INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The incidence of obesity is increasing dramatically in western society, especially in the U.S. where 22% of the population have a body mass index (BMI) greater than 30 kg/m² (1). As obesity progresses an increase in the resting partial pressure of carbon dioxide in arterial blood (Pa_CO2) may develop in some individuals (2, 3), leading to obesity hypoventilation, respiratory failure, and premature death (4, 5). In obese individuals who maintain normal ventilation, respiratory failure and death can be averted. Thus, cardiorespiratory morbidity and mortality appear related to the degree of respiratory depression that develops when people gain weight.

It has been over 40 yr since the obesity hypoventilation syndrome (Pickwickian syndrome) was first described (2). To date, there is no explanation why a minority of severely obese people develop hypoventilation and others breathe normally. Debate has centered around whether the problem is mechanical, due to loads imposed by obesity on the respiratory system or upper airway. Alternatively, a "central" component has long been suspected because the majority of severely obese people breathe normally (3) and because substantial respiratory muscle reserve is present even in those who hypoventilate (6). Furthermore, hypoventilation may be exacerbated by the reduced neural drive to ventilation during sleep (7). As yet, however, no factor has been identified that is associated with both obesity and respiratory depression.

It is now recognized that obesity is associated with elevated leptin (10), a known satiety factor that acts through neural pathways in the hypothalamus to suppress appetite (11). A mutation in the gene that encodes leptin leads to profound obesity in mice (14) and in rare cases in humans (15, 16). We have previously examined ventilatory control in the C57BL/ 6J-Lep^ob mouse. This mutant has a reduced ventilatory response to CO₂ challenge, and the response appears independent of the presence of obesity (17). These data suggest that the absence of leptin is associated with a reduced hypercapnic ventilatory response (HCVR). Conversely, it could be hypothesized that leptin replacement acts to restore ventilatory control in C57BL/6J-Lep^ob mice regardless of weight.

We tested this hypothesis by examining the relationship between leptin and ventilation in C57BL/6J-Lep^ob mice. Specifically, we predicted that: (1) the absence of leptin is associated with an elevated Pa_CO2; (2) leptin replacement reverses hypoventilation and restores ventilatory responses to CO₂; (3) the effect of leptin on central ventilatory control is not secondary to metabolic changes in peripheral tissues.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Twenty-nine mutant, obese male C57BL/6J-Lep^ob mice and 43 wild-type male C57BL/6J mice from Jackson Laboratory (Bar Harbor, ME) 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 (60 mg intraperitoneally).

Surgical Procedures

Polysomnography. Mice were instrumented with chronically implanted polysomnographic electrodes for determination of sleep/wake state as previously described (18). A midline incision was made to expose the skull and two bilateral pairs of holes were drilled through the skull in the frontal and parietal regions. Four electroencephalographic (EEG) electrodes were fashioned from Teflon-coated stainless steel wire. The end of each electrode was stripped of 0.15 cm of Teflon, bent at 90 degrees, and inserted into the skull through each predrilled hole. The four electrodes were bonded to the dorsal surface of the skull with dental acrylic (Land Dental, Wheeling, IL). Two nuchal electromyographic (EMG) electrodes were stitched (6-0 silk) flat onto the surface of the muscle immediately posterior to the dorsal area of the mouse skull. The skin overlying the skull and posterior muscles was reapposed and the six electrodes exited the skin dorsally between the shoulder blades.

Carotid arterial catheterization. Chronic catheterization of the carotid artery was performed as previously described (18). Arterial blood gas samples of 80 to 100 µl were drawn into a Hamilton syringe, placed on ice, and immediately analyzed on a blood gas analyzer (IL BG3; Instrumentation Laboratory, Lexington, MA). The blood loss was replaced with an infusion of 200 to 300 µl of saline over 2 to 3 min. A single blood gas measurement was made during periods of quiet wakefulness in conscious, unrestrained mice between 24 and 96 h postsurgery. Data were analyzed only from mice that survived and maintained weight for a 4-wk period after catheterization.

The success rate for arterial catheterization was approximately 90% in wild-type mice, but less than 20% in mutant mice. However, in all mice that survived long-term and were included in data analysis the pH was normal and the respiratory rate appeared normal. The C57BL/6J-Lep^ob mice not included in analysis developed an acute acidosis, leading to respiratory failure over a 2- to 3-day period. This was presumably related to some factor or factors associated with obesity because much higher success rates were obtained in wild-type mice. Also, the C57BL/6J mouse has an incomplete circle of Willis which may predispose to areas of ischemia with the loss of blood flow through the left carotid artery after the catheterization procedure. We were also unsuccessful in a number of attempts to measure arterial blood gases in C57BL/6J-Lep^ob mice after leptin replacement.

Insertion of osmotic pumps. A small 0.25-cm incision was made in the midline, just caudal to the shoulder blades and a 100-µl Alzet (Mt. View, CA) osmotic pump which delivers 1 µl/h inserted subcutaneously. The surgical procedure required only 5 to 10 min of anesthesia.

Ventilatory Measurements during Sleep/Wake

Plethysmography. During data collection periods, the animals were placed in a whole body barometric plethysmography chamber to measure ventilation as previously described (17). The chamber (823 ml³) was customized to allow the animal free movement while EEG and EMG electrodes were exited through a sealed port. The plethysmograph was referenced to a second chamber and flushed with compressed, humidified air (80% relative humidity) at approximately 600 ml/min. To measure ventilation, the ports through which the gases enter and exit the chamber were closed to produce a constant chamber volume. Once the chamber was at a constant volume, tidal volume (VT), and respiratory frequency (f) were measured from changes in pressure (PM15E differential pressure transducer; Statham Gould, Hato Ray, PR) caused by inspiratory and expiratory temperature fluctuations (17). The body temperature was assumed to be constant at 37° C (note: before, during, and after leptin infusion the acute mean rectal temperatures for both wild-type and mutant groups were not different and ranged between 37.4 ± 0.1° C and 37.9 ± 0.2° C). Calibration injections of 10, 20, and 30 µl of room air were made with the animal inside the constant volume chamber. Minute ventilation (E) is reported as the product of f and VT.

Respiratory control protocol. Sleep/wake state was assessed from EEG and EMG recordings as previously described (18). Wakefulness was characterized by low-amplitude, high-frequency (approximately 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 (approximately 2 to 5 Hz) EEG waves and an EMG activity considerably less than during wakefulness. Rapid eye movement (REM) sleep was characterized by low-amplitude, mixed-frequency (approximately 5 to 10 Hz) EEG waves, although the predominant pattern was a fixed amplitude theta frequency consistent with hippocampal theta rhythm. During REM sleep the EMG activity was either equal to or less than that seen during NREM sleep, but always less than that seen during wakefulness.

Ventilation was measured during wakefulness, NREM and REM sleep in response to a range of hypercapnic gases (0, 3, 5, and 8% CO₂ in 40% O₂ to ensure no hypoxic stimulus). The animals were allowed to freely cycle through sleep/wake states. In each sleep/wake state, a single gas challenge was introduced to the mouse for 3 to 4 min, before the ports through which the gases enter and exit the chamber were closed to produce a constant chamber volume and record ventilation. Over the course of the experiment, we challenged the mice twice during wakefulness and NREM sleep over a range from 0 to 8% CO₂, and once during REM sleep over a range from 0 to 5% CO₂ (the animals frequently aroused from REM sleep during 8% CO₂ challenge).

Experiments were conducted in eight male C57BL/6J-Lep^ob mice (140 ± 7 d) and eight male C57BL/6J mice (131 ± 11 d). Ventilatory control was examined under three conditions: (1) control condition after monitoring food intake and weight over 4 d; (2) leptin conditionafter 3 d of leptin infusion at 30 µg/d administered via a subcutaneous osmotic pump; (3) recovery condition4 d after removal of the osmotic pump. Experiments were conducted between 10:00 A.M. and 5:00 P.M. and lasted 5 to 7 h.

VT and frequency were assessed over 5 to 8-s periods during wakefulness and NREM sleep. In REM sleep, VT and f were highly variable. Thus, ventilatory measurements were made over the entire period of REM sleep which had to be 30 s or longer for inclusion. The hypercapnic ventilatory response (HCVR) was determined in each animal by the slope of the relationship between E and inspired CO₂ (0-8%) during wakefulness and NREM sleep via linear least-squares regression analysis. During REM sleep the HCVR was calculated over 0-5% because the animals could not consistently maintain 30-s periods of REM sleep during 8% CO₂ challenge.

Measurement of Metabolism

Metabolic parameters were measured in eight C57BL/6J-Lep^ob mice (136 ± 7 d) over a 2-h period using indirect, open-circuit calorimetry with a four chamber Oxymax system (Columbus Instruments, Columbus, OH). Data from the first half hour of equilibration were discarded. Metabolism was measured during control, leptin, and recovery conditions in a comparable protocol to that used to generate the ventilatory control data described previously.

Food Restriction

Food intake was restricted in six C57BL/6J-Lep^ob mice (122 ± 5 d) to levels comparable with leptin infusion (Day 1, 2.0 g; Day 2, 0.6 g; Day 3, 0.1 g). A control experiment (food ad libitum) was conducted first in four of the animals and second in two of the animals.

High-fat Diet

Eight wild-type C57BL/6J mice were fed a high-fat diet (49% fat; 5.8 kcal/g) for approximately 16 wk until they reached 45 to 50 g body weight (190 ± 3 d). The mice were instrumented for polysomnography and ventilation was determined by barometric plethysmography. A second group of eight wild-type C57BL/6J mice (205 ± 2 d) fed a normal diet acted as a control group.

Measurement of Plasma Leptin Levels

Arterial blood (0.8 to 1.2 ml) was obtained from the carotid artery under halothane anesthesia using catheterization techniques described previously. After blood withdrawal the animals were euthanized with pentobarbital (60 mg, intraperitoneally). Plasma leptin levels were measured with a mouse leptin radioimmunoassay kit from Linco Research, Inc. (St. Charles, MO). The intra-assay coefficient of variation was 7.2%.

Statistical Analyses

Data were analyzed using Crunch 4 (Crunch Software Corporation, Oakland, CA), and results presented for E show the mean ± SEM. Slopes for each animal were pooled and results presented show the mean ± SEM. Statistical significance between control, leptin, and recovery conditions for E and the HCVR was derived within each sleep/wake state using one-way, within-subject analysis of variance (ANOVA) with Newman-Keuls post hoc analysis.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Obesity Hypoventilation in the C57BL/6J-Lep^ob Mouse

We measured Pa_CO2 from arterial blood in conscious, unrestrained, adult C57BL/6J-Lep^ob and age-matched wild-type mice. The mutant mice exhibited a significantly elevated Pa_CO2 (> 10 mm Hg; Figure 1) compared with wild-type mice. In Figure 2 we show how raising the inspired level of CO₂ increases E less in a mutant mouse compared with a wild-type mouse. The slope of this relationship, or the HCVR, was reduced across all three sleep/wake states in the mutant mouse. These data demonstrate that the obese mutant mouse exhibits abnormalities in the relationship between ventilation and CO₂ that define obesity hypoventilation.

View larger version (15K): [in this window] [in a new window]   Figure 1.   PaCO2 is elevated in C57BL/6J-Lepob mice compared with wild-type mice. Results presented show the mean ± SEM. Statistical significance was derived using a Student's unpaired t test (two-way).

View larger version (14K): [in this window] [in a new window]   Figure 2.   The relationship between E and inspired CO2 is blunted across all three sleep/wake states in one adult obese C57BL/6J- Lepob compared with one adult lean wild-type mouse.

Leptin Replacement and Stimulation of Ventilation

To determine whether leptin is a respiratory stimulant, we examined the effect of leptin infusion (30 µg/d for 3 d) on ventilation in obese mutant mice and lean wild-type mice during wakefulness and sleep. In mutant mice leptin replacement resulted in an incremental reduction in food intake to zero by Day 3 (Figure 3). The reduced food intake was accompanied by a 6.5 ± 1.1 g decrease in weight from the control condition (Figure 3, left arrow) to the leptin condition (Figure 3, center arrow). Food intake returned to normal after withdrawal of leptin, but weight recovery was slower, allowing the animals to be weight-matched between the leptin and recovery (Figure 3, center and right arrows) conditions. As expected, leptin infusion in wild-type mice reduced food intake and weight less than in mutant mice.

View larger version (15K): [in this window] [in a new window]   Figure 3.   The effect of 3 d of leptin infusion (30 µg/d) on food intake and weight in eight C57BL/6J-Lepob mice and in eight wild-type mice. The upward arrows mark the days on which ventilatory control data were collected. Note: mutant mice were weight-matched on days in which ventilatory data were collected for the leptin condition and the recovery condition.

Leptin infusion caused pronounced increases in baseline ventilation (0% CO₂) compared with control and recovery conditions, as shown for one animal during NREM sleep (Figure 4). Pooled data demonstrated that leptin infusion significantly increased ventilation during wakefulness, NREM, and REM sleep across all levels of inspired CO₂ (Figure 5). No differences were observed in ventilation between control and recovery conditions (Figure 5), indicating that leptin is a potent, weight-independent, and reversible stimulus to ventilation in C57BL/6J-Lep^ob mice. The effect of leptin replacement on E was achieved by an increase in VT during wakefulness and NREM sleep (Table 1). In contrast, leptin replacement increased both VT and f during REM sleep (Table 1). Wild-type mice did not exhibit any change in ventilation in response to leptin infusion (Table 2).

View larger version (19K): [in this window] [in a new window]   Figure 4.   Stimulation of resting VT after leptin infusion (30 µg/d) in one C57BL/6J-Lepob mouse. Sample traces were obtained while breathing 0% CO2 in 40% O2 during NREM sleep in control, leptin, and recovery conditions. The VT, or volume change from trough to peak (inspiration), is larger during leptin infusion (373 µl) than during control (235 µl) or recovery (263 µl) conditions.

View larger version (26K): [in this window] [in a new window]   Figure 5.   Leptin infusion stimulates ventilation in eight C57BL/6J- Lepob mice (140 ± 7 d) at all levels of inspired CO2 and across all sleep/wake states compared with control and recovery conditions. Results presented show the mean ± SEM. Statistical significance between control, leptin, and recovery conditions was derived within each sleep/wake state and level of inspired CO2 using one-way, within-subject ANOVA with Newman-Keuls post hoc analysis. *p < 0.05, **p < 0.01, ***p < 0.001 compared with control and recovery conditions.

To determine the mechanism for the increase in baseline ventilation during leptin replacement, we examined the effect of leptin infusion on CO₂ production (VCO₂) during wakefulness in eight C57BL/6J-Lep^ob mice. The VCO₂ fell progressively over the course of the control, leptin, and recovery experiments (Table 3). This may reflect the animals' acclimation to the metabolic chamber with successive exposures. Nevertheless, these metabolic data indicate that the increase in baseline ventilation during wakefulness with leptin infusion, compared with control and recovery conditions (Figure 5), was not caused by an increase in VCO₂ production. Leptin infusion was accompanied by an increase in oxygen consumption (O₂) and a fall in the respiratory exchange ratio to 0.63 ± 0.02. 

A second metabolic issue relates to the effect of leptin on food intake, producing a state of acute starvation (Figure 3). We, therefore, conducted a food restriction experiment in which food intake was reduced to the mean level recorded in mutant mice during the 3 d of leptin infusion (Figure 3). In contrast to the stimulation of ventilation with leptin infusion (Figure 5), food restriction decreased ventilation across all sleep/wake states without affecting the HCVR (Table 4).

Leptin Replacement and Restoration of the HCVR

The effect of leptin infusion on the HCVR in mutant and wild-type mice is shown in Figure 6. In wild-type animals the HCVR under control conditions was higher across all sleep/ wake states than in mutant animals. Furthermore, leptin infusion in wild-type mice did not produce any change in the HCVR (Figure 6). In contrast, leptin infusion in mutant mice did cause an increase in the HCVR during sleep, but not during wakefulness (Figure 6). Thus, during NREM and REM sleep leptin infusion increased the HCVR in mutant animals compared with control and recovery conditions, and restored the HCVR compared with wild-type animals.

View larger version (22K): [in this window] [in a new window]   Figure 6.   Leptin infusion increases HCVR during NREM and REM sleep, but not during wakefulness, in eight C57BL/6J-Lepob mice (140 ± 7 d) compared with control and recovery conditions. There was no effect of leptin on HCVR in eight wild-type mice (131 ± 11 d) during any sleep/wake state. Results presented show mean ± SEM. Statistical significance between control, leptin, and recovery conditions was derived within each sleep/wake state using one-way, within-subject ANOVA with Newman-Keuls post hoc analysis. *p < 0.025, **p < 0.01 compared with control and recovery conditions.

Obesity and Elevated Endogenous Leptin Levels

We examined ventilation in wild-type mice fed a high-fat diet to induce obesity and elevate plasma leptin levels endogenously. Diet-induced obesity in wild-type mice increased weight (48.4 ± 1.2 g) compared with age-matched wild-type mice on a normal diet (32.9 ± 0.6 g). The high-fat diet in wild-type mice raised endogenous plasma leptin to concentrations comparable with the previous exogenous administration in both wild-type and mutant animals (Table 5). Baseline E was increased significantly across all sleep/wake states by the high-fat diet (Table 6). The HCVR decreased during wakefulness in mice on the high-fat diet, but was maintained at the level of the lean wild-type mice during NREM and REM sleep (Table 6). These data are consistent with the results from Figure 6 showing that in obese mutant mice replacing leptin produced an increase in HCVR during sleep, but not wakefulness. Therefore, in wild-type mice, obesity in the presence of raised endogenous leptin levels can protect against respiratory depression, particularly during sleep.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Considerable interest has surrounded the discovery of leptin as an obesity hormone and its key role in regulating food intake and weight. In the present study we report a novel role for leptin in stimulating ventilation to meet the mechanical demands imposed by obesity on the respiratory system. In experiments using the C57BL/6J-Lep^ob mouse, which is obese and mutated for the gene that encodes leptin, we have demonstrated that leptin deficiency in the presence of obesity is associated with an elevated Pa_CO2 and hypoventilation (Figures 1 and 2). By replacing leptin exogenously in these mutant mice we have established a causal relationship between leptin and respiratory control (Figure 5). The stimulating effect of leptin on ventilation was independent of weight (Figure 3), CO₂ production (Table 3), and food intake (Table 4), consistent with a direct effect of leptin on respiratory control centers in the brain. An increase in the HCVR (i.e., CO₂ chemosensitivity; Figure 6) during sleep also indicates that leptin acts centrally to stimulate ventilation. Finally, diet-induced obesity in wild-type mice, which raised endogenous leptin concentrations (Table 5), was associated with increased ventilation and maintained chemosensitivity during sleep (Table 6). These findings indicate an important role for leptin in supporting ventilation in obesity. Furthermore, the data suggest that a relative deficiency in central nervous system (CNS) leptin concentrations may be the factor responsible for producing hypoventilation in a clinical subgroup of severely obese humans.

The C57BL/6J-Lep^ob Mouse Model

The C57BL/6J mouse provides a powerful model to examine the relationship between obesity, sleep, and respiratory control. Similar to humans, wild-type mice have normal arterial blood gas levels, exhibit distinct decreases in chemosensitivity across sleep/wake states, and develop obesity on exposure to a high-fat diet. A mutation in the gene that encodes leptin in the C57BL/6J mouse induces profound obesity (Figure 3), produces abnormal blood gases (Figure 1), and blunts chemosensitivity (Figures 2, 5, and 6). This phenotype is characteristic of a minority of moderately to severely obese humans who develop alveolar hypoventilation (2). In the present study we use this model to demonstrate that the respiratory abnormalities of the mutant, obese C57BL/6J-Lep^ob mouse are leptin-mediated and independent of weight change.

Leptin Acts Neurally to Affect Ventilation

Having demonstrated that leptin stimulates respiration, we sought to determine its mechanism of action. Leptin is a hormone, secreted from adipose tissue that circulates to the brain and is known to affect neural regulatory processes controlled by the hypothalamus (12, 13). Leptin administration can decrease food intake and raise metabolic rate in C57BL/6J-Lep^ob mice (19), either of which could stimulate ventilation without altering chemosensitivity. Alternatively, leptin may act through brain respiratory centers to increase CO₂ chemosensitivity.

Although leptin replacement in C57BL/6J-Lep^ob mice can increase oxygen consumption, carbon dioxide production does not increase (Table 3). From a ventilatory perspective, CO₂ production is the important metabolic variable because alveolar ventilation (A) is dependent on VCO₂ according to the following equation:

A  VCO₂/Pa_CO2

Therefore, a rise in VCO₂ production with leptin replacement could account for our observed increases in baseline ventilation (assuming an increase in A produces a proportional increase in E). Our awake measurements of VCO₂, however, do not support a role for leptin stimulating ventilation metabolically. Although leptin infusion did increase oxygen consumption to a comparable degree to that previously reported (19), calculation of the respiratory quotient (Table 3) suggests that this may represent a switch to effectively pure fat metabolism. This seems likely given that by the third day of leptin replacement mutant mice were not consuming any food (Figure 3) and would, therefore, be dependent on body fat stores for metabolic purposes. Leptin replacement, therefore, reverses hypoventilation in C57BL/6J-Lep^ob mice by stimulating baseline ventilation independent of any increase in VCO₂.

Starvation can either cause increases or decreases in ventilation. If starvation is prolonged, a ketoacidosis can develop which will lower pH and can stimulate ventilation (20). On the other hand, acute starvation can decrease metabolism and thus depress ventilation (21). In our experiment, by the third day of leptin replacement in C57BL/6J-Lep^ob mice food intake had reached zero (Figure 3), suggesting a state of acute rather than chronic starvation. As expected, therefore, when we compared a group of obese mutant mice on a normal diet with that on a restricted diet (to match the food intake seen during leptin infusion), we saw a decrease of more than 30% in baseline ventilation across all sleep/wake states (Table 4). Thus, a period of acute starvation depressed rather than stimulated ventilation in C57BL/6J-Lep^ob mice. In summary, the above discussion argues that the stimulating effect of leptin on baseline ventilation is not secondary to either increased CO₂ production or metabolic changes associated with acute starvation. Whether leptin acts directly on respiratory control centers in the brain or acts through other pathways, such as modulating endogenous opioids (22, 23), remains to be determined.

Leptin and Sleep

Sleep/wake state is a known modulator of ventilatory control (7). Our data demonstrate that mice (Figure 6), like humans (7), decrease their HCVR across sleep/wake states. Thus, neural pathways controlling sleep interact with neural centers controlling ventilation. A third neural pathway involving leptin may also converge on those controlling respiration and sleep, because leptin replacement increased the HCVR during NREM and REM sleep, but not during wakefulness. Sleep, therefore, unmasked a leptin pathway which can stimulate ventilatory control mechanisms.

In the absence of leptin, there was no HCVR during REM sleep in obese mutant mice. This inability to increase ventilation during CO₂ challenge makes REM sleep a particularly vulnerable state for ventilatory control in obese mutant mice which lack leptin. However, obesity need not be associated with depressed ventilatory control during REM sleep. Consider that diet-induced obesity in wild-type mice did not depress the HCVR during REM sleep when accompanied by high endogenous leptin concentrations (Table 5). Similarly, leptin replacement in obese mutant mice produced a marked increase in the HCVR during REM sleep (Figure 6). Furthermore, leptin acted to increase both VT and f during REM sleep only (Table 1), suggesting a unique interaction between leptin and centers controlling sleep and respiration. Thus, the presence of high leptin concentrations can avert respiratory depression during REM sleep in obesity.

In a majority of severely obese humans, including those with alveolar hypoventilation, sleep can impact upon upper airway respiratory muscles as well as the diaphragm. During sleep a reduction in neural output to the upper airway is thought to be responsible for the development of obstructive sleep apnea (OSA) (24), which can exacerbate the hypoventilation that already occurs during sleep. In C57BL/6J-Lep^ob mice, however, we have no evidence for upper airway obstruction during sleep. This leads us to suggest that hypoventilation during sleep in this model is solely due to reductions in diaphragmatic activity. As such, the C57BL/6J-Lep^ob mouse has the distinct advantage of eliminating OSA as a confounding variable while determining the effects of obesity and leptin on ventilatory control. Whether or not leptin plays a role in controlling the human upper airway also remains to be determined.

Leptin Deficiency and Human Obesity Hypoventilation

The C57BL/6J-Lep^ob mouse model demonstrates that an inability to produce leptin causes an elevated Pa_CO2 and marked respiratory depression, particularly during sleep. However, such a model, in which obesity is associated with an absence of functional leptin, is extremely rare in humans (15, 16). Thus, a mutation in the ob gene cannot be considered the principal mechanism producing hypoventilation in obese humans.

In human populations, and in wild-type mice, obesity is associated with elevated plasma levels of leptin (10). We speculate that such elevations in leptin would be expected to stimulate ventilation, thereby averting respiratory failure. In fact, we observed just such an association between plasma leptin levels and stimulation of ventilation with diet-induced obesity in wild-type mice (Tables 5 and 6). Thus, elevated leptin levels may act as an important factor which compensates for obesity by maintaining ventilation appropriate to the degree of adiposity.

Leptin concentrations can decrease rapidly in response to dietary restrictions (25, 26). In obese humans, plasma leptin levels decreased more than 50% over a 6-wk period in response to a low-calorie diet (26). Assuming that leptin normally acts to stimulate ventilation, a rapid reduction in leptin concentrations with dietary restriction may depress ventilation. Indeed, two clinical studies demonstrated that severe dietary restriction in morbidly obese subjects, sufficient to induce significant decrements in circulating leptin levels, produced a depression in respiratory control (27, 28). Taken together, these studies suggest that an association may exist between decreased leptin concentrations and ventilatory depression in humans.

The present study indicates that leptin acts through central neural pathways to stimulate ventilatory control mechanisms. Thus, we would predict that it is the concentration of leptin in the cerebrospinal fluid (CSF) rather than the plasma that influences ventilatory control. Recent reports in humans suggest that leptin concentrations in the CSF are much lower and more variable than in the plasma (29, 30). In particular, Schwartz and coworkers (29) reported that, at a BMI of approximately 31 kg/m², the CSF leptin levels can vary by as much as fourfold between individuals. Such a variability in leptin concentrations in the brain may produce a corresponding variability in respiratory control. Thus, we speculate that the concentration of leptin in the CSF, or the sensitivity of leptin receptors in the brain, is the factor that determines whether respiration in an obese human is normal or depressed.