DOI: 10.1002/alz.081771 ISSN: 1552-5260

Amyloid β‐dependent tau phosphorylation is triggered by reactive astrocytes in preclinical Alzheimer’s disease

Bruna Bellaver, Guilherme Povala, Pamela C.L. Ferreira, João Pedro Ferrari‐Souza, Douglas Teixeira Leffa, Firoza Z Lussier, Andrea Lessa Benedet, Nicholas J. Ashton, Cécile Tissot, Joseph Therriault, Stijn Servaes, Jenna Stevenson, Nesrine Rahmouni, Oscar L. Lopez, Dana Tudorascu, Victor L Villemagne, Milos D Ikonomovic, Serge Gauthier, Eduardo R Zimmer, Henrik Zetterberg, Kaj Blennow, Howard J Aizenstein, William E Klunk, Beth E. Snitz, Pauline M Maki, Rebecca C Thurston, Ann D. Cohen, Mary Ganguli, Thomas K Karikari, Pedro Rosa‐Neto, Tharick Ali Pascoal
  • Psychiatry and Mental health
  • Cellular and Molecular Neuroscience
  • Geriatrics and Gerontology
  • Neurology (clinical)
  • Developmental Neuroscience
  • Health Policy
  • Epidemiology



A significant percentage of Aß‐positive cognitively unimpaired (CU) individuals do not develop detectable downstream tau pathology and, consequently, cognitive decline.Experimental literature suggest that reactive astrocytes are necessary to unleashing Aß effects in pathological tau phosphorylation.Here we aimed to investigate whether astrocyte reactivity is key to determining the association of Aß burden with early tau phosphorylation in preclinical Alzheimer’s disease (AD).


We assessed 1,016 CU individuals from two research and one population‐based cohort (TRIAD, Pittsburgh and MYHAT) with Aß (plasma or PET), plasma p‐tau and GFAP measures. Individuals were classified as positive (Ast+) or negative (Ast‐) for astrocyte reactivity using a cutoff based on plasma GFAP of younger Aß‐ individuals.Lowess method and linear regressions accounting for age and sex were used to model the trajectories of plasma p‐tau epitopes as a function of Aß burden. Cohen’s d corrected for age and sex was used to estimate effect sizes between groups.


We observed that plasma p‐tau181 levels increased as a function of Aß only in CU Ast+ individuals from Pittsburgh (ß = ‐0.35,t = 3.10,p = 0.003,Fig.1a,b), MYHAT (ß = ‐0.20,t = 2.26,p = 0.026, Fig.1d,e) and TRIAD (ß = 0.46,t = 2.92,p = 0.004;Fig.1g,h) cohorts. A significant interaction between Aß burden and astrocyte reactivity status on plasma p‐tau181 levels was observed in the Pittsburgh (ß = ‐0.29,t = 2.30,p = 0.022;Fig.1b), MYHAT (ß = ‐0.19,t = 2.07,p = 0.038;Fig.1e) and TRIAD (ß = 0.46,t = 2.92,p = 0.004;Fig.1h) cohorts.Cohen’s d analysis revealed that the presence of Aß+ and Ast+ has a large magnitude of effect on tau phosphorylation (Cohen’s d:Pittsburgh = 0.67; MYHAT = 0.69;TRIAD = 0.98;Fig.1c,f,i), whereas Aß+ in the absence of Ast+ presented a negligible effect size. Similar results were observed for plasma p‐tau231 and p‐tau217.Voxel‐wise analysis confirmed that Aß levels in brain regions known to present early Aß accumulation in AD associated with plasma p‐tau181 only in Ast+ individuals (Fig.2).Tau‐PET deposition occurred as a function of Aß burden only in CU Ast+ (Fig.3a), affecting 100% and 62% of the extension of the Braak I and II regions, respectively (Fig.3b)


We observed biomarker evidence across multiple cohorts that the presence of astrocyte reactivity, measured by plasma GFAP, plays a key role in the association of Aß with early tau pathology in preclinical AD.Our results might have implications for the biological definition of preclinical AD and selecting individuals for early preventive clinical trials.

More from our Archive