Inflammation from Sleep Fragmentation Starts in the Periphery Rather than Brain in Male Mice

Obstructive sleep apnea is increasing worldwide, leading to disordered sleep patterns and inflammatory responses in brain and peripheral tissues that predispose individuals to chronic disease. Pro-inflammatory cytokines activate the inflammatory response and are normally regulated by glucocorticoids secreted from adrenal glands. However, the temporal dynamics of inflammatory responses and hypothalamic-pituitary-adrenal (HPA) axis activation in relation to acute sleep fragmentation (ASF) are undescribed. Male C57BL/6J mice were exposed to ASF or control conditions (no ASF) over specified intervals (1, 2, 6, and 24 h) and cytokine gene expression (IL-1beta, TNF-alpha) in brain and peripheral tissues as well as serum glucocorticoid and interleukin-6 (IL-6) concentration were assessed. The HPA axis was rapidly activated, leading to elevated serum corticosterone from 1–24 h of ASF compared with controls. This activation was followed by elevated serum IL-6 concentration from 6–24 h of ASF. The tissue to first exhibit increased pro-inflammatory gene expression from ASF was heart (1 h of ASF). In contrast, pro-inflammatory gene expression was suppressed in hypothalamus after 1 h of ASF, but elevated after 6 h. Because the HPA axis was activated throughout ASF, this suggests that brain, but not peripheral, pro-inflammatory responses were rapidly inhibited by glucocorticoid immunosuppression.


Introduction
The increased prevalence of obesity in the United States and other developed countries has drastically increased obstructive sleep apnea (OSA) diagnoses 1 . This condition leads to sleep fragmentation (SF), reduced blood oxygen saturation, increased daytime sleepiness, and the occurrence of in ammation in the brain and peripheral tissues 2,3 . While OSA is related to the development of chronic pathologies, such as metabolic 4,5 and cardiovascular diseases 6,7 as well as neurological disorders 8 , the underlying mechanism of these studies is still unclear, although chronic in ammation is thought to play a large role in determining disease outcomes 2 .
In ammation is a pervasive phenomenon that is typically triggered during the onset of infection, injury, or exposure to pollutants 9 . Sleep loss is also a potent inducer of in ammation 2,3,10 , but the mechanisms underlying this response are unclear. Sleep promotes the clearance of metabolic waste products, such as beta-amyloid protein [11][12][13][14] , and loss of sleep, in turn, reduces clearance, leading to a build-up of waste products that may trigger the immune system to produce in ammatory mediators, termed herein the "metabolic clearance hypothesis." Based upon this possibility, the onset of in ammation from sleep loss could occur in the brain. The alternative hypothesis is that in ammation begins in the periphery with sympathetic afferents relaying this information to the brain, similar to a neuro-immune re ex to peripheral immune challenge 15 . As evidence, previous studies have shown that inhibition of the peripheral sympathetic nervous system (SNS) reduces in ammatory responses to SF in peripheral tissues 16,17 and brain 18 .
Superimposed upon these in ammatory responses is activation of the hypothalamic-pituitary-adrenal (HPA) axis, which culminates in the release of glucocorticoids from the adrenal cortices and acts as a brake on immune function and in ammation 19 . Despite this long standing dogma, glucocorticoids can in some cases prime pro-in ammatory responses in the brain 20,21 . Interestingly, HPA axis activation occurs in tandem with pro-in ammatory responses to SF in mice 10,16,22,23 . Thus, it remains unresolved whether HPA activation provides negative feedback to in ammatory responses, temporarily potentiates responses, or has no effect.
The aim of this study was to evaluate the time course of peripheral and brain in ammatory responses in relation to HPA activation in male C57BL/6J mice exposed to acute SF. We predicted an initial increase in pro-in ammatory cytokine gene expression in brain that would be negatively regulated by a rise in serum corticosterone from HPA activation occurring later in the time course. We also predicted that neuroin ammatory responses would occur earlier than peripheral in ammatory responses, as suggested by the metabolic clearance hypothesis, which would represent the initial neuro-immune response to the build-up of metabolic waste products in brain. The alternative hypotheses are that the initiation of the in ammatory response from ASF occurs rst in the periphery and then relayed to the brain through sympathetic afferents or that peripheral and brain in ammatory responses to sleep fragmentation occur independently from each other.

Methods
Animals. Male adult C57BL/6J mice between 8-12 weeks of age were used in this study (n=110; Jackson Laboratory, Bar Harbor, ME). Mice were given food and water ad libitum and housed under standard rodent colony conditions (lights on: 0800-2000 h, 21°C ± 1°C) at Western Kentucky University. Acute sleep fragmentation (ASF) experiments were performed using automated sleep fragmentation chambers (Lafayette Instrument Company; Lafayette, IN; model 80390) with a thin layer of corn bedding as previously described and each chamber contained no more than ve mice 23 . These chambers ensure that mice are subjected to sleep fragmentation and not absolute sleep deprivation 16 . Mice were acclimated to the sleep fragmentation (SF) chambers for 48 h before the commencement of experiments to minimize carryover effects from the different cage environments 24 . This study was conducted under the approval of the Institutional Animal Care and Use Committee at Western Kentucky University (#19-11), and procedures followed the National Institutes of Health's "Guide for the Use and Care of Laboratory Animals" and ARRIVE guidelines.
Acute sleep fragmentation (Acute SF) and sample collection.
Starting at 0800 (lights on), mice were exposed to 1, 2, 6, 12, or 24 h (n=110; all groups, n = 10) of ASF, which involves a sweeping bar that moves horizontally across the modi ed cage every 120 sec, simulating the rate of SF in patients with severe sleep apnea 25 (Fig. 1). For the non-sleep fragmentation (NSF) control mice, subjects were housed in SF chambers, but no sweeping bar movements occurred. The NSF groups matched collection times of ASF mice (1, 2, 6, 12, or 24 h; all groups, n = 10). Both ASF and NSF groups were compared to a baseline group (time = 0) of mice collected at 0800 (n = 10).
Sample Collection. After ASF or NSF treatments, mice were rapidly anesthetized using iso urane induction (5%) and decapitated <3 min of initial handling for tissue gene expression studies and blood collection for measurement of CORT and interleukin-6 (IL-6) levels (see below). Trunk blood was collected from mice, kept on ice for <20 min, and spun at 3000×g for 30 min at 4°C. Serum was collected and stored at -80°C for corticosterone and IL-6 ELISA assays (see below). For gene expression analyses, three brain regions (prefrontal cortex (PFC), hypothalamus, and hippocampus), liver, spleen, heart, and epididymal white adipose tissue (EWAT) were dissected from mice and stored in RNAlater solution (ThermoScienti c) in the freezer at -20°C. These particular brain regions and peripheral tissues were chosen because previous studies have demonstrated elevated pro-in ammatory gene expression from ASF 12,18 . All tissue samples were stored at -20°C before RNA extraction.
Corticosterone and Interleukin-6 ELISA. Serum levels of corticosterone (n = 9-10/group) were measured using an ELISA kit (Catalogue number ADI-901-097, EnzoLife Sciences) which had a sensitivity of 26.99 pg/mL with cross-reactivity of <30% deoxycorticosterone and <2% progesterone. Samples were diluted 1:40 before running. The reaction was carried out in duplicate according to the kit instructions, and the average absorbance of the plate was determined using a plate reader (BioTek Synergy H1 Hybrid Reader).
Average intra-and inter-assay variations were 2.85% and 2.65% respectively. IL-6 was measured in sera using ELISA MAX Deluxe kits (Catalogue number 431304; BioLegend, San Diego, CA). The assays were carried out according to the manufacturer's instructions, and the average intra-and interassay variations were 8.24% and 7.43% respectively.
Cytokine gene expression. RNA was extracted from liver, spleen, epidydimal white adipose tissue (EWAT), as well as the prefrontal cortex, hippocampus, and hypothalamus from brain using a RNeasy mini kit (Qiagen). RNA was extracted from the heart using a RNeasy Fibrous Tissue mini kit (Qiagen). All extractions were performed following the manufacturer's instructions. RNA concentrations were measured using a NanoDrop 2000 Spectrophotometer (Thermo Scienti c). Total RNA was reverse transcribed using a high-capacity cDNA reverse transcription kit (ThermoFisher Scienti c, Cat number: 4368813) according to the manufacturer's instructions and used as a template for determining relative cytokine gene expression using an ABI 7300 RTPCR system. Tissues were analyzed with cytokine primers/probes (IL1β: Mm00434228, TNFα: Mm00443258; ThermoFisher Scienti c). Assay probes were labeled with orescent marker 5-FAM and quencher MGB at the 5' end and 3' end, respectively, and VIClabeled 18S primer/probe (primer-limited; 4319413E; ThermoFisher Scienti c) was used as an endogenous control. A multiplex PCR assay which included the genes of interest, and the endogenous control was run simultaneously for each sample. Samples were run in duplicate and the fold change in mRNA level was calculated as the relative mRNA expression levels, 2 -ΔΔCt . The cycle threshold (Ct) at which the uorescence exceeded background levels was used to calculate ΔCt (Ct[target gene] -Ct[18S]). Each Ct value was normalized against the highest Ct value of a control sample (ΔΔCt), and then the negative value of this power to 2 (2 -ΔΔCt ) was used for mRNA expression analysis.
Statistical Analysis. Data are presented as mean (±SE). All statistical analyses were performed using GraphPad Prism (version 9.0). Two-way ANOVAs assessed the effect of sleep treatment (ASF or NSF), time (1h, 2h, 6h, 12h, 24h), and their interaction on mRNA expression of cytokines, serum CORT levels, and serum IL-6 concentration. One-way ANOVAs were used to assess whether ASF and NSF groups differed signi cantly from baseline levels (time 0h). Tukey's HSD and Bonferroni multiple comparisons were used for post hoc analyses for one-way ANOVA and two-way ANOVA, respectively. p < 0.05 was considered statistically signi cant.

Discussion
Sleep loss induces an in ammatory response in various tissues of the body 10 , but the temporal dynamics of in ammatory responses are unclear. Results from this time-course study fail to support the metabolic clearance hypothesis that in ammation from acute sleep fragmentation (ASF) begins in the brain. Instead, elevated pro-in ammatory gene expression was rst detected in heart tissue 1 h after ASF. The rst brain region measured that exhibited neuroin ammation was the hypothalamus after 6 h of ASF. However, there was suppression of pro-in ammatory gene expression at 1 h of ASF. These ndings support the alternative hypothesis that peripheral in ammation from sleep fragmentation occurs rst in the periphery, speci cally the heart, rather than brain. Previous studies have documented increased proin ammatory gene expression in cardiac tissue following short-term 10,16 and chronic sleep fragmentation 16,26 . We surmise that pro-in ammatory cytokines are elevated rapidly in heart due to increased SNS activity from sleep fragmentation 23,27,28 . Whether the heart directly relays in ammatory "information" to brain through sympathetic afferents or whether neuroin ammation is simply a slower, but independent, process requires further study. In addition, this study only measured select regions in brain and several organs/tissues in the periphery. Therefore, it is possible that other tissues that were not measured may have different temporal responses to ASF.
The HPA axis was rapidly activated from ASF as measured by increased corticosterone concentration in serum, which is a consistent nding from previous studies 10,18,22,23 . NSF mice exhibited a well-known diurnal rhythm in circulating corticosterone level with peak levels occurring at the onset of nocturnal activity 29 . Among ASF mice, serum corticosterone increased rapidly from 0h to 1h and remained elevated for the entire time course compared with NSF mice. As stated above, at 1h of ASF, there was suppression of TNF-α expression in hypothalamus, but an elevation of IL-1β expression in heart compared with NSF controls. These ndings imply that the increased serum corticosterone concentration at this 1h time point may be rapidly suppressing an in ammatory response in brain, such as hypothalamus, but has no suppressive effect or even possibly a pro-in ammatory effect in the periphery (Table 1). Table 1. Timing of Pro-In ammatory Cytokine mRNA Expression Levels from ASF or NSF. Summary of main effects of ASF, time, and their interaction on TNFα and IL1β cytokine gene expression levels in brain and peripheral tissues. epididymal white adipose tissue: EWAT, SF: sleep fragmentation.
ASF also elevated serum IL-6 levels concomitant with increased serum corticosterone levels, with IL-6 levels peaking after 6 h but falling thereafter. Previous studies employing various forms of sleep deprivation on mice and humans have described elevated serum IL-6 concentration 30,31 . IL-6 acts pleiotropically through multiple pro-and anti-in ammatory pathways 32 , and glucocorticoids can alter the balance of these pathways through interfering with the expression of the suppressor of cytokine signalling 3 (SOCS3) feedback inhibitor 33 , as well as repressing the transcriptional activation of nuclear factor-kappa B (NF-κB) 34 . Whether elevated serum corticosterone concentration provided negative feedback to decrease serum IL-6 levels 12 and 24 h after ASF (relative to 6 h) will require additional investigation. In addition, it is also unknown whether IL-6 reciprocally activated the HPA axis as is often the case for bi-directional neuroendocrine-immune interactions 35 . Moreover, studies that manipulate glucocorticoid action either through adrenalectomy/hormone replacement experiments or pharmacological approaches that inhibit glucorticoid synthesis and/or receptor binding are needed to pinpoint the precise modulatory effects of glucocorticoids. It is also possible that activation of the SNS from acute sleep fragmentation through release of norepinephrine/epinephrine from adrenal medullae could promote a pro-in ammatory effect in the periphery that is independent of glucocorticoid effects. As evidence, suppression of the SNS using chemical sympathectomy alleviates in ammatory responses from acute and chronic sleep fragmentation in peripheral tissues 16 as well as brain 18 .
The ndings indicate that pro-in ammatory responses to acute sleep fragmentation are tissue-speci c, which is consistent with previous studies [16][17][18]23 . ASF increased TNFα expression in heart and EWAT after exposure for 12 and 24 h, respectively, but there was no effect in liver or spleen. In addition, TNFα expression from ASF occurred earliest in EWAT at 12h while others were at 24h (Fig. 3A, C, E; Table 1). This nding is consistent with previous studies, supporting the role of EWAT in the pro-in ammatory response to acute sleep fragmentation 22,23 .
In brain, ASF signi cantly increased IL-1β expression in hypothalamus and hippocampus at 6h, 12h, and 24h time points, respectively, but did not affect IL-1β in prefrontal cortex ( Table 1). The brain responses for IL-1β gene expression were similar compared to TNFα but earlier. These results are a departure from previous studies that have shown elevated pro-in ammatory gene expression in hypothalamus, hippocampus, and pre-frontal cortex after 24-h of sleep fragmentation among female C57BL/6J mice 17,18 , but very few effects observed among males 10,23 , although there was a non-signi cant trend for increased pro-in ammatory gene expression in hippocampus in one study 12 . Reasons for this discrepancy are unclear, but could involve differences in how hippocampus was extracted from the brain in these different studies (whole vs. partial dissection).

Conclusions
Sleep fragmentation and other forms of perturbed sleep promote an in ammatory environment that predisposes individuals towards the development of chronic disease 2,3 . However, the mechanisms that lead to the onset of in ammation from sleep fragmentation are poorly understood. Does in ammation begin in the brain or the periphery? In this time-course study, we provide evidence that the heart is the rst organ to produce elevated pro-in ammatory gene expression, which suggests that in ammation from sleep fragmentation is rapidly initiated in the periphery. Because glucocorticoids are also elevated during sleep fragmentation, our ndings imply that glucocorticoids may rapidly suppress in ammatory responses in certain regions of the brain, like hypothalamus, but not in peripheral tissues, such as heart and white adipose tissue. Instead, the rapid activation of the sympathetic nervous system from sleep fragmentation is most likely promoting an in ammatory environment in peripheral tissues, while possibly overriding negative feedback from glucocorticoids, albeit further study is required. Understanding the time course of in ammatory responses to sleep fragmentation could lead to new therapeutic options for patients suffering from disrupted sleep to prevent the development of cardiovascular and metabolic syndromes. Figure 1 Experimental protocol for acute sleep fragmentation (ASF) time-course study. Mice were exposed to 1, 2, 6, 12, or 24 h of ASF, which involves a sweeping bar that moves horizontally across a modi ed cage every 120 sec. Controls (no sleep fragmentation (NSF)) mice experienced no bar movement.

Figure 2
Duration of acute sleep fragmentation (ASF) alters baseline glucocorticoid and IL-6 levels in serum. A) Corticosterone (cort) concentration in male mice subjected to acute SF (0, 1, 2, 6, 12, and 24 h of ASF or no SF (NSF)). Samples sizes are n= 9-10 per group. B) IL-6 levels in male mice subjected to acute SF (0, 1, 2, 6, 12, and 24 h of ASF or no SF (NSF)). Samples sizes are N = 9-10/group. Signi cant effect of ASF (*** and **** denote p < 0.001 and 0.0001, respectively) relative to NSF at each time point was determined by two-way ANOVA followed by Bonferroni multiple comparisons post hoc tests. Differing lowercase and upper-case letters denote p < 0.05 for NSF and ASF groups, respectively and were analyzed using a one-way ANOVA and Tukey's HSD post hoc tests. Bar plots shown as means ± 1 SE and p was set at 0.05 for statistical signi cance.

Figure 3
Effects of sleep fragmentation, time, and their interaction on TNFα and Il-1β gene expression in peripheral tissues. Panels show TNFα and Il-1β gene expression in spleen (A, B), heart (C, D), epididymal white adipose tissue (EWAT; E, F) and liver (G, H) (n = 8-10/group, time-course: 0, 1, 2, 6, 12, 24h). Signi cant effect of sleep treatment (** denotes p < 0.01) at each time point was determined by two-way ANOVA followed by Bonferroni multiple comparisons post hoc test. Differing lowercase and upper-case letters denote p< 0.05 across different time points for NSF and ASF, respectively and were analyzed using a oneway ANOVA and Tukey's HSD post hoc tests. Bar plots shown as means ± 1 SE and p was set at 0.05 for statistical signi cance.

Figure 4
Effects of sleep fragmentation, time, and their interaction on TNFα and Il-1β gene expression in brain, PFC (A, B), HIP (C, D), and HYP (E, F) (n = 9-10/group, Acute SF time-course: 0, 1, 2, 6, 12, 24h). Signi cant effect of SF (*, ** and *** denote p < 0.05, 0.01, and 0.001, respectively) was determined by two-way ANOVA followed by Bonferroni multiple comparisons post hoc test. Differing lowercase and upper-case letters denote p < 0.05 across different time points for NSF and ASF, respectively and were analyzed using a one-way ANOVA and Tukey's HSD post hoc tests. Bar plots shown as means ± 1 SE and p was set at 0.05 for statistical signi cance. HIP: Hippocampus; PCF: prefrontal cortex; HYP: Hypothalamus.