A pill or potion that dissolves away one of the primary causes of Alzheimer's disease in the brain could be in sight after ground-breaking animal studies.
Scientists who added the molecule to the drinking water of affected mice found that harmful protein fragment deposits in their brains were broken down.
The animals, genetically engineered to develop symptoms of Alzheimer's disease, also showed improvements in memory and learning.
Experts say much more work needs to be carried out before the drug, EPPS, can be tested on human patients.
But the new approach is said to look promising, especially for the treatment of people at risk from inherited forms of the disease.
The way the drug works is unclear but it is believed to make the amyloid beta "plaques" that build up in the brains of patients soluble, so they can be cleared away.
However, it may only be effective before clinical symptoms appear, highlighting the importance of finding new ways to diagnose cases earlier.
Amyloid accumulation is believed to occur first, leading on to other changes such as the formation of "tangles" of tau protein within nerve cells that actually trigger the devastating effects of Alzheimer's.
Dr Frances Edwards, a reader in neurophysiology at University College London, pointed out that most people with the disease are only diagnosed after the appearance of tau tangles and "considerable loss of brain tissue".
Commenting on the findings published in the journal Nature Communications, he said: "If we could catch Alzheimer's disease before this occurred, such an amyloid-removing drug might stop it in its tracks.
"Many more tests need to be done, but for people with the rare inherited forms where Alzheimer's disease can be predicted long in advance, this could be a very interesting drug indeed."
The research was conducted by a South Korean team led by Dr YoungSoo Kim from the Korea Institute of Science and Technology in Seoul.
Amyloid beta consists of sticky fragments of protein that clump together in the brain and are one of the key hallmarks of Alzheimer's.
They are found in the brains of all deceased patients, but appear to pre-date the onset of memory loss and other symptoms in humans.
After success with cell cultures, the Korean team tried adding large doses of EPPS to the drinking water of mice suffering from Alzheimer's.
They found that the drug broke down the amyloid beta plaques in the animals' brains, and appeared to be non-toxic. Unlike many drugs, EPPS was able to penetrate the "blood-brain barrier" - a natural "firewall" that protects the brain against potentially harmful substances.
The scientists adopted a cautious approach, concluding: "Additional studies are warranted to determine whether these favourable actions of EPPS and derivatives will translate into a therapy that might potentially be useful across a range of AD (Alzheimer's disease) stages."
The fact that the treatment improved learning and memory in the mice was evidence that amyloid beta is a cause of the disease rather than an effect, they said.
British experts were also careful not to over-hype the results.
Dr James Pickett, head of research at the charity Alzheimer's Society, said: "While new insights into potential ways to treat Alzheimer's disease are welcome, in this case, it's too early to say if this will one day benefit people living with Alzheimer's.
"This study has found a new chemical that could help to clear some forms of amyloid, the toxic protein that is a hallmark of Alzheimer's disease. However, so far this has only been shown in mice, which do not fully replicate several of the important changes that we see in the brains and behaviours of people with dementia."
Dr Tara Spires-Jones, from the Centre for Cognitive and Neural Systems at the University of Edinburgh, said: "This is a promising preclinical study in a few dozen mice that found a drug treatment that reduced levels of amyloid beta, one of the pathological proteins found in the brains of Alzheimer's patients.
"While very interesting, this potential treatment is a long way from human patients."
Professor Tom Dening, a dementia expert from the University of Nottingham, said: "From a clinician's point of view, this research is of interest, but we still don't know if removing amyloid plaques is useful in humans. It may well be that the appearance of plaques is too far down the chain of molecular processes to be beneficial."
Dr Simon Ridley, head of science at Alzheimer's Research UK, said: "This interesting study in mice uses a chemical able to break down Alzheimer's plaques in the brain after they've already formed, potentially presenting a way to circumvent the difficult issue of timing with this treatment approach.
"While this is an appealing prospect, the research is still at the early stage of being explored in animals. Research in animals is an important step in developing any new treatment, but we'll need to see the findings translated into clinical studies in people before we could know the potential of EPPS to treat Alzheimer's in humans."