The effect of leptin replacement on sleep-disordered breathing in the leptin-deficient ob/ob mouse

Abstract

Obese leptin-deficient (ob/ob) mice demonstrate defects in upper airway structural and neuromuscular control. We hypothesized that these defects predispose to upper airway obstruction during sleep, and improve with leptin administration. High-fidelity polysomnographic recordings were conducted to characterize sleep and breathing patterns in conscious, unrestrained ob/ob mice (23 wk, 67.2 ± 4.1 g, n = 13). In a parallel-arm crossover study, we compared responses to subcutaneous leptin (1 μg/h) vs. vehicle on respiratory parameters during NREM and REM sleep. Upper airway obstruction was defined by the presence of inspiratory airflow limitation (IFL), as characterized by an early inspiratory plateau in airflow at a maximum level (V̇imax) with increasing effort. The severity of upper airway obstruction (V̇imax) was assessed along with minute ventilation (V̇e), tidal volume (V_T), respiratory rate (RR), inspiratory duty cycle, and mean inspiratory flow at each time point. IFL occurred more frequently in REM sleep (37.6 ± 0.2% vs. 1.1 ± 0.0% in NREM sleep, P < 0.001), and leptin did not alter its frequency. V̇imax (3.7 ± 1.1 vs. 2.7 ± 0.8 ml/s, P < 0.001) and V̇e increased (55.4 ± 22.0 vs. 39.8 ± 16.4 ml/min, P < 0.001) with leptin vs. vehicle administration. The increase in V̇e was due to a significant increase in V_T (0.20 ± 0.06 vs. 0.16 ± 0.05 ml, P < 0.01) rather than RR. Increases in V̇e were attributable to increases in mean inspiratory flow (2.5 ± 0.8 vs. 1.8 ± 0.6 ml/s, P < 0.001) rather than inspiratory duty cycle. Similar increases in V̇e and its components were observed in non-flow-limited breaths during NREM and REM sleep. These responses suggest that leptin stabilized pharyngeal patency and increased drive to both the upper airway and diaphragm during sleep.

sleep-disordered breathing comprises a spectrum of disorders including obstructive sleep apnea and nocturnal hypoventilation. The former is characterized by recurrent episodes of upper airway obstruction during sleep, terminated by arousals and oxyhemoglobin desaturations (44), and the latter by sleep-related decreases in ventilatory drive. Obesity and central adiposity are potent risk factors for these disorders (14, 48). These factors have been associated with excess loads on upper airway (24, 41, 60). These loads can elicit neural responses in respiratory muscles that restore upper airway patency and increase ventilatory drive (10, 39). Current evidence suggests that defects in mechanical and neuromuscular control both play pivotal roles in pathogenesis of sleep-disordered breathing, although the exact mechanisms linking sleep-disordered breathing and obesity have not been elucidated.

A major impediment to investigating the pathogenesis of sleep-disordered breathing has been the paucity of suitable animal models. In early studies of upper airway function, we demonstrated that the pressure-flow dynamics and neuromechanical control of pharyngeal collapsibility (Pcrit) during anesthesia in the isolated canine and feline upper airway are comparable to that in sleeping humans (26, 28, 45, 46, 49–51, 55). Obesity can produce fixed pharyngeal narrowing without collapse in the Yucatan pig (25), and craniofacial defects can lead to obstructive apneas during REM sleep in the English bulldog (18, 19, 38, 56). Difficulties in manipulating sleep/wake state, body composition, and genetic background have curtailed the applicability of large animal models to studies of obesity and sleep apnea pathogenesis. In an effort to overcome these limitations, investigators have examined the impact of obesity in anesthetized rodents, and have demonstrated alterations in pharyngeal structure and function in the obese leptin-deficit ob/ob mice and in leptin-receptor deficient fa/fa rat (32, 35). They have discriminated effects of obesity and leptin on neuroventilatory control in conscious mice, and found that obese mice (ob/ob) lacking leptin demonstrated defects in neuroventilatory control, which were reversed by leptin administration (34, 42, 43). The effects of leptin on ventilatory patterns and the severity of upper airway obstruction, however, have not been examined during sleep.

The present study was undertaken to examine effects of leptin administration on breathing patterns in the obese leptin-deficient ob/ob mouse. We hypothesized that defects in upper airway and ventilatory control in this mouse confer susceptibility to upper airway obstruction and hypoventilation during sleep that would be ameliorated acutely by leptin replacement. To address these hypotheses, we utilized specialized, state-of-the-art whole body plethysmographic techniques to characterize the severity of upper airway obstruction and hypoventilation in non-rapid eye movement (NREM) and REM sleep.

METHODS

Mice

Male C57BL/6J-Lep^ob (ob/ob) leptin-deficient mice from Jackson Laboratory (Bar Harbor, ME) were housed for this study. Water and food were available ad libitum throughout the entire protocol. Mice were utilized at ∼23 wk of age for the baseline group. The ob/ob mice were placed in separate cages, in which their caloric intake and weight were monitored daily (34). All study protocols were approved by the Johns Hopkins University Animal Care and Use Committee (ACUC) and all animal experiments were conducted in accordance with ACUC guidelines. For all surgical procedures, 1–2% isoflurane anesthesia was administered through a facemask.

Experimental Procedures

Anesthesia protocol.

Isoflurane was used to induct and maintain anesthesia, as previously described (41). Atropine was injected (0.001 mg ip) to minimize airway secretions, rectal temperature was monitored, and body temperature was maintained at 36.5–37.5°C with a variable temperature heating pad. At the experiment's completion, the animals were euthanized by an overdose of pentobarbital (60 mg ip).

EEG/EMG headmount placement.

Mice were implanted with an EEG/EMG Headmount (no. 8201, Pinnacle Technology, Lawrence, KS) as previously described to determine sleep/wake states in polysomnography recordings. In brief, a longitudinal midline incision was made on the skull, and the underlying fascia was gently cleared from the surface of the skull. The headmount was secured with glue directly above the bregma, and two pairs of silver electrodes (Pinnacle Technology) were secured with silver conductive epoxy to frontal and parietal regions bilaterally. Insulated EMG leads were tunneled subcutaneously and placed over the nuchal muscles. The incision was sutured closed. Analgesia was administered in accordance with ACUC protocol, and the mouse was allowed at minimum 3-days observation before continuing in the protocol.

Leptin treatment.

A total of 13 obese ob/ob mice were used in the treatment protocol. In a group of 7 mice, a 100-μl Alzet (1.0 μl/h) osmotic pump (Cupertino, CA) filled with recombinant leptin (R&D, Minneapolis, MN) was inserted just caudal to the shoulder blades. Over a period of 3 days, the pump released leptin at a rate of 30 μg/day, leading to a predictable increase in leptin concentration, as previously described (34). A second group of 6 obese ob/ob were similarly inserted with a saline-filled osmotic pump.

Experimental Setup

Modified mouse whole body plethysmography.

The plethysmographic chamber (mouse whole body plethysmograph, Buxco, Wilmington, NC) consisted of a sealed animal chamber, a reference chamber, and a platform inside the chamber to support the mouse, as previously described (20). In brief, the chamber was designed to record tidal airflow, respiratory effort, and sleep-wake state continuously. Pressure differences between virtually sealed plethysmographic mouse and reference chambers were monitored to generate high-fidelity tidal volume and airflow signals.

The chamber was equipped with two ports (pneumotachographs) on the upper surface and with one large side port and three small side ports at the base, which were utilized to customize our system, as described (20). It was modified to increase the diameter of a port on the top of the chamber to accommodate the passage of EEG/EMG leads to the outside. The inflow and outflow ports were connected to positive and negative pressure sources to generate a steady biased flow through the mouse chamber. Mass flow controllers were placed at each port to match the level of inflow and outflow. High resistances with prolonged time constants were placed in series at the inflow and outflow ports to prevent significant attenuation of the tidal volume signal and maintain high-fidelity tidal volume and airflow signals. A reference chamber served to filter ambient noise. Slow leaks were created in the plethysmographic and reference chambers to maintain each chamber at atmospheric pressure. Respiratory effort was transduced from a sensor bladder on which the mouse lay. This bladder was used to transduce mechanical deformation of the torso during respiratory efforts, and was referenced to an identical bladder below the supporting platform. The sensor bladder can distinguish changes in effort within breaths and among a group of adjacent breaths. The magnitude of these respiratory effort swings were scaled in arbitrary rather than absolute units, since they reflected relative movements of the torso for periods when the mouse remained in a stationary position. Relative excursions could not be compared over prolonged periods since pressure deflections in the sensor bladder were influenced by any changes in mouse position during recordings and between recording sessions. EEG and EMG signals were also acquired by connecting electrodes to a head-stage preamplifier before passing leads through a sealed port in the roof of the chamber.

The Drorbaugh and Fenn equation was used to calculate the tidal volume signal from the plethysmography chamber pressure signal (11), as previously described (20). Application of this formula required the measurement of the following variables during each recording session: mouse rectal temperature, chamber temperature, room temperature, relative humidity, and chamber gas constant, calculated by utilizing a known volume injection and the resultant chamber pressure deflection.

During full polysomnographic recording sessions, the chamber was humidified to 90% relative humidity, and the mouse was allowed 30 min to acclimate to the chamber before recordings were initiated. All signals were digitized at 1,000 Hz (sampling frequency per channel) and recorded in LabChart 7 Pro (Version 7.2, ADInstruments, Dunedin, NZ). Tidal volume, respiratory movement, and EEG/EMG signals were recorded as previously described. The tidal volume signal was differentiated electronically to generate an airflow signal.

Study Design

A parallel-arm study design was conducted (see Fig. 1). Each mouse was recorded before, during, and after treatment with either vehicle or leptin. Respiratory parameters and sleep architecture were measured at each time point.

Fig. 1.

A parallel-arm study design comparing leptin vs. vehicle intervention was conducted with measurements of respiratory parameters and sleep architecture before, during, and after intervention. Osmotic pumps were loaded with either leptin or vehicle and implanted after completing baseline polysomnographic recordings. These recordings were repeated 3 days after initiating treatment. The osmotic pump was then explanted on day 4. Once body weight began to rise again, postintervention washout recordings were repeated. Washout recordings were also performed at the same time point in the vehicle-treated mice.

Ob/ob mice were placed in separate cages, and food consumption and body weight were monitored daily throughout the protocol. After headmount placement and recovery, all mice underwent baseline recordings at ∼23 wk from approximately 9:30 a.m. to 4:00 p.m. (see Table 1). Osmotic pumps were loaded with either leptin or vehicle and implanted after completing baseline polysomnographic recordings. These recordings were repeated 3 days after initiating treatment. The osmotic pump was then explanted on day 4. The leptin-treated mice continued to lose weight for several days, reflecting persistent leptin activity. Once body weight began to rise again, postintervention washout recordings were repeated. Washout recordings were also performed at the same time point in the vehicle-treated mice.

Table 1.

Characteristics by age, weight, and temperature by treatment

		Age, wk	Weight, g		Temperature, °C
ob/ob	n	Mean ± SE	Mean ± SE	Range	Mean ± SE	Range
Vehicle treatment
Pre	6	22.8 ± 0.4	70.0 ± 2.5	67.1–73.5	36.1 ± 1.0a	35.1–37.1
During	6	23.0 ± 0.0	69.4 ± 2.7*	66.8–73.5	35.9 ± 0.5	35.4–36.7
Post	6	25.0 ± 0.0***	70.6 ± 2.4*	68.5–74.8	35.7 ± 0.3	35.2–36.0
Leptin treatment
Pre	7	22.7 ± 0.1	66.7 ± 3.4	61.7–71.3	37.0 ± 1.2	34.7–38.0
During	7	23.7 ± 0.1	63.2 ± 3.8***	56.3–68.3	39.3 ± 0.5***	38.4–39.8
Post	7	25.0 ± 0.0***	64.6 ± 3.9***	57.3–69.6	37.5 ± 1.3	34.8–38.3

Values are presented as means ± SE. Baseline characteristics for Pre, During, and Posttreatment time points in ob/ob mice.

^*P < 0.05 vs. pretreatment.

^***P < 0.001 vs. pretreatment.

^an = 3.

Experimental Protocol

Tidal volume, airflow, and respiratory movement, and EEG/EMG signals were recorded continuously in the plethysmography chamber. Each mouse recording was evaluated in RemLogic 1.3 polysomnographic software (Embla PSG Software, Natus Medical, Pleasanton, CA), to determine sleep-wake, ventilatory parameters, and sleep-disordered breathing indexes. Respiratory signals were analyzed from all REM sleep and from 20-s periods of NREM sleep sampled every 30 min throughout the sleep recording. Custom software was used to demarcate the start and end of inspiration and expiration for subsequent calculations of timing and amplitude parameters for each respiratory cycle.

Sleep-wake.

Sleep-wake state was scored visually in 5-s epochs from 10:00 a.m. until 4:00 p.m (after ∼30 min acclimatization to the plethysmographic chamber). Standard criteria were employed to score sleep-wake state based on EEG and EMG frequency content and amplitude, as previously described (54). Wakefulness was characterized by low-amplitude, high-frequency (∼10 to 20 Hz) EEG waves and high levels of EMG activity compared with the sleep states. Non-rapid eye movement (NREM) sleep was characterized by high-amplitude, low-frequency (∼2 to 5 Hz) EEG waves with EMG activity considerably less than during wakefulness. Rapid eye movement (REM) sleep was characterized by low-amplitude, mixed-frequency (∼5 to 10 Hz) EEG waves with EMG amplitude either below or equal to that during NREM sleep.

Ventilatory parameters.

The instantaneous respiratory rate (RR, breaths/min) was calculated as the reciprocal of the respiratory period, and the instantaneous minute ventilation (V̇e, ml/min) was given by the product of the respiratory rate and tidal volume for each breath. V̇e was also represented by the product of the mean inspiratory flow rate and inspiratory duty cycle, as previously described (30). The severity of airflow obstruction was defined by the level of maximal inspiratory flow (V̇imax) during inspiratory flow-limited breaths (as defined immediately below), and was measured at the point of peak inspiratory airflow (40, 58).

We then utilized the airflow and respiratory effort signals to develop an algorithm for detecting upper airway obstruction during sleep. Obstruction was characterized by the development of inspiratory airflow limitation (IFL), which is the cardinal feature in humans who snore and have obstructive sleep apnea (15). IFL is marked by an inspiratory flow plateau at a maximal level, despite continued increases in breathing effort.

The algorithm for detecting IFL was developed based on airflow timing and amplitude parameters for each respiratory cycle as follows. Custom software outputted a discrete value for inspiratory flow (V̇i) at the midpoint (V̇i50) and values of peak inspiratory flow for the first and second half of inspiration (early V̇i max and late V̇i max). Breaths were considered to be non-flow limited when a plateau in midinspiratory flow could not be discerned from flow timing indexes. Specifically, sniffs were defined as breaths of short inspiratory duration (defined as 1.75SD less than the mean of the entire sample) and excluded. For the remaining breaths, we detected a plateau in midinspiratory flow, as defined by one of the following criteria: 1) breaths with an early peak of inspiratory flow followed by a plateau or even a decrease in flow thereafter (often referred to as negative effort dependence), 2) a midinspiratory plateau of sufficient duration (≥ 20% of inspiratory time between early and late inspiratory peaks in flow), or 3) when the late inspiratory airflow peak exceeded the early peak, these peaks (with outward convexities) flanked a midinspiratory period of inward convexity.

The computerized algorithm was initially developed and subsequently validated as follows. Two independent expert scorers visually assessed for presence of IFL in four randomly distributed NREM and REM sections (∼50 breaths per section) of airflow and effort signals for 10 different mouse recordings. Discrepancies between experts were adjudicated by consensus to develop the gold standard. The algorithm was developed and the thresholds were adjusted from a reduced sample of one NREM and one REM section from five separate recordings. The finalized algorithm was validated against the gold standard scores in the remaining recordings, and was found to have a sensitivity, specificity, positive predictive value, and negative predictive value of 86.1%, 94.1%, 80.1%, and 96.1% respectively.

Statistical Analysis

Statistical analyses were structured to examine responses in mouse characteristics, sleep architecture, and ventilatory parameters as a function of treatment (leptin vs. vehicle) and time point (before, during, and after treatment). In stratified analyses, we utilized multivariable regression methods to examine whether parameters changed significantly within each treatment groups. Specifically, fixed effects of treatment and time point (before, during, and after) on outcomes were assessed while accounting for random variations among mice with XTMIXED (STATA 12, Statacorp LP, College Station, TX). This procedure allowed us to model changes in outcomes during and after leptin and vehicle treatments from before treatment values. When outcomes before and after treatment did not differ significantly, baseline mean values for these time points were calculated, and were used subsequently for additional analyses. Mixed effects linear regression was utilized to test a priori hypotheses that ventilatory parameters responded differently to treatment with leptin vs. vehicle. XTMIXED was used to model effects of the primary independent variables (treatment, time point, and treatment by time point interaction) on outcome measures while accounting for random effects among mice (since outcome variables were assessed repeatedly over time). Separate analyses were also performed to examine leptin's effect on the severity of upper airway obstruction and ventilatory drive, respectively. In the flow-limited condition, V̇imax was considered a marker for severity of upper airway obstruction (40, 41). During non-flow-limited breathing, however, the mean inspiratory flow rate was considered to be an index of ventilatory drive to the diaphragm and respiratory pump muscles, as previously described (31, 47). Values were expressed as means ± SE, and statistical significance was inferred at a P < 0.05 level.

RESULTS

Mouse Characteristics Before, During, and After Intervention

Mouse characteristics before, during, and after intervention are illustrated in Table 1. By design, age was significantly higher (P < 0.001) before compared with after intervention. Body temperature increased in the leptin-treated group during intervention (P < 0.001), consistent with recognized effects of leptin on basal metabolic rate (6). Body weight fell significantly in leptin-treated mice during and after intervention (P < 0.001) compared with the before-intervention time point, but only fell minimally in the vehicle-treated mice during and after intervention (P < 0.05), after implanting the osmotic pump. The weight was significantly lower in the leptin- vs. vehicle-treated group during and after intervention (P < 0.01), consistent with leptin's effects on metabolism and appetite.

Representative Sleep Recording

A representative polysomnography recording during REM sleep is illustrated before and during leptin treatment in Fig. 2. Before leptin administration, airflow and effort waveforms showed evidence of inspiratory flow limitation (Fig. 2, left panel, see *). In contrast, IFL decreased and V̇imax increased after leptin administration, indicating that the severity of upper airway obstruction decreased (Fig. 2, right panel).

Fig. 2.

Sleep study recording example before and during leptin treatment. Inspiratory flow limitation (IFL, see *, left panel) and obstructive hypopneas (see horizontal bars, middle panel) were observed in ob/ob mice during rapid eye movement (REM) sleep. The events were abolished by systemic leptin administration (right panel). The shaded area on the middle panel is expanded on the left panel. a.u., Arbitrary units. See text for details.

Sleep Architecture

Sleep architecture is described before, during, and after intervention for leptin and vehicle intervention groups in Table 2. Total sleep time decreased significantly in the leptin-treated mice during intervention (P < 0.01), and decreased significantly in the vehicle-treated mice both during and after intervention (P < 0.005 and P < 0.001, respectively). These decreases can be attributed to surgical effects of implanting osmotic pumps.

Table 2.

Sleep architecture by time point and treatment

				Bouts
	Sleep, min			Number		Average Length (min)
	Total	NREM	REM	NREM	REM	NREM	REM
Vehicle Treatment
Pre	194 ± 4.7	184 ± 5.0	10.4 ± 0.5	171 ± 9	8 ± 0	1.1 ± 0.0	1.4 ± 0.9
During	132 ± 10.7**	121 ± 10.0**	11.4 ± 1.1	111 ± 5**	10 ± 1	1.1 ± 0.1	1.2 ± 0.5
Post	112 ± 3.4***	103 ± 2.8***	9.3 ± 0.9	118 ± 4*	7 ± 1	0.9 ± 0.0	1.4 ± 1.2
Leptin Treatment
Pre	238 ± 6.6	225 ± 6.8	13.0 ± 0.7	158 ± 5	9 ± 0	1.4 ± 0.0	1.4 ± 0.1
During	148 ± 6.6**	129 ± 6.9***	19.0 ± 0.7**	138 ± 7	18 ± 1***	0.9 ± 0.0***	1.1 ± 0.0
Post	193 ± 10	186 ± 9.6	7.2 ± 0.8**	153 ± 11	5 ± 1*	1.3 ± 0.0	1.7 ± 0.5

Baseline sleep characteristics are shown for Pre, During, and Posttreatment time points and are expressed as means ± SE. NREM, non-rapid eye movement sleep; REM, rapid eye movement sleep.

^*P < 0.05 vs. pretreatment.

^**P < 0.005 vs. pretreatment.

^***P < 0.001 vs. pretreatment.

The total number of NREM bouts decreased significantly in the vehicle-treated mice during (P < 0.01) and after intervention (P < 0.05), but not in leptin-treated mice. The average NREM bout length decreased significantly during intervention compared with time points before and after intervention (P < 0.001). In contrast, NREM bout length did not differ by time point in vehicle-treated mice.

The total number of REM bouts increased significantly during intervention compared with time points before and after intervention in the leptin-treated mice (P < 0.001), but did not differ across time in the vehicle-treated mice. The average REM bout length did not differ over time in leptin- or vehicle-treated mice. In leptin-treated mice, the increase in the number of REM bouts accounts for the observed increase in total REM sleep time during vs. before (P < 0.001) and after (P < 0.005) intervention. In contrast, REM sleep time did not change significantly in vehicle-treated mice.

Effect of Leptin Treatment on Ventilatory Parameters

Frequency of flow-limited breathing. As described in Table 3, IFL was highly prevalent in REM sleep and was significantly greater in REM compared with NREM sleep (P < 0.001). Nonetheless, the frequency of IFL breaths did not change significantly over the course of the intervention period in both the leptin- and vehicle-treated groups.

Table 3.

Inspiratory flow limitation prevalence by sleep stage

	Inspiratory Flow Limitation
NREM
Vehicle treatment
Pre	0.4% ± 0.1%
During	0.1% ± 0.0%
Post	2.1% ± 0.2%
Leptin treatment
Pre	0.6% ± 0.1%
During	4.3% ± 0.2%
Post	1.9% ± 0.1%
REM
Vehicle treatment
Pre	36.4% ± 0.4%
During	24.2% ± 0.3%
Post	44.3% ± 0.4%
Leptin treatment
Pre	33.9% ± 0.2%
During	36.5% ± 0.2%
Post	41.5% ± 0.4%

Values are presented as means ± SE. The frequency of inspiratory flow limitation by sleep stage: NREM and REM.

Effect of leptin treatment on flow-limited breathing (Fig. 3).

Fig. 3.

Effect of leptin on respiratory parameters during obstructed breaths (upper airway obstruction) (flow-limited breaths): non-rapid eye movement (NREM) and REM sleep. *P < 0.05, **P < 0.01, ***P < 0.001 compared with baseline. #P < 0.001 between leptin and vehicle treatment.

During flow-limited breathing, leptin increased V̇imax compared with baseline levels in NREM and REM sleep and compared with vehicle treatment in REM sleep. The latter accounts for concomitant increases in minute ventilation during leptin administration. Tidal volume also increased during leptin administration compared with baseline and compared with vehicle treatment (only in REM sleep), whereas respiratory rate did not change. The mean inspiratory flow increased during leptin treatment compared with baseline in NREM and REM sleep and compared with vehicle treatment (only in REM sleep). In addition, leptin treatment led to modest decreases in the inspiratory duty cycle compared with baseline in NREM and REM sleep, consistent with relief of upper airway obstruction (22, 47).

Effect of leptin treatment on non-flow-limited breathing (Fig. 4).

Fig. 4.

Effect of leptin on respiratory parameters during periods of unobstructed breaths (ventilatory control) (non-flow -limited breaths): NREM and REM sleep. *P < 0.05, *** P < 0.001 compared with baseline. †P < 0.05, ‡P < 0.01, #P < 0.001 between leptin and vehicle treatment.

During non-flow-limited breathing, leptin treatment increased minute ventilation compared with baseline and vehicle treatment in both NREM and REM sleep. This increase in minute ventilation can be attributed to increases in both tidal volume and respiratory rate. It was also associated with increases in mean inspiratory flow without changes in inspiratory duty cycle, suggesting increases in ventilatory drive (30, 31). In contrast, vehicle administration did not result in any significant changes in minute ventilation or its components across time points.

DISCUSSION

In this study, we found that leptin-deficient ob/ob mice demonstrated marked increases in the frequency of inspiratory flow limitation during REM compared with non-REM sleep. Leptin administration increased ventilation significantly from levels before and after intervention. These increases occurred in flow-limited and non-flow-limited breaths during both NREM and REM sleep, and were attributable to increases in tidal volume rather than respiratory rate. During non-flow-limited breathing, elevations in ventilation were associated with increases in mean inspiratory flow rather than inspiratory duty cycle, suggesting that leptin stimulated ventilatory drive to the respiratory pump muscles (30, 34). During flow-limited breathing, leptin increased maximal inspiratory airflow and decreased the inspiratory duty cycle during flow-limited breathing in REM sleep, consistent with recognized effects of leptin on pharyngeal patency (40, 52). Improvements in upper airway patency can account for increased time spent in REM sleep in leptin- compared with vehicle-treated mice. Taken together, our findings suggest that leptin can ameliorate hypoventilation and upper airway obstruction in obese mice, particularly during REM sleep.

Two types of animal models have been utilized to study the pathogenesis of obstructive sleep apnea. First, investigators have elucidated fundamental mechanisms of upper airway obstruction in highly instrumented isolated upper airway preparations of anesthetized dogs, pigs, cats, rats, and mice (25, 28, 32, 35, 40, 41, 45, 46, 49–51, 55). Specifically, they have demonstrated that stimulating pharyngeal neuromuscular activity decreases airway compliance (13), resistance (36, 37), and collapsibility (40, 41, 51). Nevertheless, extrapolation is required to predict effects of physiological changes in upper airway mechanical properties on airflow obstruction during sleep. To overcome this limitation, investigators have looked for evidence of upper airway obstruction during sleep in English bulldogs and cats, whose upper airway anatomy and/or posture make them susceptible to obstruction (33, 56). These species exhibit upper airway obstruction and frank apneas and hypopneas during REM sleep when pharyngeal neuromuscular activity is known to decline. The ob/ob mouse also demonstrated evidence for upper airway obstruction primarily during REM sleep, which could be related to excess adiposity in pharyngeal and lingual structures (4–6). Upper airway obstruction in the ob/ob mouse, however, was relieved by short-term leptin administration prior to significant alterations in body weight and composition. This finding suggests that leptin compensates for reductions in neuromuscular drive in ob/ob mice to restore pharyngeal patency (6, 12, 27, 39). Thus it is likely that both increases in upper airway adiposity and reductions in neural drive to airway muscles account for the development of upper airway obstruction during REM sleep.

Two mechanisms can account for observed increases in maximal inspiratory airflow. First, during flow-limited inspirations, dynamic collapse of the upper airway caused inspiratory airflow to plateau at a maximal level (see V̇imax on * breaths, Fig. 2) as effort continued to increase. Rigorous methods were deployed to confirm the presence of inspiratory flow limitation as airflow and effort diverged. Our assessment of IFL is based on a well-validated respiratory effort signal that provides an accurate representation of tracheal pressure swings (20), and validated objective algorithms that automate our assessment of inspiratory airflow limitation (see methods). In fact, our high-fidelity airflow and effort signals allowed us to discern cardinal features of inspiratory flow limitation including modest reductions in airflow beyond the point of maximal inspiratory flow (consistent with the recognized phenomenon of negative effort dependence in flow-limited regimes) as well as associated increases in inspiratory effort and duty cycle during flow-limited compared with non-flow-limited breaths (see Fig. 2).

Utilizing methods, we have determined that leptin increases V̇imax for flow-limited inspirations compared with baseline measurements in leptin and vehicle control groups. This increase in maximal inspiratory airflow can be largely attributed to decreases in pharyngeal collapsibility (40) rather than resistance upstream to the site of collapse (which is fixed by the caliber of the nasal airway). In a previous study, we demonstrated that reductions in pharyngeal collapsibility could be attributed to acute increases in upper airway neuromuscular activity rather than anatomic remodeling of upper airway structures (40). Although the mechanism by which leptin stimulates upper airway dilator activity has not been established, leptin signaling in the brain occurs via the long isoform of leptin receptor, ObRb, which is the only functional isoform of leptin receptor (57). Data on localization of the ObRb receptor in the hypoglossal nucleus providing the motor output to the upper airway muscles remain controversial (29). However, leptin may regulate the upper airway motor neurons via intermediate pathways such as pro-opiomelanocortin (POMC)/melanocortin receptor 4 (MC4R) (8, 9), neurotensin (16, 17), and orexin (7, 16, 53, 61) in the hypothalamus and/or nucleus tractus solitarius (1, 21). Current evidence suggests that leptin's central actions can improve metabolic profiles in skeletal muscles, which could also help restore pharyngeal patency (1).

During non-flow-limited breathing, however, the upper airway does not collapse or produce dynamic inspiratory airflow obstruction. Under these circumstances, observed increases in mean inspiratory airflow during leptin administration can be attributed to increases in respiratory drive rather than decreases in pharyngeal collapsibility (31, 47). Mean inspiratory flow rates, a recognized index of ventilatory drive, also increased in non-flow-limited breaths in NREM and REM sleep, supporting the concept that leptin increased drive to the diaphragm during sleep (34). In studies in ob/ob mice, we have demonstrated leptin replacement also increases basal CO₂ production and decreases Pa_CO2 (6, 34). The latter finding suggests that observed increases in ventilation resulted from direct effects of leptin on ventilatory drive rather than metabolic rate. Its effect on hypercapneic ventilatory drive may be mediated by receptors in the nucleus of the tractus solitarius (NTS), or the retrotrapezoid nucleus/parafacial respiratory group (RTN/pFRG) (2, 3, 21). This increase in ventilatory drive during leptin administration, however, would also be expected to increase inspiratory swings in tracheal pressure and the tendency for the upper airway to flow-limit accordingly (as described above). Our findings confirm that despite improvements in upper airway patency (increases in V̇imax), flow-limited breathing persisted during leptin administration, consistent with the notion that leptin stimulated concomitant increases in neural drive to upper airway and respiratory pump muscles.

Several limitations should be considered in interpreting our findings. First, our recording methods may have interfered with natural sleep cycles and the overall expression of sleep-disordered breathing in our mouse model. Specifically, mice were tethered to EEG and EMG leads in a relatively small whole body plethysmographic chamber, which may account for observed alterations in sleep architecture including reductions in sleep time and efficiency. Nevertheless, the primary objective of the present study was to quantify effects of leptin on ventilation during natural sleep for which well-validated, novel plethysmographic methods were adopted to generate high-fidelity measurements of tidal airflow, tidal volume, respiratory timing indexes, and respiratory effort along with sleep staging parameters (20). Second, we recognize that surgical procedures to implant and explant osmotic pumps over the course of our treatment trial may have also induced alterations in sleep architecture and respiration. Nevertheless, our study implemented a cross-over parallel design that allowed us to account for nonspecific treatment effects by comparing responses both within and between treatment groups. Third, we acknowledge that inferences drawn about leptin's effects on ventilatory control have been drawn from noninvasive measurements of respiratory parameters without direct measurements of CNS neural activity in upper airway or respiratory motor nuclei. Nevertheless, our conclusions are supported by well validated physiological correlates of CNS drive to pharyngeal and ventilatory musculature, which could warrant further invasive studies in reduced preparations. Fourth, we acknowledge that our assessment of respiratory effort relied on relative rather than absolute changes in thoracoabdominal movements, which could not be compared over prolonged periods within and between recording sessions. Nevertheless, our effort signal allowed us to discriminate obstructed (flow-limited) from unobstructed (non-flow limited) inspirations within short sequences of breaths. Finally, we recognize that concomitant weight loss may have confounded effects of leptin on the severity of upper airway obstruction during sleep. Nevertheless, the severity of upper airway obstruction did not correlate with changes in weight during the leptin administration and washout periods as follows. Upper airway obstruction decreased during leptin administration, even though weight did not change significantly at this time point. In contrast, weight decreased substantially after leptin washout, yet upper airway obstruction worsened again, and returned to baseline levels despite the loss of body weight.

Our study has significant diagnostic and therapeutic implications for our understanding of the pathogenesis of sleep-disordered breathing. The ob/ob mouse serves as a versatile model for elucidating effects of genetic, environmental and/or pharmacological factors on key neuromechanical determinants of sleep-disordered breathing. Our findings imply that anatomic alterations from excess adiposity in this model (6) predispose to upper airway obstruction (23), which develops during sleep when neuromotor tone declines (44). This model can also be utilized to quantify the effects of these factors on nocturnal ventilation and respiratory patterns during sleep. Therapeutically, respiratory and pharyngeal responses to leptin administration indicate that leptin can relieve nocturnal hypoventilation and upper airway obstruction during sleep. Current evidence suggests that in contrast to ob/ob mice, humans with sleep apnea and obesity are leptin-resistant rather than leptin-deficient, and that leptin-resistance improves with nocturnal ventilatory support (59). Nevertheless, observed responses in this mouse model suggest that therapeutic strategies to enhance leptin sensitivity may be effective in treating obstructive sleep apnea and/or nocturnal hypoventilation syndromes (e.g., caloric restriction, insulin sensitizing agents). Further work is also required to elucidate neural pathways responsible for leptin's action on upper airway and ventilatory control.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-050381, HL-080105, and 1R01-HL-128970-01.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

Author contributions: H.P., A.B.H., R.S.A., E.B.L., J.P.K., H.S., P.L.S., V.Y.P., and A.R.S. conception and design of research; H.P., A.B.H., and E.B.L. performed experiments; H.P., R.S.A., S.V.K., and A.R.S. analyzed data; H.P., R.S.A., S.V.K., J.P.K., H.S., P.L.S., V.Y.P., and A.R.S. interpreted results of experiments; H.P., V.Y.P., and A.R.S. prepared figures; H.P., V.Y.P., and A.R.S. drafted manuscript; H.P., A.B.H., R.S.A., E.B.L., S.V.K., J.P.K., H.S., P.L.S., V.Y.P., and A.R.S. edited and revised manuscript; H.P., A.B.H., R.S.A., E.B.L., S.V.K., J.P.K., H.S., P.L.S., V.Y.P., and A.R.S. approved final version of manuscript.