The glucagon-like peptide-1 receptor (GLP-1R) is an important regulator of blood glucose homeostasis. Ligand-specific differences in membrane trafficking of the GLP-1R influence its signalling properties and therapeutic potential in type 2 diabetes. Here, we have evaluated how different factors combine to control the post-endocytic trafficking of GLP-1R to recycling versus degradative pathways. Experiments were performed in primary islet cells, INS-1 832/3 clonal beta cells and HEK293 cells, using biorthogonal labelling of GLP-1R to determine its localisation and degradation after treatment with GLP-1, exendin-4 and several further GLP-1R agonist peptides. We also characterised the effect of a rare GLP1R coding variant, T149M, and the role of endosomal peptidase endothelin-converting enzyme-1 (ECE-1), in GLP1R trafficking. Our data reveal how treatment with GLP-1 versus exendin-4 is associated with preferential GLP-1R targeting towards a recycling pathway. GLP-1, but not exendin-4, is a substrate for ECE-1, and the resultant propensity to intra-endosomal degradation, in conjunction with differences in binding affinity, contributes to alterations in GLP-1R trafficking behaviours and degradation. The T149M GLP-1R variant shows reduced signalling and internalisation responses, which is likely to be due to disruption of the cytoplasmic region that couples to intracellular effectors. These observations provide insights into how ligand- and genotype-specific factors can influence GLP-1R trafficking.
The glucagon-like peptide-1 receptor (GLP-1R) is a class B G protein-coupled receptor (GPCR) with pleiotropic roles in glucose homeostasis and energy metabolism [1]. Well-characterised effects of GLP-1R activation include augmentation of glucose-stimulated insulin secretion from pancreatic beta cells, suppression of appetite and slowing of gastric emptying. These actions have been harnessed therapeutically for the treatment of type 2 diabetes (T2D) and obesity, with several pharmacokinetically optimised GLP-1R agonists now in routine clinical use [2]. Further applications of GLP-1R agonism have been proposed, including treatment of steatohepatitis [3] and neurodegenerative diseases [4].
Whilst in vitro optimisation of GLP-1R ligands has traditionally focussed on high binding affinity or signalling potency, recent work has highlighted that the balance between recruitment of signal-initiating versus signal-terminating effectors can allow low-affinity “biased” GLP-1R agonists to achieve high therapeutic efficacy [5,6,7,8]. Connected to this, agonist-specific trafficking patterns of the GLP-1R between different endomembrane compartments allow fine-tuning of the duration and spatial origin of intracellular signalling responses [9,10,11,12].
Therapeutic GLP-1R peptide ligands are typically based on the sequence of either the primary endogenous agonist, GLP-1(7-36)-NH 2, or its homologue exendin-4, a 39 amino acid peptide isolated originally from the venom of the Gila monster [13]. Whilst these are both high-potency full agonists at the GLP-1R, a number of studies have reported differential signal bias and altered trafficking characteristics of these core ligands [11,14,15,16]. In particular, recycling of internalised GLP-1R back to the plasma membrane is slower after exendin-4 treatment, resulting in more pronounced desensitisation in challenge/washout/re-challenge experiments [11,15,17]. The reasons for these differences have not been fully elucidated, but it has been suggested that the trafficking behaviours of GLP-1R ligands may relate to either intra-endosomal ligand-receptor dissociation [6] or to differential ligand sensitivity to intra-endosomal degradation by the endopeptidase endothelin converting enzyme-1 (ECE-1) [18]. Many prior studies have been performed using GLP-1R expressed in heterologous systems and, with increasing recognition that GPCR behaviours can be tissue-specific [19], their findings ideally should be confirmed using beta cell models.
In this study, we have explored the post-endocytic sorting of GLP-1R after treatment with GLP-1 or exendin-4. We focus in particular on ligand-specific differences in receptor degradation, which may in turn contribute to receptor availability and insulin secretory responses during prolonged pharmacological stimulation. We investigate the importance of intra-endosomal ligand proteolysis by ECE-1 through the use of pharmacological inhibition and a panel of chimeric peptide ligands with varying degrees of ECE-1 resistance. We further dissect this phenomenon from the contribution of binding affinity and β-arrestin recruitment by concurrently testing derivatives of these peptides bearing N-terminal substitutions. Finally, we examine the trafficking responses of a natural GLP-1R variant, T149M, previously associated with T2D [20] and known to show impaired intracellular cyclic adenosine monophosphate (cAMP) signalling responses [21].
GLP-1R is typically described as a fast-internalising receptor [22]. To provide evidence that this is the case for endogenous GLP-1R in its native cellular context, we first visualised uptake of fluorescent GLP-1 and exendin-4 conjugates (see Supplementary Figure S1) in dispersed primary islet cells, with GLP-1R expression marked by the fluorescent reporter GCaMP3 under Cre-dependent control [23] (Figure 1A). Both ligands were extensively taken up into punctate endosome-like structures. As agonist and receptor may dissociate on entering the endocytic pathway, to study the post-endocytic trafficking of GLP-1R in a beta cell context we used cultured INS-1 832/3 cells stably expressing N-terminally SNAP-tagged GLP-1R (referred to hereafter as INS-1-SNAP-GLP-1R cells) [24]. A cleavable benzyl guanine (BG)-conjugated fluorescent probe, BG-S-S-649, allowed quantification of internalised and subsequently recycled SNAP-GLP-1R by high content microscopy, as previously described [11]. Pronounced GLP-1R internalisation was observed with both ligands at high concentrations, with exendin-4 showing slightly increased potency (Figure 1B). Moreover, following agonist washout, GLP-1R recycled more slowly after exendin-4 treatment over a wide concentration range. We also visualised post-endocytic sorting of SNAP-GLP-1R by electron microscopy after treatment with either ligand in MIN6 mouse beta cells, with the receptor surface-labelled using BG-biotin and streptavidin-gold prior to treatment. This indicated a trend towards greater localisation of GLP-1R in late endosomes and lysosomes after exendin-4 treatment compared to GLP-1 (Figure 1C).
To examine the consequences of these apparent differences in post-endocytic sorting on GLP-1R degradation, experiments were performed in INS-1-SNAP-GLP-1R cells treated with cycloheximide to arrest protein translation. After continuous exposure to 10 nM of either ligand, applied in a time series in reverse order, a cell permeating SNAP-tag probe, BG-conjugated Oregon Green (BG-OG) [25], was used to label total residual intact SNAP-GLP-1R. Over an 8 h time-course, a steady reduction of BG-OG labelling was apparent with both ligands, suggesting destruction of the SNAP moiety (and by inference, the GLP-1R itself), but this was faster and more pronounced with exendin-4 treatment (Figure 1D). Note that BG-OG labelling of internalised SNAP-GLP-1R after 30 min treatment with 100 nM agonist was similar for both agonists, suggesting that the fluorescent signal reduction seen in Figure 1D did not represent the effect of endosomal pH on fluorophore emission intensity per se (Supplementary Figure S2). Moreover, densitometric Western blotting analysis following 6 h stimulation with 10 nM agonists in INS-1-SNAP-GLP-1R cells confirmed greater SNAP-GLP-1R degradation with exendin-4 (Figure 1E). Additionally, 4 h after a 30 min pulsed agonist treatment and washout, loss of GLP-1R signal was more pronounced throughout the concentration range tested with exendin-4 compared to GLP-1 (Figure 1F). Because residual extracellular agonist was removed in the latter experiment, the result is likely to represent the differential effects of internalised GLP-1R-exendin-4 versus -GLP-1 complexes on post-endocytic targeting.
It was recently reported that GLP-1R trafficking in HEK293 cells is regulated by the activity of ECE-1 [18], a pH-sensitive endopeptidase known to degrade various GPCR peptide ligands once they enter the endocytic system. The resultant dissociation of the ligand-receptor complex typically prompts preferential targeting of the receptor to a recycling pathway [26,27]. Due to the known differences in proteolytic stability of GLP-1 versus exendin-4 to the related neutral endopeptidase NEP24.11 [28], we aimed to experimentally verify whether ECE-1 sensitivity might underpin the observed differences in post-endocytic sorting with these two ligands in beta cells. We first performed in vitro digestion assays of GLP-1 using ECE-1 and ECE-2 under neutral and acidic pH conditions, to mimic the extracellular and intra-endosomal environments, respectively. Here, high performance liuid chromatography (HPLC) measurements indicated extensive peptide degradation, particularly by ECE-1 at low pH (Figure 2A). Exendin-4 was highly resistant to ECE-1 and ECE-2 degradation, remaining entirely intact after a 24 h incubation with either enzyme at low pH (Figure 2B).
