Location Bias: A “Hidden Variable” in GPCR Pharmacology

SUMMARY:

G protein-coupled receptors (GPCRs) are the largest family of transmembrane receptors and primarily signal through two main effector proteins: G proteins and β-arrestins. Many agonists of GPCRs promote “biased” responses, in which different cellular signaling pathways are activated with varying efficacies. The mechanisms underlying biased signaling have not been fully elucidated, with many potential “hidden variables” that regulate this behavior. One contributor is “location bias,” which refers to the generation of unique signaling cascades from a given GPCR depending upon the cellular location at which the receptor is signaling. Here, we review evidence that GPCRs are expressed at and traffic to various subcellular locations and discuss how location bias can impact the pharmacologic properties and characterization of GPCR agonists. We also evaluate how differences in subcellular environments can modulate GPCR signaling, highlight the physiological significance of subcellular GPCR signaling, and discuss the therapeutic potential of exploiting GPCR location bias.

Graphical Abstract

INTRODUCTION

G protein-coupled receptors (GPCRs) are the largest superfamily of receptors and share an architecture of seven transmembrane α-helices connected by three extracellular and three intracellular loops. They serve as critical nodes of communication between the extracellular and intracellular environments by responding to a diverse array of stimuli, including hormones, neurotransmitters, lipids, photons, small molecules, and mechanical force [1]. Due to their involvement in a multitude of physiological processes, GPCRs are the targets of over 30% of all Food and Drug Administration (FDA)-approved drugs [2].

GPCRs share a common activation mechanism. When a ligand binds in the receptor’s orthosteric site, the receptor undergoes a conformational change, leading to the outward movement of the transmembrane helices and rearrangement of several microswitches in the transmembrane region [3]. These conformational changes create a crevice on the cytoplasmic side of the receptor, allowing for engagement with intracellular transducers like heterotrimeric G proteins, β-arrestins, and GPCR kinases (GRKs). Upon engagement with heterotrimeric G protein, the receptor catalyzes the guanine nucleotide exchange of GTP for the bound GDP of the Gα subunit [3,4]. The GTP-bound Ga subunit dissociates from the Gβγ subunits and subsequently regulates effector proteins like adenylyl cyclase and phospholipase C, ultimately modulating the levels of second messenger effectors such as cyclic AMP (cAMP), diacylglycerol (DAG) and inositol 1,4,5-trisphosphate (IP3). These changes lead to the activation of downstream signaling pathways such as the serine/threonine kinase Akt, protein kinase A (PKA), and mitogen-activated protein kinases (MAPKs).

Following activation of G protein signaling, GRKs are recruited to the receptor to phosphorylate the intracellular loops and C-terminal tail of the receptor, which promotes the recruitment of β-arrestins [5,6]. β-arrestins can ‘arrest’ G protein signaling through steric hindrance by binding to the cytoplasmic transducer-binding pocket, and can further inhibit signaling by scaffolding components of the endocytic machinery, such as clathrin and AP2, to sequester receptors into endosomes and transport them to proteosomes or lysosomes for degradation [7-9]. The role of β-arrestins as multifunctional regulators of GPCR signaling was appreciated after the discovery that they can activate signaling pathways independent of G proteins, regulate a diverse array of cellular processes such as MAPK, Akt, the nonreceptor tyrosine kinase Src, and transactivate receptor tyrosine kinases [10].

Compared to a reference agonist, which activates both signaling pathways mediated by both G proteins and β-arrestins, certain ligands have been discovered that preferentially target one transducer pathway over the other, a phenomenon termed ‘biased agonism’ [11]. “Biased agonists” can exist endogenously, such as in the chemokine system, where multiple ligands target the same receptor and differentially activate G proteins and β-arrestins [12]. A variety of synthetic small molecules and peptide analogs that similarly demonstrate biased GPCR signaling also exist. These biased agonists have important clinical and pharmaceutical implications and have been actively pursued as a potential avenue to develop specific drugs that activate beneficial pathways while inhibiting those that lead to adverse effects.

Despite the importance of biased agonism in drug discovery, the mechanisms underlying such signaling have yet to be fully elucidated. Like in quantum mechanics, where some theoreticians have posited the presence of “hidden variables” propounded to underlie the complex behavior of observables [13-15], several hidden variables contribute to the complex phenotype observed in pharmacologic assessments of biased signaling. Initially, the largest contribution to observed biased agonism was from direct ligand bias, the actions of the biased agonist at the level of the receptor to generate unique conformations that differentially signal through transducers at the plasma membrane. But there are numerous other mechanisms that contribute to a biased response. For example, there are “biased receptors” which demonstrate an intrinsic preference for a certain signaling pathway, irrespective of the ligand used to activate the receptor [16]. In “systems bias,” a balanced ligand:receptor complex can signal differently depending on the abundance of GPCR transducers and effectors in a given cellular environment [17]. There can also be kinetic effects to bias, with some agonists generating different temporal patterns of signaling.

In “location bias,” biased agonists promote signaling from different locations within the cell, resulting in spatially and, often, functionally, distinct GPCR signaling between agonists. This signaling can range from activation of the same type of G proteins at different locations in the cell, e.g., the plasma membrane versus the endosome, but can also refer to the activation of distinct signaling pathways in different subcellular locations. For example, G proteins and β-arrestins have been found to form signaling complexes with a GPCR in endosomes, frequently with signaling profiles that are distinct from their plasma membrane counterpart [18-21]. Such processes can lead to complex patterns of spatially-localized signaling, considerably impacting a biased response. Here, we briefly review examples of location bias, and discuss the implications of this new signaling paradigm on the characterization and development of GPCR therapeutics.

The role of β-arrestins and G proteins in GPCR signaling from endosomes

While GPCRs were previously thought to signal exclusively from the plasma membrane, there is now substantial evidence supporting the activation of GPCRs at various subcellular locations, such as at endosomes, the nucleus, endoplasmic reticulum (ER), Golgi, mitochondria, and more (Figure 1) [22]. Even at the plasma membrane, there is heterogeneity in receptor expression, where hotspots for G protein:GPCR and β-arrestin:GPCR interactions can form in clathrin-coated pits due to the regional composition promoting greater complex stabilization [23,24]. Signaling from such membrane microdomains also falls under the larger umbrella of location bias. Following activation, many GPCRs undergo receptor-mediated endocytosis which can serve as one mechanism to abrogate GPCR signaling by restricting the pool of ligand-accessible receptor or targeting the receptor for degradation. However, many GPCRs can continue signaling from endosomes, and this endosomal signaling may parallel or diverge from the pattern observed at the plasma membrane [21]. Several studies have recently sought to evaluate the relative roles of β-arrestins, G proteins, and β-arrestin:G protein complexes in facilitating functionally distinct endosomal signaling at GPCRs, which we review below.

Figure 1. GPCRs are found at multiple subcellular locations.

Although GPCRs exist and are primarily studied at the plasma membrane, they are also expressed or traffick to multiple other organelles in the cell, including the endosome, endoplasmic reticulum (ER), Golgi apparatus (Golgi), nucleus, mitochondria, and more.

Historically, β-arrestins were a focus in GPCR endosomal signaling as they are known to induce receptor-mediated endocytosis through interactions with AP2, phosphoinositides, and clathrin [25-28]. Mechanistically, the interaction of β-arrestin with the receptor governs the receptor's internalization and resensitization profile at many receptors [29]. β-arrestin:receptor complexes can exist in two distinct sets of conformations: the “tail” conformation, where β-arrestin only couples to the phosphorylated receptor C-tail, and the “core” conformation, where β-arrestin also engages with the intracellular core of the receptor in addition to the C-tail [30] (Figure 2). Studies of distinct β-arrestin:GPCR complexes at the plasma membrane and at endosomes have shown that receptor internalization is primarily mediated by the “tail” conformation, while the “core” conformation is dispensable for receptor internalization, but exhibits a greater role in receptor desensitization with regards to G protein signaling [30-32]. Seminal work demonstrated distinct patterns of β-arrestin recruitment to different GPCRs fall into two broad categories: at type A receptors, β-arrestin transiently interacts with the receptor (such as the β2-adrenergic receptor (β2AR)) at the plasma membrane; while at type B receptors β-arrestins bind tightly to the receptor (such as the vasopressin type 2 receptor (V2R)) and internalize with it into endosomes [33]. When the C-terminal tails of class A and B receptors are swapped, the two receptors’ patterns of dephosphorylation, recycling, and resensitization are largely switched, indicating that the ability of a receptor to internalize is largely dictated by its C-terminal tail and its interactions with β-arrestin. These data suggest that β-arrestin conformations promote distinct functions.

Figure 2. β-arrestin conformational interactions with a GPCR.

β-arrestin primarily interacts with a GPCR in the ‘core’ conformation (left) which leads to G protein desensitization, or the ‘tail’ conformation (right) which leads to receptor internalization.

Notably, β-arrestin-mediated internalization is sometimes necessary for a GPCR to achieve maximal signaling output, despite its historical role in the abrogation of G protein signaling [34]. In addition to promoting receptor internalization, β-arrestins have also been shown to colocalize with GPCRs at endocytic vesicles, suggesting a role in modulating signaling from endosomes as well [35-37]. Initially this signaling was thought to be promoted only by β-arrestins and not G proteins. For example, GPCR signaling from endosomes has been shown to mediate rapid extracellular signal-regulated kinase 1/2 (ERK1/2) activation [38]. This β-arrestin-mediated endosomal signaling was thought to be discernable from G protein signaling, with distinct spatial and temporal patterns of activation [18,36,39]. In many of these systems, β-arrestin-mediated ERK activation occurred later and was limited at endosomes, while canonical G protein-mediated ERK occured at earlier times in the cytoplasm, although some systems demonstrated different kinetics and patterns [40]. However, it was soon shown that GPCRs could also initiate G protein-mediated signaling pathways from endosomes [41-43]. Subsequent work demonstrated endosomal activation of Gαs, Gαi/o, and Gαq at a diverse array of GPCRs [12,44-46].

GPCRs which activate the Gαs subunit and exhibit receptor signaling from endosomes have demonstrated sustained and consequential cAMP signaling upon receptor internalization [43,47-50]. Additionally, endosomal cAMP production has been shown to promote nuclear cAMP and PKA activation in ways that cAMP generated at the plasma membrane cannot [44,51,52]. At some receptors, β-arrestins appear dispensable for promotion of endosomal signaling, suggesting a complex interplay between G proteins and β-arrestins depending on the receptor [53]. These findings suggest that receptor subcellular signaling may serve to generate sufficient quantities of second messengers in discrete micro- or nanodomains within the cell to generate spatially-determined cellular functions [54].

Rather than acting as discrete signaling pathways, β-arrestins and G proteins can jointly promote signaling through downstream effectors. β-arrestins and G proteins have also been shown to form “megaplexes” with the GPCR, where a GPCR is simultaneously bound to β-arrestin in the “tail” conformation, while the core of the GPCR is occupied by G protein [20,32,55,56]. The functional relevance of these complexes has been suggested. For example, destabilization of the β-arrestin “core” conformation can selectively prevent receptor desensitization without affecting patterns of receptor internalization and β-arrestin-dependent signaling [30,31]. Thus, a system exists where both G protein and β-arrestin-dependent signaling pathways can persist in endosomes and likely other subcellular organelles. Unfortunately, some debate in the literature has focused on a paradigm that G proteins and β-arrestins act in purely an antagonistic fashion [57,58], while it appears that β-arrestins and G proteins contribute to intracellular signaling cascades in complex patterns [59], with β-arrestin promoting some signaling cascades while antagonizing others in a location-biased fashion [12].

Other Sites of Subcellular Signaling

Endosomes are only one of the various subcellular locations at which GPCR signaling has been detected. GPCRs have been extensively studied at the primary cilium, an organelle specialized to act as the “cell’s antenna,” with specific subcellular group of receptors demonstrates prominent roles in sensing mechanical and chemical signals, and integrating these signals into physiologic processes important during development [60-63]. GPCRs have also been detected in the nucleus, with some of the first evidence of nuclear GPCR localization involved explorations of the angiotensin II type I receptor (AT1R) due to its recognized nuclear localization signal (NLS) sequence [64]. Here, receptor activation with angiotensin II (AngII) promotes nuclear sequestering of AT1R but not angiotensin II type II receptor (AT2R), which does not possess an NLS sequence. Similar work has revealed the nuclear presence of prostaglandin E₂ receptor 1 (EP1) independently of ligand stimulation [65]. Recent work conducted in astrocytes has shown that norepinephrine is transported across the outer nuclear membrane via an organic cation transporter, and activates inner nuclear membrane β1-adrenergic receptor (β1AR) and, consequentially, nuclear PKA activity [66]. Over 40 GPCRs have been recognized as either localizing or having the ability to translocate to the nucleus, modulating various classical downstream signaling effectors, including phosphorylation cascades and calcium influx [21].

Because GPCRs often undergo post-translational modifications at the ER and the Golgi apparatus, it can be difficult to differentiate between GPCRs that reside in these domains, and those which are trafficked elsewhere [21]. However, studies using fluorescently tagged GPCRs, intracellular ligands, membrane impermeable inhibitors, and location markers revealed some of the first known examples of GPCR signaling at the ER [67,68]. For example, the estrogen receptor GPR30 was shown to be predominantly localized to the ER and whose activation promoted intracellular calcium mobilization and phosphatidylinositol 3,4,5-triphosphate synthesis in the nucleus [68].

GPCRs have also demonstrated the ability to signal from the membrane of the Golgi apparatus. The thyroid-stimulating hormone receptor (TSHR) has been shown to translocate to the trans-Golgi network (TGN), where it activates a late-stage cyclic AMP and PKA response necessary for adequate cAMP response element binding protein (CREB) phosphorylation and gene transcription [69]. Moreover, a preexisting receptor pool of β1AR at the Golgi contributes significantly to the global cAMP response, and commonly used β-blockers exhibit differential abilities to antagonize Golgi-localized β1AR [70]. The synthetic drug FTY720 (fingolimod), which targets the sphingosine-1-phosphate receptor (S1PR1), demonstrates sustained S1PR1 internalization and activity in the TGN, ultimately enhancing the chemokinetic migration of endothelial cells [71].

The mitochondria serve as a crucial location for several GPCRs, such as the melatonin type I receptor (MT1R) which presides on the outer mitochondrial membrane due to the mitochondrial matrix’s role in synthesizing the ligand melatonin [72].. Cannabinoid type 1 receptor (CB1R) similarly demonstrates mitochondrial expression in striated muscles as well as in the brain, where mitochondrial CB1R contributes to the amnesic effects associated with cannabinoids [73,74]. GPCRs are also known to be involved in the process of cytokinesis due to their localization to centrosomes, spindle midzones, and midbodies [75]. They have even been shown to be present on exosomes, affecting nearby cells by stimulating ERK1/2 activation and tubulogenesis [76].

Impact of subcellular location on receptor pharmacology

When determining the efficacy, potency, and affinity of a ligand, most experiments rarely consider how subcellular localization could impact these measurements, primarily due to a previous lack of ability to do so. Traditional dose-response curves characterize ligand activity by efficacy (E_max), the maximum response that can be obtained following ligand treatment, and potency (EC₅₀), the concentration of ligand needed to produce 50% of the E_max (Figure 3A). Conversely, saturation radioligand binding studies on membranes or whole cells characterize ligand affinity by labeling a fixed amount of receptor with increasing amounts of radioactively labeled ligand (Figure 3B) [77]. After accounting for both specific and non-specific ligand binding, these data will ultimately reveal a binding curve that provides the affinity of the ligand for a receptor, commonly represented by the dissociation constant (K_D), the concentration of the ligand which occupies 50% of receptors at equilibrium, as well as the maximum number of receptor binding sites (B_max). Given the advent of new tools to more precisely assess ligand activity and behavior in specific cellular contexts, we provide a brief review of data that demonstrate how subcellular localization could impact the characterization of a ligand.

Figure 3. Subcellular location can impact GPCR pharmacology.

(A) Hypothetical dose-response curve (left) and saturation-binding curve (right). In the dose-response curve, the maximum response (E_max) and concentration of ligand which produces 50% E_max (EC₅₀) are labeled. In the saturation-binding curve, the maximum number of receptor binding sites (B_max) and the equilibrium dissociation constant (K_D) are labeled. Drug concentrations are labeled as arbitrary units (A.U). Hypothetical dose-response curves measuring β-arrestin recruitment to a GPCR using three different ligands under conditions with endocytosis (B) or where (C) endocytosis is inhibited.

GPCR signaling depends on receptor location

A recent publication by our group demonstrated that G protein activation by three ligands at the chemokine receptor CXCR3 is location-dependent [12]. By modifying a recently published biosensor that detects GTP-bound Gαi [78], we measured the amount of activated G protein at the plasma membrane and endosome. At the plasma membrane, the ligand CXCL11 had the greatest efficacy, followed by CXCL10 and then CXCL9. However, at the endosome, CXCL11 and CXCL10 demonstrated nearly identical E_max, while CXCL9 remained the least efficacious ligand (Figure 3C). These data demonstrate that CXCL10 acts as a partial agonist at the plasma membrane but a full agonist at the endosome. We also demonstrated that both CXCL10 and CXCL11 promote nuclear ERK activation, which is then decreased by inhibiting endocytosis via overexpression of a dominant negative form of Dynamin (Dynamin K44A). Interestingly, CXCL10 does not recruit β-arrestin to endosomes, while CXCL11 does, suggesting that the exact mechanism of nuclear ERK is complex and can be achieved via multiple mechanisms at a single GPCR.

A similar finding was recently demonstrated in a study of spatiotemporal β-arrestin signaling at the AT1R using a variety of biased peptide agonists [79]. They found that these ligands greatly differed in their ability to recruit β-arrestin to the AT1R. However, when the authors inhibited internalization, they observed that nearly all partial agonists behaved liked full agonists (i.e. similar efficacies) (Figure 3D). Notably, the potencies of these agonists were unchanged with inhibition of internalization. These data together suggest that these ligands have the ability to recruit similar amounts of β-arrestin to the AT1R at the plasma membrane; however, GPCR activation can lead to various cellular processes, like receptor-mediated endocytosis, which can impact the measured pharmacologic properties of a ligand, and thereby, influence how a ligand is characterized (Figure 3B and 3C).

Lastly, location-specific GPCR signaling has been implicated in regulating the phosphoproteome at the β2AR [80] and CXCR3 [81], as well as the transcriptome [12]. These data demonstrate that location-specific GPCR signaling has critical effects in regulating specific pools of second messengers and other downstream signaling effectors, but also plays a role in regulating complex pools of biomolecules like protein and RNA.

The interaction between location-specific and biased GPCR signaling

The effects of subcellular localization should be considered in the context of biased GPCR signaling, which describes the phenomenon of a GPCR activating some signaling pathways over others, which is frequently used in the context of comparing G protein and β-arrestin-dependent pathways [11]. There is burgeoning evidence that changes in the subcellular localization of a GPCR can impact some signaling pathways but not others. For example, in the study on location bias at CXCR3, we found that at the plasma membrane, CXCL9, CXCL10, and CXCL11 all are capable of signaling through G proteins and β-arrestins. However, at the endosome, only CXCL11 was able to recruit β-arrestin, while all three ligands still maintained G protein activation [12]. Similar findings were obtained in the study on the AT1R, where the authors found that inhibiting internalization had a significant impact on a ligand's ability to recruit β-arrestin, but did not impact G protein activation [79]. It is plausible that these observations occur due to differential localization and availability of G proteins and β-arrestins in subcellular compartments. Nonetheless, location-specific effects can be pathway-dependent but likely impact other aspects of signaling, including temporal regulation, protein conformation, and other GPCR interacting partners beyond G proteins and β-arrestins.

Ligand properties depend on location

In addition to a cell’s subcellular location, the extracellular tissue environment where a receptor is expressed can affect the activity of its ligands. Optimization of fentanyl, an agonist of the μ-opioid receptor (MOR), to achieve better analgesics led to the development of a fentanyl analog, NFEPP, which demonstrates preferential activation of the MOR in acidic environments like the endosome, but also in inflamed, acidic tissue [82,83]. While fentanyl showed little change in binding affinity for the MOR at different pH values, NFEPP demonstrated a significant decrease in receptor affinity at physiologic pH as compared to acidic conditions [82]. This observation was accompanied by increases in potency and efficacy when measuring NFEPP-induced cAMP inhibition at acidic pH values. These data demonstrate the potential to create MOR agonists that have preferential antinociceptive activity in acidic environements like the endosome or damaged tissue, but minimal on-target effect in healthy tissue (Figure 4A) [83]. Additionally, they demonstrate the importance of characterizing ligand properties under different experimental conditions to determine its complete pharmacological profile.

Figure 4. Hypothetical Examples of GPCR Location Bias.

Location bias can be achieved in many different ways, including (A) ligands that demonstrate pH-sensitive activity, (B) ligands that are hydrophobic or hydrophilic, (C) differential location-specific activity of allosteric modulators, and (D) synthetic small molecules and peptides which demonstrate different rates of degradation.

Numerous factors beyond the lipid environment and pH of the subcellular compartment can impact the behavior of a ligand or a GPCR. For example, it was recently shown that sodium serves as a negative allosteric modulator of the ghrelin receptor [84]. Given that organelles compartmentalize electrolytes like sodium, potassium, and calcium [85], it is likely that these and other features of specific subcellular environments impact the affinity, potency, and efficacy of a ligand.

GPCRs are in multiple subcellular locations

Many GPCRs are expressed at multiple cellular locations in both normal physiologic and pathophysiologic states [21]. This is important as some ligands demonstrate the ability or inability to penetrate through the cell membrane, related to characteristics like polarity or the presence/absence of a drug transporter. For example, a variety of agonists and antagonists of the vasopressin receptor 2 (V2R) have been used to probe its location-dependent signaling [30,86]. Many patients with nephrogenic diabetes insipidus harbor V2R mutations that lead to intrinsically functional but misfolded and ER retained V2R [87]. As a result, the endogenous ligand, AVP or its analog ddAVP, cannot reach this intracellularly trapped receptor. Researchers developed nonpeptide, cell-permeable V2R agonists that could activate the ER-retained V2R mutants and initiate cAMP accumulation and restore normal translocation of aquaporin-2 water channel to the plasma membrane [88]. This work highlights how the characterization of a ligand may depend on where the receptor is located and the accessibility of that ligand to that specific compartment (Figure 4B). A further understanding of intracellular GPCR signaling will likely lead to novel therapeutic strategies that target specific pools of receptors.

Location-biased allosteric modulators

There is an increasing appreciation that GPCRs can be regulated allosterically, at sites distinct from the orthosteric site, greatly expanding the approaches to target GPCRs therapeutically [11,89]. The orthosteric binding pocket of many GPCRs is highly conserved, rendering the development of highly selective ligands difficult. However, allosteric binding sites are generally less conserved, making it easier to generate receptor-specific drugs with context-dependent activity. Allosteric regulation of GPCRs allows for fine-tuning of GPCR signaling but only when activated by the endogenous ligand [90]. Allosteric modulators are classically described across the spectrum from positive allosteric modulators (PAMs) to negative allosteric modulators (NAMs) [91], and they may also have inherent agonist activity (“ago-allosteric modulators”). There is also strong evidence for biased allosteric modulators (BAMs) which demonstrate selective modulation of some signaling pathways but not others [92]. Additionally, most allosteric modulators demonstrate probe dependence, wherein the behavior of the modulator is dependent on the specific orthosteric ligand bound to the receptor [93].

Due to the enhanced selectivity of allosteric modulators for receptor subtypes and the abundance in allosteric sites on receptors compared to orthosteric sites, allosteric modulators have been a major focus for developing therapeutics targeting several families of GPCRs. In the last decade, remarkable progress has been achived in the development and optimization of allosteric modulators with several NAMs, PAMs and BAMs being approved for clinical use or currently undergoing clinical trials. Cinacalcet, which targets the calcium-sensing receptor (CaR), is the first clinically administered allosteric modulator of GPCR in 2004 for the management of hyperparathyroidism and thus, this compound represents an imporant proof-of-concept for the potential of allosteric modulators as therapeutic targets for GPCRs [94]. Interestingly, cinacalcet was later reported to be a biased allosteric modulator of CaR as it exhibits a preference for potentiating Ca²⁺ mobilization and IP1 accumulation over ERK1/2 activation [95,96]. Other success examples of allosteric modulators in clinical use include maraviroc, which acts as a NAM of the chemokine receptor CCR5 and was FDA-approved in 2007 for the treatment of HIV infection [97,98]. Ticagrelor, approved in 2011 for antithrombotic treatment in cardiovascular diseases, is a reversible allosteric antagonist of the P2Y₁₂ receptor and thus inhibits P2Y₁₂-mediated platelet aggregation [99].

There are many endogenous allosteric modulators of GPCRs, including small molecules, amino acids, peptides, ions, lipids, autoantibodies, and other proteins, including G proteins, β-arrestins, receptor activity-modifying proteins (RAMPs), and even other GPCRs through hetero- and homodimerization/oligomerization [100,101]. Based on our previous discussion on the influence of subcellular localization on ligand behavior, it is plausible that allosteric modulators similarly demonstrate location-specific effects like their orthosteric counterparts. For example, RAMPs are single transmembrane-spanning proteins that can alter ligand selectivity, signaling profiles, and GPCR trafficking to and from the cell surface [102]. Given that cell types express different RAMP isoforms and that GPCRs can couple to multiple RAMPs with nonredundant functionality, it is possible that differences in subcellular RAMP expression and or trafficking will manifest as location-specific allosteric effects (Figure 4C) [102]. A similar framework extends to synthetic allosteric modulators that could demonstrate probe dependence, not only with the orthosteric ligand but also with the cellular environment. For example, one could hypothesize that the cellular environment at the plasma membrane leads a small molecule to act as a PAM, while in the endosome, it behaves as a NAM. The number of plausible outcomes regarding the interaction between GPCR allosteric regulation and subcellular localization is difficult to grasp, and further research will hopefully deconvolute these complex systems.

Pharmacological implications of location-biased therapeutics

Due to importance of subcellular signaling at GPCRs, efforts have begun to develop pharmaceuticals that target subcellular pools of GPCRs to obtain optimal efficacy and specificity. Limitations of drugs that cannot cross the plasma membrane are evident in their inferior efficacies. For instance, nanobodies have been developed that target the receptor US28, a virally-encoded GPCR that is homologous to human chemokine GPCRs but shows significant consistutive activitiy and internalization in the absence of agonist stimulation [103]. This receptor demonstrates oncomodulatory activity by activating G protein-mediated pathways implicated in cell proliferation, migration, survival, and inflammation [104]. The inhibitory nanobodies developed by Heukers et al. are membrane impermeable, and hence can evaluate the membrane contribution of US28 signaling on tumor growth. These nanobodies reduced US28-mediated tumor cell growth by 50% compared to stimulation with chemokine agonist, suggesting a role of US28 subcellular signaling on tumor growth [105].

Conversely, in order for a ligand to bind to the β1AR receptor pool expressed at the Golgi, the drug would either have to be hydrophobic to undergo passive diffusion, or it would require a specific membrane transporter. In contrast to hydrophilic β1AR antagonists, the hydrophobic nature of metoprolol allows for β1AR inhibition from both the plasma membrane and the Golgi, resulting in a greater cAMP reduction and subsequent lowering of heart rate and contractility [70]. At the neurokinin 1 receptor (NK1R), where endosomal NK1R signaling is partially responsible for sustained nociception, there are efforts to develop NK1R antagonists that act at both the plasma membrane and endosome, which is slightly more acidic [46,106-108]. NK1R antagonists that were more lipophilic and electrochemically neutral at low endosomal pH were significantly enriched in endosomes and able to inhibit allodynia in vivo more efficaciously and for longer than antagonists without these properties [106,108,109]. Another example of the potential utility of location-biased drugs is the MOR ligand NFEPP mentioned previously, which demonstrates significant antinociceptive ability in diseased tissue, but very little on-target effects in other tissues, thus avoiding central nervous system or gastrointestinal side effects [82,83].

Comparing the efficacies of parathyroid hormone 1-34 (PTH1-34) and parathyroid hormone-related peptide 1-36 (PTHr1-36), PTH1-34 promotes a greater parathyroid hormone receptor (PTHR) signaling output due to its ability to remain associated with PTHR following internalization, whereas PTHr1-36 is confined to interacting with PTHR only at the cell surface [47,110]. Similarly, the ability of AVP to prolong V2R endosomal signaling is argued to be the source of its superior antidiuretic/antinatriuretic effects as compared to oxytocin, which fails to induce significant V2R internalization [111,112]. The development of peptide analogs that can avoid degradation by peptidases has also been explored as an avenue for future therapeutics, as they lead to sustained endosomal signaling and prolonged ERK activation (Figure 4D) [113]. There are many different approaches to modulating GPCR signaling beyond the plasma membrane, greatly expanding the therapeutic potential of the GPCR superfamily.

CONCLUSION

The discovery of location bias and its potential to target intracellular GPCRs and GPCR-mediated signaling cascades presents new and exciting opportunities to increase drug efficacy and reduce adverse on-target effects in drug development. However, the development of location-specific GPCR pharmaceuticals requires an enhanced understanding and characterization of GPCR trafficking patterns and signaling profiles at specific subcellular locations. Additionally, more in vivo studies are required to demonstrate the physiologic relevance of this phenomenon. Elucidation of the structural differences between receptors at the plasma membrane and different intracellular compartments will also be important for designing more specific orthosteric and allosteric ligands that modulate receptor activity at desired cellular locations. Strategies to precisely target drugs to specific subcellular locations require knowledge of the unique properties of organelles, including membrane composition, pH, ion concentration, and more. Machine learning and computational simulations can assist in the structure-based drug design of peptides and small molecules that preferentially bind to receptors at specific locations. This new dimension of receptor pharmacology can help explain how complex physiologic events can be driven by seemingly simple ligand-receptor interactions, and usher in a new era of more effective, safer, and targeted GPCR therapeutics.

DATA AVAILABILITY STATEMENT

Data sharing not applicable to this article as no datasets were generated or analysed during the current study