The canonical Gα protein subunit for CB₁R and CB₂R is Gαi. Therefore, activation of these receptors leads to inactivation of the adenylate cyclase with the subsequent decrease in cAMP and deactivation of protein kinase A-mediated signaling. However, activation of CBRs may also lead to activation of the mitogen-activated protein kinases (MAPK) signaling cascade, regulation of ion channels, and recruitment of ß-arrestins, with subsequent regulation of Tyr kinase activity among others (Alexander et al., 2021).

Binding to GPCRs using radiolabeled compounds leads to detect one or two sites. Two sites reflect two different populations that, in the well-studied adenosine A₁ GPCR, correspond to the receptor uncoupled or coupled to the G protein. Uncoupled receptors display low affinity for agonists whereas G-protein coupled receptors display high affinity. These two affinity sites for the A₁ receptor can only be detected using agonists, i.e. antagonists have similar affinities for G-protein coupled and uncoupled A₁ receptors (see (Casadó et al., 1990) and references therein). To our knowledge radioligand binding to the CB₂R results in the detection of one single population. The two radioligands frequently used for measuring the binding to cannabinoid receptors, [³H]WIN55,212-2 and [³H]CP 55,940, are considered very potent orthosteric agonists of both CBRs, CB₁R and CB₂R. Competition assays using radioligands and non-labeled compounds in heterologous cells expressing CB₂R showed that affinities were consistent, i.e., WIN55,212-2 competed with similar low nanomolar affinity the binding of [³H]WIN55,212-2 and of [³H]CP 55,940. In similar conditions, a naturally occurring cannabinoid, cannabigerol, competed for the binding of [³H]WIN55,212-2 or [³H]CP 55,940 with a K_i in the micromolar range (Navarro et al., 2018b; Navarro et al., 2020b). This result did not fit with the decrease in cytosolic cAMP concentration obtained by nanomolar amounts of the compound. The main difference in the experimental setup was the use of isolated membranes for radioligand binding and of living cells for cAMP level measurements. The availability of novel approaches to obtain reliable receptor binding data in living cells is fortunately increasing, indeed, these methods do not require radiolabeled compounds. On using the SNAP-tag technology in cells expressing the tagged CB₂R and a validated “hot” compound (Martinez-Pinilla et al., 2016), the K_i for cannabigerol competition was 152 nM (Navarro et al., 2018b; Navarro et al., 2020b). These results show that the measured affinity of a given compound depends on the probe used for binding and allows identification of different states of the receptor or different modes to accommodate the ligand within the orthosteric center. In the case of the CB₁R, differences are more extreme as, in radioligand binding assays, natural cannabinoids may compete for the binding of [³H]WIN55,212-2 but not of [³H]CP 55,940. For instance, cannabigerol binding to CB₂R is similar if measured using [³H]WIN55,212-2 or [³H]CP 55,940, whereas there is no significant competition of binding to the CB₁R when [³H]CP 55,940 is used. In summary, cannabigerol binds to a subcompartment of the orthosteric site of the CB₁R, i.e., the orthosteric site of this receptor may be simultaneously occupied by cannabigerol and [³H]CP 55,940. These relatively recent findings add useful information to understand the variety of actions that different cannabinoids exert and also the experimental diversity between laboratories in the values of affinity and potency. This diversity may also underlie the enormous potential of cannabinoid receptors to combat a wide variety of diseases (see (Franco et al., 2020) and references therein).

Cannabidiol, one of the main components of Cannabis Sativa L. has been instrumental to detect non-orthosteric centers in CBRs. This phytocannabinoid exerts physiological effects via a variety of receptors, located both in the cell surface and inside cells. Apart from interacting with CBRs, it may interact with serotonin and peroxisome proliferator-activated receptors (Banerjee et al., 1975; Russo et al., 2005; O’Sullivan et al., 2009; O’Sullivan and Kendall, 2010; Espejo-Porras et al., 2013; Fernández-Ruiz et al., 2013; De Gregorio et al., 2019; Franco et al., 2019b; Franco et al., 2020; Echeverry et al., 2021). At first cannabidiol was considered an orthosteric ligand able to partially activate cannabinoid receptors although with low potency (McPartland et al., 2007). Recent results in two different laboratories have shown that this compound can interact in an allosteric mode with the two CBRs (Laprairie et al., 2015; Martínez-Pinilla et al., 2017). For both receptors, CB₁R and CB₂R, it acts as a negative allosteric modulator (NAM) when co-administered with an orthosteric ligand. At CB₂R it minimized the effects of JWH133 on the MAP kinase signaling pathway (Martínez-Pinilla et al., 2017). Thus, cannabidiol binds to an allosteric site at nanomolar concentrations while micromolar concentrations are required for significant binding to the orthosteric site. Accordingly, the in vitro results depend on the concentration while the in vivo actions at moderate doses should be mainly due to its binding to the allosteric site that has been very recently suggested to be close to the receptor entrance (Navarro et al., 2021) (See section: “Structural Insights into CB₂R Binding Modes”). As would be expected from an allosteric mode of action, the binding of the compound to the allosteric site causes conformational changes in such a way that biases the effect of orthosteric agonists (Navarro et al., 2018a). A more recent report shows that structural changes in the molecule shifts negative to positive modulation (of the CB₂R) thus confirming its allosteric nature (Navarro et al., 2021).

Novel approaches to achieve signaling diversity and addressing success in drug discovery are attempting the design of bitopic ligands that bind the orthosteric site and an allosteric site (Lane et al., 2013; Mohr et al., 2013; Bradley and Tobin, 2016). By combining experimental and in silico approaches an allosteric site was identified at the entrance of the orthosteric binding site of the ß-adrenergic GPCRs (González et al., 2011). This site has been termed the -extracellular- vestibule (Dror et al., 2011) or entrance (Wang et al., 2013), also metastable (Fronik et al., 2017) or secondary (González et al., 2011) binding site. Exosite is also used to describe such non-orthosteric sites when they are located at the lipidic-receptor interface (Masureel et al., 2018). Bitopic ligands designed according to these findings improve subfamily selectivity (Medina et al., 2014; Masureel et al., 2018); they also offer signaling bias and better off-rates (Valant et al., 2012; Lane et al., 2013). Knowing that unlike GPCRs for polar compounds, CBRs do not have the orthosteric center accessible from the extracellular milieu, we designed bitopic ligands able to enter into the CB₂R orthosteric site but also able to interact with amino acids located at the receptor transmembrane portals (Morales et al., 2020). Signaling assays in the CB₂R wild-type and specific mutants led us to discover the first CB₂R bitopic ligands. These compounds, which consist of two chromenopyrazole moieties linked by methylene spacers of different lengths, can bind to the orthosteric site and to an exosite. Bitopic ligands showed to be CB₂R selective and, as depicted in Figure 1, may likely extend from the orthosteric site, the vestibule and an “allosteric exosite” able to accommodate the same moiety that sits in the orthosteric site.

Can a given compound be more efficacious at targeting a cell that expresses CB₂R in a particular conformation? and/or can a CB₂R in a particular cell type be more likely to respond to the challenge of a given compound?

The pharmacology of cannabinoid receptors is complex. As discussed above, binding data can depend on the radioligand used as the probe, and the effects of a given compound on a given receptor are not always consistent across laboratories. At present we have enough data to realize that there are many possibilities for CB₂R-mediated responses that may turn into novel and powerful possibilities for drug discovery.

The complex pharmacology of the CB₂R has likely delayed the identification of CB₂R-containing macromolecular complexes, whose occurrence has been demonstrated in natural sources (i.e. not only in heterologous expression systems). Such interactions modify binding and/or function. Current data suggest that the receptor environment modifies the conformation and, accordingly, the binding and effects of orthosteric and non-orthosteric ligands. Interaction of the CB₂R with other GPCRs may be searched in http://www.gpcr-hetnet.com/ (using the gene name: CNR2) (Borroto-Escuela et al., 2014). Figure 2 shows the STRING analysis of the interactions of the receptor which indicates mandatory interactions with G proteins, and interactions with the CB1R and with other GPCRs. In www.gpcr-hetnet.com and in Figure 2 interactions of CB₂R with further GPCRs are not yet included (they have not yet been incorporated into the STRING database). Also missing are the recently described interactions of the CB₂R with glutamate N-Methyl-D-Asp (NMDA) ionotropic receptors (Rivas-Santisteban et al., 2021). From a therapeutic perspective, the fact that CB₂R may interact with other receptors that are also targeted by cannabinoids, for instance with GPR18 and GPR55, is of high interest (Balenga et al., 2014; Reyes-Resina et al., 2018; Martínez-Pinilla et al., 2019; Martínez-Pinilla et al., 2020; Rivas-Santisteban et al., 2021).

So far, no major change has been detected concerning the nature of the G protein coupling of CB₂R in a macromolecular environment as it occurs for D₁ and D₂ dopamine receptors. Whereas the D₁ is coupled to Gαs and D₂ to Gαi, the macromolecular complex formed when the two receptors are co-expressed in the same neuron couples to Gαq (Rashid et al., 2007; Hasbi et al., 2009; George et al., 2014; Perreault et al., 2015). Notwithstanding, conformational changes that affect the binding and signaling outputs produced by a given agonist have been shown in the interactions with the Gαi-coupled CB₁R (Callén et al., 2012; Sierra et al., 2015; Angelats et al., 2018), the Gαs-coupled adenosine A_2A receptor (Franco et al., 2019a), and the ionotropic NMDA receptor (Rivas-Santisteban et al., 2021).

In one of the first studies of biased agonism in GPCR heteromers (CB₁R/CB₂R), Navarro and co-workers showed that the allosteric effect of CBD was particularly noteworthy for the endocannabinoid anandamide but also that the effect tested using different agonists was smaller in the heteromer (Navarro et al., 2018a). These results confirmed that CBD acts as an allosteric modulator (for both receptors) also suggesting that the formation of the heteromer leads to conformational changes that make it less sensitive to the action of this phytocannabinoid. There are several examples of conformational changes induced by receptor-receptor interactions, i.e. by heteromer expression (Franco et al., 2007; Ferré et al., 2009; Franco et al., 2016). In the case of the CB₂R, indirect evidence is provided by potentiation of receptor-mediated signaling when forming heteromers with the adenosine A_2A receptor (Franco et al., 2019a).

The binding of orthosteric and non-orthosteric ligands alters the conformation of the receptor, but can ligands alter the environment? The answer to this question will take time as there is little background on the regulation of, for instance, heteromer formation.

Defining the target in the right context and delineating contextual changes due to ligand-induced regulation of the structure of the CB₂R-contaning macromolecule, may further improve the rational design of therapeutic drugs (orthosteric and non-orthosteric) targeting the CB₂R.