Imaging of neurofibrillary pathology in the brain helps in diagnosing dementia, tracking disease progression, and evaluating the therapeutic efficacy of antidementia drugs. The radiotracers used in this imaging must be highly sensitive and specific for tau protein fibrils in the human brain. We developed a novel tau PET tracer, 18 F-THK5351, through compound optimization of arylquinoline derivatives. Methods: The in vitro binding properties, pharmacokinetics, and safety of 18 F-THK5351 were investigated, and a clinical study on Alzheimer disease (AD) patients was performed. Results: 18 F-THK5351 demonstrated higher binding affinity for hippocampal homogenates from AD brains and faster dissociation from whitematter tissue than did 18 F-THK5117. The THK5351 binding amount correlated with the amount of tau deposits in human brain samples. Autoradiography of brain sections revealed that THK5351 bound to neurofibrillary tangles selectively and with a higher signal-tobackground ratio than did THK5117. THK5351 exhibited favorable pharmacokinetics and no defluorination in mice. In first-in-human PET studies in AD patients, 18 F-THK5351 demonstrated faster kinetics, higher contrast, and lower retention in subcortical white matter than 18 F-THK5117. Conclusion: 18 F-THK5351 is a useful PET tracer for the early detection of neurofibrillary pathology in AD patients.

Neurofibrillary tangles in Alzheimer disease (AD) brains are composed of the microtubule-associated protein tau. Noninvasive monitoring of tau protein aggregates in the living brain will provide useful information regarding tau pathophysiology in AD. However, no PET probes are currently available for selective detection of tau pathology in AD. We have previously reported 18 F-labeled THK-523 ( 18 F-6-(2-fluoroethoxy)-2-(4-aminophenyl)quinoline) as a tau imaging radiotracer candidate for PET. After compound optimization, we developed novel 18 F-labeled arylquinoline derivatives, 18 F-THK-5105 and 18 F-THK-5117, for use as tau imaging PET tracers. Methods: 18 F-labeled compounds were prepared from the corresponding tosylated precursors. The binding affinity of compounds to synthetic tau aggregates and tau-rich AD brain homogenates was determined by saturation and competition binding assays. The binding selectivity of compounds to tau pathology was evaluated by autoradiography of AD brain sections. The pharmacokinetics of compounds were assessed in biodistribution studies in normal mice. A 14-d toxicity study with intravenous administration of compounds was performed using rats and mice. Results: In vitro binding assays demonstrated higher binding affinity of THK-5105 and THK-5117 than THK-523 to tau protein aggregates and tau-rich AD brain homogenates. Autoradiographic analyses of AD brain sections showed that these radiotracers preferentially bound to neurofibrillary tangles and neuropil threads, which colocalized with Gallyas-positive and immunoreactive tau protein deposits. The distribution of this radiotracer binding in AD brain sections was completely different from that of 11 C-Pittsburgh compound B, showing preferential binding to amyloid plaques. Furthermore, these derivatives demonstrated abundant initial brain uptake and faster clearance in normal mice than 18 F-THK-523 and other reported 18 F-labeled radiotracers. THK-5105 and THK-5117 showed no toxic effects related to the administration of these compounds in mice and rats and no significant binding for various neuroreceptors, ion channels, and transporters at 1-mM concentrations. Conclusion: 18 Flabeled THK-5105 and THK-5117 are promising candidates as PET tau imaging radiotracers.

Clinical PET studies using F-THK5351 have demonstrated significant tracer retention in sites susceptible to tau burden in Alzheimer disease (AD). However, the in vivo PET signal to reflect tau aggregates remains controversial. We examined the spatial pattern of tracer binding, amyloid-β, tau, and gliosis in an autopsy-confirmed AD patient who underwent F-THK5351 andC-Pittsburgh compound B PET before death. Regional in vivoF-THK5351 retention was significantly correlated with the density of tau aggregates in the neocortex and monoamine oxidase-B in the whole brain, but not correlated with that of insoluble amyloid-β. Furthermore, significant association was observed between the density of tau aggregates, monoamine oxidase-B, and glial fibrillary acidic protein, suggesting that neocortical tau would strongly influence the formation of reactive astrocytes. F-THK5351 PET may have limited utility as a biomarker of tau pathology in AD; however, it could be used to monitor the neuroinflammatory processes in the living brain.

These findings suggest that [(18)F]THK-5117 provides regional information on neurofibrillary pathology in living subjects.