Ampakine pretreatment enables a single hypoxic episode to produce phrenic motor facilitation with no added benefit of additional episodes

Abstract

Repeated short episodes of hypoxia produce a sustained increase in phrenic nerve output lasting well beyond acute intermittent hypoxia (AIH) exposure (i.e., phrenic long-term facilitation; pLTF). Pretreatment with ampakines, drugs which allosterically modulate AMPA receptors, enables a single brief episode of hypoxia to produce pLTF, lasting up to 90 min after hypoxia. Here, we tested the hypothesis that ampakine pretreatment would enhance the magnitude of pLTF evoked by repeated bouts of hypoxia. Phrenic nerve output was recorded in urethane-anesthetized, mechanically ventilated, and vagotomized adult male Sprague–Dawley rats. Initial experiments demonstrated that ampakine CX717 (15 mg/kg iv) caused an acute increase in phrenic nerve inspiratory burst amplitude reaching 70 ± 48% baseline (BL) after 2 min (P = 0.01). This increased bursting was not sustained (2 ± 32% BL at 60 min, P = 0.9). When CX717 was delivered 2 min before a single episode of isocapnic hypoxia (5 min, $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ = 44 ± 9 mmHg), facilitation of phrenic nerve burst amplitude occurred (96 ± 62% BL at 60 min, P < 0.001). However, when CX717 was given 2 min before three, 5-min hypoxic episodes ($\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ = 45 ± 6 mmHg) pLTF was attenuated and did not reach statistical significance (24 ± 29% BL, P = 0.08). In the absence of CX717 pretreatment, pLTF was observed after three (74 ± 33% BL at 60 min, P < 0.001) but not one episode of hypoxia (1 ± 8% BL at 60 min, P = 0.9). We conclude that pLTF is not enhanced when ampakine pretreatment is followed by repeated bouts of hypoxia. Rather, the combination of ampakine and a single hypoxic episode appears to be ideal for producing sustained increase in phrenic motor output.

NEW & NOTEWORTHY Pretreatment with ampakine CX717 created conditions that enabled an acute bout of moderate hypoxia to evoke phrenic motor facilitation, but this response was not observed when ampakine pretreatment was followed by intermittent hypoxia. Thus, in anesthetized and spinal intact rats, the combination of ampakine and one bout of hypoxia appears ideal for triggering respiratory neuroplasticity.

INTRODUCTION

Glutamatergic synaptic transmission plays an important role in regulation of respiratory motor output. Glutamate receptors including α-amino-3-hydro-5-methyl-4-isoxazolepropionate acid (AMPA), N-methyl-d-aspartate (NMDA), and kainite are expressed on respiratory motoneurons as well as premotor respiratory control neurons. In the medullary region, AMPA receptors (AMPARs) are critical for respiratory rhythmogenesis (1, 2), and activation of AMPA receptors on respiratory motoneurons is a primary contributor to inspiratory depolarization (3). Thus, excitatory currents mediated by AMPARs are a fundamental aspect of how breathing is regulated.

Ampakines are synthetic class of small molecule drugs that act as allosteric modulators of AMPA receptors to enhance excitatory glutamatergic currents (4, 5). Ampakines can stimulate breathing in rodent models of neuromuscular disorders (6, 7) and after opioid overdose (8–11). Ampakines can also enhance expression of synaptic long-term potentiation (LTP) (12, 13) as well as neuronal expression of brain-derived neurotrophic factor (BDNF) (14, 15).

Phrenic long-term facilitation (LTF) after acute intermittent hypoxia (AIH) is a well-studied example of neuroplasticity in the respiratory neural control system (16). Defined as a sustained increase in phrenic motor output after AIH, phrenic LTF is triggered by molecular cascades in the spinal cord that ultimately lead to enhanced glutamatergic excitation of phrenic motoneurons (17). Prior work indicates that ampakines can enhance hypoxia-induced respiratory neuroplasticity (18–20). For example, pretreatment with ampakine CX717 enabled a single 5-min episode of hypoxia to produce a sustained LTF of phrenic nerve bursting in anesthetized rats, lasting up to 90 min after the hypoxic episode (20). This is a noteworthy observation as the mechanisms underlying phrenic LTF are thought to be uniquely activated only by repeated hypoxia exposures (21, 22). In a study of anesthetized mice, Turner et al. (19) reported that ampakine pretreatment also enhanced LTF of hypoglossal inspiratory bursting after three, 1-min bouts of hypoxia. Building upon the foundation of these prior studies, here we explored the impact of ampakine pretreatment on expression of phrenic LTF following repeated bouts of hypoxia. Rats were pretreated with ampakine CX717 (15 mg/kg), and then exposed to one or three 5-min bouts of moderate hypoxia (arterial partial pressure of O₂ 35–55 mmHg). The hypoxia pattern (i.e., 3 × 5 min, moderate hypoxia) was chosen because most of the foundational studies of phrenic LTF have used that paradigm (23–25). We hypothesized that following ampakine pretreatment, the expression of phrenic LTF would be enhanced in rats exposed to three versus one bout of hypoxia. Additional studies were conducted to compare the magnitude of phrenic LTF evoked by one versus three 5-min bouts of hypoxia, but without ampakine treatment. These latter studies were done to confirm earlier reports that phrenic LTF requires episodic hypoxia, and to provide a comparison to the ampakine data.

MATERIALS AND METHODS

Animals

All experiments were conducted in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals (2011) and were approved by the Animal Care and Use Committee at the University of Florida (Protocol ID 201807438). A total of 32 adult male Sprague–Dawley rats (Colony No. 208 A, Envigo), weighing 333 ± 16 g and 12 ± 1 wk old were studied. Sample sizes for each experimental group are provided in the figure legends. All animals were maintained on a 12-h light-dark cycle and had access to food and water ad libitum.

Surgical Procedure

Isoflurane (3%) mixed with 100% O₂ was used for the initial induction of anesthesia in a closed chamber. The rat was then transferred to a heating pad on a surgical table where anesthesia was maintained using a nose cone. Rectal temperature was maintained at 37 ± 1°C. After attaining a surgical plane of anesthesia as evidenced by loss of the pedal withdrawal and corneal reflexes, rats were tracheotomized and ventilated (VentElite, model 55-7040; Harvard Apparatus Inc.) with an inspiratory mixture of 50% O₂, 1% CO₂, balanced with N₂. Ventilator frequency was maintained at 65–70 breaths/min and tidal volume was set at 7 mL/kg. Tracheal pressure was continuously monitored and the lungs were periodically hyperinflated by brief occlusion of the expiratory line to reduce alveolar collapse. A final hyperinflation was done just before the start of baseline recordings. Bilateral vagotomy was performed to prevent entrainment of the phrenic efferent recording with the ventilator. End-tidal CO₂ was monitored throughout the surgery and experimental protocol (Capnogard end-tidal CO₂ monitor, Novametrix).

The tail vein was cannulated to enable intravenous fluid infusion. Animals were slowly converted from isoflurane to urethane (2.1 g/kg, 6 mL/h) anesthesia while checking adequate depth of anesthesia by withdrawal reflex to toe pinch. After completion of urethane anesthesia, a 1:4 solution of 8.4% sodium bicarbonate mixed with lactated Ringer’s was administered (2 mL/h) via the tail vein catheter to maintain acid-base balance. Neuromuscular blockade was accomplished using pancuronium bromide (3 mg/kg iv, Sigma-Aldrich, St Louis) to prevent movements and electromyogram (EMG) activity from interfering with phrenic nerve recordings. The femoral artery was cannulated using polyethylene tubing (PE 50; Intramedic) which was then connected to a transducer amplifier (TA-100, CWE) to measure arterial blood pressure. The arterial line was also used for withdrawal of blood samples for measurement of blood-gas chemistry including partial pressure of CO₂ ($\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$), O₂ ($\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$), pH, and base excess using an ABL 90 Flex, Radiometer (Copenhagen, Denmark).

The phrenic nerve was isolated using a dorsal surgical approach, cut distally, and de-sheathed. In the majority of the experiments, recordings were made from left phrenic nerve. However, in n = 3 cases, recordings were made from right phrenic nerve. The distal end was suctioned into a custom-made glass electrode (102-mm long, inner and outer diameter of 1.12 mm and 2.0 mm, respectively, WPI, 1B200F-4) filled with 0.9% saline. Phrenic nerve activity was amplified (×10 kHz) using differential AC amplifier (Model 1700, A-M systems, Everett, WA), band-pass filtered (100 Hz–3 kHz), digitized (16 bit, 25,000 samples/channel, Power 1401, CED), and integrated with (0.05 s time constant) using Spike2 software (Cambridge Electronic Design, UK).

Drugs

Ampakine CX717 in powder form was provided by RespireRx. The powder was dissolved in a 10% solution of 2-hydroxypropyl-β-cyclodextrin (HPCD, Sigma) at a concentration of 5 mg/mL and stored in 1.5 mL aliquots at −20°C. A fresh aliquot was thawed and brought to room temperature at the beginning of every experiment. The ampakine solution was administered via the tail vein at 15 mg/kg (dose calculated based on body weight on the day of experiment). The selection of dose was based on our prior report in which CX717 enabled a single episode of hypoxia to produce a sustained increase in phrenic motor output in urethane anesthetized rats (20). Vehicle-treated control groups received HPCD (10% in 0.45% saline solution) in equal volume to the ampakine dose.

Experimental Protocols

At the start of all experiments, the inspired concentration of O₂ and CO₂ were 50% and 1% respectively (balanced with N₂). The inspired levels of CO₂ or rate of ventilation was regulated to maintain the end tidal CO₂ at 40–42 mmHg. After stable baseline phrenic nerve activity was recorded for 15–20 min and obtaining arterial CO₂ partial pressure ($\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$) measures in the range of 42–45 mmHg, either ampakine or HPCD was delivered via the tail vein. Subsequent blood gas samples were taken at 20-, 40-, and 60-min postadministration. In experiments where hypoxia was delivered, an additional blood sample was taken during the 4th min of hypoxia. The baseline blood gas value was considered as reference for subsequent samples after intravenous administration or hypoxia. Care was taken to make sure that $\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$ values stayed within ±2 mmHg from the baseline values and $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ was maintained ≥150 mmHg. If $\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$ did not fit this range, changes to inspired fraction of CO₂ or rate of ventilation were made and additional blood samples were analyzed. If $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$ values dropped below 150 mmHg, the inspired fraction of O₂ was increased by 10% and another blood sample was analyzed after 5 min. If stable blood gas values could not be obtained, the preparation was excluded from analysis (n = 1). At the conclusion of the phrenic nerve recording experiments, a maximal chemoreceptor stimulation was induced by briefly stopping the mechanical ventilator. This was maintained for ∼20 s, until the phrenic bursting transformed to an irregular “gasping-like” intense discharge pattern and later, ceased completely. The ventilator was then switched back on and rats were euthanized via urethane overdose. Preparations were excluded if the response to maximal challenge was less than the previous response to acute hypoxia (n = 3), as this was taken as an indication that the nerve-electrode contact had degraded over the course of the experimental protocol.

Data Collection and Analyses

Data were collected using Spike2 software (v. 8.01 and 10.01, CED). Phrenic nerve recordings were analyzed using custom written MATLAB code that is available upon request (MathWorks, R2019a). Integrated phrenic nerve burst amplitude was normalized to baseline output and reported as percent (%) change from baseline (BL). Inspiratory phrenic burst frequency was expressed as breaths per minute. All responses were averaged over 3-min duration immediately before withdrawal of arterial blood samples.

All statistical analyses were performed using SigmaPlot 14 (Systat Software). Two-way repeated-measures analysis of variance (ANOVA) was used to statistically compare phrenic amplitude, respiratory rate, heart rate, and mean arterial pressure (MAP) values between two groups at multiple time points. Changes in arterial blood gases and pressure within each group were analyzed using one-way ANOVA. Wherever necessary, multiple comparisons were made using Student–Newman–Keuls post hoc analysis. Data were considered statistically significant if P ≤ 0.05. The mean data are presented along with one standard deviation.

RESULTS

Body weight, age, arterial blood pressure, and arterial blood gas measurements were similar between experimental groups, and are presented in Table 1.

Table 1.

Arterial blood gases, pressure, body weight, and age

Experimental Group	$\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$, mmHg	$\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, mmHg	SBEc, mmol/L	pH	Systolic BP, mmHg	Diastolic BP, mmHg	Weight, g	Age, wk
CX717 alone n = 4							328 ± 24	12 ± 1
Baseline	44.6 ± 0.9	239 ± 14.6	−0.2 ± 1.9	7.4 ± 0.03	158.3 ± 24.3	88.8 ± 22.5
20 min	44.9 ± 2.2	218 ± 14.9	−0.4 ± 1.2	7.4 ± 0.03	156.3 ± 15	89.2 ± 15.7
40 min	44.6 ± 1.9	214.2 ± 12.1	−0.3 ± 1	7.4 ± 0.02	162.4 ± 9.3	88.4 ± 13.5
60 min	44.4 ± 1.4	213 ± 24.5	0.3 ± 0.9	7.4 ± 0.02	165 ± 8.8	90.2 ± 14.9
HPCD alone n = 4							332 ± 28	11 ± 1
Baseline	43.6 ± 1.3	243.8 ± 15.9	−0.7 ± 1.1	7.4 ± 0.02	163.1 ± 11.8	89.6 ± 14.4
20 min	44.3 ± 1.8	215.3 ± 14.9*	0.1 ± 0.5	7.4 ± 0.02	164 ± 2.1	90.5 ± 7.9
40 min	45.6 ± 2	213.3 ± 13.1*	0.6 ± 0.7	7.4 ± 0.02	166.1 ± 4.9	92.9 ± 9.8
60 min	44.2 ± 1.7	209.5 ± 14.7*	0.8 ± 0.7	7.4 ± 0.01	167 ± 6.6	94.3 ± 4.6
CX717 + 1× hypoxia n = 8							330 ± 13	13 ± 1
Baseline	43.6 ± 1.9	237.8 ± 16.2	0.4 ± 1.5	7.4 ± 0.02	164.9 ± 11.8	94.8 ± 8.8
Hypoxia	45.8 ± 3.2	44.1 ± 8.5*	−0.6 ± 1.5	7.4 ± 0.03	143.9 ± 16.9	68.3 ± 9.2*
20 min	44.9 ± 1.8	159.3 ± 33.3*	0.2 ± 1.2	7.4 ± 0.02	168.2 ± 14.3	94.7 ± 10.6
40 min	44.7 ± 1.8	193.6 ± 24.9*	0.3 ± 1.5	7.4 ± 0.02	162.3 ± 18.6	90.2 ± 13.7
60 min	44.3 ± 1.6	197.6 ± 26.1*	0.3 ± 1.6	7.4 ± 0.02	161.5 ± 18.3	89.2 ± 14.7
CX717 + 3× hypoxia n = 8							340 ± 9	13 ± 1
Baseline	43.5 ± 1	246.6 ± 9.7	−0.1 ± 1	7.4 ± 0.02	173 ± 19.8	101.8 ± 12.1
Hypoxia	43.9 ± 1.7	45.3 ± 5.7*	−0.6 ± 1.3	7.4 ± 0.02	148.8 ± 21.8*	73.5 ± 13.4*
20 min	44.3 ± 2.1	139.6 ± 29.2*	−0.5 ± 1.5	7.4 ± 0.02	172.8 ± 14.7	98.1 ± 7.1
40 min	45.3 ± 1.7	192.8 ± 27.3*	0.1 ± 1.8	7.4 ± 0.02	172.9 ± 15.6	97.4 ± 8.9
60 min	44.4 ± 1.1	206 ± 23*	0.7 ± 1.4	7.4 ± 0.02	174.6 ± 13.8	99 ± 7.7
1× Hypoxia n = 4							334 ± 2	12 ± 1
Baseline	43.6 ± 0.7	252.8 ± 13.6	−1.2 ± 0.8	7.4 ± 0.01	152.9 ± 14.6	85.8 ± 12.3
Hypoxia	43.6 ± 2.8	48.1 ± 6.5*	−2.6 ± 0.7	7.3 ± 0.01	117 ± 32.7*	50.5 ± 24.7*
20 min	44.3 ± 1.5	149.3 ± 13.1*	−1.8 ± 1.5	7.3 ± 0.01	160.6 ± 14.6	88.9 ± 8
40 min	44 ± 0.8	222.3 ± 33.6	−1.3 ± 1.5	7.4 ± 0.01	154.4 ± 11.5	83.2 ± 7.4
60 min	43.8 ± 0.8	222.5 ± 15.4	−1.1 ± 1.6	7.4 ± 0.02	154.7 ± 8.7	82.13 ± 6.7
3× Hypoxia n = 4							329 ± 10	11 ± 1
Baseline	42.7 ± 0.6	272.5 ± 18.9	−1.6 ± 1.5	7.4 ± 0.01	150.8 ± 9.2	74.9 ± 12
Hypoxia	43.1 ± 1.6	47.4 ± 9.3*	−2.6 ± 0.8	7.3 ± 0.02	112.2 ± 14.9*	45.4 ± 7.7*
20 min	43.8 ± 1.2	139.3 ± 11.3*	−1.3 ± 0.9	7.4 ± 0.01	164 ± 14.5	87.5 ± 14.4
40 min	43.9 ± 0.3	220.5 ± 28.3*	−0.7 ± 1	7.4 ± 0.01	150.6 ± 22.3	77.3 ± 13.8
60 min	42.7 ± 0.9	232.5 ± 25.3*	−1.3 ± 1.6	7.4 ± 0.02	148.1 ± 18.3	73.3 ± 14.2

Mean blood gas values are shown for baseline, hypoxia, and the 20-, 40-, and 60-min posthypoxia time points. Data are presented as means ± 1 SD. 1×, 1 episode of 5-min hypoxia; 3×, 3 episodes of 5-min hypoxia; HPCD, 2-hydroxypropyl-β-cyclodextrin; $\text{P}{\text{a}}_{\text{CO}}{}_{{}_2}$, arterial CO₂ pressure; $\text{P}{\text{a}}_{\text{O}}{}_{{}_2}$, arterial O₂ pressure; SBEc, standard base excess.

Significant difference compared with baseline (*P < 0.05) within that particular group.

Intravenous Delivery of Ampakine CX717 Transiently Increases Phrenic Motor Output

Representative examples of phrenic nerve activity before and after intravenous administration of CX717 or the HPCD vehicle solution are shown in Fig. 1. Infusion of CX717 caused an immediate increase in phrenic burst amplitude, reaching a peak within ∼2 min (Fig. 1A). Phrenic burst amplitude then gradually declined to baseline values. Administration of the HPCD solution had no discernable impact on phrenic bursting (Fig. 1B). The group mean responses to CX717 or the HPCD vehicle are provided in Fig. 1C. Analysis of phrenic burst amplitude revealed an interaction between treatment (CX717 vs. HPCD) and time postinjection [F_(3,31) = 3.96, P = 0.025]. Post hoc tests indicated that burst amplitude in the CX717 group elevated at 20-min postadministration (P < 0.001 vs. HPCD).

Figure 1.

Ampakine CX717 causes a transient increase in phrenic burst amplitude. A and B: representative phrenic neurograms along with respiratory rate and arterial blood pressure for CX717 and 2-hydroxypropyl-β-cyclodextrin (HPCD) groups. Either CX717 or HPCD was given intravenously as noted on the records. A maximal chemoreceptor activation was performed at the end of each experiment (gray trace). C: mean changes in phrenic burst amplitude, respiratory rate, heart rate, and mean arterial pressure at baseline, 20-, 40-, and 60-min postdrug administration. Data represented as means ± 1 SD; n = 4 rats per group; *P < 0.05.

No between-group differences were detected in either heart rate [treatment: F_(1,31) = 1.39, P = 0.28; time: F_(3,31) = 0.99, P = 0.41] or MAP [treatment: F_(1,31) = 0.2, P = 0.66; time: F_(3,31) = 0.22, P = 0.87] (Fig. 1C). Respiratory rate was similar between the two groups across the overall duration of the experiment [Fig. 1C; treatment: F_(1,31) = 0.34, P = 0.58; time: F_(3,31) = 0.92, P = 0.44]. However, in prior work (20) and the present study we observed a transient increase in phrenic inspiratory burst frequency upon intravenous delivery of CX717. Therefore, in a separate analysis, we quantified the acute impact of CX717 on inspiratory burst frequency, and these data are shown in Table 2. Collectively, the results of the first series of experiments confirmed that intravenous delivery of ampakine CX717, at a dose of 15 mg/kg, causes a transient stimulation of phrenic inspiratory burst amplitude. This CX717 dose had minimal impact on heart rate or mean arterial pressure, but did cause an acute increase in inspiratory bursting as shown in Fig. 1A and Table 2.

Table 2.

Acute impact of CX717 on respiratory frequency

Time	Respiratory Rate for Different Experimental Groups Receiving Intravenous CX717
	CX717 Alone, n = 4	CX717 + 1× Hypoxia, n = 8	CX717 + 3× Hypoxia, n = 8
Baseline	48.8 ± 7.9	47.9 ± 4.0	49.0 ± 4.7
20 s	51.6 ± 6.9	51.5 ± 5.7*	52.2 ± 6.1*
40 s	53.6 ± 6.3	55.2 ± 5.9**	54.8 ± 6.1**
60 s	55.5 ± 6.1	56.3 ± 6.5**	55.5 ± 5.3**
80 s	57.2 ± 6.1*	56.2 ± 6.1**	55.5 ± 4.2**
100 s	57.2 ± 5.4*	55.3 ± 5.2**	55.3 ± 3.1**
120 s	56.2 ± 5.7*	54.6 ± 4.5**	54.9 ± 2.6*

Mean respiratory frequency values are shown for baseline, 20-, 40-, 60-, 80-, 100-, and 120-s post-ampakine administration. Data are presented as means ± 1 SD. 1× hypoxia, one episode of 5-min hypoxia; 3×, three episodes of 5-min hypoxia.

Significant difference compared with baseline (*P < 0.05; **P < 0.01) within that particular group.

Phrenic Motor Facilitation is Greater When Ampakine CX717 is Followed by One versus Three Episodes of Hypoxia

Representative examples of phrenic nerve activity when CX717 was followed by exposure to hypoxia are shown in Fig. 2. Persistent increases in inspiratory phrenic burst amplitude were observed when CX717 was followed by one (Fig. 2A) or three episodes of hypoxia (Fig. 2B). However, as shown by the mean data in Fig. 2C, the sustained phrenic motor facilitation was greater in the group that received one episode of hypoxia. Statistically, phrenic burst amplitude showed an interaction between treatment (e.g., 1 vs. 3 hypoxic episodes) and time [F_(3,63) = 6.01, P = 0.002]. Post hoc tests confirmed that phrenic burst amplitude was greater in CX717 + 1× hypoxia group at 20- (P = 0.021), 40- (P = 0.003), and 60-min (P < 0.001). Respiratory rate showed an effect of time [F_(3,63) = 6.77, P < 0.001] but not treatment [F_(1,63) = 0.07, P = 0.79]. Thus, both experimental groups showed a gradual increase in the rate of inspiratory phrenic bursting following the end of hypoxia exposure (Fig. 2C). Heart rate also showed an effect of time [F_(3,63) = 4.9, P = 0.005] but not treatment [F_(1,63) = 0.0001, P = 0.99]. Mean arterial pressure was similar between the two groups throughout the experimental protocol (Fig. 2B; treatment: F_(1.63) = 2.75, P = 0.22; time: F_(3,63) = 2.17, P = 0.1).

Figure 2.

Phrenic motor facilitation following pretreatment with ampakine CX717 followed by 1 vs. 3 hypoxic episodes. A and B: representative phrenic neurograms along with respiratory rate and arterial blood pressure for CX717 + 1× and 3× hypoxia, respectively. Phrenic activity was recorded for 60 min following the hypoxic episodes, followed by a maximal chemoreceptor stimulation challenge (gray trace). Statistically significant phrenic motor facilitation was observed only in CX717 + 1× hypoxia group. C: the mean changes in phrenic burst amplitude, respiratory rate, heart rate, and mean arterial pressure at baseline, 20-, 40-, and 60-min post-hypoxia. Data are presented as means ± 1 SD; n = 8 rats per group; *P < 0.05. 1× hypoxia, one episode of hypoxia; 3× hypoxia, three episodes of hypoxia.

Hypoxia Exposures without Ampakine Treatment Confirm That Phrenic Long-Term Facilitation Requires Multiple Episodes of Hypoxia

This final series of experiments was done to verify that a single bout of hypoxia, without ampakine, does not produce phrenic LTF. Thus, the response to one versus three bouts of hypoxia was directly compared. Representative examples showing the impact of one or three episodes of hypoxia on phrenic nerve activity are provided in Fig. 3, A and B; group mean responses are shown in Fig. 3C. Evaluation of phrenic burst amplitude showed a treatment (i.e., 1 vs. 3 hypoxic episodes) × time interaction [F_(3,31) = 7.7, P = 0.002]. Post hoc tests confirmed that phrenic burst amplitude was higher in the group receiving three bouts of hypoxia at 40 (P = 0.006) and 60 min (P < 0.001) after the final hypoxia exposure. Phrenic burst rate also showed a treatment × time interaction [F_(3,31) = 3.5, P = 0.036], and burst was greater in the 3 × hypoxia group at the 40-min time point (P = 0.022 vs. 1 × hypoxia). Heart rate was similar in both groups across the duration of the experimental protocol [Fig. 3C; treatment: F_(1,31) = 0.089, P = 0.77; time: F_(3,31) = 0.39, P = 0.75]. However, there was a significant time effect in mean arterial pressure among the two groups [F_(3,31) = 4, P = 0.024] but no effect of treatment [F_(1,31) = 0.26, P = 0.62]. Thus, MAP values decline over the 60-min experiment to a similar extent in both groups.

Figure 3.

Phrenic long-term facilitation is not evoked by a single episode of hypoxia. A and B: representative phrenic neurograms along with respiratory rate and arterial blood pressure. Following baseline recordings, rats were exposed to 1 or 3 episodes of hypoxia followed by a maximum chemoreceptor challenge (gray trace). Phrenic long-term facilitation (LTF) was observed only in rats treated with 3 hypoxic episodes. C: mean changes in phrenic burst amplitude, respiratory rate, heart rate, and mean arterial pressure at baseline, 20-, 40-, and 60-min post-hypoxia. Data represented as means ± 1 SD; n = 4 rats per group; *P < 0.05.

DISCUSSION

Prior reports have established that pairing ampakines with exposure to a single brief bout of hypoxia can result in a sustained increase in inspiratory motor output, lasting well beyond the impact of either treatment alone (19, 20). Here, we confirm those prior findings and also show that pairing ampakine treatment with repeated bouts of hypoxia does not have added benefit for phrenic motor plasticity, at least with the 3 × 5 min hypoxia paradigm used here. Rather, the degree of facilitation of phrenic inspiratory burst amplitude was attenuated when ampakine pretreatment was followed by three bouts of hypoxia. Thus, the number of hypoxic episodes is a key variable when using ampakines as an adjunct therapy to boost respiratory neuroplasticity (18).

Ampakine Mechanism of Action

Ampakines are a class of compounds which enhance AMPA-mediated glutamatergic neurotransmission (5, 26). These compounds are not AMPA receptor agonists, but rather are allosteric modulators of their activity (4, 27, 28). Early-stage clinical trials related to cognitive and affective disorders demonstrated that ampakines are metabolically stable, readily cross the blood brain barrier, and produce minimal side effects at therapeutic doses (9, 29–32).

In regards to breathing, ampakines can act on premotor respiratory rhythm generating neurons as well as respiratory motoneurons (8, 10, 11, 33, 34). The impact of ampakines on brainstem rhythm-related circuits is most potent when the overall level of “respiratory drive” is diminished; under such conditions, ampakines produce increases in respiratory rate, with minimal impact on burst amplitude (11, 33). A direct impact on motor neurons was shown by Lorier et al. (8), who demonstrated that ampakines, including CX717 as used in the current study, can increase postsynaptic currents recorded from hypoglossal motor neurons. Specifically, the area and duration of AMPA evoked postsynaptic currents was enhanced when CX717 was focally applied to hypoglossal motoneurons in a brainstem slice preparation (8).

Hypoxia, Phrenic Motor Facilitation, and Ampakines

Acting primarily through activation of carotid chemoafferent neurons, hypoxia can be a powerful respiratory stimulus. Many published reports show that following exposure to single bout of moderate hypoxia, phrenic motor output has shown short-term potentiation that rapidly decays back to prehypoxia baseline values (for review see Ref. 35; see also Fig. 3B). In contrast, exposure to moderate acute intermittent hypoxia in the anesthetized rat, as used here, typically evokes phrenic LTF in the range of 40%–80% increase from baseline burst amplitude (16). We validated that this occurred in our experimental paradigm when intermittent hypoxia was administered without ampakine pretreatment (e.g., Fig. 3A). Thus, hypoxia-evoked phrenic LTF shows a “pattern sensitivity,” and requires exposure to repeated bouts of hypoxia (21, 22, 36).

Evidence accumulated over the last 25 years establishes that acute intermittent hypoxia elicits phrenic LTF via a mechanism that requires episodic serotonin (5-HT) release (37), serotonin receptor (type 2) activation (24, 25, 38), ERK MAP kinase activity (39, 40), new synthesis of brain-derived neurotrophic factor (BDNF), and activation of its high affinity receptor, tyrosine kinase B (41). This cellular cascade is referred to as the “Q pathway” to phrenic motor facilitation since Gq proteins are associated with the initiating (5-HT2) metabotropic receptor (42).

Low impact ampakines, such as CX717, have been reported to increase BDNF (neuronal brain-derived neurotrophic factor) expression (43). This observation provides a potential point of convergence between the intermittent hypoxia evoked Q pathway and ampakine treatment. Since BDNF signaling is an essential component of the Q pathway (41), in the current study we hypothesized that pretreatment, before acute intermittent hypoxia, with ampakine CX717 would boost the magnitude of phrenic motor facilitation. Although our results confirmed that CX717 coupled with a single bout of hypoxia evoked motor facilitation [e.g., Fig. 2A (18, 20)], we observed no synergy when ampakine was paired with acute intermittent hypoxia. On the contrary, a greater magnitude of phrenic long-term facilitation was observed when ampakine was combined with only a single episode of hypoxia.

Our study was not designed to evaluate the underlying mechanisms, but a few relevant points can be made in this regard. First, the half-life of CX717 in the blood is estimated to be ∼60 min after intravenous administration in rats (personal communication, Dr. Arnold Lippa, RespireRx). Therefore, CX717 is likely to have been circulating, with potential to interact with AMPA receptors, for the duration of our experiments. However, the “time control” experiments in which CX717 was delivered without any hypoxia exposure show no sustained impact of the drug on phrenic motor output. Rather, the phrenic burst amplitude returned to preampakine baseline values within one hour, and the impact of CX717 was restricted to the period immediately following intravenous delivery. In contrast, coupling CX717 with a short bout of hypoxia resulted in a sustained increase in phrenic bursting. One attractive hypothesis is that the pairing of ampakine with hypoxia was sufficient to evoke the well-established Q pathway mechanisms that can produce phrenic LTF. However, it remains unknown if the enhanced respiratory neuroplasticity after ampakines requires serotonin or any of the other fundamental mechanistic components of the Q pathway. Furthermore, the attenuated phrenic LTF that occurred when ampakine was followed by AIH could indicate that exposure to additional bouts of hypoxia (i.e., beyond one, 5-min episode), may have activated competing mechanisms that diminish the expression of phrenic LTF. For instance, competing adenosinergic and serotonergic mechanisms can constrain phrenic long-term facilitation under certain conditions (44), and it is possible that a similar scenario occurred in our study.

Another potentially important consideration is the duration of hypoxia exposure. Previously, our laboratory observed that when anesthetized, mechanically ventilated mice were pretreated with ampakine CX717, LTF of inspiratory bursting recorded from the hypoglossal nerve was enhanced after three, 1-min hypoxic episodes (19). As a field, we do not know what the optimal intermittent hypoxia exposure pattern is to promote LTF even under “normal” conditions. When AMPA receptors are allosterically modulated by CX717, it is possible that certain AIH paradigms, such as 3 × 1 min hypoxia, may more effectively promote enhanced LTF. There are a number of additional differences between the prior and current study, including species (mouse vs. rat), the measured respiratory output (hypoglossal vs. phrenic), and the route of ampakine delivery (intraperitoneal vs. intravenous). All of the aforementioned variables represent avenues for future studies.

Short-term potentiation (STP) is another form of hypoxia-induced respiratory neuroplasticity (35). It is defined as a progressive increase in phrenic activity during exposure to hypoxia that is followed by a slow return to baseline activity after the hypoxic exposure is terminated. Anecdotal evaluation of our data suggests that pretreatment with CX717 might have altered the expression of hypoxia-induced phrenic STP (e.g., see Figs. 2 and 3). However, no formal comparisons between the data shown in Fig. 2 versus Fig. 3 were made because evaluation of STP was not an a priori goal of our study. Thus, there were unbalanced sample size between these groups, and more importantly these data were collected in separate cohorts, studied a few months apart. Although, the present work was not designed to test this question, future experiments can formally address the possibility that STP of respiratory motor output is impacted by ampakine treatment.

Significance

Extensive preclinical work in rodent models has led to the use of carefully controlled exposure to intermittent hypoxia as a therapy for triggering beneficial neuroplasticity in human subjects (45). The fundamental concept is to use short-duration moderate hypoxia as an adjunctive approach to neurorehabilitation, for example after spinal cord injury (46–48). To our knowledge, all current clinical testing using hypoxia in this context has utilized multiple episodes of hypoxia. Taken together, our current and past results establish that ampakine pretreatment can reduce the required number of hypoxic episodes for evoking phrenic motor facilitation to one. Additional hypoxic episodes did not add further benefit in the current study, although whether this observation holds true after neurologic injury, such as spinal cord injury, will need to be tested. Overall, it is clear that the ampakine-hypoxia interaction can enhance the expression of respiratory neuroplasticity, and more work needs to be done to elucidate the underlying mechanisms.

GRANTS

This work was supported by funding from the National Institute of Health Grants 5 R01 HL139708 02 (to D. D. Fuller) and F31 HL145831-01 (to M. D. Sunshine).

DISCLOSURES

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

AUTHOR CONTRIBUTIONS

P.P.T. and D.D.F. conceived and designed research; P.P.T. performed experiments; P.P.T., M.D.S., and D.D.F. analyzed data; P.P.T. and D.D.F. interpreted results of experiments; P.P.T. prepared figures; P.P.T. drafted manuscript; P.P.T., M.D.S., and D.D.F. edited and revised manuscript; P.P.T., M.D.S., and D.D.F. approved final version of manuscript.