Introduction

Identifying and characterizing the molecular determinants of agonist efficacy in signaling pathway activation are a vital requisite of contemporary drug design. One of the determinants of agonist efficacy is the molecular structure of the agonist, and thus the receptor conformation that it induces. However, receptor conformation is also affected by interactions with various intracellular signaling proteins.^1–3 For example, the conformation of the β2 adrenergic receptor has been demonstrated to be quite distinct in the presence of the second messenger protein G_s.^2,3 Receptor activation will therefore be both ligand and tissue specific, as the assortment and abundance of intracellular signaling constituents vary between cell types. Biased signaling is the concept that different ligands acting on the same G protein-coupled receptor (GPCR), in the same tissue, can give rise to markedly different cellular responses (Fig. 1), and this is likely due to each ligand stabilizing different receptor conformations. This concept has been given many different names—“stimulus trafficking,” “functional selectivity,” and more recently, “agonist bias” or “biased signaling.” It is important to note that differential signaling pathway activation by different agonists can probably also arise as a consequence of kinetics; if there are significant differences in agonist binding kinetics, the more slowly dissociating ligands may allow receptor conformations that favor low-affinity interactions for a particular receptor/signaling molecule pair to persist long enough for productive coupling.

Traditional approaches to demonstrating bias have focused on comparisons of EC50 or Emax values within different pathways, but such methods may not account for inherent differences between pathway stoichiometry. For example, some pathways may achieve maximum response at lower receptor occupancy, resulting in higher potency for all agonists within this pathway (pathway bias). Quantifying biased signaling involves determining the effects of two or more agonists on two or more cellular responses, and comparing the agonist profiles for each pathway. More recently, the operational model of bias has been utilized⁴; this approach compares all ligands within each pathway against a reference ligand and then makes comparisons of the relative shifts between pathways relative to the reference.

Therapeutically, it is hoped that the study of biased signaling of GPCR-directed therapeutics will enable engagement of desired pathways over those not involved in the therapeutic effect. Ideally, this would eliminate on-target but unwanted or adverse effects; agonists or allosteric modulators that induce signaling in a biased manner could potentially revolutionize future therapeutic drugs. The μ-opioid receptor presents a prime example where biased signaling might be exploited at the translational level. Although opioid receptors are useful targets of analgesics, μ-opioid receptor activation also causes respiratory depression, which is suggested to be a product of β-arrestin2 recruitment.⁵ Hence, the development of an agonist that preserves the analgesic (G protein-mediated) properties of activated μ-opioid receptors without any respiratory side effects would be beneficial to avoid adverse effects, and such compounds have recently been developed in light of this hypothesis.^6,7

The cannabinoid receptor family consists of two GPCRs, cannabinoid receptor 1 (CB₁) and cannabinoid receptor 2 (CB₂), and they will be the focus of this review. CB₁ plays a role in regulating neurotransmission in many brain regions. When activated, CB₂ regulates immune responses and inflammatory pathways. Mice lacking CB₂ often demonstrate an exacerbated inflammatory phenotype⁸ and besides roles in the periphery expression in brain microglia suggest a role for CB₂ in neuroinflammation.^9,10 Recently, CB2 has been suggested to contribute to neuronal plasticity in mouse hippocampal neurons¹¹ potentially expanding the role of this receptor in the brain. The endogenous lipids anandamide and 2-arachidonoylglycerol (2-AG) are the physiological cannabinoid receptor agonists and are hence known as endocannabinoids.^12–14 A great deal of interest has centered on the potential role of CB₁ in targeting a range of central nervous system (CNS) disorders such as pain,¹⁵ anxiety,¹⁶ multiple sclerosis,¹⁷ obesity,¹⁸ nicotine addiction,¹⁹ Huntington disease,²⁰ and Parkinson’s disease.²¹ In more recent times CB₂ has also become a focus in peripheral inflammatory disorders such as nephrotoxicity.^8,22

In addition to the two well-established cannabinoid receptors, several other GPCRs have been reported to be activated by cannabinoid drugs or endocannabinoids and related molecules, including GPR55,²³ GPR18,²⁴ and GPR119.²⁵ Furthermore, endogenous and synthetic cannabinoids can activate and potentiate transient receptor potential (TRP) channels²⁶ and glycine receptors,²⁷ respectively. The potential contribution of these receptors to the therapeutic effects of cannabinoids or the physiological effects of endocannabinoids are only beginning to be explored, and much less is known about how ligands regulate their signaling.

The widespread distribution of CB₁ in the CNS provides a strong rationale for developing ligands with biased profiles to potentially avoid the consequences of activating multiple signaling pathways in many different brain regions. Biased ligands could also potentially have context-dependent effects, providing effective modulation of pathways dysregulated by disease in restricted subsets of neurons. Modulating selected downstream pathways would result in a more targeted pharmacological response, but considerable research is still required to understand which of the characterized pathways are therapeutically desirable, and under which disease condition.

(…)