Motor-like Tics are Mediated by CB2 Cannabinoid Receptor-dependent and Independent Mechanisms Associated with Age and Sex

Δ9-Tetrahydrocannabinol (Δ9-THC) inhibits tics in individuals with Tourette syndrome (TS). Δ9-THC has similar affinities for CB1/CB2 cannabinoid receptors. However, the effect of HU-308, a selective CB2 receptor agonist, on repetitive behaviors has not been investigated. The effects of 2,5-dimethoxy-4-iodoamphetamine (DOI)-induced motor-like tics and Δ9-THC were studied with gene analysis. The effects of HU-308 on head twitch response (HTR), ear scratch response (ESR), and grooming behavior were compared between wildtype and CB2 receptor knockout (CB2−/−) mice, and in the presence/absence of DOI or SR141716A, a CB1 receptor antagonist/inverse agonist. The frequency of DOI-induced repetitive behaviors was higher in CB2−/− than in wildtype mice. HU-308 increased DOI-induced ESR and grooming behavior in adult CB2−/− mice. In juveniles, HU-308 inhibited HTR and ESR in the presence of DOI and SR141716A. HU-308 and beta-caryophyllene significantly increased HTR. In the left prefrontal cortex, DOI increased transcript expression of the CB2 receptor and GPR55, but reduced fatty acid amide hydrolase (FAAH) and α/β-hydrolase domain-containing 6 (ABHD6) expression levels. CB2 receptors are required to reduce 5-HT2A/2C-induced tics in adults. HU-308 has an off-target effect which increases 5-HT2A/2C-induced motor-like tics in adult female mice. The increased HTR in juveniles induced by selective CB2 receptor agonists suggests that stimulation of the CB2 receptor may generate motor tics in children. Sex differences suggest that the CB2 receptor may contribute to the prevalence of TS in boys. The 5-HT2A/2C-induced reduction in endocannabinoid catabolic enzyme expression level may explain the increased endocannabinoids’ levels in patients with TS. Supplementary Information The online version contains supplementary material available at 10.1007/s12035-022-02884-6.

repetitive behaviors in C57BL/6J mice were tested. Mice were randomly injected (i.p.) with SR141716A (5, 10, 20 mg/kg), while control mice were injected with the respective vehicle mixture. Each mouse was immediately placed in the middle of a clear glass cage 30 x 30 x 30 cm. Five minutes after placing a mouse into the cage center, counting started for 39 min in three minute intervals.
To test the effects of HU-308 on repetitive behaviors, each set of experiments included appropriate controls with varying concentrations of HU-308. Each mouse was injected (i.p.) with HU-308 or vehicle (as described in each Fig. legend). After 60 min, a second injection (i.p.) of SR141716A (10 mg/kg) or its respective vehicle (made with ethanol or DMSO as indicated below) was administered. Compared with ICR mice [2], SR141716A affected C57BL/6J mice at a slower rate. In our study, the effect of SR141716A on repetitive behaviors in C57BL/6J mice was more pronounced about 20 min after its injection. Therefore, 20 min after placing a mouse into the cage center, counting was started for an additional 24 min in three minute intervals (total time was 44 min).
The HTR/ESR/grooming behaviors were manually counted by an observer. In all experiments, the number of HTR, ESR and grooming behaviors were counted in the same mouse. In both model systems, HTR was counted every time the mouse had a head twitch, as described previously [1,3]. Shakes and other voluntary head movements were not counted.
ESR was counted each time the mouse scratched itself with its hind limbs, similar to methods previously described [3]. Depending on the frequency of action, there are different methods of counting ESR. In the DOI model system, a new ESR action was added to the total counts only if the mouse moved with all four paws since the previous action. Compared with DOI, SR141716A induced a higher ESR frequency. Therefore, in the SR141716A model system, a new ESR action was added to the total counts on every ESR action, similar to that described previously [2].
Self-grooming was counted each time the mouse groomed any body part with its forelimbs or hind limbs and/or licked and cleaned its tail or nails. However, in C57BL/6J mice, DOI induced rapid and brief grooming actions, while SR141716A induced mainly lengthy and long grooming actions. Therefore, in the DOI model system, a new grooming action was added to the total counts only if the mouse moved with all four paws since the previous action. In contrast, in the SR141716A model system, a new grooming action was added to the total counts when a mouse changed to groom another body part since the previous action. All areas around the head were considered as one part, such that if a mouse started to groom, for example, the ears, and then moved to groom the cheeks, this was counted as one action.
We introduced steps to minimize bias when automatic/full blinding procedures could not be applied. We called these steps 'a semi-blinding protocol' which also means the experimental person was semi-blind to the study: (1) Each mouse was randomly injected (i.p.) with the tested drug or with vehicle or with saline before the experiment started; (2) Doses were randomized in each set of experiments (according to Fig. legends); (3) Mice were tested in a random order, which reflects a random order of tested doses; (4) After injecting all the mice in the injection room and leaving them in their cage, the person tested the mice according to the order marked on their tails in the experimental room; (5) HU-308 was tested in different models in the same week; (6) Analysis was performed at the end of the entire experiment.

Open field test
The test was performed similarly to the methods previously described [4]. Mice (8) were habituated in their home cage to the experimental environment for 60 min. After injection of HU-308, each mouse was restored back to its home cage. After 60 min, the mouse was injected a second time with vehicle. Immediately after the second injection the mouse was placed at the center of the experimental cage made of a clear glass 30 X 30 X 30 cm and divided into 4 X 4 identical squares by black marks. The ambulation/rearing/grooming behaviors were manually counted by an observer. Ambulation was counted each time the mouse crossed a square line with its four paws. Rearing was counted each time the mouse reared on its hindlimbs and stretched upward with its front paws (on walls or in air) but not if it groomed itself (no stretching motion). Grooming was counted each time the mouse groomed itself with its front paws at face, tail, nails or other body part. Total grooming was counted for all 20 minutes with no intervals. Ambulation and rearing were counted for 20 min in two minute intervals.

Marble burying test
The marble burying test (MBT) has been suggested as a model for screening potential drugs for anxiety and obsessive-compulsive behavior and is sensitive to Δ 9 -THC and high doses of CBD [5,6]. The test was conducted similarly to methods previously published [5]. Mice (6-8 mice) were habituated to the experimental environment for 60 min. After injection of HU-308, each mouse was transferred to a transition cage (1). After 60 min, each mouse was exposed to the experimental cage, a white polycarbonate cage 33 X 27 X 16 cm covered with 5 cm of sawdust (clean + dirty from home cage), for 5 min, and then translocated for 5 min into a transition cage (2) while the marbles were arranged in the experiment cage. Twenty opaque black marbles (14 mm diameter) were evenly spaced in the cage at 4 X 5, at 1 cm from the cage walls, on the top of the sawdust. Using video camera and EthoVision XT 11.5, the cage was photographed before and after the experiment. After this procedure, each mouse was placed at the top left corner and recorded for 30 min.
The MBT was electronically recorded by EthoVision XT 11.5 software (Noldus Information Technology, The Netherlands), using build-in parameters, i.e. no special codes.
Using this software, the cage was divided into 9 (3 x 3) equally sized areas, the middle area was considered as the center of the cage for all the parameters of the experiment. The number of buried and moved marbles was manually analyzed from the pair of photos (before and after the experiment). A marble was considered buried if two-thirds of it was covered with sawdust.
The total distance, duration in the center of the cage, frequency and latency to center, which are related to changes in locomotor activity, were electronically quantified by EthoVision XT 11.5 software which produced raw data outputs.

Reverse transcription and RT-PCR
In juvenile mice, the effects of DOI or 5 mg/kg ∆ 9 -THC on genes of the endocannabinoid system were tested. Two days after each experiment [1], mouse brains were removed, washed and dissected into specific regions in ice-cold HEPES buffer [7]. Tissues were stored in RNAlater reagent (Ambion, U.S.) overnight at 4°C and transferred to −80°C for indefinite storage. Total RNA was isolated from the prefrontal cortex, which is enriched with 5-HT2A receptors, using Tri Reagent (Sigma-Aldrich, Israel). Residual genomic DNA was removed from RNA samples using a TURBO DNA-free Kit according to the manufacturer instruction The cycling parameters were: 10 min incubation phase at 95°C, 40 cycles at 95°C for 15 seconds, annealing at 60°C for 1 min and dissociation at 72°C for 6 seconds with a curve to examine non-specific amplification. Efficiency and specificity of the primers were examined by analysis of cDNA and primer dilution curves and by no-template controls. Gene-specific sequence oligonucleotide primers for: glyceraldehyde-3-phosphate dehydrogenase (GAPDH), monoacylglycerol lipase (MAGL), fatty acid amide hydrolase (FAAH), α/β-hydrolase domain containing 6 (ABHD6), CB1 receptor, CB2 receptor, GPR55, were designed by Primer3, verified by Blast (NCBI) and purchased from Sigma-Aldrich (Israel). The sequences of primers used in this study are provided in Table 1. A total of five experimental sets were examined for which the behavioral results were previously published [1]. Each set of experiments (i.e. experimental groups) was tested on the same plate. Each sample (i.e. animal) was tested in duplicate. Averaged gene expression level was normalized to GAPDH as its mRNA expression level was highly stable between different treatments (ΔCt). The variability between the control animals was calculated from the normalized Ct (ΔCt) values of the control group. Delta-delta Ct (ΔΔCt) was calculated relative to the normalized Ct of control group (Vehicle + Saline).
Relative quantitation analysis of gene expression levels was performed using the delta-delta Ct , such that all values were relative to the control group which was set as a delta-delta Ct (2 −ΔΔCt ) of one.

Effects of HU-308 on DOI-induced repetitive behaviors in adult mice-sex differences
Both sexes contributed to the increase in DOI-induced repetitive behaviors in CB2 -/mice (Fig.   S1). The exploratory results show sex differences for the effects of DOI on repetitive behaviors in CB2 -/mice, i.e. DOI induced lower frequency of repetitive behaviors in females (Fig. S1ac) than in males ( Fig. S1d-f). In addition, exploratory results showed no sex differences for ESR: HU-308 increased ESR in CB2 -/- (Fig. S1b,e) and WT (Fig. S1h,k) female and male mice, and no sex difference were apparent for HTR in WT mice (Fig. S1g,j). In contrast, exploratory results showed sex differences for the effects of HU-308 on DOI-induced HTR and grooming behavior in CB2 -/mice (Fig. S1). While HU-308 increased DOI-induced grooming behavior and HTR in females (Fig. S1a,c), it inhibited these behaviors in males (Fig. S1d,f). Body weights were not different between groups (Fig. S2a,b). These results suggest that in adult mice the lack of functional CB2 receptor signaling makes sex differences in the response to DOI and HU-308 more noticeable.
It is noteworthy that this is not a traditional comparison of basal behavior between knockout and their corresponding WT mice. Yet, we noticed two main differences between the two sub-strains: (1) in adult CB2 -/mice, DOI-induced repetitive behaviors HTR, ESR and grooming behavior were, respectively, 2.3, 3.4 and 1.7 times significantly higher compared with WT adult mice ( Fig. 1a-f; P < 0.05, 2-way ANOVA); (2) compared with adult WT mice, the DOI-induced HTR resulted in a vigorous and robust HTR in the CB2 -/mice.

Effects of HU-308 on DOI-induced repetitive behaviors in young adult mice
We expected to find similar results in young adult mice. However, HU-308 (0.2, 1, 5 mg/kg) had no effect on DOI-induced HTR and grooming behavior (Fig. S3d,f) in young adult male mice. While exploratory results showed that HU-308 (0.2, 1 mg/kg) reduced DOI-induced ESR (Fig. S3e). Body weights were not different between groups (Fig. S5e,f).
These observations were further supported with results from the marble burying test (MBT). In young adult male mice, HU-308 (0.2, 1, 5 mg/kg) had no effect on the distance the mice moved in the cage, latency to center or the number of buried marbles (Fig. S4a-c). There was also no difference in the time spent in the center, frequency of entries to center, or in the number of moved marbles (not shown). Similar results were obtained in young adult female mice, except that HU-308 (5 mg/kg) significantly increased the latency to center (Fig. S4d-f).
Body weights were not different between groups (Fig. S2c,d).

Effect of HU-308 on basal repetitive behaviors in juvenile mice-experimental conditions
Similar results were found under different experimental conditions i.e. in which the second injection is with saline which was used as a control in the DOI vs. SR141716A model systems.
Compared with the basal HTR of the control group (0.3 ± 0.2), HU-308 (0.2, 1, 5 mg/kg) significantly increased HTR to 2.0 ± 0.7, 1.3 ± 0.6 and 0.7 ± 0.2, resulting in an increase of 600%, 350% and 150% of basal HTR in juvenile mice, respectively (Fig. S9a). These results suggest that juvenile males are more sensitive to the stimulation of CB2 receptor-induced neck/head motor-like tics than females. It was notable that the lower dose had a higher effect, suggesting that the direction of the dose-response to HU-308 depends on the presence of salts  S9b; P < 0.05). Though HU-308 (5 mg/kg) significantly reduced ESR, by the end period of counting there was no difference (Fig. S9b). HU-308 alone had no effect on basal grooming behavior (Fig. S9c). In males and females, average body weight was not different between groups (Fig. S5a,c).
In line with our results, JWH-133 and GW833972A, both selective CB2 receptor agonists, enhance dopamine release in the presence of K + in slices from the dorsal striatum of autoreceptors (i.e. lack of dopamine), the CB2 receptor will favor the coupling to Gαs protein.

SR141716A model system
The effects of the vehicles on SR141716A-induced repetitive behaviours were compared. SR141716A (5, 10, 20 mg/kg), dissolved in ethanol, dose dependently and significantly increased HTR and ESR behaviours but not grooming behaviour in juvenile male C57BL/6J mice (Fig. S7a-c). These results replicate another study where repetitive behaviours were simultaneously counted for 20 min following SR141716A administration (2.5, 5, 10, 20 mg/kg; ethanol vehicle) to juvenile ICR male mice [2]. However, we noted major differences between the two strains of mice. While 20 mg/kg SR141716A induced about 80 ESR in juvenile male ICR mice after 20 min, the same dose induced only about 12 ESR after 20 min (85% lower), and 23 ESR (70% lower) after 39 min in juvenile male C57BL/6J mice. Similarly, the HTR -11was about 25% lower after 20 min and 50% lower after 39 min in C57BL/6J compared to ICR mice.
Compared with DMSO vehicle, the basal HTR and ESR in ethanol vehicle were higher, while grooming behaviour was lower in the control group of juvenile male C57BL/6J mice ( Fig. S7a-c vs. S7d-f, respectively). SR141716A (10, 20 mg/kg) significantly increased HTR and ESR in the presence of DMSO vehicle (Fig. S7a,b vs. S7d,e, respectively). After 20 min of SR141716A administration, there was no difference in grooming behaviour in ethanol or DMSO (Fig. S7c vs. S7f, respectively). However, a small but significant effect of SR141716A on grooming behaviour was found 39 min following its injection. Grooming behaviour was significantly increased by SR141716A (10 mg/kg) in the presence of ethanol, while SR141716A (10, 20 mg/kg) significantly reduced grooming behaviour in the presence of DMSO (Fig. S7c vs. S7f, respectively). Average body weight was not different between groups ( Fig. S8a-b).
As similar effect of SR141716A on HTR was found with both vehicles, but as ethanol increased basal HTR and ESR, we have used DMSO vehicle in the subsequent experiments at a selected dose of 10 mg/kg SR141716A.

Fig. S1
Sex comparison of the effect of HU-308 (5 mg/kg) on DOI-induced HTR (a, d, g, j), ESR (b, e, h, k) and grooming behaviour (c, f, i, l) in wildtype (WT) and adult CB2 -/knockout mice (CB2 -/mice). In a-c, the effects of HU-308 on CB2 -/females. In d-f, the effects of HU-308 on CB2 -/males. In g-i, the effects of HU-308 on WT females. In j-l, the effects of HU-308 on WT males. Two-way ANOVA analysis of variance followed by Bonferroni's test for multiple comparisons was performed by GraphPad Prism 8. Asterisks aside the graph are p value summary vs. vehicle + DOI group. Asterisks along the curve are p values of multiple comparisons (at a time point) of each dose vs. vehicle + DOI group. * P < 0.05; ** P < 0.01; *** P < 0.001 significantly different.

Fig. S2
Body weight of each adult wildtype (a) and CB2 -/-(b) mouse treated with DOI in the absence or presence of HU-308 pre-treatment. Body weight of each young adult male (c) and female (d) mouse treated with HU-308 before the MBT behaviour test. On the experimental day, each body weight was determined before drug injection. The average of body weights was not significantly different between the experimental groups.

Fig. S3
In young adult mice, HU-308 alone had no effect on repetitive behaviours. Effect of HU-308 (0.2, 1, 5 mg/kg) on basal and DOI-induced HTR (a, d), ESR (b, e) and grooming behaviour (c, f) in young adult male mice. In a-c, the effects of HU-308 alone. In d-f, the effects of HU-308 in the presence of DOI (1 mg/kg). Data represent mean ± SEM. n represent the number of animals in each group. The experiment was independently repeated a number of times according to the lowest n number.

Fig. S4
Effect of HU-308 (0.2, 1, 5 mg/kg) on young adult males and females in the marble burying test. HU-308 had no effect on the distance (a, d) and the number of buried marbles (c, f). HU-308 (5 mg/kg) significantly increased the latency to centre in females (e) but not in males (b).
Data represent mean ± SEM. n represent the number of animals in each group. The experiment was independently repeated 7 times. One-way ANOVA analysis of variance followed by Bonferroni's test for multiple comparisons was performed by GraphPad Prism 8. * P < 0.05 significantly different.

Fig. S5
Body weight of each HU-308 pre-treated mouse in the absence (a, c, e) or presence of DOI (b, d, f). In juvenile males (a, b), juvenile females (c, d), young adult males (e, f). On the experimental day, each body weight was determined before drug injection. The average of body weights was not significantly different between the experimental groups.

Fig. S6
Sex comparison of the effect of HU-308 (0.2 mg/kg) on DOI (1 mg/kg)-induced HTR (a, d), ESR (b, e) and grooming behaviour (c, f) on juvenile male (a-c) and female (d-f) mice. In males, HU-308 (0.2 mg/kg) had no effect, while it significantly inhibited DOI-induced HTR and grooming behaviour in females. Data represent mean ± SEM. n represent the number of animals in each group. The experiment was independently repeated a number of times according to the lowest n number. Two-way ANOVA analysis of variance followed by Bonferroni's test for multiple comparisons was performed by GraphPad Prism 8. *P < 0.05 significantly different as indicated. Asterisks aside the graph are p value summary vs. vehicle + DOI group. * P < 0.05; ** P < 0.01; *** P < 0.001; **** P < 0.0001 significantly different.

Fig. S10
In the open field test, HU-308 alone (1, 5 mg/kg) significantly reduced locomotor activity of ambulation (a), rearing (b) but had no effect on grooming behaviour (c) in juvenile male mice.
The average of body weights was not significantly different between the experimental groups (d). Data represent mean ± SEM. n represent the number of animals in each group. The experiment was independently repeated 5 times. Two-way ANOVA analysis of variance followed by Bonferroni's test for multiple comparisons was performed by GraphPad Prism 8.