Modelling G protein-biased agonism using GLP-1 receptor C-terminal mutations

The glucagon-like peptide-1 receptor (GLP-1R) is a major therapeutic target for type 2 diabetes and obesity. Agonists showing bias in favour of G protein signalling over β-arrestin recruitment and GLP-1R internalisation, e.g. tirzepatide and orforglipron, have favourable clinical efficacy profiles. However, understanding of the effects of biased agonism has been hampered by differences in ligand properties such as affinity, efficacy, stability and pharmacokinetics. Here we used GLP-1R C-tail mutations that inhibit phosphorylation to mimic G protein-biased GLP-1R agonism without the need for ligand modifications.

Serine doublet phosphorylation sites in the human and mouse GLP-1R C-tails were mutated to alanine. Wild-type and mutant GLP-1Rs were examined for β-arrestin recruitment, internalisation, Gα_s activation, and signalling readouts in HEK293 cells and pancreatic β-cell models. Native GLP-1 plus oppositely biased ligands exendin-phe1 (ExF1; G protein-biased) and exendin-asp3 (ExD3; β-arrestin-biased) were used to compare ligand- and receptor-mediated biased agonism.

Loss of three C-terminal phosphorylation sites reduced GLP-1- and ExD3-mediated GLP-1R internalisation and β-arrestin recruitment to that seen with ExF1. The phosphodeficient GLP-1R showed preferential plasma membrane Gα_s activation over longer stimulations, with associated increases in whole cell cAMP generation and kinomic signalling. The distal GLP-1R phosphorylation site played a larger role in β-arrestin recruitment, and the proximal sites were more important for GLP-1R internalisation and regulating cAMP production.

Genetic changes that reduce β-arrestin recruitment and slow GLP-1R internalisation can enhance GLP-1R signalling, providing conceptual support for the use of G protein bias to improve GLP-1R agonist efficacy.