Episodic hypoxia induces a persistent augmentation of respiratory activity, termed long-term facilitation (LTF). Phrenic LTF saturates in anesthetized animals such that additional episodes of stimulation cause no further increase in LTF magnitude. The present study tested the hypothesis that 1) ventilatory LTF also saturates in awake rats and 2) more severe hypoxia and hypoxic episodes increase the effectiveness of eliciting ventilatory LTF. Minute ventilation was measured in awake, male Sprague-Dawley rats by plethysmography. LTF was elicited by five episodes of 10% O2 poikilocapnic hypoxia (magnitude: 17.3 ± 2.8% above baseline, between 15 and 45 min posthypoxia, duration: 45 min) but not 12 or 8% O2. LTF was also elicited by 10, 20, and 72 episodes of 12% O2(19.1 ± 2.2, 18.9 ± 1.8, and 19.8 ± 1.6%; 45, 60, and 75 min, respectively) but not by three or five episodes. These results show that there is a certain range of hypoxia that induces ventilatory LTF and that additional hypoxic episodes may increase the duration but not the magnitude of this response.
- intermittent hypoxia
- respiratory control
repeated carotid sinus nerve stimulation induces a serotonin-dependent augmentation of respiratory activity that lasts for many minutes in anesthetized, vagotomized, paralyzed, and artificially ventilated cats (18,28) and rats (20, 24), which is termed long-term facilitation (LTF). Episodic isocapnic hypoxia has also evoked LTF in phrenic (4, 19) and hypoglossal (4) nerve activity in anesthetized rats. These forms of respiratory plasticity are quite robust and can last up to 1 h or more. In the unanesthetized preparation, awake goats (37), dogs (10), and ducks (29) exhibit a more moderate LTF after several episodes of isocapnic hypoxia. Recently, it has been shown that repeated episodes of both isocapnic and poikilocapnic hypoxia gave rise to LTF in awake rats (31). It has also been shown that episodic but not sustained hypoxia (even with longer exposures) induces LTF in both the anesthetized and unanesthetized preparations (6, 14, 16, 31). This different result indicates that episodic hypoxia exerts different effects on the hypoxic ventilatory control system compared with sustained hypoxia, suggesting that the pattern of hypoxic stimulation could be a critical determinant of this respiratory plasticity.
The protocols used in the literature to induce LTF employ a range of hypoxic episodes at differing O2 levels in different species. Some protocols were successful in inducing LTF (10, 31,37), whereas others were not. For example, LTF could not be induced in spontaneously breathing anesthetized rats by using episodic hypoxia (21) or in decerebellated rats by using repeated carotid sinus nerve stimulation (20). Also, LTF has yet to be shown in normal awake or sleeping humans (2, 25). However, whether the hypoxia selected is sufficient to induce LTF in different protocols or their differing results relate to different species, genetics, age, or gender is uncertain. Given the varying severity and number of hypoxic episodes used to induce LTF, it seems important to determine whether hypoxic severity and number of hypoxic episodes could affect LTF.
During carotid sinus nerve stimulation of the anesthetized cat, Millhorn et al. (28) observed that increasing the number of stimulations did not further increase the magnitude of the phrenic nerve activity after stimulation. Eldridge and Millhorn (15) suggested that phrenic LTF could be saturated. The potential respiratory LTF saturation in awake animals has yet to be investigated. In anesthetized rats, although different levels of hypoxia [arterial Po 2(PaO2) = 26–54 Torr] did not appear to affect the expression or magnitude of LTF, the hypoxia ventilatory response did have a weak but significant correlation with the magnitude of LTF (19). However, these investigators did not evaluate these effects of hypoxic severity in a systematic way and these observations came from experiments in anesthetized rats. The present study was designed to systematically examine the influence of hypoxic severity and episode number on ventilatory LTF in unanesthetized, spontaneously breathing rats by using barometric plethysmography. We hypothesized that ventilatory LTF saturates in awake rats and that the stimulus protocols with more severe hypoxia and hypoxic episodes are more effective in eliciting ventilatory LTF in awake rats.
The Harvard Medical Area Standing Committee on Animals approved all experimental procedures. Experiments were conducted on 62 male Sprague-Dawley rats (300–320 g, Colony 205, Harlan Sprague Dawley, Madison, WI). Ventilatory measurements were made by using a custom-made 3-liter whole body flow-through plethysmograph (Buxco Electronics, Sharon, CT). Individual unanesthetized, unrestrained rats were placed in the precalibrated plethysmographic chamber connected via a controlled leak to a reference chamber. The gas atmosphere within the animal chamber was maintained with air flowing through the chamber at a rate of 3 l/min. A bias flow was connected to an aerosol port of the chamber to maintain the O2concentration within the unit. A custom-made computer software system (Biosystem XA, Buxco Electronics) monitored the output from a differential pressure transducer (TRD5100) connected between the animal and reference chambers. This software system provided a breath-by-breath display of the respiratory frequency (fr), tidal volume (Vt), and minute ventilation (V˙e) before, during, and after the episodic hypoxia protocols used to elicit LTF. Body temperature was measured before and after each experiment.
Resting V˙e (baseline) was measured over a 10-min period. V˙e was then monitored continuously throughout the intermittent hypoxia stimulus protocols. Rats were exposed to 5 min of poikilocapnic hypoxia followed by 5 min of air. This cycle was repeated to achieve the number of episodes required for the particular protocol. During these hypoxic and normoxic episodes, only the final 2-min data were averaged and analyzed. The hypoxia ventilatory response (HVR) was defined as the change from baseline in hypoxic V˙e and normalized to a percentage of the baseline. After the last hypoxia episode, V˙e was measured at 15-min intervals (i.e., 15, 30, 45, and 60 min) with each value representing a 5-min average (e.g., the 30-min posthypoxia value is an average of the data collected between 30 and 35 min) to determine ventilatory LTF. Each rat received only one protocol.
Twenty rats were divided into three groups. Measurements were made when rats were exposed to five 5-min episodes of poikilocapnic hypoxia of either 12% O2 (n = 7), 10% O2(n = 7), or 8% O2 (n = 6). All hypoxia episodes were interspersed with 5-min intervals of normoxia (21% O2), and V˙e was monitored up to 60 min after the last hypoxic episode.
Number of episodes of hypoxia.
Thirty-four rats were exposed to poikilocapnic hypoxia (12% O2) using three (n = 7), five (n = 7), ten (n = 7), twenty (n = 7), or seventy-two (n = 6) 5-min episodes. All hypoxia episodes were interspersed with 5-min intervals of normoxia, and V˙e was monitored up to 90 min after the last hypoxic episode.
Blood-gas measurements were carried out on a separate group of rats (n = 8) because of the necessary surgery for implantation of catheter and temperature transmitter. One week before any measurements were taken, catheters were placed into the descending aorta via the femoral artery under isoflurane inhalation anesthesia. Catheters were routed under the skin to the back of the neck and were filled with heparinized saline (20 U/ml) and capped until needed. At the same time, a small temperature telemetry transmitter (Mini Mitter, Sunriver, OR) was placed into the abdominal cavity of the rat. For blood-gas measurements, the rat was placed in the plethysmographic chamber with the femoral catheter connected via a swivel to a collection chamber that housed a heparinized syringe. Blood gases were measured during resting V˙e and during each of the three hypoxia levels described above. Blood was drawn during the last minute of the first and last episodes of hypoxia and analyzed for the pH value and the PaO2 and Pco 2 (PaCO2) in arterial blood (ABL 700; Radiometer, Copenhagen, Denmark) with correction for the rat's body temperature from telemetry. These values were averaged and used as the blood gas value for that particular hypoxia.
The fr, Vt, andV˙e were measured in rats when they were observed to be awake and in a quiet state. Data recorded when the rats were not both quiet and awake (e.g., rat moving or asleep) were rejected from the analysis. This rejection was done blindly; i.e.,V˙e was not known when it was being rejected for movement artifact. Furthermore, an additional rejection algorithm was included in the computer software breath-by-breath analysis that allows for further rejection of artificial breaths. These strict rejection criteria helped to generate more consistent results.
To facilitate analysis, LTF was also numerically expressed as the magnitude and the duration. The magnitude was determined by the average of the first three V˙e values recorded at 15, 30, and 45 min posthypoxia and was expressed as a percent above the baseline value. The magnitude of LTF was thus calculated by the equation %[(V˙e15 + V˙e30 +V˙e45)/3 − baseline]/baseline. The reason for using this “three-value-average” method was because it is equivalent to an “area under the curve” quantification and yields smoother results than expression of LTF as the single largestV˙e in the entire posthypoxia period. The duration of LTF was defined as the last posthypoxia time point at whichV˙e was significantly higher than baseline.V˙e was measured for up to 90 min after the last hypoxia episode because V˙e always returned to baseline values by this time and remained at this level over 2 h in several time control experiments.
For both hypoxic severity and episode number studies, the LTF duration for each group and the differences between groups in baseline and those individual posthypoxia V˙e were determined by use of a two-way ANOVA with repeated measures (SigmaStat version 2.0, Jandel, San Rafael, CA), followed by Student-Newman-Keuls post hoc tests. Only baseline and posthypoxia data were included in this two-way ANOVA; data recorded during and between hypoxic episodes were not included. All other differences were determined by using a one-way ANOVA, including the between-group differences in LTF magnitude calculated by the three-value-average analysis method and the between/within-group differences in HVR, values during intervals, and blood-gas and body temperature values. P < 0.05 was considered significant. All values are expressed as means ± SE.
Effect of Hypoxic Severity
The baseline V˙e values in the three groups of rats were similar (42.5 ± 0.4, 42.2 ± 0.2, and 42.5 ± 0.2 ml · 100 g−1 · min−1, with 12, 10, and 8% O2 protocols, respectively; Fig.1), indicating consistent results between experimental groups. V˙e, Vt, and fr all increased during hypoxia (Figs.1 A and 2). The HVR in each group was similar throughout the five episodes (P > 0.05; Fig. 1 A) and was therefore averaged and expressed as a percentage of baseline. The average HVR was significantly different (all P < 0.05) between any two of these three groups (12% O2 group: 50.8 ± 6.0% baseline, 10% O2: 117.5 ± 8.8%, 8% O2: 151.1 ± 7.2%). The average V˙e values during the normoxic intervals that separated the hypoxic episodes in the 12% O2 (−21.1 ± 2.0% baseline) and 10% O2(−16.0 ± 3.1%) protocols were lower than the baselineV˙e (P < 0.05), but 8% O2 normoxia intervals were similar to baseline (P > 0.05). This decrease for the 12% O2group appeared to be due to a decrease in both fr (−13.7 ± 2.9% baseline) and Vt (−7.6 ± 3.4%), although only fr showed significant difference in the statistical analysis. The decrease for the 10% O2 group was due to a decrease in fr (−13.6 ± 2.8%) but not Vt (−2.1 ± 3.1%). There was no significant decrease in Vt and frin the 8% O2 group (P > 0.05).
Ventilatory LTF was not elicited by five episodes of the 12 or 8% O2 protocols (Fig. 1). But five episodes of 10% O2 induced ventilatory LTF (17.3 ± 2.8% above baseline, using the three-value-average method). Ventilatory LTF was due mainly to an increase in the fr where Vt appeared to remain at baseline values (Fig. 2). AlthoughV˙e showed a significant increase above baseline at the 15-min time point after five episodes of 8% O2, the three-value-average value (2.7 ± 1.2% above baseline) was not different from baseline (P > 0.05).
Effect of Hypoxic Episode Number
There were no differences in the baseline V˙evalues (range from 42.5 ± 0.4 to 43.6 ± 0.8 ml · 100 g−1 · min−1;P > 0.05) among the five protocols (Fig.3 A). These baselineV˙e values are also similar to the ones in the hypoxic severity study. Neither three nor five episodes of hypoxia (12% O2) induced ventilatory LTF. However, 10, 20, and 72 episodes gave rise to LTF. The LTF magnitude was similar in the three groups (19.1 ± 2.2, 18.9 ± 1.8, and 19.8 ± 1.6%, respectively; Fig. 3 B) and also similar to the LTF elicited by five episodes of 10% O2 (P > 0.05). The induced ventilatory LTF was also due to an increase in the fr with minimal change in Vt (data not presented). The duration of LTF differed between the protocols: 45 min for 10 episodes of hypoxia, 60 min for 20 episodes, and 75 min for 72 episodes, respectively (Fig. 3).
Blood Gases and Body Temperature
Table 1 shows the blood-gas and body temperature values measured in a separate group of rats under the same conditions as those used for the V˙e measurements. Body temperature remained unchanged. Blood gases were measured during the normoxic baseline (21% O2) and during episodes of each hypoxic severity. There was a significant change in PaO2 and PaCO2 between baseline and 12% O2 hypoxia. The PaO2 and PaCO2 also showed corresponding decreases at 10 and 8% hypoxic levels. The differences between groups, however, were not statistically analyzed because of the limited number of rats in the 10 and 8% groups. These blood-gas data were collected to briefly assess the hypoxemic and hypocapnic levels in these protocols compared with other LTF studies.
The present study demonstrated that in the episodic hypoxia stimulus protocols, which were commonly used in the LTF studies reported in literature, both hypoxic severity and episode number could affect the elicitation, maintenance, and size of ventilatory LTF in awake rats. More episodes of hypoxia in a stimulus protocol usually increase the effectiveness of eliciting LTF. There is a certain range of (poikilocapnic) hypoxia severity, probably near 10% O2, that is needed to successfully elicit LTF. Either too mild or too severe hypoxia would reduce or lose the effectiveness of eliciting LTF. Our results also suggest that the notion of saturation can be extended to ventilatory LTF in awake rats; however, LTF magnitude reaches the maximal level more quickly than LTF duration when episodes are increased.
In the present study, five episodes of 10% O2 induced LTF, but five episodes of 12% O2 did not induce ventilatory LTF in awake rats. The stimulus protocols at this hypoxic level needed more than five episodes to successfully elicit LTF. We believe that the failure to elicit LTF might be due to inadequate hypoxic stimulation and that the effectiveness of eliciting LTF could be somehow accumulated as episodes increased. Five episodes of 8% O2 also failed to induce LTF, although HVR was higher than that of 10% O2. We argue against the possibility that 8% O2 hypoxia was too severe to elicit LTF because the induced PaO2 was only ∼40 Torr, a level definitely within the range eliciting LTF in anesthetized animals (19). We speculate that the profound hypocapnia induced by the poikilocapnic hypoxia might have “unfavorable” effects on the serotoninergic mechanisms that are required for LTF or persistent inhibitory effects on posthypoxia ventilation that compromise LTF. Hence the above mentioned hypoxic severity range near 10% O2 should probably be limited to poikilocapnic hypoxia cases. Our experimental design and approach do not allow discrimination between the central and peripheral mechanisms that might be involved in the inhibitory effect. Repeated carotid sinus nerve stimulation, which bypasses the carotid body, elicits phrenic LTF (15, 20,24), suggesting that the carotid body is not required for elicitation or maintenance of this LTF. Recently, it was also reported that episodic hypoxia elicits phrenic LTF in carotid sinus nerve-denervated rats, although the LTF magnitude is relatively smaller (8). These results, however, do not totally rule out the possibility that the hypocapnia might exert its persistent inhibitory effects on the carotid body, thereby minimizing the peripheral chemoreceptor output after hypoxia.
Blood-gas measurements during hypoxia showed that the rats were getting adequate hypoxia and displayed a corresponding increase inV˙e. The PaO2 values during the three hypoxic levels were all within the PaO2 range (35–45 Torr) commonly used to elicit LTF in anesthetized preparations (4, 6, 19, 23). The PaO2 values appeared to be unusually higher in these three hypoxic levels, probably because of two factors:1) the same level of poikilocapnic hypoxia induced less severe hypoxemia than isocapnic hypoxia, and 2) they were not “equilibrium” values. The blood samples were taken during the last minute of 5-min hypoxia. However, it took ∼1 min to reach the target hypoxia level in the chamber and might take longer time to reach a stable level in arterial blood. Indeed, in a separate study in which PaO2 was analyzed after 10-min poikilocapnic hypoxia, the values were obviously lower [PaO2: 12% = 40.8 ± 1.3 (n = 6); 10% = 33.5 ± 0.5 (n = 4); 8% = 20.15 ± 0.6 (n = 5); unpublished data]. On the other hand, the PaCO2 values appeared to decrease more quickly with decreasing O2levels, likely because of the increased HVR.
The HVR increased as hypoxic severity rose, ∼1.5-fold increase inV˙e during 12% O2, ∼2-fold during 10% O2, and ∼2.5-fold during 8% O2. However,V˙e showed a decrease in the normoxic intervals between 12% O2 and 10% O2 hypoxia episodes. The decrease in interval V˙e might be due to the poikilocapnic hypoxia-induced hypocapnia that reducesV˙e. Blood-gas measurements taken from two (12% O2) rats during the last minute of the first normoxic interval suggested that although PaO2 had returned to baseline, PaCO2 tended to be lower (38 ± 0.5 Torr) than the baseline value (41.4 ± 0.3) for these two rats. This V˙e decrease appeared to result from a decrease in both fr and Vt, although only the fr decrease was significant in the statistical analysis. This posthypoxia frequency decline (PHFD) was expressed as the first interval fr in the present study. The PHFD, an undershoot in fr after a brief exposure to hypoxia, has been shown to occur in anesthetized, ventilated rats (3, 11, 12, 20, 23) and anesthetized, spontaneously breathing rats (21). However, PHFD was not present in the unanesthetized studies with other species. For example, awake goats (37) and dogs (10) exhibited an increase in fr during normoxic intervals, whereas awake humans showed levels similar to baseline (25). Thus PHFD might be a unique phenomenon in rats. Severe poikilocapnic hypoxia induced profound hypocapnia, which might in turn affect this PHFD. In the present study, 8% O2 appeared to induce PHFD, but the first-interval fr decrease was not significant (P = 0.09). The PHFD was also not expressed in another study on awake rats using a protocol with poikilocapnic hypoxia of 11% O2 (31).
Number of Hypoxic Episodes
Three episodes of 12% O2, a protocol that induces LTF of phrenic (4, 6) and hypoglossal (4) nerve activity in the anesthetized, paralyzed, vagotomized, and ventilated preparation, failed to elicit ventilatory LTF in awake rats in the present study. This result is similar to one study in anesthetized, spontaneously breathing rats in which three episodes of 12–13% O2 failed to induce LTF (21). Five episodes also failed to induce LTF. However, 10, 20, and 72 episodes of 12% O2 elicited ventilatory LTF. These results indicate that more episodes increased the effectiveness of protocols in inducing ventilatory LTF in awake rats. The magnitude of LTF was not significantly increased when the episode number was raised from 10 to 20 or even 72. This result is consistent with the phrenic LTF saturation data obtained from anesthetized animals (15), suggesting that LTF magnitude also saturates in awake rats; i.e., additional episodes of hypoxia do not further increase ventilatory LTF magnitude. However, LTF duration had not saturated at least at the 10-episodes level, as it increased with episode number increasing. We speculate that 75 min may be the maximal LTF duration for this animal model and can probably be achieved by hypoxic episodes <72. The LTF duration saturation has not been studied before. In most anesthetized studies, phrenic nerve activity was monitored 60 or 90 min after episodic stimulation and remained elevated above baseline.
In the present study, the parameters like hypoxic severity, duration, and episode number were selected because they were commonly used in other LTF studies reported in the literature. Five minutes is the most common duration for a hypoxic episode (4, 8, 19); 3–20 episodes of hypoxia are also mostly used (1, 2, 4, 10,21, 23, 25, 27, 29, 31, 37). The protocol with 72 episodes was selected because it could test the LTF saturation hypothesis in an extreme way and was technically feasible in our laboratory. The 72 episodes were achieved by simply treating rats in the first night of the chronic intermittent hypoxia (12% O2/air: 5 min/5 min, 12 h/night for 7 consecutive nights), a routine long-term hypoxic treatment protocol used for other purposes (23, 26). However, there are many other protocols in the literature using different hypoxic duration and intervals, e.g., those used in the studies investigating pathological effects of intermittent hypoxia (see Ref. 30 for references). Certain forms of chronic intermittent hypoxia have been implicated in systemic hypertension (7, 13, 17), sudden infant death syndrome (5, 32,38), and the pathological effects associated with obstructive sleep apnea (OSA; Ref. 34). Most stimulus protocols in those studies consist of relatively shorter episodes (<2-min) with more severe hypoxia, which might have great biological relevance because those episodes of apnea that often occurred in OSA patients could be more faithfully simulated. The LTF studies, however, were intended to study the neural plasticity involved in respiratory motor control. Thus the episodic hypoxia was not trying to mimic those episodes of apnea but to induce system neural plasticity by physiological stimulation.
V˙e measurements in the unanesthetized and unrestrained animal using plethysmography provided a more natural and physiological way to investigate this form of plasticity, in which possible confounding issues related to anesthesia, surgery, and restraint (22, 33, 36) could be eliminated.V˙e was measured under a strict rule by which the rat's activity was taken into consideration. V˙eduring a period of activity was blindly rejected, and only the data recorded during the awake and quiet state were included, similar to data reported by Borday et al. (9). BaselineV˙e in the present study was very consistent between and within experimental groups. Data of ventilatory LTF magnitude and duration were also consistent in each group.
V˙e can be influenced by changes in blood gases (PaO2 and PaCO2), metabolic rate, body temperature, and/or respiratory dead space. In the present study, 5-min hypoxic episodes did not change body temperature during and after hypoxia. In separate time-control experiments, metabolic rate (defined as CO2 production) and blood gases were measured at 15 min posthypoxia for each of the three hypoxic levels (n = 1 for each level). Although CO2production and PaO2 changed during hypoxia, they tended to return to near baseline values at 15 min posthypoxia. Respiratory dead space also appeared to be normal (∼20% in ratio of dead space to Vt) at 15 min posthypoxia (12% and 10% O2) in two separate rats. However, the rats remained slightly hypocapnic (PaCO2 = 41–43 Torr) at 15 min posthypoxia comparing to their own baseline value (44–46 Torr). Collectively, these results from separate rats suggest that the poikilocapnic hypoxia-induced changes in PaO2, metabolic rate, and respiratory dead space have little carryover influence on the posthypoxiaV˙e values that determine LTF. However, the posthypoxia hypocapnia suggests that the LTF magnitude values were slightly underestimated and could have been slightly bigger if in an isocapnic condition.
In the present study, the manifestation of ventilatory LTF was due to an increase in fr with little change in Vt. This result was also reported in the LTF of awake goats (37) and rats (31) but not in awake dogs, in whom LTF was due to an increase in the Vt with little change in fr (10). In the present study, V˙e was elevated above baseline for a duration of 45 min after five episodes of 10% O2, similar to other awake studies (10, 37). However, the pattern of LTF is different from that in other studies of anesthetized rats using similar protocols, in which phrenic nerve activity remains above baseline values for 60 min or more after the last hypoxia stimulus (4,6). It is also different from another study using awake rats, in which posthypoxia V˙e was increased above baseline for up to 60 min (31). In general, it is more difficult to elicit LTF in awake animals vs. anesthetized ones, and the LTF magnitude and duration are also relatively smaller and shorter. Poikilocapnic hypoxic episodes are easier to use but usually less effective in eliciting LTF than isocapnic hypoxia.
In humans, there is controversy as to whether LTF can be induced. Although there was a preliminary report suggesting that normal awake men exhibit LTF after seven episodes of hypoxia (35), other investigators have shown that LTF was not present in normal humans awake (25) or asleep (2). At the present time, LTF has been shown to be present only in snorers (with inspiratory flow limitation) during non-rapid eye movement (NREM) sleep. There have been reports however, that OSA patients exhibit LTF. A preliminary study (27) suggested that, during wakefulness, OSA patients exhibit LTF at 15 min after the last hypoxia stimulus. Aboubakr et al. (1) also showed that the upper airway resistance decreased after episodic hypoxia in OSA patients during NREM sleep. Although these investigators did not see any changes in ventilation after the episodic hypoxia stimulus in their OSA patients, they did suggest that the decrease in upper airway resistance might be attributed to LTF of the upper airway dilator muscles.
In short, although differences in species, genetics, age, gender, and experimental preparation may account for some of the variation in LTF, the results of the present study suggest that the parameters of the episodic hypoxic stimulus protocols should also be taken into consideration.
This work was supported by National Heart, Lung, and Blood Institute Grant HL-64912.
Address for reprint requests and other correspondence: L. Ling, Div. of Sleep Medicine, Brigham and Women's Hospital, Harvard Medical School, Boston, MA 02115 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
August 16, 2002;10.1152/japplphysiol.00405.2002
- Copyright © 2002 the American Physiological Society