Intermittent hypoxia and neurorehabilitation

Abstract

In recent years, it has become clear that brief, repeated presentations of hypoxia [i.e., acute intermittent hypoxia (AIH)] can boost the efficacy of more traditional therapeutic strategies in certain cases of neurologic dysfunction. This hypothesis derives from a series of studies in animal models and human subjects performed over the past 35 yr. In 1980, Millhorn et al. (Millhorn DE, Eldridge FL, Waldrop TG. Respir Physiol 41: 87-103, 1980) showed that electrical stimulation of carotid chemoafferent neurons produced a persistent, serotonin-dependent increase in phrenic motor output that outlasts the stimulus for more than 90 min (i.e., a "respiratory memory"). AIH elicits similar phrenic "long-term facilitation" (LTF) by a mechanism that requires cervical spinal serotonin receptor activation and de novo protein synthesis. From 2003 to present, a series of studies demonstrated that AIH can induce neuroplasticity in the injured spinal cord, causing functional recovery of breathing capacity after cervical spinal injury. Subsequently, it was demonstrated that repeated AIH (rAIH) can induce recovery of limb function, and the functional benefits of rAIH are greatest when paired with task-specific training. Since uncontrolled and/or prolonged intermittent hypoxia can elicit pathophysiology, a challenge of intermittent hypoxia research is to ensure that therapeutic protocols are well below the threshold for pathogenesis. This is possible since many low dose rAIH protocols have induced functional benefits without evidence of pathology. We propose that carefully controlled rAIH is a safe and noninvasive modality that can be paired with other neurorehabilitative strategies including traditional activity-based physical therapy or cell-based therapies such as intraspinal transplantation of neural progenitors.

Keywords: cellular transplantation; intermittent hypoxia; neurorehabilitation; spinal cord injury.

Fig. 1.

Timeline depicting some of the important milestones in the study of intermittent hypoxia and respiratory neuroplasticity. Specific emphasis is placed on the initial seminal work that established the concept of “respiratory memories,” and then on publications that have directly lead to the use of intermittent hypoxia as a potential therapeutic modality in humans with spinal cord injury (SCI). This timeline represents a summary, and many important contributions were not included. BDNF, brain-derived neurotrophic factor.

Fig. 2.

Working model of the cellular pathways contributing to long-lasting phrenic motor facilitation triggered by intermittent hypoxia. The highly simplified drawing depicts “respiratory drive” as a single axon terminal originating from the brainstem. Membrane proteins and intracellular pathways are depicted on a phrenic motoneuron. The Q pathway (left) is elicited by intermittent activation of G_q-coupled metabotropic receptors (e.g., 5-HT₂ or α₁), followed by activation of PKC, new synthesis of BDNF, activation of mature tropomyosin-related kinase B (TrkB), and activation of ERK MAP kinases (pERK). The S pathway (right) is elicited by G_s-coupled metabotropic receptors (e.g., 5-HT₇ and A_2A), followed by activation of cAMP, new synthesis of an immature TrkB isoform, and downstream signaling via Akt phosphorylation/activation (pAkt). The specific mechanisms by which pERK (Q pathway) and pAkt (S pathway) elicit persistent increases in phrenic motor output are not known but are likely to involve changes in motoneuron excitability and/or synaptic strength. pLTF, phrenic long-term facilitation.

Fig. 3.

Intermittent hypoxia can act throughout the spinal neuraxis following SCI. Initial work on acute intermittent hypoxia (AIH) focused on the respiratory system, but recent work (see Fig. 1) demonstrates that upper and lower extremity motor function is also enhanced after daily AIH. This conceptual diagram depicts strengthening of synaptic connections in the mid-cervical (phrenic-diaphragm motor system), low cervical (forelimb/upper extremities), and lumbo-sacral spinal cord (leg/ankle function). We propose that common mechanisms contribute to plasticity in these respiratory and nonrespiratory motor systems after AIH therapy (e.g., triggered by serotonin receptor activation; see Fig. 2 for a more comprehensive summary). Collectively, evidence from both human and animal models indicates that AIH is safe and easy to administer, induces robust spinal neuroplasticity, and may be an effective therapeutic approach to enhance motor function in persons with chronic SCI.

Fig. 4.

Anecdotal data suggest that rAIH can “train” a spinal cord transplant following cervical SCI. In this experiment, rat embryonic day 14 fetal spinal cord tissue was transplanted into acute C2 hemilesion (C2Hx) cavity in adult rats. Rats were then exposed to 10, 5-min hypoxic episodes (10% O₂, balance N₂) for 10 wk, beginning 1 wk post-C2Hx. After 10 wk, electrical activity of graft tissue and the ipsilateral phrenic nerve were recorded in anesthetized, vagotomized and ventilated rats during baseline (Fi_O2 = 0.5–0.6) and hypoxia (Fi_O2 = 0.13–0.15). The graft tissue was visualized using a dorsal surgical approach, and a 0.4–0.8 MΩ electrode (carbostar-3; Kation Scientific) was inserted directly into the graft. A: examples of graft and phrenic nerve activity (both “raw” and “integrated”, ∫) during baseline conditions and hypoxic challenge. The graft recording shows considerable tonic activity, but a clear respiratory-related discharge can be appreciated. B: cross-correlation analysis of the graft and phrenic bursting depicted in A. The correlogram shows a clear central peak that is consistent with the hypothesis that the graft and host phrenic motor neuron pool shared a common synaptic input. C: waveform averages (several minutes of data) of graft and phrenic signals. The averages were generated using phrenic burst onset as a trigger (arrows). Graft neurons exhibit inspiratory activity during baseline and hypoxia-triggered preinspiratory activity in the graft recording. Note also that tonic activity increased during hypoxia in the graft but decreased in the host phrenic activity.