The amyloid precursor protein (APP) is found widely throughout the body. Little is known about it's function although it is speculated that it may bind to the surface of cells and help them attach to one another. A fault with the processing of the APP in the brain leads to production of a short chain of APP called beta-amyloid. The hypothesis is that the fault may lie within the overproduction of beta-amyloid or the mechanism that removes it from the brain or both. Production of this sticky protein fragment forms clumps called amyloid plaques which triggers the destruction and disruption of the nerve cells in the brain, causing Alzheimer's.
APP is made of up to 771 amino acids and beta-amyloid is produced when betascretase and gammasecretase enzymes break the chain, forming beta-amyloid of 38, 40 or 42 amino acids long. The chain which is 42 amino acids long is chemically stickier and more likely to form plaques. There are three genetic faults, which alter the role of gammasecretase, leading to increased production of beta amyloid 42.
The plaques build up in spaces between nerve cells affecting neurotransmission, and also damage neurones. This in turn causes an inflammatory response as the brain tries to repair itself. It is also thought to cause the formation of neurofibrillary tangles, which are twisted fibres of another protein called tau, that builds up inside cells.
There is lots of evidence that beta-amyloid causes AD and destruction of nerve cells. For example beta-amyloid killed nerve cells cultured in a laboratory. Mice with human Alzheimer's genes inserted in their DNA developed proteins plaques and showed decreases in their memory and learning skills (measured by use of tests such as a water maze which requires memory). Of those mice, the ones given anti-amyloid vaccine developed the condition more slowly. Despite this evidence that amyloid is key to damage in Alzheimer's disease, the question still remains of exactly how the damage done. It seems that there are ingredients that are directly toxic to nerve cells, formed when the first few strands of protein stick together. Most damage is done during early stages, when the clumps are small and thus more mobile, enabling them to affect more nerve cells.

Treatments under development and testing aim to remove beta-amyloid or disrupt it's production from APP in the first instant.
Researchers are trying to develop drugs which prevent production of beta-amyloid from APP, by acting as inhibitors to the enzymes gamma and betasecretase. Several drugs which interfere with gamma-secretase have had some success and reached phase III trials.
Immunotherapy could be used, whereby an active vaccine is used to trigger a process that uses the immune system to clear away deposits or plaques of amyloid in the brain. After animal models were successful, human treatments were developed. In a trial in 2000, a vaccine caused 12 out of 360 to develop serious inflammation in the brain after the immune system overreacted to the treatment. This particular method of vaccine was less successful and had to be halted, however in post mortem examinations it was revealed that participants had fewer plaques than expected. If a safe vaccine can be developed this route of treatment could be very promising.