Relationship between circadian genes and memory impairment caused by sleep deprivation

Introduction

Sleep plays a vital role in the human nervous system (Keene & Duboue, 2018). Adequate sleep time is essential for the clearance of toxic metabolites produced by the body, nerve injury self-repair, and cell stress reduction (Hossain et al., 2021), preventing the occurrence of a variety of diseases. However, with the rapid pace of modern society, healthy sleep patterns are hard to achieve, and SD has become a common phenomenon. Numerous studies have shown that SD has deleterious effects on cognitive function by impairing neurological functions and development (Sabia et al., 2021) contributing to a decrease in attention, working memory, emotional signal transmission speed in the brain, and reaction time (Turner et al., 2007). SD can also lead to neurodegenerative diseases and aging (Zada et al., 2021). In addition, SD also suppresses immune and cardiovascular functions (Irwin, 2015). Short-term SD impairs short-term memory and spatial learning memory (Dai et al., 2021), while long-term SD damages novelty-related object recognition memory and object location memory (Jiao et al., 2021). Memory impairment can make daily living significantly harder and can lead to serious adverse socio-economic consequences. Therefore, it is important to clarify the specific mechanism of memory impairment caused by SD.

Circadian clock genes, a kind of gene capable of 24 h autonomous periods and driven by transcription-translation feedback loops (Ikegami et al., 2019), control daily activities and maintain organ function in many physiological processes (Patke et al., 2017). Circadian clock genes contains the core clock genes (such as Arntl, Clock, Per1, Per2, Per3, Cry1, Cry2, and Cry3) and the clock control genes (such as Rora, Rorb, Rorc, Nr1d1, and Nr1d2). In recent years, studies have shown that circadian rhythm plays an important role in many neurological disorders, including schizophrenia, bipolar disorder, depression, and autism (Johansson et al., 2016). The direct inhibition or activation of circadian clock genes plays a regulating role in diseases. For example, patients with Alzheimer’s disease (AD) lose their physiological rhythm of circadian genes, which aggravates cognitive impairment (Ni et al., 2019). Interestingly, the disorder of circadian genes affected by SD can also exacerbate the neuropathological damage of AD (Niu et al., 2021). Bmal1, found in the brain and muscle, is one of the most classic circadian genes, and participates in the hippocampal-dependent memory process by regulating the phagocytosis of microglia (Wang et al., 2021a). In addition, Dicer1 is one of the genetic factors that influence the expression of Cry1 (Lee et al., 2013) and neuronal plasticity (Tinarelli et al., 2014) after SD. Xbp1 maintains the rhythmicity of gene transcription in the mouse liver (Meng et al., 2020). In short, circadian genes play vital roles in the nervous system and in memory function.

Research on the immune system and its connection to memory has attracted considerable attention (Castellano et al., 2017). Spatial memory function is restored in immunodeficient mice after reconstitution of immune function by splenocytes. Immune cells in the central nervous system (CNS), especially T cells, are essential for maintaining hippocampal plasticity and learning memory (Ziv et al., 2006). Further research shows that CD4 T cells promote memory through the IL-4-driven regulation of synaptic function and long-term potentiation (Herz et al., 2021). In AD, eliminating NK cells enhances neurogenesis and reduces neuroinflammation, improving cognitive function in mice (Zhang et al., 2020). Changes in the immune environment of different parts of the body may also have significance (Kaiser et al., 2014). CD4 and IL-6 in the blood are associated with memory function (Wang et al., 2017). In general, the immune environment is a notable factor in memory.

Circadian genes and the immune environment are two determinants of memory function. However, their changes after SD remain insufficiently explored. Although some studies have shown the relationship between circadian genes and immune cells (Labrecque & Cermakian, 2015), their results are still far from comprehensive. The role of circadian genes and the immune environment in memory impairment caused by SD is still unclear. The present study deepens the understanding of memory impairment in SD through bioinformatics analyses and animal experiments, and provides new insights for further research on circadian genes and immune environments in SD.

Materials & Methods

Results

Gene expression of microarray datasets

The boxplot results from GEO2R showed that the median, upper quartile, lower quartile, and distribution of all samples in the standardized datasets of GSE33302 and GSE9442 were relatively consistent, indicating the position and dispersion of the data met the quality requirements. Samples from the sleep deprivation group were colored green, and the ones from the control group were colored purple (Figs. 1A; 1D). The gene expression density maps visually described the distribution of gene expression in each sample. The intensity of genes was mainly between 3 and 12, which showed their consistency in their respective datasets and ideal sample data quality (Figs. 1B; 1E). The volcano maps showed DEGs in the two datasets, with up-regulated genes marked in red and the down-regulated genes marked in blue (Figs. 1C; 1F). We displayed these DEGs in ‘Supplemental Information 1.’

Functional enrichment analysis of GSEA on SD

After SD, the gene transcription level of the mouse hippocampus and brain tissue changed. The change of genes in the nervous system was microscopic. However, a minor variation of genes in the nervous system can cause considerable changes in nervous system function. Therefore, we performed a GSEA analysis to explore the effect of SD. In GSE33302, multiple statistically significant items were enriched, including promoting systemic lupus erythematosus and asthma (immune diseases; Fig. 2A). GO and KEGG terms associated with the process of memory formation were enriched on the suppression of long-term potentiation and axon guidance after SD. Also, the inhibition of nicotine addiction, ferroptosis, and oocyte meiosis suggested that SD affects iron death and fertility. Long-term potentiation, the most statistically significant item is shown in detail in Fig. 2B; other terms are presented in ‘Supplemental Information 2.’ In GSE9442, we found a promotion of protein processing in the endoplasmic reticulum, protein export, and transcriptional misregulation in cancer. Unsurprisingly, memory-related items, including nucleotide excision repair and glycosaminoglycan degradation, were significantly suppressed. These results showed that the changes in overall gene-level after SD were mainly related to aggravated memory impairment and immune diseases.

Functional analysis of circadian rhythm-related genes after SD

To clarify the function of the genes that changed with SD, 29 differently expressed circadian rhythm-related genes were obtained by overlapping circadian rhythm-related genes (Supplemental Information 3) with DEGs in GSE33302 and GSE9442 (Fig. 3A). A GO and KEGG analysis showed that these genes were mainly involved in muscle structure development, cellular response to extracellular stimulus, selenium metabolism/selenoproteins, and circadian rhythm (Fig. 3B). These functions were closely related to each other (Fig. 3C). The MCODE module in the PPI network, with circadian rhythm genes marked in pink, showed seven circadian genes (Figs. 3D–3E).

Relationship between circadian rhythm-related genes

A correlation analysis explored the relationship between the 29 circadian rhythm-related genes. The results showed a synergistic or inhibitory relationship between these circadian rhythm-related genes, with seven direct circadian genes showing the strongest relationship (as shown in red boxes in Fig. 4A). To further verify the importance of these seven circadian genes in SD, an ROC analysis observed the role of these genes through the identification of SD injury. Generally, genes with a high AUC will play a decisive role in the pathogenesis of the disease. All seven genes shared high values in GSE33302 (Supplemental Information 4) and GSE9442 (Fig. 4B). ‘Figure 4C’ shows the expression of these seven circadian genes. The above results showed that these seven circadian genes play important roles in SD.

Relationship between circadian genes and memory

A GSEA analysis found that SD inhibited memory function. In the water maze experiment, SD significantly reduced the number of rats that crossed the platform (from 1.2 to 3.2) and significantly increased the time it took to escape the platform (from 9.94 s to 39.54 s; Figs. 5A–5B). The trajectory diagram of the rats showed less exploration of the plateau after SD (Fig. 5C), as evidenced by the varied expression of seven circadian genes (Fig. 5D), which was in agreement with GSE33302. Meanwhile, the expression of DICER1 and XBP1 changed after SD in human blood samples of GSE98582 (Fig. 5E). However, due to different sequencing platforms, other genes had no corresponding test results in the datasets. Circadian genes play roles in all organs, which indicates the need for further exploration in various organs. Therefore, we detected the expression of these genes in the heart (GSE42323), liver (GSE92913), and lungs (GSE42324), observing that some genes alternated in various organs (Fig. 5F). Seven circadian genes significantly and consistently correlated with two representative indicators of memory function (BDNF and NPAS4). These results indicate that SD leads to memory impairment and changes in seven circadian genes closely related to memory function.

SD altered the hippocampal immune infiltration

Previous studies have reported immunosuppression after SD. However, the changes in the hippocampus after SD are still implicit. M2 macrophages, naïve B cells, follicular helper T cells, plasma cells, resting memory CD4 T cells, activated NK cells, CD8 T cells, monocytes, activated dendritic cells, and resting NK cells were all validated as the most prominent cells in the hippocampal immune environment (Fig. 6A). The contents of four immune cells, including naïve B cells, follicular helper T cells, activated NK cells, activated dendritic cells, changed during SD (Figs. 6B–6D), further demonstrating that circadian genes were related to the changes in naïve B cells, follicular helper T cells, activated NK cells, M2 macrophages, monocytes, and activated dendritic cells after SD (Fig. 6E). In addition, naïve B cells were verified to be upregulated after SD in GSE34424 (Supplemental Information 7). These results showed that circadian genes also affected the immunity of the CNS.

Discussion

As a widespread chronic disease, SD severely damages learning, memory, and immune system function (Xu et al., 2021). The damage caused by sleep deprivation persists even after entering a healthy sleep regime. Although the spatial memory impairment improved 18 h after a healthy sleep regime (Wang et al., 2009a), hippocampal function in the CA1 region remained affected for more than 24 h (Kim, Mahmoud & Grover, 2005). Memory consolidation was still inhibited by sleep deprivation 48 h after normal sleep, showing the long-term effects of sleep deprivation (Li et al., 2009). Inflammatory markers, including IL-17, remained high even after seven days of normal sleep, suggesting that sleep deprivation causes an inflammatory state over a prolonged period (Yehuda et al., 2009). Further study showed that SD causes long-term damage by affecting cortical synapses (Simor et al., 2017). Disturbances in the body’s internal rhythms may be a vital cause of long-term injury. Therefore, it is necessary to study circadian genes. Recently, SD studies have made a series of achievements, with discoveries in memory function and circadian genes (Niu et al., 2021). One study found that changes in the hippocampal region after SD are critical for memory (Gaine et al., 2021). However, there is a lack of research linking the alterations of circadian genes and immune function after SD to memory function. In this study, GSE33302 and GSE9442 were applied for analysis to obtain more reproducible results, both of which showed a high correlation between their changes and memory functions. Differentially expressed circadian genes were obtained and used to analyze their relationship with memory function. Finally, the changes in the immune environment in the hippocampus after SD showed a high correlation with circadian genes.

The SD models of both datasets showed inhibited memory function. Long-term potentiation and axon guidance are functional terms essential for maintaining normal memory function (Frye et al., 2021), both of which were suppressed after SD in GSE33302. In GSE9442, nucleotide excision repair dysfunction inhibited histone recovery, resulting in epigenetic instability and memory impairment (Polo, Roche & Almouzni, 2006). The accumulation of glycosaminoglycan leads to CNS inflammation and increases the permeability of the blood–brain barrier (Wang et al., 2009b). These results highlighted the effect of SD on memory and verified the consistency and reliability of the datasets. SD increases the risk of autoimmune diseases and aggravates their symptoms (Garbarino et al., 2021), but the mechanism behind this relationship still lacks sufficient research. Our GSEA results convincingly concluded that SD affected autoimmune diseases, which were characterized by changes in the immune environment in the hippocampus, providing new ideas for the treatment of immune disorders. Notably, among these GSEA enriched entries, previous studies validated the role of SD on ferroptosis (Wang et al., 2021b), oocyte meiosis (Yong et al., 2021), and transcriptional misregulation in cancer (Barbosa et al., 2021), suggesting the extensive effects of SD on physiological functions.

The genes enrichment analysis on 29 circadian rhythm-related genes closely related to SD found significantly enriched terms, including muscle structure development, cellular response to extracellular stimulus, selenium metabolism or selenoproteins, and circadian rhythm. This indicates that sleep has an impact on both muscle (Chennaoui et al., 2021) and metabolism (Gümüştekín et al., 2004) through circadian rhythm-related genes. These findings also provided a mechanism for SD on the inhibition of myocardial ejection. Encouragingly, seven genes, including Dicer1, Xbp1, Srebf1, Crem, Top1, Sfmbt1, and Naglu, are directly enriched in circadian rhythm. Although studies have reported the role of Dicer1 (Tinarelli et al., 2014), Xbp1 (Brown, Strus & Naidoo, 2017), Crem (Wimmer et al., 2021), and Naglu (Değerliyurt et al., 2021) in SD, their relationship with memory remains unclear, as well as their relationship with immunity. Srebf1 is a gene involved in lipid metabolism and has clinical significance in human blood (Lech et al., 2016). The inhibition of Top1, recently reported in severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), can play a large part in treating fatal inflammation and reducing mortality in mice and may provide an ideal therapy for inflammatory injury caused by SD (Ho et al., 2021). To our knowledge, the role of Sfmbt1 in sleep has not yet been explored.

SD interferes with memory function and physiological circadian genes, as evidenced by the declined memory function present in the water maze experiment. Seven circadian genes significantly changed in the brain with SD. Some genes also have potential value in different organs. Dicer1 and Xbp1 also have clinical significance in human blood. BDNF (Lee, Everitt & Thomas, 2004) and NPAS4 (Ramamoorthi et al., 2011) are two classic indicators of memory formation and consolidation, contributing to hippocampal synaptogenesis and plasticity. Consistent with our prediction, a correlation analysis showed that seven circadian genes, related to BDNF and NPAS4, were the essential components of memory function. Combined with previous studies, we have sufficient evidence to believe that these genes change memory function.

The changes in the immune environment were the dominant determinant of memory in the hippocampus. Among immune cells, the expression levels of M2 macrophages and naïve B cells were the highest in the hippocampus. M2 macrophages are part of immune cells that accelerate the progression of various tumors (Garrido et al., 2020), along with promoting the termination of neuroinflammation and improving cognitive function (Zhu et al., 2016). Although M2 macrophages play a decisive role in memory, they showed no changes in the hippocampus after SD. Interestingly, naïve B cells, follicular helper T cells, activated NK cells, and activated dendritic cells changed significantly after SD, indicating a potential link with SD. Increased naïve B cells in the blood after exercise promotes attention to learning test (Sîrbulescu et al., 2019) Similar to the changes seen after SD, the content of follicular helper T cells also decreased in patients with psoriasis (Shin et al., 2016). Patients with Parkinson’s Disease also showed activated NK cells (Zhou et al., 2020). In addition, the reduction of dendritic cells in the peripheral blood of patients with Alzheimer’s disease has been shown to exacerbate clinical symptoms (Ciaramella et al., 2016). Our results revealed that four immune cells changed in the hippocampus, which further elucidates the mechanism of SD in regulating the immune environment on memory.

Finally, we explored the effects of circadian genes on the immune environment. Although studies have conclusively shown that circadian genes influence the immune environment (Boivin et al., 2003), intensive studies of SD still lag behind. We found multiple relationships between circadian genes and immune cells, some of which have appeared in other diseases. The inactivation of Dicer1 leads to NK cell differentiation dysfunction (Degouve et al., 2016). Srebf1 activates NK cells in melanoma (Fu et al., 2020). SD inhibits both Srebf1 and NK cells, according to their regulatory relationship. The other regulatory relationships we found in this study have not been reported in SD. These regulatory relationships indicate the need for further research and deserve further confirmation.

This study creatively investigated the effects of circadian genes and hippocampal immune environment on memory function after SD and elucidated the relationship between circadian genes and immune cells. This study does still have some limitations. First, the model condition of SD from the two datasets we used was not the same. For example, the duration of SD was different, which may increase the difference between the two sequencing results. Second, the samples used of the two datasets were from the brain and the hippocampus. The brain contains many organized regions, which increases the difficulty and possible inaccuracy of the analysis. Therefore, combined with the relationship between memory and the hippocampus, we only detected the immune infiltration of the hippocampus. Third, the expression profiles only showed Dicer1 and Xbp1 in human blood because of the different sequencing data platforms, but the role of other circadian clock genes in humans is still worth further study. Fourth, although we demonstrated the significance of naïve B cells in GSE34424, we are still lacking effective experimental verification. Further experimental studies are required to illuminate the relationship between circadian genes and immune cells.

Conclusions

In summary, a bioinformatics analysis of circadian genes and the hippocampal immune environment reveals that their relationship with memory may play an essential role in SD memory impairment, providing potential targets for SD treatment and profound insights into its genetic mechanisms.

Supplemental Information