Comparison of the Amyloid Load in the Brains of Two Transgenic Alzheimer’s Disease Mouse Models Quantified by Florbetaben Positron Emission Tomography

Antje Willuweit, Michael Schöneck, Sarah Schemmert, Philipp Lohmann, Saskia Bremen, Dominik Honold, Nicole Burda, Nan Jiang, Simone Beer, Johannes Ermert, Dieter Willbold, N. Jon Shah, Karl‐Josef LangenFrontiers in Neuroscience 2021 · DOI

TL;DR

This study compares two transgenic mouse models for Alzheimer's disease (AD) using florbetaben (FBB) PET imaging to assess amyloid plaque burden. The ARTE10 model demonstrated superior FBB retention and progressive amyloid accumulation with age, making it a more suitable model for FBB PET imaging compared to the APPswe/PS1ΔE9 model.

What Problem Does This Paper Address?

Alzheimer's disease (AD) is characterized by the accumulation of amyloid plaques and neurofibrillary tangles in the brain. Transgenic mouse models are crucial for studying AD pathogenesis and testing new therapies. However, not all mouse models are equally suitable for imaging techniques like Positron Emission Tomography (PET), which is used to visualize amyloid deposits. Previous studies have shown inconsistent results when using amyloid PET tracers in different AD mouse models, leading to uncertainty about which models best recapitulate human AD pathology for imaging studies. This paper aims to address this by directly comparing two commonly used AD mouse models, ARTE10 and APPswe/PS1ΔE9, using florbetaben (FBB), an 18F-labeled PET tracer approved for human use, to determine their suitability for amyloid PET imaging.

Methods

The researchers employed a multi-modal approach to characterize amyloid pathology and its correlation with PET imaging.

  • In Vivo PET Imaging: Small animal PET/CT scans were performed on ARTE10 mice at different ages (9, 15, and 21 months) and on older APPswe/PS1ΔE9 mice (24 months), along with age-matched wild-type controls. Florbetaben (FBB) was injected intravenously, and tracer uptake was measured over 120 minutes.
  • Ex Vivo Autoradiography: Immediately after PET/CT scans, brains were harvested, sectioned, and analyzed using ex vivo autoradiography to provide higher-resolution images of FBB distribution.
  • Histological Analysis: Brain sections were stained with Congo red, a dye that specifically binds to amyloid fibrils, to quantify plaque load, plaque count, and average plaque size. This provided a gold standard for assessing amyloid pathology.
  • Image Analysis: Standardized uptake value ratios (SUVRs) were calculated using cerebellum or brain stem as reference regions to quantify FBB retention. Correlation analyses (Pearson correlation) were performed to assess the relationship between PET/autoradiography data and histological findings.

Datasets

Not applicable. This study generated its own data through experimental procedures on animal models.

Key Insights and Ideas

  • ARTE10 Model is Superior for FBB PET: The ARTE10 mouse model showed a clear, progressive increase in FBB retention with age, which strongly correlated with increasing amyloid plaque burden. This makes it a more suitable model for studying amyloid progression and evaluating PET tracers.
  • APPswe/PS1ΔE9 Model Shows Lower FBB Signal: While APPswe/PS1ΔE9 mice did show some amyloid pathology and could be differentiated from wild-type controls, their FBB retention was significantly lower, comparable to much younger ARTE10 mice.
  • Congophilic Dense-Cored Plaques are Key: The study strongly suggests that the amount and type of amyloid plaques are critical for successful FBB PET imaging. Specifically, the presence of congophilic, dense-cored plaques in the ARTE10 model, as opposed to the more diffuse plaques in APPswe/PS1ΔE9, appears to be the decisive factor for high FBB tracer binding.
  • Validation is Crucial: The excellent correlation (r=0.947) between in vivo PET and ex vivo autoradiography highlights the reliability of the PET imaging protocol and the importance of cross-validation with higher-resolution techniques.
  • Reference Region Choice Matters: The study emphasizes the importance of selecting an appropriate reference region for PET analysis, as different mouse models may have varying levels of amyloid deposition in potential reference areas like the cerebellum.

Key Results

  • Progressive Amyloid in ARTE10: FBB retention in ARTE10 mice increased significantly with age in the cortex and hippocampus, detectable by both in vivo PET and ex vivo autoradiography.
  • Early Detection: FBB PET detected significant differences between ARTE10 mice and wild-type controls as early as 9 months of age.
  • High Correlation: A strong correlation (Pearson r = 0.947, p < 0.0001) was found between in vivo FBB PET data and ex vivo autoradiography.
  • APPswe/PS1ΔE9 Differentiation: Despite lower FBB signal, APPswe/PS1ΔE9 mice could still be statistically differentiated from wild-type controls (p < 0.01).
  • Histology Correlation: FBB retention correlated best with total plaque load and the number of congophilic plaques in both mouse models.
  • ARTE10 vs. APPswe/PS1ΔE9: FBB retention in the cortex and hippocampus of 24-month-old APPswe/PS1ΔE9 mice was comparable to that of 9-month-old ARTE10 mice.

Limitations and Future Work

  • Partial Volume Effect: The authors acknowledge that partial volume effects, particularly in smaller brain regions like the hippocampus, might lead to an underestimation of PET signal compared to autoradiography. They suggest that partial volume correction procedures could further refine the analysis.
  • Asymmetry: While not a major issue in this study due to strong correlations, the potential for amyloid load asymmetry between brain hemispheres is mentioned as a factor to consider.
  • Reference Region Selection: The study highlights the challenge of selecting a universal reference region when comparing different mouse models, suggesting that careful consideration and validation are necessary.
  • Future Work: The authors implicitly suggest that the ARTE10 model, due to its robust and progressive amyloid pathology detectable by FBB PET, is an ideal candidate for future preclinical studies investigating AD therapies and for validating new amyloid PET tracers. Further research could explore the impact of different plaque morphologies on tracer binding and investigate the use of advanced PET analysis techniques like partial volume correction.

Summary updated

Abstract

Alzheimer's disease (AD) is characterized by formation of amyloid plaques and neurofibrillary tangles in the brain, which can be mimicked by transgenic mouse models. Here, we report on the characterization of amyloid load in the brains of two transgenic amyloidosis models using positron emission tomography (PET) with florbetaben (FBB), an <sup>18</sup>F-labeled amyloid PET tracer routinely used in AD patients. Young, middle-aged, and old homozygous APP/PS1 mice (ARTE10), old hemizygous APPswe/PS1ΔE9, and old wild-type control mice were subjected to FBB PET using a small animal PET/computed tomography scanner. After PET, brains were excised, and <i>ex vivo</i> autoradiography was performed. Plaque pathology was verified on brain sections with histological methods. Amyloid plaque load increased progressively with age in the cortex and hippocampus of ARTE10 mice, which could be detected with both <i>in vivo</i> FBB PET and <i>ex vivo</i> autoradiography. FBB retention showed significant differences to wild-type controls already at 9 months of age by both <i>in vivo</i> and <i>ex vivo</i> analyses. An excellent correlation between data derived from PET and autoradiography could be obtained (<i>r</i> <sub><i>Pearson</i></sub> = 0.947, <i>p</i> < 0.0001). Although amyloid load detected by FBB in the brains of old APPswe/PS1ΔE9 mice was as low as values obtained with young ARTE10 mice, statistically significant discrimination to wild-type animals was reached (<i>p</i> < 0.01). In comparison to amyloid burden quantified by histological analysis, FBB retention correlated best with total plaque load and number of congophilic plaques in the brains of both mouse models. In conclusion, the homozygous ARTE10 mouse model showed superior properties over APPswe/PS1ΔE9 mice for FBB small animal amyloid PET imaging. The absolute amount of congophilic dense-cored plaques seems to be the decisive factor for feasibility of amyloidosis models for amyloid PET analysis.
BibTeX
@inproceedings{willuweit2021comparison,
  author = {Antje Willuweit and Michael Schöneck and Sarah Schemmert and Philipp Lohmann and Saskia Bremen and Dominik Honold and Nicole Burda and Nan Jiang and Simone Beer and Johannes Ermert and Dieter Willbold and N. Jon Shah and Karl‐Josef Langen},
  title = {{Comparison of the Amyloid Load in the Brains of Two Transgenic Alzheimer’s Disease Mouse Models Quantified by Florbetaben Positron Emission Tomography}},
  booktitle = {Frontiers in Neuroscience},
  year = {2021},
  doi = {10.3389/fnins.2021.699926},
}
Figure