Animal studies have shown that rhythmic breathing depends on two coupled medullary oscillators1 (Fig. 1A and 1B). The preBötzinger Complex (preBötC) generates inspiration and the parafacial respiratory group (pFRG) drives active expiration.2,3 The retrotrapezoid nucleus (RTN), a cluster of neurons adjacent to pFRG, plays a crucial role in carbon dioxide sensitivity4–6 and is involved in adapting alveolar ventilation to increased carbon dioxide expiration during exercise.7 Experimentally reducing RTN activity decreases the tidal volume (VT) responses to both hypercapnia and exercise.3,8 Of major pathophysiological significance, the RTN is almost or completely absent in rodents bearing mutations of the paired-like homeobox 2B (PHOX2B) gene.4–6

Fig. 1.

Summary of the study hypothesis. Panel A. Anatomical organization of the brainstem respiratory network in humans (left and top), inferred from analogies with rodent data and human data (see Schwarzacher, S. W., U. Rub and T. Deller 2011, reference 39 in the manuscript for the putative localization of the human pre-Bötzinger complex (preBötC); see Levy, J., F. Droz-Bartholet, M. Achour, P. Facchinetti, B. Parratte and F. Giuliano 2021, reference 15 in the manuscript for the putative location of the retrotrapezoid nucleus [RTN]). The image shows the localisation of the hypoglossal motor nucleus (XII), the nucleus ambiguus, the facial motor nucleus (VII), the inferior olive (IO), the lateral reticular nucleus (LRN), the parafacial respiratory group (pFRG) which contains the ventral parafacial nucleus (pFV, originally and still most often referred to as the RTN) and the lateral parafacial nucleus (pFL), and the ventral respiratory column (VRC) which contains the preBötC. The Fig. follows the hypothesis that the preBötC mostly drives inspiratory muscles (diaphragm and external intercostal muscles, shown in red) and that the pFRG mostly drives expiratory muscles (internal intercostal and abdominal muscles, shown in blue). Panel B. Conceptual representation of the organisation of respiratory central pattern generators, with inspiratory-related circuits shown in red and expiratory-related circuits shown in blue. The rostral ventral respiratory group (rVRG) contains inspiratory neurons and premotor neurons. The caudal ventral respiratory group (cVRG) contains expiratory premotor neurons. This organisation mainly involves three essential components: 1) an (inexorable) inspiratory oscillator in preBötC that drives inspiration by exciting inspiratory premotor neuronal populations, e.g., rVRG and parahypoglossal region (XII), and inhibits pFL; 2) a (conditional) expiratory oscillator in pFL that gates and drives expiration by exciting expiratory premotor neuronal populations, such as cVRG and XII, and assures alteration of phases by exciting neurons that inhibit preBötC, e.g., inhibitory neurons in either the preBötC or BötC; and 3) a source of tonic drive in pFV (RTN) which derives from dorsal progenitor cells that express paired-like homeobox 2B (PHOX2B) that are able to detect signals related to CO2 and/or pH levels (possibly with the involvement of glia) and integrates other sensory afferents affecting respiratory drive, via excitatory connections to preBötC, BötC, and respiratory premotor neurons, such as rVRG, cVRG, and XII. Respiratory muscles are activated at a certain frequency (FB) and intensity to generate tidal volume (VT). The excitation of rVRG activates inspiratory muscles that produce the inspired component of tidal volume (VTi), at rest and its exercise-related increases. The excitation of cVRG activates expiratory muscles that produce the expired component of tidal volume (VTe), particularly during exercise. Panel C. Tidal volume (VT) dynamics at rest and during exercise in healthy humans (“controls”, top) and patients with congenital central hypoventilation syndrome (“CCHS”, bottom). At rest, inspiration is active and expiration is passive, breathing therefore occurring above functional residual capacity (FRC) defined as the respiratory system relaxation volume (dotted horizontal line). An inspiratory reserve volume (IRV) separates the end of tidal inspiration from full inspiration, and an expiratory reserve volume (ERV) separates the end of tidal expiration (i.e. FRC) from full expiration. The human ventilatory response to exercise is primarily achieved through an increase in tidal volume (VT) that encroaches on both IRV and ERV through the recruitment of inspiratory and expiratory muscles, as depicted on the top trace. Assuming that the human organisation of respiratory rhythmogenesis is similar to that described in rodents, we reasoned that PHOX2B-related RTN/pFRG dysfunction in CCHS patients would result in defective exercise-induced expiratory recruitment. The figure is original and based on a drawing commissioned to and contributed by artist Pierre Bourcier (Paris, France).

(0.62MB).

In humans, such mutations are responsible for congenital central hypoventilation syndrome (CCHS),9 formerly referred to as Ondine's curse syndrome. CCHS patients suffer from defective breathing control, with life-threatening sleep-related alveolar hypoventilation requiring ventilatory assistance at night and during naps.10 These patients have absent or blunted responses to hypercapnia and hypoxia.10–13 They also have diminished ventilatory response to exercise.14 Although a possible human homologue of the RTN has recently been identified,15 it is currently impossible to link human PHOX2B mutations and functional respiratory perturbations to RTN abnormalities.

Bearing in mind that in the human ventilatory response to exercise VT increases through the recruitment of both inspiratory muscle to take advantage of the inspiratory reserve volume (IRV) and expiratory muscle to exploit the expiratory reserve volume (ERV) (Fig. 1C),16 and assuming that the human organisation of respiratory control is similar to that described in rodents, we reasoned that PHOX2B-related RTN dysfunction in CCHS patients would result in defective exercise-induced expiratory recruitment. To test this hypothesis (summarised in Fig. 1), we conducted a detailed analysis of the ventilatory response to exercise in adult CCHS patients and compared it to the corresponding response in healthy controls.

Participants’ characteristics and resting lung function testing are summarised in Table 1. There were no statistical differences between CCHS subjects and healthy controls in terms of sex, age, height, weight, BMI and lung function test variables (Table 1). Blood gases while breathing room air at rest and the ventilatory response to CO2 were not available for healthy controls as they were selected from a healthy-control database used in previous publications17,18 in which blood gases and the ventilatory response to CO2 were not performed. For the same reasons described above, the PHOX2B testing was not available in historical healthy subject cohort.

Our study included a genotypically and phenotypically homogeneous set of CCHS patients: twelve CCHS patients presented with PHOX2B polyalanine repeat expansion mutation (PARM) while only one CCHS patient presented with non-polyalanine repeat expansion mutation (NPARM). Among patients with PARM (n = 12): four presented with 5 additional Alanine expansion (20/25, 2 males vs 2 females), four presented with 6 additional Alanine expansion (20/26, 2 males vs 2 females), three presented with 7 additional Alanine expansion (20/27, 1 male vs 2 females), only one presented with 9 additional Alanine expansion (20/29, 1 female). None had tumors of the neural crest. One 23 years old male patient had Hirschsprung disease (PARM 20/26), a cardiac pace-maker but not a diaphragm pacing at the time of the study. All of them required ventilatory support only during sleep, while none required full-time assisted ventilation at the time of the study. None had assisted ventilation during the day, such as diaphragm pacing, at the time of the study. None had tracheostomy at the time of the study: nevertheless, one female patient had tracheostomy and had required daytime artificial ventilation until age of 8, one female patient had tracheostomy and had required daytime artificial ventilation until age of 10, one female patient had tracheostomy and had required daytime artificial ventilation until age of 15, one male patient had tracheostomy and had required daytime artificial ventilation until age of 16, one female patient had tracheostomy and had required daytime artificial ventilation until age of 17.

All patients were in fact ventilated overnight. All patients were ventilated using life support ventilator (Astral ventilator) with a pressure support mode and in approximately a third of cases with a targeted volume. All patients had a high backup rate (generally at least a minimum of 16 breaths/min). The level of expiratory pressure was usually set at 4 cmH2O and the inspiratory pressure at 18 cmH2O. Most of our cohort was ventilated using oronasal mask and none of them had oxygen entrained in their ventilator. All patients that underwent CPET had satisfactory control of overnight breathing. All patients were rigorously adherent to non-invasive ventilation use (approximately 8 hrs/day of use).

In addition, all CCHS patients were lifelong non-smokers. None had asthma, one had pulmonary arterial hypertension, two had arterial hypertension, three had cardiac pace-makers. None of these abnormalities contra-indicated CPET 26. Pulmonary function tests and room air arterial blood gases were normal (Table 1): none had elevated PETCO2 or PaCO2 at rest before exercise or low oxygen saturations (SpO2 or PaO2). The CO2 response was reduced compared with normal values published in the literature.

During exercise under standard CPET conditions, CCHS patients showed subnormal maximal oxygen uptake and reduced exercise tolerance. In most patients (11/13) the occurrence of hypercapnia during CPET indicated ventilatory limitation to exercise.30 In comparison to normal subjects who increased ventilation in response to exercise through both inspiratory and expiratory reserves, CCHS patients showed an abnormal ventilatory response without any indication of expiratory muscle recruitment.

In healthy individuals, alveolar ventilation (and thus V'E) rises linearly as a function of V'CO2 so that alveolar carbon dioxide and oxygen partial pressures (PACO2 and PAO2) and their arterial counterparts (PaCO2 and PaO2) are maintained near resting values throughout exercise.31 This involves: (i) mechanical and chemical feedback to medullary respiratory pattern generators adjusting respiratory muscle activity, and (ii) feed-forward mechanisms from higher neural centers driving the medullary generators and increasing locomotor muscle activity.32 In animals, two medullary generators drive inspiration (the preBötC) and expiration (the pFRG).2,33 Animal data show that, during exercise, the RTN, adjacent to the pFRG, is involved in adapting V’E in response to increased V'CO2 by augmenting expiratory activity,7,8 possibly via PHOX2B-positive neuronal projections to the pFRG.34 In healthy humans, V’E initially increases during exercise through increases in VT that normally expands to 50–60 % of vital capacity by encroaching on the IRV and ERV. The latter involves the early activation of expiratory muscles and results in reductions in ERV that vary with the type and intensity of exercise.16,25,31,35,36 This was observed in our control cohort. The recruitment of expiratory muscles directly increases V’E and lengthens respiratory muscles, optimizing their action and placing them at mechanical advantage, as well as storing chest wall elastic recoil energy to be released at the beginning of inspiration.37 This IRV-ERV recruitment pattern, clearly visible in our control group, was lacking in our CCHS patients in whom ERV increased (rather than decreased) from rest to 40 W. From 40 W to peak exercise ERV continued to decrease in controls but remained stable in CCHS patients. In contrast, inspiratory recruitment was higher in CCHS patients than controls. These differences were best seen when comparing CCHS and healthy controls at the same ventilatory stimulus, i.e. at iso-V’E, (Table 3 and Fig. 2). The same profile was observed with regards to V'E and VT response to exercise: V'E and VT increased in CCHS patients until 40 W, this being in line with previous published literature,14 and then decreased, which was not observed in the healthy participants (p<0.001).

Several mechanisms can impede exercise-induced expiratory reserve recruitment. Expiratory flow limitation (forced or exercise-induced expiratory flows equal to or lower than tidal expiratory flow)29 results from the bronchi closing prematurely in response to the expiratory driving pressure and is typically seen in chronic obstructive pulmonary disease (COPD)38 and in obesity.38 This was not observed in our study and has not been described in the literature for CCHS patients. In addition, the shape of the maximal flow-volume loop at rest was completely normal in CCHS patients, so that no expiratory flow limitation was observed at rest. During exercise VT increased only by encroaching on the IRV and not on the ERV, so that a flow reserve was still available on the expiratory portion of the maximal flow-volume loop, indicating that no expiratory flow limitation was noted in CCHS. Expiratory muscle weakness or fatigue can also limit expiratory reserve recruitment, as seen in patients with paraplegia.38 This mechanism does not seem likely in CCHS patients, who have normal lung function at rest. Inspiratory muscle weakness or fatigue reducing IC can give a false impression of impeded recruitment of expiratory reserve when, during CPET, ERV is calculated as resting FVC minus IC. Inspiratory muscle fatigue during exercise in our patients is made unlikely by the absence of any alteration of inspiratory reserve flow at peak exercise.29

Central neurological mechanisms can also be invoked. Although medullary sites have been identified in humans that could be homologous to the preBötC39 and RTN,15 the organisation and functioning of the respiratory network in humans is not necessarily similar to that described in rodents nor are the consequences of PHOX2B mutations necessarily the same in humans as in rodents. However, blunted CO2 chemosensitivity and a blunted ventilatory response to exercise characterized by defective expiratory recruitment that we have shown would be fully compatible with RTN dysfunction. In our study, CCHS patients showed a residual CO2 ventilatory response and responded to exercise with a progressive increase in VT encroaching on IRV, although all required ventilatory support during sleep. This indicates the persistence of wakefulness-related automatic rhythmogenesis. This may appear contradictory to the hypothesis of RTN dysfunction, with interactions between the RTN and the preBötC complex being necessary for the latter to produce inspiration.1 However, CCHS patients are known to activate respiratory-related cortical networks during wakefulness,40,41 a mechanism that is believed to maintain breathing through cortico-subcortical cooperation. In the hypothesis of RTN deficiency, the function of the cortical networks would be to allow the RTN-preBötC interaction to function despite the RTN defect, a possibility that could depend on the existence of some residual PHOX2B-positive cells, as has been observed in animals.6

In conclusion, while our study is phenomenological and does not provide formal proof of the hypothesised mechanism, we have shown defective exercise-induced expiratory recruitment in CCHS patients. The consistency of the observations made across a genotypically and phenotypically homogeneous set of patients, and their coherence with animal data, are of heuristic interest. For example, the possibility of improving breathing control pharmacologically was raised by observations of restored CO2 responses under contraceptive treatment by desogestrel.42 This has prompted various projects investigating pharmacological interventions to improve respiratory rhythmogenesis.43–45 Our data, by refining the phenotypic description of breathing control in CCHS patients, suggest the use of the ERV response to exercise as a possible novel outcome for such investigations.

PL: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing - original draft, Writing - review & editing. QF: Data curation, Formal analysis, Investigation, Writing - review & editing. MB: Data curation, Formal analysis, Writing - review & editing. MP: Investigation, Project administration, Writing - review & editing. BD: Investigation, Writing - review & editing. CL: Project administration, Writing - review & editing. CMP: Project administration, Writing - review & editing. FC: Methodology, Writing - review & editing. LB: Methodology, Writing - review & editing. CS: Conceptualization, Methodology, Writing - original draft, Writing - review & editing. TS: Conceptualization, Methodology, Writing - original draft, Writing - review & editing.

We declare that Artificial Intelligence program has not been used in writing this article.