; CB₁A antagonist) and 10 mg/kg SR 144528 (CB₂ inverse agonist; CB₂A) alone or in combination with either 30 mg/kg Δ⁹-THC or 0.3 mg/kg CP55,940. Mice were injected once weekly (on Wednesdays) with one of the following nine treatment combinations (Veh/Veh; CB₁A/Veh; CB₂A/Veh; Veh/Δ⁹-THC; Veh/CP55,940; CB₁A/Δ⁹-THC; CB₁A/CP55,940; CB₂A/Δ⁹-THC; CB₂A/CP55,940). All mice were randomly assigned and tested in a different order across each week. Briefly, mice received a pretreatment injection (IP) of either Vehicle, 10 mg/kg CB₁A, or 10 mg/kg CB₂A. Thirty minutes later, mice were assessed for basal tail-flick and body temperature measurements as described above. By baselining after the pretreatment injection, any variation in baseline attributable to the pretreatment alone was eliminated. Immediately following baseline assessment, mice were injected (IP) with either vehicle, 30 mg/kg Δ⁹-THC, or 0.3 mg/kg CP55,940 and reassessed on the tail-flick assay and for body temperature 60 min later. All results were reported as %MPE (antinociception) and %Δ body temperature (hypothermia).

Time course for Δ⁹-THC and CP55,940

In order to determine whether the pharmacodynamic half-lives of Δ⁹-THC and CP55,940 were sufficient to justify performing dose-repose curves and at what time post injection the maximum effects of our drugs occurred, separate groups of male and female wild-type (N=45) and S426A/S403A (KI; N=40) mice were assessed to determine the duration of effects of 30 mg/kg Δ⁹-THC, 0.3 mg/kg CP55,940, or vehicle of a 5 h time course. Mice were weighed and assessed for basal body temperatures are previously described immediately prior to and at 1,2, 3,4, and 5 h following a bolus (IP) injection of either 30 mg/kg Δ⁹-THC, 0.3 mg/kg CP55,940, or an equivalent volume of vehicle (18:1:1). All data is expressed as %Δ in body temperature.

Data analyses

Sample sizes appropriate for each type of experiment were estimated based on power analysis and/or previously published experiments (Morgan et al. 2014). Sample sizes differ slightly between groups in each experiment since not all litters produced identical numbers of age-matched mutants and controls. Male and female wild-type and S426A/S430A mutant mice were randomly assigned to receive vehicle, Δ⁹-THC, or CP55,940 drug treatments and randomly assigned to time-points in experiments that tested mice on different days (0, 3, and/or 7 for dose-response shifts and across weeks for the CB₁A/CB₂A experiment). Since vaginal smears were not performed in female, they were collapsed and combined into a single data set across all stages of the estrus cycle. For each experiment, the investigator performing the experiment was blinded to all mouse genotypes. Data for the dose response shifts were analyzed using SPSS version 25.0 (IBM SPSS Statistics, Armonk, New York) to enable 4-way analyses of variance (ANOVAs), while all other data was analyzed using Prism GraphPad (7.05; GraphPad, La Jolla, CA). The effective dose (ED₅₀)_S and 95% confidence intervals (CIs) were determined from initial dose response curves generated using nonlinear regression analyses. Differences between ED50s were determined to be significant if the confidence intervals did not overlap. Two-, three-, and four-way ANOVAs were run where appropriate with genotype, day/dose, sex, and/or time point as the main factors. Since we were specifically interested in examining whether there were differences in genotype as a function of sex, we followed up three-way ANOVAs with two-way ANOVAs in daily tolerance experiments. For all repeated measure analyses done with SPSS, Mauchly’s test of sphericity was calculated to assess equal variances. Where sphericity was violated, the Greenhouse‒Geisser correction was used to reduce the probability of making a type I error. When the Greenhouse‒Geisser correction is used in reporting degrees of freedom, it has been rounded off to the nearest whole number. Bonferroni post-hoc analyses were performed when significant interaction effects were detected. All data described above are expressed as the mean ± the standard error of the mean (SEM). For all analyses, significance was set at p < 0.05.

Results

Tolerance to once-daily injections of Δ⁹-THC and CP55,940

Tolerance to the antinociceptive effects of Δ⁹-THC and CP55,940

Tolerance to the antinociceptive effects of once-daily injections of either 30 mg/kg Δ⁹-THC or 0.3 mg/kg of CP55,940 was assessed in male and female S426A/S430A and wild-type mice (Fig. 1). There was an effect of time (number of Δ⁹-THC injections) (male: F_6,162 = 2.901, p = 0.0103; female: F_6,276 = 2.531, p = 0.0211) and genotype (male: F_1,27= 10.16, p = 0.0036; female: F_1,46= 10.59, p = 0.0021) but not a time x genotype interaction (male: p = 0.71; female: p = 0.073) for both male and female mice to the antinociceptive effects of Δ⁹-THC (Figs. 1a–b). Although the S426A/S430A mutants of both sexes displayed a greater antinociceptive response to Δ⁹-THC than their wild-type littermates, there were no differences as a function of genotype in the rate at which they developed tolerance to the antinociceptive effects of Δ⁹-THC as evidenced by a lack of a significant interaction. Comparison of male versus female wild-type mice revealed a main effect of day (F_6,222 = 3.135, p = 0.0057) and of sex (F_1,37= 10.91; p = 0.0021) but not a significant day x sex interaction (p = 0.47), indicating that male mice comparatively showed increased antinociception following treatment with 30 mg/kg Δ⁹-THC. Therefore, it can be concluded that there was no difference in the rate at which tolerance to Δ⁹-THC developed in these mice.

Fig. 1.

Development of antinociceptive tolerance to once-daily 30 mg/kg of Δ⁹-THC and 0.3 mg/kg CP55,940. Differences in tolerance to the antinociceptive effects (%MPE) of once-daily administration of either 30 mg/kg Δ⁹-THC (a,b) or 0.3 mg/kg CP55,940 (c,d) in both male (squares; left) and female (circles; right) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice. Error bars represent the mean ± SEM. Data were analyzed using separate two-way ANOVAs with Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001 comparing WT to KI mice). Sample sizes for each group are shown in parentheses

Analysis of antinociception in mice treated once-daily with 0.3 mg/kg CP55,940 found an effect of genotype for both male (F_1,17 = 29.20, p < 0.0001) and female (F_1,18 = 6.758, p = 0.0181) mice such that male and female mice expressing the S426A/S430A mutation showed a greater antinociceptive response to CP55,940 compared with their wild-type littermates (Figs. 1c–d). In contrast, no effect of day (number of CP55,940 injections) in either male (p = 0.1200) or female (p = 0.5263) mice nor a significant day x genotype interaction (male: p = 0.302; female: p = 0.802) were found. These results indicated that, following 7 days of once-daily CP55,940 treatment, neither male nor female mice display tolerance to the antinociceptive effects of 0.3 mg/kg CP55,940. When comparing male and female wild-type mice, there was a main effect of sex (F_1,17 = 46.11, p < 0.0001) revealing that male mice showed a greater antinociceptive response to CP55,940 compared with female littermates.

Tolerance to the hypothermic effects of Δ⁹-THC and CP55,940

Analysis of hypothermia in mice treated once-daily with 30 mg/kg Δ⁹-THC revealed main effects of both time (number of Δ⁹-THC injections) (males: F_6,162 = 36.20, p < 0.0001; females: F_1,27 = 23.11, p < 0.0001) and genotype (males: F_1,27 = 23.11, p < 0.0001; females F_6,276 = 145.3, p < 0.0001). However, there were no time x genotype interactions (males: p = 0.079 females: p = 0.77) for mice of both sexes (Figs. 2a–b). Thus, while male and female S426A/S430A mice showed an enhanced hypothermic response to Δ⁹-THC, and all mice showed evidence of tolerance, there was no difference in the rate of tolerance to the hypothermic effects of Δ⁹-THC. In contrast to the antinociceptive effects of Δ⁹-THC, there was not an effect of sex (p = 0.830) or a sex x time interaction (p = 0.7607) on the hypothermic effects of 30 mg/kg Δ⁹-THC, indicating that male and female wild-type mice did not differ either in their response or tolerance to this dose of Δ⁹-THC.

Fig. 2.

Development of hypothermic tolerance to once-daily 30 mg/kg of Δ⁹-THC and 0.3 mg/kg CP55,940. Differences in tolerance to the hypothermic effects (%ΔBT) of once-daily administration of either 30 mg/kg Δ⁹-THC (a,b) or 0.3 mg/kg CP55,940 (c,d) in both male (squares; left) and female (circles; right) wild- type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice. Error bars represent the mean ± SEM. Data were analyzed using separate two- way ANOVAs with Bonferroni post-hoc tests (*p < 0.05, ***p < 0.001 comparing WT to KI mice). Sample sizes for each group are in parentheses

Analysis of tolerance to the hypothermic effects of 0.3 mg/kg CP55,940 found an effect of day in both male (F_6,102 = 18.95, p < 0.0001) and female mice (F_6,108 = 35.51, p < 0.0001). Interestingly, an effect of genotype was found for male (F_1,17 = 4.756, p = 0.0435) but not female mice (p = 0.1278; Fig. 2c, d) and there were no time x genotype interactions (male: p = 0.8236; female: p = 0.6125). Overall, male (but not female) S426A/S430A mice showed a greater hypothermic response to 0.3 mg/kg CP55,940, but there was no genotype difference in the rate at which they were acquiring tolerance. Separate analysis of male and female wild-type mice determined that there was not an effect of sex (p = 0.9616) or a sex x time interaction (p = 0.5763), indicating that male and female mice did not differ in their hypothermic response or the rate of tolerance to 0.3 mg/kg CP55,940.

Tolerance to antinociceptive and hypothermic effects of 0.6 mg/kg CP55,940

Given the relative lack of antinociceptive response in female mice treated with 0.3 mg/kg CP55,940, a separate group of male and female wild-type and S426A/S430A mice were assessed for tolerance to the antinociceptive and hypothermic effects of once-daily treatment with 0.6 mg/kg CP55,940. Analysis of antinociception in mice treated once-daily with 0.6 mg/kg CP55,940 found an effect of genotype for both male (F_1,17 = 10.51, p = 0.0048) and female (F_1,20= 13.16, p = 0.0017) mice such that expression of the S426A/S430A mutation showed a greater antinociceptive response to CP55,940 compared to their wild-type littermates (Fig. 3a–b). In contrast to 0.3 mg/kg, there were also effects of day (number of CP55,940 injections) for both male (F_11,187 = 2.055, p = 0.0256) and female (F_11,220=2.758, p = 0.0023), suggesting that mice begin to show evidence of tolerance development at this higher dose. As with 0.3 mg/kg CP55,940, there was not a day x genotype interaction for either male (p = 0.657) or female (p = 0.3965) mice. Thus, while the S426A/S430A mutation enhanced the antinociceptive response of 0.6 mg/kg CP55,940 compared with wild-type mice, there was no difference in the rate at which tolerance developed. In contrast to treatment with 0.6 mg/kg CP55,940, when comparing male and female wild-type mice, there was an effect of day (F_11,209 = 3.930, p < 0.0001) but neither an effect of sex (p = 0.4952) nor a sex x genotype interaction (p = 0.9594). These results indicate that male and female wild-type mice did not differ in either their response or the rate at which they developed tolerance to 0.6 mg/kg CP55,940.

Fig. 3.

Development of antinociceptive and hypothermic tolerance to once-daily 0.6 mg/kg CP55,940. Differences in tolerance to the antinociceptive (%MPE)(a,b) and hypothermic (%ΔBT)(c,d) effects of once-daily administration of 0.6 mg/kg CP55,940 (c,d) in both male (squares; left) and female (circles; right) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice. Error bars represent the mean ± SEM. Data were analyzed using separate two-way ANOVAs with Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001 comparing WT to KI mice). Sample sizes for each group are in parentheses

Analysis of tolerance to the hypothermic effects of 0.6 mg/kg CP55,940 found an effect of day in both male (F_11,187 = 37.40, p < 0.0001) and female mice (F_11,209 = 33.67, p < 0.0001). Interestingly, an effect of genotype was found for both male (F_1,17 = 32.22, p < 0.0001) and female mice (F_1,19= 13.40, p = 0.0017; Figs. 3c–d). However, there was no time x genotype interaction for either male (p = 0.1465) or female (p = 0.3074) mice. Male and female S426A/S430A mice showed a greater hypothermic response to 0.6 mg/kg CP55,940 compared with wild-type littermates, however, there was no genotype difference in the rate at which they were acquiring hypothermic tolerance. Separate analysis of male and female wild-type mice determined that while there was an effect of day (F_11,209=52.50, p < 0.0001), there was not an effect of sex (p = 0.5522) or a sex x time interaction (p = 0.5474). This indicates that male and female mice did not differ in their hypothermic response or the rate of tolerance to 0.6 mg/kg CP55,940.

Δ⁹-THC and CP55,940 dose-response curve shifts

Tolerance to the antinociceptive effects of Δ⁹-THC and CP55,940

Given that females showed a much lower antinociceptive response to both 30 mg/kg Δ⁹-THC and 0.3 mg/kg CP55,940, dose-response curves were generated to assess whether these differences were consistent across the dose‒response curve. The curves also enabled us to assess whether tolerance for these cannabinoids is different by sex or genotype using a more rigorous dose‒response experimental approach. Results from four-way (dose x genotype x sex x time) ANOVA indicated a significant four-way interaction (F₇₄₇₁ = 3.46, p = 0.001)]. Bonferroni post hoc analyses revealed that Δ9-THC dose-dependently increased antinociception (F_4,471 =69.34, p < 0.001) and that female and wild-type mice were less sensitive to the antinociceptive effects of Δ⁹-THC than male (F_1,131=4.29, p = 0.040) and S426A/S430A mutant (F_1,131 = 15.71, p < 0.001) littermates.

Subsequent three- and two-way ANOVAs revealed the following interaction effects: dose x genotype x time (F_7,471 =2.26, p = 0.027), dose x sex x time (F_7,471 =2.50, p = 0.015), sex x genotype x time (F_1,131 =4.23, p = 0.017), dose x genotype (F_4,471 = 5.79, p < 0.001), dose x time (F_7,471 = 14.28, p < 0.001), sex x genotype (F_1,131 =4.96, p = 0.028), and genotype x time (F_2,131 = 5.23, p = 0.007). Bonferroni post hoc analyses indicated that male S426A/S430A mutant mice were more sensitive to the antinociceptive effects of Δ⁹-THC than male wild-type mice across all time points tested (Days: 0 (p < 0.001), 3 (p = 0.036) and 7 (p = 0.034); Figs. 4a–c). In contrast, female S426A/ S430A mutant mice only displayed greater Δ⁹-THC-induced antinociception than wild-type females on day 0 (p < 0.001) (Figs. 4d, f). Male and female S426A/S430A mutants showed greater antinociceptive responses compared with their wild-type counterparts on day 0 at doses of 3 (females only; p = 0.031), 10 (p = 0.002), 30 (females; p < 0.001; males p = 0.002), and 100 mg/kg (females p = 0.003; males p = 0.039) Δ⁹-THC. Male S426A/S430A mice remained more sensitive to the antinociceptive effects of 10 (p = 0.015) and 30 (p = 0.002) mg/kg Δ⁹-THC at 3 days and at 30 (p = 0.019) and 100 mg/kg (p = 0.004) Δ⁹-THC at days. Tolerance to the antinociceptive effects of Δ⁹-THC developed rapidly in male mutant, male wild-type, and female mutant mice, as shown by significantly diminished responses to the antinociceptive effects of Δ⁹-THC following 3 days of once-daily Δ⁹-THC (for all p < 0.003; Figs. 4b, c). Female S426A/S430A and wild-type male mice showed evidence of near complete tolerance development (as evidenced by a lack of any antinociceptive response at the doses tested) following 3 and 7 days of once-daily Δ⁹-THC, respectively. However, male S426A/S430A mice only partially developed tolerance to the doses of Δ⁹-THC tested by day 3 that persisted even after 7 days of repeated dosing (Figs. 4c, f).

Fig. 4.

Tolerance development to the antinociceptive effects of Δ9-THC assessed via shifts in dose-response curves. Tolerance development to the antinociceptive effects (%MPE) of Δ⁹-THC in both male (squares; top row) and female (circles; bottom row) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice prior to-(solid lines; a,d), 3 days (dashed lines; b,e), or 7 days (dashed lines; c,f) following once-daily treatment with 30 mg/kg Δ⁹-THC. Error bars represent the mean±SEM; data were analyzed using a four-way ANOVA with Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001 comparing WT to KI mice). Sample sizes for each group are in parentheses

Results from a four-way ANOVA (dose x genotype x sex x time) also revealed CP55,940 dose-dependently increased antinociception (F_3,383= 181.665, p < 0.001). Bonferroni post hoc analyses following-up on a significant dose x sex interaction (F_3,383 = 8.035, p < 0.001) indicated that female mice were less sensitive to the antinociceptive effects of 0.1 (p = 0.003), 0.3 (p < 0.001), and 1.0 mg/kg (p < 0.001) CP55,940 than their male littermates. While there was not an effect of genotype (p = 0.367), there was a significant genotype x dose interaction (F_3.383 = 3.85 3, p = 0.010) with post hoc analyses revealing that, overall, S426A/S430A mice were more sensitive than wild-type mice to the antinociceptive effects of 1.0 mg/kg CP55,940 (p = 0.012). Bonferroni post hoc analyses on significant 2-[sex x time (F_2,128 = 3.236, p = 0.043); dose x time (F_6,383 = 4.086, p = 0.001)] and 3-way [dose x time x sex (F_6,383 = 3.379, p = 0.003)] interaction effects revealed an effect of time among males (F_2,77 = 7.235, p = 0.001) but not females (p = 0.603) in the development of tolerance to CP55,940. Male mice did not show any evidence of tolerance following once-daily treatment with CP55,940 for 7 days (p = 0.081) but did show evidence of tolerance (diminished antinociceptive response) after 18 days (p = 0.001) of once-daily treatment with 0.3 mg/kg CP55,940 (Fig. 5). These effects were primarily driven by tolerance development at the higher doses [0.1 (p = 0.016); 0.3 (p < 0.001); 1.0 (p = 0.013)] CP55,940. Therefore, these results suggest that S426A/S430A mutation may selectively delay the development of tolerance in male mice exclusively in respect to the antinociceptive effects of Δ⁹-THC.

Fig. 5.

Tolerance development to the antinociceptive effects of CP55,940 assessed via shifts in dose‒response curves. Tolerance development to the antinociceptive effects (%MPE) of CP55,940 in both male (squares; top row) and female (circles; bottom row) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice prior to- (solid lines; a,d), 7 days (dashed lines; b,e), or 18 days (dashed lines; c,f) following once-daily treatment with 0.3 mg/kg CP55,940. Error bars represent the mean±SEM; data were analyzed using a four-way ANOVA. Sample sizes for each group are in parentheses

While the rapid development of tolerance made it difficult to ascertain ED_50s following the commencement of once-daily injections of either 30 mg/kg Δ⁹-THC or 0.3 mg/kg CP55,940, ED_50s were generated for the initial dose-response curves. Table 1 reveals that the ED_50s were significantly less for mutant mice of both sexes than their wild-type counterparts for both Δ⁹-THC and CP55,940. Likewise, the ED₅₀ was significantly less for male wild-type mice to achieve a 50%MPE for both Δ⁹-THC and CP55,940 than for their wild-type female littermates.

Table 1.

Calculated ED₅₀ values (mg/kg) assessing the antinociceptive effects of Δ⁹-THC and CP55,940 ED₅₀ values were calculated from initial dose-response curves generated using non-linear regression analysis. Values shown are mean ED₅₀ dose and 95% confidence intervals for Δ⁹-THC and CP55,940 in naive male and female S426A/ S430A (KI) and wild-type (WT) mice.

Day		Female KI Mice	Male KI Mice	Female WT Mice	Male WT Mice
Δ9-THC	ED50	13.83*	6.956*	121.7*#	31.48*#
	95% CI	10.07‒18.91	5.475‒8.812	81.4‒265	21.4‒48.55
CP55,940	ED50	0.6168*	0.1728*	0.7476*#	0.2381*#
	95% CI	0.4158‒0.886	0.1315‒0.227	0.4719‒1.16	0.1438‒0.3993

^*= a significant difference in ED₅₀ between WT and KI groups of the same sex;

^#= a significant difference in ED₅₀ between WT male and female mice

Tolerance to the hypothermic effects of Δ⁹-THC and CP55,940

In contrast to antinociception, male and female mice did not seem to differ in hypothermic tolerance development to 30 mg/kg Δ⁹-THC and 0.3 mg/kg CP55,940. To see if this effect remained consistent, dose-response curves were generated to assess whether these effects were consistent across the dose response curve. Results from a four-way ANOVA (dose x genotype x sex x time) indicated a significant three-way (dose x genotype x time) interaction (F_7,446 = 5.00, p < 0.001). Bonferroni post-hoc analyses revealed that Δ⁹-THC dose-dependently decreased body temperature (F_3,446 = 360.16, p < 0.001). Male and female S426A/S430A mutants showed a greater hypothermic response to 3,10, 30, and 100 mg/kg Δ⁹-THC (all p < 0.03; Day 0; Fig. 6a, d) than their wild-type littermates. Female S426A/S430A mice were more sensitive to the hypothermic effects of: 10 (p = 0.006), 30(p = 0.001), and 100 (p < 0.001) mg/kg Δ⁹-THC following 3 days of once-daily Δ⁹-THC and at a dose of 100 mg/kg Δ⁹-THC (p = 0.011) following 7 days of once-daily treatment with Δ⁹-THC compared with their wild types. In contrast, male S426A/S430A mice were more sensitive to the hypothermic effects of 100 mg/kg Δ⁹-THC on day 3 (p = 0.007) and did not differ at any dose following 7 days of once-daily Δ⁹-THC treatment from male wild-type mice (Figs. 6b, c). Tolerance to the hypothermic effects of Δ⁹-THC developed rapidly in both wild-type and S426A/S430A males to the doses tested with mice registering little-to-no hypothermia following 3 days of once-daily 30 mg/kg Δ⁹-THC treatment [(p < 0.001); Figs. 6a–c)]. Female mice showed evidence of partial tolerance as to the doses tested by day 3 [compared to day 0; wild-type (p < 0.001) and S426A/S430A (p < 0.001)] and little to no hypothermic response to Δ⁹-THC (indicative of full tolerance to the doses tested) by day 7 [compared to day 3: wild-type (p = 0.029) and S426A/S430A (p < 0.001); Fig. 6d, f]. Interestingly, while the S426A/S430A mutation delayed tolerance development to the antinociceptive effects of Δ⁹-THC in males, the mutation failed to differentially mediate tolerance development in males to the hypothermic effects of Δ⁹-THC. As with antinociception, the S426A/S430A mutation attenuated tolerance to Δ⁹-THC in male but not female mice.

Fig. 6.

Tolerance development to the hypothermic effects of Δ⁹-THC assessed via shifts in dose-response curves. Tolerance development to the hypothermic effects (%ΔBT) of Δ⁹-THC in both male (squares; top row) and female (circles; bottom row) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice prior to- (solid lines; a,d), 3 days (dashed lines; b,e), or 7 days (dashed lines; c,f) following once-daily treatment with 30 mg/kg Δ⁹-THC. Error bars represent the mean±SEM. Data were analyzed using a four-way ANOVA with Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001 comparing WT to KI mice). Sample sizes for each group are shown in parentheses

Results from a four-way ANOVA (dose x genotype x sex x time) revealed CP55,940 dose-dependently decreased body temperature (F_2,264 = 591.103, p < 0.001). Bonferroni post-hoc analyses following-up significant dose x genotype (F_2,264= 10.161, p < 0.001) and dose x sex (F_2,264 = 9.247, p < 0.001) interactions indicated that female mice were less sensitive to the hypothermic effects of 0.1 (p = 0.043), 0.3 (p = 0.003), and 1.0 (p < 0.001) than their male littermates. Likewise, the same analysis revealed wild-type mice were less sensitive to the hypothermic effects of 0.3 (p = 0.018) and 1.0 (p < 0.001) mg/kg of CP55,940 when compared with S426A/S430A mice. Significant dose x time (F_4,264=27.063, p < 0.001) and sex x dose x time (F_4,264 = 2.723, p = 0.029) interactions were also found. Post hoc analyses determined that female mice showed evidence of tolerance to the hypothermic effects of CP55,940 at doses of: 0.1 (p < 0.001), 0.3 (p < 0.001), and 1.0 (p = 0.039) mg/kg, while in male mice tolerance occurred at doses of 0.03 (p = 0.002), 0.1 (p = 0.031), 0.3 (p = 0.043), and 1.0 (p < 0.001) mg/kg CP55,940 (Fig. 7). Females showed the development of hypothermic tolerance to 0.1 and 0.3 mg/kg CP55,940 following 7 (p < 0.001) and 18 (p < 0.001) days but only showed evidence of any tolerance development to the highest dose (1.0 mg/kg) after 18 days (p = 0.038) of once-daily CP55,940 injections (Fig. 6d, f). While male mice showed no evidence of tolerance development after 7 days of once-daily treatment, they did show evidence of some tolerance (as indicated by a diminished hypothermic response) to the higher doses of CP55,940 [ 0.3 (p = 0.041) and 1.0 mg/kg (p < 0.001)] following 18 days of treatment. As with Δ⁹-THC, female wild-type mice were less sensitive than their male wild-type counterparts to the antinociceptive but not hypothermic responses to CP55,940. In contrast to those mice treated with Δ⁹-THC, the S426A/S430A mutation did not appear to alter either antinociceptive or hypothermic tolerance development to CP55,940 regardless of sex.

Fig. 7.

Tolerance development to the hypothermic effects of CP55,940 assessed via shifts in dose-response curves. Tolerance development to the hypothermic effects (%ΔBT) of CP55,940 in both male (squares; top row) and female (circles; bottom row) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice prior to-(solid lines; a,d), 7 days (dashed lines; b,e), or 18 days (dashed lines; c,f) following once-daily treatment with 0.3 mg/kg CP55,940. Error bars represent the mean ± SEM. Data were analyzed using a four-way ANOVA. Sample sizes for each group are shown in parentheses

Results from a four-way ANOVA (dose x genotype x sex x time) revealed CP55,940 dose-dependently decreased body temperature (F_2,264 = 591.103, p < 0.001). Post hoc analyses following significant dose-by-genotype (F_2,264 = 10.161, p < 0.001) and dose-by-sex (F_2,264 = 9.247, p < 0.001) interactions indicated that female mice were, overall, less sensitive to the hypothermic effects of 0.1 (p = 0.043) 0.3 (p = 0.003) and 1.0 (p < 0.001) than their male littermates. Likewise, these same analyses revealed that wild-type mice were less sensitive to the hypothermic effects of 0.3 (p = 0.018) and 1.0 (p < 0.001) mg/kg of CP55,940 than their S426A/S430A littermates. Significant dose-by-time (F_4,264 = 27.063, p < 0.001) and sex x dose x time (F_4,264 = 2.723, p = 0.029) interactions were also found. Post hoc analyses determined that female mice showed evidence of tolerance development to the hypothermic effects of CP55,940 at doses of: 0.1 (p < 0.001), 0.3 (p < 0.001), and 1.0 (p = 0.039) mg/kg, while in male mice tolerance occurred at doses of: 0.03 (p = 0.002), 0.1 0 = 0.031), 0.3 0 = 0.043), and 1.0 (p < 0.001) mg/ kg CP55,940 (Fig. 6). Females showed the development of hypothermic tolerance to 0.1–0.3 mg/kg CP55,940 following 7 0< 0.001) and 18 (p < 0.001) days but only showed evidence of any tolerance development to the highest dose (1.0 mg/kg) after 18 days (p = 0.038) of once-daily CP55,940 injections (Figs. 6d, f). While male mice showed no evidence of tolerance development after 7 days of once-daily treatment, they did show evidence of partial tolerance (as indicated by a diminished hypothermic response) to the higher doses of CP55,940 [ 0.3 (p = 0.041) and 1.0 mg/kg (p < 0.001)] following 18 days of treatment. As with Δ⁹-THC, female wild-type mice were less sensitive than their male wild-type counterparts to the antinociceptive, but not hypothermic responses to CP55,940. In contrast to those mice treated with Δ⁹-THC, the S426A/S430A mutation did not appear to alter either antinociceptive or hypothermic tolerance development to CP55,940 in either male or female mice.

As with antinociception, the ED_50s for the initial dose-response curves that were generated in naïve mice were calculated for hypothermia in both male and female wild-type and S426A/S430A mice (Table 2). Unlike with antinociception, however, there were no differences between mice as either a function of sex or genotype in their hypothermic responses to either Δ⁹-THC or CP55,940 (Table 2).

Table 2.

Calculated ED₅₀ values (mg/kg) assessing the hypothermic effects of Δ⁹-THC and CP55,940 ED₅₀ values were calculated from initial dose‒response curves generated using non-linear regression analysis. Values shown are mean ED₅₀ dose and 95% confidence intervals for Δ⁹-THC and CP55,940 in naive male and female S426A/S430A (KI) and wild-type (WT) mice

Day		Female KI mice	Male KI mice	Female WT mice	Male WT mice
Δ9-THC	ED50	9.408	6.510	20.45	12.47
	95% CI	6.0‒14.47	1.996‒11.05	8.557‒33.67	5.717–20.06
CP55,940	ED50	0.2874	0.2741	0.2601	0.3272
	95% CI	0.1248‒0.4501	0.2185‒0.3296	0.0868‒0.4334	0‒1.484

Δ⁹-THC and CP55/940 mediation by CB₁ versus CB₂ receptors

CB₁ versus CB₂ mediation of Δ⁹-THC and CP55,940 antinociception

Both Δ⁹-THC and CP55,940 are mixed CB₁/CB₂ agonists. As such, male and female wild-type and S426A/S430A mice were assessed to determine the extent to which the antinociceptive effects of 30 mg/kg Δ⁹-THC and 0.3 mg/kg CP55,940 were mediated by CB₁ and/or CB₂. Among male mice, there was an effect of drug treatment (F_5,90 = 10.91, p < 0.0001) but not genotype (p = 0.2926) or a treatment x genotype interaction (p = 0.4535). Pretreatment with either the CB₁ antagonist (SR141716) or CB₂ reverse agonist (SR144528) alone did not induce an antinociceptive response (Fig. 8a). Treatment with Δ⁹-THC significantly caused antinociception in both male wild-type (p = 0.0362) and S426A/S430A (p < 0.0001) mice, an effect that was blocked in both genotypes following pretreatment with SR141716 but not SR144528 (Fig. 8c). Pretreatment with the CB₂ inverse agonist SR144528 did not reduce antinociceptive responses to Δ⁹-THC in either wild-type (p = 0.6976) or S426A/S430A (p = 0.9998) males compared with treatment of these genotypes with 30 mg/kg Δ⁹-THC alone. Results from a two-way ANOVA revealed that there was an effect of both drug (F_5,95 = 2.669, p = 0.0266) and genotype (F_1,19 = 4.795, p = 0.0412) but not a treatment x genotype interaction (p = 0.6739). The antinociceptive response in females was so low that treatment with Δ⁹-THC alone failed to elicit an antinociceptive response that differed from vehicle (Fig. 8b,d).

Fig. 8.

Mediation of the antinociceptive effects of Δ⁹-THC and CP55,940 by CB₁ and/or CB₂ receptors. Male (left panes) and female (right panes) wild-type (WT; filled) and S426A/S430A (KI; unfilled) mice were assessed for differences in antinociception response via the tail-flick assay following pretreatment with either vehicle (Veh), the CB, antagonist Rimonabant (CB₁A), or the CB₂ inverse agonist (CB₂A) SRI44528 alone (a,b) or 30 min prior to treatment with either 30 mg/kg Δ⁹-THC (c,d) or 0.3 mg/kg CP55,940 (e,f). Error bars represent the mean ± SEM; data were analyzed using a two-way ANOVA and Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 comparing mice to their Veh/Veh group; #p < 0.05 compared to the Veh/Δ⁹-THC or Veh/CP55,940). Sample sizes for each group are in parentheses

Two-way ANOVA revealed a main effect of drug (F_5,80= 15.37, p < 0.0001) but not of genotype (p = 0.8289) or a treatment x genotype interaction (p = 0.4893) in male mice. Post hoc analyses revealed that pretreatment with SR141716 or SR144528 alone did not alter antinociceptive responses compared with vehicle in either male wild-type or S426A/S430A mice. Treatment with 0.3 mg/kg CP55,940 induced an antinociceptive response in both male wild-type (p = 0.0001) and S426A/S430A mice (p = 0002) compared with vehicle, the effects of which were blocked following pretreatment with SR141716A but not SR144528. The effects were predominately mediated by CB₁, as pretreatment with SR144528 did not alter the antinociception induced by CP55,940 alone in either wild-type (p = 0.1672) or S426A/S430A (p = 0.9999) mice (Fig. 8e). Results from a two-way ANOVA revealed a main effect of drug (F_5,90 = 3.535, p = 0.0058) but not of genotype (p = 0.2956) or a treatment x genotype interaction (p = 0.5815) in female mice. Post hoc analyses reveal that this effect of drug was entirely mediated by the wild-type mice, showing an antinociceptive response following treatment with CP55,940 alone (p = 0.0232).

CB₁ versus CB₂ mediation of Δ⁹-THC and CP55,940 hypothermia

In contrast to antinociception, the hypothermic effects of both Δ⁹-THC and CP55,940 appear to be mediated entirely via CB₁ receptors (Fig. 9). The results from a two-way ANOVA revealed a main effect of drug treatment (F_5,90 = 88.02, p < 0.0001), but neither a main effect of genotype (p = 0.1526) nor a treatment x genotype interaction (p = 0.3190) among male mice. Post hoc analyses revealed that among male mice, treatment with either SR141716 or SR144528 alone did not differ from those mice treated with vehicle alone (Fig. 9a). Treatment with Δ⁹-THC alone induced a significant hypothermic response in both male wild-type (p < 0.0001) and S426A/S430A (p < 0.0001) mice (Fig. 9c) that was completely reversed by pretreatment with SR141716 but not SR144528. Likewise, among female mice, there was also a main effect of drug treatment (F_5,95 = 106.1, p < 0.0001), but neither a main effect of genotype (p = 0.0977) nor a treatment x genotype interaction (p = 0.2907). Post-hoc analyses revealed that among females, neither wild-type nor S426A/S430A mice treated with SR141716 or SR144528 alone differed in their responses form vehicle. Treatment with 30 mg/kg Δ⁹-THC induced a significant hypothermic effect in both female wild-type (p < 0.0001) and S426A/S430A (p < 0.0001) mice that was completely blocked following pretreatment with SR141716 but not SR144528. Interestingly, while there was no difference in hypothermic response in either male wild-type or S426A/S430A mice treated with Δ⁹-THC alone and those pretreated with SR144528 prior to Δ⁹-THC, post-hoc analyses revealed that female mice pretreated with SR144528 prior to receiving Δ⁹-THC showed a slight decrease in hypothermic response compared to treatment with Δ⁹-THC alone [wild-type (p = 0.0136); S426A/S430A (p = 0.0122); Fig. 9d]. Thus the effects of Δ⁹-THC on hypothermic response appears to be mediated entirely by the CB₁ receptor in males and slightly less so in female mice.

Fig. 9.

Mediation of the hypothermic effects of Δ⁹-THC and CP55,940 by CB, and/or CB₂ receptors. Male (left panes) and female (right panes) wild-type (WT; filled) and S426A/S430A (KI; unfilled) mice were assessed for differences in hypothermic response following pretreatment with either vehicle (Veh), the CB₁ antagonist Rimonabant (CB₁A), or the CB₂ inverse agonist SR144528 (CB₂A) alone (a,b) or 30 min prior to treatment with either 30 mg/kg Δ⁹-THC (c,d) or 0.3 mg/kg CP55,940 (e,f). Error bars represent the mean ± SEM; data were analyzed using a two-way ANOVA and Bonferroni post-hoc tests (*p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001 comparing mice to their Veh/Veh group; #p < 0.05 compared to the Veh/Δ⁹-THC or Veh/CP55,940). Sample sizes for each group are in parentheses

There was a main effect of drug treatment (F_5,80 = 76.13, p < 0.0001), but not of genotype (p = 0.3521) or a treatment x genotype 0 = 0–4388) interaction in male mice. Post-hoc analyses determined that treatment with 0.3 mg/kg CP55,940 induced hypothermia and that pretreatment with SR141716 (but not SR144528) completely blocked CP55,940-induced hypothermia in both wild-type and S426A/S430A mice. Pretreatment with SR144528 prior to CP55,940 slightly attenuated the hypothermic effects of CP55,940 in male S426A/S430A (p = 0–0163) but not wild-type mice (p = 0.2843) compared with treatment with CP55,940 alone. Among female mice there was a main effect of drug treatment (F_5,90 = 68.66, p < 0.0001) but not of genotype (p = 0.1344) or a treatment x genotype interaction (p = 0.6150). Post hoc analyses revealed that treatment with CP55,940 alone induced a hypothermic response in both female wild-type (p < 0.0001) and S426A/S430A (p < 0.0001) mice that was blocked when pretreated with SR141716A (but not SR 144528). Pretreatment with SR 144528 did not alter the hypothermic response induced by CP55,940 in either wild-type or S426A/S430A females compared with that of CP55,940 alone (Fig. 9f).

Δ⁹-THC and CP55,940 time course

To ensure that the pharmacodynamic half-life of Δ⁹-THC and CP55,940 sufficiently justify performing cumulative dose response curves, male and female wild-type and S426A/S430A mice were assessed just prior to and for 5 subsequent hours following bolus injections of either vehicle, 30 mg/kg Δ⁹-THC, or 0.3 mg/kg CP55,940. Following a bolus injection of 30 mg/kg Δ⁹-THC, analysis using two-way ANOVA revealed main effects of time (F_4,128 = 21.15, p < 0.0001), and group (F_3,32 = 7.679, p = 0.0005) and a significant time x group interaction (F_12,128 = 3.535, p = 0.0002). Post hoc analyses revealed the S426A/S430A mutation resulted in a much greater hypothermic response to 30 mg/kg Δ⁹-THC in males compared with all other groups, which did not differ in their hypothermic responses from each other. Likewise, with the exception of the S426A/S430A males that showed a slight increase in their hypothermic response 3 h post injection, the maximal hypothermic effects following treatment with Δ⁹-THC were evident at 1 h post injection (Fig. 10a). Female S426A/S430A and male wild-type mice showed a slightly decreased hypothermic response at 5 h following a bolus injection of 30 mg/kg Δ⁹-THC compared to hour 1, while female wild-type mice showed a slightly decreased hypothermic response at both 4 and 5 h post-injection (Fig. 10a). Results from two-way ANOVAs found a main effect of drug (F KI: F_1,13 = 29.55, p = 0.0001; F WT: F_1,13 = 23.67, p = 0.0003; M KI: F_1,9 = 188.5, p < 0.0001; M WT: F_1,13 = 37.34, p < 0.0001) when comparing each sex/genotype grouping that received drug to their counterparts that received vehicle. Post hoc analyses revealed that the effects of Δ⁹-THC persisted beyond 5 h as all groups significantly differed from their vehicle counterparts at all time points assessed (Fig. 10a).

Fig. 10.

Hypothermic time course following bolus treatment with 30 mg/kg Δ⁹-THC or 0.3 mg/kg CP55,940. Male (squares) and female (circle) wild-type (WT; filled) and S426A/S430A mutant (KI; unfilled) mice were assess for the hypothermic effects just prior to and hourly for 5 h following a bolus injection of either 30 mg/kg Δ⁹-THC (a) or 0.3 mg/kg CP55,940 (b). A separate group of male and female wild-type and mutant mice were also assessed at the same intervals for alterations in body temperature following a bolus injection of an equivalent volume of vehicle (open symbols with lines through them). Error bars represent the mean±SEM; data were analyzed using a two-way ANOVA and Bonferroni post-hoc tests (*p < 0.05 compared to hour 1 of drug treated mice). Sample sizes for each group are in parentheses

Analysis following a bolus injection of CP55,940 revealed main effects of time (F_4,100 = 53.91, p < 0.0001) and group (F_3,25 = 4.260, p = 0.0146) but not a time x group interaction (p = 0.0957). Post hoc analyses revealed that the main effect of group was driven by the differences in hypothermic responses between the female wild-type and male S426A/S430A mice, with the greater response elicited by the latter (Fig. 10b). All groups showed their greatest hypothermic response approximately 1 h following CP55,940 administration. In contrast to mice in the Δ⁹-THC-treatment group, the effects of CP55,940 did not persist beyond 5 h. For each sex/genotype grouping there was a main effect of drug: (F KI: F_1,10 = 21.16, p = 0.0010; F WT: F_1,11 = 11.50, p = 0.0060; M KI: F_1,10 = 23.62, p = 0.0007; M WT: F_1,10 = 5.504, p = 0.0409) and a significant drug x time interaction: (F KI: F_{4 40} = 12.72, p < 0.0001; F WT: F_4,44 = 5.821, p = 0.0008; M KI: F_4,40 = 30.49, p < 0.0001; M WT: F_4,40 = 8.001, p < 0.0001) when comparing them to their vehicle-treated counterparts. Post hoc analyses revealed that any effect of drug had dissipated after 3 h in all groups except S426A/S430A males, which took 4 h to return to basal levels.

Discussion

The goal of this study was two-fold. The first goal was to ascertain whether there were sex differences in tolerance to cannabinoid agonists Δ⁹-THC and CP55,940. The second was to determine whether tolerance development to Δ⁹-THC and/or CP55,940 in female mice occurred via the same mechanisms demonstrated in male mice. We found that female mice were not only less sensitive to the antinociceptive effects of Δ⁹-THC and CP55,940, but that disruption of GRK-βarrestin2-induced desensitization of CB₁R (as assessed via the S426A/S430A mutation) selectively disrupted tolerance development to 30 mg/kg Δ⁹-THC but not 0.3 mg/kg CP55,940 in male mice only. These effects were specific for the antinociceptive effects of Δ⁹-THC as male and female mice did not differ in their antinociceptive responses to Δ⁹-THC, and the point mutations did not alter the rate of tolerance to the hypothermic effects of Δ⁹-THC in either male or female mice. Finally, we determined that the observed effects were mediated almost entirely via CB₁ receptors despite both Δ⁹-THC and CP55,940 being mixed CB₁/CB₂ receptor agonists.

Our findings are consistent with those previously published from our lab showing that disruption of βarrestin2-mediated desensitization through mutation of its binding sites (specifically serines 426 and 430 in the distal tail) can alter antinociceptive sensitivity and tolerance to select CB₁R agonists, including Δ⁹-THC (Morgan et al. 2014) and WIN55,212‒2 (Nealon et al. 2019) but not CP55940 (Nealon et al. 2019) in male mice. The S426A/S430A mutation decreased the ED50 in both male (6.956 mg/kg) and female (13.83 mg/kg) mutants compared with their male (31.48 mg/kg) and female (121.7 mg/kg) wild-type littermates for Δ⁹-THC, but not for CP55,940 (Table 1). In our dose-response studies we saw that the S426A/S430A mutation was able to delay tolerance to Δ⁹-THC in males, but interestingly not in females, suggesting that tolerance development in females likely occurs via other mechanisms. However, these effects were once again specific for Δ⁹-THC as there was no effect on tolerance development to CP55,940 in either male or female mice.

Our findings add credence to the idea of biased agonism which has previously been shown in the mouse brain such that different CB₁R ligands, including WIN55,212–2, Δ⁹-THC, and the synthetic anandamide analog arachidonyl-2-chloroethylamide (ACEA), were able to differentially activate intracellular signaling cascades (Diez-Alarcia et al. 2016). This is consistent with other findings showing that various cannabinoid agonists, including WIN55,212–2, Δ⁹-THC, HU210, and anandamide, were able to induce different conformations of the CB₁R which subsequently resulted in differential abilities to activate select G protein-coupled receptor (GPCR) subtypes (Glass & Northup 1999). For example, WIN55,212‒2 was shown to be a full agonist in the activation of Gα_i but only a partial agonist in the activation of Gα_o, while Δ₉-THC was only a partial agonist in the activation of both receptor subtypes (Glass and Northup 1999). As was shown in our dose-response studies and consistent with our other published studies (Morgan et al. 2014; Henderson-Redmond et al. 2020), the S426A/S430A mutation, while able to delay tolerance to the antinociceptive effects of Δ⁹-THC in male mice, did not prevent tolerance from developing to the doses tested. One mechanism that has been implicated in further mediating antinociceptive tolerance development to Δ⁹-THC is c-Jun N-terminal kinases (JNK) signaling. Our own work has shown that though blockade of JNK signaling can prevent tolerance development to the antinociceptive effects of Δ⁹-THC, it has no effect on tolerance to WIN55,212–2 and actually accelerates antinociceptive tolerance to CP55,940 in male S426A/S430A mice. This provides further evidence that different mechanisms likely coordinate to mediate tolerance to the antinociceptive effects of various CB₁R agonists (Henderson-Redmond et al. 2020). Whether or not JNK signaling may likewise mediate antinociceptive tolerance to Δ⁹-THC in females has yet to be examined and should be investigated.

It is also possible that our effects were more evident following treatment with Δ⁹-THC because Δ⁹-THC is only a partial agonist of the CB₁ receptor. In contrast, CP55,940 is considered a full CB₁R agonist and any differences, such as a low receptor reserve, could make it more difficult to ascertain differences in sensitivity and/or tolerance development to a full versus partial agonist. CB₁R reserve has been shown to be quite large in the central nervous system with maximal effects of WIN55,212‒2 occurring at only 7.5% occupancy in the hippocampus (Gifford et al. 1999). Given that antinociceptive responses to both Δ⁹-THC and CP55,940 were much more exaggerated than the hypothermic responses, it is possible that the receptor reserves may differ as a function of brain region and may be tissue specific. In fact, studies examining cannabinoid receptor activation of G-proteins found that there were significant regional differences in the brain in the ratio of cannabinoid-activated G-proteins to cannabinoid receptors, and that the cerebellum (typically a region with lower receptor levels) had a receptor reserve demonstrating that degree of relative cannabinoid activity in a region is not necessarily correlated to the relative receptor density of cannabinoid receptors (Breivogel et al. 1997). More recently, Grim et al. (2016) used transgenic mice to demonstrate that a 50% decrease in receptor reserve showed the most pronounced effects when assessing the antinociceptive effects of cannabinoids. Altering receptor reserve drastically reduced both the antinociceptive potency and efficacy of Δ⁹-THC and CP55,940. These effects were most notable with use of the weaker, partial agonist Δ⁹-THC compared with the medium efficacy agonist, CP55,940. The authors concluded that these results were consistent with lower CB₁R reserve in areas that mediate antinociception, including the periaquaductal gray and dorsal horn of the spinal cord (Grim et al. 2016). These findings are consistent with other studies that show that these regions have much lower CB₁R densities compared to other brain regions, including the cerebellum, hippocampus, and cortex (Herkenham et al. 1990; Matsuda et al. 1990).

These effects were not noted in Nealon et al. (2019) as disruption of βarrestin2-mediated desensitization resulted in a dramatic delay in antinociceptive tolerance development to WIN55,212–2. However, agonist-stimulated [³⁵S] GTPγS binding has determined that Δ⁹-THC is a weak, partial agonist (Sim et al. 1996; Burkey et al. 1997), anandaminde and CP55,940 as partial agonists in inhibiting adenylyl cyclase and calcium conductance, and WIN55,212‒2 is a full agonist (Breivogel and Childers 1998). Tolerance to the effects of Δ⁹-THC is thought to occur primarily via desensitization, while tolerance to CP55,940 may occur primarily via downregulation of internalized receptors. Indeed, while the process of desensitization is required for both desensitization and internalization, different sets of phosphorylation sites have been shown to mediate these effects (Daigle et al. 2008b; Jin et al. 1999). For example, in vitro mutation of serines 426 and 430 prevented βarrestin-mediated desensitization but not endocytosis (Jin et al. 1999), whereas mutation of 6 different serine and threonine sites (T461, S463, S465, T466, T468, and S469) interfered with βarrestin-mediated internalization, likely by interfering with βarrestin-stimulated endocytosis (Daigle et al. 2008b). Taken in combination with the observation that there are considerable regional differences in the rate and magnitude of CB₁R desensitization and downregulation (for a review, see Sim-Selley 2003), this may help explain why we see different magnitudes of antinociception with different CB₁R agonists.

In addition to biased agonism, more recent evidence suggests that sex-biased signaling may underlie sex differences both in drug sensitivity and/or vulnerability to abuse or dependence (for a review, see Valentino et al. 2013). Sex has been shown to determine the state of coupling of G proteins and βarrestin2 (βarr2) to the corticotropin releasing factor (CRF) receptor 1 subtype (CRF1; Bangasser et al. 2010). In males CRF1 is more efficiently coupled to βarr2 recruitment facilitating robust receptor internalization of this receptor in locus coeruleus, while in females coupling of this pathway and internalization is lower, rendering the locus coeruleus in females more sensitive to CRF-induced hyperstimulation during times of stress (Valentino and Bangasser 2016). Given the shared characteristics of many GPCRs, Valentino and colleagues (2013) postulate the likelihood of sex-specific signaling pathways for other GPCRs. Sex differences in the coupling of mu opioid receptors to G proteins has been suggested as a possible mechanism for differences in opioid withdrawal in the spinal cord of male rats (Chakrabarti et al. 2012). These studies raise the possibility that different signaling cascades might be engaged in agonist- and sex-specific manners by CB₁R activation in male and female rodents, a possibility that should be explored by future work.

One of our most unexpected, yet consistent findings was that wild-type female mice were much less sensitive to the antinociceptive effects of Δ⁹-THC and CP55,940 than their male littermates. Given that the majority of preclinical studies have shown female rats to be more sensitive to cannabinoid-induced antinociception than their male counterparts (Craft et al. 2012; Romero et al. 2002; Tseng and Craft 2001; Wakley et al. 2014b; Wakley and Craft 2011; Wiley et al. 2020), we anticipated similar results in our mice. As our results contrasted with what was hypothesized, it was difficult to ascertain the extent to which tolerance development may have varied in the once-daily tolerance studies as the initial lack of response likely resulted in a floor effect. To address this caveat, we generated a series of dose‒response curves in an attempt to identify differences in tolerance through the assessment of rightward ED₅₀ shifts. While we were unable to generate ED_50s following tolerance development due to the relative flatness of the curves using the doses tested, we were able to generate ED_50s for our initial dose‒response curves. We found that females required significantly more drug to achieve the same level of antinociception as males for both Δ⁹-THC (female ED₅₀: 121.7 mg/kg; male ED₅₀: 31.48 mg/kg) and CP55,940 (female ED₅₀: 0.7476; male ED₅₀: 0.2381). Interestingly, male and females did not differ in their ED_50s for hypothermia for either Δ⁹-THC (female ED₅₀: 20.45; male ED₅₀: 12.47) or CP55,940 (female ED₅₀: 0.2601; male ED₅₀: 0.3272), reinforcing the conclusion that our observed sex differences are specific for the antinociceptive effects of cannabinoids.

Although our finding of sex differences in acute response were the opposite of what we hypothesized, our findings are consistent with previous findings that sex modulates the antinociceptive response to cannabinoids and that females develop tolerance faster than males to antinociceptive effects of Δ⁹-THC. It is possible that sex difference in acute response for mice that we observe could be due to species differences between mice and rats. Recent work by Wiley et al. (2020) found that female rats were able to exhibit drug discrimination for Δ⁹-THC at lower doses than their male littermates, suggesting that female rats are more sensitive to the subjective effects of Δ⁹-THC. Interestingly, similar work by the same group found that sex did not alter discrimination for Δ⁹-THC between male and female mice (Wiley et al. 2020). Other physiological differences between male and female rodents that might impact the acute response to cannabinoids include: adipose tissue percentage and distribution that could modify the bioavailability of cannabinoids (Cortright et al. 1997), the rate of cannabinoid metabolism (Borys and Karler 1979), and pharmacokinetic differences in the production of active metabolites (Narimatsu et al. 1991; Tseng and Craft 2001).

Previous work has also found differences in CB₁R density and desensitization between male and female rodents (Castelli et al. 2014; de Fonseca et al. 1994; Farquhar et al. 2019; González and Cebeira 2005 but see Wiley et al. 2020). Another possibility that could explain our findings is that sex differences in cannabinoid response might be strain dependent and that our observations might be specific for the C57BL/6 strain upon which our mice are backcrossed. For example, Kest et al. (1999) measured morphine-induced antinociception in 11 common mouse strains and found no sex difference in seven strains, increased sensitivity to morphine in female mice in one strain, and decreased sensitivity to morphine for females in the other 3 strains, which included the C57BL/6 strain (Kest et al. 1999). Thus, it is important to repeat the current work using additional mouse strains that are commonly used for preclinical pain research.

This study shows that sex differences in cannabinoid response are also response-specific. For example, despite clear sex differences in Δ⁹-THC-induced antinociception, these same doses did not cause sex differences in the acute hypothermic response. This raises the possibility that response- and sex-specific differences in acute cannabinoid response might be due to underlying differences in cannabinoid signaling within the specific brain regions such as the periaqueductal gray (antinociception), spinal cord (antinociception), and hypothalamus (hypothermia) that mediate these responses. Although there are no sex differences in many regions of the mouse brain such as the cerebellum (Wiley et al. 2020), very little is known about possible sex differences in CB₁R levels in areas that control antinociception. It is also possible that response-specific sex differences in cannabinoid response could be due to differences in other important endocannabinoid signaling components such endocannabinoids levels, G protein coupling, CB₂R levels, and beta-arrestin expression. For example, previous work has demonstrated that endocannabinoid levels vary as a function of sex [males had lower levels of both 2-arachidonoylglycerol (2-AG) and anandamide (AEA) than females] and estrous cycle (females had significantly greater levels of 2-AG and AEA in diestrus compared to all of cycle stages) in the hypothalamus of Sprague-Dawley rats (Bradshaw et al. 2006).

Estrogen has been reported to interfere with the ability of Δ⁹-THC to effectively bind CB₁R (Wakley et al. 2014a, b) while hormone fluctuations across the estrous cycle have been shown to alter the efficacy of G protein coupling to CB₁R (Riebe et al. 2010). Acute treatment of ovariectomized (OVX) female rats with a bolus injection of estradiol has been shown to decrease CP55,940-stimu- lated [³⁵S]GTPγS binding in the cortex and hippocampus, suggesting that estradiol can suppress CB₁R signaling (Mize and Alper 2000). However, another work in rats has found that CB₁R protein and WIN55,212‒2-stimulated [³⁵S]GTPγS binding in the striatum is decreased in OVX females, suggesting an opposite finding that estrogen and/or progesterone enhance CB₁R expression and signaling (Winsauer et al. 2011). Two studies suggested a role for sex hormones in modulating the acute effects of cannabinoids in female mice. One study demonstrated that OVX increased WIN55,212-mediated antinociception in female mice and that this effect was prevented by estrogen but not progesterone replacement (Kalbasi Anaraki et al. 2008). A second study showed that mifepristone and metyrapone, two anti-progesterone drugs, enhanced the acute hypothermic effects of 25 mg/kg Δ⁹-THC in female mice (Pryce et al. 2003). These studies suggest that sex hormones may suppress cannabinoid responses in female mice, which is consistent with our finding that females display decreased antinociceptive and hypothermic responses compared to male littermates. Interestingly, despite not controlling for estrus cycle, we observe relatively low levels of variability among female mice, suggesting that hormonal fluctuations across the estrus cycle have a limited impact on cannabinoid-induced hypothermia and acute thermal antinociception.

Evidence also exists suggesting sex differences in the antinociceptive effects of Δ⁹-THC might be due to differences in the relative expression of CB₁R and CB₂R (Craft et al. 2012). In the current study, we utilized the CB₁R antagonist rimonabant and the CB₂R antagonist SR 144158 to determine the degree that sex differences in cannabinoid response were mediated by CB₁R and/or CB₂R. Our findings found that the hypothermic effects of Δ⁹-THC and CP55,940 in all mice were mediated predominately by CB₁R. Additionally, the antinociceptive effects of both Δ⁹-THC and CP55,940 were mediated almost exclusively by CB₁R in male mice of both genotypes. Given the decreased response in females, it was difficult to fully ascertain whether the antinociceptive effects were also primarily mediated through CB₁R or whether signaling at CB₂R was involved in this response.

Evidence suggests that sex influences a variety of cannabinoid-related outcomes in humans, including the prevalence of cannabinoid use disorder (CUD; Hernandez-Avila et al. 2004; Khan et al. 2013), abuse liability (Cooper and Haney 2014), withdrawal severity (Copersino et al. 2006; Levin et al. 2010; Sherman et al. 2017), treatment outcomes (Cuttler et al. 2020), and neuronal activity in those with CUD (Wetherill et al. 2015). Although men display higher prevalence of CUD in a clinical setting (Hoppe et al. 2015), women show greater escalation of intake, and effect that has been termed “telescoping” (Hernandez-Avila et al. 2004; Ehlers et al. 2010). This telescoping effect suggests that females may show more rapid tolerance development to cannabis than males. Some studies report that women exhibit greater “liking” and “wanting” associated with cannabis consumption (Cooper and Haney 2014), while others have found that men report increased euphoria and subjective effects than females (Penetar et al. 2005). Interestingly, women reported greater subjective effects than men following lower doses (5 mg) of Δ⁹-THC while the converse is seen at higher doses [15 mg; (Fogel et al. 2017)]. For nonusers of marijuana, men were more sensitive to 10‒25 mg Δ⁹-THC than women, although this difference was absent for experienced, regular marijuana smokers (Haney 2007). Therapeutically, men and women found that use of inhaled cannabis decreased their subjective pain ratings for headache and migraine (Cuttler et al. 2020); however, pain relief was greater in males than females. Additional studies in experienced cannabis users found that females are less sensitive to the tachycardic and analgesic (Cooper and Haney 2016) effects of cannabinoids. However, it remains unclear whether this decreased sensitivity might be due to increased tolerance in these experienced frequent users.

In this study, male and female desensitization-resistant S426A/S430A mice demonstrated increased sensitivity to Δ⁹-THC and CP55,940 in an acute antinociceptive pain model. Only male S426A/S430A mice demonstrated an agonist-specific delay in tolerance to the antinociceptive effects of Δ⁹-THC, suggesting that the mechanisms of cannabinoid tolerance may be both sex- and agonist-specific. Although we find that tolerance to Δ⁹-THC in male mice is partially mediated by GRK-βarrestin2-induced desensitization of CB₁R, tolerance to CP55,940 is relatively unaffected by disruption of this mechanism. As such, it is likely that tolerance development to cannabinoids is mediated by different mechanisms in males and females, and we posit that sex should be carefully and thoroughly considered when assessing the efficacy of cannabinoid-based therapies, particularly when utilized in pain management.

Abbreviations

2-AG

2-Arachidonoylglycerol

ACEA

Arachidonyl-2-chloroethylamide

AEA

Anandamide

ANOVA

Analyses of variance

βArr2

Beta-arrestin 2

%ΔBT

Percent change in body temperature

CB

Cannabinoid

CB₁/CB₁R

Type-1 cannabinoid receptor

CB₂/CB₂R

Type-2 cannabinoid receptor

CB₁A

SR141716A (CB₁R antagonist)

CB₂A

SR144528 (CB₂ inverse agonist)

CI

Confidence interval

CP55,940

5–1(1,1-Dimethylheptyl)-2-[(lR,2R,5R)-5-hydroxy-2-(3-hydroxypropyl)cyclohexyl]-phenol

CRF

Corticotropin-releasing factor

CRF1

Corticotropin-releasing factor Type I receptor

CUD

Cannabinoid use disorder(s)

Δ⁹-THC

Delta-9-tetrahydrocannabinol

ED₅₀

Mean effective dose

GPCR

G protein-coupled receptor

GRK

G protein-coupled receptor kinase

GTPγS

Guanosine 5’-O-[gamma-thio]triphosphate

HU210

(6AR, 10aR)-9-(hydroxymethyl)-6,6-dimethyl-3-(2-methyloctan-2-yl)-6a,7,10,10a-tetrahydrobenzo[c]chromen-1-ol

IP

Intraperitoneal

JNK

C-Jun N-terminal kinases

KI

Knock-in

%MPE

Percentage of maximal possible effect

OVX

Ovariectomized

SEM

Standard error of the mean

SC

Subcutaneous

Veh

Vehicle

WIN55,212‒2

R-(+)-[2,3-Dihydro-5-methyl-3-[(morpholinyl-methyl]pyrrolo[l,2,3-de]-l,4-benzoxazinyl]-(l-mapthalenyl)methanone mesylate

WT

Wild-type