Microglia activation predicts tau positivity beyond Aβ in Alzheimer's disease. (31st December 2021)

• Record Type: Journal Article
• Authors: Ottoy, Julie; Bezgin, Gleb; Savard, Mélissa; Mathotaarachchi, Sulantha; Pascoal, Tharick A.; Chamoun, Mira; Soucy, Jean‐Paul; Gauthier, Serge; Rosa‐Neto, Pedro
• Is Part Of: Alzheimer's & dementia. Volume 17(2021)Supplement 1
• Page Start: n/a
• Page End: n/a
• Publication Date: 2021-12-31
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.054667
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Deposition of amyloid‐beta (Aβ) plaques and hyperphosphorylated tau are known hallmarks of Alzheimer's disease (AD). In addition, neuroinflammation or microglia activation plays a role in Alzheimer's pathophysiology and may be a potential driver of abnormal tau deposition. This study demonstrated the important brain regions where microglia activation, in addition to Aβ, is driving tau positivity in AD. Method: 120 participants from the TRIAD cohort underwent 18F‐AZD4694 amyloid‐PET, 18F‐MK620 tau‐PET, and 11C‐PBR28 neuroinflammation‐PET. All images were normalized to the ADNI template and quantified as SUVR in 35 regions‐of‐interest using cerebellar gray as the reference region. ANCOVA was used to assess the differences in levels of microglia activation between tau‐positive (T+; n=38) vs. tau‐negative (T‐; n=82), corrected for age, sex, APOE‐e4, education, and global Aβ. Tau status was based on both visual rating with elevated uptake in Braak1and2 and quantitative cut‐off of 1.22. In addition, step‐wise logistic regression was conducted to predict T+ in three steps: 1) the covariates, i.e. age, sex, APOEε4, and education, 2) covariates + global Aβ, and 3) covariates + global Aβ + activated microglia in each of the 35 ROIs, corrected for multiple comparisons using FDR. Our prediction model was trained based on a randomly selected 70%/30% training/test set. In order to investigate the stability of our model, this procedure was repeated 500 times. Result: Levels of microglia activation were significantly increased in T+ compared to T‐ in the posterior cingulate (p=0.0001), middle (p=0.02) and inferior (p=0.0005) temporal, fusiform (p=0.0004), and orbitofrontal cortex (p=0.005). Based on logistic regression and a likelihood‐ratio test between each step, we found that the 'covariates + global Aβ model' performed significantly better in predicting T+ than the 'covariates‐only' model (χ2(1)=72.64, p<0.0001). In the 3‐step model, the addition of activated microglia showed a significantly improvement of the model's performance to predict T+, particularly the activated microglia in the hippocampus (p=0.001), middle (p=0.006) and inferior (p=0.001) temporal, and fusiform (p=0.001) cortex. Receiver operating characteristic curves (ROC) and AUCs are shown in Fig 1. Conclusion: Our study demonstrated that activated microglia in the Braak2 to Braak4 regions play an important role in predicting tau positivity in addition to Aβ.