The U.S. government is looking for the next AZT, Viagra and thalidomide — substances that washed out as treatments for one disease but later turned out to work well against a totally different ailment.

AZT failed as a cancer drug but became the first antiviral to work against HIV. Viagra didn’t succeed as a treatment for angina but turned out to be great against impotence. Thalidomide, the notorious morning sickness pill that caused thousands of birth defects in Europe, is a treatment for leprosy and multiple myeloma.

The National Institutes of Health and three large pharmaceutical companies hope there are dozens of other failed drugs waiting to be rescued and repurposed. On Wednesday, they announced a collaboration to look for them.

‘Teach old drugs new tricks’

“The goal is simple: to determine whether we can teach old drugs new tricks,” Health and Human Services Secretary Kathleen Sebelius said at a news conference.

NIH will invest $20 million a year for the next three years in the effort, the first project of the National Center for Advancing Translational Sciences, created last year by NIH.

AstraZeneca, Pfizer and Eli Lilly will make available two dozen compounds that tested safe for human use but were abandoned because they didn’t work for the intended purpose. NIH will function as matchmaker, bringing academic researchers, the compounds and the companies together. It will also provide seed money for experiments.

“The idea here is: Let’s not depend on serendipity,” said Francis S. Collins, the NIH director. “Let’s go to the ‘crowdsourcing’ model.”

If studies suggest that a compound has a previously unsuspected clinical use, the company that owns it will get first crack at bringing it to market, a process that will require the investment of millions more dollars. Profits would ultimately be divided between the company and the outside scientist who identified the new use (along with the scientist’s institution).

NIH receives no income from research it funds that leads to licensed drugs, unless the work is done by one of its in-house scientists. That is true of this arrangement as well.

For academic scientists, the advantage of the collaboration is that they get access to compounds with information that has taken years and millions of dollars to generate. That includes a molecule’s pharmacological effects — what receptors and enzymes it blocks or stimulates — as well as safe doses and side effects.

The advantage to pharmaceutical houses is the chance to recoup some of their lost investment. “This is a true business strategy for us,” said Rod MacKenzie, a senior vice president of Pfizer.

Don Frail, a vice president of AstraZeneca, remarked:“It’s not a donation. It’s a collaboration.”

There are many reasons drug companies aren’t doing this already.

A novel use may take a company into a field that isn’t part of its business strategy. The potential market may be too small, or the risk of failure too high, to warrant further investment in an abandoned drug. New or unpublished discoveries that suggest a new use may be not be known to company scientists.

“Pharma companies do not have the time and resources to test all the possible uses,” said Jan M. Lundberg, a Lilly executive at the news conference. “We’re usually testing one, perhaps two, indications and then we stop.”

AstraZeneca recently launched a similar program with the Medical Research Council, Britain’s equivalent of NIH. The company got more than 100 proposals from scientists in 37 institutions seeking access to 21 of the 22 compounds it is offering.

“We’re getting proposals that include some very novel concepts, things we hadn’t previously considered,” Frail said.

NIH will post a description of the compounds in about a month. Researchers will submit applications for a grant to study a compound. Those that pass peer review will be passed on to the companies, which decide who receives access to their materials and data.

The companies incur some expense. They must supply a compound in usable form — pills or injectables — to the researchers. There is also the cost of maintaining a compound’s patent protection. In all, the investment may run $200,000 to $800,000 a year per drug, Frail said.

Big advantage

A big advantage of the program is that the companies have agreed to use a “template agreement” outlining the financial arrangements, sparing long and expensive negotiations with researchers.

Collins said he hopes the project may lead to drugs to treat neurological diseases.

Such medicines are rare because it is hard to get molecules into the brain, and there are often unexpected side effects from drugs that do get there.

He mentioned bexarotene, now used to treat a rare form of lymphoma, as an example of a drug that might someday find a new life. A study published in February showed that it helps dissolve brain-clogging plaques in mice with a simulated form of Alzheimer’s disease.