Is sustained hypercapnia required to initiate plasticity in humans exposed to mild intermittent hypoxia?

Mild intermittent hypoxia protocols are characterized by a few (e.g. 12 episodes) brief episodes (e.g. 2 min) of mild hypoxia (e.g. 85–87 % oxygen saturation) interspersed with short recovery periods (Mateika & Sandhu 2011; Puri et al., 2021). In animal models, exposure to this stimulus leads to the initiation of long-term facilitation (LTF), which is a term used to characterize sustained increases in motor neuron, nerve or muscle activity initiated in animals following exposure to mild intermittent hypoxia. The sustained increases in motoneuronal activity have been linked to downstream increases in ventilation, upper airway patency and limb muscle function in animal models (Mateika et al., 2015). Some of the neuronal pathways and cellular mechanisms responsible for the sustained increase in activity have been determined (Mitchell & Baker 2022). Many studies completed over the past two decades have attempted to initiate LTF of ventilation, upper airway muscle or limb muscle activity in humans (Mateika & Sandhu 2011; Puri et al., 2021). However, the initiation of this phenomenon, or the magnitude of the phenomenon when it manifests, has been variable when comparisons are made across studies (see supplements in Mateika & Sandhu 2011; Puri et al., 2021). Potential reasons for this variability include differences in the number and duration of episodes that formulate mild intermittent hypoxia protocols, along with differences in the intensity of hypoxia considered to fall within the mild range (Mateika & Sandhu 2011; Puri et al., 2021). Likewise, some protocols have employed a combination of intermittent hypoxia and hypercapnia (see section below on 'Potential Mechanisms that Link Carbon Dioxide to the Initiation and Manifestation of LTF' for further discussion of this point) rather than intermittent hypoxia alone. Moreover, different human models (i.e. healthy humans, humans living with obstructive sleep apnoea (OSA), humans living with spinal cord injury) have been used to explore this phenomenon (Mateika & Sandhu 2011; Puri et al., 2021). These experimental design variations could be responsible for the variability that has been reported. On the other hand, we propose that one of the primary keys to the initiation and subsequent manifestation of LTF in humans is the presence of sustained levels of carbon dioxide during and following exposure to mild intermittent hypoxia. We and others established two decades ago that ventilatory LTF (vLTF) (Harris et al., 2006; Jordan et al., 2002; Khodadadeh et al., 2006; Mateika et al., 2004) and LTF of upper airway muscle activity (Harris et al., 2006) does not manifest in healthy humans (Harris et al., 2006; Jordan et al., 2002; Mateika et al., 2004) or humans with sleep apnoea (Khodadadeh et al., 2006) following exposure to mild intermittent hypoxia, when carbon dioxide levels are uncontrolled and hypocapnia ensues. Subsequent to these findings, ongoing studies provided additional insight by showing that sustained increases in respiratory and limb muscle function in humans, often presented in the framework of spinal cord injury, are frequently not evident following exposure to mild intermittent hypoxia in the presence of uncontrolled carbon dioxide levels (Gandevia & Butler 2024). Given the inaugural findings, these latter results are not surprising because common neural pathways and cellular mechanisms have been proposed to be involved in the initiation of LTF of respiratory and limb motor neurons based on work completed in animal models. To confirm the important role that carbon dioxide has in the initiation of LTF in humans, we took two approaches. The first approach was to show that even though vLTF was not evident during recovery from mild intermittent hypoxia in the presence of uncontrolled carbon dioxide, this phenomenon became evident in the presence of progressively increasing carbon dioxide levels (Khodadadeh et al., 2006; Mateika et al., 2004). Specifically, participants completed rebreathing studies before and following exposure to mild intermittent hypoxia that was unaccompanied by the manifestation of LTF (Khodadadeh et al., 2006; Mateika et al., 2004). During the rebreathing tests, hypoxia was sustained at 50 mmHg and carbon dioxide slowly increased over time. We showed that the ventilatory response to hypoxia at carbon dioxide levels 3 and 6 Torr above the recruitment threshold was elevated during the rebreathing tests completed after compared to before exposure to intermittent hypoxia (Khodadadeh et al., 2006; Mateika et al., 2004). This difference was not evident at carbon dioxide levels that demarcated the recruitment threshold (i.e. 0 Torr above the recruitment threshold) (Khodadadeh et al., 2006; Mateika et al., 2004). We suggested that this increased ventilatory response was due to the initiation of LTF. In subsequent experiments, we demonstrated that LTF of ventilation and genioglossus muscle activity was clearly evident when carbon dioxide levels were sustained above baseline (i.e. 5 mmHg above baseline in this study) throughout and following exposure to mild intermittent hypoxia, but was not evident if we reduced carbon dioxide levels back to baseline levels (Harris et al., 2006). Once this finding was established, we showed that exposure to acute mild intermittent hypoxia in the presence of sustained hypercapnia ( MI H C O 2 MIH_CO₂ ) resulted in vLTF in healthy men and women (Wadhwa et al., 2008) and that the magnitude of the response was independent of sex. We also showed that vLTF was initiated in individuals with OSA (Gerst et al., 2011; Syed et al., 2013) and that the magnitude of the response was greater compared to healthy individuals (Syed et al., 2013). Likewise, exposure to MI H C O 2 MIH_CO₂ lead to the initiation of LTF in individuals with spinal cord injury (Tester et al., 2014). Moreover, we initiated vLTF during wakefulness and sleep in humans (Syed et al., 2013) and showed that the magnitude of the response was greater during wakefulness. We also showed that repeated daily exposure to MI H C O 2 MIH_CO₂ enhanced the magnitude of LTF in humans (Gerst et al., 2011) and was greater in the evening compared to the morning in individuals with obstructive sleep apnoea (Gerst et al., 2011). In the studies completed over the years, the degree to which carbon dioxide was sustained above baseline ranged from 2–5 mmHg (Panza et al., 2023). In all cases, vLTF was evident, although the magnitude of the response varied (Panza et al., 2023) (see section below on 'Potential Mechanisms that Link Carbon Dioxide to the Initiation and Manifestation of LTF' for further discussion of this point). The mechanistic role(s) that sustained elevated levels of carbon dioxide have in the initiation and manifestation of LTF following exposure to MI H C O 2 MIH_CO₂ is not fully established, although there are a number of possibilities. Spinal motoneurons associated with movement and breathing are innervated in part by similar medullary neuronal pathways (e.g. raphe neurons) that are activated by peripheral chemoreflexes that respond to mild intermittent hypoxia. Thus, the inability to initiate LTF in spinal motoneurons could be a result of peripheral chemoreflex feedback that is insufficient to activate downstream mechanisms. Our recent findings support this suggestion (Panza et al., 2023). We showed in humans (n = 124) that the magnitude of LTF is predicted in part by the sensitivity of the hypoxic ventilatory response, which reflects peripheral chemoreflex sensitivity (Panza et al., 2023). It is well established that increasing the level of sustained carbon dioxide for a given level of hypoxia is associated with an increase in the hypoxic ventilatory response (Duffin & Mateika 2013). Given this relationship, the input from the peripheral chemoreceptors known to initiate LTF will be severely diminished in the presence of hypocapnia. Thus, maintaining carbon dioxide levels in the presence of mild intermittent hypoxia serves to ensure that an adequate stimulus originating from the peripheral chemoreceptors exists to initiate LTF. It is also possible that LTF is initiated but does not manifest because uncontrolled carbon dioxide is reduced significantly below a well-defined threshold (referred to as the apnoeic threshold during sleep and the recruitment threshold during wakefulness) leading to the cessation of breathing. The cessation of breathing via reductions in carbon dioxide is easily induced during sleep using artificial ventilation and is evident in some individuals during wakefulness following hyperventilation. Consequently, if carbon dioxide is significantly reduced during the application of mild intermittent hypoxia and the reduction in carbon dioxide endures after exposure, LTF might not be evident in measures of ventilation because of the withdrawal of this powerful stimulus, even though the phenomenon was initiated (Mateika & Narwani 2009). This possibility is particularly applicable to spinal motoneurons that innervate respiratory muscles and receive inputs from medullary respiratory neurons and other neuronal groups (i.e. raphe neurons) that have a role in initiating LTF. Thus, maintaining levels of carbon dioxide may impact both the initiation and manifestation of LTF. Lastly, sustained elevated levels of carbon dioxide, independent of mild intermittent hypoxia, could activate distinct neuronal and cellular pathways that independently initiate LTF (Mateika et al., 2018). Work completed using animal models has indicated that separate neuronal pathways that offset each other are activated by intermittent hypoxia and intermittent hypercapnia (Kinkead et al., 2001). On the other hand, to our knowledge, pathways activated by sustained levels of carbon dioxide in the context of initiating LTF have not been discovered (Mateika et al., 2018). Nonetheless, studies in humans have revealed that sustained levels of hypercapnia can lead to ventilatory drift which suggests that novel mechanisms may be activated by this stimulus (Harris et al., 2006). However, there is also evidence to suggest that sustained hypercapnia is not solely responsible for the magnitude of LTF. We have shown that the magnitude of the drift in ventilation recorded in response to sustained hypercapnia is much smaller than the response recorded when humans are exposed to MI H C O 2 MIH_CO₂ (Gerst et al., 2011). Thus, it is possible that MI H C O 2 MIH_CO₂ initiates LTF via the interaction of separate mechanistic pathways (Mateika et al., 2018). If so, allowing carbon dioxide levels to decrease in an uncontrolled manner would result in the removal of a stimulus that is capable of initiating LTF independently. Given that sustained levels of carbon dioxide appears to have an important role in initiating vLTF in healthy humans (Harris et al., 2006; Wadhwa et al., 2008) and humans living with sleep apnoea (Syed et al., 2013) or spinal cord injury (Tester et al., 2014), MI H C O 2 MIH_CO₂ may ultimately prove to be an effective therapeutic modality that has a multipronged effect on numerous physiological systems including respiratory and cardiovascular outcomes in individuals living with sleep apnoea (Puri et al., 2021). We have begun to explore this possibility and, to date, our work has focused on outcome measures linked to upper airway patency and cardiovascular function. We showed that acute and repeated daily exposure to MI H C O 2 MIH_CO₂ improves upper airway patency (El-Chami et al., 2017; Panza et al., 2022). Reduced critical closing pressure measures indicated that upper airway patency improves following exposure to MI H C O 2 MIH_CO₂ (Panza et al., 2022). In addition, we showed that continuous positive airway pressure (CPAP) required to treat sleep apnoea was reduced as a consequence of improved upper airway patency and as a result treatment adherence improved (Panza et al., 2022). Presently, we are exploring whether exposure to MI H C O 2 MIH_CO₂ also dampens responses to tactile pressure and auditory noise, which are stimuli often experienced when patients are treated with CPAP. If our preliminary findings are correct, these additional modifications might increase the arousal threshold to these stimuli, which could contribute to improving treatment adherence. Given the changes in the critical closing pressure that indicate improvement in upper airway patency we expected that acute exposure to MI H C O 2 MIH_CO₂ would be coupled to decreases in the apnoea/hypopnea index (Syed et al., 2013). This was not the case and we postulated that other forms of plasticity (i.e. progressive augmentation) initiated by MI H C O 2 MIH_CO₂ may override any benefits associated with improved airway patency in the absence of CPAP treatment (Mateika & Narwani 2009). We are presently exploring whether timing of exposure and length of exposure might alter the balance between these forms of plasticity and their impact on apnoea severity. Improvement in treatment adherence could have important implications for modifying cardiovascular outcome measures in individuals with sleep apnoea. Moreover, MI H C O 2 MIH_CO₂ might have direct effects on cardiovascular outcomes independent of improved adherence. To date, we have shown that repeated daily exposure to MI H C O 2 MIH_CO₂ results in a significant decrease in blood pressure in OSA patients treated nightly with CPAP (Panza et al., 2022). We are following up these studies by exploring whether this decrease is evident in the OSA population independent of nightly treatment with CPAP. Likewise, we are exploring whether blood pressure modifications are sustained for long periods of time. Our preliminary data suggests that this is the case because reductions in blood pressure following treatment with MI H C O 2 MIH_CO₂ are sustained for up to 8 weeks in most individuals. There are a number of additional studies to be performed in the coming years. The mechanisms underlying the impact of MI H C O 2 MIH_CO₂ on select outcomes measures are of interest, particularly in the context of teasing out the roles of MIH and sustained carbon dioxide individually and combined. It is also of interest to explore whether sustaining elevated carbon dioxide levels is necessary to attain maximum therapeutic value for all outcome measures or whether this requirement is limited to specific physiologic responses. In addition, studies designed to determine the efficacy of different doses of MI H C O 2 MIH_CO₂ , at the same time as considering the concurrent impact of hypoxic sensitivity on efficacy, comprise an important next step. In the context of these dosing studies, the potential safety issues related to exposure to MI H C O 2 MIH_CO₂ should be further explored. It is possible that some subgroups within a given population (e.g. low vs. high hypoxic sensitivity) will not benefit from this therapy or will experience detrimental outcomes, which could be an additional explanation for the variable responses that have been reported to date. However, it should be noted that in our studies and other published studies significant adverse events linked to MI H C O 2 MIH_CO₂ have not been reported. Likewise, based on prior work the timing of administration of this stimulus and its effectiveness on outcome measures, particularly as it relates to improvement in the severity of OSA, will be an important contribution to the existing literature. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article. No competing interests declared. J.H.M. was responsible for the conception of the work. J.H.M., R.B. and D.M.K. contributed to drafting the work or revising it critically for important intellectual content. J.H.M., R.B. and D.M.K. approved the final version of the manuscript submitted for publication and agree to be accountable for all aspects of the work by ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Lastly, all persons designated as authors qualify for authorship, and all those who qualify for authorship are listed. This work was supported by the United States Department of Veterans Affairs (I01CX00125, IK6CX002287) and the National Institute of Heart, Lung and Blood (R01HL085537).