Late last year, Denmark-based Novo Nordisk quietly started phase 1 clinical trials of a pill that it hopes to market as an alternative to insulin injections. Offering patients with diabetes the chance to avoid painful needles has long been the holy grail of some pharmaceutical companies, especially Novo, a leader in diabetes treatment for much of its 87-year history. Novo's entry into clinical testing puts the company ahead of the pack of drugmakers trying to encapsulate insulin into an easy-to-swallow dose.

The fact that Novo entered phase 1 testing with little fanfare is apropos, however, considering the checkered history of insulin drug development. Insulin is a protein that rapidly degrades in the stomach and upper portion of the small intestine, making it almost impossible to deliver orally. Several drugmakers have attempted to deliver insulin via the lungs, most notably Pfizer, which introduced its inhalable insulin drug, Exubera, in 2006. But the product flopped--patients balked at the clumsy inhaler and doctors questioned whether it might endanger the lungs. Several other drugmakers, Novo among them, abandoned their attempts to make inhalable insulin in the wake of Exubera's failure.

Theoretically, delivering insulin through the stomach should be ideal: insulin would travel directly to the liver, its primary site of action, in a way that mimics the action of endogenous insulin. Because insulin injections go into muscle and fat, patients who use them are susceptible to hypoglycemia. Mads Krogsgaard Thomsen, Novo's chief scientific officer, says that's why he believes oral insulin could be safer and more convenient than either the injectable or inhalable forms. But he urges caution for now.

The challenge of making an insulin pill that can survive the digestive system has been magnified by the characteristics of the molecule itself. Human insulin is a large and complex protein. So even if it survives being doused by stomach acid, it's unlikely to be easily absorbed by the epithelial cells of the gut. Thus the portion of oral insulin that actually makes it into the bloodstream is under 1 percent. "You can't use human insulin" in a pill, Thomsen says. "It doesn't work."

So Novo's scientists set out to increase the oral viability of insulin tenfold through protein engineering. They embarked on a painstaking, multiyear process designed to answer some key questions: What are the enzymes that attack the molecule, and where do they attack it? The scientists used the insight they gained to design a pill that can transverse the stomach without being broken down, facilitating the transfer of insulin across cells and into the bloodstream.