INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Intact upper airway (UA) function is vital for normal breathing. During the inspiratory cycle, UA dilator muscle activity occurs prior to activation of the diaphragm (1), unloading the ventilatory muscles by reducing the resistance of the UA, and thereby minimizing the work required to inflate the lungs. Serotonin (5-HT) and serotonergic neurons exert an excitatory effect on ventilatory and UA dilator motoneurons (2). The serotonergic system is especially involved in the maintenance of UA patency via activation of 5-HT_2A/2C receptor subtypes (5). Serotonergic neurons have an increased activity during wakefulness and a reduced activity during sleep (8). Serotonergic modulation of phrenic and hypoglossal motoneurons demonstrates a unique heterogeneity (9, 10). Indeed, in the English bulldog, a model of disordered breathing, the systemic administration of a 5-HT_2A/2C antagonist resulted in a selective suppression of UA dilator muscle activity, leading to a reduction in UA cross-sectional area (11).

UA resistance is often elevated in obesity. Excessive fat deposits in obese patients in areas surrounding the neck can compress the pharynx and alter its mechanical properties (12). In addition, obese subjects often have smaller lung volumes than nonobese subjects and this further contributes to UA narrowing (13). These anatomical changes in UA size with obesity may lead to long-term modifications of serotonergic control of UA patency.

The obese Zucker (Z) rat (Figure 1), a genetic model of obesity, exhibits a rapid shallow breathing pattern at rest, reduced lung volumes and chest wall compliance, and a blunted respiratory drive (14, 15). We have recently reported that obese Z rats have anatomically narrower UA compared with age-matched lean Z rats (16). Obese Z rats also display depressed serotonergic activity in the central nervous system, which has been linked to their overeating and development of obesity (17). Moreover, feeding-induced hypothalamic serotonin release is enhanced by obesity and reduced in aging (18). It is unknown, however, whether the altered serotonergic metabolism noted in the brain of obese Z rats also modulates ventilation and UA stability and whether it is affected by age.

View larger version (110K): [in this window] [in a new window]   Figure 1.   X-Ray photograph of whole body in an obese (left panel ) and a lean (right panel ) Zucker rat. The obese rat displays many fat deposits in the whole body compared with the lean rat. The excessive fat deposits in the neck can compress the pharynx.

The aim of this study, therefore, was to examine the hypothesis that a withdrawal of serotonergic excitatory drive would alter resting ventilation and ventilatory responses to chemical stimuli that may be differentially affected by obesity and age. In a first study, we measured the ventilation during room air breathing and during hypoxic and hypercapnic ventilatory challenges following systemic administration of ritanserin, a selective 5-HT_2A/2C receptor antagonist. Studies were performed in both young and older obese and lean Z rats. In addition, to clarify the mechanism responsible for the effect of 5-HT_2A/2C receptor antagonist on ventilation in older obese Z animals, a second study was performed to measure the effects of ritanserin on the mechanics of the isolated upper airway in anesthetized animals. Understanding the relative importance of active serotonergic modulation on upper airway patency may provide insight into the management of obesity-related respiratory control problems.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study I: Ventilation Experiments in Unanesthetized Rats

Animals. Two age groups of male Zucker rats were studied: (1) eight pairs of young animals at 6-8 wk of age (lean phenotype [Fa/?] n = 8, and obese phenotype [fa/fa], n = 8), and (2) eight pairs of older animals at 7-8 mo of age (lean, n = 8, and obese, n = 8). All rats were purchased from Vassar College, Poughkeepsie, NY. One lean and one obese animal were housed per cage. Ambient temperature was maintained at 26° C with a 12-h light/dark cycle. All animals were provided with standard laboratory chow (Purina) and water ad libitum. The experimental protocols were approved by the Institutional Animal Care and Use Committee of the State University of New York at Buffalo.

Techniques and measurements. Ventilation was measured by the barometric technique of plethysmography. A cylindrical Plexiglas chamber with a volume of 4 L was used for the measurements of ventilation. The rat was placed in the chamber within a restrainer, which did not permit backward rotation. The animal chamber had an inlet tube that was connected to pressurized air tanks (BOC gases). Inlet flow was regulated at 2 L/min by a flowmeter (Dwyer Instruments Inc., Michigan City, IN). The concentrations of inflowing or outflowing O₂ and CO₂ were monitored by an Ametek S-3A/I O₂ analyzer and an Ametek CD-3A CO₂ analyzer. The calibrations of the gas analyzers were checked once daily, using certified calibration gases (BOC gases). To measure ventilation, the chamber was completely sealed after momentarily interrupting the flow through it, and the oscillations in pressure caused by breathing were recorded by a sensitive pressure transducer (Statham Laboratories, Oxnard, CA; model 236). The signal was received, amplified (Grass Instrument Co., Quincy, MA) and displayed on an oscillographic strip chart recorder (Grass Polygraph). An average of 100 breaths was recorded on paper at a speed of 10 mm/s. Injection and withdrawal of 0.3 ml of air with a 1-ml syringe were performed several times during the recording, for calibration purposes. Body temperature was measured continuously by a thermometer probe (Yellow Springs Instruments, Yellow Springs, OH) placed at least 5 cm into the rectum. Chamber temperature and relative humidity were monitored by a flow-through probe (Fisher Scientific, Pittsburgh, PA) mounted within the chamber. The temperature inside the chamber was controlled at 26-27° C by heating or cooling. Barometric pressure values to the nearest hour were obtained from the Internet posting of the U.S. Weather Bureau located at the Buffalo International Airport.

Respiratory frequency (f) and inspiratory and expiratory time (TI and TE) were calculated directly from the ventilation-induced pressure swings. Tidal volume (VT) was obtained as a function of the pressure change inside the chamber. Pulmonary ventilation (E) was calculated (E = VT × f) and expressed in ml/body weight (kg)/min. Oxygen consumption (O₂) was calculated from the inflow-outflow O₂ differences multiplied by the gas flow, neglecting the small error introduced by a respiratory quotient less than unity. All O₂ values are presented at standard temperature and pressure dry (STPD), and expressed per unit of effective body mass (EBM) because lean and obese rats of the same size have different body compositions (19). EBM for lean and obese Z rats was calculated as 1.00 × M^0.75 and 0.86 × M^0.75, respectively, where M is the body weight (kg) of the animal.

Experimental protocol. Each animal was tested repeatedly at 3-d intervals following an intraperitoneal injection of equal volume of either vehicle (dimethyl sulfoxide, DMSO) or 1 mg/kg of ritanserin (Research Biochemical International, MA), a potent selective 5-HT_2A/2C receptor antagonist that can cross the blood-brain barrier. The dose of ritanserin was based on previously published studies examining the effects of this drug on brain tissue levels of biogenic amines and its effects on different types of behavior in the rat such as fear, food intake, and conditioned response (20). The solutions were prepared daily (ritanserin concentration = 1 mg/ml) and placed in vials labeled as solutions A or B. To reduce possible effects of adaptation of animals to repeated measurements, the agents were given in a randomized design on Days 1 and 4. The investigator involved in testing remained blinded to the contents of the vials until the whole study was completed and analyzed. To reduce the stress level during the study, all animals were habituated to an intraperitoneal injection of 0.4 ml of saline and the restraining device within the barometric chamber for 60 min on two successive days prior to the first experimental study. To minimize any potential differences due to circadian rhythms, experiments were begun in each animal at precisely the same time between 8:00 A.M. and 3:00 P.M.

After the administration of an agent, the rat was placed into the chamber and it breathed room air for 55 min followed by 10 min of hypoxia (10% O₂, balance N₂). The gas mixture in the chamber was then replaced with room air for 15 min followed by 10 min of hypercapnia (4% CO₂, 21% O₂, balance N₂). Ventilation and metabolic rate were measured during the last minute during room air, hypoxia, or hypercapnia breathing.

Study II: Isolated Upper Airway Experiments in Anesthetized Rats

Animals. The study was performed in male lean (n = 7) and obese (n = 7) Zucker rats at 8-9 mo of age. They were different animals from those used in Study I.

Techniques and measurements. An isolated upper airway preparation was utilized as previously described (24, 25). In this preparation, the collapsibility of the pharyngeal airway has been found to be determined by the maximal inspiratory airflow (I_max) and the pharyngeal critical pressure (Pcrit) at the flow-limiting site (FLS).

Animals were anesthetized with an intraperitoneal injection of ketamine (50 mg/kg) and xylazine (5 mg/kg). Animals were placed supine on a heating pad and rectal temperature was monitored and maintained at 37° C. Atropine sulfate (0.3 mg/kg) was subcutaneously injected to reduce airway secretions. A femoral vein was cannulated for the administration of drugs.

To prepare the isolated upper airway, the cervical trachea was cannulated with an endotracheal tube, and the animal was mechanically ventilated (Model 683; Harvard Apparatus, South Natick, MA). End-tidal CO₂ was monitored at the endotracheal tube by a CO₂ analyzer (Capstar-100; CWE, Inc., Ardmore, PA) and maintained at 4.0-4.5%. A rigid cannula was inserted into the proximal tracheal stub, through the glottic structures, and fixed in place at the level of the aryepiglottic folds. The position of this cannula was confirmed at the end of the experiments. The head was fixed in place at an angle of 10-20° from the horizontal plane.

A mobile catheter (PE 50 tubing, 0.58 mm i.d.) with a side hole midway along its length was inserted through the tracheal cannula into the pharynx and out one nostril. This catheter was used to measure the pharyngeal pressure (Pph) at different locations in the airway. In four animals (two lean and two obese) another catheter was introduced through the tracheal cannula to the lower end of the upper airway and maintained in this position to record the hypopharyngeal pressure (Php). Pph and Php were monitored with pressure transducers (P23XL; Statham Laboratories). Secretions inside the catheter were removed by inserting a small guide wire into the catheter as needed. Pharyngeal secretions were aspirated with a small suction tube connected to the tracheal cannula. The inspiratory airflow (I) through the isolated UA was measured using a pneumotachograph (Fleisch, Switzerland; No. 0.771) and a differential pressure transducer (Validyne, Northridge, CA; MP 45-1; ± 2 cm H₂O) placed in series between the tracheal cannula and a negative pressure source. The signals of Pph, Php, and I were sent out to a data-acquisition system (WinDaq DI-720; DATAQ Instruments, Akron, OH) for real-time recordings and later analysis.

I_max and Pcrit were measured as previously described (24, 25). We could determine the FLS by monitoring the shapes of I versus time and Pph versus time curves. In brief, pressure at the downstream end of the UA was rapidly lowered to 60 cm H₂O by a negative pressure source. As Pph was lowered, I rose and reached a maximum (I_max) at the onset of the inspiratory airflow limitation followed by a decrease. To localize the FLS, the catheter was slowly advanced cranially through several sites in the pharynx while monitoring side-hole pressure and I. After localization of the FLS, the mobile catheter was fixed at or immediately upstream from this site to measure Pph. We marked the catheter for later confirmation of the side-hole position. Pcrit was defined as the nadir in Pph at the onset of flow limitation as shown in Figure 4. Resistance upstream to the FLS was defined as oronasal resistance (Ron). Ron was calculated as follows: Ron = (Pon  Pcrit)/I_max, where oronasal pressure (Pon) remained atmospheric.

View larger version (14K): [in this window] [in a new window]   Figure 4.   Representative inspiratory airflow ( I), pharyngeal pressure (Pph), and hypopharyngeal pressure (Php) in isolated upper airways from a lean (A) and an obese (B) Zucker rat. Pph was measured at or immediately upstream to flow-limiting site. As Pph was lowered, I rose and reached a maximum ( Imax) at the onset of inspiratory airflow limitation followed by a decrease. Pcrit was defined as the nadir in Pph versus time curve. Imax of obese rat was smaller than that of the lean rat and Pcrit of the obese rat was greater than that of the lean rat. Imax decreased and Pcrit increased after the administration of ritanserin in each example.

For each animal, five measurements were made to obtain I_max and Pcrit 15 min after an intravenous administration of vehicle (DMSO) as a control, and again 15 min after an intravenous administration of 1 mg/kg of ritanserin. To determine possible effects of repeated measurements on the mechanics of the isolated upper airway, a preliminary study was performed. I_max and Pcrit were measured five times 15 min after an injection of DMSO and again 15 min after another injection of DMSO, and it was found that the values were highly reproducible. Average values following the administration of vehicle or ritanserin were calculated from the repeated five measurement values.

Multistranded Teflon-coated stainless steel wires (A5632; Cooner Wire Company, Chatsworth, CA) were implanted into the genioglossal (GG) and parasternal (PS) muscles in two of the lean and two of the obese animals. The GG and PS EMG signals were amplified, band-pass filtered from 30 to 2,000 Hz, processed by a Paynter filter with a time constant of 100 ms to obtain a moving average (ASC909 Amplifier, A.C. Pre-amplifier; Grass), and then recorded with a strip chart recorder (MT9500, Astro-Med, Inc.). Baseline and postritanserin EMGs were measured prior to the measurement of upper airway mechanics.

Statistical Analysis

The differences between data obtained from lean and obese Z rats or between the responses following the administration of vehicle and ritanserin within a group were analyzed by two-way analysis of variance (ANOVA) with repeated measurements (Factor A: lean versus obese; Factor B: vehicle versus ritanserin). When significance was indicated, a post-hoc t test with Bonferroni's correction for multiple comparisons was used. In all cases, a p value < 0.05 was considered significant. All data presented in the text, tables, and figures represent means ± SD.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study I

Comparisons of ventilation between lean versus obese Z rats. Young rats. Body weights of the young obese animals were 42% greater than those of age-matched lean animals (312 ± 34 g versus 219 ± 35 g, p < 0.01) (Figure 2). During room air, control values of E in the obese rats were lower than in lean rats (672 ± 68 ml/kg/min versus 817 ± 121 ml/kg/min, p < 0.05) (Table 1 and Figure 3). The low E in obese rats was due to reduced VT (Table 1). There were no significant differences in f, respiratory timing, and metabolic variables measured in room air between the two groups (Table 1 and Figure 3). Ventilatory responses to hypoxia and hypercapnia in young obese rats were significantly blunted compared with those of lean rats (Table 2).

View larger version (21K): [in this window] [in a new window]   Figure 2.   Body weights of young or older lean (L) and obese (O) Zucker rats used in Study I, and older lean and obese Zucker rats used in Study II.

View larger version (40K): [in this window] [in a new window]   Figure 3.   The effects of ritanserin on ventilation ( E), oxygen consumption ( O2), and oxygen ventilatory equivalent ( E/ O2) during room air in young and older Zucker rats. Ritanserin (R) significantly reduced E, increased O2, and decreased E/ O2 only in older obese rats compared with vehicle control (C). *p < 0.05, **p < 0.01 significantly different from obese control rats;  p < 0.05,  p < 0.01 significantly different from lean control rats.

Older rats. Older obese rats were 48% heavier than older lean rats (704 ± 43 g versus 477 ± 15 g, p < 0.01) (Figure 2). Older obese rats breathed with a higher f and a lower VT than lean rats during room air (Table 1), resulting in a lower E (439 ± 35 ml/kg/min versus 556 ± 49 ml/kg/min, p < 0.01) (Figure 3). Inspiratory time (TI), expiratory time (TE), and metabolic variables did not differ between older lean and obese animals during room air (Table 1 and Figure 3). Obese rats exhibited blunted ventilatory responses to hypercapnia compared with lean rats, and the lower E responses to hypoxia in obese rats did not achieve significance (Table 2).

Effects of 5-HT_2A/2C antagonist on ventilation. Ritanserin, relative to vehicle control, had no effect on body temperature in both young and older Z rats (Table 1). Although ritanserin did not affect E in young (lean and obese) Z rats or older lean rats, it significantly reduced E in older obese rats compared with control (439 ± 35 to 386 ± 41 ml/kg/min, p < 0.01) (Figure 3). This decrease in E was primarily due to a reduction in f (160 ± 12 to 145 ± 21 breaths/min, p < 0.05) (Table 1). Ritanserin induced a significant prolongation of TI in older obese rats without any change in TI/TTOT and VT/TI (Table 1). O₂ significantly increased after the administration of ritanserin only in older obese rats (16.4 ± 1.7 to 18.2 ± 1.9 ml/kg^0.75/min, p < 0.05) (Figure 3). As a consequence, the ventilatory equivalent for oxygen (E/O₂) was markedly reduced after ritanserin administration (28.7 ± 2.6 to 23.1 ± 2.6, p < 0.01) (Figure 3).

Ritanserin had no effects on ventilatory responses to either hypoxia or to hypercapnia in young or older lean and obese Z rats (Table 2).

Study II

Effects of 5-HT_2A/2C antagonist on upper airway collapsibility. Body weights of obese rats used in this study were 55% greater than those of age-matched lean animals (660 ± 73 g versus 424 ± 39 g, p < 0.01) (Figure 2).

These animals were significantly lighter than the older animals used in Study I (obese, p < 0.05; lean, p < 0.01, respectively). In Figure 4, representative I, Pph, and Php recordings obtained in isolated upper airways of one lean (Figure 4A) and one obese (Figure 4B) Z rat are illustrated. In this example, I_max of the lean rat was greater than that of the obese rat and Pcrit of the lean rat was more negative than that of the obese rat. I_max decreased and Pcrit increased after the administration of ritanserin in each example.

In the vehicle control condition, the Pcrit of obese rats was greater than that of lean rats (1.6 ± 0.3 versus 3.1 ± 1.0 cm H₂O, p < 0.05) and I_max of obese rats was lower than that of lean rats (7.2 ± 1.7 versus 12.9 ± 4.2 ml/s, p < 0.05), whereas the Ron of obese rats did not differ from lean rats (221 ± 28 versus 249 ± 71 cm H₂O/l/s) (Figure 5). In seven out of seven obese rats, the administration of ritanserin increased Pcrit (1.6 ± 0.3 cm H₂O to 0.8 ± 0.2 cm H₂O, p < 0.01) and decreased I_max (7.2 ± 1.7 to 4.1 ± 1.5 ml/s, p < 0.01), whereas Ron was unaffected. In lean rats, ritanserin also exhibited a significant increase in Pcrit (3.1 ± 1.0 to 2.4 ± 0.6 cm H₂O, p < 0.05) and a significant decrease in I_max (12.9 ± 4.2 to 10.7 ± 3.0 ml/s, p < 0.05) (Figure 5).

View larger version (11K): [in this window] [in a new window]   Figure 5.   Effects of ritanserin on isolated upper airway mechanics in individual obese (closed circles) and lean (open circles) Zucker rats. If the circle falls above or below the linear line of identity, that animal's index is different between ritanserin and control. Pcrit (upper panel ) of obese rats was less negative than that of lean rats (p < 0.05). Ritanserin significantly increased Pcrit in seven out of seven obese rats (p < 0.01) and in five out of seven lean rats (p < 0.05). Imax (middle panel ) of obese rats was lower than that of lean rats (p < 0.05). Ritanserin decreased Imax in obese rats (p < 0.01) and in lean rats (p < 0.05). There was no difference in Ron (lower panel ) between lean and obese rats. Ritanserin did not affect Ron in both lean and obese rats.

Effects of 5-HT_2A/2C antagonist on upper airway EMG. Raw EMG signals from the genioglossal (EMG_GG) and parasternal (EMG_PS) muscle pre- and postritanserin administration in representative lean and obese Z rat are shown in Figure 6. The administration of ritanserin induced a pronounced reduction in EMG_GG activity with little change in EMG_PS activity in both lean and obese Z rats. EMG_GG and EMG_PS in another pair of lean and obese Z rat demonstrated similar changes induced by ritanserin.

View larger version (98K): [in this window] [in a new window]   Figure 6.   Representative examples of the raw electromyogram from the genioglossal muscle (EMGGG) and the parasternal muscle (EMGPS) in a lean (upper) and an obese (lower) Zucker rat. Intravenous administration of 1 mg/kg of ritanserin reduced EMGGG activity, but not EMGPS activity.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The major findings of this study are as follows: (1) systemic administration of ritanserin resulted in a reduction of E and a rise in O₂ only in the older obese Z rats, (2) ritanserin did not affect the ventilatory responses to either hypoxia or to hypercapnia in young or older lean and obese Z rats, (3) the UA of older obese Z rats was more collapsible than that of lean rats, and (4) ritanserin increased UA collapsibility in obese and lean Z rats.

Effects of Ritanserin on Ventilation

The activation of brain stem 5-HT_2A/2C receptors has been shown to increase phrenic and hypoglossal motoneuron excitability in vitro in neonatal rats (2, 3) and in anesthetized adult cats whose phrenic and vagus nerves had been cut (4). Thus, we predicted that pretreatment with a 5-HT_2A/2C receptor antagonist would reduce phrenic motoneuron excitability, thereby reducing resting ventilation and the ventilatory response to chemical stimuli. In the present study, however, awake young and older lean Z rats demonstrated that ritanserin had no effects on resting E and O₂, as well as no effects on the ventilatory responses to hypoxia or hypercapnia. It is possible that the concentration of ritanserin used was insufficient to induce an adequate inhibition of the appropriate receptor population. However, the dose of ritanserin used in our experiments (1 mg/kg) was similar to that recommended to obtain an optimal blockade of central 5-HT receptors (20- 23). In addition, our results are consistent with the recent report of Herman and coworkers (26) who showed that the intravenous administration of ketanserin, another selective 5-HT_2A/2C receptor antagonist, did not change ventilation during room air breathing or during a hypoxic ventilatory challenge in awake goats. Thus, in awake Z rats, we suggest that an inhibition of 5-HT_2A/2C receptors by ritanserin may have little effect on phrenic motoneuron excitability, which is mainly associated with ventilation and work of breathing. In support of this, in the awake English bulldog, a natural model of sleep-disordered breathing, Veasey and coworkers reported that systemic administration of ritanserin caused only a small reduction in diaphragm electromyographic (EMG) activity whereas it induced a great suppression of UA dilator muscle EMG activity, although they did not measure ventilation (11).

In contrast to the results of young Z and old lean Z rats, older obese Z rats exhibited a reduction in E and a rise in O₂ following ritanserin administration (Figure 3). The older obese Z rats were 48% heavier and adopted a more rapid shallow breathing pattern at rest compared with age-matched lean rats. Therefore, the observations of the older obese Z rats suggest that the underlying mechanism may be related to long-term effects of increasing obesity. In older obese Z rats, the reduction of ventilation with ritanserin was accompanied by a significant prolongation in TI, a reduced f, and an unchanged VT. These findings are similar to the responses elicited by external resistive loading added to the respiratory tract. In animals with intact vagi, acute ventilatory resistive loading during inspiration induces a reduction in VT and an increase in TI followed by a decrease in f to maintain TI/TTOT through a vagal feedback mechanism (27). If the resistive loading is sustained, VT tends to return to control levels (28). Furthermore, it has been found that external resistive loading induces a rise in oxygen consumption secondary to increased work of breathing (29). Indeed, we observed that O₂ rose significantly, and the ventilatory equivalent for oxygen (E/O₂) was markedly reduced following the administration of ritanserin only in older obese animals. Thus, the effects of ritanserin on ventilation in older obese Z rats are responses consistent with sustained inspiratory resistive loading. In support of this, ritanserin elicited a significant reduction of UA cross-sectional areas in the English bulldog, leading to inspiratory collapse of the airway (11). The resistance of UA is known to increase during sleep and inadequate resistive load compensation contributes to a decline in ventilation during sleep (30). The mechanism underlying the fall of ventilation with ritanserin in older obese rats may be somewhat similar to those of the impaired load compensation during sleep.

UA cross-sectional areas and diameters of older obese rats are narrower compared with age-matched lean counterparts (16). In addition, upper airway inspiratory resistance is inversely correlated with lung volume (13). As the lung volume in obese Z rats is smaller than that of lean animals (14), apparent upper airway resistance during spontaneous breathing must be higher in obese rats than in lean animals. We speculated that the effects of ritanserin on ventilation might be related to increased UA load due to changes in either UA collapsibility or UA resistance. To confirm this hypothesis, we measured the mechanics of the isolated upper airway in anesthetized animals.

Critique of Methodology of Isolated Upper Airway

Prior to discussing our results, several limitations of the isolated UA preparation should be addressed. First, we generated a large negative suction pressure (60 cm H₂O) to achieve airflow limitation. The UA is not actually exposed to such large negative pressure in vivo because airway pressure at this portion changes in the physiological range. It is possible that these large pressures traumatized the structure of the UA, resulting in the changes in repeated measured values of Pcrit and I_max as pointed out by Rowley and coworkers (25). However, repeated measurements showed good reproducibility. Second, anesthesia can suppress motoneuronal activity to UA dilator muscles, leading to a bias in the measurement of UA mechanics. In the present study, end-tidal CO₂ was maintained constant during the experimental period. Moreover, the anesthetic levels could not account for the effects observed following ritanserin administration because the depth of anesthetic remained virtually constant between Pcrit measured pre- and postritanserin infusion. Third, the flow mechanics of the UA are dependent on the angle of the head and neck, and whether studies are conducted in the supine or prone position (31). The supine rat UA is relatively rectilinear along a rostral to caudal axis compared with that of the prone position. In our preparation, the rats were studied in the supine position with 10-20° neck flexion. Small changes in neck flexion may alter the individual values of Pcrit and I_max, but these were constant pre- and postritanserin injection. Fourth, the position of the tongue exerts differing influences on airflow dynamics of isolated UA (25). Because we did not fix the tongue, prolapse of the tongue into the pharynx may have altered UA airflow mechanics. Finally, inspiratory UA muscle activity is different between nasal breathing and oronasal breathing. Because we did not seal the lips in the present preparation, whether airflow passes through the nose or mouth may affect the values of Pcrit and I_max. However, it has been reported that baseline values of Pcrit and I_max were not different between mouth open and mouth sealed in the isolated rat UA preparation (32).

Effects of Ritanserin on Upper Airway Mechanics

In the present study, Pcrit in obese rats was significantly higher (less negative) than that of lean rats, indicating that the UA of obese rats is more collapsible than that of lean rats. This finding is consistent with a previous report (33) showing Pcrit in humans falling to more negative values after weight loss. The decrease in pharyngeal collapsibility after weight loss has been attributed to reduction in fat deposits in the pharynx and improvement in the activity of UA muscles. In the present study, older obese Z rats weighed much more than lean rats. Excessive fat deposits in the neck around the UA (Figure 1) can compress the pharynx and alter its dimension and mechanical properties (12). Fat in the abdomen can also elevate the diaphragm resulting in decreased tracheal tug on the upper airway (34). In spite of the marked differences in Pcrit and I_max between obese and lean Z rats, the values of Ron in obese Z rats were similar to those of lean rats (Figure 5), suggesting that basic structure of the segment upstream to the FLS is not different between the two rat groups. In other words, the higher Pcrit in obese rats may be attributed to a more collapsible structure of the pharynx compared with lean rats.

In the model of the upper airway as a Starling resistor system, it has been demonstrated that I_max is modulated by Pcrit and the resistance upstream to the FLS (24). In obese Z rats, decreases in I_max following the administration of ritanserin were associated with inverse rises in Pcrit without alteration in Ron (Figure 4 and 5). These results indicate that ritanserin increases UA collapsibility and acts at the FLS in the pharynx, but not at the segment upstream to the FLS. In a previous study, Pcrit has been measured to be near zero or positive in patients with obstructive sleep apnea, whereas Pcrit exhibits more negative values in normal subjects (35). The values of Pcrit both before and after ritanserin administration in lean rats were more negative than those of obese rats, indicating that their airways tended to stay open. In contrast, the values of Pcrit in obese rats increased to near zero after ritanserin, implying that their airways were more susceptible to collapse.

Considering the results of the isolated UA experiments, the effects of ritanserin on ventilation and oxygen consumption in awake older obese Z rats can be explained by an increase in pharyngeal collapsibility (positive shift of Pcrit). The increase in pharyngeal collapsibility most probably is a result of reduction in UA dilator muscle activity given the observed decreases in EMG_GG without any changes in EMG_PS. As the in vitro contractile properties of the pharyngeal muscle do not differ between lean and obese rats (36), the increase in UA collapsibility by ritanserin in obese rats suggests a reduction in motoneuronal activity projecting to pharyngeal muscles. Since ritanserin was administered systemically, we cannot determine what specific neuronal pools were affected leading to the increase in UA collapsibility. It may be due to the effects of ritanserin on hypoglossal motoneurons (5), but other neuronal pools may also have been affected and additional studies will be required.

It has been proposed that in an anatomically predisposed UA, patency requires much more use of UA dilator muscles. Hendricks and coworkers (37) demonstrated that activity of UA dilator muscles was augmented during wakefulness and non-rapid eye movement (NREM) sleep, whereas it dropped episodically during REM sleep in English bulldogs, suggesting that compensatory pharyngeal dilator hyperactivity is necessary to maintain airway patency and normal breathing. Why similar dosages of ritanserin reduced ventilation only in older obese Z rats, but not in young Z and older lean Z rats, might, in part, be due to a greater reliance on serotonergic modulation in UA dilator muscle activity superimposed upon an anatomically small UA with obesity. We, therefore, speculate that older obese Z rats may have augmented serotonergic activity on UA dilator muscles as a mechanism to prevent pharyngeal collapse.

In conclusion, ritanserin did not affect resting ventilation in young Z and older lean Z rats, but it decreased ventilation with a rise in oxygen consumption in older obese Z rats. This effect of ritanserin on ventilation was attributed to an increase in inspiratory loading due to elevated UA collapsibility. We suggest that the serotonergic system plays an important role in the active maintenance of UA stability and normal breathing in obesity. In addition, we suggest that changes in passive characteristics (UA size) with obesity may lead to long-term modifications of active serotonergic modulation of UA patency. Further research is needed to determine whether augmented UA dilator muscle recruitment via active serotonergic control is observed during wakefulness as a means of compensating for anatomically compromised UA in obesity.