Relationship between amyloid and tau levels and its impact on tau spreading

Abstract

Purpose: Previous studies have shown that Aβ-amyloid (Aβ) likely promotes tau to spread beyond the medial temporal lobe. However, the Aβ levels necessary for tau to spread in the neocortex is still unclear.

Methods: Four hundred sixty-six participants underwent tau imaging with [18F]MK6420 and Aβ imaging with [¹⁸F]NAV4694. Aβ scans were quantified on the Centiloid (CL) scale with a cut-off of 25 CL for abnormal levels of Aβ (A+). Tau scans were quantified in three regions of interest (ROI) (mesial temporal (Me); temporoparietal neocortex (Te); and rest of neocortex (R)) and four mesial temporal region (entorhinal cortex, amygdala, hippocampus, and parahippocampus). Regional tau thresholds were established as the 95%ile of the cognitively unimpaired A- subjects. The prevalence of abnormal tau levels (T+) along the Centiloid continuum was determined.

Results: The plots of prevalence of T+ show earlier and greater increase along the Centiloid continuum in the medial temporal area compared to neocortex. Prevalence of T+ was low but associated with Aβ level between 10 and 40 CL reaching 23% in Me, 15% in Te, and 11% in R. Between 40 and 70 CL, the prevalence of T+ subjects per CL increased fourfold faster and at 70 CL was 64% in Me, 51% in Te, and 37% in R. In cognitively unimpaired, there were no T+ in R below 50 CL. The highest prevalence of T+ were found in the entorhinal cortex, reaching 40% at 40 CL and 80% at 60 CL.

Conclusion: Outside the entorhinal cortex, abnormal levels of cortical tau on PET are rarely found with Aβ below 40 CL. Above 40 CL prevalence of T+ accelerates in all areas. Moderate Aβ levels are required before abnormal neocortical tau becomes detectable.

Keywords: Alzheimer’s disease; Aβ-amyloid imaging; Neurodegeneration; Positron emission tomography; Tau imaging; Tauopathies.

Fig. 1

Representative axial, coronal, and sagittal Aβ and tau PET images across the spectrum of Centiloid values

Fig. 2

Histograms and stacked plot of the percentage of tau-positive subjects among the whole population versus Centiloid scores. Cognitive groups are colour-coded (CU: blue, MCI: green, AD: red)

Fig. 3

Prevalence (percentage) of high-tau (T+) CU subjects among the cognitively unimpaired subjects

Fig. 4

Scatter plot of Centiloid versus the log tau SUVR in the Me (left) and Te (middle) and R (right) ROI in 3 different colour-coded clinical diagnostic groups, with light colours used to identify A- subjects. Thresholds are displayed in fine dash vertical and horizontal lines. Peri-thresholds are in grey shadow and a sigmoid curve has been fitted to the whole population. The thick dash horizontal line shows the CL value at which tau SUVR reaches the threshold for elevated tau.

Fig. 5

Scatter plot of tau SUVR in the Me (left) and Te (middle) and R (right) ROI versus age at tau scan in 2 different colour-coded clinical diagnostic groups (cognitively unimpaired (CU) and cognitively impaired (CI)), with light colours used to identify A- subjects