Pairing ampakine with brief hypoxia evokes phrenic neuroplasticity via a spinal NMDA receptor-mediated mechanism.

Brief hypoxic episodes drive neuroplasticity in animal models and humans. Pretreatment with an allosteric α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor (R) modulator ("ampakine") enables a single hypoxic exposure to induce sustained increases in phrenic motor activity ["phrenic motor facilitation" (pMF)]. Phrenic nerve activity was recorded in anesthetized rats to determine whether the ampakine-hypoxia (A-H) combination is unique in its ability to evoke pMF and to determine its underlying mechanisms. Pairing ampakine CX717 with brief moderate or severe hypercapnia failed to produce pMF. Pairing doxapram, a respiratory stimulant, with hypoxia did not produce pMF. We then sequentially tested the hypotheses that A-H-induced pMF requires spinal serotonin, adenosine, or NMDA receptor activation. Cervical intrathecal delivery of serotonin (methysergide) or adenosine 2 A receptor (MSX-3) antagonists before A-H failed to prevent pMF. In contrast, the NMDA-R blocker MK-801 prevented pMF when administered before but not after A-H. Finally, as a step in the translational pathway, we tested the safety and efficacy of acute A-H exposure in unanesthetized rats with indwelling diaphragm electromyogram (EMG) wires after cervical spinal cord injury (SCI). A-H was well tolerated, and at 3 mo after SCI, increased diaphragm EMG output. We conclude that the mechanism driving sustained increases in phrenic motor output after A-H is independent of spinal adenosine or serotonin receptor activation, but requires spinal NMDA-R activation for the induction, but not maintenance, of A-H pMF. Ampakine pretreatment may be useful to increase the efficacy of hypoxia-based rehabilitation paradigms after SCI, particularly since clinical trials report a substantial number of "low responders."