Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Jan 10.
Published in final edited form as: Methods Enzymol. 2017 Jul 5;593:259–279. doi: 10.1016/bs.mie.2017.06.031

Approaches to Assess Biased Signaling at the CB1R Receptor

Robert B Laprairie 1, Edward L Stahl 1, Laura M Bohn 1,1
PMCID: PMC6953168  NIHMSID: NIHMS1064750  PMID: 28750807

Abstract

G protein-coupled receptors, such as the cannabinoid type 1 receptor (CB1R), have been shown to interact with multiple binding partners to transmit signals. In both transfected cell systems and in endogenously expressing cell lines, CB1R signaling has been described as multifaceted. The question remains as to how this highly widely expressed receptor signals in a given cell at a given time in vivo. The concept of functional selectivity, or biased agonism, describes the ability of an agonist to engage the receptor in a manner that preferentially engages certain signaling interactions (e.g., G proteins) over others (e.g., β-arrestins), presumably by stabilizing certain receptor conformations. There is growing interest in using such properties of ligands to direct signaling downstream of CB1R toward desirable therapeutic outcomes and to avoid adverse side effects. While it is not currently clear what pathways should be engaged and which should be avoided, the development of biased agonist tool compounds will aid in answering these questions. In this chapter, we discuss the approaches and caveats to assessing biased agonism at the CB1R.

1. INTRODUCTION

The endocannabinoid system is composed of the cannabinoid receptors, their endogenous ligands—2-arachidonoylglycerol (2-AG) and anandamide (AEA), and the anabolic and catabolic enzymes of these endogenous ligands (Pertwee, 2008). The canonical cannabinoid receptors are the type 1 cannabinoid receptor (CB1R) and type 2 cannabinoid receptor (CB2R); both of which are G protein-coupled receptors (GPCR) (Bayewitch et al., 1995; Pertwee, 2008; Pinto et al., 1994). Cannabinoid receptor activity is modulated by a structurally diverse array of endogenous, synthetic, and plant-based ligands (Pertwee, 2008; Pinto et al., 1994). The CB1R is expressed throughout the periphery and is the most abundant GPCR in the central nervous system (Bouaboula et al., 1995; Pertwee, 2008; Pinto et al., 1994). Although the CB1R is classically considered a Gαi/o-coupled GPCR, it has also been shown to couple with GαS and Gαq/11; moreover, CB1R interactions with β-arrestins have been established (Felder et al., 1995; Glass & Northup, 1999; Hudson, Hébert, & Kelly, 2010; Mukhopadhyay & Howlett, 2005). In cell culture assay systems, activation of CB1R has been shown to activate multiple downstream pathways as well, including regulation of cyclic adenosine monophosphate (cAMP) accumulation, phosphorylation of extracellular-regulated kinase (ERK) and CREB, and fluxuations in intracellular Ca2+, among other effects (Glass & Northup, 1999; Hudson et al., 2010; Lauckner, Hille, & Mackie, 2005; Mukhopadhyay & Howlett, 2005; and “Protocols and Good Operating Practices in the Study of Cannabinoid Receptors” by Consroe-Bram et al.; “Real-Time Measurement of Cannabinoid Receptor-Mediated cAMP Signaling” by Hunter et al.; “Assessing Allosteric Modulation of CB1 at the Receptor and Cellular Levels” by Scott and Kendall). In rodent studies, CB1R activation produces multiple physiological responses, including inhibition of neurotransmitter release, antinociception, hypoactivity, catalepsy, hypothermia, increased appetite and insulin sensitivity, reduced gastrointestinal transit, and antiinflammatory effects (Felder et al., 1995; Izzo & Coutts, 2005; Motaghedi & McGraw, 2008; “Modeling Neurodegenerative Disorders for Developing Cannabinoid-Based Neuroprotective Therapies” by Fernández-Ruiz et al.). Given this wide spectrum of cellular and physiological effects, CB1R is considered a potential therapeutic target for the treatment of a wide range of ailments, including: chronic and acute pain; addiction; metabolic disorders and diabetes; epilepsy; multiple sclerosis, anxiety; and neurodegenerative disorders such as Huntington’s and Parkinson’s disease (Ranieri, Laezza, Bifulco, Marasco, & Malfitano, 2016).

Given that signaling at CB1R can propagate diverse and multifaceted physiological responses, a question arises as to whether the effects of cannabinoids could be harnessed to improve desirable therapeutic aspects while limiting unwanted side effects. Indeed, diverse signaling profiles have been identified downstream of CB1R in cell culture systems, yet a question also remains as to how CB1R expressed in different tissues respond in the endogenous system. Evidence for signaling bias at CB1R in vivo has been presented by several groups as separation between the potency of CB1R agonists to produce antinociceptive, hypothermic, and cataleptic effects (Breivogel, Lambert, Gerfin, Huffman, & Razdan, 2008; Grim et al., 2016; Nguyen et al., 2012; Wiley, Smith, Razdan, & Dewey, 2005). Therefore, if certain signaling pathways are associated with certain physiologies, then there may be opportunities to refine CB1R signaling and preserve one pathway while avoiding another (Breivogel et al., 2008; Grim et al., 2016; Nguyen et al., 2012; Wiley et al., 2005). This idea, however, only works when the desirable effects are mediated by different pathways than the adverse effects, which is not necessarily going to be the case.

The concept of inducing preferential signaling to one pathway over another has been termed functional selectivity or ligand bias (Kenakin, 2010; Kenakin & Christopoulos, 2013; Luttrell, Maudsley, & Bohn, 2015; Fig. 1). In this paradigm, the chemical structure of the ligand is proposed to interact with the receptor in a manner that preferentially promotes or stabilizes a conformation of receptor that favors the interaction with certain intracellular effectors (e.g., G proteins) more than another (e.g., β-arrestin) (Christopoulos, 2014; Kenakin, 2010; Kenakin & Christopoulos, 2013). As such a G protein-biased agonist would be expected to produce more signaling to G protein pathways than observed in a β-arrestin recruitment assay. However, the key here is the term “more,” which emphasizes that determinations of bias are measures of relative performance and that “more” is a term of comparison. Comparison in the case of agonist pharmacology requires a standard, or a reference agonist, to which is assigned the characterization of displaying “balanced” signaling, or the ability to stimulate the two pathways of interest with high efficacy. As such, bias is intrinsically context dependent—as it will rely on the biological system and upon the properties of the reference agonist. Therefore, in pharmacological assessments, the determination of bias serves as a means to rank and compare properties of new agonists as compared to the performance of a reference agonist between multiple assays.

Fig. 1.

Fig. 1

Open in a new tab

CB1R-mediated signaling. (A) CB1R interacts with multiple effector proteins to produce different intracellular signals. These effector proteins include: Gαi/o, GαS, Gαq/11, Gα12/13, Gβγ, and β-arrestins. (B) The binding of different agonists to CB1R stabilizes the receptor in conformations that preferentially enhance coupling to certain effector proteins and diminish coupling to other effector proteins. Consequently, cellular signaling pathways are effectively enhanced or diminished; this describes ligand bias. Here, three theoretical agonists bind CB1R and each facilitates a unique receptor conformation and unique signaling outputs.

For CB1R, there are an increasing number of publications describing the different performance of ligands relative to reference ligands across many assays, or, biased agonism (Delgado-Peraza et al., 2016; Khajehali et al., 2015; Laprairie, Bagher, Kelly, & Denovan-Wright, 2016). The benefit of developing biased agonists has been more apparent for some GPCRs, including the angiotensin IIa receptor (Christensen et al., 2010), kappa opioid receptor (Brust et al., 2016), and the mu opioid receptor (Luttrell et al., 2015; Rankovic, Brust, & Bohn, 2016; Zhou & Bohn, 2014). Indeed, the first clinical application of a biased mu opioid receptor agonist, oliceridine® is in late-stage clinical trials with the goal of preserving pain relief while avoiding respiratory side effects (DeWire et al., 2013; Soergel et al., 2014). For the CB1R, it is not immediately clear what benefit there will be to developing CB1R-biased agonists. However, given the wide distribution of expression of this receptor, and the diverse number of signaling cascades it has potential with which to engage, the concept of pharmacologically harnessing its downstream signaling potential may open new therapeutic avenues for controlling CB1R-mediated physiologies. Therefore, a considerable effort is ongoing for developing biased CB1R ligands to promote specific receptor-dependent signaling responses in hopes of avoiding undesirable physiologies (Kenakin & Christopoulos, 2013; Khajehali et al., 2015; Laprairie et al., 2016).

The property of bias is intrinsically context dependent; and therefore, it is highly necessary to limit the influence of the cell-based assay when comparing the performance of an agonist to a reference agonist. The adaptation of mathematical modeling allows for such intra- and intersystem normalization to be weighted when comparing dose–response derived potencies and efficacies. In this chapter we will briefly review the quantification of bias of CB1R ligands using the operational model first described by Black and Leff (1983). We will then describe assay systems that can be used to collect data for bias analyses and consider the caveats and limitations therein.

2. QUANTIFYING CB1R LIGAND BIAS

Past attempts to understand CB1R ligand bias or functional selectivity have relied on qualitative comparisons using “bias plots” or direct comparisons of maximum response achievable from an applied drug (Emax), half maximal effective concentration (EC50) (Rajagopal et al., 2011). The determination and rank order comparison of EC50 and/or Emax are extremely useful as tools to highlight instances when the two-state model does not adequately describe the data (Clarke & Berg, 2010). However, these approaches do not provide quantitative measures of ligand activity and are less useful at minimizing differences in assay system, cell type, receptor density, or assay kinetics. Due to these limitations, rank order comparison should not be used to determine “bias” (Ehlert, 2015; Luttrell, 2014; Rajagopal et al., 2011; Stahl, Zhou, Ehlert, & Bohn, 2015). Instead, the operational model (Black & Leff, 1983; Eq. 1) more appropriately handles the complexities of intrasystem comparison (assay system, cell type, receptor density, etc.) and is well suited to be applied in a straightforward manner to the familiar concentration-response data.

E=Emax[A]nτn[A]nτn+([A]+KA)n (1)

In the operational model Emax is the maximum response, A represents the agonist concentration, n the Hill slope, τ the agonist efficacy, and KA the agonist affinity (Black & Leff, 1983). A useful reparameterization of this model has been elaborated and employed over the past couple years. The transduction coefficient, defined as the (τ/KA) or synonymously the log R, provides a composite parameter value that can be produced and used to compare agonist activity between different systems (Kenakin, 2010). The transduction coefficient is not a descriptive parameter value (i.e., potency or efficacy) but rather a relative measure of the ability of a ligand to activate the receptor in the target system (Kenakin, 2010; Kenakin & Christopoulos, 2013). A reference agonist is commonly run in parallel in these experiments to provide a method of direct normalization across systems. The reference agonist is chosen based on the appreciable view that it is commonly used, widely available, and acts as a well-behaved full agonist in each system of study (Kenakin & Christopoulos, 2013; Rajagopal et al., 2011). The reference agonist should be a full agonist in each of the assay systems being used (Kenakin & Christopoulos, 2013; Rajagopal et al., 2011). For many receptors, the first choice is, therefore, the endogenous ligand(s) (Khajehali et al., 2015). CB1R may be unique in this point because AEA and 2-AG are not consistent full agonists of the receptor (Pertwee, 2008). Instead, the compound CP55,940 is a commonly used and generally agreed upon full agonist at the CB1R. Based on the concentration-response curve of CP55, 940, the relative agonist activity of test agonists can be determined when fit to the operational model (Eq. 1). A useful normalization of this analysis is to produce the relative activity of each test agonist. This normalization is produced by calculating the difference between the transduction coefficient of a test agonist and the reference agonist and results in a relative transduction coefficient (Δ log R) within an assay (Eq. 2).

ΔlogR=log(τ∕KA)Test compound−log(τ∕KA)Ref compound (2)

To best-fit data to this model and derive Δlog R, we recommend sharing log R, n, and Emax across datasets and constraining log KA between 0 and – 15 (femtomolar, i.e., an upper bound that described as extremely high-affinity ligand) (Ehlert, 2015; Stahl et al., 2015). Δlog R values can then be compared for a test compound(s) between assays to yield a system-independent bias factor for that compound(s), ΔΔlog R (Eq. 3; Ehlert, 2015; Stahl et al., 2015).

logbias=ΔΔlogR=ΔΔlog(τ∕KA)R1−R2=Δlog(τ∕KA)R1−Δlog(τ∕KA)R2 (3)

The antilog of ΔΔlog R is the bias factor of the test compound (Kenakin, Watson, Muniz-Medina, Christopoulos, & Novick, 2012; Stahl et al., 2015). ΔΔlog R values from multiple, independent, experiments can be used to calculate mean bias factors with associated error (Stahl et al., 2015). These data can then be compared using standard statistical analyses.

3. ASSAYS USED TO EXAMINE CB1R LIGAND BIAS

Several cellular effects can be compared in order to quantify ligand bias using the operational model, and multiple assays may exist to measure a cellular effect. Here, we describe some of the common downstream effectors of CB1R and the assays available to quantify their responsiveness, as well as advantages and disadvantages of each.

3.1. G Protein Coupling

One of the most proximal effectors of GPCR activation are their namesakes—the G proteins. CB1R has classically been described as a Gαi/o-coupled GPCR (Felder et al., 1995). Quantification of G protein coupling to a GPCR can be conducted using the [35S]GTPγS assay. Agonist binding to the GPCR results in the exchange of guanosine diphosphate for guanosine triphosphate (GTP). Guanosine 5′-O-[gamma-thio]triphosphate (GTPγS) is a nonhydrolyzable form of GTP. In this assay, a radiolabeled form of GTPγS, [35S]GTPγS, is incubated with dissociated membranes from cultured cells (e.g., cells stably expressing CB1R) or tissue homogenate and the compounds being tested. Compound concentration-dependent coupling of the alpha subunit of G protein (Gα) to the receptor results in the binding of [35S]GTPγS. Bound [35S]GTPγS is then separated from free [35S]GTPγS by filtration, and Gα-bound [35S]GTPγS is quantified (Bohn, Zhou, & Ho, 2015). [35S]GTPγS binding can be quantified using cell lines stably expressing CB1R or tissue from mouse brain homogenates (Bohn et al., 2015; Hua et al., 2016; Janero et al., 2015). We have described the methodology of these assays in detail (Bohn et al., 2015).

[35S]GTPγS binding reveals the efficiency of coupling to the most prevalent G protein signal available in the system, and different types of Gα protein binding cannot be readily determined or distinguished (Bohn et al., 2015). In cell models the predominant G protein is often Gαi/o, but in endogenous settings these may be less-characterized Gα proteins, such as GαZ (Hinton et al., 1990; Wettschureck et al., 2005). Additional experiments, such an immunoprecipitation and western blot, are required to determine G protein type (Hinton et al., 1990; Wettschureck et al., 2005). A major advantage [35S]GTPγS binding is that the signal is measured proximal to the receptor and can therefore be considered a more direct measurement of agonist-directed receptor signaling than other output measures discussed later (Kenakin et al., 2012).

3.2. β-Arrestin Recruitment

CB1R interacts with both β-arrestin1 and β-arrestin2 (Delgado-Peraza et al., 2016; Laprairie et al., 2016). Like G protein coupling, β-arrestin recruitment is proximal to the receptor and therefore a direct measurement of agonist-induced changes in receptor conformation. Several approaches have been developed to quantify β-arrestin recruitment, but the most widely used rely on bioluminescence resonance energy transfer (BRET), enzyme complementation, and transcriptional activation.

BRET, like FRET, relies on the transfer of energy between a donor protein and an acceptor protein that are present within close enough proximity for energy transfer to occur (James, Oliviera, Carmo, Iaboni, & Davis, 2006). Unlike in FRET, the donor protein used in BRET assays is a form of the Renilla luciferase protein that only emits blue light upon addition of its substrate, coelenterazine (James et al., 2006). Therefore, energy transfer to the acceptor protein can only occur upon addition of coelenterazine when it is quantified within the assay. BRET is advantageous because it can be used to examine kinetic effects, it is highly sensitive, adaptable to high-throughput screening, and is adaptable to measure any protein–protein interaction aside from GPCRs and β-arrestin (James et al., 2006). However, effective use of BRET requires the use of rigorous controls including BRET positive (donor–acceptor fusion proteins), BRET negative (donor protein alone), noninteracting protein controls, and determination of appropriate donor: acceptor ratios, among others (Hudson et al., 2010; James et al., 2006; Laprairie et al., 2016). Cloned BRET constructs must be in the correct orientation to facilitate energy transfer but not disrupt normal protein function (James et al., 2006), and BRET cannot be conducted in an endogenous system because it requires the expression of heterologous fusion proteins (James et al., 2006).

Enzyme fragment complementation assays utilize CB1R modified with a β-galactosidase fragment coexpressed in the same cell as β-arrestin conjugated to the remainder of β-galactosidase (van der Lee et al., 2009). Recruitment of β-arrestin produces a complete enzyme whose activity can be quantified by addition of chemiluminscent substrate, such that the amount of chemiluminscence produced is proportional to β-arrestin recruitment. This popular assay format is utilized in the PathHunter® assays commercially available from DiscoveRx and others (van der Lee et al., 2009). Enzyme complementation assays are advantageous because they are highly sensitive and produce large signal over background and are designed for high-throughput screening (van der Lee et al., 2009). Two disadvantages of these assays are (1) that they are irreversible upon addition of the substrate and are therefore not suitable for quantifying kinetic effects, and (2) it requires modification of the C-terminus of the receptor and cannot be used to quantify β-arrestin recruitment in endogenous systems.

The Tango assay utilizes transcriptional activation following β-arrestin translocation (Barnea et al., 2007). GPCRs of interest are fused to a transcription factor by a peptide containing a TEV protease cleavage sequence and β-arrestin is fused to the TEV protease (Barnea et al., 2007; Kroeze et al., 2015). Association of the GPCR with β-arrestin results in proteolytic cleavage of the transcription factor, which consequently activates transcription of a reporter gene, such as luciferase (Barnea et al., 2007; Kroeze et al., 2015).

As a result, a transient interaction between GPCR and β-arrestin is amplified and stabilized. Tango assay systems are now openly available for a wide swath of GPCRs, making this system amenable to screening compounds at many different receptor systems (Kroeze et al., 2015). As with enzyme complementation assays, Tango is advantageous because of its sensitivity and large signal over background. However, modification of the both β-arrestin and GPCR means that it cannot be used to quantify β-arrestin recruitment in endogenous systems.

3.3. Inhibition of cAMP Accumulation

Beyond proximal interactions between CB1R and G protein, and CB1R and β-arrestin, Gαi/o proteins inhibit adenylate cyclases and reduce the concentration of cAMP. Measurements of cAMP can be made from cultured cell systems (Cawston, Connor, Di Marzo, Silvestri, & Glass, 2015; Hua et al., 2016; Khajehali et al., 2015) and dissociated cell membranes from animals (Brust et al., 2017), using the many quantification assays currently available. The earliest studies of the CB1R modulation of adenylyl cyclase were performed by measuring the conversion of α-[32P]ATP to [32P]cAMP (Meschler & Howlett, 2001).

For higher throughput studies avoiding radioactivity, assay kits based on fluorophores have become available. Homogeneous time-resolved fluorescence resonance energy transfer (FRET) (HTRF) can be used to measure changes in cAMP level (Cawston et al., 2015; Hua et al., 2016). The cAMP HTRF assay utilizes an anti-cAMP antibody conjugated to a donor fluorophore and a fluorescent acceptor molecule conjugated to exogenous cAMP (Degorce et al., 2009). Excitation of the bound donor-conjugated antibody results in emission of fluorescence that then excites the acceptor molecule to emit a different wavelength of light (Degorce et al., 2009). The ratio of acceptor to donor fluorescence is FRET (Degorce et al., 2009). HTRF uses a delay between quantification of donor emission and acceptor emission (~150 μs) to enhance resolution between the two protein’s spectra, which reduces background signal allows for quantification (Degorce et al., 2009). Endogenous cAMP produced by the cell/tissue being measured competes with the acceptor molecule for antibody binding, and so FRET is inversely proportional to cAMP concentration (Degorce et al., 2009). HTRF is advantageous over other assay systems because it is amenable to measurements of cAMP accumulation (GαS signaling at CB1R; Glass & Northup, 1999) and inhibition of cAMP (Gαi/o signaling at CB1R CB1R; Mukhopadhyay & Howlett, 2005), it can be used to examine kinetic effects of cAMP modulation (Cawston et al., 2015), it is highly sensitive (Degorce et al., 2009), and assay systems are relatively easy to use. Although quantification of cAMP is less proximal to the receptor than G protein coupling, cAMP represents a major cellular effector of CB1R and is therefore important to measure.

Other approaches to measure cAMP include enzyme-linked immunosorbent assays (EIA) that utilize fluorophore-conjugated anti-cAMP anti-bodies (Jiang et al., 2007), similar to HTRF; BRET-based assays that utilize the fusion protein “cAMP sensor-using YFP-Epac-Rluc” (Cawston et al., 2015; Jiang et al., 2007); transcriptional activation in cells that express a reporter gene (e.g., luciferase) under the regulatory control of cAMP-response elements (Zhaung & Liu, 2006). EIA is similar to HTRF but is difficult to use to examine kinetic effects (Jiang et al., 2007). BRET-based quantification of cAMP is highly sensitive and amenable to kinetic experiments, but must be done in heterologous systems (Cawston et al., 2015; Jiang et al., 2007). Transcriptional assays are irreversible and are cannot be used quantifying kinetic effects, and must be conducted in heterologous systems, but provide a high signal relative to background (Zhaung & Liu, 2006). cAMP signaling is discussed further in Chapter 3 of this volume (“Real-Time Measurement of Cannabinoid Receptor-Mediated cAMP Signaling” by Hunter et al.).

3.4. Changes in Ca2+

In addition to Gαi/o, CB1R couples Gαq/11 to modulate intracellular Ca2+ levels (Lauckner et al., 2005). Ca2+ modulation may be monitored through electrophysiological approaches (Lauckner et al., 2005) or fluorescence-based approaches (Vandevoorde, Lambert, Smart, Jonsson, & Fowler, 2003). Electrophysiological techniques provide for high spatiotemporal resolution and can be easily utilized in in vivo and ex vivo neuronal preparations; however, they are not amenable to the screening of large numbers of test compounds (Lauckner et al., 2005). The FLIPR Ca2+ detection assays utilize a dye that fluoresces upon binding Ca2+ (Sirenko et al., 2013). FLIPR was specifically designed for high-throughput applications and provides some degree of spatiotemporal resolution (Sirenko et al., 2013). Due to the high cost associated with this assay, it is not a feasible approach for ex vivo neuron preparations, but can be used for cell systems whether they endogenously or heterologously express CB1R.

3.5. Receptor Internalization

Following ligand binding, CB1R is internalized through β-arrestin-mediated mechanisms and subsequently recycled or degraded. Internalization of the unmodified receptor can be detected with a fluorescent protein conjugate (e.g., β-arrestin-green fluorescent protein (GFP)) or fluorescent-conjugated antibody. Internalized receptors associated with β-arrestin-GFP are visible within the cell as puncta. The number of puncta per cell is related to the concentration of compound used in a concentration-dependent manner. To conduct this assay, cells are grown in glass-bottomed plates, treated with test compound and images of cells are captured and analyzed for the number of puncta/cell (Gasparri, 2009). This assay is advantageous because kinetics can be determined easily, they can produce large signal over background, and they are amenable to high-content analysis and compound screening (Gasparri, 2009). Because this assay relies on high-content image analysis, it provides qualitative information on the movement and localization of receptors following compound treatment. Other systems have been generated to quantify receptor internalization such as the PathHunter® assays (DiscoveRx) which can be useful if imaging systems are not available. In these assays, the enzyme fragment is localized to the endosome and so complementation and signal detection occurs at the endosome upon receptor internalization (van der Lee et al., 2009). These assays have similar advantages and disadvantages as the PathHunter® β-arrestin recruitment assays (van der Lee et al., 2009).

3.6. Protein Phosphorylation and Changes in Gene Expression

CB1R activation effects the phosphorylation and activation of signaling proteins, including ERKs, protein kinase B (Akt), activator protein 1 (AP-1), and others, downstream of changes in cAMP or Ca2+. Protein phosphorylation can be measured through standard western blot techniques using cell lysates or tissue homogenates (Bouaboula et al., 1995; Howlett et al., 2000). Although western blots are widely used, can be conducted using native tissues or nontransfected cell cultures, and an effective means of measuring protein, they are not amenable to high-throughput screening. Protein phosphorylation can also be measured by the In-cell western assay (Hudson et al., 2010; Laprairie et al., 2016, 2017; Redmond et al., 2016). In an In-cell western, cells are cultured in multiwell plates and treated with the compounds of interest. Cells are subsequently fixed, permeablized, and incubated with antibody directed against the proteins of interest (phosphoprotein and total protein), followed by a second incubation with infrared-conjugated secondary antibodies (e.g., green for phosphoprotein and red for total protein). Immunocytochemical staining is quantified using an Odyssey imaging system (Li-Cor), and protein phosphorylation is expressed relative to total protein (Hudson et al., 2010; Laprairie et al., 2016, 2017; Redmond et al., 2016). In-cell western assays can be adapted for use as a medium-throughput screen in heterologous and endogenous receptor expression systems. However, antibody conditions and specificity must be verified through standard western blot approaches before being applied in this assay.

Several groups have demonstrated that CB1R activation produces quantifiable changes in transcription factor activity (Bosier, Hermans, & Lambert, 2008; Bouaboula et al., 1995) and gene expression (Bosier et al., 2008; Laprairie et al., 2016), which can be measured by chromatin immunoprecipitation and quantitative reverse transcriptase-PCR (qRT-PCR), respectively. As with immunocytochemical approaches, quantification of transcription factor activity and qRT-PCR is amenable to medium- to high-throughput assay development and can be applied to both heterologous and endogenous receptor systems. However, the interpretation of agonist bias several steps removed from the receptor is difficult, as discussed later. Therefore, data from these assays may be physiologically relevant to the overall actions of the drug, but less relevant with regard to elucidating mechanisms of bias (Kenakin et al., 2012). Furthermore, extensive computational power may be necessary for managing and interpreting extensive datasets that can arise from large-scale transcriptome evaluation studies (Luttrell et al., 2015).

4. IMPORTANT CONSIDERATIONS AND CAVEATS

Bias is a relative term, requiring the comparison of one effect to another. As such, ligand bias describes the ability of an agonist to perform differently compared to another agonist when both agonists are examined in parallel across multiple assays. As such, some degree of bias may appear in nearly every direction examined for nearly every ligand. The key is to discern what is truly the anomaly and whether such divergences from the normality will correspond to fundamental differences in receptor function. In this regard, advances will be made as we begin to understand more about the active-, inactive-, and differently active-state structures that GPCRs assume when bound to different ligands. While the number of structures for GPCRs explode and the resolution of refined binding poses comes further into focus, we can foresee a time of structure-guided ligand design. However, the question we will still need to address is what kind of bias do we want to impart in our compounds to control what physiologies in vivo? Therefore, the pursuit of compounds with diverse pharmacological signatures will be useful for starting to answer that question in animal models. Meanwhile, to facilitate ranking our compounds and defining the differences by which they act, quantitative studies of CB1R signaling should include analyses of bias, as has been done for other GPCRs (Luttrell, 2014; Rajagopal et al., 2011).

4.1. Cell Type and Receptor Density

Unlike other approaches to quantify ligand bias, such as comparison of pEC50 or relative activity, the transducer coefficient used to determine bias (τ/KA) is a linear parameter in relation to receptor density (Kenakin et al., 2012). This is critical because it allows for comparisons to be made between most agonists as well as between tissue and cell types, which may be used to conduct various assays (Kenakin et al., 2012). Therefore, the operational model as first described by Black and Leff (1983) (see earlier) can be used in most cases so long as the compounds being studied produce an appreciable response in the assay used. If a compound(s) does not produce an appreciable response in all assays, then an operational-competitive model can be applied to determine bias (Stahl et al., 2015). The competitive model can be applied by taking advantage of the competitive nature of partial agonists in the presence of a full agonist (Ehlert, 2015; Stahl et al., 2015). This method was developed and is described in detail by Stahl et al. (2015).

In all cases, the quality of the concentration-response curves used to assess bias is critically important. Compound solubility, preparation of compound plates, and evaporation may all effect the observed response and consequently the calculated bias of a compound. For cannabinoids, as highly lipophilic compounds, the use of a carrier protein such as BSA is important because solubility can be improved and compounds do not stick to vessel walls (Howlett et al., 2002). Compound plates should be set up immediately prior to use to limit compound degradation and solvent evaporation. For live cell assays, cell number and serum deprivation should be optimized and considered across assay platforms; and protein concentration should be optimized for membrane-based assays. Early protocol optimization will improve the resolution of the assays and produce reliable concentration-response curves that can be used to determine bias.

4.2. Proximal vs Distal Effects to the Receptor

GPCR agonists facilitate cellular effects upon binding to their receptor and promoting a conformational change in that population of receptors that is conducive to signaling (Kenakin & Christopoulos, 2013). In the case of “balanced” agonists, the net effect is that of pluripotent signaling through the available repertoire of effector proteins (Kenakin et al., 2012; Luttrell et al., 2015). For biased agonists, however, a specific conformation of the receptor is favored among the ensemble of conformations and it is this shift toward a specific conformation that results in preferential signaling through specific effector proteins and therefore a biased receptor population (Luttrell et al., 2015). As a consequence of this, bias-driven effects that occur at effectors proximal to the receptor (e.g., G protein-coupling and β-arrestin recruitment) are likely representative of a change in receptor conformation producing a cellular change (Kenakin et al., 2012; Luttrell et al., 2015). As a specific signal is propagated from the GPCR to other proteins that signal is amplified.

Therefore, agonist bias becomes more pronounced at secondary and tertiary effector proteins (e.g., adenylate cyclase or G protein-coupled inwardly rectifying K+ channels) relative to primary effectors (Kenakin & Christopoulos, 2013; Kenakin et al., 2012). However, cross talk between signals results in overlap between different pathways and signal bias may be misinterpreted as the consequence of one pathway rather than another (Luttrell et al., 2015). This is particularly true in the case of ERK and Akt, whose activation is a convergence point for multiple CB1R signaling pathways including Gαi/o and β-arrestin (Khajehali et al., 2015; Laprairie et al., 2016). Another example is transcription factors such as AP-1 and Jnk that are activated through multiple CB1R-mediated signaling pathways (Bosier et al., 2008). Pathway-specific effects can be determined in such cases through careful examination of signaling kinetics (Herenbrink et al., 2016) as well as ligand potencies in different assays (Kenakin et al., 2012). Ultimately, agonist bias must be quantified such that the regulatory pathway is known and the quantified signals are amenable to bias analysis.

4.3. Ligand Kinetics

Increasing evidence from studies of other GPCRs, such as the dopamine D2 receptor, supports the hypothesis that an agonist’s dissociation rate has a pro-found effect on its observed bias (Herenbrink et al., 2016). Herenbrink et al. (2016) observed that the apparent bias of D2 agonists varied widely depending on the time at which assay measurements were made. The authors were able to relate these differences in bias to the dissociation rate of agonists and concluded that slowly dissociating agonists were more likely to display β-arrestin bias that rapidly dissociating agonists (Herenbrink et al., 2016). Although this has yet to be directly demonstrated for agonists of CB1R, Cawston et al. (2015) observed that changes in signal transduction following treatment with the CB1R allosteric modulator Org27569 changed dramatically as a function of time. Khajehali et al. (2015) reported that high-affinity CB1R agonists CP55,940 and HU-210 displayed bias for cAMP relative to pERK but did not relate these effects to ligand dissociation. The role that ligand kinetics play in bias is important to understand for two reasons: (1) it represents a possible source for discrepancy between model systems and assays in interpreting bias, and (2) it may be that biased ligands can be designed for CB1R on the basis of their dissociation rate. With respect to future studies of CB1R agonist bias, the contribution of kinetics to bias must be considered when designing experiments and interpreting data (Cawston et al., 2015; Herenbrink et al., 2016).

4.4. Applying Measurements of Bias to Allosteric Ligands

In addition to biased CB1R orthosteric agonists, allosteric modulators of CB1R can alter the signaling bias of orthosteric CB1R agonists (Khajehali et al., 2015; Laprairie et al., 2017). Khajehali et al. (2015) suggest that the negative allosteric modulatory activity of Org27569 for pERK1/2 may be due to biased allosterism and probe dependence. Laprairie et al. (2017) observed that the allosteric compound GAT228 is moderately biased for β-arrestin1, whereas its enantiomer GAT229 is biased for inhibition of cAMP. Given the growing interest in developing allosteric modulators of CB1R, evaluations of their potential bias will become increasingly important, and with it the need to establish reliable analytical approaches.

4.5. Crystal Structures of CB1R

To date two crystal structures of CB1R in inactive conformations have been described (Hua et al., 2016; Shao et al., 2016). Hua et al. (2016) first crystallized CB1R bound to the rimonabant (SR141716A) derivative, and irreversible CB1R antagonist, AM6538. The development of irreversible antagonists and subsequent crystallization of hCB1R is described in Chapter 10 of this volume (“Ligand-Assisted Protein Structure (LAPS): An Experimental Paradigm for Characterizing Cannabinoid-Receptor Ligand-Binding Domains” by Janero et al.). From these first glimpses into the structure of CB1R we can begin to understand the ligand binding landscape and form hypotheses about residues within the agonist binding pocket that may be important for mediating changes in receptor conformation and consequently bias. The next hurdle to overcome in this field will be to obtain a crystal structure of CB1R in an active, agonist-bound state. If agonist-bound crystals were obtained using differentially biased agonists, it may even be possible to observe conformational differences in receptor related to bias.

5. CONCLUSIONS

CB1R is considered an important pharmacological target because modulation of this receptor produces wide-ranging physiological effects—appetite, locomotion, mood, analgesia—that from a clinical perspective appear to be applicable to many different pathological conditions. At the receptor level, CB1R couples to multiple effector proteins and facilitates a myriad of intra- and intercellular effects. The current state of CB1R research, as described here, indicates that biased ligands can selectively enhance certain receptor-mediated effects, to a limited degree. The present questions for studying ligand bias at CB1R are (1) what is the possible extent of ligand bias for the receptor? and (2) does ligand bias facilitate physiologically relevant effects that could be therapeutically useful? The application of tools, such as the operational model, receptor crystal structures, irreversible antagonists, and a better understanding of kinetic and allosteric effects, previously not applied to CB1R or not available, will permit a more thorough understanding of CB1R pharmacology than was previously possible. From these tools we can begin to understand the importance of CB1R ligand bias.

ACKNOWLEDGMENT

This work was supported by National Institutes of Health Grant P01DA009158 (L.M.B.). R.B.L. is supported in part by a postdoctoral fellowship from the Canadian Institutes of Health Research.

ABBREVIATIONS

2-AG

2-arachidonoylglycerol

AEA

anandamide

Akt

protein kinase B

AP-1

activator protein 1

BRET

bioluminescence resonance energy transfer

cAMP

cyclic adenosine monophosphate

CB1R

type 1 cannabinoid receptor

CB2R

type 2 cannabinoid receptor

EC50

half maximal effective concentration

Emax

maximum response achievable from an applied drug

ERK

extracellular-regulated kinase

FRET

Förster resonance energy transfer

Gα

alpha subunit of G protein

GFP

green fluorescent protein

GPCR

G protein-coupled receptors

GTP

guanosine triphosphate

GTPγS

guanosine 5′-O-[gamma-thio]triphosphate

qRT-PCR

quantitative reverse transcriptase-polymerase chain reaction

REFERENCES

  1. Barnea G, Strapps W, Herrada G, Berman Y, Ong J, Kloss B, et al. (2007). The genetic design of signaling cascades to record receptor activation. Proceedings of the National Academy of Sciences of the United States of America, 105, 64–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bayewitch M, Avidor-Reiss T, Levy R, Barg J, Mechoulam R, & Vogel Z (1995). The peripheral cannabinoid receptor: Adenylate cyclase inhibition and G protein coupling. FEBS Letters, 375, 143–147. [DOI] [PubMed] [Google Scholar]
  3. Black JW, & Leff P (1983). Operational models of pharmacological agonism. Proceedings of the Royal Academy of Sciences London B, Biological Sciences, 220, 141–162. [DOI] [PubMed] [Google Scholar]
  4. Bohn LM, Zhou L, & Ho J-H (2015). Approaches to assess functional selectivity in GPCRs: Evaluating G protein signaling in an endogenous environment. Methods in Molecular Biology, 1335, 177–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bosier B, Hermans E, & Lambert D (2008). Differential modulation of AP-1- and CRE-driven transcription by cannabinoid agonists emphasizes functional selectivity at the CB1 receptor. British Journal of Pharmacology, 155, 24–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bouaboula M, Poinot-Chazel C, Bourrié B, Canat X, Calandra B, Rinaldi-Carmona M, et al. (1995). Activation of mitogen-activated protein kinases by stimulation of the central cannabinoid receptor CB1. Biochemical Journal, 312, 637–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Breivogel CS, Lambert JM, Gerfin S, Huffman JW, & Razdan RK (2008). Sensitivity to delta9-tetrahydrocannabinol is selectively enhanced in beta-arrestin2−/− mice. Behavioral Pharmacology, 19, 298–307. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brust TF, Alongkronrusmee D, Soto-Velasquez M, Baldwin TA, Ye Z, Dai M, et al. (2017). Identification of a selective small-molecule inhibitor of type 1 adenylyl cyclase activity with analgesic properties. Science Signaling, 10, eaah5381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Brust TF, Morgenweck J, Kim SA, Rose JH, Locke JL, Schmid CL, et al. (2016). Biased agonists of the kappa opioid receptor suppress pain and itch without causing sedation or dysphoria. Science Signaling, 9, ra117. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Cawston EE, Connor M, Di Marzo V, Silvestri R, & Glass M (2015). Distinct temporal fingerprint for cyclic adenosine monophosphate (cAMP) signaling of indole-2-carboxamides as allosteric modulators of the cannabinoid receptors. Journal of Medicinal Chemistry, 58, 5979–5988. [DOI] [PubMed] [Google Scholar]
  11. Christensen GL, Kelstrup CD, Lyngsø C, Sarwar U, Bøgebo R, Sheikh SP, et al. (2010). Quantitative phosphoproteomics dissection of seven-transmembrane receptor signaling using full and biased agonists. Molecular and Cellular Proteomics, 9, 1540–1553. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Christopoulos A (2014). Advances in G protein-coupled receptor allostery: From function to structure. Molecular Pharmacology, 86, 463–478. [DOI] [PubMed] [Google Scholar]
  13. Clarke WP, & Berg KA (2010). Use of functional assays to detect and quantify functional selectivity. Drug Discovery Today, 7, e31–e36. [DOI] [PubMed] [Google Scholar]
  14. Degorce F, Card A, Soh S, Tringuet E, Knapik GP, & Xie B (2009). HTRF: A technology tailored for drug discovery—A review of theoretical aspects and recent applications. Current Chemical Genomics, 3, 22–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Delgado-Peraza F, Ahn KH, Nogueras-Ortiz C, Mungrue IN, Mackie K, Kendall DA, et al. (2016). Mechanisms of biased β-arrestin-mediated signaling downstream from the cannabinoid 1 receptor. Molecular Pharmacology, 89, 618–629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. DeWire SM, Yamashita DS, Rominger DH, Liu G, Cowan CL, Graczyk TM, et al. (2013). A G protein-biased ligand at the μ-opioid receptor is potently analgesic with reduced gastrointestinal and respiratory dysfunction compared with morphine. Journal of Pharmacology and Experimental Therapeutics, 344, 708–717. [DOI] [PubMed] [Google Scholar]
  17. Ehlert FJ (2015). Functional studies cast light on receptor states. Trends in Pharmacological Science, 36, 596–604. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Felder CC, Joyce KE, Briley EM, Mansouri J, Mackie K, Blond O, et al. (1995). Comparison of the pharmacology and signal transduction of the human cannabinoid CB1 and CB2 receptors. Molecular Pharmacology, 48, 443–450. [PubMed] [Google Scholar]
  19. Gasparri F (2009). An overview of cell types in HCS: Limitations and advantages. Expert Opinion on Drug Discovery, 4, 643–657. [DOI] [PubMed] [Google Scholar]
  20. Glass M, & Northup JK (1999). Agonist selective regulation of G proteins by cannabinoid CB1 and CB2 receptors. Molecular Pharmacology, 56, 1362–1369. [DOI] [PubMed] [Google Scholar]
  21. Grim TW, Morales AJ, Gonek MM, Wiley JL, Thomas BF, Endres GW, et al. (2016). Stratification of cannabinoid 1 receptor (CB1R) agonist efficacy: Manipulation of CB1R density through use of transgenic mice reveals congruence between in vivo and in vitro assays. Journal of Pharmacology and Experimental Therapeutics, 359, 329–339. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Herenbrink KC, Sykes DA, Donthamsetti P, Canals M, Coudrat T, Shonberg J, et al. (2016). The role of kinetic context in apparent biased agonism at GPCRs. Nature Communications, 7, 10842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hinton DR, Blanks JC, Fong HK, Casey PJ, Hildebrandt E, & Simons MI (1990). Novel localization of a G protein, Gz-alpha, in neurons of brain and retina. Journal of Neuroscience, 10, 2763–2770. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Howlett AC, Barth F, Bonner TI, Carbal G, Casellas P, Devane WA, et al. (2002). International Union of Pharmacology. XXVII. Classification of cannabinoid receptors. Pharmacological Reviews, 54, 161–202. [DOI] [PubMed] [Google Scholar]
  25. Howlett AC, Wilken GH, Pigg JJ, Houston DB, Lan R, Liu Q, et al. (2000). Azido- and isothiocyanato-substituted aryl pyrazoles bind covalently to the CB1 cannabinoid receptor and impair signal transduction. Journal of Neurochemistry, 74, 2174–2181. [DOI] [PubMed] [Google Scholar]
  26. Hua T, Vemuri K, Pu M, Qu L, Han GW, Wu Y, et al. (2016). Crystal structure of the human cannabinoid receptor CB1. Cell, 167, 750–762. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Hudson BD, Hébert TE, & Kelly ME (2010). Ligand- and heterodimer-directed signaling of the CB(1) cannabinoid receptor. Molecular Pharmacology, 77, 1–9. [DOI] [PubMed] [Google Scholar]
  28. Izzo AA, & Coutts AA (2005). Cannabinoids and the digestive tract. Handbook of Experimental Pharmacology, 2005, 573–598. [DOI] [PubMed] [Google Scholar]
  29. James JR, Oliviera MI, Carmo AM, Iaboni A, & Davis SJ (2006). A rigorous experimental framework for detecting protein oligomerization using bioluminescence resonance energy transfer. Nature Methods, 3, 1001–1006. [DOI] [PubMed] [Google Scholar]
  30. Janero DR, Yaddanapudi S, Zvonok N, Subramanian KV, Shukla VG, Stahl E, et al. (2015). Molecular-interaction and signaling profiles of AM3677, a novel covalent agonist selective for the cannabinoid 1 receptor. ACS Chemical Neuroscience, 19, 1400–1410. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jiang LI, Collins J, Davis R, Lin K-M, DeCamp D, Roach T, et al. (2007). Use of a cAMP BRET sensor to characterize a novel regulation of cAMP by the sphingosine 1-phosphate/G13 pathway. Journal of Biological Chemistry, 282, 10576–10584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kenakin T (2010). Functional selectivity and biased receptor signaling. Journal of Pharmacology and Experimental Therapeutics, 336, 296–302. [DOI] [PubMed] [Google Scholar]
  33. Kenakin T, & Christopoulos A (2013). Signalling bias in new drug discovery: Detection, quantification and therapeutic impact. Nature Reviews. Drug Discovery, 12, 205–216. [DOI] [PubMed] [Google Scholar]
  34. Kenakin T, Watson C, Muniz-Medina V, Christopoulos A, & Novick S (2012). A simple method for quantifying functional selectivity and agonist bias. ACS Chemical Neuroscience, 3, 193–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Khajehali E, Malone DT, Glass M, Sexton PM, Christopoulos A, & Leach K (2015). Biased agonism and biased allosteric modulation at the CB1 cannabinoid receptor. Molecular Pharmacology, 88, 368–379. [DOI] [PubMed] [Google Scholar]
  36. Kroeze WK, Sassano MF, Huang X-P, Lansu K, McCorvy JD, Giguére PM, et al. (2015). PRESTO-Tango as an open-source resource for interrogation of the druggable human GPCRome. Nature Structural and Molecular Biology, 22, 362–369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Laprairie RB, Bagher AM, Kelly ME, & Denovan-Wright EM (2016). Biased type 1 cannabinoid receptor signaling influences neuronal viability in a cell culture model of Huntington disease. Molecular Pharmacology, 89, 364–375. [DOI] [PubMed] [Google Scholar]
  38. Laprairie RB, Kulkarni PM, Deschamps JR, Kelly ME, Janero DR, Cascio MG, et al. (2017). Enantiospecific allosteric modulation of cannabinoid 1 receptor. ACS Chemical Neuroscience, 8, 1188–1203. [DOI] [PubMed] [Google Scholar]
  39. Lauckner JE, Hille B, & Mackie K (2005). The cannabinoid agonist WIN55,212-2 increases intracellular calcium via CB1 receptor coupling to Gq/11 G proteins. Proceedings of the National Academy of Sciences of the United States of America, 102, 19144–19149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Luttrell LM (2014). Minireview: More than just a hammer: Ligand “bias” and pharmaceutical discovery. Molecular Endocrinology, 28, 281–294. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Luttrell LM, Maudsley S, & Bohn LM (2015). Fulfilling the promise of “biased” G protein-coupled receptor agonism. Molecular Pharmacology, 88, 579–588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Meschler JP, & Howlett AC (2001). Signal transduction interactions between CB1 cannabinoid and dopamine receptors in the rat and monkey striatum. Neuropharmacology, 40, 918–926. [DOI] [PubMed] [Google Scholar]
  43. Motaghedi R, & McGraw TE (2008). The CB1 endocannabinoid system modulates adipocyte insulin sensitivity. Obesity, 16, 1727–1734. [DOI] [PubMed] [Google Scholar]
  44. Mukhopadhyay S, & Howlett AC (2005). Chemically distinct ligands promote differential CB1 cannabinoid receptor-Gi protein interactions. Molecular Pharmacology, 67, 2016–2024. [DOI] [PubMed] [Google Scholar]
  45. Nguyen PT, Schmid CL, Raehal KM, Selley DE, Bohn LM, & Sim-Selley LJ (2012). Beta-arrestin-2 regulates cannabinoid CB1 receptor signaling and adaptation in a CNS region-dependent manner. Biological Psychiatry, 71, 714–724. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Pertwee RG (2008). Ligands that target cannabinoid receptors in the brain: From THC to anandamide and beyond. Addiction Biology, 13, 147–159. [DOI] [PubMed] [Google Scholar]
  47. Pinto JC, Potié F, Rice KC, Boring D, Johnson MR, Evans DM, et al. (1994). Cannabinoid receptor binding and agonist activity of amides and esters of arachidonic acid. Molecular Pharmacology, 46, 516–522. [PubMed] [Google Scholar]
  48. Rajagopal S, Ahn S, Riminger DH, Gowen-MacDonald W, Lam CM, DeWire SM, et al. (2011). Quantifying ligand bias at seven-transmembrane receptors. Molecular Pharmacology, 80, 367–377. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Ranieri R, Laezza C, Bifulco M, Marasco D, & Malfitano AM (2016). Endo-cannabinoid system in neurological disorders. Recent Patents on CNS Drug Discovery, 10, 90–112. [DOI] [PubMed] [Google Scholar]
  50. Rankovic Z, Brust TF, & Bohn LM (2016). Biased agonism: An emerging paradigm in GPCR drug discovery. Bioorganic and Medicinal Chemistry Letters, 26, 241–250. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Redmond WJ, Cawston EE, Grimsey NL, Stuart J, Edington AR, Glass M, et al. (2016). Identification of N-arachidonoyl dopamine as a highly biased ligand at cannabinoid CB1 receptors. British Journal of Pharmacology, 173, 115–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Shao Z, Yin J, Chapman K, Grzemnska M, Clark L, Wang J, et al. (2016). High-resolution crystal structure of the human CB1 cannabinoid receptor. Nature, 540, 602–606. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Sirenko O, Cromwell EF, Crittenden C, Wignall JA, Wright FA, & Rusyn I (2013). Assessment of beating parameters in human induced pluripotent stem cells enables quantitative In Vitro screening for cardiotoxicity. Toxicology and Applied Pharmacology, 273, 500–507. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Soergel DG, Subach RA, Burnham N, Lark MW, James IE, Sadler BM, et al. (2014). Biased agonism of the μ-opioid receptor by TRV130 increases analgesia and reduces on-target adverse effects versus morphine: A randomized, double-blind, placebo-controlled, crossover study in healthy volunteers. Pain, 155, 1829–1835. [DOI] [PubMed] [Google Scholar]
  55. Stahl EL, Zhou L, Ehlert FJ, & Bohn LM (2015). A novel method for analyzing extremely biased agonism at G protein-coupled receptors. Molecular Pharmacology, 87, 866–877. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. van der Lee MMC, Blomenröhr M, van der Doelen AA, Wat JWY, Smits N, Hanson BJ, et al. (2009). Pharmacological characterization of receptor redistribution and β-arrestin recruitment assays for the cannabinoid receptor 1. Journal of Biomolecular Screening, 14, 811–823. [DOI] [PubMed] [Google Scholar]
  57. Vandevoorde S, Lambert DM, Smart D, Jonsson KO, & Fowler CJ (2003). N-Morpholino- and N-diethyl-analogues of palmitoylethanolamide increase the sensitivity of transfected human vanilloid receptors to activation by anandamide without affecting fatty acid amidohydrolase activity. Bioorganic and Medicinal Chemistry, 20, 817–825. [DOI] [PubMed] [Google Scholar]
  58. Wettschureck N, Moers A, Wallenwein B, Parlow AF, Maser-Gluth C, & Offermanns S (2005). Loss of Gq/11 family G proteins in the nervous system causes pituitary somatotroph hypoplasia and dwarfism in mice. Molecular and Cellular Biology, 25, 1942–1948. [DOI] [PMC free article] [PubMed] [Google Scholar]
  59. Wiley JL, Smith FL, Razdan RK, & Dewey WL (2005). Task specificity of cross-tolerance between delta9-tetrahydrocannabinol and anandamide analogs in mice. European Journal of Pharmacology, 510, 59–68. [DOI] [PubMed] [Google Scholar]
  60. Zhaung F, & Liu YH (2006). Usefulness of the luciferase reporter system to test the efficacy of siRNA. Methods in Molecular Biology, 342, 181–187. [DOI] [PubMed] [Google Scholar]
  61. Zhou L, & Bohn LM (2014). Functional selectivity of GPCR signaling in animals. Current Opinion in Cell Biology, 27, 102–108. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES