The Contaminant That Became a Hormone

An impurity in crude pancreatic extracts, first noticed in 1923, turned out to be glucagon: a 29-amino-acid hormone made by the pancreas that pushes blood sugar up, in direct opposition to insulin. In 1923, in a laboratory in Rochester, New York, an animal injected with a pancreatic preparation that was supposed to lower its blood sugar showed the opposite first. The sugar rose, briefly, and then fell. The preparation was meant to be insulin. Something else was in the bottle 1.

What followed was a long, slow, unglamorous detective story: thirty years from the first sighting to a crystal, and nearly sixty to a place in the headlines. It is a story about how biology can progress by taking an irritating artefact seriously, and about a hormone that was, for much of the twentieth century, known mainly as the thing insulin was not.

An insulin extract that did the wrong thing

Insulin was new in 1922. The Toronto team of Frederick Banting, Charles Best, James Collip and John Macleod had shown that an extract of pancreas could pull blood sugar down in diabetic animals and people, and laboratories around the world were racing to reproduce and scale the preparation. Making insulin in bulk meant extracting animal pancreas, and the extracts were crude: a mixture in which the active principle was one component among many.

Those extracts were tested by their effect on blood sugar, and testing revealed a nuisance. Sometimes, immediately after the injection, blood sugar rose rather than fell. The effect was transient and was followed by the fall the extract was meant to produce, so it did not threaten the usefulness of insulin. It was nevertheless a puzzle, and a disturbing one for anyone trying to standardise a medicine by a blood-sugar assay 1.

Kimball and Murlin name the 'glucose agonist'

Charles Kimball and John Murlin, working at the University of Rochester, took the artefact seriously. In 1923 they published an account of pancreatic extracts in which they separated a hyperglycaemic fraction from the hypoglycaemic one, and proposed that it was a distinct substance. They called it glucagon. The name is usually explained as a contraction of 'glucose agonist', a substance that acts in the same direction as glucose, and the choice is a small piece of history in itself: they had the idea of an opposing hormone and a name before anyone had the molecule 1.

The reception was cool, and not for bad reasons. The assay was an animal one, with all the variability that implies. Pancreatic extracts contained many things. A rise in blood sugar might be produced by other substances released by the injection, by stress to the animal, or by an effect on the liver that had nothing to do with a pancreatic hormone. Without a pure substance, the claim to a new hormone could not be tested in the strong sense. It would take three decades to supply one.

Decades as a laboratory nuisance

Through the 1920s and 1930s the hyperglycaemic factor was reported, rediscovered, doubted and renamed in a handful of laboratories. For the manufacturers of insulin it was simply a contaminant. Better purification of insulin tended to remove it as a side effect, and its presence in cruder lots was one reason early preparations could behave inconsistently.

It is worth being clear about the evidence status in this long middle period. Everything known about the factor came from animal assays on impure material, supplemented by a small number of human observations of blood sugar after injection of extracts. There was no structure, no pure preparation and no way to measure the substance in blood. The thing that was attracting attention was an effect, not a molecule.

Sutherland and de Duve find where it comes from

The turning point in the argument about origin came in St Louis. In the late 1940s Earl Sutherland and Christian de Duve, working in the laboratory of Carl and Gerty Cori at Washington University, published a study of where in the body the hyperglycaemic-glycogenolytic factor was made. Their evidence, from extracts of tissues, pointed to the alpha cells of the pancreatic islets as the source, and to the same material appearing in extracts of the stomach and upper gut 2.

The two men were at the beginning of careers that would be recognised at the highest level. De Duve later shared a Nobel Prize in 1974 for his work on the organisation of cell contents. Sutherland, who kept studying how the factor raises blood sugar by stimulating the liver to break down stored glycogen, found that the process runs through a small intracellular molecule, cyclic AMP, and received the 1971 Nobel Prize for the second-messenger concept. In a small historical irony, the contaminant in the insulin bottle ended up as a route to a central idea in cell signalling.

Indianapolis, 1953: a crystal and a sequence

The company that supplied much of the world's insulin was also in a position to settle the question of purity. Eli Lilly processed enormous quantities of pig pancreas, and glucagon was a by-product of that work. In 1953 Alexander Staub and colleagues at Lilly in Indianapolis obtained glucagon as crystals, a landmark because a substance that crystallises is, to a chemist, a single substance rather than a mixture. The paper that set out the purification and crystallisation appeared in 1955 3.

The sequence followed within a few years from a team at the same company, led by William Bromer. Glucagon turned out to be a chain of 29 amino acids, about half the length of insulin, which has 51 and two chains. A peptide that had been a source of embarrassment was now one of the best-characterised hormones in existence, and its small size would later make it practical to synthesise and modify chemically. A form of the hormone for emergency use in severe low blood sugar was produced from this work, and recombinant production in yeast followed in the late 1990s.

Unger's assay and the 'bihormonal' idea

With a pure peptide in hand, the next advance was measurement. In Dallas, Roger Unger developed an immunoassay for glucagon that could detect it in blood, and with it began to ask whether the hormone changed in disease. His answer, developed across the 1960s and 1970s and set out at length in a two-part review in 1981, was that in diabetes glucagon was not a bystander 4.

Unger and his colleagues argued for a bihormonal view: that diabetes involves too little insulin and too much glucagon, so that the liver keeps releasing glucose when it should be holding on. This was a human observational and physiological argument, built on assays of plasma and on experimental manipulation of the two hormones. It was also contested, and later work refined it substantially. What it established was the principle that glucagon belonged in the explanation of glucose control as a regulatory hormone, not as a contaminant to be filtered out 4.

From rescue medicine to a component of 2026's multi-agonists

The last turn in the story came from genetics. In 1983 Graeme Bell, Joel Habener and their colleagues reported the sequence of the gene that encodes glucagon's precursor, and found that the same precursor also contained the sequences of two related peptides, which became known as glucagon-like peptide 1 and glucagon-like peptide 2 5. Cells in the pancreas and cells in the gut cut that single precursor into different products, which explained why gut extracts had shown glucagon-like activity decades earlier.

That discovery connected glucagon to the other major peptide story of recent decades. GLP-1 became the basis of the incretin-based medicines now prescribed for diabetes and obesity, and the pharmacology of several newer investigational agents combines activity at the GLP-1 receptor with activity at the glucagon receptor. Their trial results, and the regulatory status of each, belong to the trial registries and to the pharmacology articles on this network, and should be read there as they stand on the date of reading. What the history offers is the observation that the same precursor holds both a hormone raising glucose and one that helps lower it.

Where the pharmacology articles begin

This article stops where mechanism begins. It has said what was found, by whom, and on what kind of evidence: animal assays in 1923, tissue studies in 1948, chemistry in the 1950s, human plasma measurements in the 1960s, molecular genetics in the 1980s. It has not said what any of the modern receptor-targeting compounds do in people, because that question is answered, to the extent it is answered, by the trial literature rather than by history.

One pattern is worth carrying into that literature. Each stage in glucagon's rehabilitation depended on a better tool: a purer preparation, a better assay, a clone. The pace of the science was set by the instruments, and the confidence a claim deserves can still be read from the instrument behind it.