Identification of amyloid plaques in retinas from Alzheimer's patients and noninvasive in vivo optical imaging of retinal plaques in a mouse model

Abstract

Noninvasive monitoring of β-amyloid (Aβ) plaques, the neuropathological hallmarks of Alzheimer's disease (AD), is critical for AD diagnosis and prognosis. Current visualization of Aβ plaques in brains of live patients and animal models is limited in specificity and resolution. The retina as an extension of the brain presents an appealing target for a live, noninvasive optical imaging of AD if disease pathology is manifested there. We identified retinal Aβ plaques in postmortem eyes from AD patients (n=8) and in suspected early stage cases (n=5), consistent with brain pathology and clinical reports; plaques were undetectable in age-matched non-AD individuals (n=5). In APP(SWE)/PS1(∆E9) transgenic mice (AD-Tg; n=18) but not in non-Tg wt mice (n=10), retinal Aβ plaques were detected following systemic administration of curcumin, a safe plaque-labeling fluorochrome. Moreover, retinal plaques were detectable earlier than in the brain and accumulated with disease progression. An immune-based therapy effective in reducing brain plaques, significantly reduced retinal Aβ plaque burden in immunized versus non-immunized AD mice (n=4 mice per group). In live AD-Tg mice (n=24), systemic administration of curcumin allowed noninvasive optical imaging of retinal Aβ plaques in vivo with high resolution and specificity; plaques were undetectable in non-Tg wt mice (n=11). Our discovery of Aβ specific plaques in retinas from AD patients, and the ability to noninvasively detect individual retinal plaques in live AD mice establish the basis for developing high-resolution optical imaging for early AD diagnosis, prognosis assessment and response to therapies.

Keywords: Alzheimer’s disease; Aβ deposit; Aβ plaque; curcumin; fluorescence; human retina; in vivo optical imaging; mild cognitive impairment; spectral classification; vaccination.

Figure 1

Detection of retinal Aβ plaques in AD-Tg mice by ex vivo curcumin labeling. Whole-mount retinas and whole-eye sagittal cryosections from 10-month-old non-Tg (wt) and AD-Tg mice stained with various anti-human Aβ mAbs (12F4, 11A5-B10, 4G8 and 6E10; secondary Ab-Cy5 conjugate; red color) and curcumin (green color). Double-labeled Aβ plaques appear in yellow color. (a–c) Representative images of whole-mount retinas double-stained with curcumin and Aβ₄₂ mAb (12F4; specific to the C-terminal sequence ending at aa 42). (a) No retinal Aβ plaques were detected in wt mice, whereas (b) they were found in AD-Tg mice; here and below, whenever z-axis projection images are presented, the axes ZY and ZX are shown on the top and right side of the image. (c) Specific staining patterns of Aβ₄₂-containing retinal plaques at a higher magnification; (b’–c”) Separate channels for each staining; (d–f) Whole-mount retinas double-stained with anti-human Aβ₄₀ mAb (11A5-B10; specific to the C-terminal sequence ending at aa 40) and curcumin. (d) No detection of Aβ plaques in wt mice. (e) Specific Aβ plaques were found in AD-Tg mice. (f) Higher magnification image demonstrating Aβ₄₀-containing plaques and their specific staining pattern. (e’–f”) separate channels. (g–j) Whole-eye cross-sections stained with anti-Aβ mAbs (4G8 or 6E10), curcumin and DAPI nuclei staining. (g) No evidence for double-positive curcumin and anti-human Aβ plaques in wt mice. (h–j) Curcumin-positive Aβ plaques co-labeled with 4G8 or 6E10 were identified in various retinal layers, including the GCL-ganglion cell layer, IPL-inner plexiform layer, INL-inner plexiform layer, OPL-outer plexiform layer and ONL-outer Nuclear Layer. (i,j) Aβ plaques were also detected in the sclera. White Asterisks indicate DAPI nuclei staining in the GCL: undamaged in wt retinas and deficient in retinas from AD-Tg mice.

Figure 2

Bioavailability of systemically administered curcumin to the mouse eye: accumulation of curcumin-labeled retinal Aβ plaques with disease progression and detection at early pre-symptomatic stage. (a–q) Representative z-axis projection images of whole-mount retinas and brain coronal cryosections from AD-Tg and non-Tg (wt) mice at various ages, following i.v. curcumin injections for 5 days: In 2.5-month-old AD-Tg mouse (a,b) retinal curcumin-labeled in vivo of Aβ plaques (green) were visible, with further (b) validation of Aβ plaque identity in the same retinal location (arrows) by staining ex vivo with anti-human Aβ mAb 4G8 (secondary Ab-Cy5 conjugate). Co-localization of curcumin and Aβ antibody in yellow pseudocolor). Scale bar = 10 µm. (c,d) No plaques were detected in the brain hippocampus and cortex of 2.5-month-old AD-Tg mice. Scale bars = 100 µm. (e–h) 5-month-old AD-Tg mouse: (e) Presence of curcumin-labeled plaques in the retina and (f) co-labeling following 4G8 mAb staining ex vivo. Scale bar = 20 µm; (g,h) detection of plaques in the brain. Scale bars = 50 µm. (i–k) 9-month-old AD-Tg mouse: (i) Multiple plaques in the retina and (j,k) in the brain. Scale bars = 50 µm. (l–n) 17-month-old AD-Tg mouse: (l) Numerous plaques in the retina and (m,n) in the brain. Scale bars =100 µm. (o–q) 9-month-old non-Tg (wt) mouse: (o) No Aβ plaques in the retina nor (p,q) in the brain. Scale bars =100 µm.

Figure 3

Reduced Aβ-plaque burden in retinas from AD-Tg mice following MOG45D-loaded dendritic cells immunization. (a–c) Representative z-axis projection images of whole-mount retinas from (a) PBS-treated control and (b) MOG45D-immunized AD-Tg mice, and from (c) non-Tg (wt) mouse, stained ex vivo with curcumin and anti-human Aβ mAb (4G8; followed by secondary Ab-Cy5 conjugate). (d,e) Following MOG45D-loaded DCs immunization, a significant reduction in mean curcumin-positive plaque number and area was observed in retinas from immunized AD-Tg mice as compared to PBS-treated controls and non-Tg (wt) mice. (f) A significant decrease in total area covered by plaques was detected in brain hippocampus and cortex from the same mice following the immunotherapy. Curcumin staining revealed the same decrease in plaque burden following immunization in the retina and in the brain, while Aβ mAbs confirmed its specificity to Aβ, suggesting that curcumin is a suitable dye for monitoring Aβ plaques. Error bars represent SEM. Asterisks indicate statistical significance: *** P<0.0001; ** P<0.005, analyzed by one-way ANOVA followed by Bonferroni multiple comparison post-test.

Figure 4

Noninvasive in vivo optical imaging of curcumin-labeled Aβ plaques in AD-Tg mice retina. (a–c) Following i.v. administration of curcumin (7.5 mg/kg) for 5 consecutive days, retinas of 8–12 month-old AD-Tg and wt mice were in vivo fluorescently imaged using Micron II rodent retinal microscope. Representative in vivo mouse fundus images: (a) no plaques could be visualized in wt mouse; (a’) higher magnification in grayscale. (b) Retinal curcumin-labeled plaques (green spots) were visible in live AD-Tg mouse; (b’) higher magnification in grayscale. (c,c’) Representative images demonstrate in vivo detection of curcumin-labeled plaques in the retinas of 8 month-old AD-Tg mice, following a single i.v. injection of curcumin (7.5 mg/kg) two hours prior to imaging. (c’) Enlarged image of the selected area from (c) image, demonstrates that this imaging modality enables the identification of individual plaques or plaque clusters at a high spatial resolution. Note that blood vessels appear unstained (dark), possibly due to blood flow in the live mouse. (d) Whole-mount retina prepared from the same in vivo imaged mouse eye as in (c), following perfusion. An additional ex vivo staining with anti-human Aβ₄₂ mAb (12F4; secondary Ab-Cy5 conjugate) further confirmed the specificity of curcumin signals (captured by Micron II) to Aβ plaques. (d’,d”) separate channels for each staining. A similar pattern of curcumin-positive plaques imaged in vivo was identified ex vivo following antibody labeling (indicated by the numbers 1–4).

Figure 5

Identification of Aβ plaques in the human retina of AD patients via a specific curcumin labeling. All stainings of human whole-mount retinas included a Sudan Black B (SBB) pretreatment to eliminate non-specific autofluorescence signals. (a–d) Whole-mount retinas from human AD patients were first immersed with SBB and subsequently stained with curcumin and DAPI; (a,c) no plaques were observed after staining with SBB, whereas (b,d) subsequent staining of the same human retinas with curcumin revealed the presence of Aβ plaques (indicated by arrows; asterisks mark the nuclei of the same tissue location). (c,d) At higher magnification, dark spots of SBB staining are evident, and following curcumin staining a specific Aβ plaque signal is detected in the same retinal location. (e–j) Signal specificity of individual retinal Aβ plaques, single-labeled with curcumin or with anti-Aβ₄₀ (11A5-B10; secondary Ab-Cy5 conjugate), or double-labeled with both, was confirmed by spectral image analysis performed in quadruplicates in a Leica SP5 WLL double-spectral confocal microscope. Regions of interest (ROI) were marked and their corresponding signal intensity was recorded at increasing emission wavelengths from 560 nm to 750 nm to create the spectral curves for Aβ plaques versus background. (e) Representative image of a single curcumin-labeled Aβ plaque in a retinal whole-mount (ROI1) and tissue background (ROI2) captured at excitation/emission wavelengths of 550/605 nm. (f) Spectral analysis curves of individual curcumin-labeled Aβ plaque, at excitation wavelength of 550 nm, as compared to tissue background (dashed line). (g) Representative image of single retinal Aβ plaque (ROI1) and background (ROI2) after staining with Cy5-antibody 11A5-B10 conjugate captured at excitation/emission wavelengths of 640/675 nm. (h) Spectral analysis curves of individual Cy5-antibody-labeled Aβ plaque, at excitation wavelength of 640 nm, as compared to tissue background (dashed line). (i.j) Representative image and spectra curves of retinal Aβ plaque double-labeled with curcumin (ROI1; orange line) and Cy5-antibody 11A5-B10 conjugate (ROI2; purple line), and corresponding background areas (ROI3 and ROI4; dashed lines), at excitation wavelengths of 550 nm (for curcumin spectra) and 640 nm (for Ab-Cy5 conjugate). Peak wavelengths for curcumin and Cy5-antibody captured in the same individual Aβ plaque are distinct and separable; they remain the same as after single stainings. (k–m) Whole-mount retinas from AD patients and normal control stained with curcumin and 4G8; (k,l) Aβ plaques indicated by asterisks show a single-globular compacted morphology. DAPI stains nuclei. (l) Higher magnification image of an extracellular Aβ plaque with compacted morphology. (m) No Aβ plaques were detected in retinas from normal controls.

Figure 6

Characterization of retinal Aβ plaques identified in postmortem retinas of definite AD patients. (a–c) Representative z-axis projection images of whole-mount retinas of (a) normal individuals compared to (b,c) AD patients following curcumin and anti-human Aβ₄₀ mAb 11A5-B10 stainings. (a) No Aβ plaques could be detected in normal control retinas, whereas (b) clearly found in retinas from AD patients. (a’–b”) Separate channels for each staining. (c) At higher magnification, extracellular Aβ plaque with compacted large cluster is indicated by an arrow (intracellular Aβ₄₀ is demarcated by a dotted line). (d–h) Whole-mount retinas from AD patients stained with curcumin and anti-human Aβ₄₂ mAb 12F4. (d’,d”) Separate channels. Note co-localization of curcumin and antibody. (e,f) Higher magnification images of Aβ plaques demonstrated their compacted morphology, consisting of multiple small dense cores connected in larger cluster. (e) Aβ plaques containing lipid deposits indicated by arrows; right bottom image captured in DAPI channel shows dark spots of SBB staining representing lipid deposits associated with retinal Aβ plaque. (g) Aβ plaques stained with curcumin and 12F4 mAb, display either compacted single-globular (asterisk) or cluster (arrow) morphology, both lack notable lipid-associated deposits. (h) Aβ plaque with compacted morphology and associated dark SBB staining spots suggesting the presence of lipid deposits (arrows). (i,j) Immunoperoxidase staining of Aβ plaques labeled with primary mAb 12F4 (plaques are indicated by black arrows) in retinal whole-mount from (i) AD patient and (j) non-AD control. DAB was used as a chromogen. (k–m) Representative retinas from (k) normal individual compared to (l) AD patient stained with ThioS and anti-Aβ mAb 4G8. ThioS- and 4G8 double-positive parenchymal Aβ plaques are found in AD patients’ retinas but not in normal controls. (k’–l”) Separate channels. (m) Higher magnification demonstrating compacted morphology of Thio-S-positive retinal Aβ plaques in AD patients’ retinas. (n,o) Single-labeled Aβ plaques using 4G8 mAb (secondary Ab-Cy5 conjugate) in the retinal innermost layers: Aβ plaques have classical morphology consisting of a central dense-core and radiating fibrillar arms.

Figure 7

Detection of retinal Aβ plaques in suspected early AD patients. (a–f) Representative whole-mount retinas from normal individuals (Non-AD) and from possible/probable AD patients (based on the combined clinical diagnosis and postmortem brain pathology; Table S1), stained with curcumin and anti-Aβ₄₂ mAb (12F4; secondary Ab-Cy5 conjugate). (a) In the retina from cognitively normal individual with no AD pathology, Aβ plaques could not be detected. (b) In cognitively normal individual with mild Aβ-plaque brain pathology (mainly diffused plaques), sparse retinal Aβ plaques were identified. (c,d) In patients with possible AD diagnosis, clusters of Aβ plaques were observed, and (e) in retinas from patients with several years of dementia and postmortem diagnosis of probable AD, Aβ plaques were more common throughout the retina. (f) Abundant Aβ plaques were observed in the retina from severe-stage AD definite patient (based on combined clinical diagnosis and postmortem brain pathology; Table S1). (a’–f”) separate channels for each staining.