Structure and location of amyloid beta peptide chains and arrays in Alzheimer's disease: new findings require reevaluation of the amyloid hypothesis and of tests of the hypothesis.

New in situ high resolution electronmicroscopic examination of amyloid fibrils in situ indicate that in Alzheimer's disease these fibrils are not simply long chains of self aggregated amyloid beta peptide. The amyloid beta is not only associated with P protein and glycans, as was well known from previous immunohistologic studies, but is arranged in the form of short chains at right angles to a P protein backbone with the glycans wrapped around that backbone. These findings suggest that the hypothesis causally relating simple, fibrillar amyloid beta to Alzheimer's disease must be reevaluated since such simple fibrils may be absent, or not the major form of the amyloid beta in the brain. Other data shows that shorter multimers, so-called protofibrils, or dimers of amyloid beta or molecules cleaved from it can be highly toxic. Some of these may be in the soluble amyloid beta fraction. Shorter multimers or dimers of amyloid beta, either extra or intracellular, may be the real links between amyloid beta production and Alzheimer's disease. Toxicity studies employing fibrillar amyloid beta may not be relevant, even if they produce lesions, because they do not employ amyloid beta in the form in which it actually exists in the Alzheimer brain. Studies of treatments designed to remove fibrils or to prevent their formation may be ineffective or suboptimal in effectiveness because they do not reduce the relevant amyloid burden and/or fail to alter the arrangement of shorter multimers of amyloid beta around its P-protein and glycan core.