SPREAD OF BETA-AMYLOID ALONG ANATOMICALLY CONNECTED PATHWAYS IN HAPP TRANSGENIC MICE

Background: The prion-like hypothesis predicts disease spread from propagation of pathogenic misfolded proteins along anatomically-connected pathways. The epidemic spreading model (ESM) is a promising method for predicting this propagation, as was recently demonstrated using data from ADNI. It considers there is an As driving region (epicenter) and that interregional deposition patterns are dependent upon a region’s connectivity to this epicenter. The aim of the present study was to further validate ESM in As overexpressingmice using gold-standard, quantitative immunohistochemistry measures. Methods: 3D quantitative IHC volumes from 12 and 19 month-old transgenic mutant human amyloid precursor protein (hAPP) were spatially normalized to an anatomical template (PERMITS , Biospective Inc.), providing extracted mean values of As deposition in 25 gray matter regions, including 14 cortical regions. C onnectivity between the 25 regions was estimated from publicly-available viral tracer studies (http:// connectivity.brain-map.org/). An anatomical effective distances between each of pair regions i and j (Edist i,j) was then calculated as the shortest path distance between them (i.e. the more connected i and j, the lower the Edist i,j value between them). Using an iterative procedure, we then identified the epicenter region i that maximized the non-linear correlation between regional As deposition values and Edist i,j values (j 1⁄4 1 to 25). Results: The validity of an As epidemic spreading-like behavior hypothesis in mouse models was supported by the identification of the posterior cingulate cortex (PCC) as the epicenter region (same epicenter identified previously for the human subjects). The anatomical effective distance to this epicenter predicted As deposition at 12 and 19 months of age, explaining 60% and 61% of the characteristic As patterns, respectively. Conclusions: The observed correlation between connectivity and As deposition pattern supports the networkdriven hypothesis of misfolded proteins propagation. The epicenter region, PCC, is an central node of the default mode network, which, in humans, has been linked to the progression of Alzheimer’s disease. Further modeling analysis will provide us with a better understanding of disease progression, which will significantly advance translational neurodegenerative research.