In Vivo Amyloid-β Imaging in the APPPS1-21 Transgenic Mouse Model with a (89)Zr-Labeled Monoclonal Antibody

Abstract

Introduction: The accumulation of amyloid-β is a pathological hallmark of Alzheimer's disease and is a target for molecular imaging probes to aid in diagnosis and disease monitoring. This study evaluated the feasibility of using a radiolabeled monoclonal anti-amyloid-β antibody (JRF/AβN/25) to non-invasively assess amyloid-β burden in aged transgenic mice (APPPS1-21) with μPET imaging.

Methods: We investigated the antibody JRF/AβN/25 that binds to full-length Aβ. JRF/AβN/25 was radiolabeled with a [(89)Zr]-desferal chelate and intravenously injected into 12-13 month aged APPPS1-21 mice and their wild-type (WT) controls. Mice underwent in vivo μPET imaging at 2, 4, and 7 days post injection and were sacrificed at the end of each time point to assess brain penetrance, plaque labeling, biodistribution, and tracer stability. To confirm imaging specificity we also evaluated brain uptake of a non-amyloid targeting [(89)Zr]-labeled antibody (trastuzumab) as a negative control, additionally we performed a competitive blocking study with non-radiolabeled Df-Bz-JRF/AβN/25 and finally we assessed the possible confounding effects of blood retention.

Results: Voxel-wise analysis of μPET data demonstrated significant [(89)Zr]-Df-Bz-JRF/AβN/25 retention in APPPS1-21 mice at all time points investigated. With ex vivo measures of radioactivity, significantly higher retention of [(89)Zr]-Df-Bz-JRF/AβN/25 was found at 4 and 7 days pi in APPPS1-21 mice. Despite the observed genotypic differences, comparisons with immunohistochemistry revealed that in vivo plaque labeling was low. Furthermore, pre-treatment with Df-Bz-JRF/AβN/25 only partially blocked [(89)Zr]-Df-Bz-JRF/AβN/25 uptake indicative of a high contribution of non-specific binding.

Conclusion: Amyloid plaques were detected in vivo with a radiolabeled monoclonal anti-amyloid antibody. The low brain penetrance of the antibody in addition to non-specific binding prevented an accurate estimation of plaque burden. However, it should be noted that [(89)Zr]-Df-Bz-JRF/AβN/25 nevertheless demonstrated in vivo binding and strategies to increase brain penetrance would likely achieve better results.

Keywords: 89Zirconium; Alzheimer; amyloid imaging; monoclonal antibody; small animal imaging.

FIGURE 1

μPET imaging demonstrates increased brain uptake of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 in APPPS1–21 mice in comparison to WT. Mice were injected with 133 μg of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 and brain uptake was measured by μPET scanning and quantified as the %ID/g. The graphs show the VOI analysis of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 retention in WT and APPPS1–21 mice at (A) 2 days pi, (B) 4 days pi, and (C) 7 days pi. Significant genotype differences in retention were observed at 4 and 7 days pi. Data is presented as mean + SD. Student’s t-test, ^∗∗∗p < 0.001, ^∗∗p < 0.01, ^∗p < 0.05.

FIGURE 2

Voxel-wise analysis of the μPET data and ex vivo examination of brain tissue confirms higher accumulation of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 in APPPS1–21 mice. Graph (A) shows voxel-wise analysis, data is presented as the percent of each brain region that was significantly changed (increase). This data is illustrated in image (B), the statistical T-maps represent areas of significant increases. After imaging mice were sacrificed and brain radioactivity was measured ex vivo by (C) γ–counting and (D) autoradiography. Data is presented as the average + SD. (E) Representative autoradiograph of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 uptake at 4 days pi in an APPPS1–21 mouse. Student’s t-test ^∗∗∗∗p < 0.0001, ^∗∗p < 0.01.

FIGURE 3

Immunohistochemistry demonstrates that [⁸⁹Zr]-Df-Bz-JRF/AβN/25 labels plaques to a low extent after i.v injection. Sagittal brain slices (20 μm) were cut and the sections were stained for amyloid-β. Images (A–C,G–I) represent in vivo labeling arising from i.v injections of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 and are compared to adjacent slices that were stained for total plaque load (D–F,J–L). (Scale bar = 100 μm). Graph (M) depicts quantification of amyloid-β staining in the cortex, hippocampus and brain stem. Data is shown as the mean + SD.

FIGURE 4

APPPS1–21 mice did not exhibit significant uptake of a non-amyloid targeted antibody in comparison to WT mice. Mice received an i.v injection of 133 μg of [⁸⁹Zr]-trastuzumab and were underwent μPET scanning at 4 days pi. Graph (A) depicts VOI analysis of [⁸⁹Zr]-trastuzumab retention in WT and APPPS1–21 mice, data is presented as mean + SD. After scanning, mice were sacrificed and brain radioactivity of [⁸⁹Zr]-trastuzumab was measured by (B) γ–counting and (C) autoradiography. No significant differences in cerebral [⁸⁹Zr]-trastuzumab retention between genotypes were observed with either in vivo or ex vivo methods. We compared brain retention of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 in APPPS1–21 with and without transcardial perfusion by (D) γ–counting and (E) autoradiography, data is presented as mean + SD. Student’s t-test ^∗p < 0.05. Retention of [⁸⁹Zr]-Df-Bz-JRF/AβN/25 was significantly reduced in perfused WT but not perfused APPPS1–21 mice.

FIGURE 5

Pre-treatment with non-labeled Df-Bz-JRF/AβN/25 only partially blocks [⁸⁹Zr]-Df-Bz-JRF/AβN/25 uptake in APPPS1–21 mice. APPPS1–21 mice were treated with 400 μg of non-labeled Df-Bz-JRF/AβN/25 2 h prior to injection of [⁸⁹Zr]-Df-Bz-JRF/AβN/25. Brain retention was assessed at 4 days pi. Graph (A) depicts VOI analysis of the μPET imaging, data is presented as mean + SD. Graph (B) shows voxel-wise analysis, data is presented as the % of each brain region that was significantly changed (decrease). Image (C) is statistical T-maps depicting regions of significantly reduced tracer retention in pre-treated versus untreated APPPS1–21 mice. Radioactivity was measured ex vivo by (D) γ–counting and (E) autoradiography. Data is presented as the average ± SD. Student’s t-test, ^∗∗∗p < 0.001, ^∗∗p < 0.01.