Several compounds in this catalog act on one signalling system. Semaglutide, tirzepatide, retatrutide and survodutide have been compared here on receptor count, chemistry, half-life and trial design. What has never been laid out is the system itself — what the incretin axis does when nobody is administering anything.
That turns out to be more counterintuitive than the marketing suggests. The natural incretin signal is largely destroyed before it reaches the bloodstream, and much of what remains is carried by nerves rather than circulating hormone. A long-acting agonist is not a scaled-up copy of the physiological signal. It engages the same receptor from a different compartment, on a different timescale.
The Axis Was Defined by a Subtraction, Not a Molecule
The incretin effect is an observation older than the molecules that produce it. Give glucose by mouth and measure the insulin response. Then give it intravenously, titrated so the blood glucose curve matches the oral one exactly. The responses differ, and whatever accounts for the difference is not glucose, because glucose was held identical by design.
Nauck, Stöckmann, Ebert and Creutzfeldt quantified this in Diabetologia (1986;29:46-52). In healthy control subjects, the incretin contribution to the total insulin response was 72.8 ± 6.9%; in the type 2 diabetic patients, 36.0 ± 8.8%. Note what kind of number that is: a difference between two experiments in the same person, not a measurement of a hormone. The molecules were identified afterwards, to explain the gap.
The deficit is not shared equally between the two incretins. Nauck and colleagues later infused GIP and GLP-1 separately under hyperglycaemic clamp (Journal of Clinical Investigation 1993;91:301-307). Both augmented insulin secretion, but GIP's maximum effect in the diabetic patients was 54% lower than in normal subjects, while GLP-1 reached 71% of normal C-peptide increments, not a significant difference. The cohort had mild disease and the higher infusion rates were supraphysiological. That asymmetry is much of why GLP-1 became a drug target decades before GIP.
Two Hormones, Two Cell Types, One Gene Trick
The gap is filled by two peptides from two enteroendocrine cell populations scattered through the intestinal epithelium.
GIP — glucose-dependent insulinotropic polypeptide, historically gastric inhibitory polypeptide — comes from K cells, concentrated in the proximal small intestine. It was the first incretin identified.
GLP-1 comes from L cells. The old textbook picture placed these exclusively in the distal ileum and colon, which sat awkwardly with how quickly GLP-1 rises after a meal; the current picture distributes them along the gut, proximal populations included.
The elegant part is the gene. GLP-1 is encoded in proglucagon, the transcript that also yields glucagon. The difference is which enzyme cuts. In pancreatic alpha cells, prohormone convertase 2 liberates glucagon. In intestinal L cells, prohormone convertase 1/3 cuts the same precursor elsewhere and yields GLP-1, GLP-2, glicentin and oxyntomodulin. One precursor, two sets of scissors, opposite metabolic messages — the architecture described in the melanocortin system explainer, where POMC becomes ACTH or α-MSH depending on a cell's convertases.
It is also why glucagon receptor and GLP-1 receptor agonism sit together in one molecule like survodutide without contradiction. The ligands were always relatives.
An L cell is a chemosensor with its apical face in the gut lumen and its secretory machinery pointed at the tissue below. The best-supported glucose sensor is SGLT1, whose transport is electrogenic: moving sugar with sodium depolarises the cell, and knockout work shows metabolic sensing plays only a minor role. FFAR1/GPR40 and GPR119 handle lipid-derived ligands, FFAR2/FFAR3 short-chain fatty acids from microbial fermentation, and the calcium-sensing receptor peptones and amino acids. Sweet taste receptor subunits are reported here too, but their contribution is contested, with human sweetener studies cutting against a meaningful role.
Whichever the trigger, the endpoint is the same: depolarisation, calcium entry and regulated exocytosis of pre-formed granules — the machinery covered in the SNARE complex explainer.
"Glucose-Dependent" Is a Mechanism, Not a Slogan
Both incretins carry glucose dependence in their pharmacology, and it falls out of how the beta cell works rather than from anything clever about the peptides.
Insulin secretion is initiated when glucose metabolism raises the ATP/ADP ratio, closing ATP-sensitive potassium channels, depolarising the beta cell and opening voltage-gated calcium channels. Glucose owns that trigger.
Incretin receptors are class B G-protein-coupled receptors signalling through Gs and cAMP, acting via protein kinase A and Epac2 on granule priming and calcium handling. That machinery amplifies a secretory event; it does not initiate one. When glucose is low the channels stay open, the trigger is absent, and there is little for cAMP to amplify.
This is the receptor pharmacology primer point exactly: what a receptor does depends on the state of the system it is wired into, and the same occupancy produces a large effect or almost none.
Insulin is not the only output. Endogenous GLP-1 also slows gastric emptying, suppresses glucagon glucose-dependently, and acts on central intake circuits. The gastric-emptying arm is prone to tachyphylaxis under continuous exposure, so it differs between short- and long-acting agonists — the timing-as-a-variable point from the hormonal class primer.
The Spine: Most of the Endogenous Signal Never Makes It Into the Blood
Here is the finding that reorganises how the axis should be read.
Hansen, Deacon, Ørskov and Holst (Endocrinology 1999;140:5356-5363) traced newly secreted GLP-1 out of the porcine ileum. GLP-1 extractable from the tissue was 94.6 ± 1.7% intact — the L cell stores the active form. But of what had actually been secreted, only 32.9 ± 10.8% remained intact in vivo. A DPP-4 inhibitor pushed the released fraction back to 86-101%. Staining put DPP-4 in the capillary endothelium, with positive capillaries sitting directly against GLP-1-containing L cells.
The hormone is cut down as it enters the vessels draining the mucosa, not later in the general circulation. Add hepatic extraction downstream and little of an L cell's output reaches a distant tissue active.
So how does the signal travel? Substantially, it appears, through nerves. Vagal afferents and enteric neurons in the gut wall express GLP-1 receptors, positioned to detect the hormone locally before DPP-4 finishes with it.
Be careful with the anatomy, though, because a tidier version circulates than the evidence supports. Cao, Merchant and Gautron (Scientific Reports 2024;14:23666) hunted for direct contacts using super-resolution microscopy and found vagal axons touching fewer than 0.5% of GLP-1 cells, which carried generic secretory but not synaptic markers. Their conclusion: paracrine diffusion, not synaptic-like wiring.
The functional evidence is stronger than the anatomical, and dose-dependent in a telling way. Ahrén (American Journal of Physiology 2004;286:R269-R272) found that destroying sensory nerves abolished the insulin response to a low dose of GLP-1 in mice but not to a dose a hundred times higher. Nishizawa and colleagues (2013;305:E376-E387) quantified it in conscious rats: at a physiological infusion rate, close to 60% of the insulin response to intraportal GLP-1 disappeared after selective hepatic vagotomy — a component that shrank at pharmacological rates and vanished on intravenous delivery. Krieger and colleagues (Diabetes 2016;65:34-43) knocked down GLP-1 receptors in rat vagal afferents and saw larger meals, faster gastric emptying and blunted insulin release, calling it the "neuroincretin effect."
The counterweight: the neural route is not obligatory. Lamont and colleagues (Journal of Clinical Investigation 2012;122:388-402) restored the GLP-1 receptor only in the pancreas and normalised glucose tolerance in receptor-null mice. A third candidate is neither endocrine nor neural — GLP-1 made within the islet itself. What the literature agrees on is the negative claim: the signal is not simply a hormone travelling intact from gut to pancreas.
The consequence for reading this catalog is direct. A protease-resistant, albumin-anchored agonist is not simply more of the endogenous hormone. The native signal is local, pulsatile and terminated within minutes. A pharmacological agonist is a sustained systemic concentration reaching receptors the endogenous peptide engages weakly or not at all — a change of route and compartment, not only duration, and exactly what the Aib8 substitution and fatty-diacid anchor were built to produce (peptidase and clearance explainer).
The Brain Makes Its Own GLP-1, and That Is a Separate Source
GLP-1 in the central nervous system is not gut GLP-1 that crossed over. Preproglucagon neurons in the nucleus tractus solitarius express the same gene and process it the L-cell way, projecting to hypothalamic and other targets (Endocrinology 2025;166:bqaf125).
Circulating agonists reach some of these sites without a general blood-brain-barrier crossing, because the area postrema and neighbouring circumventricular structures are perfused by fenestrated capillaries. Which populations matter is contested. Huang and colleagues (Nature 2024;632:585-593) reported dissociable hindbrain circuits for satiety and aversion, and Teixidor-Deulofeu and colleagues (Cell Metabolism 2025;37:1530-1546.e6) found that ablating Adcyap1-expressing dorsal vagal complex neurons largely reversed semaglutide's effects on energy balance in mice, with those neurons driving fat rather than lean mass loss and only a modest conditioned taste aversion.
Both suggest appetite and malaise are at least partly separable. Unresolved is whether the hindbrain is the whole story, since other 2024 work locates a required contribution in the hypothalamus. This is mouse circuit neuroscience, rung 2 on the evidence hierarchy, and a mechanism finding rather than an outcome claim.
GIP: One Receptor, Two Opposite Strategies
GIP is the older incretin and the stranger one, with receptors well beyond the beta cell in adipose tissue, bone and brain.
The consensus review by Müller and colleagues (Molecular Metabolism 2025;95:102118), whose very large author list carries extensive industry disclosures, states it plainly: interest revived "in great part due to pharmacology demonstrating that both GIPR agonism and antagonism may be beneficial in treating obesity and diabetes." The review calls this an apparent paradox.
It is not wholly unexplained. The leading account is that the two strategies act in different places. Gutgesell and colleagues (Nature Metabolism 2025;7:1282-1298) found GIPR agonism and antagonism reduce weight by different mechanisms in male mice. Lewis and colleagues (2026;8:1669-1678) localised it with region-specific knockouts: the area postrema mediates a GIPR agonist's appetite suppression, while hypothalamic GIPR underlies the ability of GIPR antagonism to enhance weight loss from a GLP-1 receptor agonist. Antagonism also sensitised animals to an amylin receptor agonist, the mechanism cagrilintide works through.
So the honest statement is not "nobody knows," but that two anatomically separate populations of one receptor are in play, and blocking one is not the opposite of stimulating the other. That is mouse genetics from groups disclosing industry funding — a strong hypothesis, not a human result. See the semaglutide versus tirzepatide and retatrutide versus survodutide comparisons.
Frequently Asked Questions
Does a GLP-1 receptor agonist "restore" the incretin effect? Not like for like. The deficit measured by Nauck's protocol is a property of a person's meal response. An exogenous agonist supplies sustained occupancy from the systemic compartment, a different stimulus pattern from the local, short-lived, partly neural signal above. Mechanistically related is not equivalent.
Why do published "GLP-1 levels" disagree across papers? Because the assay decides what is counted. After DPP-4 cleavage, GLP-1(9-36)amide keeps its C-terminus, so a C-terminal assay reports total immunoreactive GLP-1 while one requiring an intact N-terminus reports only the active form. Same rule as HPLC purity, EC50 and half-life on this site: the number belongs to its method.
Related reading: the metabolic class primer, /quality/, and the /library/.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.