Sleep fragmentation affects glymphatic system through the different expression of AQP4 in wild type and 5xFAD mouse models

Abstract

Alzheimer's disease (AD) is characterized by genetic and multifactorial risk factors. Many studies correlate AD to sleep disorders. In this study, we performed and validated a mouse model of AD and sleep fragmentation, which properly mimics a real condition of intermittent awakening. We noticed that sleep fragmentation induces a general acceleration of AD progression in 5xFAD mice, while in wild type mice it affects cognitive behaviors in particular learning and memory. Both these events may be correlated to aquaporin-4 (AQP4) modulation, a crucial player of the glymphatic system activity. In particular, sleep fragmentation differentially affects aquaporin-4 channel (AQP4) expression according to the stage of the disease, with an up-regulation in younger animals, while such change cannot be detected in older ones. Moreover, in wild type mice sleep fragmentation affects cognitive behaviors, in particular learning and memory, by compromising the glymphatic system through the decrease of AQP4. Nevertheless, an in-depth study is needed to better understand the mechanism by which AQP4 is modulated and whether it could be considered a risk factor for the disease development in wild type mice. If our hypotheses are going to be confirmed, AQP4 modulation may represent the convergence point between AD and sleep disorder pathogenic mechanisms.

Keywords: Alzheimer’s disease; Amyloid-β; Aquaporin-4 channel; Neuroinflammation; Sleep fragmentation; p-tau.

Fig. 1

A Sleep fragmentation protocol time-line. B Typical hypnograms reporting sleep oscillations during baseline 24 h recording (top) and during fragmentation protocol (bottom). Thick lines indicate the dark period. C Total sleep time is represented as an average of the recordings. D The variation in the percentages relating to wakefulness, NREM sleep, and REM sleep. The two genotypes were compared for the three variables in basal and fragmentation conditions. NF not fragmented mice, F fragmented mice. The data are mean standard error of the mean (SEM), n = 3 per strain

Fig. 2

A Time spent in both closed and open arms during the EPM test. B Distance traveled in closed and open arms, and in the total arena during the EPM test. C Distance traveled in arena, border, and center in the OF test. D Time spent in both the border and the center of the arena during the OF test. E Behavioral activities during the OF test (protective and un-protective rearing). NF not fragmented mice, F fragmented mice. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.0001 versus control by one-way ANOVA followed by Fisher’s LDS post-hoc test, n = 11 per condition

Fig. 3

A Time spent near the old and the new objects. B Interaction with the old and the new objects and the total frequency of interaction with both objects. C, D Percentage of the discrimination and recognition indexes respectively. NF not fragmented mice, F fragmented mice. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.0001 versus control by one-way ANOVA followed by Fisher’s LDS post-hoc test, n = 11 per condition

Fig. 4

A Frequency of arm alternations. B Total distance traveled in the arena. C Time spent in the different arms. D Latency of the first entry in the new arm (arm 3). E Frequency of entry in the first arm (arm 1). NF not fragmented mice, F fragmented mice. *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.0001

Fig. 5

A Representative images of immunohistochemistry acquired with Axioscan microscope. Scale bar 1000 μm. Aβ plaques are labeled in green and the nuclei with DAPI in blue. B Histogram of Aβ plaques analyzed through the percentage of pixels, after the same threshold is set for all the region of interests. C Histogram of Aβ plaque number. D Representative images of immunohistochemistry of AT8. NF not fragmented mice, F fragmented mice, LS lateral septum, RSC retrosplenial cortex, MSC motor-sensory cortex, DG dentate gyrus, TH thalamus, HY hypothalamus, BLA basolateral amygdala. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. *p < 0.05; ****p < 0.0001 versus control by one-way ANOVA followed by Bonferroni post-hoc test, n = 4 per condition

Fig. 6

A Representative images of immunohistochemistry of GFAP and 6e10 antibodies in all the regions analyzed. B Representative images of immunohistochemistry of GFAP and 6e10 antibodies acquired with Axioscan microscope. Scale bar 1000 μm. C Histogram of GFAP and 6e10 shown together. D Histogram of GFAP density analyzed through the percentage of pixels, after the same threshold is set for all the region of interests. NF not fragmented mice, F fragmented mice, LS lateral septum, RSC retrosplenial cortex, MSC motor-sensory cortex, DG dentate gyrus, TH thalamus, HY hypothalamus, BLA basolateral amygdala. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. **p < 0.01; ****p < 0.0001 versus control by ANOVA followed by Bonferroni post-hoc test, n = 4 per condition

Fig. 7

A Representative images of immunohistochemistry of iba-1 and 6e10 antibodies acquired with confocal microscope in all the regions analyzed. The process of analysis includes the skeletonization of microglia for the evaluation of cell complexity. A representation of skeletonized cells is shown on the right of the iba-1 images for each brain region. B–H Analysis of the structural complexity of microglia cells by using AnalyzeSkeleton (2D/3D) ImageJ plugin. NF not fragmented mice, F fragmented mice. The data are mean standard error of the mean (SEM). **p < 0.01; ***p < 0.005; ****p < 0.0001 versus control by ANOVA followed by Bonferroni post-hoc test, n = 3 per condition

Fig. 8

A Representative images of immunohistochemistry of AQP4 antibody acquired with Axioscan microscope in 2-months-old mice. Scale bar 1000 μm. B Histogram of AQP4 density in 2-months-old mice analyzed through the percentage of pixels, after the same threshold is set for all the region of interests. C Representative images of immunohistochemistry in 2-months-old mice of AQP4 and CD31, blood vessel marker, acquired with confocal microscope. D Representative images of immunohistochemistry of AQP4 antibody acquired with Axioscan microscope in 6-months-old mice. Scale bar 1000 μm. E, F Histograms of AQP4 and GFAP densities respectively in 6-months-old mice analyzed through the percentage of pixels, after the same threshold is set for all the region of interests. NF not fragmented mice, F fragmented mice, LS lateral septum, RSC retrosplenial cortex, MSC motor-sensory cortex, DG dentate gyrus, TH thalamus, HY hypothalamus, BLA basolateral amygdala. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. *p < 0.05; **p < 0.01; ***p < 0.005; ****p < 0.0001 versus control by ANOVA followed by Bonferroni post-hoc test, n = 4 per condition

Fig. 9

A Representative images of immunohistochemistry of AQP4 antibody acquired with Axioscan microscope in wild type mice. Scale bar 1000 μm. B, C Histograms of AQP4 and GFAP density respectively analyzed through the percentage of pixels, after the same threshold is set for all the region of interests. NF not fragmented mice, F fragmented mice, LS lateral septum, RSC retrosplenial cortex, MSC motor-sensory cortex, DG dentate gyrus, TH thalamus, HY hypothalamus, BLA basolateral amygdala. The data are mean standard error of the mean (SEM). Each data point represents an individual animal. **p < 0.01; ***p < 0.005 versus control by ANOVA followed by Bonferroni post-hoc test, n = 4 per condition