For most of medical history, a peptide was something you could only get by taking it out of an animal. You could not design one, could not reliably make one, and in most cases could not say with confidence what it was. The century that followed turned each of those limits into a solved problem, and the order in which they fell explains a great deal about the medicines that exist now.
Toronto, 1921: taking it out of the pancreas
In the summer of 1921, Frederick Banting and Charles Best began work in a borrowed laboratory at the University of Toronto, under the department headed by J.J.R. Macleod. The idea they were testing was not new — physicians had suspected for decades that the pancreas produced something that controlled blood sugar, and several groups had tried to extract it. What Banting and Best had was a procedure for getting a usable extract out without the digestive enzymes destroying it along the way.
The extracts worked in depancreatized dogs. With the biochemist James Collip's purification work, they became clean enough to give to a person, and in January 1922 a fourteen-year-old patient named Leonard Thompson received the first injections that meaningfully lowered his blood glucose. Type 1 diabetes had been, until that winter, a diagnosis measured in months. Banting and Macleod received the Nobel Prize in Physiology or Medicine in 1923; Banting shared his portion with Best, Macleod with Collip.
What nobody in that laboratory knew was what insulin actually was. Its identity as a chain of amino acids, and the exact order of those amino acids, would take another three decades — Frederick Sanger's sequencing of insulin, completed in the mid-1950s, earned him the 1958 Nobel Prize in Chemistry and proved that proteins have defined sequences at all.
1953: building one from scratch
The next barrier was synthesis. Extracting a peptide from tissue gives you whatever nature made, in whatever quantity the tissue holds. Making one in a flask means you can build molecules that don't exist in nature, and you can prove a proposed structure by constructing it and checking that it behaves the same way.
Vincent du Vigneaud, working at Cornell University Medical College, achieved this for oxytocin. He and his colleagues determined the nine-residue sequence of the hormone and then synthesized it chemically, publishing in 1953 — the first synthesis of a polypeptide hormone. The synthetic material had the biological activity of the natural hormone, which settled the structure question conclusively. Du Vigneaud received the Nobel Prize in Chemistry in 1955 for this work on biochemically important sulfur compounds, including the synthesis of oxytocin.
Oxytocin is small — nine residues, one internal disulfide bridge. Even so, the synthesis was punishing work. Each coupling step required protecting some chemical groups so they wouldn't react, joining the intended pair, then removing the protection, then purifying the intermediate before starting again. Yields fell at every stage. Scaling this approach to anything much longer was, realistically, out of reach.
Extraction tells you what nature made. Synthesis lets you ask what nature might have made instead — and check the answer.
— The Backbone editors1963: Merrifield's beads
Bruce Merrifield's insight at Rockefeller University was procedural rather than chemical, and it is the reason modern peptide science exists at the scale it does. Instead of carrying an intermediate through purification after every coupling, he anchored the growing chain by one end to an insoluble polystyrene bead. Reagents could then be washed over the bead in excess, driving each reaction to completion, and everything unwanted simply rinsed away. The product stayed put because it was tethered.
Solid-phase peptide synthesis, published in 1963, collapsed work that had taken months into days, and it automated cleanly — Merrifield built a machine to run the cycles. He was awarded the Nobel Prize in Chemistry in 1984. Nearly every synthetic peptide made since, in research or in manufacturing, descends from that method.
The modern era: engineering around the weaknesses
By the 1980s the field had a new set of problems, and they were pharmacological rather than chemical. Peptides are excellent signals precisely because the body clears them quickly — enzymes called peptidases cut them apart, and the kidneys filter out what's left. That is a virtue in physiology and a serious obstacle in a medicine, where you would like the effect to last longer than a few minutes.
Recombinant DNA technology solved the supply problem first: human insulin produced in engineered bacteria was approved in the United States in 1982, ending reliance on animal pancreases. Then came deliberate structural engineering — substituting residues at the positions enzymes attack, adding fatty-acid chains that let the molecule bind circulating albumin and hide from clearance, or altering how the molecule assembles so it releases slowly from an injection site. The insulin analogues developed through the 1990s and 2000s are the clearest example: same job as the 1922 extract, redesigned timing.
The GLP-1 receptor agonists are the most visible current illustration of the same engineering logic. GLP-1 is a naturally occurring incretin hormone released by the gut after eating; in its native form it has a half-life of a couple of minutes, because an enzyme called DPP-4 clips it almost immediately. The approved drugs in this class are structurally modified versions designed to resist that cleavage and persist far longer, which is what makes a once-daily or once-weekly medicine possible at all. We describe this here as an educational example of how a mechanism becomes a therapy — these are prescription drugs with real indications, real risks, and real contraindications, and whether any of them is appropriate for a given person is a question for that person's clinician.
A note on scope
Approved peptide therapeutics have been through the FDA review pathway, which means the evidence for their safety and effectiveness in specific populations has been examined. Many other peptides circulating in fitness and biohacking discussions have not been through anything like that process — they are investigational compounds, not approved drugs and not dietary supplements, and in most cases no adequate human safety data exists. Backbone covers the science; it does not recommend, endorse, or facilitate the use of any of it. Trial records are public at ClinicalTrials.gov.
The through-line across a hundred years is a steady shift in what the limiting factor was: first getting the molecule at all, then knowing its structure, then being able to build it, and finally being able to make it last. Each of those took a different discipline to crack. None of them made the underlying chemistry any more exotic — it is still amino acids joined by amide bonds, exactly as it was in that Toronto laboratory.