Cloning Insulin

In 1978, Genentech scientist Dennis Kleid toured a factory in Indiana where insulin was being made from pigs and cattle. “There was a line of train cars filled with frozen pancreases,” he says. At the time, it took 8,000 pounds of pancreas glands from 23,500 animals to make one pound of insulin. Diabetics lack this hormone, which regulates the amount of glucose in the blood. The manufacturer, Eli Lilly, needed 56 million animals per year to meet the increasing U.S. demand for the drug. They had to find a new insulin alternative, fast.

Genentech had the expertise to make synthetic human insulin—in laboratories, from bacteria, using their recently-proven recombinant DNA technology. But could they make enough of the miniscule insulin molecules to replace these trainloads of pancreases and provide an alternative option for people living with diabetes? The scientists would have to coax the bacteria to produce insulin from the synthetic DNA at high enough concentrations to make an economically viable product. This meant that each bacteria needed to churn out so much of the protein per cell that if they could do it, they’d look like stuffed olives under a microscope. If not, Genentech’s work would have ended as a scientific curiosity, with no new option for diabetics.

I don’t want to hear that word, impossible...tell me what you need to get it done.

Kleid didn’t think they could get that kind of yield. He told Genentech founder, Bob Swanson, flat-out that it couldn’t be done. But Swanson refused to accept it. “I don’t want to hear that word, impossible,” he told Kleid. “Tell me what you need to get it done.”

The high-stakes, high-pressure race to create synthetic insulin had started over a year earlier. Eli Lilly, the main U.S. producer of insulin, had set the stage by signing contracts with competing institutions to bioengineer the hormone. Already, Harvard and the University of California, San Francisco (UCSF) had been working on rat versions of the insulin gene. With only twelve employees, Genentech joined the race against the biggest research institutions in the world—a David against a pack of Goliaths.

But the company was scrappy and imbued with Swanson’s can-do drive. After Genentech had proven the success of its biotechnology by synthesizing somatostatin in tiny amounts, Swanson had been able to raise another round of funding. He hired a team and began to outfit an airfreight warehouse in South San Francisco with a lab. Among the first to join the company were two scientists—Kleid, an organic chemist who’d been working on cloning DNA at Stanford Research Institute, and his colleague David Goeddel. The team—including scientists Roberto Crea, Arthur Riggs, and Keiichi Itakura—hit the ground running, working around the clock to be the first team to synthesize human insulin. Since the Genentech lab wasn’t yet up and running, Goeddel and Kleid commuted from their Bay Area homes to a lab in the City of Hope National Medical Center in Los Angeles. “Dave was the early person, and I was the night person,” Kleid says. They hardly slept, ate, or saw their families. “We kept the experiment going 24 hours a day.”

Dennis Kleid and a colleague in the molecular biology lab. Photo from the Genentech archives.

The stakes were high. First, there was pressure from investors: if Genentech wasn’t able to synthesize insulin, “there would be no more company,” as Kleid put it. And it wasn’t going to pay to come in second. “You either came in first or you might as well be last,” Goeddel later said.

The first challenge for the Genentech team was to improve upon the gene-splicing technique they’d developed with somatostatin on the more challenging insulin molecule, which has 51 amino acids instead of 14. And because of insulin’s more complex structure, they also needed two chains of insulin-encoding DNA working efficiently in two different bacteria instead of one. The scientists had to synthesize the genes by chemically linking together snips of DNA sequences and then stitch those genes into the plasmids—the rings of DNA found inside cells—and transplant them into benign E. coli bacteria. With powerful gene control elements, this would hijack the machinery the bacteria normally used to produce their own proteins to churn out the two insulin chains. The last step—after harvesting, isolating, and purifying these insulin protein chains—was to chemically combine the two chains to form the complete insulin molecule, identical to the one produced by the human pancreas.

The team kept running into setbacks, but continued to push ahead. Kleid remembers how intensely he and Goeddel worked. “Every time I see Tiger Woods, I think of Dave Goeddel,” he says. “He’s concentrating so hard, every shot he makes is going to be the best shot he’s done in his life. That’s the way Dave is.” Finally, in the early hours of August 21, 1978, Goeddel succeeded in reconstituting the two amino acid chains into one molecule: human insulin.

With only twelve employees, Genentech joined the race against the biggest research institutions in the world – a David against a pack of Goliaths.

It was an extraordinary moment, not only for Genentech, but for the history of medicine and the future of biotechnology. When Goeddel told the rest of the team the news, they were elated. Kleid likened the feeling to finishing a marathon. “You’re pretty much exhausted when you get to the line, and it takes a while for it to soak in that you actually won the race.”

The scientists scraped together enough of the insulin for Eli Lilly to conduct clinical trials, in which they found that not only was the synthetic insulin as effective as its chemically identical human twin, it eliminated the allergies that the animal-derived product caused in some diabetics.

Genentech's publication on insulin1 appeared in the Proceedings of the National Academy of Sciences of the United States in 1979. Photo from the Genentech archives.

But their work wasn’t done yet. After that huge moment of accomplishment, Kleid’s prediction about the low insulin yield held true. The team still had to figure out how to get the bacteria to produce enough insulin—50 times the yield—to meet the demand that had been met by those train cars full of animal pancreases. Eventually, with the help of Genentech scientists including Herbert Heyneker, Dan Yansura and Giuseppie Miozzari, they discovered a powerful control gene that, at the right moment, instructed the plasmids to reproduce in large quantities, filling the bacteria with the precious insulin peptides. It took another two years beyond their initial success to finally complete their mission.

In 1982, the FDA approved human insulin and it was on the market by 1983. Since then, millions of people have used the medicine, and it has almost completely replaced insulin created from animals.

“It was the thing I’m most proud of,” says Kleid. “We had all these statistics that there wasn’t going to be enough insulin to go around unless we made this technology work. And we did.”

1 Goeddel DV, Kleid DG, Bolivar F, Heyneker HL, Yansura DG, Crea R, Hirose T, Kraszewski A, Itakura K and Riggs AD. Proc. Natl. Acad. Sci. USA 1979 Jan;76(1):106-10.