The main finding of this study is that acute treatment with cannabidiol (CBD) at a dose of 10 mg/kg increased male BTBR mouse social interaction preference relative to the vehicle control treatment, as assessed by difference in time spent in chambers with stranger mice versus novel objects. However, acute CBD (10/mg/kg) effects were not consistent in other measurements of BTBR social interaction preference, such as sniffing times during the social interaction test (see Figure 2C). A strength of our approach was testing acute CBD (10 mg/kg) effects in males from three mouse strains. Two of the strains (BTBR and SERT knock-out) can be considered models of autism and the third is a wild-type background strain for SERT knock-out mice that typically prefers social interactions (16, 17, 19). Even in C57BL/6J mice, within group preference for strangers versus novel objects was more apparent with the CBD treatment (Figures 3A,B). CBD at 10 mg/kg had no effect on time in chambers, but within-group social interaction preference measured by sniffing in SERT knock-out mice was significant only in the CBD treated group (Figure 4B). Taken together these findings, while subtle, indicate that even at a relatively low 10 mg/kg dose CBD has therapeutic potential to enhance social interaction. This is promising since social interaction deficits are the most treatment-resistant autism symptoms (16). Despite the modest number of mice used in this acute treatment pilot study, findings are consistent with those obtained by using CBD in a mouse model of Dravet syndrome (constitutive Scn1a deficient and knock-out or Scn1a+/−) wherein doses of 10 and 20 mg/kg CBD administered an hour before testing also enhanced social interaction in three chamber preference tests (28). The primary goal of this study was to find the lowest dose of CBD that could enhance social interaction preference, and to determine if it could also enhance social dominance or reduce marble burying.

Prior murine studies at higher CBD doses also show its potential to enhance social behaviors. In a mouse model of traumatic brain injury, CBD treatment ameliorated social interaction preference deficits found in the model (29). Likewise in a genetic mouse model of Dravet syndrome (Scn1a +/−) hallmarked by severe epilepsy with co-morbid motor deficits and autism, anxiety and depression an extended CBD treatment of 100 mg/kg, twice daily for 27 days enhanced survival, and dyadic social interaction relative to vehicle controls (30). However, to the best of our knowledge this is the first report of enhanced social preference in three chamber social interaction and novelty preference tests by acute CBD treatment in BTBR or SERT knock-out mice. The vehicle-treated (10% methanol) C57BL/6J males exhibited less preference for interaction than we have seen in prior studies without alcohol administration (26). This effect may account for the unexpected improvement in their social interaction with 10 mg/kg CBD treatment (Figures 3A,B).

Social novelty preference was not enhanced by this regimen of acute CBD treatments in any strain or line of mice examined in the present study. The social novelty preference tests started 60 min and ended 70 min after CBD injection. This appears to be a peak time for CBD to impact mouse social interaction preference behaviors, consistent with findings 1 hour after treatment in other strains, however CBD effects on social novelty preference was not previously examined in mice (28). However, in rats CBD treatments of 12 and 30 mg/kg impaired social recognition memory in a parallel test (31). Intriguingly, in the present study it also appeared that in C57BL/6J as well as in SERT mice sniffing of novel stranger mice was somewhat dampened. However, since the interaction was not significant and the study may have been underpowered, these findings should be interpreted cautiously.

Marble burying behavior was unaffected by 10 mg/kg CBD in the present study at 70–100 min after administration. Prior reports indicate that a dose of 30–120 mg/kg CBD can suppress marble burying in mice acutely or when administered for several days (32). The 10 mg/kg CBD dose administered in the present study may have been insufficient to reach the threshold required to reduce marble burying. Indeed the 4 mice administered CBD at a dose of 50 mg/kg in 50% methanol vehicle did not bury any marbles or back out of the tube test, but since their locomotor and/or exploratory activity appeared to be inhibited, and this effect could not be teased apart from the higher concentration of methanol required to achieve this dose we did not pursue any higher CBD doses since it fell outside of the scope of this pilot study.

This study has three major limitations, the first is the inability to perform additional experiments due to the discontinuation of the commercial formulation of CBD in methanol used. Our use of 10% methanol as a solvent was necessitated by initial use of the United States Drug Enforcement Administration (USDEA) exempt preparation of CBD, of >98 purity that was available at the time of the study from Cayman Chemical Co. (Catalog #90081, discontinued). This formulation limited the present study to acute treatments because of the methanol solvent, and subsequently it became unavailable. The equivalent CBD reference is now sold only in an acetone solvent, so we were unable to increase our sample size or expand the study. For future studies, which should include tests of these behaviors after chronic CBD administration, a different solvent, such as 1:1:18 mixture of Tween 80, propylene glycol, and physiologic saline (33), or a 1:1:18 mixture of saline, ethanol, and emulphor (14) will provide more flexibility and stability.

The second major limitation of this study is the lack of female mouse subjects. Our justification for excluding females from this pilot study was that ASD is 4 times more common in males than in females (16). However, sex differences in autism incidence and manifestation deserve greater attention and scrutiny in future clinical and preclinical studies. Clinical manifestation of ASD can be more difficult to detect in females due to greater female conformance to gender stereotypes, gender-specific narrow interests that are more socially oriented, and greater symptom camouflaging by girls with ASD (34–36). Indeed, ASD diagnoses are typically made later in females than in males (34), and the true male/female ratio for ASD may be closer to 3:1 (37). In mice male-female comparisons have been made for autism-relevant behaviors, and in many of them the social behavioral phenotypes of both sexes is similar (16, 38). Even in response to CBD treatment, male and female behaviors were similar in the Dravet syndrome mouse model (28). However, it remains to be demonstrated if such effects are more widely generalizable in the BTBR or SERT knockout mouse or in other ASD models wherein females were not widely studied.

The third major limitation of this preliminary study is that the mechanism of action of CBD to enhance the social behavior of BTBR mice was not elucidated. CBD has more than 65 potential therapeutic targets, and comprehensive studies to discover CBDs role in shaping behaviors through them are just beginning to unfold (1, 39). Strengths of this study include the use of multiple mouse models of ASD, a variety of behavioral tests, and measurement by mass spectrometry of serum and brain homogenate CBD and our attempt to also measure the metabolite levels. The pharmacokinetics of acute CBD administration have yet to be established in the BTBR, C57BL/6J or SERT knock-out mice, and for this reason it is unclear how long sociability benefits may last. The 10 mg/kg dose of CBD was still detectable in the BTBR brain 3 h after injection, but the pharmacological targets producing the enhanced sociability remain unclear. Future studies are needed to investigate the acute CBD behavioral time course and brain regions for each type of behavior test in these specific mouse strains and lines. Chronic treatment studies are critically needed to discover other potentially beneficial behavioral effects and side effects.

We did not find any impact of CBD treatment up to 10 mg/kg on social dominance or marble burying. One possibility is that the effects of CBD acutely administered at 10 mg/kg may have worn off before these tests were performed. Alternatively, the sequence of prior testing may have dampened the effects of CBD treatment in subsequent tests. Randomized controlled clinical trials, are now underway for use of CBD, mixtures of CBD and THC, and structurally related compounds such as cannabidivarin to treat autism symptoms (6). However, isolating the mechanism(s) of action responsible for enhancing social interaction preference will still require further preclinical studies in animal models, as such efforts may yield more selective treatments for social behavior deficits.

CBD has many different pharmacological effects, and potential physiological mediators (1, 39). Based on previously established affinities, CBD administered to mice at 10 mg/kg is unlikely to bind to and compete with CB₁ or CB₂ receptors directly (10). Interestingly, the effects of CBD (15 mg/kg) to attenuate aggressive attacks of strangers by socially isolated male mice was blocked by CB₁ and 5-HT_1A antagonists AM251 and WAY100635, which demonstrated the role of these receptors in CBD modulation of aggressive behavior (15). When CBD treatment effects were tested in dyadic social interactions with an unrestrained stranger mouse in an MK-801 model of schizophrenia, they found the 5-HT_1A antagonist WAY100635, but not CB₁ antagonist AM251 or CB₂ receptor antagonist AM630 blocked the ability of CBD to enhance social interactions (40). Thus 5-HT_1A receptor modulation may be a key mediator of social boldness or dominance, and the sensitivity of BTBR mice to CBD may differ from C57BL/6J mice, since we previously found inherent upregulation of 5-HT_1A as well as CB₁ receptor expression and G-protein coupling capacity in this strain (26). Perhaps 5-HT_1A receptor impairment may account in part for why BTBR social dominance was so low in BTBR mice (Figure 5). Indirect actions of CBD at 5-HT_1A receptors is consistent with our prior discovery that the partial 5-HT_1A agonist buspirone (2 mg/kg) acutely enhanced social interaction in 3 chamber tests of male BTBR mice (41). However direct CBD binding to 5-HT_1A is unlikely to be a mediator of these effects, based on CBD’s relatively low affinity for this serotonin receptor subtype (10).

We also see in constitutive SERT knock-out mice, which have elevated extracellular 5-HT levels, that acute CBD treatment modestly enhanced social sniffing (Figure 4B). This outcome could be consistent with a corrective, albeit indirect, action of CBD at 5-HT_1A receptors, given that reductions only in 5-HT_1A autoreceptor density in the dorsal raphe were found in male SERT knock-out mice (42), or that postsynaptic 5-HT_1A feedback inhibition is enhanced in SERT knockout mice (43). For example, treatment with the 5-HT1A antagonist WAY 100635 reduced anxiety-like traits in SERT knock-out, but not wild-type mice (44). CBD modulation of 5-HT_1A receptor properties in constitutive SERT knock-out mice is worthy of further study, as it may produce compensatory modifications in the sensitivity of this receptor.

The atypical cannabinoid receptor GPR55 is another promising therapeutic target of CBD, it occurs in many brain areas including hippocampus, cortex, striatum, and cerebellum (45), and in the blood-brain barrier (46). GPR55 is involved in spatial learning and memory (45), and its expression is doubled in Rett’s syndrome mice exhibiting deficits in social behavior (47). A structurally related cannabinoid to CBD, cannabidivarin (1), acted as an antagonist at GPR55, and it enhanced social behavior in the Rett’s syndrome mice (47). However, properties of CBD at GPR55 remain unclear. Prior reports indicate CBD may be a GPR55 agonist, however, a modified form of CBD examined was not selective as a GPR55 agonist, as evidenced by its effects at a 1 µM dose in GPR55 knock-out mice (48). Now that more selective GPR55 agonists such as O-1602 and CID2440433 are available, their use to study CBD action at GPR55 warrants further investigation of GPR55 as a therapeutic target for social deficits.

Additionally, adenosine A_2A receptors are involved in CBD action to blunt cognitive impairing effects of THC in hippocampus, and CB1 receptors dimerize with A_2A receptors in vivo in hippocampal CA1, interacting in a way that may mediate such effects (49). Other promising CBD targets that may serve modulate social behaviors as well as sensory input are transient receptor potential (TRP) and other ion channels, several of which are reported to be dysfunctional in some forms of autism (50). BKCa channels (51), as well as TRP channels, especially TRPV3 and TRPV4 are prominent examples of ion channels that CBD is known to target (12, 52–54). At many TRPVs CBD was found to act as an agonist (12), and this property may be counterproductive for its use as an autism treatment, unless the TRPVs are desensitized by such treatments, as occurs with TRPV1 (55). For example, in the Shank3 mutation mouse model of ASD, social behavior was impaired and TRPV4 upregulated, so infusion of HC-067047, a TRP4 antagonist, into the nucleus accumbens increased mouse social interaction preference (56).

Finally, in addition to known receptor targets, CBD may have other indirect and unanticipated effects. For example, two-week treatment with CBD produced epigenetic changes in the mouse hippocampus, which overall favored hypomethylation (57). Studies in animals raised concerns that orally administered CBD could be converted to THC during the digestive process, but this does not appear to occur in humans (58). CBD may also inhibit uptake of or catabolism of the endogenous cannabinoid anandamide among others (59). In adult men with and without autism, CBD administration paired with magnetic resonance spectroscopy revealed that CBD treatment modulated glutamate and γ aminobutyric acid systems differently (60). This may stem from CBD action at transient receptor potential cation channel vanilloid receptor 1 (TRPV1), which it activates and de-sensitizes (55). These same researchers demonstrated TRPV1 is found near high density glutamate expressing areas of the hippocampal cornu ammonis 1 (CA1), which is potentially involved in shaping social interaction and novelty preference (26). But most likely of all mechanisms based on binding affinity in mice, CBD competitively inhibits fatty acid amide hydrolase (FAAH), the main enzyme breaking down endocannabinoids, thereby increasing brain levels of anandamide and other endocannabinoids (13). This could be the primary mechanism to acutely increase sociability, as we saw in our mice. However, there is controversy regarding the ability of CBD to reduce human FAAH activity (59). Overall, among the potential CBD mediated mechanisms, the ones activated by 10–20 mg/kg doses are of greatest interest based on the present study, and the prior findings of Kaplan et al. (28).

Due to changes in laws regarding CBD use, it is increasingly available from dispensaries, and is frequently used without physician consultation (61). However, unregulated CBD products are often unrefined, and aside from reporting CBD/THC ratio, their potential content of up to 118 other C. sativa compounds are frequently unknown (1). Heedless of lack of preclinical or clinical research evidence, CBD is perceived to be not only beneficial for alleviating pain, but also as a potential treatment of psychiatric disorders by the public (62). While mouse studies of CBD analgesic properties are more numerous and have included pharmacodynamics of oral administration (63), our knowledge of CBD effects on these and other social behaviors still lags. This is compounded by the fact that unregulated CBD products are less expensive than prescription grade purified CBD (64). However, use of unregulated CBD products is not advisable, especially if the purity and dose of such products are not strictly controlled. Even a non-intoxicating compound such as CBD may have unforeseen adverse effects in juveniles, given the potential susceptibility of the developing cannabinoid system to disruption, and its critical role in shaping social behaviors (65–69). Therefore, the need for further preclinical research on the therapeutic potential of CBD for social behavior deficits and expanding the knowledge of its underlying mechanisms is paramount to developing targeted therapies for autism and related disorders.