PATHOLOGICAL EFFECTS OF OBSTRUCTIVE APNEAS DURING THE SLEEP CYCLE IN AN ANIMAL MODEL OF CEREBRAL SMALL VESSEL DISEASE

Abstract

We tested the hypothesis that apneas during the sleep cycle exacerbate hypertension and accelerate changes that occur with cerebral small vessel disease (CSVD). Obstructive sleep apnea was modeled by intermittent inflations of a tracheal balloon to occlude the airway during the sleep cycle (termed OSA) in spontaneously hypertensive stroke prone (SHRSP) rats, a model of CSVD. SHRSP rats and their parent strain, Wistar Kyoto (WKY) rats, were exposed to OSA for two weeks (from 9–11 or 18–20 weeks). At 9 weeks, hypertension was developing in the SHRSP rats and was firmly established by 18 weeks. OSA exposure increased systolic blood pressure in SHRSP rats ~30 mm Hg in both age groups compared to shams that were surgically prepared but not exposed to OSA (p<0.05). OSA exposure also increased systolic blood pressure in WKY rats by 20 and 37 mm Hg at 11 and 20 weeks respectively (p<0.05). OSA exposure in SHRSP rats compromised blood-brain barrier (BBB) integrity in white matter at both 11 and 20 weeks of age compared to SHRSP sham rats (p<0.05). Microglia were activated in SHRSP rats exposed to OSA but not sham rats at 11 weeks (p<0.05). At 20 weeks, microglia were activated in sham SHRSP rats (p<0.05) compared to WKY sham rats and were not further activated by OSA. Neither was BBB integrity altered nor microglia activated in any of the WKY groups. We conclude that OSA accelerates the onset of the cerebral pathologies associated with CSVD in SHRSP, but not WKY rats.

Introduction

Obstructive sleep apnea (OSA) is a pathological condition in which the upper airway collapses during sleep to partially (hypopnea) or completely (apnea) restrict the movement of air into the lungs during inhalation (1). With the onset of hypopnea or apnea, the ventilatory effort increases producing arousal to a lighter stage of sleep where a patent airway can be reestablished (2). In extreme cases, the airway can repeatedly collapse throughout sleep at rates exceeding 100 per hour. Each episode of apnea produces progressive hypoxia, progressive hypercapnia, arousal from sleep, and a negative intrathoracic pressure as the breathing effort continues against a closed airway (1).

OSA is an independent risk factor or closely associated with a number of cardiovascular (including cerebrovascular) diseases (1;3–9). However, the role of OSA in the development of one cerebrovascular disease, cerebral small vessel disease (CSVD), has not been extensively studied. CSVD is characterized by pathology of the small penetrating vessels supplying subcortical structures, leading to blood-brain barrier (BBB) disruption, neuroinflammation, lacunar infarcts, microbleeds, and cognitive dysfunction (10–13). Although studies have not directly focused on OSA as a potential contributor to CSVD development or progression per se, several have attempted to determine if OSA is associated with CSVD-like injuries to the brain; that is brain injuries that can, but not necessarily, result from CSVD. Even then, some find an association between OSA and CVSD-like brain pathologies (14–18) while others find no association (19–22).

In order to better understand the relationship between OSA and CSVD, we studied spontaneously-hypertensive stroke prone rats (SHRSP), a CSVD model that closely mimics the pathological events occurring in human CSVD (23). We tested the hypothesis that apneas during the sleep cycle exacerbate hypertension and accelerate pathological changes that occur with CVSD. We used a model of OSA where a balloon, chronically implanted in the trachea, was remotely inflated to produce 60 apneas/hour (considered severe in humans) for eight hours of the sleep cycle (24;25). We evaluated rats for CSVD-related pathological changes including increases in systolic blood pressure, disruption of the BBB, neuroinflammation, white matter integrity, and cognitive deficits. We show that OSA in SHRSP rats leads to exacerbated hypertension, decreased blood brain barrier integrity, microglia activation, and cognitive impairment.

Methods

A complete description of the methods can be found in the Online Supplement. All studies were approved by the Institutional Animal Care and Use Committee at the Baylor College of Medicine. Male inbred SHRSP and WKY rats were maintained on a regular rodent chow diet and housed in a satellite facility with 12-hours light (0600 to 1800 h) and 12-hours dark (1800 to 0600 h).

At either 8 or 17 weeks of age, rats were prepared with a chronically implanted intra-tracheal balloon that could be remotely inflated to occlude the trachea in order to model OSA (24;25). One week was allowed for recovery from the surgical procedure. Rats were subjected to repeated apneas (60 per hour for eight hours during the sleep cycle, with each apnea lasting 10 sec). The repeated apneas continued during the sleep cycle for two weeks. Sham rats were instrumented with the balloon implants but did not undergo any apneas.

Systolic blood pressure was measured using tail cuff plethysmography. BBB integrity was assessed by the extravasation of Evans blue dye after an IV injection and by extravasation of IgG, a protein found only in plasma when the BBB is intact. Astrocytes were visualized with an antibody directed against glial fibrillary acidic protein and microglia were visualized using an antibody directed against ionized calcium-binding adapter molecule 1. Morphometric analysis was used to determine the activation state of microglia as previously described (26). The Kluver Barrera method was used to visualize white matter from brain sections. The novel object recognition test (27;28) was used to determine working memory by measuring the ability of each rat to discriminate between a “novel object” and an object previously presented to the rat.

Parametric data are expressed as mean ± s.e.m and non-parametric data are presented using box and whisker plots with median values. Differences between groups were determined using two-way repeated measures ANOVA or ANOVA on ranks followed by a Holm-Sidak test when appropriate. Differences were considered as statistically significant if p≤0.05.

Results

Figure 1 shows systolic blood pressures for SHRSP and WKY rats before and after two weeks of OSA or sham. Systolic blood pressure was significantly greater in the SHRSP at all times compared to that in the WKY. Hypertension was still developing in the SHRSP rats at 9 weeks (Figure 1A), as noted in the SHRSP shams, but had plateaued by 18 weeks of age (Figure 1B; also see supplemental Figure S1). In both strains and at both ages, OSA significantly increased systolic blood pressure. Systolic blood pressure increased 20 and 37 mm Hg at 11 and 20 weeks in the WKY after OSA compared to the corresponding WKY sham rats respectively. In the SHRSP, systolic blood pressures were 28 and 32 mm Hg greater in OSA rats compared to shams at 11 and 20 weeks respectively. Thus, OSA increased systolic blood pressure in both strains in both age groups.

Figure 1.

Systolic blood pressure in SHRSP and WKY rats before and after 14 days of OSA or sham, from 9 to 11 weeks old (A) or from 18 to 20 weeks old (B). n=5–10 for each experimental group. *p<0.05, **p≤0.01 and ***p<0.001 versus corresponding sham.

Weight changes for all groups of rats is shown in Supplemental Figure S2. BBB permeability as measured by Evans blue extravasation after an IV infusion is shown in Figure 2. In the SHRSP rats, Evans blue extravasation was significantly increased after OSA in white matter at 11 weeks compared to sham (Figure 2B). Although not statistically significant, the data suggest a potential increase in extravasation at 20 weeks in white and grey matter in SHRSP rats undergoing OSA (p= 0.14 and 0.08 respectively). There were no significant differences in Evans blue extravasation in WKY as a result of OSA in either age group (Figure 2A).

Figure 2.

Blood brain barrier integrity as assessed by Evan’s blue dye extravasation following two weeks of sham or OSA in WKY (A) and SHRSP rats (B). Samples were obtained from white matter (White M), striatum, and grey matter (Grey M). Rats were subjected to OSA or sham from 9–11 or 18–20 weeks. n=6–7 for each experimental group. *p<0.05 versus corresponding SHRSP sham.

Supplemental Figure S3 shows extravasation of IgG from arterioles in OSA and sham for both strains and at both ages. Note the IgG (green fluorescence) surrounding the arterioles in the SHRSP rat after OSA at 11 weeks with even greater extravasation shown in the same group at 20 weeks of age. Figure 3 summarizes the results for IgG extravasation. No IgG extravasation occurred in SHRSP sham rats at 11 weeks of age or in WKY rats in any of the groups. Although not statistically significant, there is a suggestion for BBB disruption in SHRSP after OSA at 11 weeks of age (Supplemental Figure S3 and Figure 3). By 20 weeks of age the SHRSP were showing BBB disruption in the sham group and OSA increased this even further.

Figure 3.

Quantitation of IgG-positive arterioles in a coronal section (see Supplemental Figure S3). n=7–10. *p<0.05 versus corresponding SHRSP at 11 weeks and **p<0.05 versus SHRSP sham at 20 weeks and OSA at 11 weeks.

Supplemental Figure S4 shows microglia (green) at 11 weeks of age. Microglia activation was assessed using a categorical scale from 1 to 4 where 1 was unactivated microglia and 4 was highly activated (26). Figure 4 shows a boxplot summary of the results for microglia activation. In the cingulate gyrus, microglia activation was significantly greater in the SHRSP sham rats at 11 weeks compared to WKY rats at the same age. The microglia were further activated, but not significantly, by OSA in the SHRSP at 11 weeks. Microglia were significantly activated by OSA in the corpus callosum and external capsule at 11 weeks in the SHRSP rats when compared to SHRSP sham rats or WKY sham rats. At 20 weeks of age all three white matter regions showed increased microglia activation in the SHRSP sham rats compared to WKY, and OSA did not further increase the microglia activation. WKY rats did not show an increase in microglia activation resulting from OSA in either age group.

Figure 4.

State of microglia activation in SHRSP sham, SHRSP OSA, WKY sham, and WKY OSA in three white matter regions. *p<0.05 versus WKY groups at the same time point; **p<0.05 compared to all other groups.

Supplemental Figure S5 shows that OSA did not produce any quantifiable decrease of myelin basic protein (Western blotting) with either aging from 9 to 20 weeks or OSA treatment. Additionally, we did not observe any astrogliosis as determined using GFAP fluorescence (Supplemental Figure S6). OSA did not produce disruption of the white matter tracks as determined by the Kluver Barrera staining method in either strain regardless of age or treatment (Supplemental Figure S7A). However, we did see white matter damage in SHRSP rats when 1% NaCl was added to the drinking water from age 12 to 17 weeks with OSA occurring from 18 to 20 weeks (Supplemental Figure S7B). These salt loaded SHRSP rats served as a positive control for white matter damage.

Figure 5 shows results of the novel object recognition test in SHRSP sham and OSA rats at 11 and 20 weeks of age. Neither the WKY OSA nor WKY sham rats explored and, thus, were not included in the results of the novel object recognition test. Note that during the training session rats tended to spend equal time between objects as indicated by the Discrimination Ratio near 0. The SHRSP sham rats at both ages spent more time with the novel object during the working session as indicated by a positive Discrimination Ratio (~0.3 for both age groups). However, after OSA the SHRSP in both age groups did not recognize the novel object as indicated by the Discrimination Ratios near 0.

Figure 5.

Novel object recognition, a measure of working memory, in the SHRSP sham and SHRSP OSA rats at 11 weeks and 20 weeks of age. Rats were subjected to a “Training session,” where two identical objects were placed in front of the rats for 3 min. Time spent at each object was recorded. The two objects were subsequently removed and 1 min later the rats were subjected to a “Working session” where an object with identical characteristics to the original objects used in the “Training session” and one novel object were placed in front of the rats. Time spent at each object was again recorded. With normal working memory, rats tend to spend more time exploring the novel object as indicated by a positive “discrimination ratio.” A discrimination ratio of 0 indicates equal time spent between two objects. n=5–8 for each experimental group. *p<0.01 versus “Working” memory compared to corresponding sham.

Discussion

OSA initiates a pathological cascade that consists of sympathetic activation, inflammation, oxidative stress, endothelial dysfunction, and metabolic disorders (1). Given that this cascade has similarities to the risk factors associated with CSVD (12;13;17;29), we tested the hypothesis that apneas during the sleep cycle exacerbate hypertension and accelerate pathological changes that occur with CVSD. The specific neuropathological changes included neuroinflammation, BBB integrity, white matter integrity, and cognitive function. If our hypothesis is valid, then two weeks of OSA (9–11 weeks) in 11 week old SHRSP rats should produce similar pathological changes that occurred with the aging process from 11 to 20 weeks in SHRSP sham rats (i.e., not undergoing apneas). Consistent with our hypothesis, we demonstrated both aging and OSA in SHRSP rats increased systolic blood pressure (Supplemental Figure S1 and Figure 1), accelerated breakdown of the BBB (Figures 2 and 3, and Supplemental Figure S3), and activated microglia (Figure 4 and Supplemental Figure S4). At 20 weeks of age, OSA further damaged the BBB and further increased systolic blood pressure. However, microglia were already activated by 20 weeks in sham SHRSP rats and were not further activated after being exposed to OSA (Figure 4). Additionally, we observed that SHRSP rats exposed to OSA had cognitive deficits that were not observed in sham rats at either age (Figure 5). Thus, cognitive changes occurring with OSA in SHRSP rats exceeded any cognitive changes, if they occurred, between the age of 9 and 20 weeks in sham SHRSP rats. The above changes resulting from OSA in SHRSP rats are consistent with our hypothesized changes.

We did not observe damage to deep brain structures including disorganization or disarray of white matter tracts, white matter loss, microinfarcts, microbleeds, or astrogliosis in any SHRSP rats regardless of age or treatment (Supplemental Figures S5, S6, and S7). As the complications of cerebral small vessel disease progress in SHRSP or in humans, white matter is damaged and ultimately lost (12;13;23;29). Histologically, white matter tracks can be seen to be disturbed and myelin basic protein, a major constituent of white matter, can be significantly decreased.

SHRSP rats on a conventional diet begin to develop subcortical lesions and loss of white matter starting around 20 weeks of age or older depending on the colony and/or laboratory conducting the studies (23). However, when placed on a high Na⁺/low K⁺ diet, 1% NaCl in drinking water, and/or perfusion deficits produced by carotid artery ligation, the disease process can be dramatically accelerated to the point where subcortical lesions and loss of white matter occur by 20 weeks (23;29). We were expecting OSA to accelerate the disease process in SHRSP rats (especially at 20 weeks of age) to a similar degree. However, our expectations that subcortical lesions, including disorganization and loss of white matter, would result from OSA did not materialize (Supplemental Figures S5 and S7A) except in rats which were salt-loaded (Supplemental Figure S7B). We speculate that OSA, without salt loading, was not sufficiently severe or sufficiently prolonged to produce deep brain lesions or white matter damage. Salt-loading seemed to provide the additional stress necessary to produce damage to subcortical structures.

In WKY rats, the parent strain for the SHRSP, blood pressures increased in both age groups as a result of OSA exposure (Figure 1); however, no damage to the BBB or activation of microglia was observed (Figures 2, 3, and 4). Unfortunately, the WKY rats would not explore and, as a result, could not be used in the novel object recognition test. We note that when Long Evans rats were exposed to OSA using the same model, we did not observe any increase in blood pressure even when the OSA was extended from 2 weeks to 2 months ((24;25) and unpublished observations). Two apparently normal strains of rats, WKY and Long-Evans, had different blood pressure responses when exposed to OSA.

When taking the results as a whole, OSA in SHRSP rats, an animal model for CSVD, accelerated the onset and exacerbated pathological changes associated with CSVD. Several of our endpoints for this study, including loss of BBB integrity and neuroinflammation, are considered events leading to lesions in subcortical regions of the brain in CSVD (23;30;31).

In this study, we did not attempt to delineate the exact role of the relative increase in blood pressure stemming from OSA as it related to BBB damage and microglia activation. We note that hypertension is an integral part of CVSD in many individuals and may constitute a particular inheritable form of CSVD (12;32;33).

Perspective

Our study provides strong evidence that OSA, at least in an animal model, accelerates the disease process involving CSVD. Our data also suggests that relatively short periods of OSA can exacerbate and accelerate the onset of CSVD. Although translating 2 weeks of OSA in a rat model to an equivalent time in humans is impractical, the results do point to the idea that OSA, even for short periods of time, could accelerate or exacerbate the effects of CSVD especially in those individuals with associated risk factors. The most effective treatment for OSA is continuous positive airway pressure (CPAP) which provides pressure by a mask placed over the nose or nose and mouth to prevent the airway from collapsing. If CPAP is an effective treatment for prevention of CSVD and/or allows reversal of CSVD, then early diagnoses and treatment of OSA is extremely important. One encouraging study, although limited in size, reported that cognitive impairments and white matter damage in individuals suffering from OSA could be reversed after one year of CPAP treatment (15). Given that there are few studies dealing with the potential risks of OSA on the development and/or acceleration of CSVD, and that OSA may be a potential modifiable risk factor or treatment for CSVD, this area deserves more in depth studies.

Novelty and Significance.

What Is New?

• This study demonstrates that obstructive sleep apnea (OSA) accelerates the pathological consequences of cerebral small vessel disease (CSVD).

• Only two weeks of OSA, a relatively short period of time, was needed to significantly increase systolic blood pressure, produce significant damage to the blood-brain barrier, enhance microglia activation, and produce cognitive dysfunction in spontaneously-hypertensive stroke prone rats (SHRSP), a strain that recapitulates the pathological changes occurring with human CSVD.

• OSA was modeled in the SHRSP rats and their parent strain, WKY, by remotely inflating a balloon, chronically implanted in the trachea, during eight hours of the sleep cycle each day.

What Is Relevant?

• The exact cause and effect relationship between OSA and cardiovascular diseases has been difficult to fully understand in humans due to the heterogeneity of the study population in terms of confounding comorbidities, genetic makeup, age, life styles, and gender.

• Animal models of OSA, where the confounding comorbidities can be more tightly controlled, have been instrumental in overcoming the heterogeneity of human subjects being studied.

• While OSA appears to be an independent risk factor for stroke and hypertension, its relationship to other cardiovascular diseases, such as CSVD, is less clear.

• We utilized an animal model to demonstrate that OSA accelerates the disease process involving CSVD and only relatively short periods of time, where apneas occur during the sleep cycle, are needed to exacerbate and accelerate the progression of CSVD.

Summary.

OSA accelerates the onset of the cerebral pathologies associated with cerebral small vessel disease in SHRSP rats, a model for human CSVD. Our study provides strong evidence that OSA accelerates the disease process involving CSVD and only relatively short periods of OSA are needed. In humans, early diagnosis and treatment for OSA is especially important in individuals suffering from CSVD.