NKCC1 inhibition improves sleep quality and EEG information content in a Down syndrome mouse model

Abstract

In several brain disorders, the hyperpolarizing/inhibitory effects of GABA signaling through Cl-permeable GABA_A receptors are compromised, leading to an imbalance between neuronal excitation and inhibition. For example, the Ts65Dn mouse model of Down syndrome (DS) exhibits increased expression of the Cl^- importer NKCC1, leading to depolarizing gamma aminobutyric acid (GABA) signaling in the mature hippocampus and cortex. Inhibiting NKCC1 with the Food and Drug Administration (FDA)-approved diuretic bumetanide rescues inhibitory GABAergic transmission, synaptic plasticity, and cognitive functions in adult Ts65Dn mice. Given that DS individuals and Ts65Dn mice show sleep disturbances, and considering the key role of GABAergic transmission in sleep, we investigated whether NKCC1 upregulation contributes to sleep abnormalities in adult Ts65Dn mice. Chronic oral administration of bumetanide ameliorated the spectral profile of sleep, sleep architecture, and electroencephalogram (EEG) entropy/complexity, accompanied by a lower hyperactivity in trisomic mice. These results offer a potential avenue for addressing common sleep disturbances in DS.

Keywords: Molecular biology; Neuroscience.

Graphical abstract

Figure 1

Sleep structure and power spectral density are altered in the Ts65Dn mouse model of Down syndrome (A) Scheme of the experimental procedure with surgery, healing period (2 days), pharmacological treatment, and in vivo EEG recordings during baseline (24 h) and recovery (18 h) after sleep deprivation (6 h, SD) in adult mice. (B) Quantification of the mean (±SEM) sleep duration across all recording period, binned every 2 h (WT Veh vs. Ts Veh, two-way ANOVA, F _genotype (1, 11) = 3.995, p = 0.07). (C) (Left) Quantification of the mean (±SEM) NREM sleep duration in the same recordings in (B) (WT Veh vs. Ts Veh, two-way ANOVA, F _genotype (1, 11) = 7.371, #p < 0.05; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test). (Middle, right) Quantification of the mean (±SEM) EEG power densities over the frequency spectrum (μV²/Hz) during NREM sleep at baseline (middle, baseline: WT Veh vs. Ts Veh, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 7.162, ####p < 0.0001; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test) and after SD (right, recovery: WT Veh vs. Ts Veh, two-way ANOVA, F _{frequency band X genotype} (39, 429) = 5.992, ####p < 0.0001; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test). (D) (Left) Quantification of the mean (±SEM) REM sleep duration in the same recordings in (B) (WT Veh vs. Ts Veh, two-way ANOVA, F _genotype (1, 11) = 1.67, p = 1.558). (Middle, right) Quantification of the mean ± (SEM) EEG power densities all over the frequency spectrum (μV²/Hz) during REM at baseline (middle, baseline: WT Veh vs. Ts Veh, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 4.776, ####p < 0.0001; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test) and after SD (right, recovery: WT Veh vs. Ts Veh, two-way ANOVA, F _{frequency band X genotype} (39, 429) = 5.264, ####p < 0.0001; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test). In all relevant panels, h from ZT: hours from the beginning of the recording (zeitgeber). Yellow represents the light phase, which is sleep time for nocturnal mice; white represents the dark phase, the active period; yellow and gray oblique stripes represent sleep deprivation. N indicates the number of recorded mice.

Figure 2

Bumetanide significantly rescues power spectral density in adult Ts65Dn mice (A) Scheme of the experimental procedure with surgery, healing period (2 days), pharmacological treatment, and in vivo EEG recordings during baseline (24 h) and recovery (18 h) after sleep deprivation (6 h, SD) in adult mice. (B) (Left) Quantification of the mean (±SEM) NREM sleep duration in the same set of experiments of Figure 1 with animals treated with Vehicle or bumetanide (WT Veh vs. WT Bume, two-way ANOVA, F genotype (1, 11) = 0.7476, p = 0.4057; Ts Veh vs. Ts Bume, two-way ANOVA, F genotype (1, 11) = 1.582, p = 0.2371). (Middle, right) Quantification of the mean ± SEM EEG power densities over the frequency spectrum (μV2/Hz) during NREM sleep at baseline (middle, baseline: WT Veh vs. WT Bume, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 0.8013, p = 0.7997; Ts Veh vs. Ts Bume, two-way ANOVA, F_{frequency band X genotype} (39, 390) = 2.642, ####p < 0.0001; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test) and after SD (right, recovery: WT Veh vs. WT Bume, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 0.9621, p = 0.5383; Ts Veh vs. Ts Bume, two-way ANOVA, F_{frequency band X genotype} (39, 390) = 1.921, ##p < 0.01). (C) (Left) Quantification of the mean (±SEM) REM sleep duration in the same set of experiments of Figure 1 with animals treated with Vehicle or bumetanide (WT Veh vs. WT Bume, two-way ANOVA, F genotype (1, 11) = 0.7476, p = 0.4057; Ts Veh vs. Ts Bume, two-way ANOVA, F genotype (1, 11) = 1.582, p = 0.2371). (Middle, right) Quantification of the mean (±SEM) EEG power densities over the frequency spectrum (μV2/Hz) during REM sleep at baseline (middle, baseline: WT Veh vs. WT Bume, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 0.8013, p = 0.7997; Ts Veh vs. Ts Bume, two-way ANOVA, F_{frequency band X genotype} (39, 390) = 2.642, ####p < 0.0001) and after SD (right, recovery: WT Veh vs. WT Bume, two-way ANOVA, F_{frequency band X genotype} (39, 429) = 0.9621, p = 0.5383; Ts Veh vs. Ts Bume, two-way ANOVA, F_{frequency band X genotype} (39, 390) = 1.921, ##p < 0.01; blue line indicates ∗p < 0.05 in post hoc Šídák’s multiple comparison test). In all relevant panels, h from ZT: hours from the beginning of the recording (zeitgeber); the recording time was binned every 2 h. Yellow represents the light phase, which is sleep time for nocturnal mice; white represents the dark phase, the active period; yellow and gray oblique stripes represent sleep deprivation. N indicates the number of recorded mice.

Figure 3

Bumetanide partially rescues daily baseline hyperactivity, but it does not impact on body temperature or food intake in adult Ts65Dn mice (A) Scheme of the experimental procedure with surgery, healing period, pharmacological treatment, and assessment of activity behavior and body/temperature during baseline and recovery after SD. These experiments were performed on the same animals as for Figures 1 and 2. (B) Quantification of the mean (±SEM) activity across 24 h at baseline (light, two-way ANOVA, Fgenotype (1, 20) = 7.37, #p < 0.05; dark, two-way ANOVA, Fgenotype (1, 20) = 3.27, p = 0.0856; ∗p < 0.05, post hoc Šídák’s multiple comparisons test) and 18 h after SD (recovery; light, two-way ANOVA, Fgenotype (1, 20) = 0.9379, p = 0.3444; dark, two-way ANOVA, Fgenotype (1, 20) = 2.224, p = 0.1515). (C) Quantification of the mean (±SEM) body temperature across 24 h at baseline (light, two-way ANOVA, Fgenotype (1, 21) = 0.0763, p = 0.7851; dark, two-way ANOVA, Fgenotype (1, 21) = 5.233, #p < 0.05) and 18 h after SD (light, two-way ANOVA, Fgenotype (1, 21) = 0.0411, p = 0.8413; dark, two-way ANOVA, Fgenotype (1, 21) = 2.504, p = 0.1285). (D) Quantification of the mean (±SEM) food intake index (mg of food consumed in 24 h/g of mouse weight) across 24 h at baseline (one-way ANOVA with Welch correction, F (4,24) = 6.973, ##p < 0.01; post hoc Dunnett’s T3 multiple comparisons test, ∗p < 0.05). For all graphs, each dot represents the value recorded for one mouse. Yellow represents the light phase, which is sleep time for nocturnal mice; white represents the dark phase, the active period.