Pattern sensitivity of ampakine-hypoxia interactions for evoking phrenic motor facilitation in anesthetized rat

Abstract

Repeated hypoxic episodes can produce a sustained (>60 min) increase in neural drive to the diaphragm. The requirement of repeated hypoxic episodes (vs. a single episode) to produce phrenic motor facilitation (pMF) can be removed by allosteric modulation of α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid receptors using ampakines. We hypothesized that the ampakine-hypoxia interaction resulting in pMF requires that ampakine dosing precedes the onset of hypoxia. Phrenic nerve recordings were made from urethane-anesthetized, mechanically ventilated, and vagotomized adult male Sprague-Dawley rats during isocapnic conditions. Ampakine CX717 (15 mg/kg iv) was given immediately before (n = 8), during (n = 8), or immediately after (n = 8) a 5-min hypoxic episode (arterial oxygen partial pressure 40–45 mmHg). Ampakine before hypoxia (A_prior) resulted in a sustained increase in inspiratory phrenic burst amplitude (i.e., pMF) reaching +70 ± 21% above baseline (BL) after 60 min. This was considerably greater than corresponding values in the groups receiving ampakine during hypoxia (+28 ± 47% above BL, P = 0.005 vs. A_prior) or after hypoxia (+23 ± 40% above BL, P = 0.005 vs. A_prior). Phrenic inspiratory burst rate, heart rate, and systolic, diastolic, and mean arterial pressure (mmHg) were similar across the three treatment groups (all P > 0.3, treatment effect). We conclude that the presentation order of ampakine and hypoxia impacts the magnitude of pMF, with ampakine pretreatment evoking the strongest response. Ampakine pretreatment may have value in the context of hypoxia-based neurorehabilitation strategies.

NEW & NOTEWORTHY Phrenic motor facilitation (pMF) is evoked after repeated episodes of brief hypoxia. pMF can also be induced when an allosteric modulator of AMPA receptors (ampakine) is intravenously delivered immediately before a single brief hypoxic episode. Here we show that ampakine delivery before hypoxia (vs. during or after hypoxia) evokes the largest pMF with minimal impact on arterial blood pressure and heart rate. Ampakine pretreatment may have value in the context of hypoxia-based neurorehabilitation strategies.

INTRODUCTION

Ampakines are a synthetic class of compounds that allosterically bind to the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA)-type glutamate receptors. Ampakines modulate excitatory glutamatergic synaptic neurotransmission (1, 2) and can act as respiratory stimulants (3). Ampakines most effectively stimulate breathing under conditions when respiratory drive and/or phrenic motor output is impaired, such as opioid overdose (4, 5), Rett syndrome (6), Pompe disease (7), and cervical spinal cord injury (SCI) (8, 9). In contrast, the impact of ampakines on breathing is relatively minor when respiratory drive is not impaired, such that a single dose of ampakine CX717 causes only a transient increase in phrenic nerve efferent motor output that peaks at ∼2 min after infusion and then gradually returns to baseline values (10).

Ampakine treatment can also increase the capacity for neuroplasticity in the respiratory control system (11). A well-studied example of respiratory neuroplasticity is “long-term facilitation” (LTF), in which repeated hypoxia exposures (i.e., acute intermittent hypoxia or AIH) evoke sustained increases in respiratory motor output (12). Treatment with an ampakine creates preconditions that reduce the number of hypoxic episodes required for induction of respiratory neuroplasticity (10, 13). In an anesthetized rat preparation, when intravenous delivery of an ampakine (CX717) is quickly followed by an exposure to a brief hypoxic episode, the result is a sustained increase in phrenic motor output lasting >60 min after the hypoxic episode (i.e., phrenic motor facilitation or pMF). When given separately, however, neither ampakine nor a single episode of brief hypoxia produces pMF (10). Thus, ampakine-triggered mechanisms appear to interact with hypoxia-induced mechanisms to trigger sustained increases in phrenic motoneuron output. This finding holds significance, given the prevailing consensus that recurring brief episodes of hypoxia are required to initiate the spinal molecular mechanisms responsible for triggering phrenic motor facilitation (14). Ampakine and hypoxia interactions may also have therapeutic use. For example, AIH has shown efficacy in clinical trials to improve motor function in persons with SCI (15, 16). Low-dose ampakine treatment may provide a means to boost the impact of AIH or to reduce the number of hypoxic episodes required for functional gains.

The cellular mechanisms underlying the ampakine-hypoxia interaction that leads to pMF are unknown. As a first step toward understanding these mechanisms, we evaluated whether the timing of ampakine delivery influences the expression of pMF. Prior studies show that pretreatment with ampakine can enhance expression of hippocampal long-term potentiation (17, 18). On the basis of these findings, we hypothesized that allosteric modulation of AMPA receptors, via intravenous ampakine delivery, needed to occur before (vs. during or after) hypoxia to produce pMF. For these studies, we used ampakine CX717, which is classified as a “low-impact” ampakine and has an excellent safety profile (19, 20). Our hypothesis was tested with an anesthetized rat preparation that enables rigorous arterial blood gas control and has been used in prior studies of pMF (9, 10, 13, 21). The results indicate that the order of presentation of ampakine and hypoxia impacts the magnitude of respiratory neuroplasticity, with greater pMF observed in the ampakine pretreatment group. Accordingly, low-dose ampakine treatment may provide a tool to enhance the therapeutic impact of hypoxia-based neurorehabilitation strategies (14, 22).

MATERIALS AND METHODS

Animals

All experimental protocols were approved by the Animal Care and Use Committee at the University of Florida (Protocol ID 201807438) and were in accordance with the National Institutes of Health (NIH) Guide for the Care and Use of Laboratory Animals. In total, 24 adult male Sprague-Dawley rats (Envigo, Colony No. 208A) were studied after random distribution to three groups: CX717 + hypoxia (346 ± 9 g, age 12 ± 1 wk, n = 8), hypoxia + CX717 (350 ± 10 g, age 12 ± 1 wk, n = 8), and CX717 during hypoxia (344 ± 7 g, age 12 ± 1 wk, n = 8). Rats had unrestricted access to food and water and were maintained on a 12:12-h light-dark cycle.

Surgical Procedure

For the initial induction of anesthesia, rats were placed in a chamber that was flushed with 3% isoflurane in 100% O₂. Rats were then removed from the chamber and placed on a heated pad, with isoflurane anesthesia maintained with a nose cone. A temperature probe was inserted rectally, and a servo-controller was used to keep temperature at 37 ± 1°C (TC-1000; CWE). A stable surgical plane of anesthesia was confirmed by loss of pedal withdrawal and corneal reflexes. Rats were then tracheotomized and ventilated (VentElite, model 55-7040; Harvard Apparatus Inc.) with a 50% O₂-1% CO₂ mixture (balanced with N₂). End-tidal CO₂ was maintained at 38–40 mmHg throughout the surgery and experimental protocol (Capnogard; Novametrix). Ventilator frequency was maintained between 65 and 75 breaths/min, and tidal volume was set at 7 mL/kg (23). Bilateral vagotomy was performed to prevent entrainment of the phrenic nerve efferent recordings with lung inflation. Tracheal pressure was continuously monitored with a pressure transducer (TA-100; CWE) connected to the tracheal cannula. During the surgical procedures, the lungs were periodically hyperinflated by brief occlusion of the expiratory line (2 respiratory cycles) to minimize alveolar collapse. A final hyperinflation was done immediately before the start of baseline recordings.

To enable intravenous fluid infusions, a catheter was placed in the tail vein. Rats were converted from isoflurane to urethane anesthesia over a period of 30 min (2.1 g/kg, 6 mL/h). The depth of anesthesia was consistently assessed through evaluation of the pedal withdrawal reflex. The absence of any observable withdrawal reflex was employed as an indicator to assess the sufficiency of anesthesia. After the full urethane dose was administered, a complete neuromuscular blockade was accomplished by intravenous delivery of pancuronium bromide (3 mg/kg; Sigma-Aldrich, St. Louis, MO). This was done to eliminate respiratory movements and to enable precise control of arterial blood gases via the mechanical ventilator. Furthermore, to maintain acid-base balance, a 1:4 solution of 8.4% sodium bicarbonate mixed with lactated Ringer solution was continuously infused (2 mL/h) throughout the experiment. The femoral artery was cannulated with polyethylene tubing (PE 50; Intramedic) and connected to a transducer amplifier (TA-100; CWE) to measure arterial blood pressure (ABP). Systolic and diastolic blood pressure were identified from peaks and valleys in the ABP signal. The arterial line enabled withdrawal of blood samples (65 μL) for measurement of partial pressure of CO₂ ($\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$) and O₂ ($\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$), pH, and base excess with a blood gas analyzer (ABL 90 Flex; Radiometer, Copenhagen, Denmark).

With a dorsal surgical approach, the left phrenic nerve was isolated, cut distally, and desheathed. 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 physiological saline solution. Finally, with a differential AC amplifier (model 1700; A-M Systems, Everett, WA) phrenic nerve activity was amplified (10 kHz), band-pass filtered (100 Hz to 3 kHz), digitized [16 bit, 25,000 samples/channel; Power 1401; Cambridge Electronic Design (CED), Cambridge, UK], and integrated (∫) with a time constant of 0.05 s with Spike2 software (CED).

Drugs

Ampakine CX717 in powder form was dissolved in a 10% solution of 2-hydroxypropyl-β-cyclodextrin (HPCD; Sigma) at a concentration of 5 mg/mL. The solution was stored in 1.5-mL aliquots at −20°C. Animals received ampakine CX717 via the tail vein. A fresh aliquot was first thawed and brought to room temperature before administration. The ampakine dose of 15 mg/kg was chosen on the basis of prior results that showed that this dose promotes plasticity in phrenic output after intravenous delivery (9, 10, 13).

Neurophysiology Protocol

The basic experimental paradigm has been described previously (9, 10, 21). Baseline phrenic nerve activity was established while maintaining the inspired fraction of O₂ and CO₂ at 50% and 1%, respectively (balanced with N₂).

Throughout the experiment, the end-tidal CO₂ values were maintained between 38 and 40 mmHg. The experimental protocol was started only after a stable phrenic nerve activity was recorded for a minimum of 15 min. Activity was considered stable when the burst amplitude and rate did not vary by ±5%. In the ampakine pretreatment group, CX717 was delivered 2 min before the onset of hypoxia. In the group receiving ampakine during hypoxia, CX717 was delivered 2 min after hypoxia onset. In the final group, CX717 was administered 2 min after the termination of hypoxia.

We established a priori that an experiment would be excluded from the analysis if the blood gas values could not be maintained within the specified ranges. Arterial blood samples were drawn during baseline, the final (5th) minute of hypoxia, and 20, 40, and 60 min after treatment. The baseline blood gas value was used as reference to determine if subsequent arterial blood samples were isocapnic. Minor adjustments to the inspired CO₂ and/or rate of lung inflation (mechanical ventilation) were made periodically to ensure that end-tidal CO₂ values were maintained between 38 and 40 mmHg and $\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$ values stayed within ±2.0 mmHg from the baseline values. In n = 1 experiment the $\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$ values were considerably below the acceptable range, and that preparation was excluded from formal analysis. $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$ values were maintained above 150 mmHg for all time points except hypoxia. If values fell below this threshold, the inspired fraction of O₂ was increased by 5% and another arterial blood sample was analyzed within 5 min. At the conclusion of the neurophysiology protocol, rats were given a strong chemoreceptor stimulation by briefly ceasing lung inflation induced by temporarily switching off the ventilator. This was done to establish a maximum phrenic discharge that transformed to a “gasping-like” irregular pattern within 20–30 s. The rat was then euthanized via an intravenous bolus of the urethane solution. If the increase in phrenic nerve burst amplitude that occurred in response to the “maximal challenge” was less than the response observed during hypoxia, it was considered an indication of degrading nerve-electrode contact and the preparation was excluded from analysis (n = 2). Whenever the phrenic neurophysiology preparations did not meet the a priori inclusion criteria, no aspect of that recording was included in any formal analysis. Additional experiments were conducted to make up an equal sample size across all three groups (n = 8).

Data and Analyses

All data were recorded with Spike2 software (v. 10.01; CED). Phrenic nerve recordings were analyzed with a custom-written MATLAB code (MathWorks, R2019a) that is available upon request. The goal of the study was to compare the magnitude of change of phrenic and cardiovascular parameters from baseline (pretreatment) levels between the three experimental groups. Accordingly, parameters were normalized to baseline with the following formula: [response − baseline]/baseline × 100. Values are reported in the text as percent change from baseline (Δ%BL). ∫ Phrenic nerve activity, burst rate, heart rate, and blood pressure were averaged over a 3-min period immediately before withdrawal of arterial blood samples during baseline, the final (5th) minute of hypoxia, and 20, 40, and 60 min after treatment.

Statistical analysis was performed with SigmaPlot 14 (Systat Software) and GraphPad Prism (v.9). Comparison of phrenic amplitude, respiratory rate, heart rate, and systolic, diastolic, and mean arterial pressure (MAP) between different groups at multiple points was analyzed with a two-way repeated-measures analysis of variance (2-way RM ANOVA). For analysis of selected time point of interest among different groups, a one-way analysis of variance (1-way ANOVA) was applied. Changes in arterial blood gases were compared to baseline within each group and analyzed by 1-way ANOVA. Wherever required, multiple comparisons were made with a Student–Newman–Keuls post hoc test. Data were deemed statistically significant when P ≤ 0.05. Mean data are presented along with one standard deviation.

RESULTS

Arterial blood gas measurements are presented in Table 1. Baseline parameters were similar between experimental groups. All groups showed the expected reduction in $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$ and blood pressure during acute hypoxia. The $\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$ was slightly greater, and arterial pH slightly lower, during hypoxia in the group that received ampakine immediately after hypoxia (Table 1). No differences in blood gas parameters were observed within experimental groups over the period of 20–60 min following the acute hypoxia exposure, except for $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$. The return of $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$ to baseline (prehypoxia) values was gradual, but $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$ was >200 mmHg by 40 min after hypoxia in all groups (Table 1). The raw (nonnormalized) values for phrenic burst amplitude, respiratory rate, heart rate, and arterial blood pressure are shown in Table 2.

Table 1.

Arterial blood gas measurements

Experimental Group	$\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$, mmHg	$\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$, mmHg	SBEc, mmol/L	pH
CX717 + Hy (n = 8)
Baseline	42.1 ± 1.4	249.1 ± 9.8	−0.01 ± 1	7.38 ± 0.01
Hypoxia	42.4 ± 1.7	44.7 ± 4.8*	−0.61 ± 1.4	7.38 ± 0.02
20 min	43.3 ± 2.6	177.3 ± 44.9*	−0.5 ± 0.9	7.37 ± 0.02
40 min	43 ± 1.6	223 ± 27	−0.2 ± 1	7.38 ± 0.02
60 min	42.4 ± 1.6	239.6 ± 21.1	0.5 ± 1	7.39 ± 0.02
Hy + CX717 (n = 8)
Baseline	42.3 ± 1.2	256.4 ± 16.8	−0.33 ± 1.5	7.38 ± 0.02
Hypoxia	44.5 ± 2.1*	43.7 ± 9.1*	−1.44 ± 1.5	7.35 ± 0.02*
20 min	42.9 ± 1.8	148.3 ± 17.2*	−0.8 ± 2.1	7.37 ± 0.03
40 min	42.7 ± 0.8	201.4 ± 15.6*	−0.34 ± 1.8	7.38 ± 0.02
60 min	42.1 ± 1.4	224.4 ± 18*	0.11 ± 1.8	7.39 ± 0.02
CX717 during Hy (n = 8)
Baseline	41.4 ± 1	251.1 ± 9.2	0.3 ± 1.6	7.39 ± 0.02
Hypoxia	43.4 ± 3	41.4 ± 5.7*	−0.4 ± 1.8	7.37 ± 0.03
20 min	42 ± 2.3	159.9 ± 22.6*	−0.6 ± 1.1	7.38 ± 0.03
40 min	42 ± 2.3	216.9 ± 19.8*	−0.03 ± 1.1	7.36 ± 0.07
60 min	41 ± 2	232.1 ± 19.6	0.2 ± 1.4	7.4 ± 0.03

Data are presented as means ± SD for n = 8 rats in each group. Mean blood gas values are shown for baseline, hypoxia, and the 20, 40, and 60 min posttreatment time points. Hy, hypoxia; $\mathrm{P}{\mathrm{a}}_{\mathrm{C}{\mathrm{O}}_2}$, arterial CO₂ pressure; $\mathrm{P}{\mathrm{a}}_{{\mathrm{O}}_2}$, arterial O₂ pressure; SBE_c, standard base excess. *Significant difference compared with baseline (P < 0.05) within that particular group.

Table 2.

Nonnormalized inspiratory phrenic nerve amplitude, respiratory rate, heart rate, and arterial blood pressure values in different treatment groups

Experimental Group	Phrenic Amplitude, V	Respiratory Rate, min−1	Heart Rate, min−1	MAP, mmHg	Systolic BP, mmHg	Diastolic BP, mmHg
CX717 + Hy (n = 8)
Baseline	0.057 ± 0.019	43.87 ± 4.51	410.0 ± 19.7	124.8 ± 12	176.8 ± 14.4	98.8 ± 11
Hypoxia	0.137 ± 0.056	55.90 ± 4.39	443.4 ± 14.8	110.7 ± 17.5	172 ± 23.8	80.3 ± 15
20 min	0.087 ± 0.042	45.99 ± 4.0	428.8 ± 15.7	126.1 ± 13.3	185 ± 16	97.0 ± 13
40 min	0.097 ± 0.036	45.37 ± 4.59	425.3 ± 17.2	127.4 ± 15.5	190 ± 18.8	96.1 ± 14.4
60 min	0.096 ± 0.035	46.22 ± 5.23	421.0 ± 15.9	132.2 ± 14.1	198 ± 15.5	99.5 ± 14
Hy + CX717 (n = 8)
Baseline	0.072 ± 0.044	47.64 ± 7.28	425.9 ± 22.2	129.6 ± 14.2	181.5 ± 16.6	103.7 ± 14.4
Hypoxia	0.129 ± 0.074	62.59 ± 5.14	453.1 ± 20.5	108.1 ± 22.2	160.9 ± 28.2	81.6 ± 20.7
20 min	0.088 ± 0.051	47.87 ± 5.09	447.6 ± 18.4	131.2 ± 16.5	185.4 ± 20.2	104.1 ± 15.2
40 min	0.088 ± 0.057	46.61 ± 7.45	447.3 ± 20.9	132.9 ± 12.8	188.8 ± 12.7	105 ± 13.4
60 min	0.083 ± 0.052	44.46 ± 7.73	440.5 ± 14.2	136.1 ± 12.2	193.6 ± 11.8	107.4 ± 13
CX717 during Hy (n = 8)
Baseline	0.088 ± 0.029	44.01 ± 3.0	427.5 ± 41.2	123.2 ± 15.6	181.5 ± 16.6	103.7 ± 14.4
Hypoxia	0.179 ± 0.051	62.53 ± 7.75	438.2 ± 32.9	98.5 ± 20.7	161 ± 28.2	81.7 ± 20.7
20 min	0.115 ± 0.031	44.80 ± 2.44	447.0 ± 26.3	123.5 ± 13.0	185.4 ± 20.2	104.1 ± 15.2
40 min	0.097 ± 0.028	41.99 ± 5.33	440.3 ± 21.0	127.2 ± 11.5	188.8 ± 12.7	105 ± 13.4
60 min	0.111 ± 0.042	43.19 ± 4.42	434.5 ± 24.6	127.1 ± 13.8	193.6 ± 11.8	107.4 ± 13

Data are presented as means ± SD for n = 8 rats in each group. Mean phrenic burst amplitude, respiratory rate, heart rate and arterial blood pressure (BP) values are shown for baseline, hypoxia, and the 20, 40, and 60 min posttreatment time points. Hy, hypoxia; MAP, mean arterial pressure.

Ampakine-Hypoxia: Pattern Impacts the Expression of pMF

Compressed data records providing examples of complete experimental protocols are shown in Fig. 1. These examples illustrate that pretreatment with ampakine CX717 before the onset of hypoxia resulted in the greatest magnitude of pMF. Phrenic burst amplitude showed a treatment × time interaction over the course of the experimental protocol (Fig. 2A; F_6,95 = 3.2, P = 0.009). When ampakine was delivered before hypoxia, the amplitude of the inspiratory phrenic burst was +70 ± 21% above BL after 60 min (P < 0.001 vs. BL). The other two groups also showed a sustained increase in phrenic burst amplitude, but of lesser magnitude: ampakine during hypoxia: +28 ± 47% above BL at 60 min (P = 0.047 vs. BL); ampakine after hypoxia: +23 ± 40% above BL at 60 min (P = 0.056 vs. BL). Phrenic burst amplitude values in the ampakine pretreatment group were greater than comparable values in the other two experimental groups at 40 and 60 min after hypoxia (all P ≤ 0.005). Comparison of phrenic burst amplitude at 60 min after hypoxia, showing all data points, is provided in Fig. 2Ai.

Figure 1.

Representative phrenic neurograms depicting phrenic motor facilitation (pMF). A: CX717 followed by a single 5-min episode of hypoxia (CX717 + hypoxia). B: single 5-min episode of hypoxia followed by CX717 (hypoxia + CX717). C: CX717 delivered during an ongoing episode of hypoxia (CX717 during hypoxia). For all groups, compressed records of phrenic motor output (top), respiratory rate (middle), and arterial blood pressure (bottom) are shown at baseline, during treatment (hypoxia/ampakine), and for ∼60 min after treatment. At the conclusion of the experiment, a brief maximum chemoreceptor challenge was induced (see materials and methods). ∫ Phrenic, integrated phrenic nerve activity.

Figure 2.

Phrenic motor facilitation (pMF) is greatest when ampakine treatment precedes hypoxia. A: integrated (∫) phrenic burst amplitude at baseline (BL) and 20, 40, and 60 min after -treatment. Ai: scatterplot depicting the individual data points at 60 min. Phrenic burst amplitude in CX717 + hypoxia group was greater at 40 and 60 min compared with hypoxia + CX717 and CX717 during hypoxia groups. Further evaluation of a specific time point of interest, i.e. 60 min after treatment, showed that pMF was higher in CX717 + hypoxia group compared to hypoxia + CX717 and CX717 during hypoxia groups. B and Bi: mean changes in respiratory rate over the entire duration of protocol and individual data points at 60 min alone. C and Ci: heart rate changes over time and at 60 min. D and Di: changes in mean arterial pressure (MAP) over time and at 60 min. E and Ei: systolic blood pressure (BP) changes over time and at 60 min. F and Fi: diastolic blood pressure changes over time and at 60 min. *CX717 + hypoxia statistically different than other 2 groups at 40 and 60 min. #60 min time point greater than BL. ^Statistically different from CX717 + hypoxia. Data represented as means ± 1 SD; n = 8 rats per group. Statistical difference symbols denote P < 0.05. All data are represented as % change from baseline (Δ%BL).

The phrenic inspiratory burst rate was similar over the duration of the posthypoxia recordings in all three treatment groups (Treatment: F_2,95 = 1.2, P = 0.308) (Fig. 2B). Burst rate was stable over the 60-min recording duration, with no overall effect of time (F_3,95 = 2.0, P = 0.12). A direct comparison of respiratory rate between the three experimental groups at 60 min after hypoxia is shown in Fig. 2Bi (P = 0.193).

Heart rate (Fig. 2, C and Ci) was similar between the groups over the duration of the experiment (Treatment: F_2,95 = 0.09, P = 0.911). There was an impact of time on heart rate (F_3,95 = 12.8, P < 0.001). Specifically, all groups exhibited an increased heart rate 20 min after treatment, with the elevation persisting for up to 40 min in the group that received ampakine before hypoxia and up to 60 min in the group where ampakine followed hypoxia. Figure 2Ci shows a direct comparison of heart rate at 60 min for all groups (P = 0.859). Mean arterial pressure (Fig. 2D) showed no effect of treatment (F_2,95 = 0.001, P = 0.999), but an impact of time was observed (F_3,95 = 5.7, P = 0.001). This manifested as a slight increase in the mean arterial pressure as the experiment progressed, and the magnitude of this was similar across the three experimental groups. A direct comparison of mean arterial pressure between the groups at 60 min after hypoxia is shown in Fig. 2Di (P = 0.886). The systolic blood pressure (Fig. 2, E and Ei) also showed no effect of treatment (F_2,95 = 0.99, P = 0.388), but there was an overall increase over the duration of the recordings (Time: F_3,95 = 15.8, P < 0.001). In particular, groups receiving ampakine before and during hypoxia exhibited an increased systolic blood pressure from 40 min, persisting up to 60 min. In contrast, the group that received ampakine after hypoxia only showed an elevated systolic blood pressure at 60 min. Figure 2Ei shows a direct comparison of systolic blood pressure at 60 min for all groups (P = 0.316). Diastolic blood pressure was stable over the duration of the experiments, with no effect of treatment (F_2,95 = 0.44, P = 0.646) or time (F_3,95 = 1, P = 0.37) (Fig. 2, F and Fi).

DISCUSSION

Ampakine compounds are divided into two general classes: type I (“high impact”) and type II (“low impact”) [reviewed in Arai and Kessler (19)]. The receptor kinetics of low-impact ampakines differ from high impact by having shorter decay time constant (19). Low-impact ampakines such as the CX717 compound used in the present study act as a transient respiratory stimulant, but this effect rapidly wanes (10, 13). However, when CX717 is administered before a single brief episode of moderate hypoxia, a long-lasting increase in efferent phrenic output is observed that can last more than an hour, i.e., pMF (10, 13). In contrast, an acute episode of hypoxia alone does not produce long-term changes in respiratory output. Here, we show that the order of presentation of CX717 and hypoxia impacts the magnitude of pMF in the anesthetized rat. Delivery of ampakine CX717 before an acute hypoxic exposure resulted in the most robust expression of pMF (compared with CX717 delivery during or after hypoxia). Thus, the respiratory neuroplasticity mechanisms triggered by the ampakine-hypoxia combination are sensitive to the order in which these two stimuli are presented.

AMPA Receptors and Neuroplasticity

The rapid channel kinetics of AMPA receptors enable the fast neuronal depolarization that underlies excitatory glutamatergic synaptic transmission (24). It is also firmly established that AMPA receptor activation can evoke mechanisms leading to synaptic plasticity (25). One trigger of these mechanisms is influx of Ca²⁺ ions via AMPA receptors lacking GluR2 subunit. This N-methyl-d-aspartate (NMDA) receptor-independent response is an important initiator of plasticity at glutamatergic synapses (26). AMPA receptors can also impact synaptic plasticity via mechanisms that do require activation of NMDA receptors. The classic example is the removal of the magnesium block of NMDA receptors following AMPA receptor activation, which allows further entry of Na⁺ and Ca²⁺ and can lead to postsynaptic long-term potentiation (27).

The important role of AMPA receptors in synaptic plasticity has led to studies of how ampakines impact the expression of neuroplasticity. Several prior reports describe how ampakine treatment can enhance and restore the magnitude of long-term potentiation in hippocampal neural pathways (17, 18, 28). That literature was the rationale for our initial investigations of how ampakine pretreatment could impact hypoxia-induced respiratory neuroplasticity (11). We first reported that ampakine CX717 pretreatment could substantially increase the magnitude of acute intermittent hypoxia-induced long-term facilitation of hypoglossal inspiratory activity in the mouse (29). Subsequent work, including the present study, has established that ampakine pretreatment enables a single hypoxic episode to evoke neuroplasticity in the phrenic motor system, demonstrated by the presence of pMF (9, 10, 13).

pMF Following Ampakine + Hypoxia: Potential Mechanism of Action

The molecular and synaptic mechanisms that enable the combination of ampakine and hypoxia to produce pMF are unknown. However, by considering the known actions of ampakine and hypoxia individually, we can generate evidence-based hypotheses to explain how the pairing of ampakine and hypoxia leads to pMF.

In the present study and prior reports of ampakine-hypoxia-induced pMF (10, 13), ampakines were delivered intravenously. Given the systemic distribution of ampakine, we cannot draw firm conclusions regarding the location of the relevant neurons affected by the ampakine. However, based on prior studies of ampakines conducted in vivo (8, 21) and in vitro (30), it likely that both brain stem and spinal respiratory neurons were impacted. The acute increase in respiratory rate upon intravenous ampakine delivery reported here and previously (10) strongly implicates brain stem neuronal activation. A spinal impact is also likely since phrenic motoneurons express AMPA receptors (31, 32), and activation of these receptors contributes to the inspiratory burst of phrenic activity (33). Furthermore, direct application of ampakine CX717 to the cervical spinal cord leads to an increase in phrenic nerve inspiratory burst amplitude but with no impact on the burst rate (21). Thus, modulation of spinal AMPA receptors with ampakines is likely to increase phrenic motoneuron excitability. This suggestion is consistent with in vitro studies of hypoglossal motoneurons showing that ampakines cause a direct change in membrane currents that slows the AMPA channel decay kinetics (30).

Given the known actions of ampakines and hypoxia, signaling related to brain-derived neurotrophic factor (BDNF) and the tropomyosin receptor kinase B (TrkB) could potentially be a point of convergence leading to phrenic motor plasticity. For example, several studies show that ampakine exposure can increase neuronal production of BDNF (34–36). Furthermore, intermittent hypoxia exposure can produce a BDNF-dependent pMF that requires activation of spinal TrkB receptors (37). At this time, our working hypothesis is that pretreatment with ampakines increases the overall excitability of phrenic motoneurons and thereby enables a single episode of hypoxia to produce pMF, via the well described “Q-pathway” that can be activated by acute intermittent hypoxia (38). Direct actions of ampakines on respiratory neurons, and in particular phrenic motoneurons, could act to “precondition” these cells such that the aforementioned mechanisms driving phrenic motor plasticity are activated after a single, brief hypoxia exposure. Another possibility, however, is that ampakines impact the function of the carotid body. No studies to our knowledge have examined the direct impact of ampakines on the carotid body, but AMPA-mediated glutamatergic signaling does occur in carotid bodies (39).

Significance and Future Directions

The present data add to the body of work indicating that ampakines can enhance expression of neuroplasticity (17, 18, 28, 36, 40, 41) and in particular neuroplasticity in the neural pathways that control the diaphragm muscle (10, 13, 29). On the basis of data presented here, we conclude that the impact on respiratory neuroplasticity (in this case, pMF) is greatest when ampakines are provided shortly before exposure to an acute episode of hypoxia. We suggest that this is interesting from a mechanistic standpoint, and future studies can explore the cellular pathways that lead to pMF when ampakines are provided as a preconditioning stimulus, before hypoxia. We speculate that BDNF may be a point of convergence between ampakine- and hypoxia-induced signaling.

Our results may also have implications for the field of neurorehabilitation, particularly after spinal cord injury (14). Building upon an initial report showing that intermittent hypoxia can induce changes in synaptic efficacy in the injured spinal cord (42), many studies have now shown that acute intermittent hypoxia can drive neuroplasticity and, importantly, can improve motor function in animal models of SCI (43–46). Furthermore, studies from multiple laboratories confirm that acute exposure to intermittent hypoxia can improve motor function in persons with spinal injury (15, 47–50). Thus, it appears that intermittent hypoxia-triggered mechanisms of spinal neuroplasticity that are well described in animal models can also be evoked in humans, leading to increased ability to activate motor pathways after spinal cord injury (14). The present data suggest that there may be a role for low-dose, low-impact ampakine treatment in hypoxia-based neurorehabilitation strategies.

In the context of neurorehabilitation, we suggest that providing low-dose, low-impact ampakine treatment as a “preconditioning stimulus,” before exposure to mild hypoxia, could boost the impact of hypoxia and/or reduce the number of hypoxic episodes required for benefit. Indeed, a sizable number of clinical trial participants (∼40%) are classified as minimal or low responders to therapeutic acute intermittent hypoxia (22). Ampakine pretreatment of the low-responder population may represent a novel approach to boosting the impact of hypoxia-based therapies, particularly in individuals who do not respond to standard acute intermittent hypoxia protocols. To have value in a clinical setting, however, ampakines must have minimal off-target effects. After spinal cord injury, cardiovascular and autonomic dysfunction are paramount concerns (51). In the present experiments, and prior reports in the rat model (9, 10, 13), we have observed minimal impact of ampakine and/or hypoxia on cardiovascular parameters. Nevertheless, further work is needed to confirm that the pairing of ampakine and hypoxia does not impact cardiovascular function, particularly when administered without concomitant anesthesia.

DATA AVAILABILITY

Data will be made available upon reasonable request.

GRANTS

This work was supported by funding from the National Institutes of Health: 5 R01 HL139708 02 (D.D.F.).

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. 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. and D.D.F. edited and revised manuscript; P.P.T. and D.D.F. approved final version of manuscript.