Axonal degeneration in an Alzheimer mouse model is PS1 gene dose dependent and linked to intraneuronal Aβ accumulation

Abstract

Abnormalities and impairments in axonal transport are suggested to strongly contribute to the pathological alterations underlying AD. The exact mechanisms leading to axonopathy are currently unclear, but it was recently suggested that APP expression itself triggers axonal degeneration. We used APP transgenic mice and crossed them on a hemi- or homozygous PS1 knock-in background (APP/PS1KI). Depending on the mutant PS1 dosage, we demonstrate a clear aggravation in both plaque-associated and plaque-distant axonal degeneration, despite of an unchanged APP expression level. Amyloid-β (Aβ) peptides were found to accumulate in axonal swellings as well as in axons and apical dendrites proximate to neurons accumulating intraneuronal Aβ in their cell bodies. This suggests that Aβ can be transported within neurites thereby contributing to axonal deficits. In addition, diffuse extracellular Aβ deposits were observed in the close vicinity of axonal spheroids accumulating intracellular Aβ, which might be indicative of a local Aβ release from sites of axonal damage.

Keywords: Alzheimer; amyloid; axonal degeneration; axonal transport; axonopathy; intraneuronal Abeta; presenilin; transgenic mice.

Figure 1

Immunostaining of human APP using 22C11 in spinal cords of APP (A), APP/PS1KI^he(B), and APP/PS1KI^ho(C) mice showing comparable numbers of APP-expressing cells in all three genotypes. Western-blot analyses using W0-2 revealed no differences in the amount of human full-length APP among the three genotypes, however, the levels of C99 and Aβ increased with increasing PS1KI gene dosage (D). Scale bar: 50 μm.

Figure 2

Immunostaining of Aβ using 4G8 antibody in 10-month-old APP, APP/PS1KI^he, and APP/PS1KI^ho mice in frontal cortex (A–C), pons (E–G), and spinal cord (I–K) together with respective quantifications of the area covered by Aβ staining (D,H,L). Virtually no Aβ peptide accumulation was detected in the pons and spinal cord of APP mice, whereas about 4% of the frontal cortical area was covered by Aβ staining. Upon introduction of one PS1KI gene (APP/PS1KI^he mice), the accumulation of Aβ was found to rise significantly in all three regions investigated and further approximately doubled with the introduction of an additional PS1KI gene (APP/PS1KI^ho mice), reaching percentile Aβ covered areas of 14, 8, and 6.5% in the frontal cortex, pons, and spinal cord, respectively. Data was analyzed from 6 mice of each genotype and analyzed using One-Way ANOVA followed by t-tests. All error bars represent mean ±s.e.m. ^***P < 0.001; ^**P < 0.01; ^*P < 0.05. Scale bar: 200 μm.

Figure 3

Comparison between APP (22C11) and phosphorylated APP (pAPP668) staining patterns using parallel sections in different brain regions of APP/PS1KI^ho mice of different ages. Whereas 22C11 stains dystrophic neurites as well as cell bodies, pAPP668 exclusively stains axonal compartments in cortex (A,B: 2 m), CA1 (C,D: 2 m), hippocampal mossy fibers (E,F: 2 m; G,H: 10 m; I,J: 14 m). Double-immunofluorescent staining of 22C11 (K) and pAPP668 (L) shows a high co-localization in axonal swellings (M). Arrowheads indicate same plaques (C,D) or axonal swellings (I,J). Scale bars: (A–H): 100 μm; (I,J): 33 μm; (K–M): 20 μm.

Figure 4

Double ABC-immunostaining showing Aβ in blue using 4G8 antibodies together with visualization of fibers in reddish brown using either NF-200 or anti-pT668. The number of plaque-distant dystrophic neurites (black arrowheads) was found to increase with PS1KI gene dosage; examples of stainings from APP, APP/PS1KI^he, APP/PS1KI^ho, as well as PS1KI^ho mice are shown using the NF-200 antibody in the pons (A–D) and anti-pT668 antibody in the spinal cord (E–H). Increasing numbers of NF-200 (I–K) as well as pT668-positive (L–N) plaque-distant dystrophic fibers were found with increasing PS1 gene dosage in both pons and spinal cord with practically no dystrophic fibers in the APP mice. All error bars represent mean ±s.e.m. ^***P < 0.001; ^**P < 0.01; ^*P < 0.05. Scale bar: 100 μm.

Figure 5

Intraneuronal Aβ accumulation (detected by 4G8, arrowheads) in the spinal cord was found to rise proportionally with increasing PS1KI gene dosage in 10-month-old APP (A), APP/PS1KI^he (B), and APP/PS1KI^ho mice (C). Scale bar: 50 μm.

Figure 6

Aβ-positive axonal swellings (arrows) (A–C). Axonal swellings were stained with 4G8 (A,B), Aβ [N] (C) or end-specific antibodies detecting Aβ 40 (D) or Aβ N3pE (E). Double immunostaining of Aβ (4G8, blue) together with either anti-pT668 (F) or NF-200 (G) (reddish brown). Small Aβ deposits (black arrowheads) could occasionally be detected in the vicinity of large dystrophic neurites in both APP/PS1KI^he and APP/PS1KI^ho mice. Scale bar: (A–C): 50 μm; (D,E): 33 μm; (F,G): 12.5 μm.

Figure 7

Confocal images of fibrillar Aβ oligomers and Aβ fibrils (OC antibody, red) accumulating in large dystrophic fibers (YFP, green) of 6-month-old YFP/APP/PS1KI^he mice with DAPI counterstaining confirming the lack of nuclei (blue). Merged images are shown (A,D,G) together with the isolated OC (B,E,H), and DAPI staining (C,F, I). Extracellular accumulation of Aβ (White arrowhead) could be detected near large dystrophic fibers accumulating intracellular Aβ (A–C). The intracellular accumulation of intracellular granules of OC-stained Aβ peptides were observed in many large dystrophic neurites, both plaque-distant (D–F) as well as occasionally in the vicinity of plaques (G–I). (A–C) are generated as a maximum projection of 20 confocal images within the dystrophic fibers, whereas (D–I) are images of a single confocal plane. Scale bars: 10 μm.

Figure 8

Confocal images of intraneuronal Aβ accumulation in cortical neurons of 6-month-old YFP/APP/PS1KI^he mice. Aβ accumulated inside cell bodies at the axon hillock (A,B), as well as in axons (A, arrowhead) and large apical dendrites (B, arrowhead) projecting from the cell body (OC antibody, red). Also, axonal swellings accumulating intracellular Aβ were found directly connected to cortical neurons accumulating abundant amounts of intraneuronal Aβ at the axon hillock (Aβ [N] antibody, red) (C,D: enlargement of C). Scale bars: (A,B,D): 10 μm; (C): 40 μm.

Figure 9

KLC1-positive axonal spheroids (arrows) in 10-month-old APP/PS1KI^he (A) and APP/PS1KI^ho mice (higher magnification of spheroid indicated by arrowhead in inset). RT-PCR analysis of brain mRNA levels of kinesin family members revealed significantly reduced expression of KIF1A (p < 0.05) and KIF5A (p < 0.01) in APP/PS1Ki^ho mice compared to PS1KI^ho control mice (C). All error bars represent mean ±s.e.m. ^**P < 0.01; ^*P < 0.05. Scale bar: (A,B):100 μm