Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Feb 1.
Published in final edited form as: Neuroscience. 2018 Dec 7;398:55–63. doi: 10.1016/j.neuroscience.2018.11.040

Chronic intermittent hypoxia induces robust astrogliosis in an Alzheimer’s disease-relevant mouse model

Teresa Macheda 1,2, Kelly Roberts 1,2, Danielle N Lyons 1,2, Emma Higgins 1,2, Kyle J Ritter 1,2, Ai-ling Lin 3,4, Warren J Alilain 1,2, Adam D Bachstetter 1,2,*
PMCID: PMC6402802  NIHMSID: NIHMS1516053  PMID: 30529693

Abstract

Sleep disturbances are a common early symptom of neurodegenerative diseases, including Alzheimer’s disease (AD) and other age-related dementias, and emerging evidence suggests that poor sleep may be an important contributor to development of amyloid pathology. Of the causes of sleep disturbances, it is estimated that 10% - 20% of adults in the United States have sleep-disordered breathing (SDB) disorder, with obstructive sleep apnea accounting for the majority of the SBD cases. The clinical and epidemiological data clearly support a link between sleep apnea and AD; yet, almost no experimental research is available exploring the mechanisms associated with this correlative link. Therefore, we exposed an AD-relevant mouse model (APP/PS1 KI) to chronic intermittent hypoxia (an experimental model of sleep apnea) to begin to describe one of the potential mechanisms by which SDB could increase the risk of dementia. Previous studies have found that astrogliosis is a contributor to neuropathology in models of chronic intermittent hypoxia (IH) and AD; therefore, we hypothesized that a reactive astrocyte response might be a contributing mechanism in the neuroinflammation associated with sleep apnea. To test this hypothesis, 10-11-month-old wild type (WT) and APP/PS1 KI mice were exposed to 10 hours of IH, daily for four weeks. At the end of four weeks brains were analyzed from amyloid burden and astrogliosis. No effect was found for chronic IH exposure on amyloid-beta levels or plaque load in the APP/PS1 KI mice. A significant increase in GFAP staining was found in the APP/PS1 KI mice following chronic IH exposure, but not in the WT mice. Profiling of genes associated with different phenotypes of astrocyte activation identified GFAP, CXCL10, and Ggta1 as significant responses activated in the APP/PS1 KI mice exposed to chronic IH.

Keywords: sleep-disordered breathing, sleep apnea, astrocytes, neuroinflammation, Alzheimer’s disease

Introduction

Sleep disturbances are an early symptom of a neurodegenerative disease, and maybe a contributing factor in the development of neurodegeneration, including Alzheimer’s disease (AD) (Daulatzai, 2015; Guarnieri and Sorbi, 2015; Jha et al., 2018; Peers et al., 2009). For instance, clinical and epidemiological data support a link between sleep-disordered breathing and age-related cognitive decline and dementia (Abner et al., 2015; Beebe and Gozal, 2002; Chen et al., 2016; Gelber et al., 2015; Lucey and Holtzman, 2015; Osorio et al., 2015). In the United States, the prevalence of sleep-disordered breathing affects an estimated 10-20% of the adult population and is even greater in the aged population (30-80%) (Osorio et al., 2015; Roepke and Ancoli-Israel, 2010). Postmortem investigation of patients with sleep apnea found increased microinfarcts, neuroinflammation and neuronal loss compared to the patients who did not have a recorded history of sleep apnea (Gelber et al., 2015).

There has been very little experimental research directed at understanding the mechanism associated with this correlative link between sleep-disordered breathing and the increased risk of dementia. A key component observed in both neurodegenerative diseases and sleep apnea is neuroinflammation (Daulatzai, 2015). Neuroinflammation has been seen in people with sleep apnea (Gozal, 2009; Gozal and Kheirandish-Gozal, 2008; Gozal et al., 2009; Rosenzweig et al., 2015), and in experimental models of the disease (Sapin et al., 2015; Smith et al., 2013). Accumulating experimental evidence supports the role of chronic neuroinflammation induced by sleep apnea as a potential mechanism by which sleep apnea contributes to the increased risk of developing a neurodegenerative disease (Kheirandish et al., 2005; Zhang et al., 2010; Zhang et al., 2012; Zhang et al., 2013; Zhu et al., 2012). Recent evidence has suggested that astrocytes may be critically involved in the detrimental aspects of neuroinflammation, and a new nomenclature has been developed defining astrocyte responses as A1-type or A2-type (Clarke et al., 2018; Liddelow and Barres, 2017; Liddelow et al., 2017; Yun et al., 2018). We hypothesized that a reactive astrocyte response might be a contributing mechanism in the neuroinflammation associated with sleep apnea.

Chronic intermittent hypoxia (IH) exposure given during the sleep period of the circadian cycle is a well-established model of sleep apnea (Li et al., 2003; Moreno-Indias et al., 2015; Row et al., 2003; Xu et al., 2004). We have used this model of sleep apnea to mechanistically investigate the correlative link between sleep problems and earlier onset of dementia. In particular, we exposed an APP/PS1 humanized knock-in mouse model to IH and evaluated reactive astrocyte response. Mice were exposed to IH for 10 hours a day, for a total of 4 weeks. Our results showed that IH caused robust astrogliosis in the cortex of APP/PS1 KI mice and an increase in both A1-type and A2-type astrocyte associated markers. We found no effect of IH on amyloid plaque load in the neocortex of mice exposed to 4 weeks of IH.

Materials and Methods

Animals

Adult mice (10-11 months old) APPNLh/NLh × PS-1P264L/264L (APP/PS-1 KI) (Flood et al., 2002) (n=16; 7 F and 9 M) a double gene-targeted knock-in mouse on a C57BL/6J background and C57BL/6J (n=22; 10 F and 12 M, from Jackson) males and females were used in this experiment. Mice were assigned randomly to groups before the start of the experiment. Animals were housed in a controlled humidity and temperature environment and 12/12-h light/dark cycle with free access to food and water. Body weight was measured once a week during the experiment. All procedures were approved by the Institutional Animal Care and Use Committee of the University of Kentucky, and that were conducted in accordance with the standards of proper experimentation in the Guide for the Care and Use of Laboratory Animals and ARRIVE guidelines.

Experimental design

Animals were randomly divided in control (sham) and intermittent hypoxia (IH) groups and placed in two identical commercially designed chambers (Fig. 1A) (BioSpherix Oxycycler A84 unit, W76.2 × D50.8 × H50.8 cm, Bioinstruments, Redfield, NY) that were able to accommodate up to 7 cages with a maximum of 5 mice in each cage. The two chambers were connected to supplies of pure O2 and N2 that were used to produce changes in O2 concentration and regulated by the BioSpherix computer software. During the exposure periods, the concentration of O2 was cyclically reduced from 21% to 10% in the IH group. The mice were left in the chambers for a total of 4 weeks and exposed to intermittent hypoxia (IH group) or normoxic condition (sham group) for a total of 10 hours a day (7 am- 5 pm) Fig. 1A, followed by a 21% oxygen concentration between 5pm and 7 am. Sodasorb CO2 absorbent (Grace Materials Technologies, Discovery Sciences, Chicago, IL, USA) was used during the entire duration of the experiment to remove excess of CO2 in the chamber, silicagel orange (Fisher Scientific, cat # AC392030050) nontoxic grade was used for drying purposes. Cages were changed weekly after 5 pm and the weight of the mice was recorded. Room and chamber temperature and humidity were recorded as well (average temperature: IH = 20.1°C ± 0.38°C; Sham = 20.6°C ± 0.47°C; housing room = 20.5°C ± 0.42°C; mean ± SD) (average humidity: IH = 39.3% ± 14.7%; Sham = 63.9% ± 11.3%; housing room = 38.6% ± 10.2%; mean ± SD).

Figure 1: Experimental overview.

Figure 1:

Open in a new tab

(A) Aged matched WT or APP/PS1 KI (KI) mice were exposed to intermittent hypoxic (IH) or normoxic (sham) conditions by altering the injection of oxygen or nitrogen into a BioSpherix Oxycycler chamber. (B) A representative trace from the BioSpherix Oxycycler, show the program used for adjusting the O2 levels in the camber in gray dashed line. The actual O2 saturation as monitored by the BioSpherix Oxycycler chamber is shown in the solid magenta line. (C) Changes in body weight were seen after 4 weeks of exposure to chronic IH exposure. Each circle represents an individual animal.

The IH protocol (Fig. 1B) consists of 4-minute cycles of 2 min of hypoxia (10% O2) and 2 min of normoxia (21 % O2). Hypoxia is initiated by injecting N2 into the chamber, resulting in a low point of 10 % O2 in 2 min, then O2 is injected over 2 min, resulting in 21% O2 within 2 min. Each cycle lasted 4 min, for a total of 15 cycles/hour and 150 cycles/day. Sham mice were exposed to normoxic gas in a chamber identical to the IH group. For the remaining 14 hours, O2 concentration was kept at 21% in both IH and sham group.

Immunohistochemical and Biochemical analysis

Mice were deeply anesthetized with 5% isoflurane and then transcranial perfused with ice-cold PBS for 5 min. The brains were rapidly removed and dissected. Left hemibrain was fixed in 4 % PFA and used for immunohistochemistry (IHC), the right was dissected and flash frozen for biochemical analysis. IHC staining was done following established methods, and quantified using the Aperio ScanScope XT digital slidescanner or Zesis AxioScan Z. 1 and HALO software (Indica labs) or Aperio ImageScope software positive pixel count algorithm (version 9), as previously described (Bachstetter et al., 2012; Bachstetter et al., 2013). Primary antibodies used included mouse biotin anti-β-Amyloid, 1-16 antibody (6E10 Aβ) (Covance cat#39340-200; (1:3000) Biolegend, San Diego, CA) and rabbit anti-glial fibrillary acid protein (GFAP) (Dako Cat#Z0334; (1:10000); Dako, Glostrup, Denmark). GFAP stained tissue was double labeled with 1% Thioflavin S (1326-12-1, Sigma Aldrich). Aβ1–40 and Aβ1–42 were measured by V-Plex ELISA from Meso Scale Discovery (MSD) according to the manufacturer’s instructions as we have previously described (Bachstetter et al., 2012). Transcriptional changes in astrocyte were measured in cortex homogenates using a customized low-density TaqMan gene expression array card according to the manufacturer’s instructions as previously described (Webster et al., 2015). The following TaqMan Assays were included on the array card: Amigo2- Mm00662105_s1, B3gnt5-Mm01952370_u1, C3-Mm01232779_m1, Cd109-Mm00462151_m1, Cd44-Mm01277161_m1, Clcf1-Mm01236492_m1, Clu-Mm01197002_m1, Cxcl10-Mm00445235_m1, Emp1-Mm00515678_m1,Fbln5-Mm00488601_m1, Fkbp5- Mm00487406_m1, Gbp2-Mm00494576_g1, Gfap-Mm01253033_m1, Ggta1-Mm01333302_m1, H2-T23-Mm00439246_g1, Iigp1-Mm00649928_s1, Lcn2-Mm01324470_m1, Psmb8-Mm00440207_m1, Ptgs2-Mm00478374_m1, Ptx3-Mm00477268_m1, S100a10-Mm00501458_g1, Serpina3n-Mm00776439_m1, Serping1-Mm00437835_m1, Slc10a6-Mm00512730_m1, Sphk1-Mm00448841_g1, Stat3-Mm01219775_m1, Steap4- Mm00475405_m1, Tgm1-Mm00498375_m1, Timp1-Mm01341361_m1, Tm4sf1-Mm00447009_m1, Vim-Mm01333430_m1. All endpoint analyses were conducted on coded samples, that were randomized, and the individuals involved in the endpoint analysis were blind to the experimental groups of the samples. All samples were processed and analyzed together in one batch for each endpoint.

Statistics

JMP Software version 12.1.0 (SAS Institute, Cary, NC, USA) was used for statistical analysis. Graphs were generated using GraphPad Prism version 7.0. Differences between mean were considered statistically significant at α=0.05. The normality assumption was assessed using the Shapiro-Wilk test, and heterogeneity of variances was assessed with Levene’s test. When the assumptions for parametric tests were violated the data was either transformed, or nonparametric tests were used. Changes in body weight between sham and IH were compared using an unpaired t-test. The effect of IH exposure on amyloid load in the APP/PS1 KI mice was compared using an unpaired t-test. A two-way ANOVA was used to determine the effect of genotype and IH exposure for changes in GFAP staining in the cortex. Differences in GFAP staining between groups was compared using a Tukey-Kramer test. Response screening in JMP was used to identify genes that had a statistically significant difference between groups (WT + sham, WT + IH, KI + sham, KI + IH) following false discovery rate (FDR) correction. Differences in gene expression changes between groups were compared using a Tukey-Kramer test. Values are expressed as mean ± SEM unless otherwise noted. Scatter plots represent individual mice.

Results

Two identical chambers were used to expose wild-type (WT) and APP/PS1 KI (KI) mice to normoxic or to intermittent hypoxia, which is an experimental model of the hypoxic component of sleep apnea (Fig. 1). Overall, no adverse events, such as increased aggression or mortality, occurred over the 4 weeks of chronic IH exposure. Following 4 weeks of chronic IH exposure, a decrease in body weight was found for WT (t(20) = 2.70, p ≤ 0.014) and KI mice (t(14) = 1.78, p ≤ 0.097) compared to the sham matched controls (Fig. 1C), which is consistent with previous studies (Carreras et al., 2012; Thomas et al., 2017; Wang et al., 2018). The change in body weight that occurred over the course of the experiment was not sufficient to warrant premature stoppage of any of the mice from the experiment.

Chronic Intermittent hypoxia exposure did not change amyloid plaque load

The APP/PS1 KI mice begin to deposit amyloid plaques at 9 months of age (Flood et al., 2002; Murphy et al., 2007). By 11-12 months of age amyloid plaques are evident in the neocortex and to a lesser extent in the hippocampus (Bachstetter et al., 2012). Therefore, to determine if chronic IH affected amyloid plaque burden histological quantification was evaluated in multiple brain regions using 6E10 antibody. In the WT mice no 6E10+ amyloid plaques were observed. Cerebral amyloid angiopathy was not observed. As shown in Figure 2A, amyloid plaques were seen, as expected, in the APP/PS1 KI mice. Quantification of the amyloid load using digital neuropathology (Fig. 2B) was done to determine the area of the neocortex, corpus callosum, subiculum, CA1/2, dentate gyrus, and thalamus that was 6E10+. Exposure to 4 weeks of IH did not increase amyloid plaque burden in the KI mice compared to the KI + sham mice (Fig. 2C). To determine whether determine if chronic IH affected levels or distribution of Aβ, we prepared PBS soluble, and FA soluble fractions of cortex and measured Aβ40 and Aβ42 by quantitative Aβ ELISA. Exposure to 4 weeks of IH had no effect on Aβ40 (Fig. 2D) or Aβ42 levels (Fig. 2E).

Figure 2: 6E10+ amyloid plaque load in the cortex.

Figure 2:

Open in a new tab

(A) Example photomicrographs of 6E10 immunohistochemical staining in the cortex in 11-12 month old APP/PS1 KI (KI) mice following 28 days of chronic intermittent hypoxia exposure (IH). (B) Example of the computergenerated markup created by the positive pixel algorithm demonstrating the ability of the algorithm to accurately detect the immunostaining. Colors, yellow < orange < red, indicate staining intensities low < medium < high, respectively. (C) Quantification of the 6E10+ staining found that IH exposure did not increase histological amyloid plaque load in the neocortex, corpus callosum, subiculum, CA1/2, dentate gyrus, and thalamus. Quantitative Aβ ELISA of PBS soluble and FA soluble found no effect of IH exposure on Aβ40 (D) or Aβ42 in the neocortex (E). Each circle represents an individual animal.

Chronic Intermittent hypoxia exposure caused robust astrogliosis

GFAP immunohistochemical staining was used to measure reactive astrocytes in the cortex. As shown in Figure 3A, GFAP staining was sparingly found in the cortex of WT + sham mice. Quantitatively, digital neuropathological assessment of GFAP staining was completed in the neocortex, corpus callosum, CA1/2, dentate gyrus, thalamus, and striatum as shown in the heatmap (Fig. 3B). The greatest change in GFAP staining by genotype and IH exposure was found in the neocortex. In the neocortex found an overall main effect by two-way ANOVA (F3,36=58.17; p<0.0001), as well as a significant effect of genotype (p<0.0001), IH exposure (p=0.0006), and interaction of IH and genotype (p=0.0253) (Fig. 3C)

Figure 3. Chronic intermittent hypoxia exposure (IH) cause robust astrogliosis.

Figure 3

Open in a new tab

(A) Low power photomicrographs show GFAP staining. Arrows indicate area shown in Figure 4. (B) A survey of changes in GFAP staining is shown by the heat map across multiple brain region. (C) Quantification of the GFAP staining in the cortex shows the effect of genotype, and the additive effect of IH in the context of the APP/PS1 KI genotype. ‡‡ p<0.01 compated to KI + sham. *** p<0.0001 compared to WT + sham. Each circle represents an individual animal.

Qualitatively, in the WT + IH mice, areas of strong GFAP staining can be observed along blood vessels (Fig. 4). In the KI + Sham injured mice, clusters of GFAP staining can be seen throughout the cortex, most likely corresponding to regions of amyloid plaques. In the KI + IH mice, a substantial increase in GFAP staining was found throughout the cortex. In addition to amyloid plaque associated astrogliosis. A strong reactive GFAP staining was seen around blood vessels in the KI + IH mice. To quantify GFAP staining that appeared to be associated with blood vessels, regions of interest were generated using HALO software (Indica Labs) around GFAP labeled vasculature profiles, and percentage of the neocortex that was positive of blood vessel GFAP staining was quantified. The KI + IH mice were found to have the greatest area of GFAP labeled vasculature profiles in the neocortex; however, we did not find a statistical difference between genotypes or IH exposure (Fig. 4B). Next, we sought to determine if chronic IH exposure increased the density of plaque-associated astrocytes. Tissue sections were double labeled with GFAP and Thioflavin S, and a 200μm radius circle centered on thioflavin S positive staining was generated using HALO software (Indica Labs) (Fig. 5A). Area quantification of GFAP staining in that plaque-associated region of interest found no effect of IH exposure (Fig. 5B). These results indicate that while chronic IH exposure qualitatively appeared to be most associated with blood vessels, the increase in GFAP staining seen following IH exposure was likely an increase in diffuse reactive astrocytes, and not explicitly reactive astrocytes associated with blood vessels or plaques.

Figure 4. Effects of chronic intermittent hypoxia exposure (IH) on GFAP labeled vasculature profiles.

Figure 4

Open in a new tab

(A) high power photomicrographs show GFAP staining in the cortex from areas identified in Figure 3. (B) Quantification of the area of GFAP staining in the cortex associated with vasculature profiles found no statistical differences between genotype, or IH exposure. Each circle represents an individual animal.

Figure 5. Effects of chronic intermittent hypoxia exposure (IH) on plaque-associated astrocytes.

Figure 5

Open in a new tab

(A) A Representative example of the region of interest that was generated around thioflavin S, which was used to quantify the area of GFAP staining that was associated within a 200μm radius from the center of a plaque (B). Each circle represents an individual animal.

Chronic Intermittent hypoxia exposure increased genes associated with A1-type and A2-type astrocytes

Using the recently identified genes that have been shown to be selectively upregulated following either endotoxin or an ischemic injury in astrocytes (Clarke et al., 2018; Liddelow and Barres, 2017; Liddelow et al., 2017; Yun et al., 2018), we developed a custom TaqMan low-density array to determine if those genes were responsive to chronic IH exposure. As shown in the heat map, a pattern of increased expression of pan reactive, A1-type and A2-type genes was found in the KI + IH exposed mice (Fig. 6A). Following FDR correction, five genes were found to have a significant overall effect, including: GFAP (p=0.00003), CXCL10 (p=0.001), B3gnt5 (p=0.0043), Tgm1 (p=0.0043), Ggta1 (p=0.0115). In agreement with the histological results, we found an additive effect of IH exposure and genotype on GFAP gene expression, with the KI+IH mice having the highest level of GFAP gene expression (Fig. 6B). CXCL10 was a second gene previously classified as a pan reactive astrocyte marker that showed a potential synergistic effect on IH and genotype. Similarly, Ggta1, an A1-type marker, was found to be significantly elevated in the KI + IH mice compared to all other groups. Interestingly, the ischemic injury associated gene, Tgm1, was found to be increased by IH in WT mice compared to WT + sham mice. However, there was not an additive effect of IH in the KI mice compared to KI + sham mice on Tgm1 expression in the cortex. Finally, a genotype effect was seen for B3gnt5 gene expression in the cortex with the KI mice showing less B3gnt5 compared to the WT mice, but there was no effect of IH exposure.

Figure 6. Gene expression changes in the cortex of associated reactive astrocyte markers.

Figure 6.

Open in a new tab

(A). Heat map illustrates changes in the pan reactive, A1-type, and A1-type genes in the cortex of mice following 4 weeks of sham or chronic intermittent hypoxia (IH) exposure. The asterisk identifies genes that were statistically significantly different following FDR correction and shown in the scatter plot (B). Each circle represents an individual animal. *p<0.05, **p<0.01, ***p<0.001 compared to WT + sham. §p<0.05, §§p<0.01 compared to WT + IH. ‡p<0.05, ‡‡ p<0.01 compared to KI + sham.

Discussion

We report here four key findings that highlight an underappreciated role of chronic IH in driving reactive astrogliosis in the context of AD-relevant amyloidosis. First, four weeks of chronic IH, with 150 cycles per day of O2 reaching 10% saturation for a few seconds produced robust astrogliosis in the cortex of aged APP/PS1 KI mice. Interestingly, the same IH exposure in age-matched WT mice had only marginal effects on astrocytes. Both WT and APP/PS1 KI mice were found to have a similar change in body weight. Second, in WT mice in the limited areas were increased astrogliosis was seen it was qualitatively associated with blood vessels. Similar reactive astrogliosis was seen around blood vessels in APP/PS1 KI mice exposed to chronic IH; however, the area of astrogliosis around the blood vessels was disproportionally larger in the KI + IH mice compared to the WT mice exposed to chronic IH. Third, profiling of pan, A1-type, and A2-type astrocyte associated markers found that the reactive astrocyte response seen in the KI + IH mice was statistically associated with pan reactive and A1-type markers. In contrast, in WT mice chronic IH exposure selectively increased A2-type astrocyte markers. A2-type astrocyte markers were identified as genes that increased in astrocytes in response to ischemic injury, while A1-type astrocyte markers were defined as genes that were increased in astrocyte in response to LPS endotoxin challenge (Liddelow et al., 2017). The KI + IH mice were found to have an upregulation of the A1-type astrocyte markers, suggesting that an interaction between the Aβ and the IH drives a similar subset of genes as endotoxemia. The implications for changes in the profile of astrocyte associated markers warrants further mechanistic investigation. Finally, we found that four weeks of chronic IH exposure did not increase amyloid plaque load. Overall, our results indicate that chronic IH exposure can drive reactive astrogliosis, but only if the brain is already fragile due to preexisting neuropathology.

A major goal of our study was to develop an animal model to explore how sleep-disordered breathing may contribute to AD-relevant neuropathology. To this end, we used a well-characterized mutant mouse model of amyloidosis. The APP/PS1 KI has a progressive amyloid deposition starting at six months of age that increases linearly over time (Flood et al., 2002; Murphy et al., 2007). A strength of the KI mice is that the endogenous mouse promoter drives the APP and PS1 mutations, and thus maintain a physiologically relevant expression of the APP and PS1 genes. We believe that it is important to consider the limitations of the model systems used in our study when interpreting our findings. A limitation of the APP/PS1 KI model is that it does not have tau pathology or neuronal loss, and thus we were not able to model how sleep-disordered breathing might influence these pathologies. Second, the chronic IH model, used in our study, is a well-established experimental model of sleep apnea (Li et al., 2003; Moreno-Indias et al., 2015; Row et al., 2003; Xu et al., 2004). The chronic IH model simulates the hypoxia-reoxygenation events experienced by patients with sleep apnea. However, the pattern of O2 changes induced in the IH model does not accurately reflect the amplitude and periodicity of those experienced by patients with sleep apnea. The gas exchange in the chamber was a rate-limiting step in both amplitude and periodicity of the O2 changes. While, we could have adjusted the computer program set points to induce a more variable pattern of O2 changes, as seen in people with sleep apnea, we chose to use the more established pattern of IH exposure, as previously published (Li et al., 2003; Moreno-Indias et al., 2015; Row et al., 2003; Xu et al., 2004). Third, hypoxic events could be either neuroprotective or neurotoxic depending on the severity, duration, and frequency of the hypoxia (Beebe and Gozal, 2002; Boroujerdi and Milner, 2015; Kim et al., 2015; Rosenzweig et al., 2015). Previous studies have shown that the model of chronic IH used in our experiments can contribute to neuronal injury (Gozal et al., 2001; Gozal et al., 2003; Smith et al., 2013); but there may be a component of the chronic IH that may also be neuroprotective (Satriotomo et al., 2016).

Growing evidence suggests that chronic IH exposure can accelerates not only amyloid pathology but also tau pathology in an AD mouse model (Gao et al., 2013; Li et al., 2009; Liu etal., 2016; Sun et al., 2006; Yagishita et al., 2017). We found that four weeks of chronic IH exposure did not increase amyloid plaque load. In agreement with our results, Shiota et al. (Shiota et al., 2013) demonstrated that eight weeks of IH exposure in an AD-relevant mouse model was not associated with worsening amyloid plaque deposition in the brain. In contrast to our results, some studies have reported that IH exposure increased amyloid pathology (Gao et al., 2013; Li et al., 2009; Liu et al., 2016; Yagishita et al., 2017). The protocol used in our study is fundamentally different from the groups that saw an increase in amyloid plaque load. For example, to induce hypoxia, some studies used a sealed 125ml jar (Gao et al., 2013; Li et al., 2009), a hypobaric chamber (Liu et al., 2016), or much stronger hypoxia exposure (Yagishita et al., 2017). Moreover, our study was the only one to use mice with a physiological relevant expression pattern of the APP and PS1 genes. Therefore, differences in amyloid pathology could also be a result of the transgenic overexpression patterns used in the studies that found changes in amyloid pathology.

Previous studies have shown that IH exposure in wild-type rodents could induce reactive gliosis (Gozal et al., 2001; Wang et al., 2018; Xavier Aviles-Reyes et al., 2010), and that even one day of IH exposure was sufficient to induce gliosis (Xavier Aviles-Reyes et al., 2010). Our results are in contrast to these studies, as we found that four weeks of IH exposure produced very limited astrogliosis around a subset of blood vessels in WT mice. Our observations suggest that other comorbid pathologies are necessary to see the reactive astrocyte response.

In summation, our results also suggest that the otherwise healthy brain is likely resilient to a short period of IH exposure, but the brain may have less resilience to IH exposure if amyloid pathology is present. In particular, the presence of amyloid pathology can have an additive effect with IH exposure on reactive gliosis. Future studies are warranted to determine mechanistically the detrimental versus protective aspects of the astrocyte response to IH exposure in AD-relevant mouse models.

Acknowledgements:

The corresponding author, Adam Bachstetter, PhD, had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Research reported in this publication was supported by National Institutes of Health under award numbers P30 AG028383 and R00 AG044445 (ADB). The content is solely the responsibility of the authors and does not represent the official views of the funding organizations.

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Author Disclosure Statement: No competing financial interests exist.

References:

  1. Abner EL, et al. , 2015. Baseline subjective memory complaints associate with increased risk of incident dementia: the PREADVISE trial. J Prev Alzheimers Dis. 2, 11–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bachstetter AD, et al. , 2012. Early stage drug treatment that normalizes proinflammatory cytokine production attenuates synaptic dysfunction in a mouse model that exhibits agedependent progression of Alzheimer's disease-related pathology. J Neurosci. 32, 10201–10. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Bachstetter AD, et al. , 2013. The p38alpha MAPK regulates microglial responsiveness to diffuse traumatic brain injury. J Neurosci. 33, 6143–53. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Beebe DW, Gozal D, 2002. Obstructive sleep apnea and the prefrontal cortex: towards a comprehensive model linking nocturnal upper airway obstruction to daytime cognitive and behavioral deficits. J Sleep Res. 11, 1–16. [DOI] [PubMed] [Google Scholar]
  5. Boroujerdi A, Milner R, 2015. Defining the critical hypoxic threshold that promotes vascular remodeling in the brain. Exp Neurol. 263, 132–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Carreras A, et al. , 2012. Metabolic effects of intermittent hypoxia in mice: steady versus highfrequency applied hypoxia daily during the rest period. Am J Physiol Regul Integr Comp Physiol. 303, R700–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chen JC, et al. , 2016. Sleep duration, cognitive decline, and dementia risk in older women. Alzheimers & Dementia. 12, 21–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Clarke LE, et al. , 2018. Normal aging induces A1-like astrocyte reactivity. Proc Natl Acad Sci U S A. 115, E1896–E1905. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Daulatzai MA, 2015. Evidence of neurodegeneration in obstructive sleep apnea: Relationship between obstructive sleep apnea and cognitive dysfunction in the elderly. J Neurosci Res. 93, 1778–94. [DOI] [PubMed] [Google Scholar]
  10. Flood DG, et al. , 2002. FAD mutant PS-1 gene-targeted mice: increased A beta 42 and A beta deposition without APP overproduction. Neurobiol Aging. 23, 335–48. [DOI] [PubMed] [Google Scholar]
  11. Gao L, et al. , 2013. Hypoxia increases Abeta-induced tau phosphorylation by calpain and promotes behavioral consequences in AD transgenic mice. J Mol Neurosci. 51, 138–47. [DOI] [PubMed] [Google Scholar]
  12. Gelber RP, et al. , 2015. Associations of brain lesions at autopsy with polysomnography features before death. Neurology. 84, 296–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gozal D, 2009. Sleep, sleep disorders and inflammation in children. Sleep Med. 10 Suppl 1, S12–6. [DOI] [PubMed] [Google Scholar]
  14. Gozal D, et al. , 2001. Behavioral and anatomical correlates of chronic episodic hypoxia during sleep in the rat. Journal of Neuroscience. 21, 2442–2450. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gozal D, Kheirandish-Gozal L, 2008. Cardiovascular morbidity in obstructive sleep apnea: oxidative stress, inflammation, and much more. Am J Respir Crit Care Med. 177, 369–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gozal D, et al. , 2003. Increased susceptibility to intermittent hypoxia in aging rats: changes in proteasomal activity, neuronal apoptosis and spatial function. J Neurochem. 86, 1545–52. [DOI] [PubMed] [Google Scholar]
  17. Gozal D, et al. , 2009. Plasma IGF-1 levels and cognitive dysfunction in children with obstructive sleep apnea. Sleep Med. 10, 167–73. [DOI] [PubMed] [Google Scholar]
  18. Guarnieri B, Sorbi S, 2015. Sleep and Cognitive Decline: A Strong Bidirectional Relationship. It Is Time for Specific Recommendations on Routine Assessment and the Management of Sleep Disorders in Patients with Mild Cognitive Impairment and Dementia. European Neurology. 74, 43–48. [DOI] [PubMed] [Google Scholar]
  19. Jha NK, et al. , 2018. Hypoxia-Induced Signaling Activation in Neurodegenerative Diseases: Targets for New Therapeutic Strategies. J Alzheimers Dis. 62, 15–38. [DOI] [PubMed] [Google Scholar]
  20. Kheirandish L, et al. , 2005. Apolipoprotein E-deficient mice exhibit increased vulnerability to intermittent hypoxia-induced spatial learning deficits. Sleep. 28, 1412–7. [DOI] [PubMed] [Google Scholar]
  21. Kim LJ, et al. , 2015. Hypomyelination, memory impairment, and blood-brain barrier permeability in a model of sleep apnea. Brain Res. 1597, 28–36. [DOI] [PubMed] [Google Scholar]
  22. Li L, et al. , 2009. Hypoxia increases Abeta generation by altering beta- and gamma-cleavage of APP. Neurobiol Aging. 30, 1091–8. [DOI] [PubMed] [Google Scholar]
  23. Li RC, et al. , 2003. Cyclooxygenase 2 and intermittent hypoxia-induced spatial deficits in the rat. Am J Respir Crit Care Med. 168, 469–75. [DOI] [PubMed] [Google Scholar]
  24. Liddelow SA, Barres BA, 2017. Reactive Astrocytes: Production, Function, and Therapeutic Potential. Immunity. 46, 957–967. [DOI] [PubMed] [Google Scholar]
  25. Liddelow SA, et al. , 2017. Neurotoxic reactive astrocytes are induced by activated microglia. Nature. 541, 481–487. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Liu H, et al. , 2016. Chronic hypoxia facilitates Alzheimer's disease through demethylation of gamma-secretase by downregulating DNA methyltransferase 3b. Alzheimers Dement 12, 130–43. [DOI] [PubMed] [Google Scholar]
  27. Lucey BP, Holtzman DM, 2015. How amyloid, sleep and memory connect. Nat Neurosci. 18, 933–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Moreno-Indias I, et al. , 2015. Intermittent hypoxia alters gut microbiota diversity in a mouse model of sleep apnoea. Eur Respir J. 45, 1055–65. [DOI] [PubMed] [Google Scholar]
  29. Murphy MP, et al. , 2007. Abeta solubility and deposition during AD progression and in APPxPS-1 knock-in mice. Neurobiol Dis. 27, 301–11. [DOI] [PubMed] [Google Scholar]
  30. Osorio RS, et al. , 2015. Sleep-disordered breathing advances cognitive decline in the elderly. Neurology. 84, 1964–1971. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Peers C, et al. , 2009. Hypoxia and neurodegeneration. Ann N Y Acad Sci. 1177, 169–77. [DOI] [PubMed] [Google Scholar]
  32. Roepke SK, Ancoli-Israel S, 2010. Sleep disorders in the elderly. Indian Journal of Medical Research. 131, 302–310. [PubMed] [Google Scholar]
  33. Rosenzweig I, et al. , 2015. Sleep apnoea and the brain: a complex relationship. Lancet Respir Med. 3, 404–14. [DOI] [PubMed] [Google Scholar]
  34. Row BW, et al. , 2003. Intermittent hypoxia is associated with oxidative stress and spatial learning deficits in the rat. Am J Respir Crit Care Med. 167, 1548–53. [DOI] [PubMed] [Google Scholar]
  35. Sapin E, et al. , 2015. Chronic Intermittent Hypoxia Induces Chronic Low-Grade Neuroinflammation in the Dorsal Hippocampus of Mice. Sleep. 38, 1537–46. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Satriotomo I, et al. , 2016. Repetitive acute intermittent hypoxia increases growth/neurotrophic factor expression in non-respiratory motor neurons. Neuroscience. 322, 479–88. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Shiota S, et al. , 2013. Chronic intermittent hypoxia/reoxygenation facilitate amyloid-beta generation in mice. J Alzheimers Dis. 37, 325–33. [DOI] [PubMed] [Google Scholar]
  38. Smith SM, et al. , 2013. Chronic intermittent hypoxia exerts CNS region-specific effects on rat microglial inflammatory and TLR4 gene expression. PLoS One. 8, e81584. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Sun X, et al. , 2006. Hypoxia facilitates Alzheimer's disease pathogenesis by up-regulating BACE1 gene expression. Proc Natl Acad Sci U S A. 103, 18727–32. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Thomas A, et al. , 2017. Chronic Intermittent Hypoxia Impairs Insulin Sensitivity but Improves Whole-Body Glucose Tolerance by Activating Skeletal Muscle AMPK. Diabetes. 66, 2942–2951. [DOI] [PubMed] [Google Scholar]
  41. Wang B, et al. , 2018. Curcumin attenuates chronic intermittent hypoxia-induced brain injuries by inhibiting AQP4 and p38 MAPK pathway. Respir Physiol Neurobiol. 255, 50–57. [DOI] [PubMed] [Google Scholar]
  42. Webster SJ, et al. , 2015. Closed head injury in an age-related Alzheimer mouse model leads to an altered neuroinflammatory response and persistent cognitive impairment. J Neurosci. 35, 6554–69. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Xavier Aviles-Reyes R, et al. , 2010. Intermittent hypoxia during sleep induces reactive gliosis and limited neuronal death in rats: implications for sleep apnea. Journal of Neurochemistry. 112, 854–869. [DOI] [PubMed] [Google Scholar]
  44. Xu W, et al. , 2004. Increased oxidative stress is associated with chronic intermittent hypoxiamediated brain cortical neuronal cell apoptosis in a mouse model of sleep apnea. Neuroscience. 126, 313–23. [DOI] [PubMed] [Google Scholar]
  45. Yagishita S, et al. , 2017. Treatment of intermittent hypoxia increases phosphorylated tau in the hippocampus via biological processes common to aging. Mol Brain. 10, 2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Yun SP, et al. , 2018. Block of A1 astrocyte conversion by microglia is neuroprotective in models of Parkinson's disease. Nat Med. 24, 931–938. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Zhang J-H, et al. , 2010. Apnea produces neuronal degeneration in the pons and medulla of guinea pigs. Neurobiology of Disease. 40, 251–264. [DOI] [PubMed] [Google Scholar]
  48. Zhang SX, et al. , 2012. Pathological consequences of intermittent hypoxia in the central nervous system. Compr Physiol. 2, 1767–77. [DOI] [PubMed] [Google Scholar]
  49. Zhang X, et al. , 2013. Prenatal hypoxia may aggravate the cognitive impairment and Alzheimer's disease neuropathology in APPSwe/PS1A246E transgenic mice. Neurobiol Aging. 34, 663–78. [DOI] [PubMed] [Google Scholar]
  50. Zhu B, et al. , 2012. Sleep disturbance induces neuroinflammation and impairment of learning and memory. Neurobiology of Disease. 48, 348–355. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES