The incretin axis is responsible for most postprandial insulin secretion in healthy humans, and loss of the incretin effect contributes to impaired glycemic control in people with type 2 diabetes. Based on these characteristics, the incretin axis continues to be an attractive target for drug development, and GLP-1R agonists (GLP-1RA) have emerged as potent and effective treatments to reduce blood glucose and body weight. Continued evolution of this drug class has seen the development of single peptides that activate multiple receptors, with incretin peptide sequences engineered to activate additional G-protein-coupled receptors (GPCRs). An early monomeric dual receptor agonist targeted both GLP-1R and GIPR, and the synergism between the two receptor systems was initially reported 10 years ago. In mouse models, dual agonism had superior efficacy for weight loss and glucose control compared to a GLP-1RA alone, with additive effects of the GIPR proposed to act through signaling in beta cells, alpha cells, the CNS (central nervous system) and adipocytes. Despite promising preclinical studies, a 12-week clinical trial using an iteration of this initial dual-incretin agonist failed to demonstrate superiority relative to a GLP-1R monoagonist, raising questions about multi-receptor strategies in humans.
Tirzepatide is an agonist for both incretin receptors, engineered from the human GIP (hGIP) peptide sequence. It has an average half-life of approximately 5 days, enabling once-weekly dosing. Tirzepatide is an imbalanced agonist, engaging the GIPR to a greater degree than the GLP-1R in cultured cell systems. Moreover, it has an in vitro pharmacological profile that mimics the signaling of native GIP at the GIPR, but it is biased at the GLP-1R to favor cyclic AMP (cAMP) generation over β-arrestin recruitment. In clinical trials, treatment with tirzepatide produced superior weight loss and glycemic control compared to GLP-1RAs, suggesting that agonism at both the GIPR and GLP-1R is beneficial in humans with type 2 diabetes. However, despite strong evidence that tirzepatide engages the GIPR in competition-binding assays, cell-based experiments using transfected receptors and studies of transgenic mice, there is little functional evidence that tirzepatide directly activates the GIPR in humans as part of its substantial pharmacological effect.
Here, we report the results of experiments with tirzepatide in primary islets, an experimental approach uniquely suited to assess whether tirzepatide directly activates the GIPR in humans. Beta cell incretin receptor activity drives insulin secretion, an essential component of the antidiabetic response to either GLP-1R or GIPR agonists. Additionally, the beta cell is one of only a few cell types that express both incretin receptors, providing a model to test the relative importance of GIPR versus GLP-1R signaling. The GLP-1 sequence is conserved across rodent and human species, whereas the GIP sequence differs between species. Tirzepatide is engineered from the hGIP sequence. Importantly, hGIP has reduced potency at the mGIPR, and it has been suggested that tirzepatide also has reduced potency at the mGIPR. Therefore, our initial investigation set out to provide a comprehensive analysis of the potency of tirzepatide at the mGIPR to identify potential limitations of using mouse models to study the actions of tirzepatide. We assessed target engagement of mGIP, hGIP and tirzepatide at the mGIPR with four complementary approaches: (1) ligand binding assays; (2) Gα S recruitment; (3) G-protein activation; and (4) cAMP generation (Table 1). Overall, the affinity–potency profile of tirzepatide was 3–60-fold weaker relative to mGIP at the mGIPR, with similar or slightly reduced potency compared to GLP-1 at the mGLP-1R (Extended Data Table 1). Previous measures of tirzepatide activation of human incretin receptors demonstrated increased potency at hGIPR relative to hGLP-1R. Based on these early studies, it was concluded that tirzepatide acts on the hGIPR similarly to native hGIP but engages the hGLP-1R with parameters that differ from native GLP-1. However, this profile appears to differ for tirzepatide interactions with mouse incretin receptors; for example, tirzepatide and GLP-1 behave similarly at the mGLP-1R, whereas tirzepatide is less potent at the mGIPR than mGIP. This suggests that the imbalanced activity of tirzepatide may actually favor GLP-1R signaling in murine beta cells and suggests caution when using the compound in experiments with mouse models.
To address the functional importance of tirzepatide at each incretin receptor, we investigated the effect of loss-of-function approaches on insulin secretion in mice. We first used islets isolated from mice with selective deletion of the Gipr in beta cells in combination with the GLP-1R antagonist exendin(9-39) (Ex9). Tirzepatide stimulated insulin secretion in a concentration-dependent manner (0–100 nM) in control islets (Fig. 1a). However, compared to control islets, beta cell Gipr knockout islets secreted more insulin in response to tirzepatide (Fig. 1a), whereas Ex9 completely blocked insulin secretion in response to tirzepatide in both control and knockout islets (Fig. 1a). We reasoned that the enhanced response in knockout islets was attributed to a compensatory enhancement in GLP-1R signaling that has been previously described in various GIPR knockout models. To circumvent this issue, we next used acute pharmacological antagonism of the incretin receptors in mouse islets. We applied a recently validated long-acting GIPR antagonist, which prevented glucose lowering in response to an acylated GIPR agonist in wild-type mice (Extended Data Fig. 1a). Antagonism of the GLP-1R with Ex9 prevented insulin secretion in response to tirzepatide, whereas the presence of a GIPR antagonist had no effect on tirzepatide-stimulated insulin secretion, alone or in combination with Ex9 (Fig. 1b). These findings suggest that tirzepatide works predominantly through the GLP-1R to stimulate insulin secretion in mouse islets. To determine the consequences of these findings on glucose tolerance, we pretreated mice with acylated antagonists of GLP-1R or GIPR, alone or in combination, followed by tirzepatide and an intraperitoneal glucose tolerance test (IPGTT) (Fig. 1b). We used 3 nmol kg–1 of tirzepatide, identified as a maximal dose for glucose lowering in wild-type mice (Extended Data Fig. 1b), to provide an opportunity for activity at both incretin receptors. Before glucose administration, tirzepatide reduced fasting glycemia, which was prevented by antagonism of the GLP-1R but not the GIPR (Fig. 1c). Tirzepatide robustly lowered glycemia during the IPGTT, an effect that was completely blocked by GLP-1R antagonism (Fig. 1d) or when the experiments were conducted in Glp1r-knockout mice (Extended Data Fig. 1c). In comparison, antagonism of the GIPR did not alter the actions of 3 nmol kg–1 of tirzepatide to reduce glycemia or influence the effect of the GLP-1R antagonist on glucose tolerance (Fig. 1d). It has been demonstrated that higher doses of tirzepatide show activity at the GIPR, prompting us to repeat these experiments using 30 nmol kg–1 of tirzepatide. We found that GLP-1R antagonism only partially blocked the glucose-lowering effects of tirzepatide (Extended Data Fig. 1c). Moreover, whereas GIPR antagonism alone failed to prevent the effects of this higher dose of tirzepatide, a combined effect of both antagonists was seen, preventing glucose lowering in response to tirzepatide. These data agree with previous results that show that tirzepatide can engage the mGIPR but requires high doses to do so in mice.
