APOE-ε4 synergizes with sleep disruption to accelerate Aβ deposition and Aβ-associated tau seeding and spreading

Abstract

Alzheimer's disease (AD) is the most common cause of dementia. The APOE-ε4 allele of the apolipoprotein E (APOE) gene is the strongest genetic risk factor for late-onset AD. The APOE genotype modulates the effect of sleep disruption on AD risk, suggesting a possible link between apoE and sleep in AD pathogenesis, which is relatively unexplored. We hypothesized that apoE modifies Aβ deposition and Aβ plaque-associated tau seeding and spreading in the form of neuritic plaque-tau (NP-tau) pathology in response to chronic sleep deprivation (SD) in an apoE isoform-dependent fashion. To test this hypothesis, we used APPPS1 mice expressing human APOE-ε3 or -ε4 with or without AD-tau injection. We found that SD in APPPS1 mice significantly increased Aβ deposition and peri-plaque NP-tau pathology in the presence of APOE4 but not APOE3. SD in APPPS1 mice significantly decreased microglial clustering around plaques and aquaporin-4 (AQP4) polarization around blood vessels in the presence of APOE4 but not APOE3. We also found that sleep-deprived APPPS1:E4 mice injected with AD-tau had significantly altered sleep behaviors compared with APPPS1:E3 mice. These findings suggest that the APOE-ε4 genotype is a critical modifier in the development of AD pathology in response to SD.

Keywords: Alzheimer disease; Lipoproteins; Neurodegeneration; Neuroscience.

Figure 1. SD exacerbates amyloid plaque deposition in the presence of apoE4 but not apoE3.

(A) Schematic of the experimental design. Four-month-old APPPS1:E3 and APPPS1:E4 mice were placed in an automated sleep fragmentation chamber or a normal cage (n = 12–15 per group). (B) Body weight logs under the SD condition. (C) Representative images of anti–Aβ antibody–stained (HJ3.4-biotin) brain sections from APPPS1:E3 and APPPS1:E4 mice from the NS and SD groups. Scale bars: 500 μm. Original magnification, ×1.25 (insets). (D–F) Quantification of the percentage of area covered by Aβ staining in cortex (D), hippocampus (E), and thalamus (F). Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and sleep condition). There was a significant effect of the apoE genotype and sleep condition but not of sex. **P < 0.01 and ***P < 0.001. No other statistical comparisons were significant unless indicated. See also Supplemental Table 1.

Figure 2. SD in APPPS1 mice increases fibrillar plaques in the presence of apoE4 but not apoE3.

(A) Representative images of brain sections stained with X-34 dye, which recognizes only fibrillar plaques. Brain sections were from APPPS1:E3 and APPPS1:E4 mice from the NS and SD groups. Scale bars: 400 μm. Original magnification, ×1.25 (insets). (B–D) Quantification of the percentage of area covered by X-34 staining in cortex (B), hippocampus (C), and thalamus (D) (n = 12–15 per group). Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05, **P < 0.01, and ***P < 0.001. No other statistical comparisons are significant unless indicated. See also Supplemental Table 1.

Figure 3. SD in APPPS1 mice differently affects the astrocyte population, microglia clustering, and dystrophic neurite formation around plaques in an apoE isoform–dependent manner.

(A) Confocal images of GFAP-labeled astrocytes (green) and neuritic dystrophy (BACE1, magenta costained around X34⁺ plaques (blue) in cortex. Scale bars: 40 μm. (B) Confocal images of IBA1-labeled microglia (green) and neuritic dystrophy (BACE1, magenta) costained around X34⁺ plaques (blue) in hippocampus. Scale bars: 40 μm. (C–E) Quantification of the percentage of GFAP⁺ voxels within 15 μm of plaques in cortex (C), hippocampus (D), and thalamus (E) from APPPS1:E3 and APPPS1:E4 mice from the NS and SD groups (n = 12–15 per group). (F–H) Quantification of the number of microglial cells surrounding plaques in cortex (F), hippocampus (G), and thalamus (H) from APPPS1:E3 and APPPS1:E4 mice subjected to the NS or SD condition. (I–K) Quantification of the percentage of BACE1⁺ voxels within 15 μm plaques in cortex (I), hippocampus (J), and thalamus (K) from APPPS1:E3 and APPS1:E4 mice treated with the NS or SD condition. Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Sidak’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05 and **P < 0.01. See also Supplemental Table 1.

Figure 4. SD in APPPS1 mice significantly increases NP-tau seeding and spreading in cortex in the presence of APOE4 but not APOE3.

(A) Schematic of the experimental design. Four-month-old APPPS1:E3 and APPPS1:E4 mice were injected with AD-tau in the hippocampus and overlaying cortex and were placed in sleep fragmentation chambers or normal cages for 8 weeks. All experimental mice were sacrificed at 6 months of age to evaluate tau seeding and spreading (n = 13–15 per group). (B) Body weight logs with AD-tau injection (inj.) during the SD condition. (C and D) Representative images of the ipsi- and contralateral hemisphere stained with AT8⁺ to identify NP-tau pathology in AD-tau–injected APPPS1:E3 (C) and APPPS1:E4 mice (D). Scale bars: 100 μm. Original magnification, ×10 (insets). (E–H) Quantification of the percentage of area covered by AT8⁺ staining in the ipsi- and contralateral cortices (E and F, respectively) and hippocampi (G and H, respectively) of AD-tau–injected APPPS1:E3 and APPPS1:E4 mice. Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05, **P < 0.01, and ***P < 0.001. See also Supplemental Table 1.

Figure 5. NP tau pathology is significantly increased with SD in the presence of apoE4 but not apoE3.

(A and B) Confocal images of cortical AT8⁺ NP-tau (red) around X34⁺ plaques (blue) in the ipsi- and contralateral cortices of AD-tau–injected APPPS1:E3 (A) and APPPS1:E4 (B) mice (n = 13–15 per group). Scale bars: 40 μm. (C–H) Quantification of the percentage AT8⁺ volume within 15 μm plaques in the ipsi- and contralateral cortices (C and D, respectively), hippocampi (E and F, respectively), and thalami (G and H, respectively) from AD-tau–injected APPPS1:E3 and APPPS1:E4 mice. Data are presented as the mean ± the SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05 and **P < 0.01. See also Supplemental Table 1.

Figure 6. SD in AD-tau–injected APPPS1 mice significantly affects microglia clustering and neuritic dystrophy formation in an apoE isoform–dependent manner.

(A and B) Confocal image of IBA1-labeled microglia (green) and BACE-labeled neuritic dystrophy (magenta) costained around X34⁺ plaques (blue) in ipsi- (A) and contralateral (B) cortices from APPPS1:E3 and APPPS1:E4 mice from the NS or SD groups (n = 13–15 per group). Scale bars: 20 μm. (C–H) Quantification of the number of microglia surrounding plaques in the ipsi- and contralateral cortices (C and D, respectively), hippocampi (E and F, respectively), and thalami (G and H, respectively) of AD-tau–injected APPPS1:E3 and APPPS1:E4 mice. (I–N) Quantification of the percentage of BACE1⁺ voxels within 15 μm of plaques in the ipsi- and contralateral cortices (I and J, respectively), hippocampi (K and L, respectively), and thalami (M and N, respectively) of AD-tau–injected APPPS1:E3 and APPPS1:E4 mice. Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05 and **P < 0.01. See also Supplemental Table 1.

Figure 7. SD decreases perivascular polarization of AQP4 in an apoE isoform–dependent manner.

(A and B) Confocal image of CD31-labeled blood vessel (red) and AQP4-labeled water channel AQP4 (magenta) costained with X34⁺ plaques (blue) in cortex from APPPS1:E3 (A) and APPPS1:E4 (B) mice from the NS and SD groups (n = 13–15 per group). Scale bars: 70 μm. Yellow arrows indicate a significant decrease in polarization of AQP4 around blood vessels. (C) Quantification of colocalized AQP4 and CD31 volumes in cortex from APPPS1:E3 and APPPS1:E4 mice. (D) Quantification of volume of CD31⁺ voxels in cortex from APPPS1:E3 and APPPS1:E4 mice. Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Sidak’s multiple-comparison test (sex, apoE genotype, and sleep condition). *P < 0.05 and **P < 0.01. See also Supplemental Table 1.

Figure 8. SD affects AQP4 gene and protein expression in an apoE isoform–dependent manner.

(A) Heatmap analysis of bulk RNA in cortices from APPPS1:E3 and APPPS1:E4 male mice that were subjected to NS or SD, generated by hierarchical gene clustering based on groups (n = 6 per group). (B) Selected heatmap analysis results from each cluster. Heatmap z scores were calculated for each gene and plotted instead of the normalized expression values. (C) Western blot images of AQP4, SNAP25 (neuronal marker, mainly detected in the parenchyma fraction), α-SMA (vascular marker), CD31, and the β-actin compartment from single mouse brain hemispheres (half-cerebral cortex) from APPPS1:E3 and APPPS1:E4 male mice that were subjected to NS or SD (n = 4 per group). V, vessel fraction; P, parenchyma fraction. (D and E) Quantitative analysis of AQP4 expression levels after normalization to β-actin (D) or CD31 (E). Data are presented as the mean ± SEM. Significance was determined by 2-way ANOVA followed by a Tukey’s post hoc test (apoE genotype and sleep condition) (B, D, and E). *P < 0.05. See also Supplemental Table 1.

Figure 9. Aβ deposition and peri-plaque NP-tau pathology significantly affect sleep behaviors in the presence of APOE4 but not APOE3.

(A) Schematic of the experimental design. Sleep-wake recording data were analyzed at 3 time points (14–17, 18–21, and 00–03) to investigate sleep rebound behaviors of APPPS1:E3 and APPPS1:E4 mice after SD (n = 3–6 per group). (B) Representative percentage wake plot in the SD condition. (C and D) Average sleep bout length for each group for 1400–1700, 1800–2100, and 0000–0300 hours for the first week of SD (C) and at week 8 of SD (D). F, female; M, male. (E) Schematic of the experimental design (n = 6–9 per group). (F) Representative percentage wake plot for the NS condition. (G and H) Average sleep percentage for each group for a 24-hour period, the light phase, and the dark phase in APPPS1:E4 male mice (G) and APPPS1:E4 male mice (H). 6M (control): n = 9; 6M (AD-tau inj.): n = 6. Data are presented as the mean ± SEM. Significance was determined by 3-way ANOVA with Šidák’s multiple-comparison test (sex, apoE genotype, and AD-tau injection) or Student’s t test (6M (control) versus 6M (AD-tau inj.)). *P < 0.05, **P < 0.01, and ***P < 0.001. See also Supplemental Table 1.