GLP-1 is encoded by proglucagon, a 158 amino acid precursor protein, that is predominantly expressed in the gut, pancreas, and distinct neuronal populations of the hindbrain (29). In the brain and the intestine, proglucagon is cleaved by the action of the prohormone convertase 1/3 (PC1/3) into GLP-1, GLP-2, glicentin, glicentin-related polypeptide (GRPP), and oxyntomodulin (OXM) (29–31). In the pancreatic α-cells, proglucagon is cleaved by PC2 into glucagon, GRPP and the major proglucagon fragment (MPGF). In the intestine, GLP-1 is secreted from enteroendocrine L-cells located in the gut epithelium. The density of the L-cells is low in the duodenum and jejunum and it is high in the ileum and colon (29). Nutrients stimulating the secretion of GLP-1 in the intestine include monosaccharides such as glucose, galactose, fructose (32–34), fatty acids (35, 36), as well as proteins (37) and amino acids, particularly glutamine and glycine (38, 39). The relevance of endocrine factors to promote GLP-1 secretion seem to vary across species and may include acetylcholine, insulin, ghrelin, GIP, and gastrin-releasing peptide (29). Circulating levels of total GLP-1 are low during fasting (~5 pmol/l) and rapidly rise up to 40 pmol/l shortly after a meal (40). Consistent with the ability of GLP-1 to accelerate glucose-stimulation of insulin secretion (GSIS), the meal-induced rise in plasma GLP-1 is paralleled by enhanced insulin immunoreactivity (40).

GLP-1 promotes its biological action through binding to the GLP-1 receptor (GLP-1R), a 7 transmembrane G protein-coupled receptor of the class B family (41). GLP-1R signals primarily via the Gαs pathway, and hence accelerates intracellular levels of cAMP (42). GLP-1R can also recruit, and induce signaling, via the Gαq and β-arrestin pathways and knockdown of β-arrestin-1 in rat insulinoma (INS1) cells decreases the ability of GLP-1 to promote GSIS (43). Immunohistochemical studies in tissues from humans and non-human primates show widespread distribution of GLP-1R in the brain and in the periphery (44). These data largely align with studies in rodents in which the abundance of the GLP-1R transcript was assessed using mice that express green fluorescent protein (GFP) under control of the GLP-1R promoter (45). Consistent with the ability of GLP-1R agonists to decrease homeostatic and hedonic food intake (29, 29), expression GLP-1R is found in the rodent hypothalamus (ARC, VMH, DMH, PVH, LH), hindbrain (AP, NTS, ventrolateral medulla) and telencephalon (amygdala, olfactory bulb, preoptic area, nucleus accumbens) (45, 46). In the pancreas, GLP-1R is solidly expressed in the β- and δ-cells but is only found in a small portion of α-cells (45). No expression of GLP-1R is found in the liver and the thyroid gland (44).

Albeit best known for its glycemic effects, GLP-1 is a pleiotropic hormone with a series of metabolic effects beyond the regulation of glucose metabolism ( Figure 2A ). Apart from its ability to act on the pancreas to enhance GSIS and to inhibit the secretion of glucagon, GLP-1 decreases body weight by decreasing homeostatic and hedonic food intake (29). GLP-1R agonism further inhibits gastric emptying (47); has cardio- and neuroprotective effects (48); lowers inflammation and apoptosis (49–53); stimulates β-cell proliferation in rodents (54); and exerts positive effects on learning, memory, and reward behavior (55, 56). Endogenous GLP-1 is mainly produced as GLP-1(7-36NH₂), with a low proportion produced as GLP-1(7-37) and an even a lower portion as GLP-1(1-37) or GLP-1(1-36NH₂) (29, 57). Native GLP-1 has a half-life of just ~2–3 min (58–60), which results mainly from rapid in vivo proteolysis by the dipeptidylpeptidase-4 (DPP-4) and fast renal elimination. DPP-4 cleaves GLP-1(7-36NH₂) and GLP-1(7-37) at the second N-terminal amino acid (Ala8) position, leading to metabolically metabolites GLP-1(9-36NH₂) and GLP-1(9-37) of much reduced potency (61). Despite species-related differences, GLP-1 is also subject to degradation by the neutral endopeptidase (NEP) 24.11, which cleaves GLP-1 at its central and C-terminal positions Asp15, Ser18, Tyr19, Glu27, Phe28 and Trp31 (61, 62). The relevance of NEP 24.11 to cleave GLP-1 varies among species and while it contributes substantially to GLP-1 degradation in mice and pigs (63, 64), it’s relevance in humans has long been questioned. Nonetheless, more recent data show that NEP 24.11 also plays a physiological relevant role for degradation of GLP-1 in humans (65).

Consequential due to its short half-life, native GLP-1 has only limited pharmacological potential and only ~10–15% of active GLP-1 is estimated to reach the general circulation (59, 66–68). Nonetheless, emphasizing its therapeutic implication, 6-wk continuous infusion of GLP-1(7-36NH₂) at a rate of 4.8 pmol^-1 kg^-1 min^-1 improved glycemic control and insulin sensitivity in patients with T2D (69). Despite the lack of singular attribution to enhanced GLP-1 action, similar results have been demonstrated following administration of a DPP-4 inhibitor (70–73).

Notably, as comprehensively reviewed elsewhere (20, 29, 74), GLP-1 improves glycemic control via several complementary mechanisms. In the pancreas, GLP-1 directly acts on the β-cells to promote GSIS via the action of PKA and Epac2 (29). Both pathways are equally important for the insulinotropic effect of GLP-1 and ensure that GLP-1 primarily enhances insulin secretion under conditions of hyperglycemia (29). Other than stimulating the secretion of insulin in a glucose-dependent manner, GLP-1 also promotes the production of insulin via activation of Pdx1, which binds to the insulin promoter and activates its expression (29). GLP-1 also lowers blood glucose by inhibiting the release of glucagon and thus inhibiting hepatic glucose production (29). Clamp studies in patients with T2D indicate that the insulinotropic and glucagonostatic effects of GLP-1 contribute equally to decreasing blood glucose (75). While GLP-1 directly acts on the β-cells to stimulate insulin secretion, GLP-1 inhibition of glucagon secretion seems to be indirectly triggered via paracrine effects in the islets. Accordingly, GLP-1 stimulates the secretion of insulin, zinc, GABA, amylin, and somatostatin, all of which inhibit glucagon secretion (30). Supporting this notion of a paracrine effect is that the GLP-1 receptor is only expressed in a small subset of α-cells (45). However, GLP-1 does not only decrease blood glucose via its direct effects on the islets, it also inhibits gastric emptying and thereby slows glucose entry into the circulation (47, 76–78). Emphasizing the importance of GLP-1-mediated inhibition on gastric emptying for the regulation of blood glucose, antagonizing GLP-1’s effect on gastric emptying by co-infusion of GLP-1 with erythromycin during a liquid meal diminished GLP-1’s ability to decrease post-prandial hyperglycemia in patients with T2D (79).

In summary, GLP-1 improves fasting blood glucose through its direct action on the pancreatic islets and decreases postprandial hyperglycemia through inhibition of gastric emptying, and thus reduced glucose entry into circulation (74). These differences in glucose regulation translate into important pharmacological differences. Short-acting GLP-1 mimetics (e.g. exenatide BID, lixisenatide) are self-administered prior to a meal and, in conjunction with their short half-life of 2–3 h, display substantial fluctuations in circulation, with highest levels during the prandial and early post-prandial state, and lowest levels in the fasting periods between meals. Due to their relatively high plasma levels at the time of meal intake, the short-acting GLP-1 mimetics relative to the long-acting GLP-1R agonists display a higher tendency to affect GI motility and to decrease post-prandial hyperglycemia (74). On the contrary, long-acting GLP-1R agonists are less prone to affect GI motility and primarily decrease fasting blood glucose levels via direct action in the pancreas (74). With stable plasma concentrations and lower efficacy on GI motility, the long-acting GLP-1 mimetics are less prone to incur adverse gastrointestinal side effects (80) and generally have a higher potential to decrease body weight (74).

A variety of structurally and chemically refined GLP-1R analogs have received extensive attention and have been implemented in clinical use for the treatment of T2D and obesity ( Figure 3 ) (29). Clinical success in the treatment of obesity has been established for liraglutide 3 mg (Saxenda^®, Novo Nordisk, Denmark) (81) and more recently for semaglutide 2.4 mg (Wegovy^®, Novo Nordisk, Denmark) (13, 28). Also, molecules with simultaneous activity at the receptor for GIP (82–85) or glucagon (86, 87) have shown promising results for this application. The continuous improvement of the pharmacokinetic profile of GLP-1R agonism, starting from a native hormone with a half-life of ~2–3 min to the development of twice daily (exenatide BID), daily (liraglutide, lixisenatide) and even weekly (exenatide ER, albiglutide, dulaglutide, semaglutide) formulations, highlight the pharmaceutical advancement in this arena. Additionally, the recent development of an orally available preparation of semaglutide (Rybelsus^®, Novo Nordisk), and the recruitment of GLP-1 into unimolecular pharmacology with GIP, glucagon (and others) (20, 29, 88) exemplifies how GLP-1-based drug development and innovation advanced over recent years. It emphasizes how structural refinements and biochemical modifications can extend the application of GLP-1R agonists from the treatment of T2D to obesity (13).

The short action profile of native GLP-1 imposes a major challenge towards its successful clinical utilization. To overcome this limitation, various strategies have been applied to extend the half-life of GLP-1 and to accelerate its in vivo action and potency ( Figure 4 ). The efficacy of a drug in a biological system is influenced by factors that include the stability of the active molecule and its rate of diffusion into and elimination from circulation. Methods to improve drug efficacy included structural and chemical modifications targeted to increase molecular stability and activity, to improve biodistribution/bioavailability, and to delay renal elimination (89).

Liraglutide 3 mg (Saxenda^®, Novo Nordisk, Copenhagen, Denmark) was approved in 2014 for the treatment of obesity in adults and in 2020 for the treatment of obesity in children aged 12 and older ( Figure 3 ) (141). Between 50 and 70% of the patients treated with liraglutide 3mg achieve a body weight reduction > 5% while between 6 and 35% achieve weight loss > 10% (142). Of appreciable note, in contrast to a vast majority of previously employed anti-obesity medications (13, 20), GLP-1R agonism improves cardiovascular (CV) health in patients with T2D (48, 143). Consistent with this, a recent meta-analysis assessing CV outcome of different GLP-1R agonists across eight CV outcome trials comprising 60,080 patients with T2D. It demonstrated that GLP-1R agonists reduce the risk of a major adverse cardiovascular event (MACE) by 14%, death by 13%, non-fatal stroke by 12%, and broad composite kidney outcome by 17% (144).

In June 2021, the FDA approved semaglutide 2.4 mg (Wegovy^®) in adjunct to lifestyle programs for the treatment of obesity in adults (145). In non-diabetic individuals overweight or obese, 68 weeks of treatment with semaglutide (2.4 mg QW), decreased body weight by 14.9% relative to 2.4% in placebo treated controls (28). Impressively, 86.4% of patients treated with semaglutide showed weight loss >5% (vs. 31.5% in placebo controls), while 69.1% (vs. 12.0% in placebo controls) lost >10%, 50.5% (vs. 4.9% in placebo controls) >15% and 32% (vs. 1.7% in placebo controls) >20% (28). Similar to the observation in patients with T2D (48, 143), participants with obesity receiving semaglutide 2.4 mg had greater improvement of markers indicative of cardiometabolic health (28). Like with other GLP-1R agonists, adverse effects of semaglutide are predominantly of gastrointestinal nature, with the most common being nausea, diarrhea, vomiting, and constipation (28). Notably, while a series of clinical studies confirms the ability of semaglutide 2.4 mg to lower body weight >10% in most patients, the magnitude of weight loss is considerably lower in diabetic relative to non-diabetic patients with obesity (28, 146–148).

GIP is derived from posttranslational cleavage of proGIP, a 153 amino acid preprohormone that is expressed in the enteroendocrine K-cells of the upper intestine, the pancreatic α-cells and potentially the CNS (149). The majority of circulating GIP refers to GIP(1-42), which gets cleaved from proGIP in the intestine by the action of PC1/3, but also a shorter form, GIP(1-30NH₂), is produced in the intestine and the pancreatic α-cells by cleavage of proGIP by PC2 (149). Both forms are of equal insulinotropic potency in mice (150) and both carry an alanine at their second N-terminal residue, which render the molecules susceptible for DPP-4 degradation (151). Similar to GLP-1, GIP undergoes rapid renal elimination and consistent with this, intact GIP is quickly cleared from the circulation with a half-life of four minutes and without major difference between healthy subjects and individuals with T2D (151). Primarily secreted in response to the ingestion of fat, GIP promotes its biological action through binding to the GIP receptor (GIPR), a class B GPCR that, similar to GLP-1R and belonging to the glucagon receptor family (149). In mice, Gipr is ubiquitously expressed in the endocrine pancreas with comparable mRNA quantities in the α- β- and δ-cells (152). GIPR is further expressed in adipocytes (153), myeloid cells (154), the endothelium of the heart and blood vessels, the pituitary and the inner layers of the adrenal cortex (155). In the brain, Gipr is found in the cerebral cortex, hippocampus, olfactory bulb as well as in the hypothalamus and the hindbrain (155–157).

GIP is best known for its ability to act on the pancreatic islets where depending on blood glucose concentration it stimulates the secretion of insulin or glucagon ( Figure 2B ) (158). The insulinotropic effect of GIP is diminished in patients with type-2 diabetes (159) but is restored upon near normalization of glycemia upon 4-week administration of insulin (160).

Beyond its glycemic effects, GIP decreases bone resorption (161), has neuroprotective effects in animal models of Alzheimer’s disease (162), and regulates lipid metabolism (163). As comprehensively reviewed previously (149, 163–165) GIP regulation of energy and lipid metabolism is partially conflicting, since activation and inhibition of GIPR signaling can both decrease body weight and fat mass in rodents (149, 164).

As demonstrated in vitro in cultured adipocytes (166, 167) and in vivo in Zucker rats (167), GIP enhances the activity of lipoprotein lipase (LPL), which promotes adipose tissue lipid disposal by enhancing hydrolytic cleavage of circulating triglycerides (TAG) into fatty acids (FA) and monoacylglycerol, which gets re-esterified and stored in adipose tissue. Consistent with this, GIP promotes clearance of circulating TAGs in dogs (168) and enhances adipose tissue TAG accumulation in rats (169) and humans (170). Under conditions of hyperglycemia and hyperinsulinemia, GIP increases adipose tissue blood flow and TAG clearance in healthy lean subjects (171). It further enhances FA synthesis in adipose tissue explants (172) and potentiates insulin-stimulated uptake of FA into the adipocytes (173). Collectively, GIP can act on the adipose tissue to increase adipose tissue blood flow, to enhance LPL-induced TAG clearance from the circulation and to stimulate adipocyte lipid storage. Recently, it was hypothesized that GIP may also facilitate healthy white adipose tissue expansion and thereby protect from adipocyte lipid spill-over and ectopic accumulation of lipids (163). This hypothesis is anchored on the observations that GIP is a target of PPARγ, a master regulator of adipogenesis (174), the expression of GIP increases during adipocyte differentiation (175, 176), and that knockdown of GIPR impairs adipocyte development (175).

Consistent with the proposed role of GIP in lipid storage, mice with global loss of Gipr are lean and exhibit decreased body weight gain when fed a high-fat diet (HFD) (177). Body weight is also decreased in Gipr deficient ob/ob mice relative to ob/ob controls (177) and is decreased in mice with deletion of Gipr in the adipose tissue (178) and the CNS (179). However, the body weight difference between wt mice and CNS- or adipose-specific Gipr ko mice is only mild in comparison to the global loss of Gipr (177–179), indicating that lack of GIPR signaling in these tissues does not fully explain the phenotype of the global Gipr ko mice. No difference in body weight is seen between wt mice and mice that lack Gipr in the brown adipose tissue (BAT) (180) or the pancreatic β-cells (181).

The observation that global loss of Gipr protects from diet-induced obesity (177) and that GIP can promote adipocyte lipid storage (163) has inspired the development of GIPR antagonists for the treatment of obesity (182). Among the most prominent GIPR antagonists are natural (183) and biochemically modified (fatty-acylated) (184) forms of N-terminally truncated GIP as well as GIPR neutralizing antibodies (181, 185, 186). When administered peripherally, particularly antibody-based GIPR antagonists demonstrate some potential to prevent the development of HFD-induced obesity in rodents (181, 186), but GIPR antagonists show only modest, if any ability to decrease body weight once obesity is already established (181, 183–186). However, one study demonstrated that central administration of an antibody-based GIPR antagonist remarkably decreased body weight in DIO mice, potentially through mechanisms that include restoration of leptin sensitivity (185). Enhanced leptin action can however not be the sole mechanism by which centrally applied GIPR antagonists decrease body weight, since loss of GIPR in leptin deficient ob/ob mice still decreases body weight relative to ob/ob controls (177).

While GIPR antagonism has only little effects to decrease body weight in DIO mice (181, 183–186), its combination with GLP-1R agonism decreases body weight beyond what is possible with either monotherapy alone (181). Interestingly, this observation is surprisingly similar to long-acting (acylated) GIPR agonists when co-injected with fatty-acyl GLP-1 (126). When given at a dose of 3 nmol/kg/day, fatty-acyl GIP fails to affect body weight in DIO mice, but when given as an adjunct to fatty-acyl GLP-1 (3 nmol/kg/day), weight loss in the co-therapy is synergistically greater than treatment with GIP or GLP-1 alone (126). The mechanism of how GIPR and GLP-1R agonism synergizes to enhance body weight loss are not known, but a series of experimental results indicates that GIP acts on the GIP receptor in the brain to decrease body weight via inhibition of food intake (156, 179, 187). Consistent with this, GIPR is expressed in the hypothalamus (156) and the hindbrain (157, 188) and DREADD-mediated activation of GIPR neurons/cells in the hypothalamus decreases food intake in rodents (156). Central and peripheral administration of a long-acting (fatty-acylated) GIP increases cFOS neuronal activation in key feeding areas of the hypothalamus (179) and several long-acting GIPR agonists have been shown to decreases body weight and food intake in DIO mice without affecting energy expenditure (179, 184). How pharmacological activation and inhibition of GIPR both improve energy and lipid metabolism is subject of intense scientific investigation. Relevant hypotheses include that GIPR agonists may desensitize GIP receptor signaling (189), or that GIPR agonists and antagonists improve metabolism through independent mechanisms (187). Arguing against a role of GIPR agonists as functional antagonists is the observation that single central administration of acylated GIP is sufficient to rapidly lower food intake within hours after its administration (179) and that even DREADD-mediated (non-GIPR ligand-induced) activation of GIPR neurons decreases food intake in mice (156). Also, expression of GIPR is not decreased upon chronic treatment of DIO mice with acyl-GIP in either the hypothalamus or the adipose tissue (179) and dual-agonists targeting the receptors for GLP-1 and GIP do not show enhanced internalization of the GIP receptor (42, 190).

The ability of GIPR agonists to decrease body weight and food intake vanishes in global Gipr ko mice but is preserved in GLP-1R ko mice (179, 184), indicating that GIPR agonists lower body weight and food intake independent of GLP-1R signaling via the GIP receptor. When administered centrally (icv), fatty-acyl GIP decreases body weight and food intake in HFD-fed wildtype mice but not in mice with CNS deletion of Gipr (179). When administered peripherally, fatty-acyl GIP fails to affect food intake in CNS Gipr ko mice but shows partially preserved ability to decrease body weight (179). These data indicate that fatty-acyl GIP acts in the brain to decrease body weight via inhibition of food intake but also decreases body weight independent of GIPR in the CNS (179). Although GIPR agonists do not require GLP-1R to decrease body weight (179, 184), GIPR agonism was recently shown to attenuate the emetic effects of GLP-1R agonism in mice, rats and musk shrews (188). Given the role of the caudal hindbrain in emetic/aversive behavior (191) and in mediating GLP-1 effects on food intake (56, 192, 193), these data collectively suggest that GIP potentially acts on the hindbrain-hypothalamus axis to decrease food intake and to improve tolerability of GLP-1R agonism.

The observation that weight loss in DIO mice is enhanced by co-therapy with GIPR and GLP-1R agonists (126) has resulted in the development of unimolecular GLP-1R/GIPR dual-agonist peptides (82, 126). The first of such dual-agonist, MAR709, shows nearly balanced activity at both target receptors and improves body weight and glycemia with greater potency relative to pharmacokinetically-matched GLP-1 in rodents with diet- and genetically induced obesity (126). In a 12-wk phase II study, MAR709 (a.k.a. NNC0090-2746) reduced body weight and blood glucose in patients with T2D, but the drug candidate at the single tested dose was not superior to dose titrated treatment with liraglutide (83). In 2020, Novo Nordisk announced discontinuation of MAR709 due to success of semaglutide 2.4 mg in clinical trials and in favor of other potentially more effective drugs in their clinical pipeline (13).

Another GIPR/GLP-1R co-agonist, Tirzepatide (a.k.a. LY3298176) has been developed by Eli Lilly (82). The molecule is based on the human GIP sequence, in which GLP-1R residues were introduced to yield a fivefold greater potency at the human GIPR relative to GLP-1R (82). Fatty-acylation of the lysine 20 residue with a C20 diacid allows covalent binding to albumin, which results in a half-life of ~160 hrs in humans (82). When given at equimolar concentrations, tirzepatide shows greater weight loss but equal improvement in glucose metabolism relative to treatment with semaglutide in DIO mice (82). Interestingly, when given at a daily dose of 10 nmol/kg, tirzepatide improves glucose metabolism but fails to affect body weight in GLP-1R ko mice (194). These data are seemingly in contrast to previous reports showing preserved ability of long-acting GIPR agonists to decrease body weight in GLP-1R ko mice (179, 184). Importantly, human GIP is only a weak and partial agonist at the mouse relative to the human GIP receptor (195), and all currently available data on the pharmacokinetics of tirzepatide are based on human GIPR (82). It warrants clarification whether tirzepatide shows preserved potency at the mouse GIPR.

In recent phase III clinical trials, tirzepatide showed profound ability to improve body weight and glycemia in patients with obesity and T2D. Depending on the dose (5, 10, or 15 mg QW), 40-52 weeks of treatment with tirzepatide, decreased HbA1c between -1.87 and – 2.59%, with 81 – 97% of patients achieving a HbA1c <7%, and 23 – 62% of patients achieving HbA1c <5.7% (24–27, 196). Serum levels of fasting glucose decreased between 43.6 – 67.9 mg/dl while body weight decreased relative baseline between 6.2 – 12.9% (24–27, 196). After 40 weeks of treatment, 47 – 57% of patients treated with tirzepatide 15mg QW decreased body weight >10%, relative to 1% in patients treated with placebo, while 27 – 32% of patients decreased body weight >15% (0% in placebo controls) (25, 27, 196). Treatment with tirzepatide was at all tested doses (5, 10, 15 mg QW) superior to treatment with semaglutide 1mg QW to decrease HbA1c, fasting levels of blood glucose and body weight (25). It warrants to be determined how tirzepatide compares to the recently approved semaglutide 2.4 mg. Interestingly, while tirzepatide shows comparable ability to decrease body weight in obese type-2 diabetic vs non-diabetic individuals (24–27), Semaglutide 2.4 mg is considerably less efficacious in obese diabetic relative to obese non-diabetic individuals (28, 146–148).