Eli Lilly wanted to mass produce insulin, but the Toronto team was not yet asking for help. On April 3, 1922 Macleod wrote to Clowes at Eli Lilly. He wrote that he recalled the meeting in New Haven, but that they still were not ready to have Eli Lilly begin manufacturing.
Meanwhile, the transfer and scale up of insulin production at Connaught Laboratories was stalled. No useful insulin was being produced from March to May 1922. Collip could not figure out what the problem was. He was following the same procedure as he did in January, but the insulin coming out of the process was ineffective.
His process is as follows [ from Banting, F. G., Best C. H., Collip, J. B., Campbell, W. B. and Fletcher, \ A. A., Pancreatic extracts in the treatment of diabetes mellitus. The Canadian Medical Association Journal, 12, 1922, 141-146.]:
The best method for separating the active principle from protein, salts and lipoids was found to be the following: Equal volumes of fresh minced pancreas and 95% alcohol were allowed to stand for several hours with occasional shaking, and then filtered.
The immediate addition of 95% alcohol neutralized the proteolytic enzymes, the digestive juices that destroyed the insulin. The immersion in alcohol bypassed Banting’s surgical step of tying off the pancreatic duct and waiting several weeks for the digestive part of the pancreas to atrophy. The pancreas was minced, and bits of pancreas were allowed to dissolve in this 95% alcohol. Tis solution was passed through a filter. The resulting liquid, the filtrate, should contain the insulin, but may also contain other molecules. The process was repeated.
To the filtrate was added 2 volumes of 95% alcohol. After allowing several hours for precipitation of most of the protein, filtration was repeated, and this filtrate was evaporated to small volume in vacuo at a temperature of 10 to 30 degrees C. An aqueous extract was thus obtained, which was washed twice with ether in a separatory funnel to remove lipoids, and then further distilled in vacuo to a pasty consistency. Addition of 80% alcohol and centrifugation then resulted in the formation of four layers, namely, a bottom layer of salt crystals, above it a saturated aqueous solution of salt, above this a flocculent layer of protein, and on top a clear layer of alcohol containing the entire active principle in solution.
This top layer was pipetted off, and delivered into several volumes of 95% or (preferably) absolute alcohol. Several hours were then allowed for the active principle to precipitate out. It was then dissolved in distilled water, concentrated by vacuum distillation, filtered through porcelain, tested for sterility, and delivered to the clinic. Such extracts, practically free from proteins, salts and alcohol-soluble substances, could be made isotonic and injected sub-cutaneously without local reactions.