Biologically active compounds derived from the Cannabis sativa plant are referred to as ‘phytocannabinoids’. The two most-thoroughly studied phytocannabinoids are ∆⁹-tetrahydrocannabinol (∆⁹-THC) and cannabidiol (CBD). Both of these compounds are being intensively studied for their utility in treating chronic and acute pain, epilepsy, anxiety, and modulating appetite, and their potential toxicities and side effects¹. The effects of phytocannabinoids in the human body are thought to be mediated by many different receptors including the type 1 and type 2 cannabinoid receptors (CB1R and CB2R, respectively); as well as other G protein-coupled receptors (GPCRs) such as serotonin (5HT) receptors, orphan GPCRs (e.g. GPR18, GPR55, and GPR119) and ligand-gated ion channels. The pharmacology of ∆⁹-THC is relatively well-established: THC is a CB1R and CB2R partial agonist^1–5. The pharmacology of CBD is far-less clear. CBD has been described as a CB1R negative allosteric modulator, CB2R antagonist, GPR18 agonist, GPR55 antagonist, among other effects^6–16. Beyond ∆⁹-THC and CBD, there are thought to be at least 120 other phytocannabinoids found in Cannabis for which the receptor-mediated mechanisms are still actively being investigated^2–5. Examples of these lesser-known phytocannabinoids include ∆⁹-tetrahydrocannabidiolic acid (∆⁹-THCa), tetrahydrocannabivarin (THCV), cannabigerol (CBG), cannabichromene (CBC), and cannabidivarin (CBDV)^2–5.

CB1R and CB2R, their endogenous agonists 2-arachidonoylglycerol (2-AG) and anandamide (AEA), and the associated anabolic and catabolic enzymes constitute the endocannabinoid system^1,5. CB1R activation inhibits nociception and locomotor activity, activates reward pathways, and regulates mood, memory and cognition, and central hormone homeostasis¹. CB2R activation inhibits the inflammatory response in lymphocytes and microglia¹. Interest in the development of compounds that target CB1R and CB2R is at an all-time high because of their multiple physiological roles and consequent associations with many different disease states. To date, this interest has focused mainly on synthetic compounds that target CB1R such as antagonists/inverse agonists (e.g. rimonabant), negative allosteric modulators (e.g. Org27569), and positive allosteric modulators (e.g. GAT211 and ZCZ011); as well as inhibitors of cannabinoid catabolic enzymes that increase 2-AG or AEA levels^17,18. Comparatively, little research has been done to isolate and characterize the wide variety of naturally-occurring cannabinoids of Cannabis (i.e. phytocannabinoids)^2,19. The work that has been done indicates that many of these ligands are active at CB1R, CB2R, and other receptors^2,19. Within the last three years, an influx of high-quality studies has examined the pharmacology of subsets of these phytocannabinoids for their specific activity, receptor affinity, ligand bias, and neuroprotective properties in rodent models^19–22. Previously, our group has characterized the pharmacology of endogenous, synthetic, and plant-derived cannabinoids in vitro and in vivo^{10,11,15,23,24}. In the present study, we utilized high- and medium-throughput assay systems to screen phytocannabinoids against one another and a reference compound, CP55,940, to directly assess their pharmacodynamic activity.

The primary aim of the present study was to explore the CB1R-dependent, CB2R-dependent, and in vivo pharmacology of 8 phytocannabinoids: ∆⁹-THC, ∆⁹-THCa, THCV, CBD, cannabidiolic acid (CBDa), CBDV, CBG, and CBC (Fig. 1). All compounds were assayed for the displacement of [³H]CP55,940, inhibition of forskolin-stimulated cAMP accumulation, and βarrestin2 recruitment in Chinese hamster ovary (CHO) cells stably expressing either human CB1R or CB2R. Compounds were assessed for signaling bias between inhibition of cAMP and βarrestin2 and selectivity between the two cannabinoid receptors. Compounds were further screened in the in vivo tetrad assays for cataleptic, hypothermic, anti-nociceptive, locomotive, and anxiety-modifying activities. We observed that all compounds tested displayed some degree of activity at CB1R or CB2R, with several being weak partial agonists. As Canada and other jurisdictions increasingly permit the use of Cannabis for medicinal and non-medicinal purposes, this research provides critical insight about the therapeutic potential and utility of phytocannabinoids.

In general, we observed that all phytocannabinoids tested displayed some degree of activity at either CB1R or CB2R in cell culture assays. The incomplete competition of these phytocannabinoids with [³H]CP55,940 suggests that the occupied binding site of these ligands differs slightly from that of CP55,940. It is possible that the phytocannabinoids tested bound only a subset of amino acids in the CB1R ligand binding site(s) compared to CP55,940, as has been shown for CBD, Org27569, rimonabant, and anandamide^10,18,27,28.

∆⁹-THC that is commonly consumed in cannabis products is a decarboxylated derivative of the naturally present ∆⁹-THCa¹⁹. In this study, ∆⁹-THCa partially displaced [³H]CP55,940 from CB1R and CB2R, with greater affinity for CB2R; and was a weak partial agonist of CB1R-dependent inhibition of cAMP accumulation and a weakly potent agonist of CB2R-dependent cAMP accumulation. ∆⁹-THCa produced anti-nociceptive and hypolocomotive effects at 3 and 10 mg/kg, and anxiolytic-like effects at 10 mg/kg. ∆⁹-THCa has previously been shown to bind murine cannabinoid receptors weakly²¹. Similar to our findings, Palomarés et al.²⁹ recently reported ∆⁹-THCa binding to CB1R and CB2R, with potential orthosteric and allosteric activity at CB1R. In vivo, the anti-inflammatory activity of ∆⁹-THCa in a rodent model of arthritis was CB1R- and peroxisome proliferator-activated receptor γ (PPARγ)-dependent²⁹. Similar neuroprotective and PPAR-dependent effects have been observed in rodent models of Huntington’s disease³⁰. Here, ∆⁹-THCa displayed similar efficacy to ∆⁹-THC in the cAMP inhibition assay but a lower affinity for CB1R in [³H]CP55,940 competition experiments. These data corroborate the findings of Palomarés et al.²⁹ and indicate the antinociceptive effect of ∆⁹-THCa observed here is likely in part due to CB1R agonism. Our data, together with previous studies, indicate ∆⁹-THCa may mediate neuroprotective and anti-inflammatory actions via CB1R, CB2R, and PPARγ when administered at sufficiently high concentrations and in the absence of other cannabinoids.

THCV has been posited to act at both CB1R and CB2R, and is considered by some to behave as a CB1R antagonist and CB2R agonist^31–33; reviewed in¹⁹. Here, we found that THCV was able to displace [³H]CP55,940 from both CB1R and CB2R. Unlike previous studies suggesting THCV is a CB1R antagonist, THCV did produce a weak partial agonist response in the CB1R cAMP inhibition assay. THCV was an agonist of CB2R-dependent cAMP inhibition, as described previously³⁴ and βarrestin2 recruitment. In vivo, THCV produced cataleptic (10 mg/kg), hypothermic (10 mg/kg), anti-nociceptive (3 and 10 mg/kg), hypolocomotive (3 and 10 mg/kg), and anxiolytic (3 and 10 mg/kg) effects. These effects are consistent with others’ observations of anti-epileptic, hypolocomotive, neuroprotective effects in the range of 0.25–2.5 mg/kg^32,33; reviewed in¹⁹. Bolognini et al.³⁴ demonstrated that THCV is able to reduce hyperalgesia in mice via both CB1R and CB2R, because anti-hyperalgesic effects were limited by both the CB1R antagonist SR141716A and the CB2R antagonist SR144528. Given the in vitro efficacy of THCV at both CB1R and CB2R, it is likely that the in vivo effects of THCV observed here are dependent on both cannabinoid receptors, as has been described previously³⁴.

CBD has been much more-extensively studied than the other phytocannabinoids. CBD has been described as a CB1R negative allosteric modulator^10,14,15,29; CB2R antagonist³⁵; GPR18, GPR55, and 5HT3A antagonist ^1,36; 5HT1A, 5HT2A, adenosine 1A, and PPARγ partial agonist^16,37–39; and an allosteric modulator of the µ- and δ-opioid receptors^4,9. In this study, we included CBD in an assay format that would detect agonism at CB1R and CB2R. Not surprisingly, CBD displayed low affinity at both CB1R and CB2R and displayed minimal activity at both receptors. Similarly, we observed no effects of CBD between 1 and 10 mg/kg in the tetrad of assays, as expected. CBD has been shown in animal models of disease to have anti-convulsant, anti-inflammatory, and anti-nociceptive effects; all of which appear to be independent of CB1R and CB2R¹⁹.

CBDa was inactive at CB1R and displayed weak partial agonism at CB2R in our cell culture assays. Navarro et al.¹² reported higher than anticipated affinity and activity of CBDa in both [³H]CP55,940 binding for CB1R and CB2R in CHO cells and signaling in HEK-293T cells, which were similar to our findings. In vivo, CBDa produced significant hypolocomotive and anxiolytic effects in the OFT. Beyond CB1R and CB2R, CBDa has shown efficacy in reducing inflammatory pain and nausea^40,41, seizure incidence⁴², and Parkinsonian signs in rodent models¹³. Accumulating data suggest the in vivo effects of CBDa are 5HT1A-mediated⁴⁰. Therefore, the effects observed in this study in the tetrad maybe 5HT1A-mediated, rather than cannabinoid receptor-dependent.

CBDV displayed little to no activity at CB1R but did display demonstrable affinity and activity at CB2R. Similarly, Navarro et al.¹² observed that CBDV was nearly inactive at CB1R with greater activity at CB2R. However, the binding affinity of CBDV to CB2R was greater in our experiments compared to those of Navarro et al.¹². This difference may be due to differing radioligand concentrations and/or differential expression of the receptor in our cell models. In our assays, CBDV displayed a bias toward the inhibition of cAMP relative to βarrestin2 recruitment; whereas previous reports in HEK-293T cells have described the opposite, although these bias analyses were conducted by different methods¹². CBDV produced no significant in vivo responses in our assays. Earlier reports have shown anti-convulsant effects associated with CBDV treatment in rodent models and occur at 200–400 mg/kg p.o., doses higher than were tested here (reviewed in¹⁹).

In CHO cells stably expressing human receptors, CBG was a weak partial agonist of both CB1R- and CB2R-dependent inhibition of cAMP that displaying a low affinity for CB1R and comparatively higher affinity for CB2R. Previous reports have similarly described CBG as a weak partial agonist of these two receptors, with low (> 1 µM) affinity for both^6–8,43,44. Interestingly, previous observations have also indicated that CBG-dependent recruitment of βarrestin via the cannabinoid receptors is less potent and efficacious than other signal transduction pathways^12,44, similar to our findings. Although our data indicate a higher affinity for CBG at CB2R than previous reports (500 nM vs. > 1 µM), all of our other observations are in concordance with previous findings^12,44. Whereas Navarro et al.¹² observed CBG to be a balanced, non-biased, ligand at CB1R and CB2R in HEK-293T cells; CBG displayed a bias for the inhibition of cAMP in CHO cells stably expressing CB1R or CB2R. In vivo, CBG produced a small but statistically significant anti-nociceptive effect at 3 mg/kg and an anxiolytic-like effect in the OFT at 10 mg/kg. CBG has previously been reported to have disease-ameliorating anti-inflammatory effects in mouse models of Huntington’s disease, multiple sclerosis, Parkinson’s disease, and amyotrophic lateral sclerosis^45–49. These effects were likely mediated by PPARγ, and not cannabinoid receptors^43,48,49. CBG has also been shown to have in vitro activity at GPR55, 5HT1A, the α2 adrenoceptor, and several transient receptor potential (TRP) channels^6–8. Therefore, it is possible the in vivo anti-nociceptive and anxiolytic effects observed in our study occurred via these, cannabinoid receptor-independent, mechanisms.

Similar to CBG, CBC was a partial agonist of both CB1R- and CB2R-dependent signaling, with great selectivity and potency at CB2R relative to CB1R in the assays utilized here. Previous studies have observed weak partial agonism of CBC at both CB1R and CB2R⁸. More recently, Udoh et al.⁵⁰ reported that CBC produced CB2R-dependent membrane hyperpolarization in AtT20 cells. This effect was absent in CB1R-expressing AtT20 cells, indicating CB2R-specificity for CBC⁵⁰. In vivo, CBC is able to increase neuronal viability via ERK phosphorylation in nestin-positive neural stem cells, but the receptor-specific mechanism has not been described for this effect⁵¹. CBC has also been shown to be anti-inflammatory and reduce hypermobility in a mouse model of gut inflammation, although these effects occur via TRPA1 and not cannabinoid receptors⁵². Similar to the findings of Izzo et al.⁵², CBC did produce a small, but statistically significant anti-nociceptive effect in the tail withdrawal assay and a hypolocomotive effect in the OFT; both of these effects may have occurred independently of cannabinoid receptors.

Based on these data, it is unclear whether the cannabinoids tested here bind an identical orthosteric site to that of ∆⁹-THC or CP55,940. Future experiments utilizing site-directed mutagenesis will need to assess this question directly. Moreover, the partial agonist effects displayed by these ligands suggests they may be functionally antagonistic in the presence of higher agonist concentrations and in vivo. This functional antagonism has been previously demonstrated for ∆⁹-THC itself when administered alongside full and potent CB1R agonists^10,53–56. A growing body of literature also supports the notion that phytocannabinoids such as ∆⁹-THCa, CBD, and others may be able to occupy both orthosteric and allosteric sites with varying affinity, further complicating our understanding of the cannabinoid receptors^10,15,29. Future work assessing the potential antagonist activity of these compounds in the presence of a full agonist such as CP55,940 will be able to better-classify the mechanisms of action for these compounds beyond what has been done here.

This work represents an initial step into the assessment of the pharmacology for a subset of Cannabis-derived phytomolecules at the most-thoroughly studied cannabinoid receptors, CB1R and CB2R. It is possible that if the phytocannabinoids tested here that displayed CB1R agonist and in vivo activity were present at sufficiently high concentrations in cannabis products; they may produce intoxicating effects similar to those of ∆⁹-THC. However, given the typically low content of these phytomolecules in cannabis products and their weak displacement of [³H]CP55,940, these compounds are all probably more likely to diminish ∆⁹-THC’s effects in whole organisms. Beyond CB1R, ∆⁹-THC itself is known to modulate the signaling of several proteins, including the orphan GPCR GPR55, and the TRP vanilloid 1 Ca²⁺ channel (TRPV1)¹⁷. Other cannabinoids, such as CBD, modulate the activity of a wide array of cannabinoid and non-cannabinoid receptors, including CB1R as a negative allosteric modulator, CB2R and 5HT1A as a partial agonist, the µ-opioid receptor, and PPARs¹⁷. Our in vivo observations that CBDa, CBG, and CBC (among other compounds tested) mediated changes in locomotion and time in the centre quadrant despite having little in vitro activity at CB1R, warrant further study to determine what other receptors are utilized by these ligands. Therefore, in order to comprehensively understand the poly-pharmacology of cannabinoid receptors in vivo, other receptor targets must be considered.

Finally, cannabis products contain many phytomolecules that are co-administered when Cannabis is consumed. This study represents an initial foray into potential between- and among-phytomolecule interactions that can be built upon gradually. Pharmacology is a reductionist approach to biochemical interactions that cannot always model the complex interactions occurring in nature. As initial characterizations of single ligands are made, more complex combinatorial testing of pharmacology can be carried out. Eventually, we hope to assess potential interactive effects of complex cannabinoid mixtures and assess pharmacodynamic—and pharmacokinetic—differences that stem from chemically distinct ligands.

Funding for the project was provided by an NSERC Collaborative Research and Development Grant in partnership with Aurora Cannabis to RBL and LAH (#517839-17) as well as a CIHR Partnership Grant with GlaxoSmithKline to RBL (#387577). KAM and EDK received NSERC Undergraduate Student Research Awards. TB received scholarship support from University of Saskatchewan College of Pharmacy and Nutrition.

A.Z. and K.A.M. designed, conducted, and analyzed cell culture experimental data, and contributed to the writing and editing of the manuscript. A.Z., H.-J.J.K., E.D.K., and T.B. designed, conducted, and analyzed animal experimental data and contributed to the writing and editing of the manuscript. E.D.K. and C.S.F. assisted with experimental blinding and analysis of data they were blinded to. P.J. and L.A.H. contributed reagents critical to the execution of these experiments and contributed to the writing and editing of the manuscript. R.B.L. assisted with experimental design and oversaw the writing and editing of the manuscript.

This project was partially funded through monies and the provision of essential compounds and reagents via an NSERC Collaborative Research and Development Grant in partnership with Aurora Cannabis. Aurora Cannabis did not direct, advise, or oversee the use of funds or experiments conducted.

is available for this paper at 10.1038/s41598-020-77175-y.