To the Editor:

Obstructive sleep apnea syndrome (OSAS), a respiratory disorder that affects 5–15% of the general population, is characterized by recurrent collapses of the upper airways during sleep (for review, see Reference 1). It leads to repetitive episodes of hypoxia/hypercapnia and sleep fragmentation, and has severe cardiovascular and cognitive consequences (1, 2). Instability of the upper airways can be due to anatomical causes; for example, accumulation of pharyngeal adipose tissue (3) modifies the size of the upper airways in obese individuals. Obesity is one of the main risk factors for OSAS and is reported in up to 60% of patients with OSAS (2). Treatment of severe OSAS is mainly based on continuous positive airway pressure. Weight loss also improves OSAS (1).

Several attempts have been made to reproduce the pathological features of obesity-related OSAS in animal models. The English bulldog is a natural model of OSAS because it has narrow upper airways (4), but OSAS in these animals is related to their short muzzle rather than obesity. Other models induce OSAS only temporarily, or recreate partial apnea-like conditions such as chronic intermittent hypoxia (CIH) (4). The cognitive and cardiovascular consequences of CIH, the main consequence of OSAS (4), are often studied in rodents. However, the CIH in these animals is not caused by anatomic and functional abnormalities as observed in patients, and the recurrent episodes of hypercapnia observed in OSAS are often lacking from this model. Therefore, there is a need for a natural pathophysiological model to clarify the key central and peripheral mechanisms underlying OSAS.

In this context, we investigated whether New Zealand obese (NZO/HlLtJ) mice, which have anatomic and functional characteristics similar to those of obese patients with OSAS, suffer from obstructive sleep apnea and could therefore be used as a pathophysiological model of OSAS. NZO mice display excessive fat accumulation around the pharynx (reducing the airway diameter) at 20–24 weeks of age, as shown by magnetic resonance imaging (5, 6). They also display hyperreactivity of the upper airways and metabolic syndrome (5, 6) (as previously described with hypertension, insulin resistance, and hyperleptinemia) (7).

Male NZO mice (NZO/HlLtJ #002105; n = 8; 51.7 ± 0.79 g) and New Zealand black (NZB) nonobese control mice (NZB/BlNJ #000684; n = 4; 32.31 ± 1.05 g; The Jackson Laboratory) were housed in 12-hour light/12-hour dark cycle conditions, at 24–25°C, with food and water supplied ad libitum. They were studied at the ages of 20–24 weeks. All animal procedures were performed in accordance with legislation (authorization #7323).

We quantified sleep apnea events (using the apnea–hypopnea index [AHI], the number of apnea and hypopnea events per hour of recording time) and sleep behaviors. A standard set of electrodes for polygraphic monitoring of sleeping and awakening was implanted into the animals after they were anesthetized with isoflurane (8). The animals were allowed to recover for 15 days, during which they were familiarized with whole-body plethysmographs (Emka Technologies) adapted to allow the use of a spinning collector (Air Precision). Ventilation and polygraphic electroencephalogram, electrooculogram, and electromyogram states were recorded over a period of 24 hours (8:00 a.m. to 8:00 a.m. the following day). During the light/resting period, oxygen saturation as measured by pulse oximetry (Sp_O2) was determined using a collar sensor with a MouseOx Plus pulse oximeter (Starr Life Sciences, Harvard Apparatus).

Quantification was performed over representative light/resting (9:00 a.m. to 5:00 p.m.) and dark/activity (8:00 p.m. to 4:00 a.m.) periods. A complete breath-by-breath analysis of ventilation was performed with IOX (version 2.8.053, Emka Technologies), and representative ventilatory variables (inspiratory time, Vt, and respiratory rate) were determined. Sleep/awakening states (awakening, slow-wave sleep [non-REM], and REM) were scored with Somnologica 2 EMBLA (Medcare) as previously described (8). Apnea and hypopnea were quantified as previously described (8, 9). Briefly, apnea was defined as the cessation of ventilation for at least two respiratory cycles, and hypopnea was defined as inspiratory flow 50% below the mean at rest. Data are expressed as mean ± SEM.

NZO mice presented larger numbers of spontaneous apnea and hypopnea events than NZB mice, during both the light/resting (AHI: 65.5 ± 10.9 vs. 24.3 ± 4.3; P < 0.05; Table 1; Figure 1) and dark/activity (AHI: 52.8 ± 4.5 vs. 24.5 ± 2.7; Table 1) periods. The NZO mice sometimes slept standing up, whereas the NZB mice never did (data not shown). Sp_O2 of NZO mice always showed desaturation (95.3% ± 0.2%: 92.5% < Sp_O2 < 97.3%), whereas Sp_O2 of NZB mice was consistent with normal saturation (98.9% ± 0.1%: 98.0% < Sp_O2 < 99.4%) (Mann-Whitney test, P < 0.0001). During the light/resting period, NZO mice hyperventilated (with increases in respiratory rate and decreases in inspiratory time) (Table 1). The scoring of sleep/awakening states revealed a larger number of arousals in NZO mice than in NZB mice during the light/resting period (154.3 ± 11.4 vs. 47.7 ± 14.2; P < 0.001; Table 1). During the dark/activity period, the NZO mice slept longer (49.8% ± 2.1% of the time) than the NZB mice (22.8% ± 8.7%), indicating greater sleepiness in the NZO mice (P < 0.0001; Table 1; Figure 1). This sleepiness could explain why the NZO mice exhibited lower voluntary running-wheel activity than the NZB mice during their dark/activity period (10).

Figure 1. Representative apnea/hypopnea events and hypnograms of NZO and NZB mice. From left to right and top to bottom: an apnea event (top) and a hypopnea event (bottom) observed in NZO mice, and a representative hypnogram from an NZB mouse (upper) and an NZO mouse (lower) during light/resting and dark/activity periods. AW = awakening; NREM = non-REM; NZB = New Zealand black; NZO = New Zealand obese. Arousals are shown as spots.

Download Figure | Download Powerpoint

[More] [Minimize]

We hypothesize that the sleep apnea/hypopnea events observed in NZO mice are obstructive with episodes of desaturation, due to the abnormal upper airways of these mice (5, 6) and their tendency to sleep upright occasionally (6). Hyperventilation could be a result of CIH and hypercapnia episodes during sleep interruptions. This hypothesis is supported by the increased frequency of arousals, high AHI, and significant desaturation observed during the light/resting period in NZO mice.

NZO mice may therefore constitute a pertinent pathophysiological model for OSAS, because they have several features in common with patients with OSAS, including ventilatory abnormalities, sleep fragmentation, and sleepiness during the activity period. OSAS is a complex disorder involving a multitude of anatomic, central, systemic, and metabolic aspects that have been studied separately in conventional models (4). Our results indicate that most of these features occur together in NZO mice. Validation of an animal model requires the fulfillment of at least one of three criteria: homology, predictiveness, and isomorphism with the pathophysiology of the human disease (11). NZO mice meet at least two of these criteria: homology with obese patients with OSAS and isomorphism.

In conclusion, here we report the characterization of a spontaneous, pathophysiological rodent model of obesity-related OSAS. Aside from increasing our knowledge about upper-airway pathophysiology, this mouse model could be of interest to appreciate the anatomical changes that occur in the upper airways, and even be used to test some therapies (e.g., continuous positive airway pressure) to improve the collapsibility of the upper airway assessed in these mice. This readily available model should enable the research community to gain a deep understanding of the mechanisms underlying OSAS and thus facilitate the development of better targeted treatments.

References

Section:

ChooseTop of pageReferences <<CITING ARTICLES