The Discovery of Insulin

by Michael Bliss

  The insulin situation in France was, if anything, less satisfactory than in Germany. Only a few French researchers experimented with insulin in the early months (the earliest seems to have been Dr. F.L. Blum of Stras bourg). France did not allow drugs to be patented, but had no quality control regulations, leading to many companies dabbling in poor-quality insulin while hospital labs made their own. When Banting visited France in the summer of 1923, he found the situation so “deplorable” and the insulin so bad that it was hopeless to even try to give consulting advice.44

  Despite the slow French progress, Paris was the site of one of the most bizarre incidents in the scientific rivalry for credit for work on insulin. At the December 23, 1922, meeting of the prestigious Société de Biologie, Eugène Gley, one of France’s most noted endocrinologists, asked that a sealed envelope he had deposited with the society in 1905 be opened. It contained a short statement, “Sur la Sécrétion Interne du Pancréas et son Utilisation Thérapeutique,” in which Gley, who had worked with pancreatic extracts back in the early 1890s, described experiments he had carried out in 1900–1901 with extracts of pancreas he had caused to degenerate by occluding the ducts. Intravenous injections of the extract had considerably diminished glycosuria in diabetic dogs and caused subjective improvement in their condition. Gley proposed more extensive research along the same lines and the need to isolate the active principle of his extracts, sealed his “cachet”, and never took up the work again. He had not had the resources to maintain the animal facilities required for such extensive work, he explained in 1922. Gley offered no explanation for his quixotic, irresponsible gesture (he ought to have at least published the idea, so that better-equipped researchers could pursue it), made no claim to be the discoverer of insulin, and congratulated Macleod as having achieved “une grande simplification” of his method.45

  Another researcher in a war-torn country who was also so short of resources that his lab could not afford the North American journals, learned about insulin from notices in a Paris medical publication. Early in February 1923, Nicolas Paulesco wrote to Banting from Bucharest asking for offprints and enclosing one of his papers.46 Banting was never a very good correspondent, was being deluged with mail and other obligations, would not have bothered to read letters in French carefully, and in any case had no more offprints. He did not answer Paulesco’s letter.

  IV

  August Krogh’s first experiments in insulin manufacture included making it from fish as well as hogs. This was because he had visited Toronto just at the peak of Macleod’s belief that fish were the wave of the future in insulin production.

  Macleod’s enthusiasm stemmed from his work at St. Andrew’s, New Brunswick, in the summer of 1922. He went there having known ever since the early publications of his fellow Scotsmen, Rennie and Fraser, that in the bony fishes (teleosts) the islet tissues are distinct from other pancreatic tissue. It was a fairly simple matter for Macleod to prepare separate extracts from the different tissues of teleosts; the extracts of islet tissue were potent, those of the zymogenous pancreatic tissue were not. Thus Macleod supplied the first experimental verification of the hypothesis that insulin was the secretion of the islets of Langerhans. He rushed his results into print just as purists, such as Cammidge, were questioning the propriety of his lab’s use of the term “insulin” for its extracts which were, after all, made from whole pancreases.47

  “I do not think I ever enjoyed two months work so much,” Macleod wrote about that summer by the sea.48 Even more exciting than the neat proof that insulin had been rightly named, was his finding that certain common varieties of teleosts, especially sculpin and angler- or monk-fish, seemed to contain large quantities of insulin which could be easily extracted. At a time when the Connaught Labs were producing only about five units of insulin from a kilogram of beef pancreas, Macleod found he could get three to four units from less than a gram of angler-fish pancreas.49 In September 1922, fish seemed to Macleod to be the answer to all the production problems Lilly and Connaught had been having with beef and pork insulin. The sea creatures would be all the more practical as a source of supply in such seafaring countries as Britain and Denmark, so Macleod had strongly suggested to both Dale and Krogh that they do extensive investigative work on fish.

  This search for an alternative source of supply underlines the fact that even as the volumes increased, North American production remained difficult, erratic, and expensive throughout 1922. Even with their new vacuum equipment, the Connaught Laboratories group at Toronto still could not make insulin in large quantities. “Like any of our previous experiments the first large scale trial turned out to be an almost complete failure,” Best wrote in mid-November, referring to Connaught’s use of yet another manufacturing process. During one of the recurring periods of optimism at Connaught in September, the Insulin Committee had made its decision to expand Canadian clinical testing outside of Toronto. But suddenly the Torontonians had to ask Lilly to ship the necessary insulin to Canada, for Connaught could not even meet the needs of Banting and other clinicians in the city. Canada did not stop relying on American insulin until the early summer of 1923.50

  The Americans were able to produce a supply for the Canadian clinics only because they were working so hard with vast amounts of money, pancreas, alcohol, skilled manpower, and rabbits. And they, too, had their discouragements, such as the constant complaining in the autumn of 1922 that Iletin lacked potency. Part of the difficulty was explained when it was realized that Toronto and Indianapolis had drifted into different rabbit tests for potency (Lilly began using fasting one-kilogram rabbits as opposed to the well-fed two-kilogram Toronto rabbit, which gave them a much weaker basic unit), but even then the Lilly product seemed to be suffering badly from rapid deterioration.51

  Chief chemist George Walden’s attempts to prevent the deterioration led to the company’s great advance in insulin production and purification. Comparing stable batches and those that deteriorated, Walden found that the degree of acidity, as measured by pH, varied from batch to batch. Marked deterioration took place within a pH range from about 4.0 to about 6.5. The British researchers in Dale’s laboratory came up with a similar realization early in their work, but were content simply to adjust the pH levels so that the dangerous range was avoided.52 Toronto had not made the crucial observation because it had stopped using tricresol in manufacture. The tricresol, it was later realized, had been throwing the solutions into the 4.0 to 6.5 pH range – the “preservative” was thus the catalyst causing deterioration. It was one of the information breakdowns between Toronto and Indianapolis that the Lilly chemists did not know of Toronto’s abandonment of tricresol.53

  Whereas the other labs had, by design or luck, found ways of avoiding the deterioration, Walden at Lilly took pains to study the process itself. He realized that the weakening of the insulin solution actually involved the gradual formation of a precipitate which contained the active principle, thereby reducing the activity of the remaining solution. At the “wrong” pH, insulin was being precipitated out of insulin solutions. Walden’s advance came in the discovery that this precipitate was a far purer, far more potent hormone than anything they had seen before. To get it, all you had to do was learn to profit from adversity: instead of avoiding the “isoelectric point” in the pH range at which insulin precipitated out, the thing to do was to go for it. Deliberately adjust pH’s to the isoelectric point so as to cause the maximum precipitation. Collect the precipitate, fiddle with it a bit more, and you had by far the best insulin yet, in Walden’s words, “a product having a stability many times as great and a purity ranging from ten to one hundred times as great as the best product hitherto obtainable.”54

  Walden gradually evolved his isoelectric precipitation method between October and December 1922. Once its possibilities were fully appreciated, Lilly had solved the production problem. This took place in the winter of 1922–23. By February 1923 the Americans were building up huge reserves of insulin. In his March correspon
dence with the British, Clowes offered to supply insulin in any quantities: “We can produce in Indianapolis a sufficient amount of Iletin to supply the entire needs of the civilized world.” Clowes’ one remaining problem was standardization. The company was spending $2,500 to $3,000 a week on standardization – running through thousands of rabbits, over a hundred thousand of them in the first six months* – to get a consistency which, though not at all bad in rabbit testing, still varied by up to 10 per cent from batch to batch.55

  In accordance with the Lilly-University of Toronto agreement, the American insulin had been distributed free of charge throughout the experimental period. In mid-January Lilly entered the planned second stage, selling their insulin at cost. The company had adopted the clinicians’ advice to switch to a smaller unit. Iletin was first sold at wholesale for five cents per new unit (the equivalent of twenty cents per original unit). In 1923 dollars it was an expensive treatment – the forty cents to one dollar a day that the diabetic patient was paying for insulin would be the equivalent of about $6 to $15 in 1980s money.56

  Nor was Iletin as yet available to all diabetics. There was considerable concern, vocal in Toronto but apparently shared in Indianapolis, about the risk involved in insulin treatment. An overdose, of course, could be lethal. More time was needed, it was felt, to educate ordinary physicians about insulin and the need to use it carefully. From the beginning, Clowes’ plan had included publishing the clinical results and educating the ordinary physician. Now the clinical reports on insulin had been delayed and would not be ready for publication until late in the spring. The Insulin Committee also wanted more accessible literature and training programs for general practitioners to be made available; and it would be helpful to have insulin approved by the American Medical Association’s Council on Pharmacy. Lilly, who were not yet certain of the durability of their product, accepted Toronto’s views. The policy for the early months of 1923 was to make insulin available only through selected hospitals and physicians while stepping up the educational work.57

  This cautious North American policy led to the curious anomaly that the first country in which insulin became fully available on a commercial basis was the United Kingdom. There, a joint committee of the Medical Research Council and the Ministry of Health decided in mid-May that since the insulin supply far exceeded the demand, all restrictions on its sale could be removed. British drug firms were “at liberty to supply insulin for distribution in the country and for export through the ordinary commercial channels.”58 The further irony in the British situation was that imported Lilly insulin, sold by the British drug companies at a very substantial profit, at first made up about 80 per cent of the British supply. Because of laxer control, American insulin was being sold more freely in Britain than it was in the United States or Canada.

  British health officials later regretted having been so precipitate, concluding by the end of 1923 that insulin was being “lavishly and wastefully used” by practitioners who needed to be more fully educated on the subject. By then it was too late to return to controls. It seems that the clamour for insulin resulting from its late experimental introduction in Britain had led to a too early deregulation of its distribution.59

  The crowning irony in the British situation developed in the early summer of 1923 when Burroughs Wellcome began selling the first insulin to reach Mexico, Cuba, and other Latin American countries. Insulin was still under clinical trial in Canada and the United States, and Lilly was still awaiting Toronto’s permission to export to anywhere other than Britain. Here was a competing firm first in the field – and not with a competing product, actually, but with Lilly’s own insulin! “To cap the whole situation,” Clowes wrote Toronto’s patent attorney, C.H. Riches, incredulously, “Burroughs Wellcome are actually using our product with which to establish for themselves the priority in entering the world market and all the advantages that accrue therefrom.”60

  V

  Another complicating factor affecting insulin distribution in North America was the vexing matter of rights to its manufacture. Issues involving the patent situation and the future of the Lilly-University of Toronto relationship were confused, delicate, and carried immensely important consequences for the future handling of the magic hormone.

  The patent applications had been filed in the names of Collip and Best in the spring of 1922, about the same time as Toronto gave Eli Lilly and Company its one-year exclusive licence to make insulin. The agreement with Lilly allowed the company to take out American patents on any improvements it made in the manufacturing process; it would assign patent rights for the rest of the world to Toronto. The document also contained a passing reference to trade names that Lilly might use to describe the extract. According to Clowes’ later recollections, at the May 1922 meetings in Toronto Mr. Eli Lilly had made clear how important it was to the company to have a brand name for its product. One of the Torontonians had apparently suggested Banting and Best’s original term, “isletin,” which the Americans then modified, either because of their spelling idiosyncracies or to avoid exact duplication of a possible generic name, to “Iletin.” The Canadians seem to have thought it an insignificant matter at the time.61

  It was no longer insignificant by the end of the summer of 1922, when the Torontonians began to worry that “Iletin” might be used so widely in the literature that it would effectively become the popular and scientific name for insulin, somewhat the way Bayer’s “Aspirin” and Parke Davis’s “Adrenalin” were being used.62 Then they learned that Lilly had filed a patent application in the United States, separate to the Collip and Best application, but one which covered the Toronto method. Clowes explained that the action was simply to help protect the Lilly interest; Lilly’s application would probably be thrown out as conflicting with the Toronto application, but would deter possible competitors. The explanation did not impress the university’s Insulin Committee, or Riches, its patent adviser. The disapproval was transmitted to Indianapolis.63

  H.H. Dale’s visit from Britain that autumn added to Toronto’s alarm. Dale, who had worked for a time at Burroughs Wellcome and had an insider’s alertness for machinations in the pharmaceutical industry, instantly decided that Toronto had blundered badly in its handling of Lilly. “Macleod and Best were already very suspicious of their conduct in certain directions,” Dale wrote Fletcher in late September,

  Unfortunately they missed the point of real importance and I am afraid they have given the whole game into their hands, by not only allowing but almost inviting them to register a new trade name “Iletin” for their version of “Insulin”. The Lilly game, which these people could not understand, seems to me perfectly obvious. If they can make use of their start to get the name “Iletin” used by all the clinicians as the name for the hormone, they will easily upset the patent, get clear of control & snap their fingers at the competition. They are already pursuing a policy which can only have the effect of rendering the Toronto patent doubtful, & this policy had reduced Macleod and Company to a state of complete bewilderment.64

  It is not clear that Dale, in doing all he could to sow distrust of Lilly in Toronto, fully understood the complexity of the situation. Quite apart from a clash of personalities between the Britishers, Clowes and Dale, which clouded the issues – Dale tended to condescend to the voluble upstart, Clowes, while the bemused Americans in situations like this joked about the arguments being “our Englishman against theirs” – there were serious threats of competition on the horizon to worry the Lilly company.

  Some came from quack products that fly-by-night, and not so fly-by-night, companies got out on the market remarkably quickly. In 1922 and 1923 the pharmaceutical division of the big meat-packing firm, Armour and Company, was selling “Insulase,” the Digestive Ferments Company of Detroit was selling “I’Lang-Hans,” the Philadelphia Capsule Company was offering “Insulans,” and the Harrower Laboratories of California were peddling their “Pan-Secretin” in both Britain and North America. Most of these products w
ere in capsule or tablet form and were advertised as easier to take than insulin. Most did contain extracts of pancreas; all, of course, were totally useless. Some of the manufacturers undoubtedly knew this; others took an attitude then common in the drug industry, which Best nicely described in a report on the Digestive Ferments Company:

  It is the policy of these companies to manufacture any biological a few doctors may imagine is giving encouraging results in the treatment of disease. They do not inquire into the merits of the product. They are not in a position to understand if they did inquire. They distribute anything they can sell.65

  More seriously, there was the prospect of legitimate competition and challenges to Toronto’s priority. In Britain, for example, the first articles on insulin had produced immediate claims by two doctors to have discovered effective pancreatic extracts before Toronto.66 Much closer to home, the physiologist John R. Murlin, at Rochester, was certain that as early as 1916 he had produced a pancreatic extract that could burn sugar, and that now, in 1922, he had a practical way of making his own anti-diabetic pancreatic extract. After the unfortunate experience with Havens in July, Murlin had tested his perfusate on other patients, apparently with some success. The Rochester newspapers, perhaps prodded by Murlin, proclaimed that he had conquered diabetes. These claims caused serious trouble for the Havens family, when Lilly and Toronto, believing that the Havens and Dr. Williams had co-operated with Murlin, told them to go ahead and rely on Murlin’s extract. In early September Murlin was involved in discussions with another American drug company, the Wilson Laboratories, about its possible manufacture. Murlin was outraged to learn of Toronto’s patent plans; he was certain that he had priority over the Toronto group, and that E.L. Scott also had priority by virtue of his use of alcohol as an extractive in 1912. Murlin wanted Lilly to agree not to interfere with Wilson making his extract, provisionally named “Glycopyren.” When he learned that this was unlikely, he approached Scott, suggesting that the two of them challenge Toronto’s patent application and perhaps take out a patent in their own right. Scott was lukewarm to the idea, but was surprised at Toronto’s action and felt the patent examiners should be informed of the contributions of earlier workers.67