DOI: 10.1002/alz.081639 ISSN: 1552-5260

APOEε4 facilitates the accumulation and connectivity‐mediated spreading of tau pathology at lower amyloid levels

Anna Steward, Davina Biel, Anna Dewenter, Sebastian Niclas Roemer, Fabian Wanger, Amir Dehsarvi, Diana Svaldi, Ixavier A. Higgins, Matthias Brendel, Sergey Shcherbinin, Michael Ewers, Nicolai Franzmeier
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



Targeting amyloid (Aß) may show highest clinical efficacy when preventing downstream tau spreading and subsequent neurodegeneration. Thus, it’s critical to establish Aß‐PET thresholds at which tau spreading is triggered, which may depend on ApoE4, i.e. a key genetic risk factor for Aß, tau and cognitive decline. Yet, ApoE4’s impact on Aß‐related tau spreading is unclear, which may bias intervention studies. Thus, we assessed whether ApoE4‐carriage facilitates Aß‐associated tau spreading and determined thresholds at which Aß‐related tau trajectories diverge between ApoE4‐carriers vs. non‐carriers. Lastly, we ran simulated trials using tau‐PET endpoints to test whether including ApoE4‐carriers increases the sensitivity to detect treatment effects at lower Aß‐levels.


We included non‐demented ADNI subjects (n = 237), and a replication sample (18F‐AV‐1451‐A05 trial, n = 130). We combined ApoE4‐status, amyloid‐PET and longitudinal flortaucipir tau‐PET alongside resting‐state fMRI‐connectivity templates from n = 42 Aß‐ controls to assess connectivity‐based tau spreading (subject‐specific tau epicentre connectivity ranked regions i.e. Q1‐Q4) together with tau accumulation in Braak‐stage specific regions. We used i) mediation models to test whether ApoE4 effects on Braak‐stage specific tau accumulation were mediated by centiloid and ii) linear models to determine the centiloid by ApoE4 interaction on tau accumulation in Q1‐Q4 and Braak‐stage regions and iii) determined confidence interval‐based cut‐points of ApoE4‐related divergence in tau trajectories. Lastly, we estimated ApoE4‐stratified sample sizes to detect simulated treatments effects on tau accumulation across increasing centiloid windows.


Using bootstrapped mediation we found that centiloid mediated faster ApoE4‐related tau‐PET increase in early Braak‐stage regions (ADNI/A05, BraakIII: B = 0.19/0.33, p<0.001/<0.001), and that ApoE4 facilitates Aß‐related tau spreading (ApoE4*centiloid interaction on tau‐PET increase) in early Braak‐stage and connectivity‐based ROIs. We identified centiloid cut‐points (∼10‐15) where Aß‐related tau accumulation rates diverged between ApoE4 carriers vs. non‐carriers through Braak and connectivity stages (Fig.1), which replicated across samples. Lastly, focusing on ApoE4‐carriers exclusively in simulated clinical trials increased the sensitivity for detecting treatment effects on tau accumulation at lower centiloid levels (Fig.2).


ApoE4 is related to earlier Aß‐related tau accumulation, which may drive faster ApoE4‐related disease progression. Our results are relevant for clinical trial design, illustrating that ApoE4‐carriers may require earlier anti‐Aß intervention to attenuate downstream tau spreading.

More from our Archive