GIP-derived GIP receptor antagonists – a review of their role in GIP receptor pharmacology

Surprisingly, agonists, as well as antagonists of the glucose-dependent insulinotropic polypeptide receptor (GIPR), are currently being used or investigated as treatment options for type 2 diabetes and obesity – and both, when combined with glucagon-like peptide 1 receptor (GLP-1R) agonism, enhance GLP-1-induced glycemia and weight loss further. This paradox raises several questions regarding not only the mechanisms of actions of GIP but also the processes engaged during the activation of both the GIP and GLP-1 receptors. Here, we provide an overview of studies of the properties and actions of peptide-derived GIPR antagonists, focusing on GIP(3-30)NH 2 , a naturally occurring N-and C-terminal truncation of GIP(1 (cid:0) 42). GIP(3 (cid:0) 30)NH 2 was the first GIPR antagonist administered to humans. GIP(3 (cid:0) 30)NH 2 and a few additional antagonists, like Pro3-GIP, have been used in both in vitro and in vivo studies to elucidate the molecular and cellular consequences of GIPR inhibition, desensitization, and internalization and, at a larger scale, the role of the GIP system in health and disease. We provide an overview of these studies combined with recent knowledge regarding the effects of naturally occurring variants of the GIPR system and species differences within the GIP system to enhance our understanding of the GIPR as a drug target.


Introduction
Uni-molecular compounds targeting more than one receptor are currently being developed to treat metabolic diseases.An example is the co-agonist tirzepatide, which binds to the glucagon-like peptide 1 receptor (GLP-1R) as well as the glucose-dependent insulinotropic polypeptide receptor (GIPR) that successfully improves glycaemic control and reduces body weight in patients with type 2 diabetes [1][2][3][4].The peptide hormone GIP is released postprandially from enteroendocrine K cells, primarily in the upper small intestine.It potentiates glucose-induced insulin secretion and is one of the incretin hormones of the human body [5][6][7].Moreover, endogenous GIP is suggested to be involved in bone homeostasis by reducing bone resorption [8][9][10], postprandial stimulation of intestinal blood flow [11,12], and lipid deposition in adipose tissue [13,14].Based on these physiological actions, compounds have been developed that target the GIPR to treat type 2 diabetes and obesity [1][2][3][4]15].Although initial infusions of GIP in patients with type 2 diabetes and patients with obesity revealed a severely reduced insulinotropic effect compared to healthy individuals [16][17][18][19][20], various multi-targeting therapeutic compounds in clinical development successfully target the GIPR in combination with other gut hormone receptors [4,21,22].However, whether to stimulate or block the GIPR to obtain better control of metabolic diseases is still unanswered [23,24].Recently, a phase 2 study with a GIPR agonist combined with a potent, well-established GLP-1R agonist was completed without any beneficial clinical effects of the GIPR agonist in patients with type 2 diabetes [25].Considering other successful multiple targeting strategies, to understand the actions of each therapeutic compound targeting the GIPR, the complex GIPR signalling, and recycling must be further studied and considered [26][27][28].
GIP is mainly released as a 42 amino acid peptide (GIP(1− 42)), a product of pro-GIP cleaved by prohormone convertase (PC) 1/3.Expression of another convertase, PC-2, in the gut epithelium may lead to differential cleavage of the precursor, releasing minimal amounts of GIP(1− 30)NH 2 to the circulation [29][30][31].Interestingly, the biological activities of GIP(1− 30)NH 2 and GIP 1− 42 are similar in vitro and in vivo [32][33][34][35][36].Both compounds are substrates for the ubiquitous and circulating enzyme, dipeptidyl peptidase-4 (DPP-4), that cleaves the peptides between positions 2 and 3, resulting in large amounts of GIP (3− 42) and low amounts of GIP(3− 30)NH 2 in the circulation.The metabolites have little agonist activity.The other incretin hormone, GLP-1 (7-36)NH 2 , is also cleaved by DPP-4, leading to the formation of GLP-1(9− 36)NH 2 .For both hormones, this is a rapid process, resulting in apparent half-lives of approximately 2 minutes for GLP-1 and 7 minutes for GIP in the circulation of humans [37].The inactivating N-terminal modification lead to the search for naturally occurring GIPR antagonists [15,35].The abundant metabolite GIP (3− 42) was proven early on to be an antagonist of the GIPR but with low affinity, requiring very large doses for efficient GIPR inhibition [38][39][40].The metabolite, GIP(3− 30)NH 2, however, turned out to have a much higher affinity for the receptor and >20-fold improved inhibitory properties [35].This compound was, therefore, selected for studies of the effects of blocking the GIP system in vitro and in vivo.This review will provide an overview of the studies conducted with these GIP-derived GIPR antagonists, most of which have been conducted with GIP(3− 30)NH 2 .We also discuss the implications of naturally occurring variants of the GIPR and the differences between the human and rodent GIP systems, which are important for the design of agonists and antagonists.Ultimately, we discuss the therapeutic potential of the antagonists and their impact on our understanding of the GIP system in health and disease.

Overview of human GIPR signalling at the cellular level
The GIPR belongs to class B1 of G protein-coupled receptors [41] but differs in many ways from other closely related class members.For instance, in terms of ligand selectivity, the GIPR is more selective than the more promiscuous GLP-1R and is only potently activated by GIP (1− 42) and the C-terminally cleaved variant of this (GIP(1− 30)NH 2 ) [36] while the GLP-1R can be rather potently activated by several endogenous peptides, including oxyntomodulin [42][43][44], glucagon [45], Fig. 1.Cellular effects of glucose-dependent insulinotropic polypeptide receptor (GIPR) agonists and antagonists following acute administration (A, B) and chronic GIPR targeting (C, D).A, GIPR agonist binding and subsequent G protein recruitment lead to activating signalling pathways via predominantly cAMP but also beta arrestin recruitment, which inhibits G protein association and induces internalization of receptors in endosomes, where GIPRs signals continuously and are rapidly recycled to the cell surface.B, GIPR antagonist binds and blocks the binding site for the GIPR agonist.C, Chronic GIPR activation by agonists could lead to high receptor desensitization, low receptor number on the cell surface, and, therefore, a lower degree of activation (functional antagonism).D, Chronic GIPR antagonist exposure traps receptors on the cell surface, which are then available for agonist stimulation or antagonist binding in a new agonist-antagonist equilibrium.and GLP-2 [46,47].GLP-2 also activates the GIPR but has a very low potency (EC 50 >100 nM) [48].
Like other class B1 receptors, the GIPR signals mainly via the G protein G αs , although a minor G αq recruitment has been reported [26].Likewise, the GIPR binds beta arrestins to terminate G protein-mediated signalling and initiate receptor internalization, which, in the case of the GIPR, strongly depends on the coupling to arrestins [49][50][51] (Fig. 1A).Compared to the closely related GLP-1R and GLP-2R, GIPR displays weaker arrestin recruitment (less potent and with lower efficacy) and less internalization, and its strong arrestin dependency for internalization is neither shared by the GLP-1R nor the GLP-2R [52,53].Along the same lines, GIPR recycles faster to the cell surface upon internalization than GLP-1R, but once the receptor is internalized, G αs -mediated cAMP production from early endosomes continues to rise [54,55], as also seen for GLP-1R [56].However, the GIPR displays higher internalization in its basal (agonist-free) state than the GLP-1R [50,57,58].Notably, upon binding of the novel co-agonist tirzepatide, the human GIPR is desensitized to the same extent as when it binds GIP, whereas tirzepatide does not induce the same desensitization of the GLP-1R (as GLP-1) [54,57,59].This biased action effect on the GLP-1R is considered a major advantage and at least partly explains its clinical success, whereas the role of the GIPR targeting is unclear (as reviewed recently [59][60][61]).
In terms of the interaction between GIP(1− 42) (in the following only "GIP") and the GIPR, the binding kinetics differ remarkably from those of GLP-1 to the GLP-1R, as GIP has a higher on-rate as well as off-rate resulting in short ligand residence time [62,63].Thus, despite similar affinity (K D values) in the single-digit nanomolar range, calculated as K off /K on , the agonist residence time (defined as 1/K off ) differs between these two receptors.The GLP-2R behaves more like the GLP-1R with a slower K on and K off and, thus, longer residence time than GIP at the GIPR [47].These differences in binding kinetics may contribute to the different signalling and desensitization patterns observed for the three receptors [23,64].
Thus, the GIPR is differently regulated regarding signalling initiation and termination, receptor desensitization, internalization, and resensitization compared to close homologs.In a biological context, and again compared to the GLP-1R, the GIPR signalling also appears to be more sensitive to changes in cells and tissues in which it is expressed, as reflected in the impaired response to GIP in diabetic beta cells [65] and in the subcutaneous adipose tissue of individuals with obesity [66,67].The reduced insulinotropic effect of GIP in patients with type 2 diabetes [68] is the historical reason for focusing on GLP-1-rather than GIP-based therapies [18,69,70].
The special molecular pharmacological properties of the GIPR, combined with its sensitivity to changes in the intracellular and extracellular milieu, may partly explain the similar metabolic effects of compounds with opposite GIPR actions: agonists and antagonists.For both ligands, the effects on body weight of people with obesity or diabetes are most pronounced when combined with GLP-1R agonists [23,[71][72][73] (Fig. 1).For example, exposure to the antagonist GIP(3− 30)NH 2 results in enhanced GIPR surface expression due to an inhibition of GIPR internalization [51]; hence, the antagonist resensitizes the receptor for subsequent agonist activation (Fig. 1D).The increased number of receptors available could potentially be occupied by agonists (i.e., activated) rather than antagonists in this new agonist-antagonist equilibrium (Fig. 1D).This mechanism has also been proposed to contribute to the therapeutic beneficial effects of antagonistic antibodies targeting the GIPR [23,71,74].In contrast, long-term agonist exposure would result in receptor desensitization as seen in adipose tissue and, hence, would be proposed to result in functional antagonism due to the reduced number of receptors available for activation [75][76][77] (Fig. 1C).
The receptor of the other main incretin hormone, the GLP-1R, has important functions in the beta cell, the gastrointestinal tract, and the enteric nervous system, as well as central metabolic functions in humans [70,78].In beta cells, the GIPR and GLP-1R share the insulinotropic pathways.When co-expressed in cellular systems, the signalling of one may be affected by the expression of the other receptor (Fig. 2).Thus, heterologous co-expression of the GLP-1R and GIPR in vitro showed that GIPR expression suppressed GLP-1R signalling and surface expression, whereas the GIPR system was unaffected by the presence of GLP-1R [79][80][81].One explanation for this interference could be receptor oligomerization.The GLP-1R is proposed to form homodimers crucial for its function, as disruption of the dimers impairs receptor activity [80,82].Bioluminescence (BRET)-based studies suggest GLP-1R and GIPR may form heterodimers upon GLP-1 stimulation, along with impairments of GLP-1R signaling.GIP, in contrast, impairs heterodimerization, reflected in reduced proximity between GLP-1R and GIPR [79][80][81]83,84].Hence, heteromerization of GLP-1R and GIPR induced by GLP-1 could reduce GLP-1R function by disrupting the functionally important GLP-1R homodimer.However, the relevance of heteromerization between GLP-1R and GIPR is still not completely understood.
Shared intracellular pathways may also affect the balance of GLP-1R versus GIPR signalling.Both receptors couple strongly to G αs , whereas the GLP-1R has a stronger coupling to G αq than the GIPR [26,28,85,86].The negative impact of GIPR on the GLP-1 system could be due to G protein scavenging (mainly G αs ) by an activated GIPR, leaving GLP-1R behind with only secondary pathways (like G αq ).Keeping the GIPR inactive with an antagonist could rebalance signaling pathways.Likewise, the GLP-1R is less internalized in the presence of the GIPR, possibly due to shared intracellular proteins [81,86].
Though potentially relevant, the receptor interplay is still poorly understood.
Both infusion and injection of GIP in healthy individuals will lead to insulin secretion during eu-and hyperglycaemia [5,16,48,87,88] but the response is impaired in individuals with obesity and markedly reduced in type 2 diabetes.Similar to stimulation of the GLP-1R, GIPR stimulation does not seem to cause insulin-induced hypoglycaemia [5,88,89,91], consistent with the effect being a potentiation of glucose-induced insulin secretion.GIPR stimulation also results in glucagon secretion from the pancreatic alpha cells, both in healthy individuals and patients with type 2 diabetes.This is, of course, not desirable in patients with type 2 diabetes, leading to hyperglucagonemia [19,20] and even higher plasma glucose levels but it may contribute to a glucose-stabilizing effect during low plasma glucose levels.The contribution of GIPR-mediated glucagon secretion to the actions of GIPR-GLP-1R co-agonists or tri-agonists, such as retatrutide, is so far unknown [92,93].
The long-term effects of GIP on bone tissue have yet to be evaluated.The postprandial secretion of GIP seems to promote bone formation.Assessed by the bone resorption marker carboxy-terminal telopeptide of type 1 collagen (CTX), the bone resorption is reduced min.25% by endogenous GIP [8].Mimicking a more therapeutic situation, continuous GIPR stimulation for six days (in patients with type 1 diabetes) seems to blunt the effect in bone tissue [89].Furthermore, bone-related adverse effects have not yet been reported during treatment with GIPR-targeting compounds.However, genetic GIPR missense variants (E288G and E354Q) have been associated with more fragile bone tissue (E288G with lower bone mineral density [94]) or higher fracture risk (E354Q with increased non-vertebral fracture risk in postmenopausal women [95]).E288G is a loss-of-function (LoF) mutation causing impaired GIPR activity, while E354Q mimics LoF mutations because of trafficking abnormalities, resulting in receptor desensitization.However, in a recent study with a much larger sample size, the same two genetic GIPR variants and R190Q were neither associated with increased fracture risk nor lower bone mineral density [96].Nevertheless, possible bone effects will have to be evaluated for therapeutics targeting GIPR.
Adipose tissue and the cardiovascular system are additional targets for GIPR agonists.The effects on humans are so far unclear.GIPR stimulation increases blood flow in adipose tissue [97], and the splanchnic bed [12], but the consequences and possible adverse effects of targeting the GIPR are not elucidated.
Lastly, several animal studies have shown the expression of GIPR in the central nervous system, and the effects of GIP in appetite-related cerebral areas and memory functions have been reported in rodents [97][98][99].The human relevance of the animal findings is currently unclear, and no human data support that GIPR targeting affects the central nervous system to an extent comparable to GLP-1R activation [91].
Therapy with GLP-1R agonists is consistently associated with gastrointestinal adverse effects, including nausea, diarrhoea, and vomiting [100].These effects are undoubtedly related to the activation of central nervous system receptors, including receptors in the area postrema.In comparison, although less intensely investigated and not performed in longer periods, GIP administration does not seem to be associated with gastrointestinal side effects [89].On the contrary, recent experimental and clinical studies suggest that simultaneous GIPR activation may dampen the gastrointestinal side effects of GLP-1 administration [101][102][103][104].
Here, phenotypic studies combined with genome-wide association studies (GWAS) in human carriers of LoF variants have supported protection against obesity [119].Thus, a GWAS study combining data from 718,734 individuals identified variants of the GIPR gene among the few associated with a lower BMI [119].The two identified GIPR missense variants resulting in the amino acid substitutions R190Q and E288G were revealed as LoF with respect to G αs -mediated signalling and arrestin coupling [94].Additionally, the most commonly occurring missense variant of the GIPR, E354Q, with an allele frequency of 19%, is also associated with lower BMI [96,[120][121][122].At the molecular level, E354Q acts similarly to wild-type GIPR in receptor activation and GIP binding.However, it has a faster internalization rate and enhanced receptor downregulation with slower recycling to the plasma membrane [27,62,123], mimicking a LoF variant.Another important lesson learned from the GWAS studies of class B1 receptors, supporting GIPR antagonists as future therapeutics, is the lesser deleteriousness and tolerability of LoF GIPR variants, compared to LoF variants of, for instance, the GLP-1R and the glucagon receptor [124].
Further genetic studies identified 168 naturally occurring missense variants in the human GIP gene from three independent cohorts comprising ~720,000 individuals.These studies showed that the fully processed GIP hormone sequence is more protected against mutations than the rest of the precursor protein with high species-orthologous and population-specific conservation of the GIP peptide sequence, suggesting strong evolutionary constraints to preserve the GIP peptide sequence.However, compared to the peptide-coding regions of the proglucagon gene, variants of the GIP peptide sequence occur more often, again pointing towards variants of the GIP system as more tolerable [125,126].

GIP-based GIPR antagonists -N-and C-terminal truncations
GIPR activation by an agonist requires ligand binding and stabilization of the receptor in active receptor conformations [127] (Fig. 1).Initial binding relies on the interaction of alpha-helical parts of GIP to the extracellular domains of the GIPR, while receptor activation occurs upon docking of the N-terminus of GIP into the main ligand binding Fig. 2. Interactions between the glucose-dependent insulinotropic polypeptide receptor (GIPR) and glucagon-like peptide 1 receptor (GLP-1R) when both are expressed in cellular systems reflecting the pancreatic beta cell.Besides their own activation and recycling pathways, GIPR and GLP-1R are suggested to interact (heterodimerize), and the presence of GIPR also negatively impacts the signalling of GLP-1R.In contrast, the GIPR is not affected by the presence of GLP-1R.

M.M. Rosenkilde et al.
Peptides 177 (2024) 171212 pocket in the transmembrane region [127] (Fig. 3).By modifying the N-terminus of the GIP peptide (GIP(1− 42)), the receptor binding may be intact, whereas the activation is missing or interrupted.A competitive antagonist will hinder agonist-induced GIPR signalling if it is present in sufficient amounts to fully outcompete the agonist e.g.GIP (1− 42).Results are available from studies with antagonists based on N-terminal truncations or N-terminal modifications, as described below.For a comprehensive review of all truncations, see [128].
The effects of further N-terminal truncations of GIP(3− 30)NH 2 are consistent with the current understanding of regions of the GIP molecule important for receptor binding and activation [127,[129][130][131][132][133] (Fig. 3).For example, earlier studies described GIP(6− 30)NH 2 as an antagonist of GIP-induced cAMP accumulation in vitro with 58% inhibition at 100 nM with single-digit nanomolar affinity (IC 50 of 3 nM) and GIP (7− 30) as a weak antagonist of GIP-induced insulin secretion in rats with 52 % inhibition of insulin during oral glucose administration.Likewise, rodent studies using species-specific GIP(3− 30)NH 2 have supported this truncation as causing reliable GIPR antagonism.For instance, rat GIP(3− 30)NH 2 acted as a high-affinity competitive antagonist with pA2 of 13 nM determined by Schild-plot analyses measuring cAMP production in HEK293 cells expressing the rat GIPR, confirmed by a K i of 17 nM determined by competition binding experiments.In a rat pancreas perfusion model, this translated into an efficient inhibition of GIP-stimulated insulin, glucagon, and somatostatin secretion by 1 µM rat GIP(3− 30)NH 2 [134].Similarly, a GIP(3− 30)NH 2 variant optimized for binding to the mouse GIPR and having a longer half-life (T 1/2 of ~7 hours in mice) inhibited GIP-induced glucoregulatory and insulinotropic effects in vivo and resulted in a reduction of body weight and fat mass upon sub-chronic treatment in high-fat-fed female mice [135].
In-depth in vitro molecular pharmacological studies using GIP(3− 30) NH 2 as a model antagonist have supported receptor inhibition as a therapeutic principle for metabolic diseases.In heterologous expression systems, e.g., HEK293 or COS-7 cells expressing the GIPR, GIP(3− 30) NH 2 blocks cAMP accumulation and arrestin recruitment, as expected from its competitive antagonistic nature [42].Similar findings were made in cells with endogenous GIPR expression, such as human adipocytes, where GIP(3− 30)NH 2 blocked GIP-induced cAMP production [51].Importantly, it also prevents receptor internalization and increases receptor expression at the cell surface, as described using SNAP-tagged human GIPR [51].This, in turn, may enhance the sensitivity for endogenous GIP by altering the balance of GIPR towards more receptors on the cell surface.Thus, GIPR antagonism might improve an impaired GIP system, for instance, in a beta cell during type 2 diabetes [16].
In conclusion, since GIP(3− 30)NH 2 so far seems to be the best peptide-based GIPR antagonist and is present endogenously (although in very low amounts), this molecule (or variants) has been used for a majority of the molecular, cellular, and physiological studies within the GIP system.

GIP-based GIPR antagonists -N-terminal modifications
Before the identification of GIP(3− 30)NH 2 , another N-terminal variant, Pro3-GIP was used as a model antagonist for rodent studies [136,137].This molecule was initially designed to protect the GIP N-terminus from DPP-4 activity, thereby increasing its half-life [136] (Fig. 3).It was initially presented as a GIPR antagonist based on its ability to antagonize the cAMP production in transiently transfected CHL cells, although with low potency (EC 50 value of 2.6 µM) and to inhibit GIP-stimulated insulin release in the rat pancreatic cell line BRIN-BD11.In the same study, Pro3-GIP injected intraperitoneally in obese diabetic (ob/ob) mice decreased insulin secretion by 2.4-fold [136].Administered the same way, but during chronic treatment over 60 days, the same dose of Pro3-GIP (25 nmol/kg/day) prevented the age-related development of diabetes as indicated by improved non-fasting glucose, HbA1c, glucose tolerance, and insulin sensitivity to a degree where none of these parameters differed significantly from those of normal age-matched lean controls [138].An optimized form, Pro3-GIP mini-polyethylene glycol ((Pro(3))GIP[mPEG]), with a longer half-life, supported these beneficial metabolic effects [139].Finally, administered alone or in combination with the GLP-1R antagonist (exendin(9− 39)NH 2 ), studies of Pro3-GIP in ob/ob mice indicated that Fig. 3. Sequence of human glucose-dependent insulinotropic polypeptide (GIP).DPP-4, dipeptidyl peptidase 4.

M.M. Rosenkilde et al.
GIP is the main incretin, responsible for approximately 80% of the incretin effect [136].Thus, the findings with Pro3-GIP would seem to support that anti-diabetic activity and weight loss/reduced weight gain in obese mice could be obtained with GIPR inhibition.
However, later studies focusing on a deeper molecular pharmacological analysis of Pro3-GIP revealed signs of partial agonism with this compound.Thus, in transiently transfected HEK293 cells expressing the human GIPR, Pro3-GIP stimulated cAMP production with an EC 50 of 6.7 and an E max on 83% of human GIP [52].This was confirmed in transiently transfected COS-7 cells, another cell line often used for heterologous receptor expression in molecular pharmacological studies.Here, human Pro3-GIP activated the human GIPR with an even higher potency (EC 50 of 4.7 nM) and an efficacy (E max ) of 90% [137].The partial agonism was also observed in acute in vivo studies using pancreas perfusions in rats and mice, where the specific-specific Pro3-GIP (rat and mice Pro3-GIP) induced modest insulin, glucagon, and somatostatin secretion, corresponding to the partial agonism in cAMP production [137].Importantly, using GIPR knock-out mice, this effect disappeared, supporting specific and GIPR-dependent mechanisms for Pro3-GIP [137].

Species differences in the GIP system -a major source of pharmacodynamic confusion
Studies in rodents using GIP(3− 30)NH 2 and other GIPR antagonists have revealed that GIP and its receptor are less conserved, leading to different results in experiments with heterologous receptors/ligands [137] -a difference not observed for the more conserved GLP-1 system [137,140].Also, regarding partial agonists and full antagonists, species differences matter.This is exemplified by GIP(3− 30)NH 2 .When based on the human GIP peptide sequence, it acts with >3-fold lower potency on the rat GIPR compared to rat GIP(3− 30)NH 2 [51,134].Thus, for rat studies where GIPR antagonism is warranted, rat GIP(3− 30)NH 2 is superior to the human counterpart and has been proven as an efficient antagonist of GIP-mediated insulin, glucagon, and somatostatin release [134].Likewise, regarding inhibition of the human GIPR, human GIP (3− 30)NH 2 is superior to rat-(and mouse-derived) GIP(3− 30)NH 2 [42].Intriguingly, only one amino-acid substitution distinguishes these two (position 18: histidine in the human sequence and arginine in the rat).However, with 81% identity between human and rodent GIPRs, this structural difference is enough to elicit substantial species-specific functional differences.In more closely related species, like non-human primates, human GIP(3− 30)NH 2 displays the same activities as on the human GIPR [42].Among the agonists, actions are more conserved; thus, the species-specific GIP(1− 42) agonists are more similar across and between human, rat, and mouse species [135,137].
Systematic studies of human, rat, and mouse-derived Pro3-GIP have also revealed species-dependent effects.Besides the variations in position 18 (histidine in human GIP, arginine in rat and mouse GIP), position 30 also differs with an arginine in mouse GIP compared to lysine in human and rat GIP, while position 40 is a leucine in rat, but isoleucine in human and mouse GIP (Fig. 3).Thus, overall small species differences.However, while human-derived Pro3-GIP is an almost full agonist on the human GIPR (see above), rat and mouse Pro3-GIP activated their respective receptors with much lower efficacy (E max of 64% and 59%, respectively).Similarly, the potencies decreased from 4.7 nM to 13 nM and 29 nM, respectively.Across all three receptors (human, rat, and mouse), mouse Pro3-GIP resulted in the lowest E max and human Pro3-GIP in the highest E max .The lowest efficacy and potency on the rodent receptors are consistent with the antagonistic properties reported from the early in vivo studies in these species (see above).
Taken together, studies of GIP(3− 30)NH 2 and Pro3-GIP in different species uncovered that GIPR ligands do not necessarily affect the human and non-human GIPRs similarly, considerations that are important for the translation of rodent results to the human therapeutic potential.

Human studies with GIP(3¡30)NH 2
The naturally occurring competitive GIPR antagonist GIP(3− 30)NH 2 has been administered intravenously to humans on several occasions (Table 1).Initially and most importantly, the GIPR antagonist was used to demonstrate the physiological actions of endogenous GIP via inhibition of the GIPR in the postprandial state compared to a placebo infusion [14,141] (Fig. 4).GIP(3− 30)NH 2 has been infused in both healthy individuals [6,12,14,[141][142][143] and patients with type 1 diabetes [144], type 2 diabetes [145][146][147] with or without obesity [148], and in totally pancreatectomized patients [149].In agreement with the results of early physiological infusion studies of the agonist, GIP(1− 42), endogenous GIP has been confirmed as an incretin hormone with more pronounced or at least comparable insulinotropic actions as the other incretin hormone GLP-1 [7].In contrast to previous infusions of exogenous GIP (1− 42), which did not increase insulin secretion, endogenous GIP seems to have some insulinotropic effect in patients with type 2 diabetes (although this has little effect on blood glucose) [145].Moreover, endogenous GIP also weakly affected glucose metabolism in patients with type 1 diabetes despite the complete lack of beta cells and, therefore, no possibility of stimulating endogenous insulin secretion [144].Proving the dynamic regulation of the GIPR function, as mentioned above, the insulinotropic actions of endogenous GIP in patients with type 2 diabetes can be improved by normalization of blood glucose (using anti-diabetes medications) [146].The improvement has also been confirmed for exogenous GIP [150,151], and the improved insulinotropic actions following improved glycemic control could be a possible contributory mechanism for the efficacy of GIPR agonistic compounds in combination with other glucose-lowering drugs, e.g., GLP-1R agonists.However, in patients already treated with a GLP-1R agonist (liraglutide), infusions of GIP impaired rather than improved glycemia [87].
Both in healthy individuals and patients with type 2 diabetes, GIPR agonism increases circulating glucagon levels [20,87,88,152], whereas infusions of the antagonist GIP(3− 30)NH 2 lowers glucagon [142,147].In the postprandial state, hyperglucagonemia worsens the glucose excursions in patients with type 2 diabetes, and endogenous GIP could be one of the promoting factors for this paradoxical phenomenon [153].
In the postprandial state, GIP(3− 30)NH 2 infusion resulted in less suppressed levels of the bone resorption marker CTX than placebo (saline) infusion [8,145].Remarkably, the CTX suppression due to endogenous GIP calculated from GIP(3− 30)NH 2 infusions was similar for patients with type 2 diabetes and healthy individuals.This indicates that the GIP system in bone tissue is not affected by the diabetic, hyperglycaemic state as seen in the beta cell.Although the hypothesis has not been thoroughly studied, a possible explanation for these tissue-specific changes in GIPR function could be that the loss of insulinotropic actions is not due to GIPR dysfunction but caused by GIPR downregulation [154] and/or solely beta cell exhaustion and consequently, reduced insulin secretion.
Most recently, it has been established based on GIP(3− 30)NH 2 infusions that endogenous GIP is responsible for a postprandial regulation of blood flow in the major intestinal vessels, mesenteric superior artery, and portal vein [12].The cardiovascular activities of exogenous GIP also include an increase in heart rate of approximately 5-10 bpm and, in some cases, peripheral blood pressure reductions of approximately 5 mmHg [48,155].The physiological (and therapeutic) consequences of GIP's involvement in the regulation of postprandial blood redistribution require further study.
During oral glucose or liquid mixed meal intake, GIP(3− 30)NH 2 does not affect appetite as assessed from ad libitum meal tests and visual analogue scale questionnaires [6,142,145,149].Postprandial circulating free fatty acids, triacylglycerol, and cholesterols are also not changed in response to GIP(3− 30)NH 2 compared to placebo (saline) in healthy individuals and patients with type 2 diabetes and obesity [142,145].
Taken together, infusions of the GIPR antagonist in humans have demonstrated the physiological actions of endogenous GIP in the M.M. Rosenkilde et al. pancreas, the vascular system, and bone tissue.No adverse reactions have been reported for the studies reviewed here (Table 1).Especially, gastrointestinal side effects, as seen for GLP-1R agonists, were not found.

Status on therapeutic use of GIPR antagonists
Intriguingly, treatment with GIPR antagonists has resulted in similar weight loss effects as GIPR agonists in rodents and non-human primates, both when combined with GLP1-R agonists [23,24].As discussed above, this paradox canat least partlybe understood from GIP's action on other tissues than the endocrine pancreas, for example, GIP's role in intestinal blood flow and adipose tissue metabolism, and the action of the antagonists versus agonists on GIPR expression and desensitization (as described above and in Figs. 4 and 5).Moreover, the potential crosstalk and heterodimerization between GIPR and GLP-1R in, for instance, the beta cells may be affected by GIPR agonists and antagonists given with or without GLP-1R agonists (Fig. 2).
The recent administration to humans of AMG133, a GIPR antagonistic antibody conjugated with two GLP-1-based peptides in one molecule, resulted in weight loss of 15% after four doses over 85 days of

Study
Population Main result GIP(3− 30)NH 2 is an efficacious GIP receptor antagonist in humans: a randomised, double-blinded, placebo-controlled, crossover study [141] Healthy men (n=10) GIP(3− 30)NH 2 is an efficacious and specific GIP receptor antagonist in humans GIP(3− 30)NH 2 reduces GIP-induced insulin secretion during hyperglycaemia The Gluco-and Liporegulatory and Vasodilatory Effects of Glucose-Dependent Insulinotropic Polypeptide (GIP) Are Abolished by an Antagonist of the Human GIP Receptor [14] Healthy men (n=8) GIP(3− 30)NH 2 reduces GIP-induced adipose tissue blood flow, and triacyl glyceride and glucose uptake in adipose tissue during hyperglycaemia and hyperinsulinemia GIP's effect on bone metabolism is reduced by the selective GIP receptor antagonist GIP(3− 30)NH 2 [161] Healthy men (n=10) GIP(3− 30)NH 2 reduces GIP-induced attenuation of the bone resorption marker CTX during hyperglycaemia Separate and Combined Glucometabolic Effects of Endogenous Glucose-Dependent Insulinotropic Polypeptide and Glucagon-like Peptide 1 in Healthy Individuals [6] Healthy men (n=18) GIP(3− 30)NH 2 increases plasma glucose and reduces insulin secretion during oral glucose tolerance tests compared to saline infusion.Therefore, endogenous GIP is confirmed to be insulinotropic GIP and GLP-1 Receptor Antagonism During a Meal in Healthy Individuals [142] Healthy men (n=12) GIP(3− 30)NH 2 increases plasma glucose and reduces insulin secretion during liquid mixed meal ingestion compared to saline infusion.Therefore, endogenous GIP is confirmed to be insulinotropic The role of endogenous GIP and GLP-1 in postprandial bone homeostasis [8] Healthy men (n=18, n=12 Patients with type 2 diabetes (n=15) A period of plasma glucose-normalization increases beta cell sensitivity to endogenous GIP (assessed by GIP(3− 30)NH 2 vs. saline infusions), translating into a detectable contribution of endogenous GIP to oral glucose tolerance The gut hormone GIP contributes to postprandial hyperaemia in humans [12] Healthy men (n=10) GIP(3− 30)NH 2 reduces postprandial splanchnic blood flow during oral glucose ingestion compared to saline, confirming a physiological action of endogenous GIP Glucose-dependent insulinotropic polypeptide and glucagon-like peptide 1 do not account for the entire incretin effect in healthy humans [163] Healthy participants (n=12) GIP(3− 30)NH 2 increases plasma glucose and reduces insulin secretion during intraduodenal glucose tolerance tests compared to saline Extrapancreatic effects of endogenous incretin hormones evaluated using incretin hormone receptor antagonists in totally pancreatectomised patients [149] Pancreatectomized patients administration in obese individuals with a mean BMI of 33.8 kg/m 2 (multiple ascending doses, every fourth week, phase 1 trial).The adverse effects were mild-moderate but highly prevalent (70-100%) and similar to those of other GLP-1R agonists [74].According to the registration at clinicaltrial.gov, the phase 2 trial is expected to be completed in 2026 [156].
In recent years, studies from humans receiving the naturally occurring GIPR antagonist, GIP(3− 30)NH 2 (Table 1), as reviewed above, have expanded our understanding of the GIP system postprandially.Due to the short half-life and intravenous infusion of GIP(3− 30)NH 2 , it is mainly physiological effects and pathophysiological implications of endogenous GIP, that may be addressed with this tool.Although there are limitations, the results support a role of GIP in glucose metabolism both in lean and obese individuals with or without type 2 diabetes (Table 1).

Conclusions
Recent years have brought the GIP system into the therapeutic landscape as a compelling addition to treatments based on GLP-1.The identification of exendin(9− 39)NH 2 in 1993 as a valuable tool compound to block the GLP-1R and thereby elucidate the role of the GLP-1R in metabolism, have helped to understand the actions of the GLP-1Rbased therapies, today acknowledged for their effectiveness in the treatment of type 2 diabetes and obesity [157][158][159].Similarly, the more recent identification of the endogenous GIP-derived GIPR antagonist (GIP(3− 30)NH 2 ) in 2015 [35] has contributed to a more thorough understanding of the GIP system in human physiology and metabolism.With this competitive antagonist, 15 human studies have been completed that, combined with rodent studies, have importantly improved our understanding of the role of the GIP system.We have also become aware of the importance of the species differences in the GIP system (compared to the more conserved GLP-1 system).Co-targeting of the GIPR and the GLP-1R has given striking results, possibly exceeding those obtained with GLP-1R monoagonists [1,2,24,160].However, the apparently similar results obtained by combining GLP-1R activations and either GIPR agonism or antagonism have spurred interest in the mechanism behind this (Figs. 2 and 5).At present, factors at the molecular, cellular as well as tissue levels may contribute to explaining this complexity, and it is evident that knowledge about the GIPR signalling on the parameters measured of endogenous GIP and GLP-1 could be established as assessed by separate and combined infusions of the GIPR antagonist GIP(3− 30)NH 2 and the GLP-1R antagonist exendin(9− 39)NH 2 compared to saline M.M. Rosenkilde et al.Peptides 177 (2024) 171212