Mechanisms of severe acute intermittent hypoxia-induced phrenic long-term facilitation

Abstract

Moderate acute intermittent hypoxia (mAIH; 35–55 mmHg Pa_O2) elicits phrenic long-term facilitation (pLTF) by a mechanism that requires activation of G_q protein-coupled serotonin type 2 receptors, MEK/ERK MAP kinase, and NADPH oxidase activity and is constrained by cAMP-PKA signaling. In contrast, severe AIH (sAIH; 25–35 mmHg Pa_O2) elicits G_s protein-coupled adenosine type 2 A receptor-dependent pLTF. Another G_s protein-coupled receptor, serotonin 7 receptors, elicits phrenic motor facilitation (pMF) by a mechanism that requires exchange protein activated by cyclic AMP (EPAC) and phosphatidylinositol 3-kinase/Akt (PI3K/Akt) activation and is constrained by NADPH oxidase activity. Here, we tested the hypothesis that the same downstream signaling mechanisms giving rise to serotonin 7 (vs. serotonin 2) receptor-induced pMF underlie sAIH-induced pLTF. In anesthetized rats, sAIH-induced pLTF was compared after pretreatment with intrathecal (C4) injections of inhibitors for: 1) EPAC (ESI-05); 2) MEK/ERK (UO126); 3) PKA (KT-5720); 4) PI3K/Akt (PI828); and 5) NADPH oxidase (apocynin). In partial agreement with our hypothesis, sAIH-induced pLTF was abolished by ESI-05 and PI828 and marginally enhanced by apocynin but, surprisingly, was abolished by UO126 and attenuated by KT-5720. Mechanisms of sAIH-induced pLTF reflect elements of both G_q and G_s pathways to pMF, likely as a consequence of the complex, cross-talk interactions between them.

NEW & NOTEWORTHY Distinct mechanisms give rise to pLTF induced by moderate and severe AIH. We demonstrate that, unlike moderate AIH, severe AIH-induced pLTF requires EPAC and PI3K/Akt and is marginally constrained by NADPH oxidase activity. Surprisingly, sAIH-induced pLTF requires MEK/ERK activity similar to moderate AIH-induced pLTF and is reduced by PKA inhibition. We suggest sAIH-induced pLTF arises from complex interactions between dominant mechanisms characteristic of moderate versus severe AIH-induced pLTF.

INTRODUCTION

Plasticity is a prominent feature of neuromotor systems, including the neural system controlling breathing. A general term for plasticity in the phrenic motor system is phrenic motor facilitation (pMF), which can be elicited pharmacologically (1–5) or by exposure to acute intermittent hypoxia (AIH) (6–10). AIH elicits a specialized form of pMF known as phrenic long-term facilitation (pLTF) (7, 9, 11–15). pLTF is characterized by a persistent (and progressive) increase in the amplitude of integrated phrenic bursts lasting many hours post-AIH.

Phrenic motor neurons exhibit pMF in response to spinally injected drugs targeting serotonin (2, 4, 5), adenosine (1), or adrenergic receptors (3). However, G_q (e.g., 5-HT_2A, 5-HT_2B, and α-1) versus G_s (e.g., A_2A and 5-HT₇) protein-coupled receptors elicit pMF by distinct, but interacting, cellular mechanisms (1–3, 5, 16–21). Similar pMF is revealed with intrathecal injections of downstream kinase activators, including activators of protein kinase C (PKC; 22), exchange protein activated by cAMP (EPAC; 18, 23), or protein kinase A (PKA; 23). These pathways are named for the G protein canonically coupled to the initiating neurotransmitter receptor, particularly G_q (Q pathway) and G_s (S pathway) coupled receptors (6).

AIH-induced pLTF also results from distinct cellular mechanisms, and the predominant mechanism depends on the severity of hypoxemia within each hypoxic episode of the AIH protocol. Whereas spinal 5-HT₂ receptors and the Q pathway predominate with moderate AIH (21, 24), spinal A_2A receptors and the S pathway predominate when AIH consists of more severe hypoxic episodes (sAIH; 15).

The S and Q pathways are not independent but interact via powerful cross-talk inhibition (17–19, 25, 26). Subthreshold activation of the S pathway constrains moderate AIH-induced pLTF via A_2A (27) and/or 5-HT₇ receptor activation (28). An essential mediator of such S to Q cross-talk inhibition is cyclic AMP, acting via PKA (not EPAC) (18, 19, 26). Conversely, subthreshold Q pathway activation constrains sAIH-induced pLTF via 5-HT_2A (26) or 5-HT_2B receptor activation (19). Spinal 5-HT_2A receptor activation inhibits 5-HT₇ receptor-induced pMF (i.e., S pathway) and sAIH-induced pLTF by a mechanism that requires spinal PKCδ activity (26), whereas 5-HT_2B receptors inhibit S pathway-induced pMF via NADPH oxidase activity (19). Although moderate AIH-induced pLTF requires NADPH oxidase activity and reactive oxygen species formation (29–31), the impact of NADPH oxidase inhibition on sAIH-induced pLTF has not been explored.

To date, most studies of the S pathway concern pMF induced via intrathecal drug injections (6, 19, 26), with little direct evidence concerning mechanisms contributing to sAIH-induced versus moderate AIH-induced pLTF. The major goal of this study was to investigate the role of enzymes known to play critical roles in S versus Q pathways to pMF, or their interactions. Specifically, we tested the hypothesis that spinal enzyme activities known to regulate the S pathway, including EPAC (18, 23) and phosphatidylinositol 3-kinase/Akt (PI3K/Akt) (1, 32), are necessary for sAIH-induced pLTF. We also tested the importance of an enzyme mediating S to Q pathway cross-talk inhibition: PKA (18, 28). We compared these results with outcomes following inhibition of an enzyme required for 5-HT_2A receptor-dependent, moderate AIH-induced pLTF, MEK/ERK MAP kinases (21, 32), and another mediating Q to S pathway cross-talk inhibition, NADPH oxidase (26). Our results reveal a complex picture but do confirm S pathway predominance in sAIH-induced pLTF, while leaving other questions unanswered.

MATERIALS AND METHODS

Experimental Groups

Experiments were performed using adult (3–4 mo) male Sprague Dawley rats (Harlan colony 211, Houston, TX). The Animal Care and Use Committee at the University of Wisconsin-Madison approved all experimental procedures. Each rat group received one of two treatments: 1) severe acute intermittent hypoxia (sAIH; Pa_O2: 25–30 mmHg), or 2) sham normoxia (i.e., time controls without AIH). Further, all rat groups received intrathecal drug delivery as described in detail below (Neurophysiology Protocol).

Neurophysiological Preparation

Experimental procedures have been described previously in detail in multiple publications (e.g., 15, 32). In brief, rats were anesthetized with isoflurane, tracheotomized, and ventilated (Rodent Ventilator, Model 683; Harvard Apparatus, Holliston, MA; tidal volume ∼2.5 mL, frequency ∼70–80). Isoflurane anesthesia was maintained (3.5% in 50% O₂, balance N₂) throughout surgical procedures; rats were converted to urethane anesthesia over 15–20 min (1.8 g kg⁻¹, i.v.) while slowly withdrawing isoflurane. Adequacy of anesthesia was verified before protocols commenced and immediately after completion by the lack of pressor responses or obvious respiratory neural responses to toe pinch with a hemostat. After conversion to urethane anesthesia, rats were given a continuous intravenous infusions to maintain body fluid and acid-base balance; infusions (1.5–6 mL/kg/h) consisted of a 1:2:0.13 mixture of 6% Hetastarch (in 0.9% sodium chloride), lactated Ringer’s solution, and 8.4% sodium bicarbonate.

Body temperature was assessed with a rectal thermometer (Fisher Scientific, Pittsburgh, PA) and maintained (37.5 ± 1°C) with a heated surgical table. To monitor end-tidal Pa_CO2 (Pet_CO2), a flow-through carbon dioxide analyzer was used with sufficient response time to measure Pet_CO2 in rats (Capnogard, Novametrix, Wallingford, CT). Pet_CO2 was maintained at ∼45 mmHg throughout the surgical preparation.

Rats were vagotomized bilaterally, and a polyethylene catheter (PE50 ID: 0.58 mm, OD: 0.965 mm; Intramedic, Maryland) was inserted into the right femoral artery to monitor blood pressure (Gould Pressure Transducer, P23ID) and draw samples for arterial blood gas analysis. These arterial blood samples were analyzed for partial pressures of O₂ (Pa_O2) and CO₂ (Pa_CO2) with a blood gas analyzer (ABL 800, Radiometer, Westlake, OH); blood gases were assessed during baseline, all three hypoxic episodes, and at 15, 30, 60, and 90 min post-AIH. Pa_CO2 was maintained within ± 1.5 mmHg of baseline levels by adjusting inspired CO₂ and/or ventilator rate throughout the experiment. After the third hypoxic episode, baseline O₂ levels were restored and then maintained for the duration of experiments.

The left phrenic nerves were isolated (dorsal approach), cut distally, desheathed, and covered with a saline-soaked cotton ball until placing the nerves on bipolar silver electrodes (see Neurophysiology Protocol below). Laminectomy was performed at cervical level 2 (C2) to enable intrathecal delivery of drugs. Once rats were converted to urethane anesthesia, a minimum of 1 h was allowed before protocols commenced.

Neurophysiology Protocol

To enable intrathecal drug delivery, a small incision was made in the dura at C2, and a soft silicone catheter (2 Fr; Access Technologies, Skokie, IL) was inserted caudally 3–4 mm until the tip rested approximately over the C4 segment to deliver drugs near the phrenic motor nucleus before sAIH [three 5-min episodes of isocapnic hypoxia (25–30 mmHg) separated by 5-min intervals of baseline oxygen levels (Pa_O2 ≥ 150 mmHg)]. The intrathecal catheter was attached to a 50-µL Hamilton syringe filled with drug or vehicle solutions as described below. The left phrenic nerve was submerged in mineral oil and placed on bipolar silver electrodes to record nerve activity. Neural signals were amplified (×10,000), band-pass filtered (300–10,000 Hz, Model 1800, A-M Systems, Carlsborg, WA), full-wave rectified, and integrated (Paynter filter, time constant, 50 ms, MA-821, CWE Inc., Ardmore, PA). Integrated nerve bursts were digitized (8 kHz) and analyzed using WINDAQ data acquisition system (DATAQ Instruments, Akron, OH). Rats were then paralyzed using pancuronium bromide for neuromuscular blockade (2.5 mg kg⁻¹, i.v.). To begin protocols, the apneic CO₂ threshold was determined by lowering Pet_CO2 until nerve activity ceased for approximately 1 min. The recruitment threshold was then determined by slowly increasing Pet_CO2 until nerve activity resumed (11). Pet_CO2 was raised ∼2 mmHg above the recruitment threshold, and ∼15 min were allowed to establish a stable baseline. This procedure assures comparable baseline phrenic nerve activities across individual rats, at least relative to the CO₂ apneic/recruitment thresholds. Further, this procedure assures comparability with other studies concerning molecules contributing to or regulating phrenic motor plasticity.

Rats were divided into groups to test the hypotheses that enzymes regulating the S pathway to pMF or mediating S to Q pathway cross-talk inhibition also regulate sAIH-induced pLTF (i.e., EPAC, PI3K/Akt, and PKA). As a contrast, we also tested an enzyme known to regulate the Q (not S) pathway to pMF (i.e., MEK/ERK and MAPK), and an enzyme mediating Q to S pathway cross-talk inhibition (i.e., NADPH oxidase or NOX). All drug or vehicle injections took place after a 15-min baseline was established, and were given in 2–3 µL boluses; all drug doses used were determined from previous studies (18, 28, 30, 32). Rat groups included: 1) vehicle for ESI-05 and PI828 [20% DMSO dissolved in saline, 10–12 µL, 20 min before sAIH; (18, 28) n = 3]; 2) ESI-05 [EPACi; 10 µL, 20 min before sAIH, 2 mM; BioLog Life Science Institute; (18) n = 5]; 3) ESI-05 time control (TC or no hypoxia; n = 2); 4) PI828 [2-(4-Morpholinyl)-8-(4-aminopheny)l-4H-1-benzopyran-4-one, PI3K/Akti; 12 µL, 20 min before sAIH, 100 µM; Tocris Bioscience; (32) n = 5]; 5) PI828 TC (n = 2); 6) vehicle for KT-5720 [20% DMSO in saline, 15 µL, 5 min before sAIH; (28) n = 7]; 7) KT-5720 [PKAi; 15 µL, 5 min before sAIH, 100 µM; Tocris Bioscience; (28) n = 8]; 8) vehicle TC (n = 3); 9) KT-5720 TC (n = 3); 10) vehicle for UO126 [20% DMSO dissolved in saline, 12 µL, 20 min before sAIH; (32) n = 5]; 11) UO126 [(1,4-diamino-2,3-dicyano-1,4-bis(o-aminophenylmercapto)butadiene, MEK/ERKi; 12 µL, 20 min before sAIH, 100 µM; Promega; (32) n = 5]; 12) vehicle TC (n = 1); 13) UO126 TC (n = 2); 14) vehicle for apocynin [10% DMSO dissolved in artificial cerebral spinal fluid (aCSF), 12 µL, 20 min before sAIH; (30) n = 5]; 15) apocynin [NADPH oxidase inhibitor (NOXi); 12 µL, 20 min before sAIH, 600 µM; Sigma; (30) n = 5]; 16) vehicle TC (n = 2); and 17) apocynin TC (n = 2). Since there were no apparent differences in vehicle (with sAIH) or TC groups (no AIH), and multiple previous studies demonstrating minimal experimental drift in time control experiments, all vehicle and TC rats were grouped together in S pathway or Q pathway studies for statistical analyses (S: vehicle n = 10, TC n = 10; Q: vehicle n = 10, TC n = 7). In this way, we were able to assure that drugs or vehicle did not elicit unexpected pMF (i.e., a time-dependent drift in time control experiments) while minimizing animal use consistent with national standards for animal research. The targets for the respective drugs are depicted in a working (hypothetical) cellular model of AIH-induced phrenic motor plasticity in Fig. 1.

Figure 1.

Working model of sAIH-induced pLTF. pLTF can be regulated by G_q (5-HT_2A/B) or G_s (A_2A) protein-coupled metabotropic receptors (the Q and S pathways, respectively). Following severe AIH, we hypothesize that activation of A_2A receptors activates adenyl cyclase (AC) and cAMP, with subsequent activation of EPAC, PI3K/Akt, mTORC, and iTrkB. Activation of EPAC drives PI3K activation, phosphorylating Akt to cause pLTF. In contrast, following moderate AIH, both 5-HT₂ isoforms are activated, with 5-HT_2A receptor activation increasing MEK/ERK activity. Both 5-HT₂ isoforms then utilize downstream signaling via BDNF, TrkB and PKCθ, although this convergence has not been verified. We further hypothesized that the S pathway is constrained by NADPH oxidase (NOX), whereas the Q pathway is constrained by PKA activity. We intrathecally delivered the inhibitors to key enzymes in this working cellular model, including: 1) ESI-05 (targeting EPAC), 2) PI828 (targeting PI3K/Akt), 3) KT-5720 (targeting PKA), 4) UO126 (targeting MEK/ERK), and 5) apocynin (targeting NADPH oxidase). AIH, acute intermittent hypoxia; BDNF, brain-derived neurotrophic factor; EPAC, exchange protein activated by cyclic AMP; mTORC, mammalian target of rapamycin; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; pLTF, phrenic long-term facilitation; sAIH, severe AIH.

Data Analyses

Integrated phrenic nerve burst amplitude was averaged over 1-min bins at each time (baseline, all three hypoxic episodes, and 15, 30, 60, and 90 min post-AIH) and normalized as a percent change from baseline. Respiratory frequency was expressed as a change from baseline (bursts per min). All statistical comparisons between treatment groups for nerve burst amplitude and frequency were made via two-way ANOVA with a repeated measures design (time post-AIH: 15, 30, 60, and 90 min; or hypoxic episodes for the short-term hypoxic response). A one-way ANOVA was used to compare nerve burst amplitude and frequency at 90 min posthypoxia. With statistically significant ANOVAs, individual comparisons were made using Tukey post hoc test (SigmaPlot v. 14.0; Systat Software Inc., San Jose, CA). Differences were considered significant if P < 0.05. Values are expressed as means ± 1 SE.

RESULTS

Blood Gases and Mean Arterial Pressures

During hypoxic episodes, arterial PCO₂ (Pa_CO2) was successfully maintained within 1.5 mm Hg of its baseline value at all times in all groups (Table 1). Therefore, changes in integrated phrenic nerve burst amplitude following AIH cannot be attributed to differences in chemoreceptor feedback, which can influence the apparent magnitude of pLTF (11, 14, 33). Pa_O2 was successfully maintained within the target range for severe AIH (25–30 mmHg), and mean arterial pressure (MAP) did not differ within groups during hypoxic episodes (Table 1). As expected, Pa_O2 and mean arterial pressure differed among groups when AIH versus TCs were compared during hypoxic episodes (Table 1). In addition, significant differences were detected during hypoxic episodes for MAP when comparing rats treated with KT-5720 (PKA inhibitor) or apocynin (NADPH oxidase inhibitor) to rats treated with ESI-05 (EPAC inhibitor) or PI828 (PI3K/Akt inhibitor; HX1: P = 0.026 for apocynin vs. ESI-05, P = 0.035 for apocynin vs. PI828, P = 0.017 for KT-5720 vs. ESI-05, and P = 0.024 for KT-5720 vs. PI828; HX2: P = 0.043 for apocynin vs. ESI-05, P = 0.033 for apocynin vs. PI828, P = 0.020 for KT-5720 vs. ESI-05, and P = 0.014 for KT-5720 vs. PI828; HX3: P = 0.048 for apocynin vs. ESI-05, P = 0.022 for apocynin vs. PI828, P = 0.041 for KT-5720 vs. ESI-05, and P = 0.017 for KT-5720 vs. PI828; Table 1).

Table 1.

Arterial Pco₂, Po₂, and MAP during hypoxic episodes

	PaCO2, mmHg			PaO2, mmHg			MAP, mmHg
	HX1	HX2	HX3	HX1	HX2	HX3	HX1	HX2	HX3
Spinal groups
Vehicle 1	47.6 ± 1.5	48.0 ± 1.3	47.6 ± 1.3	30 ± 1a	26 ± 1a	27 ± 1a	56 ± 7a	51 ± 7a	52 ± 7a
ESI-05	48.5 ± 0.9	50.6 ± 1.3	51.1 ± 0.8	28 ± 1a	25 ± 1a	27 ± 1a	73 ± 5	67 ± 3	64 ± 4
PI828	48.3 ± 2.9	47.8 ± 3.5	49.0 ± 2.5	27 ± 2a	27 ± 1a	28 ± 1a	72 ± 4	68 ± 3	66 ± 3
KT-5720	47.0 ± 0.9	45.0 ± 1.4	46.3 ± 1.5	29 ± 1a	28 ± 1a	27 ± 1a	45 ± 3a,b,c	39 ± 3a,b,c	38 ± 3a,b,c
(–AIH) 1	47.7 ± 1.0	48.2 ± 0.8	47.7 ± 1.0	332 ± 9	338 ± 9	333 ± 8	90 ± 5	90 ± 4	90 ± 4
Vehicle 2	45.9 ± 1.5	44.2 ± 1.4	44.1 ± 1.4	28 ± 1a	27 ± 1a	29 ± 1a	54 ± 8a	47 ± 6a	47 ± 6a
UO126	49.3 ± 1.1	48.4 ± 1.3	48.3 ± 2.4	28 ± 2a	27 ± 1a	29 ± 2a	61 ± 6	52 ± 6a	52 ± 6a
Apocynin	44.9 ± 2.1	46.1 ± 5.3	45.4 ± 2.1	27 ± 1a	26 ± 1a	25 ± 1a	43 ± 2a,b,c	39 ± 2a,b,c	36 ± 1a,b,c
(–AIH) 2	44.9 ± 1.6	45.3 ± 1.6	46.1 ± 1.7	283 ± 35	285 ± 33	294 ± 33	86 ± 2	85 ± 2	85 ± 2

All values are expressed as means ± SE. HX1, HX2, and HX3 designate hypoxic episodes 1, 2, and 3, respectively. AIH, acute intermittent hypoxia; –AIH, without AIH; EPAC, exchange protein activated by cyclic AMP; ESI-05, EPAC inhibitor; KT-5720, PKA inhibitor; MAP, mean arterial pressure; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; PI828, PI3K/Akt inhibitor; UO126, MEK/ERK inhibitor. Significant differences from: ^aappropriate time control group (-AIH), ^bESI-05, and ^cPI828. A two-way ANOVA with repeated measures design followed by a Tukey post hoc test was utilized. Differences were considered significant if P < 0.05. The n for each group is as follows: vehicle 1, n = 10; ESI-05, n = 5; PI828, n = 5; KT-5720, n = 8; (−AIH) 1, n = 10; vehicle 2, n = 10; UO126, n = 5; apocynin, n = 5; (−AIH) 2, n = 7.

No differences were detected when comparing Pa_CO2 during baseline or 90-min post-hypoxia; Pa_O2 remained above 150 mmHg at all times posthypoxia, and the only significant difference we observed was when comparing the second time control group at baseline versus 90-min post-AIH (P = 0.006; Table 2). For mean arterial pressure during baseline and at 90-min posthypoxia, we observed slight but significant differences within (both vehicle groups; P = 0.019 for vehicle 1, and P = 0.037 for vehicle 2 comparing baseline with 90-min post-AIH) and across groups [baseline: ESI-05 vs. apocynin, P = 0.047; 90-min posthypoxia: ESI-05 and PI828 vs. vehicle (P = 0.043 vs. ESI-05, and P = 0.004 vs. PI828), apocynin (P = 0.032 vs. ESI-05, and P = 0.004 vs. PI828), and KT-5720 (P = 0.045 vs. ESI-05, and P = 0.005 vs. PI828); Table 2]; changes in mean arterial pressure of <20 mmHg from baseline have minimal reported effect on respiratory activity in rats (11, 34). Thus, differential pLTF expression observed following drug treatments was not affected by differences in Pa_CO2, Pa_O2, or blood pressure regulation.

Table 2.

Arterial Pco₂, Po₂, and MAP during baseline and 90-min posthypoxia

	PaCO2, mmHg		PaO2, mmHg		MAP, mmHg
	Baseline	90 Min	Baseline	90 Min	Baseline	90 Min
Spinal groups
Vehicle 1	48.4 ± 1.1	48.7 ± 1.1	359 ± 10	329 ± 13	92 ± 5	80 ± 6a
ESI-05	49.0 ± 0.9	49.2 ± 1.0	338 ± 5	328 ± 6	110 ± 2c	105 ± 5b,c,d
PI828	48.2 ± 2.7	48.1 ± 2.6	309 ± 39	312 ± 20	105 ± 4	110 ± 5b,c,d
KT-5720	47.3 ± 0.9	47.3 ± 0.9	353 ± 10	324 ± 6	88 ± 3	79 ± 4
(–AIH) 1	48.2 ± 1.0	48.7 ± 1.0	329 ± 9	330 ± 12	90 ± 5	88 ± 4
Vehicle 2	46.4 ± 1.1	46.7 ± 1.1	313 ± 20	323 ± 9	87 ± 4	76 ± 5a
UO126	48.3 ± 0.9	48.1 ± 1.2	294 ± 25	350 ± 14	93 ± 2	84 ± 4
Apocynin	45.2 ± 2.3	45.5 ± 2.3	284 ± 40	332 ± 26	82 ± 6	75 ± 8
(–AIH) 2	46.4 ± 1.2	46.7 ± 1.0	283 ± 36	347 ± 23a	87 ± 2	83 ± 4

Values are expressed as means ± SE. AIH, acute intermittent hypoxia; –AIH, without AIH; EPAC, exchange protein activated by cyclic AMP; ESI-05, EPAC inhibitor; KT-5720, PKA inhibitor; MAP, mean arterial pressure; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; PI828, PI3K/Akt inhibitor; UO126, MEK/ERK inhibitor. Significant differences from: ^abaseline, ^bappropriate vehicle group, ^capocynin, and ^dKT-5720. A two-way ANOVA with repeated measures design followed by a Tukey post hoc test was utilized. Differences were considered significant if P < 0.05. The n for each group is as follows: vehicle 1, n = 10; ESI-05, n = 5; PI828, n = 5; KT-5720, n = 8; (−AIH) 1, n = 10; vehicle 2, n = 10; UO126, n = 5; apocynin, n = 5; (−AIH) 2, n = 7.

sAIH-Induced pLTF Is EPAC and PI3K/Akt Dependent

To test the hypothesis that enzymes required for S pathway-induced pMF are also required for sAIH-induced pLTF, we first studied sAIH-induced pLTF following spinal inhibition of EPAC (with ESI-05) or PI3K/Akt (with PI828). Representative phrenic neurograms during protocols for each treatment group are illustrated in Fig. 2, A–D. Whereas sAIH-induced pLTF was similar to previously reported values in vehicle-treated rats (15), significantly greater versus baseline (P < 0.001 at all times post-AIH), and versus time-controls (P = 0.003 at 15 min, and P < 0.001 at 30-, 60-, and 90-min post-AIH), pLTF was abolished with EPACi at 30 (P = 0.003), 60 (P < 0.001), and 90 min (P < 0.001) post-AIH, as well as PI3K/Akt inhibition at all times post-AIH (P = 0.008 at 15 min, P < 0.001 at 30-, 60-, and 90-min post-AIH) versus vehicle-treated rats (Fig. 2, A–C and E).

Figure 2.

pLTF is abolished following ESI-05 and PI828 treatment. A–D: representative traces of compressed, integrated phrenic nerve activity before and after sAIH following pretreatment with vehicle (A), ESI-05 (EPAC inhibitor, B), PI828 (PI3K/Akt inhibitor, C), and KT-5720 (PKA inhibitor, D). The dashed line in each trace represents baseline. E: phrenic burst amplitude (percent change from baseline) in vehicle (open diamonds)-, PKA inhibitor (filled squares)-, EPAC inhibitor (filled triangles)-, and PI3K/Akt inhibitor (filled circles)-treated rats after AIH, and time control (without AIH, gray diamonds)-treated rats. pLTF is significant in vehicle-treated rats versus EPACi-treated rats at 30-, 60-, and 90-min posthypoxia ($), and at all time-points post-AIH versus time control-treated rats (*), baseline (+), PKAi (#), and PI3K/Akti (&) (all P < 0.05). PKAi-treated rats showed significant pLTF versus baseline at 30-, 60-, and 90-min posthypoxia, and time control-treated rats at 90-min posthypoxia (all P < 0.05). F: frequency (change from baseline; bursts/min) in vehicle-, PKA inhibitor-, EPAC inhibitor-, and PI3K/Akt inhibitor-treated rats after AIH, and time control-treated rats. Small frequency LTF in vehicle-treated rats is observed at 30- and 60-min post-AIH versus baseline (+) (P < 0.05). PI3K/Akti-treated rats had decreased frequency LTF compared with baseline (+) at 15-min post-AIH (P < 0.05). A two-way ANOVA with repeated measures design was utilized, followed by a Tukey post hoc test. Values are means ± SE. AIH, acute intermittent hypoxia; EPAC, exchange protein activated by cyclic AMP; LTF, long-term facilitation; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; pLTF, phrenic long-term facilitation; sAIH, severe AIH. The n for each group is as follows: vehicle 1, n = 10; ESI-05, n = 5; PI828, n = 5; KT-5720, n = 8; time control, n = 10.

In vehicle-treated rats, pLTF was 51 ± 15%, 58 ± 14%, 76 ± 9%, and 78 ± 9% at 15, 30, 60, and 90 min, respectively (Fig. 2E). Following EPACi, pLTF was 16 ± 6%, 11 ± 6%, 15 ± 6%, and 10 ± 5% at 15, 30, 60, and 90 min, respectively (Fig. 2E). PI3K/Akt inhibition decreased pLTF to 7 ± 5%, 1 ± 5%, 7 ± 7%, and 3 ± 7% at 15, 30, 60, and 90 min, respectively (Fig. 2E).

Frequency LTF was statistically significant, but small, as reported before (15, 35). Phrenic burst frequency in vehicle treated rats was significantly different from baseline at 30- (P = 0.006) and 60 (P = 0.003)-min post-AIH (4 ± 2 bursts/min at 30 and 60 min; Fig. 2F). Rats treated with PI3K/Akt inhibitor exhibited less frequency change from baseline at 15 min post-AIH only (−5 ± 3 bursts/min; P = 0.017; Fig. 2F) and was unchanged at any time in rats after EPACi (P > 0.05; Fig. 2F).

PKA Inhibition Attenuates sAIH-Induced pLTF

To test the hypothesis that PKA, a protein activated by cAMP but known for its role in S to Q cross-talk inhibition (18, 23), is not necessary for sAIH-induced pLTF, we studied sAIH-induced pLTF following spinal PKA inhibition (with KT-5720). A typical phrenic neurogram during an experimental protocol is illustrated in Fig. 2. To our surprise, sAIH-induced pLTF was decreased but not abolished by PKA inhibition [20 ± 10%, 26 ± 12%, 33 ± 11%, and 34 ± 9% at 15, 30, 60, and 90 min, respectively; vs. vehicle-treated rats at all times post-AIH (P = 0.049 at 15, P = 0.033 at 30, P = 0.002 at 60, and P = 0.001 at 90 min); vs. baseline at 30 (P = 0.004), 60 (P < 0.001), and 90 (P < 0.001) min; and vs. time control rats at 90 min (P = 0.028); Fig. 2, D and E]. PKAi had not effect on frequency at any time (Fig. 2E).

sAIH-Induced pLTF Is MEK/ERK Dependent

To test the hypothesis an enzyme required for Q (not S) pathway pMF is unnecessary for sAIH-induced pLTF, we studied sAIH-induced pLTF following spinal MEK/ERK inhibition (with UO126). Contrary to our hypothesis, sAIH-induced pLTF was abolished by MEK/ERK inhibition [10 ± 3%, 8 ± 10%, 15 ± 8%, and 17 ± 4% at 15, 30, 60, and 90 min, respectively; vs. vehicle-treated rats at 60 (P = 0.006) and 90 min (P = 0.005); Fig. 3A, B, and D]. After MEK/ERK inhibition, sAIH-induced frequency LTF was −8 ± 7, −5 ± 6, −1 ± 2, and −1 ± 1 at 15, 30, 60, and 90 min, respectively, in rats treated with the MEK/ERK inhibitor (Fig. 3E). Frequency changes after MEK/ERK inhibition were significantly different from baseline and time control rats at 15-min post-AIH only (P = 0.007 vs. baseline; P = 0.008 vs. time control), and versus vehicle-treated rats at 15- (P = 0.016) and 30 (P = 0.048)-min post-AIH (Fig. 3E).

Figure 3.

pLTF is abolished following UO126, but marginally constrained by apocynin treatment. A–C: representative traces of compressed, integrated phrenic nerve activity before and after sAIH following pretreatment with vehicle (A), UO126 (MEK/ERK inhibitor, B), and apocynin (NADPH oxidase or NOX inhibitor, C). The dashed line in each trace represents baseline. D: phrenic burst amplitude (percent change from baseline) in vehicle (open diamonds)-, NOX inhibitor (filled squares)-, and MEK/ERK inhibitor (filled triangles)-treated rats after AIH, and time control (without AIH, gray diamonds)-treated rats. pLTF is significant in vehicle-treated rats at all time-points post-AIH versus time control-treated rats (*) and baseline (+), and versus MEK/ERKi-treated rats at 60- and 90-min posthypoxia ($) (all P < 0.05). NOXi-treated rats showed significant pLTF versus baseline, time control-treated rats, and MEK/ERKi-treated rats at all time-points post-AIH (all P < 0.05). E: Frequency (change from baseline; bursts/min) in vehicle-, NOX inhibitor-, and MEK/ERK inhibitor-treated rats after AIH, and time control-treated rats. Small frequency LTF in NOXi-treated rats is observed at all time points post-AIH versus MEK/ERKi-treated rats ($), but only at 30- and 90-min post-AIH versus baseline (+) and time control-treated rats (*) (P < 0.05). MEK/ERKi-treated rats had decreased frequency LTF compared with time control-treated rats (*) and baseline (+) at 15-min post-AIH and compared with vehicle-treated rats at 15- and 30-min post-AIH (#) (all P < 0.05). A two-way ANOVA with repeated measures design followed by a Tukey post hoc test was utilized. Values are means ± SE. AIH, acute intermittent hypoxia; LTF, long-term facilitation; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; pLTF, phrenic long-term facilitation; sAIH, severe AIH. The n for each group is as follows: vehicle 2, n = 10; UO126, n = 5; apocynin, n = 5; time control, n = 7.

sAIH-Induced pLTF Is Marginally Constrained by NADPH Oxidase Activity

To test the hypothesis an enzyme shown to inhibit S pathway-induced pMF also inhibits sAIH-induced pLTF, we studied sAIH-induced pLTF following spinal NADPH oxidase inhibition with apocynin. Typical phrenic neurograms during experimental protocols for each treatment group are illustrated in Fig. 3, A–C. Vehicle-treated rats exhibited similar sAIH-induced pLTF as in the first vehicle-treated group of this study (Fig. 2), and as observed in vehicle-treated rats in a previous publication (15); pLTF was 51 ± 10%, 53 ± 8%, 76 ± 9%, and 80 ± 8% at 15, 30, 60, and 90 min, respectively [P < 0.001 at all time points vs. baseline, and rats treated as time-controls at all time-points post-AIH (P = 0.017 at 15, P = 0.007 at 30, and P < 0.001 at 60 and 90 min); Fig. 3, A and D]. As predicted, sAIH-induced pLTF was present and increased following NADPH oxidase inhibition (pLTF was 94 ± 49%, 85 ± 35%, 105 ± 31%, and 118 ± 32% at 15, 30, 60, and 90 min, respectively; P < 0.001 at all time points vs. baseline and rats treated as time controls; P = 0.001 at 15, P = 0.003 at 30, and P < 0.001 at 60 and 90 min vs. rats treated with the MEK/ERK inhibitor; and P = 0.089 at 15, P = 0.271 at 30, P = 0.386 at 60, and P = 0.156 at 90 min vs. vehicle-treated rats; Fig. 3, A-D).

Again, frequency LTF was small and, consistent with the first set of studies (Fig. 2); frequency in NADPH oxidase inhibitor treated rats was significantly elevated from rats treated with the MEK/ERK inhibitor at all time points (P < 0.001 at 15-min and 30-min, P = 0.040 at 60-min, and P = 0.014 at 90-min posthypoxia), from baseline as well as versus time control rats at 30- (P = 0.022 vs. baseline; P = 0.036 vs. time control) and 90 (P = 0.023 vs. baseline; P = 0.006 vs. time control)-min post-AIH (Fig. 3E). Frequency was −1 ± 1, 2 ± 1, 1 ± 1, and 3 ± 1 bursts/min at 15, 30, 60, and 90 min, respectively in vehicle-treated rats, and 4 ± 2, 8 ± 2, 7 ± 2, and 8 ± 1 at 15, 30, 60 and 90 min, respectively in rats treated with the NOXi (Fig. 3E).

Summary of Effects on sAIH-Induced pLTF

In summary, at 90-min post-AIH, apocynin-treated rats had greater nerve burst amplitude LTF versus rats treated with inhibitors for PI3K/Akt, EPAC, PKA, and MEK/ERK [all P < 0.001; marginally significant for apocynin vs. vehicle-treated rats (P = 0.057; Fig. 4A), and greater nerve burst frequency LTF vs. rats treated with inhibitors for PI3K/Akt (P = 0.009), PKA (P = 0.022), and MEK/ERK (P = 0.004) (Fig. 4B)]. Combined vehicle-treated rats exposed to sAIH (Figs. 2 and 3, n = 20) had significantly larger nerve burst amplitudes versus rats treated as time controls (i.e., no AIH; combined group from Figs. 2 and 3, n = 17; P < 0.001), or rats pretreated with inhibitors of EPAC (P < 0.001), PI3K/Akt (P < 0.001), PKA (P = 0.002), or MEK/ERK MAPK (P < 0.001) (P < 0.05; Fig. 4A).

Figure 4.

Direct comparison of the change in burst amplitude (percent baseline) and burst frequency (bursts/min) of the phrenic nerve following sAIH at 90-min posthypoxia. A and B: change in phrenic burst amplitude (A) or burst frequency (B) following pretreatment with vehicle, EPAC inhibitor (EPACi), PI3K/Akt inhibitor (PI3K/Akti), PKA inhibitor (PKAi), MEK/ERK inhibitor (MEK/ERKi), NADPH oxidase inhibitor (NOXi), or rats treated as time controls. A: phrenic amplitude after AIH at 90-min posthypoxia is significantly higher in vehicle-treated rats and NOXi rats compared with all other groups B: phrenic burst frequency after AIH at 90-min posthypoxia is significantly higher in NOXi rats compared time control-treated rats, PKAi-treated rats, MEK/ERKi-treated rats, and PI3K/Akti-treated rats. * vs. time control-treated rats, < vs. PKAi-treated rats, # vs. vehicle-treated rats, @ vs. MEK/ERKi-treated rats, $ vs. EPACi-treated rats, and & vs. PI3K/Akti-treated rats; all P < 0.05. A one-way ANOVA followed by a Tukey post hoc test was utilized; values are means ± SE. AIH, acute intermittent hypoxia; EPAC, exchange protein activated by cyclic AMP; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; sAIH, severe AIH; TC, time control.

Drug Effects on Progressive Augmentation of the Short-Term Hypoxic Phrenic Response

Progressive augmentation is a progressive increase in the magnitude of the short-term hypoxic ventilatory (or phrenic) response from the initial to the final episode of an intermittent hypoxia protocol and has been previously described following sAIH (15, 36, 37). Here, we verified that sAIH-treated rats exhibit progressive augmentation (Fig. 5). Specifically, progressive augmentation was present in both vehicle-treated rat groups (vehicle 1: HX1 vs. HX2 and HX3, P < 0.001, and HX2 vs. HX3, P = 0.025; vehicle 2: HX1 vs. HX2 and HX3, P < 0.001, and HX2 vs. HX3, P = 0.001), PKA inhibitor-treated rats (HX1 vs. HX2, P = 0.012, and HX1 vs. HX3, P < 0.001), PI3K/Akt inhibitor-treated rats (HX1 vs. HX2, P = 0.017, and HX1 vs. HX3, P < 0.001), EPAC inhibitor-treated rats (HX1 vs. HX2, P = 0.004, and HX1 vs. HX3, P < 0.001), NADPH oxidase inhibitor-treated rats (HX1 vs. HX3, P = 0.038), and MEK/ERK inhibitor-treated rats (HX1 vs. HX3, P = 0.007; Fig. 5). All episodes were significantly different from time control rats (HX1: vehicle 1, P < 0.001; PKA inhibitor, P < 0.001; PI3K/Akt inhibitor, P = 0.008; EPAC inhibitor, P = 0.009; vehicle 2, P = 0.030; NADPH oxidase inhibitor, P = 0.004; MEK/ERK inhibitor, P = 0.049; HX2: vehicle 1, P < 0.001; PKA inhibitor, P < 0.001; PI3K/Akti, P = 0.001; EPAC inhibitor, P < 0.001; vehicle 2, P = 0.027; NADPH oxidase inhibitor, P = 0.003; MEK/ERK inhibitor, P = 0.048; HX3: vehicle 1, P < 0.001; PKA inhibitor, P < 0.001; PI3K/Akt inhibitor, P < 0.001; EPAC inhibitor, P < 0.001; vehicle 2, P = 0.012; NADPH oxidase inhibitor, P = 0.003; MEK/ERK inhibitor, P = 0.039; Fig. 5). In time control rats without AIH, there were no detectable increases in phrenic activity at times corresponding to hypoxic episodes (P > 0.05; Fig. 5).

Figure 5.

Short-term hypoxic phrenic response. A and B: progressive augmentation of the short-term hypoxic phrenic response was observed in all sAIH-treated rats including vehicle-, EPAC inhibitor (EPACi)-, PI3K/Akt inhibitor (PI3K/Akti)-, PKA inhibitor-, MEK/ERK inhibitor (MEK/ERKi)-, and NADPH oxidase inhibitor (NOXi)-treated rats. 1, 2, and 3 represent hypoxic episodes for sAIH-treated rats or normoxic episodes for time control-treated rats. In A and B, the hypoxic response progressively increases with each episode (* indicates significantly different than 1; and + indicated significantly different from 3; both P < 0.05). There were no changes in time control-treated rats, and all treatment groups were significantly different from time control-treated rats (P < 0.05). A two-way ANOVA with repeated measures design followed by a Tukey post hoc test was utilized; values are means ± SE. AIH, acute intermittent hypoxia; EPAC, exchange protein activated by cyclic AMP; PI3K/Akt, phosphatidylinositol 3-kinase/Akt; sAIH, severe AIH; TC, time control.

DISCUSSION

We demonstrate that enzymes necessary for the S pathway to pMF (defined by spinal receptor agonist or enzyme activators) also underlie sAIH-induced pLTF. Specifically, spinal EPAC and PI3K/Akt inhibition blocked sAIH-induced pLTF. Further, inhibition of NADPH oxidase, an enzyme regulating the S pathway via cross-talk inhibition, augments sAIH-induced pLTF as predicted. On the other hand, the picture is complex, since inhibition of a divergent signaling pathway (PKA), or an enzyme required for the Q pathway, but not the S pathway (MEK/ERK MAPK), either impaired (PKAi) or abolished (MEK/ERKi) sAIH-induced pLTF. We cannot explain these surprising findings at this time. However, AIH elicits multiple mechanisms that compete with one another in regulating phrenic motor plasticity (6, 16, 25, 38). The data presented here point out that we have a lot to learn concerning the regulation of phrenic motor plasticity but confirm that the severity of hypoxia during AIH is a key factor in its regulation.

Progressive Augmentation Was Not Affected by Any Drug Treatment

Unlike moderate AIH, sAIH elicits consistent progressive augmentation of the short-term hypoxic phrenic response [(15); Fig. 5, this study]. The mechanism of progressive augmentation remains unknown, but it is not dependent on serotonergic receptor activation (37, 39). We now show that sAIH-induced progressive augmentation does not require EPAC, PI3K/Akt, PKA, MEK/ERK, or NADPH oxidase activity (Fig. 5). The still unknown mechanism(s) of progressive augmentation during sAIH requires further investigation.

sAIH-Induced pLTF Requires Spinal EPAC and PI3K/Akt Activity

Unlike moderate AIH, which is serotonin-dependent and adenosine constrained (27), sAIH-induced pLTF is A_2A receptor dependent (15). However, signaling mechanisms downstream from A_2A receptors have only been inferred from studies of pMF following intraspinal injections of pharmacological agents and/or siRNAs following A_2A (e.g., 1, 20) or 5-HT₇ receptor agonists (2, 18; see Fig. 1). This is the first study verifying that two key downstream signaling kinases necessary for the S pathway to pMF, EPAC and PI3K/Akt, are also required for sAIH-induced pLTF (Figs. 2 and 4). The requirement for EPAC signaling is consistent with reports that EPAC, not PKA inhibition, abolishes 5-HT₇ receptor-induced pMF (18), whereas downstream EPAC signaling likely requires PI3K/Akt (1) and mammalian target of rapamycin (mTOR), which is indirectly activated by PI3K/Akt (40–43) and is known to be required for sAIH-induced pLTF (44).

PKA Regulates sAIH-Induced pLTF

PKA is not required for 5HT7 receptor-induced pMF (18) but does constrain moderate AIH-induced pLTF (23, 28). On the other hand, paradoxically, PKA activation alone is sufficient to elicit pMF by an unknown mechanism (23). Here, we report the unanticipated finding that spinal PKA inhibition reduces, but does not abolish, sAIH-induced pLTF (Figs. 2 and 4). Thus, 1) PKA may play an unknown role in enabling sAIH-induced pLTF, but not 5-HT_2A receptor-induced pMF (23); 2) A_2A and 5-HT₇ receptors differ in their downstream signaling, similar to 5-HT_2A versus 5-HT_2B receptors (21); and/or 3) PKA-dependent effects are “dose-dependent,” with higher PKA activity capable of eliciting pMF independently from the S or Q pathways (23).

NADPH Oxidase Activity Marginally Constrains sAIH-Induced pLTF

Moderate AIH-induced pLTF (i.e., predominant Q pathway) requires NADPH oxidase activity and reactive oxygen species formation (30, 31). Similarly, 5-HT_2B, but not 5-HT_2A receptor-induced pMF requires NADPH oxidase activity (5). On the flip side, 5-HT_2B receptors constrain 5-HT₇ receptor-induced pMF via NADPH oxidase activity (19). It has not been shown that similar, NADPH oxidase activity-dependent cross-talk inhibition constrains sAIH-induced pLTF.

Complex relationships such as the seeming dual role of NADPH oxidase in regulating phrenic motor plasticity must be interpreted in the context of the interactive, cellular network regulating phrenic motor plasticity. Our observation that NOX inhibition marginally increases sAIH-induced pLTF is reminiscent of our prior report that NADPH-dependent ROS formation underlies cross-talk inhibition from 5-HT_2B to 5-HT₇ receptors (19); similar S to Q cross-talk inhibition between A_2A receptor-induced pMF could account for the tendency of apocynin to increase pLTF in this study (Figs. 2 and 4). On the other hand, NADPH oxidase activity is necessary for 5-HT_2B receptor-induced pMF. The parsimonious explanation for this bidirectional (and seemingly paradoxical relationship) is that 5-HT_2B receptors stimulate NADPH oxidase activity, suppressing adenyl cyclase activity (45, 46), thereby minimizing cAMP formation and PKA activity. It follows that NOX inhibition would reverse these effects, increasing PKA activity and suppressing the Q pathway to pMF. In this way, NOX inhibition or ROS scavenging would block pMF elicited by intraspinal serotonin (29), 5-HT_2B receptor activation (5), and moderate AIH (30).

sAIH-Induced pLTF Requires Spinal MEK/ERK MAPK Activity

Although spinal MEK/ERK inhibition with UO126 blocks moderate AIH-induced pLTF (32), we did not anticipate that it would also block sAIH-induced pLTF (Figs. 3 and 4). We can only speculate concerning mechanisms of MEK/ERK involvement in sAIH-induced pLTF at this time. However, EPAC enhances hippocampal long-term potentiation by a mechanism that also requires ERK MAPK activity (47). Our data suggest that MEK/ERK activity is necessary for both Q and S pathway-dependent pMF and pLTF.

Possible MEK/ERK involvement in the S pathway is suggested by reports that ERK both activates and inactivates different phosphodiesterase-4 isoforms (48). Thus, it is possible that ERK enables sAIH-induced pLTF by indirect regulation of cAMP formation by, for example, attenuating phosphodiesterase-4 activity. ERK phosphorylation inactivates long phosphodiesterase-4 isoforms, which would increase cAMP levels (48–51), driving PKA and EPAC activation. In the context of this study, ERK inactivation of long phosphodiesterase 4 isoforms would drive the S pathway to pMF and sAIH-induced pLTF. Consequently, MEK/ERK inhibition would reduce cAMP levels, EPAC activation, and the driving force for S pathway-dependent pMF and pLTF. Even reduced cAMP levels may still be sufficient to activate PKA and suppress 5-HT₂ receptor-induced pLTF expected with sAIH since cAMP levels needed to activate PKA (and cross-talk inhibition) are well below those required for EPAC and S pathway activation [see (23)]. Possibilities concerning ERK-phosphodiesterase-4 interactions in regulating sAIH-induced pLTF warrant further investigation.

Significance and Conclusions

In conclusion, the S pathway to pMF is the predominant driver of sAIH-induced pLTF since this A_2A receptor-dependent form of phrenic motor plasticity requires EPAC and PI3K/Akt activity, similar to 5-HT₇ receptor-induced pMF. Further, NADPH oxidase, a cross-talk inhibitor of 5-HT₇ receptor-induced pMF, marginally constrains sAIH-induced pLTF, presumably via ROS formation. On the other hand, surprising observations that MEK/ERK activity is necessary for, and PKA activity modulates, sAIH-induced pLTF suggest there is more to learn about this important model of phrenic motor plasticity. Although spinal PKA activity is sufficient to elicit pMF in rats (23), mechanisms of this plasticity are unknown. We speculate that the MEK/ERK pathway suppresses phosphodiesterases, enabling adequate cAMP formation and accumulation to activate the high-threshold cAMP-activated kinase, EPAC. Thus, MEK/ERK inhibition would undermine the ability to initiate the S pathway to phrenic motor facilitation.

Although the complex, interacting mechanisms of phrenic motor plasticity are difficult to explain, they may offer physiological advantages, such as enabling appropriate responses to: 1) physiological challenges that differ in severity (e.g., mild to severe hypoxia) or duration (acute through lifelong), and/or 2) the ability to express emergent properties such as pattern sensitivity (25) and/or accumulation of functional benefits (i.e., metaplasticity; 17, 52). This complexity also offers potential molecular targets, as we attempt to optimize the therapeutic potential of intermittent hypoxia to elicit meaning functional recovery of breathing ability and nonrespiratory motor behaviors in clinical disorders such as spinal cord injury and motor neuron diseases (8, 53–55).

GRANTS

This work was supported by the Francis Families Foundation (N. L. Nichols), NIH K99/R00 HL119606 (N. L. Nichols), NIH R01 HL080209 (G. S. Mitchell) and R01 HL148030 (G. S. Mitchell).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

N.L.N. and G.S.M. conceived and designed research; N.L.N. performed experiments; N.L.N. analyzed data; N.L.N. and G.S.M. interpreted results of experiments; N.L.N. prepared figures; N.L.N. drafted manuscript; N.L.N. and G.S.M. edited and revised manuscript; N.L.N. and G.S.M. approved final version of manuscript.