I want to bring in the clinical pharmacology perspective. Even if we achieve perfect Gs-biased agonism, we need to consider whether the clinical endpoints actually segregate along the Gs vs. β-arrestin axis. Here's what we think each pathway contributes clinically: Gs-dependent effects: - Acute glucose-dependent insulin secretion - Glucagon suppression - Gastric emptying delay - Central appetite suppression (likely — NTS/hypothalamic cAMP) - Cardioprotection (endothelial NO, anti-inflammatory) β-arrestin-dependent effects: - β-cell proliferation/survival (ERK1/2 pathway) - Receptor internalization/desensitization - Possibly some anti-inflammatory effects via NFκB scaffolding If the primary clinical goals are glycemic control + weight loss + cardiovascular protection, these are predominantly Gs-mediated. The β-cell preservation effect is important but operates on a longer timescale. > "In a head-to-head comparison of Gs-biased versus balanced GLP-1R agonists in diabetic cynomolgus monkeys, the biased agonist achieved equivalent HbA1c reduction with 40% less nausea (as measured by kaolin consumption), suggesting that some GI side effects may be β-arrestin-mediated." > — Griffith et al., *Diabetes*, 2019; 68(Supplement 1):Abstract 81-OR

That NHP data is fascinating and suggests a possible separation of efficacy from tolerability — the holy grail of biased agonism. One more structural insight worth mentioning: the GLP-1R is a Class B GPCR with a large extracellular domain (ECD) that acts as an "affinity trap." The initial binding event involves the C-terminus of the peptide ligand docking into the ECD, followed by the N-terminus engaging the transmembrane domain (TMD) core — the "two-domain model." The bias-relevant interactions are primarily in the TMD engagement step. This means you could potentially modify the peptide N-terminus (residues 7-14) to tune bias while leaving the C-terminus and fatty acid linker (for PK) unchanged. > "The two-domain binding model for Class B GPCRs suggests that ligand bias at GLP-1R is primarily determined by TMD interactions, with the ECD serving as an affinity anchor that is agnostic to signaling pathway selection." > — Zhao et al., *Trends in Pharmacological Sciences*, 2020; 41(12):1007–1019 This is good news for drug design — it means PK optimization (via C-terminal modifications) and bias optimization (via N-terminal modifications) can potentially be performed independently.

To summarize where I think the field stands on biased agonism at GLP-1R: What we know: - Bias can be engineered into GLP-1 analogs through rational design - Gs-biased agonists show less desensitization in vitro - Preclinical data (including NHP) suggests potential for reduced GI side effects - β-arrestin signaling contributes to β-cell trophic effects What we don't know: - Whether in vitro bias factors translate to in vivo tissue-level bias - The optimal degree of bias for maximal clinical benefit - Whether long-term β-cell preservation is clinically meaningful in patients already on GLP-1RAs (since they're continuously treated anyway) - Whether CNS effects (appetite suppression) are differentially regulated by Gs vs. β-arrestin What to watch for: - Several biased GLP-1R agonists are in Phase 1 (most undisclosed structures) - Novo Nordisk and Eli Lilly both have biased agonist programs - The first clinical readouts should come in the next 2-3 years This is one of the most scientifically exciting areas in incretin pharmacology right now.

Great summary. One final thought — the dual/triple agonists (tirzepatide, retatrutide) add another layer of complexity. Tirzepatide is actually a biased agonist at GLP-1R (favoring cAMP over β-arrestin) in addition to being a GIP/GLP-1 dual agonist: > "Tirzepatide exhibited imbalanced agonism at GLP-1R, showing reduced β-arrestin-2 recruitment (approximately 4-fold reduced Emax) relative to cAMP generation compared to native GLP-1, suggesting that its superior clinical efficacy may partly derive from biased GLP-1R signaling." > — Willard et al., *Nature*, 2020; 587:560–564 So tirzepatide's advantage over semaglutide may not be JUST about GIP receptor co-agonism — it may also reflect a more favorable signaling profile at GLP-1R itself. Disentangling these contributions is going to keep pharmacologists busy for years.