Assessment of synaptic loss in mouse models of β-amyloid and tau pathology using [¹⁸F]UCB-H PET imaging

Objective: In preclinical research, the use of [¹⁸F]Fluorodesoxyglucose (FDG) as a biomarker for neurodegeneration may induce bias due to enhanced glucose uptake by immune cells. In this study, we sought to investigate synaptic vesicle glycoprotein 2A (SV2A) PET with [¹⁸F]UCB-H as an alternative preclinical biomarker for neurodegenerative processes in two mouse models representing the pathological hallmarks of Alzheimer's disease (AD).

Methods: A total of 29 PS2APP, 20 P301S and 12 wild-type mice aged 4.4 to 19.8 months received a dynamic [¹⁸F]UCB-H SV2A-PET scan (14.7 ± 1.5 MBq) 0-60 min post injection. Quantification of tracer uptake in cortical, cerebellar and brainstem target regions was implemented by calculating relative volumes of distribution (V_T) from an image-derived-input-function (IDIF). [¹⁸F]UCB-H binding was compared across all target regions between transgenic and wild-type mice. Additional static scans were performed in a subset of mice to compare [¹⁸F]FDG and [¹⁸F]GE180 (18 kDa translocator protein tracer as a surrogate for microglial activation) standardized uptake values (SUV) with [¹⁸F]UCB-H binding at different ages. Following the final scan, a subset of mouse brains was immunohistochemically stained with synaptic markers for gold standard validation of the PET results.

Results: [¹⁸F]UCB-H binding in all target regions was significantly reduced in 8-months old P301S transgenic mice when compared to wild-type controls (temporal lobe: p = 0.014; cerebellum: p = 0.0018; brainstem: p = 0.0014). Significantly lower SV2A tracer uptake was also observed in 13-months (temporal lobe: p = 0.0080; cerebellum: p = 0.006) and 19-months old (temporal lobe: p = 0.0042; cerebellum: p = 0.011) PS2APP transgenic versus wild-type mice, whereas the brainstem revealed no significantly altered [¹⁸F]UCB-H binding. Immunohistochemical analyses of post-mortem mouse brain tissue confirmed the SV2A PET findings. Correlational analyses of [¹⁸F]UCB-H and [¹⁸F]FDG using Pearson's correlation coefficient revealed a significant negative association in the PS2APP mouse model (R = -0.26, p = 0.018). Exploratory analyses further stressed microglial activation as a potential reason for this inverse relationship, since [¹⁸F]FDG and [¹⁸F]GE180 quantification were positively correlated in this cohort (R = 0.36, p = 0.0076).

Conclusion: [¹⁸F]UCB-H reliably depicts progressive synaptic loss in PS2APP and P301S transgenic mice, potentially qualifying as a more reliable alternative to [¹⁸F]FDG as a biomarker for assessment of neurodegeneration in preclinical research.