Four men in a Toronto summer — how Banting, Best, Macleod and Collip extracted insulin

The prehistory: an idea from a 2 a.m. note

In October 1920 the 28-year-old Frederick Banting — orthopaedic surgeon in a small practice in London, Ontario, with few patients and financial worries — read in the evening an article in the 'Surgery, Gynecology and Obstetrics' journal on pancreatic pathology. The article described a case in which pancreatic duct occlusion destroyed the exocrine cells but preserved the islets of Langerhans (the endocrine tissue). Banting noted in his notebook at 2 a.m. an idea: could one, by pancreatic duct ligation, destroy the exocrine digestive enzymes that had destroyed islet secretion in earlier extraction attempts, and then isolate the islet substance?

Banting was not an endocrinologist and not a researcher. Earlier attempts to isolate a pancreatic anti-diabetes substance (Minkowski 1889, Zuelzer 1908, Paulesco 1921 in Bucharest) had either remained unclear or clinically unusable. But Banting brought the idea to John James Rickard Macleod, an established physiologist at the University of Toronto. Macleod was sceptical — the idea was not new, the methodical realisation from his view naive — but agreed to give Banting a lab, dogs and an assistant for the summer of 1921. Macleod himself left for Scotland for the summer.

Summer 1921: Banting and Best on the top floor

The assigned room was on the top floor of the Medical Building, under a roof, practically uninhabitably hot in the Toronto summer. Banting lacked an assistant; Macleod had two students — Charles Best (21) and Clark Noble — who were supposed to share the summer work. After a coin flip, Best began. He was supposed to stay only a few weeks; the one summer shift became two years and the rest of the insulin story.

The first weeks were difficult. Banting and Best operated on dogs — pancreatic duct ligation for atrophy of the exocrine cells — and then tried to obtain extracts from the remaining islets of Langerhans. The first extracts produced mixed results; some diabetic dogs improved, others died of sepsis or non-specific reactions. August was a mixture of frustration and isolated positive signals.

In November 1921 Banting showed Macleod the first data. Macleod was impressed enough to expand the programme. He obtained more dogs, more funds — and brought in a biochemist: James Bertram Collip, a 28-year-old professor from Alberta on sabbatical in Toronto.

Collip's contribution: from dog pancreas extract to clinical product

Collip was the central, often underestimated factor in the jump from dog to human. The Banting/Best extracts were biologically effective but pyrogenic — they caused fever, chills, sterile abscesses at the injection site in humans. Collip developed over the winter months 1921-1922 an alcoholic fractionation method that separated the active substance from the pyrogenic pancreatic proteins. This purification was the decisive threshold: from an experimental animal extract to a clinically usable preparation.

On 11 January 1922 Leonard Thompson — 14 years old, life-threateningly diabetic, 29 kg body weight, ketoacidotic — was treated for the first time with the Banting/Best extract. The reaction was moderate, but a local sterile abscess developed. On 23 January Thompson received the second treatment — this time with Collip's purified extract. His blood glucose fell from 28.9 mmol/l to 6.7 mmol/l, ketone bodies disappeared from the urine, his general state improved dramatically within hours. It was the first documented clinical success of insulin therapy.

The personal friction

What distinguishes the honest historical account from the popular hero story are the conflicts between the four men. Banting and Macleod had a strained relationship — Banting saw Macleod as an opportunistic administrator who secured a share of an idea that was not his. Banting and Collip argued violently over the purification method — in a documented incident in January 1922, Banting grabbed Collip by the collar. Best was loyal to Banting, but Macleod's perception was that Best was methodologically less skilled than Collip.

The 1923 Nobel Prize — Banting and Macleod — exacerbated the tension. Banting considered it scandalous that Macleod and not Best was honoured with him. He publicly shared half his prize with Best, Macleod shared his with Collip. This gesture was morally important but did not erase the professional injuries. The four men never worked productively together again.

The patent transfer and its limits

In 1923 Banting, Best and Collip transferred the patents for insulin purification to the University of Toronto for one Canadian dollar each. The reasoning was morally clear: physicians should not profit from a life-saving medicine. The University of Toronto licensed manufacture to Eli Lilly (USA) and to Connaught Laboratories (Toronto), under the condition that 'reasonable' quality at a 'reasonable' price be ensured.

This gesture has received mythical status over 100 years. It was morally impressive, but it did not prevent — as also discussed in the separate reflection article 'What 100 years of insulin teach us' — insulin remaining a high-priced medicine over the following decades. Every new iteration (recombinant, analogs, pen systems) was protected by new patents; giving the original patent away is not the same as making a medicine generic-eligible when parallel patents protect the currently used form.

What the detailed story shows

Four structural observations from the Toronto summer. First: the original idea (pancreatic duct ligation) was methodically not unique — Paulesco in Bucharest had nearly simultaneously comparable data — but the Toronto group had the institutional platform (Macleod's lab) and the biochemical purification (Collip) that made the substance clinically usable. Second: the hero story 'Banting and Best' is a simplification. Without Macleod's platform, without Collip's purification, the result would not have become clinically tangible within a year. Third: personal conflicts between the researchers were substantial and shaped later memory — historical accounts follow the memoirs of those involved, and each of these memoirs is a position.

Fourth: the speed from first idea (October 1920) to clinical application (January 1922) is unusually short for modern medicine — about 15 months. That was possible in 1920 because the regulatory barrier for an experimental therapeutic practically did not exist: no FDA phase-3 programmes, no ICH-GCP standards, no ethics committee review in today's sense. A similar speed is unthinkable in contemporary pharmacological development — rightly so, but also a reminder that today's speed of substance development is limited by its precautionary standards, not by scientific speed itself.