Gene therapy has markedly improved vision in both eyes in three women who were born virtually blind. The patients can now avoid obstacles even in dim light, read large print and recognize people's faces. The operation, researchers predict, should work even better in children and adolescents blinded by the same condition.

The advance, reported in the February 8 issue of Science Translational Medicine, extends earlier work by the same group. Between 2008 and 2011, Jean Bennett of the University of Pennsylvania's Mahoney Institute of Neurological Sciences and her colleagues used gene therapy to treat blindness in 12 adults and children with Leber's congenital amaurosis (LCA), a rare inherited eye disease that destroys vision by killing photoreceptors—light-sensitive cells in the retina at the back of the eye. Typically, afflicted children start life with poor vision, which worsens as more and more photoreceptors die.

The treatment grew out of the understanding that people with the disorder become blind because of genetic mutations in retinal cells. One mutated gene that causes the disorder is named RPE65. An enzyme encoded by RPE65 helps break down a derivative of vitamin A called retinol into a substance that photoreceptors need to detect light and send signals to the brain. Mutant forms of RPE65 prevent the production of this enzyme in a "nursery" layer of cells called the retinal pigment epithelium, which is attached to the retina and nourishes photoreceptors by breaking down retinol, among other cellular services.

In the initial study, retina specialist and Bennett's co-author Albert Maguire of Penn Medicine injected a harmless virus carrying normal copies of RPE65 into an area of the retinal pigment epithelium, which subsequently began producing the enzyme. In each of the 12 patients, Maguire treated one eye—the one with worse vision. Six patients improved so much they no longer met the criteria for legal blindness.

In the new study, Maguire injected the functional genes into the previously untreated eye in three of the women from the first group. Bennett followed the patients for six months after their surgeries. The women's vision in their previously untreated eye improved as soon as two weeks after the operation: They could navigate an obstacle course, even in dim light, avoiding objects that had tripped them up before, as well as recognize people's faces and read large signs. Bennett showed that not only were the women's eyes much more sensitive to light, their brains were much more responsive to optical input as well. Functional magnetic imaging showed regions of their visual cortices that had remained offline before gene therapy began to light up.

Surprisingly, Bennett reports, the second round of gene therapy further strengthened the brain's response to the initially treated eye as well as the newly treated one. "That wasn't something we had been expecting, but it makes sense because the two eyes act in concert, and some aspects of vision rely on binocularity." In the new paper, the authors suggest that neuroplasticity plays a role: It is possible that regions of the visual cortex responding to the newly flowing channel of information from the second eye bolster activity in areas of the visual cortex responding to the initially treated eye.

An institutional review board required that Bennett work with adults in the follow-up study, but she thinks the therapy will work even better in younger patients who have not lost as many photoreceptors. She says the results "really bode well" for restoring meaningful vision to people with LCA and other forms of inherited blindness.