New applications of diffusion model based prediction of pathological brain alterations: Introducing amyloid‐tau interactions

Abstract

AbstractBackgroundAlzheimer’s disease involves widespread and progressive deposition of misfolded protein tau. There is mounting evidence for “prion‐like” trans‐neuronal transmission, whereby proteins misfold, triggering subsequent misfolding of adjacent same‐species proteins, which spread along white matter projections. It is not well understood how tau and amyloid beta interact, and how their subsequent spread lead to stereotypical progression in the Alzheimer brain.MethodHere we present a major advancement of our previously established Network Diffusion Model (Raj et al. Neuron 2012) that includes three key processes: a) Tau monomer seeding and production in the entorhinal cortex, and Ab monomer seeding in metabolically active areas. b) Interaction between Ab and tau, whereby Ab facilitates production of tau, or both facilitate each other. c) Both species then undergo network spread via our prior Network‐Diffusion model, whereby anatomic connections govern the rate at which two distant but connected brain regions can transfer pathologic species. This extended joint Ab‐tau model was thoroughly tested on empirical group‐averaged regional data (MRI‐derived atrophy, AV45‐PET for Ab and AV1451‐PET for tau) from ADNI‐3 (Table 1).ResultThe extended NDM model exhibits all hallmarks of tau and amyloid progression seen in human patients (Figure 1), accurately capturing the spatial distribution of empirical regional atrophy, Ab and tau (Table 2). Figure 1 shows evolution of the joint model with (middle column) and without amyloid‐facilitation of tau (right). Remarkably, the introduction of a 1‐way directed Aβ to tau influence recapitulated the concept of amyloid‐facilitated tauopathy, and was critical to the models’ success. In comparison, the non‐interacting model was significantly worse, while the 2‐way interaction model (bidirectional influence between tau and Ab) was essentially the same (see Table 2).ConclusionThis in silico exposition of the “pas de deux” of co‐evolving proteins supports a key role for an amyloid‐facilitated‐tau rather than the classic amyloid‐cascade or pure‐tau hypotheses, and helps explain known but poorly understood aspects of AD (Figure 2).