Tirzepatide is an imbalanced and biased dual GIP and GLP-1 receptor agonist

Abstract

Tirzepatide (LY3298176) is a dual GIP and GLP-1 receptor agonist under development for the treatment of type 2 diabetes mellitus (T2DM), obesity, and nonalcoholic steatohepatitis. Early phase trials in T2DM indicate that tirzepatide improves clinical outcomes beyond those achieved by a selective GLP-1 receptor agonist. Therefore, we hypothesized that the integrated potency and signaling properties of tirzepatide provide a unique pharmacological profile tailored for improving broad metabolic control. Here, we establish methodology for calculating occupancy of each receptor for clinically efficacious doses of the drug. This analysis reveals a greater degree of engagement of tirzepatide for the GIP receptor than the GLP-1 receptor, corroborating an imbalanced mechanism of action. Pharmacologically, signaling studies demonstrate that tirzepatide mimics the actions of native GIP at the GIP receptor but shows bias at the GLP-1 receptor to favor cAMP generation over β-arrestin recruitment, coincident with a weaker ability to drive GLP-1 receptor internalization compared with GLP-1. Experiments in primary islets reveal β-arrestin1 limits the insulin response to GLP-1, but not GIP or tirzepatide, suggesting that the biased agonism of tirzepatide enhances insulin secretion. Imbalance toward GIP receptor, combined with distinct signaling properties at the GLP-1 receptor, together may account for the promising efficacy of this investigational agent.

Keywords: Diabetes; Therapeutics.

Figure 1. Tirzepatide (TZP) is an imbalanced agonist of the GIP and GLP-1 receptors and shows biased pharmacology at the GLP-1 receptor.

(A–F) Intracellular cAMP accumulation was measured in low-density human GIPR- and GLP-1R–expressing HEK293 cells. (A) The intrinsic potency of TZP (n = 23) in the presence of casein (CAS) is equivalent to GIP(1-42) (n = 49). In the presence of human serum albumin (HSA), the potency of TZP right shifts 26-fold (n = 5), while GIP(1-42) is unaltered (n = 15). (B) The intrinsic potency of TZP (n = 22) is approximately 18-fold lower than GLP-1(7-36) (n = 57). In the presence of HSA, TZP right-shifts 81-fold (n = 5), while the potency of GLP-1(7-36) is unaltered (n = 24). (C–F) Agonist induced generation of cAMP was measured kinetically using a luminescence biosensor. Data are representative of 3 experiments. At the GIPR, GIP(1-42) (C), and TZP (E) have identical kinetic profiles. On the GLP-1R, native GLP-1(7-36) (D) has a complex profile with a bi-phasic kinetic response at high ligand concentrations, while TZP (F) is monophasic even at the highest tested concentrations. (G and H) Agonist-stimulated GTPγS binding of Gα_s in GIPR and GLP-1R in HEK293 cell membranes are presented as the mean ± SEM of 3 independent experiments. (G) On the GIPR, TZP is fully efficacious with an EC₅₀ (SEM, n) of 0.379 nM (0.070, 3) versus GIP(1-42) of 1.43 nM (0.18, 27). (H) On the GLP-1R, TZP is a partial agonist 51% stimulation (5.2, 3) with an EC₅₀ of 0.617 nM (0.190, 3) versus GLP-1(7-36) of 1.63 nM (0.21, 26). (I and J) The recruitment of ARRB2 to GIPR and GLP-1R in CHO-K1 cells. Representative data are presented. (I) The potency of TZP to recruit ARRB2 to GIPR is 2.34 nM (0.60, 7) and is comparable with GIP(1-42) of 1.58 nM (0.52, 6). (J) The potency of TZP to recruit ARRB2 to GLP-1R is difficult to determine due to low efficacy (n = 5), while GLP-1(7-36) is a full agonist with an EC₅₀ of 3.26 nM (0.71, 14).

Figure 2. Tirzepatide (TZP) differentially induces internalization of the GIP and GLP-1 receptors.

(A–D) The time course of internalization of GIPR (A and C) and GLP-1R (B and D) was assessed using changes in cell surface presentation of SNAP-tagged receptors in HEK293A cells. Receptor internalization induced by increasing concentrations of GIP(1-42) (A), GLP-1(7-36) (B), or tirzepatide (C, GIPR; D, GLP-1R) over time relative to the maximum signal for either GIP(1-42) (1 μM) or GLP-1(7-36) (1 μM) is shown. (E–H) Studies using receptors containing an N-terminal HA-epitope tag and a C-terminal EGFP fusion are presented. (E) Changes in surface GIPR 60 minutes after treatment with ligand were measured by anti-HA immunofluorescence. Data are normalized to 1 μM GIP(1-42) values. For GIPR, tirzepatide induced internalization with an EC₅₀ (SEM, n) of 18.1 nM (5.7, 4), while GIP(1-42) displayed a potency of 18.2 nM (9.7, 4). (G) Representative confocal images of HA-GIPR-EGFP cells detecting EGFP fluorescence following treatment with vehicle, 100 nM GIP(1-42), or 100 nM tirzepatide. (F) Changes in surface GLP-1R 30 minutes after treatment with ligand detected by anti-HA immunofluorescence. Data are normalized to 1 μM GLP-1(7-36) values. For GLP-1R, tirzepatide was partially efficacious at 43.6% (7.9, 3) with an EC₅₀ of 101.9 nM (29.8, 3), while GLP-1(7-36) showed a potency of 22.2 nM (1.86, 3). (H) Representative confocal images of HA–GLP-1R–EGFP cells detecting EGFP fluorescence following treatment with vehicle, 100 nM GLP-1(7-36), or 100 nM tirzepatide. Scale bars: 20 μm.

Figure 3. Deletion of β-arrestin1 in pancreatic β-cells increases GLP-1 receptor–activated insulin secretion.

Islets from littermate controls and Arrb1^{βcell–/–} mice (male) were perifused ex vivo, and insulin secretion was measured in response to glucose (A), GLP-1 (B), GIP (C), or tirzepatide (TZP; D). Islets from Arrb1^{βcell–/–} mice secreted more insulin compared with control islets in response to 16 mM glucose (A) and 300 pM GLP-1 (B). By contrast, insulin secretion was not different between the sets of islets in response to either 3 nM GIP (C) or 1 nM tirzepatide (D). Exendin-4_(9–39) (Ex9; 1 μM) was used prior to GIP stimulation to normalize the elevated glucose stimulated insulin secretion. The integrated AUC (iAUC) was determined during the stimulation period: 6–19 minutes for 16 mM glucose (A), 20–29 minutes for GLP-1 (B), 40–58 minutes for GIP (C), and 24–39 minutes for tirzepatide (D). Each panel depicts results of a representative experiment from at least 2 independent experiments. *P<0.05, values are mean ± SEM. Statistical differences in iAUC values were determined by a 2-tailed student’s t test.