## Effects of Perinatal Hyperoxia on Breathing

### Abstract

Air‐breathing animals do not experience hyperoxia (inspired O2 > 21%) in nature, but preterm and full‐term infants often experience hyperoxia/hyperoxemia in clinical settings. This article focuses on the effects of normobaric hyperoxia during the perinatal period on breathing in humans and other mammals, with an emphasis on the neural control of breathing during hyperoxia, after return to normoxia, and in response to subsequent hypoxic and hypercapnic challenges. Acute hyperoxia typically evokes an immediate ventilatory depression that is often, but not always, followed by hyperpnea. The hypoxic ventilatory response (HVR) is enhanced by brief periods of hyperoxia in adult mammals, but the limited data available suggest that this may not be the case for newborns. Chronic exposure to mild‐to‐moderate levels of hyperoxia (e.g., 30–60% O2 for several days to a few weeks) elicits several changes in breathing in nonhuman animals, some of which are unique to perinatal exposures (i.e., developmental plasticity). Examples of this developmental plasticity include hypoventilation after return to normoxia and long‐lasting attenuation of the HVR. Although both peripheral and CNS mechanisms are implicated in hyperoxia‐induced plasticity, it is particularly clear that perinatal hyperoxia affects carotid body development. Some of these effects may be transient (e.g., decreased O2 sensitivity of carotid body glomus cells) while others may be permanent (e.g., carotid body hypoplasia, loss of chemoafferent neurons). Whether the hyperoxic exposures routinely experienced by human infants in clinical settings are sufficient to alter respiratory control development remains an open question and requires further research. © 2020 American Physiological Society. Compr Physiol 10:597‐636, 2020.

Comprehensive Physiology offers downloadable PowerPoint presentations of figures for non-profit, educational use, provided the content is not modified and full credit is given to the author and publication.

 Figure 1. Ventilatory responses of preterm infants, term infants, and adult human subjects to approximately 60 s of 100% O2 (shaded region). Values are expressed as the mean (± SEM) change from baseline for small preterm infants (n = 40), large preterm infants (n = 34), term infants (n = 24), and adults (n = 16). Note that the magnitude of the ventilatory depression during 100% O2 is greater in infants than in adults; *P ≤ 0.05 versus adults. Adapted, with permission, from Al‐Matary A, et al. 2004 4. Figure 2. Ventilatory responses of healthy, preterm infants (≤8 days of age; n = 9) and adult human subjects (n = 10) to 5 min of 100% O2. Values are expressed as the mean (± SEM) change from baseline (21% O2). Although the initial ventilatory depression was greater in the preterm infants, both age groups showed a similar increase in ventilation during minutes 2 to 5 of the 100% O2 exposure. Reprinted by permission from Springer Nature Customer Service Centre GmbH: Springer Nature, Pediatric Research 257. Figure 3. Ventilatory responses to isocapnic hyperoxia (30%, 50%, and 75% O2) in adult men (n = 8) based on the data from 41. Values are expressed as the mean (± SEM) change from baseline (21% O2) during the final 5 min of each hyperoxic exposure (30 min) and each normoxic recovery period (15 min); *P ≤ 0.05 versus baseline. The ventilatory response to poikilocapnic hyperoxia (75% O2) is provided for comparison (dotted line); the increase from baseline during hyperoxia was statistically significant but much smaller than in the isocapnic 75% O2 test. The change in PetCO2 during each hyperoxic challenge is given parenthetically. Reprinted, with permission, from Dean JB, et al. 2004 91. Figure 4. The hypoxic ventilatory response of adult rats after a 10 min pre‐exposure to 100% O2 (open symbols) or 21% O2 (closed symbols). Ventilation data (mean ± SEM) are presented for the baseline period (21% O2; minute 0), during 15 min of hypoxia (10% O2, 5% CO2, balance N2), and after a 5 min recovery in room air. Panel A: Rats were injected with vehicle (1:5 DMSO‐normal saline); panel B: rats were injected with the nNOS inhibitor 7‐nitroindazole (7‐NI; 25 mg/kg). Data are for the same 17 rats studied on two separate days (i.e., first day = vehicle trials, second day = 7‐NI trials). Reprinted, with permission, from Gozal D, 1998 131. Figure 5. The hypoxic ventilatory response to 12% O2 for neonatal rats (P8) exposed for the preceding 23 to 28 h to 60% O2 (Hyperoxia) or 21% O2 (Control). Individuals were injected with saline (n = 12 Hyperoxia, 15 Control) or the antioxidant drug MnTMPyP (n = 10 Hyperoxia, 11 Control). Values are expressed as the mean (± SEM) change from baseline (21% O2) during the early (minute 1) or late (minute 8) phase of the response. Symbols denote significant (P < 0.05) main effects for hyperoxia treatment (*) or drug treatment (†) based on a 2‐way ANOVA. Reprinted from 254, with permission from Elsevier. Figure 6. (A) Normoxic ventilation (21% O2) and (B) hyperoxic ventilation (60% O2) for neonatal rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia). Minute ventilation (mean ± SEM) is presented for rats aged P4 (n = 24 Control and 19 Hyperoxia), P6‐7 (24 and 19), and P13‐14 (21 and 21); the same individuals were studied at both levels of inspired O2. *P < 0.05 versus Control within the same age group; †P < 0.05 versus 4 days of age within the same treatment group. Reprinted from 36, with permission from Elsevier. Figure 7. Spontaneous respiratory motor activity recorded from cervical (C4‐C5) nerve roots of isolated brainstem‐spinal cord preparations from neonatal rats reared in 21% O2 (Control; n = 7) or 60% O2 (Hyperoxia; n = 10). (A) Interburst interval, (B) burst frequency, and (C) coefficient of variation for interburst intervals (mean ± SEM) were determined under baseline conditions (superfused with 95% O2/5% CO2, pH 7.4, 26°C). Data are reported for the entire 70‐min protocol in 5‐min bins, but data from the first 35 min (while the preparation was stabilizing) were excluded from the statistical analysis. *P < 0.05 versus Control at the same time point; † denotes a significant main effect for developmental treatment (i.e., Control vs. Hyperoxia, P < 0.05). Reprinted from 36, with permission from Elsevier. Figure 8. The hypoxic ventilatory response to 12% O2 for neonatal rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia) until studied at (A) P4, (B) P6‐7, or (C) P13‐14. Responses are presented as mass‐specific ventilation (left panels) and as the percentage increase in ventilation from baseline (BL; 21% O2) (right panels). Values are means ± SEM. Sample sizes (P4, P6‐7, and P13‐14) are 17, 16 and 15 for Control and 15, 15 and 17 for Hyperoxia. *P < 0.05 versus Control within the same minute of hypoxia. #P < 0.05 versus minute 1 of hypoxia within the same treatment group. Reprinted, with permission, from Bavis RW, et al. 2010 38. Figure 9. Ventilatory and metabolic responses to hypoxia for adult rats previously exposed to 21% O2 (Control, •) or to 60% O2 (Hyperoxia, ▪ or ▴) for 1 month (A) during perinatal development or (B) as adults. Values are means ± SEM for minute ventilation ($V˙E$), metabolic CO2 production ($V˙CO2$), and ventilation‐to‐metabolism ratio ($V˙E/V˙CO2$) measured during 1 h of normoxia and 1 h of hypoxia. In panel A, inspired O2 was adjusted so that PaO2 during hypoxia was approximately 48 mmHg in Control and Hyperoxia rats. In panel B, both treatment groups received the same inspired O2 mixture (∼11% O2) during hypoxia. Sample sizes are 38 Control, 27 Hyperoxia during the perinatal period, and 12 Hyperoxia as adults. Reprinted with permission from 189, © 1996 The Physiological Society. Figure 10. Phrenic nerve responses to isocapnic hypoxia for adult rats previously exposed to 21% O2 (white bars; Control, C) or to 60% O2 for 1 month during perinatal development (gray bars; P) or as adults (black bars; A). Hypoxic responses were studied at two levels of hypoxia, PaO2 = 40 and 50 mmHg. Values are expressed as the mean (± SEM) change from baseline for the peak amplitude of the phrenic activity (∫Phr), phrenic burst frequency (f), and their product, minute phrenic activity (∫Phr × f). Sample sizes are 7 Control, 11 Hyperoxia during the perinatal period, and 4 Hyperoxia as adults. *P < 0.05 versus Control at the same PaO2. Reprinted from 191, with permission from Elsevier. Figure 11. Single‐unit carotid chemoafferent nerve responses to (A) 12% O2 and (B) 0% O2 in rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia) through P7 (shaded region). Rats were studied at various ages during the hyperoxic exposure (P1‐P7) or after 3 days or 7 to 8 days recovery in room air (i.e., P10 or P14‐15). Values are mean ± SEM for the peak response (3‐s moving average) recorded during the hypoxic challenge. Sample sizes are 6 to 19 per group. *P < 0.05 versus Control at the same age. For comparison, the triangles represent single‐unit chemoafferent responses recorded from P14 rats maintained in 60% O2 from birth (i.e., no normoxic recovery) in Ref. 103. Reprinted from 27, with permission from Elsevier. Figure 12. Membrane depolarization responses to (A) high extracellular K+ concentrations (1–2 min of 20 mM KCl) and (B) hypoxia (2–3 min of 0% O2) for glomus cells isolated from neonatal rat carotid bodies after 1 or 2 weeks in 21% O2 (Norm) or 60% O2 (Hyper) beginning at P1 or P8. Horizontal bars next to the symbol key summarize the oxygen exposures across ages for each treatment group: white = 21% O2 and gray = 60% O2. Values are mean ± SEM. Statistical comparisons: *P < 0.05, **P < 0.01, and ***P < 0.001 between groups connected by lines or brackets; ‡P < 0.01 between Hyper 1–8 and the normoxic control group at P8; NS, nonsignificant. Reprinted from 171, with permission from Elsevier. Figure 13. Whole‐nerve carotid sinus nerve (CSN) response to isocapnic hypoxia (PaO2 ∼45 mmHg) in adult rats previously exposed to 21% O2 (Control) or to 60% O2 for 1, 2, or 4 weeks during perinatal development. Values are mean ± SEM. Sample sizes are 14 for Control and 8 for each of the other groups. *P < 0.05 versus Control. No significant differences were detected between 1, 2, and 4 weeks of 60% O2 for the hypoxic response, although longer hyperoxic exposures tended to cause greater reductions in the CSN response to cyanide and asphyxia (data not shown; see Ref. 48). Reprinted, with permission, from Bisgard GE, et al. 2003 48. Figure 14. Carotid body growth curve for neonatal rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia). Carotid body volume (mean ± SEM) was estimated from the cumulative carotid body area × thickness of serial sections. Sample sizes are 6 per treatment group per age. *P < 0.05 versus Control at the same age. Reprinted from 97, with permission from Elsevier. Figure 15. (A) Photomicrographs showing BrdU‐positive cells in the carotid bodies of P4 rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia). Green shows staining for tyrosine hydroxylase (TH), a marker for glomus cells, and orange shows staining for BrdU, a marker for recently divided cells. Closed arrows = BrdU‐positive glomus cells; open arrows = BrdU‐positive nonglomus cells. Scale bar = 50 μm. (B) The ratio of BrdU‐positive glomus cells to the total number of glomus cells (mean ± SEM) in Control and Hyperoxia rats at P2, P4, and P6; BrdU was administered 24 h prior to tissue collection. Sample sizes are 6 per treatment group per age. *P < 0.05 versus Control at the same age. Redrawn from 97, with permission from Elsevier. Figure 16. Representative photomicrographs and cell counts of petrosal ganglion neurons retrogradely labeled by applying rhodamine dextran crystals to the carotid body of neonatal rats reared 21% O2 (Control) or 60% O2 (Hyperoxia) for 7 days followed by 2 to 3 days in room air (A,B), 14 days followed by 4 to 5 days in room air (C,D), or 14 days followed by 8 to 9 days in room air (E,F). Box and whisker plots (panels B, D, and F) report the number of labeled chemoafferent neurons observed in the petrosal ganglion (upper and lower limits of the box mark the 75th and 25th percentiles, and the thick horizontal line marks the median). Sample sizes are 7 to 10 per group. *P < 0.05 versus Control. Reprinted, with permission, from Chavez‐Valdez R, et al. 2012 72. Figure 17. Critical period for developmental plasticity of the hypoxic ventilatory response in neonatal rats. (A) Rats were exposed to 7 days of 60% O2 (black bars) during the first, second, third, or fourth postnatal week, or exposed to 60% O2 during all 4 weeks. All rats were then returned to room air (21% O2) until studied as adults. (B) The change in minute phrenic activity (∫Phr × f) in response to isocapnic hypoxia (PaO2 = 60 mmHg) for adult rats from each of the treatment groups (n = 8 per group). Responses (mean ± SEM) are expressed relative to an age‐matched control group (n = 8) reared continuously in room air, where the control group's average response would be 100% (dotted line). *P < 0.05 versus Control. Data taken from 32. Reprinted from 29, with permission from Elsevier. Figure 18. Ventilatory depression during the late phase of the hypoxic ventilatory response to 12% O2 for P4‐5 rats reared in 21% O2 (Control) or 60% O2 (Hyperoxia) treated with saline, MK‐801 (NMDA glutamate receptor antagonist; 1 mg/kg, i.p.), nNOS inhibitor I (nNOS inhibitor; 1 mg/kg, i.p.), or imatinib (PDGF‐β receptor antagonist; 100 mg/kg, i.p.). Values (means ± SEM) are expressed as the change in ventilation between the 1st and 15th minutes of hypoxia; sample sizes are 15 to 22 per group. *P < 0.05 versus Control within the same drug group; †P < 0.05 versus saline within the same treatment group. Reprinted from 24, with permission from Elsevier. Figure 19. Hypoxic ventilatory responses to 12% O2 for P13‐15 rats reared in 21% O2 (Control) or reared in intermittent hyperoxia and/or intermittent hypercapnic hyperoxia. In panel A, rats were exposed from birth to intermittent 30% O2 (IH30; n = 16) or intermittent 60% O2 (IH60; n = 18) at a rate of 5 cycles/h (4.5 min at 30% or 60% O2, 7.5 min at 21% O2; 24 h/d). In panel B, rats were exposed from birth to intermittent hypercapnic hypoxia (IHx; n = 19) or intermittent hyperoxia + intermittent hypercapnic hypoxia (IHxH; n = 23) at a rate of 5 cycles/h (2.4 min at 10% O2/6% CO2 followed by either 9.6 min at 21% O2 or by 4.5 min at 30% O2 and 5.1 min at 21% O2; 24 h/d). Sample sizes for Control groups were 16 and 24, respectively. Responses are presented as mass‐specific ventilation (left panels) and as the percentage increase in ventilation from baseline (BL; 21% O2) (right panels). Values are means ± SEM. †P < 0.05 versus Control at the same time point. Panel A adapted from 196 and panel B adapted from 30, with permission from Elsevier.

Teaching Material

Ryan W. Bavis. Effects of Perinatal Hyperoxia on Breathing. Compr Physiol 10 : 2020, 597-636

Didactic Synopsis

Major teaching points:

1) Hyperoxia affects the control of breathing across multiple time domains.

a) Acute responses are characterized by immediate changes in breathing that end once the stimulus is removed.

b) Brief and chronic exposures to hyperoxia may elicit respiratory plasticity (a change in the control of breathing that persists after return to normoxia).

2) The acute ventilatory response to hyperoxia is often biphasic in newborns and adults: initial ventilatory depression (due to reduced carotid body activity) followed by hyperpnea (attributed to CNS excitation).

3) Chronic hyperoxia elicits plasticity that is unique to exposures during critical periods of perinatal development.

a) Neonatal rats hypoventilate when returned to normoxia due to decreased peripheral and central respiratory drive.

b) The hypoxic ventilatory response is attenuated (perhaps permanently) due to abnormal carotid body development.

4) Preterm infants experience hyperoxia in hospitals, but it is not known whether these exposures elicit developmental plasticity.

Didactic Legends

The following legends to the figures that appear throughout the article are written to be useful for teaching.

Fig. 1. This figure illustrates that ventilation decreases soon after subjects begin breathing 100% O2 and that the magnitude of this ventilatory response diminishes with advancing age / maturity. The magnitude of this ventilatory depression is often used as an index for the contribution of peripheral arterial chemoreceptors to baseline (or eupneic) ventilation, but it has also been used to test the O2 sensitivity of the peripheral chemoreceptors. This figure also shows that, after the initial ventilatory depression, ventilation rapidly returns toward baseline levels despite continued exposure to 100% O2 (and may eventually exceed baseline; see Fig. 2).

Fig. 2. This figure illustrates that after the initial ventilatory depression caused by 100% O2, there is often a secondary increase in ventilation during sustained (≥ 1-2 min) exposure to hyperoxia (i.e., hyperoxic hyperpnea); the magnitude of the hyperoxic hyperpnea is often small when measured under poikilocapnic conditions. Although limited data are available to compare infants and adults using standardized techniques, the data illustrated here suggest that hyperoxic hyperpnea might be similar in magnitude and time course across ages.

Fig. 3. This figure illustrates that hyperoxic hyperpnea is more pronounced when the level of inspired O2 increases (i.e., the increase in ventilation during 75% O2 is more than 5? greater than during 30% O2 (i.e., 21% increase vs. 115% increase). Under poikilocapnic conditions, the increase in ventilation causes excessive removal of CO2 from the body as shown by a decrease in PetCO2. The resulting hypocapnia (and respiratory alkalosis) inhibits the ventilatory increase, resulting in a smaller overall ventilatory response to poikilocapnic hyperoxia than observed in isocapnic hyperoxia (in which PetCO2, and thus PaCO2, is regulated by the investigator so that it remains close to baseline). This figure also shows that ventilation remains above baseline during the 15 min recovery period. This "long-lasting facilitation" of breathing suggests that hyperoxia induces an excitatory drive to breathe that persists for a period of time after the return to normoxia; this fits the definition of respiratory plasticity (29, 208).

Fig. 4. This figure illustrates that brief periods of hyperoxia (in this case, 10 min of 100% O2) will augment the HVR of adult rats. In the vehicle-injected rats (panel A), this potentiation is evident as a greater increase in ventilation after about the fifth minute of the hypoxic challenge and is sustained through the end of the exposure (i.e., through minute 15). When pre-treated with a drug that inhibits the synthesis of nitric oxide (NO), however, hyperoxia no longer potentiates the HVR (panel B). This indicates that NO is a critical component of the mechanism underlying this plasticity. Note that the author added CO2 to the inspired air during the hypoxic test to approximate isocapnic conditions in this study, but isocapnic conditions are not necessary to detect potentiation of the HVR (e.g., (61, 239, 240)).

Fig. 5. This figure illustrates that prior exposure to moderate hyperoxia (60% O2) potentiates the HVR of neonatal rats. Although this is similar to the effect shown in Figure 4 for adult rats, the data shown here are for neonatal rats pre-exposed to hyperoxia for 23 – 28 h (vs. only 10 min in the adults); no such potentiation was observed after a 30 min pre-exposure (see (254)). It is not known whether the need for a longer hyperoxic exposure reflects the age of the rats (neonate vs. adult) or the different level of hyperoxia used (60% O2 vs. 100% O2). The magnitudes of the early and late phases of the HVR were fairly similar at P8, consistent with the observation that neonatal rats are transitioning from a biphasic HVR to a more sustained HVR by this age. Although the antioxidant drug did not prevent the potentiation of the HVR, the drug tended to diminish the late HVR in both the Hyperoxia and Control rats. This may indicate a role for ROS in modulating the HVR in neonatal rats independent of the hyperoxic pre-exposure.

Fig. 6. This figure illustrates that rats reared in moderate hyperoxia (60% O2) from have lower minute ventilation than rats reared in normoxia while breathing 60% O2 or when acutely returned to 21% O2. This difference is most pronounced at younger ages. Since neither metabolic rate nor blood gases were measured at the same time as ventilation, these data are insufficient to determine whether the hypopnea exhibited by hyperoxia-reared rats translates into a true hypoventilation (i.e., which would be marked by lower ventilation-to-metabolism ratio and/or increased PaCO2). However, subsequent studies demonstrated that the hyperoxia-reared rats do hypoventilate when returned to normoxia.

Fig. 7. This figure illustrates the effects of perinatal hyperoxia on the rhythmic respiratory motor output of isolated brainstem-spinal cord preparations. Although this in vitro preparation has limitations (e.g., below body temperature, superfused vs. perfused resulting in O2 gradients within the tissue), it is a good model for isolating central mechanisms from peripheral mechanisms. In this case, the slower, more variable bursting frequency (i.e., fictive breathing) resembles the slower, more variable breathing frequency observed in vivo after perinatal hyperoxia. This suggests that chronic hyperoxia induces plasticity in the neural networks underlying respiratory rhythm generation.

Fig. 8. This figure illustrates how the effects of chronic perinatal hyperoxia on the hypoxic ventilatory response (HVR) vary depending on the age of the neonatal rat (and/or duration of the hyperoxic exposure). Control rats (reared in normoxia) exhibit the normal pattern of postnatal maturation of the biphasic HVR: the HVR is biphasic at P4 (increase to a peak at minute 1 followed by a gradual decline toward baseline), less biphasic at P6-7, and no longer biphasic at P13-14. In addition, the magnitude of the initial increase in ventilation increases with postnatal age in the control rats. In contrast, hyperoxia-reared rats already exhibit a sustained increase in ventilation at P4 (i.e., their HVR is not biphasic). As a result, when expressed as a percentage increase from baseline, the late-phase of the HVR is augmented in hyperoxia-reared rats compared to control rats at P4. With advancing age (and/or longer durations of hyperoxia), however, the early phase of the HVR becomes progressively smaller in hyperoxia-reared rats. As a result, the HVR is blunted by P13-14 in hyperoxia-reared rats. This figure also illustrates that hyperoxia-reared rats have lower normoxic ventilation (i.e., compare values at BL in the left panels), particularly at P4 and P6-7. As a result, hyperoxia-reared rats have a lower absolute ventilation throughout the entire hypoxic challenge compared to age-matched controls in all three age groups (left panels).

Fig. 9. This figure illustrates that chronic hyperoxia causes long-lasting (for at least several months) blunting of the hypoxic ventilatory response, but only if the hyperoxic exposure occurs during perinatal development. The ventilatory response is expressed both in units of minute ventilation (V̇e) and ventilation-to-metabolism ratio (V̇e/co2); the latter accounts for the fact that metabolic rate influences breathing and may itself be influenced by acute hypoxia. For rats exposed to hyperoxia during perinatal development, both V̇e and V̇e/co2 are significantly lower than in control rats throughout the hypoxic challenge. For rats exposed to hyperoxia as adults, however, V̇e was similar to that of control rats at all time points and V̇e/co2 tended to be greater due to a more pronounced drop in V̇co2 (i.e., hypoxic hypometabolism). So, if anything, chronic hyperoxia during adulthood augments the HVR, which is opposite the effect of perinatal hyperoxia.

Fig. 10. This figure illustrates that chronic hyperoxia causes long-lasting (for at least several months) blunting of the phrenic nerve response to hypoxia, but only if the hyperoxic exposure occurs during perinatal development. This figure complements and expands on the data presented in Figure 9. Phrenic nerve activity was measured in artificially ventilated, urethane-anesthetized, paralyzed, and vagotomized rats. By monitoring respiratory motor output directly from the phrenic nerve (which innervates the diaphragm), this preparation allowed the investigator to assess the neural control of breathing independent of potential effects of hyperoxia on the neuromuscular junction, respiratory muscles, or respiratory mechanics. Moreover, this preparation allows the experimenter to carefully regulate blood gases to ensure equivalent PaO2 among treatment groups (to control for gas exchange impairment) and to maintain isocapnia (to control for potential changes in CO2 chemosensitivity). This figure also shows that the effects of perinatal hyperoxia on the hypoxic response are more prominent when tested at milder levels of hypoxia: Δ ?Phr, Δ f, and Δ (?Phr ? f) are all significantly reduced at PaO2 = 50 mmHg, but only Δ (?Phr ? f) is reduced at PaO2 = 40 mmHg.

Fig. 11. This figure illustrates that chronic perinatal hyperoxia diminishes the carotid body response to hypoxia in as little as four days. Single-unit recordings provide greater insight into O2 sensitivity of individual chemoreceptor cells than is possible using whole-nerve recordings (in which increases in activity could reflect recruitment of additional cells in addition to increased activity of neurons measured during baseline) and is not directly affected by chemoafferent neuron degeneration (i.e., fewer axons present in the carotid sinus nerve; see (72, 111)). Hypoxic responsiveness increased with age in both treatment groups over the first few days (consistent with postnatal resetting of glomus cell O2 sensitivity), but hypoxic responsiveness began to diminish in hyperoxia-reared rats between P3 and P4. After return to room air (at P7), however, hypoxic responsiveness rapidly recovered to levels similar to that of control rats. Note that this recovery is caused by the return to room air (and not simply spontaneous recovery with advancing age) since rats maintained in hyperoxia for the full 14 days are still relatively unresponsive to hypoxia (shown by the triangles).

Fig. 12. This figure illustrates that chronic perinatal hyperoxia diminishes glomus cell responses to hypoxia in as little as one week. Glomus cell membranes normally depolarize in response to hypoxia (ultimately leading to an increase in intracellular Ca2+ concentrations and neurotransmitter release), but glomus cells from rats reared in 60% O2 for one week depolarized considerably less than glomus cells from age-matched controls; the effect of chronic hyperoxia was similar whether the exposure began at P1 or P8. Glomus cells also depolarize when extracellular K+ is increases (i.e., K+ accumulates within the cell). This response was also slightly diminished after two weeks of hyperoxia, but it was not affected by one week of hyperoxia. This indicates that the blunted depolarization response observed after one week of 60% O2 is specific to the O2 transduction mechanism rather than a general loss in membrane excitability.

Fig. 13. This figure illustrates that chronic hyperoxia causes long-lasting (for at least several months) blunting of the carotid body response to hypoxia; similar data in 14-15 month old rats suggest that the blunting may be permanent (118). Unlike Figure 11, these data represent whole-nerve recordings rather than single-unit recordings. Therefore, the smaller increases in activity may reflect (at least in part) the presence of fewer chemoafferent axons in the CSN: if fewer axons increase their activity, the cumulative (population-level) increase in activity will be less. Regardless of the underlying mechanism (decreased sensitivity of individual chemoreceptor cells vs. fewer chemoreceptor cells / chemoafferent axons), a smaller increase in chemoafferent activity traveling to the brainstem is expected to result in a smaller hypoxic ventilatory response. This figure also shows that the effects of 1, 2, and 4 weeks of hyperoxia on the CSN response to hypoxia are similar in magnitude. Although there was a trend for two weeks to blunt the response more than one week, the average responses were quite similar between 2 weeks and 4 weeks of hyperoxia. This is consistent with the hypothesis that a critical developmental period closes after the second postnatal week (i.e., additional hyperoxia has no further impact on carotid body development).

Fig. 14. This figure illustrates that chronic perinatal hyperoxia diminishes the carotid body size (represented here by volume) in as little as four days. Carotid bodies initially decrease in size after birth, potentially reflecting changes to blood vessel diameters with increasing postnatal PO2 (97). In control rats, carotid bodies begin increasing in size between P4 and P5 due to high rates of cell division (see Figure 15) and continue to increase in size through P14 and beyond. In contrast, the carotid body exhibits little postnatal growth in hyperoxia-reared rats, ultimately resulting in smaller carotid body size by P4. The smaller carotid body size means fewer glomus cells are available to sense O2 and may also lead to chemoafferent neuron degeneration, both of which likely contribute to blunted hypoxic ventilatory responses after perinatal hyperoxia.

Fig. 15. This figure illustrates that chronic perinatal hyperoxia slows glomus cell proliferation, particularly in younger rats (P2 and P4 vs. P6). BrdU is stably incorporated into the DNA of dividing cells and serves as a marker for cell proliferation. When combined with a specidic marker for glomus cells (TH), it was possible to show that a smaller proportion of the glomus cells were undergoing cell division during chronic hyperoxia in each 24 h period studied. Hyperoxia also tended to decrease the proliferation of non-glomus cells within the carotid body to a comparable extent (see (97)). The decreased rate of cell division (and resulting hypoplasia) during perinatal hyperoxia likely contributes to slower carotid body growth illustrated in Figure 14. Moreover, hyperoxia may not decrease carotid body size in older rats (i.e., when hyperoxia initiated after the neonatal period) since cell proliferation is already reduced in healthy adults (26).

Fig. 16. This figure illustrates that chronic perinatal hyperoxia decreases the number of carotid chemoafferent neurons, likely reflecting chemoafferent neuron degeneration. The petrosal ganglion contains the soma of afferent neurons that project to the carotid body and synapse with glomus cells, but not all neurons within the petrosal ganglion (or the carotid sinus nerve) are chemoafferent neurons. To identify the relevant population of neurons, the authors back-labeled neurons from the carotid body and then counted the number of labeled cells. The number of putative chemoafferent neurons was decreased after 14 d of hyperoxia. Moreover, the number of cells remained lower after more than a week of recovery in room air, indicating that petrosal ganglion cells may not be able to proliferate to compensate for this cell loss. Chemoafferent neuron degeneration likely contributes to the decreased whole-nerve CSN responses illustrated in Figure 13.

Fig. 17. This figure illustrates the experimental design used to identify the first two postnatal weeks as the critical period during which chronic hyperoxia elicits long-lasting plasticity in the hypoxic ventilatory response in rats. These data show that one week of 60% O2 will not blunt the hypoxic response if initiated after the second postnatal week of life. This experimental design did not test whether the critical period extends into the prenatal period, but prenatal exposure to hyperoxia is not necessary for this plasticity since exposures during the second postnatal week alone diminished the hypoxic response as much as exposures during all four weeks (which were initiated prior to birth).

Fig. 18. This figure illustrates that chronic perinatal hyperoxia promotes a "sustained" hypoxic ventilatory response (HVR) in neonatal rats in part by influencing glutamate and PDGF signaling pathways. In neonatal rats, the postnatal transition from a biphasic HVR to a sustained HVR is thought to be governed by an age-dependent decrease in PDGF-? receptor expression and age-dependent increases in NMDA glutamate receptor expression and nNOS expression in the brainstem. At P4-5, before this transition, control rats treated only with saline exhibited a strongly biphasic HVR (as shown by a large decrease in ventilation between the 1st and 15th minute of hypoxia) while hyperoxia-reared rats exhibited a sustained HVR (as shown by no change in ventilation between the 1st and 15th minute of hypoxia). When NMDA glutamate receptors were blocked, hyperoxia-reared rats began to express a biphasic HVR, suggesting that premature strengthening of glutamatergic pathways contributed to their sustained HVR. When PDGF-? receptors were blocked, control rats exhibited a less biphasic HVR while hyperoxia-reared rats were unaffected; this suggests that PDGF signaling pathways were prematurely downregulated by perinatal hyperoxia. NO signaling, on the other hand, was not affected by perinatal hyperoxia. Collectively, these data suggest that similar mechanisms underlie hyperoxia-induced plasticity and normal maturation of the biphasic HVR. In other words, hyperoxia may cause the HVR to mature at an earlier age (i.e., physiological heterokairy).

Fig. 19. This figure illustrates the effects of a two-week exposure to intermittent hyperoxia on the hypoxic ventilatory response (HVR) of neonatal rats. In clinical settings, inspired O2 is adjusted to keep blood oxygen levels within a target range and to prevent prolonged periods of hyperoxygenation. Accordingly, intermittent hyperoxia is a more common experience than sustained hyperoxia in neonatal intensive care units. When presented alone, intermittent hyperoxia blunted the HVR of neonatal rats. But when intermittent 30% O2 alternated with intermittent hypercapnic hypoxia (to mimic periodic apneas common in preterm infants), the HVR was unaffected. The durations and frequencies of bouts of hyperoxia and hypercapnic hypoxia were selected based on published data for human infants, but they do not reflect the variability of O2 exposures experienced in clinical settings. Instead, these data demonstrate a range of possible outcomes for human infants exposed to perinatal hyperoxia.

### Related Articles:

Control of Breathing Activity in the Fetus and Newborn
Hyperbaric Environment: Oxygen and Cellular Damage versus Protection
Peripheral Chemoreceptors and Their Sensory Neurons in Chronic States of Hypo‐ and Hyperoxygenation
Peripheral Chemoreceptors: Function and Plasticity of the Carotid Body
Teaching Material