Objectives: Autoradiography on brain tissue is used to validate binding targets of newly discovered radiotracers. The purpose of this study was to correlate quantification of autoradiography signal using the novel next-generation tau positron emission tomography (PET) radiotracer [18F]PI-2620 with immunohistochemically determined tau-protein load in both formalin-fixed paraffin-embedded (FFPE) and frozen tissue samples of patients with Alzheimer’s disease (AD) and Progressive Supranuclear Palsy (PSP). Methods: We applied [18F]PI-2620 autoradiography to postmortem cortical brain samples of six patients with AD, five patients with PSP and five healthy controls, respectively. Binding intensity was compared between both tissue types and different disease entities. Autoradiography signal quantification (CWMR = cortex to white matter ratio) was correlated with the immunohistochemically assessed tau load (AT8-staining, %-area) for FFPE and frozen tissue samples in the different disease entities. Results: In AD tissue, relative cortical tracer binding was higher in frozen samples when compared to FFPE samples (CWMRfrozen vs. CWMRFFPE: 2.5-fold, p < 0.001), whereas the opposite was observed in PSP tissue (CWMRfrozen vs. CWMRFFPE: 0.8-fold, p = 0.004). In FFPE samples, [18F]PI-2620 autoradiography tracer binding and immunohistochemical tau load correlated significantly for both PSP (R = 0.641, p < 0.001) and AD tissue (R = 0.435, p = 0.016), indicating a high agreement of relative tracer binding with underlying pathology. In frozen tissue, the correlation between autoradiography and immunohistochemistry was only present in AD (R = 0.417, p = 0.014) but not in PSP tissue (R = −0.115, p = n.s.). Conclusion: Our head-to-head comparison indicates that FFPE samples show superiority over frozen samples for autoradiography assessment of PSP tau pathology by [18F]PI-2620. The [18F]PI-2620 autoradiography signal in FFPE samples reflects AT8 positive tau in samples of both PSP and AD patients.

There are two kinds of protein aggregates associated with Alzheimer’s disease: amyloid-beta and tau aggregates. Protein ligands are molecules with the ability to bind to these pathologicalprotein accumulations, and if the ligands are fluorescent, they can be used to detect the aggregates they’re bound to. The ligands can be selective and only bind to one kind of protein aggregate, or they can be general and bind to both kinds. PI-2620 is a ligand selective for tau aggregates, and in this thesis, three analogues of PI-2620 were synthesized and determined to be functional fluorescent protein ligands. Two of them, designated A1 and B1, were selective for amyloid-beta aggregates, while the third ligand, A2, obtained by ester-hydrolysis of A1, would bind to both tau and amyloid-beta aggregates. This finding suggests that introducing a charge to a ligand lowers its selectivity, since the staining experiments were carried out in a buffered solution at pH 7.4, where ligand A2 might be partially charged.