AD is the commonest form of dementia. Named after the German doctor, Alois Alzheimer, who first described it in 1907, the progressive condition leads to the destruction of nerve cells in the brain, notably in the hippocampus and the amygdala deep within its structure, and parts of the cortex in the outer part of the brain. It is estimated that around 12 million people worldwide are afflicted by AD, with nearly 400,000 patients in the UK, and about 10 times that in the US. Its prevalence is expected to escalate dramatically over the next quarter of a century, with the anticipated doubling in the worldwide population over 65 being accompanied by a doubling in the number of AD patients.

Three genes implicated in AD have so far been identified. Mutations in these lead, in the main, to the early-onset form where symptoms can begin in people as young as 30. Amyloid precursor protein (APP) is linked with this form, though mutations in it are very rare. The second gene to be identified, Presenelin-1 (PS1) was pinpointed in 1995, and at least half of all early-onset patients have a mutation in this gene. Presenelin-2 was found shortly afterwards, and this is thought to be responsible for another 10% of cases. Other genes are still being hunted. However, 95% of AD patients suffer from the late-onset form of the disease, and while some genes have been identified that increase the risk of developing the disease, they are largely not thought to cause AD directly.

The brain of a patient who has died from AD will have lost between 30 and 50% of its mass, with the proportions of grey and white matter being significantly changed, and the outer folds and inner cavities having become larger. The most dramatically affected areas are those involved in cognitive processes: the hippocampus, which is largely responsible for memory processing, and the cerebral cortex, where logical and rational thought take place.

The basic cause of these changes is loss of neurons. A new-born baby has over 100 billion neurons, connected to at least at thousand others through dendrites and synapses. Neurons are thought to be lost at a rate of a million a day, amounting throughout a normal lifespan to around 10% of the total, and accounting for the increasing forgetfulness that develops with age. Trunks of neurons connect different areas of the brain and, as some of these are particularly badly affected by AD, the result is an inability of the remaining neurons to compensate, and a consequent loss in mental processing abilities.

The neurons most affected by AD are those that rely on the neurotransmitter acetylcholine (ACh), which acts at two different types of receptors: muscarinic and nicotinic. Levels of the enzyme that makes ACh, choline acetyltransferase, or ChAT, drops in the brains of AD sufferers, leading to a drop in the level of ACh itself. Amyloid plaques form between the remaining neurons, and these often lead to swelling and distortion of neighbouring neurons.

Four drug products are currently licensed for the treatment of AD, all of which are inhibitors of acetylcholinesterase (AChE), the enzyme that breaks down ACh. The first drug to be licensed for the treatment of mild to moderate dementia of the Alzheimer's type was tacrine, developed by Parke Davis (now Pfizer) and Shire, and marketed as Cognex. It was launched as a four-times-a-day formulation in the US in 1993, and in France in 1998. However, it has marked hepatotoxic effects, and its efficacy is limited. It has never been launched in the UK.

Pfizer's launch of donepezil (Aricept) led to the cessation of the active promotion of tacrine in the US. Marketed in the US under license from Japanese company Eisai, the AChE inhibitor was launched in the US and Europe in 1997, and Japan in 1999. Donepezil is the current market leader, with sales of US$92m in 1999, escalating to US$700m in 2000. It is relatively selective for brain acetylcholine (ACh) levels, and has fewer side-effects than tacrine, with the additional advantage of being a once-daily formulation. The drug has been shown to enhance mental alertness and activity and improve behaviour, although it has no lasting effect on overall disease progression.

Also making its debut in 1997 was Novartis' rivastigmine (Exelon). First launched in Switzerland, it was marketed in the UK the following year and the US in 2000, but has thus far had little impact on the sales of donepezil. Its efficacy is similar to that of donepezil, but it must be administered twice a day. However, it has few interactions with other drugs, making it particularly useful for elderly patients who are often prescribed a large number of different medicines. Its AChE binding properties are different from those of donezepil, and it does not rely on the liver for removal from the body.

The fourth licensed treatment for AD is galantamine (Reminyl), from Shire and Janssen, which was launched in the UK last year, and received FDA approval in the US earlier this year. The molecule, extracted from daffodils, has a novel dual mechanism of action, as it is a nicotinic receptor modulator as well as being an AChE inhibitor. It is thought that nicotinic modulators may increase the release of acetylcholine and other neurotransmitters. Some patients have remained stable or even seen a slight improvement after six months of treatment. Regulatory opinions differ about its effect on the progression of the disease, with the Europeans saying that it slows the advance of dementia, and their US counterparts disagreeing. Further trials are under way, but sales could well reach US$1bn if the regulators can agree that it does indeed slow disease progression.

A further AChE inhibitor in development is TAK-147, being investigated by Takeda. Related to galantamine, it also has nerve growth factor (NGF)-like properties. The compound appears to survive for long periods of time in the brain, and has a beneficial effect on other pathways in the brain that stimulate the brain energy metabolism that is depressed in AD patients. It acts like a naturally-occurring NGF, stimulating the growth of neurons that use ACh, and the company believes it may slow or even prevent the progression of the disease in addition to improving symptoms. The drug is now in Phase III trials.

Many other AChE inhibitors are under development, including P11149 from Aventis, Chiesi's ganstigmine and phenserine at Axonyx, all of which are currently in Phase II. Arena Pharmaceuticals' T82 is in Phase I, and three candidates in preclinical testing are Pierre Fabre's F3796, FR 152558 at Fujisawa and SPH 1371 from Sanochemia Pharmazeutika.

Reactive oxygen metabolites are believed to trigger and aggravate a number of conditions and accelerate ageing, so antioxidant compounds may well have a beneficial effect in AD patients. Phase III trials are under way at Sigma-Tau of acetyl-L-carnitine, which reduces age-related damage to mitochondrial function in an inner ear model. The New York Institute for Medical Research is evaluating the effects of EGB 761, a standardised extract from Ginkgo biloba that has been shown to improve thought and reasoning. It is believed that its anti-reactive oxygen metabolite activity stimulates electrical signals, and it is already approved in Germany for the treatment of dementia.

Because of the tendency of inflammatory cells to collect around the amyloid plaques, the non-steroidal antiinflammatory drugs are another potentially useful class of medicines for treating AD. The inflammatory cells release substances that damage the neurons, speeding up the development of amyloid, so any of the standard NSAIDs might be expected to reduce the inflammation. Initial trials in the 1990s on NSAIDs including diclofenac and indomethacin showed the drugs slowed down the loss of reasoning power, but their side-effects meant that most of the elderly patients enrolled in the trial pulled out. Ibuprofen is in Phase III trials at AHP, Pharmacia and Bristol-Myers Squibb. But the most likely drugs to be of long term benefit are the selective COX-2 inhibitors, because of the reduced side-effect profile, though early Phase II trial results on Pharmacia's celecoxib are not promising. Rofecoxib is also in Phase II trials at Merck & Co, as is GlaxoSmithKline's experimental COX-2 inhibitor GW253035.

In the longer term, the aim has to be to create therapeutic agents that will either prevent the disease occurring, or reverse its course. One idea is to promote neuronal survival through modulating nerve growth factor, part of the mechanism by which galantamine operates. NGFs are small protein molecules that are essential for the health of nerves, but results using NGF itself are debatable, though attempts to deliver it directly to the brain using gene therapy were reported earlier this year.

Several products are in Phase III trials. Cerebrolysin from Ebewe Pharmaceuticals is an enzyme digest of brain proteins that is believed to protect neurons and also stimulate their growth, leading to better thought processes and function in AD patients. Sanofi-Synthélabo's xaliproden is an antineurodegenerative agent that is also in Phase III. SR 57746 from the same company is in Phase II and has properties similar to natural NGF. And a compound believed to trigger synthesis molecules for nerve growth, lateprinim, is in Phase II at NeoTherapeutics.

At a much earlier stage of investigation are hormones as nerve protectants. Novartis and Neurocrine BioSciences are conducting Phase I trials on DHEA, which has shown to reduced damage to nerves in model, and modify hormone balance in the elderly. A compound related to the thyroid hormone TRH, posatirelin, in Phase I at Dainippon Pharma, is believed to improve acetylcholine processing enzymes in damaged areas, preventing neurons and their connections from dying prematurely. Clinical trials have shown improvement in several performance scales. A third compound in Phase I, Cephalon's CEP-1347, promotes neuron survival in vitro without blocking receptors for cellular messengers, a problem that has been seen with similar drug leads. It preserves AChE levels, and trials have shown it to improve learning and attention, and protect damaged nerves linking to the forebrain.

Perhaps the most promising of the strategies aimed at reversing the progression of AD involves tau tangles and amyloid plaques. Neurofibrillary tangles, which are found within amyloid plaques, are made up of the tau protein, which helps to form and maintain the microtubules that transport nutrients and other substances to the neurons. In AD patients, the tau protein is structurally faulty and clumps together into paired helical filaments. These tangled proteins build up in the cells and upset their normal function, and the quantity present is generally related to the severity of the dementia. Although the role of tau tangles is not yet completely clear as they are also found in a number of other conditions, it is thought that inhibiting enzymes that add phosphate to the tau proteins may improve AD symptoms. No such agents have, as yet, reached the clinical trials stage.

Research into the prevention or dissolution of amyloid plaques are much further advanced, and their importance in the development of AD is clearer. The main component is beta-amyloid peptide, or BAP, which is formed by the cleavage of beta-amyloid precursor protein by secretase enzymes in the neuronal membranes. Most of the fragments are cleared from the body, but the fragment BAP42 is not, and is deposited in between the neurons as amyloid plaques. Inhibitors of beta- and gamma-secretase are looking particularly promising as AD therapies.

A couple of products are in Phase II trials. Colostrinin, isolated from colostrum, is being developed by ReGen Therapeutics. Colostrum is produced prior to milk by mothers in the first few days after childbirth, and is a rich mixture of substances that plays an important role in the development of a baby's immune system. The presence of antibodies in colostrum is believed to play a major part in its action, but it is a complex mixture, including lipids, proteins and peptides, and these compounds are also thought be significant. Colostrinin is a mixture of proline-rich polypeptides isolated from colostrum, and ReGen is investigating its possibilities as a treatment for Alzheimer's. The mechanisms by which it works are unknown, but it seems to make nerve cells more adhesive, throwing up the possibility that it may be able to reattach them, hence reversing the progression of AD, rather than just treating its symptoms. It acts as a free radical scavenger, and may also be able to dissolve amyloid plaques within the brain.

Also in Phase II is clioquinol, which binds to and removes copper from the body. This reduces, or even removes, amyloid deposits in animal studies in mice with overproduced beta-amyloid peptide. Sanofi-Synthélabo's SL-650102 acts on presenelin genes by an unknown mechanism. Bristol-Myers Squibb has a gamma-secretase inhibitor in Phase I trials, and DuPont, Amgen, Merck, Lilly and Elan are looking at other compounds of this type. The 3D structure of beta-secretase has been discovered by Amgen, and this is likely to lead to the design of further molecules to block its activity.