WASHINGTON - Three pharmaceutical giants are unlocking their freezers to see if government-funded scientists can reinvent some of their old drugs.

Pfizer, AstraZeneca and Eli Lilly & Co. entered a unique program with the National Institutes of Health on Thursday that both sides hope will speed the development of new treatments -- by dusting off two dozen old drugs that failed to treat one disease but might treat another.

"The goal is simple, to see whether we can teach old drugs new tricks," said Health and Human Services Secretary Kathleen Sebelius.

Lots of experimental drugs prove safe in early human testing but fail to help the disease their manufacturer had hoped to treat. Despite the years of work and tens of millions of dollars invested in them, "too many times these compounds, they end up sitting on shelves or they end up in somebody's freezer," said Pfizer senior vice-president Rod MacKenzie.

Some of those drugs might be able to fight other diseases, said NIH Director Dr. Francis Collins.

Consider: A failed cancer drug turned into the first effective AIDS treatment, AZT. The notorious thalidomide caused birth defects in the 1960s when some countries used it for morning sickness, but today it treats multiple myeloma. The bone drug raloxifene was found to also help prevent breast cancer.

Those discoveries "all have been sort of serendipitous. The idea here is, let's not depend on serendipity," Collins said.

In recent years, researchers have identified at the genetic level the causes of more than 4,500 diseases, many of them rare diseases, he said. But only 250 of those conditions have effective treatments.

Likewise, manufacturers have a lot of information about the specific molecules their failed drugs target. Collins' plan: Try to match those old drugs to these newly discovered disease pathways.

Under the new program, the drug companies will make at least two dozen of their shelved drugs, and the data about them, available for NIH-funded research. The NIH will award grants to scientists around the country who apply to study specific drugs, with the goal of rapidly beginning human trials of promising candidates since the required safety testing already has been done.

And rather than those scientists undergoing what Lilly executive vice-president Jan Lundberg called "endless discussions about legal agreements" before getting to work, the program provides a streamlined approach: The companies retain ownership of their drugs, but the researchers can patent and publish their own discoveries.

The NIH plans to spend about $20 million in the program's first year, and hopes other drug companies will join.

AstraZeneca said it began partnering with British researchers last December in a similar program.