Abstract
Background: One main pathological hallmark of Alzheimer’s disease (AD) is tau accumulation as neurofibrillary tangles (NFTs) in the brain. Although sandwich enzyme-linked immunosorbent assays (ELISAs) are useful for quantifying tau levels, including those in CSF, plasma and brain, it has not yet been determined which antibody combination is the most appropriate for assessing the neuropathological accumulation of tau in the brain. Methods: We developed several sandwich tau ELISAs by introducing antibodies against several tau epitopes, including from its N-terminal and C-terminal regions, and evaluated tau levels depending on disease stage, brain areas, and other AD-related changes. Results: We observed that tau levels in insoluble brain fraction determined by each ELISAs differ depending on the epitopes of the antibodies: there is a trend that non-AD control samples yield relatively high signals when an antibody against the N-terminal region of tau is used. On the other hand, ELISAs combining two antibodies against the later-middle to C-terminal regions of tau produced substantially increased signals from AD samples, compared to those from non-AD controls. Such ELISAs better distinguish AD and non-AD controls, and the results are more closely associated with Braak NFT stage, Aβ accumulation, and neuroinflammatory markers. In addition, these ELISAs can reflect the pattern of tau spread across brain regions. Conclusions: Tau ELISAs that combine two antibodies against the later-middle to C-terminal regions of tau can better reflect neuropathological tau accumulation, which would enable to evaluate tau accumulation in the brain at a biochemical level.