INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

OBESITY AMONG HUMANS is often, although not always, associated with sleep disorders of breathing. The obese Zucker rat shows many of the same obesity-related symptoms as the excessively obese human: a reduced lung capacity, decreased chest wall compliance, an increased work of breathing, blunted ventilatory responses to hypercapnia and hypoxia (6, 7, 25), and an elevated hematocrit, indicative of hypoxemic stimulation of hemopoesis (unpublished observations). Collectively, these signs suggest that the obese Zucker rat may exhibit respiratory-related dysfunctions during sleep, the incidence of which may be greater than that of its lean counterpart. In this study, we tested the hypothesis that the sleep organization of obese and lean Zucker rats differs.

Danguir (5) was the first to report that obese Zucker rats spent more time in non-rapid-eye-movement (NREM) sleep during light exposure than do Wistar rats, whereas their rapid-eye-movement (REM) sleep contents were nearly the same. He did not compare lean and obese Zucker rats. In a preliminary study, Carley et al. (4) examined the hypotensive effects of hydralazine on respiratory dysfunction during sleep of the obese Zucker rat compared with those of lean littermates. However, they did not determine whether hydralazine independently alters sleep organization.

Ambient temperature and inspired gas mixture have been shown to be important determinants of sleep organization in nonobese strains of rat. While the rat is breathing ambient air, REM sleep is maximal at 29°C (28). REM sleep content decreases progressively on either side of this temperature (28). A pilot study (unpublished observations) showed that an ambient temperature of 29°C yielded the maximal amount of REM sleep in lean and obese Zucker rats, confirming earlier findings (19, 24, 28). Hale et al. (9) showed that when rats breathe 10% O₂ at low ambient temperatures sleep organization is disrupted and body temperature (T_b) decreases. Furthermore, Pollard et al. (19) next showed that rats, breathing even 15% O₂, have their sleep organization altered without T_b being affected. Both Hale et al. (9) and Pollard et al. (19) concluded that the principal mechanism by which sleep organization is altered is the frequency of switching states, much less the duration of the respective states. Therefore, we hypothesize that sleep organization is different between obese and lean rats and that this difference is manifest predominantly by the frequency of switching from NREM sleep relative to its mean epoch duration.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Animals. We used six lean and six obese male Zucker rats that were 3 mo old at the time of surgery. They were obtained from a commercial supplier (Harlan Sprague Dawley, Indianapolis, IN). Pairs of animals, one lean and one obese (not necessarily littermates), were housed in single cages in standard conditions: 12-h light period (0700-1900) and 12-h dark period (1900-0700) at an ambient temperature of 21 ± 1°C. Food and water were provided ad libitum at all times as well as during polygraphic recording sessions. The institution's Animal Care and Use Committee of State University of New York at Buffalo approved the experimental protocol subsequent to carrying out this study.

Surgery. An intraperitoneal injection of a mixture of ketamine (60 mg/kg) and xylazine (7 mg/kg) achieved surgical anesthesia. As needed, the injection of one-fourth to one-third of the anesthetic mixture maintained anesthesia until completion of surgery. Under aseptic conditions, a patch of skin over the dorsum of the rat's skull was removed. The bony calvarium was scraped clean of soft tissue and thoroughly dried. Stainless steel screws, to which were soldered short-length Teflon-coated stainless steel wires (0.8-mm diameter), were inserted into burr holes made in the skull: 1) one 7 mm rostral to bregma in the midline and 2) one each in the center of the parietal bones. A thin coat of Super Glue was spread over the dry skull bones. Three Telfon-coated wires (Medwire, Mt. Vernon, NY, stainless steel 3T), bared at their tips (0.5-1.0 mm) and fashioned into hooks, were inserted into dorsal neck muscles as Basmajian and Stecko (2) have described. The stainless steel screw electrodes and lead wires from muscle hook electrodes were fixed to the skull with dental acrylic resin, with the heads of the screws being completely coated with the resin. The ends of the respective lead wires were then soldered to separate pins of a subminiature socket. Additional resin firmly anchored the socket to the skull and formed a pedestal. Skin was sutured snugly around the pedestal. Each rat received preoperatively a 0.25-ml subcutaneous injection of Crystiben. A topical antibiotic cream was applied to skin edges postoperatively. Animals were allowed at least 7 days to recover from surgery before being habituated to recording conditions for subsequent study. After implantation, the animals were housed in separate cages in the laboratory, which was kept at an ambient temperature of 25 ± 1°C on a 12:12-h light-dark cycle.

Polygraphic recording sessions. Each rat was kept for a minimum of 72 consecutive hours in a plastic pail (30 cm in diameter) with bedding. During this time, the rats were habituated to the mating plug and cable, the other end of which was attached to a fisherman's swivel. The latter was fixed to a platform 75-80 cm above the rat. At 0800 on the first of the next 3 days, one lean and one obese rat were placed in a dual Plexiglas enclosure, with the two rats being separated by an opaque Plexiglas wall. The enclosure was housed in a climatic chamber (VWR BOD incubator, model 2020) maintained at an ambient temperature of 29 ± 0.5°C and on a 12:12-h light-dark cycle. Each rat was connected to four channels of an eight-channel polygraph (Astro-Med, MT-9500 chart recorder/Grass P511 preamplifiers) via a mating plug, braided cable, and slip-ring connector (Air Précision). Two channels were reserved for recording the electrocorticogram (ECoG) (band pass: 0.3-30 Hz) and two for the dorsal nuchal electromyogram (EMG; band pass: 10-30 Hz). At 1600, the mating plug with cable was removed from the rat's socket until the next day while both rats remained in the climatic chamber. At 0800 on the second and third day, each rat was again connected to the respective channels of the polygraph as described above. On the third day, the polygraphic recording began at 1000 and ended at 1600. Finally, the rats were uncoupled from recording equipment and housed once again in animal quarters, each in a single cage.

Scoring of sleep-waking states. Each 30 s of polygraphic recording for each rat's ECoG and nuchal EMG was scored as belonging to one of three states: 1) wakefulness (W), characterized by low-voltage, fast-frequency of the ECoG and marked nuchal EMG activity; 2) NREM sleep, characterized by high-voltage, slow-frequency and diminished nuchal EMG activity; and 3) REM sleep, characterized by diminished ECoG amplitude and a low or absent nuchal EMG, interspersed with occasional signs of muscle twitches. Each 30-s period was denoted as an epoch. Concordance of two scorers of polygraphic traces of sleep-waking states achieved 90%.

Statistical treatment of sleep-waking data. The sleep-waking data of a rat can be displayed as an hypnogram (22). It is an example of a categorical time series. Each of the 12 categorical time series (lean: n₁ = 6 and obese: n₂ = 6 rats) was summarized in two groups of three variables. Table 1 (A) lists the number of (720) 30-s epochs designated as W, NREM sleep, and REM sleep, for a 6-h recording period. Table 1 (B) lists the mean number of consecutive epochs in each state. Table 2 lists the number of switches from each state to another state. For example, in the case of the lean rat 1 (L₁) during the 6-h recording period, this animal made 18 switches from the NREM state to the REM sleep state and 41 switches from the NREM sleep state to the W state. The total number of switches equals 59. The fraction of switches from NREM to the other states can be calculated for L₁ as 18/59 for NREM to REM and 41/59 for NREM to W. The fraction of switches from W to NREM and to REM is 58/61 and 3/61, respectively. The sum of fractions always equals 1. 

The total number of epochs (Table 1A) and all the data in Table 2 should be treated as compositional data. Their values, expressed as fractions, add up to one (see above example). Such types of data were transformed subsequent to analysis by using the additive logistic transformation (1) i = 2,..., D1 and D = 2,3, depending on the group of variables. We tested the null hypothesis of no difference in y_i values between lean and obese rats using the univariate two-sample t-test or the multivariate T² statistic with a pooled covariance matrix (11). We used the SAS statistical package to perform computations. The value of P < 0.05, or better, was accepted as statistically significant. All values are presented as means ± SD.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Mean body weights of obese animals at surgery were greater than those of lean animals: 536 ± 71 g (range: 441-619 g) vs. 342 ± 24 g (range: 320-384 g).

The total number of epochs (Table 1A) in each state and the mean number of consecutive epochs of W and REM were not significantly different between lean and obese animals (Table 1B). However, the mean consecutive number of epochs of NREM sleep was statistically significantly different between the two phenotypes (Table 1B). The frequency of switches from W and from REM sleep to the other states was not statistically significantly different between lean and obese rats, whereas the frequency of switches from NREM to the other two states was significantly different (Table 2). Stated otherwise, obese rats switched less frequently from NREM to REM sleep and more frequently from NREM sleep to W than did lean rats. The schematic of Fig. 1 summarizes our findings on the frequency of switchings in lean and obese rats.

View larger version (21K): [in this window] [in a new window]   Fig. 1.   Schematic of switches from one state, i.e., wakefulness (W), non-rapid-eye-movement (NREM), and rapid-eye-movement (REM) sleep, to another. Values on each arrow denote fraction of no. of switches from a given state to the other 2 states, sum of which equals 1. Large arrowheads indicate that switches, namely, from NREM sleep to W and from NREM sleep to REM sleep, were statistically significantly different (* P = 0.019), whereas all other switches were not (small arrowheads).

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

The principal conclusion from our findings is that the sleep organization of lean and obese male Zucker rats differs. More specifically, the frequency of switches from NREM sleep to REM sleep or from NREM to W switch is significantly different and that the obese animal spends more time in NREM sleep than does the lean animal during the light period. Explicitly, obese Zucker rats awaken with a greater frequency (0.71) than do lean rats (0.61) from NREM sleep (Fig. 1.). Second, the fatty rats have a lesser probability (0.29) of switching from NREM sleep to REM sleep than do lean rats (0.39) (Fig. 1.). The respective differences in frequencies are statistically significant (P = 0.019). It is during this phase of the circadian sleep-waking cycle when rats usually spend time sleeping, rather than during the dark period. Our finding confirms the earlier study of Danguir (5), who showed that obese Zucker rats spend more time in NREM sleep than do normal (lean) Wistar rats during both the light and dark periods, whereas Carley et al. (4) and Radulovacki et al. (20) reported no difference in volume of either NREM or REM sleep during the light period. They scored state in 10-s epochs by using the software of Bennington et al. (3).

The rationale for using an ambient temperature of 29°C was threefold: 1) to maximize the amount of REM sleep (28); 2) to minimize metabolic heat production (28); and 3) to minimize evaporative heat loss through the rat's nasal passage way (10). Deviations on either side of this ambient temperature alter the body heat balance and, presumably indirectly, sleep organization.

We consider two factors, ventilatory and thermoregulatory, that may underlie the mechanism for altering the sleep organization of the obese compared with the lean Zucker rat.

The ventilatory factor. Presumably, pulmonary gas exchange is reduced in the obese compared with the lean animal. We conjecture that the reduction is due to an increased load on the respiratory system because of excessive fat deposits around the upper airway, the chest, or both (7). Consequently, arterial PCO₂ may rise and become ever higher during sleep states, notably during REM sleep. However, hypercarbia is not likely to be the causal factor in altering sleep organization, because Megirian et al. (16) and Ryan et al. (23) showed that when normal rats breathe CO₂-enriched air their sleep organization remains unchanged compared with the breathing of ambient air. On the other hand, normal rats, breathing modest degrees of hypoxia (e.g., 15% O₂), exhibit significant disruption of their sleep organization at their neutral temperature, without reducing deep T_b (19). Such disruption is chiefly due to the difference in switching states, less so as a result in changes in the duration of each state (9, 19). Factors that may lead to alteration of the switching of states during sleep are arterial O₂ desaturations and mechanoreceptor afferent input (14), which increase during obstructive or central apneas, or both, in obese rather than in lean Zucker rats.

The thermoregulatory factor. Several groups of investigators (13, 17, 18, 21, 26, 27) have reported that obese Zucker rats have a significantly lower core T_b than do lean rats from early life into adulthood.

A thermoregulatory mechanism in ontogeny may underlie the difference we find in the sleep organization of the obese Zucker rat compared with its lean littermate. The pioneer studies of Glotzbach and Heller (8) and of Sakaguchi et al. (24) showed that systematic manipulation of hypothalamic and environmental temperatures of small adult mammals strongly influences sleep organization. From their studies and from the fact that the obese Zucker rat has an average T_b lower than that of its lean littermate from the sixth day of life (26), one may conclude that the rat has already shaped in infancy the definitive sleep organization for adult life. In the preweaning period, the obese Zucker rat's metabolic heat production shows the first signs of impairment, compared with the lean animal (13, 18, 27), and thus may partly explain its lower T_b. Also crucially important is the fact that the greatest consolidation of sleep organization is achieved in the preweaning period of the rat, which spans ~21 days (12). Therefore, it is the thermoregulatory factor in the lean vs. the obese Zucker rat that may play a role in determining the mechanism that solidifies its sleep architecture before obesity becomes severe enough to engender sleep-related disorders of breathing.

Clearly, further studies are needed to determine whether the ventilatory or the thermoregulatory factor explains the difference in sleep organization between lean and obese Zucker rats.

In summary, we provide evidence that the sleep organization of the obese Zucker rat differs from that of the lean animal. The difference is due to an increased frequency of switching from NREM sleep to W, with an accompanying decrease in the frequency of switching from NREM to REM sleep. The net effect is a greater amount of NREM sleep in obese compared with lean Zucker rats. This study lends additional support to the hypothesis (9, 18) that the frequency of switching states is a major mechanism that sculptures the sleep pattern. The resultant effect is that obese Zucker rats spend more time in NREM sleep than do lean Zucker rats.