Intermittent hypercapnic hypoxia: a model to study human respiratory motor plasticity?

Abstract

Once considered a fixed system operating solely via closed-loop negative chemofeedback, it is now widely appreciated that the respiratory control system exhibits considerable plasticity that facilitates robust homeostatic (blood-gas) regulation. A well-studied model of respiratory motor plasticity is phrenic long-term facilitation (LTF). Phrenic LTF describes sustained augmentation of phrenic nerve activity following repeated exposure to brief periods of low oxygen (i.e. intermittent hypoxia [IH]). The prolonged increase in phrenic nerve activity in response to IH is caused by episodic activation of carotid chemoafferent neurons. Distinct from the phenomenon originally referred to as phrenic ‘afterdischarge’ (Eldridge & Millhorn, 1986), IH-induced phrenic LTF is pattern sensitive and serotonin dependent. Upon activation of Hering’s nerve, brainstem raphe nuclei release serotonin on or near phrenic motor neurons in the cervical spinal cord. Binding of serotonin with Gq-coupled 5-HT2 receptors initiates intracellular signalling cascades resulting in de novo synthesis of brain-derived neurotrophic factor, ultimately leading to enhanced glutamatergic synaptic transmission/respiratory motor output (Fuller & Mitchell, 2017). Grounded in knowledge acquired from decades of rodent experiments, attempts to replicate phrenic LTF in humans have typically used changes in minute ventilation (V̇I, ventilatory LTF) as a surrogate for phrenic nerve activity. However, demonstration of ventilatory LTF has proven difficult as variations in experimental approaches (e.g. sleep versus wakefulness, duration and severity of hypoxic intervals, influence of CO2) have yielded conflicting results and confusion over the conditions necessary to induce and observe LTF. Nevertheless, it is evident that careful control of CO2 during and following IH is a critical feature of LTF expression. For example, when CO2 is variable (poikilocapnia), LTF is not observed. Conversely, when CO2 is raised slightly (∼5 mmHg) above eupnoeic levels (including baseline and post-IH recovery), LTF is revealed. In addition to phrenic/ventilatory LTF, IH also elicits plasticity at the carotid body and nucleus of the solitary tract. Long-lasting activation of sensory (peripheral chemoreceptor) discharge following IH is observed in rats preconditioned with chronic IH (Peng et al. 2003) and in reduced preparations exposed to IHwith concurrent hypercapnia (Roy et al. 2018). This form of sensory LTF may contribute to ventilatory LTF under certain circumstances. In a study recently published in The Journal of Physiology, Vermeulen et al. (2020) tested the hypothesis that IH, consisting of hypercapnic and hypoxic intervals (IHH), elicits an increase in V̇I that persists beyond the stimulus period (i.e. ventilatory LTF). It was further hypothesised that ventilatory LTF is mediated, in part, by increased tonic peripheral chemoreceptor activity (i.e. sensory LTF). Nineteen healthy adult males and females (age= 22± 3 years) participated in a single experimental testing session. Subjects rested supine as a host of respiratory and cardiovascular variables were measured. After 5 min rest, baseline end-tidal PO2 (PETO2 ) and PCO2 (PETCO2 ) were clamped using dynamic end-tidal forcing. To determine the influence of peripheral chemoreflex activation on changes in V̇I, three bouts of 1 min hyperoxia (target PETO2 = 350 mmHg) separated by 4 min normoxia/normocapnia were performed. Subjects were then exposed to 40 min IHH (1min intervals of 40 s hypercapnic hypoxia (PETO2 = ∼50 mmHg, PETCO2 = 4 mmHg above baseline) and 20 s room air) whilst continuously monitoring respiration. Upon completion of IHH, PETO2 and PETCO2 were restored to baseline (isocapnic-normoxic) and V̇I was recorded for a further 50 min. Once every 5 min into recovery, hyperoxia was delivered for 1 min to quantify the influence of peripheral chemoreceptor inhibition on ventilatory LTF. Nine subjects (all male) returned to the laboratory at least 72 h later for a time-matched control experiment breathing room air. The main findings were: (1) an increase in V̇I that persisted throughout recovery from IHH ( V̇I from baseline= 4.6± 3.7 l min−1), and (2) peripheral chemoreceptor inhibition attenuated, but did not abolish ventilatory LTF (baseline V̇I = −0.8 ± 0.9 l min−1, post-IHH V̇I = −1.7 ± 1.3 l min−1). Thus, it was observed that IHH elicits ventilatory and sensory LTF in healthy humans – the latter partially explains the former. Pertinent to basic science investigations of the respiratory control system and clinician scientists exploring the therapeutic potential of IHH, these results warrant specific discussion on: (1) the significance of CO2 in IH-induced ventilatory LTF, (2) clinical implications for neurorehabilitation, and (3) directions for future research.