The first improvement in extract preparation involved a final filtration through an unglazed porcelain, or Berkefeld, filter, which trapped clumps of bacteria, assuring sterility although apparently reducing potency. Heating or boiling the extract, Banting and Best found, seemed to destroy the active principle. With the Berkefelded extract it seemed useful to return to trying subcutaneous rather than intravenous injections; they would spread the extract’s action over a longer period of time and be less likely to cause shock. On November 23 Banting experimented with a new blood sugar test, the Shaffer-Hartman method, probably introduced to them by Collip, who had learned it in the United States that summer. More significant is a brief note in Banting’s handwriting the same day: “One of us (FGB) had 1½ cc Berk. ext. subcut. No reaction.” This is the only record of the pair’s first tentative…what the hell…why wait?…only a little bit…use of their extract on a human being. “No reaction” meant no harmful effects; they did not try taking their own blood sugars.
Banting and Best were starting to pay attention to the quantities of pancreas and Ringer’s solution being used,25 and were experimenting with more concentrated doses of the extract. The addition of tricresol, used to preserve diphtheria anti-toxin, did not seem to interfere with the active principle.26 They were beginning to have trouble getting blood from dog 33’s veins,27 but tests continued through to the end of November. The notebooks record fairly low blood sugar readings on the 27th (.048 at 2:30 p.m., followed by an injection of extract half an hour later; .05 at 7:30 p.m.) without any observations on the dog’s condition. Administration of extract through a stomach tube – to see if it might be absorbed through the gastric mucosa – produced no effect. As earlier experiments had shown, only injections seemed to work.
In the meantime the longevity experiment continued on dog 27. Although few observations of its condition were recorded, the experiment seemed to be going smoothly.28
VI
Banting and Best finished their first paper in late November, reporting the results of their work through November 10. Macleod apparently advised the pair on its format and helped polish the final draft. When the manuscript was finished, Macleod recalled ten months later, “Banting asked me if I wished my name to appear along with his and Best’s. My reply was that I thanked them but could not do so since it was their work and ‘I did not wish to fly under borrowed colours’.”29 Under the bold title, “The Internal Secretion of the Pancreas” (by F.G. Banting, M.B., and C.H. Best, B.A.), the paper was sent to the prestigious Journal of Laboratory and Clinical Medicine, published in St. Louis. It was accepted for the February 1922 issue.
The paper was a reasonably straightforward description of the work, set in the context of the background reading the pair had done. Much of the travail and disappointment of the early summer’s experiments was omitted, as were one or two particularly badly done experiments. The article does not omit the misreading of Paulesco’s work and the misinterpretation of the first experiment with whole pancreas. Like all of Banting and Best’s joint papers, the article contains minor factual errors. Figures given in the text and the charts sometimes disagree with each other and/or with figures in the notebooks. The description of the last experiment is particularly bad: the charts show one set of figures for volumes of extract injected and duration of duct ligation; the text contains a second set; the original notebooks contain a third set.
The key sentences of the paper are a sweeping summary and claim:
In the course of our experiments we have administered over seventy-five doses of extract from degenerated pancreatic tissue to ten different diabetic animals. Since the extract has always produced a reduction of the percentage sugar of the blood and of the sugar excreted in the urine, we feel justified in stating that this extract contains the internal secretion of the pancreas.
That summary is inaccurate, representing enthusiasts’ tendency to put a totally favourable gloss on their results. It was simply not true that Banting and Best’s extracts had always produced a reduction of the percentage sugar of the blood and of the sugar excreted in the urine. Sometimes the extracts had not worked at all; other times their effects had been inconclusive; a few times the necessary tests had not been done. It was not a long series of specific results, all of which were clearly successful, that was impressive. Such a series did not exist. Rather, it was the overall pattern formed by the experiments. By my rough (because partly subjective) estimate, compiled from the notebooks and charts, Banting and Best’s first 75 injections of extract of supposedly degenerated or “exhausted” pancreas, using nine dogs, produced 42 favourable results, 22 unfavourable ones, and 11 inconclusive observations.30 This is an impressive statistical picture in its own right, impressive enough to justify the work on the one hand and explain the researchers’ overly enthusiastic claims on the other. In the face of so many good results, the tendency was to forget or ignore the bad ones.
Whether the pattern of results justified the claim to have captured the internal secretion of the pancreas is another question. There is no doubt that Banting and Best thought it did, but as with all claims of discovery the difficult job was to convince other people. The effectiveness of Banting and Best’s results in inducing conviction would be clear only after publication or other public presentation. And, as we will see, it would be well after publication before anyone asked whether the results verified the researchers’ subsidiary hypothesis about the external secretion being destructive of the internal. Finally, despite their claim that their extract contained the internal secretion, Banting and Best specifically said they did not yet have a therapeutic agent. The clinical condition of their dogs had “always” distinctly improved after administration of the extract, they reported, “but it is very obvious that the results of our experimental work, as reported in this paper do not at present justify the therapeutic administration of degenerated gland extracts to cases of diabetes mellitus in the clinic.”
Before actual publication, there would be an occasion for public presentation of the work at learned society meetings held during the Christmas holidays. Macleod was a member of the American Physiological Society. When he received the call for papers for its annual meeting, to be held in New Haven, Connecticut, he suggested to Banting that a report on the work be presented. According to Macleod, Banting asked that Macleod’s name should appear on the report to draw attention to it. “I agreed to this arrangement but stipulated that Banting and Best should both attend the meeting and should themselves present the report.”31
Macleod’s attitude to Banting and Best’s work at this time (towards the end of November) is a bit obscure. While giving them advice and help he was not actively directing them, and had not yet added to their resources. There was no secrecy about the work, however, and at least one inquiry came from outside. At a meeting of the Southern Medical Association in Arkansas, Elliott Joslin talked with Dr. Lewellys Barker of Johns Hopkins Hospital, a Canadian and a University of Toronto graduate who had recently been back for a visit. Perhaps Barker had attended the Journal Club meeting; perhaps he had been talking with Macleod. He had learned what was going on at Toronto and mentioned it to Joslin. Joslin wrote Macleod asking whether he had published or was about to publish anything on the work. “Naturally if there is a grain of hopefulness in these experiments,” Joslin added, “which I can give to patients or even can say to them that you are working upon the subject, it would afford much comfort, not only to them, but to me as well, because I see so many pathetic cases.”
“It is true that we have been doing work on the influence of Pancreatic extracts, which has yielded most encouraging results,” Macleod answered on November 21.
But I would rather hesitate to attempt the application of these results in the treatment of human diabetes until we are absolutely certain of them. Dr. Banting and Mr. Best who have been doing this work, are to report their findings at the meeting of the Physiological Society at New Haven, by which time we expect to be in a position to come to
a definite conclusion. I may say privately that I believe we have something that may be of real value in the treatment of Diabetes and that we are hurrying along the experiments as quickly as possible.32
Another person Barker told about the Toronto work was Dr. George H.A. Clowes, the research director of Eli Lilly and Company, a pharmaceutical manufacturing company located in Indianapolis, Indiana. Clowes took no action at the time, but resolved to be at the New Haven meeting to hear the presentation.33
VII
Something suddenly went very wrong with the longevity experiment on the afternoon of December 2. About four hours after an injection, dog 27 began showing symptoms of “a peculiar nature” (“Convulsive twitchings – retraction of head – unconscious for several hours. Salivation & frothing – seemed to improve for a time, then repetition of symptoms”).34 It seemed improved the next day, but ninety minutes after injection the symptoms began again and were more severe. This time the dog did not recover. It died that night, bringing the longevity experiment to an abrupt end. Banting and Best recorded the death as due to an anaphylactic-like reaction.*
Instead of prolonging dog 27’s life, the extract had killed it. The idea of running a longevity experiment still seemed like a good one, though, and on December 6 the pair decided to convert the test dog, 33, later known as Marjorie,35 to longevity. She had been depancreatized on November 18. Another experimental dog, 23, was started, pancreatectomy by Best.
VIII
The first step towards another important breakthrough also came on December 6 when Banting and Best decided to try using alcohol in the preparation of their foetal calf extract. Macleod had suggested the idea months earlier, and it had been used by both Scott and Zuelzer. In 1922 Best wrote that trying alcohol was a “fairly obvious” idea, which occurred to Macleod, Banting, and himself independently.36 The trouble with an aqueous saline solution of extract, they seem to have thought, was that any attempt to concentrate it to get at the pure active principle by boiling away the water also seemed to destroy the active principle. Alcohol evaporates at much lower temperatures than water. It may also have been used in the hope that it would dissolve and remove some of the contaminating impurities in the solutions. Banting and Best ground the foetal calf pancreas up in alcohol, filtered the mixture to get out the solids, and then evaporated the alcohol by a technique Macleod had shown them of using a current of warm air flowing over porcelain dishes containing the solution. The dry residue was next redissolved in a saline solution. It was given to dog 23 on Wednesday, December 7, and worked well. Injections given to dog 33 the next day were less potent, but moderately successful.37
The realization that the active principle in a foetal pancreas was soluble in aqueous alcohol led Banting and Best to wonder whether they could get a similar result from a fresh adult pancreas. On December 8 they did a pancreatectomy on dog 35. Instead of throwing out its pancreas, they cut it up into slightly acid alcohol, macerated it, and allowed it to stand for forty-eight hours. The solution was filtered, the alcohol was evaporated off in the warm air current, and the dry residue was redissolved in saline. On Sunday morning, December 11, dog 35 was given six cc. of extract of its own whole pancreas. Its blood sugar dropped from .38 to .18 in four hours. Whole pancreas extracted with alcohol worked. Here was another major advance. No more degenerated pancreas. No more foetal pancreas. Now the research could go forward using cheap, easily obtainable supplies of fresh whole pancreas.
IX
Banting had continued to press Macleod for help, several times asking if J.B. Collip could join the team to work on the biochemistry of pancreatic extracts. Collip wanted to help. Working several blocks away in the pathology building on the grounds of Toronto General Hospital, he saw Banting and Best every few days, took a great interest in their experiments, and often left with the comment, “Well, if I can be of any assistance let me know.”38 Many years later, Dr. E.E. Shouldice, who had been working in the same lab with Collip that fall, remembered the biochemist’s anxiety to get to work on Banting and Best’s extract. Collip said, Shouldice remembered, that it would take him about two weeks to purify their crude extracts.39
James Bertram Collip was well qualified to take part in the work. At the age of twenty-nine in 1921 he was a year younger than Banting, but had far, far more experience at medical research. Born in Belleville, Ontario, a florist’s son of British descent, Collip had taken his B.A. in Honour Physiology and Biochemistry at Toronto in 1912. He had gone straight on in biochemistry at Toronto, earning his M. A. in 1913, his Ph.D. in 1916. In September 1915 he had begun work as a lecturer in physiology at the fledgling University of Alberta in Edmonton. By 1920 he had been promoted to full professor in charge of a new Department of Biochemistry.
Collip was an active researcher, working on a variety of problems most of which involved blood chemistry. By the end of 1921 he had a respectable list of twenty-three academic publications, including a very good 1916 summary article on internal secretions, a subject in which he had a longstanding interest. In 1920 he spent some time making and injecting tissue extract, accompanied by adrenalin, and studying the resulting effects on blood pressure. Although he had had to shoulder a heavy burden of teaching at Alberta, Collip’s deepest commitment was as a research scientist. Happily married and with a young family, he was a shy, sensitive, and boyish-looking young scientist, who loved nothing more than long stints in the lab, preferably late into the night, trying out this or that mixture of ingredients to produce a desired, or perhaps an undesired, physiological reaction. He was being supported on his sabbatical leave by a Rockefeller Foundation Travelling Fellowship and a temporary appointment as assistant professor at Toronto.40
Macleod finally agreed to Banting’s requests and invited Collip to help Banting and Best in the development of their extract. They needed help not, as legend has it, because they were floundering around going nowhere, but because the pace was speeding up thanks to the development of means to produce large amounts of extract. There was so much to be done. Banting wanted it done quickly. If there was any objection to Collip joining the work, it came from Best, who apparently felt slighted-a student of biochemistry seeing a full professor and Ph.D. come onto the team. “I was opposed to Collip’s participation in our work for obvious and selfish reasons,” Best said in an unpublished address in 1957, “but Fred Banting persuaded me not to protest too vigorously. This was also for obvious reasons – i.e. the urgent need of our antidiabetic material for clinical use.”41
The exact date in December on which Collip started work is not known. It would be helpful to know it because of the terrific pace of discovery in December, as well as the later dispute about who did what when. Collip appears to have been at work on pancreatic extracts by Monday, December 12. It is important to note that Banting and Best appear to have begun making alcoholic extracts of fresh whole pancreas – at least the one canine pancreas tried on December 11 – before Collip joined the group.
His first contribution to the research may have been to add a rare and totally unintentional comic note. The very first extracts he made, using alcohol and fresh whole glands obtained from the abattoir, apparently did not work. Collip went to see Banting and Best:
“There is something wrong with this whole piece of work,” Banting in 1940 remembered Collip as saying.
“What makes you think that?” asked Best.
“Well, I made some extract and did not get the results which you got.”
After much discussion of Collip’s methods, it turned out that he had told his lab boy to go to the abattoir and ask for sweetbreads. Instead of getting pancreas, which at that abattoir was being ground up for fertilizer, the boy was bringing back thymus or thyroid glands.
It was a good anecdote, one Banting liked telling to denigrate Collip’s work. Perhaps it is a true story, for an entry in Banting and Best’s notebooks later in December shows them using, as a control, an extract of “Protein free Thymus à la Collip ± Pan.”
X
Collip’s notebooks have not been found, so his work has to be pieced together from several later accounts, all of which are less detailed than Banting and Best’s records.42 There is no doubt, however, that, coming into the work just at the stage when Banting and Best had found that alcoholic extracts of whole pancreas lowered the blood sugar of diabetic dogs, Collip quickly began making important findings.
He immediately began making extract from whole beef pancreas. Then he followed up a suggestion Macleod made in the presence of the other three researchers. Why not try the extract on the blood sugar of rabbits, Macleod wondered, particularly those made “diabetic” experimentally. There were several ways of inducing hyperglycemia in rabbits, all of which would be cheaper and easier than depancreatizing dogs. Possibly because he was working a long walk away from the dog quarters in the medical building, Collip used rabbits from the beginning. As soon as he got his sweetbread confusion straightened out, Collip found that pancreatic extracts were effective on rabbits. And not necessarily diabetic rabbits, but just plain rabbits, perfectly normal ones. Extract lowered their blood sugar from normal to below normal just as Banting and Best’s extracts lowered the blood sugars of diabetic dogs from above normal to normal and below. This observation had immense practical importance, which Collip realized immediately, in giving the group a quick, easy way of testing the potency of a batch of extract. Its strength could be measured by its effect on the blood sugar of a normal rabbit, procured quickly and easily from a vein in its ear, tested by the new Shaffer-Hartman method.43
Collip also began work on depancreatized dogs. On Friday, December 9, Banting and Best had run a strange experiment on dog 23, recording the effect of an injection of extract on its blood pressure and blood sugar while it was under anesthetic. It is one of the least satisfactory of Banting and Best’s experiments: the dog was apparently dying of infection and figures given in the published account bear little relation to those in the notebooks.44 There was nothing significant in the blood pressure performance, but there was a surprising failure of the dog’s blood sugar to respond significantly to the injection of apparently potent extract. Why did the extract not seem to work on a dog under anesthesia?
The Discovery of Insulin Page 13