Imaging and Spectral Characteristics of Amyloid Plaque Autofluorescence in Brain Slices from the APP/PS1 Mouse Model of Alzheimer's Disease

Abstract

Amyloid deposits are one of the hallmark pathological lesions of Alzheimer's disease (AD). They can be visualized by thioflavin-S, silver impregnation, Congo red staining, and immunohistochemical reactions. However, that amyloid deposits generate blue autofluorescence (auto-F) has been ignored. Here, we report that visible light-induced auto-F of senile plaques (SPs) was detected and validated with conventional methods. Brain slices from APP/PS1 (amyloid precursor protein/presenilin 1) transgenic mice were mounted on slides, rinsed, coverslipped and observed for details of the imaging and spectral characteristics of the auto-F of SPs. Then the slices were treated with the above classic methods for comparative validation. We found that the SP auto-F was greatest under blue-violet excitation with a specific emission spectrum, and was much easier, more sensitive, and reliable than the classic methods. Because it does not damage slices, observation of auto-F can be combined with all post-staining techniques in slices and for brain-wide imaging in AD.

Keywords: Alzheimer’s disease; Amyloid deposits; Autofluorescence; Glial activation; Senile plaques; Spectral imaging.

Fig. 1

Autofluorescence (auto-F) detected by visible light in brain slices from APP/PS1 transgenic mice under a conventional fluorescence microscope. A Intense blue auto-F was induced in the hippocampus using U-FUN (UN, BP 360–370, BA 420IF). B, C Only faint auto-F in the hippocampus was excited by U-FBW (BW, BP 460–495, BA 510IF) and U-FGW (GW, BP 530–550, BA 575IF). D Overlapped images showing a blue dense core surrounded by halo-shaped green and red signals. E–H Amyloid plaque auto-F in piriform cortex under visible light at different wavelengths. Scale bars, 50 µm; insets, 10 µm.

Fig. 2

Auto-F visualized under a laser scanning confocal microscope and their excitation spectra in the AD mouse brain sections. A–F The intensity of auto-F was maximum at 405 nm excitation (A); the plaque core showed strong auto-F, whereas the corona was only faintly fluorescent. As the excitation wavelength increased from 488 nm to 552 nm and 638 nm (B–D), the SP auto-F showed smaller dense cores and loose halos. They were very similar when overlaid (E), but were quite different from that at 405 nm excitation (F). G Typical excitation spectra of the auto-F in AD mouse brain sections excited from 700 nm to 1040 nm in 20-nm steps using a two-photon confocal microscope. The emission intensity at 400 nm–500 nm (red) and 500 nm–600 nm (black) bands were obtained. Scale bars, 20 µm.

Fig. 3

Auto-F identified to be senile plaques by thioflavin-S staining. The 3D images of auto-F at 405 nm showed that it was irregularly spherical with a dense core and loose halo, which was then identified to be a senile plaque by Th-S staining from both superficial (B, B’) and core (C, C’) images of the auto-F. Auto-F, autofluorescence; Th-S, thioflavin-S staining; scale bars, 20 µm.

Fig. 4

SP auto-F was not attributable to DNA. A, B The blue SP auto-F was weakened after DAPI staining compared with the unstained section. C, D After DNase I treatment, strong blue SP fluorescence still occurred (C), which was further confirmed by Th-S staining (D), but the fluorescence of the cell nuclei diminished significantly (C). Auto-F, autofluorescence; Th-S, thioflavin-S staining; scale bars, 20 µm.

Fig. 5

SP fluorescence emission spectra under different conditions. The SP auto-F spectrum imaging (A, A’), after DAPI staining (B, B’), after DNase I treatment (C, C’), and after Th-S staining (D, D’). Th-S-stained SP fluorescence excited at 488 nm (E’’) was significantly brighter than that excited at 405 nm (E’) and the auto-F in unstained brain slices (E). Ex 405 nm, excited at 405 nm; Ex 488 nm, excited at 488 nm, scale bars, 20 µm (A’–D’) and 200 µm (E–E’’).

Fig. 6

Relationships between amyloid deposits and reactive microglia and astrocytes. Brain sections were pretreated with formic acid, and then triple-labeled with 6E10, GFAP, and Iba1 for immunohistochemistry. The SP auto-F was diminished after formic acid treatment (A, B). The Iba1- (C), GFAP- (D), and 6E10-positive (E) signals all occurred in the vicinity of SP auto-F. When the images were overlaid, it was evident that the auto-F was in the center, then the 6E10 signal was closely wrapped round with a certain overlap, then surrounding reactive microglia, and outermost were the activated astrocytes (F). FA, formic acid; scale bars, 50 µm.

Fig. 7

Comparison of modified Bielschowsky’s silver staining, Congo red staining, and Th-S staining with auto-F images of SP. SPs had a low signal-to-noise ratio with very high background after modified Bielschowsky’s silver staining (A, D). SPs were much smaller with a low signal-to-noise ratio after Congo red staining (B, E), and were similar to their auto-F images with a high signal-to-noise ratio after Th-S staining (C, F). Auto-F, autofluorescence; Th-S; thioflavin-S staining; scale bars, 100 µm.

Fig. 8

SP auto-F in sections from AD mouse brain with cortical vessel walls, cortex, and pia mater at 405 nm excitation (A, B), and no SP auto-F was detected in control mice (C). The magnified horizontal (A’), coronal (A”), and surrounding (B’) images showed thick plate-like aggregations that were identified with Th-S (D, E). Auto-F, autofluorescence; Th-S, thioflavin-S; scale bars, 200 µm (A–C), 20 µm (A’, A’’, B’, D, E).