Cannabinoid receptor ligand bias: implications in the central nervous system

Highlights

• Cannabinoid receptors are pleiotropically-coupled GPCRs.

• Few studies quantify cannabinoid bias using Black and Leff's operational model.

• Correlations between cannabinoid bias in vitro and in vivo are just being measured.

• Understanding cannabinoid bias could yield effective cannabinoid-based drugs.

The G protein-coupled cannabinoid receptors CB₁, CB₂, GPR18, and GPR55 regulate neurotransmission, pain, and inflammation and have been intensively investigated as potential drug targets. Each of these GPCRs is coupled to multiple effector proteins mediating divergent cellular signals. The ligand bias of cannabinoid-targeted compounds is only beginning to be quantified. Research into cannabinoid bias is now revealing correlations between bias in cell culture and functional outcomes in vivo. We present an example study of cannabinoid bias in the context of Huntington disease. In future, an understanding of cannabinoid receptor structure and quantification of ligand bias will optimize drug selection matched to patient population and disease.

Introduction

Cannabinoid receptors are G-protein-coupled receptors (GPCRs) that respond to a wide range of endogenous, synthetic and plant-derived cannabinoids [1, 2•]. The type 1 cannabinoid receptor (CB₁) is the most-abundant GPCR in the central nervous system [3]. CB₁ mediates intracellular signaling via multiple effector proteins including Gα_i/o, Gα_S, Gα_q/11, β-arrestin1, and β-arrestin2 [3, 4, 5, 6]. Activation of CB₁ directly inhibits neurotransmitter release, synapse formation, nociception, and appetite [4, 5, 6]. Consequently, pharmacotherapies targeting CB₁ have been investigated as treatments for the management of pain, addiction, energy metabolism, diabetes, movement disorders including Huntington disease (HD), Parkinson disease, multiple sclerosis, and other neurodegenerative and psychiatric conditions (Box 1) [reviewed in [8]]. The type 2 cannabinoid receptor (CB₂) and the putative cannabinoid receptors GPR18 and GPR55 interact with Gα_i/o, Gα_q/11, Gα_12/13, β-arrestin1, and β-arrestin2. These receptors are widely expressed among immunomodulatory cells [9, 10, 11, 12•]. As such, CB₂, GPR18, and GPR55 are being investigated as targets for the management of inflammation, multiple sclerosis, cancer, and hypertension (Box 1) [7, 8, 9, 10, 11, 12•]. Despite the large amount of inquiry into cannabinoid ligands as pharmacotherapeutics, no synthetic cannabinoid compounds are currently used clinically [reviewed in [8]].

Past attempts to develop cannabinoid-based therapeutics may have had limited success because the ligand bias of cannabinoids was unknown, was inappropriately matched to the pathological state or was not considered in terms of minimizing adverse effects. Ligand bias is the result of a ligand-dependent shift in a receptor's conformation that favors interaction with one effector protein at the expense of other possible effector protein(s) within the continuum of possible active receptor conformations (Box 2) [13^•]. A biased ligand binding to its receptor shifts the equilibrium of receptor-dependent signaling toward one of several possible signaling responses [13•, 14]. Differences in ligand bias between cell types or tissues is known as system bias (Box 2) [13•, 14]. Ligand bias challenges the simple classification of drugs as agonist, antagonists or inverse agonists (Box 2) [13•, 14, 15•] and opens up the possibility that specific ligands can be selected to target certain pathways for therapeutic benefit while minimizing activation of alternative pathways associated with adverse effects.

Agonist-dependent cannabinoid responses must be quantified in a way that is amenable to statistical comparison in order to draw reliable conclusions about bias [16]. Methods that compare the E_max, EC₅₀, or whole dose–response relationships of two ligands have been used as a means of indirectly assessing ligand bias [reviewed in [17]]. However, these comparisons do not control for possible differences in assay system, cell type, or receptor density and so cannot be considered as direct determinations of ligand bias [18]. The operational model described by Black and Leff [19] accounts for differences in assay system, cell type, and receptor density (Box 3) [18]. The operational model has emerged as the most-robust analytical tool for the quantification of ligand bias. Although the operational model has only been applied to cannabinoid receptor research in the last 2 years, we have used this model to demonstrate that CB₁ ligand bias can be readily defined and exploited to produce specific pharmacological effects or avoid specific side and adverse effects [2^•].

Currently, there is active investigation of biased agonists cannabinoid receptors because these receptors signal via multiple effector proteins and mediate multiple effects in vivo [5, 6]. Significant gaps remain in our understanding of pharmacodynamic properties of cannabinoids despite the long history of medicinal and recreational use of phytocannabinoids, including Δ⁹-tetrahydrocannabinol (THC) and cannabidiol (CBD). The connections between cannabinoid ligand bias, system bias, and physiologic effect have not yet been established for cannabinoid receptors, as they have been for other receptors such as β-adrenoceptor [21] and μ-opioid receptor [22]. In this review, we will discuss the current state of knowledge for cannabinoid ligand bias at the major cannabinoid receptors, provide an example of how cannabinoid ligand bias can be applied to HD, and discuss important considerations for the development of biased cannabinoids as drugs.