The molecular pathogenesis of Alzheimer’s disease (AD), humankind’s most prevalent dementia, remains unknown; correspondingly, the role of β-amyloid (Aβ) peptide in the cause and progression of AD remains an issue of controversy and dispute – a dispute unresolved by decades of cellular levels studies. AD research needs an explicit atomistic/molecular-level understanding of Aβ and the role it plays in AD. We are endeavouring to achieve this understanding through extensive in silico modelling calculations (at quantum mechanical [DFT] and molecular mechanical levels) to enable “seeing the unseeable” events that unfold during the molecular mechanistic course of this devastating neurodegenerative disease.

Based on preliminary molecular dynamics simulations, we demonstrated that Aβ peptide monomer and oligomers, implicated in AD pathology, electrostatically bind via their cationic HHQK N-terminal segment to neuronal membranes followed by destructive membrane insertion of the C-terminal; we then demonstrated an identical membrane insertion into bacterial membranes, analogous to an immunopeptide. From these computational simulations, the following new molecular-level model of AD emerges: In response to pathogen/damage-associated molecular pattern stimulating events (e.g. infection, pollution, trauma), Aβ is released as a protective peptide exhibiting immunomodulatory and antimicrobial activities. Aβ’s antimicrobial properties (whether or not bacteria are present) result in a misdirected attack upon ‘self’ neurons, arising from electrotopological similarities between neurons and bacteria in anionic charge geometries on outer leaflet membrane macromolecules (gangliosides in neurons; lipopolysaccharides in bacteria), rendering them similarly susceptible to membrane destruction. This constitutes a new mechanistic model of AD recognizing Aβ as a physiologically oligomerizing physiological peptide within a larger autoimmune conceptualization of AD.

Since immune processes are homeostatically regulated, there are endogenous control systems, offering druggable targets against immunopathic disease. To identify an endogenous "anti-AD" molecule, an in silico screen of 1137 brain molecules against Aβ membrane destruction was performed. This screen identified multiple metabolites of tryptophan and arginine as putative endogenous anti-AD agents. An extensive series of QM/MM and DFT calculations were then performed to explicitly model their mode of action in preventing Aβ-mediated neurotoxicity.

Based upon a comprehensive series of molecular modelling calculations (including molecular dynamics simulations, peptide homology modelling, in silico high throughput screening campaigns, QM/MM and DFT docking studies) at molecular mechanics and molecular quantum mechanics levels of theory, new insights into the cause of AD have been deduced. These calculations suggest that AD is caused because the brain’s immune system cannot differentiate neurons from bacteria.