In vivo coupling of tau pathology and cortical thinning in Alzheimer's disease

Abstract

Introduction: The deposition of neurofibrillary tangles in neurodegenerative disorders is associated with neuronal loss on autopsy; however, their in vivo associations with atrophy across the continuum of Alzheimer's disease (AD) remain unclear.

Methods: We estimated cortical thickness, tau ([¹⁸F]-AV-1451), and amyloid β (Aβ) status ([¹¹C]-PiB) in 47 subjects who were stratified into Aβ- (14 healthy controls and six mild cognitive impairment-Aβ-) and Aβ+ (14 mild cognitive impairment-Aβ+ and 13 AD) groups.

Results: Compared with the Aβ- group, tau was increased in widespread regions whereas cortical thinning was restricted to the temporal cortices. Increased tau binding was associated with cortical thinning in each Aβ group. Locally, regional tau was associated with temporoparietal atrophy.

Discussion: These findings position tau as a promising therapeutic target. Further studies are needed to elucidate the casual relationships between tau pathology and trajectories of atrophy in AD.

Keywords: Alzheimer's disease; Amyloid; Atrophy; Cortical thickness; MRI; Positron emission tomography; Tau.

Fig. 1

Between-group comparisons of mean cortical thickness and [¹⁸F]-AV-1451 burden. Student's t tests revealed no significant differences in mean cortical thickness between Aβ groups, although tau accumulation was significantly increased in the Aβ+ group (P < .001). Abbreviation: Aβ, amyloid β.

Fig. 2

Group comparisons of regional cortical thickness and tau accumulation between both Aβ subgroups. (Top and middle rows) Relative to the Aβ− group, the magnitude and spatial extent of tau accumulation (red, FDR P < .05) were in excess of trend-level cortical thinning (cyan, P < .05). (Bottom row) The spatial overlap in the distributions of cortical thinning and tau accumulation is visually apparent when the contrast maps are superimposed on each other. Abbreviations: Aβ, amyloid β; FDR, false discovery rate.

Fig. 3

The cortical topography of tau pathology overlaps with reduced cortical thickness in both Aβ subgroups. Mixed effect models indicated significant and negative associations between [18F]-AV-1451 binding and cortical thickness in both groups (left: Aβ−: β = −0.5, SE = 0.03, T = −14.6, P < .001; right: Aβ+: β = −0.3, SE = 0.02, T = −14.8; P < .001). The scatterplots depict individual data points of the adjusted [¹⁸F]-AV-1451 BP_ND and cortical thickness data across the subjects (i.e., data adjusted for age, gender, and scan interval days between PET and MRI). Abbreviations: Aβ, amyloid β; BP_ND, nondisplaceable binding potential; HC, healthy controls; MCI, mild cognitive impairment; MRI, magnetic resonance imaging; PET, positron emission tomography; SE, standard error.

Fig. 4

The topography of tau-associated brain atrophy. (Top row) Significant local correlations between tau and cortical thickness are overlaid on the cortical surface as parcellated by the Desikan-Killiany atlas (FDR P < .05, data adjusted for age, gender, and scan interval between PET and MRI). The color gradient represents the strength of the negative correlations, increasing in magnitude from dark blue to cyan. (Bottom row) Box plots of correlation coefficients across the major cortical lobes. The degree of local associations was significantly stronger in the temporal and parietal lobes compared with the frontal lobe (Post hoc Tukey-HSD, P < .05). Abbreviations: FDR, false discovery rate; MRI, magnetic resonance imaging; PET, positron emission tomography.

Fig. 5

Delineating the local and distributed associations of tau in inferior temporal cortex and cortical thinning. (A) Mean PET signal was extracted from bilateral inferior temporal cortex for each subject and their associations with cortical thickness were assessed with Spearman correlations due to non-normality of the inferior temporal tau ROI. (B) Significant local correlations surviving FDR correction are overlaid on the cortical surface, as parcellated by the Desikan-Killiany atlas. The color gradient represents the strength of the negative correlations, increasing in magnitude from dark blue to cyan. Abbreviations: FDR, false discovery rate; PET, positron emission tomography.