According to scientists on Sunday, far more research is needed to see if the experimental treatment - which requires a form of brain surgery - really helps.
But if the approach pans out, researchers say, delivering protective substances, called growth factors, into a diseased brain holds the potential to rescue some dying brain cells.
In one patient, the brain tissue showed new growth, which was a first, according to a study published in Sunday's edition of the journal Nature Medicine.
"It won't cure the disease," said the lead researcher, Dr Mark Tuszynski, of the University of California, San Diego. That is because Alzheimer's destroys different types of cells in different areas of the brain - the new gene therapy targets just one of those.
The preliminary success indicates that similar approaches might help other neuro-degenerative diseases such as Parkinson's Disease, Tuszynski said.
"This is, in a sense, proof of principle for the potential use of growth factors."
Doctors at Chicago's Rush University Medical Centre have begun a second small study of the approach in Alzheimer's patients.
Tuszynski, who co-founded a biotechnology company that is funding the Chicago work, hopes larger studies will begin within another year.
He took skin cells from eight patients in the early stages of Alzheimer's and modified the genes to secrete a protein found in healthy brains called nerve growth factor or NGF.
Doctors used skin cells to secrete
a protein found in healthy brains
Earlier studies had shown that injecting NGF-producing tissue into the brains of ageing monkeys could reverse deterioration. Simply injecting NGF into people would not work. If it goes into the wrong part of the brain, it can cause serious side effects.
So, doctors drilled holes into the patients' skulls and implanted the NGF-producing skin cells directly on to Alzheimer's-injured spots.
Six patients were tracked for almost two years. Tests involving memory and other skills found their rate of cognitive decline slowed by 36% to 51%, better than is usually seen with medication, Tuszynski reported.
First few results
The first two patients were awake and moved during the cell implantation, causing bleeding in their brains; one patient died five weeks later. The remaining ones received the implants under general anaesthesia to keep them still and the researchers reported no further problems.
"It is cautious optimism with a big C. It cannot be
a cure, obviously ... but maybe it will do something"
Dr David Bennett,
Rush University Medical Centre
The brain tissue of the study participant who died had new growth protruding from Alzheimer's-injured cells at the implant site, the first time that sort of recovery has been seen from a human brain treatment, Tuszynski said.
Additionally, PET scans of the other participants showed a measurable increase in their brains' metabolic activity.
"These results need to be interpreted with cautious optimism," William Thies of the Alzheimer's Association said.
With so few patients in the study, "it's really impossible to tell whether the benefit was due to the treatment or natural fluctuation in symptoms".
Moreover, he said, it would never be practical to perform brain surgery on millions of patients. Already, 4.5 million Americans have Alzheimer's, and with the country's population ageing, a staggering 14 million may have it by 2050.
But if the gene-therapy approach ultimately works, it could revive interest in finding easier methods, he said.
As for the next step, instead of genetically modifying skin cells, Dr David Bennett of the Rush University Medical Centre has begun injecting the brains of up to 12 Alzheimer's patients with an NGF-bearing virus.
Animal studies suggest the virus may penetrate injured brain cells better, producing more of the protective growth factor for longer periods, he said.
"It is cautious optimism with a big C," Bennett stressed. "It cannot be a cure, obviously ... but maybe it will do something."