The Discovery of Insulin

by Michael Bliss

  II

  There was gradual progress. As dog 387 recovered from its first stage, Best could begin making his tests. They were fairly normal, up-to-date chemical procedures – the Myers-Bailey modification of the Lewis-Benedict method of blood sugar estimation, Benedict’s qualitative and quantitative tests for sugar in the urine – but still a long way from the quick, simple, and more accurate tests available today. Best’s work involved the use of several reagents, careful preparation and measurement, and various chemical procedures involving centrifuging, filtration, evaporation, and precipitation. The blood testing was the most advanced procedure: as we have seen, it was a relatively new way of measuring the diabetic condition. But Macleod also expected Banting and Best to use more traditional tests of the urine. These involved not just finding the amount of sugar in the urine, but also determining the ratio of glucose to nitrogen. This G:N ratio, also often called D:N (dextrose to nitrogen), was thought to be a particularly accurate reflection of the diabetic state.

  Dog 387, the first to be tested, showed normal reactions. Although it had lost most of its pancreas, the presence of the pancreatic remnant was sufficient to keep its blood sugar within a normal range of .085 to. 150 per cent. There was no sugar in its urine. Then on May 28 Banting removed the pedicle, making the pancreatectomy complete. The next day the dog’s blood sugar had risen to .35 per cent and there was sugar in its urine. The next day the blood sugar was .42 and the D:N ratio was 2.7:1. These figures were both indications of a diabetic state, although Banting and Best may have been a bit disconcerted that the D:N ratio was not higher; they seem to have believed, incorrectly, that it ought to be 3.65:1 in a completely diabetic dog.4

  Dog 387 died on June 1, partly from its diabetes but probably mainly from the infection in its abdominal wall that Banting found in his autopsy. Still, this was the first experiment to have gone more or less according to plan. Banting already had another dog partially depancreatized, and soon added one more. He ligated the ducts of several other dogs. By the end of the fourth week of experiments, on Sunday, June 12, things must have seemed to be going fairly well. As soon as the two partly depancreatized dogs healed they would be ready for the second stage and more observations of the diabetic condition. Meanwhile there were now six duct-ligated dogs, and a seventh done on June 13, whose pancreases were presumably atrophying according to theory. It would be just a matter of time until they could be opened up and the critical experiments begun.

  But what exactly did Banting plan to do? All later accounts, including his own, state that he and Best planned to take out the degenerated pancreases from the ligated dogs, prepare an extract from these, and then administer the extract to other diabetic dogs. As has been mentioned, however, none of the contemporary documents refers specifically to an extract. The clearest, in fact the only, statements of the plan of work are to be found in Banting’s original notebook for June 9 and 14. On both these days he seems to have talked over the work with Macleod. The full notebook entries are as follows:

  June 9

  suggestion.

  - Have depancreatised dog pedicle.

  - graft into it remnant of degenerated pancreas

  - later remove pedicle

  - then remove graft

  - Prof. McLeod -

  - 1 put remnant free in peritoneum;

  2 put remnant subcutaneously.

  3 emulsions.

  whole remnant in one shot

  4 repeated smaller shots.

  5 50 gms glucose to totally diabetic dog

  50 gms glucose with whole gland remnant.

  June 14. Dr. McLeod’s parting instructions

  have dogs diabetic DN ratio constant for 3 days, meat diet.

  (1) intra peritoneal graft

  (2) subcutaneous graft

  (3) whole remnant intravenous injection.

  (4) Divided aq. 2 h intravenous

  (5) subcutaneous injection.

  microscopic sections of remnant before and after transplant.

  These notes are open to slightly differing interpretations, and it is not clear how the interchange of ideas between Banting and Macleod had developed. In view of the state of research at the time and the widespread interest in grafting as a technique for working with the pancreas (experimental surgery was then in a blush of enthusiasm about the possibilities of grafting) it seems that Banting and Macleod had agreed that the first approach would be to graft pieces of degenerated pancreas into diabetic dogs. The second or perhaps follow-up approach would be to make an emulsion of degenerated pancreas and inject it, in various doses, into the diabetic dogs.5

  The widely held belief that Macleod set Banting and Best to work and then immediately left town for his holidays is not true. The young men had been at the research for almost a month and seemed to have worked through their early technical problems before Macleod left. He talked over the situation with Banting in June and gave fairly explicit “parting instructions” before leaving. Banting wrote down Macleod’s summer address in Scotland. He also noted the summer address in Massachusetts of J.B. Collip, the biochemist who had been present at one of the discussions and knew roughly what Banting intended to do. Banting presumably thought Collip might be someone he could contact for help or advice on short notice.

  III

  By the third week in June Banting was working completely on his own, for Best, too, was out of the city, having gone off to the Niagara region for ten days of militia training. Charles Herbert Best, “Charley” to his friends, had the world at his feet in the summer of 1921. He had just graduated with his Bachelor of Arts degree from Toronto’s Honour Physiology and Biochemistry course, had an interesting summer job with Fred Banting, and knew he would be working again the next year for Professor Macleod while taking an M.A. He was twenty-two years old, a strikingly handsome, tall, blond-haired, blue-eyed, athletic young man. He had grown up on the coast of Maine, where his father, a Canadian with deep family roots in Nova Scotia, was a small-town general practitioner. Charley had come to Toronto in 1916 to finish his high school education and then entered university. He enlisted in the Canadian army after finally finding a homeopathic doctor who certified him as medically fit, not noticing his heart murmur, and had reached England as a sergeant in the artillery when the war ended. He somehow got home quickly enough in 1918–19 to salvage his second year in the “P&B” course at Toronto.

  Those last years at the university were a wonderful idyll of fraternity parties and dances, picnics, tennis, riding, golf, semi-pro baseball, and, of course, attention to studies. Best’s marks were not outstanding, but they got better the further he went in his course, until in the final year his second-class honours standing was worth a share of the silver medal in his class with Clark Noble. Charley Best was outgoing, sociable, and popular. In choosing him as a demonstrator, and often inviting him to his home, Professor Macleod had obviously picked Best as one of his most promising students. The young man apparently thought he might eventually train as a surgeon and dreamed romantically of going off to practise in South America.6 He would be accompanied in his adventures by the strikingly beautiful Margaret Mahon, a Presbyterian minister’s daughter whose family had recently moved from the Maritimes to Toronto. Charley met Margaret at a sorority party in 1919. They became engaged in 1920, a beautiful couple whose pictures from those years call back all the best nostalgia of the Twenties, like Scott and Zelda Fitzgerald without the crack-up. In fact Charley and Margaret were a more handsome couple than Scott and Zelda.

  Best would have enjoyed working with Banting, even though as yet the research had not produced any results. The older man’s unsettled life, country ways, and personal insecurity all contrasted sharply with Charley’s happy situation; but they were both products of small towns, both veterans, both interested in medicine, sports, and Saturday night dates. Best was learning some surgery from the partnership (Banting some chemistry), and he may have had a special interest in research relating to diabetes. In 1917 one of his au
nts, first a nurse and then a patient with Joslin in Boston, had died from the disease.

  IV

  While Best was enjoying riding with the militia, Banting found work in the lab increasingly frustrating. On Wednesday, June 15, just before Best left, Banting completed his second successful two-stage pancreatectomy, on the second dog to be numbered 386. It rapidly became diabetic. “You ought to get the 3.65:1 D:N on Sat or Sunday,” Best wrote to him in their joint notebook as he left for Niagara. Banting did not get the expected ratio. The dog’s D:N stayed constant at about 2.5:1 over the weekend and then started dropping in the next week. Later in that week there was more frustration as Banting, apparently operating without assistance, lost another dog from bleeding during the first stage of a pancreatectomy. He apparently went out on the streets and bought a dog to replace it.

  The notebooks also show changes and corrections in the figures for urinary sugar, as Banting started using different chemical solutions from those Best had left behind to use in testing. In his 1940 account, nowhere else, Banting wrote that he found Best’s solutions to have been anything but standard and the glassware filthy. Banting decided to have his nitrogen estimates done by a friend in the biochemistry building. He must have been even more frustrated when the new, revised figures showed dog 386 with an even lower D:N ratio, well under 1.0. Whatever was happening, Banting apparently decided that Best was at fault.7

  Best came back to Toronto on Saturday the 25th or Sunday the 26th. He went to see Margaret and then dropped into the lab about eleven at night. “I was waiting for him,” Banting recalled in 1940, “and on sight gave him a severe talking to….

  I told him that if he was going to work with me that he would have to show some interest, that his work was totally unsatisfactory, that he lacked accuracy and was too sloppy, and I ended up by telling him that before doing another single thing he must throw down the sink every solution that he had been using, wash every bit of glassware and make up new solutions that were truly “normal”.

  When I finished thus setting him out he gazed fiercely at me for some moments in silence – he looked very defiant – his fist opened and closed – he was very vexed. I thought that he was going to fight and I measured the height of his jaw. He delayed and I feared that he was not going to fight. Suddenly he turned on his heel, went upstairs and I heard rough usage of glassware for some time. He worked all night (for the first time in his life but it was the beginning of many). In the morning he left and I inspected. Everything was spick and span. We understood each other much better after this encounter.8

  Best apparently never referred to this incident.9 Banting did in verbal accounts, embellishing it into a kind of fight and saying he had thrown Best’s solutions down the bathtub. He tended to embellish, and no one remembered a bathtub in the medical building. Still, the incident probably took place, perhaps as Banting describes it. Crossed-out figures in the notebooks are the circumstantial evidence of Banting’s frustration. At all times he was an intensely direct man, quick to anger, ready to fight anyone on any matter of principle. This was not the last time an angry Banting, ready to fight, would confront his research associates in the next eighteen months. By the same token, Fred Banting could shake hands after the air had cleared or the dust settled, and go on to work with a person more closely than ever. His having written “we understood each other much better after this encounter,” suggests that it was an important step in the cementing of Banting and Best’s partnership.

  Best not only cleaned his glassware and re-did his solutions, he had decided to stay on to take Noble’s place in July. By all accounts, including his own, Noble fully concurred in the decision. Everyone apparently realized there was no point in a new assistant having to learn all the procedures over again at this stage of the work. Banting thought the students’ interests in being near their girl friends (Noble’s was out of town) also had something to do with the decision.10

  V

  Best’s return and their better understanding of each other did not make the work go better. Creating experimental diabetes in a totally depancreatized dog seemed to them almost impossible. Dogs that healed well enough from the first stage to have their pedicle removed would still not become diabetic; others seemed to become diabetic after the first stage and did not heal. All healing was difficult in the extreme heat and humidity of a Toronto July. Operating in the dingy little room next to the animal quarters, right under the gravel-and-tar roof of the old medical building, was a sweltering ordeal. The animal quarters, by all accounts, were fairly dingy themselves. Banting and Best found they had to help the attendant do the work of caring for their dogs and keeping their cages clean. They worked in heat and dirt and unbelievable stench.

  Down two flights of stairs in the physiology department they had been given bench space in room 221, a kind of anteroom between a corridor and one of the main laboratories. Part of it was used as a storeroom for supplies and old apparatus. They saw a lot that summer of Dr. Fidlar, the inhabitant of the main lab, who was using the most complicated apparatus they had ever seen to do intricate experiments on the respiration of one frog.11

  By early July it had been about five weeks since the pancreatic ducts had been ligated in seven of the dogs. There were two partially depancreatized dogs on hand. Perhaps thinking they were ready for the crucial experiments (or perhaps just to check on the degree of pancreatic degeneration), Banting opened up one of the ligated dogs on July 5. Its pancreas, which should have been markedly degenerated, was apparently quite normal. Something had gone drastically wrong. The duct ligation hadn’t worked at all.

  That day and the next, in temperatures over 97 degrees fahrenheit, they opened up all seven of the ligated dogs. Five of the seven had normal pancreases. Only two of the dogs showed the expected degeneration. As Banting had feared when he first did some of the dogs, he had failed to isolate and close off all of their pancreatic ducts. The fact that he had used catgut for his ligatures – a stiff material which can easily loosen – had added to his problem. The five dogs were done over again, this time with silk ligatures and sometimes two or three applied at varying tensions.12 Two of the five dogs died within a day of the operation. So did the two partially depancreatized dogs, suffering from infection in the heat. The week was a disaster.

  In fact the whole research program was not far from total failure. Banting and Best had experimented on nineteen dogs. Fourteen had died, no more than two of them according to the research plan. There were five duct-tied dogs left, and only two of them had gone according to plan. Banting and Best took a long weekend off and started over again on Monday.

  VI

  The pace was a little slower in the second week of abnormally oppressive heat. On Monday, July 11, they did a first-stage pancreatectomy on dog 410, a white short-haired mongrel terrier. While waiting for it to heal they took most of the week off. Charley wrote Margaret on the 14th that he was getting tired of the combination of heat and lab work.13 The next Monday the terrier’s pedicle was removed, completing the operation. The dog did not become very diabetic, its blood sugar ranging from .15 to .22 and its D.N ratio well under 2.0:1. Another dog, 406, a long-haired yellow collie, was started on the 22nd. Tests later in the week seemed to show that dog 410 was becoming more diabetic.

  On Saturday, July 30, they decided that the time had come to put the work to the test. The pancreases in the two original duct-tied dogs had been degenerating for seven weeks, surely enough time to destroy all the cells producing the external secretion. That morning Banting removed the pancreas, apparently atrophied, from dog 391 (whose ducts had been ligated on June 7). Whatever plans he had made for transplants or grafts were abandoned in favour of the much easier and quicker method of making an extract of 391 ‘s shrivelled pancreas.

  “We followed your directions in preparing the extract,” Best wrote to Macleod in one of their reports ten days later. They sliced the pancreas up, putting the pieces into a chilled laboratory mortar containing ice-cold “art
ificial blood” or Ringer’s solution, a mixture of salts in water commonly used to preserve tissues. The mortar was put in a freezing brine solution until its contents partly froze. The half-frozen pancreas pieces were then macerated (ground up with sand and a pestle). The solution was filtered, apparently through cheesecloth and blotting paper, to eliminate the solid particles. The filtrate, a pinkish-coloured liquid extract of degenerated pancreas, was warmed to body temperature and was ready for injection.14

  Reflecting his experience with older techniques requiring larger samples of blood, Macleod had apparently advised Banting that it would be necessary to anesthetize a dog and insert a special cannula or tube each time a sample of blood was to be taken. As Kleiner had before them, Banting and Best discovered that this was not necessary. With a fine needle they were able to get repeated blood samples from veins. After the first few punctures most dogs would lie quietly. The same technique was used for the administration of extract.15

  At 10:15 in the morning of July 30, 1921, Banting and Best injected four cc. of their extract into a vein of the white terrier, dog 410. Its blood sugar at the time of injection was .20. An hour later, at 11:15, the blood sugar had fallen by 40 per cent to .12. At that time five more cc. were injected. In the next hour the blood sugar barely moved, to .11. Despite another injection, by 2:15 it was starting to rise again, to .14.

  To see how the extract influenced the dog’s ability to metabolize sugar, Banting and Best decided to give it sugar (20 grams in 200 cc. of water) through a stomach tube. They were not practised in the technique: “Tube first passed into lung dog nearly drowned [sic] – completely recovered in 15 min.” the notebook records. This was at approximately 2:15 in the afternoon. Three five-cc. injections were made at hourly intervals. As the copy of Banting and Best’s chart shows, the terrier’s blood sugar rose and the extract did not bring it down. The increase was not as marked as the administration of sugar had caused a few days earlier, though, and very little of the sugar, less than one gram, appeared in the dog’s urine or vomit during a five-hour period. After taking a final blood sugar reading at 6:15 on that Saturday night, Banting and Best left the lab.