The National Institutes of Health said on Thursday that Pfizer Inc, AstraZeneca Plc and Eli Lilly and Co have agreed to make 24 compounds available for a pilot phase of the project, the biggest of its kind ever launched in the United States.

All of the compounds have been tried in people and found to be safe, but the drugmakers have abandoned them because they did not work for the disease they were intended to treat. The NIH will provide $20 million in grants each year to researchers trying to find new uses for the compounds.

"The goal is simple: to see whether we can teach an old drug new tricks," said Health and Human Services Secretary Kathleen Sebelius during a press conference.

Big advances in genetics have allowed researchers to identify causes of more than 4,500 diseases. But turning those discoveries into treatments has been difficult, and there are drugs to treat only about 250 of these conditions.

NIH Director Dr. Francis Collins said "diseases of the brain" represent an especially urgent unmet need.

"We need to speed the pace at which we are turning discoveries into better health outcomes," Collins said. "NIH looks forward to working with our partners in industry and academia to tackle an urgent need that is beyond the scope of any one organization or sector."

It generally takes a drugmaker $2 billion and 14 years to turn an initial idea into an effective medicine sold to patients, including the cost of testing medicines that ultimately don't lead to results.

Drugmakers may have scores of initially promising compounds that have already cost tens of millions of dollars sitting in warehouses or freezers with no apparent purpose.

Collins said there are dozens of examples of drugs originally meant for one disease being repurposed for another. In a high-profile case, azidothymidine, or AZT, was approved to treat cancer in the 1960s but shelved for lack of efficacy. More than two decades later, it became the first effective treatment for HIV, the virus that causes AIDS.

But in the past, such discoveries have occurred by chance.

"The idea here is not to depend on serendipity," Collins said, by opening up the failed drugs to researchers all over the country.

Researchers who discover a new use for a drug will be eligible for royalty payments if the drug reaches the market, though the companies will still hold the rights to the drug compound.

Pfizer has already attempted the program on a smaller scale with a collaboration with Washington University in St. Louis, while AstraZeneca launched a similar project with the British version of the NIH, the Medical Research Council.

The British project includes more than 20 drugs originally targeted at everything from tumors to Alzheimer's to diabetes and obesity. AstraZeneca said the project also includes studying drugs already on the market, not just those abandoned during testing.

The NIH said it would release the names of the 24 compounds in the pilot project next month.

(Reporting By Julie Steenhuysen in Chicago and Anna Yukhananov in Washington)

(This story was filed to correct the name of Pfizer research partner to Washington University, not University of Washington in paragraph 14)

• A Permanent Economic Emergency