Norepinephrine Drives Sleep Fragmentation Activation of Asparagine Endopeptidase, Locus Ceruleus Degeneration, and Hippocampal Amyloid-β42 Accumulation

Abstract

Chronic sleep disruption (CSD), from insufficient or fragmented sleep and is an important risk factor for Alzheimer's disease (AD). Underlying mechanisms are not understood. CSD in mice results in degeneration of locus ceruleus neurons (LCn) and CA1 hippocampal neurons and increases hippocampal amyloid-β₄₂ (Aβ₄₂), entorhinal cortex (EC) tau phosphorylation (p-tau), and glial reactivity. LCn injury is increasingly implicated in AD pathogenesis. CSD increases NE turnover in LCn, and LCn norepinephrine (NE) metabolism activates asparagine endopeptidase (AEP), an enzyme known to cleave amyloid precursor protein (APP) and tau into neurotoxic fragments. We hypothesized that CSD would activate LCn AEP in an NE-dependent manner to induce LCn and hippocampal injury. Here, we studied LCn, hippocampal, and EC responses to CSD in mice deficient in NE [dopamine β-hydroxylase (Dbh)^-/-] and control male and female mice, using a model of chronic fragmentation of sleep (CFS). Sleep was equally fragmented in Dbh ^-/- and control male and female mice, yet only Dbh ^-/- mice conferred resistance to CFS loss of LCn, LCn p-tau, and LCn AEP upregulation and activation as evidenced by an increase in AEP-cleaved APP and tau fragments. Absence of NE also prevented a CFS increase in hippocampal AEP-APP and Aβ₄₂ but did not prevent CFS-increased AEP-tau and p-tau in the EC. Collectively, this work demonstrates AEP activation by CFS, establishes key roles for NE in both CFS degeneration of LCn neurons and CFS promotion of forebrain Aβ accumulation, and, thereby, identifies a key molecular link between CSD and specific AD neural injuries.

Keywords: amyloid; hippocampus; locus ceruleus; norepinephrine; sleep deprivation; tau.

Figure 1.

Rotor platform rotation comparably fragments sleep in Dbh⁺ and Dbh⁻ mice. A, A representative polygraphic raw data image of an elicited arousal from nonrapid eye movement (NR) sleep. Top red tracing depicts the electromyographic (EMG) signal, and the white signal at the bottom shows the electroencephalographic (EEG) signal. The bottom panel bins EEG fast Fourier transform (FFT) power bins with blue representing delta or 0–4 Hz, gray for theta 6–9 Hz, and green for alpha, 11–14 Hz. The large white line marks the 10 s when the rotor platform is moving. The timing calibration bar is 4 s. B, Arousal index (number of arousals from sleep/total sleep time in hours) across two genotypes (Dbh⁺ and Dbh⁻) for two sleep conditions, rested, navy and chronic fragmentation of sleep, CFS, green/white). Shown are mean ± SE for n = 4/group. C–E, Twenty-four hour time (minutes) spent in each of the three behavioral states: wake (C), nonrapid eye movement sleep (NREMS, D), and rapid eye movement sleep (REMS, E) for Dbh⁺ and Dbh⁻ mice (n = 4/group, males and females balanced) analyzed with repeated measures two-way ANOVA and Tukey's post hoc testing. F, G, Number of wake (F) and NREMS (G) bouts/24 h in the same mice as for arousal index and behavioral state times. Values are mean ± SE. H, I, Duration of wake (H) and NREMS bout durations, as mean ± SE. In all graphs, navy represents baseline (rested) conditions and green/white diagonals represent CFS conditions. *p < 0.5; **p < 0.01; ***p < 0.001; and ****p < 0.0001.

Figure 2.

CFS upregulates and activates AEP in LCn in an NE-dependent fashion. A, Representative confocal images are presented for LCn responses across sleep and genotype conditions for AEP (green, top panels), AEP-tau (green, middle panels), and AEP-APP (green, bottom panels) with and without TH-labeled LCn (red). Scale bar, 25 μm. B–D, Data are summarized for mean gray area SE ± normalized to mean rested Dbh⁺ values in n = 7–11 (male and female balanced) mice/group for AEP (B), AEP-tau (C), and AEP-APP (D), analyzed with two-way ANOVA and Sidak's post hoc analyses. Significant differences are denoted as *p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 3.

The CFS increase in p-tau in LCn and LCn loss require NE. A, LCn confocal images of p-tau (P Ser202/Thr205, AT8, green) in TH (red)-labeled LCn neurons across groups. Scale bar, 25 μm. B, C, Summary data (mean ± SE) for LCn AT8 (B) and LCn total tau (C) for mean gray normalized to mean for rested Dbh⁺ mice (n = 6–8/group, M&F) analyzed with two-way ANOVA and Sidak's: p < 0.01; ***p < 0.001; ****p < 0.0001. D, Optical fractionator stereological estimates of LCn counts bilaterally for the four groups, mean ± SE, n = 4–6/group, analyzed two-way ANOVA and Sidak's: *p < 0.05; **p < 0.01. E, Representative coronal images of mid-LC region TH immunopositive (navy, substrate blue-labeled) neurons and dendritic field across the four conditions. Sections are counter-stained with Giemsa to delineate nuclei. Scale bar, 50 μm.

Figure 4.

CA1 hippocampal AEP responses to CFS show NE-dependent and -independent effects. A, Confocal images in CA1 of pyramidal cell layer and stratum radiatum across groups showing AEP (green, top panels,), AEP-tau (green, middle panels), and AEP-APP (green, bottom panels), Scale bar, 100 μm. B–D, Group data (mean + SE, normalized to mean of rested Dbh⁺ for mean gray area, n = 5–7) in CA1 for AEP (B), AEP-tau (C), and AEP-APP (D) for rested (navy) and CFS (green/white), analyzed with two-way ANOVA and Tukey's post hoc analyses. Significant differences are denoted as *p < 0.05; **p < 0.01; ***p < 0.001.

Figure 5.

NE-dependent and NE-independent effects for EC AEP responses to CFS. A, Confocal images in EC (layers I–VI) across the four conditions showing AEP (green, top panels), AEP-tau (green, middle panels), and AEP-APP (green, bottom panels). B–D, Group data (mean + SE, normalized to mean of rested Dbh⁺ for mean gray area, n = 5–7) in EC for AEP (B), AEP-tau (C), and AEP-APP (D) for rested (navy) and CFS (green/white), analyzed with two-way ANOVA and Sidak's post hoc analyses. Significant differences are denoted as **p < 0.01; ***p < 0.001; ****p < 0.0001.

Figure 6.

Effects of CFS and NE deficiency on p-tau in EC and Aβ₄₂ in CA1. A, Representative confocal images of p-tau (AT8) (green) and AEP (red) within the EC. Scale bar, 25 μm. B, Confocal images of Aβ₄₂ (green) and Iba-1 (red) in CA1 at pyramidal cells and dorsal stratum radiatum across groups (top panel). Arrows highlight colocalization of Aβ₄₂ in Iba-1 + cells and CD68 (green) and Iba-1 (red) images across groups, bottom panel. Scale bar, 25 μm. C–F, Group data (mean + SE; n = 6–8) in EC for AT8 (C), normalized to mean of rested Dbh⁺ for mean gray area, for Aβ₄₂ in CA1 (D) expressed as percent area above set threshold; for CA1 Iba-1 (E), expressed as % area above set threshold; and CD68 (F) normalized to rested Dbh⁺ for rested (navy) and CFS (green/white), analyzed with two-way ANOVA and Sidak's post hoc analyses. Significant differences are denoted as *p < 0.05; **p < 0.01; ***p < 0.001.