Heterozygous mutations of the transcription factor PHOX2B have been found in most patients with central congenital hypoventilation syndrome, a rare disease characterized by sleep-related hypoventilation and impaired chemosensitivity to sustained hypercapnia and sustained hypoxia. PHOX2B is a master regulator of autonomic reflex pathways, including peripheral chemosensitive pathways. In the present study, we used hyperoxic tests to assess the strength of the peripheral chemoreceptor tonic drive in Phox2b+/− newborn mice. We exposed 69 wild-type and 67 mutant mice to two hyperoxic tests (12-min air followed by 3-min 100% O2) 2 days after birth. Breathing variables were measured noninvasively using whole body flow plethysmography. The initial minute ventilation decrease was larger in mutant pups than in wild-type pups: −37% (SD 13) and −25% (SD 18), respectively, P < 0.0001. Furthermore, minute ventilation remained depressed throughout O2 exposure in mutants, possibly because of their previously reported impaired CO2 chemosensitivity, whereas it returned rapidly to the normoxic level in wild-type pups. Hyperoxia considerably increased total apnea duration in mutant compared with wild-type pups (P = 0.0001). A complementary experiment established that body temperature was not influenced by hyperoxia in either genotype group and, therefore, did not account for genotype-related differences in the hyperoxic ventilatory response. Thus partial loss of Phox2b function by heterozygosity did not diminish the tonic drive from peripheral chemoreceptors.
- control of breathing
central congenital hypoventilation syndrome (CCHS or Ondine’s curse) is a rare disease, generally present from birth and characterized by hypoventilation during sleep in the absence of primary neuromuscular or lung disease or of brain stem lesions (3). Throughout life, patients with CCHS have absent or markedly reduced ventilatory responses to sustained hypercapnia (29) and, to a lesser extent, to sustained hypoxia (29). These respiratory impairments are generally ascribed to impaired central integration of chemosensory inputs at the brain stem level, rather than to failure of chemoreceptor activity, which is at least partially present (15, 22, 37). The genetic basis for CCHS was discovered recently: most patients with CCHS have a heterozygous mutation of the PHOX2B gene (2, 23, 41, 43). PHOX2B is a master regulator of the noradrenergic phenotype and of all neuronal relays of autonomic medullary reflex pathways (30), including peripheral chemosensitive pathways (9).
Mice with a single functional Phox2b allele (i.e., Phox2b+/− mice) provide a unique opportunity to investigate the genotype-phenotype relationship in CCHS. Our laboratory’s previous studies showed that Phox2b+/− mice had longer sleep apnea times than their wild-type littermates on postnatal day (P) 5 (P5) (11) and a weaker response to CO2 on P2, followed by normalization before P10 (9). These ventilatory impairments were reminiscent of CCHS. However, the hyperpneic response to sustained hypoxia (5% O2) on P2 was normal, which is not the case with CCHS (9).
The aim of the present study was to further examine the functional deficits caused by loss of a single Phox2b allele. To this end, we investigated the ventilatory response to hyperoxia, which provides a functional estimate of the tonic chemoreceptor drive (19, 28, 40). Previous studies in newborn mammals showed divergences between the results of hypoxic and hyperoxic tests (5, 25), raising the possibility that a decrease in tonic chemoreceptor drive might coexist with normal ventilatory responses to hypoxia. The present study tested this possibility. As previously, the mice were studied soon after birth (on P2), to avoid possible confounding effects of recovery processes (9).
The generation and genotyping of Phox2b mutant mice have been reported elsewhere (31). Pups were tested at 2 days of postnatal age (P2). Mating 23 wild-type females with mutant males yielded 67 heterozygous mutant pups (weight: 1.73 ± 0.30 g) and 69 wild-type littermates (1.77 ± 0.34 g) that served as controls. In a complementary experiment investigating the effects of hyperoxia on body temperature, we used 10 mutant and 11 wild-type pups (weight: 1.52 ± 0.18 and 1.68 ± 0.12 g, respectively, P < 0.028) born to four wild-type females. The mice were housed at 24°C with a 12:12-h light-dark cycle. The experimental protocols were approved by the Institutional review committee, meet the INSERM guidelines, and were carried out in accordance with the Guide for the Care and Use of Laboratory Animals as adopted and promulgated by the U.S. National Institutes of Health.
Whole body flow plethysmography.
Respiratory variables were measured noninvasively using whole body flow barometric plethysmography, as previously described (11, 12, 24). The plethysmograph was composed of two Plexiglas cylinders serving as the animal (40 ml) and reference (70 ml) chambers, respectively, immersed in a thermoregulated water bath set that maintained their temperature at 32.8°C. A 100 ml/min flow of dry air (Bronkhorst Hi-Tec airflow stabilizer, Urlo, Holland) was divided into two 50 ml/min flows through the chambers, to avoid CO2 accumulation and water vapor condensation. The differential pressure between the two chambers (DRUCK-EFFA transducer, Asnières, France; range ±0.1 millibar) was filtered (bandwidth, 0.05–15 Hz at −3 dB), converted to a digital signal (Instrunet model 200, 14-bit converter, GW-Instruments, Somerville, MA) at a sample rate of 100 Hz, and processed using custom-written software (Software Superscope II, GW-Instruments). The time constant of the pressure decay within the system (2 s) was measured by injecting 2 μl into the measurement chamber. This allowed measurement of breathing frequencies within the 0.5- to 10-Hz range at −3 dB. Calibration was done before each session using a built-in pump incorporating a microsyringe (Ito, Fuji, Japan), which injected a sinusoidal airflow with a maximal amplitude of 2 μl and a frequency of 8 Hz into the animal chamber. Previous tests indicated that calibration coefficients did not vary as a function of pump frequency within the 2- to 8-Hz range. The pressure rise induced by this injection was of similar magnitude to that induced by the pup’s breathing. Considering the limitations of flow barometric plethysmography (39), especially in newborn mice (24), the absolute values of tidal volume (Vt) and minute ventilation (V̇e) presented here should be considered indicative only, whereas breath duration (Tt) absolute values and apnea duration are reliable.
After at least 2-min adaptation to the chamber, each pup was exposed to two consecutive poikilocapnic hyperoxic tests (test 1 and test 2), consisting of 12-min air followed by 3-min O2. To increase statistical power, we performed two tests. We performed poikilocapnic tests because available techniques fail to control normocapnia in freely moving newborn mice. Hyperoxia was achieved by switching the airflow through the plethysmograph to 100% O2 flow at the same flow rate (50 ml/min per chamber). Thus the time needed to flush the chamber (40 ml) was ∼1 min. After the second O2 exposure, the flow was switched back to air, and breathing was measured for 12 min (total duration of the session: 42 min). In a complementary experiment, we used the same design in restrained pups subjected to continuous body temperature measurement (see measurement technique described below).
Ventilatory response to O2.
Breathing variables were analyzed without previous knowledge of genotypes. Tt (s), Vt (μl/g), and V̇e (calculated as Vt/Tt and expressed in μl·g−1·s−1) were calculated on apnea-free periods (see apnea determination described below). Breathing variables were averaged over consecutive 30-s periods. The baseline levels for these variables for each test were calculated as the mean value over the 3-min of air breathing preceding the test. We expressed the O2-induced V̇e decrease as the percentage of baseline, using the formula 100 × (minimum V̇e − baseline V̇e)/baseline V̇e, where minimum V̇e was the value over the 3-min period of O2 exposure. This method took into account possible interindividual differences in time to the V̇e response to O2. We determined the Vt and Tt responses to O2 using the same formula with the Vt and Tt values measured at minimum V̇e.
Apneas were determined using an automatic classification method based on spectral analysis (24). Apneas were defined as ventilatory pauses longer than twice the duration of the preceding breath (32, 36). This definition takes into account the large interindividual differences in resting breathing frequency in newborn mice. Total apnea duration was calculated during each 3-min period, i.e., before and during O2 exposure. The apnea response to O2 was calculated as the difference between air and O2 (percentages could not be calculated because some subjects had no apneas during normoxia, so that baseline apnea duration was zero). Movements were detected based on changes in the baseline respiratory signal, using a previously validated criterion: [(Vi − Ve)/(Vi + Ve)], where Vi and Ve were the magnitudes of the inspiratory and expiratory limbs of the volume signal, respectively (24).
In the main experiment, body temperature was not continuously recorded during ventilatory measurements, as this would have required restraining the pups; instead, body temperature was measured immediately after the plethysmographic recordings. To evaluate whether body temperature changes during hyperoxia might affect V̇e and Vt measurements, we measured body temperature in 10 mutant and 11 wild-type pups exposed to the above-described protocol. Each pup was lightly anesthetized with isoflurane using a method previously shown to be associated with rapid recovery in newborn mouse pups (10). Measurements were initiated 4–5 min after anesthesia, which is sufficient for recovery from anesthesia on P2 (10). A thermocouple probe was positioned through a 2- to 3-mm incision in the interscapular region, which is the area of highest skin temperature (6). Then the pup was placed in the plethysmograph in a restraining device to ensure that temperature probe position remained unchanged. Previous studies in newborn rats showed that colonic and interscapular temperatures were closely correlated over temperatures ranging from 22.5 to 37°C (38). Because the restraining device disrupted breathing measurements in several pups, we did not analyze the respiratory data.
Breathing variables were subjected to analyses of variance with genotype (mutant Phox2b+/− vs. wild-type Phox2b+/+) as the between-subject factor and time (nine levels: six 30-s means during hyperoxia followed by three 30-s means during normoxia) and test number (two levels) as repeated factors (Superanova Software, Abacus Concepts, Berkeley, CA). Normoxic values of breathing variables and apnea durations were analyzed with genotype and periods (four levels: baseline, prehyperoxia 1, prehyperoxia 2, and final values) as the factors. To take into account the heterogeneous correlations among the repeated time measurements, we adjusted the degrees of freedom using the Greenhouse and Geisser factor, which is a very conservative downward correction to the degrees of freedom (18). Within-subject main effects and interactions are presented, together with P values based on these adjusted degrees of freedom. When the overall analysis was significant, multiple pairwise post hoc analyses using the Bonferroni/Dunn method were conducted to determine where the significant differences lay. In all tests, the critical significance level was set at 0.05.
Breathing pattern during normoxia.
Large proportions of the 42-min recordings were free of artifacts in both the mutant pups [83% (SD 9)] and the wild-type pups [86% (SD 5)]. Breathing variables were averaged over four 3-min periods of normoxia (initial baseline levels, prehyperoxic levels for test 1 and test 2, and final values; Table 1). Normoxic V̇e values tended to increase from baseline to final values in both groups, although this trend was more marked in mutants than in wild-type pups (genotype-by-period interaction: P < 0.007): 43% (SD 75) in mutant pups vs. 21% (SD 58) in wild-type pups. In particular, prehyperoxic V̇e values were slightly but significantly larger in test 2 than in test 1 in mutant pups, due to significantly higher Vt values (Table 1). This trend was mainly due to Tt, which decreased significantly in mutants but remained unchanged in wild-type pups (genotype-by-period interaction: P < 0.012, Table 1) and, to a lesser extent, to Vt (genotype-by-period interaction: P < 0.006, Table 1).
Ventilatory response to hyperoxia.
In both tests, hyperoxia caused significant V̇e depression in mutants and wild-type pups (Figs. 1 and 2A) due to the Tt increase (Fig. 2B), whereas Vt changes were small (<5%) and nonsignificant (Fig. 2C). The initial V̇e decrease (the early response to O2) generally occurred within 1 min of O2 exposure and was followed by an increase (the late response to O2) toward baseline levels (Figs. 1 and 2A).
The time courses of the V̇e and Tt responses to O2 were different in mutants and wild-type pups (genotype-by-time interaction for V̇e and Tt: P < 0.0001; Vt: nonsignificant, Fig. 2, A–C). In wild-type pups, V̇e rapidly returned to baseline levels during hyperoxia (Fig. 2A); a single mean V̇e value (30 s after O2 onset) was lower than the pre-O2 value (paired t-tests: test 1: P < 0.0007, test 2: P < 0.0001; all other differences were nonsignificant). In contrast, in mutants, V̇e remained lower than the pre-O2 level throughout O2 exposure and for 1 min following the return to normoxia (Fig. 2A, test 1 and test 2: P < 0.0001 for all comparisons with pre-O2 levels, except for the first 30-s period). Furthermore, in test 2, V̇e was significantly lower than the pre-O2 level as soon as the first 30-s period (P < 0.0001, Fig. 2A).
The minimum V̇e values, expressed as percentages of prehypoxic levels, and the corresponding Tt values were larger in mutants than in wild-type pups [V̇e: −37% (SD 13) and −25% (SD 18), respectively, P < 0.0001; Tt: 61% (SD 45) and 36% (SD 39); respectively, P < 0.0001; Vt: nonsignificant; Fig. 2, D–F]. Finally, minimum V̇e values were lower in test 2 than in test 1, irrespective of genotype, confirming the potentiation effect mentioned above. Repetition differentially affected Vt in mutants and wild-type pups, but the changes were small in both tests (test-by-genotype interaction, P < 0.038, Fig. 2F). The V̇e decreases were not significantly correlated to body weight or to body temperature measured after plethysmography.
Thus mutant pups showed a larger and more sustained ventilatory decrease in response to hyperoxia than their wild-type littermates.
Total apnea duration during normoxia was not significantly different in mutant and wild-type pups (Table 1). As with V̇e and Tt, the increase in total apnea duration (relative to normoxic levels) displayed a biphasic pattern during hyperoxia, although with greater interindividual variability (main effect for group: P < 0.0001, genotype-by-time interaction: P < 0.017 and partial comparisons, Fig. 3A). The increases in total apnea duration during hyperoxia were significantly larger in mutants than in wild-type pups in both tests (test 1: P < 0.0001; test 2: P < 0.0001, genotype-by-test interaction: not significant; Fig. 3B). These increases were not significantly correlated to body weight in either genotype group. Thus the evaluation of apneas confirmed that the hyperoxia-induced ventilatory depression was stronger in mutants than in wild-type pups. In both groups, the increase in apnea duration was weakly but significantly correlated with the V̇e decrease (mutants: test 1, r2 = 0.267; test 2, r2 = 0.276; wild-type pups: test 1, r2 = 0.337; test 2: r2 = 0.246, all values significant at P < 0.0001, group differences nonsignificant).
Body temperature varied within very narrow limits throughout the recording session, with no significant effect of genotype (Table 2). This showed that the reported changes in V̇e and Vt during hyperoxia and the genotype-related differences in breathing pattern were not due to body temperature differences.
In this study, we showed that the ventilatory decrease caused by hyperoxia was larger in newborn Phox2b+/− mutant mice than in their wild-type littermates. Furthermore, mutant pups showed a more sustained ventilatory decrease, outlasting the return to normoxia, compared with wild-type pups. In mutants, the depressed ventilation was further aggravated by a longer apnea duration in response to hyperoxia.
Limitations of whole body plethysmography in newborn mice.
Whole body plethysmography is the only available method for measuring breathing variables in unrestrained newborn mice. However, it has been validated against pneumotachography in larger animals (26), but not in newborn mice. This limitation stems from the lack of miniaturized reference devices (spirometers and pneumotachographs) and from difficulties in obtaining accurate body temperature measurements. Thus the absolute Vt and V̇e values in the present study should be considered with caution. However, this limitation does not invalidate our results, because our main finding, i.e., the genotype-related difference in ventilatory response to hyperoxia, is based on Tt and apneas, which are reliably measured by plethysmography.
Physiological chemodenervation in newborn mice.
The initial V̇e decrease caused by hyperoxia was chiefly ascribable to a Tt increase, whereas Vt changes were small and nonsignificant, in line with previous studies in mice pups of similar ages (28). In humans, the ventilatory response to hyperoxia undergoes developmental changes. In preterm infants, V̇e decreased after 3 min of O2 exposure 2 days after birth, as a result of a decrease in respiratory rate with little or no change in Vt, whereas the opposite breathing pattern was observed 6 days after birth (35) and in term infants between P2 and P6 (16). Thus the breathing strategy of 2-day-old mice pups in the present study resembled the one reported in preterm infants shortly after birth. This similarity further supports the use of newborn mice as a model of respiratory disorders in preterm infants, who contribute 20% of patients with CCHS (11). Furthermore, hyperoxia also increased apnea duration, as previously reported in human (term or preterm) infants (1). The weak correlation between V̇e decrease and apnea duration suggested that these two indexes of ventilatory depression might be complementary. The V̇e decrease and the increase in apnea duration were larger during the second than the first test, suggesting potentiation of the hyperoxic response. However, this effect was not significantly related to genotype.
Augmented tonic drive of peripheral chemoreceptors in Phox2b+/− mutant pups.
The V̇e decrease caused by hyperoxia was stronger in mutants than in wild-type pups, suggesting that the tonic activity of oxygen-sensitive peripheral chemoreceptors was greater in mutant pups. This augmented peripheral tonic input may be ascribable to low arterial Po2 (PaO2) levels, which, unfortunately, cannot be measured in newborn mice using current techniques. Our data did not indicate whether the ventilatory control disorders previously reported in mutant pups [low CO2 chemosensitivity (9) and sleep-related apneas (11)] were associated with low PaO2 values. In human term or preterm infants, periodic breathing was associated with low PaO2 values and with greater decreases in ventilation and longer apnea times in response to 100% O2 (1, 33, 34). Mutant pups may display a similar pattern of breathing disorders. The present finding that mutant pups displayed a vigorous V̇e response to hyperoxia extends our previous result that the phasic response to acute hypoxia is spared in Phox2b+/− mutant pups.
The augmented ventilatory decrease in response to hyperoxia in mutant pups may also reflect impairment of excitatory effects on the respiratory central pattern generator indirectly exerted by sustained (as opposed to short-lived) hyperoxia, considering that 1 min of O2 was needed to flush the chamber and that O2 was maintained for an additional 2 min. These excitatory effects, some involving CO2 chemosensitivity, progressively offset the initial V̇e decrease (13, 27). Among these, cerebral vasoconstriction and decreased hemoglobin transport capacity for tissue CO2 [Haldane effect (17)] are thought to decrease CO2 elimination and to increase brain tissue Pco2 at the central chemoreceptor level (20, 27). Genotype-related differences in the hyperoxic response may stem from impaired CO2 chemosensitivity in mutant pups (9). Previous experiments performed on the same colony, at the same postnatal age (P2), and using the same experimental setup showed that the ventilatory response to 8% CO2 was about twice as small in mutant as in wild-type pups (9). This impairment may account for the larger and more sustained ventilatory decrease shown by mutant pups. In line with this possibility, we found that impairment of the Tt response explained the greater ventilatory depression in mutant pups, as well as their impaired ventilatory response to CO2, compared with wild-type pups (9). Finally, genotype-related differences in the metabolic response to hyperoxia may also account for the ventilatory differences. O2 consumption and CO2 production were not measured in the present study, but the stability of body temperature during hyperoxic tests did not support a major role for metabolic changes.
Implications for the pathogenesis of CCHS.
Our laboratory previously reported that newborn Phox2b+/− mutant mice showed a weaker response to CO2 (9), a longer sleep-apnea time (11), and a normal hyperpneic response to hypoxia followed by an increased posthypoxic decline, compared with their wild-type littermates (9). The present study shows that the hyperoxic response was preserved. In CCHS children who were able to sustain adequate ventilation during wakefulness, peripheral chemosensitivity to oxygen was intact (15). However, the breathing frequency response to 100% O2 (ventilation was not measured) occurred earlier and was stronger in CCHS children than in controls (21). Taken together, these studies suggest that tonic drive to breathing from oxygen-sensitive chemoreceptors may be impaired only in the most severe cases of CCHS and that Phox2b+/− newborn mice may exhibit impairments akin to CCHS in its mildest form (9, 11).
The milder phenotype of Phox2b+/− mice compared with CCHS may be due to functional differences between the Phox2b-targeted mutation in mice (a null mutation) and the alanine expansion generally found in the PHOX2B gene of CCHS patients. Alanine expansions may result in a protein that can bind to the correct targets in the genome but is unable to carry out its normal regulatory function and, therefore, competes with the product of the wild-type allele (4). In contrast, a single functional Phox2b allele may ensure correct protein function, leading to a less severe phenotype (8). Furthermore, alanine-expanded proteins may be toxic to vulnerable cells that express them, due to aggregate formation (7).
Alternatively, the more severe phenotype in CCHS patients may be due to genetic factors. Heterozygous mutations affecting genes involved in important neural crest development pathways have been found in a subset of CCHS patients (14, 42). The phenotype in mice heterozygous for mutations in Phox2b and another gene previously found mutated in CCHS, which have not been described so far, may more closely resemble CCHS than the phenotype produced by a single mutation of Phox2b.
In conclusion, the ventilatory decrease caused by hyperoxia was larger and more sustained in mutant Phox2b+/− pups than in their wild-type littermates. Thus the tonic drive from oxygen-sensitive peripheral chemoreceptors was not disrupted, but was augmented, in mutant pups. Taken together, this and previous studies (9, 11) suggest that the impaired chemosensitivity to oxygen in CCHS may not be ascribable to PHOX2B loss of function related to heterozygosity.
This study was supported by the Association Française du Syndrome d’Ondine [grants to N. Ramanantsoa and to Institut National de la Santé et de la Recherche Médicale (INSERM) U676], the INSERM, the Paris VII University (Legs Poix, to C. Gaultier and J. Gallego), the EU (QLG2-CT-2001–01467 to C. Goridis), and the Association Française Contre les Myopathies (to C. Goridis).
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.
- Copyright © 2006 the American Physiological Society