Tau spreads across connected brain regions in progressive supranuclear palsy and corticobasal syndrome

Abstract

AbstractBackgroundCortico‐basal degeneration (CBD) and progressive supranuclear palsy (PSP) are 4R‐tauopathies characterized by progressive tau pathology spread that typically starts in the subcortex. Pre‐clinical studies suggest that tau spreads across connected neurons in an activity‐dependent manner, indicating that the brains’ connectome may mediate tau spreading. Supporting this, we found in Alzheimer’s disease that PET‐assessed cortical tau spreads across functionally connected regions. Here, we assessed whether connectivity mediates cortical/subcortical tau spreading also in 4R‐tauopathies, by combining resting‐state fMRI connectivity with i) 2nd generation in vivo tau‐PET (PI2620) in CBS and PSP and ii) post‐mortem tau assessments in PSP.MethodWe assessed PI2620 tau‐PET in 24 CBS‐patients, 28 PSP‐patients and 15 healthy controls. Voxel‐wise tau‐PET differences were determined between patients vs. controls and mean tau‐PET SUVRs were assessed for 232 cortical/subcortical‐ROIs (Fig. 1A). Semi‐quantitative post‐mortem AT8‐stained neuronal tau was assessed in two additional, non‐overlapping PSP samples (Munich: n=96; & UPENN: n=97). Neuropathological ROIs were reconstructed in MRI‐standard‐space (Figs. 1B&C). Functional connectivity was assessed between tau‐PET‐ROIs and neuropathological‐ROIs using out‐of‐sample resting‐state fMRI from 69 elderly amyloid‐ and tau‐negative controls. Using linear regression, we assessed the association between ROI‐to‐ROI connectivity and covariance in tau‐PET or post‐mortem tau levels in spatially corresponding ROI pairs. For tau‐PET, we further tested at the subject level, whether functional connectivity of tau epicenters (i.e. ROIs with highest tau‐PET in a given subject) predicted tau deposition in connected regions, using linear mixed models controlling for age, gender and random intercept.ResultPSP and CBS patients showed elevated tau‐PET compared to controls (Fig. 2). Higher functional connectivity was associated with higher covariance in tau‐PET in PSP and CBS (b=0.402‐0.715, Fig. 3A‐D, all p<0.001). For post‐mortem data, higher ROI‐to‐ROI functional connectivity was also associated higher covariance in tau in both PSP samples (b=0.468&0.765, p<0.001, Fig. 3E&F). Using tau‐PET we found further that connectivity patterns of subject‐level subcortical tau epicenters was associated with subject‐level tau‐PET uptake in connected regions (Fig. 4B‐E, all p<0.001).ConclusionHighly functionally connected brain regions share similar tau levels in 4R‐tauopathies as indicated by tau‐PET and post‐mortem assessments, suggesting that brain connectivity is associated with inter‐regional tau spreading in 4R‐tauopathies.