Type 2 diabetes is a major global health problem and there is ongoing research for new treatments to manage the disease. The GLP-1R (glucagon-like peptide-1 receptor) controls the physiological response to the incretin peptide, GLP-1, and is currently a major target for the development of therapeutics owing to the broad range of potential beneficial effects in Type 2 diabetes. These include promotion of glucose-dependent insulin secretion, increased insulin biosynthesis, preservation of β-cell mass, improved peripheral insulin sensitivity and promotion of weight loss. Despite this, our understanding of GLP-1R function is still limited, with the desired spectrum of GLP-1R-mediated signalling yet to be determined. We review the current understanding of GLP-1R function, in particular, highlighting recent contributions in the field on allosteric modulation, probe-dependence and ligand-directed signal bias and how these behaviours may influence future drug development.
Type 2 DM (diabetes mellitus) is a global epidemic, with worldwide prevalence increasing exponentially and future projections estimating that almost 10% of the adult population will suffer from the condition by the year 2030 . A complex disease, arising from multiple aetiological factors including genetic predisposition and modern lifestyle, Type 2 DM is typically diagnosed by chronic hyperglycaemia; however, the two distinct features allowing disease progression are impaired β-cell function and a target organ reduction in sensitivity to insulin. In the later stages of the condition, the continual demand for elevated insulin to compensate for insulin insensitivity results in β-cell exhaustion and glucose toxicity . Aside from these characteristic traits of Type 2 DM, there are also many other associated pathophysiologies including vascular dysfunction, the consequences of which include retinopathy, nephropathy, neuropathy and atherosclerosis, the latter increasing the risk of heart attack and stroke in addition to significantly increasing the risk of cardiovascular mortality . The evolution of understanding into both the physiology of glucose homoeostasis as well as the pathophysiology of Type 2 DM has highlighted the importance of endogenously produced incretin hormones in facilitating nutrient-induced insulin biosynthesis and secretion, as well as preserving β-cell function, decreasing β-cell apoptosis, slowing gastric emptying, and enhancing insulin sensitivity at peripheral tissues (reviewed in ). This article provides a brief overview of the GLP-1R (glucagon-like peptide-1 receptor), the major target of incretin mimetic therapies, and highlights some of the previous work on this receptor.
Physiology of the incretin system
Accounting for as much as 70% of insulin secreted from pancreatic β-cells following nutrient consumption, incretin hormones are key mediators in communicating nutrient content of the gastrointestinal tract to insulin producing pancreatic β-cells . The principal incretin hormones include GLP-1, primarily expressed in L cells of the ileum and colon, and gastric inhibitory polypeptide/glucose-dependent insulinotropic polypeptide, primarily expressed in K cells of the duodenum and jejunum. Although secreted levels of both incretin hormones are reduced in Type 2 DM subjects, only GLP-1 has been observed to retain its potent insulinotropic activity, and has therefore attracted significant interest in the development of Type 2 DM therapeutics .
The principal stimuli for GLP-1 secretion is nutrient content of the gastrointestinal tract ; however, the mechanisms behind GLP-1 secretion are complex and largely unclear, with multiple factors thought to impact on its release, including neural and endocrine factors (reviewed in ). GLP-1 is rapidly secreted postprandially, peaking at 10–15 min followed by a sustained peak at 30–60 min . The insulinotropic effects induced by secreted GLP-1 are mediated through interaction with its transmembrane expressed GPCR (G-protein-coupled receptor), the GLP-1R, promoting intracellular signalling mechanisms to aid in increasing the expression, biosynthesis and secretion of insulin from pancreatic β-cells in a glucose-dependent manner . Highlighting the importance of GLP-1-mediated signalling in the endocrine pancreas, all studies of GLP-1R−/− mice observe at least a modest reduction in glucose tolerance and impaired glucose-stimulated insulin secretion . In addition to glucoregulation, GLP-1 has a fundamental role in increasing neogenesis, proliferation and decreasing apoptosis of pancreatic β-cells in animal models, leading to an increase in β-cell mass and subsequently aiding the glucose-dependent augmentation of insulin secretion .
Aside from the pancreatic effects, there is significant evidence illustrating biological actions of GLP-1 via its receptor in other tissues (extensively reviewed in ). Briefly, GLP-1 activity suppresses appetite and inhibits gastric emptying, in turn influencing ingestive behaviour. Other roles include inhibition of glucagon release and augmenting glycogen synthase activity in muscle, adipose and hepatic cells, favouring incorporation of glucose into glycogen. Furthermore, GLP-1 and GLP-1-related peptides enhance peripheral insulin sensitivity and reduce steatosis. In the nervous system, GLP-1 augments neogenesis, proliferation and anti-apoptotic behaviour of neuronal cells, enhancing memory, and spatial and associative learning. Other documented roles include contribution to normal cardiovascular, respiratory and renal function. The diverse and beneficial actions of GLP-1 have consequently attracted significant attention in the development of therapeutics that mimic the endogenous GLP-1 system, particularly for the management of Type 2 DM.
The GLP-1R is a 463-amino-acid transmembrane-spanning protein belonging to the family B/secretin GPCRs, mediating the effects of both endogenous GLP-1 peptides [four forms: GLP-1(1–36)NH2, GLP-1(7–36)NH2, GLP-1(1–37) and GLP-1(7–37)], as well as the endogenous peptide oxyntomodulin and exogenous peptide exendin-4 (Figure 1A). Characteristic of family B GPCRs, the GLP-1R possesses a long extracellular N-terminus with an α-helical region, five β-strands forming two antiparallel β-sheets and six conserved cysteine residues that form disulfide interactions [10–12]. Together, these features allow the receptor to adopt the classic ‘Sushi domain’ or ‘short consensus repeat’, which aids N-terminal stability and confers a high level of structural homology within the N-terminal regions of family B GPCRs. The large extracellular N-terminus has a significant role in peptide binding, supported by GLP-1 binding the isolated N-terminus of the GLP-1R  and crystal structures of the isolated GLP-1R N-terminus in complex with GLP-1 and exendin peptides [11,12]. Specifically, the C-terminus of the peptide interacts with the N-terminus of the receptor, which is proposed to be responsible for ligand recognition and specificity, while the N-terminus of the peptide is proposed to associate with the core of the receptor, and is suggested to have a major influence in signalling specificity and transmission [14,15]. This widely accepted two-domain model of ligand binding is also experimentally supported by chimaeric receptors [16,17], photolabile peptide cross-linking [18–20], and to some extent, mutagenesis analysis [21–25]. However, despite the seemingly abundant data, there is still a wide knowledge gap with respect to the complete structure of any receptor in this family, as well as whether a definitive binding crevice exists that is common across all receptors of the family. Furthermore, the orientation of the receptor N-terminus in relation to the transmembrane bundle is uncertain, and has been inherently difficult to establish either experimentally or using molecular modelling [15,20].
Peptide and small molecule ligands of the GLP-1R
GLP-1R signalling and regulation
The physiological changes observed with increases in GLP-1, including increases in insulin secretion and β-cell mass, rely on signalling via GLP-1R-mediated intracellular pathways (Figure 2). The GLP-1R is a pleiotropically coupled receptor, with evidence for signalling via multiple G-protein-coupled pathways including Gαs, Gαi, Gαo and Gαq/11 [26,27]. However, the GLP-1R is most well documented for its role in Gαs coupling, favouring production of cAMP through increasing enzymatic activity of adenylate cyclase . This subsequently promotes increases in both PKA (protein kinase A) and Epac2 (exchange protein activated by cAMP-2), which is directly involved in enhancing proinsulin gene transcription . Furthermore, GLP-1R activation induces membrane depolarization of β-cells through inhibition of K+ channels, allowing VDCCs (voltage-dependent Ca2+ channels) to open and acceleration of Ca2+ influx to occur, resulting in the exocytosis of insulin from β-cells. Therefore the production of cAMP and influx of Ca2+ are vital components in the biosynthesis and secretion of insulin. GLP-1R activity also promotes EGFR (epidermal growth factor receptor) transactivation, PI3K (phosphoinositide 3-kinase) activity, IRS-2 (insulin receptor substrate-2) signalling, and subsequently, ERK1/2 (extracellular-signal-regulated kinase 1 and 2) activity, as well as nuclear translocation of PKCζ to mediate β-cell proliferation and differentiation as well as promote insulin gene transcription (reviewed in ). Aside from G-protein-coupled pathways, there are recently emerging studies suggesting that GRK (GPCR kinase) and β-arrestin recruitment are involved in optimal GLP-1R function [29–32]. Clear evidence for this is seen in β-cell knockdown of β-arrestin-1, that results in attenuated cAMP and consequently diminished insulin secretion . There is also evidence supporting β-arrestin-1-mediated ERK1/2 activation as a mechanism for β-cell preservation . Although GRKs and β-arrestins are well documented for their role in regulating cell-surface receptor function and expression through receptor desensitization and internalization, it is unclear how these scaffolding proteins regulate this process at the GLP-1R.
GLP-1R-mediated signalling in pancreatic β-cells
GLP-1 mimetics in the treatment of Type II DM
With the ability to address almost all manifestations of Type 2 DM, the GLP-1R system has become one of the most appealing targets in the development of therapies for management of the condition. However, the leading problem in enhancing this system with GLP-1 administration directly is the rapid breakdown of the peptides by the enzyme DPPIV (dipeptidyl peptidase IV) into low activity metabolites. As such, the most prominent avenue of drug development aims to imitate endogenous peptide activity but limit peptide breakdown (GLP-1 mimetics). The most well-known GLP-1 mimetic prescribed for the management of Type 2 DM is exenatide (Byetta®), a synthetically produced equivalent of the venom-derived peptide exendin-4 (Figure 1A). Similar to GLP-1, exendin-4 decreases plasma glucose levels immediately following nutrient ingestion in both healthy and diabetic subjects, promotes β-cell proliferation, and augments the synthesis and secretion of insulin . However, unlike GLP-1, exendin-4 is resistant to the proteolytic activity of DPPIV, prolonging its activity in vivo.
Unlike exendin-4, all other GLP-1 mimetics are synthetically developed, modified GLP-1 peptides that are designed to take advantage of the peptide's specificity for the receptor, but have alterations to enhance stability and/or function in vivo. These modifications typically involve the substitution of Ala8 of the GLP-1 peptide, such that the peptide becomes resistant to enzymatic degradation by DPPIV. Examples of this include (Val8)GLP-1, (Thr8)GLP-1, (Ser8)GLP-1 and (Gly8)GLP-1, each of which display insulinotropic activity and enhanced metabolic stability .
Peptide modifications through fatty acid derivatization have also been pursued in order to extend biological half-life in plasma. A well-recognized example of this is liraglutide (NN2211, Victoza®), which covalently couples a hexadecanoic fatty acid at the Lys26 residue of the GLP-1 peptide, as well as containing an arginine substitution at residue 34  (Figure 1A). Similar to GLP-1 and exendin-4, liraglutide significantly improves glycaemic control, enhances β-cell function and promotes weight loss, and, similar to exendin-4, has a significantly improved plasma half-life due to DPPIV resistance .
There are many other synthetically engineered peptide analogues for the GLP-1R that have been shown to have insulinotropic activity and enhanced metabolic stability, including the GLP-1 analogues LY315902 and CJC-1131 and the albumin-conjugated dimeric GLP-1 analogue, albiglutide [34,36], and exendin-4 analogues AC3174 and CJC-1134-PC [36,37]. In another previous study, modification through biotin and polyethylene glycol labelling of GLP-1 and exendin-4 peptides have been explored as a means to aid oral delivery of antidiabetic treatments through enhancing intestinal absorption .
Collectively, synthetically produced GLP-1 and exendin-4 analogues illustrate that biological activity can be mimicked and in some cases favourably enhanced. However, generation and application of peptides remains a difficult and complex task, with peptide stability and administration route a major challenge, as well as controversy over the long-term consequences of use, including reports of pancreatitis and C-cell hyperplasia, a precursor for thyroid cancer [39,40]. In addition, all analogous peptides are coupled to some extent with adverse side effects, the most prominent being nausea. For this reason, there is significant interest in novel treatments that have similar physiological effects to GLP-1, but which can be administered orally and eliminate, or at least minimize, side effects.
GPCRs are widely accepted to be promiscuous, signalling via multiple G-protein-dependent and -independent mechanisms on receptor activation. It has become increasingly evident that in such pleiotropically coupled receptor systems, receptor activation can engender differential effects via multiple pathways depending on the ligand present in the system. This phenomenon is termed ‘biased signalling’, but may also be referred to as ligand-directed signalling, ligand-directed stimulus bias, functional selectivity or stimulus trafficking, and is a result of different ligands stabilizing distinct receptor conformations, which subsequently influence the nature and strength of pathway coupling that may include alterations to G-protein coupling profiles, but also to non-G-protein signalling pathways such as those mediated by β-arrestins . Biased signalling has been observed at many receptors including the pituitary adenylate cyclase-activating polypeptide receptor, 5-hydroxytryptamine 2c receptor, μ-opioid receptor, dopamine receptors, V2 vasopressin receptor, β2-adrenergic receptor and recently at the GLP-1R [41,42].
Recent analytical advances in the field have demonstrated that bias in a system can be quantified through estimating τ/KA ratios, where τ equates to the efficacy in the system [24,25]. This is a novel method to determine signalling bias in a system where profound reversal of potencies is not observed. At the GLP-1R, all peptide agonists preferentially activate cAMP over ERK1/2 and Ca2+in vitro. However, the relative degree of bias is variable between ligands, with truncated GLP-1 peptides and exendin-4 having greater bias towards cAMP than full-length GLP-1 peptides and oxyntomodulin (Figure 3) [24,25,42]. This is particularly important to consider in pharmacological characterization of any receptor, and may have the potential to be exploited in the rational design of therapeutics that target pathways associated with beneficial effects over pathways that are associated with detrimental effects.
Biased signalling at the GLP-1R
Aside from both the endogenous and exogenous peptide agonists, there have been several synthetic and naturally occurring ligands of the GLP-1R that have been proposed to act allosterically, that is, at sites distinct to the endogenous ligand (Figures 1B–1F). From a therapeutic perspective, ligands acting allosterically have several major advantages, including enhanced receptor subtype selectivity, the ability to simultaneously bind to the receptor with the endogenous ligand (restoring physiologically relevant temporal control), inducing a new repertoire of receptor conformations and therefore influencing receptor activity, and in particular for peptide-activated receptors, the potential for oral administration. With respect to the GLP-1R, allosteric ligands that enhance the insulinotropic effects of the system are desired [PAMs (positive allosteric modulators)].
At present, very few allosterically acting ligands have been identified for the GLP-1R. The Novo Nordisk compounds 2-(2′-methyl)thiadiazolylsulfanyl-3-trifluoromethyl-6,7-dichloroquinoxaline (compound 1) and 6,7-dichloro-2-methylsulfonyl-3-t-butylaminoquinoxaline (compound 2)  were the first non-peptide agonists identified for the GLP-1R, the latter of which demonstrates glucose-dependent insulin release via the GLP-1R [44,45]. Similarly, the cyclobutanes Boc5 and S4P stimulate GLP-1R activity, whereas the inability to fully inhibit 125I-GLP-1(7–36)NH2 binding suggests an allosteric mechanism of action [43,46]. Although S4P is only a partial agonist in GLP-1R-expressing immortal cell lines, Boc5 is a fully efficacious agonist with maximal responses for decreasing plasma glucose and reducing nutrient intake in obese mice, comparable with the native GLP-1 peptide . Unlike the compounds detailed above, quercetin (3,3′,4,5,7-pentahydroxyflavone) is a naturally occurring compound belonging to the flavonoid family, and has been observed to allosterically enhance GLP-1 efficacy and potency in intracellular Ca2+ mobilization in vitro [42,48]. The most recently identified allosterically acting synthetic ligand of the GLP-1R, 4-(3-(benzyloxy)phenyl)-2-(ethylsulfinyl)-6-(trifluoromethyl)pyrimidine (BETP or compound B), increases glucose-dependent insulin release from normal and diabetic human islet cells . Support for an allosteric mode of action is seen in the removal of the GLP-1R N-terminus, which does not influence the activity of the compound .
Although several additional synthetic small molecule ligands have been reported to increase GLP-1R-mediated cAMP production, increase plasma GLP-1 levels or decrease acute nutrient intake (reviewed in ), they have not been fully characterized pharmacologically, and thus it remains to be determined whether they are true GLP-1R ligands. Although allosteric modulation is fast gaining traction as a desired therapeutic approach to many disorders and conditions, there are many challenges in the identification and application of allosteric modulators. One most prominent complexity is that of probe-dependence, which describes the extent and direction of allosteric modulation on an orthosteric ligand (the probe), and is correlated with the co-operativity between the allosteric and orthosteric ligand in the system (reviewed in ). Indeed, this has already been observed at the GLP-1R in vitro and in vivo, with differential effects observed between orthosterically acting peptide ligands and the allosteric ligands BETP or compound 2, with preferential enhancement of signalling via oxyntomodulin relative to GLP-1 or exendin-4 [42,52]. Intriguingly, these allosteric compounds also markedly enhance the activity of the inactive metabolite of GLP-1 (GLP-1(9–36)NH2), suggesting that therapies directed to altering metabolite activity may be possible . Although the physiological importance of probe-dependence is yet to be determined, it illustrates an important consideration when pharmacologically characterizing allosteric ligands at receptors possessing multiple orthosteric ligands.
The rapidly increasing incidence of Type 2 DM and significant impact on quality-of-life demands the development of superior therapeutics for the management of the condition. Despite the GLP-1R having a pivotal role in glucose homoeostasis and currently being a highly valued therapeutic target, there are still significant knowledge gaps that limit understanding of this complex receptor system, particularly with respect to receptor structure and the nature of allosterism. In addition, the physiological importance of biased signalling and probe-dependence remains largely unexplored. Further research into these aspects of receptor function will have an impact on the future design and development of therapeutics for the management of Type 2 DM.
G-Protein-Coupled Receptors: from Structural Insights to Functional Mechanisms: A Biochemical Society Focused Meeting co-organized with Monash University held at Monash University Prato Centre, Prato, Italy, 12–14 September 2012. Organized and Edited by Bice Chini (CNR Institute of Neuroscience, Italy), Marco Parenti (University of Milano–Bicocca, Italy), David Poyner (Aston University, U.K.) and Mark Wheatley (University of Birmingham, U.K.)
dipeptidyl peptidase IV
epidermal growth factor receptor
extracellular-signal-regulated kinase 1 and 2
glucagon-like peptide-1 receptor
G-protein-coupled receptor kinase
positive allosteric modulator
This work was supported by the National Health and Medical Research Council of Australia (NHMRC) [grant numbers 519461 and 1002180], with Senior Research Fellowship awarded to A.C. and Principal Research Fellowship awarded to P.M.S.