You are on page 1of 5


Typical Hormone

With a speed no longer seen in drug discovery and development, insulin was
isolated for the first time in 1921 from animal sources and commercialized within 12 months.
Decades later, it took just four years for developers to move from expressing recombinant
insulin in bacteria to launching the world's first biotechnology drug product.
Scientists Frederick G. Banting and Charles H. Best, working in a lab provided by John J. R.
MacLeod at the University of Toronto, isolated the polypeptide hormone and began testing it in
dogs. By 1922, with the help of James B. Collip and pharmaceutical company partners, the
researchers could purify and produce animal-based insulin in larger quantities.


SHAPELY Insulin molecules form hexamers whose

crystals appear hexagonal.

Insulin is produced by beta cells in the pancreas and is the most important hormone in the body
to regulate blood glucose levels. A partial or complete lack of insulin causes diabetes, which,
untreated, is often fatal by the teenage years. The World Health Organization reports that an
estimated 177 million people worldwide have diabetes. Although not a cure, insulin injections
have been the standard treatment since 1924.
Before insulin was discovered, diabetes was managed through diet, which allowed patients to
survive, but generally for just a few years after diagnosis. Remarkable medical results were
achieved with the first insulin injections. Doctors finally had a means to offer patients a nearly
normal quality of life, and it quickly became necessary to increase insulin production.
The Toronto scientists had trouble, however, with consistently isolating and purifying the drug.
Connaught Laboratories in Canada, now part of Sanofi-Aventis, assisted, and Eli Lilly & Co.
proposed developing large-scale production methods. The university initially rebuffed offers
from Lilly, but an agreement was reached in May 1922. By that summer, Lilly was supplying the
quantities needed for clinical trials.
To expand the supply, the university gave many royalty-free licenses. Among the licensees was
Danish Nobel Laureate August Krogh, whose wife was diabetic. In December 1922, production
started in Copenhagen, and the first patient was treated there in March 1923. Krogh later
founded the Nordic Insulin Laboratory, now part of the drug firm Novo Nordisk, with Danish
physician Hans C. Hagedorn.

Insulin was first commercialized in Great Britain in May 1922, according to the University of
Toronto, and by October 1923, it was being sold in the U.S. and Canada. By 1924, large U.S.
and British companies were marketing insulin worldwide. It became a major product for Lilly
and remains the foundation of Novo Nordisk's business.
According to accounts from the time, the scientists' fame rose rapidly. MacLeod and Banting
were awarded the 1923 Nobel Prize for Medicine or Physiology, which they shared
independently and respectively with Collip and Best. The Nobel Prize was a first for Canadian
scientists, but the relationship between Banting and MacLeod was strained over differences
about their contributions to the discovery.
Since then, the science, production, and delivery of insulin have been widely studied. It was the
first protein for which the chemical structure and molecular weight were determined. The
structure varies among species--human insulin differs by three amino acids from the bovine
form and just one from the porcine form. The resultant similarity in conformation and activity is
what, fortuitously, makes animal insulins so efficacious in humans.
FOR 60 YEARS, cattle and pigs were the sources of insulin. Although these products were
highly effective, concerns arose about the growing diabetic population, long-term supply, and
potential allergic reactions. Scientists had succeeded in synthesizing insulin in the 1960s, but
synthesis was not seen as a viable commercial alternative. With the dawn of genetic
engineering in the 1970s, new options emerged for making synthetic insulin that is chemically
identical to human insulin.
Insulin consists of a 21-amino acid A chain and a 30-amino acid B chain, linked by two disulfide
bonds. It can be produced either by generating the chains separately and chemically combining
them or by creating a single-chain precursor, human proinsulin, and cleaving out a 35-amino
acid connecting peptide. For manufacturing the drug, the proinsulin route is favored because it
requires a single fermentation and isolation step.
In 1978, the fledgling biotech company Genentech and City of Hope National Medical Center
produced human insulin in the laboratory using recombinant DNA (rDNA) technology. City of
Hope scientists had synthesized the genes for the protein's two chains before inserting them
into Escherichia coli. It was only the second time a human gene had been expressed in
bacteria; Genentech workers had expressed the hormone somatostatin in 1977.
With its expertise in purifying and handling insulin and the desire to remain a leader in the field,
Lilly immediately licensed recombinant insulin from Genentech and set about developing it.
Fortunately, Lilly also had experience isolating antibiotics from fermentation processes.
Because rDNA guidelines at the time allowed the expression of only inactive protein products,
Lilly had to use the two-chain method. But, anticipating an eventual easing in restrictions, it
simultaneously worked on the proinsulin route.
Clinical studies began in 1980. In 1982, Lilly's Humulin became the first genetically engineered
drug approved by the Food & Drug Administration. U.S. approval came just one month after
British regulatory authorities allowed its introduction in the U.K. The same year, Novo Nordisk
began selling the first semisynthetic human insulin made by enzymatically converting porcine
insulin; by 1987, Novo was using recombinant production methods as well.
Although the development of Humulin took just four years, it was a major undertaking for both
the developers and regulators, breaking new ground. FDA had quickly expanded its expertise in

genetic engineering to respond to new biotechnology products. Still, at a time when approvals
often took two years, the agency's quick endorsement of recombinant insulin, a unique and
somewhat controversial product, surprised outside observers.
"I had an excellent team that was highly motivated, and we wanted to get this done," explains
Henry I. Miller, the FDA medical officer who played a key role in advancing insulin and later
human growth hormone, the second rDNA drug, at the agency. Miller is now a research fellow
at Hoover Institution.
"The quality of the submission from Lilly was unsurpassed," Miller adds, "and the evidence of
safety and efficacy was unequivocal and copious." Like Lilly, FDA had decades of experience
approving and certifying insulin products, and the agency took just a few months to approve the
rDNA version.
A large part of the success in getting Humulin approved came from Lilly scientists being
"constantly in discussion with FDA and European regulators," says Bruce H. Frank, a retired
Lilly scientist. "We had to agree on what was going to be necessary to make everyone
comfortable with the safety of this product." Frank led the team that advanced the proinsulin
route, while Ronald E. Chance, also now retired from Lilly, headed one developing the A/B
chain process.


Stills, such as this
one from around
1931, were used to
isolate insulin from

BEING THE FIRST company to produce an rDNA product for human use involved many
unknowns and steps, starting with expressing the gene and then characterizing and
reproducibly making the drug, Frank and Chance explain. The program involved significant
resources and hundreds of scientists across many disciplines and functions within Lilly.
"The thrill of it was that nobody had ever done it before," Frank says about producing and
getting approval for the first recombinant drug. "The problem with it was that no one had ever
done it before."

Tasks included defining the biological, chemical, and physical characteristics of the recombinant
form, as well as addressing issues of immunogenicity, efficacy, and stability. Manufacturing and
purity considerations were crucial when employing the entirely new rDNA vectors and
production systems.
In addition, the chain combination process was not very efficient, Chance says, and had to be
improved to reach the scale and level of cost efficiency needed. A combination of specialized
chemistry and purification avoided contaminants from or the occurrence of a potential array of
mislinked or misfolded materials.
The researchers established a battery of nearly 20 physicochemical and analytical tests, many
of which were complex and not routinely practiced. X-ray crystallography, nuclear magnetic
resonance, ultraviolet, and other spectroscopic methods were used to confirm the correct 3-D
structure. High-performance liquid chromatography was emerging as a powerful tool and
became key in monitoring and controlling the manufacturing process and testing the product for
purity and identity.
The company also faced strong public opinion and an extremely cautious scientific and
regulatory community, according to Irving S. Johnson, former Lilly vice president for research
and front man on the project (Nat. Rev. Drug Discovery 2003, 2, 747). For example, rDNA
fermentation volumes were limited to just 10 L, well below the 40,000 L needed for production.
And there were strict rDNA hazard-containment requirements in the U.S., although Lilly also
had a lab in Europe where the restrictions were not as stringent.
Lilly did get exemptions allowing stepwise increases in fermentation capacity; by mid-1981, the
company was investing $80 million for new commercial plants in the U.S. and U.K. In the end, it
had to make a product that was at least as good as, if not better than, existing insulins. Up
through the 1970s, Lilly, Novo, and others had worked hard to increase the purity of animalbased insulins.
"The fact of the matter is that purified pork insulin was really quite an excellent product, and the
human recombinant version did not represent a huge advance," Miller points out. It was
"certainly nowhere near the advance that we saw when the animal insulins began to be used
and saved millions of lives virtually overnight." Humulin initially was also more expensive.
Since 1986, Lilly has employed a recombinant proinsulin route. More recently, producers have
used genetic engineering to create insulin analogs that differ in a few amino acids as a way to
control their onset and duration of activity. Today, more than 20 insulin and analog products are
available, including a small amount of animal-based materials. In 2004, insulin products
generated about $3.4 billion in sales for Novo, $2.1 billion for Lilly, and $1.1 for Sanofi-Aventis.
RECOMBINANT INSULIN may not have been a major medical achievement, but it was a
significant step in paving developmental, regulatory, commercial, and even political and societal
paths for genetically engineered drugs. For the first time, it married a biotech firm's new
technology with the resources of a major pharmaceutical partner. It was a proof of principle that
sparked the interest of large drug companies, generating the funding and product revenues to
fuel the emerging biotech industry.
The real challenge, however, remains in preventing or curing diabetes (C&EN, Oct. 25, 2004,
page 59). In drug therapy, work continues on analogs and delivery methods, such as inhaled
insulin. Many drugs are available to treat the more common type 2 diabetes, in which insulin

production or the body's response to insulin needs to be increased. Insulin had been the only
treatment for type 1 diabetes until March, when FDA approved Amylin Pharmaceuticals' Symlin
to control blood sugar levels in combination with the polypeptide hormone. ANN THAYER

CAS Registry
Common Product Names
Porcine and bovine forms: Iletin, Lente
Semisynthetic human insulin: Novolin
Recombinant insulin and analogs:
Apidra, Humulin, Humalog, Novolin, NovoLog, Levemir, Lantus, Velosulin
1922 in Great Britain via the Medical Research Council, and on a larger scale in 1923 by Eli
Lilly & Co.
$6.75 billion in 2004 (insulin and analogs)