The virus that causes AIDS can hide in the bone marrow, avoiding drugs and later awakening to cause illness, according to new research that could point the way toward better treatments for the disease.

Finding that hide-out is a first step, but years of research lie ahead.

Dr. Kathleen Collins of the University of Michigan and her colleagues report in this week's edition of the journal Nature Medicine that the HIV virus can infect long-lived bone marrow cells that eventually convert into blood cells.

The virus is dormant in the bone marrow cells, she said, but when those progenitor cells develop into blood cells, it can be reactivated and cause renewed infection. The virus kills the new blood cells and then moves on to infect other cells, said.

"If we're ever going to be able to find a way to get rid of the cells, the first step is to understand" where a latent infection can continue, Collins said.

In recent years, drugs have reduced AIDS deaths sharply, but patients need to keep taking the medicines for life or the infection comes back, she said. That's an indication that while the drugs battle the active virus, some of the disease remains hidden away to flare up once the therapy is stopped.

One hide-out was found earlier in blood cells called macrophages. Another pool was discovered in memory T-cells, and research began on attacking those.

But those couldn't account for all the HIV virus still circulating, Collins said, showing there were more locations to check out and leading her to study the blood cell progenitors.

Finding these sources of infection is important because eliminating them would allow AIDS patients to stop taking drugs after their infection was over. That's critical in countries where the treatment is hard to afford and deliver.

"I don't know how many people realize that although the drugs have reduced mortality we still have a long way to go," Collins said in a telephone interview. "That is mainly because we can't stop the drugs, people have to take it for a lifetime."

The research was funded by the National Institutes of Health, Burroughs Wellcome Foundation, University of Michigan, Rackham Predoctoral Fellowship, National Science Foundation and a Bernard Maas Fellowship.