Abstract
Alzheimer's disease (AD) is a complex, multifactorial disease involving many pathological mechanisms. Nonetheless, single pathogenic mutations in amyloid precursor protein (APP) or presenilin 1 or 2 can cause AD with almost all of the clinical and neuropathological features, and therefore, we believe an important mechanism of pathogenesis in AD could be revealed from examining pathogenic APP missense mutations. A comprehensive review of the literature, including clinical, neuropathological, cellular and animal model data, was conducted through PubMed and the databases of Alzforum mutations, HGMD, UniProt, and AD&FTDMDB. Pearson correlation analysis combining the clinical and neuropathological data and aspects of mutant APP processing in cellular models was performed. We find that an increase in Aβ42 has a significant positive correlation with the appearance of neurofibrillary tangles (NFTs) and tends to cause an earlier age of AD onset, while an increase in Aβ40 significantly increases the age at death. The increase in the α-carboxyl terminal fragment (CTF) has a significantly negative correlation with the age of AD onset, and β-CTF has a similar effect without statistical significance. Animal models show that intracellular Aβ is critical for memory defects. Based on these results and the fact that amyloid plaque burden correlates much less well with cognitive impairment than do NFT counts, we propose a "snowball hypothesis": the accumulation of intraneuronal NFTs caused by extracellular Aβ42 and the increase in intraneuronal APP proteolytic products (CTFs and Aβs) could cause cellular organelle stress that leads to neurodegeneration in AD, which then resembles the formation of abnormal protein "snowballs" both inside and outside of neurons.