Journal of Applied Physiology

Level and duration of developmental hyperoxia influence impairment of hypoxic phrenic responses in rats

R. W. Bavis, E. B. Olson Jr., E. H. Vidruk, G. E. Bisgard, G. S. Mitchell


Developmental hyperoxia (1-4 wk of 60% O2) causes long-lasting impairment of hypoxic phrenic responses in rats. We hypothesized that shorter or less severe hyperoxic exposures would produce similar changes. Hypoxic phrenic responses were measured in 3- to 5-mo-old, urethane-anesthetized rats exposed to 60% O2 for postnatal day 1 or week 1 or to 30% O2 for postnatal week 1. Whereas 1 day of 60% O2 had no lasting effects (P > 0.05 vs. control), both 1 wk of 60% O2 and 1 wk of 30% O2 decreased adult hypoxic phrenic responses (P < 0.05 vs. control), although the effects of 30% O2 were smaller. Hypoxic ventilatory responses (expressed as the ratio of minute ventilation to metabolic CO2 production) were also reduced in unanesthetized rats (5-10 mo old) exposed to 1 wk of 60% O2 during development (P < 0.05). An age-dependent increase toward normal hypoxic phrenic responses was observed in rats exposed to 1 wk of 60% O2 (P < 0.05), suggesting a degree of spontaneous recovery not observed after 1 mo of 60% O2. These data indicate that long-lasting effects of developmental hyperoxia depend on the level and duration of hyperoxic exposure.

  • plasticity
  • control of breathing
  • development
  • chemoreceptor
  • hypoxia

alterations in respiratory gas concentrations can have profound influences on the development of vertebrate respiratory control (5, 28, 31). For example, humans raised at high altitude or born hypoxic because of cardiorespiratory diseases may acquire blunted hypoxic ventilatory responses during development (10, 23, 33, 38, 39). There is some debate as to whether this blunting is permanent (10, 41), but many of these changes in respiratory control persist long after normoxic conditions are restored. Animal models support the conclusion that perinatal changes in O2 levels induce developmental plasticity: lasting changes in the respiratory phenotype that can be elicited only during critical periods of development (cf. Ref. 5). In rats, chronic hypoxia for the first postnatal week blunts the adult hypoxic ventilatory response for at least 6-7 wk posthypoxia, whereas hypoxic exposures in postpubertal rats have no similar, long-lasting effects on the control of breathing (2, 35).

Chronic hyperoxia during postnatal development also attenuates hypoxic ventilatory responses in rats and cats (20, 28). The influence of developmental hyperoxia on the control of breathing has primarily been studied in rats exposed to 60% O2 for the first month of life. These rats exhibit substantial reductions in adult ventilatory and phrenic responses to hypoxia (1, 16, 17, 27, 30), primarily due to impaired carotid body function (4, 11, 16, 28). In contrast, rats exposed to the same level and duration of hyperoxia as adults have normal hypoxic responses (27, 30), indicating that this plasticity is specific to development (see also Ref. 1). The functional impairment caused by developmental exposure to 1 mo of 60% O2 appears permanent because carotid body and phrenic responses to hypoxia remain depressed between 3 and 15 mo of age (16).

One month of hyperoxia may exceed the required exposure for functional impairment of the hypoxic ventilatory response. Similar to 1-mo exposures, 1 wk of 60% O2 during development reduces phrenic (1) and carotid body (4) responses to hypoxia in 3- to 5-mo-old rats. Morphological changes in the carotid body and chemoafferent neurons (e.g., reduced carotid body volume and chemoafferent neuron degeneration) are also observed after both 1 wk and 1 mo of developmental hyperoxia (Refs. 13, 16, 17; Z.-Y. Wang and G. E. Bisgard, unpublished data). Whether these changes translate into reduced ventilatory responses in awake rats, or exhibit the same persistence observed after 1 mo of developmental hyperoxia, is not known. Shorter (i.e., <1 wk) developmental exposures to hyperoxia have not been investigated. Less severe developmental hyperoxia (30% O2) may also attenuate hypoxic ventilatory responses and carotid body function (11, 13, 20), but lasting effects of mild hyperoxia have not been studied.

We tested the hypothesis that shorter and/or less severe exposures to developmental hyperoxia induce long-lasting changes in hypoxic responses similar to those reported after 1 mo of 60% O2. Initially, we investigated the effects of short periods of postnatal hyperoxia (1 day or 1 wk of 60% O2, or 1 wk of 30% O2) on hypoxic phrenic responses in young adult rats. In subsequent experiments, we examined the effects of 1 wk of 60% O2 on the hypoxic ventilatory response in awake rats. On the basis of these results, we further investigated time-dependent changes in hypoxic phrenic responses after 1 wk of developmental hyperoxia. The major new findings of this study are 1) small, but persistent, attenuation of hypoxic phrenic responses after mild developmental hyperoxia (i.e., 30% O2), and 2) gradual, spontaneous recovery of hypoxic responses after hyperoxic exposures of shorter duration. Characterization of the exposure duration and severity required for hyperoxia-induced plasticity is key to understanding the clinical and biological significance of developmental plasticity and will facilitate future studies designed to identify underlying mechanisms.


All experimental procedures were approved by the Animal Care and Use Committee at the University of Wisconsin-Madison.

Experimental Groups

Experiments were conducted on 3- to 12-mo-old, male Sprague-Dawley rats (colony 236b, Harlan Sprague Dawley, Madison, WI) from four groups: 1) 60% O2 for postnatal day 1 (day 60% O2; 4 litters), 2) 60% O2 for postnatal week 1 (week 60% O2; 20 litters), 3) 30% O2 for postnatal week 1 (week 30% O2; 5 litters), and 4) age-matched control rats born and raised in normoxia (control; >6 litters). The number of individual rats used for phrenic and ventilation measurements are indicated below. Although hypoxic responses were determined on more than one rat from a single litter in some cases (no more than 3 rats/litter), care was taken to sample from multiple litters to ensure genetic diversity in each experiment; no familial effects were apparent in the data.

Week 60% O2 and week 30% O2 hyperoxia-treated rats were produced by placing individually housed, late-gestation dams into a large chamber flushed with 30 or 60% O2 from 1 to 4 days before birth (cf. Ref. 27); dams and pups were maintained in hyperoxia for 7 days after birth. Day 60% O2 rats were born in room air and placed into the same hyperoxia chamber for 24 h at 60% O2, beginning within 12 h after birth. Exposure to 1 wk of 60% O2 beginning during the second postnatal week (i.e., normoxic throughout gestation and postnatal week 1) is sufficient to attenuate hypoxic phrenic responses in adult rats (1), indicating that it is not necessary to begin hyperoxic exposures before birth, or immediately at birth, to induce respiratory plasticity. During hyperoxic exposures, chamber temperature and humidity were similar to those experienced by control rats; CO2 was maintained below 0.4% by regulating airflow through the chamber. At the conclusion of hyperoxia treatment, rats were removed from the chamber and raised in room air under standard conditions.

The control group for this study included two groups of rats born and raised in house under normoxic conditions. Only one environmental chamber was available, so it was impossible to treat all control rats born in house in the “hyperoxia” chamber while maintaining a randomized, blinded experimental design. Thus, four litters were born and raised in parallel to the hyperoxia-treated rats but outside the chamber (i.e., in room air). At a later date, to control for hyperoxia-independent chamber effects, two litters were born and raised for the first postnatal week in the same chamber used for the hyperoxia treatment, but the chamber was flushed with air. No differences in hypoxic phrenic responses were detected for control rats reared in the chamber (n = 4) vs. room air (n = 9), so data for all normoxia-reared rats have been pooled into a single control group. For studies of ventilation in awake rats, an additional group of five adult rats was obtained directly from the supplier; because no differences were observed between these rats and control rats born in house (n = 4), these data were combined into a single control group.

Phrenic Nerve Responses to Hypoxia

Phrenic nerve responses to hypoxia were initially assessed in 3- to 5-mo-old rats from each experimental group: day 60% O2 [n = 8; 387 ± 13 (SE) g], week 60% O2 (n = 5; 426 ± 10 g), week 30% O2 (n = 11; 412 ± 12 g), and control (n = 13; 409 ± 7 g). Additional week 60% O2 rats were then studied at ∼6-7 mo (n = 5; 458 ± 10 g), 9-10 mo (n = 10; 510 ± 17 g), and 12 mo (n = 6; 553 ± 26 g). The surgical preparation and experimental protocol were identical to those used in an earlier study (1).

Surgical preparation. Rats were rapidly anesthetized with isoflurane in a closed box and then maintained with 2.5% isoflurane [inspired O2 fraction (FiO2) = 0.5-balance N2]. Once anesthetized, rats were pump ventilated (model 683 rodent respirator, Harvard Apparatus, Holliston, MA) through a tracheal cannula and bilaterally vagotomized in the midcervical region (to prevent entrainment of phrenic neurograms to the ventilator). A catheter was placed into the right femoral vein, and the rats were gradually converted to urethane anesthesia (1.6 g/kg iv) over a 30-min period. Withdrawal and/or blood pressure responses to toe pinch were monitored throughout the experiment, and supplemental urethane was given as needed. A catheter was placed into the right femoral artery to monitor blood pressure (P23ID pressure transducer and P122 amplifier, Gould, Valley View, OH) and to collect blood samples (0.2-0.3 ml). Arterial blood samples were analyzed for Po2 and Pco2 with a blood analysis system (model ABL-500, Radiometer, Copenhagen, Denmark) and corrected to rectal temperature, which was maintained at 37-38°C with a heated table. Once converted to urethane, rats were paralyzed with pancuronium bromide (3.25-3.5 mg/kg iv) to prevent spontaneous breathing movements, and thereafter they received continuous intravenous infusion (2-2.5 ml/h) of 5% sodium bicarbonate-lactated Ringer solution (1:11 vol/vol). End-tidal Pco2 (PetCO2) was monitored with a flow-through CO2 analyzer (Capnogard, Novametrix, Wallingford, CT).

The left phrenic nerve was isolated by using a dorsal approach, cut distally, desheathed, submerged in mineral oil, and placed on a bipolar silver wire electrode. After the phrenic nerve was isolated, PetCO2 was regulated near 40 Torr until the experiment was started (FiO2 = 0.50). Nerve activity was amplified (×10,000), band-pass filtered (100 Hz to 10 kHz; model 1800, A-M Systems, Carlsborg, WA), and integrated (time constant = 50 ms; model MA-821RSP moving averager, CWE, Ardmore, PA). The integrated signal was digitized and processed with commercially available computer software (WINDAQ version 2.18, Dataq Instruments, Akron, OH).

Experimental protocol. Approximately 60 min after surgery, the CO2 apneic threshold was determined by increasing ventilator rate and decreasing inspired CO2 until phrenic nerve activity ceased and then slowly raising PetCO2 until rhythmic activity reappeared (i.e., the apneic threshold). Baseline neural activity was standardized among preparations by maintaining PetCO2 3 Torr above the apneic threshold (FiO2 = 0.50).

After baseline phrenic activity (20-30 min) was established, an arterial blood sample was drawn. Isocapnic hypoxic responses were assessed at three levels of hypoxia (Pao2= 60, 50, and 40 Torr). Each hypoxic episode lasted 5 min, and arterial blood samples were collected during the final 30 s of each episode. Trials were accepted if Pao2 was ±3 Torr of the target Pao2 and if PaCO2 was ±2 Torr of baseline. Rats were returned to FiO2 = 0.50 for 5-10 min between bouts of hypoxia. After the final hypoxic challenge, the “maximal” phrenic response to hypercapnia was measured by increasing PetCO2 to 85-90 Torr. Rats were then killed via urethane overdose.

Ventilatory and Metabolic Responses to Hypoxia in Awake Rats

Ventilatory and metabolic responses to hypoxia were assessed in 5- to 12-mo-old week 60% O2 (n = 10; 441 ± 13 g) and control rats (n = 8; 468 ± 17 g) by using a whole body, flow-through plethysmograph, as previously described (27, 36). Only one rat (control) was more than 10 mo old at the time of study.

Surgical preparation. At least 1 wk before ventilation and metabolism were measured, rats were implanted with in-dwelling arterial catheters and abdominal temperature telemetry transmitters (Mini-Mitter, Sunriver, OR). Rats were anesthetized with buprenorphine (0.0025-0.005 μg/100 g sc) and pentobarbital sodium (4 mg/100 g ip); local lidocaine was applied to the femoral artery during cannulation. Catheters were placed into the left femoral artery and advanced into the descending aorta while the opposite end of the catheter was routed under the skin and externalized through a small incision at the back of the neck. Catheters were filled with sterile saline and capped until the rat was studied in the plethysmograph (see below). During the same surgical session, a temperature transmitter was placed into the peritoneal cavity of each rat.

Experimental protocol. On the day of the experiment, rats were weighed and placed unanesthetized and unrestrained into a sealed 2-liter Plexiglas chamber; all measurements were made during the light portion of the light-dark cycle. Air, saturated with water vapor, was forced into the chamber through a high-impedance orifice and exited the chamber through a variable-impedance valve to a vacuum at a rate of 2 l/min. A differential pressure transducer (model PM15E, Statham Instruments, Hato Rey, PR) was used to detect pressure fluctuations in the rat chamber relative to a 2-liter reference chamber, connected to the rat chamber by a slow leak (6-s half-time). Output from the pressure transducer was recorded by a custom-made computer system and used to calculate respiratory variables; the computer also controlled the outflow valve of the plethysmograph to maintain the average chamber pressure at ambient. The plethysmograph was calibrated before each experiment by rapid injection of 0.2 ml of air into the chamber while it contained a simulated rat. This calibration signal was used with continuous on-line measurement of respiratory-related pressure fluctuations, body temperature (Mini-Mitter) and chamber temperature to calculate tidal volume (9); barometric pressure was measured at the start of each experiment. The computer software provided breath-by-breath analysis of ventilatory parameters (respiratory frequency, tidal volume, and minute ventilation) and rejected pressure fluctuations resulting from gross body movements. Inflow and outflow gases were monitored alternately by O2 (model FCX-MV, Fujikura, Tokyo, Japan) and CO2 (model LB-2, Beckman Instruments, Schiller Park, IL) gas analyzers, and metabolic parameters were calculated at 10-min intervals.

Rats were initially placed into the plethysmographic chamber flushed with air (21% O2) and given ∼20 min to adjust to the chamber before data were recorded. Ventilatory and metabolic variables were recorded continuously for 4 h, 1 h for each of the following conditions: 21% O2 (baseline), 14.5% O2, 21% O2 (recovery), and 12.5% O2; balance N2 in each case. One-hour exposures were selected to facilitate measurements under steady-state conditions and to enable direct comparisons with previous reports from our laboratory (27). Arterial blood samples (0.2-0.3 ml) were collected at the end of the first air exposure and at the end of each hypoxic exposure and analyzed for Pao2 and PaCO2 (corrected to abdominal body temperature) with a blood analysis system (model ABL-505, Radiometer).

Data Analysis

Phrenic activity was averaged in 30-s bins (immediately preceding blood sampling) under baseline conditions, during the fifth minute of hypoxia, and at the end of the hypercapnic challenge. In 3- to 5-mo-old rats, a blinded design was used to conceal the identity of rats from the investigator during data collection and preliminary analysis. Variables measured or calculated included peak amplitude of integrated phrenic activity, phrenic burst frequency, and their product, minute phrenic activity. Changes from baseline in burst amplitude and minute activity were normalized as a percentage of baseline phrenic activity (%baseline) and as a percentage of phrenic activity during hypercapnia (%maximum). This dual-normalization procedure minimizes concerns about potential normalization artifacts that can arise when neurograms are compared within and among experiments (15). Because overall conclusions were unaffected by the normalization method, only data expressed as a percent change from baseline are reported.

Ventilatory and metabolic measurements were analyzed in 10-min bins throughout each experiment. Variables measured or calculated included tidal volume, respiratory frequency, minute ventilation (V̇e), metabolic CO2 production (V̇co2), V̇e/V̇co2, and body temperature.

Changes in phrenic activity, ventilation, metabolism, and blood gases at multiple levels of hypoxia were compared among treatment groups by using two-way repeated-measures ANOVA followed by Student-Newman-Keuls post hoc tests. Changes in phrenic activity at a single level of hypoxia and CO2 apneic thresholds were compared among treatment groups by using one-way ANOVA followed by Student-Newman-Keuls post hoc tests, or by t-tests. Age-dependent changes in hypoxic phrenic responses were assessed by linear regression. Statistical tests were run by using SigmaStat (version 2.03, SPSS, Chicago, IL) and differences were considered significant at P ≤ 0.05. Data are presented as means ± SE. Where indicated, data were logarithmically transformed to meet the assumption of normality for parametric tests.


Phrenic Responses to Hypoxia

Hypoxic phrenic responses in 3- to 5-mo-old rats. Baseline and hypoxic arterial blood gases were similar among groups of 3- to 5-mo-old rats, and the average PaCO2 remained within 1 Torr of baseline during hypoxic challenges (Table 1). The CO2 apneic threshold was lower in week 60% O2 rats relative to the other groups [PetCO2 = 39 ± 1 vs. 42 ± 1 Torr in controls (P = 0.083), 43 ± 1 Torr in day 60% O2 (P = 0.014), and 41 ± 1 Torr in week 30% O2 (P = 0.046)]. As a result, PaCO2 tended to be lower in week 60% O2 rats, although this difference was not statistically significant (P > 0.05).

View this table:
Table 1.

Arterial blood-gas values during baseline (FiO2 = 0.50) and during three levels of isocapnic hypoxia in anesthetized control rats and rats exposed to hyperoxia during development

The level of O2 (21, 30, or 60% O2) during the first postnatal week of development influences hypoxic phrenic responses in 3- to 5-mo-old rats when expressed as minute phrenic activity (P = 0.007; Fig. 1). Consistent with our earlier report (1), minute phrenic activity increased by only half as much in week 60% O2 rats vs. control rats at each level of hypoxia (P = 0.010), apparently resulting from nonsignificant reductions in the amplitude and frequency of phrenic bursts (Fig. 1). Similarly, changes in minute phrenic activity during hypoxia were reduced in week 30% O2 rats relative to control rats (P = 0.027). Although the hypoxic responses of week 30% O2 rats appear intermediate to those of week 60% O2 and control rats (Fig. 1), differences between week 30% O2 and week 60% O2 rats are not statistically significant when all levels of hypoxia are analyzed together (P = 0.150). However, when the analysis is restricted to Pao2 = 40 Torr, all three treatment groups differ from one another (1-way ANOVA after logarithmic transformation; all P < 0.03); similar results are obtained for responses to Pao2 = 50 Torr but not to 60 Torr (where week 30% O2 = week 60% O2).

Fig. 1.

Increases in phrenic amplitude, burst frequency, and minute phrenic activity during hypoxia in 3- to 5-mo-old hyperoxia-treated and untreated control rats. Values are means ± SE for control (n = 13), week 30% O2 (n = 11), and week 60% O2 (n = 4) rats; data are only presented for rats studied at all levels of hypoxia, thus excluding 1 rat from the week 60% O2 group. Pao2, arterial Po2. *Significant differences between groups (indicated by bracket) at all levels of hypoxia, P ≤ 0.05. See text explanation of groups.

In contrast to 1-wk exposures to 60% O2, 1 day of 60% O2 during early postnatal development did not alter the hypoxic phrenic responses measured at 3-5 mo of age (day 60% O2 vs. controls, all P > 0.05; Fig. 2).

Fig. 2.

Increases in phrenic amplitude, burst frequency, and minute phrenic activity during hypoxia in 3- to 5-mo-old hyperoxia-treated and untreated control rats. Values are means ± SE for control (n = 13) and day 60% O2 (n = 8) rats. Control data are the same as presented in Fig. 1. No significant differences were detected between treatment groups (P > 0.05).

Hypoxic phrenic responses in 6- to 12-mo-old rats. Because hypoxic ventilatory responses were somewhat greater than anticipated in awake week 60% O2 rats (see below), hypoxic phrenic responses were measured in additional groups of anesthetized week 60% O2 rats, ranging from 6 to 12 mo of age. Arterial blood gases for these rats are summarized in Table 1, but these data were not included in the overall statistical analysis because all treatment groups were not studied at these ages. In general, baseline PaCO2 was 3-4 Torr higher in older (6-12 mo) vs. younger (3-5 mo) week 60% O2 rats, reflecting higher CO2 apneic thresholds (PetCO2 = 43 ± 1 Torr vs. 39 ± 1 Torr, respectively; P < 0.001).

Hypoxic phrenic responses tended to be greater in 6- to 12- vs. 3- to 5-mo-old week 60% O2 rats at all three levels of hypoxia tested. To examine the effect of age on the hypoxic phrenic responses of week 60% O2 rats, changes in phrenic burst amplitude, frequency, and minute activity in response to Pao2 = 40 Torr were analyzed vs. postnatal age (Fig. 3). Data for an additional eight 3- to 4-mo-old week 60% O2 rats from an earlier study (1) were included in this analysis to increase the number of rats represented at younger ages; these data were collected by the same investigator using an identical experimental protocol. A positive correlation exists between the hypoxic responses of both amplitude (P < 0.001, R2 = 0.31) and minute activity (P = 0.005, R2 = 0.22) and age; however, no age-dependent changes in hypoxic frequency response were detected (P = 0.472, R2 = 0.02). Thus there was a slow, spontaneous functional recovery in week 60% O2 rats.

Fig. 3.

Phrenic responses to hypoxia (Pao2 = 40 Torr) in week 60% O2 rats as a function of age (n = 34). A significant relationship exists between increases in phrenic amplitude (P < 0.001, R2 = 0.31) and minute phrenic activity (P = 0.005, R2 = 0.22) during hypoxia and age but not between increases in phrenic burst frequency and age (P = 0.472, R2 = 0.02). □, Eight rats from an earlier study (1).

Ventilatory Responses in Awake Rats

e and metabolic rate (V̇co2) were measured in control and week 60% O2 rats while breathing air (21% O2) and two levels of hypoxia (14.5 and 12.5% O2). V̇e was lower in week 60% O2 rats throughout the protocol (treatment effect, P = 0.007; Fig. 4); the difference tended to be greater during hypoxia (i.e., suggesting a blunted V̇e response to hypoxia in week 60% O2 rats), but this did not reach statistical significance (treatment × time, P = 0.076). Differences in hypoxic ventilatory responses are more apparent when V̇e is normalized to metabolic rate (i.e., V̇e/V̇co2) (treatment × time, P = 0.050; Fig. 4). V̇e/V̇co2 was similar between control and week 60% O2 rats while breathing air, but it was significantly lower in week 60% O2 rats breathing 14.5 and 12.5% O2 (P < 0.05 at 4 of 6 and 5 of 6 time points, respectively). Differences in the hypoxic V̇e/V̇co2 response, and a similar trend for the V̇e response, reflect a reduced hypoxic frequency response in week 60% O2 rats (treatment × time, P = 0.038; Fig. 4). There were no significant differences in tidal volume, metabolic rate (V̇co2), or body temperature between control and week 60% O2 rats (treatment and treatment × time, all P > 0.05).

Fig. 4.

Ventilatory and metabolic measurements in unanesthetized rats. Minute ventilation (V̇e), V̇e-to-metabolic CO2 production ratio (V̇e/V̇co2), and respiratory frequency (fR) in 5- to 12-mo-old week 60% O2 (n = 10) and untreated control rats (n = 8). Data points represent the average for the preceding 10 min; periods of hypoxia (14.5 or 12.5% O2) are indicated by the horizontal lines. Values are group means ± SE. *Significantly different from control at the same time point or, for V̇e (indicated by brackets), at all time points (i.e., overall treatment effect), P ≤ 0.05.

There was no evidence for impaired gas exchange during hypoxia in week 60% O2 rats as previously reported in rats exposed to 1 mo of 60% O2. After 1 mo of 60% O2, hyperoxia-treated rats had lower Pao2 at each hypoxic FiO2 and an elevated alveolar-arterial Po2 difference during hypoxia (27). In contrast, Pao2 values for week 60% O2 rats were not significantly different from those measured in control rats (Table 2; P > 0.05), facilitating comparisons of hypoxic ventilatory responses at similar levels of hypoxemia. PaCO2 values were generally lower in week 60% O2 rats (treatment, P = 0.039), but no difference was detected between the groups for the change in PaCO2 during hypoxia (treatment × FiO2, P = 0.164). However, statistical power was low (<0.2) for comparing changes in blood gases during hypoxia between treatment groups, so nonsignificant interactions (i.e., treatment × FiO2) are difficult to interpret.

View this table:
Table 2.

Arterial blood-gas values for unanesthetized control and hyperoxia-treated rats (week 60% O2) breathing air (21% O2) and two levels of hypoxia (14.5 and 12.5% O2)


Rats chronically exposed to 60% O2 for the first month of life exhibit reduced hypoxic ventilatory and phrenic responses as adults (28), and these effects may be permanent (16). Here we demonstrate that shorter and milder hyperoxic exposures during development can also cause long-lasting attenuation of adult hypoxic responses. However, the magnitude of these effects depends on the characteristics (i.e., level and duration) of developmental hyperoxia and on the age at which the rats are studied. Mild hyperoxia (30% O2) during development causes long-lasting reduction of hypoxic phrenic responses, but these effects are small compared with an equal duration of more severe hyperoxia (60% O2). Furthermore, for a given level of hyperoxia (60% O2), the persistence of the respiratory impairment depends on the duration of hyperoxic exposure, with shorter exposures (1 wk) permitting spontaneous recovery not observed after longer exposures (1 mo). An even shorter exposure, 1 day of 60% O2, produced no detectable impairment in adult rats, although it remains possible that there was a transient functional impairment that recovered spontaneously by the time of study.

Plasticity in Hypoxic Responses After Developmental Hyperoxia

Young adult rats exposed to 60% O2 for the first week of life exhibited impaired hypoxic phrenic responses. This confirms our earlier report (1) that 1 wk of 60% O2 attenuates hypoxic phrenic responses in 3- to 4-mo-old rats, similar to the effects of 1 mo of 60% O2 (1, 16, 17, 30), provided that the hyperoxic exposure occurs during a critical developmental period (i.e., postnatal weeks 1 and 2). In the present study, we extended these earlier findings by demonstrating that 1 wk of 60% O2 also impairs hypoxic ventilatory responses in unanesthetized, adult rats. Baseline ventilation was lower in hyperoxia-treated rats, but this difference was eliminated by normalizing V̇e to V̇co2 (Fig. 4). Because impaired V̇e/V̇co2 responses were observed at similar levels of arterial hypoxemia (e.g., Pao2 = 45-50 Torr at 12.5% O2), studies of phrenic and ventilatory responses to hypoxia converge on the conclusion that developmental exposure to 60% O2 for as little as 1 wk has long-lasting effects on the neural control of breathing; alterations in neither pulmonary mechanics nor gas exchange can account for the observed plasticity (Refs. 27, 28, 30; present study).

Given their attenuated V̇e/V̇co2 response to hypoxia, it is surprising that spontaneously breathing, hyperoxia-treated rats did not have higher PaCO2 values than control rats during hypoxia (Table 2). However, a similar observation was made for rats after 1 mo of developmental hyperoxia (27). In both studies, hyperoxia-treated rats had lower PaCO2 values during normoxia than controls, and PaCO2 tended to converge between groups during hypoxia (i.e., the drop in PaCO2 during hypoxia tended to be smaller in hyperoxia-treated rats). This convergence, as well as the relatively small (20-25%) attenuation of the hypoxic ventilatory response, may make it difficult to detect significant hypoventilation of hyperoxia-treated rats during hypoxia. Using the alveolar ventilation equation (20a) and our data for control rats, we predict a 1-2 Torr smaller drop in PaCO2 for a 20% reduction in the hypoxic V̇e/V̇co2 response, assuming steady-state conditions and a constant ratio of dead space-to-tidal volume. Similarly, Mortola and Saiki (34) observed, and predicted mathematically, that a 23% reduction in the V̇e/O2 consumption response to hypoxia (10% O2) resulted in only a 2.8- to 3.0-Torr smaller drop in PaCO2 in rats. In the present study, hyperoxia-treated rats exhibited nonsignificant 3.6- and 1.4-Torr smaller drops in PaCO2 at 14.5% and 12.5% O2, respectively, similar to what might be expected. Therefore, we may have lacked sufficient statistical power to detect the effect of developmental hyperoxia on hypoxic PaCO2 values. Alternatively, because V̇co2 responses were similar between groups, an absence of differences in hypoxic PaCO2 may reflect changes in respiratory dead space (27).

The level of hyperoxia influences the degree of functional impairment after developmental hyperoxia. However, even relatively mild hyperoxic exposures during development produced long-lasting changes in respiratory control. In 3- to 5-mo-old rats born and raised for the first postnatal week in 30% O2, hypoxic phrenic responses were lower than in age-matched control rats. Hanson and colleagues (11, 20) investigated the effects of 30% O2 on the development of hypoxic ventilatory and carotid body responses. In these experiments, chronic hyperoxia (30% O2 from birth to the time of study) blunted hypoxic ventilatory responses in kittens (20) and abolished carotid body responses to hypoxia in kittens and rats (11, 20). Moreover, in a note added in proof, Erickson and colleagues (13) report that 30% O2 was as effective as 60% O2 in producing changes in carotid body morphology associated with blunted hypoxic ventilatory responses. Thus several lines of evidence indicate that 30% O2 impairs the development of hypoxic sensitivity, apparently by mechanisms similar to 60% O2 (see below). Reduced hypoxic phrenic responses in the present study are consistent with these earlier reports but extend them by demonstrating that these early effects persist into adulthood. On the other hand, the effects of 30% O2 are smaller at 3-5 mo of age than those caused by an equivalent exposure to 60% O2. It is unknown whether this difference results from a smaller initial impairment or a more rapid rate of spontaneous recovery.

The duration of the hyperoxic exposure also appears to be an important determinant of adult respiratory function. In contrast to 1 wk of 30-60% O2, a single day of 60% O2 during early postnatal life had no discernible effect on hypoxic phrenic responses measured at 3-5 mo of age. Thus either there is a limit as to how short the hyperoxic exposure can be to elicit functional impairment at 3-5 mo (i.e., between 1 and 7 days) or the specific day we chose preceded the critical developmental window. It was impractical to investigate all (or even most) combinations of levels and durations of hyperoxia. Thus it remains possible that long-lasting plasticity in hypoxic responsiveness can be elicited by short durations of hyperoxia (i.e., <1 wk) if the exposure occurs later in development or for longer than 1 day. Indeed, although the critical period for susceptibility to hyperoxia includes the first and second postnatal weeks (1), the first few postnatal days could be less influential because critical periods may start after birth (3). We avoided direct statistical comparison among rats treated with 1 wk and 1 day of postnatal hyperoxia because the “week” exposures also included 1-4 prenatal days of maternal hyperoxia. Thus differences among these groups could reflect the duration of hyperoxic exposure, the timing of the exposure, or the lack of prenatal exposure in rats exposed for only 24 h. However, because 1 wk of 60% O2 during the second postnatal week only (i.e., no prenatal exposure) is sufficient to blunt adult hypoxic phrenic responses (1), prenatal hyperoxia is not necessary for the effect.

Although 1 wk and 1 mo of 60% O2 appear equally effective at attenuating hypoxic phrenic responses if the responses are measured at 3-5 mo of age (1), this may not be the case if hypoxic responses are measured in older rats. Hypoxic ventilatory responses were reduced in awake, 5- to 12-mo-old rats exposed to 60% O2 for the first postnatal week, but this blunting was smaller than predicted by the hypoxic phrenic responses of 3- to 5-mo-old rats (and by ventilatory responses after 1 mo of developmental hyperoxia; Ref. 27). This discrepancy prompted additional studies of 1 wk 60% O2 hyperoxia-treated rats to determine whether hypoxic phrenic responses spontaneously recovered. There was a gradual increase in hypoxic phrenic responses with age such that many older hyperoxia-treated rats had near-normal hypoxic phrenic responses (>175% increase in minute phrenic activity at Pao2 = 40 Torr; Fig. 3). In contrast, there is no evidence for age-dependent changes in hypoxic phrenic responses between 3 and 15 mo in normoxia-reared (i.e., control) rats or in rats exposed to 60% O2 for the first month of life (Fig. 5; see also Ref. 16). Therefore, shorter exposures to 60% O2 may permit a greater degree of spontaneous recovery than longer exposures. However, even “permanent” impairment of hypoxic responses after 1 mo of 60% O2 can be reversed in adult rats, at least temporarily, by chronic sustained or chronic intermittent hypoxia (17).

Fig. 5.

Phrenic responses to hypoxia (Pao2= 40 Torr) in 3- to 5-mo-old and 14- to 15-mo-old rats exposed to 60% O2 for the first month of life (month 60% O2) or in untreated control rats. Data are compiled from studies using the same experimental protocol (Refs. 1, 16, 17; present study). Values are means ± SE. No significant differences were detected between age groups (P > 0.05).

There is considerable variation in hypoxic phrenic responses among older rats exposed to 60% O2 for the first postnatal week (Fig. 3), an observation that may reflect variation in the degree of spontaneous recovery. Indeed, the extent of recovery in hypoxic phrenic responses after developmental hyperoxia may depend on the extent of the initial impairment or the timing of the exposure relative to a critical developmental period. Hypoxic ventilatory and/or carotid body responses are greatly reduced, if not abolished, immediately after developmental exposure to 30% O2 (11, 20). Although care must be taken when comparing between studies, it is noteworthy that developmental exposure to 30% O2 caused a relatively small decrement in hypoxic phrenic responses at 3-5 mo of age in the present study. These observations provide additional, suggestive evidence for even more rapid spontaneous recovery of hypoxic responses after mild hyperoxic exposures during development. Similarly, our observations do not rule out functional impairment after 1 day of 60% O2 followed by rapid functional recovery. Unfortunately, we do not have the necessary data to evaluate whether variations in hypoxic phrenic responses measured at 3-5 mo of age reflect initial differences in the degree of blunting or different rates of recovery.

Mechanism of Hyperoxia-induced Plasticity

Developmental hyperoxia attenuates adult hypoxic ventilatory and/or phrenic responses by impairing carotid body function (4, 16, 28). Electrical stimulation of the carotid sinus nerve produces an equivalent increase in phrenic motor output in adult rats exposed to 1 mo of 60% O2 during development and age-matched controls (16, 29), indicating that central neural integration of carotid chemoafferent inputs is not diminished. On the other hand, rats treated with 1 mo 60% O2 exhibit substantially reduced carotid body responses to hypoxia, asphyxia, and cyanide between 3 and 15 mo of age (4, 16, 28). Recent work confirms that exposure to 60% O2 for the first postnatal week has long-lasting effects on carotid body responses to hypoxia and related stimuli (4), suggesting similar mechanisms for blunted hypoxic ventilatory and phrenic responses after 1 wk and 1 mo of hyperoxia. Moreover, chronic exposure to mild hyperoxia (30% O2) prevents the normal development of carotid body responses to hypoxia in rats and kittens (11, 20). Although the persistence of this blunting after removal from 30% O2 was not studied, it seems likely that these changes in carotid body function contribute to the apparent reduction of adult hypoxic phrenic responses observed here. Hyperoxia-induced plasticity differs in this respect from plasticity elicited by neonatal hypoxia. Like perinatal hyperoxia, 1 wk of neonatal hypoxia attenuates hypoxic ventilatory responses in adult rats (35), but this blunting does not appear to involve changes in adult carotid body function (2, 12).

It is not yet known how developmental hyperoxia alters carotid body function, although several potential mechanisms have been proposed (see Refs. 8, 13). Developmental hyperoxia (30 and 60% O2) alters multiple aspects of carotid body morphology, causing degeneration of chemoafferent neurons, loss of chemosensory tissue (i.e., glomus cells), and an overall smaller carotid body volume (Refs. 13, 16, 17; Z.-Y. Wang and G. E. Bisgard, unpublished data). Several direct and indirect mechanisms could contribute to these morphological changes, such as a specific, local hyperoxic toxicity, altered blood supply, or changes in activity-dependent gene expression (8, 13). Fewer glomus cells in the carotid body or changes in its chemoafferent innervation could attenuate carotid body responses to hypoxia, which have generally been assessed from whole carotid sinus nerve recordings (4, 11, 16, 28). It is not known whether developmental hyperoxia also causes long-lasting changes in the O2 sensitivity of individual glomus cells, although the lack of hypoxic sensitivity in single-unit carotid body recordings in chronically hyperoxic kittens suggests changes in glomus cell hypoxic chemotransduction, or changes at the synapse, immediately after hyperoxia (20). Moreover, a recent study found reduced hypoxic dopamine release from carotid bodies isolated from adult rats after developmental hyperoxia (37), consistent with reduced hypoxic chemosensitivity. Thus hyperoxia may interfere with the normal postnatal maturation of carotid body O2 sensitivity that overlaps temporally with the critical period for hyperoxia-induced plasticity (i.e, first 2 postnatal weeks; Refs. 1, 8).

Hyperoxia-induced formation of reactive O2 species can have toxic effects on neural and pulmonary tissues (14, 40). Although many tissues are susceptible to hyperoxic toxicity (21), including the carotid body (7, 26, 32), ventilatory control appears normal in rats exposed to 1 mo of 60% O2 as adults (27, 28, 30), and newborns may be less sensitive to hyperoxic toxicity than mature mammals, at least in some tissues (14, 40). The neural effects of developmental hyperoxia (60% O2) seem specific to the carotid body and related chemoafferent pathways (13), with no effect on the hypercapnic ventilatory response (27) or central neural integration of chemoafferent inputs (16, 29). Together, these data suggest that the effects of developmental hyperoxia on respiratory control are not due to widespread, nonspecific cellular toxicity. Importantly, the effects of hyperoxia are dose dependent. In the present study, we observed long-term effects after normobaric exposures to as little as 30% O2 and for durations as short as 1 wk, well below levels of O2 expected to have toxic effects (14, 26, 32). However, even 21% O2 is hyperoxic relative to prenatal conditions, and given its high blood flow relative to metabolic rate, the carotid body may be more susceptible to hyperoxic PaO2 than other tissues (26, 32). Thus we cannot rule out a role for local, development-specific reactive O2 species toxicity in carotid body developmental plasticity.

An alternative explanation, however, is that mild to moderate hyperoxia reduces carotid chemoafferent neuron activation, thereby altering the activity-dependent expression of neurotrophins or neurotransmitters associated with normal postnatal carotid body development (8, 13, 28). Indeed, many neural systems require sensory receptor activity during critical postnatal periods for normal phenotypic development (22).

Significance of Hyperoxia-induced Plasticity

Clinical disorders such as sudden infant death syndrome, congenital central hypoventilation syndrome, or sleep-disordered breathing may involve abnormal development of the respiratory control system (6, 18, 25, 42). It is therefore critical to understand how environmental conditions during early life influence the development of protective ventilatory chemoreflexes. Supplemental O2 is used clinically to treat premature infants or infants with chronic lung disease. Although care is taken to avoid pulmonary, retinal, or central nervous system hyperoxic toxicity during O2 therapy, Katz-Salamon and Lagercrantz (24) identified a relationship between the use of O2 therapy and blunted hypoxic sensitivity in preterm infants. How long these effects persist in humans is currently unknown, but our data indicate that relatively mild hyperoxic exposures (e.g., 1 wk of 30% O2) can cause long-lasting impairment of hypoxic chemoreflexes in rats. Even if this impairment is small or recovers rapidly, infants could be at risk during early life when they are particularly vulnerable to environmental and physiological insults (5, 19). The observation that 1 wk of 30 or 60% O2 has any detectable effect several months later should emphasize that the respiratory control system is highly sensitive to deviations from normoxia during development.


This work was supported by National Heart, Lung, and Blood Institute (NHLBI) Grants HL-53319, HL-65383, and HL-68255. R. W. Bavis was supported by NHLBI Training Grant HL-07654 and National Research Service Award HL-70506.


Present address of R. W. Bavis: Biology Dept., Bates College, Carnegie Science Bldg., 44 Campus Ave., Lewiston, ME 04240 (E-mail: rbavis{at}


  • 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.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 20a.
  22. 21.
  23. 22.
  24. 23.
  25. 24.
  26. 25.
  27. 26.
  28. 27.
  29. 28.
  30. 29.
  31. 30.
  32. 31.
  33. 32.
  34. 33.
  35. 34.
  36. 35.
  37. 36.
  38. 37.
  39. 38.
  40. 39.
  41. 40.
  42. 41.
  43. 42.
View Abstract