Biomarkers

The link between amyloid (Aβ) and tau accumulation in Alzheimer's disease (AD) is still unknown, hindering therapeutic efforts to attenuate the Aβ-tau axis. Preclinical studies demonstrated that Aβ promotes hyperexcitatory neuronal activity and that tau spreads trans-synaptically in an activity-dependent manner. We recently showed that tau spreads across connected brain regions, and that Aβ-related connectivity increases promote tau spreading (Roemer-Cassiano et al., 2024). Yet, it is unclear whether Aβ-related hyperconnectivity indeed represents hyperexcitatory neuronal activity. To test this, we combined resting-state fMRI, FDG-PET and post-mortem data, to determine whether Aβ promotes neuronal hyperactivity, thereby driving tau spread in AD.

We first assessed the effect Aβ on neuronal hyperactivity with a novel algorithm to estimate the excitatory to inhibitory (E/I) ratio applied to resting-state fMRI in 145 amyloid-negative controls and 441 amyloid-positive subjects across the AD spectrum, who also underwent amyloid-PET. Second, we used glucose metabolism (FDG-PET) as a marker of neuronal activity in 638 amyloid-positive AD spectrum patients, with a subset (n = 215) of them having tau-PET at a later timepoint. Lastly, we analysed post-mortem data of 5 AD patients and 4 controls stained for c-Fos as a marker of ante-mortem neuronal activity.

Resting-state fMRI-based E/I-ratio assessment in Aβ- controls showed biologically plausible stronger inhibition in association cortices (Figure 1A). In AD, we found an association between higher amyloid-PET SUVRs and a higher E/I ratio, consistent across diagnostic groups (Figure 1B-D), indicative of Aβ-associated hyperexcitatory neuronal activity. Second, we found within individuals, that higher regional amyloid-PET was linked to higher FDG-PET (correlation_{amyloid-PET vs. FDG-PET}: 95% CI [0.37,0.40] p-value <0.001), suggesting higher neuronal activity in Aβ-harbouring regions (Figure 2A). Similarly, we found post-mortem elevated neuronal c-Fos expression in AD brain tissue vs. controls, indicating higher ante-mortem neuronal activity (Figure 3G). Finally, we found that amyloid-PET-based prediction of subject-level future tau accumulation is improved when including regional FDG-PET (Figure 2B) and that FDG-PET-assessed hypermetabolism mediates subject-level effects of Aβ on subsequent tau accumulation (Figure 2C).

Aβ promotes an hyper-excitatory shift in neuronal activity that manifests in glucose hypermetabolism which promotes Aβ-related tau accumulation. Thus, Aβ-associated neuronal hyper-excitability is a potential target for attenuating the Ab-tau axis in AD.