Born as a contaminant
Glucagon entered medicine as a nuisance. The first insulin extracts of the early 1920s had an annoying habit: before blood sugar fell, it briefly went up. In 1923 Charles Kimball and John Murlin traced the rise to a second substance in the pancreatic extract and named it for what it did, the glucose agonist, glucagon (Lefèbvre, 2011). For the next quarter-century it was treated mainly as something to purify out of insulin. Only in 1948 was it pinned to its source, the alpha cells of the pancreatic islets (Sutherland & de Duve, 1948).
So from its first day glucagon was defined against insulin: the impurity that did the opposite. That framing lasted a hundred years, shaped a generation of drug programs, and turned out to be about half right.
The case for the prosecution
Once glucagon could be measured in blood, the opposition looked like more than a curiosity. People with diabetes did not just lack insulin action. They had too much glucagon, and it stayed high after meals when it should have fallen. In 1975 Roger Unger and Lelio Orci made the argument explicit: diabetes is a disease of two hormones, and the high blood sugar comes largely from glucagon driving the liver to pour out glucose that insulin can no longer restrain (Unger & Orci, Lancet, 1975).
The genetics seemed to agree. Mice engineered without the glucagon receptor had lower blood glucose than normal (Gelling et al., 2003). More dramatically, when Unger’s group destroyed the insulin-producing cells of such mice, the animals did not develop diabetes at all. No insulin, no glucagon signal, and blood sugar stayed normal (Lee et al., 2011). The conclusion wrote itself. Block the glucagon receptor and you have a diabetes drug.
The blockers worked, and that was the problem
Several companies made glucagon receptor antagonists, and on the measure they were designed for, they delivered. In phase 2, the oral antagonist LY2409021 lowered blood glucose and HbA1c in people with type 2 diabetes over 12 and 24 weeks (Kazda et al., 2016).
Then a six-month study looked at the liver directly, by MRI. Compared with placebo, people on the antagonist ended with about 4.4 percentage points more fat in their livers. Their liver enzymes rose. So did their systolic blood pressure, by about 4 mmHg, along with body weight and total cholesterol (Guzman et al., 2017). Everything reversed when the drug was stopped, but a diabetes pill that fattens the liver and raises blood pressure and weight is treating the number while worsening the patient. The approach did not survive.
The mouse had been warning about this all along. The same knockout animals with the enviable blood sugar also had swollen islets packed with extra alpha cells and enormous levels of circulating glucagon (Gelling et al., 2003). Something in the body was pushing hard against the loss of the signal. That is rarely a sign the signal was unimportant.
What glucagon was doing all along
The explanation is that raising blood sugar is one part of glucagon’s job. Its real brief is to tell the liver what to do with fuel when food is not coming in, and that brief has at least three other lines.
It clears amino acids. Glucagon drives the liver to break down amino acids, and amino acids in turn stimulate the alpha cell. It is a feedback loop between the two organs. Block the glucagon receptor and the liver stops clearing amino acids, their blood levels climb, and the excess drives alpha cells to multiply, which is exactly the islet overgrowth seen in the knockout mice (Solloway et al., 2015). After a steak, with no carbohydrate in sight, glucagon rises so the liver can process the protein without blood sugar falling.
It burns fat in the liver. The same signal that releases glucose also shifts the liver toward oxidizing fat. Take the signal away and fat accumulates, which is what the MRI scans showed.
It raises energy expenditure. This had been on record since the 1980s, when raising glucagon in insulin-deficient volunteers increased their resting metabolic rate (Nair, 1987). A later study measured it cleanly: a glucagon infusion raised energy expenditure by 15% in healthy men, about as much as cold exposure did, and without recruiting brown fat to do it (Salem et al., 2016).
Even the opposition to insulin turned out to be conditional. After a mixed meal, glucagon and insulin rise together, and glucagon acts on the neighboring beta cells to increase insulin release (Svendsen et al., 2018). In fed mice, giving glucagon lowered blood glucose, an effect that vanished when the beta cells could not sense it (Capozzi et al., 2019). Those are animal data, but they are hard to square with a hormone whose only purpose is to undo insulin.
Glucagon is not the anti-insulin. It is the liver’s fasting instruction: release glucose, clear amino acids, burn fat, spend energy. Blocking the receptor silenced all four lines to mute one.
The turn: use it instead
If glucagon burns fat and raises energy expenditure, the question flips. Could you use it, and deal with the blood sugar some other way?
Nature had already run the experiment. The gut makes a hormone called oxyntomodulin, cut from the same precursor as glucagon and GLP-1, that activates both of their receptors at once. Given to people, it reduced food intake (Cohen et al., 2003) and, over four weeks of injections, body weight (Wynne et al., 2005). GLP-1 activity supplies exactly what glucagon lacks: it raises insulin when glucose is high and suppresses appetite. Put the two in one molecule and each covers the other’s weakness.
In 2009 two groups published engineered versions of that idea within weeks of each other, single peptides that hit both receptors and reversed obesity in mice more effectively than GLP-1 activity alone (Day et al., 2009; Pocai et al., 2009). The hormone that programs had spent years trying to block was now being designed in.
| Measure | Block the receptor | Activate it, with GLP-1 |
|---|---|---|
| Blood glucose | Lower | Held down by the GLP-1 component |
| Liver fat | Higher | Lower |
| Body weight | Higher | Lower |
| Outcome | Programs stopped | Phase 3 and beyond |
Where that leaves the evidence
Two of those designs are now far along. Survodutide, a glucagon and GLP-1 dual agonist, improved liver disease in a phase 2 trial in people with MASH and fibrosis (Sanyal et al., NEJM, 2024), the liver being exactly where a glucagon component should show itself; that case is made in the arm that points at the liver. Retatrutide adds a third receptor, GIP. In its phase 3 obesity trial, 2,339 adults without diabetes were followed for 80 weeks, and mean body weight fell by 25.0% on the highest dose against 3.9% on placebo, on the treatment-regimen analysis that counts everyone randomized (Jastreboff et al., NEJM, 2026). The design logic is laid out in the triple agonist explained.
It is worth being exact about what those results do and do not show. They show that molecules containing glucagon activity can be given for over a year, at scale, with large weight loss and without the blood sugar problem the old framing predicted. They do not isolate glucagon’s share. No arm of these trials tested the same molecule with the glucagon component removed, so the claim that glucagon is what pushes these drugs past their predecessors rests on mechanism and on comparisons across different trials. That is the bottom of the evidence ladder, however good the story.
The costs of the glucagon component are also still being counted. Glucagon speeds the heart, and it breaks down amino acids, which raises a fair question about lean mass during rapid weight loss, a concern taken up in muscle loss on GLP-1 drugs. Meanwhile the hormone’s oldest job has not gone anywhere. As a rescue for severe low blood sugar, glucagon is now available as a nasal powder that needs no mixing and worked about as well as the injection in a head-to-head crossover trial (Rickels et al., 2016).
A hundred years to read one hormone
Glucagon was discovered as the thing in the vial that worked against insulin, and for most of a century it was understood that way. The framing was strong enough to launch drugs, and the drugs were good enough to show the framing was incomplete.
- Found as an impurity. Glucagon was identified in 1923 as the contaminant in insulin extracts that briefly raised blood sugar, and defined from then on as insulin's opposite.
- Blamed for diabetes. People with diabetes have too much glucagon, and mice without its receptor resist diabetes. That made blocking the receptor an obvious drug strategy.
- The blockers lowered glucose and raised liver fat. A glucagon receptor antagonist improved blood sugar but increased liver fat, liver enzymes, blood pressure, weight and cholesterol. The approach was dropped.
- It was doing four jobs. Glucagon releases glucose, clears amino acids, burns liver fat and raises energy expenditure. In the fed state it also boosts insulin release.
- Now it is designed in. Paired with GLP-1 activity to hold glucose down, glucagon agonism is part of survodutide and retatrutide. The trials show the combinations work; they do not yet isolate glucagon's contribution.
Keep going
- Glucagon reference (receptor, source, mechanism)
- What is retatrutide? The triple agonist explained
- The arm that points at the liver — survodutide and MASH
- From insulin to the peptide boom — the century this story sits in
- Is there a GLP-4? — the proglucagon family tree
- The GLP-3 bull case — glucagon as the third axis
- Ask the research agent about glucagon
Educational reference on mechanism, summarized from public scientific literature and clinical-trial disclosures and simplified in places. Not individualized medical advice or a prescription. Where a dose or regimen appears it is reported from its source as research literacy — so you can read the evidence yourself — not an instruction to dose. Specific compounds and trials are named to explain the science; verify any claim against the linked primary sources.