GLP-1R signalling: what does biased agonism at a class B GPCR actually mean here?

I can follow the basic signalling story: GLP-1 receptor is a Gs-coupled receptor, activation raises cAMP, protein kinase A does things, and in the beta cell that potentiates glucose-stimulated insulin secretion. What I cannot follow is the layer above that, where people talk about biased agonism, beta-arrestin recruitment, and receptor internalisation as though these were the interesting variables.

Concretely: ecnoglutide is described as cAMP-biased. Tirzepatide is described as having an unusual signalling profile at the GLP-1 receptor with weak beta-arrestin recruitment. Both descriptions are presented as advantages. But if cAMP is the productive signal and beta-arrestin recruitment leads to internalisation and desensitisation, then surely every agonist should be designed to be cAMP-biased, and the fact that they are not suggests I am missing a trade-off.

What is the trade-off? And is there any evidence that bias measured in a cell line predicts anything at the level of a person?

Worth adding the part of the mechanism that gets least attention and probably matters most for the weight effect: where in the body any of this is happening.

GLP-1R is expressed in pancreatic islets, in the stomach and intestine, on vagal afferent neurons, in the heart and vasculature, in the kidney, and at several central sites. The central sites that matter for intake are the area postrema and nucleus tractus solitarius in the hindbrain, and the arcuate nucleus of the hypothalamus.

The critical anatomical fact is that these acylated peptides do not broadly cross the blood-brain barrier. What they access is the circumventricular organs, where the barrier is fenestrated - the area postrema in particular - plus the arcuate, which sits adjacent to the median eminence. Imaging work with semaglutide found uptake concentrated in exactly those regions rather than distributed through the parenchyma [1]. Signalling then propagates onward through neural circuits rather than through direct drug access.

Within the arcuate the effect is bidirectional: activation of POMC and CART neurons, which are anorexigenic, plus indirect inhibition of the orexigenic NPY and AgRP population, largely through GABAergic interneurons rather than direct receptor expression on the AgRP cells themselves. That indirect step is important because it means the pharmacology cannot be read off the receptor expression map.

Three consequences worth holding:

• The area postrema is also the chemoreceptor trigger zone. Nausea and appetite suppression are driven from overlapping anatomy, which is why separating the two pharmacologically has proved so hard, and why "a version without the nausea" is a much harder ask than it sounds.
• Because access is via circumventricular organs, the free fraction matters enormously - which loops back to the albumin-binding trade-off. A molecule bound too tightly to albumin has less free drug available to reach these sites.
• The subjective phenomenon people describe as reduced food noise is most plausibly a downstream consequence of altered signalling in these circuits and their projections to mesolimbic reward regions, rather than a separate mechanism. It is not currently measurable other than by self-report, which limits what can be said about it.

A small but load-bearing correction to how the cAMP story is usually told: the glucose-dependence of the insulinotropic effect is not a property of the receptor, it is a property of where the pathway inserts itself.

cAMP and PKA do not trigger insulin exocytosis. They potentiate the machinery that glucose metabolism triggers - amplifying calcium-triggered granule fusion and increasing the size of the readily releasable pool. If glucose is low, the triggering signal is absent and the amplification has nothing to amplify. That is the entire explanation for why GLP-1 receptor agonists carry low intrinsic hypoglycaemia risk while insulin and sulfonylureas do not, and it is why combining them with a sulfonylurea reintroduces the risk - the sulfonylurea supplies a glucose-independent trigger.

Two further mechanistic points that follow from the same architecture:

• The alpha-cell effect is complementary: GLP-1 receptor agonism suppresses inappropriate glucagon secretion at hyperglycaemia but does not blunt the counter-regulatory glucagon response to genuine hypoglycaemia. Again glucose-dependent, again a property of pathway insertion rather than of receptor identity.
• Beta-cell responsiveness is a prerequisite. In advanced beta-cell failure the amplifying pathway has less to work on, which contributes to the smaller glycaemic effect observed in long-duration type 2 diabetes and is part of why the diabetes populations in the weight-loss trials respond less on both weight and glucose.

Standing note: none of this constitutes guidance about use in anyone. Combination-therapy hypoglycaemia risk in particular is a clinical matter that needs a prescriber who can see the rest of the regimen.