I think about origins more than most physicians probably should.

When I prescribe a patient a pen of tirzepatide or semaglutide, I'm aware — in a way that's almost vertiginous — that the molecule at the end of that needle is the product of a hunt that began before anyone had the vocabulary to describe what they were chasing. More than a century of work. Several dead ends. A peptide chemist whose name most people have never heard. And, eventually, a desert lizard.

In the last two years, the scientists behind that hunt have started collecting the field's highest honors — the 2024 Lasker–DeBakey Award among them. Most of my patients have heard of Ozempic. Almost none of them could name a single person who made it possible. That's worth fixing, because the real story is better than the myth, and the myth has already erased one of its most important characters.

In 1902, working at University College London, William Bayliss and Ernest Starling discovered that acidic food entering the small intestine triggered the pancreas to secrete — via something carried in the blood. They called it secretin. It was the first hormone ever described, and it established a radical idea: the gut could send chemical messages to distant organs.

By the 1960s, the idea had a sharper edge. Researchers noticed that glucose taken by mouth produced a far larger insulin response than the same amount of glucose given intravenously — even when blood sugar levels matched. The gut, in other words, was telling the pancreas to release insulin before the bloodstream demanded it. This “incretin effect” accounted for somewhere between half and two-thirds of the insulin released after a meal. Some hormone was responsible. Nobody knew which.

The criteria, eventually formalized, were precise: an incretin had to be released from the gut in response to carbohydrate, it had to stimulate insulin, and — critically — it had to act only when blood glucose was elevated. That last property is the one that matters most for what came next. It is the difference between a drug that lowers your blood sugar to dangerous levels and one that simply restores a broken signal.

For decades, the search turned up one good candidate, GIP, which fit the definition but flopped in patients with diabetes. So the hunt continued.

Joel Habener, an endocrinologist at Massachusetts General Hospital, wasn't hunting for an incretin at all. He wanted to understand how the body manufactures the hormone glucagon — insulin's counterweight, the one that raises blood sugar. To get a rich source of glucagon-producing cells, he made an unusual choice: he studied fish, whose islet cells conveniently cluster in a separate organ. In 1982, his lab reported that glucagon is carved out of a larger precursor protein — one that also encoded a second, previously unknown peptide sitting quietly alongside it.

A year later, Graeme Bell's group cloned the mammalian version of the gene and gave that mystery peptide a name: glucagon-like peptide-1. GLP-1. On paper, it looked promising. But when researchers tested the full-length molecule — GLP-1(1-37) — for an effect on insulin, they got nothing. Repeatedly. The peptide that would eventually reshape modern medicine appeared, at first, to do absolutely nothing.

The reason it looked inert is the part of the story that almost got lost.

Svetlana Mojsov was a peptide chemist — trained at Rockefeller under Bruce Merrifield, who won a Nobel Prize for inventing the method of building peptides one amino acid at a time. She arrived at MGH in 1983, just as Bell's paper appeared, and she became fixated on a hunch: that the inactive full-length GLP-1 wasn't the real hormone. The body, she suspected, was cutting it down to a shorter, active form.

In her notebook, she drew an arrow. It marked a specific spot — a cleavage site after the sixth amino acid — where she believed the body trimmed GLP-1 into something biologically alive. Then she did the painstaking work to prove it. She synthesized seven different versions of the peptide by hand, built the antibody assays to detect them, and fractionated extracts from pancreas and intestine to see which form actually showed up in tissue.

It was the trimmed version — the (7-37) and (7-36) amide forms — that the body produced in abundance. In 1986, Mojsov and colleagues published the finding that the full-length peptide was merely a prohormone, a precursor, and that the real actor was the shorter fragment. The conclusion was stated so modestly you could almost miss it. It was, in fact, the hinge on which everything turned.

Daniel Drucker, then a young postdoctoral fellow who had joined Habener's lab in 1984, contributed to this 1987 wave with a paper showing GLP-1(7-37) stimulated insulin in an islet cell line. He would go on to become one of the field's most prolific and thoughtful translational scientists. But the foundational biochemistry — the arrow in the notebook, the seven synthesized peptides, the prohormone insight — was Mojsov's. A 2023 STAT investigation documented how thoroughly she was written out of the popular Ozempic narrative, her name omitted from accounts and even from some patents. The 2024 Lasker Award put it back where it belongs, naming Habener, Mojsov, and Knudsen.

Identifying the hormone was not the same as having a drug. Native GLP-1(7-37) has a half-life in the blood of roughly one to two minutes. An enzyme called DPP-4 — characterized by Carolyn Deacon and Jens Juul Holst in Copenhagen — clips it almost the instant it's released, and the kidneys clear what's left. To get a clinical effect, early researchers had to deliver it by continuous infusion. You cannot build a once-weekly injection out of a molecule that vanishes before you finish the sentence describing it.

The field needed something with GLP-1's biology but none of its fragility. Two solutions emerged — one rational, one venomous.

Claude can be your analyst, editor, and strategist.
But most professionals are using it to fix grammar.

These 200+ Claude prompts take it from grammar tool to your most powerful AI work assistant.

Sign up for Superhuman AI and get:

Claim your free prompts

In the early 1990s, John Eng — working at a Bronx VA hospital — was screening venoms for bioactive peptides. He focused on the Gila monster (Heloderma suspectum), a venomous desert lizard that eats only a handful of meals a year. In its venom he found exendin-4: a 39-amino-acid peptide that bound the GLP-1 receptor and triggered insulin secretion, but with one decisive advantage. A quirk in its structure blocks the DPP-4 cleavage site, extending its half-life to roughly 2.5 hours — orders of magnitude longer than the native hormone.

Eng partnered with a small biotech, Amylin Pharmaceuticals, to develop synthetic exendin-4. The large drug companies had watched the GLP-1 field with interest but were wary of a compound derived from lizard venom. Amylin pressed on.

The other solution came from Lotte Bjerre Knudsen and her team at Novo Nordisk, and it is a small masterpiece of protein engineering. Instead of borrowing from a lizard, they kept the human GLP-1 sequence almost intact — minimizing the immune response — and bolted on a fatty-acid chain via a linker. That chain causes the molecule to cling reversibly to albumin, the most abundant protein in blood. Tucked against albumin, the drug is shielded from DPP-4 and from the kidney, and it circulates for far longer.

A C16 fatty acid produced liraglutide, with a half-life of about 13 hours — once-daily dosing. Swapping in a longer C18 di-acid chain with an optimized linker, plus a tweak that blocks DPP-4 cleavage outright, produced semaglutide: a half-life of roughly seven days. Once a week.

Here is the detail I love most. From the very beginning, Knudsen believed this would be an obesity drug, not just a diabetes drug — and she believed it against internal resistance. Her conviction rested on older, almost forgotten observations: certain glucagon-producing tumors caused profound loss of appetite in animals, and injecting GLP-1 directly into the brains of rats sharply reduced how much they ate. She was reading the central nervous system signal years before the trials confirmed it. She turned out to be exactly right.

When patients took these drugs for diabetes, they lost weight — and the magnitude, confirmed across large trials, was striking. The mechanism is not mainly that GLP-1 slows gastric emptying, though it does. The more clinically significant action is central. GLP-1 receptors are expressed throughout the hypothalamus and brainstem; activating GLP-1R-positive neurons in the arcuate nucleus suppresses food intake directly, independent of the stomach. This is not willpower biology. It is circuit biology — exactly the signal Knudsen had bet on.

For those of us in obesity medicine, the implication was profound. For years we spent appointments explaining to patients that their struggle with food was not a character defect but dysregulated physiology. GLP-1 receptor agonism gave us a reliable tool that operates at the level of that underlying biology, rather than asking patients to override it by force of will alone. It does not replace nutrition, sleep, or resistance training — it is the lever that finally lets those efforts hold for people who could never make them hold before.

One of the most clinically underappreciated effects of GLP-1 receptor agonism is the quieting of what patients call “food noise” — the relentless, intrusive preoccupation with eating that so many people with obesity experience but struggle to name for their physicians.

In my own practice, when I ask patients a few weeks into therapy what has changed, the most common unprompted answer isn't the number on the scale. It's some version of: “I just don't think about food all day anymore.” The first time a patient teared up describing the silence in her head after years of negotiating with every meal, I understood the discovery story differently. This wasn't a diet drug. It was relief from a symptom we'd never had a name for — and the relief traces directly back to that arrow in Mojsov's notebook.

Here is what I want you to take from this.

The medications reshaping how I practice — and reshaping the public conversation about weight, biology, and agency — were not inevitable, and they did not come from a single eureka. They came from a question nobody could answer for sixty years, a peptide that looked inert until someone guessed where to cut it, an enzyme that destroyed the answer in two minutes, a lizard that solved that problem by accident, and a chemist at Novo who bet on the brain before the data agreed with her.

That's the origin story. Not a celebrity endorsement. Not a press release. An arrow drawn in a notebook by a scientist the story almost forgot.

It's worth remembering whose work we're actually standing on.

REFERENCES