US scientists increased the activity of an enzyme called Uch-L1 which is involved in memory function. They then tested the mice and found that they had regained the ability to form new memories. Writing in the journal Cell, the team said the work was in its early stages, but could help the development of therapies for the debilitating disease.

The scientists used transgenic mice that had Alzheimer's disease. Like humans with the condition, the mice had plaques of protein amyloid beta in their brain and damaged synapses (the site where brain cells communicate with each other), and they also suffered from memory loss.

The researchers injected the mice with a substance which boosted levels of the enzyme Uch-L1. They then tested the mice's memory by putting them into a cage where they were exposed to a very mild shock from the cage floor.

Mice with normal memories stay still when they are placed in the cage for a second time, to avoid the shock, whereas mice with Alzheimer's do not because they cannot remember their first visit. The researchers found that when they placed the mice with boosted Uch-L1 levels into the cage, they stayed still like the normal mice.

Dr Ottavio Arancio, an author on the paper from Columbia University's Taub Institute for Research on Alzheimer's Disease and Aging, said: "We injected the mice with this enzyme, and we found we were restoring their memories."

He said further investigation revealed the enzyme also seemed to repair the synapse function of the mice.

Possible therapies

Michael Shelanski, another author on the paper from Columbia University, said: "While amyloid beta is certainly a key player in Alzheimer's disease — and efforts to reduce it remain a worthy goal — our results show that, even in the presence of the plaque, damage to memory can be reversed."

The authors propose that the enzyme could be targeted to look for new therapies for Alzheimer's disease; however they cautioned that the research is in its early stages.

Dr Shelanski said: "While this discovery is very promising, its proven effectiveness is limited to animal models and it will take some time before it could lead to therapies in humans. We continue to work towards that crucial goal."

In an accompanying commentary, Peter Lansbury, a neurobiologist from Harvard Medical School, Massachusetts, said: "Understanding the molecular mechanism of memory is one of the most compelling and complex challenges for the next generation of scientists. [The researchers] identify a protein that may be involved in both normal memory function and the type of memory loss characteristic of early Alzheimer's disease."

Harriet Millward, deputy chief executive of the Alzheimer's Research Trust, said: "This is a very encouraging piece of research that could potentially open up a new front in the battle against Alzheimer's disease. "It is fair to say though, that at this stage a cure is still a long way off. More research is needed to see if this result could eventually be extrapolated from mice to men."