Journal of Applied Physiology Fuel your research with LabChart
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

J Appl Physiol 91: 2758-2766, 2001;
8750-7587/01 $5.00
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (32)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Tagaito, Y.
Right arrow Articles by O'Donnell, C. P.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Tagaito, Y.
Right arrow Articles by O'Donnell, C. P.
Vol. 91, Issue 6, 2758-2766, December 2001

INNOVATIVE TECHNIQUES
A model of sleep-disordered breathing in the C57BL/6J mouse

Y. Tagaito2, V. Y. Polotsky1, M. J. Campen1, J. A. Wilson1, A. Balbir1, P. L. Smith1, A. R. Schwartz1, and C. P. O'Donnell1

1 Division of Pulmonary and Critical Care Medicine, Department of Medicine, Johns Hopkins University, Baltimore, Maryland 21224; and 2 Department of Anesthesiology, Chiba University School of Medicine, Chiba 260, Japan


    ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

To investigate the pathophysiological sequelae of sleep-disordered breathing (SDB), we have developed a mouse model in which hypoxia was induced during periods of sleep and was removed in response to arousal or wakefulness. An on-line sleep-wake detection system, based on the frequency and amplitude of electroencephalograph and electromyograph recordings, served to trigger intermittent hypoxia during periods of sleep. In adult male C57BL/6J mice (n = 5), the sleep-wake detection system accurately assessed wakefulness (97.2 ± 1.1%), non-rapid eye movement (NREM) sleep (96.0 ± 0.9%) and rapid eye movement (REM) sleep (85.6 ± 5.0%). After 5 consecutive days of SDB, 554 ± 29 (SE) hypoxic events were recorded over a 24-h period at a rate of 63.6 ± 2.6 events/h of sleep and with a duration of 28.2 ± 0.7 s. The mean nadir of fraction of inspired O2 (FIO2) on day 5 was 13.2 ± 0.1%, and 137.1 ± 13.2 of the events had a nadir FIO2 <10% O2. Arterial blood gases confirmed that hypoxia of this magnitude lead to a significant degree of hypoxemia. Furthermore, 5 days of SDB were associated with decreases in both NREM and REM sleep during the light phase compared with the 24-h postintervention period. We conclude that our murine model of SDB mimics the rate and magnitude of sleep-induced hypoxia, sleep fragmentation, and reduction in total sleep time found in patients with moderate to severe SDB in the clinical setting.

polysomnography; intermittent hypoxia; arousal; non-rapid eye movement sleep; rapid eye movement sleep


    INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

SLEEP-DISORDERED BREATHING (SDB) can result in significant morbidity and mortality due to a broad range of neurological, cardiovascular, and metabolic sequelae (9, 17, 25, 30, 38, 41). These outcomes are generally thought to be related in large part to repetitive periods of intermittent hypoxemia and the fragmentation of normal sleep architecture. Nevertheless, little data are currently available concerning the relative contributions of intermittent hypoxia and sleep fragmentation on the known sequelae of SDB. Nor have the intermediate pathways impacting on the morbidity and mortality associated with SDB been elucidated.

Various animal models have been developed to study the sequelae of SDB. In large animals, experimental obstruction of the upper airway was most commonly employed to induce SDB. Such models in sleeping dogs (4, 23, 26, 35, 36), sedated pigs (6, 7, 37), and anesthetized baboons (42), dogs (33), and cats (28) have provided important information regarding the cardiovascular and respiratory consequences of SDB. Large-animal models, however, are labor intensive and provide limited mechanistic insight into the cellular and molecular processes that mediate the pathophysiological sequelae of SDB.

To overcome the limitations inherent in large-animal models, small-animal models have been utilized to explore the effects of intermittent hypoxia. In these experiments, groups of rats have been exposed simultaneously to intermittent hypoxia to examine the mechanisms of daytime hypertension in SDB (2, 3, 10-13, 16, 21, 24). In these rat studies, however, intermittent arousals were not modeled because hypoxia was induced independent of sleep-wake state. Thus a small-animal model of sleep-related hypoxemia terminated by arousal is still required to simulate the predominant physiological effects of SDB. Furthermore, if a model could be developed in mice, this would allow unique genetic insights into the mechanisms and pathways that lead to pathophysiological sequelae in SDB.

The purpose of the present study, therefore, was to develop a murine model that simulated the patterns of oxygenation and sleep architecture characteristic of clinical SDB. A number of technical challenges had to be overcome to produce intermittent hypoxia and sleep fragmentation in a murine model of SDB. Specifically, methods were required for 1) long-term maintenance of chronic instrumentation, 2) implementation of an automated on-line sleep-wake detection system, and 3) development of a gas-delivery system that alters the inspired O2 (FIO2) level in response to changes in sleep-wake status. We, therefore, set out to address these issues and determine whether mice subjected to sleep-related hypoxia would exhibit similar patterns in sleep-related hypoxemia and sleep fragmentation as patients with moderate to severe clinical SDB.


    METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Animal Preparation

Five male C57BL/6J mice from Jackson Laboratory (Bar Harbor, ME) weighing 27-30 g were used in the study. The study was approved by the Johns Hopkins University Animal Use and Care Committee and complied with the American Physiological Society guidelines. For all surgical procedures, anesthesia was induced and maintained using halothane administered through a face mask. At the completion of experiments, animals were euthanized with pentobarbital sodium (60 mg ip).

Procedures

Polysomnography. A midline incision was made to expose the skull and muscles immediately posterior to the skull. The underlying fascia was gently cleared from the skull surface, two holes were drilled through the skull in the left frontal and parietal regions, and one hole was drilled through the skull in the right parietal region. Three electroencephalographic (EEG; E363/1, Plastics One, Roanoke, VA) electrodes were fastened via small screws inserted into the three predrilled holes, which approximated an equilateral triangle. The first electrode was placed 2-3 mm caudal to bregma and 1-2 mm to the right of the midline. The second electrode was placed 2-3 mm rostral to bregma and 2-3 mm to the right of the midline. The third electrode was placed 0-1 mm rostral to bregma and 2-3 mm to the left of the midline. Two nuchal electromyographic (EMG; E363/76, Plastics One) electrodes were stitched flat onto the surface of the muscle immediately posterior to the dorsal area of the mouse skull. The three EEG and two EMG electrodes were inserted into a pedestal (MS363, Plastics One) and cemented to the skull with dental acrylic. Each animal was allowed 5-7 days to recover from surgery and then attached via a connector cable (363-SL/6 80CM 6TCS, Plastics One) to preamplifiers.

Femoral artery catheter. In a separate group of four male C57BL/6J mice, an arterial catheter was chronically implanted in either the left or right femoral artery for sampling of arterial blood. The femoral artery was carefully exposed via a 0.5- to 1.0-cm incision and dissected free from the femoral vein, without disrupting the femoral nerve. A 60-cm catheter fashioned from Renothane tubing (MRE025, Braintree Scientific) was inserted 1-2 cm into the femoral artery, glued in place (Quicktite Superglue, Manco, OH), and routed under the skin to exit on the dorsal surface of the neck. The catheter was attached to a single-channel fluid swivel (model 375/25, Instech Laboratories) and perfused slowly by an infusion pump (0.5 ml/day) with a sterile saline solution containing heparin (1,000 U heparin/l saline). Animals were given a minimum of 48 h to recover before any sampling of blood gases.

Arterial blood-gas sampling. As previously described (34), a 100-µl glass syringe (Hamilton) was used to withdraw arterial blood for analysis of blood-gas status. An 80- to 100-µl sample of arterial blood was drawn into the syringe, placed on ice, and immediately analyzed on a blood-gas analyzer (IL BG3, Instrumentation Laboratory, Lexington, MA). In four mice, arterial blood-gas status was assessed under four conditions: 1) room air, 2) 15% O2 for 240 s, 3) 10% O2 for 240 s, and 4) 5% O2 for 90 s. The order of the four conditions was determined by block design, and each mouse successfully completed the four measurements. A maximum of one blood gas per day was obtained from any one mouse.

Animal Maintenance

The instrumentation from the animal was fixed above to a spring counterbalance arm. To allow free movement of the tethered mouse, the chamber that housed the mouse floated on water and could, therefore, rotate in response to the animal's movements. A schematic representation of our approach is shown in Fig. 1.


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1.   Schematic representation of the approach taken to chronically maintain mouse, the automatic detection of sleep-wake state, and the gas-control hardware used to deliver N2 during periods of sleep and to restore airflow at arousal and during periods of sustained wakefulness. EEG, electroencephalograph; EMG, electromyograph; FIO2, fraction of inspired O2.

The base of the chamber was circular and fashioned from 2-in.-thick polystyrene. A hole in the center of the base enabled the delivery of gases from tubing that ran underneath the water. The upper portion of the base was slightly concave in shape and covered with a flat, fine metal mesh and regular bedding material to ensure an even distribution of gases throughout the chamber. The upper portion of the chamber was a 2.0-liter clear plastic cylinder with a detachable lid that slotted into the polystyrene base. A small hole in the center of the detachable lid (0.5-in. diameter) was used to exhaust the gas delivered from below and allow the instrumentation from the mouse to exit the housing chamber. The chamber also contained a water bottle and food holder.

The mouse wore a plastic jacket connected to a spring tether (Instech Laboratories) that projected vertically out the top of the housing chamber and connected to a spring counterbalance arm. The spring tether allowed movement from the mouse to be transmitted with minimal torque to the floating housing chamber, allowing the animal free access to the entire chamber. The polysomnographic leads were loosely attached to the spring tether such that any torque-related movement was borne by the tether system rather than the electrical leads. To sample chamber levels of FIO2, a thin piece of tubing (PE-60) was attached to the spring tether with the tip positioned at a height ~2 cm directly above the mouse. The positioning of the O2-sampling tube in the immediate environment of the mouse enabled accurate determination of the specific gas stimulus experienced by the animal. The polysomnographic leads and the sampling line for FIO2 exited out of a larger holding container to connect to amplifiers. The day-night light-dark cycle inside the holding container was controlled by a timer operating two 4-W light bulbs. We determined in pilot experiments that our approach provided long-term, unrestricted movement for 15-30 days. The mice acclimated to the environment within 24 h, exhibited normal daily food and water consumption, and maintained normal body weight throughout the experiments.

Sleep-Wake Detection System

To develop a murine model of SDB, a computer-controlled automated sleep-wake detection system was implemented. For this purpose, we modified the software developed from an automated canine model of obstructive sleep apnea (23) for use in the mouse. The program uses a combination of the frequency distribution of the EEG and amplitude of the nuchal EMG to determine sleep-wake state. The epoch length was set to 5 s to minimize the time to begin reoxygenation at arousal and to allow sufficient time for the computer to accurately assess sleep-wake states.

Our laboratory has previously reported in mice (34) that wakefulness is characterized by low-amplitude, high-frequency (>10 Hz) EEG waves and high levels of EMG activity compared with the sleep states. NREM sleep is characterized by high-amplitude, low-frequency (~2-5 Hz) EEG waves and an EMG activity considerably less than during wakefulness. REM sleep is characterized by low-amplitude, mixed-frequency (~5-10 Hz) EEG waves, with the EMG activity either equal to or less than that seen during NREM sleep but always less than that seen during wakefulness

Accordingly, four polysomnographic parameters (EEG frequency, 2 EMG amplitudes, and EEG amplitude) were used for on-line computer assessment of sleep-wake states.

1) A threshold was set for the ratio of beta 2 (20-30 Hz) to delta 1 (2-4 Hz) from the EEG frequency distribution to separate NREM sleep (below beta 2/delta 1 threshold) from either wakefulness or REM sleep (above beta 2/delta 1 threshold).

2) Two EMG thresholds were set such that 1) EMG activity above the highest threshold represented wakefulness (irrespective of the beta 2/delta 1 ratio), 2) EMG activity below the lowest threshold represented either NREM or REM sleep depending on whether the beta 2/delta 1 ratio was below (NREM sleep) or above (REM sleep) its threshold, and 3) EMG activity between the high and low thresholds represented either NREM sleep or wakefulness depending on whether the beta 2/delta 1 ratio was below (NREM sleep) or above (wakefulness) its threshold.

3) A final criteria for determination of REM sleep, in addition to the two thresholds above, was that the amplitude of the EEG must also be below a set threshold. We included the amplitude of the EEG signal in the decision tree because the amplitude of the EEG signal in the mouse invariably undergoes a uniform reduction in the transition from NREM to REM sleep.

Each mouse underwent an initialization period in which threshold values for the beta 2/delta 1 ratio, the two EMG thresholds, and the EEG threshold were optimized visually from the polysomnography over several hours by one or more investigator. Once initial values for the thresholds were determined, we recorded polysomnography and computer sleep-wake assessment for a subsequent 24-48 h and made minor adjustments in the thresholds where appropriate. The threshold for the beta 2/delta 1 frequency ratio of the EEG ranged between 1.0 and 2.0 and varied little between animals. The threshold for the EEG amplitude was consistently in the 100- to 150-µV range, whereas the high and low thresholds for EMG activity varied among animals. Once the initialization period was complete, the thresholds were held constant throughout the experimental protocol.

Gas-Control System

When either NREM or REM sleep was detected for three consecutive 5-s epochs, a 5-V signal was output from the computer and used to trigger a series of effectively silent solenoid valves (LFDX05033600AA LIF series, Lee) to control the delivery of gas through the base of the housing chamber containing the mouse (Fig. 1). An axial pattern of gas distribution was employed through the bottom of the housing cage to minimize any regional differences in gas concentration. The output from the computer was changed from 5 V back to 0 V whenever wakefulness was detected. During continuous periods of wakefulness, room air was delivered through the cage at a rate of 2 l/min. After the detection of three consecutive epochs of sleep, the output signal from the computer caused the airflow to cease and the stimulus gas (100% N2) to be delivered at 2 l/min. The N2 gas delivery continued until arousal occurred, at which time the test gas was shut off and room air was once again delivered at a rate of 2 l/min. In addition, for the first 10 s after arousal, an additional solenoid was used to deliver a second source of room air at 3.5 l/min. The additional air source during the first 10 s after arousal was designed to allow reoxygenation to occur at a faster rate than deoxygenation and effectively replicate the timing, pattern, and magnitude of oxyhemoglobin desaturation that characterize human SDB.

Recording Apparatus

A pen recorder (Grass Instruments, Quincy, MA) was used to amplify and filter EEG (3-35 Hz) and EMG (10-75 Hz) activity and a Beckman analyzer (Anaheim, CA) sampled O2 levels from the mouse housing chamber. Both instruments were connected to the pen recorder, and data from the pen recorder were sampled at 300 Hz and converted to digital format (DI-200 data acquisition board, Dataq Instruments, Akron, OH) and acquired to optical disk for storage with Windaq/200 acquisition software (Dataq Instruments, Akron, OH).

Protocol

Animals were given 5-7 days to recover after surgical implantation of the polysomnographic electrodes and pedestal assembly before beginning a 3-day acclimation period in the housing chamber. At the completion of the acclimation period, sleep-wake data were collected for a 24-h control period. After the control period, animals were exposed to 5 consecutive days of sleep-related intermittent hypoxia to induce SDB, followed by a 24-h recovery period in which sleep-wake state was monitored in the absence of hypoxia in an equivalent manner to the control period.

Analyses

Sleep stage assignment. Sleep architecture data and air-N2 solenoid status data (i.e., the third and fifth channels from the top in Figs. 2 and 3) were imported into an Excel spreadsheet. A customized program determined the time and duration of each SDB event from the air-N2 channel and time and duration of periods of wakefulness, NREM sleep, and REM sleep. Furthermore, the sleep-wake state data were subjected to combining rules as follows.


View larger version (26K):
[in this window]
[in a new window]
  		Fig. 2.   Recording example from 1 mouse showing the repetitive induction of hypoxia during non-rapid eye movement (NREM) sleep and the restoration of room air at arousal. Note: 1) the decrease in amplitude and increase in frequency (not evident at compression shown) of the EEG and subtle increases in EMG activity that constitute arousal in response to induced hypoxia during NREM sleep, 2) the epoch time marker is shown as a vertical spike at 5-s intervals on the hypnogram (third trace from top) while the horizontal lines represent the states of wakefulness (bottom) and NREM sleep (top), 3) each 5-s period on the hypnogram represents the sleep-wake state detected from the pattern of EEG and EMG activity from the previous 5 s (i.e., the hypnogram is offset 5 s relative to the polysomnographic trace), and 4) 3 consecutive periods of sleep were required to activate the solenoids to switch from room air to N2. REM, rapid eye movement.



View larger version (30K):
[in this window]
[in a new window]
  		Fig. 3.   Recording example from 1 mouse showing the transition from NREM to REM sleep during a period of induced hypoxia. Note: 1) the decrease in amplitude and increase in frequency (not evident at compression shown) of the EEG and decrease in EMG activity in the transition from NREM to REM sleep; 2) the epoch time marker is shown as a vertical spike at 5-s intervals on the hypnogram (third trace from top) while the horizontal lines represent the states of wakefulness (bottom), NREM sleep (middle), and REM sleep (top); 3) each 5-s period on the hypnogram represents the sleep-wake state detected from the pattern of EEG and EMG activity from the previous 5 s (i.e., the hypnogram is offset 5 s relative to the polysomnographic trace); and 4) 3 consecutive periods of sleep were required to activate the solenoids to switch from room air to N2.

Any single epoch of REM sleep separated by eight or more epochs from the next REM period was reassigned as a NREM epoch. This rule removed all isolated periods of REM that were incorrectly assigned.

All epochs of NREM sleep flanked by two REM epochs that were four or less epochs apart were reassigned to REM. This combining rule allowed consolidation of REM periods that typically last 60-120 s in mice.

After REM combining was completed, an identical approach was used to combine periods of NREM sleep interspersed with periods of wakefulness lasting four or fewer epochs. This combining rule allowed consolidation of NREM periods that typically last 10-15 min in mice.

Scoring validation. The visual scoring of polysomnographic data by human observers was required for both the setting of appropriate thresholds during the initialization period and for the validation of the automated sleep detection system. Visual scoring was undertaken by two trained observers according to our laboratory's published standardized sleep scoring procedures for mice (34). Our laboratory has previously demonstrated consistency in scoring of sleep stages in mice between two independent observers (34), based on 96.6% agreement and a significant kappa statistic of 0.95 (P < 0.01). In the present protocol, two experienced observers were used to validate the results of the automated sleep-wake detection system.

Scoring validation of sleep-wake architecture was performed in each of the five mice over 2 h during the recovery period of the protocol. A total of 720 5-s epochs were assessed during the first hour of the light phase, and a total of 720 epochs were assessed during the first hour of the dark phase. To ensure adequate REM data were included in the validation assessment, a minimum of five REM periods were analyzed in the light phase for each mouse. If fewer than five REM periods occurred in the first hour of the light phase, then the next REM periods were assessed in chronological order until a total of five was achieved.

In addition to sleep architecture, we validated the judgment of the on-line computer detection system to correctly assess sleep and arousal during sleep-related hypoxic events during SDB. First, we assessed whether during the course of an event (which the computer by definition has scored as either NREM or REM sleep), an American Sleep Disorders Association-defined arousal (1) occurred that was undetected by the computer. Second, we assessed the performance of the sleep-wake detection system at the termination of the event (which by definition the computer had scored as awake). In each mouse, events were assessed at ~30-min intervals over a 48-h period.

Analysis of the nadir inspired FIO2 from the event duration. A least squares correlation analysis was used to assess the relationship between event duration and nadir FIO2. In each mouse, ~100-200 events ranging in duration from 5 to 65 s were assessed for the nadir FIO2.

Statistical analyses. All results are presented as means ± SE. Statistical significance between factors of SDB (control, days 1-5 SDB, recovery) and light cycle (light, dark), were determined by ANOVA. If the ANOVA was significant for a factor, a Dunnett's test was used to determine which means were significantly different.


    RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

Validation of Sleep-Wake Detection System

Sleep Architecture analysis. The results presented in Table 1 show a high agreement between human scoring and the computerized sleep-wake analyses. The computer correctly assessed wakefulness 97.2 ± 1.1% (range 93.0-99.0%), NREM sleep 96.0 ± 0.9% (range 93.0-98.9%), and REM sleep 85.6 ± 5.0% (range 67.1-94.2%) of the time.

                              
View this table:
[in this window]
[in a new window]
  		Table 1.   Validation of computerized sleep/wake detection against human scoring

Arousal and event analysis. We assessed an average of 110 ± 13.2 events in five mice and determined that in only 6.0 ± 1.3% of events did an arousal occur during the event that was undetected. An average of 93.2 ± 1.9% of events were correctly terminated with an arousal and removal of the hypoxic stimulus, whereas 6.8 ± 1.9% of events were terminated prematurely during hypoxic exposure without visual evidence of arousal.

SDB Characteristics

A recording example of a series of three consecutive events during NREM sleep is shown for one mouse in Fig. 2 during the fourth day of exposure to SDB. Figure 2 shows that, after three successive 5-epochs of NREM sleep, the gas perfusing the mouse chamber is switched from air to N2 and FIO2 level begins to fall. After the FIO2 fell to ~10%, the mouse aroused from sleep as determined by a decrease in the amplitude and increase in frequency (not visible with compression in Fig. 2) of EEG activity and subtle increases in EMG activity, at which time the chamber rapidly reoxygenated. A second recording example sample trace (Fig. 3) shows a transition from NREM to REM sleep during the course of an event.

Number of events. The total number of events per 24-h period for each of the 5 days of SDB is shown in Table 2. Approximately 500-600 events were recorded for each of the 5 days, and there was no statistical change over time. However, there was a statistical difference (P = 0.034) between the number of events during the light phase (308 ± 9) compared with the dark phase (221 ± 10) when averaged over the 5 days, which can be attributed to more sleep in the light phase, as detailed in Sleep Architecture.

                              
View this table:
[in this window]
[in a new window]
  		Table 2.   Characterization of sleep-disordered breathing events over 5 consecutive days

SDB rate. The number of events per hour of sleep averaged over 60/h throughout the 5-day protocol, and there was no difference in the SDB rate between the light and dark phases.

Event duration and nadir FIO2. There was no significant increase in the mean event duration over the 5 days of SDB. Furthermore, there was no statistical difference in the mean duration of events during the light phase (26.2 ± 1.3 s) and the dark phase (25.9 ± 1.0 s).

In five mice, the slope of the relationship between event duration and nadir FIO2 averaged (-0.196 ± 0.002% FIO2/s; range -0.187 to -0.2009% FIO2/s) and the intercept averaged (18.74 ± 0.18% FIO2; range 18.27-19.23% FIO2). The pooled data show that the mean nadir FIO2 reached 13.2 ± 0.1% on day 5 (Table 2). Although the mean nadir FIO2 was ~13-14% over the 5 days of SDB, it should be noted that many of the events that occurred were associated with severe reductions in FIO2. For example, on day 5 of SDB, there were 137.1 ± 13.2 events in which the nadir FIO2 fell below 10% O2 (event duration >45 s) at arousal and were 62.9 ± 10.8 events in which the FIO2 fell below 7% (event duration >60 s) at arousal.

Blood gases. We measured arterial blood-gas levels during exposure to FIO2 levels of 15, 10, and 5% O2 during wakefulness. The pooled data (n = 4) in which four successful blood-gas measurements were obtained from each mouse are shown in Table 3. Clearly, exposure to a 10 or 5% level of FIO2 under steady-state conditions resulted in severe arterial hypoxemia in awake mice.

                              
View this table:
[in this window]
[in a new window]
  		Table 3.   Arterial blood-gas values in response to decreasing levels of FIO2

Sleep Architecture

We examined sleep architecture during the 24-h control period, the 5 days of SDB, and the 24-h recovery period. The data presented in Fig. 4 compare the average of 5 days of SDB with the control and recovery periods. Two-way ANOVA showed an independent effect of light cycle (light vs. dark; P < 0.0001) and a significance level of P = 0.054 for SDB (control, days 1-5 SDB, recovery) for NREM and REM sleep combined. For REM sleep alone, there were independent effects for both light cycle (P < 0.0001) and SDB (P = 0.017), such that REM sleep during SDB was significantly less than either during control (P = 0.049) or recovery (P = 0.013) periods.


View larger version (20K):
[in this window]
[in a new window]
  		Fig. 4.   Effect of 5 consecutive days of sleep-disordered breathing (SDB) on the level of NREM sleep (A), REM sleep (B), and wakefulness (C) compared with 24-h control and recovery periods. Values are means ± SE for 5 mice. Statistical differences were determined by ANOVA and Dunnett's post hoc analyses. Day 1-5 SDB, time spent per 24 h in each sleep-wake state averaged over 5 days of SDB per animal and then pooled across all 5 animals.


    DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

We have developed a murine model to study the pathophysiological sequelae of SDB. We showed that, within 5 days, it was possible to induce remarkably homogeneous SDB characteristics in the C57BL/6J mouse. Our automated murine model induced intermittent sleep-related hypoxemia at a severity and rate that simulated moderate to severe obstructive sleep apnea in the clinical setting. Furthermore, we showed that the intermittent sleep-related hypoxemia resulted in reductions in sleep, particularly REM sleep. Thus our model replicates the disruptions in blood-gas homeostasis and sleep architecture that characterize obstructive sleep apnea, allowing for future examination of not only sequelae but also underlying molecular mechanisms.

Strengths of Model

Our model of SDB in the mouse has several strengths. First, we can accurately assess sleep-wake state using an on-line automated detection system, originally developed for a canine model of obstructive sleep apnea (23). The automated system demonstrated high reliability at both inducing hypoxia during sleep and terminating hypoxia at arousal. Second, the assessment of sleep architecture showed that the automated sleep-wake detection system correctly assessed wakefulness and NREM sleep >95% of the time (Table 1). Even during REM sleep, which is the most difficult state to correctly assess, the automated system was correct 85.6 ± 5.0% of the time (Table 1). Third, the reduction in FIO2 during SDB events represents a significant degree of arterial hypoxemia in the mouse. Our steady-state measurements of arterial blood-gas levels will approximate the nadir arterial PO2 levels during SDB in the mouse, because a rapid circulation, combined with a respiratory rate in the 3- to 5-Hz range, will produce a rapid plateau of the alveolar-to-arterial PO2 gradient. Thus we have developed a reliable system for inducing hypoxia during sleep as well as assessing circadian fluctuations in sleep architecture.

Limitations of Model

Although a mouse model of SDB has considerable utility, it is important to recognize the associated inherent limitations in the approach we have taken. Our model does not incorporate airway obstruction as the perturbing factor altering arterial blood-gas levels during sleep. Therefore, we have used the generic term "sleep-disordered breathing," rather than "obstructive sleep apnea," because SDB encompasses a wide variety of gas-exchange abnormalities confined to the sleep state. In our model, SDB was characterized by hypoxemia and sleep fragmentation as is commonly seen in obstructive and central sleep apnea, Cheyne-Stokes respiration, and nocturnal hypoventilation syndromes. Moreover, we demonstrate with our model that the application of hypoxic stimuli during sleep resulted in a similar duration, intensity, and frequency of SDB episodes as commonly seen in human populations.

The absence of airway obstruction in our model eliminates the large negative intrathoracic pressure swings that characterize obstructive sleep apnea. It is unclear what role changes in intrathoracic pressure play in the sequelae of obstructive sleep apnea. Potentially, large negative swings in intrathoracic pressure could alter filling of the right and left heart through its effect on venous return and intraventricular dependence (31, 32, 39) or affect afferent neural traffic from the chest wall and lung receptors (8, 14). Neural input from lung and chest wall receptors is likely different during periods of transient hypoxia in which the upper airway is patent rather than obstructed and as such may affect the arousal response (15, 20, 22). Despite the absence of changes in intrathoracic pressure, however, our model does reproduce the two other major defining characteristics of obstructive sleep apnea, namely, sleep-related hypoxemia and sleep fragmentation.

Another limitation is that investigators will not be able to elucidate mechanisms responsible for the pathogenesis of obstructive sleep apnea with our model. For this purpose, naturally occurring animal models of obstructive sleep apnea, such as the English bulldog (19) and the obese Vietnamese pot-bellied pig (40), are better suited because they provide insight into what factors can lead to upper airway collapse during sleep. To our knowledge, no small-animal models of obstructive sleep apnea have been reported. Even the severely obese C57BL/6J-Lepob mouse does not exhibit any detectable form of SDB during sleep (27, 29). Thus the utility of the mouse as a model of SDB is more likely to focus on mechanisms of pathophysiological sequelae rather than the pathogenesis of upper airway obstruction during sleep.

Implications

The development of a mouse model of SDB has both technical and experimental implications. From a technical perspective, SDB in the mouse can be modified in a number of ways to model different aspects of the syndrome. The characteristics of the stimulus gas can be altered in terms of composition (e.g., hypoxia and hypercapnia) as well as severity (e.g., 100% N2 vs. 90% N2-10% O2). Additionally, altering the flow rate at which the stimulus gas is delivered and resetting the number of sleep epochs required to trigger the gas delivery can further modify the severity of the SDB. For example, we could increase the severity of SDB by increasing the flow rate of the stimulus gas and decreasing the number of sleep epochs required to trigger gas delivery. A sample recording in Fig. 5 shows the induction of 92 events/h in one mouse as a result of increasing N2 delivery from 2 to 3.0 l/m and triggering the stimulus gas after one epoch of sleep. Alternatively, we can decrease the SDB rate in the mouse model by increasing the number of consecutive epochs of sleep required before triggering the stimulus gas. Thus in our model it will be possible to vary characteristics of SDB from very mild (0-10 events/h) to very severe (90-100 events/h).


View larger version (48K):
[in this window]
[in a new window]
  		Fig. 5.   Recording example from 1 mouse showing a 35-min period of induced SDB in which N2 was delivered at a rate of 3.0 l/min (cf. 2.0 l/min in Figs. 2 and 3) and that only one 5-s period of sleep was required to trigger the solenoids to switch from room air to N2. Note the high SDB rate of over 90 events/h and the rapid fall in FIO2 during periods of induced hypoxia.

From an experimental perspective, a mouse model of SDB has a wide range of applications. First, the power of inbred strains will reduce outcome variability and allow application of traditional genetic approaches developed for mice. Second, comorbid mouse models will determine whether specific conditions, such as obesity or hypertension, interact with SDB to exacerbate the pathophysiological sequelae. Third, specific knockout mice could test the molecular basis for specific sequelae. For example, does hypoxia-inducible factor-1alpha , which is known to mediate pulmonary hypertension during chronic hypoxia (43), contribute to pulmonary or systemic hypertension in SDB? Finally, the technique of cDNA microarray analysis (5, 18), when applied to our model, may uncover gene expression patterns implicating SDB in as yet unknown pathology. Clearly, the technical difficulties associated with the development of a model of SDB in such a small animal are offset by the broad spectrum of applications and approaches that offer the possibility of new insights.

Summary

Our study demonstrates that it is possible to model SDB in mice by inducing hypoxia at sleep onset and removing hypoxia in response to arousal from sleep. Such events occurred on average >500 times in one 24-h cycle and were of a duration that resulted in significant hypoxemia. In addition to severely fragmenting sleep, 5 days of SDB in the mouse produced significant decrements in sleep, particularly REM sleep. We anticipate that a mouse model of SDB will provide new insights into the sequelae of SDB, the interaction between comorbid conditions and SDB, and, ultimately, the cellular and molecular basis for pathophysiological outcomes from SDB.


    ACKNOWLEDGEMENTS

We acknowledge the generosity of Dr. E. A. Phillipson and his colleagues (R. J. Kimoff, H. Makino, R. L. Horner, L. F. Kozar, F. Lue, and A. S. Slutsky) from the University of Toronto in allowing us to adapt the sleep-detection software developed in his laboratory. We also thank Dr. R. L. Horner for his assistance in implementing the software and interfacing the software to our hardware.


    FOOTNOTES

Address for reprint requests and other correspondence: C. P. O'Donnell, Rm. 4B61, Johns Hopkins Asthma and Allergy Center, 5501 Hopkins Bayview Circle, Baltimore, MD 21224 (E-mail: codonnel{at}welch.jhu.edu).

The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Received 5 April 2001; accepted in final form 18 July 2001.


    REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES

1.   Atlas Task Force. EEG arousals: scoring rules and examples: a preliminary report from the Sleep Disorders Atlas Task Force of the American Sleep Disorders Association. Sleep 15: 174-184, 1996.

2.   Bakehe, M, Hedner J, Dang T, Chambille B, Gaultier CL, and Escourrou P. Role of the autonomic nervous system in the acute blood pressure elevation during repetitive hypoxic and hypercapnic breathing in rats. Blood Press 5: 371-375, 1996[Medline].

3.   Bao, G, Metreveli N, Li R, Taylor A, and Fletcher EC. Blood pressure response to chronic episodic hypoxia: role of the sympathetic nervous system. J Appl Physiol 83: 95-101, 1997[Abstract/Free Full Text].

4.   Brooks, D, Horner RL, Kozar LF, Render-Teixeira CL, and Phillipson EA. Obstructive sleep apnea as a cause of systemic hypertension. Evidence from a canine model. J Clin Invest 99: 106-109, 1997[ISI][Medline].

5.   Brown, PO, and Botstein D. Exploring the new world of the genome with DNA microarrays. Nat Genet 21: 33-37, 1999[ISI][Medline].

6.   Chen, L, and Scharf SM. Effects of aortic nerve on hemodynamic response to obstructive apnea in sedated pigs. J Appl Physiol 89: 1455-1461, 2000[Abstract/Free Full Text].

7.   Chen, L, Sica AL, and Scharf SM. Mechanisms of acute cardiovascular response to periodic apneas in sedated pigs. J Appl Physiol 86: 1236-1246, 1999[Abstract/Free Full Text].

8.   Davenport, PW, Thompson FJ, Reep RL, and Freed AN. Projection of phrenic nerve afferents to the cat sensorimotor cortex. Brain Res 328: 150-153, 1985[ISI][Medline].

9.   Epstein, L, and Weiss W. Clinical consequences of obstructive sleep apnea. Semin Respir Crit Care Med 19: 123-132, 1998.

10.   Fletcher, EC, Bao G, and Li R. Renin activity and blood pressure in response to chronic episodic hypoxia. Hypertension 34: 309-314, 1999[Abstract/Free Full Text].

11.   Fletcher, EC, Lesske J, Behm R, Miller CC, Stauss H, III, and Unger T. Carotid chemoreceptors, systemic blood pressure, and chronic episodic hypoxia mimicking sleep apnea. J Appl Physiol 72: 1978-1984, 1992[Abstract/Free Full Text].

12.   Fletcher, EC, Lesske J, Culman J, Miller CC, and Unger T. Sympathetic denervation blocks blood pressure elevation in episodic hypoxia. Hypertension 20: 612-619, 1992[Abstract/Free Full Text].

13.   Fletcher, EC, Lesske J, Qian W, Miller CC, III, and Unger T. Repetitive, episodic hypoxia causes diurnal elevation of blood pressure in rats. Hypertension 19: 555-561, 1992[Abstract/Free Full Text].

14.   Gandevia, SC, and Macefield G. Projection of low-threshold afferents from human intercostal muscles to the cerebral cortex. Respir Physiol 77: 203-214, 1989[ISI][Medline].

15.   Gleeson, K, Zwillich CW, and White DP. The influence of increasing ventilatory effort on arousal from sleep. Am Rev Respir Dis 142: 295-300, 1990[ISI][Medline].

16.   Greenberg, HE, Sica A, Batson D, and Scharf SM. Chronic intermittent hypoxia increases sympathetic responsiveness to hypoxia and hypercapnia. J Appl Physiol 86: 298-305, 1999[Abstract/Free Full Text].

17.   Grunstein, RR, Stenlof K, Hedner J, and Sjostrom L. Impact of obstructive sleep apnea and sleepiness on metabolic and cardiovascular risk factors in the Swedish Obese Subjects (SOS) Study. Int J Obes Relat Metab Disord 19: 410-418, 1995[ISI][Medline].

18.   Harkin, DP. Uncovering functionally relevant signaling pathways using microarray-based expression profiling. Oncologist 5: 501-507, 2000[Abstract/Free Full Text].

19.   Hendricks, JC, Kline LR, Kovalski RJ, O'Brien JA, Morrison AR, and Pack AI. The English bulldog: a natural model of sleep-disordered breathing. J Appl Physiol 63: 1344-1350, 1987[Abstract/Free Full Text].

20.   Issa, FG, McNamara SG, and Sullivan CE. Arousal responses to airway occlusion in sleeping dogs: comparison of nasal and tracheal occlusions. J Appl Physiol 62: 1832-1836, 1987[Abstract/Free Full Text].

21.   Kanagy, NL, Walker BR, and Nelin LD. Role of endothelin in intermittent hypoxia-induced hypertension. Hypertension 37: 511-515, 2001[Abstract/Free Full Text].

22.   Kimoff, RJ, Kozar LF, Yasuma F, Bradley TD, and Phillipson EA. Effect of inspiratory muscle unloading on arousal responses to CO2 and hypoxia in sleeping dogs. J Appl Physiol 74: 1325-1336, 1993[Abstract/Free Full Text].

23.   Kimoff, RJ, Makino H, Horner RL, Kozar LF, Lue F, Slutsky AS, and Phillipson EA. Canine model of obstructive sleep apnea: model description and preliminary application. J Appl Physiol 76: 1810-1817, 1994[Abstract/Free Full Text].

24.   McGuire, M, and Bradford A. Chronic intermittent hypoxia increases haematocrit and causes right ventricular hypertrophy in the rat. Respir Physiol 117: 53-58, 1999[ISI][Medline].

25.   Nieto, FJ, Young TB, Lind BK, Shahar E, Samet JM, Redline S, D'Agostino RB, Newman AB, Lebowitz MD, and Pickering TG. Association of sleep-disordered breathing, sleep apnea, and hypertension in a large community-based study. Sleep Heart Health Study (see comments). JAMA 283: 1829-1836, 2000[Abstract/Free Full Text].

26.   O'Donnell, CP, King ED, Schwartz AR, Robotham JL, and Smith PL. Relationship between blood pressure and airway obstruction during sleep in the dog. J Appl Physiol 77: 1819-1828, 1994[Abstract/Free Full Text].

27.   O'Donnell, CP, Schaub CD, Haines AS, Berkowitz DE, Tankersley CG, Schwartz AR, and Smith PL. Leptin prevents respiratory depression in obesity. Am J Respir Crit Care Med 159: 1477-1484, 1999[Abstract/Free Full Text].

28.   O'Donnell, CP, Schwartz AR, Smith PL, Robotham JL, Fitzgerald RS, and Shirahata M. Reflex stimulation of renal sympathetic nerve activity and blood pressure in response to apnea. Am J Respir Crit Care Med 154: 1763-1770, 1996[Abstract].

29.   O'Donnell, CP, Tankersley CG, Polotsky VP, Schwartz AR, and Smith PL. Leptin, obesity, and respiratory function. Respir Physiol 119: 163-170, 2000[ISI][Medline].

30.   Redline, S, and Strohl KP. Recognition and consequences of obstructive sleep apnea hypopnea syndrome. Clin Chest Med 19: 1-19, 1998[ISI][Medline].

31.   Robotham, JL, Rabson J, Permutt S, and Bromberger-Barnea B. Left ventricular hemodynamics during respiration. J Appl Physiol 47: 1295-1303, 1979[Abstract/Free Full Text].

32.   Scharf, SM, Brown R, Saunders N, and Green LH. Effects of normal and loaded spontaneous inspiration on cardiovascular function. J Appl Physiol 47: 582-590, 1979[Abstract/Free Full Text].

33.   Scharf, SM, Graver LM, and Balaban K. Cardiovascular effects of periodic occlusions of the upper airways in dogs. Am Rev Respir Dis 146: 321-329, 1992[ISI][Medline].

34.   Schaub, CD, Tankersely C, Schwartz AR, Smith PL, Robotham JL, and O'Donnell CP. Effect of sleep/wake state on arterial blood pressure in genetically identical mice. J Appl Physiol 85: 366-371, 1998[Abstract/Free Full Text].

35.   Schneider, H, Schaub CD, Chen CA, Andreoni KA, Schwartz AR, Smith PL, Robotham JL, and O'Donnell CP. Effects of arousal and sleep state on systemic and pulmonary hemodynamics in obstructive apnea. J Appl Physiol 88: 1084-1092, 2000[Abstract/Free Full Text].

36.   Schneider, H, Schaub CD, Chen CA, Andreoni KA, Schwartz AR, Smith PL, Robotham JL, and O'Donnell CP. Neural and local effects of hypoxia on cardiovascular responses to obstructive apnea. J Appl Physiol 88: 1093-1102, 2000[Abstract/Free Full Text].

37.   Slamowitz, D, Chen L, and Scharf SM. Effects of vagotomy on cardiovascular response to periodic apneas in sedated pigs. J Appl Physiol 86: 1890-1896, 1999[Abstract/Free Full Text].

38.   Strohl, KP. Diabetes and sleep apnea. Sleep 19: S225-S228, 1996[ISI][Medline].

39.   Tarasiuk, A, and Scharf SM. Effects of periodic obstructive apneas on venous return in closed-chest dogs. Am Rev Respir Dis 148: 323-329, 1993[ISI][Medline].

40.   Tuck, SA, Dort JC, Olson ME, and Remmers JE. Monitoring respiratory function and sleep in the obese Vietnamese pot-bellied pig. J Appl Physiol 87: 444-451, 1999[Abstract/Free Full Text].

41.   Vgontzas, AN, Papanicolaou DA, Bixler EO, Hopper K, Lotsikas A, Lin HM, Kales A, and Chrousos GP. Sleep apnea and daytime sleepiness and fatigue: relation to visceral obesity, insulin resistance, and hypercytokinemia. J Clin Endocrinol Metab 85: 1151-1158, 2000[Abstract/Free Full Text].

42.   White, SG, Fletcher EC, and Miller CC. Acute systemic blood pressure elevation in obstructive and nonobstructive breath hold in primates. J Appl Physiol 79: 324-330, 1995[Abstract/Free Full Text].

43.   Yu, AY, Shimoda LA, Iyer NV, Huso DL, Sun X, McWilliams R, Beaty T, Sham JS, Wiener CM, Sylvester JT, and Semenza GL. Impaired physiological responses to chronic hypoxia in mice partially deficient for hypoxia-inducible factor 1alpha . J Clin Invest 103: 691-696, 1999[ISI][Medline].

J APPL PHYSIOL 91(6):2758-2766
8750-7587/01 $5.00 Copyright © 2001 the American Physiological Society


This article has been cited by other articles:


Home page
 Am. J. Physiol. Heart Circ. Physiol.Home page
S. de Frutos, L. Duling, D. Alo, T. Berry, O. Jackson-Weaver, M. Walker, N. Kanagy, and L. Gonzalez Bosc
NFATc3 is required for intermittent hypoxia-induced hypertension
Am J Physiol Heart Circ Physiol, May 1, 2008; 294(5): H2382 - H2390.
[Abstract] [Full Text] [PDF]

Home page
 Proc Am Thorac SocHome page
O. S. Capdevila, L. Kheirandish-Gozal, E. Dayyat, and D. Gozal
Pediatric Obstructive Sleep Apnea: Complications, Management, and Long-term Outcomes
Proceedings of the ATS, February 15, 2008; 5(2): 274 - 282.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Respir. Crit. Care Med.Home page
D. Gozal and L. Kheirandish-Gozal
Cardiovascular Morbidity in Obstructive Sleep Apnea: Oxidative Stress, Inflammation, and Much More
Am. J. Respir. Crit. Care Med., February 15, 2008; 177(4): 369 - 375.
[Abstract] [Full Text] [PDF]

Home page
 Physiol. GenomicsHome page
J. Li, A. Nanayakkara, J. Jun, V. Savransky, and V. Y. Polotsky
Effect of deficiency in SREBP cleavage-activating protein on lipid metabolism during intermittent hypoxia
Physiol Genomics, October 19, 2007; 31(2): 273 - 280.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Physiol. Gastrointest. Liver Physiol.Home page
V. Savransky, S. Bevans, A. Nanayakkara, J. Li, P. L. Smith, M. S. Torbenson, and V. Y. Polotsky
Chronic intermittent hypoxia causes hepatitis in a mouse model of diet-induced fatty liver
Am J Physiol Gastrointest Liver Physiol, October 1, 2007; 293(4): G871 - G877.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Respir. Crit. Care Med.Home page
N. Iiyori, L. C. Alonso, J. Li, M. H. Sanders, A. Garcia-Ocana, R. M. O'Doherty, V. Y. Polotsky, and C. P. O'Donnell
Intermittent Hypoxia Causes Insulin Resistance in Lean Mice Independent of Autonomic Activity
Am. J. Respir. Crit. Care Med., April 15, 2007; 175(8): 851 - 857.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
J. Li, V. Savransky, A. Nanayakkara, P. L. Smith, C. P. O'Donnell, and V. Y. Polotsky
Hyperlipidemia and lipid peroxidation are dependent on the severity of chronic intermittent hypoxia
J Appl Physiol, February 1, 2007; 102(2): 557 - 563.
[Abstract] [Full Text] [PDF]

Home page
 Physiol. GenomicsHome page
J. Li, M. Bosch-Marce, A. Nanayakkara, V. Savransky, S. K. Fried, G. L. Semenza, and V. Y. Polotsky
Altered metabolic responses to intermittent hypoxia in mice with partial deficiency of hypoxia-inducible factor-1{alpha}
Physiol Genomics, May 16, 2006; 25(3): 450 - 457.
[Abstract] [Full Text] [PDF]

Home page
 Am. J. Physiol. Regul. Integr. Comp. Physiol.Home page
R. Lydic
Sleep disruption is related to allelic variation in the ob gene
Am J Physiol Regulatory Integrative Comp Physiol, April 1, 2006; 290(4): R892 - R893.
[Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
J. Li, D. N. Grigoryev, S. Q. Ye, L. Thorne, A. R. Schwartz, P. L. Smith, C. P. O'Donnell, and V. Y. Polotsky
Chronic intermittent hypoxia upregulates genes of lipid biosynthesis in obese mice
J Appl Physiol, November 1, 2005; 99(5): 1643 - 1648.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
M. J. Campen, L. A. Shimoda, and C. P. O'Donnell
Acute and chronic cardiovascular effects of intermittent hypoxia in C57BL/6J mice
J Appl Physiol, November 1, 2005; 99(5): 2028 - 2035.
[Abstract] [Full Text] [PDF]

Home page
 J. Appl. Physiol.Home page
M. J. Campen, Y. Tagaito, T. P. Jenkins, A. Balbir, and C. P. O'Donnell
Heart rate variability responses to hypoxic and hypercapnic exposures in different mouse strains
J Appl Physiol, September 1, 2005; 99(3): 807 - 813.
[Abstract] [Full Text] [PDF]

Home page
 J. Neurosci.Home page
D. D. Kline, M. C. F. Buniel, P. Glazebrook, Y.-J. Peng, A. Ramirez-Navarro, N. R. Prabhakar, and D. L. Kunze
Kv1.1 Deletion Augments the Afferent Hypoxic Chemosensory Pathway and Respiration
J. Neurosci., March 30, 2005; 25(13): 3389 - 3399.