Drugs that activate the glucagon-like peptide-1 (GLP-1) receptor are currently used clinically to treat type 2 diabetes mellitus (T2DM) and obesity (Nauck et al., 2021). In contrast, it is only recently that a drug, which also activates the receptor for the other incretin, glucose-dependent insulinotropic polypeptide (GIP), has been approved for the treatment of T2DM (i.e., tirzepatide, a dual GLP-1/GIP receptor agonist) (Le Roux et al., 2023). The primary action of GLP-1 and GIP is to potentiate insulin secretion in a glucose-dependent manner, a phenomenon referred to as the incretin effect (Mcintyre et al., 1964; Nauck et al., 1986). The incretin effect can account for at least 40% of the insulin produced postprandially in healthy individuals, and a loss of response to both endogenous GLP-1 and GIP, along with the subsequent deterioration of the incretin effect, is an early characteristic of T2DM (Holst et al., 2011; Gasbjerg et al., 2019). Pharmacological doses of GLP-1 can overcome this impairment in subjects with T2DM, but the same is not observed with GIP, even though in healthy individuals, GIP contributes more to the incretin effect than GLP-1 (Nauck et al., 1993; Gasbjerg et al., 2020).

The GLP-1 and GIP receptors (GLP-1R and GIPR, respectively) are closely related members of family B or the secretin class of G protein-coupled receptors (GPCRs) and share a high degree of sequence homology (Mayo et al., 2003; Graaf et al., 2016). Both receptors are expressed in pancreatic β-cells and other cell types, resulting in numerous pleiotropic effects of their peptide agonists. For example, GLP-1 decreases appetite and impairs glucagon secretion, and GIP regulates bone and adipocyte metabolism (Mayendraraj et al., 2022; Liu et al., 2022; Drucker and Holst, 2023). Studies showing that GIPR-knockout mice were resistant to weight gain when fed a high-fat diet suggested that blocking the actions of GIP may be a viable strategy for treating obesity (Miyawaki et al., 2002). This observation led to the development of (Pro3)GIP, an analog of GIP with a glutamic-acid-to-proline substitution at position 3, improving glucose homeostasis and preventing weight gain in animal models of diabetes and obesity (McClean et al., 2007; Gault et al., 2007). Although (Pro3)GIP was initially considered an antagonist, it was subsequently shown to be a low-potency agonist (Al-Sabah et al., 2014a; Sparre-Ulrich et al., 2016). Despite this finding, GIPR antagonists continued to be developed as candidates for treating T2DM and obesity. Maridebart cafraglutide (MariTide), a fully human monoclonal anti-human GIPR antagonist antibody conjugated to two GLP-1R agonist peptides, has recently exhibited robust weight loss in subjects with obesity in a phase 2 study (Véniant et al., 2024; Amgen, 2024). In contrast, tirzepatide, a dual GIPR/GLP-1R agonist, has been approved for the treatment of T2DM and additionally produces significant weight loss (Le Roux et al., 2023). It is currently unclear why both GIPR agonists and antagonists improve glucose control and reduce body weight. It is possible that the agonists act as functional antagonists by desensitizing the receptor, as demonstrated by Killion et al. (2020) and, more recently, by Davies et al. (2025).

The reasons why GLP-1, but not GIP, remains insulinotropic in T2DM are still unclear but may be related to the receptors’ differing signaling profiles (Yabe and Seino, 2011; Al-Sabah, 2016). GLP-1R and GIPR signal primarily through G_αs, resulting in an increase in intracellular cyclic adenosine monophosphate (cAMP). GLP-1R can also signal through G_αq, which may enable this receptor to function under conditions of hyperglycemia (Oduori et al., 2020), a hypothesis challenged by the observation that GIPR has also been shown to signal through G_αq (Manchanda et al., 2023). The two receptors also appear to differ in their kinetics and mechanisms of internalization and desensitization (Manchanda et al., 2023). GLP-1R undergoes rapid endocytosis following agonist stimulation, whereas GIPR has been reported to constitutively recycle between the cell surface and intracellular compartments (Shaaban et al., 2016; Mohammad et al., 2014). Although GLP-1R interacts with arrestin, its role in receptor endocytosis is still undetermined; nevertheless, knockdown experiments show that arrestin is involved in GLP-1-mediated insulin secretion (Sonoda et al., 2008). In contrast, how well GIPR interacts with arrestin is a subject of continuing investigation, with several conflicting studies in the literature (Al-Sabah et al., 2014b; Ismail et al., 2015; Gabe et al., 2018).

In this study, we investigated the role of the C-terminal tails of GLP-1R and GIPR in their differing signaling properties using chimeric receptors, where the C-terminal region of one receptor was replaced with that of the other. We hypothesized that replacing GIPR’s C-terminal tail with that of GLP-1R would result in a receptor that signals and internalizes like GLP-1R and vice versa.

In this study, we investigated the role of the C-terminal regions of GLP-1R and GIPR in cell signaling and endocytosis using chimeric receptors, where the C-terminal region of one receptor was substituted with that of the other. A chimeric receptor approach has been used previously by other groups to define the different regions of the glucagon family of GPCRs responsible for their selective binding of, and activation by, their endogenous ligands (Xiao et al., 2000; Runge et al., 2003). This two-step binding process, where the C-terminal region of the peptide ligand binds to the N-terminal domain of the receptor, allowing the N-terminal region of the peptide ligand to interact with the helical region of the receptor and resulting in receptor activation, has since been confirmed by numerous structures determined by X-ray crystallography and cryo-EM (Underwood et al., 2010; Jazayeri et al., 2017; Johnson et al., 2021). In contrast, much less is known about the molecular determinants that regulate G protein recruitment, receptor desensitization, internalization, and G protein-independent signaling within this family of receptors.

BRET-based mini-G protein recruitment assays were used to investigate the impact of swapping the C-terminal regions of GLP-1R and GIPR on G protein-dependent signaling. Both GLP-1 and GIP recruited Venus-mG_αs to their NLuc-labeled receptors in a dose-dependent manner. Although GIP was more potent in this regard, it was not significantly so. When measured as a fold-change over non-stimulated, GIP was significantly less effective than GLP-1 at recruiting Venus-mG_αs to its receptor. A similar pattern was observed when Venus-mG_αq recruitment was investigated, although in this case, no significant difference was found in E _MAX (Table 1). GIPR’s reduced ability to recruit both G_αs and G_αq compared to that of GLP-1R has been reported in previous studies, both in HEK-293 T cells and INS-1 832/3 cells (Jones et al., 2021; Manchanda et al., 2023). These results are initially surprising as GIP contributes more than GLP-1 to the incretin effect in healthy individuals (Gasbjerg et al., 2020). Manchanda et al. (2023) observed no difference between GIP- and GLP-1-stimulated cAMP production or insulin release from INS-1 832/3 cells despite reduced G_αs recruitment to GIPR. A possible explanation for these conflicting results is that GIPR is subject to less desensitization than GLP-1R.

Substituting the C-terminal tail of GLP-1R with that of GIPR resulted in a small but significant increase in the extent (E _MAX ), but not potency, to which GLP-1 stimulated Venus-mG_αs recruitment to this receptor (Figure 2A; Table 1). Thompson and Kanamarlapudi (2015a) identified residues 419–430 of GLP-1R’s C-terminus as being required for coupling to G_αs. Although this region is not completely conserved in GIPR, at least two residues (R411 and L412) are conserved. Nonetheless, based on this study, the GLP-1/GIPR chimeric receptor would be expected to be a poorer recruiter of Venus-mG_αs than the wild-type GLP-1R. A possible explanation for the slight increase in E _MAX observed with GLP-1/GIPR is that this chimeric receptor undergoes agonist-induced endocytosis at a slower rate (discussed below). The reciprocal substitution had no significant effect on Venus-mG_αs recruitment to GIPR, and neither substitution affected the ability of GLP-1 or GIP to recruit Venus-mG_αq to either the wild-type or the chimeric receptor (Figures 2C,E,H; Table 1). An important caveat is that we used mini-G proteins in our assays, which do not consider the role of the βγ-subunits.

For many GPCRs, homologous desensitization first requires the phosphorylation of the active state of the receptor by the G protein-coupled receptor kinase (GRK) family of kinases. This is then followed by the binding of arrestin, which sterically hinders coupling to G proteins, effectively desensitizing the receptor (Krasel et al., 2005). Arrestin can also act as a scaffolding protein, linking the receptor to the cellular endocytic machinery and facilitating internalization (Ferguson, 2001). It is now appreciated that arrestin can also act as a signaling molecule in its own right, linking the receptor to G protein-independent signaling pathways such as mitogen-activated protein kinase pathways (Luttrell and Getsy-Palmer, 2010). Studies performed in various cell types have shown that the activation of GLP-1R by GLP-1 results in the rapid recruitment of arrestin (Jorgensen et al., 2005; Manchanda et al., 2023). Sonoda et al. (2008) demonstrated that the knockdown of arrestin-2 in INS-1 cells impairs GLP-1-mediated insulin secretion, suggesting that arrestin is an integral component of GLP-1R’s signaling pathway. In contrast, arrestin recruitment to GIPR and its potential consequences are less well understood. Although some studies report that GIPR can recruit arrestin, others show that GIPR is either a very poor recruiter or unable to recruit arrestin to any detectable level at all (Al-Sabah et al., 2014b; Ismail et al., 2015; Gabe et al., 2018). The reasons for this disparity are unknown but may be related to the cell context and assay set-up (Al-Sabah et al., 2020). In this study, we could demonstrate a robust, dose-dependent recruitment of NLuc-labeled arrestin to GLP-1R but not to GIPR. As the intracellular C-terminal tail plays an essential role in arrestin binding for many GPCRs and because this region differs between GLP-1R and GIPR in terms of length and potential phosphorylation sites, we hypothesized that substituting the C-terminal tail of GIPR with that of GLP-1R would result in a receptor capable of recruiting arrestin (Tobin et al., 2008). We initially observed no change in arrestin recruitment to GLP-1R with C-terminal substitution (Figure 3A). In this case, the BRET pair was an SYFP2-labeled receptor and an NLuc-labeled arrestin. However, when a different BRET pair was used (RLuc-8-labeled receptor and SYFP2-labeled arrestin), a small but significant decrease in the extent of arrestin recruitment to the chimeric receptor was observed (Figure 3C). This discrepancy between BRET assays could possibly be due to the different sizes of luciferases used in each case (NLuc being 19 kDa and RLuc8 being 36 kDa) (England et al., 2016). Nonetheless, the reciprocal substitution had no significant effect on GIPR’s ability to recruit arrestin, regardless of which BRET-pair was used. This would suggest that while substituting GLP-1R’s C-terminal tail with that of GIPR could moderately impair arrestin recruitment to the chimeric receptor, regions other than the C-terminal tail are more likely to be responsible for the difference in the ability of GLP-1R and GIPR to recruit arrestin observed (Figures 3A, C, E; Table 2).

Following agonist stimulation, most GPCRs undergo endocytosis (Ferguson, 2001). This process serves to regulate the number of receptors expressed at the cell surface and attenuate the cellular response following repeated exposure to an agonist. However, some GPCRs, including GLP-1R and GIPR, continue to signal once internalized (Kuna et al., 2013; Ismail et al., 2016). Most GPCRs undergo endocytosis via either a clathrin- or caveolae-dependent pathway (Ferguson, 2001). GLP-1R undergoes rapid internalization following agonist stimulation, which occurs via clathrin-coated pits (Widmann et al., 1995). Clathrin-mediated endocytosis (CME) usually requires arrestin to act as a scaffolding protein to link the receptor to the β2-adaptin subunit of adapter protein 2 (AP2), which then targets the receptor to clathrin-coated pits (Kaksonen and Roux, 2018). Interestingly, agonist-induced endocytosis via clathrin-coated pits is an arrestin-independent process for GLP-1R (Syme et al., 2006; Thompson and Kanamarlapudi, 2015b). It is possible that both GLP-1R and GIPR can bind directly to the AP2 complex based on the existence of a potential AP2 binding site (YxxΦ) at the interface of transmembrane helix 7 and the C-terminal tail (Kelly et al., 2008). GLP-1R also contains a caveolin-binding motif in intracellular loop 2 and internalizes in a caveolin-dependent manner (Puddu and Maggi, 2021). Based on these conflicting studies, it is possible that GLP-1R can undergo endocytosis through more than one pathway.

The published data regarding agonist-induced internalization of GIPR are just as conflicting as, if not more so than, that for GLP-1R. Some studies demonstrate an arrestin-independent but AP2- and clathrin-dependent internalization mechanism, while others suggest that arrestin is required (Ismail et al., 2015; Gabe et al., 2018). Roed et al. (2014) showed no net change in cell surface expression for GIPR following agonist stimulation. Although this could be interpreted as indicating no internalization, it is also consistent with studies that show that GIPR internalizes slowly but rapidly recycles to the cell surface in a constitutive manner (Mohammad et al., 2014).

Using a BRET-based assay, our data show that GLP-1R undergoes rapid agonist-stimulated endocytosis, whereas a loss of GIPR cell surface expression occurred at a significantly slower rate. The replacement of the C-terminal tail of GIPR with that of GLP-1R significantly increased the rate of receptor endocytosis, which could suggest that this region of the receptor may be involved in GIPR’s reported constitutive recycling. The reciprocal substitution significantly decreased GLP-1R’s rate of internalization but not to the same extent to that of the wild-type GIPR. Interestingly, the kinetics of wild-type GLP-1R internalization fitted best to a two-phase exponential decay model, while all other receptors tested fitted best to a one-phase exponential decay model (Figures 5A–C; Table 2). Taken together, these observations would support the hypothesis that GLP-1R can internalize by more than one mechanism and that regions other than the C-terminal tail may be involved in this process. Moreover, this hypothesis agrees with results reported during the preparation of this manuscript that demonstrate that GLP-1R undergoes endocytosis independent of arrestin and is mediated by both clathrin- and caveolae-dependent mechanisms (Moo et al., 2025).

Substituting the C-tail region of GLP-1R with that of GIPR had no effect on either total expression or cell surface expression, whereas the reciprocal substitution impaired GIPR’s cell surface expression (Supplementary Figure 1D; Figures 4A, B). GIP/GLP-1R’s poor cell surface expression may be a result of increased constitutive endocytosis; alternatively, it may be due to this receptor being retained in either the endoplasmic reticulum or Golgi due to misfolding. The poorer total and cell surface expressions observed with GIP/GLP-1R would be expected to impair agonist-induced endocytosis; however, we observe the opposite. It is possible that this chimeric receptor’s poorer cell surface expression masks an even greater effect on the rate of internalization.

Our results using C-tail swapped GIP receptors are, in part, in agreement with the findings of a recent study that investigated the role of GIPR’s C-terminal tail in signaling and internalization in different species. Rodent GIPR was shown to be less susceptible to internalization and desensitization, as well as weaker at recruiting arrestin, than the human receptor. When the C-tails were swapped between species, there was no significant impact on G protein-mediated signaling, but arrestin recruitment to the human receptor was impaired when its tail was substituted with that of either the rat or the mouse receptor. In the reciprocal experiments, swapping the C-tail of either the rat or mouse receptor with that of the human GIPR improved arrestin recruitment to the receptor and internalization/desensitization (Gasbjerg et al., 2024). Although we observed minimal effects on G protein recruitment with our chimeric receptors and no improvement in GIPR’s arrestin recruitment with GLP-1R’s C-terminal tail, we observed significant differences in rates and extents of internalization.

GLP-1R agonists have been approved for the treatment of T2DM since 2005, but it was only with the approval of tirzepatide (a dual GLP-1R/GIPR agonist) in 2022 that a drug acting on GIPR was used clinically (Le Roux et al., 2023). To what extent GIPR agonism contributes to tirzepatide’s efficacy is still unclear (Gasbjerg et al., 2023). However, it acts as a G protein-biased agonist at GLP-1R, stimulating less internalization than GLP-1 (Willard et al., 2020). Recently, there has been a greater appreciation for targeting GLP-1R endocytosis as a means to improve efficacy. GLP-1R agonists that elicit less internalization produce less nausea, an adverse effect that often limits dosing (Jones et al., 2018). Based on these findings, a greater understanding of the mechanisms by which GLP-1R and GIPR undergo endocytosis is warranted.

The present study may be extended by measuring the interaction between the wild-type and chimeric receptors and arrestin-2 (β-arrestin-1). It would be interesting to investigate whether the C-terminal tail plays a role in differentially recruiting different isoforms of arrestin. In addition, exploring downstream signaling pathways such as cAMP and ERK would enhance our understanding of the role of these receptors’ C-terminal regions in cell signaling.

In summary, we demonstrate that the C-terminal tails of GLP-1R and GIPR are not the key determinants for their differing ability to recruit G proteins and arrestin. Substituting GIPR’s C-terminal tail with that of GLP-1R significantly increased both the rate and extent of agonist-mediated endocytosis, suggesting a role for GIPR’s C-terminal region in either its slower internalization or rapid recycling back to the plasma membrane. The reciprocal substitution significantly impaired GLP-1R internalization, but not to the same extent as that of wild-type GIPR. These data support the hypothesis that regions other than GLP-1R’s C-terminal tail may also be involved in endocytosis.