INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

IT IS INCREASINGLY RECOGNIZED that respiratory behavior exhibits a considerable degree of plasticity. Acute repetitive challenges with 3-5 min of hypoxia stimulate breathing, and the respiratory stimulation persists as long as an hour after termination of the hypoxic stimulus (5, 15, 16). The persistent elevation of respiration after episodic hypoxia (EH) has been termed as "long-term facilitation" (LTF) of breathing (19). LTF in the respiratory motor output (efferent phrenic nerve activity) was initially described in anesthetized, paralyzed, vagotomized cats (4, 12-14). Subsequent studies have demonstrated LTF of breathing in several other species including rats (6, 8, 16), mice (9), and goats (11, 23) under anesthetized as well as awake conditions. On the other hand, cumulative duration of sustained hypoxia (SH) does not elicit LTF (2), suggesting that it is unique to EH. Thus LTF is an extensively studied example of plasticity of respiratory motor behavior. It has been suggested that LTF of breathing is of considerable significance in obstructive sleep apneas wherein persistent stimulation of respiration prevents collapse of upper airways (see Ref. 19). EH associated with recurrent apneas, however, lasts only a few seconds (10-30 s; see Ref. 20). On the other hand, much of the information on LTF of breathing is based on studies wherein hypoxic exposures were as long as 5 min, which are substantially longer than that encountered during apneas. Therefore, one objective of the present study was to investigate whether brief EH lasting no more than 15 s also elicits LTF of respiration.

Recent studies reported increased generation of reactive O₂ species (ROS) from neutrophils in patients with recurrent apneas, an effect that was reversed after treatment with continuous positive airway pressure (3, 22). Our laboratory's recent studies on experimental models of EH indicated that ROS, especially O· radicals, play a critical role in cellular responses to EH in cell cultures (1) as well as in the altered carotid body function in intact animals exposed to chronic intermittent hypoxia (CIH; Ref. 18). On the basis of these studies, it has been proposed that CIH represents a form of oxidative stress (21). Recently, Ling et al. (10) reported that prior exposure to 7 days of CIH (5 min hypoxia-5 min normoxia for 12 h/night) enhances LTF of breathing evoked by acute EH. Therefore, the second objective of the present study was to determine whether conditioning with CIH with brief EH (15-s hypoxia) augments LTF of breathing, and if so whether it involves ROS-dependent mechanisms.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

All experiments were performed on male Sprague-Dawley rats weighing 250-350 g. The Institutional Animal Care and Use Committee of the Case Western Reserve University approved the experimental protocols.

General preparation of the animals. Acute experiments were performed on rats anesthetized with urethane (1.2 g/kg ip) and supplemented hourly with 15% of the initial dose. After tracheal intubation, the femoral artery and vein were cannulated for measuring arterial blood pressure (model P122, Grass) and for intravenous administration of fluids and drugs, respectively. Animals were paralyzed with pancuronium bromide (2.5 mg · kg¹ · h¹ iv) to prevent spontaneous breathing, and ventilated with a respirator (Harvard Apparatus). Bilateral vagotomy was performed in the midcervical region. Arterial blood samples were collected from femoral arterial catheter for determining blood gases [arterial PO₂ and arterial PCO₂ (Pa_CO2)] and pH (ABL-500, Radiometer, Copenhagen, Denmark). The rectal temperature of the animals was maintained at 37.5 ± 1°C by means of a heating pad. At the end of the experiments, animals were killed with intravenous administration of euthanasia solution.

Measurements of phrenic nerve activity. Efferent phrenic nerve activity was monitored as an index of respiratory motor output. The phrenic nerve was isolated unilaterally at the level of cervical 3 and 4 spinal segments. The nerve was cut distally, desheathed, and placed on bipolar silver electrodes. Action potentials were filtered (band pass 100-3,000 Hz), amplified (model P511K, Grass), and passed through Paynter filters (time constant of 100 ms, CWE, Ardmore, PA) to obtain a moving average signal. The integrated signal, as well as the raw action potentials, along with the arterial blood pressure signal, was recorded on a strip-chart recorder (Dash 10, Astro-Med, West Warwick, RI).

Exposure to CIH. Animals housed in feeding cages were placed in a special chamber (0.62 × 0.55 × 0.29 m³) for exposure to EH. The animals were unrestrained, freely mobile, and fed ad libitum. The chamber was flushed with alternating cycles of pure nitrogen followed by room air so that inspired O₂ levels reached 5% during hypoxia within 68-75 s and 21% during normoxia within 70-85 s. Ambient O₂ levels in the chamber were continuously monitored by use of a Beckman O₂ analyzer (model OM-11) by sampling the air in the chamber. Continuous vacuum was created within the chamber to balance the pressure between in and out flow of the gases. Inspired CO₂ levels were 0.2-0.5% and were monitored continuously by a medical gas analyzer (Beckman LB-2). The duration of the gas flows during each hypoxic and normoxic episode was regulated by timed solenoid valves. The paradigm of EH consisted of 15 s of hypoxia followed by 5 min of normoxia (9 episodes/h, 8 h/day). Animals were exposed to EH between 9:00 AM and 5:00 PM (8 h) for 10 consecutive days. Acute experiments were performed in the morning after the 10th day of the EH exposure, which was ~15 h after termination of CIH exposure.

Exposure to comparable, cumulative duration of SH. To examine the effects of the cumulative, comparable duration of SH, rats were exposed to isobaric hypoxia (12 or 5% O₂, see RESULTS) for 4 h. This was accomplished by placing animals in a Lucite chamber containing an inlet port for administering 12% or 5% O₂ and an outlet port connected to a vacuum sufficient to create a flow of 600 ml/min through the chamber, as measured by a rate meter. Acute experiments were performed 4 h after termination of SH.

Experimental protocols. In group 1 (n = 9), the effects of acute EH on phrenic nerve activity were examined in control rats. The protocols were as follows. In six animals, Pa_CO2 was maintained close to 35 Torr. In three animals, the CO₂ apneic threshold was determined as described (10). Briefly, the rate of the respiratory pump increased until the cessation of rhythmic phrenic nerve activity and then slowed till the appearance of rhythmic activity. Femoral arterial blood samples were collected immediately after the reappearance of rhythmic phrenic activity. Subsequently, baseline Pa_CO2 was set 2-3 Torr above the apneic threshold in each animal. Baseline phrenic activity along with arterial blood pressure was recorded for 10 min while the rats were ventilated with bleeding hyperoxic gas mixture (100% O₂) via a needle placed in the inspiratory port of the respirator. Ventilating rats with O₂-enriched room air maintained arterial blood pressure and arterial blood gases fairly stable over long periods of experimentation lasting >2-3 h. Animals were challenged with 10 episodes of EH (15 s of 12% O₂ balanced N₂ followed by 5 min of hyperoxia). Phrenic nerve activity was recorded during 10 episodes of EH and for 60 min after termination of EH (i.e., posthypoxic period). Femoral arterial blood samples were collected before and 60 min after the last hypoxic challenge. If there were variations in Pa_CO2 of >3 Torr and blood pressure fell >15-20 mmHg compared with baseline values, these experiments were excluded from analysis. The protocols described in group 1 were repeated in animals conditioned with 10 days of either CIH (group 2; n = 9) or SH (group 3; n = 8). Animals in group 4 (n = 9) received manganese (III) tetrakis (1-methyl-4-pyridyl) porphyrin pentachloride (MnTMPyP; Alexis), a potent scavenger of O· anions, via intraperitoneal injection (5 mg · kg¹ · day¹ for 10 days) and then were subjected to CIH for 8 h/day. Subsequently, the protocols described above were repeated in anesthetized animals.

Data analysis. The following variables were analyzed: 1) phrenic bursts per minute [respiratory rate (f)]; 2) amplitude of the Phr [in arbitrary units (AU)]; 3) minute neural respiration (MNR, f × Phr, AU/min); and 4) mean arterial blood pressure (mmHg), arterial PO₂, Pa_CO2, and pH. Variables in phrenic activity and arterial blood pressure were averaged over a 1-min period at the 10th, 5th, and 1st minute before EH exposure (i.e., baseline). Hypoxic responses were quantified over 15 s during each episode and expressed as percentage of baseline values (averaged over 1 min before each hypoxic episode). During the posthypoxic period, data were collected every 5 min for a 60-min period. The phrenic activity data are normalized as follows. Changes in phrenic nerve activity are expressed as percentage of baseline values (prehypoxia = 100%). In addition, in the experiments wherein apneic thresholds were measured, at the end of the EH protocols, phrenic nerve response to hypercapnia (Pa_CO2 = 70-80 Torr) were recorded. The data were analyzed as percent of hypercapnic response as well as percent of baseline values. The magnitude of responses by both ways of normalization was found to be essentially the same. Hence the data were pooled and expressed as means ± SE of percent of baseline values. Statistical analysis of the changes in phrenic nerve activity, blood pressure, and blood gases was performed by two-way ANOVA with repeated measures followed by Tukey's test. P values <0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Repetitive brief hypoxia elicits LTF in phrenic nerve activity. Phrenic nerve activity (MNR, AU/min) increased in response to each episode of hypoxia. In the posthypoxic period, phrenic nerve activity remained elevated for 60 min (Fig. 1A). Average data are summarized in Fig. 2. There was a progressive increase in baseline phrenic nerve activity (MNR, AU/min) from the fourth episode of hypoxia, and this increase in phrenic activity persisted at all time points tested during the posthypoxic period (i.e., 15, 30, and 60 min; Fig. 2). On average, MNR increased from 57 ± 4 to 79 ± 10 AU/min, which was mainly due to increases in f (from 36 ± 2 to 42 ± 2 bursts/min, P < 0.05, ANOVA). Although tidal phrenic activity () increased from 1.6 ± 0.1 to 1.9 ± 0.2 AU, it was not significant (P > 0.05).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effect of acute repetitive episodic hypoxic (EH) exposures on efferent phrenic nerve activity in control (A) and in chronic intermittent hypoxia (CIH)-conditioned (B) rats. Baseline, phrenic nerve activity before hypoxic exposures. EH #1 and EH #10, 1st and 10th episode of hypoxia. Post-EH, phrenic nerve activity during 15, 30, and 60 min of the post-EH period. A.P., action potentials; Phr, integrated phrenic nerve activity. Black bar denotes the duration of hypoxic challenge. Note the enhanced phrenic nerve activity during the posthypoxic period in the CIH-conditioned animal.

View larger version (35K): [in this window] [in a new window]   Fig. 2.   Average data showing the time course of changes in phrenic nerve activity. A: respiratory rate (f; phrenic bursts/min). B: amplitude of the integrated phrenic nerve activity [Phr; in arbitrary units (AU)]. C: minute neural respiration (MNR; AU/min). EH, 10 episodes of hypoxia represented as horizontal bar. Results are expressed as percent changes from baseline values (baseline = 100).  and , Control (n = 9) and CIH-conditioned (n = 9) rats, respectively. Data are presented as means ± SE.

CIH enhances acute EH-induced LTF in phrenic nerve activity. To determine the effects of CIH, awake rats were exposed to 10 days of intermittent hypoxia (15 s hypoxia-5 min normoxia; 9 episodes/h; 8 h/day), and then the effects of acute EH on LTF were examined in anesthetized animals as described above. In one series of experiments (n = 6), Pa_CO2 was kept close to 35 Torr, and in the other series (n = 3) apneic thresholds were determined at the beginning of the experiment. Apneic threshold was 3 Torr less in CIH-conditioned compared with control animals (28 vs. 31 Torr). The effect of acute EH was examined while Pa_CO2 was maintained 2-3 Torr above the apneic threshold. In both series of experiments, LTF was more pronounced in CIH-conditioned compared with control animals (Figs. 1B and 2). Therefore, the data were combined from both series of experiments for quantitative analysis. The magnitude of LTF (MNR, AU/min) was significantly greater in CIH-conditioned compared with control animals at 15, 30, and 60 min of posthypoxic period (Figs. 2 and 3A). Increases in f (bursts/min) as well as tidal phrenic amplitude ( phrenic burst activity, AU) contributed to enhanced LTF in CIH-conditioned animals. On average, f increased from 38 ± 3 to 54 ± 4 bursts/min and  phrenic activity from 1.8 ± 0.2 to 3.5 ± 0.4 AU during the 60-min period of posthypoxia (P < 0.01, ANOVA). The magnitude of the hypoxic response (MNR, the 1st and 10th episodes of hypoxia) was comparable between both groups when normalized to baseline (Fig. 3B). There were no significant differences in changes in arterial blood gases and mean arterial blood pressure before and after acute EH in both groups of animals (Table 1).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Average data of the magnitude of long-term facilitation (LTF; f phrenic bursts/min) during 15th, 30th, and 60th minute of posthypoxia period (A) and hypoxic response during the 1st and 10th hypoxic challenge (B). Phr, amplitude of the integrated phrenic nerve activity. Open and solid bars represent data from control and CIH-conditioned animals, respectively. Data are presented as percent changes from baseline activity. Data represent means ± SE from control (n = 9) and CIH (n = 9) animals. *P < 0.05 (ANOVA) compared with control animals. Note the significant augmentation of LTF but not the hypoxic response in CIH-conditioned animals.

Conditioning with cumulative, comparable duration of SH does not enhance acute EH-induced LTF. To assess the effect of cumulative, comparable duration of SH, rats were exposed to 4 h of continuous hypoxia (12% O₂), which is equivalent to 10 days of CIH (15 s hypoxia, 9 episodes/h, 8 h/day for 10 days). The effects of acute EH on phrenic nerve activity were analyzed in anesthetized animals while Pa_CO2 was maintained at 32 Torr. In another series of three SH-conditioned animals, apneic thresholds were determined at the beginning of experiments and the effect of acute EH was determined while the Pa_CO2 was kept ~3 Torr above apneic threshold. In SH-conditioned animals, the apneic threshold was 6 Torr less than in control rats (25 vs. 31 Torr). The effects of acute EH were found to be qualitatively the same under both experimental conditions. Therefore, the data were pooled from both groups. Although hypoxia increased phrenic nerve activity, LTF was nearly absent in SH-conditioned rats (Fig. 4A). Average data showed that LTF (expressed as MNR, AU/min) was absent in SH-conditioned animals at all time points tested (15, 30, and 60 min; Fig. 4B). The lack of LTF was due to an absence of changes both in f as well as tidal phrenic activity [baseline f (bursts/min), tidal phrenic activity ( AU) 41 ± 4, 1.3 ± 0.2 vs. 60 min posthypoxic period 41 ± 4, 1.5 ± 0.3, respectively; P > 0.05; ANOVA]. The magnitude of hypoxic response tended to be less in the SH-conditioned than the control animals (Fig. 4C). The effects of SH on acute EH-induced LTF was tested in three additional animals by using 5% instead of 12% inspired O₂ for 4 h. There was no LTF of respiration (MNR, AU/min) in all three animals conditioned with 4 h of 5% O₂. There were no significant alterations in arterial blood gases and blood pressure before and after acute EH in animals exposed to SH (Table 1).

View larger version (36K): [in this window] [in a new window]   Fig. 4.   Absence of LTF in animals conditioned with cumulative, comparable duration of sustained hypoxia. A: representative tracings showing the changes in phrenic nerve activity before episodic hypoxia challenges (baseline), during the 1st (EH #1) and 10th (EH #10) episodes of hypoxia, and during 15, 30, and 60 min after episodic hypoxia (Post-EH). Black bar indicates duration of hypoxic challenges. B and C: average data of the magnitudes of LTF (B) and hypoxic responses (C) in control (n = 9) and SH-conditioned (n = 8) rats. Data are averaged at the 15th, 30th, and 60th min of the posthypoxic period, and the 1st and 10th episodes of hypoxia and are presented as percent changes from baseline activity. Data represent means ± SE. *P < 0.05 (ANOVA) compared with control animals. Note the suppressed LTF and hypoxic responses after conditioning with cumulative, comparable duration of sustained hypoxia.

SOD mimetic prevents CIH-induced potentiation of LTF of respiratory activity. To test the role of ROS, rats were treated daily with the superoxide dismutase (SOD) mimetic MnTMPyP (5 mg · kg¹ · day¹ for 10 days), a potent scavenger of O·, and then were subjected to 8 h/day of intermittent hypoxia for 10 days (n = 9). After CIH, the effects of acute EH on phrenic nerve activity were analyzed in anesthetized animals. In six of the animals, the effects of EH were studied while Pa_CO2 was maintained at 35 Torr. In three animals, apneic thresholds were determined at the beginning of experiments and the effect of acute EH was determined while the Pa_CO2 was kept ~3 Torr above apneic threshold. Apneic thresholds in these three animals, however, were found to be the same as in control animals (CIH + SOD 30 vs. control animals 31 Torr). The results were found to be qualitatively the same in both series of experiments. Therefore the data were pooled from both series of experiments. Figure 5A depicts a representative example of the effect of EH on phrenic nerve activity in a SOD mimetic-treated, CIH-conditioned animal. Phrenic nerve activity increased with each episode of hypoxia. However, after termination of EH, phrenic nerve activity returned to baseline levels during the posthypoxic period. Analysis of the time course of the MNR showed that baseline phrenic activity gradually increased with each hypoxic episode (Fig. 5B). Before the 10th hypoxic episode, baseline phrenic activity was significantly higher than the baseline values before EH (f = 37 ± 3 vs. 45 ± 3 bursts/min;  = 2.16 ± 0.3 vs. 2.74 ± 0.4 AU; MNR = 83 ± 15 vs. 123 ± 18 AU/min; P < 0.01). After termination of EH, phrenic nerve activity returned close to baseline values in CIH-conditioned animals treated with SOD mimetic. The magnitude of LTF (MNR, AU/min) was significantly less in CIH-SOD mimetic-treated compared with CIH-conditioned animals at all time points tested during the posthypoxic period (Fig. 5C). On the other hand, the magnitude of hypoxic response to individual hypoxic challenge was comparable between both groups of animals (P > 0.05). Arterial blood gases and blood pressure values in CIH-conditioned rats treated with SOD mimetic were comparable before and after EH (Table 1).

View larger version (36K): [in this window] [in a new window]   Fig. 5.   Pretreatment with superoxide dismutase (SOD) mimetic prevents the maintenance of LTF in CIH-conditioned rats. A: representative tracings showing the changes of phrenic nerve activity before episodic hypoxia challenges (baseline), during the 1st (EH #1) and 10th (EH #10) episodes of hypoxia, and 15, 30, and 60 min after the EH challenge (Post-EH). Black bars indicate duration of hypoxic challenges. B: average data of the time course of MNR (AU/min) during 10 episodes of hypoxia and during the posthypoxic period in CIH-conditioned rats treated with SOD mimetic. Data represent means ± SE from 9 rats. Horizontal bar represents hypoxic challenges. C: average data of the magnitude of LTF during posthypoxia period at the 15th, 30th, and 60th min in CIH (vehicle control, n = 9) and CIH+SOD (n = 9) rats. Results represent percent change from baseline. Data are means ± SE. *P < 0.05 compared with vehicle control animals. Note that LTF was not maintained during the posthypoxic period in SOD mimetic-treated, CIH-conditioned animals.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The results of the present study demonstrate that LTF of breathing can be elicited even with acute repeated exposures to brief hypoxia (15 s), an effect that was markedly enhanced by prior conditioning with CIH. Furthermore, our data indicate that ROS, especially O·, play a role in CIH-induced augmentation of LTF.

Previous studies have reported LTF of breathing in response to episodic hypoxia in anesthetized rats (2, 5, 10). In these studies, each hypoxic episode lasted 5 min. It is clear from the present results that brief episodes of hypoxia, lasting no more than 15 s, are also adequate in inducing LTF in phrenic nerve activity in anesthetized rats. LTF observed in the present study was not due to changes in arterial blood gases or blood pressure, because these variables were maintained fairly constant during the posthypoxic period. However, some differences were noted between our results and the previous reports (2, 5, 10). First, the magnitude of LTF seen with our paradigm of EH was modest (~40% increase in MNR) and reflected in increases in f (bursts/min) rather than tidal phrenic nerve activity (Fig. 2 and 3). The magnitude of LTF reported in earlier studies in response to longer duration of hypoxic episode (i.e., 5 min) appears to be greater and was due primarily to increases in tidal phrenic amplitude (2, 5, 10). The lesser magnitude of LTF seen in the present study is conceivably due to the briefness of each hypoxic episode (i.e., 15 s). Nonetheless, despite some minor differences, our observations are consistent with the earlier reports that repetitive hypoxia induces LTF in respiration and further demonstrate that the duration of the individual hypoxic episodes is not a determining factor for eliciting LTF under given experimental conditions.

LTF elicited by acute EH was markedly enhanced in rats conditioned with CIH. These observations are consistent with a recent study by Ling et al. (10), who reported that prior conditioning with EH augments LTF elicited by acute EH, wherein each episode was 5 min. The CIH paradigm employed by Ling et al. consisted of 5-min hypoxia followed by 5-min normoxia; 12 h/night for 7 days, whereas we employed the paradigm more commonly encountered in recurrent apneas (15 s hypoxia-5 min normoxia; 9 episodes/h; 8 h/day). Furthermore, because apneas are more frequently encountered during sleep, our CIH exposures were done during daytime, because rats are nocturnal animals. Despite the differences in experimental paradigms, the present data along with the results by Ling et al. (10) clearly demonstrate that prior CIH conditioning enhances LTF of respiratory motor output elicited by acute EH. The enhanced LTF in CIH-conditioned animals in the present study appears not to be due to changes in either arterial blood gases or blood pressure, because changes in these variables were comparable between the control and CIH-conditioned animals. Although it is possible that the enhanced LTF is due to the effects of CIH on central neurons responsible for LTF expression (10), it is likely that carotid body inputs also contribute to the enhanced LTF, because recently Peng et al. (17) reported LTF in carotid body sensory discharge in CIH-conditioned but not in control animals. In addition, changes in threshold for ventilatory response to CO₂ (i.e., apneic threshold) might have also contributed to the enhanced LTF seen in CIH-conditioned animals. However, this is unlikely because SH-conditioned animals, in which no LTF was observed, exhibited a similar depression of CO₂ apneic threshold.

Interestingly, conditioning with comparable, cumulative duration of SH prevented the development of LTF in response to EH. Changes in arterial blood gases or blood pressure cannot account for the lack of LTF, because changes in these variables were comparable before and after acute EH in SH-conditioned animals (Table 1). Nor can the changes in CO₂ apneic threshold explain the absence of LTF in SH-conditioned animals because apneic thresholds were lower both in SH- and CIH-conditioned animals, and LTF was absent in SH, whereas it was potentiated in CIH-conditioned animals. We employed 12% O₂ for SH conditioning, whereas 5% O₂ was used for CIH conditioning. Consequently, it can be argued that the absence of LTF in SH animals was due to less severe hypoxia used for conditioning. However, such a possibility seems unlikely because a similar absence of LTF was also observed in three SH animals exposed to 5% O₂ for 4 h. However, phrenic nerve response to individual hypoxic episodes tended to be less in SH-conditioned animals. Therefore, it is possible that decreased peripheral chemoreceptor response to hypoxia might be one reason for the absence of LTF in SH animals. Alternatively, acidosis of the central neurons after 4 h of SH might have contributed to the absence of LTF. Nonetheless, these observations demonstrate that conditioning with repetitive or continuous hypoxia has profoundly different effects on LTF of the respiratory motor output, wherein the former enhances but the latter prevents LTF.

What makes the episodic pattern of hypoxia different from comparable duration of cumulative hypoxia? The feature that distinguishes episodic from the sustained pattern of hypoxia is the intervening periods of normoxia in the former but not in the latter. It has been recently suggested that episodic hypoxia represents a form of oxidative stress, wherein ROS, especially O· anions, released during the intervening normoxic periods play an important role in eliciting the physiological response to intermittent hypoxia (20, 21). Our results that LTF is nearly absent in CIH-conditioned animals treated with a daily dose of SOD mimetic, a potent scavenger of O·, support the idea that CIH represents a form of oxidative stress. The absence of LTF in SOD mimetic-treated animals is not due to changes in arterial blood pressure or blood gases (Table 1) or alterations in apneic threshold for CO₂, because changes in these variables were comparable between controls and SOD mimetic-treated animals. Because the magnitude of the response to individual hypoxic challenge was comparable, it is unlikely that reduced peripheral chemoreceptor activity contributed to the absence of LTF. Therefore, it follows that ROS are involved in the enhanced LTF elicited by CIH conditioning. It is interesting to note that baseline phrenic nerve activity progressively increased with each successive episode of hypoxia (Fig. 5). However, the increase in phrenic nerve activity was not sustained but rather returned promptly to baseline values within 10 min of the posthypoxic period (Fig. 5B). These observations suggest that LTF develops in SOD mimetic-treated animals. However, it is not maintained during the posthypoxic period. It has been suggested that induction and maintenance of LTF in breathing requires distinct mechanisms (see Ref. 15). Activation of 5-HT receptors has been proposed to be critical for the induction of LTF of respiratory motor output (7, 13). On the other hand, the mechanisms associated with maintenance of LTF have been largely unexplored. Our data suggest that ROS, especially O·, are involved in the maintenance rather than induction of LTF in CIH-conditioned animals. The mechanisms by which O· contributes to LTF of breathing are beyond the scope of the present investigation. It is possible that O· might be exerting effects by enhancing gene expression and the resulting protein synthesis and/or oxidative modification of the existing proteins that are involved in the maintenance of LTF. In addition, there might be interactions between 5-HT and O· that contribute to the CIH-induced potentiation of LTF in respiration. These possibilities, however, require further study.

In summary, the results of the present study demonstrate that acute exposure to repetitive brief hypoxia elicits LTF in phrenic nerve activity and prior conditioning with CIH-enhanced acute EH-induced LTF. Our data further demonstrate that O· plays an important role in the maintenance of the enhanced LTF in CIH-conditioned animals.