Translating biased signaling in the ghrelin receptor system into differential in vivo functions

Abstract

Biased signaling has been suggested as a means of selectively modulating a limited fraction of the signaling pathways for G-protein-coupled receptor family members. Hence, biased ligands may allow modulation of only the desired physiological functions and not elicit undesired effects associated with pharmacological treatments. The ghrelin receptor is a highly sought antiobesity target, since the gut hormone ghrelin in humans has been shown to increase both food intake and fat accumulation. However, it also modulates mood, behavior, growth hormone secretion, and gastric motility. Thus, blocking all pathways of this receptor may give rise to potential side effects. In the present study, we describe a highly promiscuous signaling capacity for the ghrelin receptor. We tested selected ligands for their ability to regulate the various pathways engaged by the receptor. Among those, a biased ligand, YIL781, was found to activate the Gα_q/11 and Gα₁₂ pathways selectively without affecting the engagement of β-arrestin or other G proteins. YIL781 was further characterized for its in vivo physiological functions. In combination with the use of mice in which Gα_q/11 was selectively deleted in the appetite-regulating AgRP neurons, this biased ligand allowed us to demonstrate that selective blockade of Gα_q/11, without antagonism at β-arrestin or other G-protein coupling is sufficient to decrease food intake.

Keywords: appetite regulation; biased signaling; food intake; gastric emptying; ghrelin receptor.

Fig. 1.

Binding properties and second-messenger responses induced by ghrelin, YIL781, and Abb13d. (A) Competition binding of ghrelin (black dashed line), YIL781 (blue line) and 13d (red line) using 3H-labeled MK-677 as a radioligand. (B) YIL781 and Abb13d antagonize Ca²⁺ mobilization induced by 10 nM ghrelin. (C) When measuring the intrinsic activity of the compounds, YIL781 turned out to be a partial agonist, and Abb13d was found to be an inverse agonist for inducing Ca²⁺ release. (D) YIL781 and Abb13d also antagonize the IP response induced by 10 nM ghrelin. (E) YIL781 is a partial agonist and Abb13d is an inverse agonist for stimulation of IP signaling. Data points are the mean ± SEM of three to five independent experiments performed in triplicate.

Fig. 2.

Differential panel of G-protein activation by ghrelin, YIL781, and Abb13d, shown by two distinct BRET biosensors. Dose-dependent activation by ghrelin (black), YIL781 (blue), and Abb13d (red) of Gα_q/11, Gα_i/o, Gα_12/13, and Gα_s proteins shown by Gα-RlucII/GFP10-Gγ₁ BRET biosensors (A) and RlucII-Gγ₅/GRK2-GFP10 BRET biosensors (B). Data points are the mean ± SEM of three to five independent experiments performed in triplicate. Ghrelin is a full agonist activating isoforms of all Gα protein families, whereas YIL781 and Abb13d signal only via proteins of the Gα_q/11 and Gα_12/13 families.

Fig. 3.

GhrR-mediated signaling through β-arrestins. (A and C) YIL781 and Abb13d antagonize β-arrestin 1 recruitment (A) and β-arrestin 2 recruitment (C) induced by 30 nM ghrelin. (B and D) YIL781 and Abb13d are weak inverse agonists in terms of their intrinsic activity to recruit β-arrestin 1 (B) and β-arrestin 2 (D). Data points are the mean ± SEM of five to eight independent experiments performed in triplicate.

Fig. 4.

Substitutions in the GhrR that affect the potency of Abb14c and YIL781. In the helical wheel of the GhrR, residues are indicated with the following color code: red, selective potency shift for Abbott 14c; blue, selective potency shift for Bayer YIL781; black, potency shift for both Abbott 14 and Bayer YIL781. Residues that are substituted and tested but do not affect the signaling properties are indicated in gray. Below the helical wheel is shown a molecular model of the ghrelin receptor based on the X-ray crystal structure of the neurotensin receptor 1 (PDB 4GRV) shown from the extracellular side, where the seven helical domains are presented as cartoon with selected residues shown in stick representation. ECL2 was omitted for clarity. The image was made with ICM, Molsoft L.L.C. Dose–response curves are shown for selected GhrR mutants (solid lines) compared with the wild-type GhrR (dotted lines). Data points are given as the mean ± SEM of three to nine independent experiments performed in triplicate.

Fig. 5.

Effects of YIL781 and Abb13d on food intake and gastric emptying. (A and B) Food intake in rats after IVC administration of YIL781 and Abb13d just before onset of dark phase (A) and during the light phase (B) (vehicle, n = 7; compound, n = 8). (C and D) Gastric emptying, measured as plasma acetaminophen, in mice after oral administration of YIL781 and Abb13d in wild-type mice (n = 8) (C) and GhrR-deficient mice (n = 8) (D). *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 based on unpaired Student’s t test (A and B) or one-way ANOVA (C and D). Data represent mean ± SEM. 13d, Abb13d; veh, vehicle; YIL, YIL781.

Fig. 6.

Food intake in Gα_q/11-KO mice. (A) Schematic illustrating the mouse model lacking Gα₁₁ globally and Gα_q specifically in AgRP neurons (Gα_q/11-KO) and the global Gα₁₁-KO mouse used as control. (B) Food intake under basal conditions, averaged over 5 d in Gα_q/11-KO mice (n = 8) and controls (n = 8). (C and D) Food intake 1 h (C) and 2 h (D) after s.c. ghrelin administration (2 mg/kg) in Gα_q/11-KO mice (n = 12) and controls (n = 18) just before the dark phase. (E) Food intake 4 h after ICV administration of YIL781 during the light phase in Gα_q/11-KO mice (n = 6) and controls (n = 7). *P < 0.05, **P < 0.01, ****P < 0.0001 based on two-way ANOVA followed by Tukey’s post hoc test. Data represent mean ± SEM.