The Gila monster and the diabetes drug
In 1992, a Detroit endocrinologist isolated a peptide from the saliva of a desert lizard that pancreata had failed to produce for decades: a GLP-1 analog with long half-life. It became exenatide — the mother of all modern GLP-1 therapies.
The GLP-1 problem
By the late 1980s it was clear: the gut hormone glucagon-like peptide-1 (GLP-1) could glucose-dependently stimulate insulin secretion in type-2-diabetes patients — only when blood glucose was high. Unlike insulin or sulfonylureas, GLP-1 caused no hypoglycaemia. It seemed the perfect diabetes drug.
The problem: human GLP-1 has a plasma half-life of about two minutes. The enzyme dipeptidyl-peptidase-IV (DPP-4) cleaves it immediately. Clinically it would only be usable as a continuous infusion or with a modification that prevents DPP-4 degradation. That was exactly what researchers were looking for in the 1980s and early 1990s.
John Eng and the venom
John Eng, an endocrinologist at the VA Medical Center in the Bronx, had an unusual research question in the early 1990s: why do some reptiles — particularly the Gila monster (Heloderma suspectum) from the south-west USA and northern Mexico — produce digestive hormones in their saliva? These lizards eat only a few times a year; their pancreas-like gland in the oral cavity (the Duvernoy gland) is inactive most of the time and must rapidly ramp up at a meal.
Eng hypothesised that the saliva might contain a long-acting GLP-1-like peptide that quickly activates the lizard's pancreas on demand. With Jean-Pierre Raufman, he analysed saliva samples — and in 1992 found a 39-amino-acid peptide they named exendin-4. It had about 53% sequence homology to human GLP-1 and — crucially — a glycine substitution at position 2 that protected it from DPP-4 cleavage. The human half-life was later measured at about 2.4 hours — more than 70-fold longer than human GLP-1.
„I literally bought a Gila monster from a pet store, milked the saliva, and ran the assays. The grant agencies thought I was crazy."
From patent to medicine
Eng patented exendin-4 in 1995 in his own name — the VA Medical Center had initially not pursued the research filing. Over several years Eng sought an industry partner. In 1996 he licensed the patent to Amylin Pharmaceuticals (San Diego), who developed a synthetic version under the code AC2993.
Clinical trials began in 1996, phase 3 from 2002 (the AMIGO programme). In 2005 the drug was FDA-approved under the name Byetta — the first GLP-1 receptor agonist ever. The twice-daily subcutaneous injection was not a comfort optimum, but the clinical effect was real: HbA1c reduction plus moderate weight loss as a bonus.
The jump to the weekly form
Exenatide's short half-life remained the limiting factor. In 2012 Bydureon was approved — a PLG microsphere depot formulation of the same active that only needed weekly injection. In parallel, Novo Nordisk worked on liraglutide (daily, approved 2010) and later semaglutide (weekly, 2017). Eli Lilly developed dulaglutide (Fc fusion, weekly, 2014) and later tirzepatide (dual GLP-1/GIP, 2022). The line is direct: each of these substances is an iteration on the idea that began in the salivary gland of a desert lizard.
The curious consequence
Heloderma suspectum is protected under US federal law — it is not directly endangered, but under strict animal protection. Ironically none of the diabetes patients who take exenatide today has ever endangered a lizard: the peptide has been synthetically produced from the start. The lizard's saliva was only the teacher.
What we take away
- Half-life often matters more than potency — exendin-4 matched GLP-1 in action, but DPP-4 resistance was the breakthrough.
- Individuals with a good question and persistence can fill gaps in industry research that institutional programmes overlook.
- The patent story shows how much science economics depends on who holds first-filer protection.
- Natural-product inspiration and synthetic production are not opposites — the lizard was teacher, not raw material.