Insulin: the first peptide medicine

Insulin is a peptide of 51 amino acids. The pancreas makes it in the beta cell, and its job is to tell muscle, fat and liver that glucose has arrived. It was the first peptide medicine: a chain you could hold in a vial, with a sequence, a mass, and a patient whose blood sugar moved. Before 1922, type 1 diabetes was a wasting disease with a short clock. After the Toronto extract, it became a disease you could treat with a molecule. That is why every later peptide drug, including the incretins, is standing on this one.

In short. Insulin is 51 amino acids, made in the pancreas. It was the first peptide you could give as a medicine, and it changed diabetes from a short illness into a treatable one.

Frederick Banting and Charles Best worked through the summer of 1921 in J. J. R. Macleod's laboratory at the University of Toronto. The idea was blunt. The pancreas both makes insulin and digests it, so if you tie off the ducts and let the digestive tissue waste, an extract of what remains might keep the hormone. J. B. Collip then cleaned that extract until it could be given more than once. On 11 January 1922 Leonard Thompson, aged 13, received the first injection at Toronto General Hospital. The preparation was impure, he reacted to it, and his sugar barely moved. On 23 January a cleaner extract dropped his blood glucose. Banting, Best, Collip, Campbell and Fletcher wrote it up that year in the Canadian Medical Association Journal.

In short. In 1921 a small Toronto lab made an extract of pancreas that lowered blood sugar. The first boy they treated, in January 1922, needed a cleaner batch before it worked.

The 1923 Nobel Prize in Physiology or Medicine went to Banting and Macleod. Best and Collip were not named. Banting split his prize money with Best, and Macleod split his with Collip. The argument about who did what has never really stopped, and it does not change the molecule. What changed medicine was an extract that lowered glucose in a child who was dying of the lack of it.

In short. The Nobel went to two of the four people in the room. The result that mattered was a child's blood sugar coming down.

Human insulin is not one string. It is two. The A chain is 21 residues, GIVEQCCTSICSLYQLENYCN. The B chain is 30, FVNQHLCGSHLVEALYLVCGERGFFYTPKT. Three disulfide bonds hold them: one inside the A chain, between cysteines 6 and 11, and two between the chains, A7 to B7 and A20 to B19. The mass is about 5,808 daltons. Frederick Sanger worked out that order in the early 1950s. Insulin was the first protein whose amino-acid sequence was known, and the 1958 Nobel Prize in Chemistry was for that work. Dorothy Hodgkin's group published the three-dimensional structure in 1969. Sequence first, shape after. That is the order the work earned.

In short. Insulin is two short chains, 21 and 30 letters, clipped together by three sulfur bridges. Sanger read the order in the 1950s. It was the first protein anyone had fully sequenced.

For decades the vials were animal insulin. Pig insulin differs from human insulin at one position, the last residue of the B chain. Beef insulin differs at three. They worked, because the receptor is forgiving of those few letters, and they were what you could extract from a slaughterhouse. The human sequence only became the ordinary medicine once bacteria could be instructed to make it.

In short. For years the medicine came from pigs and cattle. Pig insulin is almost the human chain, off by one letter. It worked well enough to carry the century.

The beta cell does not build the two chains separately and hope they find each other. Donald Steiner showed in 1967 that the cell first makes one longer chain, proinsulin. Human proinsulin is 86 residues: the B chain, a connecting peptide, and the A chain, with a few extra basic residues at the cuts. The connecting piece, the C-peptide, is 31 amino acids. Folding and the disulfide bonds happen while it is still one chain. Enzymes then cut the C-peptide out, and insulin and C-peptide leave the cell together, in the same amounts. A ribosome product is longer still. A signal peptide of 24 residues leads proinsulin into the endoplasmic reticulum and is removed. The hormone in the blood is the trimmed 51.

In short. The cell makes one long chain, folds it, then cuts out the middle. What you measure in blood is the two chains that are left, plus that discarded middle piece.

C-peptide is useful because of that cut. Injected insulin does not come with a C-peptide. Insulin from your own beta cells does. A C-peptide measurement is therefore a way to ask how much insulin a person is still making, without being fooled by insulin they were given. The half-life of insulin itself in the blood is short, on the order of five minutes, because liver and kidney clear it. The hormone is a pulse, not a bath. That is why replacement has always been a question of timing, and why a fatty-acid tail on a later incretin is a different trick: it makes a long-lived signal, not a copy of this five-minute pulse.

In short. Your own insulin arrives with a companion fragment called C-peptide. Injected insulin does not. Insulin itself only lasts a few minutes in the blood.

The insulin receptor is a receptor tyrosine kinase. It sits on the cell as a pair of pairs, two alpha subunits outside and two beta subunits crossing the membrane, and when insulin binds, the inside of the receptor phosphorylates itself and then a set of insulin-receptor substrates. That is a different machine from the receptors GLP-1 and GIP use. Those are class B G-protein-coupled receptors on the beta cell. They work through cyclic AMP, and they ask the beta cell to release insulin when glucose is already high. They are not insulin. Insulin lowers glucose whether or not you have just eaten. GLP-1 and GIP request insulin only in that high-glucose setting, and GLP-1 also slows the stomach and reduces appetite.

In short. Insulin docks on a kinase receptor and moves glucose into tissues. GLP-1 and GIP are different hormones. They ask the pancreas to release insulin when sugar is high.

Tirzepatide (LY3298176, the chain in Mounjaro) occupies the GIP receptor and the GLP-1 receptor. Retatrutide (LY3437943) occupies those two and the glucagon receptor as well. Glucagon is the other pancreatic peptide, 29 residues, from the alpha cell, and it pushes glucose out of the liver. None of these is the 51-residue insulin molecule. A trial mean for tirzepatide or retatrutide is a result about those engineered chains, in the people who were randomised, on that protocol. It is not an insulin result, and it does not rewrite 1922.

In short. Mounjaro and retatrutide copy gut hormones, not insulin. They are longer, engineered chains with a different receptor. Do not read their trial numbers as insulin numbers.

Animal insulin carried patients for sixty years. The limit was supply, allergy in some people, and the fact that you were always one species away from the human sequence. In the late 1970s Genentech built a synthetic gene for the human chains and had bacteria express them. Eli Lilly scaled it. On 28 October 1982 the US Food and Drug Administration licensed Humulin. It was the first medicine made by recombinant DNA. The sequence in the vial could finally be the human sequence, letter for letter, without a pancreas from a pig. Later analogues (lispro, aspart, glargine, degludec) edit a few residues or add a tail so the rise and fall is faster or flatter. They are still insulin. They are still that receptor.

In short. In 1982 a human insulin gene, copied into bacteria, became a licensed medicine. Later insulins change a few letters so they act faster or last longer. They are still insulin.

Growth hormone is the other early peptide hormone people put beside this story. Somatropin is 191 residues, one chain, a different receptor, a different gland. It is not a long insulin. The incretins are not a long insulin either. What they share with insulin is the class: a defined amino-acid sequence, a mass you can check, a receptor you can name. What they do not share is the 1922 result, the 51-letter structure, or the tyrosine-kinase receptor.

In short. Growth hormone is a different peptide, 191 letters, from a different gland. Same kind of molecule, different job.

A lyophilised research vial of a published incretin sequence is a chemical with a chromatogram. It is not Humulin, it is not the Toronto extract, and it was not the arm in any insulin trial. Reading the insulin papers is the right use of them. Treating a research listing as if it had inherited insulin's licence, or insulin's dose, is a different claim, and it is not the one on the label. Laboratory research only. Not for human consumption.

In short. The insulin story belongs to insulin, the licensed medicine. A research vial of a different chain is a reagent. It is not that medicine.