By inserting a corrective gene with a virus, the team induced the formation of new cochlear hair cells -- the key intermediates in converting sound waves into electrical impulses -- in the ears of artificially deafened adult guinea pigs.

They later demonstrated that the animals responded to sounds, according to the study published today in the journal Nature Medicine.

"A lot of the techniques would fairly easily translate into a clinical setting" for use in humans, said neuroscientist Matthew Kelley of the National Institute on Deafness and Other Communication Disorders, who was not involved in the research.

Humans have about 16,000 hair cells in the cochlea of each ear, where they convert sound waves into nerve impulses. The cells are easily damaged by loud noises, aging, infections and certain medications. Once damaged, they cannot regenerate on their own.

The key to the generation of new hair cells is a gene called Atoh1, first discovered in fruit flies in 1998. Variants have since been discovered in almost all species of animals.

During fetal development, the gene converts some cells in the ear into hair cells. In other ear cells, called supporting cells, its activity is suppressed.

Researchers showed that, working in laboratory dishes, the gene could convert supporting cells into hair cells.

Two years ago, Yehoash Raphael and Kohei Kawamoto of the University of Michigan Medical School reported that inserting the gene into supporting cells in live guinea pigs produced thousands of new hair cells.

But in those experiments, the researchers did not deafen the animals first. This time they did, using toxic chemicals to kill the hair cells in both ears of 10 guinea pigs. Microscopic images taken three days later confirmed that all the hair cells had been destroyed.

On the fourth day, they used gene therapy with a viral vector to insert the Atoh1 gene into the guinea pigs' left ears. Within two months, new hair cells had appeared in the treated ears. The team then tested auditory brainstem response to measure the animals' ability to hear.

Raphael and his colleagues are now trying to determine how good the restored hearing is. Experts said the required studies for safety and efficacy mean that it will be the better part of a decade, at least, before the technique can be tried in humans.

By Thomas H. Maugh Ii