Abstract
Opioid use disorder reflects a major public health crisis of morbidity and mortality in which opioid withdrawal often contributes to continued use. However, current medications that treat opioid withdrawal symptoms are limited by their abuse liability or lack of efficacy. Although cannabinoid 1 (CB1) receptor agonists, including Δ9-tetrahydrocannabinol, ameliorate opioid withdrawal in both clinical and preclinical studies of opioid dependence, this strategy elicits cannabimimetic side effects as well as tolerance and dependence after repeated administration. Alternatively, CB1 receptor positive allosteric modulators (PAMs) enhance CB1 receptor signaling and show efficacy in rodent models of pain and cannabinoid dependence but lack cannabimimetic side effects. We hypothesize that the CB1 receptor PAM ZCZ011 attenuates naloxone-precipitated withdrawal signs in opioid-dependent mice. Accordingly, male and female mice given an escalating dosing regimen of oxycodone, a widely prescribed opioid, and challenged with naloxone displayed withdrawal signs that included diarrhea, weight loss, jumping, paw flutters, and head shakes. ZCZ011 fully attenuated naloxone-precipitated withdrawal-induced diarrhea and weight loss and reduced paw flutters by approximately half, but its effects on head shakes were unreliable, and it did not affect jumping behavior. The antidiarrheal and anti–weight loss effects of ZCZ0111 were reversed by a CB1 not a cannabinoid receptor type 2 receptor antagonist and were absent in CB1 (−/−) mice, suggesting a necessary role of CB1 receptors. Collectively, these results indicate that ZCZ011 completely blocked naloxone-precipitated diarrhea and weight loss in oxycodone-dependent mice and suggest that CB1 receptor PAMs may offer a novel strategy to treat opioid dependence.
SIGNIFICANCE STATEMENT Opioid use disorder represents a serious public health crisis in which current medications used to treat withdrawal symptoms are limited by abuse liability and side effects. The CB1 receptor positive allosteric modulator (PAM) ZCZ011, which lacks overt cannabimimetic behavioral effects, ameliorated naloxone-precipitated withdrawal signs through a CB1 receptor mechanism of action in a mouse model of oxycodone dependence. These results suggest that CB1 receptor PAMs may represent a viable strategy to treat opioid withdrawal.
Introduction
Misuse of opioids and the prevalence of individuals with opioid use disorder remain a significant public health problem in the United States. Recent 2019 estimates are that 9.7 and 1.4 million individuals aged 12 or older misused opioids and developed opioid use disorder, respectively (Substance Abuse and Mental Health Services Administration, 2020). Although current medications used to treat opioid dependence attenuate withdrawal symptoms, such as diarrhea, emesis, hand tremors, and anxiety (Farrell, 1994; Wesson and Ling, 2003), they also possess abuse liability (e.g., methadone and buprenorphine) (Cicero and Inciardi, 2005) and are not fully effective for all patients (e.g., lofexidine/clonidine) (Kuhlman et al., 1998; Stolbach and Hoffman, 2020). Thus, a great need exists to develop new efficacious pharmacotherapies, which lack abuse potential, to alleviate opioid withdrawal.
A case report from the 19th century describing a patient who was opium-dependent and treated with cannabis extract (Birch, 1889) suggested that cannabinoid-based medications may effectively treat opioid dependence. A clinical trial conducted more than a century later demonstrated that the primary active constituent of cannabis, Δ9-tetrahydrocannabinol (THC), alleviates withdrawal symptoms in patients who are opioid-dependent (Bisaga et al., 2015; Lofwall et al., 2016). Nevertheless, unwanted side effects of THC, including tachycardia, somnolence, and intoxication (Jicha et al., 2015; Lofwall et al., 2016), hinder its application to treat opioid use disorder. THC produces most of its pharmacological effects through the activation of two G-protein–coupled receptors, cannabinoid receptor type 1 (CB1; Matsuda et al., 1990) and cannabinoid receptor type 2 (CB2; Munro et al., 1993). THC and other CB1 receptor agonists effectively attenuate withdrawal signs in opioid-dependent rodents (Hine et al., 1975a,b; Bhargava, 1976a,b; Vela et al., 1995; Lichtman et al., 2001; Cichewicz and Welch, 2003; Ramesh et al., 2011; Gamage et al., 2015), although this pharmacological approach elicits acute cannabimimetic side effects and results in tolerance and physical dependence after repeated administration (Wiley and Martin, 2003; Grim et al., 2016; Trexler et al., 2018, 2019). Alternatively, inhibitors of monoacylglycerol lipase (MAGL; Dinh et al., 2002) and fatty acid amide hydrolase (FAAH; Cravatt et al., 1996), which hydrolyze the respective endogenous cannabinoids 2-arachidonoylglycerol (Mechoulam et al., 1995; Sugiura et al., 1995) and anandamide (N-arachidonylethanolamine; Devane et al., 1992), attenuate naloxone-precipitated and spontaneous withdrawal signs in morphine-dependent mice with reduced cannabimimetic side effects (Ramesh et al., 2011, 2013; Gamage et al., 2015; Wills et al., 2016). However, CB1 receptor downregulation and desensitization and physical dependence after repeated administration of high doses of MAGL inhibitors (Schlosburg et al., 2010, 2014; Ramesh et al., 2011) represent a challenge for this approach.
CB1 receptor positive allosteric modulators (PAMs) offer therapeutic potential to treat opioid dependence and other conditions without eliciting side effects associated with THC. Indeed, the CB1 receptor PAMs ZCZ011 and GAT211 lack overt behavioral cannabimimetic effects but produce antinociceptive effects in multiple rodent models of pain (Ignatowska-Jankowska et al., 2015a; Slivicki et al., 2017, 2020; Thapa et al., 2020) as well as reduce withdrawal signs in cannabinoid-dependent mice (Trexler et al., 2019). These molecules are believed to bind at allosteric site(s) on the CB1 receptor that results in a conformational change of the orthosteric site to enhance the binding and efficacy of endogenous cannabinoids and elicit allosteric agonist effects on their own (Dopart et al., 2018; Tseng et al., 2019) and may also be classified as a positive allosteric agonist (Kenakin, 2013).
Here we tested whether the CB1 receptor PAM ZCZ011 attenuates naloxone-precipitated withdrawal signs using an established mouse model of oxycodone-dependence (Enga et al., 2016; Carper et al., 2021). Oxycodone represents a widely used prescription opioid misused by 3.2 million people aged 12 or older in the United States (Substance Abuse and Mental Health Services Administration, 2020). Oxycodone-dependent male and female mice were challenged with naloxone to precipitate somatic withdrawal signs (i.e., jumps, paw flutters, and head shakes), diarrhea, and body weight loss. We compared the actions of ZCZ011 on naloxone-precipitated withdrawal signs to those of oxycodone, which served as a positive control. Lastly, we employed the CB1 receptor inverse agonist/antagonist rimonabant, the CB2 receptor antagonist SR144528, and constitutive CB1 (−/−) mice (Zimmer et al., 1999) to assess cannabinoid receptor involvement of the antiwithdrawal effects of ZCZ011.
Materials and Methods
Animals
Male and female ICR mice (minimum 8 weeks of age; Envigo, Indianapolis, IN) with respective body weights of 35–45 g and 30–40 g served as subjects. ICR mice were delivered at 7 weeks followed by 1 week of habituation in VCU's vivarium before testing began. Age-matched male and female CB1 (+/+) and CB1 (−/−) mice derived from CB1 (+/−) breeding pairs bred within the Mutant Mouse Core at VCU also served as subjects. These transgenic mice were created by Zimmer and colleagues (1999) and have been backcrossed onto a C57BL/6J background for at least 15 generations. The subjects were housed up to four mice per cage, in a light- (12-hour light/dark cycle; lights on at 0600), temperature- (20–22°C), and humidity-controlled (55 ± 10%) Association for Assessment and Accreditation of Laboratory Animal Care International–approved facility at VCU. Mice received water and standard rodent chow ad libitum. Animal protocols were approved by the Institutional Animal Care and Use Committee at VCU and were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Drugs
Oxycodone hydrochloride [(5α)-4,5-epoxy-14-hydroxy-3-methoxy-17-methylmorphinan-6-one]; naloxone hydrochloride [(1S,5R,13R,17S)-10,17-dihydroxy-4-(prop-2-en-1-yl)-12-oxa-4-azapentacyclo[9.6.1.01,13.05,17.07,18]octadeca-7(18),8,10-trien-14-one]; rimonabant, the CB1 receptor inverse agonist/antagonist [N-(piperidin-1-yl)-5-(4-chlorophenyl)-1-(2,4-dichlorophenyl)-4-methyl-1H-pyrazole-3-carboxamide HCl]; and SR144528, the CB2 receptor antagonist [N-[(1S)-endo-1,3,3,-trimethylbicyclo[2.2.1]heptan-2-yl]-5-(4-chloro-3-methylphenyl)-1-(4-methylbenzyl)-pyrazole-3-carboxamide] were obtained from the National Institute on Drug Abuse Drug Supply Program (Bethesda, MD). ZCZ011, the CB1 receptor PAM [6-methyl-3-(2-nitro-1-(thiophen-2-yl)ethyl)-2-phenyl-1H-indole], was synthesized based on Ignatowska-Jankowska and colleagues (2015a) and improved upon according to an established protocol in the Lu laboratory (see Supplemental Information for revised synthesis). ZCZ011, rimonabant, and SR144528 were dissolved in a vehicle solution consisting of a mixture of ethanol, alkamuls-620 (Sanofi-Aventis, Bridgewater, NJ), and saline (0.9% NaCl; sterile) in a 1:1:18 ratio. Oxycodone and naloxone were dissolved in sterile 0.9% saline. All injections were administered in a volume of 10 ml/kg of body weight. Oxycodone and naloxone were administered via the subcutaneous route of administration, whereas ZCZ011, rimonabant, and SR144528 were given via intraperitoneal injection.
Naloxone-Precipitated Oxycodone Withdrawal Model
To induce oxycodone dependence, counterbalanced groups of male and female mice were administered escalating doses of oxycodone over an 8-day period as previously described (Enga et al., 2016; Carper et al., 2021). In brief, mice were administered subcutaneous injections of 9.0, 17.8, 23.7, and then 33.0 mg/kg oxycodone (or saline) twice a day separated by approximately 7 hours on days 1–2, 3–4, 5–6, and 7–8, respectively. On day 9, mice were administered 33 mg/kg oxycodone (or saline) followed by 1 mg/kg naloxone 2 hours later to precipitate withdrawal (Enga et al., 2016; Carper et al., 2021), and withdrawal signs were assessed as previously described with some modifications (Ramesh et al., 2011).
Plexiglas chambers were used to house mice during withdrawal assessment, with each chamber constructed of white sides and white bottom panels, a clear perforated top panel (eight 1/4-inch diameter holes for ventilation), a sliding clear front panel (23 cm H), and a sliding mirrored back panel (23 cm H). The chambers were enclosed in sound-attenuating cabinets that contained an indirect filtered light-emitting diode light source and fans for air circulation and white noise. Each cabinet contained a mounted 2.8–12.0mm varifocal lens mini-USB camera (Ailipu Technology Co., Ltd, Guangdong, China) that recorded mice through the clear front panel, and videos were saved using ANY-maze video tracking software (Stoelting Co., Wood Dale, IL). All recorded videos were randomized and scored by one primary trained observer who was blinded to treatment condition or genotype using ANY-maze software and ODLog (v2.7.2 for Windows; Macropod), respectively. A subset of 16 videos of oxycodone-dependent mice were scored by a second trained observer to ensure inter-rater reliability (Supplemental Fig. 1).
On the morning of day 9 (i.e., 0900 hours), mice were administered 33 mg/kg oxycodone and were placed in their respective Plexiglas chambers after 90 minutes for a 30-minute acclimation period. At 120 minutes, the mice were removed from the chambers, weighed, administered naloxone (1 mg/kg s.c.), and returned to the chambers for a 30-minute test session after the chambers were cleaned using a paper towel moistened with water. The mice were weighed again immediately after the test session. Chambers were changed between tests and cleaned with 10% ethanol to mitigate accumulation of residual odor between cohorts of mice.
To reduce the number of mice (Kirk, 2013) used for the acute oxycodone and ZCZ011 dose-response experiments (Figs. 1 and 2), preliminary experiments demonstrated that 75 mg/kg oxycodone and 40 mg/kg ZCZ011 did not exhibit behavioral effects (i.e., the expression of diarrhea, weight loss, jumps, paw flutters, or head shakes) in mice repeatedly administered saline and receiving a naloxone injection on day 9. In the dose-response experiment testing whether acute oxycodone administration will attenuate withdrawal responses, counterbalanced groups of male and female oxycodone-dependent ICR mice received an acute oxycodone injection (17, 33, or 75 mg/kg s.c.) or saline 30 minutes before naloxone administration with doses based on experiments conducted in C57BL/6J male mice (Carper et al., 2021). Male and female ICR mice injected repeatedly with saline received an acute saline injection 30 minutes before naloxone to serve as a negative control. The two experimental factors of repeated oxycodone and acute oxycodone were treated as a single statistical factor and were analyzed as a two-way ANOVA of oxycodone treatment by sex. In the experiment evaluating the dose-response relationship of ZCZ011, counterbalanced groups of male and female oxycodone-dependent ICR mice received ZCZ011 (5, 10, 20, or 40 mg/kg i.p.) or vehicle 75 minutes before naloxone. The pretreatment time and doses of ZCZ011 were based upon previous literature coinciding with its peak antinociceptive effects (Ignatowska-Jankowska et al., 2015a). An additional group of male and female oxycodone-dependent ICR mice was administered oxycodone (75 mg/kg s.c.) 30 minutes before naloxone to serve as a positive control. Male and female ICR mice repeatedly administered saline received vehicle 75 minutes before naloxone administration to serve as a negative control. The two experimental drug factors of oxycodone treatment and ZCZ011 treatment were also collapsed and treated as a single statistical factor.
Effects of acute oxycodone (17–75 mg/kg s.c.) on naloxone-precipitated withdrawal signs in oxycodone-dependent male and female mice. (A) illustrates the timeline of treatments mice received prior to naloxone-precipitated withdrawal. The withdrawal signs measured include the occurrence of diarrhea (B), body weight loss (C), number of jumps (D), number of paw flutters (E), and number of head shakes (F). Data are expressed as percentage scores for (B) and individual data points with mean ± S.D. for (C–F). Numbers at the bottom of bar graphs in (B) indicate the number of mice that presented with diarrhea. Fisher exact test was used to analyze the occurrence of diarrhea (B). Between-measures two-way ANOVA and Dunnett’s post hoc test were used to analyze all other withdrawal signs (C, D, E, and F). "Sal" indicates mice were administered saline and "Oxy" indicates mice were administered oxycodone. **, p < 0.01; ***, p < 0.001; ****, p < 0.0001 vs. oxycodone-saline mice. n = 12 mice/group (n = 6 male and n = 6 female mice/group).
Effects of ZCZ011 (5–40 mg/kg i.p.) on naloxone-precipitated withdrawal signs in oxycodone-dependent male and female mice. (A) illustrates the timeline of treatments mice received prior to naloxone-precipitated withdrawal. The withdrawal signs measured include the occurrence of diarrhea (B), body weight loss (C), number of jumps (D), number of paw flutters (E), and number of head shakes (F). Data are expressed as percentage scores for (B) and individual data points with mean ± S.D. for (C–F). Numbers at the bottom of bar graphs in (B) indicate the number of mice that presented with diarrhea. Fisher exact test was used to analyze the occurrence of diarrhea (B). Between-measures two-way ANOVA and Dunnett’s post hoc test were used to analyze all other withdrawal signs (C, D, E, and F). "Sal" indicates mice were administered saline, "Oxy" indicates mice were administered oxycodone, and "Veh" indicates mice were administered vehicle. *, p < 0.05; ***, p < 0.001; ****, p < 0.0001 vs. oxycodone-vehicle mice. n = 15–16 mice/group (n = 8 male and n = 7–8 female mice/group due to lost video footage).
We employed pharmacological and genetic approaches to determine the involvement of the cannabinoid receptor and the antiwithdrawal effects of ZCZ011. Employing pharmacological and genetic approaches mitigates associated pitfalls when investigating cannabinoid receptor involvement for CB1 receptor PAMs in vivo; ZCZ011 decreases equilibrium binding of rimonabant, the CB1 receptor inverse agonist/antagonist, in vitro (Ignatowska-Jankowska et al., 2015a), and constitutive CB1 (−/−) mice exhibit a lower frequency of opioid withdrawal signs in vivo (Ledent et al., 1999; Lichtman et al., 2001), which provides potential alternative explanations for cannabinoid receptor involvement when examining CB1 receptor PAMs in animal models of opioid dependence. In the rimonabant experiment, counterbalanced groups of male and female oxycodone-dependent ICR mice received two intraperitoneal injections before naloxone administration. The first injection consisted of an intraperitoneal injection of vehicle, rimonabant (3 mg/kg), or SR144528 (3 mg/kg), the CB2 receptor antagonist, and the second injection consisted of vehicle or ZCZ011 (40 mg/kg), with the respective injections given 85 and 75 minutes before naloxone administration. A dose of 3 mg/kg for rimonabant and SR144528 was chosen based upon previous literature demonstrating reversal of the effects of CB1 and CB2 receptor agonists in models of inflammatory and neuropathic pain (Kinsey et al., 2009, 2011; Ignatowska-Jankowska et al., 2015b) as well as cannabinoid and opioid withdrawal (Long et al., 2009a; Schlosburg et al., 2009; Ramesh et al., 2011, 2013). In the experiment using CB1 (−/−) mice, we used age-matched male and female oxycodone-dependent CB1 (+/+) and CB1 (−/−) mice that were administered vehicle or ZCZ011 (40 mg/kg i.p.) 75 minutes before undergoing naloxone-precipitated withdrawal. Each figure contains schematics outlining the timeline of treatments for each experiment.
Somatic signs of withdrawal were measured as previously described (Ramesh et al., 2011), which included the number of jumps, front paw flutters (including single and double flutters), head shakes, the occurrence of diarrhea, and the amount of body weight loss. Jumps were recorded as every incident when the mouse jumped from its hind legs or all four legs. Paw flutters were recorded as single or double flutters separated by at least 1 second or interrupted by any other behavior, such as jumps, head shakes, or grooming. Head shakes were recorded as every incident when the mouse quickly rotated the head clockwise and counterclockwise. The occurrence of diarrhea was assessed by the presence of increased fluid content in the fecal pellets, discoloration of the pellet (e.g., dark brown to light brown), or fragmentation of fecal pellets. Mice were also weighed before and immediately after the 30-minute test session to assess body weight loss (%).
Statistical Analyses
A power analysis from two preliminary experiments indicated that a sample size of 12 and 16 mice per group was required to detect significant effects of acute oxycodone and ZCZ011, respectively. Data for the percentage of mice present with diarrhea are reported as mean bar graphs, whereas data pertaining to other withdrawal signs are individual data points with the mean ± S.D. The occurrence of diarrhea during the 30-minute test session was scored as a quantal measure. Body weight loss (%) was calculated using the following formula:
In which postT represents the body weight of mice after the 30-minute test session, and preT represents body weight before the 30-minute test session. The Fisher exact test was used to analyze occurrence of diarrhea. Data were analyzed using two- and three-way between-measures ANOVA and Student’s t test for planned comparisons. Dunnett’s post hoc test was used to compare drug treatments with oxycodone-dependent saline-treated or vehicle-treated mice, and Tukey’s post hoc test was used to compare between various treatments. Three-way ANOVAs were used to verify whether there were relevant sex differences for the latter studies investigating the antiwithdrawal mechanism of ZCZ011 (Figs. 3 and 4). If no significant three-way interaction is observed and a two-way interaction occurs for the respective factors when excluding sex, subsequent two-way ANOVAs were conducted by pooling data across sex to discern cannabinoid receptor involvement for the effects of ZCZ011 (Figs. 3 and 4) (Kirk, 2013). To determine inter-rater reliability, a Pearson correlation was conducted between the two observers for each measure scored in the 16 videos of oxycodone-dependent mice undergoing precipitated withdrawal. The criterion of significance for all statistical tests was p < 0.05. Experimental protocols were designed to test working hypotheses with planned statistical analyses.
Effects of cannabinoid receptor antagonism (rimonabant, SR1, 3 mg/kg i.p.; SR144528, SR2, 3mg/kg i.p.) on ZCZ011 (40 mg/kg i.p.) attenuation of naloxone-precipitated withdrawal signs in oxycodone-dependent male and female mice. (A) illustrates the timeline of treatments mice received prior to naloxone-precipitated withdrawal. The withdrawal signs measured include the occurrence of diarrhea (B), body weight loss (C), number of jumps (D), number of paw flutters (E), and number of head shakes (F). Data are expressed as percentage scores for (B) and individual data points with mean ± S.D. for (C–F). Numbers at the bottom of bar graphs in (B) indicate the number of mice that presented with diarrhea. Fisher exact test was used to analyze the occurrence of diarrhea (B). Between-measures three-way ANOVA and Tukey’s post hoc test were used to analyze all other withdrawal signs (C, D, E, and F). A preplanned comparison Student’s unpaired t test was used to analyze whether ZCZ011 reliably attenuated paw flutters (E). *, p < 0.05; ***, p < 0.001; ****, p < 0.0001 vehicle-ZCZ011 or SR144528-ZCZ011 mice vs. vehicle-vehicle or SR144528-vehicle mice, respectively. ##, p < 0.01; ####, p < 0.0001 vehicle-ZCZ011 mice vs. rimonabant-ZCZ011 mice. n = 15–16 mice/group (n = 7–8 male and n = 8 female mice/group due to incorrect injection).
Effects of ZCZ011 (40 mg/kg i.p.) on oxycodone-dependent male and female CB1 (+/+) and CB1 (−/−) mice undergoing naloxone-precipitated withdrawal. (A) illustrates the timeline of treatments mice received prior to naloxone-precipitated withdrawal. The withdrawal signs measured include the occurrence of diarrhea (B), body weight loss (C), number of jumps (D), number of paw flutters (E), and number of head shakes (F). Data are expressed as percentage scores for (B) and individual data points with mean ± S.D. for (C–F). Numbers at the bottom of bar graphs in (B) indicate the number of mice that presented with diarrhea. Between-measures three-way ANOVA and Tukey’s post hoc test were used to analyze all other withdrawal signs (C, D, E, and F). *, p < 0.05; **, p < 0.01; ***, p < 0.001 CB1 (+/+)-vehicle mice vs. CB1 (+/+)-ZCZ011, CB1 (−/−)-vehicle, or CB1 (−/−)-ZCZ011 mice. ##, p < 0.01; ####, p < 0.0001 CB1 (+/+)-ZCZ011 mice vs. CB1 (−/−)-ZCZ011 mice. The $$$ indicates a main effect of genotype for jumping behavior between CB1 (+/+) mice vs. CB1 (−/−) mice. n = 15–16 mice/group (n = 8 male and n = 7–8 female mice/group due to one death during repeated oxycodone administration).
Results
Supplemental Fig. 1 compares data between a second trained observer, who scored a subset of 16 videos of oxycodone-dependent mice, and the primary observer to ensure inter-rater reliability. Pearson correlations demonstrated a high correlation between observers for naloxone-precipitated jumps (r = 0.9995; P < 0.0001; Supplemental Fig. 1A), paw flutters (r = 0.9953; P < 0.0001; Supplemental Fig. 1B), and head shakes (r = 0.9180; P < 0.0001; Supplemental Fig. 1C).
Oxycodone Dose-Dependently Attenuated Naloxone-Precipitated Withdrawal Signs in Oxycodone-Dependent ICR Mice
In the first experiment, we tested whether a positive control, acute oxycodone (17, 33, and 75 mg/kg s.c.), attenuated naloxone-precipitated diarrhea, weight loss, jumps, paw flutters, or head shakes. Table 1 summarizes Fisher exact test results for naloxone-precipitated diarrhea and two-way ANOVAs (sex × oxycodone treatment) for each of the other withdrawal signs. A schematic of treatment conditions for this experiment is shown in Fig. 1A. As illustrated in Fig. 1B, 33 mg/kg oxycodone (3 of 12 mice) and 75 mg/kg oxycodone (0 of 12 mice) elicited a lower incidence of naloxone-precipitated diarrhea than oxycodone-dependent saline-treated (12 of 12) mice. Moreover, mice repeatedly administered saline (3 of 12) exhibited a lower incidence of naloxone-precipitated diarrhea than oxycodone-dependent saline-treated (12 of 12) mice. Furthermore, 75 mg/kg oxycodone completely blocked naloxone-precipitated body weight loss, whereas 33 mg/kg oxycodone blocked half the amount of body weight loss compared with oxycodone-dependent saline-treated mice (Fig. 1C). Notably, although micturition was not quantified, weight loss often occurred in mice whose withdrawal chamber was saturated with substantial volumes of urine after the 30-minute test period. In addition, 33 mg/kg and 75 mg/kg oxycodone completely lowered the frequency of naloxone-precipitated jumps compared with oxycodone-dependent saline-treated mice. Furthermore, 75 mg/kg oxycodone completely lowered the frequency of naloxone-precipitated paw flutters, whereas 33 mg/kg oxycodone lowered the frequency by approximately half compared with oxycodone-dependent saline-treated mice (Fig. 1E). Lastly, 33 mg/kg and 75 mg/kg oxycodone lowered the frequency of naloxone-precipitated head shakes compared with oxycodone-dependent saline-treated mice. No endpoint showed a significant sex by oxycodone treatment interaction or main effect of sex (Fig. 1F). Because there were no significant interactions between oxycodone and sex in this experiment or ZCZ011 and sex throughout the study, the variable of sex is collapsed in all figures.
Summary of statistical results from Fisher exact tests and two-way ANOVA analyses for Fig. 1.
ZCZ011 Attenuates Naloxone-Precipitated Withdrawal Signs in Oxycodone-Dependent Mice
Figure 2 illustrates the effectiveness of ZCZ011 (5, 10, 20, and 40 mg/kg i.p.) in attenuating naloxone-precipitated diarrhea, weight loss, jumps, paw flutters, and head shakes. Table 2 summarizes Fisher exact test results for naloxone-precipitated diarrhea and two-way ANOVAs (sex × drug treatment) for each of the other withdrawal signs. A schematic of the treatment conditions for this experiment in which all mice were challenged with 1 mg/kg naloxone before behavioral observations is shown in Fig. 2A. As illustrated in Fig. 2B, 40 mg/kg ZCZ011 (3 of 16 mice) and 75 mg/kg oxycodone (0 of 16 mice) elicited a lower incidence of naloxone-precipitated diarrhea than oxycodone-dependent vehicle-treated (15 of 16) mice. Furthermore, naloxone did not elicit diarrhea in control mice that received repeated injections of saline (0 of 16 mice). In addition, 40 mg/kg ZCZ011 and 75 mg/kg oxycodone completely blocked naloxone-precipitated body weight loss compared with oxycodone-dependent vehicle-treated mice (Fig. 2C). As illustrated in Fig. 2D, 75 mg/kg oxycodone completely lowered the frequency of naloxone-precipitated jumps compared with oxycodone-dependent vehicle-treated mice, whereas no dose of ZCZ011 affected jumping behavior. Although a sex by ZCZ011 treatment interaction failed to achieve significance, a main effect of sex was observed with male mice averaging 61 jumps and female mice averaging 29 jumps. As shown in Fig. 2E, 20 mg/kg and 40 mg/kg ZCZ011 lowered the frequency of naloxone-precipitated paw flutters by about half, whereas 75 mg/kg oxycodone completely lowered the frequency of paw flutters compared with oxycodone-dependent vehicle-treated mice. Lastly, 20 mg/kg ZCZ011, 40 mg/kg ZCZ011, and 75 mg/kg oxycodone lowered the frequency of naloxone-precipitated head shakes compared with oxycodone-dependent vehicle-treated mice (Fig. 2F).
Summary of statistical results from Fisher exact tests and two-way ANOVA analyses for Fig. 2.
ZCZ011 Attenuates Naloxone-Precipitated Diarrhea and Weight Loss through the Activation of CB1 Receptors
These experiments investigated whether CB1 or CB2 receptors mediate the effects of ZCZ011 on withdrawal-induced diarrhea, weight loss, and paw flutters (see schematic of the treatment conditions in Figs. 3A and 4A). Subjects were administered rimonabant, the CB1 receptor inverse agonist/antagonist; SR144528, the CB2 receptor antagonist; or vehicle prior to either vehicle or 40 mg/kg ZCZ011 and were then evaluated for naloxone-precipitated diarrhea, weight loss, paw flutters, and head shakes. Table 3 summarizes Fisher exact test results for naloxone-precipitated diarrhea and three-way ANOVAs (sex × CB antagonist × ZCZ011 treatment) for each of the other withdrawal signs. As shown in Fig. 3B, oxycodone-dependent vehicle-ZCZ011–treated (4 of 16) mice and SR144528-ZCZ011–treated (3 of 15) mice exhibited a lower incidence of naloxone-precipitated diarrhea than oxycodone-dependent vehicle-vehicle–treated (15 of 15) mice and oxycodone-dependent SR144528-vehicle–treated (15 of 16) mice, respectively. Moreover, rimonabant but not SR144528 blocked the protective effects of ZCZ011 on naloxone-precipitated body weight loss, suggesting a CB1-mediated effect. Again, ZCZ011 did not significantly affect naloxone-precipitated jumps (Fig. 3D); however, a main effect of sex revealed male mice averaged 84 jumps, whereas female mice averaged 46 jumps. Significant main effects were found for ZCZ011 treatment and sex for naloxone-precipitated paw-fluttering behavior (Fig. 3E). As seen in the figure, mice in the vehicle-ZCZ011 group averaged 27 paw flutters, whereas mice in the vehicle-vehicle group averaged 46 paw flutters. Moreover, male mice averaged 34 paw flutters, whereas female mice averaged 44 paw flutters. However, no significant interactions of cannabinoid antagonists and ZCZ011 treatment were found, suggesting lack of CB1 or CB2 receptor involvement. Finally, neither ZCZ011 nor the cannabinoid receptor antagonists produced significant effects on naloxone-precipitated head shakes (Fig. 3F). On the contrary, a significant main effect of sex was found in which male mice averaged 5 head shakes, whereas female mice averaged 3 head shakes.
Summary of statistical results from Fisher exact tests and three-way ANOVA analyses for Fig. 3.
Figure 4 shows the effects of ZCZ011 (40 mg/kg i.p.) on naloxone-precipitated diarrhea, weight loss, paw flutters, and head shakes in CB1 (+/+) and CB1 (−/−) mice (see Fig. 4A for the schematic of the experimental design). Table 4 summarizes Fisher exact test results for naloxone-precipitated diarrhea and three-way ANOVAs (sex × genotype × ZCZ011 treatment) for each of the other withdrawal signs. As shown in Fig. 4B, CB1 deletion annihilated the effects of ZCZ011 in reducing the incidence of naloxone-precipitated diarrhea in oxycodone-dependent mice. Specifically, 0 of 16 CB1 (+/+)-ZCZ011–treated oxycodone-dependent mice elicited a lower incidence of naloxone-precipitated diarrhea than 10 of 16 CB1 (+/+)-vehicle–treated oxycodone-dependent mice and 13 of 15 CB1 (−/−)-ZCZ011–treated oxycodone-dependent mice. A significant two-way interaction between genotype and ZCZ011 treatment on naloxone-precipitated body weight loss indicates that CB1 receptors mediate the protective effects of ZCZ011 for this measure (Fig. 4C). Specifically, CB1 (+/+)-ZCZ011–treated mice elicited less body weight loss than CB1 (−/−)-ZCZ011–treated mice. ZCZ011 did not affect naloxone-precipitated jumping behavior, although a significant main effect for genotype revealed that CB1 (+/+) mice jumped significantly more than CB1 (−/−) mice, which respectively averaged 69 and 37 jumps (Fig. 4D). As shown in Fig. 4E, a significant 2-way interaction between genotype and ZCZ011 treatment of naloxone-precipitated paw fluttering behavior revealed CB1 (+/+)-ZCZ011–treated mice elicited little to no paw flutters compared with CB1 (+/+)-vehicle–treated mice. Furthermore, regardless of treatment, CB1 (−/−) mice elicited fewer paw flutters than CB1 (+/+)-vehicle–treated mice. Notably, no difference was observed between CB1 (+/+)-ZCZ011–treated and CB1 (−/−)-ZCZ011–treated mice, signifying a lack of a CB1 receptor–mediated effects. Lastly, Fig. 4F illustrates no three-way interaction, relevant two-way interactions, or main effects for naloxone-precipitated head shakes.
Summary of statistical results from Fisher exact tests and three-way ANOVA analyses for Fig. 4.
Discussion
Here we report that the CB1 receptor PAM ZCZ011 attenuates a subset of naloxone-precipitated withdrawal signs in oxycodone-dependent mice. Notably, ZCZ011 fully blocked withdrawal-induced diarrhea and weight loss and attenuated paw flutters. In contrast, ZCZ011 did not impact naloxone-precipitated jumping in oxycodone-dependent mice, and its attenuation of naloxone-precipitated head shakes was inconsistent. The antidiarrheal and anti–weight loss effects of ZCZ011 occurred predominantly through a CB1 receptor–dependent mechanism in both male and female mice. Thus, a CB1 receptor PAM reduces a subset of withdrawal signs in both male and female opioid-dependent mice.
Complementary pharmacological and genetic approaches used to elucidate cannabinoid receptor mediation of the antiwithdrawal effects of ZCZ011 demonstrate that ZCZ011 attenuates diarrhea and body weight loss through CB1 not CB2 receptors. However, it is difficult to draw conclusions about CB1 receptor involvement in the paw flutters measure because rimonabant did not significantly block the reduction of paw flutters produced by ZCZ011, and since CB1 (−/−) mice exhibited an overall lower frequency of paw flutters than CB1 (+/+) mice regardless of drug treatment, determining CB1 receptor involvement was hindered. The unreliability of ZCZ011 to reduce naloxone-precipitated head shakes in oxycodone-dependent mice precluded discerning the role of CB1 receptors in this measure. Consistent with previous reports (Ledent et al., 1999; Lichtman et al., 2001; Maccarrone et al., 2002), CB1 (−/−) mice showed fewer jumps and paw flutters compared with (+/+) mice. Nonetheless, both genetic and pharmacological data reveal that ZCZ011 attenuates naloxone-precipitated diarrhea and weight loss in oxycodone-dependent mice through CB1 receptors.
CB1 receptor orthosteric agonists and endocannabinoid catabolic enzyme inhibitors effectively attenuate opioid withdrawal; however, they also produce unwanted cannabimimetic effects and antinociceptive tolerance and dependence upon repeated administration, which limits their potential clinical utility. Moreover, although clonidine has shown to reduce the frequency of withdrawal-induced jumps, paw flutters, and weight loss (Carper et al., 2021), its hypotensive effects limit its clinical use (Kuhlman et al., 1998; Stolbach and Hoffman, 2020). In contrast to combined FAAH/MAGL blockade, which substitutes for THC in the drug discrimination paradigm (Long et al., 2009b), ZCZ011 does not substitute for CB1 receptor orthosteric agonists (Ignatowska-Jankowska et al., 2015a), indicating that CB1 receptor PAMs do not elicit cannabimimetic subjective effects. Thus, CB1 receptor PAMs may offer a favorable strategy to target CB1 receptors for therapeutic gain with a reduced side effect profile compared with orthosteric CB1 receptor agonists or combined FAAH/MAGL blockade. Repeated administration of THC or MAGL inhibitors leads to antinociceptive tolerance, whereas repeated administration of ZCZ011 or GAT211 produces sustained antinociceptive effects in mouse models of inflammatory and neuropathic pain (Ignatowska-Jankowska et al., 2015a; Slivicki et al., 2017). Finally, repeated administration of THC or MAGL inhibitors leads to cannabinoid dependence (Schlosburg et al., 2014; Trexler et al., 2019), which is not seen with repeated administration of GAT211 (Slivicki et al., 2017). It is also noteworthy that ZCZ011 alone does not produce a conditioned place preference (Ignatowska-Jankowska et al., 2015a; Trexler et al., 2019), suggesting that it lacks rewarding effects. The present study showing that ZCZ011 attenuates oxycodone withdrawal signs and other work showing that it attenuates rimonabant-precipitated and spontaneous withdrawal signs in THC-dependent mice (Trexler et al., 2019) without cannabimimetic side effects or tolerance supports the possibility that CB1 receptor PAMs may offer a novel strategy to treat opioid and cannabinoid dependence.
The present study employed two strains [i.e., ICR and CB1 (+/+) and (−/−) mice on a C57BL/6J background] of male and female mice and showed that ZCZ011 reliably diminished naloxone-precipitated diarrhea and weight loss in oxycodone-dependent mice regardless of strain or sex. Although female Wistar rats self-administer more oxycodone than males (Kimbrough et al., 2020), we did not observe withdrawal sex differences with respect to drug treatment. However, independent of sex, we found strain differences in which the variability in jumps and paw flutters was high in ICR mice compared with CB1 (+/+) mice or the C57BL/6J mice used in the study done by Carper and colleagues (2021). Additionally, CB1 (+/+) and (−/−) mice on a C57BL/6J background show less weight loss than the C57BL/6J mice used in the Carper and colleagues (2021) study. This solitary difference in the magnitude of weight loss is likely due to procedural differences between the studies. For example, the mice in the present study were exposed to the withdrawal chambers for an additional 30-minute acclimation period before naloxone challenge, whereas the mice in the study by Carper et al. (2021) were placed in the chamber immediately after naloxone administration.
In contrast to the present study in which ZCZ011 modestly lowered the frequency of paw flutters, Slivicki and colleagues (2020) found that GAT211 did not attenuate naloxone-precipitated paw flutters in morphine-dependent mice. They also found that GAT211 did not reduce naloxone-precipitated jumps; however, they did not report other measures of withdrawal. Thus, it will be of value to evaluate multiple CB1 receptor PAMs on a full complement of withdrawal signs in several rodent models of opioid dependence.
ZCZ011 and related compounds (i.e., GAT211) show mixed allosteric agonistic and PAM properties, which have been termed CB1 ago-PAMs (Kenakin, 2013). These compounds enhance the effects of CB1 receptor orthosteric agonists (i.e., CP55,940, WIN55,212-2, or N-arachidonylethanolamine) in a variety of functional assays, including [35S] GTPγS-binding, β-arrestin recruitment, and inhibition of cAMP production, but also activate β-arrestin recruitment and inhibit cAMP production in the absence of CB1 receptor orthosteric agonists (Ignatowska-Jankowska et al., 2015a; Slivicki et al., 2017; Saleh et al., 2018; Tseng et al., 2019; Garai et al., 2021). Thus, additional studies (e.g., site directed mutagenesis) will be required to address the receptor mechanism(s) by which ZCZ011 ameliorates withdrawal signs in opioid-dependent mice.
The differential effectiveness of ZCZ011 in ameliorating diarrhea and weight loss compared with somatic withdrawal signs merits consideration of CNS- and peripherally mediated processes. Somatic withdrawal signs in opioid-dependent mice have been mapped to various brain regions (Koob et al., 1992; Maldonado et al., 1992, 1996), including the locus coeruleus (Maldonado and Koob, 1993), the periaqueductal gray (Laschka et al., 1976), and the medial habenula (Boulos et al., 2020). It would be of value to examine whether CB1 receptors expressed in the locus coeruleus, periaqueductal gray, and medial habenula play a role in the attenuation of opioid somatic withdrawal signs by drugs targeting the endocannabinoid system, such as CB1 receptor PAMs (Kenakin, 2013). Withdrawal-induced weight loss results from a combination of diarrhea and increased micturition. Thus, activation of CB1 receptors, which are expressed in the kidneys (Shire et al., 1995; Silva et al., 2013; Lin et al., 2014) and throughout the enteric nervous system (Massa et al., 2005), particularly on neurons in the myenteric and submucosal plexuses (Storr et al., 2010; Trautmann and Sharkey, 2015; Hasenoehrl et al., 2016), may prevent withdrawal-induced weight loss. Notably, cannabinoid receptor agonists evoke diuresis (Sofia et al., 1977). Interestingly, CB1 receptor agonists produce biphasic effects on diuresis in which low doses increase urine output, whereas high doses decrease urine output (Chopda et al., 2013). Lastly, naloxone-precipitated diarrhea is heavily associated with the ileum of the small intestine (Maguma et al., 2010), where the coordination of motility and secretion occur (Ramesh et al., 2011, 2013; Smith et al., 2012), and CB1 receptor activation leads to decreased motility and secretion (Storr et al., 2010; Ramesh et al., 2011, 2013; Hasenoehrl et al., 2016). Accordingly, ZCZ011 may prevent withdrawal-induced diarrhea and weight loss through CB1 receptor activation in peripheral organs mediating gastrointestinal motility and secretion as well as urine output.
Current medications (i.e., methadone and buprenorphine) employed to treat opioid withdrawal symptoms (Bell and Strang, 2020) possess abuse liability (Cicero and Inciardi, 2005) and do not effectively ameliorate all withdrawal symptoms. Alternative adjunct medications offer benefit, such as clonidine for anxiety and loperamide for diarrhea and stomach cramps (Stolbach and Hoffman, 2020), yet also exhibit hypotension and abuse liability, respectively (Miller et al., 2017; Wu and Juurlink, 2017; Toce et al., 2018). As such, a need exists to identify alternative medications to treat the most severe withdrawal symptoms, such as diarrhea (Stolbach and Hoffman, 2020). The finding that ZCZ011 fully attenuates naloxone-precipitated diarrhea to a similar extent as 75 mg/kg oxycodone (Fig. 2), suggests that a CB1 receptor PAM may serve as a potential alternative to substitution therapy.
In conclusion, the CB1 receptor PAM ZCZ011 attenuates a subset of naloxone-precipitated withdrawal signs in oxycodone-dependent male and female mice through a CB1 receptor-dependent mechanism. As diarrhea is a severe withdrawal symptom in opioid-dependent humans and current medications possess side-effect profiles, future directions will explore the antidiarrheal mechanisms of ZCZ011. Additionally, it would be of interest to investigate ZCZ011 in other preclinical models of bowel dysfunction. Overall, the results of the present study suggest that CB1 receptor PAMs may represent an alternative strategy to treat selective and severe opioid withdrawal effects.
Acknowledgments
The authors thank Edward Dimen for the construction of the withdrawal chambers.
Authorship Contributions
Participated in research design: Dodu, Damaj, Schlosburg, Akbarali, Lichtman.
Conducted experiments: Dodu, Moncayo.
Contributed reagents or analytic tools: Schlosburg, Wu, and Lu.
Performed data analysis: Dodu, Moncayo, O’Brien, Lichtman.
Wrote or contributed to the writing of the manuscript: Dodu, Moncayo, Damaj, Schlosburg, Akbarali, O’Brien, Kendall, Wu, Lu, Lichtman.
Footnotes
- Received April 27, 2021.
- Accepted October 1, 2021.
This research was supported by National Institutes of Health National Institutes on Drug Abuse [Grants T32DA007027, R01DA039942, R00DA037344, and P30DA033934] and start-up funds from Virginia Commonwealth University (VCU) School of Medicine and VCU School of Pharmacy.
No author has an actual or perceived conflict of interest with the contents of this article.
↵
This article has supplemental material available at jpet.aspetjournals.org.
Abbreviations
- CB
- cannabinoid
- CB1
- cannabinoid receptor type 1
- CB2
- cannabinoid receptor type 2
- FAAH
- fatty acid amide hydrolase
- GAT211
- 3-(2-nitro-1-phenylethyl)-2-phenyl-1H-indole
- MAGL
- monoacylglycerol lipase
- SR144528
- N-[(1S)-endo-1,3,3,-trimethylbicyclo[2.2.1]heptan-2-yl]-5-(4-chloro-3-methylphenyl)-1-(4-methylbenzyl)-pyrazole-3-carboxamide
- THC
- Δ9 -tetrahydrocannabinol
- ZCZ011
- 6-methyl-3-(2-nitro-1-(thiophen-2-yl)ethyl)-2-phenyl-1H-indole
- Copyright © 2021 by The American Society for Pharmacology and Experimental Therapeutics
