A systems biology model studying the role of cholesterol in Alzheimer's disease pathogenesis

C. Kyrtsos and J.S. Baras

Proceedings of the 40th Annual Meeting Neuroscience 2010, San Diego, CA, November 13-17, 2010.

Abstract

Objective: To develop a systems-level, metabolic network model to study the effect that alterations in the brain cholesterol concentration levels have on downstream or interconnected sub-networks in an Alzheimer’s disease (AD) model.
Methods: A simplified network encompassing both basic metabolic networks such as glucose and amino acid metabolism, as well as APP processing, cholesterol metabolism, and cholesterol-related metabolic pathways was generated using pathways described by either the KEGG database or in the literature.  A differential equations model describing the distribution of molecules within the network over time was developed and simulated using Matlab.  The model was initialized using healthy concentrations of all pertinent and known quantities.  Individual molecules were represented by nodes in the topological mapping, while reactions between different molecules were designated by weighted links.  The weights of the links were determined stoichiometrically, where possible, or statistically, if data was not available.
Results: Our results are one of the first attempts to study Alzheimer’s disease using a systems biology approach.  This approach is unique in that it allows us to merge the creation of a systems-level quantitative model for the brain metabolic network with data obtained from in vivo experiments to generate the most accurate mathematical model using current knowledge.  The topology of the initially-generated network showed crucial overlap between the APP clearance pathway and cholesterol distribution pathway in adults, specifically with the shared usage of apolipoprotein E (apoE) as a carrier for both beta amyloid (Aβ) cholesterol, and with the LDL receptor-related protein (LRP-1) as a necessary transport protein for Aβ out of the brain and cholesterol into neurons.  Short run simulations of networks with decreased brain cholesterol synthesis showed slight increases in the concentration levels of Aβ.
Conclusion: We have developed a systems-level, pseudo-metabolic network for the basic metabolic pathways in the brain that are pertinent to the development of AD. Analysis of the network structure revealed several important topological homologies between APP and cholesterol processing.  Given our results and the important role that cholesterol plays in maintaining normal neuronal function, it is likely that decreased cholesterol will negatively affect functioning of neurons.  Our future work will use an in vivo mouse model to study the effects of decreased brain cholesterol on AD pathology.