Semaglutide Mechanism of Action: What Research Reports

Which receptor does semaglutide act on?

Every mechanism discussed in this post starts at a single target: the glucagon-like peptide-1 receptor (GLP-1R). A 2018 review in Cell Metabolism described the native hormone GLP-1, released from enteroendocrine cells in the gut, as controlling meal-related glucose excursions by augmenting insulin secretion and inhibiting glucagon secretion, while also inhibiting gastric emptying and food intake (source 1). The same review organized the field around the GLP-1 receptor-positive cell types and circuits that carry those signals, and distinguished the circuits engaged by endogenous GLP-1 from those engaged by long-acting pharmacological agonists (source 1).

Semaglutide was designed to hit that receptor for much longer than the native hormone. In the 2015 Journal of Medicinal Chemistry paper that described its discovery, Lau and colleagues reported a GLP-1R affinity of 0.38 ± 0.06 nM, identified the fatty acid moiety and its linking chemistry as the key features that gave it high albumin affinity, and measured a plasma half-life of about 46 hours in mini-pigs after intravenous administration (source 2). The structural reasoning behind that design is covered in our semaglutide research overview. This post asks a different question: once the molecule is bound, what did researchers observe downstream?

At Steadfast Research Group we read this literature in layers, because each layer was measured in a different model: a purified receptor complex, a human clamp or meal study, a rodent brain, or a mouse artery. Keeping those layers separate is the only way to say precisely what each finding shows.

What does the receptor structure show when semaglutide is bound?

The most direct look at the binding event came from cryo-electron microscopy. In a 2021 study in Cell Reports, Zhang and colleagues solved structures of the GLP-1 receptor in complex with the Gs protein, bound either to semaglutide or to taspoglutide, with consensus maps at a global resolution of 2.5 Å (source 3). The semaglutide-bound structure was deposited in the Protein Data Bank as 7KI0.

The researchers reported that both peptides adopted binding modes similar to the one previously observed for native GLP-1. The difference appeared in motion rather than in a static pose: 3D variability analysis of the cryo-EM data revealed different dynamics within the bound peptides and the receptor compared with GLP-1-bound complexes (source 3). In other words, the structure placed semaglutide in the same pocket as the native hormone, while showing that the receptor-ligand assembly did not move in an identical way. How those dynamic differences relate to signaling outcomes was left as an open question rather than a settled result.

How did researchers measure semaglutide's effect on insulin secretion?

The most detailed human mechanistic study of beta cell function was published in Diabetologia in 2017 by Kapitza and colleagues (source 4). It was a randomized, double-blind, placebo-controlled trial in 75 adults with type 2 diabetes, 37 assigned to once-weekly semaglutide and 38 to placebo, with testing after 12 weeks at steady state. The investigators used four separate tests, each probing a different aspect of insulin release:

Intravenous glucose tolerance test (IVGTT)

Separates the rapid first-phase insulin response (0 to 10 minutes) from the sustained second phase (10 to 120 minutes). The estimated ratios versus placebo were 3.02 for first-phase and 2.10 for second-phase insulin (both p < 0.0001).

Arginine stimulation test

Estimates maximal insulin secretory capacity. The study reported that this capacity increased with semaglutide.

Graded glucose infusion test

Raises glucose in steps to map insulin secretion rate against glucose level. Insulin secretion rates rose to levels similar to those reported for healthy participants.

24-hour meal test

Tracks glucose and glucagon across a full day of standardized meals. Fasting, postprandial and overall glucose and glucagon responses were lower with semaglutide (p < 0.0001).

Taken together, the four tests described two linked observations: more insulin released per unit of glucose stimulus, and less glucagon released across the day (source 4). Both match the two islet actions attributed to native GLP-1 in the 2018 review (source 1).

Was the glucagon effect still present when glucose fell?

A lower glucagon level across a day of meals raises an obvious mechanistic question: did the same suppression persist when glucose dropped below normal? A 2018 study in Diabetes, Obesity and Metabolism by Korsatko and colleagues addressed it with a hypoglycemic clamp (source 5). The trial randomized 38 adults with type 2 diabetes, all on metformin only, to two 12-week crossover periods of once-weekly semaglutide or placebo, each followed by a hypoglycemic clamp starting from a plasma glucose of 5.5 mmol/L.

From 5.5 mmol/L to the nadir, the absolute rise in mean glucagon was similar in the two periods: 88.3 pg/mL with semaglutide and 83.1 pg/mL with placebo (source 5). Other counterregulatory hormones rose in both periods, with statistically smaller increases in noradrenaline and cortisol during the semaglutide period. The glucose infusion rate needed to hold the clamp was similar between periods, and the authors described the overall counterregulatory response as similar (source 5).

Read alongside the Kapitza meal-test data, the pattern reported in these two trials was lower glucagon during fasting and meals, but a glucagon rise comparable to placebo once glucose was pushed low. That is the observation behind the description of the islet effect as glucose-dependent.

What did studies report about gastric emptying?

Slower gastric emptying is one of the classic actions attributed to GLP-1 (source 1), and two trials measured it with semaglutide using paracetamol absorption as an indirect marker. In both trials, a smaller area under the paracetamol curve was read as slower emptying (sources 6, 7).

Hjerpsted and colleagues, in a 2018 crossover trial in 30 adults with obesity, assessed glucose and lipid metabolism and gastric emptying after each 12-week period (source 6). Gastric emptying over the first hour after a standardized carbohydrate-rich breakfast was 27 percent lower with semaglutide than with placebo (ratio 0.73, 95% CI 0.61 to 0.87). Over the full five hours, the difference was not significant (ratio 0.94, 95% CI 0.88 to 1.01). The study also reported improvements in fasting and postprandial glucose and lipid measures (source 6).

A 2021 trial by Friedrichsen and colleagues ran longer, randomizing 72 adults with obesity to semaglutide or placebo for 20 weeks (source 7). Using the same paracetamol method, the investigators reported no evidence of delayed gastric emptying at week 20 (source 7). The two results are not contradictory on their face: one found an effect confined to the first postprandial hour after 12 weeks, the other found no overall delay at 20 weeks. They do indicate that slower emptying was not the dominant, sustained mechanism in these studies.

What did studies report about appetite and energy intake?

Intake was also measured directly. In a 2017 crossover trial in 30 adults with obesity, Blundell and colleagues reported that total energy intake across ad libitum meals was about 24 percent lower with semaglutide than with placebo after 12 weeks, a difference of 3,036 kJ (source 8). At lunch following a standardized breakfast, intake was 1,255 kJ lower, and intake was also lower at the evening meal and in snacks. Questionnaires recorded less hunger, fewer food cravings, better control of eating and a lower relative preference for high-fat foods (source 8).

The 2021 Friedrichsen trial reported a larger gap: ad libitum energy intake was 35 percent lower with semaglutide (1,736 versus 2,676 kJ), together with lower hunger and prospective food consumption and higher fullness and satiety ratings (source 7). Because the same trial found no delay in gastric emptying at week 20, the lower intake in that study was observed without a measurable change in emptying.

Where in the brain did semaglutide act in rodent studies?

The appetite findings raise the question of where the signal is received. Gabery and colleagues, publishing in JCI Insight in 2020, reported that semaglutide did not cross the blood-brain barrier, but reached the brainstem, septal nucleus and hypothalamus through the circumventricular organs and selected sites next to the ventricles (source 9). That access route is summarized in the overview post; the mechanistic addition here is what happened next.

The study recorded c-Fos activation in 10 brain areas, including hindbrain regions that semaglutide reached directly and secondary regions without direct GLP-1 receptor interaction, such as the lateral parabrachial nucleus (source 9). In the rodents studied, semaglutide modulated food preference and reduced food intake without decreasing energy expenditure. The authors concluded that the effect ran through direct interaction with several distinct GLP-1 receptor populations plus direct and indirect effects on neural pathways involved in food intake, reward and energy expenditure (source 9).

What did animal studies report about vascular inflammation?

Beyond islets, the stomach and the brain, one mouse study examined arterial tissue. Rakipovski and colleagues, in a 2018 paper in JACC: Basic to Translational Science, used two atherosclerosis-prone strains, ApoE-/- and LDLr-/- mice on a Western diet, and tested both liraglutide and semaglutide (source 10). Both compounds significantly attenuated plaque lesion development, and the authors reported that the attenuation was partly independent of reductions in body weight and cholesterol.

To look for a mechanism, the researchers profiled gene expression in aortic tissue. Comparing the highest semaglutide group with vehicle, 94 genes in LDLr-/- mice and 12 genes in ApoE-/- mice were differentially expressed (source 10). The Western diet had shifted pathways tied to leukocyte recruitment, rolling, adhesion and extravasation, cholesterol metabolism, lipid-mediated signaling, extracellular matrix turnover and plaque hemorrhage, and semaglutide significantly reversed those changes. In a separate lipopolysaccharide model of acute inflammation, semaglutide lowered plasma markers of systemic inflammation (source 10). The authors framed this as a mechanism that includes inflammatory pathways, observed in mice.

How do the proposed mechanisms compare?

What are the limits of the mechanism evidence?

Each layer carries its own caveat. The structure is a high-resolution view of one receptor-G protein complex in vitro; it shows where the peptide sits, not what a whole cell or tissue does. The human mechanistic trials were run by investigators using a licensed pharmaceutical product under regulatory oversight, in specific populations (adults with type 2 diabetes or obesity), over 12 to 20 weeks; they are summarized here strictly as published literature and say nothing about research-grade material. Gastric emptying was inferred from paracetamol absorption, an indirect marker, and the two trials that used it reported different results at different time points. The brain and vascular findings come from rodents, in induced or genetically modified models.

The layers also do not yet connect cleanly. The structural paper did not tie its dynamic differences to a specific downstream response, and the atherosclerosis study described its inflammatory mechanism as partial. For a broader view of how semaglutide fits among related compounds, the GLP-1 receptor agonist class overview and the semaglutide and tirzepatide comparison place these findings alongside dual-receptor agonists.