Sleep deprivation in adolescent mice impairs long-term memory till early adulthood via suppression of hippocampal astrocytes

Abstract

Sleep deficiency is a rampant issue in modern society, serving as a pathogenic element contributing to learning and memory impairment, with heightened sensitivity observed in children. Clinical observations suggest that learning disabilities associated with insufficient sleep during adolescence can persist through adulthood, but experimental evidence for this is lacking. In this study, we examined the impact of early-life sleep deprivation (SD) on both short-term and long-term memory, tracking the effects sequentially into adulthood. We employed a modified multiple-platform method mouse model to investigate these outcomes. SD induced over a 14-day period, beginning on postnatal day 28 (PND28) in mice, led to significant impairment in long-term memory (while short-term memory remained unaffected) at PND42. Notably, this dysfunction persisted into adulthood at PND85. The specific impairment observed in long-term memory was elucidated through histopathological alterations in hippocampal neurogenesis, as evidenced by bromodeoxyuridine (BrdU) signals, observed both at PND42 and PND85. Furthermore, the hippocampal region exhibited significantly diminished protein expressions of astrocytes, characterized by lowered levels of aquaporin 4 (AQP4), a representative molecule involved in brain clearance processes, and reduced protein expressions of brain-derived neurotrophic factors. In conclusion, we have presented experimental evidence indicating that sleep deficiency-related impairment of long-term memory in adolescence can endure into adulthood. The corresponding mechanisms may indicate that the modification of astrocyte-related molecules has led to changes in hippocampal neurogenesis.

Keywords: adolescence; astrocyte; early adulthood; hippocampus; neurogenesis; sleep deprivation.

Graphical Abstract

Figure 1.

Experimental design of this study. In experimental study 1 (A), behavioral tests were conducted on postnatal day 42 (PND42) following a 14-day period of sleep deprivation. Subsequently, additional behavioral tests were administered on PND85 after a 6-week period of normal sleep. In experimental study 2 (B), bromodeoxyuridine (BrdU) was injected daily for 10 days before the animals were euthanized. Blood and brain tissue samples were collected at both PND42 and PND85 to investigate the underlying mechanisms. Bodyweight measurements were recorded from postnatal day 20 to PND85 (C). The data are expressed as the means ± standard deviations (n = 8). *p < .05 and **p < .01 compared with the normal group

Figure 2.

Behavioral changes by sleep deprivation (SD). The behaviors for long-term memory in passive avoidance test (A), cognition in novel object test (B), spatial memory in Y maze test (C), anxiety (D), and locomotor activity (E) in open field test were investigated in mice underwent temporary SD, on days PND42 and PND85, respectively. The data are expressed as the means ± standard deviations (n = 6 or 8). #p < .05 and ##p < .01 compared with the PND42 normal group, *p < .05 and **p < .01 compared with the PND85 normal

Figure 3.

Changes in hippocampal neurogenesis by sleep deprivation. Hippocampal neurogenesis was evaluated by western blot analysis for DCX and NeuN in tissue lysate of the dentate gyrus (A), and its expressions were semi-quantified (B). To verify the neurogenesis using an immunofluorescent staining method, the positive signals to DCX, BrdU, NeuN were analyzed in dentate gyrus (C), and their intensities were semi-quantified (D). The data are expressed as the means ± standard deviations (n = 3; evaluated the data in triplicate by pooling the brains of five mice). ##p < .01 compared with the PND42 normal group, **p < .01 compared with the PND85 normal group.

Figure 4.

Changes in hippocampal neuroglia expression by sleep deprivation. Activities of neuroglia, including astrocyte (GFAP) and microglia (Iba-1) were evaluated by Western blot analysis in the hippocampus (A), and its expressions were semi-quantified (B). Using an immunofluorescent staining method, the cell number, aera, and morphology of GFAP and double-positive signals of GFAP/BrdU in the dentate gyrus (C) and Iba-1 positive signals in the dentate gyrus and CA1 region (G) were verified. Its intensities were semi-quantified (D to F, H, and I). The data are expressed as the means ± standard deviations (n = 3; evaluated the data in triplicate by pooling the brains of five mice). ##p < .01 compared with the PND42 normal group, **p < .01 compared with the PND85 normal group

Figure 5.

Changes in neuro-immune cytokines and molecules that may be related to astrogenesis in hippocampus by sleep deprivation. Cytokine levels of IL-6 (A) and IL-10 (B) in hippocampal tissue lysate were determined using ELISA. Hippocampal protein expressions of IL-10R, p-JAK1/JAK1, and p- STAT3/STAT3 were measured by Western blotting analysis (C), and its expressions were semi-quantified (D). The data are expressed as the means ± standard deviations (n = 3; evaluated the data in triplicate by pooling the brains of five mice). #p < .05 and ##p < .01 compared with the PND42 normal group, *p < .05 and **p < .01 compared with the PND85 normal group

Figure 6.

Changes in molecules that may be related to neuronal plasticity and glymphatic activity by sleep deprivation. Hippocampal neuronal activity-related molecules were evaluated by western blot analysis for neuronal plasticity-related molecules (BDNF and p-CREB/CREB) and glymphatic activity-related molecules (Aquaporin 4; AQP4 and p-p38/p38; A) and its expressions were semi-quantified (B). Astrocyte-derived glymphatic activity-related molecules were evaluated by immunofluorescent staining for GFAP and AQP4, and these co-localizations in hippocampus (C), and its positive signals were semi-quantified (D). The data are expressed as the means ± standard deviations (n = 3; evaluated the data in triplicate by pooling the brains of five mice). # p < .05 and ## p < .01 comparing to the PND42 normal group, *p < .05 and **p < .01 comparing to the PND85 normal group.