For many years the world relied on insulin derived from the minced pancreata of slaughtered pigs and cows. At first, allergic injection site reactions, varying from itchy lumps to anaphylactic shock, were quite common. Once it became clear that these side-effects were mainly due to contaminants, purification steps such as crystallisation were introduced which yielded a product that contained very few impurities—a few hundred parts per million.
Insulin production changed little over the next years. However, some problems persisted, the major one being lipoatrophy. This was a tendency for patients to develop disfiguring hollows at their injection sites; the fat just melted away until skin rested directly upon muscle.
Most manufacturers used beef pancreas, because of the ample supply, but cow insulin differs from human insulin in three of the 51 amino acids in the peptide chain (see figure), and the immune system can spot the difference. Allergic responses apart,
Amino-acid compositions of cow and pork insulinmany users developed antibodies to insulin, and it was natural to wonder if these interfered with its therapeutic benefits. Pig insulin differs from human insulin by only one amino acid, and people also wondered whether this was less immunogenic than beef insulin. And which was more important: the remaining impurities in the preparation, or the insulin itself?
The Swedish physician Erik Jorpes made the chance discovery that insulin allergy could be treated with insulin that had been through the purification process not once, but thrice, suggesting that the problem lay with the impurities. Jorgen Schlichtkrull, a skilled Danish insulin chemist who pioneered the zinc insulins took note of this and set out to produce a cleaner insulin.
Impurities can be detected by passing the extracted insulin through a chromatograph, which can distinguish between even closely related molecules by virtue of the rate at which they travel through a filtration system. The hallmark of a totally pure insulin is a single well-defined peak on the read-out from the machine; impurities are revealed by a cluster of untidy foothills. “Single peak” or monocomponent insulin would become the watchword of the manufacturers, and by the 1970s the Danes had a cleaner product than anyone else. And with these clean insulins, the problem of allergies and lipoatrophy almost disappeared.
Human insulin replaces animal insulin
The insulin molecule, and the receptor it interacts with, have remained relatively constant throughout vertebrate evolution, which is why the insulin produced by any animal with a backbone will lower glucose in any of the others. Since human insulin differs by only one amino acid from that produced by pigs, it was a moot question as to whether it would be any better as a treatment for diabetes—but it was inevitable that someone would try. In 1974 the company Ciba Geigy succeeded in chemically producing human insulin by total chemical synthesis, but the process was expensive and yielded only little insulin. However, in 1976, the insulin company Eli Lilly realised that bioengineering and biosynthesis were almost possible, and that the production of human insulin seemed the obvious choice when considering how to apply these new techniques. Several factors may have driven their pursuit of this goal. Firstly, there was a (probably misjudged) feeling that the demand of insulin would outdo the animal supply. Secondly, there was the unsupported feeling that the human insulin might be even less immunogenic than the purified pig insulin. And finally, of course, there were commercial motives. In 1981 Genentech and Eli Lilly managed to produce the first human insulin by inserting the human insulin gene into E. coli bacteria, soon followed by the Novo company who did the same using yeasts to produce insulin. Despite the fact that the advantages of the human insulins were rather underwhelming, the more costly human insulins soon replaced the animal insulins across Europe and North-America.